(Advances in Molecular and Cellular Microbiology, 21) Timothy D McHugh-Tuberculosis - Laboratory Diagnosis and Treatment Strategies-CAB International (2013)

Advances in Molecular and Cellular Microbiology 21
Tuberculosis: Laboratory
Diagnosis and Treatment Strategies
Edited by
Timothy D. McHugh
University College London, UK
Advances in Molecular and
Cellular Microbiology
Through the application of molecular and cellular microbiology, we now recognize

the diversity and dominance of microbial life forms on our planet, which exist in all
environments. These microbes have many important planetary roles, but for us
humans a major problem is their ability to colonize our tissues and cause disease. The
same techniques of molecular and cellular microbiology have been applied to the
problems of human and animal infection during the past two decades and have
proved to be immensely powerful tools in elucidating how microorganisms cause
human pathology. This series has the aim of providing information on the advances
that have been made in the application of molecular and cellular microbiology to
specific organisms and the diseases that they cause. The series is edited by researchers
active in the application of molecular and cellular microbiology to human disease
states. Each volume focuses on a particular aspect of infectious disease and will
enable graduate students and researchers to keep up with the rapidly diversifying
literature in current microbiological research.
Series Editor
Professor Michael Wilson

University College London
Titles Available from CABI
17. Helicobacter pylori in the 21st Century

Edited by Philip Sutton and Hazel M. Mitchell
18. Antimicrobial Peptides: Discovery, Design and Novel Therapeutic Strategies

Edited by Guangshun Wang
19. Stress Response in Pathogenic Bacteria

Edited by Stephen P. Kidd
20. Lyme Disease: an Evidence-based Approach

Edited by John J. Halperin
21. Tuberculosis: Laboratory Diagnosis and Treatment Strategies

Edited by Timothy McHugh
22. Antimicrobial Drug Discovery: Emerging Strategies

Edited by George Tegos and Eleftherios Mylonakis
24. Bacteriophages in Health and Disease

Edited by Paul Hyman and Stephen T. Abedon
Titles Forthcoming from CABI
Microbial Metabolomics
Edited by Silas Villas-Bôas and Katya Ruggiero
The Human Microbiota and Microbiome

Edited by Julian Marchesi
Meningitis: Cellular and Molecular Basis

Edited by Myron Christodoulides
Earlier titles in the series are available from Cambridge University Press (www.cup.cam.ac.uk).
CABI is a trading name of CAB International
CABI CABI
Nosworthy Way Chauncey Street
Wallingford Suite 1002
Oxfordshire, OX10 8DE Boston, MA 02111
UK USA
Tel: +44 (0)1491 832111 T: +1 800 552 3083 (toll free)

Fax: +44 (0)1491 833508 T: +1 (0)617 395 4051
E-mail: [email protected] E-mail: [email protected]
Website: www.cabi.org
© CAB International 2013. All rights reserved. No part of this publication

may be reproduced in any form or by any means, electronically, mechanically,
by photocopying, recording or otherwise, without the prior permission of the
copyright owners.
A catalogue record for this book is available from the British Library, London, UK.
Library of Congress Cataloging-in-Publication Data
Tuberculosis : laboratory diagnosis and treatment strategies / edited by

Timothy D. McHugh.
p. ; cm. -- (Advances in molecular and cellular microbiology ; 21)
Includes bibliographical references and index.
ISBN 978-1-84593-807-9 (alk. paper)
I. McHugh, Timothy D. (Timothy Daniel) II. Series: Advances in molecular
and cellular microbiology ; 21.
[DNLM: 1. Tuberculosis--diagnosis. 2. Antitubercular Agents. 3. Drug Discovery.
4. Molecular Diagnostic Techniques. WF 220]
616.99’5075--dc23
2012030958
ISBN-13: 978 1 84593 807 9
Commissioning editor: Rachel Cutts

Editorial assistant: Alexandra Lainsbury
Production editor: Simon Hill
Typeset by Columns Design XML Ltd, Reading, UK.

Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY.
Contents
Contributors vii
Abbreviations ix
Introduction xi
Timothy McHugh
Part I – Diagnosis 1
1. Improving on Sputum Smear Microscopy for Diagnosis of Tuberculosis in

Resource-poor Settings 3
Andrew Ramsay, Karen R. Steingart and Madhukar Pai
2. Molecular Diagnosis of Active Pulmonary Tuberculosis 13

Anna L. Bateson, Kate Reddington and Justin O’Grady
3. Improving on the LJ slope – Automated Liquid Culture 34

Miguel Viveiros, Diana Machado, Isabel Couto and Leonard Amaral
4. Interferon-gamma Release Assays in the Diagnosis of Latent Tuberculosis

Infection 46
Ajit Lalvani and Manish Pareek
5. Measuring Tuberculosis Immune Responses in the Lung – The Correct Target? 67

Ronan Breen and Marc Lipman
6. Sniffing Out Tuberculosis 81

Ruth McNerney and Claire Turner
Part II – Measuring Resistance 93
7. Role of Phenotypic Methods for Drug Susceptibility Testing of M. tuberculosis

Isolates in the Era of MDR and XDR Epidemics 95
Leonid Heifets
8. Genotypic Measures of Antibiotic Susceptibility 108

Vanessa Mathys, Barun Mathema and Pablo Bifani
v
vi Contents
9. Molecular Tools for Fast Identification of Resistance and Characterization

of MDR/XDR-TB 125
Simeon Cadmus and Dick van Soolingen
Part III – Understanding Treatment 137
10. Monitoring Therapy by Bacterial Load 139

Denise M. O’Sullivan
11. Modelling Responses to Tuberculosis Treatment 147

G.R. Davies
12. Measuring Gene Expression by Quantitative PCR (qPCR) 164

Isobella Honeyborne
13. Transcriptomic Approaches to Mapping Responses to Drug Therapy for

Tuberculosis 173
Simon J. Waddell and Philip D. Butcher
14. Mycobacterial Lipid Bodies and Resuscitation-promoting Factor

Dependency as Potential Biomarkers of Response to Chemotherapy 184
N.J. Garton, G.V. Mukamolova and M.R. Barer
Part IV – Treatment Strategies 195
15. Clinical Trials in Tuberculosis Chemotherapy: The Challenges 197

Stephen H. Gillespie
16. The Identification of 2-Aminothiazole-4-carboxylates (ATCs) as a

New Class of Tuberculosis Agent: A Lesson in ‘HIT’ Identification 206
Geoff Coxon
17. Rifamycins Revisited 219

Martin J. Boeree
18. Therapy of the XDR-TB Patient with Thioridazine – An Old Drug with New
Applications 232
Leonard Amaral, Marta Martins , Isabel Couto and Miguel Viveiros
19. Vaccines for Tuberculosis 243

Helen McShane
Index 261
Contributors
Leonard Amaral, Unit of Mycobacteriology, Instituto de Higiene e Medicina Tropical,

Universidade Nova de Lisboa, Rua da Junqueira, 100, 1349-008 Lisbon, Portugal. E-mail:
[email protected]
M.R. Barer, Department of Infection, Immunity and Inflammation, University of Leicester
College of Medicine, Biological Sciences and Psychology, Maurice Shock Building,
Leicester, UK. E-mail: [email protected]
Anna L. Bateson, Centre for Clinical Microbiology, Department of Infection, Royal Free
Campus, University College London, Rowland Hill Street, London NW3 2PF, UK. E-mail:
[email protected]
Pablo Bifani, The Novartis Institute for Tropical Diseases, Singapore. E-mail: pablo.bifani@
novartis.com
Martin J. Boeree, Associate Professor, Radboud University Nĳmegen Medical Centre,
Nĳmegen, The Netherlands. E-mail: [email protected]
Ronan Breen, Department of Respiratory Medicine, Guys’ & St Thomas’ NHS Trust,
Westminster Bridge Road, London SE1 7EH, UK. E-mail: [email protected]
Philip D. Butcher, Centre for Infection and Immunity, Division of Clinical Sciences, St George’s
University of London, London SW17 0RE, UK. E-mail: [email protected]
Simeon Cadmus, Tuberculosis Research Laboratory, Department of Veterinary Public Health
and Preventive Medicine, University of Ibadan, Nigeria.
Isabel Couto, Unit of Mycobacteriology, Instituto de Higiene e Medicina Tropical, Universidade
Nova de Lisboa, Rua da Junqueira, 100, 1349-008 Lisbon, Portugal.
Geoff Coxon, Strathclyde Institute of Pharmacy and Biomedical Sciences, John Arbuthnott
Building, University of Strathclyde, 27 Taylor Street, Glasgow G4 0NR, UK. E-mail: geoff.
[email protected]
G.R. Davies, Senior Lecturer in Infectious Diseases, Institute of Translational Medicine,
University of Liverpool, Liverpool, UK. E-mail: [email protected]
N.J. Garton, Department of Infection, Immunity and Inflammation, University of Leicester
College of Medicine, Biological Sciences and Psychology, Maurice Shock Building,
Stephen H. Gillespie, The Bute Medical School, St Andrews University, North Haugh, St
Andrews KY16 9TF, UK. E-mail: [email protected]
vii
viii Contributors
Leonid Heifets, Mycobacteriology Reference Laboratory, National Jewish Health, Denver,

Colorado, USA. E-mail: [email protected]
Isobella Honeyborne, Centre for Clinical Microbiology, Department of Infection, Royal Free
Campus, University College London, Rowland Hill Street, London NW3 2PF, UK. E-mail:
[email protected]
Ajit Lalvani, Tuberculosis Research Unit, Department of Respiratory Medicine, National
Heart and Lung Institute, Imperial College London, London, UK. E-mail: a.lalvani@
imperial.ac.uk
Marc Lipman, Centre for Respiratory Medicine, Royal Free Campus, University College
London, Rowland Hill Street, London NW3 2PF, UK. E-mail: [email protected]
Diana Machado, Unit of Mycobacteriology, Instituto de Higiene e Medicina Tropical,
Universidade Nova de Lisboa, Rua da Junqueira, 100, 1349-008 Lisbon, Portugal.
Ruth McNerney, Department of Pathogen Molecular Biology, London School of Hygiene and
Tropical Medicine, London, UK. E-mail: [email protected]
Helen McShane, Professor of Vaccinology and Wellcome Senior Fellow, The Jenner Institute,
University of Oxford, Old Road Campus Research Building, Roosevelt Drive, Oxford OX3
7DQ, UK. E-mail: [email protected]
Marta Martins, Unit of Mycobacteriology and UPMM, Instituto de Higiene e Medicina
Tropical, Universidade Nova de Lisboa, Rua da Junqueira, 100, 1349-008 Lisbon, Portugal.
Barun Mathema, Tuberculosis Center, Public Health Research Institute, University of Medicine
and Dentistry, Newark, New Jersey, USA.
Vanessa Mathys, Tuberculosis and Mycobacteria, Communicable and Infectious Diseases,
Scientific Institute of Public Health, Belgium.
G.V. Mukamolova, Department of Infection, Immunity and Inflammation, University of
Leicester College of Medicine, Biological Sciences and Psychology, Maurice Shock
Building, Leicester, UK. E-mail: [email protected]
Justin O’Grady, Norwich Medical School, NRP Innovation Centre, Norwich Research Park,
Colney Lane, Norwich, NR4 7GJ, UK. E-mail: [email protected]
Denise M. O’Sullivan, Molecular and Cell Biology Team, LGC, Queens Road, Teddington,
Middlesex TW11 0LY, UK. E-mail: [email protected]
Madhukar Pai, Department of Epidemiology, Biostatistics and Occupational Health, McGill
University, Montreal, Canada. E-mail: [email protected]
Manish Pareek, Department of Infection, Immunity and Inflammation, University of Leicester,
Andrew Ramsay, UNICEF/UNDP/World Bank/WHO Special Programme for Research and
Training in Tropical Diseases (TDR), World Health Organization, Geneva, Switzerland.
E-mail: [email protected]
Kate Reddington, Molecular Diagnostics Research Group, Microbiology, School of Natural
Sciences, National University of Ireland Galway, Galway, Ireland. E-mail: k.reddington1@
nuigalway.ie
Karen R. Steingart, University of Washington School of Public Health, Seattle, USA.
Claire Turner, Department of Life, Health and Chemical Sciences, The Open University, Milton
Keynes, UK.
Dick van Soolingen, Mycobacteria Reference Laboratory, National Institute of Public Health
and the Environment (RIVM), The Netherlands. E-mail: [email protected]
Miguel Viveiros, Unit of Mycobacteriology, Instituto de Higiene e Medicina Tropical,
Universidade Nova de Lisboa, Rua da Junqueira, 100, 1349-008 Lisbon, Portugal.
Simon J. Waddell, Brighton and Sussex Medical School, University of Sussex, Brighton BN1
9PX, UK. E-mail: [email protected]
Abbreviations
AFB Acid-fast bacilli

Anti-TNF Anti-tumour necrosis factor
BAL Bronchoalveolar lavage
BCG Bacillus Calmette–Guérin
CFP-10 Culture filtrate protein-10
CFU Colony-forming units
DOTS Direct observed therapy short course
DST Drug susceptibility testing
EBA Early bactericidal activity
ELISA Enzyme-linked immunosorbent assay
ELISpot Enzyme-linked immunosorbent spot
ESAT-6 Early secretory antigen target-6
GC-MS Gas chromatography–mass spectrometry
GTC Guanidine thiocyanate
GUs Growth units
IFN-γ Interferon-gamma
IGRAs Interferon-gamma release assays
IL-12 Interleukin-12
IMIDs Immune-mediated inflammatory diseases
LAMP Loop-mediated isothermal amplification
LBs Lipid bodies
LEDs Light-emitting diodes
LJ Löwenstein–Jensen
LMIC Low- and middle-income country
LPAs Line probe assays
LTBI Latent tuberculosis infection
MDR-TB Multidrug-resistant tuberculosis
MGIT Mycobacterial growth indicator tube
MIC Minimum inhibitory concentration
MPN Most probable number
MTC Mycobacterium tuberculosis complex
NAAT Nucleic acid amplification assay
NASBA Nucleic acid sequence-based amplification
NO Nitric oxide
NTM Non-tuberculous mycobacteria
ix
x Abbreviations
PBMCs Peripheral blood mononuclear cells

PD Pharmacodynamic
PK Pharmacokinetic
POC Point of care
PPD Purified protein derivative
qPCR Quantitative PCR
RD Region of difference
Rpf Resuscitation-promoting factor
RR Resistance ratio method
RRDR Rifampicin resistance-determining region
RT-PCR Reverse transcription polymerase chain reaction
SARs Structure–activity relationships
SDA Strand displacement amplification
SSCC Sputum serial colony counting
TAG Triacylglycerol
TDR-TB Totally drug-resistant tuberculosis
Th1 T-helper cell type 1
TLRs Toll-like receptors
TMA Transcription-mediated amplification
TNF-α Tumour necrosis factor alpha
TST Tuberculin skin test
TTP Time to positivity
VOC Volatile organic compounds
WE Wax ester
XDR-TB Extremely drug-resistant tuberculosis
ZN Ziehl–Neelsen stain
DRUGS
Amikacin AK
Aminoglycocide AG
Aminothiazole carboxylate ATC
Benzothiazinone BTZ
Capreomycin CM or CAP
Chlorpromazine CPZ
Ciprofloxacin CIP
Clofazimine CF
Cycloserine CS
Dithiothreitol DTT
Ethambutol EMB or E
Ethionamide ETA
Fluoroquinolone FQ
Gatifloxacin GAT
Isoniazid INH or H
Kanamycin KM
Levofloxacin LEVO
Moxifloxacin MOX or M
Ofloxacin OFLOX
Para-amino-salicylic acid PAS
Pyrazinamide PZA or Z
Rifabutin RBT
Rifampicin (rifampin) RIF, RMP or R
Rifapentine RPT
Sparfloxacin SPA
Streptomycin SM or S
Thioridazine TZ
Introduction
Tuberculosis is a disease of poverty affecting adults in their most productive years. In 2010,
there were 8.8 million incident cases with 1.1 million tuberculosis deaths in patients not
infected with HIV. A further 0.35 million deaths were associated with HIV-positive patients
(WHO, 2011). These oft-quoted figures demonstrate the challenge that tuberculosis control
presents and this book highlights the state of the art in the mainstays of tuberculosis control:
laboratory diagnosis and treatment.
Tuberculosis control is a multifaceted problem, but positive identification of the tubercle
bacillus and assessment of its drug susceptibility are essential components in the strategy. It is
telling that in 2010, 8 out of 22 high-burden countries did not meet the benchmark of 1
microscopy unit per 100,000 population. Furthermore, in a combined list of 36 countries that
were high-burden countries and/or high MDR-TB burden countries, 20 had less than the
benchmark of one laboratory capable of culture and drug sensitivity testing per 5 million
population (WHO, 2011). It is hoped that the rational approach to conventional testing and
roll-out of novel technologies described here will serve to provide cost-effective and robust
technologies that will improve delivery of diagnostic services in resource-poor settings.
For many years, there was stasis in the development of drugs for the treatment of
tuberculosis. The HIV pandemic and, more recently, the development of MDR- and XDR-TB
(and indeed now TDR-TB) have changed our perspective. New drugs are being developed and
our approaches to evaluating them require review. The spectrum of this activity is outlined in
chapters covering the fundamental biology of drug action (Chapters 12 and 13), as well those
focused on clinical trial methodology (Chapters 10, 11 and 15).
I have included three examples of approaches to new drug development; the ATC
compounds (Chapter 16) represent a design-based approach in which the ATC scaffold is
designed to inhibit the fatty acid synthase pathway. By comparison, Chapters 17 and 18 discuss
repurposing established drugs: in Chapter 17 the rifamycins are revisited and Chapter 18
discusses how the imperative of MDR-TB has driven the reconsideration of thioridazine. Of
course, matching these developments are exciting developments in vaccine design and these
are captured in Chapter 19.
xi
xii Introduction
In the past 10 years, the pace of research into new diagnostic tools and treatments has
increased, there is a new enthusiasm for tuberculosis research and while this book captures the
zeitgeist, it also aims to provide underlying information for future developments.
Tim McHugh
UCL
Reference
WHO (2011) Global Tuberculosis Report 2011.
Part I
Diagnosis
This page intentionally left blank
1
Improving on Sputum Smear
Microscopy for Diagnosis of Tuberculosis
in Resource-poor Settings
Andrew Ramsay,1 Karen R. Steingart2 and Madhukar Pai3

1UNICEF/UNDP/World Bank/WHO Special Programme for Research and Training in
Tropical Diseases (TDR), World Health Organization, Geneva, Switzerland;
2University of Washington School of Public Health, Seattle, USA; 3Department
of Epidemiology, Biostatistics and Occupational Health, McGill University,

Montreal, Canada
1.1 Introduction ground (WHO, 2011a). This chapter will

discuss the limitations of direct sputum
In 2011, despite the long-established tradition smear microscopy as practised currently,
of solid culture for Mycobacterium tuberculosis review the recent WHO recommendations
and the numerous more recent advances in and discuss how the performance of sputum
tuberculosis diagnosis, direct sputum smear smear microscopy may be improved, explore
microscopy remains the cornerstone of global how sputum smear microscopy services may
tuberculosis control. However, direct sputum be complemented by add-on tests and
smear microscopy has major limitations. It is identify which tests (now and in the future)
associated with low and variable sensitivity, may replace smear microscopy in certain
particularly in HIV-associated tuberculosis. It defined situations.
gives no information regarding the drug
resistance of the infecting tuberculosis bacilli.
It is, however, to date, the only diagnostic 1.2 Direct Sputum Smear Microscopy
method for tuberculosis that has been shown
to be implementable and sustainable in Direct sputum smear microscopy is inherently
resource-poor health-care facilities in low- insensitive since the typical acid-fast bacilli
and middle-income countries (LMICs). The (AFB) can only be detected when large
dire situation is well recognized by national numbers of them are present in the sputum
tuberculosis programmes, technical agencies (>105 organisms/ml). This impacts particularly
and donors. The World Health Organization on the diagnosis of HIV-associated tubercu-
(WHO), in an effort to improve diagnosis, has losis, which is often paucibacillary, as well as
made more than 15 global recommendations the diagnosis of tuberculosis in children, who
related to tuberculosis diagnosis in the past 4 are often unable to produce a sputum
years, and yet little has changed on the specimen. Moreover, well-trained and dedi-
© CAB International 2013. Tuberculosis: Laboratory Diagnosis and Treatment Strategies

(ed. T.D. McHugh) 3
4 A. Ramsay et al.
cated microscopists are needed to ensure a tuberculosis bacilli detected in the sputum
good quality service. Ziehl–Neelsen (ZN) smear of new tuberculosis patients and
stained smears require examination under therefore cannot diagnose multidrug-
the microscope for 5–10 min before being resistant tuberculosis (MDR-TB) or extremely
declared negative (IUATLD, 2000). To counter drug-resistant tuberculosis (XDR-TB).
the inherent lack of sensitivity of the method, In addition to the diagnosis of new cases,
the WHO recommended that three sputum direct sputum smear microscopy repeated at
specimens (including an early morning different time points following initiation of
specimen) be collected for examination from anti-tuberculosis treatment is used to monitor
each suspected tuberculosis patient over 2 patients’ response (Gilpin et al., 2007). Con-
days. This clearly has implications for the tinued presence of AFB in sputum may
laboratory workload. A microscopist can only indicate treatment failure, but this may be
examine some 20–25 smears/day (IUATLD, due to causes other than drug resistance,
2000). Smear microscopy workloads heavier such as failure to take medication as indicated.
than that impact adversely on service quality. The recommended system for external
The requirement for three sputum specimens quality assessment (EQA) of smear
also places a burden on the suspected microscopy services requires the blinded
tuberculosis patient, since they have to make rechecking of a sample of the slides examined
repeated visits to the health-care facility to by each laboratory. Usually, the national
submit specimens, receive results and start tuberculosis reference laboratory is expected
any treatment indicated. This is particularly to take on this monumental task. It is
burdensome if the patient is poor, weak or recommended that blinded rechecking is
has travelled a long distance to the health- complemented by regular visits to microscopy
care facility. Patient dropout during the laboratories to ensure a proper staining
diagnostic process is common (Kemp, 2007; technique is used and to check the condition
Botha, 2008). of microscopes (Association of Public Health
The three-specimen, multi-day standard Laboratories, 2002). Microscopes can be
approach above, which has been devised to damaged or impaired by rough handling. In
improve the sensitivity of sputum smear hot, humid climates, the lenses of microscopes
microscopy, also includes, paradoxically, a are prone to colonization by algae or mould,
requirement that reduces the sensitivity of which can render them worse than useless. A
the technique further. A patient must have fully functional national EQA scheme based
two positive smears to be defined as a smear- on this system is rarely achieved or sustained.
positive tuberculosis case, and one of these Nevertheless, most district hospitals and
smears must contain at least 10 bacilli/100 large health centres in countries with a high
high-power microscopic fields (HPF). These prevalence of tuberculosis can offer a sputum
last requirements were introduced to smear microscopy service of some kind. It is
maximize the specificity of direct sputum estimated that more than 50 million smear
smear microscopy. The requirement dates microscopy examinations are conducted
back to the times when treatment regimens globally each year (TDR/FIND, 2006). The
were longer and rifampicin was very only equipment required is a microscope
expensive and was likely a measure to reduce (relatively inexpensive as laboratory equip-
costs and inconvenience to patients resulting ment goes) and a gas, or spirit, burner.
from unnecessary treatment. Today, first-line Microscopes generally need an electricity
treatment for tuberculosis is shorter, but supply, but they can accommodate some
6-month regimens are still a considerable fluctuation in power without damage.
undertaking for patients. The cost of all the Microscopes may also be powered off car
drugs for the standard 6-month first-line batteries or solar cells or, with a mirror and
regimen is now approximately US$10–30 some experience, can be operated using
(Management Sciences for Health, 2005). directed daylight. The only consumables and
Sputum smear microscopy gives no reagents required are low cost and include
information on the drug susceptibility of specimen pots, applicator sticks for making
Diagnosis of Tuberculosis in Resource-poor Settings 5
smears, glass slides, a simple set of heat- two specimens decreased the number of
stable stains with very long shelf lives and sputum smear-positive cases defined as
small amounts of microscope immersion oil having two positive smears (≥10 AFB/100
to use with the 100 oil objective lens. HPF). A service that examined only two
Training of staff is straightforward and specimens might be more efficient and less
relatively short and, importantly, is a well- onerous for laboratory staff. However, such a
established module in laboratory assistant policy decision would require reassessment
and laboratory technician training courses of the standard definition of a smear-positive
and given importance in all countries where case. If both specimens could be collected
tuberculosis is a public health problem. (and preferably examined) on the same day,
this could reduce the number of patient visits
required, reduce patient expenditure and
1.3 Optimized Sputum Smear possibly reduce default during the diagnostic
Microscopy process.
In the absence of simple, low-cost alternatives

to sputum smear microscopy much effort has 1.3.2 Fluorescence microscopy
been expended in recent years in developing (Steingart, 2006a, 2007a)
optimized smear microscopy procedures.
Much of this effort was catalysed by a series Fluorescence microscopy (FM) was, on
of systematic reviews of smear microscopy average, 10% more sensitive than ZN micro-
conducted in 2005–2006 and which led to the scopy, with similar (98%) specificity. FM was
development of a prioritized international also much quicker than ZN microscopy, with
research agenda and to five global policy FM examination for 1 min being associated
changes by the WHO in 2007–2009 (Steingart, with a higher sensitivity and similar
2006a,b, 2007a; Mase, 2007; WHO, 2007b,c, specificity to ZN examination for 4 min. The
2010b,c). systematic review was conducted prior to the
These systematic reviews focused on recent work on FM based on light-emitting
methods of improving sputum microscopy diodes (LEDs). Conventional FM was
services in LMICs through more efficient and dependent on very expensive, sophisticated
patient-friendly specimen collection, the use equipment considered inappropriate for use
of fluorescence microscopy and the chemical outside of high throughput reference labora-
digestion of sputum followed by a physical tories in LMICs. However, the authors noted
sputum concentration method. that simple, low-cost LED-FM systems were
The main findings of the systematic being developed and should be evaluated in
reviews were as follows. resource-poor settings (Gilpin, 2007).
1.3.1 Serial sputum smear examination 1.3.3 Specimen processing (Steingart,

(Mase, 2007; Steingart, 2007a) 2006b, 2007a)
The average increase in sensitivity/incre- The systematic review indicated that, in

mental yield of tuberculosis cases through comparison with direct smears, centrifugation
examination of the third specimen was only preceded by digestion with any one of several
2–5%. Thus, evaluating 1000 persons sus- different chemicals (including household
pected of having tuberculosis in a setting bleach) was more sensitive. The review also
with 10% prevalence of the disease would revealed that overnight sedimentation
require examination of 900 third specimens to preceded by chemical processing (including
yield an additional two–five cases. However, with household bleach) was more sensitive
the analysis defined a smear-positive case as and specificity was similar. The latter method,
one with only one or more positive smears since it did not require an expensive
(≥3 AFB/100 HPF). The examination of only centrifuge, was considered to be more
6 A. Ramsay et al.
appropriate to basic laboratories in resource- on how to produce good quality sputum

poor settings. specimens increased significantly the number
A 5-year long, multi-partner, inter- of smear-positive cases detected among
nationally coordinated programme of research women (Khan, 2007).
followed, which provided evidence that Thus, an optimized sputum smear
(Ramsay, 2011): microscopy service would incorporate simple
standardized instruction on sputum speci-
1. ‘Scanty’ smears (i.e. with less than 10
men production, collection (and examination)
AFB/100 HPF) were ‘true positive’ smears
of two specimens on the same day, use of
and considering these as positives increased
LED-FM and new definitions of a positive
the sensitivity of sputum smear microscopy
smear and a smear-positive case. Wherever
without reducing specificity.
possible, smear results should be reported on
2. A single positive smear (including scanty
the same day and treatment initiated to
smears) was sufficient to confirm a patient as
reduce initial default.
a smear-positive tuberculosis case and
increased the sensitivity of sputum smear
microscopy without reducing specificity.
1.4 Serological Tests as Add-on or
3. Reducing the number of specimens exam-
Replacement Tests
ined from three to two resulted in more effi-
cient detection of cases, with only marginal
A series of systematic reviews have evaluated
reductions in sensitivity (2–5%).
the performance of various serological
4. Examining two specimens collected 1 h
(antibody detection) tests for the diagnosis of
apart on the same day was equivalent to
tuberculosis (Steingart, 2007b,c,d, 2009,
examination of two or three specimens
2011a; WHO, 2008). These tests used formats
collected over 2 days (Ramsay, 2009; Cuevas,
that required laboratory infrastructure such
2012a).
as enzyme-linked immunosorbent assays
5. Sensitivity and specificity of LED-FM was
(ELISAs), as well as rapid immuno-
comparable to that of conventional FM.21,22
chromatographic test platforms. The latter
The use of LED-FM was more sensitive than
tests, if they could perform as well as smear
ZN microscopy, and if sufficient attention
microscopy, potentially could replace it,
was paid to proper training of microscopists,
making diagnosis simpler and quicker.
it had similar specificity (including when
Moreover, the tests, being simple and robust,
used with a same-day specimen collection
could be performed by non-specialized staff
approach) (Cuevas, 2012b).
in health centres without laboratories and
6. Bleach digestion followed by overnight
closer to the community. If performance did
sedimentation led to only small increases in
not compare with smear microscopy, there
sensitivity (9%), with associated small
was the possibility that their use might
decreases in specificity (–3%).
complement smear microscopy and sero-
7. LED-FM combined with bleach digestion
logical testing of smear-negative patients
and overnight sedimentation had the same
might result in an incremental gain in
sensitivity and slightly lower specificity than
tuberculosis cases detected. A laboratory-
direct LED-FM alone (Bonnet, 2011).
based head-to-head evaluation of 19 com-
Between 2007 and 2010, the WHO endorsed a mercially available rapid antibody detection
new definition of a positive smear, a new tests for tuberculosis was included in the
definition of a smear-positive case, a re- systematic review (WHO, 2008a).
duction in the number of sputum specimens The systematic reviews and meta-
to be examined in the investigation of analyses concluded that diagnostic accuracy
pulmonary tuberculosis, same-day micro- data indicated that none of the tests could
scopy and direct LED-FM (Ramsay, 2011). replace sputum smear microscopy, nor could
A pragmatic randomized trial in Pakistan they be used as an add-on test to complement
showed that simple standardized instructions sputum smear microscopy (Steingart, 2011).
The wide availability of these tests in sophisticated biosafety laboratories. They are
India, particularly in the private health sector, 20–25% more sensitive than solid culture and
has recently been reported (Grenier, 2012; the time to positivity is reduced by an average
Steingart, 2012). Modelling of data from these of about 1 week (Chihota, 2010). Their appli-
studies has indicated enormous economic, cation in tuberculosis reference laboratories
social and public health costs associated with in LMICs has frequently been problematic,
the use of these inadequate tests. Most of and extensive international technical exper-
these tests are produced by companies based tise has been required for implementation
in Western countries (such as France, the UK, (Muyoyeta, 2009). Unacceptably high culture
Germany and the USA), where these tests are contamination rates have proved a common
not approved for use but the companies and difficult problem (Peres, 2011).
market and sell them in LMICs with high Even when implemented satisfactorily in
tuberculosis prevalence rates and poor the laboratory, the impact of tuberculosis
regulatory oversight of imported diagnostics culture on patient management is frequently
(Steingart, 2012). A very recent study has uncertain (Stall, 2011). Mathematical model-
confirmed that these tests are widely used in ling has indicated the key importance of
the private sector of most of the 22 countries patient access to diagnostics in determining
with the highest burdens of tuberculosis their public health impact (Keeler, 2006).
(high-burden countries). In about one-third Diagnostic testing in central or national
of high-burden countries, the tests were also reference laboratories far from the health
widely used in the public sector (Grenier, facilities where patients present is unlikely to
2012). be associated with any major public health
Based on expert consideration of the impact unless equitable services can be
evidence presented in the systematic reviews, developed with well-organized specimen
the WHO released a policy statement in 2011 transport, rapid turnaround times in the
actively discouraging the use of all com- laboratory and prompt, efficient com-
mercially available serological tests for munication of results. Such systems are
tuberculosis, whether as a smear microscopy difficult to build and sustain in LMICs, even
replacement test or as an add-on test (WHO, when funding is available. Experience to date
2011b). has largely confirmed these predictions
(Harries, 2004).
Several other culture-based tests have
1.5 WHO-endorsed Replacement or been endorsed by the WHO in recent years;
Add-on Tests (Culture Based) these include the microscopic observation
drug susceptibility (MODS) assay, redox
The WHO has endorsed only one culture- indicator colorimetric drug susceptibility tests
based test, liquid culture, for tuberculosis and the nitrate reductase (Greiss) test (WHO,
diagnosis. The term ‘liquid culture’ can be 2011c). However, these alternative tests have
confusing since a number of culture-based only been endorsed, as interim measures, for
tests that have not been endorsed by WHO the direct and/or indirect screening for drug
could reasonably be described as ‘liquid resistance and until capacity for commercially
culture’. The WHO endorsement, however, available broth-based culture has been built.
refers specifically to commercially available The alternative tests, like the commercially
broth-based culture systems, and most of the available broth-based culture, also require
evidence which the WHO reviewed prior to sophisticated biosafe laboratories and are
the endorsement was specific to the subject to the same concerns about a lack of
Mycobacterial Growth Indicator Tube (MGIT) major public health impact.
system produced by Becton Dickinson (USA) The recent development of mobile
(WHO, 2007a). These commercially available shipping container based culture laboratories
broth-based culture systems can be manual may help to improve patient access to these
or automated. They are expensive and require tests (WHO, 2009).
8 A. Ramsay et al.
1.6 WHO-endorsed Replacement or culture-based tests and line probe assays do.
Add-on Tests (Nucleic Acid It is simple and results can be available within
Amplification Tests) 2 h. Recent studies confirm earlier reports of
good specificity and a sensitivity (using one
To date, the WHO has endorsed two nucleic cartridge) greater than smear microscopy but
acid amplification tests for tuberculosis. The less than liquid culture (Banada, 2009;
first, endorsed in 2008, was the line probe Blakemore, 2010; Boehme 2010, 2011).
assay (Hains test, or Inno-LiPa) (WHO, Sensitivity increases and approaches that of
2008b). This was endorsed for rapid screening liquid culture if three specimens (and three
of patients at risk of MDR-TB. These line cartridges) are used. One of the major
probe assays, which detect rifampicin drawbacks of this technology is its cost. Even
resistance (Inno-LiPa) or rifampicin and at prices specially negotiated with the
isoniazid resistance, are endorsed for use company for the public sector in high
indirectly on isolates obtained from tubercu- tuberculosis burden countries, each cartridge
losis cultures or directly on smear-positive costs at least US$15. Each four-berth
patients only. Clearly, then, sputum smear instrument, which can process only 20
microscopy (particularly the more sensitive cartridges in an average working day,
optimized sputum smear microscopy) could currently costs US$15–17,000 at the specially
be combined with line probe assays for MDR negotiated price. The instruments are very
screening in line with WHO recommenda- sensitive to fluctuations in power supply and
tions. must be protected electrically with an
If line probe assays are to be performed uninterruptible power supply (UPS). At
on culture isolates, the procedures need to be present, each instrument needs to be sent to
conducted in the same standard of biosafe Europe once a year for recalibration. The
laboratory required for the handling of simplicity of this technology undoubtedly
tuberculosis culture systems. If, however, the could bring more sensitive tuberculosis
sputum from smear-positive patients will not diagnosis and MDR-tuberculosis screening
be cultured but processed directly for line closer to where patients are seeking medical
probe assay, the biosafety requirements may care and reduce time to diagnosis and to
be less stringent (WHO, 2010d). It may be appropriate treatment. It is likely that this test
possible, therefore, to conduct the line probe could be used in health centres in some
assay testing at intermediate laboratories middle-income countries and perhaps at
closer to the sputum smear microscopy district hospital level in parts of sub-Saharan
centre. The closer proximity of the advanced Africa. The WHO currently recommends that
secondary testing site (for drug resistance) to the MTB/RIF is used as the initial diagnostic
the site of the primary smear-based diagnosis test (i.e. smear microscopy replacement) in
may facilitate better interaction and promote areas with high rates of MDR-TB or high rates
a more rapid patient-centred service. of HIV-associated tuberculosis. The WHO
Unfortunately, in most cases, the implementa- further suggests that in areas where MDR-TB
tion of line probe assays has not followed this or HIV-associated tuberculosis is of less
model and secondary testing is usually importance, the MTB/RIF is used as an
confined to reference laboratories, sometimes add-on test (following smear microscopy) to
considerable distances from the site of detect smear-negative tuberculosis (WHO,
primary screening. The issues around access 2011d). Cost-effectiveness models have been
that applied to the culture-based systems developed around these two applications
would also apply here. (replacement and add-on) (WHO, 2011e). The
In 2010, the WHO endorsed the problem with the use of the MTB/RIF on
GeneXpert MTB/RIF, a cartridge-based real- smear negatives only is that smear-positive
time PCR for the simultaneous diagnosis of patients will be systematically denied a rapid
tuberculosis and detection of rifampicin rifampicin resistance test. Concerns have also
resistance (WHO, 2011d). This closed system been raised about false positive rifampicin-
does not require biosafety laboratories, as the resistant results with the test and the
associated low positive predictive value, even 2010). Very few have looked at patient-
in areas with relatively high MDR-TB important outcomes. Nearly all studies, to
prevalence (Lawn, 2011). The most pressing date, have evaluated a single diagnostic test
issue, of course, is the cost of the technology. rather than a real-world diagnostic service
Over 50 million smear microscopy investi- based on multiple tests at different health
gations are conducted annually around the service levels (Squire, 2011). It is recognized
world. Even if the MTB/RIF is to be an add-on that the current evidence on which the WHO
test for smear negatives, the volume of testing is endorsing new diagnostic tools is sufficient
will be immense, and very costly. only for a ‘technical endorsement’ – that this
tool performs with this diagnostic accuracy
when used in ideal conditions and may be
1.7 The Current Tuberculosis useful in tuberculosis control programme
Diagnostics Pipeline activities (Ramsay, 2010). More research and
more evidence is required on the performance
There is little new expected from tuberculosis of these tests when used in routine program-
diagnostics in the coming 5 years. The LAMP matic settings in combination with other
assay has been proposed as a nucleic acid diagnostic tests. This would allow the
amplification test suitable for the peripheral development of a process for the ‘program-
laboratory. Demonstration studies have been matic endorsement’ of a new tool, which
ongoing for several years. would be accompanied by evidence-based
Progress is being made in our under- guidance on how to implement the test and
standing of humoral immune responses to scale-up use (Wilson, 2011).
M. tuberculosis in different stages of the
disease and the importance of multiple anti-
gen testing and immunoresponse signatures
1.9 Conclusion
in interpreting test results (Steingart, 2009;
Kumnath-Velayudhan, 2010).4 These may
Despite exciting advances in diagnostic
result in serological tests that would fit a
technologies for tuberculosis and the creation
point-of-care (POC) test format. Technological
of a tuberculosis diagnostic toolbox of several
platforms suitable for POC application are
WHO-endorsed diagnostic technologies and
currently being revolutionized through the
approaches, global tuberculosis control for
use of microfluidics and nanoparticles (Chin,
the time being is reliant on sputum smear
2011). A number of groups are working on
microscopy. It is important that these sputum
breath analysers that detect volatile organics
smear microscopy services are optimized as
associated with tuberculosis disease in
described above to:
sampled breath (Syhre, 2009; McNerney,
2010; Phillips, 2010). • provide improved diagnostic services to
patients in LMICs while the MTB/RIF and
other diagnostic tools are rolled out. This
1.8 The Future roll-out may take several years before
adequate service coverage is achieved;
It is unlikely that tuberculosis control • provide enhanced performance of the
activities in resource-poor settings will come diagnostic service in areas where new
to rely on a single new test in the near future. diagnostics such as MTB/RIF are being
It is more probable that diagnostic services considered as an add-on test to sputum
will be dependent on multiple diagnostic smear microscopy;
technologies with different roles at different • provide the best possible baseline data on
levels of the tiered health service structure which to judge the performance and
(Ramsay, 2009). Test evaluations to date have added value of new technologies as they
been largely diagnostic accuracy studies (Pai, are introduced.
10 A. Ramsay et al.
References Cuevas, L.E., Al-Sonboli, N., Lawson, L., Yassin,

M.A., Arbide, I., Al-Aghbari, N., et al. (2011) A
Association of Public Health Laboratories (2002) multi-country evaluation of LED-fluorescence
External Quality Assessment for AFB Smear microscopy for the diagnosis of pulmonary
Microscopy. Washington, USA. tuberculosis. PLoS Medicine 8, e1001057.
Banada, P.P., Sivasubramani, S.K., Blakemore, R., Cuevas, L.E., Yassin, M.A., Al-Sonboli, N., Lawson,
Boehme, C., Perkins, M.D., Fennelly, K., et al. L., Arbide, I., Al-Aghbari, N., et al. (2011) A
(2009) Containment of bioaerosol infection risk multi-country non-inferiority cluster randomized
by the Xpert MTB/RIF assay and its applicability trial of frontloaded smear microscopy for the
to point-of-care settings. Journal of Clinical diagnosis of pulmonary tuberculosis. PLoS
Microbiology 48(10), 3551–3557. Medicine 8, e1000443.
Blakemore, R., Story, E., Helb, D., Kop, J., Banada, Gilpin, C., Kim, S.J., Lumb, R., Reider, H.L. and Van
P., Owens, M.R., et al. (2010) Evaluation of the Deun, A. (2007) Working Group on Smear
analytical performance of the Xpert MTB/RIF Microscopy. Critical appraisal of current
assay. Journal of Clinical Microbiology 48(7), recommendations and practice for tuberculosis
2495–2501. sputum smear microscopy. International Journal
Boehme, C.C., Nabeta, P., Hillemann, D., Nicol, of Tuberculosis and Lung Diseases 11(9), 946–
M.P., Shenai, S., Krapp, F., et al. (2010) Rapid 952.
molecular detection of tuberculosis and Grenier, J., Pinto, L., Nair, D., Steingart, K., Dowdy,
rifampicin resistance. The New England Journal D., Ramsay, A., et al. (2012) Widespread use of
of Medicine 363(11), 1005–1115. serological tests for tuberculosis: data from 22
Boehme, C.C., Nicol, M.P., Nabeta, P., Michael, high-burden countries. European Respiratory
J.S., Gotuzzo, E., Tahirli, R., et al. (2011) Journal 39, 502–505.
Feasibility, diagnostic accuracy, and effective- Harries, A.D., Michongwe, J., Nyirenda, T.E., Kemp,
ness of decentralized use of the Xpert MTB/RIF J.R., Squire, S.B., Ramsay, A.R., et al. (2004)
test for diagnosis of tuberculosis and multidrug Using a bus service for transporting sputum
resistance: a multicentre implementation study. specimens to the central reference laboratory:
The Lancet 377(9776), 1495–1505. effect on the routine TB culture service in
Bonnet, M., Gagnidze, L., Githui, W., Guerin, P.J., Malawi. International Journal of Tuberculosis
Bonte, L., Varaine, F., et al. (2011) Evaluation of and Lung Diseases 8(2), 204–210.
LED-based fluorescence microscopy for IUATLD [International Union Against Tuberculosis
diagnosis of tuberculosis in a peripheral and Lung Diseases] (2000) Technical Guide.
laboratory of a high HIV prevalence country. Sputum Examination for Tuberculosis by Direct
PLoS ONE 6(2), e17214. Microscopy in Low-income Countries. 5th edn.
Bonnet, M., Gagnidze, L., Guerin, P.J., Bonte, L., IUATLD, Paris.
Ramsay, A., Githui, W., et al. (2011) Evaluation Keeler, E., Perkins, M.D., Small, P., Hanson, C.,
of combined LED-fluorescence microscopy and Reed, S., Cunningham, J., et al. (2006)
bleach sedimentation for diagnosis of tubercu- Reducing the global burden of tuberculosis: the
losis at peripheral health service level. PLoS contribution of improved diagnostics. Nature
ONE 6, e20175. 444, Suppl 1, 49–57.
Botha, E., den Boon, S., Lawrence, K.A., Reuter, Kemp, J.R., Mann, G., Simwaka, B.N., Salaniponi,
H., Verver, S., Lombard, C.J., et al. (2008) From F.M. and Squire, S.B. (2007) Can Malawi’s poor
suspect to patient: tuberculosis diagnosis and access free tuberculosis services? Patient and
treatment initiation in health facilities in South household costs associated with a tuberculosis
Africa. International Journal of Tuberculosis and diagnosis in Lilongwe. Bulletin World Health
Lung Diseases 12(8), 936–941. Organization 85(8), 580–585.
Chihota, V.N., Grant, A.D., Fielding, K., Ndibongo, Khan, M.S., Dar, O., Sismanidis, C., Shah, K. and
B., van Zyl, A., Muirhead, D., et al. (2010) Liquid Godfrey-Faussett, P. (2007) Improvement of
vs solid culture for tuberculosis: performance tuberculosis case-detection and reduction of
and cost in a resource-constrained setting. discrepancies between men and women by
International Journal of Tuberculosis and Lung simple sputum submission instructions: a
Diseases 14(8), 1024–1031. pragmatic randomised controlled trial. The
Chin, C.D., Laksanasopin, T., Cheung, Y.K., Lancet 369(9577), 1955–1960.
Steinmiller, D., Linder, V., Parsa, H., et al. (2011) Kunnath-Velayudhan, S., Salamon, H., Wang, H.Y.,
Microfluidics-based diagnostics of infectious Davidow, A.L., Molina, D.M., Huynh, V.T., et al.
diseases in the developing world. Nature (2010) Dynamic antibody responses to the
Medicine 17, 1015–1019. Mycobacterium tuberculosis genome. Proceed-
ings of the National Academy of Sciences Tuberculosis and Lung Diseases 14(12), 1506–
107(33), 14703–14708. 1507.
Lawn, S.D., Brooks, S.V., Kranzer, K., Nicol, M.P., Ramsay, A., Steingart, K.R., Cunningham, J. and
Whitelaw, A., Vogt, M., et al. (2011) Screening Pai, M. (2011) Translating tuberculosis research
for HIV-associated tuberculosis and rifampicin into global policies. International Journal of
resistance before antiretroviral therapy using Tuberculosis and Lung Diseases 15, 1283–
the Xpert MTB/RIF assay: a prospective study. 1293.
PLoS Medicine 8(7). Squire, S.B., Ramsay, A., van den Hof, S., Millington,
McNerney, R., Wondafrash, B.A., Amena, K., K.A., Langley, I., Bello, G., et al. (2011) Making
Tesfaye, A., McCash, E.M. and Murray, N.J. innovations work for the poor through
(2010) Field test of a novel detection device for implementation by research. International
Mycobacterium tuberculosis antigen in cough. Journal of Tuberculosis and Lung Diseases 5(7),
BMC Infectious Diseases 8(10), 161. 862–870.
Managing Pharmaceuticals and Commodities for Stall, N., Rubin, T., Michael, J.S., Mathai, D.,
Tuberculosis: A Guide for National Tuberculosis Abraham, O.C., Mathews, P., et al. (2011) Does
Programmes (2005) Virginia, USA. solid culture for tuberculosis influence clinical
Mase, S.R., Ramsay, A., Henry, M., Ng, V., decision-making in India? International Journal
Hopewell, P.C., Cunningham, J., et al. (2007) of Tuberculosis and Lung Diseases 15(5), 641–
The incremental yield of serial sputum smears 646.
in the diagnosis of tuberculosis: a systematic Steingart, K.R., Henry, M., Ng, V., Hopewell, P.C.,
review. International Journal of Tuberculosis Ramsay, A., Cunningham, J., et al. (2006a)
and Lung Diseases 11(5), 485–495. Fluorescence versus conventional sputum
Muyoyeta, M., Schaap, J.A., De Haas, P., Mwanza, smear microscopy for tuberculosis: a systematic
W., Muvwimi, M.W., Godfrey-Faussett, P., et al. review. Lancet Infectious Diseases 6(9), 570–
(2009) Comparison of four culture systems for 581.
Mycobacterium tuberculosis in the Zambian Steingart, K.R., Ng, V., Henry, M., Hopewell, P.C.,
National Reference Laboratory. International Ramsay, A., Cunningham, J., et al. (2006b)
Journal of Tuberculosis and Lung Diseases Sputum processing methods to improve the
13(5), 460–465. sensitivity of smear microscopy for tuberculosis:
Pai, M., Minion, J., Steingart, K.R. and Ramsay, A. a systematic review. The Lancet Infectious
(2010) New and improved tuberculosis Diseases 6(10), 664–674.
diagnostics: evidence, policy, practice and Steingart, K.R., Ramsay, A. and Pai, M. (2007a)
impact. Current Opinion in Pulmonary Medicine Optimizing sputum smear microscopy for the
16, 271–284. diagnosis of pulmonary tuberculosis. Expert
Peres, R.L., Palaci, M., Loureiro, R.B., Dietze, R., Reviews in Anti-Infective Therapy 5(3), 327–
Johnson, J.L. and Maciel, E.L. (2011) Reduction 331.
of contamination of mycobacterial growth Steingart, K.R., Henry, M., Hopewell, P.C., Laal, S.,
indicator tubes using increased PANTA con- Ramsay, A., Menzies, R., et al. (2007b)
centration. International Journal of Tuberculosis Commercial serological antibody detection tests
and Lung Diseases 15(2), 281–283. for the diagnosis of pulmonary tuberculosis: a
Phillips, M., Basa-Dalay, V., Bothamley, G., systematic review. PLoS Medicine 4(6), e202.
Cataneo, R.N., Lam, P.K., Natividad, M.P., et al. Steingart, K.R., Henry, M., Hopewell, P.C., Laal, S.,
(2010) Breath biomarkers of active pulmonary Ramsay, A., Menzies, R., et al. (2007c) A
tuberculosis. Tuberculosis 90(2), 145–151. systematic review of commercial serological
Ramsay, A. and Harries, A.D. (2009) The clinical antibody detection tests for the diagnosis of
value of new diagnostic tools for tuberculosis. extrapulmonary tuberculosis. Thorax 62(10),
F1000 Medical Reports. Part ii, 36. 911–918.
Ramsay, A., Yassin, M.A., Cambanis, A., Hirao, S., Steingart, K.R., Ramsay, A. and Pai, M. (2007d)
Almotawa, A., Gammo, M., et al. (2009) Front- Commercial serological tests for tuberculosis: do
loading sputum smear microscopy services: an they work? Future Microbiology 2(4), 355–359.
opportunity to optimize smear-based case Steingart, K.R., Dendukuri, N., Henry, M., Schiller,
detection of tuberculosis in high prevalence I., Nahid, P., Hopewell, P.C., et al. (2009)
countries. Journal of Tropical Medicine 2009, Performance of purified antigens for
398767. serodiagnosis of pulmonary tuberculosis: a
Ramsay, A., Steingart, K.R. and Pai, M. (2010) meta-analysis. Clinical and Vaccine Immunology
Assessing the impact of new diagnostics on 16(2), 260–276.
tuberculosis control. International Journal of Steingart, K.R., Flores, L.L., Dendukuri, N., Schiller,
12 A. Ramsay et al.
I., Laal, S., Ramsay, A., et al. (2011) Commercial mycobacteriophage-based assays for rapid
serological antibody detection tests for the screening of patients at risk of drug-resistant
diagnosis of tuberculosis: an updated system- tuberculosis. Expert Group Meeting Report
atic review and meta-analysis. PLoS Medicine (http://www.who.int/tb/laboratory/egmreport_
8, e1001062. non-commercial_rapid_dst_nov09.pdf, accessed
Steingart, K.R., Ramsay, A., Dowdy, D.W. and Pai, 11 March 2013).
M. (2012) Serological tests for the diagnosis of WHO (2010a) WHO Report 2010. Global Tubercu-
active tuberculosis: relevance for India. Indian losis Control. WHO, Geneva, Switzerland.
Journal of Medical Research 135, 695–702. WHO (2010b) Same-day diagnosis of tuberculosis
Syhre, M., Manning, L., Phuanukoonnon, S., by smear microscopy. Policy Statement, 2010
Harino, P. and Chambers, S.T. (2009) The scent (http://www.who.int/tb/laboratory/whopolicy_
of Mycobacterium tuberculosis: Part II breath. same_day_diagnosis_bymicroscopy_mar2011.
Tuberculosis 89(4), 263–236. pdf, accessed 11 April 2011).
TDR/FIND (2006) WHO diagnostics for tuberculosis: WHO (2010c) Fluorescent light-emitting diode (LED)
global demand and market potential (http:// microscopy for diagnosis of tuberculosis. Policy
apps.who.int/tdr/svc/publications/tdr-research- Statement, 2010 (http://www.who.int/tb/labora-
publications/diagnostics-tuberculosis-global- tory/whopolicy_led_microscopy_mar2011.pdf,
demand, accessed 11 April 2011). accessed 11 April 2011).
Wilson, N., Chadha, S., Beyers, N., Claassens, M. WHO (2010d) Policy framework for implementing
and Naidoo, P. (2011) Helping the poor access new tuberculosis diagnostics (http://www.who.
innovation in tuberculosis control: using evi- int/tb/laboratory/whopolicyframework_rev_
dence from implementation research. Inter- june2011.pdf, accessed 11 March 2013).
national Journal of Tuberculosis and Lung WHO (2011a) Stop TB department. TB diagnostics
Diseases 15(7), 853. and laboratory strengthening (http://www.who.
WHO [World Health Organization] (2007a) The use int/tb/laboratory/policy_statements/en/index.
of liquid medium for culture and DST, Geneva html, accessed 3 August 2011).
(http://www.who.int/tb/laboratory/policy_liquid_ WHO (2011b) WHO policy statement: commercial
m e d i u m _ fo r _ c u l t u r e _ d s t / e n / i n d ex . h t m l , serodiagnostic tests for diagnosis of
accessed 11 March 2013). tuberculosis, Geneva (http://whqlibdoc.who.int/
WHO (2007b) Definition of a new sputum smear publications/2011/9789241502054_eng.pdf,
positive TB case (http://www.who.int/tb/ accessed 3 August 2011).
laboratory/policy_sputum_smearpositive_tb_ WHO (2011c) Non-commercial culture and drug
case/en/index.html, accessed 11 April 2011). susceptibility testing methods for screening of
WHO (2007c) Reduction of number of smears for patients at risk for multi-drug resistant tubercu-
the diagnosis of pulmonary TB, 2007 (http:// losis: policy statement, Geneva (http://whqlibdoc.
www.who.int/tb/laboratory/policy_diagnosis_ who.int/publications/2011/9789241501620_eng.
pulmonary_tb/en/index.html, accessed 11 April pdf, accessed 11 March 2013).
2011). WHO (2011d) Automated real-time nucleic acid
WHO (2008a) Laboratory-based evaluation of 19 amplification technology for rapid and
commercially-available rapid diagnostic tests for simultaneous detection of tuberculosis and
tuberculosis. UNICEF/UNDP/World Bank/WHO rifampicin resistance: Xpert MTB/RIF system.
Special Programme for Research and Training Policy statement (http://whqlibdoc.who.int/
in Tropical Diseases. publications/2011/9789241501545_eng.pdf,
WHO (2008b) Molecular line-probe assays for accessed 11 March 2013).
rapid screening of patients at risk of multi-drug WHO (2011e) Rapid implementation of the Xpert
resistant tuberculosis: policy statement, Geneva MTB/RIF diagnostic test: technical and
(http://www.who.int/tb/features_archive/policy_ operational ‘how-to’ considerations (http://
statement.pdf, accessed 11 March 2013). whqlibdoc.who.int/publications/2011/97892415
WHO (2009) Non-commercial culture methods and 01569_eng.pdf, accessed 11 March 2013).
2 Molecular Diagnosis of Active
Pulmonary Tuberculosis
Anna L. Bateson,1 Kate Reddington2 and Justin O’Grady3

1Centre for Clinical Microbiology, Department of Infection, University College
London, UK; 2Molecular Diagnostics Research Group, Microbiology, School of
Natural Sciences, National University of Ireland Galway, Ireland; 3Norwich Medical
School, Norwich, UK
2.1 Introduction efforts of agencies such as the World Health

Organization (WHO), Stop TB Partnership’s
Accurate and rapid diagnosis of tuberculosis New Diagnostics Working Group (NDWG)
is of paramount importance in establishing and the Foundation for Innovative New
appropriate clinical management and Diagnostics (FIND) have facilitated this
infection control measures (Migliori et al., process (Wallis et al., 2010). Liquid media for
2010; Wallis et al., 2010; O’Grady et al., 2011). culture and drug susceptibility testing (DST),
The most common method for diagnosing molecular line probe assays (LPAs) for
tuberculosis worldwide is sputum smear screening people at risk of multidrug-
microscopy, a technology that is over a resistant tuberculosis (MDR-TB), light-
century old and the sensitivity of which is emitting diode microscopy, non-commercial
notoriously poor, particularly in human culture and DST and automated real-time
immunodeficiency virus (HIV)-positive PCR technology for the simultaneous
patients. Culture, the gold standard diagnostic detection of tuberculosis and rifampicin (RIF)
method, is now available in most countries, resistance (Xpert MTB/RIF assay) have all
but typically only centrally in resource-poor been endorsed by the WHO since 2007 (Fig.
settings (WHO, 2011b). Culture is highly 2.1). Endorsement of technology by the
sensitive but takes between 2 and 6 weeks to WHO, combined with FIND’s negotiations on
obtain a result. Rapid and sensitive tools for pricing with industry has made new and
the diagnosis of tuberculosis are required, improved diagnostic tests more affordable
due to the increased incidence of tuberculosis and feasible for developing countries (Wallis
worldwide, particularly in HIV-endemic et al., 2010; O’Grady et al., 2011).
regions, and the length of time required for Molecular diagnostic methods for the
classical diagnostic tests. Increasingly, rapid detection of pulmonary tuberculosis have
molecular tests are being used for the been commercially available since the 1990s.
detection of pulmonary tuberculosis. The amplified MTD (Gen-Probe, Inc, San
Since the turn of the century, there have Diego, USA) test was the first nucleic acid
been significant advances in diagnostic amplification test (NAAT) to get US Food and
technologies for tuberculosis. Increase in both Drug Administration (FDA) approval for
public and private investment and the joint detection of pulmonary tuberculosis, in 1996,

(ed. T.D. McHugh) 13
14
Liquid culture and DST
Rapid speciation
LPA for MDR-TB
Non-commercial culture and DST (MODS, NRA, CRI)
LPA for XDR-TB
REFERENCE LEVEL 10–40
Access after 5 years (%)

Distance from patients
New SS+ case definition Xpert MTB/RIF

2-specimen approaches
LED microscopy Manual NAAT 1st generation
Same-day diagnosis Rapid colorimetric DST
A. Bateson et al.
INTERMEDIATE LEVEL 70
POC test (detection of TB)

Prediction (LTBI)
Manual NAAT 2nd genration
PERIPHERAL LEVEL 95
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Technologies or methods endorsed by WHO Technologies at late stages of development Technologies at early stages of development
Abbreviations: DST Drug susceptibility test; NAAT Nucleic acid amplification test; LTBI Latent TB infection; POC Point of care: MODS Microscopic observation drug susceptibility;
NRA Nitrate reductase assay; CRI Colorimetric redox indicator assay; LED Light-emitting diode; LPA Line probe assay
Fig. 2.1. The tuberculosis diagnostics pipeline in 2011 (reproduced with permission from WHO (WHO, 2011b).
Molecular Diagnosis of Active Pulmonary Tuberculosis 15
shortly followed by the Roche Amplicor MTB region have excellent correlation with pheno-
test (Nyendak et al., 2009). Early molecular typic DST results, with specificities and
methods were designed to detect the sensitivities >90%, and have been endorsed by
Mycobacterium tuberculosis complex (MTC) the WHO (WHO, 2008, 2011a). Molecular
only. These tests were followed by the detection of resistance to other anti-M.
introduction of LPAs, which combined tuberculosis drugs, such as isoniazid (INH)
nucleic acid amplification with hybridization. and ethambutol (EMB), is more complex and
LPAs have a higher capacity for multi- requires detection of mutations in multiple
parametric detection and have been designed genes for a good correlation with phenotypic
to detect both MTC and first- and second-line results (O’Grady et al., 2011). INH resistance-
anti-M. tuberculosis drug resistance. The most associated mutations have been found in the
recent commercially available WHO katG and inhA genes; however, approximately
endorsed NAAT is the Xpert MTB/RIF assay, 15–25% of INH-resistant isolates do not
which utilizes real-time PCR for the contain mutations in these regions, indicating
simultaneous detection of MTC and RIF that other sites must also be involved in
resistance. The major technological advance resistance (Parsons et al., 2011). Accurate
of this method is the automation of the genotypic DST for first- and second-line
sample analysis, from nucleic acid extraction anti-M. tuberculosis drugs is, therefore,
to result readout. Molecular diagnosis for technically challenging. Increased under-
tuberculosis is not limited to those tests that standing of the molecular mechanisms of
are commercially available and endorsed by anti-M. tuberculosis drug resistance and
the WHO. There have been numerous in- improved multi-parametric detection tech-
house molecular diagnosis assays developed nology will make genotypic DST a more
over the past two decades, with varying powerful technique in the future.
specificities and sensitivities. Tuberculosis in humans can be caused by
Resistance to anti-M. tuberculosis drugs any of the members of the MTC. The eight
arises as a result of spontaneous mutations in closely related species of the MTC have a
genes encoding either the target of the drug or wide range of natural hosts, including human
enzymes involved in drug activation (Parsons hosts (M. tuberculosis, M. africanum, M.
et al., 2011). Resistance-associated mutations canettii), bovine hosts (M. bovis), caprine hosts
have been described for all first-line drugs. (M. caprae), rodent hosts (M. microti) and
MDR-TB develops through the sequential pinniped hosts (M. pinnipedii), along with M.
acquisition of mutations at multiple loci bovis Bacillus Calmette–Guérin (BCG), the
(Parsons et al., 2011). The most common commonly used vaccine strain. All species of
mutations associated with drug resistance the MTC have been implicated in human
have been described and are publicly available infection, regardless of the natural host
on the tuberculosis drug resistance mutation (Reddington et al., 2011). While M. tuberculosis
database (Sandgren et al., 2009). Genotypic is responsible for the majority of cases of
DST methods target these well-characterized human tuberculosis, accurate identification
resistance-associated mutations to identify of other members of the MTC causing
drug-resistant M. tuberculosis. One of the most infection is not performed routinely (Pinsky
important drugs in the treatment of tubercu- and Banaei, 2008). As a result, the global
losis is rifampicin (RIF). RIF resistance is frequency and distribution of each member
particularly suitable for genotypic DST of the MTC remains largely unknown, with
because 95% of RIF resistance-associated some studies suggesting that tuberculosis
mutations are present in an 81-bp region of caused by members of the MTC other than M.
the rpoB gene known as the rifampicin tuberculosis are under-represented (Somoskovi
resistance-determining region (RRDR) (Blake- et al., 2009; Allix-Béguec et al., 2010). M.
more et al., 2010; O’Grady et al., 2011; Parsons africanum, for example, has been identified as
et al., 2011). Molecular methods, such as LPAs the causative agent of up to 50% of human
and Xpert MTB/RIF assay, targeting this tuberculosis infections in West African
16 A. Bateson et al.
countries (Kallenius et al., 1999) and M. Walker et al., 1992a; Andersen et al., 1993;
canettii was responsible for 10% of tuberculosis Verma et al., 1994; Beige et al., 1995; Del
cases in a recent study performed in the Portillo et al., 1996; Iwamoto et al., 2003; Flores
Republic of Djibouti (Koeck et al., 2010). et al., 2005; Cui et al., 2012). A published meta-
Furthermore, M. bovis remains an important analysis, which focused on PCR-based
cause of zoonotic tuberculosis worldwide amplification techniques, showed estimates
and should not be overlooked in a clinical of the accuracy of in-house NAATs were
setting (Hlavsa et al., 2008; Allix-Béguec et al., highly heterogeneous, with sensitivity
2010). In developing countries, where there is ranging from 9.4% to 100% and specificity
a paucity of data, one study suggests that the estimates of 5.6–100% (Flores et al., 2005).
prevalence of bovine tuberculosis in humans Several commercial NAATs are available
may be as high as 15% (de la Rua-Domenech, which use different amplification methods.
2006). Some members of the MTC (M. bovis, These include the Roche Amplicor M.
M. bovis BCG and M. canettii) are intrinsically tuberculosis test (Amplicor MTB), the Gen-
resistant to pyrazinamide, an important first- Probe Amplified M. tuberculosis Direct
line anti-M. tuberculosis drug (Somoskovi et (AMTD) test, the Beckton Dickinson ProbeTec
al., 2009). Therefore, specific identification of test and, most recently, the Eiken loop
these members of the MTC is clinically mediated isothermal amplification (LAMP)
important for treatment management test (Parsons et al., 2011). This chapter is not
decisions. The accurate molecular identifi- an exhaustive review of all commercial
cation of species of the MTC is therefore NAATs and will focus on the tests listed
required to guide public health and clinical above, as they are the most widely used and
decisions more effectively. have the most published clinical data
For the purposes of this chapter, mol- available. NAATs that also assess drug
ecular diagnostic methods for tuberculosis resistance are covered in a separate section.
are divided into three groups: NAATs, LPAs The Amplicor MTB test (Roche Diag-
and sequencing assays. The chapter covers nostic Systems, New Jersey, USA) is a PCR-
both commercial and in-house methods for based amplification test which targets the 16S
the detection of the MTC, differentiation of rRNA gene, and the amplified product is
the MTC and molecular DST for anti-M. detected by a colorimetric reaction after
tuberculosis drugs. The advantages, disadvan- hybridization to oligonucleotide probes
tages and future of molecular diagnosis for (Palomino, 2009). An automated version of
tuberculosis, including point-of-care (POC) the test, the COBAS Amplicor MTB test,
diagnostic tests, are also discussed. allows fully automated amplification and
detection in the same system using the
COBAS Amplicor analyser. Assay results for
2.2 Nucleic Acid Amplification Tests the COBAS Amplicor MTB test are available
in 6–7 h. This method is intended for use on
2.2.1 NAATs for the detection of the MTC decontaminated, concentrated samples and
has FDA approval for testing of smear-
A significant number of studies have been positive respiratory specimens. More recently,
published describing both commercial and the COBAS TaqMan MTB test has also been
in-house NAATs for the diagnosis of introduced using real-time PCR and hybrid-
pulmonary tuberculosis in clinical samples. ization in the COBAS TaqMan 48 analyser
In-house tests have targeted a wide range of that can run 48 samples in 2.5 h (Kim et al.,
DNA sequences specific for the MTC, 2011; Yang et al., 2011). Many studies have
including rRNA (16S and 23S); genes evaluated the performance of the Amplicor
encoding the proteins MPB64, MTP40 and MTB test (Piersimoni and Scarparo, 2003;
protein antigen b; resistance-determining Greco et al., 2006; Ling et al., 2008a), with
genes such as gyrB and rpoB; insertion overall reported sensitivity ranging from 90%
sequences IS6110 and IS986 and repetitive to 100% in smear-positive samples and from
elements (De Wit et al., 1990; Soini et al., 1992; 50% to 95.9% in smear-negative samples, and
overall specificity ranging from 91.3% to FDA (Piersimoni and Scarparo, 2003). Many
100% (Palomino, 2009; Parsons et al., 2011). studies have evaluated the ProbeTec assay
The AMTD assay (Gen-Probe) is an (Piersimoni and Scarparo, 2003; Greco et al.,
isothermal transcription-mediated amplifi- 2006; Ling et al., 2008a), with sensitivities
cation (TMA) method which also targets 16S ranging from 98.5% to 100% in smear-positive
rRNA. The Gen-Probe TMA method amplifies samples and from 33% to 85.7% in smear-
a specific rRNA target by transcription of negative samples, and an excellent overall
DNA intermediates, yielding multiple copies specificity of 98.9–100% (Piersimoni and
of mycobacterial RNA (Piersimoni and Scarparo, 2003; Parsons et al., 2011).
Scarparo, 2003). These RNA amplicons are The newest NAAT is the LAMP assay
then detected by binding to a single-stranded, (Eiken Chemical Co Ltd, Tokyo, Japan),
acridinium ester-labelled DNA probe, giving which is soon to be made commercially
a chemoluminesence signal that is read in a available. This is a novel, isothermal molecu-
luminometer (Piersimoni and Scarparo, lar method with high amplification efficacy.
2003). Amplification and detection are per- The target of the Eiken LAMP assay is an M.
formed in a single tube and test results are tuberculosis specific region of the gyrB gene,
available within 2.5 h. This method is encoding DNA gyrase subunit B. The test is a
approved by the FDA for testing on both closed system and, importantly, the amplified
smear-positive and smear-negative respira- product can be seen by direct visual
tory samples (Piersimoni and Scarparo, 2003). inspection of turbidity or fluorescence in the
The reported sensitivities ranged from 91.7% tube (Neonakis et al., 2011).
to 100% in smear-positive samples and from A full description of the LAMP methodo-
65.5% to 92.9% in smear-negative samples, logy is available in full elsewhere (Tomita et
with a specificity of 92.1–100% (Palomino, al., 2008) and a helpful animation can be
2009; Parsons et al., 2011). A recent prospective viewed online (Eiken). The LAMP assay
study has evaluated the AMTD in a high HIV requires four specially designed primers that
prevalence population in Uganda and showed recognize six distinct sites – one pair of
that even a moderate sensitivity of 39% in standard primers each recognizing a single
smear-negative patients had a substantial domain and one pair of composite primers
clinical impact in this population (Davis et al., that each recognize a double domain (two
2011). consecutive sequences). The reaction has two
The ProbeTec MTB and newer ProbeTec phases: the starting structure-producing step
ET Direct TB tests (Beckton Dickinson, and the cycle-amplification step. To generate
Sparks, Maryland, USA) use strand displace- the starting structure, all four of the primers
ment amplification (SDA) to co-amplify the are required in sequence, initiating rounds of
target DNA sequences of IS6110 and the 16S strand displacement DNA synthesis and
rRNA gene, and real-time detection of resulting in a single-stranded dumbbell struc-
fluorescent labelled detector probes using a ture with loops at either end. The amplification
ProbeTec instrument. SDA is an isothermal stage requires only the two composite
reaction that involves nicking of the DNA by primers and results in various-sized stem-
a restriction endonuclease and extension by a loop structures, consisting of alternatively
DNA polymerase with strand displacement inverted repeats of the target sequence, and
activity (Walker et al., 1992b). The newly cauliflower-like structures with multiple
replicated and displaced strands are then loops formed by annealing of the repeat
substrates for repeated rounds of primer sequences on the same strand (Notomi et al.,
annealing, nicking and strand displacement 2000; Neonakis et al., 2011). The LAMP
(Walker et al., 1992b; Piersimoni and Scarparo, reaction can also be used with RNA as the
2003). The ProbeTec MTB test assay is a closed target, with the addition of reverse tran-
system and takes approximately 3–4 h to scriptase (RT-LAMP). Visualizing of LAMP
perform. It is recommended for use on amplified DNA can be performed in two
digested, decontaminated respiratory speci- ways. As the reaction progresses, a white
mens, but is not currently approved by the precipitate of magnesium pyrophosphate is
formed as a by-product (Tomita et al., 2008). reproducible results across a range of set up
The precipitate can be observed directly by temperatures, pH and in the presence of
visual inspection or by reading tubes in a known PCR inhibitors) (Boehme et al., 2007;
turbidometer. Alternatively, calcein, a che- Francois et al., 2011). Feasibility studies in
lating agent that yields strong fluorescence microscopy centres also showed minimal
when combined with metallic ions such as technical training was required (Boehme et
magnesium, can be added to the reaction and al., 2007). Together, these practical consider-
the amplified product detected by visual ations suggest the platform is applicable for
examination under UV light or using a use in a resource-poor, peripheral laboratory
spectrophotometer (Tomita et al., 2008). setting. However, the inferior performance in
The first comprehensive clinical evalu- smear-negative patients (see Table 2.1;
ation of the LAMP assay for the detection of Boehme et al., 2010) may suggest limited
pulmonary tuberculosis showed a sensitivity utility in tuberculosis populations with high
of 97.7% in smear-positive sputum samples rates of smear-negative tuberculosis, such as
and 48.8% in smear-negative samples, and a HIV-positive individuals and children,
specificity of 99% (Boehme et al., 2007). without further modifications. More clinical
Modifications were made to the LAMP kit by data are required to assess the commercial
Eiken Chemicals to simplify the DNA LAMP assay in these groups. For a test to
extraction process, and a recent study in have maximal diagnostic impact in remote,
Japan has confirmed this new kit has similar resource-poor settings, it may be a trade-off
diagnostic performance (Mitarai et al., 2011). between practical operational characteristics
A comparison of LAMP to other molecular and performance until the next generation of
methods showed performance was similar to tests that satisfy both are available.
other molecular tests for smear-positive
samples but was slightly lower for smear
negatives, particularly in treated sputum 2.2.2 NAATs for the detection of the MTC
pellets (Mitarai et al., 2011). and drug resistance
Overall, the Eiken LAMP assay is simple
(no special reagents or equipment required), There are several commercially available
rapid (a turnaround time of 60–90 min NAATs for the detection of M. tuberculosis
including DNA extraction) and robust (yields with simultaneous detection of drug
Table 2.1. Commercially available NAATs for the detection of the MTC (Ling et al., 2008a; Parsons et al.,
2011; WHO, 2011a; Chang et al., 2012).
Sensitivity range (%)
Smear- Smear-
positive negative Specificity
NAAT name Manufacturer Amplification method samples samples range (%)
Amplicor MTB Roche Molecular PCR 90–100 50–95.9 91.3–100
test Systems
Amplified MTD Gen-Probe Inc Transcription- 91.7–100 65.5–92.9 92.1–100
test mediated
amplification
BD ProbeTec Becton Dickinson Strand displacement 98.5–100 33.3–85.7 98.9–100
ET amplification
Loopamp® Eiken Chemical Loop-mediated 97.7 48.8 99
Tuberculosis Co isothermal
Complex amplification
Detection
Reagent Kit
Xpert MTB/RIF Cepheid PCR 95–100 43.4–84.6 92–100
resistance. These include the Xpert MTB/RIF subsequent steps are performed automatic-
assay (Cepheid Inc, Sunnydale, California, ally; each cartridge incorporates microfluidics
USA), which detects RIF resistance, and LPAs technology and contains all the reagents
(discussed in a separate section) for the required for bacterial lysis, nucleic acid
detection of resistance to first-line drugs – extraction, amplification and amplicon detec-
MTBDRplus (Hain Lifesciences, GmbH, tion (Lawn and Nicol, 2011). The results are
Nehren, Germany) and INNO-LiPA (Inno- generated automatically in a user-friendly
genetics, Gent, Belgium); and second-line format. Cartridges can be loaded and
drugs – MTBDRsl (Hain Lifesciences). Array- processed independently without the need
based assays are also available to test for a for batching. The assay has a hands-on time
wider panel of resistance-determining muta- of less than 15 min and can be performed by
tions; these include QIAPlex (Qiagen, relatively low-skilled technicians, with mini-
Valencia, California, USA), TruArray MDR-TB mal training, within 2 h (Lawn and Nicol,
(Akonni Biosystems, Maryland, USA) and 2011). Safety studies have also proven the
INFINTI MDR-TB (Autogenomics, California, assay can be performed without the special-
USA) tests. Many in-house NAATs have also ized bio-containment equipment that is
been published for the detection of resistance usually required for handling viable M.
to both first- and second-line tuberculosis tuberculosis (Banada et al., 2010).
drugs (Parsons et al., 2011). These are most Two large multicentre, prospective evalu-
commonly real-time PCR methods with ation studies have been carried out using
resistance mutations detected by fluorescent reference laboratory facilities (Boehme et al.,
probes (TaqMan probes or molecular beacons) 2010), and in district and subdistrict health
or melting curve analysis (fluorescence facilities (Boehme et al., 2011), and both have
resonance energy transfer (FRET) probe shown excellent performance. In the former,
melting curve analysis or high-resolution the sensitivity of a single Xpert MTB/RIF
melting (HRM) curve analysis) (Lin et al., assay was 98.2% in smear-positive patients
2004; Wada et al., 2004; Yesilkaya et al., 2006; and 72.5% in smear-negative patients, with a
Ramirez et al., 2010). More recently, several specificity of 99.2% (Boehme et al., 2010). For
new probe-based melting curve analysis detection of RIF resistance, sensitivity was
technologies have also been evaluated, with 97.6% and specificity was 98.1% (Boehme et
promising results (Chakravorty et al., 2011; al., 2010). Importantly, performance data
Luo et al., 2011). from the study at peripheral health facilities
The Xpert MTB/RIF assay, recently showed similar results, providing important
developed by Cepheid Inc in collaboration data on the suitability of implementing the
with FIND, is the most promising molecular assay in low-resource settings (Boehme et al.,
diagnostic technology for M. tuberculosis. This 2011). A number of other studies have
is a fully automated test for the simultaneous evaluated the Xpert MTB/RIF assay in
detection of MTC and RIF resistance directly pulmonary tuberculosis (Helb et al., 2010;
from clinical samples. The assay utilizes a Armand et al., 2011; Marlowe et al., 2011;
hemi-nested PCR to amplify an M. tuberculosis Moure et al., 2011; Rachow et al., 2011) and a
specific sequence of the rpoB gene (192 bp) recent meta-analysis showed a pooled
and M. tuberculosis is detected by five sensitivity of 90.4% and specificity of 98.4%
overlapping molecular beacon probes that for the detection of tuberculosis; and a pooled
are complementary to the 81-bp RRDR (Lawn sensitivity of 94.1% and specificity of 97.0%
and Nicol, 2011). The Xpert MTB/RIF assay for the detection of RIF resistance (Chang et
procedure has been described in detail else- al., 2012). The Xpert MTB/RIF assay has also
where (Blakemore et al., 2010; Helb et al., been shown to perform well in HIV-positive
2010). Briefly, sputum samples or decon- patients, with an overall sensitivity of 82.0%
taminated sputum pellets are treated with a and specificity of 98.0% (Boehme et al., 2010;
reagent to reduce M. tuberculosis viability and Scott et al., 2011; Chang et al., 2012), out-
loaded manually into a cartridge, which is performing other molecular tests in smear-
inserted into the GeneXpert instrument. All negative patients (Scott et al., 2011). There is
potential for the Xpert MTB/RIF assay to be of graphical settings and health-care facilities,
great clinical value for tuberculosis screening and is a major advance in tuberculosis
in HIV treatment programmes (Lawn et al., diagnostic assay development. As for most
2011). Tuberculosis is the one of the main molecular tests, the major drawback to
causes of death early in antiretroviral treat- implementing this technology in low- and
ment (ART), and pre-screening using the middle-income countries is the high costs.
Xpert MTB/RIF assay prior to starting ART FIND have negotiated substantial cost
increased case detection by 45% compared to reductions, with the price of the instrument
microscopy (Lawn et al., 2011). The Xpert set at US$17,000 for a four-module instrument
MTB/RIF assay has also been evaluated for and a current cost per test of US$18, although
use in paediatric tuberculosis, with promising this is expected to drop to US$10.70 following
results (Nicol et al., 2011). the WHO endorsement and increased sales
In December 2010, the WHO endorsed (Lawn and Nicol, 2011; O’Grady et al., 2011).
the use of Xpert MTB/RIF as the initial The current costs are similar to those for
diagnostic test for individuals suspected of culture identification and drug sensitivity
having MDR-TB or HIV-associated tubercu- testing, especially factoring in biosafety
losis, and as a follow-on to microscopy in equipment costs, but higher than other
settings where MDR-TB and HIV are of lesser molecular tests (Lawn and Nicol, 2011;
concern (WHO, 2010b), with the expectation Parsons et al., 2011). There are also practical
for an initial phased implementation from limitations associated with the use of
2011 onwards and a more rapid scale-up specialized equipment and, together with the
thereafter (WHO, 2010a; Lawn and Nicol, current need to confirm drug resistance by
2011). culture-based methods, this suggests the
Although the initial evaluation studies Xpert MTB/RIF assay has good potential for
showed high specificity for RIF resistance, use in moderately equipped laboratories, but
several subsequent studies have confirmed not in most peripheral settings where
false positive RIF-resistance results, particu- infrastructure and resources are limited (Van
larly in low-prevalence settings (Lawn et al., Rie et al., 2010; Lawn and Nicol, 2011;
2011; Marlowe et al., 2011; Van Rie et al., 2012). O’Grady et al., 2011). While there is no doubt
Absolute numbers of these cases are small, this is currently the best available molecular
but this has important implications if diagnostic test for tuberculosis, it remains
resistance data are to be used for treatment open for discussion if the high costs and
decisions. As such, the WHO recommends logistical considerations required for an
that RIF resistance detected by Xpert should international roll-out programme would be
be confirmed by culture-based methods better placed waiting for a real POC solution.
(WHO, 2011c). This negates somewhat the Further studies into cost-effectiveness, impact
use of the Xpert MTB/RIF assay for rapid on treatment and programme outcomes and
resistance testing if additional facilities with tuberculosis transmission are needed to help
the capacity for phenotypic drug sensitivity answer these questions.
testing are required to validate the results. In
response to this problem, Cepheid has made
some minor changes to the Xpert MTB/RIF 2.2.3 Array-based commercial tests for
assay, including modifying one of the probes the detection of MTB and drug resistance
(FIND, 2011). The new G4 software and
cartridge combination was released outside The recently developed QIAPlex test (Qiagen,
the USA in December 2011, and clinical data Valencia, California, USA) uses a target-
are required to determine if these modifi- enriched multiplex PCR to amplify 24
cations will result in more accurate RIF mutations simultaneously in the katG, inhA,
resistance detection. rpoB, rrs, rpsL and embB genes associated with
Overall, the Xpert MTB system has been resistance to INH, RIF, streptomycin (SM)
shown to provide excellent performance in a and EMB (Gegia et al., 2008; Parsons et al.,
number of different patient groups, geo- 2011). The PCR products are characterized
using a suspension array for multiplex has reported sensitivity and specificity of 93%
detection on the Luminex 100 instrument and 99%, respectively (Antonova et al., 2008;
(Luminex, Austin, Texas, USA). One study in Gryadunov et al., 2011).
the Republic of Georgia showed good
performance for INH and RIF detection, with
sensitivity and specificity of 85.4% and 96.1%, 2.2.4 NAATs for differentiation of
94.4% and 99.4%, respectively, and 86.7% and the MTC
100% for MDR-TB (Gegia et al., 2008).
Sensitivity and specificity was 69.6% and Accurate differentiation of members of the
99.2% for SM and 50% and 98.8% for EMB, MTC is necessary to perform epidemiological
reflecting the poor understanding of the studies, to monitor whether tuberculosis
molecular mechanisms of resistance to these transmission is human-to-human or zoonotic
drugs (Gegia et al., 2008). and to administer appropriate anti-M.
Two solid microarray detection systems, tuberculosis drug treatment, as some members
TruArray MDR-TB (Akonni Biosystems) and of the complex display inherent resistance to
INFINTI MDR-TB (Autogenomics), are also PZA (Reddington et al., 2012). Differentiation
available for MDR-TB detection (Akonni of members of the MTC is difficult to achieve,
Biosystems, 2012; AutoGenomics, 2012). Both as these species are 99% similar on a
systems use multiplex PCR followed by nucleotide sequence level (Brosch et al., 2002;
detection on a microarray. The TruArray Mostowy et al., 2002). Currently, there is only
system uses gel element microfluidics array one commercially available diagnostic assay
technology (Gryadunov et al., 2011; Cooney et for differentiation of the MTC, namely the
al., 2012) to detect the most common Genotype MTBC kit (Hain Lifesciences),
mutations in the rpoB, inhA and katG genes which is an LPA (discussed in a separate
and distinguish between MTC and M. avium section). There are also a number of in-house
by the detection of insertion sequences IS6110 NAATs for MTC differentiation described in
and IS1245 (Akonni Biosystems, 2012). The the literature (Parsons et al., 2002; Huard et al.,
INFINITI MDR-TB assay uses a novel film- 2003; Djelouadji et al., 2008; Pinsky and
based microarray (BioFilmChip) to evaluate Banaei, 2008; Halse et al., 2011). The majority
INH, RIF and PZA resistance by the detection of these methods use regions of difference
of 16 mutations in rpoB, katG, inhA and pncA (RDs) as species-specific diagnostic targets. In
genes, and can also identify M. bovis mycobacteria, RDs represent regions of the
specifically (AutoGenomics, 2012). This test genome present in M. tuberculosis, which are
can be used directly from decontaminated subsequently deleted in other members of the
sediment without the need for DNA MTC (Brosch et al., 1998; Behr et al., 1999;
extraction and can deliver results within 5 h. Gordon et al., 1999).
At present, both of these commercial tests are One of the first conventional PCR methods
for experimental use only and published described for the accurate differentiation of
performance data are not yet available. How- the MTC was proposed by Parsons et al.
ever, the proprietary technology, TB-Biochip (2002). In this study, six RDs (RD 1, 9, 10, 3, 5
(used in the TruArray kits), is certified for and 11) were used to identify M. tuberculosis,
diagnosis in the Russian Federation and has M. africanum, M. bovis, M. bovis BCG and M.
been evaluated in more than 20,000 clinical microti accurately (Parsons et al., 2002). The
specimens. These studies have reported method was tested on 88 well-characterized
sensitivities between 89.5% and 96.9% for RIF MTC strains and specificity was determined
resistance and between 82.1% and 89.3% for to be 100%, but the sensitivity of the method
INH resistance, and specificities of 86.2– was not reported. Subsequently, 605 MTC
94.3% for RIF resistance and 80–95.1% for clinical isolates were evaluated to demonstrate
INH resistance (Gryadunov et al., 2011). A further the robustness of the method.
similar microarray targeting mutations that However, since the addition of M. caprae, M.
determine fluoroquinolone resistance in the canettii and M. pinnipedii to the MTC, the
gyrA gene (TB-Biochip-2) is also available and species-specific diagnostic profile obtained
using these six RDs is no longer valid. More recently, multiplex real-time PCR
In a study by Huard et al. (2003), a PCR- assays have been developed for the
based method was developed using RDs to differentiation of the MTC. One of the first
identify all members of the MTC accurately, such assays was described by Pinsky and
with the exception of M. pinnipedii. In this Banaei in 2008. In this study, a two-stage
study, seven genetic loci were used; a multiplex real-time PCR method using melt
mycobacterial amplification control (16S curve analysis targeting RDs 9, 4 and 1 was
rRNA), an MTC-specific target (Rv0577) and developed. Based on the presence or absence
species-specific diagnostic targets (IS1561, of a melt peak at a certain temperature, M.
Rv1510, Rv1970, Rv3877/8 and Rv3120) tuberculosis, M. bovis and M. bovis BCG can be
(Huard et al., 2003). For each MTC member, identified (Pinsky and Banaei, 2008). The test
seven individual amplification reactions were was optimized using 4 MTC and 10 NTM
performed in parallel and the PCR products strains and tested subsequently with 80 MTC-
were visualized on an agarose gel. Depending positive clinical isolates. This method is
on the profile observed, each individual rapid, specific (100%) and simple to perform;
species of the MTC could be identified however, there are also a number of limi-
accurately. A total of 71 MTC strains and 44 tations. For example, if a sample is positive
non-tuberculous mycobacteria (NTM) were for RD9, it is assumed M. tuberculosis is
evaluated and the specificity of the method present; however, this RD is also present in
was determined to be 100%, but the sensitivity M. canettii (Brosch et al., 2002). This method
of the method was not reported. However, in cannot identify M. africanum, M. caprae, M.
2003, M. pinnipedii was elevated to rank microti or M. pinnipedii accurately and patient
species and added to the MTC (Cousins et al., samples have not been tested directly.
2003). As it was not tested with this method, it Another multiplex real-time PCR method
was unknown if M. pinnipedii would give the using melt curve analysis for accurate
same diagnostic profile as other MTC differentiation of the MTC was described
members. A limitation of this method is that it recently by Pounder et al. (2010). In this assay,
has not been evaluated directly on clinical RDs 1, 4, 9, 10 and 12 are amplified and, based
samples. on the melting temperature (Tm) of the
Another conventional PCR approach for amplicon generated, the particular member
the differentiation of the MTC was proposed of the MTC can be identified (Pounder et al.,
by Warren et al. in 2006. Two multiplex PCRs 2010). This assay was validated against a
are performed sequentially to identify all panel of 129 MTC strains and 11 NTM and
eight members of the MTC accurately was highly specific for each MTC member
(Warren et al., 2006). The first multiplex PCR (96%), rapid and simple to perform (Pounder
targets areas of RDs 1, 4, 9 and 12 to identify et al., 2010). However, as with previous
M. tuberculosis, M. canettii, M. bovis, M. caprae studies described above, the sensitivity of the
and M. bovis BCG specifically, based on the method was not reported. M. canettii and M.
amplification patterns observed when visual- pinnipedii were not tested in this study and
ized on an agarose gel (Warren et al., 2006). To this method cannot differentiate unambigu-
differentiate between the remaining MTC ously between M. africanum and M. microti.
members, namely M. africanum, M. microti Again, as with the previous study, only
and M. pinnipedii, RD1mic and RD2seal are cultured isolates were tested.
targeted (Warren et al., 2006). This method is In a study by Halse et al., a hydrolysis
highly specific (100%) and, as the reactions probe-based real-time PCR methodology
are multiplexed, less hands-on time and (MTBC-RD real-time PCR) was developed for
consumables are required; however, the accurate differentiation between members of
sensitivity of the method has not been the MTC, both from cultured isolates and
reported. This method was not evaluated clinical specimens (Halse et al., 2011). This
directly on clinical samples, nor was an multiplex assay targets RDs 1, 4, 9 and 12 and
internal amplification control (IAC) the external junction of RD9. This is among
incorporated. the most thoroughly evaluated methods
described in the literature for differentiation MTC, a significant limitation is that it has not
of the MTC. The specificity of the method was yet been evaluated directly on clinical
tested against a panel of 727 positive samples.
mycobacterial growth indicator tubes (MGIT)
and an additional 129 clinical specimens and
found to be 100%. The analytical sensitivity of 2.3 Line Probe Assays
MTBC-RD real-time PCR was determined
using DNA isolated from an M. tuberculosis 2.3.1 Line probe assays for the detection
strain known to contain one copy of IS6110 of the MTC and drug resistance
and was determined to be 2.5 colony-forming
units (CFU) in cultured isolates, and the limit There are currently four commercially avail-
of detection was 5 CFU in sputum specimens able LPAs for the detection of tuberculosis
(Halse et al., 2011). As culturing of specimens and drug-resistant M. tuberculosis (Table 2.2).
is not required, this method does allow for Firstly, there is the GenoType Mycobacteria
detection of the MTC and identification of the Direct assay (Hain Lifesciences), which
particular species in 1 working day. However, detects the MTC and four common NTM
again, there were some limitations with this directly from decontaminated pulmonary
method; for example, M. caprae and M. and extrapulmonary patient samples (Franco-
pinnipedii were not evaluated in this study, Álvarez de Luna et al., 2006). This test is
therefore it was not known what RD PCR broken into three phases, namely isolation of
profile would be observed for these MTC RNA using a magnetic bead capture method,
members. Furthermore, there is no IAC nucleic acid sequence-based amplification
included in this method. Incorporation of an (NASBA) and reverse hybridization of ampli-
IAC or, preferably, a process control is crucial fied products to a strip containing target-
when developing molecular-based diag- specific oligonucleotide probes (Syre et al.,
nostics, particularly when working with 2009). This test procedure can be performed
clinical samples (Hoorfar et al., 2003, 2004), to in approximately 5.5 h and has reported
eliminate the reporting of false negatives due specificities and sensitivities of 90–100% and
to PCR inhibition, thermocycler malfunction, 92–93.7%, respectively, in smear-positive and
reagent problems or errors in preparing smear-negative samples (Franco-Álvarez de
mixes (Hoorfar et al., 2004). Luna et al., 2006; Neonakis et al., 2009; Syre et
In a recent study by Reddington et al., a al., 2009).
two-stage internally controlled multiplex The INNO-LiPA Rif. TB kit (Innogenetics)
real-time PCR-based method, SeekTB, was is a test for the detection of the MTC and
described for the accurate detection and determination of resistance to RIF and can be
identification of all members of the MTC. used on culture isolates or directly on patient
Depending on which MTC species is present, samples (Rossau et al., 1997). As over 90% of
the method takes between 1.5 and 3.5 h post- RIF-resistant isolates are also resistant to
DNA extraction (Reddington et al., 2012). The INH, this test can be useful in identifying
specificity of the method was determined to potential cases of MDR-TB. Using this assay,
be 100% when evaluated against a panel of the RIF-resistance determining region of the
180 well-characterized MTC, NTM and other rpoB gene is amplified using conventional
bacteria strains and species. The method PCR. Subsequently, the amplified product is
developed was also tested on 125 MGIT hybridized to a nitrocellulose strip containing
positive cultures and the results were com- ten specific probes. These probes include one
pared to the Genotype MTBC kit (Hain MTC-specific probe, five wild-type sensitive
Lifesciences) and were 100% concordant. The probes and four probes for specific mutations
analytical sensitivity was <100 genome in resistant strains (Rossau et al., 1997). On
equivalents for each diagnostic assay completion of the hybridization, a colori-
(Reddington et al., 2012). While SeekTB offers metric signal is observed on the strip and,
advantages over other methods described in based on the profile observed, it can be
the literature for the differentiation of the determined if an isolate is RIF resistant or
Table 2.2. Commercially available hybridization and line probe assays.

Target
Molecular test Manufacturer Technology sequence IAC Detects
GenoType Hain Lifescience NASBA + 23S rRNA Yes MTC, M. avium, M.
Mycobacteria reverse intracellulare, M.
Direct hybridization kansasii, M.
malmoense
INNO-LiPA Rif. TB Innogenetics PCR + reverse rpoB No MTC and resistance
kit hybridization to RIF
GenoType Hain Lifescience Multiplex PCR + rpoB, inhA, Yes MTC and resistance
MTBDRplus reverse katG to RIF and INH
hybridization
GenoType MTBDRsl Hain Lifescience Multiplex PCR + gyrA, rrs, Yes MTC and resistance
reverse embB to fluoroquinolones
hybridization and
aminoglycosides
GenoType MTBC Hain Lifescience Multiplex PCR + gyrB, RD1, Yes MTC and some
reverse 23S rRNA species
hybridization
Accuprobe Gen-probe Hybridization 16S rRNA No MTC
Mycobacterium
tuberculosis
complex culture
identification test
sensitive. In a recent meta-analysis, it was The GenoType MTBDRsl assay (Hain

determined that the specificity and sensitivity Lifesciences) was developed for the detection
ranges of this test were 92–100% and 82– of the MTC and simultaneous detection
100%, respectively, when used on cultured of the most commonly reported resistances
isolates (Morgan et al., 2005). The sensitivity to fluoroquinolones, EMB and second-
may reduce to 80% when used directly on line aminoglycosides and cyclic peptide
clinical samples (Morgan et al., 2005). (Hillemann et al., 2009). If a patient is
The GenoType MTBDRplus test is also reported to have MDR-TB, the GenoType
designed for the detection of the MTC and MTBDRsl assay can be used to identify cases
identification of MDR-TB. This second- of extremely drug-resistant tuberculosis
generation test, which was endorsed for use (XDR-TB) (Hillemann et al., 2009; Brossier et
by the WHO in 2008, allows for the al., 2010). The test procedure is the same as
identification of both RIF and INH resistance the GenoType MTBDRplus and can be used
(WHO, 2008). This test can be used on on culture or sputum smear-positive
bacterial cultures or smear-positive patient samples. In recent evaluation studies,
samples and takes approximately 5 h to fluoroquinolone resistance detection was
perform (Neonakis et al., 2009). The test demonstrated to have a sensitivity of 87–
procedure includes DNA extraction and 90.2% and a specificity of 96–100%, EMB
conventional multiplex PCR, followed by resistance detection was demonstrated to
reverse line hybridization (Hillemann et al., have a sensitivity of 57–77.3% and a
2007). It has been reported recently that the specificity of 88.2–100% and for detection of
GenoType MTBDRplus test has a specificity resistance to the second-line injectable
of 96.8–99.7% and 95.4–99.8% and a sensitivity agents, a sensitivity of 42.7–100% and a
of 95.1–99.5% and 82.4–92.8% for the specificity of 91–100% was reported
determination of RIF and INH resistance, (Hillemann et al., 2009; Brossier et al., 2010;
respectively (Ling et al., 2008b). Ignatyeva et al., 2012).
2.3.2 Line probe assays for the bacteria; however, such methods are not yet
differentiation of the MTC practical for routine use in most resource-
poor countries (Parsons et al., 2011).
The Genotype MTBC test (Hain Lifesciences) Sequencing assays can be used for the
is currently the only commercially available identification of the MTC and NTMs (Kasai et
molecular test for differentiation of the al., 2000; Cristea-Fernstrom et al., 2007),
members of the MTC. Like the other two molecular DST of first- and second-line
Hain kits, this test is based on the principle of anti-M. tuberculosis drugs (Isola et al., 2005;
DNA extraction and multiplex conventional Kontsevaya et al., 2011) and MTC dif-
PCR, followed by reverse-line hybridization ferentiation (Kahla et al., 2011). One or more
(Richter et al., 2004). This test procedure can genes are targeted, depending on the required
be performed in approximately 5.5 h and can application.
be used on culture or sputum smear-positive Sequencing assays can be broken down
samples (Somoskovi et al., 2008). This assay into two types, namely conventional sequenc-
contains probes for a genus control, a ing and pyrosequencing-based methods.
conjugate control, the collective detection of Conventional sequencing methods are
the MTC and specific probes for M. typically used on growth-positive cultures as
tuberculosis, M. africanum, M. bovis, M. caprae, culture confirmation tools. Pyrosequencing, a
M. bovis BCG and M. microti (Richter et al., technique based on the real-time sequencing
2004). However, it has been demonstrated of short sequences during DNA synthesis
that the M. tuberculosis probe can cross react (sequencing by synthesis), has been described
with M. canettii and the M. africanum probe as a promising new tool for the rapid
cross reacts with M. pinnipedii (Richter et al., identification of mycobacteria, not only in
2004; Kjeldsen et al., 2009). culture but also in smear-positive sputum
specimens (Parsons et al., 2011). Although this
method is faster, simpler and less expensive
2.3.3 Direct hybridization assays for the than conventional sequencing, it produces
detection of the MTC sequence reads <100 bp, which have less
discriminatory power than the 500-bp reads
An additional hybridization-based assay is obtained using conventional sequencing. One
the Accuprobe Mycobacterium tuberculosis of the most promising applications of
complex culture identification test (Gen- sequencing assays in tuberculosis diagnosis
probe), which detects the presence of the is the use of pyrosequencing assays for
MTC (Goto et al., 1991). Unlike all other molecular DST (Isola et al., 2005; Kontsevaya
methods discussed above, there is no et al., 2011). Several target regions can be
amplification step required in this procedure. tested and all mutations in the target regions
This test takes approximately 1.5 h to perform can be detected, including mutations not
and involves RNA extraction, hybridization previously characterized. However, a better
and detection of 16S rRNA (Goto et al., 1991). understanding of the molecular resistance
While the specificity of this method is almost mechanisms of some first- and second-line
100% (Evans et al., 1992), a significant limi- anti-M. tuberculosis drugs is required in order
tation is that it is only suitable for use with to identify the appropriate molecular targets
positive cultures (Parsons et al., 2011). and achieve high accuracy compared to
phenotypic DST.
2.4 Sequencing-based Diagnostic

Methods 2.5 Next-generation Technologies for
Tuberculosis Diagnosis
DNA sequencing assays, which target
variable genomic regions to identify species- The potential of POC diagnostic technology
specific sequence motifs, are rapid and to benefit health care is widely recognized
accurate identification methods for myco- (Mauk et al., 2011). POC devices facilitate
individualized medicine and more immediate significant number of these criteria, even
diagnostic tools at doctors’ offices, clinics and though it is recognized as not being a true
hospital bedsides. Perhaps the greatest POC test.
impact of POC tests will be in developing Over the past two decades, many
nations by bringing affordable diagnostic microfluidic lab-on-a-chip POC PCR-based
tools to resource-poor settings for diseases devices have been described, some with
such as tuberculosis and pneumonia (Mauk performance characteristics better than
et al., 2011). In recent years, the need for a existing PCR assays (Craw and Balachandran,
POC tuberculosis test has been increasingly 2012). However, the precise and repeated
recognized and some support for research heating cycles required for PCR necessitate
and development has become available, complex, power-consuming designs and
mainly through philanthropic organizations require additional engineering consider-
such as the Bill and Melinda Gates Foundation ations, such as the thermal constants of the
(Batz et al., 2011). However, despite the effort materials used and unwanted evaporation of
and resources invested, POC diagnosis for water. These characteristics ultimately render
tuberculosis has not yet materialized. The PCR inappropriate for lab-on-a-chip diag-
ideal POC tuberculosis test and its char- nostic devices (Craw and Balachandran,
acteristics have recently been defined by 2012). In order to avoid such difficulties,
Médecins Sans Frontières (MSF) and partner recent diagnostic POC test research has
organizations (Table 2.3; Batz et al., 2011). The focused on isothermal methods for nucleic
Xpert MTB/RIF assay is currently the only acid amplification such as LAMP, NASBA
tuberculosis diagnostic test that matches a and TMA. Isothermal methods use enzymes
Table 2.3. Characteristics for a good POC tuberculosis test.

POC test characteristic Minimum required value
Sensitivity (in adult pulmonary 95% for smear-positive, culture-positive patients
tuberculosis) 60–80% for smear-negative, culture-positive patients
Specificity 95% compared to culture
Time to result 3h
Throughput 20 tests/staff member/day
Specimen type Urine, oral, breath, venous blood, sputum
Sample preparation Three steps maximum
Biosafety level 1
No need for pipetting
No time-sensitive processes
Readout Easy ‘yes’ and ‘no’ answers
Readable for 1 h
Waste disposal Simple burning, no glass
Controls Positive control in test kit
Storage/shelf life Shelf life 24 months, including reagents
Stable at 30°C and higher for short periods of time
Stable in high humidity
Instrumentation If instrument, no maintenance
Works in tropical conditions
Acceptable replacement costs
Fits in backpack
Shock resistant
Power requirement Can work on battery
Training 1 day maximum
Can be used by any health-care worker
Costs <US$10/test
instead of heat to perform strand separation. tuberculosis drugs. The recent development of
These techniques offer a better solution than sample-in, answer-out molecular diagnostic
PCR for nucleic acid amplification on POC technology for tuberculosis has negated some
diagnostic platforms (Craw and Balachandran, of the problems associated previously with
2012). While biosensors for direct detection of molecular methods, such as crossover
nucleic acids without amplification have been contamination, high method complexity,
described in the literature, an integrated the necessity for skilled labour and the
system with clinically relevant specificity and requirement for dedicated molecular biology
sensitivity direct from complex samples has facilities. The majority of commercially avail-
yet to reach the market. Therefore, it is likely able molecular methods, however, are not
that some form of target and/or signal sample-in, answer-out based and continue to
amplification must be performed, given suffer such problems. Molecular methods are
existing detection technology (Craw and still relatively expensive compared to trad-
Balachandran, 2012). Hence, the future for itional methods and often require expensive
the molecular diagnosis of tuberculosis is instrumentation, which must be maintained
likely to be a POC sample-in, answer-out and which requires a relatively high test
diagnostic platform that combines isothermal volume to be cost-effective (Parrish and
nucleic acid amplification and inexpensive Carroll, 2011).
lab-on-a-chip nanobiosensor technology.
A recently published example of such
a biosensor uses thermophilic helicase- 2.7 Conclusions
dependent isothermal amplification and
dextrin-coated gold nanoparticles for the The tuberculosis diagnostics pipeline has
electrochemical detection of the IS6110 gene grown rapidly in recent years with the
of M. tuberculosis (Torres-Chavolla and development of several promising molecular
Alocilja, 2011). Gold nanoparticles and mag- diagnostic technologies. A simple, rapid,
netic particles are each functionalized with inexpensive POC test, however, is still not on
probes specific to opposite ends of a fragment the horizon (Wallis et al., 2010). The next
(105 bp) of the IS6110 gene, specific to the generation of nanobiosensor-based tubercu-
MTC. After hybridization, the gold–DNA– losis diagnostic tests, designed for POC use,
MP complex is separated magnetically from are likely to be still >5 years from com-
the solution and gold nanoparticles are mercialization, but have the potential to
detected electrochemically. The limit of revolutionize the diagnosis of tuberculosis.
detection of this method is 0.01 ng/μl of Accurate, rapid tests are also required for
isothermally amplified target. This biosensor childhood tuberculosis and latent tubercu-
system potentially can be implemented at losis, and diagnostic assays must be devel-
POC in peripheral laboratories with the use oped that can be applied not only to sputum
of a portable, hand-held potentiostat (Torres- but also to multiple sample types.
Chavolla and Alocilja, 2011).
References
2.6 Advantages and Disadvantages of
Molecular Diagnosis for Tuberculosis Akonni Biosystems (2012) TruArray® MDRTB.
Rapid, accurate identification of rifampin/
The major advantage of the molecular isoniazid susceptibility (http://www.akonni.com/
diagnosis of tuberculosis and molecular DST docs/Akonni_TruArray_MDR-TB_Datasheet.
pdf, accessed 30 April 2012).
is the turnaround time to results compared to
Allix-Béguec, C., Fauville-Dufaux, M., Stoffels, K.,
culture. The major disadvantage is the Ommeslag, D., Walravens, K., Saegerman, C.,
reduced accuracy of molecular methods et al. (2010) Importance of identifying Myco-
compared to culture, particularly in smear- bacterium bovis as a causative agent of human
negative tuberculosis and in molecular DST tuberculosis. European Respiratory Journal 35,
of many first- and second-line anti-M. 692–694.
Andersen, A.B., Thybo, S., Godfrey-Faussett, P. Boehme, C.C., Nabeta, P., Hillemann, D., Nicol,
and Stoker, N.G. (1993) Polymerase chain M.P., Shenai, S., Krapp, F., et al. (2010) Rapid
reaction for detection of Mycobacterium molecular detection of tuberculosis and rifampin
tuberculosis in sputum. European Journal of resistance. The New England Journal of
Clinical Microbiology and Infectious Diseases Medicine 363, 1005–1015.
12, 922–927. Boehme, C.C., Nicol, M.P., Nabeta, P., Michael,
Antonova, O.V., Gryadunov, D.A., Lapa, S.A., J.S., Gotuzzo, E., Tahirli, R., et al. (2011)
Kuz’min, A.V., Larionova, E.E., Smirnova, T.G., Feasibility, diagnostic accuracy, and effective-
et al. (2008) Detection of mutations in ness of decentralised use of the Xpert MTB/RIF
Mycobacterium tuberculosis genome determin- test for diagnosis of tuberculosis and multidrug
ing resistance to fluoroquinolones by hybrid- resistance: a multicentre implementation study.
ization on biological microchips. Bulletin of The Lancet 377, 1495–1505.
Experimental Biology and Medicine 145, 108– Brosch, R., Gordon, S.V., Billault, A., Garnier, T.,
113. Eiglmeier, K., Soravito, C., et al. (1998) Use of a
Armand, S., Vanhuls, P., Delcroix, G., Courcol, R. Mycobacterium tuberculosis H37Rv bacterial
and Lemaitre, N. (2011) Comparison of the artificial chromosome library for genome map-
Xpert MTB/RIF test with an IS6110-TaqMan ping, sequencing, and comparative genomics.
real-time PCR assay for direct detection of Infection and Immunity 66, 2221–2229.
Mycobacterium tuberculosis in respiratory and Brosch, R., Gordon, S., Marmiesse, M., Brodin, P.,
non-respiratory specimens. Journal of Clinical Buchrieser, C., Eiglmeier, K., et al. (2002) A new
Microbiology 49, 1772–1776. evolutionary scenario for the Mycobacterium
AutoGenomics (2012) INFINITI MDR-TB (http:// tuberculosis complex. Proceedings of the
www.autogenomics.com/infectious_MTB.php, National Academy of Sciences 99, 3684–3689.
accessed 30 April 2012). Brossier, F., Veziris, N., Aubry, A., Jarlier, V. and
Banada, P.P., Sivasubramani, S.K., Blakemore, R., Sougakoff, W. (2010) Detection by GenoType
Boehme, C., Perkins, M.D., Fennelly, K., et al. MTBDRsl test of complex mechanisms of
(2010) Containment of bioaerosol infection risk resistance to second-line drugs and ethambutol
by the Xpert MTB/RIF assay and its applicability in multidrug-resistant Mycobacterium tubercu-
to point-of-care settings. Journal of Clinical losis complex isolates. Journal of Clinical
Microbiology 48, 3551–3557. Microbiology 48, 1683–1689.
Batz, H., Cooke, G. and Reid, S. (2011) Towards Chakravorty, S., Aladegbami, B., Thoms, K., Lee,
Lab-free Tuberculosis Diagnosis. Treatment J.S., Lee, E.G., Rajan, V., et al. (2011) Rapid
Action Group, Stop TB Partnership, Imperial detection of fluoroquinolone-resistant and
College London and Médecins Sans Frontières, heteroresistant Mycobacterium tuberculosis by
New York. use of sloppy molecular beacons and dual
Behr, M.A., Wilson, M.A., Gill, W.P., Salamon, H., melting-temperature codes in a real-time PCR
Schoolnik, G.K., Rane, S., et al. (1999) assay. Journal of Clinical Microbiology 49, 932–
Comparative genomics of BCG vaccines by 940.
whole-genome DNA microarray. Science 284, Chang, K., Lu, W., Wang, J., Zhang, K., Jia, S., Li,
1520–1523. F., et al. (2012) Rapid and effective diagnosis of
Beige, J., Lokies, J., Schaberg, T., Finckh, U., tuberculosis and rifampicin resistance with
Fischer, M., Mauch, H., et al. (1995) Clinical Xpert MTB/RIF assay: a meta-analysis. Journal
evaluation of a Mycobacterium tuberculosis of Infection 64(6), 580–588.
PCR assay. Journal of Clinical Microbiology 33, Cooney, C.G., Sipes, D., Thakore, N., Holmberg, R.
90–95. and Belgrader, P. (2012) A plastic, disposable
Blakemore, R., Story, E., Helb, D., Kop, J., Banada, microfluidic flow cell for coupled on-chip PCR
P., Owens, M.R., et al. (2010) Evaluation of the and microarray detection of infectious agents.
analytical performance of the Xpert(R) MTB/RIF Biomedical Microdevices 14, 45–53.
assay. Journal of Clinical Microbiology 48, Cousins, D.V., Bastida, R., Cataldi, A., Quse, V.,
2495–2501. Redrobe, S., Dow, S., et al. (2003) Tuberculosis
Boehme, C.C., Nabeta, P., Henostroza, G., Raqib, in seals caused by a novel member of the
R., Rahim, Z., Gerhardt, M., et al. (2007) Mycobacterium tuberculosis complex: Myco-
Operational feasibility of using loop-mediated bacterium pinnipedii sp. nov. International
isothermal amplification for diagnosis of Journal of Systemic Evolutionary Microbiology
pulmonary tuberculosis in microscopy centers 53, 1305–1314.
of developing countries. Journal of Clinical Craw, P. and Balachandran, W. (2012) Isothermal
Microbiology 45, 1936–1940. Nucleic Acid Amplification Technologies for
Point-of-Care Diagnostics: A Critical Review. clinical samples. Journal of Clinical Microbiology

Lab on a Chip 12(14), 2469–2486. 44, 3025–3027.
Cristea-Fernstrom, M., Olofsson, M., Chryssanthou, FIND (2011) Performance of Xpert MTB/RIF
E., Jonasson, J. and Petrini, B. (2007) Pyro- Version G4 assay.
sequencing of a short hypervariable 16S rDNA Flores, L.L., Pai, M., Colford, J.M. Jr and Riley, L.W.
fragment for the identification of non-tuberculous (2005) In-house nucleic acid amplification tests
mycobacteria – a comparison with conventional for the detection of Mycobacterium tuberculosis
16S rDNA sequencing and phenotyping. APMIS in sputum specimens: meta-analysis and meta-
115, 1252–1259. regression. BMC Microbiology 5, 55.
Cui, Z., Wang, Y., Fang, L., Zheng, R., Huang, X., Francois, P., Tangomo, M., Hibbs, J., Bonetti, E.J.,
Liu, X., et al. (2012) Novel real-time simul- Boehme, C.C., Notomi, T., et al. (2011) Robust-
taneous amplification and testing method to ness of a loop-mediated isothermal amplification
accurately and rapidly detect Mycobacterium reaction for diagnostic applications. FEMS
tuberculosis complex. Journal of Clinical Immunology and Medical Microbiology 62,
Microbiology 50, 646–650. 41–48.
Davis, J.L., Huang, L., Worodria, W., Masur, H., Gegia, M., Mdivani, N., Mendes, R.E., Li, H.,
Cattamanchi, A., Huber, C., et al. (2011) Nucleic Akhalaia, M., Han, J., et al. (2008) Prevalence
acid amplification tests for diagnosis of smear- of and molecular basis for tuberculosis drug
negative TB in a high HIV-prevalence setting: a resistance in the Republic of Georgia: validation
prospective cohort study. PLoS One 6, e16321. of a QIAplex system for detection of drug
de la Rua-Domenech, R. (2006) Human Myco- resistance-related mutations. Antimicrobial
bacterium bovis infection in the United Kingdom: Agents and Chemotherapy 52, 725–729.
incidence, risks, control measures and review of Gordon, S.V., Brosch, R., Billault, A., Garnier, T.,
the zoonotic aspects of bovine tuberculosis. Eiglmeier, K. and Cole, S.T. (1999) Identification
Tuberculosis 86, 77–109. of variable regions in the genomes of tubercle
De Wit, D., Steyn, L., Shoemaker, S. and Sogin, M. bacilli using bacterial artificial chromosome
(1990) Direct detection of Mycobacterium arrays. Molecular Microbiology 32, 643–655.
tuberculosis in clinical specimens by DNA Goto, M., Oka, S., Okuzumi, K., Kimura, S. and
amplification. Journal of Clinical Microbiology Shimada, K. (1991) Evaluation of acridinium-
28, 2437–2441. ester-labeled DNA probes for identification of
Del Portillo, P., Thomas, M.C., Martinez, E., Mycobacterium tuberculosis and Mycobacterium
Maranon, C., Valladares, B., Patarroyo, M.E., et avium–Mycobacterium intracellulare complex in
al. (1996) Multiprimer PCR system for culture. Journal of Clinical Microbiology 29,
differential identification of mycobacteria in 2473–2476.
clinical samples. Journal of Clinical Microbiology Greco, S., Girardi, E., Navarra, A. and Saltini, C.
34, 324–328. (2006) Current evidence on diagnostic accuracy
Djelouadji, Z., Raoult, D., Daffé, M. and Drancourt, of commercially based nucleic acid amplification
M. (2008) A single-step sequencing method for tests for the diagnosis of pulmonary
the identification of Mycobacterium tuberculosis tuberculosis. Thorax 61, 783–790.
complex species. PLoS Neglected Tropical Gryadunov, D., Dementieva, E., Mikhailovich, V.,
Disease 2, e253. Nasedkina, T., Rubina, A., Savvateeva, E., et al.
Eiken Chemical Co Ltd (year?) Eiken Genome Site. (2011) Gel-based microarrays in clinical
The Principle of LAMP method. Animation diagnostics in Russia. Expert Review of
(http://loopamp.eiken.co.jp/e/lamp/anim.html, Molecular Diagnostics 11, 839–853.
accessed 30 April 2012). Halse, T.A., Escuyer, V.E. and Musser, K.A. (2011)
Evans, K.D., Nakasone, A.S., Sutherland, P.A., de Evaluation of a single-tube multiplex real-time
la Maza, L.M. and Peterson, E.M. (1992) PCR for differentiation of members of the
Identification of Mycobacterium tuberculosis Mycobacterium tuberculosis complex in clinical
and Mycobacterium avium–M. intracellulare specimens. Journal of Clinical Microbiology 49,
directly from primary BACTEC cultures by using 2562–2567.
acridinium-ester-labeled DNA probes. Journal Helb, D., Jones, M., Story, E., Boehme, C., Wallace,
of Clinical Microbiology 30, 2427–2431. E., Ho, K., et al. (2010) Rapid detection of
Franco-Álvarez de Luna, F., Ruiz, P., Gutiérrez, J. Mycobacterium tuberculosis and rifampin
and Casal, M. (2006) Evaluation of the resistance by use of on-demand, near-patient
GenoType mycobacteria direct assay for technology. Journal of Clinical Microbiology 48,
detection of Mycobacterium tuberculosis com- 229–237.
plex and four atypical mycobacterial species in Hillemann, D., Rüsch-Gerdes, S. and Richter, E.
(2007) Evaluation of the GenoType MTBDRplus Hoffner, S.E., Norberg, R., Svensson, E., et al.
assay for rifampin and isoniazid susceptibility (1999) Evolution and clonal traits of Myco-
testing of Mycobacterium tuberculosis strains bacterium tuberculosis complex in Guinea-
and clinical specimens. Journal of Clinical Bissau. Journal of Clinical Microbiology 37,
Microbiology 45, 2635–2640. 3872–3878.
Hillemann, D., Rüsch-Gerdes, S. and Richter, E. Kasai, H., Ezaki, T. and Harayama, S. (2000)
(2009) Feasibility of the GenoType MTBDRsl Differentiation of phylogenetically related slowly
assay for fluoroquinolone, amikacin-capreo- growing mycobacteria by their gyrB sequences.
mycin, and ethambutol resistance testing of Journal of Clinical Microbiology 38, 301–308.
Mycobacterium tuberculosis strains and clinical Kim, J.H., Kim, Y.J., Ki, C.S., Kim, J.Y. and Lee, N.Y.
specimens. Journal of Clinical Microbiology 47, (2011) Evaluation of Cobas TaqMan MTB PCR
1767–1772. for detection of Mycobacterium tuberculosis.
Hlavsa, M.C., Moonan, P.K., Cowan, L.S., Navin, Journal of Clinical Microbiology 49, 173–176.
T.R., Kammerer, J.S., Morlock, G.P., et al. Kjeldsen, M.K., Bek, D., Rasmussen, E.M., Priemé,
(2008) Human tuberculosis due to Myco- A. and Thomsen, V.Ø. (2009) Line probe assay
bacterium bovis in the United States, 1995– for differentiation within Mycobacterium
2005. Clinical Infectious Diseases 47, 168–175. tuberculosis complex. Evaluation on clinical
Hoorfar, J., Cook, N., Malorny, B., Wagner, M., De specimens and isolates including Myco-
Medici, D., Abdulmawjood, A., et al. (2003) bacterium pinnipedii. Scandinavian Journal of
Making internal amplification control mandatory Infectious Diseases 41, 635–641.
for diagnostic PCR. Journal of Clinical Koeck, J.L., Fabre, M., Simon, F., Daffé, M.,
Microbiology 41, 5835. Garnotel, É., Matan, A.B., et al. (2010) Clinical
Hoorfar, J., Malorny, B., Abdulmawjood, A., Cook, characteristics of the smooth tubercle bacilli
N., Wagner, M. and Fach, P. (2004) Practical ‘Mycobacterium canettii’ infection suggest the
considerations in design of internal amplification existence of an environmental reservoir. Clinical
controls for diagnostic PCR assays. Journal of Microbiology and Infection 17, 1013–1019.
Clinical Microbiology 42, 1863–1868. Kontsevaya, I., Mironova, S., Nikolayevskyy, V.,
Huard, R., de Oliveira Lazzarini, L., Butler, W., van Balabanova, Y., Mitchell, S. and Drobniewski, F.
Soolingen, D. and Ho, J. (2003) PCR-based (2011) Evaluation of two molecular assays for
method to differentiate the subspecies of the rapid detection of Mycobacterium tuberculosis
Mycobacterium tuberculosis complex on the resistance to fluoroquinolones in high-
basis of genomic deletions. Journal of Clinical tuberculosis and -multidrug-resistance settings.
Microbiology 41, 1637–1650. Journal of Clinical Microbiology 49, 2832–2837.
Ignatyeva, O., Kontsevaya, I., Kovalyov, A., Lawn, S.D. and Nicol, M.P. (2011) Xpert(R) MTB/
Balabanova, Y., Nikolayevskyy, V., Toit, K., et al. RIF assay: development, evaluation and
(2012) Detection of resistance to second-line implementation of a new rapid molecular
antituberculosis drugs by use of the Genotype diagnostic for tuberculosis and rifampicin
MTBDRsl assay: a multicenter evaluation and resistance. Future Microbiology 6, 1067–1082.
feasibility study. Journal of Clinical Microbiology Lawn, S.D., Brooks, S.V., Kranzer, K., Nicol, M.P.,
50, 1593–1597. Whitelaw, A., Vogt, M., et al. (2011) Screening
Isola, D., Pardini, M., Varaine, F., Niemann, S., for HIV-associated tuberculosis and rifampicin
Rüsch-Gerdes, S., Fattorini, L., et al. (2005) A resistance before antiretroviral therapy using
pyrosequencing assay for rapid recognition of the Xpert MTB/RIF assay: a prospective study.
SNPs in Mycobacterium tuberculosis embB306 PLoS Medicine 8, e1001067.
region. Journal of Microbiological Methods 62, Lin, S.Y., Probert, W., Lo, M. and Desmond, E.
113–120. (2004) Rapid detection of isoniazid and rifampin
Iwamoto, T., Sonobe, T. and Hayashi, K. (2003) resistance mutations in Mycobacterium
Loop-mediated isothermal amplification for tuberculosis complex from cultures or smear-
direct detection of Mycobacterium tuberculosis positive sputa by use of molecular beacons.
complex, M. avium, and M. intracellulare in Journal of Clinical Microbiology 42, 4204–4208.
sputum samples. Journal of Clinical Ling, D.I., Flores, L.L., Riley, L.W. and Pai, M.
Microbiology 41, 2616–2622. (2008a) Commercial nucleic-acid amplification
Kahla, I.B., Henry, M., Boukadida, J. and Drancourt, tests for diagnosis of pulmonary tuberculosis in
M. (2011) Pyrosequencing assay for rapid respiratory specimens: meta-analysis and
identification of Mycobacterium tuberculosis meta-regression. PLoS One 3, e1536.
complex species. BMC Research Notes 4, 423. Ling, D.I., Zwerling, A.A. and Pai, M. (2008b)
Kallenius, G., Koivula, T., Ghebremichael, S., GenoType MTBDR assays for the diagnosis of
multidrug-resistant tuberculosis: a meta- samples. Journal of Clinical Microbiology 47,

analysis. European Respiratory Journal 32, 2601–2603.
1165–1174. Neonakis, I.K., Spandidos, D.A. and Petinaki, E.
Luo, T., Jiang, L., Sun, W., Fu, G., Mei, J. and Gao, (2011) Use of loop-mediated isothermal
Q. (2011) Multiplex real-time PCR melting curve amplification of DNA for the rapid detection of
assay to detect drug-resistant mutations of Mycobacterium tuberculosis in clinical speci-
Mycobacterium tuberculosis. Journal of Clinical mens. European Journal of Clinical Microbiology
Microbiology 49, 3132–3138. and Infectious Diseases 30, 937–942.
Marlowe, E.M., Novak-Weekley, S.M., Cumpio, J., Nicol, M.P., Workman, L., Isaacs, W., Munro, J.,
Sharp, S.E., Momeny, M.A., Babst, A., et al. Black, F., Eley, B., et al. (2011) Accuracy of the
(2011) Evaluation of the Cepheid Xpert MTB/ Xpert MTB/RIF test for the diagnosis of
RIF assay for direct detection of Mycobacterium pulmonary tuberculosis in children admitted to
tuberculosis complex in respiratory specimens. hospital in Cape Town, South Africa: a
Journal of Clinical Microbiology 49, 1621–1623. descriptive study. The Lancet Infectious
Mauk, M.G., Liu, C.-C., Corstjens, P.L.A.M. and Diseases 11, 819–824.
Bau, H.H. (2011) Streamlining POC microfluidic Notomi, T., Okayama, H., Masubuchi, H., Yonekawa,
molecular diagnostics. IVD Technology (http:// T., Watanabe, K., Amino, N., et al. (2000) Loop-
www.ivdtechnology.com/article/streamlining- mediated isothermal amplification of DNA.
poc-microfluidic-molecular-diagnostics, Nucleic Acids Research 28, E63.
accessed 30 April 2012). Nyendak, M.R., Lewinsohn, D.A. and Lewinsohn,
Migliori, G.B., Dheda, K., Centis, R., Mwaba, P., D.M. (2009) New diagnostic methods for
Bates, M., O’Grady, J., et al. (2010) Review of tuberculosis. Current Opinion in Infectious
multidrug-resistant and extensively drug- Diseases 22, 174–182.
resistant TB: global perspectives with a focus on O’Grady, J., Maeurer, M., Mwaba, P., Kapata, N.,
sub-Saharan Africa. Tropical Medicine and Bates, M., Hoelscher, M., et al. (2011) New and
International Health 15(9), 1052–1066. improved diagnostics for detection of drug-
Mitarai, S., Okumura, M., Toyota, E., Yoshiyama, T., resistant pulmonary tuberculosis. Current
Aono, A., Sejimo, A., et al. (2011) Evaluation of Opinion in Pulmonary Medicine 17, 134–141.
a simple loop-mediated isothermal amplification Palomino, J.C. (2009) Molecular detection,
test kit for the diagnosis of tuberculosis. The identification and drug resistance detection in
International Journal of Tuberculosis and Lung Mycobacterium tuberculosis. FEMS Immuno-
Disease 15, 1211–1217, i. logy and Medical Microbiology 56, 103–111.
Morgan, M., Kalantri, S., Flores, L. and Pai, M. Parsons, L.M., Brosch, R., Cole, S.T., Somoskovi,
(2005) A commercial line probe assay for the A., Loder, A., Bretzel, G., et al. (2002) Rapid
rapid detection of rifampicin resistance in and simple approach for identification of
Mycobacterium tuberculosis: a systematic Mycobacterium tuberculosis complex isolates
review and meta-analysis. BMC Infectious by PCR-based genomic deletion analysis.
Diseases 5, 62. Journal of Clinical Microbiology 40, 2339–2345.
Mostowy, S., Cousins, D., Brinkman, J., Aranaz, A. Parsons, L.M., Somoskovi, A., Gutierrez, C., Lee,
and Behr, M. (2002) Genomic deletions suggest E., Paramasivan, C.N., Abimiku, A., et al. (2011)
a phylogeny for the Mycobacterium tuberculosis Laboratory diagnosis of tuberculosis in
complex. Journal of Infectious Diseases 186, resource-poor countries: challenges and
74–80. opportunities. Clinical Microbiology Reviews 24,
Moure, R., Munoz, L., Torres, M., Santin, M., 314–350.
Martin, R. and Alcaide, F. (2011) Rapid detection Piersimoni, C. and Scarparo, C. (2003) Relevance
of Mycobacterium tuberculosis complex and of commercial amplification methods for direct
rifampin resistance in smear-negative clinical detection of Mycobacterium tuberculosis com-
samples by use of an integrated real-time PCR plex in clinical samples. Journal of Clinical
method. Journal of Clinical Microbiology 49, Microbiology 41, 5355–5365.
1137–1139. Pinsky, B. and Banaei, N. (2008) Multiplex real-time
Neonakis, I.K., Gitti, Z., Baritaki, S., Petinaki, E., PCR assay for rapid identification of
Baritaki, M. and Spandidos, D.A. (2009) Mycobacterium tuberculosis complex members
Evaluation of GenoType mycobacteria direct to the species level. Journal of Clinical
assay in comparison with Gen-Probe Myco- Microbiology 46, 2241–2246.
bacterium tuberculosis amplified direct test and Pounder, J.I., Anderson, C.M., Voelkerding, K.V.,
GenoType MTBDRplus for direct detection of Salfinger, M., Dormandy, J., Somoskovi, A., et
Mycobacterium tuberculosis complex in clinical al. (2010) Mycobacterium tuberculosis complex
differentiation by genomic deletion patterns with segment of the gene coding for the 32-kilodalton
multiplex polymerase chain reaction and melting protein. Journal of Clinical Microbiology 30,
analysis. Diagnostic Microbiology and Infectious 2025–2028.
Disease 67, 101–105. Somoskovi, A., Dormandy, J., Rivenburg, J.,
Rachow, A., Zumla, A., Heinrich, N., Rojas-Ponce, Pedrosa, M., McBride, M. and Salfinger, M.
G., Mtafya, B., Reither, K., et al. (2011) Rapid (2008) Direct comparison of the Genotype
and accurate detection of Mycobacterium MTBC and genomic deletion assays in terms of
tuberculosis in sputum samples by Cepheid their ability to distinguish between members of
Xpert MTB/RIF assay – a clinical validation the Mycobacterium tuberculosis complex in
study. PLoS One 6, e20458. clinical isolates and in clinical specimens.
Ramirez, M.V., Cowart, K.C., Campbell, P.J., Journal of Clinical Microbiology 46, 1854–1857.
Morlock, G.P., Sikes, D., Winchell, J.M., et al. Somoskovi, A., Dormandy, J., Mayrer, A.R., Carter,
(2010) Rapid detection of multidrug-resistant M., Hooper, N. and Salfinger, M. (2009)
Mycobacterium tuberculosis by use of real-time Mycobacterium canettii isolated from a human
PCR and high-resolution melt analysis. Journal immunodeficiency virus-positive patient: first
of Clinical Microbiology 48, 4003–4009. case recognized in the United States. Journal of
Reddington, K., O’Grady, J., Dorai-Raj, S., Maher, Clinical Microbiology 47, 255–257.
M., van Soolingen, D. and Barry, T. (2011) Novel Syre, H., Myneedu, V.P., Arora, V.K. and Grewal,
multiplex real-time PCR diagnostic assay for H.M.S. (2009) Direct detection of mycobacterial
identification and differentiation of Myco- species in pulmonary specimens by two rapid
bacterium tuberculosis, Mycobacterium canettii, amplification tests, the Gen-Probe amplified
and Mycobacterium tuberculosis complex Mycobacterium tuberculosis direct test and the
strains. Journal of Clinical Microbiology 49, GenoType mycobacteria direct test. Journal of
651–657. Clinical Microbiology 47, 3635–3639.
Reddington, K., Zumla, A., Bates, M., van Tomita, N., Mori, Y., Kanda, H. and Notomi, T.
Soolingen, D., Niemann, S., Barry, T., et al. (2008) Loop-mediated isothermal amplification
(2012) SeekTB – a two stage multiplex real-time (LAMP) of gene sequences and simple visual
PCR based method for the differentiation of the detection of products. Nature Protocols 3, 877–
Mycobacterium tuberculosis complex. Journal 882.
of Clinical Microbiology 50(7), 2203–2206. Torres-Chavolla, E. and Alocilja, E.C. (2011) Nano-
Richter, E., Weizenegger, M., Fahr, A.-M. and particle based DNA biosensor for tuberculosis
Rusch-Gerdes, S. (2004) Usefulness of the detection using thermophilic helicase-
GenoType MTBC assay for differentiating dependent isothermal amplification. Biosensors
species of the Mycobacterium tuberculosis and Bioelectronics 26, 4614–4618.
complex in cultures obtained from clinical Van Rie, A., Page-Shipp, L., Scott, L., Sanne, I. and
specimens. Journal of Clinical Microbiology 42, Stevens, W. (2010) Xpert(R) MTB/RIF for point-
4303–4306. of-care diagnosis of TB in high-HIV burden,
Rossau, R., Traore, H., De Beenhouwer, H., Mijs, resource-limited countries: hype or hope?
W., Jannes, G., De Rijk, P., et al. (1997) Expert Review of Molecular Diagnostics 10,
Evaluation of the INNO-LiPA Rif. TB assay, a 937–946.
reverse hybridization assay for the simultaneous Van Rie, A., Mellet, K., John, M.A., Scott, L., Page-
detection of Mycobacterium tuberculosis com- Shipp, L., Dansey, H., et al. (2012) False-
plex and its resistance to rifampin. Antimicrobial positive rifampicin resistance on Xpert(R) MTB/
Agents and Chemotherapy 41, 2093–2098. RIF: case report and clinical implications. The
Sandgren, A., Strong, M., Muthukrishnan, P., International Journal of Tuberculosis and Lung
Weiner, B.K., Church, G.M. and Murray, M.B. Disease 16, 206–208.
(2009) Tuberculosis drug resistance mutation Verma, A., Rattan, A. and Tyagi, J.S. (1994)
database. PLoS Medicine 6, e1000002. Development of a 23S rRNA-based PCR assay
Scott, L.E., McCarthy, K., Gous, N., Nduna, M., Van for the detection of mycobacteria. Indian Journal
Rie, A., Sanne, I., et al. (2011) Comparison of of Biochemistry and Biophysics 31, 288–294.
Xpert MTB/RIF with other nucleic acid Wada, T., Maeda, S., Tamaru, A., Imai, S., Hase, A.
technologies for diagnosing pulmonary and Kobayashi, K. (2004) Dual-probe assay for
tuberculosis in a high HIV prevalence setting: a rapid detection of drug-resistant Mycobacterium
prospective study. PLoS Medicine 8, e1001061. tuberculosis by real-time PCR. Journal of
Soini, H., Skurnik, M., Liippo, K., Tala, E. and Clinical Microbiology 42, 5277–5285.
Viljanen, M.K. (1992) Detection and identifi- Walker, D.A., Taylor, I.K., Mitchell, D.M. and Shaw,
cation of mycobacteria by amplification of a R.J. (1992a) Comparison of polymerase chain
reaction amplification of two mycobacterial DNA DNA Test. WHO Endorsement and Recom-
sequences, IS6110 and the 65kDa antigen mendations. World Health Organization,
gene, in the diagnosis of tuberculosis. Thorax Geneva, Switzerland.
47, 690–694. WHO (2011a) Automated Real-time Nucleic Acid
Walker, G.T., Fraiser, M.S., Schram, J.L., Little, Amplification Technology for Rapid and Simul-
M.C., Nadeau, J.G. and Malinowski, D.P. taneous Detection of Tuberculosis and Rifam-
(1992b) Strand displacement amplification – an picin Resistance: Xpert MTB/RIF System. World
isothermal, in vitro DNA amplification technique. Health Organization, Geneva, Switzerland.
Nucleic Acids Research 20, 1691–1696. WHO (2011b) Global Tuberculosis Control: WHO
Wallis, R.S., Pai, M., Menzies, D., Doherty, T.M., Report 2011. World Health Organization,
Walzl, G., Perkins, M.D., et al. (2010) Biomarkers Geneva, Switzerland.
and diagnostics for tuberculosis: progress, WHO (2011c) Rapid Implementation of the Xpert
needs, and translation into practice. The Lancet MTB/RIF Dianognostic Test: Technicial and
375, 1920–1937. Operational ‘How to’: Practical Considerations.
Warren, R.M., Gey van Pittius, N.C., Barnard, M., World Health Organization, Geneva, Switzer-
Hesseling, A., Engelke, E., de Kock, M., et al. land.
(2006) Differentiation of Mycobacterium Yang, Y.C., Lu, P.L., Huang, S.C., Jenh, Y.S., Jou, R.
tuberculosis complex by PCR amplification of and Chang, T.C. (2011) Evaluation of the Cobas
genomic regions of difference. International TaqMan MTB test for direct detection of
Journal of Tuberculosis and Lung Disease 10, Mycobacterium tuberculosis complex in
818–822. respiratory specimens. Journal of Clinical
WHO [World Health Organization] (2008) Molecular Microbiology 49, 797–801.
Line Probe Assays for Rapid Screening of Yesilkaya, H., Meacci, F., Niemann, S., Hillemann,
Patients at Risk of Multidrug-resistant Tubercu- D., Rüsch-Gerdes, S., Group, L.D.S., et al.
losis (MDR-TB). World Health Organization, (2006) Evaluation of molecular-beacon,
Geneva, Switzerland. TaqMan, and fluorescence resonance energy
WHO (2010a) Roadmap for Rolling Out Xpert MTB/ transfer probes for detection of antibiotic
RIF for Rapid Diagnosis of TB and MDR-TB. resistance-conferring single nucleotide poly-
World Health Organization, Geneva, Switzer- morphisms in mixed Mycobacterium tubercu-
land. losis DNA extracts. Journal of Clinical
WHO (2010b) Tuberculosis Diagnostics Automated Microbiology 44, 3826–3829.
3 Improving on the LJ Slope – Automated
Liquid Culture
Miguel Viveiros,1,2 Diana Machado,1,3 Isabel Couto1,4 and

Leonard Amaral1,2,3
1Unit of Mycobacteriology, Instituto de Higiene e Medicina Tropical, Universidade
Nova de Lisboa (IHMT/UNL), Lisbon, Portugal; 2Cost Action BM0701 (ATENS) of

the Cost Action of the European Commission; 3UPMM, IHMT/UNL; 4Centro de
Recursos Microbiológicos (CREM), Faculdade de Ciências e Tecnologia, UNL,
Caparica, Portugal
3.1 Introduction and the guidelines recommended by the

Center for Disease Control (CDC, USA) for
The major priority for tuberculosis control the management of individual cases of active
programmes is the rapid and accurate pulmonary tuberculosis that differ with
diagnosis and treatment of individuals with respect to underlying medical conditions
active tuberculosis. This approach is intended (pregnancy, chronic liver disease, diabetes,
to cut the chain of transmission of the bacillus HIV/AIDS), status of infection, the manage-
and provide appropriate therapy to the ment of patients who have had close contact
tuberculosis patient. In this context, the with individuals known to be infected with
mycobacteriology laboratory has a crucial multidrug-resistant strains of M. tuberculosis
role in the definitive diagnosis and therapy of (preventive) and the management of patients
tuberculosis that is based on the isolation, with reactive tuberculosis indicate how
identification and antibiotic susceptibility complex the management of tuberculosis has
profile of the infecting organism. Prior to the become (ATS/CDC/IDSA, 2003; Griffith et al.,
emergence of drug-resistant strains, the role 2007; WHO, 2009). Underlying these recom-
of the laboratory was not considered as mendations is the need for the accurate isola-
important as it is today. At that time, the tion, identification and antibiotic sensitivity
demonstration of acid-fast bacteria in the testing of M. tuberculosis from clinical speci-
sputum, followed by the culture of the mens in the shortest time possible (Salfinger
specimen and the identification of Myco- and Pfyffer, 1994; Ridderhof et al., 2007;
bacterium tuberculosis was sufficient to support Tortoli and Marcelli, 2007).
the clinical diagnosis of pulmonary tubercu- New and more rapid techniques for the
losis, which, for all practical purposes, was laboratory diagnosis of mycobacterial infec-
already confidently established and sup- tions were, therefore, greatly needed if the
ported by physical clinical findings (Salfinger, clinical mycobacteriology laboratory was to
1977). Nowadays, the examination of the serve a useful purpose in the management of
International Classification of Tuberculosis tuberculosis patients, especially in areas of

34 (ed. T.D. McHugh)
Improving on the LJ slope – Automated Liquid Culture 35
the world that had large numbers of 1990s, to the development of non-radiometric
coinfected HIV/AIDS patients (WHO, 2008). broth technologies, such as the fluorescent
Despite all the technological advancements in BACTEC™ 9000MB system (Becton Dickin-
the area, culture remains the gold standard son), the fluorescent mycobacterial growth
for the detection of mycobacteria. Therefore, indicator tubes manual system BACTEC™
all clinical specimens suspected of containing MGIT™ (Becton Dickinson) and the colori-
mycobacteria should be inoculated on to metric MB/BacT®(Organon Teknika, Boxtel,
culture media. The reasons for this procedure Netherlands), later upgraded to the BacT/
are related to three main aspects: first, culture ALERT®Systems (bioMérieux SA, Marcy-
is more sensitive than microscopy, being able l’Etoile, France). Nevertheless, the Center for
to detect less than ten bacteria/ml in the Disease Control and many reference labora-
clinical sample; second, the isolation of tories around the world still maintain the
mycobacteria in culture is indispensable for recommendation that broth and agar media
proper identification of the species; and third, should be employed as a complement for the
antibiotic susceptibility tests require viable recovery of mycobacteria and further recom-
microorganisms. However, culture is a slow mend the solid egg-based (LJ) and agar-based
and expensive method, since the majority of (7H10 or 7H11) media as gold standards
these microorganisms do not grow in com- (Kent and Kubica, 1985; David, 1989; Tortoli
mon culture media, requiring the employment and Marcelli, 2007).
of specific media. The main disadvantage of For the recovery and isolation of myco-
the culture-based methods for isolation and bacteria from biological samples, three main
identification of mycobacteria is related with types of media can be employed: egg-based
the time that is required to obtain results. medium, medium with agar and liquid
Mycobacteria are a slow-growing group of medium. The introduction of antimicrobial
bacteria, with the generation time varying agents (selective media) may be very useful
with species from 2 (M. smegmatis) to more for contaminated samples, but this type of
than 20 h (M. tuberculosis). Regarding M. media should not be used alone due to the
ulcerans, for example, the primary cultures probability of inhibiting mycobacterial
are usually positive within 6–12 weeks of growth.
incubation at 29–33°C, but much longer
incubation (9 months or more) may be
necessary for some isolates (David, 1989;
3.2 Types of Media for Culture and
Griffith et al., 2007; Palomino et al., 2007).
Isolation of Mycobacteria
Growth-based methods for mycobacterial
isolation, identification and drug suscepti-
There are two types of solid media for the
bility have been in practice since the 1930s,
recovery and isolation of mycobacteria: egg-
when the mycobacteria selective egg-based
based (e.g. LJ and Ogawa) and agar-based
Löwenstein–Jensen (LJ) medium was devel-
(Middlebrook 7H10 and 7H11) (Figs 3.1 and
oped (Löwenstein, 1931; Kent and Kubica,
3.2). Among the solid culture media used for
1985). The commercial preparation of LJ
growth of mycobacteria, the LJ medium is the
medium was pioneered by Becton Dickinson
one mostly used.
(Towson, USA) in 1952, followed by agar-
based Middlebrook 7H10 (Middlebrook and
Cohn, 1958). The 1980s were marked by the
3.2.1 Solid media
development of the broth-based BACTEC™
460TB Radiometric System by Becton Dickin-
Egg-based medium
son using Middlebrook 7H12 (BACTEC 12B
vials) for specimens other than blood and Löwenstein–Jensen (LJ), an egg-based
Middlebrook 7H13 (BACTEC 13A vials) for medium, is a modification of the original
blood specimens (Strand et al., 1989). The Löwenstein medium (Löwenstein, 1931) by
strict safety regulations concerning the use of Jensen (Jensen, 1932) (Fig. 3.1). In this version
radiometric labelled material led, in the of the media firstly described by Lowestein,
36 M. Viveiros et al.
the Congo red dye was suppressed, the 5% of sodium chloride used to characterize
concentration of malachite green was some species of mycobacteria and the LJ
increased and the contents of citrate and medium deep tubes for catalase semi-
phosphate were altered. Overall, the medium quantitative assay. The Ogawa medium is
contains sources of nitrogen and vitamins another egg-based medium, which is com-
(asparagine and potato flour), enhancers of parable with LJ in its composition. It is a
growth (monopotassium phosphate and cheaper alternative to the LJ medium due to
magnesium sulfate), fatty acids and proteins the substitution of the amino acid asparagine
(glycerol and eggs) necessary for myco- by sodium glutamate. The two media also
bacterial metabolism; sodium citrate and differ in malachite green concentration,
malachite green are used as selective agents volume of egg homogenate and pH.
(Table 3.1). LJ is used for primary isolation of The two major advantages of the egg-
mycobacteria from sterile and non-sterile based media are related to the fact that they
sources. Besides the application of this type allow the growth of a wide variety of
of media for the isolation and cultivation of mycobacterial species and that this growth
mycobacteria, it is also used as a base for the can be used for biochemical tests, like niacin
selection and differentiation of mycobacteria. and catalase production testing, for species
Currently available modifications of the LJ differentiation. Besides these advantages, the
medium include the LJ without glycerol but media are easy to prepare, are the least
supplemented with pyruvate to improve the expensive of all media available for
growth of M. bovis and M. africanum; the mycobacteria culture, may be stored in the
Gruft modified LJ medium, which contains refrigerator for several weeks and possess
penicillin and nalidixic acid for a more low rates of contamination due to the addition
selective isolation of mycobacteria, the BBL™ of malachite green, which suppresses the
Mycobactosel™ Löwenstein–Jensen Medium growth of other non-mycobacterial organ-
Slants (from Becton Dickinson) supplemented isms. The major disadvantage of this type of
with cycloheximide, lincomycin and nalidixic medium is related to the time necessary for
acid for specimens particularly contaminated; the cultures to become positive, especially for
the LJ medium with iron used for the samples that contain few bacilli (either from
determination of mycobacterial iron uptake, paubacillary samples or as the result of an
which allows the differentiation and identifi- aggressive decontamination process). In
cation of some species; the LJ medium with addition, the contamination of the medium
(a) (b )
Fig. 3.1. Growth of different mycobacteria in Löwestein–Jensen media. (a) Mycobacterium marinum,
M. gordonae and M. tuberculosis; (b) M. tuberculosis and M. intracellulare.
Table 3.1. Most commonly used commercially available media for cultivation of mycobacteria from
clinical specimens.
Culture media Composition
Liquid media
Difco™ Middlebrook 7H9 Broth Ammonium sulfate; L-glutamic acid; sodium citrate; pyridoxine;
(Middlebrook 7H9) biotin; disodium phosphate; monopotassium phosphate; ferric
ammonium citrate; magnesium sulfate; calcium chloride; zinc
sulfate; copper sulfate
BBL MGIT Mycobacteria Growth Modified Middlebrook 7H9 broth base; tris 4, 7-diphenyl-1,
Indicator Tube 10-phenanthroline ruthenium chloride pentahydrate as a
fluorescent indicator; casein peptone
BACTECTM 12B Middlebrook 7H9 broth base; casein hydrolysate; bovine serum
(Middlebrook 7H12) albumin; 48,000 units of catalase; 1000 Curies 14C-substrate
(palmitic acid)
BACTECTM 13A Middlebrook 7H9 broth base; casein hydrolysate; bovine serum
(Middlebrook 7H13) albumin; sodium polyanetholesulfonate; polysorbate 80; 1440
units of catalase; 5 Curies 14C-substrate (palmitic acid)
Solid media
BBL™ Löwenstein–Jensen Asparagine; monopotassium phosphate; magnesium sulfate;
Medium magnesium citrate; potato flour; malachite green; glycerol; whole
egg
Difco™ Middlebrook 7H10 Agar Ammonium sulfate; monopotassium phosphate; disodium
(Middlebrook 7H10) phosphate; sodium citrate; magnesium sulfate; calcium chloride;
zinc sulfate; copper sulfate; L-glutamic acid; ferric ammonium
citrate; pyridoxine; biotin; malachite green; agar
Difco™ Mycobacteria 7H11 Agar Magnesium sulfate; ferric ammonium citrate; sodium citrate;
(Middlebrook 7H11) ammonium sulfate; disodium phosphate; monopotassium
phosphate; agar; pyridoxine; biotin; malachite green; pancreatic
digest of casein; L-glutamic acid
with other microorganisms may promote the When combined with a supplement, Middle-
liquefaction of the medium, which will lead brook enrichment medium (OADC), which
to culture loss. Mycobacterial growth can be consists of oleic acid, bovine albumin,
detected in less than 3 weeks, but requires an dextrose and catalase, it can be used for the
incubation of 8 weeks before the samples can qualitative procedures of isolation and culti-
be classified as negative. Their use for the vation of mycobacteria. Middlebrook 7H11
determination of drug susceptibility of agar is a modification of Middlebrook 7H10
mycobacteria differs and requires careful agar enriched by the addition of 0.1%
adjustment of drug concentrations (Kent and hydrolysed casein, which stimulates the
Kubica, 1985; David, 1989). growth of fastidious strains of M. tuberculosis
and improves susceptibility testing. The use
of agar-based media for the isolation of
Agar-based medium
mycobacteria has significant advantages: it
In 1958, Middlebrook and Cohn described a does not liquefy in the presence of proteolytic
medium based on agar that allowed a more organisms and it is translucent, which allows
rapid detection of the mycobacterial growth the visualization of growth and colony
(Middlebrook and Cohn, 1958) (Fig. 3.2). morphology with a stereo microscope even in
These types of media are available in two the presence of contaminating microorganisms
formulations, the Middlebrook 7H10 and the within a few days of incubation (10–12 days).
Middlebrook 7H11 agar, and contain a variety Besides this, susceptibility testing can be
of inorganic salts, vitamins, cofactors, performed on agar-based medium without
glycerol, malachite green and agar (Table 3.1). changing the concentration of the antibiotic,
Fig. 3.2. Growth of M.

tuberculosis in Middlebrook
7H11 medium after three
weeks at 37°C.
as is the case with the egg-based media. It is widely used for culturing mycobacteria, as
important to note that this type of media is recommended by the International Union
extremely sensitive, therefore it should be Against Tuberculosis and the WHO, liquid
prepared in small quantities at a time to avoid media have also been developed in order to
loss of quality; exposure to daylight or heat hasten the growth of mycobacteria in culture.
results in the release of formaldehyde from Examples such as the liquid Dubos’s medium
the OADC supplement with a sufficiently and Sauton’s medium, for general growth
high concentration to inhibit the growth of purposes, or the Proskauer and Beck’s
the mycobacteria. For this reason, it is medium, ideal for growth in high quantities
important not to heat the OADC (stored at and surface-pellicle formation, or M9 defined
4°C) but to allow it to warm to room media for protein isolation have been widely
temperature before adding to the agar. Also, used for specific research purposes (Parish
the use of agar-based media is considerably and Stoker, 1998). These have been the source
more expensive than egg-based media. for further improvement of the original
Several alternative solid media exist for Middlebrook formulation, which later led to
specific research purposes, such as phage the commercial liquid-based systems for
infection, Bacillus Calmette–Guérin (BCG) mycobacterial growth detection that are
production or protein purification, namely nowadays commonly used worldwide for the
the BCG, Top and Tryptic Soy Agar (Kent and routine mycobacteria laboratory (Kent and
Kubica, 1985; David, 1989; Parish and Stoker, Kubica, 1985; David, 1989; Palomino et al.,
1998; Palomino et al., 2007). 2007).
The time required for the detection of The commercial liquid-based culture
mycobacterial growth can be reduced by the systems can be manual, semi-automated or
use of liquid media for the recovery and automated using radiometric, colorimetric or
primary isolation of mycobacteria, which fluorimetric detection methods. The liquid
provides a more rapid diagnosis. media generally used is the Middlebrook 7H9
or the Middlebrook 7H12/7H13 (Table 3.1),
depending on the system and the source of
3.2.2 Liquid media the specimen. Examples of these systems
include the BACTEC™ 460TB, BACTEC™
While the LJ medium and the Middlebrook MB9000, BACTEC™ MGIT™ 960 or 320 and
7H10/7H11 agar are the solid media most the manual MGIT, the Septi-Chek AFB™
systems (all from Becton Dickinson Instru- inoculation of the sample, an antibiotic and
ment Systems, Sparks, Maryland, USA), the enrichment supplement are added and the
ESP® (Extra Sensing Power) Myco- bottle is subsequently inverted to inoculate
ESPculture System II® (Trek Diagnostic Sys- the solid medium. Bacterial growth is
tems, USA) and the BacT/ALERT MB® detected by observing the surface of the solid
(bioMérieux SA) (Salfinger and Pfyffer, 1994; media and the opacity of broth media. This
Perkins, 2000; Watterson and Drobniewski, system does not require specific instru-
2000; Fig. 3.3). mentation and is non-radioactive. However,
it is a labour-intensive, manual method.
3.2.3 Manual culture systems BBL™ MGIT™ mycobacterial growth

indicator tube
Septi-Chek™ AFB
This is a rapid and simple manual fluori-
The BBL™ Septi-Chek™ AFB is a biphasic metric method to detect mycobacterial
culture system (Becton Dickinson Micro- growth that uses BBL MGIT tubes with 4 ml
biology Systems) consisting of modified of a modified Middlebrook 7H9 media and a
Middlebrook 7H9 broth and three specific gel with a fluorochrome (ruthenium)
solid media. The BBL™ Septi-Chek™ AFB impregnated at the bottom. Using a Wood’s
Supplement and BBL™ Septi-Chek™ AFB lamp or other long-wave UV light source,
slides are used in combination with the tubes are read daily. Positive tests emit an
culture for the detection and isolation of intense orange fluorescent light at the
mycobacteria. The system consists of a bottle bottom of the tube and at the meniscus;
of modified Middlebrook 7H9 media and a negative tests show slight or no fluorescence.
slide-plate containing chocolate agar (for This method does not require specific
isolation of bacteria other than mycobacteria), instrumentation or a dedicated incubator,
LJ and Middlebrook 7H11 media (for the does not use needles or radioactivity, uses
growth of most mycobacteria). Prior to the plastic tubes that afford better security and
Fig. 3.3. Liquid media used for growth of mycobacteria.

represents an option for laboratories whose dilution fluid supplement PANTA. The
volume of work is small (Tortoli et al., 1997; BACTEC™ 460 is a semi-automated system
Watterson and Drobniewski, 2000). because the bottles incubated outside the
instrument need to be placed into the
BACTEC™ instrument, the instrument
3.2.4 Semi-automated culture systems started, calibrated and the samples read daily
and approximately at the same time by the
BACTEC™ 460TB system operator. Once placed into the system,
reading of the bottles is automated. This
The first semi-automated radiometric system
system enhances the probability of myco-
to be developed was the BACTEC™ 460TB
bacterial recovery and reduces the time
(Becton Dickinson) based on the works
required for mycobacterial growth when
performed by Cummings et al. and Middle-
compared with conventional media used for
brook et al. (Cummings et al., 1975; Middle-
isolation. With the BACTEC™ 460TB, myco-
brook et al., 1977) and whose manufacture
bacterial growth can be detected in biological
was discontinued in 2009. This system was
samples in less than 2 weeks. The disadvan-
used for isolation of mycobacteria, differen-
tages of this method are the costs associated
tiation of the M. tuberculosis complex and
with the acquisition of the instrument and the
susceptibility testing. Glass vials containing 4
need for regular maintenance, the inability to
ml of medium MB 7H12 or MB 7H13 (for
observe colony morphology and detect mixed
blood samples) (BACTEC 12B or 13A bottles)
cultures, the overgrowth of contaminating
were used for mycobacterial growth. The
bacteria, the risk associated with the use of
medium contained a radioactive substrate,
radioactive materials and the safety measures
palmitic acid labelled with 14C, which was
necessary for their discard and the extensive
metabolized by mycobacteria. Growth was
use of needles. This system can be used to
detected by measuring the release of
perform susceptibility testing of M. tubercu-
radioactive CO2 with the aid of a gas flow
losis strains.
radio counter (Siddiqi et al., 1981, 1985). The
BACTEC™ 460TB system reads and records
the radioactivity released by converting it
3.2.5 Automated culture systems
into a growth index (GI), in a scale ranging
from 0 to 999. Since the quantity of 14CO2
BACTEC™ MGIT™
released is a direct measure of respiration, it
can be correlated with a growth standard This is a totally automated fluorimetric-
curve when the values of radioactivity culture system for isolation of mycobacteria
released are plotted against time (Siddiqi et and is currently the most widely used. The
al., 1981, 1985). The bottles were incubated at incubation and reading of cultures does
37°C and the 14CO2 in the vial atmosphere not require the intervention of the operator.
sampled with an injection system with All the data obtained with the system can
needles; the counts per minute of 14CO2 were be stored and processed when equipped
determined by the instrument and converted with EpiCenter software (Becton Dickinson).
into GI units. The introduction of 5% CO2 For mycobacterial growth, the previously
during each reading will enhance the growth described growth-indicator tubes (BBL MGIT,
of the mycobacteria present. To suppress the mycobacteria growth indicator tube) are
growth of microbial contaminants, an anti- used, now containing 7 ml of modified
microbial supplement BACTEC™ PANTA™ Middlebrook 7H9 media. Ruthenium gel, in
PLUS kit consisting of PANTA (polymyxin B, the base of the tube, becomes increasingly
amphotericin B, nalidixic acid, trimethoprim fluorescent due to the reduction of the
and azlocillin) (Becton Dickinson) is added. fluorochrome as oxygen levels dissolved in
In order to enhance the growth of myco- the media drop. Thus, fluorescence intensity
bacteria, the growth promoter (POES, poly- is directly proportional to the amount of
oxyethylene stearate) is incorporated in the oxygen consumed. The MGIT tubes are
usually incubated at 37°C in the system and activity by the growing microorganisms.
the monitoring of growth by the detection of Each bottle possesses a modified Middlebrook
increased fluorescence is performed under 7H9 broth and cellulose sponges. The sponges
UV light every 60 min and converted into provide a growth support platform (Tortoli et
growth units (GUs). A positive result is al., 1998). Prior to inoculation of the specimen,
achieved at 75 GU. The time to detection the medium in each bottle is supplemented
(TTD, also known as time to positivity, TTP) with an antibiotic-containing reagent (PVNA;
for a positive culture that is obtained in the polymyxin B, vancomycin, nalidixic acid and
form of a report is provided by the system. amphotericin B) and a growth supplement.
Incubation and monitoring are carried out up This is a non-radiometric method that
to 45 days, after which cultures are reported eliminates the need for the treatment and
as negative and discarded. Since it is a culture discarding of radioactive material. The closed
system based on liquid media, it is essential system design allows the safe handling of all
to add an antimicrobial supplement (MGIT specimens throughout the laboratory. The
PANTA™) to the primary culture to suppress system is available with different capacities,
the growth of other bacterial species. The that is, for 128, 256 or 384 unitary bottles,
system has significant advantages since it is depending on the needs of each laboratory.
totally automated, has a low rate of
contamination, reduces TTD of mycobacterial
BacT/ALERT mycobacteria detection
growth to 1 week and can also be used to
systems
perform susceptibility testing of M. tubercu-
losis strains, both to first- and second-line The BacT/ALERT systems (bioMérieux SA),
antibiotics. The main disadvantage of this formerly MB/BacT (Organon Teknika, Boxtel,
system is related to the high cost of the Netherlands), are automated microbiology
equipment, reagents and media necessary for systems which use a patented colorimetric
the tests. The system is available in two sensor and constitute a non-isotopic alter-
models, both totally automated; the first to be native for the detection of mycobacteria. The
developed is the BACTEC™ MGIT™ 960 and Organon Teknika’s BacT/Alert was the first
the new version is the BACTEC™ MGIT™ system to use colorimetric technology. The
320. The BACTEC™ 960TB system has system is based on the detection of a decrease
become one of the most widely used systems in the pH of the medium by actively
for the determination of susceptibility of M. proliferating mycobacteria that metabolize
tuberculosis to antibiotics currently used to the substrates present in the culture medium,
treat tuberculosis (Siddiqi and Rüsch-Gerdes, resulting in the production of CO2, which in
2006). Based on the same principle of the turn acidifies the medium and causes a colour
original system, the BACTEC™ MGIT™ 320 change in the sensor that is present at the
could represent an option that is less expen- base of each culture bottle. The sensor will
sive for laboratories with little volume of change from dark green to yellow as micro-
work and could be an alternative to the organisms grow and this alteration is detected
BACTEC™ MGIT™ 960 in low-income by a reflectometric unit in the instrument.
countries. The system automatically performs readings
every 10 min using infrared rays. All data are
transferred to and saved with the BacT/View
Myco-ESP Culture System II
software. This simplifies and reduces the
The ESP® Culture System II (Trek Diagnostic errors associated with manual operations.
Systems, USA) incorporates a unique and Pre-processing is not required and specimen
very sensitive technology which measures volumes can be inoculated directly into the
the pressure change inside the culture bottle. culture bottle. The systems MB/Bact 240
Each bottle is connected to a sensor and (without shaking) and the BacT/ALERT 3D®
monitored continuously to detect any alter- (without shaking) use the BacT/ALERT MB
ation in the pressure due to the metabolic (non-selective media for blood specimens)
and MP culture bottles (for sterile body speci- growth usually provide the detection of
mens other than blood, and from digested colonies after 3–8 weeks, under optimum
decontaminated samples) and can be used in conditions, depending on the species, but the
combination with MB/BacT enrichment fluid use of liquid media decreases this period
or MB/BacT® antibiotic supplement. This is markedly to 1–2 weeks. Nevertheless, growth
also a non-radiometric, needle-free, fully auto- in solid media must be retained and used for
mated with continuous monitoring, ‘walk- certain mycobacterial species since the
away’ system introduced as an alternative to morphology and chromogenicity of the
its counterpart, the BACTEC™ MGIT™ 960 colonies continue to be important for the
system (Angeby et al., 2003). accurate identification of mycobacteria (Fig.
3.1; Leão et al., 2004). Since each type of
culture offers certain advantages, the
BACTEC™ 9000MB combination of different culture media (solid
The BACTEC™ 9000MB is an automated medium plus liquid medium) is recom-
non-radiometric culture system for the isola- mended for the recovery and primary
tion of mycobacteria (Becton Dickinson). This isolation of mycobacteria.
system uses the MYCO/F medium, a modified
Middlebrook 7H9 media, which can be
supplemented with an antimicrobial cocktail 3.3 Adaptation of Automated Liquid
to suppress the growth of contaminating Culture Systems for the Qualitative
microorganisms (van Griethuysen et al., and Quantitative Determination of
1996). The inoculated vials are inserted into Drug Susceptibility
the equipment for incubation and periodic
reading (every 30 min) (Sharp et al., 1997). With the increasing number of patients
The system responds to changes in oxygen infected with drug-resistant strains of M.
concentration (van Griethuysen et al., 1996). tuberculosis, in parallel with a significant
Each vial contains a silicon rubber disc, increase of infections caused by non-
impregnated with ruthenium, which serves tuberculous mycobacteria (NTM), especially
as an oxygen-specific sensor so that micro- in immunocompromised individuals, the
organism metabolism and growth are determination of the antibiotic susceptibility
detected (van Griethuysen et al., 1996; Sharp profile of the isolated mycobacteria became
et al., 1997). Oxygen consumption by micro- mandatory and the new, fast culture methods
organisms present in the medium can be quickly became the first choice to reduce the
detected by the increase in fluorescence and time between isolation and the determination
a positive reading indicates presumptive of drug susceptibility (Perkins, 2000; Hazbón,
presence of viable microorganisms (van 2004; Ridderhof et al., 2007). For this purpose,
Griethuysen et al., 1996; Sharp et al., 1997). the semi-automatic BACTEC™ 460TB and
The BACTEC™ 9000MB system monitors the BACTEC™ 960TB systems (Becton Dickin-
fluorescence levels and detects the growth of son Instrument Systems) were adapted for a
microorganisms on the basis of software- quick and reliable susceptibility test that
based positivity algorithms and is a rapid, rapidly became the reference worldwide and
sensitive and efficient non-radiometric replaced the labour-intensive proportion
method suitable for isolation of mycobacteria method in Middlebrook 7H11 (Cummings et
in a clinical laboratory (van Griethuysen et al., al., 1975; Salfinger, 1977; Siddiqi et al., 1981).
1996). The adaptation of this susceptibility test is a
Besides the recent improvement in the direct extension of the method of detection of
culture methods for mycobacteria, from solid growth in the systems and the susceptibility
to liquid and from manual to automated, all of the microorganism to the critical concen-
of them are more or less expensive and tration of the antibiotic to be tested is defined
require more than 1–2 weeks to produce in terms of diminishing GUs when compared
results. Methods that employ solid media for with the control (no antibiotic) (Siddiqi et al.,
1981). The criteria adopted for this method References

are based on the same principles of the
proportion method. If there is a sufficient American Thoracic Society, the Centers for Disease
percentage of resistant bacilli (>1%) in the Control and Prevention and the Infectious
population to obtain a normal growth curve Diseases Society of America (2003) ATS/CDC/
in vials containing the critical concentration IDSA statement: treatment of tuberculosis.
Morbidity and Mortality Weekly Report 52,
of the antibiotic, then the strain is considered
RR-11 (www.thoracic.org/sections/publications/
to be resistant. This interpretation is based on
statements/pages/mtpi/rr5211.html, accessed
the growth curves recorded and the criteria 21 December 2010).
defined in the operating manual (Lorian, Angeby, K.A., Werngren, J., Toro, J.C., Hedström,
2005). Recently, this method has evolved for G., Petrini, B. and Hoffner, S.E. (2003)
the quantitative drug susceptibility testing of Evaluation of the BacT/ALERT 3D system for
M. tuberculosis using the BACTEC™ 960TB recovery and drug susceptibility testing of
system EpiCenter software equipped with Mycobacterium tuberculosis. Clinical Micro-
the TB eXiST module and has been created biology and Infection 9, 1148–1152.
for determination of the minimum inhibitory Cummings, D.M., Ristroph, D., Camargo, E.E.,
Larson, S.M. and Wagner, H.N. Jr (1975)
concentration and evaluation of breakpoints
Radiometric detection of the metabolic activity
for each antibiotic that might be used for
of Mycobacterium tuberculosis. Journal of
therapeutic purposes against the clinical Nuclear Medicine 16, 1189–1191.
strain isolated (Springer et al., 2009). This new David, H.L. (1989) Méthodes de laboratoire pour
feature of the BACTEC™ 960TB system mycobacteriologie clinique. Institute Pasteur,
allows personalized therapies implemented Paris.
for each tuberculosis, multidrug-resistant Griffith, D.E., Aksamit, T., Brown-Elliott, B.A.,
tuberculosis (MDR-TB), extremely drug- Catanzaro, A., Daley, C., Gordin, F., et al.; ATS
resistant tuberculosis (XDR-TB) or NTM Mycobacterial Diseases Subcommittee;
infected patient, adjusting the therapy and American Thoracic Society; and Infectious
Disease Society of America (2007) An official
dosage to the susceptibility profile of the
ATS/IDSA statement: diagnosis, treatment, and
clinical isolate, providing the obvious advan-
prevention of nontuberculous mycobacterial
tages of increased efficacy and reduction of diseases. American Journal of Respiratory and
non-compliance. This technical improvement, Critical Care Medicine 175, 367–416.
although expensive, will have, in the future, a Hazbón, M.H. (2004) Recent advances in molecular
direct impact on the control of acquired methods for early diagnosis of tuberculosis and
resistance, as well as on the increase in the drug-resistant tuberculosis. Biomedica
cure rate of XDR-TB infected patients. 24(Suppl. 1), 149–162.
Furthermore, it will allow immunocom- Jensen, K.A. (1932) Rinzuchtung und
promised patients infected with NTM to Typenbestimmung von Tuberkelbazillentamen.
Zentralblatt für Bakteriologie, Parasitenkunde,
receive efficient susceptibility testing for
Infektionskrankheiten und Hygiene 1. Abt.
directed therapy as an alternative to the
Medizinisch-hygienische Bakteriologie, Virusfor-
empiric therapeutic regimens implemented schung und Parasitologie. Originale 125, 222.
today against these infections (Viveiros et al., Kent, P.T. and Kubica, G.P. (1985) Mycobacteriology:
2010). A Guide for the Level III Laboratory. US Dept of
Health and Human Services, Public Health
Service, Centers for Disease Control, Atlanta,
Georgia.
Acknowledgements
Leão, S.C., Martin, A., Mejia, G.I., Palomino, J.C.,
Robledo, J., Telles, M.A.S., et al. (2004) Practical
The authors would like to thank the Fundação handbook for the phenotypic and genotypic
Calouste Gulbenkian (Portugal) for its con- identification of mycobacteria. Vanden Broele
tinuous support in the research and develop- (ed.), Bruges, Belgium (www.esmyco
ment of new approaches for the control of bacteriology.eu/PDF%20files/foreword.pdf,
tuberculosis and drug-resistant tuberculosis. accessed 21 December 2010).
Lorian, V. (ed.) (2005) Antibiotics in Laboratory Siddiqi, S.H., Libonati, J.P. and Middlebrook, G.
Medicine, 5th edn. Lippincott Williams and (1981) Evaluation of a rapid radiometric method
Wilkins, Philadelphia, Pennsylvania. for drug susceptibility testing of Mycobacterium
Löwenstein, E. (1931) Die Zacktung der Tuber tuberculosis. Journal of Clinical Microbiology
kolbazillen aus dem stramenden Blute. 13, 908–912.
Zentralblatt für Bakteriologie, Parasitenkunde, Siddiqi, S.H., Hawkins, J.E. and Laszlo, A. (1985)
Infektionskrankheiten und Hygiene 1. Abt. Inter-laboratory drug susceptibility testing of
Medizinisch-hygienische Bakteriologie, Virusfor- Mycobacterium tuberculosis by radiometric
schung und Parasitologie. Originale 120, 127. procedure and two conventional methods.
Middlebrook, G. and Cohn, M.L. (1958) Bacteriology Journal of Clinical Microbiology 22, 919–923.
of tuberculosis: laboratory methods. American Springer, B., Lucke, K., Calligaris-Maibach, R.,
Journal of Public Health 48, 844–853. Ritter, C. and Böttger, E.C. (2009) Quantitative
Middlebrook, G., Reggiardo, Z. and Tigertt, W.D. drug susceptibility testing of Mycobacterium
(1977) Automatable radiometric detection of tuberculosis using MGIT960 and the EpiCenter
growth of Mycobacterium tuberculosis in selec- instrumentation. Journal of Clinical Microbiology
tive media. American Review of Respiratory 47, 1773–1780.
Disease 115, 1067–1069. Strand, C.L., Epstein, C., Verzosa. S., Effatt, E.,
Palomino, J.C., Leão, S.C. and Ritacco, V. (2007) Hormozi, P. and Siddiqi, S.H. (1989) Evaluation
Tuberculosis 2007 – From Basic Science to of a new blood culture medium for mycobacteria.
Patient Care, 1st edn. TuberculosisTextbook. American Journal of Clinical Pathology 91,
com (www.tuberculosistextbook.com, accessed 316–318.
21 December 2010). Tortoli, E. and Marcelli, F. (2007) Use of the INNO
Parish, T. and Stoker, N.G. (eds) (1998) Myco- LiPA Rif. TB for detection of Mycobacterium
bacteria Protocols, Vol 101. Humana Press, tuberculosis DNA directly in clinical specimens
Totowa, New Jersey. and for simultaneous determination of rifampin
Perkins, M.D. (2000) New diagnostic tools for susceptibility. European Journal of Clinical
tuberculosis. International Journal of Microbiology and Infectious Diseases 26,
Tuberculosis and Lung Disease 4, 182–188. 51–55.
Ridderhof, J.C., van Deun, A., Kam, K.M., Tortoli, E., Mandler, F., Tronci, M., Penati, V.,
Narayanand, P.R. and Aziz, M.A. (2007) Roles Sbaraglia, G., Costa, D., et al. (1997)
of laboratories and laboratory systems in Multicenter evaluation of mycobacteria growth
effective tuberculosis programmes. Bulletin of indicator tube (MGIT) compared with the
the World Health Organization 85, 354–359. BACTEC radiometric method, BBL biphasic
Salfinger, M. (1977) Diagnosis of tuberculosis and growth medium and Löwenstein–Jensen
other diseases caused by mycobacteria. medium. Clinical Microbiology and Infection 3,
Infection 25, 60–62. 468–473.
Salfinger, M. and Pfyffer, G.E. (1994) The new Tortoli, E., Cichero, P., Chirillo, M.G., Gismondo,
diagnostic mycobacteriology laboratory. Euro- M.R., Bono, L., Gesu, G., et al. (1998)
pean Journal of Clinical Microbiology and Multicenter comparison of ESP Culture System
Infectious Diseases 13, 961–979. II with BACTEC 460TB and with Löwenstein–
Sharp, S.E., Lemes, M., Erlich S.S. and Poppiti, Jensen medium for recovery of mycobacteria
R.J. Jr (1997) A comparison of the Bactec from different clinical specimens, including
9000MB system and the Septi-Chek AFB blood. Journal of Clinical Microbiology 36,
system for the detection of mycobacteria. 1378–1381.
Diagnostic Microbiology and Infectious van Griethuysen, A.J., Jansz, A.R. and Buiting,
Diseases 28, 69–74. A.G. (1996) Comparison of fluorescent
Siddiqi, S.H. and Rüsch-Gerdes, S. (2006) MGITTM BACTEC 9000 MB system, Septi-Chek AFB
Procedure Manual for BACTECTM MGIT system, and Löwenstein–Jensen medium for
960TM TB System (also applicable for Manual detection of mycobacteria. Journal of Clinical
MGIT) Mycobacteria Growth Indicator Tube Microbiology 34, 2391–2394.
(MGIT) Culture and Drug Susceptibility Demon- Viveiros, M., Martins, M., Couto, I., Rodrigues, L.,
stration Projects. Foundation for Innovative New Machado, D., Portugal, I., et al. (2010) Molecular
Diagnostics, 89 pp (http://www.finddiagnostics. tools for rapid identification and novel effective
org/export/sites/default/resource-centre/find_ therapy against MDRTB/XDRTB infections.
documentation/pdfs/mgit_manual_nov_2007. Expert Review of Anti-infective Therapy 8, 465–
pdf, accessed 9 October 2012). 480.
Watterson S.A. and Drobniewski, F.A. (2000) Modern Report No 4 (2002–2007). World Health
laboratory diagnosis of mycobacterial infections. Organization, Geneva, Switzerland, pp. 1–142.
Journal of Clinical Pathology 53, 727–732. WHO (2009) Global tuberculosis control –
World Health Organization [WHO] (2008) The epidemiology, strategy, financing. WHO Report
WHO/IUATLD Global Project on Anti- 2009. World Health Organization, Geneva,
tuberculosis Drug Resistance Surveillance – Switzerland, pp. 1–314.
4 Interferon-gamma Release Assays in
the Diagnosis of Latent Tuberculosis
Infection
Ajit Lalvani1 and Manish Pareek2

1Tuberculosis Research Unit, Department of Respiratory Medicine, National Heart
and Lung Institute, Imperial College London, UK; 2Department of Infection, Immunity
and Inflammation, University of Leicester, UK
4.1 Introduction increase the chance of being infected with M.

tuberculosis, those individuals in close contact
The failure to control the global tuberculosis with smear-positive infectious cases or with
epidemic adequately may partly reflect the underlying immunosuppressive states are
slow progress that has been made, over the often considered to be at highest risk of
last century, in developing innovative diag- acquiring infection (see Table 4.1).
nostic tools for tuberculosis (Lalvani, 2007).
However, the potential now exists to advance
tuberculosis control significantly with the Table 4.1. Risk factors for acquiring latent
development of novel diagnostic modalities tuberculosis infection (American Thoracic Society
[ATS] and Centers for Disease Control and
for latent tuberculosis infection (LTBI).
Prevention [CDC], 2000; National Collaborating
Centre for Chronic Conditions, 2006).
4.1.1 Latent tuberculosis infection: Risk factor

worldwide burden Close contact with smear-positive TB case
Immunosuppression
Estimates suggest that one-third of the HIV infection
world’s population is latently infected with Diabetes mellitus
tubercle bacilli but do not have overt clinical End-stage renal disease
disease (Dye et al., 1999). Infection with Long-term steroid use
Mycobacterium tuberculosis arises through a Immunosuppressive drugs
Silicosis
complex pathogenic process following
Malnutrition
exposure to droplet-containing bacilli which Old age
eventually concludes with containment of the Birth in high tuberculosis burden region
infection through infiltration of CD4+ T Health-care workers
lymphocytes and the release of interferon-γ, Homeless
which activate bacilli-containing macrophages Prisoners
(Cooper and Flynn, 1995; Bean et al., 1999; Travel to high-burden region
Murray, 1999). While a number of situations Smoking

Diagnosis of Latent Tuberculosis Infection 47
Latently infected individuals are at risk active tuberculosis disease (Comstock et al.,
of future progression to active tuberculosis 1974; Horsburgh, 2004).
disease if immune control fails (Dye et al., As a result, there is an increasing
1999; Corbett et al., 2003). However, evidence recognition, both in high and low tuberculosis
suggests only 5–10% of infected immuno- burden settings, of the importance of
competent people actually go on to develop specifically diagnosing and treating LTBI in
active disease over their lifetime (Comstock et those individuals with the highest risk of
al., 1974; American Thoracic Society [ATS] progression from LTBI to active tuberculosis
and Centers for Disease Control and (MacPherson and Gushulak, 2006; Cain et al.,
Prevention [CDC], 2000; Horsburgh, 2004). 2008). This is reflected in the importance
Certain factors (see Tables 4.2 and 4.3), such accorded to tracing recent contacts of smear-
as increasing age or the presence of Human positive individuals in low-burden settings
Immunodeficiency Virus (HIV) infection, (National Collaborating Centre for Chronic
increase the proportion of and rates at which Conditions, 2006) and the WHO’s policy of
latently infected individuals go on to develop recommending isoniazid-preventive therapy
Table 4.2. Incidence of active tuberculosis in persons with a positive tuberculin test, by
selected risk factors – from US and global studies (ATS and CDC, 2000).
Risk factor Incidence (per 1000 person-years)
Immunosuppressive states
HIV 35.0–162
Intravenous drug use and HIV positive 76.0
Intravenous drug use and HIV negative 10.0
Weight deviation
Underweight by 15% 2.6
Underweight by 10–14% 2.0
Underweight by 5–9% 2.2
Weight within 5% of standard 1.1
Silicosis 68.0
Recent infection with tuberculosis
<1 year ago 12.9
1–7 years ago 1.6
Table 4.3. Relative risk for developing active tuberculosis by selected clinical
conditions – from US and global studies (ATS and CDC, 2000).
Risk factor Relative risk
Diabetes mellitus 2.0–4.1
Silicosis 30.0
Smoking 2.3
End-stage renal disease 10.0–25.3
Gastrectomy 2–5
Jejunoileal bypass 27–63
Solid organ transplant

Renal 37
Cardiac 20–74
Carcinoma of head or neck 16
48 A. Lalvani and M. Pareek
in HIV-positive individuals in high-burden of a genomic segment (region of difference 1

settings (World Health Organization [WHO], – RD1) which encodes two antigens – early
1998). However, the successful, cost-effective secretory antigen target-6 (ESAT-6) and
implementation of these policies requires a culture filtrate protein-10 (CFP-10) – which
diagnostic test of sufficient specificity and are strong specific targets of Th1 T cells in M.
sensitivity. tuberculosis infection but are also deleted from
all strains of BCG (Mahairas et al., 1996) and
the majority of environmental mycobacteria
4.2 Methods of Diagnosing Latent (except M. kansasi, M. szulgai, M. marinum, M.
Tuberculosis Infection flavescens and M. gastrii) (Harboe et al., 1996).
By eliciting specific T-cell responses to these
Identifying LTBI accurately is difficult. The antigens, the frequency of false positive
combination of a low bacterial load and weak TST results in BCG-vaccinated individuals
humoral response means that diagnosis relies is reduced (Lalvani, 2007; Lalvani and
primarily on cellular immunological markers Millington, 2007, 2008b); this property forms
of infection (Lalvani, 2003). Therefore, for the basis of the two ex vivo IGRAs that are
much of the last century, a positive tuberculin currently available (see Table 4.4): the
skin test (TST) in an individual exposed to M. enzyme-linked immunosorbent spot (ELISpot),
tuberculosis with no other clinico-radiographic which directly counts the number of IFN-γ
evidence of active disease was considered secreting T cells (commercially available as
diagnostic of LTBI. the T-SPOT™.TB (Oxford Immunotec, Abing-
The TST measures a cutaneous delayed- don, UK)) and the whole blood enzyme-
type hypersensitivity response to intra- linked immunosorbent assay (ELISA), which
dermally injected purified protein derivative measures the concentration of IFN-γ secretion
(PPD), a mixture of over 200 M. tuberculosis (commercialized as the QuantiFERON™-TB
proteins (Lalvani and Pareek, 2009); the Gold In-Tube (Cellestis, Carnegie, Australia)).
resulting size of TST response determines
whether an individual has LTBI and
quantifies the risk of progression to active 4.3 Clinical Performance of IGRAs in
tuberculosis (ATS and CDC, 2000). Although Different Populations
the TST is cheap and widely used, it is neither
specific, as the antigens present in PPD cross In the absence of a gold standard reference
react with Bacillus Calmette–Guérin (BCG) test for LTBI, objective comparisons of the
and environmental mycobacteria, nor sensi- performance of TST and IGRAs are challeng-
tive, due to anergy in those individuals with a ing. Consequently, proxy markers of LTBI
compromised immune system (such as HIV, must be used, including: (i) active tuberculosis
iatrogenic immunosuppression and children) as a surrogate for LTBI; (ii) degrees of contact
(Lalvani, 2003). Logistic drawbacks include with smear-positive infectious cases; and (iii)
the need for the test to be performed by a negative IGRA results in healthy BCG-
trained health-care professional and the vaccinated individuals at low risk of
requirement for a return visit to have the tuberculosis infection due to the absence of
result read. risk factors for tuberculosis exposure.
4.2.1 T-cell interferon-gamma release

4.3.1 Performance of IGRAs in
assays
immunocompetent persons
The development of T-cell interferon-gamma
Active tuberculosis as a surrogate for LTBI in
release assays (IGRAs) as an alternative
immunocompetent patients
immunodiagnostic modality for M. tubercu-
losis (Lalvani et al., 2001b; Mori et al., 2004; In the absence of a gold standard reference
Richeldi, 2006) stems from the identification test for LTBI, a commonly used measure of
Table 4.4. The similarities and differences between the ELISpot (TSPOT.TB) and ELISA
(QuantiFERON-TB Gold In-tube).
ELISpot ELISA
Antigens ESAT-6 and CFP-10 ESAT-6, CFP-10 and TB 7.7
Positive internal control Yes Yes
Readout units IFN- spot-forming cells (SFC) International units (IU) of IFN-
Technological platform ELISpot (enzyme-linked ELISA
immunospot – modified form of
sandwich ELISA)
Test’s substrate Peripheral blood mononuclear cells Whole blood
(PBMC)
Outcome measure Number of IFN- producing T-cells Serum concentration of IFN-
produced by T-cells
Readout system Enumeration of spots by naked Measurement of optical density values
eye, magnifying lens or using an automated reader
automated reader
Technical procedures in Separation and enumeration and Tubes into which blood is drawn are
diagnostic laboratory dispensing of PBMC into ELISpot incubated directly without further
wells required before incubation procedures
the diagnostic sensitivity of IGRAs is to confirmed tuberculosis (Soysal et al., 2008). In

evaluate their performance in individuals general, most studies comparing the ELISpot
with active tuberculosis disease who, by against the TST have shown that its sensitivity
definition, are infected with M. tuberculosis. is superior to that of the TST, with values in
This was first shown for the ELISpot in a the range of 58–100% versus 46–100%,
small, proof-of-principle study among culture- respectively (Diel et al., 2010b).
confirmed cases (n = 47) where the ELISpot Numerous studies have also assessed the
had a sensitivity of 96% versus 69% for the sensitivity of the ELISA in culture-confirmed
TST (Lalvani et al., 2001b). Subsequent studies active tuberculosis. The first proof-of-prin-
have described the performance of the ciple study was conducted in 118 Japanese
ELISpot among immunocompetent indi- patients with culture-confirmed tuberculosis
viduals in both high- and low-burden settings. where the overall sensitivity was 89% (Mori et
One of the largest and most rigorously al., 2004). Since this first report in 2004, the
designed studies to date, conducted in evidence base for the ELISA has also
routine clinical practice, was undertaken in expanded rapidly. Data are now available
the UK by Dosanjh et al., who found that the from high- and low-burden settings for both
standard ELISpot had a sensitivity of 85% in children and adults. While a South African
culture-confirmed/highly probable tubercu- study of children with tuberculosis found the
losis as compared to 83% with the TST sensitivity of the ELISA to be 76% (Tsiouris et
(Dosanjh et al., 2008). A similarly large study, al., 2006b), a smaller study from the UK found
conducted in Singapore, found that the a higher sensitivity of 87% in children with
ELISpot had a sensitivity in culture-confirmed active tuberculosis (Kampmann et al., 2009).
adult tuberculosis patients of 94.1% (Chee et In adults, Chee and colleagues have con-
al., 2008). A smaller study in South Korean ducted the largest evaluation of the ELISA to
adults found the ELISpot to have a similarly date and found the sensitivity to be 83% in
high sensitivity of 95% (Lee et al., 2006). adult patients with tuberculosis (Chee et al.,
Conversely, in a mixed population of Turkish 2008); conversely, a Japanese study of 100
children and adults, Soysal and colleagues patients with active tuberculosis found that
found that the ELISpot had a slightly lower sensitivity in their population was higher at
sensitivity of 83% in those with culture- 93% (Harada et al., 2008). In general, studies
comparing the performance of the ELISA so than with the TST) with intensity of
against the TST have found the sensitivity of exposure and not affected by prior BCG
the IGRA and the TST to be in the range of vaccination (Lalvani et al., 2001a). Sub-
56–93% and 63–100%, respectively (Diel et al., sequently, many authors have used this
2010b). principle to compare the diagnostic accuracy
Two recent meta-analyses of the perform- of IGRAs and the TST in outbreak and
ance of the ELISpot and ELISA, where contact-tracing investigations – which often
culture-confirmed tuberculosis was the involve child contacts of infectious tubercu-
reference gold standard, have come to slightly losis cases (Lalvani et al., 2001a; Ewer et al.,
different calculations of the sensitivity of the 2003; Brock et al., 2004; Richeldi et al., 2004;
IGRAs – although the overall conclusions are Shams et al., 2005; Zellweger et al., 2005). One
congruent. Pai et al. calculated a pooled of the clearest demonstrations of this method
sensitivity of the ELISpot of 90% (range 83– was the outbreak investigation instigated as
100%), which was significantly higher than part of a large school outbreak comprising
the pooled sensitivities of the second- 535 students in the UK. This study found,
generation ELISA (78% – range 55–88%) and uniquely, that the ELISpot correlated signifi-
the latest-generation ELISA (QuantiFERON- cantly more closely with M. tuberculosis
Gold In-tube – pooled 70%, range 64–93%) exposure than the TST based on proximity
(Pai et al., 2008). In contrast, Diel et al. and duration of exposure to the index case
calculated the pooled sensitivity of the (Ewer et al., 2003). Other work from a variety
ELISpot and latest-generation ELISA (Quanti- of low-, intermediate- and high-prevalence
FERON-Gold In-tube) to be 88% (range 85– settings has also confirmed that the ELISpot
90%) and 81% (range 78–83%), respectively correlates significantly with tuberculosis
(Diel et al., 2010b). exposure (Kang et al., 2005; Soysal et al., 2005;
Hill et al., 2006, 2007; Detjen et al., 2007;
Connell et al., 2008; Diel et al., 2008a;
Correlation of IGRA results with tuberculosis
Dominguez et al., 2008; Nicol et al., 2009). In a
exposure in immunocompetent persons
Turkish community-based study of house-
Active tuberculosis, however, may not hold contacts, positive ELISpot and TST
represent accurately the host–bacteria results correlated significantly with the index
interaction that is present in LTBI. As a result, patient being a parent and the number of
authors have suggested that an alternative cases of smear-positive pulmonary tubercu-
method, namely correlation of IGRA responses losis per household (Soysal et al., 2005). More
with degree of exposure to tuberculosis, recently, in a study of 243 children in South
should be used to assess the performance of Africa, positive ELISpot and TST results in
IGRAs. The rationale underpinning this is children were found to be associated
that as M. tuberculosis is transmitted by the significantly with the degree of exposure to
airborne route, the risk of acquiring infection smear-positive index cases (Nicol et al., 2009).
is determined by the frequency, duration and Another recent study from a high-incidence
proximity of contact with an infectious source setting (Zimbabwe) also confirmed that
case (Houk et al., 1968; Grzybowski and positive ELISpot results were associated with
Enarson, 1978; Stead, 1978; Kenyon et al., the degree of exposure to sputum smear-
1996) and therefore the logical sequelae is positive tuberculosis cases (Mutsvangwa et
that if a new test is more sensitive and specific al., 2010).
than the TST, it should correlate more closely For the ELISA platform, the first study to
with the level of exposure to M. tuberculosis use degree of exposure as the reference
while still remaining independent of BCG standard was undertaken by Brock and
status (Lalvani et al., 2001a). colleagues in a contact-tracing study in
This was first demonstrated by Lalvani Danish children. This study showed that the
and colleagues, who showed that IGRA ELISA correlated with the degree of exposure
responses were correlated significantly (more and was unaffected by BCG status (Brock et
al., 2004). As with the ELISpot, the evidence (Hill et al., 2006), which was similar to the
base for studies correlating ELISA positivity level of agreement seen in an institutional
with exposure to tuberculosis has since outbreak in a low tuberculosis burden setting
expanded rapidly (Connell et al., 2006, 2008; (κ = 0.72) (Ewer et al., 2003). More recent work
Nakaoka et al., 2006; Chun et al., 2008; Okada from South Africa and Australia has found
et al., 2008; Lighter et al., 2009). The first study moderate levels of agreement between the
in a high tuberculosis burden setting (Nigeria) ELISpot and the TST (κ = 0.55 and 0.51,
was a case control of child contacts of adult respectively) (Connell et al., 2008; Nicol et al.,
tuberculosis patients where there was a dose– 2009).
response relationship, so that children who In contrast, concordance between the
had been in contact with the most heavily ELISA and the TST appears to be more
smear-positive index cases, rather than variable, with kappa values ranging from 0.19
smear-negative adults, were more likely to be to 0.87 (Brock et al., 2004; Dogra et al., 2007;
TST and ELISA positive (Nakaoka et al., 2006). Chun et al., 2008; Connell et al., 2008;
Similar results have been found in recent Dominguez et al., 2008; Okada et al., 2008;
work from Korea and Cambodia (Chun et al., Lighter et al., 2009). While studies in high
2008; Okada et al., 2008), However, in contrast, tuberculosis burden settings of hospitalized
a study from a South African township of children in India (κ = 0.73) (Dogra et al., 2007)
children at high risk of LTBI did not find any and childhood contacts in Cambodia (κ = 0.63)
significant relationship between levels of (Okada et al., 2008) found high concordance
exposure and ELISA positivity (Tsiouris et al., between the ELISA and the TST, other studies
2006a). In a low tuberculosis burden setting, have found considerably lower levels of
Lighter and colleagues undertook a study in agreement. In a South Korean case-control
US children and found that with increasing study, the level of agreement between the
likelihood of exposure to M. tuberculosis (from ELISA and the TST was low (κ = 0.19) (Chun et
minimal to high), the proportion of children al., 2008). Similarly, in an Australian study,
with positive ELISA results also increased there was poor agreement between the ELISA
(Lighter et al., 2009). and the TST (κ = 0.3) (Connell et al., 2006);
While the evidence base is currently more importantly, in the 21 unvaccinated
more substantial for the ELISpot than the children who had a positive TST, only 4 were
ELISA, the general consensus that can be ELISA positive. Subsequently, over half of
drawn from these studies is that both IGRAs BCG-unvaccinated children with a positive
correlate either as well as, or better than, the TST had a negative ELISA (Dominguez et al.,
TST with levels of exposure to tuberculosis, 2008). These data may imply that the ELISA
while also being independent of BCG status. may have a lower sensitivity than the TST in
An important element of these studies is diagnosing LTBI in children.
how concordant IGRAs and TSTs are when In the few studies which have compared
diagnosing individuals with suspected LTBI. the performance of the ELISA and the ELISpot
Currently available data indicates that in (CDC, 2010; Diel et al., 2010b), concordance
BCG-vaccinated individuals the level of between the two platforms has, in general,
agreement is poor, as BCG confounds the TST been good to very good (κ >0.6) (Detjen et al.,
but not IGRAs, and this is responsible for 2007; Dominguez et al., 2008; Soysal et al.,
most TST+/IGRA– discordance. Where 2008), although a notable exception is a
ELISpot and the TST have been investigated, Singaporean study by Chee et al. where
concordance between the two tests has been concordance was only fair (κ = 0.26) (Chee et
moderate to high (Ewer et al., 2003; Hill et al., al., 2008).
2006; Connell et al., 2008; Dominguez et al., Ultimately, the significance of discordant
2008; Lighter et al., 2009). For example, a results, whether they are TST/IGRA or
Gambian study of child contacts found a ELISpot/ELISA, can only be clarified reliably
moderately high level of concordance (κ = through longitudinal studies with clinical
0.62) between the ELISpot and the ELISA/TST outcomes.
4.3.2 Performance of IGRAs in tuberculosis burden settings, relatively few

immunocompromised persons studies have been conducted in HIV-positive
subjects in low tuberculosis burden settings.
Individuals with impaired cell-mediated Three studies from Europe have compared
immunity such as those with HIV infection or the ELISpot and the TST and found the
immune-mediated inflammatory diseases sensitivity of the IGRA to be in the range of
(such as rheumatoid arthritis, Crohn’s disease 84.6–90.3%, which is greater than that seen
and ankylosing spondylitis) undergoing iatro- with the TST (range 46–50%) (Clark et al.,
genic immunosuppression or anti-tumour 2007; Goletti et al., 2007; Vincenti et al., 2007).
necrosis factor (TNF) α therapy (Lalvani, The largest of these studies, which had 96
2003) are at increased risk of becoming (48%) patients with a CD4 count <200 cell/μl,
latently infection with M. tuberculosis and found that the ELISpot results were independ-
progressing to active tuberculosis disease. ent of CD4 cell counts in HIV-positive
Therefore, the evaluation of IGRAs as an individuals (Clark et al., 2007).
accurate method of identifying LTBI in these A few authors have assessed the perform-
high-risk groups is critical – especially as TST ance of the ELISA in tuberculosis- and HIV-
performance is recognized to be suboptimal coinfected individuals. In a recent Zambian
(Lalvani and Millington, 2008b). study, Raby and colleagues compared the
relative performance of the ELISA and the
TST in a cohort of HIV-positive and -negative
HIV-infected individuals
subjects with active tuberculosis. Overall, the
Since the first reports in 2002 (Chapman et al., sensitivity of the ELISA in HIV-positive
2002) outlining the performance of IGRAs in individuals (63%) was higher than the TST
HIV-positive individuals, the evidence base (55%) but significantly lower than in HIV-
has expanded very rapidly, with both active negative individuals (84%) (Raby et al., 2008).
tuberculosis and degree of exposure being In addition, the investigators found that
used as the reference standard. ELISA performance was affected negatively
by a falling CD4 count. Similar findings were
ACTIVE TUBERCULOSIS AS A SURROGATE FOR LTBI seen in a study of culture-confirmed tubercu-
IN IMMUNOSUPPRESSED PATIENTS In general, losis cases in Tanzania, where the subset of
fewer studies have used active tuberculosis patients with HIV infection had a significantly
as the reference standard when assessing the lower ELISA sensitivity than those without
performance of IGRAs in HIV-positive HIV infection (65% and 81%, respectively),
individuals. who were also affected adversely by a
One of the first studies to evaluate the decreasing CD4 count (Aabye et al., 2009). In
ELISpot in tuberculosis- and HIV-coinfected a low-burden setting, Vincenti and colleagues
individuals was undertaken in Zambian found ELISA sensitivity to be 84.6% versus
adults, where the sensitivity of the IGRA was 46% for the TST (Vincenti et al., 2007).
90% (Chapman et al., 2002). Subsequently, a
large study of South African children with CORRELATION WITH DEGREE OF EXPOSURE IN
suspected tuberculosis found that the sensi- IMMUNOSUPPRESSED PERSONS The alternative
tivity of the ELISpot was 83% versus 63% for reference standard, correlation with exposure
the TST and that IGRA responses were not to M. tuberculosis, has also been used to assess
affected adversely by the presence of HIV, the test performance of the ELISpot and
malnutrition or extremes of age (under 3 ELISA.
years old) (Liebeschuetz et al., 2004). In a Studies using this methodology to
small study of HIV-infected adults with evaluate the ELISpot have, in general, found
untreated pulmonary tuberculosis in South that in HIV-positive persons the diagnostic
Africa, the ELISpot performed better than the sensitivity of the ELISpot is higher than the
TST, with overall sensitivities for the two tests TST, and consequently concordance between
of 90% and 57%, respectively (Rangaka et al., the two tests is only moderate at best. In a
2007a). In contrast to these studies from high South African study of HIV-positive adults
and children, the ELISpot had an overall study of HIV-infected individuals in a high-
positivity of 61% as compared to 41% for the prevalence setting, the ELISA was found to
TST – although there was no reference have a lower rate of positivity than both the
standard against which the results were ELISpot and the TST, although the ELISA
compared. In this study, neither the ELISpot ESAT-6/CFP-10 responses in adults with
nor the TST correlated with exposure to M. known exposure to M. tuberculosis were
tuberculosis (Mandalakas et al., 2008). A larger significantly higher (Mandalakas et al., 2008).
study from South Africa confirmed that in In contrast, in HIV-positive Chilean adults,
HIV-infected individuals, the ELISpot had a when compared to the TST, the ELISA had a
higher positivity rate than the TST, was higher sensitivity while also correlating with
robust to changes in CD4 count and had a degree of exposure (Balcells et al., 2008).
moderate level of agreement with the TST Although this Chilean study did not find
( = 0.52) (Rangaka et al., 2007b). However, that low CD4 counts affected ELISA results
the ELISpot did not correlate with exposure adversely, there is accumulating evidence
to M. tuberculosis. In a recent study of HIV- that in HIV-positive individuals with advanced
positive individuals from a low-prevalence immunosuppression, and thus low CD4 cell
setting, the ELISpot was found to be more counts, ELISA performance is affected
sensitive than the TST and also to correlate adversely. In a study among a HIV cohort in
significantly with previous active tuberculosis San Francisco, Luetkemeyer and colleagues
disease, which was not the case for the ELISA found that decreasing CD4 count affected the
or the TST (Stephan et al., 2008). An additional CD4 count adversely – particularly the rate of
finding that ELISpot responses were not indeterminate results; when CD4 counts were
affected by advancing immunosuppression, less than 100, the odds ratio for an
as evidenced by a falling CD4 count, has been indeterminate result was 4.2 (Luetkemeyer et
confirmed by other studies (Dheda et al., al., 2007). Similar conclusions were drawn
2005). Most recently, Mutsvanga and col- from a large Danish study of HIV-positive
leagues undertook a study in Zimbabwe and individuals; despite in-tube ELISA results
found that ELISpot responses in recent being associated with risk factors for LTBI,
household contacts (who were both HIV there was also a significant association be-
positive and negative) correlated significantly tween low CD4 cell count and indeterminate
with sputum smear, and culture, positivity of ELISA results (Brock et al., 2006). These
the index case independently of contacts’ HIV findings have been confirmed by other
status; contacts’ TST results were also authors, suggesting that in advanced immuno-
associated with smear status of the index suppression, such as that seen in HIV, ELISA
cases but were affected negatively by the performance may be affected adversely
contacts’ HIV status (Mutsvangwa et al., (Jones et al., 2007; Mandalakas et al., 2008;
2010). Raby et al., 2008).
Data from evaluations of the ELISA in
HIV-positive populations are also beginning
Performance of IGRA in iatrogenically
to accumulate. In a recent study from the
immunosuppressed subjects with immune-
USA, Talati and colleagues assessed the
mediated inflammatory diseases
ELISA in HIV-positive clinic attenders. They
found that the ELISA had a higher positivity In individuals with immune-mediated inflam-
rate than the TST, poor concordance with the matory diseases (IMIDs), the evidence base
TST ( = 0.23) and, on multivariate analysis, for IGRAs as a diagnostic tool for LTBI has,
was associated with a previous history of until recently, been relatively limited. None
LTBI (Talati et al., 2009). An earlier study from the less, over the past 12–18 months there has
the USA also found a relatively low ELISA been an increase in the number of authors
positivity rate of 4.9%, although the ELISA, evaluating IGRAs in individuals with IMID.
but not the TST, was associated more closely As would be expected due to the rarity of
with risk factors for tuberculosis exposure/ finding individuals with both IMID and
LTBI (Jones et al., 2007). In a cross-sectional active tuberculosis, studies have not used
active tuberculosis as a surrogate marker for matory bowel disease undergoing screening
LTBI; instead, the majority of the work has pre anti-TNF therapy and found that agree-
been cross-sectional in design and focused on ment between the ELISpot and the TST
the concordance between the TST and IGRAs was poor ( = –0.02); in addition, TST but
and correlating IGRA responses with risk not ELISpot results were affected adversely
factors for LTBI (Lalvani and Millington, by severe iatrogenic immunosuppression
2008b). (Schoepfer et al., 2008). Although numerous
Evaluation of the ELISpot in individuals authors have confirmed these findings of
with IMID has found that concordance poor/moderate levels of concordance between
between the IGRA and the TST is, in general, TST and IGRA (Cobanoglu et al., 2007; Sellam
poor or moderate. In a study of 70 Greek et al., 2007; Takahashi et al., 2007; Ponce de
patients with a variety of rheumatological Leon et al., 2008; Gogus et al., 2009; Shovman
conditions who were being screened for LTBI et al., 2009; Qumseya et al., 2011), these
pre anti-TNF therapy, overall concordance findings are not universal. In a recent Italian
between the ELISpot and the TST was moder- study, the ELISA was used to screen indi-
ate; discordant TST positive/ELISpot negative viduals with IMID and found a relatively
results were associated with prior BCG high concordance between the TST and the
vaccination (Vassilopoulos et al., 2008). In a ELISA (87.5%, κ = 0.55), with a lower
US study, Behar and colleagues evaluated the proportion of TST+/IGRA– discordant results,
ELISpot in 200 patients with rheumatoid which was likely to relate to the low
arthritis and found poor concordance be- proportion of the population who had
tween the two tests; in addition, the ELISpot previously been BCG vaccinated (4.1%)
was not associated significantly with risk (Bartalesi et al., 2009).
factors for LTBI (Behar et al., 2009). Fewer As with the ELISpot, relatively few
authors have correlated ELISpot response authors have correlated ELISA results with
with risk factors for tuberculosis exposure. risk factors for LTBI. Bartalesi et al. studied
Recent work from Ireland found that in 398 Italian IMID patients and found that both
patients with inflammatory arthritides there ELISA and TST positivity were associated
was a significant association between ELISpot significantly with being close contacts of
positivity and risk factors for LTBI (Martin et patients with sputum smear-positive tubercu-
al., 2009). Laffitte and colleagues compared losis (Bartalesi et al., 2009). Two separate
the ELISpot and the TST in 50 patients with studies from Switzerland and Ireland in IMID
psoriasis and found that the ELISpot, but not patients undergoing pre anti-TNF screening
the TST, was associated significantly with risk found that ELISA positivity was associated
factors for LTBI (Laffitte et al., 2009). Similar significantly with risk factors for LTBI
conclusions that ELISpot results correlate (Matulis et al., 2008; Martin et al., 2009).
with risk factors for LTBI can be drawn from The general consensus that can be drawn
an Italian study of a heterogeneous group of from the published evidence on the perform-
subjects awaiting anti-TNF therapy (Bocchino ance of the ELISpot and the ELISA is that in
et al., 2008). individuals with IMID on immunosuppressive
Compared to the ELISpot, the evidence therapy, IGRAs maintain their diagnostic
base for ELISA performance in individuals sensitivity better than the TST, but false
with IMID is larger. Matulis and colleagues negative results are not uncommon and,
assessed the performance of the ELISA in therefore, given the need for a high index of
subjects with inflammatory arthritides and suspicion and a low threshold for treatment
found that concordance between the two tests initiation in this vulnerable population, an
was poor ( = 0.17) (Matulis et al., 2008). IGRA and TST should be used when screening
Similarly, in a Turkish study of individuals patients for LTBI prior to initiation of anti-
with a range of inflammatory conditions, the TNF treatment and a positive result in either
ELISpot and the TST had a poor level of test be taken as existence of LTBI and thus an
agreement ( = 0.18) (Cobanoglu et al., 2007). indication for chemoprophylaxis (Lalvani
Schoepfer et al. studied patients with inflam- and Millington, 2008b).
4.3.3 Specificity of IGRAs IGRAs and found that the proportions are
very similar for both platforms: 2.14% (95%
To quantify specificity of the IGRAs (in other confidence interval 2.0–2.3%) for the in-tube
words, the ability to identify correctly those ELISA and 3.8% (3.5–4.2%) for the ELISpot
individuals that are not latently infected), (Diel et al., 2010b).
investigators have studied BCG-vaccinated
individuals who live in low tuberculosis
burden regions with no known risk factors 4.3.5 Predictive value of IGRAs for
for tuberculosis infection who would, there- progression to active tuberculosis
fore, be reasonably expected to be at low risk
of being exposed to M. tuberculosis and hence Much of the preceding discussion has focused
acquiring LTBI. on the performance of the IGRAs in different
Two meta-analyses have been published populations. However, the question that
which address the issue of IGRA specificity. remains is – what does a positive IGRA result
In the first, Pai et al. concluded that the at baseline imply for an individual’s future
specificity of the ELISA ranged from 89% to risk of developing active tuberculosis? This is
100%, with a pooled specificity of 99% for the a critical point because providing chemo-
second-generation ELISA and 96% for the prophylaxis for IGRA-positive individuals
latest generation, in-tube ELISA (Pai et al., (especially contacts) will only be of benefit if
2008). For the ELISpot, the authors found a they are at increased risk of progressing to
similarly high diagnostic specificity of 85– active tuberculosis compared to individuals/
100%, with a pooled specificity of 93% (Pai et contacts who are IGRA negative. Data of this
al., 2008). However, a more recent meta- type from prospective studies have started to
analysis, which had different study inclusion accumulate since 2008 (Bakir et al., 2008; Diel
criteria, found slightly different estimates for et al., 2008b); recent systematic reviews and
the pooled specificity of the IGRAs: 99% meta-analyses have shown that IGRAs and
(range 98–100%) for the in-tube ELISA and TST are similarly predictive of progression to
86% (range 81–90%) for the ELISpot (Diel et active tuberculosis, although IGRAs do
al., 2010b). Although the figures may differ reduce the number of individuals requiring
slightly, the overall conclusion remains unnecessary chemoprophylaxis through
unchanged – IGRAs have a higher specificity improved specificity (Sester et al., 2011;
than the TST, particularly in BCG-vaccinated Rangaka et al., 2012).
populations. Studies from a range of settings have
evaluated the predictive power of the ELISpot
and, on the whole, have found that a positive
4.3.4 Clinical experience with IGRAs: response is prognostic for the future develop-
frequency of indeterminate results ment of active tuberculosis. A study from an
intermediate burden setting (Turkey) with
As the IGRAs begin to be more widely used, 908 child contacts in which 15 incident cases
it is important to consider their reliability in occurred found that ELISpot-positive contacts
routine clinical use. Both IGRAs are known to had a significantly increased risk of develop-
suffer from indeterminate results (Lalvani, ing active tuberculosis as compared to
2007), which usually occur either due to ELISpot-negative contacts (Bakir et al., 2008).
problems with the test/blood collection or Although the incidence of active tuberculosis
underlying cellular immunosuppression. was similar in contacts who were ELISpot-
Indeterminate results seem to be associated positive versus those who were TST-positive
particularly with extremes of age (<5 or >80) (although the prognostic value of the TST,
or, especially for the ELISA, with underlying unlike the ELISpot, did not remain statistically
immunosuppression due to HIV immuno- significant), the ELISpot predicted these from
suppressive medication (Kobashi et al., 2009). fewer contacts tested (Bakir et al., 2008).
Diel et al. have recently collated the However, in this study, a large proportion of
numbers of indeterminate results for the the children, as per national guidelines,
received isoniazid chemoprophylaxis, which investigators concluded that either the TST or
was likely to have resulted in an under- the ELISpot were equivalent in screening
estimation of the incidence rate ratios for contacts for LTBI but that neither test
both the TST and the ELISpot. A recent predicted subsequent progression to tubercu-
Senegalese study of 2679 household contacts, losis disease. This lack of predictive power
among whom 52 cases of tuberculosis for both TST and ELISpot may relate to the
occurred (40 of which were culture con- endemic setting where a substantial pro-
firmed), using an in-house ELISpot found portion of incident tuberculosis cases occur
that positive IGRA responses at baseline through de novo community transmission
predicted progression to active tuberculosis outside the household setting (Verver et al.,
disease – although on multivariate analysis 2004), although in Senegal, where endemicity
this was only significant for progression to is similar, the ELISpot was found to be
culture-confirmed tuberculosis; higher significantly predictive (Lienhardt et al.,
ELISpot responses at baseline were associated 2010).
significantly with progressing to tuberculosis Evidence for the prognostic power of the
disease (Lienhardt et al., 2010). In addition, ELISA first came via a small German study
contacts who had a positive ELISpot and TST which investigated the ELISA in 601 contacts
were more likely to progress than those who and found that a significantly higher pro-
were dual negative on these tests (Lienhardt portion of untreated ELISA-positive, com-
et al., 2010). In a Hong Kong study of 308 male pared to 5-mm TST-positive, household
silicotic patients, without recent tuberculosis contacts progressed to tuberculosis disease;
exposure, who were followed up for 2.5 all 6 incident cases were ELISA positive at
years, the ELISpot predicted progression to recruitment (Diel et al., 2008b). These data
active tuberculosis; Leung et al. found that have very recently been updated with a
15/204 (7.4%) ELISpot-positive and 13/203 longer duration of follow-up. Among 903
(6.4%) TST-positive (≥10 mm) individuals untreated contacts, a significantly higher pro-
progressed to active tuberculosis as compared portion of ELISA-positive subjects (19/147)
to only 2/104 (1.9%) of ELISpot-negative and developed active tuberculosis as compared to
4/105 (3.8%) of TST-negative subjects (Leung those subjects with a 10-mm positive TST
et al., 2010). A smaller study from Holland of (10/207); no ELISA-negative subjects de-
339 immigrant contacts, however, came to veloped active tuberculosis (Diel et al., 2010a).
different conclusions. Over a 2-year follow-up, Supporting data for these results come from a
there were nine cases among the cohort; recent Japanese study of 3102 household
ELISpot-positive immigrant contacts had a contacts; over 2 years, 20/419 (4.8%) ELISA-
higher risk of progression than those who positive contacts developed tuberculosis
were negative at baseline (Kik et al., 2010). disease as compared to only 19/2683 (0.7%) of
However, the ELISpot did not have a higher ELISA-negative contacts. The authors did not
positive predictive value for progression to undertake large-scale TST on the contacts
active tuberculosis than the TST, although the (Yoshiyama et al., 2010). Two unpublished
small numbers meant that it was not powered studies have also reported their findings in
adequately to ascertain the superiority of one recent conferences (Haldar et al., 2009;
test over another (Kik et al., 2010). Mahomed et al., 2011). Haldar et al. followed
In contrast to the above studies, a up 1309 household contacts and over a 2-year
prospective study of 2348 household contacts follow-up period found that 20/204 ELISA-
in a Gambian setting did not come to the positive contacts developed active tubercu-
same conclusions (Hill et al., 2008). During losis as compared to 0/835 ELISA-negative
the course of the study, active tuberculosis contacts (Haldar et al., 2009). A further study
occurred in 11/649 ELISpot-positive subjects from South Africa, which is still under way, is
as compared to 14/843 TST-positive subjects; evaluating the predictive power of the ELISA
in those who were ELISpot-negative or TST- in a cohort of adolescents over a 2-year
negative, 10/1087 and 11/1387, respectively, follow-up period. The investigators found
developed tuberculosis (Hill et al., 2008). The that ELISpot-positive individuals were more
likely to progress to active tuberculosis 3. At a population level, it is unknown

disease than ELISpot-negative contacts. How- whether chemoprophylaxis on the basis of
ever, when comparing the predictive power IGRA results will decrease tuberculosis inci-
of the ELISA and the TST, they found that dence more than on the basis of TST results.
similar proportions of ELISA-positive and
TST-positive adolescents progressed to
tuberculosis disease (38/3235 versus 37/2405) 4.3.6 Summary of clinical data
(Mahomed et al., 2011).
Most of the studies outlined above have As the output of IGRA research accumulates,
explored the prognostic power of the IGRA in several general conclusions can be drawn
immunocompetent subjects – mainly house- from the data that are currently available (see
hold contacts. However, the group where Table 4.5). Published studies have demon-
evidence is needed more urgently is in the strated that IGRAs are not confounded by
immunocompromised – particularly those previous BCG vaccination and are more
with HIV infection. Recently, Aichelberg et al. specific than the TST in the diagnosis of LTBI
explored the prognostic power of a positive (Pai et al., 2008; Diel et al., 2010b). With active
ELISA (QuantiFERON-Gold In-tube) result in disease used as a surrogate marker for LTBI,
HIV-positive individuals without active both IGRAs are more sensitive than the TST,
tuberculosis at recruitment (Aichelburg et al., with the ELISpot having a higher sensitivity
2009). The investigators found that over a than the ELISA (Pai et al., 2008; Diel et al.,
median follow-up period of 19 months, 3/37 2010b). In children and those with HIV
individuals with a positive ELISA at baseline infection, both the ELISpot and the ELISA
went on to develop active tuberculosis, seem to be superior to the TST – although
whereas 0/738 with a negative ELISA published data indicate that the ELISpot
subsequently developed tuberculosis (Aichel- performs better than the ELISA. In individuals
burg et al., 2009). This study, therefore, with IMID undergoing immunosuppressive
confirmed that the ELISA was prognostic for therapy, there is still a lack of evidence but, on
the future development of active tuberculosis the basis of what has been published, the
in HIV-positive subjects. ELISpot and the ELISA seem to have
These longitudinal clinical outcome comparable efficacy.
studies provide the first evidence that the
IGRAs predict the future development of
tuberculosis in a range of populations 4.4 Incorporation of IGRAs into
including: recent contacts, remotely infected
International Public Health Guidelines
individuals, children/adults, HIV negative/
positive and low/high tuberculosis burden
As the evidence base for IGRAs has evolved
settings. This supports the use of IGRAs to
and developed, they have gained regulatory
target preventive therapy, particularly for
approval in a number of countries and have
recent IGRA-positive contacts, as it enables
also been incorporated into national tuber-
prevention of a similar number of cases as
culosis control guidelines. Most European
when using the TST but necessitates
countries and Canada advise that IGRAs
treatment of significantly fewer contacts.
should be used in two situations (National
However some current areas of uncertainty,
Collaborating Centre for Chronic Conditions,
which will be addressed by future studies,
2006; Public Health Agency of Canada,
do remain:
2007):
1. It is not known how much more predictive 1. As a confirmatory test in individuals who
IGRAs are than the TST and to what extent have already tested positive with the TST.
the number needed to treat will be reduced. 2. As a direct replacement for the TST in
2. It is not known which IGRA platform is those individuals in whom the TST is likely
superior in predicting the future develop- to be unreliable (immunocompromised indi-
ment of future tuberculosis disease. viduals).
Table 4.5. Clinical utility of interferon-gamma release assays (IGRAs).

Diagnosis of active
Diagnosis of LTBI tuberculosis
Does a Does a Does a Does a
positive negative positive negative
test rule test rule test rule test rule Treatment
in? out? in? out? monitoring Test of cure
Immunocompetent
subjects
Adults Yes Yes No Nob No No
Children (under 5 years) Yes Yesa No Nob No No
Immunocompromised
subjects
Pre anti-TNF Yes Yesa No Nob No No
HIV positive Yes Yesa No Nob No No
Future directions Incorporating additional antigens improves IL-2 and IFN-

sensitivity without reducing specificity; for measurement holds
example, Rv3879c for the ELISpotPLUS promise in this setting
(Dosanjh et al., 2008). (Millington et al.,
Alternative cytokines such as IP-10 (Lalvani and 2007; Lalvani and
Millington, 2008c; Ruhwald et al., 2008). Millington, 2008a).
Notes: aWith caution and preferably should be used adjunctively with the TST to maximize sensitivity, as the TST will
detect LTBI in some patients with false negative IGRA results; bdiagnostic sensitivity of currently available IGRAs not yet
high enough to rule out tuberculosis. Use of the TST in parallel can detect some patients with false negative IGRA results
(Dosanjh et al., 2008).
Conversely, in the USA and Japan, guidelines most other European countries (National
recommend that IGRAs should be used in Collaborating Centre for Chronic Conditions,
place of (but not in addition to) the TST when 2006). Recent work has highlighted the cost
diagnosing LTBI – in all groups of individuals effectiveness of using single step IGRA testing
(Mazurek et al., 2005; CDC, 2010). in diagnosing LTBI in migrants (Pareek et al.,
Economic considerations have been an 2011).
important determinant in the disparate recom- The two-step TST and confirmatory
mendations made by different countries. approach recommended by Europe and
While IGRAs have a greater unit cost than the Canada has the potential drawback that
TST (Lalvani, 2007), health economic analyses individuals with a negative TST who may
have shown that with their increased have had a positive IGRA will not be
specificity they reduce the number of indi- diagnosed with LTBI and therefore not
viduals being treated with chemoprophylaxis offered the appropriate chemoprophylaxis.
unnecessarily (with the associated clinical and In contrast, the US and Japanese recom-
laboratory monitoring), which makes them a mendations could, potentially, result in
cost-effective option (Wrighton-Smith and overtreatment of individuals who are IGRA
Zellweger, 2006; Diel et al., 2007). The UK positive but TST negative – as the risk of
National Institute of Health and Clinical subsequently progressing to active tuber-
Excellence (NICE) also undertook their own culosis disease in this group with discordant
economic analyses and concluded that the results is not known (Fig. 4.1). Further
two-step TST and confirmatory IGRA research is urgently needed to quantify
approach would be most cost-effective, the risk of individuals with discordant
forming the basis of the recommendation IGRA/TST results progressing to active tuber-
made by NICE and subsequently adopted by culosis.
Suspected LTBI
TST and IGRA
TST + TST + TST – TST –

IGRA + IGRA – IGRA + IGRA –
*
Repeat IGRA
at 3 months
a #
IGRA + IGRA –
*
Uninfected Probable
LTBI false +ve TST LTBIb transient MTB Uninfected
due to BCG infectionc
Fig. 4.1. Algorithm of all possible outcomes resulting from parallel IGRA and TST testing.*
NICE guidance
Defines outcomes not captured by NICE guidance, which mandates undertaking the TST first
followed by the IGRA in TST-positive individuals except in the immunosuppressed population,
where IGRA only testing is recommended
Notes: Negative TST and IGRA indicates no infection. Positive TST and IGRA indicates LTBI. (a) With
strong evidence for LTBI (e.g. >25-mm TST, ulcerating Mantoux, calcified lesions on chest x-ray), the
possibility of false negative IGRA must be considered. (b) Risk of active tuberculosis development in this
group is not yet known. (c) Longitudinal contact studies have defined a group of TST-negative individuals
with transiently positive IGRAs, raising the possibility that some TST-negative contacts acquire and
spontaneously clear transient M. tuberculosis infection. *A positive TST/IGRA must be evaluated carefully
in immunocompromised patients in whom any positive result is significant. #Dual negative TST and IGRA
results normally indicate the exclusion of tuberculosis infection but negative results should be evaluated
carefully in immunocompromised patients.
This highlights the urgent need for more and evidence base. This has meant that their
longitudinal data to quantify the predictive limitations are becoming clearer. Neither
value of positive IGRA results and, in par- IGRA can differentiate between active and
ticular, the prognosis of contacts with dis- latent tuberculosis and prospective studies
cordant IGRA and TST results. which have undertake serial IGRA testing
have highlighted the dynamic nature of IGRA
responses during the course of treatment for
4.5 Treatment Monitoring and Future active tuberculosis – which means that they
Directions cannot be used to monitor treatment response
or as a test of cure (Lalvani, 2007). However,
Although IGRAs have upgraded the TST, by recent evidence that simultaneous profiling
which we can diagnose LTBI, they remain a of IL-2 and IFN-γ secretion at the single T-cell
work in progress with an evolving clinical level correlates with therapeutic response
may be useful in disease monitoring generation IGRAs will have improved

(Millington et al., 2007; Lalvani and sensitivity without compromising specificity.
Millington, 2008a).
Despite the current generation of IGRAs
having higher sensitivity than the TST, Conflict of Interest Statement
research indicates that next-generation assays
will have even higher sensitivity. This has Professor Lalvani is the inventor of several
been achieved by the next-generation ELISA patents underpinning T-cell-based diagnosis.
and ELISpot including additional antigens The ESAT-6/CFP-10 IFN-γ ELISpot was
(Liu et al., 2004). Incorporating novel antigens, commercialized by an Oxford University
Rv2645 for the in-tube ELISA (Harada et al., spin-out company (T-SPOT.TB®, Oxford
2008) and RV3879c for the ELISpotPLUS Immunotec Ltd, Abingdon, UK) in which
(Dosanjh et al., 2008), alongside ESAT-6 and Oxford University and Professor Lalvani
CFP-10 has been shown in recent studies to have a minority share of equity and royalty
improve sensitivity significantly without entitlements.
compromising specificity. Additionally, new
mycobacterial antigens are being explored,
such as heparin-binding haemagglutinin, References
which may help to distinguish active from
latent tuberculosis (Place et al., 2010). Aabye, M.G., Ravn, P., Praygod, G., Jeremiah, K.,
Studies are also exploring whether Mugomela, A., Jepsen, M., et al. (2009) The
measuring alternative, downstream chemo- impact of HIV infection and CD4 cell count on
kines secreted by IFN-γ-activated macro- the performance of an interferon gamma
phages such as inducible protein 10 (IP-10), in release assay in patients with pulmonary
combination with IFN-γ, may serve as a more tuberculosis. PLoS ONE 4, e4220.
amplified readout than IFN-γ alone, thereby Aichelburg, M.C., Rieger, A., Breitenecker, F.,
Pfistershammer, K., Tittes, J., Eltz, S., et al.
resulting in higher sensitivity (Lalvani and
(2009) Detection and prediction of active
Millington, 2008c; Ruhwald et al., 2008). tuberculosis disease by a whole-blood
A further area of ongoing research is interferon-gamma release assay in HIV-1-
focusing on whether IGRAs will be able to infected individuals. Clinical Infectious Disease
differentiate individuals who have been 48, 954–962.
infected recently and remotely; the relevance American Thoracic Society [ATS] and Centers for
of this being that those who are recently Disease Control and Prevention [CDC] (2000)
infected are more likely to progress to active Targeted tuberculin testing and treatment of
tuberculosis and should, therefore, be given latent tuberculosis infection. American Journal
chemoprophylaxis. Evidence to date, how- of Respiratory and Critical Care Medicine 161,
221S–247.
ever, suggests that IGRAs are unable to
Bakir, M., Millington, K.A., Soysal, A., Deeks, J.J.,
differentiate recent from remote infection Efee, S., Aslan, Y., et al. (2008) Prognostic value
(Hinks et al., 2009). of a T-cell-based, interferon-gamma biomarker
in children with tuberculosis contact. Annuals of
Internal Medicine 149, 777–787.
4.6 Conclusions Balcells, M.E., Perez, C.M., Chanqueo, L., Lasso,
M., Villanueva, M., Espinoza, M., et al. (2008) A
With the recent development of IGRAs, the comparative study of two different methods for
diagnosis of LTBI has moved on from the the detection of latent tuberculosis in HIV-
century-old TST. Evidence of their use in positive individuals in Chile. International
Journal of Infectious Diseases 12, 645–652.
different settings and populations has
Bartalesi, F., Vicidomini, S., Goletti, D., Fiorelli, C.,
expanded rapidly and IGRAs are an integral Fiori, G., Melchiorre, D., et al. (2009)
part of many national guidelines on LTBI QuantiFERON-TB Gold and the TST are both
diagnosis. However, further work is needed useful for latent tuberculosis infection screening
to understand the relevance of discordant in autoimmune diseases. European Respiratory
IGRA and TST results and whether next- Journal 33, 586–593.
Bean, A.G., Roach, D.R., Briscoe, H., France, M.P., Calmette–Guerin vaccinated children. Diagnostic
Korner, H., Sedgwick, J.D., et al. (1999) Microbiology and Infectious Disease 62, 389–
Structural deficiencies in granuloma formation 394.
in TNF gene-targeted mice underlie the Clark, S.A., Martin, S.L., Pozniak, A., Steel, A.,
heightened susceptibility to aerosol Myco- Ward, B., Dunning, J., et al. (2007) Tuberculosis
bacterium tuberculosis infection, which is not antigen-specific immune responses can be
compensated for by lymphotoxin. Journal of detected using enzyme-linked immunospot
Immunology 162, 3504–3511. technology in human immunodeficiency virus
Behar, S.M., Shin, D.S., Maier, A., Coblyn, J., (HIV)-1 patients with advanced disease. Clinical
Helfgott, S. and Weinblatt, M.E. (2009) Use of and Experimental Immunology 150, 238–244.
the T-SPOT.TB assay to detect latent Cobanoglu, N., Ozcelik, U., Kalyoncu, U., Ozen, S.,
tuberculosis infection among rheumatic disease Kiraz, S., Gurcan, N., et al. (2007) Interferon-
patients on immunosuppressive therapy. The gamma assays for the diagnosis of tuberculosis
Journal of Rheumatology 36, 546–551. infection before using tumour necrosis factor-
Bocchino, M., Matarese, A., Bellofiore, B., alpha blockers. The International Journal of
Giacomelli, P., Santoro, G., Balato, N., et al. Tuberculosis and Lung Disease 11, 1177–1182.
(2008) Performance of two commercial blood Comstock, G.W., Livesay, V.T. and Woolpert, S.F.
IFN-gamma release assays for the detection of (1974) The prognosis of a positive tuberculin
Mycobacterium tuberculosis infection in patient reaction in childhood and adolescence. The
candidates for anti-TNF-alpha treatment. Euro- American Journal of Epidemiology 99, 131–
pean Journal of Clinical Microbiology and 138.
Infectious Diseases 27, 907–913. Connell, T.G., Curtis, N., Ranganathan, S.C. and
Brock, I., Weldingh, K., Lillebaek, T., Follmann, F. Buttery, J.P. (2006) Performance of a whole
and Andersen, P. (2004) Comparison of blood interferon gamma assay for detecting
tuberculin skin test and new specific blood test latent infection with Mycobacterium tuberculosis
in tuberculosis contacts. American Journal of in children. Thorax 61, 616–620.
Respiratory and Critical Care Medicine 170, Connell, T.G., Ritz, N., Paxton, G.A., Buttery, J.P.,
65–69. Curtis, N. and Ranganathan, S.C. (2008) A
Brock, I., Ruhwald, M., Lundgren, B., Westh, H., three-way comparison of tuberculin skin testing,
Mathiesen, L.R. and Ravn, P. (2006) Latent QuantiFERON-TB Gold and T-SPOT.TB in
tuberculosis in HIV positive, diagnosed by the children. PLoS ONE 3, e2624.
M. tuberculosis specific interferon-gamma test. Cooper, A.M. and Flynn, J.L. (1995) The protective
Respiratory Research 7, 56. immune response to Mycobacterium tubercu-
Cain, K.P., Benoit, S.R., Winston, C.A. and losis. Current Opinion in Immunology 7, 512–
MacKenzie, W.R. (2008) Tuberculosis among 516.
foreign-born persons in the United States. Corbett, E.L., Watt, C.J., Walker, N., Maher, D.,
JAMA 300, 405–412. Williams, B.G., Raviglione, M.C., et al. (2003)
Centers for Disease Control and Prevention [CDC] The growing burden of tuberculosis: global
(2010) Updated guidelines for using interferon trends and interactions with the HIV epidemic.
gamma release assays to detect Mycobacterium Archives of Internal Medicine 163, 1009–1021.
tuberculosis infection. MMWR 59, 1–25. Detjen, A., Keil, T., Roll, S., Hauer, B., Mauch, H.,
Chapman, A.L., Munkanta, M., Wilkinson, K.A., Wahn, U., et al. (2007) Interferon-gamma
Pathan, A.A., Ewer, K., Ayles, H., et al. (2002) release assays improve the diagnosis of
Rapid detection of active and latent tuberculosis tuberculosis and nontuberculous mycobacterial
infection in HIV-positive individuals by disease in children in a country with a low
enumeration of Mycobacterium tuberculosis- incidence of tuberculosis. Clinical Infectious
specific T cells. AIDS 16, 2285–2293. Disease 45, 322–328.
Chee, C.B.E., Gan, S.H., Khinmar, K.W., Barkham, Dheda, K., Lalvani, A., Miller, R.F., Scott, G., Booth,
T.M., Koh, C.K., Liang, S. and Wang, Y.T. (2008) H., Johnson, M.A., et al. (2005) Performance of
Comparison of sensitivities of two commercial a T-cell-based diagnostic test for tuberculosis
gamma interferon release assays for pulmonary infection in HIV-infected individuals is inde-
tuberculosis. Journal of Clinical Microbiology pendent of CD4 cell count. AIDS 19, 2038–2041.
46, 1935–1940. Diel, R., Nienhaus, A. and Loddenkemper, R.
Chun, J.K., Kim, C.K., Kim, H.S., Jung, G.Y., Lee, (2007) Cost-effectiveness of interferon-
T.J., Kim, K.H., et al. (2008) The role of a whole {gamma} release assay screening for latent
blood interferon-gamma assay for the detection tuberculosis infection treatment in Germany.
of latent tuberculosis infection in Bacille Chest 131, 1424–1434.
Diel, R., Loddenkemper, R., Meywald-Walter, K., population. Clinical and Experimental Medicine
Gottschalk, R. and Nienhaus, A. (2008a) 10,173–177.
Comparative performance of tuberculin skin Goletti, D., Carrara, S., Vincenti, D. and Girardi, E.
test, QuantiFERON-TB-Gold in tube assay, and (2007) T cell responses to commercial
T-Spot.TB test in contact investigations for Mycobacterium tuberculosis specific antigens in
tuberculosis. Chest 135, 1010–1018. HIV-infected patients. Clinical Infectious
Diel, R., Loddenkemper, R., Meywald-Walter, K., Diseases 45, 1652–1654.
Niemann, S. and Nienhaus, A. (2008b) Predictive Grzybowski, S. and Enarson, D. (1978) Results in
value of a whole blood IFN-gamma assay for the pulmonary tuberculosis patients under various
development of active tuberculosis disease after treatment program conditions. Bulletin Inter-
recent infection with Mycobacterium national Union against Tuberculosis and Lung
tuberculosis. American Journal of Respiratory Disease 53, 70–75.
and Critical Care Medicine 177, 1164–1170. Haldar, P., Thuraisingham, H., Hoskyns, W. and
Diel, R., Loddenkemper, R., Niemann, S., Meywald- Woltmann, G. (2009) Contact screening with
Walter, K. and Nienhaus, A. (2010a) Negative single-step TIGRA testing and risk of active TB
and positive predictive value of a whole-blood infection: The Leicester Cohort Analysis. Thorax
IGRA for developing active TB – an update. 64, A13.
American Journal of Respiratory and Critical Harada, N., Higuchi, K., Yoshiyama, T., Kawabe, Y.,
Care Medicine 183, 88–95. Fujita, A., Sasaki, Y., et al. (2008) Comparison
Diel, R., Loddenkemper, R. and Nienhaus, A. of the sensitivity and specificity of two whole
(2010b) Evidence-based comparison of com- blood interferon-gamma assays for M.
mercial interferon-gamma release assays for tuberculosis infection. Journal of Infection 56,
detecting active TB. Chest 137, 952–968. 348–353.
Dogra, S., Narang, P., Mendiratta, D.K., Chaturvedi, Harboe, M., Oettinger, T., Wiker, H.G., Rosenkrands,
P., Reingold, A.L., Colford, J.J.M., et al. (2007) I. and Andersen, P. (1996) Evidence for
Comparison of a whole blood interferon-gamma occurrence of the ESAT-6 protein in Myco-
assay with tuberculin skin testing for the bacterium tuberculosis and virulent Myco-
detection of tuberculosis infection in hospitalized bacterium bovis and for its absence in
children in rural India. Journal of Infection 54, Mycobacterium bovis BCG. Infection and
267–276. Immunity 64, 16–22.
Dominguez, J., Ruiz-Manzano, J., De Souza- Hill, P.C., Brookes, R.H., Adetifa, I.M., Fox, A.,
Galvao, M., Latorre, I., Mila, C., Blanco, S., et al. Jackson-Sillah, D., Lugos, M.D., et al. (2006)
(2008) Comparison of two commercially Comparison of enzyme-linked immunospot
available gamma interferon blood tests for assay and tuberculin skin test in healthy children
immunodiagnosis of tuberculosis. Clinical and exposed to Mycobacterium tuberculosis.
Vaccine Immunology 15, 168–171. Pediatrics 117, 1542–1548.
Dosanjh, D.P., Hinks, T.S., Innes, J.A., Deeks, J.J., Hill, P.C., Brookes, R.H., Fox, A., Jackson-Sillah,
Pasvol, G., Hackforth, S., et al. (2008) Improved D., Jeffries, D.J., Lugos, M.D., et al. (2007)
diagnostic evaluation of suspected tuberculosis. Longitudinal assessment of an ELISPOT test
Annuals of Internal Medicine 148, 325–336. for Mycobacterium tuberculosis infection. PLoS
Dye, C., Scheele, S., Dolin, P., Pathania, V. and Medicine 4, e192.
Raviglione, M.C. (1999) Consensus statement. Hill, P.C., Jackson-Sillah, D.J., Fox, A., Brookes,
Global burden of tuberculosis: estimated R.H., De Jong, B.C., Lugos, M.D., et al. (2008)
incidence, prevalence, and mortality by country. Incidence of tuberculosis and the predictive
WHO Global Surveillance and Monitoring value of ELISPOT and Mantoux tests in
Project. JAMA 282, 677–686. Gambian case contacts. PLoS ONE 3, e1379.
Ewer, K., Deeks, J., Alvarez, L., Bryant, G., Waller, Hinks, T.S.C., Dosanjh, D.P.S., Innes, J.A., Pasvol,
S., Andersen, P., et al. (2003) Comparison of G., Hackforth, S., Varia, H., et al. (2009)
T-cell-based assay with tuberculin skin test for Frequencies of region of difference 1 antigen-
diagnosis of Mycobacterium tuberculosis infec- specific but not purified protein derivative-
tion in a school tuberculosis outbreak. The specific gamma interferon-secreting T cells
Lancet 361, 1168–1173. correlate with the presence of tuberculosis
Gogus, F., Gunendi, Z., Karakus, R., Erdogan, Z., disease but do not distinguish recent from
Hizel, K. and Atalay, F. (2009) Comparison of remote latent infections. Infection and Immunity
tuberculin skin test and QuantiFERON-TB gold 77, 5486–5495.
in tube test in patients with chronic inflammatory Horsburgh, C.R. Jr (2004) Priorities for the
diseases living in a tuberculosis endemic treatment of latent tuberculosis infection in the
United States. The New England Journal of Lalvani, A. and Millington, K.A. (2008a) T cells and
Medicine 350, 2060–2067. tuberculosis: beyond interferon-gamma. The
Houk, V.N., Baker, J.H., Sorensen, K. and Kent, Journal of Infectious Diseases 197, 941–943.
D.C. (1968) The epidemiology of tuberculosis Lalvani, A. and Millington, K.A. (2008b) Screening
infection in a closed environment. Archives of for tuberculosis infection prior to initiation of
Environmental Health 16, 26–35. anti-TNF therapy. Autoimmunity Reviews 8,
Jones, S., De Gijsel, D., Wallach, F.R., Gurtman, 147–152.
A.C., Shi, Q. and Sacks, H. (2007) Utility of Lalvani, A. and Millington, K.A. (2008c) T-cell
QuantiFERON-TB Gold in-tube testing for latent interferon-{gamma} release assays: can we do
TB infection in HIV-infected individuals. The better? European Respiratory Journal 32,
International Journal of Tuberculosis and Lung 1428–1430.
Disease 11, 1190–1195. Lalvani, A. and Pareek, M. (2009) A 100 year
Kampmann, B., Whittaker, E., Williams, A., Walters, update on diagnosis of tuberculosis infection.
S., Gordon, A., Martinez-Alier, N., et al. (2009) British Medical Bulletin 93, 69–84.
Interferon- release assays do not identify more Lalvani, A., Pathan, A.A., Durkan, H., Wilkinson,
children with active tuberculosis than the K.A., Whelan, A., Deeks, J.J., et al. (2001a)
tuberculin skin test. European Respiratory Enhanced contact tracing and spatial tracking of
Journal 33, 1374–1382. Mycobacterium tuberculosis infection by
Kang, Y.A., Lee, H.W., Yoon, H.I., Cho, B., Han, enumeration of antigen-specific T cells. The
S.K., Shim, Y.S., et al. (2005) Discrepancy Lancet 357, 2017–2021.
between the tuberculin skin test and the whole- Lalvani, A., Pathan, A.A., McShane, H., Wilkinson,
blood interferon gamma assay for the diagnosis R.J., Latif, M., Conlon, C.P., et al. (2001b) Rapid
of latent tuberculosis infection in an intermediate detection of Mycobacterium tuberculosis
tuberculosis-burden country. JAMA 293, 2756– infection by enumeration of antigen-specific T
2761. cells. American Journal of Respiratory and
Kenyon, T.A., Valway, S.E., Ihle, W.W., Onorato, I.M. Critical Care Medicine 163, 824–828.
and Castro, K.G. (1996) Transmission of Lee, J.Y., Choi, H.J., Park, I.N., Hong, S.B., Oh,
multidrug-resistant Mycobacterium tuberculosis Y.M., Lim, C.M., et al. (2006) Comparison of two
during a long airplane flight. The New England commercial interferon-gamma assays for
Journal of Medicine 334, 993–938. diagnosing Mycobacterium tuberculosis infec-
Kik, S.V., Franken, W.P.J., Mensen, M., Cobelens, tion. European Respiratory Journal 28, 24–30.
F.G.J., Kamphorst, M., Arend, S.M., et al. (2010) Leung, C.C., Yam, W.C., Yew, W.W., Ho, P.L., Tam,
Predictive value for progression to tuberculosis C.M., Law, W.S., et al. (2010) T-Spot.TB
by IGRA and TST in immigrant contacts. outperforms tuberculin skin test in predicting
European Respiratory Journal 35, 1346–1353. tuberculosis disease. American Journal of
Kobashi, Y., Sugiu, T., Mouri, K., Obase, Y., Respiratory and Critical Care Medicine 182,
Miyashita, N. and Oka, M. (2009) Indeterminate 834–840.
results of QuantiFERON TB-2G test performed Liebeschuetz, S., Bamber, S., Ewer, K., Deeks, J.,
in routine clinical practice. European Respiratory Pathan, A.A. and Lalvani, A. (2004) Diagnosis of
Journal 33, 812–815. tuberculosis in South African children with a
Laffitte, E., Janssens, J.P., Roux-Lombard, P., T-cell-based assay: a prospective cohort study.
Thielen, A.M., Barde, C., Marazza, G., et al. The Lancet 364, 2196–2203.
(2009) Tuberculosis screening in patients with Lienhardt, C., Fielding, K., Hane, A.A., Niang, A.,
psoriasis before antitumour necrosis factor Ndao, C.T., Karam, F., et al. (2010) Evaluation of
therapy: comparison of an interferon-gamma the prognostic value of IFN- release assay and
release assay vs. tuberculin skin test. British tuberculin skin test in household contacts of
Journal of Dermatology 161, 797–800. infectious tuberculosis cases in Senegal. PLoS
Lalvani, A. (2003) Spotting latent infection: the path ONE 5, e10508.
to better tuberculosis control. Thorax 58, 916– Lighter, J., Rigaud, M., Eduardo, R., Peng, C.H. and
918. Pollack, H. (2009) Latent tuberculosis diagnosis
Lalvani, A. (2007) Diagnosing tuberculosis infection in children by using the QuantiFERON-TB Gold
in the 21st century: new tools to tackle an old in-tube test. Pediatrics 123, 30–37.
enemy. Chest 131, 1898–1906. Liu, X.-Q., Dosanjh, D., Varia, H., Ewer, K., Cockle,
Lalvani, A. and Millington, K.A. (2007) T cell-based P., Pasvol, G., et al. (2004) Evaluation of T-cell
diagnosis of childhood tuberculosis infection. responses to novel RD1- and RD2-encoded
Current Opinion in Infectious Diseases 20, 264– Mycobacterium tuberculosis gene products
271. for specific detection of human tuberculosis
infection. Infection and Immunity 72, 2574– Mori, T., Sakatani, M., Yamagishi, F., Takashima, T.,
2581. Kawabe, Y., Nagao, K., et al. (2004) Specific
Luetkemeyer, A.F., Charlebois, E.D., Flores, L.L., detection of tuberculosis infection: an interferon-
Bangsberg, D.R., Deeks, S.G., Martin, J.N., et gamma-based assay using new antigens.
al. (2007) Comparison of an interferon-{gamma} American Journal of Respiratory and Critical
release assay with tuberculin skin testing in Care Medicine 170, 59–64.
HIV-infected individuals. American Journal of Murray, P.J. (1999) Defining the requirements for
Respiratory and Critical Care Medicine 175, immunological control of mycobacterial infec-
737–742. tions. Trends in Microbiology 7, 366–372.
MacPherson, D.W. and Gushulak, B.D. (2006) Mutsvangwa, J., Millington, K.A., Chaka, K.,
Balancing prevention and screening among Mavhudzi, T., Cheung, Y.-B., Mason, P.R., et al.
international migrants with tuberculosis: (2010) Identifying recent Mycobacterium
population mobility as the major epidemiological tuberculosis transmission in the setting of high
influence in low-incidence nations. Public Health HIV and TB burden. Thorax 65, 315–320.
120, 712–723. Nakaoka, H., Lawson, L., Squire, S.B., Coulter, B.,
Mahairas, G.G., Sabo, P.J., Hickey, M.J., Singh, Ravn, P., Brock, I., et al. (2006) Risk for
D.C. and Stover, C.K. (1996) Molecular analysis tuberculosis among children. Emerging Infec-
of genetic differences between Mycobacterium tious Diseases 12, 1383–1388.
bovis BCG and virulent M. bovis. Journal of National Collaborating Centre for Chronic
Bacteriology 178, 1274–1282. Conditions (2006) Tuberculosis: Clinical Diag-
Mahomed, H., Hawkridge, T., Verver, S., Abrahams, nosis and Management of Tuberculosis, and
D., Geiter, L., Hatherill, M., et al. (2011) The Measures for its Prevention and Control. Royal
tuberculin skin test versus QuantiFERON TB College of Physicians, London.
Gold® in predicting tuberculosis disease in an Nicol, M.P., Davies, M.-A., Wood, K., Hatherill, M.,
adolescent cohort study in South Africa. PLoS Workman, L., Hawkridge, A., et al. (2009)
ONE 6(3), e17984. Comparison of T-SPOT.TB assay and tuberculin
Mandalakas, A.M., Hesseling, A.C., Chegou, N.N., skin test for the evaluation of young children at
Kirchner, H.L., Zhu, X., Marais, B.J., et al.
high risk for tuberculosis in a community setting.
(2008) High level of discordant IGRA results in
Pediatrics 123, 38–43.
HIV-infected adults and children. The
Okada, K., Mao, T.E., Mori, T., Miura, T., Sugiyama,
International Journal of Tuberculosis and Lung
T., Yoshiyama, T., et al. (2008) Performance of
Disease 12, 417–423.
an interferon-gamma release assay for diag-
Martin, J., Walsh, C., Gibbs, A., McDonnell, T.,
nosing latent tuberculosis infection in children.
Fearon, U., Keane, J., et al. (2009) Comparison
Epidemiology and Infection 136, 1179–1187.
of interferon-{gamma}-release assays and
Pareek, M., Watson, J.P., Omerod, L.P., Kon, O.M.,
conventional screening tests before tumour
Woltmann, G., White, P.J., et al. (2011)
necrosis factor-{alpha} blockade in patients with
Screening of immigration in the UK for imported
inflammatory arthritis. Annals of the Rheumatic
latent tuberculosis: a multicentre cohort study
Diseases 69, 181–185.
and cost-effectiveness analysis. Lancet
Matulis, G., Juni, P., Villiger, P.M. and Gadola, S.D.
(2008) Detection of latent tuberculosis in Infectious Diseases 11, 434–444.
immunosuppressed patients with autoimmune Pai, M., Zwerling, A. and Menzies, D. (2008)
diseases: performance of a Mycobacterium Systematic review: T-cell-based assays for the
tuberculosis antigen-specific interferon gamma diagnosis of latent tuberculosis infection: an
assay. Annals of the Rheumatic Diseases 67, update. Annuals of Internal Medicine 149, 177–
84–90. 184.
Mazurek, G.H., Jereb, J., Lobue, P., Iademarco, Place, S., Verscheure, V., De San, N., Hougardy,
M.F., Metchock, B., Vernon, A., et al. (2005) J.-M., Schepers, K., Dirix, V., et al. (2010)
Guidelines for using the QuantiFERON-TB Gold Heparin-binding, hemagglutinin-specific IFN-
test for detecting Mycobacterium tuberculosis {gamma} synthesis at the site of infection during
infection, United States. MMWR Recom- active tuberculosis in humans. American
mendations and Reports 54, 49–55. Journal of Respiratory and Critical Care
Millington, K.A., Innes, J.A., Hackforth, S., Hinks, Medicine 182, 848–854.
T.S., Deeks, J.J., Dosanjh, D.P., et al. (2007) Ponce de Leon, D., Acevedo-Vasquez, E., Alvizuri,
Dynamic relationship between IFN-gamma and S.G.C. and Cucho, M., et al. (2008) Comparison
IL-2 profile of Mycobacterium tuberculosis- of an interferon-gamma assay with tuberculin
specific T cells and antigen load. Journal of skin testing for detection of tuberculosis (TB)
Immunology 178, 5217–5226. infection in patients with rheumatoid arthritis in
a TB-endemic population. Journal of Rheumato- American Journal of Gastroenterology 103,

logy 35, 776–781. 2799–2806.
Public Health Agency of Canada (2007) Interferon Sellam, J., Hamdi, H., Roy, C., Baron, G., Lemann,
gamma release assays for latent tuberculosis M., Puechal, X., et al. (2007) Comparison of in
infection: an Advisory Committee Statement. vitro-specific blood tests with tuberculin skin
Canada Communicable Disease Report 33, test for diagnosis of latent tuberculosis before
1–18. anti-TNF therapy. Annals of the Rheumatic
Qumseya, B.J., Ananthakrishnan, A.N., Skaros, S., Diseases 66, 1610–1615.
Bonner, M., Issa, M., Zadvornova, Y., et al. Sester, M., Sotgiu, G., Lange, C., Giehl, C., Girardi,
(2011) QuantiFERON TB gold testing for E., Migliori, G.B., et al. (2011) Interferon- release
tuberculosis screening in an inflammatory assays for the diagnosis of active tuberculosis: a
bowel disease cohort in the United States. systematic review and meta-analysis. European
Inflammatory Bowel Diseases 17, 77–83. Respiratory Journal 37, 100–111.
Raby, E., Moyo, M., Devendra, A., Banda, J., De Shams, H., Weis, S.E., Klucar, P., Lalvani, A.,
Haas, P., Ayles, H., et al. (2008) The effects of Moonan, P.K., Pogoda, J.M., et al. (2005)
HIV on the sensitivity of a whole blood IFN- Enzyme-linked immunospot and tuberculin skin
release assay in Zambian adults with active testing to detect latent tuberculosis infection.
tuberculosis. PLoS ONE 3, e2489. American Journal of Respiratory and Critical
Rangaka, M.X., Diwakar, L., Seldon, R., Van Care Medicine 172, 1161–1168.
Cutsem, G., Meintjes, G.A., Morroni, C., et al. Shovman, O., Anouk, M., Vinnitsky, N., Arad,
(2007a) Clinical, immunological, and epidemio- U., Paran, D., Litinsky, I., et al. (2009)
logical importance of antituberculosis T cell QuantiFERON-TB Gold in the identification of
responses in HIV- infected Africans. Clinical latent tuberculosis infection in rheumatoid
Infectious Diseases 44, 1639–1646. arthritis: a pilot study. International Journal of
Rangaka, M.X., Wilkinson, K.A., Seldon, R., Van Tuberculosis and Lung Disease 13, 1427–1432.
Cutsem, G., Meintjes, G.A., Morroni, C., et al. Soysal, A., Millington, K.A., Bakir, M., Dosanjh, D.,
(2007b) Effect of HIV-1 infection on T-cell-based Aslan, Y., Deeks, J.J., et al. (2005) Effect of
and skin test detection of tuberculosis infection. BCG vaccination on risk of Mycobacterium
American Journal of Respiratory and Critical tuberculosis infection in children with household
Care Medicine 175, 514–520. tuberculosis contact: a prospective community-
Rangaka, M.X., Wilkinson, K.A., Glynn, J.R., Ling, based study. The Lancet 366, 1443–1451.
D., Menzies, D., Mwansa-Kambafwile, J., et al. Soysal, A., Torun, T., Efe, S., Gencer, H., Tahaoglu,
(2012) Predictive value of interferon-gamma K. and Bakir, M. (2008) Evaluation of cut-off
release assays for incident active tuberculosis: values of interferon-gamma-based assays in
a systematic review and meta-analysis. The the diagnosis of M. tuberculosis infection. The
Lancet Infectious Diseases 12, 45–55. International Journal of Tuberculosis and Lung
Richeldi, L. (2006) An update on the diagnosis of Disease 12, 50–56.
tuberculosis infection. American Journal of Stead, W.W. (1978) Undetected tuberculosis in
Respiratory and Critical Care Medicine 174, prison. Source of infection for community at
736–742. large. JAMA 240, 2544–2547.
Richeldi, L., Ewer, K., Losi, M., Bergamini, B.M., Stephan, C., Wolf, T., Goetsch, U., Bellinger, O.,
Roversi, P., Deeks, J., et al. (2004) T cell-based Nisius, G., Oremek, G., et al. (2008) Comparing
tracking of multidrug resistant tuberculosis QuantiFERON-tuberculosis gold, T-SPOT
infection after brief exposure. American Journal tuberculosis and tuberculin skin test in HIV-
of Respiratory and Critical Care Medicine 170, infected individuals from a low prevalence
288–295. tuberculosis country. AIDS 22, 2471–2479.
Ruhwald, M., Bodmer, T., Maier, C., Jepsen, M., Takahashi, H., Shigehara, K., Yamamoto, M.,
Haaland, M.B., Eugen-Olsen, J., et al. (2008) Suzuki, C., Naishiro, Y., Tamura, Y., et al. (2007)
Evaluating the potential of IP-10 and MCP-2 as Interferon gamma assay for detecting latent
biomarkers for the diagnosis of tuberculosis. tuberculosis infection in rheumatoid arthritis
European Respiratory Journal 32, 1607–1615. patients during infliximab administration.
Schoepfer, A.M., Flogerzi, B., Fallegger, S., Rheumatology International 27, 1143–1148.
Schaffer, T., Mueller, S., Nicod, L., et al. (2008) Talati, N., Seybold, U., Humphrey, B., Aina, A.,
Comparison of interferon-gamma release assay Tapia, J., Weinfurter, P., et al. (2009) Poor
versus tuberculin skin test for tuberculosis concordance between interferon-gamma
screening in inflammatory bowel disease. The release assays and tuberculin skin tests in
diagnosis of latent tuberculosis infection among Vincenti, D., Carrara, S., Butera, O., Bizzoni, F.,
HIV-infected individuals. BMC Infectious Casetti, R., Girardi, E., et al. (2007) Response
Diseases 9, 15. to region of difference 1 (RD1) epitopes in
Tsiouris, S.J., Austin, J., Toro, P., Coetzee, D., human immunodeficiency virus (HIV)-infected
Weyer, K., Stein, Z., et al. (2006a) Results of a individuals enrolled with suspected active
tuberculosis-specific IFN-gamma assay in tuberculosis: a pilot study. Clinical and
children at high risk for tuberculosis infection. Experimental Immunology 150, 91–98.
The International Journal of Tuberculosis and World Health Organization [WHO] (1998) Policy
Lung Disease 10, 939–941. Statement on Preventive Therapy Against
Tsiouris, S.J., Coetzee, D., Toro, P.L., Austin, J., Tuberculosis in People Living with HIV. World
Stein, Z. and El-Sadr, W. (2006b) Sensitivity Health Organization, Geneva, Switzerland.
analysis and potential uses of a novel gamma Wrighton-Smith, P. and Zellweger, J.P. (2006) Direct
interferon release assay for diagnosis of costs of three models for the screening of latent
tuberculosis. Journal of Clinical Microbiology tuberculosis infection. European Respiratory
44, 2844–2850. Journal 28, 45–50.
Vassilopoulos, D., Stamoulis, N., Hadziyannis, E. Yoshiyama, T., Harada, N., Higuchi, K., Sekiya, Y.
and Archimandritis, A.J. (2008) Usefulness of and Uchimura, K. (2010) Use of the
enzyme-linked immunosorbent assay (Elispot) QuantiFERON-TB Gold test for screening
compared to tuberculin skin testing for latent tuberculosis contacts and predicting active
tuberculosis screening in rheumatic patients disease. The International Journal of Tubercu-
scheduled for anti-tumor necrosis factor losis and Lung Disease 14, 819–827.
treatment. The Journal of Rheumatology 35, Zellweger, J.P., Zellweger, A., Ansermet, S., De
1271–1276. Senarclens, B. and Wrighton-Smith, P. (2005)
Verver, S., Warren, R.M., Munch, Z., Richardson, Contact tracing using a new T-cell-based test:
M., Van der Spuy, G.D., Borgdorff, M.W., et al. better correlation with tuberculosis exposure
(2004) Proportion of tuberculosis transmission than the tuberculin skin test. International
that takes place in households in a high- Journal of Tuberculosis and Lung Disease 9,
incidence area. The Lancet 363, 212–214. 1242–1247.
5 Measuring Tuberculosis Immune
Responses in the Lung –
The Correct Target?
Ronan Breen1 and Marc Lipman2

1Department of Respiratory Medicine, Guys’ and St Thomas’ NHS Trust, London,
UK; 2Centre for Respiratory Medicine, University College London, UK
5.1 Introduction. Why is the Lung duced during active pulmonary tuberculosis
Important? (Loudon et al., 1969). Transmission dynamics
are poorly understood but are likely to reflect
Tuberculosis is associated with staggering physical factors such as aerosol velocity,
global morbidity and mortality. Around 9 distance between individuals at the time of
million people develop active disease and coughing, sputum composition and how
almost 2 million die each year. Tuberculosis mycobacterial strains interact with the innate
immunology is highly relevant to disease immune system and so alter virulence
control. It provides information on correlates (Fennelly et al., 2004; Reed et al., 2004). This
of immune protection and failure against variation, along with the initial protection
initial exposure to Mycobacterium tuberculosis from infection provided by the physical
infection as well as on the (possible) barrier of the upper airway and the defence
subsequent development of active disease. mechanisms of the innate immune system,
These data are of crucial importance to the probably explains why transmission from an
development of an effective vaccine and also infective source is usually only to close
have a role in diagnosis and monitoring of household contacts, with non-household
treatment response. In this chapter, we transmission thought to require at least 8 h of
discuss how the immune response in contact in a poorly ventilated environment
tuberculosis has been studied in the lung. We (Kenyon et al., 1996). This may also explain
focus on tuberculosis pathogenesis as well as why some individuals who experience
diagnosis; and evaluate the status of current multiple exposures show no evidence of
research. Finally, we reflect on the value infection (demonstrated by a repeatedly
gained from lung-based work and review negative tuberculin skin test, TST) (Morrison
future directions in tuberculosis lung et al., 2008).
immunology programmes. The physical parameters of airway angle
M. tuberculosis can cause disease in a and calibre determine where infected drop-
variety of organs, although by far the most lets are deposited within the lung (Murray,
important of these clinically is the lung. 2003). The typical site of deposition, as judged
Transmission of M. tuberculosis between by the area in which a primary focus of
individuals occurs via inhalation of cough- infection is established, is in the periphery of
generated infected airborne droplets pro- the right lung base. It is here that initial

68 R. Breen and M. Lipman
infection can be established and the adaptive transmission from person to person, the lung
immune response begins to be employed. has an intuitive appeal as a site of investi-
Adaptive immunity to M. tuberculosis both gation. However, the fairly invasive nature of
within the lung and elsewhere has been techniques that are required to sample either
studied extensively and may be altered by the the peripheral lung or the associated lymph
presence of states of lowered immunity such nodes means that reliance has been placed on
as HIV infection, the administration of the use of animal models to try to understand
immune-suppressing drugs or a previously the pathogenesis of tuberculosis.
primed response, as is postulated following
Bacille Calmette–Guerin (BCG) vaccination in
children. These factors are particularly 5.2 Animal Models
important in understanding the rapid
progression to clinical disease (which may be A useful animal model of tuberculosis should
in the lung or associated draining lymph replicate relevant aspects of latent infection
nodes; or disseminated via the bloodstream or active disease and be predictive of
to cause miliary or central nervous system outcomes (with, for example, new drugs
disease) that occurs in around 5% of infected when these are given subsequently to
cases compared to the establishment of a state humans). They should also be able to inform
of latent tuberculosis infection (LTBI) found future work and not be so expensive to use
in the other 95% of infected individuals. In that they become unfeasible economically in
the 5–10% of latently infected individuals in long-term studies. At present, no single
whom active disease develops at a later stage, model satisfies all of these criteria; and the
the most common site is the lung (either in reagents needed to work with the models are
isolated form or together with active disease not necessarily available. However, one great
at a non-pulmonary site). This, plus its advantage of tuberculosis animal systems is
importance as the source of onward trans- that tissue samples can be obtained with
mission, ensures the primacy of the lung and relative ease, although the detected tissue
its local immune environment at all stages of response may not correspond precisely to
tuberculosis. that found in humans. Further, one can
The pathological response elicited by generally standardize the model under test
infection with M. tuberculosis is to produce (something quite different to human research,
inflammation characterized by granuloma which almost always involves participants
formation. Granulomas are complex and with considerable genetic and phenotypic
immunologically active structures which diversity). Animal models differ from each
serve to control M. tuberculosis through other and also from humans in both their
interactions mediated by multiple chemo- innate and adaptive responses to myco-
kines and cytokines such as interferon- bacteria, including M. tuberculosis. Thus, they
gamma (IFN-γ), tumour necrosis factor-alpha may have clinical presentations and outcomes
(TNF-α) and interleukin-12 (IL-12), between very different to ours, and extrapolation
lymphocytes and, perhaps most importantly, between animal systems and also to humans
macrophages (Flynn et al., 2011). The function is not always straightforward or warranted
of the granuloma and how it can be disrupted (Rosenthal et al., 2007).
is crucial to our understanding of how Successful animal models include the
immunological control is maintained in those zebrafish, mouse, guinea pig, rabbit and non-
with long-term latent infection; and lost in human primates. In general, and largely
the individuals who develop reactivation depending on how closely an animal’s
tuberculosis, for example in HIV infection immune response and clinical course map
with declining CD4+ lymphocyte numbers that of adult humans, these can be used to
and function, or with therapeutic blockade of investigate specific aspects of human
TNF-α (Keane et al., 2001; Diedrich and Flynn, immunity to mycobacteria. For example, the
2011). In a disease such as tuberculosis, which (transparent) embryonic zebrafish when
is spread almost exclusively by aerosol infected with M. marinum appears to have a
Measuring Tuberculosis Immune Responses in the Lung 69
predominant innate immune response, yet source for material might be patients with
forms granulomas within days of infection previous M. tuberculosis infection who are
(Lesley and Ramakrishnan, 2008). It therefore undergoing lung resection for other
questions both the need for adaptive conditions such as malignancy in whom the
immunity in the initial generation of ongoing successful granulomatous response
granulomas and whether these cell collections can be evaluated.
are beneficial to the host (as granuloma A less invasive approach to lung biopsy
formation occurs when bacterial replication is can be performed via flexible, fibre-optic
maximal, rather than as a response to it), and bronchoscopy. In many centres worldwide,
thus has yielded important insights into the bronchoscopy is used routinely to obtain
encounter at initial infection between microbe diagnostic samples in individuals with
and immune system (Volkman et al., 2004). suspected pulmonary tuberculosis in whom
The mouse, which has a granulomatous spontaneous sputum cannot be produced or
response different from that of humans when is repeatedly acid-fast bacilli (AFB) smear
infected by aerosol (necrosis is not a negative. The physical flexibility and increas-
recognized feature of most mouse models), ingly small diameter of modern broncho-
has been used to model a wide variety of scopes allows sampling to be directed to
infection states, including acute primary specific segments and even subsegments of
infection and latent tuberculosis (Flynn, the lung. The bronchoscope has a working
2006). The rabbit and guinea pig can develop channel down which biopsy forceps can be
necrosis within granulomas, and are perhaps passed, and biopsies of the bronchial mucosa
rather better models than the mouse. How- obtained, under direct vision. To sample
ever, guinea pigs are almost too susceptible to alveolar tissue requires the forceps to be
disease (they get sick and die rapidly when passed out well beyond the vision of the
infected with M. tuberculosis); and laboratory bronchoscopist (although radiological screen-
reagents are harder to obtain than for mice. ing may be used as a guide). The use of the
Rabbits infected with M. bovis have similar transbronchial biopsy technique for research
pathology to human tuberculosis, though are is limited by the relatively frequent morbidity
a lot more expensive to house than guinea due to pneumothorax (4%), which is increased
pigs (Dannenberg, 2001). Non-human primate by the need for multiple biopsies to allow for
models such as cynomolgus macaques have the often small tissue samples obtained plus
been used and may provide valuable data the patchy distribution of granulomas within
much more relevant to understanding human the alveolar tissue (Milman et al., 1994).
tuberculosis, though are very expensive The cell populations and their associated
(Green et al., 2010). cytokines and chemokines within the alveolar
space can be sampled less invasively via
bronchoalveolar lavage (BAL) with saline.
5.3 How Can the Lung be Sampled? Here, the tip of the bronchoscope is wedged
gently into a subsegmental bronchus, fol-
Large tissue samples in human tuberculosis lowed by the instillation of aliquots of normal
are now obtained only rarely due to the saline (usually 30–60 mls up to a total volume
efficacy of modern anti-tuberculosis therapy. of typically between 120 and 360 mls). Fluid
Surgical resection is likely to be performed can then be recovered into an appropriate
only in cases of very severe pulmonary container by applying gentle suction. The
disease, often with multidrug resistance, and volume of lung wash obtained by BAL and
so the findings may not be widely applicable. the function of its contained cells is dependent
However, the study of a small number of on the skill of the operator in maintaining a
resections has demonstrated heterogeneous correct wedged position and the use of
activity in distinct parts of resected tissue and minimal suction pressures (either with wall
similar work may provide insights into the suction or manually via a syringe) to reduce
maintenance and loss of local immunological airway collapse. BAL fluid returns can be as
control (Kaplan et al., 2003). An alternative much as 75% of the instilled volume of saline,
although this is often lower in severe infection is the use of a ‘blind’ catheter, which is passed
or when airways disease is present (such as in through the airway into the lung without
smokers). It is affected also by which lung direct visualization. This would reduce sub-
area has been sampled (there being smaller stantially the need for expensive equipment,
returns from the upper lobes). The ability to but removes the directed nature of sampling
‘drive’ the bronchoscope enables more than and few data exist to support its use in
one segment or lobe to be sampled during a practice (Ashton, 1992). Much better evalu-
single procedure. Thus, useful information ated is sputum induction using inhalation of
can be obtained from both affected and hypertonic saline. A number of studies have
unaffected parts of the lung. Direct evalu- shown that this non-invasive method has a
ations against lung biopsy specimens have similar microbiological yield to BAL in
suggested that the cell populations obtained tuberculosis diagnosis, yet at a substantially
from BAL, in which up to 1 million alveoli reduced cost and in a safe and more
may be sampled, are similar to those within participant-friendly manner, which makes
the interstitium in which granulomas reside repeat sampling more feasible (McWilliams et
(Law et al., 1996). al., 2002; Dunleavy et al., 2008). The technique
Limitations on the use of bronchoscopy involves the subject inhaling hypertonic
do exist. The capital cost of equipment is saline (3–7%) delivered via an ultrasonic
high, and effective ongoing maintenance is nebulizer. This is thought to modulate an
vital to ensure adequate performance and to increase in lung lining fluid through osmosis.
prevent cross-infection (Agerton et al., 1997). The fluid and the cells contained within it are
A high degree of technical competence is carried up by the mucociliary escalator before
required to ensure consistent and high- being expectorated. Sputum induction has
volume returns of BAL fluid. Although only been used for a number of years to observe
semi-invasive, many individuals undergoing changes in leucocyte populations and liquid
bronchoscopy will require, in addition to phase markers of inflammation in airways
topical anaesthesia, intravenous therapy if disease such as asthma and chronic obstruc-
not conscious sedation for anxiolysis and to tive pulmonary disease (COPD) (Pin et al.,
make the procedure tolerable (Rees et al., 1992). Most data suggest that leucocyte
1983). This adds both a small amount of extra populations in induced sputum differ when
risk and the need for more post-procedure compared to BAL, especially in the proportion
monitoring, but may be particularly import- of neutrophils and macrophages present in
ant if a repeat procedure is contemplated. the obtained sample (Pizzichini et al., 1998;
Serious complications, however, are rare. Fireman et al., 1999). This may reflect that the
Minor adverse effects such as post-procedure cell populations arise from different levels of
cough, sore throat, worsening of fever and the bronchial tree rather than predominantly
breathlessness are common but short-lived. the alveolar space; and also that expectorated
These limitations probably explain the sputum comes from several areas of the lung
relative paucity of studies using BAL to rather than a single lung segment as with
examine the immune response in human BAL. The heterogeneous origin of induced
tuberculosis. When BAL is performed for sputum may be of advantage when, for
diagnostic purposes, there is often ‘surplus’ example, studying the immune protection
fluid available for consented research. How- produced within the lung by a candidate
ever, finding individuals who will undergo vaccine. Of interest, a study of subjects with
bronchoscopy purely as a research procedure the granulomatous condition sarcoidosis
is more difficult. Studies using this approach demonstrated similar results with both
are not common and generally are of small techniques with respect to lymphocyte
size. It is not surprising, therefore, that number and subpopulations, but comparable
longitudinal studies with repeat BAL are data are not available for individuals with
rare. tuberculosis (Moodley et al., 2000; Tsiligianni
An alternative approach to fibre-optic et al., 2002). One feature of induced sputum
bronchoscopy that simplifies lung sampling which needs to be borne in mind is that the
majority of cells expectorated are contained chemokine response characterized by a CD4+

within mucus. In order to break down this lymphocytosis in contrast to an immune-
mucus and so produce a liquid cell suspension suppressed profile found in peripheral blood
and maximize cell numbers, the sample must (Sadek et al., 1998; Schwander et al., 1998;
be treated with the potent reducing agent Hirsch et al., 1999).
dithiothreitol (DTT). This comes at the cost of Human studies have, to date, focused on
reduced viable cell numbers, as assessed by this adaptive immune response, reflecting the
immediate trypan blue staining, and variable generally better understanding of adaptive
effects on staining for cell surface markers compared to innate immunity to M.
using monoclonal antibodies (Efthimiadis et tuberculosis provided by animal studies, the
al., 1997; Loppow et al., 2000). The effects of greater apparent relevance to clinical disease
DTT on specific molecules and novel bio- in situations such as advanced HIV infection
markers will require careful evaluation. and the ability to harness adaptive response
Finally, the Lung Flute™ has been for diagnosis. The initial response to the
developed to achieve successful sputum inhalation of aerosolized M. tuberculosis is its
induction. This self-powered device requires engulfment by macrophages found in large
the subject to blow into a plastic chamber, numbers in the bronchoalveolar compart-
which contains a thin reed. Sound waves are ment. It is in these cells that the organism is
generated that vibrate bronchial secretions, able to survive and multiply. An innate
which can then be expectorated. This cheap immune response, the understanding of
plastic device requires no external power and which has accelerated in recent years through
has been shown to produce positive micro- animal studies identifying pattern recognition
biology efficiently in subjects with tubercu- receptors such as Toll-like receptors (TLRs), is
losis, though detailed cellular data are then activated, which in turn appears to have
awaited (Fujita et al., 2009). It is important to a role in generating an adaptive immune
stress that any technique which stimulates response (Akira et al., 2001; Saiga et al., 2011).
sputum production and cough can promote TLRs recognize a variety of mycobacterial
M. tuberculosis transmission. Stringent infec- components and signal via an adaptor
tion control measures such as the use of molecule, MyD88, which is also part of the
negative pressure facilities are integral, there- signalling pathway of IL-1 and IL-18. The in
fore, to such procedures. vivo importance of this molecule has been
demonstrated by the susceptibility to acute
infection of MyD88- deficient mice, despite
5.4 Why Sample the Lung Rather being able to produce an adaptive immune
Than Blood? – Human Studies response (Fremond et al., 2004). Clear differ-
ences exist between alveolar macrophages
Obtaining peripheral blood samples is and peripheral blood-derived monocytes
generally much easier than the methods with respect to expression of TLRs (Juarez et
described for lung sampling; and to justify al., 2010). Production of IL-12 and IL-18 by
this extra difficulty there must be clear macrophages and dendritic cells induces
differences in the cell composition and IFN-γ production by T-cells, natural killer
function of the two compartments. In healthy (NK) cells and NK T-cells, which in turn
subjects, this is certainly the case, with BAL activates macrophages and promotes
samples dominated by alveolar macrophages intracellular killing of M. tuberculosis. The
(90–95% of the total), while lymphocytes con- proinflammatory milieu that is produced
tribute 5–10%. This compares to peripheral facilitates chemokine attraction of further
blood where 50–70% are neutrophils, 15–45% immune cells including γδ T-cells and
lymphocytes and 0–10% monocytes neutrophils (Korbel et al., 2008). In vitro
(Schwander and Dheda, 2011). In subjects human studies have demonstrated the ability
with active tuberculosis, the typical pulmon- of neutrophils to both phagocytose M.
ary immune profile observed has been a tuberculosis and restrict its growth via
T-helper cell type 1 (Th1) cytokine and defensin peptides (Kisich et al., 2002;
Martineau et al., 2007). However, neutrophil- longitudinal responses may require simpli-
mediated inflammation may also be patho- fied lung sampling through sputum induction
genic and further work is needed on their role rather than BAL.
in vivo (Schneider et al., 2010). An interesting attempt to model a
Non-TLR pattern recognition receptors, protective response in the human lung has
such as NOD-like receptors and C-type lectin adapted the bronchoscopic challenge tech-
receptors including DC-SIGN, have also been nique used to explore the Th2-like CD4+
implicated in the recognition of M. tubercu- predominant immune response found in
losis. This can lead to the release of a number asthma. This involves bronchoscopic instil-
of locally active effector molecules such as lation of antigen to specific subsegments,
nitric oxide and inducible nitric oxide syn- with subsequent BAL allowing analysis of the
thase, and promoters of vitamin D activation immune response generated. A pilot study
and iron sequestration that restrict M. used PPD as the stimulating antigen mixture
tuberculosis growth in vitro (Saiga et al., 2011). in subjects who were TST positive or negative
Phagosome maturation and autophagy (Silver et al., 2003). Here, PPD challenge was
appear increasingly important in the innate safe in all subjects and demonstrated a huge
response to human intracellular pathogens concentration of CD4+ lymphocytes express-
and further data on their role in tuberculosis ing the CD45RO+, CCR7– phenotype of
are awaited with interest (Deretic, 2011). Our effector memory T-cells. PPD challenge
increasing understanding of the innate produced significant increases in the IFN-γ
pulmonary immune response to M. tubercu- inducible and CXCR3 ligands IP-10 and Mig,
losis will be of importance in the design of a chemokines produced by the resident PPD-
protective inhaled vaccine against tubercu- responsive CD4+ T-cells, with an enrichment
losis. of CD4+ cells expressing the α-4,β-1 integrin
It is perhaps surprising to find a paucity homing molecule (Walrath et al., 2005;
of data evaluating the pulmonary immune Walrath and Silver, 2011).
response of individuals who by virtue of The majority of studies have sampled
having been exposed to an infectious individuals with active tuberculosis in whom,
tuberculosis case though remain well, already by definition, the immune response has
appear able to generate protective immunity. changed from that found during clinically
To date, the best study of such adaptive latent tuberculosis infection. Pleural tubercu-
immunity compared healthy household losis, especially in the setting of recently
contacts of sputum smear-positive cases with acquired infection, may provide a model of
community controls (Schwander et al., 2000). relative control. Here, CD4+ lymphocytes
This demonstrated a compartmentalized with a high frequency of mycobacterial
adaptive immune response to the antigen antigen-specific responders predominate
mixture purified protein derivative (PPD) (Barnes et al., 1993). BAL from subjects with
and to M. tuberculosis Antigen 85, with miliary tuberculosis is also lymphocyte-
enhanced response frequencies observed in predominate, with some work suggesting
BAL over peripheral blood mononuclear cells that an increase in the CD8+ proportion is
(PBMCs) in household contacts and increased associated with slow resolution of the disease
responses to PPD among PBMCs compared (Ainslie et al., 1992). This connection with
to BAL cells in TST-positive community CD8+ T-cells has also been observed in slow
controls. Furthermore, ex vivo restriction of treatment responders who have pulmonary
M. tuberculosis growth in alveolar macro- tuberculosis (Yu et al., 1995). A lymphocyte-
phages was mediated by CD8+ T-cells from predominant alveolitis, with expression
the household contacts but not the controls of proinflammatory Th1 cytokines and
(Carranza et al., 2005). Future work examining associated chemokines as shown by increased
the cohorts of exposed individuals with mRNA, cytokine concentrations in BAL fluid
clinical evidence of a protective immune and ex vivo antigen stimulation assays, has
phenotype will be valuable, though the need often been reported in pulmonary tubercu-
to increase participant accrual and evaluate losis (Robinson et al., 1994). This lympho-
cytosis is associated with clinically less severe with treatment when subjects were resampled
pulmonary tuberculosis characterized by at day 30 (Almeida et al., 2009). The data
infiltrates rather than cavitation, and this may argue for more intensive longitudinal studies,
be a compartmentalized response even which in BAL have demonstrated a trend to
within the lung as differences have been increased BAL lymphocytes and IFN-γ
demonstrated when radiologically uninvolved expression with decreased TNF-α expression
and involved lung segments are compared during treatment (Condos et al., 1998).
(Ainslie et al., 1992; Condos et al., 1998). More Regulatory T-cells (CD4+CD25highFoxP3+)
advanced disease with increased cavitation is have been found at an increased frequency in
characterized by a macrophage and then the peripheral blood of individuals with
neutrophil-dominated bronchoalveolar cell active tuberculosis. Although identifiable in
population, and a consequent reduction in BAL, their importance in human tuberculosis
both lymphocyte proportion and frequency has not been firmly established (Guyot-Revol
of antigen-specific responders within the et al., 2006). In a non-human primate model,
leucocyte population (Barry et al., 2009). their frequency has been shown to be
Whether this provides an insight on the increased in animals with active tuberculosis.
failure or change occurring in the local However, reported data have suggested this
immune response or is merely a reflection of is a response to inflammation rather than a
the occurrence of cavitation is unclear. More driver of disease itself (Green et al., 2010).
relevant in this setting might be the
demonstration of elevated concentrations of
matrix metalloproteinase-1 in both BAL and 5.5 Immune Diagnosis of
induced sputum in active pulmonary Tuberculosis Using the Lung
tuberculosis compared to adult controls who
had other pulmonary infections, and in The TST, which involves subcutaneous or
tuberculosis cases with cavitation compared intradermal injection of PPD, and subsequent
to those without (Elkington et al., 2011). assessment for local tissue immunity, utilizes
The crucial importance of the adaptive the immune response to help diagnose M.
immune response in tuberculosis mediated tuberculosis infection. It has been available for
by CD4+ lymphocytes in conjunction with over a century (von Pirquet, 1907). The
proinflammatory cytokines is shown by the concerns regarding sensitivity (especially in
loss of control of M. tuberculosis seen as conditions of impaired immune function and
CD4 counts fall in advancing HIV infection, active tuberculosis) and reduced specificity
the disseminated infections described in (largely due to BCG vaccination but also to
interferon-gamma knockout mouse models cross-reactivity with non-tuberculous myco-
and humans with IFN-γ receptor deficiency bacteria) are generally highlighted rather
and who use TNF-α blocking agents (Cooper more than its manifest strengths. It has stood
et al., 1993; Flynn et al., 1993; Jones et al., 1993; the test of time because it is cheap, easy to
Keane et al., 2001). This focus on the protective perform and evaluate, repeatable and can be
arm of the immune response despite the undertaken without need of laboratory
clinical picture of loss of protective control infrastructure or a reliable power supply.
has perhaps distracted attention from the However, advances in the techniques avail-
investigation of suppressive mechanisms, able to measure antigen-specific cellular
which might explain this conundrum. A immune response, such as the development
number of studies have struggled to of the ELISpot assay and intracellular
demonstrate significant concentrations of cytokine staining techniques with flow
suppressive cytokines such as IL-4, IL-10 and cytometry, alongside sequencing of the
TGF-β, though more recent work with genomes of M. tuberculosis and BCG and
induced sputum has shown high levels of recognition of the immunogenic proteins
both suppressive and proinflammatory expressed by the former but not the latter,
mediators, and the further finding of a have facilitated the development of alter-
significant fall in relative suppressor activity natives to the TST (Andersen et al., 2000).
These assays, which measure IFN-γ An immune assay may be particularly

synthesis following antigen stimulation, have useful to diagnose active tuberculosis when a
been dubbed interferon-gamma release rapid diagnosis cannot be made by simple
assays (IGRAs). The central importance of smear for AFB. In pulmonary cases, sputum
IFN-γ in the lymphocytic immune response smear positivity is associated most commonly
to M. tuberculosis has been outlined above, with the presence of radiologically advanced
but it is pertinent to remember that this disease and cavitation. Here, neutrophils
cytokine is not essential for producing the predominate with the potential for attenuated
delayed-type hypersensitivity response antigen-specific CD4+ lymphocyte responses.
generated by the TST (Cooper et al., 1993). Sputum samples in these cases are likely to be
Initial IGRAs used the same PPD antigen least amenable to lung-based immune diag-
mixture as the TST, but soon changed to more nostic tests. However, where spontaneous
M. tuberculosis-specific antigens such as early sputum is AFB smear negative or sputum
secretory antigen target-6 (ESAT-6) and cannot be expectorated (a fairly common
culture-filtrate protein-10 (CFP-10), which are clinical problem), an accurate rapid immune-
not expressed by BCG. Two commercially assay would be of great value. Although
available assays are now widely available: obtaining a useful lung sample is more
Quantiferon™ assays use antigen stimulation challenging than either performing a TST or
of whole blood with IFN-γ measured by the taking blood for IGRA, this is already the
ELISA method; T-Spot TB™ stimulates a standard of care in many clinical settings and
fixed number of PBMCs and analysis is of thus a lung-orientated immune-assay could
IFN-γ secretion by single cells using the be performed efficiently in parallel to
ELISpot technique. In the investigation of standard microbiology testing such as culture
latent tuberculosis, they have shown superior on a single split sample.
specificity to the TST, especially in BCG- A pilot study of immune-competent
vaccinated individuals, but without great subjects, 23 of whom received a final diag-
improvements in sensitivity (Pai et al., 2008). nosis of active tuberculosis and 13 with a
IGRA performance in investigating variety of pulmonary conditions under
active tuberculosis has been more variable, investigation, analysed IFN-γ and TNF-α
with most studies reporting suboptimal synthesis by CD3+ lymphocytes in whole
sensitivity and specificity. Recent work using blood and BAL after overnight stimulation
T-Spot TB™ showed improved sensitivity with PPD (Barry et al., 2003). Flow cytometry
versus the TST (85% versus 79%), though few with intracellular cytokine staining was used,
HIV-infected subjects were included (Dosanjh and a concentration of PPD-specific produc-
et al., 2008). Given that active tuberculosis tion of both cytokines was observed in the
often is associated with failing immunity, it is lung only in the active tuberculosis cases
no surprise that the sensitivity of IGRA is (predominantly among the CD4+ rather than
reduced in active tuberculosis in a similar CD8+ lymphocyte populations), with a 100-
manner to the TST (Hirsch et al., 1999). This is fold difference in the frequency of responding
even more the case if there is coexistent HIV cells in BAL compared to blood. Interestingly,
infection characterized by progressive loss of responses did not seem to differ between
the CD4+ lymphocyte population in which cases clinically classified as pulmonary and
reside the antigen-specific effector cells that either disseminated or non-pulmonary. This
synthesize IFN-γ. Intuitively, therefore, sampl- may reflect the presence of subclinical
ing the lung in cases of active tuberculosis pulmonary disease or that, regardless of
should provide sensitivity advantages over organ involvement, all active tuberculosis
the use of blood, as the compartmentalized produces a lymphocyte-homing response to
CD4+ lymphocyte-predominant response in the lung, as the original site of infection. HIV
the lung will yield many more antigen- infection has been shown to impair Th1-
specific cells for stimulation, as has been seen cytokine responses significantly in BAL
in household contacts of infectious cases lymphocytes to both tuberculosis and other
(Schwander et al., 2000). lung pathogen-related antigens and so,
importantly, these findings have been (Jafari et al., 2009). A final diagnosis of active
extended to a cohort of 47 HIV-infected tuberculosis was made in 20% of cases (56%
individuals with a median peripheral blood culture confirmed), while in the ‘not tubercu-
CD4 count of 131 cells/μl (Breen et al., 2006; losis’ group, 46% were thought to be latently
Kalsdorf et al., 2009; Jambo et al., 2011). No infected (positive T-Spot result from
decline was observed in the proportion of peripheral blood). Reported results for BAL
IFN-γ synthetic lymphocytes among the were a sensitivity of 91% and specificity of
CD4+ population in BAL as peripheral blood 80%, while in PBMCs the sensitivity was
CD4 counts fell – suggesting an advantage to nearly identical, though specificity was below
the analysis of proportions rather than 50%. The superiority of BAL extended to
absolute responder numbers in this group. comparisons with the TST and NAAT in the
The utility of this BAL-based method complete cohort and, notably, BAL response
was demonstrated in a cohort of 250 was far less likely to be positive in subjects
spontaneous sputum smear-negative or non- with previous but not current tuberculosis
producing individuals, of whom 25% were than was seen with PBMCs. Similar large
HIV infected (median peripheral blood CD4 studies in high-burden settings have yet to be
count 153 cells/μl) (Breen et al., 2008). Again, performed. Evaluation of 85 tuberculosis
PPD was the stimulating antigen, on the basis suspects in South Africa using T-Spot TB™
of the theoretical assumptions that the with BAL (where 29% were HIV coinfected)
broader range of antigen responses elicited demonstrated sensitivity and specificity of
would be advantageous in conditions of 89% and 94% when including those with
immune suppression compared to more definite diagnoses but not the 16% of subjects
specific assays and that confounding by skin- with diagnostic uncertainty or the 34% with
based BCG vaccination would not be a inconclusive assays (Dheda et al., 2009).
problem in the lung, unlike peripheral blood. Quantiferon-TB Gold In-Tube™ was much
Using a diagnostic cut-off determined by less sensitive than T-Spot TB™ and, perhaps
receiver operating characteristic (ROC) curve surprisingly, so was a PPD-based assay. The
analysis, the assay provided 95% sensitivity superior specificity of BAL over PBMCs was
against a final clinical diagnosis of active in agreement with the European data. Repli-
tuberculosis (44% of the cohort; 75% culture cation of these findings in a larger cohort and
confirmed). Sensitivity was not altered by a more technically consistent study would be
HIV infection or site of tuberculosis, and was welcome.
significantly superior to a commercial nucleic Where BAL is performed as part of the
acid amplification assay (NAAT). In a routine work-up of spontaneous sputum
subgroup with active tuberculosis in whom smear-negative or non-producing cases of
responses to ESAT-6 were also measured, no suspected tuberculosis, its use as a one-off
response at all could be detected in 19% of source of lung-derived cells for immune
cases. What was less impressive in this single- diagnosis of tuberculosis makes sense. What
centre study in a middle tuberculosis is needed now is to demonstrate how the
incidence setting (40/100,000 cases/annum) assays may be used as part of a rapid, efficient
was a specificity of 76%, which was thought and cost-effective algorithm that also
to reflect the presence of latent or healed incorporates AFB smear and NAAT prior to
disease rather than confounding by BCG or IGRA. However, even in resource-rich
non-tuberculous mycobacteria. Comparison settings many centres routinely use sputum
of responses in BAL with those in blood was induction rather than BAL and any attempt at
not performed. repeat sampling might be best served by a
A multi-centre study from European sputum-based method. A study of 42 subjects
centres in low-incidence settings analysed (38% HIV infected) of whom 27 were
IFN-γ responses to ESAT-6 and CFP-10 using diagnosed with active tuberculosis, using
the T-Spot TB™ assay in 347 spontaneous flow cytometry and intracellular cytokine
sputum smear-negative or non-producing staining with PPD as the stimulating antigen,
individuals with suspected tuberculosis reported a fourfold reduction in lymphocyte
percentage compared to previous BAL data smaller populations of cells. This will require
(although the predominant effector memory increased use of both multi-colour flow
phenotype of CD45RO+CCR7–CD27– was cytometry and will re-enforce the advantages
the same) and a fivefold fall in CD4+IFN-γ provided by the higher-frequency antigen-
synthetic proportion (Breen et al., 2007). specific populations present in the lung in
However, post hoc ROC analysis against final tuberculosis compared to blood.
diagnosis of tuberculosis demonstrated 89%
sensitivity and 80% specificity in this small
group. Positive results using an ELISpot References
assay with the sputum-derived cells were
also reported. These data have yet to be Agerton, T., Valway, S., Gore, B., Pozsik, C.,
replicated and more work in this area is Olikaytis, B., Woodley, C., et al. (1997)
awaited with interest. Transmission of a highly drug-resistant strain
(W1) of Mycobacterium tuberculosis. Community
outbreak and nosocomial transmission via a
contaminated bronchoscope. Journal of the
5.6 Conclusion American Medical Association 278, 1073–1077.
Ainslie, G.M., Solomon, J.A. and Bateman, E.D.
There are compelling reasons to support the (1992) Lymphocyte and lymphocyte subset
use of lung- rather than blood-derived numbers in blood and in bronchoalveolar lavage
samples when investigating the immune and pleural fluid in various forms of human
response in tuberculosis. Whether these pulmonary tuberculosis at presentation and
advantages overcome the increased expense during recovery. Thorax 47, 513–518.
and complexity compared to blood testing for Akira, A., Takeda, K. and Kaisho, T. (2001) Toll-like
both routine diagnostic work and also large- receptors: critical proteins linking innate and
scale cross-sectional or longitudinal studies acquired immunity. Nature Immunology 8, 675–
680.
remains to be seen. It is clear that although
Almeida, A.S., Lago, P.M., Boechat, N., Huard,
comparative IGRA studies have shown R.C., Lazzarini, L.C., Santos, A.R., et al. (2009)
superiority of BAL versus blood, diagnostic Tuberculosis is associated with a down-
performance as an aid to clinical decision modulatory lung immune response that impairs
making is still inadequate. Specific questions Th1-type immunity. Journal of Immunology 183,
that must be addressed include how to 718–731.
distinguish between different stages of Andersen, P., Munk, M.E., Pollock, J.M. and
tuberculosis (latent infection, active disease Doherty, T.M. (2000) Specific immune-based
and previous treatment) and which markers diagnosis of tuberculosis. The Lancet 356,
robustly demonstrate an effective (and 1099–1104.
Ashton, M.R. (1992) ‘Blind’ bronchoalveolar lavage.
protective) immune response. More accurate
The Lancet 340, 1104.
diagnostic assays may use different antigens, Barnes, P.F., Lu, S., Abrams, J.S., Wang, E.,
though what will be of greater value is the Yamamura, M. and Modlin, R. (1993) Cytokine
ability to identify antigen-specific populations production at the site of disease in human
of cells which exist at lower frequency than tuberculosis. Infection and Immunity 61, 3482–
those detected through synthesis of IFN-γ. 3489.
Markers of full-effector T-cell differentiation Barry, S.M., Lipman, M.C., Bannister, B., Johnson,
such as CD27 and CD154 have shown M.A. and Janossy, G. (2003) Purified protein
promise using blood in discriminating latent derivative-activated type 1 cytokine-producing
from active tuberculosis (Streitz et al., 2007, CD4+ T lymphocytes in the lung: a characteristic
feature of active pulmonary and non-pulmonary
2011). Similarly, analysis of cell populations
TB. Journal of Infectious Diseases 187, 243–
according to their ability to produce IFN-γ, 250.
IL-2 and/or TNF-γ alone or together may Barry, S., Breen, R., Lipman, M., Johnson, M. and
serve as biomarkers of disease states and Janossy, G. (2009) Impaired antigen-specific
treatment response (Millington et al., 2007; CD4(+) T lymphocyte responses in cavitary
Caccamo et al., 2010). If this is the case, then tuberculosis. Tuberculosis (Edinburgh) 89,
future studies will need to analyse ever- 48–53.
Breen, R.A., Janossy, G., Barry, S.M., Cropley, I., Dunleavy, A., Breen, R.A., Perrin, F. and Lipman,
Johnson, M.A. and Lipman, M.C. (2006) M.C. (2008) Is bronchodilation required routinely
Detection of mycobacterial antigen responses before diagnostic sputum induction? Evidence
in lung but not blood in HIV/tuberculosis from studies with tuberculosis. Thorax 63, 473–
co-infected subjects. AIDS 20, 1330–1332. 474.
Breen, R.A., Hardy, G.A., Perrin, F.M., Lear, S., Efthimiadis, A., Pizzichini, M.M., Pizzichini, E.,
Kinloch, S., Smith, C.J., et al. (2007) Rapid Dolovich, J. and Hargreave, F.E. (1997) Induced
diagnosis of smear-negative tuberculosis using sputum cell and fluid-phase indices of
immunology and microbiology with induced inflammation: comparison of treatment with
sputum in HIV-infected and uninfected dithiothreitol vs phosphate-buffered saline.
individuals. PLoS One 2, e1335. European Respiratory Journal 10, 1336–1340.
Breen, R.A., Barry, S.M., Smith, C.J., Shorten, R.J., Elkington, P., Shiomi, T., Breen, R., Nuttall, R.K.,
Dilworth, J.P., Cropley, I., et al. (2008) Clinical Ugarte-Gil, C.A., Walker, N.F., et al. (2011)
application of a rapid lung-orientated immuno- MMP-1 drives immunopathology in human
assay in individuals with possible tuberculosis. tuberculosis and transgenic mice. Journal of
Thorax 63, 67–71. Clinical Investigation 121, 1827–1833.
Caccamo, N., Guggino, G., Joosten, S.A., Fennelly, K.P., Martyny, J.W., Fulton, K.E., Orme,
Gelsomino, G., Di Carlo, P., Titone, L., et al. I.M., Cave, D.M. and Heifets, L.B. (2004) Cough-
(2010) Multifunctional CD4(+) T cells correlate generated aerosols of Mycobacterium tubercu-
with active Mycobacterium tuberculosis infec- losis: a new method to study infectiousness.
tion. European Journal of Immunology 40, American Journal of Respiratory and Critical
2211–2220. Care Medicine 169, 604–609.
Carranza, C., Juarez, E., Torres, M., Ellner, J.J.,
Fireman, E., Topilsky, I., Greif, J., Lerman, Y.,
Sada, E. and Schwander, S.K. (2005) Myco-
Schwarz, Y., Man, A., et al. (1999) Induced
bacterium tuberculosis growth control by lung
sputum compared to bronchoalveolar lavage for
macrophages and CD8 cells from patient
evaluating patients with sarcoidosis and non-
contacts. American Journal of Respiratory and
granulomatous interstitial lung disease.
Critical Care Medicine 173, 238–245.
Respiratory Medicine 93, 827–834.
Condos, R., Rom, W.N., Liu, Y.M. and Schluger,
Flynn, J.L. (2006) Lessons from experimental
N.W. (1998) Local immune responses correlate
Mycobacterium tuberculosis infections.
with presentation and outcome in tuberculosis.
Microbes and Infection 8, 1179–1188.
American Journal of Respiratory and Critical
Flynn, J.L., Chan, J., Triebold, K.J., Dalton, D.K.,
Care Medicine 157, 729–735.
Cooper, A.M., Dalton, D.K., Stewart, T.A., Griffin, Stewart, T.A. and Bloom, B.R. (1993) An
J.P., Russell, D.G. and Orme, I.M. (1993) essential role for interferon gamma in resistance
Disseminated tuberculosis in interferon gamma to Mycobacterium tuberculosis infection. Journal
gene-disrupted mice. Journal of Experimental of Experimental Medicine 178, 2249–2254.
Medicine 178, 2243–2247. Flynn, J.L., Chan, J. and Lin, P.L. (2011)
Dannenberg, A.M. Jr (2001) Pathogenesis of Macrophages and control of granulomatous
pulmonary Mycobacterium bovis infection: basic inflammation in tuberculosis. Mucosal
principles established by the rabbit model. Immunology 4, 271–278.
Tuberculosis (Edinburgh) 81, 87–96. Fremond, C.M., Yeremeev, V., Nicolle, D.M., Jacobs,
Deretic, V. (2011) Autophagy in immunity and cell- M., Quesniaux, V.F. and Ryffel, B. (2004) Fatal
autonomous defense against intracellular Mycobacterium tuberculosis infection despite
microbes. Immunological Reviews 240, 92–104. adaptive immune response in the absence of
Dheda, K., van Zyl-Smit, R.N., Meldau, R., Meldau, MyD88. Journal of Clinical Investigation 114,
S., Symons, G., Khalfey, H., et al. (2009) 1790–1799.
Quantitative lung T cell responses aid the rapid Fujita, A., Murata, K. and Takamori, M. (2009) Novel
diagnosis of pulmonary tuberculosis. Thorax 64, method for sputum induction using the Lung
847–853. Flute in patients with suspected pulmonary
Diedrich, C.R. and Flynn, J.L. (2011) HIV-1/ tuberculosis. Respirology 14, 899–902.
Mycobacterium tuberculosis coinfection immuno- Green, A.M., Mattila, J.T., Bigbee, C.L., Bongers,
logy: how does HIV-1 exacerbate tuberculosis? K.S., Lin, P.L. and Flynn, J.L. (2010) CD4+
Infection and Immunity 79, 1407–1417. regulatory T cells in a cynomolgus macaque
Dosanjh, D.P., Hinks, T.S., Innes, J.A., Deeks, J.J., model of Mycobacterium tuberculosis infection.
Pasvol, G., Hackforth, S., et al. (2008) Improved Journal of Infectious Diseases 202, 533–541.
diagnostic evaluation of suspected tuberculosis. Guyot-Revol, V., Innes, J.A., Hackforth, S., Hinks, T.
Annals of Internal Medicine 148, 325–336. and Lalvani, A. (2006) Regulatory T cells are
expanded in blood and disease sites in patients during a long airplane flight. New England
with tuberculosis. American Journal of Journal of Medicine 334, 933–938.
Respiratory and Critical Care Medicine 173, Kisich, K.O., Higgins, M., Diamond, G. and Heifets,
803–810. L. (2002) Tumour necrosis factor alpha
Hirsch, C.S., Toosi, Z., Othieno, C., Johnson, J.L., stimulates killing of Mycobacterium tuberculosis
Schwander, S.K., Robertson, S., et al. (1999) by human neutrophils. Infection and Immunity
Depressed T-cell interferon- responses in 70, 4591–4599.
pulmonary tuberculosis: analysis of underlying Korbel, D.S., Schneider, B.E. and Schaible, U.E.
mechanisms and modulation with therapy. (2008) Innate immunity in tuberculosis. Myths
Journal of Infectious Diseases 180, 2069–2073. and truths. Microbes and Infections 10, 995–
Jafari, C., Thijsen, S., Sotgiu, G., Goletti, D., 104.
Domínguez Benítez, J.A., Losi, M., et al. (2009) Law, K.F., Jagirdar, J., Weiden, M.D., Bodkin, M.
Tuberculosis Network European Trials group. and Rom, W.N. (1996) Tuberculosis in HIV-
Bronchoalveolar lavage enzyme-linked immuno- positive patients: cellular response and immune
spot for a rapid diagnosis of tuberculosis: a activation in the lung. American Journal of
Tuberculosis Network European Trials group Respiratory and Critical Care Medicine 153,
study. American Journal of Respiratory and 1377–1384.
Critical Care Medicine 180, 666–673. Lesley, R. and Ramakrishnan, L. (2008) Insights
Jambo, K.C., Sepako, E., Fullerton, D.G., Mzinza, into early mycobacterial pathogenesis from the
D., Glennie, S., Wright, A.K., et al. (2011) zebrafish. Current Opinion in Microbiology 11,
Bronchoalveolar CD4+ T cell responses to 277–283.
respiratory antigens are impaired in HIV- Loppow, D., Böttcher, M., Gercken, G., Magnussen,
infected adults. Thorax 66, 375–382. H. and Jörres, R.A. (2000) Flow cytometric
Jones, B.E., Young, S.M., Antoniskis, D., Davidson, analysis of the effect of dithiothreitol on
P.T., Kramer, F. and Barnes, P.F. (1993) Relation- leukocyte surface markers. European Respira-
ship of the manifestations of tuberculosis to tory Journal 16, 324–329.
CD4 cell counts in patients with human Loudon, R.G., Bumgarner, L.R., Lacy, J. and
immunodeficiency virus infection. American Coffman, G.K. (1969) Aerial transmission of
Review of Respiratory Disease 148, 1292– mycobacteria. American Review of Respiratory
1297. Disease 100, 165–171.
Juarez, E., Nunez, C., Sada, E., Ellner, J.J., McWilliams, T., Wells, A.U., Harrison, A.C.,
Schwander, S.K. and Torres, M. (2010) Lindstrom, S., Cameron, R.J. and Foskin, E.
Differential expression of Toll-like receptors on (2002) Induced sputum and bronchoscopy in
human alveolar macrophages and autologous the diagnosis of pulmonary tuberculosis. Thorax
peripheral monocytes. Respiratory Research 57, 1010–1014.
11, 2. Martineau, A.R., Newton, S.M., Wilkinson, K.A.,
Kalsdorf, B., Scriba, T.J., Wood, K., Day, C.L., Kampmann, B., Hall, B.M., Nawroly, N., et al.
Dheda, K., Dawson, R., et al. (2009) HIV-1 (2007) Neutrophil-mediated innate immune
infection impairs the bronchoalveolar T-cell resistance to mycobacteria. Journal of Clinical
response to mycobacteria. American Journal of Investigation 117, 1988–1994.
Respiratory and Critical Care Medicine 180, Millington, K.A., Innes, J.A., Hackforth, S., Hinks,
1262–1270. T.S., Deeks, J.J., Dosanjh, D.P., et al. (2007)
Kaplan, G., Post, F.A., Moreira, A.L., Wainwright, Dynamic relationship between IFN-gamma and
H., Kreiswirth, B.N., Tanverdi, M., et al. (2003) IL-2 profile of Mycobacterium tuberculosis-
Mycobacterium tuberculosis growth at the specific T cells and antigen load. Journal of
cavity surface: a microenviroment with failed Immunology 178, 5217–5226.
immunity. Infection and Immunity 71, 7099– Milman, N., Faurschou, P., Munch, E.P. and Grode,
7108. G. (1994) Transbronchial lung biopsy through
Keane, J., Gershon, S., Wise, R.P., Mirabile- the fibre optic bronchoscope. Results and
Levens, E., Kasznica, J., Schwieterman, W.D., complications in 452 examinations. Respiratory
et al. (2001) Tuberculosis associated with Medicine 88, 749–753.
infliximab, a tumor necrosis factor alpha- Moodley, Y.P., Dorasamy, T., Venketasamy, S.,
neutralizing agent. New England Journal of Naicker, V. and Lalloo, U.G. (2000) Correlation
Medicine 345, 1098–1104. of CD4:CD8 ratio and tumour necrosis factor
Kenyon, T.A., Valway, S.E., Ihle, W.W., Onorato, I.M. (TNF) alpha levels in induced sputum with
and Castro, K.G. (1996) Transmission of bronchoalveolar lavage fluid in pulmonary
multidrug-resistant Mycobacterium tuberculosis sarcoidosis. Thorax 55, 696–699.
Morrison, J., Pai, M. and Hopewell, P.C. (2008) Clinical and Developmental Immunology 2011,
Tuberculosis and latent tuberculosis infection in 347594.
close contacts of people with pulmonary Schneider, B.E., Korbel, D., Hagens, K., Koch, M.,
tuberculosis in low-income and middle-income Raupach, B., Enders, J., et al. (2010) A role for
countries: a systemic review and meta-analysis. IL-18 in protective immunity against Myco-
The Lancet Infectious Diseases 8, 359–368. bacterium tuberculosis. European Journal of
Murray, J.F. (2003) Bill Dock and the location of Immunology 40, 396–405.
pulmonary tuberculosis. How bed rest might Schwander, S. and Dheda, K. (2011) Human lung
have helped consumption. American Journal of immunity against Mycobacterium tuberculosis.
Respiratory and Critical Care Medicine 168, Insights into pathogenesis and protection.
1029–1033. American Journal of Respiratory and Critical
Pai, M., Zwerling, A. and Menzies, D. (2008) Care Medicine 183, 696–707.
Systematic review: T-cell-based assays for the Schwander, S.K., Torres, M., Sada, E., Carranza,
diagnosis of latent tuberculosis infection: an C., Ramos, E., Tary-Lehmann, M., et al. (1998)
update. Annals of Internal Medicine 149, 177– Enhanced responses to Mycobacterium tubercu-
184. losis antigens by human alveolar lymphocytes
Pin, I., Gibson, P.G., Kolendowicz, R., Girgis- during active pulmonary tuberculosis. Journal of
Gabardo, A., Denburg, J.A., Hargreave, F.E., et Infectious Diseases 178, 1434–1445.
al. (1992) Use of induced sputum cell counts to Schwander, S.K., Torres, M., Carranza, C.C.,
investigate airway inflammation in asthma. Escobedo, D., Tary-Lehmann, M., Anderson, P.,
Thorax 47, 25–29. et al. (2000) Pulmonary mononuclear cell
Pizzichini, E., Pizzichini, M.M., Kidney, J.C., responses to antigens of Mycobacterium
Efthimiadis, A., Hussack, P., Popov, T., et al. tuberculosis in healthy household contacts of
(1998) Induced sputum, bronchoalveolar lavage
patients with active tuberculosis and healthy
and blood from mild asthmatics: inflammatory
controls from the community. Journal of
cells, lymphocyte subsets and soluble markers
Immunology 165, 1479–1485.
compared. European Respiratory Journal 11,
Silver, R.F., Zukowski, L., Kotake, S., Li, Q., Pozuelo,
828–834.
F., Krywiak, A., et al. (2003) Recruitment of
Reed, M.B., Domenech, P., Manca, C., Su, H.,
antigen-specific Th1-like responses to the
Barczak, A.K., Kreiswirth, B.N., et al. (2004) A
human lung following bronchoscopic segmental
glycolipid of hypervirulent tuberculosis strains
challenge with purified protein derivative of
that inhibits the innate immune response.
Mycobacterium tuberculosis. American Journal
Nature 431, 84–87.
of Respiratory, Cell and Molecular Biology 29,
Rees, P.J., Hay, J.G. and Webb, J.R. (1983)
117–123.
Premedication for fibreoptic bronchscopy.
Thorax 38, 624–627. Streitz, M., Tesfa, L., Yildirim, V., Yahyazadeh, A.,
Robinson, D.S., Ying, S., Taylor, I.K., Wangoo, A., Ulrichs, T., Lenkei, R., et al. (2007) Loss of
Mitchell, D.M., Kay, A.B., et al. (1994) Evidence receptor on tuberculin-reactive T-cells marks
for a Th1-like bronchoalveolar T-cell subset and active pulmonary tuberculosis. PLoS One 2,
predominance of interferon-gamma gene e735.
activation in pulmonary tuberculosis. American Streitz, M., Fuhrmann, S., Powell, F., Quassem, A.,
Journal of Respiratory and Critical Care Nomura, L., Maecker, H., et al. (2011)
Medicine 149, 989–993. Tuberculin-specific T cells are reduced in active
Rosenthal, I.M., Zhang, M., Williams, K.N., pulmonary tuberculosis compared to LTBI or
Peloquin, C.A., Tyagi, S., Vernon, A.A., et al. status post BCG vaccination. Journal of
(2007) Daily dosing of rifapentine cures Infectious Diseases 203, 378–382.
tuberculosis in three months or less in the Tsiligianni, J., Tzanakis, N., Kyriakou, D., Chryso-
murine model. PLoS Medicine 4, e344. fakis, G., Siafakas, N. and Bouros, D. (2002)
Sadek, M.I., Sada, E., Toosi, Z., Schwander, S.K. Comparison of sputum induction with broncho-
and Rich, E.A. (1998) Chemokines induced by alveolar lavage cell differential counts in patients
infection of mononuclear phagocytes with with sarcoidosis. Sarcoidosis, Vasculitis and
mycobacteria and present in lung alveoli during Diffuse Lung Diseases 19, 205–210.
active pulmonary tuberculosis. American Volkman, H.E., Clay, H., Beery, D., Chang, J.C.,
Journal of Respiratory, Cell and Molecular Sherman, D.R. and Ramakrishnan, L. (2004)
Biology 19, 513–521. Tuberculous granuloma formation is enhanced
Saiga, H., Shimada, Y. and Takeda, K. (2011) Innate by a mycobacterium virulence determinant.
immune effectors in mycobacterial infection. PLoS Biology 2, 1946–1956.
von Pirquet, C. (1907) Frequency of tuberculosis in (2005) Resident Th1-like effector memory cells in
childhood. Journal of the American Medical pulmonary recall responses to Mycobacterium
Association 52, 675–678. tuberculosis. American Journal of Respiratory,
Walrath, J.R. and Silver, R.F. (2011) The -41 Cell and Molecular Biology 33, 48–55.
integrin in localization of Mycobacterium Yu, C.-T., Wang, C.-H., Huang, T.-J., Lin, H.-C. and
tuberculosis-specific Th1 cells to the human Kuo, H.-P. (1995) Relation of bronchoalveolar
lung. American Journal of Respiratory, Cell and lavage T lymphocyte subpopulations to rate of
Molecular Biology 45, 24–30. regression of active pulmonary tuberculosis.
Walrath, J., Zukowski, L., Krywiak, A. and Silver, R.F. Thorax 50, 869–874.
6 Sniffing Out Tuberculosis
Ruth McNerney1 and Claire Turner2

1Department of Pathogen Molecular Biology, London School of Hygiene and
Tropical Medicine, London, UK; 2Department of Life, Health and Chemical Sciences,
The Open University, Milton Keynes, UK
6.1 Introduction Volatile compounds are those that persist

in a gaseous state. Compounds that exist
Tuberculosis, or phthisis as it was previously exclusively as vapours at ambient tem-
known, is an ancient ailment that has peratures and pressures are called gases.
perplexed health practitioners for millennia. Compounds that are predominantly liquids
One of the greatest challenges is diagnosing at ambient temperatures but that partially
the disease. Unlike other infectious diseases evaporate to the gaseous state are described
such as malaria or HIV, the pathogen is rarely as volatile. Similarly, some solid substances
found in circulating body fluids and attempts have a capacity to sublime and enter the
to test for circulating antibodies have also gaseous phase. Vapours so formed can
proved unsatisfactory. Mycobacterium tubercu- condense back to their liquid state or deposit
losis bacilli may be present in expectorated as the solid state. The propensity of a
sputum in pulmonary forms of the disease substance to convert to its gaseous form is
but detecting the bacteria is not easy and indicated by its vapour pressure; being the
specimens need to be processed in a pressure at which the gaseous and liquid or
laboratory. Poor access to diagnostic facilities solid phases are at equilibrium. Compounds
in developing countries and the low sensi- with a high vapour pressure at normal tem-
tivity of current tests combine to frustrate peratures are more volatile than those with
efforts to identify patients. Case detection low vapour pressure. In reality, most atmos-
rates are low and millions of tuberculosis pheres are mixtures of volatile compounds
cases each year are not diagnosed and and gases and individual vapours contribute
reported (McNerney and Daley, 2011). The a partial vapour pressure to the mix. In
need for rapid sensitive diagnostic tests that general, the volatility of a compound and its
are affordable in the countries where ability to escape the liquid matrix increases
tuberculosis is endemic has encouraged test with temperature. Eventually, boiling point
developers to consider less traditional diag- will be reached, where the compound’s
nostic strategies. In this chapter, we describe vapour pressure equals ambient atmospheric
the analysis of volatile compounds for disease pressure.
detection and review progress towards a Volatile compounds and gases are pro-
diagnostic test for tuberculosis. duced by living organisms as a by-product of

82 R. McNerney and C.Turner
their metabolic processes, and volatile woman’s breath and sputum was found to
organic compounds (VOCs) are of particular become increasingly obnoxious during the
interest to biologists and medical prac- month preceding her death. Less dramatic
titioners. Biological systems are water based odours have also been described, sometimes
and the solubility of a compound is another as a ‘sickly smell’ (Shattuck, 1880; Louis,
factor affecting its ability to vaporize. 1844), and Burney Yeo, a physician in 19th
Humans and other animals have evolved an century London, reported to the British
ability to recognize some compounds or mix- Medical Association ‘a peculiar nauseating
tures of compounds via their olfactory odour in the breath of many phthisical
systems (Firestein, 2001; Rouquier and patients, even before the development of
Giorgi, 2007). Olfaction, or the ‘sense of marked physical signs’ (Yeo, 1877). An
smell’, is a vital attribute that informs basic alternative approach concerned urine from
behaviours such as food selection, territorial patients with phthisis which, on storage,
activity and mating. The olfactory systems in developed a smell reminiscent of rotten
individual animal species have developed to cheese or partially treated sewage (White,
smell selectively low concentrations of 1892). Though often reported, there are no
compounds which are in their interest to published studies on the diagnostic value of
detect. This includes pheromones and semio- the human olfactory system in the manage-
chemicals which are present in very low ment of tuberculosis and evidence of its
concentrations for signalling between animals efficacy remains anecdotal. The majority of
of the same species. It also includes rancid reports date from an era prior to chemo-
odours, which humans and other animals use therapy, when tuberculosis was a common
to warn that food is not safe to eat, and faecal affliction and death rates were extremely
odours such as indole. Smells are also utilized high. Caseous necrosis of granulomatous and
in clinical situations, albeit informally at lung tissue is characteristic of tuberculosis
times (Palmer, 1993; Bomback, 2006). disease and it would not be surprising if fetid
odours were observed in the breath and
sputum of patients with late-stage terminal
6.1.1 Historical perspectives disease. Of more interest to the diagnostician
are the reports of patients on initial
The earliest description of an investigative presentation and those with less advanced
test for tuberculosis comes from ancient disease. It should be noted, however, that
Greece and the writings of Hippocrates, who other conditions may also result in foul or
recorded in his Aphorisms, ‘In persons affected sickly odours, particularly those bacterial or
with phthisis, if the sputa which they cough fungal infections that result in necrosis or pus
up have a heavy smell when poured upon formation. Thus, both the sensitivity and
coals, and if the hairs of the head fall off, the specificity of odour as a diagnostic tool for
case will prove fatal’ (Hippocrates and tuberculosis require further investigation.
Adams, 400 bc). A similar approach was
recommended by the Roman, Caelius
Aurelius (130 bc), who stated that a char- 6.2 Olfactory Detection
acteristic foul odour from phlegm over hot
coals was indicative of physical decomposition Human olfactory detection of tuberculosis is
(Drabkin, 1950). In more recent times, the beset by two problems. The first is the
sputa and breath of patients with phthisis has intrinsic risk of contracting the disease, which
been reported to have a gangrenous odour spreads through the inhalation of tiny
(Bowditch, 1889). Fetid smells similar to airborne droplet nuclei. The second is the
rotten eggs have been reported that, on relatively poor performance of the human
autopsy, were linked to the breakdown of olfactory system when compared to other
tuberculous tissue, perforation of the lungs creatures. Animals including dogs, pigs and
and the presence of purulent liquid in the rodents are believed to have olfactory senses
thoracic cavity (Louis, 1844). In one case, a superior to humans (Rouquier et al., 2000). In
Sniffing Out Tuberculosis 83
addition, they can be trained using Pavlovian Groundbreaking work on olfactory

conditioning to respond physically to specific detection of tuberculosis has been performed
odours or smells, as is demonstrated by the in Tanzania, where giant African pouched
use of dogs to detect illicit foods, narcotics rats (Cricetomys gambianus) have been trained
and explosives (Johnston, 1999). The mam- to recognize samples of sputum from
malian olfactory system is illustrated in Fig. tuberculosis patients (Weetjens et al., 2009).
6.1. Vapours enter the nasal passages, where The animals are trained to sniff individual
they interact with the olfactory epithelium. pots containing the samples and to pause for
Molecules bind to receptors on highly special- at least 5 s when they recognize a positive
ized olfactory cells, triggering an electrical signal. Conditioning was undertaken using
signal which transmits along neurons to the food rewards (banana). Though not as
olfactory bulb, which is part of the central sensitive as other diagnostic techniques, a rat
nervous system. Different receptors react can test samples at a faster rate than a
with different classes of molecules, reporting laboratory technician using conventional
to different parts of the olfactory bulb and Ziehl–Neelsen (ZN) light microscopy. The
forming a highly sophisticated neural net- yield of positives can be increased by
work which, in conjunction with the brain, is exposing samples to multiple rats. In a recent
capable of analysing complex mixtures of study, 23,101 sputum samples from 10,523
vapours. Pattern recognition combined with patients collected during a routine diagnostic
memory allows vapours to be differentiated service were tested by the rats and results
and smells to be recognized. compared to the smear microscopy result.
Fig. 6.1. Olfactory and e-nose sensing.

Samples were first frozen and then sterilized honeybee (Apis mellifera) can be conditioned
by autoclaving before being presented to ten to recognize specific odours by Pavlovian
different rats (Poling et al., 2010). conditioning, where conditioned bees are
Any sample reported as negative by stimulated to extend their proboscis when
microscopy but positive by two or more rats they encounter the olfactory signal (Bitter-
was re-evaluated by smear microscopy in the mann et al., 1983). Multiple receptors on the
laboratory performing the evaluation. Col- insect’s antennae combine with advanced
lectively, the rats identified 2274 (91%) of 2487 signal processing to provide highly sensitive
samples found positive by routine ZN. A identification of complex mixtures (Sandoz,
further 3012 samples were found positive by 2003). Recognition of organic compounds
the rats, of which 927 were found positive by occurs at a submolecular level and is influ-
smear microscopy when re-examined in the enced by carbon chain length and functional
research laboratory. Thus, the ten rats gave an group (Guerrieri et al., 2005). Preliminary
improved yield over a single routine smear work has suggested that honeybees might be
result. A high proportion of the positive trained to differentiate M. bovis BCG bacilli
results (40%) did not correlate with a positive from environmental strains such as M.
microscopy result, suggesting that specificity smegmatis, but further studies have yet to be
may not be sufficient and that confirmatory undertaken (McNerney, unpublished data).
testing might be required. Unfortunately, Although large numbers of insects may be
further conclusions cannot be drawn as cul- maintained at low cost, and training is more
ture was not undertaken, preventing esti- rapid than that for a dog or rat, they share
mation of the actual sensitivity and specificity some logistic challenges with the mammals.
of the test. Despite the shortcomings of the Maintenance of the hive and conditioning or
study, the data presented indicate that training the bees requires specialist know-
differentiation of tuberculosis and non- ledge. In addition, insects are subject to
tuberculosis patients may be achieved by environmental influences such as heat and
assessing volatile compounds emitted by humidity that affect their behaviour. Although
sputum samples. The authors did not specu- an intriguing prospect, the application of
late on the constituents of the olfactory signal insect-based diagnostic sensors for tubercu-
recognized by the rats. Previous studies losis has yet to be demonstrated.
demonstrated that the animals could recog-
nize positive sputum following training with
cultured bacteria, suggesting that volatiles 6.3 E-noses
originating from the bacteria themselves may
play a role (Weetjens et al., 2009). Although The use of volatile biomarkers to diagnose
these findings are highly interesting, it is patients will require robust and reliable
unlikely that rats could provide rapid diag- means of measuring them. Two methodo-
nosis at the point of care. The infrastructure logical strategies have been adopted. The first
to house and train them, the practical chal- is to mimic olfaction, where the vapour under
lenges of transporting live animals and the test is exposed to sensors that, on encountering
need for confirmatory testing suggest they target molecules, change their physical or
would be more suited to referral laboratories. electrical state, inducing a signal that can be
In addition, they and their handlers may measured. An instrument which does this is
suffer from fatigue and boredom, and careful called an electronic nose, or e-nose. The
management will be required to maintain second method is to determine the chemical
high productivity and performance. composition of an odour and identify and
Some of the limitations of working with quantify the compounds present in the
mammals may be overcome by using insects. mixture.
Odour recognition is integral to an insect’s E-noses are devices containing an array
life and they are constantly adapting, modify- of different sensors which each respond to a
ing and learning associations between odours different degree to the chemicals in an odour.
and their environment. Insects such as the An e-nose will not identify the components of
an odour but will compare odours and data requires significant computer processing.
classify them, for example by positive or Similar or dissimilar responses can be
negative disease state. Some sensors may clustered and, if associated with known
respond preferentially to water, for example, compounds, a library of chemical smells may
whereas others will respond to different be established. For complex odours, pattern
chemical functional groups, recognizing recognition software may be employed,
other types of compounds such as alcohols, where a comparison of positive and negative
esters, aldehydes, ketones, amines, sulfides samples allows mapping of sensor responses.
or carboxylic acids. If an odour is a complex In this way, the e-nose mimics olfactory
mixture, containing many different com- sensing, where animals or insects are trained
pounds, then each sensor will have a different to recognize and differentiate vapours.
response, producing an output unique to that E-noses can be trained in the same way as
odour. The principle of an e-nose device is people to recognize and distinguish smells
shown in Fig. 6.1. Sensors may be made from without understanding which compounds
a variety of materials, but the most common are responsible for the odour.
are constructed from metal oxides or synthetic The major difference between the e-nose
organic polymers called conducting polymers. and animal olfaction is in the sophistication
Typically, the output of a sensor will be a of the olfactory system. Typically, an e-nose
change in an electrical property. For example, might house 30 sensors, whereas each human
when compounds bind to a metal oxide has approximately 5 million olfactory sensors.
sensor, there will be a proportional change in The neural system and brain also has far
resistance and the absolute magnitude and superior networking and processing capabil-
duration of that change may be measured. ities to discern complex mixtures than the
The data collected are expressed as a curve, current e-nose technologies.
as shown in Fig. 6.2, where features such as In the way that animals smell odours, an
amplitude, area and the rate of adsorption e-nose tests vapours directly without prior
and desorption may be used to interpret manipulation of the sample. The devices are
sensor response. simple to operate and, after doing the initial
With multiple sensors, more complex computation to classify sample types, are
responses are obtained and interpreting the usually able to classify an unknown sample
(%)
Fig. 6.2. Typical output from a metal oxide sensor.

in minutes. It is not a particularly sensitive Variations were seen between two e-nose
technique, so if the odours indicative of instruments, the performance of which varied
disease are subtle, or if the natural variation from day to day. The authors speculated that
between samples is great, then differentiation the accuracy of the test might have been
will be difficult (Nicolas and Romain, 2004; affected by the sampling method used, as
Rains et al., 2004). Another factor to consider headspace vapours were taken directly from
is the stability of the sensors, as both metal sputum containers with a smaller headspace
oxides and conducting polymers are prone to rather than the sealed glass vials used
change their properties over time, causing a previously. They also reported a small but
change or drift in their output (Li and Qian, significant instrument drift. Their conclusion
1993; Romain and Nicolas, 2010). that the commercial e-nose in its current state
E-nose technology has been used to is not specific enough for tuberculosis diag-
investigate samples from animals and humans nosis is supported by other studies (Knobloch
infected with tuberculosis. A commercial et al., 2009, 2010). Variation in performance
electronic nose device (BH-114, Bloodhound across time and the undue influence of
Sensors, UK) with 14 conducting sensors environmental factors such as temperature
based on polyaniline polymers was applied and moisture suggest these instruments lack
to samples from cattle and badgers infected the reliability required of a diagnostic device.
with M. bovis and uninfected control animals In addition, the problem of making instru-
(Fend et al., 2005). Sera samples were stored ments behave reproducibly from day to day
frozen prior to dilution in saline and and the fact that no two instruments behave
incubation at 37°C in sealed vials. Samples of the same, even when manufactured at the
headspace gas were extracted by inserting a same time, suggest the technology is not
needle through a septum in the cap of the robust enough for field use. However, the fact
vials. Data obtained from the e-nose were that they are able to differentiate between
analysed using the multivariate data analysis positive and negative groups at the popu-
techniques principal component analysis and lation level (if not at the individual level)
linear discriminant function (LDF) analysis. indicates that there probably is an odour
The data obtained suggested that the e-nose associated with tuberculosis and that other
could differentiate samples from infected and techniques may be able to identify the
uninfected animals and the authors specu- components of the odour and characterize
lated that the technology had potential as a the molecules associated with tuberculosis.
diagnostic test. Further work was performed
on cultures of bacteria and showed that
mycobacteria could be discriminated from 6.4 VOC Analysis
Pseudomonas aeruginosa and that smaller but
discernable differences were observed Although studies with e-noses suggest that
between M. tuberculosis, M. avium and M. there is an odour associated with tuberculosis,
scrofulaceum (Fend et al., 2006). Differences they are not able to identify the characteristic
were also observed when bacteria were constituents of the odour. To develop a more
added to sputum in spiking experiments. The robust test, it will be necessary to identify the
application of neural networking to clinical components of the odour and to determine
samples (sputum) showed the e-nose could their significance as diagnostic markers for
discriminate samples from patients infected tuberculosis. There are five possible origins of
with tuberculosis and from non-tuberculosis VOCs in samples of body fluids of infected
patients, and that it might be useful as a individuals. The most significant group of
diagnostic tool for human tuberculosis. VOCs are those which are produced by the
Unfortunately, subsequent studies have not metabolic processes of the body. For example,
supported these findings. Sensitivities of 68% acetone is an abundant VOC found in all
and 75% and specificities of less than 70% body fluids including breath, blood and urine
were observed in a study of sputum samples (Ashley et al., 1994; Miekisch et al., 2004). It
performed in Tanzania (Kolk et al., 2010). originates through the regulation of blood
sugar by insulin. If blood sugar is low, fat is of gaseous/volatile mixtures. Chromatographic

broken down and keto-bodies are formed, methods are most frequently used for this,
leading to acetone production. Another where compounds are separated according to
example is ammonia, which is generated as a their physical or chemical characteristics,
result of protein breakdown. These so-called such as size, ionic strength or polarity. The
endogenous VOCs are important indicators sensitivity of gas chromatography is depend-
of the health of the person (Amann et al., ent on the separation capacity of the columns
2004). The second origin of VOCs is from the used and the detection method employed.
action of bacterial metabolism. Mostly, Highly sensitive instruments have been
bacteria colonize the body in a symbiotic developed where, after separation, vapours
relationship, with several trillion living in the are detected using novel sensing devices such
human gut, and these produce detectable as surface acoustic wave detectors (Watson et
VOCs (Amann et al., 2004). However, bacteria al., 2003). Traditional gas chromatographs are
that cause disease will also generate VOCs sophisticated, laboratory-based instruments
(Probert et al., 2009) and, ideally, these VOCs but rapid, portable versions have been
are those that should be targeted as potential developed that can be used in the field (Jia et
biomarkers. The third source of VOCs is from al., 2000; Staples and Viswanathan, 2008). The
the modification of the VOCs produced by retention time, or the time at which the
bacteria within the host organism. One com- compound elutes from the chromatography
pound may be converted chemically or column, may be compared to standard or
enzymatically to another compound, which reference compounds and so be used to infer
may be detectable. Fourth, the response of the the identity of a compound. However, it is
host organism to infection is a potential possible that other compounds will elute at
source of VOC. When an animal is infected the same time and chromatography alone
with a pathogen, the immune system modifies cannot be used to determine the identity of a
the metabolism, producing inflammatory compound. In order to be sure of a com-
markers and resulting in changes which can pound’s identity, mass spectrometry is
release VOC (Barker et al., 2006; Chan et al., required.
2009). The final source of VOCs is from
exogenous compounds, which are those
produced elsewhere and absorbed by the 6.6 Mass Spectrometry
body. Exposure to any compound through
inhalation, ingestion or injection could con- Mass spectrometry may be used to identify
taminate tissues of the body and sub- compounds either as a detection method
sequently be emitted as VOCs (Blount et al., following chromatography or, with some
2006). An example of this is the contamination simple mixtures, certain types of mass
of body tissues and fluids following an spectrometry may be used without pre-
anaesthetic. Traces of the anaesthetic may separation. In mass spectrometry, the VOCs
remain in the body for weeks, and the time or gases present are ionized using any one of
of removal will depend on the dose, its a number of techniques and the ions are then
solubility and whether the body will meta- separated according to the ratio of their mass
bolize the compound (Turner, unpublished to their charge (m/z ratio) by letting the ions
results). Other examples are cigarette enter a magnetic or electric field (depending
smoking and the consumption of ethanol on the type of mass spectrometry). Heavy
(Ashley et al., 1996). ions (high m/z) are deflected less than light
ions and so the ions are separated. The ions at
each m/z value are then detected and a
6.5 Chromatography spectrum is produced as counts per second at
each m/z value. Each compound will
Assessment of the concentration or relative generally have its own unique spectrum and
abundance of the constituent volatile thus mass spectrometry can be used to
compounds in an odour requires separation identify compounds by comparing spectra
obtained with library spectra. There are bacteria in vitro (Phillips et al., 2007, 2010;
several configurations of the technology, with Cunha et al., 2008; Syhre and Chambers,
some instruments tailored to examine small 2008). Attempts to validate these biomarkers
gaseous molecules, while others are used to using clinical specimens have so far met with
investigate larger semi-volatile compounds. limited success. Syhre and colleagues
In the latter case, instruments are usually examined samples of breath for derivatives of
connected to chromatography systems to nicotinic acid (Syhre et al., 2009). Breath was
separate the components. collected by asking volunteers to blow into 1 l
Mass spectrometers are primarily bags and samples were concentrated using
research tools and are used for biomarker solid-phase microextraction (SPME: absorp-
discovery, but portable versions have been tion on to activated fibres). Samples were
developed (Yang et al., 2008). An alternative then treated chemically (derivatized) prior to
detection technology, where gaseous ions are analysis by gas chromatography–mass
separated by shape and charge is differential spectrometry (GC-MS). Significant differences
mobility spectrometry (DMS) (Krylov et al., in the levels of the derivatives were observed
2007). This technology is amenable to in the breath of tuberculosis and non-
miniaturization and portable devices have tuberculosis patients (Syhre et al., 2009).
been developed that are being used as Levels were estimated to be in the femtomol/
detectors for military and industrial appli- mol range, which is below the level at which
cations (Zolotov, 2006). There is also interest they could be measured using current ‘easy
in their use for medical purposes for testing to use’ VOC detection devices. A second and
breath (Molina et al., 2008). However, DMS is perhaps more difficult problem to be resolved
not able to identify a compound with is that nicotinic acid is not specific to
certainty, as it does not produce a unique tuberculosis and is present in some food and
spectrum, but may be of use if the compounds tobacco products. Smokers were excluded
being detected are already known. from their study, but such an approach does
not offer a practical solution for the routine
diagnostic clinic. A different approach was
6.7 Volatile Biomarkers for taken by Phillips, where samples of breath
Tuberculosis were collected on sorbent traps and
transported to the laboratory. Following
Growing bacteria on various media enables thermal desorption, the vapours released
the analysis of the VOCs produced by the were analysed by GC-MS. Multivariate
bacteria as a result of their metabolism. These analysis (fuzzy logic and pattern recognition
VOCs can be collected and analysed and it methods) of data from 134 volatile com-
has been found that different bacteria, pounds suggested that tuberculosis patients
including mycobacteria, produce different could be differentiated from non-tuberculosis
combinations of VOCs (Scholler et al., 1997; patients and healthy volunteers (Phillips et
Pavlou et al., 2004); however, these com- al., 2007). Subsequent work from the same
binations vary in many cases with the type of research group reported a sensitivity of 84%
growth media used. This is perhaps not but a specificity of just 65% when testing
surprising in the case of mycobacteria, as microbiologically proven tuberculosis patients
they have the ability to alter their metabolic (Phillips et al., 2010). In this study, several
process depending on the nutritive sources compounds were found to be associated with
available (Wheeler and Radcliffe, 1994). This tuberculosis. Some were markers of oxidative
implies that when the bacteria then grow in stress, and these included alkanes. These are
vivo, having infected an animal or a human, non-specific markers of illness and may not
they may produce yet a different set of in themselves be able to distinguish between
biomarkers and this may also depend on the tuberculosis and other illnesses or ill health.
physiology of that animal. Mass spectrometry However, the researchers also found other
has led to the identification of a number of markers which might potentially be of use as
potential biomarkers produced by myco- specific biomarkers. These included those
that might be produced directly by M. tubercu- 6.9 Conclusions: The Future of

losis, namely cyclohexane and aromatic Odour Analysis for Tuberculosis
hydrocarbons. However, it should be noted Diagnosis
that cyclohexane is frequently found in
nature, so may not be specific to mycobacteria. Current evidence suggests that odour
analysis has the potential to identify tubercu-
losis disease in humans and animals either
6.8 Multivariate Statistics
from their breath or by testing the headspace
above clinical specimens such as sera or
The ubiquitous nature of small, volatile com-
sputa. If unique biomarkers are present, the
pounds suggests that individual markers are
difficulty in detecting them implies that they
unlikely to offer the specificity required for a
are present at low concentration and dis-
diagnostic test.
tinguishing such endogenous markers from
The e-nose and olfactory data suggest
that samples can be differentiated by the exogenous markers is extremely difficult.
combination of biomarkers, or the pattern of That different research groups report
VOCs, rather than by individual compounds, different markers suggests that samples need
and it may be more appropriate to consider to be examined by a variety of techniques to
the ‘VOC fingerprint’ of an odour rather than ‘rule in’ or ‘rule out’ all potential biomarkers
its constituent molecules. It is conceivable systematically. To validate such markers for
that the markers that are responsible for the diagnostic use, studies with a large number
different patterns are a combination of those of subjects will be required. E-nose
that are produced by the host’s response to technology does not yet provide sufficient
infection, such oxidative stress alkane means of detection and more analytical
markers and VOCs produced routinely by approaches are required. Pattern recognition
the body (e.g. ammonia, acetone, methanol, provides an attractive alternative, but the
etc.) but present in different proportions. approach is dependent on knowledge of the
VOC originating from the M. tuberculosis patterns characteristic to the disease to aid
bacilli growing in the body may also the design of sensors of high sensitivity and
contribute, although current evidence sug- specificity. Potential diagnostic biomarker
gests they may be present at very low and patterns shall need to be verified with a large
sometimes undetectable levels. number of individuals with different
Multivariate statistics and bioinformatics backgrounds, lifestyles, ages, gender, health
provide us with the tools to discern patterns status, environment, diet and nutritional
in large complex data sets. This approach status. An additional problem is the col-
was taken by Spooner and colleagues, who lection of samples. Volatiles are, by their
used selected ion flow tube mass nature, unstable and samples need to be
spectrometry (SIFT-MS) to analyse the collected and stored in a manner to avoid
headspace above 250 samples of badger loss of volatile material and to avoid
serum (Spooner et al., 2009). They found that contamination with VOC in the local
by using the method of partial least squares environment. Analysis of body fluids
discriminant analysis on the raw MS data, potentially containing tuberculosis bacilli
they could distinguish between tuberculosis- also requires careful health and safety
positive and tuberculosis-negative badgers assessment. Breath samples in particular
with 88% sensitivity. However, a false- need to be collected in a safe manner, as they
positive rate of 38% was reported. The study may contain infectious droplet nuclei that
was limited by inappropriate sample storage could spread tuberculosis. Whether the
and low volumes of sample; nevertheless, ancient physicians will be proved correct in
the results were encouraging, again imply- their assessment of odours as a tool to
ing that there were patterns of VOCs that diagnose tuberculosis remains to be seen, but
could distinguish tuberculosis-positive from current evidence suggests this is a distinct
tuberculosis-negative samples. possibility.
References Fend, R., Geddes, R., Lesellier, S., Vordermeier,

H.M., Corner, L.A., Gormley, E., et al. (2005)
Amann, A., Poupart, G., Telser, S., Ledochowski, Use of an electronic nose to diagnose Myco-
M., Schmid, A. and Mechtcheriakov, S. (2004) bacterium bovis infection in badgers and cattle.
Applications of breath gas analysis in medicine. Journal of Clinical Microbiology 43, 1745–1751.
International Journal of Mass Spectrometry Fend, R., Kolk, A.H., Bessant, C., Buijtels, P.,
239, 227–233. Klatser, P.R. and Woodman, A.C. (2006)
Ashley, D.L., Bonin, M.A., Cardinali, F.L., McCraw, Prospects for clinical application of electronic-
J.M. and Wooten, J.V. (1994) Blood-con- nose technology to early detection of Myco-
centrations of volatile organic-compounds in a bacterium tuberculosis in culture and sputum.
nonoccupationally exposed US population and Journal of Clinical Microbiology 44, 2039–2045.
in groups with suspected exposure. Clinical Firestein, S. (2001) How the olfactory system
Chemistry 40, 1401–1404. makes sense of scents. Nature 413, 211–218.
Ashley, D.L., Bonin, M.A., Cardinali, F.L., McCraw, Guerrieri, F., Schubert, M., Sandoz, J.C. and Giurfa,
J.M. and Wooten, J.V. (1996) Measurement of M. (2005) Perceptual and neural olfactory
volatile organic compounds in human blood. similarity in honeybees. PLoS Biology 3, e60.
Environmental Health Perspectives 104, 871– Hippocrates and Adams, T.B.F. (400 BC) Aphorisms.
877. eBooks@Adelaide, Adelaide, Australia.
Barker, M., Hengst, M., Schmid, J., Buers, H.J., Jia, M.Y., Koziel, J. and Pawliszyn, J. (2000) Fast
Mittermaier, B., Klemp, D., et al. (2006) Volatile field sampling/sample preparation and quantifi-
organic compounds in the exhaled breath of cation of volatile organic compounds in indoor
young patients with cystic fibrosis. European air by solid-phase microextraction and portable
Respiratory Journal 27, 929–936. gas chromatography. Field Analytical Chemistry
Bittermann, M.E., Menzel, R., Fietz, A. and Schäfer, and Technology 4, 73–84.
S. (1983) Classical conditioning of proboscis Johnston, J.M. (1999) Canine detection capabilities:
extension in honeybees (Apis mellifera). Journal operational implications of recent R & D findings
of Comparative Psychology 97, 107–119. (http://www.barksar.org/K-9_Detection_Capabil
Blount, B.C., Kobelski, R.J., McElprang, D.O., ities.pdf, accessed March 2011).
Ashley, D.L., Morrow, J.C., Chambers, D.M., et
Knobloch, H., Turner, C., Spooner, A. and
al. (2006) Quantification of 31 volatile organic
Chambers, M. (2009) Methodological variation
compounds in whole blood using solid-phase
in headspace analysis of liquid samples using
microextraction and gas chromatography-mass
electronic nose. Sensors and Actuators
spectrometry. Journal of Chromatography
B-Chemical 139, 353–360.
B-Analytical Technologies in the Biomedical and
Knobloch, H., Schroedl, W., Turner, C., Chambers,
Life Sciences 832, 292–301.
M. and Reinhold, P. (2010) Electronic nose
Bomback, A. (2006) The physical exam and the
responses and acute phase proteins correlate
sense of smell. New England Journal of
in blood using a bovine model of respiratory
Medicine 354, 327–329.
infection. Sensors and Actuators B-Chemical
Bowditch, V.Y. (1889) Two cases of phthisis treated
by intra-pulmonary injections. Boston Medical 144, 81–87.
and Surgical Journal 120, 455–458. Kolk, A., Hoelscher, M., Maboko, L., Jung, J.,
Chan, H.P., Lewis, C. and Thomas, P.S. (2009) Kuijper, S., Cauchi, M., et al. (2010) Electronic-
Exhaled breath analysis: novel approach for nose technology in diagnosis of TB patients
early detection of lung cancer. Lung Cancer 63, using sputum samples. Journal of Clinical
164–168. Microbiology 48, 4235–4238.
Cunha, M.G., Hoenigman, S., Kanchagar, C., Krylov, E.V., Nazarov, E.G. and Miller, R.A. (2007)
Rearden, P., Sassetti, C.S., Trevejo, J.M., et al. Differential mobility spectrometer: model of
(2008) Joint analysis of differential mobility operation. International Journal of Mass
spectrometer and mass spectrometer features Spectrometry 266, 76–85.
for tuberculosis biomarkers. 2008 30th Annual Li, Y.F. and Qian, R.Y. (1993) Stability of conducting
International Conference of the Ieee Engineer- polymers from the electrochemical point-of-
ing in Medicine and Biology Society, Vols 1–8. view. Synthetic Metals 53, 149–154.
Engineering in Medicine and Biology Society, Louis, P.C.A. (1844) Organs of respiration. Section
New Jersey, pp. 359–362. 1: The lungs. In: Society, T.S. (ed.) Researches
Drabkin, I.E. (ed.) (1950) Caelius Aurelianus, On on Phthisis: Anatomical, Pathological and
Acute Diseases and on Chronic Diseases. Therapeutical, 2nd edn. Translated by W.H.
University of Chicago Press, Chicago, Illinois. Walshe. C & J Adlard, London, pp. 1–35.
McNerney, R. and Daley, P. (2011) Towards a point- comparative study. Transactions of the ASAE
of-care test for active tuberculosis: obstacles 47, 2145–2152.
and opportunities. Nature Reviews Microbiology Romain, A.C. and Nicolas, J. (2010) Long-term
9, 204–213. stability of metal oxide-based gas sensors for
Miekisch, W., Schubert, J.K. and Noeldge- e-nose environmental applications: an overview.
Schomburg, G.F.E. (2004) Diagnostic potential Sensors and Actuators B-Chemical 146, 502–
of breath analysis – focus on volatile organic 506.
compounds. Clinica Chimica Acta 347, 25–39. Rouquier, S. and Giorgi, D. (2007) Olfactory
Molina, M.A., Zhao, W., Sankaran, S., Schivo, M., receptor gene repertoires in mammals. Mutation
Kenyon, N.J. and Davis, C.E. (2008) Design-of- Research 616, 95–102.
experiment optimization of exhaled breath Rouquier, S., Blancher, A. and Giorgi, D. (2000) The
condensate analysis using a miniature olfactory receptor gene repertoire in primates
differential mobility spectrometer (DMS). and mouse: evidence for reduction of the
Analytica Chimica Acta 628, 155–161. functional fraction in primates. Proceedings of
Nicolas, J. and Romain, A.C. (2004) Establishing the National Academy of Sciences 97, 2870–
the limit of detection and the resolution limits of 2874.
odorous sources in the environment for an array Sandoz, J.C. (2003) Olfactory perception and
of metal oxide gas sensors. Sensors and learning in the honey bee (Apis mellifera):
Actuators B-Chemical 99, 384–392. calcium imaging in the antenna lobe. Journal de
Palmer, R. (1993) In bad odour: smell and its la Société de Biologie 197, 277–282.
significance in medicine from antiquity to the Scholler, C., Molin, S. and Wilkins, K. (1997) Volatile
seventeenth century. In: Bynum, W.F. and Porter, metabolites from some gram-negative bacteria.
R. (eds) Medicine and the Five Senses. Chemosphere 35, 1487–1495.
Cambridge University Press, Cambridge, UK. Shattuck, F.C. (1880) Fibroid phthisis. Boston
Pavlou, A.K., Magan, N., Jones, J.M., Brown, J., Medical and Surgical Journal 102, 241–243.
Klatser, P. and Turner, A.R. (2004) Detection of Spooner, A.D., Bessant, C., Turner, C., Knobloch,
Mycobacterium tuberculosis (TB) in vitro and in H. and Chambers, M. (2009) Evaluation of a
situ using an electronic nose in combination combination of SIFT-MS and multivariate data
with a neural network system. Biosensors and analysis for the diagnosis of Mycobacterium
Bioelectronics 20, 538–544. bovis in wild badgers. Analyst 134, 1922–1927.
Phillips, M., Cataneo, R.N., Condos, R., Ring Staples, E.J. and Viswanathan, S. (2008) Detection
Erickson, G.A., Greenberg, J., La Bombardi, V., of contrabands in cargo containers using a
et al. (2007) Volatile biomarkers of pulmonary high-speed gas chromatograph with surface
tuberculosis in the breath. Tuberculosis acoustic wave sensor. Industrial and Engineer-
(Edinburgh) 87, 44–52. ing Chemistry Research 47, 8361–8367.
Phillips, M., Basa-Dalay, V., Bothamley, G., Syhre, M. and Chambers, S.T. (2008) The scent of
Cataneo, R.N., Lam, P.K., Natividad, M.P., et al. Mycobacterium tuberculosis. Tuberculosis (Edin-
(2010) Breath biomarkers of active pulmonary burgh) 88, 317–323.
tuberculosis. Tuberculosis (Edinburgh) 90, 145– Syhre, M., Manning, L., Phuanukoonnon, S.,
151. Harino, P. and Chambers, S.T. (2009) The scent
Poling, A., Weetjens, B.J., Cox, C., Mgode, G., of Mycobacterium tuberculosis – part II breath.
Jubitana, M., Kazwala, R., et al. (2010) Using Tuberculosis (Edinburgh) 89, 263–266.
giant African pouched rats to detect tuberculosis Watson, G.W., Staples, E.J. and Viswanathan, S.
in human sputum samples: 2009 findings. (2003) Performance evaluation of a surface
American Journal of Tropical Medicine and acoustic wave analyzer to measure VOCs in air
Hygiene 83, 1308–1310. and water. Environmental Progress 22, 215–
Probert, C.S.J., Ahmed, I., Khalid, T., Johnson, E., 226.
Smith, S. and Ratcliffe, N. (2009) Volatile organic Weetjens, B.J., Mgode, G.F., Machang’u, R.S.,
compounds as diagnostic biomarkers in Kazwala, R., Mfinanga, G., Lwilla, F., et al.
gastrointestinal and liver diseases. Journal of (2009) African pouched rats for the detection of
Gastrointestinal and Liver Diseases 18, 337– pulmonary tuberculosis in sputum samples.
343. International Journal of Tuberculosis and Lung
Rains, G.C., Tomberlin, J.K., D’Alessandro, M. and Disease 13, 737–743.
Lewis, W.J. (2004) Limits of volatile chemical Wheeler, P.R. and Radcliffe, C. (1994) Metabolism
detection of a parasitoid wasp, Microplitis of Mycobacterium tuberculosis. In: Bloom, B.R.
croceipes, and an electronic nose: a (ed.) Tuberculosis: Pathogenesis, Protection,
and Control. American Society for Microbiology, mass spectrometer. Journal of the American
Washington. Society for Mass Spectrometry 19, 1442-1448.
White, W.H. (1892) On a condition of the urine Yeo, B.I. (1877) On the results of recent researches
met with in phthisis. British Medical Journal 1, in the treatment of phthisis. British Medical
1638. Journal 1, 195–197.
Yang, M., Kim, T.Y., Hwang, H.C., Yi, S.K. and Kim, Zolotov, Y.A. (2006) Ion mobility spectrometry.
D.H. (2008) Development of a palm portable Journal of Analytical Chemistry 61, 519–519.
Part II
Measuring Resistance
7 Role of Phenotypic Methods for Drug
Susceptibility Testing of M. tuberculosis
Isolates in the Era of MDR and XDR
Epidemics
Leonid Heifets
Mycobacteriology Reference Laboratory, National Jewish Health, Denver,
Colorado, USA
7.1 Introduction resistant Mycobacterium tuberculosis in most

parts of the United States, the cost of routine
7.1.1 History lessons testing of all initial isolates is difficult to
justify’ (ATS, 1990).
The broad application of antimicrobial Severe outbreaks of drug-resistant
therapy of tuberculosis in the 1950s has not tuberculosis in the USA in the late 1980s and
only changed the course of history of this early 1990s changed this attitude and testing
disease but has also influenced social and of initial isolates for drug susceptibility
political life in Europe and the USA. became mandatory (Tenover et al., 1993). For
Continuing the broad application of anti- many years, this lesson did not influence
tuberculosis drugs resulted in the inevitability attitudes outside the USA. Moreover, with
of emerging drug resistance of tubercle bacilli the introduction of the directly observed
(Gupta et al., 2004). The importance of this therapy short-course (DOTS) strategy in
problem, in particular the role of patients 1994–1995, new arguments against the
with primary drug resistance as a source of worldwide use of drug susceptibility testing
infection with drug-resistant tuberculosis, emerged. One argument was that many
has been underestimated for many years. countries could not afford the laboratory
Therefore, drug susceptibility testing of arrangement needed for such testing.
patients’ isolates (particularly in new Another reason was that, according to the
patients) was not considered a priority in the original DOTS strategy, the focus was on
measures of control. Statements from leading using smear microscopy as the main
national and international organizations diagnostic tool to detect most infectious
illustrated this neglect. For example, the patients. It did not matter that this testing
American Thoracic Society (ATS) and the was only able to detect less than 50% of the
Centers for Disease Control and Prevention culture-positive patients, but culture
(CDC) published the following joint isolation and drug susceptibility testing were
statement, ‘Given the low prevalence of drug- not considered for broad implementation.

96 L. Heifets
Table 7.1. Abbreviations of the drugs’ names. why primary resistance hardly affects the
Isoniazid INH or H
outcome of treatment with a WHO standard
regimen …’.
Rifampin RMP, RIF or R
Have we learned lessons from this
Pyrazinamide PZA or Z
approach?
Streptomycin SM or S
Ethambutol EMB or E
Ethionamide ETA 7.1.2 The alarming situation today and
Amikacin AK new developments
Kanamycin KM
Capreomycin CM Tuberculosis today is more widespread than
Cycloserine CS at any other time in the history of humankind.
Para-amino-salicylic acid PAS A recent report by the WHO indicates that, in
Levofloxacin Levo 2008, there were estimated 440,000 cases of
Moxifloxacin Moxi MDR-TB, resulting in the death of 150,000
Ofloxacin Oflox patients (WHO, 2010). MDR-TB cases have
Clofazimine CF been reported in more than 50 countries, but
the WHO considered 27 of them as ‘high MDR
burden countries’. For example, in seven areas
of the former Soviet Union, the proportion of
The argument was that proper imple- MDR-TB in new cases ranged between 12%
mentation of the DOTS strategy (including a and 27% and between 50% and 60% in
standard treatment regimen for all patients) previously treated cases. At the beginning of
was supposed to prevent the spread of drug 2010, 58 countries had reported the emergence
resistance, and even make it ‘virtually of extremely drug-resistant tuberculosis
impossible’ for a patient to develop (XDR-TB). It is estimated that by 2015, we will
multidrug-resistant tuberculosis (MDR-TB) face nearly 1.6 million new cases of drug-
(World Health Organization [WHO], 1997). resistant tuberculosis annually, and it is hard
A group of authors referring to their to predict what proportion of them will
experience in Peru stated, ‘It is not true that represent MDR-TB and XDR-TB cases.
DOTS makes it virtually impossible to cause There is a new important development
a patient to develop MDR-TB’ (Farmer et al., (compared to reports of previous years): the
1998). They demonstrated how patients with WHO 2010 report states that, ‘The laboratory
initial resistance to rifampin (RIF) or plays a central role in patient care and
isoniazid (INH) developed additional surveillance’ and ‘Establishing reference
resistance to other drugs through DOTS laboratory facilities to supervise DST (sic
treatment, a phenomenon they labelled as Drug Susceptibility Testing) surveillance
‘amplified effect of short-course therapy’. activities in the country is a critical step in
Escalation of this phenomenon into the MDR MDR-TB control and care’. Yes, it is a ‘critical
epidemics became unavoidable through step’ forward, but is this general statement
blindfolded application of either a standard sufficient to address the issue of timely
regimen or through the addition of any drug detection of patients that harbour drug-
to the failing treatment regimens, without resistant bacteria? The report does say that
drug susceptibility results. In 1997, the WHO the WHO ‘will promote new and rapid
guidelines for the management of drug- technologies within appropriate laboratory
resistant tuberculosis contained the following services through 2013 …’. It is not clear from
statements, ‘Susceptibility testing is not this report whether these new and rapid
recommended in all new cases of smear- technologies are planned for another sur-
positive pulmonary tuberculosis… since it is veillance programme only or if there is a plan
not practical, it is expensive, and it is useless’ to implement these methods widely enough
and ‘The level of resistance… is lower in to check all new patients in communities for
primary than in acquired resistance. This is drug-resistant tubercle bacilli. The report
Role of Phenotypic Methods for Drug Susceptibility Testing 97
does not address the issue of the organization laboratory reports depends on the specific
of laboratory services beyond having national conditions of the laboratory.
reference laboratories. In other WHO docu-
ments, there were also recommendations on
the enhancement of laboratory services, 7.2 Choices of the Phenotypic
including drug susceptibility testing, but so Methods
far, there have been no specifics on this
‘enhancement’. This overview is based on the update of our
In the USA, the Federal Tuberculosis previous reports (Heifets, 1999, 2000; Heifets
Task Force published the ‘Action Plan to and Gangelosi, 1999, 2009). Based on the type
Combat Extensively Drug-Resistant Tubercu- of culture medium, solid or liquid, all drug
losis’ (CDC, 2009), in which two statements susceptibility methods based on cultivation
addressed the need for drug susceptibility can be divided into two groups. Solid media
testing: ‘The laboratory plays a critical role in include the following types: Löwenstein–
the diagnosis and management of drug- Jensen (LJ) or Ogawa egg-based media,
resistant TB’ and ‘To combat the growing Middlebrook 7H10 or 7H11 agar media and
problem of resistance to TB drugs, the most HSTB agar medium. Liquid media are used
current methods need to be applied to their fullest in automated or semi-automated liquid
capacity while better diagnostic tests are medium systems, such as BACTEC 460 and
developed’. The reason for such an endorse- BACTEC 960 (MGIT), MB/BacT and
ment of the ‘current methods’ is an attempt to VersaTREK. The advantage of the solid media
confront the widespread tendency of post- over the liquid systems is that they can be
poning broad implementation of drug used for a direct drug susceptibility test with
susceptibility testing until the development smear-positive sputum specimens, which
and introduction of new rapid and affordable provides a relatively short turnaround time
methods that would not require standard of the laboratory reports. Although liquid
tuberculosis laboratory settings. media cannot be used for a direct test, it has
The phenotypic methods for drug the advantage of rapid growth detection. The
susceptibility testing are among these ‘current advantages of solid media are that, historic-
methods’, with or without combination with ally, they have been in use for a longer period,
molecular methods. Progress in molecular with more experience and better validation of
biotechnology makes it possible to detect procedures using a large variety of drugs.
MDR and XDR directly in raw specimens. The lower cost of solid media is another
Unfortunately, even the best of these methods advantage compared to the liquid medium
are not sensitive enough to provide reliable system. Liquid medium systems can be used
results with smear-negative specimens. There- not only for qualitative tests with critical
fore, for more than 50% of new patients concentrations but also for a quantitative
having small contents of mycobacteria in determination of the MICs. These values can
their sputum (acid-fast bacilli smear-negative be compared with the pharmacokinetic
specimens), particularly among those with parameters (for example with the Cmax) as an
dual tuberculosis+HIV infection, these option for interpretation of the drug
methods are not applicable. Therefore, there susceptibility test results and the predict-
is a need for phenotypic cultural methods to ability of the patient’s response to therapy.
test cultures grown from smear-negative Qualitative tests, either in solid or liquid
specimens. A combination of molecular and media, are performed with the so-called
phenotypic methods may provide the best critical drug concentrations that were pro-
results, combining rapidity for approximately posed for distinguishing between ‘resistant’
half of the specimens and addressing speci- and ‘susceptible’ isolates. These critical con-
mens with low mycobacterial content. The centrations are not intended for comparison
choice of specific algorithm testing with a with the pharmacokinetic parameters, and
general goal of obtaining results in the they are different for various media (Tables
shortest possible turnaround time of the 7.2–7.4). The concentrations shown in Tables
98 L. Heifets
Table 7.2. Critical concentrations (μg/ml) for testing M. tuberculosis in solid

media (National Jewish Health Mycobacteriology Laboratory, Denver,
Colorado).
Drug LJ 7H10 agar 7H11 agar HSTB agar
Isoniazid 0.2 0.2, 1.0 0.2, 1.0 0.2, 1.0
Rifampin 40.0 1.0 1.0 1.0
Ethambutol 2.0 5.0, 10.0 7.5 12.0
Pyrazinamide – – – 900.0
Streptomycin 4.0 2.0, 10.0 2.0, 4.0 5.0, 10.0
Amikacin 20.0 4.0 6.0 6.0
Kanamycin 20.0 5.0 6.0 6.0
Capreomycin 20.0 10.0 10.0 10.0
Ethionamide 20.0 5.0 10.0 10.0
Cycloserine – 20.0 60.0 60.0
PAS 0.5 2.0 8.0 8.0
Levofloxacin – – 4.0 4.0
Moxifloxacin – – 4.0 4.0
Table 7.3. Critical concentrations (μg/ml) of first-line drugs for testing M. tuberculosis in liquid medium
systems.
BACTEC
MB/BacT
460 BACTEC
(Heifets and 960 Bemer Tortoli Heifets’
BACTEC Gangelosi, (MGIT) VersaTREK et al., et al., proposal
Drug 460 insert 2009) insert insert 2004 2000 2010
Isoniazid 0.1, 0.4 – 0.1 0.1, 0.4 0.09 1.0 0.1, 0.4
Rifampin 2.0 0.5, 2.0 1.0 1.0 0.9 1.0 2.0
Ethambutol 2.5, 7.5 5.0 5.0 5.0, 8.0 3.5 2.0 5.0
Streptomycin 2.0, 6.0 5.0 1.0 – 0.45 1.0 2.0
Pyrazinamide 100.0 300.0 100.0 300.0 200.0 50.0 300.0
Table 7.4. Suggested critical concentrations of second-line drugs for BACTEC 460 and BACTEC MGIT
960 systems.
Suggested for the
Suggested for the Suggested for the BACTEC MIGIT 960
BACTEC 460 system BACTEC 460 system system (Krüüner et al.,
Drug (Heifets, 1991) (Pfyffer et al., 1999) 2006)
Ethionamide 2.5 1.25 2.5, 5.0
Kanamycin 5.0 5.0 –
Amikacin 5.0 1.0 1.0
Capreomycin 5.0 1.25 1.25
Levofloxacin 4.0 – –
Moxifloxacin 4.0 – 0.5, 1.0
Ofloxacin 4.0 2.0 1.0
7.2–7.4 are those incorporated into the media, concentrations, as well as for selection of the
not the concentrations actually interacting quality control strains.
with the bacteria. Due to the various degrees A quantitative test to determine the MIC
of degradation and deterioration of the drugs as the lowest drug concentration that inhibits
in different media during their preparation, the bacterial growth completely has many
storage and incubation, the actual concen- advantages and fewer problems over the
trations of the drugs interacting with the qualitative test with critical concentrations.
bacteria are variable and are not defined in So far, such a methodology has been
the descriptions of the procedures. developed for testing in 7H12 broth in the
The critical concentrations are rather BACTEC 460 system (Heifets, 1988, 1991).
empirical values and some of them have not There is no known significant deterioration/
been calibrated properly, even for traditional degradation of drugs in this simple medium,
solid media such as 7H10 and 7H11 agar and there is the assumption that the drug
(Mitchison, 1998). The issue is even more concentrations incorporated into 7H12 broth
problematic for the liquid medium systems are actually those that interact with the
(Tables 7.3 and 7.4). bacterial inoculum. Therefore, unlike testing
There were problems with the definition in solid medium, the MIC determined in this
of susceptible and resistant strains. Originally, system can be correlated with the concen-
the classical definition of the drug-resistant trations attainable in blood. This approach
strain of M. tuberculosis was that it was was used for interpretation of the MICs
significantly different by a degree of sus- shown in Table 7.5. Complete inhibition of
ceptibility from a wild strain that had never growth by the lowest of the three concen-
come into contact with the drug (Canetti et al., trations is indicative of interpretation as
1969; Mitchison, 1969). In reality, over a ‘susceptible’. No inhibition of growth by the
period of many years, this definition became highest concentration is reported as MIC
unreliable, particularly in determining critical equal or higher than this concentration and is
concentrations. interpreted as ‘resistant’. Inhibition of growth
The general idea of a correct critical by the highest and the middle concentrations
concentration is that it should be somewhere without growth inhibition by the lowest
between the highest MIC found for fully concentration gives an MIC that is interpreted
susceptible isolates (‘wild’ strains) and the as ‘moderately susceptible’.
lowest MIC for resistance to the particular
drug strains. Unfortunately, this principle is
rarely used for establishing the critical 7.2.1 Löwenstein–Jensen medium
concentration for the new cultivation systems.
Instead, some authors select these concen- This medium is the most traditional and
trations on the basis of comparison with other oldest medium used for both cultivation and
media or systems. The worst scenario, when drug susceptibility testing of M. tuberculosis.
based on MICs found for susceptible strains, Unfortunately, because of the strong
is that a too low concentration is selected as a attachment to the historically based tradition,
critical concentration. Use of such low this medium is often the only medium used
concentrations often results in false resistance for drug susceptibility testing in many parts
reports. It is known that resistance of tubercle of the world. The major problem is that
bacilli to some drugs, if it is a true resistance, cultivation on any egg-based medium (LJ or
can be demonstrated at any high concen- Ogawa) takes a longer period for growth
trations due to the loss of the genetic target recovery as compared to agar or liquid media.
through the mutation. Therefore, the pos- That is why the results obtained on this
sibility of false susceptible results is lower medium are often reported too late for a
than that of false resistance. It seems that, physician to make any adjustment to the
currently, there is the possibility of addressing treatment regimen. The situation is aggravated
this issue by genetic testing for mutations as a by the tendency to use this technology as an
basis for better choice of the correct critical indirect rather than a direct test.
100 L. Heifets
Table 7.5. Suggested guidelines for interpretation of MICs (μg/ml)

determined in the BACTEC 460 system for M. tuberculosis isolates
(in use at the National Jewish Health Mycobacteriology Reference
Laboratory, Denver, Colorado).
Moderately
Drug Susceptible susceptible Resistant
Isoniazid 0.1 0.4 1.6
Rifampin 0.5 2.0 8.0
Rifabutin 0.12 0.25 0.5
Ethambutol 4.0 8.0 16.0
Ethionamide 1.25 2.5 5.0
Streptomycin 2.5 5.0 10.0
Amikacin 2.5 5.0 10.0
Kanamycin 2.5 5.0 10.0
Capreomycin 2.5 5.0 10.0
Ofloxacin 2.0 4.0 8.0
Levofloxacin 2.0 4.0 8.0
Moxifloxacin 2.0 4.0 8.0
Cycloserine 4.0 8.0 16.0
Clofazimine 0.12 0.25 0.5
Historically, three methods based on slant. The tubes are incubated at 37°C for 1–2
cultivation of M. tuberculosis on starch-free LJ days in a slope position with caps slightly
medium have been proposed: (i) the pro- ajar. Afterwards, the caps are tightened and
portion method; (ii) the resistance ratio (RR) the tubes are incubated in an upright position.
method; and (iii) the absolute concentration On the 28th day of incubation, the colonies
method (Canetti et al., 1963, 1969). Each of are counted to calculate the proportion of
these methods can be used as either a direct colonies grown on drug-containing media to
or indirect test, but in reality a direct test is the number of colonies in the controls. Any
rarely used with this medium. proportion that exceeds 1% for INH, RIF and
The proportion method in its simplified PAS (para-amino-salicylic acid), and 10% for
version has become the most popular among other drugs, indicates ‘resistant’, and the
these three options. Both drug-free (controls) results are final. If the proportion is less than
and drug-containing egg-based medium 1%, then a second reading is required on the
suspensions are coagulated in tubes in a slope 42nd day of cultivation to confirm that the
position at 85°C for 50 min. This process isolate is ‘susceptible’ to the specific drug.
alone inactivates a substantial proportion of The RR method is perhaps the most
some drugs in addition to their absorption by accurate and most labour-intensive and
the medium. Therefore, the critical drug costly of the three methods. The RR is defined
concentrations developed for this medium as a ratio of MIC for the patient’s isolate to the
(Table 7.2) reflect the amounts of drugs added MIC for the drug-susceptible reference strain
and not those that remain in the medium to (H37Rv), both tested in the same experiment.
interact with bacteria. Two sets of tubes with MIC determination requires a large number
drug-containing and drug-free medium are of medium tubes containing a broad range of
inoculated with a bacterial suspension. There drug concentrations. For this method, the
are various techniques to calibrate the MIC is defined as the lowest drug concen-
inoculum, with a goal of being sufficient to tration in the presence of which the number
produce a growth of 100–200 colonies on a of colonies is less than 20 after 4 weeks of
incubation. RR of 2 or less is an indication of albumin-dextrose-catalase) for 7H10 or 7H11

‘susceptible’ and RR of 8 or greater indicates medium, and calf or horse serum for HSTB
‘resistant’. agar. The appropriate drug solutions are
The absolute concentration method is added to ensure the intended critical concen-
based on the comparison of growth intensity trations (Table 7.2). The medium is distributed
in the presence of critical concentrations into plastic plates, usually into quadrant
(adjusted for each laboratory) and on drug- plates, approximately 5.0 ml/quadrant. It is
free controls. The original H37Rv strain now known that some isolates may have
preserved in aliquots at our laboratory is genetically predetermined low levels of
tested in parallel with the same drugs in resistance to INH (Madison et al., 2004).
multiple concentrations for QC of reproduci- Therefore, two concentrations of INH are
bility. The results are determined after 4 needed to distinguish between low and high
weeks of incubation, or after 5–6 weeks if levels of resistance to this drug. A higher
growth is insufficient at the 4-week reading. concentration of streptomycin (SM) in
A ‘susceptible’ result is reported if the number addition to the critical concentration of this
of colonies in the presence of a drug is less drug is usually employed in agar media for
than 20, while there are more than 100 confirmation of resistance on some occasions
colonies grown in drug-free controls. when the presence of CO2 in the incubator
(for 7H10 and 7H11 agar) may have affected
the activity of SM.
7.2.2 Agar proportion method To detect MDR, only one quadrant plate
of any of the agar medium type (with a drug-
The major advantage of performing drug free quadrant) is needed for a test with RIF
susceptibility tests in agar plates is related to and two concentrations of INH. In the case of
the transparency of the medium, which using HSTB agar, two quadrant agar plates
makes it possible to observe the growing can be used for a test with INH, rifampin
colonies at the beginning of their formation. (RMP), ethambutol (EMB) and pyrazinamide
Therefore, final results can be reported within (PZA). Four quadrant plates of any agar type
3 weeks for most isolates, instead of 4–6 are needed for a test with ten drugs. For a
weeks or more when using egg-based media. direct test after processing the specimen,
There are at least three types of agar medium commonly described as the digestion–
that can be used for either direct or indirect decontamination procedure, the concentrated
tests: 7H10, 7H11 and HSTB. A detailed sputum is inoculated on to plates, 0.1 ml/
description of their preparation and use can quadrant. For an indirect test, two sets of
be found in the appropriate publications plates are used: one inoculated with 10–3-fold
(David, 1971; Kent and Kubica, 1985; Heifets, and the other with 10-5-fold dilutions of the
2000; Heifets and Sanchez, 2000, 2003). These bacterial suspension adjusted to the optical
media are made from commercially available density of the McFarland No 1 standard. The
7H10 or 7H11 agar base. The HSTB agar is plates are incubated at 35–37°C for 3 weeks,
made from the 7H11 agar base with the protected from light, in an atmosphere of
addition of monosodium phosphate (0.6 5–10% CO2 for 7H10 and 7H11 agar plates
g/100 ml of medium) to change the pH of the and in a regular incubator (without CO2) for
medium from 6.6 ± 0.2 to 6.2 ± 0.2. the HSTB agar. The plates are incubated in
To prepare agar media, one should fol- sealed plastic bags. After incubation, the
low the manufacturer’s instruction marked plates are placed on the bench at room
on the label placed on each container of the temperature for 3 h or overnight in an upside
base powder. The base powder is suspended position (agar up) for elimination of the
in distilled water containing 0.5% glycerol, condensate. The colonies are counted and the
autoclaved at 121°C for 15 min, and then results are reported as the percentage (pro-
cooled in a water bath to 52–54°C. After- portion) based on comparison of the number
wards, 10% enrichment supplement is added: of colonies on drug-containing and drug-free
the commercially available OADC (oleate- quadrants. The isolate is considered ‘resistant’
102 L. Heifets
if this proportion is 1% or greater for all drugs monitors the increase of fluorescence. Com-
except PZA. The established criterion for parison of these patterns in drug-containing
PZA is 10%. For 10–15% of isolates, the and drug-free tubes is analysed automatically
growth may not be sufficient at the 3-week by the instrument and reported as ‘sus-
reading. Then, the plates are re-examined at ceptible’ or ‘resistant’.
the 6-week reading, but in such cases only In the MB/BacT system, a colorimetric
‘susceptible’ results are considered valid. This sensor detecting the release of CO2 by the
is because growth at 6 weeks in drug- growing bacteria is embedded in the bottom
containing quadrants may be related to drug of the vial. The instrument continuously
degradation during the prolonged incubation monitors and reports the colour as it changes
period rather than to the occurrence of true from dark green to yellow. A ‘susceptible’
drug resistance. result is reported if no colour changes occur
in the drug-containing vial or if a positive
detection time in this vial is greater than in
7.2.3 Qualitative indirect test in liquid the drug-free control. A ‘resistant’ result is
media reported if growth is detected in the drug-
containing vial or if the detection time for this
The need for expedited detection of drug vial is shorter than for the drug-free control.
resistance was first addressed in the 1980s by In the VersaTREK system, growth
developing the drug susceptibility testing monitoring is based on the reduction of
procedure for the semi-automated BACTEC pressure in the vials due to the consumption
460 system introduced by Becton Dickinson, of oxygen by the growing bacteria. The
Sparks, Maryland, USA (Siddiqi et al., 1981, conclusion is based on the comparison of
1985; Roberts et al., 1983). The liquid medium drug-free controls and drug-containing vials
for this system, 7H12 broth in 12B vials, after positive readings have occurred for 3
contains 14C-substrate, consumption of which consecutive days in the drug-free vial.
by the growing bacteria results in the release ‘Susceptible’ is reported if no growth is
of 14CO2, measured by the instrument and detected in the drug-containing vials and
expressed as the growth index (GI). The ‘resistant’ is reported if growth is detected in
turnaround time for an indirect test in this the drug-containing vials at this time point.
system was 9.3 days, and the overall mean For all four liquid medium systems,
time for primary isolation plus indirect test critical concentrations have been either
was 18 days in a cooperative study by five developed or suggested for four to five first-
institutions (Roberts et al., 1983; Heifets, line drugs: INH, RIF, EMB, SM, PZA (4.3).
1986). The major disadvantage of this system Two concentrations of INH are needed to
was the problem of disposal of radioactive distinguish between low and high levels of
materials (12B vials), which stimulated the resistance to this drug (Madison et al., 2004).
development of new non-radiometric systems. A lower concentration of EMB (2.5 μg/ml) by
Three such systems, all fully automated Becton Dickinson for the BACTEC 460 system
and computerized, became commercially appeared to cause false resistance results, and
available: BACTEC MGIT 960 (Becton therefore the manufacturer suggested an
Dickinson Microbiology Systems, Sparks, additional higher concentration of 7.5 μg/ml.
Maryland, USA), MB/BacT (bioMérieux, Dur- Instead, we introduced a concentration of 4.0
ham, North Carolina, USA) and VersaTREK, μg/ml and, later, 5.0 μg/ml, which appeared
formerly the ESP-II Culture System by Difco to be quite reliable in distinguishing between
(TREK Diagnostic Systems, Westlake, Ohio, resistant and susceptible strains. For address-
USA). ing the problem of false resistance to PZA, we
In the BACTEC MGIT 960 system, the replaced the concentration of 100.0 μg/ml
bacterial growth detection is based on the suggested by the manufacturer with 300.0
consumption of oxygen, which causes the μg/ml (Heifets, 1991, 2002).
indicator embedded in the bottom of tubes to Critical concentrations for the MB/BacT
fluoresce, and the instrument continuously system shown in Table 7.3 include references
to two publications to illustrate the diversity gested by the manufacturer or various

in the opinions of different authors regarding authors.
the critical concentrations needed (Tortoli et The validity of results by non-radiometric
al., 2000; Bemer et al., 2004). Critical concen- systems has been analysed in some publi-
trations of various second-line drugs were cations based on comparison with the
suggested for the BACTEC 460 and BACTEC BACTEC 460 system and/or with the agar
MGIT 960 systems (Table 7.4) (Heifets, 1991; proportion method. Both are used as
Pfyffer et al., 1999; Krüüner et al., 2006). We reference methods. Unfortunately, even the
are proposing now to use for this system the agar proportion method has been criticized
same concentrations as we are using for the as having improperly calibrated critical
BACTEC 460. There are debates about concentrations (Mitchison, 1998). As men-
concentrations of quinolones to be defined as tioned above, critical concentrations for the
critical for testing in vitro. Unfortunately, BACTEC 460 system are also far from perfect.
there have been no clinical trails completed Therefore, acceptance of these two systems as
yet to indicate any possible correlation reference methods may have a relative value.
between resistance to various concentrations It would be more reasonable to use strains,
of quinolones in vitro and a patient’s response resistant and susceptible, with an established
to therapy with these drugs. Often, testing genetic characterization for the development
with very low concentrations results in false of critical concentrations in new systems.
resistance reports. One should keep in mind Such an approach would resolve conflicting
that for these drugs, as is true for some other results and suggestions from different
agents as well, loss of a genetic target due to laboratories.
mutation makes the isolate resistant to any A review of the literature regarding drug
drug concentration. Therefore, it is reasonable susceptibility testing in non-radiometric
to use a drug concentration significantly automated liquid medium systems suggests
higher (at least four- or eightfold higher) than that each of the available systems provides an
the highest MIC found for wild strains opportunity for relatively rapid turnaround
obtained from patients never treated with the time of the indirect drug susceptibility test,
agent in question. although time ranges reported by different
In addition to the qualitative tests, a authors have been broad, and that each of
quantitative MIC test was proposed for the these methods has certain advantages and
BACTEC 460 system (Table 7.5), in which the disadvantages (Piersimoni et al., 2006).
category ‘moderately susceptible’ cor- Authors of this review concluded that the
responded to the critical concentrations in BACTEC MGIT 960 system appears to be the
Tables 7.3 and 7.4. The critical concentrations most reliable option among the non-radio-
listed in the tables reflect the situation at the metric liquid medium systems for testing
time of preparation of this paper. However, drug susceptibility to the first-line anti-
one should be aware that the manufacturers tuberculosis drugs in an indirect test.
of liquid medium systems review and change
these concentrations periodically according
to the progress in their evaluation. Up-to-date 7.3 Optimal Algorithms
suggestions are usually listed in the manu-
facturer’s inserts. In addition, some authors Described below are various combinations of
have used their own ‘critical concentrations’. different procedures presented in the order of
Recommendations by the Clinical and affordability/cost, from the simplest to the
Laboratory Standards Institute (NCCLS, most complex.
2003; CLSI, 2009) may not help to resolve The least expensive protocol, and still
inevitable controversies. Therefore, each with concern for minimizing the turnaround
laboratory should re-evaluate the protocol time of the laboratory report, is the use of a
suggested by the manufacturer and, based on single HSTB agar plate with four quadrant
appropriate studies, consider different or (segments) containing a drug-free control
additional concentrations than those sug- and three quadrants containing INH 0.2, INH
104 L. Heifets
1.0 and RIF 1.0 μg/ml. Use of this type of plate significantly. Perhaps it is reasonable to
is the simplest way to detect MDR. The costs combine the direct test on agar plates for the
of materials for such a plate is less than smear-positive specimens with culture
US$1.00 (USA costs, 2010), and it can be isolation from all specimens in the liquid
cultivated in a regular incubator (without medium, followed by the indirect test in
CO2). Instead of the HSTB agar, the test can liquid medium with cultures isolated from
be performed on 7H10 or 7H11 agar, if the smear-negative specimens or in cases of
laboratory is equipped with the CO2 failure of the direct test on agar plates. Total
incubators and has a reliable supply of the turnaround time of the laboratory reports,
OADC supplement. The turnaround time of including culture isolation and the indirect
the direct test with a smear-positive sputum test is about 3 weeks, but it is highly variable
specimen is 3 weeks for 80–85% of specimens. depending on the bacterial contents in the
For culture isolation, parallel with or without specimens.
(for smear-negative specimens) the direct The most desirable approach for obtain-
test, the specimen should be inoculated on to ing maximum information within the shortest
an agar bi-plate consisting of plain and turnaround time is to combine molecular and
selective agar. The isolated culture is neces- phenotypic methods; these are discussed in
sary for setting up an indirect test when the Chapters 2 and 9. Therefore, it should be
direct test fails, as well as for cultures isolated combined with cultivation in both liquid
from smear-negative specimens. The total medium and agar plates. Cultures obtained
turnaround time of culture isolation plus an from smear-negative specimens can be
indirect test is 6 weeks for 80–85% of smear- subjected to the Hain test for rapid detection
negative specimens. of MDR/XDR. The total turnaround time in
The next option is an expansion of the this case would be less than 3 weeks for most
agar proportion method through testing not of these smear-negative specimens. Parallel
only against INH and RIF, but, with any of testing by phenotypic methods, in liquid
the other drugs listed in Table 7.1, under medium and on agar plates, is essential not
conditions that in each plate one quadrant is only for confirmation of the results obtained
used as a control containing a drug-free by the Hain technology but also for obtaining
medium. An agar proportion method, either a broader picture of the drug resistance/
a one-plate test as described above or its susceptibility pattern of the isolate. Thus,
expanded version, should be recommended combined use of molecular and phenotypic
as a replacement for the LJ medium in methods may provide the most desirable
countries where the traditional use of the combination of the rapidity and reliability of
egg-based medium is still in place. The the results.
advantages of the agar plates over the egg-
based media are obvious in regard to the
turnaround time, cost of materials and labour. 7.4 Conclusion
Unlike with supplies sold by manufacturers
for agar plates, the basic cost for egg-based Phenotypic methods of drug susceptibility
media is the local cost of eggs. This is one of testing of the M. tuberculosis isolates can play
the reasons for the cost variability among a significant role in timely detection of
countries for this type of medium. It is likely patients with MDR and XDR. An important
that such a cost may be significantly higher element of the application of these methods is
than the cost of an agar plate. the direct test for the smear-positive speci-
Incorporation of a liquid medium system mens, preferably on agar plates, as a minimal
into the laboratory protocol for both culture algorithm for countries with limited
isolation and drug susceptibility testing resources. This option can provide a rapid
shortens the turnaround time of the drug turnaround time (3 weeks) to report on
susceptibility reports for smear-negative susceptibility to INH and RIF for 80–85% of
specimens, but it also increases the cost specimens. It can be applied to other drugs as
well, and it can be complemented with the testing and for shortening the turnaround
indirect test for cultures isolated from smear- time of the laboratory reports, as well as for
negative specimens. When resources allow, better cost-efficiency of the operation.
the addition of the liquid medium system
would expedite the results, particularly for
smear-negative specimens. Finally, a com- References
bination of phenotypic and molecular
methods is the best, although the most American Thoracic Society [ATS] (1990) Diagnostic
expensive, option for obtaining the most standards and classification of tuberculosis.
complete information in the shortest American Review of Respiratory Diseases 142,
turnaround time 3 (Supplement), 725–735.
It must be stressed that, regardless of the Bemer, P., Bodmer, T., Munzinger, J., Perrin, M.,
methods used, any of the procedures for drug Vincent, V. and Drugeon, H. (2004) Multicenter
susceptibility testing of M. tuberculosis require evaluation of the MB/BacT system for sus-
implementation of biosafety standards, ceptibility testing of Mycobacterium tubercu-
losis. Journal of Clinical Microbiology 42,
usually defined as BSL-3 practices (Richmond
1030–1034.
et al., 1996; Heifets and Richmond, 2006). Brossier, F., Veziris, N., Aubry, A., Jarlier, V. and
There are no phenotypic methods for drug Sougakoff, W. (2010) Detection by GenoType
susceptibility testing of M. tuberculosis that MTBDRsl test of complex mechanism of
can be implemented without these standards. resistance to second line drugs and ethambutol
Taking into account the need for costly in multidrug-resistant Mycobacterium tubercu-
equipment to comply with biosafety stand- losis complex isolates. Journal of Clinical
ards, as well as the application of any level of Microbiology 48, 1683–1689.
sophistication of the protocol, the most Canetti, G., Froman, S., Grosset, J., Hauduroy P.,
economical approach to drug susceptibility Langerova, M., Mahler, H.T., et al. (1963)
Mycobacteria: laboratory methods for testing
testing, along with bacteriological diagnosis
drug sensitivity and resistance. Bulletin of The
in general, is to have large laboratories with World Health Organization 29, 565–578.
direct centralized services to a country, or to Canetti, G., Fox, W., Khomenko, A., Mahler, H.T.,
significant parts of that country, and not just Menon, M.K., Mitchison, D.A., et al. (1969)
reference laboratories that would provide a Advances in techniques of testing mycobacterial
reference service to the lower level of drug sensitivity and the use of sensitivity tests
laboratories. in tuberculosis control programs. Bulletin of
Direct microscopy smear examination The World Health Organization 41, 21–43.
was recommended as the main diagnostic CDC [Centers for Disease Control and Prevention]
tool when the DOTS strategy was initially (2009) Plan to combat extensively drug-
resistant tuberculosis: recommendations of the
introduced. Under the provisions of this
Federal Tuberculosis Task Force. Morbidity and
approach, the decentralization of laboratory Mortality Weekly Report 58(RR-3), 1–43.
services in developing countries with high CLSI [Clinical and Laboratory Standard Institute]
tuberculosis prevalence was inevitable, along (2009) Susceptibility Testing of Mycobacteria,
with the needs of supervision by the reference Nocardiae, and Other Aerobic Actinimycetes.
laboratories of the quality of work at the Approved standard – Second Edition. CLSI
microscopy stations. Today, the new enhanced document M24-A2. Wayne, Pennsylvania.
DOTS strategy recommends the phased David, H.L. (1971) Fundamentals of Drug Sus-
introduction of drug susceptibility testing as ceptibility Testing in Tuberculosis. CDC/HEW
a major tool to combat MDR and XDR Publication No 00-2165. Atlanta, Georgia.
Farmer, P., Bayona, J., Becerra, M., Furin, J.,
epidemics (WHO, 2010). Implementation of
Henry, C., Hiatt, H., et al. (1998) The dilemma
this approach will require qualified labora- of MDR-TB in the global era. International
tories being able to perform culture isolation Journal of Tuberculosis and Lung Disease 2,
and drug susceptibility testing, as well as 869–876.
some molecular testing. The direct delivery of Gupta, R., Espinal, M.A. and Raviglione, M.C.
raw specimens to such laboratories will (2004) Tuberculosis as a major global health
become necessary for the sake of quality of problem in 21st century. In: Heifets, L. (ed.)
106 L. Heifets
Tuberculosis and other mycobacterial Krüüner, A., Yates, M.D. and Drobniewski, F.A.
infections. Seminars in Respiratory and Critical (2006) Evaluation of MGIT 960-based
Care Medicine 25(3), 245–253. antimicrobial testing and determination of
Heifets, L.B. (1986) Rapid automated methods critical concentrations of first- and second-line
(BACTEC System) in clinical mycobacteriology. antimicrobial drugs with drug-resistant clinical
Seminars in Respiratory Infections 1(4), 242– strains of M. tuberculosis. Journal of Clinical
249. Microbiology 44, 811–818.
Heifets, L.B. (1988) Qualitative and quantitative Lacoma, A., Garcia-Sierra, N., Prat, C., Ruitz-
drug susceptibility tests in mycobacteriology. Manzano, J., Haba, L., Rosés, S., et al. (2008)
American Review of Respiratory Diseases 137, Genotype MTBDplus assay for molecular
1217–1222. detection of rifampin and isoniazid resistance in
Heifets, L. (1991) Drug susceptibility tests in the Mycobacterium tuberculosis strains and clinical
management of chemotherapy of tuberculosis. specimens. Journal of Clinical Microbiology 46,
In: Heifets, L.B. (ed.) Drug Susceptibility in 3660–3667.
Chemotherapy of Mycobacterial Infections. Madison, B.M., Siddiqi, S.H., Heifets, L., Gross, W.,
CRC Press, Boca Raton, Florida, pp. 89–121. Higgins, M., Warren, N., et al. (2004) Identification
Heifets, L. (2000) Conventional methods for of a Mycobacterium tuberculosis strain with
antimicrobial susceptibility testing of M. stable, low-level resistance to isoniazid. Journal
tuberculosis. In: Bastian, I. and Portaels, F. (eds) of Clinical Microbiology 42, 1294–1295.
Multidrug-Resistant Tuberculosis. Kluwer Mitchison, D.A. (1969) What is drug resistance?
Academic Publisher, Dordrecht, The Tubercle 50 (Supplement), 44–47.
Netherlands, pp. 133–143. Mitchison, D.A. (1998) Standardization of sensitivity
Heifets, L. (2002) Susceptibility testing of M. tests (letter). International Journal of Tubercu-
tuberculosis to pyrazinamide. Journal of Medical losis and Lung Disease 2, 69.
Microbiology 51(1), 11–12. NCCLS [National Committee for Clinical Laboratory
Heifets, L.B. and Gangelosi, G.A. (1999) Drug Standards] (2003) Susceptibility Testing of
susceptibility testing of Mycobacterium Mycobacteria, Nocardiae, and Other Aerobic
tuberculosis: a neglected problem at the turn of Actinomycetes. Approved standard M24-A.
the century. International Journal for Tubercu- NCCLS, Wayne, Pennsylvania.
losis and Lung Disease 3(7), 564–581. Pfyffer, G.E., Bonato, D.A., Ebrahimzade, A.,
Heifets, L. and Gangelosi, G. (2009) Chapter 81: Gross, W, Higgins, M., Warren, N., et al. (1999)
Drug resistant assays for Mycobacterium Multicenter laboratory validation of susceptibility
tuberculosis. In: Mayers, D.I. (ed.) Antimicrobial testing of M. tuberculosis against second-line
Drug Resistance. Humana Press, pp. 1161– and newer antimicrobial drugs by using the
1170. radiometric Bactec-460 technique and the
Heifets, L.B. and Richmond, J.Y. (2006) Modern proportion method with solid media. Journal of
biological safety standards in tuberculosis Clinical Microbiology 37, 3179–3186.
diagnostic laboratories. In: Anthology of Bio- Piersimoni, C., Olivieri, A., Benacchio, L. and
safety, IX, Chapter 13, pp. 155–166. Scarparo, C. (2006) Current perspectives on
Heifets, L. and Sanchez, T. (2000) New agar drug susceptibility testing of Mycobacterium
medium for testing susceptibility of Myco- tuberculosis complex: the automated non-
bacterium tuberculosis to pyrazinamide. Journal radiometric systems. Journal of Clinical
of Clinical Microbiology 38, 1498–14501. Microbiology 44, 20–28.
Heifets, L. and Sanchez, T. (2003) and (2005) New Richmond, J.Y., Knudsen, R.C. and Good, R.C.
agar medium for mycobacteria (HSTB agar). (1996) Bio-safety in the clinical mycobacteriology
Patents: No US 6,579,694 B2, and No US laboratory. Clinics in Laboratory Medicine
6,951,733 B2. (Clinical Mycobacteriology) 16, 527–550.
Hilleman, D., Rüsch-Gerdes, S. and Richter, E. Roberts, G.D., Goodman, N.L., Heifets, L., Larsh,
(2010) Feasibility of the GenoType MTBDRst H.W., Lindner, T.H., McClatchy, J.K., et al.
assay for fluoroquinolone, amikacin-capreo- (1983) Evaluation of the BACTEC radiometric
mycin, and ethambutol resistance testing of method for recovery mycobacteria and drug
Mycobacterium tuberculosis strains and clinical susceptibility testing of M. tuberculosis acid-fast
specimens. Journal of Clinical Microbiology 47, smear-positive specimens. Journal of Clinical
1767–1772. Microbiology 18, 689–696.
Kent, P.T. and Kubica, G.P. (1985) Public Health Siddiqi, S.H., Libonati, J.P. and Middlebrook, G.
Mycobacteriology: A Guide for the Level III (1981) Evaluation of a rapid radiometric
Laboratory. CDC, Atlanta, Georgia. method for drug susceptibility testing of M.
tuberculosis. Journal of Clinical Microbiology Tortoli, E., Mattei, R., Savarino, A., Bartolini, L. and
13, 908–912. Beer, J. (2000) Comparison of M. tuberculosis
Siddiqi, S.H., Hawkins, J.E. and Laszlo, A. (1985) susceptibility testing performed with BACTEC
Interlaboratory drug susceptibility testing of M. 460-TB (Becton Dickinson) and MB/BacT
tuberculosis by radiometric and two conventional (Organon Teknika) systems. Diagnostic Micro-
methods. Journal of Clinical Microbiology 22, biology and Infectious Diseases 38, 83–88.
919–923. World Health Organization [WHO] (1997) TB Treat-
Tenover, F.C., Crawford, J.T., Heubner, R.E., Geiter, ment Observer. WHO, Geneva, Switzerland.
I.J., Horsburg, C.R. and Good, R.C. (1993). The WHO (2010) Multidrug and Extensively Drug-
resurgence of tuberculosis: is your laboratory resistant TB (M/XDR-TB): 2010 Global Report
ready? Journal of Clinical Microbiology 31, on Surveillance and Response. WHO/HTM/TB
767–770. 2010.3, Geneva, Switzerland.
8 Genotypic Measures of Antibiotic
Susceptibility
Vanessa Mathys,1 Barun Mathema2 and Pablo Bifani3*

1Tuberculosis and Mycobacteria, Communicable and Infectious Diseases, Scientific
Institute of Public Health, Belgium; 2Tuberculosis Center, Public Health Research
Institute, University of Medicine and Dentistry, Newark, New Jersey, USA; 3The
Novartis Institute for Tropical Diseases, Singapore
8.1 Introduction: The Challenge of novo. The need to maintain high drug levels
Drug-resistant Tuberculosis over many months of treatment, combined
with the inherent toxicity of the agents,
Tuberculosis remains the leading cause of results in reduced patient compliance and the
adult mortality attributable to a single in- subsequently higher likelihood of selecting
fectious disease, despite effective chemo- drug-resistant mutants (Davies, 1998). There-
therapy having been available for over 50 fore, in addition to identifying new anti-
years and the development of the tuberculosis tuberculosis agents, the means to shorten the
control strategy of directly observed therapy, length of chemotherapy is paramount, as it
short course (DOTS) (Murray et al., 1991; would impact clinical management greatly
Bloom and Murray, 1992; Raviglione et al., and the emergence of drug resistance.
1995). Tuberculosis chemotherapy is markedly The emergence of increasingly drug-
different from other bacterial pathogens. resistant forms of M. tuberculosis represents a
Mycobacterium tuberculosis has an excep- growing problem that stands to derail much
tionally long generation time and a capacity progress made by global and local tubercu-
for dormancy, when its low metabolic activity losis control programmes. In recent years,
renders it a difficult therapeutic target. substantial advances have been made in
Moreover, M. tuberculosis may reside in understanding mechanisms of action of
pulmonary cavities, solid caseous material or antimycobacterial agents and the biochemical
empyema pus where drug penetration is basis of drug resistance (reviewed in this
inefficient or the pH sufficiently low to inhibit chapter) (Ramaswamy and Musser, 1998;
antibiotic activity (Kaufmann, 2001). Selection Webb and Davies, 1999). In contrast to other
for drug-resistant mutants in the patient pathogenic bacteria, which generally acquire
occurs mainly when patients are treated drug-resistance determinants by horizontal
inappropriately or are exposed to, even transfer, the acquisition of drug resistance in
transiently, subtherapeutic drug levels; con- M. tuberculosis occurs via de novo point
ditions that provide adequate positive selec- mutations, small deletions or insertions in
tion pressure for emergence and, on occasion, specific chromosomal genes. The genetic
maintenance of drug-resistant organisms de basis of drug resistance in M. tuberculosis has

Genotypic Measures of Antibiotic Susceptibility 109
been defined for many of the commonly used pathogens. For instance, in South Africa,
first- and second-line antibiotics (Davies, invasive streptococcal infection that is
1998). Once the bacilli have mutated, selection resistant to levofloxacin has been associated
for drug-resistant mutants occurs in the with a history of tuberculosis treatment
presence of subtherapeutic drug levels or (Gottberg et al., 2008). Likewise, fluoro-
inappropriate therapy. An alarming trend, quinolones that are widely used to treat
and a growing source of public health con- community-acquired bacterial infections
cern, has been the emergence of multiple may also act inadvertently on undiagnosed
drug-resistance tuberculosis (MDR-TB), tuberculosis, in essence by monotherapy and
defined as isolates that are resistant to at least thereby selecting for resistance (Drlica et al.,
isoniazid (INH) and rifampicin (RIF), the two 2008).
most potent anti-tubercular drugs (Iseman In this chapter, we review the genetic
and Madsen, 1989; Vareldzis et al., 1994). The basis of drug resistance in M. tuberculosis and
World Health Organization (WHO) estimates correlates to phenotypic antibiotic suscepti-
that 50 million individuals worldwide bility. We focus on anti-tubercular agents and
harbour MDR M. tuberculosis (WHO, 1996). drug resistance-conferring mutations that are
MDR-TB is associated with high morbidity most relevant in the clinical and epidemiologic
and mortality, prolonged treatment to cure setting.
and an increased risk of spreading infection
with drug-resistant isolates in the community
(Iseman and Madsen, 1989; Bifani et al., 1996). 8.2 Rifampicin
A recent global survey of MDR-TB incidence
estimated that an average of 3.4% of all new Rifampicin (RIF), a semi-synthetic derivative
cases (no documented prior history of of rifamycin B, was first introduced as an
tuberculosis) had MDR-TB (Dye et al., 2002). anti-tubercular in 1966. Together with INH,
Compounding the tuberculosis crisis, RIF comprises the cornerstone of present-day
XDR (extensively drug resistant) M. tubercu- anti-tubercular therapy. RIF is the fastest
losis isolates that are defined as resistant to acting, sterilizing anti-tubercular agent, due
INH, RIF plus a fluoroquinolone and at least in part to its ability to kill semi-dormant
one of the injectable drugs (e.g. kanamycin, bacterial subpopulations and hence active
capreomycin and amikacin) commonly used against both actively growing and slowly
in second-line treatment for MDR-TB have metabolizing M. tuberculosis (Zhang et al.,
emerged. The Centers for Disease Control 2005). The loss of RIF as a therapeutic option
and Prevention (CDC) conducted a global is associated with higher rates of treatment
survey of XDR-TB prevalence in 2000–2004 failure (Espinal et al., 2000). Since the mid-
that estimated 9.9% of all MDR-TB isolates 1990s, RIF resistance (RIFR) has been used as
were additionally resistant to at least two a surrogate marker for multidrug-resistant
classes of second-line anti-tuberculosis agents M. tuberculosis isolates, given that mono-RIFR
(Shah et al., 2007). This report suggested that isolates are rare (Bifani et al., 1996; Ridzon et
XDR-TB might be emerging in many countries al., 1998). Various other rifamycin derivatives
with and without an expanding DOTS have been evaluated for activity and reduced
tuberculosis control programme. While treat- cytotoxicity. Of these, rifapentine and rifa-
ment for MDR-TB has improved greatly butin have been approved by the FDA for the
(mainly in resource-rich settings), outcomes treatment of mycobacterial infections; while
remain far below drug-susceptible counter- others, such as KRM 1648, are under in-
parts; indices that are worse for XDR-TB vestigation. Rifapentine poses improved
(Bifani et al., 1996; Crofton et al., 1997; Iseman, pharmacokinetics accumulating in macro-
2000; CDC, 2003; Gandhi et al., 2006). phages at concentrations 60 times higher than
Additionally, using second-line agents, which in the extracellular fluid as well as longer (4–5
are often broader-spectrum antibiotics, there times) half-life, allowing for intermittent
is a concern of collateral damage in terms of therapy. Based on clinical trials, rifapentine is
selection of resistance among other coexisting now recommended twice a week for the
110 V. Mathys et al.
intensive phase of treatment and once a week to correlate to drug resistance. Often, other
for continuation therapy. Similarly, rifabutin mutations within the RRDR are associated
has a longer half-life than RIF and better in with low levels of RIF resistance and low
vitro activity; however, it has lower in vivo fitness. RIFR spontaneous mutants of M.
activity and increased side effects. Rifabutin tuberculosis and M. bovis BCG selected for in
is currently used for the prevention of dis- vitro mirror the clinical distribution of
seminated M. avium infection, particularly mutations in type and frequency (Morlock et
among HIV/AIDS patients. Both rifapentine al., 2000; Mathys, unpublished). In some
and rifabutin share extensive cross-resistance instances, RIFR isolates encoding two or more
with RIF. mutations within the RRDR have been
RIF inhibits transcription by binding the reported (Bahrmand et al., 2009). It is possible
DNA-dependent RNA-polymerase (RNAPol) that these additional mutations are associated
and suppresses the initial chain formation/ with compensatory mechanism(s) or higher
elongation of the nascent mRNA synthesis. levels of resistance; however, adequate experi-
RIF interferes by binding deep in the DNA– mental work is required before drawing any
RNA channel, both through hydrophobic conclusions. Finally, about 3–5% of all RIFR
side-chain interaction and by hydrogen clinical isolates have mechanisms of resist-
bonding to the key hydroxyl groups of the ance not encoded by the rpoB. Efflux pumps,
RIF and the β-subunit of the RNAPol (rpoβ). mechanisms of ADP ribosylation and proteins
In the process, RIF blocks the elongation of unknown functions have been proposed;
selectively at the initiation site, within the however, the epidemiological significance of
addition of the first 2–5 nucleotides (Campbell these has yet to be determined (Baysarowich
et al., 2001). et al., 2008).
8.2.1 Molecular genetics of rifampicin 8.3 Isoniazid

resistance
Isoniazid (INH, isonicotinic acid hydrazide),
The correlation of RIFR and mutations in rpoB a structural analogue of nicotinamide, is a
were first described in E. coli in 1980 synthetic anti-tubercular agent whose activity
(Ovchinnikov et al., 1981) and in M. tubercu- was first highlighted in 1952 (Bernstein et al.,
losis in 1993 (Honore and Cole, 1993; Telenti et 1952; Fox, 1952) and soon after introduced in
al., 1993). The codon numbering used in M. clinic. INH was first shown to inhibit the
tuberculosis corresponds to the E. coli protein, synthesis of cell wall mycolic acids (Winder
which differs by 81 codons. Numerous studies and Collins, 1970) and later to inhibit the
have shown that >95% of all RIFR isolates catalysis of the synthesis of mycolic acids
encode a mutation within the rifampicin leading to the accumulation of long-chain
resistant-determining region (RRDR), an fatty acids in INH-treated samples (Takayama
81-bp long fragment (codons 507–533). Fur- et al., 1972, 1975). Twenty years later, Zhang
ther, within the RRDR, mutations at His526 and colleagues reported that M. tuberculosis
and Ser531 account for ~80% (~40% each) of all clinical isolates resistant to INH presented
RIFR clinical isolates and Asp516 for a further reduced catalase activity, leading to the
~8%. Noteworthy, substitutions at His526 and identification of mutations within the bifunc-
Ser531 correlate to little to no cost in bacterial tional catalase-peroxidase enzyme (KatG)
fitness (Gagneux et al., 2006b). Single nucleo- (Zhang et al., 1992). Resistance and suscepti-
tide polymorphisms (SNPs) and in-frame bility could be reinstated in a wild-type strain
deletions, insertions in other positions of the through homologous recombination (Zhang
RRDR and sometimes in the N-terminus of et al., 1992, 1993). Subsequent investigations
the rpoB account for ~4% of the remaining revealed that INH was a prodrug merely
RIFR isolates. About ~60 different alterations activated by KatG, while the target of the
have been reported, of which ~40 are within active form of INH was the enoyl-ACP
the RRDR, and few have yet to be confirmed reductase (InhA) (Banerjee et al., 1994). InhA
encodes a protein implicated in the elongation strains encode mutations within the katG
of the fatty-acid chain in mycolic acid gene. SNPs account for most of the mutants;
synthesis in the FAS II (fatty acid synthase) however, in- and out-of-frame deletions,
system (Banerjee et al., 1994). Recombination insertions and terminations have also been
of the inhA mutant Ser94Ala in a wild-type M. observed (Ramaswamy and Musser, 1998;
tuberculosis is sufficient to confer INH Slayden and Barry, 2000). SNPs in katG-Ser315
resistance and inhibition of mycolic acid are by far the most frequently encountered,
synthesis (Vilcheze et al., 2006), and likewise, whereby Ser315Thr alone accounts for ~40–
overexpression of InhA in a susceptible strain 50% of all INHR strains. Mutations on Ser315
results in INHR (Larsen et al., 2002). The have been shown to bypass INH activation
crystallization of InhA in the presence of INH while retaining a 50% functional catalase-
and NADH made it possible to establish that peroxidase activity (Wengenack et al., 1997),
the activated form of INH did not bind to the giving a selective advantage to the organism.
InhA directly (Rozwarski et al., 1998). Instead, Mutations in KatG315 are associated with
a covalently bound INH–NAD complex resistance to >4 μg/ml in most instances,
competes for the active site with NADH, the depending on the substitution. The W-MDR
natural substrate, resulting in the inhibition strain from New York City encodes an
of the synthesis of the mycolic acid α-chain unusual double nucleotide substitution katG-
(Argyrou et al., 2006). Consequently, it is AGC315>ACA315 (Bifani et al., 1996). Several
hypothesized that INH penetrates the cell by other INHR associated mutations localized
passive diffusion, where it is activated by within this same locus of katG-Ser315. Over 80
KatG to form a hypothetical isonicotinoyle different mutations in the katG have been
radical which binds to NAD, resulting in an reported, of which ~60 are within the 300 aa
INH–NAD complex that would inhibit InhA. flanking katG-Ser315; thus allowing for ~80%
These successive steps inhibit mycolic acid detection of all INHR strains by only sequenc-
synthesis, resulting in the accumulation of ing the direct flanks. Out-of-frame deletions,
long-chain fatty acids and cell death insertions and terminations lead to a non-
(Timmins and Deretic, 2006; Vilcheze and functional catalase-peroxidase enzyme associ-
Jacobs, 2007). In addition to KatG and InhA, ated with low bacterial fitness and very
several other less frequent targets have elevated levels of resistance (>256 μg/ml).
been implicated in INHR isolates. Some of Alternatively, mutations in the promoter
the other targets include an NADH- of the mabA-inhA operon occur in 10–25% of
dehydrogenase (ndhII), an arylamine clinical isolates resistant to INH (Ramaswamy
N-acetyltransferase (nhoA) and a ferric and Musser, 1998; Slayden and Barry, 2000).
uptake regulator (furA), to mention a few Mutations -15C>T and -8T>C flanking the ribo-
(Ramaswamy et al., 2003). Although in some somal binding site resulting in upregulation
instances the implication of these alternative of inhA and thus increasing the target protein
targets are supported by extensive experi- are by far the most common. Intragenic
mental data, their prevalence and significance mutations in the InhA remain rare and are
in the clinical setting remain to be explored usually limited to Ser94Ala and a few others.
further. In addition, targets such as the Overall, mutations in InhA are associated
β-ketoacyl-ACP-synthase (KasA) and the with low levels of INHR, <1 μg/ml. Although
dyhydrofolate reductase (DHFR) have been there are numerous studies on the molecular
challenged experimentally and clinically biology of INHR, there exist some important
(Wang et al., 2010). limitations on the available data. Most labora-
tories limit their analysis to the 300 aa flank-
ing katG315 and only examine mabA-inhA or
8.3.1 Molecular genetics of isoniazid the remaining of the katG gene in the absence
resistance of mutations in katG315 and seldom explore
other genes. An inherent limitation of this
Review of the literature indicates that approxi- approach dictated by cost and complexity is
mately 70–80% of all INHR M. tuberculosis that little is known about the frequency,
combinations of and contribution to the (Dover et al., 2004; Frenois et al., 2004). This
resistance of multiple coexisting mutations ligand renders the conformation of EthR
associated with INHR. Furthermore, in con- incompatible with the function of a repressor,
trast to observations on RIFR, the distribution leading to constitutive expression of EthA in
and frequency of INHR spontaneous mutants vivo and hence enhancing the activation of
does not reflect what is observed in the clinic, ETA. Consequently, coadministration of such
hence restricting inferences generated from a ligand–ETA complex could reduce the
experimental data (Bergral et al., 2009; Bifani, therapeutic dose of ETA required and its
unpublished). accompanying side effect (Frenois et al., 2004,
2006).
8.4 Ethionamide
8.4.1 Molecular genetics of ethionamide
Ethionamide (ETA, 2-ethylthioisonicotina- resistance
mide), a structural analogue of INH pre-
scribed since 1960, is today an important Analysis of clinical isolates resistant to ETA
second-line anti-tubercular agent which is as has only been correlated to mutations in
potent as INH (Rist, 1960). ETA, or its either the ethA and/or inhA genes. As for
analogue prothionamide (PTH, 2-propyl- other non-essential genes, mutations are
thioisonicotinamide), administration is limited scattered throughout the ethA and include
for the treatment of MDR-TB, given its signifi- out-of-frame insertions/deletions, truncations
cant side effects, including gastrointestinal associated with high levels of ETAR (DeBarber
and hepatic toxicity. et al., 2000; Morlock et al., 2003; Brossier et al.,
As for INH, ETA is a prodrug requiring 2010; Mathys, unpublished). Our own ana-
activation for inhibitory activity. Recently, a lysis of diverse clinical isolates indicate ~40%
two-gene operon (ethA-ethR) implicated in bear a mutation within ethA, in contrast to
the activation of ETA has been identified others (Brossier et al., 2010; Mathys, unpub-
(Baulard et al., 2000; DeBarber et al., 2000). lished). Cross-resistance with INH has been
The ethA gene encodes a monooxygenase associated with mutations in mabA-inhA
(Baeyer–Villiger flavin-containing), which promoter (primarily -15C>T) and some SNPs
also catalyses the activation of ETA, thiaceta- within the inhA coding region (62I>T, 61A>G,
zone and isoxyl (Fraaĳe et al., 2004; Dover et 130C>T, 280T>G) ( Lee et al., 2000; Morlock et al.,
al., 2007). The difference in the mechanisms of 2003; Guo et al., 2006; Brossier et al., 2010).
activation between INH and ETA explains MabA-inhA display both low and high levels
why KatG mutants remain susceptible to ETA of resistance, probably a consequence of
but not to INH (Morlock et al., 2003). The concurrent mutations elsewhere (Morlock
expression of EthA is under the control of a et al., 2003). Prevalence of InhA-ETA cross-
repressor EthR (Engohang-Ndong et al., resistance varies depending on the study and,
2004). Overexpression of EthA in M. smegmatis once again, interpretation is limited by the
induces a hypersusceptibility to ETA, absence of genotyping data. Further, ETA
whereas the overexpression of EthR induces a susceptibility testing is performed only on
resistance to ETA. EthR controls the expres- MDR isolates. Cross-resistance associated to
sion of EthA, which in turn activates ETA the InhA can also stem from increased levels
(Baulard et al., 2000). The crystallization of of NADH due to mutations in ndhII (NADH-
EthR revealed a homodimer of helical struc- dehydrogenase) (Vilcheze et al., 2005). Other
ture similar to a transcriptional regulator examples of cross-resistance involve EthA, as
belonging to the family of TetR/CamR an activator of thiocarbamide-containing
repressors capable of binding DNA and prodrugs including isoxyl and thiacetazone
ligands (Aramaki et al., 1995). Interestingly, a (Dover et al., 2007). Recently, mshA has been
fortuitous ligand (hexadecyl octanoate, proposed as a potential additional target for
1,4-dioxane) was found to fix each EthR ETA based on extensive genetic evidence
monomer during the process of crystallization (Vilcheze et al., 2008). As for other alternative
GyrApolymorphismsassociatedwithfluoroquinoloneresistance
Ala
74
Ser95
*
STOP8389
Met
1

Ala74ArgSerValAlaGluThr80MetGlyAspTyrHisProHisGly88Asp89Ala90Ser91IleTyrAsp94Ser95
74 80 88 89 90 91 94
Ser ...............................Ala .............................................Ala His Val Pro ...........Gly ..........
........................................Cys80...............................................................Gly 90......................Asn94........
94
............................................................................................................................... ..............Tyr .........
............................................................................................................................... ...............Ala94.......
94
............................................................................................................................... ...............His .......
EthApolymorphismsassociatedwithethionamideresistance

H22P P51L Q269STOP A381P STOP489
D58A R207S FMS270 FMS351

G43C
Y84C Q165P FMS226 G343A C403G
G11A Y84D FMS464
FMS252 S329L
T61M G124D Y386C Y461H
M1R T186K L272P
Y386STOP

ThyApolymorphismsassociatedwithPASresistance

L183V STOP264R

M1 T202A* V261G

L172P
R127L FMS217
G15R FMS72 G91E
Y164STOP Q191R V263G
G91R L118STOP
FMS11 Y153STOP A259P

W83STOP Q111STOP C146R
L143P A182P Y251STOP
Fig. 8.1. Distribution of polymorphisms associated with fluoroquinolones, ethionamide, p-aminosalicylic

acid resistance in the GyrA, EthA and ThyA proteins, respectively. Mutations in essential proteins (e.g.
GyrA) are restricted to key amino acids and comprise few variations. In contrast, polymorphisms in non-
essential proteins are dispersed throughout the corresponding genes (e.g. ethA and thyA) and include
truncations of the proteins. Only selected polymorphisms are shown for EthA and ThyA. FMS =
frameshift; A = alanine; C = cysteine; D = aspartic acid; E = glutamic acid; F = phenylalanine; G =
glycine; H = histadine; I = isoleucine; K = lysine; L = leucine; M = methionine; N = asparagine; P =
proline; Q = glutamine; R = arginine; S = serine; T = threonine; W = tryptophan; V = valine. GyrA-Ser95
and thyA-Thr202 are phylogenetic SNPs.
targets, further evaluation of clinical isolates mutations in the pncA gene, resulting in the
is required, in particular in light of the bacilli’s inability to activate the prodrug
observation that both the W-Beĳing and the (Butler and Kilburn, 1983). As for EthA, PncA
Haarlem strain families encode phylogenetic is a non-essential gene allowing for multiple
mutations not associated with ETAR within possible mutations dispersed throughout the
mshA (Mathys, unpublished). gene and promoter. No evidence has been
found of possible decrease in bacterial fitness
associated with PZA resistance, even in the
8.5 Pyrazinamide event of a truncation, downregulation or
deletion of the gene (Scorpio and Zhang,
Pyrazinamide (PZA) is a structural analogue 1996; Sreevatsan et al., 1997a; Ramaswamy
of the nicotinamide, INH and ETA and, and Musser, 1998). All M. bovis isolates are
likewise, it is a prodrug. None the less, PZA intrinsically resistant to PZA because they
has some distinctive features, including a encode a phylogenetic mutation C169G which
different mechanism of activation and a inactivates pncA (Singh et al., 2006). Mutations
unique sterilizing activity on semi-dormant in the pncA and promoter present an ideal
(persistent) tubercular bacilli (McCune et al., situation for molecular diagnostics through
1966; Mitchison, 1985) and is only active in sequencing, in particular in light of the poor
acidified medium (pH 5.5) (McDermott, reproducibility of phenotypic assays. Never-
1954; Zhang, 1999). Moreover, PZA does not theless, a small number of PZAR strains are
present a bactericidal effect on M. tuberculosis independent of PncA mutations, underlining
in vitro, nor on rapidly growing myco- the need to identify other potential direct or
bacteria. No specific target for pyrazinoic indirect targets.
acid (POA) has been discovered; however, a
model of activation has been proposed
(Zhang and Mitchison, 2003) whereby PZA 8.6 Ethambutol
penetrates the mycobacterial cell wall by
passive diffusion and is converted into POA Ethambutol ((S)-2,2’-(ethylenediimino)-di-1-
by the pyrazinamidase in the cytoplasm. butanol, EMB), as for INH, ETA and PZA, is
POA leaves the cell by either passive an antibiotic specifically active on myco-
diffusion or a possible defective efflux bacteria (Cambau et al., 2003). Side effects
mechanism (Konno et al., 1967; Trivedi and comprising retrobulbair neuritis (Melamud et
Desai, 1987). If the extracellular medium al., 2003) render this drug counter-indicated
presents an acidic pH, POA is protonated to for treatment in children (Trebucq, 1997).
form HPOA, which is reabsorbed by the cell. EMB targets an arabinosyl transferase (EmbB)
As the efflux mechanism of POA is defective, involved in the polymerization of arabinans
POA accumulates in the cytoplasm of the in the cell wall, though its exact mode of
mycobacterium and causes cellular damage. action remains unknown (Mikusova et al.,
Moreover, the influx of protons in the cell, 1995). Inhibition of the enzyme results in the
via the HPOA, acidifies the cytoplasm and termination of arabinogalactan and lipoarabi-
disturbs cellular activity. This model could nomannan synthesis, major polysaccharides
explain the particular characteristics of the of the mycobacterial wall (Takayama and
PZA which is only active in an acidic Kilburn, 1989; Deng et al., 1995). Accumulation
environment. of carbohydrate intermediaries and mycolic
acids has been attributed to the loss of
anchoring scaffold in the cell wall (Mikusova
8.5.1 Molecular genetics of pyrazinamide et al., 1995). EmbB is part of the embA-embB
resistance operon in M. avium and embC-embA-embB in
M. tuberculosis and other mycobactera
Pyrazinamidase is a small protein of 186 aa, (Alcaide et al., 1997). The EmbB proteins of
encoded by the gene pncA. Approximately various mycobacterial species share 65%
70–90% of PZAR cases are associated with identity and are all predicted as being integral
membrane proteins with 12 transmembrane and sparfloxacin) are potent synthetic anti-
regions (Telenti et al., 1997; Ramaswamy and bacterial agents originally derived from
Musser, 1998). nalidixic acid. FQs, currently used as second-
line drugs, are under consideration as a
potential substitute for INH in the first-line
8.6.1 Molecular genetics of ethambutol regiment. This will likely reduce the duration
resistance of therapy to 4 months (Chang et al., 2010;
Nuermberger et al., 2004). Current human
Resistance to EMB in M. tuberculosis correlates trials with moxifloxacin show promising
primarily to mutations in the embB gene results (Rosenthal et al., 2008); however, a
(Sreevatsan et al., 1997b; Telenti et al., 1997). recent report by the WHO describes the
EmbB are found in ~70% of all the EMBR presence of XDR-TB (MDR isolates add-
clinical strains (Ramaswamy et al., 2000) and itionally resistant to an FQ and an amino-
are clustered in the ERDR (ethambutol glycoside or capromycin in 58 countries;
resistance-determining region), likely located WHO, 2009). FQs exert their effects by
in a cytoplasmic loop of the transmembrane inhibiting bacterial topoisomerases II and IV
protein (Ramaswamy and Musser, 1998). The (DNA gyrases). Topoisomerase II facilitates
frequency of SNPs at embB-Met306 range from DNA unwinding and topoisomerase IV
47% to 65%, of which Met306Ile (ATG306>ATA, (notably absent in M. tuberculosis) activates
ATG306>ATC and ATG306>ATT) account for decatenation (Champoux, 2001; Ginsburg et
the majority, followed by ATG306>Leu or al., 2003). Consequently, FQs inhibit DNA
ATG306>Val (Sreevatsan et al., 1997b; Rama- gyrase, a tetramer constituted of 2 A and B
swamy and Musser, 1998). The implication of subunits encoded by gyrA and gyrB (Drlica,
mutations at embB-Met306 on resistance were 1999). The exact mechanisms of action of FQs
first demonstrated through homologous remain elusive; however, strand breakage,
recombination in 1997 (Alcaide et al., 1997), SOS-mediated autolysis and blockage of
but challenged as EMB-susceptible isolates replication by the gyrase–FQ complex may
encoding the same mutations were also enable bacterial inhibition without lethality
reported (Mokrousov et al., 2002). Finally, (Guillemin et al., 1998; Drlica, 1999; Hooper,
analysis of EMBR spontaneous mutants and 2001; Drlica and Malik, 2003; Ginsburg et al.,
allelic exchange reconfirmed that mutations 2003).
in embB-Met306 were sufficient to render 55%
of the isolates EMB. Other mutations within
embB as well as those in the embCAB-operon 8.7.1 Molecular genetics of
have been linked with EMBR, albeit at low fluoroquinolone resistance
levels (Safi et al., 2010; Plinke et al., 2010).
Further, rmlD-rmlA2, involved in the modifi- As for other bacteria, the main mechanism of
cation of rhamnose and the iniBAC promoter FQR in M. tuberculosis results from missense
may also contribute to EMBR (Ramaswamy et mutations in gyrA within the QRDR
al., 2000). In spite of these recent discoveries, (quinolone resistance-determining region)
30% of the EMBR strains do not present any and rarely in gyrB, corresponding to the
polymorphisms within the genes proposed DNA–protein complex (Takiff et al., 1994;
(Ramaswamy et al., 2000; Cambau et al., 2003). Alangaden et al., 1995; Sullivan et al., 1995;
SQ109, a homologue of EMB, is presently Kocagoz et al., 1996; Williams et al., 1996; Xu et
under evaluation. Interestingly, embCAB al., 1996; Perlman et al., 1997). No insertions/
mutants do not confer cross-resistance. deletions, even in-frame, have been reported
in gyrA or gyrB. Instead, 13 possible SNPs in 8
codons within the QRDR account for 42–85%
8.7 Fluoroquinolones of all FQR isolates, of which substitution at
D94 (D94>A94, D94>N94, D94>G94, D94>Y94 and
Fluoroquinolones (FQs: ciprofloxacin, gati- D94>H94) comprise ~60% of all resistant iso-
floxacin, levofloxacin, moxifloxacin, ofloxacin lates (Sullivan et al., 1995; Yew et al., 1995;
Kocagoz et al., 1996). A correlation between 8.8.1 Molecular genetics of para-

specific mutations in the QRDR and levels of aminosalicylic acid resistance
resistance has been described (Kocagoz et al.,
1996; Xu et al., 1996; Zhou et al., 2000). Cross- Recently, Rengarajan and colleagues showed
resistance exists only among the FQs in most by transposon mutagenesis that resistance to
instances (Alangaden et al., 1995; Yew et al., PAS was associated with mutations in the
1995; Ruiz-Serrano et al., 2000). Mutations thyA. Subsequently, it has been shown that
outside QRDR, decreased cell wall permea- approximately 40% of all PASR isolates
bility, active drug efflux pump mechanism, encode a mutation of the thyA, while the
sequestration of drug or drug inactivation are mechanism of resistance remains unknown
some of the hypotheses proposed for non- for the remaining ~60% of clinical isolates
QRDR FQR (Jarlier and Nikaido, 1994; (Mathys et al., 2009). thyA is a non-essential
Cambau and Jarlier, 1996; Liu et al., 1996; gene in M. tuberculosis, given the redundancy
Takiff et al., 1996). of thyX (Myllykallio et al., 2002). As such, as
for other non-essential drug targets in M.
tuberculosis, SNPs, deletions and insertions
8.8 Para-aminosalic acid are diverse and scattered throughout the
target gene.
Para-amino-salicylic acid (PAS) was dis-
covered by Lehmann in 1943 (Lehmann, 1946;
Youmans et al., 1947). In 1951, INH, PAS and 8.9 Aminoglycosides and
SM constituted the birth of triple therapy Macrocyclic Polypeptide Antibiotics
(PAS-SM-INH), used until the mid-1960s.
Although the efficacy of triple therapy proved Streptomycin (SM) was the first amino-
positive, PAS was discontinued due to gastro- glycoside (AG) discovered and drug approved
intestinal toxicity, the need for multiple and for the treatment of tuberculosis following
elevated doses and the advent of more lethal human trials in 1947. AGs (SM, amikacin
anti-tubercular agents. PAS was only reintro- (AK) and kanamycin (KM)) and two macro-
duced in the USA in 1992, following several cyclic polypeptides (viomycin and capreo-
outbreaks of MDR isolates (CDC, 1992). mycin, CAP) share both a common target, the
Currently, a new, less toxic formulation of bacterial ribosome, and require intramuscular
PAS is used as a second-line antibiotic for the administration. AG remains an important
treatment of MDR-TB (WHO, 2000). In spite drug in the tuberculosis therapy arsenal, SM
of this renewed interest, the mechanism of being used regularly in many developing
action of PAS remains elusive. Since its dis- countries, particular for retreatment (WHO
covery, PAS has been assumed to interfere Category II) cases, while other AGs and CAPs
with the folate pathway, given that it presents are used as second line (Peloquin et al., 2004).
structural similarities with the sulfonamides, SM is the least toxic of the three AGs (Peloquin
which are analogues of the p-aminobenzoic et al., 2004), while AK is the most active in the
acid, the substrate of the dihydropteroate mouse efficacy model of tuberculosis (Lounis
synthase (FolP1) (Lehninger, 1977). Recent et al., 1997). AG and macrocyclic polypeptides
experimental and epidemiological data have hinder protein synthesis by inhibiting the
not been able to implicate FolP PASR (Nop- initiation of translation and the translocation
ponpunth et al., 1999; Mathys et al., 2009). reaction, respectively (Benveniste and Davies,
Instead, PAS resistance has been shown to be 1973). Extensive molecular data link SMR to
associated in part, downstream of the folate missense mutations in either the S12 protein
pathway, with the thymidylate synthase (rpsL) or 16S RNA (rrs) genes, which are
(ThyA), but not its analogue ThyX components of the 30S subunits of the
(Rengarajan et al., 2004; Mathys et al., 2009). ribosome. AG tightly binds the 30S ribosomal
ThyA and ThyX are responsible for the subunit, while MCP binds the 50S (Benveniste
methylation of uracil. and Davies, 1973; Ho et al., 1997; Chan, 2003).
Table 8.1. Selected characteristics of drug resistance in M. tuberculosis.

Anti-tuberculars Targets, activating enzymes, cross-resistance and comments
Rifamycin Target: RNA polymerase β-subunit: RpoB (rpoB)
– rifampicin (RIF) FoR associated with the rpoB: >95% within the RRDR, of which His526 ~40%,
– rifabutin Ser531 ~40% and Asp516 ~8%
– rifapentine FoR associated with other targets or outside the RRDR: <5%
Shared cross-resistance within the RRDR for RIF, RFB, RFP and KRM-1648
FoM: 1  108
Isoniazid Activating enzyme: bifunctional enzyme catalase-peroxidase: KatG (katG)
(prodrug) FoR associated with the KatG: ~70–80%
Target: enoyl-ACP reductase: InhA (inhA)
FoR associated with the inhA: 5–25%
KatG: no known cross-resistance
InhA: cross-resistance with isoniazid and ethionamide
Cross-resistance with isoniazid and thiacetazone reported
FoM: 1  106
Ethionamide Activating enzyme: monooxygenase: EthA-EthR operon (ethA)
(prodrug) FoR associated with ethA mutations: ~40%
Cross-resistance (EthA): ETA with isoxyl and thiacetazone (THI)
Target: enoyl-ACP reductase InhA (inhA)
FoR (InhA): 5–25%
Cross-resistance (InhA): with isoniazid and ethionamide
FoM: 1  106
Pyrazinamide Activating enzyme: pyrazinamidase: PncA (pncA)
(prodrug) FoR associated with the PncA: ~70–90%
Active against slow metabolism or dormant bacilli
FoM: 1  105
Ethambutol Targets: arabinosyl transferase operon embCAB
FoR (embB): ~70%, of which embB-Met306 ~47–65%
FoM: 1  106
Fluoroquinolones Target: DNA topoisomerase: GyrA and GyrB subunits (gyrA and gyrB)
– ciprofloxacin (CIP) Cross-resistance: GyrA and GyrB: CIP, GAT, LEV, MOX, OFL, SPA
– gatifloxacin (GAT) FoR(GyrA): 80–95% within the QDDR
– levofloxacin (LEVO) FoR(GyrB): 10–15%
– moxifloxacin (MOX) FoM: 1  107–108
– ofloxacin (OFLOX)
– sparfloxacin (SPA)
p-Aminosalicylic acid
Targets: thymidylate synthase: ThyA (thyA)
(possible prodrug)Some PASR ThyA mutants share cross-resistance with fluorouracil (FU), but
FU-resistant mutants are not resistant to PAS
High levels PASR share cross-resistance with THI; but THIR mutants are not
resistant to PAS
FoR(ThyA): ~40%
FoM: 1  105
Aminoglycosides and Targets: SM: S12 ribosomal proteins (rpsL); SM, AMI and KAN: 16S rRNA (rrs);
macrocyclic CAP: methyltransferase (tlyA)
peptides FoR(rpsL): ~65%; primarily Lys43 and Lys88
– streptomycin (SM) No cross-resistance associated with rpsL-Lys43 or Lys88
– kanamycin (KM) FoR (rrs): ~10%
– amikacin (AMI) Mutation rrs1401 associated with high levels of cross-resistance with AMI, CAP,
– capreomycin (CAP) KM and VIO
– viomycin (VIO)
Note: FoR = frequency of resistance; FoM: = frequency of mutation.

8.9.1 Molecular genetics of the problem is the long treatment required

aminoglycosides and macrocyclic to cure tuberculosis with inherent drug-
resistance associated toxicity hampering therapeutic
adherence and thus increasing the likelihood
Mutations at either Lys43>Arg(Thr) or of selecting and maintaining drug-resistant
Lys88>Arg account for ~65% of all STRR iso- mutants.
lates. While SNPs at positions 43 and 88 are Drug resistant-conferring mutations
associated with elevated levels of resistance, among clinical and laboratory isolates are
other rpsL mutations are phylogenetic and being used increasingly to study bacterial
share no correlation to SMR (Sreevatsan et al., fitness (Gagneux et al., 2006a,b). These studies
1996). In contrast, mutations within the rrs have indicated, not surprisingly, a greater
have been associated with low levels of SMR diversity of mutations among in vitro-
but high levels of resistance to AMI and KM generated spontaneous mutants compared to
(Alangaden et al., 1998). All together, SNPs in those commonly observed from pheno-
rpsL and rrs account for only ~75% of all SMR typically drug-resistant patient isolates. These
isolates, and no cross-resistance has been observations suggest a strong relationship
reported with either AK, KM or the macro- between particular mutants and their
cyclic polypeptides in mutants with SNPs at inherent fitness, facilitating a progressive in
position 43 or 88. Cross-resistance with other vivo infection. Complicating the analysis of
class members (KM and AK) and the macro- bacterial fitness among drug-resistant M.
cycle polypeptide CAP and viomycin has tuberculosis is the increasingly immuno-
been correlated to mutation rrs1401. Other compromised host population (e.g. HIV),
cross-resistances have not always been com- where less fit organisms proliferate unchal-
plete or reciprocal and depend largely on the lenged and possibly allow acquiring com-
mutation (Ho et al., 1997). CAPR and vio- pensatory alterations to regain fitness (Bifani
mycinR can result from mutations in the et al., 2008).
2-O-methyltransferase (tlyA) and the rrs; Recently, much progress has been made
though these do not account for all resistant in rapid diagnostics of drug resistance, par-
strains and the full clinical implications have ticularly first-line agents. Rapid methods
yet to be determined. typically rely on PCR or other molecular
techniques that target hotspot regions. While
it is not ideal to study some drug resistance-
8.10 Conclusion conferring genomic regions, the majority of
mutations in the clinical setting are restricted,
Despite progress made in elucidating the facilitating high-throughput rapid analysis.
underlying mechanism that confers drug For instance, genotypic (i.e. 81-bp RRDR)
resistance in M. tuberculosis, genotypic correlates for RIFR are upwards of 95%, mak-
measures of antibiotic susceptibility remain ing it highly amenable for rapid molecular-
deceptively simple. This, in part, has to do based testing (Boehme et al.). This is also the
with a confluence of factors, including the case with INHR, where the majority of
lifestyle of M. tuberculosis. That is, the bacilli’s mutations are restricted to a few sites in KatG
long doubling time, subversion of the host and mabA-inhA (Lacoma et al., 2008). How-
immune response, residence in macrophages ever, interpretation must bear in mind the
and in anatomical sites (e.g. cavities) where possibility of heteroresistant subpopulations
bacterial populations are dense, thereby that collectively may give false positive
rendering drug penetration inefficient and results.
allowing for selection of drug-resistant In this chapter, we have focused on anti-
mutants. In addition to de novo acquisition of tubercular agents and the corresponding
drug resistance (i.e. acquired resistance), the genotypic measures that are of primary
mainly aerosol transmission route enables clinical and epidemiologic importance. A
rapid dissemination of already resistant organ- number of other mutations, some confirmed
isms (i.e. primary resistance). Compounding experimentally, have been described in the
literature and have been deposited recently Griffiths, E., Zhang, K., et al. (2008) Rifamycin
on a website (www.tbdreamdb.com). Of note antibiotic resistance by ADP-ribosylation:
are mutations in the regions discussed above structure and diversity of Arr. Proceedings of the
that have not been confirmed experimentally. National Academy of Sciences 105, 4886–4891.
Benveniste, R. and Davies, J. (1973) Mechanisms
Paramount is the experimental confirmation
of antibiotic resistance in bacteria. Annual
of these genetic alterations, as some are Review of Biochemistry 42, 471–506.
phylogenetic scars or do not confer pheno- Bernstein, J., Lott, W.A., Steinberg, B.A. and Yale,
typic resistance (Motiwala et al., 2010; H.L. (1952) Chemotherapy of experimental
Feuerriegel et al., 2011). tuberculosis. V. Isonicotinic acid hydrazide
(nydrazid) and related compounds. American
Reviews of Tuberculosis 65, 357–364.
References Bifani, P.J., Plikaytis, B.B., Kapur, V., Stockbauer,
K., Pan, X., Lutfey, M.L., et al. (1996) Origin and
Alangaden, G.J., Manavathu, E.K., Vakulenko, interstate spread of a New York City multidrug-
S.B., Zvonok, N.M. and Lerner, S.A. (1995) resistant Mycobacterium tuberculosis clone
Characterization of fluoroquinolone-resistant family. Journal of the American Medical
mutant strains of Mycobacterium tuberculosis Association 275, 452–457.
selected in the laboratory and isolated from Bifani, P., Mathema, B., Kurepina, N., Shashkina,
patients. Antimicrobial Agents and Chemo- E., Bertout, J., Blanchis, A.S., et al. (2008) The
therapy 39, 1700–1703. evolution of drug resistance in Mycobacterium
Alangaden, G.J., Kreiswirth, B.N., Aouad, A., tuberculosis: from a mono-rifampin-resistant
Khetarpal, M., Igno, F.R., Moghazeh, S.L., et al. cluster into increasingly multidrug-resistant
(1998) Mechanism of resistance to amikacin variants in an HIV-seropositive population.
and kanamycin in Mycobacterium tuberculosis. Journal of Infectious Diseases 198, 90–94.
Antimicrobial Agents and Chemotherapy 42, Bloom, B.R. and Murray, C.J. (1992) Tuberculosis:
1295–1297. commentary on a re-emergent killer. Science
Alcaide, F., Pfyffer, G.E. and Telenti, A. (1997) Role 257, 1055–1064.
of embB in natural and acquired resistance to Boehme, C.C., Nabeta, P., Hillemann, D., Nicol,
ethambutol in mycobacteria. Antimicrobial M.P., Shenai, S., Krapp, F., et al. (2010) Rapid
Agents and Chemotherapy 41, 2270–2273. molecular detection of tuberculosis and rifampin
Aramaki, H., Yagi, N. and Suzuki, M. (1995) resistance. New England Journal of Medicine
Residues important for the function of a 363, 1005–1015.
multihelical DNA binding domain in the new Brossier, F., Veziris, N., Truffot-Pernot, C., Jarlier, V.
transcription factor family of Cam and Tet and Sougakoff, W. (2011) Molecular investi-
repressors. Protein Engineering 8, 1259–1266. gation of resistance to the antituberculous drug
Argyrou, A., Jin, L., Siconilfi-Baez, L., Angeletti, ethionamide in multidrug-resistant clinical iso-
R.H. and Blanchard, J.S. (2006) Proteome-wide lates of Mycobacterium tuberculosis. Anti-
profiling of isoniazid targets in Mycobacterium microbial Agents and Chemotherapy 55,
tuberculosis. Biochemistry 45, 13947–13953. 355–360.
Bahrmand, A.R., Titov, L.P., Tasbiti, A.H., Yari, S. Butler, W.R. and Kilburn, J.O. (1983) Susceptibility
and Graviss, E.A. (2009) High-level rifampin of Mycobacterium tuberculosis to pyrazinamide
resistance correlates with multiple mutations in and its relationship to pyrazinamidase activity.
the rpoB gene of pulmonary tuberculosis Antimicrobial Agents and Chemotherapy 24,
isolates from the Afghanistan border of Iran. 600–601.
Journal of Clinical Microbiology 47, 2744–2750. Cambau, E. and Jarlier, V. (1996) Resistance to
Banerjee, A., Dubnau, E., Quemard, A., Bala- quinolones in mycobacteria. Research in Micro-
subramanian, V., Um, K.S., Wilson, T., et al. biology 147, 52–59.
(1994) inhA, a gene encoding a target for Cambau, E., Lemaitre, N., Sougakoff, W. and
isoniazid and ethionamide in Mycobacterium Jarlier, V. (2003) Résistance aux antituberculeux.
tuberculosis. Science 263, 227–230. Antibiotiques 5, 29–37.
Baulard, A.R., Betts, J.C., Engohang-Ndong, J., Campbell, E.A., Korzheva, N., Mustaev, A.,
Quan, S., McAdam, R.A., Brennan, P.J., et al. Murakami, K., Nair, S., Goldfarb, A., et al. (2001)
(2000) Activation of the pro-drug ethionamide is Structural mechanism for rifampicin inhibition of
regulated in mycobacteria. Journal of Biological bacterial rna polymerase. Cell 104, 901–912.
Chemistry 275, 28326–28331. CDC (1992) Update: availability of streptomycin
Baysarowich, J., Koteva, K., Hughes, D.W., Ejim, L., and para-aminosalicylic acid – United States.
Morbidity and Mortality Weekly Report (MMWR) Drlica, K., Zhao, X. and Kreiswirth, B. (2008)
41, 482. Minimising moxifloxacin resistance with
CDC (2003) Treatment of tuberculosis. Morbidity tuberculosis. The Lancet Infectious Diseases 8,
and Mortality Weekly Report (MMWR). June 20 273–275.
2003, Vol 52, No RR-11. Dye, C., Espinal, M.A., Watt, C.J., Mbiaga, C. and
Champoux, J.J. (2001) DNA topoisomerases: Williams, B.G. (2002) Worldwide incidence of
structure, function, and mechanism. Annual multidrug-resistant tuberculosis. Journal of
Review of Biochemistry 70, 369–413. Infectious Diseases 185, 1197–1202.
Chan, E. (2003) Pyrazinamide, ethambutol, ethiona- Engohang-Ndong, J., Baillat, D., Aumercier, M.,
mide, and aminoglycosides. In: Rom, W. and Bellefontaine, F., Besra, G.S., Locht, C., et al.
Garay, S. (eds) Tuberculosis, 2nd edition. (2004) EthR, a repressor of the TetR/CamR
Lippincott Williams and Wilkins, Philadelphia, family implicated in ethionamide resistance in
Pennsylvania. mycobacteria, octamerizes cooperatively on its
Chang, K.C., Leung, C.C., Yew, W.W., Lau, T.Y., operator. Molecular Microbiology 51, 175–188.
Leung, W.M., Tam, C.M., et al. (2010) Newer Espinal, M.A., Kim, S.J., Suarez, P.G., Kam, K.M.,
fluoroquinolones for treating respiratory Khomenko, A.G., Migliori, G.B., et al. (2000)
infection: do they mask tuberculosis? European Standard short-course chemotherapy for drug-
Respiratory Journal 35, 606–613. resistant tuberculosis: treatment outcomes in 6
Crofton, J., Chaulet, P., Maher, D., Grosset, J., countries. Journal of the American Medical
Harris, W., Horne, N., et al. (1997) Guidelines Association 283, 2537–2545.
for the management of drug-resistant tubercu- Feuerriegel, S., Koser, C., Trube, L., Archer, J.,
losis. WHO/TB/96.210. Rusch Gerdes, S., Richter, E., et al. (2010)
Davies, J. (1998) Antibiotic resistance in myco- Thr202Ala in thyA is a marker for the Latin
bacteria. In: Chadwick, D.J. and Cardew, G. American Mediterranean lineage of the Myco-
(eds) Genetics and Tuberculosis. John Wiley, bacterium tuberculosis complex rather than
Chichester, UK. para-aminosalicylic acid resistance. Anti-
DeBarber, A.E., Mdluli, K., Bosman, M., Bekker, microbial Agents and Chemotherapy 54, 4794–
L.G. and Barry, C.E. 3rd (2000) Ethionamide 4798.
activation and sensitivity in multidrug-resistant Fox, H.H. (1952) The chemical approach to the
Mycobacterium tuberculosis. Proceedings of control of tuberculosis. Science 116, 129–134.
the National Academy of Sciences 97, 9677– Fraaije, M.W., Kamerbeek, N.M., Heidekamp, A.J.,
9682. Fortin, R. and Janssen, D.B. (2004) The prodrug
Deng, L., Mikusova, K., Robuck, K.G., Scherman, activator EtaA from Mycobacterium tuberculosis
M., Brennan, P.J. and McNeil, M.R. (1995) is a Baeyer–Villiger monooxygenase. Journal of
Recognition of multiple effects of ethambutol on Biological Chemistry 279, 3354–3360.
metabolism of mycobacterial cell envelope. Frenois, F., Engohang-Ndong, J., Locht, C.,
Antimicrobial Agents and Chemotherapy 39, Baulard, A.R. and Villeret, V. (2004) Structure of
694–701. EthR in a ligand bound conformation reveals
Dover, L.G., Corsino, P.E., Daniels, I.R., Cocklin, therapeutic perspectives against tuberculosis.
S.L., Tatituri, V., Besra, G.S., et al. (2004) Molecular Cell 16, 301–307.
Crystal structure of the TetR/CamR family Frenois, F., Baulard, A.R. and Villeret, V. (2006)
repressor Mycobacterium tuberculosis EthR Insights into mechanisms of induction and
implicated in ethionamide resistance. Journal of ligands recognition in the transcriptional
Molecular Biology 340, 1095–1105. repressor EthR from Mycobacterium tubercu-
Dover, L.G., Alahari, A., Gratraud, P., Gomes, J.M., losis. Tuberculosis (Edinburgh) 86, 110–1104.
Bhowruth, V., Reynolds, R.C., et al. (2007) Gagneux, S., Burgos, M.V., Deriemer, K., Encisco,
EthA, a common activator of thiocarbamide- A., Munoz, S., Hopewell, P.C., et al. (2006a)
containing drugs acting on different myco- Impact of bacterial genetics on the transmission
bacterial targets. Antimicrobial Agents and of isoniazid-resistant Mycobacterium tubercu-
Chemotherapy 51, 1055–1063. losis. PLoS Pathogens 2, e61.
Drlica, K. (1999) Mechanism of fluoroquinolone Gagneux, S., Long, C.D., Small, P.M., Van, T.,
action. Current Opinion in Microbiology 2, 504– Schoolnik, G.K. and Bohannan, B.J. (2006b)
508. The competitive cost of antibiotic resistance in
Drlica, K. and Malik, M. (2003) Fluoroquinolones: Mycobacterium tuberculosis. Science 312,
action and resistance. Current Topics in 1944–1946.
Medicinal Chemistry 3, 249–282. Gandhi, N.R., Moll, A., Sturm, A.W., Pawinski, R.,
Govender, T., Lalloo, U., et al. (2006) Extensively Konno, K., Feldmann, F.M. and McDermott, W.
drug-resistant tuberculosis as a cause of death (1967) Pyrazinamide susceptibility and amidase
in patients co-infected with tuberculosis and activity of tubercle bacilli. American Reviews in
HIV in a rural area of South Africa. The Lancet Respiratory Disease 95, 461–469.
368, 1575–1580. Lacoma, A., Garcia-Sierra, N., Prat, C., Ruiz-
Ginsburg, A.S., Grosset, J.H. and Bishai, W.R. Manzano, J., Haba, L., Roses, S., et al. (2008)
(2003) Fluoroquinolones, tuberculosis, and GenoType MTBDRplus assay for molecular
resistance. The Lancet Infectious Diseases 3, detection of rifampin and isoniazid resistance in
432–442. Mycobacterium tuberculosis strains and clinical
Gottberg, K., Einarsson, U., Ytterberg, C., samples. Journal of Clinical Microbiology 46,
Fredrikson, S., Von Koch, L. and Holmqvist, 3660–3667.
L.W. (2008) Use of health care services and Larsen, M.H., Vilcheze, C., Kremer, L., Besra, G.S.,
satisfaction with care in people with multiple Parsons, L., Salfinger, M., et al. (2002) Over-
sclerosis in Stockholm county: a population- expression of inhA, but not kasA, confers resist-
based study. Multiple Sclerosis 14, 962–971. ance to isoniazid and ethionamide in
Guillemin, I., Jarlier, V. and Cambau, E. (1998) Mycobacterium smegmatis, M. bovis BCG and
Correlation between quinolone susceptibility M. tuberculosis. Molecular Microbiology 46,
patterns and sequences in the A and B subunits 453–466.
of DNA gyrase in mycobacteria. Antimicrobial Lee, H., Cho, S.N., Bang, H.E., Lee, J.H., Bai, G.H.,
Agents and Chemotherapy 42, 2084–2088. Kim, S.J., et al. (2000) Exclusive mutations
Guo, H., Seet, Q., Denkin, S., Parsons, L. and related to isoniazid and ethionamide resistance
Zhang, Y. (2006) Molecular characterization of among Mycobacterium tuberculosis isolates
isoniazid-resistant clinical isolates of from Korea. International Journal of Tuberculosis
Mycobacterium tuberculosis from the USA. and Lung Disease 4, 441–447.
Journal of Medical Microbiology 55, 1527–1531. Lehmann, J. (1946) Para-aminosalicylic acid in the
Ho, Y.I., Chan, C.Y. and Cheng, A.F. (1997) In-vitro treatment of tuberculosis. The Lancet 247,
activities of aminoglycoside-aminocyclitols 15–16.
against mycobacteria. Journal of Antimicrobial Lehninger, A.L. (1977) Biochemistry: The Molecular
Chemotherapy 40, 27–32. Basis of Cell Structure and Function. Worth
Honore, N. and Cole, S.T. (1993) Molecular basis of Publishers, Inc, New York.
rifampin resistance in Mycobacterium leprae. Liu, J., Takiff, H.E. and Nikaido, H. (1996) Active
Antimicrobial Agents and Chemotherapy 37, efflux of fluoroquinolones in Mycobacterium
414–418. smegmatis mediated by LfrA, a multidrug efflux
Hooper, D.C. (2001) Mechanisms of action of pump. Journal of Bacteriology 178, 3791–3795.
antimicrobials: focus on fluoroquinolones. Lounis, N., Ji, B., Truffot-Pernot, C. and Grosset, J.
Clinical Infectious Diseases 32 (Suppl 1), S9– (1997) Comparative activities of amikacin
S15. against Mycobacterium avium complex in nude
Iseman, M. (2000) Drug-Resistant Tuberculosis. A and beige mice. Antimicrobial Agents and
Clinican’s Guide to Tuberculosis. Lippincott Chemotherapy 41, 1168–1169.
Williams and Wilkins, Philadelphia, Pennsyl- McCune, R.M., Feldmann, F.M. and McDermott, W.
vania. (1966) Microbial persistence. II. Characteristics
Iseman, M.D. and Madsen, L.A. (1989) Drug- of the sterile state of tubercle bacilli. Journal of
resistant tuberculosis. Clinics in Chest Medicine Experimental Medicine 123, 469–486.
10, 341–353. Mathys, V., Wintjens, R., Lefevre, P., Bertout, J.,
Jarlier, V. and Nikaido, H. (1994) Mycobacterial cell Singhal, A., Kiass, M., et al. (2009) Molecular
wall: structure and role in natural resistance to genetics of para-aminosalicylic acid resistance
antibiotics. FEMS Microbiology Letters 123, in clinical isolates and spontaneous mutants of
11–18. Mycobacterium tuberculosis. Antimicrobial
Kaufmann, S.H. (2001) How can immunology Agents and Chemotherapy 53, 2100–2109.
contribute to the control of tuberculosis? Nature Melamud, A., Kosmorsky, G.S. and Lee, M.S.
Reviews in Immunology 1, 20–30. (2003) Ocular ethambutol toxicity. Mayo Clinic
Kocagoz, T., Hackbarth, C.J., Unsal, I., Rosenberg, Proceedings 78, 1409–1411.
E.Y., Nikaido, H. and Chambers, H.F. (1996) Mikusova, K., Slayden, R.A., Besra, G.S. and
Gyrase mutations in laboratory-selected, fluoro- Brennan, P.J. (1995) Biogenesis of the
quinolone-resistant mutants of Mycobacterium mycobacterial cell wall and the site of action of
tuberculosis H37Ra. Antimicrobial Agents and ethambutol. Antimicrobial Agents and Chemo-
Chemotherapy 40, 1768–1774. therapy 39, 2484–2489.
Mitchison, D.A. (1985) The action of antituberculosis Peloquin, C.A., Berning, S.E., Nitta, A.T., Simone,
drugs in short-course chemotherapy. Tubercle P.M., Goble, M., Huitt, G.A., et al. (2004)
66, 219–225. Aminoglycoside toxicity: daily versus thrice-
Mokrousov, I., Otten, T., Vyshnevskiy, B. and weekly dosing for treatment of mycobacterial
Narvskaya, O. (2002) Detection of embB306 diseases. Clinical Infectious Diseases 38,
mutations in ethambutol-susceptible clinical 1538–1544.
isolates of Mycobacterium tuberculosis from Perlman, D.C., El Sadr, W.M., Heifets, L.B., Nelson,
Northwestern Russia: implications for genotypic E.T., Matts, J.P., Chirgwin, K., et al. (1997)
resistance testing. Journal of Clinical Susceptibility to levofloxacin of Myocobacterium
Microbiology 40, 3810–3813. tuberculosis isolates from patients with HIV-
Morlock, G.P., Plikaytis, B.B. and Crawford, J.T. related tuberculosis and characterization of a
(2000) Characterization of spontaneous, in strain with levofloxacin monoresistance. Com-
vitro-selected, rifampin-resistant mutants of munity Programs for Clinical Research on AIDS
Mycobacterium tuberculosis strain H37Rv. Anti- 019 and the AIDS Clinical Trials Group 222
microbial Agents and Chemotherapy 44, 3298– Protocol Team. AIDS 11, 1473–1478.
3301. Plinke, C., Cox, H.S., Zarkua, N., Karimovich, H.A.,
Morlock, G.P., Metchock, B., Sikes, D., Crawford, Braker, K., Diel, R., et al. (2010) embCAB
J.T. and Cooksey, R.C. (2003) ethA, inhA, and sequence variation among ethambutol-resistant
katG loci of ethionamide-resistant clinical Mycobacterium tuberculosis isolates without
Mycobacterium tuberculosis isolates. Anti- embB306 mutation. Journal of Antimicrobial
microbial Agents and Chemotherapy 47, 3799– Chemotherapy 65, 1359–1367.
3805. Ramaswamy, S. and Musser, J.M. (1998) Molecular
Motiwala, A.S., Dai, Y., Jones-Lopez, E.C., Hwang, genetic basis of antimicrobial agent resistance
S.H., Lee, J.S., Cho, S.N., et al. (2010) in Mycobacterium tuberculosis: 1998 update.
Mutations in extensively drug-resistant Tuberculosis and Lung Disease 79, 3–29.
Mycobacterium tuberculosis that do not code Ramaswamy, S.V., Amin, A.G., Goksel, S., Stager,
for known drug-resistance mechanisms. Journal C.E., Dou, S.J., El Sahly, H., et al. (2000)
of Infectious Diseases 201, 881–888. Molecular genetic analysis of nucleotide
Murray, C.J., Dejonghe, E., Chum, H.J., Nyangulu, polymorphisms associated with ethambutol
D.S., Salomao, A. and Styblo, K. (1991) Cost resistance in human isolates of Mycobacterium
effectiveness of chemotherapy for pulmonary tuberculosis. Antimicrobial Agents and Chemo-
tuberculosis in three sub-Saharan African therapy 44, 326–336.
countries. The Lancet 338, 1305–1308. Ramaswamy, S.V., Reich, R., Dou, S.J., Jasperse,
Myllykallio, H., Lipowski, G., Leduc, D., Filee, J., L., Pan, X., Wanger, A., et al. (2003) Single
Forterre, P. and Liebl, U. (2002) An alternative nucleotide polymorphisms in genes associated
flavin-dependent mechanism for thymidylate with isoniazid resistance in Mycobacterium
synthesis. Science 297, 105–107. tuberculosis. Antimicrobial Agents and Chemo-
Nopponpunth, V., Sirawaraporn, W., Greene, P.J. therapy 47, 1241–1250.
and Santi, D.V. (1999) Cloning and expression Raviglione, M.C., Snider, D.E. Jr and Kochi, A.
of Mycobacterium tuberculosis and Myco- (1995) Global epidemiology of tuberculosis.
bacterium leprae dihydropteroate synthase in Morbidity and mortality of a worldwide epidemic.
Escherichia coli. Journal of Bacteriology 181, Journal of the American Medical Association
6814–6821. 273, 220–226.
Nuermberger, E.L., Yoshimatsu, T., Tyagi, S., Rengarajan, J., Sassetti, C.M., Naroditskaya, V.,
Williams, K., Rosenthal, I., O’Brien, R.J., et al. Sloutsky, A., Bloom, B.R. and Rubin, E.J. (2004)
(2004) Moxifloxacin-containing regimens of The folate pathway is a target for resistance to
reduced duration produce a stable cure in the drug para-aminosalicylic acid (PAS) in
murine tuberculosis. American Journal of mycobacteria. Molecular Microbiology 53, 275–
Respiratory and Critical Care Medicine 170, 282.
1131–1134. Ridzon, R., Whitney, C.G., McKenna, M.T., Taylor,
Ovchinnikov Yu, A., Monastyrskaya, G.S., Gubanov, J.P., Ashkar, S.H., Nitta, A.T., et al. (1998) Risk
V.V., Lipkin, V.M., Sverdlov, E.D., Kiver, I.F., et al. factors for rifampin mono-resistant tuberculosis.
(1981) Primary structure of Escherichia coli American Journal of Respiratory and Critical
RNA polymerase nucleotide substitution in the Care Medicine 157, 1881–1884.
beta subunit gene of the rifampicin resistant Rist, N. (1960) L’activite antituberculeuse de
rpoB255 mutant. Molecular and General l’ethionamide. Advances in Tuberculosis
Genetics 184, 536–538. Research 10, 69–126.
Rosenthal, I.M., Zhang, M., Almeida, D., Grosset, Sreevatsan, S., Stockbauer, K.E., Pan, X.,
J.H. and Nuermberger, E.L. (2008) Isoniazid or Kreiswirth, B.N., Moghazeh, S.L., Jacobs, W.R.
moxifloxacin in rifapentine-based regimens for Jr., et al. (1997b) Ethambutol resistance in
experimental tuberculosis? American Journal of Mycobacterium tuberculosis: critical role of
Respiratory and Critical Care Medicine 178, embB mutations. Antimicrobial Agents and
989–993. Chemotherapy 41, 1677–1681.
Rozwarski, D.A., Grant, G.A., Barton, D.H., Jacobs, Sullivan, E.A., Kreiswirth, B.N., Palumbo, L., Kapur,
W.R. Jr and Sacchettini, J.C. (1998) Modification V., Musser, J.M., Ebrahimzadeh, A., et al. (1995)
of the NADH of the isoniazid target (InhA) from Emergence of fluoroquinolone-resistant tubercu-
Mycobacterium tuberculosis. Science 279, losis in New York City. The Lancet 345, 1148–
98–102. 1150.
Ruiz-Serrano, M.J., Alcala, L., Martinez, L., Diaz, Takayama, K. and Kilburn, J.O. (1989) Inhibition of
M., Marin, M., et al. (2000) In vitro activities of synthesis of arabinogalactan by ethambutol in
six fluoroquinolones against 250 clinical isolates Mycobacterium smegmatis. Antimicrobial
of Mycobacterium tuberculosis susceptible or Agents and Chemotherapy 33, 1493–1499.
resistant to first-line antituberculosis drugs. Takayama, K., Wang, L. and David, H.L. (1972)
Antimicrobial Agents and Chemotherapy 44, Effect of isoniazid on the in vivo mycolic acid
2567–2568. synthesis, cell growth, and viability of Myco-
Safi, H., Fleischmann, R.D., Peterson, S.N., Jones, bacterium tuberculosis. Antimicrobial Agents
M.B., Jarrahi, B. and Alland, D. (2010) Allelic and Chemotherapy 2, 29–35.
exchange and mutant selection demonstrate Takayama, K., Schnoes, H.K., Armstrong, E.L. and
that common clinical embCAB gene mutations Boyle, R.W. (1975) Site of inhibitory action of
only modestly increase resistance to ethambutol
isoniazid in the synthesis of mycolic acids in
in Mycobacterium tuberculosis. Antimicrobial
Mycobacterium tuberculosis. Journal of Lipid
Agents and Chemotherapy 54, 103–108.
Research 16, 308–317.
Scorpio, A. and Zhang, Y. (1996) Mutations in pncA,
Takiff, H.E., Salazar, L., Guerrero, C., Philipp, W.,
a gene encoding pyrazinamidase/nicotinami-
Huang, W.M., Kreiswirth, B., et al. (1994)
dase, cause resistance to the antituberculous
Cloning and nucleotide sequence of Myco-
drug pyrazinamide in tubercle bacillus. Nature
bacterium tuberculosis gyrA and gyrB genes
Medicine 2, 662–667.
and detection of quinolone resistance mutations.
Shah, N.S., Wright, A., Bai, G.H., Barrera, L.,
Antimicrobial Agents and Chemotherapy 38,
Boulahbal, F., Martin-Casabona, N., et al.
773–780.
(2007) Worldwide emergence of extensively
drug-resistant tuberculosis. Emerging Infectious Takiff, H.E., Cimino, M., Musso, M.C., Weisbrod, T.,
Diseases 13, 380–387. Martinez, R., Delgado, M.B., et al. (1996) Efflux
Singh, P., Mishra, A.K., Malonia, S.K., Chauhan, pump of the proton antiporter family confers
D.S., Sharma, V.D., Venkatesan, K., et al. (2006) low-level fluoroquinolone resistance in Myco-
The paradox of pyrazinamide: an update on the bacterium smegmatis. Proceedings of the
molecular mechanisms of pyrazinamide resist- National Academy of Sciences 93, 362–366.
ance in mycobacteria. Journal of Communicable Telenti, A., Imboden, P., Marchesi, F., Lowrie, D.,
Disease 38, 288–298. Cole, S., Colston, M.J., et al. (1993) Detection of
Slayden, R.A. and Barry, C.E. 3rd (2000) The rifampicin-resistance mutations in Myco-
genetics and biochemistry of isoniazid resist- bacterium tuberculosis. The Lancet 341, 647–
ance in Mycobacterium tuberculosis. Microbes 650.
and Infection 2, 659–669. Telenti, A., Philipp, W.J., Sreevatsan, S.,
Sreevatsan, S., Pan, X., Stockbauer, K.E., Williams, Bernasconi, C., Stockbauer, K.E., Wieles, B., et
D.L., Kreiswirth, B.N. and Musser, J.M. (1996) al. (1997) The emb operon, a gene cluster of
Characterization of rpsL and rrs mutations in Mycobacterium tuberculosis involved in
streptomycin-resistant Mycobacterium tubercu- resistance to ethambutol. Nature Medicine 3,
losis isolates from diverse geographic localities. 567–570.
Antimicrobial Agents and Chemotherapy 40, Timmins, G.S. and Deretic, V. (2006) Mechanisms
1024–1026. of action of isoniazid. Molecular Microbiology
Sreevatsan, S., Pan, X., Zhang, Y., Kreiswirth, B.N. 62, 1220–1227.
and Musser, J.M. (1997a) Mutations associated Trebucq, A. (1997) Should ethambutol be
with pyrazinamide resistance in pncA of recommended for routine treatment of
Mycobacterium tuberculosis complex organ- tuberculosis in children? A review of the
isms. Antimicrobial Agents and Chemotherapy literature. International Journal of Tuberculosis
41, 636–640. and Lung Disease 1, 12–15.
Trivedi, S.S. and Desai, S.G. (1987) Pyrazinamidase WHO (2009) Global tuberculosis control 2009 –
activity of Mycobacterium tuberculosis – a test surveillance, planning, financing. WHO Report
of sensitivity to pyrazinamide. Tubercle 68, 221– 2009. WHO/HTM/TB/2009.411. WHO, Geneva,
224. Switzerland.
Vareldzis, B.P., Grosset, J., De Kantor, I., Crofton, Williams, K.J., Chan, R. and Piddock, L.J. (1996)
J., Laszlo, A., Felten, M., et al. (1994) Drug- gyrA of ofloxacin-resistant clinical isolates of
resistant tuberculosis: laboratory issues. World Mycobacterium tuberculosis from Hong Kong.
Health Organization recommendations. Tubercu- Journal of Antimicrobial Chemotherapy 37,
losis and Lung Disease 75, 1–7. 1032–1034.
Vilcheze, C. and Jacobs, W.R. Jr (2007) The Winder, F.G. and Collins, P.B. (1970) Inhibition by
mechanism of isoniazid killing: clarity through isoniazid of synthesis of mycolic acids in
the scope of genetics. Annual Reviews of
Mycobacterium tuberculosis. Journal of General
Microbiology 61, 35–50.
Vilcheze, C., Weisbrod, T.R., Chen, B., Kremer, L.,
Xu, C., Kreiswirth, B.N., Sreevatsan, S., Musser,
Hazbon, M.H., Wang, F., et al. (2005) Altered
J.M. and Drlica, K. (1996) Fluoroquinolone
NADH/NAD+ ratio mediates coresistance to
isoniazid and ethionamide in mycobacteria. resistance associated with specific gyrase
Antimicrobial Agents and Chemotherapy 49, mutations in clinical isolates of multidrug-
708–720. resistant Mycobacterium tuberculosis. Journal
Vilcheze, C., Wang, F., Arai, M., Hazbon, M.H., of Infectious Diseases 174, 1127–1130.
Colangeli, R., Kremer, L., et al. (2006) Transfer Yew, W.W., Chau, C.H., Wong, P.C., Lee, J., Wong,
of a point mutation in Mycobacterium tubercu- C.F., Cheung, S.W., et al. (1995) Ciprofloxacin
losis inhA resolves the target of isoniazid. in the management of pulmonary tuberculosis
Nature Medicine 12, 1027–1029. in the face of hepatic dysfunction. Drugs Under
Vilcheze, C., Av-Gay, Y., Attarian, R., Liu, Z., Experimental and Clinical Research 21, 79–83.
Hazbon, M.H., Colangeli, R., et al. (2008) Youmans, G.P., Raleigh, G.W. and Youmans, A.S.
Mycothiol biosynthesis is essential for ethiona- (1947) The tuberculostatic action of para-
mide susceptibility in Mycobacterium tubercu- aminosalicylic acid. Journal of Bacteriology 54,
losis. Molecular Microbiology 69, 1316–1329. 409–416.
Wang, F., Jain, P., Gulten, G., Liu, Z., Feng, Y., Zhang, Y. and Mitchison, D. (2003) The curious
Ganesula, K., et al. (2010) Mycobacterium characteristics of pyrazinamide: a review.
tuberculosis dihydrofolate reductase is not a International Journal of Tuberculosis and Lung
target relevant to the antitubercular activity of Disease 7, 6–21.
isoniazid. Antimicrobial Agents and Chemo- Zhang, Y., Heym, B., Allen, B., Young, D. and Cole,
therapy 54, 3776–3782. S. (1992) The catalase-peroxidase gene and
Webb, V. and Davies, J. (1999) Antibiotics and isoniazid resistance of Mycobacterium tubercu-
antibiotic resistance in mycobacteria. In: losis. Nature 358, 591–593.
Ratledge, C. and Dale, J. (eds) Mycobacteria:
Zhang, Y., Garbe, T. and Young, D. (1993) Trans-
Molecular Biology and Virulence. Blackwell
formation with katG restores isoniazid-sensitivity
Science, Oxford, UK, pp. ??.
in Mycobacterium tuberculosis isolates resistant
Wengenack, N.L., Uhl, J.R., St Amand, A.L.,
to a range of drug concentrations. Molecular
Tomlinson, A.J., Benson, L.M., Naylor, S., et al.
(1997) Recombinant Mycobacterium tubercu-
losis KatG(S315T) is a competent catalase- Zhang, Y., Vilcheze, C. and Jacobs, W.R. (2005)
peroxidase with reduced activity toward Mechanisms of drug resistance in Myco-
isoniazid. Journal of Infectious Disease 176, bacterium tuberculosis. In: Cole, S.T., Eisenach,
722–727. K.D., McMurray, D.N. and Jacobs, W.R. Jr (eds)
World Health Organization [WHO] (1996) TB: Tuberculosis and the Tubercle Bacillus. ASM
groups at risk. WHO Report on the Tuberculosis Press, Washington, DC.
Epidemic. WHO, Geneva, Switzerland. Zhou, J., Dong, Y., Zhao, X., Lee, S., Amin, A.,
WHO (2000) Guidelines for establishing DOTS- Ramaswamy, S., et al. (2000) Selection of
Plus pilot projects for the management of antibiotic-resistant bacterial mutants: allelic
multidrug-resistant tuberculosis (MDR-TB). diversity among fluoroquinolone-resistant
WHO/CDS/TB/2000.279. WHO, Geneva, mutations. Journal of Infectious Diseases 182,
Switzerland. 517–525.
9 Molecular Tools for Fast Identification
of Resistance and Characterization of
MDR/XDR-TB
Simeon Cadmus1 and Dick van Soolingen2

1TuberculosisResearch Laboratory, Department of Veterinary Public Health and
Preventive Medicine, University of Ibadan, Nigeria; 2Mycobacteria Reference
Laboratory, National Institute of Public Health and the Environment (RIVM), The
Netherlands
9.1 Introduction losis (MDR-TB) and other challenges towards

reaching the 2015 Millennium Development
In a bid to provide meaningful solution to the Goals (MDGs) (Migliori et al., 2007).
persistent burden of tuberculosis in most Globally, there were an estimated 8.8
developing nations and the world at large, million incident cases of tuberculosis and
the strategy for the directly observed therapy 650,000 MDR-TB cases in 2010 (WHO, 2012a).
short course (DOTS) was developed by Dr However, only about 16% of the estimated
Karel Styblo in the 1980s. The DOTS strategy MDR-TB patients and 1.4 million tuberculosis
(combined five key components: government patients who suffer from infection with HIV
commitment, diagnosis through microscopy, worldwide each year have access to suffi-
standardized and supervised treatment, un- ciently sensitive case detection or drug
interrupted drug supply and regular pro- susceptibility testing (WHO, 2012a). There-
gramme monitoring) contributed immensely fore, a major underreporting in the problem
to the improvement of global tuberculosis of MDR-TB is conceivable. Also, significant
control over a decade. Specifically, it involves diagnostic delay aggravated by the dispro-
the inclusion of the standardized short-course portionate frequency of smear-negative
chemotherapy regimens with first-line drugs disease in HIV-associated tuberculosis is
(isoniazid (INH), rifampicin (RIF), pyrazina- common (Getahun et al., 2007; Perkins and
mide (PZA) and streptomycin (SM) or etham- Cunningham, 2007; Uys et al., 2007; Havlir et
butol (EMB), or both) under direct observation, al., 2008). The failure to recognize and treat
at least in the intensive treatment phase, MDR-TB patients quickly leads to increased
regardless of patient drug susceptibility mortality, acquired resistance (including
pattern (WHO, 1997). In furtherance to this extensively drug-resistant tuberculosis, or
programme, in 2006, the World Health XDR-TB) and ongoing transmission (Farmer
Organization (WHO) launched the Stop TB et al., 1998; Van Rie and Enarson, 2006). There
Strategy to pursue, among other things, are 27 high-burden MDR-TB countries; the
DOTS expansion and enhancement and to countries that rank first to fifth in terms of
address TB/HIV, multidrug-resistant tubercu- total numbers of MDR-TB cases are China

126 S. Cadmus and D. van Soolingen
(100,000), India (99,000), the Russian Feder- of tuberculosis with a bad prognosis for the
ation (38,000), Pakistan (15,000) and South concerned patient is defined as MDR-TB with
Africa (13,000) (WHO, 2010). However, as additional resistance to at least one fluoro-
stated earlier, presumably only the tip of the quinolone (FQ) and any of the following
iceberg of the MDR-TB problem is currently injectable second-line drugs: kanamycin
being notified. By the end of 2008, 55 countries (KM), amikacin (AK) or capreomycin (CM).
and territories had also reported at least one Surprisingly, it is estimated that 0–30% of
case of XDR-TB (WHO, 2009). However, the MDR-TB isolates currently tested for second-
problem of XDR-TB most likely has been line drug resistance are, in fact, XDR-TB
developing for many years but has not been (WHO, 2008a). It is currently unknown
adequately recognized. In Europe, for instance, whether XDR-TB is emerging or has been
in the period from 2003 to 2007, about 10% of addressed sufficiently in the previous period.
the MDR-TB cases in the European Union Coupled with the emergence of XDR-TB are
(EU) were, in fact, already XDR-TB (Devaux the technically controversial and recently
et al., 2009). Moreover, on the basis of DNA reported cases of ‘totally drug-resistant
fingerprint surveillance in the EU, it was tuberculosis’ (TDR-TB) or ‘extremely drug
shown that about half of the MDR-TB in this resistant’ (XDR-TB), known to be resistant to
region was caused by transmission and not all first- and second-line anti-tuberculosis
acquired by low-quality treatment (Devaux et drugs that have been found in some patients
al., 2009). In addition, as shown in other in Italy, Iran and India (WHO, 2012b). It is
geographic regions, the Beĳing genotype of currently unknown whether XDR-TB is emerg-
Mycobacterium tuberculosis was associated ing or has been addressed sufficiently in the
with the significant MDR-TB problem in, for previous period. None the less, resistance in
example, the Baltic States and the Former general is emerging and is influencing
Soviet Union States (Devaux et al., 2009) and significantly the development of the current
South Africa (Johnson et al., 2010). Therefore, worldwide tuberculosis epidemic.
it is assumed there is also an important
bacteriological component in the current
emergence of drug resistance globally, 9.2 Principles of Drug Resistance
although not sufficiently addressed (Parwati
et al., 2010). However, in order to meet the In M. tuberculosis and other members of the
global public health challenges related to M. tuberculosis complex, multiple mechanisms
current tuberculosis control, as described by can underlie resistance to anti-tuberculosis
the WHO, the diagnosis and treatment of agents. First, the mycobacterial cell is sur-
MDR-TB and XDR-TB need to be scaled up rounded by a specific, highly hydrophobic
rapidly. cell wall that has a low permeability to many
The treatment of pulmonary tuberculosis compounds (Somoskovi et al., 2001). In
involves the prolonged administration of addition, active drug efflux pumps and
multiple drugs, but is usually highly effective. degrading or inactivating enzymes, and the
However, M. tuberculosis may be resistant to genes that are associated with these functions,
one or more of the drugs, causing a slow have been disclosed in M. tuberculosis (Ramas-
conversion, treatment failures and relapses wamy and Musser, 1998). The concept of
after curative treatment. The four common drug resistance in M. tuberculosis is believed
first-line drugs used in anti-tuberculosis to be mediated almost exclusively by chromo-
therapy are INH, RIF, EMB and PZA. When somal mutations, which affect either the drug
this drug combination fails, MDR-TB (i.e. a target itself or bacterial enzymes that activate
tuberculous disease caused by a bacterial prodrugs (Sandgren et al., 2009). The resist-
strain that is resistant to the two most effective ance mutations emerge naturally at particular
first-line anti-tuberculosis agents, namely paces (1 in 106 for INH and 1 in 108 for RIF);
INH and RIF, ensues). Further down the however, if combinations of effective drugs
spectrum, the previously mentioned problem are applied, the chance that a resistant mutant
of XDR-TB is increasingly reported. This form is selected for is much lower (106  108 = 1014
Fast Identification of Resistance 127
in the case of combined use of INH and RIF), 2007). The mechanism of resistance in the
because the bacteria resistant to one drug will remaining 5% of resistant isolates so far
be killed by the other drug. However, if remains undetermined, with the exception of
suboptimal treatment is applied, the bacteria further mutations at codons 381 (Taniguchi et
with mutations associated with resistance are al., 1996), 481 (Nash et al., 1997), 505 (Matsiota-
selected for and such bacteria will have a Bernard et al., 1998), 508 (Matsiota-Bernard et
relative advantage over the bacteria with the al., 1998) and 509 (Nash et al., 1997) of the rpoB
wild-type sequence, and they will form a new gene. Molecular assays that have been used
bacterial population in the diseased body to screen the rpoB gene for RIF-resistance
sites. Since the early 1990s, numerous studies mutations include DNA sequencing (Kapur
have described the genetic mechanisms of et al., 1994), heteroduplex analysis (Williams
drug resistance in M. tuberculosis, and a et al., 1994), PCR single-stranded con-
wealth of data has accumulated on the formational polymorphism (PCR-SSCP)
mutations found in isolates resistant to (Telenti et al., 1993a,b), line probe assay (LPA)
specific drugs (Ramaswamy and Musser, (De Beenhouwer et al., 1995; Cooksey et al.,
1998). An increasing number of studies have 1997; Van Rie et al., 2010), mismatch analysis
identified the genomic mechanisms which (Nash et al., 1997) and a recently introduced
confer INH and RIF resistance in the majority semi-automated real-time PCR device, the
of clinical isolates; while for other drugs such GeneXpert (Boehme et al., 2010).
as SM and many of the second-line drugs, Although the detection of INH resistance
known resistance mutations occur in only in M. tuberculosis is more complex because a
part of the resistant isolates (Ramaswamy number of genes can be implicated, up to 95%
and Musser, 1998). Hence, because most of this resistance may be due to mutations in
mutations associated with INH and RIF are katG (Hazbón et al., 2006). If the mutations
generally only found in restricted genomic occurring in the inhA gene are also included
sites, molecular tests for these drugs reach a in the test, a reliable detection of INH
high level of sensitivity and specificity, while resistance can be reached in most areas. The
for the other drugs mentioned above the most frequently observed alteration in katG is
positive predictive value can be high, while a serine-to-threonine substitution at codon
the sensitivity is low. As documented, RIF 315 (S315T), located in the active site of the
resistance is rarely encountered as a catalase moiety of the katG gene. Additionally,
monoresistance and therefore usually, in mutations in the promoter region of inhA
about 95% of cases, this indicates resistance to account for 8–20% of INH resistance in M.
a number of other anti-tuberculosis drugs tuberculosis (Hazbón et al., 2006). A C-to-T
(Caws and Drobniewski, 2001; WHO, 2008a). substitution at nucleotide –15 results in the
Accordingly, since RIF monoresistance is overexpression of inhA, an NADH-dependent
relatively rare, RIF resistance is a good enoyl-acyl reductase involved in mycolic acid
indicator of MDR-TB. Point mutations in the synthesis, and INH resistance arises as a
rpoB gene encoding the β-subunit of DNA- result of drug titration (Hazbón et al., 2006).
dependent RNA polymerase have been Although, in general, fitness is reduced by
shown to account for the majority of RIF mutations in the katG gene, because the
resistance worldwide, though with regional bacteria are believed to suffer from a reduced
variations (Gillespie, 2002). More specifically, capability to withstand exposure to oxygen
95% of these RIF resistance-causing mutations radicals in the intracellular environment, this
are located in an 81-bp hotspot region of the phenomenon varies by mutation (Zhang et
rpoB gene, spanning codons 507–533, known al., 1992). Also, the level of resistance is
as the RIF resistance-determining region associated with the type of mutation under-
(RRDR) (Ramaswamy and Musser, 1998). lying INH resistance. Bacteria with codon 315
Mutations in codons 516, 526 and 531 of the mutations have a higher level of resistance
rpoB gene are most commonly associated than bacteria that utilize other mutations
with high-level RIF resistance (Cavusoglu et associated with INH resistance (van Doorn et
al., 2006; Huitric et al., 2006; Rigouts et al., al., 2006). Moreover, while INH resistance in
general is a negative risk factor for trans- and phage-based tests for rapid diagnosis of
mission, codon 315 resistant strains are as MDR-TB (Van Rie et al., 2010).
transmissible as susceptible ones, while the Though recently introduced, the new
mutations at this codon are also associated Genotype M. tuberculosis drug resistance
more frequently with multiple resistance second-line (MTBDRsl) assay (Hain Life-
(van Doorn et al., 2006). science, Nehren, Germany) is another very
promising LPA for the screening of second-
line drugs (i.e. FQs, injectable drugs – AK,
9.3 Novel and Rapid Molecular CM, KM). The implementation of this assay
Diagnostic Tools becomes very useful with the emergence of
XDR-TB.
New molecular methods for rapid molecular
detection of resistance in M. tuberculosis need
to be widely deployed and integrated into 9.3.1 Need for molecular assays
case-finding strategies to reduce transmission
of MDR-TB and to diagnose and treat patients M. tuberculosis generally acquires drug
adequately. Recent advances in availability resistance via the selection of mutants with de
should ensure optimal use of true point-of- novo non-synonymous single nucleotide poly-
care (POC) tests for MDR-TB (Keshavjee and morphisms (nsSNPs), small deletions or inser-
Seung, 2009). With the emergence of resist- tions in specific chromosomal loci, unlike
ance to anti-tuberculosis drugs, several high- most other pathogenic bacteria, which often
throughput sequencing methods and acquire drug resistance via horizontal gene
genotyping strategies have been developed, transfer. This feature of M. tuberculosis drug
and studies to identify the mutations resistance, coupled with fast and efficient
associated with resistance have been under- DNA molecular methods, makes studying
taken globally in the last few years (Sandgren drug resistance highly amenable in molecular
et al., 2009). The WHO Stop TB Partnership’s epidemiology (Mathema et al., 2006).
New Diagnostics Working Group and Calls from the WHO and other inter-
Foundation for Innovative New Diagnostics national public health authorities translate
(FIND) have classified tools for the diagnosis the demand to expand access rapidly to cul-
of active tuberculosis and resistance in three ture and drug susceptibility tests (DST) in
categories, as (i) ‘WHO-endorsed’ tools response to the dual threats that HIV and
(favourably evaluated), (ii) in a ‘late-stage MDR-TB pose as significant challenges to
development or evaluation’ and (iii) tools in tuberculosis control programmes and tubercu-
‘early-stage development’ (Pai et al., 2009). losis laboratory services (WHO, 2007). The
WHO-endorsed tools include molecular costs and complexity of establishing culture
LPAs for MDR-TB diagnosis (GenoType® and DST capacity to meet the anticipated
MTBDRplus, Hain Lifescience, Nehren, need, especially in low-income countries
Germany, and INNO-LiPA Rif. TB, Inno- where these services are currently scarcely
genetics, Gent, Belgium) and a rapid-detec- available, present overwhelming challenges
tion and -speciation assay (Capilia TB-Neo, (see Chapter 1). Consequently, other diag-
TAUNS, Numazu, Japan). A commercially nostic methods, and in particular fast mol-
available tool in late-stage development or ecular approaches, are considered as
evaluation is the GeneXpert MTB/RIF assay alternatives (Barnard et al., 2008). The fast
(Cepheid, California, USA), a nucleic acid developments of molecular approaches in
amplification test (NAAT) for M. tuberculosis tuberculosis diagnosis, especially identifi-
detection and MDR-TB screening. Examples cation and drug susceptibility, question the
of tools in the early phase of development necessity to upgrade all tuberculosis labora-
include the lipoarabinomannan assay, the tories to Biosafety Level 3. It has become
breathalyser screening test, loop-mediated likely that within a few years, microscopy
isothermal amplification technology (TB- and culture can be avoided for the majority of
LAMP, Eiken Chemical Co Ltd, Tokyo, Japan) suspects. Only in cases in which resistance is
detected may additional culture and extended region), the oxyR-ahpC intergenic region
resistance testing still be required. However, (ahpC) and the rpoB gene after PCR amplifi-
with the current increasing initiatives to cation and sequencing using specific primers
apply whole genome sequencing techniques, for all resistance genes. Techniques, such as
it is likely that the vast majority of resistance the real-time polymerase chain reaction, that
mechanisms can be disclosed within the com- make use of primers complementary to resist-
ing years, facilitating the development of ance gene sequences to amplify genomic
complete full-coverage Mycobacterium diag- regions where mutations occur and use
nostic methods. specific probes (i.e. molecular beacons) to
identify mutations, are now available, though
still expensive and complicated, even if
9.3.2 Advantages in molecular highly sensitive and specific (Huang et al.,
diagnostic tools 2009). Reverse hybridization-based assays,
referred to as LPAs, represent useful tools
One of the main progressions in molecular because of their superior cost-effectiveness
diagnostic assays is their enhancing sensi- and speed. These tests are based on the
tivity and the increasing ability to diagnose hybridization of PCR products to specific
tuberculosis in clinical material with nearly probes for wild-type and mutated sequences
the same sensitivity as culture (Davies and of genes involved in drug resistance, and they
Pai, 2008). Again, NAAT can be performed on show a high specificity and medium/high
at least one respiratory specimen from each sensitivity. LPAs, direct DNA sequencing,
patient with the signs and symptoms of molecular beacon analysis and biprobe
pulmonary tuberculosis, for whom a diag- analysis rapidly detect drug resistance
nosis of tuberculosis is being considered but accurately against the most important first-
has not yet been established and for whom line drugs INH and RIF, but these methods
the test result would alter case management are expensive and require extensive training
or tuberculosis control activities (CDC, 2009). of specialized personnel. The recently intro-
In the last decade, molecular tools have duced reversed line blot method for the
been developed based on nucleic acid amplifi- detection of resistance against FQs and
cation in conjunction with electrophoresis, aminoglycosides is promising (Hillemann et
sequencing or hybridization. Though most of al., 2009) and relatively easy to implement,
them are directed at detecting drug resistance providing caution is taken regarding analysis
in M. tuberculosis complex isolates, they are of the results, which can be more difficult
currently also being evaluated for direct than assumed.
detection of DNA of these bacteria and
identification of alleles related to drug
resistance in clinical specimens (such as 9.4 Line Probe Assays (LPAs)
sputum). This current trend is aimed at
generating laboratory results within the Recently, the WHO recommended the use of
shortest time interval in order to enhance molecular LPAs for rapid resistance screening
patient care and prevent further spread to the in low- and middle-income settings (WHO,
larger population. 2008b). LPAs use multiple-target PCR
One of the methods of detecting mutations amplification and reverse line blot hybrid-
is direct sequencing; but although very useful, ization for the identification of Mycobacterium
this approach is still rather expensive and isolates and detection of resistant mutants.
time-consuming. In addition, the technology There are kits available for the distinction of
involved is still limited to few laboratories the M. tuberculosis (sub) species and the most
globally. In most instances, DNA sequencing frequently encountered non-tuberculous
of the katG, inhA, inhA locus, oxyR-ahpC and mycobacteria. In addition, there are reliable
rpoB genes is explored. Huang et al. (2009) and favourably evaluated kits for the
conducted a study involving mutations in detection of mutants associated with RIF and
katG, inhA, the inhA locus (inhA regulatory INH resistance. Lately, in 2009, an LPA kit
was also introduced for the detection of similar caution should be taken. In a study by
resistance against second-line drugs (Hille- Huang et al. (2009), the performances of the
mann et al., 2009), as discussed above. LPAs LPA and DNA sequencing in detecting RIF
can be performed directly from acid-fast and INH resistance associated mutations in
bacilli (AFB) in smear-positive sputum, or the rpoB, katG, inhA regulatory region and
from positive cultures, and provide results in oxyR-ahpC genes were compared to that of a
1 day. A recent systematic review concluded conventional agar proportion DST. A total of
that LPAs were highly sensitive and specific 242 MDR and 30 pan-susceptible M. tubercu-
for detection of RIF resistance (97% and 99%, losis isolates were evaluated in this study. The
respectively) and INH resistance (90% and sensitivities obtained for RIF resistance
99%, respectively) on culture isolates and detection by the GenoType® MTBDRplus test
smear-positive sputum. Overall agreement and by resistance gene sequencing were
with conventional DST for detection of 95.5% and 97.9%, respectively. The sen-
MDR-TB was 99% (Ling et al., 2008). sitivities for INH resistance detection by the
GenoType® MTBDRplus test and by resist-
ance gene sequencing were 81.8% and 93.4%,
9.4.1 Limitations of the Hain respectively. Together, the sensitivity for detec-
MTBDRplus test tion of MDR-TB was 78.5% with the Geno-
Type® MTBDRplus test and 91.3% by
Issues to be considered regarding the resistance gene sequencing. The specificity
implementation of LPAs in high tuberculosis for RIF resistance, INH resistance and
burden countries include the supply of con- MDR-TB detection was 100% by both methods.
sumables and reagents not provided as part However, the GenoType® MTBDRplus test has
of the kit, such as pipette tips and molecular the important advantage of a short turn-
grade water. Furthermore, the necessary around time for the diagnosis of drug-
infrastructure for performing LPAs should be resistant M. tuberculosis and requires only
considered prior to implementation – a mini- relative simple equipment. Overall, the two
mum of three separate rooms is recommended assays performed equally well in detecting
to minimize the risk of contamination of tests RIF resistance (P _ 0.13), although DNA
with previously produced amplicons, which sequencing demonstrated superior perform-
induces false positive results. Restricted ance in detecting INH resistance (P < 0.001)
access to the molecular laboratories and strict and MDR-TB (P < 0.001). The study suggested
adherence to standard operating procedures that additional alleles associated with INH
are essential to reduce the risk of amplicon resistance should be evaluated to improve the
contamination (Albert et al., 2010). sensitivity of the GenoType® MTBDRplus
An additional point of consideration is test, especially in areas with genetically
the interpretation of less obvious test results. diverse and unknown M. tuberculosis strains.
For instance, a caveat in the interpretation of The findings of this report are similar to those
the GenoType® MTBDRplus assay with published by other colleagues. For example,
respect to RIF resistance detection is that katG mutations were found in 97% (77/79)
resistance may be suggested by the absence and inhA mutations in 24% (19/79) of the
of a wild-type hybridization signal, without INH-resistant isolates from KwaZulu-Natal
confirmation of hybridization on a mutant (Kiepiela et al., 2000), whereas Van Rie and
probe. As a wild-type probe may not hybrid- colleagues reported katG mutations in 72% of
ize due to a mutation in the RRDR that is not INH-resistant isolates (41/57) and mutation in
associated with a resistance phenotype (Ma et the inhA gene in only 2% (1/57) of the isolates
al., 2006), such RIF-susceptible isolates could in the Western Cape province of South Africa
be interpreted wrongly as resistant and lead (Van Rie et al., 2001). Studies from other
to the unnecessary removal of RIF from treat- countries have confirmed this variability in
ment regimens. In addition, isolates indicated the distribution of mutations associated with
as resistant due to a mutation at codon 533 INH resistance (Mokrousov et al., 2002; Baker
may be susceptible (Ma et al., 2006), and et al., 2005). Mutations in inhA were, for
instance, also reported as rare in Germany below an acceptable level, otherwise such a
(Hillemann et al., 2007). A high prevalence of molecular test becomes useless.
katG mutations has been reported to account Growth inhibition tests for M. tuberculosis
for a high proportion of INH resistance in isolates therefore remain highly important as
high tuberculosis prevalence countries and a control of the performance of molecular
for a much lower proportion in lower tubercu- tests. The molecular approach may, in the
losis prevalence settings, presumably due to near future, become nearly perfect when
ongoing transmission of these strains in high- whole genome sequencing will yield more
burden settings (Mokrousov et al., 2002). insight into all genomic mechanisms under-
Furthermore, the GenoType® MTBDRplus lying resistance. Translation to simple, cheap
test may need to include new alleles of RIF- and rapid detection of the respective
and INH-resistant genes in order to improve mutations remains a challenge.
sensitivity in the detection of resistance in Overall, LPAs are accurate and useful for
genetically diverse M. tuberculosis strains the rapid detection of drug resistance directly
circulating in various geographic areas. in clinical specimens, providing sufficient
In conclusion, the results available illus- AFB are present. Currently, if the microscopy
trate that the distribution of mutations is positive, it will generally be possible to
resulting in drug resistance in M. tuberculosis amplify the DNA in reversed line blot
differs by geographic regions. This may have methods and therefore direct the therapy in
important implications for the roll-out of an early stage of treatment on the basis of the
rapid genotypic tests for the diagnosis of results. In general, LPAs are expensive and
drug-resistant M. tuberculosis, which may require sophisticated laboratory infrastructure.
impact the progression of M. tuberculosis to Their role and utility in low-income, high-
MDR- and, ultimately, XDR-TB. In particular, burden countries will need to be evaluated
a high proportion of what may be genetically further in field studies.
related INH- and RIF-monoresistant strains
that are not recognized as resistant by the
GenoType® MTBDRplus assay have already 9.4.2 GenoType® MTBDRsl assay
been identified. If rapid genotypic assays for
the detection of drug resistance are to be Studies have shown that resistance to FQs,
widely used, there is a need to monitor local AK, CM and EMB in M. tuberculosis is most
distribution of drug-resistance mutations frequently attributed to mutations in the
continually to ensure that if clonal groups of gyrA, rrs and embB genes, respectively (Hille-
M. tuberculosis with alternative resistance mann et al., 2009). By targeting mutations in
mutations emerge, they are diagnosed prop- codons 90, 91 and 94 in the gyrA gene,
erly as drug resistant. In general, potential approximately 70–90% of all FQ-resistant
users of molecular tests should realize that strains can be detected correctly (Antonova et
the visualization of mutants associated with al., 2008; Mokrousov et al., 2008; van Doorn et
resistance in M. tuberculosis is an indirect al., 2008). Previous reports have linked
approach; it is likely, but not certain, that mutations A1401G, C1402T and G1484T in
these mutations are associated with resistance the rrs gene to AK, CM and KM resistance
and it is uncertain whether negative findings (Alangaden et al., 1998; Maus et al., 2005a,b),
exclude all possibilities of resistance prob- each of them being responsible for a specific
lems. In other words, a large part of the resistance pattern. Mutations G1484T and
mutations that confer resistance are covered A1401G were found to cause high-level resist-
by the current molecular tests, but not all. The ance to all drugs, whereas C1402T caused
reliability of molecular tests varies by geo- resistance against CM and KM only. In order
graphic region. This implies the sensitivity of to increase the capacity to detect further drug
molecular testing will be lower than 100%. resistance in M. tuberculosis, apart from those
However, this is not a problem if the positive due to the first-line drugs (particularly INH
predictive value is sufficiently high. The and RIF, the cause of MDR-TB), the
negative predictive value should not drop GenoType® M. tuberculosis drug-resistance
second-line (MTBDRsl) assay was developed samples in less than 2 h. The rapid detection
with a specific focus on the most prevalent of M. tuberculosis and RIF resistance in clinical
gyrA, rrs and embB gene mutations. Hille- material allows the physician to make critical
mann et al. (2009) reported that this assay patient management decisions regarding
represented a reliable tool for detection of FQ therapy during the same medical encounter.
and AK/CM resistance and it could be applied The GeneXpert system integrates and
directly to smear-positive specimens as well automates sample processing, nucleic acid
as smear-negative specimens known to be amplification and detection of the amplified
culture positive. With a turnaround time of target sequences in simple or complex
approximately 6 h, the diagnosis of second- samples using real-time PCR and reverse
line drug resistance can be shortened from transcriptase PCR. The system consists of an
weeks (conventional DST) to a single day. The instrument, personal computer, barcode
new MTBDRsl assay is therefore a major scanner and preloaded software for inter-
improvement in the routine detection of FQ pretation of the results. The system requires
and AK/CM resistance and, to a lesser extent, the use of single-use disposable GeneXpert
to EMB-resistant M. tuberculosis strains. cartridges that hold the PCR reagents and
Especially in combination with the host the PCR process. Because the cartridges
MTBDRplus assay for the detection of MDR are self-contained, the possibility of cross-
M. tuberculosis, the detection of XDR M. contamination between samples is eliminated.
tuberculosis is made possible within 1 or 2 This is a major step forward.
days (Hillemann et al., 2009). Therefore, the
introduction of the MTBDRsl assay will have
a great impact on the diagnosis of drug 9.5.1 Strengths and limitations of the
resistance to strengthen both the management GeneXpert
of patient therapy and the prevention of
transmission. Van Rie et al. (2010), in an expert commentary
The limitations of this assay are that: review, concluded that, ‘Xpert MTB/RIF is the
first, it has to be used in conjunction with first novel TB diagnostic that has the potential
other LPAs, hence it cannot be used as a for use as a true point of care (POC) tool in
standalone test; and second, its low sensitivity resource-limited settings. Although the
for the detection of EMB resistance: presum- performance characteristics determined by
ably, gene loci other than embB codon 306 are experimental and clinical studies create hope
also associated with EMB resistance. There- for dramatic improvement in the diagnosis of
fore, novel mutations with locations besides TB and drug-resistant TB, the current Xpert
the probe regions or outside the amplified MTB/RIF assay lacks the robustness (espe-
targets can lead to misinterpretations (Hille- cially with regard to its need for continuous
mann et al., 2009). Finally, the MTBDRsl assay electrical power) and the sensitivity in smear-
should be evaluated in different settings negative TB to become the “magic bullet” for
before routine use, since the prevalence of TB diagnosis in high-TB burden, low-resource
mutations associated with FQs, AK, CM and settings. In addition it is doubtful whether
EMB resistance varies in different locations the sole detection of RIF is sufficient to guide
(Hillemann et al., 2009). the treatment in areas where poly-resistance
is prevalent. The automation and simplicity
of the system, however, allow for its use by
9.5 The Semi-automated Real-time low-skilled personnel in the absence of
PCR: The GeneXpert biohazard containment infrastructure, and its
high specificity for the detection of M. tubercu-
The GeneXpert is based on the molecular losis and RIF resistance have already caught
detection of M. tuberculosis complex and rpoB the interest of clinicians, clinical researchers
gene mutations associated with RIF resist- and funding agencies. The POC diagnosis of
ance. This is accomplished in the Xpert MTB/ RIF resistance, a marker for MDR-TB, and the
RIF test by analysing liquefied sputum 72.5% sensitivity of a single Xpert MTB/RIF
assay to diagnose smear-negative culture- Kuz’min, A.V., Larionova, E.E., Smirnova, T.G.,
positive TB cases have created, for the first et al. (2008) Detection of mutations in Myco-
time, the long-awaited possibility of a POC bacterium tuberculosis genome determining
diagnostic tool to reduce the morbidity and resistance to fluoroquinolones by hybridization
on biological microchips. Bulletin of Experi-
mortality of TB.’ Happily, with the endorse-
mental Biology and Medicine 145, 108–113.
ment of the WHO in 2010, over 26 of the 145 Baker, L.V., Brown, T.J., Maxwell, O., Gibson, A.L.,
countries eligible to purchase the Xpert MTB/ Fang, Z., Yates, M.D., et al. (2005) Molecular
RIF kits have done so, with greater prospects analysis of isoniazid-resistant Mycobacterium
in the future. tuberculosis isolates from England and Wales
reveals the phylogenetic significance of the
ahpC-46A polymorphism. Antimicrobial Agents
9.6 Conclusions: Future Perspectives and Chemotherapy 49, 1455–1464.
of Molecular Diagnosis of MDR-TB/ Barnard, M., Albert, H., Coetzee, G., O’Brien, R.
XDR-TB and Bosman, M. (2008) Rapid molecular
screening for multidrug-resistant tuberculosis in
a high-volume public health laboratory in South
Owing to the reality of expanding knowledge Africa. American Journal of Respiratory and
and advances in molecular and genomic Critical Care Medicine 177, 787–792.
research in the area of drug resistance in Boehme, C.C., Nabeta, P., Hillermen, D., Nicol,
tuberculosis, it is envisaged that newer, M.P., Shenai, S., Krapp, F., et al. (2010) Rapid
simpler, more rapid and cheaper molecular molecular detection of tuberculosis and rifampin
inventions and innovations will be brought to resistance. New England Journal of Medicine
bear in unlocking the jigsaw in the diagnosis 363(11), 1005–1015.
of MDR-TB and XDR-TB. It is expected that Cavusoglu, C., Turhan, A., Akinci, P. and Soyler, I.
the current developments in whole genome (2006) Evaluation of the GenoType MTBDR
assay for rapid detection of rifampin and
sequencing will enlarge our knowledge on
isoniazid resistance in Mycobacterium
resistance mechanisms in M. tuberculosis tuberculosis isolates. Journal of Clinical
significantly, and that this will facilitate the Microbiology 44, 2338–2342.
development of more sensitive molecular Caws, M. and Drobniewski, F.A. (2001) Molecular
tests. When this is achieved, it is anticipated techniques in the diagnosis of Mycobacterium
that the problem of MDR-TB and the emerg- tuberculosis and the detection of drug
ing pandemic of XDR-TB can be dealt with in resistance. Annals of the New York Academy of
a more adequate way. However, solving the Science 953,138–145.
problem of tuberculosis is not a matter of Centers for Disease Control and Prevention [CDC]
waiting for innovative miracles, but working (2009) Updated guidelines for the use of nucleic
acid amplification tests in the diagnosis of
extremely hard in an international but also
tuberculosis. Morbidity and Mortality Weekly
local multidisciplinary approach. Report 58, 7–10.
Cooksey, R.C., Morlock, G.P., Glickman, S. and
Crawford, J.T. (1997) Evaluation of a line probe
References assay kit for characterization of rpoB mutations
in rifampin-resistant Mycobacterium tubercu-
Alangaden, G.J., Kreiswirth, B.N., Aouad, A., losis isolates from New York City. Journal of
Khetarpal, M., Igno, F.R., Moghazeh, S.L., et al. Clinical Microbiology 35, 1281–1283.
(1998) Mechanism of resistance to amikacin Davies, P.D. and Pai, M. (2008) The diagnosis and
and kanamycin in Mycobacterium tuberculosis. misdiagnosis of tuberculosis. International
Antimicrobial Agents and Chemotherapy 42, Journal of Tuberculosis and Lung Disease 12,
1295–1297. 1226–1234.
Albert, H., Bwanga, F., Mukkada, S., Nyesiga, B., De Beenhouwer, H., Lhiang, Z., Jannes, G., Mijs,
Ademun, J.P., Lukyamuzi, G., et al. (2010) W., Machtelinckx, L., Rossau, R., et al. (1995)
Rapid screening of MDR-TB using molecular Rapid detection of rifampin resistance in sputum
line probe assay is feasible in Uganda. BMC and biopsy specimens from tuberculosis
Infectious Diseases, 10:41doi:10.1186/1471- patients by PCR and line probe assay. Tubercle
2334-10-41. and Lung Disease 76, 425–430.
Antonova, O.V., Gryadunov, D.A., Lapa, S.A., Devaux, I., Kremer, K., Heersma, H. and van
Soolingen, D. (2009) Clusters of multidrug- E.M., et al. (2010) Drug resistant tuberculosis
resistant Mycobacterium tuberculosis cases, epidemic in the Western Cape driven by a
Europe. Emerging Infectious Diseases 15, virulent Beijing genotype strain. International
1052–1060. Journal of Tuberculosis and Lung Disease 14,
Farmer, P., Bayona, J., Becerra, M., Furin, J., 119–121.
Henry, C., Hiatt, H., et al. (1998) The dilemma of Kapur, V., Li, L.L., Iordanescu, S., Hamrick, M.R.,
MDR-TB in the global era. International Journal Wanger, A., Kreiswirth, B.N., et al. (1994)
of Tuberculosis and Lung Disease 2, 869–876. Characterization by automated DNA sequenc-
Getahun, H., Harrington, M., O’Brien, R. and Nunn, ing of mutations in the gene (rpoB) encoding
P. (2007) Diagnosis of smear-negative pulmon- the RNA polymerase b subunit in rifampin-
ary tuberculosis in people with HIV infection or resistant Mycobacterium tuberculosis strains
AIDS in resource-constrained settings: inform- from New York City and Texas. Journal of
ing urgent policy changes. The Lancet 369, Clinical Microbiology 32, 1095–1098.
2042–2049. Keshavjee, S. and Seung, K. (2009) Stemming the
Gillespie, S.H. (2002) Evolution of drug resistance tide of multidrug-resistant tuberculosis: major
in Mycobacterium tuberculosis: clinical and barriers to addressing the growing epidemic. In:
molecular perspective. Antimicrobial Agents Giffin, R. and Robinson, S. (eds) Institute of
and Chemotherapy 46, 267–274. Medicine. Addressing the Threat of Drug
Havlir, D.V., Getahun, H., Sanne, I. and Nunn, P. Resistant Tuberculosis: A Realistic Assessment
(2008) Opportunities and challenges for HIV of the Challenge: Workshop Summary. National
care in overlapping HIV and TB epidemics. Academies Press, Washington, DC, pp. 141–
Journal of American Medical Association 300, 235.
423–430. Kiepiela, P., Bishop, K.S., Smith, A.N., Roux, L. and
Hazbón, M.H., Brimacombe, M., del Valle, M.B., York, D.F. (2000) Genomic mutations in the
Cavatore, M., Guerrero, M.I., Varma-Basil, M., katG, inhA and aphC genes are useful for the
et al. (2006) Population genetics study of prediction of isoniazid resistance in
isoniazid resistance mutations and evolution of Mycobacterium tuberculosis isolates from Kwa-
multidrug-resistant Mycobacterium tuberculosis. Zulu Natal, South Africa. Tubercle and Lung
Antimicrobial Agents and Chemotherapy 50, Disease 80, 47–56.
2640–2649. Ling, D., Zwerling, A. and Pai, M. (2008) GenoType
Hillemann, D., Rusch-Gerdes, S. and Richter, E. MTBDR assays for the diagnosis of multidrug-
(2007) Evaluation of the Genotype MTBDRplus resistant tuberculosis: a meta-analysis.
assay for rifampin and isoniazid susceptibility European Respiratory Journal 32, 1165–1174.
testing of Mycobacterium tuberculosis strains Ma, X., Wang, H., Deng, Y., Liu, Z., Xu, Y., Pan, X.,
and clinical specimens. Journal of Clinical et al. (2006) rpoB gene mutations and molecular
Microbiology 45, 2635–2640. characterization of rifampin-resistant Myco-
Hillemann, D., Rusch-Gerdes, S. and Richter, E. bacterium tuberculosis isolates from Shandong
(2009) Feasibility of the GenoType MTBDRsl/ Province, China. Journal of Clinical Microbiology
Assay for fluoroquinolone, amikacin-capreo- 44, 3409–3412.
mycin, and ethambutol resistance testing of Mathema, B., Kurepina, N.E., Bifani, P.J. and
Mycobacterium tuberculosis strains and clinical Kreiswirth, B.N. (2006) Molecular epidemiology
samples. Journal of Clinical Microbiology 47, of tuberculosis: current insights. Clinical
1767–1772. Microbiology Reviews 19, 658–685.
Huang, W., Chen, H., Kuo, Y. and Jou, R. (2009) Matsiota-Bernard, P., Vrioni, G. and Marinis, E.
Performance assessment of the GenoType (1998) Characterization of rpoB mutations in
MTBDRplus test and DNA sequencing in rifampin-resistant clinical Mycobacterium
detection of multidrug-resistant Mycobacterium tuberculosis isolates from Greece. Journal of
tuberculosis. Journal of Clinical Microbiology Clinical Microbiology 36, 20–23.
47, 2520–2524. Maus, C.E., Plikaytis, B.B. and Shinnick, T.M.
Huitric, E., Werngren, J., Juréen, P. and Hoffner, S. (2005a) Molecular analysis of cross resistance
(2006) Resistance levels and rpoB gene to capreomycin, kanamycin, amikacin, and
mutations among in vitro-selected rifampin- viomycin in Mycobacterium tuberculosis.
resistant Mycobacterium tuberculosis mutants. Antimicrobial Agents and Chemotherapy 49,
Antimicrobial Agents and Chemotherapy 50, 3192–3197.
2860–2862. Maus, C.E., Plikaytis, B.B. and Shinnick, T.M.
Johnson, R., Warren, R.M., van der Spuy, G.D., (2005b) Mutation of tlyA confers capreomycin
Gey van Pittius, N.C., Theron, D., Streicher, resistance in Mycobacterium tuberculosis.
Antimicrobial Agents and Chemotherapy 49, isoniazid, rifampicin, and pyrazinamide in Myco-
571–577. bacterium tuberculosis. Respiratory Research
Migliori, G.B., Loddenkemper, R., Blasi, F. and 2, 164–168.
Raviglione, M.C. (2007) 125 years after Robert Taniguchi, H., Aramaki, H., Nikaido, Y., Mizuguchi,
Koch’s discovery of the tubercle bacillus: the Y., Nakamura, M., Koga, T., et al. (1996)
new XDR-TB threat. Is ‘science’ enough to Rifampicin resistance and mutation of the rpoB
tackle the epidemic? European Respiratory gene in Mycobacterium tuberculosis. FEMS
Journal 29, 423–427. Microbiology Letters 144, 103–108.
Mokrousov, I., Narvskaya, O., Otten, T., Telenti, A., Imboden, P., Marchesi, F., Lowrie, D.,
Limenschenko, E., Steklova, L. and Vyshnevskiy, Cole, S., Colston, M.J., et al. (1993a) Detection
B. (2002) High prevalence of katG Ser315Thr of rifampin resistance mutations in Myco-
substitution among isoniazid-resistant Myco- bacterium tuberculosis. The Lancet 341, 647–
acterium tuberculosis clinical isolates from 650.
Northwestern Russia, 1996–2001. Antimicrobial Telenti, A., Imboden, P., Marchesi, F., Schmidheini,
Agents and Chemotherapy 46, 1417–1424. T. and Bodmer, T. (1993b) Direct, automated
Mokrousov, I., Otten, T., Manicheva, O., Potapova, detection of rifampin-resistant Mycobacterium
Y., Vishnevsky, B., Narvskaya, O., et al. (2008) tuberculosis by polymerase chain reaction and
Molecular characterization of ofloxacin-resistant single-strand conformation polymorphism
Mycobacterium tuberculosis strains from analysis. Antimicrobial Agents and Chemo-
Russia. Antimicrobial Agents and Chemotherapy therapy 37, 2054–2058.
52, 2937–2939. Uys, P.W., Warren, R.M. and van Helden, P.D.
Nash, K.A., Gaytan, A. and Inderlied, C.B. (1997) (2007) A threshold value for the time delay to TB
Detection of rifampin resistance in Myco- diagnosis. PLoS ONE 2, e757.
bacterium tuberculosis by means of a rapid, van Doorn, H.R., de Haas, P.E.W., Kremer, K.,
simple, and specific RNA/RNA mismatch assay. Vandenbroucke-Grauls, C.M.J.E., Borgdorff,
Journal of Infectious Diseases 176, 533–536. M.W. and van Soolingen, D. (2006) Public health
Pai, M., Minion, J., Sohn, H., Zwerling, A. and impact of isoniazid-resistant Mycobacterium
Perkins, M.D. (2009) Novel and improved tuberculosis strains with a mutation at amino-
technologies for tuberculosis diagnosis: acid position 315 of katG: a decade of
progress and challenges. Clinical Chest experience in the Netherlands. Clinical
Medicine 30, 701–716. Microbiology and Infection 12, 769–775.
Parwati, I., van Crevel, R. and van Soolingen, D. van Doorn, H.R., An, D.D., de Jong, M.D., Lan,
(2010) Possible underlying mechanisms for N.T.N., Hoa, D.V., Quy, H.T., et al. (2008)
successful emergence of the Mycobacterium Fluoroquinolone resistance detection in
tuberculosis Beijing genotype strains. The Mycobacterium tuberculosis with locked nucleic
Lancet Infectious Disease 10, 103–111. acid probe real-time PCR. International Journal
Perkins, M.D. and Cunningham, J. (2007) Facing of Tuberculosis and Lung Disease 12, 736–742.
the crisis: improving the diagnosis of Van Rie, A. and Enarson, D. (2006) XDR tubercu-
tuberculosis in the HIV era. Journal of Infectious losis: an indicator of public-health negligence.
Disease 196 (Suppl 1), S15–S27. The Lancet 368, 1554–1556.
Ramaswamy, S. and Musser, J.M. (1998) Molecular Van Rie, A., Warren, R., Mshanga, I., Jordaan,
genetic basis of antimicrobial agent resistance A.M., van der Spuy, G.D., Richardson, M., et al.
in Mycobacterium tuberculosis: 1998 update. (2001) Analysis for a limited number of gene
International Journal of Tuberculosis and Lung codons can predict drug resistance of
Disease 79, 3–29. Mycobacterium tuberculosis in a high-incidence
Rigouts, L., Nolasco, O., de Rijk, P., Nduwamahoro, community. Journal of Clinical Microbiology 39,
E., Van Deun, A., Arevalo, J., et al. (2007) Newly 636–641.
developed primers for comprehensive amplifi- Van Rie, A., Page-Shipp, L., Scott, L., Sanne, I. and
cation of the rpoB gene and detection of rifampin Stevens, W. (2010) Xpert® MTB/RIF for point-of
resistance in Mycobacterium tuberculosis. care diagnosis of TB in high-HIV burden,
Journal of Clinical Microbiology 45, 252–254. resource-limited countries: hype or hope? Expert
Sandgren, A., Strong, M., Muthukrishnan, P., Review of Molecular Diagnostics 10, 937–946.
Weiner, B.K., Church, G.M. and Murray, M.B. Williams, D.L., Waguespack, C., Eisenach, K.,
(2009) Tuberculosis drug resistance mutation Crawford, J.T., Portaels, F. and Salfinger, M.
database. PLoS Medicine 6(2), e1000002. (1994) Characterization of rifampin resistance
Somoskovi, A., Parsons, L.M. and Salfinger, M. in pathogenic mycobacteria. Antimicrobial
(2001) The molecular basis of resistance to Agents and Chemotherapy 38, 2380–2386.
World Health Organization [WHO] (1997) Global (WHO/HTM/TB/2009.411). World Health Organ-
Tuberculosis Programme. Treatment of Tubercu- ization, Geneva, Switzerland.
losis: Guidelines for National Programmes, 2nd WHO (2010) Multidrug and extensively drug-
edn. Publication WHO/GTB/96.210. World resistant TB (M/XDR-TB) 2010. Global Report
Health Organization, Geneva, Switzerland. on Surveillance and Response (http://whqlib
WHO (2007) The global MDR-TB and XDR-TB doc.who.int/publications/2010/9787, accessed
response plan. WHO/HTM/TB/2007.387. World 18 September 2012).
Health Organization, Geneva, Switzerland. WHO (2012a) Tuberculosis Global Facts (http://
WHO (2008a) Antituberculosis drug resistance in www.who.int/tb/publications/2011/factsheet,
the world. Fourth global report (http://www.who. accessed 18 September 2012).
int/tb/publications/2008/drs_report4_26feb08. WHO (2012b) Totally drug-resistant tuberculosis: a
pdf, accessed 5 January 2009). WHO consultation on the diagnostic definition
WHO (2008b) Policy statement. Molecular line and treatment options (http://www.who.int/tb/
probe assays for rapid screening of patients at challenges/xdr/xdrconsultation/en/, accessed
risk of multidrug resistant tuberculosis (MDR-TB) 18 September 2012).
(http://www.who.int/tb/dots/laboratory/lpa_policy. Zhang, Y., Heym, B., Allen, B., Young, D. and Cole,
pdf, accessed 18 November 2009). S. (1992) The catalase-peroxidase gene and
WHO (2009) Global tuberculosis control – isoniazid resistance of Mycobacterium tubercu-
epidemiology, strategy, financing: WHO report losis. Nature 358, 591–593.
Part III
Understanding Treatment
10 Monitoring Therapy by Bacterial Load
Denise M. O’Sullivan
Molecular and Cell Biology Team, Laboratory for the Government Chemist,
Teddington, UK
10.1 Introduction is useful for monitoring the patient during

treatment. Acid-fast microscopy gives a semi-
The determination of bacterial load allows quantitative estimate as to the burden of
the monitoring of antimicrobial therapy; how tuberculosis in the patient. The examination
the patient and bacteria are responding. The of the smear provides an estimate as to the
monitoring of bacterial load could lead to the number of bacilli that are visible per micro-
challenging of an antibiotic regimen as to scopy field. Up to 100 fields are examined to
whether it is delivering the optimal dose. determine smear positivity and then the
Bacterial load could identify individuals who smear is scored as to the number of acid-fast
could benefit from more aggressive therapy bacilli. According to WHO guidelines for
management. In order to evaluate new Ziehl–Neelsen (ZN) staining, >10 bacilli/field
therapeutics, the monitoring of bacterial load is a 3+, 1–10/field is a 2+, 10–99/100 fields is a
is critical in Phase IIB studies. The measure- 1+, 1–9/100 fields is reported as scanty and 0
ment of bacterial load could indicate patients bacilli/100 fields is reported as smear nega-
who are at increasing risk of treatment failure. tive. There are also smear scoring guidelines
In drug-susceptible patients who are receiv- for fluorochrome microscopy using auramine
ing the standard regimen recommended by staining, and there is a separate scoring
the World Health Organization (WHO) (2- system provided by the American Thoracic
month intensive phase and 4-month con- Society which is based on the same principle
tinuation phase), the relapse rate is 7% or less as the WHO system (2000). In order for a ZN-
and the failure rate is 1–4% (Dye et al., 2005). stained smear to be positive, there must be
approximately 10,000 bacilli/ml of sputum.
Fluorescence microscopy of auramine-
10.2 Microscopy stained smears is approximately 10% more
sensitive than light microscopy of ZN-stained
Smear microscopy is the gold standard in the smears (Steingart et al., 2006a). Fluorescence
identification of Mycobacterium tuberculosis in microscopy using light-emitting diodes could
sputum from patients with pulmonary improve tuberculosis case detection in high
tuberculosis. It has been reported to have tuberculosis burden settings, as they have
variable sensitivity (30–80%) (Steingart et al., advantages over conventional fluorescence
2006b). Microscopic examination of sputum microscopy. They are cheaper, can be used in

140 D.M. O’Sullivan
resource-poor settings as they do not require 10.3 Culture

a darkroom, require less examination time
and they last longer (Albert et al., 2010). Bacterial load can also be measured by cul-
Smears should be prepared when the ture, which is more specific and sensitive
patient first presents. The preparation of a than smear. However, this is time-consuming,
smear is only possible in the case of pulmonary costly and requires appropriate facilities,
tuberculosis when the patient is producing which are not universally available. It can
sputum, which contains the bacilli (infectious take up to 12 weeks to get a result by culture,
period). Smears should be examined as the which negates the method’s usefulness when
patient is progressing through treatment it comes to the real-time monitoring of
during both the intensive and the continuation therapy. Culturing of sputum samples requires
phases of treatment. The WHO recommends liquefaction to release the mycobacteria from
examination of the smear after the intensive the cells and mucus in the isolate. Reagents
phase of treatment has been completed such as N-acetyl-l-cysteine (NALC) and
(WHO, 2009). It is imperative that sputum is sputasol (dithiotheritol) followed by sample
examined at the end of treatment, to confirm vortexing are used to treat the sputum.
cure. The term ‘conversion rate’ is used when Culturing also requires decontamination of
monitoring treatment programmes to describe the sputum sample, which usually involves
the proportion of patients who have smear- exposure to strong bases such as sodium
positive disease at the beginning of treatment hydroxide or acids followed by neutralization
which becomes smear negative when they are (Kubica et al., 1963). This results in removing
on treatment. Typically, a conversion rate of organic debris and killing the contaminating
2–3 months is an indicator of a good treatment bacterial and fungal flora, but also a 62–78%
programme (Blumberg et al., 2003; WHO, reduction in the number of bacilli present
2010). The period of sputum smear positivity (Mitchison et al., 1972; Grandjean et al., 2008).
can typically last 2 months after initiation of Sodium hydroxide can be used as a mucolytic
treatment (Blumberg et al., 2003), but patients and decontaminating agent. There are strict
who are responding to treatment can still incubation periods that must be adhered to,
have positive smears (Telzak et al., 1997; Al- to minimize the reduction in mycobacterial
Moamary et al., 1999). Even in drug- cells while ensuring sufficient killing of the
susceptible cases where the patient is receiving contaminating flora. When smears are
optimum therapy, 10% remain smear positive prepared from concentrated specimens rather
after 2 months (Fitzwater et al., 2010). This than directly from the sputum, it has been
limits the usefulness of smear microscopy in shown to increase the sensitivity of the test
monitoring bacterial load, as patients can still (Woods and Witebsky, 1995). Unlike in smear
expectorate bacilli which are dead in their microscopy, where 105 bacteria are required
sputum even though they are responding to to obtain a positive result, culture is much
treatment. If the sputum smear is positive at more sensitive, with approximately 10–100
the end of the intensive phase of treatment, it viable bacilli required to obtain a positive
could, however, indicate that the patient has a result (Rouillon et al., 1976). Sputum sample
drug-resistant strain of M. tuberculosis, subculture also provides sample for drug
optimal drug dosing, poor-quality drugs, susceptibility testing, species identification
poor adherence to the regimen or the patient and determining the strain genotype.
could have extensive disease with cavitation. Methods which avoid decontamination use
Smear microscopy is not specific for M. selective growth media to inhibit the growth
tuberculosis and can detect mycobacteria other of contaminating flora and directly quantify
than tuberculosis (Schluger and Rom, 1994). the bacterial load in the sputum sample. An
Approximately 40–60% of patients that have assay which uses selective media is the
culture-positive tuberculosis are smear nega- sputum serial colony-counting assay. This
tive (Mase et al., 2007) and co-infection with method is commonly used when evaluating
HIV increases the likelihood of smear- new therapeutics in clinical trials. It deter-
negative disease (Getahun et al., 2007). mines the number of colony-forming units
Monitoring Therapy by Bacterial Load 141
(CFU) per ml throughout the treatment at set treatment (Epstein et al., 1998). The MGIT
time points (Mitchison, 2006). The selective system detects O2 consumption by
growth reagents required for this method are fluorescence and the MB/BacT and BacT/
costly and it requires a long incubation time ALERT detection systems monitor the amount
for the results. Early bactericidal studies use of CO2. When the level of CO2 reaches a
serial culture samples collected for up to 5 threshold, the machine signals the sample is
days after the start of therapy to estimate the positive and reports a TTP reading, which
potency of different anti-tubercular agents can then be correlated to the number of
(Jindani et al., 1980; Gillespie et al., 2002). bacteria present (Joloba et al., 2001). The
Sputum culture conversion after 2 months liquid culture systems report a positive
is used to monitor treatment progression and sample when there are 105–106 bacteria pre-
to determine the relapse rate (Mitchison, sent. There is an acceptable range of con-
1993). Examining changes in CFU from serial tamination of 3–5% for liquid culture systems
sputum samples at weekly or biweekly (Della Latta, 2004). If the contamination rate
intervals could also provide a marker for is less than 3%, this would suggest an overly
relapse (Brindle et al., 2001). In the work by severe decontamination procedure, and a con-
Brindle et al., the rate of decline in bacterial tamination rate of greater than 5% would
load was measured as the slope of log CFU suggest decontamination was inadequate or
counts and it was observed that there was a sample digestion was incomplete. There is a
higher rate of decline in one drug regimen good correlation between viable counts and
(streptomycin, isoniazid, rifampicin and TTP, as determined by liquid culture systems
pyrazinamide) compared to another (strepto- (O’Sullivan et al., 2007). The relationship
mycin, isoniazid and thiacetazone). They between TTP and CFU in serial samples can
observed the steepest decline in CFU counts be used as a marker of treatment response
during the initial days of therapy, which was (the TTP increases as the CFU decreases)
likely to be due to the potent sterilizing (Pheiffer et al., 2008). The relationship be-
activity of isoniazid. Following the first 2 tween cavitation as observed radiographically,
days of therapy, the effect of rifampicin is TTP as determined by liquid culture and
thought to be involved in the reduction of the bacterial load has been investigated (Perrin et
bacterial load as well as the bactericidal effect al., 2010). It was observed that patients with
of pyrazinamide. They also observed a dif- cavities on their thoracic computed tomo-
ference between HIV-negative and -positive graphy (CT) scan had significantly shorter
patients, with a greater decline in bacterial TTP results and therefore a higher CFU
load in the initial phases of treatment among compared to patients with no cavitation. The
HIV-negative patients. presence of cavities could therefore require
Liquid culture systems such as the BD prolonged treatment; however, this is likely
BACTEC™ Mycobacteria Growth Indicator to be because of the association between
Tube (MGIT™) mycobacterial detection sys- cavitation and high bacterial load. Predicting
tem and the bioMérieux MB/BacT® and BacT/ treatment outcome using culture widely
ALERT® 3D are used in diagnostic clinical accepts the evaluation of sputum specimens
laboratories, as they use a culture media at 2 months (Wallis et al., 2009). However, a
which is optimized for the rapid growth of systematic review of this subject by Horne et
mycobacteria that measure time to positivity al. found this method to have poor sensitivity;
(TTP) in days. These automated liquid culture 57% for predicting failure and 24% for
systems can shorten TTP significantly predicting relapse (Horne et al., 2010). This
compared to traditional culture, but this can indicator is still of clinical use, as a positive
still take 12.6–17.7 days (mean) (Somoskovi et culture will trigger a follow-up culture at 3
al., 2000). It has been used to monitor months and it could also provide an early
tuberculosis patients and has been shown to indication of a poor treatment outcome.
correlate more closely, when compared to It has been shown that a subpopulation of
other bacteriological, radiological and clinical bacterial cells exists in vitro which can only be
evaluations, with the patient response to cultured by adding resuscitation-promoting
factors (Rpfs), which are mycobacterial to the shedding of dead or non-replicating

proteins that are involved in stimulating bacilli from the site of infection, and also
growth among non-replicating cells in vitro DNA which is resistant to degradation.
(Mukamolova et al., 2002; Shleeva et al., 2002). Large-scale studies of the usefulness of PCR
As a result, it may be important to add these in a diagnostic setting have shown it to be less
proteins to the culture to increase the accuracy sensitive than culture (Clarridge et al., 1993).
in measuring the patient’s response to chemo- The target most often used in PCR is the
therapy, especially as this subpopulation is insertion sequence (IS) element 6110
thought to increase relative to the replicating (Eisenach et al., 1991; Greco et al., 2009).
M. tuberculosis cells (Mukamolova et al., 2010). However, careful primer selection is required
to increase the specificity of using this target
(Kent et al., 1995; Mulcahy et al., 1996). Also,
10.4 Molecular Tools there are strains of M. tuberculosis which lack
IS6110 (Das et al., 1995), so it could be useful
Methods other than traditional microbio- when combined with another target such as
logical techniques can be used to monitor hsp65, rpoB or the M. tuberculosis complex
bacterial load in clinically useful timescales, specific protein MBP-64 (Dinnes et al., 2007).
which can then predict clinical outcome. Using a real-time system for PCR product
These methods described include molecular detection removes the need for gel electro-
tests which can use DNA or RNA species. phoresis, which makes the PCR process
These tests could overcome the low sensitivity simpler, less cumbersome and there is less
and specificity of microscopy and the time- risk of cross-contamination. Real-time PCR
consuming culture methods. The use of PCR can use specific fluorescently labelled probes
in monitoring patients undergoing treatment or molecular beacons, which also make the
has been investigated (Kennedy et al., 1994; assay more specific.
Levée et al., 1994). Thomsen et al. investigated A new real-time PCR-based system, the
the optical density ratio of PCR for M. Cepheid Xpert MTB/RIF assay, has achieved
tuberculosis to an internal inhibition control high sensitivity and specificity in clinical
for 22 patients, compared this to baseline settings (Boehme et al., 2010; Rachow et al.,
levels which had been previously established 2011). Although a qualitative test, it has
and found that the ratio correlated with the potential to relate Ct (crossing threshold)
extent of the disease (Thomsen et al., 1999). value generated by target amplification to
For instance, patients with extensive disease bacterial load (Rachow et al., 2011). Results
did not reach baseline until at least 1 year are determined by their Ct value, so a low Ct
after treatment initiation. So, PCR could be value would relate to a sample with a high
used in this setting combined with smear and concentration of M. tuberculosis complex.
culture results to indicate suboptimal treat- Rachow et al. reported good correlation
ment, non-compliance, resistant M. tubercu- among the Xpert MTB/RIF assay result, TTP
losis and reduced drug absorption, and could determined by liquid culture and grade of
then predict patients who are at risk of smear positivity. The limit of sensitivity of the
relapse. The use of PCR in a semi-quantitative Xpert system is 131 CFU/ml (Marlowe et al.,
manner as described appears to have more 2011).
promise compared to using it qualitatively Messenger RNA (mRNA) methods could
(Chierakul et al., 2001). A negative result from overcome the false positivity by only detect-
a nucleic acid amplification test could be ing live bacilli compared to dead and live
false, as there may be inhibition of amplifi- bacilli estimated by DNA methods. Most
cation or the limit of detection is too high. mRNA species have a short half-life, so they
These false positives can be controlled for by are unstable (Hu and Coates, 1999). Reverse
confirmation by smear microscopy result. transcriptase PCR targeting the M. tuberculosis
The PCR result can continue to remain posi- antigen 85B has been shown to disappear
tive even after patients become culture from sputum from patients who are respond-
negative (Yuen et al., 1993). This could be due ing to treatment, as indicated by a reduction
in viable colony counts (Desjardin et al., 1999). still have a strong early secretory antigenic
In another study, the levels of antigen 85 in target 6 (ESAT-6) response for up to several
sputum samples were able to predict which years (Wu-Hsieh et al., 2001). ESAT-6, in
patients were ‘persisters’ (Wallis et al., 1998). conjunction with culture filtrate protein 10
Quantitative reverse transcriptase real-time (CFP-10), is an M. tuberculosis secreted protein
PCR has investigated isocitrate lyase (icl) which is synthesized and used in IFN-γ
mRNA and found it correlates with M. release assays such as T-SPOT®.TB and the
tuberculosis viability as detected by culturing QuantiFERON®-TB Gold to measure immune
methods (Li et al., 2010). Following 1 month response in patients.
of a standard 4-drug regimen, icl mRNA
correlated with both liquid and solid culture;
however, after 2 months of therapy, it 10.6 Conclusion
correlated more closely with liquid culture.
The use of RNA species is discussed further The principal measure of treatment outcome
in Chapter 12. in patients is conversion of positive sputum
culture to negative after 2 months. Clinical
decisions can be made with respect to the
10.5 Immunological Measures results from culture-based methods. How-
ever, these methods are lengthy and require
It is unclear how well the immunological long incubation times. New methods can be
status of a patient correlates with bacterial used to determine treatment outcome; how-
load. Several immunological assays have ever, currently, culture remains the standard
attempted to characterize patient improve- method. The monitoring of bacterial load is
ment during and after treatment. The measur- important in anti-tubercular therapy due to
ing of cell-mediated immunity has used lengthy treatment regimens and thus the risk
assays to determine interleukin (IL)-10 and of relapse.
IL-12 levels (Sai Priya et al., 2009). Millington
et al. found a decline in T-cells secreting
interferon-gamma (IFN-γ) and an increase in References
T-cells secreting IL-2 in relation to antigen
load (Millington et al., 2007). Methods have Al-Moamary, M.S., Black, W., Bessuille, E., Elwood,
used ex vivo enzyme-linked immunospot R.K. and Vedal, S. (1999) The significance of
assay (ELISpot) to quantify the secretion of the persistent presence of acid-fast bacilli in
IFN-γ from Th1-type T-cells (Ewer et al., sputum smears in pulmonary tuberculosis.
2006). It may be possible to estimate bacterial Chest 116, 726–731.
load using ELISpot to quantify the M. Albert, H., Manabe, Y., Lukyamuzi, G., Ademun, P.,
Mukkada, S., Nyesiga, B., et al. (2010)
tuberculosis-specific T-cells (Lalvani, 2004).
Performance of three LED-based fluorescence
However, studies have shown conflicting microscopy systems for detection of tuberculosis
results and have shown that individuals vary in Uganda. PLoS ONE 5, e15206.
in the production of M. tuberculosis-specific American Thoracic Society (2000) Diagnostic
T-cells in response to bacterial burden Standards and Classification of Tuberculosis in
(Tibayrenc, 2007). Current immunological Adults and Children. This official statement of
assays measuring T-cell response could the American Thoracic Society and the Centers
measure bacterial load in a patient during for Disease Control and Prevention was
treatment, although this has also been found adopted by the ATS Board of Directors, July
to vary over time, but could not be used to 1999. This statement was endorsed by the
Council of the Infectious Disease Society of
measure bacterial load between patients. It is
America, September 1999. American Journal of
clear that for an immunological test to prove Respiratory and Critical Care Medicine 161,
useful in determining mycobacterial load, it 1376–1395.
will need to include multiple immunological Blumberg, H.M., Burman, W.J., Chaisson, R.E.,
markers. It has been shown that patients who Daley, C.L., Etkind, S.C., Friedman, L.N., et al.
have completed treatment successfully could (2003) American Thoracic Society/Centers for
Disease Control and Prevention/Infectious J.H. and Crawford, J.T. (1991) Detection of
Diseases Society of America: treatment of Mycobacterium tuberculosis in sputum samples
tuberculosis. American Journal of Respiratory using a polymerase chain reaction. American
and Critical Care Medicine 167, 603–662. Reviews of Respiratory Disease 144, 1160–
Boehme, C.C., Nabeta, P., Hillemann, D., Nicol, 1163.
M.P., Shenai, S., Krapp, F., et al. (2010) Rapid Epstein, M.D., Schluger, N.W., Davidow, A.L.,
molecular detection of tuberculosis and rifampin Bonk, S., Rom, W.N. and Hanna, B. (1998) Time
resistance. New England Journal of Medicine to detection of Mycobacterium tuberculosis in
363, 1005–1015. sputum culture correlates with outcome in
Brindle, R., Odhiambo, J. and Mitchison, D. (2001) patients receiving treatment for pulmonary
Serial counts of Mycobacterium tuberculosis in tuberculosis. Chest 113, 379–386.
sputum as surrogate markers of the sterilising Ewer, K., Millington, K.A., Deeks, J.J., Alvarez, L.,
activity of rifampicin and pyrazinamide in treat- Bryant, G. and Lalvani, A. (2006) Dynamic
ing pulmonary tuberculosis. BMC Pulmonary antigen-specific T-cell responses after point-
Medicine 1, 2–9. source exposure to Mycobacterium tuberculosis.
Chierakul, N., Chaiprasert, A., Tingtoy, N., American Journal of Respiratory and Critical
Arjratanakul, W. and Pattanakitsakul, S.-N. Care Medicine 174, 831–839.
(2001) Can serial qualitative polymerase chain Fitzwater, S.P., Caviedes, L., Gilman, R.H.,
reaction monitoring predict outcome of Coronel, J., Lachira, D., Salazar, C., et al. (2010)
pulmonary tuberculosis treatment? Respirology Prolonged infectiousness of tuberculosis
6, 305–309. patients in a directly observed therapy short-
Clarridge, J.E. 3rd, Shawar, R.M., Shinnick, T.M. course program with standardized therapy.
and Plikaytis, B.B. (1993) Large-scale use of Clinical Infectious Diseases 51, 371–378.
polymerase chain reaction for detection of Getahun, H., Harrington, M., O’Brien, R. and Nunn,
Mycobacterium tuberculosis in a routine P. (2007) Diagnosis of smear-negative
mycobacteriology laboratory. Journal of Clinical pulmonary tuberculosis in people with HIV
Microbiology 31, 2049–2056. infection or AIDS in resource-constrained
Das, S., Paramasivan, C.N., Lowrie, D.B., settings: informing urgent policy changes. The
Prabhakar, R. and Narayanan, P.R. (1995) Lancet 369, 2042–2049.
IS6110 restriction fragment length poly- Gillespie, S.H., Gosling, R.D. and Charalambous,
morphism typing of clinical isolates of B.M. (2002) A reiterative method for calculating
Mycobacterium tuberculosis from patients with the early bactericidal activity of antituberculosis
pulmonary tuberculosis in Madras, South India. drugs. American Journal of Respiratory and
Tubercle and Lung Disease 76, 550–554. Critical Care Medicine 166, 31–35.
Della Latta, P. (2004) Mycobacteriology and Grandjean, L., Martin, L., Gilman, R.H., Valencia,
antimicrobial susceptibility testing. In: Isenberg, T., Herrera, B., Quino, W., et al. (2008)
H.D. (ed.) Clinical Microbiology Procedures Tuberculosis diagnosis and multidrug resistance
Handbook, Vol 2. ASM Press, Washington DC, testing by direct sputum culture in selective
pp. 7111–7883. broth without decontamination or centrifugation.
Desjardin, L.E., Perkins, M.D., Wolski, K., Haun, S., Journal of Clinical Microbiology 46, 2339–2344.
Teixeira, L., Chen, Y., et al. (1999) Measurement Greco, S., Rulli, M., Girardi, E., Piersimoni, C. and
of sputum Mycobacterium tuberculosis Saltini, C. (2009) Diagnostic accuracy of
messenger RNA as a surrogate for response to in-house PCR for pulmonary tuberculosis in
chemotherapy. American Journal of Respiratory smear-positive patients: meta-analysis and
and Critical Care Medicine 160, 203–210. metaregression. Journal of Clinical Microbiology
Dinnes, J., Deeks, J., Kunst, H., Gibson, A., 47, 569–576.
Cummins, E., Waugh, N., et al. (2007) A Horne, D.J., Royce, S.E., Gooze, L., Narita, M.,
systematic review of rapid diagnostic tests for Hopewell, P.C., Nahid, P., et al. (2010) Sputum
the detection of tuberculosis infection. Health monitoring during tuberculosis treatment for
Technology Assessment 11, 1–196. predicting outcome: systematic review and
Dye, C., Watt, C.J., Bleed, D.M., Hosseini, S.M. and meta-analysis. The Lancet Infectious Diseases
Raviglione, M.C. (2005) Evolution of tuberculosis 10, 387–394.
control and prospects for reducing tuberculosis Hu, Y. and Coates, A.R.M. (1999) Transcription of
incidence, prevalence, and deaths globally. the stationary-phase-associated hspX gene of
JAMA: The Journal of the American Medical Mycobacterium tuberculosis is inversely related
Association 293, 2767–2775. to synthesis of the 16-kilodalton protein. Journal
Eisenach, K.D., Sifford, M.D., Cave, M.D., Bates, of Bacteriology 181, 1380–1387.
Jindani, A., Aber, V.R., Edwards, E.A. and tuberculosis: a systematic review [Review
Mitchison, D.A. (1980) The early bactericidal Article]. The International Journal of Tubercu-
activity of drugs in patients with pulmonary losis and Lung Disease 11, 485–495.
tuberculosis. American Review of Respiratory Millington, K.A., Innes, J.A., Hackforth, S., Hinks,
Disease 121, 939–949. T.S., Deeks, J.J., Dosanjh, D.P., et al. (2007)
Joloba, M.L., Johnson, J.L., Namale, A., Morrissey, Dynamic relationship between IFN-gamma and
A., Assegghai, A.E., Rusch-Gerdes, S., et al. IL-2 profile of Mycobacterium tuberculosis-
(2001) Quantitative bacillary response to specific T cells and antigen load. Journal of
treatment in Mycobacterium tuberculosis Immunology 178, 5217–5126.
infected and M. africanum infected adults with Mitchison, D.A. (1993) Assessment of new
pulmonary tuberculosis. International Journal of sterilizing drugs for treating pulmonary tubercu-
Tuberculosis and Lung Disease 5, 579–582. losis by culture at 2 months. American Review
Kennedy, N., Gillespie, S.H., Saruni, A.O.S., of Respiratory Disease 147, 1062–1063.
Kisyombe, G., McNerney, R., Ngowi, F.I., et al. Mitchison, D.A. (2006) Clinical development of anti-
(1994) Polymerase chain reaction for assessing tuberculosis drugs. Journal of Antimicrobial
treatment response in patients with pulmonary Chemotherapy 58, 494–495.
tuberculosis. Journal of Infectious Diseases Mitchison, D.A., Allen, B.W., Carrol, L., Dickinson,
170, 713–716. J.M. and Aber, V.R. (1972) A selective oleic acid
Kent, L., McHugh, T.D., Billington, O., Dale, J.W. albumin agar medium for tubercle bacilli.
and Gillespie, S.H. (1995) Demonstration of Journal of Medical Microbiology 5, 165–175.
homology between IS6110 of Mycobacterium Mukamolova, G.V., Turapov, O.A., Young, D.I.,
tuberculosis and DNAs of other Mycobacterium Kaprelyants, A.S., Kell, D.B. and Young, M.
spp. [published erratum appears in Journal of (2002) A family of autocrine growth factors in
Clinical Microbiology 1995 Nov 33(11), 3082]. Mycobacterium tuberculosis. Molecular Micro-
Journal of Clinical Microbiology 33, 2290–2293. biology 46, 623–635.
Kubica, G.P., Dye, W.E., Cohn, M.L. and Middle- Mukamolova, G.V., Turapov, O., Malkin, J.,
brook, G. (1963) Sputum digestion and Woltmann, G. and Barer, M.R. (2010)
decontamination with N-acetyl-l-cysteine-sodium Resuscitation-promoting factors reveal an
hydroxide for culture of mycobacteria. The occult population of tubercle bacilli in sputum.
American Review of Respiratory Disease 87, American Journal of Respiratory and Critical
775–779. Care Medicine 181, 174–180.
Lalvani, A. (2004) Counting antigen-specific T cells: Mulcahy, G.M., Kaminski, Z.C., Albanese, E.A.,
a new approach for monitoring response to Sood, R. and Pierce, M. (1996) IS6110-based
tuberculosis treatment? Clinical Infectious PCR methods for detection of Mycobacterium
Diseases 38, 757–759. tuberculosis. Journal of Clinical Microbiology
Levée, G., Glaziou, P., Gicquel, B. and Chanteau, 34, 1348–1349.
S. (1994) Follow-up of tuberculosis patients O’Sullivan, D.M., Sander, C., Shorten, R.J.,
undergoing standard anti-tuberculosis chemo- Gillespie, S.H., Hill, A.V.S., McHugh, T.D., et al.
therapy by using a polymerase chain reaction. (2007) Evaluation of liquid culture for
Research in Microbiology 145, 5–8. quantitation of Mycobacterium tuberculosis in
Li, L., Mahan, C.S., Palaci, M., Horter, L., murine models. Vaccine 25, 8203–8205.
Loeffelholz, L., Johnson, J.L., et al. (2010) Perrin, F.M., Woodward, N., Phillips, P.P., McHugh,
Sputum Mycobacterium tuberculosis mRNA as T.D., Nunn, A.J., Lipman, M.C., et al. (2010)
a marker of bacteriologic clearance in response Radiological cavitation, sputum mycobacterial
to antituberculosis therapy. Journal of Clinical load and treatment response in pulmonary
Microbiology 48, 46–51. tuberculosis. International Journal of
Marlowe, E.M., Novak Weekley, S.M., Cumpio, J., Tuberculosis and Lung Disease 14, 1596–1602.
Sharp, S.E., Momeny, M.A., Babst, A., et al. Pheiffer, C., Carroll, N.M., Beyers, N., Donald, P.,
(2011) Evaluation of the Cepheid Xpert MTB/ Duncan, K., Uys, P., et al. (2008) Time to
RIF assay for the direct detection of Myco- detection of Mycobacterium tuberculosis in
bacterium tuberculosis complex from respiratory BACTEC systems as a viable alternative to
specimens. Journal of Clinical Microbiology 49, colony counting. International Journal of
1621–1623. Tuberculosis and Lung Disease 12, 792–798.
Mase, S.R., Ramsay, A., Ng, V., Henry, M., Rachow, A., Zumla, A., Heinrich, N., Rojas-Ponce,
Hopewell, P.C., Cunningham, J., Urbanczik, R., G., Mtafya, B., Reither, K., et al. (2011) Rapid
et al. (2007) Yield of serial sputum specimen and accurate detection of Mycobacterium
examinations in the diagnosis of pulmonary tuberculosis in sputum samples by Cepheid
Xpert MTB/RIF assay – a clinical validation tuberculosis. Clinical Infectious Diseases 25,
study. PLoS ONE 6, e20458. 666–670.
Rouillon, A., Perdrizet, S. and Parrot, R. (1976) Thomsen, V.O., Kok-Jensen, A., Buser, M., Philippi-
Transmission of tubercle bacilli: the effects of Schulz, S. and Burkardt, H.J. (1999) Monitoring
chemotherapy. Tubercle 57, 275–299. treatment of patients with pulmonary tubercu-
Sai Priya, V.H., Anuradha, B., Latha Gaddam, S., losis: can PCR be applied? Journal of Clinical
Hasnain, S.E., Murthy, K.J. and Valluri, V.L. Microbiology 37, 3601–3607.
(2009) In vitro levels of interleukin 10 (IL-10) Tibayrenc, M. (2007) Encyclopedia of Infectious
and IL-12 in response to a recombinant Diseases: Modern Methologies. Wiley and
32-kilodalton antigen of Mycobacterium bovis Sons, Inc, Hoboken, New Jersey.
BCG after treatment for tuberculosis. Clinical Wallis, R.S., Perkins, M., Phillips, M., Joloba, M.,
Vaccine Immunology 16, 111–115. Demchuk, B., Namale, A., et al. (1998) Induction
Schluger, N.W. and Rom, W.N. (1994) Current of the antigen 85 complex of Mycobacterium
approaches to the diagnosis of active pulmonary tuberculosis in sputum: a determinant of
tuberculosis. American Journal of Respiratory outcome in pulmonary tuberculosis treatment.
and Critical Care Medicine 149, 264–267. Journal of Infectious Diseases 178, 1115–1121.
Shleeva, M.O., Bagramyan, K., Telkov, M.V., Wallis, R.S., Doherty, T.M., Onyebujoh, P., Vahedi,
Mukamolova, G.V., Young, M., Kell, D.B., et al. M., Laang, H., Olesen, O., et al. (2009) Bio-
(2002) Formation and resuscitation of ‘non- markers for tuberculosis disease activity, cure,
culturable’ cells of Rhodococcus rhodochrous and relapse. The Lancet Infectious Diseases 9,
and Mycobacterium tuberculosis in prolonged 162–172.
stationary phase. Microbiology 148, 1581–1591. Woods, G.L. and Witebsky, F.G. (1995) Myco-
Somoskovi, A., Kodmon, C., Lantos, A., Bartfai, Z., bacterial testing in clinical laboratories that
Tamasi, L., Fuzy, J., et al. (2000) Comparison of participate in the College of American
recoveries of Mycobacterium tuberculosis using Pathologists’ Mycobacteriology E survey:
the automated BACTEC MGIT 960 system, the results of a 1993 questionnaire. Journal of
BACTEC 460 TB system, and Lowenstein– Clinical Microbiology 33, 407–412.
Jensen medium. Journal of Clinical Microbiology World Health Organization [WHO] (2009) Treatment
38, 2395–2397. of Tuberculosis: Guidelines, 4th edn. World
Steingart, K.R., Henry, M., Ng, V., Hopewell, P.C., Health Organization, Geneva, Switzerland.
Ramsay, A., Cunningham, J., et al. (2006a) WHO (2010) Monitor TB Case Detection and
Fluorescence versus conventional sputum Treatment. Management of Tuberculosis:
smear microscopy for tuberculosis: a systematic Training for Health Facility Staff, 2nd edn. World
review. The Lancet Infectious Diseases 6, 570– Health Organization, Geneva, Switzerland.
581. Wu-Hsieh, B.A., Chen, C.-K., Chang, J.-H., Lai,
Steingart, K.R., Ng, V., Henry, M., Hopewell, P.C., S.-Y., Wu, C.H.H., Cheng, W.-C., et al. (2001)
Ramsay, A., Cunningham, J., et al. (2006b) Long-lived immune response to early secretory
Sputum processing methods to improve the antigenic target 6 in individuals who had
sensitivity of smear microscopy for tuberculosis: recovered from tuberculosis. Clinical Infectious
a systematic review. The Lancet Infectious Diseases 33, 1336–1340.
Diseases 6, 664–674. Yuen, K.Y., Chan, K.S., Chan, C.M., Ho, B.S., Dai,
Telzak, E.E., Fazal, B.A., Pollard, C.L., Turett, G.S., L.K., Chau, P.Y., et al. (1993) Use of PCR in
Justman, J.E. and Blum, S. (1997) Factors routine diagnosis of treated and untreated
influencing time to sputum conversion among pulmonary tuberculosis. Journal of Clinical
patients with smear-positive pulmonary Pathology 46, 318–322.
11 Modelling Responses to Tuberculosis
Treatment
G.R. Davies
Institute of Translational Medicine, University of Liverpool, UK
11.1 Introduction and this hierarchy of end points relates

directly to the historical development of
Renewed interest in drug development for tuberculosis treatment (Fox et al., 1999).
improved treatment of tuberculosis has The earliest trials of monotherapy with
recently led to a re-evaluation of existing streptomycin (SM), para-amino-salicylic acid
measures of treatment response. While it is (PAS) and isoniazid (INH) provided dramatic
widely accepted that the definitive outcome evidence that such agents could prevent early
for tuberculosis trials is the absence of death from the disease (MRC, 1948, 1952).
bacteriological relapse up to 2 years after the Typically, the case fatality rate in the com-
cessation of treatment, a key debate in the parator arm of these studies was 30% during
field centres on the extent to which inter- the first few months of the trial. These
mediate surrogates can capture treatment desperate circumstances provided investi-
effects on this reference end point and which gators with an unequivocal and statistically
statistical representation offers optimal powerful end point on which to establish the
power for their use in clinical trials. This efficacy of such drugs even prior to com-
chapter gives an overview of the challenges pletion of treatment by the majority of
this problem poses for the development of participants in the trial. The same clinical
new drugs, summarizes the state of current situation has recently returned with the
efforts to overcome them and gives some advent of extensively drug-resistant tubercu-
insight into future developments. losis (XDR-TB), which may carry an even
higher case-fatality rate (Gandhi et al., 2010).
Even when such unambiguous results
11.2 Conceptualizing Treatment could be obtained quickly, however, the
Response effects of treatment were noticed to be time
dependent. Though SM, for instance, pro-
Response to treatment for tuberculosis can be vided durable benefits, after 5 years of
conceptualized as a dependent chain of pos- follow-up these were marginal by comparison
sible composite clinical and laboratory events with the early results in these trials (Fox et al.,
(Lienhardt and Davies, 2010) (Fig. 11.1). The 1954). Detailed microbiological study of the
most appropriate end points for quantifying evolution of Mycobacterium tuberculosis strains
the success or failure of a particular regimen during these trials revealed the rapid emer-
can therefore depend on its overall potency, gence of resistance in a significant proportion

148 G.R. Davies
Hierarchy of Phase III endpoints
Increasing potency of regimens
Remain/revert Remain/revert Revert to

Die from to culture to culture culture positive
Survive long
tuberculosis positive with positive without without new
term without
during new resistance new resistance resistance
disease
treatment during at end of after
treatment treatment treatment
Treatment Treatment
Death Relapse Cure
failure failure
‘Bactericidal activity’ ‘Sterilizing activity’

Fig. 11.1. Chain of outcomes in Phase III trials of anti-tuberculosis treatment.
of patients, with subsequent failure of therapy Hence, the distinct ways in which
(Mitchison, 1950). It was soon widely acknow- tuberculosis treatment can fail offer some
ledged that monotherapy could not be relied empirical insight into the pharmacodynamic
on to achieve a permanent cure and interest mechanisms underlying successful therapy.
began to focus on the capacity of multiple In particular, a combination of drugs must be
drugs to prevent the emergence of resistance collectively potent enough to result in a
during therapy (MRC, 1950, 1953). While significant net decrease in bacillary load in
drug combinations with even modest potency order to prevent death of the host and reduce
could achieve this, they still required the probability that resistance mutation
prolonged therapy of up to 24 months to events will occur. Both these goals are depend-
ensure stable cure (MRC, 1973). ent to some extent on the speed with which
Once more potent agents became avail- particular drugs can achieve this. Historically,
able, offering the possibility of shortening the this property of a regimen has often been
duration of therapy, a new problem was referred to as ‘bactericidal activity’ (Mitchison,
identified. Though drugs such as rifampicin 1992). It is also clear, however, that bactericidal
(RIF) and pyrazinamide (PZA) halved the activity alone may not suffice for a stable
duration of treatment at a stroke, if the dur- cure. The ability of a combination regimen to
ation of the regimen was too short, apparent reduce bacillary numbers in itself may
cure was soon followed by an unacceptably effectively prevent the emergence of resist-
high rate of relapse (EAMRC, 1974, 1981). ance, but if it cannot ultimately eliminate
Characteristically, however, the organisms them (or at least reduce them below some
responsible were not drug resistant, giving threshold at which the host’s immune system
rise to the concept of subpopulations of can complete the task) then relapse will occur.
organisms capable of persisting during This capability has typically been referred to
therapy in the absence of genotypic resistance as ‘sterilizing activity’ and understanding
(Mitchison, 1979). and quantifying this phase of the pharma-
Modelling Responses to Tuberculosis Treatment 149
codynamics of anti-tuberculosis drugs has surrogacy is clearly a matter of degree, with

attracted much research interest in recent the dual consequence that evaluation is a
years (Davies, 2010). From a practical per- cumulative process across many trials and
spective, it must currently be equated with ultimately unlikely to provide theoretically
relapse rates post-treatment and the extent ‘perfect’ surrogate end points capable of com-
to which earlier measures reflect this pletely substituting for any reference end
remains uncertain. Whether bactericidal and point (Prentice, 1989). Secondly, a critical
sterilizing activities should be conceived of as distinction has to be made between surrogacy
simply ends of a spectrum created by at the level of individual participants and that
increasing drug potency or as qualitatively of trial cohorts. While the former may provide
distinct, arising from the nature of the targets a useful clinical predictor of response, it is
of the different drug classes, has also usually the latter which is of interest in drug
remained an open question. development. These aspects of a surrogate
end point are not always closely linked in
practice and are, in fact, logically and
11.3 The Problem of Surrogate statistically independent (Baker and Kramer,
End Points 2003), which again suggests that evaluation
across many trials should be preferred to
Early prediction of efficacy or toxicity in large cohorts of individuals within a single
Phase I/II trials is a generic problem in most trial. Another important consideration is that
areas of clinical trial activity. The challenge is even biomarkers which initially show some
to select surrogate end points correctly from empirical statistical association will be un-
among candidate biomarkers which can likely to be taken up if they lack a plausible
reliably capture changes in a given clinical or causal explanation for the relationship. In-
reference end point appropriate for Phase III creasingly, this must be framed in terms of
trials (Burzykowski et al., 2005). Typically, specific biological mechanisms of action and
such reference end points occur infrequently pharmacodynamic considerations for the
enough that a very large sample size would class of drugs concerned.
be required to achieve adequately powerful
comparisons between treatments, and they
may require prolonged follow-up to obtain. 11.3.1 Bacteriological versus non-
Hence, they are usually not appropriate for bacteriological biomarkers
the purposes of proof-of-concept, dose find-
ing and defining optimum combinations of Bacteriological data have always formed part
drugs in early development. A practically of the definition of composite ‘clinical’ end
useful surrogate end point can be charac- points since the earliest trials. Since the
terized by the following (De Gruttola et al., pharmacodynamic target of current anti-
2001): tuberculosis drugs is to kill viable M.
tuberculosis organisms, measures of bacillary
• a biologically plausible relationship with
elimination based on laboratory culture
the reference end point, preferably mech-
would be expected a priori to have some
anism based
usefulness as surrogate end points. As the
• useful statistical properties
potency of regimens has improved, these
• successful evaluation against the reference
have received more emphasis and priority
end point.
over other aspects of response. In general,
Recently, a general statistical framework has because development of anti-tuberculosis
been proposed for understanding and drugs has focused on pulmonary disease,
evaluating the relationship between proposed where repeated clinical samples are readily
surrogate and reference end points (Buyse et accessible, data on this approach come almost
al., 2000). In particular, two key concepts have exclusively from studies of sputum.
been clarified by these advances. Firstly, That bacteriological end points at an
150 G.R. Davies
early stage of treatment could be used to 11.3.2 Empirical versus model-based

predict the ultimate outcome was proposed approaches
initially on the basis of the aggregated data
from the extensive series of short-course Traditionally, bacteriological response has
chemotherapy trials conducted by the British been represented as culture positivity at fixed
Medical Research Council (Aber and Nunn, time-points as treatment progresses (Fox et
1978; Mitchison, 1996). Data from these trials al., 1999). The structure of short-course chemo-
(12 trials comprising 6974 patients in 49 therapy regimens with an initial intensive and
treatment arms) were recently retrieved and simplified continuation phase tended to focus
re-analysed at the trial level and this analysis the attention of trialists and, later, tuberculosis
confirmed that, particularly for RIF-based programme managers on results at 2 and
regimens, there was a moderately strong subsequently 5 and 6 months. The choice of
relation between 2- and 3-month culture these occasions was based arbitrarily on con-
results and poor treatment outcome (R2 was venience and clinical considerations. Though
0.67 and 0.46, respectively) (Phillips and simple to interpret, when expressed as a
Fielding, 2007, 2008). However, the relation- simple proportion, this approach is statisti-
ship appeared less strong over all the regi- cally inefficient (Fedorov et al., 2009; Yoo,
mens included in the analysis and there must 2009). An additional problem is that as the
therefore remain some caution over the potency of regimens improves and the rate of
generalizability of this observation to future culture conversion at 2 months in the com-
combination regimens containing drugs with parator arm increases, statistical power to
new mechanisms of action. detect treatment differences diminishes. Such
Given the imperfections of historical an empirical approach employs no specific
bacteriological data, several other techniques model of the pharmacodynamics of treatment.
and biomarkers have been developed and Hence, it is insensitive both quantitatively in
proposed as a means of supplementing or the sense that it will only reliably detect large
supplanting them for use as surrogate end treatment effects and qualitatively in that
points. Those most advanced have been differences in the pattern of bacillary
reviewed recently (Perrin et al., 2007; Wallis et elimination due to different classes of drugs
al., 2009) and can be considered broadly as cannot be detected using only a single or very
organism or host based. Organism-based sparse measurements.
biomarkers include new methods of Typically, however, samples for culture
automated broth-based culture and nucleic have often been obtained on at least a monthly
acid-based techniques (discussed further basis in Phase III trials and on a weekly basis
below), while relevant host-based biomarkers throughout the intensive phase in Phase II
include serial peripheral blood IFN- release trials. The analysis of such repeated measure-
assays and transcriptomic approaches. To ments has frequently been expressed as a
date, very few of these alternative biomarkers series of univariate analyses with or without
have undergone any large-scale evaluation correction for multiplicity and usually
against the current standard bacteriological prioritizing the 2-month result as primary.
approach. Whether the majority of these This is clearly not faithful to the correlated
novel techniques will find any secure place as nature of these data in time, and in recent
surrogate end points in the development years it has become more common to make
process remains to be seen. While it seems use of more modern survival techniques
currently unlikely that any single one will (Holtz et al., 2006; Conde et al., 2009). By
replace sputum bacteriology in the near representing bacillary elimination as a hazard
future, it may be that some could be useful of culture conversion over time, this approach
adjuncts as covariates to such data, perhaps implicitly adopts a non-parametric model of
providing better insight into additional pharmacodynamics which makes use of all
bacillary or host determinants of response the available data and hence has greater
that influence sterilizing activity. power to detect treatment effects. Within a
regression framework, it also becomes ing use of it have become the standard design
simpler to adjust the model for important for Phase IIa studies for all new agents
covariates, and logistic regression models are (Donald et al., 2003).
also increasingly used for this purpose, using EBA studies have provided useful
the 2-month culture conversion end point insights into the pharmacology of many
alone (Burman et al., 2006). classes of drugs. Since they can be conducted
Within the context of modern drug with as few as 10–20 patients per arm, they
development, the continued use of these are well suited to initial proof-of-concept and
empirical approaches in tuberculosis appears dose-finding studies. For INH, which exhibits
anomalous. In the early stages of clinical the highest EBA (approximately –0.5 log10
development, it is critical to refine and colony-forming units (CFU)/ml/day), a com-
expand information about drug response and plete dose–response curve can be obtained
this is often achieved more effectively using which clearly identifies the minimum and
continuous, time-dependent measures and maximum effective doses, and activity can be
modelling and simulation methodology related to pharmacokinetic exposure and
(Zhang et al., 2006). This ‘learning’ approach acetylator status (Donald et al., 1997, 2004).
is to be contrasted with the shift in focus in For rifamycins, EBA studies have drawn
Phase III to a ‘confirming’ paradigm under attention to incomplete dose-finding data
which unambiguous and simple response (Diacon et al., 2007) and differences in activity
measures are collected when or after treat- across members of the class according to
ment is completed (Sheiner, 1997). A strictly differences in the free drug concentration
empirical approach to response in Phase II (Sirgel et al., 2005). The technique has also
trials is a significant obstacle to implementing been used to compare the relative potencies
the learning paradigm in tuberculosis, mak- of antimycobacterial fluoroquinolones and
ing the achievement of the goals of early identify equivalent dose regimens (Johnson et
development difficult or impossible at reason- al., 2006). However, the activity of INH is
able sample sizes. clearly time dependent, waning after only 2–5
days of treatment, while some classes of
drugs such as PZA, aminoglycosides and
11.4 Early Bactericidal Activity nitroimidazopyrans have delayed or very
weak EBA (Jindani et al., 1980; Donald et al.,
An alternative method of assessing drug 2001, 2002; Diacon et al., 2010), suggesting
activity in tuberculosis is based on explicitly that EBA may represent imperfectly the
quantitative bacteriological methods. By capability of drugs to achieve long-term cure.
standardizing sputum preparation and using Since no ethics committee has agreed to
solid media highly selective for mycobacteria, longer than 14 days of monotherapy with
a reproducible count of viable colonies in anti-tuberculosis drugs, for reasons of patient
sputum specimens can be obtained and fol- safety and the possible emergence of resist-
lowed over time (Mitchison, 1950; Mitchison ance, this is a major limitation of the approach.
et al., 1971). The first study to use this serial The EBA methodology is highly depend-
sputum colony-counting approach followed ent on the variance of the measurements on
patients on individual and pairs of drugs which it is based. Since this ranges more than
over the first 14 days of therapy (Jindani et al., twofold over different centres that have
1980). Using an analysis of variance approach attempted such studies (Sirgel et al., 2000), the
based on the linear slopes of the decline in sample sizes typically proposed can some-
colony counts summarized over different times be unduly optimistic. Both laboratory
time periods for individual patients, clear experience with the technique and careful
differences in the potency of the drugs patient selection are important, since efforts
studied could be demonstrated. Such measure- to correct for sputum quality have not been
ments have since become known as ‘early successful (Sirgel et al., 2001). Furthermore,
bactericidal activity’ (EBA) and studies mak- variance of the measurements increases with
152 G.R. Davies
time, which diminishes the efficiency of the stein et al., 2002) and pharmacokinetics for
technique to detect delayed treatment effects. many years (Sheiner et al., 1977). The basic
The traditional statistical approach to analysis principle is that instead of computing sum-
of EBA data is based on computation of a mary estimates of the parameters of interest
summary statistic for each patient based on for each individual, the mixed effects
their measurements in a series of arbitrary approach focuses on obtaining estimates of
time intervals (e.g. days 0–2, 2–7, 7–14, etc.), these parameters for the population as a
followed by analysis of the variance of these whole (the ‘fixed’ effects), while representing
statistics across treatment groups. This is a the variability of those parameters with a
simple and effective way to circumvent the probability distribution whose parameters in
problem of correlation between repeated turn are to be estimated (the ‘random’ effects).
measurements in study subjects (Matthews et Thus, estimates of a particular individual’s
al., 1990). However, it embodies strong parameters are not obtained directly using
assumptions of piecewise linearity within the this method (but can be obtained indirectly as
selected time intervals and begs the question model predictions) and if, as is usually the
of the pharmacodynamic significance of those case in clinical trials, these individual values
intervals. A particular problem is that the are not the primary focus of the analysis, the
summary statistics absorb the inter- mixed effects representation of the data is
individual variability in colony counts as well both parsimonious and efficient. The total
as the intra-individual variability, which com- variance is partitioned in the model between
promises significantly their power to detect inter and intra-individual components and it
treatment effects. This is exacerbated when is the explicit modelling of inter-individual
combination regimens are evaluated, because variability that lends the technique its power
they inevitably incorporate INH, which has when applied to repeated measures data. The
the highest variance in its EBA among all methodology can be applied to any type of
drugs studied to date. Since the correlation outcome measure, though in pharma-
between early EBA measures (such as EBA0-2 cokinetic–pharmacodynamic analyses it is
0-5) and the baseline measurement is fairly typically applied in a continuous context. To
high (usually ~0.5–0.7), one way to improve date, no published examples of applying this
this situation would be use the baseline approach to EBA data have appeared.
measurement as a covariate in the analysis in
order to account for inter-individual varia-
bility in the counts (analysis of covariance, or 11.5.1 Extended serial sputum
ANCOVA). Interestingly, many study proto- colony-counting studies
cols incorporate 1 or 2 pre-treatment samples,
but these measurements are never used in the The phenomenon of delayed or unexpectedly
analysis. A number of linear summary stat- weak EBA, specifically the anomalous case of
istics based on the ANCOVA approach have PZA and recent discrepant results in the
been proposed (Frison and Pocock, 1992, development programme of TMC207 (beda-
1997), which could be used to take advantage quiline), in the context of other evidence of
of this additional data. possibly strong sterilizing activity, has raised
concerns about the interpretation of such
studies (Rustomjee et al., 2008a; Diacon et al.,
11.5 Mixed Models 2009). Since there is no possibility of extending
EBA studies of monotherapy beyond 14 days,
An alternative approach to representing one proposal has been to study instead
multiple hierarchical levels of variability in combination regimens over a longer period in
statistical modelling is that of mixed effects an attempt to define more clearly the pattern
modelling (Davidian and Giltinan, 1995). of response and give better assurance that this
Though relatively novel in medical statistics corresponds to long-term outcome.
(Diggle et al., 2002), this idea has been used The first such study compared two dis-
extensively in the fields of education (Gold- tinct combination regimens in a cohort of 100
patients in Nairobi, Kenya, over the first possible to show that, though the two regi-
month of therapy, sampling their sputum mens in the study did not differ in terms of
at days 0, 2, 7, 14 and 28 (Brindle et al., the rate of early phase elimination, the SHRZ
1993). Approximately half were receiving regimen accelerated late-phase activity signifi-
streptomycin-isoniazid-thiacetazone (SHT), a cantly. Given the size of the treatment groups
‘standard’ regimen, and half streptomycin- and the already established superior efficacy
isoniazid-rifampicin-pyrazinamide (SHRZ), of the SHRZ regimen in prior clinical trials,
a ‘short-course’ regimen. The initial analysis this result was of great interest. Lastly, though
of these data concluded that response to HIV-positive subjects in this study had lower
treatment was similar in both groups and did baseline bacillary loads, their status did not
not differ by HIV status (Brindle et al., 2001). affect the rate of bacillary elimination from
A recent review and re-analysis of these data their sputum. A great deal of the variability in
applied instead a mixed effects modelling the model was accounted for by random
approach, with different results (Davies et al., effects in the intercepts rather than the rates
2006a). Firstly, the profile of response was of the two phases of decay.
noticed to be unequivocally non-linear, with The biphasic nature of elimination in this
an early phase of rapid elimination of bacilli model raises interesting questions of interpret-
lasting up to 7 days succeeded by a later ation. The same profile of response could, in
phase where elimination proceeded at less principle, be due to two different scenarios:
than 20% of the initial rate, continuing until simultaneous elimination of distinct and
the end of the study period (Fig. 11.2). heterogeneous subpopulations of organisms
Secondly, using a biphasic model with two present prior to the commencement of therapy
distinct phases of exponential decay, it was or rapid adaptation to drug action and
Fig. 11.2. Serial sputum colony-counting data and fitted model predictions from three studies using
similar short-course regimens. The figures refer to the estimated rates of late-phase elimination of
organisms (log CFU/mL/day).
154 G.R. Davies
selection from a single more homogeneous masivan et al., 2005). Once again, HIV did not
initial population. These two scenarios are affect the rate of elimination materially, but a
not identifiable in this basic model due to the simple measure of radiological extent of
lack of information with which to classify disease was associated with lower activity.
organisms into the two putative subpopu- The results of this study are in agreement
lations and hence estimate any rate of with those obtained under a similarly in-
transition between them. In addition, extrapo- tensive sampling scheme in another study of
lation and simulation of the model predictions substitution of moxifloxacin for ethambutol
suggested that even organisms undergoing in Brazil (Conde et al., 2009). This slightly
slow elimination were not likely to comprise larger study relied on a survival analysis of
the subpopulation of organisms capable of the rate of culture conversion, which sug-
causing relapse after treatment was stopped. gested significantly faster conversion in the
Given the findings of this analysis, moxifloxacin arm. Another multicentre study
further evaluation of the biphasic model for of the same regimens used 2-month culture
response to tuberculosis treatment was conversion as its primary end point and con-
carried out. Conditional on the estimates of cluded that there was no difference between
the model parameters estimated in the the regimens, despite suggestions of more
original study, features of the study design rapid culture conversion in the moxifloxacin
were optimized using an analysis of the arm at earlier time points (Burman et al.,
Fisher’s information matrix of the model and 2006).
an approximate linearization method for com- A third observational pharmacokinetic–
putation of sample size (Davies et al., 2006b). pharmacodynamic study based on a colony-
This suggested that greatly improved preci- counting design of two balanced blocks of
sion of studies based on such a model could five sampling points each was completed
be achieved at sample sizes as low as 40–60/ recently in Thailand (Davies et al., 2008).
arm, provided that the total number of Though baseline bacillary load was lower
sampling points be doubled to 10 and the than in the two African studies, the rate
duration of sampling be increased to include parameters of the model estimated from this
the entire intensive phase of 56 days. data set were similar. Again, HIV status did
Subsequently, a study employing a simi- not appear to influence the rate of bacillary
lar design was conducted in South Africa elimination. In this study, INH exposure was
(Rustomjee et al., 2008b). This study evaluated related significantly to early-phase elimin-
substitution of three different fluoro- ation in the presence of the other companion
quinolones (ofloxacin, moxifloxacin and gati- drugs, but there was no apparent relationship
floxacin) for ethambutol in the first-line between RIF or PZA exposure and late-phase
regimen. The design thus comprised four elimination.
arms of 50 patients each and obtained sputum Given a particular study design, the
samples at 0, 2 days and then weekly until relative performance of the possible analytical
day 56. Analysis of the colony-counting data approaches has also been evaluated using
using the biphasic mixed effects model trial simulation techniques (Davies, 2009).
separated the four arms into two groups on Assuming that the biphasic model of elimin-
the basis of late-phase elimination, with the ation is a reasonable description of the data,
control and ofloxacin arms distinct from the the late phase of the model can be approxi-
moxi-/gatifloxacin arms. The estimated para- mated by a linear mixed effects model from
meters for the control regimen (HRZE) were day 7 onwards. This approach has been used
similar to those observed in the previous to simulate large numbers of trial scenarios
analysis. In the moxi-/gatifloxacin arms, the under the alternative hypothesis of a real
elimination rate was approximately 24% treatment effect and the resulting data sets
faster than the former, a highly statistically analysed using the different possible stat-
significant result compatible with the rank istical methods to compute the power of each
order of activity obtained in an in vitro model approach. The results of this study appear to
of stationary phase M. tuberculosis (Para- bear out the limited experience to date. An
analysis of covariance approach based on the 11.5.2 Liquid culture

quantitative information in the colony counts
appeared to be uniformly the most powerful In the most recent clinical trials, automated
method under all scenarios. However, non- liquid culture techniques have increasingly
and semi-parametric survival methods based been used alongside more traditional solid
on culture conversion also performed quite media (see Chapter 3). The technique may
well and were reasonably robust to interval have a lower limit of detection with respect to
censoring caused by sparse sampling schemes. bacillary load and may, in addition, enable
Not unexpectedly, restricting the analysis to the resuscitation of organisms in altered
proportions positive at 2 months alone reduced growth states that lack the capacity to grow
the expected power of these designs severely, on solid media, particularly later in treatment.
generally doubling the sample size required Another attractive feature is that, in vitro at
to detect a given treatment effect size (Fig. least, time to positivity (TTP) in the liquid
11.3). culture system is proportional to the load of
Taken together, these early data on organisms inoculated. This inherently quan-
colony-counting techniques support its fur- titative aspect of liquid culture data raises the
ther evaluation and raise the possibility of possibility of another measure of bacillary
efficient, possibly factorial Phase II designs load that is not as labour-intensive to obtain
which may be capable of sequential or fully as colony counts. Unfortunately, a number of
adaptive dose and companion drug opti- alternative culture systems exist with dif-
mization in humans. ferent detection systems and protocols for
LOD = 1 log10
1.0
0.8
0.6
1–`
0.4
[2
0.2
Cox
SLOPE
SLAIN
0.0
0 20 40 60 80 100 120
N per arm
Fig. 11.3. Simulated power curves for four different analytical approaches over three different treatment
effect sizes: a 2 test of culture conversion at 2 months, Cox proportional hazards modelling, the SLOPE
summary statistic based on ANOVA and the SLAIN summary statistic based on ANCOVA. Within each
panel colour refers to the effect size used in the simulations: mid-grey = 0.015; light grey = 0.020; black =
0.025. The dotted horizontal lines correspond to 80% and 90% power.
156 G.R. Davies
processing of the samples, which can make has centred on the application of molecular
interpretation of existing data problematic. methods to try to overcome this limitation,
When used simply as an alternative to since this technology offers potentially greater
solid media for the purposes of evaluating scope for improved sensitivity. Extensive
culture conversion at specific time points, studies under in vitro conditions have identi-
liquid culture tends to result in higher culture fied a tractable number of genes implicated in
positivity rates overall, which persist later in responses to hypoxia, nutrient limitation and
the course of treatment. Paradoxically, this drug pressure (Butcher, 2004; see Chapter 12).
can, in fact, result in a loss of sensitivity of a Several of these have also been identified as
simple analysis of proportions due to a undergoing upregulation in sputum during
reduction in the culture conversion rate in the treatment in humans (Garton et al., 2008; see
comparator arm, which is a critical determin- Chapter 13). Such genes could conceptually
ant of its power. However, the prolonged be used either as an alternative quantitative
period over which culture positivity persists measure of total bacillary load or as a means
in liquid media is generally an advantage for of monitoring the changing metabolic state of
survival approaches, provided that sampling the whole population of organisms over time.
persists long enough to capture a large Most studies of this approach have addressed
enough number of culture conversion events. the former application. For this purpose,
The only published analyses of liquid cul- DNA hybridization has not proved to be a
ture data have focused mostly on its use for useful measure, since it persists long after the
prediction of solid culture results. In a recent destruction of viable organisms. mRNA
example, it was demonstrated that very short reflects ongoing transcription better and is
baseline TTP was associated with delayed believed to continue constitutively at low but
culture conversion at 2 months and with sub-
measurable levels even in stationary phase
sequent relapse of disease (Hesseling et al.,
bacilli.
2010). The radiological extent of disease was
Several studies have reported results
also an important covariate in this analysis. A
with mRNA amplification from sputum dur-
previous analysis by the same investigators
ing treatment. One recent study compared
proposed a response ratio based on TTP
PCR of the DNA IS6110 element with RT-PCR
measurements obtained during the first 2
of mRNA for antigen 85B against culture
weeks as an alternative means of predicting
conversion (Mdivani et al., 2009). The former
poor treatment outcome (Pheiffer et al., 2008).
remained persistently positive late into
However, in this study, the positive predictive
value of this ratio was only 14% and the treatment, suggesting that it would be
results appeared to depend on the timing of unlikely to be a useful predictor of treatment
sampling during treatment. failure. Antigen 85B mRNA, however, declined
More extensive data on changes in TTP in tandem with culture results, becoming
during treatment are available from recent undetectable at a similar rate (see Chapter
clinical trials, but much remains unpublished 12).
to date. In a recently completed clinical trial Another study evaluated RT-PCR of
of adjuvant vitamin D treatment, extensive multiple targets (fbpB, hspX, icl1 and
liquid culture data were collected over the rrnA-P1) in the context of response to differ-
first 2 months of therapy. Over time, these ent single and combination drug regimens (Li
data appeared to follow a linear decline in et al., 2010). Expression of hspX and icl1, both
TTP and were adequately modelled using a relating to proteins well-known to be
linear mixed effects approach (Martineau et upregulated in the stationary phase, cor-
al., 2011). related well with decline in bacillary load,
though the contrasting sensitivity of the
amplification of the different targets appeared
11.5.3 Molecular markers to be responsible for variable calibration
against rates of culture conversion. The
Given the potentially altered growth charac- analysis of all of these studies relied on
teristics of bacilli in vivo, considerable interest empirical statistical techniques such as cor-
relation and repeated measures analysis of some of the inter-individual variability in

variance and did not employ a modelling treatment response. Recent advances in
approach. None of these studies clearly animal and in vitro models of tuberculosis
expressed the relative variability of the therapy have established concentration
mRNA measurements against culture results response relationships successfully and the
and it is difficult to comment on their nature of the relevant pharmacokinetic–
suitability for use in clinical trials without pharmacodynamic index using dose
this information. However, the ability to fractionation methods (Jayaram et al., 2003,
examine qualitatively different aspects of M. 2004; Gumbo et al., 2004, 2007). So far, in
tuberculosis physiology simultaneously is an clinical studies this has only been clearly
attractive aspect of this approach and could demonstrated for INH, a drug with a strong
give some insight into differing patterns of pharmacokinetic–pharmacodynamic relation-
pharmacodynamics induced by various ship in EBA studies and highly polymorphic
classes of drugs. pharmacogenetic determinants of metabolism,
resulting in a wide range of exposures. To
date, this goal has not been reproduced con-
11.6 Adjuncts to Modelling vincingly for any other TB drug, particularly
Responses and Joint not those with known or putative sterilizing
Modelling activity (Davies and Nuermberger, 2008).
Several possible reasons for this have been
Several studies have now observed that the advanced: lack of concomitant information
performance of quantitative models of treat- about the susceptibility of individual organ-
ment response is frequently enhanced by the isms in terms of their minimum inhibitory
addition of other important covariates which concentration (MIC), assay and sampling
capture some additional aspect of pharmaco- variability in the pharmacokinetic assays,
dynamics, whether this reflects factors related narrow dose range and limited sample size of
to the host or the organism. Clearly, correctly the relevant studies.
identifying and incorporating this infor- The extent of variability in the charac-
mation into such models is critical to their teristics of M. tuberculosis in individual
success. patients remains uncertain. Though animal
Radiological measures of severity of models support the use of area under the
disease at baseline have often been identified curve (AUC)/MIC as a general index of treat-
as independent predictors of disease outcome ment success for many tuberculosis drugs,
since the earliest clinical trials (MRC, 1948). usually only a limited number of strains are
However, precisely which features are evaluated. There is currently no convincing
important has not been clearly established evidence that the different natural lineages of
and various approaches have been developed M. tuberculosis have any impact on treatment
to give an overall measure of severity. The success (Nahid et al., 2010). In the absence of
proportion of lung involvement, especially genotypic high-level resistance, the variance
when expressed as bilateral or unilateral in the in vitro distribution of MIC90s for INH
disease, has been a consistent predictor of and RIF is less than tenfold (Schön et al., 2009),
poor response (Davies et al., 2006a; Rustomjee but since no human studies have yet evaluated
et al., 2008b). The only qualitative feature it extensively as a covariate, no data exist as to
with similar predictive power appears to be its predictive value. Another approach has
the presence or absence of cavitation (Burman been to attempt to quantify phenotypic drug
et al., 2006; Dorman et al., 2009). Both these tolerance by studying the time-kill properties
features are usually correlated with high of patient strains on re-exposure to drug
sputum bacillary load and hence help to pressure in liquid culture systems. One small
explain the variability of this measure. study claimed to show that this could be
Where pharmacokinetic information is related to rapidity of treatment response
available, it may be expected that measure- (Wallis et al., 1999). Observations that
ments of drug exposure could also explain stationary phase bacilli exhibiting phenotypic
158 G.R. Davies
drug tolerance frequently accumulate However, the rapid growth in infor-

droplets of triacylglycerol during changes of mation about all aspects of the biology of M.
intermediary metabolism in vitro (Garton et tuberculosis has resulted recently in the
al., 2002) have prompted investigations of capability to study the total transcriptome
whether staining with neutral lipid dyes and proteome dynamically under different
could be used to discriminate and perhaps conditions in vitro, and even in limited
quantify different subpopulations of bacilli in numbers of clinical samples. These advances
vivo (Garton et al., 2008). Independent have enabled the construction of a complete
information of this kind could be used to metabolic map of M. tuberculosis (http://tbcyc.
improve selection among the models referred tbdb.org/MTH37RVV) and the development
to above, which cannot currently discriminate of related dynamic systems models of these
between the simultaneous and evolutionary processes (Jamshidi and Palsson, 2007; Fang
models of bacillary heterogeneity. et al., 2009), which have enabled global
In addition to the identification of identification of critical metabolic subsystems
covariates for the models of response, another which could be targeted by new drugs.
statistical approach gaining increasing cur- Impressive as this is, the local mechanisms of
rency is that of the joint modelling of more action of many classes of drugs have already
than one response (Diggle et al., 2008). Con- been clearly defined and the critical com-
ceptually, this involves simultaneous model- ponents of the implicated pathways can some-
ling of the joint distribution of two distinct times be quantified as biomarkers (Raman et
outcome variables, and in practical terms has al., 2005, 2009; Wang and Marcotte, 2008).
usually involved combining clinical data These new techniques could be exploited to
expressed as survival times and repeated provide considerable supplementary infor-
measurements of some continuous biomarker mation about drug response. How best to
which is related to survival. Provided that the incorporate such information into pharma-
biomarker does explain an important pro- cokinetic–pharmacodynamic models has
portion of the variability in the other outcome, attracted recent attention under the rubric of
this can lead to more precise inferences. For ‘disease modelling’ and overlaps with similar
example, joint modelling of the profiles of developments in systems biology where the
culture conversion could be considered idea of ‘multi-scale modelling’ has facilitated
alongside some other meaningful measure of a flexible approach to incorporating mol-
host response; for example, T cell ELISpot ecular mechanisms into the lowest level of
measurements or measurements of mRNA hierarchical representations of biological
transcripts for specific bacillary genes. structures (Young et al., 2008). This kind of
approach will certainly help to integrate
current scientific knowledge derived from
11.6.1 Mechanistic modelling related disciplines and, by making use of a
approaches simulation strategy, help to generate and
refine new research questions. Relatively
There have been attempts by theoretical simple disease models supported by add-
biologists to construct systems models repre- itional biomarker data may even be capable
senting assumptions about how tuberculosis of being fitted to data from clinical studies.
therapy works and drawing indirectly from
the results of animal and clinical data
(Lipsitch and Levin, 1998). Usually, they have 11.7 Conclusion
only been capable of a sensitivity analysis
approach to their assumptions, particularly Treatment response in tuberculosis is best
regarding such phenomena as post-antibiotic thought of as a process rather than a single
effect and bacillary heterogeneity, and have event, different aspects of which are appro-
not incorporated explicit modelling of drug priate to different treatment situations and
effects. phases of clinical drug development. Among
currently proposed biomarkers, only bacterio- and Critical Care Medicine 174(3), 331–338.
logical responses have so far received signifi- Burzykowski, T., Molenberghs, G. and Buyse, M.
cant, albeit imperfect, support as surrogate (2005) The Evaluation of Surrogate Endpoints.
end points. However, the laboratory methods Springer, New York.
Butcher, P. (2004) Microarrays for Mycobacterium
used and the way in which the data are
tuberculosis. Tuberculosis 84, 131–137.
handled are crucial to getting the best perform- Buyse, M., Molenberghs, G., Burzykowski, T.,
ance out of these measures. Statistical model- Renard, D. and Geys, H. (2000) The validation
ling approaches based on capturing treatment of surrogate endpoints in meta-analyses of
effects over the entire period of therapy randomized experiments. Biostatistics (Oxford,
should be preferred over simpler and less England) 1(1), 49–67.
efficient end points, especially in the early Conde, M.B., Efron, A., Loredo, C., De Souza,
phases of development, when learning about G.R., Graça, N.P., Cezar, M.C., et al. (2009)
pharmacokinetics and pharmacodynamics is Moxifloxacin versus ethambutol in the initial
prioritized over definitive confirmation of treatment of tuberculosis: a double-blind,
randomised, controlled phase II trial. The
efficacy. New biomarkers currently in develop-
Lancet 373(9670), 1183–1189.
ment are unlikely to replace bacteriological Davidian, M. and Giltinan, D. (1995) Nonlinear
end points, but may have a significant role to Models for Repeated Measurement Data.
play in improving their performance. How Chapman and Hall, London.
this new information can be combined with Davies, G. (2009) Effect of analytical approach,
current approaches remains to be seen, but sampling scheme and laboratory method on
a number of different novel modelling power of surrogate endpoints in tuberculosis
approaches capable of incorporating it have Phase II trials: a simulation study. In: Gordon
been developed recently and promise to Conference on Tuberculosis Drug Development.
reduce the uncertainty that currently plagues Oxford.
Davies, G.R. (2010) Early clinical development of
early phase clinical development of new anti-
anti-tuberculosis drugs: science, statistics and
tuberculosis drugs. sterilizing activity. Tuberculosis (Edinburgh)
90(3), 171–176.
Davies, G.R. and Nuermberger, E.L. (2008)
References Pharmacokinetics and pharmacodynamics in
the development of anti-tuberculosis drugs.
Aber, V. and Nunn, A. (1978) Factors affecting Tuberculosis (Edinburgh) 88 (Suppl 1), S65–74.
relapse following short-course chemotherapy. Davies, G.R., Brindle, R., Khoo, S.H. and Aarons,
Bulletin of the International Union Against L.J. (2006a) Use of nonlinear mixed-effects
Tuberculosis 53(4), 260–264. analysis for improved precision of early
Baker, S.G. and Kramer, B.S. (2003) A perfect pharmacodynamic measures in tuberculosis
correlate does not a surrogate make. BMC treatment. Antimicrobial Agents and Chemo-
Medical Research Methodology 3, 16. therapy 50(9), 3154–3156.
Brindle, R.J., Nunn, P.P., Githui, W., Allen, B.W., Davies, G., Khoo, S. and Aarons, L. (2006b)
Gathua, S. and Waiyaki, P. (1993) Quantitative Optimal sampling strategies for early pharma-
bacillary response to treatment in HIV- codynamic measures in tuberculosis. Journal of
associated pulmonary tuberculosis. American Antimicrobial Chemotherapy 58, 594–600.
Review of Respiratory Diseases 147, 958–961. Davies, G., Cheirakul, N., Saguenwong, N.,
Brindle, R., Odhiambo, J. and Mitchison, D. (2001) Chaiprasert, A., Chuchuttaworn, C., Eompo-
Serial counts of Mycobacterium tuberculosis in kalap, B., et al. (2008) A factorial study of the
sputum as surrogate markers of the sterilizing effect of HIV, tuberculosis and pharmacogenetics
activity of rifampicin and pyrazinamide in treat- on the pharmacokinetics and pharma-
ing pulmonary tuberculosis. BMC Pulmonary codynamics of anti-tuberculosis drugs. In:
Medicine 1, 2. Conference on Retroviruses and Opportunistic
Burman, W.J., Goldberg, S., Johnson, J.L., Infections. Boston, USA.
Muzanye, G., Engle, M., Mosher, A.W., et al. De Gruttola, V.G., Clax, P., DeMets, D.L., Downing,
(2006) Moxifloxacin versus ethambutol in the G.J., Ellenberg, S.S., Friedman, L., et al. (2001)
first 2 months of treatment for pulmonary Considerations in the evaluation of surrogate
tuberculosis. American Journal of Respiratory endpoints in clinical trials. summary of a
160 G.R. Davies
National Institutes of Health workshop. et al. (2009) Substitution of moxifloxacin for

Controlled Clinical Trials 22(5), 485–502. isoniazid during intensive phase treatment of
Diacon, A., Patientia, R.F., Venter, A., van Helden, pulmonary tuberculosis. American Journal of
P.D., Smith, P.J., McIlleron, H., et al. (2007) Early Respiratory and Critical Care Medicine 180(3),
bactericidal activity of high-dose rifampin in 273–280.
patients with pulmonary tuberculosis evidenced EAMR Council (1974) Controlled clinical trial of four
by positive sputum smears. Antimicrobial Agents short-course (6-month) regimens of chemo-
and Chemotherapy 51(8), 2994–2996. therapy for treatment of pulmonary tuberculosis.
Diacon, A.H., Pym, A., Grobusch, M., Patientia, R., Third report. The Lancet 2, 237–240.
Rustomjee, R., Page-Shipp, L., et al. (2009) The EAMR Council (1981) Controlled clinical trial of five
diarylquinoline TMC207 for multidrug-resistant short-course (4 month) chemotherapy regimens
tuberculosis. The New England Journal of in pulmonary tuberculosis. Second report of the
Medicine 360(23), 2397–2405. fourth study. American Review of Respiratory
Diacon, A.H., Dawson, R., Hanekom, M., Narunsky, Disease 123, 165–170.
K., Maritz, S.J., Venter, A., et al. (2010) Early Fang, X., Wallqvist, A. and Reifman, J. (2009) A
bactericidal activity and pharmacokinetics of systems biology framework for modeling meta-
PA-824 in smear-positive tuberculosis patients. bolic enzyme inhibition of Mycobacterium
Antimicrobial Agents and Chemotherapy 54(8), tuberculosis. BMC Systems Biology 3, 92.
3402–3407. Fedorov, V., Mannino, F. and Zhang, R. (2009)
Diggle, P., Heagerty, P., Liang, K.Y. and Zeger, S. Consequences of dichotomization. Pharma-
(2002) Analysis of Longitudinal Data. Oxford ceutical Statistics 8(1), 50–61.
University Press, Oxford. Fox, W., Sutherland, I. and Daniels, M. (1954) A five
Diggle, P.J., Sousa, I. and Chetwynd, A.G. (2008) year assessment of patients in a controlled trial
Joint modelling of repeated measurements and of streptomycin in pulmonary tuberculosis.
time-to-event outcomes: the fourth Armitage Quarterly Journal of Medicine 23, 347–366.
lecture. Statistics in Medicine 27(16), 2981– Fox, W., Ellard, G.A. and Mitchison, D.A. (1999)
2998. Studies on the treatment of tuberculosis
Donald, P., Sirgel, F.A., Botha, F.J., Seifart, H.I., undertaken by the British Medical Research
Parkin, D.P., Vandenplas, M.L., et al. (1997) The Council Tuberculosis Units 1946–1986, with
early bactericidal activity of isoniazid related to subsequent relevant publications. International
it’s dose size in pulmonary tuberculosis. Journal of Tuberculosis and Lung Disease
American Journal of Respiratory and Critical 3(10), S231–S279.
Care Medicine 156, 895–900. Frison, L. and Pocock, S.J. (1992) Repeated
Donald, P., Sirgel, F.A., Venter, A., Smit, E., Parkin, measures in clinical trials: analysis using mean
D.P., Van de Wal, B.W., et al. (2001) The early summary statistics and its implications for
bactericidal activity of amikacin in pulmonary design. Statistics in Medicine 11(13), 1685–
tuberculosis. International Journal of Tubercu- 1704.
losis and Lung Disease 5(6), 533–538. Frison, L.J. and Pocock, S.J. (1997) Linearly
Donald, P., Sirgel, F.A., Venter, A., Smit, E., Parkin, divergent treatment effects in clinical trials with
D.P., Van de Wal, B.W., et al. (2002) The early repeated measures: efficient analysis using
bactericidal activity of streptomycin. Inter- summary statistics. Statistics in Medicine
national Journal of Tuberculosis and Lung 16(24), 2855–2872.
Disease 6(8), 693–698. Gandhi, N.R., Shah, N.S., Andrews, J.R., Vella, V.,
Donald, P., Sirgel, F.A., Venter, A., Parkin, D.P., Moll, A.P., Scott, M., et al. (2010) HIV coinfection
Seifart, H.I., van de Wal, B.W., et al. (2003) in multidrug- and extensively drug-resistant
Early bactericidal activity of antituberculosis tuberculosis results in high early mortality.
agents. Expert Reviews in Anti-Infective Therapy American Journal of Respiratory and Critical
1(1), 141–155. Care Medicine 181(1), 80–86.
Donald, P., Sirgel, F.A., Venter, A., Parkin, D.P., Garton, N.J., Christensen, H., Minnikin, D.E.,
Seifart, H.I., Van de Wal, B.W., et al. (2004) The Adegbola, R.A. and Barer, M.R. (2002)
influence of human N-acetyltransferase geno- Intracellular lipophilic inclusions of mycobacteria
type on the early bactericidal activity of in vitro and in sputum. Microbiology (Reading,
isoniazid. Clinical Infectious Diseases 39, England) 148(Pt 10), 2951–2958.
1425–1430. Garton, N.J., Waddell, S.J., Sherratt, A.L., Lee,
Dorman, S.E., Johnson, J.L., Goldberg, S., S.M., Smith, R.J., Senner, C., et al. (2008)
Muzanye, G., Padayatchi, N., Bozeman, L., Cytological and transcript analyses reveal fat
and lazy persister-like bacilli in tuberculous pulmonary tuberculosis. International Journal of

sputum. PLoS Medicine 5(4), e75. Tuberculosis and Lung Disease 10, 605–612.
Goldstein, H., Browne, W. and Rasbash, J. (2002) Li, L., Mahan, C.S., Palaci, M., Horter, L.,
Multilevel modelling of medical data. Statistics Loeffelholz, L., Johnson, J.L., et al. (2010)
in Medicine 21, 3291–3315. Sputum Mycobacterium tuberculosis mRNA as
Gumbo, T., Louie, A., Deziel, M.R., Parsons, L.M., a marker of bacteriologic clearance in response
Salfinger, M. and Drusano, G.L. (2004) Selection to antituberculosis therapy. Journal of Clinical
of a moxifloxacin dose that supresses drug Microbiology 48(1), 46–51.
resistance in Mycobacterium tuberculosis by Lienhardt, C. and Davies, G. (2010) Methodological
use of an in vitro pharmacodynamic infection issues in the design of clinical trials for the
model and mathematical modelling. Journal of treatment of multidrug-resistant tuberculosis:
Infectious Diseases 190, 1642–1651. challenges and opportunities. The International
Gumbo, T., Louie, A., Deziel, M.R., Liu, W., Parsons, Journal of Tuberculosis and Lung Disease
L.M., Salfinger, M., et al. (2007) Concentration- 14(5), 528–537.
dependent Mycobacterium tuberculosis killing Lipsitch, M. and Levin, B.R. (1998) Population
and prevention of resistance by rifampin. dynamics of tuberculosis treatment: mathe-
Antimicrobial Agents and Chemotherapy matical models of the roles of non-compliance
51(11), 3781–3788. and bacterial heterogeneity in the evolution of
Hesseling, A.C., Walzl, G., Enarson, D.A., Carroll, drug resistance. The International Journal of
N.M., Duncan, K., Lukey, P.T., et al. (2010) Tuberculosis and Lung Disease 2(3), 187–199.
Baseline sputum time to detection predicts Martineau, A.R., Timms, P.M., Bothamley, G.R.,
month two culture conversion and relapse in Hanita, Y., Islam, K., Claxton, A.P., et al. (2011)
non-HIV-infected patients. The International High dose vitamin D3 during intensive-phase
Journal of Tuberculosis and Lung Disease antimicrobial treatment of pulmonary
14(5), 560–570. tuberculosis: a double-blind randomised con-
Holtz, T.H., Sternberg, M., Kammerer, S., Laserson, trolled trial. The Lancet 377, 242–250.
K.F., Riekstina, V., Zarovska, E., et al. (2006) Matthews, J.N., Altman, D.G., Campbell, M.J. and
Time to sputum culture conversion in multidrug- Royston, P. (1990) Analysis of serial measure-
resistant tuberculosis: predictors and ments in medical research. BMJ (Clinical
relationship to treatment outcome. Annals of Research Ed.) 300(6719), 230–235.
Internal Medicine 144(9), 650–659. Mdivani, N., Li, H., Akhalaia, M., Gegia, M.,
Jamshidi, N. and Palsson, B. (2007) Investigating Goginashvili, L., Kernodle, D.S., et al. (2009)
the metabolic capabilities of Mycobacterium Monitoring therapeutic efficacy by real-time
tuberculosis H37Rv using the in silico strain detection of Mycobacterium tuberculosis mRNA
iNJ661 and proposing alternative drug targets. in sputum. Clinical Chemistry 55(9), 1694–1700.
BMC Systems Biology 1(1), 26. Mitchison, D. (1950) Development of streptomycin
Jayaram, R., Gaonkar, S., Kaur, P., Suresh, B.L., resistant strains of tubercle bacilli in pulmonary
Mahesh, B.N., Jayashree, R., et al. (2003) tuberculosis. Results of simultaneous sensitivity
Pharmacokinetics–pharmacodynamics of rifam- tests in liquid and on solid media. Thorax 5,
pin in an aerosol infection model of tuberculosis. 144–161.
Antimicrobial Agents and Chemotherapy 47(7), Mitchison, D. (1979) Basic mechanisms of chemo-
2118–2124. therapy. Chest 76(6 Suppl), 771–781.
Jayaram, R., Shandil, R.K., Gaonkar, S., Kaur, P., Mitchison, D. (1992) The Garrod Lecture. Under-
Suresh, B.L., Mahesh, B.N., et al. (2004) standing the chemotherapy of tuberculosis-
Isoniazid pharmacokinetics–pharmacodynamics current problems. Journal of Antimicrobial
in an aerosol infection model of tuberculosis. Chemotherapy 29, 477–493.
Antimicrobial Agents and Chemotherapy 48(8), Mitchison, D. (1996) Modern methods for assessing
2951–2957. the drugs used in the chemotherapy of
Jindani, A., Aber, V.R., Edwards, E.A. and mycobacterial disease. Journal of Applied
Mitchison, D.A. (1980) The early bactericidal Bacteriology 81, 72S–80S.
activity of drugs in patients with pulmonary Mitchison, D., Allen, B.W., Carrol, L., Dickinson,
tuberculosis. American Review of Respiratory J.M. and Aber, V.R. (1971) A selective oleic acid
Diseases 121, 939–949. albumin agar medium for tubercle bacilli.
Johnson, J., Hadad, D.J., Boom, W.H., Daley, C.L., Journal of Medical Microbiology 5(2), 165–175.
Peloquin, C.A., Eisenach, K.D., et al. (2006) MRC (1948) Streptomycin treatment of pulmonary
Early and extended early bactericidal activity of tuberculosis. British Medical Journal 2, 769–
levofloxacin, gatifloxacin and moxifloxacin in 782.
162 G.R. Davies
MRC (1950) Treatment of pulmonary tuberculosis Raman, K., Vashisht, R. and Chandra, N. (2009)
with streptomycin and para-aminosalicylic acid. Strategies for efficient disruption of metabolism
British Medical Journal 2, 1073–1085. in Mycobacterium tuberculosis from network
MRC (1952) The treatment of pulmonary analysis. Molecular bioSystems 5(12), 1740–
tuberculosis with isoniazid. British Medical 1751.
Journal 2, 735–746. Rustomjee, R., Diacon, A.H., Allen, J., Venter, A.,
MRC (1953) Isoniazid in combination with strepto- Reddy, C., Patientia, R.F., et al. (2008a) Early
mycin or with PAS in the treatment of pulmonary bactericidal activity and pharmacokinetics of
tuberculosis. British Medical Journal 2, 1005– the diarylquinoline TMC207 in treatment of
1014. pulmonary tuberculosis. Antimicrobial Agents
MRC (1973) Co-operative controlled trial of a and Chemotherapy 52(8), 2831–2835.
standard regimen of streptomycin, PAS and Rustomjee, R., Lienhardt, C., Kanyok, T., Davies,
isoniazid and three alternative regimens of G.R., Levin, J., Mthiyane, T., et al. (2008b) A
chemotherapy in Britain. Tubercle 54, 99–129. Phase II study of the sterilising activities of
Nahid, P., Bliven, E.E., Kim, E.Y., MacKenzie, W.R., ofloxacin, gatifloxacin and moxifloxacin in
Stout, J.E., Diem, L., et al. (2010) Influence of pulmonary tuberculosis. The International
M. tuberculosis lineage variability within a Journal of Tuberculosis and Lung Disease 12(2),
clinical trial for pulmonary tuberculosis. PloS 128–138.
One 5(5), e10753. Schön, T., Juréen, P., Giske, C.G., Chryssanthou,
Paramasivan, C., Sulochana, S., Kubendiran, G., E., Sturegård, E., Werngren, J., et al. (2009)
Venkatesan, P. and Mitchison, D.A. (2005) Evaluation of wild-type MIC distributions as a
Bactericidal action of gatifloxacin, rifampin, and tool for determination of clinical breakpoints for
isoniazid on logarithmic- and stationary-phase Mycobacterium tuberculosis. The Journal of
cultures of Mycobacterium tuberculosis. Antimicrobial Chemotherapy 64(4), 786–793.
Antimicrobial Agents and Chemotherapy 49, Sheiner, L.B. (1997) Learning versus confirming in
627–631. clinical drug development. Clinical Pharma-
Perrin, F., Lipman, M.C., McHugh, T.D. and cology and Therapeutics 61(3), 275–291.
Gillespie, S.H. (2007) Biomarkers of treatment Sheiner, L.B., Rosenberg, B. and Marathe, V.V.
response in clinical trials of novel (1977) Estimation of population characteristics
antituberculosis agents. The Lancet Infectious of pharmacokinetic parameters from routine
Diseases 7(7), 481–490. clinical data. Journal of Pharmacokinetics and
Pheiffer, C., Carroll, N.M., Beyers, N., Donald, P., Biopharmaceutics 5(5), 445–479.
Duncan, K., Uys, P., et al. (2008) Time to Sirgel, F., Donald, P.R., Odhiambo, J., Githui, W.,
detection of Mycobacterium tuberculosis in Umapathy, K.C., Paramasivan, C.N., et al. (2000)
BACTEC systems as a viable alternative to A multicentre study of the early bactericidal
colony counting. The International Journal of activity of anti-tuberculosis drugs. Journal of
Tuberculosis and Lung Disease 12(7), 792–798. Antimicrobial Chemotherapy 45, 859–870.
Phillips, P. and Fielding, K. (2007) The evaluation of Sirgel, F., Venter, A. and Mitchison, D. (2001)
culture conversion during treatment for tubercu- Sources of variation in studies of the early
losis as a surrogate marker for relapse. In: bactericidal activity of antituberculosis drugs.
Conference of the International Union Against Journal of Antimicrobial Chemotherapy 47,
Tuberculosis and Lung Disease, 2007, Cape 177–182.
Town. Sirgel, F., Fourie, P.B., Donald, P.R., Padayatchi, N.,
Phillips, P.P.J. and Fielding, K. (2008) Surrogate Rustomjee, R., Levin, J., et al. (2005) The early
markers for poor outcome to treatment for bactericidal activities of rifampin and rifapentine
tuberculosis: results from extensive multi-trial in pulmonary tuberculosis. American Journal of
analysis. In: Conference of the International Respiratory and Critical Care Medicine 171,
Union against Tuberculosis and Lung Disease, 1–8.
2008, Cape Town. Wallis, R., Patil, S., Cheon, S.H., Edmonds, K.,
Prentice, R.L. (1989) Surrogate endpoints in clinical Phillips, M., Perkins, M.D., et al. (1999) Drug
trials: definition and operational criteria. tolerance in mycobacterium tuberculosis.
Statistics in Medicine 8(4), 431–440. Antimicrobial Agents and Chemotherapy
Raman, K., Rajagopalan, P. and Chandra, N. (2005) 43(11), 2600–2606.
Flux balance analysis of mycolic acid pathway: Wallis, R.S., Doherty, T.M., Onyebujoh, P., Vahedi,
targets for anti-tubercular drugs. PLoS Com- M., Laang, H., Olesen, O., et al. (2009)
putational Biology 1(5), e46. Biomarkers for tuberculosis disease activity,
cure, and relapse. The Lancet Infectious Young, D., Stark, J. and Kirschner, D. (2008)
Diseases 9(3), 162–172. Systems biology of persistent infection:
Wang, R. and Marcotte, E.M. (2008) The proteomic tuberculosis as a case study. Nature Reviews.
response of Mycobacterium smegmatis to anti- Microbiology 6(7), 520–528.
tuberculosis drugs suggests targeted pathways. Zhang, L., Sinha, V., Forgue, S.T., Callies, S., Ni, L.,
Journal of Proteome Research 7(3), 855–865. Peck, R., et al. (2006) Model-based drug
Yoo, B. (2009) The impact of dichotomization in development: the road to quantitative
longitudinal data analysis: a simulation study. pharmacology. Journal of Pharmacokinetics
Pharmaceutical Statistics 9(4), 298–312. and Pharmacodynamics 33(3), 369–393.
12 Measuring Gene Expression by
Quantitative PCR (qPCR)
Isobella Honeyborne
Centre for Clinical Microbiology, Department of Infection,
University College London, UK
12.1 Introduction scriptome changes is therefore important

in understanding how the bacteria are
Quantification of gene expression requires adapting and surviving in the host. To obtain
isolation and detection of RNA. There are a broad picture of the changes in gene path-
two main ways in which extracted RNA can ways, microarray technology is generally
be used to study Mycobacterium tuberculosis. used. Microarrays can screen a large number
The first is to compare changes in the of genes simultaneously. Reverse tran-
transcriptome (mRNA expression level) dur- scription quantitative PCR (RT-qPCR) is
ing phases of growth or with different stress then used to confirm the findings for changes
factors; for example, drug treatment, hypoxia in the expression of specifically identified
and starvation. genes.
The ability to elucidate how the bacteria The second use for measuring gene
respond under different circumstances will expression by qPCR reflects quantification of
inform our understanding of the basic cell numbers. It is theoretically possible to
biology of the organism and help to identify determine the number of M. tuberculosis
novel drugs which can subvert these survival bacteria in a clinical sample by detecting the
mechanisms. Like many bacteria, M. tubercu- amount of a particular high abundance,
losis is remarkably adaptable. M. tuberculosis constitutively expressed gene. Sputum is a
is able to tolerate reduced oxygen tension in useful sample for this purpose, since it can be
the granuloma, host adaptive immune collected non-invasively and is usually
responses and subvert post-phagocytic readily available from individuals with
destruction within the macrophage to use pulmonary tuberculosis. Use of a standard
this cell as a living space. These different curve with qPCR cycle threshold values for
conditions are likely to induce changes in sputum samples with a known bacterial load
genes important in adapting to a generic should then allow the unknown sputum
stress response. Additionally, particular bacterial load to be determined. The re-
gene pathways may respond by increased or mainder of this chapter will focus on using
decreased expression in order to cope with a qPCR of such constitutively expressed genes
specific set of conditions. Detecting tran- to determine the bacterial number.

Measuring Gene Expression by Quantitative PCR (qPCR) 165
12.2 The Need for Molecular-based system. Despite the decontamination process,
Quantification of Bacterial Load a proportion of samples will be overgrown
and therefore invalidated by growth of other
To date, although there are several com- microorganisms. Additionally, there may be
mercially available molecular-based kits for M. tuberculosis bacteria present that are
the detection and diagnosis of tuberculosis, refractive to the particular culture media but
there are none available that offer quantifi- which are not dead. Recent studies have
cation. This is surprising since real-time PCR suggested that these may represent an
technology has been available for upwards of important population (Shleeva et al., 2002;
a decade. The reasons why a commercial Davies et al., 2008; Mukamolova et al., 2010).
assay is not already available will be explored If decontamination is not used, then the
later in this section. solid culture media can be made selective for
Assessment of bacterial load in sputum tuberculosis by the addition of antimicrobials,
is generally used as a research tool or to and can be used to quantitate bacilli load
ascertain the efficacy of novel drugs. For directly in sputum samples. In some cases,
clinical diagnosis of tuberculosis, the bacterial the results are still invalidated by the growth
burden of the sample is only assessed at the of other microorganisms. The sputum serial
level of the sputum smear. Although it is of colony-counting (SSCC) assay uses such
limited use in clinical decision making, a selective media to determine the number of
smear can be graded according to how many M. tuberculosis colony-forming units (CFU) in
bacilli are seen per microscope field of view. sputum at multiple time points during
This can then be used to approximate the treatment (see Chapter 11).
bacterial load of the sample. Notably, this The drawbacks of the currently available
does not differentiate between live and dead culture-based methods for detecting bacterial
bacilli and the detection of tuberculosis by load highlight how useful a molecular-based
sputum smear requires approximately 104 assay would be for quantification. There are
bacilli/ml sputum. There are a large number two main advantages of using molecular
of tuberculosis-positive individuals who are quantitation over culture. The first would be
smear negative, and of this group, paediatric the use of primer and probe sets specific for
tuberculosis is particularly difficult to diag- M. tuberculosis complex bacteria to circumvent
nose by this method. the problems with invalidation by growth
The use of culture in many laboratories from other organisms. The second that
increases the detection rate of tuberculosis. samples could be enumerated rapidly, regard-
Inoculation of decontaminated sputum sedi- less of the number of bacteria present.
ment on to solid culture, such as Löwenstein– Although bacterial load is not commonly
Jensen slopes, can be used to semi-quantify used during routine clinical management of
the bacterial load by counting the number of tuberculosis disease, culture-based methods
colonies that grow. Liquid culture can be are used to determine the efficacy of novel
used similarly by assessing how long the tuberculosis drugs by measuring the decline
culture takes to be become positive. These in bacterial numbers during early therapy
methods, however, are confounded by several (Donald et al., 2000; Brindle et al., 2001).
effects, with the major drawback being the Previous studies have identified culture
slow replication rate of tuberculosis. In solid positivity after 2 months of therapy as a risk
culture, bacterial colonies are unlikely to be factor (Aber and Nunn, 1978; Mitchison,
seen for at least 2–3 weeks, and for liquid 1993) associated with later relapse. Recently,
culture the speed of detection is dependent high presenting bacterial load has been
on the number of bacteria present in the highlighted as a hazard factor for risk of later
starting inoculum. Those samples most likely relapse (Hesseling et al., 2010). It is important
to be missed by smear, due to a low bacilli to detect all bacteria with the potential to
load, will also likely take the longest to flag cause disease, regardless of the life cycle they
positive in the automated liquid culture are in. As discussed in Chapter 11, mathe-
166 I. Honeyborne
matical modelling of SSCC measurement this time point, all patients were found
during the first 2 months of treatment has simultaneously to be negative for M.
found the data are best fitted to a biexponential tuberculosis by culture, as would be expected,
decay curve. This has been suggested as since after 6 months of treatment the vast
being reflective of two principle subpopu- majority of individuals have been cured of
lations of bacteria which respond to treatment active disease.
differently. It is widely accepted that RNA is a
A recent study shortened treatment from shorter-lived species than DNA and therefore
6 to 4 months for those with sputum culture its degradation might be expected to match
negative for tuberculosis at 2 months of disappearance in bacteria as they are killed.
treatment (Johnson et al., 2009). A higher Either ribosomal (rRNA) or messenger
proportion of this group went on to relapse (mRNA) can be followed. mRNA is a useful
with the identical strain to the first episode molecule, since it has been shown to decay
during the 2-year follow-up period. This rapidly after cell death (Hellyer et al., 1999a);
suggests that either there were bacteria however, it is not very abundant. There are an
present in the sputum that were refractive to estimated 0.1 copies of the highly expressed
culture or that the bacteria were archived in antigen 85B mRNA per CFU (Desjardin et al.,
the lung in a place inaccessible to expectorated 1999). Previous studies and our data suggest
sputum. that you cannot detect abundant mRNA
In summary, it is clear that bacterial load species in samples containing <104 bacilli/ml
ascertained from culture-based methods is a (Desjardin et al., 1999). Additionally, some in
useful method for determining bacterial load vitro data found 85B mRNA disappearance to
but is likely not to give an accurate reflection be pre-emptive of the decline in bacterial
of the true viable bacteria in the sample. numbers (Hellyer et al., 1999a). rRNA is much
more available for detection, with an esti-
mated 800 molecules/CFU (Desjardin et al.,
12.3 Why There Is Not Currently a 1999). It would be expected, therefore, that it
Molecular Assay for Enumeration of would be detected more readily than even
M. tuberculosis in Sputum non-repetitive DNA sequences in samples
containing few bacteria. The literature, how-
Ideally, molecular techniques would identify ever, has reported that the 16S rRNA molecule
the number of bacteria present in a sample does not decay rapidly enough to be useful as
that were at any stage of the bacterial life a marker of cell viability (Desjardin et al.,
cycle and capable of causing disease. The 1999; Hellyer et al., 1999a). However, the in
perfect molecule for measuring this would be vitro study did find changes in the 16S rRNA
both abundant, and therefore readily detect- level after 72 h of isoniazid treatment (Hellyer
able, but would also decay rapidly following et al., 1999a).
cell death, so that only live cells would be Other studies of mycobacteria also sup-
quantified. port the rapid decline of 16S rRNA following
DNA has been shown to be detected bacterial cell death. One such study found
following culture negativity in a controlled that M. aurum ingestion by peritoneal macro-
murine model, which is suggestive that this phages resulted in ribosome destruction
molecule is unreflective of live cell status (de being an early event, occurring after 4 days
Wit et al., 1995). This observation has also (Silva et al., 1987). Macrophages form a major
been confirmed in several studies using part of the granuloma, although they are not
longitudinal data from tuberculosis-infected present in the acellular necrotic centre. It is
individuals during chemotherapy in Tanzania hypothesized that dead bacteria will be
(Kennedy et al., 1994), and DNA was detected expelled from the sputum and those
in 25% of specimens collected in tuberculosis- remaining destroyed by macrophages present
positive patients administered with at least in the periphery of the granuloma structure.
180 days of treatment (Hellyer et al., 1996). At Further work needs to be done, however, to
ascertain the extent to which rRNA is detected and enumerated accurately between
detectable following ingestion of tuberculosis 107 and 102/ml sputum using this method
by alveolar macrophages following bacterial (Honeyborne et al., 2011).
cell death. In ex vivo samples, detection of
rRNA beyond culture negativity could
represent a population of live bacteria 12.3.2 Problems and preservation of
refractive to culture (Moore et al., 1996). In a sputum RNA
recent study of 111 patients with severe
tuberculosis disease, we measured tubercu- Rapid decay of RNA is a useful attribute for
losis-specific 16S rRNA longitudinally during measuring live cells; however, the friability of
treatment. We found a rapid decline of 16S this molecule means that preserving it so that
rRNA after 72 h of therapy (0.99 log10). In the it is available must be addressed. It is
same study, we found that only 1 of 43 important for transcriptome studies that
patients was positive for tuberculosis using RNA transcripts are preserved immediately
the 16S assay after 168 days of treatment following expectoration, otherwise bacteria
(unpublished observation). present in sputum may change their expres-
sion in response to altered environmental
conditions not relevant to the experiment.
12.3.1 Sensitivity of nucleic acid-based Treating bacteria with solutions containing
detection guanidine thiocyanate (GTC) has been found
to preserve the RNA effectively (Monahan et
mRNA has been demonstrated to decay al., 2001).
rapidly following in vitro drug treatment of GTC penetrates host cells and the myco-
M. tuberculosis (Hellyer et al., 1999a) and in bacterial cell wall protects the tuberculosis
sputum (Desjardin et al., 1999; Hellyer et al., bacteria, preserving the intracellular com-
1999b; Li et al., 2010). Genes that have ponents including RNA. In the method used
previously been measured include fbpB in our laboratory, based on the method of
(encodes 85B protein), hspX (encodes alpha- Monahan et al. (2001), successful amplification
crystalline homologue protein), rrnA-P1 (a of RNA has been performed from sputum
non-encoding region for the ribosomal samples frozen in GTC. This is somewhat
promoter) and icl (encodes isocitrate lyase). surprising since it would be expected that the
icl expression appears to be the highest of the bacterial cell wall would lose its integrity on
genes tested to date; however, it does not thawing. However, we have extracted RNA
appear to correlate with decline of CFU repeatedly and successfully from sputum
between days 0 and 2 (Li et al., 2010). icl and frozen at –80°C in GTC-containing solutions,
rrnA-P1 showed some correlation with CFU even with repeat freeze–thaw cycles of the
between days 2 and 7. In the same study, after sputum in GTC. We have extracted M.
2 months of therapy, 66% (25 of 38) of tuberculosis RNA successfully and amplified it
individuals tested were still culture positive; in RT-qPCR reactions after the sputum has
detection of icl, however, was found in only been frozen at –80°C for up to 10 years.
28% of the same samples (7 of 25). Detection Following the treatment of sputum with
of these genes may have some use during GTC, the solution is centrifuged at 2000 g to
very early therapy and, potentially, com- pellet the bacteria out of the sample. Nucleic
binations of genes could prove useful. How- acid in aqueous solutions will not pellet at
ever, sensitivity of mRNA detection still this centrifugal force. Some nucleic acid
remains an issue. Due to the greater abund- released from cells is, however, recovered
ance of rRNA compared to mRNA, we have during this process. We have shown this by
developed an assay based on detection of 16S spiking a naked in vitro transcribed mRNA
rRNA. Using a serial dilution from a known 1957-bp fragment into sputum in GTC and
number of bacteria and spiking M. tuberculosis find that it is possible to detect the gene by
into negative sputum samples, we have RT-qPCR following extraction. Approximately
168 I. Honeyborne
500-fold of the mRNA is lost during this and Taq polymerase. However, these effects
process, however. were the same for the internal control and the
bacterial genes measured. This therefore
allows the internal control to be used to
12.3.3 Extraction of RNA and inhibitors normalize cycle threshold values obtained for
bacterial genes in unknown samples.
There are inhibitors present in sputum samples In summary, the low abundance of
and if these are not accounted for, then a count mRNA, the questionability of whether rRNA
based on direct correlation between cycle is a short-lived molecule, preservation of
threshold as measured by RT-qPCR will not RNA in sputum and inhibitors are likely to be
be accurate. the main reasons why a commercially avail-
The loss of RNA and inhibitors present able molecular-based assay of quantification
in a sample are proportional for the synthetic has not been readily available to date.
spiked mRNA and bacterial genes sigA and If RNA is extracted carefully without the
16S rRNA tested (Fig. 12.1). This was deter- introduction of RNases, it does not degrade
mined when 51 tuberculosis-negative sputum rapidly at 4°C. We measured samples stored
samples were spiked with the same number at 4°C 3 months later and identical cycle
of M. tuberculosis bacteria and internal control threshold values were obtained. In all likeli-
and RNA extracted from each sample. When hood, multiple freeze–thaw cycles degrade
sigA mRNA, 16S rRNA and the internal RNA more rapidly and should be avoided.
control gene were measured, there was a The use of 16S rRNA and an internal
large range of cycle threshold values obtained control has allowed us to develop a molecular
for each measured gene. As would be assay to quantify tuberculosis bacteria in
expected, comparison of the detection of two sputum samples.
bacterial genes (16S rRNA and sigA) found If a serial dilution of tuberculosis is made
there was a correlation between their recovery and spiked into tuberculosis-negative sputum,
(R2 = 0.87). When detection of the internal RNA can be extracted and detected. If inhib-
control was similarly compared to that of the ition is taken into account by normalizing
16S rRNA and sigA, a similar correlation was against the internal control, then a repro-
found between the internal control and 16S ducible method of quantification is available
rRNA and between the internal control and between 107 and 102 bacilli/ml sputum (Fig.
sigA of R2 = 0.84 and 0.81, respectively. The 12.2). Samples tested were between 107 and
range in cycle threshold values was caused 102 bacilli/ml sputum were shown to be
by RNA loss and inhibitors, transferred with enumerated within 0.5 log10 in 98% (n = 86) of
the RNA that affected the reverse transcriptase cases.
slope = 0.8562 ± 0.05 slope 0.7834 ± 0.05 slope 0.7076 ± 0.05

35
r2 = 0.87 35
r2= 0.84 r2 = 0.81
35
16s rRNA (CT)
30 30 30
sigA (CT)
sigA (CT)
25 25 25
20 20 20
15 15 15
10 10 10
5 5 5
0 0 0
5 10 15 20 25 5 10 15 20 25 30 35 5 10 15 20 25 30 35
16S rRNA (CT) Internal control (CT) Internal control (CT)
Fig. 12.1. Detection of tuberculosis mRNA (sigA), 16S rRNA and artificially spiked internal control. Each
point is a sample with 107 bacilli and 50 ng internal control. (Adapted from Copyright © American Society
for Microbiology, Journal of Clinical Microbiology, Vol 49, 2011, pp. 3905–3911, DOI 10.1128/JCM.00547-
11.)
40 represent the corresponding gradients of the

Cycle threshold
decline of counts during that phase. There

30
was a significant difference in the fitted lines
20 between relapsed and cured patients (P =
0.0031), with the difference coming in the A
10 and B parameters, reflecting the bacterial
load at presentation (Fig. 12.3). There was no
0 7 6 5 4 3 2 1 0
10 10 10 10 10 10 10 10 0 significant difference in the gradient α in the
first (P = 0.122) or second phase. Comparison
Sputum bacterial load (bacilli/ml)
between the 16S rRNA measured bacterial
load and other studies measuring bacterial
Fig. 12.2. Serial dilutions of H37Rv and 50 ng decline using solid agar counts found the
internal control were spiked into tuberculosis- transition times were longer using 16S rRNA
negative sputum samples. RNA was extracted and
assay than solid culture, occurring at 8.15
the cycle threshold measured for 16S rRNA and
days and 9.36 days for cured and relapsed,
the internal control. The recovery of the internal
control was used to normalize the detection of the respectively, compared to other studies of
16S rRNA and the cycle threshold (CT) value for cured patients, which found the transition
each dilution plotted (the mean ± standard time to be around 2.62 and 2.12 days (Fig.
deviations are defined by the lines at each 12.3). These differences may be due to a
bacterial concentration). (Adapted from Copyright plethora of reasons, including 16S rRNA not
© American Society for Microbiology, Journal of decaying immediately following bacterial cell
Clinical Microbiology, Vol 49, 2011, pp. 3905–3911, death or measurement of bacteria that are
DOI 10.1128/JCM.00547-11.) refractive to culture.
At day 56, we compared liquid culture
positivity to samples positive by the 16S rRNA
assay. However, a culture result was not
available as 14% of samples (15 of 110) were
12.4 Measurement of 16S rRNA in found to have contaminated cultures. This
Patient Sputum Samples was in comparison to 0.9% (1 of 110) where
the internal control had failed and a 16S rRNA
16S rRNA normalized against our novel assay result was not available; 64% (60 of 94)
internal control has been used successfully to had a culture and 16S rRNA result that
screen patient sputum samples over time matched. There were an additional 24 samples
(Honeyborne, 2011). Modelling of bacterial that were found to be positive by culture but
decay in response to chemotherapy has negative by 16S rRNA, and 10 that contained
previously been performed using solid agar 102 bacilli/ml sputum by 16S rRNA but were
counting on media selective for tuberculosis negative by culture (Table 12.1).
by the addition of antimicrobials, as des- Ninety-six per cent (25 of 26) of sputum
cribed. These studies typically found biphasic samples that were found to contain 103
decay in bacterial load when measured over bacilli/ml sputum as measured by the abun-
time. Modelling of bacterial load using 43 dance of 16S rRNA were also found to be
patients and measuring the bacterial load culture negative.
using the 16S rRNA assay found similar
biphasic decay. Comparison of the decay of
16S rRNA determined bacterial load in our 12.5 Conclusion
study compared to bacterial load ascertained
by SSCC in other studies. The biexponential In summary, future studies measuring
mixed effects model with parameters bacterial numbers in sputum samples might
differing by relapse status resulted in the best find it worthwhile to have data for both the
fit (model 4) (Fig. 12.3). A and B represent the 16S rRNA assay and culture. Those that are
intercepts on the log10 scale of the first and negative by both assays at day 56 might be
second phases of elimination, and α and β better candidates for treatment shortening.
170 I. Honeyborne
8
Relapse
Cure
Rustomjee et al., 2008
Davies et al., 2006
6
Log CFU
4
2
Limit of detection
0
0 20 40 60 80
Days
Notes: Lower panel: aAIC = Akaike Information Criterion – a lower number indicates a model that fits the
data better; bcomparing models 1 and 2; ccomparing models 2 and 3; dcomparing models 3 and 4; eWald
test comparing parameter estimates between models for cured and relapse patients.
Fig. 12.3. Mathematical modelling of longitudinal data of bacterial decline for 43 patients using data from
day 0 to day 56. Upper panel: the best fit, mixed effect biphasic decay model for patients who went on to
cure and those who relapsed, showing the OLFLOTUB (Rustomjee et al., 2008) and Davies streptomycin
and isoniazid or rifampin and pyrazinamide (SHRZ) (Davies et al., 2006) studies where bacterial decline
was measured using solid agar colony and also found the best fit was a biphasic decay. Lower panel: fit
of nested sums of various exponential models. Parameters A and B are the intercepts for the various
phases of killing (log10 bacterial load/ml sputum),  and  are the corresponding rates of decrease in
bacterial load (log10 bacterial load/ml sputum), as derived from the model. (Adapted from Copyright ©
American Society for Microbiology, Journal of Clinical Microbiology, Vol 49, 2011, pp. 3905–3911, DOI
10.1128/JCM.00547-11.)
Table 12.1. Liquid culture and 16S rRNA sputum treatment. Antimicrobial Agents and Chemo-
tuberculosis detection after 8 weeks of therapy for therapy 50(9), 3154–3156.
110 individuals. Upper panel: total sputum samples de Wit, D., Wootton, M., Dhillon, J. and Mitchison,
positive and negative for each assay. Lower panel: D.A. (1995) The bacterial DNA content of mouse
distribution of positive and negative culture at each organs in the Cornell model of dormant
bacterial load measured using 16S rRNA. tuberculosis. Tuberculosis and Lung Disease
(Adapted from Copyright © American Society for 76(6), 555–562.
Microbiology, Journal of Clinical Microbiology, Vol Desjardin, L.E., Perkins, M.D., Wolski, K., Haun, S.,
49, 2011, pp. 3905–3911, DOI 10.1128/ Teixeira, L., Chen, Y., et al. (1999) Measurement
JCM.00547-11.) of sputum Mycobacterium tuberculosis mes-
senger RNA as a surrogate for response to
16S rRNA Liquid culture chemotherapy. Amercian Journal of Respiratory
+ + 37 and Critical Care Medicine 160(1), 203–210.
– – 23 Donald, P.R., Sirgel, F.A., Kanyok, T.P., Danziger,
+ – 10 H., Venter, A., Botha, F.J., et al. (2000) Early
– + 24 bactericidal activity of paromomycin (amino-
Result Contaminated 15 sidine) in patients with smear-positive pulmon-
ary tuberculosis. Antimicrobial Agents and
Failed Result 1 Chemotherapy 44(12), 3285–3287.
Hellyer, T.J., Fletcher, T.W., Bates, J.H., Stead,
W.W., Templeton, G.L., Cave, M.D., et al. (1996)
16S rRNA measured
Strand displacement amplification and the
bacterial load/ml sputum Liquid culture
polymerase chain reaction for monitoring
+ – response to treatment in patients with pulmon-
1–5  102 12 9 ary tuberculosis. Journal of Infectious Diseases
1–5  103 22 1 173(4), 934–941.
1–5  104 2 0 Hellyer, T.J., DesJardin, L.E., Hehman, G.L., Cave,
1–5  105 1 0 D. and Eisenach, K.D. (1999a) Quantitative
analysis of mRNA as a marker for viability of
Mycobacterium tuberculosis. Journal of Clinical
Microbiology 37(2), 290–295.
Hellyer, T.J., DesJardin, L.E., Teixeira, L., Perkins,
M.D., Cave, M.D. and Eisenach, K.D. (1999b)
Detection of viable Mycobacterium tuberculosis
References by reverse transcriptase-strand displacement
amplification of mRNA. Journal of Clinical
Aber, V.R. and Nunn, A.J. (1978) Short-term Microbiology 37(3), 518–523.
chemotherapy of tuberculosis. Factors affecting Hesseling, A.C., Walzl, G., Enarson, D.A., Carroll,
relapse following short term chemotherapy. N.M., Duncan, K., Lukey, P.Y., et al. (2010)
Bulletin of the International Union against Baseline sputum time to detection predicts
Tuberculosis and Lung Disease 53(4), 276–280. month two culture conversion and relapse in
Brindle, R., Odhiambo, J. and Mitchison, D. (2001) non-HIV-infected patients. The International
Serial counts of Mycobacterium tuberculosis in Journal of Tuberculosis and Lung Disease
sputum as surrogate markers of the sterilising 14(5), 560–570.
activity of rifampicin and pyrazinamide in Honeyborne, I., McHugh, T.D., Phillips, P.P.,
treating pulmonary tuberculosis. BMC Pulmon- Bannoo, S., Bateson, A., Carroll, N., et al.
ary Medicine 1, 2. (2011) Molecular bacterial load assay, a culture-
Davies, A.P., Dhillon, A.P., Young, M., Henderson, free biomarker for rapid and accurate
B., McHugh, T.D. and Gillespie, S.H. (2008) quantification of sputum Mycobacterium tubercu-
Resuscitation-promoting factors are expressed losis bacillary load during treatment. Journal of
in Mycobacterium tuberculosis-infected human Clinical Microbiology 49(11), 3905–3911.
tissue. Tuberculosis (Edinburgh) 88(5), 462– Johnson, J.L., Hadad, D.J., Dietze, R., Maciel, E.L.,
468. Sewali, B., Gitta, P., et al. (2009) Shortening
Davies, G.R., Brindle, R., Khoo, S.H. and Aarons, treatment in adults with noncavitary tuberculosis
L.J. (2006) Use of nonlinear mixed-effects and 2-month culture conversion. American
analysis for improved precision of early Journal of Respiratory and Critical Care
pharmacodynamic measures in tuberculosis Medicine 180(6), 558–563.
172 I. Honeyborne
Kennedy, N., Gillespie, S.H., Saruni, A.O., Protocols: Methods in Molecular Medicine.
Kisyombe, G., McNerney, R., Ngowi, F.I., et al. Humana Press, Totowa, New Jersey.
(1994) Polymerase chain reaction for assessing Mukamolova, G.V., Turapov, O., Malkin, J., Wolt-
treatment response in patients with pulmonary mann, G. and Barer, M.R. (2010) Resuscitation-
tuberculosis. Journal of Infectious Diseases promoting factors reveal an occult population of
170(3), 713–716. tubercle bacilli in sputum. American Journal of
Li, L., Mahan, C.S., Palaci, M., Horter, L., Respiratory and Critical Care Medicine 181(2),
Loeffelholz, L., Johnson, J.L., et al. (2010) 174–180.
Sputum Mycobacterium tuberculosis mRNA as Rustomjee, R., Lienhardt, C., Kanyok, T., Davies,
a marker of bacteriologic clearance in response G.R., Levin, J., Mthivane, T., et al. (2008) A
to antituberculosis therapy. Journal of Clinical Phase II study of the sterilising activities of
Microbiology 48(1), 46–51. ofloxacin, gatifloxacin and moxifloxacin in
Mitchison, D.A. (1993) Assessment of new pulmonary tuberculosis. International Journal of
sterilizing drugs for treating pulmonary tubercu- Tuberculosis and Lung Disease 12(2), 128–138.
losis by culture at 2 months. American Review Shleeva, M.O., Bagramyan, K., Telkov, M.V.,
Respiratory Disease 147(4), 1062–1063. Mukamolova, G.V., Young, M., Kell, D.B., et al.
Moore, D.F., Curry, J.I., Knott, C.A. and Jonas, V. (2002) Formation and resuscitation of ‘non-
(1996) Amplification of rRNA for assessment of culturable’ cells of Rhodococcus rhodochrous
treatment response of pulmonary tuberculosis and Mycobacterium tuberculosis in prolonged
patients during antimicrobial therapy. Journal of stationary phase. Microbiology 148(Pt 5), 1581–
Clinical Microbiology 34(7), 1745–1749. 1591.
Monahan, I.M., Mangan, J.A. and Butcher, P.D. Silva, M.T., Appelberg, R., Silva, M.N. and Macedo,
(2001) Extraction of RNA from intracellular P.M. (1987) In vivo killing and degradation of
Mycobacterium tuberculosis: methods, con- Mycobacterium aurum within mouse peritoneal
siderations, and applications. In: Parish, T. and macrophages. Infection and Immunity 55(9),
Stoker, N.G. (eds) Mycobacterium tuberculosis 2006–2016.
13
Transcriptomic Approaches to
Mapping Responses to Drug Therapy for
Tuberculosis
Simon J. Waddell1 and Philip D. Butcher2

1Brighton and Sussex Medical School, University of Sussex, Brighton, UK;
2Centre for Infection and Immunity, Division of Clinical Sciences,
St George’s University of London, UK
13.1 Introduction cell screens or mechanisms of resistance in

drug-resistant strains. Third, as an unsuper-
Understanding how anti-tuberculosis drugs vised global approach, to confirm compound-
work is fundamental to developing novel target selectivity and detect mutagenic
chemotherapy strategies; transcriptional pro- signatures during hit-to-lead medicinal chem-
filing of the response of Mycobacterium istry. As such, while the experimental set-up
tuberculosis bacilli to drug exposure offers is involved and the results often difficult to
important clues to the mode of action of both interpret, transcriptional profiling contributes
new and existing compounds (Sacchettini et considerably to multiple stages of early
al., 2008; Barry and Blanchard, 2010). Global antibacterial drug discovery; as summarized
gene expression profiling by microarray was in Fig. 13.1. This chapter focuses on the
first applied to M. tuberculosis in 1999, when applications of gene expression profiling,
Wilson and colleagues explored the action of derived predominantly from microarray
isoniazid (INH) and ethionamide (ETA) on analyses, towards the development of novel
log-phase bacilli (Wilson et al., 1999). Since compounds and a greater understanding of
then, contrasting and dissecting the tran- drug action in M. tuberculosis.
scriptional responses of M. tuberculosis bacilli
to antimycobacterial agents has become an
integral element of the drug discovery 13.2 Identifying Compound Class
toolbox. Global mRNA profiling is particu- and Mechanism of Action
larly suited to tackling three key scenarios in
the drug development pipeline. First, to The genome-wide measurement of transcript
identify targets for rational drug design by abundance offers a molecular landscape of
defining pathways used by bacilli in axenic, pathways used by M. tuberculosis bacilli in the
intracellular or in vivo models of M. tubercu- particular microenvironment modelled. Thus,
losis infection. Second, in a hypothesis- the comparison of mRNA profiles in the
generating capacity, to characterize the mode presence or absence of a compound that
of action of compounds isolated from whole inhibits a key target captures the effects that

174 S.J. Waddell and P.D. Butcher
Characterize drug target
Define drug class
Identify resistance mechanisms
Investigate drug-tolerant states
Recognize mutagenic signatures
Determine target drift
Create interpretive framework
Fig. 13.1. The applications of M. tuberculosis transcriptional profiling through the drug development
process, from the comparison of drug-treated with drug-free mRNA signatures derived from log-phase
bacilli.
abrogation of this enzyme might have on generated analogous transcriptional signa-

every cellular process. This global approach tures (Boshoff et al., 2004), therefore con-
has therefore been employed to reveal, at the firming that the class of an antibiotic could be
transcriptional level, which pathways are defined by the bacterial transcriptional
influenced by the actions of antimycobacterial response to drug exposure.
compounds. The first proof-of-principle Contrasting these drug class-specific
study, by Wilson et al. (1999), demonstrated gene expression signatures may also reveal
that the mechanism of action of a known novel modes of action for existing (and newly
drug, INH, could be inferred from the genes discovered) compounds. For example, the
expressed divergently after INH treatment. profile derived after exposure to triclosan, an
Genes associated with the FAS-II cycle (fabD, inhibitor of mycolic acid synthesis (InhA),
acpM, kasA, kasB, accD6), involved in the did not cluster with known cell wall-targeting
elongation of mycolic acids, were induced drugs but rather with a group of compounds
after INH (or ETA) exposure. Thus, although that acted by inhibiting mycobacterial respira-
the enoyl-ACP reductase InhA (encoded by tion. This led the authors to propose an
inhA), the target of INH, was not itself alternative mechanism for triclosan-mediated
upregulated in the presence of INH, genes in killing (Boshoff et al., 2004). Likewise, the
the FAS-II pathway were significantly transcriptional response of M. tuberculosis to
differentially expressed. This approach was the nitroimidazole PA-824 revealed signa-
expanded by the studies of Betts et al. (2003) tures representative of both cell wall inhib-
and Waddell et al. (2004), where common and ition and respiratory poisoning. This
drug-specific signatures were defined for a multifactorial profile highlighted to the
range of antimycobacterial compounds. The authors that PA-824 might have dual mechan-
former investigation employed a stepwise isms of action; targeting cell wall biosynthesis
linear-discriminant analysis that was able to in log-phase growth and mediating killing
predict the antibiotic applied (from three via nitric oxide in hypoxic non-replicating
compounds all targeting fatty acid bio- bacilli (Manjunatha et al., 2009). The use of
synthesis) based on the expression pattern of such an interpretive framework of drug-
21 genes. The seminal work of Boshoff and induced transcriptional patterns to classify
colleagues categorized the responses of M. novel compounds has been highlighted
tuberculosis bacilli to 75 different drugs or recently (Makarov et al., 2009). The myco-
drug combinations and demonstrated that bacterial response to benzothiazinones (BTZs)
compounds with similar modes of action was highly similar to the signature derived
Transcriptomic Approaches to Mapping Responses to Drug Therapy 175
after exposure of M. tuberculosis bacilli to ations are also summarized in Fig. 13.2; for a
ethambutol (EMB). Subsequent investigations detailed review of INH activity, see Vilcheze
demonstrated that this new class of anti- and Jacobs (2007). INH enters the bacterium
mycobacterial compound prevented the by passive diffusion, where the prodrug is
generation of arabinogalactan precursors, converted to an active INH–NAD adduct by a
specifically decaprenylphosphoryl arabinose catalase-peroxidase-peroxynitritase (encoded
mediated by Rv3790, a decaprenylphosphoryl- by katG). The transformation of INH into the
beta-d-ribose 2’-epimerase (Makarov et al., INH–NAD adduct requires NAD+ and KatG
2009; Trefzer et al., 2010). Furthermore, the catalase activity, which might usually
BTZ-target epimerase Rv3790 acts in the perform antioxidative/nitrosative functions.
same pathway two enzymatic steps upstream This process may also generate nitric oxide
of the arabinosyl-transferases (encoded by radicals, with toxic consequences (Timmins
embC/A/B) that are inhibited by EMB and Deretic, 2006). Thus, conversion of INH
(Belanger et al., 1996). Thus, BTZ and EMB, to the active INH–NAD adduct may influence
both targeting the mycobacterial cell wall by the mycobacterial NAD+/NADP+ pool and
preventing formation of arabinogalactan- the ability of bacilli to respond to reactive
mycolyl structures, exert similar stresses on oxygen and nitrogen intermediates. Efflux
multiple processes within the cell. Cor- pumps (such as IniB/A/C) may remove both
respondingly, the transcriptional profiles INH and INH–NAD moieties from the cell.
derived from these two compounds overlap The INH–NAD adduct acts primarily to
considerably (Makarov et al., 2009). In such a inhibit elongation of mycolic acids (targeting
way, a drug-induced transcriptional signature the enoyl-ACP reductase, encoded by inhA),
alongside profiles of target knockdown or although additional targets (products of kasA,
conditional overexpression linked to changes fabG1/mabA and dfrA) have been proposed
in minimum inhibitory concentration (MIC) (Argyrou et al., 2006; Timmins and Deretic,
becomes an effective method of target valid- 2006; Vilcheze and Jacobs, 2007; Wang et al.,
ation for drug discovery programmes. 2010). Abrogation of enoyl-ACP reductase
A useful feature of these analyses is the activity as part of fatty acid synthesis (FAS-II)
ability to scrutinize crude extracts results in defective mycolic acid biosynthesis.
demonstrating antimycobacterial activity. This results in the build-up of precursors
This is particularly relevant to natural from FAS-I and intermediate saturated fatty
product screens, where, for example, the acids that feed into FAS-II, and perhaps also
RNA signature derived after exposure of M. an excess of cofactors that normally form
tuberculosis to an unrefined sample of the complex structures with full-length mycolic
marine tunicate Eudistoma amplum is acids. Finally, the inability to generate mature
analogous to that of the purified active full-length mycolic acids presumably leads to
compound, ascididemin (Boshoff et al., 2004). a breakdown in the structure of the
mycobacterial cell wall and cell death as
membrane integrity is impaired. It is a
13.3 The Multifactorial Nature of INH combination of these effects that are captured
Exposure as a Paradigm of Complex in the INH-transcriptional signature, reflect-
M. tuberculosis Responses ing changes in the flux of fatty acid inter-
mediates (fadB, acpM, kasA, kasB, accD6, fbpC),
The effects of drug-mediated enzyme inhib- removal or impact of toxic intermediates
ition may influence multiple metabolic pro- (aphC, efpA, iniB/A/C) and maintenance of
cesses in the mycobacterial cell, all of which membrane potential and the cell wall
may be reflected in the changing M. tubercu- structure.
losis transcriptome. Here, we describe briefly Thus, besides identifying drug mode of
the mechanism of INH killing to exemplify action, mapping the transcriptional responses
the complexity of drug-induced stresses and to antimicrobial compounds acts as a tool for
how this may contribute to the INH-induced exploring complex patterns of differential
transcriptional signature. These consider- gene regulation in response to multiple
FAS-I products
FAS-II
Fig. 13.2. The multifactorial impact of INH exposure on M. tuberculosis bacilli, highlighting the multiple
stimuli that likely contribute to the complex drug-induced transcriptional signature.
stimuli. For example, in addition to identi- inhA also in the FAS-II cycle) was after 6 h
fying molecular markers for cell cycle arrest, treatment with 5 MIC INH. Thus, the con-
chemical inhibition of FtsZ and FtsI high- trasting kinetics of these responses highlights
lighted regulatory networks associated with the different primary targets of these FAS-II
the cessation of septum formation and myco- inhibitory drugs (Betts et al., 2003). The length
bacterial replication (Slayden et al., 2006; of antimycobacterial drug treatment clearly
Slayden and Belisle, 2009). impacts on the transcriptional response
observed, in most cases with a greater number
of genes (with increasing magnitude) differ-
13.4 Temporal and Concentration- entially expressed with time compared to
dependent Signatures drug-free control cultures. Commonly selected
time points range from 2 to 6 h post-exposure
A number of additional parameters, other to define the initial drug-induced modifi-
than mechanism of action, influence the cations. This relatively discrete primary
bacterial response to drug exposure, includ- response is likely to be most useful for
ing secondary effects that result from bacterial recognizing the key pathways targeted. Later
killing. Modification of experimental con- time intervals of 12–48 h, incorporating the
ditions may be used to define drug action secondary effects of drug exposure as bacilli
further and the regulation of gene expression are killed, may be more applicable to
in M. tuberculosis. For example, the kas operon distinguishing between compound classes.
(Rv2243–Rv2247) was induced maximally The concentration of compound administered
after 2 h exposure to 1 MIC thiolactomycin may also influence differential expression
(which targets the β-ketoacyl-ACP synthases patterns, with an increase in dose often
encoded by kasA and kasB), while the highest correlating with greater transcriptional regu-
induction of the kas operon by INH (which lation (both magnitude of expression and
inhibits the enoyl-ACP reductase coded by number of genes). MICs from 1 to 50 are
routinely used, with 10 MIC a common feedback or compensatory mechanisms that
starting point (Boshoff et al., 2004; Waddell et manifest in detectable differences in relative
al., 2004). Subinhibitory concentrations usually transcript abundance.
accompanied by longer time intervals have
also been assessed; such as 0.5 MIC
mefloquine for 24 h (Danelishvili et al., 2005), 13.5 Characterizing M. tuberculosis
or 0.5 MIC and 0.25 MIC ciprofloxacin for Drug Resistance and Drug Tolerance
24 h (O’Sullivan et al., 2008). Both temporal
and dose-dependent signatures are contingent The gene expression signatures of drug-
on the dynamics of mycobacterial killing that resistant and drug-sensitive M. tuberculosis
will vary for each compound profiled. The strains have been compared successfully to
metabolic state of bacilli will also mediate the highlight putative mechanisms of drug
efficacy and transcriptional adaptations to resistance. Milano and colleagues contrasted
drug exposure. This will be discussed further the transcriptional profiles of a spontaneous
in the following sections; suffice to say that bifonazole-resistant M. tuberculosis mutant
most studies describe signatures derived with sensitive wild-type M. tuberculosis (in
from log-phase bacilli, as this is the setting the absence of drug) to reveal overexpression
where maximal antimicrobial activity is of the gene cluster Rv0676c (mmpL5), Rv0677c
usually achieved. (mmpS5) and Rv0678 in the resistant mutant.
Technical parameters such as the choice This gene cluster has been hypothesized to
of assay (quantitative RT-PCR, microarray function as a mycobacterial efflux system.
platform, RNAseq) also contribute to varia- Mutations were identified in the resistant
tion between studies and must be considered. strain within and upstream of Rv0678 which
However, a high degree of correlation could be linked to the overexpression of these
between gene expression patterns determined genes, thus conferring a resistant phenotype
by qRT-PCR and microarray methodologies (Milano et al., 2009). Interestingly, the same
and the identification of core sets of genes gene cluster was also observed to be
induced by the same drug in multiple reports, upregulated in drug-sensitive bacilli after
for example INH (Wilson et al., 1999; Betts et exposure to tetrahydrolipstatin (Waddell et
al., 2003; Boshoff et al., 2004; Waddell et al., al., 2004) and a range of other antibiotics
2004; Fu, 2006), demonstrate that the technical (Boshoff et al., 2004), perhaps highlighting
variation between methodologies can be this transport system as a multi-solute efflux
minimized. Other experimental details such pump with a role in mediating drug tolerance
as the stabilization of RNA during extraction in M. tuberculosis. Where drugs require con-
may contribute significantly to variation in version into active compounds, examining
signatures (Waddell and Butcher, 2010). Gene the response of resistant mutants to drug
expression analyses are, in most cases, com- exposure may help to define the mechanism
parative; therefore, the choice of comparator of resistance. For example, no genes were
condition also impacts on the divergent significantly differentially expressed after
transcriptional patterns generated. Studies exposure of a catalase-negative, INH-resistant
usually contrast drug-exposed to drug-free M. tuberculosis strain to INH (Wilson et al.,
or, more usefully, carrier-control cultures, 1999). In contrast, a characteristic INH signa-
where the same volume of solvent used to ture was observed after treatment of an INH-
re-suspend the antimicrobial compounds resistant M. tuberculosis strain with partial
(usually sterilized water, ethanol or DMSO) catalase activity with high dose INH (10 μg/
acts as the control. Finally, the gene expression ml), which was not present after exposure to
signature will be biased fundamentally low dose INH (0.2 μg/ml), reflecting the
towards the pathways that are regulated at functional significance of KatG in drug
the transcriptional level. The accumulation of activation (Fu and Shinnick, 2007).
precursors or the deprivation of intermediates The response of drug-sensitive bacilli to
resulting from drug action will only be antimicrobial compounds may also be ex-
reflected through transcriptionally regulated ploited to recognize processes that influence
drug tolerance, mechanisms through which example, identifying co-targets for drug
genetically drug-sensitive bacilli are refrac- development to help prevent the emergence
tory to killing. Alland et al. (1998) identified a of resistance to drugs targeting mycolic acid
cluster of genes Rv0341–Rv0343 (iniB/A/C) biosynthesis (RecA, Rv0823c, Rv0892 and
that were induced after exposure to INH and DnaE1).
which subsequently were demonstrated to be In an alternative approach to delineating
upregulated by multiple cell wall-targeting M. tuberculosis tolerance, Tudo et al. (2010)
drugs (Wilson et al., 1999; Alland et al., 2000; mapped pathways that might be responsible
Betts et al., 2003; Boshoff et al., 2004; Waddell for a drug-tolerant phenotype by comparing
et al., 2004). Functional characterization re- the transcriptional signatures derived from
vealed that deletion of this gene cluster, likely log-phase bacilli (which are effectively killed
encoding an MDR-like pump, resulted in by INH) with non-replicating persistent
hypersensitivity to INH, while overexpression (NRP) state 2 bacilli (which are tolerant to
conferred resistance to INH and EMB INH) in the presence and absence of the drug.
(Colangeli et al., 2005). Factors affecting drug As expected, the INH-tolerant NRP2 bacilli
tolerance are not restricted to efflux systems, did not respond transcriptionally to INH
but may also act by detoxifying specific treatment. Furthermore, the differential
compounds. For example, the gene cluster expression of genes implicated in resistance
Rv3160c–Rv3162c (encoding a putative tran- to INH (such as inhA, katG, iniB/A/C, fbpC)
scriptional regulator, probable ring de- that might affect the action of INH could not
hydroxylating dioxygenase and possible account for the abrogation of INH-mediated
membrane protein, respectively) was observed killing, suggesting that the changing redox
to be induced by triclosan (Betts et al., 2003), status or reduced requirement for newly
SRI No 967, SRI No 9190 (Waddell et al., 2004) synthesized mycolic acids likely influenced
and thioridazine (Dutta et al., 2010b). INH efficacy in non-replicating bacilli (Tudo
Mediators that affect global tran- et al., 2010). Similarly, Karakousis and co-
scriptional responses have also been identi- workers demonstrated that the INH-inducible
fied to influence the basal levels of transcriptional signature diminished in bacilli
mycobacterial tolerance to antibiotics. For isolated from a mouse hollow fibre model
example, induction of the transcriptional representing late-stage infection. The loss of
regulatory protein WhiB7, controlling the this INH signature correlated with the
expression of several genes implicated in reduced capacity of INH to kill bacilli as
intrinsic antibiotic resistance (Rv1258c, infection continued, and as such may be a key
Rv1473 and Rv1988 encoding a putative efflux indicator of drug tolerance in vivo (Karakousis
pump, a probable macrolide transporter and et al., 2008).
a possible ribosomal methyltransferase,
respectively), was observed after exposure of
M. tuberculosis to sub-inhibitory concen- 13.6 Defining Relevant In Vitro
trations of erythromycin or tetracycline Models for Whole Cell Screening
(Morris et al., 2005). Likewise, the gene
encoding sigma factor E (which responds to Besides exploring mechanisms of drug killing
cell wall-related stresses) was upregulated and mycobacterial resistance, transcriptional
after peptidoglycan-targeting vancomycin profiling has been used to characterize axenic
treatment; furthermore, deletion of sigE conditions modelling in vivo environments.
resulted in an increased sensitivity to These analyses serve to define and mimic the
vancomycin (Provvedi et al., 2009). The global conditions that M. tuberculosis bacilli are
nature of mRNA profiling was utilized by exposed to during natural infection. Thus, in
Raman and Chandra (2008), who integrated vitro culture systems have allowed the
the transcriptional responses of M. tuberculosis dissection of transcriptional profiles of M.
after drug exposure with protein–protein tuberculosis isolated from intracellular or in
interaction networks to identify novel path- vivo models of infection (Schnappinger et al.,
ways that might influence drug efficacy; for 2003; Talaat et al., 2004; Rachman et al., 2006;
Garton et al., 2008; Tailleux et al., 2008). These metabolism after macrophage infection has
models reveal complex patterns of divergent revealed exciting new enzymatic steps for
gene expression, reflecting changes in M. target-based high-throughput screening
tuberculosis respiratory and metabolic state, (McKinney et al., 2000; Schnappinger et al.,
recently reviewed in Waddell (2010). In a 2003; van der Geize et al., 2007). These and
complementary approach, chemostat models similar studies, defining the adaptations of
have been used to capture the responses of M. M. tuberculosis bacilli to intracellular and in
tuberculosis bacilli to specific challenges, such vivo microenvironments, also promote the
as carbon starvation (Hampshire et al., 2004) identification of novel M. tuberculosis targets
and low oxygen (Bacon et al., 2004), while centred on exploiting the interactions of
nutrient starvation (Betts et al., 2002), oxygen bacilli with the host immune system (Nathan
limitation (Rustad et al., 2008), phosphate et al., 2008). Beyond this, the integration of
depletion (Rifat et al., 2009) and the transition these transcriptional data sets with global
to non-replicating persistence (Muttucumaru gene essentiality studies (Sassetti and Rubin,
et al., 2004; Garton et al., 2008) have also been 2003; Sassetti et al., 2003; Dutta et al., 2010a)
characterized by microarray analyses. In has allowed the suitability of every enzyme
addition, multi-stress in vitro culture con- in the M. tuberculosis genome to be evaluated,
ditions (low oxygen, high CO2, low nutrients formulating hierarchies of potentially drug-
and acidic pH) have been developed and gable targets (Raman et al., 2008). Further-
modelled using transcriptional profiling to more, additional key enzymatic steps that
mimic in vivo conditions in axenic culture may be amenable to drug inhibition have
(Deb et al., 2009). In this way, global RNA been proposed by modelling the interactions
profiling may identify, define and then verify of metabolic pathways or the changing flux of
in vitro modelling conditions that imitate the carbon after drug exposure or during
stresses that bacilli experience during natural infection (Raman et al., 2005; Shi et al., 2010).
infection, enabling the development of rele-
vant whole cell screens for novel antimyco-
bacterial compounds. The requirement for 13.8 Conclusion: Transcriptional
appropriate screening conditions was Profiling Through the Drug Discovery
highlighted recently by Pethe and colleagues, Pipeline
who identified inhibitors that were effective
in vitro; with no activity in the murine model The global response of M. tuberculosis to
of infection, however. These novel pyrimidine- antimicrobial compounds is useful in
imidazole compounds targeted glycerol meta- multiple stages of the drug discovery process,
bolism used by mycobacteria in vitro, but illustrated in Fig. 13.3. Early in the pipe-
evidently not essential to bacterial metabolism line, in vitro models optimized for RNA
in vivo (Pethe et al., 2010). Therefore, tran- profiling are often the first opportunity to
scriptional profiling offers the opportunity to assess compound attributes such as solubility
recognize and characterize conditions that M. or stability outside the whole cell or high-
tuberculosis bacilli must adapt to in vivo which throughput screens from which the com-
are amenable to high-throughput screening. pounds are selected. The kinetics of
mycobacterial killing in relation to con-
centration and time may also highlight key
13.7 Transcriptomic Approaches to features of drug action. Drug-specific mRNA
Target Identification signatures may then identify or confirm
mode of action (a key step in drug develop-
Global RNA abundance data sets provide an ment), contrasting M. tuberculosis responses
interpretive framework to discover, assess to an interpretive transcriptional framework
and rank targets in a pathway-led approach defining compound class or target knock-
to drug development. For example, the down (Boshoff et al., 2004). These profiles also
induction of genes involved in the β-oxidation allow potentially problematic mutagenic or
of fatty acids, glyoxylate shunt and cholesterol nitrosative reactions to be recognized early in
the drug development process (Makarov et led to the development of EthR inhibitors
al., 2009). Finally, as new analogues for a lead (which block the repression of EthA by EthR),
compound are developed and tested, global resulting in increased efficacy of thio-
transcriptome analysis provides a useful carbamide derivatives (Willand et al., 2009).
checkpoint enabling target drift to be Moreover, β-lactam antibiotics have been
detected. This was the case in a screen of demonstrated to be effective against M.
diamine analogues (modelled on the structure tuberculosis if combined with β-lactamase
of EMB) that identified a number of hit inhibitors that limit the destruction of peni-
compounds (including SQ109) with M. cillins by the hydrolysing activity of chromo-
tuberculosis-inhibitory properties. Tran- somally encoded β-lactamase, blaC (Cynamon
scriptional profiling revealed related but and Palmer, 1983; Hugonnet et al., 2009).
distinct cell wall-targeting signatures between As the drug discovery pipeline is applied
these diamine analogues and EMB, high- increasingly to search for novel compounds
lighting a shift in the mechanism of action of to control the growing threat of MDR-TB,
these new effective compounds (Boshoff et al., transcriptional profiling will continue to play
2004; Barry and Blanchard, 2010). an important role characterizing the complex
A greater understanding of drug action in interactions between drug action and myco-
M. tuberculosis facilitates novel chemo- bacterial physiological state.
therapeutic approaches; for example, dis-
covery of the mechanism of BTZ-mediated
killing (Makarov et al., 2009) revealed DprE1/ Acknowledgements
E2 (encoded by Rv3790/3791) to be an excellent
target that may be amenable to target-based SJW was supported by the European Com-
screening (Manina et al., 2010). Similarly, mission ‘New Medicines for TB – NM4TB’
characterization of gene products involved in programme (LHSP-CT-2005-018923). PDB
the conversion or inactivation of antimicrobial would like to thank the Wellcome Trust for
compounds that are pinpointed by transcrip- funding the Bacterial Microarray Group at St
tional profiling may precipitate alternative George’s (Grant Nos 062511, 080039 and
therapeutic strategies. For example, demon- 086547).
stration that the monooxygenase EthA is The authors declare no conflicting
required for activation of the prodrug ETA has financial interests.
New
Confirmation of drug-target selectivity; drugs
profiling drug versus gene knockout/down
Fig. 13.3. The utility of transcriptional profiling through the drug discovery pipeline. Defining
mycobacterial responses to test compounds by microarray analysis is of use at multiple stages of both
phenotype and target-based screening approaches; adapted from Terstappen et al. (2007).
References Journal of Biological Chemistry 279, 40174–

40184.
Alland, D., Kramnik, I., Weisbrod, T.R., Otsubo, L., Colangeli, R., Helb, D., Sridharan, S., Sun, J.,
Cerny, R., Miller, L.P., et al. (1998) Identification Varma-Basil, M., Hazbon, M.H., et al. (2005)
of differentially expressed mRNA in prokaryotic The Mycobacterium tuberculosis iniA gene is
organisms by customized amplification libraries essential for activity of an efflux pump that
(DECAL): the effect of isoniazid on gene confers drug tolerance to both isoniazid and
expression in Mycobacterium tuberculosis. ethambutol. Molecular Microbiology 55, 1829–
Proceedings of the National Academy of 1840.
Sciences 95, 13227–13232. Cynamon, M.H. and Palmer, G.S. (1983) In vitro
Alland, D., Steyn, A.J., Weisbrod, T., Aldrich, K. and activity of amoxicillin in combination with
Jacobs, W.R. Jr (2000) Characterization of the clavulanic acid against Mycobacterium tubercu-
Mycobacterium tuberculosis iniBAC promoter, a losis. Antimicrobial Agents and Chemotherapy
promoter that responds to cell wall biosynthesis 24, 429–431.
inhibition. Journal of Bacteriology 182, 1802– Danelishvili, L., Wu, M., Young, L.S. and Bermudez,
1811. L.E. (2005) Genomic approach to identifying the
Argyrou, A., Vetting, M.W., Aladegbami, B. and putative target of and mechanisms of resistance
Blanchard, J.S. (2006) Mycobacterium tubercu- to mefloquine in mycobacteria. Antimicrobial
losis dihydrofolate reductase is a target for
Agents and Chemotherapy 49, 3707–3714.
isoniazid. Nature Structural and Molecular
Deb, C., Lee, C.M., Dubey, V.S., Daniel, J.,
Biology 13, 408–413.
Abomoelak, B., Sirakova, T.D., et al. (2009) A
Bacon, J., James, B.W., Wernisch, L., Williams, A.,
novel in vitro multiple-stress dormancy model
Morley, K.A., Hatch, G.J., et al. (2004) The
for Mycobacterium tuberculosis generates a
influence of reduced oxygen availability on
lipid-loaded, drug-tolerant, dormant pathogen.
pathogenicity and gene expression in Myco-
PLoS ONE 4, e6077.
bacterium tuberculosis. Tuberculosis (Edin-
Dutta, N.K., Mehra, S., Didier, P.J., Roy, C.J., Doyle,
burgh) 84, 205–217.
L.A., Alvarez, X., et al. (2010a) Genetic
Barry, C.E. 3rd and Blanchard, J.S. (2010) The
requirements for the survival of tubercle bacilli
chemical biology of new drugs in the
development for tuberculosis. Current Opinion in primates. Journal of Infectious Diseases 201,
in Chemical Biology 14, 456–466. 1743–1752.
Belanger, A.E., Besra, G.S., Ford, M.E., Mikusova, Dutta, N.K., Mehra, S. and Kaushal, D. (2010b) A
K., Belisle, J.T., Brennan, P.J., et al. (1996) The Mycobacterium tuberculosis sigma factor
embAB genes of Mycobacterium avium encode network responds to cell-envelope damage by
an arabinosyl transferase involved in cell wall the promising anti-mycobacterial thioridazine.
arabinan biosynthesis that is the target for the PLoS ONE 5, e10069.
antimycobacterial drug ethambutol. Proceed- Fu, L.M. (2006) Exploring drug action on
ings of the National Academy of Sciences 93, Mycobacterium tuberculosis using affymetrix
11919–11924. oligonucleotide genechips. Tuberculosis (Edin-
Betts, J.C., Lukey, P.T., Robb, L.C., McAdam, R.A. burgh) 86, 134–143.
and Duncan, K. (2002) Evaluation of a nutrient Fu, L.M. and Shinnick, T.M. (2007) Understanding
starvation model of Mycobacterium tuberculosis the action of INH on a highly INH-resistant
persistence by gene and protein expression Mycobacterium tuberculosis strain using
profiling. Molecular Microbiology 43, 717–731. Genechips. Tuberculosis (Edinburgh) 87,
Betts, J.C., McLaren, A., Lennon, M.G., Kelly, F.M., 63–70.
Lukey, P.T., Blakemore, S.J., et al. (2003) Garton, N.J., Waddell, S.J., Sherratt, A.L., Lee,
Signature gene expression profiles discriminate S.M., Smith, R.J., Senner, C., et al. (2008)
between isoniazid-, thiolactomycin-, and Cytological and transcript analyses reveal fat
triclosan-treated Mycobacterium tuberculosis. and lazy persister-like bacilli in tuberculous
Antimicrobial Agents and Chemotherapy 47, sputum. PLoS Medicine 5, e75.
2903–2913. Hampshire, T., Soneji, S., Bacon, J., James, B.W.,
Boshoff, H.I., Myers, T.G., Copp, B.R., McNeil, Hinds, J., Laing, K., et al. (2004) Stationary
M.R., Wilson, M.A. and Barry, C.E. 3rd (2004) phase gene expression of Mycobacterium
The transcriptional responses of Mycobacterium tuberculosis following a progressive nutrient
tuberculosis to inhibitors of metabolism: novel depletion: a model for persistent organisms?
insights into drug mechanisms of action. The Tuberculosis (Edinburgh) 84, 228–238.
Hugonnet, J.E., Tremblay, L.W., Boshoff, H.I., Barry, Pethe, K., Sequeira, P.C., Agarwalla, S., Rhee, K.,
C.E. 3rd and Blanchard, J.S. (2009) Meropenem- Kuhen, K., Phong, W.Y., et al. (2010) A chemical
clavulanate is effective against extensively genetic screen in Mycobacterium tuberculosis
drug-resistant Mycobacterium tuberculosis. identifies carbon-source-dependent growth
Science 323, 1215–1218. inhibitors devoid of in vivo efficacy. Nature
Karakousis, P.C., Williams, E.P. and Bishai, W.R. Communications 1, 1–8.
(2008) Altered expression of isoniazid-regulated Provvedi, R., Boldrin, F., Falciani, F., Palu, G. and
genes in drug-treated dormant Mycobacterium Manganelli, R. (2009) Global transcriptional
tuberculosis. Journal of Antimicrobial Chemo- response to vancomycin in Mycobacterium
therapy 61, 323–331. tuberculosis. Microbiology 155, 1093–1102.
McKinney, J.D., Honer zu Bentrup, K., Munoz-Elias, Rachman, H., Strong, M., Ulrichs, T., Grode, L.,
E.J., Miczak, A., Chen, B., Chan, W.T., et al. Schuchhardt, J., Mollenkopf, H., et al. (2006)
(2000) Persistence of Mycobacterium tubercu- Unique transcriptome signature of Myco-
losis in macrophages and mice requires the bacterium tuberculosis in pulmonary tubercu-
glyoxylate shunt enzyme isocitrate lyase. Nature losis. Infection and Immunity 74, 1233–1242.
406, 735–758. Raman, K. and Chandra, N. (2008) Mycobacterium
Makarov, V., Manina, G., Mikusova, K., Mollmann, tuberculosis interactome analysis unravels
U., Ryabova, O., Saint-Joanis, B., et al. (2009) potential pathways to drug resistance. BMC
Benzothiazinones kill Mycobacterium tubercu- Microbiology 8, 234.
losis by blocking arabinan synthesis. Science Raman, K., Rajagopalan, P. and Chandra, N. (2005)
324, 801–804. Flux balance analysis of mycolic acid pathway:
Manina, G., Pasca, M.R., Buroni, S., De Rossi, E.
targets for anti-tubercular drugs. PLoS
and Riccardi, G. (2010) Decaprenylphosphoryl-
Computational Biology 1, e46.
beta-d-ribose 2’-epimerase from Mycobacterium
Raman, K., Yeturu, K. and Chandra, N. (2008)
tuberculosis is a magic drug target. Current
TargetTB: a target identification pipeline for
Medicinal Chemistry 17, 3099–3108.
Mycobacterium tuberculosis through an inter-
Manjunatha, U., Boshoff, H.I. and Barry, C.E. (2009)
actome, reactome and genome-scale structural
The mechanism of action of PA-824: novel
analysis. BMC Systems Biology 2, 109.
insights from transcriptional profiling. Com-
Rifat, D., Bishai, W.R. and Karakousis, P.C. (2009)
municative and Integrative Biology 2, 215–218.
Phosphate depletion: a novel trigger for
Milano, A., Pasca, M.R., Provvedi, R., Lucarelli,
Mycobacterium tuberculosis persistence.
A.P., Manina, G., Ribeiro, A.L., et al. (2009)
Azole resistance in Mycobacterium tuberculosis Journal of Infectious Diseases 200, 1126–1135.
is mediated by the MmpS5-MmpL5 efflux Rustad, T.R., Harrell, M.I., Liao, R. and Sherman,
system. Tuberculosis (Edinburgh) 89, 84–90. D.R. (2008) The enduring hypoxic response of
Morris, R.P., Nguyen, L., Gatfield, J., Visconti, K., Mycobacterium tuberculosis. PLoS ONE 3,
Nguyen, K., Schnappinger, D., et al. (2005) e1502.
Ancestral antibiotic resistance in Mycobacterium Sacchettini, J.C., Rubin, E.J. and Freundlich, J.S.
tuberculosis. Proceedings of the National (2008) Drugs versus bugs: in pursuit of the
Academy of Sciences 102, 12200–12205. persistent predator Mycobacterium tuberculosis.
Muttucumaru, D.G., Roberts, G., Hinds, J., Stabler, Nature Reviews Microbiology 6, 41–52.
R.A. and Parish, T. (2004) Gene expression Sassetti, C.M. and Rubin, E.J. (2003) Genetic
profile of Mycobacterium tuberculosis in a non- requirements for mycobacterial survival during
replicating state. Tuberculosis (Edinburgh) 84, infection. Proceedings of the National Academy
239–246. of Sciences 100, 12989–12994.
Nathan, C., Gold, B., Lin, G., Stegman, M., De Sassetti, C.M., Boyd, D.H. and Rubin, E.J. (2003)
Carvalho, L.P., Vandal, O., et al. (2008) A Genes required for mycobacterial growth
philosophy of anti-infectives as a guide in the defined by high density mutagenesis. Molecular
search for new drugs for tuberculosis. Microbiology 48, 77–84.
Tuberculosis (Edinburgh) 88 (Suppl 1), S25–33. Schnappinger, D., Ehrt, S., Voskuil, M.I., Liu, Y.,
O’Sullivan, D.M., Hinds, J., Butcher, P.D., Gillespie, Mangan, J.A., Monahan, I.M., et al. (2003)
S.H. and McHugh, T.D. (2008) Mycobacterium Transcriptional adaptation of Mycobacterium
tuberculosis DNA repair in response to tuberculosis within macrophages: insights into
subinhibitory concentrations of ciprofloxacin. the phagosomal environment. The Journal of
Journal of Antimicrobial Chemotherapy 62, Experimental Medicine 198, 693–704.
1199–1202. Shi, L., Sohaskey, C.D., Pfeiffer, C., Datta, P., Parks,
M., McFadden, J., et al. (2010) Carbon flux profiling. Future Medicinal Chemistry 2, 1371–
rerouting during Mycobacterium tuberculosis 1383.
growth arrest. Molecular Microbiology 78, Van der Geize, R., Yam, K., Heuser, T., Wilbrink,
1199–1215. M.H., Hara, H., Anderton, M.C., et al. (2007) A
Slayden, R.A. and Belisle, J.T. (2009) Morphological gene cluster encoding cholesterol catabolism in
features and signature gene response elicited a soil actinomycete provides insight into
by inactivation of FtsI in Mycobacterium Mycobacterium tuberculosis survival in macro-
tuberculosis. Journal of Antimicrobial Chemo- phages. Proceedings of the National Academy
therapy 63, 451–457. of Sciences 104, 1947–1952.
Slayden, R.A., Knudson, D.L. and Belisle, J.T. Vilcheze, C. and Jacobs, W.R. Jr (2007) The
(2006) Identification of cell cycle regulators in mechanism of isoniazid killing: clarity through
Mycobacterium tuberculosis by inhibition of the scope of genetics. Annual Review of
septum formation and global transcriptional Microbiology 61, 35–50.
analysis. Microbiology 152, 1789–1797. Waddell, S.J. (2010) Reprogramming the Myco-
Tailleux, L., Waddell, S.J., Pelizzola, M., Mortellaro, bacterium tuberculosis transcriptome during
A., Withers, M., Tanne, A., et al. (2008) Probing host pathogenesis. Drug Discovery Today:
host pathogen cross-talk by transcriptional Disease Mechanisms 7, e67–e73.
profiling of both Mycobacterium tuberculosis Waddell, S.J. and Butcher, P.D. (2010) Use of DNA
and infected human dendritic cells and arrays to study transcriptional responses to
macrophages. PLoS ONE 3, e1403. antimycobacterial compounds. Methods in
Talaat, A.M., Lyons, R., Howard, S.T. and Johnston, Molecular Biology 642, 75–91.
S.A. (2004) The temporal expression profile of Waddell, S.J., Stabler, R.A., Laing, K., Kremer, L.,
Mycobacterium tuberculosis infection in mice. Reynolds, R.C. and Besra, G.S. (2004) The use
Proceedings of the National Academy of of microarray analysis to determine the gene
Sciences 101, 4602–4607. expression profiles of Mycobacterium tubercu-
Terstappen, G.C., Schlupen, C., Raggiaschi, R. and losis in response to anti-bacterial compounds.
Gaviraghi, G. (2007) Target deconvolution Tuberculosis (Edinburgh) 84, 263–274.
strategies in drug discovery. Nature Reviews Wang, F., Jain, P., Gulten, G., Liu, Z., Feng, Y.,
Drug Discovery 6, 891–903. Ganesula, K., et al. (2010) Mycobacterium
Timmins, G.S. and Deretic, V. (2006) Mechanisms tuberculosis dihydrofolate reductase is not a
of action of isoniazid. Molecular Microbiology target relevant to the antitubercular activity of
62, 1220–1227. isoniazid. Antimicrobial Agents and Chemo-
Trefzer, C., Rengifo-Gonzalez, M., Hinner, M.J., therapy 54, 3776–3782.
Schneider, P., Makarov, V., Cole, S.T., et al. Willand, N., Dirie, B., Carette, X., Bifani, P., Singhal,
(2010) Benzothiazinones: prodrugs that A., Desroses, M., et al. (2009) Synthetic EthR
covalently modify the decaprenylphosphoryl- inhibitors boost antituberculous activity of
beta-d-ribose 2’-epimerase DprE1 of Myco- ethionamide. Nature Medicine 15, 537–544.
bacterium tuberculosis. Journal of the American Wilson, M., Derisi, J., Kristensen, H.H., Imboden,
Chemical Society 132, 13663–13665. P., Rane, S., Brown, P.O., et al. (1999) Exploring
Tudo, G., Laing, K., Mitchison, D.A., Butcher, P.D. drug-induced alterations in gene expression in
and Waddell, S.J. (2010) Examining the basis of Mycobacterium tuberculosis by microarray
isoniazid tolerance in non-replicating Myco- hybridization. Proceedings of the National
bacterium tuberculosis using transcriptional Academy of Sciences 96, 12833–12838.
14 Mycobacterial Lipid Bodies and
Resuscitation-promoting Factor
Dependency as Potential Biomarkers of
Response to Chemotherapy
N.J. Garton,1 G.V. Mukamolova1 and M.R. Barer1,2

1Department
of Infection, Immunity and Inflammation, University of Leicester
College of Medicine, Biological Sciences and Psychology, Leicester, UK;
2Department of Clinical Microbiology, University Hospitals of Leicester NHS Trust,
Leicester, UK
14.1 Introduction non-replicating state in host tissues. Although

genetically susceptible to antibiotics, M.
Current chemotherapy regimens for the tuberculosis bacilli in a non-replicating persist-
treatment of tuberculosis are lengthy and ent state exhibit phenotypic antibiotic resist-
generally require a combination of four anti- ance (Connolly et al., 2007; Dhar and
biotics. This can lead to a breakdown in McKinney, 2007), and it is this trait which
patient compliance and an increased risk of complicates tuberculosis therapy. The
developing multidrug-resistant Mycobacterium ‘persister’ population in infection has never
tuberculosis strains. The development of new been directly identified, but is inferred from
therapeutic agents and regimes with the the biphasic reduction of viable counts
potential to reduce the length of treatment is recovered from serial sputum samples
therefore an ongoing major focus of research. collected during therapy (Mitchison, 2004).
An additional goal is the identification and The factors which promote M. tuberculosis
assessment of prognostic biomarkers of persistence are not understood, but hostile
treatment response and clinical end points host conditions within the caseating lung
(Wallis et al., 2010). granulomas such as hypoxia, low pH and
The current requirement for a prolonged nutrient limitation may contribute. When
multidrug regimen is a likely consequence these lesions liquefy and erode into the lower
of the heterogeneous physiological states of airways, cavities are formed. The cavitary
the causative M. tuberculosis bacilli in the lesions support rapid bacillary replication,
infected host (Mitchison, 1979, 2004). Rapidly resulting in continuous discharge of bacilli
replicating M. tuberculosis bacilli are killed expectorated in sputum. In addition to being
relatively quickly through the action of the representative of those bacilli that go on to
frontline antibiotics, isoniazid (INH) and transmit disease, the bacilli in sputum
rifampicin (rifampin; RIF); however, it is represent a sample of those which must be
believed that M. tuberculosis can persist in a eliminated by chemotherapy. The findings of

Lipid Bodies and Resuscitation-promoting Factor Dependency 185
our studies of the phenotypes of M. tubercu- (NRP2) of the Wayne dormancy model
losis bacilli in sputum challenge the long-held (Wayne and Hayes, 1996; Muttucamaru et al.,
belief that the bacilli are replicating rapidly; 2004; Voskuil et al., 2004) showed significant
instead, available evidence supports the overlaps, with genes downregulated in the
hypothesis that they are in a slowly growing sputum transcriptome. Second, there were
or non-replicating state, which we have significant similarities with the transcriptome
termed persister-like (Garton et al., 2002, of M. tuberculosis during chronic murine
2008; Mukamolova et al., 2010). As smear- infection, a point at which exponential
positive sputum must be populated regularly bacterial replication had ceased. The M.
with M. tuberculosis, we hypothesize that at tuberculosis transcripts associated with
some point during the relatively short time chronic murine infection reflect a shift from
taken for discharge from the cavity to aerobic to anaerobic respiration (Shi et al.,
expectoration, the bacilli adapt to a persister- 2005).
like state. The environmental signals which
lead to this adaptation are not known. Here,
we discuss the sputum phenotypes we have 14.2.2 Lipid bodies
identified and their potential as biomarkers
for the assessment of patient response to Lipid bodies (LBs) are neutral lipid-con-
therapy. taining (triacylglycerol (TAG) and wax ester
(WE)) inclusions in actinobacteria; they were
recently recognized as the bacterial equiva-
14.2 Sputum Phenotypes lent of the LBs that occur in eukaryotes
(Wältermann and Steinbüchel, 2005). We first
14.2.1 Slow/non-growth identified LBs by fluorescence microscopy
following Nile Red staining of the rapidly
Microarray analysis of the global tran- growing non-pathogenic strain M. smegmatis
scriptome of M. tuberculosis in sputum and noted that the presence and appearance
expectorated by four individual smear- of LBs, and associated TAG fatty acid content,
positive patients, in comparison with that of reflected the culture conditions (Garton et al.,
in vitro aerobic log-phase growth of the M. 2002). By application of Nile Red in com-
tuberculosis strain H37Rv, revealed many bination with auramine acid-fast staining, we
transcript signatures suggestive of slow or observed that, in contrast to growing labora-
non-growth (Garton et al., 2008). Strikingly tory cultures, M. tuberculosis cells in sputum
similar, the transcriptomes of the four M. contained well-defined LBs at frequencies
tuberculosis samples had a unique profile; no varying from 3% to 87% of the auramine-
single or combination of defined conditions staining population (Garton et al., 2002, 2008).
previously reported in vitro or in vivo cor- We were able to identify an LB-positive
responded. Compared with the log-phase subpopulation in most smear-positive sputum
H37Rv growth, there were significant samples examined this way, a complicating
decreases in the transcription of genes factor being high Nile Red background stain-
required for aerobic respiration, ribosomal ing of host-derived material in some samples.
function and ATP synthesis, with increased The same LBs can be revealed with an alter-
expression of genes belonging to the DosR native lipophilic reagent, LipidTox red neutral
regulon, which has been widely linked to lipid stain. Figure 14.1 shows LB-positive
dormancy (Voskuil et al., 2004). Two lines of acid-fast cells revealed with LipidTox red
evidence support the view that the sputum neutral lipid labelling of a smear-positive
transcriptome represents a population domin- tuberculous sample.
ated by slowly or non-replicating cells. First, The transcriptome of M. tuberculosis in
the in vitro transcriptomes of M. tuberculosis in sputum revealed the upregulation of Rv3130c
well-recognized growth arrest, those of (tgs1) (Garton et al., 2008), which encode the
nutrient starvation (Betts et al., 2002) and the most active of 15 TAG synthases (Daniel et al.,
second non-replicating persistence phase 2004). Overexpression of tgs1 in M. smegmatis
186 N.J. Garton et al.
phosphate limitation in a chemostat (Sherrat

et al., unpublished results). An increasing
population of LB-positive M. tuberculosis cells
has been identified throughout the duration
of a multiple-stress model of dormancy (Deb
et al., 2009), which involved incubation at low
O2, high CO2, nutrient deprivation and low
pH. Intriguingly, during the Deb multiple-
stress model, the increase in the proportion of
LB-positive cells was concomitant with a
decrease in the auramine acid-fast cell popu-
lation. We have made observations that are
compatible with the possible presence of such
cells in sputum (unpublished results) and are
developing methods for quantifying the
microscopically distinct sputum subpopu-
lations accurately. Together, these results
Fig. 14.1. Lipid bodies within acid-fast bacilli in an describing the LB-positive M. tuberculosis
auramine/LipidTox red neutral lipid labelled, fixed cells in conditions which limit growth and
tuberculous sputum smear. model persistence support our hypothesis
that such persister-like cells are present in
sputum (Garton et al., 2008).
and analysis of Nile Red staining of the
overexpressing and plasmid control strains
revealed that tgs1 was involved in LB 14.2.3 Resuscitation-promoting factor
formation; the tgs1 overexpressing strain dependency
contained larger, more intensely stained LBs
(Garton et al., 2008). A recent study aimed at The resuscitation-promoting factor (Rpf) was
identifying LB-associated proteins reported first described as a protein which stimulated
that tgs1 in addition to tgs2 (Rv3734c, the the growth and resuscitation of non-replicat-
second most active TAG synthase) were both ing cells of Micrococcus luteus (Mukamolova et
associated physically with LBs isolated from al., 1998). Rpf is an essential gene in M. luteus
hypoxic non-replicating M. bovis bacillus which is secreted into culture medium and
Calmette-Guérin (BCG) (Low et al., 2010). is also detected on the surface of growing
We have so far reported inability to cells (Mukamolova et al., 2002a). Following
detect LBs in growing M. tuberculosis cultures molecular characterization of the M. luteus
in vitro and their induction by conditions that Rpf gene, five homologues (RpfA, RpfB, RpfC,
produce growth arrest and stimulation of the RpfD and RpfE) were identified in M.
DosR regulon (of which tgs1 is a constituent). tuberculosis (Mukamolova et al., 2002b). The
These conditions include exposure to nitric M. tuberculosis recombinant Rpf proteins
oxide and Wayne-type non-replicating persist- showed similar activity, activating growth of
ence (Garton et al., 2008). This pattern of LB non-replicating cultures at picomolar con-
occurrence we have described has recently centrations. Following construction of Rpf
been substantiated in M. bovis BCG taken gene knockout strains, Rpfs have been
from the Wayne model (Low et al., 2009, demonstrated to be important for in vivo
2010). Rifat and colleagues report that phos- persistence and reactivation of chronic
phate limitation, a nutrient stress likely infection in mice (Downing et al., 2005;
encountered by M. tuberculosis in macro- Tufariello et al., 2006; Biketov et al., 2007;
phages, results in a dose-dependent restric- Russell-Goldman et al., 2008). The precise
tion in M. tuberculosis growth (Rifat et al., mechanism of action of these growth-
2009). We have also identified LB-positive stimulatory proteins is unknown, but they
cells in M. tuberculosis biomass produced by have been shown to possess structural motifs
similar to that of lysozyme and soluble lytic from 25 smear-positive tuberculosis patients
transglycosylases and to possess muralytic before initiation of chemotherapy and
activity (Cohen-Gonsaud et al., 2004, 2005; measurements of culturable cells by most
Keep et al., 2006; Mukamolova et al., 2006; probable number (MPN) assay were made
Kana and Mizrahi, 2010). Site-directed with, and without, the addition of recombin-
mutagenesis of the catalytic glutamate ant RpfE, RpfB or fresh Rpf-containing
diminishes this activity and the ability of the culture supernatants. Strikingly, these assays
proteins to stimulate growth and resusci- revealed that Rpf-dependent M. tuberculosis
tation, thereby suggesting that cleavage of cells were not only present but were, in fact,
peptidoglycan is related directly to growth the dominant population in 20 of the samples,
stimulation and resuscitation (Mukamolova measured at 80–99.9% of the culturable cells
et al., 2006; Telkov et al., 2006). The stresses present (Mukamolova et al., 2010). The
experienced by M. tuberculosis from the host specificity of these effects was further demon-
immune system during infection are likely to strated by evidence that the addition of fresh
result in cell wall damage and possible cell culture supernatant from a quintuple M.
wall remodelling. Accordingly, in vitro, it has tuberculosis Rpf mutant strain (Kana et al.,
been shown that, following infection in 2008) did not lead to the recovery of increased
macrophages, the morphology and surface cell numbers.
properties of M. tuberculosis change and Rpf dependency in sputum is a pheno-
culturability is reduced (Biketov et al., 2000). typic adaptation of M. tuberculosis cells to
Enhanced numbers of cells were recovered specific in vivo signals, since it is lost after
following the addition of Rpf to the culture primary isolation. One possibility is that
media. Similar morphological changes have these signals are present in sputum. A
been associated with the development of a candidate for such a signal was the inhibitory
population of M. tuberculosis which could not activity against M. tuberculosis growth in
be cultured on agar following extended liquid medium we observed in sputum.
incubation in stationary phase (Shleeva et al., Although we have been able to detect this
2002). Again, enhanced numbers of cells were inhibitory activity directly by the inoculation
recovered in liquid growth assays on the of M. tuberculosis H37Rv into two tuberculous
addition of Rpf to the culture media. Our sputum samples which had previously been
interpretation of these results is that Rpf rendered culture negative by freezing,
activity may be required to degrade a incubation of M. tuberculosis in this milieu did
modified peptidoglycan (associated with the not lead to the development of Rpf-dependent
altered cell morphology) to assist cell division cells. With the exception of extended culture
and regrowth. Alternatively, the degradation in stationary phase (Shleeva et al., 2002), the
of peptidoglycan by Rpfs will release muro- only other demonstration of Rpf-dependent
peptide fragments, which may have a role in cells was in a population recovered from the
providing a resuscitation signal via activation peritoneal macrophages of M. tuberculosis-
of a serine-threonine protein kinase signalling infected BALB/c mice (Biketov et al., 2000).
cascade (Jones and Dyson, 2006; Shah et al., These observations suggest that the adap-
2008; Shah and Dworkin, 2010). Evidence that tation leading to Rpf dependence is a
these events are significant in human response to as yet unidentified environmental
infection comes from the demonstration that signals in vivo.
Rpfs are expressed in M. tuberculosis-infected
human tissues (Davies et al., 2008).
Our transcriptome and LB studies of M. 14.3 Evidence for a Relationship
tuberculosis in sputum which provided Between Sputum M. tuberculosis
evidence to support the presence of non- Phenotypes and Antibiotic Tolerance
replicating M. tuberculosis bacilli (Garton et
al., 2008) led us to hypothesize that Rpf- Phenotypic antibiotic tolerance is a well-
dependent cells might be present in these recognized microbial trait and has been
samples. Sputum samples were collected recognized in many bacterial infections
(Levin and Rozen, 2006; Connolly et al., 2007; following NO treatment of a log-phase M.
Dhar and McKinney, 2007). The cidal action tuberculosis culture, the LB-positive sub-
of many antibiotics is proportional to the population increased to 65% compared with
growth rate of bacteria and metabolic activity; <1% in an untreated control. This is in accord
log-phase cells are killed efficiently, whereas with the reported maximal increase in TAG
slow-growing or non-replicating cells show content 4 h following NO stimulation (Daniel
phenotypic tolerance (Wallis et al., 1999; et al., 2004). The NO-stimulated cells sampled
Gomez and McKinney, 2004; Paramasivan et at this time showed nearly 50-fold greater
al., 2005; Dhar and McKinney, 2007). M. survival following exposure to 1 μg/ml RIF
tuberculosis exhibits phenotypic tolerance to than the untreated control (Garton et al.,
various antibiotics, including the frontline 2008). A sample of the same culture 24 h after
agent INH, in stationary phase compared NO stimulation had a reduced LB content of
with log-phase growth (Herbert et al., 1996; 22% of the total population and approxi-
Paramasivan et al., 2005). M. tuberculosis and mately 10-fold greater survival following RIF
M. bovis BCG conditioned in models designed treatment. This is consistent with the
to reflect the non-replicating state of the dissipation of NO from the culture, resulting
organism during latent infection, for example, in a resumption of growth and decrease in
Wayne’s model of non-replicating persistence TAG content (Daniel et al., 2004). In
in which cells gradually adapt to hypoxic subsequent experiments, we have repeatedly
conditions (Wayne and Hayes, 1996; Low et demonstrated enhanced TAG and LB levels
al., 2009) or the multiple stress model of Deb in NO treated cells and that these properties
and colleagues (Deb et al., 2009), are are associated with tolerance to INH; in
phenotypically tolerant to INH and RIF. The contrast, results with RIF have been
measured degree of tolerance to INH is inconsistent (Sherratt et al., unpublished
greater than to RIF. However, when these results). This would suggest that although
mycobacteria are subject to conditions which rapid growth arrest in response to NO
induce a more rapid growth arrest, such as the stimulation results in an increase in LB (TAG)
addition of nitric oxide (NO) (Hussain et al., content of the M. tuberculosis (and stimulation
2009) and phosphate limitation (Rifat et al., of the dormancy regulon), additional stimuli
2009), only tolerance to INH is observed. This or a more gradual adaption to slow/non-
is consistent with the target of RIF, RNA growth is required to promote tolerance to
polymerase, being active during growth arrest RIF in addition to INH. It is not known
(Wayne and Hayes, 1996; Hu et al., 2000). whether LBs are linked mechanistically to
antibiotic tolerance or are a biomarker for an
as yet unidentified property which would
14.3.1 Lipid bodies explain this. It is interesting to note, however,
that the RIF tolerance of M. tuberculosis in the
As non-replicating Wayne-type M. multiple-stress dormancy model reported by
tuberculosis cultures, which show the most Deb and colleagues, was reduced, but not
similar transcriptome signatures to M. eliminated, in a tgs1 mutant and restored by
tuberculosis in sputum and enhanced levels of complementation (Deb et al., 2009). This
LBs compared with aerobic growth, are suggests that TAG content of the M.
tolerant to the action of INH and RIF (Wayne tuberculosis does directly influence the RIF
and Hayes, 1996), we have examined the tolerance of the cells in these stress
action of these antibiotics against M. conditions. In addition to the development of
tuberculosis cultures exposed to NO. NO is an LBs, it is possible that there are associated
inhibitor of aerobic respiration and results in changes, for example the accumulation of
growth arrest and induction of the dormancy TAG, in the cell envelope, which may reduce
regulon of M. tuberculosis (Ohno et al., 2003; permeability to antibiotics. In this regard,
Voskuil et al., 2003; Daniel et al., 2004). We Cunningham and Spreadbury reported that
reported initial results showing that 4 h M. tuberculosis sampled from the Wayne
model, a condition in which we have whether this observation is reproducible and

demonstrated LB positivity, have thickened whether it is reflected in clinical responses.
cell envelopes (Cunningham and Spreadbury,
1998).
It remains to be investigated whether 14.3.2 Rpf dependency
there are any clinical correlates with the LB-
positive M. tuberculosis population observed Following the demonstration that Rpf-
in the sputum of untreated tuberculosis dependent cells were a dominant population
patients. Studies to determine whether there in the sputum samples of tuberculosis
is a relationship between the LB-positive M. patients, the response of this population was
tuberculosis population in sputum and frac- monitored by assessing sputum samples
tional exhaled NO at the time of sputum collected from eight patients 7–11 days after
collection are ongoing. If LB positivity is a the initiation of chemotherapy (Mukamolova
biomarker of a non-replicating population of et al., 2010). The decrease in the viable M.
M. tuberculosis cells in a sputum sample, it tuberculosis counts in these samples was
would be expected that this population assessed by measuring CFUs and the Rpf-
would become enriched in sputum during dependent population by MPN counts in the
the course of treatment. In a preliminary presence of fresh culture supernatant. A good
study, assessment of sputum samples taken correlation was found between a decline in
from seven patients prior to commencement CFU counts and the measured Rpf-dependent
and during chemotherapy did show a population. However, as would be expected
positive relationship (R2 = 0.53) between LB if the Rpf-dependent M. tuberculosis cells
positivity and killing (reduction of colony- were a non-replicating, persisting population
forming unit (CFU) counts), which can be phenotypically tolerant to antibiotics, the
seen in Fig. 14.2. This supports the hypothesis Rpf-dependent population was preserved
that LB positivity may be predictive of relative to CFU counts in all the post-chemo-
treatment response; however, large-scale therapy samples. This trend was observed to
clinical studies are required to determine continue in samples that were collected from
3.5
Killing log10 [CFU T = 0/CFU
2.5
treatment]
1.5
1
R² = 0.5252
0.5
0
35 40 45 50 55 60 65 70 75
LB count (%)
Fig. 14.2. The log10-fold reduction in viable counts in samples of patients’ sputum following 7 or 8 days of
antibiotic treatment show a correlation with the LB count of AFB in sputum samples evaluated before
initiation of treatment. M. tuberculosis viable counts (CFU/ml) were assessed in early-morning samples of
patients’ sputum before the onset of treatment and at either 7 or 8 days of treatment. A separate aliquot of
the pre-treatment samples were treated for LB assessment with auramine/Nile red staining.
four patients further into their treatment. At 14.4 Outstanding Questions

time points when no CFU assessments could
be made, MPN counts were still measurable There remain a number of unanswered
and the addition of culture supernatant questions with regard to the M. tuberculosis
revealed substantial populations of Rpf- phenotypes we have assessed in sputum.
dependent cells. These results suggest that an First, it is unclear whether the LB-positive
Rpf-dependent population present in sputum cells in sputum are indeed growth arrested or
at the time of the initiation of chemotherapy slow growing. We remain unable to demon-
are killed more slowly than a colony-forming strate convincingly LBs in growing batch
population. However, an alternative argument cultures of M. tuberculosis H37Rv in rich
that treatment induces Rpf dependence must media, but have observed LBs in M.
be noted. tuberculosis biomass harvested from chemo-
The addition of RIF directly to decon- stat cultures under conditions of low O2, low
taminated smear-positive samples from three iron or low phosphorus (Sherratt et al.,
patients and incubation for 1 week revealed unpublished results). Supplementing a grow-
that the Rpf-dependent M. tuberculosis popu- ing M. tuberculosis batch culture with oleic
lation was RIF tolerant (Mukamolova et al., acid (conditions which lead to rapid LB
2010). Although in each sample the CFU accumulation in M. smegmatis) led to an
count fell below detectable levels after 7 days increase in TAG content but not LBs (Garton,
exposure to RIF, MPN assessment in the unpublished results). This suggests that TAG
presence of culture supernatant revealed that may accumulate in the cell envelope until
the Rpf-dependent population decreased, at saturation is reached, at which point
most, by 0.5 log. A culture of M. tuberculosis cytoplasmic LBs may develop. It follows that
H37Rv which contained no Rpf-dependent the LB-positive cell numbers may not reflect
cells showed a 2-log decline in the supple- accurately total TAG accumulation in a popu-
mented MPN count over the same duration lation and we are working to develop more
of RIF exposure. accurate reporters of M. tuberculosis lipid con-
The demonstration of an Rpf-dependent tent. Following the identification of the
population which appears to be relatively sputum LB-positive M. tuberculosis popu-
well conserved during treatment in samples lation and the microarray study of the
collected during therapy and the direct transcriptome, we confirmed in samples
measurement of RIF tolerance of Rpf- collected from nine tuberculosis patients the
dependent cells support the hypothesis that apparent upregulation of tgs1 as compared
sputum contains a population of non-repli- with M. tuberculosis H37Rv aerobic growth by
cating cells. This study highlights the quantitative reverse transcription PCR (qRT-
potential importance of this previously PCR) (Garton et al., 2008). A larger-scale study
overlooked population. The addition of Rpf of 30 tuberculosis patients failed to confirm
in assessments of culturability may provide an association between tgs1 expression, as
an early indication of treatment response. measured by qRT-PCR and LB positivity (Lee
Furthermore, early bactericidal activity ana- et al., unpublished results). This was also the
lyses of potential chemotherapeutic agents case for two other genes associated with non-
currently use CFU assessment to measure the replicating persistence, hspX and icl1. We do
efficacy of various agents and, therefore, such not know the age of M. tuberculosis cells in
studies will not reveal the emergence of Rpf- sputum samples, or the stability of tgs1
dependent populations. There is potential for transcripts in vivo. Daniel and colleagues
the measurement of the Rpf-dependent reported maximal tgs1 expression 4 h follow-
population to enhance such studies. A major ing NO stimulation of replicating M.
focus of future research is to determine tuberculosis culture; this time point was also
whether there is any relationship between the coincidental with maximal TAG accumulation
Rpf-dependent population in a patient’s (Daniel et al., 2004). Our own NO studies
sputum with the clinical response during revealed induction of LB positivity at 4 h,
treatment and reported relapse rates. followed by a threefold reduction in LB
positivity 24 h post-stimulation, a point at between LB and Rpf-dependent sputum M.

which growth would be resumed (Garton et tuberculosis populations.
al., 2008). High and enduring tgs1 (and hspX)
expression was reported throughout 30 days
of the Wayne model (Voskuil et al., 2004). 14.5 Conclusions
Later studies of the expression of genes of the
DosR regulon in response to a rapid transfer LB positivity and Rpf dependency are two
into hypoxia revealed that, unlike the quantifiable phenotypes which, in addition to
majority of genes, the expression levels of the transcriptome signatures, provide evi-
tgs1 and hspX (two of the most highly induced dence for the presence of non-replicating M.
genes) remained high over a period of 7 days tuberculosis cells in sputum. Although our
(Rustad et al., 2008). Recent investigation of knowledge of the biological and clinical
the stability of M. tuberculosis transcripts significance of these phenotypes is limited
revealed an overall enhanced stability in and remains to be developed, both LB posi-
hypoxic conditions compared with aerobic tivity and Rpf dependence can be linked to
conditions (T.R. Rustad, San Diego, 2010, phenotypic antibiotic tolerance. The evidence
personal communication). In such conditions, to support the view that one or either of these
cells which have produced LB in response to measurable phenotypes may constitute a
the induction of tgs1 will retain high levels of potential biomarker for treatment response is
tgs1 transcripts. Regrowth of M. bovis BCG gathering. Large-scale clinical studies are
from hypoxia-induced dormancy results in required to test this hypothesis and to deter-
the assimilation of accumulated TAG and LBs mine any clinical correlates of these pheno-
(Low et al., 2009). The M. tuberculosis cells in types. In the laboratory, greater understanding
sputum must, at some point, have undergone of the stimuli leading to the development of
growth within the nutrient-rich environment these sputum phenotypes will allow us to
of the patent cavity; it would be impossible to develop in vitro conditions in which we can
maintain ongoing smear positivity otherwise. study the adaption in M. tuberculosis. In com-
Therefore, unless the LB-positive cells in bination, these studies will provide exciting
sputum are dead or injured, we hypothesize new opportunities to address the problem of
that rather than a slow adaptation to non- persisters in the treatment of tuberculosis.
replicating conditions, LB positivity is a
reflection of a more rapid growth arrest in
response to a change of environment. This References
remains to be determined, along with identifi-
cation of the specific growth-arresting Betts, J.C., Lukey, P.T., Robb, L.C., McAdam, R.A.
signal(s). and Duncan, K. (2002) Evaluation of a nutrient
starvation model of Mycobacterium tuberculosis
What, if any, is the relationship between
persistence by gene and protein expression
the sputum M. tuberculosis LB-positive and profiling. Molecular Microbiology 43, 717–731.
Rpf-dependent populations? Although the Biketov, S., Mukamolova, G.V., Potapov, V.,
sputum LB and Rpf-dependency sputum Gilenkov, E., Vostroknutova, G., Kell, D.B., et al.
studies have been largely independent, it is (2000) Culturability of Mycobacterium tubercu-
clear that the numbers of Rpf-dependent M. losis cells isolated from murine macrophages: a
tuberculosis can potentially be orders of bacterial growth factor promotes recovery.
magnitude higher than LB-positive cells. In FEMS Immunology and Medical Microbiology
the limited number of sputum samples in 29, 233–240.
which LB positivity was correlated with Biketov, S., Potapov, V., Ganina, E., Downing, K.,
Kana, B.D. and Kaprelyants, A. (2007) The role
killing as measured by a drop in the sputum
of resuscitation promoting factors in patho-
CFU count following treatment (Fig. 14.2), genesis and reactivation of Mycobacterium
Rpf-dependent M. tuberculosis cells had been tuberculosis during intra-peritoneal infection in
identified in six out of seven samples. Larger- mice. BMC Infectious Diseases 7, 146.
scale studies are required to determine Cohen-Gonsaud, M., Keep, N.H., Davies, A.P.,
whether there is any significant connection Ward, J., Henderson, B. and Labesse, G. (2004)
Resuscitation-promoting factors possess a Gomez, J.E. and McKinney, J.D. (2004) Myco-
lysozyme-like domain. Trends in Biochemical bacterium tuberculosis persistence, latency,
Sciences 29, 7–10. and drug tolerance. Tuberculosis (Edinburgh)
Cohen-Gonsaud, M., Barthe, P., Bagneris, C., 84, 29–44.
Henderson, B., Ward, J., Roumestand, C., et al. Herbert, D., Paramasivan, C.N., Venkatesan, P.,
(2005) The structure of a resuscitation- Kubendiran, G., Prabhakar, R. and Mitchson,
promoting factor domain from Mycobacterium D.A. (1996) Bactericidal action of ofloxacin,
tuberculosis shows homology to lysozymes. sulbactam-ampicillin, rifampin, and isoniazid on
Nature Structural and Molecular Biology 12, logarithmic- and stationary-phase cultures of
270–273. Mycobacterium tuberculosis. Antimicrobial
Connolly, L.E., Edelstein, P.H. and Ramakrishnan, Agents and Chemotherapy 40, 2296–2299.
L. (2007) Why is long-term therapy required to Hu, Y., Mangan, J.A., Dhillon, J., Sole, K.M.,
cure tuberculosis? PLoS Medicine 4, e120. Mitchison, D.A., Butcher, P.D., et al. (2000)
Cunningham, A.F. and Spreadbury, C.L. (1998) Detection of mRNA transcripts and active tran-
Mycobacterial stationary phase induced by low scription in persistent Mycobacterium tubercu-
oxygen tension: cell wall thickening and losis induced by exposure to rifampin or
localization of the 16-kilodalton alpha-crystallin pyrazinamide. Journal of Bacteriology 182,
homolog. Journal of Bacteriology 180, 801–808. 6358–6365.
Daniel, J., Deb, C., Dubey, V.S., Sirakova, T.D., Hussain, S., Malik, M., Shi, L., Gennaro, M.L. and
Abomoelak, B., Morbidoni, H.R., et al. (2004) Drlica, K. (2009) In vitro model of mycobacterial
Induction of a novel class of diacylglycerol growth arrest using nitric oxide with limited air.
acyltransferases and triacylglycerol accumu- Antimicrobial Agents and Chemotherapy 53,
lation in Mycobacterium tuberculosis as it goes 157–161.
into a dormancy-like state in culture. Journal of Jones, G. and Dyson, P. (2006) Evolution of
Bacteriology 186, 5017–5030. transmembrane protein kinases implicated in
Davies, A.P., Dhillon, A.P., Young, M., Henderson, coordinating remodeling of Gram-positive
B., McHugh, T.D. and Gillespie, S.H. (2008) peptidoglycan: inside versus outside. Journal of
Resuscitation-promoting factors are expressed Bacteriology 188, 7470–7476.
in Mycobacterium tuberculosis-infected human Kana, B.D. and Mizrahi, V. (2010) Resuscitation-
tissue. Tuberculosis (Edinburgh) 88, 462–468. promoting factors as lytic enzymes for bacterial
Deb, C., Lee, C.M., Dubey, V.S., Daniek, J., growth and signaling. FEMS Immunology and
Abomoelak, B., Sirakova, T.D., et al. (2009) A Medical Microbiology 58, 39–50.
novel in vitro multiple-stress dormancy model Kana, B.D., Gordhan, B.G., Downing, K.J., Sung,
for Mycobacterium tuberculosis generates a N., Vostroktunova, G., Machowski, E.E., et al.
lipid-loaded, drug-tolerant, dormant pathogen. (2008) The resuscitation-promoting factors of
PLoS One 4, e6077. Mycobacterium tuberculosis are required for
Dhar, N. and McKinney, J.D. (2007) Microbial virulence and resuscitation from dormancy but
phenotypic heterogeneity and antibiotic are collectively dispensable for growth in vitro.
tolerance. Current Opinion in Microbiology 10, Molecular Microbiology 67, 672–684.
30–38. Keep, N.H., Ward, J.M., Cohen-Gonsaud, M. and
Downing, K.J., Mischenko, V.V., Shleeva, M.O., Henderson, B. (2006) Wake up! Peptidoglycan
Young, D.I., Young, M., Kaprelyants, A.S., et al. lysis and bacterial non-growth states. Trends in
(2005) Mutants of Mycobacterium tuberculosis Microbiology 14, 271–276.
lacking three of the five Rpf-like genes are Levin, B.R. and Rozen, D.E. (2006) Non-inherited
defective for growth in vivo and for resuscitation antibiotic resistance. Nature Reviews.
in vitro. Infection and Immunity 73, 3038–3043. Microbiology 4, 556–562.
Garton, N.J., Christensen, H., Minnikin, D.E., Low, K.L., Rao, P.S., Shui, G., Bendt, A.K., Pethe,
Adegbola, R.A. and Barer, M.R. (2002) K., Dick, T., et al. (2009) Triacylglycerol utilization
Intracellular lipophilic inclusions of mycobacteria is required for regrowth of in vitro hypoxic
in vitro and in sputum. Microbiology 148, 2951– nonreplicating Mycobacterium bovis bacillus
2958. Calmette–Guerin. Journal of Bacteriology 191,
Garton, N.J., Waddell, S.J., Sherratt, A.L., Lee, 5037–5043.
S.M., Smit, R.J., Senner, C., et al. (2008) Low, K.L., Shui, G., Natter, K., Yeo, W.K., Kohlwein,
Cytological and transcript analyses reveal fat S.D., Dick, T., et al. (2010) Lipid droplet-
and lazy persister-like bacilli in tuberculous associated proteins are involved in the bio-
sputum. PLoS Medicine 5, e75. synthesis and hydrolysis of triacylglycerol in
Mycobacterium bovis bacillus Calmette–Guerin. Russell-Goldman, E., Xu, J., Wang, X., Chan, J.
Journal of Biological Chemistry 285, 21662– and Tufariello, J.M. (2008) A Mycobacterium
21670. tuberculosis Rpf double-knockout strain
Mitchison, D.A. (1979) Basic mechanisms of exhibits profound defects in reactivation from
chemotherapy. Chest 76, 771–781. chronic tuberculosis and innate immunity
Mitchison, D.A. (2004) The search for new sterilizing phenotypes. Infection and Immunity 76, 4269–
anti-tuberculosis drugs. Frontiers of Bioscience 4281.
9, 1059–1072. Rustad, T.R., Harrell, M.I., Liao, R. and Sherman,
Mukamolova, G.V., Yanopolskaya, N.D., Kell, D.B. D.R. (2008) The enduring hypoxic response of
and Kaprelyants, A.S. (1998) On resuscitation Mycobacterium tuberculosis. PLoS One 3,
from the dormant state of Micrococcus luteus. e1502.
Antonie Van Leeuwenhoek 73, 237–243. Shah, I.M. and Dworkin, J. (2010) Induction and
Mukamolova, G.V., Turapov, O.A., Kazarian, K., regulation of a secreted peptidoglycan
Telkov, M., Kaprelyants, A.S., Kell, D.B., and hydrolase by a membrane Ser/Thr kinase that
Young, M. (2002a) The Rpf gene of Micrococcus detects muropeptides. Molecular Microbiology
luteus encodes an essential secreted growth 75, 1232–1243.
factor. Molecular Microbiology 46, 611–621. Shah, I.M., Laaberki, M.H., Popham, D.L. and
Mukamolova, G.V., Turapov, O.A., Young, D.I., Dworkin, J. (2008) A eukaryotic-like Ser/Thr
Kaprelyants, A.S., Kell, D.B. and Young, M. kinase signals bacteria to exit dormancy in
(2002b) A family of autocrine growth factors in response to peptidoglycan fragments. Cell 135,
Mycobacterium tuberculosis. Molecular Micro- 486–496.
biology, 46, 623-35. Shi, L., Sohaskey, C.D., Kana, B.D., Dawes, S.,
Mukamolova, G.V., Murzin, A.G., Salina, E.G., North, R.J., Mizrahi, V., et al. (2005) Changes in
Demina, G.R., Kell, D.B., Kaprelyants, A.S., et energy metabolism of Mycobacterium tuber-
al. (2006) Muralytic activity of Micrococcus culosis in mouse lung and under in vitro
luteus Rpf and its relationship to physiological conditions affecting aerobic respiration.
activity in promoting bacterial growth and Proceedings of the National Academy of
resuscitation. Molecular Microbiology 59, Sciences 102, 15629–15634.
84–98. Shleeva, M.O., Bagramyan, K., Telkov, M.V.,
Mukamolova, G.V., Turapov, O., Malkin, J., Mukamolova, G.V., Young, M., Kell, D.B., et al.
Woltmann, G. and Barer, M.R. (2010) (2002) Formation and resuscitation of ‘non-
Resuscitation-promoting factors reveal an culturable’ cells of Rhodococcus rhodochrous
occult population of tubercle bacilli in sputum. and Mycobacterium tuberculosis in prolonged
American Journal of Respiratory and Critical stationary phase. Microbiology 148, 1581–1591.
Care Medicine 181, 174–180. Telkov, M.V., Demina, G.R., Voloshin, S.A., Salina,
Muttucamaru, D.G., Roberts, G., Hinds, J., Stabler, E.G., Dudik, T.V., Stekhanova, T.N., et al. (2006)
R.A. and Parish, T. (2004) Gene expression Proteins of the Rpf (resuscitation promoting
profile of Mycobacterium tuberculosis in a non- factor) family are peptidoglycan hydrolases.
replicating state. Tuberculosis (Edinburgh) 84, Biochemistry (Moscow) 71, 414–422.
239–246. Tufariello, J.M., Mi, K., Xu, J., Manabe, Y.C.,
Ohno, H., Zhu, G., Mohan, V.P., Chu, D., Kohno, S., Kesavan, A.K., Drumm, J., et al. (2006) Deletion
Jacobs, W.R. Jr, et al. (2003) The effects of of the Mycobacterium tuberculosis resuscitation-
reactive nitrogen intermediates on gene promoting factor Rv1009 gene results in
expression in Mycobacterium tuberculosis. delayed reactivation from chronic tuberculosis.
Cellular Microbiology 5, 637–648. Infection and Immunity 74, 2985–2995.
Paramasivan, C.N., Sulochana, S., Kubendiran, G., Voskuil, M.I., Schnappinger, D., Visconti, K.C.,
Venkatesan, P. and Mitchison, D.A. (2005) Harrell, M.I., Dolganov, G.M., Sherman, D.R.,
Bactericidal action of gatifloxacin, rifampin, and et al. (2003) Inhibition of respiration by nitric
isoniazid on logarithmic- and stationary-phase oxide induces a Mycobacterium tuberculosis
cultures of Mycobacterium tuberculosis. Anti- dormancy program. Journal of Experimental
microbial Agents and Chemotherapy 49, 627– Medicine 198, 705–713.
631. Voskuil, M.I., Visconti, K.C. and Schoolnik, G.K.
Rifat, D., Bishai, W.R. and Karkousis, P.C. (2009) (2004) Mycobacterium tuberculosis gene
Phosphate depletion: a novel trigger for expression during adaptation to stationary
Mycobacterium tuberculosis persistence. phase and low-oxygen dormancy. Tuberculosis
Journal of Infectious Diseases 200, 1126–1135. (Edinburgh) 84, 218–227.
Wallis, R.S., Patil, S., Cheon, S.H., Edmonds, K., Wältermann, M. and Steinbüchel, A. (2005) Neutral
Phillips, M., Perkins, M.D., et al. (1999) Drug lipid bodies in prokaryotes: recent insights into
tolerance in Mycobacterium tuberculosis. structure, formation, and relationship to
Antimicrobial Agents and Chemotherapy 43, eukaryotic lipid depots. Journal of Bacteriology
2600–2606. 187, 3607–3619.
Wallis, R.S., Pai, M., Menzies, D., Doherty, T.M., Wayne, L.G. and Hayes, L.G. (1996) An in vitro
Walzl, G., Perkins, M.D., et al. (2010) Biomarkers model for sequential study of shiftdown of
and diagnostics for tuberculosis: progress, Mycobacterium tuberculosis through two stages
needs, and translation into practice. The Lancet of nonreplicating persistence. Infection and
375, 1920–1937. Immunity 64, 2062–2069.
Part IV
Treatment Strategies
15 Clinical Trials in Tuberculosis
Chemotherapy: The Challenges
Stephen H. Gillespie
Sir James Black Professor of Medicine, The Bute Medical School,
St Andrews University, Scotland, UK
15.1 Introduction receive the current best treatment with or

without SM. The initial results showed
The first antibiotics, penicillins and sul- significant benefit for the patients who had
fonamides, were not active against Myco- received SM, but found evidence of toxicity
bacterium tuberculosis and some commentators and development of drug resistance (MRC,
at the time considered the organism might be 1948). At 5 years follow-up, the benefit of the
beyond the reach of chemotherapy (Ryan, added antibiotic had disappeared as the
1992; Dormandy, 1999). Selman Waksman did organisms had developed resistance and
not take that view and rather considered the infection had relapsed. This trial and its
closely related organisms might produce results were seminal in several ways. Not
antibiotics that would inhibit mycobacteria, only was it the first randomized placebo
an idea that might have come from his controlled clinical trial in medical history but
medical student son (Daniel, 2005). His PhD also it was a pioneering study that established
student, Albert Schatz, was given a culture of the way new drugs were developed. It also
Streptomyces griseus that had been isolated demonstrated the importance of long-term
from a chicken and he isolated the compound follow-up in tuberculosis clinical trials
that became known as streptomycin (SM) – a (McDermott, 1960; Fox and Mitchison, 1975;
revolution in anti-tuberculosis chemotherapy Fox et al., 1999).
and one which opened the floodgates for In subsequent years, a series of trials
isoniazid (INH), pyrazinamide (PZA), para- were performed that evaluated each new
amino-salicylic acid and eventually rifampi- anti-tuberculosis drug as it became ready for
cin (RIF) (Ryan, 1992; Dormandy, 1999). introduction in clinical practice. INH’s potent
When SM became available, it was bactericidal activity was realized, and when
rushed into use in the USA, where it had first RIF was introduced and was used in a three-
been produced. In the UK, a small amount drug regimen with SM, treatment duration
was imported and given to the Medical dropped from 24 to 9 months (EABMR Coun-
Research Council (MRC), who formed a trials cil, 1973; Grosset, 1978). In the late 1970s, PZA
committee. There was not enough of this was re-evaluated in anti-tuberculosis regi-
wonder drug to treat all of the patients. The mens. It had been developed at the same time
MRC sponsored a study to evaluate the use of as INH, but it was withdrawn from therapy
SM and allocated patients randomly to due to toxicity. The later trials used a lower

198 S.H. Gillespie
dose, and a series of trials sponsored by the described as ‘early bactericidal studies’ are
MRC and the British Thoracic Society showed performed where the new drug is given as
that treatment was possible in 6 months monotherapy for a short period (Jindani et al.,
(EACA/BMR Councils, 1986). Trials to shorten 1980, 2003; Gillespie et al., 2002). Since the
treatment further resulted in unacceptable drug is novel and the duration of therapy is
rates of relapse, and the duration of treatment short, there is little chance that this will result
using an optimal regimen was fixed at 6 in the emergence of resistance (Gillespie et al.,
months. Other regimens in resource-poor 2002; Gosling, 2003a,b). This phenomenon of
settings using less potent agents needed a the higher activity of INH in the first 2 days of
treatment duration of at least 8 months. The therapy was named ‘early bactericidal
success of anti-tuberculosis chemotherapy activity’, which is somewhat of a misnomer
was deemed complete and many countries (Jindani et al., 1980, 2003) as subsequent
with efficient treatment and control pro- studies with other drugs did not find the
grammes did witness a year-on-year decline same characteristic. Thus, these studies
in the prevalence of tuberculosis. This rosy should, more properly, be described as
picture neglected several important problems: monotherapy studies. They have considerable
in many countries there was no effective benefits in the drug development process, as
control programme and tuberculosis raged they provide the first indication of whether
unchecked. When Human Immunodeficiency the new drug is active in killing tubercle
Virus appeared, it formed a malignant bacilli in the human host. It also provides the
alliance with the tubercle bacillus, reversing opportunity to identify a dose that is likely to
the favourable trend in countries that had be effective (Rustomjee et al., 2008a). Some
seen success. Importantly, for patients authors also suggest that it may be possible to
infected with bacteria that had become evaluate combinations in regimens but, in
resistant or that were not able to tolerate the combinations with RIF and INH, the effect
current regimen, there were few alternative size of the established drugs are usually too
regimens available. large to see the contribution of the novel
The Cape Town Conference, held in agent reliably (Gillespie, 2005). Moreover,
Somerset West, South Africa, in 2000, regimen building is more complex and
addressed the pressing need to develop depends on the pharmacokinetic interaction
shorter treatment regimens and new drugs between all of the drugs in the regimens.
for resistant disease. Global bodies formed Finding a solution to evaluate combinations
the Global Alliance for TB Drug Development rapidly, reliably and cheaply is a major
to address the paucity of alternative anti- challenge. Once this stage is complete, more
tuberculosis drugs (Pablos-Mendez, 2000). conventional Phase II studies can be devised,
The spotlight that this has thrown on tubercu- but tuberculosis poses particular challenges
losis drug development means that there are of its own, as set out in the paragraphs below.
now many new drugs in development,
although only a few have reached the clinical
trials stage (Duncan and Barry, 2004; 15.2.1 Additive or substitutive
Ginsberg, 2010). This immediately poses the
question of how, 60 years after the first golden It is essential that the treatment of tuberculosis
age, will the tuberculosis community manage is with a multiple drug regimen, because M.
the clinical evaluation of new drugs for tuberculosis develops resistance rapidly
tuberculosis? through mutation of chromosomal genes
(Gillespie, 2002). Although this occurs at a
relatively low rate among the majority of the
15.2 Tuberculosis Clinical Trials most effective drugs (e.g. once in a billion cell
divisions for RIF), the large number of bacilli
Tuberculosis clinical trials differ from those in the body would mean that monotherapy
for other indications as, at the first stage after inevitably would result in resistance emerg-
Phase I studies are completed, what are often ing (Shimao, 1986). This is solved by using
Clinical Trials in Tuberculosis Chemotherapy: The Challenges 199
multiple-drug regimens, which means that 2008b; Dorman et al., 2009). This represents a
the risk of a bacterium being spontaneously model for the tuberculosis drug development
resistant to three of the agents in the regimen pathway (Nunn et al., 2008b).
is essentially nil. The added advantage is that
the various drugs target the bacteria in
different ways and have differing pharma- 15.3 Trial Design
cokinetics, reducing the risk that a meta-
bolically inactive bacterium in a hard-to-reach Although we know that the control of
site, such as an empyema, would elude tuberculosis globally will require improved
therapy. treatment that is shorter and easier to tolerate,
How should new drugs be evaluated in when designing tuberculosis clinical trials we
the context of a multiple-drug regimen? are confronted with the fact that, for sus-
Should they be added to the regimen or ceptible disease, currently recommended
should one of the current drugs be sub- regimens have more than 95% efficacy under
stituted? controlled clinical trial conditions (Fox et al.,
Adding a new agent to a three- or four- 1999; Nunn et al., 2008b). It is very unlikely
drug regimen has a number of difficulties. that a better result could be achieved with
The more drugs that are present, the higher one or a combination of new drugs. Thus,
the risk of pharmacokinetic interaction or clinical trials of new drugs must focus on
adverse events. Care must also be taken to developing a regimen that is shorter in
ensure that combining drugs with a given duration. In order for any proposed regimen
mode of action is rational. It has been shown to be licensed, it will have to prove that it is
for other organisms that agents which sup- comparable to the existing regimen in terms
press protein synthesis can inhibit the activity of its efficacy. Although this issue has
of agents that depend on protein synthesis for occasionally proved controversial, future
their mechanism of action. Careful mind trials of treatment in susceptible disease will
experiments and in vitro synergy studies have a non-inferiority design (Nunn et al.,
supplemented by mouse interaction studies 2008a,b). In planning such studies, con-
might be used to select the components of siderable care must be taken to choose and
novel regimens. justify the choice of the margin of non-
There is some experience of additive inferiority, which should be justified both
regimens: a trial of levofloxacin added to statistically and, in terms of the impact,
conventional anti-tuberculosis failed to show clinically and programmatically. The trial
benefit because the standard regimen was methodology must be rigorous, with the
already sufficiently effective (el-Sadr et al., effective recruitment and retention of patients.
1998). If there are substantial losses to follow-up or
An alternative approach is to substitute unsatisfactory laboratory procedures, non-
the new agent with established members of inferiority could be concluded falsely. The
the regimen. This is a more attractive prop- sample size should be calculated carefully to
osition, as several of the components of ensure that the study has sufficient power to
standard anti-tuberculosis chemotherapy are answer the clinical question posed (Nunn et
toxic. In a series of in vitro and mouse al., 2008b).
experiments, it was shown that substituting
the new fluoroquinolone, moxifloxacin, for
either ethambutol or INH in the regimen 15.3.1 Multiple drug-resistant
resulted in a stable cure more rapidly than tuberculosis
with standard therapy (Nuermberger et al.,
2004a,b). This prompted the series of clinical Multiple drug-resistant (MDR-TB) and exten-
trials in the evaluation of moxifloxacin and sively drug-resistant tuberculosis (XDR-TB)
gatifloxacin in which these various sub- pose unique challenges for clinical trials
stitutions were trialled in Phase IIb and Phase (Sacks and Behrman, 2008; WHO, 2008).
III trials (Burman et al., 2006; Rustomjee et al., Although a growing problem, patients are
200 S.H. Gillespie
scattered, with few clinical centres having a These basic constraints on the relationship
large number of cases or, alternatively, the are often forgotten. Clearly, a patient who is
number of sites with a large number of MDR-/ rendered culture negative by an effective
XDR-TB cases is small and the multiplicity of regimen and then goes on to a less satisfactory
sites enhances the complexity of running regimen for the remainder of the treatment
clinical trials in order to ensure uniform period or who stops therapy for whatever
clinical and laboratory practice. Moreover, reasons will have a poorer outcome than an
although there are clear definitions of MDR- individual on an effective consolidation
and XDR-TB, the susceptibility pattern of regimen taken completely. The importance of
individual strains differs significantly, with this relationship was demonstrated in ‘Study
the effect that a ‘standard’ regimen cannot be A’, organized by the International Union
established easily for comparison. Alternative against Tuberculosis and Lung Disease, in
approaches include the use of the best avail- which an inadequate consolidation regimen
able regimen, as defined by WHO guidelines, resulted in an excess of relapses in the
compared to this regimen plus the new drug 8-month regimen against the standard
(WHO, 2008). This approach is expected to 6-month regimen, even though there was no
work well, as the activity of many second- or significant difference in culture negativity
third-line regimens is modest and the addition data at 2 months (Jindani et al., 2004).
of a potent novel agent may have a significant In tuberculosis clinical trials, the con-
impact (Sacks and Behrman, 2008). Trials that ventional end point for a Phase III study is
have taken this approach have been able to combined bacteriological failure at the end of
yield a positive outcome with modest treatment and relapse in the 2 years following.
numbers of patients, for example TMC207, In practice, most of the relapses occur within
which was shown to have only modest the first year of treatment completion. In a
activity in the first monotherapy trial, made a review of 15 treatment trials, 68% of relapses
very significant impact when tested in MDR. occurred within 6 months and 91% within 1
The addition of TMC207 to standard therapy year (Nunn et al., 2010). Also, the availability
for MDR-TB reduced the time to conversion to of highly discriminatory typing methods
a negative sputum culture, as compared with have shown that many of the cases that were
placebo (hazard ratio, 11.8; 95% confidence thought to be relapses were, in fact, new
interval, 2.3–61.3; P = 0.003 by Cox regression infections with different strains (Verver et al.,
analysis) and increased the proportion of 2005; van Helden et al., 2008). There is
patients with conversion of sputum culture increasing evidence that the risk of reinfection
(48% versus 9%) (Diacon et al., 2009, 2010). is related to the incidence of disease in the
community (van Helden et al., 2008). In other
words, in areas where rates of transmission
15.3.2 End points are high, the proportion of reinfections to
relapses increases. Thus, a relapse can be
For the early stage/Phase II clinical trial, the identified only if the strain has been subjected
usual end point is defined as the proportion to a modern typing technique such as IS6110
of patients who are deemed culture negative RFLP (restrictive fragment length poly-
after 2 months of therapy (Fox et al., 1999). morphism) analysis, MIRU (mycobacterial
Although this has become a standard intergenic repeat unit) typing or whole
approach, the reasoning behind may not be genome sequencing that demonstrates that
as sound. A series of studies have indicated the initial and relapse strain are indistinguish-
that there is an association between the able (Nunn et al., 2008b).
2-month culture negativity rate and relapse
(Aber and Nunn, 1978). In this statement,
there are a number of assumptions: that the 15.4 Subjects and Locations
laboratory methods are the same as in the
studies that identified the relationship and For a clinical trial to be truly useful, it should
that the post 2-month therapy is appropriate. be representative of the human population.
For reasons of practicality and applicability, it will be a challenge of languages, so it is

is more efficient to perform clinical trials in essential that the informed consent docu-
areas where tuberculosis is a common diag- ments are translated into the relevant local
nosis. This is often a country that is resource languages and then translated back into the
poor, both financially and in terms of original language, to ensure that the meaning
community medical facilities. This poses its has been rendered accurately. Participation in
own challenges for the management of Phase the trial should not be an added financial
III clinical trials. burden for the patient, but investigators
To ensure patient safety and to detect should ensure that payments are appropriate
mycobacteria safely to perform susceptibility and do not, in themselves, provide an
tests and provide a full range of safety, blood inducement to participate (Perrin and
test and electrocardiographic and radiology Gillespie, 2006).
facilities are necessary. This limits the sites in
which trials can be performed and recent
studies have shown that the number of 15.5 Laboratory Methodology
suitable sites is relatively limited (van
Niekerk and Ginsberg, 2009). Beyond the Since the series of trials performed by the
basic physical infrastructure, the differences MRC and the US Public Health Service, there
between the approach for clinical practice has been a revolution in mycobacterial labora-
and the constraints essential for a clinical trial tory methods. The trials were conducted when
can be a challenge for many sites in resource- diagnosis was defined by examination of
poor settings where the large patient burden Zhiel–Neelsen stained smears and cultivation
means that diagnosis and treatment are of the organism on Löwenstein–Jensen (LJ)
stripped down to their bare essentials. slopes (Fox et al., 1999). Modern myco-
Accreditation and quality assurance of clinical bacteriology includes liquid culture and
and laboratory facilities can be a major automates susceptibility testing, molecular
challenge in any setting. For a location to be susceptibility testing and molecular typing
suitable for a Phase III clinical trial, the site (Wallis et al., 2010).
must be highly committed to developing and Modern liquid culture is monitored con-
adapting their processes to the constraints of tinuously, signalling positive when changes
the clinical trial paradigm. Central to the in CO2 concentration generated by meta-
success of this programme is the effective bolizing bacteria pass a critical threshold,
training of staff, not only in the requirements sensed by the machine. The machine is moni-
of Good Clinical Practice and Good Labora- tored constantly: this shortens the time to
tory Clinical Practice but also to understand diagnosis, as solid cultures are only moni-
the nature of the disease and the research tored intermittently. A positive result can be
being undertaken (Perrin and Gillespie, 2006). expected approximately 2 weeks quicker
Short cuts taken in the training aspects of trial than could be achieved using LJ slopes
development will result in unreliable data. (Rusch-Gerdes et al., 1999; Hillemann et al.,
It might be argued that tuberculosis is a 2006; Chihota et al., 2010). In addition, liquid
disease of poverty. It is important, therefore, culture is significantly more sensitive than
that care is taken with the recruitment of solid culture: the effect of this is that a positive
patients who are clinically and financially culture on an LJ slope and in a liquid culture
vulnerable. It is essential that patients flask relate to a different number of organisms
entering a pivotal Phase III study should give in the specimen (Nunn et al., 2008). In effect, a
fully informed consent. This means that the patient will become culture negative more
risks of participation should be explained quickly if the measure is LJ culture than if it
clearly and simply, using non-specialist were liquid culture. Thus, it is essential that
language. Ideally, consent should be taken by laboratories use a single method for determin-
someone who is not directly involved in the ing culture positivity in the protocol, but
study and who is supported by trained since all of the data that we have to date are
counsellors. In any international study, there based on solid culture, it would be prudent to
202 S.H. Gillespie
have both methods running in parallel. This ing concern that they may not be culture
is the approach that is being used in the negative, a question that is left tantalizingly
current REMoxTB clinical trial (Bateson et al., out of sight as their sputum samples are often
2011). contaminated with other organisms. The
It is also important to recognize that question is often raised whether sputum
liquid culture is more susceptible to growth should be induced to identify cryptic positive
of non-tuberculosis organisms, resulting in patients. The evidence that is available does
the loss of that data point (Chihota et al., suggest that those patients who have reverted
2010). This should be tackled by ensuring that to sputum negativity rarely yield a positive
the decontamination process is controlled culture following sputum induction, which
carefully and adjusted to ensure standard suggests that treated patients who have
performance. It also means that additional difficulty producing sputum should be
samples may be necessary to ensure that data classed a treatment successes (Perrin et al.,
are available for the crucial post-treatment 2009).
end points.
Liquid-based susceptibility testing means
that patients that have resistances can be 15.6 Conclusion
removed from a trial of susceptible disease
rapidly and classified as a late protocol There is encouraging news that there are a
exclusion. The availability of rapid molecular large number of new anti-tuberculosis drugs
hybridization-based methods means that it is in development. This brings the challenge
now possible to identify patients with that we need to expand the number of sites
MDR-TB within a few hours and exclude that are capable of performing high-quality
them from the trial. This is especially import- studies internationally. Further work is
ant when performing trials of susceptible required to understand the impact of new
disease in areas where the rate of drug laboratory technology on clinical trials
resistance is high (Boehme et al., 2010; Kiet et design. Importantly, we need to harmonize
al., 2010). the methodologies that are used in trials to
Smear diagnosis remains a cornerstone ensure that results are reliable and that they
of the recruitment process, as demonstrating can be translated in other environments.
that a patient is smear positive increases the
likelihood, in a high-burden country, that the
patient has tuberculosis (Bonnet et al., 2010).
Modern molecular methods based on a References
number of different techniques allow for
smear-positive samples to be screened to ex- Aber, V.R. and Nunn, A.J. (1978) Short term
chemotherapy of tuberculosis. Factors affecting
clude non-tuberculosis mycobacteria (NTM),
relapse following short term chemotherapy.
which are being recognized increasingly in Bulletin of the International Union against
high-burden countries. NTM can be identified Tuberculosis 53(4), 276–280.
more frequently when liquid culture is used. Bateson, A.L., Batt, S., Betteridge, M., Bongard, E.,
This is important if patients are found to Ciesielczuk, H., Gillespie, S.H., et al. (2011)
revert from smear negativity to positivity REMoxTB Laboratory Manual (http://www.ucl.
later in the trial or their liquid culture is ac.uk/infection-immunity/research/res_ccm/
positive with an acid-fast bacterium. In these ccm_files/CCM_REMox, accesssed 4 October
situations, it is important to perform a test to 2012).
ensure that a positive result comes from M. Boehme, C.C., Nabeta, P., Hillemann, D., Nicol,
M.P., Shenai, S., Krapp, F., et al. (2010) Rapid
tuberculosis and not from an NTM (Boehme et
molecular detection of tuberculosis and rifampin
al., 2010); Bateson et al., 2011). resistance. New England Journal of Medicine
Questions are raised frequently about 363(11), 1005–1015.
patients who stop producing sputum towards Bonnet, M., Tajahmady, A., Hepple, P., Ramsay, A.,
the end of or after treatment. There is a linger- Githui, W., Gagdnidze, L., et al. (2010) Added
value of bleach sedimentation microscopy for regimen and duration of short-course treatment
diagnosis of tuberculosis: a cost-effectiveness for human immunodeficiency virus-related
study. International Journal for Tuberculosis and pulmonary tuberculosis. Terry Beirn Community
Lung Disease 14(5), 571–577. Programs for Clinical Research on AIDS
Burman, W.J., Goldberg, S., Johnson, J.L., (CPCRA) and the AIDS Clinical Trials Group
Muzanye, G., Engle, M., Mosher, A.W., et al. (ACTG). Clinical Infectious Diseases 26(5),
(2006) Moxifloxacin versus ethambutol in the 1148–1158.
first 2 months of treatment for pulmonary Fox, W. and Mitchison, D.A. (1975) Short-course
tuberculosis. American Journal of Respiratory chemotherapy for pulmonary tuberculosis.
and Critical Care Medicine 174(3), 331–338. American Review of Respiratory Disease
Chihota, V.N., Grant, A.D., Fielding, K., Ndibongo, 111(3), 325–353.
B., van Zyl, A., Muirhead, D., et al. (2010) Liquid Fox, W., Ellard, G.A. and Mitchison, D.A. (1999)
vs. solid culture for tuberculosis: performance Studies on the treatment of tuberculosis
and cost in a resource-constrained setting. undertaken by the British Medical Research
International Journal for Tuberculosis and Lung Council tuberculosis units, 1946–1986, with
Disease 14(8), 1024–1031. relevant subsequent publications. International
Daniel, T.M. (2005) Selman Abraham Waksman Journal for Tuberculosis and Lung Disease
and the discovery of streptomycin. International 3(10 Suppl 2), S231–279.
Journal for Tuberculosis and Lung Disease 9(2), Gillespie, S.H. (2002) Evolution of drug resistance
120–122. in Mycobacterium tuberculosis: clinical and
Diacon, A.H., Pym, A., Grobusch, M., Patientia, R., molecular perspective. Antimicrobial Agents
Rustomjee, R., Page-Shipp, L., et al. (2009) The and Chemotherapy 46(2), 267–274.
diarylquinoline TMC207 for multidrug-resistant Gillespie, S.H. (2005) Early bactericidal activity of
tuberculosis. New England Journal of Medicine a moxifloxacin and isoniazid combination in
360(23), 2397–2405. smear-positive pulmonary tuberculosis. Journal
Diacon, A.H., Dawson, R., Hanekom, M., Narunsky, of Antimicrobial Chemotherapy 56(6), 1169–
K., Maritz, S.J., Venter, A., et al. (2010) Early 1171.
bactericidal activity and pharmacokinetics of Gillespie, S.H., Gosling, R.D. and Charalambous,
PA-824 in smear-positive tuberculosis patients. B.M. (2002) A reiterative method for calculating
Antimicrobial Agents and Chemotherapy 54(8), the early bactericidal activity of antituberculosis
3402–3407. drugs. American Journal of Respiratory and
Dorman, S.E., Johnson, J.L., Goldberg, S., Critical Care Medicine 166(1), 31–35.
Muzanye, G., Padayatchi, N., Bozeman, L., et Ginsberg, A.M. (2010) Tuberculosis drug
al., the Tuberculosis Trials Consortium (2009) development: progress, challenges, and the
Substitution of moxifloxacin for isoniazid during road ahead. Tuberculosis (Edinburgh) 90(3),
intensive phase treatment of pulmonary 162–167.
tuberculosis. American Journal of Respiratory Gosling, R.D. (2003a) A multicentre comparison of
and Critical Care Medicine 180(3), 273–280. a novel surrogate marker for determining the
Dormandy, T. (1999) The White Death: A History of specific potency of anti-tuberculosis drugs.
Tuberculosis. Hambleton Press, London. Journal of Antimicrobial Chemotherapy 52(3),
Duncan, K. and Barry, C.E. 3rd (2004) Prospects 473–476.
for new antitubercular drugs. Current Opinion in Gosling, R.D. (2003b) The bactericidal activity of
Microbiology 7(5), 460–465. moxifloxacin in patients with pulmonary
EABR Council (1973) Controlled clinical trial of four tuberculosis. American Journal of Respiratory
short-course (6-month) regimens of chemo- and Critical Care Medicine 168(11), 1342–
therapy for treatment of pulmonary tuberculosis. 1345.
Second report. The Lancet 1(7816), 1331–1338. Grosset, J. (1978) The sterilizing value of rifampicin
EACA/BMR Councils (1986) Controlled clinical trial and pyrazinamide in experimental short-course
of 4 short-course regimens of chemotherapy chemotherapy. Bulletin of the International
(three 6-month and one 8-month) for pulmonary Union Against Tuberculosis 53(1), 5–12.
tuberculosis: final report. East and Central Hillemann, D., Richter, E. and Rüsch-Gerdes, S.
African/British Medical Research Council Fifth (2006) Use of the BACTEC Mycobacteria
Collaborative Study. Tubercle 67(1), 5–15. Growth Indicator Tube 960 automated system
el-Sadr, W.M., Perlman, D.C., Matts, J.P., Nelson, for recovery of mycobacteria from 9,558
E.T., Cohn, D.L., Salomon, N., et al. (1998) extrapulmonary specimens, including urine
Evaluation of an intensive intermittent-induction samples. Journal of Clinical Microbiology
204 S.H. Gillespie
44(11), 4014–4017. Timing of relapse in short-course chemotherapy

Jindani, A., Aber, V.R., Edwards, E.A. and trials for tuberculosis. International Journal for
Mitchison, D.A. (1980) The early bactericidal Tuberculosis and Lung Disease 14(2), 241–
activity of drugs in patients with pulmonary 242.
tuberculosis. American Review of Respiratory Pablos-Mendez, A. (2000) Working alliance for TB
Disease 121(6), 939–949. drug development, Cape Town, South Africa,
Jindani, A., Dore, C.J. and Mitchison, D.A. (2003) February 8th, 2000. International Journal for
Bactericidal and sterilizing activities of Tuberculosis and Lung Disease 4(6), 489–490.
antituberculosis drugs during the first 14 days. Perrin, F.M. and Gillespie, S.H. (2006) A new world
American Journal of Respiratory and Critical approach to an old disease. Good Clinical
Care Medicine 167(10), 1348–1354. Practice 8, 18–21.
Jindani, A., Nunn, A.J. and Enarson, D. (2004) Two Perrin, F.M., Breen, R.A., McHugh, T.D., Gillespie,
8-month regimens of chemotherapy for treat- S.H. and Lipman, M.C.I. (2009) Are patients on
ment of newly diagnosed pulmonary tubercu- treatment for pulmonary TB who stop
losis: international multicentre randomised trial. expectorating sputum genuinely culture
The Lancet 364(9441), 1244–1251. negative? Thorax 64(11), 1009–1010.
Kiet, V.S., Lan, N.T.N., An, D.D., Dung, N.H., Hoa, Rusch-Gerdes, S., Domehl, C., Nardi, G.,
D.V., van Vinh Chau, N., et al. (2010) Evaluation Gismondo, M.R., Welscher, H.M. and Pfyffer,
of the MTBDRsl test for detection of second- G.E. (1999) Multicenter evaluation of the
line-drug resistance in Mycobacterium tubercu- mycobacteria growth indicator tube for testing
losis. Journal of Clinical Microbiology 48(8), susceptibility of Mycobacterium tuberculosis to
2934–2939. first-line drugs. Journal of Clinical Microbiology
McDermott, W. (1960) Antimicrobial therapy of 37(1), 45–48.
pulmonary tuberculosis. Bulletin of the World Rustomjee, R., Diacon, A.H., Allen, J., Venter, A.,
Health Organization 23(4–5), 427–461. Reddy, C., Patientia, R.F., et al. (2008a) Early
MRC (1948) Streptomycin in the treatment of bactericidal activity and pharmacokinetics of
pulmonary tuberculosis. British Medical Journal the diarylquinoline TMC207 in treatment of
ii, 769–783. pulmonary tuberculosis. Antimicrobial Agents
Niekerk, C. van and Ginsberg, A. (2009) Assess- and Chemotherapy 52(8), 2831–2835.
ment of global capacity to conduct tuberculosis Rustomjee, R., Lienhardt, C., Kanyok, T., Davies,
drug development trials: do we have what it G.R., Levin, J., Mthiyane, T., et al. (2008b) A
takes? International Journal for Tuberculosis Phase II study of the sterilising activities of
and Lung Disease 13(11), 1367–1372. ofloxacin, gatifloxacin and moxifloxacin in
Nuermberger, E.L., Yoshimatsu, T., Tyagi, S., pulmonary tuberculosis. International Journal
O’Brien, R.J., Vernon, A.N., Chaisson, R.E., et for Tuberculosis and Lung Disease 12(2), 128–
al. (2004a) Moxifloxacin-containing regimen 138.
greatly reduces time to culture conversion in Ryan, F. (1992) The Forgotten Plague: How the
murine tuberculosis. American Journal of Battle Against Tuberculosis Was Won and Lost.
Respiratory and Critical Care Medicine 169(3), Little Brown, Boston, Massachusetts.
421–426. Sacks, L.V. and Behrman, R.E. (2008) Developing
Nuermberger, E.L., Yoshimatsu, T., Tyagi, S., new drugs for the treatment of drug-resistant
Williams, K., Rosenthal, I., O’Brien, R.J., et al. tuberculosis: a regulatory perspective.
(2004b) Moxifloxacin-containing regimens of Tuberculosis (Edinburgh) 88(Suppl 1), S93–
reduced duration produce a stable cure in 100.
murine tuberculosis. American Journal of Shimao, T. (1986) Review of tuberculosis control
Respiratory and Critical Care Medicine 170(10), programmes in the Far East Region of the
1131–1134. International Union Against Tuberculosis.
Nunn, A.J., Meredith, S.K., Spigelman, M.K., Bulletin of the International Union Against
Ginsberg, A.M. and Gillespie, S.H. (2008a) The Tuberculosis 61(1–2), 7–27.
ethics of non-inferiority trials. The Lancet van Helden, P.D., Warren, R.M. and Uys, P. (2008)
371(9616), 895; author reply 896–897. Predicting reinfection in tuberculosis. Journal of
Nunn, A.J., Phillips, P.P. and Gillespie, S.H. (2008b) Infectious Disease 197(1), 172–173; author
Design issues in pivotal drug trials for drug reply 173–174.
sensitive tuberculosis (TB). Tuberculosis Verver, S., Warren, R.M., Beyers, N., Richardson,
(Edinburgh) 88(Suppl 1), S85–92. M., van der Spuy, G.D., Borgdorff, M.W., et al.
Nunn, A.J., Phillips, P.P. and Mitchison, D.A. (2010) (2005) Rate of reinfection tuberculosis after
successful treatment is higher than rate of new and diagnostics for tuberculosis: progress,
tuberculosis. American Journal of Respiratory needs, and translation into practice. The Lancet
and Critical Care Medicine 171(12), 1430– 375(9729), 1920–1937.
1435. WHO (2008) Guidelines for the Programmatic
Wallis, R.S., Pai, M., Menzies, D., Doherty, T.M., Management of MDRTB Emergency Update
Walzl, G., Perkins, M.D., et al. (2010) Biomarkers 2008. WHO, Geneva, Switzerland.
16 The Identification of 2-Aminothiazole-
4-carboxylates (ATCs) as a New Class of
Tuberculosis Agent:
A Lesson in ‘HIT’ Identification
Geoff Coxon
Strathclyde Institute of Pharmacy and Biomedical Sciences, University of
Strathclyde, Glasgow, UK
16.1 Introduction: Strategies to Find directly and use them as templates to design
New Tuberculosis Drug Scaffolds new inhibitors or ligands. The expectation
using this approach is that the discovered
So, how and where does one begin when compound will confer a similar effect to the
looking for a new class of anti-Mycobacterium organism as observed during the genetic
tuberculosis drug? This is a question asked by validation experiments which were used to
many a scientist looking for new therapeutic identify the target. Known as the target or
interventions to eradicate diseases of all types genetic-based approach, this technique, as
and one to which there is no single, or simple, yet, has not afforded the plethora of new
answer. Finding the answer can be achieved chemical entities that were expected from its
using a number of methods and technologies exploitation (Projan, 2003; Payne et al., 2007;
and can be split broadly into two categories Barry, 2009). Nevertheless, this approach still
(Balganesh et al., 2004, 2008). receives much attention and has received
The first method or classical approach to some credit in the antibiotic drug discovery
finding compounds with anti-tubercular activ- field in the pharmaceutical industry (Lerner
ity involves taking a compound or mixture of et al., 2007).
compounds and assessing their ability to The result of these assessments usually
inhibit the growth of the pathogen (in this yields a quantitative indication of how the
case, mycobacteria) at given, or a series of, three-dimensional structural features of the
concentrations. This is often referred to as the compound(s) affect its activity. In the case of
phenotypic approach, as the inhibition of the phenotypic approach, this is normally
growth is the phenotypic result of the action expressed as the minimum concentration
of the compound against one or more gene required to inhibit the growth of mycobacteria
products in the organism. (MIC) or, in the target-based approach, the
A more modern approach, and one which concentration required to reduce the turnover
has been used most over the last decade since of an enzymatic reaction by 50% (IC50).
the complete sequencing of the M. tuberculosis It is the role of the medicinal chemist to
genome, is to target the gene products interpret these data and develop structure–

The Identification of 2-Aminothiazole-4-carboxylates (ATCs) 207
activity relationships (SARs) from which an exist in several forms of differing chemical
iterative modification process may increase functionality. Indeed, the first-line anti-
the potency of inhibition of the target enzyme tubercular drug, isoniazid (INH), works by
or mycobacterial growth. Of course, this is inhibiting their biosynthesis. The complete
only the start of the process and the medicinal sequencing of the M. tuberculosis genome
chemist must ensure that selectivity over the (Cole et al., 1998) has revealed significant
host cells and good physico-chemical prop- biochemical and genetic insight into mycolic
erties are not sacrificed in the pursuit of acid biosynthesis that will aid the search for
potency in order to retain good drug-like new druggable targets. These unique lipids
characteristics. are biosynthesized by both fatty acid synthase
enzyme systems I and II (FAS-I and FAS-II) to
produce C56-64 meromycolic acids and the C26
16.2 Validated Drug Targets for -branch (Portevin et al., 2004; Takayama et
Mycobacterial Killing al., 2005) after a series of biotransformations
(Yuan et al., 1995; Takayama et al., 2005).
Our research began looking for validated The logical step at this stage was to search
drug targets and possible mechanisms of the literature for known inhibitors of this
action that would lead to mycobacterial pathway to exploit this biochemical know-
killing action. It is known that the cell wall of ledge. In looking for compounds that will
M. tuberculosis is rich with many unique key inhibit an enzyme as part of a series of
structural components that are necessary for biochemical steps, it is best to target validated
the mycobacteria to survive and grow within enzymes early in the sequence. As part of our
the human host, and has long been a target investigations, we were fortunate that the
for anti-M. tuberculosis drug development naturally occurring antibiotic, thiolactomycin
(Fig. 16.1). Essential to the cell wall are the 1 (TLM; Fig. 16.2), had already been identified
mycolic acids, which are high molecular as such an inhibitor. This compound, a
weight 2-alkyl, 3-hydroxy fatty acids that secondary metabolite produced by the
hydroxyacyl AcpM
Fig. 16.1. Mycolic acid biosynthesis as a target for anti-M. tuberculosis drug design.
208 G. Coxon
the thiazole ring scaffold which led to the

S O identification of methyl 2-amino-5-benzyl-
thiazole-4-carboxylate (Fig. 16.3) and methyl
2-(2-bromoacetamido)-5-(3-chlorophenyl)
HO
thiazole-4-carboxylate (Fig. 16.4). The com-
Fig. 16.2. The structure of thiolactomycin (TLM). pound in Fig. 16.3 was discovered to be a
potent inhibitor of M. tuberculosis H37Rv, with
an MIC of 0.06 μg/ml (0.24 μm), which was
soil-dwelling Nocardia sp. (Slayden et al., 1996) more effective than both TLM and the first-
acts primarily by inhibiting the line drug INH (MICs of 13 μg/ml (62.5 μm)
FAS-II β-ketoacyl-ACP synthase-condensing and 0.25 μg/ml (1.8 μm), respectively) (Kamal
enzymes, halting mycolic acid biosynthesis et al., 2005). The compound in Fig. 16.4 was
and subsequently leading to M. tuberculosis found to inhibit mtFabH with an IC50 of 0.95
cell death (Choi et al., 2000; Kremer et al., 2000; μg/ml (2.43 μm), which compares well with
Schaeffer et al., 2001; Kremer et al., 2002). TLM and its most potent racemic analogue in
Importantly, TLM is also orally available Fig. 16.5 (IC50 values of 16 μg/ml (75 μm) and
and non-toxic in the mouse model, which 1.1 μg/ml (3.0 μm), respectively) (Bhowruth
makes it an attractive compound for develop- et al., 2007).
ment. However, one of the disadvantages of
this compound in this regard is that the
S NH2
chemical scaffold of TLM possesses a chiral
centre at the 5-position. This structural
O N
feature makes the synthesis of a series of TLM
analogues lengthy and costly, and complicates O
the optimization process. When one considers
the target product profile of a drug to treat Fig. 16.3. The structure of methyl 2-amino-5-
tuberculosis, a key feature is affordability, as benzylthiazole-4-carboxylate.
no commercial market exists to sustain the
development and subsequent distribution of
Cl
drugs at the clinic. With the current cost of
the drugs required to treat the disease being
very low, and in some cases free, a simple H
molecular scaffold for development is vital. S N
Br
In recent years, this issue of synthetic
tractability has focused researchers’ efforts O N O
towards the synthesis of either racemic
analogues or derivatives of TLM that contain
O
simple modifications and has yielded limited Fig. 16.4. The structure of methyl
improvements in activity against M. tubercu- 2-(2-bromoacetamido)-5-(3-chlorophenyl)thiazole-
losis and modest activity against mtFabH 4-carboxylate.
(Douglas et al., 2002; Senior et al., 2003, 2004;
Kamal et al., 2005; Kim et al., 2006; Bhowruth
et al., 2007). Instead, we focused on identifying S O
alternative, easily accessible 5-membered ring
isosteres potentially to generate large com-
HO
pound libraries targeted against the con- O
densing enzyme mtFabH and M. tuberculosis,
and our attention, after reviewing the litera-
ture, focused on the thiazole ring as a
HO
template for exploitation. Fig. 16.5. The structure of 4 ((3-hydroxy-2,4-
The following section of this chapter will dimethyl-5-oxo-2,5-dihydrothiophen-2-yl)methyl)-
detail the rationale, design and synthesis of [1,1-biphenyl]-4-carboxylic acid.
16.2.1 The target-based approach using 4-hydroxyl group H-bonds to the carbonyl
mtFabH oxygen of Val270 and the amide NH of
Gly305 through a lattice of three water
Ligand design molecules; and the sulfur is adjacent to the
active site Cys residue, although without any
Our strategy was based initially on the active obvious interaction. The strong H-bonding
site geometry and mechanism of action of between the active site His residues of ecFabB
mtFabH, a homodimer (PDB: 1M1M) that and the TLM carbonyl group is believed to be
converts C12-20 acyl-CoA substrates to the crucial for effective inhibitory activity against
corresponding β-ketoacyl-AcpM product this enzyme. Based on this analysis, we
after reaction with mal-AcpM in a two-step postulated that the closely related active site
process (Brown et al., 2005) (Figs 16.6a and b). of mtFabH could be expected to form an
At the molecular level, the acyl-CoA equivalent H-bonding network with TLM,
substrate enters an L-shaped binding pocket and any new isosteric scaffold would need to
consisting of a lateral and longitudinal
maintain many of these important inter-
channel, with the active site catalytic triad of
actions. The equivalent condensing enzyme
Cys112-His244-Asn274 located at the
from E. coli (ecFabH) is also closely related to
junction. Transacylation of the Cys122 residue
mtFabH, and has been co-crystallized with
occurs when the adjacent His244 deprotonates
the very potent inhibitor, 2-hydroxy-6-(3-
the thiol group (directly or via a molecule of
phenoxy-4-phenyl-benzamido) benzoic acid
water, as postulated by Brown et al. (2005)) to
(Ashek and Cho, 2006). This complex reveals
generate a thiolate nucleophile that attacks
an important role for a carboxylic acid moiety
the carbonyl group of the acyl chain occupy-
in the ligand, as it forms specific interactions
ing the longitudinal channel and releases
in the active site with the His250 residue
CoA-SH from the lateral channel through
which the substrate entered. The second from ecFabH (Ashek and Cho, 2006). We
substrate, mal-AcpM, then enters the lateral considered inclusion of this moiety to be
channel and is decarboxylated by the catalytic important for our inhibitors and proposed
residues, His244 and Asn274, and condensed the achiral 2-aminothiazole-5-carboxylate
with the thioester formed at Cys112 to scaffold as an alternative to the TLM
generate the β-ketoacyl-AcpM, which also substructure to combine pharmacodynamic
dissociates via the lateral channel. Recently, potency with essential pharmacokinetic
however, Sachdeva et al. have postulated that considerations such as solubility. Finally,
the overall reaction may occur simultaneously synthetic tractability and subsequent diverse
in both active sites of the dimer (Sachdeva library generation were considered possible
and Reynolds, 2008; Sachdeva et al., 2008a,b). by exploiting this scaffold.
As no inhibitor-mtFabH co-crystal struc- Using the molecular modelling software
tures had been solved, we investigated the package, GOLD, docking studies were per-
binding pattern of TLM with the closely formed to investigate the poses for our
related analogue ecFabB from Escherichia coli scaffold in the mtFabH active site (Jones et al.,
(Price et al., 2001). TLM inhibits ecFabB 1995; Al-Balas et al., 2009). We observed that
reversibly by forming a number of non- the carboxyl group of the thiazole ring
covalent interactions: the methyl group at formed H-bonds with the NH of Cys112,
carbon 3 of TLM is positioned in a hydro- while the NH2 was proximal to and H-bonds
phobic pocket defined by residues Phe229 with the imidazole ring of His244. This
and Phe392; the 5-isoprenoid moiety is allowed us to generate a hypothetical tem-
wedged between two peptide bonds – from plate for the development of inhibitors of
above by residues Val271 and Phe272 and mtFabH (Fig. 16.7a). Further docking studies
from below by Gly391 and Phe392, which are were carried out with phenyl, m-chlorophenyl
important for specificity (Brown et al., 2005); (Fig. 16.7b) and benzyl substituents in the
the carbonyl oxygen forms two H-bonds with 5-position, as both the 4-methyl ester and free
the two histidines in the active site; the acid forms of the thiazoles. In these cases, the
(a) (b)
210
N274
O
G306 N274 O
N G306 O NH
H R O NH N
H R
H
H H
O H
O
R N O O
S CH3 - CO2 R N
S CH3
ACP H2C ACP
O O
HN
C112 O O C112 HN O
N H NH
H2C ACP
H244 H244
O O
G. Coxon
G306 NH G306 NH N274
2 N274
H O NH
O NH2
H
H2N O H
O
SH R N
S
O O
CH3
N O H
C112
N ACP
NH C112
O ACP
NH
H244
H244
Fig. 16.6. (a) Transacylation mechanism in mtFabH, as proposed by Brown et al. (2005); (b) proposed scheme for the decarboxylation and condensation
mechanism in mtFabH.
longitudinal channel appears to be the pre- required to build large numbers of com-
ferred residence for the 5 substituents, while pounds with structural diversity required to
the 2-amino group and potentially amide build SARs, while retaining the potential to
derivatives thereof could be accommodated yield a candidate drug which would meet the
in the lateral channel. We also noted that a cost specifications of the TTP.
2-bromoacetamido substituent in this posi- The synthesis of our initial fragment-
tion would place the thiol group of Cys112 in sized (Mwt <250) analogues (Fig. 16.8) was
a position to become alkylated via a achieved starting with the Darzens reaction
nucleophilic SN2 reaction, and could lead to between methyl dichloroacetate 5 and the
irreversible inhibition of the enzyme (Fig. appropriate aldehyde. This afforded a mix-
16.7c). While such a strategy normally would ture of the α-chloro glycidic ester and
be performed once a selective and relatively β-chloro α-oxoester, which was extracted
potent inhibitor had been found, we postu- with diethylether and reacted immediately
lated that it would be reasonable at this early with thiourea dissolved in methanol to
stage of inhibitor design to attempt this in generate the methyl ester thiazoles 2, 6–8. To
order to establish if ligand inhibition was at investigate whether a free carboxylic acid
all possible before addressing selectivity in functionality would increase binding within
later rounds of ligand optimization. Based on the active site through facilitating electrostatic
this rationale, we decided to prepare a series interactions proposed by the modelling
of thiazoles that included the substituents studies, we hydrolysed the esters with 0.1 M
studied for evaluation against the enzyme sodium hydroxide solution followed by
and M. tuberculosis. work-up with dilute hydrochloric acid to
generate compounds 9–12. In order to
generate 2-bromoacetamido analogues, the
Chemical synthesis
free amines 2, 6–12 were reacted with
When reviewing the literature to identify 2-bromoacetylchloride in anhydrous tetra-
suitable 5-membered ring isosteres of TLM, hydrofuran at 0°C to afford the amides 3, 13–
the flexible synthetic procedure described by 15. As described previously, hydrolysis with
Barton et al. was identified (Barton et al., 0.1 M sodium hydroxide gave the correspond-
1982). This would facilitate the flexibility ing carboxylic acids 16–19 for comparison.
Fig. 16.7. The modelling studies of the 2-aminothiazole-4-carboxylate analogues with mtFabH. (a) The
hypothetical template of the 2-aminothiazole-4-carboxylates for mtFabH inhibitor development. This
illustrates the key H-bonding interactions with the catalytic triad amino acid residues. (b) The binding
pose of methyl 2-amino-5-methylthiazole-4-carboxylate in the active site of mtFabH showing the NH2
group proximal to His244 and directed towards the lateral channel, with the 5-methyl group directed
towards the longitudinal channel. (c) The binding pose of methyl 2-(2-bromoacetamido)-5-(3-
chlorophenyl)thiazole-4-carboxylate with the bromomethylene portion in the vicinity of the Cys112 thiol
group.
212 G. Coxon
˚
˚C
Fig. 16.8. The flexible synthetic route to the ATCs.
16.2.2 Enzymatic assay versus whole cell are unable to maximize the hydrophobic
assay conflict observed in the interactions with these channels necessary to
development of in vitro SARs facilitate enzyme inhibition. However, insert-
ing a phenyl group at position 5 and
The compounds were first assessed against augmenting it with an m-Cl, as in 14 and 3,
the target enzyme mtFabH using the pro- respectively, allows the appropriate hydro-
cedure developed by Brown et al. (Brown et phobic interactions with the enzyme to occur
al., 2005). Despite many of them not demon- and enables effective inhibition. The flexibility
strating any inhibitory activity at a con- of the ligand also appears to be an important
centration of 200 μg/ml, it was pleasing to see factor; the m-Cl phenyl carboxylic acid
the bromoacetamido analogues that were analogue 18 is inactive, whereas its 5-benzyl
prepared to facilitate an SN2 type substitution counterpart 19, which is less lipophilic, inhibits
between the ligand and the Cys112 residue the enzyme with an IC50 of 225 ± 2.81 μg/ml
were active. The bromoacetamido esters 3, 14 (718 ± 8.97 μm). Comparison with its benzyl
and 15 inhibited the enzyme with IC50 values ester 15, which inhibits mtFabH with an IC50
of 0.95 ± 0.05 μg/ml (2.43 ± 0.13 μm), 1.1 ± 0.1 of 59 ± 1.1 μg/ml (159.8 ± 3.0 μm) suggests that
μg/ml (3.22 ± 0.29 μm) and 59 ± 1.1 μg/ml both hydrophobicity and flexibility at the 5
(159.8 ± 3.0 μm), respectively, while the cor- position of the thiazole ring are instrumental
responding carboxylic acid 19 inhibited the in orientating the ligand to achieve effective
enzyme at 225 ± 2.81 μg/ml (718 ± 8.97 μm). inhibition. It is interesting that 13 fails to
Interestingly, the ester 13 and carboxylic acids inhibit mtFabH, as it appears that the free
16, 17 and 18 failed to inhibit the enzyme. acids are weaker inhibitors than the esters.
It is clear that while the electrophilic This may be accounted for by the molecule
bromomethyl substituent establishes activity possessing a methyl group at position 5, which
against the enzyme, its effect is modified by does not enable it to maximize the hydrophobic
different substituents at the 4- and 5-positions. interactions necessary to bind to the enzyme.
Assuming that the inhibition observed in- Conversely, such interactions may be per-
volves reaction with the Cys112 residue, then mitted by the flexible benzyl substituent on the
the ligands must be situated in the vicinity of free acid 19, which enables weak inhibition.
the catalytic triad. As the longitudinal and While it was encouraging to find that a
lateral tunnels are composed of lipophilic small number of compounds inhibited the
amino acid residues, we suggest that 13 and target enzyme, it was important to know if
16, which possess methyl groups at position 5, this activity would lead to inhibition of the
whole cell organism. From the data obtained, bacteria through an inability to access the
it was clear that all of the bromoacetamido target enzyme due to inappropriate physico-
analogues failed to inhibit M. tuberculosis chemical properties or because the bromo-
H37Rv (Table 16.1). We did not know whether methyl moiety was inactivated chemically or
these compounds did not inhibit the myco- metabolically by the mycobacterium.
Table 16.1. The in vitro activity and molecular properties of the 2-aminothiazole-4-carboxylates.
a,bCompounds regarded as not active (N/A) if no inhibition is observed at 200 μg/ml. cFAS-I/II assay
conducted at 200 μg/ml and compounds regarded as not active is <50% inhibition observed. dAlogP and
logD calculated using Pipeline Pilot (SciTegic) software. eFrom Makarov et al., 2009. fFrom Projan, 2003.
R1 S
R3
N
O
R2
O
TB MIC
mtFabH IC50 μg/ml AlogPd
MWt R1 R2 R3 μg/ml (mM)a (mM)b FAS-Ic FAS-IIc (logD)d
1 (TML) 210.29 – – – 16 (75)e 13 (62.5)f N/Ae Activee 2.62 (1.94)
2 248.3 C6H5CH2 CH3 NH2 N/A 0.06 N/A N/A 2.20 (2.21)
(0.24)
3 389.65 m-Cl- CH3 NHCOCH2Br 0.95 ± 0.05 N/A N/A Active 3.38 (3.38)
C6H5 (2.43 ± 0.13)
6 172.2 CH3 CH3 NH2 N/A 16 (93) N/A N/A 0.51 (0.51)
7 234.27 C6H5 CH3 NH2 N/A N/A N/A N/A 2.17 (2.17)
8 268.72 m-Cl- CH3 NH2 N/A N/A N/A N/A 2.84 (2.84)
C6H5
9 158.1 CH3 H NH2 N/A 0.06 N/A N/A 0.28 (–1.18)
(0.35)
10 220.25 C6H5 H NH2 N/A N/A N/A N/A 1.95 (1.82)
11 254.69 m-Cl- H NH2 N/A 32 (125) N/A N/A 2.61 (2.48)

C6H5
12 234.27 C6H5CH2 H NH2 N/A N/A N/A N/A 1.98 (0.83)
13 293.13 CH3 CH3 NHCOCH2Br N/A N/A Active Active 1.05 (1.05)
14 355.21 C6H5 CH3 NHCOCH2Br 1.1 ± 0.1 N/A N/A Active 2.72 (2.72)
(3.22 ± 0.29)
15 369.23 C6H5CH2 CH3 NHCOCH2Br 59 ± 1.1 N/A N/A N/A 2.75 (2.78)
(159.8 ± 3.0)
16 279.11 CH3 H NHCOCH2Br N/A N/A N/A Active 0.83 (–0.56)
17 341.18 C6H5 H NHCOCH2Br N/A N/A Active Active 2.49 (2.46)
18 375.63 m-Cl- H NHCOCH2Br N/A N/A Active N/A 3.16 (3.13)

C6H5
19 355.21 C6H5CH2 H NHCOCH2Br 225 ± 2.81 N/A Active Active 2.56 (1.38)
(718 ± 8.97)
214 G. Coxon
In contrast to the 2-bromoacetamido vitally important and ensures that the com-
analogues, four of the free amine compounds pound will not be potentially toxic in the
(2, 6, 9 and 11) inhibited M. tuberculosis H37Rv patient and assists with complicated regu-
with MIC values of 0.06, 16, 0.06 and 32 μg/ml latory issues associated with progressing the
(0.24, 93, 0.35 and 125 μm), respectively, while class of compounds through the later stages
the other analogues of this group showed no of drug development.
activity. Given that these compounds did not It is important to recall that the intended
inhibit mtFabH, their mechanism of action site of action of the ATCs is the enzyme
must involve other targets within the mtFabH, which provides the pivotal link in
organism. the mycobacterial FAS-II system. It was thus
Although these compounds exhibit important that selectivity over the related
excellent activity, in the absence of any mammalian FAS-I system was achievable. To
recognizable trends in the series, it is difficult examine selectivity, the compounds were
to ascertain clear SARs. The best activity was assessed using the procedures of Slayden et
obtained with 2, with a benzyl group in the al. (1996) and Brown et al. (2005).
5-position and a methyl ester in the 4-position, When studying the SAR in this regard,
whereas the carboxylic acid analogue was no inhibition of the FAS enzymes was
shown to be inactive. The opposite obser- observed, either as the acid or ester, when the
vations were seen with the inactive methyl 2-position was the free amine (Table 16.1).
ester analogue 8 possessing an m-Cl phenyl These data support the possibility that
group at the 5-position and the corresponding activity against M. tuberculosis H37Rv of
active acid 11. A similar trend was observed compounds 2, 6, 9 and 11 involves a target
for the 5-methyl analogues 9 and 6, with MIC other than mtFabH, or indeed the other
values of 0.06 and 16 μg/ml (0.35 and 93 μm), FAS-II enzymes. However, with the exception
respectively. We speculate that these obser- of 15, all of the compounds possessing the
vations may result from the compounds’ bromoacetamido group at the 2-position
ability to enter the cell. In all cases, the showed activity against the FAS enzymes,
primary amine at the 2-position would be although no obvious trends were observed
associated with a dissociation equilibrium at across this series. This is perhaps not surpris-
physiological pH which could penetrate ing, as these compounds may be inhibiting
cellular membranes in the unionized state. other enzymes in the FAS-II system or indeed
Compounds 9 and 11, on the other hand, with at different sites in the multifunctional FAS-I
both carboxylic acid and amino substituents, complex. However, it was evident from the
would exist as zwitterions and would not data that the phenyl 14 and m-Cl-phenyl 3
normally be expected to penetrate the lipo- analogues were the only compounds active
philic cell wall. The uptake of these com- against mtFabH and selective against FAS-II.
pounds could involve a cellular uptake The mtFabH inhibitor 19 inhibited both FAS-I
mechanism, such as mycobacterial porins and FAS-II, whereas 16 inhibited only FAS-II
that the inactive zwitterionic compounds 10 and 18 FAS-I. Intriguingly, compound 15
and 12 are not substrates for. However, we failed to inhibit the FAS enzymes, although
postulate that the inactivity of the 2-amino inhibited mtFabH with an IC50 of 59 ± 1.1 μg/
analogue 7 is more likely, due to its inability ml (159.8 ± 3.0 μm). While 15 may have shown
to interact structurally with the target in the activity against purified mtFabH, its inhib-
organism. itory potential against mtFabH in the crude
FAS-II assay may be difficult to detect, as the
related enzyme KasA (present in the crude
16.2.3 Specificity and selectivity of the reaction mix) has been shown to have FabH-
2-aminothiazole-4-carboxylates type activity and could have had a com-
pensatory effect (Kremer et al., 2002).
It is of key importance that when one designs Having demonstrated that the killing
a series of molecules in a compound class, action of the ATCs against M. tuberculosis was
they confer a high degree of specificity. This is clearly not via fatty acid metabolism, it was
critical to establish that the compounds much work to be done in the development of
would not confer toxicity against mammalian these compounds. In doing so, care must be
cells. To ascertain this knowledge, and thus taken to retain the non-toxic and oral bio-
understand if this class of compounds still availability properties of the molecule. How-
warranted further investigation in future ever, while the zwitterionic, low molecular
studies, cytoxicity was evaluated against weight compound 9 has a structure that is
human foreskin fibroblast HS-27 cells (Fig. highly hydrophilic, which has implications
16.9) to establish toxicity profiles for our for both absorption and excretion by patients,
compounds. It was reassuring to find that this should be viewed in the context of INH, a
none of the 2-amino analogues or the free successful and routinely administered anti-
carboxylic analogues of the bromoacetamido tuberculosis drug, which also has a low
compounds showed significant cytotoxicity molecular weight and a logP value of –1.1.
at a concentration of 100 μg/ml. Conversely, Further information is needed regarding
the carboxylic esters of the bromoacetamido how these compounds work. For example, it
compounds 3, 13, 14 and 15 all showed signs is clear it is more advantageous that the action
of significant cytotoxicity. We speculate that of the compounds is to destroy the myco-
this may be due to the indiscriminate alky- bacteria outright rather than simply inhibit
lation of essential cellular components rather its growth. Similarly, it is important to
than the increased ability of these esters to understand if the activity of these compounds
penetrate the cells over the carboxylic acids. in a multi-drug regime with existing anti-
This is supported by the fact that the non- tubercular drugs such as INH or rifamipicin
cytotoxic acids 17, 18 and 19 have comparable provide an additive, synergistic or antagon-
logD values to those of the cytotoxic esters 13, istic effect.
14 and 15, and thus possess similar physico- Pharmacodynamic (PD) and pharma-
chemical characteristics. cokinetic (PK) considerations also require
investigation. Early, in vitro PD can be
investigated quickly in the mouse macro-
16.3 Conclusions and Future phage model, where one would hope to see
Direction similar killing action against the mycobacteria
as observed in vitro, and also evidence of
Although a promising new scaffold has been SARs is required in order to optimize activity.
identified with excellent in vitro activity Critical information about the initial PK
against M. tuberculosis and promising selec- drivers, or routes of metabolism, for these
tivity versus mammalian cells, there remains compounds may be obtained using HPLC,
Conc.
Cytotoxicity
(% control)
c
Fig. 16.9. The cytotoxic effects of the compounds against HS-27 human fibroblast cells.
216 G. Coxon
MS and NMR methodology following incu- physico-chemical knowledge (pKa, logD,

bation of the compounds in rat hepatocyte or polar surface area) and could be used to select
microsomal assays. Information about the compounds for in vivo work. As many of the
potential formation of reactive metabolites ATCs have polarizable groups, and in some
may also be generated using similar experi- cases are zwitterionic, then these may not be
ments where levels of glutathione are moni- expected to have good passive diffusion
tored and structural elucidation of the formed properties. However, it is possible that an
metabolites is performed. Hepatic cytotoxicity active uptake mechanism may help adsorb
may also be established and comparison of highly charged molecules through the gut
this data in the presence of a blanket Cyp450 wall and thus the Caco-2 model may be an
inhibitor will indicate whether toxicity is a appropriate experiment.
phenomenon of the intrinsic structure of the Of course, further preclinical develop-
compound or its metabolite(s). All of these ment will be more time-consuming and
data can help determine the progression of a resource-intensive and thus knowledge of the
chemical series. mechanism of action of the compounds,
Of course, obtaining in vivo data is critical which is a regulatory requirement, should be
and a murine model may be used in the first gained before embarking on further develop-
instance to understand acute toxicity, initial ment. To address this, in a similar strategy
PK and then efficacy. However, being labour- used to identify the target of the benzo-
and time-intensive and requiring significant thiazinones, mutants may be selected and
resources, there are a number of in vitro comparison of their DNA wild-type
experiments which may be further used to sequences performed (Makarov et al., 2009).
‘triage’ promising compounds. These include It is clear that the development of the
plasma stability models, which give an ATCs is at an early stage. However, they are
indication of structural stability in the very active against M. tuberculosis, simple to
presence of plasma esterase that may make and are non-toxic against human cell
hydrolyse ester groups. This is, of course, lines. Moreover, important lessons have been
important if the retention of ATC ester learned during their discovery in that while a
functionality is critical for activity. single or multiple target approach may be
Plasma protein binding should also be used as a template for ligand design, this
measured in vitro. Compounds can bind to often does not correlate with whole cell
albumin (HAS), α1-acid glycoprotein (AGP) activity. Additionally, it can be hard to
or lipoproteins in the blood, and this reduces develop meaningful SARs from a small set of
the amount of free drug in solution for compounds and thus a synthesis of larger
penetration into tissue to reach the therapeutic libraries of the same structural class with
target or to the liver and kidney for increased structural diversity may be re-
elimination. As the ATCs feature acid and quired. There is no doubt, however, that the
basic moieties in their structure, this is an basic ATC scaffold yields promise for further
important experiment as these features may investigation as a potential new tuberculosis
be expected to lead to binding to HAS drug.
(potential 2-amino and 4-carboxyl group
binding) and AGP (potential 2-amino bind-
ing), and the extent of this binding needs to References
be identified.
The permeability properties of the ATCs Al-Balas, Q., Anthony, N.G., Al-Jaidi, B., Alnimr, A.,
may be evaluated in vitro by two possible Abbott, G., Brown, A.K., et al. (2009)
Identification of 2-aminothiazole-4-carboxylate
methods, by using a parallel artificial mem-
derivatives active against Mycobacterium
brane permeability assay (PAMPA) or Caco-2 tuberculosis H37Rv and the beta-ketoacyl-ACP
methods. This will give important infor- synthase mtFabH. PLoS One 4(5), e5617.
mation about the potential of the molecules’ Ashek, A. and Cho, S.J. (2006) A combined
ability to move across the gut and into the approach of docking and 3D QSAR study of
bloodstream and may be used along with beta-ketoacyl-acyl carrier protein synthase III
(FabH) inhibitors. Bioorganic and Medicinal Kim, P., Zhang, Y.M., Shenoy, G., Nguyen, Q.A.,
Chemistry 14(5), 1474–1482. Boshoff, H.I., Manjunatha, U.H., et al. (2006)
Balganesh, T.S., Balasubramanian, V. and Kumar, Structure–activity relationships at the 5-position
S.A. (2004) Drug discovery for tuberculosis: of thiolactomycin: an intact (5R)-isoprene unit is
bottlenecks and path forward. Current Science required for activity against the condensing
86(1), 167–176. enzymes from Mycobacterium tuberculosis and
Balganesh, T.S., Alzari, P.M. and Cole, S.T. (2008) Escherichia coli. Journal of Medicinal Chemistry
Rising standards for tuberculosis drug 49(1), 159–171.
development. Trends in Pharmacological Kremer, L., Douglas, J.D., Baulard, A.R., More-
Sciences 29(11), 576–581. house, C., Guy, M.R., Alland, D., et al. (2000)
Barry, C.E. 3rd (2009) Unorthodox approach to the Thiolactomycin and related analogues as novel
development of a new antituberculosis therapy. anti-mycobacterial agents targeting KasA and
New England Journal of Medicine 360(23), KasB condensing enzymes in Mycobacterium
2466–2467. tuberculosis. Journal of Biological Chemistry
Barton, A., Breukelman, S.P., Kaye, P.T., Meakins, 275(22), 16857–16864.
G.D. and Morgan, D.J. (1982) The preparation of Kremer, L., Dover, L.G., Carrere, S., Nampoothiri,
thiazole-4-carboxylates and thiazole-5-carboxy- K.M., Lesjean, S., Brown, A.K., et al. (2002)
lates, and an infrared study of their rotational Mycolic acid biosynthesis and enzymic charac-
isomers. Journal of the Chemical Society, terization of the beta-ketoacyl-ACP synthase
Perkin Transactions 1(1), 159–164. A-condensing enzyme from Mycobacterium
Bhowruth, V., Brown, A.K., Senior, S.J., Snaith, J.S. tuberculosis. Biochemical Journal 364(Pt 2),
and Besra, G.S. (2007) Synthesis and biological 423–430.
evaluation of a C5-biphenyl thiolactomycin Lerner, C.G., Hajduk, P.J., Wagner, R., Wagenaar,
library. Bioorganic and Medicinal Chemistry F.L., Woodall, C., Gu, Y.G., et al. (2007) From
Letters 17(20), 5643–5646. bacterial genomes to novel antibacterial agents:
Brown, A.K., Sridharan, S., Kremer, L., Lindenberg, discovery, characterization, and antibacterial
S., Dover, L.G., Sacchettini, J.C., et al. (2005) activity of compounds that bind to HI0065
Probing the mechanism of the Mycobacterium (YjeE) from Haemophilus influenzae. Chemical
tuberculosis beta-ketoacyl-acyl carrier protein Biology and Drug Design 69(6), 395–404.
synthase III mtFabH: factors influencing Makarov, V., Manina, G., Mikusova, K., Mollmann,
catalysis and substrate specificity. Journal of U., Ryabova, O., Saint-Joanis, B., et al. (2009)
Biological Chemistry 280(37), 32539–32547. Benzothiazinones kill Mycobacterium tubercu-
Choi, K.H., Kremer, L., Besra, G.S. and Rock, C.O. losis by blocking arabinan synthesis. Science
(2000) Identification and substrate specificity of 324(5928), 801–804.
beta-ketoacyl (acyl carrier protein) synthase III Payne, D.J., Gwynn, M.N., Holmes, D.J. and
(mtFabH) from Mycobacterium tuberculosis. Pompliano, D.L. (2007) Drugs for bad bugs:
Journal of Biological Chemistry 275(36), confronting the challenges of antibacterial
28201–28207. discovery. Nature Reviews Drug Discovery 6(1),
Cole, S.T., Brosch, R., Parkhill, J., Garnier, T., 29–40.
Churcher, C., Harris, D., et al. (1998) Portevin, D., De Sousa-D’Auria, C., Houssin, C.,
Deciphering the biology of Mycobacterium Grimaldi, C., Chami, M., Daffe, M., et al. (2004)
tuberculosis from the complete genome A polyketide synthase catalyzes the last
sequence. Nature 393(6685), 537–544. condensation step of mycolic acid biosynthesis
Douglas, J.D., Senior, S.J., Morehouse, C., in mycobacteria and related organisms. Pro-
Phetsukiri, B., Campbell, I.B., Besra, G.S., et al. ceedings of the National Academy of Sciences
(2002) Analogues of thiolactomycin: potential 101(1), 314–319.
drugs with enhanced anti-mycobacterial activity. Price, A.C., Choi, K.H., Heath, R.J., Li, Z., White,
Microbiology 148(Pt 10), 3101–3109. S.W. and Rock, C.O. (2001) Inhibition of beta-
Jones, G., Willett, P. and Glen, R.C. (1995) Molecu- ketoacyl-acyl carrier protein synthases by
lar recognition of receptor sites using a genetic thiolactomycin and cerulenin. Structure and
algorithm with a description of desolvation. mechanism. Journal of Biological Chemistry
Journal of Molecular Biology 245(1), 43–53. 276(9), 6551–6559.
Kamal, A., Shaik, A.A., Sinha, R., Yadav, J.S. and Projan, S.J. (2003) Why is big Pharma getting out
Arora, S.K. (2005) Antitubercular agents. Part 2: of antibacterial drug discovery? Current Opinion
new thiolactomycin analogues active against in Microbiology 6(5), 427–430.
Mycobacterium tuberculosis. Bioorganic and Sachdeva, S. and Reynolds, K.A. (2008) Myco-
Medicinal Chemistry Letters 15(7), 1927–1929. bacterium tuberculosis beta-ketoacyl acyl
218 G. Coxon
carrier protein synthase III (mtFabH) assay: tuberculosis mtFabH fatty acid condensing
principles and method. Methods in Molecular enzyme. Bioorganic and Medicinal Chemistry
Medicine 142, 205–213. Letters 13(21), 3685–3688.
Sachdeva, S., Musayev, F., Alhamadsheh, M.M., Senior, S.J., Illarionov, P.A., Gurcha, S.S.,
Neel Scarsdale, J., Tonie Wright, H. and Campbell, I.B., Schaeffer, M.L., Minnikin, D.E.,
Reynolds, K.A. (2008a) Probing reactivity and et al. (2004). Acetylene-based analogues of
substrate specificity of both subunits of the thiolactomycin, active against Mycobacterium
dimeric Mycobacterium tuberculosis FabH using tuberculosis mtFabH fatty acid condensing
alkyl-CoA disulfide inhibitors and acyl-CoA enzyme. Bioorganic and Medicinal Chemistry
substrates. Bioorganic Chemistry 36(2), 85–90. Letters 14(2), 373–376.
Sachdeva, S., Musayev, F.N., Alhamadsheh, M.M., Slayden, R.A., Lee, R.E., Armour, J.W., Cooper,
Scarsdale, J.N., Wright, H.T. and Reynolds, K.A. A.M., Orme, I.M., Brennan, P.J., et al. (1996)
(2008b) Separate entrance and exit portals for Antimycobacterial action of thiolactomycin: an
ligand traffic in Mycobacterium tuberculosis inhibitor of fatty acid and mycolic acid synthesis.
FabH. Chemistry and Biology 15(4), 402–412. Antimicrobial Agents and Chemotherapy
Schaeffer, M.L., Agnihotri, G., Volker, C., Kallender, 40(12), 2813–2819.
H., Brennan, P.J. and Lonsdale, J.T. (2001) Takayama, K., Wang, C. and Besra, G.S. (2005)
Purification and biochemical characterization of Pathway to synthesis and processing of mycolic
the Mycobacterium tuberculosis beta-ketoacyl- acids in Mycobacterium tuberculosis. Clinical
acyl carrier protein synthases KasA and KasB. Microbiology Reviews 18(1), 81–101.
Journal of Biological Chemistry 276(50), Yuan, Y., Lee, R.E., Besra, G.S., Belisle, J.T. and
47029–47037. Barry, C.E. 3rd (1995) Identification of a gene
Senior, S.J., Illarionov, P.A., Gurcha, S.S., involved in the biosynthesis of cyclopropanated
Campbell, I.B., Schaeffer, M.L., Minnikin, D.E., mycolic acids in Mycobacterium tuberculosis.
et al. (2003) Biphenyl-based analogues of Proceedings of the National Academy of
thiolactomycin, active against Mycobacterium Sciences 92(14), 6630–6634.
17 Rifamycins Revisited
Martin J. Boeree
Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands
17.1 Introduction highly bactericidal (Maggi et al., 1966; Lester,

1972; Sensi, 1983). With rifampicin (RIF or R)
In 1957, Sensi and co-workers at Group available, the use of rifamycin B was almost
Lepetit SpA Laboratories in Milan, Italy, fully abolished, although it remained
isolated a new drug with antibiotic properties available for use in some (mainly European)
which they named rifomycin (Sensi et al., countries (van Ingen et al., 2011b).
1959); the name was later changed to rifa- After the introduction of RIF, optimism
mycin. The rifamycins are a member of the came into the tuberculosis scientific society
ansamycin family of antibiotics. Rifamycin and tuberculosis control institutions. The
was extracted from fermentation cultures of possibility of elimination was even becoming
Amycolatopsis rifamycinica (previously Strepto- a feasible target. This led to several years of
myces mediterranei). It actually consisted of relative apathy in tuberculosis drug discovery.
seven substances and therefore was renamed Yet several derivatives of RIF were syn-
rifamycin A, B, C, D, E, S and SV. These thesized and tested for clinical use in search
substances were poorly absorbed and were for even more efficiency in terms of duration,
first developed as parenteral agents. Rifa- dosing intervals and tolerability. Rifapentine
mycin B was the most stable, least toxic and (RPT) (2008d) was approved for use in
active against a broad spectrum of bacteria, humans in 1998. It is attractive because of its
mainly Gram-positive cocci and Myco- long half-life. It needs to be dosed once
bacterium tuberculosis (Lester, 1972; Sensi, weekly, or even once fortnightly. Rifalazil
1983). Peculiarly, the name rifamycin was (2008b) has an equally long half-life, but
dedicated to the popular 1955 French ‘film never made it to registration because of
noir’ movie, Rififi (see Fig. 17.1) (Sensi, 1983). adverse events. Rifabutin (RBT) (2008a) was
Rifamycin B was not much used clinically in first released in 1983 on a compassionate use
tuberculosis treatment, because of its toxicity basis for the treatment of disseminated M.
and parenteral administration. In the years to avium infection. It is not registered for use in
follow, the Sensi group tried to synthesize an tuberculosis (Davies et al., 2007). RBT is
oral equivalent with good intestinal absorp- probably as equally efficient as RIF, has more
tion. In 1965, rifampicin, a hydrazone of a adverse events but has favourable pharmaco-
rifamycin B derivate with N-amino-N´- logical properties to use concurrently with
methylpiperazine, showed to be well antiretroviral drugs.
absorbed orally and, moreover, was still All rifamycins express their anti-tubercu-

220 M.J. Boeree
Fig. 17.1. The Rififi film poster, French version. The rifamycins were named after the novel and film.
losis activity by inhibiting the essential the rifamycins works in >97% of cases,
β-subunit of RNA polymerase (RNAP) by through mutations in the RIF resistance-
binding to it. This subunit is produced by the determining region (RRDR) around amino
rpoB gene (2008c). In addition, the rifamycins acids 513–531 of the rpoB gene (Huitric et al.,
may act bactericidally by activating the 2006). Some mutations, especially in codon
‘suicide gene module’, mazEF, and hence 531, cause resistance to all rifamycins and
inducing apoptosis of the mycobacteria. A some mutations in codons 511 and 516 result
vexing problem in tuberculosis treatment in resistance specific to RIF and RPT, but not
with all currently used tuberculosis drugs is to RBT.
the rapid emergence of resistance, especially In this chapter, the several rifamycins
if used in monotherapy. The existing rifa- will be reviewed, with a special emphasis on
mycins are no exception. Drug resistance for rifampicin, its chosen dosage, the potential
Rifamycins Revisited 221
role of rifapentine, the place rifabutin has high cost (Hong Kong and Singapore) (Fox et
taken and the rifamycins that have not yet al., 1999). At the same time, the introduction
made it to clinical use. of RIF containing first-line regimens was
advocated for four reasons: easier adminis-
tration, less toxicity, faster conversion and –
17.2 Rifampicin the most important – a shorter duration of
therapy. The shortening of treatment would
17.2.1 History therefore justify (and, in part, make up for)
the expense (Houk, 1972). In the second half
Rifampicin (also rifampin, rifamycin AMP or the 1970s, RIF gained a foothold as a first-line
rifaldazine, abbreviated RIF, RMP or R; Fig. drug, at least for countries that could afford
17.2) was registered by the FDA in 1971 its use (McConville and Rapoport, 1976;
(Sensi, 1983). By this time, several trials and Angel, 1977; Robinson, 1978). The body of
case series had established efficacy for RIF- evidence to support the use of RIF and
containing regimens in tuberculosis treatment introduce successful treatment with 6-month
(Maggi et al., 1966; De et al., 1968; Gyselen et short-course regimens overruled the associ-
al., 1968; Newman et al., 1971; Nitti et al., 1971; ated cost (McConville and Rapoport, 1976;
Constans et al., 1972; Corpe and Sanchez, Angel, 1977; Long et al., 1979). Still, use of this
1972; Davidson et al., 1972; Favez et al., 1972). costly drug was, at times, met with con-
Originally, RIF was reserved to be a second- troversy (Gosling et al., 2003). Eventually, RIF
line drug, since first-line triple therapy treatment made its way into tuberculosis
(streptomycin (S or SM), isoniazid (INH or H) treatment guidelines, but it did so in the
and para-amino-salicylic acid (PAS or P)) was lowest effective dosage of 10 mg/kg (Angel,
already highly effective and RIF was con- 1977; Fox et al., 1999). RIF had been shown to
sidered to be too dire (Citron, 1972). An be efficient, expressed by favourable cure
editorial in the British Medical Journal stated rates and low relapse rates, with a con-
that the price for 10 g of sodium PAS was siderable shorter duration of the standard
£02.5, 1 g of ethambutol was £0.22 and 600 mg treatment of at least 12–18 months. Hence,
of RIF was £0.68 (1973) (the latter comparable these RIF-containing regimens of short dur-
to £3.79 or US$6.05 today). The British ation were called – at present, paradoxically
Medical Research Council (BMRC) preferred – short-course chemotherapy.
INH and pyrazinamide (PZA or Z) regimens
over equally active INH–RIF regimens, as RIF
cost over £110 for 6 months, more than four 17.2.2 Microbiological activity
times the cost of PZA. In a respected review
on tuberculosis treatment, Fox et al. repeatedly RIF has potent antibiotic activity against M.
remind us that the first short-course therapy tuberculosis. In vitro, the minimum inhibitory
trials were designed ‘to use the minimum concentration (MIC) against the laboratory
amount of expensive rifampicin’ (in East strain H37Rv is 0.1–0.4 μg/ml (Rastogi et al.,
Africa) or did not include RIF because of the 1996). The MIC90 (the MIC in which 90% of
the bacterial population is inhibited) is 0.25
μg/ml. RIF also acts as exposure dependent in
macrophages and other cells (Jayaram et al.,
2003). RIF is heavily protein bound (83%),
which leads to increased MIC in serum. It is
bactericidal. There are in vitro synergies with
several quinolones, macrolides, ethambutol
and streptomycin (Bhusal et al., 2005).
Recently, a synergy with the efflux pump
inhibitor, thioridazine, has been demonstrated
(see Chapter 18). For INH, there is no synergy
Fig. 17.2. Rifamycin. observed. In the animal model, RIF has
222 M.J. Boeree
shown its sterilizing efficacy repeatedly, and 600 mg and 750 mg arms. The regimen
in the mouse model, it has demonstrated its containing 450 mg was significantly less
sterilizing activity repeatedly. In one study, effective than the other two regimens, with a
there was a complete sterilization of the mice lower rate of sputum conversion and a higher
given INH 25 mg/kg and RIF 25 mg/kg for 9 rate of treatment failures (Long et al., 1979).
months. If RIF was withdrawn in the last 3 Interestingly, the 600 mg dose had become
months, there was a 20% relapse rate. The standard practice even before the FDA’s
existing evidence shows that there is a dose approval of RIF for tuberculosis treatment in
relation; in the extreme, a complete steriliza- 1971, and it is still recommended as the
tion was achieved in the infected mice in 6 maximum daily dose (World Health
days with a dose of 810 mg/kg (2008c). Organization, 2009). Other trials had studied
doses other than 600 mg, but the differences
were either small (±150 mg in the USPHS
17.2.3 Clinical efficacy study) or were part of regimens that deviated
so far from common practice that it was
RIF is an efficient drug against drug-sensitive difficult to single out the effect of RIF. An
tuberculosis in humans. In a combination important example is the study by Kreis et al.,
regimen with INH and PZA for 2 months in published in 1976, which tested two 3-month
an intensive phase, followed by a continuation regimens of high-dose RIF (1200 mg daily or
phase of 4 months, cure rates are high: in every other day) combined with high daily
programme conditions, 85% is considered to doses of INH and S. These regimens had
be realistic and in the early trials, cure rates almost 100% sputum culture conversion, but
are >95% at a dose of 10 mg/kg (Long et al., 16% of patients relapsed after 12–24 months
1979). In the studies that showed efficacy for (Kreis et al., 1976). From today’s perspective,
RIF, eventually a daily dosage of 10 mg/kg the most likely contributor to the high initial
was chosen, for several reasons (1972; Gyselen cure rate achieved is the high dose of RIF.
et al., 1968; Newman et al., 1971; Nitti et al., This also makes the daily dose of 600 mg RIF
1971; Constans et al., 1972; Corpe and worth being reconsidered. More importantly,
Sanchez, 1972; Davidson et al., 1972). Whether what renders a dose ‘optimal’ from today’s
this was the right choice is currently being point of view? According to current stand-
debated: there are several arguments to ards, the optimal dose of RIF would be
assume that the dosage is at the lower end of derived from the relationships between dos-
the dose–response curve. Higher dosages of ing, drug exposure achieved and desirable
RIF are currently being investigated for their and undesirable responses in Phase I and II
potential to shorten the duration of tubercu- studies with clearly differing doses of RIF,
losis treatment. Van Ingen et al., 2011b) have followed by pivotal Phase III studies of
performed a literature search focusing on the regimens. No such studies have been done. In
dosing question, reviewing papers published 2010, 2011 and 2012 three Phase II trials were
in the first two decades after the development executed in Africa to establish a more accurate
of RIF. The most explicit comment came from maximum tolerated dose with a proof a
Richard O’Brien and Andrew Vernon, who, concept for more efficacy in terms of shorter
in their 1998 editorial (O’Brien and Vernon, duration of treatment.
1998), stated that determination of the
optimal dose of RIF ‘was done by the United
States Public Health Service (USPHS) 17.2.4 Pharmacokinetics
Tuberculosis Study 19 from 1979 which
established 600 mg as the optimal dose for Early pharmacokinetic studies showed that a
most adults’. Study 19 evaluated a daily single daily dose of 600 mg of RIF resulted in
dosage of 450 (7.5 mg/kg), 600 (10 mg/kg) and serum concentrations of 7.0 μg/ml 90 min
750 mg (12.5 mg/kg) of RIF with a fixed dose after ingestion, or 8.80–12.0 μg/ml after 2 h,
of INH and observed no significant difference i.e. well above 0.2 μg/ml, the mean MIC of M.
in sputum conversion or relapse between the tuberculosis (Furesz et al., 1967; Verbist and
Gyselen, 1968; Nitti et al., 1972). Higher doses dependent; in other words, the area under the
were explored and showed significantly concentration–time curve (AUC) divided by
higher serum concentrations and half-life, the MIC correlates better with killing than
suggesting a non-linear association (Constans Cmax/MIC (Burman et al., 2001; Peloquin,
et al., 1968; Acocella et al., 1971; Ruslami et al., 2001; Jayaram et al., 2003; Nuermberger and
2006); one early study in France applied 900 Grosset, 2004; Gumbo et al., 2007).
mg RIF once daily and found average serum Pharmacokinetic studies generally have
concentrations of 16.2 μg/ml 3 h after intake been based on serum assays. M. tuberculosis is
(Constans et al., 1968). Since peak serum predominantly an intracellular organism, so
concentrations (Cmax) are reached 2 h after serum may not be the right compartment. A
intake (Acocella et al., 1971; Nitti et al., 1972; few studies have examined RIF concentrations
Ruslami et al., 2006), the Cmax was likely to be in alveolar macrophages and epithelial lining
higher. A year later, Furesz and co-workers fluid (ELF) (Ziglam et al., 2002; Goutelle et al.,
measured peak serum concentrations (after 2 2009). RIF concentrations in ELF and
h) of 20.87 ± 3.25 after a single 750 mg dose bronchial biopsies are slightly below those in
and 27.70 ± 4·16 μg/ml after administration of serum, whereas those in alveolar macro-
900 mg of RIF in healthy volunteers. Yet they phages are over ten times higher (Ziglam et
stated that the 600 mg doses already ensured al., 2002). As a result, AUC/MIC or Cmax/MIC
therapeutic (i.e. equal to or above MIC) blood ratios compatible with bacterial killing are
levels for 24 h after administration (Furesz et generally achievable in serum and alveolar
al., 1967). Although this pharmacokinetic macrophages, but not in ELF. The use of a
argument is compelling, it is based on the 1200 mg RIF dose improved significantly the
assumption that RIF is active when serum attainment of AUC/MIC or Cmax/MIC ratios
concentrations exceed the MIC throughout compatible with bacterial killing (Goutelle et
the dosing interval, which means that the al., 2009). Importantly, if the 600 mg dose was
ratio of the trough concentration (Cmin) to applied, RIF concentrations attained at the
MIC is the relevant pharmacodynamic index site of infection were too low. Recently,
(‘time-dependent inhibition’) (Pallanza et al., Gumbo et al. showed – as in serum – that
1967; Verbist and Gyselen, 1968; Favez et al., exposure to higher concentrations of RIF
1972; Lester, 1972). Similar reasoning is might lead to a non-linear increase in RIF
illustrated by the statement of Constans et al. concentrations inside the bacteria. The exact
that the 900 mg dose applied in a pilot of their mechanism is not known; it may be related to
trial was ‘unnecessarily high’ (Constans et al., saturation of bacterial efflux pumps (Ruslami
1972). et al., 2006; Gumbo et al., 2007). The question
In the years following, this assumption is whether increasing the concentrations in
changed into the view that the efficacy of RIF the bacillus and at the site of infection actually
was concentration dependent, i.e. correlating improves killing and improves treatment
with peak plasma concentration (Cmax) outcome. Studies of higher doses of RIF and
divided by the MIC (Pallanza et al., 1967; other rifamycins are therefore important and
Nuermberger and Grosset, 2004). This view are being performed at the time of writing of
was based on the intracellular mode of action this chapter.
of RIF; other drugs with intracellular targets, In conclusion, the current dose of RIF
such as aminoglycosides and fluoroquino- used in internationally recommended regi-
lones, were shown to have concentration- mens has not been based on careful evaluation
dependent activity. Concentration-dependent of the relationships between dosing, drug
activity and the associated post-antibiotic exposure achieved at the site of infection and
effect also explained the efficacy of RIF better desirable and undesirable responses. By the
when it was administered intermittently time of the introduction of the drug, and
(Pallanza et al., 1967; Nuermberger and because of its success, a proper dose-
Grosset, 2004). Recent studies have challenged escalating study had not been performed.
this assumption again and established that Recent studies suggest that the current dose
the activity of RIF is more exposure of RIF is at the lower end of the dose–response
224 M.J. Boeree
curve (Mitchison, 2000; Jayaram et al., 2003; mittent therapy with high doses of RIF (once
Peloquin, 2003), as suggested already by the or twice weekly) has been associated with
study of Kreis et al. mentioned previously increased toxicity (Poole et al., 1971; Peloquin,
(Kreis et al., 1976). The probable non-linear 2001, 2003; Ruslami et al., 2006). Poole and co-
increase of RIF concentrations means that a authors noted that intermittent high doses of
relatively small increase in dose is associated RIF (1200 mg twice weekly with 900 mg INH)
with a more than proportional increase in RIF led to RIF sensitization and antibody for-
AUC. Clearly, this characteristic should be mation; 11 (22%) of the 49 patients dis-
assessed further in future studies that explore continued treatment after developing mostly
the utility of higher doses of RIF. fever, thrombocytopaenia or renal failure,
designated as the ‘flu-like syndrome’. The
high incidence of flu-like syndrome has been
17.2.5 Adverse events ascribed to the intermittency of dosing rather
than the height of the dose. Nevertheless, this
Most frequently and remarkably, RIF causes a experience resulted in an end to clinical trials
red-orange staining of all body fluids. Other with high-dose RIF. In retrospect, it appears
adverse effects of RIF are mainly gastro- that this was an overreaction that might have
intestinal discomfort (nausea, vomiting, prevented the research community identify-
diarrhoea) and, more infrequently, hepatitis ing the correct dosage regimen for RIF. It has
and hypersensitivity reactions such as been said that the ‘baby was thrown out with
thrombocytopaenia, haemolytic anaemia and the bath water’. It is notable that higher doses
interstitial nephritis. Quite rare side effects of RIF (900–2400 mg) have been used in other
include hypotension and shock, shortness of diseases, for example brucellosis, osteo-
breath, organic brain syndrome and per- myelitis and leishmaniasis, without signifi-
ipheral neuropathy. cant tolerability problems. In these contexts,
One argument for not using higher doses RIF is co-administered with less toxic drugs
of RIF is that of these feared side effects. such as co-trimoxazole, and it is used for
There is little evidence that higher daily doses weeks rather than months (Solera et al., 1995;
lead to increased toxicity, which may also be Kochar et al., 2000).
expected because of the non-linear increase of RIF may not be the most toxic drug in the
pharmacokinetic (PK) parameters. However, multi-drug regimen currently used for
this has not been observed in the few studies tuberculosis. In studies of 4-month RIF
that have looked at higher doses of RIF. The monotherapy for latent tuberculosis, hepato-
few that have been performed report little or toxicity was rare compared with its frequency
no safety and toxicity data (Constans et al., in RIF–PZA treatment. This may be due to the
1968; Acocella et al., 1971; Favez et al., 1972; PZA or to drug interactions rather than to RIF
Kreis et al., 1976; Ruslami et al., 2006). In the itself (Ziakas and Mylonakis, 2009). The
study with 900 mg of RIF, Constans and co- arrival of novel anti-tuberculosis drugs and
workers said that this was ‘liable to induce novel combinations may shed more light on
slight disorders’, without providing additional this issue and allow higher doses of RIF to be
data (Constans et al., 1972); Favez and co- used safely.
workers drew similar conclusions (Favez et
al., 1972). In the USPHS Study 19, drug-
induced hepatitis frequency did not differ 17.3 Rifapentine
between patients who were given 450, 600 or
750 mg of RIF (Long et al., 1979). Recent 17.3.1 History
studies of higher RIF doses (13 mg/kg and 20
mg/kg) did not report increased hepatotoxicity In 1975, the same LePetit group from Milan
or other adverse effects (Ruslami et al., 2006; who discovered RIF developed a new
Diacon et al., 2007). The patient numbers of derivative with a long half-life: DL 473, later
the studies were small. In contrast, inter- called rifapentine (cyclopentylrifamycin, RPT;
Consortium (TBTC) Study 25 in 35 patients,

in which three doses of RPT once weekly,
combined with INH of 15 mg/kg, were
evaluated (Weiner et al., 2004). Similarly, it
was demonstrated that low INH caused an
unacceptable failure rate in the HIV-positive
population. The current role for RPT has been
assessed recently in a systematic review (Gao
et al., 2009). Based on nine evaluable clinical
trials, they concluded that once or twice
Fig. 17.3. Rifapentine. weekly RPT and daily RIF had a similar
outcome on cure rates and safety for the
treatment of HIV-negative pulmonary
tuberculosis, but once-weekly or less frequent
Fig. 17.3) (Arioli et al., 1981). RPT was approved use of RPT increased the risk of bacteriological
by the FDA in 1993 for the treatment of relapse. Additionally, they observed that the
tuberculosis. The long half-life permitted RPT risk of resistance to RIF in HIV-positive
to be dosed once or twice weekly. patients was increased. The hope for a role for
RPT in the treatment in drug-sensitive
tuberculosis is being focused on its potential
17.3.2 Microbiological activity to shorten treatment with higher dosages,
based mainly on the exciting mouse experi-
RPT has the same mechanism of action as ments of Grosset’s group (Rosenthal et al.,
RIF. In vitro, the MIC against the M. tubercu- 2006, 2007). There are several ongoing trials
losis strain H37Rv is 0.031 μg/ml and is about by various research groups, such as the TBTC
tenfold lower than RIF (2008d). There are and InterTB, to investigate this concept. Very
post-antibiotic effects measured with RPT, recently, RPT in combination with INH for 3
especially with INH and moxifloxacin. In the months was shown to be non-inferior to 9
mouse model, RPT was clearly efficient, months of INH alone in the treatment of
especially in higher doses (10–15 mg/kg). RPT latent tuberculosis infection (Sterling and
was not able to sterilize the mice completely, TBTC, 2011).
but did so in combination with INH and PZA
(Daniel et al., 2000). In another experiment,
the activity of RPT was especially enhanced 17.3.4 Pharmocokinetics
when INH was added in the same weekly
dose frequency (Chapuis et al., 1994). Also, a RPT was introduced as a more potent and
combination with moxifloxacin showed good longer-acting rifamycin. The notable charac-
sterilizing activity, though only when MOX teristic of RPT is its high ratio of protein
was dosed on a daily basis (Veziris et al., binding (97%), with a long half-life as result.
2005). In a series of experiments, it was shown Bioavailability is about 70%. Tissue levels are
that the duration of treatment could be generally higher than in plasma concen-
reduced considerably by higher doses of RPT trations (Burman et al., 2001). Like RIF, there
in mice (Rosenthal et al., 2006, 2007). is a long post-antibiotic effect (75 h) (Chan et
al., 2004). The other PK characteristics are
outlined in Table 17.1.
17.3.3 Clinical efficacy
The registration for RPT is restricted for use 17.3.5 Adverse events
in HIV-negative patients who are sputum
negative at 2 months treatment. This is based The adverse events profile very much
on clinical data from the Tuberculosis Trials resembles that of RIF. Hepatitis is as frequent
226 M.J. Boeree
Table 17.1. PK parameters of rifapentine.

Volume
distribution Clearance PK
Species Half-life (h) AUC (mg/h/l) Cmax (μg/ml) (l/kg) (l/h) methodology
Mouse – 309 ± 35.5a 11.1 ± 3.09a – – Single dose of
474 ± 5.8b 16.7 ± 1.1b MOXI 100
mg/kg in
mice of a10
mg/kg and
b15 mg/kg.
Human 13.18 ± 7.38 319.54 ± 91.52 15.05 ± 4.62 – 2.03 ± 0.6 Dose 300 mg/
day. PK
determined
at day 10.
25-deactyl 13.35 ± 2.67 215.88 ± 86 6.26 ± 2.06 – –
RIFAP
metabolite
Notes: a10 mg/kg; b15 mg/kg.
as for RIF. There is a difference in the

occurrence of the flu-like syndrome, as seen
in intermittent admission of higher doses of
RIF. For RPT, this phenomenon is observed
infrequently.
17.4 Rifabutin
17.4.1 History
In the search for better and safer rifamycin

derivatives, Marsili – again, from Milan –
Fig. 17.4. Rifabutin.
described a spiropiperidyl derivative of
rifamycin S in 1981 (Marsili et al., 1981). It was
called ansamycine LM 427 and later rifabutin
(RBT) or mycobutin (Fig. 17.4). The drug had and half-life were lowest (Ji et al., 1993). M.
good properties against mycobacteria, tuberculosis may retain sensitivity to RBT in
especially against M. avium, and was initially 30% of the strains resistant to RIF, because of
for compassionate use in disseminated M. a point mutation at a different codon in the
avium infection in HIV patients (O’Brien et al., rpoB gene.
1987).
17.4.3 Clinical efficacy

17.4.2 Microbiological activity
In clinical trials investigating the clinical
RBT works similarly to RIF and RPT. In vitro, efficacy of RBT, there were no significant
the MIC against M. tuberculosis strain H37Rv differences between RBT and RIF in curing
is <0.015 μg/ml (2008a). In the mouse model, tuberculosis and preventing relapse. How-
RBT was clearly efficient, comparable with ever, higher doses of RBT may be associated
RPT, though pharmacokinetically the Cmax with more adverse effects and discontinuation.
RBT is a considerably weaker inducer of the 17.5 Rifalazil

CYP450 system and therefore has less
influence on the concentration of other drugs. Recently, rifalazil (KRM-1648 or ABI-1648)
This makes it especially suitable for use in was developed by Kaneka Corp, Japan. Its
HIV. It can be combined better with anti- MIC for M. tuberculosis was considerably
retroviral drugs, especially with the class of lower than for other rifamycins (MIC 0.0125).
protease inhibitors (PIs). Consequently, RBT It had more potential than RIF. Unfortunately,
is now reserved specifically for use in HIV its development had to be suspended because
patients. In a recent Cochrane review, Davies of the side effects in clinical trials (Aristoff et
et al. conclude that ‘RBT containing regimens al., 2010).
perform as well as RIF-containing regimen …
There is no evidence currently to support the
replacement of RIF by RBT for the treatment 17.6 Conclusions and Looking
of new cases of tuberculosis on the basis of Forward
efficacy….’ (Davies et al., 2007).
It is important to engage with the question of
the optimal dose of RIF and RPT and, in
17.4.4 Pharmocokinetics addition, the potential of newer rifamycins.
The tuberculosis pandemic is still un-
RBT was very promising in terms of its PK controlled; there is a pressing need to improve
profile: a relatively high Cmax and AUC. The the treatment of tuberculosis, to render
Cmax/MIC ratio is lower than for RIF (7.5 patients smear and culture negative as
versus 67), but protein binding is less than quickly as possible. There is a need to shorten
RIF (85%) and RPT (98%) with 70%, and there the duration of treatment, ideally to 3 months
is a relatively long half-life. It accumulates or even less. In the future, treatment will be of
better in the cell than RIF. There is no food short duration, while some patients may be
effect, though bioavailability is only 20%. For identified with biomarkers that predict
an overview of the PK characteristics of RBT, failure or relapse. These patients can then
see Table 17.2 (2008a). receive a tailor-made treatment, probably of
longer duration.
First, in small-scale studies, high (20 mg/
17.4.5 Adverse events kg) doses of RIF have already been shown to
have bactericidal activity up to twice that of
The adverse events profile of RBT resembles the 600 mg (10 mg/kg) dose (Gosling et al.,
that of RIF: hepatitis, gastrointestinal symp- 2003; Diacon et al., 2007). In addition, RIF has
toms and, rarely, cardiavascular, respiratory a sterilizing effect, i.e. the capacity to kill the
and neurological events. Hypersensitivity remaining mycobacteria that undergo
reactions are also reported, as is the flu-like sporadic metabolism and that remain after
syndrome. Unusual RBT-specific toxicity the initial phase of treatment (Jindani et al.,
with myositis and uveitis is reported; this is 2003). This provides the most exciting reason
dose related (2008a). to address this issue: if higher doses of RIF
Table 17.2. PK parameters of rifabutin.

Volume
AUC distribution Clearance
Species Half-life (h) (mg/h/l) Cmax (μg/ml) (l/kg) (l/h/kg) PK methodology
Human 45 ± 17 – 375 ± 267 – 0.69 ± 0.32 300 mg single
(range 16–69) oral dose to
healthy
volunteers
228 M.J. Boeree
and RPT are safe and well tolerated, and kill derivatives in terms of in vitro MICs.
tubercle bacilli more rapidly, we may be able Unfortunately, the bactericidal activity was
to shorten tuberculosis treatment further. inferior to RIF. A promising development in
This will have an immediate effect on the search for better rifamycin derivatives is
completion and cure rates. The original dose the synthesis of crystal structures of inhibitors
has helped to reduce treatment duration of with RNAP, which in theory may block RNA
the multi-drug regimen to 6 months but has polymerase and will not be sensitive to point
failed to reduce the duration of treatment to 4 mutations in the rpoB gene (Ho et al., 2009).
months, as 5–15% of the patients experienced Consequently, resistance will not emerge
bacteriological relapses in the BMRC Study 4 through this pathway. In fact, it may be
trials (Fox et al., 1999). A further reduction possible to develop inhibitors that bind
should still be our goal. In addition, faster tighter to a resistant mutant than to a wild-
smear and culture conversion decreases the type RNAP (Barluenga et al., 2006). These are
size of the infectious pool of patients in the rifamycin derivatives with a C3/C4 tail (in
community and can reduce further trans- contrast with RIF and RPT, which have C3
mission of tuberculosis. Second, higher doses tails). RBT and rifalazil are examples of such
of RIF and RPT may expose the bacteria to derivatives. Rifalazil derivatives such as ABI-
AUC/MIC values that prevent the emergence 0418, 0299, 11331 and 0043, all from scientists
of resistance or, according to the concept of at ActivBiotics (Tucker, Georgia, USA), are
mutant prevention concentration (MPC), described in a series of patent applications
raise concentrations to a level that prevents (Murphy et al., 2007). Finally, researchers at
the emergence of RIF-resistant mutants, and Cumbre Pharmaceuticals (Dallas, Texas,
thus risk of the emergence of MDR-TB. Third, USA) have prepared novel rifamycin–
there is a recent report that higher-dose RIF quinolone hybrids (the lead component being
may have a role in infections caused by low- CBR-2092) that are very active against S.
level resistant strains (i.e. MICs of 1–2 mg/l, aureus (Robertson et al., 2008).
just above the 1 mg/l breakpoint concen-
tration). Such resistance may be overcome by
applying higher doses of RIF (van Ingen et al., References
2011a). If higher doses prove to be more
efficacious and tolerable, a new regimen (1972) Controlled clinical trial of short-course
could be implemented quickly. Finally, (6-month) regimens of chemotherapy for treat-
financial restrictions are no longer warranted; ment of pulmonary tuberculosis. The Lancet
the price of a complete treatment regimen for 2(7760), 1079–1085.
a single patient, based on the 2HRZE4HR (1973) Editorial: Rifampicin or ethambutol in the
regimen, is now US$38.15 through the Stop routine treatment of tuberculosis. British Medical
Journal 4(5892), 568.
TB Partnerships’ Global Drug Facility. This
(2008a) Rifabutin. Tuberculosis (Edinburgh) 88(2),
price equals the price of 1 week of RIF 145–147.
monotherapy in 1973. (2008b) Rifalazil. Tuberculosis (Edinburgh) 88(2),
Since the development of the rifamycins 148–150.
in the 1950s and 1960s, and especially in the (2008c) Rifampin. Tuberculosis (Edinburgh) 88(2),
last decades, there are several groups who 151–154.
have developed new rifamycin derivatives (2008d) Rifapentine. Tuberculosis (Edinburgh)
with potentially better properties (Aristoff et 88(2), 155–158.
al., 2010). Ideally, new rifamycins need to be Acocella, G., Pagani, V., Marchetti, M., Baroni, G.C.
able to: (i) shorten treatment; (ii) be effective and Nicolis, F.B. (1971) Kinetic studies on
rifampicin. I. Serum concentration analysis in
against MDR-TB; and (iii) be effective against
subjects treated with different oral doses over a
latent tuberculosis. Researchers have period of two weeks. Chemotherapy 16(6), 356–
expanded on the structure of RIF. On the 370.
basis of such studies, RPT was developed; Angel, J.H. (1977) Short-course chemotherapy in
later azinomethylrifamycins such as FCE pulmonary tuberculosis. Journal of Antimicrobial
22250 and rifametane were promising Chemotherapy 3(4), 290–294.
Arioli, V., Berti, M., Carniti, G., Randisi, E., Rossi, E. Journal of Respiratory and Critical Care
and Scotti, R. (1981) Antibacterial activity of DL Medicine 161(5), 1572–1577.
473, a new semisynthetic rifamycin derivative. Davidson, P.T., Goble, M. and Lester, W. (1972) The
Journal of Antibiotics (Tokyo) 34(8), 1026–1032. antituberculosis efficacy of rifampin in 136
Aristoff, P.A., Garcia, G.A., Kirchhoff, P.D. and Hollis patients. Chest 61(6), 574–578.
Showalter, H.D. (2010) Rifamycins – obstacles Davies, G., Cerri, S. and Richeldi, L. (2007)
and opportunities. Tuberculosis (Edinburgh) Rifabutin for treating pulmonary tuberculosis.
90(2), 94–118. Cochrane Database Systematic Reviews 4,
Barluenga, J., Aznar, F., Garcia, A.B., Cabal, M.P., CD005159.
Palacios, J.J. and Menendez, M.A. (2006) New De, M.G., Iodice, F. and Lappa, B. (1968)
rifabutin analogs: synthesis and biological Preliminary observations on the therapeutic
activity against Mycobacterium tuberculosis. effect of rifampycin associated with streptomycin
Bioorganic and Medicinal Chemistry Letters or isoniazid. Archivio di Tisiologia e delle Malatie
16(22), 5717–5722. dell’Apparato Respiratorio 23(5), 377–386.
Bhusal, Y., Shiohira, C.M. and Yamane, N. (2005) Diacon, A.H., Patientia, R.F., Venter, A., van
Determination of in vitro synergy when three Helden, P.D., Smith, P.J., McIlleron, H., et al.
antimicrobial agents are combined against (2007) Early bactericidal activity of high-dose
Mycobacterium tuberculosis. International rifampin in patients with pulmonary tuberculosis
Journal of Antimicrobial Agents 26(4), 292–297. evidenced by positive sputum smears. Anti-
Burman, W.J., Gallicano, K. and Peloquin, C. (2001) microbial Agents and Chemotherapy 51(8),
2994–2996.
Comparative pharmacokinetics and pharmaco-
Favez, G., Chiolero, R. and Willa, C. (1972)
dynamics of the rifamycin antibacterials. Clinical
Rifampin-isoniazid compared with streptomycin-
Pharmacokinetics 40(5), 327–341.
isoniazid in the original treatment of infectious
Chan, C.Y., Au-Yeang, C., Yew, W.W., Leung, C.C.
pulmonary tuberculosis. Results of a controlled
and Cheng, A.F. (2004) In vitro post-antibiotic
study. Chest 61(6), 583–586.
effects of rifapentine, isoniazid, and moxifloxacin
Fox, W., Ellard, G.A. and Mitchison, D.A. (1999)
against Mycobacterium tuberculosis. Anti-
Studies on the treatment of tuberculosis
microbial Agents and Chemotherapy 48(1),
undertaken by the British Medical Research
340–343.
Council tuberculosis units, 1946–1986, with
Chapuis, L., Ji, B., Truffot-Pernot, C., O’Brien, R.J.,
relevant subsequent publications. International
Raviglione, M.C. and Grosset, J.H. (1994) Journal of Tuberculosis and Lung Disease 3(10,
Preventive therapy of tuberculosis with Suppl 2), S231–S279.
rifapentine in immunocompetent and nude Furesz, S., Scotti, R., Pallanza, R. and Mapelli, E.
mice. American Journal of Respiratory and (1967) Rifampicin: a new rifamycin. 3.
Critical Care Medicine 150(5, Pt 1), 1355–1362. Absorption, distribution, and elimination in man.
Citron, K.M. (1972) Tuberculosis – chemotherapy. Arzneimittelforschung 17(5), 534–537.
British Medical Journal 1(5797), 426–428. Gao, X.F., Li, J., Yang, Z.W. and Li, Y.P. (2009)
Constans, P., Saint-Paul, M., Morin, Y., Bonnaud, G. Rifapentine vs. rifampicin for the treatment of
and Bariety, M. (1968) Rifampicin: initial study of pulmonary tuberculosis: a systematic review.
plasma levels during prolonged treatment of International Journal of Tuberculosis and Lung
pulmonary tuberculosis patients. Reviews in Disease 13(7), 810–819.
Tuberculosis and Pneumology (Paris) 32(8), Gosling, R.D., Heifets, L. and Gillespie, S.H. (2003)
991–1006. A multicentre comparison of a novel surrogate
Constans, P., Baron, A., Parrot, R. and Coury, C. marker for determining the specific potency of
(1972) A study of 200 cases of active, recent anti-tuberculosis drugs. Journal of Antimicrobial
pulmonary tuberculosis treated with rifampin- Chemotherapy 52(3), 473–476.
isoniazid. A follow-up history of one and one- Goutelle, S., Bourguignon, L., Maire, P.H., Van,
half to three years. Chest 61(6), 539–549. G.M., Conte, J.E. Jr and Jelliffe, R.W. (2009)
Corpe, R.F. and Sanchez, E.S. (1972) Rifampin in Population modeling and Monte Carlo
initial treatment of advanced pulmonary simulation study of the pharmacokinetics and
tuberculosis. Chest 61(6), 564–573. antituberculosis pharmacodynamics of rifampin
Daniel, N., Lounis, N., Ji, B., O’Brien, R.J., Vernon, in lungs. Antimicrobial Agents and Chemo-
A., Geiter, L.J., et al. (2000) Antituberculosis therapy 53(7), 2974–2981.
activity of once-weekly rifapentine-containing Gumbo, T., Louie, A., Deziel, M.R., Liu, W., Parsons,
regimens in mice. Long-term effectiveness with L.M., Salfinger, M., et al. (2007) Concentration-
6- and 8-month treatment regimens. American dependent Mycobacterium tuberculosis killing
230 M.J. Boeree
and prevention of resistance by rifampin. McConville, J.H. and Rapoport, M.I. (1976)
Antimicrobial Agents and Chemotherapy Tuberculosis management in the mid-1970s.
51(11), 3781–3788. Journal of the American Medical Association
Gyselen, A., Verbist, L., Cosemans, J., Lacquet, 235(2), 172–176.
L.M. and Vandenbergh, E. (1968) Rifampin and Maggi, N., Pasqualucci, C.R., Ballotta, R. and
ethambutol in the retreatment of advanced Sensi, P. (1966) Rifampicin: a new orally active
pulmonary tuberculosis. American Review of rifamycin. Chemotherapy 11(5), 285–292.
Respiratory Diseases 98(6), 933–941. Marsili, L., Pasqualucci, C.R., Vigevani, A., Gioia,
Ho, M.X., Hudson, B.P., Das, K., Arnold, E. and B., Schioppacassi, G. and Oronzo, G. (1981)
Ebright, R.H. (2009) Structures of RNA New rifamycins modified at positions 3 and 4.
polymerase-antibiotic complexes. Current Synthesis, structure and biological evaluation.
Opinion in Structural Biology 19(6), 715–723. Journal of Antibiotics (Tokyo) 34(8), 1033–1038.
Houk, A.N. (1972) Rifampin: its role in the treatment Mitchison, D.A. (2000) Role of individual drugs in
of tuberculosis. Chest 61(6), 518–519. the chemotherapy of tuberculosis. International
Huitric, E., Werngren, J., Jureen, P. and Hoffner, S. Journal of Tuberculosis and Lung Disease 4(9),
(2006) Resistance levels and rpoB gene 796–806.
mutations among in vitro-selected rifampin- Murphy, C.K., Karginova, E., Sahm, D. and
resistant Mycobacterium tuberculosis mutants. Rothstein, D.M. (2007) In vitro activity of novel
Antimicrobial Agents and Chemotherapy 50(8), rifamycins against gram-positive clinical
2860–2862. isolates. Journal of Antibiotics (Tokyo) 60(9),
Jayaram, R., Gaonkar, S., Kaur, P., Suresh, B.L., 572–576.
Mahesh, B.N., Jayashree, R., et al. (2003) Newman, R., Doster, B., Murray, F.J. and Ferebee,
Pharmacokinetics–pharmacodynamics of S. (1971) Rifampin in initial treatment of
rifampin in an aerosol infection model of pulmonary tuberculosis. A U.S. Public Health
tuberculosis. Antimicrobial Agents and Chemo- Service tuberculosis therapy trial. American
therapy 47(7), 2118–2124. Review of Respiratory Diseases 103(4), 461–
Ji, B., Truffot-Pernot, C., Lacroix, C., Raviglione, 476.
M.C., O’Brien, R.J., Olliaro, P., et al. (1993) Nitti, V., Catena, E., Delli, V.F., De, M.G. and Marra,
Effectiveness of rifampin, rifabutin, and A. (1971) Rifampin in association with isoniazid,
rifapentine for preventive therapy of tuberculosis streptomycin, and ethambutol, respectively, in
in mice. American Review of Respiratory the initial treatment of pulmonary tuberculosis.
Diseases 148(6, Pt 1), 1541–1546. American Review of Respiratory Diseases
Jindani, A., Dore, C.J. and Mitchison, D.A. (2003) 103(3), 329–337.
Bactericidal and sterilizing activities of anti- Nitti, V., Delli, V.F., Ninni, A. and Meola, G. (1972)
tuberculosis drugs during the first 14 days. Rifampicin blood serum levels and half-life
American Journal of Respiratory and Critical during prolonged administration in tuberculous
Care Medicine 167(10), 1348–1354. patients. Chemotherapy 17(2), 121–129.
Kochar, D.K., Aseri, S., Sharma, B.V., Bumb, R.A., Nuermberger, E. and Grosset, J. (2004) Pharmaco-
Mehta, R.D. and Purohit, S.K. (2000) The role of kinetic and pharmacodynamic issues in the
rifampicin in the management of cutaneous treatment of mycobacterial infections. European
leishmaniasis. Quarterly Journal of Medicine Journal of Clinical Microbiology and Infectious
93(11), 733–737. Diseases 23(4), 243–255.
Kreis, B., Pretet, S., Birenbaum, J., Guibout, P., O’Brien, R.J. and Vernon, A.A. (1998) New
Hazeman, J.J., Orin, E., et al. (1976) Two three- tuberculosis drug development. How can we do
month treatment regimens for pulmonary better? American Journal of Respiratory and
tuberculosis. Bulletin of the International Union Critical Care Medicine 157(6, Pt 1), 1705–1707.
Against Tuberculosis 51(1), 71–75. O’Brien, R.J., Lyle, M.A. and Snider, D.E. Jr (1987)
Lester, W. (1972) Rifampin: a semisynthetic Rifabutin (ansamycin LM 427): a new
derivative of rifamycin – a prototype for the rifamycin-S derivative for the treatment of
future. Annual Reviews of Microbiololgy 26, mycobacterial diseases. Reviews in Infectious
85–102. Diseases 9(3), 519–530.
Long, M.W., Snider, D.E. Jr and Farer, L.S. (1979) Pallanza, R., Arioli, V., Furesz, S. and Bolzoni, G.
U.S. Public Health Service Cooperative trial of (1967) Rifampicin: a new rifamycin. II.
three rifampin–isoniazid regimens in treatment Laboratory studies on the antituberculous
of pulmonary tuberculosis. American Review of activity and preliminary clinical observations.
Respiratory Diseases 119(6), 879–894. Arzneimittelforschung 17(5), 529–534.
Peloquin, C.A. (2001) Pharmacological issues in Grupo de Estudio de Castilla-la Mancha de

the treatment of tuberculosis. Annals of the New Enfermedades Infecciosas. Antimicrobial
York Academy of Science 953, 157–164. Agents and Chemotherapy 39(9), 2061–2067.
Peloquin, C. (2003) What is the ‘right’ dose of Sterling, T.R., Villarino, M.E., Borisov, A.S., Shang,
rifampin? International Journal of Tuberculosis N., Gordin, F., Bliven-Sizemore, E., et al. (2011)
and Lung Disease 7(1), 3–5. Three months of rifapentine and isoniazid for
Poole, G., Stradling, P. and Worlledge, S. (1971) latent tuberculosis infection. New England
Potentially serious side-effects of high-dose Journal of Medicine 365(23), 2155–2166.
twice-weekly rifampicin. Postgraduate Medical van Ingen, J., Aarnoutse, R., de Vries, G., Boeree,
Journal 47(553), 727–747. M.J. and van Soolingen, D. (2011a) Low-level
Rastogi, N., Labrousse, V. and Goh, K.S. (1996) In rifampicin-resistant Mycobacterium tuberculosis
vitro activities of fourteen antimicrobial agents strains raise a new therapeutic challenge.
against drug susceptible and resistant clinical International Journal of Tuberculosis and Lung
isolates of Mycobacterium tuberculosis and Disease 15(7), 990–992.
comparative intracellular activities against the van Ingen, J., Aarnoutse, R.E., Donald, P.R.,
virulent H37Rv strain in human macrophages. Diacon, A.H., Dawson, R., Plemper van, B.G.,
Current Microbiology 33(3), 167–175. et al. (2011b) Why do we use 600 mg of
Robertson, G.T., Bonventre, E.J., Doyle, T.B., Du, rifampicin in tuberculosis treatment? Clinical
Q., Duncan, L., Morris, T.W., et al. (2008) In vitro Infectious Diseases 52(9), e194–e199.
evaluation of CBR-2092, a novel rifamycin- Verbist, L. and Gyselen, A. (1968) Antituberculous
quinolone hybrid antibiotic: microbiology profil- activity of rifampin in vitro and in vivo and the
ing studies with staphylococci and streptococci. concentrations attained in human blood.
Antimicrobial Agents and Chemotherapy 52(7), American Review of Respiratory Diseases
2324–2334. 98(6), 923–932.
Robinson, D. (1978) Treatment of tuberculosis. Veziris, N., Lounis, N., Chauffour, A., Truffot-Pernot,
British Medical Journal 1(6119), 1053–1054. C. and Jarlier, V. (2005) Efficient intermittent
Rosenthal, I.M., Williams, K., Tyagi, S., Peloquin, rifapentine-moxifloxacin-containing short-course
C.A., Vernon, A.A., Bishai, W.R., et al. (2006) regimen for treatment of tuberculosis in mice.
Potent twice-weekly rifapentine-containing regi- Antimicrobial Agents and Chemotherapy
mens in murine tuberculosis. American Journal 49(10), 4015–4019.
of Respiratory and Critical Care Medicine Weiner, M., Bock, N., Peloquin, C.A., Burman, W.J.,
174(1), 94–101. Khan, A., Vernon, A., et al. (2004) Pharmaco-
Rosenthal, I.M., Zhang, M., Williams, K.N., kinetics of rifapentine at 600, 900, and 1,200 mg
Peloquin, C.A., Tyagi, S., Vernon, A.A., et al. during once-weekly tuberculosis therapy.
(2007) Daily dosing of rifapentine cures American Journal of Respiratory and Critical
tuberculosis in three months or less in the Care Medicine 169(11), 1191–1197.
murine model. PLoS Medicine 4(12), e344. World Health Organization, S.T.D. (2009) Treatment
Ruslami, R., Nijland, H., Aarnoutse, R., Alisjahbana, of Tuberculosis: Guidelines, 4th edn. WHO,
B., Soeroto, A.Y., Ewalds, S., et al. (2006) Geneva, Switzerland.
Evaluation of high- versus standard-dose Ziakas, P.D. and Mylonakis, E. (2009) 4 months of
rifampin in Indonesian patients with pulmonary rifampin compared with 9 months of isoniazid
tuberculosis. Antimicrobial Agents and Chemo- for the management of latent tuberculosis
therapy 50(2), 822–823. infection: a meta-analysis and cost-effectiveness
Sensi, P. (1983) History of the development of study that focuses on compliance and liver
rifampin. Reviews of Infectious Diseases 5 toxicity. Clinical Infectious Diseases 49(12),
(Suppl 3), S402–S406. 1883–1889.
Sensi, P., Margalith, P. and Timbal, M.T. (1959) Ziglam, H.M., Baldwin, D.R., Daniels, I., Andrew,
Rifomycin, a new antibiotic; preliminary report. J.M. and Finch, R.G. (2002) Rifampicin
Farmaco Sci 14(2), 146–147. concentrations in bronchial mucosa, epithelial
Solera, J., Rodriguez-Zapata, M., Geijo, P., Largo, lining fluid, alveolar macrophages and serum
J., Paulino, J., Saez, L., et al. (1995) following a single 600 mg oral dose in patients
Doxycycline-rifampin versus doxycycline- undergoing fibre-optic bronchoscopy. Journal
streptomycin in treatment of human brucellosis of Antimicrobial Chemotherapy 50(6), 1011–
due to Brucella melitensis. The GECMEI Group. 1015.
18 Therapy of the XDR-TB Patient with
Thioridazine – An Old Drug with New
Applications
Leonard Amaral,1,2,3 Marta Martins,1,2,4 Isabel Couto1,5 and

Miguel Viveiros1,3
1Unit of Mycobacteriology and 2UPMM, Instituto de Higiene e Medicina Tropical,
Universidade Nova de Lisboa (IHMT/UNL), Lisbon, Portugal; 3Cost Action BM0701

(ATENS) of Cost Action of the European Commission; 4UCD Centre for Food Safety,
School of Agriculture, Food Science and Veterinary Medicine, University College
Dublin, Ireland; 5Centro de Recursos Microbiológicos (CREM), Faculdade de
Ciências e Tecnologia, UNL, Caparica, Portugal
18.1 Introduction and the advent of HIV/AIDS in the early

1980s, were frequent, and wherever they
Pulmonary tuberculosis produced by the occurred, the incidence of new cases soared
human bacterial pathogen Mycobacterium (Amaral et al., 2010). In New York City during
tuberculosis has plagued modern man since he the early 1990s, it became clear that new cases
moved from foraging and gathering to of pulmonary tuberculosis had quadrupled,
becoming one dependent on agricultural and that more than half of these cases were
animal husbandry as sources of food. resistant to the two most effective anti-
Nevertheless, with the advent of antibiotics tuberculosis drugs, INH and RIF (Moss et al.,
such as isoniazid (INH) and rifampicin (RIF), 1997), and that the vast majority of these new
and their availability throughout the Western cases of multidrug-resistant M. tuberculosis
world, with few exceptions, it seemed (MDR-TB) were found in the migrant
possible that this essentially lethal infection population (Tornieporth et al., 1997). Rapid
could be eradicated (Martins et al., 2008b). measures were taken and within 5 years the
However, because the cost of these antibiotics incidence of new cases was decreased to
was beyond the means of most of the world’s levels below that of the 1950s (Driver et al.,
population, the number of new cases of 2007). These measures involved vast improve-
tuberculosis remained high in sub-Saharan ments in the manner by which the laboratory
Africa, South and Latin America, South-east identified M. tuberculosis in the sputum of
Asia, Eastern Europe, India and Portugal infected patients, the rapid culture of the
(Amaral et al., 2001a). Moreover, conditions products of sputum in liquid media and
which predispose the spread of the infection, antibiotic susceptibility to at least four anti-
such as war, poverty, overcrowding, famine biotics commonly employed for the therapy

Therapy of the XDR-TB Patient with Thioridazine 233
of MDR-TB, using such systems as the antibiotic target may take place. However, it
BACTEC 460TB System (Salfinger and is well known that in M. tuberculosis,
Pfyffer, 1994). The establishment of a fast- chromosomal mutations can be responsible
track programme that placed the sputum for the resistance seen in several strains and
specimen at the highest level of priority for that the MDR phenotype is due to the
subsequent laboratory examinations and the accumulation of random mutations in the
creation of directly observable therapy (DOT) genes involved in the resistance to the most
that ensured the patient would indeed take effective anti-tuberculous drugs (INH and
the medication were the landmark pro- RIF). Resistance to RIF is attributed to point
grammes that yielded success (Salfinger and mutations in the rpoB gene and is almost
Pfyffer, 1994). But this was not to last, for always associated with resistance to INH
within a few years, clinical isolates that were (Viveiros et al., 2005; Migliori et al., 2010). The
resistant to INH and RIF, to any fluoro- main mechanism of resistance to INH is due
quinolone and to at least one of the injectable to mutations in the KatG gene (Migliori et al.,
anti-M. tuberculosis drugs (capreomycin, kana- 2009, 2010). The rates of mutation in these
mycin and amikacin) began to appear with genes contribute to approximately 75% of the
increasing frequency (Migliori et al., 2009). cases of resistance to INH obtained in the
These extensively drug-resistant strains were clinical setting (Guo et al., 2006).
termed XDR-TB, and the therapy of the Therapy consisting of the antibiotic to
XDR-TB patient is far more problematic than which the strain is now resistant promotes
that posed by infections produced by selection of that strain if the effectiveness of
MDR-TB (London, 2009). In areas of the USA other antibiotics is reduced due to non-
where the frequency of HIV/AIDS is high, co- compliance (Lipsitch and Levin, 1998), poor
infection with XDR-TB strains and HIV result management of the patient (Victor et al., 2002)
in mortality within a year of diagnosis, or establishment of mutations that contribute
regardless of the therapy employed (Chan to the drug-resistance mechanisms in the
and Iseman, 2008). At the time of writing, strains.
there is no effective anti-XDR-TB standardized The selected strain may undergo further
treatment, unless we consider alternative mutation that renders the strain resistant to
therapies such as thioridazine (TZ), an another antibiotic within the same infected
inexpensive drug used for over 40 years for patient (Victor et al., 2002), or within another
the therapy of psychosis and as safe as any patient who was infected by the first mutated
other more current neuroleptic (Thanacoody, strain (Iademarco and Castro, 2003). Because
2007). It is the intent of this chapter to provide the two most effective anti-M. tuberculosis
the rationale and experimental basis for its antibiotics are INH and RIF, resistance to
consideration as a new effective anti-XDR-TB these two antibiotics (MDR-TB) from serial
drug. mutations constitutes the major threat for
successful therapy. Further serial mutations
result in further resistance to other antibiotics,
18.2 Development of Antibiotic such that today, XDR-TB strains are emerging
Resistance in Mycobacterium globally (Migliori et al., 2010). The dis-
tuberculosis semination of these highly drug-resistant
strains is of obvious concern. The need for
The development of resistance to antibiotics effective anti-MDR- and XDR-TB agents is
used for the therapy of tuberculosis may take urgent.
place via a variety of ways. Because of the However, resistance to one antibiotic
extraordinary length of the therapy period, may not involve a mutation. As early as 2001,
involving a minimum of 2–3 months for experimentally induced resistance of INH
patients whose tuberculin or purified protein susceptible M. tuberculosis to INH could be
derivative (PPD) skin test converted from brought about by prolonged exposure of the
negative to positive, the spontaneous strain to increasing concentrations of INH
mutation occurred in a gene for a given (Viveiros et al., 2002). Because this induced
234 L. Amaral et al.
resistance could be eliminated by transfer to seminate to other sites of the body once it
drug-free medium or by the addition of escapes its macrophage prison (Mogga et al.,
known efflux pump inhibitors, the putative 2002; El-Masry et al., 2007; Yoshida et al.,
presence of an overexpressed efflux pump 2009). Prior to escape, the organism lays
that extruded the INH prodrug, or its prodrug dormant for many years, if not decades.
product, prior to reaching its intended target Diagnosis may be made with a simple skin
seemed plausible. Since that demonstration, test (PPD) in countries that do not immunize
others have shown that the overexpression of routinely and boost the young with the
given efflux pumps in mycobacteria render Bacillus Calmette–Guérin (BCG). This test,
resistance to given antibiotics such as INH, also known as the Mantoux test, is now
clarithromycin, RIF and clofazimine (Ramón- standardized by the WHO and continues to
García et al., 2009; Rodrigues et al., 2009). be used as the standard method for detecting
The role of efflux pumps in MDR-TB and latent infection with M. tuberculosis. It is
XDR-TB is of obvious interest and, at the time widely used to support clinical and radio-
of writing, is pursued vigorously. Agents that logical findings in the evaluation of patients
can inhibit the efflux pumps responsible for with suspected tuberculosis. A positive result
antibiotic resistance are needed for possible can help in the decision to start treatment
adjuvant use. while bacteriological confirmation is awaited
Dormant M. tuberculosis strains commonly or lacking. The test uses PPD, a combination
reside in macrophages of the lung and are of mycobacterial antigens obtained from M.
considered to be in a non-replicating stage tuberculosis culture and which share a large
(Ehlers, 2009). Dormant M. tuberculosis can be number of antigens, both with BCG and with
induced readily from the incubation of environmental mycobacteria. It consists of an
antibiotic-susceptible M. tuberculosis in vitro, intradermal injection of 0.1 ml of tuberculin
by the addition of oxygen-coupling agents (100 units/ml). This combination elicits
into the medium and incubation of the culture delayed-type hypersensitivity response or
in the absence of oxygen (Sohaskey, 2008). type IV hypersensitivity reaction that is
Because dormant M. tuberculosis are resistant mediated by specific T lymphocytes. Such
to most antibiotics (Sohaskey, 2008) and effector cells function in essentially the same
because they may be the prominent strains way as during a response to an infectious
that infect the human but do not produce an agent. When small amounts of PPD are
active infection until many decades later injected, a T cell-mediated local inflammatory
(reactivated tuberculosis) (Wayne and reaction evolves in individuals who have
Sohaskey, 2001), therapy with INH and RIF of previously responded to M. tuberculosis. This
a patient who has sero-converted to a positive indicates the presence of antibodies or
PPD test may not provide a cure and, lymphocytes that are specific for PPD. This
therefore, the possibility of reactivated cell-mediated immunity can then be detected
tuberculosis in later life is a possibility. Agents as a local response when the individual’s skin
that have activity against dormant M. is injected with a small amount of PPD. The
tuberculosis are therefore of great interest for response typically appears a day or two after
use whenever PPD sero-conversion has taken the injection and consists of a raised, red and
place. hard (or indurated) area. This induration can
be measured 2–7 days afterwards and
disappears as the PPD is degraded.
18.3 Criteria for an Effective Anti- Vaccination with BCG can result in a
MDR/XDR-TB, Anti-latent positive PPD test and even though many
M. tuberculosis Agent claim that the size of indurations is related to
a recent infection by M. tuberculosis of a BCG-
Pulmonary tuberculosis is essentially an intra- immunized subject, this claim is not faithfully
cellular infection of the pulmonary macro- realized. The poor sensitivity of the Mantoux
phage of the alveolar sac caused by a strain of test in young children and immunosuppressed
M. tuberculosis with the capacity to dis- people makes it impossible to interpret
negative results in these groups. Moreover, To do this, the agent must first be shown to
up to 25% of immunocompetent adults with have in vitro activity against all forms of M.
active tuberculosis may have a negative or tuberculosis, it must traverse the plasma
ambiguous result (Gooding et al., 2007). membrane of the macrophage, reach the
Regardless, until the organism breaks phagosome (vacuole formed by the fusion of
free of its macrophage prison, the infected the cell membrane around the phagocytosed
patient is not infectious and therapy is bacteria) or phagolysosome (cellular compart-
desirable only to obviate the progression of ment that results from the fusion of the
latent infection to active disease – the phagosome with lysosomes during their
infectious phase of tuberculosis. This stage is maturation process) where the organism
accompanied by the presence of acid-fast resides and retain its proven in vitro activity
stained bacteria present in the sputum of the against the organism in situ. Is there such a
subject who is presenting with some or all of compound? Is there such a compound that
the symptoms of active disease. Confirmation exists and which has been used relatively
that the sputum does indeed contain M. safely for the past 40 years for the therapy of
tuberculosis is made from culture and identifi- a non-infectious pathology? Is this agent
cation, or by direct identification with the use ready for clinical trials at sites of the globe
of specific molecular probes or PCR-type where tuberculosis, MDR, XDR and their
tests (Parsons et al., 2004; Viveiros et al., 2005). latent forms produce a high prevalence of
If the laboratory and radiological tests pulmonary tuberculosis infections? The
support the diagnosis of a pulmonary answer is yes.
tuberculosis infection and the responsible
strain is susceptible to INH and RIF, the
patient is treated with these antibiotics. This 18.4 Phenothiazines and Their
therapy may apparently cure the patient of Anti-tuberculosis Activity
overt symptoms related to the destruction of
pulmonary tissue, such as the immune Phenothiazines have a long history of
responses that result in inflammation and antimicrobial activity that began with studies
physical symptoms such as night sweats, loss conducted by Paul Ehrlich in the late 19th
of appetite, loss of weight, morbidity, respira- century (Kristiansen and Amaral, 1997;
tory stress, due to killing of intracellular Amaral and Kristiansen, 2000). However,
bacteria that have begun their replicative because the first phenothiazine studied was
phase and, hence, are susceptible to INH and methylene blue, and this turned patients
RIF. However, those that are still in the blue, it was only after its colourless derivative,
dormant phase are not affected and, therefore, chlorpromazine (CPZ), was introduced in
when therapy is deemed ‘complete’, and 1957 for the therapy of psychosis that its anti-
ended, the recurrence of symptoms may take mycobacteriological properties were noted
place due to the progression of the intra- (Pleasure, 1956; Alcozer and Lingiardi, 1957;
cellular strain from the quiescent to the Hyvert et al., 1957; Marchand and Reuter,
replicative state, and destruction of pulmon- 1957; Shubin et al., 1957; Zamfir and Ionesco,
ary tissue and the extracellular phase begins 1958; Santopadre and Silanos, 1959; Filippov,
anew. If the infection is caused by MDR or 1960; Hollister et al., 1960; Amaral et al.,
XDR M. tuberculosis, therapy is problematic, 2001a). However, because therapy of tubercu-
and even for those few successful cases, a real losis was very successful with IHN and RIF,
cure is not probable due to the presence of and CPZ produced a series of nasty side
dormant mycobacteria. effects (Amaral et al., 2001b), interest in CPZ
The criterion that an agent must satisfy as an anti-tuberculosis agent was limited to
to be a truly successful anti-tuberculosis drug intellectual curiosity (Molnár et al., 1977;
that provides a cure for antibiotic-susceptible Kristiansen and Vergmann, 1986). However,
MDR, XDR and latent M. tuberculosis is to with the global resurgence of tuberculosis
have activity against these forms of M. during the 1980s and the emergence of
tuberculosis when the organism is intracellular. MDR-TB, especially at the beginning of the
1990s, and its problematic therapy, attention known as post-marketing surveillance trials
was now being paid to the possibility that and involve the safety surveillance, namely
CPZ could have a special role for therapy of pharmacovigilance, and ongoing technical
MDR-TB. That the anti-mycobacterial activity support of a drug after it receives permission
of CPZ took place at concentrations that were to be sold. Phase IV clinical trials may be
extremely high (minimum inhibitory concen- required by regulatory authorities, or alter-
trations (MICs) in excess of 20–30 mg/l) and natively may be undertaken by the sponsoring
clinically irrelevant (maximum plasma con- companies (for example, in the case that the
centration that can be achieved safely is c.0.4– drug has not been tested for interactions with
0.5 mg/l), meant the use of CPZ for therapy other drugs). In this case, safety surveillance
was out of the question. However, resurgence is designed to detect any rare or long-term
in the interest of CPZ took place momentarily adverse effects over a much larger patient
with the demonstration that CPZ could population and longer period than was
promote the killing of intracellular M. tubercu- possible during the Phase I–III clinical trials.
losis when its concentration in the macrophage Harmful effects identified during Phase IV
was lower than that present in the plasma of a trials may result in a drug no longer being
CPZ chronically treated patient (Amaral et al., sold, or restricted to certain uses.
2001b). Nevertheless, the problems posed by
the serious side effects produced by CPZ
presented an insurmountable barrier for its
18.4.1 Targeting the macrophage: a
use. The demonstration that a derivative of
unique concept for the therapy of
this compound, TZ, the equal to CPZ for
MDR- and XDR-TB infections
therapy of psychosis but far less problematic
with respect to serious side effects, could
The mechanism by which TZ enhances the
inhibit the in vitro replication of XDR-TB
killing of intracellular mycobacteria appears
strains (Amaral et al., 1996) led to the con-
to be one related to the inhibition of the Ca2+/
sideration of a phenothiazine for use, albeit
K+ transport of the macrophage (Martins et
limited only by the fact that its in vitro activity
al., 2008a). The sequence of events by which
took place with concentrations that were well
this mechanism is believed to take place is as
beyond those achievable in the patient. This
follows:
latter limitation was eliminated with the
demonstration that CPZ and TZ could 1. The mycobacterium binds to the plasma
enhance the killing of intracellular antibiotic- membrane of the macrophage, thereby acti-
susceptible and antibiotic-resistant strains vating the phagocytosis process (García et al.,
(MDR- and XDR-TB) at concentrations that 2005; Plaza et al., 2007; Chapeton-Montes et
were well below those present in the plasma al., 2008; Link et al., 2010). This process results
of patients treated initially with TZ (Crowle in the invagination of the plasma membrane,
et al., 1992; Ordway et al., 2003; Amaral et al., resulting in a vacuole that contains the myco-
2007; Martins et al., 2007a). Demonstration of bacterium; the phagosome. It is important to
the ability of TZ to cure mice infected with M. note that the plasma membrane contains
tuberculosis soon followed (Martins et al., Ca2+/K+ transporters that bind Ca2+/K+ and
2007b; van Ingen et al., 2009). The ability of TZ pump it into the cytoplasm of the cell against
to effect complete cures of 10 out of 12 a gradient. The invaginated plasma mem-
XDR-TB patients when used as an adjuvant to brane therefore contains Ca2+/K+ transport-
therapy with three antibiotics, as previously ers. However, after the internalization of the
recommended (Bettencourt et al., 2000), was bacteria, they now pump Ca2+/K+ from the
demonstrated in 2007 by Eduardo Abbate phagosome to the cytoplasm of the cell.
and his group (Abbate et al., 2007). Since that 2. The phagosome and the lysosome will
time, a global call for clinical trials for therapy fuse eventually, but because Ca2+/K+ is
of XDR-TB with TZ has been made (Amaral et pumped out from the phagolysosome, the
al., 2010) and has resulted in a number of acidification of the phagolysosome needed
Phase IV studies in India. These trials are also for the activation of hydrolases does not take
place (Reeves et al., 2002; Ahluwalia et al., 18.4.2 In vitro activity of TZ against
2004; Segal, 2005). This could be one of the induced latent M. tuberculosis
reasons why the pulmonary macrophage
does not kill phagocytosed mycobacteria. TZ has been shown to kill in vitro non-
3. TZ (Eilam, 1983; Kongsamut et al., 2002), replicating M. tuberculosis as well as induced
as well as other phenothiazines (Moriyama et dormant M. tuberculosis (Sohaskey, 2008).
al., 1993; Chattopadhyay et al., 1998; Sampaio- This suggests that if the agent is equally
Maia et al., 2001; Choi et al., 2005), is known effective against these forms of M. tuberculosis
to inhibit the Ca2+/K+ transporters (Traykov et when the organism is intracellular, then the
al., 1997). use of TZ for therapy of an infection that had
4. TZ, as well as other neuroleptic phenothia- been demonstrated decades earlier, and
zines, is concentrated by macrophages that which is probably latent, has the potential to
are rich in vesicles (Wójcikowski and Daniel, reduce, if not obviate, reactivated tubercu-
2002; Daniel, 2003). This process takes place losis.
via pinocytosis and results in a TZ-containing
vesicle formed from the invagination of the
plasma membrane. 18.4.3 TZ as an inhibitor of efflux pumps
5. The phenothiazine is concentrated by the in antibiotic resistance in M. tuberculosis
macrophage vesicles (Daniel and Wójcikowski,
1999; Daniel et al., 2001). INH-susceptible M. tuberculosis can be
6. The fusion of the phagosome and the induced readily to high-level resistance to
TZ-containing vesicle takes place. this antibiotic (Viveiros et al., 2002). The
7. TZ inhibits the Ca2+/K+ flux of the phago- induced resistance is due to the overexpression
lysosome and therefore the energy needed of efflux pumps of the organism that extrudes
by the Ca2+/K+ transporter for pumping INH prior to either its conversion to the active
out Ca2+/K+ from the phagolysosome is obvi- drug by peroxidases or prior to the active
ated. drug reaching its intended target (Viveiros et
8. The build-up of Ca2+/K+ takes place, with al., 2003). A number of efflux pumps of M.
cytosolic homeostasis mechanisms activated, tuberculosis have been characterized (Aínsa et
leading to an increased activity of V-ATPases al., 1998; Choudhuri et al., 1999; De Rossi et al.,
that will acidify the phagolysosome, and 2002). Moreover, these efflux pumps appear
subsequently the activation of the hydrolases to bestow antibiotic resistance to the organism
takes place and the mycobacterium is finally (De Rossi et al., 2002, 2006; De La Iglesia et al.,
degraded and killed. 2006; Gupta et al., 2006, 2010; Escribano et al.,
The mechanism by which a phenothiazine 2007; Gumbo et al., 2007; Jiang et al., 2008;
enhances the killing of intracellular myco- Ramón-García et al., 2008a). TZ has been
bacteria invokes a totally new concept for the shown to inhibit efflux pumps of mycobacteria
therapy of a pulmonary tuberculosis infec- (Viveiros et al., 2002; Amaral et al., 2008;
tion, as well as for other bacterial infections Rodrigues et al., 2009) and so therapy of MDR
of the non-killing pulmonary macrophage. infections caused by M. tuberculosis that over-
This concept targets the macrophage rather express efflux pumps is promising.
than the intracellular bacterium (Amaral
et al., 2008; Martins et al., 2008b, 2009; Amaral
and Molnar, 2010) and so bypasses any 18.4.4 Compassionate therapy of the
mutational response directed by the XDR-TB patient with TZ
organism against TZ. One would therefore
expect that the use of TZ for the therapy of Therapy of antibiotic-susceptible infections
MDR- and XDR-TB infections would not with TZ is not recommended at this time.
suffer the expected consequences of resist- Rather, therapy should be restricted to
ance that result from exposure of the selected cases of MDR- and XDR-TB that have
mycobacterium to the agent. not responded to therapy and whose prog-
nosis is poor (Amaral et al., 2010). At the time pumps are inactivated. CPZ and TZ have
of writing, 10 XDR-TB patients in Mumbai, been shown to be effective efflux pump
India, who were deemed terminal, were inhibitors with clinically relevant activity
selected for monotherapy with TZ. Prior to against mycobacteria. These compounds are
therapy, the patients were monitored for ideal candidates that can be tested in
cardiac functions for at least 4 consecutive combination with conventional antibiotics.
days, for the first 2 weeks of therapy and Combined therapies will decrease the critical
periodically thereafter. Monitoring of cardiac concentration of antibiotics in vitro and in
functions is mandatory since TZ may, on rare infected human macrophages, thereby
occasions, produce a prolongation of the QT reducing toxicity, one of the main problems
interval, which, if significant, can cause death. associated with clinically significant levels of
The therapeutic protocol consisted of a daily these drugs. Since they can transform the
dose of 25 mg/day for 1 week, followed by non-killing macrophage into an effective
weekly increments of 25 mg/day until a daily killer, they provide an exciting new approach
dose of 75 mg/day was reached. Within 1 for the therapy of these MDR-TB infections
week, symptoms such as night sweats, loss of and may contribute to limiting the emergence
appetite, depression and anxiety were of XDR-TB.
obviated and patients began to gain weight.
The number of acid-fast positive organisms
in sputa was decreased. Acknowledgements
However, at the time of writing, 6 weeks
of therapy had not yet resulted in negative This work was supported by grants SDH.
cultures. It is still too early to tell if a cure is IC.I.01.17-TB Task Force for Greater Lisbon
under way. However, the quality of life of the (2001), TB-Fast-Track Programme (2004) and
patients could be improved significantly. XDRTB Early Detection in Greater Lisbon
(2009) from the Calouste Gulbenkian
Foundation. The work was also supported by
18.5 Conclusion the EU-FSE/FEDER-POCI/SAU-MMO/59370/
2004 and the EU-FSE/FEDER-PTDC/BIA-
The emergence of MDR- and XDR-TB MIC/71280/2006 grants provided by the
represents a major threat to public health Fundação para a Ciência e a Tecnologia (FCT)
worldwide, as many of these strains are of Portugal. M. Martins was a recipient of
untreatable with the current arsenal of drugs. grant SFRH/BD/14319/2003 provided by FCT
Treatment of these infections with existing (Portugal).
antibiotics is only marginally effective.
Nevertheless, they remain the only available
therapeutic option in the clinical setting. To References
design and implement new approaches that
can reduce the concentration of these drugs at Abbate, E., Vescovo, M., Natiello, M., Cufré, M.,
the site of infection while maintaining efficacy Garcia, A., Ambroggi, M., et al. (2007) Tubercu-
is a substantial challenge. It is now widely losis extensamente resistente (XDR-TB) en
accepted that the intrinsic drug resistance (i.e. Argentina: aspectos destacables epidemio-
low-level resistance) exhibited by many lógicos, bacteriológicos, terapéuticos y
eukaryotic and prokaryotic cells is due to the evolutivos. Revista Argentina de Medicina
maintenance of active efflux pumps, and the Respiratoria 1, 19–25.
Ahluwalia, J., Tinker, A., Clapp, L.H., Duchen, M.R.,
magnitude of this resistance can be reduced
Abramov, A.Y., Pope, S., et al. (2004) The large-
by the use of effective efflux pump inhibitors. conductance Ca2+-activated K+ channel is
Enhanced killing by the human macrophage essential for innate immunity. Nature 427, 853–
using compounds that can act as efflux pump 858.
inhibitors proceeds via a mechanism that Aínsa, J.A., Blokpoel, M.C., Otal, I., Young, D.B., De
promotes the acidification of the phago- Smet, K.A. and Martín, C. (1998) Molecular
lysosome wherein the eukaryotic Ca2+ and K+ cloning and characterization of Tap, a putative
multidrug efflux pump present in Mycobacterium bacterium tuberculosis. International Journal of

fortuitum and Mycobacterium tuberculosis. The Antimicrobial Agents 16, 69–71.
Journal of Bacteriology 180, 5836–5843. Chan, E.D. and Iseman, M.D. (2008) Multidrug-
Alcozer, G. and Lingiardi, G. (1957) Chlorpromazine resistant and extensively drug-resistant tubercu-
in therapy of pulmonary tuberculosis. Archivio losis: a review. Current Opinion in Infectious
‘E. Maragliano’ di Patologia e Clinica 13, 1247– Diseases 21, 587–595.
1264. Chapeton-Montes, J.A., Plaza, D.F., Curtidor, H.,
Amaral, L. and Kristiansen, J.E. (2000) Pheno- Forero, M., Vanegas, M., Patarroyo, M.E., et al.
thiazines: an alternative to conventional therapy (2008) Characterizing the Mycobacterium
for the initial management of suspected multi- tuberculosis Rv2707 protein and determining its
drug resistant tuberculosis. A call for studies. sequences which specifically bind to two human
International Journal of Antimicrobial Agents cell lines. Protein Science 17, 342–351.
14, 173–176. Chattopadhyay, D., Mukherjee, T., Pal, P., Saha, B.
Amaral, L. and Molnar, J. (2010) Therapy of XDRTB and Bhadra, R. (1998) Altered membrane
with thioridazine a drug beyond patent permeability as the basis of bactericidal action
protection but eligible for patent ‘As New Use’. of methdilazine. Journal of Antimicrobial
Recent Patents on Anti-infective Drug Discovery Chemotherapy 42, 83–86.
5, 109–114. Choi, S.Y., Koh, Y.S. and Jo, S.H. (2005) Inhibition
Amaral, L., Kristiansen, J.E., Abebe, L.S. and of human ether-a-go-go-related gene K+
Millett, W. (1996) Inhibition of the respiration of channel and IKr of guinea pig cardiomyocytes
multi-drug resistant clinical isolates of Myco- by antipsychotic drug trifluoperazine. Journal of
bacterium tuberculosis by thioridazine: potential Pharmacology and Experimental Therapeutics
use for initial therapy of freshly diagnosed 313, 888–895.
tuberculosis. Journal of Antimicrobial Chemo- Choudhuri, B.S., Sen, S. and Chakrabarti, P. (1999)
therapy 38, 1049–1053. Isoniazid accumulation in Mycobacterium
Amaral, L., Kristiansen, J.E., Viveiros, M. and smegmatis is modulated by proton motive force-
Atouguia, J. (2001a) Activity of phenothiazines driven and ATP-dependent extrusion systems.
against antibiotic-resistant Mycobacterium Biochemical and Biophysical Research Com-
tuberculosis: a review supporting further studies munications 256, 682–684.
that may elucidate the potential use of Crowle, A.J., Douvas, G.S. and May, M.H. (1992)
thioridazine as anti-tuberculosis therapy. Chlorpromazine: a drug potentially useful for
Journal of Antimicrobial Chemotherapy 47, treating mycobacterial infections. Chemotherapy
505–511. 38, 410–419.
Amaral, L., Viveiros, M. and Kristiansen, J.E. Daniel, W.A. (2003) Mechanisms of cellular
(2001b) Phenothiazines: potential alternatives distribution of psychotropic drugs. Significance
for the management of antibiotic resistant for drug action and interactions. Progress in
infections of tuberculosis and malaria in Neuro-Psychopharmacology and Biological
developing countries. Tropical Medicine and Psychiatry 27, 65–73.
International Health 6, 1016–1022. Daniel, W.A. and Wójcikowski, J. (1999) The role of
Amaral, L., Martins, M. and Viveiros, M. (2007) lysosomes in the cellular distribution of
Phenothiazines as anti-multi-drug resistant thioridazine and potential drug interactions.
tubercular agents. Infectious Disorders – Drug Toxicology and Applied Pharmacology 158,
Targets 7, 257–265. 115–124.
Amaral, L., Martins, M., Viveiros, M., Molnar, J. and Daniel, W.A., Wójcikowski, J. and Pałucha, A.
Kristiansen, J.E. (2008) Promising therapy of (2001) Intracellular distribution of psychotropic
XDR-TB/MDR-TB with thioridazine an inhibitor drugs in the grey and white matter of the brain:
of bacterial efflux pumps. Current Drug Targets the role of lysosomal trapping. British Journal of
9, 816–819. Pharmacology 134, 807–814.
Amaral, L., Boeree, M.J., Gillespie, S.H., Udwadia, Danilchanka, O., Mailaender, C. and Niederweis, M.
Z.F. and van Soolingen, D. (2010) Thioridazine (2008a) Identification of a novel multidrug efflux
cures extensively drug-resistant tuberculosis pump of Mycobacterium tuberculosis. Anti-
(XDR-TB) and the need for global trials is now! microbial Agents and Chemotherapy 52, 2503–
International Journal of Antimicrobial Agents 2511.
35, 524–526. Danilchanka, O., Pavlenok, M. and Niederweis, M.
Bettencourt, M.V., Bosne-David, S. and Amaral, L. (2008b) Role of porins for uptake of antibiotics
(2000) Comparative in vitro activity of pheno- by Mycobacterium smegmatis. Antimicrobial
thiazines against multidrug-resistant Myco- Agents and Chemotherapy 52, 3127–3134.
De La Iglesia, A.I. and Morbidoni, H.R. (2006) of the emergence of resistance, not depletion of
Mechanisms of action of and resistance to Mycobacterium tuberculosis in the log phase of
rifampicin and isoniazid in Mycobacterium growth. The Journal of Infectious Diseases 195,
tuberculosis: new information on old friends. 194–201.
Revista Argentina de Microbiología 38, 97–109. Guo, H., Seet, Q., Denkin, S., Parsons, L. and
De Rossi, E., Arrigo, P., Bellinzoni, M., Silva, P.A., Zhang, Y. (2006) Molecular characterization of
Martín, C., Aínsa, J.A., et al. (2002) The isoniazid-resistant clinical isolates of
multidrug transporters belonging to major Mycobacterium tuberculosis from the USA.
facilitator superfamily in Mycobacterium Journal of Medical Microbiology 55, 1527–1531.
tuberculosis. Molecular Medicine 8, 714–724. Gupta, A.K., Chauhan, D.S., Srivastava, K., Das,
De Rossi, E., Aínsa, J.A. and Riccardi, G. (2006) R., Batra, S., Mittal, M., et al. (2006) Estimation
Role of mycobacterial efflux transporters in of efflux mediated multi-drug resistance and its
drug resistance: an unresolved question. FEMS correlation with expression levels of two major
Microbiology Reviews 30, 36–52. efflux pumps in mycobacteria. The Journal of
Driver, C.R., Kreiswirth, B., Macaraig, M., Clark, C., Communicable Diseases 38, 246–254.
Munsiff, S.S., Driscoll, J., et al. (2007) Molecular Gupta, A.K., Katoch, V.M., Chauhan, D.S., Sharma,
epidemiology of tuberculosis after declining R., Singh, M., Venkatesan, K., et al. (2010)
incidence, New York City, 2001-2003. Epidemio- Microarray analysis of efflux pump genes in
logy and Infection 135, 634–643. multidrug-resistant Mycobacterium tuberculosis
Ehlers, S. (2009) Lazy, dynamic or minimally during stress induced by common anti-
recrudescent? On the elusive nature and tuberculous drugs. Microbial Drug Resistance
location of the mycobacterium responsible for 16, 21–28.
latent tuberculosis. Infection 37, 87–95. Hollister, Le., Eikenberry, Dt., and Raffel, S. (1960)
Eilam, Y. (1983) Membrane effects of pheno- Chlorpromazine in nonpsychotic patients with
thiazines in yeasts. I. Stimulation of calcium and pulmonary tuberculosis. American Review of
potassium fluxes. Biochimica et Biophysica Acta Respiratory Disease 81, 562–566.
733, 242–248. Hyvert, M., Le Cloarec, J. and Carrer, J. (1957)
El-Masry, S., Lotfy, M., Nasif, W.A., El-Kady, I.M. Chlorpromazine and atypical tuberculosis in
and Al-Badrawy, M. (2007) Elevated serum level 578 chronic mental patients; need for combined
of interleukin (IL)-18, interferon (IFN)-gamma therapy. Annales Médico-Psychologiques
and soluble Fas in patients with pulmonary (Paris) 115, 296–300.
complications in tuberculosis. Acta Microbio- Iademarco, M.F. and Castro, K.G. (2003)
logica et Immunologica Hungarica 54, 65–77. Epidemiology of tuberculosis. Seminars in
Escribano, I., Rodríguez, J.C., Llorca, B., García- Respiratory Infections 18, 225–240.
Pachon, E., Ruiz, M. and Royo, G. (2007) James, C.E., Mahendran, K.R., Molitor, A., Bolla,
Importance of the efflux pump systems in the J.M., Bessonov, A.N., Winterhalter, M., et al.
resistance of Mycobacterium tuberculosis to (2009) How beta-lactam antibiotics enter
fluoroquinolones and linezolid. Chemotherapy bacteria: a dialogue with the porins. PLoS One
53, 397–401. 4, e5453.
Filippov, Mi. (1960) Aminazin therapy of active Jiang, X., Zhang, W., Zhang, Y., Gao, F., Lu, C.,
forms of tuberculosis in mental patients. Zhang, X., et al. (2008) Assessment of efflux
(Preliminary communication). Zhurnal nevro- pump gene expression in a clinical isolate
patologii i psikhiatrii imeni S.S. Korsakova 60, Mycobacterium tuberculosis by real-time
1024–1026. reverse transcription PCR. Microbial Drug
García, J., Puentes, A., Rodríguez, L., Ocampo, Resistance 14, 7–11.
M., Curtidor, H., Vera, R., et al. (2005) Kongsamut, S., Kang, J., Chen, X.L., Roehr, J. and
Mycobacterium tuberculosis Rv2536 protein Rampe, D. (2002) A comparison of the receptor
implicated in specific binding to human cell binding and HERG channel affinities for a series
lines. Protein Science 14, 2236–2245. of antipsychotic drugs. European Journal of
Gooding, S., Chowdhury, O., Hinks, T., Richeldi, L., Pharmacology 450, 37–41.
Losi, M., Ewer, K., et al. (2007) Impact of a T Kristiansen, J.E. and Amaral, L. (1997) The
cell-based blood test for tuberculosis infection potential management of resistant infections
on clinical decision-making in routine practice. with non-antibiotics. Journal of Antimicrobial
Journal of Infection 54, e169–174. Chemotherapy 40, 319–327.
Gumbo, T., Louie, A., Liu, W., Ambrose, P.G., Kristiansen, J.E. and Vergmann, B. (1986) The
Bhavnani, S.M., Brown, D., et al. (2007) antibacterial effect of selected phenothiazines
Isoniazid’s bactericidal activity ceases because and thioxanthenes on slow-growing myco-
bacteria. Acta Pathologica et Microbiologica expression in infected macrophages in slowly

Scandinavica – Section B 94, 393–398. progressive primary murine Mycobacterium
Link, T.M., Park, U., Vonakis, B.M., Raben, D.M., tuberculosis infection. Scandinavian Journal of
Soloski, M.J. and Caterina, M.J. (2010) TRPV2 Immunology 56, 383–391.
has apivotal role in macrophage particle binding Molnár, J., Béládi, I. and Földes, I. (1977) Studies
and phagocytosis. Nature Immunology 11, 232– on antituberculotic action of some phenothiazine
239. derivatives in vitro. Zentralblatt für Bakteriologie,
Lipsitch, M. and Levin, B.R. (1998) Population Parasitenkunde, Infektionskrankheiten und
dynamics of tuberculosis treatment: mathe- Hygiene. Erste Abteilung Originale. Reihe A:
matical models of the roles of non-compliance Medizinische Mikrobiologie und Parasitologie
and bacterial heterogeneity in the evolution of 239, 521–526.
drug resistance. International Journal of Moriyama, Y., Tsai, H.L. and Futai, M. (1993)
Tubercle and Lung Diseases 2, 187–199. Energy-dependent accumulation of neuron
London, L. (2009) Confinement for extensively blockers causes selective inhibition of neuro-
drug-resistant tuberculosis: balancing protection transmitter uptake by brain synaptic vesicles.
of health systems, individual rights and the Archives of Biochemistry and Biophysics 305,
public’s health. International Journal of Tubercle 278–281.
and Lung Diseases 13, 1200–1209. Moss, A.R., Alland, D., Telzak, E., Hewlett, D. Jr,
Marchand, H. and Reuter, C. (1957) Phenothiazine Sharp, V., Chiliade, P., et al. (1997) A city-wide
derivatives in treatment of pulmonary tubercu- outbreak of a multiple-drug-resistant strain of
losis. Tuberkulosearzt 11, 19–27. Mycobacterium tuberculosis in New York.
Martins, M., Schelz, Z., Martins, A., Molnar, J.,
International Journal of Tubercle and Lung
Hajös, G., Riedl, Z., et al. (2007a) In vitro and ex
Diseases 1, 115–121.
vivo activity of thioridazine derivatives against
Ordway, D., Viveiros, M., Leandro, C., Bettencourt,
Mycobacterium tuberculosis. International
R., Almeida, J., Martins, M., et al. (2003) Clinical
Journal of Antimicrobial Agents 29, 338–340.
concentrations of thioridazine kill intracellular
Martins, M., Viveiros, M., Kristiansen, J.E., Molnar,
multidrug-resistant Mycobacterium tuberculosis.
J. and Amaral, L. (2007b) The curative activity of
Antimicrobial Agents and Chemotherapy 47,
thioridazine on mice infected with Myco-
917–922.
bacterium tuberculosis. In Vivo 21, 771–775.
Parsons, L.M., Somoskövi, A., Urbanczik, R. and
Martins, M., Viveiros, M. and Amaral, L. (2008a)
Salfinger, M. (2004) Laboratory diagnostic
Inhibitors of Ca2+ and K+ transport enhance
intracellular killing of M. tuberculosis by non- aspects of drug resistant tuberculosis. Frontiers
killing macrophages. In Vivo 22, 69–75. in Bioscience 9, 2086–2105.
Martins, M., Viveiros, M. and Amaral, L. (2008b) Plaza, D.F., Curtidor, H., Patarroyo, M.A., Chapeton-
The TB laboratory of the future: macrophage- Montes, J.A., Reyes, C., Barreto, J., et al. (2007)
based selection of XDR-TB therapeutics. Future The Mycobacterium tuberculosis membrane
Microbiology 3, 135–144. protein Rv2560 – biochemical and functional
Martins, M., Viveiros, M., Couto, I. and Amaral, L. studies. FEBS Journal 274, 6352–6364.
(2009) Targeting human macrophages for Pleasure, H. (1956) Chlorpromazine (thorazine) for
enhanced killing of intracellular XDR-TB and mental illness in the presence of pulmonary
MDR-TB. The International Journal of Tubercu- tuberculosis. The Psychiatric Quarterly 30,
losis and Lung Disease 13, 569–573. 23–30.
Migliori, G.B., D’Arcy Richardson, M., Sotgiu, G. Purdy, G.E., Niederweis, M. and Russell, D.G.
and Lange, C. (2009) Multidrug-resistant and (2009) Decreased outer membrane permeability
extensively drug-resistant tuberculosis in the protects mycobacteria from killing by ubiquitin-
West. Europe and United States: epidemiology, derived peptides. Molecular Microbiology 73,
surveillance, and control. Clinics In Chest 844–857.
Medicine 30, 637–665. Ramón-García, S., Martín, C., Thompson, C.J. and
Migliori, G.B., Centis, R., Lange, C., Richardson, Aínsa, J.A. (2009) Role of the Mycobacterium
M.D. and Sotgiu, G. (2010) Emerging epidemic tuberculosis P55 efflux pump in intrinsic drug
of drug-resistant tuberculosis in Europe, Russia, resistance, oxidative stress responses, and
China, South America and Asia: current status growth. Antimicrobial Agents and Chemotherapy
and global perspectives. Current Opinion in 53, 3675–3682.
Pulmonary Medicine 16, 171–179. Reeves, E.P., Lu, H., Jacobs, H.L., Messina, C.G.,
Mogga, S.J., Mustafa, T., Sviland, L. and Nilsen, R. Bolsover, S., Gabella, G., et al. (2002) Killing
(2002) Increased Bcl-2 and reduced Bax activity of neutrophils is mediated through
activation of proteases by K+ flux. Nature 416, activated macrophage-induced luminol-

291–297. dependent chemiluminescence. General
Rodrigues, L., Sampaio, D., Couto, I., Machado, D., Physiology and Biophysics 16, 3–14.
Kern, W.V., Amaral, L., et al. (2009) The role of van Ingen, J., van der Laan, T., Amaral, L.,
efflux pumps in macrolide resistance in Myco- Dekhuijzen, R., Boeree, M.J. and van
bacterium avium complex. International Journal Soolingen, D. (2009) In vitro activity of
of Antimicrobial Agents 34, 529–533. thioridazine against mycobacteria. International
Salfinger, M. and Pfyffer, G.E. (1994) The new Journal of Antimicrobial Agents 34, 190–191.
diagnostic mycobacteriology laboratory. Euro- Victor, T.C., Lee, H., Cho, S.N., Jordaan, A.M., van
pean Journal of Clinical Microbiology and Infec- der Spuy, G., van Helden, P.D., et al. (2002)
tious Diseases 13, 961–979. Molecular detection of early appearance of
Sampaio-Maia, B. and Soares-Da-Silva, P. (2001) drug resistance during Mycobacterium
Inhibition of calcium-independent luminal tuberculosis infection. Clinical Chemistry and
uptake of L-dopa by calmodulin antagonists in Laboratory Medicine 40, 876–881.
immortalized rat capillary cerebral endothelial Viveiros, M., Portugal, I., Bettencourt, R., Victor,
cells. Cell Biology International 25, 245–252. T.C., Jordaan, A.M., Leandro, C., et al. (2002)
Santopadre, I. and Silanos, G. (1959) Neuroplegics Isoniazid-induced transient high-level
as a therapeutic aid in treatment of tuberculous resistance in Mycobacterium tuberculosis.
meningitis. La Clinica Pediatrica (Bologna) 41, Antimicrobial Agents and Chemotherapy 46,
925–936. 2804–2810.
Segal, A.W. (2005) How neutrophils kill microbes. Viveiros, M., Leandro, C. and Amaral, L. (2003)
Annual Review of Immunology 23, 197–223. Mycobacterial efflux pumps and chemo-
Shubin, H., Heiken, Ca., Glaskin, A., Pennes, E. therapeutic implications. International Journal
and Arsenian, J. (1957) Chlorpromazine as an of Antimicrobial Agents 22, 274–278.
adjunct in managing tuberculous patients. Viveiros, M., Leandro, C., Rodrigues, L., Almeida,
International Record of Medicine and General J., Bettencourt, R., Couto, I., et al. (2005) Direct
Practice Clinics 170, 369–373. application of the INNO-LiPA Rif.TB line-probe
Siroy, A., Mailaender, C., Harder, D., Koerber, S., assay for rapid identification of Mycobacterium
Wolschendorf, F., Danilchanka, O., et al. (2008) tuberculosis complex strains and detection of
Rv1698 of Mycobacterium tuberculosis repre- rifampin resistance in 360 smear-positive
sents a new class of channel-forming outer respiratory specimens from an area of high
membrane proteins. The Journal of Biological incidence of multidrug-resistant tuberculosis.
Chemistry 283, 17827–17837. Journal of Clinical Microbiology 43, 4880–4884.
Sohaskey, C.D. (2008) Nitrate enhances the Viveiros, M., Dupont, M., Rodrigues, L., Couto, I.,
survival of Mycobacterium tuberculosis during Davin-Regli, A., Martins, M., et al. (2007)
inhibition of respiration. The Journal of Bacterio- Antibiotic stress, genetic response and altered
logy 190, 2981–2986. permeability of E. coli. PLoS One 2, e365.
Svetlíková, Z., Skovierová, H., Niederweis, M., Wayne, L.G. and Sohaskey, C.D. (2001) Non-
Gaillard, J.L., McDonnell, G. and Jackson, M. replicating persistence of Mycobacterium
(2009) Role of porins in the susceptibility of tuberculosis. Annual Review of Microbiology
Mycobacterium smegmatis and Mycobacterium 55, 139–163.
chelonae to aldehyde-based disinfectants and Wójcikowski, J. and Daniel, W.A. (2002)
drugs. Antimicrobial Agents and Chemotherapy Thioridazine-fluoxetine interaction at the level
53, 4015–4018. of the distribution process in vivo. Polish
Thanacoody, H.K. (2007) Thioridazine: resurrection Journal of Pharmacology 54, 647–654.
as an antimicrobial agent? British Journal of Yoshida, A., Inagawa, H., Kohchi, C., Nishizawa, T.
Clinical Pharmacology 64, 566–574. and Soma, G. (2009) The role of toll-like
Tornieporth, N.G., Ptachewich, Y., Poltoratskaia, N., receptor 2 in survival strategies of Myco-
Ravi, B.S., Katapadi, M., Berger, J.J., et al. bacterium tuberculosis in macrophage phago-
(1997) Tuberculosis among foreign-born somes. Anticancer Research 29, 907–910.
persons in New York City, 1992–1994: Zamfir, D. and Ionesco, G. (1958) Healing of
implications for tuberculosis control. Inter- tuberculous caverns of the lung, with persist-
national Journal of Tubercle and Lung Diseases ence of the cavitary image, after treatment by
1, 528–535. antibiotics associated with chlorpromazine and
Traykov, T., Hadjimitova, V., Goliysky, P. and phenylsemicarbazide. Le Poumon et le Coeur
Ribarov, S. (1997) Effect of phenothiazines on 14, 1093–1097.
19 Vaccines for Tuberculosis
Helen McShane
The Jenner Institute, University of Oxford, Oxford, UK
19.1 Existing Vaccines result of this trial, routine BCG administration

to adolescent schoolchildren was introduced
The only licensed vaccine against tuberculosis throughout the UK, a practice which has only
is an attenuated strain of Mycobacterium bovis, recently been revised (Fine, 2005). In contrast,
named Bacille Calmette–Guérin (BCG) after many large-scale efficacy trials conducted in
the French scientists who were responsible the USA and in various countries in the
for developing it. BCG was first used per os in developing world failed to demonstrate any
1921 and has been routinely administered protective efficacy at all (Comstock and
intradermally or subcutaneously in many Palmer, 1966; Baily, 1980).
countries throughout the world for many Understanding the scientific basis for
decades. It forms an integral part of the this variability in protective efficacy poses
Expanded Programme on Immunization one of the biggest challenges in tuberculosis
(EPI) infant vaccination schedule and is vaccine development. An understanding of
administered at birth throughout the develop- why BCG appears to work well in some
ing world. BCG has never been routinely used populations but not in others is important for
in the USA for the prevention of tuberculosis, the design of new tuberculosis vaccines. Any
but is widely used in a variety of vaccination new vaccine ideally would work across all
schedules throughout Europe. Since BCG populations throughout the world.
was first developed, many efficacy trials have Many potential hypotheses have been
been conducted with the aim of evaluating proposed to explain why BCG is less effective
protective efficacy in different populations. It in tropical climates than in temperate ones.
is clear that BCG immunization in infancy Some of these key hypotheses are listed in
confers significant protection against dis- Box 19.1.
seminated disease (Rodrigues et al., 1993; There are many different substrains of
Trunz et al., 2006). However, the level of BCG in use throughout the world that have
efficacy demonstrated against pulmonary evolved through distribution and repeated
disease varies widely across the many trials subculture (Ritz and Curtis, 2009). With
conducted in tuberculosis-endemic countries modern genetic sequencing tools, it is pos-
throughout the world (Colditz et al., 1994). In sible to identify genetic differences in these
temperate climates, trials such as the British BCG subtypes (Behr et al., 1999). Levels of
MRC study, conducted in the 1950s, show a gene expression also differ, particularly
clear and significant degree of protective between BCG strains derived early on in the
efficacy (Hart and Sutherland, 1977). As a development and strains which were derived

244 H. McShane
Box 19.1.
Potential explanations for variability in BCG across geographical areas
Genetic sequence differences between BCG strains in clinical use
Microbiological differences in BCG formulation
Nutritional status of population
Pre-existing exposure to non-tuberculous mycobacteria
later (Brosch et al., 2007). It is not a simple malnourished and there is no evidence that
task to relate these genetic differences to the BCG protects less well in malnourished
variability in protective efficacy, as the effi- individuals (Fine and Rodrigues, 1990).
cacy trials have been conducted in genetically Another potential explanation for the
diverse populations with differing exposure variability in efficacy across temperate and
to mycobacteria and different trial designs. tropical climates is that exposure to non-
As well as evaluating different substrains of tuberculous mycobacteria (NTM) may offer
BCG, these trials used different formulations some degree of protective immunity. Such
of BCG, with different routes of delivery. exposure varies according to geographical
Some of the early studies used a mid-log latitude, being greater in more tropical
phase of growth formulation, where the climates. This NTM exposure may also
mycobacteria were most likely to be interfere directly with the protective efficacy
replicating, whereas some of the later studies conferred by subsequent BCG vaccination.
used a lyophilized formulation. These differ- The first demonstration that NTM exposure
ent formulations may have different pro- can protect against a subsequent M. tubercu-
portions of live and dead bacilli, which may losis challenge was in guinea pigs (Palmer
impact on immunogenicity and protective and Long, 1966). In clinical studies, work by
efficacy (Behr, 2002). Data from preclinical Dockrell et al. has demonstrated a high degree
animal models where direct comparisons can of pre-existing exposure to NTM in BCG-
be performed more easily demonstrate that naïve adolescents in Malawi, where BCG is
the different BCG substrains result in known not to work well, and very little pre-
different levels of immunogenicity and existing exposure in a population of BCG-
protective efficacy in some, but not all, studies naïve adolescents in the UK, where BCG has
(Freudenstein et al., 1979; Smith et al., 1979; previously been shown to be highly effective
Lagranderie et al., 1996; Horwitz et al., 2009). (Black et al., 2002). In this study, BCG
Data from clinical studies where the vaccination in each of these populations had
immunogenicity of different BCG strains has a very different effect on anti-mycobacterial
been directly compared also reveal some immune responses. In the UK, where there
studies which demonstrate differences and was little pre-existing exposure, vaccination
some which do not (Brindle et al., 1972; with BCG induced a potent Th-1 cellular
Davids et al., 2006; Gorak-Stolinska et al., immune response. In contrast, in the
2006). Nevertheless, strain differences cannot Malawian adolescents who had a high pre-
explain all the variability in protective existing level of anti-mycobacterial immune
efficacy, as Danish BCG gave significant responses, vaccination with BCG did not alter
protection in the UK but not in Southern this level of immunity significantly (Black et
India (Hart and Sutherland, 1977; Baily, 1980). al., 2002). This phenomenon is referred to as
Nutritional factors may also be relevant ‘masking’, whereby the anti-mycobacterial
in contributing to an explanation regarding immunity induced by NTM exposure ‘masks’
the variability in BCG efficacy. However, the immunity induced by BCG. The impli-
while it is likely that no vaccine would work cation from these data is also that vaccination
well in a profoundly malnourished popu- with BCG cannot boost this NTM-induced
lation, many of the populations in which immunity. Subsequent work has demon-
BCG does not work well are not severely strated a dramatic difference in the pro- and
Vaccines for Tuberculosis 245
anti-inflammatory profile induced by BCG in killer cells and neutrophils have both been
the UK and Malawian adolescents studied implicated in this early protective effect (Feng
(Weir et al., 2004). A related hypothesis is that et al., 2006; Korbel et al., 2008). Gamma-delta
NTM-induced immunity inhibits BCG T cells, a subpopulation of T cells thought to
replication, and hence interferes with or have a role in the non-adaptive, early host
‘blocks’ the induction of a BCG-induced immune response to foreign pathogens, also
protective immune response. There are some appear to have a protective role against
data from preclinical studies which support mycobacteria (Shen et al., 2002; Spencer et al.,
this explanation, and it is known that 2008). Once infected, M. tuberculosis is an
inhibiting BCG replication with isoniazid intracellular pathogen, residing primarily
inhibits the induction of a protective immune inside macrophages. Such an intracellular
response (Dworski, 1973; Brandt et al., 2002). organism is resistant to antibody-mediated
More recently, a further limitation of immune defence mechanisms, and the
routine BCG vaccination in infancy has been primary protective immune response against
identified. A rare but well-recognized side mycobacteria is a T cell-mediated immune
effect of BCG vaccination is disseminated response. Herein lies the first challenge. With
BCG disease. Over the past few years, it has the exception of BCG, there are no vaccines
become clear that the incidence of dis- licensed and in routine use which induce the
seminated BCG disease is higher than initially cellular arm of the immune response. All of
thought in populations with a high prevalence the vaccines used today work via, or were
of HIV infection (Hesseling et al., 2007). Such licensed on the basis of, the induction of a
data have led the World Health Organization humoral immune response. Developing the
(WHO) to amend the BCG vaccination guide- next generation of vaccines which induce
lines, and these guidelines now recommend strong cellular immunity represents one of
withholding BCG vaccination in infancy in the significant hurdles in the field.
areas where HIV is prevalent, until the HIV Within the cellular immune response,
status of the infant is known (WHO, 2007). there are many subsets of T cells and cyto-
Such a strategy is far from ideal, as many kines which are implicated in this protective
areas where HIV is prevalent do not have immune response. We know of the importance
optimal testing facilities and any strategy of Class II-restricted CD4+ T cells from
which requires determining HIV status preclinical studies in gene-deleted animals
before vaccination will result in a reduced and from adoptive transfer experiments
vaccine uptake rate. There is, therefore, an (Orme, 1987; Caruso et al., 1999). However,
urgent need for a safe vaccination regimen the best evidence for an essential role for
for all infants, including those who are HIV CD4+ T cells comes from HIV-infected popu-
infected. lations: as the absolute CD4+ count declines,
the risk of developing tuberculosis disease
increases (Jones et al., 1993). HIV-uninfected
19.2 Nature of the Protective Immune people have a 10% lifetime risk of reactivating
Response latent infection. In contrast, HIV-infected
people have a 10% annual risk of reactivation
In order to design an effective vaccine against (Corbett and De Cock, 1996).
M. tuberculosis, it is necessary to understand There is also evidence for the importance
the nature of protective immunity to myco- of Class I-restricted CD8+ T cells in protective
bacteria. Only one in ten immunocompetent immunity against M. tuberculosis. Evidence
people infected with M. tuberculosis develop for a role for CD8+ T cells comes from
disease. Ideally, a new vaccine would mimic preclinical studies in gene-depleted mice and
this naturally occurring protective immunity. adoptive transfer experiments (Sousa et al.,
An effective early innate immune 2000). In addition, there is some evidence
response at the site of infection is responsible from studies in non-human primates for an
for the clearance of pathogen in most people essential role for CD8+ T cells in BCG-induced
who are exposed to M. tuberculosis. Natural protection against M. tuberculosis challenge
246 H. McShane
(Chen et al., 2009). While there is accumulating tuberculosis (Keane et al., 2001). Other
evidence for a role for CD8+ T cells, the exact cytokines such as interleukin-2 (IL-2) are
mechanism by which they confer protection known to be important for the generation and
remains unclear. There is some evidence that maintenance of central memory T-cell
they may be more important in controlling responses, which are necessary for durable
latent infection than in containing the acute protection against M. tuberculosis (Sallusto et
phase of infection (van Pinxteren et al., 2000). al., 2004; Harari et al., 2005).
The primary action of CD8+ T cells may be More recently, there has been increasing
via the secretion of cytokines (Tascon et al., interest in so-called multifunctional T cells,
1998; Brookes et al., 2003). However, human which express the Th-1 cytokines IFN-γ,
cytotoxic T lymphocytes can kill M. tubercu- TNF-α and IL-2 on intracellular cytokine
losis, and this is dependent on the presence of staining. Such multifunctional T cells appear
granulysin in cytotoxic vesicles (Stenger et al., to be important in protection in preclinical
1998). models of leishmania and tuberculosis
A newly defined subset of T cells, called (Darrah et al., 2007; Forbes et al., 2008).
Th-17 cells, may also be involved in protective However, recent evidence from a randomized
immunity. There is some evidence from controlled clinical trial of BCG vaccination in
preclinical and clinical studies to support a South African infants found that the level of
role for these cells, which secrete the cytokine BCG-induced multifunctional T cells did not
IL-17, in protection against M. tuberculosis. differ in protected and unprotected infants
Levels of IL-17 pre-challenge correlate with (Kagina et al., 2010). Additional work in both
protection in cattle; in mice, IL-17 was clinical and preclinical studies is required to
necessary for the accumulation of interferon define further the role of these T cells in
gamma (IFN-γ)-secreting CD4+ T cells in the protective immunity.
lungs; and in humans, tuberculin skin test
(TST)-positive subjects had lower levels of
IL-17 than TST-negative subjects (Khader et 19.3 Target Populations for a New
al., 2007; Vordermeier et al., 2009; Babu et al., Tuberculosis Vaccine
2010).
There are many cytokines which are There are several different points on the
required for protective immunity against tuberculosis clinical disease spectrum at
mycobacteria. The essential nature of IFN-γ which one might intervene with an effective
has been demonstrated by studies in both vaccine, and hence a number of different
preclinical animal models and in human target populations for such a vaccine.
subjects deficient in this pathway (Flynn et al., Vaccines administered at or soon after birth
1993; Newport et al., 1996). However, it is would be administered predominantly to M.
increasingly clear that IFN-γ alone is not tuberculosis-uninfected infants and therefore
sufficient for protection and, moreover, abso- could potentially prevent infection and/or
lute levels of this cytokine, as measured by a primary disease. Infants in the developing
variety of immunological assays, do not world are an important target population, as
appear to correlate with protection (Langer- disease rates in this group are high (Moyo et
mans et al., 2001; Mittrucker et al., 2007; al., 2010). A vaccine administered during
Wedlock et al., 2007). There is evidence from adolescence/early adulthood would be
preclinical studies for the importance of administered to a mixture of M. tuberculosis
tumour necrosis factor alpha (TNF-α) in infected and M. tuberculosis uninfected
protective immunity against M. tuberculosis. adolescents, depending on the prevalence of
Mice deficient in TNF-α are more susceptible infection in any particular population.
to M. tuberculosis challenge than their wild- Adolescents and young adults form a second
type counterparts (Flynn et al., 1995). important target population, as disease
Furthermore, monoclonal antibodies against incidence throughout most of the world is
TNF-α, used as therapy for rheumatoid highest in this age group. Finally, HIV-
arthritis, increase reactivation of latent M. infected people form an important population
for deployment of an effective tuberculosis correlates of protection, which can be further

vaccine, given the high rate of disease in HIV evaluated in clinical efficacy trials to (i)
co-infected subjects. The HIV-infected popu- validate the relevance of the preclinical
lation would also consist of a mixture of M. animal model for human disease and (ii)
tuberculosis-infected and -uninfected indi- validate the correlate of protection. Most of
viduals. these animal models are used to evaluate the
One further potential use for an effective effects of prophylactic immunization in naïve
tuberculosis vaccine is as a therapeutic animals, and this represents vaccination best
vaccine for administration as an adjunct to in M. tuberculosis-uninfected subjects, most
chemotherapy in patients with tuberculosis obviously in infancy. However, with
disease. There are safety concerns about increased interest in developing vaccines
administering vaccines which induce a potent which specifically target the latent phase of
anti-mycobacterial immune response in M. tuberculosis infection, there are animal
patients with a heavy bacillary burden, as models of latency in development (Scanga et
there is some preclinical data which demon- al., 1999; Botha and Ryffel, 2002; Capuano et
strate the induction of immunopathology (so- al., 2003).
called Koch phenomenon) in animals with a Once sufficient preclinical efficacy data
heavy bacillary burden (Taylor et al., 2003). It have been obtained, preclinical safety and
is likely that any potential therapeutic vaccine toxicology studies conducted under Good
would be evaluated in conjunction with Laboratory Practice conditions are also
conventional chemotherapy. required before any new vaccine can be
evaluated in early-stage clinical trials. The
clinical trials typically begin in small numbers
19.4 Vaccine Evaluation of subjects and primarily evaluate safety and
immunogenicity in different target popu-
In general, most candidate tuberculosis lations. Once sufficient data have been
vaccines follow a well-established develop- obtained, these trials then progress through
ment programme through preclinical animal to Phase IIb/Phase III efficacy testing in large
testing in different species to early-stage numbers of subjects within a particular target
clinical testing in Phase I/IIa safety and population.
immunogenicity studies and finally through
to Phase IIb/III efficacy testing in the target
populations. Any new vaccine is first 19.5 Potential Approaches to the
evaluated in a series of preclinical animal Development of a New Tuberculosis
models, and some level of efficacy within Vaccine
more than one species is usually required
before clinical evaluation begins. All the The established protective efficacy of BCG
animal models used in tuberculosis vaccine against disseminated tuberculosis, when
development have utility but also limitations, administered in infancy, has resulted in most
and until we have an effective human vaccine, global effort focusing on the development of
we will not know how representative any of vaccination strategies to improve BCG rather
these models are of human disease. than replace it completely. Efforts to improve
In general, preclinical evaluation begins BCG have been either by adding a second
in the murine model and then moves in subunit vaccine, designed to enhance the
sequence through into guinea pigs and non- immune response and protective efficacy of
human primates. Cattle are also a relevant BCG, or by improving BCG itself in some
animal model for a new tuberculosis vaccine way. A further approach is to replace BCG
and, importantly, are a target species for a with an attenuated strain of M. tuberculosis in
new tuberculosis vaccine in their own right. order to improve the efficacy. The leading
As well as evaluating the protective efficacy new vaccination approaches currently being
of a new vaccine, these animal models can be explored in preclinical and clinical studies are
used to identify potential immunological outlined in Table 19.1.
248 H. McShane
Table 19.1. Summary of tuberculosis vaccine candidates undergoing clinical evaluation.

Stage of
Type of vaccine Vaccine name development References
BCG replace- Recombinant rBCG30 Phase I Hoft et al., 2008
ments BCG rBCGureChly Phase I Grode et al., 2005
Subunit vaccines Recombinant MVA85A Phase IIb Brookes et al., 2003; McShane et
viral vectors al., 2004; Ibanga et al., 2006;
Pathan et al., 2007; Hawkridge
et al., 2008; Sander et al.,
2009; Scriba et al., 2010
Aeras-402 Phase IIb Abel et al., 2010
Protein and M72 Phase IIa Von Eschen et al., 2009
adjuvant Hybrid I Phase IIa van Dissel et al., 2010
HyVAC IV Phase I Dietrich et al., 2005
Other M. vaccae Phase III von Reyn et al., 2010
RUTI Phase I Vilaplana et al., 2010
The general strategies and leading clinical pathway for product registration
vaccine candidates are discussed in detail which is sufficiently robust.
below, with a focus on those candidates
currently being evaluated, or close to com-
Recombinant BCG strains
mencing evaluation, in clinical trials.
Broadly, there are two general approaches
being investigated. The first is whether the
19.5.1 BCG replacements construction of recombinant strains of BCG
which overexpress certain immunodominant
There have been two WHO working group antigens will lead to an enhancement in
meetings on so-called ‘live vaccines’ which protective efficacy. One such vaccine, which
encompassed the BCG replacement vaccines demonstrates improved efficacy over the
in development, in 2004 and 2009, in order to wild-type parent strain in guinea pigs, is
facilitate development and help define a rBCG30 (Horwitz et al., 2000). A Phase I
regulatory pathway forward (Kamath et al., clinical trial with this vaccine in the USA
2005; Walker et al., 2010). An important goal demonstrated comparable safety profile and
of the live vaccine approach is to demonstrate enhanced immunogenicity when compared
improved safety over BCG, particularly given to the parentral strain of BCG (Tice) used
the concerns with using BCG in HIV-infected alone (Hoft et al., 2008). A second potential
infants (Hesseling et al., 2007). Preclinical candidate, BCG::RD1-2F9, replaces the
models such as severe combined immuno- region of deletion 1 (RD-1) which is missing
deficiency (SCID) mice can be used to demon- from BCG (Pym et al., 2003). Although this
strate improved safety when compared to recombinant BCG confers greater protection
wild-type BCG strains. In addition, levels of against aerosol M. tuberculosis challenge in
efficacy equivalent or superior to the wild- mice and guinea pigs, there are safety
type strain should also be demonstrated in concerns regarding the virulence of such a
preclinical models. The scale of clinical trial strain of BCG and, in addition, this vaccine
needed to demonstrate non-inferiority over could be likely to confound the new
wild-type strains means that such a trial may interferon gamma release assays (IGRAs),
not be feasible. One of the challenges for the Quantiferon Gold and T-Spot TB, which are
field is, therefore, to devise an alternative based on a T-cell response to two antigens in
the RD-1 gene, ESAT-6 and CFP-10 and are 19.5.2 BCG booster vaccines
now in routine clinical use as diagnostic tests
(see Chapter 4) (Pym et al., 2002; Pai et al., These vaccine candidates are designed to be
2004). The second approach being pursued is given after BCG immunization, in order to
to generate recombinant strains of BCG enhance the protective efficacy of BCG. They
which express molecules such as listeriolysin are more advanced than the potential BCG
or perfringolysin (Grode et al., 2005; Sun et replacements described above and there are
al., 2009). The scientific rationale behind currently five vaccine candidates undergoing
these candidates is to enhance cross-priming evaluation in clinical trials. These candidates
and hence improve the induction of a CD8+ involve the delivery of an immunodominant
T-cell response. One of these vaccines has antigen or antigens, either as a recombinant
been evaluated in a Phase I clinical trial in protein with adjuvant or as a recombinant
Europe (clinicaltrials.gov trial identifier viral vector. Both antigen(s) and the antigen
NCT00749034) and is currently in a Phase I delivery system need to be identified.
trial in South African adults (clinicaltrials. Regarding antigen selection, there has been a
gov trial identifier NCT01113281). It is focus on the well-characterized and immuno-
possible to combine these two approaches, dominant antigens such as the antigen 85
and recombinant strains of BCG which complex, and RD-1 genes such as ESAT-6 and
express perfringolysin and immunodominant TB10.4. The antigen 85 complex (A, B and C)
antigens are currently in development (Sun constitutes a major portion of the secreted
et al., 2009). proteins from M. tuberculosis, BCG and all
other mycobacteria sequenced to date
(Andersen et al., 1992). The antigen 85
Attenuated M. tuberculosis
complex is protective against M. tuberculosis
strains
challenge in small animals, either as a DNA
An alternative approach to modifying BCG is vaccine or as protein (Horwitz et al., 1995;
to develop a rationally attenuated strain of M. Huygen et al., 1996). Inclusion of genes from
tuberculosis. The potential advantage in using the RD-1, either in a subunit vaccine or in a
M. tuberculosis as the starting organism rather recombinant BCG strain, may potentially
than BCG is that some of the genes which are confound the new IGRA diagnostic tests and
present in M. tuberculosis but deleted in BCG this may limit the development of candidates
may be relevant for protection (Hernandez containing these antigens, as these IGRAs
Pando et al., 2006). There are two distinct rely on detecting a T-cell response to the RD-1
approaches currently being developed. One antigens, ESAT-6 and CFP-10 (Pym et al.,
candidate is an attenuated strain of M. 2002; Pai et al., 2004).
tuberculosis with a disrupted Pho-P gene, There is increasing interest in the
which has demonstrated superior efficacy inclusion of so-called ‘dormancy’ or hypoxia-
when compared to BCG in the guinea pig and induced antigens, which may be preferentially
non-human primate models (Williams et al., upregulated and expressed during latent
2005b; Verreck et al., 2009). The safety and infection. Inclusion of such antigens into a
efficacy of two other attenuated M. tuberculosis subunit vaccine may be necessary for optimal
vaccine strains, mc26020 and mc26030, in non- efficacy of a post-exposure vaccine admin-
human primates has also been demonstrated istered to latently infected subjects (Roupie et
(Larsen et al., 2009). al., 2007). In addition, while the most
There are significant safety concerns advanced candidates have focused on protein
regarding the clinical evaluation of attenuated antigens, there is now an increasing interest
strains of M. tuberculosis. However, the WHO in non-protein, glycolipid antigens as well
working group has defined a pathway (Guiard et al., 2009).
forward for early-stage clinical evaluation of In terms of antigen delivery systems,
such vaccines and these vaccine candidates there are two main approaches being pursued
represent an important proof-of-concept in and the leading candidates using each
tuberculosis vaccine development. approach are discussed below.
250 H. McShane
Recombinant viral vectors response. Recombinant MVAs are known to

be effective at boosting pre-existing T-cell
This approach to antigen delivery requires
responses (Schneider et al., 1998; Hanke et al.,
the selection of antigen. This approach clones
1999; Amara et al., 2001; McConkey et al.,
the antigen(s) of choice into a viral vector and
2003). In MVA85A, the antigen expressed by
uses the virus itself as an endogenous
the MVA vector is antigen 85A. Preclinical
adjuvant. Many viruses have been used as
studies with MVA85A have demonstrated
vectors for antigen delivery, such as pox
that MVA85A can improve BCG-induced
viruses, adenoviruses, flaviviruses and
protection in animal models (McShane et al.,
lentiviruses (Prevec et al., 1989; Paoletti, 1996;
2002; Williams et al., 2005a; Verreck et al.,
Bonaldo et al., 2010; Draper and Heeney, 2010;
2009; Vordermeier et al., 2009). The last of
Negri et al., 2010). Safety concerns with the
these references demonstrates that both
use of replicating vectors means that the
MVA85A and an adenovirus expressing
challenge is to develop an attenuated viral
antigen 85A can both improve BCG-induced
vector which induces a persistent immune
protection against M. bovis challenge in cattle
response. One significant limitation of virally
(Vordermeier et al., 2009). This cattle study
vectored vaccines is the presence of immunity
identified two potential correlates of
to the vector (virus), which limits the ability
protection: both levels of antigen 85A-specific
of the recombinant virus to induce or boost
IL-17 and cultured ELISPOT responses to
the immune response. This is relevant for
antigen 85A on the day of challenge correlated
naturally occurring antivector immunity (as
with subsequent protection in this experi-
with human adenoviral strains, particularly
ment. These correlates are now being
AdHu5) (Tatsis and Ertl, 2004) but also for
evaluated in the ongoing clinical studies.
repeated use of a single vector, both for
MVA85A was the first new tuberculosis
boosting an immune response to a single
subunit vaccine to enter into clinical trials in
pathogen and for multiple usage of the same
September 2002 and is currently the most
vector as a vaccine for different pathogens.
clinically advanced new tuberculosis vaccine.
Heterologous prime-boost immunization
Since 2002, a series of Phase I/IIa clinical trials
regimens, where two different viral vectors
in the UK, The Gambia and South Africa have
encoding the same antigen are delivered
been conducted with this vaccine candidate
sequentially, are an effective way to induce
in populations including BCG naïve adults,
high levels of cellular immunity for both
BCG vaccinated adults, adolescents, children
CD4+ and CD8+ T cells and circumvent the
and infants, M. tuberculosis-infected adults
problems of antivector immunity (Schneider
and, most recently, HIV-infected adults
et al., 1998; Hanke et al., 1999; McShane et al.,
(Brookes et al., 2003; McShane et al., 2004;
2001). For the development of a tuberculosis
Ibanga et al., 2006; Pathan et al., 2007; Hawk-
vaccine, BCG becomes the priming
ridge et al., 2008; Sander et al., 2009; Scriba et
immunization in such a strategy and many
al., 2010; Minassian et al., unpublished data).
groups, including our own, have focused on
These trials have all demonstrated this
developing an effective ‘boost’ for BCG. Such
vaccine candidate to be safe and immuno-
a strategy retains the protective efficacy of
genic, and to be particularly potent at
BCG conferred against disseminated disease
stimulating an antigen-specific CD4+ T-cell
and aims to improve protection against
response. More detailed characterization of
pulmonary disease. There are two leading
the vaccine-induced immune response
approaches using viral vectors currently
demonstrates a high level of polyfunctional
being developed.
CD4+ T cells, which are positive on
intracellular cytokine staining performed on
MVA85A (OXFORD UNIVERSITY; OXFORD EMERGENT either peripheral blood mononuclear cells
TUBERCULOSIS CONSORTIUM) This approach (PBMCs) or whole blood samples for
uses the modified vaccinia virus Ankara interferon gamma, TNF-α, IL-2, IL-17 and
(MVA), an attenuated strain of vaccinia, to MiP1b (Beveridge et al., 2007). These antigen-
induce a potent cell-mediated immune specific T cells proliferate and have a
non-terminally differentiated phenotype after the primary immunization did not

(Beveridge et al., 2007). It is also possible to result in higher levels of antigen-specific T
detect a modest antigen-specific CD8+ T-cell cells, likely because of antivector immunity
response after vaccination with this candidate to the AdHu35 induced by the first dose
(Whelan et al., 2009). Furthermore, it is clear (Abel et al., 2010). This vaccine is now being
that a regulatory T-cell response is also evaluated in a Phase II safety, immunogenicity
induced after vaccination (Fletcher et al., and efficacy study in HIV-infected adults in
2008). This last study demonstrates the South Africa (clinicaltrials.gov trial identifier
importance of characterizing the regulatory NCT01017536; http://clinicaltrials.gov/ct2/
T-cell response as well as the effector T-cell show/NCT01017536?term=NCT01017536&r
response in early-stage vaccine trials, as any ank=1).
effective immune response will be the result
of a balance between these different aspects
Protein/adjuvant combinations
of the cellular immune response. A Phase IIb
proof-of-concept efficacy trial with MVA85A Using recombinant protein is one other
commenced in South African BCG-vaccinated method of developing a subunit vaccine
infants in 2009 (clinicaltrials.gov trial identi- based on one or a few immunodominant
fier NCT00953927; http://clinicaltrials.gov/ antigens. Recombinant protein delivered
ct2/show/NCT00953927?term=NCT00953927 alone is a poor inducer of a cellular immune
&rank=1). response. However, recombinant protein
which is co-administered with an adjuvant
AERAS-402 (AERAS; CRUCELL) This vaccine can be an effective way to induce a cellular
candidate is a recombinant, replication- immune response. The challenge is in
deficient adenovirus, serotype 35, expressing developing adjuvants which are good at
a fusion protein created from the sequences stimulating the Th-1 pathway. The most
of the antigens Ag85A, B and TB10.4 from M. widely used adjuvant, aluminium salts,
tuberculosis. One of the limitations of using induces primarily antibody responses and a
adenoviruses as viral vectors for vaccines in Th-2 biased cellular immune response
general is the level of pre-existing, naturally (HogenEsch, 2002). There are new adjuvants
occurring antivector immunity. Immunity to being developed which induce a Th-1 biased
the commonest serotype, AdHu5, exists at cellular immune response and some of these
levels of up to 90% in some populations are being evaluated with mycobacterial
(Xiang et al., 2006). Pre-existing immunity to proteins for use as a tuberculosis vaccine.
AdHu35, a rarer serotype, is known to be These protein/adjuvant combinations are
considerably lower (Kostense et al., 2004). also being developed as booster vaccines to
Aeras-402 is a recombinant strain of AdHu35 enhance the protective efficacy conferred by
expressing antigens 85A, B and TB10.4 (Abel BCG. To date, there are three such protein/
et al., 2010). Preclinical studies in mice have adjuvant candidate tuberculosis vaccines
demonstrated this vaccine confers levels of which have entered into clinical testing:
protection comparable with BCG when
administered by either the systemic or the HYBRID I (STATUM SERUM INSTITUTE, SSI) This
mucosal route (Radosevic et al., 2007). This vaccine consists of a fusion protein of antigen
vaccine is also immunogenic in non-human 85B and ESAT-6. Preclinical studies with this
primates following either BCG or recom- protein, administered intranasally with the
binant BCG prime (Magalhaes et al., 2008). A mucosal adjuvant LTK63, a non-toxic mutant
clinical trial with this vaccine in of the Escherichia coli heat labile enterotoxin,
BCG-vaccinated South African adults has demonstrated an improvement on BCG-
demonstrated a good safety profile, together induced protection in mice (Dietrich et al.,
with high levels of antigen-specific CD8+ T 2006). However, no improvement on
cells and more modest levels of CD4+ T cells BCG-induced protection was demonstrated
(Abel et al., 2010). In this trial, a second in the guinea pig (Williams et al., 2005b). A
booster vaccination administered 2 months clinical trial of this vaccine candidate in BCG
252 H. McShane
naïve subjects has now been completed (van stered with the adjuvant AS02A, a proprietary
Dissel et al., 2010). When administered alone, oil in water emulsion with the immuno-
this protein was only very weakly immuno- stimulants, monophosphoryl lipid A and
genic. When co-administered with a high Quillaja saponaria fraction 21, has demon-
dose of a novel adjuvant, IC31, a synthetic strated comparable efficacy to BCG in mice
formulation which combines an antimicrobial and guinea pigs (Skeiky et al., 2004) and to
peptide, KLK, and an immunostimulatory improve BCG-induced protection in guinea
oligodeoxynocleotide, ODN1a, significantly pigs but not mice (Brandt et al., 2004). A Phase
higher immunogenicity was demonstrated I clinical trial with M72, administered with
using a short-term cultured ELISPOT assay AS02A, demonstrated this candidate to be
(van Dissel et al., 2010). Further studies are modestly reactogenic, with a significant
now being conducted in purified protein induction of an antigen-specific CD4+ T-cell
derivative (PPD)-positive subjects (clinical response but no detectable CD8+ T-cell
trials.gov trial identifier NCT00929396; http:// responses (Von Eschen et al., 2009). Further
clinicaltrials.gov/ct2/show/NCT00929396?ter studies are under way in tuberculosis-
m=NCT00929396&rank=1); and with another endemic regions (clinicaltrials.gov trial
novel adjuvant, CAF01 (clinicaltrials.gov trial identifier NCT00600782; http://clinicaltrials.
identifier NCT00922363; http://clinicaltrials. gov/ct2/show/NCT00600782?term=NCT0060
gov/ct2/show/NCT00922363?term=NCT0092 0782&rank=1).
2363&rank=1). One potential limitation with
this protein/adjuvant vaccine candidate is the
Other tuberculosis vaccines in development
potential for the T-cell response to ESAT-6 to
interfere with the new IGRAs, as discussed
above. To date, the available clinical data MYCOBACTERIUM VACCAE M. vaccae is a whole
demonstrate a low level of Quantiferon cell lysate of a culture of M. vaccae, which has
positivity (one of the new IGRA tests) after been evaluated in many clinical trials as a
vaccination with Hybrid I; however, this therapeutic vaccine over the past 10 years.
construct has not yet been evaluated in The results from most of these trials have
tuberculosis-endemic countries (van Dissel et been variable (Stanford et al., 1990; Durban
al., 2010). A Phase I study evaluating the Immunotherapy Trial Group, 1999; Johnson
safety of this candidate administered intra- et al., 2000). More recently, a Phase III study in
nasally was terminated for safety reasons HIV-infected adults in Tanzania evaluating
(clinicaltrials.gov trial identifier NCT00440544; repeated immunizations with M. vaccae
http://clinicaltrials.gov/ct2/show/NCT0044054 demonstrated a modest degree of efficacy
4?term=NCT00440544&rank=1). against culture-positive tuberculosis (von
Reyn et al., 2010). The primary end point in
HYVAC IV (SSI) In this second-generation this study was disseminated disease, and too
fusion protein from SSI, the ESAT-6 antigen few cases were accrued for this end point to
has been replaced with another immuno- be met. Further studies are required to
dominant antigen from M. tuberculosis, demonstrate the reproducibility of this result,
TB10.4, in order to avoid the problems of where culture-positive disease is the primary
confounding the diagnostic IGRA assays end point. If these results are reproducible,
(Dietrich et al., 2005). Preclinical data with this study represents an important proof-of-
this construct show levels of protection concept that vaccine-induced protection
comparable to BCG in mice and guinea pigs against tuberculosis, even in immuno-
(Dietrich et al., 2005; Skeiky et al., 2010). suppressed subjects, is achievable.
M72 (GSK) This vaccine candidate consists of RUTI RUTI is a potential therapeutic vaccine
a fusion protein of two antigens from M. for tuberculosis which consists of detoxified
tuberculosis, the 32 kDa and the 39 kDa and liposomal cellular fragments of M.
antigens (Alderson et al., 2000). In preclinical tuberculosis (Vilaplana et al., 2010). A Phase I
models, this fusion protein, when admini- clinical trial in healthy subjects has recently
demonstrated that this vaccine is safe and any immune correlate is vaccine specific.
modestly immunogenic for the induction of Results from a recent randomized controlled
antigen-specific IFN-γ (Vilaplana et al., 2010). trial of BCG vaccination in South African
infants demonstrated that neither frequency
nor function of antigen-specific T cells,
19.6 Major Challenges to measured 10 weeks after BCG vaccination,
Tuberculosis Vaccine Development correlated with protection (Kagina et al.,
2010). The challenge with these studies,
As vaccine candidates are advanced through particularly in infants where blood volumes
preclinical studies to clinical evaluation, the are limited, is to identify the optimal time
major challenges in the field become more point for evaluation. BCG is a complex whole
apparent. These are discussed below. organism vaccine, known to induce a robust
innate immune response. Further work is
required to evaluate correlates of protection
19.6.1 Lack of immunological correlates in the ongoing efficacy trials with the new
of protection candidates.
The identification of an immunological

correlate of protection would greatly facilitate 19.6.2 Lack of validated animal models
vaccine development. Although we under-
stand broadly the nature of a protective Although there are many preclinical animal
immune response to M. tuberculosis, we do models of tuberculosis vaccine evaluation, we
not yet have a simple, validated measure of do not yet know which, if any, of these models
immunogenicity that will predict vaccine predicts vaccine success in humans. It is only
success in human clinical trials. The only way once we have a vaccine which protects against
to evaluate whether a vaccine is effective is in human disease that we can go back to the
large-scale, expensive and time-consuming animal models and evaluate which one(s)
efficacy trials. The early-stage clinical trials predicts success best in humans. It is only by
are essential for the demonstration of safety, moving forward in an iterative way, evalu-
and also immunogenicity, in order to be sure ating several candidates in proof-of-concept
that the vaccine candidate in question is efficacy trials, that we can reach a clearer
inducing what is thought to be a protective understanding as to the nature of protective
immune response. However, in the absence immunity and the most useful animal models
of predefined correlates, such data do not with which to evaluate subsequent vaccine
necessarily predict efficacy. It is only possible candidates. Such models can be utilized for
to validate an immunological correlate once the identification of potential immunological
we have an effective vaccine. It would be too correlates, which can then be evaluated in
resource- and time-inefficient to perform clinical efficacy trials.
detailed immunology on all subjects in
efficacy trials. However, if peripheral blood
mononuclear cells and serum samples from 19.6.3 Lack of high-incidence field sites
all subjects are stored in such trials, at the end
of the trial samples from diseased (i.e. not Although the global prevalence of tubercu-
protected) subjects and matched healthy losis disease is high, the incidence in any
exposed controls can be analysed in order to particular population is relatively low. Any
identify potential immunological correlates. efficacy trial needs to be conducted in a
Any potential correlates can then be validated population with as high an incidence as
subsequently in a Phase III trial. However, possible, in order for the trial to be conducted
protective immunity to mycobacteria is as quickly as possible. In addition, the
complex, and it is perhaps unlikely that a epidemiological data which underpin the
single, simple immunological correlate will incidence data need to be robust and recent,
be identified in such trials. It may also be that in order that the efficacy trials can be powered
254 H. McShane
with confidence. Such epidemiological studies stitute a case definition of tuberculosis. Some
in themselves take considerable time and link to the causative organism is important
resources. A more pragmatic way forward from a regulatory viewpoint. In the absence
may be to use less robust incidence data in an of a positive M. tuberculosis culture, factors
adaptive trial design, and to plan for a Data such as TST positivity, IGRA positivity (or
and Safety Monitoring Board review of the conversion) or Ziehl–Neelsen (ZN)-positive
real-time case accrual rate in the control arm sputum smears, need to be considered.
of any vaccine study, and for the sample size Newer molecular diagnostic tests such as
and/or follow-up time to be adjusted PCR-based tests offer real promise in
accordingly (Luce et al., 2009). providing specificity without a loss of
There are very few sites throughout the sensitivity. One of the important aspects to
world which are currently capable of explore in Phase IIb proof-of-concept trials is
conducting an efficacy trial with a new which end points work best and which are
tuberculosis vaccine. The limitations are due most acceptable to regulatory agencies. The
to lack of sites both with a sufficiently high challenge of defining appropriate and
(and well-defined) incidence and with a meaningful end points is not just in infant
sufficient clinical trial infrastructure for the trials. With other indications, for example
conduct of a regulatory standard, Inter- HIV-infected adults, the end points can also
national Conference on Harmonization Good be challenging, as HIV-associated tuberculosis
Clinical Practice (ICH-GCP)-compliant is also less likely to be sputum smear positive
clinical trial. It is important that, in parallel than in HIV-negative people, and HIV-
with vaccine development, a commitment is infected adults are more likely to have extra-
made to the funding and development of pulmonary disease (Sterling et al., 2010).
clinical trial sites capable of evaluating these
vaccine candidates. Such sites will necessarily
be in high-burden countries; however, it is 19.7 Conclusions
important that such sites are not limited to
sub-Saharan Africa. The BCG efficacy data The last decade has been an exciting time in
demonstrate a striking variability in efficacy tuberculosis vaccine development. After
across different geographical latitudes and it decades of inadequate funding, there are now
is important that efficacy testing of the new several vaccine candidates being evaluated in
vaccine candidates is conducted in different clinical trials and the most advanced of these
continents. are being evaluated in efficacy trials. These
proof-of-concept efficacy trials are essential
for vaccine development, and ultimately for
19.6.4 End point definitions in efficacy the registration of a new tuberculosis vaccine.
trials However, there is much more than individual
vaccine efficacy that we can learn from such
The end points in a clinical trial are critical. trials. These trials allow the opportunity to
Such end points must be clearly defined and feedback iteratively into immunological ana-
acceptable to the regulatory agencies. For lysis and preclinical animal models, to allow
efficacy trials, the end points must relate to us to establish models which are more
tuberculosis disease and the gold standard is predictive of human disease and also to
a positive culture for M. tuberculosis. For identify potential correlates within which
clinical trials in infants, this poses a problem subsequent vaccines can be evaluated.
as, in infants, most of the tuberculosis disease
diagnosed in routine clinical practice is not
culture positive (Moyo et al., 2010). Efficacy References
trials in infants therefore require end points
which are defined by a complex algorithm of Abel, B., Tameris, M., Mansoor, N., Gelderbloem,
clinical, radiological, microbiological and S., Hughes, J., Abrahams, D., et al. (2010) The
immunological factors, which together con- novel tuberculosis vaccine, AERAS-402,
induces robust and polyfunctional CD4+ and Botha, T. and Ryffel, B. (2002) Reactivation of latent
CD8+ T cells in adults. American Journal of tuberculosis by an inhibitor of inducible nitric
Respiratory and Critical Care Medicine 181, oxide synthase in an aerosol murine model.
1407–1417. Immunology 107, 350–357.
Alderson, M.R., Bement, T., Day, C.H., Zhu, L., Brandt, L., Feino Cunha, J., Weinreich Olsen, A.,
Molesh, D., Skeiky, Y.A., et al. (2000) Expression Chilima, B., Hirsch, P., Appelberg, R., et al.
cloning of an immunodominant family of (2002) Failure of the Mycobacterium bovis BCG
Mycobacterium tuberculosis antigens using vaccine: some species of environmental
human CD4(+) T cells. Journal of Experimental mycobacteria block multiplication of BCG and
Medicine 191, 551–560. induction of protective immunity to tuberculosis.
Amara, R.R., Villinger, F., Altman, J.D., Lydy, S.L., Infection and Immunity 70, 672–678.
O’Neil, S.P., Staprans, S.I., et al. (2001) Control Brandt, L., Skeiky, Y.A., Alderson, M.R., Lobet, Y.,
of a mucosal challenge and prevention of AIDS Dalemans, W., Turner, O.C., et al. (2004) The
by a multiprotein DNA/MVA vaccine. Science protective effect of the Mycobacterium bovis
292, 69–74. BCG vaccine is increased by coadministration
Andersen, P., Askgaard, D., Gottschau, A., with the Mycobacterium tuberculosis
Bennedsen, J., Nagai, S. and Heron, I. (1992) 72-kilodalton fusion polyprotein Mtb72F in M.
Identification of immunodominant antigens tuberculosis-infected guinea pigs. Infection and
during infection with Mycobacterium tubercu- Immunity 72, 6622–6632.
losis. Scandinavian Journal of Immunology 36, Brindle, T.W., Griffiths, M.I., Holme, T., Stalker, R.,
823–831. Burland, W.L., Coates, G.A., et al. (1972) A trial
Babu, S., Bhat, S.Q., Kumar, N.P., Kumaraswami, V. to compare United Kingdom strength BCG
and Nutman, T.B. (2010) Regulatory T cells vaccine with vaccine of double the moist weight
modulate Th17 responses in patients with content. Tubercle 53, 106–110.
positive tuberculin skin test results. Journal of Brookes, R.H., Pathan, A.A., McShane, H.,
Infectious Diseases 201, 20–31. Hensmann, M., Price, D.A. and Hill, A.V. (2003)
Baily, G.V. (1980) Tuberculosis prevention trial, CD8+ T cell-mediated suppression of intra-
Madras. Indian Journal of Medical Research
cellular Mycobacterium tuberculosis growth in
72(Suppl), 1–74.
activated human macrophages. European
Behr, M.A. (2002) BCG – different strains, different
Journal of Immunology 33, 3293–3302.
vaccines? The Lancet Infectious Diseases 2,
Brosch, R., Gordon, S.V., Garnier, T., Eiglmeier, K.,
86–92.
Frigui, W., Valenti, P., et al. (2007) Genome
Behr, M.A., Wilson, M.A., Gill, W.P., Salamon, H.,
plasticity of BCG and impact on vaccine efficacy.
Schoolnik, G.K., Rane, S., et al. (1999)
Proceedings of the National Academy of
Comparative genomics of BCG vaccines by
Sciences 104, 5596–5601.
whole-genome DNA microarray. Science 284,
Capuano, S.V. 3rd, Croix, D.A., Pawar, S., Zinovik,
1520–1523.
A., Myers, A., Lin, P.L., et al. (2003) Experimental
Beveridge, N.E., Price, D.A., Casazza, J.P., Pathan,
Mycobacterium tuberculosis infection of
A.A., Sander, C.R., Asher, T.E., et al. (2007)
cynomolgus macaques closely resembles the
Immunisation with BCG and recombinant
MVA85A induces long-lasting, polyfunctional various manifestations of human M. tuberculosis
Mycobacterium tuberculosis-specific CD4+ infection. Infection and Immunity 71, 5831–5844.
memory T lymphocyte populations. European Caruso, A.M., Serbina, N., Klein, E., Triebold, K.,
Journal of Immunology 37, 3089–3100. Bloom, B.R. and Flynn, J.L. (1999) Mice deficient
Black, G.F., Weir, R.E., Floyd, S., Bliss, L., in CD4 T cells have only transiently diminished
Warndorff, D.K., Crampin, A.C., et al. (2002) levels of IFN-gamma, yet succumb to
BCG-induced increase in interferon-gamma tuberculosis. Journal of Immunology 162, 5407–
response to mycobacterial antigens and 5416.
efficacy of BCG vaccination in Malawi and the Chen, C.Y., Huang, D., Wang, R.C., Shen, L., Zeng,
UK: two randomised controlled studies. The G., Yao, S., et al. (2009) A critical role for CD8 T
Lancet 359, 1393–1401. cells in a nonhuman primate model of
Bonaldo, M.C., Martins, M.A., Rudersdorf, R., tuberculosis. PLoS Pathogens 5, e1000392.
Mudd, P.A., Sacha, J.B., Piaskowski, S.M., et al. Colditz, G.A., Brewer, T.F., Berkey, C.S., Wilson,
(2010) Recombinant yellow fever vaccine virus M.E., Burdick, E., Fineberg, H.V., et al. (1994)
17D expressing simian immunodeficiency virus Efficacy of BCG vaccine in the prevention of
SIVmac239 gag induces SIV-specific CD8+ tuberculosis. Meta-analysis of the published
T-cell responses in rhesus macaques. Journal literature. Journal of the American Medical
of Virology, 84, 3699–3706. Association 271, 698–702.
256 H. McShane
Comstock, G.W. and Palmer, C.E. (1966) Long-term Fine, P.E. and Rodrigues, L.C. (1990) Modern
results of BCG vaccination in the southern vaccines. Mycobacterial diseases. The Lancet
United States. American Review of Respiratory 335, 1016–1020.
Disease 93, 171–183. Fletcher, H.A., Pathan, A.A., Berthoud, T.K.,
Corbett, E.L. and De Cock, K.M. (1996) Tuberculosis Dunachie, S.J., Whelan, K.T., Alder, N.C., et al.
in the HIV-positive patient. British Journal of (2008) Boosting BCG vaccination with MVA85A
Hospital Medicine 56, 200–204. down-regulates the immunoregulatory cytokine
Darrah, P.A., Patel, D.T., De Luca, P.M., Lindsay, TGF-beta1. Vaccine 26, 5269–5275.
R.W., Davey, D.F., Flynn, B.J., et al. (2007) Flynn, J.L., Chan, J., Triebold, K.J., Dalton, D.K.,
Multifunctional TH1 cells define a correlate of Stewart, T.A. and Bloom, B.R. (1993) An
vaccine-mediated protection against Leishmania essential role for interferon gamma in resistance
major. Nature Medicine 13, 843–850. to Mycobacterium tuberculosis infection. Journal
Davids, V., Hanekom, W.A., Mansoor, N., of Experimental Medicine 178, 2249–2254.
Gamieldien, H., Gelderbloem, S.J., Hawkridge, Flynn, J.L., Goldstein, M.M., Chan, J., Triebold, K.J.,
A., et al. (2006) The effect of bacille Calmette– Pfeffer, K., Lowenstein, C.J., et al. (1995) Tumor
Guerin vaccine strain and route of administration necrosis factor-alpha is required in the protective
on induced immune responses in vaccinated immune response against Mycobacterium
infants. Journal of Infectious Diseases 193, 531– tuberculosis in mice. Immunity 2, 561–572.
536. Forbes, E.K., Sander, C., Ronan, E.O., McShane,
Dietrich, J., Aagaard, C., Leah, R., Olsen, A.W., H., Hill, A.V., Beverley, P.C., et al. (2008)
Stryhn, A., Doherty, T.M., et al. (2005) Multifunctional, high-level cytokine-producing
Exchanging ESAT6 with TB10.4 in an Ag85B Th1 cells in the lung, but not spleen, correlate
fusion molecule-based tuberculosis subunit with protection against Mycobacterium
vaccine: efficient protection and ESAT6-based tuberculosis aerosol challenge in mice. Journal
sensitive monitoring of vaccine efficacy. Journal of Immunology 181, 4955–4964.
of Immunology 174, 6332–6339. Freudenstein, H., Pranter, W. and Schweinsberg, H.
Dietrich, J., Andersen, C., Rappuoli, R., Doherty, (1979) Assessment of several BCG vaccines in
T.M., Jensen, C.G. and Andersen, P. (2006) different animal test systems (additional studies
Mucosal administration of Ag85B-ESAT-6 to an IABS collaborative assay). Journal of
protects against infection with Mycobacterium Biological Standardisation 7, 203–212.
tuberculosis and boosts prior bacillus Calmette– Gorak-Stolinska, P., Weir, R.E., Floyd, S., Lalor,
Guerin immunity. Journal of Immunology 177, M.K., Stenson, S., Branson, K., et al. (2006)
6353–6360. Immunogenicity of Danish-SSI 1331 BCG
Draper, S.J. and Heeney, J.L. (2010) Viruses as vaccine in the UK: comparison with Glaxo-Evans
vaccine vectors for infectious diseases and 1077 BCG vaccine. Vaccine 24, 5726–5733.
cancer. Nature Reviews Microbiology 8, 62–73. Grode, L., Seiler, P., Baumann, S., Hess, J.,
Durban Immunotherapy Trial Group (1999) Immuno- Brinkmann, V., Nasser Eddine, A., et al. (2005)
therapy with Mycobacterium vaccae in patients Increased vaccine efficacy against tuberculosis
with newly diagnosed pulmonary tuberculosis: a of recombinant Mycobacterium bovis bacille
randomised controlled trial. The Lancet 354, Calmette–Guerin mutants that secrete
116–119. listeriolysin. Journal of Clinical Investigation 115,
Dworski, M. (1973) Efficacy of bacillus Calmette– 2472–2479.
Guerin and isoniazid-resistant bacillus Guiard, J., Collmann, A., Garcia-Alles, L.F., Mourey,
Calmette–Guerin with and without isoniazid L., Brando, T., Mori, L., et al. (2009) Fatty acyl
chemoprophylaxis from day of vaccination. I. structures of mycobacterium tuberculosis sulfo-
Experimental findings in guinea pigs. American glycolipid govern T cell response. Journal of
Review of Respiratory Disease 108, 294–300. Immunology 182, 7030–7037.
Feng, C.G., Kaviratne, M., Rothfuchs, A.G., Hanke, T., Samuel, R.V., Blanchard, T.J., Neumann,
Cheever, A., Hieny, S., Young, H.A., et al. (2006) V.C., Allen, T.M., Boyson, J.E., et al. (1999)
NK cell-derived IFN-gamma differentially Effective induction of simian immunodeficiency
regulates innate resistance and neutrophil virus-specific cytotoxic T lymphocytes in
response in T cell-deficient hosts infected with macaques by using a multiepitope gene and
Mycobacterium tuberculosis. Journal of DNA prime-modified vaccinia virus Ankara boost
Immunology 177, 7086–7093. vaccination regimen. Journal of Virology 73,
Fine, P. (2005) Stopping routine vaccination for 7524–7532.
tuberculosis in schools. British Medical Journal Harari, A., Vallelian, F., Meylan, P.R. and Pantaleo,
331, 647–648. G. (2005) Functional heterogeneity of memory
CD4 T cell responses in different conditions of tuberculosis DNA vaccine. Nature Medicine 2,
antigen exposure and persistence. Journal of 893–898.
Immunology 174, 1037–1045. Ibanga, H.B., Brookes, R.H., Hill, P.C., Owiafe, P.K.,
Hart, P.D. and Sutherland, I. (1977) BCG and vole Fletcher, H.A., Lienhardt, C., et al. (2006) Early
bacillus vaccines in the prevention of clinical trials with a new tuberculosis vaccine,
tuberculosis in adolescence and early adult life. MVA85A, in tuberculosis-endemic countries:
British Medical Journal 2, 293–295. issues in study design. The Lancet Infectious
Hawkridge, T., Scriba, T.J., Gelderbloem, S., Smit, Diseases 6, 522–528.
E., Tameris, M., Moyo, S., et al. (2008) Safety Johnson, J.L., Kamya, R.M., Okwera, A., Loughlin,
and immunogenicity of a new tuberculosis A.M., Nyole, S., Hom, D.L., et al. (2000)
vaccine, MVA85A, in healthy adults in South Randomized controlled trial of Mycobacterium
Africa. Journal of Infectious Diseases 198, 544– vaccae immunotherapy in non-human immuno-
552. deficiency virus-infected Ugandan adults with
Hernandez Pando, R., Aguilar, L.D., Infante, E., newly diagnosed pulmonary tuberculosis. The
Cataldi, A., Bigi, F., Martin, C., et al. (2006) The Uganda-Case Western Reserve University
use of mutant mycobacteria as new vaccines to Research Collaboration. Journal of Infectious
prevent tuberculosis. Tuberculosis (Edinburgh) Diseases 181, 1304–1312.
86, 203–210. Jones, B.E., Young, S.M., Antoniskis, D., Davidson,
Hesseling, A.C., Marais, B.J., Gie, R.P., Schaaf, P.T., Kramer, F. and Barnes, P.F. (1993)
H.S., Fine, P.E., Godfrey-Faussett, P., et al. Relationship of the manifestations of tuberculosis
(2007) The risk of disseminated Bacille to CD4 cell counts in patients with human
Calmette–Guerin (BCG) disease in HIV-infected immunodeficiency virus infection. American
children. Vaccine 25, 14–18. Review of Respiratory Disease 148, 1292–1297.
Hoft, D.F., Blazevic, A., Abate, G., Hanekom, W.A., Kagina, B.M., Abel, B., Scriba, T.J., Hughes, E.J.,
Kaplan, G., Soler, J.H., et al. (2008) A new Keyser, A., Soares, A., et al. (2010) Specific T
recombinant bacille Calmette–Guerin vaccine cell frequency and cytokine expression profile
safely induces significantly enhanced tubercu- do not correlate with protection against
losis-specific immunity in human volunteers. tuberculosis, following BCG vaccination of
Journal of Infectious Diseases 198, 1491–1501. newborns. American Journal of Respiratory and
HogenEsch, H. (2002) Mechanisms of stimulation Critical Care Medicine 182(8), 1073–1079.
of the immune response by aluminum adjuvants. Kamath, A.T., Fruth, U., Brennan, M.J., Dobbelaer,
Vaccine 20(Suppl 3), S34–39. R., Hubrechts, P., Ho, M.M., et al. (2005) New
Horwitz, M.A., Lee, B.W., Dillon, B.J. and Harth, G. live mycobacterial vaccines: the Geneva
(1995) Protective immunity against tuberculosis consensus on essential steps towards clinical
induced by vaccination with major extracellular development. Vaccine 23, 3753–3761.
proteins of Mycobacterium tuberculosis. Pro- Keane, J., Gershon, S., Wise, R.P., Mirabile-Levens,
ceedings of the National Academy of Sciences E., Kasznica, J., Schwieterman, W.D., et al.
92, 1530–1534. (2001) Tuberculosis associated with infliximab, a
Horwitz, M.A., Harth, G., Dillon, B.J. and Maslesa- tumor necrosis factor alpha-neutralizing agent.
Galic, S. (2000) Recombinant bacillus New England Journal of Medicine 345, 1098–
Calmette–Guerin (BCG) vaccines expressing 1104.
the Mycobacterium tuberculosis 30-kDa major Khader, S.A., Bell, G.K., Pearl, J.E., Fountain, J.J.,
secretory protein induce greater protective Rangel-Moreno, J., Cilley, G.E., et al. (2007)
immunity against tuberculosis than conventional IL-23 and IL-17 in the establishment of protective
BCG vaccines in a highly susceptible animal pulmonary CD4+ T cell responses after
model. Proceedings of the National Academy of vaccination and during Mycobacterium
Sciences 97, 13853–13858. tuberculosis challenge. Nature Immunology 8,
Horwitz, M.A., Harth, G., Dillon, B.J. and Maslesa- 369–377.
Galic, S. (2009) Commonly administered BCG Korbel, D.S., Schneider, B.E. and Schaible, U.E.
strains including an evolutionarily early strain (2008) Innate immunity in tuberculosis: myths
and evolutionarily late strains of disparate and truth. Microbes and Infection 10, 995–1004.
genealogy induce comparable protective Kostense, S., Koudstaal, W., Sprangers, M.,
immunity against tuberculosis. Vaccine 27, Weverling, G.J., Penders, G., Helmus, N., et al.
441–445. (2004) Adenovirus types 5 and 35
Huygen, K., Content, J., Denis, O., Montgomery, seroprevalence in AIDS risk groups supports
D.L., Yawman, A.M., Deck, R.R., et al. (1996) type 35 as a vaccine vector. AIDS 18, 1213–
Immunogenicity and protective efficacy of a 1216.
258 H. McShane
Lagranderie, M.R., Balazuc, A.M., Deriaud, E., boost setting with an adenovirus 35 tuberculosis
Leclerc, C.D. and Gheorghiu, M. (1996) Com- vaccine vector. PLoS One 3, e3790.
parison of immune responses of mice Mittrucker, H.W., Steinhoff, U., Kohler, A., Krause,
immunized with five different Mycobacterium M., Lazar, D., Mex, P., et al. (2007) Poor
bovis BCG vaccine strains. Infection and correlation between BCG vaccination-induced T
Immunity 64, 1–9. cell responses and protection against
Langermans, J.A., Andersen, P., Van Soolingen, D., tuberculosis. Proceedings of the National
Vervenne, R.A., Frost, P.A., Van der Laan, T., et Academy of Sciences 104, 12434–12439.
al. (2001) Divergent effect of bacillus Calmette– Moyo, S., Verver, S., Mahomed, H., Hawkridge, A.,
Guerin (BCG) vaccination on Mycobacterium Kibel, M., Hatherill, M., et al. (2010) Age-related
tuberculosis infection in highly related macaque tuberculosis incidence and severity in children
species: implications for primate models in under 5 years of age in Cape Town, South
tuberculosis vaccine research. Proceedings of Africa. International Journal of Tuberculosis and
the National Academy of Sciences 98, 11497– Lung Disease 14, 149–154.
11502. Negri, D.R., Michelini, Z., Baroncelli, S., Spada, M.,
Larsen, M.H., Biermann, K., Chen, B., Hsu, T., Vendetti, S., Bona, R., et al. (2010) Non-
Sambandamurthy, V.K., Lackner, A.A., et al. integrating lentiviral vector-based vaccine
(2009) Efficacy and safety of live attenuated efficiently induces functional and persistent
persistent and rapidly cleared Mycobacterium CD8+ T cell responses in mice. Journal of
tuberculosis vaccine candidates in non-human Biomedicine and Biotechnology 2010, Article
primates. Vaccine 27, 4709–4717. 534501.
Luce, B.R., Kramer, J.M., Goodman, S.N., Connor, Newport, M.J., Huxley, C.M., Huston, S.,
J.T., Tunis, S., Whicher, D., et al. (2009) Hawrylowicz, C.M., Oostra, B.A., Williamson,
Rethinking randomized clinical trials for com- R., et al. (1996) A mutation in the interferon-
parative effectiveness research: the need for gamma-receptor gene and susceptibility to
transformational change. Annals of Internal mycobacterial infection. New England Journal
Medicine 151, 206–209. of Medicine 335, 1941–1949.
McConkey, S.J., Reece, W.H., Moorthy, V.S., Orme, I.M. (1987) The kinetics of emergence and
Webster, D., Dunachie, S., Butcher, G., et al. loss of mediator T lymphocytes acquired in
(2003) Enhanced T-cell immunogenicity of response to infection with Mycobacterium
plasmid DNA vaccines boosted by recombinant tuberculosis. Journal of Immunology 138, 293–
modified vaccinia virus Ankara in humans. 298.
Nature Medicine 9, 729–735. Pai, M., Riley, L.W. and Colford, J.M. Jr (2004)
McShane, H., Brookes, R., Gilbert, S.C. and Hill, Interferon-gamma assays in the immuno-
A.V. (2001) Enhanced immunogenicity of diagnosis of tuberculosis: a systematic review.
CD4(+) t-cell responses and protective efficacy The Lancet Infectious Diseases 4, 761–776.
of a DNA-modified vaccinia virus Ankara prime- Palmer, C.E. and Long, M.W. (1966) Effects of
boost vaccination regimen for murine tubercu- infection with atypical mycobacteria on BCG
losis. Infection and Immunity 69, 681–686. vaccination and tuberculosis. American Review
McShane, H., Behboudi, S., Goonetilleke, N., of Respiratory Disease 94, 553–568.
Brookes, R. and Hill, A.V. (2002) Protective Paoletti, E. (1996) Applications of pox virus vectors
immunity against Mycobacterium tuberculosis to vaccination: an update. Proceedings of the
induced by dendritic cells pulsed with both National Academy of Sciences 93, 11349–
CD8(+)- and CD4(+)-T-cell epitopes from 11353.
antigen 85A. Infection and Immunity 70, 1623– Pathan, A.A., Sander, C.R., Fletcher, H.A., Poulton,
1626. I., Alder, N.C., Beveridge, N.E., et al. (2007)
McShane, H., Pathan, A.A., Sander, C.R., Keating, Boosting BCG with recombinant modified
S.M., Gilbert, S.C., Huygen, K., et al. (2004) vaccinia Ankara expressing antigen 85A:
Recombinant modified vaccinia virus Ankara different boosting intervals and implications for
expressing antigen 85A boosts BCG-primed efficacy trials. PLoS ONE 2, e1052.
and naturally acquired antimycobacterial Prevec, L., Schneider, M., Rosenthal, K.L., Belbeck,
immunity in humans. Nature Medicine 10, L.W., Derbyshire, J.B. and Graham, F.L. (1989)
1240–1244. Use of human adenovirus-based vectors for
Magalhaes, I., Sizemore, D.R., Ahmed, R.K., antigen expression in animals. Journal of
Mueller, S., Wehlin, L., Scanga, C., et al. (2008) General Virology 70(Pt 2), 429–434.
rBCG induces strong antigen-specific T cell Pym, A.S., Brodin, P., Brosch, R., Huerre, M. and
responses in rhesus macaques in a prime- Cole, S.T. (2002) Loss of RD1 contributed to the
attenuation of the live tuberculosis vaccines tuberculosis vaccine, is safe in adolescents and
Mycobacterium bovis BCG and Mycobacterium children, and induces polyfunctional CD4+ T
microti. Molecular Microbiology 46, 709–717. cells. European Journal of Immunology 40,
Pym, A.S., Brodin, P., Majlessi, L., Brosch, R., 279–290.
Demangel, C., Williams, A., et al. (2003) Shen, Y., Zhou, D., Qiu, L., Lai, X., Simon, M.,
Recombinant BCG exporting ESAT-6 confers Shen, L., et al. (2002) Adaptive immune
enhanced protection against tuberculosis. response of Vgamma2Vdelta2+ T cells during
Nature Medicine 9, 533–539. mycobacterial infections. Science 295, 2255–
Radosevic, K., Wieland, C.W., Rodriguez, A., 2258.
Weverling, G.J., Mintardjo, R., Gillissen, G., et Skeiky, Y.A., Alderson, M.R., Ovendale, P.J.,
al. (2007) Protective immune responses to a Guderian, J.A., Brandt, L., Dillon, D.C., et al.
recombinant adenovirus type 35 tuberculosis (2004) Differential immune responses and
vaccine in two mouse strains: CD4 and CD8 protective efficacy induced by components of a
T-cell epitope mapping and role of gamma tuberculosis polyprotein vaccine, Mtb72F,
interferon. Infection and Immunity 75, 4105– delivered as naked DNA or recombinant protein.
4115. Journal of Immunology 172, 7618–7628.
Ritz, N. and Curtis, N. (2009) Mapping the global Skeiky, Y.A., Dietrich, J., Lasco, T.M., Stagliano, K.,
use of different BCG vaccine strains. Tubercu- Dheenadhayalan, V., Goetz, M.A., et al. (2010)
losis (Edinburgh) 89, 248–251. Non-clinical efficacy and safety of HyVac4:IC31
Rodrigues, L.C., Diwan, V.K. and Wheeler, J.G. vaccine administered in a BCG prime-boost
(1993) Protective effect of BCG against regimen. Vaccine 28, 1084–1093.
tuberculous meningitis and miliary tuberculosis: Smith, D., Harding, G., Chan, J., Edwards, M.,
a meta-analysis. International Journal of Hank, J., Muller, D., et al. (1979) Potency of 10
Epidemiology 22, 1154–1158. BCG vaccines as evaluated by their influence
Roupie, V., Romano, M., Zhang, L., Korf, H., Lin, on the bacillemic phase of experimental
M.Y., Franken, K.L., et al. (2007) Immunogenicity airborne tuberculosis in guinea-pigs. Journal of
of eight dormancy regulon-encoded proteins of Biological Standardisation 7, 179–197.
Mycobacterium tuberculosis in DNA-vaccinated Sousa, A.O., Mazzaccaro, R.J., Russell, R.G., Lee,
and tuberculosis-infected mice. Infection and F.K., Turner, O.C., Hong, S., et al. (2000)
Immunity 75, 941–949. Relative contributions of distinct MHC class
Sallusto, F., Geginat, J. and Lanzavecchia, A. I-dependent cell populations in protection to
(2004) Central memory and effector memory T tuberculosis infection in mice. Proceedings of
cell subsets: function, generation, and mainten- the National Academy of Sciences 97, 4204–
ance. Annual Reviews of Immunology 22, 745– 4208.
763. Spencer, C.T., Abate, G., Blazevic, A. and Hoft, D.F.
Sander, C.R., Pathan, A.A., Beveridge, N.E., (2008) Only a subset of phosphoantigen-
Poulton, I., Minassian, A., Alder, N., et al. (2009) responsive gamma9delta2 T cells mediate
Safety and immunogenicity of a new tubercu- protective tuberculosis immunity. Journal of
losis vaccine, MVA85A, in Mycobacterium Immunology 181, 4471–4484.
tuberculosis-infected individuals. American Stanford, J.L., Bahr, G.M., Rook, G.A., Shaaban,
Journal of Respiratory and Critical Care M.A., Chugh, T.D., Gabriel, M., et al. (1990)
Medicine 179, 724–733. Immunotherapy with Mycobacterium vaccae as
Scanga, C.A., Mohan, V.P., Joseph, H., Yu, K., an adjunct to chemotherapy in the treatment of
Chan, J. and Flynn, J.L. (1999) Reactivation of pulmonary tuberculosis. Tubercle 71, 87–93.
latent tuberculosis: variations on the Cornell Stenger, S., Hanson, D.A., Teitelbaum, R., Dewan,
murine model. Infection and Immunity 67, P., Niazi, K.R., Froelich, C.J., et al. (1998) An
4531–4538. antimicrobial activity of cytolytic T cells mediated
Schneider, J., Gilbert, S.C., Blanchard, T.J., Hanke, by granulysin. Science 282, 121–125.
T., Robson, K.J., Hannan, C.M., et al. (1998) Sterling, T.R., Pham, P.A. and Chaisson, R.E.
Enhanced immunogenicity for CD8+ T cell (2010) HIV infection-related tuberculosis: clini-
induction and complete protective efficacy of cal manifestations and treatment. Clinical
malaria DNA vaccination by boosting with Infectious Disease 50(Suppl 3), S223–230.
modified vaccinia virus Ankara. Nature Medicine Sun, R., Skeiky, Y.A., Izzo, A., Dheenadhayalan, V.,
4, 397–402. Imam, Z., Penn, E., et al. (2009) Novel
Scriba, T.J., Tameris, M., Mansoor, N., Smit, E., Van recombinant BCG expressing perfringolysin O
der Merwe, L., Isaacs, F., et al. (2010) Modified and the over-expression of key immunodominant
vaccinia Ankara-expressing Ag85A, a novel antigens; pre-clinical characterization, safety
260 H. McShane
and protection against challenge with inactivated whole-cell mycobacterial vaccine.

Mycobacterium tuberculosis. Vaccine 27, 4412– AIDS 24, 675–685.
4423. Vordermeier, H.M., Villarreal-Ramos, B., Cockle,
Tascon, R.E., Stavropoulos, E., Lukacs, K.V. and P.J., McAulay, M., Rhodes, S.G., Thacker, T., et
Colston, M.J. (1998) Protection against Myco- al. (2009) Viral booster vaccines improve
bacterium tuberculosis infection by CD8+ T Mycobacterium bovis BCG-induced protection
cells requires the production of gamma against bovine tuberculosis. Infection and
interferon. Infection and Immunity 66, 830–834. Immunity, 77, 3364–3373.
Tatsis, N. and Ertl, H.C. (2004) Adenoviruses as Walker, K.B., Brennan, M.J., Ho, M.M., Eskola, J.,
vaccine vectors. Molecular Therapy 10, 616– Thiry, G., Sadoff, J., et al. (2010) The second
629. Geneva Consensus: recommendations for
Taylor, J.L., Turner, O.C., Basaraba, R.J., Belisle, novel live TB vaccines. Vaccine 28, 2259–2270.
J.T., Huygen, K. and Orme, I.M. (2003) Pulmon- Wedlock, D.N., Denis, M., Vordermeier, H.M.,
ary necrosis resulting from DNA vaccination Hewinson, R.G. and Buddle, B.M. (2007)
against tuberculosis. Infection and Immunity 71, Vaccination of cattle with Danish and Pasteur
2192–2198. strains of Mycobacterium bovis BCG induce
Trunz, B.B., Fine, P. and Dye, C. (2006) Effect of different levels of IFNgamma post-vaccination,
BCG vaccination on childhood tuberculous but induce similar levels of protection against
meningitis and miliary tuberculosis worldwide: a bovine tuberculosis. Veterinary Immunology
meta-analysis and assessment of cost- and Immunopathology 118, 50–58.
effectiveness. The Lancet 367, 1173–1180. Weir, R.E., Black, G.F., Dockrell, H.M., Floyd, S.,
Van Dissel, J.T., Arend, S.M., Prins, C., Bang, P., Fine, P.E., Chaguluka, S.D., et al. (2004)
Tingskov, P.N., Lingnau, K., et al. (2010) Ag85B- Mycobacterial purified protein derivatives
ESAT-6 adjuvanted with IC31 promotes strong stimulate innate immunity: Malawians show
and long-lived Mycobacterium tuberculosis enhanced tumor necrosis factor alpha,
specific T cell responses in naive human interleukin-1beta (IL-1beta), and IL-10
volunteers. Vaccine 28, 3571–3581. responses compared to those of adolescents in
Van Pinxteren, L.A., Cassidy, J.P., Smedegaard, the United Kingdom. Infection and Immunity 72,
B.H., Agger, E.M. and Andersen, P. (2000) 1807–1811.
Control of latent Mycobacterium tuberculosis Whelan, K.T., Pathan, A.A., Sander, C.R., Fletcher,
infection is dependent on CD8 T cells. European H.A., Poulton, I., Alder, N.C., et al. (2009)
Journal of Immunology 30, 3689–3698. Safety and immunogenicity of boosting BCG
Verreck, F.A., Vervenne, R.A., Kondova, I., Van vaccinated subjects with BCG: comparison with
Kralingen, K.W., Remarque, E.J., Braskamp, boosting with a new TB vaccine, MVA85A.
G., et al. (2009) MVA.85A boosting of BCG and PLoS One 4, e5934.
an attenuated, phoP deficient M. tuberculosis WHO (2007) Weekly Epidemiological Record 3,
vaccine both show protective efficacy against 17–24.
tuberculosis in rhesus macaques. PLoS One 4, Williams, A., Goonetilleke, N.P., McShane, H.,
e5264. Clark, S.O., Hatch, G., Gilbert, S.C., et al.
Vilaplana, C., Montane, E., Pinto, S., Barriocanal, (2005a) Boosting with poxviruses enhances
A.M., Domenech, G., Torres, F., et al. (2010) Mycobacterium bovis BCG efficacy against
Double-blind, randomized, placebo-controlled tuberculosis in guinea pigs. Infection and
Phase I Clinical Trial of the therapeutical Immunity 73, 3814–3816.
antituberculous vaccine RUTI. Vaccine 28, Williams, A., Hatch, G.J., Clark, S.O., Gooch, K.E.,
1106–1116. Hatch, K.A., Hall, G.A., et al. (2005b) Evaluation
Von Eschen, K., Morrison, R., Braun, M., Ofori- of vaccines in the EU TB Vaccine Cluster using
Anyinam, O., De Kock, E., Pavithran, P., et al. a guinea pig aerosol infection model of
(2009) The candidate tuberculosis vaccine tuberculosis. Tuberculosis (Edinburgh) 85,
Mtb72F/AS02A: tolerability and immunogenicity 29–38.
in humans. Human Vaccines 5, 475–482. Xiang, Z., Li, Y., Cun, A., Yang, W., Ellenberg, S.,
Von Reyn, C.F., Mtei, L., Arbeit, R.D., Waddell, R., Switzer, W.M., et al. (2006) Chimpanzee
Cole, B., Mackenzie, T., et al. (2010) Prevention adenovirus antibodies in humans, sub-Saharan
of tuberculosis in Bacille Calmette–Guerin- Africa. Emerging Infectious Diseases 12,
primed, HIV-infected adults boosted with an 1596–1599.
Index
Acid-fast bacilli (AFB) 3, 5, 69, 74, 75, 130 cytotoxic effects, HS-27 human fibroblast cells
Active pulmonary tuberculosis 215
advantages and disadvantages, molecular FAS enzymes 214
diagnosis 27 genetic validation experiments 206
anti-M. tuberculosis drugs 15 in vitro activity and molecular properties 213
biosensor 27 modelling studies 211
diagnostics pipeline, 2011 13, 14 pharmacodynamic (PD) and pharmacokinetic
isoniazid (INH) and ethambutol (EMB) 15 (PK) considerations 215
isothermal methods 26–27 and SARs 206–207
LPAs see Line probe assays stages, drug development 214
MTC 15 validated drug targets see Drug targets,
MTD test 13 mycobacterial killing
Mycobacterium bovis 16 Amplicor Mycobacterium tuberculosis (MTB)
NAATs see Nucleic acid amplification tests 15, 16
(NAATs) Amplified M. tuberculosis direct (AMTD) test 16,
nanobiosensor 27 17
POC 25–26 Analysis of covariance (ANCOVA) approach 152
rifampicin (RIF) 15 Animal models, tuberculosis
sequencing-based diagnostic methods 25 advantages 68
sputum smear microscopy 13 embryonic zebrafish 68–69
AERAS-402 251 mouse 69
AFB see Acid-fast bacilli (AFB) non-human primate models 69
AG see Aminoglycosides (AG) Antibiotic susceptibility
Agar proportion method AG and macrocyclic polypeptide antibiotics
drug resistance 102 116–118
drug susceptibility tests 101 drug-resistant tuberculosis 108–109
HSTB agar 101 ethambutol 114–115
isoniazid (INH) 101 ethionamide (ETA) 112–114
MDR 101 fluoroquinolones (FQs) 115–116
pyrazinamide (PZA) 101–102 isoniazid (INH) 110–112
Aminoglycosides (AG) 116–118 PAS 116
2-Aminothiazole-4-carboxylates (ATCs) pyrazinamide (PZA) 114
anti-tubercular activity 206 rifampicin (RIF) 109–110
261
262 Index
Antibiotic tolerance vs. sputum phenotypes Bacterial load see Monitoring therapy, bacterial
latent infection 188 load
LBs Bactericidal activity 148
aerobic growth 188 Bacteriological vs. nonbacteriological biomarkers
CFU 189 149–150
clinical responses 189 BAL see Bronchoalveolar lavage (BAL)
multiple-stress dormancy model 188 BCG see Bacillus Calmette–Guérin (BCG)
NO stimulation 188 Bronchoalveolar lavage (BAL)
Rpf dependency 189–190 alveolar macrophages 71
Automated liquid culture–LJ slope vs. blood 76
acid-fast bacteria 34 CD4+ population 75
BacT/ALERT systems 41–42 cytokine concentrations 72
BACTEC™ 9000MB 42 immune response in human tuberculosis 70
BACTEC™ MGIT™ 40–41 miliary tuberculosis 72
Centre for Disease Control (CDC) 34 neutrophils and macrophages, proportion 70
clinical specimens 35 PBMCs 75
drug susceptibility 42–43 saline 69
growth-based methods 35 sensitivity 75
laboratory diagnosis 34 spontaneous sputum smear-negative 75
liquid media 38–39 T-Spot TBT 75
manual culture systems TST-positive community 72
BBL™ MGIT™ tube 39–40 tuberculosis diagnosis 70
Septi-Chek™ AFB 39
Myco-ESP Culture System II 41 CFU see Colony-forming units (CFU)
non-radiometric broth technologies 35 Chemotherapy
semi-automated 40 antibiotic tolerance vs. sputum phenotypes
solid media see Solid media 187–190
bacillary replication 184
Bacillus Calmette–Guérin (BCG) multidrug regimen 184
antigen selection 249 qRT-PCR 190
ELISA 50 sputum phenotypes 185–187
IGRA 48 therapeutic agents 184
immunization 249 Chronic obstructive pulmonary disease (COPD)
latent infection 249 70
protein/adjuvant combinations 251–252 Colony-forming units (CFU) 141, 189
recombinant strains 248–249 COPD see Chronic obstructive pulmonary disease
recombinant viral vectors 250–251 (COPD)
regulatory pathway 248 Culture methods, bacterial load
RUTI 252–253 liquid systems 141
SCID 248 measurement 140
T-cell response 249 Rpfs 141–142
TST 48 selective growth media 140–141
vaccinated individuals 51, 55
vaccination Differential mobility spectrometry (DMS) 88
children 68 Directly observed therapy short-course (DOTS)
infancy 245 95, 96, 108
M. tuberculosis 73 Direct sputum smear microscopy
non-tuberculous mycobacteria 73 AFB 3
skin 75 EQA 4
TST 74 HPF 4
BacT/ALERT mycobacteria detection systems MDR-TB/XDR-TB 4
41–42 public health problem 4–5
BACTEC™ 9000MB 42 ZN stained smears 4
BACTEC™ MGIT™ 40–41 DMS see Differential mobility spectrometry (DMS)
BACTEC MGIT 960 system 102 DOTS see Directly observed therapy short-course
BACTEC 460 system 103 (DOTS)
Index 263
Drug resistance enzymatic assay vs. whole cell assay 212–214

active drug efflux pumps 126 FAS-I and FAS-II 207
anti-tuberculosis agents 126 methyl 2-amino-5-benzylthiazole-4-carboxylate.
chromosomal mutations 126 208
codon 315 resistant strains 127–128 methyl 2-(2-bromoacetamido)-5-(3-
DNA sequencing 127 chlorophenyl)thiazole-4-carboxylate 208
genetic mechanisms 127 mycolic acid biosynthesis 207
LPA 127 target-based approach, mtFabH 209–212
molecular tests 127 thiolactomycin (TLM) 207, 208
multiple resistance 128 DST see Drug susceptibility testing (DST)
PCR-SSCP 127
RIF resistance-causing mutations 127 Early bactericidal activity (EBA)
and tolerance defined 151
antibiotics 178 INH 151
differential expression, genes 178 RIF 151
drug-sensitive bacilli 177–178 statistical approach, analysis 152
functional characterization 178 EBA see Early bactericidal activity (EBA)
mycobacterial efflux system 177–178 Electronic nose see E-noses
protein–protein interaction networks 178 ELISpot see Enzyme-linked immunosorbent spot
resistant phenotype 177 (ELISpot)
Drug-resistant tuberculosis ELISpot assay 73, 74, 76
de novo 108 Empirical vs. model-based approaches 150–151
DOTS 108 E-noses
public health concern 109 animal olfaction 85
streptococcal infection 109 animals and humans infection 86
Drug susceptibility testing (DST) application 86
Agar proportion method see Agar proportion complex responses 85
method compounds 85
antimicrobial therapy 95 defnition 84
critical concentrations 97, 98 discernable differences 86
degradation and deterioration 99 making instruments 86
DOTS 95, 96 metal oxide sensor 85
DST 96 methodological strategies 84
growth inhibition 99 sensors 84–85
liquid media 97 Enzymatic assay vs. whole cell assay
Löwenstein–Jensen medium see 2-aminothiazole-4-carboxylates 213
Löwenstein–Jensen medium Cys112 residue 212
MDR and XDR 97 electrophilic bromomethyl substituent 212
MIC 99 flexibility, ligand 212
M. tuberculosis 42 zwitterions 214
NTM 42 Enzyme-linked immunosorbent assay (ELISA)
qualitative indirect test culture-confirmed active tuberculosis 49
BACTEC MGIT 960 system 102 and ELISpot see Enzyme-linked immunosorbent
liquid medium 102 spot (ELISpot)
MB/BacT system 102–103 HIV-positive populations 53
MIC test 103 sensitivity 52
non-radiometric automated liquid medium Enzyme-linked immunosorbent spot (ELISpot)
systems 103 culture-confirmed tuberculosis 50
VersaTREK system 102 ELISA positivity 51
semi-automatic BACTEC™ 460TB and the IMID 54
BACTEC™ 960TB systems 42–43 immunocompetent individuals 49
solid media 97 positive /negative 56
susceptible and resistant strains 99 T cells 48
tuberculosis (TB) drugs 97 TST 51
WHO 96 tuberculosis- and HIV-coinfected individuals 52
Drug targets, mycobacterial killing EPI see Expanded Programme on Immunization
2-aminothiazole-4-carboxylates 214–215 (EPI)
264 Index
EQA see External quality assessment (EQA) High-power microscopic fields (HPF) 4, 5
ETA see Ethionamide (ETA) HPF see High-power microscopic fields (HPF)
Ethambutol Human olfactory detection
description 114–115 animals 82–83
resistance, molecular genetics 115 e-noses see E-noses
Ethionamide (ETA) and e-nose sensing 83
description 112 environmental influences 84
resistance, molecular genetics 112–114 groundbreaking work 83
Expanded Programme on Immunization (EPI) insect’s life 84
243 intrinsic risk 82
External quality assessment (EQA) 4 mammalian 83
Extremely drug-resistant (XDR) non-tuberculosis patients 84
drug susceptibility testing 97, 105 organic compounds 84
smear-negative specimens 104 problems 82
tuberculosis (TB) 96 rats 84
Extremely drug-resistant tuberculosis (XDR-TB) 4, ZN light microscopy 83
234–235 Human studies, M. tuberculosis in lung
BAL 73
FAS see Fatty acid synthesis (FAS) blood samples 71
Fatty acid synthesis (FAS) cavitation 73
FAS-I 207 interferon-gamma knockout mouse models 73
FAS-II 207 non-human primate model 73
mtFabH inhibitor 214 non-TLR 72
FIND see Foundation for Innovative New PPD 72
Diagnostics (FIND) pulmonary tuberculosis 72–73
Fluorescence microscopy (FM) 5 TLRs 71–72
Fluoroquinolones (FQs)
description 115 IFN- see Interferon- gamma (IFN-)
resistance, molecular genetics 115–116 IGRAs see Interferon-gamma release assays
FM see Fluorescence microscopy (FM) (IGRAs)
Foundation for Innovative New Diagnostics In vitro models, whole cell screening
(FIND) 128 gene expression 179
FQs see Fluoroquinolones (FQs) modelling conditions 179
natural infection 178
Gas chromatography-mass spectrometry (GC-MS) nutrient starvation 179
88 pyrimidine-imidazole compounds 179
GC-MS see Gas chromatography-mass IMIDs see Immune-mediated inflammatory
spectrometry (GC-MS) diseases (IMIDs)
Gene expression measurement, qPCR Immune diagnosis, lung
gene pathways 164 AFB 74
molecular-based quantification, bacterial load BAL 74–75
165–166 BCG vaccination 73
molecular techniques 166–169 ELISpot assay 73
quantification, cell membranes 164 IGRAs 74
16S rRNA, patient sputum samples 169–171 multi-centre study, European centres 75
GeneXpert NAAT 75
molecular detection 132 PBMCs 75
nucleic acid amplification 132 post hoc ROC analysis 76
performance characteristics 132 PPD 74
real-time and reverse transcriptase PCR 132 spontaneous sputum smear-negative/non-
GenoType® M. tuberculosis drug-resistance producing cases 75
second-line (MTBDRsl) assay TST 73
GeneXpert system 132 Immune-mediated inflammatory diseases (IMIDs)
patient therapy 132 53–54, 57
PCR 132 Immune responses in lung
GTC see Guanidine thiocyanate (GTC) animal models see Animal models, tuberculosis
Guanidine thiocyanate (GTC) 167 BAL 69–70
Index 265
BCG 68 NAD 111

biopsy specimens 69, 70 resistance, molecular genetics 111–112
bronchoscopy 69 and RIF 157
description 67
fibre-optic bronchoscopy 70 Joint modelling
full-effector T-cell differentiation 76 characterization, M. tuberculosis 157
human studies see Human studies, M. mechanistic modelling approaches 158
tuberculosis in lung pharmacokinetic–pharmacodynamic index 157
IGRA studies 76
immune diagnosis see Immune diagnosis, lung -Lactam antibiotics 180
immunology 67 LAMP see Loop-mediated isothermal
initial infection 67–68 amplification (LAMP)
LTBI 68 Latent tuberculosis infection (LTBI)
Lung FluteT 71 American Thoracic Society (ATS)/Centers for
minor adverse effects 70 Disease Control (CDC) 47
M. tuberculosis 67 HIV 47–48
multiple chemokines and cytokines 68 IGRAs see Interferon-gamma release assays
physical barrier 67 (IGRAs)
sputum induction 70–71 Lalvani ELISpot 60
surgical resection 69 Mycobacterium tuberculosis 46
INH see Isoniazid (INH) PPD 48
INH exposure smear-positive infectious 46
drug-mediated enzyme inhibition 175 T-cell IGRAs 48
FAS-II 175 treatment monitoring 59–60
mapping, transcriptional responses 175–176 TST 48
M. tuberculosis bacilli 175, 176 WHO’s policy 47
regulatory networks 176 LBs see Lipid bodies (LBs)
Interferon- gamma (IFN-) Ligand design, mtFabH
blood 74 H-bonding network 209
expression 73 molecular modelling software package, GOLD
lymphocytic immune response 74 209
production 71 optimization 211
receptor deficiency 73 thiazoles 211
synthesis 74 transacylation mechanism 209, 210
Interferon-gamma release assays (IGRAs) Linear discriminant function (LDF) analysis 86
active tuberculosis Line probe assays (LPAs)
BCG 50–51 AFB 130
children 51 detection, mutants 129
ELISA 49–51 direct hybridization assays 25
ELISpot 49–51 drug resistance and MTC 23–24
host-bacteria interaction 50 GenoType® MTBDRsl assay 131–132
LTBI 48–49 MTBDRplus test 130–131
TST 50 MTC 25
BCG-vaccinated individuals 55 resistance screening 129
cell-mediated immunity 52 Lipid bodies (LBs)
clinical utility 57, 58 acid-fast bacilli 185, 186
ELISpot and ELISA 48, 49 actinobacteria 185
HIV-infected individuals 52–53 antibiotic tolerance vs. sputum phenotypes see
IMIDs 53–54 Antibiotic tolerance vs. sputum phenotypes
indeterminate results 55 chronic murine infection 185
International Public Health Guidelines 57–59 multiple-stress model, dormancy 186
Mycobacterium tuberculosis 55 stimulation, DosR regulon 186
predictive value, active tuberculosis 55–57 Liquid media 38–39
T-cell 48 LMICs see Low- and middle-income countries
Isoniazid (INH) (LMICs)
activity 151 Loop-mediated isothermal amplification (LAMP)
description 110 9, 17, 18
266 Index
Low- and middle-income countries (LMICs) 3, 5, 7 reversed line blot method 129
Löwenstein–Jensen (LJ) medium reverse hybridization-based assays 129
cultivation and drug susceptibility testing 99 whole genome sequencing techniques 129
drug-containing egg-based medium 100 Molecular techniques
H37Rv strain 101 extraction, RNA and inhibitors 168–169
MIC 100–101 identification, bacteria 166
RR method 100 macrophages 166–167
starch-free LJ medium 100 RNA and DNA 166
Löwenstein–Jensen (LJ) slope see Automated sensitivity, nucleic acid-based detection 167
liquid culture–LJ slope sputum RNA 167–168
LPAs see Line probe assays (LPAs) Molecular tools, bacterial load
LTBI see Latent tuberculosis infection (LTBI) mRNA methods 142–143
PCR 142
Macrocyclic polypeptide antibiotics 116–118 Monitoring therapy, bacterial load
Manual culture systems culture methods 140–142
BBL™ MGIT™ tubes 39–40 immunological measurements 143
Septi-Chek™ AFB 39 molecular tools 142–143
MB/BacT system 102–103 smear microscopy 139–140
MDR see Multidrug-resistant (MDR) MPC see Mutant prevention concentration
MDR-TB see Multidrug-resistant M. tuberculosis (MPC)
(MDR-TB) mRNA see Messenger RNA (mRNA)
Messenger RNA (mRNA) MTBDRplus test
detection, TB 168 distribution, drug-resistance mutations 131
methods 142–143 growth inhibition tests 131
MGIT see Mycobacterial Growth Indicator Tube molecular tests 131
(MGIT) resistance gene sequencing 130
MIC see Minimum inhibitory concentration (MIC) whole genome sequencing 131
Minimum inhibitory concentration (MIC) 99, 100, wild-type probe 130
103, 157, 175, 221–222, 236 MTC see Mycobacterium tuberculosis complex
MIRU see Mycobacterial intergenic repeat unit (MTC)
(MIRU) MTD see Mycobacterium tuberculosis direct (MTD)
Mixed effects modelling mtFabH
molecular markers 156–157 chemical synthesis 211, 212
serial sputum colony-counting studies 152–155 ligand design 209–211
Modelling responses, TB treatment Multidrug-resistant (MDR)
drug development 147 agar medium 101
EBA see Early bactericidal activity (EBA) drug susceptibility testing 97
and joint modelling 157–158 smear-negative specimens 104
mixed effects see Mixed effects modelling tuberculosis (TB) 96
outcomes, anti-tuberculosis treatment 147, 148 Multidrug-resistant M. tuberculosis (MDR-TB)
RIF and PZA 148 acquired resistance 125
SM 147 antibiotics, therapy 232–233
‘sterilizing activity’ 148–149 chemotherapy regimens 125
surrogate end points 149–151 clinical and laboratory practice 200
Modified vaccinia virus Ankara (MVA) DOTS 125
BCG-induced protection, animal models 250 drug resistance 126–128
cell-mediated immune response 250 global public health challenges 126
cellular immune response. 251 LPAs 129–132
intracellular cytokine staining 250 MDR-TB 232
Molecular diagnostic assays molecular and genomic research 133
direct sequencing 129 molecular diagnostic tools see Molecular
DST 128 diagnostic assays
FIND 128 role, efflux pumps 234
MTBDRsl 128 semi-automated real-time PCR 132–133
NAAT 128 sputum culture 200
nsSNPs 128 treatment, pulmonary tuberculosis 126
Index 267
tuberculosis control 125 PBMCs see Peripheral blood mononuclear cells

Mutant prevention concentration (MPC) 228 (PBMCs)
Mycobacteria growth indicator tube (MGIT™) PCR see Polymerase chain reaction (PCR)
system 141 PCR single-stranded conformational
Mycobacterial Growth Indicator Tube (MGIT) 7 polymorphism (PCR-SSCP) 127
Mycobacterial intergenic repeat unit (MIRU) Peripheral blood mononuclear cells (PBMCs)
200 BAL cells 72, 75
Mycobacterium tuberculosis M. tuberculosis Antigen 85 72
characteristics 157 sensitivity 75
evolution 147–148 Pharmacokinetics
metabolic map 158 RBT 227
moxi-/gatifloxacin arms 154 RIF 222–224
physiology 157 rifapentine 225
Mycobacterium tuberculosis complex (MTC) Phenothiazines
direct hybridization assays 25 antimicrobial activity 235
drug resistance 18–20 anti-tuberculosis agent 235
LPAs 23–25 chlorpromazine (CPZ) 235
NAATs 16–18, 21–23 compassionate therapy 237–238
natural hosts 15 inhibitor, efflux pumps 237
polymerase chain reaction (PCR) 21–23 intrinsic drug resistance 238
tuberculosis transmission 21 macrophage targeting 236–237
Mycobacterium tuberculosis direct (MTD) 13 safety surveillance 236
Myco-ESP Culture System II 41 Point-of-care (POC) test 9, 25–26
Polymerase chain reaction (PCR)
NAATs see Nucleic acid amplification tests optical density ratio 142
(NAATs) real-time system 142
Non-synonymous single nucleotide reverse transcriptase 142–143
polymorphisms (nsSNPs) 128 PPD see Purified protein derivative (PPD)
Non-tuberculous mycobacteria (NTM) 42, 202, Purified protein derivative (PPD)
244 challenges 72
nsSNPs see Non-synonymous single nucleotide intracellular cytokine staining 74
polymorphisms (nsSNPs) M. tuberculosis Antigen 85 72
NTM see Non-tuberculous mycobacteria pilot study 72
(NTM) stimulating antigen 75
Nucleic acid amplification tests (NAATs) subcutaneous/intradermal injection 73
Amplicor MTB test 16 Pyrazinamide (PZA)
AMTD assay 17 description 114
array-based commercial tests 20–21 resistance, molecular genetics 114
drug resistance and MTC 18–20 and RIF 148
LAMP assay 17–18 PZA see Pyrazinamide (PZA)
MTC 21–23
polymerase chain reaction (PCR) 16 Quantitative PCR (qPCR) see Gene expression
pulmonary tuberculosis 16 measurement, qPCR
resistance-determining genes 16 Quantitative reverse transcription PCR (qRT-PCR)
SDA 17 190
smear positive/negative samples 18
Recombinant viral vectors
Optimized sputum smear microscopy AERAS-402 251
fluorescence microscopy 5 antivector immunity 250
serial examination 5 cellular immunity 250
specimen processing 5–6 cloning, antigens 250
MVA85A 250–251
Para-amino-salicylic acid (PAS) pulmonary disease 250
description 116 Resistance ratio (RR) method 100
resistance, molecular genetics 116 Restrictive fragment length polymorphism (RFLP)
PAS see Para-amino-salicylic acid (PAS) 200
268 Index
Resuscitation-promoting factors (Rpfs) pharmacokinetics 225

antibiotic tolerance vs. sputum phenotypes PK parameters 226
189–190 RIF resistance-determining region (RRDR) 220
chronic infection, mice 186 RNA polymerase (RNAP) 219–220
dependency, sputum 187 RNA profiling 179
description 186 Rpfs see Resuscitation-promoting factors (Rpfs)
environmental signals in vivo 187 RRDR see RIF resistance-determining region
host immune system 187 (RRDR)
M. luteus 186 rRNA (16S) assay
morphological changes 187 and liquid culture 169, 171
muralytic activity 186–187 mathematical modelling, bacterial decline 169,
site-directed mutagenesis 187 170
Reverse transcription quantitative PCR (RT-qPCR) RT-qPCR see Reverse transcription quantitative
164, 167, 168 PCR (RT-qPCR)
RFLP see Restrictive fragment length RUTI
polymorphism (RFLP) induction, antigen-specific IFN- 252–253
Ribosomal RNA (rRNA) see rRNA (16S) assay therapeutic vaccine 252
RIF see Rifampicin (RIF)
Rifabutin (RBT) SARs see Structure-activity relationships (SARs)
adverse events 227 SCID see Severe combined immunodeficiency
clinical efficacy 226–227 (SCID)
description 226 SDA see Strand displacement amplification (SDA)
microbiological activity 226 Selected ion flow tube mass spectrometry (SIFT-
pharmacokinetics 227 MS) 89
PK parameters 227 Semi-automated culture systems 40
Rifalazil 227 Semi-automated real-time PCR see GeneXpert
Rifampicin (RIF) Serial sputum colony-counting studies
adverse events 224 and fitted model predictions 153
clinical efficacy 222 liquid culture 156
description 109, 221 moxi/gatifloxacin arms 154
DNA-dependent RNA-polymerase (RNAPol) pharmacokinetic-pharmacodynamic 155
110 SHRZ regimen 153
EBA studies 151 Severe combined immunodeficiency (SCID) 248
microbiological activity 221–222 SHRZ see Streptomycin-isoniazid-rifampicin-
pharmacokinetics pyrazinamide (SHRZ)
concentration-dependent activity 223 SIFT-MS see Selected ion flow tube mass
drug exposure 223 spectrometry (SIFT-MS)
non-linear increase 224 SM see Streptomycin (SM)
and PZA 148 Smear microscopy
resistance, molecular genetics 110 ‘conversion rate’ 140
Rifamycins defined 139
Amycolatopsis rifamycinica 219 preparation 140
microbiological activity 221–222 Solid media
N-amino-N´-methylpiperazine 219 agar-based medium 37–38
pharmacokinetics 222–224 egg-based medium 35–37
RBT 226–227 Sputum phenotypes
RIF 221–224 vs. antibiotic tolerance see Antibiotic tolerance
rifalazil 227 vs. sputum phenotypes
rifapentine 224–226 chronic murine infection 185
RNAP 219–220 DosR regulon 185
RRDR 220 fluorescence microscopy 185
tuberculosis drug discovery 219 LBs 185–186
Rifapentine microarray analysis 185
adverse events 225–226 nutrient starvation 185
clinical efficacy 225 Rpf 186–187
description 224–225 Sputum serial colony-counting (SSCC) assay
microbiological activity 225 165–166, 169
Index 269
Sputum smear microscopy Transcriptomic approaches, drug therapy

diagnostic testing 7 antimycobacterial agents 173
direct see Direct sputum smear microscopy anti-tuberculosis drugs 173
LAMP assay 9 applications, M. tuberculosis transcriptional
liquid culture 7 profiling 174
LMICs 7 -lactam antibiotics 180
MGIT 7 compound class and mechanism, action
mobile shipping container 7 ascididemin 175
Mycobacterium tuberculosis 3 genome-wide measurement of transcript 173
nucleic acid amplification tests 8–9 linear-discriminant analysis 174
optimized see Optimized sputum smear MIC 175
microscopy mycolic acids 174
POC 9 respiratory poisoning. 174
‘programmatic endorsement’ 9 drug discovery pipeline 179, 180
serological tests 6–7 drug-specific mRNA signatures 179
‘technical endorsement’ 9 gene expression profiling, microarray 173
tuberculosis diagnostic toolbox 9 global mRNA profiling 173
SSCC assay see Sputum serial colony-counting INH exposure 175–176
(SSCC) assay microarray analyses 173
Strand displacement amplification (SDA) 17 mono oxygenase EthA 180
Streptomycin (SM) 147 mutagenic/nitrosative reaction 179–180
Streptomycin-isoniazid-rifampicin-pyrazinamide novel chemotherapeutic approaches 180
(SHRZ) 153 resistance and tolerance 177–178
Structure-activity relationships (SARs) RNA profiling 179
enzymatic assay vs. whole cell assay see target-based screening 180
Enzymatic assay vs. whole cell assay target identification 179
modification process 206–207 temporal and concentration-dependent
structural diversity 216 signatures 176–177
whole cell screening 178–179
Temporal and concentration-dependent Trial design, TB
signatures early stage/phase II clinical trial 200
drug action 176 MDR-TB 199–200
kas operon 176 MIRU 200
qRT-PCR and microarray methodologies 177 RFLP 200
subinhibitory concentrations 177 TST see Tuberculin skin test (TST)
transcriptional regulation 176 TTP see Time to positivity (TTP)
Thioridazine (TZ) Tuberculin skin test (TST)
antibiotic resistance, Mycobacterium tuberculosis ELISA 51, 53
233–234 ELISpot 49, 52, 54, 56
drug-resistant strains 233 HIV-positive and -negative subjects 52
effective anti-MDR/XDR-TB 234–235 IGRAs 48, 54, 58, 59
efflux pump inhibitors 238 Mycobacterium tuberculosis 50
intrinsic drug resistance 238 NAAT 75
isoniazid (INH) 232 positive community controls 72
MDR-TB 232 PPD antigen 74
phenothiazines and anti-tuberculosis activity vs. sensitivity 74
see Phenothiazines tuberculosis disease 57
rifampicin (RIF) 232 Tuberculosis (TB)
Time to positivity (TTP) Aphorisms 82
automated liquid culture systems 141 breath samples 89
changes, treatment 156 case detection rates 81
liquid culture system 155 chemotherapy
measurements 156 additive/substitutive 198–199
TNF- see Tumour necrosis factor-alpha (TNF-) antibiotics 197
Toll-like receptors (TLRs) drug development 198
activation 71–72 human population 200
non-TLR pattern recognition receptors 72 laboratory methodology 201–202
270 Index
Tuberculosis (TB) continued chemotherapy, patients 247

patient safety 201 clinical disease spectrum 246
Streptomyces griseus 197 clinical evaluation 248
trial design 199–200 end point definitions, efficacy trials 253–254
chromatography 87 EPI 243
clinical specimens 89 evaluation 247
diagnostician 82 genetic sequencing tools 243
human olfactory system 82 growth formulation 244
infectious diseases 81 immunological correlates, protection 253
investigative test 82 induction, immunopathology 247
less dramatic odours 82 lack, high-incidence field sites 253–254
mass spectrometry 87–88 NTM 244
multivariate statistics 89 nutritional factors 244
olfactory detection see Human olfactory protective immune response 245–246
detection tuberculosis vaccine development 253–254
pattern recognition 89 vaccination strategies 247
pheromones and semiochemicals 82 validated animal models 253
phthisis 82 WHO 245
vaccines see Vaccines, TB VersaTREK system 102
vapours 81 Volatile organic compounds (VOCs) analysis
VOC analysis see Volatile organic compounds anaesthetic 87
(VOCs) analysis bacterial metabolism 87
volatile biomarkers 88–89 endogenous 86–87
volatile compounds and gases 81–82 odour 86
Tumour necrosis factor-alpha (TNF-)
blocking agents 73 XDR see Extremely drug-resistant (XDR)
expression 73 XDR-TB see Extremely drug-resistant tuberculosis
synthesis 73 (XDR-TB)
therapeutic blockade 68
Ziehl–Neelsen (ZN)
Vaccines, TB light microscopy 83, 84
BCG 248–253 stained smears 4

(Advances in Molecular and Cellular Microbiology, 21) Timothy D McHugh-Tuberculosis - Laboratory Diagnosis and Treatment Strategies-CAB International (2013)

Uploaded by

Copyright:

Available Formats

(Advances in Molecular and Cellular Microbiology, 21) Timothy D McHugh-Tuberculosis - Laboratory Diagnosis and Treatment Strategies-CAB International (2013)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Advances in Molecular and Cellular Microbiology, 21) Timothy D McHugh-Tuberculosis - Laboratory Diagnosis and Treatment Strategies-CAB International (2013)

Uploaded by

Copyright:

Available Formats

What is the aim of this series?

What is the aim of this series?

What are some of the titles available in this series?

What are some of the titles available in this series?

Advances in Molecular and Cellular Microbiology 21

Through the application of molecular and cellular microbiology, we now recognize

Professor Michael Wilson

17. Helicobacter pylori in the 21st Century

18. Antimicrobial Peptides: Discovery, Design and Novel Therapeutic Strategies

19. Stress Response in Pathogenic Bacteria

20. Lyme Disease: an Evidence-based Approach

21. Tuberculosis: Laboratory Diagnosis and Treatment Strategies

22. Antimicrobial Drug Discovery: Emerging Strategies

24. Bacteriophages in Health and Disease

Titles Forthcoming from CABI

The Human Microbiota and Microbiome

Meningitis: Cellular and Molecular Basis

Tel: +44 (0)1491 832111 T: +1 800 552 3083 (toll free)

© CAB International 2013. All rights reserved. No part of this publication

Library of Congress Cataloging-in-Publication Data

Tuberculosis : laboratory diagnosis and treatment strategies / edited by

ISBN-13: 978 1 84593 807 9

Commissioning editor: Rachel Cutts

Typeset by Columns Design XML Ltd, Reading, UK.

1. Improving on Sputum Smear Microscopy for Diagnosis of Tuberculosis in

2. Molecular Diagnosis of Active Pulmonary Tuberculosis 13

3. Improving on the LJ slope – Automated Liquid Culture 34

4. Interferon-gamma Release Assays in the Diagnosis of Latent Tuberculosis

5. Measuring Tuberculosis Immune Responses in the Lung – The Correct Target? 67

6. Sniﬃng Out Tuberculosis 81

Part II – Measuring Resistance 93

7. Role of Phenotypic Methods for Drug Susceptibility Testing of M. tuberculosis

8. Genotypic Measures of Antibiotic Susceptibility 108

9. Molecular Tools for Fast Identification of Resistance and Characterization

Part III – Understanding Treatment 137

10. Monitoring Therapy by Bacterial Load 139

11. Modelling Responses to Tuberculosis Treatment 147

12. Measuring Gene Expression by Quantitative PCR (qPCR) 164

13. Transcriptomic Approaches to Mapping Responses to Drug Therapy for

14. Mycobacterial Lipid Bodies and Resuscitation-promoting Factor

Part IV – Treatment Strategies 195

15. Clinical Trials in Tuberculosis Chemotherapy: The Challenges 197

16. The Identification of 2-Aminothiazole-4-carboxylates (ATCs) as a

17. Rifamycins Revisited 219

19. Vaccines for Tuberculosis 243

Leonard Amaral, Unit of Mycobacteriology, Instituto de Higiene e Medicina Tropical,

Leonid Heifets, Mycobacteriology Reference Laboratory, National Jewish Health, Denver,

AFB Acid-fast bacilli

PBMCs Peripheral blood mononuclear cells

Andrew Ramsay,1 Karen R. Steingart2 and Madhukar Pai3

of Epidemiology, Biostatistics and Occupational Health, McGill University,

1.1 Introduction ground (WHO, 2011a). This chapter will

© CAB International 2013. Tuberculosis: Laboratory Diagnosis and Treatment Strategies

In the absence of simple, low-cost alternatives

1.3.1 Serial sputum smear examination 1.3.3 Specimen processing (Steingart,

The average increase in sensitivity/incre- The systematic review indicated that, in

appropriate to basic laboratories in resource- on how to produce good quality sputum

References Cuevas, L.E., Al-Sonboli, N., Lawson, L., Yassin,

Anna L. Bateson,1 Kate Reddington2 and Justin O’Grady3