Drug Discovery and Development

Senior Content Strategist: Pauline Graham

Content Development Specialist: Carole McMurray
Senior Project Manager: Beula Christopher
Designer: Miles Hitchen
Illustration Manager: Jennifer Rose
Drug Discovery and
Development
Technology in Transition

Edited by

RG Hill BPharm PhD DSc (Hon) FMedSci

President Emeritus, British Pharmacological Society; Visiting Professor of Pharmacology, Department
of Medicine, Imperial College, London; Formerly, Executive Director and Head, Licensing and External
Research, Europe, MSD/Merck Research Laboratories

HP Rang MB BS MA DPhil FMedSci FRS

President-Elect British Pharmacological Society, Emeritus Professor of Pharmacology, University
College, London; Formerly Director, Novartis Institute for Medical Sciences, London

Foreword by
Patrick Vallance BSc MBBS FRCP FMedSci
President, Pharmaceuticals Research and Development, GlaxoSmithKline, Brentford, UK

Edinburgh  London  New York  Oxford  Philadelphia  St Louis  Sydney  Toronto  2013
© 2013 Elsevier Ltd. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and retrieval system, without permission in writing
from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies
and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency,
can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than
as may be noted herein).

First edition 2006

Second edition 2013

ISBN 978-0-7020-4299-7

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging in Publication Data

A catalog record for this book is available from the Library of Congress

Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our under-
standing, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any
information, methods, compounds, or experiments described herein. In using such information or methods they
should be mindful of their own safety and the safety of others, including parties for whom they have a professional
responsibility.

With respect to any drug or pharmaceutical products identified, readers are advised to check the most current informa-
tion provided (i) on procedures featured or (ii) by the manufacturer of each product to be administered, to verify the
recommended dose or formula, the method and duration of administration, and contraindications. It is the responsibility
of practitioners, relying on their own experience and knowledge of their patients, to make diagnoses, to determine
dosages and the best treatment for each individual patient, and to take all appropriate safety precautions.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for
any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any
use or operation of any methods, products, instructions, or ideas contained in the material herein.

The
Working together to grow publisher’s
policy is to use
libraries in developing countries paper manufactured
from sustainable forests
www.elsevier.com | www.bookaid.org | www.sabre.org

Printed in China
 Foreword

Foreword

Medicines have done more to alter the practice of medicine and improve the outlook
for patients than any other intervention. HIV infection has been transformed from
a death sentence to a chronic managed disorder, many childhood leukaemias can be
cured, surgery for peptic ulcers is a thing of the past, treatment of high blood pres-
sure and the introduction of statins has dramatically reduced cardiovascular disease
burden. Despite these and many other examples, the pharmaceutical industry has
been in a crisis. The number of new medicines approved each year has not kept pace
with either the rate of scientific discovery or the huge investment made in R&D. There
appears to be a discrepancy between the ability to make biological discoveries and
ability to translate these into meaningful inventions – the new medicines that will
improve the health of this and future generations. There have been gloom and doom
headlines for the past decade or more and an erosion of trust in the industry. Many
will argue that the low hanging fruit has been picked and that we are now in a period
where we are left with the more difficult problems to solve. I do not accept this
argument.
This new edition comes at a time when I believe the tide has turned and that the
opportunity is in fact greater than it has ever been. Witness the number of new
oncology medicines making it through the pipeline of discovery and development.
Here we have new treatments, some of which have a dramatic effect and many of
which are true exemplars of what we mean by stratified treatments targeted towards
specific patient populations or specific subsets of tumours. Look also at the rise in
antibody treatments and the impact being made in auto-immune and inflammatory
disorders. Many pipelines are beginning to fill up and with a different type of medi-
cine. Medicines that aren’t just small molecules but include antibodies, nucleic acid
based treatments and even cell-based therapies. Medicines that do not target all-
comers across an entire broad disease but rather aim to treat where the mechanism
can have the biggest impact, whether that be in a smaller disease market or in seg-
ments of one or more larger disease area. These changes are not surprising and reflect
the evolution of understanding as traditional definitions of disease based on clinical
observation become more refined by modern biology. There is, I believe a conver-
gence of interests in these changes. Scientists want to make a medicine that has a big
impact on a specific disease area. This of course is also what patients and doctors
want. It is also what underlies the process of health technology assessment, health
economics and the various forms of reimbursement and market access panels that
have emerged across the globe. We will see many more medicines emerging over the
next few decades and many will have bigger effects in smaller populations.
This book tackles these issues head on and in an accessible and authoritative
manner. It is written by experts and has something for everyone, from the beginner
to the seasoned professional. There is one common theme running through the book

v
 Foreword

that I hope will be heeded. Drug discovery and development is about judgement and
about integration of disciplines. There are still too many unknowns for this to be a
routine process. Making a medicine requires that the disciplines from genetics, chem-
istry, biology, protein engineering, pharmacology, experimental medicine, clinical
medicine, modelling and simulation, epidemiology and many others come together
in an integrated way and that bold decisions are made on the basis of what is inevi-
tably incomplete information. I think this book will help those aiming to be part of
this exciting and challenging mission.
Patrick Vallance BSc MBBS FRCP FMedSci

vi
 Foreword

Preface
to 2nd Edition

The intention that we set out with when starting the task of editing this new edition
was to bring the book fully up to date but to try to keep the style and ethos of the
original. All of the chapters have been updated and many have been extensively
revised or completely rewritten by new authors. There are two completely novel
topics given chapters of their own (Protein Scaffolds as Antibody Replacements;
Imaging in Clinical Drug Development) as these topics were judged to be of increas-
ing importance for the future. There are a greater number of authors contributing to
this new edition and an attempt has been made to recruit some of our new authors
from the traditional stronghold of our subject in the large pharmaceutical companies
but also to have contributions from individuals who work in biotechnology compa-
nies, academic institutions and CROs. These segments of the industry are of increas-
ing importance in sustaining the flow of new drugs yet have a different culture and
different needs to those of the bigger companies. We are living in turbulent times
and the traditional models of drug discovery and development are being questioned.
Rather than the leadership position in research being automatically held by tradi-
tional large pharmaceutical companies we now have the creative mixture of a small
number of very large companies (often with a focus on outsourcing a large part of
their research activities), a vigorous and growing biotechnology / small specialty
pharmaceutical company sector and a significant number of academic institutions
engaged in drug discovery. It is interesting to speculate on how the face of drug dis-
covery may look 5 years from now but it will certainly be different with movement
of significant research investment to growth economies such as China and a reduc-
tion in such investment being made in the US and Western Europe. We hope that
this new edition continues to fill the need for a general guide and primer to drug
discovery and development and has moved with the times to reflect the changing
face of our industry.
RG Hill
HP Rang

vii
 Preface
to 1st edition

A large pharmaceutical company is an exceptionally complex organization. Several

features stand out. First, like any company, it must make a profit, and handle its
finances efficiently in order to survive. Modern pharmaceutical companies are so
large that they are financially comparable to a small nation, and the average cost of
bringing a new drug to market – about $800m – is a sum that any government would
think twice about committing.
Second, there is an underlying altruism in the mission of a pharmaceutical company,
in the sense that its aim is to provide therapies that meet a significant medical
need, and thereby relieve human suffering. Though cynics will point to examples
where the profit motive seems to have prevailed over ethical and altruistic concerns,
the fact remains that the modern pharmacopeia has enormous power to alleviate
disease, and owes its existence almost entirely to the work of the pharmaceutical
industry.
Third, the industry is research-based to an unusual extent. Biomedical science has
arguably advanced more rapidly than any other domain of science in the last few
decades, and new discoveries naturally create expectations that they will lead on
improved therapies. Though discoveries in other fields may have profound implica-
tions for our understanding of the natural world, their relevance for improving the
human condition is generally much less direct. For this reason, the pharmaceutical
industry has to stay abreast of leading-edge scientific progress to a greater extent than
most industries.
Finally, the products of the pharmaceutical industry have considerable social
impact, producing benefits in terms of life expectancy and relief of disability, risks
of adverse effects, changes in lifestyle – for example, the contraceptive pill – and
financial pressures which affect healthcare policy on a national and international
scale. In consequence, an elaborate, and constantly changing, regulatory system exists
to control the approval of new drugs, and companies need to devote considerable
resources to negotiating this tricky interface.
This book provides an introduction to the way a pharmaceutical company goes
about its business of discovering and developing new drugs. The first part gives a
brief historical account of the evolution of the industry from its origins in the medi-
aeval apothecaries’ trade, and discusses the changing understanding of what we mean
by disease, and what therapy aims to achieve, as well as summarizing case histories
of the discovery and development of some important drugs.
The second part focuses on the science and technology involved in the discovery
process, that is the stages by which a promising new chemical entity is identified,
from the starting point of a medical need and an idea for addressing it. A chapter
on biopharmaceuticals, whose discovery and development tend to follow routes
somewhat different from synthetic compounds, is included here, as well as accounts

viii
 Preface to 1st edition

of patent issues that arise in the discovery phase, and a chapter on research manage-
ment in this environment. Managing drug discovery scientists can be likened in some
ways to managing a team of huskies on a polar journey. Huskies provide the essential
driving force, but are somewhat wayward and unpredictable creatures, prone to fight-
ing each other, occasionally running off on their own, and inclined to bite the expe-
dition leader. Success in husky terms means gaining the respect of other huskies, not
of humans. (We must not, of course, push the analogy too far. Scientists, unlike
huskies, do care about reaching the goal, and the project management plan – in my
experience at least – does not involve killing them to feed their colleagues.)
The third section of the book deals with drug development, that is the work that
has to be undertaken to turn the drug candidate that emerges from the discovery
process into a product on the market, and a final chapter presents some facts and
figures about the way the whole process operates in practice.
No small group on its own can produce a drug, and throughout the book there is
strong emphasis on the need for interdisciplinary team-work. The main reason for
writing it was to help individual specialists to understand better the work of col-
leagues who address different aspects of the problem. The incentive came from my
own experience when, after a career in academic pharmacology, I joined Sandoz as
a research director in 1985, motivated by a wish to see whether my knowledge of
pharmacology could be put to use in developing useful new medicines. It was a world
startlingly different from what I was used to, full of people – mostly very friendly
and forthcoming – whose work I really did not understand. Even the research labo-
ratories worked in a different way. Enjoyable though it was to explore this new ter-
ritory, and to come to understand the language and preoccupations of colleagues in
other disciplines, it would have been a lot quicker and more painless had I been able
to read a book about it first! No such book existed, nor has any appeared since.
Hence this book, which is aimed not only at scientists who want to understand better
the broad range of activities involved in producing a new drug, but also non-scientists
who want to understand the realities of drug discovery research. Inevitably, in cover-
ing such a broad range, the treatment has to be superficial, concentrating on general
principles rather than technical details, but further reading is suggested for those
seeking more detail. I am much indebted to my many friends and colleagues, espe-
cially to those who have taken the time to write chapters, but also to those with
whom I have worked over the years and who taught me many valuable lessons.
It is hoped that those seeking a general guide to pharmaceutical R & D will find
the book helpful.
H P Rang

ix
 Contributors

Peter Ballard, PhD Åsa Holmgren, MSc (Pharm)

AstraZeneca R&D, Alderley Park, Cheshire, UK Senior Vice President Regulatory Affairs, Orexo AB, Uppsala,
Sweden
Julian Bertschinger-Ehrler, PhD
CEO, Covagen AG, Zurich-Schlieren, Switzerland Charlotte Keywood, MBBS, MRCP, FFPM
Chief Medical Officer, Addex Pharma SA, Plan les Ouates,
Paul Beswick, BSc, PhD, FRSC Switzerland
Head, Department of Medicinal Chemistry, Almirall S.A.,
Barcelona, Spain Vincent Lawton, PhD
Chairman, Aqix Limited, London, UK
Patrick Brassil, PhD
Theravance Inc, South San Francisco, California, USA Harry LeVine, III, PhD
Associate Professor, Center on Aging, Center for Structural
Susanne Bredenberg, BSc (Pharm), PhD Biology, Department of Molecular and Cellular Biochemistry,
Pharmaceutical Science Director, Orexo AB, Uppsala, Sweden University of Kentucky, Lexington, Kentucky, USA

Khanh H Bui, PhD Thomas Lundqvist, MSc (Pharm)

AstraZeneca R&D, Wilmington, Delaware, USA EVP, Orexo AB, Uppsala, Sweden

David Cronk, CBiol MSB Paul M Matthews, MA, MD, DPhil, FRCP
Senior Director, Biology, BioFocus, Saffron Walden, Essex, UK Head, Division of Brain Sciences and Professor, Imperial
College; Vice President, GlaxoSmithKline Research and
Hugues Dolgos, PhD Development, Ltd., London, UK
Global head of DMPK, Merck-Serono R&D, Geneva,
Switzerland Alan Naylor, BSc, PhD, FRSC
Independent Consultant, Alan Naylor Consultancy Ltd.,
Dragan Grabulovski, PhD Royston, UK
CSO, Covagen AG, Zurich-Schlieren, Switzerland
Robert McBurney, PhD
Philip Grubb, MA, BSc, DPhil Chief Executive Officer, Accelerated Cure Project for Multiple
Chartered Patent Attorney (UK), European Patent Attorney, Sclerosis, Waltham, Massachusetts, USA
Formerly Intellectual Property Counsel, Novartis International,
Basel, Switzerland Carl Petersson, PhD
Principal Scientist, Medicinal Chemistry, CNSP iMed Science
Inger Hägglöf, MSc (Pharm) Södertälje, AstraZeneca Research and Development,
Regulatory Affairs Manager, AstraZeneca AB, Södertälje, Innovative Medicines, Södertälje, Sweden
Sweden
Humphrey P Rang, MB BS, MA, DPhil, FMedSci, FRS
Raymond G Hill, BPharm, PhD, DSc (Hon), FMedSci President-Elect British Pharmacological Society, Emeritus
President Emeritus, British Pharmacological Society; Visiting Professor of Pharmacology, University College, London;
Professor of Pharmacology, Department of Medicine, Imperial Formerly Director, Novartis Institute for Medical Sciences,
College, London; Formerly, Executive Director and Head, London
Licensing and External Research, Europe, MSD/Merck
Research Laboratories

x
 Contributors

Barry Robson, BSc Hons, PhD, DSc Anders Tunek, PhD

University Director of Research and Professor of Biostatistics, Senior Principal Scientist, AstraZeneca, Mölndal, Sweden
Epidemiology and Evidence Based Medicine, St. Matthew’s
University Grand Cayman, Cayman Islands; Distinguished Peter J H Webborn, PhD
Scientist, Mathematics and Computer Science, University of AstraZeneca R&D, Macclesfield, UK
Wisconsin-Stout, Wisconsin, USA; Chief Scientific Officer,
Quantal Semantics Inc., Raleigh, North Carolina, USA; Chief
Executive Officer, The Dirac Foundation, Oxfordshire, UK

xi
 Chapter 1 
The development of the pharmaceutical industry
H P Rang

discovered, through empiricism and common sense, many

ANTECEDENTS AND ORIGINS plant extracts whose pharmacological properties we recog-
nize and still use today; their active principles include
Our task in this book is to give an account of the principles opium alkaloids, ephedrine, emetine, cannabis, senna and
underlying drug discovery as it happens today, and to many others1.
provide pointers to the future. The present situation, of In contrast to the ancient Egyptians, who would, one
course, represents merely the current frame of a long- feels, have been completely unsympathetic to medical
running movie. To understand the significance of the dif- science had they been time-warped into the 21st century,
ferent elements that appear in the frame, and to predict the ancient Greeks might have felt much more at home in
what is likely to change in the next few frames, we need the present era. They sought to understand nature, work
to know something about what has gone before. In this out its rules and apply them to alleviate disease, just as
chapter we give a brief and selective account of some of we aim to do today. The Hippocratic tradition had little
the events and trends that have shaped the pharmaceutical time for theistic explanations. However, the Greeks were
industry. Most of the action in our metaphorical movie not experimenters, and so the basis of Greek medicine
happened in the last century, despite the film having remained essentially theoretical. Their theories were philo-
started at the birth of civilization, some 10 000 years ago. sophical constructs, whose perceived validity rested on
The next decade or two will certainly see at least as much their elegance and logical consistency; the idea of testing
change as the past century. theory by experiment came much later, and this aspect of
Many excellent and extensive histories of medicine present-day science would have found no resonance in
and the pharmaceutical industry have been published, ancient Greece. The basic concept of four humours – black
to which readers seeking more detailed information are bile, yellow bile, blood and phlegm – proved, with the
referred (Mann, 1984; Sneader, 1985; Weatherall, 1990; help of Greek reasoning, to be an extremely versatile
Porter, 1997; see also Drews, 2000, 2003). framework for explaining health and disease. Given the
Disease has been recognized as an enemy of human­kind right starting point – cells, molecules and tissues instead
since civilization began, and plagues of infectious diseases of humours – they would quickly have come to terms with
arrived as soon as humans began to congregate in settle- modern medicine. From a therapeutic perspective, Greek
ments about 5000 years ago. Early writings on papyrus medicine placed rather little emphasis on herbal remedies;
and clay tablets describe many kinds of disease, and list a they incorporated earlier teachings on the subject, but
wide variety of herbal and other remedies used to treat made few advances of their own. The Greek traditions
them. The earliest such document, the famous Ebers formed the basis of the prolific writings of Galen in the
papyrus, dating from around 1550BC, describes more 2nd century AD, whose influence dominated the practice
than 800 such remedies. Disease was in those times of medicine in Europe well into the Renaissance. Other
regarded as an affliction sent by the gods; consequently,
the remedies were aimed partly at neutralizing or purging 1
There were, it should be added, far more – such as extracts of asses’
the affliction, and partly at appeasing the deities. Despite testicles, bats’ eyes and crocodile dung – that never found their way
its essentially theistic basis, early medicine nevertheless into modern pharmacology.

© 2012 Elsevier Ltd. 3

Section | 1 | Introduction and Background

civilizations, notably Indian, Arabic and Chinese, simi- We can obtain an idea of the state of therapeutics at the
larly developed their own medical traditions, which – time from the first edition of the British Pharmacopoeia,
unlike those of the Greeks – still flourish independently published in 1864, which lists 311 preparations. Of these,
of the Western ones. 187 were plant-derived materials, only nine of which were
Despite the emphasis on herbal remedies in these early purified substances. Most of the plant products – lemon
medical concepts, and growing scientific interest in their juice, rose hips, yeast, etc. – lacked any components we
use as medicines from the 18th century onwards, it was would now regard as therapeutically relevant, but some
only in the mid-19th century that chemistry and biology – digitalis, castor oil, ergot, colchicum – were pharmaco-
advanced sufficiently to give a scientific basis to drug logically active. Of the 311 preparations, 103 were ‘chemi-
therapy, and it was not until the beginning of the 20th cals’, mainly inorganic – iodine, ferrous sulfate, sodium
century that this knowledge actually began to be applied bicarbonate, and many toxic salts of bismuth, arsenic, lead
to the discovery of new drugs. In the long interim, the and mercury – but also a few synthetic chemicals, such as
apothecary’s trade flourished; closely controlled by guilds diethyl ether and chloroform. The remainder comprised
and apprenticeship schemes, it formed the supply route miscellaneous materials and a few animal products, such
for the exotic preparations that were used in treatment. as lard, cantharidin and cochineal.
The early development of therapeutics – based, as we have
seen, mainly on superstition and on theories that have
been swept away by scientific advances – represents prehis-
tory as far as the development of the pharmaceutical
AN INDUSTRY BEGINS TO EMERGE
industry is concerned, and there are few, if any, traces of
it remaining2. For the pharmaceutical industry, the transition from pre-
history to actual history occurred late in the 19th century
(3Q19C, as managers of today might like to call it), when
three essential strands came together. These were: the
THERAPEUTICS IN THE   evolving science of biomedicine (and especially pharma-
19TH CENTURY cology); the emergence of synthetic organic chemistry; and
the development of a chemical industry in Europe, coupled
Although preventive medicine had made some spectacular with a medical supplies trade – the result of buoyant
advances, for example in controlling scurvy (Lind, 1763) entrepreneurship, mainly in America.
and in the area of infectious diseases, vaccination (Jenner,
1798), curtailment of the London cholera epidemic of Developments in biomedicine
1854 by turning off the Broad Street Pump (Snow), and
control of childbirth fever and surgical infections using Science began to be applied wholeheartedly to medicine
antiseptic techniques (Semmelweis, 1861; Lister, 1867), – as to almost every other aspect of life – in the 19th
therapeutic medicine was virtually non-existent until the century. Among the most important milestones from the
end of the 19th century. point of view of drug discovery was the elaboration in
Oliver Wendell Holmes – a pillar of the medical estab- 1858 of cell theory, by the German pathologist Rudolf
lishment – wrote in 1860: ‘I firmly believe that if the whole Virchow. Virchow was a remarkable man: pre-eminent as
materia medica, as now used, could be sunk to the bottom a pathologist, he also designed the Berlin sewage system
of the sea, it would be all the better for mankind – and and instituted hygiene inspections in schools, and later
the worse for the fishes’ (see Porter, 1997). This may have became an active member of the Reichstag. The tremen-
been a somewhat ungenerous appraisal, for some contem- dous reductionist leap of the cell theory gave biology –
porary medicines – notably digitalis, famously described and the pharmaceutical industry – the scientific foundation
by Withering in 1785, extract of willow bark (salicylic it needed. It is only by thinking of living systems in terms
acid), and Cinchona extract (quinine) – had beneficial of the function of their cells that one can begin to under-
effects that were well documented. But on balance, Holmes stand how molecules affect them.
was right – medicines did more harm than good. A second milestone was the birth of pharmacology as a
scientific discipline when the world’s first Pharmacological
Institute was set up in 1847 at Dorpat by Rudolf Buchheim
2
Plenty of traces remain outside the pharmaceutical industry, in – literally by Buchheim himself, as the Institute was in his
the form of a wide variety of ‘alternative’ and ‘complementary’ own house and funded by him personally. It gained such
therapeutic procedures, such as herbalism, moxibustion, reflexology
and acupuncture, whose underlying principles originated in the recognition that the university built him a new one 13
prescientific era and remain largely beyond the boundaries of science. years later. Buchheim foresaw that pharmacology as a
It may not be long, given the growing appeal of such approaches in science was needed to exploit the knowledge of physiol-
the public’s eye, before the mainstream pharmaceutical industry
decides that it must follow this trend. That will indeed be a challenge ogy, which was being advanced by pioneers such as
for drug discovery research. Magendie and Claude Bernard, and link it to therapeutics.

4
 The development of the pharmaceutical industry Chapter |1|

When one remembers that this was at a time when organic the chemical and commercial developments that were
chemistry and physiology were both in their cradles, and going on simultaneously.
therapeutics was ineffectual, Buchheim’s vision seems
bold, if not slightly crazy. Nevertheless, his Institute was a
spectacular success. Although he made no truly seminal
Developments in chemistry
discoveries, Buchheim imposed on himself and his staff The first synthetic chemicals to be used for medical pur-
extremely high standards of experimentation and argu- poses were, ironically, not therapeutic agents at all, but
ment, which eclipsed the empiricism of the old therapeu- anaesthetics. Diethyl ether (’sweet oil of vitriol’) was first
tic principles and attracted some exceptionally gifted made and described in 1540. Early in the 19th century,
students. Among these was the legendary Oswald it and nitrous oxide (prepared by Humphrey Davy in
Schmiedeberg, who later moved to Strasbourg, where he 1799 and found – by experiments on himself – to have
set up an Institute of Pharmacology of unrivalled size and stupefying properties) were used to liven up parties and
grandeur, which soon became the Mecca for would-be sideshows; their usefulness as surgical anaesthetics was
pharmacologists all over the world. demonstrated, amid much controversy, only in the 1840s3,
A third milestone came with Louis Pasteur’s germ theory by which time chloroform had also made its appearance.
of disease, proposed in Paris in 1878. A chemist by train- Synthetic chemistry at the time could deal only with very
ing, Pasteur’s initial interest was in the process of fermen- simple molecules, made by recipe rather than reason, as
tation of wine and beer, and the souring of milk. He our understanding of molecular structure was still in its
showed, famously, that airborne infection was the under- infancy. The first therapeutic drug to come from synthetic
lying cause, and concluded that the air was actually alive chemistry was amyl nitrite, prepared in 1859 by Guthrie
with microorganisms. Particular types, he argued, were and introduced, on the basis of its vasodilator activity, for
pathogenic to humans, and accounted for many forms of treating angina by Brunton in 1864 – the first example of
disease, including anthrax, cholera and rabies. Pasteur suc- a drug born in a recognizably ‘modern’ way, through the
cessfully introduced several specific immunization proce- application of synthetic chemistry, physiology and clinical
dures to give protection against infectious diseases. Robert medicine. This was a landmark indeed, for it was nearly
Koch, Pasteur’s rival and near-contemporary, clinched the 40 years before synthetic chemistry made any further sig-
infection theory by observing anthrax and other bacilli in nificant contribution to therapeutics, and not until well
the blood of infected animals. into the 20th century that physiological and pharmaco-
The founder of chemotherapy – some would say the logical knowledge began to be applied to the invention of
founder of molecular pharmacology – was Paul Ehrlich new drugs.
(see Drews, 2004 for a brief biography). Born in 1854 and It was during the latter half of the 19th century that
trained in pathology, Ehrlich became interested in histo- the foundations of synthetic organic chemistry were laid,
logical stains and tested a wide range of synthetic chemical the impetus coming from work on aniline, a copious
dyes that were being produced at that time. He invented byproduct of the coal-tar industry. An English chemist,
‘vital staining’ – staining by dyes injected into living Perkin, who in 1856 succeeded in preparing from aniline
animals – and described how the chemical properties of a vivid purple compound, mauvein, laid the foundations.
the dyes, particularly their acidity and lipid solubility, This was actually a chemical accident, as Perkin’s aim had
influenced the distribution of dye to particular tissues and been to synthesize quinine. Nevertheless, the discovery
cellular structures. Thence came the idea of specific binding gave birth to the synthetic dyestuffs industry, which played
of molecules to particular cellular components, which a major part in establishing the commercial potential of
directed not only Ehrlich’s study of chemotherapeutic synthetic organic chemistry – a technology which later
agents, but much of pharmacological thinking ever since. became a lynchpin of the evolving pharmaceutical indus-
‘Receptor’ and ‘magic bullets’ are Ehrlich’s terms, though try. A systematic approach to organic synthesis went hand
he envisaged receptors as targets for toxins, rather than in hand with improved understanding of chemical struc-
physiological mediators. Working in Koch’s Institute, ture. Crucial steps were the establishment of the rules
Ehrlich developed diphtheria antitoxin for clinical use, of chemical equivalence (valency), and the elucidation of
and put forward a theory of antibody action based on the structure of benzene by Von Kekulé in 1865. The first
specific chemical recognition of microbial macromole- representation of a structural formula depicting the bonds
cules, work for which he won the 1908 Nobel Prize.
Ehrlich became director of his own Institute in Frankfurt,
close to a large dye works, and returned to his idea of using 3
An event welcomed, in his inimitable prose style, by Oliver Wendell
the specific binding properties of synthetic dyes to develop Holmes in 1847: ‘The knife is searching for disease, the pulleys are
selective antimicrobial drugs. dragging back dislocated limbs – Nature herself is working out the
At this point, we interrupt the biological theme at the primal curse which doomed the tenderest of her creatures to the
sharpest of her trials, but the fierce extremity of suffering has been
end of the 19th century, with Ehrlich in full flood, on the steeped in the waters of forgetfulness, and the deepest furrow in the
verge of introducing the first designer drugs, and turn to knotted brow of agony has been smoothed forever’.

5
Section | 1 | Introduction and Background

between atoms in two dimensions, based on valency rules, working in Paris in the period 1810–1825. The recognition
also appeared in 18654. that medicinal plants owed their properties to their indi-
The reason why Perkin had sought to synthesize quinine vidual chemical constituents, rather than to some intangi-
was that the drug, prepared from Cinchona bark, was much ble property associated with their living nature, marks a
in demand for the treatment of malaria, one of whose critical point in the history of the pharmaceutical industry.
effects is to cause high fever. So quinine was (wrongly, as It can be seen as the point of origin of two of the three
it turned out) designated as an antipyretic drug, and used strands from which the industry grew – namely the begin-
to treat fevers of all kinds. Because quinine itself could not nings of the ‘industrialization’ of the apothecary’s trade,
be synthesized, fragments of the molecule were made and the emergence of the science of pharmacology. And
instead; these included antipyrine, phenacetin and various by revealing the chemical nature of medicinal prepara-
others, which were introduced with great success in the tions, it hinted at the future possibility of making medi-
1880s and 1890s. These were the first drugs to be ‘designed’ cines artificially. Even though, at that time, synthetic
on chemical principles5. organic chemistry was barely out of its cradle, these dis-
coveries provided the impetus that later caused the chemi-
cal industry to turn, at a very early stage in its history, to
The apothecaries’ trade making drugs.
Despite the lack of efficacy of the pharmaceutical prepara- The first local apothecary business to move into large-
tions that were available in the 19th century, the apothe- scale production and marketing of pharmaceuticals was
cary’s trade flourished; then, as now, physicians felt the old-established Darmstadt firm Merck, founded in
themselves obliged to issue prescriptions to satisfy the 1668. This development, in 1827, was stimulated by the
expectations of their patients for some token of remedial advances in purification of natural products. Merck was
intent. Early in the 19th century, when many small apoth- closely followed in this astute business move by other
ecary businesses existed to satisfy the demand on a local German- and Swiss-based apothecary businesses, giving
basis, a few enterprising chemists undertook the task of rise to some which later also became giant pharmaceut­
isolating the active substances from these plant extracts. ical companies, such as Schering and Boehringer. The
This was a bold and inspired leap, and one that attracted American pharmaceutical industry emerged in the middle
a good deal of ridicule. Although the old idea of ‘signa- of the 19th century. Squibb began in 1858, with ether
tures’, which held that plants owed their medicinal proper- as its main product. Soon after came Parke Davis (1866)
ties to their biological characteristics6, was falling into and Eli Lilly (1876); both had a broader franchise
disrepute, few were willing to accept that individual chem- as manufacturing chemists. In the 1890s Parke Davis
ical substances could be responsible for the effects these became the world’s largest pharmaceutical company, one
plants produced, such as emesis, narcosis, purgation or of whose early successes was to purify crystalline adrena-
fever. The trend began with Friedrich Sertürner, a junior line from adrenal glands and sell it in ampoules for injec-
apothecary in Westphalia, who in 1805 isolated and puri- tion. The US scientific community contested the adoption
fied morphine, barely surviving a test of its potency on of the word ‘adrenaline’ as a trade name, but industry
himself. This was the first ‘alkaloid’, so named because of won the day and the scientists were forced to call the
its ability to neutralize acids and form salts. This discovery hormone ‘epinephrine’.
led to the isolation of several more plant alkaloids, includ- The move into pharmaceuticals was also followed by
ing emetine, strychnine, caffeine and quinine, mainly by several chemical companies such as Bayer, Hoechst, Agfa,
two remarkably prolific chemists, Caventou and Pelletier, Sandoz, Geigy and others, which began, not as apothecar-
ies, but as dyestuffs manufacturers. The dyestuffs industry
at that time was also based largely on plant products,
4
Its author, the Edinburgh chemist Alexander Crum Brown, was also a which had to be refined, and were sold in relatively small
pioneer of pharmacology, and was the first person to use a chemical quantities, so the commercial parallels with the pharma-
reaction – quaternization of amines – to modify naturally occurring
substances such as strychnine and morphine. With Thomas Fraser, in ceutical industry were plain. Dye factories, for obvious
1868, he found that this drastically altered their pharmacological reasons, were usually located close to large rivers, a fact
properties, changing strychnine, for example, from a convulsant to a that accounts for the present-day location of many large
paralysing agent. Although they knew neither the structures of these
molecules nor the mechanisms by which they acted, theirs was the
pharmaceutical companies in Europe. As we shall see, the
first systematic study of structure–activity relationships. link with the dyestuffs industry later came to have much
5
These drugs belong pharmacologically to the class of non-steroidal more profound implications for drug discovery.
anti-inflammatories (NSAIDs), the most important of which is aspirin From about 1870 onwards – following the crucial dis-
(acetylsalicylic acid). Ironically, aspirin itself had been synthesized many
years earlier, in 1855, with no pharmacological purpose in mind. covery by Kekulé of the structure of benzene – the dye-
Aspirin was not developed commercially until 1899, subsequently stuffs industry turned increasingly to synthetic chemistry
generating huge revenues for Bayer, the company responsible. as a source of new compounds, starting with aniline-
6
According to this principle, pulmonaria (lungwort) was used to treat
respiratory disorders because its leaves resembled lungs, saffron to based dyes. A glance through any modern pharmacopoeia
treat jaundice, and so on. will show the overwhelming preponderance of synthetic

6
 The development of the pharmaceutical industry Chapter |1|

aromatic compounds, based on the benzene ring structure, Almost simultaneously, Einthorn in Munich synthesized
among the list of useful drugs. Understanding the nature procaine, the first synthetic local anaesthetic drug, which
of aromaticity was critical. Though we might be able to followed the naturally occurring alkaloid cocaine. The
dispense with the benzene ring in some fields of applied local anaesthetic action of cocaine on the eye was dis­
chemistry, such as fuels, lubricants, plastics or detergents, covered by Sigmund Freud and his ophthalmologist col-
its exclusion would leave the pharmacopoeia bankrupt. league Koeller in the late 19th century, and was heralded
Many of these dyestuffs companies saw the potential of as a major advance for ophthalmic surgery. After several
the medicines business from 1880 onwards, and moved chemists had tried, with limited success, to make synthetic
into the area hitherto occupied by the apothecaries. The compounds with the same actions, procaine was finally
result was the first wave of companies ready to apply produced and introduced commercially in 1905 by
chemical technology to the production of medicines. Hoechst. Barbitone and procaine were triumphs for chem-
Many of these founder companies remained in business ical ingenuity, but owed little or nothing to physiology, or,
for years. It was only recently, when their cannibalistic indeed, to pharmacology. The physiological site or sites of
urges took over in the race to become large, that mergers action of barbiturates remain unclear to this day, and their
and takeovers caused many names to disappear. mechanism of action at the molecular level was unknown
Thus the beginnings of a recognizable pharmaceutical until the 1980s.
industry date from about 1860–1880, its origins being in From this stage, where chemistry began to make an
the apothecaries and medical supplies trades on the one impact on drug discovery, up to the last quarter of the 20th
hand, and the dyestuffs industry on the other. In those century, when molecular biology began to emerge as a
early days, however, they had rather few products to sell; dominant technology, we can discern three main routes
these were mainly inorganic compounds of varying by which new drugs were discovered, namely chemistry-
degrees of toxicity and others best described as concoc- driven approaches, target-directed approaches, and acci-
tions. Holmes (see above) dismissed the pharmacopoeia dental clinical discoveries. In many of the most successful
in 1860 as worse than useless. case histories, graphically described by Weatherall (1990),
To turn this ambitious new industry into a source of the three were closely interwoven. The remarkable family
human benefit, rather than just corporate profit, required of diverse and important drugs that came from the ori­
two things. First, it had to embrace the principles of ginal sulfonamide, lead, described below, exemplifies this
biomedicine, and in particular pharmacology, which pattern very well.
provided a basis for understanding how disease and
drugs, respectively, affect the function of living organisms.
Second, it had to embrace the principles of chemistry, Chemistry-driven drug discovery
going beyond the descriptors of colour, crystallinity taste,
volatility, etc., towards an understanding of the structure Synthetic chemistry
and properties of molecules, and how to make them in The pattern of drug discovery driven by synthetic chem­
the laboratory. As we have seen, both of these fields had istry – with biology often struggling to keep up – became
made tremendous progress towards the end of the 19th the established model in the early part of the 20th century,
century, so at the start of the 20th century the time was and prevailed for at least 50 years. The balance of research
right for the industry to seize its chance. Nevertheless, in the pharmaceutical industry up to the 1970s placed
several decades passed before the inventions coming from chemistry clearly as the key discipline in drug discovery,
the industry began to make a major impact on the treat- the task of biologists being mainly to devise and perform
ment of disease. assays capable of revealing possible useful therapeutic
activity among the many anonymous white powders that
arrived for testing. Research management in the industry
The industry enters the 20th century
was largely in the hands of chemists. This strategy pro-
By the end of the 19th century various synthetic drugs had duced many successes, including benzodiazepine tran-
been made and tested, including the ‘antipyretics’ (see quillizers, several antiepileptic drugs, antihypertensive
above) and also various central nervous system depres- drugs, antidepressants and antipsychotic drugs. The survi­
sants. Chemical developments based on chloroform had ving practice of classifying many drugs on the basis of their
produced chloral hydrate, the first non-volatile CNS chemical structure (e.g. phenothiazines, benzodiazepines,
depressant, which was in clinical use for many years as a thiazides, etc.), rather than on the more logical basis of
hypnotic drug. Independently, various compounds based their site or mode of action, stems from this era. The
on urea were found to act similarly, and von Mering fol- development of antiepileptic drugs exemplifies this
lowed this lead to produce the first barbiturate, barbitone approach well. Following the success of barbital (see
(since renamed barbital), which was introduced in 1903 above) several related compounds were made, including
by Bayer and gained widespread clinical use as a hypnotic, the phenyl derivative phenobarbital, first made in 1911. This
tranquillizer and antiepileptic drug – the first blockbuster. proved to be an effective hypnotic (i.e. sleep-inducing)

7
Section | 1 | Introduction and Background

drug, helpful in allowing peaceful nights in a ward full of Westminster Abbey to improve its acoustics. But the fact
restive patients. By chance, it was found by a German remains that Nature unexpectedly provides some of our
doctor also to reduce the frequency of seizures when tested most useful drugs, and most of its potential remains
in epileptic patients – an example of clinical serendipity untapped.
(see below) – and it became widely used for this purpose,
being much more effective in this regard than barbital
itself. About 20 years later, Putnam, working in Boston,
Target-directed drug discovery
developed an animal model whereby epilepsy-like sei- Although chemistry was the pre-eminent discipline in
zures could be induced in mice by electrical stimulation drug discovery until the 1970s, the seeds of the biological
of the brain via extracranial electrodes. This simple model revolution had long since been sown, and within the
allowed hundreds of compounds to be tested for potential chemistry-led culture of the pharmaceutical industry these
antiepileptic activity. Phenytoin was an early success of this developments began to bear fruit in certain areas. This
programme, and several more compounds followed, as happened most notably in the field of chemotherapy,
chemists from several companies embarked on synthetic where Ehrlich played such an important role as the first
programmes. None of this relied at all on an understand- ‘modernist’ who defined the principles of drug specificity
ing of the mechanism of action of these compounds – in terms of a specific interaction between the drug mole-
which is still controversial; all that was needed were teams cule and a target molecule – the ‘receptive substance’ – in
of green-fingered chemists, and a robust assay that pre- the organism, an idea summarized in his famous Latin
dicted efficacy in the clinic. catchphrase Corpora non agunt nisi fixata. Although we now
take it for granted that the chemical nature of the target
molecule, as well as that of the drug molecule, determines
Natural product chemistry
what effects a drug will produce, nobody before Ehrlich
We have mentioned the early days of pharmacology, with had envisaged drug action in this way7. By linking chem-
its focus on plant-derived materials, such as atropine, istry and biology, Ehrlich effectively set the stage for drug
tubocurarine, strychnine, digitalis and ergot alkaloids, which discovery in the modern style. But despite Ehrlich’s seminal
were almost the only drugs that existed until well into role in the evolution of the pharmaceutical industry, dis-
the 20th century. Despite the rise of synthetic chemistry, coveries in his favourite field of endeavour, chemotherapy,
natural products remain a significant source of new drugs, remained for many years empirical rather than target
particularly in the field of chemotherapy, but also in other directed8.
applications. Following the discovery of penicillin by The fact is that Ehrlich’s preoccupation with the binding
Fleming in 1929 – described by Mann (1984) as ‘the most of chemical dyes, as exemplified by biological stains, for
important medical discovery of all time’ – and its develop- specific constituents of cells and tissues, turned out to be
ment as an antibiotic for clinical use by Chain and Florey misplaced, and not applicable to the problem of achieving
in 1938, an intense search was undertaken for antibac­ selective toxicity. Although he soon came to realize that
terial compounds produced by fungi and other microor- the dye-binding moieties of cells were not equivalent to
ganisms, which yielded many useful antibiotics, including the supposed drug-binding moieties, neither he nor
chloramphenicol (1947), tetracyclines (1948), streptomycin anyone else succeeded in identifying the latter and using
(1949) and others. The same fungal source that yielded them as defined targets for new compounds. The history
streptomycin also produced actinomycin D, used in cancer of successes in the field of chemotherapy prior to the
chemotherapy. Higher plants have continued to yield antibiotic era, some of which are listed in Table 1.1,
useful drugs, including vincristine and vinblastine (1958),
and paclitaxel (Taxol, 1971).
Outside the field of chemotherapy, successful drugs
7
derived from natural products include ciclosporin (1972) Others came close at around the same time, particularly the British
physiologist J. N. Langley (1905), who interpreted the neuromuscular
and tacrolimus (1993), both of which come from fungi and blocking effect of ‘curari’ in terms of its interaction with a specific
are used to prevent transplant rejection. Soon after came ‘receptive substance’ at the junction between the nerve terminal and
mevastatin (1976), another fungal metabolite, which was the muscle fibre. This was many years before chemical transmission
at this junction was discovered. Langley’s student, A. V. Hill (1909),
the first of the ‘statin’ series of cholesterol-lowering drugs first derived the equations based on the Law of Mass Action, which
that act by inhibiting the enzyme HMG CoA reductase. describe how binding varies with drug concentration. Hill’s
Overall, the pharmaceutical industry continues to have quantitative theory later formed the basis of ‘receptor theory’,
elaborated by pharmacologists from A. J. Clark (1926) onwards.
something of a love–hate relationship with natural prod- Although this quantitative approach underlies much of our current
ucts. They often have weird and wonderful structures that thinking about drug–receptor interactions, it was Ehrlich’s more
cause hardened chemists to turn pale; they are often near- intuitive approach that played the major part in shaping drug
impossible to synthesize, troublesome to produce from discovery in the early days.
8
Even now, important new chemotherapeutic drugs, such as the
natural sources, and ‘optimizing’ such molecules to make taxanes, continue to emerge through a combination of a chance
them suitable for therapeutic use is akin to remodelling biological discovery and high-level chemistry.

8
 The development of the pharmaceutical industry Chapter |1|

Table 1.1  Examples of drugs from different sources: natural products, synthetic chemistry
and biopharmaceuticals

Natural products Synthetic chemistry Biopharmaceuticals produced by

recombinant DNA technology
Antibiotics (penicillin, streptomycin, Early successes include: Human insulin (the first biotech
tetracyclines, cephalosporins, etc.) Antiepileptic drugs product, registered 1982)
Anticancer drugs (doxorubicin, Antihypertensive drugs Human growth hormone
bleomycin, actinomycin, vincristine, Antimetabolites α-interferon, γ-interferon
vinblastine, paclitaxel (Taxol), etc.) Barbiturates Hepatitis B vaccine
Atropine, hyoscine Bronchodilators Tissue plasminogen activator (t-PA)
Ciclosporin Diuretics Hirudin
Cocaine Local anaesthetics Blood clotting factors
Colchicine Sulfonamides Erythropoietin
Digitalis (digoxin) [Since c.1950, synthetic chemistry has G-CSF, GM-CSF
accounted for the great majority of
new drugs]
Ephedrine
Heparin
Human growth hormone*
Insulin (porcine, bovine)*
Opium alkaloids (morphine,
papaverine)
Physostigmine
Rauwolfia alkaloids (reserpine)
Statins
Streptokinase
Tubocurarine
Vaccines
*Now replaced by material prepared by recombinant DNA technology.
G-CSF, granulocyte colony-stimulating factor; GM-CSF, granulocyte macrophage colony-stimulating factor.

actually represents a series of striking achievements in Eckhardt, 1990; Lednicer, 1993). A few selected case his-
synthetic chemistry, coupled to the development of assay tories exemplify this general trend.
systems in animals, according to the chemistry-led model
that we have already discussed. The popular image of
‘magic bullets’ – a phrase famously coined by Ehrlich – The sulfonamide story
designed to home in, like cruise missiles, on defined Ehrlich’s major triumph was the discovery in 1910 of Sal-
targets is actually a misleading one in the context of the varsan (Compound 606), the first compound to treat
early days of chemotherapy, but there is no doubt that syphilis effectively, which remained in use for 40 years.
Ehrlich’s thinking prepared the ground for the steady Still, bacterial infections, such as pneumonia and wound
advance of target-directed approaches to drug discovery, a infections, proved resistant to chemical treatments for
trend that, from the 1950s onwards, steadily shifted the many years, despite strenuous effort on the part of the
industry’s focus from chemistry to biology (Maxwell and pharmaceutical industry. In 1927, IG Farbenindustrie,

9
Section | 1 | Introduction and Background

Pteridine PABA H 2N COOH

Sulfonamides H2N SO2NHR

Pteroic acid

Glutamate

H2N N N
COOH

N CH2 NH CO NH CH
N
CH2
OH
COOH
Pteridine PABA Glutamate
Folic acid

Fig. 1.1  Folic acid synthesis and PABA.

which had a long-standing interest in discovering antimi- were made in the next few years and they dramatically
crobial drugs, appointed Gerhard Domagk to direct their improved the prognosis of patients suffering from infec-
research. Among the various leads that he followed was a tious diseases.
series of azo dyes, included among which were some sul- The mechanistic light began to dawn in 1940, when
fonamide derivatives (a modification introduced earlier D. D. Woods, a microbiologist in Oxford, discovered that
into dyestuffs to improve their affinity for certain fibres). the antibacterial effect of sulfonamides was antagonized
These were much more effective in animals, and less toxic, by p-aminobenzoic acid (PABA), a closely related com-
than anything that had gone before, and Prontosil – a dark- pound and a precursor in the biosynthesis of folic acid
red azo dye – was introduced in 1935. In the same year, (Figure 1.1). Bacteria, but not eukaryotic cells, have to
it saved the life of Domagk’s daughter, who developed synthesize their own folic acid to support DNA synthesis.
septicaemia after a needle prick. It was soon discovered Woods deduced that sulfonamides compete with PABA for
that the azo linkage in the Prontosil molecule was rapidly a target enzyme, now known to be dihydropteroate syn-
cleaved in the body, yielding the colourless compound thase, and thus prevent folic acid synthesis.
sulfanilamide, which accounted for the antibacterial effect The discovery of sulfonamides and the elucidation of
of Prontosil9. their mechanism of action had great repercussions, scien-
With chemistry still firmly in the driving seat, and little tifically as well as clinically. In the drug discovery field, it
concern about mechanisms or targets, many sulfonamides set off two major lines of inquiry. First, the sulfonamide
structure proved to be a rich source of molecules with
many different, and useful, pharmacological properties –
9
Sulfanilamide, a known compound, could not be patented, and an exuberant vindication of the chemistry-led approach to
so many companies soon began to make and sell it in various drug discovery. Second, attacking the folic acid pathway
formulations. In 1937 about 80 people who took the drug died as a proved to be a highly successful strategy for producing
result of solvent-induced liver and kidney damage. It was this accident
that led to the US Food and Drug Act, with the Food & Drug therapeutically useful drugs – a powerful boost for the
Administration (FDA) to oversee it (see Chapter 20). ‘targeteers’, who were still few in number at this time.

10
 The development of the pharmaceutical industry Chapter |1|

Prontosil (1932)
First antibacterial drug

Sulfanilamide (1935) Other sulfonamides

Active metabolite of Prontosil

Carbutamide (1955) Acetazolamide (1949)

First oral hypoglycaemic drug First carbonic anhydrase inhibitor
Other sulfonylureas Other carbonic
anhydrase inhibitors

Chlorothiazide (1957)
Major improvement in diuretics
Other
thiazide diuretics

Diazoxide (1961) Furosemide (1962)

Novel hypotensive drug First loop diuretic
Other
loop diuretics

Fig. 1.2  Sulfonamide dynasty.

The chemical dynasty originating with sulfanilamide is modifications of the thiazide structures led to the acciden-
shown in Figure 1.2. An early spin-off came from the clini- tal but important discovery of a series of hypotensive
cal observation that some sulfonamides produced an alka- vasodilator drugs, such as hydralazine and diazoxide. In yet
line diuresis, associated with an increased excretion of another development, carbutamide, one of the sulfona-
sodium bicarbonate in the urine. Carbonic anhydrase, an mides synthesized by Boehringer in 1954 as part of an
enzyme which catalyses the interconversion of carbon antibacterial drug programme, was found accidentally to
dioxide and carbonic acid, was described in 1940, and its cause hypoglycaemia. This drug was the first of the sulfo-
role in renal bicarbonate excretion was discovered a few nylurea series, from which many further derivatives, such
years later, which prompted the finding that some, but not as tolbutamide and glibenclamide, were produced and used
all, sulfonamides inhibit this enzyme. Modification of the successfully to treat diabetes. All of these products of the
sulfonamide structure led eventually to acetazolamide the sulfonamide dynasty are widely used today. Their chemi-
first commercially available carbonic anhydrase inhibitor, cal relationship to sulfonamides is clear, though none of
as a diuretic in 1952. Following the diuretic trail led in them has antibacterial activity. Their biochemical targets
turn to chlorothiazide (1957), the first of the thiazide diu- in smooth muscle, the kidney, the pancreas and elsewhere,
retics, which, though devoid of carbonic anhydrase inhibi- are all different. Chemistry, not biology, was the guiding
tory activity, was much more effective than acetazolamide principle in their discovery and synthesis.
in increasing sodium excretion, and much safer than the Target-directed approaches to drug design have played a
earlier mercurial diuretics, which had until then been the much more significant role in areas other than antibacte-
best drugs available for treating oedema associated with rial chemotherapy, the approaches being made possible by
heart failure and other conditions. Still further modifica- advances on two important fronts, separated, as it happens,
tions led first to furosemide (1962) and later to bumetanide by the Atlantic Ocean. In the United States, the antime-
(1984), which were even more effective than the thiazides tabolite principle, based on interfering with defined meta-
in producing a rapid diuresis – ‘torrential’ being the adjec- bolic pathways, proved to be highly successful, due largely
tive applied by clinicians with vivid imaginations. Other to the efforts of George Hitchings and Gertrude Elion at

11
Section | 1 | Introduction and Background

Burroughs Wellcome. In Europe, drug discovery took its James Black and  
lead more from physiology than biochemistry, and sprung
receptor-targeted drugs
from advances in knowledge of chemical transmitters and
their receptors. The names of Henry Dale and James Black As already mentioned, the concept of ‘receptors’ as recog-
deserve special mention here. nition sites for hormones and other physiological media-
tors came from J. N. Langley’s analysis of the mechanism
of action of ‘curari’. Henry Dale’s work on the distinct
Hitchings and Elion and the ‘muscarinic’ and ‘nicotinic’ actions of acetylcholine also
antimetabolite principle pointed to the existence of two distinct types of choliner-
George Hitchings and Gertrude Elion came together in gic receptor, though Dale himself dismissed the receptor
1944 in the biochemistry department of Burroughs Well- concept as an abstract and unhelpful cloak for ignorance.
come in Tuckahoe, New York. Their biochemical interest During the 1920s and 1930s, major discoveries high­
lay in the synthesis of folic acid, based on the importance lighting the role of chemical mediators were made by
of this pathway for the action of sulfonamides, and they physiologists, including the discovery of insulin, adrenal
set about synthesizing potential ‘antimetabolites’ of steroids and several neurotransmitters, and the realization
purines and pyrimidines as chemotherapeutic agents. At that chemical signalling was crucial for normal function
the time, it was known that this pathway was important focused attention on the receptor mechanisms needed to
for DNA synthesis, but the role of DNA in cell function decode these signals. Pharmacologists, particularly A. J.
was uncertain. It turned out to be an inspired choice, and Clark, J. H. Gaddum and H. O. Schild, applied the Law of
theirs was one of the first drug discovery programmes to Mass Action to put ligand–receptor interactions on a
focus on a biochemical pathway, rather than on a series of quantitative basis. Schild’s studies on drug antagonism
chemical compounds10. in particular, which allowed the binding affinity of com-
Starting from a series of purine and pyrimidine ana- petitive antagonists to receptors to be estimated from
logues, which had antibacterial activity, Hitchings and pharmacological experiments, were an important step
Elion identified a key enzyme in the folic acid pathway, forward, which provided the first – and still widely used
namely dihydrofolate reductase, which was necessary – quantitative basis for classifying drug receptors. On the
for DNA synthesis and was inhibited by many of their basis of such quantitative principles, R. P. Ahlquist in
antibacterial pyrimidine analogues. Because all cells, not 1948 proposed the existence of two distinct classes of
just bacteria, use this reaction to make DNA, they won- adrenergic receptor, α and β, which accounted for the
dered why the drugs showed selectivity in their ability to varied effects of epinephrine and norepinephrine on the
block cell division, and found that the enzyme showed cardiovascular system. This discovery inspired James
considerable species variation in its susceptibility to inhib- Black, working in the research laboratories of Imperial
itors. This led them to seek inhibitors that would selec- Chemical Industries in the UK, to seek antagonists that
tively attack bacteria, protozoa and human neoplasms, would act selectively on β-adrenoceptors and thus block
which they achieved with great success. Drugs to emerge the effects of epinephrine on the heart, which were
from this programme included the antituberculosis drug thought to be harmful in patients with coronary disease.
pyrimethamine, the antibacterial trimethoprim, and the anti- His chemical starting point was dichloroisoprenaline, which
cancer drug 6-mercaptopurine, as well as azathioprine, an had been found by Slater in 1957 to block the relaxant
immunosuppressant drug that was later widely used to effects of epinephrine on bronchial smooth muscle – of
prevent transplant rejection. Another spin-off from the no interest to Slater at the time, as he was looking for
work of Hitchings and Elion was allopurinol, an inhibitor compounds with the opposite effect. The result of Black’s
of purine synthesis that is used in the treatment of gout. efforts was the first β-adrenoceptor blocking drug, prone-
Later on, Elion – her enthusiasm for purines and pyrimi- thalol (1960), which had the desired effects in humans but
dines undiminished – led the research group which in was toxic. It was quickly followed by propranolol (regis-
1977 discovered one of the first effective antiviral drugs, tered in 196411) – one of the earliest blockbusters, which
aciclovir, an inhibitor of DNA polymerase, and later the found many important applications in cardiovascular
first antiretroviral drug, zidovudine (AZT), which inhibits medicine. This was the first time that a receptor, identified
reverse transcriptase. Hitchings and Elion, by focusing on pharmacologically, had been deliberately targeted in a
the metabolic pathways involved in DNA synthesis, drug discovery project.
invented an extraordinary range of valuable therapeutic Black, after moving to Smith Kline and French, went on
drugs, an achievement unsurpassed in the history of drug from this success to look for novel histamine antagonists
discovery.
11
Ironically, propranolol had been made in 1959 in the laboratories of
10
They were not alone. In the 1940s, a group at Lederle laboratories Boehringer Ingelheim, as part of a different, chemistry-led project.
made aminopterin and methotrexate, folic acid antagonists which Only when linked to its target could the clinical potential of
proved effective in treating leukaemia. propranolol be revealed – chemistry alone was not enough!

12
 The development of the pharmaceutical industry Chapter |1|

that would block the stimulatory effect of histamine on children became much less agitated. From this chance
gastric acid secretion, this effect being resistant to the then- observation he went on to set up one of the first controlled
known antihistamine drugs. The result of this project, in clinical trials, which demonstrated unequivocally that
which the chemistry effort proved much tougher than the amphetamine had a calming effect – quite unexpected for
β-adrenoceptor antagonist project, was the first H2-receptor a drug known to have stimulant effects in other circum-
antagonist, burimamide (1972). This compound was a stances. From this developed the widespread use, validated
major clinical advance, being the first effective drug for by numerous controlled clinical trials, of amphetamine-
treating peptic ulcers, but (like pronethalol) was quickly like drugs, particularly methylphenidate (Ritalin) to treat
withdrawn because of toxicity to be replaced by cimetidine attention deficit hyperactivity disorder (ADHD) in chil-
(1976). In 1988 Black, along with Hitchings and Elion, dren. Other well-known examples include the discovery of
was awarded the Nobel Prize. the antipsychotic effects of phenothiazines by Laborit in
Black’s work effectively opened up the field of receptor 1949. Laborit was a naval surgeon, concerned that patients
pharmacology as an approach to drug discovery, and the were dying from ‘surgical shock’ – circulatory collapse
pharmaceutical industry quickly moved in to follow his resulting in irreversible organ failure – after major opera-
example. Lookalike β-adrenoceptor antagonists and H2- tions. Thinking that histamine might be involved, he
receptor antagonists followed rapidly during the 1970s tested the antihistamine promethazine combined with
and 1980s, and many other receptors were set up as targets autonomic blocking drugs to prevent this cardiovascular
for potential therapeutic agents, based on essentially the reaction. Although it was ineffective in treating shock,
same approach – though with updated technology – that promethazine caused some sedation and Laborit tried
Black and his colleagues had introduced (Page et al., 2011; some chemically related sedatives, notably promazine,
Walker, 2011). which had little antihistamine activity. Patients treated
Drews (2000) estimated that of 483 identified targets with it fared better during surgery, but Laborit particularly
on which the current set of approved drugs act, G-protein- noticed that they appeared much calmer postoperatively
coupled receptors – of which β-adrenoceptors and H2- He therefore persuaded his psychiatrist colleagues to test
receptors are typical examples – account for 45%. Many the drug on psychotic patients, tests that quickly revealed
other successful drugs have resulted from target-directed the drug’s antipsychotic effects and led to the development
projects along the lines pioneered by Black and his col- of the antipsychotic chlorpromazine. In a sequel, other
leagues. In recent years, of course, receptors have changed phenothiazine-like tricyclic compounds were tested for
from being essentially figments in an operational scheme antipsychotic activity but were found accidentally to
devised by pharmacologists to explain their findings, to relieve the symptoms of depression. After Bradley and
being concrete molecular entities that can be labelled, Laborit, psychiatrists had become alert to looking for the
isolated as proteins, cloned and expressed, just like many unexpected.
other proteins. As we shall see in later chapters, these Astute clinical observation has revealed many other
advances have completely transformed the techniques unexpected therapeutic effects, for example the efficacy of
employed in drug discovery research. various antidepressant and antiepileptic drugs in treating
certain intractable pain states.
Accidental clinical discoveries
The regulatory process
Another successful route to the discovery of new drugs has
been through observations made in the clinic. Until drug In the mid-19th century restrictions on the sale of
discovery became an intentional activity, such serendipi- poisonous substances were imposed in the USA and UK,
tous observations were the only source of knowledge. but it was not until the early 1900s that any system of
Withering’s discovery in 1785 of the efficacy of digitalis ‘prescription-only’ medicines was introduced, requiring
in treating dropsy, and Wenkebach’s discovery in 1914 of approval by a medical practitioner. Soon afterwards,
the antidysrhythmic effect of quinine, when he treated a restrictions began to be imposed on what ‘cures’ could be
patient with malaria who also happened to suffer from claimed in advertisements for pharmaceutical products
atrial tachycardia, are two of many examples where the and what information had to be given on the label; legisla-
clinical efficacy of plant-derived agents has been discov- tion evolved at a leisurely pace. Most of the concern was
ered by highly observant clinicians. More recently, clinical with controlling frankly poisonous or addictive substances
benefit of unexpected kinds has been discovered with syn- or contaminants, not with the efficacy and possible
thetic compounds developed for other purposes. In 1937, harmful effects of new drugs.
for example, Bradley tried amphetamine as a means of alle- In 1937, the use of diethylene glycol as a solvent for a
viating the severe headache suffered by children after sulfonamide preparation caused 107 deaths in the USA,
lumbar puncture (spinal tap) on the grounds that the and a year later the 1906 Food and Drugs Act was revised,
drug’s cardiovascular effects might prove beneficial. The requiring safety to be demonstrated before new products
headache was not alleviated, but Bradley noticed that the could be marketed, and also allowing federal inspection

13
Section | 1 | Introduction and Background

of manufacturing facilities. The requirement for proven

efficacy, as well as safety, was added in the Kefauver–Harris CONCLUDING REMARKS
amendment in 1962.
In Europe, preoccupied with the political events in the In this chapter we have followed the evolution of ideas
first half of the century, matters of drug safety and efficacy and technologies that have led to the state of drug discov-
were a minor concern, and it was not until the mid-1960s, ery research that existed circa 1970. The main threads,
in the wake of the thalidomide disaster – a disaster averted which came together, were:
in the USA by an assiduous officer, who used the provi-
sions of the 1938 Food and Drugs Act to delay licensing • Clinical medicine, by far the oldest of the
approval – that the UK began to follow the USA’s lead in antecedents, which relied largely on herbal remedies
regulatory laws. Until then, the ability of drugs to do harm right up to the 20th century
– short of being frankly poisonous or addictive – was not • Pharmacy, which began with the apothecary trade in
really appreciated, most of the concern having been about the 17th century, set up to serve the demand for
contaminants. In 1959, when thalidomide was first put on herbal preparations
the market by the German company Chemie Grünenthal, • Organic chemistry, beginning in the mid-19th
regulatory controls did not exist in Europe: it was up to century and evolving into medicinal chemistry via
the company to decide how much research was needed to dyestuffs
satisfy itself that the drug was safe and effective. Grü- • Pharmacology, also beginning in the mid-19th
nenthal made a disastrously wrong judgement (see century and setting out to explain the effects of
Sjöstrom and Nilsson, 1972, for a full account), which plant-derived pharmaceutical preparations in
resulted in an estimated 10 000 cases of severe congenital physiological terms.
malformation following the company’s specific recom- Some of the major milestones are summarized in
mendation that the drug was suitable for use by pregnant Table 1.2.
women. This single event caused an urgent reappraisal, The pharmaceutical industry as big business began
leading to the introduction of much tighter government around the beginning of the 20th century and for 60 or
controls. more years was dominated by chemistry. Gradually, from
In the UK, the Committee on the Safety of Drugs was the middle of the century onwards, the balance shifted
established in 1963. For the first time, as in the USA, all towards pharmacology until, by the mid-1970s, chemistry
new drugs (including new mixtures and formulations) and pharmacology were evenly balanced. This was a
had to be submitted for approval before clinical trials highly productive period for the industry, which saw many
could begin, and before they could be marketed. Legally, new drugs introduced; some of them truly novel but also
companies could proceed even if the Committee did not many copycat drugs, which found an adequate market
approve, but very few chose to do so. This loophole was despite their lack of novelty. The maturation of the scien-
closed by the Medicines Act (1968), which made it illegal tific and technological basis of the discovery process to its
to proceed with trials or the release of a drug without 1970s level coincided with the development of much
approval. Initially, safety alone was the criterion for more stringent regulatory controls, which also reached a
approval; in 1970, under the Medicines Act, evidence of degree of maturity at this time, and an acceptable balance
efficacy was added to the criteria for approval. It was the seemed to be struck between creativity and restraint.
realization that all drugs, not just poisons or contami- We ended our historical account in the mid-1970s,
nants, have the potential to cause harm that made it essen- when drug discovery seemed to have found a fairly serene
tial to seek proof of therapeutic efficacy to ensure that the and successful equilibrium, and products and profits
net effect of a new drug was beneficial. flowed at a healthy rate. Just around the corner, however,
In the decade leading up to 1970, the main planks in lay the arrival on the drug discovery scene of molecular
the regulatory platform – evidence of safety, efficacy and biology and its commercial wing, the biotechnology
chemical purity – were in place in most developed coun- industry, which over the next 20 years were to transform
tries. Subsequently, the regulations have been adjusted in the process and diversify its products in a dramatic fashion
various minor ways, and adopted with local variations in (see Chapters 3, 12 and 13). Starting in 1976, when the
most countries. first biotechnology companies (Cetus and Genentech)
A progressive tightening of the restrictions on the licens- were founded in the USA, there are now about 1300 such
ing of new drugs continued for about two decades after companies in the USA and another 900 in Europe, and
the initial shock of thalidomide, as public awareness of the products of such enterprises account for a steadily
the harmful effects of drugs became heightened, and the rising proportion – currently about 25% – of new thera-
regulatory bodies did their best to respond to public peutic agents registered. As well as contributing directly in
demand for assurance that new drugs were ‘completely terms of products, biotechnology is steadily and radically
safe’. The current state of licensing regulations is described transforming the ways in which conventional drugs are
in Chapter 20. discovered.

14
 The development of the pharmaceutical industry Chapter |1|

Table 1.2  Milestones in the development of the pharmaceutical industry

Year Event Notes

c.1550 BC Ebers papyrus The earliest known compendium of medical remedies
1540 Diethyl ether synthesized ‘Sweet oil of vitriol’, arguably the first synthetic drug
1668 Merck (Darmstadt) founded The apothecary business which later (1827) evolved
into the first large-scale pharmaceutical company
1763 Lind shows that lack of fruit causes scurvy
1775 Nitrous oxide synthesized
1785 Withering describes use of digitalis extract to treat The first demonstration of therapeutic efficacy
‘dropsy’
1798 Jenner shows that vaccination prevents smallpox
1799 Humphrey Davy demonstrates anaesthetic effect of
nitrous oxide
1803 Napoleon established examination and licensing
scheme for doctors
1806 Sertürner purifies morphine and shows it to be the A seminal advance – the first evidence that herbal
active principle of opium remedies contain active chemicals. Many other plant
alkaloids isolated 1820–1840
1846 Morton administers ether as anaesthetic at The first trial of surgical anaesthesia
Massachusetts General Hospital
1847 Chloroform administered to Queen Victoria to
control labour pain
1847 The first Pharmacological Institute set up by
Bucheim
mid-19C The first pharmaceutical companies formed:
Merck (1827)
Squibb (1858)
Hoechst (1862)
Parke Davis (1866)
Lilley (1876)
Burroughs Wellcome (1880) In many cases, pharmaceutical companies evolved
from dyestuffs companies or apothecaries
1858 Virchow proposes cell theory
1859 Amyl nitrite synthesized
1865 Benzene structure elucidated (Kekule), and first use Essential foundations for the development of organic
of structural fromulae to describe organic molecules synthesis
1867 Brunton demonstrates use of amyl nitrite to relieve
anginal pain
1878 Pasteur proposes germ theory of disease
1898 Heroin (diacetylmorphine) developed by Bayer The first synthetic derivative of a natural product.
Heroin was marketed as a safe and non-addictive
alternative to morphine

Continued

15
Section | 1 | Introduction and Background

Table 1.2  Continued

Year Event Notes

1899 Aspirin developed by Bayer
1903 Barbital developed by Bayer
1904 Elliott demonstrates biological activity of extracts of The first evidence for a chemical mediator – the
adrenal glands, and proposes adrenaline release as basis of much modern pharmacology
a physiological mechanism
1910 Ehrlich discovers Salvarsan The first antimicrobial drug, which revolutionized the
treatment of syphilis
1912 Starling coins the term ‘hormone’
1921 MacLeod, Banting and Best discover insulin Produced commercially from animal pancreas by Lilly
(1925)
1926 Loewi demonstrates release of ‘Vagusstoff’ from The first clear evidence for chemical
heart neurotransmission
1929 Fleming discovers penicillin Penicillin was not used clinically until Chain and
Florey solved production problems in 1938
1935 Domagk discovers sulfonamides The first effective antibacterial drugs, and harbingers
of the antimetabolite era
1936 Steroid hormones isolated by Upjohn company
1937 Bovet discovers antihistamines Subsequently led to discovery of antipsychotic drugs
1946 Gilman and Philips demonstrate anticancer effect of The first anticancer drug
nitrogen mustards
1951 Hitchings and Elion discover mercaptopurine The first anticancer drug from the antimetabolite
approach
1961 Hitchings and Schwartz discover azathioprine Also from the antimetabolite programme, the first
effective immunosuppressant able to prevent
transplant rejection
1962 Black and his colleagues discover pronethalol The first β-adrenoceptor antagonist to be used
clinically
1972 Black and his colleagues discover burimamide The first selective H2 antagonist
1976 Genentech founded The first biotech company, based on recombinant
DNA technology
c.1990 Introduction of combinatorial chemistry

The last quarter of the 20th century was a turbulent • Genomics as an approach to identifying new drug
period which affected quite radically the scientific basis of targets (Chapters 6 and 7)
the development of new medicines, as well as the com- • Increasing use of informatics technologies to store
mercial environment in which the industry operates. The and interpret data (Chapter 7)
changes that are occurring in the first quarter of the 21st • High-throughput screening of large compound
century show no sign of slowing down, and it is too soon libraries as a source of chemical leads
to judge which developments will prove genuinely suc- (Chapter 8)
cessful in terms of drug discovery, and which will not. In • Computational and automated chemistry as a means
later chapters we discuss in detail the present state of the of efficiently and systematically synthesizing
art with respect to the science and technology of drug collections of related compounds with drug-like
discovery. The major discernible trends are as follows: properties (Chapter 9)

16
 The development of the pharmaceutical industry Chapter |1|

70 12

Number of new biopharmaceuticals

60
NCEs marketed in UK

10
50
40 8
30
6
20
10 4
0
1950 1960 1970 1980 1990 2000 2
A Year
0
80

1984

1986

1988

1990

1992

1994

1996

1998

2000
1982
70
Number of NMEs launched

Year
60
50 Fig. 1.4  Growth of biopharmaceuticals. Between 2000 and
40 2010 an average of 4.5 new biopharmaceuticals were
30 introduced each year (see Chapter 22).
20
10 whether new drugs have inherent novelty and represent a
0 significant therapeutic advance, or are merely the result of
1970 1975 1980 1985 1990 1995 2000 one company copying another. There are reasons for
B Year
thinking that new drugs are now more innovative than
they used to be, as illustrated by the introduction of selec-
Fig. 1.3  Productivity – new drugs introduced 1970–2000. tive kinase inhibitors for treating particular types of cancer
(a) Number of new chemical entities (NCEs) marketed in (see Chapter 4). One sign is the rapid growth of biophar-
the UK 1960–1990 showing the dramatic effect of the maceuticals relative to conventional drugs (Figure 1.4) –
thalidomide crisis in 1961. (Data from Griffin (1991) and it has been estimated that by 2014 the top six drugs
International Journal of Pharmacy 5: 206–209.) (b) Annual in terms of sales revenues will all be biologicals (Avastin,
number of new molecular entities marketed in 20 countries
Humira, Rituxan, Enbrel, Lantus and Herceptin) and that
worldwide 1970–1999. The line represents a 5-year moving
average. In the subsequent years up to 2011 the number of biotechnology will have been responsible for 50% of all
new compounds introduced has levelled off to about 24 per new drug introductions (Reuters Fact Box 2010); see also
year (see Chapter 22). Chapters 12 and 22). Most biopharmaceuticals represent
Data from Centre for Medicines Research Pharma R&D Compendium, novel therapeutic strategies, and there may be less scope
2000. for copycat projects, which still account for a substantial
proportion of new synthetic drugs, although biosimilars
are starting to appear. Adding to the disquiet caused by the
• Increased emphasis on ‘drugability’ – mainly centred downward trend in Figure 1.3 is the fact that research and
on pharmacokinetics and toxicology – in the development expenditure has steadily increased over the
selection of lead compounds (Chapter 10) same period, and that development times – from discov-
• Increased use of transgenic animals as disease ery of a new molecule to market – have remained at 10–12
models for drug testing (Chapter 11) years since 1982. Costs, times and success rates are dis-
• The growth of biopharmaceuticals (Chapters 12 cussed in more detail in Chapter 22.
and 13) Nobody really understands why the apparent drop in
• The use of advanced imaging techniques to aid drug research productivity has occurred, but speculations
discovery and development (Chapter 18). abound. One factor may be the increasing regulatory
What effect are these changes having on the success of hurdles, which mean that development takes longer and
the industry in finding new drugs? Despite the difficulty costs more than it used to, so that companies have become
of defining and measuring such success, the answer seems more selective in choosing which compounds to develop.
to be ‘not much so far’. Productivity, measured by the Another factor may be the trend away from ‘me-too’ drugs
flow of new drugs (Figure 1.3) seems to have drifted (drugs that differ little, if at all, from those already in use,
downwards over the last 20 years, and very markedly so but which nevertheless in the past have provided the
since the 1960s, when regulatory controls began to be company with a profitable share of the market while pro-
tightened. This measure, of course, takes no account of viding little or no extra benefit to patients).

17
Section | 1 | Introduction and Background

The hope is that the downward trend will be reversed as In the remainder of this book, we describe drug discov-
the benefit of new technologies to the drug discovery ery at the time of writing (2011) – a time when the molecu-
process works through the system, as long development lar biology revolution is still in full swing. In a few years’
times mean that novel technologies only make an impact time our account will undoubtedly look as dated as the
on registrations some 20 years after their introduction. 1970s scenario seems to us today.

REFERENCES

Drews J. Drug discovery: a historical analysis. Clifton, NJ: Humana Press; companies. Harmondsworth:
perspective. Science 2000;287: 1990. Penguin; 1972.
1960–4. Page C, Schaffhausen J, Shankley NP. Sneader W. Drug discovery: the
Drews J. In quest of tomorrow’s The scientific legacy of Sir James W evolution of modern medicines.
medicines. 2nd ed. New York: Black. Trends in Pharmacological Chichester: John Wiley; 1985.
Springer; 2003. Science 2011;32:181–2. Walker MJA. The major impacts of
Drews J. Paul Ehrlich: magister mundi. Porter R. The greatest benefit to James Black’s drug discoveries on
Nature Reviews Drug Discovery mankind: a medical history of medicine and pharmacology. Trends
2004;3:797–801. humanity from antiquity to the in Pharmacological Science
Lednicer D, editor. Chronicles of drug present. London: Harper Collins; 2011;32:183–8.
discovery. vol 3. Washington, DC: 1997. Weatherall M. In search of a cure: a
American Chemical Society; 1993. Reuters Fact Box. Apr 13th 2010 http:// history of pharmaceutical discovery.
Mann RD. Modern drug use: an enquiry www.reuters.com/article/2010/04/13 Oxford: Oxford University Press;
on historical principles. Lancaster: (accessed July 7th 2011). 1990.
MTP Press; 1984. Sjöstrom H, Nilsson R. Thalidomide
Maxwell RA, Eckhardt SB. Drug and the power of the drug
discovery: a casebook and

18
 Chapter 2 
The nature of disease and the purpose of therapy
H P Rang, R G Hill

steadily accelerating since, has changed our concept of

INTRODUCTION disease quite drastically, and continues to challenge it,
raising new ethical problems and thorny discussions of
In this book, we are concerned mainly with the drug dis- principle. For the centuries of prescientific medicine, codes
covery and development process, proudly regarded as the of practice based on honesty, integrity and professional
mainspring of the pharmaceutical industry. In this chapter relationships were quite sufficient: as therapeutic interven-
we consider the broader context of the human environ- tions were ineffective anyway, it mattered little to what
ment into which new drugs and medicinal products are situations they were applied. Now, quite suddenly, the
launched, and where they must find their proper place. language of disease has changed and interventions have
Most pharmaceutical companies place at the top of their become effective; not surprisingly, we have to revise our
basic mission statement a commitment to improve the ideas about what constitutes disease, and how medical
public’s health, to relieve the human burden of disease intervention should be used. In this chapter, we will try to
and to improve the quality of life. Few would argue with define the scope and purpose of therapeutics in the context
the spirit of this commitment. Nevertheless, we need to of modern biology. In reality, however, those in the
look more closely at what it means, how disease is defined, science-based drug discovery business have to recognize
what medical therapy aims to alter, and how – and by the strong atavistic leanings of many healthcare profes-
whom – the effects of therapy are judged and evaluated. sions1, whose roots go back much further than the age of
Here we outline some of the basic principles underlying science.
these broader issues. Therapeutic intervention, including the medical use of
drugs, aims to prevent, cure or alleviate disease states. The
question of exactly what we mean by disease, and how we
distinguish disease from other kinds of human affliction
CONCEPTS OF DISEASE and dysfunction, is of more than academic importance,
because policy and practice with respect to healthcare
The practice of medicine predates by thousands of years provision depend on where we draw the line between
the science of medicine, and the application of ‘therapeu- what is an appropriate target for therapeutic intervention
tic’ procedures by professionals similarly predates any sci- and what is not. The issue concerns not only doctors, who
entific understanding of how the human body works, or have to decide every day what kind of complaints warrant
what happens when it goes wrong. As discussed in Chapter treatment, the patients who receive the treatment and
1, the ancients defined disease not only in very different all those involved in the healthcare business – including,
terms, but also on a quite different basis from what we of course, the pharmaceutical industry. Much has been
would recognize today. The origin of disease and the meas-
ures needed to counter it were generally seen as manifesta-
1
tions of divine will and retribution, rather than of physical The upsurge of ‘alternative’ therapies, many of which owe nothing
to science – and, indeed, reject the relevance of science to what its
malfunction. The scientific revolution in medicine, which practitioners do – perhaps reflects an urge to return to the
began in earnest during the 19th century and has been prescientific era of medical history.

© 2012 Elsevier Ltd. 19

Section | 1 | Introduction and Background

written on the difficult question of how to define health Health cannot, therefore, be regarded as a definable
and disease, and what demarcates a proper target for thera- state – a fixed point on the map, representing a destination
peutic intervention (Reznek, 1987; Caplan, 1993; Caplan that all are seeking to reach. Rather, it seems to be a con-
et al., 2004); nevertheless, the waters remain distinctly tinuum, through which we can move in either direction,
murky. becoming more or less well adapted for survival in our
One approach is to define what we mean by health, and particular environment. Perhaps the best current defini-
to declare the attainment of health as the goal of all tion is that given by Bircher (2005) who states that ‘health
healthcare measures, including therapeutics. is a dynamic state of wellbeing characterized by physical,
mental and social potential which satisfies the demands
of life commensurate with age, culture and personal
What is health? responsibility’. Although we could argue that the aim of
In everyday parlance we use the words ‘health’, ‘fitness’, healthcare measures is simply to improve our state of
‘wellbeing’ on the one hand, and ‘disease’, ‘illness’, ‘sick- adaptation to our present environment, this is obviously
ness’, ‘ill-health’, etc., on the other, more or less inter- too broad. Other factors than health – for example wealth,
changeably, but these words become slippery and evasive education, peace, and the avoidance of famine – are at
when we try to define them. The World Health Organiza- least as important, but lie outside the domain of medicine.
tion (WHO), for example, defines health as ‘a state of What actually demarcates the work of doctors and health-
complete physical, mental and social wellbeing and not care workers from that of other caring professionals – all
merely the absence of sickness or infirmity’. On this basis, of whom may contribute to health in different ways – is
few humans could claim to possess health, although the that the former focus on disease.
majority may not be in the grip of obvious sickness
or infirmity. Who is to say what constitutes ‘complete
physical, mental and social wellbeing’ in a human being? What is disease?
Does physical wellbeing imply an ability to run a mara- Consider the following definitions of disease:
thon? Does a shy and self-effacing person lack social
wellbeing?
• A condition which alters or interferes with the
normal state of an organism and is usually
We also find health defined in functional terms, less
characterized by the abnormal functioning of one
idealistically than in the WHO’s formulation: ‘…health
or more of the host’s systems, parts or organs
consists in our functioning in conformity with our natural
(Churchill’s Medical Dictionary, 1989).
design with respect to survival and reproduction, as deter-
mined by natural selection…’ (Caplan, 1993). Here the
• A morbid entity characterized usually by at least two
of these criteria: recognized aetiologic agents,
implication is that evolution has brought us to an optimal
identifiable groups of signs and symptoms, or
– or at least an acceptable – compromise with our envi-
consistent anatomical alterations (elsewhere,
ronment, with the corollary that healthcare measures
‘morbid’ is defined as diseased or pathologic)
should properly be directed at restoring this level of func-
(Stedman’s Medical Dictionary, 1990).
tionality in individuals who have lost some important
element of it. This has a fashionably ‘greenish’ tinge, and
• ‘Potential insufficient to satisfy the demands of life’
as outlined by Bircher (2005) in his definition of
seems more realistic than the WHO’s chillingly utopian
health above.
vision, but there are still difficulties in trying to use it as a
guide to the proper application of therapeutics. Environ- We sense the difficulty that these thoughtful authorities
ments differ. A black-skinned person is at a disadvantage found in pinning down the concept. The first definition
in sunless climates, where he may suffer from vitamin D emphasizes two aspects, namely deviation from normality,
deficiency, whereas a white-skinned person is liable to and dysfunction; the second emphasizes aetiology (i.e. caus-
develop skin cancer in the tropics. The possession of ative factors) and phenomenology (signs, symptoms, etc.),
a genetic abnormality of haemoglobin, known as sickle- which is essentially the manifestation of dysfunction.
cell trait, is advantageous in its heterozygous form in
the tropics, as it confers resistance to malaria, whereas
Deviation from normality does not
homozygous individuals suffer from a severe form of
haemolytic anaemia (sickle-cell disease). Hyperactivity
define disease
in children could have survival value in less developed The criterion of deviation from normality begs many
societies, whereas in Western countries it disrupts families questions. It implies that we know what the ‘normal state’
and compromises education. Obsessionality and com­ is, and can define what constitutes an alteration of it. It
pulsive behaviour are quite normal in early motherhood, suggests that if our observations were searching enough,
and may serve a good biological purpose, but in other we could unfailingly distinguish disease from normality.
walks of life can be a severe handicap, warranting medical But we know, for example, that the majority of 50-year-
treatment. olds will have atherosclerotic lesions in their arteries, or

20
 The nature of disease and the purpose of therapy Chapter |2|

that some degree of osteoporosis is normal in postmeno- as pathological, and determined attempts were made to
pausal women. These are not deviations from normality, treat it.
nor do they in themselves cause dysfunction, and so they A definition of disease which tries to combine the con-
do not fall within these definitions of disease, yet both are cepts of biological malfunction and harm (or disvalue)
seen as pathological and as legitimate – indeed important was proposed by Caplan et al. (1981):
– targets for therapeutic intervention. Furthermore, as dis-
cussed below, deviations from normality are often benefi-
’States of affairs are called diseases when they
cial and much prized. are due to physiological or psychological
processes that typically cause states of disability,
pain or deformity, and are generally held to be
Phenomenology and aetiology are
below acceptable physiological or psychological
important factors – the naturalistic view
norms.’
Setting aside the normality criterion, the definitions quoted
above are examples of the naturalistic, or observation-based, What is still lacking is any reference to aetiology, yet this
view of disease, defined by phenomenology and backed up can be important in recognizing disease, and, indeed, is
in many cases by an understanding of aetiology. It is now increasingly so as we understand more about the underly-
generally agreed that this by itself is insufficient, for there ing biological mechanisms. A patient who complains of
is no general set of observable characteristics that distin- feeling depressed may be reacting quite normally to a
guishes disease from health. Although individual diseases bereavement, or may come from a suicide-prone family,
of course have their defining characteristics, which may suggestive of an inherited tendency to depressive illness.
be structural, biochemical or physiological, there is no The symptoms might be very similar, but the implications,
common feature. Further, there are many conditions, based on aetiology, would be different.
particularly in psychiatry, but also in other branches of In conclusion, disease proves extremely difficult to
medicine, where such physical manifestations are absent, define (Scully, 2004). The closest we can get at present to
even though their existence as diseases is not questioned. an operational definition of disease rests on a combina-
Examples would include obsessive-compulsive disorder, tion of three factors: phenomenology, aetiology and dis-
schizophrenia, chronic fatigue syndrome and low back value, as summarized in Figure 2.1.
pain. In such cases, of which there are many examples, the Labelling human afflictions as diseases (i.e. ‘medicaliz-
disease is defined by symptoms of which only the patient ing’ them) has various beneficial and adverse conse-
is aware, or altered behaviour of which he and those around quences, both for the affected individuals and for
him are aware: defining features at the physical, biochemi- healthcare providers. It is of particular relevance to the
cal or physiological level are absent, or at least not yet pharmaceutical industry, which stands to benefit from the
recognized. labelling of borderline conditions as diseases meriting
therapeutic intervention. Strong criticism has been lev-
elled at the pharmaceutical industry for the way in which
Harm and disvalue – the normative view it uses its resources to promote the recognition of ques-
The shortcomings of the naturalistic view of disease, which tionable disorders, such as female sexual dysfunction or
is in principle value free, have led some authors to take social phobia, as diseases, and to elevate identified risk
the opposite view, to the extent of denying the relevance factors – asymptomatic in themselves but increasing the
of any kind of objective criteria to the definition of disease. likelihood of disease occurring later – to the level of dis-
Crudely stated, this value-based (or normative) view holds eases in their own right. A pertinent polemic (Moynihan
that disease is simply any condition the individual or et al., 2004) starts with the sentence: ‘there’s a lot of
society finds disagreeable or harmful (i.e. disvalues). Taken money to be made from telling healthy people they’re
to extremes by authors such as Szasz and Illich, this sick’, and emphasizes the thin line that divides medical
view denies the relevance of the physical manifestations education from marketing (see Chapter 21).
of illness, and focuses instead on illness only as a mani-
festation of social intolerance or malfunction. Although
few would go this far – and certainly modern biologists THE AIMS OF THERAPEUTICS
would not be among them – it is clear that value-laden
judgements play a significant role in determining what
Components of disvalue
we choose to view as disease. In the mid-19th century
masturbation was regarded as a serious disease, to be The discussion so far leads us to the proposition that the
treated if necessary by surgery, and this view persisted proper aim of therapeutic intervention is to minimize the
well into the 20th century. ‘Drapetomania’, defined as disvalue associated with disease. The concept of disvalue
a disease of American slaves, was characterized by an is, therefore, central, and we need to consider what com-
obsessional desire for freedom. Homosexuality was seen prises it. The disvalue experienced by a sick individual has

21
Section | 1 | Introduction and Background

AETIOLOGY

Genetic Environmental
factors factors

Dysfunction

Prognosis:
Diagnostic Symptoms • Non-recovery
signs and • Worsening symptoms
disabilities • Reduced life expectancy

PHENOMENOLOGY DISVALUE

Fig. 2.1  Three components of disease.

two distinct components2 (Figure 2.1), namely present components of disvalue are present and both need to be
symptoms and disabilities (collectively termed morbidity), addressed with therapeutic measures – different measures
and future prognosis (namely the likelihood of increasing may be needed to alleviate morbidity and to improve
morbidity, or premature death). An individual who is suf- prognosis. Of course, such measures need not be confined
fering no abnormal symptoms or disabilities, and whose to physical and pharmacological approaches.
prognosis is that of an average individual of the same age, The proposition at the beginning of this section sets
we call ‘healthy’. An individual with a bad cold or a clear limits to the aims of therapeutic intervention, which
sprained ankle has symptoms and disabilities, but prob- encompass the great majority of non-controversial appli-
ably has a normal prognosis. An individual with asymp- cations. Real life is, of course, not so simple, and in the
tomatic lung cancer or hypertension has no symptoms next section we consider some of the important exceptions
but a poor prognosis. Either case constitutes disease, and and controversies that healthcare professionals and policy-
warrants therapeutic intervention. Very commonly, both makers are increasingly having to confront.

2
These concepts apply in a straightforward way to many real-life Therapeutic intervention is not
situations, but there are exceptions and difficulties. For example, in
certain psychiatric disorders the patient’s judgement of his or her
restricted to treatment or
state of morbidity is itself affected by the disease. Patients suffering prevention of disease
from mania, paranoid delusions or severe depression may pursue an
extremely disordered and self-destructive lifestyle, while denying that The term ‘lifestyle drugs’ is a recent invention, but the
they are ill and resisting any intervention. In such cases, society often
imposes its own judgement of the individual’s morbidity, and may use
concept of using drugs, and other types of interventions,
legal instruments such as the Mental Health Act to apply therapeutic in a medical setting for purposes unrelated to the treat-
measures against the patient’s will. ment of disease is by no means new.
Vaccination represents another special case. Here, the disvalue Pregnancy is not by any definition a disease, nor are skin
being addressed is the theoretical risk that a healthy individual will
later contract an infectious disease such as diphtheria or measles. This wrinkles, yet contraception, abortion and plastic surgery
risk can be regarded as an adverse factor in the prognosis of a are well established practices in the medical domain. Why
perfectly normal individual. are we prepared to use drugs as contraceptives or aborti-
Similarly, a healthy person visiting the tropics will, if he is wise, take
antimalarial drugs to avoid infection – in other words, to improve his facients, but condemn using them to enhance sporting
prognosis. performance? The basic reason seems to be that we attach

22
 The nature of disease and the purpose of therapy Chapter |2|

disvalue to unwanted pregnancy (i.e. we consider it

harmful). We also attach disvalue to alternative means of 90
avoiding unwanted pregnancy, such as sexual abstinence 80
or using condoms. Other examples, however, such as cos-

Life expectancy at birth

70
metic surgery to remove wrinkles or reshape breasts, seem 60
to refute the disvalue principle: minor cosmetic imperfec- 50
tions are in no sense harmful, but society nonetheless
40
concedes to the demand of individuals that medical tech-
30
nology should be deployed to enhance their beauty. In
other cases, such as the use of sildenafil (Viagra) to 20
improve male sexual performance, there is ambivalence 10
about whether its use should be confined to those with 0
evidence for erectile dysfunction (i.e. in whom disvalue 1900 1920 1940 1960 1980 2000
exists) or whether it should also be used in normal men. Year of birth
It is obvious that departures from normality can bring
benefit as well as disadvantage. Individuals with above- Fig. 2.2  Human lifespan in the USA.
average IQs, physical fitness, ball-game skills, artistic Data from National Centre for Health Statistics, 1998.
talents, physical beauty or charming personalities have an
advantage in life. Is it, then, a proper role of the healthcare
system to try to enhance these qualities in the average ultimate social disaster4. A particular consequence of
person? Our instinct says not, because the average person improved survival into old age is an increased incidence
cannot be said to be diseased or suffering. There may be of dementia in the population. It is estimated that some
value in being a talented footballer, but there is no harm 700 000 people in the UK have dementia and world wide
in not being one. Indeed, the value of the special talent prevalence is thought to be over 24 million. The likeli-
lies precisely in the fact that most of us do not possess it. hood of developing dementia becomes greater with age
Nevertheless, a magical drug that would turn anyone into and 1.3% of people in the UK between 65 and 69 suffer
a brilliant footballer would certainly sell extremely well; from dementia, rising to 20% of those over 85. In the UK
at least until footballing skills became so commonplace alone it has been forecast that the number of individuals
that they no longer had any value3. with dementia could reach 1.7 million by 2051 (Nuffield
Football skills may be a fanciful example; longevity is Council on Bioethics, 2009).
another matter. The ‘normal’ human lifespan varies enor-
mously in different countries, and in the West it has
Conclusions
increased dramatically during our own lifetime (Figure
2.2). Is lifespan prolongation a legitimate therapeutic We have argued that that disease can best be defined in
aim? Our instinct – and certainly medical tradition – sug- terms of three components, aetiology, phenomenology
gests that delaying premature death from disease is one of and disvalue, and that the element of disvalue is the most
the most important functions of healthcare, but we are important determinant of what is considered appropriate
very ambivalent when it comes to prolonging life in the to treat. In the end, though, medical practice evolves in a
aged. Our ambivalence stems from the fact that the aged more pragmatic fashion, and such arguments prove to be
are often irremediably infirm, not merely chronologically of limited relevance to the way in which medicine is actu-
old. In the future we may understand better why humans ally practised, and hence to the therapeutic goals the drug
become infirm, and hence more vulnerable to the envi- industry sees as commercially attractive. Politics, econom-
ronmental and genetic circumstances that cause them to ics, and above all, social pressures are the determinants,
become ill and die. And beyond that we may discover how and the limits are in practice set more by our technical
to retard or prevent aging, so that the ‘normal’ lifespan will capabilities than by issues of theoretical propriety.
be much prolonged. Opinions will differ as to whether Although the drug industry has so far been able to take
this will be the ultimate triumph of medical science or the a pragmatic view in selecting targets for therapeutic inter-
vention, things are changing as technology advances. The
increasing cost and sophistication of what therapeutics
3
The use of drugs to improve sporting performance is one of many can offer mean that healthcare systems the world over are
examples of ‘therapeutic’ practices that find favour among
being forced to set limits, and have to go back to the issue
individuals, yet are strongly condemned by society. We do not, as a
society, attach disvalue to the possession of merely average sporting of what constitutes disease. Furthermore, by invoking the
ability, even though the individual athlete may take a different view. concept of disease, governments control access to many
4
Jonathan Swift, in Gulliver’s Travels, writes of the misery of the other social resources (e.g. disability benefits, entry into
Struldbrugs, rare beings with a mark on their forehead who, as they
grew older, lost their youth but never died, and who were declared the armed services, insurance pay-outs, access to life insur-
‘dead in law’ at the age of 80. ance, exemption from legal penalties, etc.).

23
Section | 1 | Introduction and Background

So far, we have concentrated mainly on the impact of for its therapeutic effect may be very indirect. We can see
disease on individuals and societies. We now need to how it has come about that molecular biology, and in
adopt a more biological perspective, and attempt to put particular genomics, has come to figure so largely in the
the concept of disease into the framework of contempo- modern drug discovery environment.
rary ideas about how biological systems work.

Levels of biological organization

FUNCTION AND DYSFUNCTION:   Figure 2.3 shows schematically the way in which the
THE BIOLOGICAL PERSPECTIVE genetic constitution of a human being interacts with his
or her environment to control function at many different
levels, ranging from protein molecules, through single
The dramatic revelations of the last few decades about the cells, tissues and integrated physiological systems, to the
molecular basis of living systems have provided a new way individual, the family and the population at large. For
of looking at function and dysfunction, and the nature of simplicity, we will call this the bioaxis. ‘Disease’, as we have
disease. Needless to say, molecular biology could not have discussed, consists of alterations of function sufficient to
developed without the foundations of scientific biology cause disability or impaired prognosis at the level of the
that were built up in the 19th century. As we saw in individual. It should be noted that the arrows along the
Chapter 1, this was the period in which science came to bioaxis in Figure 2.3 are bidirectional – that is, distur-
be accepted as the basis on which medical practice had to bances at higher levels of organization will in general
be built. Particularly significant was cell theory, which affect function at lower levels, and vice versa. Whereas it
established the cell as the basic building block of living is obvious that genetic mutations can affect function
organisms. In the words of the pioneering molecular bio­ further up the bioaxis (as in many inherited diseases, such
logist, François Jacob: ‘With the cell, biology discovered its as muscular dystrophy, cystic fibrosis or thalassaemia), we
atom’. It is by focusing on the instruction sets that define should not forget that environmental influences also affect
the form and function of cells, and the ways in which these gene function. Indeed, we can state that any long-term
instructions are translated in the process of generating the phenotypic change (such as weight gain, muscle weakness
structural and functional phenotypes of cells, that molecu- or depressed mood) necessarily involves alterations of gene
lar biology has come to occupy centre stage in modern expression. For example:
biology. Genes specify proteins, and the proteins a cell
produces determine its structure and function. • Exposure to a stressful environment will activate the
hypothalamopituitary system and thereby increase
From this perspective, deviations from the norm, in
adrenal steroid secretion, which in turn affects gene
terms of structure and function at the cellular level, arise
transcription in many different cells and tissues,
through deviations in the pattern of protein expression by
affecting salt metabolism, immune responses and
individual cells, and they may arise either through faults
many other functions.
in the instruction set itself (genetic mutations) or through
environmental factors that alter the way in which the • Smoking, initiated as result of social factors such as
peer pressure or advertising, becomes addictive as a
instruction set is translated (i.e. that affect gene expres-
result of changes in brain function, phenotypic
sion). We come back to the age-old distinction between
changes which are in turn secondary to altered gene
inherited and environmental factors (nature and nurture)
expression.
in the causation of disease, but with a sharper focus:
altered gene expression, resulting in altered protein syn- • Exposure to smoke carcinogens then increases the
probability of cancer-causing mutations in the DNA
thesis, is the mechanism through which all these factors
of the cells of the lung. The mutations, in turn, result
operate. Conversely, it can be argued5 that all therapeutic
in altered protein synthesis and malignant
measures (other than physical procedures, such as surgery)
transformation, eventually producing a localized
also work at the cellular level, by influencing the same
tumour and, later, disseminated cancer, with damage
fundamental processes (gene expression and protein syn-
to the function of tissues and organs leading to
thesis), although the link between a drug’s primary target
symptoms and premature death.
and the relevant effect(s) on gene expression that account
The pathogenesis of any disease state reveals a similar
level of complexity of such interactions between different
5
This represents the ultimate reductionist view of how living levels of the bioaxis.
organisms work, and how they respond to external influences. Many
still hold out against it, believing that the ‘humanity’ of man demands There are two important conclusions to be drawn from
a less concrete explanation, and that ‘alternative’ systems of the bidirectionality of influence between events at dif­
medicine, not based on our scientific understanding of biological ferent levels of the bioaxis. One is that it is difficult to
function, have equal validity. Many doctors apparently feel most
comfortable somewhere on the middle ground, and society at large pinpoint the cause of a given disease. Do we regard the
tends to fall in behind doctors rather than scientists. cause of lung cancer in an individual patient as the lack

24
 The nature of disease and the purpose of therapy Chapter |2|

Inherited Environmental
characteristics factors

Physiological
Gene Protein Cell Tissue Individual Family Society
system

Mutation Amount Growth Growth Hypoxia Symptoms Relationships Disease prevalence

Expression Function Shape Atrophy Hyperglycaemia Disabilities Finances Healthcare costs
Differentiation Repair Hypertension Prognosis Activities Social costs
Movement Inflammation Heart failure Life expectancy Care burden etc.
Division Infection Renal failure
Apoptosis Ischaemia Epilepsy
Metabolism Necrosis Hemiplegia
Secretion etc. Obesity
Excitability etc.
etc.

Drugs act here Therapeutic aims directed here

Fig. 2.3  The nature of disease.

of control over tobacco advertising, the individual’s sus- together with the ‘proteome’ (which describes the array of
ceptibility to advertising and peer pressure, the state of proteins present in a cell or tissue), provides a uniquely
addiction to nicotine, the act of smoking, the mutational detailed description of how the cell or tissue is behaving.
event in the lung epithelial cell, or the individual’s inher- Molecular biology is providing us with powerful methods
ited tendency to lung cancer? There is no single answer, for mapping the changes in gene and protein expression
and the uncertainty should make us wary of the stated associated with different functional states – including
claim of many pharmaceutical companies that their aim disease states and therapeutic responses – and we discuss
is to correct the causes rather than the symptoms of in more detail in Chapters 6 and 7 the way these new
disease. The truth, more often than not, is that we cannot windows on function are influencing the drug discovery
distinguish them. Rather, the aim should be to intervene process (see Debouck and Goodfellow, 1999).
in the disease process in such a way as to minimize the
disvalue (disability and impaired prognosis) experienced
Therapeutic targets
by the patient.
The second conclusion is that altered gene expression Traditionally medicine has regarded the interests of the
plays a crucial role in pathogenesis and the production of individual patient as paramount, putting them clearly
any long-term phenotypic change. If we are thinking of ahead of those of the community or general population.
timescales beyond, at maximum, a few hours, any change The primacy of the patient’s interests remains the guiding
in the structure and function of cells and tissues will be principle for the healthcare professions; in other words,
associated with changes in gene expression. These changes their aim is to address disvalue as experienced by the
will include those responsible for the phenotypic change patient, not to correct biochemical abnormalities, nor
(e.g. upregulation of cytokine genes in inflammation, to put right the wrongs of society. The principal aim of
leading to leukocyte accumulation), and those that are therapeutic intervention, as shown in Figure 2.3, is there-
consequences of it (e.g. loss of bone matrix following fore to alleviate the condition of the individual patient.
muscle paralysis); some of the latter will, in turn, lead to Genetic, biochemical or physiological deviations which
secondary phenotypic changes, and so on. The pattern of are not associated with any disvalue for the patient (e.g.
genes expressed in a cell or tissue (sometimes called the possession of a rare blood group, an unusually low heart
‘transcriptome’, as distinct from the ‘genome’, which rep- rate or blood pressure, or blood cholesterol concentra-
resents all of the genes present, whether expressed or not), tion) are not treated as diseases because they neither cause

25
Section | 1 | Introduction and Background

symptoms nor carry an unfavourable prognosis. High show how disease affects function at the level of cells,
blood pressure, or high blood cholesterol, on the other tissues, organs and individuals. The links between the two
hand, do confer disvalue because they carry a poor prog- nevertheless remain tenuous, a fact which greatly limits
nosis, and are targets for treatment – surrogate targets, in our ability to relate drug targets to therapeutic effects. Not
the sense that the actual aim is to remedy the unfavourable surprisingly, attempts to bridge this Grand Canyon form
prognosis, rather than to correct the physiological abnor- a major part of the work of many pharmaceutical and
mality per se. biotechnology companies. Afficionados like to call them-
Although the present and future wellbeing of the selves ‘postgenomic’ biologists; Luddites argue that they
individual patient remains the overriding priority for are merely coming down from a genomic ‘high’ to face
medical care, the impact of disease is felt not only by once more the daunting complexities of living organisms.
individuals, but also by society in general, partly for eco- We patient realists recognize that a biological revolution
nomic reasons, but also for ideological reasons. Reducing has happened, but do not underestimate the time and
the overall burden of disease, as measured by rates of money needed to bridge the canyon. More of this later.
infant mortality, heart disease or AIDS, for example, is a
goal for governments throughout the civilized world, akin
to the improvement of educational standards. The disease-
related disvalue addressed in this case, as shown by the
THERAPEUTIC INTERVENTIONS
secondary arrow in Figure 2.3, is experienced at the
national, rather than the individual level, for individuals Therapeutics in its broadest sense covers all types of inter-
will in general be unaware of whether or not they have vention aimed at alleviating the effects of disease. The term
benefited personally from disease prevention measures. As ‘therapeutics’ generally relates to procedures based on
the therapeutic target has come to embrace the population accepted principles of medical science, that is, on ‘conven-
as a whole, so the financial burden of healthcare has tional’ rather than ‘alternative’ medical practice6. The
shifted increasingly from individuals to institutional pro- account of drug discovery presented in this book relates
viders of various kinds, mainly national agencies or large- exclusively to conventional medicine – and for this we
scale commercial healthcare organizations. Associated make no apology – but it needs to be realized that the
with this change, there has been a much more systematic therapeutic landscape is actually much broader, and
focus on assessment in economic terms of the burden of includes many non-pharmacological procedures in the
disease (disvalue, to return to our previous terminology) domain of conventional medicine, as well as quasi-
in the community, and the economic cost of healthcare pharmacological practices (e.g. homeopathy and herbal-
measures. The new and closely related disciplines of phar- ism) in the ‘alternative’ domain.
macoeconomics and pharmacoepidemiology, discussed later, As discussed above, the desired effect of any therapeutic
reflect the wish (a) to quantify disease-related disvalue and interventions is to improve symptoms or prognosis or both.
therapeutic benefit in economic terms, and (b) to assess From a pathological point of view, therapeutic interven-
the impact of disease and therapy for the population as a tions may be directed at disease prevention, alleviation of the
whole, and not just for the individual patient. effects of existing disease, or permanent cure (i.e. restora-
tion to a state of function and prognosis equivalent to
The relationship between drug those of a healthy individual of the same age, without
the need for continuing therapeutic intervention). In
targets and therapeutic targets practice, there are relatively few truly curative interven-
There are very few exceptions to the rule, shown in Figure tions, and they are mainly confined to certain surgical
2.3, that protein molecules are the primary targets of drug procedures (e.g. removal of circumscribed tumours, fixing
molecules. We will come back to this theme repeatedly of broken bones) and chemotherapy of some infectious
later, because of its prime importance for the drug discov- and malignant disorders. Most therapeutic interventions
ery process. We should note here that many complex bio- aim to alleviate symptoms and/or improve prognosis, and
logical steps intervene between the primary drug target there is increasing emphasis on disease prevention as an
and the therapeutic target. Predicting, on the one hand, objective.
whether a drug that acts specifically on a particular protein It is important to realize that many types of inter­
will produce a worthwhile therapeutic effect, and in what ventions are carried out with therapeutic intent whose
disease state, or, on the other hand, what protein we efficacy has not been rigorously tested. This includes not
should choose to target in order to elicit a therapeutic
effect in a given disease state, are among the thorniest 6
Scientific doctors rail against the term ‘alternative’, arguing that if a
problems for drug discoverers. Molecular biology is pro- therapeutic practice can be shown to work by properly controlled
viding new insights into the nature of genes and proteins trials, it belongs in mainstream medicine. If such trials fail to show
efficacy, the practice should not be adopted. Paradoxically, whereas
and the relationship between them, whereas time- ‘therapeutics’ generally connotes conventional medicine, the term
honoured biochemical and physiological approaches can ‘therapy’ tends to be used most often in the ‘alternative’ field.

26
 The nature of disease and the purpose of therapy Chapter |2|

only the myriad alternative medical practices, but also that now regulate the provision of healthcare, such as
many accepted conventional therapies for which a good formulary committees, insurance companies, health man-
scientific basis may exist but which have not been sub- agement organizations, and bodies such as the grandly
jected to rigorous clinical trials. titled National Institute for Health and Clinical Excellence
(NICE), set up by the UK Government in 1999 to advise,
on the basis of cost-effectiveness, which drugs and other
therapeutic procedures should be paid for under the
MEASURING THERAPEUTIC National Health Service.
OUTCOME Benefit comprises effectiveness expressed in monetary
terms. It is popular with economists, as it allows cost
Effect, efficacy, effectiveness   and benefit to be compared directly, but treated with
deep suspicion by many who find the idea of assigning
and benefit monetary value to life and wellbeing fundamentally
These terms have acquired particular meanings – more abhorrent.
limited than their everyday meanings – in the context of Returning to the theme of Figure 2.3, we can see that
therapeutic trials. whereas effect and efficacy are generally measured at the
Pharmacological effects of drugs (i.e. their effects on level of cells, tissues, systems and individuals, effectiveness
cells, organs and systems) are, in principle, simple to and benefit are measures of drug action as it affects popula-
measure in animals, and often also in humans. We can tions and society at large. We next consider two growing
measure effects on blood pressure, plasma cholesterol disciplines that have evolved to meet the need for informa-
concentration, cognitive function, etc., without difficulty. tion at these levels, and some of the methodological prob-
Such measures enable us to describe quantitatively the lems that they face.
pharmacological properties of drugs, but say nothing
about their usefulness as therapeutic agents.
Efficacy describes the ability of a drug to produce a
PHARMACOEPIDEMIOLOGY AND
desired therapeutic effect in patients under carefully con-
trolled conditions. The gold standard for measurements PHARMACOECONOMICS
of efficacy is the randomized controlled clinical trial,
described in more detail in Chapter 17. The aim is to Pharmacoepidemiology (Strom, 2005) is the study of the
discover whether, based on a strictly defined outcome use and effects of drugs in human populations, as distinct
measure, the drug is more or less beneficial than a stand- from individuals, the latter being the focus of clinical
ard treatment or placebo, in a selected group of patients, pharmacology. The subject was born in the early 1960s,
under conditions which ensure that the patients actually when the problem of adverse drug reactions came into
receive the drug in the specified dose. Proof of efficacy, as prominence, mainly as a result of the infamous thalido-
well as proof of safety, is required by regulatory authorities mide disaster. The existence of rare but serious adverse
as a condition for a new drug to be licensed. Efficacy tests drug reactions, which can be detected only by the study of
what the drug can do under optimal conditions, which is large numbers of subjects, was the initial stimulus for the
what the prescriber usually wants to know. development of pharmacoepidemiology, and the detec-
Effectiveness describes how well the drug works in real tion of adverse drug reactions remains an important
life, where the patients are heterogeneous, are not rand- concern. The identification of Reye’s syndrome as a serious,
omized, are aware of the treatment they are receiving, are albeit rare, consequence of using aspirin in children is just
prescribed different doses, which they may or may not one example of a successful pharmacoepidemiological
take, often in combination with other drugs. The desired study carried out under the auspices of the US Department
outcome is generally less well defined than in efficacy of Health and published in 1987. The subject has gradu-
trials, related to general health and freedom from symp- ally become broader, however, to cover aspects such as the
toms, rather than focusing on a specific measure. The focus variability of drug responses between individuals and
is not on the response of individual patients under con- population groups, the level of compliance of individual
trolled conditions, but on the overall usefulness of the patients in taking drugs that are prescribed, and the overall
drug in the population going about its normal business. impact of drug therapies on the population as a whole,
Studies of effectiveness are of increasing interest to the taking all of these factors into account. The widely used
pharmaceutical companies themselves, because effective- antipsychotic drug clozapine provides an interesting
ness rather than efficacy alone ultimately determines how example of the importance of pharmacoepidemiological
well the drug will sell, and because effectiveness may issues in drug evaluation. Clozapine, first introduced in
depend to some extent on the companies’ marketing strat- the 1970s, differed from its predecessors, such as haloperi-
egies (see Chapter 21). Effectiveness measures are also dol, in several ways, some good and some bad. On the
becoming increasingly important to the many agencies good side, clozapine has a much lower tendency than

27
Section | 1 | Introduction and Background

haloperidol to cause extrapyramidal motor effects (a et al. (1996), Johannesson (1996) and McCombs (1998).
serious problem with many antipsychotic drugs), and it The aim of pharmacoeconomics is to measure the benefits
appeared to have the ability to improve not only the posi- and costs of drug treatments, and in the end to provide a
tive symptoms of schizophrenia (hallucinations, delu- sound basis for comparing the value for money of differ-
sions, thought disorder, stereotyped behaviour) but also ent treatments. As might be expected, the subject arouses
the negative symptoms (social withdrawal, apathy). Com- fierce controversy. Economics in general is often criticized
pliance is also better with clozapine, because the patient for defining the price of everything but appreciating the
usually has fewer severe side effects. On the bad side, in value of nothing, and health economics particularly tends
about 1% of patients clozapine causes a fall in the blood to evoke this reaction, as health and quality of life are such
white cell count (leukopenia), which can progress to an ill-defined and subjective, yet highly emotive, concepts.
irreversible state of agranulocytosis unless the drug is Nevertheless, pharmacoeconomics is a rapidly growing
stopped in time. Furthermore, clozapine does not produce discipline and will undoubtedly have an increasing influ-
benefit in all schizophrenic patients – roughly one-third ence on healthcare provision.
fail to show improvement, and there is currently no way Pharmacoeconomic evaluation of new drugs is often
of knowing in advance which patients will benefit. Cloza- required by regulatory authorities, and is increasingly
pine is also more expensive than haloperidol. Considered being used by healthcare providers as a basis for choosing
from the perspective of an individual patient, and with how to spend their money. Consequently, pharmaceutical
hindsight, it is straightforward to balance the pros and companies now incorporate such studies into the clinical
cons of using clozapine rather than haloperidol, based on trials programmes of new drugs. The trend can be seen as
the severity of the extrapyramidal side effects, the balance a gradual progression towards the right-hand end of the
of positive and negative symptoms that the patient has, bioaxis in Figure 2.3 in our frame of reference for assessing
whether clozapine is affecting the white cell count, and the usefulness of a new drug. Before 1950, new drugs were
whether the patient is a responder or a non-responder. often introduced into clinical practice on the basis of
From the perspective of the overall population, evaluating studies in animals and a few human volunteers; later,
the pros and cons of clozapine and haloperidol (or indeed formal randomized controlled clinical trials on carefully
of any two therapies) requires epidemiological data: how selected patient populations, with defined outcome meas-
frequent are extrapyramidal side effects with haloperidol, ures, became the accepted standard, along with postmar-
what is the relative incidence of positive and negative keting pharmacoepidemiological studies to detect adverse
symptoms, what is the incidence of agranulocytosis with reactions. Pharmacoeconomics represents the further shift
clozapine, what proportion of patients are non-responders, of focus to include society in general and its provisions for
what is the level of patient compliance with haloperidol healthcare. A brief outline of the main approaches used in
and clozapine? pharmacoeconomic analysis follows.
In summary, pharmacoepidemiology is a special area of Pharmacoeconomics covers four levels of analysis:
clinical pharmacology which deals with population, rather
than individual, aspects of drug action, and provides the
• Cost identification
means of quantifying variability in the response to drugs.
• Cost-effectiveness analysis
Its importance for the drug discovery process is felt mainly
• Cost-utility analysis
at the level of clinical trials and regulatory affairs, for two
• Cost–benefit analysis.
reasons (Dieck et al., 1994). First, allowing for variability Cost identification consists of determining the full cost in
is essential in drawing correct inferences from clinical monetary units of a particular therapeutic intervention,
trials (see Chapter 17). Second, variability in response to including hospitalization, working days lost, etc., as well
a drug is per se disadvantageous, as drug A, whose effects as direct drug costs. It pays no attention to outcome,
are unpredictable, is less useful than drug B which acts and its purpose is merely to allow the costs of different
consistently, even though the mean balance between ben- procedures to be compared. The calculation is straight­
eficial and unwanted effects may be the same for both. forward, but deciding exactly where to draw the line
From the population perspective, drug B looks better than (e.g. whether to include indirect costs, such as loss of
drug A, even though for many individual patients the income by patients and carers) is somewhat arbitrary.
reverse may be true. Nevertheless, cost identification is the least problematic
Pharmacoeconomics, a branch of health economics, is part of pharmacoeconomics.
a subject that grew up around the need for healthcare Cost-effectiveness analysis aims to quantify outcome as
providers to balance the ever-growing costs of healthcare well as cost. This is where the real problems begin. The
against limited resources. The arrival of the welfare state, outcome measure most often used in cost-effectiveness
which took on healthcare provision as a national rather analysis is based on prolongation of life, expressed as life-
than an individual responsibility, was the signal for econo- years saved per patient treated. Thus if treatment prolongs
mists to move in. Good accounts of the basic principles the life expectancy of patients, on average, from 3 years to
and their application to pharmaceuticals are given by Gold 5 years, the number of life-years gained per patient is 2.

28
 The nature of disease and the purpose of therapy Chapter |2|

Comparing cost and outcome for different treatments then

Perfect
allows the cost per life-year saved to be determined for health 1
each. For example, a study of various interventions in Normal
coronary heart disease, cited by McCombs (1998), showed healthy
individual
that the cost per life-year saved was $5900 for use of
a β-adrenoceptor blocker in patients who had suffered QALYs
gained Treated
a heart attack, the corresponding figure for use of a

Quality of life
type II
cholesterol-lowering drug in patients with coronary heart diabetes
0.5
disease was $7200, while coronary artery bypass surgery
cost $34 000 per life-year saved. As these drugs have
reached the end of their patent life and become low-priced
Untreated Theoretical
generic medicines, the cost difference changes in favour of ‘ideal’
type II
their use. Any kind of all-or-nothing event, such as prema- diabetes
ture births prevented, hospital admissions avoided, etc.,
can be used for this kind of analysis. Its weakness is that Death 0
it is a very crude measure, making no distinction between 0 20 40 60 80 100
Years
years of life spent in a healthy and productive mode and
years of life spent in a state of chronic illness.
Cost-utility analysis is designed to include allowance for Fig. 2.4  Quality of life affected by disease and treatment.
quality of life, as well as survival, in the calculation, and
is yet more controversial, for it becomes necessary
somehow to quantify quality – not an endeavour for the
which, to the untrained observer, have a distinctly chilling
faint-hearted. What the analysis seeks to arrive at is an
and surreal quality. For example, the standard gamble
estimate known as quality-adjusted life-years (QALYs). Thus
approach, which is well grounded in the theory of welfare
if the quality of life for a given year, based on the results
economics, involves asking the individual a question of
of the questionnaire, comes out at 70% of the value for an
the following kind:
average healthy person of the same age, that year repre-
sents 0.7 QALYs, compared with 1 QALY for a year spent Imagine you have the choice of remaining in your
in perfect health, the assumption being that 1 year spent present state of health for 1 year or taking a gamble
at this level of illness is ‘worth’ 0.7 years spent in perfect between dying now and living in perfect health for
health. 1 year. What odds would you need to persuade you
to take the gamble?7
Many different questionnaire-based rating scales have
been devised to reflect different aspects of an individual’s If the subject says 50 : 50, the implication is that he
state of health or disability, such as ability to work, mobil- values a year of life in his present state of health at 0.5
ity, mental state, pain, etc. Some relate to specific disease QALYs. An alternative method involves asking the patient
conditions, whereas others aim to provide a general how many years of life in their present condition he or she
‘quality-of-life’ estimate (Jaeschke and Guyatt, 1994), would be prepared to forfeit in exchange for enjoying
some of the best-known being the Sickness Impact Profile, good health until they die. Although there are subtle ways
the Nottingham Health Profile, the McMaster Health Index, of posing this sort of question, such an evaluation, which
and a 36-item questionnaire known as SF-36. In addition most ordinary people find unreal, is implicit in the QALY
to these general quality-of-life measures, a range of disease- concept. Figure 2.4 shows schematically the way in which
specific questionnaires have been devised which give quality of life, as a function of age, may be affected by
greater sensitivity in measuring the specific deficits asso­ disease and treatment, the area between the curves for
ciated with particular diseases. Standard instruments of untreated and treated patients representing the QALYs
this kind are now widely used in pharmacoeconomic saved by the treatment. In reality, of course, continuous
studies. measurements spanning several decades are not possible,
To use such ratings in estimating QALYs it is necessary so the actual data on which QALY estimates are based in
to position particular levels of disability on a life/death practice are much less than is implied by the idealized
scale, such that 1 represents alive and in perfect health and diagram in Figure 2.4. Cost-utility analysis results in an
0 represents dead. This is where the problems begin in estimate of monetary cost per QALY gained and it is
earnest. How can we possibly say what degree of pain is becoming widely accepted as a standard method for
equivalent to what degree of memory loss, for example, pharmacoeconomic analysis. Examples of cost per QALY
or how either compares with premature death? This
problem has, of course, received a lot of expert attention
(Gold et al., 1996; Johannesson, 1996; Drummond et al., 7
Imagine being asked this by your doctor! ‘But I only wanted
1997) and various solutions have been proposed, some of something for my sore throat’, you protest weakly.

29
Section | 1 | Introduction and Background

gained range from £3700 for the use of sildenafil (Viagra) As a result, cost–benefit analysis has been largely shunned
in treating erectile dysfunction (Stolk et al., 2000) to as a practical approach for evaluating medicines, but may
£328 000 for the treatment of multiple sclerosis with be unavoidable as more personalized and expensive medi-
β-interferon (Parkin et al., 2000), this high value being cines become available (Hunter and Wilson, 2011 and
accounted for by the high cost and limited therapeutic Chapter 22).
efficacy of the drug. ‘Acceptable’ thresholds for cost-
effectiveness are suggested to be in the range of £8000–
£30 000 per QALY gained (Hunter and Wilson, 2011). It is
hard, if not impossible, to make sure that available funds SUMMARY
are spent in the most appropriate way. For example,
Avastin (bevacizumab) is a monoclonal antibody used in In this chapter we have discussed concepts of disease and
the treatment of colorectal cancer and NICE estimate that the aims of therapeutics, the needs newly introduced drugs
some 6500 patients per year would be eligible for treat- have to satisfy, and the ways in which their ability to satisfy
ment with this drug. The cost of a course of treatment is those needs are judged in practice. There are many uncer-
£20 800 per patient and overall average survival is increased tainties and ambiguities surrounding the definition of
by between 1.4 and 2.2 months depending on whether it disease, and ideas are constantly shifting, but the two
is being used for first- or second-line therapy. Funding the components that most satisfactorily define it are dysfunc-
use of Avastin in this way would cost £135 million per tion and disvalue. Disvalue, which therapeutic interven-
year or >1% of the total NHS drugs budget of £11 billion tions aim to mitigate, in turn has two main components,
per year (Hunter and Wilson, 2011). In principle, cost- namely morbidity and prognosis.
utility analysis allows comparison of one form of treat- We have described the bioaxis, which represents the
ment against another, and this explains its appeal to those various levels in the organizational heirarchy of living
who must make decisions about the allocation of health- systems in general, and human beings in particular, and
care resources. It has been adopted as the method of emphasized that disease inevitably affects all levels on the
choice for pharmacoeconomic analysis of new medicines bioaxis. The drugs that we invent home in very specifically
by several agencies, such as the US Public Health Service on one level, namely proteins, although the effects we
and the Australian Pharmaceutical Benefits Advisory want to produce are at another level, namely individuals.
Committee. Furthermore, we emphasize that healthcare issues are
Hardline economists strive for an absolute scale by increasingly being viewed from the perspective of popula-
which to judge the value of healthcare measures compared tions and societies, and so the impact of drugs at these
with other resource-consuming initiatives that societies levels – even further removed from their primary targets
choose to support. Cost–benefit analysis fulfils this need in – has to be evaluated. Evaluation of drug effects from these
principle, by translating healthcare improvements into rather lofty perspectives, through the application of the
monetary units that can be directly balanced against costs, emerging disciplines of pharmacoepidemiology and phar-
to assess whether any given procedure is, on balance, ‘prof- macoeconomics, although fraught with problems, is an
itable’. The science of welfare economics has provided important trend the pharmaceutical industry cannot
various tools for placing a monetary value on different ignore.
experiences human beings find agreeable or disagreeable, Having taken this rather nervy look at the world about
based generally on the ‘willingness-to-pay’ principle. Not us, we turn in the next chapter to discuss the different
surprisingly, attempts to value human life and health in therapeutic modalities on which the pharmaceutical and
cash terms lead rapidly into an ethical and moral mine- biotechnology industries have focused, before retreating
field, dangerous enough in the context of a single nation to the safer ground at the left-hand end of the bioaxis,
and its economy, but much more so in the global context. where the modern-day drug discovery business begins.

REFERENCES

Bircher J. Towards a dynamic definition Caplan AL. The concepts of health, DC: Georgetown University Press;
of health and disease. Medical illness, and disease. In: Bynum WF, 2004.
Health Care and Philosophy Porter R, editors. Companion Debouck C, Goodfellow PN. DNA
2005;8:335–41. encyclopedia of the history of microarrays in drug discovery and
Caplan AL, Engelhardt Jr HT, McCartney medicine. vol. 1. London: Routledge; development. Nature Genetics
JJ, editors. Concepts of health and 1993. 1999;(Suppl. 21):48–50.
disease: interdisciplinary Caplan AL, McCartney JJ, Sisti DA, Dieck GS, Glasser DB, Sachs RM.
perspectives. London: Addison- editors. Health, disease, and illness: Pharmacoepidemiology: a view from
Wesley; 1981. concepts in medicine. Washington, industry. In: Strom BL, editor.

30
 The nature of disease and the purpose of therapy Chapter |2|

Pharmacoepidemiology. Chichester: pharmacoepidemiology research. In: Parkin D, Jacoby A, McNamee P, et al.

John Wiley; 1994. p. 73–85. Strom BL, editor. Treatment of multiple sclerosis with
Drummond MF, O’Brien B, Stoddart GI, Pharmacoepidemiology. Chichester: interferon beta: an appraisal of
et al. Methods for the economic John Wiley; 1994. p. 495–505. cost-effectiveness and quality of life.
evaluation of healthcare Johannesson M. Theory and methods of Journal of Neurology, Neurosurgery
programmes. Oxford: Oxford economic evaluation of health care. and Psychiatry 2000;68:144–9.
University Press; 1997. Dordrecht: Kluwer Academic; 1996. Reznek L. The nature of disease. New
Gold MR, Siegel JE, Russell LB, et al, McCombs JS. Pharmacoeconomics: York: Routledge & Kegan Paul; 1987.
editors. Cost-effectiveness in health what is it and where is it going? Scully JL. What is a disease? EMBO
and medicine. New York: Oxford American Journal of Hypertension Reports 2004;5:650–3.
University Press; 1996. 1998;11:112S–9S. Stolk EA, Busschbach JJ, Caffa M, et al.
Hunter D, Wilson J. Hyper-expensive Moynihan R, Heath I, Henry D. Selling Cost utility analysis of sildenafil
treatments. Background paper for sickness: the pharmaceutical industry compared with papaverine-
Forward Look 2011. London: and disease mongering. British phentolamine injections. British
Nuffield Council on Bioethics; 2011. Medical Journal 2004;324:886–91. Medical Journal 2000;320:1156–7.
p. 23. Nuffield Council on Bioethics. Strom BL, editor.
Jaeschke R, Guyatt GH. Using quality- Dementia – ethical issues. London: Pharmacoepidemiology 4th Edition.
of-life measurements in NCOB; 2009. p. 172. Chichester: John Wiley; 2005.

31
 Chapter 3 
Therapeutic modalities
H P Rang, H LeVine, R G Hill

accepted conventional therapies for which a sound scien-

INTRODUCTION tific basis may exist but which have not been subjected to
rigorous clinical trials.
Therapeutic interventions that lie within the field of
Therapeutics in its broadest sense covers all types of
conventional medicine can be divided into the following
intervention aimed at alleviating the effects of disease.
broad categories:
The term ‘therapeutics’ generally relates to procedures
based on accepted principles of medical science, that • Advice and counselling (e.g. genetic counselling)
is, on ‘conventional’ rather than ‘alternative’ medical • Psychological treatments (e.g. cognitive therapies for
practice. anxiety disorders, depression, etc.)
The account of drug discovery presented in this book • Dietary and nutritional treatments (e.g. gluten-free
relates exclusively to conventional medicine – and for diets for coeliac disease, diabetic diets, etc.)
this we make no apology – but it needs to be realized • Physical treatments, including surgery, radiotherapy
that the therapeutic landscape is actually much broader, • Pharmacological treatments – encompassing the
and includes many non-pharmacological procedures in whole of conventional drug therapy
the domain of conventional medicine, as well as quasi- • Biological and biopharmaceutical treatments,
pharmacological practices in the ‘alternative’ domain. a broad category including vaccination,
As discussed in Chapter 2, the desired effect of any transplantation, blood transfusion,
therapeutic intervention is to improve symptoms or progno- biopharmaceuticals (see Chapters 12 and 13), in
sis or both. From a pathological point of view, therapeutic vitro fertilization, etc.
interventions may be directed at disease prevention, allevia- On the fringe of conventional medicine are prepara-
tion of the effects of existing disease, or permanent cure tions that fall into the category of ‘nutriceuticals’ or
(i.e. restoration to a state of function and prognosis equiv- ‘cosmeceuticals’. Nutriceuticals include a range of dietary
alent to those of a healthy individual of the same age, preparations, such as slimming diets, and diets supple-
without the need for continuing therapeutic intervention). mented with vitamins, minerals, antioxidants, unsaturated
In practice, there are relatively few truly curative interven- fatty acids, fibre, etc. These preparations generally have
tions, and they are mainly confined to certain surgical some scientific rationale, although their efficacy has not,
procedures (e.g. removal of circumscribed tumours, fixing in most cases, been established by controlled trials. They
of broken bones) and chemotherapy of some infectious are not subject to formal regulatory approval, so long as
and malignant disorders. Most therapeutic interventions they do not contain artificial additives other than those
aim to alleviate symptoms and/or improve prognosis, and that have been approved for use in foods. Cosmeceuticals
there is increasing emphasis on disease prevention as an is a fancy name for cosmetic products similarly supple-
objective. mented with substances claimed to reduce skin wrinkles,
It is important to realize that many types of interven- promote hair growth, etc. These products achieve very
tion are carried out with therapeutic intent whose efficacy large sales, and some pharmaceutical companies have
has not been rigorously tested. This includes not only expanded their business in this direction. We do not
the myriad alternative medical practices, but also many discuss these fringe ‘ceuticals’ in this book.

© 2012 Elsevier Ltd. 33

Section | 1 | Introduction and Background

Within each of the medical categories listed above lies

a range of procedures: at one end of the spectrum are
Box 3.1  Advantages and disadvantages of
procedures that have been fully tried and tested and are small-molecule drugs
recognized by medical authorities; at the other is outright
Advantages
quackery of all kinds. Somewhere between lie widely used
• ‘Chemical space’ is so vast that synthetic chemicals,
‘complementary’ procedures, practised in some cases
according to many experts, have the potential to bind
under the auspices of officially recognized bodies, which
specifically to any chosen biological target: the right
have no firm scientific foundation. Here we find, among
molecule exists; it is just a matter of finding it.
psychological treatments, hypnotherapy and analytical
• Doctors and patients are thoroughly familiar with
psychotherapy; among nutritional treatments, ‘health
conventional drugs as medicines, and the many
foods’, added vitamins, and diets claimed to avoid ill-
different routes of administration that are available.
defined food allergies; among physical treatments, acu-
Clinical pharmacology in its broadest sense has
puncture and osteopathy; among chemical treatments, become part of the knowledge base of every
homeopathy, herbalism and aromatherapy. Biological practising doctor, and indeed, part of everyday culture.
procedures lying in this grey area between scientific medi- Although sections of the public may remain suspicious
cine and quackery are uncommon (and we should prob- of drugs, there are few who will refuse to use them
ably be grateful for this) – unless one counts colonic when the need arises.
irrigation and swimming with dolphins. • Oral administration is often possible, as well as other
In this book we are concerned with the last two treat- routes where appropriate.
ment categories on the list, summarized in Table 3.1, and • From the industry perspective, small-molecule drugs
in this chapter we consider the current status and future make up more than three-quarters of new products
prospects of the three main fields; namely, ‘conventional’ registered over the past decade. Pharmaceutical
therapeutic drugs, biopharmaceuticals and various bio- companies have long experience in developing,
logical therapies. registering, producing, packaging and marketing such
products.
• Therapeutic peptides are generally straightforward to
design (as Nature has done the job), and are usually
CONVENTIONAL THERAPEUTIC non-toxic.
DRUGS Disadvantages
• As emphasized elsewhere in this book, the flow of
Small-molecule drugs, either synthetic compounds or new small-molecule drugs seems to be diminishing,
natural products, have for a long time been the mainstay despite increasing R&D expenditure.
of therapeutics and are likely to remain so, despite the • Side effects and toxicity remain a serious and
rapid growth of biopharmaceuticals in recent years. For unpredictable problem, causing failures in late
their advantages and disadvantages see Box 3.1. development, or even after registration. One reason
Although the pre-eminent role of conventional small- for this is that the selectivity of drug molecules with
molecule drugs may decline as biopharmaceutical prod- respect to biological targets is by no means perfect,
ucts grow in importance, few doubt that they will continue and is in general less good than with
to play a major role in medical treatment. New technolo- biopharmaceuticals.
gies described in Section 2, particularly automated chem- • Humans and other animals have highly developed
istry, high-throughput screening and genomic approaches mechanisms for eliminating foreign molecules, so drug
to target identification, have already brought about an design often has to contend with pharmacokinetic
acceleration of drug discovery, the fruits of which are only problems.
just beginning to appear. There are also high expectations • Oral absorption is poor for many compounds. Peptides
that more sophisticated drug delivery systems (see Chapter cannot be given orally.
16) will allow drugs to act much more selectively where
they are needed, and thus reduce the burden of side
effects. (Walsh, 2003), although smaller nucleotide assemblies are
now being made using a chemical approach. Although
proteins such as insulin and growth hormone, extracted
from human or animal tissues, have long been used thera-
BIOPHARMACEUTICALS peutically, the era of biopharmaceuticals began in 1982
with the development by Eli Lilly of recombinant human
For the purposes of this book, biopharmaceuticals are insulin (Humulin), made by genetically engineered
therapeutic protein or nucleic acid preparations made Escherichia coli. Recombinant human growth hormone
by techniques involving recombinant DNA technology (also produced in E. coli), erythropoietin (Epogen) and

34
 Therapeutic modalities Chapter |3|

Table 3.1  The main types of chemical therapeutic agent

Type Source Examples Notes

Conventional Synthetic organic Most of the pharmacopoeia The largest category of drugs
small-molecule compounds* in use, and of new
drugs registrations
Natural products Paclitaxel (Taxol), many antibiotics Continues to be an important
and anticancer drugs (e.g. penicillins, source of new therapeutic
aminoglycosides, erythromycin), drugs
opiates (e.g. morphine), statins (e.g.
lovastatin), ciclosporin, fujimycin
Semisynthetic compounds Penicillin derivatives (e.g. ampicillin), Strategy for generating
(i.e. compounds made by second-generation statins (e.g. improved ‘second-generation’
derivatizing natural simvastatin) drugs from natural products
products)
Peptide and Synthetic Somatostatin, calcitonin, vasopressin Peptides up to approximately
protein 20 residues can be reliably
mediators made by solid-phase synthesis
Extracted from natural Insulin, growth hormone, human At one time the only source of
sources (human, animal, γ-globulins, botulinum toxin such hormones. Now largely
microbial) replaced by recombinant
biotechnology products.
γ-globulins still obtained from
human blood
Recombinant DNA Human insulin, erythropoietin, Many different expression
technology human growth hormone, GM-CSF systems in use and in
TNF-α, hirudin development
Antibodies Animal antisera, human Antisera used to treat infections
immunoglobulins such as hepatitis A and B,
diphtheria, rabies, tetanus. Also
poisoning by botulinum, snake and
spider venoms, etc.
Monoclonal antibodies Trastuzumab (directed against A rapidly growing class of
epidermal growth factor receptor), biopharmaceuticals, with many
rituximab (directed against B-cell products in development
surface antigen)
Enzymes Recombinant DNA Cerebrosidase, dornase,
technology galactosidase
Vaccines Infecting organism Smallpox, diphtheria, measles, The conventional approach,
(killed, attenuated or tuberculosis, tetanus, influenza and still widely used. Some risk of
non-pathogenic strains) many others introducing viable pathogens
Antigens produced by Many of the above vaccines now Advantages are greater
recombinant DNA available as recombinant antigens consistency and elimination of
technology risk of introducing pathogens
DNA products Recombinant DNA Antisense and siRNA Many products in clinical
technology oligonucleotides (e.g. Vitravene) development. Vitravene (for
treating cytomegalovirus
infection) is the only marketed
product so far

Continued

35
Section | 1 | Introduction and Background

Table 3.1  Continued

Type Source Examples Notes

Cells Human donors, Various stem cell therapies in
engineered cell lines development
Tissues Human donors, animal Apligraf Bilayer of human skin cells
tissues, engineered
tissues
Organs Human donors Transplant surgery
*Not considered here are many ‘adjunct’ therapies, such as oxygen, antiseptic agents, anaesthetic agents, intravenous salts, etc., which are
outside the scope of this book.

tissue plasminogen activator (tPA) made by engineered

mammalian cells followed during the 1980s. This was the
Box 3.2  Advantages and disadvantages
birth of the biopharmaceutical industry, and since then of biopharmaceuticals
new bioengineered proteins have contributed an increas-
Advantages
ing proportion of new medicines to be registered (see Table
• The main benefit offered by biopharmaceutical
3.1 for some examples, and Chapters 12 and 22 for more
products is that they open up the scope of protein
details). The scope of protein biopharmaceuticals includes
therapeutics, which was previously limited to proteins
copies of endogenous mediators, blood clotting factors,
that could be extracted from animal or human sources.
enzyme preparations and monoclonal antibodies, as well
• The discovery process for new biopharmaceuticals is
as vaccines. See Box 3.2 for their advantages and disadvan-
often quicker and more straightforward than with
tages. This field has now matured to the point that we are
synthetic compounds, as screening and lead
now facing the prospect of biosimilars consequent on the
optimization are not required.
expiry of the first set of important patents in 2004 and
• Unexpected toxicity is less common than with
notwithstanding the difficulty of defining ‘difference’ in
synthetic molecules.
the biologicals space (Covic and Kuhlmann 2007).
• The risk of immune responses to non-human proteins
Immunization against infectious diseases dates from
– a problem with porcine or bovine insulins – is
1796, when Jenner first immunized patients against
avoided by expressing the human sequence.
smallpox by infecting them with the relatively harmless
• The risk of transmitting virus or prion infections is
cowpox. Many other immunization procedures were
avoided.
developed in the 19th century, and from the 20th century
Disadvantages
onwards pharmaceutical companies began producing
standardized versions of the antigens, often the attenuated • Producing biopharmaceuticals on a commercial scale is
expensive, requiring complex purification and quality
or modified organisms themselves, as well as antisera
control procedures.
which would give immediate passive protection against
disease organisms. Vaccines and immune approaches to • The products are not orally active and often have
short plasma half-lives, so special delivery systems may
controlling disease are still a major concern, and increas-
be required, adding further to costs. Like other
ingly biotechnology-derived vaccines are being developed
proteins, biopharmaceutical products do not cross the
to improve the efficacy of, and reduce the risks associated
blood–brain barrier.
with, preparations made from infectious material.
• For the above reasons, development generally costs
Overall, biopharmaceuticals offer great promise for the
more and takes longer, than it does for synthetic drugs.
future, and rapid progress is being made in the technolo-
• Many biopharmaceuticals are species specific in their
gies used to produce them (Scrip Report, 2001; Covic and
effects, making tests of efficacy in animal models
Kuhlmann 2007). Currently, nearly all approved biophar-
difficult or impossible.
maceuticals are proteins, the majority being copies of
endogenous mediators, monoclonal antibodies or vac-
cines. Many of the clinically useful hormones and media- much broader possibilities, and progress is being facili-
tors that we currently know about have already been tated by identifying the genes for important functional
produced as biopharmaceuticals, but future advances are proteins, such as key enzymes, transporters, etc. Once the
rapidly being made as basic research discovers new protein DNA sequence of a putative target is known, its amino
signalling mechanisms. Monoclonal antibodies and anti- acid sequence can be inferred and an antibody or antibody-
body mimicking scaffold proteins (see Chapter 13) offer mimetic protein can be produced, even if the target protein

36
 Therapeutic modalities Chapter |3|

is of such low abundance that it cannot be isolated and effectively eliminated for good. The serious risks and
biochemically. ethical objections to such human genetic engineering,
Following the wave of successes by the biotechnology however, have led to a worldwide ban on germ-cell gene-
industry in producing biopharmaceuticals such as human therapy experiments, and efforts are restricted to somatic
insulin, erythropoietin and growth hormone during the cell treatments.
1980s and 1990s, medical biotechnology expanded into How much impact has gene therapy had so far as a
many other fields, including the development of therapeu- therapeutic approach, and what can be expected of it in
tic modalities beyond therapeutic proteins and antibodies. the future? The first successful trial of gene therapy to be
Next we briefly discuss two important developments still reported was by Anderson and colleagues, who used it in
in the experimental phase, namely gene-based and cell- 1990 to replace the dysfunctional gene for the enzyme
based therapies, which are under very active investigation. adenosine deaminase (ADA). ADA deficiency causes severe
combined immunodeficiency syndrome (SCID), a rare condi-
tion which prevents the normal immune response to
pathogens, and means that the child can only survive in a
GENE THERAPY germ-free environment. This first gene-therapy trial was
successful in partly restoring ADA function, but by no
Recombinant DNA technology offers the promise of alter- means curative. Hundreds of clinical trials were performed
ing the genetic material of cells and thereby correcting the during the 1990s, mainly in three clinical areas, namely
results of genetic defects, whether inherited or acquired. cancer, AIDS and single-gene inherited disorders such as
The techniques for manipulating cellular DNA that under- cystic fibrosis, haemophilia and SCID. Most of these used
pin much of modern molecular biology have great versatil- viral vectors to deliver the DNA, though some used
ity, and can in principle be applied to therapeutic as well liposome-packaged DNA or other non-viral vectors for
as experimental endeavours. Even where the genetic basis this purpose. The genetic material was delivered systemi-
of the disease is not well understood, it should be possible cally in some cases, by intravenous or subcutaneous injec-
to counteract its effects by genetic, as distinct from phar- tion; in other cases it was injected directly into solid
macological, means. Further technical information about tumours. An alternative strategy was to harvest bone
gene therapy is given in Chapter 12, and in reference works marrow cells from the patient, transfect these with the
such as Meager (1999), Templeton and Lasic (2000), necessary DNA construct ex vivo, and return them to the
Kresina (2001), Brooks (2002) and Sheridan (2011). Gene patient so that the genetically modified cells would recolo-
therapy has been actively investigated for more than two nize the bone marrow and provide the required protein.
decades, and many clinical trials have been performed. So These techniques had been extensively worked out in labo-
far, however, the results have proved disappointing, and ratory animals, but the clinical results were uniformly dis-
there are currently (at the time of publishing) no gene appointing, mainly because transfection rates were too
therapy products approved for clinical use, although many low and expression was too transient. Repeat administra-
clinical trials are still in progress (Sheridan 2011). tion of viral vectors often elicited an immune response
The most widely investigated approach involves intro- which inactivated the vector. So the very high expectation
ducing new genes to replace missing or dysfunctional in the early 1990s that gene therapy would revolutionize
ones; this is most commonly done by engineering the new treatment in many areas of medicine, from arthritis to
gene into a modified virus (the vector), which has the mental illness, quickly gave way to much more guarded
ability to enter the host cell, causing expression of the optimism, and in some cases pessimistic dismissal of the
artificially introduced gene until the cell dies or expels whole concept. There were, however, a few cases in which
the foreign DNA. Such non-integrated DNA is usually SCID in children was successfully – and apparently per-
eliminated quite quickly and is not passed on to the cell’s manently – cured by gene therapy, and there were other
progeny, and so this type of transfection is generally only trials in haemophilia and certain cancers where results
appropriate in situations where transient expression is all looked promising. Alarm bells sounded, first in 1999
that is required. Retroviral vectors are able to incorporate when a teenager, Jesse Gelsinger, who was participating in
the new DNA into the host cell’s chromosomes, where it a gene therapy trial in Philadelphia, developed an intense
will remain and be expressed during the lifetime of the immunological reaction and suddenly died 4 days after
cell and will be passed on to any progeny of that cell. More treatment. Official scrutiny uncovered many other cases of
elaborate gene therapy protocols for treating single-gene adverse reactions that had not been reported as they
disorders are designed actually to correct the disease- should have been. Many ongoing trials were halted, and
producing sequence mutation in the host genome, or to much tighter controls were imposed. Subsequently, in
alter gene expression so as to silence dysfunctional genes. 2000, immune function was successfully restored in 18
At one time gene therapy directed at germline cells was SCID children, 17 of whom were alive 5 years later (the
considered a possibility, the advantage being that an inher- first therapeutic success for human gene therapy), but
ited gene defect could be prevented from affecting progeny, two later developed leukaemia, thought to be because

37
Section | 1 | Introduction and Background

integration of the retroviral transgene occurred in a way a roadblock that is very hard to overcome and most of the
that activated a cancer-promoting gene, raising even more early clinical trials currently in progress are using direct
serious concerns about the long-term side effects of gene injection to the target tissues (Ledford, 2010).
therapy. In addition to their chequered clinical trials history,
In the much more cautious atmosphere now prevailing, gene therapy products share with other biopharmaceuti-
some carefully controlled trials are beginning to give posi- cals many features that cause major pharmaceutical com-
tive results, mainly in the treatment of haemophilia, but panies to shy away from investing heavily in such products.
overall, the general view is that gene therapy, while The reagents are large molecules, or viruses, that have to
showing great theoretical potential, has so far proved dis- be delivered to the appropriate sites in tissues, often to
appointing in its clinical efficacy, amid concerns about its particular cells and with high efficiency. Supplying gene
long-term safety and ongoing problems in designing effec- therapy reagents via the bloodstream is only effective for
tive delivery systems (see commentaries by Cavazzana- luminal vascular targets, and topical administration is
Calvo et al., 2004; Relph et al., 2004; Sheridan, 2011). usually needed. Viral vectors do not spread far from the
Pessimists refer to a decade of failure and note that hun- site of injection, nor do they infect all cell types. The
dreds of trials have failed so far to produce a single vectors have their separate toxicology issues. Commercial
approved therapy. A quick survey of the literature, however, production, quality control, formulation and delivery
shows a profusion of laboratory studies aimed at improv- often present problems.
ing the technology, and exploring many new ideas for In summary, the theoretical potential of gene therapy is
using gene therapy in numerous conditions, ranging from enormous, and the ingenuity being applied to making it
transplant rejection to psychiatric disorders. work is very impressive. Still, after 30 years of intense
The main problems to be overcome are (a) to find deliv- research effort no product has been developed, and many
ery vectors that are efficient and selective enough to trans- of the fundamental problems in delivering genes effec-
fect most or all of the target cells without affecting other tively and controllably still seem far from solution. Most
cells; (b) to produce long-lasting expression of the thera- likely, a few effective products for a few specific diseases
peutic gene; and (c) to avoid serious adverse effects. Addi- will be developed and marketed in the next few years, and
tionally, a method for reversing the effect by turning the this trickle will probably grow until gene therapy makes a
foreign gene off if things go wrong would be highly desir- significant contribution to mainstream therapeutics.
able, but has not so far been addressed in trials. Whether it will grow eventually to a flood that supplants
Antisense DNA and small interfering RNA (siRNA) have much of conventional therapeutics, or whether it will
been investigated as an alternative to the DNA strategies remain hampered by technical problems, nobody can say
outlined above. Antisense DNA consists of an oligonucleo­ at this stage. In the foreseeable future, gene therapy is
tide sequence complementary to part of a known mRNA likely to gain acceptance as a useful adjunct to conven-
sequence. The antisense DNA binds to the mRNA and, by tional chemotherapy for cancer and viral infections, par-
mechanisms that are not fully understood, blocks expres- ticularly AIDS. The Holy Grail of a cure for inherited
sion very selectively, though only for as long as the anti- diseases such as cystic fibrosis still seems some way off.
sense DNA remains in the cell. The practical problems of
developing therapeutic antisense reagents are considera-
ble, as unmodified oligonucleotides are quickly degraded
in plasma and do not enter cells readily, so either chemical CELL-BASED THERAPIES
modification or special delivery systems such as liposomal
packaging are required (see Kole et al., 2012). So far only Cell-replacement therapies offer the possibility of effective
one antisense preparation has been approved for clinical treatment for various kinds of degenerative disease, and
use, an oligonucleotide used to treat an ocular virus infec- much hope currently rests on the potential uses of stem
tion in AIDS patients. Ribozymes, specific mRNA sequences cells, which are undifferentiated progenitor cells that can
that inactivate genes by catalysing DNA cleavage, are being be maintained in tissue culture and, by the application of
investigated as an alternative to antisense DNA, but so far appropriate growth factors, be induced to differentiate into
none has been approved for clinical use. The recent past functional cells of various kinds. Their ability to divide in
has seen an explosion of the evaluation of siRNAs follow- culture means that the stock of cells can be expanded as
ing the award of the Nobel Prize for its discovery in 2006 required (see Atala et al., 2011; Lanza et al., 2007).
and the demonstration that it can be used to silence Autologous cell grafts (i.e. returning treated cells to the
important genes in non-human primates (Zimmerman same individual) are quite widely used for treating leukae-
et al., 2006). Most major pharmaceutical companies made mias and similar malignancies of bone marrow cells. A
large investments in the technology in the hope that it sample of the patient’s bone marrow is taken, cleansed
would have broad utility and lend itself to systemic medi- of malignant cells, expanded, and returned to colonize
cation against a variety of targets but it now seems that the the bone marrow after the patient has been treated with
problems of delivery of the synthetic nucleotide 23mers is high-dose chemotherapy or radiotherapy to eradicate all

38
 Therapeutic modalities Chapter |3|

resident bone marrow cells. Bone marrow is particularly produced, but trials in humans have so far been ruled out
suitable for this kind of therapy because it is rich in stem because of the risk of introducing pig retroviruses into
cells, and can be recolonized with ‘clean’ cells injected into humans. Despite much discussion and arguments on both
the bloodstream. sides, there is no sign of this embargo being lifted. Organ
Apart from this established procedure for treating bone transplantation requires such a high degree of organiza-
marrow malignancies, only two cell-based therapeutic tion to get the correct matched organs to the right patients
products have so far gained FDA approval: preparations of at the right time, as well as advanced surgical and follow-up
autologous chondrocytes used to repair cartilage defects, resources, that it will remain an option only for the privi-
and autologous keratinocytes, used for treating burns. leged minority.
Other potential applications which have been the focus of Building two- (e.g. skin) and three-dimensional (e.g. a
much experimental work are reviewed by Fodor (2003). heart valve) structures that are intended to function
They include: mechanically, either from host cells or from banked, certi-
• Neuronal cells injected into the brain (Isaacson, fied primordial or stem cell populations, is at the cutting
2003) to treat neurodegenerative diseases such as edge of tissue engineering efforts. The aim is to fashion
Parkinson’s disease (loss of dopaminergic neurons), these tissues and organ parts around artificial scaffold
amyotrophic lateral sclerosis (loss of cholinergic materials, and to do this in culture under the control
neurons) and Huntington’s disease (loss of GABA of appropriate growth and differentiation factors. The
neurons) development of biocompatible scaffolding materials,
• Insulin-secreting cells to treat insulin-dependent and achieving the right growth conditions, are problems
diabetes mellitus where much remains to be done. Artificial skin prepara-
• Cardiac muscle cells to restore function after tions recently became available and others will probably
myocardial infarction. follow.
Also in an early, albeit encouraging, state of develop-
The major obstacle to further development of such cell-
ment are bionic devices – the integration of mechanical
based therapies is that the use of embryonic tissues – the
and electronic prostheses with the human body – which
preferred source of stem cells – is severely restricted for
will go beyond what has already been accomplished with
ethical reasons. Although stem cells can be harvested from
devices such as cochlear implants and neurally controlled
adult tissues and organs, they are less satisfactory. Like
limb prostheses. The role of the major pharmaceutical
gene therapy, cell-based therapeutics could in principle
companies in this highly technological area is likely to be
have many important applications, and the technical
small. The economic realities of the relatively small patient
problems that currently stand in the way are the subject
populations will most likely be the limiting factor govern-
of intensive research efforts. Biotechnology companies are
ing the full implementation of integrated bionics.
active in developing the necessary tools and reagents that
are likely to be needed to select and prepare cells for
transplantation.
SUMMARY

Small organic molecule drugs of molecular weight

TISSUE AND ORGAN
<500 Da are the preferred therapeutic modality of the
TRANSPLANTATION major pharmaceutical companies for most disease appli-
cations but the landscape is now changing and it has been
Transplantation of human organs, such as heart, liver, predicted that by 2014 the most important therapies for
kidneys and corneas, is of course a well established proce- treating human diseases will be biologicals (Reuters
dure, many of the problems of rejection having been Factbox, 2010). The development over the years of large,
largely solved by the use of immunosuppressant drugs chemically diverse small-molecule libraries, many already
such as ciclosporin and fujimycin. Better techniques for with ‘drug-like’ properties (see Chapters 8 and 9) built
preventing rejection, including procedures based on gene into their structure, reinforces the commitment of the
therapy, are likely to be developed, but the main obstacle industry to this approach. However, protein and peptide
remains the limited supply of healthy human organs, and therapeutics also have their place in the pharmaceutical
there is little reason to think that this will change in the armamentarium, especially with respect to the immune
foreseeable future. The possibility of xenotransplantation system and hormonal dysregulation. Many pharmaceuti-
– the use of non-human organs, usually from pigs – has cal companies began with immune antisera and vaccines,
received much attention. Cross-species transplants are but the first specialized biotechnology companies took
normally rejected within minutes by a process known as advantage of recombinant DNA methods to produce ther-
hyperacute rejection. Transgenic pigs whose organs are apeutic proteins. Although the major pharmaceutical
rendered resistant to hyperacute rejection have been companies have almost all adopted protein therapeutics

39
Section | 1 | Introduction and Background

in addition to their traditional small molecule approach, more difficult than expected. Nevertheless, there is reason
many of the advances in the field have been made by for optimism in the long term. Currently, gene therapy
specialist biotechnology companies. development is being directed mainly at life-threatening
Protein- and DNA- and RNA-based biopharmaceuticals disorders such as cancer, AIDs and haemophilia, where the
often face difficult pharmacokinetic problems, in particu- need is greatest and the risks are balanced by the severity
lar poor absorption, rapid degradation, and inability to of the diseases. It is likely to be another decade or two
enter cells or cross the blood–brain barrier. Their success- before gene therapy begins to make a broader clinical
ful development therefore often depends on developing impact.
suitable delivery systems that help to overcome these The involvement of pharmaceutical companies in the
problems. For this reason (and also to improve the per- transplantation field is largely confined to improving the
formance of conventional therapeutic drugs) drug delivery immunosuppressant drugs that are needed to protect
technology (see Chapter 16) is currently receiving a great transplants from immune rejection. The use of transplants
deal of attention, with many new polymer- and liposome- is severely restricted by the availability of human organs,
based formulations being invented and tested. The right and hopes for improving the situation by the use of
delivery system is as necessary as the right drug, and for xenografts are unlikely to be realized in the foreseeable
biopharmaceuticals the two will generally need to be future. Stem-cell technologies are likely to be used success-
developed in tandem, rather than first developing a com- fully for certain kinds of tissue repair and cell replacement;
pound and then optimizing the delivery system (which biotechnology companies, rather than pharmaceutical
is the development strategy usually adopted for small- companies, are likely to make the running in these new
molecule drugs). fields although some large firms are establishing depart-
Somatic (non-germline) gene therapy initially was ments of regenerative medicine. Currently, techniques
thought to have great promise for curing inborn errors that such as bone marrow transplants are being developed and
lead to disease. More than 30 years later, although the used successfully by clinical teams without any necessary
technology and our understanding have greatly improved, input from commercial research. Probably their use will
clinical success has proved elusive. Optimizing vectors and become more routine, but it seems unlikely that the
delivery systems so as to produce long-lasting gene expres- market size for commercial products in this area will be
sion in the tissues where it is needed has proved much enough for a large pharmaceutical company.

REFERENCES

Atala A, Lanza R, Thomson JA, Nerem Kole R, Krainer AR, Altman S. RNA British Medical Journal 2004;329:
R, editors. Principles of regenerative therapeutics: beyond RNA 839–42.
medicine. 2nd ed. New York: interference and antisense Reuters Factbox. Apr 13th 2010 Worlds
Academic Press; 2011. oligonucleotides. Nature Rev Drug top 2014 vs 2010 www.reuters.com/
Brooks G, editor. Gene therapy: the use Discov 2012;11:125–40. article/2010/04/13
of DNA as a drug. New York: John Kresina TF, editor. An introduction to Scrip Report. (2001) Biopharmaceuticals:
Wiley and Sons; 2002. molecular medicine and gene a new era of discovery in the
Cavazzana-Calvo M, Thrasher A, Mavilio therapy. New York: John Wiley and biotechnology revolution.
F. The future of gene therapy. Nature Sons; 2001. Sheridan C. Gene therapy finds its
2004;427:779–81. Lanza R, Langer R, Vacanti JP, editors. niche. Nature Biotechnology
Covic A, Kuhlmann MK. Biosimilars: Principles of tissue engineering. New 2011;29:121–8.
recent developments. International York: Academic Press; 2007. Templeton NS, Lasic DD. Gene therapy:
Urology and Nephrology Ledford H. Drug giants turn their backs therapeutic mechanisms and
2007;39:261–6. on RNA interference Nature strategies. New York: Marcel Dekker;
Fodor WL. Tissue engineering and cell 2010;468:487. 2000.
based therapies, from the bench to Meager A. Gene therapy technologies: Walsh G. Biopharmaceuticals. 2nd ed.
the clinic: the potential to replace, applications and regulations from Chichester: John Wiley and Sons Ltd;
repair and regenerate. Reproductive laboratory to clinic. Chichester: John 2003.
Biology and Endocrinology Wiley and Sons; 1999. Zimmerman TS, Lee ACH, Akinc A,
2003;1:102–7. Relph K, Harrington K, Pandha H. et al. RNAi-mediated gene silencing
Isaacson O. The production and use of Recent developments and current in non-human primates. Nature
cells as therapeutic agents in status of gene therapy using viral 2006;441:111–4.
neurodegenerative diseases. Lancet vectors in the United Kingdom.
Neurology 2003;2:417–24.

40
 Chapter 4 
The drug discovery process: general principles
and some case histories
H P Rang, R G Hill

same area, or from drugs already in clinical use? Is it likely

INTRODUCTION that an esoteric drug delivery system will be required, and
if so, can this be developed? If the drug is successfully
developed, is the expected market sufficient to justify the
The creation of a new drug can be broadly divided into
cost of development? The answers to questions of this
three main phases (Figure 4.1):
kind are likely to change, for better or for worse, during
• Drug discovery – from therapeutic concept to the course of the project, so it is essential to keep such
molecule issues constantly under review, and to adapt the project
• Drug development – from molecule to registered plan if necessary.
product To integrate successfully the different interests – and
• Commercialization – from product to therapeutic cultures – of research, development and marketing is one
application to sales. of the major challenges for a pharmaceutical company,
Traditionally, these functions are performed by Research, and the need for such integration is a relatively modern
Development and Marketing, respectively, reflecting the development in the industry. As recently as 25–30 years
different professional training and expertise required to do ago in most companies, the process was much more com-
the job. Figure 4.1 greatly oversimplifies what is actually a partmentalized: scientists produced molecules with inter-
very complex process. For example, development activi- esting pharmacological properties, development functions
ties, in the form of additional clinical trials, or testing of were responsible for checking their safety and turning
new formulations, continue well beyond the point of reg- them into registrable drugs, and the marketing department
istration, with the aim of extending the range of applica- generated sales and turned them into revenues. At the time
tions of the compound or complying with regulatory this worked well, and many companies prospered. The
requirements. The discovery team, having delivered the drop-out rate was not excessive, because regulatory require-
first candidate drug, will carry on looking for others, to ments were less stringent, and the failures that did occur
serve as back-ups in case the lead compound should fail were not unduly expensive in terms of time and resources
in development, or as follow-up compounds intended to lost. Since then, biomedical science has advanced dramati-
have advantages over the lead compound. The three com- cally, drug discovery and development have become
ponents of the overall process are not independent and more technology-driven and, hence, expensive, regulatory
consecutive stages, but have to be closely coordinated at requirements are much more stringent, and the competi-
all stages of the project. At the outset of any new project, tion is more intense. With bigger teams, and more complex
the criteria against which the plan will be judged include multidisciplinary tasks, effective project management has
not only its scientific strength and originality but, impor- become much more important than it used to be to keep
tantly, development and marketing issues. For example, if costs and delays to a minimum. An additional complica-
the therapeutic target is an ill-defined clinical disorder, tion is that of the increased amount of work being done
such as chronic fatigue syndrome, will it be possible to in partnership with other companies or with academic
measure clinical efficacy objectively? Does the project face groups, adding the need for alliance as well as project
stiff competition from other companies working in the management (see also Chapter 22).

© 2012 Elsevier Ltd. 43

Section | 2 | Drug Discovery

Commercialization

Discovery Development

Therapeutic Target Target Lead Lead Preclinical Clinical Regulatory

Product
concept selection validation finding optimization development development approval

Lead Candidate Registration

compound drug

Fig. 4.1  Three main phases of the creation of a new drug: discovery, development and commercialization.

A more detailed overview of the drug discovery phase error prone to place at the front end of a drug discovery
of a typical project aimed at producing a new synthetic project.
drug is shown in Figure 4.2. It starts with the choice of a The foregoing remarks apply to the discovery of conven-
disease area and defining the therapeutic need that is to tional ‘small-molecule’ therapeutics, and the strategy for
be met. It proceeds to the identification of the biochemi- developing biopharmaceuticals – an increasing propor-
cal, cellular or pathophysiological mechanism that will be tion of new drugs appearing on the market – is generally
targeted, and, if possible, the identification and validation different. Biopharmaceutical agents (see Chapter 3) are
of a molecular ‘drug target’. Next comes the identification very diverse, including endogenous mediators, mono-
of a lead structure, followed by the design, testing and clonal antibodies and vaccines, and in the future, no
fine-tuning of the drug molecule to the point where it is doubt, products for siRNA and gene therapy applications.
deemed suitable for development, discussed in more Where endogenous molecules are involved, the concept
detail in subsequent chapters. of targets and lead compounds has much less relevance,
The strong emphasis on defined molecular (normally as nature has done the discovery part of the work, so
protein) drug targets as a starting point is too recent to once the therapeutic relevance of the substance has been
have culminated so far in many actual drugs on the established, the problems mainly revolve around the
market. Many drugs now being registered have their production, purification and formulation of the material
origins in research going back 20 years or more, in the in a form suitable for the market. With other kinds of
‘premolecular’ drug discovery era, when the selected biopharmaceuticals, such as therapeutic antibodies, the
targets were mainly pathophysiological or biochemical molecular target will generally be chosen in advance, and
mechanisms, such as blood pressure regulation, inflam- the main task is to obtain an antibody with the required
mation or cholesterol metabolism, of which the molecu- properties.
lar components were not yet defined. This earlier period, A glance at the pharmacopoeia will show that many
roughly from 1960 to 1980, was actually highly pro­ therapeutic agents, particularly anti-infective and anti-
ductive in terms of drug discovery, representing a return tumour drugs, originate from natural products, rather than
on R&D investment considerably greater than what can synthetic molecules (Table 4.1). Until about 1950, when
be achieved today, and the discovery approaches used synthetic chemistry really came into its own as a source of
then remain very much alive despite the increasing new drugs, most of the pharmacopoeia consisted of natural
emphasis on molecular targets. Nevertheless, we now products, and they continue to be important, as the
think increasingly in terms of defined molecular targets as example of paclitaxel described below shows. It is reason-
the necessary starting point for drug discovery, and turn able to suppose that such ready-made, highly evolved bio-
automatically to molecular technologies to provide the molecules stand a better chance of interacting with selected
necessary tools. Until about 1980 this was rarely feasible; drug targets than do random synthetic molecules, and the
even when the ‘target’ was defined, for example as an pool from which they come is huge and largely untapped.
enzyme or a receptor, it was seldom available in sufficient Exploiting such a ready-made compound library is seen as
quantities in a purified functional form to be used as the an attractive strategy which has led to some important
basis for screening assays. Instead, the functional activity therapeutic breakthroughs, such as the anti-malarial drug
of the target was measured by indirect means in isolated artemesinin, immunosuppressants such as ciclosporin fujimy-
tissue preparations, or even in whole animals, methods cin (FK506) and rapamycin, as well as paclitaxel and other
which we nowadays regard as too slow, laborious and recently introduced anticancer drugs such as epothilones. In

44
 The drug discovery process: general principles and some case histories Chapter |4|

Table 4.1  Examples of therapeutic drugs derived

Choice of project Ch.5 from natural products
Warfarin Anticoagulant. Synthetic
Choice of starting compound derived from
points dicoumarol, found in spoiled
sweet clover
Heparin Anticoagulant, occurring
naturally in mammalian tissues
Biological Chemical
Hirudin Anticoagulant from leech, now
Target Custom- produced by genetic engineering
Available
identificatation synthesized
compound Opiates Analgesic compounds from
compound
Chs 6,7 libraries poppies
libraries
Target Methylxanthines Phosphodiesterase inhibitors and
validation (caffeine, adenosine receptor antagonists.
theophylline) Produced by tea, coffee and
Ch.9 coca plants
Selected target
Statins HMG CoA reductase inhibitors
used to reduce plasma
cholesterol. Lovastatin is a
Assay
fungal metabolite. Later
development
compounds (mevastatin,
pravastatin) synthesized from
lovastatin
Ch.8 Screening assays Screening libraries
Cromoglycate Asthma prohylaxis. Synthetic
compound based on khellin, a
HTS plant product used as a herbal
medicine
Vinca alkaloids Anticancer drugs produced by
Primary hits (vincristine, plants of the periwinkle family
vinblastine)
Paclitaxel Anticancer drug from yew tree
Secondary
assays Etoposide Anticancer drug synthesized
from podophyllotoxin, produced
by mandrake plant; used in folk
Confirmed hits medicine
Artemether Antimalarial drug, semisynthetic
derivative of artemesin,
Lead produced by Chinese herb
Chs 9,10,11 identification
Ivermectin Antihelminthic drug,
semisynthetic derivative of
Leads Chs 9,10 avermectin, a fungal metabolite
Antibiotics Too numerous to list. The
majority of current antibiotics
Lead are derived from fungal
optimization metabolites
In earlier times the pharmacopoeia consisted very largely of
plant-derived compounds (e.g. opiates, atropine, ephedrine, ergot
Drug candidate
alkaloids, strychnine, tubocurarine, digoxin, quinine, veratridine,
reserpine, etc.), many of which remain in therapeutic use or
provide valuable research tools.
Fig. 4.2  Drug discovery phase of a typical project aimed at
producing a new synthetic drug.

45
Section | 2 | Drug Discovery

Flecainide

Structures of
endogenous Structures of
mediators existing drugs

Genomic Molecular Structure Drug design

approaches target elucidation

Imatinib
Disease Pharmacological or Chemical
Screening assay Product
pathophysiology biochemical target leads
Omeprazole

Erythyopoietin Natural Chemical

products libraries
Biopharmaceutical
molecules
Pacilitaxel

Fig. 4.3  Discovery pathways of some successful projects.

2008–2010, 8 out of 64 novel compounds registered were 2000). The brief case histories of five successful drugs,
natural products or derived from natural products. In prac- paclitaxel (Taxol), flecainide (Tambocor), omeprazol (Losec),
tice, the theoretical advantages of natural products are imatinib (Gleevec/Glivec) and trastuzumab (Herceptin), are
balanced by several practical disadvantages. Access to summarized in Figure 4.3 and Table 4.2, and described in
source material in remote places can be troublesome for more detail below. Each represents a highly innovative
geographical reasons, as well as being politically sensitive, ‘breakthrough’ project rather than an incremental devel-
and the continuing availability of the active compound, if opment based on an existing therapy, and they illustrate
it cannot be synthesized on a commercial basis, may be the variety of different approaches taken by successful
uncertain. Microorganisms have an advantage over higher projects over the past 35 years. However, for several reasons
species in this regard, but initial positive test data on we should avoid interpreting these as guidelines for success
microbial samples frequently cannot be replicated, pre- in the future. For one thing, the approach changes as the
sumably because of inconsistencies in the culture condi- underlying technologies advance; furthermore, pharma-
tions. Purification and structure determination of natural ceutical companies generally publicize only their suc-
products is now fairly routine, but is often difficult and cesses, and even then the accounts are often somewhat
time-consuming. A recent example is the introduction of sanitized, and fail to describe the errors that were made,
Sativex, a standardized preparation of cannabinoids, given the deadlines missed and the blind alleys that were
as intra oral drops for the treatment of spasticity associated encountered – the full ‘shaggy drug stories’ generally
with multiple sclerosis. remain discreetly hidden. It must be remembered that,
of drug discovery projects begun, only about 1 in 50 is
successful in terms of bringing a compound to market.
Some case histories Only at the point when official approval for trials in man
It is clear that there are many starting points and routes to is granted does the project become visible to the outside
success in drug discovery projects (Lednicer, 1993; Drews, world, so data on success rates, timelines, etc., are much

46
 The drug discovery process: general principles and some case histories Chapter |4|

Table 4.2  Timelines for some successful drug discovery projects

Drug Paclitaxel Flecainide Omeprazol Imatinib Trastuzumab

(Taxol) (Tambocor) (Losec) (Gleevec/ (Herceptin)
Glivec)
Mechanism Natural product Antidysrhythmic Inhibitor of Inhibits Abl Humanized
inhibitor of drug. Blocks gastric acid kinase monoclonal antibody.
tubulin cardiac Na+ secretion. Blocks Blocks Her2 oestrogen
depolymerization channels proton pump receptor
Company US National 3M Astra Novartis Genentech/Roche
Cancer Institute/
Bristol Myers
Squibb
Indication Ovarian cancer Cardiac Peptic ulcer Chronic myeloid Breast cancer
dysrhythmias leukaemia
Project start 1964 1965 1966 1983 1988
Compound 1971 (7 years) 1974 (9 years) 1978 (12 years) 1990 (7 years) 1990 (2 years)
synthesized
or structure
determined
Phase I 1983 (19 years) 1976 (11 years) 1981 (15 years) 1991 (3 years)
Registration 1992 (28 years) 1984 (19 years) 1988 (22 years) 2001 (18 years) 1998 (10 years)

more accessible for the minority of projects that progress contained a high proportion of the solubilizing agent Cre-
to Phase I or beyond than for the majority that never get mophor EL, causing frequent severe allergic reactions
that far. Analysing the success factors for early-stage drug when given as a bolus intravenous injection. After consid-
discovery projects is therefore difficult. erable delay, the problem was overcome by the use of slow
infusions and development was resumed. The second
problem was the supply of material for clinical trials, and
Paclitaxel (Taxol) the uncertainty that it could ever be produced on a com-
Paclitaxel is an interesting example of a project based on mercial basis. The Pacific Yew grows slowly and has a
the development of a natural product (Cragg, 1998). It restricted habitat, and conservationists were opposed to
began in the early 1960s, when the US National Cancer commercial harvesting. As a result, there was only enough
Institute, responding to the Nixon-inspired ‘war on cancer’, material for limited Phase II studies, on patients with
set up one of the first directed screening programmes – ovarian cancer. The improvement in these patients was
still running – to seek new anticancer drugs from plant dramatic, but continuation of the project, now in collabo-
sources. The sample of bark from the Pacific Yew was col- ration with Bristol Myers Squibb, was seriously hindered
lected in 1962 and found to have modest activity against by the limited supplies of yew tree bark. The conservation
various tumour cell lines. The active substance was iso- concerns were overcome when a census showed that the
lated in 1969 and joined a collection of moderately active, tree population was not in fact threatened, and industrial
but not particularly interesting, lead compounds. When supplies of bark were collected to support the trials pro-
this collection was dusted off in 1975 and tested on a new gramme right through to 1992, when the drug was offi-
assay, a melanoma cell line, paclitaxel stood out as highly cially approved.
active. Its activity was confirmed in animal models, and it Commercialization of the material extracted from bark
was soon chosen as a development candidate. Interest was was seen as a major problem, but was solved when it
further stimulated when its novel mechanism of action, was found that the needles of many yew species contain
the promotion of microtubule polymerization, was very baccatin, from which paclitaxel can be produced. This
elegantly demonstrated. Development was difficult, for semisynthetic paclitaxel, made from an abundant and
two main reasons. Paclitaxel is insoluble in water, and the renewable source, was officially approved in 1999 and is
early formulations for injection used in Phase I trials a highly successful and clinically valuable form of cancer

47
Section | 2 | Drug Discovery

therapy. The obstacles to progress in this case were (a) the the rate-limiting factor for the whole project, is typical of
failure of the primary screen to reveal the compound as many projects conducted in the 1970s (including many
anything out of the ordinary; (b) the appearance of serious that were, like flecainide, ultimately very successful).
side effects resulting from the properties of the excipient;
and (c) the supply problem.
Omeprazole (Losec)
Omeprazole, developed by Astra, was the first proton pump
Flecainide (Tambocor) inhibitor, and transformed the treatment of peptic ulcers
The story of flecainide (Banitt and Schmid, 1993) repre- when it was launched in 1988, quickly becoming the com-
sents a completely different route to success, variations of pany’s best-selling drug. The project, however, graphically
which gave rise to many innovative drugs (e.g. antihyper- described by Östholm (1995) had a chequered and death-
tensive drugs, antidepressants and antipsychotics) during defying history. In 1966 Astra started a project aimed at
the 1960s. In the early 1960s, the drugs used to treat developing inhibitors of gastric acid secretion, having pre-
cardiac dysrhythmias were mainly quinidine, procainamide, viously developed profitable antacid preparations. They
digoxin (for supraventricular tachycardias) and lidocaine started a chemistry programme based on carbamates, and
(given i.v. for ventricular dysrhythmias). The first three had collaborated with an academic group to develop a suitable
many troublesome side effects, whereas lidocaine’s use in vivo screening assay in rats. Compounds with weak
was largely confined to intensive care settings. In 1964, the activity were quickly identified; initial hepatotoxicity
3M company decided to seek better antidysrhythmic problems were overcome, and a potential development
drugs. Their chemists had developed a new synthetic compound was tested in humans in 1968. It had no effect
pathway for introducing –CF3 groups, and they started a on acid secretion, and the project narrowly escaped termi-
chemistry programme based on fluorinated derivatives of nation. In the meantime, good progress was being made
known local anaesthetic and antidysrythmic drugs. Assays by Smith, Kline and French in developing histamine H2
for antidysrhythmic activity at the time involved elaborate antagonists for the same indication, thereby adding to the
studies on anaesthetized dogs, which were quite unsuita- anxiety within Astra. At the same time Searle reported a
ble for screening, and so the group developed a simple new class of inhibitory compounds, benzimidazoles,
primary screening assay based on the ability of com- which were active but toxic. Astra began a new chemistry
pounds to prevent ventricular fibrillation induced by chlo- programme based on this series, and in 1973 produced a
roform inhalation in mice, which was used to screen highly active compound which was proposed for further
hundreds of compounds. Secondary assays on selected development. To their dismay, they found that a Hungar-
compounds were carried out on anaesthetized dogs in the ian company had a patent on this compound (for a com-
then conventional fashion. Questions of mechanism were pletely different application). However, upon entering
not addressed, it being (correctly) assumed that efficacy in licensing negotiations they found that the Hungarian
these animal models would serve as a good predictor of patent had actually lapsed because the company had
clinical efficacy irrespective of the cellular mechanisms defaulted on payment of the fees to the patent office!
involved. A potential development compound was synthe- Further studies with this compound revealed problems
sized in 1969, but abandoned on account of CNS side with thyroid toxicity, however, and more demands to ter-
effects. After a further 5 years of painstaking chemistry, minate this hapless project were narrowly fought off. The
during which many different structural classes were tested, thyroid toxicity was thought to be associated with the
flecainide was synthesized (1974) and found to have a thiouracil structure, and further chemistry aimed at elimi-
much improved therapeutic window compared to its pred- nating this resulted, in 1976, in the synthesis of picoprazole,
ecessors. The first clinical studies were performed in 1976, the forerunner of omeprazole. After yet another toxico­
and development proceeded quite smoothly until the logical alarm – this time vasculitis in dogs – which turned
compound was registered in 1984. It was the first deliber- out to be an artefact, picoprazole was tested in human
ate effort to develop an improved antidysrhythmic drug patients suffering from Zollinger–Ellison syndrome and
and proved highly successful in the clinic, now accepted was found to be highly effective in reducing acid secretion.
as the standard Class 1c antidysrhythmic agent according At around the same time, an academic group showed that
to the current classification. acid secretion involved a specific transport mechanism,
With the benefit of hindsight, we can see that the main the proton pump, which was strongly inhibited by the
delaying factor in the flecainide project was simply slow Astra compounds, so their novel mechanism of action was
chemistry, guided largely by empiricism. One result of this established. Omeprazole, an analogue of picoprazole, was
was that, after encountering side-effect problems with the synthesized in 1979, and was chosen for development
lead compound, it took 5 years to find the solution (during instead of picoprazole. The chemistry team had by then
which, one suspects, the biologists on the team were made over 800 compounds during the 13-year lifetime
growing a little bored!). This model of drug discovery of the project. The chemical development of omeprazole
research, where chemistry was both the driving force and was complicated by the compound’s poor stability and

48
 The drug discovery process: general principles and some case histories Chapter |4|

sensitivity to light, requiring special precautions in formu- blocking Abl and PDGF-receptor kinases, and systematic
lation. Phase II/III clinical trials began in 1981, but were chemical derivatization led to the synthesis of imatinib in
halted for 2 years as a result of yet another toxicology scare 1992, roughly 8 years after starting the project. Although
– carcinogenicity – which again proved to be a false alarm. crystallographic analysis of the kinase structure played no
Omeprazole was finally registered in 1988. part in guiding the chemistry that produced imatinib, a
That omeprazole, one of the most significant new drugs later structural study (Schindler et al., 2000) provided an
to appear in the early 1990s, should have survived this explanation for its selectivity for Abl kinase by showing
frightful odyssey is something of a miracle. One setback that its binding site extends beyond the ATP site to other,
after another was faced and overcome, a tribute to the less conserved domains. Imatinib proved to have no major
sheer determination and persuasive skills of the discovery shortcomings in relation to pharmacokinetics or toxicol-
team. Nowadays, when research managers pride them- ogy, and was highly effective in suppressing the growth of
selves on their decisiveness and courage in terminating cells engineered to express Bcr-Abl, and of human tumour
projects at the first hint of trouble, omeprazole would cells transplanted into mice. Importantly, it also inhibited
surely stand little chance. the growth in culture of peripheral blood or bone marrow
cells from CML patients (Druker et al., 1996). The latter
result was particularly valuable for the project, as it is
Imatinib (Gleevec/Glivec)
rarely possible to carry out such ex vivo tests on material
Imatinib (Druker and Lydon, 2000; Capdeville et al., from patients – normally, it is necessary to wait until the
2002), registered in 2001, is the most recent example in compound enters Phase II trials before any evidence relat-
these brief histories, and exemplifies the shift towards ing to clinical efficacy emerges. On that basis the project
defined molecular targets that has so altered the approach was given high priority and an accelerated clinical trials
to drug discovery over the last 20 years. In the mid-1980s, programme was devised. The first trials (Druker et al.,
it was discovered that a rare form of cancer, chronic 2001), beginning in 1998, were performed not on normal
myeloid leukaemia (CML), was almost invariably associ- subjects, but on 83 CML patients who had failed to
ated with the expression of a specific oncogene product, respond to treatment with interferon. Different doses were
Bcr-Abl kinase. The enhanced tyrosine kinase activity of tested in groups of six to eight patients, and the pharma-
this mutated protein was shown to underlie the malignant cokinetic parameters, adverse effects and clinical response
transformation of the white blood cells. The proven asso- were measured in parallel. These highly streamlined
ciation between the gene mutation1, the enhanced kinase studies showed an unequivocal clinical effect, with 100%
activity and the distinct clinical phenotype, provided a of patients receiving the higher doses showing a good
particularly clear example of cancer pathogenesis. On this haematological response. As a result, and because the
basis, the oncology team of Ciba-Geigy (later Novartis) regulatory procedures were dispatched particularly rapidly,
began a project seeking specific inhibitors of Abl kinase. It the drug was registered in record time, in May 2001, just
is known that there are many different kinases involved in 3 years after being tested for the first time in humans.
cellular regulatory mechanisms, all using ATP as a phos- Imatinib is the first ‘designer’ kinase inhibitor to be regis-
phate donor and possessing highly conserved ATP-binding tered (other drugs, such as rapamycin, probably act by
domains. Interest in kinase inhibitors as drugs was, and kinase inhibition, but this was not known at the time).
remains, high (Cohen, 2002), but at the time the known Imatinib has proved efficacious also in certain gastrointes-
inhibitors were all relatively non-specific and distinctly tinal tumours, and is the first of a number of kinase inhibi-
toxic, and the widely held view was that, as the known tors now used for treating cancer.
compounds all acted at the highly conserved ATP-binding In retrospect, the imatinib project owes its success to
site, specific subtype selective kinase inhibitors would be several factors, most obviously to the selection of a pre-
difficult to produce. The commercial potential also cisely defined molecular target which was known to be
appeared weak, as CML is a rare disease. Undaunted, the disease relevant, and was amenable to modern assay tech-
team started by developing routine biochemical assays for nologies. Setting up the various kinase assays took 4–5
this and other kinases, based on purified enzymes pro- years, but thereafter screening produced the lead series of
duced in quantity by a genetic engineering technique compounds rather quickly, and imatinib itself was made
based on the baculovirus expression system. Screening of within about 4 years of starting the screening programme.
synthetic compound libraries revealed that compounds of Avoiding the pitfalls of pharmacokinetics and toxicology,
the 2-phenylaminopyrimidine class showed selectivity in which so often hinder development, was very fortunate.
What was quite exceptional was the speed of clinical devel-
opment and registration. This was possible partly because
CML is resistant to conventional anticancer drugs, and
1
Identifiable by a chromosomal staining technique able to detect the so imatinib did not need to be compared with other treat-
translocation of DNA between two chromosomes, producing the
characteristic ‘Philadelphia chromosome’ which gives rise to the ments. Also, the designation of imatinib as an ‘orphan
abnormal kinase. drug’ (see Chapter 20), based on the rarity of CML, allowed

49
Section | 2 | Drug Discovery

the trials programme to be simplified and accelerated. Its out that the dose was about 30 times what was later found
action is readily monitored by haematological tests, per- effective in humans. The experiment revealed the unex-
mitting a rapid clinical readout. The therapeutic effect of pected hypotensive action of clonidine upon which its
the drug on circulating white cells is directly related to its subsequent commercial development was based. More
plasma level, which is often not the case for drugs acting recently, it is well known that sildenafil (Viagra) was origi-
on solid tumours. It is an example where the choice of nally intended as a vasodilator for treating angina, and
indication, initially made on the basis of a solid biological only during clinical testing did its erection-inducing effect
hypothesis, proved highly advantageous in allowing the become evident and matters have gone full circle with the
clinical development to progress rapidly. recent realization that it can be used to treat pulmonary
hypertension as well.
It might be supposed that the increasing emphasis now
Trastuzumab (Herceptin)
being placed on defined molecular targets as starting
Trastuzumab is a humanized monoclonal antibody which points for drug discovery projects would reduce the likeli-
selectively blocks the oestrogen receptor Her2. This project, hood of such therapeutic surprises. However, the molecu-
which took 8 years from compound discovery to registra- lar targets used for screening nowadays lie further, in the
tion, shows the speed with which biopharmaceuticals, functional sense, from the therapeutic response that is
under the right conditions, can be developed. The Her2 being sought than do the physiological responses relied
receptor was first cloned in 1985, and 2 years later it was on previously (see Chapter 2, Figure 2.3). Thus com-
found to be strongly over-expressed in the most aggressive pounds aimed with precision at well-defined targets com-
breast cancers. Genentech used its in-house technology to monly fail to produce the desired therapeutic effect,
develop a humanized mouse monoclonal antibody that evidently because we do not sufficiently understand the
blocked the function of the receptor and suppressed the pathophysiological pathway linking the two. Recent exam-
proliferation of receptor-bearing cells. Compared with ples of such failures include ondansetron, a 5HT3-receptor
conventional lead finding and lead optimization of syn- antagonist conceived as an antimigraine drug but ineffec-
thetic molecules, this took very little time – only 2 years tive in this indication (developed instead as an antiemetic),
from the start of the project. Antibodies generally exhibit and substance P receptor antagonists which were expected
much simpler and more predictable pharmacological to have analgesic properties in humans, but which proved
effects than synthetic compounds, and run into fewer ineffective, although again these agents have found a
problems with chemical development, formulation and useful niche as antiemetics. Lack of efficacy in clinical trials
toxicology, so that trastuzumab was able to enter Phase I – a measure of our inability to predict therapeutic efficacy
within 2 years. Clear-cut efficacy was evident in Phase II, on the basis of pharmacological properties – remains one
and the rest of the clinical development was rapid and of the commonest causes of project failure, accounting for
straightforward. Trastuzumab represents a significant step 30% of failures (Kennedy, 1997), second only to pharma-
forward in the treatment of breast cancer, as well as a com- cokinetic shortcomings (39%).
mercial success. Given the right circumstances, biophar-
maceuticals can be developed more quickly and more
cheaply than conventional drugs, a fact reflected in the
growing proportion (approximately 35% in 2001 but THE STAGES OF DRUG DISCOVERY
approaching 50% in 2011) of biopharmaceuticals among
new chemical entities being registered. In this section we are concerned with the initial stages of
the overall project outlined in Figure 4.1, up to the point
at which a molecule makes its solemn rite of passage from
Comments and conclusions research to development. Figure 4.4 summarizes the main
One common feature that emerges from a survey of the stages that make up a ‘typical’ drug discovery project, from
many anecdotal reports of drug discovery projects is that the identification of a target to the production of a candi-
they often have outcomes quite different from what was date drug2.
originally intended. The first tricyclic antidepressant drug, A huge number of ‘theoretical’ compounds (far too
imipramine, emerged from a project aimed at developing many to be physically accessible) is first ‘filtered’ in silico
antihistamines based on the structure of promethazine. to reduce it to a practicable number of compounds that is
Clonidine was synthesized in the early 1960s as part of a available, or can be synthesized, in screening libraries.
project intended to develop α-adrenoceptor agonists as High-throughput screening is then used to identify ‘hits’,
vasoconstrictors for use as decongestant nose drops. The
physician involved tested the nose drops on his wife, who
had a cold, and was surprised by the fact that her blood 2
The process of developing a new biopharmaceutical generally follows
pressure plummeted. She also slept for 24 hours. It turned a different path, and is described more fully in Chapter 12.

50
 The drug discovery process: general principles and some case histories Chapter |4|

Approximate
number of
compounds

Theoretical: >1020
Compound sources
Available: c.10 7

Filtering Preparation of
combinational
libraries
In silico assessment of likely
fit to target, compliance with Screening libraries Typically 105–106
Lipinski rules, chemical
feasibility, etc.

High-throughput screens Screening

Actives 102–103

Secondary assays, testing

for selectivity, interpretable Validation
SAR

Hits ~10 Hit series

‘Lead identification’
(or hit-to-lead phase)
Indentify DMPK issues Screening and profiling Preparation of
Preliminary tox, etc. combinational
Assessment of synthetic libraries
feasibility, etc.
Leads ~3 Lead series

Preparation of
‘Lead optimization’ combinational
Evaluation
Solve DMPK issues libraries. One-off
Evaluation of potency, synthesis
selectivity and DMPK to
establish detailed SAR Drug candidates 1–3

Preclinical development
(or ‘pre-nomination phase’)

Development compounds 1

Fig. 4.4  Stages of drug discovery.

51
Section | 2 | Drug Discovery

which show significant activity in the chosen screen. This discovery in some companies as it has been realized that
may throw up hundreds or thousands of compounds, mass screening is not always the best way to generate
depending on the nature of the screen and the size and good-quality chemical leads.
quality of the library. Normally, a significant proportion
of these prove to be artefacts of one sort or another, for
example results that cannot be repeated when the com-
pound is resynthesized, positives in the assay resulting
from non-specific effects, or simply ‘noise’ in the assay TRENDS IN DRUG DISCOVERY
system. ‘Validation’ of hits is, therefore, necessary to elimi-
nate artefacts, and this may involve repeating the screening Increasingly, drug discovery has become focused on cloned
of the hits, confirming the result in a different assay molecular targets that can be incorporated into high-
designed to measure activity on the chosen target, as well throughput screens. Target selection, discussed in Chapter
as resynthesis and retesting of the hit compounds. Valida- 6, has, therefore, assumed a significance that it rarely had
tion will also entail assessment of the structure–activity in the past: of the projects described above, three began
relationships within the screening library, and whether the without any molecular target in mind, which would
hit belongs to a family of compounds – a ‘hit series’ – seldom be the case nowadays.
which represents a reasonable starting point for further The examples summarized in Table 4.2 show that it took
chemistry. 7–12 years from the start of the project to identificaton of
In the next stage, lead identification, the validated hits the compound that was finally developed. That interval
are subjected to further scrutiny, particularly in respect of has now been substantially reduced, often to 3 years or
their pharmacokinetic properties and toxicity as well as less once a molecular target has been selected, mainly as
addressing in more detail the feasibility of building a syn- a result of (a) high-throughput screening of large com-
thetic chemistry programme. In this process, the handful pound libraries (including natural product libraries) to
of ‘hit series’ is further reduced to one or a few ‘lead series’. identify initial lead compounds; (b) improvements at the
Up to this point, the aim has been to reduce the number lead optimization stage, including the use of automated
of compounds in contention from millions to a few – synthesis to generate large families of related compounds,
‘negative chemistry’, if you will. and increased use of molecular modelling techniques,
Synthetic chemistry then begins, in the ‘lead optimiza- whereby the results of compound screening are analysed
tion’ stage. This usually involves parallel synthesis to gen- to reveal the molecular configurations that are associated
erate derivatives of the lead series, which are screened and with biological activity.
profiled with respect to pharmacology, pharmacokinetics There is strong motivation to improve not only the
and toxicology, to home in on a small number of ‘drug speed of lead optimization, but also the ‘quality’ of the
candidates’ (often a single compound, or if the project is compound selected for development. Quality, in this
unlucky, none at all) judged suitable for further develop- context, means a low probability that the compound will
ment, at which point they are taken into preclinical fail later in development. The main reasons that com-
development. pounds fail, apart from lack of efficacy or unexpected side
The flow diagram in Figure 4.4 provides a useful basis effects, are that they show toxicity, or that they have unde-
for the more detailed accounts of the various activities that sirable pharmacokinetic properties (e.g. poor absorption,
contribute to the drug discovery process, described in the too long or too short plasma half-life, unpredictable
chapters that follow. It should be realized, though, that metabolism, accumulation in tissues, etc.). In the past,
many variations on this basic scheme take place in real these aspects of a drug’s properties were often not investi-
life. A project may, for example, work on one lead series gated until the discovery/development hurdle had been
and, at the same time, go back to library screening to find crossed, as investigating them was traditionally regarded
new starting points. New biological discoveries, such as as a responsibility of ‘development’ rather than ‘research’.
the identification of a new receptor subtype, may cause the Frequently a compound would fail after several months in
project to redefine its objectives midstream. preclinical development – too late for the problem to be
Increasingly, large companies have tended to centralize addressed by the drug discovery team, which had by then
the early stages of drug discovery, up to the identification moved on. This highlighted the need to incorporate phar-
of lead series, and to carry out many screens in parallel macokinetic and toxicological studies, as well as pharma-
with very large compound libraries, focusing on putative cological ones, at an earlier stage of the project, during the
drug targets often identified on the basis of genomics, lead optimization phase. As described in Chapter 10,
rather than on pharmacology and pathophysiology relat- studies of this kind are now routinely included in most
ing to specific disease indications. In this emerging sce- drug discovery projects. Inevitably this has a cost in terms
nario, the work of the drug discovery team effectively of time and money, but this will be more than justified by
begins from the chemical lead. In the last few years there a reduction in the likelihood of compounds failing during
has been a return to a more intuitive approach to drug clinical development.

52
 The drug discovery process: general principles and some case histories Chapter |4|

Table 4.3  Trends in drug discovery

c.1800 c.1900 c.2000

Target finding Sources of Insights from Known pharmacological Defined human molecular
targets pathophysiology targets targets based on genomics
Known pharmacological Defined molecular
targets targets
Serendipitous findings
Hit finding Compound Available natural Large compound Massive virtual libraries, then
sources products ‘one-at-a- libraries, selected on focused combinatorial
time’ synthesis basis of availability libraries
Natural product libraries
Screens In vitro and in vivo High-throughput screen Virtual library screened in silico
pharmacological in vitro for predicted target affinity
screens. Radioligand Hits validated by ‘rule-of-5’ compliance,
binding assays secondary functional chemical and metabolic
assays in vitro stability, toxic groups, etc.
Selection of leads mainly Combinatorial libraries
‘in cerebro’* screened by HTS, several
screens in parallel
Validated hits
Compound Custom synthesis based Analogues synthesized Combinatorial synthesis of
sources on medicinal one at a time or analogues
chemistry insights combinatorially
Natural products
Lead finding
Screens Low throughput Functional assays in vitro Medium throughput in vitro
pharmacological and in vivo screens for target affinity
assays in vitro and DMPK characteristics
in vivo Preliminary measurements of
DMPK in vivo
Leads
Compound Analogues of active Combinatorial or one at Combinatorial or one at a
sources compounds a time synthesis of time synthesis of analogues
synthesized one at a analogues
time
Lead optimization Screens Animal models Efficacy in animal models Efficacy in animal models
Simple PK measurements Detailed DMPK analysis
in vivo
Drug candidate Safety pharmacology Safety pharmacology
In vitro genotoxicity In vitro genotoxicity
*The cerebrum required being that of an experienced medicinal chemist.

The main trends that have occurred over the last two during the 1990s, to massive virtual libraries of tens
decades are summarized in Table 4.3, the key ones being, of millions
as discussed above: • Use of automated synthesis methods to accelerate
• A massive expansion of the compound collections lead optimization
used as starting points, from large ‘white-powder’ • To deal with large, compound libraries, the
libraries of 100 000 or more compounds created introduction of high-throughput screens for actual

53
Section | 2 | Drug Discovery

compounds, and in silico screens for virtual criteria that might apply to a typical drug acting on a target
compounds such as an enzyme or receptor, and intended for oral use.
• Increasing reliance on in silico predictions to Where appropriate (e.g. potency, oral bioavailability),
generate leads quantitative limits will normally be set. Such a list of
• Focus on cloned human targets as starting points necessary or desirable features, based on results from
• Progressive ‘front-loading’ of assessment of DMPK many independent assessments and experimental tests,
(drug metabolism and pharmacokinetic) provides an essential focus for the project. Some of these,
characteristics, including in silico assessment of such as potency on target, or oral bioavailability, will be
virtual libraries. absolute requirements, whereas others, such as water solu-
bility or lack of in vitro genotoxicity, may be highly desir-
able but not essential. There are, in essence, two balancing
Project planning components of a typical project:
When a project moves from the phase of exploratory • Designing and synthesizing novel compounds
research to being an approved drug discovery project to • Filtering, to eliminate compounds that fail to satisfy
which specific resources are assigned under the direction the criteria.
of a project leader, its objectives, expected timelines Whereas in an earlier era of drug discovery these two
and resource requirements need to be agreed by the activities took place independently – chemists made com-
members of the project team and approved by research pounds and handed over white powders for testing, while
management. pharmacologists tried to find ones that worked – nowa-
The early drug discovery phase of the project will typi- days the process is invariably an iterative and interactive
cally begin when a target has been selected and the neces- one, whereby the design and synthesis of new compounds
sary screening methods established, and its aim will be continuously takes into account the biological findings
to identify one or a few ‘drug candidates’ suitable for and shortcomings that have been revealed to date. Formal
progressing to the next stage of preclinical development. project planning, of the kind that would be adopted for
The properties required for a drug candidate will vary the building of a new road, for example, is therefore inap-
from project to project, but will invariably include chemi- propriate for drug discovery. Experienced managers conse-
cal, pharmacological, pharmacokinetic and toxicological quently rely less on detailed advance planning, and more
aspects of the compound. Table 4.4 summarizes the on good communication between the various members of

Table 4.4  Typical selection criteria for drug candidates intended for oral use*

Chemical Pharmacological Pharmacokinetic Toxicological

Patentable Defined potency on Cell-permeable in vitro In vitro genotoxicity tests
structure target negative
Water-soluble Selectivity for specific Adequate oral bioavailability Preliminary in vivo toxicology
target relative to other tests showing adequate margin
related targets between expected ‘therapeutic’
dose and maximum No Adverse
Effect Dose
Chemically stable Pharmacodynamic activity For CNS drugs: penetrates
in vitro and in vivo blood–brain barrier
Large-scale No adverse effects in Appropriate plasma half-life
synthesis feasible standard safety
pharmacology tests
Non-chiral Active in disease models Defined metabolism by
human liver microsomes
No known No inhibition or induction
‘toxophoric’ groups of cytochrome P450
*Criteria such as these, which will vary according to the expected therapeutic application of the compound, would normally be applied to
compound selection in the drug discovery stage of a project, culminating in lead optimization and the identification of a drug candidate,
when preclinical development begins, focusing mainly on pharmacokinetics, toxicology, chemistry and formulation.

54
 The drug discovery process: general principles and some case histories Chapter |4|

the project team, and frequent meetings to review progress commercial aims in their mission statements, the latter
and agree on the best way forward. necessarily take priority. The company will, therefore, wish
An important aspect of project planning is deciding to ensure that the research it supports is in some way
what tests to do when, so as to achieve the objectives as relevant to its commercial objectives. Clearly, studies
quickly and efficiently as possible. The factors that have to aimed directly at drug discovery present no problems. At
be taken into account for each test are: the other end of the spectrum lies pure curiosity-driven
• Compound throughput (‘blue-skies’) research. Although such work may – and
• Cost per compound often does – lead to progress in drug discovery in the
• Amount of compound required in relation to long term, the commercial pay-off is highly uncertain
amount available and inevitably long delayed. Generally speaking, only a
• Time required. Some in vivo pharmacodynamic tests minimal amount of such long-term research is performed
(e.g. bone density changes, tumour growth) are in a commercial setting. Between the two lies a large
inherently slow. Irrespective of compound territory of ‘applied’ research, more clearly focused on
throughput, they cannot provide rapid feedback drug discovery, though still medium-term and somewhat
• ’Salience’ of result (i.e. is the criterion an absolute uncertain in its applicability. Many technological projects
requirement, or desirable but non-essential?) come into this category, such as novel high-throughput
• Probability that the compound will fail. In a screening methods, imaging technologies, etc., as well
typical high-throughput screen more than 99% as research into pathophysiological mechanisms aimed
of compounds will be eliminated, so it is essential at the identification of new drug targets. The many appli­
that this is done early. In vitro genotoxicity, in cations of genomics and molecular biology in drug dis­
contrast, will be found only occasionally, so it covery make up an increasing proportion of work in this
would be wasteful to test for this early in the applied research category. Small biotechnology companies
sequence. which started during the 1980s and 1990s moved quickly
into this territory, and several large pharmaceutical com-
The current emphasis on fast drug discovery, to increase
panies, notably SmithKlineBeecham (now part of GSK),
the time window between launch and patent expiry, and
also made major investments. The extent to which phar-
on decreasing the rate of failure of compounds during
maceutical companies invest in medium-term applied
clinical development, is having an important effect on the
research projects varies greatly. A growing tendency seems
planning of drug discovery projects. As discussed above,
to be for larger companies to set up what amounts to an
there is increasing emphasis on applying fast-result, high-
in-house biotechnology facility, often incorporating one
throughput methods of testing for pharmacokinetic and
or more acquired biotech companies. The technological
toxicological properties at an early stage (‘front loading’;
revolution in drug discovery, referred to frequently in this
see Chapter 10), even though the salience (i.e. the ability
book, has yet to pay dividends in terms of improved per-
to predict properties needed in the clinic) of such assays
formance, so it is uncertain whether this level of commit-
may be limited. The growth of high-throughput test
ment to medium-term applied research can be sustained.
methods has had a major impact on the work of chemists
More recently the view has emerged that much of the
in drug discovery (see Chapter 9), where the emphasis has
information obtained from industry driven genomic
been on preparing ‘libraries’ of related compounds to feed
studies of disease should be freely available to all and,
the hungry assay machines. These changes have undoubt-
for example, Merck in the USA has spun out its database
edly improved the performance of the industry in finding
to a not-for-profit organization called Sage (see http://
new lead compounds of higher quality for new targets. The
www.sagebase.org) for this purpose.
main bottlenecks now in drug discovery are in lead opti-
The cultural difference between academic and commer-
mization (see Chapter 9) and animal testing (see Chapter
cial research is real, but less profound than is often
11), areas so far largely untouched by the high-throughput
thought. Quality, in the sense of good experimental
revolution.
design, methodology and data interpretation, is equally
important to both, as is creativity, the ability to generate
new ideas and see them through. Key differences are that,
in the industry environment, freedom to choose projects
RESEARCH IN THE  
and to publish is undoubtedly more limited, and effective
PHARMACEUTICAL INDUSTRY interdisciplinary teamwork is obligatory, rather than a
matter of personal choice: there is little room in industry
Pharmaceutical companies perform research for commer- for the lone genius. These differences are, however, becom-
cial reasons and seek to ensure that it produces a return ing narrower as the bodies funding academic research take
on investment. The company owns the data, and is free to a more ‘corporate’ approach to its management. Individ-
publish it or keep it secret as it sees fit. Although most ual researchers in academia are substantially constrained
pharmaceutical companies include altruistic as well as to work on projects that attract funding support, and are

55
Section | 2 | Drug Discovery

increasingly being required to collaborate in interdiscipli- Within a pharmaceutical company, research findings are
nary projects. They are certainly more free to publish, but aimed at a more limited audience and their purpose is
are also under much more pressure to do so, as, in contrast more focused. The immediate aim is to provide the data
to the situation in industry, publications are their only needed for prediction of the potential of a new compound
measure of research achievement. Pharmaceutical com­ to succeed in the clinic. The internal audiences are the
panies also have incentives to publish their work: it research team itself and research management; the exter-
gives their scientists visibility in the scientific community nal audiences are the external scientific community and,
and helps them to establish fruitful collaborations, as well importantly, the regulatory authorities.
as strengthening the company’s attractiveness as an We should, however, resist the tendency to think of drug
employer of good scientists. The main restrictions to discovery as a commercial undertaking comparable devel-
publication are the company’s need to avoid compromis- oping a new sports car, where the outcome is assured
ing its future patent position, or giving away information provided that operatives with the necessary skill and expe-
that might help its competitors. Companies vary in their rience fulfil what is asked of them. They are not likely by
attitude to publication, some being much more liberal chance to invent a new kind of aeroplane or vacuum
than others. cleaner.
Scientists in industry have to learn to work, and com- Like all research, drug discovery is more likely than not
municate effectively, with the company’s management. to come up with unexpected findings, and these can have
Managers are likely to ask ‘Why do we need this informa- a major impact on the plan and its objectives. So, frequent
tion?’ which, to scientists in academia, may seem an irritat- review – and, if necessary, amendment – of the plan is
ingly silly question. To them, the purpose of any research essential, and research managers need to understand (as
is to add to the edifice of human knowledge, and working most do, having their own research experiences to draw
towards this aim fully justifies the time, effort and money on) that unexpected results, and the problems that may
that support the work. The long-term benefits to humanity follow for the project, reflect the complexity of the
are measured in terms of cultural richness and technologi- problem, rather than the failings of the team. Finding by
cal progress. In the short term the value of what has been experiment that a hypothesis is wrong is to be seen as a
achieved is measured by peer recognition and grant creditable scientific achievement, not at all the same as
renewal; the long-term judgment is left to history. designing a car with brakes that do not work.

REFERENCES

Banitt EH, Schmid JR. Flecainide. In: development. Medical Research kinase in chronic myeloid leukemia.
Lednicer D, editor. Chronicles of Review 1998;18:315–31. New England Journal of Medicine
drug discovery. vol. 3. Washington Drews J. Drug discovery: a historical 2001;344:1031–7.
DC: American Chemical Society; perspective. Science Kennedy T. Managing the drug
1993. 2000;287:1960–4. discovery/development interface.
Capdeville R, Buchdunger E, Druker BJ, Lydon NB. Lessons learned Drug Discovery Today 1997;2:
Zimmermann J, et al. Glivec (STI571, from the development of an Abl 436–44.
imatinib), a rationally developed, tyrosine kinase inhibitor for chronic Lednicer D. Chronicles of drug
targeted anticancer drug. Nature myelognous leukemia. Journal of discovery. vol. 3. Washington DC:
Reviews Drug Discovery 2002;1: Clinical Investigation 2000;105:3–7. American Chemical Society; 1993.
493–502. Druker BJ, Tamura S, Buchdunger E, Östholm I. Drug discovery: a
Cohen P. Protein kinases – the major et al. Effects of a selective inhibitor pharmacist’s view. Stockholm:
drug targets of the twenty-first of the Abl tyrosine kinase on the Swedish Pharmaceutical Press; 1995.
century? Nature Reviews Drug growth of Bcr-Abl positive cells. Schindler T, Bornmann W, Pellicana P,
Discovery 2002;1:309–16. Nature Medicine 1996;2:561–6. et al. Structural mechanism for
Cragg GM. Paclitaxel (Taxol): a success Druker BJ, Talpaz M, Resta DJ, et al. STI-571 inhibition of Abelson
story with valuable lessons for Efficacy and safety of a specific tyrosine kinase. Science 2000;289:
natural product drug discovery and inhibitor of the Bcr-Abl tyrosine 1938–42.

56
 Chapter 5 
Choosing the project
H P Rang, R G Hill

INTRODUCTION MAKING THE DECISION

In this chapter we discuss the various criteria that are The first, and perhaps the most important, decision in the
applied when making the initial decision of whether or life history of a drug discovery project is the decision to
not to embark on a new drug discovery project. The point start. Imagine a group considering whether or not to climb
at which a project becomes formally recognized as a drug a mountain. Strategically, they decide whether or not they
discovery project with a clear-cut aim of delivering a can- want to get to the top of that particular mountain; techni-
didate molecule for development and the amount of cally, they decide whether there is a feasible route; and
managerial control that is exercised before and after this operationally they decide whether or not they have the
transition vary considerably from company to company. wherewithal to accomplish the climb. In the context of a
Some encourage – or at least allow – research scientists to drug discovery project, the questions are:
pursue ideas under the general umbrella of ‘exploratory • Should we do it? (strategic issues)
research’, whereas others control the research portfolio • Could we do it? (scientific and technical issues)
more tightly and discourage their scientists from straying • Can we do it? (operational issues).
off the straight and narrow path of a specific project. Gen-
The main factors that need to be considered are sum-
erally, however, some room is left within the organization
marized in Figure 5.1.
for exploratory research in the expectation that it will
generate ideas for future drug discovery projects. When
successful, this strategy results in research-led project Strategic issues
proposals, which generally have the advantage of starting
from proprietary knowledge and having a well-motivated Strategic issues relate to the desirability of the project from
and expert in-house team. Historically, such research-led the company’s perspective, reflecting its mission (a) to
drug discovery projects have produced many successful make a significant contribution to healthcare, and (b) to
drugs, including β-blockers, ACE inhibitors, statins, make a profit for its shareholders. These translate into
tamoxifen and many others, but also many failures, for assessments respectively of medical need and market
example prostanoid receptor ligands, which have found potential.
few clinical uses. Facing harder times, managements are
now apt to dismiss such projects as ‘drugs in search of a
disease’, so it is incumbent upon research teams to align
Unmet medical need
their work, even in the early exploratory stage, as closely Unmet medical need represents what many would regard
as possible with medical needs and the business objectives as the most fundamental criterion to be satisfied when
of the company, and to frame project proposals accord- evaluating any drug discovery project, though defining and
ingly. Increasingly, research is becoming collaborative with evaluating it objectively is far from straightforward. Of
large pharmaceutical companies partnering with small course there are common and serious diseases, such as
biotechnology companies or with academic groups. many cancers, viral infections, neurodegenerative diseases,

© 2012 Elsevier Ltd. 57

Section | 2 | Drug Discovery

SCIENTIFIC AND TECHNICAL

Scientific opportunity: Development hurdles
• Strength of hypothesis Patent situation • Formulation and routes
• Availibilty of targets and freedom to of administration
• Availability of assays operate • Toxicology
• Animal models • Clinical development
• Chemical starting points Competition • Regulatory hurdles
• Production

DECISION TO
START PROJECT
STRATEGIC OPERATIONAL
Unmet Company Resource needs:
Market Company • Staff and expertise
medical strategy and Timescale
predictions pipeline • Facilities
need franchise
• Costs

Fig. 5.1  Criteria for project selection.

certain developmental abnormalities, etc., for which • How close do we come to ‘free’? In other words, can
current treatments are non-existent or far from ideal, and we be more cost-effective?
these would be generally accepted as areas of high unmet • Compliance is important: a drug that is not taken
need. Nevertheless, trying to rank potential drug discovery cannot work. Once-daily oral dosing is convenient,
projects on this basis is full of difficulties, because it but would a long-lasting injection be a better choice
assumes that we can assess disease severity objectively, and for some patients?
somehow balance it against disease prevalence. Does a rare • If an oral drug exists, can it be improved, for
but catastrophic disease represent a greater or a lesser example by changing the formulation?
medical need than a common minor one? Is severity best • The reduction of side effects is an area where there
measured in terms of reduced life expectancy, or level of is often a medical need. Do we have a strategy for
disability and suffering? Does reducing the serious side improving the side-effect profile of existing drugs?
effects of existing widely used and efficacious drugs (e.g. • Curing a condition or retarding disease progression
gastric bleeding with conventional non-steroidal anti- is preferable to alleviating symptoms. Do we have a
inflammatory drugs) meet a need that is more important strategy for achieving this?
than finding a therapy for a condition that was previously
untreatable? Does public demand for baldness cures, anti-
wrinkle creams and other ‘lifestyle’ remedies constitute a
Market considerations
medical need? These overlap to some extent with unmet medical need,
Assessing medical need is not simply a matter of asking but focus particularly on assessing the likelihood that the
customers what they would regard as ideal; this usually revenue from sales will succeed in recouping the invest-
results in a product description which falls into the cate- ment. Disease prevalence and severity, as discussed above,
gory of ‘a free drug, given once orally with no side effects, obviously affect sales volume and price. Other important
that cures the condition’. Nevertheless, this trite response factors include the extent to which the proposed com-
can serve a useful purpose when determining the medical pound is likely to be superior to drugs already on the
need for a new product: how closely do existing and future market, and the marketing experience and reputation
competitors approach this ideal? How big is the gap of the company in the particular disease area. The more
between the reality and the ideal, and would it be profit- innovative the project, the greater is the degree of uncer-
able to try and fill it? Using this type of ‘gap analysis’ we tainty of such assessments. The market for ciclosporin, for
can ask ourselves the following questions: example, was initially thought to be too small to justify its

58
 Choosing the project Chapter |5|

development – as transplant surgery was regarded as a In general, the management of a pharmaceutical

highly specialized type of intervention that would never company needs to choose and explain which therapeutic
become commonplace – but in the end the drug itself gave areas it wishes to cover, and to communicate this effec-
a large boost to organ transplantation, and proved to be a tively to the drug discovery research organization, but
blockbuster. There are also many examples of innovative most will avoid placing too much emphasis on market
and well-researched drugs that fail in development, or analysis and commercial factors when it comes to select-
perform poorly in the marketplace, so creativity is not an ing specific projects. This is partly because market ana­
infallible passport to success. Recognizing the uncertainty lysis is at best a very imprecise business and the
of market predictions in the early stages of an innovative commercial environment can change rapidly, but also
project, companies generally avoid attaching much weight because success in the past has often come from exploit-
to these evaluations when judging which projects to ing unexpected scientific opportunities and building a
support at the research stage. A case in point is the discov- marketing and commercial strategy on the basis of what
ery of H2 receptor blockers for the treatment of peptic ulcer is discovered, rather than the other way round. The
by James Black and his colleagues where throughout the interface between science and commerce is always some-
project he was being told by his sales and marketing col- what turbulent.
leagues that there was no market for this type of product
as the medical professional treated ulcers surgically. For-
Legislation, government policy,
tunately the team persisted in its efforts and again the
resulting drug, cimetidine, became a $1 billion per year reimbursement and pricing
blockbuster. Unlike the products of many other industries, pharmaceu-
ticals are subject to controls on their pricing, their pro­
motional activities, and their registration as ‘safe and
Company strategy and franchise effective’ products that can be put on the market, and
The current and planned franchise of a pharmaceutical so their commercial environment is significantly altered
company plays a large part in determining the broad by government intervention. The socialized medicine that
disease areas, such as cancer, mental illness, cardiovas­ is a feature of European and Canadian healthcare means
cular disease, gastroenterology, etc., addressed by new that the government is not only the major provider of
projects, but will not influence the particular scientific healthcare, but also a monopoly purchaser of drugs for
approach that is taken. All companies specialize to a use in the healthcare system. The USA is the only major
greater or lesser extent on particular disease areas, and this market where the government does not exert price con-
is reflected in their research organization and scientific trols, except as a feature of Medicare, where it acts as
recruitment policies. Biotechnology companies are more a major buyer and expects discounts commensurate with
commonly focused on particular technologies and scien- its purchasing power. Governments have great influence
tific approaches, such as drug delivery, genomics, growth on the marketing of drugs, and they have an interest
factors, monoclonal antibodies, etc., rather than on par- in limiting the prescription of newer, generally more
ticular disease areas. They are, therefore, more pragmatic expensive products, and this affects where the industry
about the potential therapeutic application of their dis- can sell premium-priced products. Governments and
coveries, which will generally be licensed out at an appro- their associated regulatory bodies have gradually extended
priate stage for development by companies within whose their power from a position of regulation for public
franchise the application falls. In the biotech environ- safety to one of guardians of the public purse. Products
ment, strategic issues, therefore, tend to be involved more deemed to be ‘clinically essential or life-saving’ now may
with science and technology than may be the case in be fast-tracked, whereas those that are thought to be ines-
larger pharmaceutical companies – a characteristic that is sential may take twice as long to obtain approval. Many
often appealing to research scientists. countries, such as Canada, Australia and the UK, have
The state of a company’s development pipeline, and its additional reimbursement hurdles, based on the need for
need to sustain a steady flow of new compounds entering cost-efficacy data. As a consequence, the choice of drug
the market, sometimes influences the selection of drug development candidate will have to reflect not only the
discovery projects. The company may, for example, need a clinical need of the patient, but also government health-
product to replace one that is nearing the end of its patent care policies in different countries. The choice of indica-
life, or is losing out to competition, and it may endeavour tions for a new drug is thus increasingly affected by the
to direct research to that end. In general, though, in the willingness of national healthcare systems to offer acceler-
absence of an innovative scientific strategy such top-down ated approval and reimbursement for that drug. This is a
commercially driven priorities often fail. A better, and changing landscape at the time of writing, but the indica-
quicker, solution to the commercial problem will often be tions are that governments worldwide will exert more
to license in a partly developed compound with the rather than less control over drug registration and pricing
required specification. in the future.

59
Section | 2 | Drug Discovery

Scientific and technical issues unguarded remarks, and scientists are expected to keep
their antennae well tuned to these signals – a process
These issues relate to the feasibility of the project, and referred to euphemistically as ‘competitive intelligence’.
several aspects need to be considered: There are also professional agencies that specialize in
• The scientific and technological basis on which the obtaining commercially sensitive information and provid-
success of the drug discovery phase of the project ing it to subscribers in database form. Such sources may
will depend reveal which companies are active in the area, and often
• The nature of the development and regulatory which targets they are focusing on, but will rarely indicate
hurdles that will have to be overcome if the how close they are to producing development compounds,
development phase of the project is to succeed and they are often significantly out of date.
• The state of the competition At the drug discovery stage of a project overlap with
• The patent situation. work in other companies is a normal state of affairs and
should not, per se, be a deterrent to new initiatives. The
fact that 80–90% of compounds entering clinical develop-
The scientific and technological basis ment are unsuccessful should be borne in mind. So, a
Evaluation of the scientific opportunity should be the main single competitor compound in early clinical develop-
factor on which the decision as to whether or not to ment theoretically reduces the probability of our new
embark on a project rests once its general alignment with project winning the race to registration by only 10–20%,
corporate strategy, as described above, is accepted. Some say from about 2% to 1.7%. In practice, nervousness
of the key elements are listed in Figure 5.1. In this context tends to influence us more than statistical reasoning, and
it is important to distinguish between public knowledge there may be reluctance to start a project if several com­
and proprietary knowledge, as a secure but well-known panies are working along the same lines with drug dis­
scientific hypothesis or technology may provide a less covery projects, or if a competitor is well ahead with
attractive starting point than a shakier scientific platform a compound in development. The incentive to identify
based on proprietary information. Regardless of the novel proprietary drug targets (Chapter 6) stems as much
strength of the underlying scientific hypothesis, the other from scientific curiosity and the excitement of breaking
items listed in Figure 5.1 – availability of targets, assays, new ground as from the potential therapeutic benefits of
animal models and chemical starting points – need to be the new target.
evaluated carefully, as weaknesses in any of these areas are
likely to block progress, or at best require significant time Development and regulatory hurdles
and resources to overcome. They are discussed more fully
in Chapters 6–9. Although many of the problems encountered in develop-
ing and registering a new drug relate specifically to the
compound and its properties, others are inherent to the
Competition therapeutic area and the clinical need being addressed,
Strictly speaking, competition in drug discovery is unim- and can be anticipated at the outset before any candidate
portant, as the research activities of one company do not compound has been identified. Because development con-
impede those of another, except to the limited extent that sumes a large proportion of the time and money spent on
patents on research tools may restrict a company’s ‘freedom a project, likely problems need to be assessed at an early
to operate’, as discussed below. What matters greatly is stage and taken into account in deciding whether or not
competition in the marketplace. Market success depends to embark on the discovery phase.
in part on being early (not necessarily first) to register a The various stages of drug development and registration
drug of a new type, but more importantly on having a are described in more detail in Section 3. Here we give
product which is better. Analysis of the external competi- some examples of the kinds of issue that need to be con-
tion faced by a new project therefore involves assessing the sidered in the planning phase.
chances that it will lead to earlier registration, and/or a • Will the drug be given as a short course to relieve an
better product, than other companies will achieve. Making acute illness, or indefinitely to relieve chronic symptoms
such an assessment involves long-range predictions based or prevent recurrence? Many anti-infective drugs come
on fragmentary and unreliable information. The earliest into the former category, and it is notable that
solid information comes from patent applications, gener- development times for such drugs are much shorter
ally submitted around the time a compound is chosen for than for, say, drugs used in psychiatry or for lowering
preclinical development, and made public soon thereafter plasma cholesterol levels. The need for longer
(see Chapter 19), and from official clinical trial approvals. clinical trials and more exhaustive toxicity testing
Companies are under no obligation to divulge informa- mainly accounts for the difference.
tion about their research projects. Information can be • Is the intended disease indication rare or common?
gleaned from gossip, publications, scientific meetings and Definitive trials of efficacy in rare diseases may be

60
 Choosing the project Chapter |5|

slow because of problems in recruiting patients. existing patents on compounds or research tools may limit
(For recognized ‘orphan indications’ (see Chapter the company’s ‘freedom to operate’ in the research phase.
20), the clinical trials requirements may be less Existing patents on compounds of a particular class (see
stringent.) Chapter 19) may rule out using such compounds as the
• What clinical models and clinical end-points are starting point for a new project.
available for assessing efficacy? Testing whether a drug By ‘research tool’ is meant anything that contributes
relieves an acute symptom, such as pain or nausea, is to the discovery or development of a drug, without being
much simpler and quicker than, for example, testing part of the final product. Examples include genes, cell
whether it reduces the incidence of stroke or lines, reagents, markers, assays, screening methods, animal
increases the life expectancy of patients with a rare models, etc.
cancer. Where accepted surrogate markers exist (e.g. A company whose business it is to sell drugs is not
lowering of LDL cholesterol as an indicator of usually interested in patenting research tools, but it is the
cardiovascular risk, or improved bone density on business of many biotech companies to develop and com-
X-ray as a marker of fracture risk in osteoporosis), mercialize such tools, and these companies will naturally
this enables trials to be carried out more quickly and wish to obtain patent protection for them. For pharma-
simply, but there remain many conditions where ceutical companies, such research tool patents and appli-
symptoms and/or life expectancy provide the only cations raise issues of freedom to operate, particularly if
available measures of efficacy. they contain ‘reach-through’ claims purporting to cover
• How serious is the disease, and what is the status of drugs found by using the patented tools.
existing therapies? Where no effective treatment of Some scientists may believe that research activities, in
a condition exists, comparison of the drug with contrast to manufacture and sale of a product, cannot lead
placebo will generally suffice to provide evidence of to patent infringement. This is not the case. If I have
efficacy. Where standard therapies exist several invented a process that is useful in research and have a
comparative trials will be needed, and the new drug valid patent for it, I can enforce that patent against anyone
will have to show clear evidence of superiority before using the process without my permission. I can make
being accepted by regulatory authorities. For serious money from my patent by granting licenses for a flat fee,
disabling or life-threatening conditions the or a fee based on the extent to which the process is used;
regulatory hurdles are generally lower, and the or by selling kits to carry out the process or reagents for
review process is conducted more quickly. use in the process (for example, the enzymes used in the
• What kind of product is envisaged (synthetic compound, PCR process). What I am not entitled to do is to charge a
biopharmaceutical, cell or gene-based therapy, etc.), and royalty on the sale of drugs developed with the help of my
what route of administration? The development track process. I can patent an electric drill, but I should not
and the main obstacles that are likely to be expect a royalty on everything it bores a hole in!
encountered depend very much on the type of Nevertheless, some patents have already been granted
product and the route of administration. With containing claims that would be infringed, for example,
protein biopharmaceuticals, for example, toxicology by the sales of a drug active in a patented assay, and
is rarely a major problem, but production may be, although it is hoped that such claims would be held
whereas the reverse is generally true for synthetic invalid if challenged in court, the risk that such claims
compounds. With proteins, the route of might be enforceable cannot be dismissed.
administration is often a difficult issue, whereas with Should research tool patents be the subject of a freedom
synthetic small molecule compounds the expectation to operate search at an early stage of project planning?
is that they will be developed for oral or topical use. Probably not. There are simply too many of them, and if
Cell- or gene-based products are likely to face serious no research project could be started without clearance on
safety and ethical questions before clinical trials can the basis of such a search, nothing would ever get done.
be started. The development of a special delivery At least for a large company, it is an acceptable business
system, such as a skin patch, nasal spray or slow- risk to go ahead and assume that if problems arise they
release injectable formulation, will add to the time can be dealt with at a later stage.
and cost of development.
Operational issues
The patent situation It might seem that it would be a straightforward manage-
As described in Chapter 19, patent protection in the phar- ment task to assess what resources are likely to be needed
maceutical industry relates mainly to specific chemical to carry through a project, and whether they can be made
substances, their manufacture and their uses. Nevertheless, available when required. Almost invariably, however,
at the start of a project, before the drug substance has been attempts to analyse the project piece by piece, and to esti-
identified, it is important to evaluate the extent to which mate the likely manpower and time required for each

61
Section | 2 | Drug Discovery

phase of the work, end up with a highly over-optimistic invariably an overrun in terms of time, money and
prediction, as they start with the assumption that every- manpower.
thing will run smoothly according to plan: targets will be Experienced managers understand this and make due
established, leads will be found and ‘optimized’, animal allowance for it, but the temptation to approve more
models will work as expected, and the project will most projects than the organization can actually cope with is
likely be successful in identifying a development com- hard for most to resist.
pound. In practice, of course, diversionary or delaying
events almost invariably occur. The gene that needs to be
cloned and expressed proves difficult; obtaining an engi-
neered cell line and devising a workable assay runs into A FINAL WORD
problems; new data are published which necessitate revi-
sion of the underlying hypothesis, and possibly a switch To summarize the key message in a sentence: The drug
to an alternative target; a key member of the team leaves discovery scientist needs to be aware that many factors
or falls ill; or an essential piece of equipment fails to be other than inherent scientific novelty and quality affect the
delivered on time. However, project champions seeking decision whether or not to embark on a new project. And
management support for their ideas do their case no good the subtext: The more the drug discovery scientist under-
if they say: ‘We plan to set up this animal model and vali- stands about the realities of development, patenting, reg-
date it within 3 months, but I have allowed 6 months just istration and marketing in the pharmaceutical industry,
in case …’. Happy accidents do occur, of course, but the more effective he or she will be in adapting research
unplanned events are far more likely to slow down a plans to the company’s needs, and the better at defending
project than to speed it up, so the end result is almost them.

62
 Chapter 6 

Choosing the target

H P Rang, R G Hill

How many drug targets are there?

INTRODUCTION: THE SCOPE FOR
Even when we restrict our definition to defined protein
NEW DRUG TARGETS targets, counting them is not simple. An obvious starting
point is to estimate the number of targets addressed by
The word ‘target’ in the context of drug discovery has existing therapeutic drugs, but even this is difficult. For
several common meanings, including the market niche many drugs, we are ignorant of the precise molecular
that the drug is intended to occupy, therapeutic indica­ target. For example, several antiepileptic drugs apparently
tions for the drug, the biological mechanism that it will work by blocking voltage-gated sodium channels, but there
modify, the pharmacokinetic properties that it will possess, are many molecular subtypes of these, and we do not know
and – the definition addressed in this chapter – the mole­ which are relevant to the therapeutic effect. Similarly,
cular recognition site to which the drug will bind. For the antipsychotic drugs block receptors for several amine
great majority of existing drugs the target is a protein mediators (dopamine, serotonin, norepinephrine, acetyl­
molecule, most commonly a receptor, an enzyme, a trans­ choline), for each of which there are several molecular
port molecule or an ion channel, although other proteins subtypes expressed in the brain. Again, we cannot pinpoint
such as tubulin and immunophilins are also represented. the relevant one or ones that represent the critical targets.
Some drugs, such as alkylating agents, bind to DNA, and A compendium of therapeutic drugs licensed in the UK
others, such as bisphosphonates, to inorganic bone matrix and/or US (Dollery, 1999) lists a total of 857 compounds
constituents, but these are exceptions. The search for new (Figure 6.1), of which 626 are directed at human targets.
drug targets is directed mainly at finding new proteins Of the remainder, 142 are drugs used to treat infectious
although approaches aimed at gene-silencing by antisense diseases, and are directed mainly at targets expressed by
or siRNA have received much recent attention. the infecting organism, and 89 are miscellaneous agents
Since the 1950s, when the principle of drug discovery such as vitamins, oxygen, inorganic salts, plasma substi­
based on identified (then pharmacological or biochemi­ tutes, etc., which are not target-directed in the conven­
cal) targets became established, the pharmaceutical indus­ tional sense.
try has recognized the importance of identifying new Drews and Ryser (1997) estimated that these drugs
targets as the key to successful innovation. As the industry addressed approximately 500 distinct human targets, but
has become more confident in its ability to invent or dis­ this figure includes all of the molecular subtypes of the
cover candidate molecules once the target has been defined generic target (e.g. 14 serotonin receptor subtypes and
– a confidence, some would say, based more on hubris seven opioid receptor subtypes, most of which are not
than history – the selection of novel targets, preferably known to be therapeutically relevant, or to be specifically
exclusive and patentable ones, has assumed increasing targeted by existing drugs), as well as other putative targets,
importance in the quest for competitive advantage. such as calcineurin, whose therapeutic relevance is unclear.
In this chapter we look at drug targets in more detail, Drews and Ryser estimated that the number of potential
and discuss old and new strategies for seeking out and targets might be as high as 5000–10 000. Adopting a
validating them. similar but more restrictive approach, Hopkins and Groom

© 2012 Elsevier Ltd. 63

Section | 2 | Drug Discovery

89

142
Drugs directed at human
targets

Drugs for infectious diseases

Miscellaneous agents

626
Drugs by category (n = 857)

8
5
36 G-protein coupled receptors
11
Ionotropic receptors
Kinase-linked receptors
Nuclear receptors
Enzymes
Transporters
9 Ion channels
38
Miscellaneous
9
Human target classes
9

Fig. 6.1  Therapeutic drug targets.

(2002) found that 120 drug targets accounted for the were already well known. The small target base of existing
activities of compounds used therapeutically, and esti­ drugs, and the low rate of emergence of new targets, sug­
mated that 600–1500 ‘druggable’ targets exist in the gests that the number yet to be discovered may be consider­
human genome. From an analysis of all prescribed drugs ably smaller than some optimistic forecasters have
produced by the 10 largest pharmaceutical companies, predicted. They estimated that 100–150 high-quality
Zambrowicz and Sands (2003) identified fewer than 100 targets in the human genome might remain to be discov­
targets; when this analysis was restricted to the 100 best- ered. However, a recent analysis of 216 therapeutic drugs
selling drugs on the market, the number of targets was registered during the decade 2000–2009 (Rang unpub­
only 43, reflecting the fact that more than half of the suc­ lished) showed that 62 (29%) were first in class com­
cessful targets (i.e. those leading to therapeutically effective pounds directed at novel human targets. This is clearly an
and safe drugs) failed to achieve a significant commercial increase compared with the previous decade and may give
impact. More recently, Imming et al. (2006) identified 218 grounds for optimism that many more new targets remain
targets of registered drugs, including those expressed by to be found. In the earlier analysis by Drews and Ryser
infectious agents, whereas Overington et al. (2006) found (1997), about one-quarter of the targets identified were
324 (the difference depending on the exact definition of associated with drugs used to treat infectious diseases, and
what constitutes a single identified target). How many belonged to the parasite, rather than the human, genome.
targets remain is highly uncertain and Zambrowicz and Their number is even harder to estimate. Most anti-infective
Sands (2003) suggested that during the 1990s only two or drugs have come from natural products, reflecting the fact
three new targets were exploited by the 30 or so new drugs that organisms in the natural world have faced strong
registered each year, most of which addressed targets that evolutionary pressure, operating over millions of years, to

64
 Choosing the target Chapter |6|

develop protection against parasites. It is likely, therefore, The above discussion relates to drug targets expressed by
that the ‘druggable genome’ of parasitic microorganisms human cells, and similar arguments apply to those of infec­
has already been heavily trawled – more so than that of tive organisms. The therapy of infectious diseases, ranging
humans, in whom many of the diseases of current concern from viruses to multicellular parasites, is one of medicine’s
are too recent for evolutionary counter-measures to have greatest challenges. Current antibacterial drugs – the largest
developed. class – originate mainly from natural products (with a few,
More recent estimates of the number of possible drug e.g. sulfonamides and oxazolidinediones, coming from
targets (Overington et al., 2006; Imming et al., 2006) are synthetic compounds) first identified through screening on
in line with the figure of 600–1500 suggested by Hopkins bacterial cultures, and analysis of their biochemical mecha­
and Groom (2002). The increased use of gene deletion nism and site of action came later. In many cases this
mutant mice and si RNA for gene knock down coupled knowledge remains incomplete, and the molecular targets
with a bioinformatic approach to protein characterization are still unclear. Since the pioneering work of Hitchings
is starting to increase the reliability of target identification and Elion (see Chapter 1), the strategy of target-directed
(Imming et al., 2006; Wishart et al., 2008; Bakheet and drug discovery has rarely been applied in this field1;
Doig, 2009). New approaches, in which networks of genes instead, the ‘antibiotic’ approach, originating with the dis­
are associated with particular disease states, are likely to covery of penicillin, has held sway. For the 142 current
throw up additional targets that may not have been dis­ anti-infective drugs (Figure 6.1), which include antiviral
covered otherwise (e.g. see Emilsson et al., 2008). and antiparasitic as well as antibacterial drugs, we can
identify approximately 40 targets (mainly enzymes and
structural proteins) at the biochemical level, only about
The nature of existing drug targets half of which have been cloned. The urgent need for drugs
The human targets of the 626 drugs, where they are known, that act on new targets arises because of the major problem
are summarized in Figure 6.1. Of the 125 known targets, of drug resistance, with which the traditional chemistry-led
the largest groups are enzymes and G-protein-coupled strategies are failing to keep up. Consequently, as with
receptors, each accounting for about 30%, the remainder other therapeutic classes, genomic technologies are being
being transporters, other receptor classes, and ion chan­ increasingly applied to the problem of finding new anti­
nels. This analysis gives only a crude idea of present-day microbial drug targets (Rosamond and Allsop, 2000;
therapeutics, and underestimates the number of molecular Buysse, 2001). In some ways the problem appears simpler
targets currently addressed, since it fails to take into than finding human drug targets. Microbial genomes are
account the many subtypes of these targets that have been being sequenced at a high rate, and identifying prokaryotic
identified by molecular cloning. In most cases we do not genes that are essential for survival, replication or patho­
know which particular subtypes are responsible for the genicity is generally easier than identifying genes that are
therapeutic effect, and the number of distinct targets will involved in specific regulatory mechanisms in eukaryotes.
certainly increase as these details become known. Also
there are many drugs whose targets we do not yet know.
There are a growing number of examples where drug CONVENTIONAL STRATEGIES FOR
targets consist, not of a single protein, but of an oligomeric FINDING NEW DRUG TARGETS
assembly of different proteins. This is well established for
ion channels, most of which are hetero-oligomers, and
recent studies show that G-protein-coupled receptors Two main routes have been followed so far:
(GPCRs) may form functional dimers (Bouvier, 2001) or • Analysis of pathophysiology
may associate with accessory proteins known as RAMPs • Analysis of mechanism of action of existing
(McLatchie et al., 1998; Morphis et al., 2003), which therapeutic drugs.
strongly influence their pharmacological characteristics. There are numerous examples where the elucidation
The human genome is thought to contain about 1000 of pathophysiological pathways has pointed to the exist­
genes in the GPCR family, of which about one-third could ence of novel targets that have subsequently resulted in
be odorant receptors. Excluding the latter, and without successful drugs, and this strategy is still adopted by most
taking into account the possible diversity factors men­ pharmaceutical companies. To many scientists it seems
tioned, the 40 or so GPCR targets for existing drugs repre­ the safe and logical way to proceed – first understand
sent only about 7–10% of the total. Roughly one-third of the pathway leading from the primary disturbance to the
cloned GPCRs are classed as ‘orphan’ receptors, for which
no endogenous ligand has yet been identified, and these 1
A notable exception is the discovery (Wengelnik et al., 2002) of a
could certainly emerge as attractive drug targets when more new class of antimalarial drug designed on biochemical principles to
is known about their physiological role. Broadly similar interfere with choline metabolism, which is essential for membrane
synthesis during the multiplicatory phase of the malaria organism
conclusions apply to other major classes of drug target, within the host’s red blood cells. Hitchings and Elion would have
such as nuclear receptors, ion channels and kinases. approved.

65
Section | 2 | Drug Discovery

appearance of the disease phenotype, then identify par­ biochemical approach led to a remarkable series of thera­
ticular biochemical steps amenable to therapeutic inter­ peutic breakthroughs in antibacterial, anticancer and
vention, then select key molecules as targets. The pioneers immunosuppressant drugs. The work of Black and his col­
in developing this approach were undoubtedly Hitchings leagues, based on mediators and receptors, also described
and Elion (see Chapter 1), who unravelled the steps in in Chapter 1, was another early and highly successful
purine and pyrimidine biosynthesis and selected the example of this approach, and there have been many
enzyme dihydrofolate reductase as a suitable target. This others (Table 6.1a). In some cases the target has emerged

Table 6.1  Examples of drug targets identified, (a) by analysis of pathophysiology, and (b) by analysis of
existing drugs

(a) Targets identified via pathophysiology

Disease indication Target identified Drugs developed

AIDS Reverse transcriptase HIV protease Zidovudine
Saquinavir
Asthma Cysteinyl leukotriene receptor Zafirlukast
Bacterial infections Dihydrofolate reductase Trimethoprim
Malignant disease Dihydrofolate reductase 6-mercaptopurine
Methotrexate
Depression 5HT transporter Fluoxetine
Hypertension Angiotensin-converting enzyme Captopril
Type 5 phosphodiesterase Sildenafil
Angiotensin-2 receptor Losartan
Inflammatory disease COX-2 Celecoxib
Rofecoxib
Alzheimer’s disease Acetylcholinesterase Donepezil
Breast cancer Oestrogen receptor Tamoxifen
Herceptin
Chronic myeloid leukemia Abl kinase Imatinib
Parkinson’s disease Dopamine synthesis Levodopa
MAO-B Selegiline
Depression MAO-A Moclobemide

(b) Targets identified via drug effects

Drug Disease Target

Benzodiazepines Anxiety, sleep disorders BDZ binding site on GABAA receptor
Aspirin-like drugs Inflammation, pain COX enzymes
Ciclosporin, FK506 Transplant rejection Immunophilins
Vinca alkaloids Cancers Tubulin
Dihydropyridines Cardiovascular disease L-type calcium channels
Sulfonylureas Diabetes KATP channels
Classic antipsychotic drugs Schizophrenia Dopamine D2 receptor
Tricyclic antidepressants Depression Monoamine transporters
Fibrates Raised blood cholesterol PPARα

66
 Choosing the target Chapter |6|

Drug effects

Expression Regulatory Compensatory

level factors mechanisms

Gene Functional Physiological Disease

Genes
products activity effects phenotype

Fig. 6.2  Drug effects in relation to disease aetiology.

from pharmacological rather than pathophysiological phenotype, namely by altering gene expression, by altering
studies. The 5HT3 receptor was identified from pharmaco­ the functional activity of gene products, or by activating
logical studies and chosen as a potential drug target for compensatory mechanisms. This is of course an oversim­
the development of antagonists, but its role in pathophysi­ plification, as changes in gene expression or the activation
ology was at the time far from clear. Eventually animal and of compensatory mechanisms are themselves indirect
clinical studies revealed the antiemetic effect of drugs such effects, following pathways similar to that represented by
as ondansetron, and such drugs were developed mainly to the primary track in Figure 6.2. Nevertheless, it provides a
control the nausea and vomiting associated with cancer useful framework for discussing some of the newer
chemotherapy. Similarly, the GABAB receptor (target for genomics-based approaches. A useful account of the
relaxant drugs such as baclofen) was discovered by analysis various genetic models that have been developed in dif­
of the pharmacological effects of GABA, and only later ferent organisms for the identification of new drug targets,
exploited therapeutically to treat muscle spasm. and elucidating the mechanisms of action of existing
The identification of drug targets by the ‘backwards’ drugs, is given by Carroll and Fitzgerald (2003).
approach – involving analysis of the mechanism of action
of empirically discovered therapeutic agents – has pro­ Trawling the genome
duced some major breakthroughs in the past (see exam­
ples in Table 6.1b). Its relevance is likely to decline as drug The conventional route to new drug targets, starting from
discovery becomes more target focused, though natural pathophysiology, will undoubtedly remain as a standard
product pharmacology will probably continue to reveal approach. Understanding of the disease mechanisms in
novel drug targets. the major areas of therapeutic challenge, such as Alzhe­
It is worth reminding ourselves that there remain several imer’s disease, atherosclerosis, cancer, stroke, obesity, etc.,
important drug classes whose mechanism of action we still is advancing rapidly, and new targets are continually
do not fully understand (e.g. acetaminophen (paraceta­ emerging. But this is generally slow, painstaking,
mol)2, valproate). Whether their targets remain elusive hypothesis-driven work, and there is a strong incentive to
because their effects depend on a cocktail of interactions seek shortcuts based on the use of new technologies to
at several different sites, or whether novel targets will select and validate novel targets, starting from the genome.
emerge for such drugs, remains uncertain. In principle, it is argued, nearly all drug targets are pro­
teins, and are, therefore, represented in the proteome, and
also as corresponding genes in the genome. There are,
New strategies for identifying  
however, some significant caveats, the main ones being the
drug targets following:
Figure 6.2 summarizes the main points at which drugs • Splice variants may result in more than one
may intervene along the pathway from genotype to pharmacologically distinct type of receptor being
encoded in a single gene. These are generally
predictable from the genome, and represented as
2
COX-3, a novel splice variant of cyclooxygenase-1, was suggested as distinct species in the transcriptome and proteome.
apossible target underlying the analgesic action of acetaminophen
(Chandrasekharan et al., 2002), spurring several companies to set up • There are many examples of multimeric receptors,
screens for new acetaminophen-like drugs. This now seems to have made up of non-identical subunits encoded by
been a dry well and recent suggestions on mechanism have included different genes. This is true of the majority of
activation of TRPV1 and cannabinoid receptors by a metabolite
(Mallet et al., 2010) and the inhibition of prostaglandin H synthase ligand-gated ion channels, whose pharmacological
(Aronoff et al., 2006). characteristics depend critically on the subunit

67
Section | 2 | Drug Discovery

composition (Hille, 2001). Recent work also shows of rare inherited disorders of this type. Information of this
that G-protein-coupled receptors often exist as kind, important though it is for the diagnosis, manage­
heteromeric dimers, with pharmacological properties ment and counselling of these patients, has so far had little
distinct from those of the individual units (Bouvier, impact on the selection of drug targets. None of the gene
2001). Moreover, association between receptors products identified appears to be directly ‘targetable’.
and non-receptor proteins can determine the Much more common than single-gene disorders are con­
pharmacological characteristics of certain G-protein- ditions such as diabetes, hypertension, schizophrenia,
coupled receptors (McLatchie et al., 1998). bipolar depressive illness and many cancers in which there
Despite these complications, studies based on the is a clear genetic component, but, together with environ­
appealingly simple dogma: mental factors, several different genes contribute as risk
factors for the appearance of the disease phenotype. The
one gene → one protein → one drug target
methods for identifying the particular genes involved (see
have been extremely productive in advancing our knowl­ Chapter 7) were until recently difficult and laborious, but
edge of receptors and other drug targets in recent years, are becoming much easier as the sequencing and annota­
and it is expected that genome trawling will reveal more tion of the human genome progresses. The Human
‘single-protein’ targets, even though multiunit complexes Genome Consortium (2001) found 971 ‘disease genes’
are likely to escape detection by this approach. already listed in public databases, and identified 286 para­
How might potential drug targets be recognized among logues of these in the sequenced genome. Progress has
the 25 000 or so genes in the human genome? been rapid and Lander (2011) contrasts the situation in
Several approaches have been described for homing in 2000, when only about a dozen genetic variations outside
on genes that may encode novel drug targets, starting with the HLA locus had been associated with common diseases,
the identification of certain gene categories: with today, when more than 1100 loci associated with 165
• ’Disease genes’, i.e. genes, mutations of which diseases have been identified. Notably most of these have
cause or predispose to the development of human been discovered since 2007. The challenge of obtaining
disease. useful drug targets from genome-wide association studies
• ’Disease-modifying’ genes. These comprise (a) genes remains considerable (Donnelly, 2008).
whose altered expression is thought to be involved The value of information about disease genes in better
in the development of the disease state; and (b) understanding the pathophysiology is unquestionable,
genes that encode functional proteins, whose activity but just how useful is it as a pointer to novel drug targets?
is altered (even if their expression level is not) in the Many years after being identified, the genes involved in
disease state, and which play a part in inducing the several important single-gene disorders, such as thalas­
disease state. saemia, muscular dystrophy and cystic fibrosis, have not
• ’Druggable genes’, i.e. genes encoding proteins likely so far proved useful as drug targets, although progress is
to possess binding domains that recognize drug-like at last being made. On the other hand, the example of
small molecules. Included in this group are genes Abl-kinase, the molecular target for the recently intro­
encoding targets for existing therapeutic and duced anticancer drug imatinib (Gleevec; see Chapter 4),
experimental drugs. These genes and their paralogues shows that the proteins encoded by mutated genes can
(i.e. closely related but non-identical genes occurring themselves constitute drug targets, but so far there are few
elsewhere in the genome) comprise the group of instances where this has proved successful. The findings
druggable genes. that rare forms of familial Alzheimer’s disease were associ­
On this basis (Hopkins and Groom, 2002), novel targets ated with mutations in the gene encoding the amyloid
are represented by the intersection of disease-modifying precursor protein (APP) or the secretase enzyme responsi­
and druggable gene classes, excluding those already tar­ ble for formation of the β-amyloid fragment present in
geted by therapeutic drugs. Next, we consider these catego­ amyloid plaques were strong pointers that confirmed the
ries in more detail. validity of secretase as a drug target (although it had
already been singled out on the basis of biochemical
studies). Secretase inhibitors reached late-stage clinical
Disease genes development as potential anti-Alzheimer’s drugs in 2001,
The identification of genes in which mutations are associ­ but none at the time of writing have emerged from the
ated with particular diseases has a long history in medicine development pipeline as useful drugs. Similar approaches
(Weatherall, 1991), starting with the concept of ‘inborn have been successful in identifying disease-associated
errors of metabolism’ such as phenylketonuria. The strate­ genes in defined subgroups of patients with conditions
gies used to identify disease-associated genes, described in such as diabetes, hypertension and hypercholesterolae­
Chapter 7, have been very successful. Examples of common mia, and there is reason to hope that novel drug targets
diseases associated with mutations of a single gene are will emerge in these areas. However, in other fields, such
summarized in Table 7.3. There are many more examples as schizophrenia and asthma, progress in pinning down

68
 Choosing the target Chapter |6|

disease-related genes has been very limited. The most Gene expression profiling
promising field for identifying novel drug targets among
The principle underlying gene expression profiling as a
disease-associated mutations is likely to be in cancer thera­
guide to new drug targets is that the development of any
pies, as mutations are the basic cause of malignant trans­
disease phenotype necessarily involves changes in gene
formation and real progress is being made in exploiting
expression in the cells and tissues involved. Long-term
our knowledge of the cancer genome (Roukos, 2011). In
changes in the structure or function of cells cannot occur
general, one can say that identifying disease genes may
without altered gene expression, and so a catalogue of all
provide valuable pointers to possible drug targets further
the genes whose expression is up- or down-regulated in
down the pathophysiological pathway, even though their
the disease state will include genes where such regulation
immediate gene products may not always be targetable.
is actually required for the development of the disease
The identification of a new disease gene often hits the
phenotype. As well as these ‘critical path genes’, which
popular headlines on the basis that an effective therapy
may represent potential drug targets, others are likely to
will quickly follow, though this rarely happens, and never
be affected as genes involved in secondary reinforcing of
quickly.
compensatory mechanisms following the development of
In summary, the class of disease genes does not seem to
the disease phenotype (which may also represent potential
include many obvious drug targets.
drug targets), but many will be irrelevant ‘bystander genes’
(Figure 6.3). The important problem, to which there is no
Disease-modifying genes simple answer, is how to eliminate the bystanders, and
In this class lie many non-mutated genes that are directly how to identify potential drug targets among the rest.
involved in the pathophysiological pathway leading to the Zanders (2000), in a thoughtful review, emphasizes the
disease phenotype. The phenotype may be associated with importance of focusing on signal transduction pathways
over- or underexpression of the genes, detectable by as the most likely source of novel targets identifiable by
expression profiling (see below), or by the over- or under­ gene expression profiling.
activity of the gene product – for example, an enzyme – The methods available for gene expression profiling are
independently of changes in its expression level. described in Chapter 7. DNA microarrays (‘gene chips’)
This is the most important category in relation to drug are most commonly used, their advantage being that they
targets, as therapeutic drug action generally occurs by are quick and easy to use. They have the disadvantage that
changing the activity of functional proteins, whether or the DNA sequences screened are selected in advance, but
not the disease alters their expression level. Finding new this is becoming less of a limitation as more genomic
ones, however, is not easy, and there is as yet no shortcut information accumulates. They also have limited sensitiv­
screening strategy for locating them in the genome. ity, which (see Chapter 7) means that genes expressed at
Two main approaches are currently being used, namely low levels – including, possibly, a significant proportion
gene expression profiling and comprehensive gene knock­ of potential drug targets – can be missed. Methods based
out studies. on the polymerase chain reaction (PCR), such as serial

Reinforcing genes

Environmental factors

Altered gene Critical path genes Functional Disease

expression changes phenotype

Genetic make-up
Bysta

Compensator y genes
nde
rg

ne
s
e

No disease-related effect

Fig. 6.3  Gene expression in disease.

69
Section | 2 | Drug Discovery

analysis of gene expression (SAGE; Velculescu et al., 1995), development of the disease phenotype. With many
are, in principle, capable of detecting all transcribed RNAs acute interventions, several distinct temporal patterns
without preselection, with higher sensitivity than microar­ are detectable in the expression of different genes.
rays, but are more laborious and technically demanding. ‘Clustering’ algorithms (see Chapter 7; Quackenbush,
Both technologies produce the same kind of data, namely 2001) are useful to identify genes which are
a list of genes whose expression is significantly altered as co-regulated and whose function might point to a
a result of a defined experimental intervention. A large particular biochemical or signalling pathway relevant
body of gene expression data has been collected over the to the pathogenesis.
last few years, showing the effects of many kinds of distur­ • The study of several different clinical and animal
bance, including human disease, animal models of disease models with similar disease phenotypes can reveal
and drug effects; some examples are listed in Table 6.2. In networks of genes whose expression is consistently
most cases, several thousand genes have been probed with affected and hence representative of the phenotype
microarrays, covering perhaps 20% of the entire genome, (see Emilsson et al. 2008; Schadt 2009).
so the coverage is by no means complete. Commonly, it • The effects of drug or other treatments can be
is found that some 2–10% of the genes studied show studied in order to reveal genes whose expression is
significant changes in response to such disturbances. Thus, normalized in parallel with the ‘therapeutic’ effect.
given a total of about 25 000 genes in man, with perhaps
20% expressed in a given tissue, we can expect several
hundred to be affected in a typical disease state. Much Gene knockout screening
effort has gone into (a) methods of analysis and pattern Another screening approach for identifying potential drug
recognition within these large data sets, and (b) defining target genes, based on generating transgenic ‘gene knock­
the principles for identifying possible drug targets within out’ strains of mice, as performed by biotechnology
such a group of regulated genes. Techniques for analysis company Lexicon Pharmaceuticals (Walke et al., 2001;
and pattern recognition fall within the field of bioinfor­ Zambrowicz and Sands, 2003), aimed to study 5000 genes
matics (Chapter 7), and the specific problem of analysing within 5 years. The 5000 genes were selected on several
transcription data is discussed by Quackenbush (2001) criteria, including existing evidence of association with
and Butte (2002), and also Lander (2011). Because altered disease and membership of families of known drug targets
mRNA levels do not always correlate with changes in (GPCRs, transporters, kinases, etc.). The group has devel­
protein expression, protein changes should be confirmed oped an efficient procedure for generating transgenic
independently where possible. This normally entails the knockout mice, and a standardized set of tests to detect
production and validation of an antibody-based assay or phenotypic effects relevant to a range of disease states. In
staining procedure. a recent review, Zambrowicz and Sands (2003) present
Identifying potential drug targets among the array of many examples where the effects of gene knockouts in
differentially expressed genes that are revealed by expres­ mice produce effects consistent with the known actions
sion profiling is neither straightforward nor exact. Possible and side effects of therapeutic drugs. For example, inactiva­
approaches are: tion of the genes for angiotensin converting enzyme (ACE)
or the angiotensin receptor results in lowering of blood
• On the basis of prior biological knowledge, genes pressure, reflecting the clinical effect of drugs acting on
that might plausibly be critical in the pathogenesis these targets. Similarly, elimination of the gene encoding
of the disease can be distinguished from those (e.g. one of the subunits of the GABAA receptor produces a state
housekeeping genes, or genes involved in of irritability and hyperactivity in mice, the opposite of the
intermediary metabolism) that are unlikely to be effect of benzodiazepine tranquillizers, which enhance
critical. This, the plausibility criterion, is in practice GABAA receptor function. However, using transgenic gene
the most important factor in identifying likely drug knockout technology (sometimes dubbed ‘reverse genet­
targets. As genomic analysis proceeds, it is becoming ics’) to confirm the validity of previously well-established
possible to group genes into functional classes, e.g. drug targets is not the same as using it to discover new
tissue growth and repair, inflammation, myelin targets. Success in the latter context will depend greatly on
formation, neurotransmission, etc., and changes in the ability of the phenotypic tests that are applied to detect
expression are often concentrated in one or more of therapeutically relevant changes in the knockout strains.
these functional groups, giving a useful pointer to Some new targets have already been identified in this way,
the biological pathways involved in pathogenesis. and deployed in drug discovery programmes; they include
• Anatomical studies can be used to identify whether cathepsin K, a protease involved in the genesis of osteoporo­
candidate genes are regulated in the cells and tissues sis, and melanocortin receptors, which may play a role in
affected by the disease. obesity. The Lexicon group has developed a systems-based
• Timecourse measurements reveal the relationship panel of primary screens to look for relevant effects on the
between gene expression changes and the main physiological systems (cardiovascular, CNS, immune

70
 Choosing the target Chapter |6|

Table 6.2  Examples of gene expression profiling studies, showing numbers of mRNAs affected under
various conditions

Test system Method Number Number of Number Number Ref

of genes genes up down
analysed affected
(threshold)
Activated vs quiescent Microarray 8600 517 (2.2-fold)) 148 242 Iyer et al
human fibroblasts (Some genes (1999)
showed both
up- and down-
regulation at
different times)
Differentiated vs Microarray 8700 156 (twofold) 94 62 Somogyi
undifferentiated mouse (1999)
neural stem cells
Small diameter vs large Microarray 477 40 (1.5-fold) 26 14 Luo et al.
diameter rat sensory (1999)
neurons
Ischaemic vs control rat Microarray 1179 20 (1.8-fold) 6 14 Keyvani
brain et al. (2002)
Inflamed vs normal Microarray 6000 470 (twofold) Kaminski
mouse lung (Many genes et al. (2000)
showed both
up- and down-
regulation at
different times)
Old vs young mouse Microarray 6347 110 (1.7-fold) 63 47 Lee et al.
brain (cortex) (2000)
Old vs young mouse Microarray 6300 113 (twofold) 58 55 Lee et al.
skeletal muscle (1999)
Human colorectal SAGE ~49 000 289 (difference 108 181 Zhang et al.
cancer vs normal tissue significant at (1997)
p < 0.01)
Human MS plaques vs Microarray >5000 (~twofold) 29 ND Whitney
normal brain et al. (1999)
Human ovarian cancer Microarray 5766 726 (threefold) 295 431 Wang et al.
vs normal ovary (1999)
Human Alzheimer’s Microarray 6800 18 (1.8-fold) – 18 Ho et al.
disease plaques vs (2001)
non-plaque regions
Mouse liver, thyroid Microarray 2225 55 (twofold) 14 41 Feng et al.
treated vs hypothyroid (2000)
Mouse, hypertophied vs Microarray >4000 47 (~1.8-fold) 21 26 Friddle et al.
normal heart muscle (2000)
Human postmortem Microarray ~7000 179 (1.9-fold) 97 82 Mirnics
prefrontal cortex, et al. (2000)
schizophrenia vs control

Continued

71
Section | 2 | Drug Discovery

Table 6.2  Continued

Test system Method Number Number of Number Number Ref

of genes genes up down
analysed affected
(threshold)
Human postmortem Microarray ~6000 88 (1.4-fold) 71 17 Hakak et al.
prefrontal cortex, (2001)
schizophrenia vs control
Single neurons from Microarray >18 000 4139 (2-fold) 2574 1565 Hemby
human postmortem et al. (2002)
entorhinal cortex,
schizophrenia vs control

system, etc.). On the basis of their experience with the first

750 of the planned 5000 gene knockouts, they predicted TARGET VALIDATION
that about 100 new high-quality targets could be revealed
in this way. Programmes such as this, based on mouse The techniques discussed so far are aimed at identifying
models, are time-consuming and costly, but have the great potential drug targets within the diversity warehouse rep­
advantage that mouse physiology is fairly similar to resented by the genome, the key word being ‘potential’.
human. The use of species such as flatworm (Caenorhabditis Few companies will be willing to invest the considerable
elegans) and zebra-fish (Danio rerio) is being explored as a resources needed to mount a target-directed drug discov­
means of speeding up the process (Shin and Fishman, ery project without more direct evidence that the target is
2002) and more recently studies in yeast have successfully an appropriate one for the disease indication. Target vali­
facilitated the discovery of molecules that interact with dation refers to the experimental approaches by which a
mammalian drug targets (Brown et al., 2011). potential drug target can be tested and given further cred­
ibility. It is an open-ended term, which can be taken to
’Druggable’ genes embrace virtually the whole of biology, but for practical
purposes the main approaches are pharmacological and
For a gene product to serve as a drug target, it must possess genetic.
a recognition site capable of binding small molecules. Although these experimental approaches can go a long
Hopkins and Groom (2002) found a total of 399 protein way towards supporting the validity of a chosen target, the
targets for registered and experimental drug molecules. ultimate test is in the clinic, where efficacy is or is not
The 399 targets belong to 130 distinct protein families, confirmed. Lack of clinical efficacy causes the abandon­
and the authors suggest that other members of these fami­ ment of roughly one-third of drugs in Phase II, reflecting
lies – a total of 3051 proteins – are also likely to possess the unreliability of the earlier surrogate evidence for target
similar binding domains, even though specific ligands validity.
have not yet been described, and propose that this total
represents the current limit of the druggable genome. Of
course, it is likely that new targets, belonging to different
Pharmacological approaches
protein families, will emerge in the future, so this number The underlying question to be addressed is whether drugs
may well expand. ‘Druggable’ in this context implies only that influence the potential drug target actually produce
that the protein is likely to possess a binding site for a the expected effects on cells, tissues or whole animals.
small molecule, irrespective of whether such an interac­ Where a known receptor, for example the metabotropic
tion is likely be of any therapeutic value. To be useful as glutamate receptor (mGluR), was identified as a potential
a starting point for drug discovery, a potential target needs target for a new indication (e.g. pain) its validity could be
to combine ‘druggability’ with disease-modifying proper­ tested by measuring the analgesic effect of known mGluR
ties. One new approach is to look at the characteristics of antagonists in relevant animal models. For novel targets,
a typical human protein drug target (e.g. hydrophobic, of course, no panel of active compounds will normally be
high length, signal motif present) and to then look for available, and so it will be necessary to set up a screening
these characteristics in a non-target set of proteins so as to assay to identify them. Where a company is already active
identify new potential targets (Bakheet and Doig, 2009). in a particular line of research – as was the case with Ciba

72
 Choosing the target Chapter |6|

Geigy in the kinase field – it may be straightforward to validate putative drug targets include a range of recent
refine its assay methods in order to identify selective inhib­ studies on the novel cell surface receptor uPAR (urokinase
itors. Ciba Geigy was able to identify active Abl kinase plasminogen-activator receptor; see review by Wang,
inhibitors within its existing compound collection, and 2001). This receptor is expressed by certain malignant
hence to show that these were effective in cell proliferation tumour cells, particularly gliomas, and antisense studies
assays. have shown it to be important in controlling the tendency
A variant of the pharmacological approach is to use of these tumours to metastasize, and therefore to be a
antibodies raised against the putative target protein, rather potential drug target. In a different field, antisense studies
than small-molecule inhibitors. have supported the role of a recently cloned sodium
Many experts predict that, as high-throughput screening channel subtype, PN3 (now known to be Nav 1.87)
and automated chemistry develop (see Chapters 8 and 9), (Porreca et al., 1999) and of the metabotropic glutamate
allowing the rapid identification of families of selective receptor mGluR1 (Fundytus et al., 2001) in the pathogen­
compounds, these technologies will be increasingly used esis of neuropathic pain in animal models. Antisense oli­
as tools for target validation, in parallel with lead finding. gonucleotides have the advantage that their effects on gene
‘Hits’ from screening that show a reasonable degree of expression are acute and reversible, and so mimic drug
target selectivity, whether or not they represent viable lead effects more closely than, for example, the changes seen
compounds, can be used in pharmacological studies in transgenic animals (see below), where in most cases the
designed to test their efficacy in a selection of in vitro and genetic disturbance is present throughout life. It is likely
in vivo models. Showing that such prototype compounds, that, as more experience is gained, antisense methods
regardless of their suitability as leads, do in fact produce based on synthetic oligonucleotides will play an increas­
the desired effect greatly strengthens the argument for ingly important role in drug target validation.
target validity. Although contemporary examples are gen­
erally shrouded in confidentiality, it is clear that this
approach is becoming more common, so that target vali­ RNA interference (RNAi)
dation and lead finding are carried out simultaneously. This technique depends on the fact that short lengths of
More recently the move has been towards screening double-stranded RNA (short interfering RNAs, or siRNAs)
enriched libraries containing compounds of a particular activate a sequence-specific RNA-induced silencing complex
chemical phenotype known to, for example, be disposed (RISC), which destroys by cleavage the corresponding
towards kinase inhibition rather than screening entire functional mRNA within the cell (Hannon, 2000; Kim,
compound collections in a random fashion. 2003). Thus specific mRNAs or whole gene families can
be inactivated by choosing appropriate siRNA sequences.
Gene silencing by this method is highly efficient, particu­
Genetic approaches larly in invertebrates such as Caenorhabditis elegans and
These approaches involve various techniques for suppress­ Drosophila, and can also be used in mammalian cells
ing the expression of specific genes to determine whether and whole animals. Its use for studying gene function and
they are critical to the disease process. This can be done validating potential drug targets is increasing rapidly, and
acutely in genetically normal cells or animals by the use it also has potential for therapeutic applications. It has
of antisense oligonucleotides or RNA interference, or con­ proved to be a rapid and effective tool for discovering new
stitutively by generating transgenic animals in which the targets and automated cell assays have allowed the rapid
genes of interest are either overactive or suppressed. screening of large groups of genes.

Antisense oligonucleotides Transgenic animals

Antisense oligonucleotides (Phillips, 2000; Dean, 2001) The use of the gene knockout principle as a screening
are stretches of RNA complementary to the gene of inter­ approach to identify new targets is described above. The
est, which bind to cellular mRNA and prevent its transla­ same technology is also valuable, and increasingly being
tion. In principle this allows the expression of specific used, as a means of validating putative targets. In principle,
genes to be inhibited, so that their role in the development deletion or overexpression of a specific gene in vivo can
of a disease phenotype can be determined. Although provide a direct test of whether or not it plays a role in the
simple in principle, the technology is subject to many sequence of events that gives rise to a disease phenotype.
pitfalls and artefacts in practice, and attempts to use it, The generation of transgenic animal – mainly mouse –
without very careful controls, to assess genes as potential strains is, however, a demanding and time-consuming
drug targets are likely to give misleading results. As an process (Houdebine, 1997; Jackson and Abbott, 2000).
alternative to using synthetic oligonucleotides, antisense Therefore, although this technology has an increasingly
sequences can be introduced into cells by genetic engineer­ important role to play in the later stages of drug
ing. Examples where this approach has been used to discovery and development, some have claimed it is too

73
Section | 2 | Drug Discovery

cumbersome to be used routinely at the stage of target model that replicated the amyloid deposits typical of this
selection, though Harris and Foord (2000) predicted that disease. In summary, transgenic animal models are often
high-throughput ‘transgenic factories’ may be used in this helpful for post hoc target validation, but their main – and
way. One problem relates to the genetic background of the increasing – use in drug discovery comes at later stages of
transgenic colony. It is well known that different mouse the project (see Chapter 11).
strains differ in many significant ways, for example in their
behaviour, susceptibility to tumour development, body
weight, etc. For technical reasons, the strain into which the
transgene is introduced is normally different from that SUMMARY AND CONCLUSIONS
used to establish the breeding colony, so a protocol of
back-crossing the transgenic ‘founders’ into the breeding In the present drug discovery environment most projects
strain has to proceed for several generations before a begin with the identification of a molecular target, usually
genetically homogeneous transgenic colony is obtained. one that can be incorporated into a high-throughput
Limited by the breeding cycle of mice, this normally takes screening assay. Drugs currently in therapeutic use cover
about 2 years. It is mainly for this reason that the genera­ about 200–250 distinct human molecular targets, and the
tion of transgenic animals for the purposes of target vali­ great majority of new compounds registered in the last
dation is not usually included as a stage on the critical decade are directed at targets that were already well known;
path of the project. Sometimes, studies on transgenic on average, about three novel targets are covered by drugs
animals tip the balance of opinion in such a way as to registered each year. The discovery and exploitation of new
encourage work on a novel drug target. For example, the targets is considered essential for therapeutic progress and
vanilloid receptor, TRPV1, which is expressed by nocicep­ commercial success in the long term. Estimates from
tive sensory neurons, was confirmed as a potential drug genome sequence data of the number of potential drug
target when the knockout mouse proved to have a marked targets, defined by disease relevance and ‘druggability’,
deficit in the development of inflammatory hyperalgesia suggest that from about 100 to several thousand ‘drugga­
(Davis et al., 2000), thereby confirming the likely involve­ ble’ new targets remain to be discovered. The uncertainty
ment of this receptor in a significant clinical condition. reflects two main problems: the difficulty of recognizing
Most target-directed projects, however, start on the basis ‘druggability’ in gene sequence data, and the difficulty of
of other evidence for (or simply faith in) the relevance of determining the relevance of a particular gene product in
the target, and work on developing transgenic animals the development of a disease phenotype. Much effort is
begins at the same time, in anticipation of a need for them currently being applied to these problems, often taking the
later in the project. In many cases, transgenic animals have form of new ‘omic’ disciplines whose role and status are
provided the most useful (sometimes the only available) not yet defined. Proteomics and structural genomics are
disease models for drug testing in vivo. Thus, cancer expected to improve our ability to distinguish druggable
models based on deletion of the p53 tumour suppressor proteins from the rest, and ‘transcriptomics’ (another
gene are widely used, as are atherosclerosis models based name for gene expression profiling) and the study of trans­
on deletion of the ApoE or LDL-receptor genes. Alzheim­ genic animals will throw light on gene function, thus
er’s disease models involving mutation of the gene for improving our ability to recognize the disease-modifying
amyloid precursor protein, or the presenilin genes, have mechanisms wherein novel drug targets are expected to
also proved extremely valuable, as there was hitherto no reside.

REFERENCES

Aronoff DM, Oates JA, Boutaud O. Bouvier M. Oligomerization of Butte A. The use and analysis of
New insights into the mechanism of G-protein-coupled transmitter microarray data. Nature Reviews
action of acetaminophen: its clinical receptors. Nature Reviews Drug Discovery 2002;1:951–60.
pharmacologic characteristics reflect Neuroscience 2001;2:274–86. Buysse JM. The role of genomics in
its inhibition of the two Brown AJ, Daniels DA, Kassim M, et al. antibacterial drug discovery. Current
prostaglandin H2 synthases. Clinical Pharmacology of GPR-55 in yeast Medicinal Chemistry
Pharmacology and Therapeutics and identification of GSK494581A as 2001;8:1713–26.
2006;79:9–19. a mixed-activity glycine transporter Carroll PM, Fitzgerald K, editors.
Bakheet TM, Doig AJ. Properties and subtype 1 imhibitor and GPR55 Model organisms in drug
identification of human protein agonist. Journal of Pharmacology discovery. Chichester: John Wiley;
drug targets. Bioinformatics 2009; and Experimental Therapeutics 2003.
25:451–7. 2011;337:236–46.

74
 Choosing the target Chapter |6|

Chandrasekharan NV, Dai H, Roos KL, therapeutic target selection. its retardation by caloric restriction.
et al. COX-3, a cyclooxygenase-1 Pharmacogenomics 2000;1:433–43. Science 1999;285:1390–3.
variant inhibited by acetaminophen Hemby SE, Ginsberg SD, Brunk B, et al. Lee CK, Weindruch R, Prolla TA.
and other analgesic/antipyretic Gene expression profile for Gene-expression profile of the ageing
drugs: cloning, structure, and schizophrenia: discrete neuron brain in mice. Nature Genetics
expression. Proceedings of the transcription patterns in the 2000;25:294–7.
National Academy of Sciences of the entorhinal cortex. Archives of Luo L, Salunga RC, Guo H, et al. Gene
USA 2002;99:13926–31. General Psychiatry 2002;59:631–40. expression profiles of laser-captured
Davis JB, Gray J, Gunthorpe MJ, et al. Hille B. Ion channels of excitable adjacent neuronal subtypes. Nature
Vanilloid receptor-1 is essential for membranes. 3rd ed. Sunderland, Medicine 1999;5:117–22.
inflammatory thermal hyperalgesia. MA: Sinauer; 2001. McLatchie LM, Fraser NJ, Main MJ, et al.
Nature 2000;405:183–7. Ho L, Guo Y, Spielman L et al. Altered RAMPs regulate the transport and
Dean NM. Functional genomics and expression of a-type but not b-type ligand specificity of the calcitonin-
target validation approaches using synapsin isoform in the brain of receptor-like receptor. Nature
antisense oligonucleotide patients at high risk for Alzheimer’s 1998;393:333–9.
technology. Current Opinion in disease assessed by DNA microarray Mallet C, Barriere DA, Ermund A, et al.
Biotechnology 2001;12:622–5. technique. Neuroscience Letters TRPV1 in brain is involved in
Dollery CT, editor. Therapeutic drugs. 2001;298:191–4. acetaminophen-induced
Edinburgh: Churchill Livingstone; Hopkins AL, Groom CR. The druggable antinociception. PLoS ONE
1999. genome. Nature Reviews Drug 2010;5:1–11.
Donnelly P. Progress and challenges in Discovery 2002;1:727–30. Mirnics K, Middleton FA, Marquez A,
genome-wide association studies. Houdebine LM. Transgenic animals et al. Molecular characterization of
Nature 2008;456:728–31. – generation and use. Amsterdam: schizophrenia viewed by microarray
Drews J, Ryser S. Classic drug targets. Harwood Academic; 1997. analysis of gene expression in
Nature Biotechnology Imming P, Sinning C, Meyer A. Drugs, prefrontal cortex. Neuron 2000;
1997;15:1318–9. their targets and the nature and 28:53–67.
Emilsson V, Thorleifsson G, Zhang B, number of drug targets. Nature Morphis M, Christopoulos A, Sexton
et al. Genetics of gene expression Reviews. Drug Discovery 2006; PM. RAMPs: 5 years on, where to
and its effect on disease. Nature 5:821–34. now? Trends in Pharmacological
2008;452:423–8. Iyer VR, Eisen MB, Ross DT, et al. The Sciences 2003;24:596–601.
Feng X, Jiang Y, Meltzer P, et al. Thyroid transcriptional program in the Overington JP, Bissan A-L, Hopkins AL.
hormone regulation of hepatic genes response of human fibroblasts to How many drug targets are there?
in vivo detected by complementary serum. Science 1999;283:83–7. Nature Reviews. Drug Discovery
DNA microarray. Molecular Jackson IJ, Abbott CM. Mouse genetics 2006;5:993–6.
Endocrinology 2000;14:947–55. and transgenics. Oxford: Oxford Phillips MI. Antisense technology, Parts
Friddle CJ, Koga T, Rubin EM, et al. University Press; 2000. A and B. San Diego: Academic Press;
Expression profiling reveals distinct Kaminski N, Allard JD, Pittet JF, et al. 2000.
sets of genes altered during Global analysis of gene expression in Porreca F, Lai J, Bian D, et al. A
induction and regression of cardiac pulmonary fibrosis reveals distinct comparison of the potential role of
hypertrophy. Proceedings of the programs regulating lung the tetrodotoxin-insensitive sodium
National Academy of Sciences of the inflammation and fibrosis. channels, PN3/SNS and NaN/SNS2,
USA 2000;97:6745–50. Proceedings of the National in rat models of chronic pain.
Fundytus ME, Yashpal K, Chabot JG, Academy of Sciences of the USA Proceedings of the National
et al. Knockdown of spinal 2000;97:1778–83. Academy of Sciences of the USA
metabotropic receptor 1 (mGluR1) Keyvani K, Witte OW, Paulus W. Gene 1999;96:7640–4.
alleviates pain and restores opioid expression profiling in perilesional Quackenbush J. Computational analysis
efficacy after nerve injury in rats. and contralateral areas after ischemia of microarray data. Nature Reviews
British Journal of Pharmacology in rat brain. Journal of Cerebral Genetics 2001;2:418–27.
2001;132:354–67. Blood Flow and Metabolism Rosamond J, Allsop A. Harnessing the
Hakak Y, Walker JR, Li C, et al. 2002;22:153–60. power of the genome in the search
Genome-wide expression analysis Kim VN. RNA interference in functional for new antibiotics. Science 2000;
reveals dysregulation of myelination- genomics and medicine. Journal of 287:1973–6.
related genes in chronic Korean Medical Science 2003;18: Roukos DH. Trastuzumab and beyond:
schizophrenia. Proceedings of the 309–18. sequencing cancer genomes and
National Academy of Sciences of the Lander ES. Initial impact of the predicting molecular networks.
USA 2001;98:4746–51. sequencing of the human genome. Pharmacogenomics Journal 2011;11:
Hannon GJ. RNA interference. Nature Nature 2011;470:187–97. 81–92.
2000;418:244–51. Lee CK, Klopp RG, Weindruch R, et al. Schadt EE. Molecular networks as
Harris S, Foord SM. Transgenic gene Gene expression profile of aging and sensors and drivers of common
knock-outs: functional genomics and

75
Section | 2 | Drug Discovery

human diseases. Nature 2009;461: activator receptor gene expression in microarrays. Annals of Neurology
218–23. cancer invasion and metastasis. 1999;46:425–8.
Shin JT, Fishman MC. From zebrafish to Medicinal Research Reviews 2001;21: Wishart DS, Knox, C, Guo AC, et al.
human: molecular medical models. 146–70. DrugBank: a knowledgebase for
Annual Review of Genomics and Wang K, Gan L, Jeffery E, et al. drugs, drug actions and drug targets.
Human Genetics 2002;3:311–40. Monitoring gene expression profile Nucleic Acids Research 2008;36:
Somogyi R. Making sense of gene- changes in ovarian carcinomas D901–D906.
expression data. In: Pharma using cDNA microarray. Gene Zambrowicz BP, Sands AT. Knockouts
Informatics, A Trends Guide. Elsevier 1999;229:101–8. model the 100 best-selling drugs
Trends in Biotechnology 1999;17 Weatherall DJ. The new genetica – will they model the next 100?
(Suppl. 1):17–24. and clinical practice. 3rd ed. Nature Reviews. Drug Discovery
Velculescu VE, Zhang L, Vogelstein B, Oxford: Oxford University Press; 2003;2:38–51.
et al. Serial analysis of gene 1991. Zanders ED. Gene expression profiling
expression. Science 1995;270:484–7. Wengelnik K, Vidal V, Ancelin ML, et al. as an aid to the identification of
Walke DW, Han C, Shaw J, et al. In vivo A class of potent antimalarials and drug targets. Pharmacogenomics
drug target discovery: identifying the their specific accumulation in 2000;1:375–84.
best targets from the genome. infected erythrocytes. Science Zhang L, Zhou W, Velculescu VE, et al.
Current Opinion in Biotechnology 2002;295:1311–4. Gene expression profiles in normal
2001;12:626–31. Whitney LW, Becker KG, Tresser NJ, and cancer cells. Science 1997;276:
Wang Y. The role and regulation of et al. Analysis of gene expression in 1268–72.
urokinase-type plasminogen mutiple sclerosis lesions using cDNA

76
 Chapter 7 
The role of information, bioinformatics
and genomics
B Robson, R McBurney

information theoretic network, from the computations of

THE PHARMACEUTICAL INDUSTRY probabilities from genetics and other sources, to the prob-
AS AN INFORMATION INDUSTRY ability that a drug will play a useful role in the market-
place. As may easily be imagined, such a description is
rich, complex, and evolving (and rather formal), and the
As outlined in earlier chapters, making drugs that can chapter was rapidly reaching the size of a book while
affect the symptoms or causes of diseases in safe and ben- barely scratching the surface of such a description. It must
eficial ways has been a substantial challenge for the phar- suffice to concentrate on the nodes (old and new) of that
maceutical industry (Munos, 2009). However, there are network, and largely on those nodes that represent the
countless examples where safe and effective biologically sources and pools of data and information that are brought
active molecules have been generated. The bad news is that together in multiple ways.
the vast majority of these are works of nature, and to This chapter begins with general concepts about infor-
produce the examples we see today across all biological mation then focuses on information about biological
organisms, nature has run a ‘trial-and-error’ genetic algo- molecules and on its relationship to drug discovery and
rithm for some 4 billion years. By contrast, during the last development. Since most drugs act by binding to and
100 years or so, the pharmaceutical industry has been modulating the function of proteins, it is reasonable for
heroically developing handfuls of successful new drugs in pharmaceutical industry scientists to want to have as much
10–15 year timeframes. Even then, the overwhelming information as possible about the nature and behaviour
majority of drug discovery and development projects fail of proteins in health and disease. Proteins are vast in
and, recently, the productivity of the pharmaceutical numbers (compared to genes), have an enormous dynamic
industry has been conjectured to be too low to sustain its range of abundances in tissues and body fluids, and have
current business model (Cockburn, 2007; Garnier, 2008). complex variations that underlie specific functions. At
A great deal of analysis and thought is now focused on present, the available information about proteins is
ways to improve the productivity of the drug discovery and limited by lack of technologies capable of cost-efficiently
development process (see, for example, Paul et al., 2010). identifying, characterizing and quantifying the protein
While streamlining the process and rebalancing the effort content of body fluids and tissues. So, this chapter deals
and expenditures across the various phases of drug discov- primarily with information about genes and its role in
ery and development will likely increase productivity to drug discovery and development.
some extent, the fact remains that what we need to know to
optimize productivity in drug discovery and development far
exceeds what we do know right now. To improve productivity INNOVATION DEPENDS  
substantially, the pharmaceutical industry must increas-
ON INFORMATION FROM  
ingly become what in a looser sense it always was, an
information industry (Robson and Baek, 2009). MULTIPLE SOURCES
Initially, the present authors leaned towards extending
this idea by highlighting the pharmaceutical process as Information from three diverse domains sparks innovation
an information flow, a flow seen as a probabilistic and in drug discovery and development (http://dschool.

© 2012 Elsevier Ltd. 77

Section | 2 | Drug Discovery

utility of available, but diverse, information across all

aspects of drug discovery and development is to connect
apparently disparate information sets using a common
format and a collection of rules (a language) that relate
Genbank elements of the content to each other. Such connections
make it possible for users to access the information in
any domain or set then traverse across information in
multiple sets or domains in a fashion that sparks innova-
tion. This promising approach is embodied in the Seman-
EMBL
DDBJ tic Web, a vision for the connection of not just web pages
database
but of data and information (see http://en.wikipedia.org/
wiki/Semantic_Web and http://www.w3.org/2001/sw/),
which is already beginning to impact the life sciences,
generally, and drug discovery and development, in particu-
Unigene Ensembl
lar (Neumann and Quan, 2006; Stephens et al., 2006;
Ruttenberg et al., 2009).
Stack trEMBL

TIGR HGI
BIOINFORMATICS
Swissprot
The term bioinformatics for the creation, analysis and man-
agement of information about living organisms, and par-
Fig. 7.1  Flow of public sequence data between major ticularly nucleotide and protein sequences, was probably
sequence repositories. Shown in blue are the components of first coined in 1984, in the announcement of a funding
the International Nucleotide Sequence Database programme by the European Economic Community (EC
Collaboration (INSDC) comprising Genbank (USA), the COM84 Final). The programme was in response to a
European Molecular Biology Laboratory (EMBL) Database
memo from the White House that Europe was falling
(Europe) and the DNA Data Bank of Japan (DDBJ).
behind America and Japan in biotechnology.
The overall task of bioinformatics as it was defined in
that announcement is to generate information from bio-
stanford.edu/big_picture/multidisciplinary_approach. logical data (bioinformation) and to make that information
php): accessible to humans and machines that are in need of
• Feasibility (Is the product technically/scientifically information to advance toward the achievement of an
possible?) objective. Handling all the data and information about life
• Viability (Can such a product be brought cost- forms, even the currently available information on nucle-
effectively to the marketplace?) otide and protein sequences, is not trivial. Overviews of
• Desirability (Is there a real need for such a established basic procedures and databases in bioinfor-
product?). matics, and the ability to try one’s hand at them, are
provided by a number of high-quality sources many of
The viability and desirability domains include informa-
which have links to other sources, for example:
tion concerning public health and medical need, physi-
cian judgment, patient viewpoints, market research, • The European Bioinformatics Institute (EMBL-EBI),
healthcare strategies, healthcare economics and intellec- which is part of the European Molecular Biology
tual property. The feasibility domain includes information Laboratory (EMBL) (http://www.ebi.ac.uk/2can/
about biology, chemistry, physics, statistics and mathemat- home.html)
ics. Currently, the totality of available information can be • The National Center for Biotechnology Information
obtained from diverse, unconnected (’siloed’) public or of the National Institutes of Health (http://
private sources, such as the brains of human beings, www.ncbi.nlm.nih.gov/Tools/)
books, journals, patents, general literature, databases • The Biology Workbench at the University of San
available via the Internet (see, for example, Fig. 7.1) and Diego (http://workbench.sdsc.edu/).
results (proprietary or otherwise) of studies undertaken Rather than give a detailed account of specific tools for
as part of drug discovery and development. However, bioinformatics, our focus will be on the general ways by
due to its siloed nature, discreet sets of information which bioinformation is generated by bioinformatics, and
from only one domain are often applied, suboptimally, the principles used to analyse and interpret information.
in isolation to particular aspects of drug discovery and Bioinformatics is the management and data analytics
development. One promising approach for optimizing the of bioinformation. In genetics and molecular biology

78
 The role of information, bioinformatics and genomics Chapter |7|

applications, that includes everything from classical statis- interest to an error of e percent (and e is usually much less
tics and specialized applications of statistics or probability than 50%) then this escalates to (100/e)N. Clearly, there-
theory, such as the Hardy–Weinberg equilibrium law and fore, data mining is a strategy, not a guaranteed solution,
linkage analysis, to modelling interactions between the but, equally clearly, delivers a lot more than one query.
products of gene expression and subsequent biological Insomuch as issuing one query is simply a limiting case
interpretation (for interpretation tool examples see of the ultimate in highly constrained data mining, both
www.ingenuity.com and www.genego.com). According to modes can be referred to as data mining.
taste, one may either say that a new part of data analytics Both querying and data mining seem a far cry from
is bioinformatics, or that bioinformatics embraces most if predicting what regions of DNA are likely to be a gene, or
not all of data analytics as applied to genes and proteins the role a pattern of gene variants might play in disease or
and the consequences of them. Here, we embrace much drug response, or the structure of a protein, and so on. As
as bioinformatics, but also refer to the more general field it happens, however, prediction of what regions of DNA
of data analytics for techniques on which bioinformatics are genes or control points, what a gene’s function is, of
draws and will continue to borrow. protein secondary and tertiary structure, of segments in
The term bioinformatics does have the merit of address- protein primary structure that will serve as a basis for a
ing not only the analysis but also the management of synthetic diagnostic or vaccine, are all longstanding exam-
data using information technology. Interestingly, it is not ples of what we might today call data mining followed by
always easy to distinguish data management from the its application to prediction. In those activities, the mining
processing and analysis of data. This is not least because of data, as a training set, is used to generate probabilistic
computers often handle both. Sometimes, efficiency and parameters often called ‘rules’. These rules are preferably
insight can be gained by not segregating the two aspects validated in an independent test set. The further step
of bioinformatics. required is some process for using the validated rules to
draw a conclusion based on new data or data sets, i.e.
Bioinformatics as data mining   formally the process of inference, as a prediction.
An Expert System also uses rules and inference except the
and inference rules and their probabilities are drawn from human
Data mining includes also analysis of market, business, experts at the rate of 2–5 a day (and, by definition, are
communications, medical, meteorological, ecological, essentially anecdotal and likely biased). Computer-based
astronomical, military and security data, but its tools have data mining can generate hundreds of thousands of
been implicit and ubiquitous in bioinformatics from unbiased probabilistic rules in the order of minutes to
the outset, even if the term ‘data mining’ has only fairly hours (which is essentially in the spirit of evidence based
recently been used in that context. All the principles medicine’s best evidence). In the early days of bioinfor­
described below are relevant to a major portion of tradi- matics, pursuits like predicting protein sequence, signal
tional bioinformatics. The same data mining programme polypeptide sequences, immunological epitopes, DNA
used by one of the authors has been applied to both joint consensus sequences with special meaning, and so forth,
market trends in South America and the relationship of were often basically like Expert Systems using rules and
protein sequences to their secondary structure and immu- recipes devised by experts in the field. Most of those pur-
nological properties. In the broader language of data ana- suits have now succumbed to use rules provided from
lytics, bioinformatics can be seen as having two major computer-based mining of increasingly larger amounts of
modes of application in the way it obtains information data, and those rules bear little resemblance to the original
from data – query or data mining. In the query mode expert rules.
(directed analysis), for example, a nucleotide sequence
might be used to find its occurrence in other gene
General principles for data mining
sequences, thus ‘pulling them from the file’. That is similar
to a Google search and also somewhat analogous to Data mining is usually undertaken on a sample dataset.
testing a hypothesis in classical statistics, in that one spe- Several difficulties have dogged the field. At one end of the
cific question is asked and tested as to the hypothesis that spectrum is the counterintuitive concern of ‘too much
it exists in the data. (relevant) information’. Ideally, to make use of maximum
In the data mining mode (undirected or unsupervised data available for testing a method and quality of the rules,
analysis), one is seeking to discover anything interesting it quickly became clear that one should use the jackknife
in the data, such as a hidden pattern. Ultimately, finding method. For example, in predicting something about each
likely (and effectively testing) hypotheses for combina- and every accessible gene or protein in order to test a
tions of N symbols or factors (states, events, measure- method, that gene or protein is removed from the sample
ments, etc.) is equivalent to making 2N-1 queries or data set used to generate the rules for its prediction. So
classical statistical tests of hypotheses. For N = 100, this is for a comprehensive test of predictive power, the rules
1030. If such an activity involves continuous variables of are regenerated afresh for every gene or protein in the

79
Section | 2 | Drug Discovery

database, or more correctly put, for the absence of each in perhaps, lethal) combination. Yet, traditionally, unicorn
the database. The reason is that probabilistic rules are events are not even seen to justify consideration, and in
really terms in a probabilistic or information theoretic computing terms they are not even ‘existentially qualified’,
expansion that, when brought together correctly, can no variables may even be created to consider them. If we
predict something with close to 100% accuracy, if the gene do force creation of them, then the number of things to
or protein was in the set used to derive the rules. That has allow for because they just might be, in principle, can be
practical applications, but for most purposes would be astronomical.
‘cheating’ and certainly misleading. Once the accuracy is Data mining algorithms can yield information about:
established as acceptable, the rules are generated from all • the presence of subgroups of samples within a
genes or proteins available, because they will typically be sample dataset that are similar on the basis of
applied to new genes or proteins that emerge and which patterns of characteristics in the variables for each
were not present in the data. On the other hand, once sample (clustering samples into classes)
these become ‘old news’, they are added to the source data, • the variables that can be used (with weightings
and at intervals the rules are updated from it. reflecting the importance of each variable) to classify
At the other end of the scale are the concerns of too little a new sample into a specified class (classification)
relevant information. For example, data may be too sparse • mathematical or logical functions that model the
for rules with many parameters or factors (the so-called data (for example, regression analysis) or
‘curse of high dimensionality’), and this includes the case • relationships between variables that can be used to
where no observations are available at all. Insight or pre- explore the behaviour of the system of variables as a
dictions may then be incorrect because many relevant whole and to reveal which variables provide either
rules may need to come together to make a final pro- novel (unique) or redundant information (association
nouncement. It is rare that a single probabilistic rule will or correlation).
say all that needs to be said. Perhaps the greatest current
Some general principles for data mining in medical
concern, related to the above, is that the information
applications are exemplified in the mining of 667 000
obtained will only be of general utility if the sample
patient records in Virginia by Mullins et al. (2006). In that
dataset is a sufficient representation of the entire population.
study, which did not include any genomic data, three main
This is a key consideration in any data mining activity and
types of data mining were used (pattern discovery, predic-
likely underlies many disputes in the literature and else-
tive analysis and association analysis). A brief description
where about the validity of the results of various studies
of each method follows.
involving data mining. To generate useful information
from any such study, it is essential to pay particular atten-
tion to the design of the study and to replicate and validate
the results using other sample datasets from the entire
Pattern discovery
population. The term data dredging is often used in refer- Pattern discovery is a data mining technique which seeks to
ence to preliminary data mining activities on sample data- find a pattern, not necessarily an exact match, that occurs
sets that are too small to generate results of sufficient in a dataset more than a specified number of times k. It is
statistical power to likely be valid but can generate hypoth- itself a ‘number of times’, say n(A & B & C &…). Pattern
eses for exploration using large sample datasets. discovery is not pattern recognition, which is the use of these
A further concern is that sparse events in data can be results. In the pure form of the method, there is no nor-
particularly important precisely because they are sparse. malization to a probability or to an association measure
What matters is not the support for a rule in terms of the as for the other two data mining types discussed below.
amount and quality of data concerning it, but (in many Pattern discovery is excellent at picking up complex pat-
approaches at least) whether the event occurred with an terns with multiple factors A, B, C, etc., that tax the next
abundance much more, or much less, than expected, say two methods. The A, B, C, etc. may be nucleotides (A, T,
on a chance basis. Negative associations are of great G, C) or amino acid residues (in IUPAC one letter code),
importance in medicine when we want to prevent some- but not necessarily contiguous, or any other variables. In
thing, and we want a negative relationship between a practice, due to the problems typically addressed and the
therapy and disease. The so-called unicorn events about nature of natural gene sequences, pattern discovery for
observations never seen at all are hard to handle. A simple nucleotides tends to focus on runs of, say, 5–20 symbols.
pedagogic example is the absence of pregnant males in a Moreover, the patterns found can be quantified in the
medical database. Whilst this particular example may only same terms as either of the other two methods below.
be of interest to a Martian, most complex rules that are However, if one takes this route alone, important relation-
not deducible from simpler ones (that are a kind of subset ships are missed. It is easy to see that this technique cannot
to them) might be of this type. Of particular concern is pick up a pattern occurring less than k times, and k = 1
that drugs A, B, and C might work 100% used alone, and makes no sense (everything is a pattern). It must also miss
so might AB, BC, and AC, but ABC might be a useless (or, alerting to patterns that statistically ought to occur but do

80
 The role of information, bioinformatics and genomics Chapter |7|

not, i.e. the so-called unicorn events. Pattern discovery is I( A : ∼ A ; B) = I( A ; B) − I(∼ A ; B) (5)
an excellent example of an application that does have
where ~A is a negative of complementary state or event
an analogue at operating system level, the use of the
such that
regular expression for partial pattern matching (http://
en.wikipedia.org/wiki/Regular_expression), but it is most P(∼ A) = 1 − P( A) (6)
efficiently done by an algorithm in a domain-specific is familiar as log predictive odds, while
application (the domain in this case is fairly broad,
however, since it has many other applications, not just in I( A : ∼ A ; B) − I( A : ∼ A ; ∼ B) (7)
bioinformatics, proteomics, and biology generally). An is the log odds ratio.
efficient pattern discovery algorithm developed by IBM is The association analysis approach handles positive,
available as Teireisis (http://www.ncbi.nlm.nih.gov/pmc/ zero, and negative associations including treatment sparse
articles/PMC169027/). joint events. To that end, it may use the more general defi-
nition of information in terms of zeta functions, ζ. Unlike
predictive analysis, the approach used in this way returns
Predictive analysis
expected information, basically building into the final value
Predictive analysis encompasses a variety of techniques the idea of support. In the Virginia study, using the above
from statistics and game theory that analyse current and ‘zeta approach’, one could detect patterns of 2–7 symbols
historical facts to predict future events (http://en. or factors, the limit being the sparseness of data for many
wikipedia.org/wiki/Predictive_analytics). ‘Predictive anal- such. As data increase, I(males; tall) = ζ(1, observed[males,
ysis’ appears to be the term most frequently used for the tall]) − ζ(1, expected[males, tall]) will rapidly approach
part of data mining that involves normalizing n(A & B & loge (observed[males, tall]) − loge (expected[males, tall]),
C) to conditional probabilities, e.g. but unlike log ratios, ζ(1, observed[males, pregnant]) −
P( A | B & C) = n( A & B & C) / n(B & C) (1) ζ(1, expected[males, pregnant]) works appropriately with
In usual practice, predictive analysis tends to focus on the data for the terms that are very small or zero. To handle
2–3 symbols or factors. This approach is an excellent form unicorn events still requires variables to be created in the
for inference of the type that uses ‘If B & C then A’, but takes programme, but the overall approach is more natural.
no account of the fact that A could occur by chance anyway. The above seem to miss out various forms of data mining
In practice, predictive analysis usually rests heavily on the such as time series analysis and clustering analysis, although
commonest idea of support, i.e. if the pattern is not seen a ultimately these can be expressed in the above terms. What
significant number of times, the conditional probability of seems to require an additional comment is correlation.
it is not a good estimate. So again, it will not detect impor- While biostatistics courses often use association and correla-
tant negative associations and unicorn events. Recently, tion synonymously, data miners do not. Association relates
this approach has probably been the most popular kind of to the extent to the number of times things are observed
data mining for the general business domain. together more, or less, than on a chance basis in a ‘presence
or absence’ fashion (such as the association between a
categorical SNP genotype and a categorical phenotype).
Association analysis This is reminiscent of the classical chi square test, but
Association analysis is often formulated in log form as revealing the individual contributions to non-randomness
Fano’s mutual information, e.g. within the data grid (as well as the positive or negative
nature of the association). In contrast, correlation relates
I( A ; B & C) = log e{P( A & B & C)/[P( A)P(B & C)]} (2)
to trends in values of potentially continuous variables
(Robson, 2003, 2004, 2005, 2008; Robson and Mushlin, (independence between the variances), classically exempli-
2004; Robson and Vaithiligam, 2010). Clearly it does take fied by use of Pearson’s correlation. Correlation is impor-
account of the occurrence of A. This opens up the full tant in gene expression analysis, in proteomics and in
power of information theory, of instinctive interest to the metabolomics, since a gene transcript (mRNA) or a protein
pharmaceutical industry as an information industry, and or a small molecule metabolite, in general, has a level of
of course to other mutual information measures such as: abundance in any sample rather than a quantized presence/
I( A ; B | C) = log e{P( A & B & C)/[P( A & C)P(B & C)]} (3) absence. Despite the apparent differences, however, the
implied comparison of covariance with what is expected
and
on independent, i.e. chance, basis is essentially the same
I( A ; B; C) = log e{P( A & B & C)/[P( A)P(B)P(C)]} (4) general idea as for association. Hence results can be
The last atomic form is of interest since other measures expressed in mutual information format, based on a kind
can be calculated from it (see below). Clearly it can also of fuzzy logic reasoning (Robson and Mushlin, 2004).
be calculated from the conditional probabilities of predic- Much of the above may not seem like bioinformatics,
tive analysis. To show relationship with other fields such but only because the jargon is different. That this ‘barrier’
as evidence based medicine and epidemiology, is progressively coming down is important, as each

81
Section | 2 | Drug Discovery

discipline has valuable techniques less well known in the should indeed be considered in their own right but may
other. Where they do seem to be bioinformatics it is essen- also be seen as subsets or derivatives of the genome
tially due to the fact that they come packaged in distinct concept. The remainder of this chapter is largely devoted
suites of applications targeted at bioinformatics users, and to genomic information and its use in drug discovery and
where they do not seem to be bioinformatics, they do not development.
come simply packaged for bioinformatics users. While the term genome has recently spawned many
offspring ‘-omes’ relating to the disciplines that address
various matters downstream from inherited information
in DNA, e.g. the proteome, these popular -ome words have
GENOMICS an even earlier origin in the 20th century (e.g. biome and
rhizome). Adding the plural ‘-ics’ suffix seems recent. The
The genome and its   use of ‘omics’ as a suffix is more like an analogue of the
earlier ‘-netics’ and ‘-onics’ in engineering. The current
offspring ‘-omes’
rising hierarchy of ‘-omes’ is shown in Table 7.1, and these
In contrast to bioinformatics, the term genome is much and others are discussed by Robson and Baek (2009).
older, first believed to be used in 1920 by Professor Hans There are constant additions to the ‘-omes’.
Winkler at the University of Hamburg, as describing the The new ‘-omes’ are not as simple as a genome. For one
world or system within the discipline of biology and thing, there is typically no convenient common file format
within each cell of an organism that addresses the inher- or coding scheme for them comparable with the linear
ited executable information. The word genome (Gk: DNA and protein sequence format provided by nature.
γ  νοµαι) means I become, I am born, to come into being, and Whereas the genome contains a static compilation of the
the Oxford English Dictionary gives its aetiology as being sequence of the four DNA bases, the subject matters of the
from gene and chromosome. This aetiology may not be other ‘-omes’ are dynamic and much more complex,
entirely correct. including the interactions with each other and the sur-
In this chapter, genomes of organisms are in computer rounding ecosystem. It is these shifting, adapting pools of
science jargon the ‘primary objects’ on which bioinformat- cellular components that determine health and pathology.
ics ‘acts’. Their daughter molecular objects, such as the Each of them as a discipline is currently a ‘mini’ (or,
corresponding transcriptomes, proteomes and metabolomes, perhaps more correctly, ‘maxi’) genome project, albeit

Table 7.1  Gene to function is paved with ‘-omes’

Commonly used terms

Genome Full complement of genetic information (i.e. DNA sequence, including coding and Static
non-coding regions)
Transcriptome Population of mRNA molecules in a cell under defined conditions at a given time Dynamic
Proteome Either: the complement of proteins (including post-translational modifications) Static
encoded by the genome
or: the set of proteins and their post-translational modifications expressed in a cell Dynamic
or tissue under defined conditions at a specific time (also sometimes referred to
as the translatome)

Terms occasionally encountered (to be interpreted with caution)

Secretome Population of secreted proteins produced by a cell Dynamic
Metabolome Small molecule content of a cell Dynamic
Interactome Grouping of interactions between proteins in a cell Dynamic
Glycome Population of carbohydrate molecules in a cell Dynamic
Foldome Population of gene products classified by tertiary structure Dynamic
Phenome Population of observable phenotypes describing variations of form and function in Dynamic
a given species

82
 The role of information, bioinformatics and genomics Chapter |7|

with ground rules that are much less well defined. Some concerning the behaviour of an organism, understanding
people have worried that ‘-omics’ will prove to be a spiral which ones are altered in disease or play a role in drug
of knowing less and less about more and more. Systems responses is most relevant to drug discovery and develop-
biology seeks to integrate the ‘-omics’ and simulate the ment. Furthermore, while detailed mechanisms may never
detailed molecular and macroscopic processes under the be fully understood, physicians can certainly make use of
constraint of as much real-world data as possible (van der validated relationships between genome features and the
Greef and McBurney, 2005; van der Greef et al., 2007). efficacy or safety outcomes of various treatments to tailor
For the purposes of this chapter, we follow the United treatment strategies for individual patients, often referred
States Food and Drug Administration (FDA) and Euro- to as personalized medicine.
pean Medicines Evaluation Agency (EMEA) agreed defini-
tion of ‘genomics’ as captured in the definition of a
A few genome details
‘genomic biomarker’, which is defined as: A measurable
DNA and/or RNA characteristic that is an indicator of normal A great deal of technology had to be invented to accom-
biologic processes, pathogenic processes, and/or response to plish the sequencing of genomes (Cantor and Smith,
therapeutic or other interventions (see http://www.fda.gov/ 1999). Recent years have seen dramatic progress. By the
downloads/Drugs/GuidanceComplianceRegulatory middle of 2010, the sequences of the genomes of more
Information/Guidances/ucm073162.pdf). A genomic bio­ than 2400 organisms were either complete (770 prokaryo-
marker can consist of one or more deoxyribonucleic acid tes; 37 eukaryotes), in draft assembly (568 prokaryotes;
(DNA) and/or ribonucleic acid (RNA) characteristics. 240 eukaryotes) or in progress (551 prokaryotes; 266
DNA characteristics include, but are not limited to: eukaryotes) (http://www.ncbi.nlm.nih.gov/sites/genome).
• Single nucleotide polymorphisms (SNPs); Most sequenced genomes are for bacteria, but the eukary-
• Variability of short sequence repeats ote genomes sequenced include Homo sapiens (Lander
• Haplotypes et al., 2001; Venter et al., 2001) and a wide variety of
• DNA modifications animals that play a role in the discovery of disease mecha-
• Deletions or insertions of (a) single nucleotide(s) nisms or in drug discovery and development, such as the
• Copy number variations fruit fly (Drosophila simulans (melanogaster)), flatworm
• Cytogenetic rearrangements, e.g., translocations, (Caenorhabditis elegans), guinea pig (Cavia porcellus), mouse
duplications, deletions, or inversions. (Mus musculus), rat (Rattus norvegicus), dog (Canis lupus)
and monkey (Macaca mulatta). The massive amount of
RNA characteristics include, but are not limited to:
information generated in the various genome projects and
• RNA sequences gene mapping studies has to be stored, disseminated and
• RNA expression levels analysed electronically, relying heavily on bioinformatics
• RNA processing, e.g. splicing and editing software systems and the Internet (see, for example, Fig.
• MicroRNA levels. 7.2). Here we may note that defining the nucleotide
The definition of a genomic biomarker does not include sequence is only the starting point of the genomic approach.
the measurement and characterization of proteins or While the percentage of human DNA spanned by genes
small molecule metabolites and, therefore, to limit the is 25.5–37.8%, only about 2% of the human genome is
scope of this chapter, the roles of proteomics and metabo- nucleotide sequence that actually codes for protein, or
lomics in drug discovery and development will not be indeed for structural or catalytic RNA. The other 98%,
discussed. initially considered to be ‘junk’ DNA, may indeed include
It remains that the genome currently rules in the sense evolution’s junk, i.e. fossil relics of evolution or, more
of being the basic instruction set from which various kindly put, non-functional genes that might serve as the
subsets of gene products are derived (Greenbaum et al., basis of evolution in the longer term. However, informa-
2001), and hence largely responsible for the manifestation tion about this ‘junk’ DNA is not uninteresting or useless.
of all the ‘-omes’, albeit the details of that manifestation Even truly non-functional DNA may have significant
are contingent upon the environment of the cell and medical and forensic value. Significant human individual
organism (see below). The coded information is a strength differences represented by genomic biomarkers are an
in terms of simplicity, and as an essentially invariant exploding opportunity for improving healthcare and trans-
feature of a patient, while all else on the lifelong ‘health forming the activities of the pharmaceutical industry
record’ may change. And whereas the downstream ‘-omes’ (Svinte et al., 2007), but these genomic biomarkers fre-
seen can have a huge role in interpreting the genome, quently are found not to lie in protein coding regions. They
inspection of features in the genome alone will often have often travelled with the genes about which they carry
inform what is not possible, or likely to be a malfunction. information, in the course of the migrations of human
Although it will ultimately be important to know all of prehistory and history. The so-called ‘junk’, including the
the possible gene products that a species is capable of introns discussed below and control segments, is involved
producing and to understand all the mechanistic details in chromatin structure, recombination templating, and is

83
Fig. 7.2  Screenshot from the Ensembl website. Shown is a chromosomal overview and a regional view of chromosome 21
surrounding the gene APP, the Alzheimer’s disease amyloid A4 protein precursor. In the detailed view the precise exon–intron
structure of the gene is shown, including sequence homology matches to a variety of databases or de novo gene predictions.
The viewer is highly customizable and users can easily navigate along the genomic axis.
Reproduced with kind permission of European Bioinformatics Institute and Wellcome Trust Sanger Institute.

84
 The role of information, bioinformatics and genomics Chapter |7|

now seen as representing hotspots for sponsoring hetero- changes in gene function … that do not entail a change
geneity (see comment on epigenetics below). in DNA sequence’ (Wu and Morris, 2001). The word was
There was, for a time, a surprising reduction in the coined by Conrad Waddington in 1939 (Waddington,
promised amount of what constitutes really interesting 1939) in recognition of relationship between genetics and
genomic information in humans. Although identifying developmental biology and of the sum of all mechanisms
genes within the genome sequence is not straightforward, necessary for the unfolding of the programme for develop-
earlier estimates of about 100 000 genes in the human ment of an organism.
genome based on the variety of proteins produced came The modifications of chromatin structure that can influ-
down, rather startlingly, to about 35 000 when the genome ence gene expression comprise histone variants, post-
was determined, and later (International Human Genome translational modifications of amino acids on the
Sequencing Consortium, 2004) to about 25 000. A clue as amino-terminal tail of histones, and covalent modifica-
to how this can be so was already well known in the late tions of DNA bases – most notably methylation of cyto-
1970s and 1980s – that genes in complex organisms are sine bases at CpG sequences. CpG-rich regions are not
suspiciously much larger than would be expected from the evenly distributed in the genome, but are concentrated in
number of amino acid residues in the proteins for which the promoter regions and first exons of certain genes
they code. It was found that only the regions of a gene (Larsen et al., 1992). If DNA is methylated, the methyl
called exons actually code for proteins, and the interven- groups protrude from the cytosine nucleotides into the
ing introns are removed by RNA splicing at the messenger major groove of the DNA and inhibit the binding of tran-
RNA (mRNA) level. The average human gene has some scription factors that promote gene expression (Hark
30 000 nucleotides and 7 exons. The largest known human et al., 2000). Changes in DNA methylation can explain the
gene, for titin, has 363 introns. fact that genes switch on or off during development and
the pattern of methylation is heritable (first proposed by
Holliday and Pugh, 1975). It has been shown that DNA
Genome variability and   methylation can be influenced by external stressors, envi-
individual differences ronmental toxins, and aging (Dolinoy et al., 2007; Hanson
Personalized medicine in the clinical setting is an emerging and Gluckman, 2008), providing a key link between the
field. Human individuality underpins susceptibility to environment and life experience and the regulation of
disease and response to treatment. Fundamental to human gene expression in specific cell types.
individuality is the content of each individual’s genome. A key finding in epigenetics was reported by Fraga et al.
All humans have 99.5–99.9% of nucleotide sequence in 2005. In a large cohort of monozygotic twins, DNA
identity in their genomes, depending upon whether one methylation was found to increase over time within dif-
takes into account just SNPs or all forms of genetic varia- ferent tissues and cell types. Importantly, monozygotic
tion (see http://www.genome.gov/19016904 and also twins were found to be indistinguishable in terms of DNA
Freeman et al., 2006). However, the 0.1% difference in methylation early in life but were found to exhibit sub-
SNPs represents 6 million bases in a genome of 6 billion stantial differences in DNA methylation with advancing
bases – plenty of scope for individuality. Even mono­ age. The authors of the landmark paper stated that ‘distinct
zygotic twins can develop individuality at the genome profiles of DNA methylation and histone acetylation pat-
(epigenome) level in individual body cells or tissues terns that among different tissues arise during the lifetime
through ‘life experience’. Understanding human individu- of monozygotic twins may contribute to the explanation
ality at the genomic level has the potential to enable great of some of their phenotypic discordances and underlie
advances in drug discovery and development. Pharmaco­ their differential frequency/onset of common disease.’
genomics addresses the matter of whether the same drug
will cure, do little, or harm according to the unique The transcriptome
genomics of the patient.
The fact that the cell must cut out the introns from mRNA
and splice together the ends of the remaining exons is the
The epigenome ‘loophole’ that allows somatic diversity of proteins,
In addition to the linear sequence code of DNA that con- because the RNA coding for exons can be spliced back
tains the information necessary to generate RNA and, ulti- together in different ways. Alternative splicing might there-
mately, the amino acid sequences of proteins, there is an fore reasonably be called recombinant splicing, though
additional level of information in the DNA structure in that term offers some confusion with the term ‘recom-
the form of the complex nucleoprotein entity chromatin. binant’ as used in classical genetics. That some 100 000
Genetic information is encoded not only by the linear genes were expected on the basis of protein diversity and
sequence of DNA nucleotides but by modifications of about 25 000 are found does not mean that the diversity
chromatin structure which influence gene expression. Epi- so generated is simply a matter of about four proteins
genetics as a field of study is defined as ‘… the study of made from each gene on average. Recent work has shown

85
Section | 2 | Drug Discovery

intergenic length is about 3 million nucleotides. While 3%

Control Experimental of DNA is short repeats and 5% is duplicated large pieces,
tissue tissue
45% of intergenic DNA comprises four classes of ‘parasitic
elements’ arising from reverse transcription. In fact it may
Prepare cDNA
be that this DNA is ‘a boiling foam’ of RNA transcription
and reverse transcription. It now appears that both repeats
Prepare microarray and boiling foam alike appear to harbour many control
features.
At first, the fact that organisms such as most bacteria
RT-PCR have 75–85% of coding sequences, and that even the
Green Label with Red puffer fish has little intergenic DNA and few repeats, may
dye fluorescent dye bring into question the idea that the intergenic DNA and
dyes its transcriptome can be important. However, there is evi-
dently a significant difference in sophistication between
humans and such organisms. This thinking has now
altered somewhat the original use of the term genome. The
human genome now refers to all nuclear DNA (and ideally
mitochondrial DNA as well). That is, the word has been
increasingly interpreted as ‘the full complement of genetic
Combine
information’ coding for proteins or not. The implication
equal amounts
is that ‘genome’ is all DNA, including that which does not
Hybridize probe appear to carry biological information, on the basis of the
to microarray growing suspicion that it does. As it happens, the term
Scan ‘coding’ even traditionally sometimes included genes
that code for tRNA and rRNA to provide the RNA-based
Fig. 7.3  Microarray experiment: technology for analysing protein-synthesizing machinery, since these are long
mRNA expression. known. The key difference may therefore mainly lie in
departure from the older idea that RNA serves universal
housekeeping features of the cell as ‘a given’ that do not
that alternative splicing can be associated with thousands reflect the variety of cell structure and function, and dif-
of important signals. ferentiation. That, in contrast, much RNA serves contingency
The above is one aspect of the transcriptome, i.e. the functions relating to the embryonic development of
world of RNA produced from DNA. Whereas one may organisms is discussed below.
defer discussion of other ‘omes’ (e.g. of proteins, except as There are still many unidentified functionalities of RNA
the ‘ultimate’ purpose of many genes), the above exempli- coding regions in the ‘junk DNA’, but one that is very
fies that one makes little further progress at this stage by topical and beginning to be understood is microRNA (or
ignoring the transcriptome. A primary technology for meas- miRNA). miRNAs are short lengths of RNA, just some
uring differences in mRNA levels between tissue samples 20–25 nucleotides, which may base pair with RNA and in
is the microarray, illustrated in Figs. 7.3 & 7.4. some cases DNA – possibly to stop it binding and compet-
The RNA world now intermediates between the genome ing with its own production, but certainly to enable
and the proteome, including carrying transcripts of DNA control process that hold the miRNAs in check.
in order to determine the amino acid sequences of pro- The maturation of miRNAs is quite complicated. Two
teins, but is believed to be the more ancient genomic complementary regions in the immature and much larger
world before DNA developed as a sophisticated central miRNA associate to form a hairpin, and a nuclear enzyme
archive. Not least, since unlike a protein any RNA is a Drosha cleaves at the base of the hairpin. The enzyme Dicer
direct copy of regions of DNA (except for U replacing T), then extracts the mature miRNA from its larger parent. A
much of an account of the transcriptome looks like and typical known function of a mature miRNA is as an anti-
involves an account of the genome, and vice versa. sense inhibitor, i.e. binding messenger mRNA, and then
To be certain, the transcriptome does go beyond that. initiating its degradation, so that no protein molecule can
The RNA produced by DNA can perform many complex be made from the mRNA. Note that in the above, one
roles, of which mRNA and its alternative splicing provide could think largely of recognition regions in the proto-
just one example. Many of the sites for producing these miRNA transcript as simply mapping to patterns of nucle-
RNAs lie not in introns but outside the span of a gene, i.e. otides in DNA. But it would be impossible to meaningfully
in the intergenic regions. Like elucidation of the exon/ discuss, or at least not be insightful, without consideration
intron structure within genes, intergenic regions are simi- of the biochemical processes downstream of the first tran-
larly yielding to more detailed analysis. The longest known scription step.

86
 The role of information, bioinformatics and genomics Chapter |7|

Fig. 7.4  Cluster analysis of gene expression experiment.

Expression levels of 54 genes (listed on right) in eight control
and eight multiple sclerosis brain samples (listed above).
Relative expression levels are indicated in shades of red for
high values and green for low values. A computer algorithm

MS-4
MS-3
MS-1
MS-8
MS-7
MS-5
MS-6
MS-2
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
was used to calculate similarity between the expression
IL–12Rb2 pattern of each gene for all subjects, and the expression
HSCalBR
IFN–b
pattern for each subject for all genes, and to display the
IFN–a results as a dendrogram. The top dendrogram shows that MS
IL–11 and control samples are clearly distinguishable, the intragroup
IL–6 distance being much greater than the intersample difference
IL–12p35
within each group. The left-hand dendrogram shows the
MBP
MAG grouping of genes into clusters that behave similarly. A group
MOG–b of genes with highly similar expression patterns is outlined
MOG–a in yellow.
TGF–b3
Adapted, with permission, from Baranzini et al., 2000.
IL–12p40
MIP–1a
IL–9
IL–1b
IL–1a
IL–2Rg
IL–15
C1r
LMP2
TNFRp75
TGF–b2
IL–3
PRL
IFN–aR
LMP7 E–2
IL–2Rb
RANTES
CCR1
CCR5
IL–18
Caspase–1
beta–2 micro
IL–6R
IL–12Rb1
CD4
DRA
TNFRp55
TNFR1
IL–7
IL–5
LMP7 E–1
CCR4
FcRI
IL–10
PFR–1
HLA–C
TGF–b1
IL–1R
HLA–E
CCR8
IL–13
IL–8

magnetic tapes, floppy disks, CDs, and Internet downloads

In defence of the genome
that we feed into our computers, indispensible and
The complexities contained in the discussion above chal- primary, but cold and lifeless without all the rest. While
lenge the authority of the genome as ‘king’. At present, the genomic data now represent a huge amount of informa-
genome rules much because of our ignorance. DNA is tion that is relevant to the pharmaceutical industry, it
potentially no more, and no less, important than the cannot be denied that complete pharmaceutical utility lies

87
Section | 2 | Drug Discovery

in understanding how the information in DNA, including

Table 7.2  Steps in the drug discovery and
interaction between many genes and gene products, trans- development process
lates into function at the organism level. Not least, drugs
influence that translation process. The rise of systems
Process step
biology reflects this awareness of the organism as a
dynamic system dwelling in three dimensions of space and Understanding human disease intrinsic mechanisms
one of time. Understanding the biology of infectious agents
The problem is that the child ‘-omes’ are still immature Identifying potential drug targets
and rather weak. We are assembling a very rich lexicon Validating drug targets
from the genomes, but are left with the questions of when, Selecting targets for assay development
Assay development
how, and why, the patterns of life, molecular and macro-
Discovering hits
scopic, unfold. And one may say fold too, since the folding
Hits to leads
of proteins, the key step at which linear information in
Preclinical pharmacokinetics
DNA becomes three-dimensional biology, remains largely
Preclinical efficacy pharmacology
unsolved. Perhaps for all such reasons, the earlier pharma- Preclinical safety pharmacology
ceutical industry hopes for genomics and bioinformatics Preclinical toxicology
(Li and Robson, 2000) have not yet been realized (Drews, Phase 0 clinical studies for low dose agents (e.g.
2003). molecular imaging)
But low-hanging fruits have been picked to keep the Phase I clinical studies – pharmacokinetics
pharmaceutical industry buoyant. Biotechnology compa- Phase I clinical studies – safety
nies drink of biological knowledge much closer to the Phase II clinical studies – pharmacokinetics
genome than pharmaceutical companies usually do, and Phase II clinical studies – efficacy
they largely enabled the methodologies that made the Phase II clinical studies – safety
Human Genome Project a reality. Thus, they currently Phase III clinical studies – efficacy
position themselves as leaders in managing and tapping Phase III clinical studies – safety
this vast new information resource. Companies specializ- Phase IV clinical studies – efficacy
ing in genomic technology, bioinformatics, and related Phase IV clinical studies – safety
fields were the fastest expanding sector of the biotechnol- Drug repurposing – additional/alternate uses
ogy industry during the 1990s, aspiring to lead the phar- Drug rescue
maceutical industry into an era of rapid identification and
exploitation of novel targets (Chandra and Caldwell,
2003).
Not the least because the biotechnology industry has With regard to the relevance of genomic information
provided proof of concept, there is still consensus that the in the drug discovery and development process, a key
Human Genome Project will be worth to the pharmaceutical consideration is how best (cost-effectively) to use the
industry the billions of dollars and millions of man-hours information that underpins human individuality to
that nations have invested. But it still remains unclear as improve the process such that:
to exactly how all the available and future information can • Diseases are characterized, or defined, by molecular
be used cost-effectively to optimize drug discovery and means rather than by symptoms – a key advance
development. since pharmaceutical scientists make drug candidate
So just how prominent should genome-based strategies molecules which interact with other molecules not
be, right now, in the research portfolio of pharmaceutical with symptoms
companies, and what should those strategies be? The prin- • Diseases are detected by molecular means earlier
ciple of using genomic information as the starting point than currently possible, at a time when the
for identifying novel drug targets was described in Chapter underlying pathologies are reversible
6. Following is an overview of the ways in which genomic • Subgroups of patients with similar molecular
information can, or might, impact almost every aspect of characteristics can be identified within the larger
the drug discovery and development process. group of patients with similar symptoms, enabling
clinical trials to be focused on those subgroups of
Genomic information in drug patients most likely to benefit from a particular
pharmacological approach
discovery and development
• The best drug targets are selected for specific diseases
Drug discovery and development is mankind’s most com- or subgroups of patients
plicated engineering process, integrating a wide variety of • Individual human variation in drug exposure or
scientific and business activities, some of which are pre- drug response in terms of efficacy or side effects can
sented in Table 7.2. be anticipated or explained in the drug discovery

88
 The role of information, bioinformatics and genomics Chapter |7|

and development process, with the resulting by the genes associated with a specific disease are mapped
reduction in the frequency of clinical trials that onto known biochemical pathways (see, for example,
fail because of unanticipated ‘responder versus www.ingenuity.com or www.genego.com), it is sometimes
non-responder’ treatment outcomes and overall possible to gain an overall picture of likely disease mecha-
increase in productivity of drug discovery and nisms and to identify and prioritize potential drug targets
development. for the disease (see Chapter 6).
Following are comments on the impact that genomics
and genomic information can, or will, have on some of Validating drug targets
the steps in Table 7.2.
Transgenic animals, in which the gene coding for the drug
target protein can be knocked out or its expression down-
Understanding human disease regulated, can provide validation for the pharmacological
effect of an inhibitor drug prior to availability of any can-
intrinsic mechanisms
didate drug molecules (Sacca et al., 2010).
Many common human diseases are thought to result from
the interaction of genetic risk profiles with life style and
Assay development for selected
environmental risk factors. Over the past few decades,
large scale studies of healthy humans and human suffering drug targets
from diseases have associated specific gene variants with An analysis of variation in the gene coding for the drug
diseases. In many cases, these studies have revealed only target across a diverse human population can identify
relatively weak associations between many individual potential drug response variability in patients (Sadee
gene variants and diseases (Robinson, 2010). To date, this et al., 2001; Niu et al., 2010; Zhang et al., 2010). For
area of investigation has not returned the immediate example, at SNP locations, the relative frequencies of
benefit hoped for, and perhaps ‘hyped for’, by the genomic occurrence of each of the genotypes (for example, AA, AG
community. Nevertheless, insights have been gained into or GG) can provide information on likely differences in
intrinsic biochemical mechanisms underlying certain the abundance or structure of the drug target across the
diseases. human population. Different protein isoforms with even
Insight into the biochemical mechanisms of common slightly different structures could be associated with
diseases has definitely been gained from the study of altered drug binding or activation/inhibition of down-
inherited monogenic diseases in special families (Pelto- stream biochemical mechanisms. To avoid the discovery
nen et al., 2006). In these cases, the association between of drug candidates which interact effectively with only one
the gene variant and disease phenotype is very strong, isoform of the protein drug target or, alternatively, to focus
although the application of the information to common on discovering such drug candidates, multiple assays
diseases must be undertaken with caution. based on different protein isoforms expressed from vari-
ants of each drug target gene could be established. These
parallel assays could be used for screening chemical librar-
Understanding the biology of ies to select hits which were selective for one isoform of
infectious agents the drug target protein, for enabling personalized medicine,
The availability of complete genome sequences for many or were able to interact effectively with more than one
infectious agents has provided a great deal of insight into isoform, for enabling population-focused medicine.
their biology (Pucci, 2007). Furthermore, such genetic
information has also revealed the mutations in the genetic Phase 0 clinical studies – understanding from
material of these organisms that are responsible for drug compounds in low concentration
resistance (Kidgell and Winzeler, 2006) – of great interest This relatively recent term relates to trials which are safe
to those involved in the discovery of new drugs to treat because they need such low doses of a new chemical entity
major infections. There is a strong interplay with the being tested. Usually this applies to testing the very low
human genome, especially the HLA genes, which deter- concentrations of radioactively labelled compounds for
mine whether the T cell system will recognize certain fea- molecular imaging; concentrations still are sufficient to see
tures of proteins expressed by pathogens. how they travel, where they go, where they accumulate,
and how they are metabolized. A compound can be
labelled at several points to track its components and
Identifying potential drug targets derivatives formed in metabolism. It is inevitable that
Studies of the association of gene variants with the many differences in transport and metabolism will occur
phenotypes of common diseases have revealed, in some in individuals, so genomics will play a strong supporting
cases, hundreds of gene variants that can predispose role. It is likely that these kinds of trial will increasingly
humans to certain diseases. When the proteins coded for frequently be applied not just to develop a research agent,

89
Section | 2 | Drug Discovery

but to gather data for the compound to be used in the For example, drug treatment response could be altered in
following trials. patients with variants in individual genes encoding the
drug receptors or with variants in individual genes encod-
Phase I clinical studies – pharmacokinetics ing associated signalling proteins or with patterns of vari-
and safety ants in the genes encoding multiple proteins in relevant
biochemical signalling pathways.
Preclinical drug metabolism studies conducted on
At present there are relatively few clear examples of
drug candidates in line with the guidance provided by
the influence of genomic variation on clinical trial out­
the US FDA (see http://www.fda.gov/downloads/Drugs/
come (see http://www.fda.gov/Drugs/ScienceResearch/
GuidanceComplianceRegulatoryInformation/Guidances/
ResearchAreas/Pharmacogenetics/ucm083378.htm). Nev-
ucm072101.pdf) can provide a wealth of information on
ertheless, this area of investigation has a substantial place
the likely routes of metabolism of the drug, the enzymes
in drug discovery and development strategies in the future.
and transporters involved in the metabolism, and the drug
All clinical studies should now incorporate DNA collec-
metabolites generated. Enzymes of particular interest are
tion from every subject and patient. The costs of whole
the cytochromes P450 (CYPs) which introduce a polar
genome analysis are constantly diminishing and bioinfor-
functional group into the parent molecule in the first
matic methods for data mining to discover associations
phase of biotransformation of a drug.
between genomic variables and treatment–response vari-
An analysis of the genotypes of Phase I human subjects
ables are continually improving.
with respect to drug metabolizing enzymes can provide
A research effort undertaken in Phase I and Phase II
explanations for unusual pharmacokinetics and even
clinical studies to discover information about genetic vari-
adverse drug effects in certain subjects (Crettol et al.,
ants which influence the response to a drug candidate can
2010). In particular, subjects with a CYP2D6 gene variant
contribute substantially to the planning for pivotal Phase
associated with lower than normal enzyme expression or
III trials and the interpretation of the results of those trials,
activity (referred to as ‘poor metabolizers’) might have an
enhancing the overall prospects for successful completion
exaggerated response to a standard dose of the drug. In
of the development programme for a drug candidate.
those subjects, drug exposure might reach toxic levels and
More information on clinical trials can be found later in
trigger adverse effects. On the other hand, subjects with a
the book.
different CYP2D6 variant, associated with higher than
normal enzyme expression or activity (referred to as ‘rapid
metabolizers’ or ‘ultra rapid metabolizers’) might not Genomic information and drug
exhibit any drug effects whatsoever. Clearly, information regulatory authorities
concerning each individual’s genotype, with respect to
Successful drug discovery and development culminates in
drug metabolizing enzymes, in concert with information
the approval for marketing of a product containing the
about the routes of metabolism of a drug candidate, can
drug. Approval for marketing in a specific country or
be used to interpret the response to drug dosing in Phase
region is based on a review by the appropriate drug regula-
I studies but, more comprehensively, can potentially be
tory authority of a substantial amount of information
used to optimize the probability of success for the devel-
submitted to the regulatory authority by the sponsor (e.g.
opment of the drug and even the strategies for using an
pharmaceutical company) – on manufacturing, efficacy
approved drug in medical practice.
and safety of the product.
How has the growing availability of vast amounts of
Phase II and III clinical studies – efficacy genomic information affected the drug approval process?
and safety What changes are taking place at the regulatory authori-
In clinical studies to determine the efficacy and safety of ties? What new types of information do pharmaceutical
drug candidates in patients, it is important to recognize and biotechnology companies need to provide in the regu-
and take steps to deal with the likelihood that the different latory submissions associated with the drug discovery and
genotypes across the clinical trial population could influ- development process?
ence overall trial outcome. A not uncommon feature of Over the past 10 years, in response to the changing
clinical trials is that, with respect to treatment outcome landscape of genomics information and its potential to
(efficacy or safety), patients fall into two broad groups – improve the drug discovery and development process and
‘responders’ and ‘non-responders’. The balance between the cost-effectiveness of new medicines, there has been a
these two groups of patients can have a profound effect steep learning curve around the world within regulatory
on the overall results of a trial. While variants in the genes authorities and pharmaceutical/biotechnology compa-
encoding drug metabolizing enzymes might explain some nies. Fortunately, representatives of the regulatory authori-
such results, as mentioned above, other genetic variations ties and the companies have joined together to explore
across the population could also play a role in diverse and define the role of genomic information in drug dis-
responses to drug treatment in a clinical trial population. covery and development, and marketing approval.

90
 The role of information, bioinformatics and genomics Chapter |7|

Critical Path Initiative and the EMEA’s continue to shape the ways that pharmaceutical companies
equivalent programme incorporate, and regulators review, genomic information
in the drug discovery, development and approval process.
In the early 2000s, the US Food and Drug Administration The FDA and EMEA guidance documents and meeting
(FDA) and the European Medicines Evaluation Agency reports are presented in Table 7.3.
(EMEA) separately initiated efforts to improve the drug
discovery and development process. In part, these efforts
were stimulated by a trend in the preceding years of fewer
and fewer drug approvals each year. Additionally, there CONCLUSION
was an understanding at these regulatory agencies that
new developments in scientific and medical research, The biopharmaceutical industry depends upon vast
such as the explosion of genomic information, were not amounts of information from diverse sources to support
being incorporated into the drug evaluation processes. the discovery, development and marketing approval of its
The FDA kicked-off its initiative in March 2004 with an products. In recent years, the sustainability of the indus-
insightful white paper entitled ‘Innovation/Stagnation: try’s business model has been questioned because the fre-
Challenge and Opportunity on the Critical Path to quency of successful drug discovery and development
New Medical Products’ (http://www.fda.gov/downloads/ projects, in terms of approved products, is considered too
ScienceResearch/SpecialTopics/CriticalPathInitiative/ low in light of the associated costs – particularly in a social
CriticalPathOpportunitiesReports/ucm113411.pdf) and context where healthcare costs are spiralling upwards and
the initiative became known as the Critical Path Initiative. ‘out of control’.
Also in 2004, the EMEA set up the EMA/CHMP Think- The lack of successful outcome for the overwhelming
Tank Group on Innovative Drug Development (http:// majority of drug discovery and development projects
www.ema.europa.eu/ema/index.jsp?curl=pages/special_ results from the fact that we simply do not know enough
topics/general/general_content_000339.jsp&murl= about biology in general, and human biology in particu-
menus/special_topics/special_topics.jsp&mid=WC0b01ac lar, to avoid unanticipated ‘roadblocks’ or ‘minefields’ in
05800baed8&jsenabled=true). the path to drug approval. In short, we do not have enough
bioinformation.
Most drugs interact with protein targets, so it would be
Voluntary exploratory data submissions and
natural for pharmaceutical scientists to want comprehen-
guidance documents sive information about the nature and behaviour of pro-
In November 2003, the FDA published a draft guidance teins in health and disease. Unfortunately, because of the
document on Pharmacogenomic Data Submissions (final complexity of protein species, availability proteomic tech-
guidance published in March 2005, http://www.fda.gov/ nologies have not yet been able to reveal such comprehen-
downloads/RegulatoryInformation/Guidances/ucm sive information about proteins – although substantial
126957.pdf) and introduced a programme to encourage advances have been made over the past few decades.
pharmaceutical/biotechnology companies to make Volun- However, as a result of tremendous advances in genomic
tary Genomic Data Submissions (VGDS) concerning and bioinformatic technologies during the second half of
aspects of their drug discovery or development projects, the 20th century and over the past decade, comprehensive
under ‘safe harbour’ conditions without immediate regu- information about the genomes of many organisms is
latory impact. The goal of the VGDS programme was to becoming available to the biomedical research community,
provide ‘real-world’ examples around which to create dia- including pharmaceutical scientists. Genomic information
logue between the FDA and pharmaceutical/biotechnology is the primary type of bioinformation and the nature and
companies about the potential roles of genomic informa- behaviour of the genome of an organism underpins all of
tion in the drug development process and to educate both its activities. It is also ‘relatively’ simple in structure com-
regulators and drug developers. Within a short time of its pared to proteomic or metabolomic information.
initiation, the programme was expanded to include other Genomic information, by itself, has the potential to
molecular data types and is now known as the Voluntary transform the drug discovery and development process, to
Exploratory Data Submissions (VXDS) programme and enhance the probability of success of any project and to
also involves collaboration with the EMEA. A recent article reduce overall costs. Such a transformation will be brought
(Goodsaid et al., 2010) provides an overview of the first 5 about through new drug discovery and development strat-
years of operation of the programme and presents a egies focused on understanding molecular subtypes
number of selected case studies. underpinning disease symptoms, discovering drug targets
Overall the VXDS programme has been of mutual benefit relevant to particular disease molecular subtypes and
to regulators, pharmaceutical companies, technology pro- undertaking clinical studies informed by the knowledge of
viders and academic researchers. As envisioned by the how human genetic individuality can influence treatment
FDA’s Critical Path Initiative, the VXDS programme will outcome. Onward!

91
Section | 2 | Drug Discovery

Table 7.3  EMEA and FDA Genomics guidance documents, concept papers and policies

EMEA guidance documents and concept papers

Reflection paper on co-development of pharmacogenomic http://www.ema.europa.eu/docs/en_GB/document_library/
biomarkers and assays in the context of drug development Scientific_guideline/2010/07/WC500094445.pdf
Use of pharmacogenetic methodologies in the http://www.ema.europa.eu/docs/en_GB/document_library/
pharmacokinetic evaluation of medicinal products Scientific_guideline/2010/05/WC500090323.pdf
ICH concept paper on pharmacogenomic (PG) biomarker http://www.ema.europa.eu/docs/en_GB/document_library/
qualification: Format and data standards Scientific_guideline/2009/09/WC500003863.pdf
Reflection paper on pharmacogenomics in oncology http://www.ema.europa.eu/docs/en_GB/document_library/
Scientific_guideline/2009/09/WC500003866.pdf
ICH Topic E15: Definitions for genomic biomarkers, http://www.ema.europa.eu/docs/en_GB/document_library/
pharmacogenomics, pharmacogenetics, genomic data and Scientific_guideline/2009/09/WC500003888.pdf
sample coding categories
Reflection paper on pharmacogenomic samples, testing and http://www.ema.europa.eu/docs/en_GB/document_library/
data handling Scientific_guideline/2009/09/WC500003864.pdf
Guiding principles: Processing joint FDA EMEA voluntary http://www.ema.europa.eu/docs/en_GB/document_library/
genomic data submissions (VGDSs) within the framework of Scientific_guideline/2009/09/WC500003887.pdf
the confidentiality arrangement
Reflection paper on the use of genomics in cardiovascular http://www.ema.europa.eu/docs/en_GB/document_library/
clinical trials Scientific_guideline/2009/09/WC500003865.pdf
Reflection paper on the use of pharmacogenetics in the http://www.ema.europa.eu/docs/en_GB/document_library/
pharmacokinetic evaluation of medicinal products Scientific_guideline/2009/09/WC500003890.pdf
Pharmacogenetics briefing meeting http://www.ema.europa.eu/docs/en_GB/document_library/
Scientific_guideline/2009/09/WC500003886.pdf
Position paper on terminology in pharmacogenetics http://www.ema.europa.eu/docs/en_GB/document_library/
Scientific_guideline/2009/09/WC500003889.pdf

FDA guidance documents, concept papers and policy documents

Guiding Principles for Joint FDA EMEA Voluntary Genomic http://www.fda.gov/downloads/Drugs/ScienceResearch/
Data Submission Briefing Meetings ResearchAreas/Pharmacogenetics/UCM085378.pdf
Pharmacogenetic Tests and Genetic Tests for Heritable http://www.fda.gov/MedicalDevices/
Markers DeviceRegulationandGuidance/GuidanceDocuments/
ucm077862.htm
Guidance for Industry: Pharmacogenomic Data Submissions http://www.fda.gov/downloads/Drugs/
GuidanceComplianceRegulatoryInformation/Guidances/
ucm079849.pdf
Pharmacogenomic Data Submissions – Companion Guidance http://www.fda.gov/downloads/Drugs/
GuidanceComplianceRegulatoryInformation/Guidances/
ucm079855.pdf
E15 Definitions for Genomic Biomarkers, http://www.fda.gov/downloads/RegulatoryInformation/
Pharmacogenomics, Pharmacogenetics, Genomic Data and Guidances/ucm129296.pdf
Sample Coding Categories
Class II Special Controls Guidance Document: Drug http://www.fda.gov/MedicalDevices/
Metabolizing Enzyme Genotyping System – Guidance for DeviceRegulationandGuidance/GuidanceDocuments/
Industry and FDA Staff ucm077933.htm
Drug-Diagnostic Co-Development Concept Paper http://www.fda.gov/downloads/Drugs/ScienceResearch/
ResearchAreas/Pharmacogenetics/UCM116689.pdf

92
 The role of information, bioinformatics and genomics Chapter |7|

Table 7.3  Continued

Guiding principles Processing Joint FDA EMEA Voluntary http://www.fda.gov/downloads/Drugs/ScienceResearch/
Genomic Data Submissions (VGDSs) within the framework ResearchAreas/Pharmacogenetics/UCM085378.pdf
of the Confidentiality Arrangement
Management of the Interdisciplinary Pharmacogenomics http://www.fda.gov/downloads/AboutFDA/CentersOffices/
Review Group (IPRG) CDER/ManualofPoliciesProcedures/UCM073574.pdf
Processing and Reviewing Voluntary Genomic Data http://www.fda.gov/downloads/AboutFDA/CentersOffices/
Submissions (VGDSs) CDER/ManualofPoliciesProcedures/UCM073575.pdf
Examples of Voluntary Submissions or http://www.fda.gov/downloads/Drugs/
Submissions Required Under 21 CFR GuidanceComplianceRegulatoryInformation/Guidances/
312, 314, or 601 UCM079851.pdf

REFERENCES

Baranzini SL, Elfstrom C, Chang SY, Freeman JL, Perry GH, Feuk L, et al. basis of drug resistance. Future
et al. Transcriptional analysis of Copy number variation: new insights Microbiology 2006;1:185–99.
multiple sclerosis brain lesions in genome diversity. Genome Lander ES, Linton LM, Birren B, et al.
reveals a complex pattern of cytokine Research 2006;16:949–61. Initial sequencing and analysis of the
expression. Journal of Immunology Garnier JP. Rebuilding the R&D engine human genome. Nature
2000;165:6576–82. in big pharma. Harvard Business 2001;409:860–921.
Cantor CR, Smith CL. Genomics. The Review 2008;86:68–70, 72–6, 128. Larsen F, Gundersen G, Lopez R, et al.
science and technology behind the Goodsaid FM, Amur S, Aubrecht J, et al. CpG islands as gene markers in the
Human Genome Project. New York: Voluntary exploratory data human genome. Genomics
John Wiley and Sons; 1999. submissions to the US FDA and the 1992;13:1095–107.
Chandra SK, Caldwell JS. Fulfilling the EMA: experience and impact. Nature Li J, Robson B. Bioinformatics
promise: drug discovery in the Reviews Drug Discovery and computational chemistry
post-genomic era. Drug Discovery 2010;9:435–45. in molecular design. Recent
Today 2003;8:168–74. Greenbaum D, Luscombe NM, Jansen advances and their application.
Cockburn IM. Is the pharmaceutical R, et al. Interrelating different types In: Peptide and protein drug
industry in a productivity crisis? of genomic data, from proteome to analysis. NY: Marcel Dekker; 2000.
Innovation Policy and the Economy secretome: ‘oming in on function. p. 285–307.
2007;7:1–32. Genome Research 2001;11:1463–8. Mullins IM, Siadaty MS, Lyman J, et al.
Crettol S, Petrovic N, Murray M. Hanson, MA, Gluckman, PD. Data mining and clinical data
Pharmacogenetics of Phase I and Developmental origins of health repositories: insights from a 667,000
Phase II drug metabolism. Current and disease: new insights. Basic patient data set. Computers in
Pharmaceutical Design and Clinical Pharmacology and Biology and Medicine 2006;36:
2010;16:204–19. Toxicology 2008;102:90–3. 1351–77.
Dolinoy DC, Huang D, Jirtle RL. Hark AT, Schoenherr CJ, Katz DJ, et al. Munos BH. Lessons from 60 years of
Maternal nutrient supplementation CTCF mediates methylation-sensitive pharmaceutical innovation. Nature
counteracts bisphenol A-induced enhancer-blocking activity at the Reviews Drug Discovery 2009;8:
DNA hypomethylation in early H19/Igf2 locus. Nature 959–68.
development. Proceedings of the 2000;405:486–9. Neumann EK, Quan D. Biodash: a
National Academy of Sciences of the Holliday R, Pugh JE. DNA modification semantic web dashboard for drug
USA 2007;104:13056–61. mechanisms and gene activity during development. Pacific Symposium on
Drews J. Strategic trends in the drug development. Science Biocomputing 2006;11:176–87.
industry. Drug Discovery Today 1975;187:226–32. Niu Y, Gong Y, Langaee TY, et al.
2003;8:411–20. International Genome Sequencing Genetic variation in the β2 subunit
Fraga MF, Ballestar E, Paz MF, et al. Consortium. Finishing the of the voltage-gated calcium channel
Epigenetic differences arise during euchromatic sequence of the and pharmacogenetic association
the lifetime of monozygotic twins. human genome. Nature 2004; with adverse cardiovascular
Proceedings of the National 431:915–6. outcomes in the International
Academy of Sciences of the USA Kidgell C, Winzeler EA. Using the VErapamil SR-Trandolapril STudy
2005;102:10604–9. genome to dissect the molecular GENEtic Substudy (INVEST-GENES).

93
Section | 2 | Drug Discovery

Circulation. Cardiovascular Genetics Robson B. Clinical and Sadee W, Hoeg E, Lucas J, et al. Genetic
2010;3:548–55. phamacogenomic data mining. 4. variations in human g protein-
Paul SM, Mytelka DS, Dunwiddie CT, The FANO program and command coupled receptors: implications for
et al. How to improve R&D set as an example of tools for drug therapy. AAPS PharmSci
productivity: the pharmaceutical biomedical discovery and evidence 2001;3:E22.
industry’s grand challenge. Nature based medicine. Journal of Proteome Stephens S, Morales A, Quinlan M.
Reviews Drug Discovery Research 2008;7:3922–47. Applying semantic web technologies
2010;9:203–14. Robson B, Baek OK. The engines of to drug safety determination. IEEE
Peltonen L, Perola M, Naukkarinen J, Hippocrates: from medicine’s early Intelligent Systems 2006;21:82–6.
et al. Lessons from studying dawn to medical and pharmaceutical Svinte M, Robson B, Hehenberger M.
monogenic disease for common informatics. New York: John Wiley & Biomarkers in drug development
disease. Human Molecular Genetics Sons; 2009. and patient care. Burrill 2007 Person.
2006;15:R67–R74. Robson B, Mushlin R. Clinical and Medical Report 2007;6:3114–26.
Pucci MJ. Novel genetic techniques and pharmacogenomic data mining. 2. A van der Greef J, McBurney RN. Rescuing
approaches in the microbial simple method for the combination drug discovery: in vivo systems
genomics era: identification and/or of information from associations pathology and systems
validation of targets for the discovery and multivariances to facilitate pharmacology. Nature Reviews Drug
of new antibacterial agents. Drugs analysis, decision and design in Discovery 2005;4:961–7.
R&D 2007;8:201–12. clinical research and practice. Journal van der Greef J, Martin S, Juhasz P, et al.
Robinson R. Common disease, multiple of Proteome Research The art and practice of systems
rare (and distant) variants. PLoS 2004;3:697–711. biology in medicine: mapping
Biology 2010;8(1): e1000293. Robson B, Vaithiligam A. Drug gold patterns of relationships. Journal of
Robson B. Clinical and and data dragons: Myths and Proteome Research 2007;4:1540–59.
pharmacogenomic data mining. 1. realities of data mining in the Venter JC, Adams MD, Myers EW, et al.
The generalized theory of expected pharmaceutical industry. In: Balakin The sequence of the human genome.
information and application to the KV, editor. Pharmaceutical Data Science 2001;291:1304–51.
development of tools. Journal of Mining. John Wiley & Sons; 2010.
Waddington CH. Introduction to
Proteome Research 2003;2:283–301. p. 25–85.
modern genetics. London: Allen and
Robson B. The dragon on the gold: Ruttenberg A, Rees JA, Samwald M, Unwin; 1939.
Myths and realities for data mining et al. Life sciences on the semantic
Wu CT, Morris JR. Genes, genetics, and
in biotechnology using digital and web: the neurocommons and
epigenetics: a correspondence.
molecular libraries. Journal of beyond. Briefings in Bioinformatics
Science 2001;293:1103–5.
Proteome Research 2004;3:1113–9. 2009;10:193–204.
Zhang JP, Lencz T, Malhotra AK. D2
Robson B. Clinical and Sacca R, Engle SJ, Qin W, et al.
receptor genetic variation and
pharmacogenomics data mining. 3 Genetically engineered mouse
clinical response to antipsychotic
Zeta theory as a general tactic for models in drug discovery research.
drug treatment: a meta-analysis.
clinical bioinformatics. Journal of Methods in Molecular Biology
American Journal of Psychiatry
Proteome Research 2005;4;445–55. 2010;602:37–54.
2010;167:763–72.

94
 Chapter 8 
High-throughput screening
D Cronk

diverse drugs, including β-adrenoceptor agonists and

antagonists, benzodiazepines, angiotensin receptor antag-
INTRODUCTION: A HISTORICAL AND
onists and ultimately monoclonal antibodies.
FUTURE PERSPECTIVE Today’s marketed drugs are believed to target a range
of human biomolecules (see Chapters 6 and 7), ranging
Systematic drug research began about 100 years ago, when from various enzymes and transporters to G-protein-
chemistry had reached a degree of maturity that allowed coupled receptors (GPCRs) and ion channels. At present
its principles and methods to be applied to problems the GPCRs are the predominant target family, and more
outside the field, and when pharmacology had in turn than 800 of these biomolecules have been identified
become a well-defined scientific discipline. A key step in the human genome (Kroeze et al., 2003). However, it
was the introduction of the concept of selective affinity is predicted that less than half are druggable and in reality
through the postulation of ‘chemoreceptors’ by Paul proteases and kinases may offer greater potential as targets
Ehrlich. He was the first to argue that differences in chem- for pharmaceutical products (Russ and Lampel, 2005).
oreceptors between species may be exploited therapeuti- Although the target portfolio of a pharmaceutical company
cally. This was also the birth of chemotherapy. In 1907, can change from time to time, the newly chosen targets
Ehrlich identified compound number 606, Salvarsan are still likely to belong to one of the main therapeutic
(diaminodioxy-arsenobenzene) (Ehrlich and Bertheim, target classes. The selection of targets and target families
1912), which was brought to the market in 1910 by (see Chapter 6) plays a pivotal role in determining the
Hoechst for the treatment of syphilis, and hailed as a success of today’s lead molecule discovery.
miracle drug (Figure 8.1). Over the last 15 years significant technological progress
This was the first time extensive pharmaceutical screen- has been achieved in genomic sciences (Chapter 7), high-
ing had been used to find drugs. At that time screening throughput medicinal chemistry (Chapter 9), cell-based
was based on phenotypic readouts e.g. antimicrobial assays and high-throughput screening. These have led to a
effect, a concept which has since led to unprecedented ‘new’ concept in drug discovery whereby targets with ther-
therapeutic triumphs in anti-infective and anticancer ther- apeutic potential are incorporated into biochemical or
apies, based particularly on natural products. In contrast, cell-based assays which are exposed to large numbers of
today’s screening is largely driven by distinct molecular compounds, each representing a given chemical structure
targets and relies on biochemical readout. space. Massively parallel screening, called high-throughput
In the further course of the 20th century drug research screening (HTS), was first introduced by pharmaceutical
became influenced primarily by biochemistry. The domi- companies in the early 1990s and is now employed rou-
nant concepts introduced by biochemistry were those of tinely as the most widely applicable technology for iden-
enzymes and receptors, which were empirically found to tifying chemistry starting points for drug discovery
be drug targets. In 1948 Ahlquist made a crucial, further programmes.
step by proposing the existence of two types of adrenocep- Nevertheless, HTS remains just one of a number of pos-
tor (α and β) in most organs. The principle of receptor sible lead discovery strategies (see Chapters 6 and 9). In
classification has been the basis for a large number of the best case it can provide an efficient way to obtain

© 2012 Elsevier Ltd. 95

Section | 2 | Drug Discovery

collections in sufficient quantities and of sufficient quality

to file them by electronic systems, and store them in the
most appropriate way in compound archives. This resulted
in huge collections that range from several hundred thou-
sand to a few million compounds. Today’s focus has
shifted to the application of defined electronic or physical
filters for compound selection before they are assembled
into a library for testing. The result is a customized ensem-
ble of either newly designed or historic compounds for
use in screening, otherwise known as ‘cherry picking’.
However, it is often the case that the HTS departments
have sufficient infrastructure to enable routine screening
of the entire compound collection and it is only where the
assay is complex or relatively expensive that the time to
create ‘cherry picked’, focused, compound sets is invested
(Valler and Green, 2000).
In assay development there is a clear trend towards
mechanistically driven high-quality assays that capture the
relevant biochemistry (e.g. stochiometry, kinetics) or cell
biology. Homogeneous assay principles, along with sensi-
tive detection technologies, have enabled the miniaturiza-
tion of assay formats producing a concomitant reduction
of reagent usage and cost per data point. With this evolu-
tion of HTS formats it is becoming increasingly common
to gain more than one set of information from the same
assay well either through multiparametric analysis or mul-
tiplexing, e.g. cellular function response and toxicity
(Beske and Goldbard, 2002; Hanson, 2006; Hallis et al.,
2007). The drive for information rich data from HTS cam-
Fig. 8.1  In France, where Salvarsan was called ‘Formule paigns is no more evident than through the use of imaging
606’, true miracles were expected from the new therapy. technology to enable subcellular resolution, a methodol-
ogy broadly termed high content screening (HCS). HCS
assay platforms facilitate the study of intracellular phar-
macology through spatiotemporal resolution, and the
quantification of signalling and regulatory pathways. Such
useful data on the biological activity of large numbers of techniques increasingly use cells that are more phenotypi-
test samples by using high-quality assays and high-quality cally representative of disease states, so called disease-
chemical compounds. Today’s lead discovery departments relevant cell lines (Clemons, 2004), in an effort to add
are typically composed of the following units: (1) com- further value to the information provided.
pound logistics; (2) assay development and screening Screening departments in large pharmaceutical com­
(which may utilize automation); (3) tool (reagent) pro- panies utilize automated screening platforms, which in the
duction; and (4) profiling. Whilst most HTS projects focus early days of HTS were large linear track systems, typically
on the use of synthetic molecules typically within a molec- five metres or more in length. The more recent trends have
ular weight range of 250–600 Da, some companies are been towards integrated networks of workstation-based
interested in exploring natural products and have dedi- instrumentation, typically arranged around the circumfer-
cated research departments for this purpose. These groups ence of a static, rotating robotic arm, which offers greater
work closely with the HTS groups to curate the natural flexibility and increased efficiency in throughput due to
products, which are typically stored as complex mixtures, reduced plate transit times within the automated workcell.
and provide the necessary analytical skills to isolate the Typically, the screening unit of a large pharmaceutical
single active molecule. company will generate tens of millions of single point
Compared with initial volume driven HTS in the 1990s determinations per year, with fully automated data acqui-
there is now much more focus on quality-oriented output. sition and processing. Following primary screening, there
At first, screening throughput was the main emphasis, but has been an increased need for secondary/complementary
it is now only one of many performance indicators. In the screening to confirm the primary results, provide informa-
1990s the primary concern of a company’s compound tion on test compound specificity and selectivity and to
logistics group was to collect all its historic compound refine these compounds further. Typical data formats

96
 High-throughput screening Chapter |8|

include half-maximal concentrations at which a com-

pound causes a defined modulatory effect in functional
LEAD DISCOVERY AND HIGH-
assays, or binding/inhibitory constants. Post-HTS, broader
selectivity profiling may be required, for active compounds THROUGHPUT SCREENING
against panels of related target families. As HTS technolo-
gies are adopted into other related disciplines compound A lead compound is generally defined as a new chemical
potency and selectivity are no longer the only parameters entity that could potentially be developed into a new drug
to be optimized during hit-finding. With this broader by optimizing its beneficial effects and minimizing its side
acceptance of key technologies, harmonization and stand- effects (see Chapter 9 for a more detailed discussion of the
ardization of data across disciplines are crucial to facilitate criteria). HTS is currently the main approach for the iden-
analysis and mining of the data. Important information tification of lead compounds, i.e. large numbers of com-
such as compound purity and its associated physico- pounds (the ‘compound library’) are usually tested in a
chemical properties such as solubility can be derived very random approach for their biological activity against a
quickly on relatively large numbers of compounds and disease-relevant target. However, there are other tech-
thus help prioritize compounds for progression based on niques in place for lead discovery that are complementary
overall suitability, not just potency (Fligge and Schuler, to HTS.
2006). These quality criteria, and quality assessment at all Besides the conventional literature search (identifica-
key points in the discovery process, are crucial. Late-stage tion of compounds already described for the desired activ-
attrition of drug candidates, particularly in development ity), structure-based virtual screening is a frequently
and beyond, is extremely expensive and such failures must applied technique (Ghosh et al., 2006; Waszkowycz,
be kept to a minimum. This is typically done by an exten- 2008). Molecular recognition events are simulated by
sive assessment of chemical integrity, synthetic accessibil- computational techniques based on knowledge of the
ity, functional properties, structure–activity relationship molecular target, thereby allowing very large ‘virtual’ com-
(SAR) and biophysicochemical properties, and related pound libraries (greater than 4 million compounds) to
absorption, distribution, metabolism and excretion be screened in silico and, by applying this information,
(ADME) characteristics, as discussed further in Chapters 9 pharmacophore models can be developed. These allow
and 10. the identification of potential leads in silico, without
In summary, significant technological progress has been experimental screening and the subsequent construction
made over the last 15 years in HTS. Major concepts such of smaller sets of compounds (‘focused libraries’) for
as miniaturization and parallelization have been intro- testing against a specific target or family of targets (Stahura
duced in almost all areas and steps of the lead discovery et al., 2002; Muegge and Oloff, 2006). Similarly, X-ray
process. This, in turn, has led to a great increase in screen- analysis of the target can be applied to guide the de novo
ing capacity, significant savings in compound or reagent synthesis and design of bioactive molecules. In the absence
consumption, and, ultimately, improved cost-effectiveness. of computational models, very low-molecular-weight
More recently, stringent quality assessment in library man- compounds (typically 150–300 Da, so-called fragments),
agement and assay development, along with consistent may be screened using biophysical methods to detect low-
data formats in automated screening, has led to much affinity interactions. The use of protein crystallography
higher-quality screening outcomes. The perception of HTS and X-ray diffraction techniques allows elucidation of the
has also changed significantly in the past decade and is binding mode of these fragments and these can be used
now recognized as a multidisciplinary science, encom­ as a starting point for developing higher affinity leads by
passing biological sciences, engineering and information assemblies of the functional components of the fragments
technology. HTS departments generate huge amounts of (Rees et al., 2004; Hartshorn et al., 2005; Congreve et al.,
data that can be used together with computational chem- 2008).
istry tools to drive compound structure–activity relation- Typically, in HTS, large compound libraries are screened
ships and aid selection of focused compound sets for (‘primary’ screen) and numerous bioactive compounds
further testing from larger compound libraries. Where (‘primary hits’ or ‘positives’) are identified. These com-
information rich assays are used complex analysis algo- pounds are taken through successive rounds of further
rithms may be required to ensure the relevant data are screening (’secondary’ screens) to confirm their activity,
extracted. Various statistical, informatics and filtering potency and where possible gain an early measure of spe-
methods have recently been introduced to foster the inte- cificity for the target of interest. A typical HTS activity
gration of experimental and in silico screening, and so cascade is shown in Figure 8.2 resulting in the identifica-
maximize the output in lead discovery. As a result, lead- tion of hits, usually with multiple members of a similar
finding activities continue to benefit greatly from a more chemical core or chemical series. These hits then enter into
unified and knowledge-based approach to biological the ‘hit-to-lead’ process during which medicinal chemistry
screening, in addition to the many technical advances teams synthesize specific compounds or small arrays of
towards even higher-throughput screening. compounds for testing to develop an understanding of the

97
Section | 2 | Drug Discovery

Assay development
Validate assay conditions and undertake assay optimization for screening if required
Validate pharmacology and demonstrate assay robustness for screening, usually
including testing of a limited number of compounds in duplicate on separate days

Compound selection and plating

Selection of compounds for screening
and preparation of screen ready plates

Primary screening
Screening of all selected compounds at a single, fixed concentration as n = 1
Selection of ‘hit compounds’ for further testing (typically 1% of compounds progressed)

Hit confirmation
Screening of ‘hit compounds’ at a single fixed concentration in duplicate against
selected target and a suitable control assay to eliminate false positives
Selection of ‘confirmed hit compounds’

Potency determination
Test ‘confirmed hit compounds’ over a concentration range to determine potency
against selected target and in a suitable control assay

Fig. 8.2  The typical high throughput screening process.

Data variability band Data variability band

Separation band
Frequency

3 σs 3 σc

Low controls Assay signal High controls

Fig. 8.3  Illustration of data variability and the signal window, given by the separation band between high and low controls.
Adapted, with permission, from Zhang et al., 1999.

structure–activity relationship (SAR) of the underlying improved by medicinal chemistry in a ‘lead optimization’
chemical series. The result of the hit-to-lead phase is a process (Figure 8.3). Often the HTS group will provide
group of compounds (the lead series) which has appropri- support for these hit-to-lead and lead optimization stages
ate drug-like properties such as specificity, pharmacokinet- through ongoing provision of reagents, provision of assay
ics or bioavailability. These properties can then be further expertise or execution of the assays themselves.

98
 High-throughput screening Chapter |8|

Assay development and validation Once a decision on the principal format and readout
technology is taken, the assay has to be validated for its
The target validation process (see Chapters 6 and 7) estab- sensitivity and robustness. Biochemical parameters, rea-
lishes the relevance of a target in a certain disease pathway. gents and screening hardware (e.g. detectors, microtitre
In the next step an assay has to be developed, allowing the plates) must be optimized. To give a practical example, in
quantification of the interaction of molecules with the a typical screen designed for inhibitors of protease activity,
chosen target. This interaction can be inhibition, stimula- test compounds are mixed together with the enzyme and
tion, or simply binding. There are numerous different finally substrate is added. The substrate consists of a cleav-
assay technologies available, and the choice for a specific able peptide linked to a fluorescent label, and the reaction
assay type will always be determined by factors such as is quantified by measuring the change in fluoresecence
type of target, the required sensitivity, robustness, ease of intensity that accompanies the enzymic cleavage. In the
automation and cost. Assays can be carried out in different process of validation, the best available labelled substrate
formats based on 96-, 384-, or 1536-well microtitre plates. (natural or synthetic) must be selected, the reaction condi-
The format to be applied depends on various parameters, tions optimized (for example reaction time, buffers and
e.g. readout, desired throughput, or existing hardware in temperature), enzyme kinetic measurements performed to
liquid handling and signal detection with 384- (either identify the linear range, and the response of the assay
standard volume or low volume) and 1536-well formats to known inhibitors (if available) tested. Certain types
being the most commonly applied. In all cases the homo- of compound or solvent (which in most cases will be
geneous type of assay is preferred, as it is quicker, easier dimethylsulfoxide, DMSO) may interfere with the assay
to handle and cost-effective, allowing ‘mix and measure’ readout and this has to be checked. The stability of assay
operation without any need for further separation steps. reagents is a further important parameter to be determined
Next to scientific criteria, cost is a key factor in assay during assay validation, as some assay formats require a
development. The choice of format has a significant effect long incubation time.
on the total cost per data point: the use of 384-well low- At this point other aspects of screening logistics have to
volume microtitre plates instead of a 96-well plate format be considered. If the enzyme is not available commercially
results in a significant reduction of the reaction volume it has to be produced in-house by process development,
(see Table 8.1). This reduction correlates directly with and batch-to-batch reproducibility and timely delivery
reagent costs per well. The size of a typical screening have to be ensured. With cell-based screens it must be
library is between 500 000 and 1 million compounds. guaranteed that the cell production facility is able to
Detection reagent costs per well can easily vary between deliver sufficient quantities of consistently functioning,
US$0.05 and more than U$0.5 per data point, depending physiologically intact cells during the whole screening
on the type and format of the assay. Therefore, screening campaign and that there is no degradation of signal or loss
an assay with a 500 000 compound library may cost either of protein expression from the cells with extended periods
US$25 000 or US$250 000, depending on the selected of subculture.
assay design – a significant difference! It should also be The principal goal of developing HTS assays is the fast
borne in mind that these costs are representative for rea- and reliable identification of active compounds (‘posi-
gents only and the cost of consumables (assay plates and tives’ or ‘hits’) from chemical libraries. Most HTS pro-
disposable liquid handling tips) may be an additional grammes test compounds at only one concentration. In
consideration. Whilst the consumables costs are higher for most instances this approximates to a final test concentra-
the higher density formats, the saving in reagent costs and tion in the assay of 10 micromolar. This may be adjusted
increased throughput associated with miniaturization depending on the nature of the target but in all cases must
usually result in assays being run in the highest density be within the bounds of the solvent tolerance of the assay
format the HTS department has available. determined earlier in the development process. In order
to identify hits with confidence, only small variations
in signal measurements can be tolerated. The statistical
parameters used to determine the suitability of assays for
HTS are the calculation of standard deviations, the coef-
Table 8.1  Reaction volumes in microtitre plates
ficient of variation (CV), signal-to-noise (S/N) ratio or
signal-to-background (S/B) ratio. The inherent problem
Plate format Typical assay volume
with using these last two is that neither takes into account
96 100–200 µL the dynamic range of the signal (i.e. the difference between
the background (low control) and the maximum (high
384 25–50 µL
control) signal), or the variability in the sample and refer-
384 low volume 5–20 µL ence control measurements. A more reliable assessment of
1536 2–10 µL assay quality is achieved by the Z’-factor equation (Zhang
et al., 1999):

99
Section | 2 | Drug Discovery

(3(SD of High Control) + 3(SD of Low Control)) present in the compound wells, and assuming a low
Z’ = 1 − number of active compounds, the Z-value is usually lower
[Mean of High Control − Mean of Low Control]
than the Z’-value.
where SD = standard deviation and the maximum possible
Whilst there are been several alternatives of Zhang’s
value of Z is 1. For biochemical assays a value greater than
proposal for assessing assay robustness, such as power
0.5 represents a good assay whereas a value less than 0.5
analysis (Sui and Wu, 2007), the simplicity of the equation
is generally unsatisfactory for HTS. A lower Z’ threshold of
still make the Z’-value the primary assessment of assay
0.4 is usually considered acceptable for cell-based assays.
suitability for HTS.
This equation takes into account that the quality of
The Assay Guidance Website hosted by the National
an assay is reflected in the variability of the high and low
Institutes of Health Center for Translational Therapeutics
controls, and the separation band between them (Figure
(NCTT) (http://assay.nih.gov/assay/index.php/Table_of_
8.3). Z’-factors are obtained by measuring plates contain-
Contents) provides comprehensive guidance of factors to
ing 50% low controls (in the protease example: assay plus
consider for a wide range of assay formats.
reference inhibitor, minimum signal to be measured) and
50% high controls (assay without inhibitor; maximum
signal to be measured). In addition, inter- and intra-plate Biochemical and cell-based assays
coefficients of variation (CV) are determined to check for There is a wide range of assays formats that can be deployed
systematic sources of variation. All measurements are nor- in the drug discovery arena (Hemmilä and Hurskainen,
mally made in triplicate. Once an assay has passed these 2002), although they broadly fall into two categories: bio-
quality criteria it can be transferred to the robotic screen- chemical and cell-based.
ing laboratory. A reduced number of control wells can be Biochemical assays (Figure 8.4) involve the use of cell-
employed to monitor Z’-values when the assay is pro- free in-vitro systems to model the biochemistry of a subset
gressed to HTS mode, usually 16 high- and 16 low-controls of cellular processes. The assay systems vary from simple
on a 384-well plate, with the removal of no more than two interactions, such as enzyme/substrate reactions, receptor
outlying controls to achieve an acceptable Z’-value. The binding or protein–protein interactions, to more complex
parameter can be further modified to calculate the Z-value, models such as in-vitro transcription systems. In contrast
whereby the average signal and standard deviation of test to cell-based assays, biochemical assays give direct infor-
compound wells are compared to the high-control wells mation regarding the nature of the molecular interaction
(Zhang et al., 1999). Due to the variability that will be (e.g. kinetic data) and tend to have increased solvent

Biochemical assays

Homogeneous Heterogeneous Alternative

Radioactive Non-radioactive Radioactive Non-radioactive

Scintillation Fluorescence Filtration ELISA Capillary electrophoresis

Absorption DELFIA Frontal affinity chromatography
Flash plate Intensity Precipitation Biophysical (NMR, SPR)
Bead FRET Radioimmunoassay Quantitative-PCR
TRF
FP
FC
Alpha-screen

Absorbance
(Chemi) Luminescence

Fig. 8.4  Types of biochemical assay.

100
 High-throughput screening Chapter |8|

tolerance compared to cellular assays, thereby permitting Cell-based assays frequently lead to higher hit rates,
the use of higher compound screening concentration if because of non-specific and ‘off-target’ effects of test com-
required. However, biochemical assays lack the cellular pounds that affect the readout. Primary hits therefore
context, and are insensitive to properties such as mem- need to be assessed by means of secondary assays such as
brane permeability, which determine the effects of com- non- or control-transfected cells in order to determine the
pounds on intact cells. mechanism of the effect (Moore and Rees, 2001).
Unlike biochemical assays, cell-based assays (Figure 8.5) Although cell-based assays are generally more time-
mimic more closely the in-vivo situation and can be consuming than cell-free assays to set up and run in
adapted for targets that are unsuitable for screening in high-throughput mode, there are many situations in which
biochemical assays, such as those involving signal trans- they are needed. For example, assays involving G-protein
duction pathways, membrane transport, cell division, coupled receptors (GPCRs), membrane transporters and
cytotoxicity or antibacterial actions. Parameters measured ion channels generally require intact cells if the functional-
in cell-based assays range from growth, transcriptional ity of the test compound is to be understood, or at least
activity, changes in cell metabolism or morphology, to membranes prepared from intact cells for determining
changes in the level of an intracellular messenger such as compound binding. In other cases, the production of bio-
cAMP, intracellular calcium concentration and changes in chemical targets such as enzymes in sufficient quantities
membrane potential for ion channels (Moore and Rees, for screening may be difficult or costly compared to cell-
2001). Importantly, cell-based assays are able to distin- based assays directed at the same targets. The main pros
guish between receptor antagonists, agonists, inverse ago- and cons of cell-based assays are summarized in Table 8.2.
nists and allosteric modulators which cannot be done by
measuring binding affinity in a biochemical assay.
Many cell-based assays have quite complex protocols,
for example removing cell culture media, washing cells, Table 8.2  Advantages and disadvantages of
adding compounds to be tested, prolonged incubation at cell-based assays
37°C, and, finally, reading the cellular response. There-
fore, screening with cell-based assays requires a sophisti- Advantages Disadvantages
cated infrastructure in the screening laboratory (including
Cytotoxic compounds can be Require high-capacity
cell cultivation facilities, and robotic systems equipped to
detected and eliminated at cell culture facilities
maintain physiological conditions during the assay proce-
the outset and more challenging
dure) and the throughput is generally lower. to fully automate
In receptor studies, agonists Often require specially
can be distinguished from engineered cell lines
antagonists and/or careful selection
Cell-based assays of control cells
Detection of allosteric Reagent provision and
modulators control of variability of
reagent batches
Homogeneous Heterogeneous
Binding and different Cells liable to become
ELISA functional readouts can be detached from support
Alpha-LISA used in parallel – high
information content
Non-radioactive Radioactive
Phenotypic readouts are High rate of false
enabling when the molecular positives due to
Biomarkers Filtration
target is unknown (e.g. to non-specific effects of
Reporter gene Radioimmunoassay detect compounds that affect test compounds on cell
High content screening cell division, growth, function
Yeast 2-hybrid differentiation or metabolism)
Growth and proliferation
More disease relevant than Assay variability can
Second messenger (e.g.cAMP, Ca2+)
biochemical asays make assays more
Label free technology difficult to miniaturize
High throughput electrophysiology
Membrane potential No requirement for protein Assay conditions (e.g.
production/scale up use of solvents, pH)
limited by cell viability
Fig. 8.5  Types of cell-based assay.

101
Section | 2 | Drug Discovery

Assay readout and detection the radioligand to the target brings it into close proximity
to the scintillant, resulting in light emission, which can be
Ligand binding assays quantified. Free radioactive ligand is too distant from the
Assays to determine direct interaction of the test com- scintillant and no excitation takes place. Isotopes such as
pound with the target of interest through the use of radio­ 3
H or 125I are typically used, as they produce low-energy
labelled compounds are sensitive and robust and are particles that are absorbed over short distances (Cook,
widely used for ligand-binding assays. The assay is based 1996). Test compounds that bind to the target compete
on measuring the ability of the test compound to inhibit with the radioligand, and thus reduce the signal.
the binding of a radiolabelled ligand to the target, and With bead technology (Figure 8.8A), polymer beads of
requires that the assay can distinguish between bound and ~5 µm diameter are coated with antibodies, streptavidin,
free forms of the radioligand. This can be done by physical receptors or enzymes to which the radioligand can bind
separation of bound from unbound ligand (heterogeneous (Bosworth and Towers, 1989; Beveridge et al., 2000).
format) by filtration, adsorption or centrifugation. The Ninety-six- or 384-well plates can be used. The emission
need for several washing steps makes it unsuitable for fully wavelength of the scintillant is in the range of 420 nm and
automated HTS, and generates large volumes of radioactive is subject to limitations in the sensitivity due to both
waste, raising safety and cost concerns over storage and colour quench by yellow test compounds, and the variable
disposal. Such assays are mainly restricted to 96-well efficiency of scintillation counting, due to sedimentation
format due to limitations of available multiwell filter plates of the beads. The homogeneous platforms are also still
and achieving consistent filtration when using higher subject to limitations in throughput associated with the
density formats. Filtration systems do provide the advan- detection technology via multiple photomultiplier tube-
tage that they allow accurate determination of maximal based detection instruments, with a 384-well plate taking
binding levels and ligand affinities at sufficient throughput in the order of 15 minutes to read.
for support of hit-to-lead and lead optimization activities. The drive for increased throughput for radioactive assays
In the HTS arena, filtration assays have been superseded led to development of scinitillants, containing europium
by homogeneous formats for radioactive assays. These have yttrium oxide or europium polystyrene, contained in
reduced overall reaction volume and eliminate the need beads or multiwell plates with an emission wavlength
for separation steps, largely eliminating the problem of shifted towards the red end of the visible light spectrum
waste disposal and provide increased throughput. (~560 nm) and suited to detection on charge-coupled
The majority of homogenous radioactive assay types are device (CCD) cameras (Ramm, 1999). The two most
based on the scintillation proximity principle. This relies widely adopted instruments in this area are LEADseeker™
on the excitation of a scintillant incorporated in a matrix, (GE Healthcare) and Viewlux™ (Perkin Elmer), using
in the form of either microbeads (’SPA’) or microplates quantitative imaging to scan the whole plate, resulting in
(Flashplates™, Perkin Elmer Life and Analytical Sciences) a higher throughput and increased sensitivity. Imaging
(Sittampalam et al., 1997), to the surface of which the instruments provide a read time typically in the order
target molecule is also attached (Figure 8.6). Binding of of a few minutes or less for the whole plate irrespective
of density, representing a significant improvement in
throughput, along with increased sensitivity. The problem
of compound colour quench effect remains, although blue
Scintillant compounds now provide false hits rather than yellow. As
impregnated bead CCD detection is independent of plate density, the use of
imaging based radioactive assays has been adopted widely
in HTS and adapted to 1536-well format and higher (see
Bays et al., 2009, for example).
In the microplate form of scintillation proximity assays
the target protein (e.g. an antibody or receptor) is coated
on to the floor of a plate well to which the radioligand
and test compounds are added. The bound radioligand
causes a microplate surface scintillation effect (Brown
et al., 1997). FlashPlate™ has been used in the investiga-
tion of protein–protein (e.g. radioimmunoassay) and
Free radioligand – Bound radioligand – receptor–ligand (i.e. radioreceptor assay) interactions
energy absorbed by energy absorbed by (Birzin and Rohrer, 2002), and in enzymatic (e.g. kinase)
medium. NO LIGHT bead. LIGHT
assays (Braunwaler et al., 1996).
Due to the level of sensitivity provided by radioactive
Fig. 8.6  Principle of scintillation proximity assays. assays they are still widely adopted within the HTS setting.
Reproduced with kind permission of GE Healthcare. However, environmental, safety and local legislative

102
 High-throughput screening Chapter |8|

considerations have led to the necessary development of

absorbed by quenc
alternative formats, in particular those utilizing fluorescent- ergy her
ligands (Lee et al., 2008; Leopoldo et al., 2009). Through En
careful placement of a suitable fluorophore in the ligand
via a suitable linker, the advantages of radioligand binding Fluorophore Quencher
assays in terms of sensitivity can be realized without the
obvious drawbacks associated with the use of radioiso-
topes. The use of fluorescence-based technologies is dis- FRET peptide
cussed in more detail in the following section.

Fluorescence technologies
Protease
The application of fluorescence technologies is wide-
spread, covering multiple formats (Gribbon and Sewing,
2003) and yet in the simplest form involves excitation of Fluorophore Quencher
a sample with light at one wavelength and measurement
of the emission at a different wavelength. The difference
between the absorbed wavelength and the emitted wave- +
length is called the Stokes shift, the magnitude of which
depends on how much energy is lost in the fluorescence Fig. 8.7  Protease assay based on FRET. The donor
process (Lakowicz, 1999). A large Stokes shift is advanta- fluorescence is quenched by the neighbouring acceptor
geous as it reduces optical crosstalk between photons from molecule. Cleavage of the substrate separates them,
the excitation light and emitted photons. allowing fluorescent emission by the donor molecule.
Fluorescence techniques currently applied for HTS can
be grouped into six major categories:
• Fluorescence intensity or a quenching group; this results in measurable photon
• Fluorescence resonance energy transfer emission by the acceptor. In simple terms, the amount of
• Time-resolved fluorescence energy transfer from donor to acceptor depends on the
• Fluorescence polarization fluorescent lifetime of the donor, the spatial distance
• Fluorescence correlation between donor and acceptor (10–100 Å), and the dipole
• AlphaScreen™ (amplified luminescence proximity orientation between donor and acceptor. The transfer effi-
homogeneous assay). ciency for a given pair of fluorophores can be calculated
using the equation of Förster (Clegg, 1995).
Fluorescence intensity Usually the emission wavelengths of donor and accep-
In fluorescence intensity assays, the change of total light tor are different, and FRET can be determined either by the
output is monitored and used to quantify a biochemical quenching of the donor fluorescence by the acceptor (as
reaction or binding event. This type of readout is fre- shown in Figure 8.7) or by the fluorescence of the acceptor
quently used in enzymatic assays (e.g. proteases, lipases). itself. Typical applications are for protease assays based on
There are two variants: fluorogenic assays and fluorescence quenching of the uncleaved substrate, although FRET has
quench assays. In the former type the reactants are not also been applied for detecting changes in membrane
fluorescent, but the reaction products are, and their forma- potential in cell-based assays for ion channels (Gonzalez
tion can be monitored by an increase in fluorescence and Maher, 2002). With simple FRET techniques interfer-
intensity. ence from background fluorescence is often a problem,
In fluorescence quench assays a fluorescent group is which is largely overcome by the use of time-resolved fluo-
covalently linked to a substrate. In this state, its fluores- rescence techniques, described below.
cence is quenched. Upon cleavage, the fluorescent group
is released, producing an increase in fluorescence intensity
Time resolved fluorescence (TRF)
(Haugland, 2002). TRF techniques (Comley, 2006) use lanthanide chelates
Fluorescence intensity measurements are easy to run (samarium, europium, terbium and dysprosium) that
and cheap. However, they are sensitive to fluorescent inter- give an intense and long-lived fluorescence emission
ference resulting from the colour of test compounds, (>1000 µs). Fluorescence emission is elicited by a pulse of
organic fluorophores in assay buffers and even fluores- excitation light and measured after the end of the pulse,
cence of the microplate itself (Comley, 2003). by which time short-lived fluorescence has subsided. This
makes it possible to eliminate short-lived autofluorescence
Fluorescence resonance energy transfer (FRET) and reagent background, and thereby enhance the signal-
In this type of assay a donor fluorophore is excited and to-noise ratio. Lanthanides emit fluorescence with a large
most of the energy is transferred to an acceptor fluorophore Stokes shift when they coordinate to specific ligands.

103
Section | 2 | Drug Discovery

Typically, the complexes are excited by UV light, and emit

light of wavelength longer than 500 nm. Donor fluorophore Acceptor fluorophore
Europium (Eu3+) chelates have been used in immu-
noassays by means of a technology called DELFIA
(dissociation-enhanced lanthanide fluoroimmuno assay).
DELFIA is a heterogeneous time-resolved fluorometric
assay based on dissociative fluorescence enhancement. Cell- +
and membrane-based assays are particularly well suited to Analyte
the DELFIA system because of its broad detection range
and extremely high sensitivity (Valenzano et al., 2000).
High sensitivity – to a limit of about 10−17 moles/well
– is achieved by applying the dissociative enhancement
principle. After separation of the bound from the free
label, a reagent is added to the bound label which causes Excitation Emission Emission
the weakly fluorescent lanthanide chelate to dissociate and light source 615nm 665nm
form a new highly fluorescent chelate inside a protective
micelle. Though robust and very sensitive, DELFIA assays
are not ideal for HTS, as the process involves several
binding, incubation and washing steps.
The need for homogeneous (‘mix and measure’) assays
led to the development of LANCETM (Perkin Elmer Life
Sciences) and HTRF® (Homogeneous Time-Resolved Fluo-
rescence; Cisbio). LANCETM, like DELFIA®, is based on
chelates of lanthanide ions, but in a homogeneous format.
The chelates used in LANCETM can be measured directly Fig. 8.8  HTRF assay type: the binding of a europium-labelled
without the need for a dissociation step, however in an ligand (= donor) to the allophycocyanine (APC = acceptor)-
aqueous environment the complexed ion can spontane- labelled receptor brings the donor–acceptor pair into close
ously dissociate and increase background fluorescence proximity and energy transfer takes place, resulting in
(Alpha et al., 1987). fluorescence emission at 665 nm.
Reproduced with kind permission of Cisbio.
In HTRF® (Figure 8.8) these limitations are overcome by
the use of a cryptate molecule, which has a cage-like struc-
ture, to protect the central ion (e.g. Eu+) from dissociation. depolarized emission can be used to determine the extent
HTRF® uses two separate labels, the donor (Eu)K and the of binding of a labelled ligand (Figure 8.11; Nasir and
acceptor APC/XL665 (a modified allophycocyanine from Jolley, 1999). The rotational relaxation speed depends on
red algae) and such assays can be adapted for use in plates the size of the molecule, the ambient temperature and the
up to 1536-well format. viscosity of the solvent, which usually remain constant
In both LANCETM and HTRF®, measurement of the ratio during an assay.
of donor and acceptor fluorophore emission can be The method requires a significant difference in size
applied to compensate for non-specific quenching of assay between labelled ligand and target, which is a major
reagents. As a result, the applications of both technologies restriction to its application (Nosjean et al., 2006) and the
are widespread, covering detection of kinase enzyme activ- reliance on a single, non-time resolved fluorescence output
ity (Jia et al., 2006), protease activity (Karvinen et al., makes the choice of fluorphore important to minimize
2002), second messengers such as cAMP and inositiol tri- compound interference effects (Turek-Etienne et al.,
phosphate (InsP3) (Titus et al., 2008; Trinquet et al., 2003). FP-based assays can be used in 96-well up to 1536-
2006) and numerous biomarkers such as interleukin 1β well formats.
(IL-1β) and tumour necrosis factor alpha (TNFα) (Achard
et al., 2003). Fluorescence correlation methods
Although an uncommon technique in most HTS depart-
Fluorescence polarization (FP) ments, due the requirement for specific and dedicated
When a stationary molecule is excited with plane-polarized instrumentation, this group of fluorescence technologies
light it will fluoresce in the same plane. If it is tumbling provide highly sensitive metrics using very low levels of
rapidly, in free solution, so that it changes its orientation detection reagents and are very amendable to ultra-high
between excitation and emission, the emission signal will throughput screening (uHTS) (Eggeling et al., 2003). The
be depolarized. Binding to a larger molecule reduces the most widely applied readout technology, fluorescence cor-
mobility of the fluorophore so that the emission signal relation spectroscopy, allows molecular interactions to be
remains polarized, and so the ratio of polarized to studied at the single-molecule level in real time. Other

104
 High-throughput screening Chapter |8|

proprietary technologies such as fluorescence intensity dis-

tribution analysis (FIDA and 2-dimensional FIDA (Kask Excitation AlphaScreen©
et al., 2000) also fall into this grouping, sharing the 680 nm Emission
common theme of the analysis of biomolecules at
1O
2
520–620nm AlphaLISA©
extremely low concentrations. In contrast to other fluores- Emission
615nm
cence techniques, the parameter of interest is not the emis-
sion intensity itself, but rather intensity fluctuations. By
confining measurements to a very small detection volume A B
(achieved by the use of confocal optics) and low reagent
concentrations, the number of molecules monitored is
kept small and the statistical fluctuations of the number
contributing to the fluorescence signal at any instant
become measurable. Analysis of the frequency compo- Alpha donor bead Alpha acceptor bead
nents of such fluctuations can be used to obtain informa-
tion about the kinetics of binding reactions. Fig. 8.9  Principle of AlphaScreen™ assays.
With help of the confocal microscopy technique and Reproduced with kind permission of Perkin Elmer.
laser technologies, it has become possible to measure
molecular interactions at the single molecule level. Single
molecule detection (SMD) technologies provide a number
of advantages: significant reduction of signal-to-noise ratio, when excited by light at a wavelength of 680 nm, releases
high sensitivity and time-resolution. Furthermore, they singlet oxygen that is absorbed by an acceptor bead, and
enable the simultaneous readout of various fluorescence assuming it is in sufficiently close proximity (<200 nm)
parameters at the molecular level. SMD readouts include this results in the emission of light between 520 and
fluorescence intensity, translational diffusion (fluorescence 620 nm (Figure 8.9). This phenomenon is unusual in that
correlation spectroscopy, FCS), rotational motion (fluores- the wavlength of the emitted light is shorter and therefore
cence polarization), fluorescence resonance energy trans- has higher energy than the excitation wavelength. This is
fer, and time-resolved fluorescence. SMD technologies of significance since it reduces the potential for compound
are ideal for miniaturization and have become amenable inner filter effects; however, reactive functionality may still
to automation (Moore et al., 1999). Further advantages inhibit the energy transfer.
include very low reagent consumption and broad applica- As with other bead-based technologies the donor and
bility to a variety of biochemical and cell-based assays. acceptor beads are available with a range of surface treat-
Single molecular events are analysed by means of confo- ments to enable the immobilization or capture of a range
cal optics with a detection volume of approximately 1 fL, of analytes. The range of immobolization formats and the
allowing miniaturization of HTS assays to 1 µL or below. distance over which the singlet oxygen can pass to excite
The probability is that, at any given time, the detection the donor bead provide a suitable format for developing
volume will have a finite number of molecular events homogeneous antibody-based assays similar to enzyme-
(movement, intensity, change in anisotropy), which can linked immunosorbent assays (ELISA) which are generally
be measured and computed. The signal-to-noise ratio typi- avoided in the HTS setting due to multiple wash, addition
cally achieved by these methods is high, while interference and incubation steps. These bead-based ELISA, such as
from scattered laser light and background fluorescence are AlphaLISA™ (Perkin Elmer), provide the required sensitiv-
largely eliminated (Eigen and Rigler, 1994). ity for detection of biomarkers in low concentration and
Fluorescence lifetime analysis (Moger et al., 2006) is a can be configured to low volume 384-well format without
relatively straightforward assay methodology that over- loss of signal window.
comes many of the potential compound intereference
effects achieved through the use of TRF, but without the Cell-based assays
requirement for expensive fluorophores. The technique
utilizes the intrinsic lifetime of a fluorophore, correspond- Readouts for cell-based assays
ing to the time the molecule spends in the excited state. This Readouts that can be used for cell-based assays are many
time is altered upon binding of the fluorophore to a com- and varied. In some cases, such as radioligand binding or
pound or protein and can be measured to develop robust enzyme activity, the readouts are essentially the same as
assays that are liable to minimum compound intereference those described above. Here we describe five cell-based
using appropriate detection instrumentation. readout technologies that have found general application
in many types of assay, namely fluorometric methods, reporter
AlphaScreen™ Technology gene assays, yeast complementation assays, high-throughput
The proprietary bead-based technology from Perkin Elmer electrophysiology assays and more recently label free detection
is a proximity-based format utilizing a donor bead which, platforms. Some informative case histories of cell-based

105
Section | 2 | Drug Discovery

assays based on different readout principles have been simultaneous application of reagents and test compounds
presented by Johnston and Johnston (2002). to multiwell plates and the capture of the fluorescence
signal from each well was a key advance in allowing cel-
Fluorometric assays lular assays to be utilized in the HTS arena. Early instru-
Fluorometric assays are widely used to monitor changes in ments employed an argon laser to deliver the excitation
the intracellular concentration of ions or other constitu- light source with the emission measured using a CCD
ents such as cAMP. A range of fluorescent dyes has been imaging device. In more recent models the laser has been
developed which have the property of forming reversible replaced with an LED light source (www.moleculardevices.
complexes with ions such as Ca2+ or Tl+ (as a surrogate for com) and overcomes some of the logistical considerations
K+). Their fluorescent emission intensity changes when the for deploying these instruments in some laboratories.
complex is formed, thereby allowing changes in the free Repeated measurements can be made at intervals of less
intracellular ion concentration to be monitored, for than 1 s, to determine the kinetics of the cellular response,
example in response to activation or block of membrane such as changes in [Ca2+]i or membrane potential, which
receptors or ion channels, Other membrane-bound dyes are often short-lasting, so that monitoring the time profile
are available whose fluorescence signal varies according to rather than taking a single snapshot measurement is
the cytoplasmic or mitochondrial membrane potential. essential.
Membrane-impermeable dyes which bind to intracellular
structures can be used to monitor cell death, as only dying Reporter gene assays
cells with leaky membranes are stained. In addition to Gene expression in transfected eukaryotic cells can be
dyes, ion-sensitive proteins such as the jellyfish photo- quantified by linking a promoter sequence to a reporter
protein aequorin (see below), which emits a strong fluores- gene, whose level of expression is readily monitored, and
cent signal when complexed with Ca2+, can also be used reflects the degree of activation or inhibition of the pro-
to monitor changes in [Ca2+]i. Cell lines can be engineered moter (Naylor, 1999). Compounds activating or inhibit-
to express this protein, or it can be introduced by electro- ing the promoter itself, or interfering with a signal pathway
poration. Such methods find many applications in cell connected to that promoter, can thus be detected. By using
biology, particularly when coupled with confocal micro­ two different reporter constructs e.g. firefly and Renilla
scopy to achieve a high level of spatial resolution. For luciferase, different targets can be screened simultaneously
HTS applications, the development of the Fluorescence (Kent et al., 2005). The principle of a reporter gene assay
Imaging Plate Reader (FLIPR™, Molecular Devices Inc., for GPCR activity, based on luciferase, is shown in Figure
described by Schroeder and Negate, 1996), allowing the 8.10. Reporter readouts can also be duplexed with more

Extracellular
stimuli

Cytoplasm
Signalling
pathway
Nucleus

Transcription
factor

TRE Reporter gene

Fig. 8.10  Reporter gene assay principle for the detection of a ligand mediated signalling event in a cell based assay. Upon
binding of a small molecule to the receptor the signalling cascade is initiated, resulting in the binding of a signalling mediator
to a specific transcription factor response element, controlling the expression of the reporter protein.

106
 High-throughput screening Chapter |8|

immediate readouts of cell signalling, such as calcium High throughput electrophysiology assays
sensitive dyes, to reduce the false positive liability associ-
The progression of ion channels, and in particular voltage-
ated with using a single assay readout (Hanson, 2006)
gated ion channels, as druggable targets using screening
Commonly used reporter genes are CAT (chlorampheni-
approaches was, until recently, severely limited by the
col acetyltransferase), GAL (β-galactosidase), LAC (β-
throughput of conventional electrophysiology techniques
lactamase) (Zlokarnik et al., 1998), LUC (luciferase, Kolb
and lack of suitable higher throughput assay platforms.
and Neumann, 1996) and GFP (green fluorescence
Although fluorescence methods using membrane poten-
protein, Kain, 1999), usually employing a colorimetric or
tial sensitive dyes such as DiBAC4(3) and the FLIPR™ vari-
fluorescent readout and each having relative merits (Suto
ants of this and the FRET based voltage sensor probes
and Ignar, 1997). The number of reporter genes is dwarfed
(Gonzalez and Maher, 2002) were widely used, the meth-
compared to the range of promoters that can be employed
odology could not provide accurate voltage control and
in this format, covering a diversity of signalling events.
the temporal resolution of the evoked responses was poor.
Whilst having been widely deployed in the drug discov-
The introduction of planar patch-clamp instruments, par-
ery process there are several limitations of reporter gene
ticularly systems such as IonWorks Quattro which record
technology, not least because of the measurement of a
using multihole planar substrate consumables (Finkel
response distal to the ligand interaction and the longer
et al., 2006, Southan and Clark, 2009) has to a certain
compound incubation times required, increasing the
extent overcome the throughput hurdle. The operating
potential for cytotoxic events (Hill et al., 2001).
principle of this instrument is shown in Figure 8.11 and
Yeast complementation assay whilst the data point generation is not as high throughput
so as to compete with fluorescence methods (a maximum
Yeast is a well-characterized organism for investigating
of approximately 3000 data points per day per instrument
mammalian systems, and the yeast two-hybrid assay is a
compared with > 20 000 per day for a FLIPR™) it is suffi-
powerful genetic screening technique for measuring the
cient for screening of targeted libraries, diverse compound
protein–protein and protein–DNA interactions that under-
decks up to around 100 000 compounds and the confirma-
lie many cellular control mechanisms (Tucker, 2002 and
tion of a large number of hits identified in less physiologi-
reviewed in Brückner et al., 2009). Widely applied in cell
cally relevant platforms.
and systems biology to study the binding of transcription
factors at the sequence level, it can also be used to screen
small molecules for their interference with specific protein–
protein and protein–DNA interactions, and has recently
been adapted for other types of drug–target interactions
(Fields and Song, 1989; Young et al., 1998; Serebriiskii
et al., 2001). Conventional in vitro measurements, such as Electrode
immunoprecipitation or chromatographic co-precipitation
(Regnier, 1987; Phizicky and Fields, 1995), require the
interacting proteins in pure form and at high concentra- Extra-cellular
tions, and therefore are often of limited use.
The yeast two-hybrid system uses two separated peptide
domains of transcription factors: a DNA-specific binding
Cell
part (DNB) and a transcription activation domain (AD).
The DNB moiety is coupled to one protein (the ‘bait’),
and the AD moiety to another (the ‘prey’). If the prey
protein binds to the bait protein, the AD moiety is Small pore High-resistance seal
brought into close association with the reporter gene, Antibiotic Intra-cellular
which is thereby activated, producing a product (e.g. GAL
or LAC, as described above, or an enzyme which allows
the yeast to grow in the presence of cycloheximide). The Low-resistance
addition of a test compound that blocks the specific Common pathway
protein–protein interaction prevents activation of the ground
reporter gene. Serebriiskii et al. (2001) describe a project electrode
in which lead compounds able to block the activation of
a specific N-type voltage-gated Ca2+ channel have been
identified with a yeast two-hybrid assay. The bait and prey
proteins contained domains of two different channel Fig. 8.11  Planar patch clamp, the underlying principle of
subunits which need to associate to form a functional high throughput electrophysiology.
channel. Reproduced with kind permission of Molecular Devices.

107
Section | 2 | Drug Discovery

Label free detection platforms at a preselected spatial resolution. The spatial resolution
is largely defined by the instrument specification and
The current drive for the drug discovery process is to move
whether it is optical confocal or widefield. Confocal
towards as physiologically relevant systems as possible and
imaging enables the generation of high-resolution images
away from target overexpression in heterologous expres-
by sampling from a thin cellular section and rejection of
sion systems and, in the case of G-protein coupled recep-
out of focus light; thus giving rise to improved signal : noise
tors, to avoid the use of promiscuous G-proteins where
compared to the more commonly applied epi-fluorescence
possible. The downside to this is that endogenous receptor
microscopy. There is a powerful advantage in confocal
expression levels tend to be lower and therefore more
imaging for applications where subcellular localization or
sensitive detection methods are required. Also, for the
membrane translocation needs to be measured. However,
study of targets where the signalling mechanism is
for many biological assays, confocal imaging is not ideal
unknown, e.g. orphan GPCRs, multiple assay systems
e.g. where there are phototoxicity issues or the applica-
would need to be developed which would be time con-
tions have a need for a larger focal depth.
suming and costly. Consequently, the application of assay
HCS relies heavily on powerful image pattern recogni-
platforms which detect gross cellular responses, usually
tion software in order to provide rapid, automated and
cell morphology due to actin cytoskeleton remodelling,
unbiased assessment of experiments.
to physiological stimuli have been developed. These fall
The concept of gathering all the necessary information
into two broad categories, those that detect changes
about a compound at one go has obvious attractions, but
in impedance through cellular dielectric spectroscopy
the very sophisticated instrumentation and software
(Ciambrone et al., 2004), e.g. CellKey™, Molecular Devices
produce problems of reliability. Furthermore, the principle
Corporation; xCelligence, Roche Diagnostics or the use of
of ‘measure everything and sort it out afterwards’ has its
optical biosensors (Fang, 2006), e.g. Epic™, Corning Inc;
drawbacks: interpretation of such complex datasets often
or Octect, Fortebio. The application of these platforms in
requires complex algorithms and significant data storage
the HTS arena is still in its infancy largely limited by
capacity. Whilst the complexity of the analysis may seem
throughput and relatively high cost per data point com-
daunting, high content screening allows the study of
pared to established methods. However, the assay develop-
complex signalling events and the use of phenotypic rea-
ment time is quite short, a single assay may cover a broad
douts in highly disease relevant systems. However, such
spectrum of cell signalling events and these methods are
analysis is not feasible for large number of compounds
considered to be more sensitive than many existing
and unless the technology is the only option for screening
methods enabling the use of endogenous receptor expres-
in most instances HCS is utilized for more detailed study
sion and even the use of primary cells in many instances
of lead compounds once they have been identified (Haney
(Fang et al., 2007; Minor, 2008).
et al., 2006).
High content screening
Biophysical methods in
High content screening (HCS) is a further development of
cell-based screening in which multiple fluorescence read-
high-throughput screening
outs are measured simultaneously in intact cells by means Conventional bioassay-based screening remains a main-
of imaging techniques. Repetitive scanning provides tem- stream approach for lead discovery. However, during
porally and spatially resolved visualization of cellular recent years alternative biophysical methods such as
events. HCS is suitable for monitoring such events as nuclear magnetic resonance (NMR) (Hajduk and Burns,
nuclear translocation, apoptosis, GPCR activation, recep- 2002), surface plasmon resonance (SPR) (Gopinath, 2010)
tor internalization, changes in [Ca2+]i, nitric oxide produc- and X-ray crystallography (Carr and Jhoti, 2002) have been
tion, apoptosis, gene expression, neurite outgrowth and developed and/or adapted for drug discovery. Usually in
cell viability (Giuliano et al., 1997). assays whose main purpose is the detection of low-affinity
The aim is to quantify and correlate drug effects on cel- low-molecular-weight compounds in a different approach
lular events or targets by simultaneously measuring mul- to high-throughput screening, namely fragment-based
tiple signals from the same cell population, yielding data screening. Hits from HTS usually already have drug-like
with a higher content of biological information than is properties, e.g. a molecular weight of ~ 300 Da. During the
provided by single-target screens (Liptrot, 2001). following lead optimization synthesis programme an
Current instrumentation is based on automated digital increase in molecular weight is very likely, leading to
microscopy and flow cytometry in combination with hard poorer drug-like properties with respect to solubility,
and software systems for the analysis of data. Within the absorption or clearance. Therefore, it may be more effec-
configuration a fluorescence-based laser scanning plate tive to screen small sets of molecular fragments (<10 000)
reader (96, 384- or 1536-well format), able to detect fluo- of lower molecular weight (100–250 Da) which can then
rescent structures against a less fluorescent background, be chemically linked to generate high-affinity drug-like
acquires multicolour fluorescence image datasets of cells compounds. Typically, such fragments have much weaker

108
 High-throughput screening Chapter |8|

binding affinities than drug-like compounds and are Beyond the 96-, 384- and 1536-well arena there remains
outside the sensitivity range of a conventional HTS assay. a drive for further increased density to 3456-well format
NMR-, SPR- or X-ray crystallography-based assays are (Kornienko et al., 2004) and micro-fluidics/lab-on-a-chip
better suited for the identification of weak binders as these (Pihl et al., 2005) approaches to further reduce reagent
methodologies lend themselves well to the area of frag- usage.
ment based screening. As the compound libraries screened Regardless of the microplate format adopted by a
are generally of limited size throughput is less important screening laboratory, the advances in microplates, liquid
than sensitive detection of low-affinity interactions. Once handling, plate stacking and handling devices, and sensi-
the biophysical interactions are determined, further X-ray tive reagents and detection instrumentation (such as
protein crystallographic studies can be undertaken to CCD imagers) have advanced to the point where execution
understand the binding mode of the fragments and this of a high throughput screen is rarely the bottleneck in
information can then be used to rapidly drive the fragment- drug discovery. Although as the density increases the time
to-hit or fragment-to-lead chemistry programme (Carr to develop a robust assay with low variability can also
et al., 2005). increase as the challenges of reagent evaporation and
As discussed in Chapter 9, the chemical linkage of weak mixing are overcome.
binding fragments can generate a high-affinity lead
without violating the restrictions in molecular weight. The
efficiency of this strategy has been demonstrated by several
Robotics in HTS
groups (Nienaber et al., 2000; Lesuisse et al., 2002). In many dedicated HTS facilities automation is employed
to varying degrees to facilitate execution of the screen. This
varies from the use of automated work stations with some
Assay formats – miniaturization manual intervention to the use of fully automated robotic
Multiwell plates began to be used for screening in the platforms. During the assay development phases the key
early 1980s, before which time tube-based assays were pieces of automation present in the automated platform
routinely used in a low-throughput mode. The introduc- will be used to ensure the assay is optimized correctly fol-
tion of 96-well plates allowed the automation and mini- lowed by transfer of the assay to a robotic workstation that
aturization of biochemical experiments and was rapidly can operate in high-throughput mode (typically up to
followed by the transition of many cell-based assays into 100 000 compounds per day at a single concentration).
the same density formats. The drive to increase throughput The robotic system consists of devices for storage, incuba-
and reduce associated reagent costs has seen great advances tion and transportation of plates in different format;
in liquid handling and detection technologies since the instruments for liquid transfer; and a series of plate readers
1990s with most vendors basing their approach on adap- for the various detection technologies. In many instances,
tation of the 96-well format to reduce the space between these devices will be replicated on the same system to
wells (the pitch) whilst maintaining an overall standard allow the screen to continue, albeit at lower throughput,
footprint and depth for the plate for the ease and con- should one device fail mid-run.
stancy of instrument design (see Society for Laboratory A typical robotic system is illustrated in Figure 8.12.
Automation and Screening, www.slas.org/education/ Robotic arms and/or automated transport systems move
microplate.cfm). Reducing the pitch by one-half yields plates between different devices. Plate storage devices
a four-fold increase in density to 384 wells per plate (‘hotels’) and incubators are used for storage and incuba-
and a further two-fold reduction gives rise to 1536-well tion of microplates. Incubators can be cooled or heated;
format plates. In both cases, the wells of these increased for mammalian cell cultivation they can also be supplied
density plates can still be addressed using standard 96-well with CO2 and are designed to facilitate automatic transfer
technology, although for liquid handling the reduced of plates in and out. Compound plates are typically sup-
volumes and well area associated with the 1536-well plied with seals that can be perforated by the liquid hand­
formats present challenges for tip-based, displacement ling devices to allow easy dilution and compound transfer
dispensing. To overcome this, the low-volume 384-well to the assay plates. The assay plates themselves may either
plate (Garyantes, 2002) has emerged as an important have removable lids or be sealed automatically once all
format, offering lower reagent usage per well whilst over- reagents have been added. Stackers are sequential storage
coming the issues of well access. In turn, the liquid- units for microtitre plates, connected to automated pipet-
handling technologies to support 1536-well plates has ting instruments and are equally common in laboratories
developed significantly through the use of fixed tips or where HTS is conducted without the large scale applica-
non-contact dispensing using piezo-electric dispensers or tion of automation as they allow the scientists to walk
acoustic dispensing. These latter formats do not rely on away and return when the pipetting steps are complete.
tip-based technology and can dispense volumes as low Various detection devices (enabling different modes of
as 2.5 nanolitres (Dunn and Feygin, 2000; Ellson et al., detection) are located at the output of the system before
2003). the assay plate, and usually the compound plate as well,

109
Section | 2 | Drug Discovery

Liquid handling
Detection devices devices

Incubators

Temperature
controlled
storage

Bulk reagent
dispensers/
dilutors Plate sealer

Plate conveyor

Fig. 8.12  Typical layout of fully automated high throughput screening platform. (A) Computer-aided design of the automated
HTS work station and (B) photograph of the final installation.
Images supplied courtesy of the RTS Group, a leading supplier or laboratory automation, and Novartis.

110
 High-throughput screening Chapter |8|

are automatically discarded to waste. Central to the plat- ‘end point’ read rather than more time-consuming multi-
form is the control (scheduling) software which controls ple or kinetic readings. However, the instrument itself may
the overall process including correct timing of different perform some initial calculations and these heterogeneous
steps during the assay. This is critical, and ensures types of raw data are automatically transferred into the
co-ordination of the use of the different devices (pipetters, data management software. Assay plates are typically iden-
incubators, readers etc.) to produce maximum efficiency. tified by a unique bar-code to relate the data to the com-
This software will also control recovery and/or continued pound plate layout. Ideally the plate reader will have an
operation of the platform, in the event of an error, without integral bar-code reader and the data file will be automati-
manual intervention. As one can imagine, programming cally named with the bar-code to provide an error-free
and testing of the different process steps for individual association of the correct data file with the compounds
screens can be a time-consuming part of the operation and tested.
frequently dedicated automation teams exist in large HTS In a next step raw data are translated into contextual
departments to expedite this. information by calculating results. Data on percentage
Before primary screening can start sample plates have to inhibition or percentage of control are normalized with
be prepared, which is usually done offline by separate values obtained from the high and low controls present in
automated liquid transfer systems (384-tip pipettes or each plate. In secondary screening IC50/EC50 and Ki values
acoustic dispensing devices). Compound storage plates, are also calculated. The values obtained depend on the
containing the library to be screened prepared as DMSO method used (e.g. the fitting algorithm used for
solutions, are delivered from the compound library ware- concentration–response curves) and have to be standard-
house and samples are further diluted with aqueous buffer ized for all screens within a company. Once the system
to reach the desired compound and DMSO concentration captures the data it is then necessary to apply validation
for the assay. Samples are usually transferred to the assay rules and techniques, such as trimmed means, to eliminate
plates by the robot during the assay run. outliers (ideally using an automated algorithm) and to
During the screening itself, all processes have to be apply predetermined acceptance criteria to the data, for
monitored online to ensure the quality of the data example, the signal-to-noise ratio, the Z’-value, or a test
obtained. The performance of the assay is continuously for gaussian distribution of the data. All plates that fail
measured by calculating Z’ values for each plate (see earlier against one or more quality criteria are flagged and
section). For this purpose, each screening plate includes discarded.
high and low controls for quality analysis, in addition to The process may also involve a step to monitor visually
the compounds for screening. the data that have been flagged, as a final check on quality.
For the selection of positives a variety of methods may This is to ensure the system has performed correctly, i.e.
be applied based upon the control wells present on the no missed reagent dispense or patterns indicative of
assay plate. This hit threshold may be an arbitrary activity blocked dispenser tips or edge effects which may lead to
cut-off set across the screen or statistically based on a false positives or negatives (Gunter et al., 2003). Whilst
plate-by-plate or screen-wide basis and is then usually set this inspection may be performed manually there are a
at least three standard deviations away from the mean of number of software packages available, e.g. GeneData
the library signal (Brideau et al., 2003). Assay Analyzer (www.genedata.com), which flag such
errors and, in the case of edge effects, apply mathematical
Data analysis and management corrections to overcome them. Examples of validation
data obtained in a typical screening of 100,000 com-
Owing to the large volume of data generated in HTS, effi-
pounds and some common data patterns revealed by
cient data management is essential. Software packages for
tracking such parameters are shown in Figure 8.13. In
HTS (e.g. ActivityBase, Spotfire, GeneData) are available to
addition to tracking high and low control values, most
carry out the principal tasks:
HTS departments also include quality-control plates at
• Storage of raw data regular intervals throughout an assay run which contain
• Association of raw data with compound information standard compounds to allow target pharmacology to be
• Quality control monitored.
• Transformation of data into information On completion of the screen it is advisable to plot a
• Visualization distribution histogram of the compound data against the
• Documentation control well populations and any other quality-control
• Reporting. wells that have been included (Figure 8.14). The median
In HTS each biochemical experiment in a single well is of the test wells should be the same as the null control
analysed by an automated device, typically a plate reader population and, if the assay is robust as demonstrated
or other kind of detector. The output of these instruments by the Z’ value, there will be good separation between
comes in different formats depending on the type of the two control well populations. The variability of the
reader. Where possible the HTS favours the use of a single null control population can be used to determine an

111
Section | 2 | Drug Discovery

120 120
Non-corrected percentage activity value

Corrected percentage activity value

100 100

80 80

60 60

40 40

20 20

0 0

–20 –20

A Control wells in assay plate run order B Control wells in assay plate run order
1.0 6
0.9
0.8 5

Signal/background
0.7
4
0.6
Z’ value

0.5 3
0.4
0.3 2
0.2
1
0.1
0.0 0
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350

C Plate number D Plate number

Fig. 8.13  Data validation checks in a typical screening assay. Distribution of high (green symbols) and low controls (red
symbols) (% inhibition) in a set of screening plates. Each marker represents one well. Panel (A) shows the original data and (B)
is the same data following treatment with GeneData Assay Analyzer to take account of spatially induced effects. Note the
reduction in variability of the low control well data. (C) Z’ values and (D) signal to background ratios from an HTS campaign.
The plate run order is shown on the x-axis and vertical dashed line divide each screening run, the horizontal lines at Z’ = 0.5
and signal-to-background = 1.5 represents the quality control pass levels and each data point represents on assay plate. The
data demonstrate a decline in assay performance during each screening day. The marked change in rate of assay performance
deterioration after plate 150 correlates with a change in the batch of a key assay component.
Data kindly supplied by BioFocus, with permission.

acceptable cut-off level for selecting the hit compounds for

further progression. SCREENING LIBRARIES AND
In addition to registering the test data, all relevant infor- COMPOUND LOGISTICS
mation about the assay has to be logged, for example the
supplier and batch of reagents, storage conditions, a
Compound logistics
detailed assay protocol, plate layout, and algorithms for the
calculation of results. Each assay run is registered and its In the drug discovery value chain the effective manage-
performance documented. If assay performance is charted ment of compound libraries is a key element and is usually
any significant change in assay quality can be reviewed and handled by a dedicated compound logistics group, either
if caused by reagent change readily identified within the screening organization or at a specialist out-
HTS will initially deliver hits in targeted assays. Retrieval sourcing provider. The compound management facility is
of these data has to be simple, and the data must be the central port for transshipment of compounds in lead
exchangeable between different project teams to generate discovery, not only for primary screening, but also during
knowledge from the mass of data. the hit-to-lead phase. It is the unit responsible for

112
 High-throughput screening Chapter |8|

Test compounds High controls Low controls Reference inhibitor

1000

800
1.3% HR at 30% I
Number of compounds

600 0.95% HR at 35% I

0.7% HR at 40% I
400

0.5% HR at 50% I
200

0
–30 –20 –10 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
% Inhibition

Fig. 8.14  Histogram analysis of compound activity against control well populations from a screen to identify antagonists of a
G-protein coupled receptor. The median of the compound wells is in line with the null control population. The hit rate (% HR)
associated with selecting each of the percentage inhibition value (% I) is as indicated. As the percentage inhibition is lowered
the confidence in the reality of the hit is reduced; however, these compounds can provide useful information on the hit series
being identified. Note the reference compound controls appear to fall into two populations. This is associated with the change
in reagent batch as described in Figure 8.13.
Data kindly supplied by BioFocus, with permission.

registration of samples, their preparation, storage and storage conditions. Different sets of compound libraries
retrieval. This facility has to ensure global compound are needed, depending on the target and project specifica-
accessibility, both individually and in different formats; tions with the preferred storage format of the compound
maintain compound integrity (quality control and proper collection being dictated by the chosen screening plate
storage); guarantee error-free compound handling; guar- density, i.e. 384- or 1536-well.
antee efficient compound use; and guarantee a rapid Because library sets are not static, as new compounds
response to compound requests. are continuously being added, samples in the repository
Many pharmaceutical companies have greatly increased need to be individually addressable to allow a flexible and
both the size and quality of their compound collection. quick rearrangement of existing libraries to more specific,
Screening libraries frequently exceed 1,000,000 com- focused collections. Advanced compound logistic systems
pounds and originate from many different sources with store compound libraries in a single tube system so that
variable quality, although there has been great attention individual tubes can be accessed without the need to take
to how the assays themselves are impacted by poor com- out a whole plate from the storage facility.
pound quality, particularly solubility (Di and Kerns, This functionality is a prerequisite for efficient ‘cherry
2006). This has necessitated both hardware and software picking’. After primary screening, positive compounds have
automation of compound management in order to cope to be confirmed in secondary assays and concentration–
with the increasing demands of HTS and lead discovery. response curves determined. The individual active com-
Most advanced systems use fully automated robotics for pounds have to be located and reformatted in microtitre
handling compounds (Figure 8.15). Compounds used for plates. With a large number of targets and projects running
screening are stored as liquids in microtitre plates in a at any one time, a highly automated compound handling
controlled environment (temperature −4°C to −20°C; low systems is needed to do this efficiently.
humidity or storage under an inert atmosphere and con­
trol of freeze-thaw cycles) and numerous studies have been
undertaken to identify the most appropriate conditions to
maintain compound stability (e.g. Blaxill et al., 2009). PROFILING
Splitting the collection into a number of copies in dif-
ferent formats secures a balance between fast response High-throughput screening, the subject of this chapter, has
times to the various compound requests and optimal as its first objective the identification of a few ‘validated

113
Section | 2 | Drug Discovery

The technology involved in miniaturization, automa-

tion and assay readouts required for HTS has developed
rapidly and continues to do so. As this technology evolves,
the laboratory set-ups installed in HTS facilities are stead-
ily broadening their capabilities beyond their primary
function of identifying hits to apply HTS techniques to
more diverse compound profiling assays relating not only
to the target selectivity of compound libraries, but also to
their pharmacokinetic characteristics. Increasingly, there-
fore, early compound profiling tasks on ‘hit’ compounds
are being carried out in HTS laboratories where the neces-
sary technological expertise is concentrated. Such assays
are also very helpful in the ‘lead identification’ stage of a
project, where focused synthetic compound libraries based
on the initial hits need to be assessed. As this work gener-
ally involves testing small compound libraries, usually
fewer than 1000 compounds at a time, in several different
assays, small dedicated robotic workstations are needed,
rather than the fast but inflexible factory-style robotic
assemblies used for large-scale HTS.
In vitro pharmacokinetic assays (see Chapter 10), which
are not generally project specific and can be automated to
run in medium-throughput fashion, are very suitable for
running in this environment. This extension of the work
of HTS laboratories beyond the primary task of finding
hits is a clear and continuing trend, for which the term
‘high-throughput profiling’ (HTP) has been coined. It
brings the work of HTS laboratories into a close and
Fig. 8.15  Storage of a compound screening collection in healthy relationship with drug discovery teams. The highly
384-well plates at −4°C. A robot is able to move around and disciplined approach to assay formats and data logging
collect the plates or individual samples specified in the
that is essential for HTS, but not second nature to many
compound management software.
laboratory scientists, brings the advantage that profiling
data collected over a wide range of projects and drug
targets is logged in standard database formats, and is there-
hits’ (defined in Chapter 9) within large compound librar- fore a valuable company-wide tool for analysing structure–
ies. The decision as to whether a particular hit is worth activity relationships. This necessity to handle and visualize
pursuing as a chemical lead in a drug discovery project such data has driven the development software packages
depends on several factors, important ones being its chem- such as Spotfire (spotfire.tibco.com).
ical characteristics and its pharmacodynamic and pharma- In summary, it is clear that pharmacological profiling
cokinetic properties. These aspects, broadly covered by the will be an increasing activity of HTS units in the future,
term ‘compound profiling’, are discussed in detail in the and will help to add further value in the drug discovery
next three chapters. chain.

REFERENCES

Achard S, Jean A, Lorphelin D, et al. Angewandte Chemie Beske O, Goldbard S. High-throughout

Homogeneous assays allow direct ‘in 1987;99:259–61. cell analysis using multiplexed array
well’ cytokine level quantification. Bays N, Hill A, Kariv I. A simplified technologies. Drug Discovery Today
Assay and Drug Development scintillation proximity assay for 2002;7:S131–5.
Technologies 2003;1(supplement 2): fatty acid synthase activity: Beveridge M, Park YW, Hermes J, et al.
181–5. development and comparison Detection of p56(lck) kinase activity
Alpha B, Lehn JM, Mathis G. Energy- with other FAS activity assays. using scintillation proximity assay in
transfer luminescence of Journal of Biomolecular 384-well format and imaging
europium(III) and terbium(III) with Screening 2009;14: proximity assay in 384- and
macrobicyclic polypyridine ligands. 636–42. 1536-well format. Journal of

114
 High-throughput screening Chapter |8|

Biomolecular Screening Comley J. TR-FRET based assays Journal of Biomolecular Screening

2000;5:205–12. – getting better with age. Drug 2006;11:488–96.
Birzin ET, Rohrer SP. High-throughput Discovery World Spring 2006;22–37. Fligge T, Schuler A. Integration of a
receptor-binding methods for Congreve M, Chessari G, Tisi D, et al. rapid automated solubility
somatostatin receptor 2. Analytical Recent developments in fragment- classification into early validation of
Biochemistry 2002;307:159–66. based drug discovery. Journal of hits obtained by high throughput
Blaxill Z, Holland-Crimmen S, Lifely R. Medicinal Chemistry 2008;51: screening. Journal of Pharmaceutical
Stability through the ages: the GSK 3661–80. and Biomedical Analysis
experience. Journal of Biomolecular Cook ND. Scintillation proximity assay: 2006;42:449–54.
Screening 2009;14:547–56. a versatile high-throughput screening Garyantes TK. 1536-well assay
Bosworth N, Towers P. Scintillation technology. Drug Discovery Today plates: when do they make sense?
proximity assay. Nature 1996;1:287–94. Drug Discovery Today 2002;7:
1989;341:167–8. Di L, Kerns E. Biological assay 489–90.
Braunwaler AF, Yarwood DR, Hall T, challenges from compound Ghosh S, Nie A, An J, et al. Structure-
et al. A solid-phase assay for the solubility: strategies for bioassay based virtual screening of chemical
determination of protein tyrosine optimization. Drug Discovery Today libraries for drug discovery. Current
kinase activity of c-src using 2006;11:446–51. Opinion in Chemical Biology
scintillating microtitration plates. Dunn D, Feygin I. Challenges and 2006;10:194–202.
Analytical Biochemistry solutions to ultra-high-throughput Giuliano KA, de Basio RL, Dunlay RT,
1996;234:23–6. screening assay miniaturization: et al. High-content screening: a new
Brideau C, Gunter B, Pikounis B, et al. submicroliter fluid handling. Drug approach to easing key bottlenecks
Improved statistical methods for hit Discovery Today 2000;5(Suppl. 12): in the drug recovery process. Journal
selection in high-throughput S84–91. of Biomolecular Screening
screening. Journal of Biomolecular Eggeling C, Brand L, Ullmann D, et al. 1997;2:249–59.
Screening 2003;8:634–47. Highly sensitive fluorescence Gonzalez J, Maher M. Cellular
Brown BA, Cain M, Broadbent J, et al. detection technology currently fluorescent indicators and voltage/
FlashPlate technology. In: Devlin PJ, available for HTS. Drug Discovery ion probe reader (VIPR(tm)): tools
editor. High-throughput screening. Today 2003;8:632–41. for ion channel and receptor drug
New York: Marcel Dekker; 1997. Ehrlich P, Bertheim A. Berichte. discovery. Receptors and Channels
p. 317–28. 1912;4:756. 2002;8:283–95.
Brückner A, Polge C, Lentze N, et al. Eigen M, Rigler R. Sorting single Gopinath S. Biosensing applications of
Yeast two-hybrid, a powerful tool molecules: application to diagnostics surface plasmon resonance-based
for systems biology. International and evolutionary biotechnology. Biacore technology. Sensors and
Journal of Molecular Science Proceedings of the National Actuators B 2010;150:722–33.
2009;10:2763–88. Academy of Sciences of the USA Gribbon P, Sewing A. Fluorescence
Carr R, Congreve M, Murray C, et al. 1994;91:5740–7. readouts in HTS: no gain without
Fragment-based lead discovery: leads Ellson R, Mutz M, Browning B, et al. pain? Drug Discovery Today
by design. Drug Discovery Today Transfer of low nanoliter acoustics 2003;8:1035–43.
2005;10:987–92. – automation considerations. Journal Gunter B, Brideau C, Pikounis B, et al.
Carr R, Jhoti H. Structure-based of the Association for Laboratory Statistical and graphical methods
screening of low-affinity compounds. Automation 2003;8:29–34. for quality control determination
Drug Discovery Today 2002;7: Fang Y. Label-free cell-based assays with of high-throughput screening
522–7. optical biosensors in drug discovery. data. Journal of Biomolecular
Ciambrone G, Liu V, Lin D, et al. Assay and Drug Development Screening 2003;8:624–33.
Cellular dielectric spectroscopy: a Technologies 2006;4:583–95. Hajduk PJ, Burns DJ. Integration of
powerful new approach to label-free Fang Y, Guangshan L, Ferrie A. NMR and high-throughput
cellular analysis. Journal of Non-invasive optical biosensor for screening. Combinatorial Chemistry
Biomolecular Screening assaying endogenous G protein- High Throughput Screening
2004;9:467–80. coupled receptors in adherent cells. 2002;5:613–21.
Clegg RM. Fluorescence resonance Journal of Pharmacological and Hallis T, Kopp A, Gibson J, et al. An
energy transfer. Current Opinion Toxicological Methods 2007; improved beta-lactamase reporter
in Biotechnology 1995;6: 55:314–22. assay: multiplexing with a
103–10. Fields S, Song O. A novel genetic system cytotoxicity readout for enhanced
Clemons P. Complex phenotypic assays to detect protein–protein accuracy of hit identification. Journal
in high-throughput screening. interactions. Nature of Biomolecular Screening
Current Opinions in Chemical 1989;340:245–6. 2007;12:635–44.
Biology 2004;8:334–8. Finkel A, Wittel A, Yang N, et al. Haney S, LaPan P, Pan J, et al. High-
Comley J. Assay interference a limiting Population patch clamp improves content screening moves to the front
factor in HTS? Drug Discovery World data consistency and success rates in of the line. Drug Discovery Today
Summer 2003;91. the measurement of ionic currents. 2006;11:889–94.

115
Section | 2 | Drug Discovery

Hanson B. Multiplexing Fluo-4 NW and Kornienko O, Lacson R, Kunapuli P, Combinatorial Chemistry High
a GeneBLAzer® transcriptional assay et al. Miniaturization of whole live Throughput Screening
for high-thoughput screening of cell-based GPCR assays using 1999;2:177–90.
G-protein-coupled receptors. Journal microdispensing and detection Naylor L. Reporter gene technology: the
of Biomolecular Screening systems. Journal of Biomolecular future looks bright. Biochemical
2006;11:644–51. Screening 2004;9:186–95. Pharmacology 1999;58:749–57.
Hartshorn M, Murray C, Cleasby A, Kroeze W, Sheffler D, Roth B. G-protein- Nienaber VL, Richardson PL, Klighofer
et al. Fragment-based lead discovery coupled receptors at a glance. V, et al. Discovering novel ligands for
using X-ray crystallography. Journal Journal of Cell Science 2003;116: macromolecules using X-ray
of Medicinal Chemistry 4867–9. crystallographic screening. Nature
2005;48:403–13. Lakowicz JR. Principles of fluorescence Biotechnology 2000;18:1105–8.
Haugland RP. Handbook of fluorescent spectroscopy. New York: Plenum Nosjean O, Souchaud S, Deniau C,
probes and chemical research, 9th Press; 1999. et al. A simple theoretical model for
edn. Molecular Probes. Web edition: Lee P, Miller S, Van Staden C, et al. fluorescence polarization binding
www.probes.com/handbook, Development of a homogeneous assay development. Journal of
2002. high-throughput live-cell G-protein- Biomolecular Screening
Hemmilä I, Hurskainen P. Novel coupled receptor binding assay. 2006;11:949–58.
detection strategies for drug Journal of Biomolecular Screening Phizicky EM, Fields S. Protein–protein
discovery. Drug Discovery Today 2008;13:748–54. interactions: methods for detection
2002;7:S150–6. Leopoldo M, Lacivita E, Berardi F, et al. and analysis. Microbiology Review
Hill S, Baker G, Rees S. Reporter-gene Developments in fluorescent probes 1995;59:94–123.
systems for the study of G-protein- for receptor research. Drug Discovery Pihl J, Karlsson M, Chiu D.
coupled receptors. Current Opinion Today 2009;14:706–12. Microfluidics technologies in drug
in Pharmacology 2001;1:526–32. Lesuisse D, Lange G, Deprez P, et al. discovery. Drug Discovery Today
Jia Y, Quinn C, Gagnon A, et al. SAR and X-ray. A new approach 2005;10:1377–83.
Homogeous time-resolved combining fragment-based screening Regnier FE. Chromatography of
fluorescence and its applications for and rational drug design: application complex protein mixtures. Journal
kinase assays in drug discovery. to the discovery of nanomolar of Chromatography 1987;418:
Analytical Biochemistry inhibitors of Src SH2. Journal of 115–43.
2006;356:273–81. Medicinal Chemistry Ramm P. Imaging systems in assay
Johnston PA, Johnston PA. Cellular 2002;45:2379–87. screening. Drug Discovery Today
platforms for HTS: three case studies. Liptrot C. High content screening 1999;4:401–10.
Drug Discovery Today – from cells to data to knowledge. Rees D, Congreve M, Murray C, et al.
2002;7:353–63. Drug Discovery Today 2001;6:832–4. Fragment-based lead discovery.
Kain RK. Green fluorescent protein Minor L. Label-free cell-based functional Natures Reviews Drug Discovery
(GFP): applications in cell-based assays. Combinatorial Chemistry and 2004;3:660–72.
assays for drug discovery. Drug High Throughput Screening Russ A, Lampel S. The druggable
Discovery Today 1999;4:304–12. 2008;11:573–80. genome: an update. Drug Discovery
Karvinen J, Hurskainen P, Moger J, Gribon P, Sewing A, et al. The Today 2005;10:1607–10.
Gopalakrishnan S, et al. application of fluorescence lifetime Schroeder KS, Negate BD. FLIPR, a new
Homogeneous time-resolved readouts in high-throughput instrument for accurate high-
fluorescence quenching assay screening. Journal of Biomolecular throughput optical screening.
(LANCE) for Caspase-3. Journal of Screening 2006;11:765–72. Journal of Biomolecular Screening
Biomolecular Screening Moore K, Rees S. Cell-based versus 1996;1:75–80.
2002;7:223–31. isolated target screening: how lucky Serebriiskii IG, Khazak V, Golemis EA.
Kask P, Palo K, Fay N, et al. Two- do you feel? Journal of Biomolecular Redefinition of the yeast two-hybrid
dimensional fluorescence intensity Screening 2001;6:69–74. system in dialogue with changing
distribution analysis: theory and Moore KJ, Turconi S, Ashman S, et al. priorities in biological research.
applications. Biophysics Journal Single molecule detection Biotechniques 2001;30:634–55.
2000;78:1703–13. technologies in miniaturized high Sittampalam GS, Kahl SD, Janzen WP.
Kent T, Thompson K, Naylor L. throughput screening: fluorescence High-throughput screening: advances
Development of a generic dual- correlation spectroscopy. Journal of in assay technologies. Current
reporter gene assay for screening Biomolecular Screening Opinion in Chemistry and Biology
G-protein-couped receptors. Journal 1999;4:335–54. 1997;1:384–91.
of Biomolecular Screening Muegge I, Oloff S. Advances in virtual Southan A, Clark G. Recent
2005;10:437–46. screening. Drug Discovery Today: technological advances in
Kolb AJ, Neumann K. Luciferase Technologies 2006;3:405–11. electrophysiology based screening
measurements in high throughput Nasir MS, Jolley ME. Fluorescence technology and the impact upon ion
screening. Journal of Biomolecular polarization: an analytical tool for channel discovery research. High
Screening 1996;1:85–8. immunoassay and drug discovery. Throughput Screening: Methods and

116
 High-throughput screening Chapter |8|

Protocols (Methods in Molecular Trinquet E, Fink M, Bazin H, et al. discovery. Drug Discovery Today
Biology) 2009;565:187–208. D-myo-Inositol 1-phosphate as a 2000;5:286–93.
Stahura F, Xue L, Godden J, et al. surrogate of D-myo-inositol 1,4,5-tris Waszkowycz B. Towards improving
Methods for compound selection phosphate to monitor G protein- compound selection in structure-
focused on hits and application in coupled receptor activation. based virtual screening. Drug
drug discovery. Journal of Molecular Analytical Biochemistry Discovery Today 2008;13:
Graphics and Modelling 2006;358:126–35. 219–26.
2002;20:439–46. Tucker C. High-throughput cell-based Young K, Lin S, Sun L, et al.
Sui Y, Wu Z. Alternative statistical assays in yeast. Drug Discovery Identification of a calcium
parameter for high-thoughput Today 2002;7:S125–30. channel modulator using a high
screening assay quality assessment. Turek-Etienne T, Small E, Soh S, et al. throughput yeast two-hybrid screen.
Journal of Biomolecular Screening Evaluation of fluorescent compound National Biotechnology 1998;16:
2007;12:229–34. interference in 4 fluorescence 946–50.
Suto C, Ignar D. Selection of an optimal polarization assays: 2 kinases, 1 Zhang JH, Chung TD, Oldenburg KR. A
reporter gene for cell-based high protease and 1 phosphatase. Journal simple statistical parameter for use
throughput screening assays. Journal of Biomolecular Screening in evaluation and validation of
of Biomolecular Screening 2003;8:176–84. high-throughput screening assays.
1997;2:7–9. Valenzano KJ, Miller W, Kravitz JN, Journal of Biomolecular Screening
Titus S, Neumann S, Zheng W, et al. et al. Development of a fluorescent 1999;4:67–73.
Quantitative high-throughput ligand-binding assay using the Zlokarnik G, Negulescu PA, Knapp TE,
screening using a live-cell cAMP AcroWell filter plate. Journal of et al. Quantitation of transcription
assay identifies small-molecule Biomolecular Screening and clonal selection of single living
agonists of the TSH receptor. Journal 2000;5:455–61. cells with beta-lactamase as reporter.
of Biomolecular Screening Valler MJ, Green D. Diversity screening Science 1998;279:84–8.
2008;13:120–7. versus focussed screening in drug

117
 Chapter 9 
The role of medicinal chemistry in the drug
discovery process
P Beswick, A Naylor

many of the preclinical animal models of human disease.

INTRODUCTION Thus there is a pressing need to improve the quality of
validation of novel targets, placing less emphasis on pre-
Clinically the unmet need for novel and safe drugs is clinical in vivo models and single gene ‘knock-outs’, and
becoming more significant due to increasing longevity in a greater emphasis on cellular pathways, genetically associ-
the developed world and the consequential increased ated targets and innovative clinical trial design. This
prevalence of age-related diseases such as cancer and approach will require the availability of selective chemical
dementia. Whilst few would dispute that the role of the tools with which to validate targets in human tissue and
medicinal chemist is pivotal to the discovery of new medi- thus the role of the medicinal chemist in the future may
cines to address this growing need, and will continue to incorporate a greater component of chemical biology. The
be so for the foreseeable future, the environment in which increasing trend towards academic target and drug discov-
the medicinal chemist operates has changed dramatically ery is also likely to place greater emphasis on chemical
over the last 5–10 years. Despite many years of rising biology skills.
investment in R&D, the productivity of major Pharma, in Despite stringent toxicity evaluation during the discov-
terms of New Chemical Entities (NCEs) delivered to the ery phase, an unacceptable level of failure due to safety
patient, has not improved over the past 10 years. In 2009 issues remains. The present situation is arguably worse
only nineteen NCEs were approved by the FDA, two less than 20 years ago as there has been a trend towards mol-
than in 2008 (Hughes, 2010) (Figure 9.1; see also Chapter ecules failing later in the safety evaluation process, incur-
22). The approval of six biological licence applications in ring greater cost and time. If significant impact is to be
2009, compared with three in 2008, is perhaps the first made on this source of attrition it will be important for
indication of the greater emphasis which Pharma is the medicinal chemist to develop a rational understanding
placing on the discovery of biological therapeutic agents. of possible structural and physicochemical determinants
However, the challenges and current limitations of such of toxicity in much the same way as the determinants of
treatments are such that small molecule discovery is likely oral bioavailability have been recognized. However, this
to remain the mainstay of therapeutic innovation for at objective is likely to be much more challenging due to the
least the next 20 years (but see also Chapter 22). diverse and complex nature of the multiple pathologies.
Whilst a detailed discussion of the reasons behind Phar- Preliminary studies to evaluate this approach have been
ma’s poor level of productivity is beyond the scope of this recently reported and will be discussed later in the chapter.
chapter (see Chapter 22), it is clear that the high level of It is clear that addressing the challenges described above
attrition in the clinic, due to lack of clinical efficacy and and raising productivity will require long-term investment
insufficient safety margins, is a major component and it is and persistence. However, due to the immediate financial
appropriate to consider what role the medicinal chemist pressures, the industry has responded to the productivity
might play in addressing this challenge in the future. The challenge by minimizing costs, shifting focus to cheaper
high level of attrition in the clinic due to lack of sufficient sources of chemistry, predominantly in the Far East, and
efficacy reflects an inadequate understanding of disease in-licensing assets. Thus today’s medicinal chemist is
pathophysiology in man and the poor predictability of required to optimally integrate a network of both internal

© 2012 Elsevier Ltd. 119

Section | 2 | Drug Discovery

60
53 New molecular entities
50 Biological licence applications
Number of drugs approved

40 39
35
30 31
30 27
24
21 21
20 18 18 19
17 16

10 7 7
6 5 6 5 6
3 3 4 3
2 2 2
0
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Fig. 9.1  Recent trends in small molecule approvals and biological licience applications – reproduced with the kind permission of
Nature publishing.

and external resources to prosecute the identification of foreseeable future and that, whatever infrastructure evolves
development candidates. within the industry, the medicinal chemist will have a key
Whilst medicinal chemistry is facing unprecedented role to play. This chapter will discuss some of the recent
pressures due to the volatility of the sector, it is clear that scientific advances in the field which will enable the
small molecule therapeutics will be required for the medicinal chemist to rise to the challenges ahead.

TARGET SELECTION AND VALIDATION

Target Lead Lead Preclinical Clinical

selection identification optimization development development

A key challenge in increasing the industry’s productivity and clinical counterparts, in the process of selecting and
will be the selection of targets which have compelling prioritizing future protein targets. In particular the devel-
validation in terms of both efficacy and safety in the opment of an in-depth understanding of the structural
human context and, which, consequently, have a greater biology of the target and the biophysics of its interaction
probability of achieving a successful clinical proof of with low-molecular-weight ligands is a critical component
concept. The role of the medicinal chemist has previously at the outset of a project and very much in the chemist’s
focused primarily on ensuring the quality of preclinical domain. The necessity for a rigorous analysis of potential
candidate molecules in terms of potency, selectivity and targets cannot be over-emphasized since this occurs at the
physicochemical properties such as those highlighted start of the 12–15-year period of intensive effort and
initially by Lipinski et al. (Lipinski CA, 1997) and, more investment required to achieve the launch of a new medi-
recently, by others (Leeson PD and Springthorpe B, 2007). cine into large scale clinical usage. The post genomic era
As a result, the level of attrition due to poor ADME proper- has provided the industry with a plethora of potential drug
ties fell dramatically between 1991 and 2000 (Kola I and targets; however, the selection of tractable targets with a
Landis J, 2004), but the level of clinical success with new high probability of delivering safe and effective treatments
mechanisms remains poor. However, chemists are now represents a huge challenge requiring a multidisciplinary
increasingly becoming engaged, along with their biology approach.

120
 The role of medicinal chemistry in the drug discovery process Chapter |9|

Whilst a detailed discussion of the complex facets of Various criteria for the selection and prioritization of
target validation is beyond the scope of this chapter (see targets are utilized within major Pharma in the process of
Chapter 6), it is clear that targets with strong genetic asso- building sustainable portfolios of viable potential targets
ciations, such as the voltage-gated sodium channel NaV1.7 for the discovery scientists to address. A recent illustration
(Dib-Hajj SD et al., 2007), with potential utility in the (Wehling M, 2009) has highlighted one such approach
treatment of pain, and the chemokine CCR5 receptor which objectively applies weighted scores to available
(Westby M and van der Ryst E, 2005), which has led to the target information.
discovery of treatments for HIV infection, exemplify an With the emergence of exciting new insights into disease
aspirational level of target validation. pathologies, the importance of rigorous target selection
Whilst genetic disease-association data provide strong will become even more important in the future. The avail-
evidence of target involvement, target modulation, ideally ability of the human genome sequence coupled with
within a human cellular pathway, provides compelling impressive advances in biology is uncovering challenging
validation, and the identification of early chemical probes targets for the medicinal chemist to address. As an example,
in addition to subsequent lead molecules will be an epigenetic phenomena are increasingly being recognized
important challenge for the medicinal chemist in the as a potentially important area in the development of
future. chronic diseases (Gluckman PD et al., 2008) and the dis-
Frye has recently described (Frye SV, 2010) essential fea- covery and characterization of specific histone-modifying
tures of a quality chemical probe which are summarised enzyme subfamilies (Cole PA, 2008) offers the medicinal
below: chemist the opportunity to design specific enzyme modu-
lators with which to regulate gene expression. The availa-
Molecular profiling. Sufficient in vitro potency and
bility of selective chemical probes for specific domains
selectivity data to confidently associate its in vitro profile
within these enzyme classes will be a critical factor in the
to its cellular or in vivo profile.
identification and selection of the most relevant targets.
Mechanism of action. Activity in a cell-based or cell-free
The promise of this area is illustrated by the successful
assay influences a physiologic function of the target in a
discovery of the histone deacetylase (HDAC) inhibitor
dose-dependent manner.
Vorinostat (McGuire C and Lee J, 2010; Figure 9.2).
Identity of the active species. Has sufficient chemical
Historically natural products have been considered as a
and physical property data to interpret results as
source of both new drugs and potential targets. More
due to its intact structure or a well-characterized
recently this approach has suffered demise due to its dis-
derivative.
appointing productivity. However, there have been a
Proven utility as a probe. Cellular activity data available
number of recent success stories which are illustrated by
to confidently address at least one hypothesis about the
the following examples. Ziconotide (Prialt®); (Schmidtko
role of the molecular target in a cell’s response to its
A et al., 2010) is a potent and selective N-type calcium
environment.
channel blocker approved by the FDA in December 2004
Availability. Is readily available to the academic
for the treatment of severe chronic pain. Ziconotide has
community with no restrictions on use.
subsequently demonstrated efficacy in patients with
The first four criteria attempt to describe ideal properties refractory pain, i.e. pain which has proven difficult to
of a probe whilst the last point addresses a more philo- control using conventional analgesics. Ziconotide is a
sophical viewpoint on the sharing of precompetitive infor- synthetic peptide derived from a toxin extracted from
mation. Further evolution of this ‘wish list’ will inevitably the marine snail Conus magus. The success of ziconotide
take place in the future. has highlighted the potential of natural toxins in drug
It is well known that target classes are associated with
different levels of tractability with respect to the probabil-
ity of being able to identify quality lead molecules. For
example the G-protein coupled receptor (GPCR) super-
family is generally considered to be one of the most trac-
table target classes with approximately 40% of prescription O
H
drugs in current practice producing efficacy via modula- N
tion of a GPCR. In contrast, voltage-gated ion channels are NHOH
generally regarded as one of the more challenging variety
O
of protein targets to modulate, particularly with respect to
subunit specificity. Progress in this area was for many years
limited by lack of suitable screening platforms to facilitate Vorinostat
drug discovery; however, recent technological develop-
ments offer the potential to more fully explore this target
class (Kaczorowski GJ et al., 2008). Fig. 9.2 

121
Section | 2 | Drug Discovery

discovery and has stimulated renewed interest from a address these pathologies will require significant
number of groups in the area (Halai R and Craik DJ, innovation.
2009). Peptides offer high levels of both potency and Encouragingly great strides have recently been made in
selectivity which make them valuable tools for target vali- our understanding of the factors which cause proteins to
dation and the toxins of a number of venomous species misfold, primarily through the use of biophysical and
have proved to be a rich source of such molecules. The computational techniques that enable systematic and
identification of the endogenous peptide hormone ghrelin quantitative analysis of the effects of a range of different
(Helstroem PM, 2009), and subsequent understanding of perturbations in proteins (Luheshi LM et al., 2008). Recent
its function, has highlighted the potential clinical value of advances in the design of small molecule inhibitors
ghrelin receptor agonists in a number of conditions such which inhibit protein–protein interactions has been
as gastroparesis (Ejskjaer N et al., 2009) and postoperative amply demonstrated by the discovery of the serum
ileus (Popescu I, Fleschner PR et al., 2010) in which the amyloid P cross-linking agent CPHPC (Pepys MB et al.,
ghrelin receptor agonist TZP-101 (Figure 9.3) has recently 2002; Figure 9.4) and the C-reactive protein inhibitor
demonstrated clinical efficacy. Several of the major disease 1,6-bis(phosphocholine)-hexane (Pepys MB et al., 2006;
challenges facing society, such as Alzheimer’s disease, Figure 9.5).
cancer and heart disease, etc., have been shown to be It is clear from the above selected examples that there is
associated, at least in part, with aberrant protein folding likely to be a plethora of novel biological targets to chal-
and/or protein–protein interactions and the identification lenge the innovation of medicinal chemists for the foresee-
of both small molecule and peptidic therapeutics to able future.

N H O
N
O
N O
O N H HO2C
H
O N
N

CO2H
F O

TZP-101 CPHPC

Fig. 9.3  Fig. 9.4 

O
O
+
Me3N P O +
O P NMe3
O
O

1,6-bis(phosphocholine)hexane

Fig. 9.5 

122
 The role of medicinal chemistry in the drug discovery process Chapter |9|

LEAD IDENTIFICATION/GENERATION

Target Lead Lead Preclinical Clinical

selection identification optimization development development

The identification of high-quality lead molecules is a criti- by this approach (Fox S et al., 2006), although a more
cal phase in the discovery of drug candidates since deci- recent review suggests that, over the past 3 years, the
sions made at this point in the process are likely to have success rate may have increased to approximately 60%
a significant impact on the outcome of the project as it (Macarron R et al., 2011). However, it is noteworthy that,
sets the framework for future lead optimization and devel- despite the massive growth in the numbers of compounds
opment (Bleicher KH et al., 2003). screened over the past 20 years, no corresponding increase
Reducing attrition has already been alluded to as a in the number of successful registered NCEs has resulted.
major challenge in the discovery of new therapeutic agents It has been argued that, in view of increasing development
and the lead generation phase is the earliest point in the times, it is still too early to evaluate the true success
drug discovery process where the incorporation of appro- of high-throughput screening (Macarron R et al., 2011).
priate physicochemical properties can be addressed. It is Nevertheless, there has been greater focus on improving
important to establish a clear target lead profile at this the ‘drug and lead-like’ properties and the structure/
early stage of the process which takes into account the property diversity of screening collections. ‘Drug-likeness’
critical properties required of the ultimate preclinical can- can be broadly defined as the overall profile of biophysico-
didate and extrapolates back into an appropriate minimum chemical properties of a molecule which facilitate its
acceptable profile for a lead series. Such a profile should access to and effective mode of interaction at the site of
not simply take into account target potency and selectivity action at a biologically relevant and safe concentration for
but should also include consideration of a wide range of sufficient duration to elicit the desired therapeutic effect.
physicochemical characteristics which would facilitate Leeson and Springthorpe discuss the importance of con-
favourable ADMET properties later in the process. sidering the properties conferring ‘drug-likeness’, particu-
The initial aim at the lead identification/generation larly lipophilicity, in a recent survey of selected development
stage is to identify ‘hits’ which are molecules that interact candidates in leading drug companies (Leeson PD and
with the chosen biological target in an initial screen, often Springthorpe B, 2007). An analysis of the structural rela-
carried out at a high concentration, to elicit a measurable tionship of launched drugs to their corresponding lead
response in a reproducible assay. Once validated by estab- molecules revealed that in general the structure of the
lishing the molecular integrity of the hit molecule(s), and marketed molecule was very closely related to the lead
confirming the robustness of the biological response, a series (Proudfoot JR, 2002). Thus one could infer that the
‘validated hit’ may be further investigated to establish apparent disappointing success rate of high-throughput
whether it possesses the potential to be regarded as a lead screening has been related to the lack of ‘drug likeness’ in
series. the compound collections and significant effort is now
Several strategies have been adopted to identify these being devoted to address this.
early hits as listed below: Major companies have, therefore, filtered their screening
collections to exclude compounds which do not possess
• Existing drugs
physicochemical properties that are generally accepted to
• Natural ligands
confer ‘drug-like’ attributes (Bleicher KH et al., 2003).
• Natural products
However, it should be noted that, although the discovery
• Focused screens
of drug-like preclinical development candidates requires
• Rational structure-based design
the initial identification of quality leads, the properties
• Knowledge-based design
which confer ‘drug-likeness’ and those associated with
• Fragment screening
‘lead-likeness’ are not the same (Rishton GM, 2003). It is,
• Virtual screening
therefore, important to understand the components con-
• High-throughput screening (including phenotypic
ferring ‘lead-likeness’ when assessing screening collec-
screens).
tions. A publication from the Astra-Zeneca group addressed
The widespread use of high-throughput screening has the design of lead-like combinatorial libraries and sug-
generally been reported to have enabled the identification gested that leads suitable for further optimization are most
of leads for approximately 50% of the projects addressed likely to be relatively polar, low molecular weight (MW =

123
Section | 2 | Drug Discovery

200–350) and of relatively low lipophilicity (clogP ca.1.0– validation in the disease of interest, there has been signifi-
3.0) (Oprea TI et al., 2001). The authors point out that cant activity in establishing screens based on the cell phe-
when beginning from such a tractable low-molecular- notype, i.e. where biochemical pathways are targeted as
weight lead molecule, subsequent optimization to increase opposed to specific proteins within the pathway. The
potency and selectivity is likely to increase molecular major challenge associated with this approach is the sub-
weight and lipophilicity to values in the order of the sequent identification of the specific molecular target or
accepted ‘drug-likeness’ parameters. targets with which the agents interact to produce the
Whilst significant progress is being made in improving desired cellular response (Hart CP, 2005). Whilst in some
the quality of compound collections with respect to the cases further optimization of leads has been carried using
enrichment of ‘lead-like’ components and the elimination the phenotypic assay itself, because of the numerous vari-
of potentially toxic or undesirable functionality, improving ables involved, e.g. multiple targets and cellular distribu-
chemical diversity is an ongoing preoccupation since the tion, etc., precise SAR can be difficult to obtain from the
concept of diversity is open to a variety of interpretations. phenotypic assay. Thus a detailed analysis and elucidation
Whilst in its original conception combinatorial chemistry of the mechanism of action is often required to enable
aimed to generate large numbers of molecules providing rapid progress to be achieved using a protein specific assay.
a high level of diversity, the early libraries tended to be Over the past few years fragment-based drug discovery
bolstered with molecules having significant molecular has become an established technique for lead generation,
complexity and only modest diversity due to the limita- often in conjunction with other strategic approaches
tions of the synthetic methodologies which could be uti- (Whittaker M et al., 2010). Fragment-based screening con-
lized in a high-throughput mode. Furthermore, compound sists of screening a chemical library of low-molecular-
collections with less than optimal physicochemical proper- weight compounds, usually at high concentration, using
ties were previously employed in high-throughput screen- sensitive assay methods. In addition to biochemical assays
ing campaigns. Thus more attention is now being paid to such campaigns often incorporate biophysical methodol-
the generation of focused libraries containing subsets of ogy such as NMR, surface plasmon resonance (SPR), iso-
compounds having so-called ‘privileged scaffolds’ for spe- thermal titration calorimetry (ITC) and X-ray crystallography
cific target classes. This has in turn shifted the emphasis (Carr R and Jhoti H, 2002; Rees DC et al., 2004; Orita M
from combinatorial synthesis to rapid automated synthesis et al., 2009). The considerable progress achieved with
of small quantities of target compounds. in silico design of fragment libraries using a variety of
High-throughput screening (Chapter 8) continues to be techniques has recently been reviewed (Konteatis ZD,
a valuable strategy where there is little information around 2010). Essentially the fragment-based approach aims to
the structure of the biological target or its ligands; however, identify ligands for the target protein which have relatively
where such knowledge exists, it has generally been more low affinity but high ligand/ligand-lipophilicity efficiency.
successful to utilize this information to identify potential Ligand efficiency (LE) is a measure of the binding energy
molecular starting points. For example, structural informa- of a molecule normalized by its size. This is often calcu-
tion on ligands which are known to interact with a target lated by dividing the pKi or IC50 by the molecular weight
can be used to construct a pharmacophore, which may be or number of heavy atoms. Similarly ligand-lipophilicity
used in an initial virtual screening programme of propri- efficiency (LLE) can be calculated by determining the dif-
etary or commercial compound collections to enrich the ference between the pKi or pIC50 value and the clogP, which
sample set of compounds ultimately selected for actual provides a measure of the binding energy of a molecule
screening. A critical review assessing the past effectiveness normalized by its lipophilicity (Smith GF, 2009). Thus
of virtual screening and proposing considerations for compounds with high LE and LLE interact highly effi-
future improvements has recently appeared (Schneider G, ciently with the biological target. Subsequent optimization
2010). Computational tools are widely used to generate in is then focused on increasing potency whilst maintaining
silico pharmacophores, using either molecular or elec- high LE/LLE, thus avoiding the pitfall of increasing potency
tronic overlays. The introduction of electronic field overlay by increasing size and lipophilicity which often leads to
technology has allowed chemists to interrogate electronic promiscuous ‘off-target’ interactions and subsequent toxic-
properties in a relatively accessible manner. Similarly the ity. Although the concepts of LE and LLE have been particu-
latter technology may be usefully applied to ‘scaffold larly applied in fragment-based design they are now widely
hopping’ where a structurally distinct compound class has applied across all lead generation strategies. There are now
been derived from an active series via application of a field numerous examples of a fragment-base approach being
pharmacophore (Cheeseright TJ et al., 2008). An extreme employed in drug discovery projects and several have been
example of utilizing structural information is where an reviewed recently (Congreve M, Chessari G et al., 2008).
existing known drug molecule is used as the starting point In stark contrast to the fragment-based approach to lead
to discover an improved agent (Naik P et al., 2010). generation, natural products have been, and arguably con-
Although the majority of high-throughput screens are tinue to be, a fruitful source of lead molecules and drugs
aimed at known biological targets with supporting target over the past 25–30 years (Newman DJ and Cragg JM,

124
 The role of medicinal chemistry in the drug discovery process Chapter |9|

2007). However, despite the demonstrable historical suc- important role in the identification of lead molecules (Carr
cesses of natural product research, major Pharma has R and Jhoti H, 2002). Similarly, NMR is proving to be an
largely reduced its investment in natural product screening effective, accessible, complementary technique in identify-
in favour of the small molecule library approaches ing ligand-protein interactions particularly for membrane
described above. The anticipation that combinatorial bound proteins such as GPCRs, for which detailed X-ray
chemistry/automated synthesis would provide an excess of structures are relatively rare (Bartoschek S et al., 2010).
small molecule leads for most biological targets was partly Recent developments in the acquisition of bioengi-
responsible for the demise of this strategy. Furthermore neered protein and subsequent X-ray crystal structures
the technical challenges associated with the deconvolution of stabilized GPCRs (Tate GC and Schertler GFX, 2009)
of multicomponent natural product mixtures and subse- may provide exciting future opportunities for medicinal
quent chemical simplification of large and complex struc- chemists to actively use ligand-bound co-crystal structures
tures has also deterred continued investment in this field. of GPCRs in an analogous fashion to the enzyme
Nevertheless a number of smaller companies have recog- co-crystallization structures which have greatly facilitated
nized this as an opportunity and are pursuing this strategy the discovery of selective enzyme inhibitors.
for lead identification and it has recently been argued that Industry pressure to improve productivity and efficiency
the technical challenges associated with this approach in the identification of new chemical entities has added
have been lessened (Harvey AI, 2008; Harvey AI, 2010). stimulus to the exploration of new technologies in the
Despite the relative demise of natural product screening identification of lead molecules. In particular miniaturi­
over the past few years, natural products have inspired the zation and automation of chemical synthesis and biologi-
development of new strategies in the identification of cal screening assays are currently under intense focus
privileged structures from which small molecule libraries (Lombardi D and Dittrich PS, 2010).
have been constructed. Herbert Waldmann has received The identification of screening platforms which obviate
notable acclaim for his biology-oriented synthesis (BIOS) the requirement for artificial labels or reporter systems is
strategy which encompasses the development of small also receiving considerable attention (Shiau AK et al.,
molecular entities derived from natural product and drug 2008). Biophysical methods such as surface-plasmon reso-
families, the so-called ‘structural classification of natural nance and impedance-based technologies are developing
products’ (SCONP) (Keiser M et al., 2008) and an associ- rapidly and will become more widely used, particularly
ated ‘scaffold hunter’ approach (Wetzel S et al., 2009). since these approaches are less likely to identify false posi-
X-ray crystallography has been effectively applied over tives which result from labelling artefacts and aggregation
many years to optimize small molecule ligands primarily phenomena (Giannetti AM et al., 2008). The successful
for soluble enzyme targets. With the introduction of development of these new technologies has the potential
fragment-based screening high-throughput X-ray crystal- to transform the lead identification process and address,
lography has over the past decade also assumed an in part, the challenge of improving productivity.

LEAD OPTIMIZATION

Target Lead Lead Preclinical Clinical

selection identification optimization development development

The lead optimization stage of drug discovery is at the before the initiation of a programme, thus giving the
pivotal interface between lead identification and the early medicinal chemist and the programme team a clear focus.
development phase of a molecule. The ultimate aim of From the candidate profile a focused screening strategy can
the lead optimization phase is the identification of a be developed and should only contain screens which
compound which possesses the required properties of facilitate decision-making. Careful consideration should
the targeted preclinical development candidate. Thus the be given to the components and order of assays within the
medicinal chemist must be cognizant of the wide range of screening cascade which should include higher through-
parameters which will need to be optimized in the lead put assays, for which high attrition can be expected, early
thus ensuring a high probability of success during subse- in the triage and lower throughput and more discerning
quent development. It is important that the desired can- assays, including those involving animals, later in the
didate profile is clearly defined as early as possible, ideally cascade. A major consideration in constructing an effective

125
Section | 2 | Drug Discovery

screening cascade is the choice of selectivity assays which evidence of the desired selectivity profile, (c) have activity
need to be included. The selectivity panel normally con- in cellular systems if necessary, (d) have acceptable meta-
sists of proteins both within the target family and selected bolic stability in vitro, (e) be free of any structural toxicity
‘off-target’ proteins which, if modulated, would result in alerts, and (f) not have any major cytotoxicity issues. It is
undesirable toxicity or side effects, discussed in more highly advantageous for a lead to demonstrate some evi-
detail below. dence of oral exposure even at this early stage.
In addition to fulfilling potency and selectivity criteria It is important also to consider how many lead series
a potential development candidate should possess appro- should be investigated in the early stages of the optimiza-
priate ADME properties to deliver acceptable exposure tion process and the requisite level of structural diversity
of the compound at the target site by the preferred route in order to mitigate against failure of a particular series.
of administration allowing data obtained from subse- Ideally three structurally distinct lead series would be
quent in vivo profiling in preclinical disease models to selected for lead optimization at the commencement of
be interpreted with confidence and facilitate predictions the process.
for future clinical efficacy. Much of this section will focus A lead molecule will have a predefined level of potency
on molecules intended for oral administration, but a at the chosen target together with the initial target selec­
brief discussion on the different requirements of drugs tivity for both related targets and common ‘off-target’
intended for administration via other routes will also be liabilities which are considered important to avoid. The
included. primary assays in most screening cascades focus on activity
A number of excellent reviews have recently appeared at the target protein and selectivity to allow the medicinal
which describe the lead optimization process in detail chemist to quickly optimize these important parameters
(Baringhaus KH and Matter H 2005; Lindsley GW et al., early in the process. At this stage rudimentary structure–
2009). This section will focus on recent discoveries and activity relationships will have already been established
developments that have enhanced or have the potential to during the lead identification phase providing an insight
further improve the traditional process. as to which moieties of the lead compound may be modi-
The process commences with the careful selection of fied to achieve the desired result. Furthermore the availa-
lead series which, as previously stated, is of great impor- bility of other information such as protein X-ray crystal
tance since this selection has consequences for the future structures or docking predictions from target protein
lengthy and expensive discovery and development pro- homology models can provide strategic guidance for the
gramme. It is essential that great attention is paid to select- optimization process. In addition to considering selectiv-
ing robust leads with the appropriate physicochemical ity within the target protein family it is important to con-
properties. Recent analyses (Kola I and Landis J, 2004; sider selectivity more broadly (Bass AS et al., 2009) since
Keserü GM and Makara GM, 2009) have clearly shown a growing number of ‘liability targets’ are being identified
that leads chosen from the ‘wrong’ area of chemical space as being potentially responsible for safety issues which are
have a much higher probability of failure than those from often only detected later in the development process.
the ‘correct’ area. Often failure occurs late in development Screening compounds against recombinant proteins has
and can be extremely costly. Far too often medicinal chem- become widely accepted since this allows the configura-
ists have been seduced by high levels of potency in ‘non tion of robust high-throughput assays, which perform
drug-like’ lead molecules only to find that the physical consistently over time. However, caution should be exer-
properties of the molecule are incompatible with the cised when interpreting data generated from such assays
desired candidate profile. since the biological activity data from these assays do not
Lipinski has highlighted (Lipinski CA et al., 1997) the necessarily correlate well with activity from native tissue
importance of the physical properties of a molecule and, systems (Eglen RM et al., 2008). It is therefore important,
whilst this paper is often quoted, the guidelines contained where feasible, to periodically interrogate the relationship
within are too often disregarded. between activities derived from each assay system. Ideally
It is important to measure key physical properties such the native tissue used should be closely related to the
as logD, solubility and pKa (where appropriate), etc., of target tissue and from the relevant disease situation wher-
novel compounds throughout the drug discovery process ever possible.
to ensure that molecules continue to reside within ‘drug- A further consideration when designing a screening
like’ chemical space and to regularly analyse data to inves- cascade is the potential for species differences to exist in
tigate potential correlations between physicochemical which the biological activity at the target protein varies
properties, ADME parameters and biological activities (see between different species. Species differences can lead to
also Chapter 10). problems in the extrapolation of in vitro to in vivo data
In addition to having the desired physicochemical prop- and the interpretation of data from in vivo efficacy and
erties a lead molecule should also have the following char- safety studies (Swanson R and Beasley JR, 2010). Whilst an
acteristics: (a) be a member of a distinct structural series indication of potential species differences can be obtained
with confirmed structure–activity relationships, (b) show by performing a bioinformatics analysis early in the

126
 The role of medicinal chemistry in the drug discovery process Chapter |9|

project, differences in activity can occur even when there are being developed and are anticipated to be in routine
is a high degree of homology between proteins. It is there- use as screening assays in the near future.
fore important to consider the incorporation of animal The assessment of potential metabolic instability is also
orthologue assays in the screening cascade which would an important early evaluation in a screening cascade. Thus,
typically be the species of choice for the primary animal typically compounds are incubated with human and/or
model. rodent liver microsomes and their metabolic stability is
Much of the focus in lead optimization is on optimiza- expressed as the percentage of compound remaining after
tion of ADME parameters which are clearly important to a given time or, more commonly, as an intrinsic clearance
(a) select the most appropriate compounds for evaluation parameter (Nagai N, 2010). For a more thorough evalua-
in in vivo disease models, (b) ensure that the drug is opti- tion of in vitro metabolic turnover human and rodent
mally delivered to its intended site of action, (c) allow a hepatocytes may be used which possess the capacity to
prediction of the human pharmacokinetics and estima- carry out both phase 1 and 2 metabolic processing. In
tion of clinical dose, (d) minimize the potential for drug– addition an assessment of plasma protein binding is often
drug interactions, (e) avoid selecting compounds with the included since the degree to which a molecule binds to
potential to form reactive metabolites and thus reduce the plasma proteins can influence both the efficacy and the
potential for idiosyncratic toxicity, and (f) provide confi- tissue distribution of the compound (Chang G et al,
dence that safety studies can be performed in several 2010).
species. As molecules progress to the later stages of lead optimi-
Several in vitro ADME screens feature early in virtually zation an understanding of the route and specific sites of
all screening strategies. The introduction of cytochrome metabolism of the compounds can be important informa-
P450 (CYP450) inhibition assays has allowed potent tion to assist the medicinal chemist to further improve the
inhibitors of these key metabolizing enzymes to be identi- metabolic stability of the compound. For example, the
fied early in the optimization process. The elucidation of introduction of blocking groups at sites of metabolism or
crystal structures of a number of CYP450 enzymes has electron withdrawing substituents at appropriate sites on
alerted medicinal chemists to specific chemical modifica- the molecule, which will reduce electrophilicity, can sig-
tions with which to mitigate such liabilities from candidate nificantly reduce the rate of metabolic turnover. An early
molecules, thus reducing the risk of drug–drug interactions assessment of the potential of a compound to form reac-
in the clinic (Ekroos M and Sjögren T, 2006; Foti RS et al., tive and, therefore, potentially toxic metabolites is also
2010). For most CYP450 isoforms a strong correlation valuable and can be achieved by using a combination of
exists between the lipophilicity of a compound and its time-dependent CYP450 inhibition (Howard M et al,
CYP450 inhibitory potential (Lewis DFV et al., 2007). 2010) and trapping experiments, typically with either glu-
In addition to determining the potential of a molecule tathione or cyanide (Riley RJ et al., 2007). For molecules
to inhibit CYP450 enzymes, it is also important to assess with the potential for reactive metabolite formation addi-
the potential of a compound to induce CYP450 expres- tional in vitro and in vivo studies to detect covalent
sion. CYP450 induction can potentially lead to increased binding to tissue protein, using radiolabelled compound,
clearance of the compound on chronic dosing and toxicity may be required (Evans DC et al., 2004).
due to the possible formation of reactive oxygen species. A preclinical pharmacokinetic study (see Chapter 10) is
For some years the pregnane X receptor (PXR), a nuclear often the first occasion at which a compound is tested in
hormone receptor assay has been used to determine the an in vivo experiment. These studies are normally per-
potential of a compound to induce CYP3A4. More recently, formed in rats since this is the species of choice for the
additional assays have become available for mechanisms majority of disease models and for rodent safety assess-
by which other CYP isoforms are induced (Chu V et al, ment studies. Compounds targeted as oral therapies are
2009) and could enter more routine use in the future. administered by both intravenous (i.v.) and oral (p.o.)
Currently such CYP450 induction assays are not com- routes, each experiment providing valuable information
monly included early in the screening cascade unless the for the medicinal chemist. An i.v. pharmacokinetic study
particular liability has been identified, but are employed allows the clearance (the rate by which the drug is elimi-
in the later stages of lead optimization, particularly if the nated from the system) of the drug to be measured together
level of systemic exposure falls following chronic dosing. with the volume of distribution (a measure of how well
Much attention is currently being given to the role the drug distributes into tissue). Both the clearance and
of transporter proteins (Chu V et al, 2009), which have the volume of distribution determine the half-life of the
been implicated in the unfavourable disposition of drugs, drug. A p.o. study is used to determine the circulating
either by active efflux from target tissue, potentially limit- levels that can be achieved after a given p.o. dose. Studies
ing efficacy, or through accumulation leading to toxicity frequently use both i.v. and p.o. routes of administration
and potential drug–drug interactions (Lai Y et al, 2010). to gain a full profile of a compound and allow the abso-
Currently only P-glycoprotein (Pgp) assays are used in lute bioavailability to be determined (White RE, 2009;
routine screening, but assays for many other transporters Chapter 10).

127
Section | 2 | Drug Discovery

The volume of distribution is a parameter which quanti- pharmacokinetic and pharmacodynamic half-lives for the
fies the distribution of the drug between the plasma and compounds, thus obviating the need to maintain high
body tissue, and ideal molecules should have low clearance drug levels in the plasma over an extended time period to
from the blood and a volume of distribution indicative of produce an efficacious response. A potential key advantage
distribution into total body water (i.e. > 1). It is thus of these so-called ‘tight binders’ is that they may possess
important to understand the properties which affect the an improved safety profile (Packeu A et al., 2010), as the
volume of distribution of a compound. Common factors total body burden of drug is reduced.
which can affect this parameter are plasma protein binding, Molecules which fulfil the targeted potency, selectivity
the physical properties (particularly the presence or and ADME properties will usually progress to in vivo
absence of either an acidic or basic centre) and the involve- studies in animal models to assess their potential for
ment of transporters (Grover A and Benet LZ, 2009). in vivo efficacy. The relevance of animal models for both
Common strategies to increase the volume of distribution efficacy and, in some cases, safety is a matter of much
include the introduction of a basic centre to increase the debate within the drug discovery community. With an
pKa, if appropriate, and to increase the LogD. However, increasing number of compounds failing to demonstrate
unless the lead molecule is hydrophilic in nature, increas- clinical efficacy, despite having shown excellent results in
ing lipophilicity can result in increased metabolism and preclinical models of disease, it is clear that the animal
CYP450 inhibition as discussed above and thus the iden- models are not totally predictive and a better understand-
tification of the optimal properties is often a trade-off of ing of the translation of data from animal model to
physicochemical properties. In addition to blocking sites human disease for a particular mechanism is required.
of metabolism to reduce clearance, reducing the lipophilic- Within areas where validated predictive models exist for
ity of a molecule is often effective at reducing the rate of specific mechanisms in vivo data can be very powerful
metabolism. Drugs which are specifically aimed at targets both in aiding the selection of advanced molecules as
within the central nervous system (CNS) require more potential development candidates and in predicting the
stringent control of their physicochemical properties to efficacious clinical dose through the use of pharmacokinetic/
facilitate passive permeation of the blood–brain barrier pharmacodynamic (PK/PD) studies which correlate the
which is comprised of endothelial tight junctions. In systemic exposure of the compound with the degree of
comparison with Lipinski guidelines, molecules designed efficacy.
to cross the blood–brain barrier would typically have In addition to the optimization of ADME properties
a lower molecular weight (<450), lower clogP (1–3), during the lead optimization phase, compounds are also
reduced hydrogen bond acceptors (<6) and donors (<2) evaluated for potential toxicity with subsequent chemical
and lower total polar surface area (TPSA) (<75A2). The modification being applied to mitigate potential risks. A
blood–brain barrier also contains efflux transporters such detailed discussion of toxicological evaluation in drug dis-
as P-glycoprotein which serve to prevent access of sub- covery is beyond the scope of this section (see Chapter 15);
strates into the CNS. The free fraction of the compound, however, it is pertinent to mention the function of such
i.e. that proportion which is not bound to plasma protein, assays in the process. Cytotoxicity assays are commonly
has also been reported to have a profound effect on both used alongside cellular functional screens to identify mol-
access to the CNS and subsequent distribution to the site ecules which produce biological activity by virtue of cel-
of action (Jeffrey P and Summerfield S, 2010). lular toxicity. An area of intense current interest is that of
The desired pharmacokinetic profile of a drug will the potential for compounds to produce adverse cardiovas-
depend on the site of action and the nature of the target. cular effects in the clinic, since cardiovascular toxicity is
Most commonly used drugs are required to penetrate responsible for a significant number of drug withdrawals.
tissue to reach their target protein and to be present at An increasing number of targets, currently confined to ion
therapeutically efficacious concentrations for long enough channels, have been identified in recent years and are now
to support once, or at most twice, daily dosing. commonly included in ‘off-target’ selectivity panels during
The majority of discovery programmes have targeted lead optimization, the most well characterized of these
orally well-absorbed molecules with high metabolic sta- being the hERG channel (Hancox JC et al., 2008; see also
bility to maintain high circulating blood levels for suffi- Chapter 15). The panel of cardiac liability channels being
cient time to deliver the required duration of action. Thus incorporated into screening cascades is increasing with
the duration of action for these agents is related to their many groups routinely including other channels such as
pharmacokinetic properties, particularly the rate of clear- Nav1.5, Kv1.5, Cav1.2, KCNQ1 (hmink) amongst others
ance from the plasma. An alternative approach has been (Cao X et al., 2010). Later in the optimization phase, when
to identify molecules which have an extended residence potential candidates have been identified, assays to detect
time, due to a slow ‘off rate’, at the target protein which is genotoxicity are routinely employed. Typically compounds
significantly longer than the time over which an effica- are tested in both a bacterial mutagenicity assay such as the
cious plasma concentration of the compound is main- Ames test initially and also in a mammalian screen such
tained. In such cases there is a clear mismatch between the as the mouse micronucleus assay (Reifferscheid G and

128
 The role of medicinal chemistry in the drug discovery process Chapter |9|

Buchinger S, 2010). Prior to being accepted into preclinical unacceptable level. In the last 10 years 17 drugs have
development it is usual to conduct a preliminary in vivo been withdrawn (http://en.wikipedia.org/wiki/List_of_
safety study, typically dosing for 7 or 14 days, to provide withdrawn_drugs) due to adverse drug reactions. Further-
an initial assessment of the maximum tolerated dose in more, adverse drug reactions have been estimated to
preparation for subsequent more precise regulatory safety account for approximately 5% of deaths amongst hospital
studies. patients and 3% of the general population (Wester K et al.,
Much of this section has focused on the lead optimiza- 2008). The most common toxicities are associated with
tion process for oral drugs and it is important to recognize the liver, the cardiovascular system, skin and blood. Sig-
other routes of administration where the molecule profiles nificant efforts are being made to understand the reasons
will be significantly different, particularly with respect to responsible for these toxicities and thus to develop assays
ADME and physical properties. Other common routes of capable of identifying molecules possessing the potential
administration are inhaled or intranasal, intravenous and to cause these adverse events.
topical. There are four main factors which can give rise to organ
For inhaled drugs, whose site of action is in the lung, a toxicities: the overall physical properties of a drug mole-
poorly soluble molecule which is highly metabolized in cule, the primary pharmacology, the secondary or ‘off-
the circulation and of low membrane permeability is often target’ pharmacology and the presence of structural
the preferred profile. This profile is aimed at ensuring elements known to cause toxicity. The medicinal chemist
maximum residence time in the lung for optimal efficacy needs to consider all of these factors when designing
and minimal systemic exposure thus reducing the poten- potential drug molecules.
tial for any side effects (Ritchie TJ et al., 2009). A similar Medicinal chemists have long been aware of the rela-
profile is desirable for topical drugs for application to the tionship between physical properties and ‘drug likeness’
skin (Sloan KB et al., 2006). (Lipinski CA et al., 1997); however, recent reviews (Leeson
For intravenously administered drugs a very different PD and Springthorpe B, 2007) suggest that a significant
profile is required where a highly soluble molecule is pre- percentage of compounds being prepared by major phar-
ferred (Shi Y et al., 2009) to allow a low dose volume to maceutical companies are too highly lipophilic and it has
be employed. In the latter case the focus of optimization been proposed that high lipophilicity is likely to be a
is generally on the identification of highly potent agents major cause of toxicity. A study by Pfizer (Price DA et al.,
(to facilitate low dose), with high metabolic stability and 2009) which investigated potential correlations between
an acceptable duration of action. For drugs projected to be physical properties and observed in vivo toxicity for 245
administered via routes other than the oral route, consid- of their preclinical development compounds showed a
erations such as Lipinski guidelines, which address oral marked correlation between cLogP and toxicity and an
bioavailability, are inappropriate. inverse correlation between total polar surface area (TPSA).
They concluded that ideally a molecule should have a
cLogP < 3 and TPSA > 75 to reduce the probability of
producing toxicity. The importance of reducing lipophilic-
ADDRESSING ATTRITION ity cannot be over-emphasized as it has been shown to
influence all parameters that the medicinal chemist needs
Attrition is currently the major issue facing scientists to control in discovering an orally bioavailable drug
engaged in the drug discovery process. During the 1990s including solubility, membrane permeability, plasma
the pharmaceutical industry took great steps to identify protein binding, metabolism (Waring MJ, 2010) and the
the reasons responsible for the high attrition rate in devel- potential to cause drug–drug interactions (Lewis D and Ito
opment and introduced measures to overcome the issues Y, 2010). High lipophilicity has been postulated to increase
(Kola I and Landis J, 2004). The incorporation of both the probability for promiscuous binding of the compound
in vitro and in vivo ADME assays in discovery screening to ‘off-target’ proteins thus increasing the likelihood of
cascades significantly reduced the subsequent level of attri- toxicity (Leeson PD and Springthorpe B, 2007).
tion due to poor pharmacokinetic properties during clini- Adverse drug reactions have been characterized into
cal evaluation between 1990 and 2000. However, in 2000 three groups (Pirmohamed M et al., 1998). Type A (aug-
an increased number of molecules were failing due to mented) adverse reactions are the so-called target-mediated
toxicity (Kola I and Landis J, 2004). Ten years later the toxicities and are related to the pharmacological effect
level of attrition due to toxicity is still high in the discovery of the drug and such adverse events are predictable and
phase and much effort is currently being devoted to iden- dose-related. Type B (idiosyncratic) adverse reactions are
tifying the key reasons responsible for this. unpredictable, often non-dose related, less common, but
Of more concern is the level of late stage attrition, frequently more severe than Type A. Type C (chemical)
i.e. that occurring subsequent to the discovery and adverse reactions are related to structural elements in the
development phase and a worrying number of registered drug molecule and may be predictable from the molecular
drugs continue to be withdrawn from the market at an structure.

129
Section | 2 | Drug Discovery

Type A, augmented or target-mediated, toxicity is an which could lead to an adverse cardiovascular effect. Many
important factor to take into account, particularly when groups are now screening compounds against a panel of
choosing a target or during a target validation exercise. If cardiac ion channels throughout the discovery process (see
interacting with the target is likely to cause an adverse event Lead optimization section).
then the implications of the adverse event in the context A number of medicinal chemistry strategies have been
of the disease requires careful consideration based on the adopted to reduce the potential for hERG inhibition,
potential benefit of the entity versus its risk. For example which include reducing lipophilicity, modulating the pKa
many anticancer agents are cytotoxic and would not be to reduce π-cationic interactions and reducing the poten-
considered suitable treatments for less serious conditions. tial for π-π interactions. The relative success of these strate-
Type B, idiosyncratic toxicity is a growing area of con- gies has been recently reviewed (Jamieson C et al., 2006).
siderable interest (Ulrich RG, 2007) and the major cause There is currently no available X-ray crystal structure of the
of drug withdrawals. This type of toxicity is often severe hERG channel; however, the value of homology models
and can result in hospitalization and, in extreme cases, using known potassium channel structures to guide struc-
death. Recent data suggest that some of the mechanisms tural modifications has been effectively demonstrated in
responsible are becoming better understood, but there is the discovery of the CCR5 receptor antagonist, maraviroc
still a need for much further work. In particular the forma- (Price DA et al., 2006).
tion of reactive metabolites which may bind covalently to Drug-induced liver injury is also a major cause of com-
cellular protein and precipitate an immunological pound attrition, often through the formation of reactive
response is currently a leading hypothesis to explain at metabolites and this will be discussed later in this section.
least some of the idiosyncratic toxicity (Kaplowitz N, A greater understanding is emerging on the possible role
2005). The potential for a compound to form reactive of drug–transporter interactions in the development of
metabolites can be assessed by a combination of time- hepatotoxicity (Lai Y et al., 2010) which may result in cel-
dependent CYP450 inhibition studies and electrophile lular accumulation, impaired efflux, alteration of nutrient
trapping experiments with an agent such as glutathione transport and altered drug disposition leading to drug–
(Kalgutkar AS and Didiuk MT, 2009) followed, if neces- drug interactions. For example the toxicity of troglitazone
sary, by detailed covalent protein binding studies (Evans has been partially attributed to its interaction with the
D et al., 2004). The major organs affected by idiosyncratic bile salt export pump (BSEP) (Funk C et al., 2001). The
toxicity are the heart, liver, kidneys and skin (Caldwell, number of known transporters is currently growing and
2006). It is important that any potential for idiosyncratic assays are being developed to assess the potential of drugs
toxicity should be considered in the context of the pre- to cause toxicity through interaction with these proteins
dicted efficacious plasma concentration of the drug and (Greer MI et al., 2010).
the targeted indication. Empirically it is assumed that if Another area of current growing interest, which has been
the dose is 10 mg or less then the probability of idiosyn- implicated in liver toxicity, and, indeed, toxicity in other
cratic toxicity is reduced significantly (Uetrecht J, 2001). organs, is the potential of a drug to impair mitochondrial
Cardiovascular toxicity still remains one of the major function (Dykens JA and Will Y, 2010). This is a complex
causes of drug withdrawal and the recent removal of clob- and developing area which is beyond the scope of this
utinol from clinical usage as a result of cardiac QT interval chapter, however 80% of drugs which have received FDA
prolongation (Takahara A et al., 2009) highlights the need ‘Black Box Warnings’ for hepatotoxicity and cardiovascular
to study potential drug molecules carefully to assess the toxicity have been shown to inhibit mitochondrial function
risk of this phenomenon. The hERG channel, a potassium (Dykens JA and Will Y, 2007). Assay systems are currently
channel involved in the repolarization phase of the heart, available and a suggested protocol for their use in drug
has been known for some time to be associated with discovery has been published (Dykens JA and Will Y, 2010).
cardiac QT interval prolongation and there is a good cor- Recent studies have suggested biomarkers that are
relation between compounds which have high hERG potentially predictive of liver injury that could be used
binding activity and clinical arrhythmogenic activity, par- both preclinically and clinically. Two such biomarkers are
ticularly the torsade des pointes syndrome (Polak S et al., the high mobility group box1 and keratin 18 proteins
2009). A variety of assays are available to assess the poten- (Antoine DJ et al., 2009).
tial of compounds to inhibit the hERG channel, but it is Whilst renal toxicity has not been responsible for a large
important to choose the most appropriate in vitro assay number of drug withdrawals, it is still a major cause for
and not to rely on data from one assay in isolation (Pollard concern. Interaction with transporter proteins has recently
CE et al., 2010). It is advisable, for advanced compounds, been implicated in the renal toxicity of a number of sub-
to obtain data in both in vitro and in vivo assays to allow stances. Rosuvastatin (Crestor®) was shown to cause pro-
an informed decision on the potential progression of the teinuria at a dose of 80 mg in Phase 3 clinical trials. This
molecule (see Chapter 14). The hERG channel is one phenomenon has subsequently been shown to be due to
of several ion channels (Witchel HJ, 2010) involved in the a combination of the target pharmacology of the drug
repolarization phase of the heart, interference with any of (inhibition of protein endocytosis) and the fact that it is

130
 The role of medicinal chemistry in the drug discovery process Chapter |9|

a substrate for the organic anion transporter 3 (OAT3),

which causes it to accumulate in the kidney (Windass AS SUMMARY
et al., 2007). Until recently there has been a lack of predic-
tive biomarkers for renal toxicity; however, a recent study The environment in which drug discovery is taking place
by an international consortium, has identified seven is currently undergoing considerable change and will con-
biomarkers which can be used both preclinically and in tinue to do so for some time to come. The downsizing of
human studies (Hewitt P and Herget T, 2009). major Pharma and the increased focus on academic drug
Type C or structural-based toxicity is clearly an area that discovery are attempts to deal with the escalating costs,
the medicinal chemist should particularly consider before high attrition rate and patent expiries which are threaten-
embarking on a chemical modification. Structural alerts, ing the viability of the current discovery model (see
of which the chemist should be aware, can be divided into Chapter 22).
four groups, chemically reactive functionality, structural This chapter is intended to convey the message that the
elements which present a risk following metabolic activa- medicinal chemist has an extremely important role to play
tion, DNA binders and CYP450 inhibitors. There are in all aspects of the discovery process to address the chal-
several excellent reviews in the recent literature covering lenges of increasing efficiency and productivity and reduc-
this area (Blagg J, 2006; Kalgutkar AS and Didiuk MT, ing attrition. Whilst traditionally the medicinal chemist
2009) in detail and describe many structural features has not become engaged with the drug discovery process
which should be given careful consideration before incor- until the lead identification phase, there is a compelling
poration within a drug molecule. The medicinal chemist argument for their involvement at the target selection and
is advised to read these carefully and to be aware of new validation stage. The identification of high-quality chemi-
toxicophores as they appear in the literature. cal probes as early as possible will enable robust target
Despite the increasing volume of information, the theo- validation and the selection of those biological targets
retical prediction of toxicity is still imprecise. It is, there- which are more likely to succeed.
fore, important to incorporate ‘predictive’ assays early in Furthermore, an increasing array of technological
screening strategies. Despite recent progress, current advances are becoming available which will enable the
in vitro safety testing is conducted in animal cells or, at chemist to develop a detailed understanding of the bio-
best, in recombinant human cell lines. It would be prefer- physics associated with compound-target engagement,
able to conduct these studies in native human cells and which can be effectively incorporated into the design of
recent advances in stem cell research facilitate this oppor- suitable preclinical candidates.
tunity in the future (Balls M, 2010). Whilst toxicity is a Whilst significant progress has been made over the past
major cause of compound failure, the failure of potential 10–15 years in reducing attrition due to poor ADME prop-
drug molecules to demonstrate efficacy upon clinical eval- erties, toxicity remains to be a major source of attrition
uation is a more difficult issue to address since the factors and the chemist has an important role to play in reducing
which contribute to these failures are complex. Inadequate the late-stage attrition due to compound-related toxicity.
target validation, poor predictability of animal models, Whilst the current predictive packages are imperfect, there
poor clinical trial design and patient selection are poten- is a developing understanding of the impact of chemical
tial causes. Whilst the medicinal chemist cannot address structure and physicochemical properties on toxicity. It is
all of these issues, it is important that potent, selective and anticipated that an increasing number of screening assays
safe molecules with appropriate physical properties are will become available for incorporation into project
identified for such studies to ensure that it is the mecha- screening cascades which will help to identify potential
nism of action of the drug that is being tested and not the risks as early as possible in the discovery process.
quality of the molecule. Greater involvement of the medic- Thus, whilst the industry is entering uncharted territory,
inal chemist at the initial target identification and valida- there is an unprecedented opportunity for the medicinal
tion stage may provide better tool or probe compounds chemist to make a major contribution to the future success
(see Target selection and validation section) with which to of the discovery of new and important medicines to meet
validate novel targets. the growing level of unmet clinical need.

REFERENCES

Antoine DJ, Williams DP, Kipar A, vivo. Toxicological Sciences 2009; Baringhaus KH, Matter H. Efficient
et al. High-mobility group 112:521–31. strategies for lead optimization by
box-1 protein and keratin-18, Balls M. Adult human stem cells and simultaneously addressing affinity,
circulating serum proteins toxicity – realising the potential. selectivity and pharmacokinetic
informative of acetaminophen- Alternatives to Laboratory Animals parameters. Methods and Principles
induced necrosis and apoptosis in 2010;38:91. in Medicinal Chemistry 2005;

131
Section | 2 | Drug Discovery

23(Chemoinformatics in Drug Cole PA. Chemical probes for histone- Frye SV. The art of the chemical probe.
Discovery):333–79. modifying enzymes. Nature Nature Chemical Biology
Bartoschek S, Klabunde T, Defossa E, Chemical Biology 2008;4: 2010;6:159–61.
et al. Drug design for G-protein- 590–7. Funk C, Ponelle C, Scheuermann G,
coupled receptors by a ligand-based Congreve M, Chessari G, Tisi D, et al. et al. Cholestatic potential of
NMR method. Angewandte Chemie, Recent developments in fragment- troglitazone as a possible factor
International Edition 2010;49:1426– based drug discovery. Journal of contributing to troglitazone-induced
9, S1426/1–S1426/5. Medicinal Chemistry hepatotoxicity: in vivo and in vitro
Bass AS, Cartwright ME, Mahon C, et al. 2008;51:3661–80. interaction at the canalicular bile salt
Exploratory drug safety – a drug Dib-Hajj SD, Cummins TR, Black JA, export pump (BSEP) in the rat.
discovery strategy to reduce attrition et al. From genes to pain: Nav1.7 Molecular Pharmacology
in development. Journal of and human pain disorders. Trends 2001;59:627–35.
Pharmacological and Toxicological in Neurosciences 2007;30:555–63. Giannetti AM, Koch BD, Browner MF.
Methods 2009;60:69–78. Dykens JA, Will Y. The significance of Surface plasmon resonance based
Blagg J. Structure-activity relationships mitochondrial toxicity testing in assay for the detection and
for in vitro and in vivo toxicity. drug development. Drug Discovery characterization of promiscuous
Annual Reports in Medical Today 2007;12:777–85. inhibitors. Journal of Medicinal
Chemistry 2006;41:358. Dykens JA, Will Y. Drug-induced Chemistry 2008;51:574–80.
Bleicher KH, Boehm HJ, Mueller K, mitochondrial dysfunction: an Gluckman PD, Hanson MA, Cooper C,
et al. A guide to drug discovery: emerging model of idiosyncratic et al. In utero and early-life
hit and lead generation: beyond drug toxicity. International Drug conditions and adult health and
high-throughput screening. Nature Discovery 2010;5:32–6. disease – in reply. New England
Reviews Drug Discovery Eglen RM, Gilchrist A, Reisine T. The Journal of Medicine 2008;359:61–3.
2003;2:369–78. use of immortalized cell lines in Greer ML, Barber J, Eakins J, et al. Cell
Caldwell GW, Yan Z. Screening for GPCR screening: the good, bad and based approaches for evaluation of
reactive intermediates and toxicity ugly. Combinatorial Chemistry & drug induced liver injury. Toxicology
assessment in drug discovery. High Throughput Screening 2010;268:125.
Current Opinion in Drug Discovery 2008;11:560–5. Grover A, Benet LZ. Effects of drug
& Development 2006;9:47–60. Ejskjaer N, Vestergaard ET, Hellström transporters on volume of
Cao X, Lee YT, Holmqvist M, et al. PM, et al. Ghrelin receptor agonist distribution. AAPS Journal
Cardiac ion channel safety profiling (TZP-101) accelerates gastric 2009;11:250–61.
on the IonWorks Quattro automated emptying in adults with diabetes and Halai R, Craik DJ. Conotoxins: natural
patch clamp system. Assay and Drug symptomatic gastroparesis. product drug leads. Natural Product
Development Technologies Alimentary Pharmacology and Reports 2009;26:526–36.
2010;8:766–80. Therapeutics 2009;29:1179–87. Hancox JC, McPate MJ, El Harchi A,
Carr R, Jhoti H. Structure-based Ekroos M, Sjögren T. Structural basis for et al. The hERG potassium channel
screening of low-affinity compounds. ligand promiscuity in cytochrome P and hERG screening for drug-
Drug Discovery Today 2002;7:522–7. 450 3A4. Proceedings of the induced torsades de pointes
Chang G, Steyn SJ, Umland JP, et al. National Academy of Sciences of the Pharmacology & Therapeutics
Strategic use of plasma and United States of America 2008;119:118–32.
microsome binding to exploit in 2006;103:13682–7. Hart CP. Finding the target after
vitro clearance in early drug Evans DC, Watt AP, Nicoll-Griffith DA, screening the phenotype. Drug
discovery. ACS Medicinal Chemistry et al. Drug-protein adducts: an Discovery Today 2005;10:513–9.
Letters 2010;1:50–3. industry perspective on minimizing Harvey AL. Natural products in drug
Cheeseright TJ, Mackey MD, Melville JL, the potential for drug bioactivation discovery. Drug Discovery Today
et al. FieldScreen: virtual screening in drug discovery and development. 2008;13:894–901.
using molecular fields. Application Chemical Research in Toxicology Harvey AL, Clark RL, Mackay SP, et al.
to the DUD data set. Journal of 2004;17:3–16. Current strategies for drug discovery
Chemical Information and Modeling Foti RS, Wienkers LC, Wahlstrom JL. through natural products. Expert
2008;48:2108–17. Application of cytochrome P450 Opinion on Drug Discovery
Chu V, Einolf HJ, Evers R, et al. In drug interaction screening in drug 2010;5:559–68.
vitro and in vivo induction of discovery. Combinatorial Chemistry Helstroem PM. Faces of ghrelin
Cytochrome P450: A survey of & High Throughput Screening – research for the 21st century.
the current practices and 2010;13:145–58. Neurogastroenterology and Motility
recommendations: a pharmaceutical Fox S, Farr-Jones S, Sopchak L, et al. 2009;21:2.
research and manufacturers of High-throughput screening: update Hewitt P, Herget T. Value of new
America perspective. Drug on practices and success. Journal of biomarkers for safety testing in drug
MetAbolism and Disposition biomolecular screening development. Expert Review of
2009;37:1339. 2006;11:864–9. Molecular Diagnostics 2009;9:531–6.

132
 The role of medicinal chemistry in the drug discovery process Chapter |9|

Howard ML, Hill JJ, Galluppi GR, et al. chemistry. Nature Reviews Drug drugs? A historical perspective.
Plasma protein binding in drug Discovery 2007;6:881–90. Journal of Chemical Information
discovery and development. Lewis D, Ito Y. Human CYPs involved and Computer Sciences
Combinatorial Chemistry & High in drug metabolism, structures, 2001;41:1308–15.
Throughput Screening substrates and binding affinities. Orita M, Warizaya M, Amano Y, et al.
2010;13:170–87. Expert Opinion on Drug Metabolism Advances in fragment-based drug
Hughes B. 2009 FDA drug approvals. and Toxicology 2010;6:661–74. discovery platforms. Expert Opinion
Nature Reviews Drug Discovery Lewis DFV, Lake BG, Dickins M. on Drug Discovery 2009;4:1125–44.
2010;9:89–92. Quantitative structure-activity Packeu A, Wennerberg M, Balendran A.
Jamieson C, Moir EM, Rankovic Z, et al. relationships (QSARs) in inhibitors Estimation of the dissociation rate
Medicinal chemistry of hERG of various cytochromes P450: the of unlabelled ligand-receptor
optimizations: highlights and importance of compound complexes by a ‘two-step’
hang-ups. Journal of Medicinal lipophilicity. Journal of Enzyme competition binding approach.
Chemistry 2006;49:5029–46. Inhibition and Medicinal Chemistry British Journal of Pharmacology
Jeffrey P, Summerfield S. Assessment of 2007;22:1–6. 2010;161:1311–28.
the blood–brain barrier in CNS drug Lindsley CW, Weaver D, Bridges TM, Pepys MB, Herbert J, Hutchinson WL,
discovery. Neurobiology of Disease et al. Lead optimization in drug et al. Targeted pharmacological
2010;37:33–7. discovery. Wiley Encyclopedia of depletion of serum amyloid P
Kaczorowski GJ, McManus OB, Priest Chemical Biology 2009;2:511–9. component for treatment of
BT, et al. Ion channels as drug Lipinski CA, Lombado F, Dominy BW, human amyloidosis. Nature
targets: the next GPCRs. Journal of et al. Experimental and (London, United Kingdom)
General Physiology computational approaches to 2002;417:254–9.
2008;131:399–405. estimate solubility and permeability Pepys MB, Hirschfield GM, Tennent GA,
Kalgutkar AS, Didiuk MT. Structural in drug discovery and development et al. Targeting C-reactive protein for
alerts, reactive metabolites, and settings. Advanced Drug Delivery the treatment of cardiovascular
protein covalent binding: how Reviews 1997;23:3–25. disease. Nature (London, United
reliable are these attributes as Lombardi D, Dittrich PS. Advances in Kingdom) 2006;440:1217–21.
predictors of drug toxicity? microfluidics for drug discovery. Pirmohamed M, Breckenridge AM,
Chemistry & Biodiversity Expert Opinion on Drug Discovery Kitteringham NR, et al. Adverse drug
2009;6:2115–37. 2010;5:1081–94. reactions. British Medical Journal
Kaplowitz N. Idiosyncratic drug Luheshi LM, Crowther DC, Dobson (Clinical Research Edition)
hepatotoxicity. Nature Reviews Drug CM. Protein misfolding and disease: 1998;316:1295–8.
Discovery 2005;4:489–99. from the test tube to the organism. Polak S, Wisniowska B. Brandys
Keiser M, Wetzel S, Kumar K, et al. Current Opinion in Chemical Collation, assessment and analysis of
Biology-inspired synthesis of Biology 2008;12:25–31. literature in vitro data on hERG
compound libraries. Cellular and Macarron R, Banks MN, Bojanic D, receptor blocking potency for
Molecular Life Sciences et al. Impact of high-throughput subsequent modeling of drugs
2008;65:1186–201. screening in biomedical research. cardiotoxic properties. Journal of
Keserü GM, Makara GM. The influence Nature Reviews Drug Discovery Applied Toxicology 2009;29(3):
of lead discovery strategies on the 2011;10:188–95. 183–206.
properties of drug candidates. Nature McGuire C, Lee J. Brief review of Pollard CE, Abi Gerges N, Bridgland-
Reviews Drug Discovery vorinostat. Clinical Medicine Taylor MH, et al. An introduction to
2009;8:203–12. Insights: Therapeutics 2010;2:83–8. QT interval prolongation and
Kola I, Landis J. Opinion: can the Nagai N. Drug interaction studies on non-clinical approaches to assessing
pharmaceutical industry reduce new drug applications – current and reducing risk. British Journal of
attrition rates? Nature Reviews Drug situation and regulatory view in Pharmacology 2010;159:12–21.
Discovery 2004;3:711–6. Japan. Drug Metabolism and Popescu I, Fleshner PR, Pezzullo JC,
Konteatis ZD. In silico fragment-based Pharmacokinetics 2010;25:3–15. et al. The ghrelin agonist TZP-101
drug design. Expert Opinion on Naik P, Murumkar P, Giridhar R, et al. for management of postoperative
Drug Discovery 2010;5:1047–65. Angiotensin II receptor type 1 (AT1) ileus after partial colectomy: a
selective nonpeptidic antagonists – a randomized, dose-ranging, placebo-
Lai Y, Sampson KE, Stevens JC.
perspective. Bioorganic & Medicinal controlled clinical trial. Diseases of
Evaluation of drug transporter
Chemistry 2010;18:8418–56. the Colon and Rectum
interactions in drug discovery and
2010;53:126–34.
development. Combinatorial Newman DJ, Cragg JM. Natural
Chemistry & High Throughput products as sources of new drugs Price DA, Armour D, de Groot M, et al.
Screening 2010;13:112–34. over the last 25 years. Journal of Overcoming HERG affinity in the
Natural Products 2007;70:461–77. discovery of the CCR5 antagonist
Leeson PD, Springthorpe B. The
maraviroc. Bioorganic & Medicinal
influence of drug-like concepts on Oprea TI, Davis AM, Teague SJ, et al. Is
Chemistry Letters 2006;16:4633–7.
decision-making in medicinal there a difference between leads and

133
Section | 2 | Drug Discovery

Price DA, Blagg J, Jones L, et al. Shi Y, Merdan T, Li LC. Recent advances Discovery & Development
Physicochemical drug properties in intravenous delivery of poorly 2001;4:55–9.
associated with toxicological water-soluble compounds. Expert Ulrich RG. Idiosyncratic toxicity: a
outcomes – a review. Expert Opinion Opinion on Drug Delivery convergence of risk factors. Annual
on Drug Metabolism & Toxicology 2009;6:1261–82. Review of Medicine 2007;58:17–34.
2009;5:921–31. Shiau AK, Massari ME, Ozbal CC. Back Waring MJ. Lipophilicity in drug
Proudfoot JR. Drugs, leads, and to basics: label-free technologies for discovery. Expert Opinion on Drug
drug-likeness: an analysis of some small molecule screening. Discovery 2010;5:235–48.
recently launched drugs. Bioorganic Combinatorial Chemistry & High Wehling M. Assessing the translatability
& Medicinal Chemistry Letters Throughput Screening of drug projects: what needs to be
2002;12:1647–50. 2008;11:231–7. scored to predict success? Nature
Rees DC, Congreve M, Murray CW, Sloan KB, Wasdo SC, Rautio J. Design Reviews Drug Discovery 2009;8:
et al. Fragment-based lead discovery. for optimized topical delivery: 541–6.
Nature Reviews Drug Discovery prodrugs and a paradigm change. Westby M, van der Ryst E. CCR5
2004;3:660–72. Pharmaceutical Research antagonists: host-targeted antivirals
Reifferscheid G, Buchinger S. Cell-based 2006;23:2729–47. for the treatment of HIV infection.
genotoxicity testing. Genetically Smith GF. Medicinal chemistry by the Antiviral Chemistry & Chemotherapy
modified and genetically engineered numbers: the physicochemistry, 2005;16(6):339–54.
bacteria in environmental thermodynamics and kinetics of Wester K, Jönsson AK, Spigset O, et al.
genotoxicology. Advances in modern drug design. Progress in Incidence of fatal adverse drug
Biochemical Engineering/ Medicinal Chemistry 2009;48:1–29. reactions: a population based study.
Biotechnology 2010;118(Whole Cell Source http://en.wikipedia.org/wiki/ British journal of Clinical
Sensing Systems II) 85–112. List_of_withdrawn_drugs Pharmacology 2008;65:573–9.
Riley RJ, Grime K, Weaver R. Time- Swanson R, Beasley JR. Pathway-specific, Wetzel S, Klein K, Renner S, et al.
dependent CYP inhibition. Expert species, and sub-type Interactive exploration of chemical
Opinion on Drug Metabolism & counterscreening for better GPCR space with scaffold hunter. Nature
Toxicology 2007;3:51–66. hits in high throughput screening. Chemical Biology 2009;5: 581–3.
Rishton GM. Reactive compounds Current Pharmaceutical White RE. Pharmacokinetics of drug
and in vitro false positives in HTS. Biotechnology 2010;11:757–63. candidates. Wiley Encyclopedia of
Drug Discovery Today 2003;8: Takahara A, Sasaki R, Nakamura M, Chemical Biology 2009;3:652–61.
86–96. et al. Clobutinol delays ventricular Whittaker M, Law RJ, Ichihara O, et al.
Ritchie TJ, Luscombe CN, Macdonald repolarisation in the guinea pig heart Fragments: past, present and future.
SJ. Analysis of the calculated – comparison UIT cardiac effects of Drug Discovery Today: Technologies
physicochemical properties of hERG K+ channel inhibitor E-4031. 2010;7:E163–E171.
respiratory drugs: can we design for Journal of Cardiovascular
Windass AS, Lowes S, Wang Y, et al. The
inhaled drugs yet? Journal of Pharmacology 2009;54:552–9.
contribution of organic anion
Chemical Information and Modeling Tate CG, Schertler GFX. Engineering G transporters OAT1 and OAT3 to the
2009;49:1025–32. protein-coupled receptors to renal uptake of rosuvastatin. Journal
Schmidtko A, Lötsch J, Freynhagen R, facilitate their structure of Pharmacology and Experimental
et al. Ziconotide for treatment of determination. Current Opinion Therapeutics 2007;322:1221–7.
severe chronic pain. Lancet in Structural Biology 2009;19:
Witchel HJ. Emerging trends in ion
2010;375:1569–77. 386–95.
channel-based assays for predicting
Schneider G. Virtual screening: an Uetrecht J. Prediction of a new drug’s the cardiac safety of drugs. IDrugs
endless staircase? Nature Reviews potential to cause idiosyncratic 2010;13:90–6.
Drug Discovery 2010;9:273–6. reactions. Current Opinion in Drug

134
 Chapter 10 
Metabolism and pharmacokinetic optimization
strategies in drug discovery
P Ballard, P Brassil, K H Bui, H Dolgos, C Petersson, A Tunek, P J H Webborn

reach the target organ at high enough concentration to

INTRODUCTION engage the target.
In this chapter, strategies to optimize key DMPK chal-
Optimization of drug metabolic and pharmacokinetic lenges using appropriate in silico, in vitro and in vivo
properties is an integral component of the modern drug DMPK tools during drug discovery are presented. The
discovery process. The objective of the drug metabolism rational use of these strategies will help ‘drug hunting’
and pharmacokinetics (DMPK) discipline in drug discov- projects to advance drug candidates with attractive DMPK
ery is to aid design and selection of candidate drugs with target profile and with low potential for failure in develop-
properties that yield the required efficacy and safety for ment due to DMPK issues. In addition, as prediction of
effective clinical use. The roles of DMPK at the various human PK and safe and effective dose is probably the most
stages of drug discovery are summarized in Table 10.1. important activity in drug discovery to ensure that the
DMPK in vitro and in vivo information are used through- candidate drug has the attribute to test the biological
out the drug discovery process to facilitate target validation hypothesis in patients, strategy to integrate information in
and safety assessment, and to guide the conversion of early discovery to holistically predict PK properties in man will
screening hits and leads into drug candidates. Indeed, the be discussed.
frontloading of DMPK in drug discovery has resulted in a
reduction of drug attrition rate due to undesirable DMPK
properties from approximately 40% in 1990 to 10% in OPTIMIZATION OF  
2000 (Kola and Landis, 2004). DMPK PROPERTIES
To help drug hunting teams focus on the key issues and
goals, in a multitude of screening options, it is important
Optimization principles are described for six key DMPK
to prescribe, as early as possible, a candidate drug target
areas:
profile in terms of efficacy, potency, safety and ease of
use. DMPK plays a central role in defining this target • Absorption and bioavailability
profile. For instance to be commercially attractive in terms • Avoiding PK-based drug–drug interactions
of ease of use, a compound has to be orally active, has • Achieving/avoiding CNS exposure
a convenient dosing regimen and be able to be adminis- • Clearance
tered without effect from food and other medications. The • Role of metabolite identification:
physicochemical and DMPK attributes that will allow a  Active metabolites

compound to meet this target profile would be: good  Minimizing risk for reactive metabolites.

solubility and permeability, high oral bioavailability, low The following sections summarize current understanding/
clearance and reasonable half-life (if PD half-life is not available tools and suggest best practice for each of the
much longer than PK half life), and absence of ‘drug–drug above. Each section introduces the challenges, outlines
interaction’ potential. Likewise to be orally active, a com- tactics for dealing with them and identifies areas requiring
pound should have good oral bioavailability and able to caution.

© 2012 Elsevier Ltd. 135

Section | 2 | Drug Discovery

Table 10.1  Roles of drug metabolism and pharmacokinetics (DMPK) in various phases of drug discovery

Discovery phase DMPK roles

Target identification Selection and characterization of tool compounds
Partner with in vivo pharmacology and toxicology in target validation and safety assessment
activities
Hit identification In silico and in vitro DMPK profiling to help prioritize hit series
Lead identification Identify DMPK liabilities of lead series
Determine whether if DMPK properties in lead series are optimizable (DMPK liabities not linked to
pharmacophore)
Develop structure (DMPK) property relationship (SPR)
Develop ‘in vitro in vivo’ correlation (IVIVC)
Guide selection of lead series
Contribute to development of PD assay and develop early PK/PD understanding
Contribute to development of PD assays and develop early PK/PD understanding
Lead optimization Guide optimization of DMPK properties toward target profile
Comprehensive DMPK characterization of candidate compounds
Develop biomarker based and/or efficacy/disease related PK/PD models
Prediction of human PK
Prediction of efficacious human dose
Integration of predicted profile to calculate key margins against side effects

Absorption and oral bioavailability In humans, the combinations of high to low solubility
and permeability have led compounds to be characterized
Introduction according to the Biopharmaceutical Classification System
As the preferred route of administration for most indica- (Amidon et al., 1995). Class 1 compounds, with high
tions is oral, it is important to characterize oral bioavaila- solubility and permeability, generally have very good
bility (F) of a compound during drug discovery. In addition, absorption properties. Those in Class 4, with poor solubil-
F must be optimized, as a low F is often associated with ity and permeability, are likely to present significant for-
poor and variable exposure and lack of efficacy. F is defined mulation challenges and/or variable and poor exposure. It
as the percentage of dosed drug that reaches the systemic can be important to characterize the maximum absorbable
circulation compared to the IV route. As shown below, it dose (MAD) of a compound relative to its predicted thera-
can be considered to be dependent on three serial steps: peutic dose, as this will determine the risk of being able
the fraction of dosed drug absorbed (fa), the fraction escap- to deliver an efficacious dose to humans and guide whether
ing intestinal metabolism (fg) and the fraction extracted by high exposure in safety studies is achievable.
the liver as it passes from the portal vein to the systemic Table 10.2 gives guidance on acceptable pharmaceutical
circulation (fh) (see Rowland and Tozer, 1989): properties for typical oral drug candidates.
F = fa × f g × fh
fa is influenced by a number of factors including the Tactics
gastrointestinal (GI) solubility (dose, particle size, pH
In theory, it is relatively simple to obtain good absorp-
solubility profile and formulation), the effective permea-
tion and to ensure that solubility and permeability fall
bility (both passive permeability and active transport
within the right ranges. However, the reality is more
processes) and GI stability. fg and fh are affected by meta-
complex. From an in vivo perspective, the product of
bolic enzymes in the intestinal wall and liver, respectively.
absorption and intestinal metabolism is often assessed
In addition, fh can be influenced by transporters if a
by accounting for first-pass hepatic clearance in bioavail-
drug is excreted unchanged into the bile. Both metabolic
ability estimations:
and active transport processes are saturable, generally
obeying Michaelis–Menten kinetics. Hence, fa, fg and fh f a × f g = F/f h
can all be non-linear if relevant concentrations are above If fa x fg is low, it is important to understand the relative
the Michaelis–Menten constant (Km) for the particular contribution of both fa and fg to F so that this can be
enzyme/transporter–drug interaction. designed out of the project. Generally it is simplest to

136
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

Table 10.2  Pharmaceutical development candidate drug target profile (CDTP)

Property CDTP Impact on pivotal safety Impact of not meeting on

studies and initial clinical later drug development
evaluations if not met timelines
Crystallinity Crystalline Significant risk of non-robustness Significant issues with
of manufacture and robustness of manufacture,
Thermal property No melt or other event
performance of drug substance performance and ease of
< 80°C
with the potential for variable handling of both drug
Hygroscopicity < 2% uptake at exposure in safety and clinical substance and drug product
25 °C/80% relative studies due to physical instability Potential of compromising
humidity Crystalline material required for clinical studies due to physical
assessment of formulation instability as drug substance
approaches and exposure and in drug product requiring
additional time and resources
Chemical stability >90% (in solution Insufficient stability for SA studies. Consideration of the learning in
formulation after at More time and resources needed phase 1 and assess impact
least 1 day at room to understand degradation and and options in line with the
temperature) mitigate in line with the formulation strategy. Risk of
formulation strategy (storage, short shelf-life of the product
pack or formulation) for safety
and first time in man studies
Maximum absorbable MAD ≥ predicted dose Significant risk for not reaching High risk for not reaching
dose (MAD) to man required exposures in early clinically relevant exposures
clinical and safety studies. More using conventional
Predicted fraction Fa ≥ 50 %
time and resources needed to formulations. Strategy
absorbed (fa) at
develop enhancing formulations with non-conventional
predicted dose to
as well as a risk of failure formulations will require
man
more time and resources in
development. Obvious risk of
development failure

estimate the likely fa and if this does not account for The primary in vitro tool for assessing absorption is the
the poor fa x fg value, fg should be investigated. Poor cell based Caco-2 permeability assay, although MDCK-
absorption can be a result of slow dissolution rate, low MDR1, PAMPA (parallel artificial membrane permeability
solubility in the GI tract, poor effective permeability assay), or in silico predictions may also provide valuable
(passive or active efflux), or instability in GI fluids or information about efflux transporter risk and permeabil-
in the wall of the GI tract. If absorption is adequate but ity. This permeability assessment, in combination with a
bioavailability is poor, hepatic clearance (metabolic or solubility measurement (ideally using crystalline mate-
biliary elimination) and/or intestinal metabolism may rial), is used to estimate fa using commercially available
need to be optimized. modelling tools like GastroPlus (www.simulations-plus
When maximizing the chances of good absorption, the .com) or SIMCYP (www.simcyp.com). For actively trans-
starting point is to ensure that the physicochemical prop- ported (efflux or uptake) compounds, bi-directional per-
erties of the compound/series are in the optimal space as meability assays can be used as a guide to possible in vivo
described by Lipinski and others (Lipinski et al., 1997; effects. However, it should be noted that it is often difficult
Wenlock et al., 2003; Johnson et al., 2009; Waring, 2009). to extrapolate the results from these in vitro transport
Generally, this requires minimizing the number of H-bond assays to accurately quantify effects in vivo. Because of its
donors and acceptors, restricting lipophilicity in the range reasonable throughput, the Caco-2 assay can be posi-
LogD7.4 0 to 3, and limiting molecular weight to <500. tioned as an early screen if absorption or permeability/
Both the solubility and passive permeability of a com- efflux is found to be an issue in the project. If further
pound should be assessed prior to any in vivo study. Pre- evaluation of absorption or efflux is warranted, it is pos-
dictive models for these should be assessed for suitability sible to use more physiological models such as sections of
in each project and considered in compound design if intestinal tissue in an Ussing Chamber (Ungell et al.,
either is an issue for a chemical series. 1998), or transfected cell lines over-expressing particular

137
Section | 2 | Drug Discovery

• Predicted fraction absorbed (fa) acceptable based on solubility and permeability (predicted or measured)
• Rat CL <50% LBF

Rat fa <30% Caco-2 Dog*

oral PK ABBA PK
• Paap predicts good absorption
• Progress if compound of
sufficient interest
Subsequent compounds:
fa >30% LI Establish IVIVC/SAR with
Paap low fa >50%
fa >50% LO Caco-2 to improve rat fa

Stop and investigate Progress if acceptable human fa, MAD and exposure in 2 toxicology
Progress compounds
issues and SAR species predicted
Subsequent compounds: Subsequent compounds:
Assess directly in rat PK Use Caco-2 to select for
as necessary rat PK Yes No

Consider other risks Stop

(e.g. intestinal metabolism
* or other acceptable non-rodent species or human Ussing GI fluid stability)

Fig. 10.1  Absorption troubleshooting decision tree. Caco-2 ABBA, transport assays; CL, clearance; fa, fraction of dose
absorbed; GI, gastro-intestinal; IVIVC, in vitro to in vivo correlation; LBF, liver blood flow; LI, lead identification; LO,
optimization; MAD, maximum absorbable dose; Papp, apparent permeability; SAR, structure–activity relationship.

efflux transporters (e.g. P-gp in the MDCK-MDR1 cell available, it is difficult to make an accurate
line). The Ussing Chamber technique can help in under- quantitative assessment of the contribution of
standing cross-species differences and, because the tissue intestinal metabolism to in vivo fa. However, attempts
used is enzymatically competent (metabolic and trans- have been made to use CLint from human liver
porters), the output represents the product of fa and fg. microsomes or S9 fraction to estimate the relative
Compounds with low hepatic clearance in the rat, good contributions of fa and fg to F (Gertz et al., 2010).
solubility and high effective permeability should exhibit • It is often very difficult to pinpoint why compounds
good oral absorption and bioavailability in that species. have poor absorption characteristics and, therefore,
However, if this is not the case, the troubleshooting deci- to resolve this design issue. Thus, it is often
sion tree in Figure 10.1 can be used to help determine the reasonable to prioritize series with good fa x fg, in
cause(s) of poor absorption, identify assays to aid in opti- early discovery even if other properties (e.g. potency,
mizing compound design and understand if the com- clearance) are less attractive.
pound is of sufficient quality to progress in the value • Typically, oral doses are formulated as suspensions
chain, despite its non-optimal absorption properties. which, on many occasions, may be derived from
Table 10.3 is an aid to selecting the assays and tech- amorphous material. However, it is important to
niques to use in the decision tree (Figure 10.1), to explain assess absorption periodically using crystalline
potential issues and risks, and the parameter (fa, fg and/or material, as physical form may have substantial
fh) the assays impact on. effects on absorption profiles.
• Particularly for compounds likely to proceed into
development, it is important to determine the effect
Cautions of the polymorphic solid states on absorption. It is
• As scaling factors for intestinal microsomes or other also important to ensure that the formulations used
subcellular intestinal fractions are currently not are discussed with Pharmaceutical Development (see

138
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

Table 10.3  Assays and techniques used when troubleshooting absorption

Assay/technique Potential issue/risk addressed Impacted

Intestinal microsomes or S9 fraction Assess cross species differences in intestinal metabolism. Nature fg
and source of metabolites can give key information about
potential for gut metabolism, as CYP3A and UGTs account for
most gut metabolites
Absorption profiling on Caco-2 cells Varying apical to basolateral pH gradient fa
Concentration dependency
Use of proteins (e.g. BSA) at various percentages in apical chamber
Use of efflux transporter inhibitors
PAMPA Passive permeability fa
High dose PK studies Saturation of efflux or metabolism fa, fg, fh
Ussing chamber technique Effective permeability (including transporters) fa, fg
Intestinal tissue metabolism
GI stability test Degradation of drug in stomach or intestinal lumen is possible fa
explanation (usually the case if predicted F much greater than
measured F)
Human metabolic phenotyping Gut metabolism fg, fh
In situ/in vivo portal vein cannulation Determine amount of drug and metabolites passing through fa, fg
preparation intestine
SIMCYP; GastroPlus Predict absorption rate and extent (software methods) fa
Transfected cell lines; vesicles Determine involvement of specific efflux transporters fa, fh
expressing specific transporters;
Caco-2 efflux assay with and
without specific transporter inhibitors
Knock-out (KO) animals; chemical Determine involvement of specific efflux transporters fa, fg, fh
KO with inhibitors

Chapter 16), and are appropriate for safety and early of PK-based DDI, in which the compound may be a per-
clinical development studies. petrator or be a victim of DDI:
• If in vitro and in vivo (rat and dog) assessments • Competitive (reversible) cytochrome P450 (CYP)
of fa do not agree, the risk of an inaccurate estimate inhibition
of absorption in man will increase. Sometimes • Mechanism based/time dependent CYP inhibition
other species have been investigated to mitigate this • Uptake and efflux transporter inhibition
risk, but we caution that there can, for example, • CYP induction.
be marked discrepancies in fa x fg between
CYP, notably CYP3A4 and CYP2D6, based DDI is the
cynomolgus monkeys and humans (Takahashi
most important and most common, and may occur in
et al., 2009).
the liver or intestine. Transporter based DDI is mainly
related to renal clearance, although specific issues can
Avoidance of PK based   arise with CNS compounds and hepatic uptake of statins.
drug–drug interactions The science and regulatory guidance to support risk assess-
ment of CYP-based DDI are well established, but is less
Introduction advanced for transporter-related issues (Bjornsson et al.,
Aid in the design and selection of candidate drugs with a 2003; US Food and Drug Administration, 1997, 2006;
low potential for PK-based drug–drug interactions (DDI) Huang et al., 2008; International Transporter Consortium,
is a key role of discovery DMPK. There are four main forms 2010).

139
Section | 2 | Drug Discovery

Tactics
Competitive (reversible) CYP inhibition IC50 <20µm for any CYP in fluorescence based assay

Two main types of CYP inhibition assays with different

Yes No
capabilities are in general use. Fluorescence based assays
are relatively cheap and have enhanced throughput, but in
a small but significant number of cases can lead to mis- IC50 < 20 µm with No Low risk –
representing DDI risk (Bell et al., 2008). They are best drug-like substrates stop further evaluation
used for initial profiling of large numbers of compounds,
with the data being acceptable in the early phase such as Fu,mics Yes No
lead generation. LCMS based assays are more expensive,
have lower throughput, but are more predictive. They
should be used in optimization cycles once CYP inhibition
Predicted C max/Ki,u > 0.1
issues have been identified, and for generating compound
profiles during more advanced project phases.
Inhibition of the five major CYP isoforms 1A2, 2C9, Yes
2C19, 2D6 and 3A4, should be evaluated in the earliest
phases, while later it would be prudent to assess potential
Evaluate with tools
interactions with isoforms 2B6, 2C8 and 3A5. AUC ratio >2
like SIMCYP
Reduction of CYP inhibition potential is facilitated by
the fact that strong QSAR relationships are often obtained.
Yes
Various computational models that allow prediction of
CYP DDI risk are available within most drug development
companies. It is well established that lipophilicity, aroma- Potential clinical risk
ticity and charge type are major drivers for inhibiting
various CYP enzymes (Gleeson et al., 2007).
The risk of DDIs based on Phase 2 metabolism (e.g.
Fig. 10.2  Decision tree for reversible DDI CYP inhibition.
glucuronidation and sulphation) is usually small, resulting
AUC, area under the concentration-time curve; Cmax,
in less than a two-fold increase in area under the con­ maximum concentration; CYP, cytochrome P450;
centration versus time curve (AUC), and they are rarely DDI, drug-drug interaction; fu,mics, fraction unbound in
observed, possibly due in part to the nature of the enzy- microsomes); IC50, concentration producing 50% inhibition;
matic reaction (high Vmax and moderate to high Km values). Ki,u, unbound dissociation rate constant; SIMCYP, software
Such DDIs are not generally evaluated in lead optimization platform.
(Williams et al., 2004). Evidence suggesting the need to
do so at this stage should prompt re-evaluation of the risk.
The decision tree in Figure 10.2 can be used to assess a medium throughput screening assay can be used to
the potential DDI risk of a competitive CYP inhibitor. Hits screen for TDI. However, for selecting candidates at later
identified in a fluorescence-based assay should be con- phases, the method employed should provide accurate
firmed with an LCMS-based assay using druglike sub- determination of Kinact and Ki to properly evaluate the DDI
strates as probes for the different CYP isoforms. Although risks of compounds in which preliminary screening indi-
the ratio Cmax/Ki,u can be used to obtain a preliminary cated the potential for TDI. A positive TDI finding also
estimate of the DDI risk, more accurate evaluation should suggests that the compound or its metabolites may be
be conducted using PBPK modelling (e.g. SIMCYP plat- reactive, and further evaluation should be conducted as
form) to predict the potential clinical risk (expanded specified by reactive metabolite strategies.
below in ‘Prediction of DDI risk’ ”). A decision tree to help evaluate the potential DDI risk
of a TDI CYP inhibitor is shown in Figure 10.3. If a com-
pound is found to be at risk for CYP inhibition (IC50
Mechanism-based/time-dependent <20 µM) or is flagged to have a potential liability for reac-
CYP inhibition tive metabolites, it should be screened for TDI. If it is
The inhibition of CYP enzymes may be irreversible (due found to have potential TDI risk based on screening data,
to irreversible or covalent binding to the prosthetic haem Ki and Kinact values should be generated to help accurately
or the enzyme) or quasi-irreversible (due to the formation predict the risk using prediction tools like SIMCYP.
of transient complexes with the iron of the haem pros-
thetic group). Time-dependent inhibition (TDI) methods Uptake and efflux transporter inhibition
can be used to determine this (Riley et al., 2007; Fowler Uptake transporter inhibition assays are emerging as
and Zhang, 2008). During early phases of drug discovery, being of real value (Ward, 2008), but they are highly

140
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

IC50 <20 µm for any CYP in Reactive metabolite potential

fluorescence based assay (in vitro/structural alert)
Yes Yes

In vitro TDI assay

Determine mechanism
with K3Fe(CN)6
No TDI detected TDI detected displacement

No further investigation Make DDI predictions based on screening data:

use [I]/IC50 and Ki/Kinact plot
set Kinact = 0.2 min–1 and calculate Ki

Yes
Determine Ki and Kinact in vitro Potential DDI risk

No

Assess risk with tools like SIMCYP Screening data enough

Fig. 10.3  Decision tree to assess time dependent inhibition (TDI) risk. CYP, cytochrome P450; DDI, drug–drug interaction; Ki,
dissociation rate constant; Kinact, inactivation rate constant; [I], inhibitor concentration; IC50, concentration producing 50%
inhibition; SIMCYP, assay method; TDI, time dependent inhibition.

dependent on chemotype (e.g. acids being primarily trans- the liver or in the gut can be estimated based on its esti-
ported by organic anion transporters (OATs)) and likely mated local exposure and the IC50.
co-medications (e.g. OCT2 and metformin). Acids and
zwitterions should be assessed for inhibition of OATP1B1
during drug discovery. Other inhibition assays (OAT1,
Determination of clearance mechanism and
OAT3, OATP1B1, OCT1 and OCT2) should be used on a CYP phenotyping
case-by-case basis. The potential for a compound to be a victim of a DDI with
Efflux transporters are believed to serve a protective a co-medication is greatly reduced if there are multiple
function, and prevent molecules perceived as foreign from clearance mechanisms, particularly involving metabolism
gaining entry in cells or tissues. Of the various efflux trans- by multiple CYP enzymes. This should be a consideration
porters, P-glycoprotein (P-gp) is the most prevalent and if the therapeutic window is low or if the clinical/marketing
well understood. As specific locations of the P-gp trans- disadvantage of the interaction would be significant.
porter include small intestine enterocytes, hepatocytes, the Quantitative assessment of multiple clearance mecha-
kidney and the blood–brain barrier, P-gp can affect oral nisms and phenotyping of CYP metabolism should be
bioavailability, biliary and renal clearance, and brain established for any candidate drug. Methods for pheno-
uptake of compounds that are substrates of P-gp. In addi- typing the individual CYP enzymes responsible for a
tion, compounds that modulate P-gp can influence the drug’s metabolism include the use of (1) specific chemi-
clearance and distribution of drugs that are substrates of cals or antibodies as specific enzyme inhibitors, (2) indi-
P-gp. The bi-directional transport assay using either Caco-2 vidual human recombinant CYPs, and (3) a bank of
(P-gp, BCRP, and MRP2) or MDCK-MDRI (specifically for human liver microsomes characterized for CYP activity
P-gp) cells is widely available. Evaluation of the potential prepared from individual donor livers. At nomination of
to inhibit P-gp should be evaluated in early discovery a candidate drug, human phenotyping work should
using a bi-directional transport assay with a probe P-gp include all eight major CYPs (1A2, 2B6, 2C8, 2C9, 2C19,
substrate (e.g. digoxin). If a compound is found to be a 2D6, 3A4/3A5), and be followed by other enzymes if nec-
P-gp inhibitor, its potential impact for P-gp inhibition in essary. Another way to investigate clearance mechanisms

141
Section | 2 | Drug Discovery

in vitro is a so-called ‘fractional Clint’ assay. In such an Prediction of DDI risk

assay the rate of parent disappearance is measured in
Once inhibition (of CYP enzymes and/or transporters)
human hepatocytes with and without the presence of an
potential has been assessed in vitro, a risk assessment is
enzyme inhibitor. Most common is the addition of keto-
made by examining the data in relation to the likely clini-
conazole to block out the CYP3A4 contribution to metab-
cal exposures. At candidate drug nomination, a thorough
olism. The remaining rate of metabolism in such an assay
evaluation of CYP-based DDI risk (both reversible and
can be attributed non-CYP3A4 pathways like phase 2
irreversible) should be made using PBPK modelling (e.g.
enzymes, other CYPs or any other possible mechanism.
SIMCYP platform). PBPK modelling can be used to esti-
mate relevant concentration of inhibitors at the inlet to
CYP induction mediated risk for DDI the liver or in the gut and to assess DDI risks in various
Induction of specific CYP enzymes may not only change patient populations.
a drug’s metabolic profile but also have toxicological con- During early drug discovery, a simple criterion can be
sequences as CYP enzymes are also involved in the metab- used to determine if CYP inhibition presents little or no
olism and synthesis of important endogenous compounds. risk. For this purpose, an IC50 of >20 µM provides a reason-
Although close collaboration with Safety Assessment (see able cut-off. Should the predicted therapeutic exposures be
Chapter 15) is needed to evaluate the full impact of CYP very low, then a lower cut-off might be rationally adopted.
induction, it is DMPK’s primary responsibility to predict To assess CYP induction potential, the Emax and EC50
the DDI potential caused by CYP induction in man. This obtained from human hepatocytes (e.g. HepaRG cells)
should be done during optimization with the use of induction assays, can be used in conjunction with pre-
HepaRG cells which are a good surrogate of primary dicted human exposure (free Cmax or free liver inlet concen-
human hepatocytes for AhR-mediated CYP1A induction tration) to calculate a relative induction score. This is then
and PXR- and CAR-mediated CYP3A4 and CYP2B6 induc- compared against the relative induction scores of known
tion. If higher throughput is needed during the optimiza- inducers to estimate the percent human AUC change.
tion phase, the PXR reporter gene assay may be used to A compound is likely to be a victim of clinically relevant
minimize PXR-dependent CYP3A4 induction liability. DDI with a co-medication if it has a narrow therapeutic

Trigger to evaluate DDI risk for compound as substrate (victim)

High victim DDI sensitivity as Available data on minimum toxic

Yes
judged from understanding dose and minimum effective dose
of elimination mechanisms suggest narrow therapeutic range

No
No
Stop further investigation
Yes

Evaluate risk with in house Predicted Yes Evaluate for poor metabolizers
tool or SIMCYP AUC ratio >2 Look for outliers in population

No

Report
• Estimated AUC ratio
• Identified high risk groups if any

Fig. 10.4  Decision tree to assess risk as a drug–drug interaction (DDI) victim. AUC, area under the concentration-time curve;
CYP, cytochrome P450; fm, fraction metabolized; fu, fraction unbound; SIMCYP, software platform.

142
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

window and is a sensitive substrate to an inhibited and/or, more importantly, compounds with high affinity
enzyme/transporter, as indicated by high values of fraction for efflux transporters.
metabolized (fm), fraction of CYP metabolized (fm,CYP), The quantitation of CNS exposure is necessary for both
and fraction unbound (fu), and by plasma and hepatic CNS and peripherally targeted (i.e. that require restriction
extraction. This risk can be reduced by ensuring that the from the CNS) therapies as a means to estimate the
clearance mechanism in question represents <50% of total unbound brain concentration required for efficacy and
clearance (resulting in less than a two-fold change in the CNS-mediated side effects. Unless other information is
AUC of the victim compound). At candidate drug nomina- available, the unbound concentration is assumed to be the
tion, software tools such as SIMCYP could be used to relevant exposure measure for both desired and undesired
evaluate the impact of known inhibitors or inducers on pharmacological effects. The reason for this is that meas-
the relevant candidate compounds if their clearance is urement of total concentration can be very misleading,
driven mainly by a single enzyme or transporter. Figure being to a large extent driven by non-specific binding to
10.4 is a decision tree to help in assessing the potential brain tissue and thus strongly correlated with physico-
risk that a compound will be a DDI victim. A compound chemical properties (e.g. lipophilicity and charge). The
is deemed to have this potential risk if it has a narrow tactics below therefore focus on assessing unbound con-
therapeutic window or is cleared predominantly by a centration in the brain relative to the free concentration
single CYP enzyme. A simple Excel based DDI template or in blood or plasma.
SIMCYP should be used to estimate the risk. If the CYP
enzyme involved is polymorphic, evaluate the impact of
polymorphism to identify high-risk populations. Tactics
Screening cascades for both CNS targeted and CNS-
restricted compounds should contain an efflux/
Cautions permeability assay (Di et al., 2008). MDCK-MDR1 or
• Software tools such as SIMCYP can be used to Caco-2 cells may be used (Table 10.2) and MDCK-MDR1
estimate the potential and magnitude of DDI, may offer better sensitivity for P-gp substrates, while
particularly in populations at most risk. DDI risk is Caco-2 cells, which are a constitutive system, have the
currently best assessed by relating IC50 to free drug advantage of identifying substrates of several efflux mecha-
concentrations at the sites of action (liver or gut). nisms not necessarily limited to P-gp.
However, because of the current regulatory position, Peripheral restriction can be obtained either by efflux or
it may be necessary to conduct clinical studies where very low passive permeability. The latter is often difficult
the perceived risk is based on total drug to combine with oral pharmacological activity, as it
concentrations. requires extreme physical properties (e.g. very low/negative
• CYP based DDI due to CYP3A4 inhibition may have LogP).
an impact on the bioavailability of co-medications Efflux ratios obtained in vitro should be used for risk
through inhibition of intestinal CYP3A4. Although assessment in the early phases and once validated in vivo,
the area is complex, SIMCYP has the capability of can be used for progression of compounds through the
assessing likely clinical impact. screening cascade. The ratio may be optimized using in
• Animal studies are of little value for assessing silico models (Fridén et al., 2009b) or simple rule sets.
CYP DDI risk. However, they may be of use in Roughly, P-gp substrates can be estimated by the ‘rule of
developing an understanding of potential transporter fours’ (Didziapetris et al., 2003), where compounds with
mediated DDI. a sum of nitrogen and oxygen atoms ≥8, molecular weight
>400 and acid pKa >4 are likely to be P-gp substrates,
whereas compounds with a sum of nitrogen and oxygen
Central nervous system uptake atoms ≤4, molecular weight <400 and base pKa <8 are
likely to be non-substrates. Additionally, it has been sug-
Introduction gested that LogP should be >2 to allow for sufficient
The brain is a protected organ, being separated from the passive permeability, i.e. lower sensitivity to efflux (Hitch-
systemic circulation by three barriers: the blood–brain cock and Pennington, 2006).
barrier formed primarily by cerebrovascular endothelial Commencing during lead generation, it is important to
cells between the blood and brain tissue, the choroid establish an in vitro/in vivo correlation using in vivo
plexus epithelium between the blood and ventricular cer- experiments to determine the ratio Cu,br/Cu,pl (Table 10.4).
ebrospinal fluid (CSF), and the arachnoid epithelium The fraction unbound in brain may be determined ex vivo
between the blood and subarachnoid CSF. These barriers, using brain binding assays such as the brain homogenate
which exhibit low paracellular permeability and express or brain slice (Fridén et al., 2009a; Wan et al., 2007)
multiple drug transporters, restrict the entry of compounds methods and, together with determined total brain con-
with either very low transcellular (passive) permeability centration in vivo, enables the calculation of Cu,br.

143
Section | 2 | Drug Discovery

Table 10.4  Assay and CNS measurement attributes at various milestones for compounds requiring CNS access
or restriction

Phase Assay/study Attribute if access required to Attribute required for CNS

CNS restriction
Hit to lead Bidirectional Profile hit series for permeability/efflux Profile hit series for permeability/
Caco-2 assay liabilities efflux liabilities
Prioritize and select hit series based on Prioritize and select hit series
absence of liabilities based on absence of liabilities
Lead identification Bidirectional Determine optimization feasibility Determine optimization feasibility
Caco-2 assay
Early Evaluate usefulness of in silico tools Evaluate usefulness of in silico
and develop project specific in silico tools and develop project-specific
model in silico model
Develop series-specific CNS penetrant
tool compounds
Late Protein and brain Estimate free brain concentration to Estimation of free brain
binding assay help develop PD model and establish concentration to help develop a
IVIVC PD model and establish IVIVC
In vivo CNS Estimate free brain concentration to Estimation of free brain
distribution study help develop PD model and establish concentration to help develop a
IVIVC PD model and establish IVIVC
Lead optimization Bi-directional Optimize away from efflux substrates Optimize away towards efflux
Caco-2 assay Optimize free ratio Cu,br/Cu,pl at steady substrates
Protein and brain state towards unity, build IVIVC Optimize free ratio Cu,br/Cu,pl at
binding assays steady state away from unity, build
IVIVC
In vivo CNS Continue to build IVIVC
Continue to build IVIVC
distribution study Understand impact of efflux in vivo
In vivo P-gp Establish IVIVC
inhibitor/KO studies

Milestones Assay/study Attribute if access required to Attribute required for CNS

CNS restriction
CD nomination Bi-directional Assess risk of involvement of efflux Assess risk of involvement of efflux
Caco-2 assay transporters in human transporters in human (DDI focus)
In vivo P-gp Assess risk of involvement of efflux Assess risk of involvement of efflux
inhibitor/KO studies transporters in human transporters in human (DDI focus)
In vitro mechanistic Evaluate species difference in efflux if Evaluate species difference in
studies tools exist efflux if tools exist
Protein and brain For central targets, free brain
binding assays concentrations should be <3 times free
plasma concentrations at steady state*
In vivo CNS Understand potential interspecies Knowledge of pharmacodynamics
distribution study differences in plasma protein binding for CNS related side effects to
establish required ratio
Microdialysis/CSF To determine rates, and better To determine rates, and better
understand pharmacodynamics and understand pharmacodynamics
translation to man and translation to man
*Measurement of free ratios is a composite of one in vivo and two in vitro experiments. Total experimental error may be as high as
three-fold.
CSF, cerebrospinal fluid; DDI, drug–drug interaction; IVIVC, in vitro to in vivo correlation; KO, knock-out; PD, pharmacodynamic

144
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

Initial brain distribution studies may be performed with few data are available regarding the distribution of
single compounds or cassettes using subcutaneous or oral compounds within CSF.
administration, preferably at steady state or with more • It may be difficult to assess free levels of acidic
than one time point. Samples generated in pharmacody- compounds in CNS with standard methodologies
namic models may also be used. In late stages, longer (Fridén et al., 2010), because the high plasma
intravenous infusion studies may be used to generate a protein binding and low tissue binding of such
ratio closer to steady state. compounds result in inaccurate data due to blood
Failure to establish in vitro/in vivo correlations may contamination of the brain tissue. The brain
provide a trigger for investigations using: homogenate binding method has limited use for
• In vitro/in vivo interaction studies with inhibitors drugs that reside predominantly in the interstitial
• CNS distribution studies in alternative species to space or compounds that are accumulated
gain an understanding of differences in transporter intracellularly. In such cases, the brain slice method
expression/specificity may provide a better alternative.
• In vitro studies (e.g. alternate cell lines) with
over-expression of animal transporters or additional Clearance optimization
human transporters
• In vivo studies in knock-out (KO)/polymorph Introduction
animals.
Optimization of clearance is one of the major challenges
In vivo intracerebral microdialysis may be useful for in drug discovery because clearance must be suitably low
indications where the rate of CNS distribution is critical, in order to achieve an appropriate half-life and bioavail-
and for simultaneous studies of biomarkers/effect and ability, and therefore dose, in man. However, human clear-
CNS compound levels. Because this method enables direct ance cannot be optimized directly because it is never
and repeated measurement of extracellular fluid, it is pos- measured preclinically, which presents two key challenges.
sible to determine both the rate and extent of CNS pene- Firstly, there is the need to identify the key elimination
tration. The utility of the technique is limited by the processes that determine clearance, and secondly, those
physicochemical properties of the compounds studied. processes must be optimized through appropriate use of
Although the concentration in CSF may differ from that human in vitro systems and/or animal data.
in extracellular fluid, sampling of CSF in animals could be Identification of the likely key clearance mechanism(s)
used since it may offer translatability of CNS distribution in man is supported by a detailed understanding of elimi-
in man. In addition, associations between in vitro potency nation kinetics in animals and demonstration of the
and unbound plasma EC50s of robust PD endpoints such ability of human in vitro tools to predict hepatic clearance
as receptor occupancy should be made, to confirm CNS in man. This puts emphasis on animal kinetics in all
distribution. phases of drug discovery to select the appropriate tools
and models for both compound optimization and human
Cautions PK prediction.
• It is important to note that a relevant brain to plasma There are advantages in having elimination via multiple
ratio will only be generated at steady state like clearance mechanisms, notably for mitigation of DDI
conditions, so the Cu,br/Cu,pl may be very different in risks. This extends to metabolism by multiple enzymes,
steady state and non-steady state conditions. For CNS which helps mitigate against significant metabolism
acting compounds, it is therefore critical not only to (>70%) by highly variable enzyme activities in a popula-
appreciate any acute/steady state differences in CNS tion (i.e. CYP3A4 – see Rawden et al., 2005) and against
distribution, but also to consider the steady state polymorphic enzyme metabolism (e.g. CYP2D6 – see
CNS distribution kinetics at pharmacodynamic Zhou, 2009a, 2009b, and CYP2C19 – see Damle et al.,
endpoints using human dose estimates. For 2009). The severity of variability of human PK depends on
peripheral targets where CNS restriction is important, therapeutic area and on margins to side effects.
high Cu,br/Cu,pl ratios may be unacceptable. Tactics for optimizing clearance are highly dependent
• There are indications that levels in CSF overestimate on the nature of the clearance mechanism, as outlined
free levels in brain, especially for efflux substrates below in sections dealing with metabolic, renal and biliary
(Fridén et al., 2009b), although the latter has been clearance, respectively.
challenged (Doran et al., 2005). The translatability
of CSF concentrations from animal to human is also Optimization of metabolic clearance
somewhat limited because CSF samples in animals
are usually collected from the cisterna magna Introduction
(proximal part of CSF) while those in man are Optimizing metabolic clearance is one of the most
collected from the more distal lumbar region. Very common and challenging activities in drug discovery

145
Section | 2 | Drug Discovery

projects, because high metabolic clearance can be associ- and CYPs. Such knowledge should, therefore, be used to
ated with various metabolic pathways. These include CYP assist in the rational design of novel compounds.
mediated (NADPH dependent) oxidation or reduction in For more lipophilic compound series, CYP3A4 is usually
the liver and, to a lesser extent, in other organs such as the the CYP responsible for most of the metabolism. The rate
small intestine (Galetin et al., 2008). In addition, FMOs of such metabolism is sometimes difficult to reduce with
(Cashman, 2008) and MAO can be involved in oxidative specific modifications of soft spots. Commonly, the overall
metabolism (Bortolato et al., 2008). High metabolic lipophilicity of the molecules needs to be modified. Once
clearance can also be associated with direct or phase 2 a project has established a means to predict lipophilicity
conjugation via UGTs, sulphotransferases, or glutathione- and pKa, correlations with clearance measurements in
S-transferases (Jana and Mandlekar, 2009). Although less microsomes or hepatocytes can be established, and in
common, whole blood and tissue amidases, esterases, turn, used in early screening phases to improve metabolic
various amine oxidases (diamine oxidase, semi-carbazine stability and to influence chemical design.
sensitive amine oxidase), adenosine deaminase, and Screening of compounds in rat or human microsomes
alcohol and aldehyde dehydrogenase may come into play, or hepatocytes is an effective way to determine improve-
depending on the chemotype in question (Cossum, 1988). ments in metabolic clearance. Once alternative clearance
During project work, metabolic clearance can be efficiently mechanisms have been ruled out, total body clearance
screened with available front-line metabolic tools (i.e. greater than hepatic blood flow may indicate an extrahep­
microsomes and hepatocytes). Differential results between atic metabolic clearance mechanism. While not always
microsomes and hepatocytes can highlight the involve- warranted, this can be investigated with a simple plasma
ment of phase 2 processes or the involvement of uptake or blood stability assay if the offending chemotype con-
transporters (Soars et al., 2009). Whole blood and/or tains esters, amide linkages, aldehydes or ketones.
plasma stability can easily be screened where appropriate Softwares that predict sites of metabolism can be used
when the chemotype dictates. to guide experiments to identify potential labile sites on
High metabolic clearance is known to be associated molecules. In addition, structure-based design has taken a
with physicochemical properties of molecules such as a major step forward in recent years with the crystallization
high LogD7.4 (i.e. LogD7.4 values >2.5), where lipophilicity and subsequent elucidation of major CYP isoform struc-
can correlate with metabolic liability (Van de Waterbeemd tures. In silico methods based on molecular docking can
et al., 2001). facilitate an understanding of metabolic interactions with
Failure to clarify and control metabolic clearance is a CYP’s 3A4, 2D6 and 2C9. Although most of the tools,
major issue in discovery because it can: software and modelling expertise to perform structure-
• result in very high total body clearance, low oral based design reside within computational chemistry, there
bioavailability or short half-life is an increasing role for DMPK personnel in this process.
• result in an inability to deliver an oral therapeutic Sun and Scott (2010) provide a review of the ‘state of the
drug level art’ in structure-based design to modify clearance.
• lead to excessively long discovery timelines
• ultimately lead to the demise of a project that Cautions
cannot find an alternative chemotype.
In theory, scaling to in vivo from metabolic stability in
vitro, should in most cases underpredict clearance since
Tactics almost always other mechanisms, more or less, contribute
In order to optimize metabolic clearance it is essential to to overall clearance in vivo. Where common in vitro
understand the enzymatic source of the instability and to models (e.g. hepatocytes) significantly underestimate in
exclude other clearance mechanisms. It is also necessary vivo hepatic clearance and where the chemotype is either a
to have an in vivo assessment of clearance in selected carboxylic acid or a low LogD7.4 base, poor scaling may be
compounds to establish that the in vitro system is predic- due to the compound being a substrate for hepatic uptake
tive of the in vivo metabolic clearance. transporters (e.g. OATPs) which can significantly influence
Addressing high rates of metabolism requires an under- clearance. In such instances, a hepatic uptake experiment
standing of metabolic soft spots and their impact on may prove useful to predict in vivo clearance (see
clearance. For instance, reductions of electron density or Soars et al., 2007 for methodologies). In other instances,
removal of metabolically labile functional groups are suc- carbonyl-containing compounds metabolized by carbonyl
cessful strategies to reduce CYP-related clearance. Knowing reductases in microsomes have been shown to significantly
where in the structure catalysis takes place also gives infor- underestimate clearance when in vitro reactions are driven
mation about how the compound is oriented in the active by direct addition of NADPH, versus the addition of an
site of a CYP. As a result, the site of metabolism could be NADPH regenerating system (Mazur et al., 2009).
blocked for catalysis, or other parts of the molecule could Reductions in in vivo clearance may not be due to
be modified to reduce the affinity between the substrate increased metabolic stability, but driven by increased

146
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

plasma protein binding. Either may be used to reduce For acids and zwitterions, early determination of renal
plasma clearance, but the implications of both need to be clearance in rat and dog is recommended as part of estab-
appreciated. lishing the primary clearance mechanism of a compound.
Once these data are available two possible scenarios exist:
Optimization of renal clearance firstly, a simple allometric relationship can be used to
predict human renal clearance or secondly, if such a rela-
Introduction tionship is not apparent, this prediction should be based
Renal clearance is most commonly a feature of polar, low on fu and renal blood flow corrected dog data. On balance,
LogD7.4 (<~1) compounds due to their low plasma protein there are enough examples to suggest that for acids and
binding, lack of passive permeability (and inability to zwitterions, human renal clearance is best predicted from
facilitate re-absorption in the distal tubule), and the high the dog (McGinnity et al., 2007).
expression in the kidney of a number of transporter
proteins (e.g. OATs and OCTs). Passively cleared com- Cautions
pounds are generally well predicted from animal data Accurately predicting renal clearance of acids and zwitte-
because CLr = GFR × fu (where CLr is renal clearance, GFR rions (i.e. within three-fold) is important, as volumes of
glomerular filtration rate, fu the fraction unbound). Hence, distribution tend to be low and acceptable half-lives are
simple, related, allometric relationships can be readily at risk.
established. The potential for renal DDI of renal transporter sub-
By introducing the potential for large interspecies differ- strates should be considered.
ences, transporters significantly complicate the issue,
making prediction of renal clearance in man more diffi- Optimization of biliary clearance
cult. While there are in vitro assays for animal and human Introduction
transporters (e.g. OATs) that can generate useful QSAR
information; their utility for human PK prediction is yet The biliary clearance of compounds involves active secre-
to be fully established. tion from hepatocytes by ATP-dependent transporter pro-
teins (primarily P-gp, MRP2 and BCRP) into the biliary
Tactics canaliculus, and drainage into the small intestine. It is
commonly a feature of acidic and zwitterionic compounds
Reducing renal clearance in order to reduce total body
and can be very efficient, resulting in hepatic blood flow
clearance is generally achieved by increasing lipophilicity
limited clearances. This efficiency is partly a consequence
through modulation of LogP or pKa. All data are amenable
of many biliary transporter substrates also being substrates
to QSAR analysis, and particularly those from cell lines
of sinusoidal uptake transporters (e.g. OATPs).
expressing individual transporter proteins.
Biliary clearance is a major issue in discovery because it:
Based on a few descriptors, a relatively simple diagram
(Figure 10.5) can be used to classify drugs as having high • can result in very high drug clearance and low
(>1 mL/min/kg), medium (>0.1 to <1 mL/min/kg) or low bioavailability
(≤0.1 mL/min/kg) renal clearance (Paine et al., 2010). • can result in low systemic exposure and small safety
margins
• is difficult to optimize chemically and predict across
species
Acid • is often the synergistic product of two transporter
and Medium systems in series (uptake and efflux).
Compound class

neutrals
Tactics
High
Regardless of the compound, no specific investigation of
Bases biliary clearance is recommended if total clearance is well
and Low
predicted from in vitro metabolic systems, although bile
zwitterions
collected for other reasons can be used to confirm that
≤ –0.38 > –0.38 biliary clearance is low.
In an extensive evaluation of biliary clearance, Yang
Log D et al. (2009) showed molecular weight to be a good pre-
6.5 dictor of biliary clearance in anionic (but not cationic or
neutral) compounds, with a threshold of 400 Da in rat
Fig. 10.5  Simple diagram to classify compounds for and 475 Da in man.
risk of renal clearance. Low renal clearance is defined as If total clearance is not well predicted from in vitro
≤ 0.1 mL/min/kg, moderate as >0.1 to <1 mL/min/kg, and metabolic systems, a biliary elimination study in the rat
high as 1> mL/min/kg should be considered. High biliary clearance in the rat

147
Section | 2 | Drug Discovery

presents a significant hurdle to a discovery project, as being their major effect on the action of many statins
much of time and resource can be spent establishing SARs (Garcia et al., 2003). Knowledge about the SAR for
in the rat, and in considering likely extrapolations to the potency should be kept in mind during drug optimization
dog and man. High biliary clearance is considered a major when evaluating pathways of biotransformation. Active
defect in high clearance chemotypes at candidate drug metabolites with equal or better potency, more favourable
nomination. Tactics for resolving biliary clearance include distribution and/or longer half-lives than the parent can
the following: have profound effects on PKPD relationships. Time–
• Consider molecular weight and ionic charge as response studies will indicate the presence of such metab-
determinants of biliary CL olites provided an appropriate PD model is available.
• Track progress in rat and dog using intravenous PK Failure to discover the influence of such metabolites may
studies without bile collection lead to erroneous dose predictions. Assuming similar
• Make use of in vitro hepatic uptake or uptake potency, metabolites with shorter half-lives than the
transporter models parent will have more limited effects on dose predictions
• If rat and dog provide ambiguous data, an and are difficult to reveal by PKPD modelling. The
intravenous PK study in another non-rodent species presence of circulating active metabolites of candidate
(possibly a non-human primate) may help establish drugs can be beneficial for the properties of a candidate,
a higher species trend but also adds significant risk, complexity and cost in
• Consider assessing efflux in sandwiched cultured rat development.
and human hepatocytes. An emerging body of Prodrugs are special cases where an inactive compound
literature supports this approach, although is designed in such a way as to give rise to pharmacologi-
quantification is challenging (Li et al., 2009). cally active metabolite(s) with suitable properties (Smith,
2007). The rationale for attempting to design oral prod-
Cautions rugs is usually that the active molecule is not sufficiently
Vesicle-based efflux transporter systems are available, but soluble and/or permeable. Remarkably successful prod-
they are not usually of use in detecting/monitoring sub- rugs include omeprazole, which requires only a chemical
strates. Vesicle-based transporter assays have greater utility conversion to activate the molecule (Lindberg et al.,
in screening of inhibition potential (Yamazaki et al., 1996). 1986), and clopidogrel (Testa, 2009) which is activated
Biliary secretion does not necessarily result in the elimi- through CYP3A4. Both omeprazole and clopidogrel
nation of compounds from the body, as there is the poten- became the world’s highest selling drugs.
tial for re-absorption from the GI tract, which can lead to
overestimation of Vss. Tactics
In vitro metabolite identification studies, combined with
Role of metabolite identification the SAR, may suggest the presence of an active metabolite.
studies in optimization Studies of plasma metabolite profiles from PKPD animals
may support their hypothesized presence and relevance.
Introduction Disconnect between the parent compound’s plasma
Knowledge of the metabolites formed from a compound/ profile and its PD in PKPD studies may be a trigger for in
compound series is often highly beneficial or even essential vitro metabolite identification studies.
during drug discovery, because knowing the sites of metab- If active metabolites cannot be avoided during com-
olism facilitates rational optimization of clearance and pound optimization, or are considered beneficial for effi-
aids in understanding DDI, particularly time-dependent cacy, their ADME properties should be determined.
DDI. As these issues are considered elsewhere (sections on Advancing the metabolite rather than the parent com-
clearance optimization and avoidance of PK-based drug– pound is an option to be considered.
drug interactions), they are not discussed further here. The decision tree in Figure 10.6 highlights ways to
However, two other major areas of drug discovery which become alerted to the potential presence of active metabo-
are dependent on metabolite identification and on under- lites, and suitable actions to take.
standing metabolite properties are considered below:
• Assessment of pharmacologically active metabolites Cautions
• Avoiding/assessing risk from reactive metabolites. Using preclinical data to predict exposure to an active
metabolite in man is usually accompanied by considera-
Active metabolites ble uncertainty (Anderson et al., 2009). Plasma and tissue
concentrations of metabolites in man are dependent on a
Introduction number of factors. Species differences in formation, distri-
Active metabolites can influence PD and influence PKPD bution and elimination need to be included in any
relationships (Gabrielsson and Green, 2009), one example assessment.

148
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

In vitro or in vivo metabolism combined with SAR and/or

PKPD indicates/shows hysterisis or unexpected potency
suggests active metabolite likely to circulate

No Yes Yes No

No further Yes Plasma profiles in PD species indicate No No further

action active circulating metabolite(s) action

Yes
Other candidates available? Progress other candidates Seek other reasons for observed PKPD
(e.g. hit-and-run mechanisms)
No Consider radiolabel for adequate quantification
Fully characterize parent
No and metabolite for
Make and characterize active metabolite
next milestone
Can metabolite progress instead of parent?
Get buy-in for increased
cost and risk
Yes

Switch to metabolite as candidate

Fig. 10.6  Active metabolite decision tree. PKPD, pharmacokinetics-pharmacodynamics; SAR, structure–activity relationship.

Minimizing risk for reactive metabolites to avoid/design away from reactive metabolite
during drug discovery formation.
• Trapping studies in human liver microsomes to
Introduction enable the detection of reactive metabolites. Agents
Many xenobiotics are converted by drug metabolizing used to trap unstable reactive intermediates are
enzymes to chemically reactive metabolites which may glutathione (for soft electrophiles) and radiolabelled
react with cellular macromolecules. The interactions may potassium cyanide (K14CN) (for iminium ions) as
be non-covalent (e.g. redox processes) or covalent, and can first-line assays. Structural information can be
result in organ toxicity (liver being the most common obtained from further analysis of GSH and CN
organ affected), various immune mediated hypersensitiv- adduct mass spectrometry data. For very reactive
ity reactions, mutagenesis and tumour formation. Muta- aldehydes, i.e. α,β-unsaturated aldehyde,
genesis and carcinogenesis arise as a consequence of DNA methoxylamine can be used as trapping reagent.
damage, while other adverse events are linked to chemical • Metabolite identification in human liver
modification of proteins and in some instances lipids. microsomes or hepatocytes. Interpretation of the
Avoiding, as far as possible, chemistry giving rise to such metabolite patterns may give information about
reactive metabolites is, therefore, a key part of optimiza- existence of putative short-lived intermediates
tion in drug discovery. preceding the observed stable metabolites (e.g.
dihydrodiols are likely to be result of hydration of
Tactics epoxides).
Minimizing the reactive metabolite liability in drug dis- • Time-dependent inhibition (TDI) studies in human
covery is based on integrating DMPK and safety screening liver microsomes to flag the likelihood of a
into the design-make-test-analyse process. A number of mechanism based inhibitor (mainly inhibition of
tools are available to assist projects in this work (Thomp- CYPs).
son, et al., 2011 submitted to Chem-Biol Interact, and ref- • Formation and degradation of acyl glucuronides
erences therein): from carboxylic acids in activated human liver
• Search tools/databases: this includes in silico microsomes. This gives an overall estimate of the
tools to (1) identify substructures associated with acylating capability of acyl glucuronides.
potential reactive metabolite formation, (2) identify If it proves difficult or not possible to optimize away
potential bacterial mutagenicity, and (3) learn how from reactive metabolite signals using the screen assays

149
Section | 2 | Drug Discovery

listed above, yet compounds in the series for other reasons Despite these and other fundamental shortcomings in
are regarded as sufficiently promising, then a reactive the reactive metabolite science, most pharmaceutical com-
metabolite risk assessment will have to be undertaken. panies invest significant resources into screening away
Such assessment includes covalent binding to human from chemistry giving rise to such molecular species.
hepatocytes in vitro and/or to rat in vivo. Predicted dose Taking a compound with such liabilities into man will
to man and fraction of the dose being predicted to be require complex, and probably even more costly and time-
metabolized over the reactive metabolite pathway will consuming risk assessment efforts.
have to be taken into account. Other experimental systems
which might be used for assaying metabolite-mediated
cytotoxicity are cell systems devoid of, or overexpressing, HUMAN PK AND DOSE PREDICTION
various CYPs. Details of such an assessment are beyond
the scope of this review but are discussed by Thompson
The overriding goal of in silico, in vitro, and in vivo DMPK
et al., 2011 (Chem-Biol Interact, submitted).
methods conducted at all stages of drug discovery is to
help design and select compounds with acceptable human
Caution pharmacokinetics. Hence the prediction of human PK is
Although reactive metabolites, beyond reasonable doubt, an important component of modern drug discovery and
constitute a risk worthwhile to screen away from, the is employed throughout the drug discovery process. In
underlying science of potential toxicity is complex; e.g. the early stages, it is used to estimate how far project
overall covalent binding to proteins is a crude measure chemistry is from its target profile and to identify the most
indeed. It is likely covalent binding to some proteins critical parameters for optimization. At candidate selec-
might give rise to antigenic conjugates, while binding to tion, accurate PK and dose prediction is required to deter-
other proteins will be quite harmless. Formation of glu- mine not only whether the drug candidate meets the target
tathione adducts as such is not alarming, particularly profile but also to estimate safety margins, compound
when catalysed by glutathione transferases. The relevance requirements for early development phases and potential
of trapping with an unphysiological agent like cyanide DDI risk.
could be even more challenged. Simply quantifying cova- The prediction of human PK involves two distinct
lent binding in vivo or in vitro may be misleading, but stages – firstly, the estimation of individual kinetic para­
since the science of estimating risks associated with differ- meters, and secondly, the integration of these processes
ent patterns of binding is still in its infancy, there is little into a model to simulate a concentration-time profile (see
choice. Figure 10.7).

Oral absorption Tissue distribution Metabolism/elimination

1. In silico 1. In silico 1. In silico

2. In vitro (e.g. MDCK) 2. In vitro (PPB, Kp) 2. In vitro (heps, mics)
3. Animal data (F, ka) 3. Animal data (Vss,u) 3. Animal data (CL)

Human scaling

Human F, Ka Human Vss Human CL

1 compartment model or PBPK

Stimulated human concentration-time profile

Fig. 10.7  Process for predicting human pharmacokinetics.

150
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

To predict the human PK profile following oral admin- extraction is a function of hepatic clearance and hepatic
istration, the following fundamental PK parameters must blood flow and can be estimated using the following
be determined: (a) absorption rate (ka) and bioavailability equation:
(F), (b) volume of distribution (Vss), and (c) clearance f h = 1 − CL h /Q
(CL, total, hepatic, renal and other routes). Methods
where CLh is the hepatic clearance and Q is the hepatic
used to predict individual PK parameters can be classified
blood flow.
into two approaches: empirical or mechanistic. Mechanis-
tic methods are based on knowledge of the underlying
mechanisms or processes that define the PK parameters
while empirical methods rely on little or no a priori Prediction of clearance
knowledge. Data required for these methods can be in
Clearance is a primary pharmacokinetic parameter that is
silico, in vitro, or in vivo data and in general, methods
used to characterize drug disposition. Total or systemic
that integrate and use both in vitro and in vivo data.
clearance, the sum of all individual organ clearance
PBPK (physiologically based pharmacokinetic) models
responsible for overall elimination of a drug, can be
utilising measured parameters (in vivo or in vitro) tend to
predicted empirically using allometry. Allometric scaling
be more accurate than methods that rely on in silico
is based on the empirical observations that organ sizes
predictions.
and physiological processes in the mammalian body
are proportional to body weight by the following
equation:
Prediction of absorption rate and
Y = aW b
oral bioavailability
where Y is the parameter of interest, W the body
Although frequently ignored in scaling, accurate predic- weight, and a and b are the coefficient and exponent
tion of oral absorption rates (ka) is required to predict of the allometric equation, respectively. For rates, flows
Cmax and to project the potential for drug–drug inter­ and clearance processes, the exponent should be 0.75. If
actions. Oral absorption rate is heavily dependent on the calculated exponent deviates significantly from the
the final physical form of the compound and the pre­ above, the prediction may not be accurate (Mahmood,
diction of this parameter in the early phase of discovery 2007).
may be of limited value. ka can be scaled empirically using For individual organ clearance, allometry is useful for
ka values determined from preclinical species. ka values flow-dependent clearance processes, e.g. renal and high
are usually obtained from rat or dog, i.v. and p.o. PK hepatic CL. However, for low CL compounds that exhibit
studies via de-convolution analysis. Oral absorption deter- large species differences in metabolism, hepatic clearance
mined in the rat was usually found to be in good agree- and hence total clearance cannot be reliably predicted
ment to that in human. However, in the dog, oral by allometry. Hepatic clearance can be mechanistically
absorption rates for hydrophilic compounds were usually scaled using an in vitro and in vivo correlation (IVIVC)
much faster than in human (Chiou et al., 2000a). Hence, approach. With this approach, intrinsic in vitro clearance
if there is discrepancy in the absorption rate between the is measured using in vitro metabolic stability assays
two species, the value from the rat should be used for the (microsomes or hepatocytes). The measured in vitro clear-
prediction. ance values are scaled to clearance by the whole liver
ka can also be predicted mechanistically using various and then converted to an hepatic clearance using a liver
variations of a physiologically based transit model origi- model that incorporates blood flow (a well-stirred model
nally developed by Amidon et al. (Yu and Amidon, most often used). If IVIVC predicts clearance in animal
1999). The model is termed compartment absorption species, it is likely that the method is applicable for
and transit (CAT) model and characterizes the fraction of human predictions. Historically, the method for pre­
compound absorbed per unit time (based on its dissolu- dicting intrinsic hepatic metabolic clearance does so to
tion rate, pH solubility profile, effective permeability, within a factor of 3, unless confounded by transporter
intestinal metabolism and efflux, etc.) as the compound or other mechanistic effects that are not incorporated
moves through the different compartments of the GI tract. in the modelling. For any chemical series, a key compo-
This approach is used in a number of commercially avail- nent of gaining confidence in the use of IVIVC to predict
able tools such as GastroPlus and SIMCYP and can be human clearance comes from establishing the ability of
used to estimate fraction of the dose absorbed and (fa), analogous tools to predict metabolic clearance in the
and the fraction escaping intestinal metabolism (fg), in rat and dog. It is important to establish such relationships
addition to ka. and this requires an accurate estimation of the metabolic,
As mentioned earlier in the optimizing oral bioavaila- and other clearance mechanisms in each species, as well
bility section, oral bioavailability is a composite product as prediction of all significant clearance mechanisms
of (fa), (fg) and hepatic ‘first-pass’ extraction (fh). Hepatic in man.

151
Section | 2 | Drug Discovery

Prediction of volume of   are estimated by logD and pKa of the compound (Poulin
and Theil, 2000; Rodgers et al., 2005; Rodgers and
distribution
Rowland, 2006). A key assumption is that unbound tissue
Volume of distribution (V) is a primary PK parameter Kps are constant across species. This underpins the
that relates drug concentration measured in plasma or assumption described above, that the unbound volume of
blood to the amount of drug in the body and is used to distribution is constant across species.
characterize drug distribution. It is a key parameter as it is
a primary determinant (together with clearance) of drug Prediction of plasma concentration
half-life. Higher volume compounds will have longer half-
lives. In a simple system: time profile
T1/2 = ln(2) ∗ V/CL To accurately estimate Cmax and Cmin, which are important
Volume of distribution is a measure of the relative affin- exposure parameters for assessing safety and efficacy, the
ity of the compound for plasma/blood constituents and plasma concentration-time profile has to be accurately
tissue constituents. In general, for moderately lipophilic predicted.
compounds, acids have a high affinity for albumin and, For compounds displaying mono-exponential decreases
therefore, have a low volume of distribution (0.1–0.5 L/ in plasma concentrations over time, the concentration-
kg), based have a high affinity for tissues and, therefore, time profile in man can be predicted using the following
have high volumes (>3 L/kg), while neutral compounds equation:
have volumes around 1 L/kg. Correcting V for plasma FDka
C(t ) =
protein binding yields a very useful parameter ‘unbound CL   − V ∗ t
CL

V  ka − e − e − ka ∗ t 
volume’:  V   
Vu = V/fu where C is the concentration at time t, F is bioavailability,
Unbound volume is a measure of the average tissue D is the dose, ka is the absorption rate, CL is the clearance,
affinity of a compound, but more importantly should be and V is the volume of distribution.
relatively constant across species. If this consistency is For compound showing biphasic or multi-exponential
observed in preclinical species, human predictions are concentration time profile in animals, PBPK modelling
relatively straightforward, and a sound basis for the use of is the recommended approach to simulate the profile
more sophisticated tools (e.g. physiologically based in man.
(PBPK) models, see below) to predict complex distribu-
tion profiles is established. Prediction of human  
In PBPK modelling, the body is portrayed as a series of
compartments that represent tissues and organs. The com- efficacious dose
partments are arranged to reflect anatomical layout and Estimation of the likely therapeutic dose and dosing fre-
are connected by arterial and venous pathways. Each tissue quency in patients, requires not only the prediction of the
(i) has an associated blood flow rate (Q), volume (Vt) and human pharmacokinetics, but also a robust understand-
a tissue partition coefficient (Kp) and the rate of change ing of the concentration-time-response relationship. The
of the drug in each tissue is mathematically described by quality of the prediction depends on how well the PD
a differential equation. effect scales from animals to man and the linkage between
The total volume of distribution is the sum of the the effect and the clinical outcome. The linkage between
volume of distribution of all the organs. The volume of the effect and the clinical outcome is based on a series of
distribution in each organ is simply the actual volume of translational biomarkers. The different types of biomark-
the organ multiplied by its corresponding tissue distribu- ers are classified (Figure 10.8, adapted from Danhof et al.,
tion coefficient (Kp). Kp can be determined experimen- 2005). The classification is based on the mechanism of
tally in animals for each organ; however, this is very drug action and the relationship to the disease process. In
resource intensive and impractical. There are two general, the closer the relationship of the biomarker is to
approaches for estimating Kp in various organs. In silico the disease process, the more relevant and predictive is the
estimation of Kp can be done using various tissue compo- biomarker.
sition methods. This is the approach used in commercial
software packages such as GastroPlus and SIMCYP. The
other approach was developed by Arundel (1997) using
rat Vss data and is based on the observation that Kps are SUMMARY
constant for a given Vss and that they can be predicted from
Vss (Arundel, 1997). With the tissue composition methods, To be a successful drug candidate, a compound, in addi-
partitions to various components in the tissue (neutral tion to having good efficacy and safety profile, has to have
lipid, acidic phospholipids, lipoproteins, tissue albumin) acceptable DMPK properties. This chapter highlights the

152
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

Type 4A
Physiological
response

Type 0 Type 1 Type 2 Type 3 Type 4B Type 5 Type 6

Genotype/ Drug Target Target Physiological Patho- Clinical
phenotype concentration occupancy function response physiological scale
process

Fig. 10.8  Classification of biomarkers useful to quantify a project’s therapeutic model.

roles of DMPK in drug discovery and provides strategies

to resolve key DMPK challenges such as improving oral ACKNOWLEDGMENTS
bioavailability, optimizing clearance, avoiding drug–drug
interaction, achieving or avoiding CNS exposure, and The authors wish to thank Lovisa Afzelius, Tommy B.
avoiding risks from metabolites. In addition, approaches Andersson, Madeleine Antonsson, Ulf Bredberg, Anne
used to scale individual PK parameters and strategy to Cooper, Doug Ferguson, C. Edwin Garner, Ken Grime,
integrate these parameters to predict human pharmacoki- Anshul Gupta, Lena Gustavsson, Ramon Hendrickx,
netics and dose are discussed in the chapter. The proposed Suzanne Iversson, Owen Jones, Etienne Lessard, Stefan
strategy is based on best practices within AstraZeneca as Lundquist, Yan Li, Nektaria Markoglou, Jan Neelisen, Ken
well as current science, technology and understanding of Page, Stuart Paine, Jan Paulson, Denis Projean, Maria Rib-
DMPK. It is our hope that the proposed integrated strategy, adeneira, Caroline Rivard, Ralf Schmidt, Patricia Schroeder,
which focus DMPK efforts toward the optimization and Anna-Karin Sternbeck, Per Strandberg, Richard Thomp-
prediction of DMPK properties in human will provide a son, Anna-Lena Ungell, and Lucas Utley for their inputs
sound basis to efficiently progress drug discovery projects. to the strategy.

REFERENCES

Arundel P. A multi-compartment model spectrometry-based CYP inhibition Chiou WL, Ma C, Chung SM, et al.
generally applicable to assays for use in drug discovery. Similarity in the linear and
physiologically based Journal of Biomolecular Screening non-linear oral absorption of drugs
pharmacokinetics. 3rd IFAC 2008;13(5):343–53. between human and rat.
Symposium: modelling and control Bjornsson T, Callaghan J, Einolf JH, International Journal of Clinical
in biological systems, Warwick, UK, et al. The conduct of in vitro and in Pharmacolology Therapy
March 23rd–26th, 1997. vivo drug-drug interaction studies: a 2000a;38:532–9.
Amidon GL, Lennernas H, Shah VP, PhRMA perspective. Journal of Chiou W, Jeong Y, Chung S, et al.
et al. A theoretical basis for a Clinical Pharmacology Evaluation of dog as an animal
biopharmaceutic drug classification: 2003;43:443–69. model to study the fraction of oral
the correlation of in vitro drug Bortolato M, Chen K, Shih JC. dose absorbed for 43 drugs in
product dissolution and in vivo Monoamine oxidase inactivation: humans. Pharmacology Research
bioavailability. Pharmacology from pathophysiology to 2000b;17:135–40.
Research 1995;12:413–20. therapeutics. Adverse Drug Cossum PA. Role of the red blood cell
Anderson S, Luffer-Atlas D, Knadle MP. Delivery Review 2008;60: in drug metabolism.
Predicting circulating human 1527–33. Biopharmaceutics and Drug
metabolites: how good are we? Cashman JR. Role of flavin-containing Disposition 1988;9:321–36.
Chemical Research Toxicology monooxygenase in drug Damle BD, Uderman H, Biswas P, et al.
2009;22:243–56. development. Expert Opinion on Influence of CYP2C19
Bell L, Bickford S, Nguyen PH, et al. Drug Metabolism and Toxicology polymorphism on the
Evaluation of fluorescence- and mass 2008;4:1507–21. pharmacokinetics of nelfinavir and

153
Section | 2 | Drug Discovery

its active metabolite. British Journal Galetin A, Gertz M, Houston JB. Lindberg P, Nordberg P, Alminger T,
of Clinical Pharmacology Potential role of intestinal first-pass et al. The mechanism of action of
2009;68:682–9. metabolism in the prediction of the gastric acid secretion inhibitor
Danhof M, Alvan G, Dahl SG, et al. drug-drug interactions. Expert omeprazole. Journal of Medicinal
Mechanism-based pharmacokinetic- Opinion on Drug Metabolism and Chemistry 1986;29:1327–9.
pharmacodynamic modeling – a Toxicology 2008;4:909–22. Lipinski CA, Lombardo F, Dominy BW,
new classification of biomarkers. Garcia MJ, Reinoso RF, Sanchez Navarro et al. Experimental and
Pharmaceutical Research A, et al. Clinical pharmacokinetics of computational approaches to
2005;22:1432–7. statins. Methods and Findings in estimate solubility and permeability
Di L, Kerns H, Carte G. Strategies to Experimental and Clinical in drug discovery and development
assess blood–brain barrier Pharmacology 2003;25:457–81. settings. Advanced Drug Delivery
penetration. Expert Opinion on Gertz M, Harrison A, Houston JB, et al. Reviews 1997;23:3–25.
Drug Discovery 2008;3:677–87. Prediction of human intestinal Mahmood I. Application of allometric
Didziapetris R, Japertas P, Avdeef A, first-pass metabolism of 25 CYP3A principles for the prediction of
et al. Classification analysis of substrates from in vitro clearance pharmacokinetics in human and
P-glycoprotein substrate specificity. and permeability data. Drug veterinary drug development.
Journal of Drug Targeting Metabolism and Disposition Advanced Drug Delivery Reviews
2003;11:391–406. 2010;38:1147–58. 2007;59:1177–92.
Doran A, Obach RS, Smith BJ, et al. Gleeson P, Davis A, Chohan K, et al. Mazur CS, Kenneke JF, Goldsmith MR,
The impact of P-glycoprotein on Generation of in silico cytochrome et al. Contrasting influence of
the disposition of drugs targeted for P450 1A2, 2C9, 2C19, 2D6, 3A4 NADPH and a NADPH-regenerating
indications of the central nervous inhibition QSAR models. Journal of system on the metabolism of
system: evaluation using the Computer Aided Molecular Design carbonyl-containing compounds in
MDR1A/1B knockout mouse model. 2007;21:559–73. hepatic microsomes. Drug
Drug Metabolism and Disposition Hitchcock S, Pennington L. Structure- Metabolism and Disposition
2005;33:165–74. brain exposure relationships. Journal 2009;37:1801–5.
Fowler S, Zhang H. In vitro evaluation of Medicinal Chemistry McGinnity DF, Collington J, Austin RP,
of reversible and irreversible 2006;49:7559–83. et al. Evaluation of human
cytochrome P450 inhibition: current Huang S, Strong J, Zhang L, et al. New pharmacokinetics, therapeutic dose
status on methodologies and their era in drug interaction evaluation: and exposure predictions using
utility for predicting drug-drug US Food and Drug Administration marketed oral drugs. Current Drug
interactions. AAPS Journal update on CYP enzymes, Metabolism 2007;8:463–79.
2008;10:410–24. transporters, and the guidance Paine SW, Barton P, Bird J, et al. A rapid
Fridén M, Ducrozet F, Middleton B, et al. process. Journal of Clinical computational filter for predicting
Development of a high-throughput Pharmacology 2008;48:662–70. the rate of human renal clearance.
brain slice method for studying drug International Transporter Consortium. Journal of Molecular Graphics and
distribution in the central nervous Membrane transporters in drug Modelling 2010;29:529–37. Epub
system. Drug Metabolism and development. Nature Reviews. Drug 2010 Oct 20.
Disposition 2009a;37:1226–33. Discovery 2010;9:215–36. Poulin P, Theil FP. A priori prediction of
Fridén M, Ljungqvist H, Middleton B, Jana S, Mandlekar S. Role of phase II tissue:plasma partition coefficients of
et al. Improved measurement of drug drug metabolizing enzymes in drugs to facilitate the use of
exposure in brain using drug-specific cancer chemoprevention. Current physiologically-based
correction for residual blood. Journal Drug Metabolism 2009;10: pharmacokinetic models in drug
of Cerebral Blood Flow and 595–616. discovery. Journal of Pharmaceutical
Metabolism 2010;30:150–61. Johnson TW, Dress KR, Edwards M. Sciences 2000;89:16–35.
Fridén M, Winiwarter S, Jerndal G, et al. Using the golden triangle to Rawden HC, Carlile DJ, Tindall A, et al.
Structure-brain exposure optimise clearance and oral Microsomal prediction of in vivo
relationships in rat and human using absorption. Bioorganic and clearance and associated
a novel data set of unbound drug Medicinal Chemistry Letters interindividual variability of six
concentrations in brain interstitial 2009;19:5560–1554. benzodiazepines in humans.
and cerebrospinal fluids. Journal of Kola I, Landis J. Can the pharmaceutical Xenobiotica 2005;35:603–25.
Medicinal Chemistry industry reduce attrition rates? Riley R, Grime K, Weaver R. Time-
2009b;52:6233–43. Nature Reviews. Drug Discovery dependent CYP inhibition. Expert
Gabrielsson J, Green AR. Quantitative 2004;3:711–5. Opinion on Drug Metabolism and
pharmacology or pharmacokinetic Li N, Bi YA, Duignan DB, et al. Toxicology 2007;3:51–66.
pharmacodynamic integration Quantitative expression profile of Rodgers T, Leahy D, Rowland M.
should be a vital component in hepatobiliary transporters in Physiologically based
integrative pharmacology. Journal of sandwich cultured rat and human pharmacokinetic modeling 1:
Pharmacology and Experimental hepatocytes. Molecular predicting the tissue distribution of
Therapeutics 2009;331:767–74. Pharmacology 2009;6:1180–9. moderate-to-strong bases. Journal of

154
 Metabolism and pharmacokinetic optimization strategies in drug discovery Chapter | 10 |

Pharmaceutical Sciences Chemico-Biological Interactions Bioorganics and Medicinal

2005;94:1259–76. 2011;192:65–71. Chemistry Letters 2009;19:
Rodgers T, Rowland M. Physiologically Ungell AL, Nylander S, Bergstrand S, 2844–51.
based pharmacokinetic modelling 2: et al. Membrane transport of Wenlock MC, Austin RP, Barton P, et al.
predicting the tissue distribution of drugs in different regions of the A comparison of physiochemical
acids, very weak bases, neutrals and intestinal tract of the rat. Journal property profiles of development
zwitterions. Journal of of Pharmaceutical Sciences and marketed oral drugs. Journal
Pharmaceutical Sciences 1998;87:360–6. of Medicinal Chemistry 2003;46:
2006;95:1238–57. US Food and Drug Administration. 1250–6.
Rowland M, Tozer TN. Clinical Draft Guidance for Industry: Drug Williams J, Hyland R, Jones B, et al.
pharmacokinetics – concepts and Interaction Studies – Study Design, Drug-drug interactions for UDP-
applications. 2nd ed. London: Data Analysis and Implications for glucuronosyltransferase substrates; a
Lea and Febiger; 1989. Dosing and Labeling, 2006. pharmacokinetic explanation for
Smith DA. Do prodrugs deliver? Current Available from: URL: http://www. typically observed low exposure
Opinion on Drug Discovery and fda.gov/downloads/Drugs/Guidance (AUCi/AUC) ratios. Drug
Development 2007;10:550–9. ComplianceRegulatoryInformation/ Metabolism and Disposition
Soars MG, Grime K, Sproston JL, et al. Guidances/ucm072101.pdf. 2004;32:1201–8.
Use of hepatocytes to assess the US Food and Drug Administration. Yamazaki M, Kobayashi K, Sugiyama Y.
contribution of hepatic uptake to Guidance for Industry: Drug Primary active transport of
clearance in vivo. Drug Metabolism Metabolism/Drug Interactions in the pravastatin across the liver
and Disposition 2007;35:859–65. Drug Development Process: Studies canalicular membrane in normal
Soars MG, Webborn PJ, Riley RJ. Impact in vitro, 1997. Available from: URL: and mutant Eisai hyperbilirubinemic
of hepatic uptake transporters on http://www.fda.gov/downloads/ rats. Biopharmaceutics and Drug
pharmacokinetics and drug-drug Drugs/GuidanceCompliance Disposition 1996;17:607–21.
interactions: use of assays and RegulatoryInformation/Guidances/ Yang X, Gandhi YA, Duignan DB, et al.
models for decision making in the ucm072104.pdf. Prediction of biliary excretion in rats
pharmaceutical industry. Molecular Van de Waterbeemd H, Smith DA, Jones and humans using molecular weight
Pharmacology 2009;6:1662–77. BC. Lipophilicity in PK design: and quantitative structure-
Sun H, Scott DO. Structure-based drug methyl, ethyl, futile. Journal of pharmacokinetic relationships. AAPS
metabolism predictions for drug Computer Aided Molecular Design Journal 2009;11:511–25.
design. Chemical Biology and Drug 2001;15(3):273–86. Yu LX, Amidon GL. A compartmental
Design 2010;75:3–17. Wan H, Rehngren M, Giordanetto F, absorption and transit model for
Takahashi M, Washio T, Suzuki N, et al. et al. High-throughput screening of estimating oral absorption.
The species differences of intestinal drug-brain tissue binding and in International Journal of
drug absorption and first-pass silico prediction for assessment of Pharmaceutics 1999;186:
metabolism between cynomolgus central nervous system drug delivery. 119–25.
monkeys and humans. Journal of Journal of Medicinal Chemistry Zhou SF. Polymorphism of human
Pharmaceutical Sciences 2007;50:4606–15. cytochrome P450 2D6 and its
2009;98:4343–53. Ward P. Importance of drug transporters clinical significance: part I. Clinical
Testa B. Drug metabolism for the in pharmacokinetics and drug safety. Pharmacokinetics
perplexed medicinal chemist. Toxicology Mechanisms and 2009a;48:689–723.
Chemistry and Biodiversity Methods 2008;18:1–10. Zhou SF. Polymorphism of human
2009;6:2055–70. Waring MJ. Defining optimum cytochrome P450 2D6 and its
Thompson RA, Isin EM, Yan L, et al. lipophilicity and molecular weight clinical significance: part II. Clinical
Risk assessment and mitigation ranges for drug candidates – Pharmacokinetics
strategies for reactive metabolites in molecular weight dependent lower 2009b;48:761–804.
drug discovery and development. logD limits based on permeability.

155
 Chapter 11 
Pharmacology: its role in drug discovery
H P Rang

related to the target (i.e. side effects). At the

laboratory level the two are often not clearly
INTRODUCTION distinguishable, and the borderline between
secondary pharmacodynamic and safety
Pharmacology as an academic discipline, loosely defined pharmacology studies (see below) is also uncertain.
as the study of the effects of chemical substances on living Nevertheless, for the purposes of formal
systems, is so broad in its sweep that it encompasses all documentation, the distinction may be useful
aspects of drug discovery, ranging from the molecular • Testing in animal models of disease to determine
details of the interaction between the drug molecule and whether the compound is likely to produce
its target to the economic and social consequences of therapeutic benefit
placing a new therapeutic agent on the market. In this • Safety pharmacology, consisting of a series of
chapter we consider the more limited scope of ‘classical’ standardized animal tests aimed at revealing
pharmacology, in relation to drug discovery. Typically, undesirable side effects, which may be unrelated to
when a molecular target has been selected, and lead com- the primary action of the drug. This topic is
pounds have been identified which act on it selectively, discussed in Chapter 15.
and which are judged to have ‘drug-like’ chemical attributes The pharmacological evaluation of lead compounds
(including suitable pharmacokinetic properties), the next does not in general follow a clearly defined path, and often
stage is a detailed pharmacological evaluation. This means it has no clearcut endpoint but will vary greatly in its
investigation of the effects, usually of a small number of extent, depending on the nature of the compound, the
compounds, on a range of test systems, up to and includ- questions that need to be addressed and the inclinations
ing whole animals, to determine which, if any, is the most of the project team. Directing this phase of the drug dis-
suitable for further development (i.e. for nomination as a covery project efficiently, and keeping it focused on the
drug candidate). Pharmacological evaluation typically overall objective of putting a compound into develop-
involves the following: ment, is one of the trickier management tasks. It often
• Selectivity screening, consisting of in vitro tests on happens that unexpected, scientifically interesting data are
a broad range of possible drug targets to determine obtained which beg for further investigation even though
whether the compound is sufficiently selective for they may be peripheral to the main aims of the project.
the chosen target to merit further investigation From the scientists’ perspective, the prospect of opening
• Pharmacological profiling, aimed at evaluating in up a new avenue of research is highly alluring, whether
isolated tissues or normal animals the range of the work contributes directly to the drug discovery aims
effects of the test compound that might be or not. In this context, project managers need to bear in
relevant in the clinical situation. Some authorities mind the question: Who needs the data and why? – a
distinguish between primary pharmacodynamic studies, question which may seem irritatingly silly to a scientist in
concerning effects related to the selected therapeutic academia but totally obvious to the commercial mind. The
target (i.e. therapeutically relevant effects), and same principles apply, of course, to all parts of a drug
secondary pharmacodynamic studies, on effects not discovery and development project, but it tends to be at

© 2012 Elsevier Ltd. 157

Section | 2 | Drug Discovery

Table 11.1  Characteristics of pharmacological test systems

Test Molecular/cellular In vitro Whole animal Whole animal

system assays pharmacology pharmacology disease models
attribute (normal animals)
Throughput High (thousands/day) Moderate (c. 10/day) Low (<10/day) Generally low or
very low, depending
on nature of model
Quantitative Good Good, but may be Relatively poor, due to As for whole animal
precision subject to uncontrolled pharmacology, plus
environmental and pharmacokinetic and added variability of
physiological physiological factors disease model
variation phenotype
Cost Low Fairly low depending High, depending on High, depending on
on number and cost number and cost of number and cost of
of animals needed animals needed animals needed
Flexibility of Generally inflexible. Highly adaptable Adaptable, but
experimental Washout effects, limitations imposed by
design repeat dose effects, pharmacokinetics
etc., difficult to study
Suitability for Unsuitable Unsuitable Depends on model. As for whole
chronic Suitable if repeated animal, provided
experiments non-invasive readouts are disease phenotype
feasible. Possible, but remains stable
expensive for one-off
terminal readouts
Species Often performed on Rarely possible with Animal studies may not Animal studies may
dependence human cell lines or human tissues be applicable to humans not be applicable to
cloned human targets human
Usefulness OK for me-too drugs. OK for me-too As above Variable, depending
for predicting Poor for drugs acting drugs. Poor for on characteristics of
therapeutic through novel drugs acting through model
efficacy mechanisms novel mechanisms
Usefulness Useful if broad Sometimes useful Generally useful as basis Usually not
for predicting selectivity screen is for ‘safety pharmacology’ informative
side effects performed screening

the stage of pharmacological evaluation that conflicts first The strengths and weaknesses of these test systems are
arise between scientific aspiration and commercial need. summarized in Table 11.1.
An important principle in pharmacological evaluation Pharmacological characterization of a candidate com-
is the use of a hierarchy of test methods, covering the range pound often has to take into account active metabolites,
from the most reductionist tests on isolated molecular based on information from drug metabolism and phar-
targets to much more elaborate tests of integrated physi- macokinetics (DMPK) studies (see Chapter 10). If a major
ological function. Establishing and validating such a series active metabolite is identified, it will be necessary to syn-
of tests appropriate to the particular target and indication thesize and test it in the same way as the parent compound
being addressed is one of the most important functions of in order to determine which effects (both wanted and
pharmacologists in the drug discovery team. In general, unwanted) relate to each. Particular problems may arise if
assays become more complicated, slow and expensive, and the metabolic fate of the compound shows marked species
more demanding of specialist skills as one moves up this differences, making it difficult to predict from animal
hierarchy. studies what will happen in humans.

158
 Pharmacology: its role in drug discovery Chapter | 11 |

Although most of the work involved in pharmacological analysing drug binding experiments are available (Keen,
characterization of a candidate drug takes place before 1999; Vogel, 2002). Generally, the aim of the assay is to
clinical studies begin, it does not normally end there. Both determine the dissociation constant, KD, of the test com-
ongoing toxicological studies and early trials in man may pound, as a measure of its affinity for the receptor. In most
reveal unpredicted effects that need to be investigated cases, the assay (often called a displacement assay) measures
pharmacologically, and so the discovery team needs to the ability of the test compound to inhibit the binding of
remain actively involved and be able to perform experi- a high-affinity radioligand which combines selectively
ments well into the phase of clinical development. They with the receptor in question, correction being made for
cannot simply wave the compound goodbye once the dis- ‘non-specific’ binding of the radioligand.
covery phase is completed. In the simplest theoretical case, where the radioligand
and the test compound bind reversibly and competitively
to a homogeneous population of binding sites, the effect
of the test ligand on the amount of the radioligand specifi-
SCREENING FOR SELECTIVITY cally bound is described by the simple mass-action
equation:
The selectivity of a compound for the chosen molecular B/Bmax = ([A]/K A )/([A]/K A + [L]/K L + 1)
(1)
target needs to be assessed at an early stage. Compounds
selected for their potency, for example on a given amine where B = the amount of radioligand bound, after correct-
receptor, protease, kinase, transporter or ion channel, are ing for non-specific binding, Bmax = the maximal amount
very likely to bind also to related – or even unrelated – of radioligand bound, i.e. when sites are saturated, [A] =
molecular targets, and thereby cause unwanted side effects. radioligand concentration, KA = dissociation constant for
Selectivity is, therefore, as important as potency in choos- the radioligand, [L] = test ligand concentration, and KL =
ing potential development candidates, and a ‘selectivity dissociation constant for the test ligand.
screen’ is usually included early in the project. The range By testing several concentrations of L at a single con­
of targets included in such a screen depends very much on centration of A, the concentration, [L]50, needed for 50%
the type of compound and the intended clinical indica- inhibition of binding can be estimated. By rearranging
tion. Ligands for monoamine receptors and transporters equation 1, KL is given by:
form a large and important group of drugs, and several K L = [L]50 /([A]/K A + 1) (2)
contract research organizations (e.g. CEREP, MDL) offer a This is often known as the Cheng–Prusoff equation, and
battery of assays – mainly binding assays, but also a range is widely used to calculate KL when [L]50, [A] and KA are
of functional assays – designed to detect affinity for a known. It is important to realize that the Cheng–Prusoff
wide range of receptors, transporters and channels. In the equation applies only (a) at equilibrium, (b) when the
field of monoamine receptors, for example, it is usually interaction between A and L is strictly competitive, and (c)
important to avoid compounds that block or activate when neither ligand binds cooperatively. However, an [L]50
peripheral muscarinic receptors, adrenergic receptors or value can be measured for any test compound that inhibits
histamine (particularly H1) receptors, because of the the binding of the radioligand by whatever mechanism,
side effects that are associated with these actions, and a irrespective of whether equilibrium has been reached.
standard selectivity test battery allows such problems to Applying the Cheng–Prusoff equation if these conditions
be discovered early. Recently, several psychotropic and are not met can yield estimates of KL that are quite mean-
anti-infective drugs have been withdrawn because of ingless, and so it should strictly be used only if the condi-
sudden cardiac deaths, probably associated with their tions have been shown experimentally to be satisfied – a
ability to block a particular type of potassium channel fairly laborious process. Nevertheless, Cheng–Prusoff
(known as the hERG channel; see Chapter 16) in myocar- estimates of ligand affinity constants are often quoted
dial cells. This activity can be detected by electrophysio- without such checks having been performed. In most cases
logical measurements on isolated myocardial cells, and it would be more satisfactory to use the experimentally
such a test is now usually performed at an early stage of determined [L]50 value as an operational measure of
development of drugs of the classes implicated in this type potency. A further important caveat that applies to binding
of adverse reaction. studies is that they are often performed under conditions
of low ionic strength, in which the sodium and calcium
concentrations are much lower than the physiological
Interpretation of binding assays
range. This is done for technical reasons, as low [Na+]
Binding assays, generally with membrane preparations commonly increases both the affinity and the Bmax of the
made from intact tissues or receptor-expressing cell lines, radioligand, and omitting [Ca2+] avoids clumping of the
are widely used in drug discovery projects because of their membrane fragments. Partly for this reason, ligand affini-
simplicity and ease of automation. Detailed technical ties estimated from binding studies are often considerably
manuals describing the methods used for performing and higher than estimates obtained from functional assays

159
Section | 2 | Drug Discovery

5HT3 receptors

100
Ki (nM) functional assay

10

1
0.01 0.1 1 10 100
A Ki (nM) binding assay

10000
5HT4 receptors

1000
Ki (nM) functional assay

100

10
0.1 1 10 100 1000
B Ki (nM) binding assay

Fig. 11.1  Correlation of binding and functional data for 5HT receptor ligands. (A) 5HT3 receptors, (B) 5HT4 receptors.
Data from Heidempergher et al., 1997.
Data from Yang et al., 1997.

(Hall, 1992), although the effect is not consistent, presum- often rather poor (see below). Figure 11.1 shows data
ably because ionic bonding, which will be favoured by the obtained independently on 5HT3 and 5HT4 receptors; in
low ionic strength medium, contributes unequally to the both cases the estimated KD values for binding are on
binding of different ligands. Consequently, the correlation average about 10 times lower than estimates from func-
between data from binding assays and functional assays is tional assays, and the correlation is very poor.

160
 Pharmacology: its role in drug discovery Chapter | 11 |

• What happens if the drug is given continuously or

PHARMACOLOGICAL PROFILING repeatedly to an animal over the course of days or
weeks? Does it lose its effectiveness, or reveal effects
Pharmacological profiling aims to determine the pharma- not seen with acute administration? Is there any kind
codynamic effects of the new compound – or more often of ‘rebound’ after effect when it is stopped?
of a small family of compounds – on in vitro model
systems, e.g. cell lines or isolated tissues, normal animals,
and animal models of disease. The last of these is particu- In vitro profiling
larly important, as it is intended to give the first real Measurements on isolated tissues
pointer to therapeutic efficacy as distinct from pharmaco-
dynamic activity. It is valuable to assess the activity of the Studies on isolated tissues have been a mainstay of phar-
compounds in a series of assays representing increasingly macological methodology ever since the introduction of
complex levels of organization. The choice of test systems the isolated organ bath by Magnus early in the 20th
depends, of course, on the nature of the target. For exam­ century. The technique is extremely versatile and applica-
ple, characterization of a novel antagonist of a typical G- ble to studies on smooth muscle (e.g. gastrointestinal
protein-coupled receptor might involve the following: tract, airways, blood vessels, urinary tract, uterus, biliary
tract, etc.) as well as cardiac and striated muscle, secretory
• Ligand-binding assay on membrane fragments from epithelia, endocrine glands, brain slices, liver slices, and
a cell line expressing the cloned receptor
many other functional systems. In most cases the tissue
• Inhibition of agonist activity in a cell line, based is removed from a freshly killed or anaesthetized animal
on a functional readout (e.g. raised intracellular
and suspended in a chamber containing warmed oxygen-
calcium)
ated physiological salt solution. With smooth muscle
• Antagonism of a selective agonist in an isolated preparations the readout is usually mechanical (i.e.
tissue (e.g. smooth muscle, cardiac muscle). Such
tension, recorded with a simple strain gauge). For other
assays will normally be performed with non-human
types of preparation, various electrophysiological or bio-
tissue, and so interspecies differences in the receptor
chemical readouts are often used. Vogel (2002) and
need to be taken into account. Sometimes specific
Enna et al. (2003) give details of a comprehensive range
questions have to be asked about effects on human
of standard pharmacological assay methods, including
tissues for particular compounds and then collecting
technical instructions.
viable tissues to use becomes a major challenge
Studies of this kind have the advantage that they are
• Antagonism of the response (e.g. performed on intact normal tissues, as distinct from iso-
bronchoconstriction, vasoconstriction, increased
lated enzymes or other proteins. The recognition mole-
heart rate) to a selective receptor agonist in vivo.
cules, signal transduction machinery and the mechanical
Prior knowledge about species specificity of the
or biochemical readout are assumed to be a reasonable
agonist and antagonist is important at this stage.
approximation to the normal functioning of the tissue.
Pharmacological profiling is designed as a hypothesis- There is abundant evidence to show that tissue responses
driven programme of work, based on the knowledge pre­ to GPCR activation, for example, depend on many factors,
viously gained about the activity of the compound on including the level of expression of the receptor, the type
its specific target or targets. In this respect it differs from and abundance of the G proteins present in the cell, the
safety pharmacology (see below), which is an open- presence of associated proteins such as receptor activity-
minded exercise designed to detect unforeseen effects. The modifying proteins (RAMPs; see Morfis et al., 2003), the
aim of pharmacological profiling is to answer the follow- state of phosphorylation of various constituent proteins
ing questions: in the signal transduction cascade, and so on. For com-
• Do the molecular and cellular effects measured in pounds acting on intracellular targets, functional activity
screening assays actually give rise to the predicted depends on permeation through the membrane, as well as
pharmacological effects in intact tissues and whole affinity for the target. For these reasons – and probably
animals? also for others that are not understood – the results of
• Does the compound produce effects in intact tissues assays on isolated tissues often differ significantly from
or whole animals not associated with actions on its results found with primary screening assays. The discrep-
principal molecular target? ancy may simply be a quantitative one, such that the
• Is there correspondence between the potency of the potency of the ligand does not agree in the two systems,
compound at the molecular level, the tissue level or it may be more basic. For example, the pharmacological
and the whole animal level? efficacy of a receptor ligand, i.e. the property that deter-
• Do the in vivo potency and duration of action match mines whether it is a full agonist, a partial agonist, or an
up with the pharmacokinetic properties of the antagonist, often depends on the type of assay used
compound? (Kenakin, 1999), and this may have an important bearing

161
Section | 2 | Drug Discovery

100

10
IC50 (µM) cellular assays

0.1

0.01 Tyrosine phosphorylation

MK cell proliferation

0.001
0.001 0.01 0.1 1 10
IC50 (µM) isolated enzyme

Fig. 11.2  Correlation of cellular activity of EGFR receptor kinase inhibitors with enzyme inhibition.
Data from Traxler et al., 1997.

on the selection of possible development compounds. muscle contraction, cell motility, secretion and release of
Examples that illustrate the poor correlation that may exist mediators, transmembrane ion fluxes, vascular resistance
between measurements of target affinity in cell-free assay and permeability, and epithelial transport and permeabil-
systems, and functional activity in intact cell systems, are ity. This versatility and the relative technical simplicity of
shown in Figures 11.1 and 11.2. Figure 11.1 shows the many such methods are useful attributes for drug discov-
relationship between binding and functional assay data ery. Additional advantages are that concentration–effect
for 5HT3 and 5HT4 receptor antagonists. In both cases, relationships can be accurately measured, and the design
binding assays overestimate the potency in functional of the experiments is highly flexible, allowing rates of
assays by a factor of about 10 (see above), but more impor- onset and recovery of drug effects to be determined, as well
tantly, the correlation is poor, despite the fact that the as measurements of synergy and antagonism by other
receptors are extracellular, and so membrane penetration compounds, desensitization effects, etc.
is not a factor. Figure 11.2 shows data on tyrosine kinase The main shortcomings of isolated tissue pharmacology
inhibitors, in which activity against the isolated enzyme is are (a) that tissues normally have to be obtained from
plotted against inhibition of tyrosine phosphorylation in small laboratory animals, rather than humans or other
intact cells, and inhibition of cell proliferation for a large primates; and (b) that preparations rarely survive for
series of compounds. Differences in membrane penetra- more than a day, so that only short-term experiments are
tion can account for part of the discrepancy between feasible.
enzyme and cell-based data, but the correlation between
intracellular kinase inhibition and blocking of cell prolif-
In vivo profiling
eration is also weak, which must reflect other factors.
It is worth noting that these examples come from very As already mentioned, experiments on animals have
successful drug discovery projects. The quantitative dis- several drawbacks. They are generally time-consuming,
crepancies that we have emphasized, though worrying to technically demanding and expensive. They are subject to
pharmacologists, should not therefore be a serious distrac- considerable ethical and legal constraints, and in some
tion in the context of a drug discovery project. countries face vigorous public opposition. For all these
A very wide range of physiological responses can be reasons, the number of experiments is kept to a bare
addressed by studies on isolated tissues, including meas- minimum, and experimental variability is consequently
urements of membrane excitability, synaptic function, often a problem. Animal experiments must, therefore, be

162
 Pharmacology: its role in drug discovery Chapter | 11 |

used very selectively and must be carefully planned and

designed so as to produce the information needed as effi-
Box 11.1  Pharmacological profiling of beraprost
ciently as possible. In the past, before target-directed
In vitro studies
approaches were the norm, routine in vivo testing was
often used as a screen at a very early stage in the drug Binding to PGI2 receptors of platelets from various
discovery process, and many important drugs (e.g. thiazide species, including human
diuretics, benzodiazepines, ciclosporin) were discovered PGI2 agonist activity (cAMP formation) in platelets
on the basis of their effects in vivo. Nowadays, the use of Dilatation of arteries and arterioles in vitro, taken from
in vivo methods is much more limited, and will probably various species
decline further in response to the pressures on time and Increased red cell deformability (hence reduced blood
costs, as alternative in vitro and in silico methods are viscosity and increased blood flow) in blood taken from
developed, and as public attitudes to animal experimenta- hypercholesterolaemic rabbits
tion harden. An additional difficulty is the decreasing In vivo studies
number of pharmacologists trained to perform in vivo
Increased peripheral blood flow in various vascular regions
studies1.
(dogs)
Imaging technologies (Rudin and Weissleder, 2003; see
Cutaneous vasodilatation (rat)
also Chapter 18) are increasingly being used for pharma-
Reduced pulmonary hypertension in rat model of
cological studies on whole animals. Useful techniques
drug-induced pulmonary hypertension (measured by
include magnetic resonance imaging (MRI), ultrasound
reduction of right ventricular hypertrophy)
imaging, X-ray densitometry tomography, positron emis-
Reduced tissue destruction (gangrene) of rat tail induced
sion tomography (PET) and others. They are proving
by ergotamine/epinephrine infusion
highly versatile for both structural measurements (e.g.
cardiac hypertrophy, tumour growth) and functional Reduction of vascular occlusion resulting from intra-
arterial sodium laureate infusion in rats
measurements (e.g. blood flow, tissue oxygenation). Used
in conjunction with radio-active probes, PET can be used Reduction of vascular occlusion and thrombosis following
for studies on receptors and other targets in vivo. Many of electrical stimulation of femoral artery in anaesthetized
dogs and rabbits
these techniques can also be applied to humans, providing
an important bridge between animal and human pharma- Reduction of vascular damage occurring several weeks
cology. Apart from the special facilities and equipment after cardiac allografts in immunosuppressed rats
needed, currently the main drawback of imaging tech-
niques is the time taken to capture the data, during which
the animal must stay still, usually necessitating anaesthe-
sia. With MRI and PET, which are currently the most ver-
satile imaging techniques, data capture normally takes a plethora of possible studies that might be performed to
few minutes, so they cannot be used for quick ‘snapshots’ characterize a particular drug can be difficult.
of rapidly changing events. A typical example of pharmacological profiling is sum-
A particularly important role for in vivo experiments is marized in Box 11.1. The studies were carried out as part
to evaluate the effects of long-term drug administration on of the recent development of a cardiovascular drug, berap-
the intact organism. ‘Adaptive’ and ‘rebound’ effects (e.g. rost (Melini and Goa, 2002). Beraprost is a stable analogue
tolerance, dependence, rebound hypertension, delayed of prostaglandin I2 (PGI2) which acts on PGI2 receptors of
endocrine effects, etc.) are often produced when drugs are platelets and blood vessels, thereby inhibiting platelet
given continuously for days or weeks. Generally, such aggregation (and hence thrombosis) and dilating blood
effects, which involve complex physiological interactions, vessels. It is directed at two therapeutic targets, namely
are evident in the intact functioning organism but are not occlusive peripheral vascular disease and pulmonary
predictable from in vitro experiments. hypertension (a serious complication of various types of
The programme of in vivo profiling studies for charac- cardiovascular disease, drug treatment or infectious dis-
terization of a candidate drug depends very much on the eases), resulting in hypertrophy and often contractile
drug target and therapeutic indication. A comprehensive failure of the right ventricle. The animal studies were,
catalogue of established in vivo assay methods appropriate therefore, directed at measuring changes (reduction in
to different types of pharmacological effect is given by blood flow, histological changes in vessel wall) associated
Vogel (2002). Charting the appropriate course through the with peripheral vascular disease, and with pulmonary
hypertension. As these are progressive chronic conditions,
it was important to establish that long-term systemic
1
administration of beraprost was effective in retarding the
The rapid growth in the use of transgenic animals to study the
functional role of individual gene products has recently brought in development of the experimental lesions, as well as moni-
vivo physiology and pharmacology back into fashion, however. toring the acute pharmacodynamic effects of the drug.

163
Section | 2 | Drug Discovery

1000
Ki (nM) human receptor binding assay

100

10

1
0.01 0.1 1 10 100 1000
Ki (nM) rat receptor binding assay

Fig. 11.3  Species differences in bradykinin B2 receptors.

Data from Dziadulewicz et al., 2002.

Species differences Species differences are, of course, one of the main argu-
ments used by animal rights activists in opposing the
It is important to take species differences into account at use of animals for the purpose of drug discovery. Their
all stages of pharmacological profiling. For projects based claim – misleading when examined critically (see Under-
on a defined molecular target – nowadays the majority – standing Animal Research website) – is that animal data
the initial screening assay will normally involve the human actually represent disinformation in this context. While
isoform. The same target in different species will generally being aware of the pitfalls, we should not lose sight of the
differ in its pharmacological specificity; commonly, there fact that non-human data, including in vivo experiments,
will be fairly small quantitative differences, which can be have actually been an essential part of every major drug
allowed for in interpreting pharmacological data in experi- discovery project to date. The growing use of transgenic
mental animals, but occasionally the differences are large, animal models has led to an increase, rather than a
so that a given class of compounds is active in one species decrease, in animal experimentation, as even breeding
but not in another. An example is shown in Figure 11.3, such animals is counted as an experiment for statistical
which compares the activities of a series of bradykinin purposes.
receptor antagonists on cloned human and rat receptors.
The complete lack of correlation means that, for these
compounds, tests of functional activity in the rat cannot
be used to predict activity in man. ANIMAL MODELS OF DISEASE
Species differences are, in fact, a major complicating
factor at all stages of drug discovery and preclinical devel- The animal models discussed earlier were used to investi-
opment. The physiology of disease processes such as gate the pharmacodynamic effects of the drug and to
inflammation, septic shock, obesity, atherosclerosis, etc., answer the question: How do the effects observed at the
differs markedly in different species. Most importantly molecular and cellular levels of organization translate into
(see Chapter 10), drug metabolism often differs, affecting physiological effects in the whole animal?
the duration of action, as well as the pattern of metabo- The next, crucial, question is: Can these physiological
lites, which can in turn affect the observed pharmacology effects result in therapeutic benefit? Animal experiments
and toxicity. can never answer this conclusively – only clinical trials

164
 Pharmacology: its role in drug discovery Chapter | 11 |

can do that – but the use of animal models of human • ‘Kindling’ and other procedures for inducing
disease provides a valuable link in the chain of evidence, ongoing seizures as models for epilepsy
and there is strong pressure on drug discovery teams to • Self-administration of opiates, nicotine or other
produce data of this sort as a basis for the important deci- drugs as a model for drug-dependence
sion to test a new compound in man. Despite the immense • Cholesterol-fed rabbits as a model for
range and diversity of animal models that have been hypercholesterolaemia and atherosclerosis
described, this is often the most problematic aspect of a • Immunization with myelin basic protein as a model
drug discovery project, particularly where a novel target or for multiple sclerosis
mechanism is involved, so that there is no mechanistic • Administration of the neurotoxin MPTP, causing
precedent among established drugs. The magnitude of the degeneration of basal ganglia neurons as a model of
difficulties varies considerably among different therapeu- Parkinson’s disease
tic areas. Many inflammatory conditions, for example, are • Transplantation of malignant cells into
straightforward to model in animals, as are some cancers. immunodeficient animals to produce progressive
Animal models of hypertension generally predict very well tumours as a model for certain types of cancer.
the ability of compounds to lower blood pressure in man. Details of these and many other examples of physiologi-
Endocrine disorders involving over- or undersecretion of cal and pharmacological models can be found in Vogel
particular hormones can also be simply modelled in (2002). As discussed above, species differences need to be
animals. Psychiatric disorders are much more difficult, as taken into account in the selection of animal models, and
the symptoms that characterize them are not observable in the interpretation of results. In septic shock, for example,
in animals. In most therapeutic areas there are certain rodents show a much larger elevation of nitric oxide (NO)
disorders, such as migraine, temporal lobe epilepsy, metabolites than do humans, and respond well to NO
asthma or irritable bowel syndrome, for which animal synthesis inhibitors, which humans do not. Rodents and
models, if they exist at all, are far from satisfactory in rabbits transgenically engineered to favour cholesterol
predicting clinical efficacy. deposition nevertheless develop atherosclerosis only when
Here we consider, with a few selected examples, the fed high-cholesterol diets, whereas humans often do so
main experimental approaches to generating animal even on low-cholesterol diets. Genetically obese mice are
models, and the criteria against which their ‘validity’ as deficient in the hormone leptin and lose weight when
models of human disease need to be assessed. treated with it, whereas obese humans frequently have
high circulating leptin concentrations and do not respond
to treatment with it. It is often not clear whether such
Types of animal model
discrepancies reflect inherent species differences, or simply
Animal models of disease can be divided broadly into failure of the model to replicate satisfactorily the predomi-
acute and chronic physiological and pharmacological nant human disease state (see Validity criteria below).
models, and genetic models.
Acute physiological and pharmacological models are
intended to mimic certain aspects of the clinical disorder.
Genetic models
There are many examples, including: There are many examples of spontaneously occurring
animal strains that show abnormalities phenotypically
• Seizures induced by electrical stimulation of the
resembling human disease. In addition, much effort is
brain as a model for epilepsy (see below)
going into producing transgenic strains with deletion or
• Histamine-induced bronchoconstriction as a model
over-expression of specific genes, which also exhibit
for asthma
disease-like phenotypes.
• The hotplate test for analgesic drugs as a model for
Long before genetic mapping became possible, it was
pain
realized that certain inbred strains of laboratory animal
• Injection of lipopolysaccharide (LPS) and cytokines
were prone to particular disorders, examples being spon-
as a model for septic shock
taneously hypertensive rats, seizure-prone dogs, rats
• The elevated maze test as a model for testing
insensitive to antidiuretic hormone (a model for diabetes
anxiolytic drugs.
insipidus), obese mice and mouse strains exhibiting a
Chronic physiological or pharmacological models involve the range of specific neurological deficits. Many such strains
use of drugs or physical interventions to induce an ongoing have been characterized (see Jackson Laboratory website,
abnormality similar to the clinical condition. Examples www.jaxmice.jax.org) and are commercially available, and
include: are widely used as models for testing drugs.
• The use of alloxan to inhibit insulin secretion as a The development of transgenic technology has allowed
model for Type I diabetes inbred strains to be produced that over- or under-express
• Procedures for inducing brain or coronary ischaemia particular genes. In the simplest types, the gene abnormal-
as models for stroke and ischaemic heart disease ity is present throughout the animal’s life, from early

165
Section | 2 | Drug Discovery

development onwards, and throughout the body. More from rats. Success in producing gene knockout strains by
recent technical developments allow much more control an alternative method has now been achieved (Zan et al.,
over the timing and location of the transgene effect. For 2003), and the use of transgenic rats is increasing, this
reviews of transgenic technology and its uses in drug dis- being the favoured species for pharmacological and physi-
covery, see Polites (1996), Rudolph and Moehler (1999), ological studies in many laboratories.
Törnell and Snaith (2002) and Pinkert (2002).
The genetic analysis of disease-prone animal strains, or
of human families affected by certain diseases, has in The choice of model
many cases revealed the particular mutation or mutations Apart from resource limitations, regulatory constraints on
responsible (see Chapters 6 and 7), thus pointing the animal experimentation, and other operational factors,
way to new transgenic models. Several diseases associated what governs the choice of disease model?
with single-gene mutations, such as cystic fibrosis and As discussed in Chapter 2, naturally occurring diseases
Duchenne muscular dystrophy, have been replicated in trans- produce a variety of structural biochemical abnormalities,
genic mouse strains. Analysis of the obese mouse strain and these are often displayed separately in animal
led to the identification of the leptin gene, which is models. For example, human allergic asthma involves:
mutated in the ob/ob mouse strain, causing the production (a) an immune response; (b) increased airways resistance;
of an inactive form of the hormone and overeating by (c) bronchial hyperreactivity; (d) lung inflammation;
the mouse. Transgenic animals closely resembling ob/ob and (e) structural remodelling of the airways. Animal
mice have been produced by targeted inactivation of the models, mainly based on guinea pigs, whose airways
gene for leptin or its receptor. Another example is the behave similarly to those of humans, can replicate each
discovery that a rare familial type of Alzheimer’s disease is of these features, but no single model reproduces the
associated with mutations of the amyloid precursor whole spectrum. The choice of animal model for drug
protein (APP). Transgenic mice expressing this mutation discovery purposes, therefore, depends on the therapeutic
show amyloid plaque formation characteristic of the effect that is being sought. In the case of asthma, existing
human disease. This and other transgenic models of bronchodilator drugs effectively target the increased
Alzheimer’s disease (Yamada and Nabeshima, 2000) rep- airways resistance, and steroids reduce the inflammation,
resent an important tool for drug discovery, as there had and so it is the other components for which new drugs
hitherto been no animal model reflecting the pathogenesis are particularly being sought.
of this disorder. A similar need for a range of animal models covering a
The number of transgenic animal models, mainly range of therapeutic targets applies in many disease areas.
mouse, that have been produced is already large and is
growing rapidly. Creating and validating a new disease
model is, however, a slow business. Although the method- Validity criteria
ology for generating transgenic mice is now reliable and
Obviously an animal model produced in a laboratory can
relatively straightforward, it is both time-consuming and
never replicate exactly a spontaneous human disease state,
labour-intensive. The first generation of transgenic animals
so on what basis can we assess its ‘validity’ in the context
are normally hybrids, as different strains are used for the
of drug discovery?
donor and the recipient, and it is necessary to breed several
Three types of validity criteria were originally proposed
generations by repeated back-crossings to create animals
by Willner (1984) in connection with animal models of
with a uniform genetic background. This takes 1–2 years,
depression. These are:
and is essential for consistent results. Analysis of the phe-
notypic changes resulting from the transgene can also • Face validity
be difficult and time-consuming, as the effects may be • Construct validity
numerous and subtle, as well as being slow to develop as • Predictive validity.
the animal matures. Despite these difficulties, there is no Face validity refers to the accuracy with which the model
doubt that transgenic disease models are playing an reproduces the phenomena (symptoms, clinical signs and
increasing part in drug testing, and many biotechnology pathological changes) characterizing the human disease.
companies have moved into the business of developing Construct validity refers to the theoretical rationale on
and providing them for this purpose. The fields in which which the model is based, i.e. the extent to which the
transgenic models have so far had the most impact are aetiology of the human disease is reflected in the model.
cancer, atherosclerosis and neurodegenerative diseases, A transgenic animal model in which a human disease-
but their importance as drug discovery tools extends to producing mutation is replicated will have, in general,
all areas. good construct validity, even if the manifestations of the
Producing transgenic rat strains proved impossible until human disorder are not well reproduced (i.e. it has poor
recently, as embryonic stem (ES) cells cannot be obtained face validity).

166
 Pharmacology: its role in drug discovery Chapter | 11 |

Predictive validity refers to the extent to which the effect Epilepsy models
of manipulations (e.g. drug treatment) in the model is
The development of antiepileptic drugs, from the pioneer-
predictive of effects in the human disorder. It is the most
ing work of Merritt and Putnam, who in 1937 developed
pragmatic of the three and the most directly relevant to
phenytoin, to the present day, has been highly dependent
the issue of predicting therapeutic efficacy, but also the
on animal models involving experimentally induced sei-
most limited in its applicability, for two main reasons.
zures, with relatively little reliance on knowledge of the
First, data on therapeutic efficacy are often sparse or non-
underlying physiological, cellular or molecular basis of the
existent, because no truly effective drugs are known (e.g.
human disorder. Although existing drugs have significant
for Alzheimer’s disease, septic shock). Second, the model
limitations, they have brought major benefits to sufferers
may focus on a specific pharmacological mechanism, thus
from this common and disabling condition – testimony
successfully predicting the efficacy of drugs that work by
to the usefulness of animal models in drug discovery.
that mechanism but failing with drugs that might prove
Human epilepsy is a chronic condition with many
effective through other mechanisms. The knowledge that
underlying causes, including head injury, infections,
the first generation of antipsychotic drugs act as dopamine
tumours and genetic factors. Epileptic seizures in humans
receptor antagonists enabled new drugs to be identified by
take many forms, depending mainly on where the neural
animal tests reflecting dopamine antagonism, but these
discharge begins and how it spreads.
tests cannot be relied upon to recognize possible ‘break-
Some of the widely used animal models used in drug
through’ compounds that might be effective by other
discovery are summarized in Table 11.2. The earliest
mechanisms. Thus, predictive validity, relying as it does on
models, namely the maximal electroshock (MES) test and
existing therapeutic knowledge, may not be a good basis
the pentylenetetrazol-induced seizure (PTZ) test, which are
for judging animal models where the drug discovery team’s
based on acutely induced seizures in normal animals, are
aim is to produce a mechanistically novel drug. The basis
still commonly used. They model the seizure, but without
on which predictive validity is judged carries an inevitable
distinguishing its localization and spread, and do not
bias, as the drugs that proceed to clinical trials will nor-
address either the chronicity of human epilepsy or its
mally have proved effective in the model, whereas drugs
aetiology (i.e. they score low on face validity and construct
that are ineffective in the model are unlikely to have been
validity). But, importantly, their predictive validity for con-
developed. As a result, there are many examples of tests
ventional antiepileptic drugs in man is very good, and the
giving ‘false positive’ expectations, but very few false nega-
drugs developed on this basis, taken regularly to reduce
tives, giving rise to a commonly held view that conclusions
the frequency of seizures or eliminate them altogether, are
from pharmacological tests tend to be overoptimistic.
of proven therapeutic value. Following on from these acute
seizure models, attempts have been made to replicate the
Some examples processes by which human epilepsy develops and contin-
ues as a chronic condition with spontaneous seizures, i.e.
We conclude this discussion of the very broad field of to model epileptogenesis (Löscher, 2002; White, 2002) by
animal models of disease by considering three disease the use of models that show greater construct and face
areas, namely epilepsy, psychiatric disorders and stroke. validity. This has been accomplished in a variety of ways
Epilepsy-like seizures can be produced in laboratory (see Table 11.2) in the hope that such models would be
animals in many different ways. Many models have been helpful in developing drugs capable of preventing epi-
described and used successfully to discover new anti- lepsy. Such models have thrown considerable light on the
epileptic drugs (AEDs). Although the models may lack pathogenesis of epilepsy, but have not so far contributed
construct validity and are weak on face validity, their pre- significantly to the development of improved antiepileptic
dictive validity has proved to be very good. With models drugs. Because there are currently no drugs known to
of psychiatric disorders, face validity and construct validity prevent epilepsy from progressing, the predictive validity
are very uncertain, as human symptoms are not generally of epileptogenesis models remains uncertain.
observable in animals and because we are largely ignorant
of the cause and pathophysiology of these disorders; nev-
ertheless, the predictive validity of available models of Psychiatric disorders
depression, anxiety and schizophrenia has proved to be Animal models of psychiatric disorders are in general
good, and such models have proved their worth in drug problematic, because in many cases the disorders are
discovery. In contrast, the many available models of stroke defined by symptoms and behavioural changes unique
are generally convincing in terms of construct and face to humans, rather than by measurable physiological,
validity, but have proved very unreliable as predictors of biochemical or structural abnormalities. This is true in
clinical efficacy. Researchers in this field are ruefully aware conditions such as schizophrenia, Tourette’s syndrome
that despite many impressive effects in laboratory animals, and autism, making face validity difficult to achieve.
clinical successes have been negligible. Depressive symptoms, in contrast, can be reproduced to

167
Section | 2 | Drug Discovery

Table 11.2  Epilepsy and epileptogenesis models

Model Procedure Face validity Construct Predictive validity

validity
Acute seizure No acute seizure models
models show ‘treatment-resistance’
to conventional antiepileptic
drugs, though this occurs in
~30% of human cases
Maximal Acute seizures evoked Weak. No Weak. Good. Predictive of activity
electroshock by whole-brain spontaneous Production of of drugs against partial
model stimulation. Measure: seizures. No seizures not seizures and generalized
proportion of mice neuropathological related to tonic–clonic seizures (with
responding with changes epileptogenesis some false positives). Poor
seizures prediction of drugs effective
in absence seizures
Pentylenetetrazole Seizures induced by s.c. As above As above Quite good as predictor of
(PTZ)-induced injection of the efficacy in absence seizures.
seizure model convulsant drug PTZ. Unreliable for other clinical
Measure: proportion of types
mice responding with
seizures
Epileptogenesis
models
Kindling model Weak electrical Moderate, though Uncertain Moderate, but model is
stimulation of spontaneous seizures generally more drug-
amygdala repeated are rarely produced. responsive than human
over several days. Histological, epilepsy
Evoked full-blown electrophysiological
seizures develop and biochemical
gradually changes similar to
human epilepsy
Post-seizure Various procedures Good. Spontaneous Probably good As MES for anti-seizure
models (e.g. injection of seizures, latent for some drugs. Uncertain for
kainate, lithium or period after initial clinical forms antiepileptogenic drugs, of
other agents into brain, trigger. Replicates of epilepsy which there are no proven
sustained stimulation histological and clinical examples
of amygdala or other other changes,
pathways) evoke including
sustained seizures, with neurodegeneration
ongoing spontaneous
seizures appearing days
or weeks later. Can be
used to test drug
effects on fully kindled
seizures, or on the
kindling process
Surgical Cortical undercutting. Good as model of Probably good As above
procedures Isolated region of post-traumatic for post-
cortex gradually epilepsy. Replicates traumatic
develops spontaneous histological and epilepsy
seizure activity other changes

168
 Pharmacology: its role in drug discovery Chapter | 11 |

some extent in animal models (Willner and Mitchell, advances. Drugs of many types, including glutamate antag-
2002), and face validity is therefore stronger. The aetiology onists, calcium and sodium channel blocking drugs, anti-
of most psychiatric conditions is largely unknown2, inflammatory drugs, free radical scavengers and others,
making construct validity questionable. produced convincing degrees of neuroprotection in animal
Models are therefore chosen largely on the basis of pre- models, even when given up to several hours after the
dictive validity, and suffer from the shortcomings men- ischaemic event. Many clinical trials were undertaken (De
tioned above. Nonetheless, models for some disorders, Keyser et al., 1999), with uniformly negative results. The
particularly depression, have proved very valuable in the only drug currently known to have a beneficial – albeit
discovery of new drugs. Other disorders, such as autism small – effect is the biopharmaceutical ‘clot-buster’ tissue
and Tourette’s syndrome, have proved impossible to model plasminogen activator (TPA), widely used to treat heart
so far, whereas models for others, such as schizophrenia attacks. Stroke models thus represent approaches that have
(Lipska and Weinberger, 2000; Moser et al., 2000), have revealed much about pathophysiology and have stimu-
been described but are of doubtful validity. The best predic- lated intense efforts in drug discovery, but whose predic-
tion of antipsychotic drug efficacy comes from pharmaco- tive validity has proved to be extremely poor, as the drug
dynamic models reflecting blockade of dopamine and sensitivity of the animal models seems to be much greater
other monoamine receptors, rather than from putative than that of the human condition. Surprisingly, it appears
disease models, with the result that drug discovery has so that whole-brain ischaemia models show better predictive
far failed to break out of this mechanistic straitjacket. validity (i.e. poor drug responsiveness) than focal ischae-
mia models, even though the latter are more similar to
Stroke human strokes.
Many experimental procedures have been devised to
produce acute cerebral ischaemia in laboratory animals,
resulting in long-lasting neurological deficits that resem- Good laboratory practice  
ble the sequelae of strokes in humans (Small and Buchan, (GLP) compliance in  
2000). Interest in this area has been intense, reflecting pharmacological studies
the fact that strokes are among the commonest causes
of death and disability in developed countries, and that GLP comprises adherence to a set of formal, internation-
there are currently no drugs that significantly improve ally agreed guidelines established by regulatory authori-
the recovery process. Studies with animal models have ties, aimed at ensuring the reliability of results obtained
greatly advanced our understanding of the pathophys­ in the laboratory. The rules (see GLP Pocketbook, 1999;
iological events. Stroke is no longer seen as simple EEC directives 87/18/EEC, 88/320/EEC, available online:
anoxic death of neurons, but rather as a complex series of pharmacos.eudra.org/F2/eudralex/vol-7/A/7AG4a.pdf)
events involving neuronal depolarization, activation of cover all stages of an experimental study, from planning
ion channels, release of excitatory transmitters, disturbed and experimental design to documentation, reporting and
calcium homeostasis leading to calcium overload, release archiving. They require, among other things, the assign-
of inflammatory mediators and nitric oxide, generation of ment of specific GLP-compliant laboratories, certification
reactive oxygen species, disturbance of the blood–brain of staff training to agreed standards, certified instrument
barrier and cerebral oedema (Dirnagl et al., 1999). Glial calibration, written standard operating procedures cover-
cells, as well as neurons, play an important role in the ing all parts of the work, specified standards of experimen-
process. Irreversible loss of neurons takes place gradually tal records, reports, notebooks and archives, and much
as this cascade builds up, leading to the hope that inter- else. Standards are thoroughly and regularly monitored by
vention after the primary event – usually thrombosis – an official inspectorate, which can halt studies or require
could be beneficial. Moreover, the biochemical and changes in laboratory practice if the standards are thought
cellular events involve well-understood signalling mecha- not to be adequately enforced. Adherence to GLP stand-
nisms, offering many potential drug targets, such as ards carries a substantial administrative overhead and
calcium channels, glutamate receptors, scavenging of reac- increases both the time and cost of laboratory studies, as
tive oxygen species and many others. Ten years ago, on the well as limiting their flexibility.
basis of various animal models with apparently good con- The regulations are designed primarily to minimize the
struct and face validity and a range of accessible drug risk of errors in studies that relate to safety. They are,
targets, the stage seemed to be set for major therapeutic therefore, not generally applied to pharmacological pro­
filing as described in this chapter. They are obligatory
for toxicological studies that are required in submissions
2
Many psychiatric disorders have a strong genetic component in for regulatory approval. Though not formally required for
their aetiology, and much effort has gone into identifying particular safety pharmacology studies, most companies and con-
susceptibility genes. Success has so far been very limited, but the
expectation is that, in future, success in this area will enable improved tract research organizations choose to do such work under
transgenic animal models to be developed. GLP conditions.

169
Section | 2 | Drug Discovery

REFERENCES

Current protocols in pharmacology. Lipska BK, Weinberger DR. To model a in pharmacological research:
New York: John Wiley and Sons. psychiatric disorder in animals: future trends. European Journal
[Published as looseleaf binder and schizophrenia as a reality test. of Pharmacology 1999;375:
CD-ROM, and regularly updated.) Neuropsychopharmacology 327–37.
De Keyser J, Sulter G, Luiten PG. 2000;23:223–39. Small DL, Buchan AM. Stroke: animal
Clinical trials with neuroprotective Löscher W. Animal models of epilepsy models. British Medical Bulletin
drugs in ischaemic stroke: are we for the development of 2000;56:307–17.
doing the right thing? Trends in antiepileptogenic and disease- Törnell J, Snaith M. Transgenic systems
Neurosciences 1999;22:535–40. modifying drugs. A comparison of in drug discovery: from target
Dirnagl U, Iadecola C, Moskowitz MA. the pharmacology of kindling and identification to humanized mice.
Pathobiology of ischaemic stroke: post-status epilepticus models of Drug Discovery Today 2002;7:
an integrated view. Trends in temporal lobe epilepsy. Epilepsy 461–70.
Neurosciences 1999;22:391–7. Research 2002;50:105–23. Traxler P, Bold G, Frei J, et al. Use of a
Dziadulewicz EK, Ritchie TJ, Hallett A, Melini EB, Goa KL. Beraprost: a review pharmacophore model for the
et al. Nonpeptide bradykinin B2 of its pharmacology and therapeutic design of EGF-R tyrosine kinase
receptor antagonists: conversion of efficacy in the treatment of inhibitors: 4-(phenylamino)
rodent-selective bradyzide analogues peripheral arterial disease and pyrazolo[3,4-d]pyrimidines. Journal
into potent orally active human pulmonary hypertension. Drugs of Medicinal Chemistry
bradykinin B2 receptor antagonists. 2002;62:107–33. 1997;40:3601–16.
Journal of Medicinal Chemistry Morfis M, Christopoulos A, Sexton PM. Vogel WH, editor. Drug discovery
2002;45:2160–72. RAMPs: 5 years on, where to now? and evaluation: pharmacological
Enna S, Williams M, Ferkany JW, et al., Trends in Pharmacological Sciences assays. Heidelberg: Springer-Verlag;
editors. Current protocols in 2003;24:596–601. 2002.
pharmacology. Hoboken, NJ: Wiley; Moser PC, Hitchcock JH, Lister S, et al. White HS. Animal models of
2003. The pharmacology of latent epileptogenesis. Neurology
GLP Pocketbook. Contains the inhibition as an animal model of 2002;59:S7–14.
Good Laboratory Practice (GLP) schizophrenia. Brain Research Willner P. The validity of animal models
Regulations 1999 and the Guide to Reviews 2000;33:275–307. of depression. Psychopharmacology
the GLP Regulations. London: MCA Pinkert CA. Transgenic animal 1984;83:1–16.
Publications; 1999. technology. 2nd ed. San Diego, Willner P, Mitchell PJ. The validity of
Hall JM. Bradykinin receptors: CA: Academic Press; 2002. animal models of predisposition
pharmacological properties and Polites HG. Transgenic model to depression. Behavioural
biological roles. Pharmacology and applications to drug discovery. Pharmacology 2002;13:169–88.
Therapeutics 1992;56:131–90. International Journal of Yamada K, Nabeshima T. Animal
Heidempergher F, Pillan A, Pinciroli V, Experimental Pathology models of Alzheimer’s disease and
et al. Phenylimidazolidin-2-one 1996;77:257–62. evaluation of anti-dementia drugs.
derivatives as selective 5-HT3 Research Defense Society website: Pharmacology and Therapeutics
receptor antagonists and refinement www.rds-online.org.uk/ethics/ 2000;88:93–113.
of the pharmacophore model for arclaims – a reasoned rebuttal of the Yang D, Soulier J-L, Sicsic S, et al. New
5-HT3 receptor binding. Journal of view put forward by opponents of esters of 4-amino-5-chloro-2-
Medicinal Chemistry 1997;40: animal experimentation that the methoxybenzoic acid as potent
3369–80. use of such experiments in drug agonists and antagonists for 5-HT4
Keen M, editor. Receptor binding discovery is at best unnecessary, receptors. Journal of Medicinal
techniques. Totowa, NJ: Humana if not positively misleading. Chemistry 1997;40:608–21.
Press; 1999. Rudin M, Weissleder R. Molecular Zan Y, Haag JD, Chen KS, et al.
Kenakin T. The measurement of efficacy imaging in drug discovery and Production of knockout rats using
in the drug discovery agonists development. Nature Reviews Drug ENU mutagenesis and a yeast-based
selection process. Journal of Discovery 2003;2:122–31. screening assay. Nature
Pharmacologic and Toxicologic Rudolph U, Moehler H. Biotechnology 2003;21:645–51.
Methods 1999;42:177–87. Genetically modified animals

170
 Chapter 12 
Biopharmaceuticals
H LeVine

arise commonly because biomolecules tend to be large

INTRODUCTION and unstable, and considerable ingenuity is often needed
to improve their pharmacokinetic properties, and to target
The term ‘biopharmaceutical’ was originally coined to their distribution in the body to where their actions are
define therapeutic proteins produced by genetic engineer- required.
ing, rather than by extraction from normal biological It is beyond the scope of this book to give more than a
sources. Its meaning has broadened with time, and the brief account of the very diverse and rapidly developing
term now encompasses nucleic acids as well as proteins, field of biopharmaceuticals. More detail can be found in
vaccines as well as therapeutic agents, and even cell-based textbooks (Buckel, 2001; Ho and Gibaldi, 2003; Walsh,
therapies. In this chapter we describe the nature of bio­ 2003). As the field of biopharmaceuticals moves on from
pharmaceuticals, and the similarities and differences in being mainly concerned with making key hormones,
discovery and development between biopharmaceuticals antibodies and other signalling molecules available as
and conventional small-molecule therapeutic agents. The therapeutic agents, efforts – many of them highly ingen-
usual starting point for biopharmaceuticals is a naturally ious – are being made to produce therapeutic effects in
occurring peptide, protein or nucleic acid. The ‘target’ is other ways. These include, for example, using antisense
thus identified at the outset, and the process of target nucleic acids, ribozymes or RNAi (see below and Chapter
identification and validation, which is a major and often 6) to reduce gene expression, the use of catalytic antibod-
difficult step in the discovery of conventional therapeutics ies to control chemical reactions in specific cells or tissues,
(see Chapter 6), is much less of an issue for biopharma- and the development of ‘DNA vaccines’. So far, very few of
ceuticals. Equally, the process of lead finding and optimi- these more complex ‘second-generation’ biopharmaceuti-
zation (Chapters 7, 8, 9) is generally unnecessary, or at cal ideas have moved beyond the experimental stage, but
least streamlined, because nature has already done the job. there is little doubt that the therapeutic strategies of the
Even if it is desirable to alter the properties of the naturally future will be based on more sophisticated ways of affect-
occurring biomolecule, the chemical options will be much ing biological control mechanisms than the simple ‘ligand
more limited than they are for purely synthetic com- → target → effect’ pharmacological principle on which
pounds. In general, then, biopharmaceuticals require less most conventional drugs are based.
investment in discovery technologies than do conven-
tional drugs. Toxicity associated with reactive metabolites
– a common cause of development failure with synthetic RECOMBINANT DNA  
compounds – is uncommon with biopharmaceuticals. On
TECHNOLOGY: THE ENGINE  
the other hand, they generally require greater investment
in two main areas, namely production methods and formula- DRIVING BIOTECHNOLOGY
tion. Production methods rely on harnessing biological
systems to do the work of synthesis, and the problems of The discovery of enzymes for manipulating and engineer-
yield, consistency and quality control are more complex ing DNA – the bacterial restriction endonucleases, poly-
than they are for organic synthesis. Formulation problems nucleotide ligase and DNA polymerase – and the invention

© 2012 Elsevier Ltd. 171

Section | 2 | Drug Discovery

of the enabling technologies of DNA sequencing and Extracts of tissues provided hormones, many of which
copying of DNA sequences by using the polymerase chain were polypeptides. After its discovery in 1921, insulin
reaction (PCR) allowed rapid determination of the amino extracted from animal pancreas replaced a starvation
acid sequence of a protein from its mRNA message. Versa- regimen for treating diabetes. The size of the diabetic
tile systems for introducing nucleic acids into target cells population, and the activity in humans of the hormone
or tissues and for the control of host nucleic acid metabo- isolated from pancreas of pigs and cows, permitted early
lism brought the potential for correction of genetic defects commercial success. Generally, however, the low yield of
and for new therapeutic products for disorders poorly many hormones and growth factors from human or
served by conventional small-molecule drugs. The impor- animal sources made industrial scale isolation difficult
tance of these discoveries for the biological sciences was and often uneconomic. Nevertheless, several such hor-
indicated by the many Nobel Prizes awarded for related mones were developed commercially, including follicle-
work. No less critical was the impact that these reagents stimulating hormone (FSH) extracted from human urine to
and technologies had on applied science, especially in treat infertility, glucagon extracted from pig pancreas to
the pharmaceutical industry. The business opportunities treat hypoglycaemia, and growth hormone, extracted from
afforded by biotechnology spawned thousands of startup human pituitary to treat growth disorders. Some enzymes,
companies and profoundly changed the relationship such as glucocerebrosidase, extracted from human placenta
between academia and industry. The mainstream pharma- and used to treat an inherited lipid storage disease
ceutical industry, with its 20th-century focus on small- (Gaucher’s disease), and urokinase, a thrombolytic agent
molecule therapeutic agents, did not immediately embrace extracted from human urine, were also developed as com-
the new methodologies except as research tools in the mercial products.
hands of some discovery scientists. Entrepreneurs in bio- Some serious problems emerged when proteins extracted
technology startup firms eventually brought technology from human or animal tissues were developed for thera-
platforms and products to large pharmaceutical compa- peutic use. In particular:
nies as services or as products for full-scale development. • Repeated dosage of non-human proteins generated
This alliance allowed each party to concentrate on the part an immune response in some patients against the
they did best. foreign sequences, which differed by several amino
Biotechnology products were naturally attractive to the acids from the human sequence. Such immune
small startup companies. Because the protein or nucleic responses could cause illnesses such as serum
acid itself was the product, it was unnecessary to have sickness, or loss of efficacy of the protein.
medicinal chemists synthesize large collections of organic • Human tissue was in short supply and was subject
small-molecule compounds to screen for activity. A to potential contamination with infectious agents.
small energetic company with the right molecule could Growth hormone extracted from human cadaver
come up with a profitable and very useful product. The pituitary glands was contaminated with prions that
niche markets available for many of the initial protein or cause Creutzfeld–Jakob disease, a dementing
nucleic acid products were sufficient to support a small brain-wasting disease similar to bovine spongiform
research-based company with a high-profit-margin thera- encephalitis (BSE) and sheep scrapie. Human blood
peutic agent. plasma-derived products have been tainted with
hepatitis B virus and HIV.
• Many agents (e.g. cytokines) cannot be extracted in
THE EARLY DAYS OF   sufficient quantities to be used therapeutically.
PROTEIN THERAPEUTICS • Batch-to-batch variability was considerable, requiring
standardization by bioassay in many cases.
The recombinant DNA revolution and the subsequent
Along with plant-derived natural products, proteins and
development of biotechnology resolved many of these
peptides were some of the first therapeutic agents pro-
issues. Many vaccines and antisera, however, are still pre-
duced by the fledgling pharmaceutical industry in the
pared from blood products or infectious organisms, rather
latter half of the 19th century, before synthetic chemistry
than by recombinant DNA methods.
became established as a means of making drugs. Long
before antibiotics were discovered, serum from immune
animals or humans was successfully used to treat a variety
of infectious diseases. Serotherapy was the accepted treat-
CURRENTLY AVAILABLE CLASSES OF
ment for Haemophilus influenzae meningitis, measles,
diphtheria, tetanus, hepatitis A and B, poliovirus, cyto­ BIOPHARMACEUTICALS
megalovirus and lobar pneumonia. Antisera raised in
animals were used to provide passive protection from The major classes of biopharmaceuticals currently on
diphtheria and tetanus infection. the market include hormones, cytokines, growth factors,

172
 Biopharmaceuticals Chapter | 12 |

antibodies, enzymes, vaccines and nucleotide-based or misregulation. Insulin isolated from biological sources
agents. Examples of therapeutic proteins, including anti- is administered to diabetics to control blood glucose
bodies, enzymes and other proteins approved for clinical levels. Immunological reaction to non-human (porcine or
use, are presented in Table 12.1. Others not included in bovine) insulin preparations, which occur in a significant
the compilation include therapeutic preparations such as number of patients, are avoided by recombinant products
serum albumin, haemoglobin and collagen, which are not incorporating parts of the human insulin sequence. A
drugs in the conventional sense. variety of human insulin and glucagon preparations are
In addition to the ‘mainstream’ biopharmaceuticals now on the market.
considered here are numerous speciality products for Human growth hormone (somatotropin) was originally
niche markets that are under investigation or in develop- developed for treating paediatric growth failure and Turn-
ment by small ‘boutique’ companies. er’s syndrome. Originally, growth hormone was extracted
from human pituitary tissue post mortem, but this mate-
rial carried a significant risk of transmitting Creutzfeld-
Growth factors and cytokines Jakob disease, a fatal neurodegenerative condition now
The production, differentiation and survival of the various known to be transmitted by a prion, an abnormal protein
types of blood cell are tightly regulated by an interacting found in affected nervous tissue, whose existence was
network of hormones, cytokines and growth factors. unsuspected when human-derived growth hormone was
Species-specific activities of many of these chemical medi- introduced as a therapeutic agent. The production of
ators, and their very low abundance, highlighted the need human growth hormone by recombinant methods rather
for biopharmaceutical products. than extraction avoids this serious problem as well as
The most common uses of haemopoietic factors are for providing a much more abundant source. Growth hormone
the treatment of various types of neutropenia, where spe- has acquired notoriety since the potential for misuse to
cific white cell levels are depressed as a result of infection, produce taller and stronger athletes was realized.
immune disorders, recovery from chemotherapy or reac- Human gonadotropin-releasing hormones have been
tion to various drug regimens. They are especially useful extensively used in fertility management as well as in treat-
in aiding recovery from the dose-limiting side effects of ing endometriosis and precocious puberty. Originally iso-
cancer chemotherapy. Granulocyte colony-stimulating factor lated from urine, there are now numerous recombinant
(G-CSF) and granulocyte-macrophage colony-stimulating products on the market.
factor (GM-CSF) are used for this purpose to improve
patient quality of life and allow the continuation of
Coagulation factors
chemotherapy.
Erythropoietin (EPO), normally secreted by the kidney to Coagulation factors are a group of plasma proteins
stimulate the production of red blood cells, is the most required for proper haemostasis. Deficiencies are associ-
successful biotechnology product so far marketed. EPO ated with genetic lesions and occur as complications of
boosts red cell counts and reduces transfusion require- viral infections such as hepatitis C and HIV. They were
ments for patients rendered anaemic by cancer chemo- initially treated with plasma-derived concentrates, which
therapy or renal disease. Various forms of EPO with carried a significant risk of viral and prion contamination.
clearance profiles – and hence duration of action – modi- Recombinant factor VIII (Recombinate, Kogenate) and factor
fied by linkage with polyethylene glycol (PEGylation) or IX (BeneFix) avoid this risk and are now widely used.
alteration of its glycosylation (see below and Chapter 17)
are also available. Off-label use of EPO by athletes to
Antithrombotic factors
improve performance has caused controversy.
Interferons are a complex group of proteins that augment To reduce blood coagulation, when conventional heparin
immune effector cell function. Interferon-α was the first or warfarin therapy is contraindicated, recombinant
recombinant biotherapeutic agent approved by the FDA thrombin inhibitors (Leprudin, Bivalirudin) and antiplatelet
for cancer treatment. Recombinant interferons have been gpIIb/IIIa antagonists (Eptifibatide) are now available. In
approved for melanoma, hepatitis C, Karposi’s sarcoma, conditions such as stroke, coronary thrombosis and pul-
T-cell lymphoma, chronic myelogenous leukaemia, monary embolism early treatment with thrombolytic
multiple sclerosis and severe malignant osteopetrosis. agents to relieve the vascular block is highly beneficial.
Mechanism-based side effects have thus far restricted their Streptokinase and urokinase are proteases that process plas-
utility to these severe disorders. minogen to plasmin, activating its thrombolytic activity to
dissolve fibrin clots. Newer products include tissue plas-
minogen activator (tPA) and TNKase, which catalyse the
Hormones
same reaction but in a fibrin-dependent fashion, so that
Hormone replacement or augmentation is a commonly plasmin production is concentrated in the region of the
accepted medical practice in certain diseases of deficiency clot. These proteins are also less immunogenic than the

173
Section | 2 | Drug Discovery

Table 12.1  Examples of approved therapeutic proteins produced by recombinant technology

Product Trade name Date approved Indications

Blood clotting factors and plasminogen activators
Human factor VIII Kogenate, Recombinate 1992 Haemophilia A
Human factor IX Benefix 1997 Haemophilia B
Human tissue plasminogen Activase 1987 Heart attacks, stroke
activator (tPA)
Haemopoietic factors
Erythropoietin Epogen, Procrit 1989 Anaemia
Granulocyte-macrophage Leukine 1991 Neutropenia
colony-stimulating factor
(GM-CSF)
Hormones
Human insulin Humulin, Novolin, Protropin 1982 Diabetes mellitus
Human glucagons GlucaGen 1998 Hypoglycaemia
Human growth hormone Humatrope, Nutropin 1987 Growth hormone deficiency
Human thyroid-stimulating Thyrogen 1998 Thyroid deficiency
hormone (TSH)
Human follicle-stimulating Gonal F, Follistim 1995 Infertility
hormone (FSH)
Interferons and interleukins
Human interferon-α Intron A, Viraferon 1986 Hepatitis B and C
Human interferon-β Betaferon 1995 Multiple sclerosis
Human interferon-γ Actimmune 1990 Chronic inflammatory
disease
Modified human interleukin-2 Proleukin 1992 Renal carcinoma
Modified human interleukin-11 Neumega 1997 Thrombocytopenia
Monoclonal antibodies
Abciximab (against platelet ReoPro 1994 Blood clot prevention
GPIIb/IIIa)
Trastuzumab (against human Herceptin 1998 Breast cancer
EGF receptor)
Infliximab (against TNF-α) Remicade 1998 Crohn’s disease, arthritis
Rituximab (against CD20 Rituxan 1997 Non-Hodgkin’s lymphoma
lymphocyte antigen)
Others
Hirudin Revasc, Refludan 1998 Prevention of thrombosis
Human β-cerebrosidase Cerezyme 1994 Gaucher’s disease (lipid
storage disorder)
DNase Pulmozyme 1993 Cystic fibrosis

174
 Biopharmaceuticals Chapter | 12 |

bacterial streptokinase, which is important in cases where biology, in which bacteriophages, gene libraries in plas-
repeated administration of the thrombolytic agent is mids and bacterial hosts are engineered to produce either
required. whole antibodies or derivatives of antibodies with desired
Another regulator of coagulation is the serine protease properties. ‘Humanizing’ the antibodies, or replacing the
protein C. The activated form of protein C breaks down the rodent constant domains with human sequences (chi­
clotting factors Va and VIIIa and plasminogen activator merization), limits hypersensitivity reactions to foreign
inhibitor-1, tipping the balance in the favour of throm- protein. Chimerization increases the half-life of the anti-
bolysis. These enzymes are important bio-pharmaceutical bodies in human plasma up to six-fold and improves their
products that have no small-molecule counterpart. function within the human immune network. The human
Fc domain reacts with Fc receptors on human cells more
avidly. Chimeric antibodies with human constant regions
Therapeutic antibodies also interact optimally with human complement proteins,
and are thus more effective in destroying target cells in
Monoclonal antibodies
patients than are their rodent counterparts.
A breakthrough in high-quality reproducible and scalable
production of antibodies came with the development by
Kohler and Milstein in 1975 of monoclonal antibodies. Antibody selection by phage display
Fusion of primed T cells from an immunized mouse with Phage display technology (Benhar, 2001) is a useful way to
an immortalized mouse myeloma (B-cell) line that secretes identify antigen combining regions to produce mono-
immunoglobulin light chains provided a cell-culture clonal antibodies that bind to therapeutically relevant
system that could produce unlimited quantities of anti- antigens. Bacteriophages (or phages) are viruses that rep-
body with defined specificity. Single-cell clones secreted licate in bacteria, Escherichia coli being the organism of
antibody against a single epitope of the antigen. The initial choice in most cases. For antibody selection (Figure 12.1)
immunization could, for toxic or scarce immunogens, be a large DNA library, encoding millions of different puta-
replaced by in vitro stimulation of isolated mouse thymo- tive antigen-binding domains, is incorporated into phage
cytes for fusion with the myeloma cells. DNA, so that each phage particle encodes a single antigen
The technology for antibody production has now gone combining region. The mixed phage population is added
mouseless. It has moved into the realm of molecular to E. coli cultures, where the phage replicate, each phage

Antigen-combining
region gene library
Expressed
antigen-combining
region

Incorporate
into phage
DNA
Replicate Panning of phage
in E.coli suspension
Immobilized
antigen

Phage Isolate bound

virus phage

Recombinant
antibody
Produce antibody Isolate DNA

Fig. 12.1  Antibody selection by phage display technique.

175
Section | 2 | Drug Discovery

particle expressing copies of a single antigen combining treated with monoclonal antibodies, have replaced the
region on its surface. The phage suspension is applied to (Fab)2 fragments generated by proteolysis from the intact
plates coated with the antigen of interest (‘panning’) and antibody. Single sFv chains containing the H and L vari-
those phage particles expressing antigen combining able regions are the smallest antibodies containing a high-
regions recognizing the antigen stick to the plates. The affinity antigen-combining site. Still in the experimental
adherent phages are isolated, allowing the antigen com- stage, ‘di-antibodies’ with two H-L units connected by a
bining region-encoding DNA to be identified and inserted 15-amino acid linker (Gly4Ser)3 peptide have greatly
into the DNA sequence encoding an appropriate full-size increased affinity for antigen (see also Chapter 13).
antibody scaffold to produce a specific antibody, or into a
reduced size, simplified monomeric framework to produce
a single-chain (sFv) antibody. Catalytic antibodies
Antibodies can be used to enhance the chemical reactivity
of molecules to which they bind. Such catalytic antibodies
Uses of antibodies as therapeutic agents
(‘abzymes’) created with transition state analogs as immu-
Cancer immunotherapy nogens specifically enhance substrate hydrolysis by factors
Although high-affinity mouse monoclonal antibodies to of 102–105 over the rate in their absence, and this principle
target antigens can be reproducibly produced in industrial has been applied to the development of therapeutic agents
quantities, and bind to specific human targets, they gener- (Tellier, 2002). Both esterase and amidase activities have
ally function poorly in recruiting human effector func- been reported. Catalytic turnover of substrates by abzymes
tions. The therapeutic potency of these antibodies can be is low in comparison to true enzymes, as high-affinity
enhanced by taking advantage of the targeting selectivity binding impedes the release of products. Attempts to
of antibodies (see Chapter 17) for the purposes of drug improve catalytic efficiency and to identify therapeutic
delivery. Linking other agents, such as cytotoxic drugs, uses for catalytic antibodies have engrossed both academic
biological toxins, radioisotopes or enzymes to activate and biotech startup laboratories. Targets being approached
prodrugs to targeting antibodies enhances their delivery with these antibodies include cocaine overdose and drug
to the target cells and reduces side effects by directing addiction, bacterial endotoxin, and anticancer mono-
the toxic agents to the tumour and minimizing clearance. clonal antibody conjugated with a catalytic antibody
Bispecific antibodies with one H-L chain pair directed designed to activate a cytotoxic prodrug. Attempts are also
against a target cell antigen and the other against a soluble being made to develop proteolytic antibodies containing
effector such as a complement component, or against a a catalytic triad analogous to that of serine proteases,
cell surface marker of an effector cell type, have also been designed to cleave gp120 (for treatment of HIV), IgE (for
developed to bring the components of the reaction treatment of allergy), or epidermal growth factor receptor
together. A number of these are in clinical trials for a (for treatment of cancer).
variety of different malignancies, but have in general
proved less successful than had been expected on the basis
of animal studies. Therapeutic enzymes
Enzymes can be useful therapeutic agents as replacements
Antibody use in transplantation for endogenous sources of activity that are deficient as a
and immunomodulation result of disease or genetic mutation. Because enzymes are
The major obstacle in the transplantation of cells, tissues large proteins they generally do not pass through cellular
or organs is the body’s recognition of foreign material. A membranes, and do not penetrate into tissues from the
strong immune response to rid the body of the non-self bloodstream unless assisted by some delivery system. Lys-
antigens leads to rejection of the transplant. Selective osomal hydrolases such as cerebrosidase and glucosidase
immunological suppression can ablate components of the have targeting signals that allow them to be taken up by
antitransplant response. Fab fragments of antibodies to a cells and delivered to the lysosome. Genetic lysosomal
number of surface antigens on T-cells are used to partially storage diseases (Gaucher’s, Tay–Sachs) are treated by
block T-cell responses to donor cells. Fab fragments are enzyme replacement therapy. However, penetration of the
required because the Fc portion of the complete antibody enzymes into the nervous system, where the most severe
targets the cell for destruction, resulting in unwanted com- effects of the disease are expressed, is poor. Cystic fibrosis,
plete immunosuppression. a genetic disorder characterized by deficits in salt secretion,
Even human antibodies face disadvantages as therapeu- is treated with digestive enzyme supplements and an
tics. Because of their size (MW ~150 kDa), their ability to inhaled DNase preparation (dornase-α) to reduce the
penetrate into tissues, such as solid tumours, is limited. extracellular viscosity of the mucous layer in the lung to
Engineered versions lacking the Fc region, which is largely ease breathing. All of these are produced as recombinant
responsible for hypersensitivity responses of patients proteins.

176
 Biopharmaceuticals Chapter | 12 |

time of writing, recombinant vaccines against hepatitis A

Vaccines
and B (e.g. Twinrix), papillomavirus virus for genital warts
Using the immune system to protect the body against and cervical cancer (Gardisil), and against Haemophilus
certain organisms or conditions is a powerful way to b/meningococcal protein for meningitis (Comvax) are
provide long-term protection against disease. Unlike with approved. Single or multiple proteins from an organism
small-molecule pharmaceuticals, which are administered can be included, and proteins that interfere with the
when needed, once immunity is present subsequent expo- immune response (common in many infectious agents)
sure to the stimulus automatically activates the response. excluded.
Most current vaccines are against disease-causing organ-
isms such as bacteria, viruses and parasites. More complex
conditions where the antigens are not so well defined are
GENE THERAPY
also being addressed by immunization. Vaccines are being
tested for cancer, neurodegenerative diseases, contracep-
tion, heart disease, autoimmune diseases, and alcohol and The development and present status of gene therapy are
drug addiction (Rousseau et al., 2001; Biaggi et al., 2002; discussed in Chapter 3. Here we focus on the technical
BSI Vaccine Immunology Group, 2002; Kantak, 2003). problems of gene delivery and achieving expression levels
Immune induction is complex. Pioneering experiments sufficient to produce clinical benefit, which are still major
with attenuation of disease organisms showed that illness obstacles.
was not required for immunity. The goal in immunization
is to retain enough of the disease-causing trait of an Introduction of genetic material
antigen to confer protection without causing the disease.
For infectious agents, various methods of killing or
into cells
weakening the organism by drying or exposure to inacti- Nucleic acid polymers are large, highly charged poly-
vating agents still dominate manufacturing processes. Iso- anions at physiologic pH. The mechanism by which such
lation of antigens from the organisms, or modifying their cumbersome hydrophilic molecules traverse multiple
toxins, can be used in some cases. Vaccine production of membrane lipid bilayers in cells is poorly understood,
isolated antigen ‘subunit’ vaccines benefits from protein although a variety of empirical techniques for getting
engineering. DNA into cells have been devised by molecular biologists.
Novel approaches to presenting antigens are expected to Eukaryotic cells use many strategies to distinguish between
have an impact on vaccination. An example is the phage self and foreign DNA to which they are continually
display technique (Benhar, 2001), described earlier as a exposed. Some mechanisms apparently interfere with the
technique for antibody selection. The same approach can incorporation and expression of engineered genes in gene
be used to provide the protein antigen in a display frame- therapy and other DNA transfer applications. Despite
work that enhances its immunogenicity and reduces the advances in technology, expression of foreign genes in
requirement for immune adjuvants (of which there are cells or animals in the right place, at the right moment, in
few approved for human use). Viruses encoding multiple the right amount, for a long enough time remains difficult.
antigens (‘vaccinomes’) can also be employed. Doing so in a clinical situation where so many more of
Genetic vaccination (Liu, 2003) employs a DNA plasmid the variables are uncontrollable is yet more challenging.
containing the antigen-encoding gene, which is delivered The high expectations of gene therapies, which conceptu-
to the host tissue by direct injection of DNA, or by ally seemed so straightforward when the first trials were
more exotic techniques such as attaching the DNA to started, have run aground on the shoals of cellular genetic
microparticles which are shot into tissues at high speed regulatory mechanisms. Effective gene therapy awaits the
by a ‘gene gun’, or introduced by other transfection selective circumventing of these biological controls to
methods (Capecchi et al., 2004; Locher et al., 2004; Manoj deliver precise genetic corrections.
et al., 2004). When the DNA is transcribed, the mRNA
translated and the protein expressed in the host tissue,
eukaryotic sequences will undergo appropriate post-
Delivery of nucleic acids
translational modification, which does not occur with A potential gene therapy requires three major compo-
conventional protein vaccines. In general, partial sequences nents: the payload to accomplish the mission, a targeting
of pathogen-derived proteins are fully immunogenic, system to direct the vehicle and its payload to the correct
despite lacking the toxicity of full-length transcripts. The body compartment in the correct cell population, and
first human trial for a genetic vaccine against HIV took finally gene regulatory element(s), such as promoters/
place in 1995, and others quickly followed, including enhancers, to control the time and place of expression of
hepatitis, influenza, melanoma, malaria, cytomegalovirus, the payload sequence(s). Table 12.2 compares the proper-
non-Hodgkin’s lymphoma, and breast, prostate and colo­ ties of some gene therapy vectors that have been used in
rectal tumours. Initial results have been promising. At the clinical trials.

177
Section | 2 | Drug Discovery

Table 12.2  Gene therapy vectors

Naked DNA Liposome- Adenovirus Adenoassociated Retrovirus

encapsulated virus (AAV)
DNA
Genetic material DNA or RNA DNA (≥50 kb) DNA (7.5 kb) DNA (5 kb) RNA (8 kb)
(max size insert) (≥50 kb)
Efficiency Low Low Very high Medium High (low in vivo)
Transform Possibly, weak Possibly, weak Yes, transient Yes, site-specific No, transient
non-dividing cells integration integration expression, integration integration
non-integrating
Safety issues None, good None, good Viral Viral recombination, Viral
safety profile safety profile recombination, immune reactions recombination,
immune reactions tumorogenesis

To be clinically useful, a vehicle must elude host immu- integrase protein can be altered to target specific
nological or toxicological responses, must persist in the DNA sequences
body, and must avoid sequestration in non-target organs • MicroRNA (miRNA) regulation.
such as the liver or kidney. Ideally, for clinical use one Transcription of DNA into mRNA can also be controlled
would want the ability to modulate or remove the genetic either through transcription factors or through antisense
modification in case of adverse effects – the equivalent of oligonucleotide triplex formation. Direct interference with
discontinuing administration of a traditional drug. In genomic DNA through quadruplex formation is another
practice, major difficulties are experienced in obtaining strategy. These oligonucleotide sequences can be provided
sufficient expression of the desired product for long by biologically derived vector systems or by chemical syn-
enough to measure a positive clinical outcome. thesis. Short double- and single-strand oligonucleotide
sequences are readily taken up by cells. Produced as single
Strategies for nucleic acid-mediated strands by automated solid-phase chemical synthesis,
intervention these short, single-stranded 8–20-nucleotide sequences
can be chemically modified on the base, sugar, or phos-
There are several points of attack for nucleic acid therapeu- phate linker to enhance their stability against nuclease
tics. Most diseases are not due to an identified gene muta- degradation and cellular penetration (Figure 12.2).
tion, but rather reflect secondary cellular malfunction, Table 12.3 lists a number of clinical trials with nucleic
often in response to events entirely external to the cell acid-based therapeutics that are at different stages of
being targeted. Overexpressing proteins or fragments, completion.
either normal or mutated so as to compete with the
normal protein for participation in cellular functions –
known as the ‘dominant negative’ strategy – is a commonly RNA interference – gene silencing by RNAi
used approach. RNAi is a technique for reducing the expression of or
RNA metabolism can be modulated in many different silencing specific genes. Its usefulness as an experimental
ways, including: tool, particularly for identifying potential drug targets, is
• Antisense RNA discussed in Chapter 6, and it also has considerable poten-
• Small interfering mRNA (siRNA) tial as a means of silencing genes for therapeutic purposes.
• RNA decoys for viral RNA-binding proteins It is similar to antisense modulation of mRNA, but extends
• Specific mRNA-stabilizing proteins further into the command and control of cellular nucleic
• Interference with mRNA splicing to induce exon acids. Short double-stranded RNA (21–23 nucleotides) is
skipping or to correct abnormal splicing used to silence homologous gene activity. Specific RNA-
• Sequence-specific cleavage by catalytic RNAs such as induced silencing complexes (RISC) guide multiple activities
hammerhead and hairpin ribozymes. Bacterial that target mRNA for degradation, block translation, or
introns that code for a multifunctional reverse block gene transcription by methylation of chromatin
transcriptase/RNA splicing/DNA endonuclease/ (Denli and Hannon, 2003).

178
 Biopharmaceuticals Chapter | 12 |

5’ H2N H 2N
BASE
O
O O N N
H H N NH O
N O N O NH2
N
H H
O H (e) (f) (g) O
O P O– BASE Base modifications

O O
O O BASE
H H O

H H O
O H H 2C N

O P O– O P O,S
O P N(CH3)2
O
O O
O P S O
3’
O (i) (j)

O
DNA structure

(h) BASE
(a) F
N
O
(b) OCH3 H2C
NH
N
(c) O O BASE
CH3 O N
O
(d) O (k) (l)
OCH3

Sugar modifications Phosphodiester backbone modifications

Fig. 12.2  Chemical modification of nucleotides. Key for modifications: Sugar modifications: (a) fluoro-, (b) methoxy-,
(c) methoxyethyl-, (d) propoxy-. Base modifications: (e) 5-methyl cytosine, (f) 5-propyne cytosine, (g) tricyclic cytosine.
Phosphodiester backbone modifications: (h) phosphorothioate, (i) mopholino-, (j) methylene- (on PDF), (k) methylene-
methylimino, (l) peptide nucleic acid (PNA).

Examples of genes that have been targeted by RNAi still remain to be overcome. Unlike single-stranded anti-
include: sense oligonucleotides, the double-stranded siRNAs are
unable to penetrate cell membranes effectively without
• Viruses such as HIV-1, poliovirus, respiratory assistance. They are also highly susceptible to plasma deg-
syncytial virus and hepatitis C
radation. Unfortunately, the same chemical modifications
• Oncogenes and tumour suppressors, such as Ras, in the phosphodiester backbone used to stabilize anti-
bcl-abl, p53, p53bp and p73Dn
sense oligonucleotides reduces or eliminates silencing
• Cell surface receptors for HIV-1-CD4, CCR5 and activity of interfering RNA constructs. Modification of the
CXCR4, and the IL2 receptor α CD25 (Dykxhoorn
2’ position of uridines and cytosines with fluorine increases
et al., 2003).
plasma half-life and preserves their inhibitory capacity.
Although RNA interference is highly efficient as a gene- Although initial results suggested a high degree of spe-
silencing technique under laboratory conditions, most of cificity for siRNA in down-regulating a target molecule and
the difficulties with antisense or expression technology the technique is widely used in cell culture studies, there

179
Section | 2 | Drug Discovery

Table 12.3  Clinical trials with antisense oligonucleotides

Agent Target Indication Company

ISIS 2302 ICAM-1 Crohn’s disease, ulcerative ISIS Pharmaceuticals/Boehringer
colitis, rheumatoid arthritis, Ingelheim
psoriasis, renal transplant
ISIS 2105* HPV 6 and 11 E2 Genital warts ISIS Pharmaceuticals
gene product
ISIS 2922 (Formivirsen) CMV IE2 gene CMV retinitis ISIS Pharmaceuticals
GEM 132 CMV UL36 gene CMV retinitis Hybridon
CGP 64128A/ISIS 2531 Protein kinase C-α Miscellaneous cancers Novartis/ISIS Pharmaceuticals
CGP 69846A/ISIS 5132 c-raf kinase Miscellaneous cancers Novartis/ISIS Pharmaceuticals
GEM 91* HIV gag protein AIDS Hybridon
Genta 3139 Bcl-2 protein Non-Hodgkin’s lymphoma Genta
LR3280 c-myc Restenosis Lynx Therapeutics
OL(1)p53 P53 Haemopoetic malignancy University of Nebraska/Lynx Therapeutics
c-myb Chronic myelogenous University of Pennsylvania/Lynx
leukaemia Therapeutics
*Trials terminated.
Source: Bennett CF, Dean NM, Monia BP (1998) In: Harvey A L (ed) Advances in drug discovery techniques. Chichester: John Wiley; 174.

is a surprising tolerance of mismatching leading to effects Successful development faces both the known challenges
on unintended targets (Lin et al., 2005). Even if microar- of chemical matter and delivery as well as additional safety
ray analysis indicates that two siRNAs to the same protein and efficacy hurdles because of the web of interactions
give similar expression patterns, siRNAs with multiple among the genome and individual miRNAs (Seto, 2010).
mismatches in the centre of the sequence can operate as
microRNAs (miRNA) that can inhibit expression of off-
target proteins by arresting their translation without alter-
ing cellular mRNA levels, or affect mRNA stability (Saxena COMPARING THE DISCOVERY
et al., 2003). Despite these challenges, several companies PROCESSES FOR
are moving ahead with siRNA candidates in clinical trials
BIOPHARMACEUTICALS AND SMALL
(Table 12.3), and other products in this class are in the
pipeline for approval for testing. RNAi will probably dom- MOLECULE THERAPEUTICS
inate the next wave of gene therapeutics moving into clini-
cal trials. The approach to discovering new protein/peptide thera-
miRNAs are naturally occurring small RNA species that peutics differs in many ways from the drug discovery
have been shown to have profound regulatory effects on approach for synthetic compounds described in other
the genome. Multiple genes may be regulated by a single chapters in this section.
miRNA and each gene may also be connected to multiple Protein/peptide drug discovery usually starts with
miRNAs, creating a vast network of interactions that may known biomolecules, which it may be desirable to trim
provide fine control to multiple processes from tissue down to smaller domains with desired activities. Alterna-
development to a variety of disease states. They operate tively, native proteins are often optimized for desirable
analogously to siRNA, but with a twist. The RISC complex properties by mutation or other kinds of modification.
differs in that noncomplementary nucleotides create These modifications can also provide the basis for patent
bulges in the double-stranded nucleotide preventing the protection. Unlike small-molecule synthesis, where the
miRNA : RISC complex from continuing along the siRNA possibilities are virtually limitless, there is a finite number
pathway. Instead, mRNA translation is affected and the of natural proteins and a limited range of feasible modi-
mRNA is destabilized. Targeting microRNAs as a thera­ fications that may be introduced to improve their biologi-
peutic strategy is in the very early stages of evaluation. cal properties. The race to protect important proteins as

180
 Biopharmaceuticals Chapter | 12 |

starting points for therapeutics is critical for the future in eukaryotic expression systems, which express exogenous
application of biotechnology. proteins at lower levels than do bacterial systems, where
The production of biotechnology-based therapeutics the expressed protein can be 10–50% of the total cellular
also involves technologies and facilities quite different protein. The tags are designed to be removed from the
from those used for producing synthetic compounds. protein once it has been purified.

Manufacture of biopharmaceuticals
Expression systems Cut

N-Tag

C-Tag
Promoter rbs ATG MCS poly-A+
Full-size proteins, including a number of hormones, site
growth factors and mediators, are produced recombinantly
by inserting the cDNA version of the mRNA sequence
coding for the protein of interest into specially designed Gene
plasmid vectors that orchestrate the expression of the
desired protein product in the cell type of choice (Figure
12.3). The characteristics of different expression systems
are compared in Table 12.4.
These vectors contain a selection marker to maintain the
Sequ
vector in the host cell along with regulatory DNA sequences ences fo ance
r vector mainten
that direct the host cell transcriptional machinery to
produce mRNA from the inserted cDNA sequence, with
appropriate initiation and termination codons and a
ribosome-binding site. A polyA+ tail is added to eukaryotic Fig. 12.3  Generic vector for protein expression. The cDNA
expression systems to stabilize the mRNA. mRNA tran- for the protein to be expressed is cloned into restriction sites
scription, and hence protein production, is controlled by in the multiple cloning site (MCS) of an expression vector
an induction element, so that protein is produced only which contains appropriate sequences for maintenance in
when required. This is important, as some expressed pro- the chosen host cell. Expression of the gene is controlled by
teins are toxic to growing and dividing host cells. Other inducible promoter. A ribosome-binding site (rbs) optimized
for maximum expression directs translation, beginning at the
sequences, such as protein export signal peptides, hexahis-
ATG initiation codon. Either N-terminal or C-terminal tags
tidine tags, immunologic epitopes, or fusion with protein (N-, C-Tag) can be used to aid in purification or stabilization
partners such as glutathione S-transferase, thioredoxin, of the protein, and a protease cut site may be used to
maltose-binding protein or IgG Fc, facilitate secretion or remove the tag after purification. A translation terminator
provide handles for the detection/purification of the codon is followed (in eukaryotic cells) by a poly-A+ tail to
expressed protein. Purification tags are particularly useful stabilize the mRNA.

Table 12.4  Comparison of protein expression systems

Property Similarity to human situation

E. coli Yeast Aspergillus Insect Mammalian cell culture

Folding Some Some Some Yes Yes
Subunit assembly No Yes ? Yes Yes
Secretion Some Yes Yes Yes Yes
Modifications:*
  acetylation Yes Yes Probably Probably Yes
  myristoylation No Yes ? Yes Yes
  phosphorylation No Yes ? Yes Yes
  glycosylation No Incomplete Incomplete Incomplete Yes
*Co-transfection of appropriate enzymes into cells expressing the protein of interest can provide post-translational modifications.

181
Section | 2 | Drug Discovery

Bacterial protein expression, usually in Escherichia coli, Site-directed mutagenesis

is the system of choice for high expression levels of many
Changes in the sequence of a protein at the level of a single
proteins, and because scale-up technology is well devel-
amino acid can change the biochemical activity or the
oped. However, not every protein is amenable to bacterial
stability of the protein. It can also affect interactions with
expression. Proteins that contain disulfide bonds or are
other proteins, and the coupling of those interactions to
comprised of multiple subunits are problematic, as are
biological function. Stabilizing proteins to better with-
proteins containing critical carbohydrate, lipid or other
stand the rigours of industrial production and formulation
post-translational modifications. In some cases the
is commercially important. Microorganisms adapted to
enzymes responsible for post-translational modification,
extreme environments have evolved modified forms of
such as fatty acylation, can be co-transfected with the
enzymes and other functional proteins, such as increased
desired protein. Highly expressed proteins are frequently
disulfide (cystine) content and more extensive core inter-
produced as insoluble inclusion bodies in the bacterial
actions, to withstand high temperatures. Similar tricks are
cytoplasm or periplasmic space. Active proteins must be
used to stabilize enzymes for use at elevated temperatures,
recovered from these inclusion bodies after denaturation
such as in industrial reactors, or in the home washing
and in vitro refolding. Intrinsic membrane-bound pro-
machine with enzyme-containing detergents. Other muta-
teins can pose additional problems.
tions alter the response of the protein to environmental
Eukaryotic microbial systems include Saccharomycetes
conditions such as pH and ionic strength. Mutations to
cerevisiae and Picha pastoralis – yeasts that perform many
add or remove glycosylation signal sequences or to increase
mammalian post-translational modifications, including
resistance to proteolytic enzymes may be introduced to
partial glycosylation. Insect cells, notably from Spodoptera
alter immunogenicity or improve stability in body fluids.
frugiperda (Sf9), are infected with engineered baculovirus
Protein activation by conformational changes due to
expression vectors carrying the protein of interest. This
phosphorylation, binding of protein partners or other
is a particularly efficient system for membrane protein
signals may sometimes be mimicked by amino acid
expression. Many G-protein-coupled receptors, a large
changes in the protein. The solubility of a protein can
class of integral membrane proteins that includes many
often be improved, as can the ability to form well-ordered
targets for currently marketed drugs, are often produced
crystals suitable for X-ray structure determination.
in these cells for use in screening assays for receptor
ligands.
Cultured mammalian cells are used to express proteins
Fused or truncated proteins
on a research or production scale where faithful post- Sometimes the full-size form of a protein or peptide is too
translational modification is required. Favourites include difficult or expensive to produce or deliver to the target.
fibroblasts, anchorage-independent lymphocytes, Chinese Domains of proteins can often be stripped to their essen-
hamster ovary (CHO) cells, and human embryonic kidney tials, so that they retain the required characteristics of
(HEK) cells. Mammalian cells grow more slowly and are activity or targeting specificity. They may require fusion
more fastidious than the bacteria, yeasts or insect cells, with a partner for proper folding or stability, the partner
and are therefore more expensive for production. Human being removed as part of purification.
cells are used in situations where there is some mamma- Ligands and receptor-binding domains can be attached
lian species dependence of bound carbohydrate. This can to the active biomolecule in order to target it to specific
be particularly important for growth factors and the large sites. Pathogens and infectious agents, in addition to
glycoprotein hormones, as their glycosylation modulates appropriating intracellular targeting motifs, have evolved
the activity and stability of the protein product in vivo. protein sequences that enable them to penetrate the cell
Potential contaminating human pathogens in the human membrane to access the machinery for their propagation.
cell lines used for expression must be controlled. So must Some of these sequences can be incorporated into bio­
non-primate mammalian pathogens, as the possibility of pharmaceuticals in order to deliver peptides and proteins
transfer to humans cannot be ignored. into cells. Hydrophobic portions of Kaposi fibroblast
The use of transgenic animals and plants as ‘factories’ growth factor, Grb2 SH2 domain, and human integrin α
for genetically engineered proteins is discussed below. and β subunits have been employed in this way as cell-
penetrating agents. The fusion sequence (1–23) of HIV-1
gp41 has been used to deliver synthetic antisense oligo-
Engineering proteins nucleotides and plasmid DNA into cells. Amphipathic
Nature has devoted millions of years to optimizing protein sequences containing periodic arrangements of polar and
structures for their biological roles. The natural configura- hydrophobic residues, such as those found in the fusion
tion may not, however, be optimal for pharmaceutical peptide of the influenza haemagglutinin-2 protein,
applications. Recombinant DNA methods can be used to promote fusion with cell membranes and the delivery of
re-engineer the protein structure so as to better fulfil the associated agents. Polycationic sequences in penetratin
requirements for a pharmaceutical agent. (residues 43–58 from the Antp protein), HIV-1 tat protein

182
 Biopharmaceuticals Chapter | 12 |

(47–57), and the HIV-1 transcription factor VP22 (267– folding or intermolecular bond formation, thereby pro-
300), permeabilize membranes and increase the uptake of moting aggregation and precipitation of protein. Exam-
associated material. ples are interferon, and the acidic and basic forms of FGF.
Beta-elimination at cystine, cysteine, serine and threonine
leads to dehydroalanine formation and reaction with
Protein production nucleophilic side chains. Anhydride formation between
Although relatively small quantities of highly potent bio- the α amino group on a protein and the C terminus is a
logicals are normally required, production often presents major degradation pathway for insulin stored between
problems. Only synthetic peptides and oligonucleotides pH3 and pH5. Sometimes peroxides contaminate the
and their derivatives can be produced by a scalable chemi- excipients and stabilizing agents, such as the polyethylene
cal synthesis. The others require large-scale fermentation glycols or the surfactants Polysorbate 20 and Polysorbate
or cell culture. The proteins and nucleic acids are then 80 used in product formulation (see Chapter 16).
isolated and purified from the cells or media. Biotechno- The first therapeutic proteins produced on an industrial
logical process development is a complex field (Sofer and scale were antibodies, which initially relied on immuniza-
Zabriskie, 2000; Buckel, 2001; Gad, 2007). High-level tion of large animals such as horses. Monoclonal technol-
production of recombinant proteins can result in hetero- ogy paved the way for the production of murine antibodies
geneity because of the misincorporation of amino acids in cell culture. With the ability to produce human anti-
such as norvaline into the product, depending on the body proteins and derivatives through molecular biologi-
culture conditions, the nutrient composition of the cal manipulation, the organism used to manufacture the
medium and the producing organism. Post-translational protein is chosen on the basis of economics and require-
modifications depend on the cultured organism and the ments for post-translational modification. The character-
subcellular localization of the protein. In cultured eukary- istics of some commonly used organisms for protein
otic cells proteins undergo many kinds of post-translational production are summarized in Table 12.4.
modification, including glycosylation, sulfation, phospho- Eukaryotic and prokaryotic microorganisms and eukary-
rylation, lipidation, acetylation, γ-carboxylation and pro- otic cells in culture are the most common source of thera-
teolytic processing. These and other modifications increase peutic proteins. Prokaryotic expression is generally the first
the heterogeneity of the product, affecting its biological choice, whereas higher organisms produce protein prod-
half-life, immunogenicity and activity. The most homoge- ucts that are most similar to human material and thus are
neous preparation consistent with retention of the impor- more likely to be biologically active and pharmacokineti-
tant biological properties is standardized for production cally stable. The production of proteins on a commercial
and a reproducible analytical profile is established. Never- scale by eukaryotic cell culture is, however, extremely
theless, protein products are invariably less homogeneous expensive. The cells grow slowly (18–24 hours doubling
in composition than synthetic compounds, and the quality time, compared with about 20 minutes for bacteria) and
criteria for approval of clinical materials have to be most require a serum source or expensive growth factor/
adjusted to take account of this. hormone cocktails in the culture medium.
Formulation and storage conditions are critical because ‘Pharming’ of protein expression in whole animals or
the protein primary amino acid sequences naturally plants is used for large-scale production of proteins for
display a variety of chemical instabilities. A major degra­ some therapeutics and for industrial applications. Trans-
dation route of protein products is through unfolding of genic farm animals – goats, cows, sheep and rabbits – have
the polypeptide chain and aggregation. Proteins have a been engineered to secrete human proteins into their milk
finite half-life in an organism, governed partly by built-in (van Berkel et al., 2002). Annual milk yields range from
chemical and conformational instabilities. Adsorption to about 8 L (rabbit) to 8–10 000 L (cow). Expression levels
surfaces of glass or plastic containers, particulates and air– of from 1 to 20 g of product per litre of milk have been
liquid interfaces is often a cause of product loss. achieved in favourable instances, allowing small herds of
The chemical reactivity of the various amino acid resi- transgenic animals to do the work of large fermentation
dues contributes to protein loss and heterogeneity. Suc- facilities. Many biopharmaceuticals are under develop-
cinimide formation, isomerization and racemization, as ment by specialized biotechnology companies (Table
well as peptide bond cleavage, tend to occur at aspartate 12.5). The first drug produced in genetically engineered
residues, to an extent that depends somewhat on the sur- livestock (Atryn in goat milk) was approved by the FDA in
rounding amino acid sequence. Examples include hACTH, February 2009. Thirty or more such products are expected
soluble CD4, OKT-3 monoclonal antibody, hGH-releasing to be registered in the next decade.
factor, IL-1β and hEGF. Oxidation, particularly at methio- Plants can be directed to concentrate antibodies and
nine residues, is catalysed by transition metal ions, pH, other proteins in their leaves, seeds or fruit (Giddings
light, and various oxygen free radicals. Examples are et al., 2000; Daniell et al., 2001; Powledge, 2001; Streat-
relaxin A, hIGF-I and enkephalin analogs. Disulfide field and Howard, 2003), and are beginning to be used for
exchange (cysteine/cystine) can generate inappropriate the commercial production of antibodies and vaccines.

183
Section | 2 | Drug Discovery

plant genome has been successful for hepatitis B, Norwalk,

Table 12.5  Biopharmaceuticals produced in
transgenic farm animals. (These products are in
cholera and rabies virus vaccines.
development; none have yet reached the market) Feeding antigen-producing plants to animals has shown
promise for inducing immune protection (Judge et al.,
2004), despite the differences in the type of immune
Protein Indication Host
response from oral exposure and the standard humoral
Antithrombin III Inflammation Goat administration. The oral route would be ideal for human
α1-Antitrypsin Inflammation, Cow, goat, vaccines, particularly in countries where refrigeration is
inherited deficiency sheep uncertain. Although antigen expressed in potato for use
in humans was effective in animals, the cooking required
α-Glucosidase Glycogen storage Rabbit
for human consumption of the vegetable destroyed the
disease type II
immunogenicity of the antigen. Current efforts are devel-
h-Chorionic Infertility Cow, goat oping the technology for human foods that are eaten raw
gonadotropin and which are suitable for tropical climates, such as
tomato and banana (Korban et al., 2002).
Factor VIII Haemophilia Pig
There are many developments in plant-derived biophar-
Factor IX Haemophilia Pig maceuticals for human use, but none has yet become com-
Factor XI Haemophilia Sheep mercially available. Widespread environmental and other
concerns among the general public about transgenic crop
Fibrinogen Burns, surgery Pig, sheep production are a significant impediment, especially outide
Lactoferrin GI bacterial infection Cow of the United States.

Monoclonal Colon cancer Goat

antibody
PHARMACOKINETIC, TOXICOLOGICAL
Protein C Deficiency, adjunct Pig, sheep
to tPA AND DRUG-DELIVERY ISSUES WITH
PROTEINS AND PEPTIDES
Serum albumin Surgery, burns, shock Cow, goat
Tissue Infarction, stroke Goat The main source of the bias against biologicals (proteins)
plasminogen as therapeutic agents in major pharmaceutical companies
activator (tPA) comes from the difficulties in delivering these expensive,
large, highly charged molecules to the target in sufficient
concentrations for long enough to achieve pharmacologic
Monoclonal antibodies expressed in plants – ‘plantibod- effects. The complexities faced in taking small molecules
ies’ – have been produced in maize, tobacco and soybeans. through preclinical development into clinical trials are
Antibodies produced against Streptococcus mutans, the compounded with biologicals. More detailed expositions
main cause of tooth decay in humans, have demonstrated of pharmacokinetics, toxicology and the delivery of thera-
efficacy against dental caries. CaroRX is in Phase II clinical peutic proteins are available (Frokjaer and Hovgaard,
trials for topical application for tooth decay. Other pro- 2000; Sofer and Zabriskie, 2000; Ho and Gibaldi, 2003;
teins, such as mouse and human interferon, human Walsh, 2003; Steffansen et al., 2010).
growth hormone, haemoglobin, human serum albumin Figure 12.4 illustrates the pathways that govern the
and human epidermal growth factor, are also being distribution and elimination of a pharmacological agent.
targeted. Merispace, a recombinant mammalian gastric Factors that assume particular importance for biopharma-
lipase, is being produced in corn for oral administration ceuticals are (a) the choice of formulation and route of
to counteract lipid maladsorption in cystic fibrosis and administration so as to achieve consistent absorption,
chronic pancreatitis Vaccine production for both animals (b) control of distribution by targeting to the required site
and humans appears to be a most promising application of action, and (c) protection against rapid inactivation or
of ‘pharming’. By splicing the gene for the antigen into a elimination.
modified plant virus (cowpea mosaic virus), companies
such as Agricultural Genetics (Cambridge, UK) have pro-
duced effective vaccines for animals against mink entero- Absorption of protein/peptide
virus, HIV-1 and foot and mouth disease, obtaining up to therapeutics
200 doses of vaccine from a single cowpea (black-eyed
pea) leaf. Other companies are pursuing vaccines for Mucosal delivery
hepatitis A and B, cold and wart viruses and Plasmodium Peptide and protein drugs are in general poorly absorbed
(malaria). Splicing the antigen genes directly into the by mouth because of their high molecular weight, charge,

184
 Biopharmaceuticals Chapter | 12 |

through the cytosol, and then released by exocytosis at

Plasma the apical membrane. However, much of the protein is
DOSE Absorption degraded by the lysosomal system during transit.
compartment
Formulations have been devised to optimize the stabil-
ity and delivery of protein and peptide therapeutics
(Frokjaer and Hovgaard, 2000). Protection from proteoly-
Metabolism Distribution sis is a major concern for many routes of administration.
excretion Various entrapment and encapsulation technologies have
Blood–brain been applied to the problem (see Chapter 16). The greater
barrier the protection, though, the less rapidly the protein is
Elimination Peripheral released from the preparation. Hydrogels, hydrophilic
tissues Brain polymers formed around the protein, avoid proteolysis
but limit absorption. Microcapsule and microsphere for-
mulations are made by enclosing the active compound
within a cavity surrounded by a semipermeable mem-
Action brane or within a solid matrix. The surface area for absorp-
tion is much greater than that provided by hydrogel or
solid matrix formulations.
Fig. 12.4  Simplified scheme showing the processes involved A variety of penetration-enhancing protocols have been
in drug absorption, distribution and elimination. Pink boxes, applied to increase the rate and amount of protein absorp-
body compartments; grey boxes, transfer processes.
tion. These include surfactants (sodium dodecyl sulfate,
polyoxyethylene fatty acyl ethers), polymers (polyacrylic
acid, chitosan derivatives), certain synthetic peptides
(occulin-related peptides), bile salts (sodium deoxycho-
Table 12.6  Physical properties of some therapeutic late, sodium glycocholate, sodium taurocholate), fatty
proteins acids (oleic acid, caprylic acid, acyl carnitines, acylcho-
lines, diglycerides), and chelating agents (EDTA, citric
Protein Molecular Isoelectric acid, salicylates, N-amino acyl β-diketones).
wt (Da) point The nasal mucosal route of administration is emerging
Somatostatin 1600 10.4
as an acceptable and reasonably effective method of drug
delivery. Some peptide drugs, such as nafarelin, oxytocin,
Insulin 6000 5.6 lypressin, calcitonin and desmopressin, are effective as nasal
Human growth 20 000 5.3 sprays, although the fraction absorbed in humans is gener-
hormone ally low (≤10%). Pulmonary administration is surprisingly
effective when –3 µm particles of the therapeutic are dis-
t-PA 60 000 9.1 persed deep in the lung. The vast alveolar surface area and
Human serum albumin 66 000 5.3 blood supply allow even macromolecules to be absorbed.
Transdermal delivery across the stratum corneum, the
IgG 150 000 Variable outermost, least-permeable skin layer, is being increas-
Factor VIII 270 000 7.4 ingly used for proteins and other drugs. Transport of drug
molecules across the skin is facilitated by transient disrup-
Typical small-molecule <500 Variable
tion of epithelial barrier function with sonic energy (sono-
drug
phoresis), electric current (iontophoresis) or electric field
(electroporation). Table 12.7 compares different routes of
non-parenteral dosing for the relative proteolytic activity
to which a protein drug would be exposed.
and susceptibility to proteolytic enzymes. The main barri-
ers to absorption of proteins are their size and charge
(Table 12.6).
Parenteral delivery
Large hydrophilic and charged molecules do not readily Oral or mucosal absorption of native, full-length proteins
pass the cellular lipid bilayer membrane. Small peptides is inefficient even with enhancing strategies. Thus, many
can be absorbed if they are relatively hydrophobic, or if protein therapeutics are delivered parenterally by intra­
specific transporters exist. Proteins can be transported by venous, intramuscular or subcutaneous injection. With
transcytosis, in which proteins enter cells at the basal intramuscular and subcutaneous administration sustained-
surface via receptor-mediated endocytosis or fluid-phase release formulations can be used to increase the duration
pinocytosis into vesicles, are transported across the cell of action.

185
Section | 2 | Drug Discovery

glomerulus decreases to about 30% for proteins of 30 kDa,

Table 12.7  Exposure of protein/peptide drugs to
protease activity by different routes of
and to less than 1% for 69 kDa proteins. Some proteins,
administration such as calcitonin, glucagon, insulin, growth hormone,
oxytocin, vasopressin and lysozyme, are reabsorbed by the
Oral Very high proximal tubules via luminal endocytosis, and then are
Rectal High hydrolysed in lysosomes. Linear peptides less than 10
amino acids long are hydrolysed by proteases on the
Buccal Medium surface of brush border membranes of the proximal
Nasal Medium tubules.
Liver metabolism is highly dependent on specific
Vaginal Medium amino acid sequences in proteins. Cyclic peptides, small
Transdermal Low (<1.4 kDa) and hydrophobic peptides are taken up by
carrier-mediated transport and degraded. Large proteins
Pulmonary Low use energy-dependent carrier-mediated transport, includ-
Ocular Low ing receptor-mediated endocytosis and transcytotic path-
ways (polymeric IgA), and are targeted to lysosomes.

The blood–brain barrier is impermeable to most pro- SUMMARY

teins and peptides, but possesses transporters that facili-
tate the entry of some, such as apolipoprotein E, transferrin In the late 1970s recombinant DNA technology boosted
and insulin. Linking active peptides to an antibody biotechnology into a prominence it has not relinquished.
directed against the transferrin receptor has been shown Apart from fears about human genetic modification, the
in experimental animals to allow neurotrophic factors outlook has been positive for the production of new and
such as nerve growth factor (NGF) to enter the brain and better biopharmaceuticals. Whereas the traditional phar-
produce neurotrophic effects when given systemically and maceutical companies were slow to embrace the new tech-
this strategy may prove applicable for human use. nology, a number of small biotech companies sprang up
which have provided the innovative driving force for the
Elimination of protein/peptide protein and gene product industry. The genomic revolu-
tion is the latest manifestation of the technology, stimulat-
therapeutics
ing efforts to mine the information coming out of the
Most therapeutic proteins are inactivated by proteolysis, various species genome projects for new products and
producing short peptides and amino acids which then therapies.
enter dietary pathways. Unlike small-molecule drugs, Recombinant DNA technology allows the production
metabolites of protein therapeutics are not considered to of individual proteins and peptides on a large scale
be a safety issue. The rate and extent of degradation of regardless of their natural abundance. There are many
protein therapeutics before excretion depend on the route advantages to recombinant production. Human sequence
of administration. Breakdown occurs in both plasma and proteins reduce potential immunological problems and
tissues. Attempts to minimize these losses include the use avoid potential infectious agents in materials isolated
of glycosylated proteins, which more closely resemble the from natural sources. Proteins are expressed in a variety
native protein, or chemical modification of the protein of cellular systems, determined by the properties of the
with polyethylene glycol (PEG) chains or other entities. molecule required. Bacterial systems are used wherever
PEGylation can increase the plasma half-life of a protein possible because of the high level of expression that can
between three- and several hundredfold. Besides reducing be obtained and the simplicity and availability of produc-
proteolysis, PEGylation can shield antigenic sites, enhance tion. Bacteria, however, do not provide many of the post-
protein solubility and stability, and prevent rapid uptake translational modifications required for human protein
by organs. With small peptides, cyclization or the inclu- stability and function, such as glycosylation, lipid modi­
sion of D-amino acid residues, particularly at the N termi- fication, phosphorylation, sulfation and disulfide bond
nus, can be used to protect against exoproteolysis, though formation. Protein aggregation often occurs, and even
this approach is of course applicable only to synthetic though denaturation and refolding are successful for many
peptides and not to biopharmaceuticals. proteins, some are problematic. In these cases eukaryotic
The kidney is an important organ for the catabolism systems, including various yeast species and cultured
and elimination of small proteins and peptides. Proteins insect and mammalian cells, are used. Transgenic farm
of 5–6 kDa, such as insulin, are able freely to pass the animals and plants are also coming into their own as
glomerular filter of the kidney. Passage through the protein ‘factories’.

186
 Biopharmaceuticals Chapter | 12 |

Antibodies in the form of animal sera were the first charge and instability in the gastrointestinal tract, special
widely used protein therapeutics. Although immunization delivery systems often must be used to achieve efficacious
is still used as a prophylactic and as a therapeutic, recom- blood levels. Penetration through the blood–brain barrier
binant DNA technology is used to produce both immuno- can be even more challenging. Small-molecule drugs of
gen and antibodies. Hypersensitivity reactions are reduced <500 Da molecular weight have a much better record in
by producing altered antibodies with the required epitope this regard. Once in the blood, proteins are often rapidly
specificity. They are ‘chimerized’ (human constant domain) eliminated by a variety of mechanisms, although there are
or ‘humanized’ (rodent complementarity-determining modifications that can slow degradation.
region (CDR)), with human Fc portions for optimal inter- Even though protein and peptide therapeutics faces sig-
action with the human complement system. Single-chain nificant pharmacokinetic and pharmacodynamic liabili-
and multichain antibodies can be expressed intracellularly ties, much work is still being done on agents for which
to attack previously sequestered antigens. These immuno- small molecules are not available to perform the same
logical agents are also used to target infectious organisms, function.
deliver radioisotopes or cytotoxic agents to cancer cells, Nucleic acids can be delivered into cells in a variety of
and to moderate tissue rejection in transplantation. ways. Complexation of the highly charged DNA with
Proteins and enzymes have been engineered to hone lipophilic cations, or derivitization of nucleic acid bases,
their therapeutic usefulness. Protein structure can be sta- sugars or the phosphodiester backbone can be efficient in
bilized or modified to slow degradation or increase uptake, cell culture, but less so in whole organisms. Ingenious
and multiple functions can be built into the same mole- schemes of antisense, RNAi, ribozyme, decoy sequences
cule. Recombinant vaccine production of cloned protein and RNA splicing inhibition are used to attain therapeutic
domains uses these modifications to increase immuno- effects. Viruses, nature’s DNA delivery machines, modified
genicity and avoid exposure to infectious agents. Delivery to eliminate their pathogenic capacity, are used in organ-
of synthetic vectors coding for antigenic epitopes into host isms to introduce the therapeutic gene into target cells. The
tissues promotes endogenous antigen expression. inserted DNA directs the synthesis of a protein in the host
The biggest obstacle to the use of proteins and peptides cell, in the case of a growth factor deficiency in the brain
as therapeutics is the delivery to the site of action of suf- providing a secreted product for the support of surround-
ficient agent to create a biological effect. Formulating pro- ing cells. Numerous gene therapy clinical trials are in
teins and peptides to penetrate epithelial barriers in a progress, but the need to avoid side effects and host sup-
biologically active form is difficult. Because of their size, pression of viral-based therapeutics has slowed progress.

REFERENCES

Benhar I. Biotechnological applications biopharmaceuticals and edible Ho RJY, Gibaldi M. Biotechnology and
of phage and cell display. vaccines in plants. Trends in Plant biopharmaceuticals. Transforming
Biotechnology Advances Science 2001;6:219–26. proteins and genes into drugs.
2001;19:1–33. Denli AM, Hannon GJ. RNAi: an ever Hoboken, NJ: Wiley-Liss;
Biaggi E, Rousseau RF, Yvon E, et al. growing puzzle. Trends in 2003.
Cancer vaccines: dream, reality, Biochemical Sciences Judge NA, Mason HS, O’Brien AD. Plant
or nightmare? Clinical and 2003;28:196–201. cell-based intimin vaccine given
Experimental Medicine Dykxhoorn DM, Novina CD, Sharp PA. orally to mice primed with intimin
2002;2:109–18. Killing the messenger: short RNAs reduces time of Escherichia coli
BSI Vaccine Immunology Group. that silence gene expression. Nature O157:H7 shedding in feces.
Vaccination against non infectious Reviews Molecular Cell Biology Infection and Immunity
disease (BSI Vaccine Immunology 2003;4:457–67. 2004;72:168–75.
Group/BSACI session). Immunology Frokjaer S, Hovgaard L. Pharmaceutical Kantak KM. Anti-cocaine vaccines:
2002;107(Suppl 1):67–70. formulation: development of antibody protection against relapse.
Buckel P, editor. Recombinant protein peptides and proteins. Philadelphia, Expert Opinion in Pharmacotherapy
drugs. Basel: Birkhauser Verlag; PA: Taylor & Francis; 2000. 2003;4:213–18.
2001. Gad SC, editor. Handbook of Korban SS, Krasnyanski SF, Buetow DE.
Capecchi B, Serruto D, Adu-Bobie J, Pharmaceutical Biotechnology. Foods as production and delivery
et al. The genome revolution in Hoboken, NJ: John Wiley and Sons; vehicles for human vaccines. Journal
vaccine research. Current Issues in 2007. of the American College of Nutrition
Molecular Biology 2004;6:17–27. Giddings G, Allison G, Brooks D, et al. 2002;21:212S–17S.
Daniell H, Streatfield SJ, Wycoff K. Transgenic plants as factories for Lin X, Ruan X, Anderson MG, et al.
Medical molecular farming: biopharmaceuticals. Nature SiRNA-mediated off-target gene
production of antibodies, Biotechnology 2000;18:1151–5. silencing triggered by a 7 nt

187
Section | 2 | Drug Discovery

complementation. Nucleic Acids Res Rousseau RF, Hirschmann-Jax C, aspects of drug characterization, drug
2005;33:4527–35. Takahashi S, et al. Cancer vaccines. delivery, and dosage form evaluation.
Liu MA. DNA vaccines: a review. Hematology/Oncology Clinics of Chicago: Pharmaceutical Press; 2010.
Journal of Internal Medicine North America 2001;15:741–73. Streatfield SJ, Howard JA. Plant
2003;253:402–10. Saxena S, Jonsson ZO, Dutta A. Small production systems for vaccines.
Locher CP, Soong NW, Whalen RG, RNAs with imperfect match to Expert Review on Vaccines
et al. Development of novel vaccines endogenous mRNA repress 2003;2:763–75.
using DNA shuffling and screening translation. Implications for Tellier C. Exploiting antibodies as
strategies. Current Opinion in off-target activity of small inhibitory catalysts: potential therapeutic
Molecular Therapy 2004;6:34–9. RNA in mammalian cells. J Biol applications. Transfusion Clinique
Manoj S, Babiuk LA, van Drunen Chem 2003;278:44312–19. et Biologique 2002;9:1–8.
Littel-van den Hurk S. Approaches Seto AG. The road toward microRNA van Berkel PH, Welling MM, Geerts M,
to enhance the efficacy of DNA therapeutics. Int J Biochem Cell Biol et al. Large scale production of
vaccines. Critical Reviews in 2010;42:1298–305. recombinant human lactoferrin in
Clinical Laboratory Science Sofer G, Zabriskie DW. the milk of transgenic cows. Nature
2004;41:1–39. Biopharmaceutical process Biotechnology 2002;20:484–7.
Powledge TM. Tobacco pharming. validation, vol. 25. New York: Walsh G. Biopharmaceuticals,
A quest to turn the killer crop into Marcel Dekker; 2000. biochemistry, and biotechnology.
a treatment for cancer. Scientific Steffansen B, Brodin B, Nielson CU. 2nd ed. Chichester: John Wiley;
American 2001;285:25–6. Molecular biopharmaceuticals: 2003.

188
 Chapter 13 
Scaffolds: small globular proteins as
antibody substitutes
D Grabulovski, J Bertschinger

antibody is a Y-shaped complex of two identical light

INTRODUCTION chains, each with a variable and constant domain, and two
heavy chains, each with one variable and three constant
An emerging field in pharmaceutical biotechnology is rep- domains (Fig. 13.1H). Therefore, full-size IgGs are always
resented by the generation of novel binding molecules bivalent molecules and elicit effector functions through
based on small globular domains serving as protein frame- their Fc part, which may not be ideal for certain applica-
works (’scaffolds’). In this approach, certain amino acid tions. With regards to expression, there is the requirement
residues of the surface of a single domain are combinatori- for a rather expensive mammalian cell production system
ally mutated to produce a protein library, which can then and, in addition, antibodies depend on disulfide bonds
be screened for binding specificities of interest. Thus, the for stability. Through the versatile techniques of genetic
concept of a universal binding site from the antibody engineering, smaller antibody fragments have been pro-
structure is transferred to alternative protein frameworks duced to overcome some of the limitations of full-size
with suitable biophysical properties. On one hand these IgGs (Holliger and Hudson, 2005), but some antibody
new proteins provide further insights into the processes of fragments tend to aggregate and display limited solubility.
molecular recognition, on the other hand they also have Finally, the success, and consequently the extensive use, of
commercial applications, as therapeutic agents, diagnostic antibodies has led to a complicated patent situation for
reagents or affinity ligands. antibody technologies and applications. Therefore, several
research groups and small and medium-sized companies
have recently focused on the development of small globu-
lar proteins as scaffolds for the generation of a novel class
SCAFFOLDS of versatile binding proteins.
As therapeutics, globular proteins with engineered
Monoclonal antibodies (mAbs) are an increasingly impor- binding properties might be particularly interesting, (i) if
tant class of therapeutic agents as discussed in preceding the neutralization of a target protein is the desired phar-
chapters. Over 25 recombinant antibodies are currently on macological effect (in contrast to a full length antibody,
the market, the top 10 best-selling of these products earned where the Fc portion may stimulate immune processes),
over $32 billion in global sales during 2008, and five mAb (ii) as fusion proteins, for the targeted delivery of bioactive
products (rituximab, infliximab, bevacizumab, trastuzu- molecules to sites of disease, (iii) as receptor-binding
mab and adalimumab) were included in the top 15 best- drugs, thus interfering with the cell-signalling and (iv) as
selling prescription drugs for that year (Reichert, 2010). enzyme inhibitors.
The future prospects for therapeutic antibodies are clearly Common to all approaches of finding a suitable scaffold
bright, the annual growth rate estimated to be 12% during are the following steps (Nygren and Uhlen, 1997; Smith
2010–2012. However, even antibody blockbusters suffer 1998; Binz and Pluckthun, 2005):
from certain drawbacks: human immunoglobulin G • Choosing a small protein (domain) that is well
(IgG) antibody is a complex molecule composed of four expressed in bacteria or yeast and has good
protein chains with attached carbohydrates. A full-size IgG biophysical properties (stability, solubility)

© 2012 Elsevier Ltd. 189

 FG
P1 region CDR-H1
Second loop M
I A BG
K K 15 CDR-H3
R I G
47 CY
DE
S E F F P G P
L
C CDR-H2
E F E
R G E D G
E C F
C N N D
K O P A
N O
K M R T P 26 K
C I 8
T F
C A
R D F
M H
A S B 10Fn3
VHH

C C

Ca Library
C Exon
Exon C
C
C

Scaffold
Linker

C D

+ +
n

E N-terminus F

1
4
3 VL CL CH VH Fab
F R
C
A
E
N
H D CH2 scFv
G
C
Fc

CH3 VH
G H
 Scaffolds: small globular proteins as antibody substitutes Chapter | 13 |

• Creating a library of protein mutants by introducing treatment of hereditary angioedema (see kunitz type
diversity at a contiguous patch of the surface of the domains), others are being investigated in clinical trials.
protein (e.g. loops) Therefore, we will learn more in the near future about this
• Using a genotype-phenotype display system (e.g. interesting field of antibody alternatives.
phage display) to select a range of binders to a
therapeutic target of interest
• Screening the binders obtained by the selection Kunitz type domains
procedure for the desired biological activity, e.g. by Proteases have important biological functions, they have
enzyme-linked immunosorbent assay (ELISA). been identified as key regulators of cellular processes such
Mutations introduced in the protein scaffold to produce as ovulation, fertilization, wound healing, angiogenesis,
diversity may compromise the three-dimensional struc- apoptosis, peptide hormone release, coagulation and
ture, stability and solubility of the protein scaffold, thus complement activation (Nixon and Wood, 2006). Because
making the isolation of protein binders based on folding unregulated proteolysis in any of these processes will have
frameworks other than the immunoglobulin fold difficult. undesirable effects, effective control of protease activity is
Nevertheless, more than 50 scaffolds have been described critical. Indeed, unregulated activity has been implicated
for the generation of new protein binders (reviewed in in the pathogenesis of many diseases (e.g. cancer, inflam-
Binz et al., 2005; Gebauer and Skerra, 2009). Being mainly matory diseases, chronic obstructive pulmonary disease
a topic of academic interest at first, nowadays the develop- (COPD), chronic pancreatitis, muscular dystrophy and
ment and use of non-antibody classes of proteins is being Alzheimer’s disease (Naruse et al., 1999; Egeblad and
pursued by small- and medium-sized biotechnology com- Werb, 2002; Nunan and Small, 2002; Shapiro, 2002; Laval
panies (Hey et al., 2005). The common basis for the suc- and Bushby, 2004; Parks et al., 2004), making proteases
cessful development of well-designed protein drugs lies in important target proteins. It is notable, that there are
the availability of both a well-behaving protein scaffold only a few marketed drugs (such as HIV protease inhibi-
and an efficient selection or screening technology in order tors, angiotensin-converting enzyme inhibitors, the pro-
to find suitable candidate proteins in the large repertoire teasome inhibitor bortezomib and the renin inhibitor
created by mutagenesis. The next sections will introduce aliskiren), since a key challenge in the discovery of new
selected classes of protein scaffolds (see also Figure 13.1) inhibitors is the ability to identify drugs which are both
that have been successfully used and commercialized by potent and specific. Structural similarities within the active
small- and medium-sized companies. The first scaffold site of most proteolytic enzyme families often result in a
derived drug has been recently approved by the FDA for simultaneous inhibition of several family members, which

Fig. 13.1  Schematic representation of representative scaffolds. A, Kunitz type domain (LAC–D1) (Ley et al., 1996). Varied
positions are depicted in black, the P1 and second loop positions are enclosed. B, Structural comparison of a llama VHH domain
and the wild-type human 10Fn3 domain (Xu et al., 2002). Despite the lack of significant sequence identity, both domains fold
into similar beta sheet sandwiches. The CDRs of the VHH domain and the residues randomized in the 10Fn3 domain are shown
in color. C, Fixed and variable positions of the A-domain library, as well as the disulfide topology, are indicated (Silverman
et al., 2005). Each circle represents an amino acid position. Calcium-coordinating scaffold residues are yellow, structural
scaffold residues are blue, cysteine residues are red and variable positions are green. A ribbon diagram of a prototypical
A-domain structure is included (Protein Data Bank entry 1AJJ). D, Fyn SH3 wt protein structure (Protein Data Bank entry 1M27)
The RT-Src loop is in red, and the n-Src loop is in green. E, Electron density for all 13 mutated residues in the affibody
(Hogbom et al., 2003). For clarity, only the electron density around the side chains is displayed. F, Schematic representation of
the library generation of designed ankyrin repeat proteins (Binz et al., 2004). (upper part) Combinatorial libraries of ankyrin
repeat proteins were designed by assembling an N-terminal capping ankyrin (green), varying numbers of the designed ankyrin
repeat module (blue) and a C-terminal capping ankyrin (cyan); side chains of the randomized residues are shown in red. (lower
part) Ribbon representation of the selected MBP binding ankyrin repeat protein (colours as above) This binder was isolated
from a library of N-terminal capping ankyrin repeat, three designed ankyrin repeat modules and a C-terminal capping ankyrin
repeat. G, General structure of human retinol-binding protein, a prototypic lipocalin (Schlehuber and Skerra, 2005). Ribbon
diagram of the crystal structure of RBP with the bound ligand retinol (magenta, ball and stick representation). The eight
antiparallel strands of the conserved β-barrel structure are shown in blue with labels A to H, and the four loops, which are
highly variable among the lipocalin family, are colored red and numbered. The typical α-helix that is attached to the central
β-barrel in all lipocalins, the loops at the closed end the N- and C-terminal peptide segments are shown in grey. The three
disulfide bonds of RBP are depicted in yellow. H, Schematic representation of full-sized antibodies, multidomain and
single-domain antigen-binding fragments (Saerens et al., 2008). Classical IgG consists of two light chains and two
heavy chains. The light chain is composed of one variable (VL) and one constant (CL) domain, whereas the heavy chain has
one variable (VH) and three constant domains (CH1, CH2, and CH3). The smaller antigen-binding antibody fragments are
shown, that is, Fab, scFv, and the single-domain antibody (sdAb) fragment, VH.

191
Section | 2 | Drug Discovery

can lead to an unacceptable toxicity profile. One solution improving it by protein engineering works well for certain
to this problem represents the use of endogenous, engi- therapeutic applications. At the same time, kunitz type
neered protease inhibitors that make more contacts with domains have the limitation of binding only to proteases,
the target protease and therefore allow for tighter control and thus, at present, they do not represent a universal
over specificity. One type of endogenous inhibitors include source for the generation of binding proteins to a variety
the kunitz domain inhibitors, e.g. bovine pancreatic of targets.
trypsin inhibitor (BPTI) (Ascenzi et al., 2003).
In 1992, Ladner and co-workers produced a small Tenth type III domain of  
library of mutants of kunitz domain BPTI, displayed on
fibronectin (Adnectins)
phage and screened for the ability to bind to elastase
(Roberts et al., 1992), a serine protease that is believed to The tenth type III domain of human fibronectin (10Fn3)
play a causative role in a variety of lung diseases, including is a 94 aminoacid residue structural protein with an
COPD and cystic fibrosis (Hautamaki et al., 1997). Iso- immunoglobulin-like fold that is used as an antibody-
lates that were able to bind to elastase were expressed, mimic scaffold. High thermal stability (TM = 90°C) and
purified and tested for inhibition of elastase. The P-loop solubility (>15 mg/mL), high expression yields in
(see Figure 13.1A), the region that was varied in the library, Escherichia coli and the lack of cysteines in its structure are
of the best inhibitor (Ki ∼1 pM) was transferred to a attractive properties that make the 10Fn3 a good scaffold
human kunitz domain without subsequent loss of potency candidate (Xu et al., 2002). Three of the solvent exposed
(Ley et al., 1997). The resulting engineered molecule, loops in 10Fn3, named BC, DE and FG are structurally
depelestat (Dyax/Debiopharm SA), inhibited human neu- analogous to the VH complementary-determining regions
trophile elastase (HNE) activity in the sputum of cystic (CDR) (Figure 13.1B).
fibrosis patients, was demonstrated to be safe in an The first report on the use of the 10Fn3 domain as a
aerolized format in monkeys and no adverse effects were protein with artificial binding sites was published in 1998
reported in Phase I clinical trials (Delacourt et al., 2002; (Koide and Koide, 1998). In that study, a library of 108
Grimbert et al., 2003; Nixon and Wood, 2006). Pharma- distinct 10Fn3 mutants was made by deleting three residues
codynamic reports of a Phase IIa clinical trial have been in the FG loop and by randomizing five residues in the
shown to inhibit completely sputum HNE in 52.6% of BC loop and four residues in the FG loop. The library,
the patients and to decrease interleukin-8 levels, a bio­ displayed on phage, was used in affinity selection experi-
marker of inflammation, in the sputum of treated patients ments with immobilised ubiquitin as model target protein.
(Saudubray et al., 2003). After five rounds of panning, clones were randomly picked
In 1996, another human kunitz domain, the lipoprotein- for sequencing. A clone designated as Ubi4-Fn3 domi-
associated coagulation inhibitor (LACI-D1), was used as a nated the population of selected protein mutants and was
scaffold for the successful isolation of a very specific and subjected to further analysis. Ubi4-Fn3 was demonstrated
potent inhibitor of human kallikrein (Markland et al., to bind in phage enzyme-linked immunosorbent assay
1996). Kallikrein is believed to be an important mediator (ELISA) experiments. However, when expressed as a single
of hereditary angioedema (HAE), a rare disorder with domain protein in E. coli, Ubi4-Fn3 exhibited low solubil-
attacks of oedema in the hands, face, feet, abdomen and/ ity at neutral pH, unspecific binding to the dextran matri-
or throat. This condition is caused by a genetic deficiency ces of the size-exclusion chromatography column and the
of C1 esterase inhibitor (C1-Inh), which is an endogenous dextran coated biosensor chips used for the subsequent
inhibitor of plasma kallikrein. In Europe, HAE is treated characterization of Ubi-Fn3. More importantly, the affinity
with plasma-derived C1-Inh which attenuates attacks and was relatively low; the IC50 determined by competition
may prove life-saving, but C1-Inh is expensive and requires ELISA was 5 µM.
the use of pooled blood product (Nixon and Wood, In order to select 10Fn3 domains binding to a target of
2006). After an iterative selection and screening approach interest with improved properties than Ubi4-Fn3, an alter-
chosen by the authors (Markland et al., 1996), the plasma native library design in combination with a fully in vitro
kallikrein inhibitor DX-88 (Dyax Corp/Genzyme Corp) selection system (mRNA display (Wilson et al., 2001)) was
was isolated and proved to be a potent inhibitor (Ki = used for the isolation of 10Fn3 variants binding to tumour
40 pM) with high selectivity (Table 1 in Williams and necrosis factor-α (TNF-α) (Xu et al., 2002). Clones from
Baird, 2003). The use of DX-88 in a C1-Inh deficient the ninth and tenth round of selection were cloned in
mouse model demonstrated that DX-88 is effective in E. coli, sequenced and expressed. Affinity constants were
preventing changes in vascular permeability (Han et al., determined by incubating the in vitro translated
35
2002). In December 2009, DX-88 (ecallantide) was S-methionine labelled proteins with biotinylated TNF-α
approved by FDA for the treatment of HAE. at different concentrations. The solution containing the
10
With the first approved drug kunitz type domains rep- Fn3 mutants bound to TNF-α were aspirated by vacuum
resent the most advanced scaffold in this field. The strategy onto a membrane coated with streptavidin, therefore cap-
of taking a human protease inhibitor as a scaffold and turing protein complexes on the membrane. Binding was

192
 Scaffolds: small globular proteins as antibody substitutes Chapter | 13 |

analysed by measuring the radioactivity on the membrane, interest and significant potential of antibody alternatives
the KD for selected mutants were found in the range of in the field of pharmaceutical biotechnology.
1-24 nM. Further affinity maturation procedures revealed
KD values around 100 pM, the best being 20 pM, and
A-domains (Avimers)
analytical gel filtration showed that the apparent molecu-
lar weight of purified, soluble wild-type (wt) and mutant A family of A-domains (Huang et al., 1999; North and
domains was consistent with the variants being mono- Blacklow, 1999) has been described to be suitable as a
meric. The reported affinities, especially for TNF-α, are scaffold for the generation of new binding proteins (Silver-
very high. However, these values should be evaluated criti- man et al., 2005). A-domains occur as strings of multiple
cally because first, the capture step of the biotinylated domains in several cell-surface receptors (Figure 13.1C).
TNF-α-Fn3 protein complexes was performed in a solid Domains of this family bind over 100 different known
phase format, and second, TNF-α is homotrimeric protein, targets, including small molecules, proteins and viruses
so avidity effects may contribute to the high apparent (Krieger and Herz, 1994; Gliemann, 1998). Such a target
affinity observed. is typically contacted by multiple A-domains with each
Other 10Fn3 variants were shown to recognize ανβ3- domain binding independently to a unique epitope,
integrin expressed on cell surface and inhibiting ανβ- thereby binding with high avidity. Each of the 217 human
dependent cell adhesion (Richards et al., 2003) and to A-domains comprises approximately 35 amino acids and
bind to vascular endothelial growth factor receptor 2 the domains are separated by linkers that average five
(VEGF-R2) inhibiting the activation by VEGF-A, -C and –D amino acids in length. Native A-domains fold quickly and
(Parker et al., 2005; Getmanova et al., 2006). Indeed, the efficiently to a uniform, stable structure mediated by
test case of 10Fn3 for the in vivo behaviour in human calcium binding and disulfide formation. A conserved
patients is the VEGF-R2 binding CT-322 (KD = 11 nM), scaffold motif of only 12 amino acids is required for this
currently in clinical development by Bristol Myers Squibb common structure (Koduri and Blacklow, 2001). Stemmer
in cancer patients (Choy et al., 2002; Molckovsky and and co-workers have designed a phage display library of
Siu, 2008). The CT-322 10Fn3 domain is coupled via a A-domains, consisting of a conserved consensus sequence
C-terminal cysteine to a 40-kDa polyethylene glycol (PEG) in the scaffold motif and variable amino acids at different
to increase the size of the molecule above the threshold positions (Silverman et al., 2005). Like other directed evo-
for kidney-mediated clearance. In a Phase I study in lution technologies, the process the authors developed to
patients with solid tumours, CT-322 administered intrave- isolate binding proteins, included multiple recursive
nously at 1 mg/kg showed a terminal half-life of 68.7 cycles, each consisting of library generation and screening
hours, based on the first dose of a weekly dosing regimen in functional assays. However, rather than mutagenizing
(Molckovsky and Siu, 2008). The half-life in humans is the protein between cycles (e.g. for affinity maturation
substantially shorter than that of IgG (7–26 days), as purposes), they added a new domain library adjacent to
expected for a molecule lacking FcRn binding, and also the domain(s) selected in previous rounds.
shorter than the 14-day half-life reported for Cimzia, a The result of this process is a single protein chain
pegylated Fab that binds to TNF-α (Choy et al., 2002). containing multiple domains, each of which having a
Still, the pharmacokinetics of CT-322 supported weekly separate binding function. The A-domains are therefore
intravenous dosing, and a biomarker for VEGF-R2 block- called ‘avimers’, from avidity multimers. A heterotetramer
ade showed sustained elevation over several repeated consisting of three IL-6 binding A-domains and an IgG
weekly doses. Immunogenicity was also assessed in this binding domain (named C326) showed a remarkable
Phase I trial showing that 31 out of 39 patients developed affinity (picomolar range) to its target and exhibited sub-
antibodies to CT-322 (Beyond Antibodies Conference, picomolar IC50s in cell-proliferation assays (Silverman
2009, San Diego) but these did not lead to a clinical sig- et al., 2005). Moreover, C326 completely abrogated acute-
nificant immunogenicity (Bloom and Calabro, 2009). As phase protein induction by human IL-6 in mice in a
for other biologics, immunogenicity should be monitored dose-dependent manner, suggesting that C326 inhibited
carefully for this scaffold, especially for the generation of IL-6 functions in vivo. The technology of using the
cross-reactive antibodies binding not only to the injected modular platform of A-domains was commercialized by
mutant, but also to the endogenous fibronectin domain. Avidia Inc. that was acquired by Amgen for $290 million
In the near future we will learn more about this scaffold, in 2006. A placebo-controlled Phase I study of C326 was
since patients are currently being enrolled in a Phase II registered in the United States in September 2006, with
trial in glioblastoma multiforme, in which CT-322 will be patients suffering from Crohn’s disease. Since then, to our
tested alone or in combination with irinotecan (Campto, knowledge, new results have not been published in the
Pfizer), a cytostatic agent inhibiting DNA-topoisomerase public domain. Recent data (Beyond Antibodies Confer-
I. The technology of using the 10Fn3 domain was com- ence, 2009, San Diego) indicate that Amgen is further
mercialized by Adnexus, which was acquired by Bristol engineering avimers for the neutralization of IL-6 medi-
Myers Squibb for $430 million in 2007, proving the ated effects.

193
Section | 2 | Drug Discovery

mechanism of alternative splicing at the level of the

SH3 domain of Fyn kinase (Fynomers)
primary transcript (Zardi et al., 1987). The Fynomer clone
Fyn kinase is a 59 kDa member of the Src family of tyro- termed D3 exhibited low nM affinities to EDB (Grabu-
sine kinases (Cooke and Perlmutter, 1989; Resh, 1998). lovski et al., 2007). Furthermore, genetically fused homo-
As a result of alternative splicing, the Fyn protein is dimers of D3 were cloned, expressed and analysed for
expressed as two isoforms, differing in approximately 50 their capability to target tumoural neo-vasculature in vivo.
amino acids in a region between their SH2 and kinase In addition, we could demonstrate that neither Fyn SH3
domain. One form is found in thymocytes, splenocytes WT nor the D3 mutant was immunogenic in mice after
and some hematolymphoid cell lines (FynT), while the repeated intravenous injections (Grabulovski et al., 2007).
second form accumulates principally in the brain (Cooke In another study, Bertschinger et al. successfully used
and Perlmutter, 1989). Its biological functions are diverse Covalent DNA Display technology (Bertschinger and Neri,
and related to Fyn’s ability to associate and to phosphor- 2004) to isolate Fynomers binding to mouse serum
ylate a variety of intracellular signaling molecules. One of albumin (Bertschinger et al., 2007). More recently, we
the best known functions of Fyn kinase is the phosphor- have shown the construction of a novel, large Fyn SH3
ylation of SLAM (signalling lymphocyte activation mole- library comprising more than 8 × 1010 individual clone
cule) in T cells, inducing a signalling complex modulating members (Advancing Protein Therapeutics Conference,
interferon-γ expression (Li, 2005). The interaction between 2010, Frankfurt, Germany). In addition, a large collection
Fyn and SLAM occurs via the SH3 domain of Fyn and the of formats (homo-dimers, homo-trimers, bispecific
adaptor protein SAP (SLAM associated protein), forming dimers, bivalent and tetravalent Fc fusions) and sub-
a ternary complex. Interestingly, Fyn SH3 associates with nanomolar Fynomer derivatives binding to human IL-17A
SAP through a surface–surface interaction that does not have been presented. Overall, the single-pot Fyn SH3
involve the canonical PXXP recognition motif of SH3 library may provide useful reagents for many biochemical
domains (Chan et al., 2003). The Fyn SH3 domain com- and biomedical applications as an alternative to more
prises 63 residues (amino acids 83–145 of the sequence conventional IgG-based immunochemical technologies.
reported by Kawakami et al., 1986; Semba et al., 1986)
and shows the typical SH3 topology of two perpendicular Affibodies: Z-domain of
β-sheets and a single turn of 310 helix (Figure 13.1D).
staphylococcal protein A
Interacting regions of the Fyn SH3 domain include the
RT-loop, n-src loop and single residues of the β3- and β4 The immunoglobulin binding domain of staphylococcal
strands (Chan et al., 2003). With its biophysical properties protein A (SPA) is widely used as an immunochemical
and its primary structure, Fyn SH3 perfectly matches ligand for the purification or detection of antibodies and
important criteria for a scaffold to be used as alternative serves as affinity tag in fusion proteins (Uhlen et al., 1992).
to antibodies: (i) it is expressed in bacteria at high level in The IgG binding domain of SPA consists of five highly
soluble, monomeric form (Morton et al., 1996), (ii) it is homologous three-helix bundle domains of approximately
stable (TM: 70.5° C) (Filimonov et al., 1999), (iii) it does 58 amino acid residues each. Because SPA is known as a
not contain any cysteine residues, (iv) it is from human non-cysteine containing, highly soluble, proteolytically
origin and (v) its amino acid sequence is conserved in and thermally stable protein, an engineered version of one
mouse, rat, monkey (gibbon) and man. As mentioned of the SPA domains, the so-called Z-domain (Nilsson et al.,
above, its binding mode to SAP has been resolved and 1987), was chosen as a scaffold for the development of
indicates that this particular domain does not necessarily novel affinity proteins designated as ‘affibodies’ (Figure
need a PXXP core binding motif (Chan et al., 2003), 13.1E). Utilizing structural data available for the complex
which is an important prerequisite to be well suited for between a native SPA domain and the Fc fragment of
the isolation of binders against a variety of different target human IgG1, 13 positions distributed across two helices,
epitopes. Recently, we presented the design, construction, located at the surface of the domain and involved in the
and characterization of a human Fyn SH3 phage library Fc interaction, were chosen for random mutagenesis, and
containing more than 1 billion individual clones, where subsequently, for the creation of phage libraries (Nord
both loops of the Fyn SH3 domain (the RT- and n-Src- et al., 1995). In 1997, from two medium-sized libraries
loop) were randomized (Grabulovski et al., 2007). Using comprising about 4 × 107 individual clones, binding pro-
phage display technology, the isolation and in vitro char- teins against three model target proteins (Taq DNA
acterization of Fyn SH3 proteins (termed Fynomers) polymerase, human insulin, and a human apolipoprotein
binding to the extra-domain B (EDB) of fibronectin (Cas- A-1 variant) were isolated, with binding affinities in the
tellani et al., 1994; Kaspar et al., 2006), a marker of ang- micromolar range (Nord et al., 1997). In the meantime,
iogenesis, was described. Fibronectin is a large glycoprotein affibodies have been isolated against a variety of protein
that is present in large amounts in plasma and body targets (e.g. human factor VIII (Nord et al., 2001), human
tissues. EDB is a 91-amino acid type III homology domain immunoglobulin A (Ronnmark et al., 2002) and CD28
that becomes inserted into the fibronectin molecule by a (Sandstrom et al., 2003)). In a more recent paper (Wikman

194
 Scaffolds: small globular proteins as antibody substitutes Chapter | 13 |

et al., 2004), an affibody ligand binding to human epider- numbers of this repeat module were cloned between
mal growth factor receptor 2 (Her2) was isolated with a capping repeats, which are special terminal repeats of
dissociation constant (KD) of 50 nmol/L. Dimerization of ankyrin domains shielding the hydrophobic core. Two
this affibody molecule resulted in improved target binding libraries were created with two and three, respectively, ran-
affinity (KD = 3 nmol/L), and radioiodination of the dimer domized ankyrin repeats in between an N-terminal and a
allowed selective targeting and imaging of Her2-expressing C-terminal cap.
xenografts in vivo (Steffen et al., 2006). Further affinity Using a library with more than 1010 individual members
maturation strategies led to an affibody molecule with a in combination with the ribosome display selection meth-
dissociation constant in the range of 20 pmol/L and to odology, binding variants (termed DARPins) were selected
better tumour targeting and imaging results in the same by performing four or five rounds of selection. Dissocia-
tumour mouse model (Orlova et al., 2006). A fully syn- tion constants were determined by surface plasmon reso-
thetic affibody molecules has also been described, site- nance and found to be in the range of 2–20 nM against
specifically and homogenously conjugated with a DOTA these model target proteins. More recently, a so-called SRP
(1,4,7,10-tetra-azacyclododecane-N,N’,N’’,N’’’-tetraacetic phage display methodology was employed for efficient
acid) chelator, produced in a single chemical process by filamentous phage display of DARPins. Using a phage
peptide synthesis (Orlova et al., 2007). DARPin library with more than 1010 individual members
Importantly, the first clinical investigation of 68Ga- and resulted in isolation of well-behaved and highly specific
111
In-labelled anti Her2 affibody molecule (ABY-002) in DARPins against a broad range of target proteins (such as
three patients with metastatic breast cancer showed that the Fc domain of human IgG, TNFalpha, ErbB1 (EGFR),
Her2-specific affibody molecules have the potential to ErbB2 (Her2) and ErbB4 (Her4)) having affinities as low
visualize Her2-expressing metastatic lesions (Löfblom as 100 pM directly from this library, without affinity matu-
et al., 2010). High contrast SPECT or PET images were ration (Steiner et al., 2008).
obtained already 2–3 h after injection. Over-expression of In April 2010, the first cohorts of patients were enrolled
Her2 in two metastases was confirmed on biopsy tissue in two separate Phase I trials with the first DARPin
samples using the HercepTest™ indicating that ABY-002 (MP0112) neutralizing VEGF for a single intravitreal
specifically targets Her2 also in humans. Further clinical injection in wet age-related macular degeneration (wet
studies are warranted to assess the sensitivity and specifi- AMD) and diabetic macular edema (DME) (www.
city of radiolabelled Her2-targeting affibody molecules molecularpartners.ch). The trials are investigating the
(Löfblom et al., 2010). For therapeutic applications, con- safety and tolerability of MP0112, but the studies will also
cerns about the immunogenic potential of this class of allow a preliminary assessment of efficacy and the dura-
proteins should be seriously considered, owing to its bac- tion of action of MP0112. The DARPin was engineered to
terial origin. have a long ocular half-life, but fast systemic clearance.
In general, as the DARPins were newly designed and are
not found in nature in this format, their immunogenic
Ankyrin repeat proteins (DARPins) potential should be investigated in detail, especially after
Ankyrin repeat (AR) proteins, first isolated in mammalian repeated systemic administration.
erythrocytes, are involved in the targeting, mechanical
stabilization and orientation of membrane proteins to
Lipocalins
specialized compartments within the plasma membrane
and endoplasmic reticulum. Natural ankyrin repeat pro- The lipocalins represent a family of functionally diverse,
teins consist of many 33-amino-acid modules, each com- small proteins that comprise 160–180 amino acid resi-
prising a β-turn and two anti-parallel α-helices (Sedgwick dues and have weak sequence homology, but high similar-
and Smerdon, 1999). In most known complexes, the β- ity at the tertiary structural level (Skerra, 2000). Members
turn and the first α-helix mediate the interactions with the of this family have important biological functions in a
target, and different numbers of adjacent repeats are variety of organisms, from bacteria to humans. The major-
involved in binding. The reported target binding affinities ity of lipocalins are responsible for the storage and trans-
of natural AR proteins are in the low nanomolar range port of compounds that have low solubility or are
(Suzuki et al., 1998). Ankyrin repeat proteins were built chemically sensitive, such as vitamins, steroids and meta-
and diversified to create a library from which ankyrin vari- bolic products (Flower, 1995). An example of human
ants were selected binding to maltose-binding-protein lipocalin is the retinol binding protein (RBP), of which
and two eukaryotic kinases (Binz et al., 2004; Amstutz the three-dimensional-structure has been elucidated by
et al., 2005; Zahnd et al., 2006). In the approach chosen X-ray cristallography (Cowan et al., 1990). RBP transports
by the authors, a consensus ankyrin repeat module con- the poorly soluble and oxidation-prone vitamin A from
sisting of six diversified potential interaction residues and the liver, where it is stored as a fatty acid ester, to several
27 framework residues was designed based on sequence target tissues. Despite their extremely poor sequence
alignments and structural analyses (Figure 13.1F). Varying homology, the lipocalins share a structurally conserved

195
Section | 2 | Drug Discovery

β-barrel as their central folding motif, which consists of target cryptic epitopes normally hidden for whole anti-
eight antiparallel β-strands that are arranged in a cylindri- bodies and even for smaller fragments thereof. However,
cal manner (Figure 13.1G). The binding specificity for they were described to be poorly soluble and often prone
low-molecular-weight compounds is well-characterized to aggregation, because of their exposed hydrophobic
for many lipocalins: some bind their ligands with high surface normally ‘capped’ by the VL domain. Interest was
specificity, whereas others form complexes with a consid- recently revived when it was discovered that at least two
erable range of lipophilic molecules (Schlehuber and types of organisms, the camelids and cartilaginous fish
Skerra, 2005). Typically, the ligand affinities of lipocalins have evolved high-affinity V-like single domains, mounted
are moderate with dissociation constants of approximately on an Fc-equivalent constant domain framework as an
1 µM, which is consistent with their presumed function as integral and crucial component of their immune system
a physiological buffer for the bound substance; however, (De Genst et al., 2004; Dooley and Flajnik, 2005).
there are notable exceptions, e.g. the tick histamine- Unlike mouse VH domains (Ward et al., 1989), camelid
binding protein (HBP)-2 with a KD of 1.7 nM (Paesen VhH (termed nanobodies) and shark V-NAR domains are
et al., 1999). The insect bilin binding protein (BBP) from in general soluble and can be produced as stable in vitro
Pieris brassicae served as a model lipocalin in initial studies reagents. However, for in vivo administration, humaniza-
to create artificial binding sites for several ligands. Sixteen tion (or deimmunization) may be crucial to reduce immu-
amino acid positions located within the four loops at the nogenicity. Human single domain antibodies (dAbs)
open end of the β-barrel and adjoining regions of the (Figure 13.1H) would be even more preferable, and the
β-strands were subjected to targeted random mutagenesis problems of poor stability and solubility have been solved
(Beste et al., 1999). The resulting molecular random for some human V domains by the identification and
library with about 4 × 108 members was subjected to selec- design of mutations that minimize the hydrophobic VH/
tion towards several low-molecular-weight compounds VL interface (Dottorini et al., 2004; Jespers et al., 2004a).
using phage display. BBP variants, so-called ‘anticalins’, Moreover, in a publication of Jespers and co-workers
that specifically recognize fluorescein, digoxigenin, (2004b), aggregation-resistant domain antibodies could
phthalic acid esters and doxorubicin were obtained, with be directly selected on phage by heat denaturation. Start-
affinities in the nM range (Beste et al., 1999; Schlehuber ing from a DP47d domain antibody (a typical human VH
et al., 2000; Mercader and Skerra, 2002). In order to dAb), which unfolds irreversibly and forms aggregates if
extend the concept of anticalins, several human lipocalins heated above 55°C, a repertoire containing approximately
were subjected to random mutagenesis generating libraries 1 billion different mutants was cloned by diversification
that enable recognition of macromolecular protein targets. of the CDR loops. The library was multivalently displayed
Because proteins have larger molecular dimensions than on phage and after three rounds of heat denaturation fol-
small compounds, they cannot penetrate into the ligand- lowed by selection on protein A (a ligand common to
binding pocket of lipocalins. Consequently, side chains of folded dAbs), 179 out of 200 colonies secreted dAb phage
lipocalins at more exposed positions, close to the tips of that retained more than 80% of protein A-binding activity
the four loops at the open end of the β-barrel, were sub- after heating (in contrast to the starting protein DP47d,
jected to random mutagenesis. Following this strategy, a which loses binding by a factor of 560 after heating).
panel of anticalins based on human lipocalins with spe- Twenty clones were sequenced and revealed many unique
cificities for several therapeutic targets such as cytotoxic dAb sequences with a large variability in the CDR length
T-lymphocyte-associated antigen (CTLA-4), c-met, IL- and sequences, which shows that mutations located only
4Ra, VEGF and other undisclosed targets were isolated in the CDR loops of the dAbs are sufficient to confer resist-
(www.pieris-ag.com). The VEGF binding anticalin (PRS- ance to aggregation. Interestingly, the TM of the mutants
050) is about to enter clinical trials and we will learn more was not higher than the one of the parental clone DP47d,
about this scaffold in the near future. which shows that the selected dAbs are not heat stable, but
can fold reversibly.
Most data in the public domain describe serum albumin-
Single domain antibodies binding domain antibodies (AlbudAbs) fusions, which
Although not belonging to the strict definition of antibody extend the serum half-life of otherwise short-lived pro-
alternative protein ‘frameworks’, single domain antibody teins, such as the interleukin-1 receptor antagonist IL-1ra
fragments are included as well in this chapter since they (Holt et al., 2008) and interferon (IFN)-α2b (Walker
share a lot of common properties with the ‘classical’ scaf- et al., 2010). In the latter study, IFN-α2b was fused to both
folds (e.g. small domains with high expression yields in human serum albumin (termed HSA-IFN-α2b) and to an
bacteria). Moreover, they represent a valuable source for AlbuDab (termed IFN-α2b-DOM7h-14). The fusion pro-
the generation of new binding proteins. teins were subjected to a pharmacokinetic (PK) analysis in
In a seminal early publication (Ward et al., 1989), rats following intravenous dosing. The in vivo half-life of
mouse single variable domains (VH) were shown to be both fusion proteins was significantly increased in com-
functional, and it was proposed that they could potentially parison with the IFN-α standard (22.6 and 14.2 h for

196
 Scaffolds: small globular proteins as antibody substitutes Chapter | 13 |

IFN-α2b-DOM7h-14 and HSA-IFN-α2b, respectively com- Aptanomics; Kunz et al., 2006)). Antibody fragments of
pared with 1.2 h for IFN-α standard). The t1/2 of IFN- the constant domain have also been used as scaffolds, e.g.
α2b-DOM7h-14, interestingly, is approximately 1.5 times the CH2 domain of human IgG1 has been used as a library
longer than that of HSA-IFN-α2b. Moreover, similar AUC scaffold, and isolated CH2 domains with specific binding
values were obtained for IFN-α2b-DOM7h-14 and HSA- affinity to a HIV-1 gp120-CD4 complex have been
IFN-α2b, 737.5 and 689.2 h mg/mL, respectively, which described (Xiao et al., 2009). There, the new binding site
in both cases represents a significant increase over the is formed by engineered sequences in the BC and FG loops
value of 18.8 h mg/mL observed with the IFN-α standard. of the CH2 domain located at the N-terminal tip of the
A further study was carried out to determine PKs following domain. Another approach was chosen by Wozniak-Knopp
subcutaneous administration in rats. As with the i.v. study and co-workers (2010), where the whole Fc fragment of
the in vivo half-life of both fusion proteins was increased IgG1 was used as an alternative small-size antibody format.
in comparison with IFN-α standard. Interestingly, the anti- With the exception of the antigen binding site, the full Fc
viral efficacy of the AlbuDab fusion IFN-α2b-DOM7h-14 of an IgG1 has all the properties of a complete antibody,
was 5.8 fold greater in comparison with albumin fusion i.e. the ability to elicit effector functions via binding to
HSA-IFN-α2b, as determined by an in vitro antiviral assay Fc-gamma receptors and to the complement activator C1q,
with A549 human lung carcinoma cells, indicating that in as well as the long half-life of antibodies mediated through
this particular case fusion of IFN-α to the Albudab repre- binding to FcRn. Using their CH3 library and yeast display
sents an attractive avenue to prolong its half-life. selection technologies these authors described recently the
Today, the single domain antibody technology is com- isolation of a Her2 binding Fc mutant with low nM affini-
mercialized by Glaxo Smith Kline, which acquired Doman- ties (Wozniak-Knopp et al., 2010).
tis Ltd. in December 2006 for £230 million, and certainly A few other scaffolds have also been commercialized by
reflects the interest of big pharmaceutical companies in small- and medium-sized companies; however, no refer­
single domain protein scaffolds. eed publications are available (ubiquitin by Scil Proteins,
C-type lectin domain by Anaphore, and a not disclosed
scaffold by Amunix; information obtained from the cor-
Other domains responding websites).
The preceding sections introduced important domain scaf-
folds that have been commercialized and for which pub-
lished data are available. A recent review by Gebauer and
Skerra (2009) listed more than 50 scaffolds that have been CONCLUDING REMARKS
described for the generation of new binding proteins. The
majority of them have been used for research purposes Overall, the engineering and practical uses of binding pro-
only (e.g. probing the specificity determinants of WW- teins derived from non-Ig scaffolds are established and
domains to their ligands (Dalby et al., 2000), elucidating these will push biological chemistry toward both basic
target recognition rules of SH3 domains (Hiipakka et al., research and applied science. Although some rational jus-
1999; Panni et al., 2002) or identifying signal transduction tification may be given for the choice of almost any scaf-
pathways with the staphylococcal nuclease as scaffold fold described so far, the true potential of the various
(Norman et al., 1999)). Other domains have been com- protein scaffolds for human therapy, including in vivo
mercialized by companies for in vitro applications (e.g. diagnostics, and considering the different approaches
PDZ domains for the generation of high-affinity detection to the implementation of effector functions, will only
reagents (Biotech Studio; Ferrer et al., 2005) or for intra­ become clear once a number of additional Phase I/II trials
cellular signalling interference applications (thioredoxin, have been completed in the near future.

REFERENCES

Amstutz P, Binz HK, Parizek P, et al. Protein & Peptide Science 2003;4: Protein Engineering, Design &
Intracellular kinase inhibitors 231–51. Selection 2007;20:57–68.
selected from combinatorial libraries Bertschinger J, Neri D. Covalent DNA Beste G, Schmidt FS, Stibora T, et al.
of designed ankyrin repeat proteins. display as a novel tool for directed Small antibody-like proteins with
Journal of Biological Chemistry evolution of proteins in vitro. prescribed ligand specificities derived
2005;280:24715–22. Protein Engineering, Design & from the lipocalin fold. Proceeding
Ascenzi P, Bocedi A, Bolognesi M, Selection 2004;17:699–707. of the National Academy of Sciences
et al. The bovine basic pancreatic Bertschinger J, Grabulovski D, Neri D. of USA 1999;96:1898–903.
trypsin inhibitor (Kunitz inhibitor): Selection of single domain binding Binz K, Pluckthun A. Engineered
a milestone protein. Current proteins by covalent DNA display. proteins as specific binding reagents.

197
Section | 2 | Drug Discovery

Current Opinion in Biotechnology Dooley H, Flajnik MF. Shark immunity Grimbert D, Vecellio L, Delépine P,
2005;16:459–69. bites back: affinity maturation and et al. Characteristics of EPI-hNE4
Binz HK, Amstutz P, Kohl A, et al. memory response in the nurse shark, aerosol: a new elastase inhibitor
High-affinity binders selected from Ginglymostoma cirratum. European for treatment of cystic fibrosis.
designed ankyrin repeat protein Journal of Immunology Journal of Aerosol Medicine
libraries. Nature Biotechnology 2005;35:936–45. 2003;16:121–9.
2004;22:575–82. Dottorini T, Vaughan CK, Walsh MA, Han ED, MacFarlane RC, Mulligan AN,
Binz HK, Amstutz P, Plückthun A. et al. Crystal structure of a human et al. Increased vascular permeability
Engineering novel binding proteins VH: requirements for maintaining a in C1 inhibitor-deficient mice
from nonimmunoglobulin domains. monomeric fragment. Biochemistry mediated by the bradykinin type 2
Nature Biotechnology 2004;43:622–8. receptor. Journal of Clinical
2005;23:1257–68. Dyax Corp. (Ley AC, L. R, Guterman SK, Investment 2002;109:1057–63.
Bloom L, Calabro V. FN3: a new protein Roberts Bruce L, Markland W, Ken Hautamaki RD, Kobayashi DK, Senior
scaffold reaches the clinic. Drug RB) Engineered human-derived RM, et al. Requirement for
Discovery Today 2009;14:949–55. kunitz domains that inhibit human macrophage elastase for cigarette
Castellani P, Viale G, Dorcaratto A, et al. neutrophil elastase. US056633143 smoke-induced emphysema in mice.
The fibronectin isoform containing patent. 1997. Science 1997;277:2002–4.
the ED-B oncofetal domain: a Egeblad M, Werb Z. New functions for Hey T, Fiedler E, Rudolph R, et al.
marker of angiogenesis. International the matrix metalloproteinases in Artificial, non-antibody binding
Journal of Cancer 1994;59:612–18. cancer progression. Nat Review. proteins for pharmaceutical and
Chan B, Lanyi A, Song HK, et al. SAP Cancer 2002;2:161–74. industrial applications. Trends in
couples Fyn to SLAM immune Ferrer M, Maiolo J, Kratz P, et al. Biotechnology 2005;23:514–22.
receptors. Nature Cell Biology Directed evolution of PDZ variants Hiipakka M, Poikonen K, Saksela K.
2003;5:155–60. to generate high-affinity detection SH3 domains with high affinity and
Choy EH, Hazleman B, Smith M, et al. reagents. Protein Engineering, Design engineered ligand specificities
Efficacy of a novel PEGylated & Selection 2005;18:165–73. targeted to HIV-1 Nef. Journal of
humanized anti-TNF fragment Filimonov VV, Azuaga AI, Viguera AR, Molecular Biology 1999;293:
(CDP870) in patients with et al. A thermodynamic analysis of a 1097–106.
rheumatoid arthritis: a phase II family of small globular proteins: Hogbom, M, Eklund M, Nygren PA,
double-blinded, randomized, SH3 domains. Biophysical Chemistry et al. Structural basis for recognition
dose-escalating trial. Rheumatology 1999;77:195–208. by an in vitro evolved affibody.
(Oxford) 2002;41:1133–7. Flower DR. Multiple molecular Proceeding of the National Academy
Cooke MP, Perlmutter RM. Expression recognition properties of the of Sciences of USA 2003;100:3191–6.
of a novel form of the fyn proto- lipocalin protein family. Journal of Holliger P, Hudson PJ. Engineered
oncogene in hematopoietic cells. Molecular Recognition antibody fragments and the rise of
New Biology 1989;1:66–74. 1995;8:185–95. single domains. Nature
Cowan SW, Newcomer ME, Jones TA. Gebauer M, Skerra A. Engineered Biotechnology 2005;23:1126–36.
Crystallographic refinement of protein scaffolds as next-generation Holt LJ, Basran A, Jones K, et al.
human serum retinol binding antibody therapeutics. Current Anti-serum albumin domain
protein at 2A resolution. Proteins Opinion in Chemical Biology antibodies for extending the
1990;8:44–61. 2009;13:245–55. half-lives of short lived drugs.
Dalby PA, Hoess RH, DeGrado WF. Getmanova EV, Chen Y, Bloom L, et al. Protein Engineering, Design &
Evolution of binding affinity in a Antagonists to human and mouse Selection 2008;21:283–8.
WW domain probed by phage vascular endothelial growth factor Huang, W, Dolmer K, Gettins PG. NMR
display. Protein Science receptor 2 generated by directed solution structure of complement-
2000;9:2366–76. protein evolution in vitro. Chemical like repeat CR8 from the low density
De Genst E, Handelberg F, Van Biology 2006;13:549–56. lipoprotein receptor-related protein.
Meirhaeghe A, et al. Chemical basis Gliemann J. Receptors of the low Journal of Biological Chemistry
for the affinity maturation of a camel density lipoprotein (LDL) receptor 1999;274:14130–6.
single domain antibody. Journal of family in man. Multiple functions Jespers L, Schon O, Famm K, et al.
Biological Chemistry of the large family members via Aggregation-resistant domain
2004;279:53593–601. interaction with complex ligands. antibodies selected on phage by heat
Delacourt, C, Hérigault S, Delclaux C, Biological Chemistry 1998;379: denaturation. Nature Biotechnology
et al. Protection against acute lung 951–64. 2004a;22:1161–5.
injury by intravenous or intratracheal Grabulovski D, Kaspar M, Neri D. Jespers L, Schon O, James LC, et al.
pretreatment with EPI-HNE-4, a A novel, non-immunogenic Fyn Crystal structure of HEL4, a soluble,
new potent neutrophil elastase SH3-derived binding protein with refoldable human V(H) single
inhibitor. American Journal of tumor vascular targeting properties. domain with a germ-line scaffold.
Respiratory Cell Molecular Biology Journal of Biological Chemistry Journal of Molecular Biology
2002;26:290–7. 2007;282:3196–204. 2004b;337:893–903.

198
 Scaffolds: small globular proteins as antibody substitutes Chapter | 13 |

Kaspar M, Zardi L, Neri D. Fibronectin display. 2. Plasma kallikrein and Nunan J, Small DH. Proteolytic
as target for tumor therapy. thrombin. Biochemistry processing of the amyloid-beta
International Journal of Cancer 1996;35:8058–67. protein precursor of Alzheimer’s
2006;118:1331–9. Mercader JV, Skerra A. Generation of disease. Essays in Biochemistry
Kawakami T, Pennington CY, Robbins anticalins with specificity for a 2002;38:37–49.
KC. Isolation and oncogenic nonsymmetric phthalic acid ester. Nygren PA, Uhlen M. Scaffolds for
potential of a novel human src-like Annals of Biochemstry engineering novel binding sites in
gene. Molecular Cell Biology 2002;308:269–77. proteins. Current Opinion in
1986;6:4195–201. Molckovsky A, Siu LL. First-in-class, Structural Biology 1997;7:463–9.
Koduri V, Blacklow SC. Folding first-in-human phase I results of Orlova A, Magnusson M, Eriksson TL,
determinants of LDL receptor targeted agents: highlights of the et al. Tumor imaging using a
type A modules. Biochemistry 2008 American Society of Clinical picomolar affinity HER2 binding
2001;40:12801–7. Oncology meeting. Journal of affibody molecule. Cancer Research
Koide A, Koide S. The fibronectin type Hematology and Oncology 2006;66:4339–48.
III domain as a scaffold for novel 2008;1:20. Orlova A, Tolmachev V, Pehrson R,
binding proteins. Journal of Morton CJ, Pugh DJ, Brown EL, et al. et al. Synthetic affibody molecules:
Molecular Biology 1998;284: Solution structure and peptide a novel class of affinity ligands for
1141–51. binding of the SH3 domain from molecular imaging of HER2-
Krieger M, Herz J. Structures and human Fyn. Structure 1996;4: expressing malignant tumors.
functions of multiligand lipoprotein 705–14. Cancer Research 2007;67:
receptors: macrophage scavenger Naruse S, Kitagawa M, Ishiguro H. 2178–86.
receptors and LDL receptor-related Molecular understanding of chronic Paesen GC, Adams PL, Harlos K, et al.
protein (LRP). Annual Reviews in pancreatitis: a perspective on the Tick histamine-binding proteins:
Biochemstry 1994;63:601–37. future. Molecular Medicine Today isolation, cloning, and three-
Kunz C, Borghouts C, Buerger C, et al. 1999;5:493–9. dimensional structure. Molecular
Peptide aptamers with binding Nilsson B, Moks T, Jansson B, et al. Cell 1999;3:661–71.
specificity for the intracellular A synthetic IgG-binding domain Panni S, Dente L, Cesareni G. In vitro
domain of the ErbB2 receptor based on staphylococcal protein A. evolution of recognition specificity
interfere with AKT signaling and Protein Engineering 1987;1:107–13. mediated by SH3 domains reveals
sensitize breast cancer cells to Taxol. Nixon AE, Wood CR. Engineered target recognition rules. Journal of
Molecular Cancer Researchearch protein inhibitors of proteases. Biological Chemistry 2002;277:
2006;4:983–98. Current Opinion in Drug Discovery 21666–74.
Laval SH, Bushby KM. Limb-girdle and Development 2006;9:261–8. Parker MH, Chen Y, Danehy F, et al.
muscular dystrophies – from Nord K, Nilsson J, Nilsson B, et al. Antibody mimics based on human
genetics to molecular pathology. A combinatorial library of an fibronectin type three domain
Neuropathology and Applied alpha-helical bacterial receptor engineered for thermostability and
Neurobiology 2004;30:91–105. domain. Protein Engineering high-affinity binding to vascular
Ley AC, LR, Guterman SK, Roberts 1995;8:601–8. endothelial growth factor receptor
Bruce L, et al. US056633143 patent Nord K, Gunneriusson E, Ringdahl J, two. Protein Engineering, Design &
1997. et al. Binding proteins selected from Selection 2005;18:435–44.
Ley AC, Markland W, Ladner RC. combinatorial libraries of an Parks WC, Wilson CL, López-Boado YS.
Obtaining a family of high-affinity, alpha-helical bacterial receptor Matrix metalloproteinases as
high-specificity protein inhibitors domain. Nature Biotechnology modulators of inflammation and
of plasmin and plasma kallikrein. 1997;15:772–7. innate immunity. Nature Review.
Molecular Divers 1996;2:119–24. Nord K, Nord O, Uhlén M, et al. Immunology 2004;4:617–29.
Li SS. Specificity and versatility of SH3 Recombinant human factor Reichert JM. Editorial for PEDS special
and other proline-recognition VIII-specific affinity ligands selected issue on antibodies. Protein
domains: structural basis and from phage-displayed combinatorial Engineering, Design & Selection
implications for cellular signal libraries of protein A. European 2010;23:153–4.
transduction. Biochemical Journal Journal of Biochemistry Resh MD. Fyn, a Src family tyrosine
2005;390:641–53. 2001;268:4269–77. kinase. International Journal of
Löfblom J, Feldwisch J, Tolmachev V, Norman TC, Smith DL, Sorger PK, et al. Biochemistry and Cell Biology
et al. Affibody molecules: engineered Genetic selection of peptide 1998;30:1159–62.
proteins for therapeutic, diagnostic inhibitors of biological pathways. Richards J, Miller M, Abend J, et al.
and biotechnological applications. Science 1999;285:591–5. Engineered fibronectin type III
FEBS Letters. Epub ahead of print 11 North CL, Blacklow SC. Structural domain with a RGDWXE sequence
April 2010. independence of ligand-binding binds with enhanced affinity and
Markland W, Ley AC, Ladner RC. modules five and six of the LDL specificity to human alphavbeta3
Iterative optimization of high-affinity receptor. Biochemistry integrin. Journal of Molecular
protease inhibitors using phage 1999;38:3926–35. Biology 2003;326:1475–88.

199
Section | 2 | Drug Discovery

Roberts BL, Markland W, Ley AC, et al. kinase family. Proceeding of the Ward ES, Güssow D, Griffiths AD, et al.
Directed evolution of a protein: National Academy of Sciences of Binding activities of a repertoire of
selection of potent neutrophil USA 1986;83:5459–63. single immunoglobulin variable
elastase inhibitors displayed on M13 Shapiro SD. Proteinases in chronic domains secreted from Escherichia
fusion phage. Proceeding of the obstructive pulmonary disease. coli. Nature 1989;341:544–6.
National Academy of Sciences of Biochemical Society Transactions Wikman M, Steffen AC, Gunneriusson
USA 1992;89:2429–33. 2002;30:98–102. E, et al. Selection and
Ronnmark J, Grönlund H, Uhlén M, Silverman J, Liu Q, Bakker A, et al. characterization of HER2/neu-
et al. Human immunoglobulin A Multivalent avimer proteins evolved binding affibody ligands. Protein
(IgA)-specific ligands from by exon shuffling of a family of Engineering, Design & Selection
combinatorial engineering of protein human receptor domains. Nature 2004;17:455–62.
A. European Journal of Biochemistry Biotechnology 2005;23:1556–61. Williams A, Baird LG. DX-88 and HAE:
2002;269:2647–55. Skerra A. Lipocalins as a scaffold. a developmental perspective.
Saerens D, Ghassabeh GH, Biochimica et Biophysica Acta Transfusion and Apheresis Science
Muyldermans S. Single-domain 2000;1482:337–50. 2003;29:255–8.
antibodies as building blocks for Smith G. Patch engineering: a general Wilson DS, Keefe AD, Szostak JW. The
novel therapeutics. Current Opinion approach for creating proteins that use of mRNA display to select
in Pharmacology 2008;8:600–8. have new binding activities. Trends high-affinity protein-binding
Sandstrom K, Xu Z, Forsberg G, et al. in Biochemical Science 1998;23: peptides. Proceedings of the
Inhibition of the CD28–CD80 457–60. National Academy of Sciences of
co-stimulation signal by a CD28- Steffen AC, Orlova A, Wikman M, et al. USA 2001;98:3750–5.
binding affibody ligand developed Affibody-mediated tumour targeting Wozniak-Knopp G, Bartl S, Bauer A,
by combinatorial protein of HER-2 expressing xenografts in et al. Introducing antigen-binding
engineering. Protein Engineering mice. European Journal of Nuclear sites in structural loops of
2003;16:691–7. Medicine and Molecular Imaging immunoglobulin constant domains:
Saudubray F, Labbé A, Durieu I. Phase 2006;33:631–8. Fc fragments with engineered HER2/
IIa clinical study of a new human Steiner D, Forrer P, Plückthun A. neu-binding sites and antibody
neutrophil elastase inhibitor (HNE), Efficient selection of DARPins with properties. Protein Engineering,
EPI-HNE4 (DX-890), with repeated sub-nanomolar affinities using SRP Design & Selection 2010;23:289–97.
administration by inhalation in phage display. Journal of Molecular Xiao X, Feng Y, Vu BK, et al. A large
adult cystic fibrosis patients. Biology 2008;382:1211–27. library based on a novel (CH2)
Journal of Cystic Fibrosis scaffold: identification of HIV-1
Suzuki F, Goto M, Sawa C, et al.
2003;2:A85. inhibitors. Biochemical and
Functional interactions of
Schlehuber S, Beste G, Skerra A. A novel transcription factor human GA- Biophysical Research
type of receptor protein, based on binding protein subunits. Journal of Communications 2009;387:
the lipocalin scaffold, with specificity Biological Chemistry 387–92.
for digoxigenin. Journal of Molecular 1998;273:29302–8. Xu L, Aha P, Gu K, et al. Directed
Biology 2000;297:1105–20. evolution of high-affinity antibody
Tomlinson IM. Next-generation protein
Schlehuber S, Skerra A. Lipocalins in drugs. Nature Biotechnology mimics using mRNA display.
drug discovery: from natural 2004;22:521–2. Chemical Biology 2002;9:933–42.
ligand-binding proteins to anticalins. Zahnd C, Pecorari F, Straumann N,
Uhlen M, Forsberg G, Moks T, et al.
Drug Discovery Today et al. Selection and characterization
Fusion proteins in biotechnology.
2005;10:23–33. of Her2 binding-designed ankyrin
Current Opinion in Biotechnology
Sedgwick SG, Smerdon SJ. The ankyrin 1992;3:363–9. repeat proteins. Journal of Biological
repeat: a diversity of interactions on Chemistry 2006;281:35167–75.
Walker A, Dunlevy G, Rycroft D, et al.
a common structural framework. Zardi L, Carnemolla B, Siri A, et al.
Anti-serum albumin domain
Trends in Biochemical Science Transformed human cells produce a
antibodies in the development of
1999;24:311–16. new fibronectin isoform by
highly potent, efficacious and
Semba K, Nishizawa M, Miyajima N, long-acting interferon. Protein preferential alternative splicing of a
et al. yes-related protooncogene, syn, Engineering, Design & Selection previously unobserved exon. The
belongs to the protein-tyrosine 2010;23:271–8. EMBO Journal 1987;6:2337–42.

200
 Chapter 14 

Drug development: introduction

H P Rang, R G Hill

INTRODUCTION THE NATURE OF DRUG

DEVELOPMENT
Drug development comprises all the activities involved in
transforming a compound from drug candidate (the end- Drug discovery, as described in Section 2, is invariably an
product of the discovery phase) to a product approved for exploration of the unknown, and successful projects may
marketing by the appropriate regulatory authorities. Effi- end up with compounds quite different from what had
ciency in drug development is critical for commercial originally been sought: there is a large component of
success, for two main reasons: ‘unplannability’. In contrast, drug development has a very
• Development accounts for about two-thirds of the clear-cut goal: to produce the drug in a marketable form,
total R&D costs. The cost per project is very much and to gain regulatory permission to market it for use in
greater in the development phase, and increases the target indication(s) as quickly as possible. The work
sharply as the project moves into the later phases of required to do this falls into three main parts, respectively
clinical development. Keeping these costs under technical, investigative and managerial:
control is a major concern for management. Failure • Technical development – solving technical problems
of a compound late in development represents a lot relating to the synthesis and formulation of the drug
of money wasted. substance, aimed mainly at ensuring the quality of
• Speed in development is an important factor in the end-product:
determining sales revenue, as time spent in  Main functions involved: chemical development,
development detracts from the period of patent pharmaceutical development.
protection once the drug goes to market. As soon as • Investigative studies – establishing the safety and
the patent expires, generic competition sharply efficacy of the product, including assessment of
reduces sales revenue. whether it is pharmacokinetically suitable for clinical
Despite a high level of awareness in the pharmaceutical use in man:
industry of the need to reduce the money and time spent  Main functions involved: safety pharmacology,
on development, both have actually increased significantly toxicology, pharmacokinetics, clinical
over the last two decades (see Chapter 22). This is mainly development.
due to external factors, particularly the increased strin- • Managerial functions:
gency applied by regulatory authorities in assessing the  Coordination – managing quality control,
safety and efficacy of new compounds (see Chapter 20). logistics, communication and decision making
The development burden is, therefore, tending to increase, in a large multidisciplinary project to ensure
thereby increasing the need for companies to improve high-quality data and to avoid unnecessary
their performance in this area in order to remain profitable delays:
and competitive.  Main function involved: project management

© 2012 Elsevier Ltd. 203

Section | 3 | Drug Development

 Documentation and liaison with regulatory group selected for the trial, so as to maximize the chance
authorities – collating and presenting data of the of success in obtaining a clear-cut result. Experience shows
type, quality and format needed to secure that inconclusive clinical trials resulting from poor deci-
regulatory approval sions of this sort are a common cause of failure or delay
 Main function involved: regulatory affairs. in drug development. Where the indication allows this an
adaptive trial design may allow a more efficient evaluation
An important distinction between the technical and of the drug (see Chapter 17).
investigative aspects of development is that, in tackling
technical problems, it is assumed that a solution does
exist, and so the team’s task is to find and optimize it
COMPONENTS OF DRUG
as quickly as possible, whereas in assessing safety and
efficacy it cannot be assumed that the compound reaches DEVELOPMENT
the required standards – rather, the object is to discover
this as quickly and cheaply as possible. In other words, Figure 14.1 summarizes the main activities involved in
technical development is essentially an exercise in problem developing a typical synthetic compound. It shows the
solving, whereas clinical and toxicological development main tasks that have to be completed before the com-
is a continuing investigation of the properties of the pound can be submitted for regulatory approval, but
compound. Although technical problems, such as an needs to be translated into an operational plan (Figure
unacceptably complex and poor-yielding synthesis route, 14.2) that will allow the project to proceed as quickly and
or difficulty in developing a satisfactory formulation, can efficiently as possible. It is obvious that certain tasks have
result in abandonment of the project, this is relatively to be completed in a particular order. For example, a
uncommon. Failure on account of the drug’s biological supply of pure compound, prepared in an acceptable for-
properties, such as toxicity, poor efficacy or unsatisfactory mulation, has to be available before Phase I clinical studies
pharmacokinetics, is, however, very common, and largely can begin. Animal toxicity data must also be available
accounts for the fact that only some 10% of compounds before the compound can be given to humans. Deciding
entering Phase I clinical trials are eventually marketed. An on the dosage schedule to be used in efficacy trials requires
important aspect of the management of drug-development knowledge of the pharmacokinetics and metabolism of
projects, therefore, is to establish firm ‘no-go’ criteria, and the compound in humans. Because the data generated will
to test the compound against them as early as possible. be included in the final registration proposal, it is essential
Development proceeds along much more clearly defined that each part of the work should be formally reported and
lines than discovery, and is consequently more ‘planna- ‘signed off’ by the group responsible and archived for
ble’, particularly the non-clinical studies, where standard future reference. A typical development project is likely to
experimental protocols exist for most of the work that involve several hundred individuals, expert in different
needs to be carried out. This applies also in Phase I clinical disciplines and working on different aspects of the project,
studies. Delays can nevertheless occur if unexpected find- and coordinating their work is a complex and demanding
ings emerge, for example poor oral absorption in humans, task. For this reason, most companies assign specialist
or species-specific toxic effects, which require additional project managers to this task. Their role is to design a
work to be carried out before clinical trials can proceed. If project plan, based on input from the experts involved, to
the drug has a completely novel mechanism of action this monitor progress and to adapt the plan accordingly. As
often prolongs the technical phase whilst off-target effects well as being good organizers, project managers need to
are explored (sometimes at the insistence of the regulatory be excellent communicators, diplomatic, and with a good
authority). understanding of the scientific and technological aspects
Beyond Phase I, the route to be followed is generally of the project. Figure 14.2 is a much-simplified outline of
much less well charted, and success depends to a much a project plan of the development of a typical orally active
greater extent on strategic decisions by the project team as drug. Each ‘task’, represented by an arrow, starts and ends
to which clinical indications should be investigated (see at a circular symbol (representing an ‘event’), and decision
Chapter 17). They will need to assess, for example, whether points are marked by diamond symbols. This type of
recruiting patients to the trial will be easy or difficult, what graphical format, which is widely used as a project man-
exclusion criteria should apply, what clinical outcome agement tool and implemented in many commercially
measures should be used, and how long the treatment and available software packages, is known as a PERT (project
assessment periods will need to be. To achieve registration evaluation and review technique) chart. By assigning times
as quickly as possible, it may, for example, be expedient – shortest possible, maximum, and expected – to each
to select a relatively low-market, but quick-to-test, clinical task, the timing of the whole project can be assessed and
indication for the initial trials, and to run these trials in the critical path – i.e. the sequence of tasks that need to be
parallel with more prolonged trials in the major indica- completed on time in order to avoid an overall delay –
tion. Careful attention needs to be given to the patient defined. In Figure 14.2 the process has been reduced to a

204
 Drug development: introduction Chapter | 14 |

Technical part
Chemical development
• Check chemical stability under various conditions
• Optimize and scale up synthesis and purification
• Produce GMP batches as required for biological and clinical testing

Pharmaceutical development (see Ch.17)

• Develop and deliver formulations for biological testing
(i.e., safety pharmacology, toxicology, clinical trials)
• Develop dosage forms as required (e.g., i.v. solution, oral capsules,
nasal sprays, skin creams, etc.)
• Develop special delivery systems if required (e.g., slow release,
skin patch, liposomes, etc.)
Project management

Drug metabolism/pharmacokinetics (DMPK) (see Ch.10)

Documentation
Development • Develop analytical techniques for measuring compound and
Registration
compound metabolite concentrations in tissues
• Determine major metabolites in relevant test species, including man
• Determine PK parameters in test species

Safety assessment (see Ch.16)

• Safety pharmacology
• In vitro toxicology
• In vivo toxicology in various species

Clinical development (see Ch.17)

• Phase I (assessment of PK characteristics and side effects in
normal volunteers)
• Phase II (dose-finding and tests of efficacy in small groups of
patients)
• Phase III (definitive trials of efficacy in large groups of patients,
including comparison with standard therapies)

Investigative part

Fig. 14.1  The main technical and investigative components of a typical drug development project.

bare minimum to allow representation on a single page; completion dates for each task, many of which will be
in practice, each of the ‘tasks’ shown (e.g. develop formula- running simultaneously on any given date1.
tions, perform Phase I studies, etc.) needs to be further In this section of the book we outline the main technical
subdivided into a series of subtasks and timings to enable and experimental parts of the work that goes into drug
the project to be planned and monitored at the opera- development, namely toxicology, pharmaceutical develop-
tional level. The complete diagram for a typical drug ment and clinical studies. Chapter 19 discusses the prin-
development project will be of such size and complexity ciples underlying the patenting of drugs. Chapter 20
as to frighten all but the most hardened project manage-
ment professionals. Software tools, fortunately, are avail-
able which allow the project to be viewed in different 1
In Robert Burns’ words ‘The best-laid plans of mice and men gang
ways, such as Gantt charts, which are barcharts set against aft agley’. Drug development is no exception to this principle
a calendar timescale, showing the expected start and – managers prefer the euphemism ‘slippage’.

205
 Decision point Decision point Decision point
Start preclinical Go/No go OK to test in humans
development

GMP sythesis ~10 kg Analytical Oral/i.v. human

Analysis and release method formulation

Preliminary Stability tests Genotoxicity

Synthesis formulations
4-week toxicology
~30g Safety 2 species
Toxicology pharmacology
Dose-finding
Analytical method
Synthesis Drug and
A Rat ADME/PK ~300 g metabolites

Decision point Decision point Decision point

OK to test in humans Proceed with Ph II Proceed with Ph III

Develop formulations Synthesize Prepare final

For Ph II and Ph III ~100 kg formulations

Plan Ph I and select

study centres Decide target
indications
Plan Ph II
Prepare IND and
Ph I trials Recruit Ph II trials
investigator brochure
investigators
Plan marketing
4-week strategy
toxicology

Reproductive 6-month
toxicology toxicology

Decision point
Regulatory
Proceed with Ph III decision

Prepare final
formulations

Plan Ph III
Prepare and submit
Recruit Ph III trials
registration documents
investigators

24-month
carcinogenicity
B

Fig. 14.2  Simplified flowchart showing the main activities involved in drug development. The nodes indicated by circles
represent the start and finish points of each activity, and the diagram indicates which activities need to be completed before
the next can begin. By assigning timescales to each activity, the planned overall development time can be determined and
critical path activities identified. (A) Preclinical development. (B) Clinical development.

206
 Drug development: introduction Chapter | 14 |

Drug Preclinical Early clinical Full

Phases Launch
discovery development development development

Compound Marketed
Research ESC FSC Ph I Ph IIa Ph IIb Ph III
status Ph IV

Decision points ESP FSP DDM FDP SDP

Fig. 14.3  Strategic decision points in drug development. ESP, early selection point; FSP, final selection point; DDM, decision to
develop in man; FDP, full development decision point; SDP, submission decision point.

describes how regulatory bodies go about evaluating new Phase I trials themselves, will need to be performed on
compounds for registration, and Chapter 21 presents an a group of candidate compounds. This is clearly more
introduction to the principles of pharmaceutical market- expensive than choosing the development compound
ing. Chemical development, covering the specialized before Phase I, but may be justified as a strategy for reduc-
technical aspects of producing the drug substance eco- ing the risk of failure at a later (and more expensive) stage.
nomically and safely on a large scale, as well as the control
measures needed to ensure consistent high quality of the
final product, is beyond the scope of this book (see
Gadamasetti, 2007; Repic, 1998 for a full account of this
DECISION POINTS
subject).
The decision to advance a drug candidate into early devel-
opment is the first of several key strategic decision points
in the history of the drug development project. The timing,
THE INTERFACE BETWEEN nomenclature and decision-making process vary from
DISCOVERY AND DEVELOPMENT company to company, and Figure 14.3 shows a typical
scheme, developed by Novartis:
For the purposes of this book, drug development is pre- • The early selection point (ESP) is the decision
sented as an operation separate from discovery and fol- to take the drug candidate molecule into early
lowing on from it, but the distinction is actually not (preclinical) development. The proposal will
clear-cut. Increasingly, as has been stressed in Chapters 9 normally be framed by the drug discovery team and
and 10, activities previously undertaken during develop- evaluated by a research committee, which determines
ment are taking place earlier, as an integral part of the whether the criteria to justify further development
discovery process. The emphasis on the ‘druggability’ of have been met. After this checkpoint, responsibility
leads (Chapter 9) reflects a concern for focusing on struc- normally passes to a multidisciplinary team with
tures that are least likely to have unsatisfactory pharma- representatives from research, various development
cokinetic, toxicological or stability problems. Whereas in functions, patents, regulatory affairs and marketing,
vitro tests of absorption and metabolism were tradition- under the direction of a professional project
ally performed on individual compounds during the manager. In a large multinational company the team
development phase, they are now being incorporated will have international representation, and the
into medium-throughput screens in the discovery phase development plan will be organized to meet global
(Chapter 10), as are in vitro toxicity tests. Formulation development standards as far as possible.
work also often begins during the discovery phase, particu- • The decision to develop in man (DDM) controls
larly if the characteristics of the lead compounds suggest entry of the compound into Phase I, based on the
that specialized formulations are likely to be required for additional information obtained during the
use in pharmacological profiling of compounds in vivo. preclinical development phase (i.e. preliminary
Furthermore, the selection of a single compound for full toxicology, safety pharmacology, pharmacokinetics,
development will sometimes be deferred until data from etc.). An important task once this decision point is
Phase I trials on several candidates have been obtained. passed is normally the production of a sufficient
Thus, the preliminary work needed before Phase I, and the quantity (usually 2–5 kg) of clinical-grade material.

207
Section | 3 | Drug Development

Passing this decision point takes the project into with disaster applies, if at all, only to the technical parts
Phase I and Phase IIa clinical studies, described in of development, not to the investigational parts, which
more detail in Chapter 17, which are designed to are, as discussed above, a continuation of research into the
reveal whether the drug has an acceptable properties of the drug. Surprises in this arena can be good
pharmacokinetic and side-effect profile in normal or bad for the outcome of the project. Finding, to the
volunteers (Phase I), and whether it shows evidence company’s surprise, that sildenafil (Viagra) improved the
of clinical efficacy in patients (Phase IIa). For drugs sex life of trial subjects, set the development project off in
acting by novel mechanisms, Phase IIa provides the a completely new, and very successful, direction.
all-important first ‘proof of concept’, on which the The value attached to stopping projects reflects the frus-
decision whether or not to proceed with the serious tration – commonly felt in large research organizations –
business of full development largely rests2. that projects that have little chance of ending in success
• The full development decision point (FDP) is tend to carry on, swallowing resources, through sheer
reached after the Phase I and Phase IIa (‘proof-of- inertia, sustained mainly by the reluctance of individuals
concept’) studies have been completed, this being to abandon work to which they may have devoted many
the first point at which evidence of clinical efficacy years of effort. In practice, of course, the value of stopping
in man is obtained. It is at this point that the project a project depends only on the possibility of redeploying
becomes seriously expensive in terms of money and the resources to something more useful, i.e. on the ‘oppor-
manpower, and has to be evaluated strictly in tunity cost’. If the resources used cannot be redeployed, or
competition with other projects. Evaluation of the if no better project can be identified, there is no value in
likely commercial returns as well as the chances of stopping the project. What is certain is that it is only by
successful registration and the time and cost of the carrying projects forward that success can be achieved.
‘pivotal’ Phase III studies, are, therefore, important Despite the aphorism, it is no surprise that managers who
considerations at this point. regularly lead projects into oblivion achieve much less
• The submission decision point (SDP) is the final favourable recognition than those who bring them to
decision to apply for registration, based on a check fruition!
that the amount and quality of the data submitted
are sufficient to ensure a smooth passage through
the regulatory process. Hold-ups in registration can
carry serious penalties in terms of the cost and time THE NEED FOR IMPROVEMENT
required to perform additional clinical studies, as
well as loss of confidence among financial analysts, As emphasized elsewhere in this book, innovative new
who will have been primed to expect a new product drugs are not being registered as fast as the spectacular
to bring in revenues according to the plan as developments in biomedical science in the last decades of
originally envisaged. the 20th century had led the world to expect. The rate of
A couple of aphorisms are often applied to drug devel- new drug approvals in recent years has shown a disheart-
opment, namely: ening decline and there is little sign of the anticipated
surge (see Chapter 22), despite a steadily rising R&D
• In research, surprise = discovery; in development, spend. This worrying problem was analysed in a 2004
surprise = disaster
report by the FDA, which lays the blame firmly on the
• It is as valuable to stop a project as to carry one failure of the development process to keep up with
forward.
advances in biomedicine. The report comments: ‘the
Like most aphorisms they contain a grain of truth, but applied sciences needed for medical product development
only a small one. With regard to the first, equating surprise have not kept pace with the tremendous advances in the
basic sciences’. The report outlines a historical success rate
(i.e. chance of reaching the market) for new compounds
2
The division of Phase II clinical trials, which are exploratory tests entering Phase I clinical trials of 14% and comments that
involving relatively small numbers of patients (100–300), and lack the this figure did not improve between 1985 and 2000. Fur-
statistical power of the later ‘pivotal’ Phase III trials, into two stages (a
and b) is a fairly recent idea, now widely adopted. Phase IIa is
thermore, a recent analysis cited in this report shows that
essentially a quick look to see whether the drug administered in a the cost of development, per compound registered, almost
dose selected on the basis of pharmacokinetic, pharmacodynamic and doubled in 2000–2002 compared with 1995–2000,
toxicological data has any worthwhile therapeutic effect. Phase IIb, whereas the cost of discovery changed very little. In their
also on small numbers, is aimed at refining the dosage schedule to
optimize the therapeutic benefit, and to determine the dosage to be view, too little effort is being made to develop an improved
tested in the large-scale, definitive Phase II trials. It marks the ‘product development toolkit’ that places more reliance on
beginning of full development, following the key decision (FDP) early laboratory data, and relies less on animal models and
whether to proceed or stop. For the discovery team, the outcome of
Phase IIa largely determines whether they throw their hats in the air clinical testing in assessing safety and efficacy. The FDA
or retire to lick their wounds. and other regulatory bodies hold a large amount of data

208
 Drug development: introduction Chapter | 14 |

which could be used to analyse in a much more systematic possibility of replacing or improving existing procedures,
way than has so far been done by the predictive value of with substantial savings of money and time. The task is
particular laboratory tests in relation to clinical outcome. beyond the capabilities of any one pharmaceutical
New screening technologies and computer modelling company, but needs collaboration and funding at the
approaches need to be brought into the same frame. Cur- national or international level. The FDA has continued to
rently, extrapolation from laboratory and animal data to issue reports aimed at improving the regulatory process
the clinical situation relies largely on biological intuition and reducing delays in approval of important new drugs
– it is assumed, for example, that a compound that does (see CDER 2011; FDA, 2010).
not cause hepatotoxicity in animals is unlikely to do so in The remaining chapters in this section give a simple
man – but there may well be cheaper and quicker tests that overview of the main activities involved in drug develop-
would be at least as predictive. There are many other exam- ment. Griffin and O’Grady (2002) describe the drug devel-
ples, in the FDA’s view, where new technologies offer the opment process in more detail.

REFERENCES

CDER. Identifying CDER’s science and FDA Report. Advancing regulatory Griffin JP, O’Grady JO. The textbook of
research needs – report July 2011, science for public health. 2010. pharmaceutical medicine, 4th ed.
2011. p. 24 www.fda.gov/scienceresearch/ London: BMJ Books; 2002.
FDA Report. Innovation stagnation: specialtopics/regulatoryscience/ Repic O. 1998 Principles of process
challenge and opportunity on ucm228131.htm research and chemical development
the critical path to new medical Gadamasetti K, editor. Process chemistry in the pharmaceutical industry. New
products. 2004. www.fda.gov/ in the pharmaceutical industry, vol. York: John Wiley.
oc/initiatives/criticalpath/ 2. Chichester: CRC Press/Taylor and
whitepaper.html Francis; 2007. p. 520.

209
 Chapter 15 
Assessing drug safety
H P Rang, R G Hill

is not hypothesized cannot be tested. For this reason, the

INTRODUCTION problem of safety assessment is fundamentally different
from that of efficacy assessment, where we can define
Since the thalidomide disaster, ensuring that new medi- exactly what we are looking for.
cines are safe when used therapeutically has been one of Here we focus on non-clinical safety assessment – often
the main responsibilities of regulatory agencies. Of course, called preclinical, even though much of the work is done
there is no such thing as 100% safety, much as the public in parallel with clinical development. We describe the
would like reassurance of this. Any medical intervention various types of in vitro and in vivo tests that are used to
– or for that matter, any human activity – carries risks as predict adverse and toxic effects in humans, and which
well as benefits, and the aim of drug safety assessment is form an important part of the data submitted to the regu-
to ensure, as far as possible, that the risks are commensu- latory authorites when approval is sought (a) for the new
rate with the benefits. compound to be administered to humans for the first time
Safety is addressed at all stages in the life history of a (IND approval in the USA; see Chapter 20), and (b) for
drug, from the earliest stages of design, through pre- permission to market the drug (NDA approval in the USA,
clinical investigations (discussed in this chapter) and MAA approval in Europe; see Chapter 20).
preregistration clinical trials (Chapter 17), to the entire The programme of preclinical safety assessment for a
post-marketing history of the drug. The ultimate test new synthetic compound can be divided into the follow-
comes only after the drug has been marketed and used in ing main chronological phases, linked to the clinical trials
a clinical setting in many thousands of patients, during the programme (Figure 15.1):
period of Phase IV clinical trials (post-marketing surveil- • Exploratory toxicology, aimed at giving a rough
lance). It is unfortunately not uncommon for drugs to quantitative estimate of the toxicity of the
be withdrawn for safety reasons after being in clinical use compound when given acutely or repeatedly over a
for some time (for example practolol, because of a rare short period (normally 2 weeks), and providing an
but dangerous oculomucocutaneous reaction, troglitazone indication of the main organs and physiological
because of liver damage, cerivastatin because of skeletal systems involved. These studies provide information
muscle damage, terfenadine because of drug interactions, for the guidance of the project team in making
rofecoxib because of increased risk of heart attacks), reflect- further plans, but are not normally part of the
ing the fact that safety assessment is fallible. It always regulatory package that has to be approved before
will be fallible, because there are no bounds to what the drug can be given to humans, so they do not
may emerge as harmful effects. Can we be sure that drug need to be perfomed under good laboratory practice
X will not cause kidney damage in a particular inbred (GLP) conditions.
tribe in a remote part of the world? The answer is, of • Regulatory toxicology. These studies are performed to
course, ‘no’, any more than we could have been sure that GLP standards and comprise (a) those that are
various antipsychotic drugs – now withdrawn – would not required by regulatory authorities, or by ethics
cause sudden cardiac deaths through a hitherto unsus- committees, before the compound can be given for
pected mechanism, hERG channel block (see later). What the first time to humans; (b) studies required to

© 2012 Elsevier Ltd. 211

Section | 3 | Drug Development

IND NDA

Preclinical Clinical Clinical

Discovery
development Phase I/II Phase IV

Exploratory (non-GLP) toxicology

In vitro screens, e.g., Single and repeat dose range-

• In silco screens finding studies in 2 species
• Mutagenicity
• Cytotoxicity Regulatory (GLP) toxicology
• Immunotoxicity
• Hepatotoxicity Safety pharmacology 3–12 month chronic toxicity 24–month carcinogenicity
• Embryotoxicity Genotoxicity (in vitro and in 2 species in 2 species
in vivo) Reproductive toxicity in
28-day repeat dose toxicity 1 species, covering:
and recovery in 2 species • Fertility and implantation
• Fetal development
• Pre- and postnatal effects

Fig. 15.1  Timing of the main safety assessment studies during drug discovery and development.

support an application for marketing approval, company to anticipate and exclude any unwanted effects
which are normally performed in parallel with based on the specific chemistry, pharmacology and
clinical trials. Full reports of all studies of this kind intended therapeutic use of the compound in question.
are included in documentation submitted to the The regulatory authorities will often ask companies
regulatory authorities. developing second or third entrants in a new class to
Regulatory toxicology studies in group (a) include perform tests based on findings from the class leading
28-day repeated-dose toxicology studies in two species compound.
(including one non-rodent, usually dog but sometimes There are many types of new drug applications that do
monkey especially if the drug is a biological), in vitro and not fall into the standard category of synthetic small mol-
in vivo genoxocity tests, safety pharmacology and repro- ecules, where the safety assessment standards are different.
ductive toxicity assessment. In vitro genotoxicity tests, These include most biopharmaceuticals (see Chapter 12),
which are cheap and quick to perform, will often have as well as vaccines, cell and gene therapy products (see
been performed much earlier in the compound selection below). Drug combinations, and non-standard delivery
phase of the project, as may safety pharmacology studies. systems and routes of administration are also examples of
The nature of the tests in group (b) depends greatly on special cases where safety assessment requirements differ
the nature and intended use of the drug, but they will from those used for conventional drugs. These special
include chronic 3–12-month toxicological studies in two cases are not discussed in detail here; Gad (2002) gives a
or more species, long-term (18–24 months) carcinogenic- comprehensive account.
ity tests and reproductive toxicology, and often interaction
studies involving other drugs that are likely to be used for
the same indications.
TYPES OF ADVERSE DRUG EFFECT
The basic procedures for safety assessment of a single
new synthetic compound are fairly standard, although the
regulatory authorities have avoided applying a defined Adverse reactions in man are of four general types:
checklist of tests and criteria required for regulatory • Exaggerated pharmacological effects – sometimes
approval. Instead they issue guidance notes (available referred to as hyperpharmacology or mechanism-related
on relevant websites – USA: www.fda.gov/cder/; EU: effects – are dose related and in general predictable
www.emea.eu.int; international harmonization guide- on the basis of the principal pharmacological effect
lines: www.ich.org), but the onus is on the pharmaceutical of the drug. Examples include hypoglycaemia caused

212
 Assessing drug safety Chapter | 15 |

by antidiabetic drugs, hypokalaemia induced by about 1% of patients and was detected in early
diuretics, immunosuppression in response to clinical trials. The bone marrow effect, though
steroids, etc. potentially life-threatening, is reversible, and
• Pharmacological effects associated with targets other clozapine was successfully registered, with a
than the principal one – covered by the general term condition that patients receiving it must be
side effects or off-target effects. Examples include regularly monitored.)
hypotension produced by various antipsychotic drugs Safety pharmacology testing and dose range-finding studies
which block adrenoceptors as well as dopamine are designed to detect pharmacological adverse effects;
receptors (their principal target), and cardiac chronic toxicology testing is designed to detect dose-related
arrythmias associated with hERG-channel inhibition toxic effects, as well as the long-term consequences of
(see below). Many drugs inhibit one or more forms pharmacological side effects; idiosyncratic reactions may
of cytochrome P450, and hence affect the be revealed in Phase III clinical trials, but are likely to
metabolism of other drugs. Provided the remain undetected until the compound enters clinical use.
pharmacological profile of the compound is known
in sufficient detail, effects of this kind are also
predictable (see also Chapter 10).
• Dose-related toxic effects that are unrelated to the SAFETY PHARMACOLOGY
intended pharmacological effects of the drug.
Commonly such effects, which include toxic effects The pharmacological studies described in Chapter 11 are
on liver, kidney, endocrine glands, immune cells and exploratory (i.e. surveying the effects of the compound
other systems, are produced not by the parent drug, with respect to selectivity against a wide range of possible
but by chemically reactive metabolites. Examples targets) or hypothesis driven (checking whether the
include the gum hyperplasia produced by the expected effects of the drug, based on its target selectivity,
antiepileptic drug phenytoin, hearing loss caused are actually produced). In contrast, safety pharmacology
by aminoglycoside antibiotics, and peripheral comprises a series of protocol-driven studies, aimed spe-
neuropathy caused by thalidomide1. Genotoxicity cifically at detecting possible undesirable or dangerous
and reproductive toxicity (see below) also fall into effects of exposure to the drug in therapeutic doses (see
this category. Such adverse effects are not, in general, ICH Guideline S7A). The emphasis is on acute effects
predictable from the pharmacological profile of the produced by single-dose administration, as distinct from
compound. It is well known that certain chemical toxicology studies, which focus mainly on the effects of
structures are associated with toxicity, and so these chronic exposure. Safety pharmacology evaluation forms
will generally be eliminated early in the lead an important part of the dossier submitted to the regula-
identification stage, sometimes in silico before any tory authorities.
actual compounds are synthesized. The main ICH Guideline S7A defines a core battery of safety phar-
function of toxicological studies in drug macology tests, and a series of follow-up and supplementary
development is to detect dose-related toxic effects of tests (Table 15.1). The core battery is normally performed
an unpredictable nature. on all compounds intended for systemic use. Where they
• Rare, and sometimes serious, adverse effects, known are not appropriate (e.g. for preparations given topically)
as idiosyncratic reactions, that occur in certain their omission has to be justified on the basis of informa-
individuals and are not dose related. Many examples tion about the extent of systemic exposure that may occur
have come to light among drugs that have entered when the drug is given by the intended route. Follow-up
clinical use, e.g. aplastic anaemia produced by studies are required if the core battery of tests reveals
chloramphenicol, anaphylactic responses to penicillin, effects whose mechanism needs to be determined. Sup-
oculomucocutaneous syndrome with practolol, bone plementary tests need to be performed if the known chem-
marrow depression with clozapine. Toxicological tests istry or pharmacology of the compound gives any reason
in animals rarely reveal such effects, and because to expect that it may produce side effects (e.g. a compound
they may occur in only one in several thousand with a thiazide-like structure should be tested for possible
humans they are likely to remain undetected in inhibition of insulin secretion, this being a known side
clinical trials, coming to light only after the drug has effect of thiazide diuretics; similarly, an opioid needs to
been registered and given to thousands of patients. be tested for dependence liability and effects on gastroin-
(The reaction to clozapine is an exception. It affects testinal motility). Where there is a likelihood of significant
drug interactions, this may also need to be tested as part
of the supplementary programme.
1
This notorious drug is undergoing a therapeutic revival in the The core battery of tests listed in Table 15.1 focuses on
treatment of myeloma, a serious type of bone marrow cancer,
progressive muscle weakness and sensory loss due to peripheral acute effects on cardiovascular, respiratory and nervous
neuropathy being the most common and troublesome side effects. systems, based on standard physiological measurements.

213
Section | 3 | Drug Development

Table 15.1  Safety pharmacology

Type Physiological system Tests

Core battery Central nervous system Observations on conscious animals
Motor activity
Behavioural changes
Coordination
Reflex responses
Body temperature
Cardiovascular system Measurements on anaesthetized animals
Blood pressure
Heart rate
ECG changes
Tests for delayed ventricular
repolarization (see text)
Respiratory system Measurements on anaesthetized or
conscious animals
Respiratory rate
Tidal volume
Arterial oxygen saturation
Follow-up tests (examples) Central nervous system Tests on learning and memory
More complex test for changes in
behaviour and motor function
Tests for visual and auditory function
Cardiovascular system Cardiac output
Ventricular contractility
Vascular resistance
Regional blood flow
Respiratory system Airways resistance and compliance
Pulmonary arterial pressure
Blood gases
Supplementary tests (examples) Renal function Urine volume, osmolality pH
Proteinuria
Blood urea/creatinine
Fluid/electrolyte balance
Urine cytology
Autonomic nervous system Cardiovascular, gastrointestinal and
respiratory system responses to agonists
and stimulation of autonomic nerves
Gastrointestinal system Gastric secretion

214
 Assessing drug safety Chapter | 15 |

Table 15.1  Continued

Type Physiological system Tests

Gastric pH
Intestinal motility
Gastrointestinal transit time
Other systems (e.g. endocrine, blood Tests designed to detect likely acute
coagulation, skeletal muscle function, etc.) effects

The follow-up and supplementary tests are less clearly on ferrets or guinea pigs, as well as larger mammalian
defined, and the list given in Table 15.1 is neither prescrip- species, such as dog, rabbit, pig or monkey, in which
tive nor complete. It is the responsibility of the team to hERG-like channels control ventricular repolarization,
decide what tests are relevant and how the studies should rather than in rat and mouse. In vivo tests for proarrhyth-
be performed, and to justify these decisions in the submis- mic effects in various species are being developed (De
sion to the regulatory authority. Clerck et al., 2002), but have not yet been evaluated for
regulatory purposes.
Because of the importance of drug-induced QT prolon-
Tests for QT interval prolongation gation in man, and the fact that many diverse groups of
The ability of a number of therapeutically used drugs to drugs appear to have this property, there is a need for
cause a potentially fatal ventricular arrhythmia (’torsade de high-throughput screening for hERG channel inhibition
pointes’) has been a cause of major concern to clinicians to be incorporated early in a drug discovery project. The
and regulatory authorities (see Committee for Proprietary above methods are not suitable for high-throughput
Medicinal Products, 1997; Haverkamp et al., 2000). The screening, but alternative methods, such as inhibition of
arrhythmia is associated with prolongation of the ven- binding of labelled dofetilide (a potent hERG-channel
tricular action potential (delayed ventricular repolariza- blocker), or fluorimetric membrane potential assays on
tion), reflected in ECG recordings as prolongation of the cell lines expressing these channels, can be used in high-
QT interval. Drugs known to possess this serious risk, throughput formats, as increasingly can automated patch
many of which have been withdrawn, include several tri- clamp studies (see Chapter 8). It is important to note that
cyclic antidepressants, some antipsychotic drugs (e.g. thiori- binding and fluorescence assays are not seen as adequately
dazine, droperidol), antidysrhythmic drugs (e.g. amiodarone, predictive and cannot replace the patch clamp studies
quinidine, disopyramide), antihistamines (terfenadine, astem- under the guidelines (ICH 7B). These assays are now
izole) and certain antimalarial drugs (e.g. halofantrine). The becoming widely used as part of screening before selecting
main mechanism responsible appears to be inhibition of a clinical candidate molecule, though there is still a need
a potassium channel, termed the hERG channel, which to confirm presence or absence of QT prolongation in
plays a major role in terminating the ventricular action functional in vivo tests before advancing a compound into
potential (Netzer et al., 2001). clinical development.
Screening tests have shown that QT interval prolonga-
tion is a common property of ‘drug-like’ small molecules,
and the patterns of structure–activity relationships have
EXPLORATORY (DOSE RANGE-
revealed particular chemical classes associated with this
effect. Ideally, these are taken into account and avoided at FINDING) TOXICOLOGY STUDIES
an early stage in drug design, but the need remains for
functional testing of all candidate drug molecules as a The first stage of toxicological evaluation usually takes the
prelude to tests in humans. form of a dose range-finding study in a rodent and/or a non-
Proposed standard tests for QT interval prolongation rodent species. The species commonly used in toxicology
have been formulated as ICH Guideline S7B. They com- are mice, rats, guinea pigs, hamsters, rabbits, dogs, mini-
prise (a) testing for inhibition of hERG channel currents pigs and non-human primates. Usually two species (rat
in cell lines engineered to express the hERG gene; (b) and mouse) are tested initially, but others may be used if
measurements of action potential duration in myocardial there are reasons for thinking that the drug may exert
cells from different parts of the heart in different species; species-specific effects. A single dose is given to each test
and (c) measurements of QT interval in ECG recordings animal, preferably by the intended route of administration
in conscious animals. These studies are usually carried out in the clinic, and in a formulation shown by previous

215
Section | 3 | Drug Development

pharmacokinetic studies to produce satisfactory absorp- As mentioned above, these toxicology studies are pre-
tion and duration of action (see also Chapter 10). Gener- liminary, and usually they are not sufficient for supporting
ally, widely spaced doses (e.g. 10, 100, 1000 mg/kg) will the first human evaluation of the medicine. Very often,
be tested first, on groups of three to four rodents, and the they are not conducted under good laboratory practice
animals will be observed over 14 days for obvious signs (GLP) conditions, nor are they conducted with test mate-
of toxicity. Alternatively, a dose escalation protocol may be rial which has been produced according to good manufac-
used, in which each animal is treated with increasing turing practice (GMP).
doses of the drug at intervals (e.g. every 2 days) until signs Subsequent work is guided by regulatory requirements
of toxicity appear, or until a dose of 2000 mg/kg is elaborated by the International Conference of Harmoniza-
reached. With either protocol, the animals are killed at the tion or by national regulatory authorities. Their published
end of the experiment and autopsied to determine if any guidelines specify/recommend the type of toxicological
target organs are grossly affected. The results of such dose evaluation needed to support applications for carrying out
range-finding studies provide a rough estimate of the studies in humans. These documents provide guidance, for
no toxic effect level (NTEL, see Toxicity measures, below) in example, on the duration of administration, on the design
the species tested, and the nature of the gross effects seen of the toxicological study, including the number of animals
is often a useful pointer to the main target tissues and to be studied, and stipulate that the work must be carried
organs. out under GLP conditions and that the test material must
The dose range-finding study will normally be followed be of GMP standard. The test substance for the toxicology
by a more detailed single-dose toxicological study in two evaluation has to be identical in terms of quality and
or more species, the doses tested being chosen to span the characteristics to the substance given to humans.
estimated NTEL. Usually four or five doses will be tested,
ranging from a dose in the expected therapeutic range to
doses well above the estimated NTEL. A typical protocol
GENOTOXICITY
for such an acute toxicity study is shown in Figure 15.2.
The data collected consist of regular systematic assessment
of the animals for a range of clinical signs on the basis of Foreign substances can affect gene function in various
a standardized checklist, together with gross autopsy find- ways, the two most important types of mechanism in rela-
ings of animals dying during a 2-week observation period, tion to toxicology being:
or killed at the end. The main signs that are monitored are • Mutagenicity, i.e. chemical alteration of DNA
shown in Table 15.2. sufficient to cause abnormal gene expression in the
Single-dose studies are followed by a multiple dose- affected cell and its offspring. Most commonly, the
ranging study in which the drug is given daily or twice mutation arises as a result of covalent modification
daily, normally for 2 weeks, with the same observation and of individual bases (point mutations). The result
autopsy procedure as in the single-dose study, in order to may be the production of an abnormal protein if the
give preliminary information about the toxicity after mutation occurs in the coding region of the gene, or
chronic treatment. altered expression levels of a normal protein if the
The results of these preliminary in vivo toxicity studies mutation affects control sequences. Such mutations
will help in the planning and design of the next steps of occur continuously in everyday life, and are
the development programme; they will also help to decide counteracted more or less effectively by a variety of
whether or not it is worthwhile to continue the research DNA repair mechanisms. They are important
effort on a given chemical class. particularly because certain mutations can interfere
with mechanisms controlling cell division, and
thereby lead to malignancy or, in the immature
organism, to adverse effects on growth and
Test dose development. In practice, most carcinogens are
mutagens, though by no means all mutagens are
8 days 14 days carcinogens. Evidence of mutagenicity therefore
Acclimatization Test period sounds a warning of possible carcinogenicity, which
must be tested by thorough in vivo tests.
• Chromosomal damage, for example chromosome
Control Daily observation for Animals breakage (clastogenesis), chromosome fusion,
observations weight, mortality killed for translocation of stretches of DNA within or
and morbidity gross between chromosomes, replication or deletion
Autopsy of dead animals autopsy of chromosomes, etc. Such changes result from
alterations in DNA, more extensive than point
Fig. 15.2  Typical protocal for single-dose toxicity study. mutations and less well understood mechanistically;

216
 Assessing drug safety Chapter | 15 |

Table 15.2  Clinical observations in acute toxicity tests

System Observation Signs of toxicity

Nervous system Behaviour Sedation
Restlessness
Aggression
Motor function Twitch
Tremor
Ataxia
Catatonia
Convulsions
Muscle rigidity or flaccidity
Sensory function Excessive or diminished response to stimuli
Respiratory Respiration Increased or decreased respiratory rate
Intermittent respiration
Dyspnoea
Cardiovascular Cardiac palpation Increase or decrease in rate or force
?Electrocardiography Disturbances of rhythm.
Altered ECG pattern (e.g. QT prolongation)
Gastrointestinal Faeces Diarrhoea or constipation
Abnormal form or colour
Bleeding
Abdomen Spasm or tenderness
Genitourinary Genitalia Swelling, inflammation, discharge, bleeding
Skin and fur Discoloration
Lesions
Piloerection
Mouth Discharge
Congestion
Bleeding
Eye Pupil size Mydriasis or miosis
Eyelids Ptosis, exophthalmos
Movements Nystagmus
Cornea Opacity
General signs Body weight Weight loss
Body temperature Increase or decrease

217
Section | 3 | Drug Development

they have a similar propensity to cause cancerous positive results with non-carcinogenic substances.
changes and to affect growth and development. These tests are not possible with compounds that are
The most important end results of genotoxicity – inherently toxic to mammalian cells.
carcinogenesis and impairment of fetal development (ter- • An in vivo test for chromosomal damage in rodent
atogenicity) – can only be detected by long-term animal haemopoietic cells. The mouse micronucleus test is
studies. There is, therefore, every reason to pre-screen com- commonly used. Animals are treated with the test
pounds by in vitro methods, and such studies are routinely compound for 2 days, after which immature
carried out before human studies begin. Because in many erythrocytes in bone marrow are examined for
cases the genotoxicity is due to reactive metabolites rather micronuclei, representing fragments of damaged
than to the parent molecule, the in vitro tests generally chromosomes.
include assays carried out in the presence of liver micro- If these three tests prove negative, no further tests of
somes or other liver-derived preparations, so that metabo- genotoxicity are generally needed before the compound
lites are generated. Often, liver microsomes from rats can be tested in humans. If one or more is positive, further
treated with inducing agents (e.g. a mixture of chlorinated in vitro and in vivo genotoxicity testing will usually be
biphenyls known as Arochlor 1254) are used, in order to carried out to assess more accurately the magnitude of the
enhance drug metabolizing activity. risk. In a few cases, where the medical need is great and
the life expectancy of the patient population is very limited,
development of compounds that are clearly genotoxic –
Selection and interpretation of tests and, by inference, possibly carcinogenic – may still be justi-
fied, but in most cases genotoxic compounds will be
Many in vitro and in vivo test systems for mutagenicity
abandoned without further ado. An updated version of the
have been described, based on bacteria, yeast, insect and
guideline is open for consultation at the time of writing.
mammalian cells (see Gad, 2002 for details). The ICH
Whereas early toxicological evaluation encompasses
Guidelines S2A and S2B, stipulate a preliminary battery of
acute and subacute types of study, the full safety assesse-
three tests:
ment is based on subchronic and chronic studies. The
• Ames test: a test of mutagenicity in bacteria. The basis focus is more on the harmful effects of long-term exposure
of the assay is that mutagenic substances increase to ‘low’ doses of the agent. The chronic toxic effects can
the rate at which a histidine-dependent strain of be very different from the acute toxic effects, and could
Salmonella typhimurium reverts to a wild type that can accumulate over time. Three main categories of toxicologi-
grow in the absence of histidine. An increase in the cal study are required according to the regulatory guide-
number of colonies surviving in the absence of lines, namely chronic toxicity special tests and toxicokinetic
histidine therefore denotes significant mutagenic analysis.
activity. Several histidine-dependent strains of the
organism which differ in their susceptibility to
particular types of mutagen are normally tested in
parallel. Positive controls with known mutagens that CHRONIC TOXICOLOGY STUDIES
act directly or only after metabolic activation are
routinely included when such tests are performed. The object of these studies is to look for toxicities that
• An in vitro test for chromosomal abnormalities in appear after repetitive dosing of the compound, when a
mammalian cells or an in vitro mouse lymphoma tk cell steady state is achieved, i.e. when the rate of drug admin-
assay. To test chromosomal damage Chinese hamster istration equals the rate of elimination. In long-term toxic-
ovary (CHO) cells are grown in culture in the ity studies three or more dose levels are tested, in addition
presence of the test substance, with or without liver to a vehicle control. The doses will include one that is
microsomes. Cell division is arrested in metaphase, clearly toxic, one in the therapeutic range, and at least one
and chromosomes are observed microscopically in between. Ideally, the in-between doses will exceed the
to detect structural aberrations, such as gaps, expected clinical dose by a factor of 10 for rodents and 5
duplications, fusions or alterations in chromosome for non-rodents, yet lack overt toxicity, this being the
number. The mouse lymphoma cell test for ‘window’ normally required by regulatory authorities. At
mutagenicity involves a heterozygous (tk +/−) cell least one recovery group is usually included, i.e. animals
line that can be killed by the cytotoxic agent BrDU. treated with the drug at a toxic level and then allowed to
Mutation to the tk −/−) form causes the cells to recover for 2–4 weeks so that the reversibility of the
become resistant to BrDU, and so counts of changes observed can be assessed. The aim of these studies
surviving cells in cutures treated with the test is to determine (a) the cumulative biological effects pro-
compound provide an index of mutagenicity. The duced by the compound, and (b) at what exposure level
mouse lymphoma cell mutation assay is more (see below) they appear. The initial repeated-dose studies,
sensitive than the chromosomal assay, but can give by revealing the overall pattern of toxic effects produced,

218
 Assessing drug safety Chapter | 15 |

also give pointers to particular aspects (e.g. liver toxicity, Regulatory authorities stipulate the duration of repeat-
bone marrow depression) that may need to be addressed dose toxicity testing required before the start of clinical
later in more detailed toxicological investigations. testing, the ICH recommendations being summarized in
The standard procedures discussed here are appropriate Table 20.1. These requirements can be relaxed for drugs
for the majority of conventional synthetic compounds aimed at life-threatening diseases, the spur for this change
intended for systemic use. Where drugs are intended only in attitude coming from the urgent need for anti-HIV
for topical use (skin creams, eye drops, aerosols, etc.) the drugs during the 1980s. In such special cases chronic toxic-
test procedures are modified accordingly. Special consid- ity testing is still required, but approval for clinical trials
erations apply also to biopharmaceutical preparations, – and indeed for marketing – may be granted before they
and these are discussed briefly below. have been completed.

Experimental design Evaluation of toxic effects

Chronic toxicity testing must be performed under GLP During the course of the experiment all animals are
conditions, with the formulation and route of administra- inspected regularly for mortality and gross morbidity,
tion to be used in humans. Tests are normally carried out severely affected animals being killed, and all dead animals
on one rodent (usually rat) and one non-rodent species subjected to autopsy. Specific signs (e.g. diarrhoea, saliva-
(usually dog), but additional species (such as monkey or tion, respiratory changes, etc.) are assessed against a
pig) may be tested if there are special reasons for suspect- detailed checklist and specific examinations (e.g. blood
ing that their responses may predict effects in man more pressure, heart rate and ECG, ocular changes, neurological
accurately than those of dogs. Pharmacokinetic measure- and behavioural changes, etc.) are also conducted regu-
ments are included, so that extrapolation to humans can larly. Food intake and body weight changes are monitored,
be done on the basis of the concentration of the drug in and blood and urine samples collected at intervals for
blood and tissues, allowing for differences in pharmacoki- biochemical and haematological analysis.
netics between laboratory animals and humans (see At the end of the experiment all animals are killed and
Chapter 10). Separate male and female test groups are examined by autopsy for gross changes, samples of all
used. The recommended number of animals per test group major tissues being prepared for histological examination.
(see Gad, 2002) is shown in Table 15.3. There are strong Tissues from the high-dose group are examined histologi-
reasons for minimizing the number of animals used in cally and any tissues showing pathological changes are
such studies, and the figures given take this into account, also examined in the low-dose groups to enable a dose
representing the minimum shown by experience to be threshold to be estimated. The reversibility of the adverse
needed for statistically reliable conclusions to be drawn. effects can be evaluated in these studies by studying
An average study will require 200 or more rats and about animals retained after the end of the dosing period.
60 dogs, dosed with compound and observed regularly, As part of the general toxicology screening described
including blood and urine sampling, for several months above, specific evaluation of possible immunotoxicity is
before being killed and autopsied, with tissue samples required by the regulatory authorities, the main concern
being collected for histological examination. It is a massive being immunosuppression. If effects on blood, spleen or
and costly experiment requiring large amounts of com- thymus cells are observed in the 1-month repeated-dose
pound prepared to GMP standards, the conduct and studies, additional tests are required to evaluate the
results of which will receive detailed scrutiny by regulatory strength of cellular and humoral responses in immunized
authorities, and so careful planning and scrupulous execu- animals. A battery of suitable tests is included in the FDA
tion are essential. Guidance note (2002).

Table 15.3  Recommended numbers of animals for chronic toxicity studies (Gad, 2002)

Study duration Rats (per sex) Dogs (per sex) Primates (per sex)
4 weeks 5–10 3–4 3
3 months 20 6 5
6 months 30 8 5
12 months 50 10 10
2-year (carcinogenicity) 50–80

219
Section | 3 | Drug Development

The large body of data collected during a typical toxicol- proliferation); and (3) specific tests for immunogenicity,
ogy study should allow conclusions to be drawn about the expressed as sensitization, or the development of neutral-
main physiological systems and target organs that underlie izing antibodies.
the toxicity of the compound, and also about the dose Newer types of biopharmaceuticals, such as DNA-, RNA-
levels at which critical effects are produced. In practice, the and cell-based therapies (see Chapter 3), pose special
analysis and interpretation are not always straightforward, questions in relation to safety assessment (see FDA Guid-
for a variety of reasons, including: ance Note, 1998). Particular concerns arise from the use
• Incorrect choice of doses of genetically engineered viral vectors in gene therapy,
• Variability within the groups of animals which can induce immune responses and in some cases
• Spontaneous occurrence of ‘toxic’ effects in control retain potential infectivity. The possibility that genetically
or vehicle-treated animals engineered foreign cells might undergo malignant trans-
• Missing data, owing to operator error, equipment formation is a further worrisome risk. These specialized
failure, unexpected death of animals, etc. topics are not discussed further here.
• Problems of statistical analysis (qualitative data,
multiple comparisons, etc.).
Overall, it is estimated that correctly performed chronic SPECIAL TESTS
toxicity tests in animals successfully predict 70% of toxic
reactions in humans (Olson et al., 2000); skin reactions In addition to the the general toxicological risks addressed
in humans are the least well predicted. by the testing programme discussed above, the risks of
carcinogenicity and effects on reproduction (particularly
on fertility and on pre- and postnatal development) may
Biopharmaceuticals be of particular concern, requiring special tests to be
Biopharmaceuticals now constitute more than 25% of performed.
new drugs being approved. For the most part they are
proteins made by recombinant DNA technology, or mono­ Carcinogenicity testing
clonal antibodies produced in cell culture. As a rule, pro-
Carcinogenicity testing is normally required before a com-
teins tend to be less toxic than synthetic compounds,
pound can be marketed – though not before the start of
mainly because they are normally metabolized to smaller
clinical trials – if the drug is likely to be used in treatment
peptides and amino acids, rather than to the reactive com-
continuously for 6 months or more, or intermittently for
pounds formed from many synthetic small molecule com-
long periods. It is also required if there are special causes
pounds, which are the cause of most types of drug toxicity,
for concern, for example if:
especially genotoxicity. Biopharmaceuticals are, therefore,
generally less toxic than synthetic compounds. Their • the compound belongs to a known class of
unwanted effects are associated mainly with ‘hyperphar- carcinogens, or has chemical features associated with
macology’ (see above), lack of pharmacological specificity carcinogenicity; nowadays such compounds will
or immunogenicity. Many protein therapeutics are highly normally have been eliminated at the lead
species specific and this often means that safety studies identification stage (see Chapter 9)
have to be performed in monkeys or that a rodent-selective • chronic toxicity studies show evidence of
analogue has to be made. This applies to biological media- precancerous changes
tors, such as hormones, growth factors, cytokines, etc., as • the compound or its metabolites are retained in
well as to monoclonal antibodies. The presence of impuri- tissues for long periods.
ties in biopharmaceutical preparations has an important If a compound proves positive in tests of mutagenicity
bearing on safety assessment, as unwanted proteins or (see above), it must be assumed to be carcinogenic and its
other cell constituents are often present as contaminants, use restricted accordingly, so no purpose is served by in
and may vary from batch to batch. Quality control presents vivo carcinogenicity testing. Only in very exceptional cases
more problems than with conventional chemical prod- will such a compound be chosen for development.
ucts, and is particularly critical for biopharmaceuticals. ICH Guidelines S1A, S1B and S1C on carcinogenicity
The safety assessment of new biopharmaceuticals is dis- testing stipulate one long-term test in a rodent species
cussed in ICH Guidance Note S6 and the Addendum (usually rat), plus one other in vivo test, which may be
issued in June 2011. The main aspects that differ from the either (a) a short-term test designed to show high sensitiv-
guidelines relating to conventional drugs are: (1) careful ity to carcinogens (e.g. transgenic mouse models) or to
attention to choice of appropriate species; (2) no need for detect early events associated with tumour initiation or
routine genotoxicity and carcinogenicity testing (though promotion; or (b) a long-term carcinogenicity test in a
carcinogenicity testing may be required in the case of second rodent species (normally mouse). If positive results
mediators, such as growth factors, which may regulate cell emerge in either study, the onus is on the pharmaceutical

220
 Assessing drug safety Chapter | 15 |

company to provide evidence that carcinogenicity will not disaster of the 1960s. The second was the high incidence
be a significant risk to humans in a therapeutic setting. of cervical and vaginal cancers in young women whose
Until recently, the normal requirement was for long-term mothers had been treated with diethylstilbestrol (DES) in
studies in two rodent species, but advances in the under- early pregnancy with the aim of preventing early abortion.
standing of tumour biology and the availability of new DES was used in this way between 1940 and 1970, and
models that allow quicker evaluation have brought about the cancer incidence was reported in 1971. These events
a change in the attitude of regulatory authorities such that led to the introduction of stringent tests for teratogenicity
only one long-term study is required, together with data as a prerequisite for the approval of new drugs, and from
from a well-validated short-term study. this flowed concern for other aspects of reproductive toxi-
Long-term rat carcinogenicity studies normally last for cology which now must be fully evaluated before a drug
2 years and are run in parallel with Phase III clinical trials. is marketed. Current requirements are summarized in ICH
(Oral contraceptives require a 3-year test for carcinogenic- Guideline S5A.
ity in beagles.) Drugs can affect reproductive performance in three
Three or four dose levels are tested, plus controls. Typi- main ways:
cally, the lowest dose tested is close to the maximum
recommended human dose, and the highest is the maxi­
• Fertility (both sexes, fertilization and implantation),
addressed by Segment 1 studies
mum tolerated dose (MTD) in rats (i.e. the largest dose
that causes no obvious side effects or toxicity in the
• Embryonic and fetal development or teratology,
addressed by Segment 2 studies
chronic toxicity tests). Between 50 and 80 animals of each
sex are used in each experimental group (see Table 15.3),
• Peri- and postnatal development, addressed by
Segment 3 studies.
and so the complete study will require about 600–800
animals. Premature deaths are inevitable in such a large It is usually acceptable for Phase I human studies on
group and can easily ruin the study, so that housing the male volunteers to begin before any reproductive toxicol-
animals under standard disease-free conditions is essen- ogy data are available, so long as the drug shows no evi-
tial. At the end of the experiment, samples of about 50 dence of testicular damage in 2- or 4-week repeated-dose
different tissues are prepared for histological examination, studies. The requirement for reproductive toxicology data
and rated for benign and malignant tumour formation by as a prelude to clinical trials differs from country to
experienced pathologists. Carcinogenicity testing is there- country, but, as a general rule, clinical trials involving
fore one of the most expensive and time-consuming women of childbearing age should be preceded by rele-
components of the toxicological evaluation of a new com- vant reproductive toxicology testing. In all but exceptional
pound. New guidelines for dose selection were adopted cases, such as drugs intended for treating life-threatening
from 2008 which allow a more rational approach to the diseases, or for use only in the elderly, registration will
selection of the high dose on the basis of a 25-fold higher require comprehensive data from relevant toxicology
exposure of the rodent than that seen in human subjects studies so that the reproductive risk can be assessed.
on the basis of AUC measurements in plasma rather than Segment 1 tests of fertility and implantation involve
MTD (Van der Laan, 2008). treating both males (for 28 days) and females (for 14 days)
Several transgenic mouse models have been developed with the drug prior to mating, then measuring sperm
which provide data more quickly (usually about 6 months) count and sperm viability, numbers of implantation sites
than the normal 2-year carcinogenicity study (Gad, 2002). and live and dead embryos on day 6 of gestation. For drugs
These include animals in which human proto-oncogenes, that either by design or by accident reduce fertility, tests
such as hRas, are expressed, or the tumour suppressor gene for reversibility on stopping treatment are necessary.
P53 is inactivated. These mice show a very high incidence Segment 2 tests of effects on embryonic and fetal devel-
of spontaneous tumours after about 1 year, but at 6 months opment are usually carried out on two or three species (rat,
spontaneous tumours are rare. Known carcinogens cause mouse, rabbit), the drug being given to the female during
tumours to develop in these animals within 6 months. the initial gestation period (day 6 to day 16 after mating
Advances in this area are occurring rapidly, and as they in the rat). Animals are killed just before parturition, and
do so the methodology for carcinogenicity testing is the embryos are counted and assessed for structural abnor-
expected to become more sophisticated and faster than malities. In vitro tests involving embryos maintained in
the conventional long-term studies used hitherto (www. culture are also possible. The main stages of early embryo-
alttox.org/ttrc/toxicity-tests/carcinogenicity/). genesis can be observed in this way, and the effects of
drugs added to the medium can be monitored. Such in
Reproductive/developmental vitro tests are routinely performed in some laboratories,
but are currently not recognized by regulatory authorities
toxicology studies
as a reliable measure of possible teratogenicity.
Two incidents led to greatly increased concern about the Segment 3 tests on pre- and perinatal development entail
effects of drugs on the fetus. The first was the thalidomide dosing female rats with the drug throughout gestation and

221
Section | 3 | Drug Development

lactation. The offspring are observed for motility, reflex toxicology studies in specific areas. It must be remembered
responses, etc., both during and after the weaning period, that the regulatory authorities put the onus firmly on the
and at intervals some are killed for observations of struc- pharmaceutical company to present a dossier of data that
tural abnormalities. Some are normally allowed to mature covers all likely concerns about safety. Thus, where toxic
and are mated, to check for possible second-generation effects are observed in animals, evidence must be pre-
effects. Mature offspring are also tested for effects on learn- sented to show either that these are seen only at exposure
ing and memory. levels well outside the therapeutic range, or that they
Reproductive and developmental toxicology is a involve mechanisms that will not apply in humans. If the
complex field in which standards in relation to pharma- compound is observed to cause changes in circulating
ceuticals have not yet been clearly defined. The experimen- immune cells, or in lymphoid tissues, further tests for
tal studies are demanding, and the results may be immunosuppression will be needed, as well as studies to
complicated by species differences, individual variability determine the mechanism. Potential toxicological prob-
and ‘spontaneous’ events in control animals. lems, not necessarily revealed in basic toxicology testing,
It is obvious that any drug given in sufficient doses to must also be anticipated. Thus, if the compound is poten-
cause overt maternal toxicity is very likely to impair fetal tially immunogenic (i.e. it is a peptide or protein, or
development. Non-specific effects, most commonly a belongs to a known class of haptens), tests for sensitiza-
reduction in birthweight, are commonly found in animal tion will be required. For drugs that are intended for
studies, but provided the margin of safety is sufficient – say topical administration, local tissue reactions, including
10-fold – between the expected therapeutic dose and that allergic sensitization, need to be investigated. For some
affecting the fetus, this will not be a bar to developing the classes of drugs skin photosensitization will need to be
compound. The main aim of reproductive toxicology is to tested.
assess the risk of specific effects occurring within the thera- Interaction toxicology studies may be needed if the
peutic dose range in humans. Many familiar drugs and patients targeted for the treatment are likely to be taking
chemicals are teratogenic in certain species at high doses. another medicine whose efficacy or toxicity might be
They include penicillin, sulfonamides, tolbutamide, diphenyl- affected by the new therapeutic agent.
hydantoin, valproate, imipramine, acetazolamide, ACE inhibi- In summary, it is essential to plan the toxicology testing
tors and angiotensin antagonists, as well as many anticancer programme for each development compound on a case-
drugs and also caffeine, cannabis and ethanol. Many of these by-case basis. Although the regulatory authorities stipulate
are known or suspected teratogens in humans, and their the core battery of tests that need to be performed on every
use in pregnancy is to be avoided. A classification of drugs compound, it is up to the development team to anticipate
based on their safety during pregnancy has been devel- other safety issues that are likely to be of concern, and to
oped by the FDA (A, B, C, D or X). Category A is for drugs address them appropriately with experimental studies. In
considered safe in human pregnancy, that is, adequate and planning the safety assessment programme for a new drug,
well-controlled studies in pregnant women have failed to companies are strongly advised to consult the regulatory
demonstrate a risk to the fetus in any trimester of preg- authorities, who are very open to discussion at the plan-
nancy. Few drugs belong to this category. Category X is ning stage and can advise on a case-by-case basis.
reserved for drugs (e.g. isotretinoin, warfarin) that have been
proved to cause fetal abnormalities in man, and are there-
fore contraindicated in pregnancy. Category B covers drugs
with no evidence of risk in humans; category C covers TOXICOKINETICS
drugs in which a risk cannot be ruled out; and category D
covers drugs with positive evidence of risk. Requirements Toxicokinetics is defined in ICH Guideline S3A as ‘the
for the reproductive safety testing of biologicals are laid generation of pharmacokinetic data, either as an integral
out in the Addendum to S6 (2011) and there are specific component in the conduct of non-clinical toxicity studies,
considerations, e.g. the inability of some high-molecular- or in specially designed supportive studies, in order to
weight proteins to cross the placenta which do not apply assess systemic exposure’. In essence, this means pharma-
to small molecules. cokinetics applied to toxicological studies, and the meth-
odology and principles are no different from those of
conventional absorption, distribution, metabolism and
Other studies
excretion (ADME) studies. But whereas ADME studies
The focus of this chapter is on the core battery of tests address mainly drug and metabolite concentrations in dif-
routinely used in assessing drug safety at the preclinical ferent body compartments as a function of dose and time,
level. In practice, depending on the results obtained from toxicokinetics focuses on exposure. Although exposure is
these tests, and on the particular therapeutic application not precisely defined, the underlying idea is that toxic
and route of administration intended for the drug, it is effects often appear to be a function of both local concen-
nearly always necessary to go further with experimental tration and time, a low concentration persisting for a long

222
 Assessing drug safety Chapter | 15 |

time being as likely to evoke a reaction as a higher but

more transient concentration. Depending on the context, VARIABILITY IN RESPONSES
exposure can be represented as peak plasma or tissue con-
centration, or more often as average concentrations over a The response to a given dose of a drug is likely to vary
fixed period. As with conventional pharmacokinetic meas- when it is given to different individuals or even to the
urement, attention must be paid to plasma and tissue same individual on different occasions. Factors such as
binding of the drug and its metabolites, as bound mate- age, sex, disease state, degree of nutrition/malnutrition,
rial, which may comprise 98% or more of the measured co-administration of other drugs and genetic variations
concentration, will generally be pharmacologically and may influence drug response and toxicity. As elderly
toxicologically inert. The principles of toxicokinetics as people have reduced renal and hepatic function they may
applied to preclinical studies are discussed in detail by metabolize and excrete drugs more slowly, and, therefore,
Baldrick (2003). may require lower doses of medication than younger
Despite the difficulties of measuring and interpreting people. In addition, because of multiple illnesses elderly
exposure, toxicokinetic measurements are an essential part people often may be less able than younger adults to toler-
of all in vivo toxicological studies. Interspecies compari- ate minor side effects. Likewise, children cannot be
sons and extrapolation of animal data to humans are best regarded as undersized adults, and drug dosages relative
done on the basis of measured plasma and tissue concen- to body weight may be quite different. Drug distribution
trations, rather than administered dose, and regulatory is also different between premature infants and children.
authorities require this information to be provided. The dosages of drugs for children are usually calculated on
As mentioned earlier, the choice of dose for toxicity tests the basis of weight (mg/kg) or on the basis of body surface
is important, and dose-limiting toxicity should ideally be area (mg/m2). These important aspects of drug safety
reached in toxicology studies. This is the reason why the cannot be reliably assessed from preclinical data, but the
doses administered in these studies are always high, unless major variability factors need to be identified and
a maximum limit based on technical feasibility has been addressed as part of the clinical trials programme.
reached.

TOXICITY MEASURES CONCLUSIONS AND FUTURE TRENDS

Lethal dose (expressed as LD50, the estimated dose required No drug is completely non-toxic or safe2. Adverse effects
to kill 50% of a group of experimental animals) has been can range from minor reactions such as dizziness or skin
largely abandoned as a useful measure of toxicity, and no reactions, to serious and even fatal effects such as anaphy-
longer needs to be measured for new compounds. Meas- lactic reactions. The aims of preclinical toxicology are (a)
ures commonly used are: to reduce to a minimum the risk to the healthy volunteers
and the patients to whom the drug will be given in clinical
• no toxic effect level (NTEL), which is the largest dose trials; and (b) to ensure that the risk in patients treated
in the most sensitive species in a toxicology study of
with the drug once it is on the market is commensurate
a given duration which produced no observed toxic
with the benefits. The latter is also a major concern during
effect
clinical development, and beyond in Phase IV.
• no observed adverse effect level (NOAEL), which is the It is important to understand the margin of safety that
largest dose causing neither observed tissue toxicity
exists between the dose needed for the desired effect and
nor undesirable physiological effects, such as
the dose that produces unwanted and possibly dangerous
sedation, seizures or weight loss
side effects. But the extrapolation from animal toxicology
• maximum tolerated dose (MTD), which usually applies to safety in man is difficult because of the differences
to long-term studies and represents the largest dose
between species in terms of physiology, pathology and
tested that caused no obvious signs of ill-health
drug metabolism.
• no observed effect level (NOEL), which represents the Data from preclinical toxicity studies may be sufficiently
threshold for producing any observed
discouraging that the project is stopped at that stage. If the
pharmacological or toxic effect.
project goes ahead, the preclinical toxicology data provide
The estimated NTEL in the most sensitive species is a basis for determining starting doses and dosing regimens
normally used to determine the starting dose used in the
first human trials. The safety factor applied may vary from
100 to 1000, depending on the information available, the 2
Nor, of course, are any other everyday technologies, such as ladders,
type and severity of toxicities observed in animals, and kitchen knives, pots of paint or trains. None the less, public opinion
seems particularly sensitive to iatrogenic medical risks and is inclined
whether these anticipated toxicities can be monitored by to demand what is impossible, namely ‘proof that this drug/vaccine/
non-invasive techniques in man. procedure is 100% safe’.

223
Section | 3 | Drug Development

for the initial clinical trials, and for identifying likely target effects, such as liver damage (Nuwaysir et al., 1999;
organs and surrogate markers of potential toxicity in Pennie, 2000), so that detecting such a change
humans. Two particular trends in preclinical toxicology produced by a novel compound makes it likely that
are noteworthy: the compound will prove toxic, thereby ruling it out
• Regulatory requirements tend to become increasingly as a potential development candidate. Such high-
stringent and toxicology testing more complex as throughput genomics-based approaches enable large
new mechanisms of potential toxicity emerge. This databases of gene expression information to be built
has been one cause, over the years, of the steady up, covering a diverse range of chemical structures.
increase in the cost and duration of drug The expectation is that the structure–activity patterns
development (see Chapter 22). Only in the last 5–10 thus revealed will enable the prediction in silico of
years have serious efforts been made to counter this the likely toxicity of a wide range of hypothetical
trend, driven by the realization that innovation and compounds, enabling exclusion criteria to be applied
therapeutic advances are being seriously slowed very early in the discovery process. This field of
down, to the detriment, rather than the benefit, of endeavour, dubbed ‘toxicogenomics’, is expected by
human healthcare. The urgent need for effective many to revolutionize pharmaceutical toxicology
drugs against AIDS was the main impetus for this (Castle et al., 2002).
change. At present, extensive toxicity testing in vivo is required
• In an effort to reduce the time and cost of testing, by regulatory authorities, and data from in vitro studies
and to eliminate development compounds as early carry little weight. There is no likelihood that this will
as possible, early screening methods are being change in the near future (Snodin, 2002). What is clearly
increasingly developed and applied during the drug changing is the increasing use of in vitro toxicology screens
discovery phase of the project, with the aim of early in the course of a drug discovery programme to
reducing the probability of later failure. One such reduce the risk of toxicological failures later (see Chapter
approach is the use of cDNA microarray methods 10). New guidelines have recently (2009) been introduced
(see Chapter 7) to monitor changes in gene to avoid delays in the evaluation of anticancer drugs (S9
expression resulting from the application of the test ICH). The main impact of new technologies will therefore
compound to tissues or cells in culture. Over- or be – if the prophets are correct – to reduce the attrition
underexpression of certain genes is frequently rate in development, not necessarily to make drug devel-
associated with the occurrence of specific toxic opment faster or cheaper.

REFERENCES

Baldrick P. Toxicokinetics in preclinical FDA Guidance Note. Cell therapy ICH Guideline S1C: Dose selection
evaluation. Drug Discovery Today and gene therapy products. 1998. for carcinogenicity studies of
2003;8:127–33. http://www.fda.gov/cber/gdlns/ pharmaceuticals and S1C(R):
Castle AL, Carver MP, Mendrick DL. somgene.pdf. Addendum: addition of a limit dose
Toxicogenomics: a new revolution in FDA Guidance Note. Immunotoxicology and related notes. www.ich.org.
drug safety. Drug Discovery Today evaluation of investigational new ICH Guideline S2A: Genotoxicity:
2002;7:728–36. drugs. 2002. www.fda.gov/guidance/ guidance on specific aspects of
Committee for Proprietary Medicinal index.htm. regulatory tests for pharmaceuticals.
Products. Points to consider for the Gad SC. Drug safety evaluation. New www.ich.org.
assessment of the potential QT York: Wiley Interscience; 2002. ICH Guideline S2B: Genotoxicity: a
prolongation by non-cardiovascular Haverkamp W, Breitlandt G, Comm AJ, standard battery for genotoxicity
medicinal products. Publication et al. The potential for QT testing for pharmaceuticals.
CPMP 986/96. London: Human prolongation and proarrhythmias by www.ich.org.
Medicines Evaluation Unit; 1997. non-anti-arrhythmic drugs: clinical ICH Guideline S3A: Note for guidance
De Clerck, F, Van de Water A, D’Aubiol and regulatory implications. on toxicokinetics: the assessment of
J, et al. In vivo measurement of European Heart Journal systemic exposure in toxicity studies.
QT prolongation, dispersion and 2000;21:1232–7. www.ich.org.
arrhythmogenesis: application to ICH Guideline S1A: Guideline on the ICH Guideline S5A: Detection of
the preclinical cardiovascular need for carcinogenicity studies of toxicity to reproduction for
safety pharmacology of a new pharmaceuticals. www.ich.org. medicinal products. www.ich.org.
chemical entity. Fundamental ICH Guideline S1B: Testing for ICH Guideline S6: Preclinical safety
and Clinical Pharmacology 2002; carcinogenicity of pharmaceuticals. evaluation of biotechnology-derived
16:125–40. www.ich.org. pharmaceuticals. www.ich.org.

224
 Assessing drug safety Chapter | 15 |

ICH Guideline 6 – Addendum 2011 Netzer R, Ebneth E, Bischoff U, et al. Pennie WD. Use of cDNA microassays
www.ich.org. Screening lead compounds for QT to probe and understand the
ICH Guideline S7A: Safety interval prolongation. Drug toxicological consequences of altered
pharmacology studies for human Discovery Today 2001;6:78–84. gene expression. Toxicology Letters
pharmaceuticals. www.ich.org. Nuwaysir EF, Bittner M, Trent J, et al. 2000;112:473–7.
ICH Guideline S7B: Safety Microassays and toxicology; the Snodin DJ. An EU perspective on the
pharmacology studies for assessing advent of toxicogenomics. Molecular use of in vitro methods in regulatory
the potential for delayed ventricular Carcinogenesis 1999;24:152–9. pharmaceutical toxicology.
repolarization (QT interval Olson H, Betton G, Robinson D, et al. Toxicology Letters 2002;127:
prolongation) by human Concordance of the toxicity of 161–8.
pharmaceuticals. www.ich.org. pharmaceuticals in humans and Van der Laan JW. Dose selection for
ICH Guideline S9: Non clinical animals. Regulatory Toxicology carcinogenicity studies of
evaluation for anticancer and Pharmacology 2000;32: pharmaceuticals. 2008. www.ich.org.
pharmaceuticals. www.ich.org. 56–67.

225
 Chapter 16 
Pharmaceutical development
T Lundqvist, S Bredenberg

need to be documented. These investigations are termed

INTRODUCTION preformulation studies. Most synthetic drugs are either weak
bases (~75%) or weak acids (~20%), and will generally
Active pharmaceutical ingredients (API) in pharmaceutical need to be formulated as salts. Salts of a range of accept-
products must be formulated to dosage forms suitable for able conjugate acids or bases therefore need to be prepared
handling, distribution and adminstration to patients. and tested. Intravenous formulations of relatively insolu-
Common dosage forms are tablets and capsules, liquids ble compounds may need to include non-aqueous sol-
for injection, patches for transdermal administration and vents or emulsifying agents, and the compatibility of the
semisolids for dermal application. Development and doc- test substance with these additives, as well as with the
umentation of dosage forms is complex involving several commonly used excipients that are included in tablets or
stages and areas of expertise. In the preformulation phase capsules, will need to be assessed.
the chemical and physicochemical properties of the drug The main components of preformulation studies
substance are characterized. This information serves are:
together with information about the disease, intended • Development of a suitable spectroscopic assay
doses and preferred route of administration as the starting method for determining concentration and
point for the formulation development. Stability of the purity
drug substance is a cornerstone in the preformulation • Determination of solubility and dissolution rates of
phase. Analytical pharmaceutical chemistry is crucial for parent compound and salts in water and other
successful formulation development. Impurities and deg- solvents
radation products must be identified and quantified. Uni- • Chemical stability of parent compound and salts in
formity of content in unit dosage forms is vital to assure solution and solid state
that the patient is treated with correct doses. • Determination of pKa and pH dependence of
This chapter describes the main steps involved in phar- solubility and chemical stability
maceutical development and is illustrated with examples • Determination of lipophilicity (i.e. oil:water partition
of the most common formulations, tablets and capsules, coefficient, expressed as Kd)
since development of these dosage forms contains most • Determination of particle morphology, melting point
of the steps involved in formulation development per se. and suitability for milling
Development of solutions for injection/infusion or other • Characterizations of importance for the dosage forms
sterile dosage forms are not specifically discussed in this of choice, e.g. bulk density, powder flow, angle of
chapter. repose and compression properties.
Theoretical treatments of these molecular properties,
and laboratory methods for measuring them, which are
PREFORMULATION STUDIES beyond the scope of this book, are described in textbooks
such as Burger and Abraham (2003), Aulton (2007) and
As a starting point to developing dosage forms, various Allen (2008). Here we consider some issues that com-
physical and chemical properties of the drug substance monly arise in drug development.

© 2012 Elsevier Ltd. 227

Section | 3 | Drug Development

solubility of the compound, the diffusion coefficient of

Solubility and dissolution rate
the solute, the surface area of the boundary layer, and the
The question of solubility, already emphasized in Chap- geometry of the path leading from the boundary layer to
ters 9 and 10, is particularly important in relation to the the bulk phase.
development of pharmaceutical formulation. It is meas- In practice, dissolution rates depend mainly on:
ured by standard laboratory procedures and involves • Intrinsic solubility (since this determines the
determining the concentration of the compound in solu- boundary layer concentration)
tion after equilibration – usually after several hours of • Molecular weight (which determines diffusion
stirring – with the pure solid. In general, compounds coefficient)
whose aqueous solubility exceeds 10 mg/mL present no • Particle size and dispersion of the solute (which
problems with formulation (Kaplan, 1972). Compounds determine the surface area of the boundary layer and
with lower solubility are likely to require conversion to the length of the diffusion path).
salts, or the addition of non-aqueous solvents, in order to
In pharmaceutical development, dissolution rates are
achieve satisfactory oral absorption. Because the extreme
often manipulated intentionally by including different
pH values needed to induce ionization of very weak acids
polymers, such as methylcellulose into tablets or capsules,
or bases are likely to cause tissue damage, the inclusion of
to produce ‘slow-release’ formulations of drugs such as
a miscible solvent of relatively low polarity, such as 20%
diclofenac, allowing once-daily dosage despite the drug’s
propylene glycol or some other biocompatible solubiliz-
short plasma half-life.
ing agent (see below), will often be required for preparing
injectable formulations. Complications may arise with
oral formulations if the solubility is highly dependent on Stability
pH, because of the large pH difference between the
stomach and the small intestine. Gastric pH can range For routine use, a drug product is expected to have a shelf-
from near neutrality in the absence of any food stimulus life, representing less than 5% decomposition and no sig-
to acid secretion, to pH 1–2, whereas the intestinal pH is nificant physical change under normal storage conditions,
around 8. Basic substances that dissolve readily in the of at least 3 years.
stomach can therefore precipitate in the intestine and fail During the preformulation studies stability tests are
to be absorbed. Compounds that can exist in more than often carried out for 1–4 weeks. The chemical stability of
one crystal form can also show complex behaviours. The the solid is measured at temperatures ranging from 4 to
different lattice energies of molecules in the different 75°C, and moisture uptake at different relative humidities
crystal forms mean that the intrinsic solubility of the com- is also assessed.
pound is also different. Different crystal forms may cor- Measurements of stability in solution at pH values
respond to different hydration states of the compound, so ranging from 1 to 11 at room temperature and at 37°C
that a solution prepared from the unhydrated solid may will be performed, including formulations with solubiliz-
gradually precipitate as hydrated crystals. Selecting the best ing agents where appropriate. Sensitivity to UV and visible
salt form to avoid complications of this sort is an impor- light, and to exposure to oxygen, is also measured.
tant aspect of preformulation studies. The rate of degradation in short-term studies under
Compounds that have low intrinsic solubility in these harsh conditions is used to give a preliminary esti-
aqueous media can often be brought into solution by the mate of the likely rate of degradation under normal
addition of a water-miscible solubilizing agent, such as storage conditions. Sensitivity to low pH means that deg-
polysorbates, ethanol or polyethylene glycol (PEG). Pre- radation is likely to occur in the stomach, requiring meas-
formulation studies may, therefore, include the investiga- ures to prevent release of the compound until it reaches
tion of various solubilizing agents, such as methylcellulose the intestine.
or cyclodextrin, which are known to be relatively free of These preformulation stability tests serve mainly to
adverse effects in man. warn that further development of the compound may be
As well as intrinsic solubility, dissolution rate is important difficult or even impossible. Definitive tests of the long-
in determining the rate of absorption of an oral drug. The term (3 years or more) stability of the formulated prepara-
process of dissolution involves two steps: (a) the transfer tion will be required at a later stage of development for
of molecules from the solid to the immediately adjacent regulatory purposes.
layer of fluid, known as the boundary layer; and (b) escape
from the boundary layer into the main reservoir of fluid,
which is known as the bulk phase and is assumed to be well
Particle size and morphology
stirred so that its concentration is uniform. Step (a) is Ideally, for incorporation into a tablet or capsule the drug
invariably much faster than step (b), so the boundary layer substance needs to exist in small, uniformly sized parti-
quickly reaches saturation. The overall rate of dissolution cles, forming a smoothly flowing powder which can be
is limited by step (b), and depends on the intrinsic uniformly blended with the excipient material. Rarely, the

228
 Pharmaceutical development Chapter | 16 |

compound will emerge from the chemistry laboratory as

non-hygroscopic crystals, and the melting point will be
ROUTES OF ADMINISTRATION AND
sufficiently high that it can be reduced to a uniform fine
powder by mechanical milling. More often it will take the DOSAGE FORMS
form of an amorphous, somewhat waxy solid, and addi-
tional work will be needed later in development to With the knowledge of the characteristics of the drug sub-
produce it in a form suitable for incorporation into tablets stance from preformulation studies, the administration
or capsules. route has to be selected and the substance to be included
Preformulation studies are designed to reveal how far into a dosage form which is effective and convenient for
the available material falls short of this ideal. Various labo- the patient. The preferred dosage form for therapeutic
ratory methods are available for analysing particle size and agents is almost always an oral tablet or capsule, either
morphology, but in the preformulation stage simple taken as needed to control symptoms, or taken regularly
microscopic observation is the usual method. once or several times a day. However, there are many
Particles larger than a few micrometres, particularly if alternatives, and Table 16.1 lists some of the main ones.
the particle size is very variable, are difficult to handle and An important consideration is whether it is desirable to
mix uniformly. Hygroscopic materials and polymorphic achieve systemic exposure (i.e. distribution of the drug to all
crystal forms are a disadvantage, as already mentioned. organs via the bloodstream) or selective local exposure (e.g.
These issues are unlikely to matter greatly in the early to the lungs, skin or rectum) by applying the drug topi-
stages of development, as Phase I studies can usually be cally. In most cases systemic exposure will be required, and
carried out with liquid formulations if necessary, but can an oral capsule or tablet will be the desired final dosage
be a major hurdle later, so the main aim of the preformu- form. Even so, an intravenous formulation will normally
lation studies is to give a warning of likely problems be required for use in safety pharmacology, toxicology and
to come. pharmacokinetic studies in man.

Table 16.1  The main routes of administration and dosage forms

Exposure required Routes of Dosage forms

administration
Systemic Oral Tablet, capsule, solution, Liquid forms are particularly suitable for
suspension, emulsion children, and for patients unable to
swallow tablets. Unsuitable for
foul-tasting medicines
Parenteral
Injection (intravenous, Solution, emulsion, Examples: cytotoxic drugs liable to
subcutaneous, suspension, implant damage GI tract, drugs needed for
intramuscular) unconscious patients, drugs unstable in
Needle-free injection GI tract (e.g. peptide hormones)
Percutaneous Skin patches
Inhalation Gas, vapour Applicable mainly to anaesthetic agents
Intranasal Aerosol Used for some hormone preparations
that are not absorbed orally, e.g.
vasopressin analogues, gonadotropin-
releasing hormone
Topical Skin Ointment, cream, gel, aerosol
Respiratory tract Aerosol, inhaled powder
Rectum, vagina Suppository
Eye Solution, ointment

229
Section | 3 | Drug Development

The main routes of administration for drugs acting sys- rapidly than if they are given intravenously, possibly
temically, apart from oral and injectable formulations, are bypassing the blood–brain barrier by reaching the CNS
transdermal, intranasal and oromucosal. Rectal, vaginal and directly via the nerves to the olfactory bulb.
pulmonary routes are also used in some cases, though these Oromucosal delivery (Madhav et al., 2009) and espe-
are used mainly for drugs that act locally. cially utilizing the buccal and sublingual mucosa as the
Transdermal administration of drugs formulated as absorption site is a drug delivery route which promotes
small adhesive skin patches has considerable market rapid absorption and almost immediate pharmacological
appeal, even though such preparations are much more effect. The sublingual mucosa, especially, is highly vascu-
expensive than conventional formulations. To be admin- larized and this route bypasses the gastrointestinal tract
istered in this way, drugs must be highly potent, lipid and, thus, the first-pass metabolism. However, not all
soluble and of low molecular weight. Examples of com- drugs can be efficiently absorbed through the oral mucosa
mercially available patch formulations include nitroglyc- because of physicochemical properties or enzymatic
erin, scopolamine, fentanyl, nicotine, testosterone, estradiol breakdown of the drug and the amount of drug that could
and ketoprofen, and several others in development. The be absorbed is limited to a few milligrams. The drug has
main limitation is the low permeability of the skin to to be soluble, stable and able to easily permeate the
most drugs and the small area covered, which mean that mucosal barrier at the administration site. Also some for-
dosage is limited to a few milligrams per day, so only mulation aspects have to be taken into consideration,
very potent drugs can be given systemically via this route using a tablet formulation for a rapid onset of effect a
of administration. Variations in skin thickness affect the prerequisite is a fast disintegration and dissolution in the
rate of penetration, and the occurrence of local skin reac- oral cavity resulting in an optimal exposure of active sub-
tions is also a problem with some drugs. Various penetra- stance to the small volume of dissolving fluids. Swallow-
tion enhancers, mainly surfactant compounds of the sort ing of the drug could be a potential problem, but can be
discussed above, are used to improve transdermal absorp- minimized by using technologies using mucoadhesive
tion. The transfer rate can be greatly enhanced by apply- components (Bredenberg et al., 2003; Brown and Hamed,
ing a small and painless electric current (about 0.5  mA/ 2007). An optimized formulation and using this adminis-
cm2), and this is effective in achieving transfer of peptides tration route has the potential for very fast absorption
(e.g. calcitonin) and even insulin through the skin. It also (Kroboth et al., 1995) and obtaining peak blood levels
offers promise as a route of administration of oligonucleo­ within 10 to 15 minutes. It is thereby potentially a more
tides in gene therapy applications (see Chapter 12). These comfortable and convenient alternative to the intravenous
procedures are used for administration of nucleotides route of administration.
and in gene therapy and are being used experimentally
in the clinic, but are not yet available as commercial
products for routine clinical use. Ultrasonic irradiation
is also under investigation as a means of facilitating FORMULATION
transdermal delivery. These procedures would also allow
the administration to be controlled according to need.
Formulation of an active substance into a dosage form,
Intranasal drug administration (Illium, 2002, 2003) is
where there are no special requirements for modified
another route that has been used successfully for a few
release, involves a good deal of engineering. As already
drugs. The nasal epithelium is much more permeable
mentioned, the ‘ideal’ drug substance, intended for use as
than skin and allows the transfer of peptide drugs as well
an oral preparation, has the following characteristics:
as low-molecular-weight substances. Commercially avail-
able preparations have been developed for peptide hor- • Water solubility
mones, such as vasopressin analogues, calcitonin, buserelin • Chemical stability (including stability at low pH)
and others, as well as for conventional drugs such as • Permeability across the gastrointestinal epithelium
triptans, opioids, etc. The main disadvantages are that sub- • Good access to site of action (e.g. blood–brain
stances are quickly cleared from the nasal epithelium by barrier penetration, if intended to work in the brain)
ciliary action, as well as being metabolized, and the epi- • Resistance to first-pass metabolism.
thelial permeability is not sufficient to allow most pro- If these conditions are met the formulation of oral for-
teins to be given in this way. Ciliary clearance can be mulations presents no special problems, but they rarely
reduced by the use of gel formulations, and surfactant are, and it falls to the pharmaceutical development group
permeability enhancers can be used to improve the pen- to develop formulations that successfully overcome the
etration of larger molecules. The possibility of adminis- shortcomings of the compound. Firstly, the drug substance
tering insulin, growth factors or vaccines by this route is must be dried and converted to a powder form that can
the subject of active research efforts. Some studies have be precisely dispensed. Further, depending on the dosage
suggested (see Illium, 2003) that substances absorbed form and the desired properties, other substances called
through the nasal epithelium reach the brain more excipients, have to be included.

230
 Pharmaceutical development Chapter | 16 |

As mentioned above, tablets and capsules are the most years. The presence of moisture is the main contributor to
common dosage form (probably due to a combination of the degradation of the drug substance. Tablets normally
convenience for the patient and a cost-effective manufac- have a longer shelf-life than other formulations such as
turing process, at least for conventional tablets). Even for oral and parental liquids since they are a dry dosage form.
a tablet without special requirements for its drug release However, it is important to choose excipients which are
profile a range of excipients are needed, e.g. the tablet not hygroscopic since small amounts of moisture could
should be sufficiently strong to withstand handling but decrease the stability of the drug. Therefore, selection of
also has to disintegrate after intake in order to release the packing material is also an important aspect to take
the drug. into consideration: for example, moisture-sensitive freeze-
Inert diluents or fillers, such as lactose or starch, are dried tablets are often packed in almost impenetrable alu-
added to produce tablets of a manageable size (generally minium blisters.
50–500 mg). The filler should have good compactability In reality, formulation development has to take into
and flow properties, be non-hygroscopic and have accept- account not only the properties of the drug substance, but
able taste. Binders, such as cellulose and other polymeric also the desired delivery system and the form of the final
materials, may be needed to assist compaction into a solid product. In developing a new nitrate preparation for treat-
tablet that will not crumble. A binder could be added both ing angina, for example, the preferred delivery system
in dry form and in the granulation liquid depending on might be a skin patch to be packaged in a foil sachet, or
the manufacturing process. Disintegrating agents, such as a nasal spray to be packaged as a push-button aerosol can.
starch and cellulose, ensure that the tablet disintegrates Although a simple oral preparation may be feasible, the
rapidly in the gastrointestinal tract. For a very fast disinte- medical need and patient convenience might require the
gration, so-called super disintegrants acting to produce development of different dosage forms, and the develop-
extensive swelling could be used. Slippery, non-adherent ment plan would have to be directed towards this more
materials, such as magnesium stearate, may be needed to demanding task.
ensure that the powder runs smoothly in the tablet
machine, by reduction of friction between particles and
between particles and parts of the machine in contact. The
tablet may need to be coated with cellulose or sugar to PRINCIPLES OF DRUG  
disguise its taste. DELIVERY SYSTEMS
Capsule formulations are often used for initial clinical
trials, as they are generally simpler to develop than com-
pressed tablets, but are less suitable for controlled-release In recent years, drug delivery systems have become pro-
formulations (see below). A two-piece gelatin capsule can gressively more sophisticated, for three main reasons.
be used to contain drugs also in semisolid or liquid form. First, biopharmaceuticals represent an increasing propor-
Other advantages are that capsules are easy to swallow and tion of new drugs. They are very unlikely to conform to
provide effective taste-masking. Drug dissolution rate the ‘ideal’ profile summarized above, and so ingenuity in
could, with a fast dissolving capsule, be increased com- formulation is often needed to turn them into viable prod-
pared to a conventional tablet resulting in improved drug ucts. Second, there is increasing emphasis on selective
absorption, especially for poorly soluble substances. targeting of drugs to sites of disease via the use of special-
The choice of excipients and the manufacturing process ized delivery systems. This is particularly relevant for anti-
is very much dependent on the characteristics of drug cancer drugs. Third, controllable delivery systems are
substance and the desired properties of the dosage form. being developed for specific indications. The use of elec-
Drug release profile is just one aspect, while the homogen­ trophoresis to provide a steady flux of drug through the
eity or uniformity of content is another. If the drug sub- skin (see below) appears promising and has been applied,
stance is very potent and cohesive then mixing a small for example, to the treatment of pain using fentanyl as the
amount of drug with a high amount of filler could lead to active drug. A further step is to incorporate sensors (e.g.
a product with low homogeneity. This problem is most for blood glucose concentration) into devices that control
severe if the drug particles are micronized to improve the the delivery of insulin, in order to provide feedback
dissolution rate. Then it is important to also choose the control of blood glucose.
most appropriate manufacturing process, such as dry
mixing with larger filler paticles, so-called ordered or inter-
active mixing, wet granulation or drug coating of placebo Polymers and surfactants
tablets. Different processes used in drug development and Combining the drug substance with different polymers
manufacturing are described in textbooks such as Aulton and surfactants permits it to adopt states that are interme-
(2007). diate between the pure solid and a free aqueous solution.
As mentioned in the preformulation section the shelf- Polymers in colloidal, gel or solid form can be used to
life of the drug product should preferably be at least 3 entrap drug molecules, and have many applications in

231
Section | 3 | Drug Development

drug formulations (Torchilin, 2001; Dimitriu, 2002; Kim The range of polymers available for drug formulation is
et al., 2009; Savic et al., 2010): vast. A few important examples are shown in Figure 16.1.
• Polymers and surfactants in liquid form give rise to By combining hydrophilic and hydrophobic domains in a
micelles or emulsions, which can greatly enhance the single polymer, as in ‘Pluronic’, micelle formation is
solubility of drug molecules while at the same time encouraged (see below). The inclusion of acidic or basic
protecting them from chemical degradation, and side chains, as in polyaspartate or polylysine, enables the
sometimes improving permeation through tissue polymer to bind oppositely charged drug molecules,
barriers, such as the gastrointestinal epithelium and thereby increasing their effective solubility.
the blood–brain barrier.
• Polymers that form soft hydrated gels are used
mainly in topical dermatological preparations. Micelles
• Solid gel formulations can be used as implantable Micelles (Figure 16.2) consist of aggregates of a few
depot preparations which can be inserted under the hundred amphiphilic molecules that contain distinct
skin to give sustained release of the drug (see below). hydrophilic and hydrophobic regions. In an aqueous
Skin patches can be made from sheet of such a medium, the molecules cluster with the hydrophilic
flexible polymer, loaded with the drug substance. regions facing the surrounding water and the hydrophobic
• Polymers are commonly used in tablets and regions forming an inner core. Micelles typically have
capsules for several purposes, either to facilitate the diameters of 10–80 nm, small enough not to sediment
manufacturing process or for a specific drug release under gravity, and to pass through most filters. Micelle-
profile, e.g. binders, disintegrants, film coating agents. forming substances have limited aqueous solubility, and

Polyethyleneglycol (PEG) H–(–OCH2CH2–)n–OH

O
Polycaprolactone
H–(–OCH2CH2 C–)n–OH

O
Polyglycolic acid (PGA)
H–(–OCH2C–)n–OH

Polylactic acid H–(–OCH2C–)n–OH

CH3

Polyoxyethylene-polyoxypropylene H–(–OCH2CH2–)n–(–OCH2CH2–)m–(–OCH2CH2–)n–OH
block copolymer (Pluronic)
CH3

H–(–NHCHC–)n–OH
Polyaspartate
CH2

COO–

H–(–NHCHC–)n–OH
Polylysine
(CH2)4

NH+

Fig. 16.1  Amphiphilic polymer.

232
 Pharmaceutical development Chapter | 16 |

Micellar absorption by the gastrointestinal tract is directed

mainly to the lymphatic system rather than the vascular
system. Thus, unlike substances absorbed directly from
aqueous solution, micelle-associated compounds tend to
bypass the hepatic portal circulation, and thereby escape
first-pass metabolism.
Micelle formation is a general property of amphiphilic
molecules, and many chemical forms have been devel-
oped for pharmaceutical use (Pillai and Panchagnula,
2001; Torchilin, 2001). Some examples are shown in
Figure 16.2, and additional information on the many
types of polymers used in pharmaceutical formulation is
given by Kumar and Kumar (2001), as well as in many
textbooks (e.g. Ansel et al., 1999; Dimitriu 2002; Aulton,
2007). A particularly versatile group is that of copolymers,
containing more than one type of polymer unit, one of
which, typically, is hydrophilic (e.g. polyethylene glycol),
whereas the other is hydrophobic (e.g. polypropylene
glycol). Alternating blocks of these two units form a
Micelle copolymer (known commercially as Pluronic, see Figure
16.2) which is commonly used in drug formulations.
Amphiphilic polymer Copolymers of this sort at low concentrations form a
liquid micellar suspension, but at higher concentrations
the micelles may aggregate in an ordered array to form a
Hydrophilic part Hydrophobic part water-containing gel. Such gel formulations are com-
(e.g., polyoxyethylene) (e.g, polyoxypropylene) monly used to prepare controlled-release preparations
(see below). The polymer components may include
Lipophilic drug anionic or cationic groups, which have the effect of alter-
ing their affinity for charged drug molecules, and also of
High MW drug
altering their pharmacokinetic behaviour.
Micelles and other drug vehicles, such as cyclodextrins
and liposomes (see below), have a considerable – and
generally beneficial – effect on the pharmacokinetic
Fig. 16.2  Structures of some common pharmaceutical
properties of the drug. Often, but not invariably, absorp-
polymers.
tion from gastrointestinal tract is improved, though the
reasons for this are not well understood. Preferential
uptake into lymphatics, as mentioned above, reduces the
when the free aqueous concentration reaches a certain extent of first-pass metabolism. Circulating micelles
point – the critical micelle concentration – typically in the protect the drug from metabolic degradation, so the
millimolar range, micelles begin to form; further addition plasma half-life is generally prolonged. Micelles are too
of the substance increases their abundance. With some large to cross ‘tight’ capillary endothelium, so transfer
compounds, as the density of micelles increases a gel is across the blood–brain barrier is not increased. They are
formed, consisting of a loosely packed array of micelles able to cross the fenestrated capillaries that occur in most
interspersed with water molecules. Lipophilic drug mole- tissues, but the rate of permeation is less than that of the
cules often dissolve readily in the inner core allowing uncomplexed drug. Malignant tumours and inflamed
concentrations to be achieved that greatly exceed the tissues generally have rather leaky capillaries with large
aqueous solubility limit of the drug. Amphiphilic sub- fenestrations, so that transfer of micellar drug complexes
stances also tend to associate with micelles, as do high- into such tissues is more rapid than into normal tissues.
molecular-weight substances such as peptides and proteins, This mechanism results in a degree of selectivity in the
which have affinity for surfaces, on account of the large distribution of the drug to diseased tissues, a phenomenon
surface area which micelles present. known as passive targeting (see below).
Micelle formation is a natural property of bile acids,
which are secreted into the duodenum under physiologi-
cal conditions and which play an important role in fat
Liposomes
absorption by the intestine. Micellar drug formulations Liposomes were first discovered in 1965 and proposed as
are thus an extension of a normal physiological process. drug carriers soon afterwards (see Samad et al., 2007 for a

233
Section | 3 | Drug Development

disintegrate in the circulation, and so fail to retain the

Targeting Polymer coat Aqueous drug satisfactorily; and their circulating half-life is short,
molecules (PEG) pore because they are rapidly taken up by tissue macrophages,
(antibody)
mainly in liver and spleen, so these tissues receive most of
the drug as a brief bolus. The biological characteristics of
liposomes, however, can be improved by chemical modi-
fications of various kinds:
• The inclusion of cholesterol, and other alterations
of the phospholipid composition, improves the
stability of liposomes and renders them less
permeable to drug molecules.
• Attaching other substances to their surface. Such
substances include polyethylene glycol (PEG; see
Figure 16.3), charged compounds, or antibodies
directed at antigens expressed by tissues in which the
drug is intended to act.
Altering the size, lipid composition and charge on the
vesicle surface affects the rate at which circulating lipo-
somes are taken up by tissue macrophages. Liposomes are
Phospholipid/ Drug substance not generally suitable for oral administration, as they are
cholesterol (soluble or crystalline) destroyed by enzymes and bile acids in the small intestine.
bilayer Nor do they cross the blood–brain barrier, so they cannot
be used to improve the access of drugs to the brain.
Despite an extensive literature acclaiming liposomes as the
Fig. 16.3  Structure of drug delivery liposome.
answer to almost every imaginable formulation problem,
the application of liposome technology in commercial
products is so far limited to a very few examples, in the
recent short review). They are microscopic vesicles formed field of anticancer and antifungal drugs. Liposome tech-
when an aqueous suspension of phospholipid is exposed nology has received much attention in the design of drug
to ultrasonic agitation. Depending on the conditions, delivery systems (e.g. Harrington et al., 2002; Sapra and
large multilayered vesicles 1–5 µm in diameter, or small Allen, 2003; Gabizon et al., 2006), though the develop-
single-layered vesicles 0.02–0.1 µm in diameter, may be ments mostly are at the experimental stage and remain to
formed. The vesicles are bounded by a phospholipid show added value in clinical applications. Recent reviews
bilayer, which is impermeable to non-lipophilic com- providing information on trends and expectations in this
pounds. They can act as drug carriers in various ways field include Barratt (2003), Sapra and Allen (2003),
(Figure 16.3): Goyal et al. (2005) and Samad et al. (2007).
• Non-lipophilic drugs are carried in solution in the
aqueous core, by adding them to the aqueous
medium in which the liposomes are produced. Nanontechnology more then nanoparticles
Techniques for introducing drug molecules into Nanotechnology, using materials in the nanometre scale,
preformed liposomes have also been described. is a new rapidly growing scientific field. The first approved
• Lipophilic drugs occupy the phospholipid membrane products using formulations based on nanoparticles were
phase. liposomes, polymer protein conjugates polymeric susb-
• Some peptides and proteins, as well as amphiphilic stances or suspensions (EMEA, 2006). Medical applica-
drugs, can be sequestered at the lipid–water interface. tions of nanontechnology (nanomedicine) has also been
The main purpose of using liposomes is to improve the suggested to have the potential to create new ‘intelligent
pharmacokinetic behaviour of drugs. Drugs contained in materials’ in nano scale to be used as diagnostic materials,
liposomes are inaccessible to metabolizing enzymes and in stents with and without drugs. In drug development the
transport systems, and their effective biological half-life is technology has been suggested to have a great potential
determined by the rate of clearance of the liposomes. for targeted drug delivery in the treatment of cancer and
These drug-containing vesicles are easy and cheap to organ-specific drug delivery, e.g. to CNS (Singh and Lillard,
manufacture, and in most cases are stable as aqueous 2009). However, the safety issue of delivering such small
suspensions. Simple phospholipid-based liposomes are particles is currently also under discussion and evaluation
unsatisfactory as drug carriers for several reasons: they (e.g. EMEA 2006; De Jong and Borm, 2008). Several new
are relatively permeable to drug molecules and tend to drug-delivery technologies using a number of materials

234
 Pharmaceutical development Chapter | 16 |

There are many different ways of producing sustained-

Table 16.2  Drug delivery applications
of nanotechnology
release oral preparations, some of which are shown
in Figure 16.4. They rely mainly on the use of imperme-
able coatings that are slowly eroded as the pill passes
Drug delivery system Application
through the gut, or water-absorbing polymer gels which
Liposomes Targeting drug/ slowly become hydrated. Injectable implants operate in
oligonucleotide/gene delivery the same way over a longer period, and have the advantage
Micelles Targeting drug/ that they can be removed if necessary. Depot injections
oligonucleotide/gene delivery of drugs dissolved in oil can also be used to provide
long-lasting sustained release, but these generally produce
Nanoemulsions and CNS disorders, targeted a less constant rate of administration and cannot be
nanogels across blood–brain barrier removed.
Nanoparticles, e.g. Carrier, site-specific delivery Controlled release represents a stage beyond sustained
metallic, mesoporous and contrast agents in release and involves coupling a sensing mechanism,
silica, solid lipids cancer therapy responding to changes in temperature or pH, for example,
Oligonucleotides to the drug release mechanism. Examples of the many
kinds of device that could meet this need include
Nanoprobes Biomarkers
temperature-sensitive liposomes, which disintegrate when
the temperature is increased to, say, 40°C, and temperature-
sensitive polymers which aggregate into a gel when the
such as polymers, metals and ceramics are under develop- temperature is increased. The idea is that local heating of
ment where surfaces or pores in nanoscale are loaded a tumour will cause the drug to be released at that site.
with pharmaceutical substances. The drug release is rate Another example is the pH-sensitive system, which can be
determined or programmed using nanotechnology and used to delay the release of acid-sensitive drugs (particu-
conventional pharmaceutical formulation principles in larly peptides) until they have passed beyond the stomach.
combination. Besides the size other physicochemical Drug delivery can also, in principle, be targeted to regions
properties, such as surface properties, particle morphology of low pH, such as inflamed or hypoxic tissues, by the use
and structure, and drug release are of importance for of pH-sensitive polymers. A particularly ingenious
understanding and interpreting in vivo results (Putheti approach is to incorporate insulin into pH-sensitive gels
et al., 2008). Table 16.2 exemplifies drug delivery applica- loaded with glucose oxidase (known to specialists in this
tions of nanotechnology. field as GOD-gels). If the ambient glucose concentration
increases, enzymic oxidation causes a fall in pH and the
release of insulin. (For more details of ‘responsive’ poly-
Modified-release drug formulations mers and their application in controlled drug delivery, see
The most common challenge in drug formulation is to Soppimath et al. (2002), Gupta et al. (2002) and Kim
shorten the time to onset of action or prolong the duration et al. (2009).)
of action of a drug. In either situation the rate of absorp-
tion must be determined by the rate of release of the active
Delivery and formulation
substance from the dosage form. The most common
method is to delay the rate of release of the drug substance of biopharmaceuticals
from the dosage form causing it to be absorbed gradually. In contrast to traditional small drug molecules biophar-
Ideally, the rate of absorption should reach a steady level maceutical drugs are protein-based, built up of amino acid
that is maintained for hours or days, depending on the chains, and structurally complex with many functional
application, until the reservoir is used up. This is known groups. Examples are: human insulin in diabetes therapy,
as sustained release. It is widely used to produce once-daily different vaccines and interferons in, e.g., lung cancer, leu-
oral preparations, or long-lasting depot injections (e.g. kaemia and hepatitis therapy (see Chapter 12). Since these
contraceptives, hormone replacements, antipsychotic molecules are large, the diffusional transport across epi-
drugs) where the drug effect is required to last for weeks thelial barriers in the gastrointestional tract is slow, and
or months. enzymes present there result in fast degradation of the
Other types of modified release include delayed release, molecules; a majority of the biotech drugs are, therefore,
used mainly for oral drugs that are unstable at the low delivered via the parenteral route. Despite that, the market
pH of the stomach, and controlled release, produced by for biotech products is growing and they accounted for
specialized devices that allow the rate of release to be one-fifth of all blockbuster drugs in the US market in 2008
adjusted according to need. Some of the approaches used (Malik, 2008).
to develop controlled-release formulations are discussed Formulation of proteins and peptides are a real chal-
briefly below. lenge since they are usually more unstable compared to

235
Section | 3 | Drug Development

Drug particles

Slowly eroding
1. Eroding impermeable matrix 3. Multilayer
matrix composition
2. Variable coating
thickness

Water Drug Drug

Dry polymer 4. Hydratable dry Hydrated gell

containing polymer containing
immobile drug diffusible drug
molecules molecules

Fig. 16.4  Types of sustained-release preparation.

small molecules and the formulation strategy needs a high oxidation. Liposomes, micelles and the addition of poly-
focus on stabilization. The structures are often both chem- mers and surfactants, as described above, are used to over-
ically and physically unstable. These molecules are often come these stability problems with biopharmaceuticals.
designed for specific mechanisms of action; loss of activity
can arise with, for example, increased temperature or a Drug delivery to the central nervous system
change in pH and, therefore, heating should obviously be
avoided and the right pH conditions need to be selected. Brain capillaries, unlike those in most parts of the body,
Biopharmaceuticals also have fast degradation in aque­ are non-fenestrated, so that drug molecules must traverse
ous solutions and, therefore, freeze-drying is a common the endothelial cells, rather than passing between them,
technique to transfer them into a dry state for longer shelf- to move from circulating blood to the extracellular space
life. However, it is important to select the right tempera- of the brain (see Chapter 10). Three main routes of access
ture and pressure to avoid damage of the molecule during are important (Scherrmann, 2002):
the process. Further instability problems that could arise • Lipophilic compounds of low molecular weight cross
with biopharmaceuticals are protein aggregation and the membrane of endothelial cells very easily, and

236
 Pharmaceutical development Chapter | 16 |

comprise the great majority of CNS-acting drugs. marketable product that will perform reliably when
Peptides, proteins, non-lipophilic or ionized drugs used in real life. Preformulation studies consist of a
are, for the most part, unable to cross the endothelial series of chemical and physicochemical investigations
cell membrane. on the drug substance which indicate the kinds of
• The endothelial cells also possess various active formulation that are likely to be satisfactory. In some
transport mechanisms that can allow certain cases problems (for example, poor solubility or chemical
non-lipophilic compounds to enter the brain. instability) will emerge at this stage, requiring modi­
Examples include levodopa, used for treating fication of the drug molecule before development can
Parkinson’s disease, baclofen, a GABA analogue used proceed – in other words, back to the drawing board.
to treat spasticity, and the cytotoxic drug melphalan, Therefore it is important to run preformulation activities
all of which are transported across the blood–brain in parallel with early drug development of new chemical
barrier by the amino acid transporter. Attempts have entitites.
been made to couple other drugs with amino acids The process of formulation will depend greatly not only
or sugars which are carried via this transport system. on information from preformulation activites, but also on
Despite being successful in animal models, however, the medical condition and on the intended route of
such compounds have not been developed for administration of the drug. In most cases, where the inten-
clinical use. Active transport out of the brain tion is to produce a tablet or capsule for oral use, an
also occurs with many compounds, including drugs intravenous formulation will also be developed for use in
such as penicillins, which are able to enter passively. clinical trials, and the oral form used in initial efficacy
• Molecules can be carried as endocytotic vesicles trials (clinical Phase II) may not be the same as the
across the endothelial cells. This type of transcytosis intended marketed form.
occurs with molecules that are bound to receptors or Even in cases where no problems are encountered, for-
other proteins on the endothelial cell surface. An mulation studies require considerable time and resources.
approach that has been tested extensively, though The end result has to be a product that can be manufac-
not yet applied for clinical use, is to couple active tured on a large scale and meet strict quality-control stand-
peptides and proteins to the monoclonal antibody ards, and can be stored in thousands of homes under
OX26, which recognizes the endothelial transferrin varying conditions of temperature and humidity without
receptor. Binding to this receptor stimulates significant deterioration.
transcytosis, carrying the antibody and its cargo Very often pharmaceutical development is called on to
across the blood–brain barrier. Transcytosis can also improve the characteristics of the drug substance, for
be stimulated by the non-specific binding of small example by improving its solubility using amorphous
cationic peptides to the acidic glycoprotein substances or new salts, disguising its taste, increasing
components of the endothelial cell surface. its plasma half-life or reducing unwanted effects, and
Conjugates of various cytotoxic and antimicrobial work of this kind increases the value of a substance
drugs to such peptides show improved brain clinically and commercially. Increasing use is being made
penetration in animal models, and may prove to be of colloidal systems, such as micelles, polymers and lipo-
applicable clinically. somes as vehicles for drug molecules. Such formulations
have a considerable effect on the drug’s pharmacokinetic
Recent approaches for improving drug delivery to the
properties, and can also be used to achieve a degree of
brain are described in more detail by Denora et al. (2009),
targeting of the drug to the tissues on which it is required
Patel et al. (2009) and Pardridge (2010).
to act. Drug targeting based on these and other principles,
Enabling impermeant drugs to reach the brain repre-
e.g. antibodies, are currently the subject of much experi-
sents a major challenge for formulation chemists, and
mental work. The principles have only proved applicable
there are actually very few examples where it has been
so far to a few anticancer and antifungal drugs, but
overcome. Most often, the drug molecule has to be rede-
many more applications are expected in the foreseeable
signed to increase its lipophilicity. Formation of a lipid-
future.
soluble prodrug is one strategy, but is rarely effective in
Overall, work on more sophisticated formulations,
this context because conversion to the active, non-
new routes of administration and new delivery systems,
lipophilic compound is likely to take place in the circula-
e.g. based on nanotechnology for currently used drugs
tion before the drug reaches the brain.
and biopharmaceuticals, is thought likely to contribute
as much to improved therapeutics as the discovery of
new drugs, and is seen by the pharmaceutical industry
SUMMARY as an important parallel approach to drug discovery,
particularly at times – like the present – when drug
Pharmaceutical development comprises all the activities discovery runs into a phase of disappointingly low
needed to turn a therapeutic drug substance into a productivity.

237
Section | 3 | Drug Development

REFERENCES

Allen LV. The art, science and nanotechnology-based medicinal systems: a review. Journal of
technology of pharmaceutical products for human use. EMEA/ Controlled Release 2009;140:2–11.
compounding. 3rd ed. Washington CHMP/79769/2006. Malik NN. Drug discovery: past, present
DC: American Pharmaceutical Gabizon AA, Shmeeda H, Zalipsky S. and future. Drug Discovery Today
Association; 2008. Pros and cons of the liposome 2008;13:909–12.
Ansel HC, Allen LV, Popovich NG. platform in cancer drug targeting. Pardridge WM. Biopharmaceutical drug
Pharmaceutical dosage forms and Journal of Liposome Research targeting to the brain. Journal of
drug delivery systems. 7th ed. 2006;16:175–83. Drug Targeting 2010;18:157–67.
Baltimore: Lippincott Williams & Goyal P, Goyal K, Vijaya Kumar SG, Patel MM, Goyal BR, Bhadada SV, et al.
Wilkins; 1999. et al. Liposomal drug delivery Getting into the brain: approaches to
Aulton ME, editor. Aulton’s systems – clinical applications. Acta enhance brain drug delivery. CNS
pharmaceutics. The design and Pharmaceutica 2005;5:1–25. Drugs 2009;23:35–58.
manufacture of medicines. 3rd ed. Gupta P, Vermani K, Garg S. Hydrogels: Pillai O, Panchagnula R. Polymers in
Edinburgh: Churchill Livingstone; from controlled release to drug delivery. Current Opinion in
2007. pH-responsive drug delivery. Chemical Biology 2001;5:447–51.
Barratt G. Colloidal drug carriers: Drug Discovery Today 2002;7: Putheti RR, Okigbo RN, Sai advanapu
achievements and perspectives. 569–79. M, et al. Nanotechnology
Cellular and Molecular Life Sciences Harrington KJ, Syrigos KN, Vile RG. importance in the pharmaceutical
2003;60:21–37. Liposomally targeted cytotoxic drugs industry. African Journal of Pure and
Bredenberg S, Duberg M, Lennernäs B, for the treatment of cancer. Journal Applied Chemistry 2008;2:27–31.
et al. In vitro and pharmacokinetic of Pharmaceutics and Pharmacology Samad A, Sultana Y, Aqil M. Liposomal
evaluation of a new sublingual tablet 2002;54:1573–600. drug delivery systems: an update
system for rapid oromucosal Illium L. Nasal drug delivery: new review. Current Drug Delivery
absorption using fentanyl citrate as developments and strategies. Drug 2007;4:297–305.
the active substance. European Discovery Today 2002;7:1184–9. Sapra P, Allen TM. Ligand-targeted
Journal of Pharmaceutical Sciences Illium L. Nasal drug delivery – liposomal anticancer drugs. Progress
2003;20:327–34. possibilities, problems in Lipid Research 2003;42:439–62.
Brown BA-S, Hamed E. oraVescent® and solutions. Journal of Controlled Savic S, Tamburic S, Savic MM. From
Technology offers greater oral Release 2003;87:187–98. conventional towards new – natural
transmucosal delivery. Drug Delivery Kaplan SA. Relationships between surfactants in drug delivery systems
Technology 2007;7:42–5. aqueous solubility of drugs and their design: current status and
Burger A, Abraham D, editors. Burger’s bioavailability. Drug Metabolism perspectives. Expert Opinion in Drug
medicinal chemistry and drug Reviews 1972;1:15–32. Delivery 2010;7:353–69.
discovery. New York: John Wiley and Kim S, Kim J-H, Jeon O, et al. Scherrmann JM. Drug delivery to brain
Sons; 2003. Engineered polymers for advanced via the blood–brain barrier. Vascular
Denora N, Trapani A, Laquintana V, drug delivery. European Journal of Pharmacology 2002;38:349–54.
et al. Recent advances in medicinal Pharmaceutics and Biopharmaceutics Singh R, Lillard Jr JW. Nanoparticle-
chemistry and pharmaceutical 2009;71:420–30. based targeted drug delivery.
technology – strategies for drug Kroboth PD, McAuley JW, Kroboth FJ, Experimental and Molecular
delivery to the brain. Current Topics et al. Triazolam pharmacokinetics Pathology 2009;86:215–23.
in Medicinal Chemistry after intravenous, oral, and
2009;9:182–96. Soppimath KS, Aminabhavi TM, Dave
sublingual administration. Journal AM, et al. Stimulus-responsive
De Jong WH, Borm PJA. Drug delivery of Clinical Psychopharmacology ‘smart’ hydrogels as novel drug
and nanoparticles: applications and 1995;15:259–62. delivery systems. Drug Development
hazards. International Journal of Kumar RMNV, Kumar H. Polymeric and Industrial Pharmacy
Nanomedicine 2008;3:133–49. controlled drug-delivery systems: 2002;28:957–74.
Dimitriu S, editor. Polymeric perspective issues and opportunities. Torchilin V. Structure and design of
biomaterials. New York: Marcel Drug Development and Industrial polymeric surfactant-based drug
Dekker; 2002. Pharmacy 2001;27:1–30. delivery. Journal of Controlled
EMEA (European Medicines Agency) Madhav NVS, Shakya AK, Shakya P, Release 2001;73:137–72.
(2006) Reflection paper on et al. Orotransmucosal drug delivery

238
 Chapter 17 
Clinical development: present and future
C Keywood

market, Di Masi (2003) estimated the costs to be around

INTRODUCTION US$802 million and Adams and Brantner (2006) made an
estimate of US$868 million but within a range of $500
Clinical development is the art of turning science into million to $2 billion, depending on the indication pursued
medicine. It is the point at which all the data from basic and the company performing the development. Of this
science, preclinical pharmacology and safety are put into total amount the clinical development accounts for just
medical practice to see whether scientific theory can trans- over half, i.e. about US$480 million.
late into a valuable new medicine for patients. The funda- In spite of heavy investment in research and develop-
mental purpose of the clinical development programme is ment, new product approvals have been decreasing in the
to provide the clinical information to support the product last 10 years (Woodcock and Woosley, 2008). Pipelines of
labelling, which ultimately tells the healthcare profes- large pharmaceutical companies have been declining and
sional and patient how to use the drug effectively and the productivity of large pharmaceutical companies in
safely. new drug development has been diminishing in an envi-
The segments of product label coming from clinical ronment of increasing costs of clinical development and
trials are the pharmacokinetics, the dosing regimen in increasing risk aversion of companies, regulators and the
the main population and in special populations, e.g. the general public. New strategies are needed to overcome this
elderly or those with hepatic and renal impairment, the problem, both in the way companies source and develop
clinical pharmacology/mechanism of action in man, drug new drugs and also in the way regulators approach the
interactions, contraindications, warnings, precautions, evaluation of efficacy and safety of new medicines.
efficacy in the indication, safety and side effects. All this This new environment is leading to an evolution in
information has to be generated from a development pro- clinical development strategy, with the type of indications
gramme designed to investigate these specific properties. being pursued and in the sourcing and development of
Clinical development has to satisfy the demands of new compounds. There is a move away from blockbuster
regulators who will grant product approval, government medicines, i.e. ‘one size fits all’ in major indications,
organizations responsible for reimbursement in countries towards more specialized unmet medical needs and
where healthcare is state subsidized, managed care organi- patient-tailored therapies, i.e. personalized medicine. The
zations in the USA and also the marketing team who will success of the human genome project was supposed to
sell the product. The needs are sometimes conflicting and have heralded a new era for patient-specific drug develop-
a challenge of clinical development is to design a trials ment. However, for now, the art of medicine appears to
programme that not only demonstrates that the new drug continue to outwit the theory of basic science and drug
is effective and safe, but also balances the various desires development based purely upon genomic approaches has
and requirements of the different parties, for the product yet to live up to its promises.
profile. There is an increasing trend for large pharmaceutical
Bringing new drugs to the market is not only complex companies to collaborate with small pharmaceutical com-
but also costly, time and resource consuming. Taking into panies and biotech companies in order to enhance discov-
account failure of drugs in development to make it to ery pipelines and drug development productivity. Large

© 2012 Elsevier Ltd. 239

Section | 3 | Drug Development

companies have a great deal of resources to put behind including first in man – Phase I; exploratory, proof of
development programmes but internal competition for concept – Phase IIa; confirmatory efficacy and dose range
resources, risk aversion and political pressures within finding – Phase IIb; confirmatory, large scale efficacy and
these organizations can mean they are less flexible and safety – Phase III; marketing authorization application,
creative in clinical development. Small pharmaceutical license extension and post-marketing surveillance – Phase
companies can provide the flexibility, innovation and crea- IIIb/IV.
tivity to complement the large pharmaceutical company
development activities and there is an increasing trend for
new drug development to be performed in partnership. Phase I – Clinical pharmacology
In the United Sates the Food and Drug Administration
Phase I is somewhat of a misnomer for, although the first
(FDA) published a white paper in 2004 (FDA, 2004a) that
studies in man are performed as part of Phase I, many of
identified that the current methods of drug development
the other components of Phase I, for example drug–drug
were partly behind the decline of new drug applications
interactions, special populations and human radiolabelled
and has set up the Critical Path Initiative (FDA, 2004b) in
studies, are conducted in parallel with later phase studies.
order to address some of the problems of low productivity
Hence it is probably more appropriate to call these ‘clini-
and high late-stage attrition rates, the idea being to encour-
cal pharmacology studies’.
age novel approaches to clinical development in trial
A typical Phase I programme may contain around 20
design and measuring outcomes. The Innovative Medi-
clinical pharmacology studies. The main objectives of the
cines Initiative (EFPIA, 2008), underway in Europe, is also
programme are to define the pharmacokinetics, metabo-
looking to encourage public and private collaboration
lism and safety of the intended formulation given alone
with small and large enterprises and academia, to share
and with other drugs that have potential to interact either
knowledge, enhance the drug discovery and development
kinetically or dynamically with the new drug. How the
process, reduce late-stage attrition and, ultimately, bring
drug is handled by certain populations, such as the elderly,
good medicines to patients in a cost- and time-effective
ethnic groups or those with hepatic or renal impairment,
manner.
is also studied. All of this information goes into the pre-
These factors are changing the way clinical development
scribing information to guide the safe and effective use
is being performed and will be performed in the future. In
and dosing of the drug. Some efficacy information can
this chapter the conventional path of clinical development
also be gathered in human pharmacological models in
will be explained along with the new strategies for merging
order to assist dose selection for trials in patients. The
phases and using adaptive trial design to enhance the
principal components are as follows:
development of new medicines.
• First in man single ascending dose pharmacokinetics
and safety
• Multiple ascending repeat dose pharmacokinetics
and safety
CLINICAL DEVELOPMENT PHASES
• Pharmacodynamic studies
• Bioavailability absolute and bioequivalence of new
The conventional path of clinical development involves formulations
four phases: • Absorption distribution metabolism excretion in
• Phase I, pharmacokinetics, safety and tolerability and man (radiolabelled studies)
clinical pharmacology • Drug–drug interaction studies
• Phase IIa, exploratory efficacy • Safety pharmacology, e.g. thorough QT studies and
• Phase IIb, efficacy and dose range finding abuse liability studies
• Phase III, pivotal efficacy and safety and larger • Elderly pharmacokinetics and safety
population • Ethnic groups pharmacokinetics and safety
• Phase IV, post-marketing safety and efficacy • Hepatic and renal impairment, pharmacokinetics
evaluation. and safety.
A summary of these phases is given in Table 17.1. Clinical pharmacology studies and in particular ‘first in
Traditionally these phases have been conducted in a man’ studies are conducted by specialist medical staff,
step-wise manner with decisions to proceed to the next trained in clinical pharmacology, in specialized units
stage being made once the preceding stage was completed. either within or close to a major hospital. The units are
However, more recently, the margins between phases are specifically equipped for the preparation and correct
becoming less distinct as more seamless drug develop- administration of the test drugs (and in some cases manu-
ment programmes are being performed and adaptive facture of the drug product), collection and storage of
trial designs adopted. Therefore, it may be more appropri- biological samples and management of subject safety,
ate to describe the phases as: clinical pharmacology, including full resuscitation facilities. The subjects selected

240
 Table 17.1  Phases of clinical development

Clinical General aim Subjects/design Data collected Approx Approx Approx

phase number of cost ($ 000) time
subjects required
Ia Exploratory; safety, Healthy subjects. Escalating Adverse events PK parameters 40–60 200–400 6 months
tolerability and PK to single-dose, placebo-controlled, PD measures (sometimes)
support patient studies randomized, double-blind
Ib Exploratory; safety, Healthy subjects. Escalating Adverse events PK parameters 30–50 200–400 6 months
tolerability and PK to repeat-dose, placebo-controlled, PD measures (sometimes)
support patient studies randomized double-blind
IIa Exploratory; preliminary Patients; intended clinical dose and Adverse events PK parameters 50–200 500–1000 9 months–
safety and efficacy to regimen based on Ph I results Preliminary evidence of 2 years
support go/no-go decision Usually placebo-controlled efficacy. ‘Proof of concept’
randomized double-blind, but
sometimes open label
IIb Confirmatory; dose selection Patients in one or more indications Statistically rigorous analysis of 200–500 2000–5000+ 2–3 years
to support registration Selected dose levels/regimens dose–response relationships
compared to placebo and/or Confirmation of clinical dose
standard treatment, randomized and regimen for optimum
double-blind efficacy, safety and
tolerability
III Confirmatory; efficacy and Patients in target indication(s) but Statistically rigorous 500–1000+ 2000–10 000+ 2–5+ years
safety data to support including different groups (age, measurements
registration; may include ethnicity, etc.). Selected dose demonstrating safety and
pharmacoeconomic level compared with placebo efficacy in comparison with
evaluation and/or standard treatment(s) placebo or existing therapies
randomized double-blind May include pharmacoeco­
nomic analysis
IV Obligatory post-marketing Treated patients Adverse events 10 000+ 10 000+ 2–4+ years
Clinical development: present and future

surveillance to reveal
unexpected adverse
effects or toxicity
Chapter | 17 |

241
Section | 3 | Drug Development

for studies are usually enrolled from the unit’s subject Subjects
volunteer panel. Typically a clinical pharmacology unit In most cases the subjects in first in man studies will be
will advertise for subjects to join their database. People conducted in healthy young men usually aged 18–45
who are interested in taking part in trials are then carefully years. The idea behind this is to have a fairly homogeneous
screened for their physical health, their ability to compre- population in which to study the effects of the new drug
hend the requirements of taking part in the studies and and so limit variability, and also to have a population who
their motivation to comply with the constraints of the will be more able to withstand any unexpected toxicity
studies, to check whether they are suitable to join the caused by the test drug. Healthy subjects are those who
volunteer panel. In some countries (e.g. France), the sub- have no underlying diseases that could interfere with the
jects have to be registered on a national database to make conduct of the study or confound the interpretation of the
sure they comply with laws governing participation in safety or pharmacokinetic data. Criteria for inclusion into
clinical trials. For most large professional units, the data- the study based upon medical history, physical examina-
base will comprise a variety of healthy subjects including tion, use of concomitant medications, alcohol, cigarettes
young men and women (18–45 years), older healthy sub- and recreational drugs, as well as the results of blood
jects (over 55 years), subjects of non-Caucasian origin (e.g. testing 12-lead ECG, blood pressure heart rate are laid out
Japanese) and subjects with genetic polymorphisms for in the study protocol. Male subjects are generally pre-
CYP metabolism. As the subjects gain no medical benefit ferred, because at this early stage of development, repro-
from taking part in the studies and have to undergo pro- ductive toxicology testing in animals will not have been
cedures which can be mildly unpleasant or inconvenient, completed and the risk to the fetus of female subjects who
they are paid for taking part. However, the amount they might be pregnant or become pregnant shortly before or
can be paid is restricted to be commensurate with the after the study, has not been characterized. Once the
inconvenience and discomfort of the study and is not so segment 2 reproductive toxicology has been completed
high as to be an inducement to take part. The subjects’ (see Chapter 15), female subjects of non-childbearing
reimbursement is reviewed and has to be approved by the potential, i.e. post-menopausal, surgically sterilized,
ethics committee, before the study can go ahead. The sexually abstinent or using effective methods of contracep-
number of studies that subjects can take part in annually tion, may be included in European studies. In the USA,
is also restricted. Most protocols demand that a subject female subjects may be included prior to reproductive
cannot be exposed to another investigational drug within toxicology data being available if they are of non-
90 days of taking part in a study and so that limits how childbearing potential. In some cases it is not appropriate
many studies in which a given subject can participate, in to use healthy young men, for example, studies of female
any one year. hormone products or products for oncology, and in these
cases the subjects will be selected from the appropriate
First in man, single ascending dose, population.
pharmacokinetics and safety
First dose in man (FDIM), also known as single ascending Design
dose (SAD), is a red-letter day for the drug development The classical design of the SAD study is a double-blind,
team, when single doses of the drug are given to small placebo-controlled sequential cohort design; where the
cohorts of subjects in a sequential manner until the first cohort takes the lowest dose and then the dose is
maximum tolerated dose is achieved. escalated through subsequent cohorts, provided the toler-
ability and safety in the preceding cohort are acceptable.
Objectives The size of each group is usually six to eight subjects, with
The objectives of the SAD study are to evaluate the safety two of the subjects being randomized to placebo. The use
(physiological effects on body systems), tolerability of placebo and the double-blinding, where neither the
(occurrence of side effects) and pharmacokinetics of dif- investigator nor the subject knows the treatment being
ferent doses of the test drug, and to identify the maximum taken, allows for a more objective evaluation of the safety
tolerated dose (MTD). That is the dose at which either the and tolerability of the test drug.
occurrence of intolerable side effects and/or unacceptable The choice of doses to be administered in the SAD trials
safety findings are encountered. The MTD is important to should be based on the highest dose at which no adverse
identify for estimation of the therapeutic window, which effects were seen in the most sensitive species tested in
is the dose range within which the drug is effective but toxicology studies (the no observed adverse effect level, or
safe and well tolerated. In some cases it is not possible to NOAEL; see Chapter 15) and on the nature of the toxicity
achieve the MTD, perhaps if the drug is very well tolerated, observed. The starting dose should be at least 10-fold
or has poor bioavailability, in which case the maximum lower than the NOAEL, but specific guidelines exist for the
feasible dose can be used to estimate the therapeutic accurate determination of the starting dose on a case-by-
window. case basis (FDA; http:www.fda.gov/cber/gdlns/dose.htm).

242
 Clinical development: present and future Chapter | 17 |

The dose escalation plan will depend on the character- Outcome measures
istics of the drug and its metabolites, and especially on the In the SAD study the principal outcome measures are
nature of any toxicity seen in preclinical toxicology testing safety, tolerability and pharmacokinetics.
at doses above the NOAEL. It will also be influenced by Throughout the study, for an appropriate period after
the relationship between dose, systemic exposure as deter- each dose, the subjects are intensively monitored for signs
mined by its pharmacokinetic (PK) profile in animals, and and/or symptoms of toxicity (adverse events). The safety
the pharmacodynamic (PD) effects observed in preclinical parameters measured include, blood pressure, heart rate
pharmacology studies. Each dose escalation step will be and rhythm, 12-lead ECG (intervals and morphology),
dependent on satisfactory safety data from the previous body temperature, haematology, liver function and renal
dose level, according to the clinical judgment of the function, as well as observation for any other unwanted
investigator. effects. Tolerability is evaluated by the documentation of
Usually the drug is administered only once to each adverse events which are collected throughout the study
subject so that each dose group comprises separate sub- and then categorized by severity, duration, outcome and
jects. This has the advantage of maximizing the population causality. If an adverse event meets certain specific criteria
exposed to the test drug. Sometimes, however, it may be (for instance if it is life-threatening or necessitates hospi-
appropriate for each subject to receive two or three of the talization) it is classified as a serious adverse event (SAE) and
planned doses at successive visits, with the proviso that must be reported without delay to the ethics committee,
each dose is administered only after the response to the and usually also to the regulatory authority.
preceding dose in the series has been evaluated. In this Whereas assessment of safety and tolerability is the
way the required number of subjects is reduced, but each primary objective, pharmacokinetic evaluation is the sec-
has to attend more than once. ondary objective of a SAD study. Blood samples will nor-
Typical dose escalation schedules from a starting dose mally be taken before dosing and at specified intervals
of X are: after dosing to measure the amount of drug in the blood
• Dose escalation schedule 1: X, 2X, 4X, 8X, 16X, 32X or plasma at various time points after each dose. A typical
• Dose escalation schedule 2: X, 2X, 4X, 6X, 8X, 10X. sampling schedule is shown in Figure 17.1A. In addition,
urine and/or faeces may be collected to measure the excre-
The dose escalation schedule is guided primarily by tion of the drug via the kidneys and/or liver (in bile). The
safety considerations. Dose escalation schedule 1, where results enable the rate of absorption, metabolism and
the dose is increased exponentially, would be appropriate excretion of the drug to be explored, and the PK param-
for a drug that has shown low toxicity in animal testing. eters to be estimated (see also Chapter 10).
Schedule 2, where the dose increments are constant, is The plasma or serum PK parameters usually derived are:
more conservative and might be more appropriate for a
drug which has a toxicology profile that calls for a more • Cmax: peak drug and/or metabolite(s) concentration
cautious approach to dose escalation in man. There are • Tmax: time to peak drug and/or metabolite(s)
many other possible dose escalation patterns, and each is concentration
considered on its own merits. The ideal is to exceed the • AUC0–∞: area under the concentration–time curve of
dose predicted to effective from preclinical pharmacology the drug and/or metabolite(s), extrapolated to
studies, by a good margin. A drug for which the MTD is infinity
close to the dose predicted for therapeutic effect is less • AUC0-T: area under the concentration–time curve of
likely to be successful than one that has a wide therapeutic the drug and/or metabolite(s), calculated to a
margin. specific time point T
As it is common for the PK profiles of orally adminis- • t1/2: time taken for levels of drug and/or
tered drugs to be affected by the presence or absence of metabolite(s) to decrease by half (a measure of the
food in the stomach at the time of dosing, the food effect rate of elimination of the drug from plasma).
is usually investigated at the end of the SAD study. A dose Other PK parameters may also be determined, including:
level that is safe and well tolerated (e.g. one-quarter of the
MTD) will be given to healthy subjects on two occasions • VD: volume of distribution of drug and/or
once in the fed state (following a standard high-fat break- metabolite(s)
fast) and once fasted (overnight fast). The order of fed and • Cl: clearance of drug and/or metabolite, i.e. the
fasted administration is randomized among the subjects volume of plasma/serum cleared of drug and/or
and the two dose periods are separated by an appropriate metabolite(s) per unit time, e.g. mL/min, L/h
washout period, so that the results of the second period • MRT: mean residence time, i.e. the average time a
are not affected by the drug administration on the first drug molecule remains in the body after rapid i.v.
period. The results of the fed/fasted comparison will form injection.
the basis of the dosing instructions for all future studies Specialist pharmacokineticists perform the calculation
with the drug. of these parameters.

243
Section | 3 | Drug Development

Single-dose study

15 30
Time 0 1h 11⁄2 h 2h 4h 6h 8h 12 h 24h 48 h
min min

Dose

Blood sample
A

Repeat-dose study

15 30
0 1h 11⁄2 h 2h 4h 6h 8h 12 h
min min
Day 1

15 30
0 1h 11⁄2 h 2h 4h 6h 8h 12 h
min min
Days 2–7

15 30
0 1h 11⁄2 h 2h 4h 6h 8h 12 h 24 h 48 h
min min
Day 8

Fig. 17.1A, B  A, A typical sampling schedule. B, A typical Phase MAD blood sampling schedule.

A comparison of the PK parameters at each dose level, two placebo) to 12 (eight active four placebo) subjects
e.g. AUC∞′, Cmax, will indicate whether they increase pro- taking successively higher doses for several days. As for the
portionately (linear kinetics) or disproportionately (non- SAD, the decision to escalate to the next dose level is based
linear kinetics) with increasing dose. This information will upon the safety and tolerability of the preceding dose
influence the selection of dose levels, regimen and dura- level. The dose regimen in the MAD study is based upon
tion of dosing for the multiple, ascending repeat-dose the PK characteristics seen in the SAD and designed to give
study (MAD). The single-dose PK data can also be used to the PK profile necessary to allow the drug to exert its
predict the drug/metabolite(s) concentrations expected on therapeutic effect and to achieve ‘steady state’ in which the
repeated dosing, based on the assumption that the kinetics drug’s input rate is balanced by its rate of elimination. The
do not change with time. duration of the dosing is based upon the desire to achieve
steady state but also to collect sufficient safety information
Multiple ascending repeat-dose studies to support use in Phase II studies. For indications where
After review of the safety data and the single-dose PK chronic dosing is required a duration of around 7–10 days
profile, two or three safe and well-tolerated dose levels are is usual in MAD studies.
chosen and the multiple ascending (repeated-dose) study
(MAD) is designed. Its purpose is to test safety, tolerability Outcome measures
and PK when the drug is given repeatedly. Safety assessment is necessary under these conditions, as
steady-state blood levels are usually higher than blood
Design levels following a single administration. It is also impor-
A typical MAD study design will be double-blind, placebo- tant to know whether the PK of the drug and/or
controlled with two or three cohorts of eight (six active metabolite(s) changes on repeated dosing. For instance,

244
 Clinical development: present and future Chapter | 17 |

saturation of elimination pathways (e.g. metabolizing can be very useful to confirm that a new drug is actually
enzyme systems) could cause the drug to accumulate to having the pharmacological effect in man that was pre-
toxic levels in the body, or alternatively stimulation dicted from animal studies. As well as studying pharmaco-
(induction) of drug-metabolizing enzyme systems could dynamics in healthy subjects, patients with a mild form of
cause the levels of drug and/or metabolite(s) to decrease disease can be studied. An example is asthma, where a new
to subtherapeutic levels. A comparison of the predicted drug could be studied for a short period of time in mild
and observed plasma–serum concentration time curves asthmatics to observe a pharmacological response, without
will provide evidence of any such non-linear or time- necessarily intending to provide a long-term therapeutic
dependent kinetics for the drug and/or metabolite(s). benefit. Such subjects are known as patient volunteers and,
The general requirements and procedures for MAD like healthy subjects, as they are not entering the study to
studies are the same as those for Phase I SAD. A typical seek a cure for their disease, they can receive some finan-
Phase MAD blood sampling schedule is shown in Figure cial compensation for their time and inconvenience.
17.1B. Blood samples for PK profiling will generally be
taken on the first and last days of dosing, with additional Drug–drug interaction studies
single samples taken immediately before the first morning
The objective of drug–drug interaction studies (DDI) is to
dose each day, to measure the levels of drug remaining in
determine whether the test drug’s efficacy, safety or phar-
the blood immediately before the next dose is adminis-
macokinetics will be altered if it is given with other drugs
tered, i.e. the trough level.
that the target patient population may also be taking. The
The results of the Phase I SAD and MAD studies together
timing of drug interaction studies depends partly on the
support the decision as to whether to administer the drug
importance of understanding interactions prior to treating
to patients and, if so, at what dose and regimen and for
patients. Some DDIs may need to be performed prior to
how long.
Phase IIa if the patient population in the study cannot be
excluded from taking concomitant medications that might
Pharmacodynamic studies interact with the test drug. Otherwise DDIs can be con-
ducted later in the development programme when more is
Pharmacodynamic (PD) assessments may be included in
known about the target treatment population and the effi-
the MAD studies, or specific PD studies can be performed
cacy and safety of the new drug. The choice of which DDI
separately. Usually prior to Phase IIa, the objective of these
studies to perform is based on the following: the metabo-
studies is to establish whether the drug has some pharma-
lism of the new drug (for example if it inhibits, induces or
cological effect in man that may be relevant to its thera-
is metabolized by certain cytochrome p450 enzymes – see
peutic effect, and to determine at what doses and plasma
Chapter 10); the pharmacodynamic action of the drug; and
concentrations the effects are seen, with a view to optimiz-
which drugs the target population may be taking concom­
ing dose selection for Phase IIa. Such studies are termed
itantly. Drug interactions can occur on a metabolic level,
PK/PD studies.
so that the exposure to a certain dose may be changed by
This approach must still be treated with some caution,
a drug interaction or on a pharmacodynamic level so that
as the physiology in patients may differ from that in
pharmacological effects may be increased or diminished
healthy subjects, and clinical efficacy may therefore not be
by the interacting drug. It is important to known whether
reliably predicted from Phase I results. It is not uncom-
co-administration of the relevant drugs leads to a change
mon for drugs that are highly effective in patients suffering
in exposure or a change in pharmacodynamic effect of
from a certain condition to have little or no effect on the
either the test drug or the interacting drug.
same body system in healthy subjects. This is particularly
In a typical DDI study, subjects are dosed with the inter-
true of drugs acting on psychiatric diseases as these condi-
acting drug until steady state is achieved and then a single
tions are very difficult to emulate in healthy subjects. The
dose of the test drug is given and the PK, safety, and some-
ideal PD assessments in Phase I are those that have bio-
times PD effects are evaluated. The information from the
logical or surrogate markers that are measurable in healthy
DDI studies gives guidance as to whether any dose altera-
subjects and are relevant to drug’s mechanism of action
tions are necessary when the new drug is co-administered
and/or therapeutic effect. For example, the ability of a
with a drug with which it interacts and the information is
β-adrenoceptor antagonist to inhibit exercise-induced
included on the product label.
tachycardia, or the effect of a proton pump inhibitor on
acid gastric secretion are relevant effects that can easily be
Absolute bioavailability and bioequivalence
measured objectively in volunteers. However, such markers
are not always available (e.g. in the case of many psychiat- of new formulations
ric diseases) or may be misleading (pain is a good example Absolute bioavailability studies are performed to compare
because the endpoints are subjective rather than objec- the exposure of an intravenous preparation of the study
tive), and the interpretation of such data is usually drug which has 100% bioavailability, with the formulation
approached with care. Nonetheless, Phase I PK/PD studies intended for clinical use which is to be given by another

245
Section | 3 | Drug Development

route, e.g. oral or subcutaneous. The absolute bioavailabil- new compound to cause QT prolongation, which may
ity information is needed for the product label, but also carry a risk of a potentially fatal ventricular arrhythmia,
assists further formulation development as modifications called torsade de pointes, or for some central nervous
to the formulation can be made in order to optimize system drugs, abuse liability studies are needed. There are
exposure to the drug. The studies are conducted in healthy specific study designs to look for whether a drug has
subjects in a crossover fashion where each subject receives potential for abuse. Other studies such as effect on reac-
an intravenous dose and then one or more doses of the tion time or driving ability may also be needed, particu-
study drug given by its intended clinical route of admin- larly for central nervous system compounds.
istration. Standard pharmacokinetic parameters are The need for thorough QT studies has become a general
measured and the absolute bioavailability calculated by requirement for all NCEs in the past few years, whether
comparing the pharmacokinetics of the intravenous and or not the compound demonstrates any potential to
non-intravenous doses. cause cardiac conduction abnormalities in vitro or in non-
In the initial clinical pharmacology trials and in the clinical studies. The studies are usually conducted in mid-
exploratory efficacy studies it is not usual to use a formula- stage of development when efficacy has been shown, the
tion that would be the final commercial formulation. likely therapeutic dose is known and prior to large scale
Development of a commercial formulation is performed confirmatory studies. They are conducted in healthy male
in parallel with the early phase studies and the data from and female subjects with no evidence of heart disease or
those studies guides the formulation development activity. concomitant medication that could affect cardiac conduc-
Once a commercial type formulation or formulations has tion. The study typically comprises four arms in a cross­
been identified, the safety and pharmacokinetics are com- over, that is the test drug at the intended clinical dose, the
pared with the prototype formulation in comparative test drug dosed at or near the MTD, an active comparator
bioavailability studies. As for absolute bioavailability, the known to cause QTc prolongation, e.g. moxifloxacin and
studies are conducted in healthy subjects in a crossover placebo. Drug dosing on each treatment day is followed
fashion. If there is no more than 5% difference in exposure by the collection of multiple ECGs at multiple time points,
(AUC) between two formulations, they are considered to especially around Cmax, that are read by hand by a blinded
be bioequivalent. Dosing information obtained from central observer. As the studies require approximately 50
exploratory efficacy studies using a prototype formulation subjects and hundreds of ECGs to be read they are labori-
which is bioequivalent to the commercial type formula- ous and expensive. The Guidance ICH E14 from 2005 sets
tion, can, therefore, be applied directly to confirmatory out the conduct and interpretation of thorough QT studies.
efficacy dose range finding studies. If the formulations are However, the true predictiveness of these studies for torsade
not bioequivalent, further multiple dose pharmacokinet- de pointes remains to be proven and it is likely that a
ics and pharmacodynamic studies may be needed to deter- number of useful drugs may be halted in development due
mined dosing regimens for the dose range finding studies. to an observation of prolongation of QTc that may not
Depending on the intricacy of the formulation develop- represent a risk of dangerous arrhythmia to patients.
ment, comparative bioavailability studies may need to be
performed on more than one occasion.
Special populations
Absorption distribution metabolism In order to provide accurate dosing information in the
product label, to cover administration to different patient
excretion (ADME) in man types in the target population, the pharmacokinetics and
(radiolabelled studies) safety are studied in patient subgroups. These include
The objective of the human ADME is to identify precisely elderly subjects, specific ethnic groups, subjects belonging
the handling of the drug by the body and to look for and to a defined genetic subgroup for metabolism (fast or slow
quantify metabolites that might also have effects on its metabolizers) and also subjects with hepatic impairment
efficacy and safety. The study involves giving a small and subjects with renal impairment. These studies are
number (usually four to six) of male subjects a dose of usually conducted once the clinically effective dose is fairly
radio­labelled compound and then sampling blood urine certain, as they are performed with the intended clinical
and faeces over a period commensurate with its elimina- dose. As many news drugs will be given to patients over
tion half-life. Unless information about metabolite pres- 65 years old, elderly subjects are required in order to assess
ence and activity is needed more urgently, the study is safety and kinetics in a group of healthy subjects more
usually performed in late Phase II or early in Phase III. representative of the target patient population. Specific
ethnic groups, on the other hand, will be required when
a drug being developed in Caucasian populations is also
Other safety pharmacology studies being submitted for approval in a different population
The clinical pharmacology programme also contains (e.g. Japanese). In this case, it is to determine whether
studies to examine safety aspects, notably the ability of the significant differences exist in the pharmacokinetics and

246
 Clinical development: present and future Chapter | 17 |

pharmacodynamics of the two ethnic groups, and whether any meaningful therapeutic activity in the case of first in
different dosage regimens will be necessary. Specific class compounds, or whether it has any useful differentiat-
genetic metabolic subtypes might be selected in order to ing features in the case of follow-on compounds. As the
ensure that where slow metabolizer subtypes exist this studies are not fully statistically powered to demonstrate
does not cause accumulation of the drug and related tox­ treatment differences it is important to determine a priori
icity in those individuals. Finally, as most drugs are elimi- what are the criteria for success or failure in the study.
nated via the kidney and/or the liver, alteration of the This requires objectivity and an experienced develop­
function of these organs could change the kinetics and ment team.
safety when the drug is given to patients with hepatic or The objectives of a large pharmaceutical company in
renal impairment; hence these studies are performed to conducting PoC with a novel mechanism of action may
make dosing recommendations for those patients. be different to those of a small pharmaceutical company.
For the large company, the most important aspect will be
Phase IIa – Exploratory efficacy to show as soon as possible that they have a novel mecha-
nism of action that is safe and well tolerated and works in
Objectives man, to feedback to the discovery teams working on the
The exploratory efficacy studies have different objectives back-up programme. To that end they may choose a drug
depending on whether the drug in development has a which is good enough for exploratory clinical trials, but
novel mechanism of action, i.e. is a first in class drug, or may have features that are not optimal for a final com-
if it has a known mechanism and others in the class are mercial medicinal product, for example a short half-life or
already available for patient use in the indication, i.e. a poor solubility. However, once PoC has been demon-
‘follow-on drug’. In the case of the former, The ‘proof of strated with the novel mechanism of action they can accel-
concept’ (PoC) studies in Phase IIa will be the first time erate development of back-up molecules, which have
the drug is tested in patients and this is when scientific better properties to become medicinal products. On the
theory is tested in clinical reality. It is interesting for the other hand if the PoC fails, then this line of research can
non-clinical and the clinical teams to see how transla- be abandoned.
tional the animal models turn out to be in patients. As it For a small company with more limited resources and
is the first time the drug is given to patients, safety evalu- fewer pipeline programmes, the Phase IIa PoC studies may
ation is the key objective. The design of PoC for novel be make or break for the entire company. Start-up compa-
mechanisms is crucial to ensure that meaningful clinical nies funded by venture capital may only have one lead
effect can be identified, if it exists and likewise if there is product going into Phase IIa, no products on the market
no clinical effect this also needs to be identified, so that a and other compounds in their pipeline may well be a long
decision about the future of the drug in the indication can way from the clinic. In that case a great deal of the com-
be determined. pany’s fortune rides on the drug being tested in Phase IIa
In the case of a follow-on drug with a known mecha- becoming a medicinal product. The small company may
nism of action, the PoC studies will be more orientated well have done their best to select a molecule with good
towards establishing points of differentiation with drugs drug-like properties which they believe could make it to
on the market that have the same mechanism of action. the market, prior to entering into man. The objectives of
The existing product may have weaknesses of efficacy, side Phase IIa for this type of company are to demonstrate
effects or pharmacokinetics which the follower drug can convincing therapeutic effect with acceptable safety so that
improve upon. An example of this is the class of oral the compound can be advanced further into development.
triptans for acute treatment of migraine, where speed of It is not generally the intention to go back to other mol-
onset of action and duration of action, manifest by a lower ecules that might look a bit better, as in the case of large
headache recurrence rate, were differentiators of interest companies, unless there is a major problem with the lead
compared to the market leader sumatriptan. Follow-on compound. Whatever the size of the company, the basic
compounds in the class concentrated on having either a tenet of Phase IIa is to gain robust data upon which to
more rapid onset of action or a long half-life leading to base decisions about the future of the drug. Given that the
lower headache recurrence. The design of studies for PoC studies are fundamental to the go/no-go decision for
follow on compounds will be orientated towards drawing an entire development project in both large and small
out these differentiating features rather than answering the companies, the design of the studies is crucial.
question ‘does it work or not?’
Design
In most cases, a double-blind placebo-controlled study is
Design consideration for first in class
an ideal approach as this allows for a more objective
compounds – large pharma vs small pharma measure of efficacy and safety. However, there are notable
Phase IIa PoC studies are small in size and designed care- exceptions, for example in oncology, where a comparison
fully to answer the questions about whether the drug has with, or add-on to a standard therapy is usually needed as

247
Section | 3 | Drug Development

it would be unethical to withhold known effective treat- If the compound is poorly soluble more formulation
ments from patients with cancer. It is possible to perform development will have taken place prior to entry into
open-label small ‘look see’ studies in Phase IIa using his- man. Nevertheless, in Phase IIa it is not usual to use a
torical comparison with efficacy of other drugs used in the final commercial type formulation as the results of the
indication, but this is not an ideal strategy as there is a initial clinical studies do influence the final formulation
substantial risk of bias if there is no control arm. The development.
dosing regimen will have been determined from the Phase The duration of Phase IIa studies for novel drugs is
I studies and also from supporting non-clinical pharma- influenced by the indication, but is usually shorter than
cology data. There are numerous possibilities for choosing for larger scale later phase studies, for two reasons. Firstly
the dose, but the principal objective is to be as sure as pos- it is important to gain information on clinical effect
sible that the dose or doses chosen for the study have a as soon as possible so that decisions on compound
high likelihood of showing an effect if there is one and development can be made. Secondly, the toxicology pro-
that the doses have acceptable safety and tolerability. A gramme to support the shorter studies, e.g. up to 1 month
maximum tolerated dose approach is often used, espe- in duration, will also be shorter. Clinical trials of 3 and 6
cially for drugs with good safety and tolerability. Doses can months dosing duration require toxicology studies in two
be selected to achieve plasma concentrations that have species of corresponding duration to support the human
been shown to be effective in animal studies or which are dosing. To make this type of investment in a toxicology
likely to give a particular receptor occupancy if PET studies programme for a novel mechanism of action drug before
have been performed previously. The number of dose its clinical effects are known, is not always desirable, par-
groups in Phase IIa studies is usually limited to one or two ticularly for smaller companies. For follow-on compounds,
active groups and a placebo group. The treatment groups however, studies of more than 1 month in duration might
may be parallel or the study can be done as a crossover, be needed in Phase IIa for indications in which clinical
which is when the patient is randomized to receive each of differentiation will only become apparent after longer-
the treatments in a random order, separated by a suitable term administration, for example in Parkinson’s disease or
washout period. The advantage of a crossover is that the psychiatric indications.
patient acts as his/her own control and the number of
patients can be reduced. The disadvantage is that there can Patients to be studied
be an order effect where the previous treatment affects the The PoC is usually the first time that the drug will be used
outcome of the subsequent treatment. Also crossovers are in patients with the relevant disease; therefore, careful
generally more suitable for indications where the outcome patient selection is vital to the success of a PoC study,
variables are rapidly measurable, e.g. pain, or have objec- especially those of novel compounds. It important to be
tive endpoints, e.g. blood pressure. Crossovers are less clear about what question is being asked in the study and
suitable for indications where the efficacy measurement is what patient group should be targeted so that question
highly dependent on patient reported outcomes, e.g. can be answered. Typically in the exploratory phase, as the
anxiety, because they are more prone to bias from order population is small, a more homogeneous patient group
effects. As the patients have to have more than one treat- is selected to reduce variability that might dilute the pos-
ment the crossover study may also take longer to complete sibility of seeing a result. The inclusion and exclusion
than a parallel group. However, if patient recruitment for criteria at this stage may be more restrictive than at later
a particular indication is difficult (for example rarer dis- stages of development. In the exploratory efficacy phase
eases or highly competitive therapeutic trial areas) a study less is known about the safety of the drug in diverse clini-
with a larger parallel group population may take longer to cal situations and the drug–drug interactions will not have
complete. As they are exploratory, Phase IIa studies lend been fully explored. A subgroup of patients within a
themselves to adaptive trial designs where the dosing may disease entity who are considered most likely to show a
be adjusted during the study according to the response of benefit can be studied at this early stage. For example,
the study population. Adaptive trial designs are discussed to evaluate the benefit of reflux inhibition with a novel
in more detail below. drug in patients with gastroesophageal reflux disease
In Phase IIa, as clinical data are required rapidly, the (GORD), patients with non-erosive reflux disease and
preclinical programme may well be lean, and focused on classical symptoms of heartburn and regurgitation would
getting only the data needed to support initial study in be selected, in order to increase the likelihood of seeing a
humans, therefore the formulation development is likely response to the drug. Once convincing efficacy and satis-
to be incomplete. If the compound is highly soluble, an factory safety/tolerability have been demonstrated, subse-
intravenous or simple oral formulation (solution or quent studies can include patients with more severe
simple tablet) can be used in early Phase I and Phase IIa. disease, or more diverse symptoms. In the case of GORD
Information from these studies, in particular the PK-PD this could extend to patients with erosive oesophagitis and
data, will help with the development of a commercial type symptoms other than heartburn and regurgitation that are
formulation, to be used in confirmatory efficacy studies. thought to be due to GORD.

248
 Clinical development: present and future Chapter | 17 |

Outcome measures
Phase IIb to III dose range finding
and confirmatory efficacy studies
Thorough safety and adverse event monitoring is per-
formed in these initial patient studies, with close attention Objectives
paid to looking for untoward effects in patients that might
The purpose of this phase of development is to confirm
not have been seen in healthy subjects. With regards to
the initial efficacy seen in the Phase IIa PoC studies and
efficacy, as the PoC studies have a relatively small popula-
to provide the pivotal data which will support the applica-
tion and short duration, it is necessary to have robust
tion for market approval, as such extensive, comprehen-
outcome measures. In the exploratory phase it can be
sive data from well-designed and adequately powered
tempting to try and answer a lot of questions in one clini-
studies are required. The data from the confirmatory
cal trial. This is not a good idea because putting too much
studies will be used to make the claims in the product
into the trial increases the complexity of its execution and
label upon which the drug can be promoted. Therefore,
can confound the interpretation of the data. It is by far
during this part of the development, the clinical team
better to have one clear principal objective and, at most,
needs to liaise closely with Regulatory and Marketing so
two smaller less important objectives. The ease of having
that the trials can be designed to fulfil the desired product
robust measurable endpoints depends a lot on the
label. Usually the first step is a Phase IIb dose range
indication. Those indications with objective measures,
finding study to confirm the preliminary safety and effi-
e.g. hypertension and diabetes, are straightforward to
cacy data and to explore the dose–response relationship
demonstrate in terms of proof of concept. However, for
in detail, to select the dose that has the optimal efficacy
licensing in such indications it is not sufficient to show
and safety profile to be used for large-scale Phase III effi-
that the drug lowers blood pressure or glucose alone, it
cacy and long-term safety studies. The eventual marketing
has to be shown that the drug improves patient outcome
strategy is a major guiding factor for the design of the
in reducing mortality or cardiovascular morbidity, which
Phase III programme, especially for follow-on drugs where
presents a substantial challenge in late phase develop-
product differentiation to existing treatments, whether it
ment. In the middle are pain-type indications, as pain
be on efficacy, safety or cost effectiveness, is crucial to its
is rapidly treatable. However, as one is relying on the
success.
patient to report the severity of pain, the outcome can be
subject to bias and a high placebo response. For other
indications where long-term treatment is needed to evalu- Design
ate the full benefit, a good surrogate marker can be used The standard design of Phase IIb dose-finding studies is
in the PoC. One example would be in the case of the treat- parallel-group randomized double-blind and, placebo-
ment of rheumatoid arthritis where measuring inflam­ controlled and/or active comparator controlled. The
matory mediators in the blood can indicate evidence placebo group in theory allows for a comparison of active
of meaningful pharmacological effect. Another example intervention with ‘no treatment’ and, therefore, should
comes from a study of the prevention of vasospasm post provide a clear view of the efficacy benefits and safety of
subarachnoid haemorrhage, with an endothelin antago- the test drug. However, it is well documented that patients
nist. Ultimately it needs to be known if treatment with the in clinical trials do have a response to placebo (Beecher,
drug improves patient outcome, i.e. mortality and morbid- 1955). Strictly speaking, taking placebo does not mean
ity, but in the PoC trial the occurrence of vasospasm was that the patient has no treatment. The very act of taking
measured with angiography, transcranial Doppler and the part in a clinical trial and receiving extra medical attention
presence of infarcts on CT scanning, to see if the drug was can contribute to an improvement to a patient’s underly-
able to prevent physiological vasospasm and its anatomi- ing condition. This is particularly true for psychiatric con-
cal sequelae. ditions or those exacerbated by stress or anxiety. Trials in
Pharmacokinetic samples may also be taken in the psychiatry and those which rely heavily on patient reported
Phase IIa studies to examine the PK-PD response and also outcomes have notoriously high placebo response rates.
to see if the pharmacokinetics in patients are different Occasionally the enthusiasm of the investigating physician
from healthy subjects; migraine patients, for example, for the new treatment can colour their view of the efficacy,
have gastric stasis during an attack and this can alter the which can also contribute to high placebo response rates.
absorption of orally administered drugs. Phase IIa may be In some medical conditions, for example oncology or
the first time that the PK–PD relationship is studied and other serious conditions for which effective treatment
the information can be used to guide the dose selection exists, it is not appropriate for patients to receive placebo.
for Phase IIb. In this case the new drug can be tested either against a
In summary, the characteristics of Phase IIa PoC studies gold standard, single active comparator or against a
are short, focused, using carefully selected patient popula- standard-care treatment regimen. In the latter case the test
tions and robust endpoints to enable go/no-go decisions drug might also be added to standard care in one group
for new drugs in development. while the other group just receives standard care alone.

249
Section | 3 | Drug Development

Another way to minimize placebo response in the active studies. If the Phase IIb study is successful it will have
treatment phase, is to have a single blind, placebo run-in identified the dose level and dosing regimen which gives
period, where all patients take placebo but only the inves- the optimal balance between efficacy, safety and tolerabil-
tigator knows that they are on placebo. At the end of the ity. This dosing regimen is subject to confirmation in
period, those who continue to fulfil the predefined eligi- Phase III.
bility criteria can be randomized to the double-blind
phase. For pivotal efficacy Phase IIb or III studies in Patients and study setting
Europe, it is a requirement to have at least one study with For the later phases of development it is important to
an active comparator arm. The choice of comparator will study the drug in a more ‘real-life’ setting, because the
depend on marketing considerations and what the poten- safety and efficacy data from these studies has to support
tial advantages the new drug can offer, whether it be in the use in the patient population in the target market.
terms of safety, tolerability, efficacy or cost effectiveness. Therefore, the trial eligibility criteria are expanded in
Active comparator studies may be designed to show supe- Phases IIb and III to enroll patients that are more repre-
riority, equivalence or non-inferiority compared to the sentative of the final target population. By this stage more
gold standard treatment. The choice will depend upon the safety and drug interaction data are available, so reducing
efficacy and the safety of the drug. For example, if the drug the restrictions on patient recruitment can be done with
is thought to have similar efficacy but a superior safety more confidence. The countries, study sites and doctors
profile to the existing treatment an equivalence or non- selected for the late phase studies will generally be more
inferiority design may be chosen for the primary efficacy, diverse than in Phase IIa and influenced by marketing as
because the main objective is to demonstrate the better well as regulatory needs. It is perfectly possible to apply
safety and tolerability. The treatment selected for compari- for a product licence in a country or region without con-
son has to be justified to the regulatory authorities when ducting pivotal efficacy trials there, because the trials are
submitting the trial application and can be discussed with conducted according to ICH GCP so the trials should be
them in advance, for example at an end of Phase II acceptable as long as the data generated cover any intrinsic
meeting. The only exception to this is for indications (genetic) or extrinsic (e.g. diet, medical practice) differ-
where no treatment is currently licensed. An example of ences that exist in that region. However, for the major
this is levodopa-induced dyskinesia in Parkinson’s disease. territories of USA, Europe and Japan it is usual to conduct
Even so, the non-use of an active comparator needs to be at least one study in that region and, for Japan, bridging
justified in the clinical trial application. studies may be needed to demonstrate that the properties
These trials form the basis of the marketing authoriza- of the drug demonstrated in a ‘European type’ population
tion application and so they have to have a sufficient are applicable to Japanese patients.
sample size to demonstrate efficacy with confidence. Also
sufficient safety data to support exposure to the target Outcome measures
patient population needs to be generated. Statistical cal- In the large-scale confirmatory studies the surrogate end-
culations are made to ensure enough patients are enrolled points of Phase IIa give way to demonstrating clinically
to have robust efficacy results that will support the desired meaningful benefit. This means that simply demonstrat-
claims on the product label. In the case of a parallel group, ing a reduction in blood pressure, or preventing cerebral
multiple, dose-range finding Phase IIb study, although the vasospasm or reflux events from occurring, to name but a
power of the study might be lower than required for a few examples, has to be shown to provide a benefit on
Phase III pivotal efficacy study (e.g. 85% rather than 90%), disease-free survival or a benefit to the patient’s ability to
sample size calculations have to take into account adjust- function in their daily life. It also has to be shown in many
ment for comparisons of the multiple dose arms. Typically cases that the new medicine will be cost effective. This
the studies will have several hundred patients or in the means that the outcome measures in Phase IIb and III are
case of some cardiovascular intervention (e.g. hyperten- more orientated towards patient and physician reported
sion) studies, a few thousand patients may be required. outcomes. For example, in oncology tumour markers used
The dosing duration must be sufficient not only to dem- in Phase IIa will give way to demonstrating survival over
onstrate efficacy but also to establish safety. For indica- a predefined period of time. In hypertension, the measure-
tions where long-term chronic use is intended, even if it ment of blood pressure may need to be accompanied by
is intermittent dosing, the safety database should contain data on the incidence of cardiovascular morbidity and
information on dosing for at least 12 months in an ade- mortality, and in neurology and psychiatry there are a host
quate number of patients. These data are often generated of validated questionnaires to demonstrate meaningful
by enrolling patients from the Phase IIb and III studies clinical benefit. In addition, in many cases pharmacoeco-
into a long-term, open-label, extension study at the nomic data needs to be collected, if not to justify market-
intended clinical dose. ing authorization, then to justify the pricing of the drug
The dose range for Phase IIb will be selected based upon to reimbursement committees and managed care plans in
the efficacy and safety data from the Phase I and Phase IIa the various countries where the drug is to be sold. As these

250
 Clinical development: present and future Chapter | 17 |

reported outcomes have various factors that can influence The Phase IV studies, which take place once the product
them the sample size needed to show a positive effect is is on the market, may take the form of collecting safety
generally much larger than when objectively measured data as required by authorities, but they may also be used
surrogate markers are used. For many indications where to collect additional simple efficacy data which are used
treatments exist there are outcome measures that are rec- to support the marketing effort. Phase IV studies are gener-
ognized by the regulatory authorities as being the gold ally large in scale, conducted at widespread sites but
standard efficacy measure for the particular indication. usually of simple design, and are orientated towards
These measures are not always appropriate for drugs with looking at how the drug is used in everyday practice within
novel mechanism of action or for a subset of patients the confines of the existing product licence. As for any
within an indication. It is possible to stray from the path clinical trial, Phase IIIb and IV studies are subject to the
of the standard efficacy measure and use outcomes more conditions of ICH GCP.
adapted to the clinical scenario, but it requires good jus-
tification and careful negotiation with the relevant com-
petent authorities. Bridging studies
Data generated in the USA and/or Europe and used to
support marketing approval of the drug in Japan present
Phase IIIb and IV studies
more problems, owing to the more significant intrinsic
The data included in the submission package from the (e.g. genetic) differences between Caucasian and Japanese
Phase I to III studies is very comprehensive; however, it is populations. An example is the greater frequency of poly-
not possible to guarantee that all adverse effects that could morphisms in the cytochrome P450 enzyme CYP2A6
be observed when the drug is on the market, have been (Oscarson, 2001) seen in Chinese and Japanese popula-
identified in the pre-marketing studies; especially for those tions, and the related differences in nicotine metabolism.
events that occur with an incidence of less than 1 in It is therefore usually necessary for a study to be carried
10 000. Indeed there are many cases of drugs being with- out to ‘bridge’ Caucasian data into Japan, i.e. to test the
drawn from the market for safety reasons, or having sig- validity of Caucasian data in a Japanese population. This
nificant labelling amendments applied to them, once is done by studying the drug’s pharmacokinetics, and
more information has become available. Regulatory usually its pharmacodynamic properties, in both popula-
authorities recognize that to ask companies to collect tens tions and analysing the results to determine whether they
of thousands of patients’ safety data in the initial submis- are comparable.
sion package would be costly, time consuming, and could In considering the issue of extrapolating drug safety and
cause unnecessary delay, or even prevent the marketing of efficacy data from one population to the other, the objec-
useful new medicines. Therefore, in order to protect tive is to do this safely while not repeating clinical trials
patient safety once the drug is on the market, the authori- unnecessarily. Specific ICH guidelines exist for the extrapo-
ties request that the companies perform post-marketing lation of data from one ethnic group to another (ICH, E5
safety surveillance, through specific studies designed to latest update 2006). There are several strategies and
evaluate safety of the marketed drug and by means of filing approaches for exploring ethnic differences in drug
Periodic Safety Update Reports (PSURS). The data for response, but none is appropriate for all circumstances
these reports are collected by the company’s pharmacovig- and each situation must be carefully evaluated, with a
ilance group, whose members are specially trained in drug detailed understanding of the requirements of the target
safety surveillance. The reports are required to be filed at regulatory authority and its acceptance of foreign data.
regular intervals and include data from ongoing trials and
spontaneous adverse event reports, which can arise from
a variety of sources, e.g. doctors, patients, clinical trials, Patient recruitment in efficacy studies
journal articles, etc. In exploratory efficacy studies, as the efficacy and safety in
Clinical trials collecting additional data can be started the patient population are yet to be determined, it is desir-
while the initial marketing authorization submission is able to have a relatively homogeneous population in order
being reviewed. Authorities will sometimes accept dossiers to minimize variability which might confound the results.
for review with limited duration safety data if the company The entry criteria for patients in exploratory studies will
continues the collection of data during the review period therefore be more stringent than in the later phases where
and has sufficient patient exposure by the time the author- it is then desirable to study patients who are more repre-
ity comes to make their decision. Studies conducted in this sentative of the population, which will use the drug once
peri-approval period are known as Phase IIIb studies. If a it is on the market. Patients can be recruited from the
company wishes to expand the indication, they may also patient pool known to the clinic if they are regular outpa-
conduct additional pivotal efficacy studies in the peri- tients or if they have already participated in a clinical trial
approval period or shortly after product launch and these in the indication. In some studies the patients may already
studies also fall into the category of Phase IIIb. be in the hospital (e.g. in studies of interventions in acute

251
Section | 3 | Drug Development

medical emergencies, such as myocardial infarction, important. Differences in metabolism exist between adults
stroke, or inflammatory bowel disease). Patients may be and children and at different times in childhood, e.g.
referred to the investigator from other clinics, or it is pos- during adolescence, that may alter the kinetics or safety of
sible to source patients externally by means of advertising. the adult medicine. Also formulations used by adults
There are various possibilities for advertising. Media such many not be suitable for use in children. The purpose of
as radio and television are popular in the USA but less so the plan, therefore, is to ensure that companies will
in Europe, partly because the cost is not always commen- develop medicines that can also safely and effectively
surate with the return. Publicity in local papers and posters treat important conditions in children. The draft plan is
and fliers in hospital or general practice clinics are often usually requested by the end of Phase II. The timing of the
effective, as are advertisements on a hospital website. The start of the paediatric trials is usually within a year after
appropriateness of external advertising will depend on the the product licence for the adult use has been granted.
indication. For common, straightforward ailments, such as However, this can be varied according to the indication
migraine or allergies, external advertising can be a good or need. Exemptions to paediatric development can be
way to recruit patients, but for more complex indications, obtained if the disease does not exist in children, for
for example in neurology or oncology, it may not be example Parkinson’s disease or Alzheimer’s disease, or in
appropriate. The disadvantage of advertising is that the certain age categories, e.g. migraine, which does not really
patient coming in ‘off the street’, is not already known to occur in children less than 6 years old.
the investigator and so a thorough check of the patient’s
medical history and trial eligibility, including likelihood
of good compliance, needs to be made. All advertising
REGULATORY AND  
materials, including the scripts for radio and television
advertisements, have to be reviewed and approved by ETHICAL ENVIRONMENT
ethics committees before they can be used. Patients who
take part in the efficacy trials may anticipate deriving some In most parts of the world, clinical trials are a legal require-
clinical benefit; therefore, they are not volunteers as such, ment before a new drug can be sold or any claims made
and cannot be paid for their participation, as is the case for its therapeutic benefit or safety. All clinical trials,
for subjects in Phase I studies. However, the patients can including Phase I studies, are subject to international,
receive a modest sum to cover out-of-pocket costs for national and, sometimes, also local regulation. Interna-
transport and subsistence associated with the study clinic tional regulatory requirements for human administration
visits, the amount of which has to be approved by the of a new active substance (NAS) are set out in a series of
ethics committee. guidelines published by the International Committee on
Harmonization (ICH; see Chapter 20). This committee
was formed to harmonize the regulation of clinical trials
in the three major pharmaceutical development regions
CLINICAL TRIALS IN CHILDREN (European Union, USA and Japan), with the aim of avoid-
ing duplication of clinical research programmes when
As part of the overall clinical development plan it is now applying for approval in all three ICH regions. National
a requirement in the EU and USA to include a paediatric and local regulations still exist and may vary from country
drug development plan. The objective is to have thorough to country within these regions, but ICH guidelines still
testing of the safety, pharmacokinetics and efficacy of apply and usually have the force of law. Clinical develop-
drugs that may also be used in children, so that correct ment of new drugs for registration in any of these regions
dosing information, using an appropriate formulation, must comply, and be seen to comply, with ICH guidelines
can be given for this population. In the past, companies if the data are to be accepted for registration purposes in
tended to shy away from testing their drugs in children the ICH regions, irrespective of where in the world they
due to ethical concerns and also because the paediatric were generated. This means it is not possible to sidestep
market was not seen to be particularly commercially the requirements of the ICH guidelines by developing
attractive. Adult drugs were used off-label in children, drugs outside the ICH regions in a manner that would not
often by extrapolating the adult dose to the weight of the comply with them. As the EU, USA and Japan represent
child with no proper guidance in the product label for the three biggest world markets for new drugs, there is
paediatric use. However, there are many medical condi- little incentive for non-compliance.
tions that occur in children as well as in adults. Examples All human studies are performed according to strict
of drugs commonly used by children as well as adults ethical requirements. The Declaration of Helsinki (World
include anti-epileptics, asthma drugs, anti-inflammatory Medical Association, 2008) and ICH Guidance E6 (CPMP/
and anti-infectives. The therapeutic margin of some of ICH/135/95) 2002 set the rules for all clinical develop-
these compounds can be quite narrow and so precise ment. In one sentence, the message comes through: ‘It
information about safety and dosing in children is very is the duty of the physician in medical research to protect

252
 Clinical development: present and future Chapter | 17 |

the life, health, privacy, and dignity of the human subject’. diaries, questionnaires, etc.) and decide whether it is justi-
The major points of the Declaration of Helsinki and ICH fied on ethical grounds. They evaluate the study in terms
E6 Good Clinical Practice are summarized below: of the risk to the subjects, the appropriateness of any
• The risks and potential benefits must be assessed remuneration offered to both subjects and the clinical
before trials are initiated, and the benefits must investigator, the design of the study and its ability to fulfil
outweigh the risks. its primary objective(s), the qualifications, experience and
• The interests of the individual study subjects must clinical trial performance of the clinical investigator, the
take precedence over those of science or society. text of any advertising used to recruit subjects, etc., and
• All trial subjects must freely give their informed approve or reject the proposed trial on the basis of these
consent prior to participation. ethical considerations. If it is approved, the EC/IRB
• Trials must be scientifically sound and clearly remains involved with the trial until it is completed: for
described in a trial protocol. example, it must be informed of any serious adverse events
• The trial must be carried out according to the (see below) that occur, and of any other major issues that
protocol, which must be reviewed and approved by a cause concern during the study. Changes to an approved
properly constituted ethics committee. protocol may not normally be implemented without the
• Only properly qualified physicians may provide EC/IRB approval.
medical care to trial subjects, and all other staff In the case of regulatory authorities the degree of
involved in clinical trials must be appropriately involvement varies from country to country (see Chapter
educated, qualified and experienced for the tasks 20), some reviewing all of the available data on the drug
they carry out. and some only requiring notification that EC/IRB approval
• Human administration must be supported by the has been given and that the trial will take place.
results of adequate preclinical testing in compliance
with ICH guidelines for the administration of drugs
to man. Clinical trial operations and  
• Data from clinical trials must be recorded, handled quality assurance
and stored in a way that allows accurate reporting,
interpretation and verification. As described above, all clinical trials have to be conducted
• Trial subjects’ privacy and confidentiality must be to ICH Good Clinical Practice (GCP) by investigators and
respected and assured. staff properly trained in the conduct of clinical trials. The
• The material to be administered must be of sponsor has a duty to ensure that the trials are being con-
acceptable quality and purity, as defined by the ducted according to GCP both at the investigational site
relevant ICH guidelines. This means both the drug and within the sponsor’s organization. The constraints of
substance and the formulated drug product must be ICH and their local rules place a large organizational and
manufactured in compliance with ICH Guidelines administrative burden on the trial centres, which need to
(2000) for good manufacturing practice (GMP) and be selected with care, based on inspection of their facilities
used in accordance with the trial protocol. and an assessment of the investigator’s and site staff’s
• The trial must be registered in a publicly available qualifications, experience and availability to carry out the
database prior to the first subject being recruited. study to the standard required.
The sponsor has specific responsibilities to train those
involved in the trial in the requirements of the study pro-
tocol and ICH GCP standards, and then to monitor the
Ethical procedures
project on site to ensure compliance monitor the perform-
The safety of subjects is always the paramount considera- ance of the sites during the study. Study staff training and
tion in clinical trials. There are two main bodies responsi- monitoring may be performed by the sponsor company
ble for evaluating proposals for trials: the national drug or delegated to a subcontractor, i.e. a clinical research
regulatory authority in the country where the trial will take organization (CRO), in which case the Sponsor has to
place, and the local ethics committee (EC) or, in the USA, oversee the activities of the CRO to ensure and be seen to
the institutional review board (IRB) of the clinic in which ensure, that they are compliant with GCP. The study moni-
the study will be performed. All clinical trials must be toring is performed by specially trained clinical monitors
approved by a properly convened and correctly function- who visit the sites regularly throughout the trial and, using
ing EC or IRB before they can be initiated. Ethics com­ specific written procedures, verify the study data, check
mittees are composed of independent experts and lay that the study is being conducted in accordance with the
members, who review the proposed study (protocol, protocol and that there are no breaches of GCP. To further
investigator’s brochure, insurance arrangements, patient ensure compliance, additional quality assurance (QA) is
information sheet, informed consent documentation and carried out in the form of an independent audit. Study
any other patient facing materials, e.g. electronic or paper sites will be selected for auditing. The clinical monitor is

253
Section | 3 | Drug Development

instrumental in identifying the sites for auditing. Typically inevitably means that good news receives more publicity
those sites who are high recruiters or who have had par- than does bad news. The most recent Declaration of Hel-
ticular problems with the study are selected. In addition, sinki, from 2008, states that authors should publish trial
if the monitor suspects any irregularity in the data, or even results or make them publicly available whether they are
fraud, a site may undergo a ‘for cause’ audit. Auditors are positive, negative or inconclusive. GlaxoSmithKline, has a
responsible for thoroughly inspecting the data and docu- publicly available database of trial results summaries for
mentation files to validate the site and the data it produces its marketed drugs and www.clinicalstudyresults.org also
in terms of ICH compliance. contains study results for marketed compounds. For drugs
Finally, the regulatory authority itself may choose to that fail in development there is no obligation to put the
inspect one or more clinical sites and/or the files of the results into a public database, nevertheless, there is increas-
sponsoring company or its CRO, as a final check on data ing pressure for companies to publish data about their
quality. Issues arising at this late stage cast suspicion on drugs in development whether positive or negative. Such
the whole dossier and can delay or prevent the granting of transparency could help drug development in general as
marketing approval. A detailed overview of the regulatory much can be learned from the development programmes
requirements for product development and licensing are of drugs that do not make it to market.
given in Chapter 20.

Issues of confidentiality   SEAMLESS DRUG DEVELOPMENT

and disclosure WITH ADAPTIVE CLINICAL TRIAL
DESIGN: THE FUTURE OF  
Since 2008 it has become a requirement for all clinical
trials to be registered on a publicly accessible database CLINICAL DEVELOPMENT?
before the first patient is enrolled. In the USA and
Europe clinical trials can be registered on http//www. The disadvantage of the conventional step-wise approach
clinicaltrials.gov or http//pharmacos.eudra.org which to clinical development where drugs move to the next
provide details of the purpose and plan of the trial, the Phase once the preceding Phase is finished is that it is
clinical centres and investigators involved, and the stage time-consuming, costly and drugs may fail in late Phase
the trial has reached. For every trial a registration number development. It is estimated currently that approximately
(International Standardized Randomized Controlled Trial 40% of drugs entering Phase III do not make it through
Number, ISRCTN) is assigned, allowing public access to registration (Di Masi et al., 2010). The success rate is some-
information on all prospective studies involving experi- what dependent on the indication, with GI, CNS and
mental and registered compounds. Its main aim is to avoid cardiovascular drugs having the lowest overall success rates
unnecessary repetition of clinical trials. These databases do (Di Masi et al., 2010). This is a great waste of resources for
not, however, include the results of completed trials, which the pharmaceutical industry and a missed opportunity for
frequently remain unpublished. medical practice. Late-stage attrition rates may be reduced
Regulatory authorities (see Chapter 20) require detailed by adopting a more seamless approach to the clinical
results of all clinical trials approved by them – including development programme using adaptive clinical trial
trials of marketed compounds – to be reported to them, designs so that ineffective drugs can be identified earlier
but this information is not, in general, publicly accessible, and their development terminated and those drugs with
except to the extent that the Summary Basis of Approval promising efficacy can be accelerated through develop-
(USA) or Centralized Evaluation Report (EU) that is pub- ment, with a higher likelihood of successful registration.
lished when a drug is approved, includes a summary of There are a number of ways in which the development
the clinical trials results on which the approval was based. stages can be merged or overlapped. Broadly speaking the
Clinical trial sponsors are under an obligation to report to development path is looking to have initial safety testing
the regulatory authority any safety issues that come to with exploratory efficacy leading to a go/no-go decision
light, and in the case of marketed drugs the regulatory and then confirmatory efficacy and safety, leading up to
authority may respond by altering the terms of the market- filing the registration dossier. It is becoming more common
ing approval (for example by including a warning in the for initial Phase I protocols to contain more than one
package insert, or, in an extreme case, by withdrawing stage. For example, within the same protocol it is possible
the approval altogether). Publication of trial results in the to do a single ascending dose study, a food effect crossover
open literature is not obligatory. Trial sponsors often at a dose determined from the SAD, and then, based upon
choose to publish their data in refereed journals, although these results, go into the multiple ascending dose study,
they are not obliged to do so, and both they and journal which itself might contain pharmacodynamic evaluations
editors tend to give preference to positive rather than nega- relevant to the desired indication. Using this combined
tive findings (Hopewell et al., 2009). Publication bias approach saves time in making repeated applications to

254
 Clinical development: present and future Chapter | 17 |

regulators and ethics committees. Provided the decision IIb dose range finding study, there needs to be sufficient
points to proceed to the next stage within the study are confidence that the results of exploratory studies will be
clear and subjects’ safety is maintained throughout, com- replicated in the larger scale trials at the doses chosen.
bination Phase I protocols can be approved at a single Adaptive design trials are those where an interim adap-
regulatory and ethical review. If efficacy in an indication tation of treatment or endpoint during the study is decided
can be shown with a single dose, an initial proof of before the study is started. The a priori protocol design
concept study in patients could commence after the SAD incorporates the changes within it. The FDA’s draft guid-
and be conducted in parallel with the MAD, thereby over- ance from February 2010 (FDA, 2010) defines adaptive trial
lapping Phase IIa with Phase I. This always assumes that design studies as ‘a prospectively planned opportunity for
safety and tolerability in the SAD were acceptable. A good modification of one or more specific aspects of the study
example of this is acute treatment of migraine, where effi- design and hypotheses based on analysis of the data
cacy can be shown with a single dose of study medication (usually interim data) from subjects in the study’.
to treat a single attack of migraine. For indications where The purpose is to make studies more efficient, either
repeat dosing and/or steady state plasma levels are consid- by having a shorter duration, fewer patients, more
ered necessary to see a treatment effect one can consider appropriate patients, increasing the likelihood of identify-
performing the MAD study in patients, again assuming ing drug activity if it exists, or making the study more
that the level of tolerability is acceptable for proceding informative, for example in understanding better the dose
straight to patients after the SAD. Most likely a relevant response, the patients, or part of the indication most likely
surrogate marker of efficacy will be required as the full to benefit.
effect on symptom control or control of the disease may The adaptations may include the following:
not be evident until after several weeks’ treatment. An
example of this could be treatment of gastro-esophageal
• Change of sample size to either increase or decrease
it according to results of an interim analysis
reflux with a reflux inhibitor. The effect of a study drug
on the occurrence of reflux events and transient lower
• Reallocate treatment distribution; dose groups may
be added or dropped according to predefined criteria
oesophageal sphincter relaxations, following a challenge
for doing so
meal, can be monitored with oesophageal impedance-pH
monitoring and manometry respectively, on a single day
• Changing the treatment duration
at steady state. Clinical symptoms could be monitored as
• Refining the study entry criteria: certain patient types
may be dropped if they appear to have no benefit,
a secondary objective because a significant effect on clini-
alternatively the population can be enriched with
cal symptom control would not necessarily be anticipated
patient types who appear to be deriving benefit.
with such a short duration study. In the case of oncology
studies it is routine to look for clinical effects in patients The types of adaptations used may differ according to
using surrogate markers in clinical pharmacology studies, whether the trial is in the exploratory or confirmatory
as it is not possible to test most anticancer drugs in healthy phase (Orloff et al., 2009). For example, surrogate end-
subjects. points such as biomarkers may be used to evaluate patient
Once a go/no-go decision has been achieved in the progress in an exploratory study, whereas clinical outcome
exploratory phase, dose range finding Phase IIb studies measures would be needed in a confirmatory study. In
can be combined with Phase III to achieve the objective an exploratory study it may be possible to use a single
of finding the optimal dose and having adequate, well- (patient) blind approach to reallocate treatments. An
controlled, pivotal efficacy and safety data. An example of example of this was a study of migraine. The investigator
this approach is to have a parallel group, dose range knew the treatment allocation but the patients did not.
finding study that has sufficient sample size and power so The first patient at each site was allocated to start on a
that it can be considered pivotal. Patients on the optimal mid-range dose of the treatment. If the patient had a head-
dose can then continue either into an open-label safety ache response, the next patient was allocated the next dose
phase or generate further efficacy and safety data in a down, if not, the next patient was allocated the next dose
double-blind controlled study extension Adaptive trial up. The doses were changed up or down by the investiga-
designs are useful in this regard, because the optimal dose tor according to each patient’s response. The results of this
could be selected from several dose arms on an interim study turned out to closely predict the dose that was
measure, and then the study continued with the optimal chosen as the optimal effective dose by a subsequent,
dose to achieve full power (Orloff et al., 2009). Using the conventional, parallel, dose-range finding study.
seamless Phase II to Phase III design, the second pivotal It is crucial that making the adaptations while the study
efficacy study can be started with the optimal dose as soon is ongoing does not compromise the integrity or outcome
as it has been identified and the start of this study would of the study. The FDA draft Guidance 2010 (FDA, 2010)
overlap with the end of, or continuation of the dose range lays out the major design and performance considerations
finding study. As the sample size for this type of study is to be taken into account for adaptive clinical trials. The
likely to be larger than a conventional stand-alone Phase study needs to be carefully designed and the protocol

255
Section | 3 | Drug Development

clearly written to incorporate any interim analyses, the throughout the development. Migraine attacks occur fre-
conduct of which must be performed in such a way as to quently in the study population and the gold standard
not compromise the study blinding, or influence interpre- efficacy parameter ‘pain free at 2 hours post dose’ is easily
tation of the data. Close liaison with statisticians, who will identifiable, even though it relies on the patient to report
have to model the effects of proposed adaptations on it. Indications where drug benefit takes months or years to
statistical outcomes, will be required when designing the become apparent and rely heavily on patient reported
study. Likewise the regulatory affairs group needs to be outcome are more challenging. This includes studies in
involved to help craft a protocol that will be acceptable oncology, neurology and psychiatry. Nevertheless, in the
and justify the use of the adaptive design to competent exploratory phase it may well be possible to perform an
authorities. This means that the study design and protocol adaptive design using surrogate markers, to improve the
writing stage may take longer than a conventional study, dose and patient selection for confirmatory efficacy in
but the reward should come in the form of a study that these indications.
may be shorter to perform and have a higher chance of a To summarize, the schematic in Figure 17.2A shows the
successful outcome. stages and timelines of a conventional clinical develop-
Adaptive clinical trials lend themselves better to some ment path and Figure 17.2B shows how a new form of
indications than others. The ideal scenario is to have an clinical development might look, using seamless develop-
indication where the effect of treatment is rapidly evident ment with adaptive clinical trial design.
and objectively measurable. For example, as blood glucose
and blood pressure are objective and respond rapidly to
effective intervention, Phase IIa studies in diabetes or
hypertension could use adaptive design to good advan-
CONCLUSIONS
tage, to rapidly identify doses and the target patient popu-
lation. Acute treatment of migraine is a good example Clinical development is the cornerstone of bringing new
of an indication where adaptive design could be used medicines to the market. It is a complex, highly regulated,

Phase 1 Clinical pharmacological

FIM, MAD, PK/PD DDI Form Dev/biovailability, DDI ADME, Special pops, Safety pharm

Phase 2A
Exploratory PoC

Go/NoGo

Phase 2B
Confirmatory DRF

Phase 3 Confirmatory – 2 Pivotal efficacy studies

MAA/NDA approval
with one year OL safety

Phase 3B/4

Year 1 Years 2 to 4 Years 4 to 6 Years 6 to 9 Year 10

Fig. 17.2  A, Conventional path of clinical development.

256
 Clinical development: present and future Chapter | 17 |

Phase 1 – Exploratory Phase 1 – Clinical pharmacological

Combined SAD/MAD
and PKPD and overlap Form Dev/biovailability, DDI ADME, Special pops, Safety

Phase 2A – Proof of concept

Use statistical
modelling, PK/PD Go/NoGo
data and adaptive
design to establish Phase 2B – Confirmatory pivotal
effect size doses DRF with OL safety extension
for P2B
Adaptive design to confirm optimal
dose and patient selection for P3
pivotal efficacy study

Phase 3 – Single pivotal efficacy and safety MAA/NDA

with OL safety extension approval Phase 3B/4

Year 1 to 2 Year 3 Year 4 Year 5 Year 6 Year 7

Fig. 17.2  Continued B, Seamless development with adaptive trial designs.

time-consuming and very costly part of the overall drug input end to improve the sourcing of the drug develop-
development process, with a substantial risk of failure. The ment pipeline through public and private partnership,
traditional path of clinical development hitherto has partnership with academia and partnership between small
served the pharmaceutical industry reasonably well and and large pharmaceutical companies and at the output
has enabled high-quality innovative medicines to be deliv- end, by modifying the traditional clinical development
ered to patients. However, the current process has a high pathway. It is hoped that adopting these new initiatives
attrition rate which is wasteful of financial, medical and will enable companies to bring new medicines to the
patient resources. New initiatives are being launched at the market and to patients more effectively and rapidly.

REFERENCES

Adams CP, Brantner VV. Estimating the drug development: success rates for critical path opportunities report and
cost of new drug development: is it investigational drugs. Clinical list. www.fda.gov/oc/initiatives/
really 802 million dollars? Health Pharmacology and Therapeutics criticalpath/reports/opp_report.pdf;
Affairs 2006;25:420–8. 2010;87:272–7. 2004b.
Beecher HK. The powerful placebo. EFPIA. The Innovative Medicines FDA (2010) FDA Guidance for
Journal of the American Medical Initiative. www.imi-europe.org; 2008. industry: adaptive design clinical
Association 1955;159:1602–6. FDA. US Food and Drug Administration trials for drugs and biologics. Feb
Di Masi JA, Hansen RW, Grabowski HG. 2004. Innovation or stagnation: 2010.
The price of innovation: new challenge and opportunity on the Hopewell S, Loudon K, Clarke MJ, et al.
estimates of drug development costs. critical path to new medicinal Publication bias in clinical trials
Journal of Health Economics products. www.fda.gov/oc/initiatives/ due to statistical significance or
2003;22:151–85. criticalpath/whitepaper.html; 2004a. direction of trial results. Cochrane
Di Masi JA, Feldman L, Seckler A, et al. FDA. US Food and Drug Administration Database Systematic Review 2009;1:
Trends in risks associated with new 2004. Innovation or stagnation: MR000006.

257
Section | 3 | Drug Development

Orloff J, Douglas F, Pinheiro J, et al. The nicotine metabolism. Drug World Medical Association. Declaration
future of drug development: Metabolism and Disposition of Helsinki. Ethical principles for
advancing clinical trial design. 2001;29:91–5. medical research involving human
Nature Reviews Drug Discovery Woodcock J, Woosley R. The FDA subjects. 59th WMA General
2009;8:949–57. critical path initiative and its Assembly, Seoul, October 2008.
Oscarson M. Nicotine metabolism by influence on new drug development. www.wma.net/en/30publications/
the polymorphic cytochrome P450 Annual Review of Medicine 10policies/b3/index.html; 2008.
2a6 (cyp2a6) enzyme: implications 2008;59:1–12.
for interindividual differences in

258
 Chapter 18 
Clinical imaging in drug development
P M Matthews

1. Target validation: does the chosen therapeutic target

INTRODUCTION potentially play a central role in determining the
disease or symptom of interest?
Preclinical and clinical imaging has already made a signifi- 2. Biodistribution: does the molecule reach the tissue
cant contribution to drug development. Almost 30% of of interest in potentially pharmacologically active
new molecular entities approved for neuropsychiatric concentrations?
indications by the FDA between 1995 and 2004 were 3. Target interactions: does the molecule interact with
developed with contributions from imaging (Zhang and the target of interest? What is the relationship
Raichle, 2010) and there will be a growing need in drug between administered dose and interaction with the
development for information provided by imaging. New target?
ways of thinking about clinical development are putting 4. Pharmacodynamics: what are the effects of the drug
a premium on integration of early biology and clinical and how long do they last?
development in the context of ‘experimental medicine’. 5. Patient stratification and personalized medicine: how can
Therapeutics development increasingly involves research the most responsive patient population be identified
that is hypothesis led, performed across levels of biological for more efficient clinical trials? How can a medicine
complexity (e.g. cells to the whole human), or inspired be given in the clinic to patients who will experience
by concepts translated across species (e.g. mouse to man). the greatest benefit?
Similar non-invasive imaging methods can be applied This chapter will provide a brief overview of clinical
in both clinical and preclinical applications. Regulatory imaging applications in drug development. Preclinical
authorities also are putting an increasing emphasis on imaging is a larger topic, beyond the scope of this
the need to have a deeper understanding of pharmacology review, although reference will be made to some aspects
and the biology of disease in approvals for new chemical of imaging for direct translation of pharmacological
entities, as well as the potential for delivering individual- hypotheses from preclinical to clinical applications. Pre-
ized therapies. The Critical Path Initiative (CPI) (www.fda. clinical studies also enable the validation of innovative
gov/ScienceResearch/SpecialTopics/CriticalPathInitiative/ imaging methods that can be applied to clinical develop-
default.htm) sets out a strategy for transforming the ment applications.
way FDA-regulated products – drugs, biological products,
medical devices and veterinary drugs – are developed
and manufactured. Central to this is developing better
evaluation methods (including specifically better imaging IMAGING METHODS
methods, as well as advancing genetics and bioinformatics
– see also Chapter 7) and applying these to the accelera-
Positron emission tomography (PET)
tion of development of personalized medicine.
A useful way of considering how imaging tools can PET imaging relies on the design and manufacture of
contribute to clinical drug development is through their radiolabelled ligands which can bind selectively to a target
roles in addressing major questions in new drug of interest with minimal non-specific binding. These
development: ligands are most typically labelled with positron-emitting

© 2012 Elsevier Ltd. 259

Section | 3 | Drug Development

The effect of the environment of the hydrogen atoms is

Table 18.1  Some examples of PET radioisotopes
useful in drug development applications
expressed as two parameters: the T1 and T2 relaxation
times. Changes in the way in which the radiofrequency is
applied and the signal received by the scanner, mix the
Radioisotope Half-life (min)
relative contributions of T1 and T2 parameters to signal in
15
O 2.1 different ways. This allows image contrast (grey scale vari-
11
C 20.4
ation) between different tissues (e.g. grey matter and white
matter in the brain) or regions of a heterogenous tissue to
68
Ga 68 be generated. Images thus allow tissue size and shape to
18
F 109 be measured and are sensitive to the state of tissue (e.g.
changing with evolution of the pathology of stroke in the
brain).
The signal of blood relative to tissue can be enhanced
on a T1-weighted imaged by intravenous injection of a
radioisotopes that decay with a relatively short half-life gadolinium chelate ‘contrast agent’ that alters local relaxa-
(Table 18.1). The short half-life allows high enough doses tion properties of water in the blood. Quantitative assess-
to be administered for a strong imaging signal without ments of signal change after the injection of contrast
substantially increasing long-term health risks associated agent provides one way by which MRI can measure blood
with the ionizing radiation. volume and flow. Leak of plasma across the vascular
PET imaging is based on the principle that emitted endothelium (e.g. with a damaged blood–brain barrier or
positrons collide with local electrons to produce pairs with tumour neovascularization) can be detected as
of photons that travel at 180° to each other and can be abnormal, sustained tissue signal enhancement after this
detected as coincident events by γ-detectors surrounding contrast administration.
the subject. The relative positions of detection of coinci- MRI images usually measure volumes between 1 and 5
dent events and their precise timing enable localization of mm3. Morphological measures can be conducted with
the original annihilation events for reconstruction of the high precision because of the high soft tissue contrast.
spatial distribution of the radiolabelled ligand. By follow- Moreover, the method is non-ionizing and without known
ing the time course of the emissions (and appropriate health risks.
instrument corrections) across different tissues, the rates
of delivery of the radiotracer and the amount retained can
be modelled. Functional magnetic resonance
Only microdoses of radioligands or other radiolabelled
molecules need to be used; PET is exquisitely sensitive and imaging (fMRI)
even nanomoles of labelled material (e.g. with 11C-labelling) fMRI is based on indirect measures of neuronal response
can be detected. However, spatial resolution is limited by being sensitive to changes in relative blood oxygenation
intrinsically by the distance over which the annihilation (Jezzard, 2001). Increased neuronal activity is associated
event occurs (and, in practice, is typically ~4 mm). The with a local haemodynamic response involving both
PET data can be co-registered with structural data from increased cerebral blood flow and blood volume. This
computed tomography (CT) or MRI images to localize the neurovascular coupling appears to be a consequence pre-
signal anatomically. dominantly of presynaptic neurotransmitter release and
thus reflects local signalling.
The most commonly used fMRI (or pHMRI) imaging
Magnetic resonance imaging (MRI)
method applies blood oxygen level dependent contrast
MRI imaging conventionally relies on the interaction of (BOLD). MRI is sensitive to changes in blood oxygenation
the weak magnetic dipole of the hydrogen nucleus with a because deoxyhaemoglobin is paramagnetic and, there-
strong applied magnetic field varying in a well-defined way fore, locally distorts the static magnetic field used for MR
across the body. Energy in the radiofrequency range mod- imaging. In the MRI magnet, the magnetic field is made
ulates this with a specific frequency that depends on the highly homogeneous, but the presence of deoxyhaemo-
precise magnetic field at each point in the body. As most globin leads to small magnetic field inhomogeneities around
hydrogen atoms in the body are in water (or fat), the blood vessels, the magnitude of which increases with the
frequencies of the mix of signals detected can be used to amount of paramagnetic deoxyhaemoglobin. A relation-
reconstruct tissue morphology as an image. Additional ship between neuronal activation and blood oxygenation
information comes from the intensity and duration of the is observed because blood flow increases with higher neu-
signal detected. These are determined by the concentration ronal activity and this increase in blood flow is larger
of hydrogen atoms (e.g. how much water or fat) and their than is needed simply for increased oxygen delivery
local environment, respectively. with greater tissue demands: the local oxygen extraction

260
 Clinical imaging in drug development Chapter | 18 |

fraction decreases with synaptic signalling. The signal, endophenotypes (heritable quantitative traits). For example,
therefore, is not a measure of blood flow directly. Note indirect evidence has suggested that glycogen synthase
also that these signal changes are small (typically, 0.5–5% kinase-3beta (GSK3β) and canonical Wnt pathway func-
at 3T). tion contribute to the molecular pathology of major
A typical experiment would involve acquisition of a depressive disorder (MDD). Brain structural changes also
series of brain images during infusion of a drug or over have been associated with MDD. To test the hypothesis
the course of a changing cognitive state (e.g. performing a that GSK3β is relevant to the disease, variations in brain
visually presented working memory task vs. attending to grey volume were associated with GSK3β polymorphisms
a simple matched visual stimulus). Regions of significant in a mixed population of healthy controls and MDD
signal change with drug infusion or between cognitive patients to demonstrate an interaction between genetic
states then are defined by statistical analysis of the time variation and MDD (Inkster et al., 2009). Supporting evi-
series of signal change. Quantitative measurement of this dence for a functional association also can come from
change allows measures relevant to drug action on the similar analyses linking brain structural variation to
brain to be defined. genetic polymorphisms related to genes encoding multi-
ple proteins contributing to the same signalling pathway
(Inkster et al., 2010).
Functional imaging methods can be used in similar
HUMAN TARGET VALIDATION ways. Patients with a history of affective disorders carrying
the S allele of the common 5-HTTLPR polymorphism in
Confidence in progression of drug development from the serotonin transporter gene (SLC6A4) have an exagger-
target validation in preclinical models (e.g. by demonstra- ated fMRI response (in the amygdala) to environmental
tion of a phenotype plausibly related to the human disease threat relative to L allele homozygotes (Hariri et al., 2002).
with knockout of the gene of interest) is often limited. The Other work has supported hypotheses regarding genetic
approach arguably is particularly problematic for chronic variation associated with other disorders. For example,
diseases, for diseases that are determined by the interac- polymorphisms linked to the genes DISC1, GRM3 and
tion of multiple biological factors and the environment COMT all have been related to imaging endophenotypes
and particularly for those with uniquely human pheno- for schizophrenia and associated with altered hippocam-
types (e.g. most neurological or psychiatric disorders). pal structure and function (Callicott et al., 2005), gluta-
This has brought an increasing interest in validation of matergic fronto-hippocampal function (Egan et al., 2004)
new therapeutic targets using experimental medicine and and prefrontal dopamine responsiveness (Egan et al.,
human disease ‘models’. Imaging supports this by provid- 2004), respectively.
ing a range of methods for directly assessing molecular Application of functional MRI approaches that define
interactions, biochemistry and physiology in humans neurobiological bases for general cognitive processes
non-invasively. To date, most applications have been to (such as in the context of psychiatric disease, motivation
targets for neuropsychiatric diseases. or reward) facilitate understanding of the general impor-
Human models can support target validation for tance of targets relevant to more than one disease. For
symptom management. For example, sleep deprivation example, fMRI approaches have contributed to the current
has been used as a model for mild cognitive impairment. appreciation for neural mechanisms common to addictive
FMRI can be applied as a probe for physiological changes behaviours across a wide range of substance abuse states.
specific to sleep deprivation-associated cognitive impair- Studies of cue-elicited craving have defined similar activi-
ment to enable assessment of any responses to a test agent ties of the mesolimbic reward circuit in a range of addic-
interacting with the target of interest (Chuah and Chee, tions (e.g. nicotine (David et al., 2005)). Combination of
2008). In this instance, the ability of fMRI to report quan- fMRI with PET receptor mapping on the same subjects has
titatively on physiological modulation in specific func- the potential to relate systems-level dysfunction directly
tional anatomical regions relevant both to diseases of with the molecular targets of drug therapies to further
cognitive impairment and the model (e.g. the hippocam- speed target validation in appropriate circumstances.
pus) adds specificity to associated behavioural measures. With the potential to define the relationship between
Modulation of impaired hippocampal activation during in vivo molecular pathology and disease expression, the
memory tasks after sleep deprivation by a molecule inter- relevance of a target can be inferred more confidently in
acting with a novel target provides compelling evidence some instances than is possible based on post-mortem
supporting validation of the target for symptomatic treat- studies only. For example, central to current therapeutic
ment of disorders of memory. hypotheses for schizophrenia is targeting of dopamine
An alternative concept for target validation in humans receptor signalling. PET imaging with a receptor-specific
involves testing for modulation of disease-related brain radiotracer allows the receptor densities and distributions
systems by allelic variation at candidate target loci. This to be mapped in vivo in patient and healthy control popu-
approach employs structural MRI or fMRI outcomes as lations. Using this approach, D2/D3 binding potential

261
Section | 3 | Drug Development

values have been shown to be abnormal in schizophrenics

in both striatal and extrastrial regions and to vary with age BIODISTRIBUTION
(Kegeles et al., 2010).
A limitation of the PET measure, however, is that it Microdialysis can provide accurate measurements of the
does not distinguish between effects of abnormal receptor free concentration of a drug in the brain or other organ or
availability or dopamine release (and neurotransmitter direct assays of tissue uptake can be performed on biopsies
occupancy of the receptor that reduces the free receptor or performed post mortem. However, because of its relative
available for binding to the radiotracer). To more specifi- inaccessibility, whether a drug intended for a CNS target
cally test the therapeutic hypothesis that dopamine recep- crosses the blood–brain barrier in sufficient amounts
tor antagonism is relevant to schizophrenia, dynamic to be pharmacologically active can be very difficult to
changes in receptor binding potential can be studied answer early in new drug development using conventional
before and after an intervention modulating dopamine approaches to Phase I and IIa studies (see also Chapter
release. For example, dopamine depletion leads to a larger 10). Confidence in extrapolation of measures directly from
increase in PET D2 receptor availability in patients with rodents to humans is limited (Figure 18.1). Recognized
schizophrenia than in healthy controls, suggesting a species differences in blood–brain barrier penetration are
higher synaptic dopamine concentration in the patients related to species-specific patterns of expression of trans-
(Kegeles et al., 2010). port enzymes, for example. PET provides the most general
The relevance of a target to symptoms or behaviours method for assessing distribution of a drug molecule.
can be validated in a similar fashion. For example, Imaging biodistribution can answer the question: does a
because dopamine is known to be an important mediator molecule reach the tissue of interest in concentrations
of the reinforcing effects of cocaine, it was hypothesized high enough to be potentially pharmacologically active?
that alterations in dopamine function are involved in The principles for a PET biodistribution study are
cocaine dependence. To validate dopamine receptor mod- straightforward. The time course of data from the blood
ulation as a therapeutic target for the treatment of drug and tissue allow the clearance from plasma to tissue (a
dependence, pre- and postsynaptic dopamine function function of the blood flow and the tissue extraction of the
were characterized by assessing the receptor-specific molecule from the blood) to be estimated. If the tissue
binding of a PET radiotracer in recently detoxified cocaine- uptake is low, then separate estimates of the blood volume
dependent subjects and related directly to drug-seeking and allowance in the kinetics for the amount of the
behaviour in the same group of subjects (Martinez labelled molecule in the blood at any point are needed.
et al., 2007). An important caution, however, is that it is only the

Fig. 18.1  Species differences in blood–brain barrier penetration. A novel CNS active molecule was labelled with 11C for PET
biodistribution studies in a rodent, pig and human. The studies illustrate how poorly predictive rodent studies can be regarding
distribution into the human brain. The small rodent brain lies at the arrowhead. A detected signal scale (standardized uptake
value or SUV) is shown to the right.
Images courtesy of the GSK Clinical Imaging Centre, London.

262
 Clinical imaging in drug development Chapter | 18 |

distribution of the positron-emitting isotope ‘label’ that is However, as the labelled drug molecule is used in micro-
being measured with PET. Information also is needed doses only, toxicity testing for IMP filing can be limited
regarding the concentration and the nature of any metabo- to studies in a single species, allowing even human
lites carrying the isotope that are generated during the volunteer studies to progress early in a drug development
imaging period and appropriate corrections made to the programme.
uptake model. While the greatest application of biodistribution studies
Molecules in the tissue will distribute to varying extents thus far has been in the development of CNS drugs, bio-
into tissue ‘free’ and ‘bound’ compartments. Binding can distribution studies should also play an important role in
be either specific (e.g. binding to a receptor) or non- optimizing anticancer drugs, as up-regulation of pumps
specific, reflecting, for example, lipophilic interactions or that may exclude drugs from tumours is well described.
the action of non-specific uptake mechanisms. In the When a pump mechanism is suspected, then the depend-
general case, a non-linear relationship between the relative ence of distribution on the dose of the unlabelled
tissue distribution of a molecule and the amount that is molecule can be explored. This approach was used to
specifically bound is expected. To define this, a kinetic characterize the brain and tumour distribution of temo-
analysis of the tissue compartment signal change over time zolomide, an alkylating agent used in the treatment of
is needed, ideally with respect to another compartment brain tumours. Normal brain and brain tumour temozolo-
in which there is similar non-specific, but no specific, mide concentration profiles were estimated for different
binding (a ‘reference’ region). temozolomide dosing regimens. A relatively impaired
Moreover, if the plasma free concentration of the plasma–tissue barrier in the tumour (presumed to be
labelled molecule is measured and it is assumed that the related to both breakdown of the blood–brain barrier and
molecule distributes passively, measures over the time tumour angiogenesis) was demonstrated for the drug
course to a steady-state distribution allow an estimation (Rosso et al., 2009).
of the tissue free concentration (Slifstein and Laruelle, There is a potential for integration of PET with conven-
2001). Defining the volume of distribution of a molecule tional drug metabolism and pharmacokinetic (DMPK)
along with measurement of the plasma free concentration radiotracer studies of new molecules, although it is not an
allows the occupancy of a receptor (OR) to be estimated approach that has been used widely. Subjects can be
if the assumption that the in vitro and in vivo KD are administered simultaneously both 11C and 14C-labelled
equivalent, where: molecules. The radiation burden added by the 14C mole-
cule at radiotracer doses is small. The long half-life of 14C
OR = C free plasma /(C free plasma + K D ) (5700 years) means that there is no time pressure on
sample handling for the additional analyses! This poten-
The passive distribution assumption can be tested with tially allows the 14C to be used to provide more detailed
separate, invasive preclinical experiments (ideally in a information on PK behaviour, which can improve model-
non-human primate for brain studies) exploring the rela- ling in the PET experiment. An alternative would be to use
13
tionship between plasma concentration of the molecule C label with GC-MS, allowing additional information on
and the concen­tration in the tissue of interest as demon- molecule absorption, distribution, metabolism and excre-
strated by microdialysis. tion (ADME) to be obtained. The value of the information
The passive distribution model does not hold in situa- can be enhanced by performing the study after administer-
tions in which there is high expression of active transport- ing varying pharmacological doses of the unlabelled ‘cold’
ers for the molecule of interest, such as P-glycoprotein, at drug. However, the cost of this combined approach is
the blood–brain barrier. Evidence for transporters can be high, limiting its use to specialized applications.
derived from demonstration of exclusion of tracer doses A creative extension of the traditional biodistribution
of the radiolabelled molecule but increasing tissue uptake experiment was demonstrated in use of the differential
with increasing plasma concentrations of the unlabelled distribution of alternatively labelled molecules to provide
molecule or after a transporter inhibitor is administered information on drug metabolism directly from the PET
(Loscher and Potschka, 2005). experiment. The approach demands sophisticated radio-
Reaching the tissue of interest in amounts sufficient to chemistry, but provides an elegant reminder that what is
have a pharmacological effect is such a fundamental measured by PET is the distribution of the label, not the
requirement for drug action that, if there is uncertainty, molecule specifically. Temozolomide undergoes decar-
biodistribution data should be acquired at the earliest boxylation and ring opening in the 3–4 position to
stages of new drug development. Relative biodistribution produce the highly reactive methyldiazonium ion that
can be a factor contributing to selection of the lead mol- alkylates DNA. To evaluate this directly in humans, a dual
ecule for development at candidate selection. Non-human radiolabelling strategy was employed in which [11C]temo-
primate or other preclinical studies can be performed effi- zolomide was radiolabelled separately both in the
ciently prior to filing for registration of a labelled drug 3-N-methyl and 4-carbonyl positions. 11C in the C-4 posi-
molecule as an investigational medicinal product (IMP). tion of [4-11C-carbonyl]temozolomide will be converted to

263
Section | 3 | Drug Development

[11C]CO2 and an inactive metabolite. Paired studies were inhibitors of G-protein coupled receptors, preclinical (and
performed with both forms of [11C]temozolomide in 6 clinical) studies suggest that free concentrations sufficient
patients with gliomas. A third PET scan was performed to provide at least 70% receptor occupancy are needed. By
with 11C-radiolabelled bicarbonate to provide data allow- demonstrating the relationship between plasma concen-
ing quantitative modelling of the labelled CO2 release tration and target interaction, doses sufficient to achieve
from the independently performed temozolomide experi- this degree of interaction can be defined. If this informa-
ment. Data were obtained on activities of [11C]temozolo- tion is available before dose ranging studies, the range of
mide and [11C]metabolites in plasma collected during doses that need to be explored is reduced, allowing the
scanning and [11C]CO2 was measured in the expired air. number of volunteers exposed to the experimental mole-
Greater amounts of [11C]CO2 in the plasma and exhaled cule (and cost) to be minimized at this early stage.
air and lower tumour [11C]temozolomide signal with the Target interaction studies demand availability of a radio-
[4-11C-carbonyl]temozolomide relative to that labelled in ligand that has good affinity and relative specificity for the
the 3-N-methyl position confirmed ring-opening as a target (Cunningham, 2005). The selected radioligand with
mechanism for metabolic activation of temozolomide relative target selectivity then can be used for imaging the
(Saleem et al., 2003). extent of target available before and after a pharmacologi-
Monoclonal antibodies and other biopharmaceuticals cal dose of the molecule of interest. The ratio of the specifi-
are becoming an increasingly important part of new thera- cally bound radioligand to its free concentration (estimated
peutics development. An increasing range of methods are from the plasma free concentration) in the tissue is termed
available for labelling such large molecules with positron the ‘binding potential’ (BP). The estimated BP at equilib-
emitting isotopes in well-defined ways (van Dongen and rium is the ratio of the target availability (Bavail) relative to
Vosjan, 2010). Because of the much longer distribution the radioligand dissociation constant from the receptor
times for these large molecules, long-lived position emit- (Bavail/KD). With prior administration of the unlabelled
ters such as 89zirconium, 64copper or 124iodine have been (drug) molecule that binds to the same target, the meas-
used. Considerable information potentially is available, ured BP varies with the local concentration and the affinity
but the slow approaches to steady-state and the distinct of the unlabelled molecule. Over a range of doses of the
range of non-specific interactions and metabolism of these unlabelled molecule, the variation in radioligand BP thus
large molecules makes the conduct and interpretation of allows binding affinity of the unlabelled molecule to be
these studies more challenging than for small molecules. estimated. As a rule of thumb, the radiotracer BP before
While promising, this area still is in an early stage of devel- any drug administration should be greater than about 0.5
opment. There are special challenges in defining a mean- for sufficient signal to detect binding changes.
ingful distribution for radiotracers with very high affinities Estimates of in vivo BP for alternative candidate mole-
(especially if the binding site availability [Bavail] also is cules for a target can provide particularly important infor-
high), as the distribution may reflect delivery (blood flow) mation if there are dose limiting toxicities (Figure 18.2).
more than the distribution of specific binding sites. In such cases, the BP measures and simultaneous measures
of plasma concentration allow estimation of the pharma-
cologically active dose range, the upper limit of which can
be related directly to the minimum dose at which adverse
TARGET INTERACTION events are expected.
PET studies are expensive and theoretically confer some
Does a potential drug interact with the intended target additional long-term health risks for volunteers arising
to an extent that could be pharmacologically active? from the additional radiation exposure. It therefore is
What is the relationship between plasma concentrations important to optimize the efficiency of study designs to
(and, thus, administered dose) and the extent of target use the smallest number of subjects and the optimal selec-
interaction? tion of doses for defining the in vivo binding affinity. In
Direct demonstration of interaction of a molecule with contrast to the approach with a fixed dose design, adaptive
the tissue target confirms distribution into the tissue. designs use information gained from each experiment to
Target interaction studies also allow estimation of relevant improve the selection of the subsequent dose (see also
tissue binding of a molecule without making assumptions Chapter 17). Adaptive designs are well suited to PET meas-
regarding correspondences between in vivo and in vitro urements of this type, as outcomes data can be provided
measures or between human and preclinical in vivo recep- soon after completion of a study of any one subject to
tor affinities. contribute to dose selection for the next. The less prior
Target interaction studies are more informative if there knowledge concerning dosing that is available, the greater
is a strong hypothesis concerning the degree of target inter- the potential efficiency gains with adaptive designs
action needed for a pharmacological effect. In such cases, (Zamuner et al., 2010) .
data relating plasma concentration to target occupancy Is there ‘added value’ from human in vivo target interac-
can guide dose selection directly. For example, for tion studies in drug development? In some cases, human

264
 Clinical imaging in drug development Chapter | 18 |

100
Receptor occupancy (%)

Fig. 18.2  A, Images illustrating the specific binding of a

50 receptor-specific radioligand before and after administration
of pharmacological doses of a drug targeting the same
receptor binding site. The reduced specific signal on the PET
scans after dosing reflects decreased receptor site availability
in the presence of the drug. B, Quantitative modelling of the
0 PET signal to measure specific binding of the radioligand as a
0 100 200 300 400 500 600 700 800
function of plasma concentration for the drug allows the in
Plasma concentration (ng/mL) vivo receptor affinity of the drug to be estimated.
B
Images courtesy of the GSK Clinical Imaging Centre, London.

in vivo binding potentials can be very different from those Potentially important information for prioritization of
measured in vitro or in preclinical models. A histamine H3 candidate molecules prior to selection for clinical develop-
subtype antagonist, for example, was shown to have an in ment can be obtained in this way. Data from preclinical
vivo binding potential in humans substantially greater (especially non-human primate) studies also can establish
than that estimated on the basis of preclinical studies. This ‘priors’ that will reduce the number of volunteers needed
observation had a substantial impact on the related drug- for estimation of the human in vivo BP.
development programme, as it gave a rationale and confi- An interesting variant of this application is to the ‘reverse
dence for a major reduction in dosing and subsequent engineering’ of empirically established treatments to
evaluation of doses that were well tolerated by patients. better define factors that may be driving therapeutic effi-
The study also highlighted unexpectedly slow receptor ‘off’ cacy. A recent example was to better understand differ-
rates for the molecule, leading to a re-estimation of the ences in dopamine receptor subtype interactions between
optimal dosing frequency. In general, tissue PK or target atypical and typical antipsychotics. PET imaging was used
interactions and plasma PK are not the same except for the in baboons with [11C]PHNO to determine the binding of
limiting case of molecules with fast equilibrium binding clozapine and haloperidol to dopaminergic D2 and D3
properties that diffuse passively between the plasma and receptor subtypes. Scans were acquired following single
relevant tissue compartments. doses of antipsychotic drugs and compared with baseline
Non-human primate or preclinical studies of other scans. The percent changes in binding (ΔBP(ND)) follow-
species can be used to estimate relevant plasma ing challenges with antipsychotic drugs were measured. A
concentration-time-target interaction relationships prior regression model, based on published values of regional
to obtaining approval of IMP approval for a molecule. D2 and D3 fractions of [11C]PHNO BP(ND) in six brain

265
Section | 3 | Drug Development

regions then was used to infer the specific occupancy sepa-

rately on the D2 and D3 receptors (Girgis et al., 2011).
Box 18.1  Selected applications of imaging in
Target interaction studies typically have been conducted drug development
with single dose studies for experimental convenience. A
Marketed drugs
follow-up study, informed by the prior data, then is pos-
sible after repeat dosing to confirm dose–occupancy rela- • Differentiation between available treatments
tionships with more chronic drug administration. • Earlier detection of disease or associated pathology:
However, if it is known (or can reasonably be assumed) – Improved disease classification/diagnosis
that repeat dosing does not induce changes in receptor
– Diagnosis of pre-symptomatic or minimally
expression or availability, a consistent relationship symptomatic disease
between plasma concentration and target occupancy can
– Improved identification of chronic disease
be assumed and repeat-dose brain target occupancy can
exacerbation/recurrence
be estimated based on the basis of the combined occu-
pancy data obtained after administration of a single dose – Patient stratification based on disease sub-
and plasma pharmacokinetic (PK) data. The principles phenotype or early treatment response
behind this kind of analysis were illustrated with a study Late phase development
of single and repeat dose target interactions for the anti-
• Surrogate markers of response more sensitive than
depressant duloxetine. Integrated plasma concentration- clinical measures
time-target occupancy models were fitted to the single
• Stratification of patients based on potential for
dose data to characterize the model parameters and then
treatment efficacy
applied to an estimated repeat dose duloxetine plasma
time course to predict the 5-HTT occupancy after repeat • Pharmacological differentiation of asset from marketed
dosing (Abanades et al., 2011). drugs or new competitor compounds
Early phase development
• Biodistribution studies confirming molecule reaches
PHARMACODYNAMICS the target tissue and does not accumulate in
non-target sites of potential toxicity
• Target PK (dose-target occupancy) measurements
Imaging methods can provide information on in vivo
guiding dose selection
tissue structure, physiology or biochemistry. Based on
questions derived from the pharmacological hypotheses • Pharmacodynamic biomarkers for proof of
for the study, these approaches thus can be used to provide pharmacology, stronger ‘reasons to believe’ or
pharmacodynamic information for proof of principle or contributing key rationale for proof of concept
molecule differentiation. The full range of imaging • Translational preclinical imaging to identify or
methods has been – or could be – applied with the validate new imaging biomarkers or provide early
common objective of answering the question: does a mol- differentiation between candidates based on target PK
ecule exert a pharmacological effect on the biological or PD responses
system of therapeutic interest? The potential range of Safety and toxicity
applications is illustrated by some general principles and
• Imaging biomarkers as non-invasive, in vivo surrogates
specific examples (Box 18.1). for direct, ex vivo studies of tissues and their
Pharmacological studies of diseases of the brain have translation from preclinical to clinical studies
been a particularly important area for the development of
new imaging pharmacodynamic markers, in part simply
because the privileged localization of the brain limits 2009) have given confidence that early phase trials with
access to tissue and the informativeness of circulating MRI endpoints for anti-inflammatory therapies can be reli-
biomarkers. MRI fundamentally changed how clinical ably informative, demand smaller numbers of patients for
trials in multiple sclerosis (MS) were performed. An initial meaningful endpoints and can be completed more quickly
therapeutic goal was to limit new inflammatory lesions, than trials with clinical outcome measures. An additional
which are sensitively marked by gadolinium-enhancement advantage of the close MRI monitoring is that adverse con-
of T1-weighted images in the hyperacute stage and hyper- sequences of immune modulation also can be identified
intensity on T2-weighted images in the acute and chronic early. In such cases, the costs and complexity of serial MRI
stages. The observation that the frequency of new gadolin- scans for the evaluation of the frequency of new enhancing
ium enhancement is as much as 10-fold the rate of disease lesions or the volume of T2 hyperintense lesions is well
relapse (Arnold and Matthews, 2002) and validation of the offset by the gains in trial efficiency and informativeness.
correlation between changes in these imaging markers and Additional endpoints are now in various stages of valida-
changes in relapse rate with treatments (Sormani et al., tion, allowing specific neuropathological changes to be

266
 Clinical imaging in drug development Chapter | 18 |

monitored to report on drug actions for remyelination, For example, a number of preclinical and pilot clinical
axonal loss or neurodegeneration, for example. studies had suggested that peroxisome proliferator acti-
Alzheimer’s disease (AD) has been a major recent area vated receptor gamma (PPAR-γ) -gamma agonism could
of new research for evaluation of imaging markers in ther- reverse bioenergetic defects related to abnormal insulin
apeutics development because of the slow rate of progres- signalling and reduced glucose uptake in AD that might
sion of the disease and thus long periods over which be contributing to neurodegeneration. The brain has a
clinical measures need to be followed for meaningful high metabolic rate and glucose is the preferred substrate.
outcomes. The most striking neuropathological feature A reduced rate of glucose utilization, particularly in the
of AD is the progressive brain atrophy related to neuronal hippocampus and temporal-parietal cortex, is an early dis-
dystrophy (retraction of the extensive ramification of neu- criminant of people at risk of developing AD. The cerebral
rites extending from the neuronal cell body) and death. glucose metabolic rate slows progressively over time with
This is reflected in shrinking of the cortex (and other grey development of the disease. Changes are also dispropor-
matter), which leads to generalized brain atrophy. Meas- tionate to decreases in volume, suggesting that they repre-
urement of the rate of brain atrophy provides a pharma- sent neuronal dysfunction prior to the neuronal dystrophy
codynamic index related to neurodegeneration (assuming or loss assessed by structural measures. Together, this prior
that changes in water content or the relative size of associ- knowledge allowed a precise pharmacodynamic hypoth-
ated cell compartments, such as glia, are not significant). esis to be generated a proof of principle study of the effi-
Robust, automated brain MRI measures of volume and cacy of the PPAR-gamma agonist rosiglitazone in AD: does
volume change provide reproducible indices of the pro- treatment enhance uptake of the glucose analogue [18F]-
gression of disease (Smith et al., 2009). Recent interest has fluoro-dexoyglucose (FDG) as measured by PET and is this
focused on development of similarly robust approaches associated with slowing of the progressive reduction in
for the measurement of regional brain volumes defining FDG uptake expected with the natural history of the
atrophy specifically of the hippocampus, for example, disease? A novel multi-centre FDG PET trial with 80 sub-
which begins in the earliest stages of the disease and jects demonstrated a trend for an improvement in FDG
progresses faster than that for the whole brain. uptake over the first month of treatment, but provided no
Pharmacodynamic biomarkers ideally should be tai- evidence for slowing of the progression of neurodegenera-
lored very specifically to the drug development question. tion (Tzimopoulou et al., 2010; Fig. 18.3). These data were

Fig. 18.3  FDG PET as an imaging biomarker for the progression of neuropathology in Alzheimer’s disease. Rendering of the
location of voxels with significant decreases in glucose metabolism (red) over a 12-month period are overlaid on the surface of
a reference brain (grey). Changes in the rate of decrease of brain glucose metabolism can be used as a pharmacodynamic
marker for therapies intended to slow neurodegeneration (Tzimopoulou et al., 2010).
Images courtesy of the GSK Clinical Imaging Centre, London.

267
Section | 3 | Drug Development

consistent with clinical results from two large phase III the basis of the current, widely accepted RECIST criteria
trials (involving many hundreds of patients) that had been for assessment of longitudinal changes with treatment.
conducted in parallel. The experience highlights how Several MRI markers of tumour volume, tissue microstruc-
directly well-selected imaging pharmacodynamic markers ture or metabolism promise additional approaches
can be related to underlying pharmacological hypotheses. (Workman et al., 2006). PET radiotracers are being devel-
It also suggests the potential cost and trial efficiency gains oped to image hypoxic, apoptotic or dying cells or other
that could come with optimal staging of imaging studies parameters relevant to potential tissue responses.
able to address key pharmacodynamic questions like this fMRI is an emerging imaging pharmacodynamic marker
in a clinical development plan. for drugs that have actions – direct or indirect – on brain
This is further illustrated by a recent study of treatment activity, for example, for measurement of brain activity
effects on brain amyloid deposition in AD. Genetic find- while the symptom of interest is experienced, or during
ings in families with familial disease have suggested that performance of tasks that engage cognitive processes puta-
excessive or abnormal amyloid protein may be a cause of tively related to the symptom. The method is of particular
neurodegeneration in AD. This has led a number of phar- interest as a way of providing an objective, neurophysio-
maceutical companies to develop anti-amyloid antibodies logical measure of brain events related to subjective
intended to provide a ‘peripheral sink’ binding amyloid experiences such as pain. This could be of special value
and lowering the free plasma and brain amyloid con­ in patients who are unwilling or – as for those with AD
centrations. Brain amyloid deposits can be assessed by – unable to report accurately. An advantage of fMRI is that
PET measures of brain binding of the radiotracer [11C]- similar imaging approaches can be adapted to address a
Pittsburgh compound B (PIB), which shows a relative affin- broad range of pharmacodynamic questions simply by
ity for the beta sheet structure of the deposits. In a recent applying them in the context of different behavioural task
Phase IIa study, time- and dose-dependent reductions in challenges.
brain PIB binding were reported with use of the amyloid An emerging alternative to task-related fMRI is the use
binding antibody bapineuzumab (Rinne et al., 2010). of resting state fMRI, which probes a class of functional
Imaging in this case is molecularly specific and provides interactions between brain regions (Zhang and Raichle,
a pharmacodynamic marker. PIB PET does not report 2010). There is a continued background activity in the
directly on the free amyloid peptide or amyloid oligomers, brain, the modulation of which is measured with task-
however, and would not necessarily provide a useful phar- constrained fMRI. A component of this background activ-
macodynamic marker for a treatment addressing another, ity can be monitored as fMRI signal fluctuations at rest
independent mechanism of genesis, expression or progres- that occur at low frequencies (0.01–0.05 Hz) and show
sion of AD. Pharmacodynamic markers – whether imaging with coherent changes between widely separated brain
or other kinds of biomarkers – need to be tailored to meet regions. These consistent spatiotemporal coherence pat-
the demands of the specific treatment question. terns define common activity networks that correspond
Similar considerations hold for the development of functional anatomically with those engaged by task-
therapeutics in oncology. A growing ‘toolkit’ of imaging constrained activations (Smith et al., 2009). These and
markers for activity of biological processes commonly related coherence measures can define drug effects (Cole
altered by many therapies is becoming available. For et al., 2010).
example, quantitative FDG PET provides a marker for the Pharmacodynamic measures can be integrated with
elevated glucose metabolism in tumours as a concern of those for biodistribution to define PK-pharmacodynamic
the upregulation of glycolysis (the ‘Warburg’ effect). relationships directly. This innovative strategy was applied
Changes in FDG uptake can reflect specific effects on to the characterization of a novel antisense oligonucleo­
insulin signalling and glucose uptake (e.g. of use with Akt tide strategy for tumour treatment. The antisense oligonu-
or PI3 kinase specific inhibitors), as well as non-specific cleotide was labelled with 11C and a PET biodistribution
effects on gly­colytic enzyme expression or cell viability. study performed to demonstrate its accumulation within
Relative cell turnover rates can be probed through assess- the tumour tissue, while biochemical studies performed
ment of the activity of the thymidine salvage pathway with on samples obtained from biopsies of the same tumours
[18F] fluoro-L-thymidine (FLT). Dynamic contrast enhanced were used to relate these measures to direct tests of the
(DCE) MRI uses the time course of signal change in a pharmacodynamic hypothesis (Talbot et al., 2010).
tumour after injection of a bolus of gadiolinium contrast
into the circulation to model tissue blood volume, per-
fusion and the contrast agent permeability, which is
increased with tumour neoangiogenesis. Pharmacody-
PATIENT STRATIFICATION AND
namic effects of angiogenesis inhibitors should be reflected PERSONALIZED MEDICINE
in these parameters.
Less specific markers also can be useful. X-ray computed A critical issue in early drug development is to establish
tomography (CT) estimates of tumour size changes are at an appropriate level of confidence in the potential of a

268
 Clinical imaging in drug development Chapter | 18 |

new molecule to become a therapy. One way in which this generally has not been needed for good diagnostic
can be facilitated is by better controlling for intrinsic vari- imaging. A recent report also has highlighted the possibil-
ations in therapeutic responses between individuals. As ity that a stratified population behaves differently from the
well demonstrated in oncology, stratification of patients general population, so that gains from the investment in
based on specific disease characteristics can allow for more stratification are reduced. In this example, modelling
powerful trial designs. Consider, hypothetically, the differ- showed that an increase in outcome variance in a subgroup
ence in outcomes of trials first for a population in which of mild cognitive impairment (MCI) patients selected
a new molecule has a 50% treatment effect in 20% of according to proposed criteria significantly offset the gains
patients (leading to a 10% net treatment effect) and then expected from the population as a whole (Schneider et al.,
in a stratified population enriched so that 70% are 2010).
responders (net 35% treatment effect). If able to predict
potential responders, imaging could suggest ways of best
selecting optimal patient groups for applications of new Towards personalized medicine
molecules. Of course, developing such a strategy is predi-
Personalized medicine (PM) is an extension of the concept
cated on having an understanding of pharmacology suf-
of stratification of patients in trials to their stratification
ficient to make useful guesses regarding the selection of
in medical care delivery. The President’s Council of Advi-
stratification criteria.
sors on Science and Technology (PCAST), in their Septem-
Imaging-based stratification is already applied in clini-
ber 2008 report entitled ‘Priorities for Personalized
cal practice to better ensure efficacy and limit adverse out-
Medicine’, define ‘personalized medicine’ as ‘the tailoring of
comes, e.g. to limit surgical treatments to patients with
medical treatment to the individual characteristics of each
localized neoplastic disease, or tissue plasminogen activa-
patient. The intent is for treatments to be concentrated on those
tor (tPA) therapy to patients presenting within a few hours
who will benefit, sparing expense and side effects for those who
after ischaemic stroke. Enrichment of clinical trials based
will not.’
on imaging indices also is well established in specific
The underlying concepts are not novel. Healthcare
areas, e.g. enrichment of MS trial populations for active
delivery has long relied on physiological or pathological
inflammatory disease by screening for gadolinium-
indices, as well as their clinical assessments to optimize
enhancing lesions at trial entry.
(personalize) the diagnosis and treatment of patients.
There are obvious cautions to the use of enrichment or
Current efforts in PM are evaluating the use of new
stratification methods. First, they can increase trial cost or
biomarkers (including imaging based measures) together
complexity to an undesirable extent. Individual scans and
with an understanding of both pharmacology and disease
the additional demands for image data management,
to make more rational treatment choices for individual
quality control and expert support in analysis add to trial
patients. Potential applications of this approach in an
cost. The demands of research imaging in busy clinical
imaging context would include:
settings can limit times for scheduling subjects in a trial,
introducing constraints that make trial execution more • Selection of patients for a treatment based on
difficult. Information sheets for volunteers and consent imaging
forms inevitably become more difficult to understand as • Dose adjustment based on imaging measures
imaging procedures and their safety concerns are explained, • Identification of risks or early markers of adverse
potentially reducing recruitment. Specialized types of events with use of the drug based on imaging
imaging (e.g. PET scanning with advanced radiotracers) • Using imaging to monitor treatment responses
may be able to be performed at a very limited number of • Selection of pre-symptomatic people with
sites. A consequence can be less efficient trial design and developing pathology for treatment based on
execution. These consequences need to be factored into imaging criteria.
decisions to use imaging for stratification. Optimally effi- PM approaches ultimately will demand a common
cient statistical methods able to evaluate likelihoods of vision from the pharmaceutical industry and regulatory
causality to derive maximum utility from the data and an agencies, and likely will be most powerfully applied
understanding of the clinical relevance of strategies add when a PM strategy can be incorporated into new drug
value. In addition, for situations where imaging metrics development. Treatments ultimately approved for use
derived from imaging that is part of routine clinical care with imaging support will need to be able to include
can be used for stratification, immediate gains may be sufficient information in the drug label to ensure that
expected. the correct patients can be selected for treatment in
However, the gains can be lower than might initially be routine practice. This will demand that measures are
expected. There are a number of reasons for this. One is both available and able to be appropriately standardized.
that use of quantitative imaging methods demands explicit Instrument manufacturers and relevant professional
consideration of ways of controlling for inter-session or bodies are already implementing approaches to facilitate
inter-site variance in measurements with a care that this.

269
Section | 3 | Drug Development

IMAGING AS A   IMAGING IN THE REAL WORLD –  

SURROGATE MARKER CHALLENGES TO IMPLEMENTATION

Enthusiasts of imaging in drug development often explain As with any aspect of trial conduct, conceptually elegant,
the benefits of any imaging approach through its potential robust implementation of clinical imaging in the context
to become a surrogate marker for clinical treatment of a new drug trial brings substantial challenges. While
responses in pivotal trials. However, the instances in which imaging markers can have a high statistical power in some
this will be possible are likely to be rather limited. pharmacological applications, the measures are quantita-
The Food, Drug and Cosmetics Act is interpreted by the tive, continuous and often dependent on multiple factors.
FDA as demanding evidence that the drug confers a clini- Controlling for variance in data between sites or even
cal benefit. This establishes a clear need to acquire data for between examinations at a single site can be demanding.
efficacy, as the usual, most meaningful clinical measures Reasonable prior estimates of the expected treatment effect
which, in general, are not expressed simply as imaging size and the reproducibility of measures is needed to
outcomes. While this does not mean that key questions estimate the potential study power. Progression of over-
cannot be addressed with primary imaging endpoints in powered studies incorporating complex and expensive
early phase development or that imaging data may not outcomes will be not be sustainable.
become a general part of regulatory packages to support In general, studies with designs based on reproducibility
the therapeutic rationale, it does not suggest that primary of measures within a single site can substantially underes-
outcomes for late phase studies will often be able to be timate true variance. For example, structural measures
framed as an imaging measure. based on MRI are dependent on the homogeneity and
There are circumstances in which registration based on linearity of the MR gradients, which can vary between
imaging markers could be possible. For example, espe- instruments and installations of even the same instru-
cially if the relationship between amyloid and Alzheimer’s ments. Fortunately, considerable effort has been expended
disease becomes firmly established, therapies to lower in standardizing measurements (Boellaard et al., 2008;
brain amyloid might be approved on the basis of PET Friedman and Glover, 2006; Stocker et al., 2005). In anal-
molecular imaging markers of brain amyloid. ogous ways, the sensitivity of PET scanners varies. In both
This and other special situations in which imaging may cases, calibrations with standard phantoms used across
provide a primary outcome measure may be considered sites can be used to minimize variances. Even so, instru-
within the provisions of subpart H, 21 CFR 314.50 of the ment or site bias in measurements can be introduced in
Accelerated Approval Provision, which allows approval on many ways, e.g. differences in radiofrequency coil cou-
the basis of a surrogate marker likely to predict clinical pling for MRI, detection mode for PET scanners or injec-
benefit. However, a central tenet of this regulation is that tion timing for radiopharmaceuticals or contrast agents.
it is applied in a situation in which there is an ‘inability’ Analysis of results allowing for confounding of site-to-site
to assess a clinical outcome with a feasible trial. Applica- variation helps to ensure that site bias can be recognized
tions to the evaluation of treatments for rare diseases with and optimally accounted for.
known pathophysiological mechanisms, for which there More fundamentally, introduction of more sophisti-
are well established imaging markers, may be the most cated imaging endpoints can limit the availability sites.
obvious areas for immediate application, e.g. in rare Training on specialized techniques may not be accessible
storage disorders such as adrenomyeloneuropathy. widely, e.g. cardiac MRI demands cardiologist and radiog-
Even in such situations, for the first agent in a new rapher training and ongoing experience to maintain skills
therapeutic class, the use of imaging markers is unlikely to if it is to be performed well. Some methods may be imple-
meet generally acceptable criteria as a measure of efficacy. mented only at institutions with relevant research inter-
The most obvious example of how such an approach ests. PET methods relying on molecular probes other than
would be developed would be to follow imaging observa- the few that are obtainable commercially are only availa-
tional studies of the disease and validation of the biomar- ble at a limited number of sites and shorter-lived isotopes
ker against usual measures of progression or even (e.g., for [11C] PET radiotracers) can only be used where
against a ‘gold standard’ intervention. However, a simple there are radiopharmaceutical manufacturing facilities
correlation between the natural history of progression and and a cyclotron are accessible on site.
a measure will not establish predictive likelihood in the The outcomes from many methods depend on precisely
context of a new treatment. Even validation against a how they are implemented. Consensus criteria have been
‘gold standard’ treatment does not establish the general developed to minimize site variation in DCE-MRI meas-
relationship because of the potential for multiple mech­ ures, for example (Leach et al., 2003). Analytical tech-
anisms or therapeutic effects other than those originally niques can be designed that are relatively robust to many
postulated. aspects of patient or site variability. The issues have been

270
 Clinical imaging in drug development Chapter | 18 |

discussed before in different contexts, e.g. for precision in Imaging supported clinical drug development is still an
the assessment of brain atrophy rates or dynamic changes immature field and experience is limited. Validation of
in FDG measures of brain glucose utilization in AD (Smith methods in the context of novel targets or classes of thera-
et al., 2002; Tzimopoulou et al., 2010). peutic molecules is even more limited. Potential applica-
Imaging approaches are variably complex (or can confer tions need to be anticipated sufficiently and investments
additional health risks) for patients or volunteers. These made well in advance of need if imaging is to be applied
factors limit volunteer interest and complicate recruit- most powerfully in these contexts. This is most easily illus-
ment. MRI studies can be as short as 7–10 min (e.g. a trated by considering molecular imaging markers. Biodis-
whole body scan for fat-water assessment) or less, but tribution studies demand the site-specific labelling and
multi-sequence studies demand cooperation for periods quantitative modelling of the distribution of signal in the
of 30–60 min in what to the volunteers may be the alien tissues of interest. Implementation demands evaluation of
environment of the high field magnet bore and with the labelling feasibility under the constraints of automated or
sometimes distressing noise from the MR gradients. Prior semi-automated radiochemical methods, tailored for the
familiarization of subjects with both the magnet and the appropriate half-life and the development and validation
noise, and attentive support from staff, can reduce com- of the specific radiopharmaceutical and metabolite quan-
plaints of claustrophobia and increase tolerance of more titative analysis methods for the radiotracer of interest.
extended procedures. Contrast or radiopharmaceutical While timelines are highly variable and depend on both
injections demand placement of intravenous catheters, so the molecule and on particular capabilities of the imaging
are not strictly non-invasive. While intravenous catheters centre, typical estimates from our site with recent mole-
are generally well tolerated, the intra-arterial (typically cules suggest that even an optimal progression timeline
placed in the radial artery) cannulation needed for serial demands at least 3–4 months for feasibility to be estab-
blood sampling to allow accurate estimation of the input lished, and 3 months for validation of a novel-labelled
function for fully quantitative PET can be uncomfortable. molecule. Thus, the decision to incorporate a PET biodis-
Finally, all of the imaging methods using ionizing radia- tribution study needs to be made well in advance of the
tion carry some estimated additional health risks for sub- need for information. Fortunately, as a microdosing exper-
jects. It is important, therefore, to minimize demands iment, this planning can be done and implementation
on volunteers to those strictly needed to answer the completed as one of the earliest experiments in the
question(s) of interest and to ensure that the procedures sequence of clinical development studies.
and risks are explained well, that procedures are optimized Target occupancy studies are more demanding, espe-
to give subjects the best experience possible, and that they cially if a radioligand suitable for use as a reporter for the
are conducted by highly trained staff to minimize risk or target needs to be developed. Radioligand development is
discomfort. as challenging as the first stages of any drug development
Some kinds of clinical studies for drug development programme, as affinity and specificity need to be balanced
already are incorporating imaging routinely to provide for the molecule. Candidate radioligands show high attri-
data for development decisions because of the demands tion along the course of development and validation. Lim-
of usual clinical care delivery or the need for imaging- iting non-specific binding is a particular challenge. In
based safety readouts. For these, the extra care in site general, useful radioligands should be lipophilic enough
setup and data analysis needed for quantitative imaging to diffuse well between tissue compartments without
endpoints may add minimally to either trial cost or being so lipophilic that they accumulate in fatty tissues to
complexity. However, in other instances, imaging end- a substantial extent. However, recent advances in a design
points typically add substantially to cost and trial com- based biomathematical modelling approaches (Guo et al.,
plexity and can make recruitment more difficult. Feasibility 2009) and the great potential for radioligand development
of the design of any imaging supported clinical trial needs to benefit from medicinal chemistry experience in the syn-
to be carefully explored with potential trial sites. Some- theses of related drug candidates suggest that there are
times, relatively small issues – such as the extra time substantial opportunities for more efficient development
needed in the clinical centre for a full imaging examina- in the future.
tion – add so much to the burden of the trial for patients Pharmacological fMRI offers the advantages of a poten-
communicating into a research centre, that recruitment tially general marker for brain pharmacodynamic studies
within desired time frames is not possible. It is important, and has been used for a broad range of applications in
therefore, that the value of the imaging outcome is healthy volunteers and patients. However, major chal-
high, commensurate with these direct and indirect costs. lenges to the meaningful quantitative interpretations of
In the planning stage, designs and development plans pHMRI measures remain. First, the relationship of blood
that do not rely on imaging need to be considered to flow changes with altered presynaptic activity depends on
balance the information available with respect to both the physiological (and, potentially, pharmacological)
ease of implementation by the sites and participation by context. Even the direction of changes in relative activation
volunteers. in pathological states may be difficult to interpret

271
Section | 3 | Drug Development

precisely. For example, reduced activation in the aging

brain may reflect either brain functional impairment or SUMMARY
improved efficiency. Experimental designs need to accom-
modate this, for example by studying dose–response rela- Clinical imaging has an important role in drug develop-
tions and behavioural correlates in individual studies. ment. The utility already has been well demonstrated in
Additionally, BOLD signal changes arising from any direct several applications reviewed here. It can facilitate more
(or indirect, e.g. with changes in ventilation rate (Wise seamless transitions from preclinical to clinical develop-
et al., 2004)) need to be controlled. Nonetheless, experi- ment. Imaging also will facilitate answering critical ques-
ence has shown that fMRI provides pharmacologically tions earlier in development and directly with human
discriminative measures (Matthews, 2009) and that studies studies, adding to confidence in decision-making. While
can be implemented across multiple sites and the resulting implementation of imaging-supported protocols can add
data meaningfully integrated (Bosnell et al., 2008). to trial complexity and cost per patient, where imaging has
Future methodological developments promise to make greater sensitivity, the gains can translate into smaller
fMRI even more informative. Computational advances numbers of subjects in trials. This would have a particu-
already allow robust analyses in real time. In the context larly large impact on the potential to pursue studies for
of pHMRI, this could enable improved quality control rare diseases for which it may be possible to recruit only
during an examination or more precise tailoring of the small numbers of subjects or for highly stratified popula-
protocol to the question being asked about an individual tions amongst even prevalent diseases for which PM is
patient. Some limitations to interpretation of the BOLD sought.
response can also be addressed with use of complemen- There now is an opportunity to substantially extend the
tary forms of MRI contrast or through the integration of range of situations in which, and the extent to which,
BOLD MRI and other measures in simultaneous data imaging is used in clinical drug development. However,
acquisition. For example, direct measures of brain blood optimal use will demand the development of preclinical
flow can be made using non-invasive ‘arterial spin label- and clinical imaging strategies to support molecule devel-
ling’ (ASL) MRI methods which have greater stability over opment from the earliest stages of programme planning
time for better assessment of slow (of the order of a to ensure that methods are validated and ready to be
minute or more) changes in brain responses now can be implemented as critical decision points in projects are
implemented robustly (Xie et al., 2010). With care for being approached. Wider use in development should cata-
safety issues and correction of the artefacts induced by the lyse new applications for imaging-based PM to better
shifting magnetic field gradients used for MRI, high- ensure that the right medicine is used by the right patient.
quality EEG now can be obtained simultaneously during
an fMRI examination (Lemieux, 2004), allowing simulta-
neous pHMRI and pharmaco-EEG studies. Variance in
measures can be reduced by correcting for variations ACKNOWLEDGMENTS
in pCO2 across the ventilatory cycle (Wise et al., 2004). In
the near future, advances in positron detection methods The author would like to thank his colleagues in the GSK
should support commercial availability of combined Clinical Imaging Centre for their substantial contributions
human PET/MRI scanners that will allow integrated exper- to the ideas that have developed into this chapter. He also
iments examining target occupancy and pharmacodynam- wishes to thank Mrs Rachel Green for careful editorial
ics (Herzog et al., 2010). assistance.

REFERENCES

Abanades S, van der Aart J, Barletta JA, multiple sclerosis. Neurology clinical setting: implications for trial
et al. Prediction of repeat-dose 2002;58:S23–31. designs. Neuroimage
occupancy from single-dose data: Boellaard R, Oyen WJ, Hoekstra CJ, 2008;42:603–10.
characterisation of the relationship et al. The Netherlands protocol for Callicott JH, Straub RE, Pezawas L, et al.
between plasma pharmacokinetics standardisation and quantification of Variation in DISC1 affects
and brain target occupancy. Journal FDG whole body PET studies in hippocampal structure and function
of Cerebral Blood Flow and multi-centre trials. European Journal and increases risk for schizophrenia.
Metabolism 2011;31:944–52, Epub of Nuclear Medicine and Molecular Proceedings of the National
Oct 13 2010. Imaging 2008;35:2320–33. Academy of Sciences of the USA
Arnold DL, Matthews PM. MRI in the Bosnell R, Wegner C, Kincses ZT, et al. 2005;102:8627–32.
diagnosis and management of Reproducibility of fMRI in the

272
 Clinical imaging in drug development Chapter | 18 |

Chuah LY, Chee MW. Cholinergic Archives of General Psychiatry temozolomide in glioma patients.
augmentation modulates visual task 2009;66:721–8. Cancer Research 2009;69:120–7.
performance in sleep-deprived young Inkster B, Nichols TE, Saemann PG, et Saleem A, Brown GD, Brady F, et al.
adults. Journal of Neuroscience al. Pathway-based approaches to Metabolic activation of
2008;28:11369–77. imaging genetics association studies: temozolomide measured in vivo
Cole DM, Beckmann CF, Long CJ, et al. Wnt signaling, GSK3beta substrates using positron emission tomography.
Nicotine replacement in abstinent and major depression. Neuroimage Cancer Research 2003;63:2409–15.
smokers improves cognitive 2010;53:908–17. Schneider LS, Kennedy RE, Cutter GR.
withdrawal symptoms with Jezzard P, Matthews PM, Smith S. Requiring an amyloid-beta1-42
modulation of resting brain network Functional magnetic resonance biomarker for prodromal
dynamics. Neuroimage imaging: methods for neuroscience. Alzheimer’s disease or mild cognitive
2010;52:590–9. Oxford: Oxford University Press; impairment does not lead to more
Cunningham VJ, Parker CA, Rabiner EA, 2001. efficient clinical trials. Alzheimers
et al. PET studies in drug Kegeles LS, Abi-Dargham A, Frankle and Dementia 2010;6:367–77.
development: methodological WG, et al. Increased synaptic Slifstein M, Laruelle M. Models and
considerations. Drug Discovery dopamine function in associative methods for derivation of in vivo
Today: Technologies 2005;2:311–5. regions of the striatum in neuroreceptor parameters with PET
David SP, Munafo MR, Johansen-Berg schizophrenia. Archives of General and SPECT reversible radiotracers.
H, et al. Ventral striatum/nucleus Psychiatry 2010;67:231–9. Nuclear Medicine and Biology
accumbens activation to smoking- Leach MO, Brindle KM, Evelhoch JL, et 2001;28:595–608.
related pictorial cues in smokers and al. Assessment of antiangiogenic and Smith SM, Zhang Y, Jenkinson M, et al.
nonsmokers: a functional magnetic antivascular therapeutics using MRI: Accurate, robust, and automated
resonance imaging study. Biology recommendations for appropriate longitudinal and cross-sectional
and Psychiatry 2005;58:488–94. methodology for clinical trials. brain change analysis. Neuroimage
Egan MF, Straub RE, Goldberg TE, et al. British Journal of Radiology 2002;17:479–89.
Variation in GRM3 affects cognition, 2003;1(76 Spec No):S87–91. Smith SM, Fox PT, Miller KL, et al.
prefrontal glutamate, and risk for Lemieux L. Electroencephalography- Correspondence of the brain’s
schizophrenia. Proceedings of the correlated functional MR imaging functional architecture during
National Academy of Sciences of the studies of epileptic activity. activation and rest. Proceedings of
USA 2004;101:12604–9. Neuroimaging Clinics of North the National Academy of Sciences of
Friedman L, Glover GH. Report on a America 2004;14:487–506. the USA 2009;106:13040–5.
multicenter fMRI quality assurance Loscher W, Potschka H. Role of drug Sormani MP, Bonzano L, Roccatagliata
protocol. Journal of Magnetic efflux transporters in the brain for L, et al. Magnetic resonance imaging
Resonance Imaging 2006;23: drug disposition and treatment of as a potential surrogate for relapses
827–39. brain diseases. Progress in in multiple sclerosis: a meta-analytic
Girgis R, Xu X, Miyake N, et al. In vivo Neurobiology 2005;76:22–76. approach. Annals in Neurology
binding of antipsychotics to D(3) Martinez D, Narendran R, Foltin RW, et 2009;65:268–75.
and D(2) receptors: a PET study in al. Amphetamine-induced dopamine Stocker T, Schneider F, Klein M, et al.
baboons with [(11)C]-(+)-PHNO. release: markedly blunted in cocaine Automated quality assurance
Neuropsychopharmacology dependence and predictive of the routines for fMRI data applied to a
2011;36:887–95, Epub Dec 22 2010. choice to self-administer cocaine. multicenter study. Human Brain
Guo Q, Brady M, Gunn RN. A American Journal of Psychiatry Mapping 2005;25:237–46.
biomathematical modeling approach 2007;164:622–9. Talbot DC, Ranson M, Davies J, et al.
to central nervous system Matthews PM, Filippi M. Tumor survivin is downregulated by
radioligand discovery and Pharmacological applications of the antisense oligonucleotide
development. Journal of Nuclear fMRI. In: fMRI Techniques and LY2181308: a proof-of-concept,
Medicine 2009;50:1715–23. Protocols. Heidelberg and New York: first-in-human dose study. Clinical
Hariri AR, Mattay VS, Tessitore A, et al. Springer-Verlag; 2009. Cancer Research 2010;16:6150–8.
Serotonin transporter genetic Rinne JO, Brooks DJ, Rossor MN, et al. Tzimopoulou S, Cunningham VJ,
variation and the response of the 11C-PiB PET assessment of change in Nichols TE, et al. A multi-center
human amygdala. Science fibrillar amyloid-beta load in randomized proof-of-concept clinical
2002;297:400–3. patients with Alzheimer’s disease trial applying [(1)F]FDG-PET for
Herzog H, Pietrzyk U, Shah NJ, et al. treated with bapineuzumab: a phase evaluation of metabolic therapy with
The current state, challenges and 2, double-blind, placebo-controlled, rosiglitazone XR in mild to moderate
perspectives of MR-PET. Neuroimage ascending-dose study. Lancet Alzheimer’s disease. Journal of
2010;49:2072–82. Neurology 2010;9:363–72. Alzheimer’s Disease
Inkster B, Nichols TE, Saemann PG, Rosso L, Brock CS, Gallo JM, et al. A 2010;22:1241–56.
et al. Association of GSK3beta new model for prediction of drug van Dongen GA, Vosjan MJ. Immuno-
polymorphisms with brain structural distribution in tumor and normal positron emission tomography:
changes in major depressive disorder. tissues: pharmacokinetics of shedding light on clinical antibody

273
Section | 3 | Drug Development

therapy. Cancer Biotherapy and pharmacokinetic and Resonance in Medicine

Radiopharmaceuticals pharmacodynamic technologies in 2010;64:203–10.
2010;25:375–85. hypothesis-testing clinical trials of Zamuner S, Di Iorio VL, Nyberg J, et al.
Wise RG, Ide K, Poulin MJ, et al. Resting innovative therapies. Journal of the Adaptive-optimal design in PET
fluctuations in arterial carbon National Cancer Institute occupancy studies. Clinical
dioxide induce significant low 2006;98:580–98. Pharmacology and Therapeutics
frequency variations in BOLD signal. Xie J, Clare S, Gallichan D, et al. 2010;87:563–71.
Neuroimage 2004;21:1652–64. Real-time adaptive sequential design Zhang D, Raichle ME. Disease and the
Workman P, Aboagye EO, Chung YL, for optimal acquisition of arterial brain’s dark energy. Nature Reviews.
et al. Minimally invasive spin labeling MRI data. Magnetic Neurology 2010;6:15–28.

274
 Chapter 19 
Intellectual property in drug discovery
and development
P Grubb

As is clear from the rest of this book, it costs a great further infringement. Note that although the state grants
deal of effort and money to discover and develop a new the patent right, the state does not check whether the right
drug. No one would make such an investment if the is being infringed – the patent owner must do that.
results could simply be copied by an imitator who had It is important to realize that the rights given by a patent
invested nothing. The best way to protect the investment do not include the right to practise the invention, but only
is by obtaining a patent for the drug. In this chapter we to exclude others from doing so. Many inventors and busi-
will look at what patents are, what kinds of inventions ness managers think that having a patent gives them
can be patented, and how a patent may be obtained and freedom to operate, but this is not so. The patentee’s
enforced. freedom to use the invention may be limited by laws or
Patents are not the only form of intellectual property regulations having nothing to do with patents, or by the
(IP), but they are by far the most important for the phar- existence of other patents. For example, owning a US
maceutical industry. Many patents that are filed and patent for a new drug does not give the right to market
granted prove to be worth nothing, but a patent protecting that drug in the USA without permission from the FDA
a blockbuster drug against generic competition may be (see Chapter 20).
worth millions of dollars for each day that it is in force. What is less obvious is that having a patent does not
An unexpected loss of patent protection may have a much give the right to infringe an earlier existing patent. To take
larger effect upon the market value of the company a simple example, if A has a patent for a process using an
holding the patent. In August 2000, when a US patent acid catalyst, and B later finds that nitric acid (not dis-
covering Prozac was held invalid by the Court of Appeal closed in A’s patent) gives surprisingly good results, B may
for the Federal Circuit, thereby reducing by about 3 years be able to get a patent for the process using nitric acid as
the expected term of exclusivity for this drug, 29% of the catalyst. However, because this falls under the broad
value of Eli Lilly stock was lost in 1 day – over $35 billion. description of acid catalysis covered in A’s patent, B is not
This is serious money by any standards. free to use his invention without the permission of A. On
the other hand, A cannot use nitric acid without a licence
from B, and in this situation, cross-licensing may allow
both parties to use the improved invention.
WHAT IS A PATENT? Patents are important to industry because they give the
innovator a period during which imitations can be
A patent is the grant by a nation state of the exclusive right excluded and the investment in R&D can be recovered.
to commercialize an invention in that state for a limited They are of particular importance to the pharmaceutical
time. During that time (the ‘term’ of the patent, usually industry because once the chemical structure of a drug is
20 years from the filing date) the patent owner can go to published it is usually rather easy to copy the product, and
the courts and enforce his or her rights by suing an because the manufacturing cost of a pharmaceutical is
infringer. An owner who wins the infringement suit can only a small part of the selling price, an imitator who has
get damages or other compensation, and more impor- no R&D costs to recover can sell the product cheaply and
tantly can obtain a court order (an injunction) to stop any still make a profit.

© 2012 Elsevier Ltd. 275

Section | 3 | Drug Development

‘equivalence’. The consequence is that before doing any-

THE PATENT SPECIFICATION thing in the USA that is even close to the claims of a
granted US patent, you must make sure that you get a
A patent (which strictly speaking is just a one-page certifi- written opinion from a US patent attorney that you are
cate of grant) is in most countries published with a printed not infringing any valid claims. If you do not, and infringe-
patent specification, which typically will be 10–100 pages ment is found, you may find yourself having to pay triple
long, or even more. The patent specification consists of damages for ‘willful infringement’.
three parts, the bibliographic details and abstract, the
description, and the claims. Each part has a different
purpose. WHAT CAN BE PATENTED?

Bibliographic details There are basically only two categories of subject matter
that can be patented – products and processes. Products
The title page usually sets out the bibliographic details, are broadly anything having physical reality, including
giving information such as the names of the inventors, the machines, manufactured articles, chemical compounds,
owner or assignee of the patent, the title, the dates of prior- compositions comprising a mixture of substances, and
ity, filing, publication and grant, and the name of the even living organisms. A process may be a process for
attorney, if any, who acted for the patentee. It may also manufacturing an article or synthesizing a compound, or
give the international search classification, and a list of may be a method of using or testing a product. However, a
prior published documents considered by the Patent patent for a process for making something, for example
Office when examining the application. Generally it will a chemical compound, also covers the direct product of
also have an abstract summarizing the invention; this is that process. A patent claiming simply ‘the compound of
meant as a tool for searching purposes and is not used in formula X’ covers X however it is made, but a process claim
determining the scope of protection given by the patent. to ‘a method of production of X by reacting Y and Z’ covers
X only when made by that process, and not in any other
Description way. A claim to the compound itself covers the compound
not only however it is made but also however it is used.
The longest part of the specification is the description, the Thus a claim to a compound invented as a dyestuff will
purpose of which is to give enough information about the also cover the compound when used as a pharmaceutical.
invention to enable a reader who is technically qualified There are also some types of subject matter for which
in the relevant field to reproduce it. This ensures that when the grant of patents is specifically excluded, and these
the patent is no longer in force the invention will be fully exclusions vary from country to country. For example,
in the public domain and able to be used by anyone some countries do not grant patents on any plants or
having the necessary skills. The description will usually animals, whereas in Europe only specific plant and animal
start with a brief account of the background to the inven- varieties are excluded, and in the USA there is no such
tion, followed by a summary of the invention, then present restriction. Similarly, the USA allows patents for methods
full details, with actual examples where appropriate. There of surgical or medical treatment or diagnosis, whereas
may also be figures (drawings, structural formulae, graphs, most other countries do not. Nevertheless, the invention
photographs, etc.) and if DNA or amino acid sequences that a known drug may be used for a new indication may
are disclosed there will be sequence identifiers in standard usually be protected in these countries by patents having
form. a different form of claim. Generally, patents will not be
granted in any country for aesthetic creations, mathemati-
cal and scientific theories, and discoveries without any
Claims practical application.
At the end of the specification come one or more claims,
which have the legal purpose of setting out exactly what
Pharmaceutical inventions
is covered by the scope of the exclusionary right. Readers
who see that what they wish to do clearly falls within the Within the pharmaceutical field, patentable inventions
claims of someone else’s patent are put upon notice that may include not only new chemical compounds of known
if they go ahead they may be sued for infringement, and structure, but also, for example, biopolymers and mixtures
will have to stop their activities unless they can prove that the structure of which has not been fully elucidated. Iso-
the patent is invalid. Unfortunately the reverse situation is lated DNA sequences and genes are also patentable as
not so clear. In many countries, particularly the USA, even chemical compounds, although in some countries the
an activity that does not fall within the literal wording of scope of protection given by the patent is limited to the
a patent claim may nevertheless be held to infringe by disclosed use. Even if a chemical compound is already

276
 Intellectual property in drug discovery and development Chapter | 19 |

known, it may be possible to patent variants, such as new which a later patent application is invalidated by written
optical isomers and crystal forms of the compound, as publication anywhere in the world, but by oral publica-
well as new galenic formulations, mixtures with other tion or use of the invention only in the home country.
active ingredients, manufacturing and purification proc- Thus a US patent would not be invalidated by the lecture
esses, assay processes, etc. in Ulan Bator, but would be by an account of it published
If a known compound, not previously known to have in a newspaper there. Similarly, prior use in a country
any pharmaceutical use, is found to be useful as a drug, outside the USA would not invalidate a US patent if there
this invention may be protected by claiming a pharmaceu- was no written description, whereas a European patent
tical composition containing the compound, or, in Europe, would be invalidated by prior use anywhere in the world,
the use of the compound as a pharmaceutical. Such claims so long as the use made the invention available to the
will cover all pharmaceutical uses of the compound, not public – for example the sale of a chemical compound that
only the one found by the inventor. If the invention is that could be analysed.
a known drug has a new and unexpected indication, such
an invention may be protected in the USA by a ‘method
Novelty in the USA
of medical treatment’ claim (‘method of treating a human
suffering from disease Y by administering an effective A more basic difference between the USA and all other
amount of a compound X’), or in Europe by a use claim countries is that all countries other than the USA have a
(‘use of compound X for the treatment of disease Y’). ‘first-to-file’ system, whereby if two persons make the same
invention the first one to file a patent application gets the
patent. The present US system is ‘first-to-invent’, so that
REQUIREMENTS FOR PATENTABILITY irrespective of who files the first application, the person
who can prove the earlier invention date gets the patent. A
consequence of this is that in most countries, prior art is
For an invention in any of the above categories to be pat- what is published before the first filing date (the priority
entable, it must meet three basic criteria: date). In the USA, however, prior art is what is published
• It must be novel before the invention date. Since by definition an inventor
• It must involve an inventive step (must not be cannot publish his or her invention before it is invented,
obvious) self-publication cannot normally be prior art. However, if
• It must be industrially applicable (must have utility). an invention has been published, by the inventor or by
another person, after the invention date, a US patent appli-
cation for the invention is regarded as lacking novelty
Novelty unless it is filed within 12 months of the date of publica-
The first and clearest requirement is that nothing can be tion. This means that an inventor may publish the inven-
patentable which is not new. If a patent were to be granted tion and still obtain a valid US patent so long as a US
for something already known, then the grant of a patent application is filed within this 12-month period. In the
in respect of this information would violate the funda- past, many US inventors, particularly those working in
mental principle that a patent cannot deprive the public universities, sought to take advantage of this so-called grace
of rights that it already has. There are, however, different period, only to find that by so doing they had destroyed
definitions of ‘novelty’. The most straightforward is that of their chances of getting any protection in other countries.
‘absolute novelty’ applied in Europe, Japan, China, and Now most US inventors are aware of the dangers, and
the majority of other countries, which provides that an unless they are interested only in obtaining a US patent,
invention is new if it is not part of the ‘state of the art’, the they will adhere to the first-to-file principle and file an
state of the art being defined as everything that was avail- application before publishing their results. Following
able to the public by written or oral publication, use or passage of the America Invents Act in September 2011, the
any other way, in any country in the world, before the US system will change to first-to-file on March 16 2013.
priority date of the invention. For example, if it could be
proved that the invention had been described before that
Inventive step (non-obviousness)
date in a public lecture given in the Mongolian language
in Ulan Bator, a European patent application for the Whereas the concept of novelty is (or should be) an objec-
invention would lack novelty even if no European had tive matter, the question of whether or not something
heard or understood the lecture. involves an inventive step is intrinsically much more dif-
A few countries still have the system of ‘local novelty’, ficult, as subjective judgement is involved. The basic prin-
under which a disclosure of the invention before the ciple to remember is that the reason for requiring the
priority date destroys novelty only if it is available within presence of an inventive step is that ordinary workers in
that country. Rather more countries, including the USA, that field should remain free to apply their normal skills
have an intermediate ‘mixed novelty’ system, according to to making minor variations of old products.

277
Section | 3 | Drug Development

Thus the person to whom the invention must be non- in this chapter. However, some patent issues need to be
obvious in order to be patentable is the ‘person skilled in considered at an earlier stage. Here we consider issues that
the art’, i.e. a worker who is competent but lacks imagina- concern the selection of a compound as a development
tion or inventive capability. In the days when the majority candidate.
of patents were for simple mechanical devices, the person The two questions that need to be answered before
skilled in the art was usually described as an ‘ordinary significant sums are invested in development activities are:
workman’. However, for complex inventions in pharma- • What sort of protection can we get for this
ceutical chemistry and biotechnology, the ‘person skilled compound?
in the art’ may be considered to be a team of highly quali- • What patent rights of others could prevent us from
fied scientists. marketing this compound?
It is a legal fiction to suppose that such a team could be
These are two completely different issues. As explained
competent but non-inventive, considering that its members
above, it is perfectly possible to have strong patent protec-
would, if employed in industry, be expected by their
tion for one’s own invention, yet still to be blocked by
company to make inventions, and if academic scientists,
earlier dominating patent rights owned by someone else.
would be expected by their university to produce original
The ‘patent situation’ for a compound should attempt to
scientific work, which amounts to much the same thing.
give the answers to both questions.
The point is that obviousness should be judged by a
person with qualifications and imagination that are
average for those in the field. It is tempting for a party
attacking a patent on the ground of obviousness to use an THE STATE OF THE ART
expert witness with the highest possible qualifications, but
it is not very helpful to have a Nobel laureate testify that The answers to the two basic questions depend upon the
something is obvious. It may be obvious to a genius, but state of the art – patent jargon for all material relating to
the real question is whether it is obvious to the normal the technical field that has been published at the relevant
worker in the field. date. The state of the art (sometimes called the prior art)
It is often very easy to reconstruct an invention with the includes not only published scientific papers, but also, for
benefit of hindsight, as a series of logical steps from the example, what is in textbooks, manufacturers’ brochures,
prior art, but it does not necessarily follow that the inven- newspaper articles, web pages on the Internet and oral
tion was obvious, especially if there is evidence that the presentations at conferences. It also includes patent docu-
invention was commercially successful, or satisfied a need. ments, which may be either granted patents or published
The question ‘If the invention was so obvious, why did no patent applications which have not been examined as of
one do it before?’ is usually a relevant one to ask. the date of publication.
The requirements that a patentable invention must be
novel and must have an inventive step mean that nothing
Industrial applicability (utility)
can be patented that is already part of the state of the art,
In Europe it is a requirement that the invention should be and that anything that is very close to the state of the art
capable of industrial application, which is broadly defined may be very difficult to patent.
and includes making or using the invention in any kind
of industry, including agriculture. In the USA, patentable
inventions are defined as any new and useful process, PATENT DOCUMENTS AS STATE OF
machine, manufacture or composition of matter, or any
new and useful improvement thereof. The US requirement
THE ART
that the invention be useful has generally been applied no
more strictly than the corresponding European require- A granted patent is not only a description, which, like any
ment, but recently US examination guidelines have been kind of prior publication, is part of the state of the art. It
tightened up so as to make more difficult the patenting of also contains claims defining the scope of protection. Pub-
DNA sequences for which no real function is known, and lished patent applications also contain claims, but these
the courts both in the USA and the UK have held such are often much broader than the claims (if any) that will
patents invalid. finally be granted.
Patent documents in the state of the art are therefore
important in answering the second basic question, which
concerns freedom to operate. Granted patents may be invali-
PATENT ISSUES IN DRUG DISCOVERY dated by a court, but as a rule they have a presumption of
validity which is hard to challenge. Patent applications do
The strategies that should be used to obtain patent protec- not give exclusionary rights, but may act as a warning flag
tion for a compound in development are described later for rights that may be granted in the future.

278
 Intellectual property in drug discovery and development Chapter | 19 |

proposed manufacturing process and the chosen pharma-

EVALUATION BY THE SCIENTIST ceutical formulation are also free from third-party patent
rights.
The first person to evaluate the patent situation of a pos-
sible lead compound must be the research scientist
responsible for the project. He or she should be aware of
SOURCES OF INFORMATION
the work being published in the area, and should know
who are the main players in the field, what journals and
other information sources contain relevant information, For patent literature there are now a number of databases
and what competitor companies are likely to be filing available online which allow full-text searching by key-
relevant patent applications. To the extent that the chemi- words. One, available on the website of the US Patent and
cal structure of the lead compound is a matter of choice, Trademark Office (www.uspto.gov), contains fully search-
the research chemist should try to steer away from the able texts of all US patents since 1976, as well as image
known state of the art. files of all US patents back to 1790. A similar database
Where the research is in a field in which a number of (Esp@cenet) available through the home page of the
competitors are active, it is clear that the closer you are to European Patent Office (www.epo.org) allows searching of
the competition, the closer you are to the prior art, and European patent applications and PCT applications. For
the more difficult it will be to obtain patent protection. Japan, the database JAPIO, available through the website
Research managers are sometimes struck by what seems of the Japanese Patent Office (www.jpo.go.jp), gives
like a great idea – ask the patent department to check English-language abstracts of all Japanese early-published
where there are gaps in the competitors’ patent protection, applications from 1976 onwards. Use of these databases
and then try to work in these gaps. This is actually a very is free.
bad idea if the intention is to produce something innova- Other databases, maintained by commercial firms
tive. Research must drive patenting, not the other way which charge user access fees, add value by high-quality
around, and the initial work should be done before the abstracts and additional indexing possibilities, and down-
patent situation is checked. loading and printing information from these may be
quicker and easier than it is from public domain Internet
databases.
Chemical Abstracts (CA), published by the Ohio State
University, abstracts both patents and scientific literature
EVALUATION BY THE PATENT
in the chemical field. The information retrieval system is
PROFESSIONAL based on a CA registry number allocated to every pub-
lished chemical compound; once this has been identified,
A professional evaluation of the patent situation of a new abstracts of all patents or literature articles mentioning the
chemical entity (NCE) needs to be made at about the same compound can be listed, and printed out if required.
time that the filing of a patent application is being con- Derwent Publications Ltd provides a wide range of
sidered. The timing of this will depend upon the patent abstracting and information retrieval systems for both sci-
policy of the company or organization owning the inven- entific and patent literature. The latter includes the WPI
tion, but will usually be at the time the compound is ready (World Patent Index) database, covering all patents in the
to enter the development process, i.e. at the time of transi- major countries issued since 1974. Searches can be made
tion of the compound to ‘drug candidate’ status. on the basis of keywords, or of partial structures of chemi-
The patent situation must be established on the basis of cal compounds. Derwent also has a database covering all
a search of the scientific and patent literature. Ideally, the publications and patents in the field of biotechnology
search should be carried out by a professional patent since 1982.
searcher and evaluated by a patent attorney or patent
agent. However, it is becoming more and more easy for a
patent attorney or a scientist to carry out searches online,
and while these are unlikely to be as complete as those RESULTS OF THE EVALUATION – NCES
done by a professional searcher, such a ‘quick and dirty’
search may be all that is required at this early stage. At If the search shows that the compound lacks novelty, that
some stage after a patent application has been filed, is, it has already been published, then the best course is
searches will be carried out in the major patent offices, and to pick a different one for development. Even though it
these can be used to supplement the search made at the may be possible to obtain some form of secondary patent
time of filing. protection, for example the use of the compound as a
More complete ‘freedom to operate’ searches must medicament if it was previously known for a non-
be made at later stages, for example to ensure that the pharmaceutical use, most companies will not invest in the

279
Section | 3 | Drug Development

development of a compound unless it will be possible to newly discovered mechanism have been held invalid as
patent the compound itself. claiming ‘not an invention but a research programme’. It
If the NCE appears to be novel but the search shows that is probably an acceptable risk to ignore patents of these
very similar compounds are known, so that the compound types, although some risk of litigation is unavoidable.
may lack an inventive step, the best advice is to go ahead Should other types of research tool patents be the
anyway. If the compound proves to have superior proper- subject of a freedom to operate search at an early stage of
ties compared to the known product, these can be used to product development? Probably not. There are simply too
establish the inventive step. If it does not, it will drop out many of them, and if no research project could be started
of development whatever the patent situation. without clearance on the basis of such a search, nothing
If the NCE, despite being apparently novel and inven- would ever get done. At least for a large company, it is an
tive, appears to be covered by a third-party patent, the acceptable business risk to go ahead and assume that if
advice would be to go ahead only if the third-party patent problems arise they can be dealt with at a later stage, nor-
appears to be invalid or will have expired before your mally by taking a licence under any relevant patent.
product can reach the market, or if you are sure that you
will be able to obtain a licence on acceptable terms.

OBTAINING PATENT PROTECTION

PATENTING OF RESEARCH TOOLS FOR A DEVELOPMENT COMPOUND

In addition to patent issues relating to the compound Filing a patent application

itself, research tools may also be covered by patents.
By ‘research tool’ is meant anything that contributes to When to file
the discovery or development of a drug, without being Given that in most countries the first to file an application
part of the final product. Examples include genes, cell gets the patent, it would seem to make sense to file as early
lines, reagents, markers, assays, screening methods, animal as possible as soon as an invention is made. It is not quite
models, etc. as simple as this, however. For one thing, the earlier a
A company whose business it is to sell drugs is not patent is filed the earlier it will expire, and particularly in
usually interested in patenting research tools, but it is the the pharmaceutical field, the last year or two of patent life
business of many biotech companies to develop and com- for a major product can be worth hundreds of millions of
mercialize such tools, and these companies will naturally dollars. For another, a patent application filed at a very
wish to obtain patent protection for them. For pharma- early stage may lack sufficient enabling disclosure to
ceutical companies, such research tool patents and appli- support claims of the desired scope. Too much delay,
cations raise issues of freedom to operate, particularly if however, and another party may have filed an earlier appli-
they contain ‘reach-through’ claims purporting to cover cation or published a paper that destroys the novelty of
drugs found by using the patented tools. the invention.
Some scientists may believe that research activities, in For pharmaceutical inventions, the decision when to
contrast to the manufacture and sale of a product, cannot make a first filing will depend on a number of factors,
be patent infringement. This is not the case. If I have including the intensity of competition in the relevant field.
invented a process that is useful in research and have a As a general rule, however, it is generally best to wait at
valid patent for it, I can enforce that patent against anyone least until one or more lead compounds within the scope
using the process without my permission. I can make have shown clear activity in a validated in vitro assay, or
money from my patent by granting licences for a flat fee, in an animal model, i.e. close to the point at which a drug
or a fee based on the extent to which the process is used; candidate (see Chapter 4) is identified.
or by selling kits for carrying out the process or reagents
for use in the process (for example the enzymes used in
the polymerase chain reaction (PCR) process). What I am Where to file
not entitled to do is to charge a royalty on the sale of drugs Normally a single filing in one country will be made,
developed with the help of my process. I can patent an which, under the Paris Convention for the Protection of
electric drill, but I should not expect to get a royalty on Industrial Property, can form the basis for a claim to prior-
everything it bores a hole in. ity in other countries. Some national laws, such as those
Although some patents have been granted containing of the USA and France, require that, for reasons of national
claims that would be infringed, for example, by the sale of security, an application for any invention made in that
a drug active in a patented assay, such patents have been country must first be filed in that country (unless special
successfully challenged in the courts, both in the USA and permission is obtained). The UK now limits this require-
the UK. Similarly, patents claiming all drugs acting by a ment to certain categories of inventions, but it may be

280
 Intellectual property in drug discovery and development Chapter | 19 |

safer to file all UK-originating inventions in the UK first. application in his home country long enough for it to
Other countries, for example Switzerland, are less para- issue as a published application (see below) or by sending
noid, and allow a first filing to be made in any country. it to a journal such as Research Disclosure, in which any
The Paris Convention, now adhered to by the great disclosure may be rapidly published for a reasonable fee.
majority of countries, provides that a later application
filed for the same invention in another Convention Refiling
country within 12 months of the first filing in a Conven- It frequently happens that by the time the foreign filing
tion country may claim the priority of the original applica- decision must be taken it is not yet possible to decide
tion. This means that the first filing date (the priority date) whether or not to invest time and money in foreign pat-
is treated for prior art purposes as if it were the filing date enting. Commercial interest may be low but could increase
of the later application, so that a publication of the inven- later, more testing may have to be done, or the inventors
tion before the later application but after the priority date may not have done any more work on the invention since
does not invalidate claims for the same invention in the the first application was filed. In such cases the best solu-
later application. If it were not for the Paris Convention, tion is to start from the beginning again. The existing
it would be necessary to make simultaneous filings in all application is abandoned, a new application is filed, and
the countries of interest at a very early stage, which would the 12-month countdown starts all over again. In this case
be extremely wasteful of time and money. Instead, a single it is essential to meet the requirements of the Paris Con-
priority filing may be made and a decision taken before vention that the first application be explicitly abandoned
the end of the priority year on what to do with the before the second application is filed.
application. Of course, refiling always entails a loss of priority,
During the priority year, work on the invention will usually of 8–10 months, and if someone else has pub-
normally continue, and for example further compounds lished the invention or filed a patent application for it
will be made and tested, new formulations compounded, during this time, the refiled application cannot lead to a
or new process conditions tried. All this material can be valid patent. Consequently, in a field where competitors
used in preparing the patent applications to be filed are known to be active refiling may involve an unaccept-
abroad, and, where possible, a subsequent application in able risk, and, naturally, if there has been any known
respect of the country of first filing. It is also possible to publication of the invention since the priority date, aban-
file new patent applications for further developments donment and refiling is ruled out. Such publication most
made during the priority year, and then at the foreign frequently arises from the inventor himself. Most inven-
filing stage to combine these into a single application. The tors know that they should not publish inventions before
advantage of this is that the new developments will then a patent application is filed; it is not so generally realized
have an earlier priority date (the date of filing of the new that publication within the priority year can also be very
application) than they otherwise would have (the date of damaging.
filing of the foreign text).
Home-country patenting
The foreign filing decision If the applicant is an individual or a small company having
There are four options to be considered: no commercial interests or prospects of licensing outside
• Abandon the home country (which will usually be the country in
• Abandon and refile which the first filing is made), the expense of foreign filing
• Obtain a patent in the country of first filing only would be wasted, and the applicant will wish only to
• File corresponding applications in one or more obtain a patent in the home country. Even where the
foreign countries. applicant is a larger company that would normally file any
commercially interesting case in several countries, indi-
Abandonment vidual applications may be of such low interest that pro-
If there is no commercial interest in the invention at all, tection in the home country is all that is needed. This
or if a search has shown that it lacks novelty, one can option is, of course, more attractive if the home country
simply do nothing. Sooner or later a fee must be paid or is a large market such as the USA, rather than a small
some action taken to keep the application in being, and country such as Switzerland.
when this is not done the application will lapse. It is best
not to withdraw the application explicitly, as such a posi- Foreign filing
tive abandonment is usually irrevocable and applicants Finally, if an invention appears likely to be commercially
have been known to change their minds. important, the decision may be to file corresponding
If the applicant wants to ensure that he retains freedom applications in a number of other countries. For the phar-
to operate and that no one else can patent the invention, maceutical industry one can assume that the costs of
he should have it published, either by continuing an patent protection would be small compared with the value

281
Section | 3 | Drug Development

of protection for any compound that actually reaches the Patent cooperation treaty (PCT)
market, but at the time when a foreign filing decision must
The PCT allows rights to be established in a large number
be taken, it is usually impossible to estimate the chance
of countries (144 as of April 2012) by a single inter­
that the product in question will progress that far. Accord-
national application. Search and optional preliminary
ingly, one must rely upon some rule of thumb such that
examination are carried out before the application goes to
if the product is being developed further, foreign filing
the national or regional patent offices. This system gives
should be carried out as a matter of course. High patenting
the maximum flexibility and allows the costs associated
costs are a necessary part of the high research overheads
with translations, etc., to be significantly postponed. There
of the pharmaceutical industry.
are now very few economically significant countries that
are not members of the PCT, of which the most important
Procedures on foreign filing is Taiwan. An initial international phase, in which a search
and possibly also a preliminary examination is carried
National filings out, is followed after 18 months by a national phase, in
It is possible to file patent applications (in the local lan- which selected national or regional patent offices con-
guage) in the national patent offices of each selected clude the examination process and grant (or refuse) the
country individually. This involves a large outlay of money patent. The PCT procedure is described in more detail in
at a relatively early stage, and also means that all necessary Box 19.1.
translations must be prepared in good time before the end
of the priority year. It is also very labour-intensive, as the Selection of countries
application must be prosecuted separately before each In deciding the list of countries in which patent protection
national patent office. Fortunately, there are ways to sim- should be obtained, the main criteria are the strength of
plify the procedure. patent protection in the country and the size of the market.
Now that most countries have joined the World Trade
Regional patent offices Organization and are obliged by the TRIPs (Trade-Related
Aspects of Intellectual Property Rights) agreement to intro-
One is that there are certain regional patent offices by
duce strong patent protection, the most important crite-
which patents in a number of countries can be granted
rion has become market size. There is no point in filing
based on a single application filed and prosecuted in one
patents in a country if the size of the market does not
patent office. By far the most important of these is the
justify the costs, no matter how strong its patent laws may
European Patent Office, which as of April 2012 grants
be. Nevertheless, for a new chemical entity that may
patents for a total of 37 countries. These are all the 27
become a market product, filing in 40–60 countries is
current EU states plus Albania, Croatia, Iceland, Liechten-
normal practice. To avoid long discussions each time a
stein, Macedonia, Monaco, Norway, San Marino, Switzer-
decision must be taken, the use of standard filing lists to
land and Turkey. The European application can be filed in
cover most situations makes a lot of sense.
English, French or German, and translations into other
languages are required only at the time of grant. Once the
Maintenance of patents
European patent is granted, opposition to the patent may
be filed by any other party within 9 months of the date of In nearly all countries, periodic (usually annual) renewal
grant. If the opposition is wholly or partly successful, the fees must be paid to keep a patent in force. These generally
patent is invalidated or limited in scope for all of the increase steeply towards the end of the patent term, thus
designated countries. encouraging patent owners who are not making commer-
Although the European Patent Convention provides for cial use of their patents to make the invention available to
a central filing, grant and opposition procedure, once the the public earlier than would otherwise be the case. To
European patent is granted it is treated as if it were a save costs, pharmaceutical patents should be abandoned
bundle of national patents in the designated contracting as soon as they no longer provide protection for a com-
states, so that, for example, the European patent may be pound that is on the market or is being developed. Main-
invalidated by the courts in one country without directly taining a collection of patents that are not being used is
affecting its validity in other countries. Proposals have an expensive luxury.
repeatedly been made by the European Commission for a
single unitary patent to cover all EU countries, just as a Extension of patent term
single US patent covers all 50 states, but little real progress The standard patent term provided in the TRIPs agreement
has been made in this direction. is at least 20 years from the filing date. However, because
Other regional patent offices are the Eurasian Patent it takes a long time to bring a drug to market, the effective
Office (Russia and certain former Soviet countries), and term (the term during which a drug is sold with patent
ones for English-speaking and French-speaking African protection) is much less than this. To compensate for these
countries. regulatory delays, a number of countries, including EU

282
 Intellectual property in drug discovery and development Chapter | 19 |

for FDA approval during the patent term, so that they can
Box 19.1  PCT procedure come on the market as soon as patent protection expires.
In Europe, extension is provided by means of a separate
International phase
form of intellectual property right known as a Supplemen-
Filing tary Protection Certificate (SPC).
An international application can be filed by any national
or resident of a PCT country, at a national or regional Enforcement of patent rights
patent office competent to act for that applicant, or at
the International Bureau (World Intellectual Property Governments grant patents, but do not enforce them. The
Office, or WIPO) in Geneva. A single filing fee can give patent owner must take action against infringement by
rights in all Contracting States. suing an infringer in the civil courts. If successful, the
patentee can obtain an injunction to restrain further
International publication and search report
infringement, as well as other remedies such as damages
The PCT application is published 18 months from the first and costs. Usually the alleged infringer will counterclaim
priority date, and the search report drawn up by the that the patent is invalid, and if the patentee loses the case
International Searching Authority (selected from one of a the patent may be revoked. This risk, as well as the high
number of patent offices including the USPTO and the cost of litigation, must be weighed against the benefit
EPO) is published at the same time or as soon as possible
gained if the infringer is forced out of the market. As an
afterwards. At the same time, a Written Opinion on
alternative to litigation, the patentee may choose to exploit
Patentability is drawn up, indicating on the basis of the
the patent by granting exclusive or non-exclusive licences
search report whether or not the invention appears to be
new and non-obvious. If no further steps are taken, this
for royalties or other forms of compensation, or in
will be issued as the International Preliminary Report on exchange for a cross-licence.
Patentability (IPRP). Although the procedure for obtaining a patent has been
harmonized to a large extent by the PCT and other means,
International preliminary examination the procedure for enforcement, as well as the cost and the
If the applicant wishes to contest the findings of the chance of success, varies enormously from one country to
Written Opinion, he may within 22 months from the another. In the USA, patent cases are heard at first instance
priority date file a Demand for International Preliminary in the Federal District Courts, in which the judges are not
Examination, pay a fee and respond to the Written specialized in intellectual property law and in which many
Opinion, possibly also making amendments. This will then cases are decided by jury verdicts. At the appeal stage,
be taken into account in the final form of the IPRP. however, the Court of Appeal for the Federal Circuit is a
National phase specialized and technically competent court. In England
After 30 months from the priority date the application (a separate jurisdiction from Scotland), on the other hand,
may be sent to any of the national or regional patent patent cases are heard either in the Patents County Court
offices, translated into the local language as necessary. or, more usually, in the Patents Court, which is part of the
The individual patent offices may rely on the international High Court. Both of these are specialized courts with tech-
search and examination reports to any extent they choose nically literate judges, but appeals from them go to the
in deciding whether or not to grant a patent. This varies general Court of Appeal, where the majority of the judges
from offices which usually ignore the IPRP altogether (e.g. are not patent experts. Which of these two systems gives
the USPTO), to those which will grant a patent without the fairest results is a matter of debate.
further examination only if the IPRP is positive (e.g. In both the USA and the English systems issues of patent
Turkey), to Singapore, which will automatically grant a validity are dealt with by the same court that deals with
patent on any PCT application with an IPRP, whether it is the issue of infringement, and this is also the case in the
positive or negative. Singapore very sensibly puts the majority of European and Asian countries. In Germany,
burden on the applicant, who, if he wishes to enforce the Japan, China and Korea, however, these issues are kept
patent, would have to prove to the court that the separate, and a patent may be invalidated only by a special
negative IPRP was incorrect. court or by a branch of the patent office.
It is a problem in many parts of the world that even if
the country has a good patent law on paper, enforcement
of patent rights may be very difficult for a number of
states, the USA, Switzerland and Japan, allow for patent
reasons, ranging from lack of experienced judges to inef-
term extensions of up to 5 years for pharmaceutical (and
ficiency and even corruption.
sometimes agricultural) products. In the USA, patent term
extension is one part of the Hatch–Waxman Act, in which
the interests of the innovative companies are balanced Other forms of intellectual property
against those of the generic companies. The former get a A trademark is a word, design, shape or colour used to
longer patent term, the latter are allowed carry out testing distinguish the goods of the trademark owner from those

283
Section | 3 | Drug Development

of another manufacturer. Unlike patents, registered trade- Non-proprietary Name (INN) or his own trademark, not
marks may be renewed at the end of their term and may that of the originator.
be kept alive indefinitely, although they may be liable to Additional forms of IP include copyright (e.g. for the text
cancellation if they are not used. Thus, once a patent for of advertisements and package inserts), and Internet
a drug has expired a competitor will be able to sell a domain names, which may, for example, incorporate the
generic version, but must sell it under the International name of a product and may be a useful marketing tool.

FURTHER READING

Dutfield G. Intellectual property edn. Oxford: Oxford University Reid B. A practical guide to patent
rights and the life science Press; 2010. law. London: Sweet & Maxwell;
industries. Aldershot: Ashgate Press; Kleemann A, Engel J. Pharmaceutical 1999.
2003. substances: syntheses, patents, Rosenstock J. The law of chemical
Grubb P, Thomsen P. Patents for applications. New York: Thieme and pharmaceutical invention:
chemicals, pharmaceuticals and Medical; 2001. patent and nonpatent protection.
biotechnology: fundamentals of Old F. Inventions, patents, brands and New York: Aspen Publishers;
global law, practice and strategy. 5th designs. Sydney: Patent Press; 1993. 1998.

USEFUL WEBSITES

Patent offices Professional organizations Lists of links

EPO http://www.epo.org Chartered Institute of Patent Attorneys: http://www.epo.co.at/online/index.htm
UK http://www.patent.gov.uk or http://www.cipa.org.uk http://portico.bl.uk/collections/patents/
www.ukpats.org.uk European Patents Institute: http:// html
USA http://www.uspto.gov www.patentepi.com
Japan http://www.jpo.go.jp American Intellectual Property Law
WIPO http://www.wipo.int Association: http://www.aipla.org

284
 Chapter 20 
Regulatory affairs
I Hägglöf, Å Holmgren

uses. These developed into official pharmacopoeias, of

INTRODUCTION which the earliest was probably the New Compound Dis-
pensatory of 1498 issued by the Florentine guild of physi-
This chapter introduces the reader to the role of the regula- cians and pharmacists.
tory affairs (RA) department of a pharmaceutical company, The pharmacopoeias were local rules, applicable in a
outlining the process of getting a drug approved, and particular city or district. During the 19th century national
emphasizing the importance of interactions of regulatory pharmacopoeias replaced local ones, and since the early
affairs with other functions within the company, and with 1960s regional pharmacopoeias have successively replaced
the external regulatory authorities. national ones. Now work is ongoing to harmonize – or
To keep this chapter to a reasonable size the typical at least mutually recognize – interchangeable use of the
examples given refer to the first registration of a new chem- US Pharmacopeia, the European Pharmacopoeia and the
ical compound. The same way of reasoning also applies, Japanese Pharmacopoeia.
however, to any subsequent change to the approval of As described in Chapter 1, the development of experi-
products. Depending on the magnitude of the change, the mental pharmacology and chemistry began during the
new documentation that needs to be compiled, submitted second half of the 19th century, revealing that the effect of
and approved by health authorities is variable, ranging the main botanical drugs was due to chemical substances
from a few pages of pharmaceutical data (e.g. for an in the plants used. The next step, synthetic chemistry, made
update to product stability information) to a complete it possible to produce active chemical compounds. Other
new application for a new clinical use in a new patient important scientific developments, e.g. biochemistry, bac-
group in a new pharmaceutical form. teriology and serology during the early 20th century, accel-
It needs also to be said that, as every drug substance and erated the development of the pharmaceutical industry
every project is unique, the views expressed represent the into what it is today (see Drews, 1999).
opinion of the authors and are not necessarily shared by Lack of adequate drug control systems or methods to
others active in the field. investigate the safety of new chemical compounds became
a great risk as prefabricated drug products were broadly
and freely distributed. In the USA the fight against patent
medicines led to the passing of the US Pure Food and
BRIEF HISTORY OF
Drugs Act against misbranding as long ago as 1906. The
PHARMACEUTICAL REGULATION Act required improved declaration of contents, prohibited
false or misleading statements, and required content and
Control of pharmaceutical products has been the task of purity to comply with labelled information. A couple of
authorized institutions for thousands of years, and this decades later, the US Food and Drug Administration
was the case even in ancient Greece and Egypt. (FDA) was established to control US pharmaceutical
From the Middle Ages, control of drug quality, composi- products.
tion purity and quantification was achieved by reference Safety regulations in the USA were, however, not enough
to authoritative lists of drugs, their preparation and their to prevent the sale of a paediatric sulfanilamide elixir

© 2012 Elsevier Ltd. 285

Section | 3 | Drug Development

containing the toxic solvent diethylene glycol. In 1937, The 1960s and 1970s saw a rapid increase in laws, regu-
107 people, both adults and children, died as a result of lations and guidelines for reporting and evaluating the
ingesting the elixir, and in 1938 the Food Drug and Cos- risks versus the benefits of new medicinal products. At the
metics Act was passed, requiring for the first time approval time the industry was becoming more international and
by the FDA before marketing of a new drug product. seeking new global markets, but the registration of medi-
The thalidomide disaster further demonstrated the lack cines remained a national responsibility.
of adequate drug control. Thalidomide (Neurosedyn®, Although different regulatory systems were based on the
Contergan®) was launched during the last years of the same key principles, the detailed technical requirements
1950s as a non-toxic treatment for a variety of conditions, diverged over time, often for traditional rather than scien-
such as colds, anxiety, depression, infections, etc., both tific reasons, to such an extent that industry found it neces-
alone and in combination with a number of other com- sary to duplicate tests in different countries to obtain
pounds, such as analgesics and sedatives. global regulatory approval for new products. This was a
The reason why the compound was regarded as harmless waste of time, money and animals’ lives, and it became
was the lack of acute toxicity after high single doses. After clear that harmonization of regulatory requirements was
repeated long-term administration, however, signs of neu- needed.
ropathy developed, with symptoms of numbness, paraes- European (EEC) efforts to harmonize requirements for
thesia and ataxia. But the overwhelming effects were the drug approval began 1965, and a common European
gross malformation in infants born to mothers who had approach grew with the expansion of the European Union
taken thalidomide in pregnancy: their limbs were partially (EU) to 15 countries, and then 27. The EU harmonization
or totally missing, a previously extremely rare malforma- principles have also been adopted by Norway and Iceland.
tion called phocomelia (seal limb). Altogether around This successful European harmonization process gave
12 000 infants were born with the defect in those few years. impetus to discussions about harmonization on a broader
Thalidomide was withdrawn from the market in 1961/62. international scale (Cartwright and Matthews, 1994).
This catastrophe became a strong driver to develop
animal test methods to assess drug safety before testing
compounds in humans. Also it forced national authorities
to strengthen the requirements for control procedures INTERNATIONAL HARMONIZATION
before marketing of pharmaceutical products (Cartwright
and Matthews, 1991). The harmonization process started in 1990, when repre-
Another blow hit Japan between 1959 and 1971. The sentatives of the regulatory authorities and industry asso-
SMON (subacute myelo-optical neuropathy) disaster was ciations of Europe, Japan and the USA (representing the
blamed on the frequent Japanese use of the intestinal majority of the global pharmaceutical industry) met,
antiseptic clioquinol (Entero-Vioform®, Enteroform® or ostensibly to plan an International Conference on Harmo-
Vioform®). The product had been sold without restrictions nization (ICH). The meeting actually went much further,
since early 1900, and it was assumed that it would not be suggesting terms of reference for ICH, and setting up an ICH
absorbed, but after repeated use neurological symptoms Steering Committee representing the three regions.
appeared, characterized by paraesthesia, numbness and The task of ICH was ‘… increased international harmo-
weakness of the extremities, and even blindness. SMON nization, aimed at ensuring that good quality, safe and
affected about 10 000 Japanese, compared to some 100 effective medicines are developed and registered in the
cases in the rest of the world (Meade, 1975). most efficient and cost-effective manner. These activities
These tragedies had a strong impact on governmental are pursued in the interest of the consumer and public
regulatory control of pharmaceutical products. In 1962 health, to prevent unnecessary duplication of clinical trials
the FDA required evidence of efficacy as well as safety as in humans and to minimize the use of animal testing
a condition for registration, and formal approval was without compromising the regulatory obligations of safety
required for patients to be included in clinical trials of and effectiveness’ (Tokyo, October 1990).
new drugs. ICH has remained a very active organization, with
In Europe, the UK Medicines Act 1968 made safety substantial representation at both authority and industry
assessment of new drug products compulsory. The Swedish level from the EU, the USA and Japan. The input of other
Drug Ordinance of 1962 defined the medicinal product nations is provided through World Health Organization
and required a clear benefit–risk ratio to be documented representatives, as well as representatives from Switzerland
before approval for marketing. All European countries and Canada.
established similar controls during the 1960s. ICH conferences, held every 2 years, have become a
In Japan, the Pharmaceutical Affairs Law enacted in 1943 forum for open discussion and follow-up of the topics
was revised in 1961,1979 and 2005 to establish the current decided. The important achievements so far are the scien-
drug regulatory system, with the Ministry of Health and tific guidelines agreed and implemented in the national/
Welfare assessing drugs for quality, safety and efficacy. regional drug legislation, not only in the ICH territories

286
 Regulatory affairs Chapter | 20 |

The regulatory authority:

Step 5 Implementation in the three regions • approves clinical trial applications

• gives procedural and scientific advice to companies
during drug development
Step 4 Agreement on a harmonized ICH guideline
Adopted by regulators
• approves for marketing drugs that have been
scientifically evaluated to provide evidence of a
Step 3 Regulatory consultation in the three regions satisfactory benefit/risk ratio
Consolidation of the comments • monitors the safety of the marketed product, based
on (a) reports of adverse reactions from healthcare
Step 2 Agreement by the Steering Committee to release providers, and (b) from compiled and evaluated
the draft consensus text for wider consultation safety information from the company that owns the
Step 1 Building scientific concensus in joint Regulatory/Industry product
Expert Working Groups • can withdraw the licence for marketing in serious
cases of non-compliance (e.g. failure on inspections,
failure of adequate additional warnings in
Fig. 20.1  Five steps in the ICH process for harmonization of prescribing information after clinical adverse
technical issues. reactions are reported, or failure of the company to
consider serious findings in animal studies).
The company:
but also in other countries around the world. So far some • owns the documentation that forms the basis for
50 guidelines have reached ICH approval and regional assessment, is responsible for its accuracy and
implementation, i.e. steps 4 and 5 (Figure 20.1). For a correctness, for keeping it up to date, and for ensuring
complete list of ICH guidelines and their status, see the that it complies with standards set by current scientific
ICH website (website reference 1). development and the regulatory authorities
The process described in Figure 20.1 is very open, and • collects, compiles and evaluates safety data, and
the fact that health authorities and the pharmaceutical submits reports to the regulatory authorities at
industry collaborate from the start increases the efficiency regular intervals – and takes rapid action in serious
of work and ensures mutual understanding across regions cases. This might involve the withdrawal of the entire
and functions; this is a major factor in the success of ICH. product or of a product batch (e.g. tablets containing
the wrong drug or the wrong dose), or a request to
the regulatory authority for a change in prescribing
information
ROLES AND RESPONSIBILITIES  
• has a right to appeal and to correct cases of
OF REGULATORY AUTHORITY   non-compliance.
AND COMPANY
The role of the regulatory  
The basic division of responsibilities for drug products is
affairs department
that the health authority is protecting public health and
safety, and the pharmaceutical company is responsible for The regulatory affairs (RA) department of a pharmaceuti-
all aspects of the drug product. The regulatory approval of cal company is responsible for obtaining approval for new
a pharmaceutical product permits marketing and is a con- pharmaceutical products and ensuring that approval is
tract between the regulatory authority and the pharmaceu- maintained for as long as the company wants to keep the
tical company. The conditions of the approval are set out product on the market. It serves as the interface between
in the dossier and condensed in the prescribing informa- the regulatory authority and the project team, and is the
tion. Any change that is planned must be forwarded to the channel of communication with the regulatory authority
regulatory authority for information and, in most cases, as the project proceeds, aiming to ensure that the project
new approval before being implemented. plan correctly anticipates what the regulatory authority
To protect the public health, regulatory authorities also will require before approving the product. It is the respon-
develop regulations and guidelines for companies to sibility of RA to keep abreast of current legislation, guide-
follow in order to achieve a balance between the possible lines and other regulatory intelligence. Such rules and
risks and therapeutic advantages to patients. The authori- guidelines often allow some flexibility, and the regulatory
ties’ work is partly financed by fees paid by pharmaceutical authorities expect companies to take responsibility for
companies. Fees may be reduced, under certain condi- deciding how they should be interpreted. The RA depart-
tions, to stimulate research. This may be driven, e.g., by ment plays an important role in giving advice to the
company size or size of target patient groups. project team on how best to interpret the rules. During

287
Section | 3 | Drug Development

the development process sound working relations with proposed. In the ICH environment there is a greater pos-
authorities are essential, e.g. to discuss such issues as diver- sibility to exert influence at an early stage.
gence from guidelines, the clinical study programme, and
formulation development.
Most companies assess and prioritize new projects
based on an intended Target Product Profile (TPP). The RA
THE DRUG DEVELOPMENT PROCESS
professional plays a key role in advising on what will be
realistic prescribing information (‘label’) for the intended An overview of the process of drug development is given
product. As a member of the project team RA also contrib- in Chapters 14–18 and summarized in Figure 20.2. As
utes to designing of the development programme. The RA already emphasized, this sequential approach, designed to
department reviews all documentation from a regulatory minimize risk by allowing each study to start only when
perspective, ensuring that it is clear, consistent and com- earlier studies have been successfully completed, is giving
plete, and that its conclusions are explicit. The department way to a partly parallel approach in order to save develop-
also drafts the core prescribing information that is the ment time.
basis for global approval, and will later provide the plat- All studies in the non-clinical area – chemistry, pharma-
form for marketing. The documentation includes clinical cology, pharmacokinetics, pharmaceutical development
trials applications, as well as regulatory submissions for and toxicology – aim to establish indicators of safety and
new products and for changes to approved products. The efficacy sufficient to allow studies and use in man. Accord-
latter is a major task and accounts for about half of the ing to ICH nomenclature, documentation of chemical and
work of the RA department. pharmaceutical development relates to quality assessment,
An important proactive task of the RA is to provide animal studies relate to safety assessment, and studies in
input when legislative changes are being discussed and humans relate to efficacy.

Preclinical studies Clinical studies

Discovery Development

Chemistry/
pharmacology IND* Phase I Phase II Phase III NDA** Phase IV

Search for active Regulatory Efficacy studies Clinical studies Comparative Regulatory Continued
substances view on healthy on a limited scale studies on a review comparative
Toxicology, volunteers 100 – 200 large number studies
efficacy studies 50 –150 patients of patients Registration,
on various persons 500 – 5000 market
types of animals patients introduction
*Investigational
new drug
Application for Knowledge level
permission to
administer a new **New drug
drug to humans application
Application for
Knowledge level permission to
market
a new drug

Time span

2–4 years 2–6 months 3–6 years 1–3 years

10 –13 years from idea to marketable drug

Fig. 20.2  The drug development process.

288
 Regulatory affairs Chapter | 20 |

Quality assessment (chemistry and not to create local rules to avoid, as far as possible, differ-
ent interpretations and duplicate work.
pharmaceutical development)
As previously said, all changes to the originally submit-
The quality module of a submission documents purity and ted dossier must be made known to the approving regula-
assay for the drug substance, and purity data for all the tory authority. Since the majority of changes are made in
inactive ingredients. The formulation must fulfil require- the quality section, very small changes take lots of resources
ments for consistent quality and allow storage, and the for the company as well as for the authority. The US leg-
container must be shown to be fit for its purpose. These islation has allowed the submission of annual reports col-
aspects of a pharmaceutical product have to be kept under lecting those changes that have no impact on quality. This
control throughout the development process, as toxicol- possibility has also been introduced in EU in 2010 with
ogy and pharmacology results are reliable only for sub- the purpose of saving time and resources.
stances of comparable purity. Large-scale production,
improved synthetic route, different raw material supply,
etc., may produce a substance somewhat different from Safety assessment (pharmacology
the first laboratory-scale batches. Any substantial change and toxicology)
must be known and documented. Next we consider how to design and integrate pharmaco-
The formulation of a product is a challenge. For initial logical and toxicological studies in order to produce ade-
human studies, simple i.v. and oral solutions are needed quate documentation for the first tests in humans. ICH
for straightforward results, whereas for the clinical pro- guidelines define the information needed from animal
gramme in patients, bioequivalent formulations are essen- studies in terms of doses and time of exposure, to allow
tial for comparison of results across studies, and so it is clinical studies, first in healthy subjects and later in
preferable to have access to the final formulation already patients. The principles and methodology of animal
during Phase II. studies are described in Chapters 11 and 15. The questions
If the formulation intended for marketing cannot be discussed here are when and why these animal studies are
completed until late in the clinical phase, bioequivalence required for regulatory purposes.
studies showing comparable results with the preliminary
and final market formulations will be necessary to support
the use of results with the preliminary formulation. There Primary pharmacology
may even be situations when clinical studies must be The primary pharmacodynamic studies provide the first
repeated. evidence that the compound has the pharmacological
The analytical methods used and their validation must effects required to give therapeutic benefit. It is a clear
be described. Manufacturing processes and their valida- regulatory advantage to use established models and to be
tion are also required to demonstrate interbatch uniform- able at least to establish a theory for the mechanism of
ity. However, full-scale validation may be submitted when action. This will not always be possible and is not a firm
sales production has eventually started. requirement, but proof of efficacy and safety is helped by
Studies on the stability of both substance and products a plausible mechanistic explanation of the drug’s effects,
under real-life conditions are required, covering the full and this knowledge will also become a very powerful tool
time of intended storage. Preliminary stability data are for marketing. For example, understanding the mecha-
sufficient for the start of clinical studies. The allowable nism of action of proton pump inhibitors, such as
storage time can be increased as data are gathered and omeprazole (see Chapter 4), was important in explaining
submitted. Even marketing authorizations can be approved their long duration of action, allowing once-daily dosage
on less than real-time storage information, but there is a of compounds despite their short plasma half-life.
requirement to submit final data when available.
Inactive ingredients as well as active substances need to
be documented, unless they are well known and already General pharmacology
documented. Even then it may become necessary to General pharmacology1 studies investigate effects other
perform new animal studies to support novel uses of com- than the primary therapeutic effects. Safety pharmacology
monly used additives. studies (see Chapter 15), which must conform to good
Although the quality module of the documentation is laboratory practice (GLP) standards, are focused on iden-
the smallest, the details of requirements and the many tifying the effects on physiological functions that in a
changes needed during development and maintenance of clinical setting are unwanted or harmful.
a product make it the most resource intensive module
from a regulatory perspective. Also, legislation differs most
1
in this area, so it will often be necessary to adapt the docu- There are a number of widely used terms with similar meanings,
e.g. secondary pharmacology, safety pharmacology, high-dose
mentation for the intended regional submission. RA pro- pharmacology, regulatory pharmacology and pharmacodynamic
fessionals, however, try to convince regulatory authorities safety.

289
Section | 3 | Drug Development

Although the study design will depend on the properties Toxicology

and intended use of the compound, general pharmacol-
The principles and methodology of toxicological assess-
ogy studies are normally of short duration (i.e. acute,
ment of new compounds are described in Chapter 15.
rather than chronic, effects are investigated), and the
Here we consider the regulatory aspects.
dosage is increased until clear adverse effects occur. The
In contrast to the pharmacological studies, toxicological
studies also include comparisons with known compounds
studies generally follow standard protocols that do not
whose pharmacological properties or clinical uses are
depend on the compound characteristics. Active compara-
similar.
tors are not used, but the drug substance is compared at
When required, e.g. when pharmacodynamic effects
various dose levels to a vehicle control, given, if possible,
occur only after prolonged treatment, or when effects seen
via the intended route of administration.
with repeated administration give rise to safety concerns,
the duration of a safety pharmacology study needs to be
Single and repeated-dose studies
prolonged. The route of administration should, whenever
possible, be the route intended for clinical use. The acute toxicity of a new compound must be evaluated
There are cases when a secondary pharmacological prior to the first human exposure.
effect has, eventually, been developed into a new indica- This information is obtained from dose escalation
tion. Lidocaine, for example, was developed as a local studies or dose-ranging studies of short duration. Lethality
anaesthetic agent and its cardiac effects after overdose were is no longer an ethically accepted endpoint. The toxicol-
considered a hazard. Later that cardiac effect was exploited ogy requirements for the first exploratory studies in man
as a treatment for ventricular arrhythmia. are described in the ICH guidline M3 which reached step
All relevant safety pharmacology studies must be 5 in June 2009 (see website reference 1). Table 20.1 shows
completed before studies can be undertaken in patients. the duration of repeated-dose studies recommended by
Complementary studies may still be needed to clarify ICH, to support clinical trials and therapeutic use for dif-
unexpected findings in later development stages. ferent periods.

Genotoxicity
Pharmacokinetics: absorption, distribution, Preliminary genotoxicity evaluation of mutations and
metabolism and excretion (ADME) chromosomal damage (see Chapter 15) is needed before
Preliminary pharmacokinetic tests to assess the absorp- the drug is given to humans. If results from those studies
tion, plasma levels and half-life (i.e. exposure information) are ambiguous or positive, further testing is required. The
are performed in rodents in parallel with the preliminary entire standard battery of tests needs to be completed
pharmacology and toxicology studies (see Chapter 10). before Phase II (see Chapter 17).
Studies in humans normally start with limited short-
term data, and only if the results are acceptable are detailed Carcinogenicity
animal and human ADME studies performed. The objective of carcinogenicity studies is to identify any
Plasma concentrations observed in animals are used to tumorigenic potential in animals, and they are required
predict the concentrations that may be efficacious/ only when the expected duration of therapy, whether con-
tolerated in humans, under the assumption that similar tinuous or intermittent, is at least 6 months. Examples
biological effects should be produced at similar plasma include treatments for conditions such as allergic rhinitis,
levels across species. This is a reasonable assumption pro- anxiety or depression.
vided the in vitro target affinity is similar. Carcinogenicity studies are also required when there is
Investigations during the toxicology programme give particular reason for concern, such as chemical similarities
the bulk of the pharmacokinetic information due to the to known carcinogens, pathophysiological findings in
long duration of drug exposure and the wide range of animal toxicity studies, or positive genotoxicity results.
doses tested in several relevant species. They also give data Compounds found to be genotoxic by in vitro as well as
about tissue distribution and possible accumulation in the in vivo tests are presumed to be trans-species carcinogens
body, including placental transfer and exposure of the with hazards to humans.
fetus, as well as excretion in milk. Carcinogenicity studies normally run for the lifespan of
Metabolic pathways differ considerably between species, the test animals. They are performed quite late in the
often quantitatively but sometimes also qualitatively. development programme and are not necessarily com-
Active metabolites can influence study results, in particular pleted when the application for marketing authorization
after repeated use. A toxic metabolite with a long half-life is submitted. Indeed, for products for which there is a
may accumulate in the body and disturb results. The char- great medical need in the treatment of certain serious
acterization and evaluation of metabolites are long proc- diseases, the regulatory authority may agree that submis-
esses, and are generally the last studies to be completed in sion of carcinogenicity data can be delayed until after
a development programme. marketing approval is granted.

290
 Regulatory affairs Chapter | 20 |

Table 20.1  Duration of repeated-dose toxicity studies

Recommended minimum duration of repeated-dose toxicity studies

Maximum duration To support clinical trials To support marketing

of clinical trials

Rodents Non-rodents Rodents Non-rodents

Up to 2 weeks 2 weeksa 2 weeks 1 month 1 month
>2 weeks to 1 month Same as the clinical trialb Same as the clinical trial b
3 months 3 months
>1 month to 3 months 6 months 3 months
>3 months to 6 months 6 months c
9 monthscd
>6 months 6 monthsbc 9 monthsbcd
a
In the USA, as an alternative to 2-week studies, single-dose toxicity studies with extended examination can support single-dose human
trials.
b
Data from 3-month studies in rodents and non-rodents may be sufficient to start clinical trials longer than 3 months provided longer-term
data are made available before extension of the clinical trial.
c
If paediatric patients are the target population, long-term toxicity studies in juvenile animals may be required.
d
6 months’ studies are acceptable in non-rodents in certain cases, e.g. intermittent treatment of migraine, chronic treatment to prevent
recurrence of cancer, indications for which life expectancy is short, animals cannot tolerate the treatment.

Reproductive and developmental toxicity Box 20.1  Requirement for reproduction toxicity
These studies (see Chapter 15) are intended to reveal related to clinical studies in fertile women
effects on male or female fertility, embryonic and fetal
development, and peri- and postnatal development. EU: Embryo/fetal development studies are
An evaluation of effects on the male reproductive system required before Phase I, and female fertility
is performed in the repeated-dose toxicity studies, and this should be completed before Phase III
histopathological assessment is considered more sensitive USA: Careful monitoring and pregnancy testing
in detecting toxic effects than are fertility studies. Men can may allow fertile women to take part
therefore be included in Phase I–II trials before the male before reproduction toxicity is available.
fertility studies are performed in animals. Female fertility and embryo/fetal assessment
Women may enter early studies before reproductive tox- to be completed before Phase III
icity testing is completed, provided they are permanently Japan: Embryo/fetal development studies are
sterilized or menopausal, and provided repeated-dose tox- required before Phase I, and female fertility
icity tests of adequate duration have been performed, should be completed before Phase III.
including the evaluation of female reproductive organs.
For women of childbearing potential there is concern
regarding unintentional fetal exposure, and there are
regional differences (Box 20.1) in the regulations about Efficacy assessment (studies in man)
including fertile women in clinical trials. When the preclinical testing is sufficient to start studies in
man, the RA department compiles a clinical trials submis-
Local tolerance and other toxicity studies sion, which is sent to the regulatory authority and the
The purpose of local tolerance studies is to ascertain ethics committee (see Regulatory procedures, below).
whether medicinal products (both active substances and The clinical studies, described in detail in Chapter 17,
excipients) are tolerated at sites in the body that may come are classified according to Table 20.2.
into contact with the product in clinical use. This could
mean, for example, ocular, dermal or parenteral adminis-
tration. Other studies may also be needed. These might be Human pharmacology
studies on immunotoxicity, antigenicity studies on meta­ Human pharmacology studies refer to the earliest human
bolites or impurities, and so on. The drug substance and exposure in volunteers, as well as any pharmacological
the intended use will determine the relevance of other studies in patients and volunteers throughout the develop-
studies. ment of the drug.

291
Section | 3 | Drug Development

Table 20.2  ICH classification of clinical studies

Type of Study objectives Traditional

study terminology
Human Assess tolerance; describe or define pharmacokinetics/pharmacodynamics; Phase I
pharmacology explore drug metabolism and drug interactions; estimate activity
Therapeutic Explore use for the targeted indication; estimate dosage for subsequent Phase II
exploratory studies; provide basis for confirmatory study design, endpoints, methodologies
Therapeutic Demonstrate or confirm efficacy; establish safety profile; provide an adequate Phase III
confirmatory basis for assessing benefit–risk relationship to support licensing (drug (a and b)
approval); establish dose–response relationship
Therapeutic use Refine understanding of benefit–risk relationship in general or special Phase IV
populations and/or environments; identify less common adverse reactions;
refine dosing recommendations

The first study of a new drug substance in humans has factors such as concomitant drug treatment, diet, social
essentially three objectives: habits (e.g. tobacco or alcohol), age, gender, ethnic origin
• To investigate tolerability over a range of doses and, if and time of administration.
possible, see the symptoms of adverse effects Interaction studies can be performed in healthy volun-
• To obtain information on pharmacokinetics, and to teers looking at possible metabolism changes when
measure bioavailability and plasma concentration/ co-administering compounds that share the same enzy-
effect relations matic metabolic pathway. Also, changed pharmacokinetic
• To examine the pharmacodynamic activity over a range behaviour can be investigated in combinations of drugs
of doses and obtain a dose–response relationship, that are expected to be used together. Generally, such
provided a relevant effect can be measured in healthy human volunteer studies are performed when clinical
volunteers. findings require clarification or if the company wishes to
avoid standard warning texts which would otherwise
Further human pharmacology studies are performed
apply to the drug class.
to document pharmacodynamic and pharmacokinetic
effects. Examples of the data needed to support an applica-
tion for trials in patients are the complete pharmacoki- Therapeutic exploratory studies
netic evaluation and the performance of bioavailability/ After relevant information in healthy volunteers has
bioequivalence studies during the development of new been obtained, safety conclusions from combined animal
formulations or drug-delivery systems. Information is also and human exposure will be assessed internally. If these
obtained on the possible influence of food on absorption, are favourable, initial patient studies can begin. To obtain
and that of other concomitant medications, i.e. drug inter- the most reliable results, the patient population should
action. Exploration of metabolic pattern is also performed be as homogeneous as possible – similar age, no other
early in the clinical development process. diseases than the one to be studied – and the design
Special patient populations need particular attention should, when ethically justified, be placebo controlled.
because they may be unduly sensitive or resistant to treat- For ethical reasons only a limited number of closely
ment regimens acceptable to the normal adult population monitored patients take part in these studies. Their impor-
studied. One obvious category is patients with renal or tance lies in the assumption that any placebo effect in the
hepatic impairment, who may be unable to metabolize or group treated with active drug should be eliminated by
excrete the drug effectively enough to avoid accumulation. comparison with blinded inactive treatment. They are
The metabolic pattern and elimination route are impor- used primarily to establish efficacy measured against no
tant predictors for such patients, who are not included in treatment.
clinical trials until late in development.
Gender differences may also occur, and may be detected
Studies in special populations: elderly,
by the inclusion of women at the dose-finding stage of
clinical trials. children, ethnic differences
An interaction is an alteration in the pharmacodynamic Clinically significant differences in pharmacokinetics
or the pharmacokinetic properties of a drug caused by between the elderly and the young are due to several

292
 Regulatory affairs Chapter | 20 |

factors related to aging, such as impaired renal function, Ethnic differences

which can increase the variability in drug response, as well
In order for clinical data to be accepted globally, the risk
as increasing the likelihood of unwanted effects and drug
of ethnic differences must be assessed. ICH Efficacy guide-
interactions. Bearing in mind that the elderly are the
line E5 defines a bridging data package that would allow
largest group of consumers, this category should be studied
extrapolation of foreign clinical data to the population in
as early as possible in clinical trials.
the new region. A limited programme may suffice to
Clinical trials in children confirm comparable effects in different ethnic groups.
Ethnic differences may be genetic in origin, as in the
Studies in children require experience from adult human
example described in Chapter 17, or related to differences
studies and information on the pharmacokinetic profile
in environment, culture or medical practice.
of the substance. Because of the difficulties, and the often
small commercial return, companies have seldom consid-
ered it worthwhile to test drugs in children and to seek Therapeutic confirmatory studies
regulatory approval for marketing drugs for use in chil- The therapeutic confirmatory phase is intended to confirm
dren. Nevertheless, drugs are often prescribed ‘off-label’ efficacy results from controlled exploratory efficacy studies,
for children, on the basis of clinical experience suggesting but now in a more realistic clinical environment and with
that they are safe and effective. Such off-label prescribing a broader population.
is undesirable, as clinical experience is less satisfactory In order to convincingly document that the product is
than formal trials data as a guide to efficacy and safety, efficacious, there are some ‘golden rules’ to be aware of in
and because it leaves the clinician, rather than the phar- terms of the need for statistical power, replication of the
maceutical company, liable for any harm that may result. results, etc. Further, for a product intended for long-term
Recently, requirements to include a paediatric population use, its performance must usually be investigated during
early have forced the development of new guidance as to long-term exposure. All these aspects will, however, be
how to include children in clinical development. Market influenced by what alternative treatments there are, the
exclusivity prolongation has been successfully tried for intended target patient population, the rarity and severity
some years in the USA, and in July 2003 the Federal Food of the disease as well as other factors, and must be evalu-
Drug and Cosmetics Act was amended to request paediat- ated case by case.
ric studies in a new submission unless omission is justi-
fied. In September of 2007, the Paediatric Research Equity
Act replaced the 2003 Act and an assessment was made to
Clinical safety profile
evaluate the quality of the studies submitted. The results An equally important function of this largest and longest
were positive and between September 2007 and June 2010 section of the clinical documentation is to capture all
more than 250 clinical trials comprising more than adverse event information to enable evaluation of the rela-
100 000 children have been performed. (see website refer- tive benefit–risk ratio of the new compound, and also to
ence 2). detect rare adverse reactions. To document clinical safety,
In Europe the European Commission adopted the Pae- the ICH E1 guideline on products intended for chronic use
diatric Regulation in February 2007 (see website reference stipulates that a minimum of 100 patients be treated for
3). This stipulates that no marketing approval will be at least 1 year, and 300–600 treated for at least 6 months.
granted unless there is an agreed paediatric investigation In reality, however, several thousand patients usually form
plan in place or, alternatively, there is a waiver from the the database for safety evaluation for marketing approval.
requirement because of the low risk that the product will Not until several similar studies can be analysed together
be considered for use in children. The paediatric studies can a real estimate be made of the clinical safety of the
may, with the agreement of the regulatory authority, be product.
performed after approval for other populations. For new The collected clinical database should be analysed
compounds or for products with a Supplementary Protec- across a sensible selection of variables, such as sex, age,
tion Certificate (see p. 300), paediatric applications will race, exposure (dose and duration), as well as concomitant
be given a longer market exclusivity period, and off-patent diseases and concomitant pharmacotherapy This type of
products will be given a special paediatric use marketing integrated data analysis is a rational and scientific way to
authorization (PUMA); funds will be available for paedi- obtain necessary information about the benefits and risks
atric research in these products. of new compounds, and has been required for FDA sub-
To further encourage paediatric research, authority sci- missions for many years, though it is not yet a firm require-
entific advice is free. Paediatric clinical data from the EU ment in the EU or Japan.
and elsewhere are collected in a common European data- To further emphasize the accountability for the product
base to avoid repetition of trials with unnecessary expo- by the pharmaceutical company a more proactive legisla-
sure in children. The US FDA and the EMA in Europe tion for risk evaluation has been introduced in USA and
exchange information on paediatric clinical research. EU. The term Risk Evaluation and Mitigation Strategy

293
Section | 3 | Drug Development

(REMS) in the USA is matched by Risk Management stringent, but procedures and standards are flexible and
System in EU. generally established on a case-by-case basis. Conse-
The difference from the established method of reporting quently, achieving regulatory approval can often be a
adverse reactions that have occurred is that any possible greater challenge for the pharmaceutical company, but
or potential risks for a patient being harmed should be there are also opportunities to succeed with novel and
foreseen or identified early and mitigation/minimization relatively quick development programmes.
activities be planned in advance. The risk management Published guidelines on the development of conven-
document is generally part of the regulatory approval and tional drugs need to be considered to determine what parts
follows the product’s entire lifecycle. The true risk profile are relevant for a particular biopharmaceutical product. In
will develop along with the increased knowledge base. The addition, there are, to date, seven ICH guidelines dealing
goal is at any time to be able to demonstrate how and why exclusively with biopharmaceuticals, as well as numerous
the treatment benefits outweigh the risks (see website ref- FDA and CHMP guidance documents. These mostly deal
erences 4 and 5). with quality aspects and, in some cases, preclinical safety
aspects. The definition of what is included in the term
‘biopharmaceutical’ varies somewhat between documents,
Regulatory aspects of novel types
and therefore needs to be checked. The active substances
of therapy include proteins and peptides, their derivatives, and prod-
As emphasized in earlier chapters, the therapeutic scene is ucts of which they are components. Examples include (but
moving increasingly towards biological and biopharma- are not limited to) cytokines, recombinant plasma factors,
ceutical treatments, as well as various innovative, so-called growth factors, fusion proteins, enzymes, hormones and
advanced therapies. There are also broadening definitions monoclonal antibodies (see also Chapters 12 and 13).
of what constitutes medical devices. The regulatory frame-
Quality considerations
work established to ensure the quality, safety and efficacy
of conventional synthetic drugs is not entirely appropriate A unique and critically important feature for biopharma-
for many biopharmaceutical products, and even less so for ceuticals is the need to ensure and document viral safety
the many gene- and cell-based products currently in devel- aspects. Furthermore, there must be preparedness for
opment. Recombinant proteins have been in use since potentially new hazards, such as infective prions. There-
1982 and the regulatory process for such biopharmaceu- fore strict control of the origin of starting materials and
ticals is by now well established, hence this chapter will expression systems is essential. The current battery of ICH
mainly focus on these. The EU also has, for example, new quality guidance documents in this area reflects these
legislation in force since December 2008 specifically on points of particular attention (ICH Q5A-E, and Q6B).
Advanced Therapy Medicinal Products (ATMP), meaning At the time when biopharmaceuticals first appeared, the
somatic cell therapy, gene therapy and tissue engineering. ability to analyse and exactly characterize the end-product
This involves requirements for the applicant, as part of a was not possible the way it was with small molecules.
marketing authorization, to establish a risk management Therefore, their efficacy and safety depended critically on
system. In addition to the standard requirements in terms the manufacturing process itself, and emphasis was placed
of safety follow-up for an approved product, this system on ‘process control’ rather than ‘product control’.
should also include an evaluation of the product’s effec- Since then, much experience and confidence has been
tiveness (see website reference 6). In the USA, there are a gained. Bioanalytical technologies for characterizing large
number of FDA guidelines relating to cellular and gene molecules have improved dramatically and so has the field
therapies. The regulatory framework for even newer thera- of bioassays, which are normally required to be included
peutic modalities relating, for example, to nanotechnolo- in such characterizations, e.g. to determine the ‘potency’
gies is not yet clearly defined, and the regulatory authorities of a product. As a result, the quality aspects of biophar-
face a difficult task in keeping up with the rapid pace of maceuticals are no longer as fundamentally different from
technological change. those of synthetic products as they used to be. The quality
documentation will still typically be more extensive than
it is for a small-molecule product.
Biopharmaceuticals Today the concept of ‘comparability’ has been estab-
Compared with synthetic compounds, biopharmaceuti- lished and approaches for demonstrating product compa-
cals are by their nature more heterogeneous, and their rability after process changes have been outlined by
production methods are very diverse, including complex regulatory authorities. Further, the increased understand-
fermentation and recombinant techniques, as well as pro- ing of these products has in more recent times allowed the
duction via the use of transgenic animals and plants, approval of generic versions also of biopharmaceuticals,
thereby posing new challenges for quality control. This has so-called follow-on biologics or biosimilars. A legal frame-
necessarily led to a fairly pragmatic regulatory framework. work was established first in the EU, in 2004, and the first
Quality, safety and efficacy requirements have to be no less such approvals (somatropin products) were seen in 2006.

294
 Regulatory affairs Chapter | 20 |

Approvals are so far slightly behind in the USA since there Efficacy considerations
was actually not a regulatory pathway for biosimilars until
The need to establish efficacy is in principle the same for
such provisions were signed into law in March 2010 (see
biopharmaceuticals as for conventional drugs, but there
website reference 7). The FDA is, at the time of writing,
are significant differences in practice. The establishment of
working on details of how to implement these provisions
a dose–response relationship can be irrelevant, as there
to approve biosimilars. Also, in Japan, the authorities have
may be an ‘all-or-none-effect’ at extremely low levels. Also,
published ‘Guidelines for the Quality, Safety and Efficacy
to determine a maximum tolerated dose (MTD) in humans
Assurance of Follow-on Biologics’, the most recent version
may be impractical, as many biopharmaceuticals will not
being in March 2009.
evoke any dose-limiting side effects. Measuring pharma-
Safety considerations cokinetic properties may be difficult, particularly if the
substance is an endogenous mediator. Biopharmaceuticals
The expectations in terms of performing and documenting may also have very long half-lives compared to small mol-
a non-clinical safety evaluation for biotechnology-derived ecules, often in the range of weeks rather than hours.
pharmaceuticals are well outlined in the ICH guidance For any biopharmaceutical intended for chronic or
document S6. This guideline was revised in 2011 driven by repeated use, there will be extra emphasis on demonstrat-
scientific advances and experience gained since publica- ing long-term efficacy. This is because the medical use of
tion of the original guidance. It indicates a flexible, case- proteins is associated with potential immunogenicity and
by-case and science-based approach, but also points out the possible development of neutralizing antibodies, such
that a product needs to be sufficiently characterized to that the intended effect may decrease or even disappear
allow the appropriate design of a preclinical safety with time. Repeated assessment of immunogenicity may
evaluation. be needed, particularly after any process changes.
Generally, all toxicity studies must be performed accord- To date, biopharmaceuticals have typically been devel-
ing to GLP. However, for biopharmaceuticals it is recog- oped for serious diseases, and certain types of treatments,
nized that some specialized tests may not be able to such as cytotoxic agents or immunomodulators, cannot be
comply fully with GLP. The guidance further comments given to healthy volunteers. In such cases, the initial dose-
that the standard toxicity testing designs in the commonly escalation studies will have to be carried out in a patient
used species (e.g. rats and dogs) are often not relevant. population rather than in normal volunteers.
To make it relevant, a safety evaluation should include
a species in which the test material is pharmacologically Regulatory procedural considerations
active. Further, in certain justified cases one relevant
In the EU, only the centralized procedure can be used for
species may suffice, at least for the long-term studies. If no
biopharmaceuticals as well as biosimilars (see Regulatory
relevant species at all can be identified, the use of trans-
procedures, below). In the USA, biopharmaceuticals are in
genic animals expressing the human receptor or the use of
most cases approved by review of Biologics License Appli-
homologous proteins should be considered.
cations (BLA), rather than New Drug Applications (NDA).
Other factors of particular relevance with biopharmaceu-
ticals are potential immunogenicity and immunotoxicity.
Long-term studies may be difficult to perform, depending
Personalized therapies
on the possible formation of neutralizing antibodies in the It is recognized that an individual’s genetic makeup influ-
selected species. For products intended for chronic use, the ences the effect of drugs. Incorporating this principle into
duration of long-term toxicity studies must, however, therapeutic practice is still at a relatively early stage,
always be scientifically justified. Regulatory guidance also although the concept moves rapidly towards everyday
states that standard carcinogenicity studies are generally reality, and it presents a challenge for regulatory authori-
inappropriate, but that product-specific assessments of ties who are seeking ways to incorporate personalized
potential risks may still be needed, and that a variety of medicine (pharmacogenomics) into the regulatory
approaches may be necessary to accomplish this. process. How a drug product can or should be co-developed
In 2006 disastrous clinical trial events took place with with the necessary genomic biomarkers and/or assays is
an antibody (TGN 1412), where healthy volunteers suf- not yet clear. Another regulatory uncertainty has been how
fered life-threatening adverse effects. These effects had not the generation of these kinds of data will translate into
been anticipated by the company or the authority review- label language for the approved product. Already in 2005
ing the clinical trial application. The events triggered the FDA issued guidance on a specific procedure for ‘Phar-
intense discussions on risk identification and mitigation macogenomic Data Submissions’ to try and alleviate
for so-called first-in-human clinical trials, and in particular potential industry concerns and instead promote scientific
for any medicinal product which might be considered progress and the gaining of experience in the field. In the
‘high-risk’. Since then, specific guidelines for such clinical EU there is, at the time of writing, draft guidance pub-
trial applications have been issued by the regulatory lished on the use of pharmacogenetic methodologies in
authorities. the pharmacokinetic evaluation of medicinal products.

295
Section | 3 | Drug Development

However, still today, there are relatively few approved however, be excluded from this requirement under certain
products to learn from. conditions (see website reference 10, p. 301), for example
if the estimated concentration in the aquatic environment
of the active substance is below 1 part per billion, or if the
Orphan drugs
substance occurs naturally. In Europe a corresponding
Orphan medicines are those intended to diagnose, prevent general guideline was implemented in December 2006 for
or treat rare diseases, in the EU further specified as life- human medicinal products (see website reference 11).
threatening or chronically debilitating conditions. The
concept also includes therapies that are unlikely to be
developed under normal market conditions, where the REGULATORY PROCEDURES
company can show that a return on research investment
will not be possible. To qualify for orphan drug status,
there should be no satisfactory treatment available or, alter- Clinical trials
natively, the intended new treatment should be assumed It is evident that clinical trials can pose unknown risks to
to be of significant benefit (see website references 8 and 9) humans: the earlier in the development process, the
To qualify as an orphan indication, the prevalence of greater the risk. Regulatory and ethical approvals are based
the condition must be fewer than 5 in 10 000 individuals on independent evaluations, and both are required before
in the EU, fewer than 50 000 affected in Japan, or fewer investigations in humans can begin.
than 200 000 affected in the USA. Financial and scientific The ethical basis for all clinical research is the Declara-
assistance is made available for products intended for use tion of Helsinki (see website reference 12, p. 301), which
in a given indication that obtain orphan status. Examples states that the primary obligation for the treating physician
of financial benefits are a reduction in or exemption from is to care for the patient. It also says that clinical research
fees, as well as funding provided by regulatory authorities may be performed provided the goal is to improve treat-
to meet part of the development costs in some instances. ment. Furthermore, the subject must be informed about
Specialist groups within the regulatory authorities provide the potential benefits and risks, and must consent to par-
scientific help and advice on the execution of studies. ticipate. The guardians of patients who for any reason
Compromises may be needed owing to the scarcity of cannot give informed consent (e.g. small children) can
patients, although the normal requirements to demon- agree on participation.
strate safety and efficacy still apply. Regulatory authorities are concerned mainly with the
The most important benefit stimulating orphan drug scientific basis of the intended study protocol and of
development is, however, market exclusivity for 7–10 years course the safety of subjects/patients involved. Regulatory
for the product, for the designated medical use. In the EU authorities require all results of clinical trials approved by
the centralized procedure (see Regulatory procedures, them to be reported back.
below) is the compulsory procedure for orphan drugs. All clinical research in humans should be performed
The orphan drug incentives are fairly recent. In the USA according to the internationally agreed Code of Good
the legislation dates from 1983, in Japan from 1995, and Clinical Practice, as described in the ICH guideline E6 (see
in the EU from 2000. The US experience has shown very website reference 1, p. 301).
good results; a review of the first 25 years of the Orphan
Drug Act resulted in 326 marketing approvals, the majority
of which are intended to treat rare cancers or metabolic/ Europe
endocrinological disorders. Until recently, the regulatory requirements to start and
conduct clinical studies in Europe have varied widely
between countries, ranging from little or no regulation to
Environmental considerations
a requirement for complete assessment by the health
Environmental evaluation of the finished pharmaceutical authority of the intended study protocol and all support-
products is required in the USA and EU, the main concern ing documentation.
being contamination of the environment by the com- The efforts to harmonize EU procedures led to the devel-
pound or its metabolites. The environmental impact of the opment of a Clinical Trial Directive implemented in May
manufacturing process is a separate issue that is regulated 2004 (see website reference 13, p. 301). One benefit is
elsewhere. that both regulatory authorities and ethics committees
The US requirement for environmental assessment must respond within 60 days of receiving clinical trials
(EA) applies in all cases where action is needed to mini- applications and the requirements for information to be
mize environmental effects. An Environmental Assess- submitted are being defined and published in a set of
ment Report is then required, and the FDA will develop guidelines. Much of the information submitted to the
an Environmental Impact Statement to direct necessary regulatory authority and ethics committee is the same. In
action. Drug products for human or animal use can, spite of the Clinical Trial Directive, however, some national

296
 Regulatory affairs Chapter | 20 |

differences in format and information requirements have The requirements for beginning a clinical study in Japan
remained, and upon submission of identical clinical are similar to those in the USA or Europe. Scientific
trial protocols in several countries the national reviews summary information is generally acceptable, and ethics
may reach different conclusions. Therefore, at the time of committee approval is necessary.
writing, there is a pilot Voluntary Harmonization Proce-
dure (VHP) available for applications involving at least
three EU member states. This comprises two steps, where Application for marketing
in the first round one coordinator on the regulatory authorization
authority side manages national comments on the appli-
cation and provides the applicant with a consolidated list The application for marketing authorization (MAA in
of comments or questions. In the second step the usual Europe, NDA in the USA, JNDA in Japan) is compiled and
national applications for the clinical trial are submitted submitted as soon as the drug development programme
but national approval should then be gained very rapidly has been completed and judged satisfactory by the
without any further requests for change. company. Different authorities have differing require-
ments as to the level of detail and the format of submis-
sions, and it is the task of the RA department to collate all
USA the data as efficiently as possible to satisfy these varying
An Investigational New Drug (IND) application must be requirements with a minimum of redrafting.
submitted to the FDA before a new drug is given to The US FDA in general requires raw data to be submit-
humans. The application is relatively simple (see website ted, allowing them to make their own analysis, and thus
reference 14, p. 301), and to encourage early human they request the most complete data of all authorities.
studies it is no longer necessary to submit complete phar- European authorities require a condensed dossier contain-
macology study reports. The toxicology safety evaluation ing critical evaluations of the data, allowing a rapid review
can be based on draft and unaudited reports, provided the based on conclusions drawn by named scientific experts.
data are sufficient for the FDA to make a reliable assess- These may be internal or external, and are selected by the
ment. Complete study reports must be available in later applicant.
clinical development phases. Japanese authorities have traditionally focused predom-
Unless the FDA has questions, or even places the product inantly on data generated in Japan; studies performed
on ‘clinical hold’, the IND is considered opened 30 days elsewhere being supportive only.
after it was submitted and from then on information Below, the procedures adopted in these three regions are
(study protocol) is added to it for every new study planned. described in more detail.
New scientific information must be submitted whenever
changes are made, e.g. dose increases, new patient catego- Europe
ries or new indications. The IND process requires an
annual update report describing project progress and addi- Several procedures are available for marketing authoriza-
tional data obtained during the year. tion in the EU (Figure 20.3, see website reference 13):
Approval from Institutional Review Board (IRB) is • National procedure, in which the application is
needed for every institution where a clinical study is to be evaluated by one regulatory authority. This procedure
performed. is allowed for products intended for that country
only. Also, it is the first step in a mutual recognition
procedure.
Japan
• Mutual recognition, in which a marketing approval
Traditionally, because of differences in medical culture, in application is assessed by one national authority, the
treatment traditions and the possibility of significant racial Reference Member State (RMS), which subsequently
differences, clinical trials in Japan have not been useful for defends the approval and evaluation in order to
drug applications in the Western world. Also, for a product gain mutual recognition of the assessment from
to be approved for the Japanese market, repetition of clini- other European authorities. The pharmaceutical
cal studies in Japan was necessary, often delaying the avail- company may select countries of interest. These,
ability of products in Japan. the Concerned Member States, have 90 days to
With the introduction of international standards under recognize the initial assessment. The Mutual
the auspices of ICH, data from Japanese patients are Recognition procedure is used for harmonization
increasingly becoming acceptable in other countries. The and conversion of nationally approved products.
guideline on bridging studies to compensate for ethnic After mutual recognition the final marketing
differences (ICH E5; see website reference 1, p. 301) authorizations are given as national decisions, but
may allow Japanese studies to become part of global the scientific assessment is quicker and requires
development. fewer resources from all national authorities. In the

297
Section | 3 | Drug Development

After close of procedure, national approval in each CMS within 30 days

Submission in
Applicant
RMS and CMS

For info

Day 70, To arbitration

Preliminary
Assessment
report
No concensus
For response

Assessment step II Day 210

Day 100, Close of procedure
Consultation if consensus
RMS, CMS, No
Applicant Consensus

Clock stop, 3 months Until Day 120

RMS prepares Day 120,
Close of procedure
Draft Assessment Consensus To National approval
Report

RMS = Reference Member State, CMS = Concerned Member State

Fig. 20.3  One of the submission processes for marketing approval in EU, the Decentralized Procedure.

case of non-agreement, referral to EMA for for any reason is not submitted to follow the
arbitration is done as a last resort. But before centralized procedure.
arbitration is considered, the Coordination Group • Centralized procedure is a ‘federal’ procedure carried
for Mutual Recognition and Decentralised Procedures out by the EMA, with scientists selected from CHMP
(CMDh/CMDv) will try to resolve outstanding to perform the review, the approval body being the
issues. This is a group composed of expert members European Commission. This procedure is mandatory
from European regulatory authorities. The worst for biotechnological products, biosimilars, orphan
outcome of an arbitration would be that the drugs as well as products intended to treat diabetes,
marketing authorizations obtained are withdrawn in AIDS, cancer, neurodegenerative disorders,
all EU countries, including the RMS. autoimmune diseases and viral diseases.
• Decentralized procedure is similar to the Mutual • The centralized procedure starts with the nomination
Recognition Procedure. It is a modernization the aim of one CHMP member to act as rapporteur, who
of which is to share the work among authorities selects and leads the assessment team. A selected
earlier in the process, with the possibility of a co-rapporteur and team make a parallel review. The
decision being reached before 120 days have European Commission approves the application
passed from receipt of a valid submission. It is the based on a CHMP recommendation, which in turn
procedure of choice for a new chemical entity that is based on the assessment reports by the two

298
 Regulatory affairs Chapter | 20 |

rapporteur teams. Products approved in this way can is made by the MHLW based on the Evaluation Centre’s
be marketed in all EU countries with the same report.
prescribing information, packs and labels. It is worth mentioning that health authorities, in par-
CHMP is prepared to give scientific advice to companies ticular in the ICH regions, have well-established commu-
in situations where published guidance on the European nications and often assist and consult each other.
position is not available, or when the company needs to
discuss a possible deviation from guidelines. Such advice, The common technical document
as well as advice from national regulatory authorities, may
be very valuable at any stage of the development pro- Following the good progress made by ICH in creating
gramme, and may later be incorporated in new guidelines. scientific guidelines applicable in the three large regions,
Providing advice requires considerable effort from CHMP discussions on standardizing document formats began in
specialists, and fees have to be paid by the pharmaceutical 1997. The aim was to define a standard format, called the
company. Common Technical Document (CTD), for the application
for a new drug product. It was realized from the outset
that harmonization of content could not be achieved,
USA owing to the fundamental differences in data require-
The FDA is more willing than other large authorities to ments and work processes between different regulatory
take an active part in planning the drug development authorities. Adopting a common format would, nonethe-
process. Some meetings between the FDA and the spon- less, be a worthwhile step forward.
soring company are more or less compulsory. One such The guideline was adopted by the three ICH regions in
example is the so-called end-of-Phase II meeting. This is November 2000 and subsequently implemented, and it
in most cases a critically important meeting for the has generally been accepted in most other countries. This
company in which clinical data up until that point is saves much time and effort in reformatting documents for
presented to the FDA together with a proposed remaining submission to different regulatory authorities. The struc-
phase III programme and the company’s target label. The ture of the CTD (see website reference 1, p. 301) is sum-
purpose is to gain FDA feedback on the appropriateness marized in Figure 20.4.
of the intended NDA package and whether, in particular, Module 1 (not part of the CTD) contains regional infor-
the clinical data are likely to enable approval of a desirable mation such as the application form, the suggested pre-
product label. It is important that the advice from the FDA scribing information, the application fee, and also other
is followed. At the same time, these discussions may make information that is not considered relevant in all territo-
it possible in special cases to deviate from guidelines by ries, such as environmental assessment (required in the
prior agreement with the FDA. Furthermore, these discus- USA and Europe, but not in Japan). Certificates of different
sion meetings ensure that the authority is already familiar regional needs are also to be found in Module 1, as well
with the project when the dossier is submitted. as patent information not yet requested in the EU.
The review time for the FDA has decreased substantially Module 2 comprises a very brief general introduction,
in the last few years (see Chapter 22). Standard reviews followed by summary information relating to quality,
should be completed within 10 months, and priority safety (i.e. non-clinical studies) and efficacy (i.e. clinical
reviews of those products with a strong medical need studies). Quality issues (purity, manufacturing process,
within 6 months. stability, etc.) are summarized in a single document of a
The assessment result is usually communicated either as maximum 40 pages. The non-clinical and clinical sum-
an approval or as a complete response letter. The latter is maries each consist of a separate overview (maximum 30
in essence a request for additional information or data pages) and summaries of individual studies. The overviews
before approval. in each area are similar to the previous EU Expert Reports
in that they present critical evaluations of the programme
performed. Detailed guidelines (see website references 1,
Japan 13 and 14), based on existing US, European and Japanese
Also in Japan the regulatory authority is today available requirements, are available to indicate what tabulated
for consultation, allowing scientific discussion and feed- information needs to be included in these summaries, and
back during the development phase. These meetings how the written summaries should be drafted. The non-
tend to follow a similar pattern to those in the USA and clinical section has been fairly non-controversial. The
EU and have made it much easier to address potential guidance is very similar to the previous US summary, with
problems well before submission for marketing approval clear instructions on how to sort the studies regarding
is made. animals, doses, durations of treatment and routes of
These changes have meant shorter review times and a administration.
more transparent process. The review in Japan is per- The clinical summary is similar to what was required
formed by an Evaluation Centre, and the ultimate decision by the FDA, incorporating many features taken from the

299
Section | 3 | Drug Development

ADMINISTRATIVE RULES

Not part Patent protection and  

of the
CTD data exclusivity
Module 1 Supplementary protection certificate
Regional
administrative During the 1980s, the time taken to develop and obtain
information 1.0 marketing approval for a new drug increased so much that
the period of market exclusivity established by the original
patent could be too short to allow the company to recoup
CTD table of contents
its R&D costs. To overcome this problem (see Chapter 19),
2.1
the EU Council in 1992 introduced rules allowing com-
CTD introduction 2.2 panies to apply for a Supplementary Protection Certificate
(SPC), matching similar legislation in the USA and Japan.
Module 2 This can prolong market protection by a maximum of
Non-clinical Clinical
5 years to give an overall period of exclusivity of up to
overview overview
Quality 2.4 2.5 15 years.
CTD
overall The application for an SPC has to be submitted within
summary Non-clinical Clinical 6 months of first approval anywhere in Europe, within or
2.3 summaries summaries outside the EU, and from then on the clock starts. It is thus
2.6 2.7 strategically important to obtain first approval in a finan-
cially important market for best revenue.
Module 3 Module 4 Module 5
Quality Non-clinical Clinical
3.0 study reports study reports Data exclusivity
4.0 5.0
Data exclusivity should not be confused with patent pro-
tection. It is a means to protect the originators data,
Fig. 20.4  Diagrammatic representation of the organization meaning that generic products cannot be approved by
of the Common Technical Document (CTD). referring to the original product documentation until the
exclusivity period has ended. For a new chemical entity
the protection may be up to 10 years, depending on region.
The generation of paediatric data may extend it even
Integrated Summaries of Efficacy (ISE) and Safety (ISS). further, between 6 and 12 months.
ISE will generally fit in the clinical summary document.
The ISS document has proved very useful in drawing con-
clusions from the clinical studies by sensible pooling and Pricing of pharmaceutical  
integration, but is too large (often more than 400 pages
products – ‘the fourth hurdle’
in itself) to be accepted in the EU and Japan. This problem
may be resolved by including the ISS as a separate report The ‘fourth hurdle’ or ‘market access’ are expressions for
in Module 5. the ever growing demand for cost-effectiveness in phar­
Modules 3–5 comprise the individual study reports. maceutical prescribing. Payers, whether health insurance
Most reports are eligible for use in all three regions, pos- companies or publicly funded institutions, now require
sibly with the exception at present of Module 3, Quality, more stringent justification for the normally high price of
which may need regional content. a new pharmaceutical product. This means that, if not in
There is a gradual shift to fully electronic submissions the original application, studies have to demonstrate that
– known as e-CTD – in place of the large quantities of the clinical benefit of a new compound is commensurate
paper which have comprised an application for marketing with the suggested price, compared to the previous therapy
approval. As this gets fully implemented experimental of choice in the country of application.
data will be lodged mainly in databases, allowing the To address this concern, studies in a clinical trials pro-
information to be exchanged between pharmaceutical gramme from Phase II onwards now normally include
companies and regulatory authorities much more easily. health economic measures (see Chapter 18). In the USA,
Guidelines on the structure and interface between data- certainly, it is advantageous to have a statement of health
bases in the industry setting and those at the authorities economic benefit approved in the package insert, to justify
are available. reimbursement. Reference pricing systems are also being

300
 Regulatory affairs Chapter | 20 |

implemented in Europe. A favourable benefit–risk evalua- EU European Union

tion is no longer sufficient: value for money must also be FDA Food and Drug Administration; the US
demonstrated in order to have a commercially successful Regulatory Authority
product (see also Chapter 21). GCP Good clinical practice
GLP Good laboratory practice
GMP Good manufacturing practice
ICH International Conference on Harmonization
LIST OF ABBREVIATIONS IND Investigational New Drug application (US)
IRB Institutional Review Board (US), equivalent
CHMP Committee for Medicinal Products for to Ethics Committee in Europe
Human Use, the new name to replace ISS Integrated Summary of Safety (US)
CPMP early 2004 JNDA Japanese New Drug Application
CTD Common Technical Document MAA Marketing Authorization Application (EU),
CMDh/ equivalent to NDA in USA
CMDv Coordination Group for Human/Veterinary MHLW Ministry of Health, Labor and Welfare (Jpn)
Medicinal Products for Mutual Recognition MTD Maximum tolerated dose
and Decentralised Procedure (EU) NDA New Drug Application (USA)
EC Ethics Committee (EU), known as NTA Notice to Applicants; EU set of
Institutional Review Board (IRB) in USA pharmaceutical regulations, directives and
eCTD Electronic Common Technical Document guidelines
EMA European Medicines Agency SPC Supplementary Protection Certificate
ERA Environmental Risk Assessment (EU) WHO World Health Organization

REFERENCES

Cartwright AC, Matthews BR. product registration. London: Taylor Meade TW. Subacute myelo-optic
Pharmaceutical product licensing and Francis; 1994. neuropathy and clioquinol. An
requirements for Europe. London: Drews J. In quest of tomorrow’s epidemiological case-history for
Ellis Horwood; 1991. medicines. New York: Springer- diagnosis. British Journal of
Cartwright AC, Matthews BR, editors. Verlag; 1999. Preventive and Social Medicine
International pharmaceutical 1975;29:157–69.

WEBSITE REFERENCES

1. http://www.ich.org/ 6. http://www.ema.europa.eu/docs/ 10. http://www.fda.gov/downloads/

2. http://www.fda.gov/Drugs/ en_GB/document_library/ Drugs/GuidanceCompliance
DevelopmentApprovalProcess/ Regulatory_and_procedural_ RegulatoryInformation/Guidances/
DevelopmentResources/ guideline/2009/10/ ucm070561
ucm049867.htm WC500006326.pdf 11. http://www.ema.europa.eu/docs/
3. http://ec.europa.eu/health/files/ 7. http://www.fda.gov/Drugs/ en_GB/document_library/
eudralex/vol-1/reg_2006_1902/ GuidanceComplianceRegulatory Scientific_guideline/2009/10/
reg_2006_1902_en.pdf Information/ucm215031.htm WC500003978.pdf
4. http://www.fda.gov/downloads/ 8. http://www.ema.europa.eu/ema/ 12. http://www.wma.net/
Drugs/GuidanceCompliance index.jsp?curl=pages/special_topics/ en/30publications/10policies/b3/
RegulatoryInformation/Guidances/ general/general_content_000034. index.html
UCM184128.pdf jsp&murl=menus/special_topics/ 13. http://ec.europa.eu/health/
5. http://www.ema.europa.eu/docs/ special_topics.jsp&mid=WC0b01ac documents//eudralex/index_en.htm
en_GB/document_library/ 058002d4eb 14. http:/www.fda.gov
Regulatory_and_procedural_ 9. http://www.fda.gov/ForIndustry/
guideline/2009/10/ DevelopingProductsforRare
WC500004888.pdf DiseasesConditions/default.htm

301
 Chapter 21 
The role of pharmaceutical marketing
V M Lawton

management thought and practice – such as flexible man-

INTRODUCTION ufacturing systems, flat organizational structures and an
increased emphasis on service – all of which are designed
This chapter puts pharmaceutical marketing into the to make businesses more responsive to customer needs
context of the complete life-cycle of a medicine. It describes and preferences.
the old role of marketing and contrasts that with its con- A modern definition of marketing could be:
tinually evolving function in a rapidly changing environ- ‘A flexible customer focused business planning
ment and its involvement in the drug development function, including a wide variety of activities,
process. It examines the move from the product-centric
focus of the 20th century to the customer-centric focus of
aimed at achieving profitable sales’.
the 21st century.
Pharmaceutical marketing increased strongly in the
10–15 years following the Second World War, during which
HISTORY OF PHARMACEUTICAL
time thousands of new molecules entered the market ‘over-
whelming’ physicians with new scientific facts to learn in MARKETING
order to safely and appropriately prescribe these break-
throughs to their patients. There was a great dependence on There are not many accessible records of the types of
the pharmaceutical companies’ marketing departments marketing practised in the first known drugstore, which
and their professional sales representatives to give the full was opened by Arab pharmacists in Baghdad in 754,
information necessary to support the prescribing decision. (http://en.wikipedia.org/wiki/Pharmaceutical_industry-
Evidence-based marketing is based on data and research, cite_note-1) and the many more that were operating
with rigorous examination of all plans and follow-up to throughout the medieval Islamic world and eventually in
verify the success of programmes. Marketing is a general medieval Europe. By the 19th century, many of the drug
term used to describe all the various activities involved in stores in Europe and North America had developed into
transferring goods and services from producers to consum- larger pharmaceutical companies with large commercial
ers. In addition to the functions commonly associated functions. These companies produced a wide range of
with it, such as advertising and sales promotion, market- medicines and marketed them to doctors and pharmacists
ing also encompasses product development, packaging, for use with their patients.
distribution channels, pricing and many other functions. In the background, during the 19th century in the USA,
It is intended to focus all of a company’s activities upon there were the purveyors of tonics, salves and cure-alls,
discovering and satisfying customer needs. the travelling medicine shows. These salesmen specialized
Management guru Peter F. Drucker claimed that market- in selling sugared water or potions such as Hostetter’s Cel-
ing ‘is so basic it cannot be considered a separate function. ebrated Stomach Bitters (with an alcoholic content of 44%,
… It is the whole business seen from the point of view of which undoubtedly contributed to its popularity). In the
its final result, that is, from the customer’s point of view.’ late 1800s, Joseph Myers, the first ‘snake oil’ marketer, from
Marketing is the source of many important new ideas in Pugnacity, Nebraska, visited some Indians harvesting their

© 2012 Elsevier Ltd. 303

Section | 3 | Drug Development

knowledge gap. This rush of innovative medicines and

promotion activity was named the ‘therapeutic jungle’ by
Goodman and Gilman in their famous textbook (Goodman
and Gilman, 1960). Studies in the 1950s revealed that
physicians consistently rated pharmaceutical sales repre-
sentatives as the most important source in learning about
new drugs. The much valued ‘detail men’ enjoyed lengthy,
in-depth discussions with physicians. They were seen as a
valuable resource to the prescriber. This continued through-
out the following decades.
A large increase in the number of drugs available neces-
sitated appropriate education of physicians. Again, the
industry gladly assumed this responsibility. In the USA,
objections about the nature and quality of medical infor-
mation that was being communicated using marketing
Fig. 21.1  Picture of old style advertising.
tools (Podolsky and Greene, 2008) caused controversy in
Reproduced with kind permission of the Wellcome Trust.
medical journals and Congress. The Kefauver–Harris Drug
Control Act of 1962 imposed controls on the pharmaceu-
‘medicine plant’, used as a tonic against venomous stings tical industry that required that drug companies disclose
and bites. He took the plant, Echinacea purpurea, around to doctors the side effects of their products, allowed their
the country and it turned out to be a powerful antidote to products to be sold as generic drugs after having held the
rattlesnake bites. However, most of this unregulated mar- patent on them for a certain period of time, and obliged
keting was for ineffective and often dangerous ‘medicines’ them to prove on demand that their products were, in fact,
(Figure 21.1). effective and safe. Senator Kefauver also focused attention
Medical innovation accelerated in the 1950s and early on the form and content of general pharmaceutical mar-
1960s with more than 4500 new medicines arriving on keting and the postgraduate pharmaceutical education of
the market during the decade beginning in 1951. By 1961 the nation’s physicians. A call from the American Medical
around 70% of expenditure on drugs in the USA was on Association (AMA) and the likes of Kefauver led to the
these newly arrived compounds. Pharmaceutical compa- establishment of formal Continued Medical Education
nies marketed the products vigorously and competitively. (CME) programmes, to ensure physicians were kept objec-
The tools of marketing used included advertising, mailings tively apprised of new development in medicines.
and visits to physicians by increasing numbers of profes- Although the thrust of the change was to provide medical
sional sales representatives. education to physicians from the medical community, the
Back in 1954, Bill Frohlich, an advertising executive, and newly respectable CME process also attracted the interest
David Dubow, a visionary, set out to create a new kind of and funding of the pharmaceutical industry. Over time the
information company that could enable organizations to majority of CME around the world has been provided by
make informed, strategic decisions about the marketplace. the industry (Ferrer, 1975).
They called their venture Intercontinental Marketing Serv- The marketing of medicines continued to grow strongly
ices (IMS), and they introduced it at an opportune time, throughout the 1970s and 1980s. Marketing techniques,
when pharmaceutical executives had few data to consult perceived as ‘excessive’ and ‘extravagant’, came to the
when in the throes of strategic or tactical planning. By attention of the committee chaired by Senator Edward
1957, IMS had published its first European syndicated Kennedy in the early 1990s. This resulted in increased
research study, an audit of pharmaceutical sales within the regulation of industry’s marketing practices, much of it
West German market. Its utility and popularity prompted self-regulation (see Todd and Johnson, 1992); http://
IMS to expand into new geographies – Great Britain, www.efpia.org/Content/Default.asp?PageID=296flags).
France, Italy, Spain and Japan among them. Subsequent The size of pharmaceutical sales forces increased dra-
acquisitions in South Africa, Australia and New Zealand matically during the 1990s, as major pharmaceutical com-
strengthened the IMS position, and by 1969 IMS, with an panies, following the dictum that ‘more is better’,
annual revenue of $5 million, had established the gold bombarded doctors’ surgeries with its representatives.
standard in pharmaceutical market research in Europe and Seen as a competitive necessity, sales forces were increased
Asia. IMS remains the largest supplier of data on drug use to match or top the therapeutic competitor, increasing
to the pharmaceutical industry, providers, such as HMOs frequency of visits to physicians and widening coverage to
and health authorities, and payers, such as governments. all potential customers. This was also a period of great
The medical associations were unable to keep the doctors success in the discovery of many new ‘blockbuster’ prod-
adequately informed about the vast array of new drugs. It ucts that addressed many unmet clinical needs. In many
fell, by default, upon the pharmaceutical industry to fill the countries representatives were given targets to call on eight

304
 The role of pharmaceutical marketing Chapter | 21 |

to 10 doctors per day, detailing three to four products in

the same visit. In an average call on the doctor of less than
Product
10 minutes, much of the information was delivered by rote Intro Growth Maturity Decline
development
with little time for interaction and assessment of the point
of view of the customer. By the 2000s there was one rep-
resentative for every six doctors. The time available for

Volume $ units
representatives to see an individual doctor plummeted
and many practices refused to see representatives at all.
With shorter time to spend with doctors, the calls were
seen as less valuable to the physician. Information gath-
ered in a 2004 survey by Harris Interactive and IMS Health
(Nickum and Kelly, 2005) indicated that fewer than 40% 0
of responding physicians felt the pharmaceutical industry Time
was trustworthy. Often, they were inclined to mistrust pro-
Sales
motion in general. They granted reps less time and many
closed their doors completely, turning to alternative forms Profits
of promotion, such as e-detailing, peer-to-peer interaction,
and the Internet.
Something had to give. Fewer ‘blockbusters’ were hitting Fig. 21.2  General product life cycle graph.
the market, regulation was tightening its grip and the
downturn in the global economy was putting pressure on
public expenditure to cut costs. The size of the drug indus- translating various pieces of information and incorporat-
try’s US sales force had declined by 10% to about 92 000 ing them into a new product. A product undergoes several
in 2009, from a peak of 102 000 in 2005 (Fierce Pharma, changes and developments during this stage. With phar-
2009). This picture was mirrored around the world’s phar- maceutical products, this depends on efficacy and safety.
maceutical markets. ZS Associates, a sales strategy consult- Pricing strategy must be developed at this stage in time
ing firm, predicted another drop in the USA – this time of for launch. In a number of markets for pharmaceuticals,
20% – to as low as 70 000 by 2015. It is of interest, there- price approval must be achieved from payers (usually
fore, that the USA pharmaceutical giant Merck ran a pilot government agencies). Marketing plans, based on market
programme in 2008 under which regions cut sales staff by research and product strengths, for the launch and devel-
up to one-quarter and continued to deliver results similar opment of the product, are established and approved
to those in other, uncut regions. A critical cornerstone of within the organization. Research on expectations of cus-
the marketing of pharmaceuticals, the professional repre- tomers from new products should form a significant part
sentative, is now under threat, at least in its former guise. of the marketing strategy. Knowledge, gained through
market research, will also help to segment customers
according to their potential for early adoption of a product.
Production capacity is geared up to meet anticipated
PRODUCT LIFE CYCLE market demands.

It is important to see the marketing process in the context

of the complete product life cycle, from basic research to
Introduction phase
genericization at the end of the product patent. A typical The introduction phase of a product includes the product
product life cycle can be illustrated from discovery to dec­ launch phase, where the goal is to achieve maximum
line as follows (William and McCarthy, 1997) (Figure 21.2). impact in the shortest possible time, to establish the
The product’s life cycle period usually consists of five product in the market. Marketing promotional spend is
major steps or phases: highest at this phase of the life cycle. Manufacturing and
• Product development supply chains (distribution) must be firmly established
• Introduction during this phase. It is vital that the new product is available
• Growth in all outlets (e.g. pharmacies) to meet the early demand.
• Maturity
• Decline.
Growth phase
The growth phase is the time the when the market accepts
Product development phase
the new entry and when its market share grows. The matu-
The product development phase begins when a company rity of the market determines the speed and potential for
finds and develops a new product idea. This involves a new entry. A well-established market hosts competitors

305
Section | 3 | Drug Development

who will seek to undermine the new product and protect markets around the world, the length of the maturity
its own market share. If the new product is highly innova- phase is longer in some than in others, depending on the
tive relative to the market, or the market itself is relatively propensity of physicians to switch to new medicines. For
new, its growth will be more rapid. example, French doctors are generally early adopters of
Marketing becomes more targeted and reduces in new medicines, where British doctors are slow.
volume when the product is well into its growth. Other
aspects of the promotion and product offering will be Decline phase
released in stages during this period. A focus on the effi-
ciency and profitability of the product comes in the latter The decline phase usually comes with increased competi-
stage of the growth period. tion, reduced market share and loss of sales and profitabil-
ity. Companies, realizing the cost involved in defending
against a new product entry, will tend to reduce marketing
Maturity phase to occasional reminders, rather than the full-out promo-
tion required to match the challenge. With pharmaceutical
The maturity phase comes when the market no longer products, this stage is generally realized at the end of the
grows, as the customers are satisfied with the choices avail- patent life of a medicine. Those companies with a strong
able to them. At this point the market will stabilize, with R&D function will focus resources on the discovery and
little or no expansion. New product entries or product development of new products.
developments can result in displacement of market share
towards the newcomer, at the expense of an incumbent
product. This corresponds to the most profitable stage for
PHARMACEUTICAL PRODUCT LIFE
a product if, with little more than maintenance marketing,
it can achieve a profitable return. A product’s branding and CYCLE (Figure 21.3)
positioning synonymous with the quality and reliability
demanded by the customer will enable it to enjoy a longer As soon as a promising molecule is found, the company
maturity phase. The achievement of the image of ‘gold applies for patents (see Chapter 19). A patent gives the
standard’ for a product is the goal. In pharmaceutical company intellectual property rights over the invention for

Phase I – Phase III Phase IV/market expansion Patent

expiry/generic
4–5 years 9–11 years 10+ years competition

Product approval
Product market and customer knowledge

and
marketing launch

Marketing, sales
and distribution

Manufacturing
and supply chain

Drug
development

Drug discovery Price approval

File for patent HTA

Life cycle time

Fig. 21.3  Pharmaceutical life cycle graph.

306
 The role of pharmaceutical marketing Chapter | 21 |

around 20 years. This means that the company owns the Marketing will be closely involved in the identification
idea and can legally stop other companies from profiting of the gold standard, through its own research. It is
by copying it. Sales and profits will enable the manufac- common practice that certain Phase III trials will continue
turer to reinvest in R&D: while the regulatory submission is pending at the appro-
• For a pharmaceutical product the majority of patent priate regulatory agency. This allows patients to continue
time could be over before the medicine even hits the to receive possibly life-saving drugs until the drug can be
market obtained by purchase. Other reasons for performing trials
• Once it is produced it takes some time for the at this stage include attempts by the sponsor at ‘label
medicine to build up sales and so achieve market expansion’ (to show the drug works for additional types
share; in its early years the company needs to of patients/diseases beyond the original use for which the
persuade doctors to prescribe and use the medicine drug would have been approved for marketing), to obtain
• Once the medicine has become established then it additional safety data, or to support marketing claims for
enters a period of maturity; this is the period when the drug.
most profits are made Phase IV studies are done in a wider population after
• Finally as the medicine loses patent protection, copies launch, to determine if a drug or treatment is safe over
(generic products) enter the market at a lower cost; time, or to see if a treatment or medication can be used in
therefore, sales of the original medicine decline other circumstances. Phase IV clinical trials are done after
rapidly. a drug has gone through Phases I, II and III, has been
approved by the Regulator and is on the market.
The duration and trend of product life cycles vary among
Phase IV is one of the fastest-growing area of clinical
different products and therapeutic classes. In an area of
research today, at an annual growth rate of 23%. A chang-
unmet clinical need, the entry of a new efficacious product
ing regulatory environment, growing concerns about the
can result in rapid uptake. Others, entering into a relatively
safety of new medicines, and various uses for large-scale,
mature market, will experience a slower uptake and more
real-world data on marketed drugs’ safety and efficacy are
gradual growth later in the life cycle, especially if experi-
primary drivers of the growth seen in the Phase IV research
ence and further research results in newer indications for
environment. Post-marketing research is an important
use (Grabowski et al., 2002).
element of commercialization that enables companies to
expand existing markets, enter new markets, develop and
deliver messaging that directly compares their products
with the competition. Additionally, payer groups and regu-
TRADITIONAL PHARMACEUTICAL lators are both demanding more post-marketing data from
MARKETING drug companies.
Advantages of Phase IV research include the develop-
ment of data in particular patient populations, enhanced
Clinical studies
relationship with customers and development of advocacy
Pharmaceutical marketing depends on the results from among healthcare providers and patients. These studies
clinical studies. These pre-marketing clinical trials are con- help to open new communication channels with health-
ducted in three phases before the product can be submit- care providers and to create awareness among patients.
ted to the medicines regulator for approval of its licence Phase IV data can also be valuable in the preparation of a
to be used to treat appropriate patients. The marketing health economics and HTA file, to demonstrate cost
planner or product manager will follow the development effectiveness.
and results from these studies, interact with the R&D func-
tion and will develop plans accordingly. The three phases
Identifying the market
(described in detail in Chapter 17) are as follows.
Phase I trials are the first stage of testing in human At the stage when it appears that the drug is likely to be
subjects. approved for marketing, the product manager performs
Phase II trials are performed on larger groups (20–300) market research to identify the characteristics of the
and are designed to assess how well the drug works, as well market. The therapeutic areas into which the new product
as to continue Phase I safety assessments in a larger group will enter are fully assessed, using data from standard
of volunteers and patients. sources to which the company subscribes, often IMS
Phase III studies are randomized controlled multicentre (Intercontinental Marketing Services). These data are
trials on large patient groups (300–3000 or more depend- generally historic and do not indicate what could happen
ing upon the disease/medical condition studied) and are nor do they contain much interpretation. The product
aimed at being the definitive assessment of how effective manager must begin to construct a story to describe the
the drug is, in comparison with current ‘gold standard’ market, its dynamics, key elements and potential. This is
treatment. the time to generate extensive hypotheses regarding the

307
Section | 3 | Drug Development

market and its potential. Although these hypotheses can material. The Limitations of the product, where they could
be tested, they are not extensively validated. Therefore the lead to the medicine being given to a patient for whose
level of confidence in the data generated at this stage condition the medicine is not specifically indicated, must
would not normally enable the marketer to generate a be clearly indicated in all materials and discussions with
reliable plan for entry to the market, but will start to indi- the doctor. The potential risk of adverse reactions of a
cate what needs to be done to generate more qualitative medicine must also be clearly addressed in marketing
and quantitative data. interactions with medical professionals.
When the product label is further developed and tested, The analysis of the product in this way enables the
further market research, using focus groups (a form of marketer to produce a benefit ladder, beginning with the
qualitative research in which a group of people are asked about product features and attributes and moving to the various
their perceptions, opinions, beliefs and attitudes towards a new benefits from the point of view of the customer. Emotional
concept and the reasons for current behaviour) of physicians benefits for the customer in terms of how they will feel by
to examine their current prescribing practice and rationale introducing the right patient to the right treatment are at
for this behaviour. These data are extensively validated the top of the ladder. It is also important for the product
with more qualitative research. The identification of the manager to construct a benefit ladder for competitive
target audiences for the new medicines is critical by this products in order to differentiate.
stage. Armed with these data and the limited quantitative data,
Traditionally, for the most part, the pharmaceutical the product manager defines the market, identifies the
industry has viewed each physician customer in terms of target audience and their behaviour patterns in prescribing
his/her current performance (market share for a given and assesses what must be done to effect a change of
product) and potential market value. Each company uses behaviour and the prescription of the new medicine once
basically the same data inputs and metrics, so tactical launched. A key classification of a target physician is his
implementation across the industry is similar from or her position in the Adoption Pattern of innovative new
company to company. products. There are three broad classifications generally
used:

The product • Early adopters/innovators: those who normally are

quick to use a new product when it is available
Features, attributes, benefits, • Majority: most fall into this category. There are the
limitations (FABL) early majority and the late majority, defined by
their relative speed of adoption. Some are already
A product manager is assigned a product before its launch satisfied and have no need for a new product. Some
and is expected to develop a marketing plan for the product wait for others to try first, preferring to wait until
that will include the development of information about unforeseen problems arise and are resolved.
the existing market and a full understanding of the prod- Improved products (innovative therapies) may move
uct’s characteristics. This begins with Features, such as the this group to prescribe earlier
specific mechanism of action, the molecular form, the • Laggards/conservatives: this group is the slowest
tablet shape or colour. Closely aligned to this are the adopter, invariably waiting for others to gain the
product Attributes, usually concrete things that reside in early experience. They wait for something compelling
the product including how it functions. These are charac- to occur, such as additional evidence to support the
teristics by which a product can be identified and differen- new product, or a new indication for which they do
tiated. For example, these can include the speed of onset not currently have a suitable treatment.
of action, absence of common side effects, such as drowsi-
ness, or a once per day therapy rather than multiple doses.
Assessing the competition
Then come the Benefits of the new product. Benefits are
intrinsic to the customer and are usually abstract. A product It is critical for a product marketer to understand the com-
attribute expressed in terms of what the doctor or patient petition and learn how to out-manoeuvre them. Market
gets from the product rather than its physical characteris- research is the starting point. A complete analysis of the
tics or features. A once per day therapy can lessen the competitive market through product sales, investment
intrusion of medicine taking on patient’s lives. Lack of levels and resource allocation helps the marketer to
drowsiness means the patient complies with the need to develop the marketing plan for his own product. But it is
take the medicine for the duration of the illness and is able important to understand the reason for these behaviours
to function effectively during their working day. The ben- and critically evaluate their importance and success.
efits may also accrue to the physician as patient compliance A comprehensive review of competitor’s activities and
indicates that the patient is feeling well, thanks to the rationale for them is also necessary. Tools are available to
actions of the doctor. It is critically important for the phar- facilitate this in the U.S., for example, PhRMA (the indus-
maceutical marketer to maintain balance in promotional try association), publish a New Medicines Database that

308
 The role of pharmaceutical marketing Chapter | 21 |

tracks potential new medicines in various stages of devel- have been effective in a number of cases (Kaiser Family
opment. The database includes medicines currently in Foundation, 2001) with increased sales volumes. Then
clinical trials or at the FDA for evaluation (http://www. practically every major company jumped on the band-
phrma.org/newmeds/). However, it is not just a desk wagon and competed for prime-time slots and the most
bound exercise for a group of bright young marketers. attractive actors to play the role of patient. The appearance
Hypotheses, thus formed, must be tested against the cus- of these messages is practically identical and one washes
tomer perception of them. over another. Probably the only people to pay them much
The appearance of promotional aids, such as ‘detail attention, because of the volume and sameness, are the
aids’, printed glossies containing key messages, desktop pharmaceutical marketers and the advertising agencies
branded reminder items (e.g. calendars), or ‘door openers’ that are competing for the business. As for the physicians,
(pens, stick-it pads, etc.) to get past the ‘dragon on the little attention was paid to their reaction to this visual-
door’ (receptionist), are common, but do they work? In media-informed patient.
the pharmaceutical industry, for many years, marketing In assessing how successful DTC has been, early data
has been a little like the arms war. Rep numbers are from the USA (Kaiser Family Foundation, 2001) observed,
doubled or tripled because that’s what the competition is on average, a 10% increase in DTC advertising of drugs
doing. Identical looking detail pieces, clinical study papers within a therapeutic drug class resulted in a 1% increase
wrapped in a shiny folder with the key promotional mes- in sales of the drugs in that class. Applying this result to
sages and product label printed on it, appear from all the 25 largest drug classes in 2000, the study found that
competitors. Regulations known as ‘medical-legal control’ every $1 the pharmaceutical industry spent on DTC adver-
are in place, internally and at a national level, to ensure tising in that year yielded an additional $4.20 in drug
that promotional messages are accurately stated, balanced sales. DTC advertising was responsible for 12% of the
and backed by well referenced evidence, such as clinical increase in prescription drugs sales, or an additional $2.6
trial data and the approved product label. billion, in 2000. DTC advertising did not appear to affect
Are promotional activities being focused on key leverage the relative market share of individual drugs within their
points and appropriate behavioural objectives? Before drug class. In the decade to 2005 in the USA, spend on
embracing what seems to be a good idea from the com- DTC practically tripled to around $4.2 billion (Donohue
petition, the marketer needs to understand how this is et al., 2007). This level of expenditure and impact on pre-
perceived by the customer. New qualitative research needs scriptions has caused a great deal of controversy driven by
to be conducted to gain this insight. One of the principal traditional industry critics. However, some benefits have
aims of successful marketing is to differentiate one product accrued, including the increased interaction between phy-
from another. It would, therefore, seem to be counter- sicians and patients. In surveys, more than half of physi-
intuitive for an assessment of competitor behaviour to cians agree that DTC educates patients about diseases and
conclude that copying the idea of a competitor is somehow treatments. Many physicians, however, believe that DTC
going to automatically enable you to out-compete them. encourages patients to make unwarranted requests for
Much of traditional pharmaceutical marketing implemen- medication.
tation has been outsourced to agencies. These organiza- From a marketing point of view, new concepts like DTC
tions are briefed about the product, the market into which can prove helpful. Sales increase and patients become a
it will enter and the key therapeutic messages the company new and legitimate customer. The cost of entry, particu-
wishes to convey. The companies retain the preparation larly in financially difficult times can be prohibitive for all
and maintenance of the marketing plan and, in general, but a few key players. For those who are in this group, it
leave the creative, behavioural side of the preparation to becomes imperative that they stay there, as it is still a
the agency. It is not, therefore, surprising that a formulaic viable if costly segment. The naysayers are successfully
approach to marketing delivery is the norm. containing DTC to the USA and New Zealand. The clarion
A classic example of this effect came in the 1990s with calls for a moratorium and greater FDA regulation have
Direct to Consumer advertising (DTC). This form of direct still been avoided by those who market in this way. Will
marketing, in the printed media, television and radio, is the investment in this type of marketing continue to pass
legally permitted in only a few countries, such as the USA the test of cost effectiveness and revenue opportunity?
and New Zealand. Its original conception, for the first Competitive marketing demands a continuous critical
time, was to talk directly to patients about the sorts of assessment of all expenditure according to rigorous ROI
medication that they should request from their physician. (return on investment) criteria.
Anyone glued to breakfast television in the USA will be DTC advertising expenditure decreased by more than
bombarded with attractive messages about what you could 20% from 2007 to 2009. Economic pressures and the
feel like if you took a particular medicine. These messages global financial meltdown have resulted in a tighter bud­
always contain a balance about contraindications and pos- getary situation for the pharmaceutical industry. These
sible adverse reactions. Patients are encouraged to consult pressures have been strongly affected by rapidly disappear-
their physician. This concept was original and seems to ing blockbuster drugs and decreasing R&D productivity.

309
Section | 3 | Drug Development

e-Marketing decision-making process wants and benefits from these

programmes, as long as they are balanced, approved and
e-Marketing is a process, using the internet and other elec- helpful. If the pharmaceutical industry wants to continue
tronic media, of marketing a brand or concept by directly with the funding and involvement, then it should be seen
connecting to the businesses of customers. as a legitimate part of the marketing template.
In pharmaceuticals, electronic marketing has been
experimented with since the late 1990s. As a branch of
Key opinion leaders
DTC, it has been limited to only a couple of markets, the
USA and New Zealand, because local laws elsewhere do Key opinion leaders (KOL), or ‘thought leaders’, are
not permit DTC communication about medicines. Some respected individuals in a particular therapeutic area, such
companies have created websites for physician-only access as prominent medical school faculty, who influence physi-
with some success, in terms of usage, but difficult to assess cians through their professional status. Pharmaceutical
in terms of product uptake. Electronic detailing of physi- companies generally engage key opinion leaders early in
cians, that is use of digital technology: the Internet and the drug development process to provide advocacy and
video conferencing, has been used in some markets for a key marketing feedback. These individuals are identified
number of years. There are two types of e-detailing: inter- by a number of means, reputation, citations, peer review
active (virtual) and video. Some physicians find this type and even social network analysis.
of interaction convenient, but the uptake has not been Pharmaceutical companies will work with KOLs from
rapid or widespread, nor has its impact been accurately the early stage of drug development. The goal is to gain
measured in terms of utility. early input from these experts into the likely success and
While e-Marketing has promise and theoretically great acceptability of these new compounds in the future
reach, the pharmaceutical industry in general has done market. Marketing personnel and the company’s medical
little more than dabble in it. It is estimated to occupy function work closely on identifying the best KOLs for
around 1–3% of DTC budgets. Anecdotally staff recruited each stage of drug development. The KOL has a network
for their e-Marketing expertise have not integrated well which, it is hoped, will also get to know about a new
into the highly regulated media market of the pharmaceu- product and its advantages early in the launch phase. The
tical industry. KOL can perform a number of roles, in scientific develop-
ment, as a member of the product advisory board and an
advisor on specific aspects of the product positioning.
CME The marketing manager is charged with maintaining
Continuing Medical Education is a long established and coordinating the company’s relationship with the
means for medical professionals to maintain and update KOL. Special care is taken by the company of the level of
competence and learn about new and developing areas of payment given to a KOL and all payments have to be
their field, such as therapeutic advances. These activities declared to the medical association which regulates the
may take place as live events, written publications, online individual, whatever the country of origin. KOLs can be
programmes, audio, video, or other electronic media (see divided into different categories. Global, national and local
http://www.accme.org/dir_docs/doc_upload/9c795f02- categorizations are applied, depending on their level of
c470-4ba3-a491-d288be965eff_uploaddocument.pdf). influence. Relationships between the company and a par-
Content for these programmes is developed, reviewed, and ticular opinion leader can continue throughout the life
delivered by a faculty who are experts in their individual cycle of a medicine, or product franchise, for example
clinical areas. rheumatology, cardiovascular disease.
Funding of these programmes has been largely provided
by the pharmaceutical industry, often through Medical
Education and Communications Companies or MECCs. A PRICING
number of large pharmaceutical companies have with-
drawn from using these sorts of third-party agencies in the Pharmaceutical pricing is one of the most complex ele-
USA, due to the controversy concerning their objectivity. ments of marketing strategy. There is no one answer as to
The Swedish health system puts a cap of 50% on the level how members of the industry determine their pricing strat-
of contribution to costs of the pharmaceutical industry. egy. Elements to consider are whether there is freedom of
This type of activity is beneficial to the medical com- pricing at launch, such as in the USA, the UK and Germany,
munity and provides, if only through networking, an or regulated price setting as in France, Italy and Spain.
opportunity to interact with the physicians. The regulation
of content of these programmes is tight and generally
Freedom of pricing
well controlled, but a strong voice of discontent about
the involvement of the industry continues to be sounded. The criteria for price setting are many and varied,
The fact is that a key stakeholder in the medical including consideration of the market dynamics. Pricing

310
 The role of pharmaceutical marketing Chapter | 21 |

of established therapies, the ability, or otherwise to Prices with which a product goes to market rarely grow
change prices at a later stage than the product launch, the during the product life cycle in all but a few markets.
value added by the new product to the market and the However, there are frequent interventions to enforce price
perception of its affordability, are just some of the con­ reductions, usually affecting the whole industry, in regu-
siderations. A vital concern of the company is the recovery lated markets over the marketed life of a medicine.
of the costs of bringing the product to the market, esti-
mated at an average of $1.3 billion for primary care
medicines.
The marketing team, health economic analysts, financial
HEALTH TECHNOLOGY  
function and global considerations are all part of the
pricing plan. Cost-effectiveness analyses and calculations ASSESSMENT (HTA)
of benefit to the market are all factored into the equation.
When a global strategy is proposed, market research is Health technology assessment (HTA) is a multidiscipli-
conducted along with pricing sensitivity analyses. nary activity that systematically examines the safety, clini-
cal efficacy and effectiveness, cost, cost-effectiveness,
organizational implications, social consequences and
Regulated pricing legal and ethical considerations of the application of a
health technology – usually a drug, medical device or
In the majority of markets where the government are also
clinical/surgical procedure (see http://www.medicine.
the payers, there is a formal process after the medical
ox.ac.uk/bandolier/painres/download/whatis/What_is_
approval of a drug for negotiating a reimbursement price
health_tech.pdf).
with the industry. The prices for multinational companies
The development of HTAs over the 20 years prior to
are set globally and the goal is to achieve the same or
2010 has been relatively strong. Before that, in the ’70s
closely similar prices in all markets.
outcomes research assessed the effectiveness of various inter-
However, in matters of health and its perceived afford-
ventions. The Cochrane Collaboration is an international
ability each nation retains its sovereignty. Bodies such as
organization of 100 000 volunteers that aims to help
the OECD gather and compare data from each of its 33
people make well-informed decisions about health by pre-
member states. The reports it issues highlight trends in
paring, maintaining and ensuring the accessibility of sys-
pharmaceutical pricing and give each country easy access
tematic reviews of the benefits and risks of healthcare
to the different approaches each regulator uses.
interventions (http://www.cochrane.org/). INAHTA (Inter-
Different approaches abound throughout the regulated
national Network of Agencies for Health Technology
markets and include:
Assessment) is a non-profit organization that was estab-
• Therapeutic reference pricing, where medicines to lished in 1993 and has now grown to 50 member agencies
treat the same medical condition are placed in from 26 countries including North and Latin America,
groups or ‘clusters’ with a single common Europe, Asia, Australia and New Zealand. All members are
reimbursed price. Underpinning this economic non-profit-making organizations producing HTAs and are
measure is an implicit assumption that the linked to regional or national government. Three main
products included in the cluster have an forces have driven the recent developments of HTA: con-
equivalent effect on a typical patient with this cerns about the adoption of unproven technologies, rising
disease costs and an inexorable rise in consumer expectations. The
• Price-volume agreements, where discounts could be HTA approach has been to encourage the provision of
negotiated on volume commitments research information on the cost effectiveness of health
• Pay for performance models, where rebates could be technologies, including medicines.
achieved if the efficacy is not as expected NICE, the National Institute for Health and Clinical
• Capitation models where the expenditure is capped Excellence was established in the UK in 1999, primarily to
at a certain level. help the NHS to eliminate inequality in the delivery of
There are a myriad of different approaches across the certain treatments, including the more modern and effec-
nations. The goal of these systems is generally cost con- tive drugs, based upon where a patient lived. This became
tainment at the government level. From a marketing point known as the ‘postcode lottery’. Over time, NICE also
of view, the Global Brand Manager must be familiar with developed a strong reputation internationally for the
and sensitive to the variety of pricing variables. At the local development of clinical guidelines. In terms of the entry
level, a product manager needs to be aware of the concerns of new medicines and the assessment of existing treat-
of the payer and plan discussions with key stakeholders in ments, NICE’s activities and influence attracted the atten-
advance of product launch. Increasingly, the onus will be tion of the global pharmaceutical industry. In the UK,
on the marketer to prove the value of the medicine to the NICE drew up and announced lists of medicines and other
market. technologies it planned to evaluate. A number of health

311
Section | 3 | Drug Development

authorities used the impending reviews to caution practi- Manufacturing and distribution
tioners to wait for the results of the evaluations before
using the treatments. This became known, particularly in Product must be made for shipment, often through whole-
the pharmaceutical industry, as ‘NICE blight’. The phe- salers, to retail outlets. Sufficient stock must be available
nomenon was magnified in importance by the length of to match the anticipated demand. Samples for distribu-
time taken to conduct reviews and the time between tion to physicians by reps will be made ready for launch.
announcement that a review was to take place and it
taking place. Resource allocation
A NICE review includes the formation of a team of
experts from around the country, each of whom contrib- This is planned to ensure the product has the largest share
utes to the analysis. Data for these reviews are gathered of voice compared to the competition when launched. That
from a number of sources including the involved pharma- is in terms of numbers and types of sales force (specialist
ceutical company. These reports involve a number of func- and GP) according to the target audiences, and medical
tions within a company, usually led by a health economics media coverage with advertising. In addition representa-
expert. The marketing and medical functions are also tive materials, including detail aids, leave behind bro-
involved and, in a global company, head office personnel chures and samples.
will join the team before the final submission, as the
results could have a material impact on the successful Launch meeting
uptake of a medicine throughout the world. The impor-
tance of the launch period of a medicine, the first 6 Sales representatives will have been trained in the thera-
months, has been well documented by IMS (IMS, 2008b). peutic area, but the launch meeting will normally be their
Delays in the HTA review, having been announced, inevi- first sight of the approved product label. The positioning
tably affect uptake in certain areas. of the product and key messages will be revealed with the
HTA bodies abound throughout the world. Their influ- first promotional tools they will be given. The overall
ence is growing and they are increasingly relied upon by strategy will be revealed and training will take place. The
governments and health services. The methodology used training of reps includes practice in using promotional
by these agencies is not consistent. Different governments materials, such as detail aids, in order to communicate the
use HTAs in different ways, including for politically moti- key messages. Sometimes practising physicians are brought
vated cost-containment ends. For the pharmaceutical into the meeting to rehearse the first presentations (details)
company, the uncertainty caused by the addition of these with the reps. Each rep must attain a high level of under-
assessments requires a new approach to the cost effective- standing, technical knowledge and balanced delivery
ness and value calculations to ensure an expeditious before he or she is allowed to visit doctors.
entry into a market. New expertise has been incorporated
into many organizations with the employment of health Media launch
economists. Marketing managers and their teams have
started to understand the importance of HTA in the life Marketing will work with public affairs to develop a media
cycle of a new medicine. The added value and cost effec- release campaign. What can be communicated is strictly
tiveness of medicines are now a critical part of the market- limited in most markets by local regulation and limita-
ing mix. It is clear that, in one form or another, HTAs are tions in direct to consumer (DTC) rules. It is a matter of
here to stay. public interest to know that a new medicine is available,
even though the information given about the new medi-
cine is rather limited.

NEW PRODUCT LAUNCH

Target audience
The marketing plan is finalized and approved, the strategy The launch preparations completed, it is time for the
and positioning are established, the price is set, the target product to enter the market. The bright, enthusiastic sales
audiences are finalized and segmented according to the force will go optimistically forward. In general, the matu-
key drivers in their decision-making process. It is now time rity of the market and level of satisfaction with currently
to focus on the product launch. available therapy will determine the level of interest of the
target audiences.
It is important to cover the highest potential doctors in
Registration
the early stages with the most resource. The innovators and
When the medicine has been medically approved, there early adopters prescribe first and more than the others. The
are formalities, different in each market concerning the proportion varies by therapy area and by country, making
registration and pricing approval. around 5% and 10% of total prescriptions respectively.

312
 The role of pharmaceutical marketing Chapter | 21 |

2.0
Benefit of targeting
1.8 practice with innovators/
early adopters ~50%
1.6
increase in larger
1.4 practices
Prescription per practice

1.2

1.0

0.8

0.6 With innovators/

early adopters
0.4
Without innovators/
0.2 early adopters

0
Solo practice 2–4 GPs 5–7 GPs 8–10 GPs

Fig. 21.4  Innovators graph.

Reproduced with kind permission of IMS Health Consulting.

In a study from IMS covering launches from 2000 to patients. In most treatment areas there are a cohort of
2005, an analysis of the effect of innovators, early adopters patients who are not responding well to existing therapy,
and early majority revealed these groups as being responsi- or who suffer side effects from the treatment. The role of
ble for more than half of all prescriptions in the first 6 marketing in helping doctors to identify pools of existing
months from launch. After the first 6 months the late patients is important, even before launch of the new
majority and conservative prescribers start to use the product.
product. From this period on, the number of prescriptions Early penetration of the market by targeting the right
begins to resemble that of the physicians in each of the type of physicians and the right type of patients is predic-
groups (IMS, 2009). tive of continued success in the post-launch period,
Another advantage of early focus on innovators is their according to the IMS launch excellence study (IMS, 2009).
influence on others in their group practice. In this case, A key predictor of success is the size of the dynamic market
prescribing per practice is higher than for equivalent-sized (new prescriptions, switches, plus any add on prescrip-
practices that lack an innovator prescriber (Figure 21.4). tions) that a launch captures.
An excellent launch requires:
The influence of innovators and early • That the right stakeholders be addressed in the right
adopters on group prescribing sequence
• Winning leading market share
Patients driving launch success • Outperformance of the competition in prescriber-
Three types of patients make up the early uptake of focused promotion.
medicines:
• New patients, who have not received previous drug The first 6 months
therapy (e.g. Viagra; Gardasil) IMS studies of launch excellence (IMS, 2008b; IMS, 2009)
• Switch patients, who have received another drug illustrate clearly that in a global product launch, the first
therapy 6 months are critical. Those that achieve launch excellence,
• Add-on patients, who have the new therapy added to according to the criteria above, continue to do well through
their current treatment. the following stage of the life cycle. Those that do not
The success with which these types of patient are cap- achieve launch excellence rarely recover. Fewer than 20%
tured will determine the success of the launch and subse- of launches significantly improve their uptake trajectory
quent market share growth. An excellent launch tends to between 6 and 18 months on the market after an unsuc-
come as a result of the level of success in achieving switch cessful early launch period.

313
Section | 3 | Drug Development

Decline in prescriber decision-making power

A significant trend which represents a major challenge to
CHANGING ENVIRONMENT –
traditional pharmaceutical marketing is the observed CHANGING MARKETING
ongoing decline in the impact and predictability of the
relationship between prescriber-focused promotional Traditional pharmaceutical marketing has served the phar-
investment and launch market share achievement, in the maceutical industry well over the past 60 years, but less so
early years of launch. as we move on. The development of the market over time
has the majority of the industry doing the same things and
effectively cancelling each other out. The importance of
Implementing the market plan differentiation from the competition is still high, whether
Once the product has been successfully launched and is the product or market is weak or strong. How can it be
gaining market share, the role of marketing is to continue achieved? (See Box 21.1.)
to implement the marketing plan, adjusting strategy as
necessary to respond to the competitive challenge.
The process is of necessity dynamic. Results and changes
in the market have to be continuously tracked and, as
necessary, decisions should be re-evaluated. At the outset, Powerful
the product manager develops a market map, according to
the original market definition. Experience in the market
will often necessitate the redrawing of parts of the map,
according to the opportunity and barriers to entry. Prescribers
Payers
A customer portrait developed pre-launch will also need Health
to be re-evaluated with experience in the market. The technology Patient
changing environment is identifying new customers and assessors
stakeholders, who are traditionally not engaged in the
marketing process, including non-prescribers, e.g. payers
and patients (Figure 21.5)
Life cycle management of the product is no longer an Distribution
automatic process that gradually sorts itself out. It needs channels
Weak
to be managed with the same enthusiasm and systematic
rigor as a new product launch. The contribution to country Stakeholders
growth from launches of products in the previous 3 years
indicates a continued decline across the top eight pharma- Fig. 21.5  Power shift graph.
ceutical markets (Figure 21.6). Reproduced with kind permission of IMS Health Consulting.

9
Germany
8
Spain
Sales growth contribution (%)

7
6 Japan

5 UK

4 USA
3 Canada
2 Italy
1 France
0
2003 2004 2005 2006 2007 2008 2009

Fig. 21.6  Country launch contribution graph.

Reproduced with kind permission of IMS Health Consulting.

314
 The role of pharmaceutical marketing Chapter | 21 |

an entity performs to create value for its customers and

Box 21.1  Competitive differentiation thus for the entity itself.’ The pharmaceutical value propo-
sition ‘starts with the raising of capital to fund R&D, con-
Product status/opportunities
cluding with the marketing and resulting sale of the
Weak or disappearing: Strong or emerging:
• Promotional spending • Ethical promotion and
products. In essence, it is about making innovative medi-
and sales force size regulatory conformity cines that command a premium price’.
The payer’s value system begins with raising revenue,
• Product innovation and • Utilisation of medical
differentiation technology
through taxation or patient contribution. The value created
for patients and other stakeholders is by managing admin-
• Managed care and • Internet
istration and provision of healthcare. For the payer, the
contracting relationships • Marketing sophistication
goal is to have a cost-effective health system and to
• DTC
enhance its reputation with its customers, or voters in the
case of governments. It aims to minimize its costs and
keep people well. At the time of writing, the UK health
The philosophy of the 1990s and early 2000s was that department is proposing a system of ‘value based pricing’
‘more is better’. More reps, more physician visits, more for pharmaceuticals, although no one seems to be able to
samples, more congresses and so on. All the while the old define what it means.
customer, the physician, was closing doors on sales reps, The provider, in this case the physician, wants to
and losing trust in the industry’s motives and integrity. provide high quality of care, economically and for as long
Additionally, the physician as provider is losing power as as necessary. This stakeholder values an assessment of the
the principal source of therapeutic decision making. health of a given population and to know what measures
Within the traditional pharmaceutical marketplace, over can be taken to manage and prevent illness.
70% of marketing spend is focused on physicians. Does The convergence of these three stakeholders’ value
this allocation any longer make sense? Companies are systems, says the PWC report, is as follows. ‘The value
structured and resourced in order to serve this stakeholder, healthcare payers generate depends on the policies and
not the other customers who are increasingly vital to the practices of the providers used’; whereas providers gener-
success or otherwise of new medicines. ate value based on the revenues from payers and drugs that
pharmaceutical companies provide. As for the pharmaceu-
The key stakeholder tical company the value it provides is dependent on access
to the patients who depend on the payers and physicians.
Even more important, a new customer base is appearing, In this scenario each of the partners is by definition
one of whose goals is to contain costs and reduce the rapid dependent on the others. An antagonistic relationship
uptake of new medicines without any ‘perceived’ innova- between these parties is, therefore, counterproductive and
tive benefit. At the time of introduction, this payer may be can result in each of the stakeholders failing to achieve
told of the ‘benefit’ claimed by the manufacturer. If the their goals. A telling conclusion for the pharmaceutical
communicator of this message is an official of a health industry is that they must work with payers and providers,
department, it is unlikely that they will be able, or moti- to determine what sort of treatments the market wants
vated, to convince the payer of the claimed benefit. As the to buy.
payers are rapidly taking over as the key stakeholder, it is
imperative that the marketer talks directly to them. The
pharmaceutical industry has viewed this stakeholder as
Innovation
something of a hindrance in the relationship between Another conundrum for the pharmaceutical industry is the
them and their ‘real’ customer, the doctor. It is likely that acceptance from the stakeholders of their definition of
the feeling may be mutual. Industry has generally failed innovation. Where a patient, previously uncontrolled on
to understand the perspective and values of the payer, one statin, for example, may be well controlled by a later
willing to accept an adversarial relationship during pricing entry from the same class, the payer may not accept this
and reimbursement discussions. as innovative. They are unwilling to accept the fact that
some patients may prefer the newer medicine if it means
paying a premium to obtain it. With the aid of HTA
Values of the stakeholders
(health technology assessment), the payer may use statis­
Decisions made on whether a new medicine adds value tical data to ‘disprove’ the cost effectiveness of the new
and is worth the investment depend on what the customer statin and block its reimbursement. The pharmaceutical
values. The value systems of the key stakeholders in the company will find it difficult to recoup the cost of R&D
pharmaceutical medicine process need to be considered, associated with bringing the drug to the market. The PWC
to see where they converge or otherwise. A value system report (2009) addresses this problem by suggesting that
has been defined by PWC (2009) as ‘the series of activities the industry should start to talk to the key stakeholders,

315
Section | 3 | Drug Development

payers, providers and patients during Phase II of the devel- complete care packages surrounding their medical discov-
opment process. It is at this stage, they suggest, that pricing eries, such as diagnostic tests and delivery and monitoring
plans should start to be tested with stakeholders, rather devices.
than waiting until well into Phase III/launch. It is possible Marketers will need to learn new specialties and prepare
to do this, but the real test of whether a medicine is a real for more complex interactions with different providers.
advance comes in the larger patient pool after launch. Traditional sales forces will not be the model for this type
Post-marketing follow-up of patients and Phase IV research of marketing. Specialist medicines are around 30% more
can help to give all stakeholders reassurance about the expensive to bring to market. They treat smaller popula-
value added by a new medicine. It was in the 1990s, after tions and will, therefore, demand high prices to recoup
the launch of simvastatin, that the 6-year-long 4S (Scan- the development cost. The debate with payers, already
dinavian Simvastatin Survival Study Group, 1994) study somewhat stressed, will become much more important.
showed that treatment with this statin significantly reduced The skills of a marketer and seller will of necessity stretch
the risk of morbidity and mortality in at-risk groups, thus far beyond those needed for primary care medicines. The
fundamentally changing an important aspect of medical structure and scope of the marketing and sales functions
practice. At Phase II, it would have been difficult to predict will have to be tailored to the special characteristics of
this. With today’s HTA intervention, it is difficult to see if these therapies.
the product would have been recommended at all for use
in treatment.

R&D present and future THE FUTURE OF MARKETING

It has been said that the primary care managed ‘block-
buster’ has had its day. Undoubtedly R&D productivity is If the industry and the key stakeholders are to work more
going through a slow period. Fewer breakthrough treat- effectively together, an appreciation of each other’s value
ments are appearing (IMS, 2008a) as, in Germany ‘only systems and goals is imperative. The industry needs to win
seven truly innovative medicines were launched in 2007’, back the respect of its customers for its pivotal role in the
from 27 product launches. global healthcare of the future.
More targeted treatment solutions are being sought for The excess of over-enthusiastic marketing and sales per-
people with specific disease subtypes. This will require sonnel in the ‘more is better’ days of pharmaceutical pro-
reinvention at every step of the R&D value chain. Therapies motion has been well documented and continues to grab
developed using advances in molecular sciences, genomics the headlines. Regulation and sanctions have improved
and proteomics could be the medicines of the future. the situation significantly, but there is still lingering suspi-
Big pharmaceutical companies are changing research cion between the industry and its customers.
focus to include biologicals and protein-based compounds The most effective way to change this is greater trans­
(http://www.phrma.org/files/Biotech%202006.pdf). Part- parency and earlier dialogue throughout the drug dis­
nerships between big pharma and biotech companies are covery and development process. Marketing personnel
forming to extend their reach and to use exciting technolo- are a vital interface with the customer groups and it is
gies such as allosteric modulation to reach targets which important that they are developed further in the new tech-
have been elusive through traditional research routes. The nologies and value systems upon which the customer
possible demise of the traditional discovery approach may relies. Although they will not be the group who conduct
be exaggerated, but it highlights the need to rethink and the HTA reports from the company, they must be fully
revise R&D. cognizant of the process and be able to discuss it in depth
with customers.
Products of the future Pharmaceutical marketing must advance beyond sim-
plistic messages and valueless giveaways to have the aim
Generic versions of previous ‘blockbuster’ molecules, sup- of adding value in all interactions with customers. Mass
ported by strong clinical evidence, are appearing all the marketing will give way to more focused interactions to
time, giving payers and providers choice without jeopard- communicate clearly the value of the medicines in specific
izing patient care. This is clearly the future of the majority patient groups.
of primary care treatment. But if a sensible dialogue and
partnership is the goal for all stakeholders in the process,
generics should not only provide care, but should also
The new way of marketing
allow headroom for innovation.
It would seem reasonable that a shift in emphasis from Marketing must focus on the benefits of new medicines
primary to secondary care research will happen. It would from the point of view of each of the customers. The past
also seem reasonable that the pharmaceutical industry will has been largely dominated by a focus on product
need to think beyond just drugs and start to consider attributes and an insistence on their intrinsic value. At one

316
 The role of pharmaceutical marketing Chapter | 21 |

point, a breakthrough medicine almost sold itself. All that needed from their point of view. Price and value added
was necessary was a well-documented list of attributes and must be topics of conversation in an open and transparent
features. That is no longer the case. manner, to ensure full understanding.
It seems as if the provider physician is not going to be The future of marketing is in flux. As Peter Drucker said,
the most important decision maker in the process anymore. ‘Marketing is the whole business seen from the customer’s
Specialist groups of key account managers should be point of view. Marketing and innovation produce results;
developed to discuss the benefits and advantages of the all the rest are costs. Marketing is the distinguishing,
new medicine with health authorities. unique function of the business.’ (Tales from the market-
Payers should be involved much earlier in the research ing wars, 2006). This is a great responsibility and an excit-
and development process, possibly even before research ing opportunity for the pharmaceutical industry. It is in
begins, to get the customer view of what medicines are the interests of all stakeholders that it succeeds.

REFERENCES

Donohue JM, Cevasco M, Rosenthal IMS. Intelligence. 360: global doi:300/7/831 [pii]10.1001/
MB. A decade of direct-to-consumer pharmaceutical perspectives 2007, jama.300.7.831.
advertising of prescription drugs. March 2008. Norwalk, Connecticut: Price Waterhouse Cooper. Pharma 2020,
New England Journal of Medicine IMS Health, 2008a. which path will you take? Available
2007;357:673–81. IMS. Launch excellence 2008, study of online at: http://www.pwc.com/gx/
Ferrer JM. How are the costs of 3000 launches across 8 markets from en/pharma-life-sciences/pharma-
continuing medical education to be 2000–7. Norwalk, Connecticut: IMS 2020/pharma-2020-vision-
defrayed? Bulletins of the New York Health, 2008b. path.jhtml; 2009.
Academy of Medicine IMS. Achieving global launch excellence, Scandinavian Simvastatin Survival Study
1975;51:785–8. January 2009. Norwalk, Connecticut: Group. Randomized trial of
Fierce Pharma. Sales force cuts have IMS Health, 2009. cholesterol lowering in 4444 patients
only just begun – FiercePharma. Kaiser Family Foundation. with coronary heart disease: the
Available online at: http:// Understanding the effects of Scandinavian Simvastatin Survival
www.fiercepharma.com/story/ direct-to-consumer prescription drug Study (4S). Lancet 1994;344:1383–9.
sales-force-cuts-have-only-just- advertising, November 2001. PMID 7968073.
begun/2009-01-20#ixzz15Rcr0Vn0; Available online at: www.kff.org/ Todd JS, Johnson KH. American
2009. content/2001/3197/DTC%20Ad%20 Medical Association Council on
Goodman LS, Gilman A. The Survey.pdf; 2001. Ethical and Judicial Affairs.
pharmacological basis of Nickum C, Kelly T. Missing the Annotated guidelines on gifts to
therapeutics: a textbook of mark(et). PharmExecutive.com, physicians from industry. Journal of
pharmacology, toxicology, and Sep 1, 2005; 2005. Oklahoma State Medical Association
therapeutics for physicians and Pharmaceutical Research and 1992;5:227–31.
medical students. 2nd ed. New York: Manufacturers of America (PhRMA). Trout, Jack. Tales from the marketing
Macmillan; 1960. New Medicines Database and wars: Peter Drucker on marketing
Grabowski H, Vernon J, DiMasi J. Medicines in Development: (http://www.forbes.com/2006/06/30/
Returns on research and Biotechnology (http://www.phrma. jack-trout-on-marketing-
development for 1990s new drug org/newmeds/). cxjt0703drucker.html).
introductions, pharmacoeconomics, Podolsky SH, Greene JA. A historical William D, McCarthy JE. Product life
2002. Adapted by IBM Consulting perspective of pharmaceutical cycle: ‘essentials of marketing’.
Services. Available online at: http:// promotion and physician education. Richard D Irwin Company; 1997.
www.aei.org/docLib/20040625_ JAMA 2008;300:831–3,
Helms.pdf; 2002.

317
 Chapter 22 
Drug discovery and development:
facts and figures
H P Rang, R G Hill

In this chapter we present summary information about the situation. For these reasons, it is very difficult to obtain
costs, timelines and success rates of the drug discovery and anything more than crude overall measures of the effec­
development operations of major pharmaceutical com­ tiveness of drug discovery research. There are, for example,
panies. The information comes from published sources, few published figures to show what proportion of drug
particularly the websites of the Association for the British discovery projects succeed in identifying a compound fit
Pharmaceutical Industry (www.abpi.org.uk), the Euro­ to enter development, whether this probability differs
pean Federation of Pharmaceutical Industries and Associa­ between different therapeutic areas, and how it relates to
tions (www.efpia.eu) and the Pharmaceutical Research the resources allocated. We attempt to predict some trends
and Manufacturers of America (www.phrma.org). Much of and new approaches at the end of this chapter.
this information comes, of course, directly or indirectly As will be seen from the analysis that follows, the most
from the pharmaceutical companies themselves, who are striking aspects of drug discovery and development are
under no obligation to release more than is legally that (a) failure is much more common than success, (b)
required. In general, details of the development process it costs a lot, and (c) it takes a very long time (Figure 22.1).
are quite well documented, because the regulatory author­ By comparison with other research-based industries, phar­
ities must be notified of projects in clinical development. maceutical companies are playing out a kind of slow-
Discovery research is less well covered, partly because com­ motion and very expensive arcade game, with the odds
panies are unwilling to divulge detailed information, but heavily stacked against them, but offering particularly rich
also because the discovery phase is much harder to codify rewards.
and quantify. Development projects focus on a specific
compound, and it is fairly straightforward to define the
different components, and to measure the cost of carrying
out the various studies and support activities that are SPENDING
needed so that the compound can be registered and
launched. In the discovery phase, it is often impossible to Worldwide, the R&D spending of the American-based
link particular activities and costs to specific compounds; pharmaceutical companies has soared, from roughly $5
instead, the focus is often on a therapeutic target, such as billion in 1982 to $40 billion in 1998 and $70 billion in
diabetes, Parkinson’s disease or lung cancer, or on a both 2008 and 2009. Figure 22.2 shows total R&D expend­
molecular target, such as a particular receptor or enzyme, iture for US-based companies over the period 1995–2009.
where even the therapeutic indication may not yet be The R&D expenditure of the 10 largest pharmaceutical
determined. The point at which a formal drug discovery companies in 2009 averaged nearly $5 billion per company
project is recognized and ‘managed’ in the sense of having over the range of Roche at $8.7 billion to Lilly at $4.13
a specific goal defined and resources assigned to it, varies billion (Carroll, 2010). Marketing and administration
greatly between companies. A further complication is costs for these companies are about three times as great as
that, as described in Section 2 of this book, the scientific R&D costs. The annual increase in R&D spending in the
strategies applied to drug discovery are changing rapidly, industry in most cases has exceeded sales growth over the
so that historic data may not properly represent the current last decade, reflecting the need to increase productivity

© 2012 Elsevier Ltd. 321

Section | 4 | Facts and Figures

It takes ~10–15 years and $802 million to develop one new medicine1

Post-marketing

FDA review 1 2

Phase III
n = 1000–5000
Phase II

Years
Clinical trials 5 compounds 6
n = 100–500
Phase I
n = 20–100

Preclinical 250 compounds 1.5

Drug discovery 5000–10 000 compounds 5

1 DiMasi JA, Hansen RW, Grabowski HG Journal of Health Economics 2003, 22, 151–185

Fig. 22.1  The attrition of compounds through the discovery and development pipeline (PhRMA).
Reproduced, with permission, from DiMasi et al., 2003.

70 65.3
63.2 63.7 Entire pharma
sector
60 56.1
51.8
47.6 47.9 47.4
50 PhRMA member
Expenditures (billions $)

45.8
43.4
39.9
companies’ R&D
40 37.0 expenditures
34.5
29.8 31.0
30 26.0 Total NIH budget
29.3 30.6
22.7 28.5 28.5 29.0
21.0 27.1 27.9
19.0
20 15.2 16.9 23.3
20.5
17.8
15.6
10 12.7 13.7
11.3 11.9

0
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 (est.)

Fig. 22.2  Private vs public spending on R&D in the USA (PhRMA).

(see Kola and Landis, 2004). Today a minimum of 15– money invested in pharmaceutical R&D and some com­
20% of sales revenues are reinvested in R&D, a higher panies (notably Pfizer with a 25% reduction over the next
percentage than in any other industry (Figure 22.3). The 2 years) have announced plans to make further cuts in
tide may have turned, however, and in 2010 the overall their budgets. The overall cost of R&D covers discovery
global R&D spend for the industry fell to $68 billion (a research as well as the various development functions,
3% reduction from 2009) (Hirschler, 2011). This fall described in Section 3 of this book. The average distribu­
reflects a growing disillusion with the poor returns on tion of R&D expenditure on these different functions is

322
 Drug discovery and development: facts and figures Chapter | 22 |

R&D expenditures per employee, by manufacturing sub-sector and industry, 2000–2007 ($)

Biopharmaceuticals* 105428
Communications equipment* 62995
Semiconductors 40341
Computers and electronics 37980
Chemicals 34978
Navigational measuring equipment* 22 262
Aerospace products* 21162
Motor vehicles, trailers, parts* 15 704
Transportation equipment 15 693
Petroleum, coal 13 319
All manufacturing 9956
Electrical equipment, appliances 6411
Machinery 5663
Paper, printing 2238
* Asterisks indicate manufacturing subsectors

Fig. 22.3  The pharmaceutical industry spends more on R&D than any other industry (PhRMA).

Table 22.1  R&D expenditure by function (2008)

1.4
Billions $ (constant dollars, year 2000)

Function Total R&D costs (%) $1.3b

1.2
US
1.0
Discovery and preclinical 27
development 0.8
$800m
Phase I & II 21.1 0.6
Clinical development 53.6
Phase III 32.5
0.4
Approval 4.7 $300m
0.2
Phase IV / 14.4
$100m
pharmacovigilance 0.0
1979 1991 2000 2005
Data from PhRMA Membership Survey 2010 (calculated from
2008 data).
Fig. 22.4  The costs (corrected for inflation) of drug
development continue to increase (PhRMA, 2011).

shown in Table 22.1. These proportions represent the

overall costs of these different functions and do not neces­ industry, and is dominated by four therapeutic areas,
sarily reflect the costs of developing an individual com­ namely:
pound. As will be discussed later, the substantial drop-out • Central nervous system disorders (including
rate of compounds proceeding through development Alzheimer’s disease, schizophrenia, depression,
means that the overall costs cover many failures as well as epilepsy and Parkinson’s disease)
the few successes. Overall cost of developing a compound • Cancer, endocrine and metabolic diseases (including
to launch and how much this has increased between 1979 osteoporosis, diabetes and obesity)
and 2005 is shown in Figure 22.4. • Cardiovascular diseases (including atherosclerosis,
Numbers of compounds in clinical development accord­ coronary disease and heart failure)
ing to therapeutic classes is shown in Figure 22.5, based on • Infectious diseases (including bacterial, viral and
a survey of the American research-based pharmaceutical other infections such as malaria).

323
Section | 4 | Facts and Figures

$897 million in 20031 (DiMasi et al., 2003; Figure 22.1),

Number of medicines compared with $31.8 million (inflation-adjusted) in 1987,
Condition in development and concluded that the R&D cost of a new drug is increas­
ing at an annual rate of 7.4% above inflation.
Alzheimer’s and other dementias 98 Based on different assumptions, even larger figures –
$1.1 billion per successful drug launch over the period
Arthritis 74
1995–2000, and $1.7 billion for 2000–2002 – were calcu­
Cancer 878 lated by Gilbert et al. (2003).
• Breast cancer 125 The fact that more than 70% of R&D costs represents
• Colorectal cancer 82 discovery and ‘failure’ costs, and less than 30% represents
• Lung cancer 120 direct costs (which, being determined largely by require­
• Leukemia 119 ments imposed by regulatory authorities, are difficult to
• Skin cancer 86 reduce), has caused research-based pharmaceutical com­
panies in recent years to focus on improving (a) the effi­
Cardiovascular disorders 237 ciency of the discovery process, and (b) the success rate of
the compounds entering development (Kola and Landis,
Diabetes mellitus 193
2004). A good recent example of expensive failure of a
HIV/AIDS 81 compound in late-stage clinical development is illustrated
by the experience of Pfizer with their cholesterol ester
Mental and behavioral disorders 252 transfer protein (CETP) inhibitor, torcetrapib (Mullard,
2011). This new approach to treating cardiovascular disease
Parkinson’s disease 25 by raising levels of the ‘good’ HDL cholesterol had looked
very promising in the laboratory and in early clinical
Respiratory disorders 334
testing, but in late 2006 development was halted when a
Rare diseases 303 large Phase III clinical trial revealed that patients treated
with the drug had an increased risk of death and heart
problems. At the point of termination of their studies
Fig. 22.5  Numbers of compounds in clinical development for Pfizer had already spent $800 million on the project. The
different therapeutic targets (PhRMA, 2011). impact extended beyond Pfizer, as both Roche (dal­
cetrapib) and Merck (anacetrapib) had their own com­
pounds acting on this mechanism in clinical development.
It was necessary to decide if this was a class effect of the
mechanism or whether it was just a compound-specific
The trend in the last decade has been away from cardio­ off-target effect limited to torcetrapib. Intensive compari­
vascular disease towards other priority areas, particularly sons of these agents in collaboration with academic groups
cancer and metabolic disorders, and also a marked increase and a published full analysis of the torcetrapib clinical
in research on biopharmaceuticals, whose products are trial data allowed the conclusion that this was most likely
used in all therapeutic areas. There is also a marked to be an effect limited to torcetrapib, possibly mediated
increase in the number of drugs being introduced for the by aldosterone, and it was not seen with the other two
treatment of rare or orphan diseases driven by enabling CETP blockers (Mullard, 2011). Roche and Merck were
legislation in the USA and EU (see Chapter 20). able to restart their clinical trials, but using very large
numbers of patients and focusing on clinical outcomes to
How much does it cost to develop   establish efficacy and safety. The Pfizer experience proba­
bly added 4 years to the time needed to complete develop­
a drug?
ment of the Merck compound anacetrapib and made this
Estimating the cost of developing a drug is not as straight­ process more expensive (for example the pivotal study
forward as it might seem, as it depends very much on what that started in 2011 will recruit 30 000 patients) with no
is taken into account in the calculation. Factors such as guarantee of commercial success. In addition the patent
‘opportunity costs’ (the loss of income that theoretically term available to recover development costs and to make
results from spending money on drug development rather any profit is now 4 years shorter (see Chapter 19). It is
than investing it somewhere else) and tax credits (contri­ not unreasonable to estimate that collectively the pharma­
butions from the public purse to encourage drug develop­ ceutical industry will have spent more than $3 billion
ment in certain areas) make a large difference, and are the
source of much controversy. The Tufts Centre for Drug 1
Quote from a commentary on this analysis (Frank RG (2003) Journal
Development Studies estimated the average cost of devel­ of Health Economics 22: 325): ‘These are impressively large numbers,
oping a drug in 2000 to be $802 million, increasing to usually associated with a purchase of jet fighters – 40 F16s in fact’.

324
 Drug discovery and development: facts and figures Chapter | 22 |

investigating CETP as a mechanism before we have either its revenues fall below the average cost. Second, the devel­
a marketed product or the mechanism is abandoned. opment money is spent several years before any revenues
come in, and sales predictions made at the time develop­
ment decisions have to be taken are far from reliable. At
any stage, the decision whether or not to proceed is based
SALES REVENUES on a calculation of the drug’s ‘net present value’ (NPV) –
an amortised estimate of the future sales revenue, minus
Despite the stagnation in compound registrations over the the future development and marketing costs. If the NPV is
past 25 years, the overall sales of pharmaceuticals have positive, and sufficiently large to justify the allocation of
risen steadily. In the period 1999–2004 the top 14 com­ development capacity, the project will generally go ahead,
panies showed an average sales growth of 10% per year, even if the money already spent cannot be fully recouped,
whereas in 2004–2009 growth was still impressive but at as terminating it would mean that none of the costs would
a lower rate of 6.7% per year. However, the boom years be recouped. At the beginning of a project NPV estimates
may be behind us and the prediction is that for 2009–2014 are extremely unreliable – little more than guesses – so
the growth will be a modest 1.2% per year (Goodman, most companies will not pay much attention to them for
2009). Loss of exclusivity as patents expire is the most decision-making purposes until the project is close to
important reason for reduced sales growth (Goodman, launch, when sales revenues become more predictable.
2009). The total global sales in 2002 reached $400.6 Furthermore, unprofitable drugs may make a real contri­
billion and in 2008 had risen to $808 billion (EFPIA, bution to healthcare, and companies may choose to
2010). In Europe the cost of prescribed drugs accounts for develop them for that reason.
some 17% of overall healthcare costs (EFPIA, 2010). Even though only 34% of registered drugs in the
Grabowski et al. (2002) study made a profit, the profits on
those that did so more than compensated for the losses
Profitability
on the others, leaving the industry as a whole with a large
For a drug to make a profit, sales revenue must exceed overall profit during the review period in the early 1990s.
R&D, manufacture and marketing costs. Grabowski et al. The situation is no longer quite so favourable for the
(2002) found that only 34% of new drugs introduced industry, partly because of price control measures in
between 1990 and 1994 brought in revenues that exceeded healthcare, and partly because of rising R&D costs. In the
the average R&D cost (Figure 22.6). This quite surprising future there will be more emphasis on the relative efficacy
result, which at first sight might suggest that the industry of drugs and it will only be those new agents that have
is extremely bad at doing its sums, needs to be interpreted demonstrable superiority over generic competitors that
with caution for various reasons. First, development costs will be able to command high prices (Eichler et al., 2010).
vary widely, depending on the nature of the compound, It is likely that the pharmaceutical industry will adapt to
the route of administration and the target indication, and changed circumstances, although whether historical profit
so a drug may recoup its development cost even though margins can be maintained is uncertain.

2000 1880
After-tax present value of sales
(millions of 2000 dollars)

1500

1000
701 After-tax average R&D costs
500 434
299
162
87 39 21 (–1)
6
0
1 2 3 4 5 6 7 8 9 10
New medicines introduced between 1990 and 1994, grouped by tenths, by lifetime sales

Fig. 22.6  Few drugs are commercially successful (PhRMA, 2011).

325
Section | 4 | Facts and Figures

Pattern of sales market accounts for only 10%, with 3% in emerging

markets (EFPIA, 2010). The approaching patent cliff (see
The distribution of sales (2008 figures) by major pharma­ later) suggests that with the current business model there
ceutical companies according to different therapeutic cat­ may be a 5–10% reduction in sales and a 20–30% reduction
egories is shown in Table 22.2. Recent changes reflect a in net income over 2012–2015 for the 13 largest companies
substantial increase in sales of anticancer drugs and drugs (Munos, 2009). Not all predictions are pessimistic and IMS
used to treat psychiatric disorders. The relatively small have recently opined that global drug sales will top $1 tril­
markets for drugs used to treat bacterial and parasitic infec­ lion in 2014 (Berkrot, 2010). Whilst accepting that growth
tions contrasts sharply with the worldwide prevalence and in USA and other developed markets is slowing, they
severity of these diseases, a problem that is currently being believe that this will be more than compensated for by
addressed at government level in several countries. The USA growth in developing markets (China is growing at 22–
represents the largest and fastest-growing market for phar­ 25% per year and Brazil at 12–15% per year).
maceuticals (61% of global sales in 2010). Together with
Europe (22%) and Japan (5%), these regions account for
88% of global sales. The rest of the established world Blockbuster drugs
The analysis by Grabowski et al. (2002) showed that 70%
of the industry’s profits came from 20% of the drugs mar­
Table 22.2  Global sales by therapeutic area keted, highlighting the commercial importance of finding
(Maggon, 2009) ‘blockbuster’ drugs – defined as those achieving annual
sales of $1 billion or more – as these are the ones that
Therapeutic area $US billions actually generate significant profits. A recent analysis of
pharmaceutical company pipelines showed that 18 new
Arthritis 35
blockbusters were registered in the period 2001–2006,
Infectious disease 24 roughly four per year globally among the 30 or so new
compounds that are currently being registered each year.
Cancer 70
In 2000, a total of 44 marketed drugs achieved sales
Cardiovascular disease 105 exceeding $1 billion, compared with 17 in 1995 and 35 in
1999. The contribution of blockbuster drugs to the overall
Central nervous system 118
global sales also increased from 18% in 1997 to 45% in
Diabetes 21 2001, reflecting the fact that sales growth in the blockbuster
Respiratory disease 25 sector has exceeded that in the market overall. The top 10
best selling drugs in 2010 are shown in Table 22.3 and this

Table 22.3  Global sales of top ten drugs (from IMS, 2011)

Drug Trade Type Indication Sales 1998 Sales 2010

name ($bn) ($bn)
Atorvastatin Lipitor HMG CoA reductase nhibitor Cholesterol lowering 1.9 12.7
Clopidogrel Plavix Adenosine receptor Anti-platelet 8.8
antagonist anticoagulant
Fluticasone/ Seretide Steroid/β adrenoceptor Asthma 8.5
salmeterol agonist
Esomeprazole Nexium Proton pump inhibitor Gastric ulcer 8.3
Quetiapine Seroquel Atypical antipsychotic drug Schizophrenia 6.8
Cerivastatin Crestor HMG CoA reductase inhibitor Cholesterol lowering 6.7
Etanercept Enbrel TNF R2 fusion protein Arthritis/psoriasis 6.2
Infliximab Remicade Anti-TNF mab Arthritis/psoriasis 6.0
Adulimumab Humira Anti-TNF mab Arthritis/psoriasis 5.9
Olanzapine Zyprexa Atypical antipsychotic Schizophrenia 2.37 5.7

326
 Drug discovery and development: facts and figures Chapter | 22 |

Table 22.4  Top 10 drugs worldwide (blockbusters) 2000

Drug Trade name Class Launch Sales 2000 Change since

year (US$bn) 1999 (%)
Omeprazole Prilosec Proton pump inhibitor 89 6.26 5.9
Simvastatin Zocor HMGCR inhibitor 91 5.21 17.5
Atorvastatin Lipitor HMGCR inhibitor 96 5.03 32.6
Amlodipine Norvasc Vasodilator 92 3.36 12.4
Loratadine Claritin Non-sedating antihistamine 93 3.01 12.6
Lansoprazole Prevacid Proton pump inhibitor 95 2.82 27.0
Epoetin-α Procrit Erythropoietic factor 90 2.71 29.4
Celecoxib Celebrex Cyclooxygenase 2 inhibitor 98 2.61 77.7
Fluoxetine Prozac SSRI 87 2.5 –0.6
Olanzapine Zyprexa Atypical antipsychotic 96 2.37 26.6

should be compared with the top 10 best selling drugs in competition-free sales window, will cost the company
2000 shown in Table 22.4. Only atorvastatin and olanza­ roughly $8 million.
pine appear in both tables as many of the other drugs have Despite increasing expenditure on R&D costs and
become generics in the interim and have, therefore, lost decreased output, the mean development time from first
much of their sales revenue due to competition. It is also synthesis or isolation (i.e. excluding discovery research
noteworthy that there was only one biological in the top preceding synthesis of the development compound) to
10 in 2000 whereas in 2010 there were three. Munos (2009) first launch was over 14 years in 1999, having increased
analysed the peak sales achieved by 329 recently intro­ somewhat over the preceding decade. Half of this time was
duced drugs and calculated that the probability of a new taken up by discovery and preclinical development and
product achieving blockbuster status was only 21%, even half by clinical studies. A further 2 years was required for
though companies take a drug into clinical development FDA review and approval. The long FDA review times
only if they believe it has blockbuster potential. during the 1980s have since come down substantially
The spate of pharmaceutical company mergers in the (Reichert, 2003), mainly because user fees and fast-track
last decade has also been driven partly by the need for procedures for certain types of drug were introduced.
companies to remain in blockbuster territory despite the Current estimates of time taken for R&D leading to a new
low rate of new drug introductions and the difficulty of marketed product are in the region of 10 years with a
predicting sales. By merging, companies are able to further 2 or 3 years need for preparation and submission
increase the number of compounds registered and thus of the regulatory package and its approval (EFPIA, 2010).
reduce the risk that the pipeline will contain no blockbust­ There are, of course, wide variations in development
ers. This may have a negative effect on R&D performance times between individual projects, although historically
and creativity, however (Kneller, 2010; LaMattina, 2011) there has been little consistent difference between differ­
ent therapeutic areas (with the exception of anti-infective
drugs, for which development times are somewhat shorter,
and anticancer drugs, for which they have been longer by
TIMELINES about 2 years). During the 1990s, most biopharmaceuti­
cals were recombinant versions of human hormones, and
One important factor that determines profitability is the their clinical development was generally quicker than that
time taken to develop and launch a new drug, in particular of small molecules. More recently, biopharmaceuticals
the time between patent approval and launch, which will have become more diverse and many monoclonal anti­
determine the length of time during which competitors bodies have been developed, and these have generally
are barred from introducing cheap generic copies of the encountered more problems in development because their
drug. A drug that is moderately successful by today’s therapeutic and unwanted effects are unpredictable, and
standards might achieve sales of about $400 million/year, so clinical development times for biopharmaceuticals
so each week’s delay in development, by reducing the have tended to increase.

327
Section | 4 | Facts and Figures

Table 22.5  Discovery timelines

Target Screening Lead Total Preclinical Total to

identification optimization discovery development clinic
and validation
McKinsey (1997)
10–36 1–3 7–14 18–53 8–16 26–69
2000 predicted 6–17 1 5–9 12–27 7–15 19–42
Andersen (1997)
10–24 6–24 24–36 40–84 6–30 46–114
2000 predicted 5–8 1–12 12–18 18–38 5–8 23–46

Inderal (1965) Lopressor (1978)

Tagamet (1977) Zantac (1983)
Capoten (1980) Vasotec (1985)
Seldane (1985) Hismanal (1989)
AZT (1987) Videx (1991)
Mevacor (1987) Pravachol (1991)
Prozac (1988) Zoloft (1992)
Diflucan (1990) Sporanox (1992)
Recombinate (I992) Kogenate (1992)
Invirase (1995) Norvir (1996)
Celebrex (1999) Vioxx (1999)
0 2 4 6 8 10 12 14
Interval between launches (years)

Fig. 22.7  Period of exclusivity is shrinking (PhRMA, 2011).

Information about the time taken for discovery – from time a premium can be charged to recover the R&D
the start of a project to the identification of a development expenditure. For example, cimetidine (Tagamet), an ulcer
compound – is sparse in the public literature. Manage­ drug introduced in 1977, had an exclusivity period of 6
ment consultants McKinsey and Arthur Andersen estimate years before another drug in the same class, ranitidine
that in 1995, the time taken from the start of the discovery (Zantac), was introduced. In contrast, celecoxib (Cele­
project to the start of clinical studies was extremely vari­ brex), the first selective cyclooxygenase-2 inhibitor (COX-
able, ranging from 21 to 103 months (Table 22.5). Both 2), which had a significant advantage over established
studies predicted a reduction in discovery time by the non-steroidal anti-inflammatory drugs, was on the market
year 2000 to 46 months or less, owing to improved dis­ only 3 months before a second, very similar, drug,
covery technologies. Although solid data are lacking, few rofecoxib (Vioxx), was approved (Figure 22.7). (Vioxx was
believe that such a dramatic improvement has actually withdrawn in 2004, because it was found to increase the
occurred. risk of heart attacks: other COX-2 inhibitors were subse­
Intensifying competition in the pharmaceutical market­ quently withdrawn or given restricted labels.)
place also is demonstrated by the shrinking period of Loss of patent protection opens up the competition to
exclusivity during which the first drug in a therapeutic generic products (the same compound being manufac­
class is the sole drug in that class, thereby reducing the tured and sold at a much lower price by companies that

328
 Drug discovery and development: facts and figures Chapter | 22 |

Table 22.6  Selected drugs facing patent expiry in the USA 2010–2012 (from Harrison, 2011)

Branded drug (INN drug Indication Worldwide 2009 Expected

name; company) sales (billion)* patent expiry
Aricept (donepezil; Eisai/Pfizer) Alzheimer’s-type dementia ¥303.8 (US$3.61) Nov 2010
Lipitor (atorvastatin; Pfizer) High cholesterol US$11.43 2011
Zyprexa (olanzapine; Eli Lilly & Schizophrenia, bipolar 1 US$4.92 2011
Company) disorder
Lexapro (escitalopram; Forest Depression and anxiety DKK 7.77 (US$1.37) 2012
Laboratories/Lundbeck)
Actos (pioglitazone; Takeda) Type 2 diabetes ¥334.5 (US$3.98)° 2012
Plavix (clopidogrel; Sanofi-Aventis/ Clot-related cardiovascular US$6.15 2012
Bristol-Myers Aquibb) events
Lovenox (enoxaparin; Sanofi-Aventis) Acute deep vein thrombosis C3.04($4.03) 2012

Seroquel (quetiapine; AstraZeneca) Schizophrenia, bipolar disorder, US$4.87 2012

major depressive disorder
*Data from company annual reports. ° Europe and the Americas. INN, international nonproprietary name.

have not invested in the R&D underlying the original dis­ deficit’. According to these calculations, to sustain a
covery), and the sales revenue from the branded com­ revenue growth rate of 10% – considered to be a healthy
pound generally falls sharply. Over the period 2010–2014 level – the 10 largest pharmaceutical companies each
drugs generating combined revenues of $78 billion needed to launch on average 3.1 new compounds each
(Harrison, 2011) will lose patent protection (Table 22.6). year, compared to 1.8 launches per company actually
Half of the revenue erosion is predicted to be in 2011 as achieved in 2000 – a rate insufficient to maintain even
Lipitor (atorvastatin) loses patent protection. This drug zero growth. As mergers increase the size of companies
(see Table 22.3) earned >$12 billion for Pfizer in 2010 and and their annual revenues these numbers increase propor­
accounts for more than 20% of their total revenues. The tionally (e.g. in 2003 Pfizer had revenues of $45 billion
loss of patent protection and the introduction of low- and to sustain this level of income it would need to gener­
priced generic atorvastatin will not only affect Pfizer, as ate nine NCEs per year (Kola and Landis, 2004)). A recent
other drugs with longer patent life in the statin class will survey indicated that the top 10 pharmaceutical compa­
now face significant competition on price as well as on nies are producing 1.17 NCEs per year if based in the USA
efficacy. These global figures vary from year to year as and 0.83 NCEs per year if based in Europe – a considerable
individual high-revenue drugs drop out of patent protec­ shortfall (Pammolli et al., 2011). Goodman (2009) con­
tion, and the revenue swings for an individual company cluded that most pharmaceutical companies are not
are of course much more extreme than the global average, replacing the products they lose from patent expiry suffi­
so maintaining a steady pipeline of new products and ciently quickly and can only remain competitive by rigor­
timing their introduction to compensate for losses as ous cost-cutting initiatives. The large numbers of layoffs
products become open to generic competition is a key part in the industry in the last 2 years would support his
of a company’s commercial strategy. conclusion.
The number of clinical trials being carried out in each
therapeutic area (Figure 22.5) provides a measure of
potential future launches. The figures shown may some­
PIPELINES AND ATTRITION RATES what overestimate the numbers, because official notifica­
tion of the start of clinical projects is obligatory, but
As we have already seen, the number of new chemical projects may be terminated or put on hold without formal
entities registered as pharmaceuticals each year has notification. Estimates of the number of active preclinical
declined over the last decade or at best has stayed flat, and projects are even more unreliable, as companies are under
various analyses (Drews, 2003a,b; Kola and Landis, 2004; no obligation to reveal this information, and the defini­
Munos, 2009) have pointed to a serious ‘innovation tion of what constitutes a project is variable. The number

329
Section | 4 | Facts and Figures

of clinical trials appears large in relation to the number of 2000 the pharmacokinetic issues had largely been
new compounds registered in each year, partly because addressed. As discussed in Chapter 10, determined efforts
each trial usually lasts for more than 1 year, and partly have been made to control for pharmacokinetic properties
because many trials (i.e. different indications, different in the discovery phase, and this appears to have reduced
dosage forms) are generally performed with each com­ the failure rate during development. Accurate prediction
pound, including previously registered compounds as well of therapeutic efficacy in the discovery phase remains a
as new ones. problem, however, particularly in disease areas, such as
The fact that in any year there are more Phase II than psychiatric disorders, where animal models are unsatisfac­
Phase I clinical projects in progress reflects the longer tory. The analysis in Figure 22.8A shows that large numbers
duration of Phase II studies, which more than offsets the of drugs with novel mechanisms of action are failing in
effect of attrition during Phase I. The number of Phase III areas of high medical need such as cancer and neurode­
projects is smaller, despite their longer duration, because generation. In Figure 22.8B it can be seen that 66% of
of the attrition between Phase II and Phase III. failures are due to lack of clinical efficacy regardless of
High attrition rates – which would horrify managers in therapeutic target and 21% due to inadequate safety
most technology-based industries – are a fact of life in (Arrowsmith, 2011). It has been estimated that in a typical
pharmaceuticals, and are the main reason why drug dis­ research project 20–30% of the time is spent fine-tuning
covery and development is so expensive and why drug molecules to fit the available animal models of disease
prices are so high in relation to manufacturing costs. The perfectly even though in many cases these models are not
cumulative attrition based on projects in the mid-1990s, predictive of clinical efficacy (Bennani, 2011). It is espe­
predicts an overall success rate of 20% from the start of cially concerning that drugs are still failing due to lack of
clinical development (Phase I). A more recent analysis efficacy in Phase III clinical trials (see Figure 22.9) and
quoted by the FDA (FDA Report, 2004) suggest a success that, although increasing numbers of compounds are
rate of compounds entering Phase I trials of only 8%, and reaching Phase I and II clinical evaluation, the number of
this can be even lower for particular therapeutic areas (e.g. active Phase III compounds has not increased between
5% for oncology (Kola and Landis, 2004)). 1997 and 2007. It has been suggested that the remedy is
The main reasons currently for failure of compounds are to make Phase II trials more rigorous on the grounds that
summarized in Figure 22.8 (Arrowsmith, 2011). It can be Phase II failures are less disruptive and less costly than
seen that the situation has changed little since Kola and Phase III failures (Arrowsmith, 2011). Even having reached
Landis (2004) showed that unsatisfactory pharmacoki­ the registration phase, 23% of compounds subsequently
netic properties and lack of therapeutic efficacy in patients fail to become marketed products (Kola and Landis,
were the commonest shortcomings in 1991 but that by 2004). The problem of lack of predictability of long-term

Anticancer
Financial and/or
(n=23) Safety (including
commercial
risk-benefit)
Not disclosed
28% 7%
21%
6%
Nervous
system
18%
Other (n=15)
19%
(n=16)

8% 13%
13% 66%
Cardiovascular Alimentary
(n=7) and/or
metabolism Efficacy
Anti-infectives
(n=11) • Versus placebo: 32%
(n=11)
• As add-on therapy: 29%
A B • Versus active control: 5%

Fig. 22.8  Reasons for compound failure in development.

Reproduced, with permission, from Arrowsmith, 2011.

330
 Drug discovery and development: facts and figures Chapter | 22 |

Number of worldwide active R&D projects in development 1650

1500
Phase II
1350

1200

1050 Phase I
900

750

600

450 Phase III

300
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Fig. 22.9  Numbers of compounds in Phase III trials is not increasing year on year (Parexel Handbook data).

toxicity such that compounds need to be withdrawn after money and experience to undertake development projects.
launch is still largely intractable. This has resulted in increased in-licensing activities,
whereby a large company licenses in and develops the
substance, for which it pays fees, milestone payments and,
eventually, royalties to the biotechnology company.
BIOTECHNOLOGY-DERIVED
MEDICINES

An increasing share of research and development projects RECENT INTRODUCTIONS

is devoted to the investigation of therapies using
biotechnology-derived molecules (see Chapters 12 and An analysis of the 58 new substances approved by the FDA
13). The first such product was recombinant human in the period 2001–2002 showed that about half of the
insulin (Humulin), introduced by Lilly in 1982. From new synthetic compounds were directed at receptor targets
1991 to 2003, 79 out of 469 (17%) new molecules regis­ (mostly GPCRs, and some steroid receptors), all of which
tered were biopharmaceuticals. Recently (2002–2008) the had been identified pharmacologically many years earlier,
proportion was around 30% – partly reflecting reduced as had the transporter and enzyme targets. Only a minor­
numbers of small molecules registered in recent years – ity of the new compounds were ‘first-in-class’ drugs. In
and this is predicted to increase to about 50% over the 2006 only about 30% of the drugs approved by the FDA
next few years. In 2002 an estimated 371 biopharmaceu­ were new chemical compounds the rest being modified
ticals were in clinical development (PhRMA Survey, 2002), formulations or new uses for existing drugs (Austin, 2006).
with strong emphasis on new cancer therapies (178 prepa­ The 20% of compounds directed at infectious agents in
rations) and infectious diseases, including AIDS (68 prep­ 2002 were also mainly ‘follow-up’ compounds, directed at
arations). One growth area is monoclonal antibodies familiar targets, and so it is clear that during the 1990s,
(MAbs). The first agent in this class, adulimumab, was when these drugs were being discovered and developed,
approved by the FDA in 2002, but between then and 2010 the industry was operating quite conservatively (DiMasi
an additional six were registered with three under review. and Faden, 2011). Although, as described in Section 2,
In 2010 there were seven in Phase III and 81 in Phase I or many new technologies are being applied in the hope of
II clinical trials (Nelson et al., 2010). Overall biologicals improving the speed and efficiency of drug discovery, drug
have had a higher probability of success than small mol­ targets and therapeutic approaches had not really changed.
ecules with 24% of agents entering Phase I trials becoming You are more likely to produce blockbusters, it was argued,
marketed products (Kola and Landis, 2004). by following the routes that produced blockbusters in the
Many biotechnology-based projects originate not in past than by trying new approaches with the aim of being
mainstream pharmaceutical companies, but in specialized ‘first-in-class’ (DiMasi and Faden, 2011). In a survey of 259
small biotechnology companies, which generally lack the drugs registered by the FDA between 1999 and 2008

331
Section | 4 | Facts and Figures

(Swinney and Anthony, 2011) it was found that only 75 yet paid off. We are learning more about disease processes
of the 259 were first-in-class agents with novel mecha­ day by day, but much of this knowledge is difficult to
nisms of action. Of the small molecule drugs in the survey translate into drug discovery. More and more senior figures
target-based screening had produced 17 with 28 being in the pharmaceutical industry are standing up in public
discovered by using phenotypic disease model screening. and saying we need to reinvent the industry (Bennani,
Thus, even though the industry is becoming more centred 2011; Paul et al., 2010). It seems inevitable (as mentioned
on molecular target-based screening, this is not the source above) that if we cannot close the gap between drugs
of the majority of new drugs. One worrying trend is the losing patent protection and new product introductions
reduction in effort across the industry in the search for then the industry will shrink. We have seen for the first
new drugs to treat psychiatric and neurological disorders, time in living memory a reduction in the R&D spending
as, although this is the most difficult research area, the of the industry driven by the realization that spending
medical need for drugs has never been higher (Abbott, more on R&D year on year did not work (Hirschler, 2011).
2010; Kaitin and Milne, 2011). The driver now is to reduce costs where possible, get the
Biopharmaceuticals registered between 1999 and 2008 investment right and only invest where probability of
include several protein mediators produced by recom­ success is high (Kola and Landis, 2004; Paul et al., 2010).
binant methods and a number of monoclonal antibodies. This is coupled with a need to streamline development
Vaccines are also strongly represented. Of the 75 first-in- (FDA, 2011) and kill drugs that are not clearly an improve­
class drugs registered by the FDA in this period, 50 (67%) ment over what we already have as early as is feasible (Paul
were small molecules and 25 (33%) were biologicals. et al., 2010; Figure 22.10). Some companies are seeing the
Judging from the portfolio of new biotechnology products need to empower creative talent and admit that drug dis­
currently undergoing clinical evaluation, the trend towards covery is as much an art as a science (Bennani, 2011;
biopharmaceuticals is set to continue, with cancer and Douglas et al., 2010; Paul et al., 2010). The crucial reor­
inflammation/autoimmune diseases as the main therapeu­ ganization may already have happened with much of the
tic targets. A milestone was reached in 2012 when a biologi­ creative part of drug discovery moving over to the biotech
cal became the world’s top selling drug as a consequence of sector or to academic centres for drug discovery (Kotz,
patent expiry for the previous top sellers (Hirschler, 2012). 2011; Stephens et al., 2011). This in turn is leading to more
Some new drug introductions are aimed at small patient partnerships between different pharmaceutical compa­
populations sometimes coupled with a biomarker to iden­ nies, and between pharmaceutical companies, biotech and
tify sensitive groups of patients who will benefit, and, as academia. It is interesting to note that the academic con­
a consequence, these drugs are hyper-expensive (Hunter tribution to drug discovery may have been underestimated
and Wilson, 2011). One recent example is a new treatment in the past (Stephens et al., 2011).
for multiple sclerosis from Novartis where a course of The economics of our business will have shifted geo­
treatment will cost $48 000 in the USA (Von Schaper and graphically by 2015 (IMS, 2011) with the US share of the
Kresge, 2011). It is not clear whether such treatments will global market declining from 41% in 2010 to 31% in 2015.
be affordable (Hunter and Wilson, 2011) or whether they At that time the US and EU combined are likely to reflect
will be beneficial to the profitability of the industry in the only 44% of global spending on medicines. This in turn
long term. is driving where investment in R&D and manufacturing
facilities is being made. It is already evident that cuts in
R&D spending in the UK and USA are paralled by an
increase in spending in Asia, especially in China (see Zang,
PREDICTING THE FUTURE? 2011). The other important shift that is happening is from
proprietary medicines to generics and in 2015 it is likely
Many models have been proposed as the ideal way to that a major growth area will be generics which will
discover drugs, but there is now general cynicism and the account for 39% of the total drug spend compared with
intrinsic cyclical nature of drug discovery is likely to have 27% in 2010 (IMS, 2011). We are thus victims of our own
been the real reason for swings from failure to success success as the blockbusters of 2005 become the generics
rather than the approach taken by the industry. Predicting of 2015. Around 50% of the new drugs that will drive
the future, especially the facts and figures is therefore a profitability in 2015 will be biologicals and the majority
dangerous game! A number of factors have become appar­ of these will be monoclonal anti­bodies (IMS, 2011).
ent in the recent past that will change our world in ways Our world will look very different 5 years from now
which we probably could not have guessed at 5 years ago. with an increasingly complex social, legal, scientific and
The belief that drug discovery could be industrialized is political environment. The pharmaceutical industry will
clearly a mistaken one. Access to larger numbers of molec­ continue to be important and governments will have to
ular targets, larger numbers of drug like compounds and ensure that a fair reward for innovation (see Aronson
faster screening technology has not led to more registered et al., 2012) is still achievable but how this will be done
drugs. Similarly our knowledge of human genome has not is at present unclear.

332
 Drug discovery and development: facts and figures Chapter | 22 |

Preclinical Test each scarce

development Phase I molecule thoroughly
Phase II
Phase III
Scarcity
of drug $ $ $$ $$$$
discovery
PD Launch
FED
CS FHD
A Traditional • Increase critical information
content early to shift attrition
to cheaper phase
• Use savings from shifted
attrition to re-invest in the
Preclinical R&D ‘sweet spot’
development
POC
Confirmation, Higher p(TS)
dose finding
Commercialization
Abundance
of drug
discovery
PD Launch

FHD
CS
R&D ‘sweet spot’
B Quick win, fast fail

Fig. 22.10  The quick win/fast fail model for drug development.
Reproduced, with permission, from Paul et al., 2010.

REFERENCES

Abbott A. The drug deadlock. Nature Business and Financial News Apr Drews J. In quest of tomorrow’s
2010;468:178–9. 20th 2010; 2010. medicines. 2nd ed. New York:
Arthur Andersen Consultants. Report on Carroll J. Top 15 Pharma R&D budgets. Springer-Verlag; 2003b.
innovation and productivity; 1997. FiercePharma 12th August 2010 EFPIA. The pharmaceutical industry in
Aronson JK, Ferner RE, Hughes DA. (www.fiercepharma.com/ figures. Brussels: European Federation
Defining rewardable innovation in node/35385/print); 2010. of Pharmaceutical Industries and
drug therapy. Nature Rev Drug DiMasi JA, Hansen RW, Grabowski HG. Associations; 2010. p. 40.
Discov 2012;11:253–4. The price of innovation: new Eichler H-G, Bioechl-Daum B, Abadie E,
Arrowsmith J. Phase III and submission estimates of drug development costs. et al. Relative efficacy of drugs: an
failures: 2007–10. Nature Reviews. Journal of Health Economics emerging issue between regulatory
Drug Discovery 2011;10:1. 2003;22:151–85. agencies and third-party payers.
Association of the British DiMasi JA, Faden LB. Competitiveness Nature Reviews. Drug Discovery
Pharmaceutical Industry (website: in follow-on drug R&D: a race or 2010;9:277–91.
www.abpi.org.uk). imitation? Nature Reviews. Drug FDA. Report. Challenge and opportunity
Austin DH. Research and development Discovery 2011;10:23–7. on the critical path to new medical
in the pharmaceutical industry. Douglas FL, Narayanan VK, Mitchell L, products. (www.fda.gov/oc/
Congressional Budget Office Study; et al. The case for entrepreneurship initiatives/criticalpath/
2006. p. 55. in R&D in the pharmaceutical whitepaper.html); 2004.
Bennani YL. Drug discovery in the next industry. Nature Reviews. Drug FDA. Advancing regulatory science at
decade: innovation needed ASAP. Discovery 2010;9:683–9. FDA: a strategic plan, 2011. p. 34
Drug Discovery Today – in press; 2011. Drews J. Strategic trends in the drug Gilbert J, Hnaske P, Singh A. Rebuilding
Berkrot B. Global drug sales to top $1 industry. Drug Discovery Today big pharma’s business model. InVivo,
trillion in 2014: IMS. Reuters 2003a;8:411–20. the Business & Medicine Report,

333
Section | 4 | Facts and Figures

21(10). Norwalk, Conn: Windhover Nature Reviews. Drug Discovery Paul SM, Mytelka DS, Dunwiddie CT,
Information; 2003. 21(10). 2010;9:867–82. et al. How to improve R&D
Goodman M. Pharmaceutical industry Kola I, Landis J. Can the pharmaceutical productivity: the pharmaceutical
financial performance. Nature industry reduce attrition rates? industry’s grand challenge. Nature
Reviews. Drug Discovery Nature Reviews. Drug Discovery Reviews. Drug Discovery 2010;9:
2009;8:927–8. 2004;3:711–6. 203–14.
Grabowski H, Vernon J, DiMasi JA. Kotz J. Small (molecule) thinking in PhRMA Survey. Biotechnology
Returns on research and development academia. SciBX 2011;4; doi:10,1038/ medicines in development.
for 1990s new drug introductions. scibx.2011.617. (www.phrma.org/newmedicines/
Pharmacoeconomics 2002;20:11–29. LaMattina JL. The impact of mergers on surveys.cfm); 2002.
Harrison C. The patent cliff steepens. pharmaceutical R&D. Nature Pharmaceutical Research and
Nature Rev Drug Discov 2011;10: Reviews. Drug Discovery Manufacturers of America, 2011.
12–3. 2011;10:559–60. (www.phrma.com).
Hirschler B. Drug R&D spending fell in McKinsey. Report. Raising innovation to Reichert JM. Trends in development and
2010 and heading lower. Reuters new heights in pharmaceutical R&D; approval times for new therapeutics
Business and Financial News 30th 1997. in the United States. Nature Reviews.
June 2011 (www.reuters.com/assets/ Maggon K. Global Pharmaceutical Drug Discovery 2003;2:695–702.
print?aid=USL6E7HO1BL20110626); Market Intelligence Monograph Stephens AJ, Jensen JJ, Wyller K, et al.
2011. http://knol.google.com/k/global- The role of public sector research
Hirschler B. Abbott drug tops sales as pharmaceutical-market-intelligence- in the discovery of drugs and
Lipitor, Plavix era ends. Reuters monograph; 2009. vaccines. New England Journal of
Business and Financial News April Mullard A. Learning lessons from Medicine 2011;364:535–41.
11th 2012 (www.reuters.com/assets/ Pfizer’s $800 million failure. Nature Swinney DC, Anthony J. How were
print?aid=USL6E8FA3WO20120411); Reviews. Drug Discovery 2011;10: new medicines discovered? Nature
2012. 163–4. Reviews. Drug Discovery 2011;10:
Hunter D, Wilson J. Hyper-expensive Munos B. Lessons from 60 years of 507–19.
treatments. Background paper for pharmaceutical innovation. Nature Von Schaper E, Kresge N. Novartis
Forward Look 2011. London: Nuffield Reviews. Drug Discovery 2009;8: $48,000 pill spurs price increases
Council on Bioethics; 2011. p. 23. 959–68. for MS drugs. Bloomberg
IMS. The global use of medicines: Nelson AL, Dhimolea E, Reichert JM. Business Week March 23rd 2011
outlook through 2015. IMS Institute Development trends for human (www.businessweek.com/news/2011-
for Healthcare Informatics; 2011. p27. monoclonal antibody therapeutics. 03-21/novartis-s-48-000-pill-spurs-u-
Kaitin K, Milne CP. A dearth of new Nature Reviews. Drug Discovery s-price); 2011.
meds. Scientific American 2011 2010;9:767–74. Zang J. Moving from made in China to
August:16. Pammolli F, Magazzini L, Riccaboni M. discovered in China. Lecture at
Kneller R. The importance of new The productivity crisis in Academy of Medical Sciences 2011,
companies for drug discovery: pharmaceutical R&D. Nature Reviews. May 2011.
origins of a decade of new drugs. Drug Discovery 2011;10:428–38.

334
 Index

Index

NB: Page numbers in italics refer to Adoption pattern, new products, 308 Ankyrin repeat (AR) proteins, 190–191,
boxes, figures and tables. Adrenaline, as trade name, 6 195
Adulimumab, 326 Antibiotics, 45
Advanced Therapy Medicinal Products Antibodies
A (ATMP), 294 catalytic, 176
Absorption, 136–139, 137 Adverse drug reactions, 129 monoclonal (mAbs), 175, 189, 264,
cautions, 138–139 Aequorin, 106 332
protein/peptides, 184–186, 185, Aetiology, 21, 22 selection, phage display, 175–176,
185–186, 290 drug effects, 67 175
rate prediction, 151 Affordability (health), 311 substitutes see Scaffolds
tactics, 136–138 Ahlquist, R., 95 therapeutic agents, 35–36, 175–176
troubleshooting decision tree, 138 Algorithms, clustering, 70 Anticalins, 195–196
Absorption, distribution, metabolism, Alkaloids, plant, 6 Anticancer drugs, 222
excretion (ADME), 246 Alleviation, disease effects, 26, 33 Antimetabolite principle, 12
in-vitro screens, 127 Allopurinol, 12 Antiplatelet gpIIb/IIIa antagonists,
parameters, 126–129 Alloxan, 165 173–175
Abzymes, 176 AlphaLISA™, 105 Antipsychotic drugs, classic, 66
Accelerated Approval Provision, 270 AlphaScreen™ Technology, 105, 105 Antipyretics, 7
ACE inhibitors, 222 Alzheimer’s disease, 66, 166, 267–268, Antisense DNA, 38
Acetazolamide, 11, 11, 222 267, 270 Antisense oligonucleotides, 73
Aciclavir, 12 chromosome 21, 84 Antithrombotic factors, 173–175
Acquired immune deficiency syndrome American Medical Association (AMA), Apothecaries’ trade, 6–7
(AIDS), 66 304 Artemether, 45
Actinomycin D, 8 Ames test, 218 Arterial spin labelling (ASL), 272
Active metabolites identification, 148 p-Aminobenzoic acid (PABA), 10, 10 Aspirin-like drugs, 66
cautions, 148 Amlodipine, 327 Assay development, 95–96
tactics, 148, 149 Amphetamine, 13 drug targets, 89–90
Active pharmaceutical ingredients Amphiphilic molecules, 232–233 miniaturization, 109
(API), 227 Amphiphilic polymer, 232 readout/detection, 102–105
Acute illness, 60 Amyl nitrite, 5 validation, 99–112, 99
Acute pharmacological animal model, Anatomical studies, 70 Assay Guidance Website, 100
165 Ancient Greeks, 3–4 Association analysis, 81–82
Acute physiological animal model, Angiotensin agonists, 222 Asthma, 66
165 Animal disease models, 157, 158, Atherosclerosis, 165
Acute seizure models, 168 164–169 Atomic form, 81
Acyl glucuronides, 149 choice, 166–169 Atorvastatin, 326–327, 329
Adenoassociated virus (AAV), 178 examples, 167–169 Atropine, 8
Adenovirus, 178 types, 165–166 Attrition, drug, 129–131, 322,
Administrative rules, regulation, Animal pharmacology, 158 329–331, 330
300–301 profiling, 157, 161–164 Augmented toxicity, 130
Adnectins, 190–191, 192–193 species differences, 164, 164 Autonomic nervous system, tests,
A-domains, 190–191, 193 Animals, transgenic, 73–74, 184 214–215

335
 Index

Azathioprine, 12 Biophysical methods, high-throughput Cells, therapeutic agents, 35–36

AZT (zidovudine), 12, 328 screening (HTS), 108–109 Cellular DNA transfer, 177–180
Biotechnology, 171–172 Central nervous system
biotechnically-derived medicines, 331 cautions, 145
B 1,6-Bis(phosphocholine)-hexane, 122, core battery tests, 214–215
Baclofen, 237 122 drug delivery, 236–237
Bacterial infections, 66 Black Box Warnings (FDA), 130 follow-up tests, 214–215
Bacterial protein expression, 182 Black, James, 12–13 tactics, 143–145, 144
Barbitone (barbital), 7 Blockbuster drugs, 7, 331–332 uptake, 143–145
Benefit, defines, 25, 27 marketing, 304–305 Centralized Evaluation Report (EU),
Benzene, 5–7 molecules, 316 254
Benzodiazepines, 66 sales, 326–327, 326–327 Ceruvastatin, 326
Beraprost, profiling, 163, 163 Blood clotting factors, 174 Chemical Abstracts (CA) (Ohio State
Bile Salt Export Pump (BSEP), 130 Blood oxygen level dependent (BOLD) University), 279
Biliary clearance, 147–148 fMRI, 260–261, 272 Chemical Genomics Center (NIH), 100
cautions, 148 Blood-brain barrier penetration, 262 Chemistry
tactics, 147–148 Boundary layer, 228 20th century discoveries, 7
Binding assays, 159–160, 160 Bradykinin B2 receptors, 164, 164 developments, 5–6
Bioavailability Breast cancer, 66 natural product, 8, 9
absolute, 245–246 Bridging studies, 251 synthetic, 7–8, 9
oral see Oral bioavailability British Pharmacopoeia, 4 see also Medicinal chemistry
Bioaxis, 24–25 Bucheim, Rudolf, 4–5 Chemoreceptors, 95
Biochemical assays, 100–101, 100 Bulk phase, 228 Chemotherapy, 5
Biodistribution, imaging, 259, Bumetanide, 11 Cheng, Prusoff equation, 159–160
262–264, 262 Burimamide, 12–13 ‘Cherry picking’, compounds, 96
Bioequivalence, 245–246 Buserelin, 230 Children
Bioinformatics, 78–82 clinical trials, 252, 293
Bioinformation, 78 special populations, 246–247,
C
Biological perspective, 24–26 292–293
organization levels, 24–25, 25 Caenorhabditis elegans, 73 Chloramphenicol, 8
Biology-oriented synthesis (BIOS), 125 Caffeine, 45, 222 Chloroform, 5, 7
Biomarkers, 130 Calcitonin, 185, 230 Chlorothiazide, 11, 11
classification, 152, 153 Cancer, 165 Chlorpromazine, 13
genomics, 83–85 immunotherapy, 176 Chromosomes, 82
imaging, 266, 267, 268 Cannabis, 222 abnormalities, 218
Biomedicine, developments, 4–5 Capitation models, pricing, 311 chromosome 21, 84
Biomolecules, 95 Capofen, exclusivity period, 328 damage, 216–218
Biopharmaceutical Classification Capsule formulations, 231 Chronic myeloid leukemia, 66
System, 136 Carbutamide, 11, 11 Chronic pharmacological animal
Biopharmaceuticals, 34–37, 171–188 Carcinogenicity, 290 model, 165
advantages/disadvantages, 36 testing, 220–221 Chronic physiological animal model,
chronic toxicology studies, 220 Cardiac muscle cells, 39 165
delivery/formulation, 235–236 Cardiovascular system Chronic symptoms, 60
development, 9, 17, 17 core battery tests, 214–215 Chronic toxicology studies, 218–220
discovery, 44 follow-up tests, 214–215 biopharmaceuticals, 220
gene therapy, 177–180 toxicity, 130, 217 evaluating toxic effects, 219–220
major classes, 172–177, 174 Catalytic antibodies, 176 experimental design, 219, 219
manufacture, 181–182 Cathepsin K, 70–72 segment 1, 221
protein therapeutics, 172 CDNA microarray methods, 224 segment 2, 221
proteins/peptides difficulties, Celebrex, exclusivity period, 328, 328 segment 3, 221–222
184–186 Celecoxib, 327, 328 Ciclosporin, 8, 44–46, 66
quality assessment, 294–295 Cell theory (Virchow), 4 Cimetidine, 12–13, 328
recombinant DNA technology, Cell-based assays, 95, 100–101, 101, Cinchona extract (quinine), 4, 6, 13
171–172 101, 105–108 Claims, patents, 276, 278
regulation, 294–295 readouts, 105–106 ‘reach-through’, 280
research & development, 332 test system characteristics, 158 Clearance
safety, 295 Cell-based therapies, 38–39 mechanisms, 141–142
vs small molecule therapeutics, autologous cell grafts, 38–39 optimization, 145–148
180–184 cell-replacement, 38 prediction, 151

336
 Index

Clinical development, 239–258 Construct validity, 166, 168 Data mining, 79

children, 252 Continued Medical Education (CME), general principles, 79–82
clinical pharmacology, 240–247, 241 304, 310 pattern discovery, 80–81
discoveries, 13 Controlled release, drugs, 235 Databases, 149
dose range finding studies, 249–251 Conus magus, 121–122 Decision to develop in man (DDM),
efficacy, 247–251 Conventional drugs, 34, 34, 35–36 207–208
64
future trends, 254–256, 256–257 Copper, 264 Declaration of Helsinki (World Medical
phases, 240–252, 241 Core battery, safety tests, 213, 214–215 Association), 252–253
regulation/ethics, 252–254 Coronary ischaemia, 165 Delayed release, drugs, 235
Clinical imaging, 259–274 Corpora non agunt nisi fixata (Ehrlich), DELFIA (dissociation-enhanced
biodistribution, 259, 262–264, 262 8 lanthanide fluoroimmuno assay),
implementation, 270–272 Correlation, 81 104
methods, 259–261 Cosmeceuticals, 33 Depression, 66
patient stratification, 245 Costs Derwent Publications Ltd, patents,
personalized medicine (PM), identification, 28 279
245–246 research and development (R&D), Desirability, 78
pharmacodynamics, 259, 266–268, 321–325, 322–323, 323 Desmopressin, 185
266, 267 Cost-benefit analysis, 30 Development process, new drug
as surrogate marker, 246 Cost-effectiveness, 28–29 approvals, 208–209
target interaction, 259, 264–266, 265 Cost-utility analysis, 29 Developmental toxicology studies,
target validation, 259, 261–262 Covalent DNA Display technology, 221–222, 291, 291
Clinical research organization (CRO), 194 Diabetes Type I, 165
253–254 CPHPC crosslinking agent, 122, 122 Diazoxide, 11, 11
Clinical studies Creutzfeldt–Jakob disease, 172–173 Dichloroisoprenaline, 12
compounds, low concentration, Critical micelle concentration, Diclofenac, 228
89–90 232–233 Diethyl ether, 5
efficacy, 90 Critical path, 204–205 Diethylene glycol, 13–14
marketing, 307 Critical Path Initiative (CPI) (FDA), Diethylstilbestrol (DES), 221
models, 61 240, 259 Diflucan, exclusivity period, 328
pharmacokinetic, 90 Cromoglycate, 45 Digitalis, 8, 13
safety, 90 Cure, permanent, 26 Digoxin, 48
Clinical trials, 253–254 Customer portrait, market planning, Dihydropyridines, 66
adaptive design, 254–256, 256–257 314 Diphenylhydantoin, 222
children, 252, 293 CYP Direct to Consumer (DTC) advertising,
regulation, 296–297 induction, 142 309, 312
Clonidine, 50 phenotyping, 141–142 Disabilities, present, 21–22
Clopidogrel, 326, 329 see also Cytochrome P450 (CYP450) Disclosure, 254
Clozapine, 27–28 CYP inhibition Disease
Cluster analysis, 81, 87 competitive (reversible), 140–143, aetiology, drug effects, 67
Coagulation factors, 173 140 alleviation of effects, 26, 33
Cochrane Collaboration, 311 cytochrome P450 (CYP450), common, 60–61
Colour quench effect, 102 127–128 components, 22
Commercialization, 43, 44 time-dependent (mechanism based) concepts of, 19
Committee on the Safety of Drugs (TDI), 140, 141, 149 defined, 20–21
(UK), 14 Cystic fibrosis, 166 genes, 68–72
Common disease, 60–61 Cytochrome P450 (CYP450), 127 intrinsic mechanisms, 89
Common Technical Document (CTD), inhibition, 127–128 nature of, 25
299–300, 300 Cytokines, 173 permanent cure, 26
Comparability, 294–295 Cytotoxicity assays, 128–129 prevention, 26, 33
Competition, 60 rare, 60–61
Complementary procedures, 34 Displacement assay, 159
D
Compound logistics, 95–96 Dissociative fluorescence enhancement,
Compound number 606 (salvarsan), Darmstadt, 6 104
9–10, 95, 96 DARPins, 195 Dissolution rate, 228
Compounds, low concentration, 89–90 Data Disvalue, 21, 22, 30
Concerned Member States (EU), analysis vs data management, 79 components, 21–22, 22
297–298 dredging, 80 DNA
Confidentiality, 254 exclusivity, 300 liposome-encapsulated, 178
Confocal imaging, 108 variability, 98 naked, 178

337
 Index

‘vaccines’, 171 genomics, 88–90, 88 exploratory, 247–249

see also Recombinant DNA patents, 278 in man, 291–294, 292
technology phases, 43–50, 44–45 patient recruitment, 251–252
Documents, patents, 278 pipelines, 322, 329–331, 331 Efflux transporter inhibition, 140–141
Dofetilide, 215 quick win/fast fail model, 333 EGFR receptor kinase inhibitors, 162
Dominant negative strategy, 178 research, 55–56 Ehrlich, Paul, 5, 8–9, 95
Donepezil, 329 stages, 50–52, 51 Elderly, special populations, 246–247,
Dose timelines, 47, 328 292–293
escalation protocol, 215–216 trends, 52–55, 53 Elimination, proteins/peptides, 185,
forms, 229–230, 229 see also Pharmaceuticals 186
lethal, 223 Drug metabolism and Elion, Gertrude, 12
maximum tolerated (MTD), 221, pharmacokinetics (DMPK), E-marketing, 310
223, 242 135–155, 136 Enoxaparin, 329
measures, 223 human PK/dose prediction, Environment, drug development, 296
prediction, 150–152, 150 150–152, 150 Enzyme-linked immunosorbant assay
range-finding studies, 215–216, 216, optimization, 135–150 (ELISA), 105, 192
217, 249–251 Drug Ordinance (1962) (Sweden), 286 Enzymes, 95
see also First dose in man (FDIM); Drug safety see Safety therapeutic agents, 35–36, 176
Multiple ascending repeat-dose Drug targets, 8–12, 9, 63–76 Epigenome, 85
(MAD) studies assay development, 89–90 Epilepsy, 165
Drapetomania, 21 candidate profile (CDTP), 137 models, 162, 167
Drosophila spp, 73 identification, 65–72, 67, 89 Epileptogenesis, 167, 168
Drugs nature of, 65 Epinephrine, 6, 12
adverse reactions, 129 new, scope for, 63–65 Epoetin-α, 327
inhaled, 129 number of, 63–65, 64 Epothilones, 44–46
intravenously administered, 129 receptors, 12–13 Equivalence, patents, 276
lifestyle, 22 vs therapeutic targets, 25, 26 Ergot alkaloids, 8
Drug delivery systems validation, 72–74, 89 Erythropoietin (EPO), 46, 173, 174
nanotechnology, 234–235, 235 Drug-delivery issues, 184–186 Escherichia coli, 34–36, 175–176, 182
principles, 231–237 Drug-drug interactions (DDI), Escitalopram, 329
Drug development, 43, 44, 203–209 139–143, 140, 245 Esemeprazole, 326
attrition rates, 129–131, 322, cautions, 143 Etanercept, 326
329–331, 330 prediction, 142–143, 142 Ethanol, 222
components, 204–207, 205–206 tactics, 140–143 Ethics
costs, 324–325 Drug-induced liver injury, 130 clinical development, 252–254
decision points, 207–208, 207 Drug-likeness, 123–124, 129, 157 procedures, 253
discovery, interface with, 207 ‘Druggability’, 207 Ethics Committee (EU), 254–255
efficacy see Efficacy studies ‘Druggable’ genes, 68, 72 Ethnic differences, 292–293
environment, 296 Duchenne muscular dystrophy, 166 Etoposide, 45
exclusivity period, 328, 328 Dyestuffs industry, 6–7 Eukaryotic microbial systems, 182
genomics, 88–90, 88 Dysfunction, 30 Europe
improvement need, 208–209 Dysfunction, function and, 24–26 marketing, 297–299, 298
nature of, 203–204, 205 regulation, 296–297
orphan drugs, 296 European Bioinformatics Institute
E
pipelines, 322, 329–331, 331 (EBI), 78
process, 288 Early adopters, 312–314 European Commission, 298
quality assessment, 289 Early majority, 312–313 European Economic Community,
recent introductions, 331–332 Early selection point (ESP), 207 bioinformatics, 78
regulation, 288–296 Ebers papyrus, 3 European Medicines Agency (EMA),
research see Research and Effect, defined, 25, 27 293
development (R&D) Effectiveness, defined, 25, 27 European Medicines Evaluation Agency
safety assessment, 289–291 Efficacy (EMEA), 83
seamless, 254–256, 256–257 defined, 61 guidance document, genomics,
see also Pharmaceuticals doses, prediction, 152, 153 92–93
Drug discovery, 43–56 Efficacy studies, 25, 27 initiatives, 91
case histories, 46–50, 46 biopharmaceuticals, 295 European Molecular Biology
development, interface with, 207 clinical, 90 Laboratory (EMBL), 78
future trends, 332, 333 confirmatory, 249–251 European Patent Office, 279

338
 Index

Europium (Eu3+) chelates, 104, 104 draft Guidance 2010 for adaptive drug targets, 73–74
Expert Systems, 79 clinical trials, 255–256 vaccination, 177
Eyes, toxicity tests, 217 efficacy, 286 Genitourinary system, toxicity tests,
genomics, guidance documents, 217
92–93 Genome, 25, 82
F imaging, 259 Genomics, 67–72, 82–91, 95
Face validity, 166, 168 IND approval, 211, 297 biomarkers, 83–85
Feasibility, 78 initiatives, 91 cluster analysis, 87
Fentanyl, 230 marketing, 299 drug discovery/development, 88–90,
Fibrates, 66 paediatric studies, 293 88
Fibronectin, 194 patents, 275, 282–283 guidance documents, 92–93
tenth type III domain, 190–191, pregnancy, 222 regulatory authorities, 90–91
192–193 recent introductions, 331–332 sequences, 83–85
First dose in man (FDIM), 242–244 Voluntary Exploratory Data terminology, 82–83, 82
design, 242–243 Submissions (VXDS), 91 usefulness, 87–88
objectives, 242–244 Food, Drug and Cosmetics Act, 270 variability, 85–86
outcome measures, 243–244 Food and Drugs Act (USA) Genotoxicity, 216–218, 290
subjects, 242 1906, 13–14 Germ theory of disease (Pasteur), 5
5HT receptor ligands, 159–162, 160 1937, 13–14 Ghrelin, 121–122, 122
FK506 (fujimycin, tacrolimus), 8, Formulation, 230–231 Gleevec see Imatinib
44–46, 66 biopharmaceuticals, 171 Glibenclamide, 11
FlashPlate™, 102 process, 228 Global brand management, 311
Flecainide (Tambocor), 46, 47, 48 Four humours, 3–4 Glucagon, 172
case history, 46–47, 46, 47 Fractional Clint assay, 141–142 Glucocerebrosidase, 172
Fluorescence Fragment-based screening, 108–109 Glycome, 82
intensity, 103 Franchise, 59 GOD-gels (glucose oxidase), 235
lifetime analysis, 105 Freedom to operate, 61, 278 Good Clinical Practice (GCP)
polarization (FP), 104, 107 Fujimycin (FK506, tacrolimus), 8, standards (ICH), 253–254
quench assays, 103 44–46, 66 Good laboratory practice (GLP), 169
resonance energy transfer (FRET), Full development decision point Government policy, 59
103, 103 (FDP), 208 G-protein-coupled receptors (GPCRs),
technologies, 103–105 Function and dysfunction, 24–26 95, 101, 108, 161
Fluorescence correlation methods, Functional magnetic resonance Granulocyte colony-stimulating factor
104–105 imaging (fMRI), 268, 271–272 (G-CSF), 173
spectroscopy, 104–105 Fur, toxicity tests, 217 Granulocyte-macrophage colony-
Fluorescence Imaging Plate Reader, Furosemide (frusemide), 11, 11 stimulating factor (GM-CSF), 173,
106 Fyn kinase, SH3 domain, 190–191, 194 174
[18F]-fluorodeoxyglucose (FDG), Fynomer clone, 194 Growth factors, 173
267–268, 267 Growth hormone, 172
Fluorogenic assays, 103 Guidelines for the Quality, Safety and
G
Fluorometric assays, 106 Efficacy Assurance of Follow-on
Fluorophore, 103 GABAA receptor functions, 70–72 Biologics (Japan), 294–295
Fluoxetine, 327 Galantamine, 44–46
Fluticasone/salmeterol, 326 Gastroesophageal reflux disease
H
Focus groups, 308 (GORD), 248
Foldome, 82 Gastrointestinal system Haemophilus influenzae, 172
Folic acid synthesis, 10 supplementary tests, 214–215 Haemopoietic factors, 174
Follicle-stimulating hormone (FSH), toxicity tests, 217 Haloperidol, 27–28
172, 174 ‘Gene chips’, 69–70 Harm, 21
Follow-up tests, safety pharmacology, Gene expression, profiling, 69–70, 69, Harris Interactive, 304–305
213, 214–215 71–72 Health
Food and Drug Administration (FDA), Gene knockout screening, 70–72 defined, 20
83, 119 Gene silencing, 178–180, 180 foods, 34
Black Box Warnings, 130 Gene therapy, 37–38, 177–180 Heparin, 45
clinical benefit, 270 vectors, 178 Hepatoxicity, 130
compounds, success rate, 330 General signs, toxicity tests, 217 Herbal remedies, 4
development process and, 208–209, Genetic approaches Herceptin see Trastuzumab
240 animal models, 165–166 hERG channel, 215

339
 Index

hERG gene, 215 Industrial applicability, invention, 278 Invirase, exclusivity period, 328
124
High content screening (HCS), 108 Infection theory (Koch), 5 Iodine, 264
High Mobility Group Box1, 130 Infectious agents, biology, 89 IonWorks Quattro, 107
High-throughput electrophysiology Inference, 79 Ivermectin, 45
assays, 107, 107 Inflammatory disease, 66
High-throughput medicinal chemistry, Infliximab, 326
J
95 Information industry, 77
High-throughput screening (HTS), bioinformatics, 78–82 Jackknife method, 79–80
95–117 innovation, 77–78, 78 Japan
activity cascade, 97–98, 98 Information sources, patents, 279 marketing, 299
compound logistics, 112–113 Inhaled drugs, 129 regulation, 297
data analysis, 111–112, 112–113 Inhomogeneities, 260–261 Japanese Patent Office, 279
future trends, 95–97 Innovation, 315–316
historical perspective, 95–97, 96 Innovators, 312–314, 313
lead discovery, 97–112 Innovative Medicines Initiative
K
profiling (HTP), 113–114 (EFPIA), 240 Kefauver–Harris Drug Control Act
robotics, 109–111, 110 Insulin-secreting cells, 39 (1962), 304
screening libraries, 112–113, 114 Integrated Summaries of Efficacy (ISE), Keratin 18 proteins, 130
Hirudin, 45 299–300 Ketoprofen, 230
Hismanal, exclusivity period, 328 Integrated Summaries of Safety (ISS), Key opinion leaders (KOL), 310
Hitchings, George, 12 299–300 Key stakeholders, 315
Holmes, Oliver Wendell, 4 Intellectual property (IP), 283–284 Kindling model, 165, 168
Hormones, 173, 174 see also Patents Koch, Robert, 5
Hostetter’s Celebrated Stomach Bitters, Intelligent materials, 234–235 Kogenate, exclusivity period, 328
303–304 Interactome, 82 Kunitz type domains, 190–191,
HTRF® (Homogeneous Time-Resolved Intercontinental Marketing Services 191–192
Fluorescence), 104, 104 (IMS), 304
Human factor VIII, 173, 174 Interferons, 173, 174
Human factor IX, 173, 174 interferon-α, 173
L
Human Genome Project, 88 Interleukins, 174 Label free detection platforms, 108
Human gonadotrophin-releasing International Conference on LANCE™, 104
hormones, 173 Harmonization (ICH), 252, 286 Law of Mass Action, 12
Human growth hormone Good Clinical Practice (GCP) Lead discovery, high-throughput
(somatotropin), 173, 174 standards, 253–254 screening (HTS), 97–112
Human immunoglobulin G (IgG) Guidance E6, 252–253 Lead identification, 52
antibody, 189 Guideline M3, 290 Lead optimization stage, 52, 97–98,
Human lifespan, 23, 23 Guideline Q5A-E, 294 98
Human pharmacology studies, Guideline Q6B, 294 Lead-like property, 123–124
291–292, 292 Guideline S1A, 220–221 LEADseeker™, 102
Hydralazine, 11 Guideline S1B, 220–221 Legislation, 59
Hypercholesterolaemia, 165 Guideline S1C, 220–221 Leptin, 165
Hyperpharmacology, 212–213 Guideline S5A, 221 Levodopa, 237
Hypertension, 66 Guideline S7A, 213 Lidocaine, 48
Guideline S7B, 215 Lifestyle drugs, 22
Guideline S9, 224 Life-years saved per patient treated,
I harmonization process, 287 28–29
Idiosyncratic reactions, 213 quality, 294 Ligand binding assays, 102–103
Idiosyncratic toxicity, 130 International Nucleotide Sequence Ligand efficiency (LE), 124
Imatinib (Gleevec), 46, 47, 49–50 Database Collaboration (INSDC), Ligand-lipophilicity efficiency (LLE),
case history, 46–47, 46, 47 78 124
Imipramine, 50, 222 Intranasal administration, 230 Lipocalins, 190–191, 195–196
Immunomodulation, antibody use, Intravenously administered drugs, 129 Lipophilicity, 129
176 Inventive step, 277–278 Liposomes, 233–234, 234, 235
IMS Health Inc., 304–305 Investigational medicinal product Local tolerance studies, 291
new product launch, 312–313 (IMP), 263 Lopressor, exclusivity period, 328
INAHTA (International Network of Investigational New Drug (IND) Loratadine, 327
Agencies for Health Technology approval (FDA), 211 Losec see Omeprazole
Assessment), 311 regulation, 297 Lymphoma tk cell assay, 218
Inderal, exclusivity period, 328 Investigative studies, 203, 205 Lypressin, 185

340
 Index

M Medicines Act (1968) (UK), 14, 286 Natural products, 8, 9

Melanocortin receptors, 70–72 drug examples, 44–46, 45
McMaster Health Index, 29 Melphalan, 237 Naturalist view, 21
‘Magic’ bullets, 8–9 6-Mercaptopurine, 12 Nerve growth factor (NGF), 186
Magnetic resonance imaging (MRI), Metabolic clearance optimization, Nervous system, toxicity tests, 217
260 145–147 Net present value (NPV), 325
functional (fMRI), 260–261 cautions, 146–147 Neuronal cells, 39
Malignant disease, 66 tactics, 146 New active substance (NAS), 252
Managerial functions, 203–204 Metabolite identification, 148–150 New Chemical Entities (NCEs), 119,
Markers Metabolome, 82, 82 120, 279–280, 329
imaging, 246 Methylphenidate (Ritalin), 13 New Drug Application (NDA) approval
see also Biomarkers Methylxanthines, 45 (USA), 211
Market Mevacor, exclusivity period, 328 Nicotine, 165, 230
considerations, 58–59 Mevastatin, 8 Nitroglycerin, 230
identification, 307–308 Micelles, 232–233, 233, 235 Nitrous oxide, 5
map, 314 Microbeads, 102, 104 No observed effect level (NOEL), 223
research, 308 Microplates, 102, 109 no observed adverse effect level
Marketing, 303–317 Microtitre plates, 99 (NOAEL), 223, 242–243
authorization, 297–299 Miniaturization, 97, 109 No toxic effect level (NTEL), 215–216,
competition assessment, 308–309 Modified-disease drug formulations, 223
defined, 303 235 Norepinephrine, 12
e-marketing, 310 Molecular assays, 158 Normality, deviation from, 20–21
environment, changing, 314–316, 315 Monoclonal antibodies (mAbs), 174, Normative view, 21
Europe, 297–299, 298 175, 189, 264, 332 Norvir, exclusivity period, 328
future trends, 316–317 Morbidity, 21–22, 30 Nottingham Health Profile, 29
health technology assessment (HTA), Morphology, 228–229 Novelty, 277
311–312 Mouth, toxicity tests, 217 regulation, 294–296
historical perspective, 304, 311–312 MPTP neurotoxin, 165 in USA, 277
imaging marketed drugs, 266 mRNAs, 71–72 Nucleic acid-mediated intervention,
Japan, 299 expression analysis, 86 178, 179, 180
life cycle, pharmaceutical, 306–307, Mucosal delivery, 184–185 Nucleic acids, 177–178, 178, 187
306 Multiple ascending repeat-dose (MAD) polymers, 177
market considerations, 58–59 studies, 244–245, 244, 290 Nucleotides, 179
market identification, 307–308 design, 244 Nutriceuticals, 33
plan implementation, 314, 314 outcome measures, 244–245
plans, 305
pricing see Pricing
Multiple sclerosis, 165 O
Multiwell plates, 109
traditional, 307–310 Mutagenesis, site-directed, 182 Oestradiol, 230
USA, 299 Mutagenicity, 216, 218 Off-target effects, 213
see also Product launch; Product Myelin, 165 Olanzapine, 326–327, 329
life-cycle Oligonucleotides, 178, 180
Marketing Authorisation Application Omeprazole (Losec), 48–49, 327
(MAA) approval (Europe), 211
N
case history, 46–47, 46, 47
Mauvein, 5–6 Nafarelin, 185 Ondansetron, 50
Maximal electroshock (MES) model, Nanoemulsions, 235 Operational issues, 58, 61–62
167, 168 Nanogels, 235 Opiates, 45, 165
Maximum tolerated dose (MTD), 221, Nanoparticles, 235 Opioids, 230
223, 242 Nanoprobes, 235 Opportunity costs, 324
Mechanism-related effects, 212–213 Nanotechnology, 234–235, 235 Oral bioavailability, 136–139, 137
Media launch, product launch, 312 National Center for Biotechnology prediction, 151
Medical need, unmet, 57–58 Information, 78 Oral drugs, project planning, 54
Medicinal chemistry, 119–134 National Health Service (NHS), Organic Anion Transporter 3 (OAT3),
attrition, 129–131 311–312 130–131
lead identification/generation, National Institute for Health and Organs
123–125 Clinical Excellence (NICE), 29–30, therapeutic agents, 35–36
lead optimization, 125–129 311–312 transplantation, 39
New Chemical Entities (NCEs), 119, ‘NICE blight’, 311–312 Oromucosal administration, 230
120 National Institutes of Health (NIH), 78 Oromucosal delivery, 230
target selection/validation, 120–122 Chemical Genomics Center, 100 Orphan drugs, 296

341
 Index

Outcomes research, 311 recruitment, 248, 250–252 regulatory, 290

Oxytocin, 185 special populations, 246–247, 285 see also Drug metabolism and
stratification, imaging, 245 pharmacokinetics (DMPK)
switch, 313 Pharmacological Institute, Dorpat,
P Pattern discovery, 80–81 4–5
PABA (p-aminobenzoic acid), 10, 10 Pay for performance models, pricing, Pharmacology, 4–5, 157–170
Paclitaxel (Taxol), 8, 45 311 clinical, 240–247, 241
case history, 46–47, 46, 47 Payer, value system, 315 drug target approaches, 72–73
discovery, 46, 47–48, 47 Penicillin, 8, 222, 237 efficacy, 161–162
Paediatric Regulation (2007) (EU), Pentylenetetrazol-induced seizure general, 289–290
293 (PTZ) model, 167, 168 good laboratory practice (GLP), 169
Paediatric Research Enquiry Act (2007) Peptides human, studies, 291–292, 292
(USA), 293 difficulties, 184–186 primary, 289
Paediatric use marketing authorization mediators, 35–36 profiling, 157, 161–164
(PUMA), 293 Periodic Safety Update Reports safety, 157, 213–215, 214–215, 246,
p-Aminobenzoic acid (PABA), 10, 10 (PSURS), 251 289–291
Parallelization, 97 Perkin, William Henry, 5–6 selectivity screening, 157, 159–160
Parenteral delivery, 185–186 Personalized medicine (PM), 83, 85, test systems, characteristics, 158
Paris Convention for the Protection of 89, 295–296 in vitro, 158, 161–162, 163
Industrial Property, 280–281 imaging, 245–246 in vivo, 162–163, 163
Parkinson’s disease, 66, 165 PERT (project evaluation and review see also Animal disease models
Particle size, 228–229 technique), 204–205 ‘Pharming’, 183–184
Passive distribution model, 263 P-glycoprotein (Pgp) assays, 127–128 Phenobarbital, 7–8
Passive targeting, 233 Phage display technique, 177 Phenome, 82
Pasteur, Louis, 5 antibody selection, 175–176, 175 Phenomenology, 21, 22
Patent Cooperation Treaty, 282, 283 Pharma 2020 (PWC), 315–316 Phenytoin, 7–8
Patent and Trademark Office (US), Pharmaceutical Affairs Law (1943) Phocomelia, 286
279 (Japan), 286 Picha pastoralis, 182
Patents, 275–284 Pharmaceutical Benefits Advisory Picoprazole, 48–49
bibliographic details, 276 Committee (Australia), 29–30 Pipelines, 322, 329–331, 331
claims, 276, 278 Pharmaceuticals, 3–18, 227–238 Planar patch-clamp instruments, 107,
defined, 275 19th century, 4 107
description, 276 antecedents/origins, 3–4 Plantibodies, 183–184
development compound protection, dosage forms, 229–230, 229 Plasma concentration time profile, 152
280–284 drug delivery systems, 231–237 Plasminogen activators, 174
documents, 278 emerging, 4–14 Plasmodium spp, 183–184
drug discovery, 278 formulation, 230–231 Ploglitazone, 329
expiry, 329 as information industry, 77 Polymers, 231–237, 232–233
filing applications, 280–284 major trends, 16–17 Population-focused medicine, 89
information sources, 279 marketing see Marketing Positron emission tomography (PET),
maintenance, 282 milestones, 15–16 163, 259–260, 260, 267–268, 267
New Chemical Entities (NCEs), patents, 276–277 Postcode lottery, 311–312
279–280 preformulation studies, 227–229 Post-seizure models, 168
offices, regional, 282 pricing, 300–301 Practolol, 211
professional evaluation, 279 regulation see Regulation Pravachol, exclusivity period, 328
project planning, 61 research, 55–56 Prediction
requirements, 277–278 routes of administration, 61, 186, predictive analysis, 79, 81
research tools, 280 229–230, 229 of toxicity, 131
rights enforcement, 283 see also Drug development; Drug validity, 167, 168
scientific evaluation, 279 discovery Preformulation studies, 227–229, 237
specification, 276 Pharmacodynamics (PD), 292 Prescriber decision-making power, 314
state of the art (prior art), 278 imaging, 259, 266–268, 266, 267 President’s Council of Advisors on
subject categories, 276–277 studies, 245 Science and Technology (PCAST),
Supplementary Protection Certificate, Pharmacoeconomics, 26–30 269
300 Pharmacoepidemiology, 26–30 Prevacid, 327
term extension, 282–283 Pharmacokinetics (PK), 139–143, 140, Prevention, disease, 26, 33
Patients 242–244, 292 Price-volume agreements, 311
add-on, 313 clinical studies, 90 PriceWaterhouseCooper (PWC),
new, 313 protein/peptide issues, 184–186 Pharma 2020, 315–316

342
 Index

Pricing, 59, 300–301, 310–311 production, 183–184, 184 drug development process, 288–296,
freedom of, 310–311 therapeutics, 172, 174 288
regulated, 311 see also Scaffolds Europe, 296–297
Primary pharmacodynamic studies, Protein expression genomics, 90–91
157 bacterial, 182 historical perspective, 285–286
Prior art (state of the art), patents, ‘pharming’ of, 183, 184 international harmonization,
278 Proteome, 25, 82, 82 286–287, 287
Procainamide, 48 Provider Japan, 297
Procaine, 7 therapeutic decision making, 315 pricing, 311
Processes, patents, 276 value system, 315 procedures, 295–300
Product launch, 312–314 Prozac, 275 process, 13–14
launch meeting, 312 exclusivity period, 328 project planning, 60–61
manufacture/distribution, 312 Psychiatric disorders, 167–169 regulatory affairs (RA) department,
media launch, 312 Public Health Service (USA), 29–30 287–288
registration, 312 Pulmonary administration, 230 roles/responsibilities, 287–288
resource allocation, 312 Pyrimethamine, 12 toxicity testing, 224
target audience, 312–314 USA, 297
Product life-cycle, 305–306, 305 Reimbursement, 59
Q
introduction phase, 305 Renal clearance optimization, 147
development phase, 305 QT interval prolongation tests, 215 cautions, 147
growth phase, 305–306 Quality assessment, 289 tactics, 147, 147
maturity phase, 306 Quality assurance, 253–254 Renal function, tests, 214–215
decline phase, 306 Quality chemical probe, 121 Renal toxicity, 130–131
Production methods, Quality module, 289 Repeated-dose studies, 290, 291
biopharmaceuticals, 171 Quality-adjusted life years (QALYs), Reporter gene assays, 106–107, 106
Productivity (1970-2000), 17 29–30, 29 Reproductive toxicology studies,
Products Quality-of-life estimate, 29, 29 221–222, 291, 291
features/attributes/benefits/ Query mode, 79 Research and development (R&D), 316,
limitations (FABL), 308 Quetiapine, 326, 329 321–325
future, 316 Quinidine, 48 biotechnically-derived medicines,
patents, 276 Quinine (Cinchona extract), 4, 6, 13 331
Profiling, 95–97 by function, 323
Profitability, 325, 325 compound attrition, 322
Prognosis, 21–22, 26, 30
R compounds in development, current,
therapeutic modalities, 33 Radioactive assays, homogenous 323
Project planning, 57–62 formats, 102 pharmaceutical vs other industries,
desirability, 57 Radio-labelled studies, 246 323
drug discovery, 54–55, 54 Ranitidine, 328 private vs public spending, 322
feasibility, 60 Rapamycin, 44–46, 49 recent introductions, 331–332
selection criteria, 44, 46–50 Rare disease, 60–61 Research tools, 61
Promazine, 13 ‘Reach through’ claims, 61 patents, 280
Promethazine, 13 Reactive metabolites identification, Resource allocation, product launch,
Pronethalol, 12 149–150 312
Prontosil, 9–10, 11 cautions, 150 Respiratory system
Proof of concept (PoC) studies, 120 tactics, 149–150 core battery tests, 214–215
design, 247–249 Reagent production, 95–96 follow-up tests, 214–215
objectives, 247 Receptors, 95 toxicity tests, 217
outcome measures, 249 Receptor-targeted drugs, 12–13 Retrovirus, 178
subjects, 248 RECIST criteria, 268 Reyes’ syndrome, 27–28
Propranolol, 12 Recombinant DNA technology, 9, Ribosome display selection
Protease protein C, 175 35–36, 171–172 methodology, 195
Protein approved proteins, 174 Ritalin (methylphenidate), 13
difficulties, 184–186 Recombinate, exclusivity period, 328 RNA interference (RNAi), 73, 178–180,
engineering, 182–183 Rectal administration, 230 180
expression systems, 181–182, 181, Reference Member State (RMS) (EU), RNA-induced silencing complexes
181 297–298 (RISC), 73, 178
folding, 88 Regulation, 285–301 Robotics, high-throughput screening
fused/truncated, 182–183 administrative rules, 300–301 (HTS), 109–111, 110
mediators, 35–36 clinical development, 252–254 Rofecoxib, 211, 328

343
 Index

Rosuvastatin, 130–131 Septic shock, 165 Strychnine, 8

Routes of administration, 61, 186, Sertürner, Friedrich, 6 Submission decision point (SDP), 208
229–230, 229 Severe combined immunodeficiency Sulfanilamide, 9–11, 11
syndrome (SCID), 37–38 Sulfonamide, 9–12, 10–11, 222
SH3 domain of fyn kinase, 190–191, dynasty, 11
S 194 Sulfonylureas, 66
Saccharomycetes cerevisiae, 182 Sickness Impact Profile, 29 Summary Basis of Approval (USA), 254
Safety, 211–225, 242–244 Side effects, 213 Supplementary Protection Certificate
adverse effects, types, 212–213 Signal window, 98 (SPC), 282–283
biopharmaceuticals, 295 ‘Signatures’, biological, 6 Supplementary tests, safety, 213,
chronological phases, 211–212, 212 Sildenafil (Viagra), 50 214–215
clinical studies, 90, 293–294 Simvastatin, 327 Surfactants, 231–237
dose range-finding studies, 215–216, Single ascending dose (SAD) see First Surgical procedures, epilepsy/
216, 217 dose in man (FDIM) epileptogenesis models, 168
future trends, 223–224 Single domain antibodies, 196–197 ‘Surgical shock’, 13
genotoxicity, 216–218 Single molecule detection (SMD) Sustained release, drugs, 235, 236
imaging, 266 technologies, 105 Symptoms, 26
pharmacology, 157, 213–215, Single-dose studies, 290 present, 21–22
214–215, 246, 289–291 Sirolimus, 44–46 therapeutic modalities, 33
special tests, 220–222 Skin, toxicity tests, 217 Synthetic chemistry, 7–8, 9
toxicity measures, 223 SLAM (signalling lymphocyte Systemic exposure, 229
toxicokinetics, 222–223 activation molecule), 194
variability, response, 223 Small interfering RNA (siRNA), 38, 73
T
see also Chronic toxicology studies Small-molecule drugs, 34, 34, 35–36
Sales vs biopharmaceuticals, 180–184 Tacrolimus (FK506, fujimycin), 8,
pattern, 326, 326 discovery, 44 44–46, 66
revenues, 325–327 SmithKline Beecham, 55 Tagamet, exclusivity period, 328, 328
Salmonella typhimurium, 218 Smoking, gene expression, 24–25 Tambocor see Flecainide
Salvarsan (compound number 606), SMON (subacute myelo-optical Target
9–10, 95, 96 neuropathy), 286 audience, product launch, 312–314
Sample dataset, data mining, 79–80 ‘Snake oil’, 303–304 -directed drug discovery, 8–12, 9
Scaffolds, 189–200 Solubility, 228 interaction, imaging, 259, 264–266,
A-domains, 190–191, 193 Solubilizing agent, 228 265
ankyrin repeat (AR) proteins, Somatotropin (human growth occupancy studies, 271
190–191, 195 hormone), 173, 174 validation, imaging, 259, 261–262
kunitz type domains, 190–191, Special populations, 246–247, 285 see also Drug targets
191–192 Species differences, 164, 164, 262, 262 Target Product Profile (TPP), 288
lipocalins, 190–191, 195–196 Spending, 321–325, 322–323, 323 Taxol see Paclitaxel
privileged, 124 Spodoptera frugiperda (Sf9), 182 Technical developments, 203, 205
‘scaffold hunter’ approach, 125 Sporanox, exclusivity period, 328 Technical issues, 58, 60–61
SH3 domain of fyn kinase, 190–191, Stability, 228 Terfenadine, 211
194 Stakeholders Testosterone, 230
single domain antibodies, 196–197 key, 315 Tests
tenth type III domain of fibronectin, values, 315 characteristics, 158
190–191, 192–193 Stapphylococcal protein A (SPA), methods, hierarchy, 158
Z-domain of stapphylococcal protein Z-domain, 190–191, 194–195 selection, 218
A (SPA), 190–191, 194–195 State of the art (prior art), patents, 278 special, 220–222
Scientific issues, 58, 60–61 Statins, 45 Tetracyclines, 8
opportunity, 60 Stokes shift, 103 Thalidomide, 14, 286
Scintillation proximity assays, 102, 102 Strategic issues, 57–59, 58 Theophylline, 45
Scopolamine, 230 Streptococcus mutans, 183–184 Therapeutic studies
Screening process, 95–97, 98 Streptokinase, 173–175 confirmatory, 293
fragment-based, 108–109 Streptomycin, 8 exploratory, 292, 292
libraries, 112–113, 114 Stroke models, 169 Therapeutics
Search tools, 149 Structural classification of natural aims, 21–24
Secondary pharmacodynamic studies, products (SCONP), 125 interventions, 22–23, 26–27
157 Structural-based toxicity, 131 modalities, 33–40, 35–36
Secretome, 82 Structure-activity relationship (SAR), outcomes, 27
Seldane, exclusivity period, 328 96–98 reference pricing, 311

344
 Index

targets, 25–26, 25 Training set, 79 Valproate, 222

‘therapeutic jungle’, 304 Transcriptome, 25, 82, 82, 85–86 Variability, 28, 78, 223
‘Thought leaders’, 310 Transdermal administration, 230 Vasopressin analogues, 230
Thrombin inhibitors, 173–175 Transgenic animals, 73–74 Vasotec, exclusivity period, 328
Tight binders, 128 factories, 73–74 Viagra (sildenafil), 50
‘Tight’ capillary endothelium, 233 founders, 73–74 Videx, exclusivity period, 328
Time resolved fluorescence (TRF), Transplantation Viewlux™, 102
103–104, 104 antibody use, 176 Vinblastine, 8, 45
Time series analysis, 81 tissue/organ, 39 Vinca alkaloids, 45
Timecourse measurements, 70 Transporters, 95 Vincristine, 8, 45
Time-dependent inhibition (TDI), 140, Trapping studies, 149 Vinea alkaloids, 66
141, 149 Trastuzumab (Herceptin), 47, 50 Vioxx, exclusivity period, 328, 328
Timelines, 327–329, 328, 328–329 case history, 46–47, 46, 47 Virchow, Rudolf, 4
Tissue plasminogen activator (tPA), Tricyclic antidepressants, 66, 215 Volume of distribution (V), prediction,
173–175, 174 Trimethoprim, 12 152
Tissues Triptans, 230 Voluntary Exploratory Data
measurement on isolated, 160, Troglitazone, 211 Submissions (VXDS), 91
161–162, 162 Tubocuranne, 8 Voluntary Genomic Data Submissions
therapeutic agents, 35–36 Tumours, 165 (VGDS), 91
transplantation, 39 Von Kekulé, Friedrich August, 5–6
Tolbutamide, 11, 222 Vorinostat, 121, 121
U
Tolerability, 292
Tool (reagent) production, 95–96 Unicorn events, 80
W
Torsade des pointes syndrome, 130, United States of America (USA)
215 marketing, 299 Warfarin, 45
Toxicity, 129–131 regulation, 297 Willful infringement, patents, 276
augmented, 130 University of San Diego (Biology Willingness-to-pay principle, 30
cardiovascular, 130 Workbench), 78 World Health Organization, 20, 286
dose range-finding studies, 215–216, Unmet medical need, 57–58
216, 217 Unplannability of drug development,
dose-related effects, 213
X
203–204
idiosyncratic, 130 Uptake transporter inhibition, 140–141 X-ray crystallography, 125
imaging, 266 Urokinase, 172–175
measures, 223 Ussing Chamber technique, 137–138
prediction, 131 Utility, invention, 278
Y
protein/peptide issues, 184–186 Yeast complementation assay, 107
renal, 130–131
structural-based, 131
V
Toxicokinetics, 222–223 Vaccines, 177
Z
Toxicology DNA, 171 Zantac, exclusivity period, 328, 328
chronic see Chronic toxicology research and development (R&D), Z-domain of stapphylococcal protein A
studies 332 (SPA), 190–191, 194–195
exploratory, 211, 215–216, 216, 217 therapeutic agents, 35–36 Zeta functions ζ, 81
interaction, 222 Vaccinomes, 177 Z’-factor equation, 99–100
regulatory, 211–212, 289–291, 291, Vaginal administration, 230 Ziconotide, 121–122
291 Validation Zidovudine (AZT), 12, 328
reproductive/developmental, of hits, 50–52 89
Zirconium, 264
221–222, 291, 291 target, 259, 261–262 Zollinger–Ellison syndrome, 48–49
Trademarks, 283–284 validity criteria, 166–167, 168 Zoloft, exclusivity period, 328

345