Advances in Molecular Toxicology

Visit the link below to download the full version of this book:

https://medipdf.com/product/advances-in-molecular-toxicology-9/

Click Download Now

PREFACE

This sixth volume continues in the tradition established by the first five in
which contributions emphasize the diversity of subjects, views, techniques,
and perspectives within this large discipline. It is hoped that the readers’
own unique perspective is challenged and enlightened by the stimulating
chapters within. The editor is pleased to welcome Jacqueline M. Heilman,
Ph.D. as co-editor. She brings a novel expertise as a toxicology consultant
from Exponent, Inc. in the areas of food and chemical regulation and risk
assessment.
Chapter 1 surveys the nonnucleoside reverse transcriptase inhibitors
(NNRTIs) that are key components of current combined antiretroviral
therapies used in the treatment of HIV AIDS across the globe. The toxicities
and bioactivation pathways of the major NNRTIs—Nevirapine, Efavirenz,
Etravirine, and Rilpivirine—are the focus of the review. Biotransformation
gives rise to numerous reactive metabolites that might contribute to toxic-
ity, and elaboration of these bioactivation pathways can contribute to
development of more effective therapeutic agents.
As in many parts of the world, in South America, genetically modified
strains of plants that are resistant to herbicides and pesticides are now widely
cultivated in an effort to improve crop yields to sate an ever increasing
human population. Glyphosate-based herbicides are very widely and exten-
sively employed. Andrés E. Carrasco and coauthors review in Chapter 2 the
toxic effects on humans and in animal models. An emphasis is on the
identification and development of biomarkers for exposure and safety
assessment.
Chapter 3 contains a thorough discussion of fetal alcohol syndrome and
fetal alcohol syndrome spectrum disorder with particular emphasis on the
utility of the Medaka fish (Oryzias lapites) as a model organism for studying
this disease. Haron et al. detail the importance of developmental timing
with respect to the various effects of ethanol exposure on this model system
resulting from exposure during very specific developmental windows. In
the course of describing the developmental effects of ethanol on Medaka,
comparison to other model systems, especially zebrafish (Danio rerio), and
relevance to the human disease are discussed.
Epigenetic factors—DNA methylation, histone modifications, and
micro RNAs—are significant factors modulating gene expression. Epige-
netic changes across the genome are known to occur in cancers and
developmental defects. A number of well-known toxins have been demon-
strated to manifest epigenetic changes that may contribute to their

xi
xii Preface

subsequent toxicities. Kathryn A. Bailey and Rebecca C. Fry review epi-

genetics and the epigenetic changes wrought by three distinct toxins:
various metals, bisphenol A, and benzo(a)pyrene in Chapter 4.
In Chapter 5, Gema Arribas-Lorenzo and Francisco J. Morales present
the current status of acrylamide toxicity. Once thought to be chiefly a
workplace toxin with additional human exposure from polyacrylamides
used in many applications, food is now known to be the chief source of
human exposure with cooking of starchy foods now identified as a prime
generator. Neurotoxicity, reproductive toxicity, and carcinogenesis are
major manifestations of exposure, but linkages between normal long-term
human exposure and these manifestations are weak. The need for better,
more sensitive exposure assessments and consequences in human popula-
tions is emphasized.
Hassan K. Obied et al. review the pharmacology of olive oil biophenols
in Chapter 6. The authors describe olive oil as a complex mixture contain-
ing a variety of molecules with possible biological functions including
antioxidant activity as well as potential interactions with targets such as
lipids, proteins, carbohydrates, and nucleic acids. The major biophenols
found in olive oil include hydroxytyrosol, tyrosol, secoiridoid derivatives of
tyrosols, verbascoside, lignans, and flavonoids.
Mitomycin C is a potent cytotoxin that was discovered in the 1950s and
whose mechanism of action began to be elucidated in the following decade.
It is still employed in the clinic as a useful anticancer agent. The reductive
activation and DNA cross-linking activities have been well documented,
and much is known about sequence specificity, polymerase interactions,
and the repair process engaged. Manuel M. Paz and Chris A. Pritsos lay
all this out in rich detail in Chapter 7. Importantly, they identify
remaining unknowns and make the case for alternative biological targets
for mitomycin C.
 C H A P T E R O N E

Insights into the Role of

Bioactivation Mechanisms in the Toxic
Events Elicited by Non-nucleoside
Reverse Transcriptase Inhibitors
Sofia A. Pereira,1 Riccardo Wanke,2 M. Matilde Marques,2
Emı́lia C. Monteiro,1 and Alexandra M. M. Antunes2,*

Contents
1. Introduction 2
2. The Non-nucleoside Reverse Transcriptase Inhibitors 3
2.1. Nevirapine 5
2.2. Efavirenz 16
2.3. Etravirine 21
2.4. Rilpivirine 23
3. Conclusions 26
Acknowledgments 26
References 27

Abstract
The indisputable benefits of combined antiretroviral therapies (cARTs) have
lead to a dramatic change in the prognosis of human immunodeficiency virus
(HIV) infection; a life-threatening disease a few decades ago is now perceived
as a chronic illness in developed countries. However, as the eradication of HIV
seems unlikely in the near future, chronic treatment with cART is unavoidable
and increased concerns regarding the long-term adverse effects of these thera-
pies are emerging. According to the World Health Organization, the most
globally prescribed initial cART includes a non-nucleoside reverse transcriptase
inhibitor (NNRTI). The currently approved NNRTIs are a class of chemically
distinct compounds that share the possibility of undergoing biotransformation
into electrophilic metabolites capable of reacting with biomacromolecules to

1
Centro de Estudos de Doenças Crónicas (CEDOC), Departamento de Farmacologia, Faculdade de Ciências
Médicas, Universidade Nova de Lisboa, Lisboa, Portugal
2
Centro de Quı́mica Estrutural, Instituto Superior Técnico, Universidade Técnica de Lisboa, Lisboa, Portugal
*Corresponding author: Tel.: þ 351-21-8417627; Fax: þ 351-21-8464455
E-mail address: [email protected]

Advances in Molecular Toxicology, Volume 6 # 2012 Elsevier B.V.

ISSN 1872-0854, http://dx.doi.org/10.1016/B978-0-444-59389-4.00001-X All rights reserved.

1
2 Sofia A. Pereira et al.

afford covalent DNA and protein adducts that could be at the genesis of toxicity.
Insights into the bioactivation mechanisms of the NNRTIs nevirapine, efavirenz,
etravirine, and rilpivirine are presented in this review.

1. Introduction
The human immunodeficiency virus (HIV) is the causative agent of
acquired immunodeficiency syndrome (AIDS), an epidemiological global
crisis with over 33 millions of infected individuals worldwide. HIV/AIDS-
related morbidity and mortality have been largely reduced by the use of
combined antiretroviral therapy (cART) [1]. As a result, this disease has
become perceived as a chronic illness in developed countries. However, as
the eradication of HIV seems unlikely in the near future, chronic treatment
with cART is unavoidable and new concerns regarding the toxicity of these
long-term treatments are emerging.
The premature onset of conditions that resemble aging and aging-related
comorbidities such as cancer, cardiovascular, neurocognitive, and bone
diseases in HIV-infected people [2] is becoming increasingly prevalent.
Moreover, as the infection prevails among persons of reproductive age,
infertility is also a growing concern in the cART era [3]. The increased
incidence of these conditions, which have been credited to the chronic use
of antiretrovirals, leads both to concerns regarding the long-term adverse
effects of cART in children and to larger needs for suitable enduring
treatment options for an aging patient population. Therefore, understand-
ing the molecular basis underlying the toxicities induced by antiretrovirals is
a pressing issue, not only to develop safer and more effective drugs but also
to achieve accurate risk/benefit estimations that can guide decisions on the
currently available therapeutic choices.
Drugs, as any other xenobiotics, undergo biotransformation to more
hydrophilic derivatives so that excretion can occur. However, metabolic
pathways can also be responsible for the formation of reactive (primarily
electrophilic) species capable of reacting with biomacromolecules to afford
covalent protein and DNA adducts that may elicit direct cell toxicity,
trigger an immune response, and/or initiate mutagenicity/carcinogenicity.
Drug bioactivation is, therefore, a frequent event at the onset of drug-
induced toxicity [4,5]. For instance, in drug-induced hypersensitivity reac-
tions (HSRs), which are immune-mediated processes, drug–protein
adducts, unlike the parent drugs or their metabolites, interact directly with
immune receptors acting as antigens [6–8]. Likewise, DNA modification by
toxic electrophiles can be an early event in mutagenic/carcinogenic pro-
cesses, should the DNA lesions remain unrepaired or undergo erroneous
repair [9].
Bioactivation in Toxic Events Elicited by NNRTIs 3

Whereas the chemical structure of a drug (or its Phase I metabolites) is

the primary factor ruling its propensity to be converted into a reactive/toxic
metabolite [5], other concurrent risk factors and patient characteristics, such
as genetic polymorphisms in metabolizing enzymes and disease status, can
exacerbate drug toxicity. HIV/AIDS disease has been shown to play a
significant role in drug bioactivation and detoxification [10]; being asso-
ciated with a glutathione (GSH)-depleted state [11], the presence of HIV
can be seen as a predisposing condition for drug-induced toxic outcomes.
Indeed, whereas GSH conjugation is a typical detoxification pathway for
reactive electrophiles, its depletion may hamper an efficient scavenging of
reactive metabolites, which will conceivably become available to react with
DNA and proteins, eliciting toxic responses.

2. The Non-nucleoside Reverse Transcriptase

Inhibitors
The non-nucleoside reverse transcriptase inhibitors (NNRTIs) are a
group of structurally diverse compounds that noncompetitively prevent the
conversion of viral RNA into DNA. According to the World Health
Organization, the most globally prescribed initial cART includes an
NNRTI [12] (Scheme 1); as such, the worldwide exposure to this class of
antiretrovirals is very significant.
In 1996, nevirapine (11-cyclopropyl-5,11-dihydro-4-methyl-6H-
dipyrido[3,2-b:20 ,30 -e][1,4]diazepin-6-one, NVP, 1), a dipyridodiazepinone,
was the first drug of this class to be approved by the U.S. Food and Drug
Administration (FDA) for use in combination therapy for HIV-1 infection.
The benzoxazinone efavirenz [(S)-6-chloro-4-(cyclopropylethynyl)-1,4-
dihydro-4-(trifluoromethyl)-2H-3,1-benzoxazin-2-one, EFV, 2] was
approved 2 years later.
NVP and EFV are known as the first-generation NNRTIs and are still
keystones of first-line cART. However, despite their efficacy and although
individual susceptibilities to adverse effects differ among patients, adminis-
tration of these two NNRTIs is associated with a variety of toxic responses.
Both EFV and NVP may cause skin rash, hepatotoxicity, and neuropsy-
chiatric events, albeit they differ in individual propensities to induce each of
these adverse effects [13]. NVP shows a higher risk of cutaneous and hepatic
reactions, and EFV a higher risk of central nervous system (CNS) effects
[13]. Whereas NVP and EFV toxicities have been advocated as idiosyncratic
in nature, increasing evidence supports the involvement of bioactivation
mechanisms.
Although the toxic effects associated with prolonged treatment with
these antiretrovirals remain mostly unaddressed, recent epidemiological
4 Sofia A. Pereira et al.

O
HN
CF3
Cl
O
N N
N
N O
H

NVP, 1 EFV, 2

H H
O N N N
N
Br N
N
N N
N N NH2
H
ETV, 3 RPV, 4

Scheme 1 Structures of the NNRTIs currently available for use in combined antire-
troviral therapy. NVP, nevirapine; EFV, efavirenz; ETV, etravirine; RPV, rilpivirine.

studies of the occurrence and incidence of non-AIDS-defining cancers

(NADCs) in HIV-positive individuals suggested that, among typical
cART components, only the first-generation NNRTIs are associated
with an increased risk of NADCs [14]. This implies that these drugs may
themselves be carcinogenic, which adds a further concern about their
chronic use.
More recently, two second-generation NNRTIs were added to the set
of treatment possibilities with the aim to overcome the most common
adverse effects of first-generation NNRTIs, as well as their viral cross-
resistance [15]. Etravirine (4-[[6-amino-5-bromo-2-[(4-cyanophenyl)
amino]4-pyrimidinyl]oxy]-3,5-dimethylbenzonitrile, ETV, 3) and rilpivir-
ine (4-[[4-[[4-[(E)-2-cyanoethenyl]-2,6-dimethylphenyl]amino]-2-pyri-
midinyl]amino]benzonitrile, RPV, 4) were approved in 2008 and 2011,
respectively. These pyrimidine derivatives are active against both wild-type
and first-generation NNRTI-resistant HIV strains, most likely because their
conformational flexibility allows different binding modes to the NNRTI
binding site of the HIV-1 reverse transcriptase (RT) when RT mutations
are present [16]. However, postmarketing reports of severe skin rash and
HSRs to ETV [17] and of RPV-associated depressive disorders [18] are
a cause for concern when contemplating chronic administration regimens.
It should be emphasized that patients who start second-generation
NNRTI-based cART have a history of virus resistance and typically
Bioactivation in Toxic Events Elicited by NNRTIs 5

experienced several drug-induced side effects. Hence, the administration of

a potentially (geno)toxic drug to these already debilitated patients must be
considered carefully, and the development of reliable prognostic tools for
early risk/benefit estimations is urgent.
As recently approved drugs, limited information is available about the
involvement of bioactivation in the initiation of toxic outcomes associated
with second-generation NNRTIs. Nevertheless, a critical discussion of
their structural features and/or metabolites potentially prone to react with
bionucleophiles is included in this review. Additionally, we present a
detailed appraisal of recent evidence for the contribution of NVP and
EFV bioactivation to the onset of drug-induced toxicity.

2.1. Nevirapine
NVP is one of the most prescribed antiretroviral drugs in low-resource
countries, both as a single drug, to prevent mother-to-child HIV transmis-
sion, and as a component of cART [19,20]. The low cost of the drug and its
availability as a generic prescription are plausible explanations for the
widespread use of this NNRTI in developing countries, where HIV is
more prevalent. Moreover, in developed countries, NVP is still a first-line
choice of initial therapy regimens for children younger than 3 years of age,
who have not been exposed to NVP as part of maternal–infant prophylaxis
[21]. Indeed, the high-efficacy levels of the drug, favorable lipid profile [22],
and suitability for use during pregnancy [23], as well as in contexts of
neuropsychiatric abnormalities or drug dependence [24], have granted
NVP-based regimens a significant role in HIV-1 treatment strategies.
The cART NVP schedules are recommended to be initiated with a 200-
mg dose for the first 14 days, to minimize toxicity risks, and then are
followed by a 400 mg daily dose [25]. Until 2011, NVP was only available
on the market in an immediate-release formulation approved for twice-
daily dosing, which had the disadvantage of poor adherence, particularly in
low-resource settings. However, the worldwide use of NVP is expected
to be heightened as a result of the recent approval of an extended-release
once-daily formulation. Nonetheless, the prospective increase in NVP use
worldwide should be regarded with caution, partly on account of toxicity
considerations.

2.1.1. Nevirapine toxicity

Despite unequivocal benefits, a pitfall of NVP regimens is their association
with serious and clinically restrictive idiosyncratic side effects. Severe, life-
threatening, and in some cases, fatal liver and skin toxicity have been
reported in patients treated with NVP [26–30]. The most likely dangerous
toxic events in the skin are a morbilliform eruption (13–28% of patients) and
a systemic HSR or severe rash (with an incidence of 8%) that is the major
6 Sofia A. Pereira et al.

cause of drug discontinuation [26–28,31]. Due to the potentially fatal toxic

reactions, the FDA issued a black box label warning on NVP [32], and a
similar recommendation was issued by the European Agency for Evaluation of
Medicinal Products (EMEA) [33].
These adverse events occur most frequently during the first 6 weeks of
therapy, and women, including those who are pregnant or of Asian ethnic-
ity, seem to have an increased risk of developing NVP-induced toxicity
[29,31,34–36]. A high CD4þ lymphocyte count and a detectable viral load
are additional risk factors [30,36]. Therefore, NVP therapy should be
initiated only in cART-naive men and women with CD4þ lymphocyte
counts below 400 and 250 cells/mm3, respectively [25]. Moreover, due to
immunocompetence-linked toxicity, NVP is not recommended as part of
postexposure prophylaxis [37]. This association with higher CD4 cell
counts suggests a role for an immune response at the onset of NVP adverse
reactions, and two different allergenic pathways have been proposed [38]:
cutaneous adverse events are thought to be most likely major histocompati-
bility complex (MHC) class I-mediated (CD8 T cells) while NVP-induced
hepatotoxicity is probably MHC class II-mediated (CD4 T cells).
Since susceptibilities to NVP-induced toxicity differ among patients, it
has been postulated that idiosyncratic events can be genetically conditioned
[39,40]. Several reports have described a correlation between HLA variants
and NVP adverse reactions [39–42]. Moreover, a recent toxicogenomics
study has indicated that, besides immune response pathways, polymorphisms
in drug metabolism should also be considered as risk factors in NVP-related
adverse events [38].
A recent study with male Wistar rats showed that NVP-induced testic-
ular toxicity was detected with both biochemical and histopathology para-
meters [43]. Genetic toxicology tests, including microbial and mammalian
cell gene mutation assays and cytogenetic assays, have provided no evidence
that NVP is either mutagenic or clastogenic in vitro [44]. However, while
conclusive evidence for NVP carcinogenicity in humans has yet to be
presented, long-term administration to mice and rats showed increased
incidences of hepatocellular adenomas and carcinomas in NVP-treated
animals [44].

2.1.2. Nevirapine pharmacokinetics

As a highly lipophilic molecule, NVP is readily (greater than 90%) absorbed
after oral administration. The extent of NVP absorption is not affected by
food intake or pH. It has a long elimination half-life (t1/2 ¼  45 h) and its
peak plasma concentration is attained within 4 h, following a single 200-mg
oral dose [45]. NVP is about 60% bound to plasma proteins, crosses the
placenta and blood–brain barrier, and is found in breast milk [46]. Despite
the lack of data on NVP transporters [47,48], it was recently shown that
Bioactivation in Toxic Events Elicited by NNRTIs 7

ABCC10 (MRP7), an efflux transporter highly expressed in liver, intestine,

and peripheral blood cells, influences NVP pharmacokinetics [49].
NVP interacts with cytochrome P450 (CYP) enzymes both as a sub-
strate and as an inducer. It is extensively metabolized by CYP into oxidized
metabolites that undergo subsequent glucuronidation. Urinary excretion
of glucuronide conjugates represents the primary route of NVP biotrans-
formation and elimination in humans [50]. In all species investigated,
CYP-mediated oxidation of NVP consistently involves the formation of
2-, 3-, and 8-hydroxy-NVP(2-, 3-, and 8-OH-NVP), 12-hydroxy-NVP
(12-OH-NVP), and 4-carboxy-NVP [50–54] (5–9, Scheme 2).
Different human CYP isoforms are involved in NVP biotransformation
into its hydroxylated metabolites [52]. Whereas the formation of
2-OH-NVP (5) and 3-OH-NVP (6) is exclusively mediated by CYP3A
and CYP2B6, respectively, the conversion to 8-OH-NVP (7) is attributed
to a group of subfamilies: CYP3A4, CYP2B6, and CYP2D6. The formation

O
HO O
HO O
HN
HN
Oxidation
N N N N
Glucuronide N N
UGT

9
8

UGT
CYP3A4 UGT

O
O
HN
HN Glucuronide

CYP3A4 N N N CYP2B6
N N N
HO
O
NVP, 1 HN
5 HO
CYP3A4, CYP2B6,
N N N
CYP2D6
O
HN
UGT 6
OH

N N N Glucuronide
UGT

Scheme 2 Phase I NVP metabolites and their subsequent Phase II glucuronidation.

CYP, cytochrome P450; UGT, uridine-50 -diphospho-glucuronosyltransferase.
8 Sofia A. Pereira et al.

of 12-OH-NVP (8) is credited to CYP3A4, with the possible involvement

of CYP2D6 and CYP2C9 [52]. Secondary oxidation of 12-OH-NVP (8)
yields 4-carboxy-NVP (9) [50], but the enzymes involved remain to be
elucidated.
NVP induces CYP3A4 and CYP2B6 by approximately 20–25% [45];
this autoinduction results in a 1.5- to 2-fold increase in the apparent
systemic clearance upon chronic treatment, with concurrent decrease in
the terminal phase half-life in plasma. The induction is complete within
28 days [45,50], at which point a steady-state NVP plasma concentration is
attained [55]. Different studies have addressed the impact of CYP2B6
variation on NVP pharmacokinetics [56–58]. Recently, a pharmacometric
analysis quantified this impact, and a new hypothesis, regarding the contri-
bution of CYP2C19 to NVP biotransformation, was proposed [59]. Skin
rash has been related to a CYP2B6 516TT polymorphism [38], consistent
with loss of CYP2B6 function, and shunting of NVP metabolism to
alternative pathways. In addition, a decreased risk of NVP-induced hepato-
toxicity has been associated with the MDR1 3435C!T allele [47,48], a
polymorphism linked to an altered expression of the multidrug efflux pump,
P-glycoprotein.

2.1.3. Nevirapine bioactivation

Among the four NNRTIs currently on the market, NVP is the most
extensively studied regarding the molecular mechanisms underlying its
idiosyncratic toxicity [60]. Although the reasons for the adverse effects
of NVP are still unclear, increasing evidence suggests that NVP bioac-
tivation is required to initiate the toxic events induced by the parent
drug.
The reported fast recovery of a patient suffering from NVP-induced
toxic epidermal necrolysis and toxic hepatitis, upon treatment with a
continuous infusion of human immunoglobulins and a high dose
(300 mg/day) of N-acetylcysteine (NAC) [61], provided relevant clinical
support for both the involvement of metabolic activation in the NVP toxic
response and the importance of adequate GSH levels in the detoxification of
reactive NVP metabolites. Although circumstantial, this fast recovery may
have stemmed from GSH replenishment in the HIV-infected patient [62],
leading to decreased oxidative stress and/or induced detoxification of reac-
tive metabolites. Additionally, it is also plausible that the administered
NAC, an excellent nucleophile, reacted with electrophilic NVP metabolites
to form readily excreted mercapturate conjugates.
Takakusa et al. [63] have provided evidence for covalent binding of
[14C]NVP to rat and human liver microsomal proteins in vitro and to rat
liver tissue and plasma proteins in vivo, in male rats administered 20 mg
NVP/kg bw as a single oral dose. Although the reactive metabolites were
not identified, these data represent unequivocal evidence of the capacity of
Bioactivation in Toxic Events Elicited by NNRTIs 9

NVP to undergo protein haptenation upon metabolic activation. Indeed,

the generation of reactive metabolites with the potential to bind to proteins
and DNA and initiate toxic events can be attributed to different NVP
metabolic pathways (Scheme 3). Phase II conjugations, involving

O O O GS O
S HN HN
HO O

N N N
N N N

11 +H+, -H2SO4 12 NAC O

HN
SULT

HO O O N N
HN N N

N N N 14
N N N

8 13

O
HN
H O
N CYP •
O
• N N HN
O N
N N H
N
N N
18 Å HO
N
1 H [o] O
CYP HN
H O 5
N O
HÅ
O N N
O N
N N
N
10
GSH 16
Michael addition
Schiff base formation

H O H O
GS N N
NAC Adducts

N N N N
N N

17 15

Scheme 3 Hypothetic pathways for in vivo generation of covalent adducts from

12-OH-NVP (8) [involving 12-sulfoxy-NVP (11) and/or NVP quinone-methide
(13)] and phenolic NVP metabolites [involving arene oxide (16) ring opening and/or
quinone (10) formation]. CYP, cytochrome P450; GSH, glutathione; SULT, sulfo-
transferase; NAC, N-acetylcysteine.
10 Sofia A. Pereira et al.

acetylation or sulfonation, or further oxidation of the phenolic metabolites

to quinone/semiquinone species (e.g., 10, Scheme 3), are possible routes to
NVP-induced toxicity [5].
It has been proposed that 12-OH-NVP (8) is the primary metabolite
responsible for an idiosyncratic NVP-induced skin rash in Brown Norway
rats that resembles the rash occurring in man [64–67]. In particular, the
identification of the Phase II metabolite, 12-sulfoxy-NVP (11), in this rat
strain [64] suggested that sulfotransferase (SULT) activity may have a role at
the onset of immune-mediated NVP-induced skin rash. The presence of
SULTs in the skin is consistent with this hypothesis, since protein haptena-
tion could stem from direct reaction of this electrophilic Phase II metabolite
with proteins [68,69]. This mechanism is consistent with the mass spectro-
metric (MS) detection of a putative 12-(glutathion-S-yl)-NVP conjugate
(12, Scheme 3) following incubation of NVP with human liver microsomes
in the presence of GSH [70], although NVP activation to an electrophilic
quinone-methide (13, Scheme 3) could also occur without Phase II conju-
gation, either by dehydration of 12-OH-NVP (8) or by CYP-induced
dehydrogenation of NVP.
Contrasting with skin rash, NVP-induced hepatotoxicity suffers from
lack of a suitable animal model. Nevertheless, a potential role of NVP
metabolic activation in liver injury has also been suggested by Walubo
et al. [71], who observed that increased hepatotoxicity is closely associated
with enzyme induction.
The first unequivocal evidence for the binding capacity of NVP meta-
bolites to proteins in vivo was provided by Srivastava et al. [72], who
identified two NVP mercapturates, the minor one through NVP-C12
and the major one through NVP-C3 (14 and 15, respectively, Scheme 3),
in the urine of HIV-positive individuals administered NVP as part of
cART. The minor mercapturate was suggested to be formed upon
SULT-mediated bioactivation of 12-OH-NVP (8) and subsequent nucleo-
philic attack by GSH to yield 12, which then underwent catabolic degrada-
tion to 14. The involvement of Phase II electrophiles in NVP
mutagenicity/carcinogenicity may be a possible explanation for the nega-
tive results of NVP in mutagenicity assays in vitro compared to the positive
carcinogenicity in vivo [44]. Indeed, standard mutagenicity assays use exter-
nally added liver-derived metabolic systems to mimic biotransformation
absent from the majority of the mutagenicity-indicator cell lines. Such
systems generate reactive sulfate conjugates externally, with limited capacity
to penetrate the target cells. Thus, SULT-mediated bioactivation is not
detected in standard mutagenicity tests [73].
Identification of the mercapturate through NVP-C3 (15) represented
the first evidence that reactive metabolites other than 12-OH-NVP (8)
could have a role in the onset of NVP-induced toxic effects. Adduct 15 was
proposed to be formed by nucleophilic attack of GSH on an arene oxide
Bioactivation in Toxic Events Elicited by NNRTIs 11

intermediate (16, Scheme 3), followed by dehydration to yield 17, which

then underwent catabolism to 15. In addition, it cannot be excluded that
phenolic NVP metabolites [e.g., 2-OH-NVP (5)], either generated from
CYP-mediated oxidation to an arene oxide (e.g., 16, Scheme 3) and/or a
diradical (e.g., 18) precursor, may undergo metabolic activation to qui-
none/semiquinone electrophiles (e.g., 10) capable of reacting with bionu-
cleophiles through Michael-type addition and/or Schiff base formation
[74,75], leading to covalent adducts (Scheme 3).
To gain insight into the potential significance of phenolic metabolites to
NVP toxicity, we have recently investigated the chemical and enzymatic
(lactoperoxidase) oxidation of 2-OH-NVP (5) in vitro [76]. The product
profile obtained in aqueous media (e.g., the spiro derivative 19 and the
nicotinamide 20, Scheme 4) suggests the formation of a quinone-imine (21)
as an electrophilic intermediate. Since the toxicological significance of
quinone-imines has long been recognized [77], these results support the
hypothesis that an activation pathway from phenolic NVP metabolites, in
particular 2-OH-NVP, may be involved in NVP-induced toxicities. Even
though evidence for such mechanisms in vivo or in metabolically competent
systems in vitro has yet to be reported, the formation of a quinone-imine
electrophile under lactoperoxidase catalysis, together with the presence of
NVP in breast milk [46] and the frequent administration of the drug
concurrently with breastfeeding, suggests that it could have a role in the
perinatal setting. Moreover, given the presence of lactoperoxidase in tears
and saliva [78], an activation pathway involving a quinone-imine appears as
a plausible explanation for NVP-induced oral and ocular toxicity [79,80].
Despite the evidence for the role of NVP metabolism in the onset of
NVP-induced clinically restrictive effects, no consistent correlations have
been observed between the plasma concentration of any individual NVP
metabolite and the occurrence of rash and liver function abnormalities
in patients exhibiting these adverse effects within the first 6 weeks of
treatment [81]. Likewise, when the plasma concentration of NVP was

O
H O
N O O
N
NH
[o] H2O N N
NH2
HO N N O N N HN
N N N NH
O
O

2-OH-NVP (5) Quinone-imine (21) 19 20

Scheme 4 Oxidation of the phenolic NVP metabolite 2-OH-NVP (5) in aqueous

solution to the spiro derivative 19 and the nicotinamide 20, involving the transient
generation of a quinone-imine (21) [76].
12 Sofia A. Pereira et al.

correlated with NVP-induced toxic effects, the results were not conclusive
[55,82,83]. These conflicting observations may stem from a variety of
confounding factors that can conceivably affect the systemic exposure to
NVP, and probably its metabolite profile [84,85]; sex [55,86], genetics
[56–58,87], hepatitis coinfection, and pre-existing altered hepatic function
[55,88] are some of these variables. Nevertheless, these data suggest that
more suitable biomarkers of toxicity than the parent drug and its primary
metabolites should be used to assess unequivocally the role of bioactivation
in NVP-induced toxic outcomes and ultimately achieve better risk–benefit
estimations. Ideally, we should monitor the electrophiles generated from
primary NVP metabolites (e.g., 12-sulfoxy-NVP, quinone-methide, qui-
none-imine, quinones), since these are the reactive intermediates capable of
binding to biomacromolecules; however, they are very difficult to detect in
humans or experimental animals, due to their short lifetimes. Nonetheless,
the products from their reaction with bionucleophiles, such as GSH, pro-
tein, and DNA adducts, can be readily detected and quantified in body
fluids and tissues [89], giving an indirect measure of the biologically effec-
tive dose of the reactive intermediate that has reached the tissue or biofluid
under study [90]. Our group has conducted a considerable amount of work
toward the establishment of reliable and fully characterized potential bio-
markers of NVP toxicity, and the development of analytical tools to assess
these biomarkers at levels expected in vivo.
As part of this program, and with the aim of mimicking in vitro the
putative NVP metabolic pathways to reactive electrophiles, we have
conducted several studies, using 12-mesyloxy-NVP (22, Scheme 5) as
a synthetic surrogate for 12-sulfoxy-NVP [91, 92]. The choice of this
model electrophile for our in vitro studies was based on its higher stability
and easy synthesis, combined with an expected similar reactivity toward
bionucleophiles. Indeed, adducts 12 and 14 were obtained in high yields
upon reaction with GSH and NAC, respectively [91]. As indicated
above, adduct 12 was detected in NVP incubations with human liver
microsomes supplemented with GSH [70], and adduct 14 was reported
to be present in the urine of HIV-positive patients undergoing NVP
therapy [72]. This identity between our synthetic adducts and those
reported to be formed in vivo or under biologically plausible conditions
supports the validity of our model for mimicking NVP bioactivation and
adduct formation with bionucleophiles. Moreover, given that mercaptu-
rate conjugates are useful biomarkers to probe human exposure to elec-
trophiles [93], we used this methodology to prepare adduct 14 in proper
amounts to be used as standard, enabling a quantitative measure of expo-
sure to electrophiles derived from the 12-OH-NVP pathway in HIV-infected
patients.
Human serum albumin (HSA) is frequently used as a model to investi-
gate protein haptenation by skin allergens, given that approximately 40% of
Bioactivation in Toxic Events Elicited by NNRTIs 13

O
NH2

HO

O
HN
O
S
O HN

O O O GSH NH
S HN N N N
O O
HO
N N N
12

NAC
12-Mesyloxy-NVP (22) H3C

HN O

S O
OH
HN

N N N

14

Scheme 5 In vitro reaction of the model electrophile 12-mesyloxy-NVP (22) with

glutathione (GSH) and N-acetylcysteine (NAC) yielding adducts 12 and 14, respectively
[91]. Adduct 12 was detected in incubations of NVP with human liver microsomes
supplemented with GSH [70] and adduct 14 was detected in the urine of HIV-positive
patients [72].

extravascular HSA is located in the skin [94]. Therefore, the association

of NVP with skin rash reactions prompted us to investigate the capacity
of 12-mesyloxy-NVP to modify this blood protein. Different MS-based
strategies were followed toward this goal, so that complementary information
about modified binding sites could be obtained [95]: (i) enzymatic hydrolysis
of the modified protein to amino acids, followed by liquid chromatography
electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) analysis
(by comparison with fully characterized adduct standards prepared in vitro
from reaction of 12-mesyloxy-NVP with the individual amino acids [91]);
(ii) determination of the intact protein mass by matrix-assisted laser desorption
ionization time-of-flight mass spectrometry (MALDI-TOF-MS); and
(iii) specific digestion of the modified HSA with trypsin, followed by
MALDI-TOF-TOF-MS analysis of the obtained peptides.
14 Sofia A. Pereira et al.

Our model electrophile efficiently modified multiple amino acid resi-

dues (e.g., cysteine, lysine, histidine, tryptophan) in HSA. The consistent
detection of an NVP-tryptophan adduct (23, Scheme 6) by LC-ESI-MS
and MALDI-TOF-MS is noteworthy. Indeed, tryptophan is not a common
site of protein adduction and HSA has only one tryptophan residue, which
suggests a remarkable affinity of the indole ring toward 12-OH-NVP-
derived electrophiles. This was further confirmed by the detection of
tryptophan modification when the same methodology was applied to human
hemoglobin (Hb) [95]. Therefore, the usefulness of monitoring this modi-
fication in vivo is anticipated, given its potential as specific biomarker of
NVP activation and/or toxicity.
DNA adducts are considered suitable biomarkers of exposure to carcino-
gens and are typically detected upon enzymatic or thermal hydrolysis of the
DNA, followed by LC-ESI-MS/MS analysis, through comparison with
fully characterized synthetic standards [92]. Using either a biomimetic strat-
egy (with 12-mesyloxy-NVP as a model electrophile) or a palladium-based
catalysis method, we prepared a library of eight fully characterized adducts
with 20 -deoxynucleosides [deoxyguanosine (dG), deoxyadenosine (dA), and
deoxycytidine (dC)]. These adducts were subsequently used as standards to
investigate adduct formation by LC-ESI-MS upon reaction in vitro of the
same model electrophile with DNA, followed by either enzymatic or ther-
mal hydrolysis. This strategy enabled the identification of a depurinating
guanine adduct, N7-NVP-Gua (25, Scheme 7), in thermal DNA hydro-
lysates, and of the deoxynucleoside adducts N1-NVP-dA (26), N3-NVP-
dC (27), O6-NVP-dG (28), and N6-NVP-dA (29) in enzymatic DNA
hydrolysates. Although the formation of NVP-DNA adducts in vivo remains
to be demonstrated, some of these adducts could have considerable
mutagenic potential should they be present in human patients.

O
O
HO
NH2 NH N N O
O Ph
HN
HN
S

N N N
N N N

24
23

Scheme 6 Structures of the NVP-tryptophan adduct 23 (identified by both LC-ESI-

MS/MS and MALDI-TOF-TOF-MS in human serum albumin and hemoglobin reacted
in vitro with 12-mesyloxy-NVP) and of the N-terminal valine adduct 24 (identified in
human hemoglobin following reaction in vitro with 12-mesyloxy-NVP and subsequent
N-alkyl Edman degradation) [95].