Current perspectives

New genetic findings lead the way to a better understanding

of fundamental mechanisms of drug hypersensitivity
Munir Pirmohamed, PhD, FRCP,a David A. Ostrov, PhD,b and B. Kevin Park, PhDa Liverpool, United Kingdom, and
Gainesville, Fla

Drug hypersensitivity reactions are an important clinical

problem for both health care and industry. Recent advances in Abbreviations used
genetics have identified a number of HLA alleles associated with ADR: Adverse drug reaction
a range of these adverse reactions predominantly affecting the p-i: Pharmacologic interaction
skin but also other organs, such as the liver. The associations SJS: Stevens-Johnson syndrome
TEN: Toxic epidermal necrolysis
between abacavir hypersensitivity and HLA-B*57:01 and
carbamazepine-induced Stevens-Johnson syndrome and HLA-
B*15:02 have been implemented in clinical practice. There are
many different mechanisms proposed in the pathogenesis of
drug hypersensitivity reactions, including the hapten
hypothesis, direct binding to T-cell receptors (the Adverse drug reactions (ADRs) remain a major problem in the
pharmacologic interaction hypothesis), and peptide-binding clinic and in drug development.1 Of particular concern are unpre-
displacement. A problem with all the hypotheses is that they are dictable ADRs, which are often referred to as type B or idiosyn-
largely based on in vitro findings, with little direct in vivo cratic reactions.2 They can affect any organ system, most
evidence. Although most studies have focused on individual commonly the skin but also the liver, lungs, bone marrow, and
mechanisms, it is perhaps more important to consider them all kidneys. Such reactions are still poorly understood, but many of
as being complementary, potentially occurring at the same time them are often assumed to be immunologic in nature,3 so-called
with the same drug in the same patient. This might at least drug hypersensitivity reactions.
partly account for the heterogeneity of the immune response Recent advances in the pharmacogenetics of drug hypersen-
seen in different patients. There is a need to develop novel sitivity reactions, most notably HLA restriction, have provided
methodologies to evaluate how the in vitro mechanisms relate to persuasive evidence that some of these drug reactions have an
the in vivo situation and how the highly consistent genetic immunologic component. Furthermore, such advances are
findings with different HLA alleles can be more consistently beginning to provide the experimental tools to investigate the
used for both prediction and prevention of these serious adverse interplay between the genetic, cellular, and chemical mecha-
reactions. (J Allergy Clin Immunol 2015;136:236-44.) nisms postulated to be involved in drug hypersensitivity. We now
need to build on such advances in a critical and constructive
Key words: Drug hypersensitivity, HLA, genetic polymorphisms, fashion to:
mechanisms, crystallography
d develop diagnostic tests to define whether an ADR to a new
drug is truly immunologic;
d develop pharmacogenetics tests (which might be point of
a
From the Department of Molecular and Clinical Pharmacology, MRC Centre for Drug care) for patients to prevent these reactions, which would
Safety Science, University of Liverpool, and bthe Department of Pathology, Immu- fit in with the evolving paradigm of personalized or preci-
nology and Laboratory Medicine, University of Florida College of Medicine, Univer- sion medicine; and
sity of Florida, Gainesville.
Both M.P. and B.K.P. wish to thank the MRC for their support through the MRC Centre
d provide the science to develop novel in vitro model systems
for Drug Safety Science. M.P. is a NIHR Senior Investigator. D.A.O. is supported with which to (1) eliminate immunologic liabilities from
through National Institutes of Health/National Institute of Allergy and Infectious Dis- drugs known to be associated with hypersensitivity and
eases grant 5R01AI103348-02. (2) develop preclinical test systems to minimize the immu-
Disclosure of potential conflict of interest: M. Pirmohamed has received research support
nologic liabilities of new drugs in development.
from the Medical Research Council and the National Institute for Health Research.
D. A. Ostrov has received research support from the National Institutes of Health/ To achieve these goals, we need to investigate further the best
National Institute of Allergy and Infectious Diseases (R01 AI103348). B. K. Park
declares that he has no relevant conflicts of interest.
understood case histories ‘‘from man to molecule and back
Received for publication May 18, 2015; revised June 22, 2015; accepted for publication again.’’
June 23, 2015.
Corresponding author: Munir Pirmohamed, PhD, FRCP, MRC Centre for Drug Safety
Science, Department of Molecular and Clinical Pharmacology, University of Liver- GENETIC BASIS OF DRUG HYPERSENSITIVITY
pool, Liverpool, L69 3GL, United Kingdom. E-mail: [email protected]. Association with HLA alleles
0091-6749 Since the completion of the human genome project, there have
Ó 2015 The Authors. Published by Elsevier Inc. on behalf of the American Academy
of Allergy, Asthma & Immunology. This is an open access article under the CC BY
been some remarkable advances in elucidating the genetic basis
license (http://creativecommons.org/licenses/by/4.0/). of drug hypersensitivity reactions.4,5 This has largely focused on
http://dx.doi.org/10.1016/j.jaci.2015.06.022 the HLA alleles in the MHC on the short arm of chromosome 6.

236
J ALLERGY CLIN IMMUNOL PIRMOHAMED, OSTROV, AND PARK 237
VOLUME 136, NUMBER 2

FIG 1. HLA associations reported with serious ADRs caused by many different drugs. This is not an
exhaustive list of reported drug-HLA associations but illustrates that the pattern of drug–HLA–tissue injury
interactions varies and that there is no general rule for predicting susceptibility.

This is the region of the genome that is highly polymorphic and HLA alleles can be more highly expressed in particular organs.
has been linked to both autoimmune diseases and infectious dis- Consistent with this, a recent genome-wide association study
eases. Since 2001, at least 24 ADRs have been linked to different showed that HLA-DRB1*15:01 is also associated with alcohol-
HLA alleles (Fig 1). These associations afford a fascinating induced liver cirrhosis.14
insight into the complexity of the immune response, highlighting Fifth, although most of the associations illustrated in Fig 1 are
our incomplete understanding of the pathogenesis of these reac- consistent with the time lag needed to induce an immune response
tions. However, they also provide several important insights (and are consistent with clinical and histologic evidence), for some
that need to be addressed in future research. forms of organ injury, the HLA association was surprising and
First, different organ systems are affected by these ADRs, most highlighted that the general rules of immune-mediated injury
commonly the skin and liver, but also muscles (statin-induced occurring soon after the start of the drug or being worse on rechal-
autoimmune myopathy and HLA-DRB1*11:01)6 and neutrophils lenge are not necessarily correct in all cases. For example,
(clozapine-induced agranulocytosis and HLA-DQB1).7 Whether lumiracoxib-induced liver injury typically occurs after more than
the HLA allele determines which organ system is affected is un- 100 days on the drug12 which is not the usual timeframe associated
clear, but it is likely to play a role in association with other factors, with liver injury caused by other drugs, such as co-amoxiclav. For
such as other genetic variants, expression of organ-homing recep- clozapine agranulocytosis, rechallenge does not necessarily result
tors, and T-cell clonotypes. in recurrence of the reaction much sooner,15 which is inconsistent
Second, the associations sometimes show marked ethnic with evidence from other forms of immune-mediated drug injury,
variation reflecting the background prevalence of the implicated such as abacavir hypersensitivity.16
HLA allele. The most impressive example of this is the associ- Finally, it is interesting to note that the drugs most commonly
ation of HLA-B*15:02 with carbamazepine-induced Stevens- associated with cutaneous ADRs, such as the penicillins and
Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) sulfonamides, have not convincingly been shown to have
in Han Chinese, Thai, and Malay patients8 but not in Northern Eu- associations with HLA alleles. This might be because they form
ropeans.9 The background prevalence of HLA-B*15:02 varies multiple epitopes and therefore interact with multiple HLA
from 4% to 15% in the affected populations but is less than 1% alleles, but this cannot be the full explanation given the very
in Japanese and Korean subjects and extremely rare in Northern strong association between flucloxacillin hepatotoxicity and
Europeans (<0.01%). HLA-B*57:01. A possible explanation is that studies thus far
Third, the same HLA allele can be associated with adverse have not phenotyped the patients precisely with mixing of pheno-
reactions to therapeutically and structurally unrelated compounds types, poor assessment of causality, and inadequate sample sizes.
but with effects in different organs. The best example of this is the This is perhaps illustrated by a recent genome-wide study in
association of HLA-B*57:01 with both abacavir hypersensitiv- which careful phenotyping of an adequate number of patients
ity10 and flucloxacillin-induced hepatotoxicity.11 This is consid- with penicillin-induced type 1 IgE-mediated reactions showed
ered in more detail below. an association within the HLA-DRA region.17
Fourth, the same type of organ injury can occur with the same
HLA allele, even with therapeutically and structurally unrelated
compounds. For example, HLA-DRB1*15:01 is associated with Association with non-HLA alleles
liver injury with both lumiracoxib12 and co-amoxiclav.13 Whether Observational studies have shown with some drugs, such as
there are some HLA alleles that predispose to certain forms of or- phenytoin, that higher doses, particularly at commencement, can
gan injury will become clear as more research is undertaken with increase the risk of cutaneous eruptions.18 Similarly, with lamotri-
other drugs but seems likely based on the possibility that certain gine, higher doses, particularly when the patient is cotreated with
238 PIRMOHAMED, OSTROV, AND PARK J ALLERGY CLIN IMMUNOL
AUGUST 2015

sodium valproate, which inhibits the glucuronidation of lamotri- with serious ADRs. This is covered in more detail in the article
gine,19 lead to a higher frequency of serious cutaneous ADRs, by White et al34 in this issue of the Journal.
including SJS.20 This led to revised prescribing information to start
patients on lower doses of lamotrigine and escalate the dose slowly.
These clinical observations would suggest that genetic poly- Carbamazepine-induced hypersensitivity reactions
morphisms in drug metabolic pathways can increase the risk of Carbamazepine is widely used in the treatment of epilepsy,
serious idiosyncratic reactions. However, candidate gene studies trigeminal neuralgia, and bipolar disorder. The clinical manifesta-
and, more recently, genome-wide association studies have not tions of carbamazepine hypersensitivity are more complex and
shown any convincing associations with polymorphisms in genes diverse than those observed with abacavir and include simple
coding for drug-metabolizing enzymes or transporters.21 The maculopapular exanthema, hypersensitivity syndrome, or drug
exception to this seems to be phenytoin; a recent genome-wide as- reaction with eosinophilia and systemic symptoms syndrome and
sociation study22 in Taiwanese, Japanese, and Malaysian patients SJS/TEN. Carbamazepine-induced SJS/TEN has shown a strong
showed a strong association with CYP2C9*3, an allelic variant (odds ratio >1000) association with HLA-B*15:02 in the Han Chi-
associated with reduction in CYP2C9 activity by about 90%, nese population.35-37 This association has also been replicated in
which reached genome-wide significance (P 5 10217). There several other Asian populations, including Thai,38,39 Malay,40
are perhaps several lessons to be learned from this. and Indian subjects41 but not in white9,42,43 and Japanese sub-
First, phenytoin has a narrow therapeutic index because its jects.44,45 Apart from being ethnicity specific, the association
fractional clearance is largely dependent on CYP2C9. Thus the with HLA-B*15:02 is phenotype specific in that it is only valid
lower activity of the CYP2C9*3 allelic variant leads to reduced for SJS/TEN but has not been shown to be important for maculo-
metabolic clearance22 and higher plasma concentrations, which papular exanthema and hypersensitivity syndrome.36 The utility
are consistent with clinical observations of a dose-dependent rela- of HLA-B*15:02 testing for preventing carbamazepine-induced
tionship in the risk of cutaneous eruptions.18 It also suggests that a SJS/TEN was also shown in a prospective study from Taiwan.46
relative deficiency in one pathway might not necessarily predis- The drug label for carbamazepine has been changed by many
pose to a serious ADR if multiple metabolic pathways are drug regulatory agencies worldwide, including the European Med-
involved in the clearance of that drug. Alternatively, the effect icines Agency and US Food and Drug Administration. Even
size may be so low that very large numbers of patients would though genetic testing for HLA-B*15:02 is reimbursed in Hong
be required to show genome-wide significance. Kong, advice that patients should be tested for this allele before
Second, type B or idiosyncratic reactions have typically been the use of carbamazepine led physicians to use alternative anticon-
classified as being dose independent.2 However, these more vulsants, such as phenytoin and lamotrigine, which did not reduce
recent findings suggest that even these reactions have some the overall incidence of SJS-TEN in Hong Kong.47 This provides a
form of dose-response relationship, but this might not be depen- salutary lesson as to the difficulties that can be encountered in im-
dent on the external dose but more on the systemic exposure to plementing genetic tests into clinical practice.
the parent drug or 1 or more of its metabolites. For other serious cutaneous ADRs caused by carbamazepine in
Third, this association with reduced metabolic clearance of Chinese patients48 and for maculopapular eruptions, drug reac-
phenytoin does not necessarily mean that the parent drug is tion with eosinophilia and systemic symptoms syndrome, and
responsible for the reaction because inability to metabolize SJS/TEN in white49 and Japanese50 subjects, HLA-A*31:01 has
phenytoin through its main route can lead to diversion through been shown to be a predisposing factor. A recent study has
more toxic pathways. demonstrated that preprescription genotyping for HLA-A*31:01
in the context of the United Kingdom National Health Service
would be cost effective.51 HLA-A*31:01 is mentioned in drug la-
Clinical translation bels in many countries, but testing is not mandatory, and thus it is
The effect size and therefore the strength of association with not routinely used in clinical practice. A recent functional study in
HLA alleles varies with different drugs,4,5 which has influenced a carbamazepine-hypersensitive patient showed the activation of
translation into clinical practice. To date, for abacavir hypersen- both HLA-A*31:01–restricted, carbamazepine-specific CD81 T
sitivity and carbamazepine-induced SJS in Southeast Asians, cells and HLA-DRB1*04:04–restricted, carbamazepine-specific
the drug label has been changed, stating that HLA testing needs CD41 T cells, indicating that cooperation between different T-
to be done before commencing the drug.21 cell subsets within the extended genetic haplotype plays a role
in the clinical manifestations seen in patients.52

Abacavir hypersensitivity
Abacavir hypersensitivity occurs in approximately 5% to 7% of PATHOGENESIS OF DRUG HYPERSENSITIVITY
patients.23 The reaction is CD81 T cell mediated.24 The associa- REACTIONS
tion between the HLA class I allele HLA-B*57:01 and abacavir Our current understanding of drug hypersensitivity is outlined
hypersensitivity was first reported from Australia25 and North in Fig 2. From a chemical perspective, the drug can act as:
America26 and subsequently by a study from the United d a hapten—a chemical that forms covalent attachment to a
Kingdom.27 Genetic testing has been demonstrated to be cost- protein;
effective, first in the United Kingdom27 and subsequently in other d a prohapten—a chemical that can be converted to a hapten;
countries.28-30 Furthermore, preprescription testing for HLA- d an antigen—a chemical, normally a peptide, that acts as a
B*57:01 reduced the frequency of hypersensitivity in Australia,31 ligand between MHC molecules and cognate immunologic
the United Kingdom,32 and France,33 showing that genetic testing receptors in both the afferent and efferent limbs of the
can have a powerful influence in reducing the burden associated immune response;
J ALLERGY CLIN IMMUNOL PIRMOHAMED, OSTROV, AND PARK 239
VOLUME 136, NUMBER 2

FIG 2. A scheme to illustrate current hypotheses concerning how a drug might interject in natural immunity
to cause a hypersensitivity ADR. It is likely that multiple molecular initiating events, including reactive
metabolite formation, conjugation of drug to protein, and direct interaction of the drug and MHC molecules,
can provide signal 1. For a response to proceed, there will be regulation by costimulatory (signal 2) and
coinhibitory molecules, which might ultimately activate various adverse pathways, resulting in the highly
variable clinical manifestations of drug hypersensitivity. It is important to note that no mechanism has been
proven unequivocally in human subjects to date. Furthermore, little is known about the signal in the target
cell or tissue that provokes effector T cells to cause tissue damage.

d a costimulatory agent—a chemical or patient factor that in- reactivity53: trinitrophenyl-modified MHC peptides can generate
teracts with dendritic cells, polarizing and maturing an im- a T-cell response, and trinitrophenyl-specific T cells can recog-
mune response; nize both trinitrophenyl-modified and unmodified peptides. This
d an immunogen—a chemical that can initiate an immune shows how chemical modification can break tolerance. Further-
response (in human subjects); and more, such skin sensitizers indirectly activate innate inflamma-
d a sensitogen—a chemical that can elicit hypersensitivity (in tory signaling through production of reactive oxygen species
human subjects). and hyaluronic acid degradation.54 These findings in experi-
mental animals have been translated into human experimental
From an immunologic perspective, it is important to note that
models through topical application of dinitrochlorobenzene to
the whole process is likely to be subject to immune modulation by
volunteers,55-57 in whom both chemical-specific CD41 and
costimulatory signals, regulatory T cells, and the nature of the
CD81 T cells have been detected, which is consistent with the
antigen in the target tissue, which pulls the destructive immune
ability of 2,4-dinitrochlorobenzene to form haptens with both
response into the target tissue.
intracellular and extracellular proteins.
Penicillins are b-lactam antibiotics that react spontaneously with
The chemical basis of drug hypersensitivity nucleophilic lysine residues in proteins in a dose- and time-
Until recently, the hapten hypothesis was the most widely dependent fashion.58,59 Antibodies have been detected against
assumed mechanism for drug hypersensitivity. The ability of penicillin-protein conjugates and the benzylpenicilloyl group but
low-molecular-weight nitrophenyl derivatives to cause skin not the parent drug.60-63 Use of specific mass spectrometric
sensitization in guinea pigs is a function of their chemical methods has shown that penicillin-protein conjugates can be
240 PIRMOHAMED, OSTROV, AND PARK J ALLERGY CLIN IMMUNOL
AUGUST 2015

detected in all patients exposed to the drug, and T cells that recog- cell assay that simultaneously discriminates between hypersen-
nize protein conjugates64,65 have been demonstrated in hypersensi- sitive and nonhypersensitive patients. The first step toward such
tive patients. Interestingly, specific lysine residues are targeted in an assay system has been the development of novel sensitive mass
the albumin molecule at low concentrations, with the number of spectrometric methods that enable the full characterization and
epitopes increasing as the drug concentration increases.66 In pa- quantification of drug-modified proteins in patients. Such
tients with cystic fibrosis, which is characterized by a high inci- methods can be used not only to detect which proteins have
dence of piperacillin hypersensitivity, mass spectrometric been modified but also the precise chemistry of the hapten and the
techniques have revealed the presence of multiple haptens on albu- specific amino acid modified within the target protein. Such
min in patients with the presence of drug-specific T-lymphocyte re- platforms provide a direct and unequivocal definition of the
sponses.59 Given that antibiotic courses are often given over 14 relationship between drug metabolism and hapten formation
days in these patients and that the half-life of albumin is 19 days, in vivo and provide sequence information for the synthesis and
modified protein is likely to accumulate during the course and immunologic evaluation of potential antigenic drug-peptide
will persist after the course. How this relates to the likelihood of hy- conjugates.59
persensitivity occurring during the course of the treatment or some- In summary, the above examples provide evidence that
times after the course is completed requires further investigation. bioactivation (either spontaneous or metabolism dependent) to
A key unanswered question for penicillins and the hapten reactive metabolites has a role to play in the pathogenesis of
hypothesis is why one penicillin, such as piperacillin, targets serious ADRs, but whether it is the predominant mechanism or
the skin primarily, whereas another, such as flucloxacillin or one that plays an accessory role will depend on the drug, host, and
co-amoxiclav, can also target the liver. Organ-selective dispo- environmental factors.
sition of the drug antigen could be a factor (ie, albumin is
synthesized in the liver),67 but it is also possible that
organ-selective costimulatory signals could play a role given The pharmacologic interaction hypothesis of drug
the ubiquitous nature of penicillin hapten formation. In relation hypersensitivity
to flucloxacillin, it is metabolized by CYP3A4 to 59-hydroxy- Pichler74 has proposed an alternative hypothesis based on
methylflucloxacillin,68 but the overall ratio of parent drug to direct binding of the parent drug to the T-cell receptor or the
metabolite is 25:1, and therefore whether this is important in HLA molecule. This is termed the pharmacologic interaction
hepatic toxicity is unclear. (p-i) hypothesis and requires the drug to bind to these receptors
The human safety risk posed by the generation of reactive with high affinity but in the absence of innate immune activation,
metabolites with the potential to form haptens in vivo remains a metabolism, or antigen processing.75 Indeed, it has been proposed
major concern for drug development.69 A large number of drugs that HLA restriction, as has been demonstrated with many drugs
associated with hypersensitivity have been shown to form reactive (outlined above), is likely to occur only when a drug is able to bind
metabolites that can modify proteins both in vitro and in vivo.3 to one HLA allele but not when the drug is converted to a reactive
However, the ‘‘reactive metabolite theory of drug hypersensitiv- metabolite because this will generate a large number of antigenic
ity’’ is still largely based primarily on a ‘‘global association’’ be- determinants.75 However, this ignores the increasing number of
tween in vitro bioactivation and clinical outcome. General HLA alleles that have been associated with drugs causing liver
anesthetics were one of the first paradigms used to investigate injury in which metabolism might play an important role (Fig 1).
the role of metabolism in drug-induced immune hepatitis. Halo- Interaction with T-cell receptors, which results in T-cell
thane is metabolized in the liver by CYP2E1 to yield highly reac- activation (but not antibody formation), has been shown with a
tive trifluoroacetyl chloride, which binds spontaneously to number of drugs, including sulfamethoxazole, abacavir, carba-
proteins. Circulating IgG antibodies that recognize modified liver mazepine, lamotrigine, lidocaine, and allopurinol.76 However, the
proteins and autologous proteins were implicated in the mecha- evidence is largely based on in vitro experiments, and although
nism of halothane hepatitis because they were detected in all pa- direct T-cell interaction is likely to occur in vivo as well, it does
tients with severe reactions.70,71 Interestingly, the overall extent not fully simulate the complex situation seen in vivo in which
of metabolism (bioactivation) of enflurane and isoflurane, which drug concentrations will vary in different tissues and there will
have replaced halothane in clinical practice, is around 10-fold be differing amounts of metabolites present in different subjects,
lower, and these are associated with a much lower risk of which will include variability in the formation of drug-protein
hepatitis.2 Similarly, IgG antibodies have been detected that conjugates. Thus there is a need to rationalize how the 2 potential
recognize drug-modified proteins in patients who had immune- mechanisms might operate together under different circum-
mediated hepatitis on isoniazid.72 More recently, a number of stances. This can be done by considering the drug sulfamethoxa-
drugs known to cause liver injury have been shown to have strong zole, an antimicrobial used for the treatment of urinary tract
genetic associations with specific HLA alleles (Fig 1). All these infections and Pneumocystis jirovecii pneumonia.
drugs are known to undergo P450-mediated metabolism, but Sulfamethoxazole has been used to define the chemical basis of
whether the parent drug, stable metabolite, or chemically reactive drug (metabolite) hypersensitivity reactions because its reactive
metabolites are involved is undetermined at present. For example, metabolite, nitroso sulfamethoxazole, can be synthesized77 and
with lapatinib, bioactivation to a quinoneimine intermediate has integrated into functional in vitro T-cell assays at nontoxic con-
been shown,73 but again, the final proof of its role in hepatic injury centrations.78 In vivo cytochrome P450–mediated metabolism
is not available. The main issue is that at present, the technology of sulfamethoxazole generates an unstable but not protein-
to investigate T-cell responses to short-lived chemically reactive reactive hydroxylamine metabolite that can be detected in
intermediates does not exist. plasma.79 The hydroxylamine undergoes rapid spontaneous
The ultimate proof of the hapten hypothesis would be detection oxidation to nitroso sulfamethoxazole, an electrophile that can
and characterization of effector drug–peptide conjugates in a T- bind covalently to cysteine residues in both extracellular and
J ALLERGY CLIN IMMUNOL PIRMOHAMED, OSTROV, AND PARK 241
VOLUME 136, NUMBER 2

cellular proteins.80-82 Formation of adducts can induce apoptotic and randomly at all other positions. For 9-mer peptides (20 amino
and necrotic cell death and the release of mediators that promote acids 3 9 positions), 180 peptide mixtures can be used in HLA-
dendritic cell maturation.83-85 The binding of nitroso sulfameth- binding assays in the presence and absence of drug. For example,
oxazole to protein produces multiple adduct structures and there- peptide libraries in which position 9 (P9) was fixed to be V or I
fore a multitude of potential antigens. Whether this is a potential (and random at all other positions) were found to bind HLA-
explanation for the lack of an HLA association with sulfamethox- B*57:01 with higher affinities in the presence of abacavir
azole hypersensitivity is unclear.86 compared with the absence of the drug.95
Nitroso sulfamethoxazole is chemically unstable in cell culture Elution of HLA-bound peptides from drug-treated cells is an
medium; its half-life is less than 30 minutes,77,78 and therefore informative method to characterize drug/HLA interactions. In this
cellular exposure to nitroso sulfamethoxazole in T-cell assays is method cells expressing the HLA molecule of interest are grown to
likely to be 10- to 50,000-fold lower than the parent compound. large numbers (>109) when treated with either the drug or vehicle
Despite this, nitroso sulfamethoxazole–responsive T cells have control. The HLA molecules are immunopurified, and bound pep-
been detected in the periphery and skin of 100% of patients tides are subsequently eluted and then characterized by means of
with sulfamethoxazole hypersensitivity.87-90 In patients with mass spectrometry and sequencing. Comparison of peptides eluted
milder skin reactions (eg, maculopapular exanthema), CD41 T from HLA in the presence and absence of drug can be used to mea-
cells are preferentially activated, whereas CD81 T cells are sure the effects of the drug on the HLA-bound peptide repertoire.
more readily detected in patients with SJS/TEN. T cells from hy- For example, HLA-B*57:01 typically exhibits a preference for 9-
persensitive patients are also activated with the parent drug sulfa- mer peptides with bulky aromatic side chains (such as W or F) at
methoxazole through direct interaction with MHC molecules, P9. However, peptides eluted from HLA-B*57:01 after isolation
specific T-cell receptors, or both.91 Some of these T cells were from abacavir-treated cells exhibit an altered binding preference:
found to be highly structurally specific in that different sulfon- nonamer peptides with short aliphatic side chains at P9 (I or V).
amide derivatives can bind to the T-cell receptor, but only sulfa- These data show that abacavir treatment of cells results in alter-
methoxazole results in activation. This has been attributed to its ation of the repertoires of HLA-B*57:01–bound peptides.95,96
orientation being similar to that of the hapten formed with nitroso Alteration of the repertoire of HLA-bound peptides might be
sulfamethoxazole.92 due to direct (eg, by drug/HLA binding) or indirect (eg, by altering
After in vitro priming of naive T cells with nitroso sulfameth- protein expression, metabolism, or both) effects. The peptide
oxazole, it is possible to prime naive CD41 T cells from 100% elution method can be used to measure the effects of a drug on
of healthy human donors in 7 days if (1) the drug-derived anti- the HLA-bound peptide repertoire, regardless of whether the effect
gen was presented by dendritic cells, (2) regulatory T cells were is direct or indirect. These aspects in relation to abacavir, together
removed from the assay, and (3) costimulatory signals were pro- with the concepts of heterologous immunity, alloreactivity, and
vided through the use of LPS/TNF-a.93 Furthermore, the polyspecificity, are covered by White et al34 in this issue of the
magnitude of the T-cell response could be enhanced dramati- Journal and will not be repeated here.
cally through blockade of programmed cell death 1/pro-
grammed cell death ligand 1 signlling.94 Feeding regulatory T
cells back into the system effectively blocked nitroso sulfameth- CONCLUSIONS
oxazole–specific T-cell priming. Hence levels of costimulatory/ The discovery of MHC restriction in drug hypersensitivity has
coinhibitory signaling at the time of drug exposure might be provided an opportunity to delve further into the fundamental
another important determinant of whether a pathogenic T-cell chemical basis of how a drug might initiate such events in a
response develops. patient and raise broader pharmacologic questions with regard to
Taken together, these observations show that T cells circulating the disposition and effector functions of drug-related antigens.
in the same hypersensitive patients are activated by different Proteins are processed by heterogeneous intracellular pathways,
chemical entities through different pathways (both stimulatory leading to the generation of peptide–MHC class I complexes and
and inhibitory), which is consistent with drug- and metabolite- peptide–MHC class II complexes. For both pathways, the
specific T cells being part of the spectrum of a polyclonal T-cell peptides are mostly buried in the peptide-binding groove, and
response. Therefore it is impossible to identify the drug-derived the peptide is an integral part of the folded protein complex, part
antigen that actually initiated the initial T-cell response in a of which is recognized by T-cell receptors. Endogenous peptides
patient from retrospective analysis of the frequency of memory T can be altered posttranslationally through a variety of chemical
cells (recall). Furthermore, at present, we cannot be certain which processes, some of which might be disease induced,97,98 and thus
particular T cells are the actual effector cells during a hypersen- contribute to allergy or autoimmune disease. The hapten hypoth-
sitivity reaction in vivo when the target organ is usually the skin esis would suggest that a drug (metabolite) covalently modifies a
or, less commonly, the liver. protein before processing with presentation of drug peptide.
Although drug-hapten protein complexes have now been detected
in vivo for a number of drugs, characterization of drug peptides in
Methods to characterize drug/HLA interactions the context of MHC binding and T-cell activation has not been
Drugs that show HLA associations can be tested for HLA defined. The p-i hypothesis provides an alternative mechanism
binding by using in vitro assays. A cell-free system with purified that has been demonstrated to operate in vitro, but there is no
HLA molecules can be used to measure drug effects on peptide- direct experimental evidence that it operates in vivo. More
binding affinity by using a positional scanning peptide library recently, studies with abacavir have shown that it might occupy
approach.95 In this approach binding assays are established in a space below peptides expressed by particular HLA alleles,
which HLA molecules are incubated with drug and peptide mix- thus leading to altered peptide presentation and autoimmunity.
tures in which a peptide position is fixed to a single amino acid The difference between the altered peptide/p-i hypotheses and
242 PIRMOHAMED, OSTROV, AND PARK J ALLERGY CLIN IMMUNOL
AUGUST 2015

2. Park BK, Pirmohamed M, Kitteringham NR. Role of drug disposition in drug hy-
persensitivity: a chemical, molecular, and clinical perspective. Chem Res Toxicol
1998;11:969-88.
3. Uetrecht J, Naisbitt DJ. Idiosyncratic adverse drug reactions: current concepts.
Pharmacol Rev 2013;65:779-808.
4. Pavlos R, Mallal S, Ostrov D, Buus S, Metushi I, Peters B, et al. T cell-mediated
hypersensitivity reactions to drugs. Annu Rev Med 2015;66:439-54.
5. Pirmohamed M. Personalized pharmacogenomics: predicting efficacy and
adverse drug reactions. Annu Rev Genomics Hum Genet 2014;15:349-70.
6. Mammen AL, Gaudet D, Brisson D, Christopher-Stine L, Lloyd TE, Leffell MS,
et al. Increased frequency of DRB1*11:01 in anti-hydroxymethylglutaryl-
coenzyme A reductase-associated autoimmune myopathy. Arthritis Care Res
(Hoboken) 2012;64:1233-7.
7. Goldstein JI, Jarskog LF, Hilliard C, Alfirevic A, Duncan L, Fourches D, et al.
Clozapine-induced agranulocytosis is associated with rare HLA-DQB1 and
HLA-B alleles. Nat Commun 2014;5:4757.
8. Yip VL, Marson AG, Jorgensen AL, Pirmohamed M, Alfirevic A. HLA genotype
and carbamazepine-induced cutaneous adverse drug reactions: a systematic re-
view. Clin Pharmacol Ther 2012;92:757-65.
9. Lonjou C, Thomas L, Borot N, Ledger N, de Toma C, LeLouet H, et al. A marker
for Stevens-Johnson syndrome.: ethnicity matters. Pharmacogenomics J 2006;
FIG 3. A simple scheme to illustrate how a ternary complex between the
6:265-8.
drug, MHC molecule, and peptide can be assembled to stimulate a T-cell
10. Mallal S, Phillips E, Carosi G, Molina JM, Workman C, Tomazic J, et al. HLA-
response. The interaction between the peptide and drug can be either
B*5701 screening for hypersensitivity to abacavir. N Engl J Med 2008;358:
covalent (D-) or noncovalent (D.). It is possible that in the in vivo situation
568-79.
both models could be occurring simultaneously, but we do not have the
11. Daly AK, Donaldson PT, Bhatnagar P, Shen Y, Pe’er I, Floratos A, et al. HLA-
precise chemical tools to determine this.
B*5701 genotype is a major determinant of drug-induced liver injury due to flu-
cloxacillin. Nat Genet 2009;41:816-9.
12. Singer JB, Lewitzky S, Leroy E, Yang F, Zhao X, Klickstein L, et al. A genome-
wide study identifies HLA alleles associated with lumiracoxib-related liver injury.
the hapten hypothesis is that the former do not depend on covalent Nat Genet 2010;42:711-4.
interaction in vitro. 13. Lucena MI, Molokhia M, Shen Y, Urban TJ, Aithal GP, Andrade RJ, et al. Sus-
Although it is important to identify the mechanisms of drug ceptibility to amoxicillin-clavulanate-induced liver injury is influenced by multi-
hypersensitivity, perhaps too much emphasis is being placed on ple HLA class I and II alleles. Gastroenterology 2011;141:338-47.
14. Hebbring SJ, Schrodi SJ, Ye Z, Zhou Z, Page D, Brilliant MH. A PheWAS
the differences between the various hypotheses rather than approach in studying HLA-DRB1*1501. Genes Immun 2013;14:187-91.
focusing on the fact that the mechanisms are complementary 15. Pirmohamed M, Park K. Mechanism of clozapine-induced agranulocytosis: cur-
and might at least partly account for the heterogeneity of the rent status of research and implications for drug development. CNS Drugs
immune response seen in different patients. For example, even 1997;7:139-58.
16. Clay PG. The abacavir hypersensitivity reaction: a review. Clin Ther 2002;24:
with abacavir, although there is evidence of peptide-binding
1502-14.
displacement95 and alloimmunity,99 abacavir protein haptens 17. Gueant JL, Romano A, Cornejo-Garcia JA, Oussalah A, Chery C, Blanca-Lopez
have also been detected in vivo100; the role of these different pro- N, et al. HLA-DRA variants predict penicillin allergy in genome-wide fine-
cesses in vivo in patients with abacavir hypersensitivity is thus un- mapping genotyping. J Allergy Clin Immunol 2015;135:253-9.
clear, and it is of course possible that all 3 are important. Thus 18. Wilson JT, Hojer B, Tomson G, Rane A, Sjoqvist F. High incidence of a
concentration-dependent skin reaction in children treated with phenytoin. BMJ
what is being suggested (Fig 3) is that an assembly of the MHC 1978;1:1583-6.
molecule, peptide, and drug leads to a complex maintained by a 19. Yuen AWC, Land G, Weatherley BC, Peck AW. Sodium valproate acutely inhibits
variety of chemical bonds (which could include covalent, coordi- lamotrigine metabolism. Br J Clin Pharmacol 1992;33:511-3.
nate, hydrogen, and hydrophobic bonds). The thermodynamic 20. Wong IC, Mawer GE, Sander JW. Factors influencing the incidence of
lamotrigine-related skin rash. Ann Pharmacother 1999;33:1037-42.
stability of the complex will be determined by the totality of the
21. Pirmohamed M. Genetics and the potential for predictive tests in adverse drug re-
binding energies and not the type of bond per se. Furthermore, actions. Chem Immunol Allergy 2012;97:18-31.
it seems highly likely that the epitopes recognized by various T 22. Chung WH, Chang WC, Lee YS, Wu YY, Yang CH, Ho HC, et al. Genetic var-
cells will have variable composition of substructures in the iants associated with phenytoin-related severe cutaneous adverse reactions.
drug, peptide, and MHC molecule, some of which might not JAMA 2014;312:525-34.
23. Easterbrook PJ, Waters A, Murad S, Ives N, Taylor C, King D, et al. Epidemio-
contain the drug at all. logical risk factors for hypersensitivity reactions to abacavir. HIV Med 2003;4:
Finally, we know little or nothing about the nature of the MHC- 321-4.
drug-peptide complex that elicits target organ toxicity in the 24. Phillips EJ, Sullivan JR, Knowles SR, Shear NH. Utility of patch testing in pa-
afferent arm of the immune response in vivo. We need to know tients with hypersensitivity syndromes associated with abacavir. AIDS 2002;16:
2223-5.
more about the relationship between total drug disposition and
25. Mallal S, Nolan D, Witt C, Masel G, Martin AM, Moore C, et al. Association be-
the constitution of drug-peptide-MHC complexes in both the tween presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity
afferent and efferent arms of the immune response to define thresh- to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet 2002;359:727-32.
olds (drug, metabolite, and drug-protein conjugate) in early drug 26. Hetherington S, Hughes AR, Mosteller M, Shortino D, Baker KL, Spreen W, et al.
discovery that can be reasonably expected not to pose a risk in Genetic variations in HLA-B region and hypersensitivity reactions to abacavir.
Lancet 2002;359:1121-2.
the clinic. 27. Hughes DA, Vilar FJ, Ward CC, Alfirevic A, Park BK, Pirmohamed M. Cost-
effectiveness analysis of HLA B*5701 genotyping in preventing abacavir hyper-
REFERENCES sensitivity. Pharmacogenetics 2004;14:335-42.
1. Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ, et al. 28. Kauf TL, Farkouh RA, Earnshaw SR, Watson ME, Maroudas P, Chambers MG.
Adverse drug reactions as cause of admission to hospital: prospective analysis Economic efficiency of genetic screening to inform the use of abacavir sulfate in
of 18 820 patients. BMJ 2004;329:15-9. the treatment of HIV. Pharmacoeconomics 2010;28:1025-39.
J ALLERGY CLIN IMMUNOL PIRMOHAMED, OSTROV, AND PARK 243
VOLUME 136, NUMBER 2

29. Nieves Calatrava D, Calle-Martin Ode L, Iribarren-Loyarte JA, Rivero-Roman A, 53. Landsteiner K, Jacobs J. Studies on the sensitization of animals with simple
Garcia-Bujalance L, Perez-Escolano I, et al. Cost-effectiveness analysis of HLA- chemical compounds. J Exp Med 1935;61:643-56.
B*5701 typing in the prevention of hypersensitivity to abacavir in HIV1 patients 54. Esser PR, Wolfle U, Durr C, von Loewenich FD, Schempp CM, Freudenberg MA,
in Spain. Enferm Infecc Microbiol Clin 2010;28:590-5. et al. Contact sensitizers induce skin inflammation via ROS production and hyal-
30. Schackman BR, Scott CA, Walensky RP, Losina E, Freedberg KA, Sax PE. The uronic acid degradation. PLoS One 2012;7:e41340.
cost-effectiveness of HLA-B*5701 genetic screening to guide initial antiretroviral 55. Friedmann PS, Moss C, Shuster S, Simpson JM. Quantitative relationships be-
therapy for HIV. AIDS 2008;22:2025-33. tween sensitizing dose of DNCB and reactivity in normal subjects. Clin Exp Im-
31. Rauch A, Nolan D, Martin A, McKinnon E, Almeida C, Mallal S. Prospective munol 1983;53:709-15.
genetic screening decreases the incidence of abacavir hypersensitivity reactions 56. Pickard C, Louafi F, McGuire C, Lowings K, Kumar P, Cooper H, et al. The cuta-
in the Western Australian HIV cohort study. Clin Infect Dis 2006;43:99-102. neous biochemical redox barrier: a component of the innate immune defenses
32. Waters LJ, Mandalia S, Gazzard B, Nelson M. Prospective HLA-B*5701 against sensitization by highly reactive environmental xenobiotics. J Immunol
screening and abacavir hypersensitivity: a single centre experience. AIDS 2009;183:7576-84.
2007;21:2533-4. 57. Pickard C, Smith AM, Cooper H, Strickland I, Jackson J, Healy E, et al. Investi-
33. Zucman D, Truchis P, Majerholc C, Stegman S, Caillat-Zucman S. Prospective gation of mechanisms underlying the T-cell response to the hapten 2,4-
screening for human leukocyte antigen-B*5701 avoids abacavir hypersensitivity dinitrochlorobenzene. J Invest Dermatol 2007;127:630-7.
reaction in the ethnically mixed French HIV population. J Acquir Immune Defic 58. Jenkins RE, Meng X, Elliott VL, Kitteringham NR, Pirmohamed M, Park BK.
Syndr 2007;45:1-3. Characterisation of flucloxacillin and 5-hydroxymethyl flucloxacillin haptenated
34. White KD, Chung WH, Hung SI, Mallal S, Phillips EJ. Evolving models of the HSA in vitro and in vivo. Proteomics Clin Appl 2009;3:720-9.
immunopathogenesis of T-cell mediated drug allergy: the role of host, pathogens, 59. Whitaker P, Meng X, Lavergne SN, El-Ghaiesh S, Monshi M, Earnshaw C, et al.
and drug response. J Allergy Clin Immunol 2015;136:219-34. Mass spectrometric characterization of circulating and functional antigens
35. Chung WH, Hung SI, Hong HS, Hsih MS, Yang LC, Ho HC, et al. Medical ge- derived from piperacillin in patients with cystic fibrosis. J Immunol 2011;187:
netics: a marker for Stevens-Johnson syndrome. Nature 2004;428:486. 200-11.
36. Hung SI, Chung WH, Jee SH, Chen WC, Chang YT, Lee WR, et al. Genetic sus- 60. Batchelor FR, Dewdney JM, Gazzard D. Penicillin allergy: the formation of the
ceptibility to carbamazepine-induced cutaneous adverse drug reactions. Pharma- penicilloyl determinant. Nature 1965;206:362-4.
cogenet Genomics 2006;16:297-306. 61. Christie G, Coleman JW, Newby S, McDiarmaid-Gordon A, Hampson JP, Breck-
37. Man CB, Kwan P, Baum L, Yu E, Lau KM, Cheng AS, et al. Association between enridge AM, et al. A survey of the prevalence of penicillin-specific IgG, IgM and
HLA-B*1502 allele and antiepileptic drug-induced cutaneous reactions in Han IgE antibodies detected by ELISA and defined by hapten inhibition, in patients
Chinese. Epilepsia 2007;48:1015-8. with suspected penicillin allergy and in healthy volunteers. Br J Clin Pharmacol
38. Locharernkul C, Loplumlert J, Limotai C, Korkij W, Desudchit T, Tongkobpetch S, 1988;25:381-6.
et al. Carbamazepine and phenytoin induced Stevens-Johnson syndrome is associ- 62. Levine BB, Ovary Z. Studies on the mechanism of the formation of the penicillin
ated with HLA-B*1502 allele in Thai population. Epilepsia 2008;49:2087-91. antigen. III. The N-(D-alpha-benzylpenicilloyl) group as an antigenic determinant
39. Tassaneeyakul W, Tiamkao S, Jantararoungtong T, Chen P, Lin SY, Chen WH, responsible for hypersensitivity to penicillin G. J Exp Med 1961;114:875-904.
et al. Association between HLA-B*1502 and carbamazepine-induced severe cuta- 63. Ley AB, Harris JP, Brinkley M, Liles B, Jack JA, Cahan A. Circulating antibody
neous adverse drug reactions in a Thai population. Epilepsia 2010;51:926-30. directed against penicillin. Science 1958;127:1118-9.
40. Ding WY, Lee CK, Choon SE. Cutaneous adverse drug reactions seen in a tertiary 64. Brander C, Mauri-Hellweg D, Bettens F, Rolli H, Goldman M, Pichler WJ. Het-
hospital in Johor, Malaysia. Int J Dermatol 2010;49:834-41. erogeneous T cell responses to beta-lactam-modified self-structures are observed
41. Mehta TY, Prajapati LM, Mittal B, Joshi CG, Sheth JJ, Patel DB, et al. Associa- in penicillin-allergic individuals. J Immunol 1995;155:2670-8.
tion of HLA-B*1502 allele and carbamazepine-induced Stevens-Johnson syn- 65. El-Ghaiesh S, Monshi MM, Whitaker P, Jenkins R, Meng X, Farrell J, et al. Char-
drome among Indians. Indian J Dermatol Venereol Leprol 2009;75:579-82. acterization of the antigen specificity of T-cell clones from piperacillin-
42. Alfirevic A, Jorgensen AL, Williamson PR, Chadwick DW, Park BK, Pirmo- hypersensitive patients with cystic fibrosis. J Pharmacol Exp Ther 2012;341:
hamed M. HLA-B locus in Caucasian patients with carbamazepine hypersensitiv- 597-610.
ity. Pharmacogenomics 2006;7:813-8. 66. Meng X, Jenkins RE, Berry NG, Maggs JL, Farrell J, Lane CS, et al. Direct ev-
43. Lonjou C, Borot N, Sekula P, Ledger N, Thomas L, Halevy S, et al. A European idence for the formation of diastereoisomeric benzylpenicilloyl haptens from ben-
study of HLA-B in Stevens-Johnson syndrome and toxic epidermal necrolysis zylpenicillin and benzylpenicillenic acid in patients. J Pharmacol Exp Ther 2011;
related to five high-risk drugs. Pharmacogenet Genomics 2008;18:99-107. 338:841-9.
44. Ikeda H, Takahashi Y, Yamazaki E, Fujiwara T, Kaniwa N, Saito Y, et al. HLA 67. Carey MA, van Pelt FN. Immunochemical detection of flucloxacillin adduct for-
class I markers in Japanese patients with carbamazepine-induced cutaneous mation in livers of treated rats. Toxicology 2005;216:41-8.
adverse reactions. Epilepsia 2010;51:297-300. 68. Andrews E, Daly AK. Flucloxacillin-induced liver injury. Toxicology 2008;254:
45. Kaniwa N, Saito Y, Aihara M, Matsunaga K, Tohkin M, Kurose K, et al. HLA-B 158-63.
locus in Japanese patients with anti-epileptics and allopurinol-related Stevens- 69. Park BK, Boobis A, Clarke S, Goldring CE, Jones D, Kenna JG, et al. Managing
Johnson syndrome and toxic epidermal necrolysis. Pharmacogenomics 2008;9: the challenge of chemically reactive metabolites in drug development. Nat Rev
1617-22. Drug Discov 2011;10:292-306.
46. Chen P, Lin JJ, Lu CS, Ong CT, Hsieh PF, Yang CC, et al. Carbamazepine- 70. Aithal GP, Ramsay L, Daly AK, Sonchit N, Leathart JB, Alexander G, et al. He-
induced toxic effects and HLA-B*1502 screening in Taiwan. N Engl J Med patic adducts, circulating antibodies, and cytokine polymorphisms in patients
2011;364:1126-33. with diclofenac hepatotoxicity. Hepatology 2004;39:1430-40.
47. Chen Z, Liew D, Kwan P. Effects of a HLA-B*15:02 screening policy on antiep- 71. Eliasson E, Kenna JG. Cytochrome P450 2E1 is a cell surface autoantigen in
ileptic drug use and severe skin reactions. Neurology 2014;83:2077-84. halothane hepatitis. Mol Pharmacol 1996;50:573-82.
48. Genin E, Chen DP, Hung SI, Sekula P, Schumacher M, Chang PY, et al. HLA- 72. Metushi IG, Sanders C. Acute Liver Study Group, Lee WM, Uetrecht J. Detection
A*31:01 and different types of carbamazepine-induced severe cutaneous adverse of anti-isoniazid and anti-cytochrome P450 antibodies in patients with isoniazid-
reactions: an international study and meta-analysis. Pharmacogenomics J 2014; induced liver failure. Hepatology 2014;59:1084-93.
14:281-8. 73. Hardy KD, Wahlin MD, Papageorgiou I, Unadkat JD, Rettie AE, Nelson SD.
49. McCormack M, Alfirevic A, Bourgeois S, Farrell JJ, Kasperaviciute D, Carring- Studies on the role of metabolic activation in tyrosine kinase inhibitor-
ton M, et al. HLA-A*3101 and carbamazepine-induced hypersensitivity reactions dependent hepatotoxicity: induction of CYP3A4 enhances the cytotoxicity of la-
in Europeans. N Engl J Med 2011;364:1134-43. patinib in HepaRG cells. Drug Metab Dispos 2014;42:162-71.
50. Ozeki T, Mushiroda T, Yowang A, Takahashi A, Kubo M, Shirakata Y, et al. 74. Pichler WJ. Delayed drug hypersensitivity reactions. Ann Intern Med 2003;139:
Genome-wide association study identifies HLA-A*3101 allele as a genetic risk 683-93.
factor for carbamazepine-induced cutaneous adverse drug reactions in Japanese 75. Yun J, Adam J, Yerly D, Pichler WJ. Human leukocyte antigens (HLA) associated
population. Hum Mol Genet 2011;20:1034-41. drug hypersensitivity: consequences of drug binding to HLA. Allergy 2012;67:
51. Plumpton CO, Yip VLM, Alfirevic A, Marson AG, Pirmohamed M, Hughes DA. 1338-46.
Cost-effectiveness of screening for HLA-A*31:01 prior to initiation of carbamaz- 76. Adam J, Pichler WJ, Yerly D. Delayed drug hypersensitivity: models of T-cell
epine in epilepsy. Epilepsia 2015;56:556-63. stimulation. Br J Clin Pharmacol 2011;71:701-7.
52. Lichtenfels M, Farrell J, Ogese MO, Bell CC, Eckle S, McCluskey J, et al. HLA 77. Naisbitt DJ, ONeill PM, Pirmohamed M, Park BK. Synthesis and reactions of ni-
restriction of carbamazepine-specific T-Cell clones from an HLA-A*31:01- troso sulphamethoxazole with biological nucleophiles: implications for immune
positive hypersensitive patient. Chem Res Toxicol 2014;27:175-7. mediated toxicity. Bioorg Med Chem Lett 1996;6:1511-6.
244 PIRMOHAMED, OSTROV, AND PARK J ALLERGY CLIN IMMUNOL
AUGUST 2015

78. Naisbitt DJ, Hough SJ, Gill HJ, Pirmohamed M, Kitteringham NR, Park BK. 88. Castrejon JL, Berry N, El-Ghaiesh S, Gerber B, Pichler WJ, Park BK, et al. Stim-
Cellular disposition of sulphamethoxazole and its metabolites: implications for ulation of human T cells with sulfonamides and sulfonamide metabolites.
hypersensitivity. Br J Pharmacol 1999;126:1393-407. J Allergy Clin Immunol 2010;125:411-8.e4.
79. Gill HJ, Maggs JL, Madden S, Pirmohamed M, Park BK. The effect of flucona- 89. Nassif A, Bensussan A, Boumsell L, Deniaud A, Moslehi H, Wolkenstein P, et al.
zole and ketoconazole on the metabolism of sulphamethoxazole. Br J Clin Phar- Toxic epidermal necrolysis: effector cells are drug-specific cytotoxic T cells.
macol 1996;42:347-53. J Allergy Clin Immunol 2004;114:1209-15.
80. Callan HE, Jenkins RE, Maggs JL, Lavergne SN, Clarke SE, Naisbitt DJ, et al. 90. Schnyder B, Burkhart C, Schnyder-Frutig K, von Greyerz S, Naisbitt DJ, Pirmo-
Multiple adduction reactions of nitroso sulfamethoxazole with cysteinyl residues hamed M, et al. Recognition of sulfamethoxazole and its reactive metabolites by
of peptides and proteins: implications for hapten formation. Chem Res Toxicol drug-specific CD41 T cells from allergic individuals. J Immunol 2000;164:
2009;22:937-48. 6647-54.
81. Naisbitt DJ, Gordon SF, Pirmohamed M, Burkhart C, Cribb AE, Pichler WJ, et al. 91. Pichler WJ, Naisbitt DJ, Park BK. Immune pathomechanism of drug hypersensi-
Antigenicity and immunogenicity of sulphamethoxazole: demonstration of tivity reactions. J Allergy Clin Immunol 2011;127(suppl):S74-81.
metabolism-dependent haptenation and T-cell proliferation in vivo. Br J Pharma- 92. Pichler W, Watkins S. Interaction of small molecules with specific immune recep-
col 2001;133:295-305. tors: the p-i concept and its consequences. Curr Immunol Rev 2014;10:7-18.
82. Ogese MO, Jenkins RE, Maggs JL, Meng X, Whitaker P, Peckham D, et al. Char- 93. Faulkner L, Martinsson K, Santoyo-Castelazo A, Cederbrant K, Schuppe-Koisti-
acterization of peroxidases expressed in human antigen presenting cells and anal- nen I, Powell H, et al. The development of in vitro culture methods to characterize
ysis of the covalent binding of nitroso sulfamethoxazole to myeloperoxidase. primary T-cell responses to drugs. Toxicol Sci 2012;127:150-8.
Chem Res Toxicol 2015;28:144-54. 94. Gibson A, Ogese M, Sullivan A, Wang E, Saide K, Whitaker P, et al. Negative
83. Elsheikh A, Lavergne SN, Castrejon JL, Farrell J, Wang H, Sathish J, et al. Drug regulation by PD-L1 during drug-specific priming of IL-22-secreting T cells and
antigenicity, immunogenicity, and costimulatory signaling: evidence for forma- the influence of PD-1 on effector T cell function. J Immunol 2014;192:2611-21.
tion of a functional antigen through immune cell metabolism. J Immunol 2010; 95. Ostrov DA, Grant BJ, Pompeu YA, Sidney J, Harndahl M, Southwood S, et al.
185:6448-60. Drug hypersensitivity caused by alteration of the MHC-presented self-peptide
84. Naisbitt DJ, Farrell J, Gordon SF, Maggs JL, Burkhart C, Pichler WJ, et al. Co- repertoire. Proc Natl Acad Sci U S A 2012;109:9959-64.
valent binding of the nitroso metabolite of sulfamethoxazole leads to toxicity and 96. Illing PT, Vivian JP, Dudek NL, Kostenko L, Chen Z, Bharadwaj M, et al. Im-
major histocompatibility complex-restricted antigen presentation. Mol Pharmacol mune self-reactivity triggered by drug-modified HLA-peptide repertoire. Nature
2002;62:628-37. 2012;486:554-8.
85. Sanderson JP, Naisbitt DJ, Farrell J, Ashby CA, Tucker MJ, Rieder MJ, et al. Sul- 97. Jensen PE. Recent advances in antigen processing and presentation. Nat Immunol
famethoxazole and its metabolite nitroso sulfamethoxazole stimulate dendritic 2007;8:1041-8.
cell costimulatory signaling. J Immunol 2007;178:5533-42. 98. Neefjes J, Ovaa H. A peptide’s perspective on antigen presentation to the immune
86. Alfirevic A, Vilar FJ, Alsbou M, Jawaid A, Thomson W, Ollier WE, et al. TNF, system. Nat Chem Biol 2013;9:769-75.
LTA, HSPA1L and HLA-DR gene polymorphisms in HIV-positive patients with 99. Adam J, Wuillemin N, Watkins S, Jamin H, Eriksson KK, Villiger P, et al. Aba-
hypersensitivity to cotrimoxazole. Pharmacogenomics 2009;10:531-40. cavir induced T cell reactivity from drug naive individuals shares features of allo-
87. Burkhart C, von Greyerz S, Depta JP, Naisbitt DJ, Britschgi M, Park KB, et al. In- immune responses. PLoS One 2014;9:e95339.
fluence of reduced glutathione on the proliferative response of sulfamethoxazole- 100. Meng X, Lawrenson AS, Berry NG, Maggs JL, French NS, Back DJ, et al. Aba-
specific and sulfamethoxazole-metabolite-specific human CD41 T-cells. Br J cavir forms novel cross-linking abacavir protein adducts in patients. Chem Res
Pharmacol 2001;132:623-30. Toxicol 2014;27:524-35.