REACTIVE IMINOQUINONE METABOLITES OF

INDOMETHACIN AND CARBAMAZEPINE

Implications for drug-induced idiosyncratic reactions

Changqing Ju

A thesis subrnitted in conforrnity with the requirements

for the Degree of Doctor of Philosophy

Faculty of Pharmacy

University of Toronto

O Copyright by Changqing Su 1999

National Library 1*1 oi Canada Bibliothèque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395, rue Wellington Ottawa ON K I A ON4 Ottawa ON KI A ON4 Canada Canada

Your (ik Votre relerence

Our K& Notre retêrenw

The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sell reproduire, prêter, distribuer ou copies of this thesis in microfom, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/film, de

reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts £iom it Ni la thèse ni des extraits substantiels may be p ~ t e d or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.

REACTIVE IMINOQUINONE METABOLITES OF INDOMETHACIN AND

CARBAMAZEPINE

Implications for drug-induced idiosyncratic reactions

Degree of Doctor of Philosophy, Faculty of Pharmacy, University of

Toronto

Changqing Ju, 1999

ABSTRACT

Indomethacin is a potent nonsteroidal anti-infiammatory drug. Its use has been

Limited because of a relatively high incidence of adverse reactions including

agranulocytosis. Carbarnazepine is an anticonvulsant and its use is associated with a range

of serious idiosyncratic adverse reactions including blood disorders. We postulated that

indomethacin and carbamazepine can be oxidized to reactive intermediates by

myeloperoxidase in activated neutrophils, which may be responsible for the adverse effects

caused by these two dmgs.

We found that desmethyldeschlorobenzoylindomethacin (DMBI), a major

metabolite of indomethacin, was oxidized to a reactive iminoquinone intermediate. This

irninoquinone intermediate could be trapped with glutathione (GSH) and N-acetylcysteine

(NAC) to form conjugates. The sarne oxidation was observed when DMBI was oxidized

by the MPO system or hypochlorous acid. The iminoquinone intermediate with a MH+ ion

at d z 204 was directly detected by mass spectrometry in the reaction between DMBI and

hypoc hlorous acid.

In the urine of patients taking carbamazepine, we detected 2-hydroxyiminostilbene

as its glucuronide conjugate and found that it is a major metabolite of carbarnazepine. We

iii

have also demonstrated that 2-hydroxyiminostilbene can undergo autoxidaiion to a reactive

iminoquinone intermediate. The irninoquinone intermediate reacts with sulfhydryl-

containing nucleophiles, such as GSH and NAC. To further investigate the role of the

iminoquinone intermediate in carbarnazepine-induced adverse reactions, we developed a

polyclonai antibody against the irninoquinone epitope and used it to study covalent binding

of the irninoquinone to proteins. The antibodies were able to detect covaient binding in

bone marrow cells of rats treated with carbarnazepine M vivo. This binding was inhibited

by 2-hydroxyiminostilbene but not by carbamazepine or iminostilbene, which suggested

that the polypeptides recognized by the antibodies were likely to be covalently modified by

the iminoquinone reactive intermediate. Furthemore, we also observed increased protein

oxidation in the liver from carbarnazepine-treated rats compared to that of control rats. This

implied that the iminoquinone, as an oxidizing agent. may undergo redox cycling and lead

to protein oxidation itz VIVO.

Our studies suggest that the iminoquinone intemediates inay react with protein

sulfhydryl groups NI vivo and play an important role in indomethacin- and carbamazepine-

induced idiosyncratic reactions.

ACKNOWLEDGMENTS

Dr. Jack Uetrecht is a great mentor. I am forninate to have had such an

understanding, supportive, patient and knowledgeable supervisor to guide me through my

Ph.D. degree. 1 owe a debt of gratinide to him for al the help he has given me and for an

example of a dedicated scientist he has set for me.

1 also would like to acknowledge and show my appreciation to Dr. Robert A.

McClelland. He was so generous with his thne and expertise whenever I asked him for

advice or had technicd problems in various experimental procedures.

1 am grateful to Dr. Iain Gardner and Mr. Nasir Zahid for sharing their technical

knowledge, without which it would have been impossible for me to complete this thesis.

1 would like to dearly thank Dr. Denis M. Grant and Dr. Peter J. O'Brien for their

roles as cornmittee members of my Ph.D. project and for their constructive criticism which

helped me better my work.

1 would also like to thank Dr. Steven Leeder for his vaiuable advice and

collaboration with the covalent binding studies.

1 am also grateful to Dr. Alastair E. Cnbb and Dr. Allan B. Okey for attending my

Senate Defense and their roles as my extemal appraiser.

1 would like to, finally, thank the following colleagues for giving their support and

friendship and sharing experiences with me through my Ph.D. degree: Ebrahim, Ian.

George, Miyamoto, Terrence, Hak. Kenny, John, Inga and Taras.

TABLE OF CONTENTS

Abstract

Ackno w ledamen ts

Table of content

List of Tables

List of Figures

List of Abbreviations

List of Publications

Chapter 1 General Introduction

1.1 An overview of research interests, rationale and objectives

1.2 Hypothesis

1.3 Dmg-induced idiosyncratic reactions

1.3.1 Clinical manifestations of drug-induced idiosyncratic reactions

1.3.1 a Systemic reactions

1,3. I b Cutaneous reactions

1.3. l c Hepatotoxicity

1.3.1 d Hernatological reactions

1.3.2 Carbarnazepine-induced idiosyncratic reactions

1.3.3 Indomethacin-induced idiosyncratic reactions

1.4 Possible mechanisms of idiosyncratic reactions

1.4.1 Relevance of the immune system to idiosyncratic dnig reactions

1.4.2 Hapten hypothesis

1.4.3 Reactive metabolites involved in idiosyncratic reactions

1.4.3a Hydroxylamines

1.4.3 b Quinones

1 .43 Cations and radicals

Page

ll

i v

v

ix

X

xiii

XV

1

3

4

5

6

6

8

9

10

13

14

15

15

18

3 1

24

29

37

1.4.3d Arene oxides

1.4.3e Miscellmeous

1.5 Metabolism of indomethacin and carbarnazepine

1 S. 1 Indomethacin

1 S.2 Carbarnazepine

Chapter 2 Oxidation of a metabolite of indomethacin (DMBI)

to reactive intermediates by activated neutrophils,

HOCI and the myeloperoxidase systern

2.1 Abstract

2.2 Introduction

2.3 Materials and Methods

2.3.1 Synthesis of DM1

2.3.2 Synthesis of DiMBI

2.3.3 Oxidation of DMBI by HOC1

2.3.4 Oxidation of indomethacin and DM1 by HOC1

2.3.5 Oxidation of D.MBI by the MPO system

2.3.6 Oxidation of D.WI, indomethacin and DM1 by activated neutrophils

2.3.7 Synthesis of reacrive intermediate-NAC adduct

2.3.8 Reaction kinetics determined by spectrophotornetry

2.3.9 Analytical Methods

2.4 Results

2.4.1 Oxidation of D . B I by activated neutrophils

2.4.2 Oxidation of DLMBI by HOCI

2.4.3 Oxidation of DMBI by MPO system

2.4.4 Oxidation of indomethach by HOC1 and activated neutrophils

2.4.5 Oxidation of D-W by HOC1 and activated neutrophils

2.5 Discussion

vi i

Chapter 3 Detection of 2-hydroxyiminostilbene in the urine

of patients taking car barnazepine and its oxidation

to a reactive iminoquinone intermediate

3.1 Abstract

3.2 Introduction

3.3 Materids and Methods

3.3.1 Synthesis of the irninoquinone intermediate

3.3.2 Synthesis of 2-hydroxyiminostilbene

3.3.3 Synthesis of NAC-adduct of the iminoquinone

3.3.4 Synthesis of 2-hyroxycarbamazepine

3.3.5 Treatment of the urine sample

3.3.6 Analytical methods

3.4 Results

3.4.1 Oxidation of 2-hydroxy irninostilbene

3 A.2 Reactivity of the irninoquinone intemediate

3.4.3 Patient urine sample analysis

3.5 Discussion

Chapter 4 lnvestigating the role of the reactive iminoquinone

metabolite of carbarnazepine in the mechanism of

carbarnazepine-induced idiosyncratic reactions

4.1 Abstract

4.2 Introduction

4.3 iMaterials and Methods

4.3.1 Synthesis of irninoquinone-NAC-conjugaied KLH and RSA

1.3.2 Production of anti-iminoquinone-NAC-KLH antiserum

S . .

V l l l

1.3.3 Activity of the anti-iminoquinone-NAC-KLH antiserum

4 3 . 4 Dosing of rats with carbarnazepine in vivo

4.3.5 Isolation of rat bone marrow cells

4.3.6 Preparation of subcellular fractions of rat liver

4.3.7 Preparation of rat skin homogenates

4 3 . 8 SDS-PAGE and imrnunoblotting

4.3.9 Determination of oxidatively modified proteins

4.4 Results

1.4.1 S ynthesis of irninoquinone-NAC-KLH (RS A) conjugate

44.2 Determination of the activity of the antiserum

4.4.3 hunoblot t ing of bone marrow cells isolated from rats

treated with carbarnazepine

4.4.4 Protein oxidation in subcellular fractions of rat liver

-1.45 hunoblot t ing of skin homogenate and subcellular

liver fractions from rats treated with carbarnazepine

4.5 Discussion

Chapter 5 Summary and Discussion

5.1 Surnmary

5.2 Discussion

LIST OF TABLES

Chapter 1 General Introduction

Table 1 - 1 Fonation of the reactive hydroxylarnine intermediates during

bioactivation of drugs by various oxidizing enzymes.

Table 1-2 Formation of the reactive iminoquinone intermediates during

bioactivation of drugs by various oxidizing enzymes.

Table 1-3 Formation of cation or free radicals during bioactivation

of drugs by various oxidizing enzymes.

Table 13 Formation of the reactive arene oxide intermediates during

bioactivation of xenobiotics by CYP450.

Table 1-5 Formation of the reactive metabolites during bioactivation of

drugs by various oxidizing enzymes.

Chapter 3 Detection of 2-hydroxyiminostilbene in the urine

of patients taking carbarnazepine and its oxidation

to a reactive iminoquinone intermediate

Table 3- 1 The ratios of the TIC of 2-hydroxyiminostilbene-glucuronide

( d z 386) to that of carbarnazepine-glucuronide (m/z 4 13),

monohydroxycarbamazepine-glucuronide (m/z 429) and

10.1 1 -dihydrodiol-carbarnazepine (m/z 27 1).

Page

19

LIST OF FIGURES

Page

Chapter 1 General Introduction

Figure 1 - 1 The major and minor antigenic determinants of benzylpenicillin. 20

Figure 1-2 The role of bioactivation in rnediating either direct (metabolic) 23

idiosyncrasy or immune-mediated idiosyncrasy.

Figure 1-3 Metabolism of procainamide by MPO in activated neutrophils. 27

Figure 1-4 Oxidation of arnodiaquine to a reactive irninoquinone interrnediate. 3 1

Figure 1-5 Proposed pathway of the oxidation of diclofenac to a reactive 33

irninoquinone.

Figure 1-6 Proposed parhway of the oxidation of vesnarinone to reactive 34

irninoquinone intermediates.

Figure 1 -7 Proposed mechanism for 1 ,Cbenzoquinone formation from the 36

MPO-catalyzed metabolism of CGS 12094.

Figure 1-8 Metabolic pathway for the bioactivation of tacrine to its posnilated 37

quinone methide reactive metabolite by CYPl A?.

Figure 1-9 Proposed scheme of the oxidation of arninopyrine to a cation 39

radical via a dication interrnediate.

Figure 1-10 Proposed pathway for the oxidation of clozapine to a reactive 3 1

nitreniurn ion or a cation radical by HRP and MPO.

Figure 1 - 1 1 Proposed mechanisms of forming the imidoirninium ion 43

intermediate of vesnarinone.

Figure 1 - 12 B ioactivation of benzo[a]pyrene to the diol-epoxides carcinogens. 17

Figure 1-13 Metabolism of tienilic acid to the sulfoxide metabolite by CYP2C9. 50

Figure 1-14 Metabolism of propylthiouracil by MPO to the reactive sulfonic 51

acid.

Chapter 2 Oxidation of a metabolite of indomethacin (DMBI)

Figure 2- 1

Figure 2-2

Figure 2-3

Figure 2-4

Figure 2-5

Figure 2-6

Figure 2-7

Figure 2-8

Figure 7-9

to reactive intermediates by activated neutrophils,

HOCl and the myeloperoxidase system

Major metabolic pathway of indomethacin.

Metabolisrn of DMBI by activated neutrophils as a fùnction

of t h e , initial DMBI concentration and neutrophil concentration.

MSMS spectra of the metabolites with MH+ ions at m/z 409 and 67

rn/z 407, formed dunng DMBI oxidation by activated neutrophils.

W absorption spectnim of the metabolite with a MH+ ion at 107, 65

during DM31 oxidation by activated neutrophils.

Proposed pathway of the formation of the two stable metabolites

dunng DMBI oxidation by activated neutrophils.

Repetitive absorption spectra from the reaction of DMBI with

HOCl.

Proton NMR spectmm of the iminoquinone-NAC conjugate.

Proposed scheme for DiMBI metabolism by activated neutrophils

and trapping the reactive intermediate by NAC.

Cosy specrrum of the iminoquinone-NAC conjugate.

Chapter 3 Detection of 2-hydroxyiminostilbene in the urine

of patients taking carbarnazepine and its oxidation

to a reactive iminoquinone intermediate

Figure 3- 1 Proposed bioactivation pathway of carbarnazepine which leads

to the formation of a reactive iminoquinone intermediate.

Figure 3-2 The W absorption spectra of the synthesized iminoquinone and

2-hydroxyiminostilbene before and after reaction with HOC1.

Figure 3-3 Proton NMR of the iminoquinone-NAC conjugate

xii

Figure 3-4 Oxidation of 2-hydroxyiminostilbene to the iminoquinone 93

intermediate and its reaction with NAC.

Figure 3-5 Cosy spectnirn of the iminoquinone-NAC conjugate 95

Figure 3-6 MSMS spectrum of the metabolite, which has the same molecular 98

weight as that of 4-methylthio-7-hydroxyirninostilbene, in its

glucuronide form (with a MH+ ion at m/z 432).

Figure 3-7 Selected ion current of the urinary metabolites from the LCfMS of a 99

patient's urine.

Fipre 3-9 MS/MS spectrum of the glucuronide conjugate of 1 O0

2-hydroxyiminostilbene.

Figure 3-9 LC/MS ion current at m/z 208 corresponding to the iminoquinone 10 1

intermediate in a urine sample before and after hydrolysis by

P-glucuronidase.

Chapter 4 Investigating the role of the reactive iminoquinone

metabolite of carbarnazepine in the mechanism of

carbarnazepine-induced idiosyncratic reactions

Figure 4- 1 The structure of the irninoquinone-NAC-Protein conjugate. 117

Figure 4-2 ELISA analysis showing binding of the anti-irninoquinone-NAC-KLH 120

antiserum to wells of microtiter plates coated with an

iminoquinone-NAC-RSA conjugate or RSA.

Figure 4-3 Immunochernical detection of protein binding in the bone marrow 12 1

from rats treated in vivo.

Fi jure 4-4 Immunoblotting expenment show ing protein oxidation in 122

subcellular fractions of rat liver in vivo.

Chapter 5 Summary and Discussion

Figure 5-1 Proposed pathway of carbarnazepine bioactivation to an arene

oxide intermediate and its reactions with GSH.

xiii

LIST OF ABBREVIATIONS

Abbreviation

ANA

BP

CGS 12094

CYP450

DBI

DL

DMBI

DMI

EDC

ELISA

FDA

GSH

HLA

HRP

HSR

k KLH

LCMS

MH+

MHC

MPO

NAC

NADPH

NSAID

Narne

an tinuclear antibody

benzo[a] pyrene

para-hydroxy metabolite of p ~ o m i d e

cytochrome P45O

deschlorobenzoylindomeihacin

drug-induced lupus

desrnethyldeschlorobenzoylindomethacin

desmethy lindomethacin

1 -ethyl-3-(3-dimethyi~nopropy1)carbodiimide

enzyme-linked immunosorbent assay

Food and Dmg Administration of U.S.

glu tathione

human leukocyte antigen

horseradish peroxidase

hypersensitivity syndrome reactions

imrnunoglobulin

Keyhole limpet hemocyanin

liquid chromatography/Mass spectrometry

protonated molecular ion

major histocompatibility genes

myeloperoxidase

n-acetylcysteine

nicotinamide adenine dinucleotide phosphate, reduced form

nonsteroidal anu-inflammatory dmg

xiv

PHs

RSA

SJS

SLE

SSLR

TCPO

TEN

TIC

prostaglandin H synthase

rabbit semm aibumin

Stevens-Johnson syndrome

systernic lupus erythematous

semm sickness-like reactions

trichloropropene oxide

toxic epidermal necrolysis

total ion current

LIST OF PUBLICATIONS

lu C and Uetrecht JP (1998) Detection of 2-hydroxyiminostilbene in the urine of patients

taking carbarnazepine and its oxidation to a reactive iminoquinone intermediate. J.

Pharmacol. Exp. Ther. 288: 5 1-56.

Ju C and Uetrecht JP (1998) Oxidation of a metabolite of indomethacin

(desmethyldeschlorobenzoylindomethacin) to reactive intermediates by activated

neutrophils, hypochlorous acid, and the myeloperoxidase system. Dnlg Merab. Dispos.

26: 676-680.

Chapter 1

General Introduction

1.1 An overview of research objectives and rationale

Adverse drug reactions represent a significant health problem. In a recent study by

Bruce Pomeranz and Jason Lazarou (Lazarou et al., 1998). they estirnated that adverse

dmg reactions h l1 an average of 100,000 people annually in the US.. making them the

founh leading cause of death. after heart disease. cancer and stroke (The Financial Post

Magazine, June 1998 Page 9). Some of the adverse dmg reactions are preventable, in

principle. because they are due to known pharmacological effects of the dmgs and would

be expected to occur in most patients at a sufficiently high dose. Another type of adverse

drug reaction is the so-called "idiosyncratic reaction" for which the mechanism is mostly

unknown. Idiosyncratic dnig reactions, which account for approximately 10% of al1

adverse drug reactions (Goldstein and Patterson. 1984), are often quite severe and

sometimes life threatening. Due to their unpredictable nature, drug-induced idiosyncratic

reactions have posed a difficult problern, both in medical practice and for present and Future

dmg development. Although the rnechanisms of most dmg-induced idiosyncratic reactions

are unknown. a large number of studies have suggested that reactive metabolites are

involwd. Several clinical characteristics of this type of adverse reaction ülso suggest the

involvernent of the immune system.

Indomethacin and carbarnazepine have been s h o w to cause idiosyncratic reactions

including drug-induced agranulocytosis. The formation of reactive metabolites is thought

to be a necessary step in the pathogenesis of these idiosyncratic reactions. Although liver is

the major organ responsible for drug metabolism. short-lived reactive metabolites generated

in the liver rnay not be able to escape the liver and induce clinical manifestations in other

areas of the body which are involved in idiosyncratic drug reactions. Therefore. it is

important to investigrite the formation of reactive metabolites in the txget organs of dmg-

induced adverse reactions. In indomethacin- and carbarnazepine-induced hematological

disorders reactive metabolites are more likely to be generated by the oxidative

myeloperoxidase system in neutrophils or other cells in the bone marrow than in the liver.

The objectives of this project were: a) to determine if indomethacin and

carbarnazepine c m be oxidized by myeloperoxidase in activated neutrophils; b) to identib

the reactive intermediates generated during the oxidation and characterize their chernical

reactivity as electrophiles; c) and to further investigate the possible role of the reactive

intermediates in the mechanism of idiosyncratic dmg reactions.

1.2 Hypothesis

Reactive iminoquinone metabolites are generated during the oxidation of

indomethacin and carbarnazepine in the targets of dmg-induced toxicity. These reactive

iminoquinone intermediates play an important role in the mechanism of indomethacin- and

carbarnazepine-inducrd idiosyncratic reactions via covalent binding to endogenous proteins

that induce an immune reaction.

1.3 Drug-induced idiosyncratic reactions

An adverse dmg reaction is defined as any undesirable response to a drug that

occurs at therapeutic doses (Park and Kitteringham, 1990). Adverse dru: reactions are

major complications of modem drug therapy; they account for significant patient rnorbidity

and death (Park et al. , 1992). Drug-induced adverse reactions can be classified into

predictable reactions and unpredictable reactions (Rieder, 1994). Predictable reactions +

include toxicity from overdose; dmg-dmg or dmg-disease interactions; and side effects that

occur at usual therapeutic doses such as nausea and dizziness. These reactions can often be

predicted from the chernical properties and the pharmacology of the drug; they ore often

host-independent, dose-dependent; and usually reproducible in animals (Pohl et al., 1988).

Such reactions should be anticipated and can often be eliminated by dose reduction.

Unpredictable reactions include hypersensitivity (or "allergie") reactions and idiosyncratic

reactions. The terms "dmg hypersensitivity reaction" and "idiosyncratic drug reaction"

have been used interchangeabl y, for example, "drug-induced hypersensitiviry reactions"

has been used to describe systemic idiosyncratic reactions with fever. rash and multiorpan

involvement. Ideally, "hypersensitivity reaction" should be used to describe adverse

reacrions associated with drug therapy which have a known immunological etioiogy;

however, The term "idiosyncratic dmg reactions" will be used, in this thesis. to descnbe

reactions with cornmon characteristics that include:

Occurrence in a relatively small number of patients. These reactions have been

rstimated to account for 6% to 10% of ail adverse drug reactions caused by a total of 90

million courses of dru; therapy per year among medical inpatients (Goldstein and

Patterson, 1984)

An apparent lack of dose-response curve for the risk of toxicity. In principle these

reactions must be dose-dependent; however, in many cases the dose that can cause

roxicity is much srnaller than the dose for desired pharmacological effect.

;ui unpredictable nature and lack of animal models. Idiosyncratic reactions occur in

humans with a relatively low incidence; therefore it is diffïcult to reproduce the s m e

reactions in laboratory anirnals, which are inbred strains, even if a large number of

animais are tested.

Involvement of specific organs, such as the skin, bone marrow and liver. These

reactions are often lirnited to specific organs, although generalized reactions may also

occur in which there is fever, skin rash and lymphadenopathy, dong with involvement

of specific organs.

1.3.1 Clinical manifestations of drug-induced idiosyncratic reactions

The patterns of idiosyncratic reactions are variable depending on both the patients

and the drugs. These reactions, although not very common, can be life-threatening and

have resulted in removal from the market of many potentially useful dmgs. For example,

practolol, benoxaprofen, ticrynafen, zomepirac and nomifensine were withdrawn shortly

after they were released because of an unacceptable risk of toxicity (Uetrecht, 1990). The

most comrnonly occumng idiosyncratic dmg reacrions involve the skin, various elements

of the blood and the liver. The kidney and nervous system may also be affected and

autoimmune syndromes such as drug-induced lupus can also occur.

1.3. l a Systemic reactions

In some cases idiosyncratic reactions show a rernarkable degree of organ

specificity. In other cases, multiple ogans may be affected simultaneously, such as in

hypersensitivity syndrome reaction (HSR), semm sickness-like reaction (SSLR) and drug-

induced lupus (DL).

CI

/

HSRs are serious and possibly life-threatening reactions that usually occur 2 to 4

weeks af'ter the stan of therapy. Although this type of reactions is termed "hypersensitivity

syndrome", their mechanisms are not fully understood. Drugs that are most often

associated with this type of adverse reaction include anticonvulsants, dapsone,

sulfonamides and allopurinol (Knowles et al., 1996). The initial sign of HSRs is often

fever, which is followed by a rash. The rash may be exanthematous; however, in severe

cases it c m be erythema multiforme, Stevens-Johnson syndrome (SJS), or toxic epidermal

necrolysis (TEN). Intemal organ involvement often develops 1 to 2 weeks later. Hepatic

injury is a commoo manifestation, although pulmonary, hematological or renal impairment

may also occur.

SSLRs are characterized clinically by fever, rash, anhralgia and occasionally with

associated lymphadenopathy (Parker, 1982). Clinical signs and symptoms of SSLRs

typically develop 1 to 3 weeks after the start of drug treatment. Unlike cllissical serum

sickness reactions, immune complexes have not been identified in SSLRs although they

share many clinical symptoms (Knowles er al., 1996). SSLRs have been shown to occur

with a nurnber of drugs including penicillin, cephalosporin, sulfonamides,

propylthiouracil, phenytoin, streptomycin, barbiturates, isoniazid, salicylates, captopril.

hydralazine and nonsteroidal antiinflammatory dmgs (NSAIDs) (Füeder, 1994).

Similar to idiopathic systernic lupus erythematous (SLE), D L is an autoimmune

disease characterized by antibodies that bind to "selft-antigens. However, in the case of

DL the symptorns are often milder (Anderson, 1992). In particular, the most serious

manifestations in SLE. involving the central nervous system and kidney, are uncornmon in

DIL (Anderson, 1992). Symptoms of D L are self-limiting once the offending drug has

been discontinued. Because it usually requires several months of therapy before lupus is

induced, a dnig would cause lupus only if it is used for a chronic illness. The drugs

comrnonly associated with DL are procainamide, isoniazid, hydralazine, chlorpromazine,

phenytoin, methyldopa, penicillamine, quinidine and minocycline (Gilliland, 1991;

8 Knowles et al., 1996). Of the dmgs associated with lupus, procainamide induces the

highest incidence (10 to 20% of patients on continuous procainamide therapy over 2 years)

(Rieder, 1994). The mechanism of D U is unknown. it has been shown to be associated

with antinuclear antibodies (ANA); however. only a minority of patients who develop a

positive ANA titer related to drug ingestion develop clinicai disease consistent with DIL

(Knowles et al., 1996).

1.3.lb Cutaneous Reactions

In cornparison with other organs, the skin is quite frequently a target of drug-

induced idiosyncratic reactions. The combination of metabolic activity and imrnunological

responsiveness may explain why skin is the organ most frequently affected by adverse

dmg reactions. In contrast with allergic contact dermatitis, dmg-induced skin reactions

have a large variability with regard io their pathophysiological pathways, clinical signs and

symptoms, severity and the drugs that can elicit these reactions (Merk and Henl, 1996).

Alrnost any type of skin disease or characteristic symptoms can be mimicked by drug-

induced skin reactions. Probably the most comrnon cutaneous manifestations of drug

reactions are a variety of maculopapular, morbilliform. scarlatiniform or exanthematous

emptions. Some of the most cornmon offenders being various antibiotics, anticonvulsants,

allopurinol, barbiturates, carbarnazepine, sulfonamides, gold salts, and antituberculosis

agents (Mathews, 1984). The more serious skin reactions due to drugs include erythema

multiforme, SJS and TEN. The incidence of SJS, TEN and SJSiTEN overlap has been

estirnated to be approximately 1.89 cases per one million people per year (Mockenhaupt

and Schopf, 1996). Although SJS and TEN occur very rareiy, a mortaiity rate of more

than 40% c m be calculated for patients suffering from TEN (Mockenhaupt and Schopf,

1996). The highest incidence has been implicated to be associated with three groups of

9

drugr, antibacterials (e.g. sulfonamides, penicillin), NSAIDs and anticonvulsants (Park et

al., 1992).

These severe type of skin reactions are thought to be immune-mediated. The

evidence for this include: the occurrence of TEN in acute graft-versus-host disease in which

the immune system is definitely involved. demonstration of anti-epidermal antibodies and

positive lymphocyte transformation tests and demonstration of epidermal infiltration by

CD8+ T-cells (Park et al., 1992).

1.3.1~ Hepatotoxicity

Adverse drug reactions affecting the liver are common with more than 600 drugs

having been reponed to cause hepatic injury (Stricker and Spoelsua, 1985). The types of

liver injury are variable and include hepatocellular necrosis, steatosis. cholestasis,

granuloma, chronic hepatitis and cirrhosis (Pohl, 1990). The fatality rate of drus-induced

hepatic injury is often hiph such that drugs are a major cause of hepatic failure

(Zirnmerman, 1978). For example, the case fatality rate from halothane-induced hepritic

failure may be as high as 50% (TimbreIl. 1983).

Drug-induced hepatotoxicities have been classified into two catesories: the

predictable acute response after an overdose and the idiosyncratic reaction following

chronic therapeutic doses. Predictable injury resulting in hepatocellular necrosis is the best

understood variety. It is dose-related. host-independent and usually reproducible in animal

models. It has been shown to be due ro the intrinsic toxicity of a drug or its metabolite

produced in the b e r . Hepatic necrosis induced by acetarninophen and carbon tetrachloride

are typical examples of this type (Kaplowitz e t al., 1986). Clinically, idiosyncratic hepatic

injury is seen more commonly than predictable toxicity (Kaplowitz et al., 1986). With

idiosyncratic hepatotoxicity, there seems to be no dose relationship, usually no animal

mode1 and it is often host dependent (Pohl, 1990). Dmgs that are associated with this type

10 of reaction include rifampin, halothane, chlorpromazine, isoniazid, hydantoins and

sulfonamides (Mathews, 1984; Timbrell, 1983).

Idiosyncntic hepatotoxicity may be due to direct toxicity of a chemically reactive

metabolite, such as isoniazid-induced hepatotoxicity. This is called "metabolic"

idiosyncrasy (Zimmerman, 1978). Another type of idiosyncratic hepatotoxicity is

secondary to an immune reaction (Pohl, 1990). A diagnosis of immunological liver injury

is supported by the association with multiple exposures, the pattern of fever, occasional

eosinophilia and skin rash and the reoccurrence of the same pathologic condition upon

rechallenging with the culprit dmg (Pohl, 1990).

1.3.ld Hematologicai Reactions

Blood cells are a cornmon target in idiosyncratic dmg reactions. Dmg-induced

hematological toxicity can affect either platelets, red cells, white cells or al1 fomed elements

of the bone marrow resulting in thrombocytopenia, hemolytic anemia, agranulocytosis or

aplastic anernia, respectively.

Agranulocytosis is characterized by an almost complete lack of granulocytic

leukocytes in the circulation - less than 500 cells compared to a normal count of 5.000-

10,000 cells per pl of blood. The granulocytic leukocytes are neutrophils, eosinophiis, iuid

basophils; the most abundant ce11 being the neutrophils. Agranulocytosis carries a high

monality rate because a patient without neutrophils has a very high nsk of infection that

cannot be controlled by antibiotics (Uetrecht. 1992). Drug-induced agranulocytosis is

usuaily reversible with an "overshoot" of the peripheral neutrophils approximately 1 week

after the dmg is stopped. Depletion of al1 elements of the bone marrow results in aplastic

anemia. The monality rate of aplastic anernia is much higher than that of agranulocytosis

and the time to recovery is usually longer. The estimates of eiiologic fractions of

agranulocytosis and aplasric anernia due to drus use have been reported in three case-

control studies: the International Agranulocytosis and Aplastic Anernia S tudy (IAAAS)

conducted in Israel and Europe, a study conducted in the northeast U.S. and a study

conducted in Thailand (Kaufman et al.. 1996). The overall etiologic fractions of

agranulocytosis due to drug use were 62% in IAAAS. 77% in the U.S. and 70% in

Thailand. Compare to agranulocytosis. the etiologic fractions of aplastic anemia due to

drug use were much less and they were 27% in the IAAAS, 17% in the U.S. and 2% in

Thailand. Drugs that have been associated with agranulocytosis include: acetaminophen,

aminopyrine, arnodiaquine, benzene, captopril, carbamazepine, chloramphenicol,

chlorpromazine, chlozapine, dapsone, indomethacin. methimazole, mianserin,

penicillamine, phenylbutazone, procainamide, propylthiouracil, sulfonamides,

trimethoprim, trimethoprim/sulfamethoxazole and vesnarinone (Kaufman et al., 199 1 ;

Uetrecht, 1992). Drugs that are often associated with aplastic anemia include allopurinol,

cûptopril, carbamazepine, chloramphenicol, dapsone, indomethacin, mephenytoin,

methimazole, phenylbutazone, phenytoin, propylthiouracil. sulfonamides, trimethadione

and trimethopnm/sulfamethoxazole (Kaufman et al., 199 1 ).

The mechanism of dmg-induced aganulocytosis is unknown; however, either an

imrnunologically-mediated destruction of neutrophils or a direct cytotoxici ty to bone

marrow has been sug_oested (Isildar et al., 1988). The clearest example of a drug that

causes antibody-mediated peripheral destruction of neutrophils is aminopyrine (Magis et

al., 1968: Moeschlin and Wagner, 1952; Uetrecht. 1990). The evidence for an immune-

mediated reaction includes: the discovery of a serum factor in patients with aminopyrine-

induced agranulocytosis that agglutinates neutrophils (Magis et al., 1968; Moeschlin and

Wagner, 1952), a rapid decline in neutrophils when blood from the patient was infused into

normal subjects with the same blood type and the rapid onset of agranulocytosis on

reexposure of a patient to even a small quantity of arninopyrine. In contrast,

chlorpromazine-induced agranulocytosis is considered an exarnple of a direct bone marrow

toxicity (Pisciotta et al., 1958). One piece of evident to suppon this is that the onset is

delayed upon reexposure. In addition, bone marrow examination demonstrated depletion

of al1 elements of the granulocyte senes (Uetrecht, 1992). The clinical characteristics of

drug-induced agranulocytosis associated with other drugs appear to fa11 somewhere

between those of arninopyrine and chlorpromazine, with some characteristics suggesting

more direct toxicity and others suggesting an immune-mediated reaction.

Thrornbocytopenia is characterized by bleeding and petechial hemorrhages in the

skin that are related directly to the fa11 in circulating platelet levels. The most commonly

proved or strongly suspected causes of immunologically-mediated thrombocytopenia are

chlorothiazide, digitoxin. gold, heparin, meprobarnate, quinidine, quinine, rifarnpin,

stibophen and sulfonamides (Petz and Fudenberg, 1976). A marked decrease in

rnegakaryocytes in the early phases of a reaction is usually considered to be indicative of a

toxic rather than imrnunologic reaction. However, decreased platelet synthesis and release

may also occur to a limited extent in immunologically mediated thrombocytopenia,

particularly when the reaction is prolonged. The immunological nature of the reaction has

been demonstrated both by passive transfer and irz vitro studies in thrombocytopenia

produced by apronalide, quinine, and quinidine (Ackroyd, 1975; Ackroyd, 1962;

Shulman, 1958). The sera and immunoglobulin (Ig) fraction of affected patients contain

antibodies that. in the presence of the drug, produce platelet lysis or agglutination.

A variety of dnigs, pmicularly penicillin and a-methyldopa. have been reponed to

produce an immune-mediated hemolytic anemia. Other agents include dipyrone, p-

aminosalicylic acid, stibophen, quinine, quinidine, phenacetin and nomifensine (Parker,

1982). a-methyldopa produces its effect on red blood cells by inducinj an autoantibody.

Penicillin-induced hemolytic anemia js an exarnple of an irnrnunodestnictive response in

which the dmg participates as a hapten. It is an occasional complication in individuals

receiving prolonged high dose penicillin therapy (Parker, 1982). In about 3% of

individuals enough antihapten antibody is produced to give a positive antiglobulin reaction

13 on the red cells. If enough of the right kind of antihapten antibody is present and the bone

marrow cannot compensate, anernia develops.

1.3.2 Carbarnazepine-induced idiosyncratic reactions

Carbarnazepine has become for many clinicians a dnig of choice for the treatment of

borh partial and generalized convulsive seizures. It has challenged the use of phenytoin

because it is considered by many to cause less cognitive impairment then phenytoin (Dodrill

and Troupin, 1977; Smith et al., 1987; Thompson and Trimple, 1982), although the two

dmgs share similar profiles of adverse effects. The incidence of side effects associated

with carbarnazepine treatment in adults and children ranges from 33 to 50%. Most of the

effects are rnild, transient and reversible if the dosage is reduced or if initiation of treatrnent

is gradual; however, 5% of the adverse reactions associated with carbamazepine can be

classified as idiosyncratic reactions (Durelli et al., 1989). These idiosyncratic adverse

reactions include skin rash (Crill, 1973), blood disorders (Gerson et al., 1983). hepatitis

(Horowitz et al., 1988) and the carbamazepine hypersensitivity syndrome (Carmela et al.,

1995; Shear and Spielberg, 1988). A Swedish survey of 505 reports of 7 13 idiosyncratic

reactions to carbamazepine from 1965 to 1987 reported skin reactions (48%).

hematological (12%) and hepatic disorders (10%) to b e the most frequent (Askmark and

Wholm, 1990).

S kin reactions were the most frequent adverse reaction associated with

carbarnazepine treatment that has been reported to Geigy Phmaceuticals and the Food and

Drug Administration (FDA) in the United States. For the years 1975 to 1986 there were a

total of 396 cases of rash and nonspecific derrnatological reactions reported. These

included 13 reported cases of SJS. 2 cases of TEN, 10 cases of exfoliative dermatitis and

12 cases of eryrhema multiforme (Pellock, 1987). In addition, carbamazepine is one of the

drugs most frequently implicated ro cause TEN (Park et ai., 1992).

In addition to derrnatological reactions, hematological adverse reactions to

carbamazepine, although rare, are important because of the serious and potentially fatal

consequences of protracted bone marrow depression. Of 80 significant cases of

carbarnazepine-associated hematological reactions reponed to Geigy Pharmaceuticals over

12 year, thrornbocytopenia ranked first with 3 1 cases. followed by aplastic anernia (27

cases), agranulocytosis (10 cases), pancytopenia (8 cases) and bone marrow depression (4

cases) (Pellock, 1987).

Most carbarnazepine-associated hepatic abnormalities are transient elevations of

liver enzymes (Pellock. 1987). However, carbarnazepine-induced hepatic injury has been

reponed to show a cholestatic or mixed pattern as well as a hepatocellular pattern. Cases of

granuiomatous hepatitis and cholangitis have also been descnbed (Askmark and Wiholm,

1 990).

Carbarnazepine hypersensitivity syndrome. similar to the hypersensitivity syndrome

associated with phenytoin and phenobarbital, is a potentially fatal reaction. In most cases,

the reaction invoives multiorgan abnormalities accompanied by fever, rash and

lymphadenopathy. The onset of the reaction usually occurs within 3 months of the initial

treatment. most ofren within 2 to 4 weeks. It occurs sooner in previously sensitized

individuals and occurs within 1 day on rechallenge in those with a history of carbamazepine

hypersensitivity syndrome (Carmela et ni., 1995).

1.3.3. Indomethacin-induced idiosyncratic reactions

indomethacin is a very effective NSAID, but its use is lirnited by a high incidence

of adverse reactions, including blood disorders. Adverse effecrs have been estimated to

occur in 30-604 of patients treated with indomethacin, and serious reactions requiring

discontinuance of the drug occur in about 10% of patients. Most comrnon side effects,

15 which appear to be dose-dependent, include gastro-intestinal disturbances, headache and

dizziness (Cuthbert, 1974).

Indomethacin also causes idiosyncratic hematological adverse effects in less than

1% of patients. These reactions include hemolytic anemia, bone marrow depression,

aplastic anemia, agranulocytosis, leukopenia, thrombocytopenia and thrombocytopenic

purpura. Of a total of 1,261 reports of adverse reactions to indomethacin reported to the

U.K. Cornmittee on Safety of Medicines between Iune 1964 and January 1973, blood

disorders were recorded in 157 cases (25 fatal) which included thrombocytopenia (35; 5

fatal), aplastic anemia ( 17; no fatalities), and agranulocytosis or leukopenia (2 1 ; 3 fatal)

(Cuthbert, 1 974). Subsequently . the First Report from the International Agranulocytosis

and Aplastic Anemia Study confirmed a signifiant relationship between the use of

indomethacin and agranulocytosis and aplastic anemia. In this population-based case-

control study conducted in Europe and Israel, the excess risk estimated for agranulocytosis

and aplastic anemia was 0.6 per million for indomethacin exposure in a one-week period

(Shapiro, 1986).

1.4 Possible mechanisms of idiosyncratic reactions

1.4.1 Relevance of the immune system to idiosyncratic drug reactions

Adverse drug reactions that involve a specific drug-induced immune response are

referred to as drug hypersensitivity reactions and they can be classified according to the

general scheme of Gel1 and Coombs (1963):

Type 1 (imrnediate hypersensitivity or anaphylacllc) reactions

This type of reactions result from the interaction (cross-linking) of drug antigen with

specific IgE antibody on the surface of mast cells and basophils to trigger the release of

chernical mediators such as histamine and leukotrienes. The chnical features include

generalized erythema. urticaria, angioedema, gastrointestinal disturbances,

bronchospasm, laryngoedema, and hypotension. To most clinicians, the term "allergy "

is synonymous with type 1 hypersensitivity. Anaphylactic reactions to penicillin are an

example of a type I reaction.

Type II (cytotoxic) reactions

This type of reactions are caused by binding of specific IgG or IgM antibody with

antigen associated with the surface of a cell, or to a ce11 surface altered antigenically by

a drug. The ce11 is destroyed by phagocytic cells or by activation of the complement

system leading to ce11 lysis. This may result in thrombocytopenia, anemia or

leucopenia. depending on the disposition of the antigen. Hemolytic anemia due to

quinine and penicillin are type II reactions.

Type IQ (immune compiex) reactions

This type of reaction involves the formation of circularing immune complexes, from

specific antigen and IgG or IgM antibodies, which c m precipitate in the vasculature and

rend glomeruli; tissue damage results from activation of complement and recruitment of

phagocytes. The classical example of type III hypersensitivity is serum sickness

reactions, w hich rnay presrnt clinically with lymphadenopathy. arthralgia. fever and

rash.

4 Type IV (hg-induced delayed hypersensitivity) reactions

This type of reaction usuaily occurs in the skin. The major ce11 types involved are T

lymphocytes and macrophages. Lymphokines are released following interaction of antigen

with specific T-cells; tissue darnage is caused by infiltrating mononuclear cells. Topically

applied dmg becomes nttached to a carrier protein in the skin; induces a cell-mediated

response and causes contact dermatitis. Such reactions can also occur nfter either oral or

parenteral administration of the drug.

Idiosyncratic dmg reactions have often been referred as drug hypersensitivity

reactions. However, there is not enough evidence to definitely prove that the immune

system is involved. The most convincing evidence for the involvement of the immune

system is demonstration that specific antiboaies or sensitized T lymphocytes are reacting

wirh the dmg or its metabolites. Dmg-related antibodies have been described in some

idiosy ncratic dmg reactions. For example, the evidence for an immune-mediated reaction

is strong in penicillin-induced anaphylaxis and aminopyrine-induced agranulocytosis.

Penicillin is chernically reactive due to the b-lactam ring and it can act as a hapten. There is

strong evidence that most of the antibodies in patients who are allergic to penicillin

recognize penicillin bound to lysine (Parker, 1982). In some patients the major type of

antibody induced is IgE, and this c m result in an anaphylactic reaction. Antibody-mediated

peripheral destruction of neutrophils has been suggested to be the mechanism of

aminopyrine-induced agranulocytosis by several pieces of evidence. For example,

incubation of semm from a patient who has the active disease with normal neutrophils Ied

to agglutination of the neutrophils (Moeschlin and Wagner, 1952). In addition, infusion of

blood frorn a sensitive patient 3 hours after a dose of aminopyrine into a normal subject

rnatched for blood type led to granulocytopenia (Moeschlin and Wagner, 1952). In another

study, the serum from a patient with aminopyrine-induced agranulocytosis inhibited

colonies cultured from a rnononuclear fraction of the patient's blood in the presence of

d r y , while serum from a normal control did not (Barrett et al., 1976). However. probably

due to the complexities of the immune systern. in most cases of idiosyncratic reacrions rhere

is no direct evidence for the involvement of the immune system, even when effort has been

made to search for such evidence. Furthemore, since not al1 drug-induced antibodies or

sensitized T lymphocytes will necessarily cause tissue damage, a definite proof that ri

toxicity has an allergic basis requires that it be produced in animals. where the

imrnunologica1 mechanisms of cellular damage can be thoroughly studied (Pohl et al.,

1988). However, there have been few animai models developed for studying the

mechanism of idiosyncratic dnig reactions.

1s Despite this lack of proof that the immune system is involved in the majority of

idiosyncratic drug reactions, the idiosyncratic nature of such reactions has led many

investigators to believe that most of these reactions are mediated by the immune system

(Parker, 1982: Pohl et al., 1988). Sorne of the clinical characteristics of many idiosyncratic

reactions also suggest that the immune systern is involved. Such characteristics include:

A requirement of prior exposures to the drug, or a lag period of more than a week

between starting the drug and the development of toxicity.

A lack of delay in toxicity on reexposure of a patient to the offending dmg.

No apparent correlation between dose and the risk of toxicity.

The presence of a skin rash and peripheral blood eosinophilia.

However, there is no reason to presume thar al1 idiosyncratic reactions have an

imrnunological mechanism. The role of direct cytotoxicity has not been ruled out and

remains a controversial issue. However, i t is unlikeiy that many idiosyncratic drug

reactions involve direct cytotoxicity of the drug or its metabolites. If such were tme, the

reaction would be expected to be more predictable and also to be detected in animal toxicity

tests with high doses. Direct cytotoxicity also may not explain the large intenndividual and

interspecies differences in the toxicity since differences in drug rnetabolism do not appear to

be the sole bais for the idiosyncratic character of these drug reactions.

1.4.2 Hapten hypothesis

Unfortunately, not enough is h o w n about the mechanism of many dmg-induced

idiosyncratic reactions to place them into the Gel1 and Coornbs classification scheme.

However, for a drug to cause any imrnunological reaction, it would probably have to act as

a hapten. The term "hapten" has been used to describe a substance which is not

imrnunogenic per se, but becomes immunogenic when conjugated to a rnacromolecuiar

carrier. In generd, in order for a molecule to be recognized as "nonself' and to induce an

immune response, it must have a minimum molecular weight of approximately 1000

daltons (Pohl et al., 1988). Since most drugs are not this large, they must bind to an

endogenous macromolecule. such as protein, to form a drug-protein conjugate before they

c m interact with the immune system. This binding of a drug to a macrornolecular carrier

must be essentially irreversible, usually covalent, for the adduct to be immunogenic. The

reason for the requirement of a covalent interaction is that. for an immune response to be

induced, the irnrnunogen must be processed and presented by antigen presenting cells (i.e.

macrophages). If the interactions between hapten and macromolecule were reversible, the

hapten would diffuse away from the macromolecule before fragments of the adduct could

be presented to T-cells. Therefore, central to the hapten hypothesis is the requirement for

the ability of a hapten to form a stable adduct with nucleophilic groups on proteins (e.g.

lysyl and cysteinyl residues) in aqueous conditions.

Some drugs are chemically reactive themselves so that they can react directly to

proteins, such as penicillin. Penicillin has a reactive structure and can bind covalently to

proteins and carbohydrates by reaction with nucleophilic arnino, hydroxyl, mercapto and

histidine groups. The principal mode of conjugation, referred as the "major" ontigenic

determinant, is the penicilloyl moiety bound to e-amino groups of lysine residues of

proteins. There are two possible routes for the formation of penicilloyl derivatives. One

pathway is the direct aminolysis of penicillin by the protein. An alternative pathway is

through remangement of penicillin to penicillenic acid (Schwartz, 1969). Penicillenic acid

contains a free thiol Croup that can also form mixed-disuifide bonds with cysteine residues

and this is one of the so called "minor determinants" (Figure 1-1). There is also evidence

that penicilloic acid and penamaldate are specific antigenic determinants in penicillin allergy

and they are included in the "rninor determinants" (Schwartz, 1969).

penicillin

1 Protein

Major antigenic determinant

Isornerization

penicillenic acid

penicilloic acid

Penamaldate

Minor antigenic determinants

Figure 1- 1 : The major and rninor antigenic deterrninants of benzylpenicillin (Park

and Kitteringham, 1990; Schwartz, 1969).

21 Some dmgs have a free sulfhydryl group that can react with disulfide linkages to

form mixed disulfides. For example, both D-penicillamine and captopril have a free

sulfhydryl group that can react with cysteine residues on proteins via disulfide linkages.

Imrnunochemical studies have shown that captopil linked by disulfide bonds to proteins

cm function as a hapten in humans and experimental animals (Park er al., 1987). Besides a

free sulfhydryl group, D-penicillamine also possesses a free amino group so that it can

react wirh biological intermediates that contain a free carbonyl group to form thiazolidine

derivatives. It has been shown that penicillamine interferes with collagen maturation by

chernicd reaction with free aldehyde group in pro-collagen (Park er al., 1987).

Unlike penicillin. captopril and penicillamine, most drugs are not chemically

reactive nor sufficiently electrophilic to bind to macromolecules. Therefore most dmgs

must be convened to reactive species within the body in order to act as haptens.

1.3.3 Reactive metabolites involved in idiosyncratic reactions

Although we do not have a complete understanding of the mechanism of drug-

induced idiosyncratic reactions, there are a large number of studies to suggesr that reüctive

metabolites are involved (Hinson and Roberts, 1992; Park et al., 1995) and scxeral

chxactenstics suggest the involvement of the immune system (Pohl et al., 1988).

In generaf, drug metabolism can be considered as a detoxification process in which

it convens therüpeutically active compounds to inactive. more polar metabolites that can

then be removed from the body. However, in cenain circumstances, the drug metabolizing

enzymes can convert a dnig to a chernically reactive metabolite. This process is referred as

"bioacrivaion" (Pirmohamed er al., 1994). Chemically reactive agents cm be cytotoxic by

reacting with critical biological molecules (Jerina and Daiay, 1974; Stadtman and Oliver,

199 1). In most cases. the chemically reactive metabolite will be detoxified before it can

initiate tissue damage. If inadequateiy detoxified, formation of a reactive metabolite as a

22 result of drug bioactivation is often the first step, although not necessarily the ultimate step,

in initiating idiosyncratic drug toxicity. The reactive metabolite may cause direct cellular

toxicity; cause ce11 stress and initiate innate immune responses; or it can act as a hapten and

initiate an immune reaction which may be due to a specific humoral (antibody) response, a

cellular response (T lymphocytes) or a combination of both (Park et al., 1987; Pohl et al.,

1988) (Figure 1-2).

Me tabolic Bioactivation

binding

Detoxification

/ Stable metabolite ]

T ce11 urnora response

Immune-mediated idiosyncrasy

Figure 1-2: The role of drug bioactivation in mediating either direct (metabolic)

idiosyncrasy or immune-mediated idiosyncrasy (Pirmohamed et 01.. 1996).

24 Reactive metabolites can be formed by most of the enzymes that are involved in

drug metabolism. the most important being the enzymes of cytochrome P450 (CYP450)

mixed function oxidase system. Other enzymes also play an important role in drug

bioactivation. In white blood cells and bone marrow cells, for example. MPO has been

shown to bioactivate a wide range of dmgs (Pirmohamed et of., 1996: Uetrecht, 1992) In

other tissues low in CYP450 activity (e.g. kidney, h g , skin), prostaglandin H synthase

(PHs) may also be responsible for bioactivation; e.g. in the kidney acetarninophen toxicity

is thought to result from activation via this enzyme (Pirmohamed et al., 1996).

For most drugs, the occurrence of bioactivation and the chemical nature of the

reactive metabolite have not been definitely determined. This is largely due to the lack of

techniques that allow the detection of in vivo bioactivation and determination of the nature

of the metabolites formed. However, a large arnount of indirect evidence suggests that

many dmgs are bioactivated to structurally different reactive metabolites that may play a

role in the mechanisrn of drug-induced idiosyncratic reactions.

In this discussion, dmgs will be categorized according to the chernical nature of

their reactive metabolites that are generated during the bioactivation of these drugs. Tables

1- 1, 1-2, 1-3, 1-4 and 1-5 are lisrs of dmgs and xenobiotics, the types of toxicity that they

cause and their reactive metabolites generated during bioactivation by various metabolizing

enzymes.

1.4.3a Hydroxylamines

Many drugs have an arylarnine moiety or can be metabolized to arylamines (either

primary or secondary). Studies by various research groups have demonsîrated that

arylamine-containing dmgs can usuaiiy be oxidized to reactive hydroxylamines or further to

the closely related nitroso derivatives. These studies have been carried out in various

enzyme systems (e.g. CYP15O and MPO), cells (i.e. human neutrophils) or tissue

25 homogenates (eg . human and mouse liver microsomes). Typical examples of this class of

dmgs include procainamide, dapsone and sulfonamides (Table 1- 1).

Table 1-1: Formation of the reactive hydroxylarnine intermediates during bioactivation of

dmgs by various oxidizing enzymes.

Idiosyncratic

Toxicity

Agranulocytosis,

Hepatotoxicity,

Drug-induced lupus

Methernoglobinernia,

"Hypersensitivity ",

Mononucleosis-like

syndrome,

Agranulocytosis

Hepatoxicity,

Skin rash,

Blood disorders,

"Hypersensitivity "

Reactive Metabolites

hydroxylarnine

hydroxy lamine

hydroxy lamine

Enzvmes

Procainamide is an

incidence of drug-induced

26 antiarrhythmic agent. but its chronic use is limited by a high

lupus and agranulocytosis (Ellrodt et al., 1984). It has been

shown that procainamide was metabolized to a hydroxylamine by both rat and human

hepatic microsornes and by activated human neutrophils (Uetrecht et al.. 1988a; Uetrecht et

al.. 1984). In vitro the hydroxylamine was oxidized nonenzymatically by hydrogen

peroxide to a nitroso denvative and funher to the nitro derivative (Figure 1-3). Of these

metabolites, the nitroso-procainamide was the most reactive and accounted for most of the

covalent binding to proteins (Uetrecht, 1985). When procainamide was oxidized by the

combination of myeloperoxidase and hydrogen peroxide in the presence of chloride ion, a

reactive N-chloroprocainamide metabolite was formed (Uetrecht and Zahid, 199 1) (Fig 1 -

3) .

Procainamide Hydroxy lamine

MPO/H,O,, CI' nonenzymatic 1

N-Chloro Nitroso

Figure 1-3: Metabolism of procainamide by MPO in activated neutrophils (Uetrecht

et al., 1988a; Uetrecht and Zahid, 199 1).

28 Dapsone is used in the treatment of leprosy and it is also effective in malarial

prophylaxis. Adverse reactions to dapsone include agranulocytosis and severe hemolytic

mernia (Coleman et al., 1989). The rnetabolism of dapsone by activated neutrophils leads

to the formation of the hydroxylamine and nitroso metabolites (Uetrecht et al.. 1988b).

Dapsone has also been shown to undergo bioactivation to a cytotoxic metabolite, in a

nicotinamide adenine dinucieotide phosphate (NADPH)-dependent reaction catalyzed by

microsornes prepared from nine individual human livers (Coleman et al.. 1989). The

kinetics of dapsone hydroxylamine formation in human liver microsornes were biphasic.

111 vitro metabolic studies using enzyme inhibitors, inhibitory antibodies, enzyme

expression systems and correlations with the levels of individual isozymes have detemiined

that CYP3A4 and CYP2El accounted largely for dapsone hydroxylamine formation at high

(100 pmol/l) and low (5 pmolA) substrate concentrations, respectively. These data also

suggested that C Y P X isoforms contribute to dapsone hydroxylamine formation at low

substrate concentration (Fleming et al.. 1992; Gill et al., 1995; Mitra et al., 1995).

Sulfonamides are antimicrobid agents that are associated with various idiosyncratic

reactions in humans which include hepatitis, nephritis, blood dyscrasias. drug-induced

lupus, serum sickness and the suifonamide hypersensitivity syndrome. It has been shown

that in humans. sulfarnethoxazole was metabolized to its hydroxylamine in the liver

predominantly by CYP2C9 (Cribb and Spielberg, 1990b; Cribb et al., 1995).

Sulfamethoxazole was also shown io be oxidized by activated monocytes and neutrophils

(human and canine) to form sulfarnethoxazole hydroxylarnine and nitrosulfamethoxazole;

the latter was presumably formed via the reactive nitroso-sulfamethoxazole intermediate

(Cribb et al., 1990a).

1.4.3b Quinones

A large number of drugs or their metabolites contain a phenyl ring with a hydroxyl

group in the para (or onho) position to another hydroxyl group or an amine or methyl

group. This structure renders the ability of these h g s (or their metabolites) to be oxidized

and convened to reactive quinones. iminoquinones or quinone methides depending on the

hinctional group that is para (or ortho) to the hydroxyl group on the arornatic ring. The

dmgs (or their metabolites) of this category are listed in Table 1-2.

Table 1-2: Formation of the reactive irninoquinone intermediate during bioactivation of

dmgs by various oxidizing enzymes.

Drug

Acetaminophen

Diclofenac

(5- hydroxyl

metabolite)

Vesnarinone

Benzene

(hydroquinone)

Prinornide

(parahydroxy

metabolite)

Idiosy ncratic

Toxicitv

Agrariulocytosis,

Hepato toxici ty

Hepatotoxicity ,

Nephrotoxicity

Hepatotoxicity,

Blood disorders

Aplastic anemia

Agranulocytosis

Reactive Metabolites

iminoquinone

N-ace ty lbenzo-

auinoneimine

irninoquinone

- - -

imidoiminium ion,

irninoquinone

quinone

quinone

--

quinone methide

Metabolizing

Enzvmes

MPO,

HRP

CYP 1A2, CYPSE1,

CYP3A4

MPO

MPO

32 Amodinquine is an antirnalarial agent associated with a high incidence of

agranulocytosis. A number of cases of fatal agranulocytosis, as well as hepatotoxicity. in

patients during prophylactic administration has led to its withdrawal fiom use (Hatton et

al., 1986; Laney et al., 1986; Neftel et al., 1986). Bioactivation of amodiaquine and

covalent binding to various proteins have been demonstrated in several systems: human

liver microsomes, horseradish peroxidase (HRP) I Hz02 or in the presence of chlorine and

activated neutrophils (Jewell et al., 1995; Maggs et al., 1988; Naisbitt et al., 1997). Ir was

shown that a reactive iminoquinone intermediate was formed from amodiaquine by

autoxidation at neutral pH, by peroxidases and by silver oxide oxidation, and by activated

neutrophils in vitro (Christie et al., 1989; Clarke et al.. 1990; Maggs et al., 1987; Maggs et

al., 1988) (Figure 1-4).

Amodiaquine

HRP or MPO/H202

nonenzymatic

lminoquinone

Figure 1-4: Oxidation of amodiaquine to a reactive irninoquinone intemediate

(Park and Kitteringham, 1990).

and hepatotoxicity after

large amount of evidence

Acetarninophen is an anaigesic agent. Renal toxicity

overdoses have been associated with acetaminophen. There is a

to support the involvement of the reactive N-acetylbenzoquinone imine in this toxicity. At

least three f o m of CYP450 (CYPlA2, CYP2El and CYP3A4), as well as MPO and PHs

enzyme systems, have been shown to be involved in acetaminophen bioactivation and the

formation of the reactive iminoquinone (Boyd and Eling, 198 1; Moldeus and Rahimtula,

1980; Potter and Hinson, 1987; Raucy et al., 1989; Thummel et al., 1993).

Formation of the reactive iminoquinone intermediates during bioactivation of

indomethacin and carbarnazepine will be discussed in detail in Chapter 2 and Chapter 3,

respec tively . Diclofenac is a NSAID, and is widely used in the treatment of rheumatoid anhntis

and other inflammatory diseases (Dunk et al., 1982). In rare incidences, diclofenac

appears to be associated with hepatitis (Dunk et al., 1982; Helfgott et al., 1990) and

hematological toxicity, such as aplastic anemia, neutropenia (Eustace et al., 1989; Kaufman

et al., 1991; Kim and Kovacs, 1995; Salama et al., 1989), hemolytic anemia and

thrombocytopenia (Epstein et al., 1990; Kim and Kovacs, 1995; Kramer et al., 1986;

Salama et al., 1991). It has been shown that when diclofenac is oxidized by MPO or

activated neutrophils, the major product was an iminoquinone (Figure 1-5). This

iminoquinone reacted with glutathione and it is also likely to react with protein sulfhydryl

groups which would contribute to covalent binding (Miyarnoto et al., 1997). In addition,

the phenolic precursor of the iminoquinone, 5-hydroxydiclofenac, is a significant hepatic

metabolite of diclofenac (Sawchuk et al., 1995). It was found that the irninoquinone was

also formed by the oxidation of 5-hydroxydiclofenac with rat hepatic microsomes, so it is

likely that the iminoquinone is formed both in the liver and in the bone marrow of humans

(Miyamoto et al., 1997).

Diclofenac

5- hydroxydiclofenac Iminoquinone

Figure 1-5: Proposed pathway of the oxidation of diclofenac to a reactive

iminoquinone (Miyarnoto et al., 1997).

Vesnarinone is used in congestive heart failure and its use is associated with

agranulocytosis. It was found that vesnarinone was metabolized to a reactive metabolite

that covalently bound to activated neutrophils (Uetrecht et al., 1994). The reactive

metabolite was identified as an imidoiminium ion (Figure 1-6). This reactive species can

either hydrolyze to form veratrylpiperazinamide or react with other nucleophiles, such as

34 GSH. Hydrolysis also leads to another reactive intermediate, an iminoquinone, which also

reacts with GSH (Uetrecht et al., 1994).

Vesnarinone

OCHB Imidoiminium ion

OCHj

Veratrylpiperazinamide

Figure 1-6: Proposed pathway of oxidation of vesnarinone

iminoquinone intermediates (Uetrecht et al., 1994).

to reactive

35 Benzene is an environmental toxicant. The most frequently observed toxic effect of

benzene in humans and in animal models has been bone marrow depression leading to

aplastic anemia (Snyder and Kocsis, 1975). Benzene is initially metabolized in the liver by

CYP2E 1 to hydroxylated products: phenols, catechols and hydroquinones. The bone

marrow has been shown to accumulate hydroquinone and catechol (Greenlee et al., 198 1;

Rickert et al., 1979). It has been suggested that the reactive intermediates, p-benzoquinone

and O-benzoquinone, are formed in the bone marrow during oxidation of hydroquinone and

catechol by MPO (Eastmond et al., 1987; Eastmond et al., 1986; Sadler et al., 1988).

These quinone reactive intermediates may be responsible for benzene-induced bone marrow

toxiciiy.

Prinornide is an mti-inflammatory dmg which was associated with a relatively high

incidence of agranulocytosis (< 0.3%) in clinical trials (Parrish et al., 1997). Metabolism

of pnnornide and its parahydroxy metabolite, CGS 12094, by MPO was studied (Parrish et

al., 1997). CGS12094 was found to be rapidly metabolized (> 90%; 2 min) and the

reaction was dependent on Hz02 and MPO and was inhibited by azide (a MPO inhibitor).

During the MPO-catalyzed metabolism of CGS 12094, reactive intermediates that

irreversibly bound to protein and cysteine were generated. One of the reactive metabolites

has been identified as 1,4-benzoquinone which is the same reactive intemediate implicated

in benzene toxicity (Figure 1-7).

Prinomide CGS 12094

Figure 1-7: Proposed mechanism for 1,4-benzoquinone formation from the MPO-

catalyzed metabolism of CGS 120% (Parrish et al., 1997).

Tacnne is a potent, centraily acting anticholinesterase currently used in the treatment

of Alzheimer's disease (Madden et al., 1995a) Tacrine therapy is associated with an

unpredictable hepatotoxic reaction in up to 50% of patients (Pimohamed et al., 1996).

initial in vitro studies dernonstrated hepatic rnicrosomal enzyme-dependent bioactivation of

tacnne to both cytotoxic and protein-reactive species (Madden et al., 1993). The formation

of these metabolites was inhibited by the addition of glutaihione and/or ascorbic acid to the

incubation medium, indicating that electrophilic species were being formed. The enzyme

involved in the generation of chernically reactive metabolites was shown to be CYPlA2

(Madden er al., 1993; Spaldin et al., 1994). A stable thioether conjugate of the reactive

37 species was formed when tacrine (or its 7-hydroxyi metabolite) was coincubated with

microsornes and mercaptoethanol. The reactive intermediate was identified as a quinone

methide (Madden et al., 1995a; Madden et al., 1995b) (Figure 1-8).

Tacrine 7-hydroxytacrine quinone methide

Figure 1-8: Metabolic pathway for the bioactivation of tacrine to its postulated

quinone methide reactive metabolite by CYPlA2 (Pimoharned et al., 1996).

1 .4 .3~ Cations and radicals

A large number of compounds, including many drugs can be oxidized under

appropriate conditions to potentially toxic free radical (or cation) intermediates. The cations

and radicals may cause protein adduct formation or oxidative stress that rnay further lead to

toxicity. Table 1-3 shows sorne examples of this category.

38 Table 1-3: Formation of cation or free radical intermediates dunng bioactivation of dmgs by

various oxidizing enzymes.

Idiosyncratic

Toxicitv

Aminopyrine

Clozapine

Agranulocytosis

Chlorpromazine

Agranulocytosis

Agranulocytosis

dication,

Rex tive Me taboli tes

cation radical 1 I

Metabolking

HRP

cation radical

nitrenium ion

imidoiminiurn ion 1 MPO 1

MPO 1

MPO,

Aminopyrine is an analgesic agent but its use has been lirnited by its association

with a high incidence of agranulocytosis (Madison and Squier, 1934). Aminopyrine has

been shown to be oxidized to a blue cation radical (Figure 2-8) by PHs, MPO and

hypochlorous acid (Eling et al., 1985; Sayo and Saito, 1990; Uetrecht et al., 1995). It has

been proposed by Sayo and Saito that the mechanism by which hypochlorous acid oxidizes

aminopyrine to a cation radical involves N-chlorination followed by the loss of a chlorine

radical (Sayo and Saito, 1990). However, a study by Uetrecht et al. (1995) later observed

a dication intermediate as the precursor for the cation radical. The dication was formed by

the loss of chlonde ion from N-chloroarninopyrine and it is very reactive (tl/z = 15 ms)

(Figure 1-9). The same pathway was suggested when arninopyrine was oxidized by the

combination of MPO, and chloride or activated neutrophils (Uetrecht et al., 1995).

Aminopyrine

Aminopyrine

Cation radical dication

Figure 1-9: Proposed scheme of the oxidation of arninopyrine to a cation radical via

a dication intemediate (Uetrecht et al., 1995).

40 Chlorpromazine, an antipsychotic agent, is known to induce agranulocytosis. It

was found that chlorpromazine can be oxidized by hypochlorous acid or MPO of activated

neutrophils to a free radical (Kalyanaraman and Sohnle, 1985). Covalent binding of 3H-

labeled chlorpromazine to proteins of stimulated polymorphonuclear leukocytes was also

observed (Kelder et al., 199 1). Van Zyl et al. ( 1990) also found that chlorpromazine was

oxidized to a relatively stable radical cation by the combination of MPO and H202.

However, it is not ciear whether the reaction involves a direct one-electron oxidation or

chlorination to forrn the two-electron oxidation product followed by coproportionation with

the parent dmg to fonn two radical cations.

Clozapine is a unique antipsychotic drug that is more effective than standard

neuroleptic drugs in the treatment of refractory schizophrenia. Unfortunately, its use has

been restricted due to a 0.8% incidence of drug induced agranulocytosis. Meiabolism of

clozapine by activated neutrophils, HRP/H202, and MPOIH202 has been studied (Fischer

et al., 1991; Liu and Uetrecht, 1995). Formation of a reactive metabolite was supported by

two pieces of evidence: covalent binding to activated neutrophils and formation of

glutathione conjugate of the reactive metabolites. The identity of the reactive metabolite hris

been suggested to be a free radical or a nitrenium ion (Figure 1- 10). The nitrenium ion

intermediate, which is 2 mass units less than the parent dmg, was detected by mass

spec trometry (Liu and Uetrecht, 1995). Indirect evidence t hat peroxidase systems also

activate clozapine to free radical intermediates stems from the detection of thiyl and ascorbyl

radicals in the presence of glutathione and ascorbate, respectively (Fischer et al., 1991).

Clozapine

,m'O HRP

Nitrenium ion

Cation radical

Fisure 1-10: Proposed pathway for the oxidation of clozapine to a reactive

nitrenium ion or a cation radical by HRP and MPO (Fischer er al., 1991: Liu and Uetrecht.

1995).

Studies of the metabolic pathway of vesnarinone by MPO or by HOC1 suggested

that a reactive imidoiminium ion was involved (Uetrecht et al., 1994). Formation of the

imidoiminium ion could involve a radical cation or chlorination of the nitrogen followed by

loss of HC1 (Figure 1-1 1).

A number of studies have suggested that free radical rnetabolic pathways exist for a

wide variety of compounds and free radical metabolites do seem to be implicated in the

toxic effects of those xenobiotics which are metabolized to free radicals (Mason, 1979).

However, it is rnuch more difficult to obtain direct evidence for free radical formation than

that for the formation of electrophilic intemediates. Even if the free radical intermediates

do exist, they are generally extremely rectctive so that it will be quenched by anything they

encounter first before reacting with important molecules (e.g. proreins). In addition, as has

been shown in many studies, the major reaction of free radicals with GSH is the abstraction

of a hydrogen atom rather than covalent binding. Presumably the same is true of its

reaction with other biological molecules. Therefore, the pattern of toxic effects caused by

free radical intermediates and by electrophilic intermediates are very likely to be different.

0CH3 Vesnarinone

L OCH, Radical cation

N-C hloro

0CH3

Imidoiminium ion

Figure 1 - 1 1 : Proposed mechanisms of forming the imidoiminium ion intermediate

of vesnarinone.

1.4.3d Arene Oxide

A major metabolic pathway for compounds that contain arornatic rings is aromatic

hydroxylation. It has been suggested that arene oxides may be intermediates in this

reaction. Arene oxides are chemically reactive species and their half-lives are usually very

short. For example, the half-life of benzene oxide is less than 2 minutes in water at pH 7

(lerina and Dalay, 1974). Arene oxides. once formed, rapidly undergo isomerization to

phenols or readily react with a variety of nucleophiles including such cellular

macromolecules as DNA, RNA and protein. Thus, arene oxides have become prime

candidates for the "reactive intermediates" responsible for the binding of aromatic

compounds to macrornolecules in the cell. Table 1-4 lists some examples of this category.

45 Table 1-1: Formation of the reactive arene oxide intermediates during bioactivation of

xenobiotics by CYP450.

Xenobiotics

PoIycycIic

hvdrocarbons

Phenytoin

Toxicity

Hepatotoxicity,

Bone marrow

toxicity

Carcinogenicity

- - - - - - - - - - - - - - -

Hepatotoxicity,

Blood disorders,

"Hypersensitivity "

Teratogenicity,

"Hypersensitivity"

- -

Reactive Metabolites

Arene oxide

*ne oxide

Arene oxide

Arene oxide

It has been shown that hepatotoxicities of simple

halobenzenes being metabolized to arene oxides which

(Brodie et al., 197 1). When radioactive brornobenzene

Metabolizing

Enzymes

CYP450

benzene compounds are caused by

covalently bind to hepatic protein

was administered to rats, massive

centrolobular necrosis resulted and the necrotic regions contained significant arnounts of

covalently bound radioactivity (Brodie et al.. 1971). Pior treatment of the animals with

phenobarbitd enhanced both the bornobenzene-induced necrosis and the hepatic binding

(Brodie et al., 1971; Reid and Krishna, 1973). Uthough bioactivation to form an arene

oxide intermediate may lead to benzene compound-induced liver damage, it is less likely to

be responsible for the bone marrow toxicity associated with these compounds since even if

46 smdl amounts can escape the liver, their concentrations at distant sites such as the bone

marrow are likely to be very low. In contrat, catechol-type metabolites can be readily

oxidised to quinone-tvpe metabolites in the bone marrow. Even in the liver microsomes, it

has been demonstrated that for bromobenzene, which does form an arene oxide, most of

the covalent binding is due to a quinone intermediate (Hanzlik et al., 1984; Rietjens et al.,

1997). This suggesrs that arene oxide, although it may be formed, may not be the most

likely reactive metabolite responsible for toxicity.

Polycyclic hydrocarbons are widespread environmental contaminants. Many of

these compounds are known to be carcinogenic in experimental mimals (Freudenthal and

Jones, 1976; Miller, 1978). It is now well established that metabolic activation is required

for the obsewed biolo_oical effects of polycyclic hydrocarbons (Gelboin and Ts'o, 1975;

Sims and Grover, 1974). Benzo[a]pyrene (BP), for example, is metabolized to a wide

variety of metabolites. such as epoxides, dihydrodiols. quinones and phenols (Gelboin.

1980). Evidence suggests that two of the BP-diol-epoxide metabolites (Figure 1 - 12), are

the ultimate carcinogenic forms of this polycyclic hydrocarbon. The two diastereomers,

BP-7. 8-diol-9,lO-epoxides. are highly carcinogenic to newbom mice (Kapitulnik er al.,

1977; Sims and Grover. 1974); are potential mutagens and cytotoxins to both mammalian

and bacterial cells (,Huberrnan et cd., 1976; Marnett and Bienkowski, 19S0; Wood rr rd.,

1977) and are chernically very reactive in binding to DNA (Jeffrey et al., 1977).

Figure 1- 12: Bioactivation of benzo[a]pyrene to the diol-epoxide cürcinogens.

It has been proposed that the reactive metabolite of carbamazepine. which may be

responsible for carbarnazepine-induced adverse reactions, is an arene oside. The evidence

to support this hypothesis is that carbamazepine can be bioactivated by mouse and human

liver microsornes to a cytotoxic and protein-reactive metabolite (Pirmohamed et al., 19921:

Pirmohamed e t ni., 1992b). The cytotoxicity and covalent binding were enhanced by

trichloropropene oxide (TCPO), which is an epoxide hydrolase inhibitor (Riley et al.,

48 1989). However, TCPO is not a specific inhibitor for epoxide hydrolase, it is also known

to deplete glutathione (Pessayre et cil.. 1979) and inhibit CYP4.50 (Ivanetich er al.. 1982).

Furthemore, studies have failed to find a consistent mutation, or pattern of mutations. in

the microsomal epoxide hydrolase gene which is common in patients with a history of

carbamazepine hypersensitivity reactions (Gaedigk et al., 1994; Green et al.. 1995). The

only direct evidence for the arene oxide formation in vivo is detection, in the bile, of the

GSH conjugate of the postulated arene oxide in rodents (Arnore et al., 1997; Madden et al.,

1996). However, in humans, the same GSH-conjugate was not detected nor were any of

its funher metabolized products, such as the NAC, cysteinyl or thiomethyl derivatives

Waggs et al., 1997).

Phenytoin is a cornmonly used anticonvuisant for the treatment of epilepsy.

Phenytoin causes a wide range of adverse reactions including hypersensitivity and i t has

been shown to be teratogenic in animals and humans (Winn and Wells. 1995). It has been

postulated that CYP450 bioactivation to an electrophilic arene oxide intemediate may play

an important role in the mechanism of phenytoin-induced toxicities. Evidence for the

formation of the arene oxide intermediate is supported by the ability of phenytoin to bind

covalently to molecular tnrgets. by the formation of the trans-dihydrodiol metabolite of

phenytoin (Ocsch. 1973) and by the formation of a catechol or its rnethylated derivative

(Billings, 1985; Billings and Fischer. 1983: .Maguire et al.. 1979). Nevenheless. none of

these observations are direct proof of the formation of an arene oxide intermediate (Brown

rr al.. 1986). It is known that arene oxides zenerally react readily with GSH to form stable

conjugates (Jerina and Dalay, 1974): however, to date neither the GSH adduct of the

putative arene oxide, nor its urinary mercapturate metabolite, has been found (Hansen,

1991 J. Recently, an alternative reactive metabolite, an O-quinone, has been proposed by

Gillarn and coworkers and it was demonstrated that the phenol and catechol metabolites

lead to more covalent binding than the parent drug in vitro (Munns et al., 1997).

1.4.3e Miscellaneous

There are many other drugs that can be bioactivated to chemically reactive

metabolites. The chernical nature of these reactive metabolites, however, are different from

those discussed earlier in this section. They are classified as miscellaneous drugs (Table 1-

5) and will be discussed individually.

Table 1-5: Formation of reactive metabolites during bioactivation of dmgs by various

oxidizing enzymes.

Tienilic acid

Halothane

Idiosyncntic

Toxic itv

Hepatotoxicity 1 trifluomacetyl

Reactive Metaboli tes

Hepato toxici ty sui foxide

Hepatotoxicity, 1 sulfonic acid

AgranuIocytosis,

Drug-induced lupus 1

sulfenyl chlonde,

Metabolizing

Enzvmes

Tienilic acid, a diuretic agent, was withdrawn from general use because of a high

incidence of hepatitis (Dansette er al., 1991). Bioactivation of tienilic acid has been shown

to involve CYP2C9 and to be inhibited by anti-human CYP2C antibodies. The reactive

metabolite has been postulated to be the sulfoxide of tienilic acid (Mansuy et al., 1991)

(Figure 1-13). It is likely that the bioactivation of tienilic acid to a sulfoxide metabolite by.

50 and subsequent binding to, the C Y P X proteins causes the antigenic alteration of this

CYP450 isoform resulting in antibody production.

Tienilic acid Sulfoxide

Figure 1 - 13: Metnbolism of tienilic x i d to the sulfoxide metabolite by CYP7C9

(Mansuy et ai., 199 1 ).

Halothane, a widely used anesthetic, i n thought to cause two forms of liver damage

(Neuberger and Kenna. 1987). The minor form is characterized by slightly raised

transaminases and perhaps mild necrosis which occurs in up to 20% of patients. The more

severe form of halothane hepatitis (Le. massive liver ce11 dmage) is observed in 1 :35,000

patients. The incidence of this reaction is increased (l:3.700) after multiple exposure to

halothane (Park and Kitteringham, 1990). Halothane is metabolized in the liver by

cytochrome P450 enzymes by oxidation and reduction to two chemically reactive

metabolites. trifluoroacetyl chloride (TFA) and the 1-chloro-2.2.3-trifluorocetyl radical. It

has been demonstrated that it is the trifluoroacetyl group which forms the neoantigen on

rnicrosomal protein and that sera from the majority of patients with halothane hepatitis

contain antibodies that recognize the trifluoroacetyl epitopes (Kenna et al.. 1988).

Propylthiouracil is a thioureylene antithyroid dru, and its use is associated with

several hypersensitivity reactions. The most cornmon serious reaction is agranulocytosis;

others include toxic hepatitis and a syndrome similar to SLE (Cooper, 1984)

51 Propylthiouracil is accumulated not only in the thyroid gland but also in phagocytizing

poiyrnorphonuclear leukocytes ( L m and Lindsay, 1979a; Lam and Lindsay, 1979b).

Propylthiouracil was found to be degraded by MPO or eosinophil peroxidase, purified

from rat bone marrow. in the presence of H2Oz and Cl-. One of the metabolites was

identified as propylthiouracil-sulfonic acid (Lee et al.. 1988). A study by Waldhauser and

Uetrecht (1991) has found that propyithiouraci1 was metabolized by activated neutrophils or

iMPO/H202/Cl- systrm. The first step is presumably the formation of the sulfenyl chloride.

This metabolite was expected to be reactive and was not detected directly. The disulfide

metabolite was isolated and it was funher oxidized to the sulfonic acid (Figure 1-14). The

sulfonic acid is also reactive and reacrs with thiols. such as glutathione, to form a thioether.

Propylthiouracil Sulfenyl chloride Suifonic acid

Figure 1- 14: Metabolism of propylthiouracil by MPO to the reactix sulfonic acid

(Wddhauser and Uetrecht, 199 1).

1.5 -Metabolism of indomethacin and carbamazepine

This section summarizes briefly the general pharmacokinetics and metabolism of

indomethacin and carbamazepine.

1.5.1 Indomethacin

Indomethacin is rapidly and almost compleiely absorbed from the gastro-intestinal

tract in healthy adults. Following oral administration, bioavailability is virtually 100%.

with 90% of a single dose being absorbed within 4 hrs. At therapeutic concentrations,

indomethacin is approximately 99% bound to plasma proteins.

In various laboratory anirnals, indomethacin has been found to undergo extensive

biodegradation via O-demethylation, N-deacylation or both (Harman et al., 1961: Hucker et

al., 1966). The respective products of these reactions are desmerhylindomethacin (DMII),

descNorobenzoylindomethacin (DBI) and DMBI. A study by Duggan et ni. ( 1972) has

also shown that in humans demethylation, followed by deacylation, constitues the major

pathivay. and direct deacylation is a minor parhway. Demethylation oppears to be

exclusively mediated by CYP2C9 in humans (Nakajima et al., 1998). An enzyme in pig

liver rnicrosomes that catalyzes the hydrolysis of amide-linkage of indomethacin has been

purified and characterized (Terashima et al., 1996). Due to the similar activity of

indomethacin hydrolysis in pigs and in humans, it is likely that the hydrolysis of amide-

linkage of indomethacin in humans would be associated with an enzyme similar to the

indomethacin hydrolyzing enzyme found in pig liver microsornes. Indomethacin and its

metabolites are excreted in urine and bile both in the free form and as conjugates. Fecal

excretion of radioactivity was shown to vary from 21 to 42% of the dose over (i period of

four days, but it was invariably comprised predominantly of the two deme thy lated

metabolites (> go%), almost entirely in their unconjugated forms.

1.5.2 Carbamazepine

Carbamazepine is slowly but almost completely absorbed from the gastrointestinal

tract. The peak concentration of the dmg occurs 4-10 hrs after a single-dose ingestion. In

normal healthy volunteers the apparent plasma half-life of carbarnazepine ranges from 25-

65 hrs (Morselli and Frigeno, 1975). Carbamazepine is extensively metabolized both in

human and rat and more than 30 metabolites have been isolated and identified. Following

the administration of 14C-carbarnazepine, 72% of the 14C-activity was recovered in urine;

the remaining radioactivity being recovered in feces (Eichelbaum et al., 1984). In ilirro and

NI vivo studies of the bioactivation of carbamazepine in man and rat have demonstrated that

epoxidation of the double bond in the 10,ll-position and hydration of this 10, l l -

carbamazepine epoxide to trans- l0,ll-dihydrodihydroxycarbamazepine is a major route of

carbamazepine metabolism (Eichelbaum et al., 1984). Other important pathways include

hydroxylation of the aromatic rings, N-giucuronidation of the carbamoyl side chain and

substitution of sulfur-containing groups on the aromatic rings (Lertratanangkoon and

Horning, 1982; Lynn et al., 1978). One of the major merabolites, which is generared frorn

hydroxylation of the aromatic rings, is 2-hydroxycarbamazepine (Eichelbaum et al., 1984).

It can be funher metabolized to 2-hydroxyiminostilbene, rither in the iiver or in other

tissues. Both 2-hydroxycarbmazepine and 2-hydroxyiminostilbene have been detected in

the urine of rats and hurnans after hydrolyzed from their glucuronide conjugates

(Lenratanangkoon and Homing, 1982).

It has been postulated that the reactive metabolite, which may be responsible for

carbarnazepine-induced toxicity, is an arene oxide (Pirmohamed et al., l992b; Shear and

Spielberg, 1988). The arene oxide is chemically remive and cannot be characterized by

routine analytical methods such as mass spectrometry and HPLC; therefore, its existence

can only be implied from the urinary excretion of unquantified phenols and catechols of

54 carbarnazepine (Lenratanangkoon and Horning, 1982; Regnaud et al., 1988). However,

the aromatic hydroxyl metabolites are not necessarily derived from an arene oxide

intermediate (Hanzlik et al., 1984). Furthermore, although the detection of glutathione

(GSH) conjugate of the postulated arene oxide intermediate in rodents has been reported

(Amore et al., 1997: Madden et al., 1996). in humans. the same GSH-conjugate was not

detected nor were any of its funher metabolized products, such as NAC or thiomethyl

derivatives (Maggs et al., 1997).

Chapter 2

Oxidation of a metabolite of indomethacin (DMBI) to reactive

intermediates by activated neutrophils, HOC1 and the

myeloperoxidase system.

The use of indomethacin is associated with a relatively high incidence of adverse

reactions such as ûgranulocytosis. Many other drugs associated with agranulocytosis are

metabolized to reactive metabolites by activnted neutrophils. Therefore. we studied the

oxidation of indomethacin and its metabolites by activated neutrophils, MPO (the major

oxidizing enzyme in neutrophils), and HOCl (the major oxidant produced by activated

neutrophils). No significant oxidation of indomethacin by activated neutrophils was

observed. However. DMBI, a major metabolite of indomethacin. was oxidized to a

reactive iminoquinone which could be trapped with GSH or NAC to form conjugates with

W H + ions at m/z 5 1 1 and m/z 367, respectively. No metabolism was detrcted in

neutrophils that had not been activated and the oxidation was inhibited by azide, which

inhibits .VPO, and by catalase, which catalyzes the breakdown of H202. In reactions with

HOCI. the same reactive intermediate was formed; its mass spectrum, with a MH+ ion at

m/z 201, was obtained by using a fIow system in which the reactants were fed into a

rnixing chamber and the products flowed directly into the mass spectrometer. The same

GSH and NAC conjugates were also observed when DMBI was oxidized by HOCl or by

the MPO system followed by addition of GSH or NAC. NMR data for the NAC conjugate

indicated that the sulfur was substituted in the 4 position on the aromatic ring. The reactive

intermediate generated from DMBI by activated neutrophils may be responsible for the

indomethacin-induced agranuloc y tosis.

2.2 Introduction

Indornethacin is a very effective nonsteroidal antiinflarnrnatory dmg, but its use is

limited by a high incidence of adverse reactions, including blood disorders. Of a total of

1,261 reports of adverse reactions to indomethacin reported to the United Kingdom.

Cornmittee on Safety of Medicines between June 1964 and January 1973, blood disorders

were recorded in 157 cases (with 25 fatalities) including thrombocytopenia (35 cases, with

5 fatalities), aplastic anemia (17 cases, with no fatalities), and agranulocytosis or

leukopenia (2 1 cases, with 3 fatalities) (Cuthbert, 1974). Subsequently, the First Report

from the International Agranulocytosis and Aplastic Anemia Study confirmed a significant

relationship between the use of indomethacin and agranulocytosis and aplastic anemia. In

this population-based case-control study conducted in Europe and Israel, the excess risk

estimated for agranulocytosis and aplastic anemia was 0.61106 for indomethacin exposure

in a 1-week period (Shapiro, 1986). The rnechanism of drug-induced agranulocytosis is

unknown: howe\*er. i t is speculated that reactive metabolites produced by activated

neutrophils or neutrophil precursors may be responsible for agranulocytosis, cither by

direct toxicity or via an immune rnechanism. It has been shown that several drugs that are

associated with agranulocytosis are metabolized to reactive intermediates by activated

neutrophils and monocytes (Uetrecht. 1990: Uetrecht. 1996; Uetrecht, 1992). When

activnted. neutrophils release MPO and generate Hz02 (Klebanoff, 1965; Weiss. 1989).

.MPO is the major oxidizing enzyme in neutrophils: HOCl is generated by the MPO system.

which uses Hz02 to oxidize Cl- to HOCl.

A study by Duggan et al. (1977) has demonstrated that the dominant pathways of

indomethacin metabolism in humans are demethylation and deacylation. The major

metabolites are D-MI and DMBI (Figure 3-1). The objectives of this work were to

derennine whether indomethacin and/or its major metabolites can be metaboIized to reactive

58 intermediates by activated neutrophils, HOC1 and the MPO system and, if so, to identify

the proposed reactive metabolite and to characterize its chernical reactivity.

Indomethacin

amide hydrolase or nonenzyrnatically

DiMBI

Fijure 7- 1 : Major metabolic pathway of indomethacin.

2.3 ,Materials and Methods

2.3.1 Synthesis of DMI.

DM1 was synthesized by demethyintion of indomethacin using a modification of the

method of McOmie et rd . (1 968). Indomethacin (0.7 g, Sigma Chernical Co., St. Louis.

MO) was dissolved in 200 ml of dichloromethane. To this solution, 10 ml of 1 M BBr3

was added dropwise. The reaction was protected from moisture by bubbling N? through

the reaction mixture. The reaction was carried out at room temperature for 3 hr. The

reaction mixture was then shaken with water to hydrolyze excess reagent and boron

complexes. DM1 was obtained by extraction into ether. The mass spectmm of the

synthesizrd DM1 included of a MH+ ion ar m/z 3-11. The synthesized DM1 demonstrated

the same HPLC retention time as did an authentic sample provided by Merck Research

Laboratories, and it was used to synthesize DMBI without further purification.

2.3.2 Synthesis of DMBI.

D.Ml (15 mg) was hydrolyzed in 3 ml of 1 'i sodium hydroxide. The reaction wüs

carried out at room temperature for 5 min. The reaction mixture was made ocidic with

concentrated hydrochloric acid and then extracted with ethyl acetate. The ethyl acerate layer

was drird with magnesium sulfate, and the solvent was evaporated under a Stream of 81.

The product was rhen purified by silica gel TLC. using a solvent of dichloromethane /

methmol (9: 1, v / v). The yield was approximately 10% [m.p.= 15 1 - 152°C; Rf = 0.35;

MS: m/z 706 (MH*) and 160 [(MH-HCOOH)']; 'H NMR (dirnethylçulfoxide-d6) 6 10.64

pprn ( iH. s); 6 8.50 pprn ( lH, bs); H-7. S 7.00 pprn ( 1H, d, J=8.54 Hz); H-4, 6 6.72

pprn ( 1H' d. J=2.20 Hz); H-6, 6 6.48 pprn (1H. dd, J=8.30 Hz, 2.20 Hz); methylene

protons. S 3.42 pprn (2H. s); methyl protons on the indole ring, 6 2.28 ppm (3H, s)].

2.3.3 Oxidation of DMBI by HOCI.

Sodium hypochlorite (various concentrations) was added to DMBI (final

concentration, 0.1 rnM) in 100 pl of 0.1 M phosphate buffer at room temperature. The pH

of the phosphate buffer was varied from 4.0 to 7.0. Aliquots (20 pl) of the solution were

analyzed by HPLC.

2.3.4 Oxidation of indomethacin and DM1 by HOCI.

Sodium hypochlorite (50 pl of a 4 mM solution) was added to indomethacin or

DM1 (final concentration, 1 mM) in 150 pl of 0.1 M phosphate buffer (pH = 6) at room

temperature. Aliquots (20 pl) of the solution were analyzed by HPLC.

2.3.5 Oxidation of DMBI by the MPO system.

DMBI (5 pl of a 10 mM methanolic solution) was added to 500 pL of phosphate

buffer (O. 1 M, pH 6 or 7) to yield a final concentration of 0.1 mM. MPO (2.5 units / mL)

was added. and then the renction was initiated by the addition of hydrogen peroxide (final

concentration. 0.4 miM1. Incubations were carried out in the absence or in the presence of

added chloride. Aliquots (20 pl) of the solution were analyzed by HPLC.

2.3.6 Oxidation of DMBI, indomethacin and DM1 by activated neutrophils.

Blood was drawn from normal volunteers, into heparinized syringes, and the

neutrophils were isolated by differential centrifugation on Ficoll-paque (Pharmacia,

Uppsala. Sweden) according to a standard procedure (Boyum, 1984). Briefly, whole

blood was rnixed with equal volumes of 3 8 dextran (wlv) made up in 0.9% saline. This

62 mixture was allowed to settle for 30 minutes and the supematant was then centrifuged

(SOOxg, 5 minutes). The pellet was resuspended and residual erythrocytes were lysed by a

brief exposure to 0.2% saline. The neutrophils were suspended in phosphate buffer saline,

underlaid w ith Ficoll-Paque and centri fuged for 25 minutes (500xg). Following

centrifugation, the neutrophil pellet was washed twice with Hank's buffer before use.

Neutrophils (1 x 106 cells / ml) were suspended in phosphate-buffered saline (pH 7.4.0.5

ml). Phorbol myristate acetate (10 ng / ml ; Sigma) was added. followed by DMBI (5 pl of

a 10 mM methanolic solution, to a finai concentration of IOOpM). The suspension was

then incubated at 37OC in a shaking water bath. After incubation for vanous periods, the

cells were centrifuged and 30 pl of the supematant was analyzed by HPLC. Indomethacin

and DM1 were dso incubated with activated neutrophils under the same conditions, except

that their concentrations were twice that of DMBI.

2.3.7 Synthesis of reactive intermediate-NAC adduct.

Sodium hypochlorite ( 1 ml of a 100 mM aqueous solution) was added to DMBI (50

mi of a 2 mM aqueous solution). The reacrion mixture was vortex-mixed for 10 sec, and

then NAC (0.5 ml of a 200 miM aqueous solution) was added. The mixture was then

concentrated on a rotary evaporator. The NAC adduct was purified by HPLC using a 10-

x 150-mm preparative column packed with 3-pM Spherisorb ODS 2 material. The solvent

consisted of water, acetonitrile, acetic acid, and triethylamine (89: 10: 1:0.05, v/v). The

flow rate was 3 rni/rnin, and the retention time of the NAC adduct was 9.4 min.

2.3.8 Reaction kinetics determined by spectrop hotometry.

A Hewlett-packard diode-array spectrophotometer (HP8452A, Hew lett-Packard,

Pa10 Alto, CA) was used to detemine the rate of oxidation of DMBI by HOCl. Scanning

63 of the reaction mixture was initiated immediately after the addition of sodium hypochlorite

(10 pl of a 10 mM aqueous solution) to DMBI (2 ml of a 50 pM aqueous solution), with

rapid stirring. The reaction was monitored at 1.0-sec intervals for S s e c . over a

wavelength range of 200-500 nm.

2.3.9 Analytical Methods.

HPLC was performed using a 2 x 100 mm Phenomenex Column (Phenornenex,

Torrance, CA) packed with 5-pM Ultracarb ODS 30 matenal. The solvent used to analyze

the products of DMBI oxidation by HOCl, MPO and activated neutrophils consisted of

water, acetonitnle and acetic acid (54: 15: 1. v 1 v) containing 1 rnM ammonium acetate. The

solvents used to analyze the products of indomethacin and DM1 oxidation by HOCl and

activated neutrophils consisted of water, acetonitrile and acetic acid (4950: 1 and 59:40: 1,

V/V, respectively). The flow rate was 0.2 N m i n .

A Sciex API III mass spectrometer (Perkin-Elmer Sciex, Thornhill. Ontario,

Canada) was used for LC/MS. Collisional activation spectra were obtained using the

LCLvfSMS mode, with argon as the target sas. A splitter was used to decrease the flow

inro the mass spectrometer to approximately 10 pl I min.

I H NMR spectra were recorded at 500 MHz with a Brucker AM 500 spectrometer

(Brucker Canada; Milton, Ontario). Spectra were obtained in D20 except for that for

D.MBI, which was obtained in a dimethylsulfoxide-dg. The structure of the NAC conjugate

of the iminoquinone intermediate was assigned by the conventional I H NMR as well as the

!H-1H and I H - ~ ~ C correlation experiments.

2.4 Results

2.4.1 Oxidation of DMBI by activated neutrophils.

DMBI was metabolized by neutrophils to a major metabolite with a MH+ ion at rn/z

409 and a minor metaboiite with a MHf ion at m/z 407. The retention times of the two

products with MHf ions at m/z 409 and m/z 407 were 5.8 min and 24.6 min. respectively.

No signifiant metabolism occurred in the absence of activation of the cells. The

metabolism increased with time, initial DMBI concentration and ce11 number (Figure 2-2).

and i t was inhibited by azide and catalase. The ûmount of the product with a MH+ ion at

m/z 409 was decreased by a factor of 2 and 10 when catalase (200 units./ ml) or azide (0.2

mM), respectively, was included in the incubation mixture.

The M S M S spectra of the metabolites with MH+ ions at rn/z 409 and m/z 407 are

shown in Figure 2-3. To obtain enough material for UV and NMR analysis. rnanganese

dioxide was used to oxidize DMBI; this procedure yielded the same products, except that

the amount of the product with a MH+ ion at m/z 107 was greater than that of the product

with a MH+ ion at rnfz 409. The UV absorption spectrum of the product with a MHf ion

at m/z 109 demonstrated hm,, values of 215 and 300 nm. The UV absorption spectrum of

the product with a MH+ ion at rnlz 407 showed an additional broad peak with a hm,, value

of approximately 480 nm (Figure 2-4). When the CV absorption spectrum was obtained at

a higher concentration, this broad peak was more obvious. The 'H NMR spectra of both

products were obtained. The NMR of the product with a MH+ ion at m/z 407 showed that

the aromatic region consisted of five protons, as follows: H-7. 6 7.38 ppm ( lH , d. 1 =

9.77 Hz); H-7', 6 7.29 ppm (1H. d, J = 8.55 Hz): H-6. 6 6.7 1 ppm ( lH, dl J = 8.55 Hz);

H-4, 6 6.70 pprn ( lH, d, J = 1.71 Hz); H-6', 6 6.65 pprn (IH, dd, I = 9.88 Hz, 1.58

Hz). The loss of the proton signal for the +position of the aromatic ring suggested the

stmcture shown in Figure 2-5. A GSH conjugate wirh a MHf ion at m/z 714 was formed

when this product (bWf ion at mlz 407) was reacted with GSH. This result is consistent

with the structure of this product having an iminoquinone moiety (Figure 2-5). The

simplicity of the NMR cf the product with a MH+ ion at m/z 409 indicated a highly

symmetncal moiecule [H-7. 6 7.32 ppm ( l H , d, J = 8.55 Hz); H-6. ô 6.78 ppm (1 H, a, J

= 8.55 Hz); methylene protons of the side chain, 6 2.88 pprn (2H, dd, J = 18.07 Hz);

protons of the methyl group, 62.22 pprn (3H. s)]. The NMR of the aromatic region

suggested that the proton in the 4 position was substituted by another identical molecule.

Figure 2-5 shows the postuiated pathway of the formation of the two stable metabolites

generated in the oxidation by activated neutrophils.

When GSH or NAC was included in the incubation mixture, the GSH or NAC

conjugates of a reactive intermediate were observed, with MH+ ions at rnfz 5 1 1 and m/z

367, respectively. At the same time, the amount of the product with a MH+ ion at rn/z 409

was markedly decreased (data not shown).

O 20 40 60 80 Time (min)

10 30 50 70 90

[DMBI] (PM)

0.0 2.0 4.0 6.0 8.0

[Neutrophil] (x 106 cells/mL)

Figure 2-2: Metabolism of DMBI by activated neutrophils as a function of time

(A), initial DMBI concentration (B), and neutrophil concentration (C) . showing decreases

in the DMBI concentration (open circle) and the production of the major product (filled

square), which has a MHf ion at m/z 409 by MS. Standard conditions were as follows:

neutrophil concentration, Ix 106 cells/mi; DMBI concentration, 40 pM; phorbol mynstate

acetate concentration. 40 ngml; incubation time, 30 min. The peak areas are in arbitrary

units, because we did not have enough materid to determine a standard curve. The values

are the mean + SD from three experiments.

Figure 2-3: i M S M S spectra of the metabolites with MHf ions at m/z 409 (.A) and

d z 107 (B), formed during DMBI oxidation by iictivated neutrophils.

Wavelength (nm)

Figure 2-4: UV absorption spectrurn of the metabolite wirh a MHf ion at 407.

during DM31 oxidation by activated neutrophils.

(M+1= 204) Iminoquinone

DMBI intermediate

H

U f' I AT -

Figure 2-5: Proposed pathway of the formation of the two stable metabolites

during DMBI oxidation by activated neutrophils.

2.4.2 Oxidation of DMBI by HOCl

DMBI was readily oxidized by HOCl. Scanning of the UV absorption spectrum of

this mixture at 1-sec intervals resulted in the pattern shown in Figure 2-6. The absorbance

of DMBI (Amax = 280 nm) decreased rapidly and a new species was generated, with

increased absorption at longer wavelengths.

A----O sec BI---1 sec C----2 sec D----3 sec E-0-4 sec F m - - 5 sec G---10 sec H--45 sec

Wavelength (nm)

Figure 2-6: Repetitive absorption spectra from the reaction of DMBI with HOCI.

The concentrations of both DMBI and HOCI were 50 pM.

When the products of the reaction between DMBI and HOC1 were analyzed by MS

in the flow system, so thar the short-lived intermediate could be detected, an intermediate

with a MH+ ion at m/z 204 was observed. Using the MSIMS mode. the major fragment

ions were at d z 176 [(MH- CO)^ and 158 [(MH-HCOOH)+J.

When GSH or NAC was added imrnediately to the reaction mixture of DMBI and

HOCI, the GSH or NAC conjugate of the intermediate with a MH+ ion at mlz 204 was

observed. The GSH and NAC conjugates formed by trapping the reactive intermediate

generated during DMBI oxidation by neutrophils and HOC1 were found to have the sarne

retention time and molecular weights and identical MS/MS spectra. In addition. a n-

butylamine conjugate of this intermediate with a MH+ ion at m/z 277 was observed by

LClMS when n-butylarnine, instead of NAC was added to the mixture of DMBI and HOCI.

The NAC conjugate was isolated and the NMR analysis was perfomed (Figure 2-7). The

H NMR data, combined with the finding from IH-IH and I H - ~ ~ C correlation experiments

(Figure 2-9), confirmed the structure of the conjugate where the sulfhydryl group of NAC

was substituted in the 4-position on the aromatic ring (Figure 2-8).

When DMBI was oxidized by HOCI, another product was formed, with a MH+ ion

at m/z 138 and a chlorine isotope peak for the MHf ion at m/z 240. The MS/MS specuum

of this product showed a major fragment ion at rn/z 202, resulting from the loss of HCl,

which suggested that this was a chlorinated species. This product was not observed when

the HOC1 concentration was lower than one fifth of the D m 1 concentration. It wris also

not observed when GSH or NAC was added irnrnediately to the reaction mixture of DLMBI

and HOC1. These results suggested that this chlorinated species was formed by funher

oxidation with excess HOCI.

(M+1 = 206) DMBI

O

(M+1= 204)

Iminoquinone intermediate

H

(M+1 = 367) Iminoquinone-NAC conjugate

Figure 2-8: Proposed scheme for DMBI rnetabolism by activated neutrophils and

trapping the reactive intermediate by NAC.

2.4.3 Oxidation of DMBI by MPO system

Compared with the oxidation by neutrophils, a sirnilar pattern of products was

obtained when DMBI was oxidized by the MPOM202 system without added Cl-. The

major product showed a MH+ ion at m/z 409. When GSH was included in the incubation

mixture, the sarne GSH conjugate, with a MH+ ion at m/z 5 1 1, was observed. When 0.15

rnM Cl- was added to the phosphate buffer, the reaction was more rapid and the major

product observed had a MH+ ion at d z 238. This is the same product as that observed

when DMBI was oxidized by higher concentrations of HOCl.

2.4.4 Oxidation of indomethacin by HOC1 and activated neutrophils

Indomethacin was oxidized by HOC1 to a major product with a molecular weight of

39 1, corresponding to indomethacin plus chlorine minus hydrogen. However, no GSH

conjugates were observed when GSH was added to the reaction mixture of indomethacin

and HOCI. When indomethacin was incubated with activated neutrophils under the same

conditions as those used for DMBI, no oxidation was observed.

2.4.5 Oxidation of DM1 by HOCl and activated neutrophils

DMI was oxidized by HOCl to two major products, with retention tirnes of 16 and

28 min. The molecular weight of the product with a retention time of 16 min was 2 rnass

units less than that of DMI. The other product, with a retention time of 28 min, showed a

molecular weight of 378, corresponding to DM1 plus chlorine minus hydrogen. No GSH

conjugates were observed when GSH was added to the reaction mixture of DM1 and

HOCI. When DM1 was incubated with activated neutrophils, only the product wirh a

retention time of 16 min was observed. When GSH was included in the incubation

mixture, no GSH conjugates were detected.

2.5 Discussion

DMBI was oxidized by neutrophils to a reactive intermediate that could be trapped

by glutathione. MPO appears to be the enzyme in activated neutrophils responsible for this

oxidation. This is supported by our observations that the MPO/H202 system generated the

sarne metabolites as did neutrophils, the metabolism by neutrophils required activation of

the cells (resulting in release of MPO and the generation of HzOz), and that the metabolism

by neutrophils was inhibited by azide (which inhibits MPO) and catalase (which catalyzes

the breakdown of H102).

The mechanism presumably involved N-chlorination. f~'olhved by the loss of HCl

to form an iminoquinone, as shown in Figure 2-8. The N-chloro-DMBI was not directly

observed. However, strong evidence for the identity of the reactive iminoquinone

intermediate was obtained. The diode-array UV spectrophotometric data showed that

DMBI absorbance (kmax= 280 nm) decreased rapidly upon addition of HOCl and a new

species, with absorbance at longer wavelengths (suggesting a quinone-like structure), was

formed within seconds. When the products of the reaction between DMBI and HOCl were

analyzed by MS in the flow system, the iminoquinone intermediate. with a MH+ ion at m/z

204, was observed. This irninoquinone intermediate was reactive with sulfhydryl-

containing molecules. such as GSH and NAC. The proton K-MR data confimed the

structure of the imnoquinone-NAC conjugate, in which the sulfur was substituted in the

ortlzo-position (relative to the hydroxyl group) on the aromatic ring.

When DMB 1 was oxidizrd by activated neutrophils. the final stable metabolites,

with MHf ions at m/z 409 and m/z 407, were formed. Both products were generated from

the irninoquinone intermediate reacting with another molecule of DLMBI. In the case of the

product with a MH+ ion at d z 409, the iminoquinone intermediate reacted with another

DMBI molecule in the 4-position, which is orrho to the hydroxyl group. The aromatic

carbon orrho to the hydroxyl group of DMBI rnolecule is expecred to be electron-rich and

78

could act as a nucleophile to attack the carbon-4 of the irninoquinone intermediate (Figure

2-5). Another possible mechanism for the formation of the product with a MHf ion of 409

involves a coproportionation reaction between DMBI and the iminoquinone to f o m two

radicals that couid dimerite. However, because of the solvent cage effect. the free radical

might not be able to react with other molecules and conuibute to the toxicity; even if a free

radical is formed, the major reaction with GSH involves the electrophilic attack of the

irninoquinone, rather than the abstraction of a hydrogen atom by a free radical. Presumably

the same is tme of the reaction between the reactive metabolite and other biological

molecules. In the case of the rninor product with a MH+ ion at rn/z 407, the iminoquinone

intermediate reacted with another molecule of DMBI, on the nitrogen of the indole ring, to

form a dimer. This is consistent with Our observation that the iminoquinone intermediate

can react with nitrogen-contoining nucleophiles (e.g. n-butylamine), in addition to

sulfhydryl-containing nucleophiles. The initially formed dimer can be further oxidized to

yield the product with a MH+ ion at rn/z 407 (Figure 2-5).

A product with a MH+ ion at rn/z 238 was observed only when DMBI was

oxidized by HOC1 at high concentrations (presumably yielding an cxcess of chlorine).

When GSH or NAC wiis added immediately into the reaction mixture of DMBI and HOCI,

this product with a iMH+ ion at m/z 238 was not observed, which suggested that it was a

downstream product where the iminoquinone was the initial reactive metabolite. This

product. with MHf ion at m/z 238, was not observed when DMBI was oxidized by

activated neutrophils or by purified MPO in the absence of added CI-. This product is

probably not a significant N i vivo metabolite.

The formation of reactive metabolites is centrai to most mechanistic theories of dnig

hypersensitivity reactions (Uerrecht, 1990). Studies frorn Our laboratory have found that

several dmgs that are associated with a relatively high incidence of hypersensitivity

reactions (especially agranulocytosis and lupus) are metabolized to reactive metabolites by

activated leukocytes. Similarly, we propose that the oxidation of DMBI to the

79 iminoquinone reactive intermediate by activated neutrophils is responsible for the

hypersensitivi ty reactions to indomethacin (i.e. agranulocytosis). There is evidence to

suggest that other iminoquinones generated by the MPO system or activated neutrophils

may cause hypersensitivity reactions (Maggs el al., 1987; Miyarnoto et al., 1997: Parrish et

al., 1997; Smith et al., 1989; Uetrecht et al.. 1994).

In summary, DMBI. a major metabolite of indomethacin in the Iiver, is metabolized

by the MPO enzyme system in activated neurrophils, forming a reactive iminoquinone

intermediate. We propose that the iminoquinone intermediate is responsibie for some of the

idiosyncratic reactions associated with indomethacin, especially agranulocytosis.

Chapter 3

Detection of 2-hydroxyiminostilbene in the urine of patients

taking carbarnazepine and its oxidation to a reactive

iminoquinone intermediate.

3.1 Abstract

Carbamazepine is one of the rnost widely used anticonvulsants in Nonh America;

however. its use is associated with a range of serious idiosyncratic adverse reactions.

These reactions are thought to result from the formation of chernically reactive metabolites.

Carbamazepine is extensively metabolized in the liver and one of the major metabolites is 2-

hydroxycarbamazepine. whic h has previously been detected as a urinary me tabolite

excreted by rats and hurnans along with its further metabolized product, 2-

hydroxyiminostilbene.

In this study, we found that the urine of patients taking carbarnazepine appeared to

contain more of the glucuronide of 2-hydroxyiminostilbene than that of 2-

hydroxycarbamazepine. We have also demonstrated that 2-hydroxyiminos til bene can be

oxidized readily to an irninoquinone species by HOCI, H202 or even on exposure to air.

The reactivity of this irninoquinone as an electrophile was studied. It was shown to react

with sulfhydryl-containing nucleophiles, such as GSH and NXC. We also found a

metabolite with the same molecular weight as 4-rnethylthio-2-hydroxyiminostilbene. but

not the correspondin_o cnrbamazepine derivative. in the urine of patients taking

carbnmazepine and this presurnrnably reflects the formation of a GSH conjugate of the

reactive irninoquinone i ~ i i*ir*o. This irninoquinone intermediate may play a role in

carbarnazepine-induced idiosyncratic reactions.

3.2 Introduction

Carbamazepine (jH-dibenzob,flazepine-5-carboxamide) is an effective dmg in the

trerttment of convu1sii.e disorders (Cereghino et al., 1974; Cereghino et cil.. 1973)

However. carbarnazepine-induced adverse reactions have been reponed in as many as one-

third to one-half of a11 patients treated with this drug (Durelli et rd. , 1989; Gram and

82 Jensen. 1989; Pellock, 1987). Arnong these adverse reactions. 5% of them can be

classified as idiosyncratic reactions (Askmark and Wiholrn, 1990). These idiosyncratic

adverse reactions include skin rash (Criil. 1973), blood disorders (Gerson et al., 1983) and

hepatitis (Horowitz et al.. 1988). ..\ Swedish survey of 505 reports of 7 13 idiosyncratic

reactions to carbamazepine from 1965 to 1987 reported skin reactions (48%),

hematological (12%) and hepatic disorders (10%) to be the most frequent. (Askmark and

Wiholm, 1990) Although the mechanism of carbarnazepine-induced adverse reactions is

not clear, they are thought to result from the formation of chemically reactive metabolites

(Riley et al., 1989; Shear et al.. 1988). An arene oxide intermediate has been postulated to

be responsible for the idiosyncratic toxicity of carbamazepine. This hypothesis is mostly

based on the observation that the metabolism-dependent cytotoxicity of carbamazepine in

vitro can be enhanced by an epoxide hydrolase inhibitor, TCPO (Riley et al., 1989).

However, this evidence is complicated by the fact that TCPO is also known to deplete

glutathione (Pessayre et al.. 1979) and inhibit cytochrome P350 (Ivanetich et al., 1982).

Funbermore, since the arene oxide is chernically reactive, it may not reach targets, such as

the skin and bone marrow, distanr from the liver in sufficient concentrations to cause

toxicities. It appears that most drugs associated with bone marrow toxicity are metabolized

to reactive metabolites by myeloperoxidase or other enzymes present in the bone marrow.

such as alcohol dehydrogenase (Uetrecht, 1996). These considerations led us to search for

alternative bioactivation pathways of carbarnazepine.

Carbarnazepine is extensively metabolized and more than 30 metabolites have been

identified in urine of patients taking the drug (Lertratanangkoon and Horning, 1982). The

major pathways involve N-glucuronidation on the carbamoyl side-chain of carbarnazepine,

formation of carbarnazepine-l0,l 1-epoxide and hydroxylation on the aromatic rings. One

of the major metabolites is 2-hydroxycarbamazepine (Eichelbaum el al., 1984) , and it can

be funher rnetabolized to 2-hydroxyiminostilbene, either in the liver or in other tissues.

Both 2-hydroxycarbamazepine and 2-hydroxyiminostilbene have been detected as

olucuronide conjugates in the urine of rats and humans (Lertratanangkoon and Horning, Li

1982). Ive hypothesized that Mydroxy irninostilbene can be funher oxidized to a reactive

iminoquinone intemediate (Figure 3- 1). which may be responsible for carbarnazepine-

induced idiosyncratic reactions.

Hydrolysis

fyJ=Je-, % / f-J==&OH \ /

Oxidation I Iminoquinone

H 2- hydroxyiminostilbene

Fisure 3- 1 : Proposed bioactivation pathway of carbarnazepine which leads to the

formation of a reactive iminoquinone intermediate.

3.3 Materials and Methods

3.3.1 Synthesis of the iminoquinone intermediate

The iminoquinone was synthesized by oxidation of iminostilbene using a

modification of the method of Islam e t al. (Islam and Skibo, 1991). Fremy's salt

((KS03)2NO, 2.6 g, Aldrich Chernical Co., Milwaukee, WI) was dissolved in 360 ml of

phosphate buffer (pH 7.1). To this solution, 1 g of iminostilbene (Aldrich), which was

dissolved in 260 ml of acetone. was added. The mixture was stirred at room temperature

for 5 min, and the resulting red solution was extracted with 3 x 200 ml portions of

chloroform. The combined extracts were washed with 3 x 100 ml portions of water and

then dried over magnesium sulfate. The solvent was evaporated by a rotary evaporator.

The iminoquinone was then purified by flash chromatography using hexane / ethyl acetate

(7:3, v/v) as the eluting solvent. The product was a red solid with a yield of 23% [rn.p. =

135-136°C; MS: d z 208 (MH+) and 180 [(MH-CO)+]; 'H-NMR (chloroform-d): 6 7.94

pprn (Hg or Hg. dl J = 7 . 8 1 Hz); 8 7.62 pprn (H7 or Hg, dd. J = 7.57 Hz; 1.63 Hz): 6

7.59 pprn (HJ, d, J = 9.77 Hz); 6 7.54 pprn (Hg or H7, dd, J = 7.57 Hz; 1.47 Hz); 6 7.48

pprn (H9 or &j, d. 1 = 7.51 Hz); 6 6.95 pprn (H3, dd, J = 9.76 Hz: 2.70 Hz); 6 6.96 pprn

(Hi0 or H l i , d, I = 11.00 Hz); 6 6.78 pprn (Hl1 or Hia, d. J = 11.48 Hz); 8 6.58 pprn

(Hi . ci. J = 2.14 Hz)]. This irninoquinone species has been synthesized and the chernical

analyses have been reponed: m.p. 135-136°C: the NMR spectrum consisted only of a

series of multiplets between r 1.8 and 3.6; the base peak in MS war the M-CO ion (Haque

and Proctor, 1968).

3.3.2 Synthesis of 2-Hydroxyiminostilbene

2-Hydroxyiminostilbene was synthesized by reduction of the irninoquinone using

the method of Chang (1983). A chloroform solution containing 40 mg of iminoquinone

was extracted with freshly prepared sodium hydrosulfite solution (in excess. Aldrich) until

the color of the organic layer changed from red to yellow. The chloroforrn layer was

washed with water and dried over magnesium sulfate. The solvent was then removed by a

rotary evaporator to yield 36 mg (89% yield) of 2-hydroxyiminostilbene as a yellow solid.

1rn.p. = 225-226OC; MS: mlz 210 (MH+); 'H NMR(chlorofom-d): 6 7.05 pprn (H7 or

H8, dd, J = 7.50 Hz; 1.41 Hz); 6 6.90 pprn (Hg or Hg, d, J = 7.57 Hz); 6 6.85 pprn (Hg

or H7, dd. J = 7.45 Hz: 1.60 Hz); 6 6.55 pprn (H3, dd, J = 8.3 Hz; 2.93 Hz); 6 6.54 pprn

(H9 or Hg, d. J = 7.81 Hz): 6 6.41 pprn (H4, d, J = 8.3 Hz); 6 6.42 pprn (Hi0 or H i 1 , d,

J = 12.00 Hz); 6 6.41 pprn ( H l , d, J = 2.68 Hz); 6 6.32 pprn (Hi1 or Hlo, ci, J = 11.72

Hz) and two broad peaks in the range of 6 4.9 ppm to 6 4.4 pprn due to the proton on the

nitrogen and the proton of the hydroxyl groupl. The chernical analyses of 2-

hydroxyirninostilbene reported in the literature were: m.p. 225226°C; NMR (chloroform-

d/dirnethylsulfoxide-dg): 6 8.27 pprn (s, IH, -OH), 6 7.5-6.2 pprn (m, 9H. ûromatic and

olefin), 6 5.41 (s, broad. 1H, -Ni); MS: rn/e 209 (M+) (Chang, 1983).

3.3.3 Synthesis of NAC-adduct of the iminoquinone

A methanolic solution (4.8 mlj of the iminoquinone (10 mg) was added to NAC (4

g), which was dissolved in 43.2 mi of phosphate buffer (pH 8). The mixture was stirred

for 10 min at room temperature and then concentrated by a rotary evaporator. The NAC

adduct of the iminoquinone was purified by silica gel TLC using a solvent of ethyl acetate /

methanol (7:3, v / v). The TLC band (RF = 0.3) that contained the irninoquinone-NAC

adduct was scraped off the TLC plate and the adduct was extracted with ethyl acetate. The

y ield was approxirnately 33%. The mass spectrum of the iminoquinone-NAC adduct

showed a MH+ ion at m/z 37 1.

3.3.4 Synthesis of 2-Hydroxycarbamazepine

The hydroxyl group on 2-hydroxyiminostilbene was first protected using a

modification of the method of Kendall et al. (Kendall et al., 1979). Imidazole (12 mg,

0.18 rnrnol, Aldrich) was added to a solution of 2-hydroxyiminostilbene ( 18.4 mg, 0.09

m o l ) and teri.-butyldimethylchlorosilane (26.4 mg, 0.18 mmol, Aldrich) in dry N, N-

dimethylformarnide (0.1 d). The mixture was stirred for 3 hr, followed by dilution with

5% aqueous sodium bicarbonate (0.5 ml). The reaction mixture was then extracted four

tirnes with hexane / dichloromethane (9 : 1, vhr) . The combined extracts were washed with

water and dried, the solvent was evaporated by a rotary evaporator.

Without further purification, the silylated 2-hydroxyiminostilbene (27 mg, 0.083

mmol) was reacted with 4-nitrophenylchloroformate (50 mg, 0.25 rnmol, Aldrich) in

chloroform ( 1 ml). Triethylamine (36 41, 0.076 rnrnol) was added at the beginning of the

reaction to neutralize the hydrochloric acid generated from the reaction. The mixture was

stirred for 3 days and water (1 ml) was then added to quench rhe excess 1-

nitrophenylchloroformate. The product was esuacted with chloroform three times and the

solvent was evaporated by a rotary evaporator. The product was dissolved in acetonitrile

(1 ml) and the ammonia solution (in excess) was added. The mixture was stirred for 3

days before it was neutralized by concentrûted hydrochloric acid. The product was

extracted into ethyl acetate (3 x 10 ml) and the solvent evaporated. 2-

hydroxycarbamazepine was purified by TLC using ethyl acetate : chloroform (8 : 2 , v I v)

as the eluring solvent. The product was a white solid and the yield was approximately 20%

[rn.p. = 231-237OC; RF = 0.35; MS: m/z 213 ()lHf) and 210 [(MH-HNCO)+]; 'H-XMR

(chloroform-d): 8 7.47 pprn (Hg or Hg. d, J = 7.08 Hz); 6 7.43 pprn (H7 or Hg, dd. J =

6.83 Hz; 1.71 Hz); 6 7.38 pprn (Hg or Hg, d. J = 7.08 Hz); 6 7.34 ppm (Hg or H7 , dd. J

= 6.84 Hz; 1.56 Hz); 6 7.32 ppm (H4, d, J = 8.54 Hz); 6 6.93 ppm (Hi0 or H i 1 , d, J =

11.72 Hz); 6 6.84 pprn (H3, d. J = 8.06 Hz); 6 6.83 ppm (Hl 1 or Hl0 , d, J = 11.32 Hz):

6 6.76 pprn (Hi , s): protons on the amide nitrogen, 6 4.5 ppm (2H,s) and a broad peak at

6 5.4 pprn due to the proton on the hydroxyl groupl. The chemical analyses of 2-

hydroxycarbamazepine reported in the literature were: m.p. 239-242°C; NMR

(dimethylsulfoxide-dg): 6 8.25 ppm (s, lH, -OH), 6 7.5-6.6 (m. 9H, aromatic and olefin),

8 5.45 (s, 2H, -CONH): MS: rn/e 252 (M+), 209 [(M+)-HNCO] (Chang, 1983).

3.3.5 Treatment of the urine sample

Random urine sarnples were obtained from two male patients (ages, 93 and 53,

respectively) who had received carbamazepine (250 and 200 mg/day, respectiveiy) for

more than 1 year. The urine sample (-100 ml) was concentrated to about 5 ml by

lyophilization. The concentrated urine sarnple was acidified to pH 4.5 and purified by C-

18 Prep-Sep solid phase extraction column (Fisher Scientific, Unionville, ON). The

sample was loaded ont0 the column followed by washing with 2 x IO-ml portions of

water. The desired urinary metabolites were then eluted from the column by 7 x IO-ml

portions of methanol. The methanol effluents were cornbined and concentrated to

approximately 1 ml. Aliquots (10 pl) were analyzed by LC/MS. Enzymatic hydrolysis

was carried out by incubating the urine samples with P-glucuronidase (100 U, Sigma) in

0.2 ml of pH 5.1 buffer for 20 hr at 37OC. The carbamazepine metabolites were then

extracted with ethyl acetate and the solvent was evaporated under a Stream of N2. The final

products were dissolved in methanol and aliquots of 10 pl were analyzed by LC/MS.

3.3.6 Analytical ~Methods

The analyses of carbarnazepine urinary metabolites were performed by interfacing

an Ultracarb C-18 colurnn (2- x 100-mm, Phenomenex, Torrance, CA) to a Sciex API III

mass spectrometer (Perkin-Elmer Sciex, Thomhill, Ontario, Canada). Aliquots of a urine

sample (IO pl) were eluted at a flow rate of 0.2 d m i n , with a solvent consisted of water,

acetonitrile and acetic acid (49:50: 1, v/v) containing 1 rnM ammonium acetate. When the

metabolites with s h o ~ e r retention times (e.g. the glucuronide conjugate of 4-methylthio-2-

hydroxyiminostilbene) were analyzed, the eluting solvent consisted of water, acetonitrile

and acetic acid (74% 1, vlv) containing 1 rnM ammonium acetate. When P-glucuronide

conjugates were analyzed. the eluting solvent consisted of water, acetonitrile and acetic acid

(84: 15: 1, v/v) includin; 1 mM ammonium acetate. A splitter was used to decrease the flow

to the LC/MS interface at approximately 10 pumin. The collisional activation spectra were

obtained by using the LCMSMS mode, with argon as the target gas.

~ H N M R spectra were recorded at 500 MHz with a Bmcker AM 500 spectrometer

(Bmcker Canada. Milton. Ontario, Canada). Spectra were obtained in chloroform-d except

that for the iminoquinone-NAC ndduct, which was obtained in deuterium oxide (DzO).

3.4 Results

3.4.1 Oxidation of 2-hydroxyiminostilbene

2-hydroxyiminostilbene was readily oxidized to the irninoquinone intermediate by

HOCI. When 2-hydroxyiminostilbene was reacted with HOCl at equal concentrations, the

color of the reaction mixture changed instantaneously from yellow to red, and a new Peak

due to the irninoquinone was obsewed by HPLC. Figure 3-2 shows the UV absorption

specua of 2-hydroxyiminosiilbene and that of the reaction mixture after HOCl was added.

The UV absorption spectrum of the oxidation product was similar to that of the

irninoquinone. H 2 0 2 was also able to oxidize 2-hydroxyiminostilbene to the irninoquinone

species; however, at a much slower rate than did HOCI. It was also observed thai 2-

hydroxyiminostilbene underwent autoxidation to the iminoquinone intermediate by air at

room temperature. The half-life of 2-hydroxyiminostilbene was approxirnately 2 hr in

phosphate buffer (pH 7.4).

200 300 400 500

Wavelength (nm)

Fioure - 3-2: The UV absorption spectra of the synthesized irninoquinone (dashed

line) and 2-hydroxyiminostilbene before (O sec) and after (10 sec) reaction with HOCI.

The concenuation of both 2-hydroxyiminostilbene and HOC1 were O. 1 mM.

3.4.2 Reactivity of the iminoquinone intermediate

The reactivity of the iminoquinone as an electrophile was studied. It was found to

react with sulfhydryl-containing nucieophiles, such as GSH and NAC, to f o n conjugates.

The iminoquinone-NAC conjugate was isolated and anaiyzed by NMR (Figure 3-3). The

'H NMR (D20) consisted of protons as follows: 6 7.17 pprn (H7 or Hg, dd, J = 7.9

Hz; 1.65 Hz); 6 6.99 pprn (Hg or Hg, d, J = 7.48 Hz); 6 6.97 ppm (H8 or H7, dd, J = 7.80

Hz; 1.70 Hz); 6 6.9 1 pprn (H3, d. J = 2.77Hz); 6 6.89 pprn (Hg or Hg, d, J = 7.56 Hz); 6

6.5 1 ppm (HIo or Hi 1, d, J= 1 1.72 Hz); 6 6.44 pprn (H,, d, J = 2.78 Hz); 6 6.40 pprn

(HI or H 10, d, J=11.75 Hz); NAC-CH, 6 4.28 pprn ( IH, dd, J = 7.05 Hz, 3.66 Hz);

NAC-CHzct, 6 3.34 pprn ( 1 H, dd, J = 14.1 1 Hz, 3.84 Hz); NAC-CH$, 6 3.13 pprn (1H.

dd. J = 14.10 Hz, 7.05 Hz); NAC-CH3,6 1.67 pprn (3H, s). Compared with the NMR

spectrum of 2-hydroxyiminostilbene, the proton signal on C4 disappeared in the

iminoquinone-NAC conjugate and the proton signal on C3 changed from a doublet of

doublets to a doublet. This suggested that the sulfur of NAC was substituted in the 4-

position on the aromatic ring (Figure 3-1). This assignment was also confirmed with the

findings from the 1H-lH and ' H - 1 3 ~ correlation expenments (Figure 3-5).

10 11

HOCI,H2O2 or O2 t

4

. . GSH or NAC

2- hydroxyiminostilbene Iminoquinone

NAC

Figure 3 4 : Oxidation of 3-hydroxyiminostilbene to the iminoquinone intermediate

and its reaction with NAC.

3.4.3 Analysis of urine samples from patients

When the urine sample was analyzed by LC/MS. neither the GSH nor the NAC

conjugate of irninoquinone was detected. However, a metabolite with MH+ ion at m/z 432

(RI = 5.7 min) was detected. This metabolite has the same molecular weight as the

glucuronide conjugate of 4-methylthio-2-hydroxyirninostilbene, which is a probable further

metabolite of the irninoquinone-GSH conjugate. The MSMS of the metabolite with MH+

ion at m/z 432 showed a fragment ion at m/z 256, corresponding to the loss of

dehydroglucuronic acid moiety ([M+I-176]+) (Figure 3-6). When the urine sample was

hydrolyzed by P-glucuronidase, a peak with a MHf ion at rnlz 254 was detected by

LC/MS. The molecular weight of this species was 2 mass units less than that of 4-

methylthio-3-hydroxyirninostilbene suggesting that it could be a further oxidized product of

4-methyl thio-3-hydroxyiminostilbene.

A glucuronide conjugate of 2-hydroxyiminostilbene was also detected, which

showed a MHf ion at m/z 386. Figure 3-7 shows the cornparison of the total ion curent

(TIC) of the 2-hydroxyiminostilbene glucuronide conjugate with that of other metabolites

excreted in the urine including the glucuronide conjugate of carbamazepine (with a MH+

ion at m/z 4 13). the glucuronide conjugates of monohydroxylated carbarnazepines. as well

as the N-glucuronide of carbamazepine 10.1 1-epoxide (with MH+ ions at m/z 429), and

the 10,l I-dihydrodiol-carbamazepine (with a MH+ ion at rn/z 27 1). The ratios of the TIC

of 2-hydroxyiminosrilbene glucuronide to that of other urinary metabolites are shown in

Table 3- 1.

MS/MS spectrum of 2-hydroxyiminostilbene glucuronide conjugate dernonstrated

that it undenvent charactenstic fragmentation with the major fra,gnent ion at rn/z 210 which

represents the loss of the dehydroglucuronic acid moiety ([M+ 1- 1 XI+) (Figure 3-8).

hterestingly, the irninoquinone intermediate was also observed by LC/MS and it showed

the same retention rime (Rt = 8.4 min) and MS/MS fragmentation pattern as that of the

97 synthesized iminoquinone. When NAC was added to the sample, the peak due to the

irninoquinone disappeared, suggesting that it reacted with sulfhydryl-containing

nucleophiles. When the urine samples were hydrolyzed by P-glucuronidase. the peak of

the irninoquinone increased dramatically (Figure 3-9). In addition, when the glucuronide

conjugates of rnonohydroxylated carbarnazepines were hydrolyzed. a peak due to 2-

hydroxycarbamazepine was observed, which has the same retention rime and MSlMS

specuum as that of the synthesized 2-hydroxycarbamazepine.

Figure 3-6: MS/MS spectrum of the metabolite, which has the same molecular

weight as that of 4-methylthio-2-hydroxyiminostilbene. in its glucuronide form (with a

.MH+ ion at d z 432).

0.0 10.0 20.0 30.0 40.0 50.0

Time (min)

Figure 3-7: Selected ion current of the urinary metabolites from the LC/MS of a

patient's urine: 7-hydroxyiminostilbene-glucuronide (A), carbarnazepine-glucuronide (B),

monohydroxycarbarnazepine-glucuronide (C) and 10.1 1 -dihydrodiol-carbarnazepine (D).

The ion current for each ion is provided in the upper right corner of each trace.

Figure 3-8: MS/MS spectrum of the glucuronide conjugate of 2-

hydroxyiminostilbene.

Time (min)

Figure 3-9: LCMS ion current at m/z 208 corresponding to the irninoquinone

intermediate in a urine sample before (A) and after (B) hydrolysis by P-glucuronidase.

Table 3-1: The ntios of the TIC of 2-hydroxyiminostilbene-glucuronide (m/z 386) to that of carbarnazepine-glucuronide (rn/z 4 13); rnonohydroxyc arbamazepine-gl ucuronide (m/z 429) and 10,ll- dihydrodiol-carbarnazepine (m/z 27 1).

Patient #1

Patient #2

miz 386

* The third urine sample was collected from a patient in chronic use of carbarnazepine. However. derailed information about this patient was not obtained.

3.5 Discussion

The mechanism(s) of the idiosyncratic reactions associated with carbamazepine

therapy are poorly understood. It has been postulated that bioactivation to a reactive arene

oxide metabolite is a prerequisite to toxicity (Pimohamed et al.. 1992; Shear et al., 1988)

and a risk factor for the adverse reactions is a deficiency of epoxide hydrolase (Spielberg et

al., 198 1) However, two good studies have failed to find a consistent mutation, or pattern

of mutations, in the rnicrosomal epoxide hydrolase gene that is common in patients with a

history of carbamazepine hypersensitivity reactions (Gaedigk et al., 1994: Green et c d . ,

1993). Carbarnazepine has been shown to be bioactivated to a protein-reactive metabolite

by human liver microsomes. however the covalent binding of 14C-carbarnazepine to human

liver microsomes was relatively low (Pirmohamed et al., 1992). The formation of the

arene oxide metabolite has only been inferred from the urinary excretion of unquantified

phenols and catechols of carbarnazepine (Lertratmangkoon and Horning, 1982; Regnaud et

al., 1988). However, the aromatic hydroxyl metabolites are not necessady derived from

an arene oxide intermediate (Hanzlik et al., 1984). A study by Lillibridge et al. ( 1996)

using mouse liver microsomes has shown that a quinone intermediate may also be forined

in carbarnazepine bioactivation. Furtherrnore, i t has been demonstrated thnt for

bromobenzene (which does form an arene oside) most of the covalent binding is due to a

quinone (Hanzlik el al., 1984; Rietjens et oz.. 1997). Although the detection of GSH

conjugate of the postulated arene oxide intemediate in rodrnts has been reported (Amore et

al.. 1997: .Madden et ni., 1996), in humans. the same GSH-conjugate was not detected nor

were any of its further metabolized products, such as the NAC or thiomethyl derivatives

(Mqgs er al., 1997). In addition, due to their reactivity, arene oxides are short lived and

probably would not be able to reach sites (e.g. the bone marrow or skin) distant from the

major site of production, i.e. liver, at sufficient concentrations to cause toxicity. Therefore,

104 the objective of this work was to investigate alternative bioactivation pathways where

reactive metabolites might be generated at the targets of toxicity.

One of the major routes of carbarnazepine metabolisrn in the liver is hydroxylation

of the aromatic ring to yield one of the major carbarnazepine metabolites. 2-

hydroxycarbamazepine, which c m be hirther hydrolyzed to 2-hydroxyiminostilbene, either

in the liver or in other tissues (Lertratanangkoon and Horning, 1982). In this study, we

were able to detect the glucuronide conjugate of the 2-hydroxyiminostilbene by LC/MS.

Two isomers of hydroxyirninostilbene (2-OH and 3-OH-iminostilbene) had been detected

in human urine during an earlier study in which they were characterized as their

trimethylsilyl (TMS) derivatives after enzymatic hydrolysis of the urine sample with

Glusulase (Lertratanangkoon and Horning, 1982). One of the TMS derivatives of the

hydroxyirninostilbene isomers was observed to have the same GC/MS properties as the

TMS derivative of 2-hydroxyirninostilbene (Lertratanangkoon and Horning, 1982).

Although the arnount of the metabolite was too small for us to identify the structure by

NMR, the fact that we also observed a peak due to the iminoquinone in the same sample

suggested that 2-hydroxyiminostilbene glucuronide conjugate must be excreted in the urine

because other hydroxyiminostilbene isomers would not generate the iminoquinone

intermediate upon oxidation. In addition, when the sample was hydrolyzed by P-

glucuronidase to liberate 2-hydroxyiminostiibene, a much larger peak due to the

iminoquinone was observed instead of a peak due to 2-hydroxyiminostilbene (Figure 3-8).

This was not unexpected since we had shown that 2-hydroxyiminostilbene was readily

oxidized to the iminoquinone by HOCl, H102 or even by air. The free 3-

hydroxyiminostilbene released from enzymatic hydrolysis of the glucuronide conjugate was

presurnably oxidized to iminoquinone readily, before it was detected by LC/MS. We also

detected a unnary metabolite with the same molecular weight as the glucuronide conjugate

of 4-methylthio-7-hydroxyiminostilbene. However, we did not obtain enough material to

identim its structure.

105 In this work, we also demonstrated that the iminoquinone reacted with sulfhydryl-

containing nucleophiles. such as GSH and NAC, to form conjugates. The NMR data

confirmed that the sulfur was substituted on the aromatic ring in the meta-position to the

hydroxyl group. This resuit suggested that the iminoquinone is an reactive electrophile.

and it may bind to macromolecules (i.e. proteins) in vivo to cause direct toxicity or act as a

hapten to modulate the immune system. The iminoquinone intermediate could also undergo

redox cycling and generate reactive oxygen species, since it is an oxidizing agent.

In summary, we showed that the P-glucuronide conjugate of 2-

hydroxyiminostilbene was excreted in the unne of patienis taking carbamazepine. We have

also demonstrated that 2-hydroxyiminostilbene was readily oxidized to a reactive

iminoquinone intermediate that can be trapped by GSH and NAC. h i vivo, 2-

hydroxyirninostilbene could be readily oxidized in target organs, such as bone marrow and

skin. by peroxidases or even by oxygen. The ease of oxidation of 2-hydroxyiminostilbene

makes the iminoquinone an attractive candidate for the reactive metabolite of carbamazepine

responsibie for idiosyncratic reactions in the bone marrow and skin.

Chapter 4

Investigating the role of the reactive iminoquinone metabolite

of carbarnazepine in the mechanism of carbarnazepine-induced

idiosyncratic reactions

4.1 Abstract

We have previously shown that 7-hydroxyirninostilbene, a major metabolite of

carbamazepine. can undergo autoxidation to a reactive iminoquinone intermediate. The

iminoquinone intermediate reacrs with sulfhydryl-containing nucleophiles such as GSH and

NAC. It is thought that this iminoquinone intermediate may play a role in carbamazepine-

induced idiosyncratic reactions. Immunoblotting studies with anti-iminoquinone-NAC-

KLH antibodies detected protein binding in the bone marrow of rats treated with

carbamazepine for 4 days. The major modified polypeptides had molecular masses of 1 15

and 133 KDa. The recognition of these polypeptides by the antiserum could be inhibited

by preincubation of the antiserum with 2-hydroxyiminostilbene (1OpM) but not by the

sarne concentration of carbarnazepine or irninostilbene. Protein oxidation possibly caused

by reactive oxygen species generated by redox cycling of the irninoquinone intermediate

was also investigated. Increased protein oaidation was found in the nuclear and cytosol

fractions of the liver from rats treated with carbamazepine comparing to that of the controls.

Several polypeptides with rnolecular masses of 43. 3 1, 30 and 28 KDa were found to be

oxidized in these two subcellular liver fractions isolated from carbamazepine trented rats but

not frorn control rats.

1.2 Introduction

Idiosyncratic drug reactions are frequently life threatening, and they are almost

totally unpredictable. Although the mechanism of idiosyncratic drug reactions are mostly

unknown. there is a large amount of circumsrantial evidence to suggest that reactive

metabolites are involved (Hinson and Roberts. 1992; Park et al., 1995) and some of the

clinical characteristics also suggrsted the invol~ement of the immune system (Pohl et al.,

1988). One proposed mechanism for idiosyncratic dnig reactions is based on the hapten

108 hypothesis. The hapten hypothesis postulates that the drug, or a metabolite. becomes

covalently bound to a macromolecular carrier, such as a protein. and that the resulting dmg-

protein conjugate can be recognized as an irnmunogen (Park et al., 1987). Some drugs,

such as penicillin, captopril and penicillarnine, are chemically reactive and react directly

with proteins. Howevcr, most drugs that cause hypersensitivity rractions are not

intrinsically reactive and it is, therefore, assurned that bioactivation must occur ~ I I vivo.

There is now good evidence that bioactivation of dmgs is involved in immunogen

formation (Park and Kitteringharn, 1990).

Acetaldehyde, the prim;iry metabolite of ethanoi. was shown to bind covalently to

microsomai proteins, and the binding was shown to increase after chronic alcohol

consumption (Nomura and Lieber, 1981). It was found that, in man, chronic abuse is

associated with the appearance of an increased titer of antibodies against acetaldehyde

protein conjugates (Hoerner e t al., 1986). The specificity of antibodies against

acetaldehyde adducts produced in vitro suggested that in vivo sensitization occurs as a

result of acetaidehyde binding to proteins.

Amodiaquine has been associated with two major side effects, rigranulocytosis and

hepatitis (Hatton et ul., 1986; Larrey et al., 1986; Neftel et cil., 1986). S tudies have shown

that amodiaquine is readily oxidized to an electrophilic quinone irnine intermediate and that

both amodiaquine and the quinone imine metabolite are irnrnunogenic in the rat (Clarke et

al., 1990; Maggs e t ai., 1987; Maggs et ai., 1988). The antibodies raised recognize a

common antigenic deterrninant. The clinical relevance of these results obtained in animal

studies is shown by the detection of specific anti-amodiaquine IgG antibodies in man (Park

and Kitteringham. 1990). The specificity of the antibody binding again indicated an

important role of the quinone imine metabolite in dmg sensitization.

Tienilic acid is another example of a drug which requires bioactivation to induce

idiosyncratic hepatotoxic effects. It was shown that reactive metabolite forrned by

oxidation of 1%-tienilic acid by human liver rnicrosomes were irreversibly bound to

109

cytochrome P450-8 (now known as CYP2C9), the liver enzyme responsible for

bioactivation of the drug (Dansette et al., 1989). Sera of patients with tienilic acid-induced

hepatitis contained antibodies which specifically recognized a protein of human liver

microsomes determined to CYP2C9.

Halothane, a widely used anesthetic, is known to cause idiosyncratic liver damage.

Halothane is oxidized in the liver by CYP450 enzymes to a reactive trifluoroacety!chloride

which forms the neoantigen on microsorniil proteins. Immunochernical studies have shown

that sera from the majoriiy of halothane hepatitis patients contain antibodies of the IgG

isotype that recognize trifluoroacetyl modified proteins expressed in the liver of animals or

humans treated with halothane (Pohl et al., 1988). This is the only case where the identity

of the bound metabolite and characterization of the endogenous carrier molecules have been

defined.

In the sera of nine patients with carbarnazepine-induced hypersensitivity syndrome,

autoantibodies (IgG) were found to recognize a 53-Kd protein in rat liver microsomes

(Leeder et OZ., 1992; Riley er al., 1993). No such reactivity was observed in sera from

controls. The protein target was identified as CYP3Al; however, these antibodies did not

recognize related human CYP3A proteins despite the high degree of structural similarity

(Leeder et al., 1996).

We have shown that 2-hydroxyirninostilbene, a major metabolite of carbarnazepine

found in patient's urine, can undergo autoxidation to a reactive irninoquinone intermediate.

Funhermore, the iminoquinone intemediate reacts wit h sulfhydryl-containing nucleophiles

such as GSH and NAC. As discussed earlier, covalent modification of critical endogenous

proteins by reactive metabolites is thought to be important in drug-induced idiosyncratic

toxicity. To further investigate the role of this reactive iminoquinone metabolite in

carbarnazepine-induced idiosyncratic reactions, we raised an antibody against the

iminoquinone epitope and used it to investigate covalent binding of the iminoquinone

intermediate to proteins hi vivo . We demonstrated that 2-hydroxyiminostilbene can be

110 readily oxidized to the iminoquinone intermediate and which not only reacted with GSH or

SAC to form a GSH or NAC conjugate, but mostly was reduced to form 2-

hydroxyirninostilbene. This result suggested that redox cycling may occur in vivo and

result in oxygen activation and possible protein oxidation. Protein oxidation in vivo was

therefore investigated.

4.3 mate rials and Methods

1.3.1 Synthesis of iminoquinone-NAC-conjugated KLH and RSA

The iminoquinone-NAC conjugare (15 mg) was synthesized as described in

Chapter 3 and dissolved in 10 ml N,N-dimethylformarnide which contained

hydroxysuccinamide (230 mg, Sigma). This solution was then slowly added to a 25 ml

dichloromethane solution of 1 -ethyl-3-(3-dimethy1aminopropyi)carbodiimide (EDC, 3 10

mg, Sigma). Dichlorornethane was evaporated under a sueam of N:! after the reaction was

stined for 0.5 hr. The reaction mixture was then added dropwise with stirring to either

Ksyhole limpet hemocyanin (ELH. 9 mg, Sigma) or to rabbit semm albumin (RSA, 9 mg,

Sigma) dissolved in 5 ml phosphate buffer (pH 7.4). The reactions were cmi rd out by

srimng at room temperature for 3 hr. The protein conjugates were dialyzed cxtensively

qainst warer, and then lyophilized.

4.3.2 Production of anti-iminoquinone-NAC-KLH antiserum

Polyclonal anti-irninoquinone-NAC-KLH antibodies were raised in 2.0-2.5 kg

male, pathogen-free New Zealand white rabbits (Charles River, Quebec, Canada). A pre-

immune serum sample was obtained, and the rabbits were immunized with 500 pg of the

iminoquinone-NAC-KLH conjugate emulsified with an equal volume of Freund's complete

111

adjuvant at multiple subcutaneous sites. Injections with 500 pg antigen ernulsified with

Freund's incomplete adjuvant divided into 6-8 subcutaneous sites were repeated 2.4, and 6

weeks after the initial immunization. Exsanguinarion under pentobarbital anaesthesia was

conducted 2 weeks after the final injection and the blood was allowed to clot overnight at

4°C and then centrifuged at 400 g. The serum was recovered and heat-inactivated at 56°C

for 30 min before being aliquoted and stored at -20°C.

43.3 Activity of the anti-iminoquinone-N.4C-KLH antiserum

The activity of the anti-iminoquinone-NAC-KLH antibodies against the

iminoquinone epitope was determined by an enzyme-linked immunosorbent assay

(ELISA). Lminoquinone-NAC-RS A or RSA ( 1ûû u1 of a 15 pglml solution) was incubated

in flat-bottom 96-well plates (Costar, Cambridge. MA) at 4°C overnight. The plates were

then emptied and washed four times with ELISA wash buffer [10 mM Tris-HCl, pH 7.5,

153 rnM NaCI, 0.5% (w/v) casein, and 0.03% (w/v) thimerosal]. After the last wash, the

plates were tapped dry. Various dilutions (11100 to 1/100,000) of the anti-iminoquinone-

SAC-KLH antiserum (100 pl in PBS) were added to the plates and incubated at room

temperature for 3 hr. Plates were again washed four times with ELISA wash buffer and

tapped dry. Alkaline phosphate-conjugated goat anti-rabbit IgG (diluted 15000 in PBS)

was added to each well (100 $I/well) and incubated at room temperature for 2 hr. Plates

subsequently were washed four times with ELISA wash buffer and twice with PBS. A

stock solution of methyl umbelliferyl phosphate (10 mg/& in dimethylsulfoxide, kept at

-20°C) was diluted 1 : 100 in PBS, and this solution was added to each well of the plates

( 1 v e l l ) . Plates were incubated at room temperature for 10 min before fluorescence

was measured with a Fluorescence Concentration Xndyzer (Pandex, Mundelein, IL) set at

365/450 nm (excitation/emission).

4.3.4 Dosing of rats with carbamazepine in vivo

Female Lewis rats (= 200 g) were obtained from Charles River and housed in

standard cages with free access to water and powdered lab chow. After one week

acclimation period. the rats were either treated daily with carbamazepine (250 mg/kg,

gavage) as a suspension in hydroxypropyl cellulose or with hydroxypropyl cellulose alone

(control). At the end of the 4-day treatment, the rats were killed by an i-p. injection of

ketamine (100 m m ) and xylazine (20 mgkg).

43.5 Isolation of rats bone rnarrow cells

Rats femurs were removed. The bone marrow was obtained by pushing a needle

through the femurs and was collected into RPMI 1640 culture medium (University of

Toronto, Media Service). The bone marrow was suspended by passing through a 1-ml

automatic pipette tip a few rimes. After washing the cells twice in RPMI 1640 and once in

Hanks' buffer (HBSS), contaminating red blood cells were lysed by incubation in

ammonium chloride (0.16 M)/Tris (17rnM; pH 7.2) buffer for 10 min. The bone marrow

cells were ihen washed twice in HBSS and counted using a hemocytometer. Bone marrow

cells were lysed in ce11 lysis buffer [10 mM Tris-HCl, pH 7.4. 1 mM EDTA, 0.2% Triton

X-1001. Protein concentration was determined using a BCX protein assay kit (Pierce.

Rockford, IL) with bovine semm albumin as the standard. The samples were diluted to

give protein concentrations of 3 mg/ml; then, an equal volume of sample buffer [8% (w/v)

SDS, 10% ( v h ) glycerol, 0.002% bromophenol blue, 125 mM Tris-HCI, pH 6.81

containing dithiothreitol(6 mghl ) . The samples were boiled at 100°C for 10 min before

analysis by SDS-PAGE.

4.3.6 Preparation of subcelluiar fractions of rat liver

Subcellular fractions of rat liver were prepared frorn freshly isolated livers by

differential centrifugation. Rat livers from control and carbarnazepine treated groups were

pooled, weighed. and minced in five volumes of sucrose buffer (0.25 M sucrose, 15 m M

Tris-HC1, O. 1 rnM EDTA. pH 6.8), then homogenized (Janke-Kunkel hornogenizer, 1000

rpm, 10 strokes), and filtered through two layers of muslin. Subcellular fractions, nuclear

(1000xg pellet, 10 min). mitochondria (10,000xg pellet, 20 min), microsomes (100,000xg

pellet, 90 min) and cytosol(100,000xg supernatant, 90 min), were prepared by sequential

centrifugation of the crude whole homogenate os described by Wade e l ni. (1997) Each

pellet was washed twice by resuspending in buffer using a Dounce homogenizer (loose

fitting pestle) followed by recentrifugation. At each step, the supematants decanted from

the "crude" pellets and those obtained from the subsequent wash steps were cornbined and

then used in the next centrifugation step to obtain the next fraction. The pelleted subcellular

fractions were resuspended in one volume of sucrose buffer, aliquoted and stored at -70°C.

4.3.7 Preparation of rat skin homogenates

A piece of dorsal s h n of rat (2x2 cm) was shaved and cut into small pieces. The

skin samples were hornogenized in sucrose buffer and then filtered through two layers of

rnuslin. Protein concentration was determined using a BCA protein assay kit.

4.3.8 SDS-PAGE and immunoblotting

Samples for analysis by SDS-PAGE were solubilized by boiling for 10 min in

sample buffer [8% (wh) SDS. 20% (vlv) glycerol, 0.002% bromophenol blue, 125 mM

114

Tris-HC1, pH 6.81 containing dithiothreitol (6 mg/ml). Protein (30 pg) was loaded on each

lane for al1 samples (i.e. bone marrow cells. skin homogenates and liver subcellular

fractions). SDS-PAGE was carried out using a minigel system (Mini-PROTEAN II;

BioRad. Mississnuga, Ontario). Stacking and resolving gels were 4% and 10% (or 12%)

acrylamide. respectively. Gels were run at 100 V until the dye front reached the bottom of

the resolving gel (= 45 min). Resolved proteins were transferred ont0 nitrocellulose using

a buffer of 15.7 mM Tris, 120 rnM glycine. pH 8.3, containing 20% (v/v) methanol, for

75 rnin at 100 V using a mini Trans-blot transfer ce11 (BioRad). The subsequent steps were

conducted at room temperature with constant shaking. Before exposure to the antiserum.

the nitrocellulose was blocked by incubation in a solution containing 10 mM Tris-HC1. pH

7.5, 153 mM NaCl, 2.5% (w/v) casein and 0.02% (wlv) thimerosal for 2 hr. The blocked

nitrocellulose was incubated overnight with anti-iminoquinone-NAC-KLH antiserum

( 1500) diluted with ELISA wash buffer [!O rnM TrisaHCI, pH 7.5, 154 mM NaCl, 0.5%

(whl) casein and 0.02% (wlv) thimerosal]. Unbound antibodies were rernoved by washing

the nitrocellulose in wash buffer (3 x 10 min). The nitrocellulose then was incubated for 2

hr wirh horseradish peroxidase-conjugated goat anti-rabbit IgG (H+L chain) antiserum

(diluted 1:10,000 in wash buffer). The nitrocelluIose was washed with ELIS.4 wash

buffer (3 x 10 rnin) followed by washing with Tris buffer [50 mM Tris-HCI. pH 7.1 and

154 m i NaCI] (3 x 5 min). The nitrocellulose sheers were incubated in Super Signal ECL

reagent for 5 min, and bound antibodies were visualized by exposinp the nitrocellulose to

ECL film under safe-light conditions.

1.3.9 Determination of oxidatively-rnodified proteins

Protein oxidation was determined utilizing the method described by Levine et al.

(1994). Protein concentrations of subcehlar rat Iiver fractions were determined and

115 diluted to 10 mg/mi. One volume of 12% sodium dodecyl sulfate (Sigma) was added to the

samples with mixing followed by adding 2 volumes of the 2,4-dinitrophenylhydrazine

solution [20 mM in 10% (v/v) trifluoroacetic acid]. The mixture was allowed to stand for 5

min at room temperature before i t was neutralized by adding 1.5 volume of 2 M Tris

containing 30% glycerol and 19% 2-mercaptoethanol. When the solution was neutralized,

the color changed from light yellow to orange. For SDS-PAGE analysis, the samples (36

mg/lane) were loaded directly ont0 the gel without heating. The subsequent steps for

irnmunoblotting were the sarne as described above. Rabbit polyclonal antibodies against

the 7.4-dinitrophenyl moiety (Sigma. diluted 1: 1000) were used to recognize oxidatively

modified proteins.

4.4 Results

4.4.1 Synthesis of the iminoquinone-NAC-KLH (RSA) conjugate

As shown in Chapter 3. the iminoquinone intermediate reacted with NAC and

formed a conjugate. This inrroduced a carboxyl group on the molecule so that it can be

coupled with amino groups on proteins. (Fig 4-1). In this case EUH was used as the

carrier protein because it is a large foreign protein and has been commonly used in

immunogen synthesis. The irninoquinone-NAC-RSA conjugate was also synthesized in

the same way and i t was used as an artificial nntigen to determine the activity of the

antiserum.

KLH or RSA

Iminoquinone-NAC-Protein conjugate

Fig 4- 1: The structure of the iminoquinone-NAC-Protein conjugate.

4.42 Determination of the activity of the antiserum

The polyclonal antibodies were raised by immunizing rabbits with the

iminoquinone-NAC-KLH conjugate. The activity of the antibodies nsainst the

irninoquinone epitope was detennined by ELISA. ELISA anaiysis (Fig 4-2) demonstrated

that the anti-iminoquinone-NAC-KLH antibodies recognized the iminoquinone-NAC-RSA

conjugate but not RSA itself. However, the misera could not detect covalent binding of

the iminoquinone intermediate to other proteins, such as reduced lysozymes.

4.4.3 Immunoblotting of bone marrow cells isolated from rats treated with

carbamazepine

Using the anti-iminoquinone-NAC-KLH antiserum, it was possible to detect

covalent binding in bone marrow cells of rats dosed with carbamazepine (250 mg/kg/day)

(Fig 4-3A). The major modified polypeptides had molecular masses of 1 15 KDa and 133

KDa. The recognition of these polypeptides by the antiserum could be inhibited by

preincubation of the antiserum with 3-hydroxyimincstiibene (10 PM). Howe~er. the same

amount of carbamazepine and iminostilbene ( I O PM) were not able to inhibit the binding

(Fig 4-3B). Surprisingly, carbamazepine not only did not inhibit the bindin: bu t also

increased the background signal.

4.4.4 Protein oxidation in subcellular fractions of rat liver

Female Lewis rats were rreated with carbamazepine (250 mg/kg/dayl and protein

oxidation in various subcellular fractions of rats liver was studied. An irnrnunochemical

assay (Keller er al., 1993; Levine et al.. 1994) was used to detect carbonyi moieties

(ddehydes and ketones) that result from oxidative darnage to proteins. The carbonyls were

119 reacted with 2.4-dinitrophenylhydrazine, givin; the corresponding hydrazones, which

were detected by Western blotting using a polyclonal anti-dinitrophenyl antibodies. In

nuclear and cytosoi fractions. an increase in protein aldehydes was found in rats treated

with carbamazepine comparing to that in the controls (Fig 4-4). Furthermore, in cytosol

and nuclear fractions, oxidized polypeptides with molecular masses of 43, 3 1, 30 and 28

KDa were detected in the samples from rais treated with carbamazepine but not in the

samples from control rats (Fig 4-4).

4.4.5 Irnrnunoblotting of skin homogenate and subcellular liver fractions

from rats treated with carbamazepine

Proteins of skin homogenates and subcellular liver fractions were separated by

SDS-PAGE. Unfortunately, no covalent binding of the iminoquinone to skin and liver

proteins was detected using the anti-iminoquinone-NAC-EUH antisera.

- T'O

- T O ' O

- T O O ' O

- T000'0

4.5 Discussion

Bioactivation of xenobiotics to highly reactive intermediates in vivo which can

subsequently undergo covalent binding to proteins has been widely implicated as a critical

event in target orgm toxicity (Boelsterli, 1993; Hinson e t al., 1994; Nelson and Pearson.

1990; Vermeulen et al.. 1992). We hypothesized that bioactivation of carbarnazepine to a

reactive metabolite and the subsequent covalent binding of this reactive intermediate to

proteins may play a role in carbarnazepine-induced idiosyncratic reactions. We have shown

that a reactive irninoquinone intermediate was formed readily from oxidation of 2-

hydroxyiminostilbene, a major metabolite of carbarnazepine. B y trapping the irninoquinone

intermediate with NAC and then coupling this conjugate to KLH, i t was possible to

produce an immunogen that subsequently was used to raise the polyclonal antibodies. The

antibodies were shown to recognize the irninoquinone epitope by ELISA experiment. The

possible role of the iminoquinone in carbarnazepine-induced toxicity was investigated using

the raised antiserum. Fernale Lewis rats were treated with carbamazepine (250 rnglkgJday,

gavage) for four days and the liver and bone marrow were isolated at the end of the dmg

treatment. üsing the developed anti-iminoquinone-NAC-KLH antiserum. i t was possible

for us to detect protein binding in the bone marrow cells of rats treated with carbamazepine

but not in that of the control animais. The major modified polypeptides had molecular

masses of 1 15 KDa and 133 KDa. This binding was inhibited by 2-hydroxyiminostilbene

(10 PM) but not by carbamazepine or iminostilbene at the same concentration. This result

suggested that the polypeptides that were recognized by the antiserum were likely to be

covalently modified by the iminoquinone reactive intermediate. Nonspecific binding was

detected in the bone marrow isolated from control rats that had not been treated with

carbamazepine. This c m b e explained by two reasons. First, if the rabbits used to raise

antibodies were not kept in a properly stenlized environment, they could develop antibodies

against various environmental bacteria and pathogens that will cross-react with proteins in

124 the bone marrow. Second. to detect the iminoquinone-modifïed proteins, a low dilution

(1500) of the antiserum had to be used, which rnay also contribute to the background

binding.

In a previous study, we have shown that when the irninoquinone intermediate of

carbarnazepine reacted with GSH or NAC, the GSH or NAC conjugate was formed.

However, the major pathway of this reaction was reduction of the irninoquinone to 2-

hydroxyirninostilbene. This sugpested that the irninoquinone intermediate is both an

electrophile and an oxidizing agent: therefore, it rnay covalently bind to critical proteins or it

may initiate oxidative damage. The iminoquinone intermediate. as an oxidizing agent, may

deplete the ce11 of glutathione and cause oxidative stress. Oxidative stress is generally

associated with production of reactive oxygen species, such as H202, 0 2 - and the highly

reactive -OH. These species may cause lipid peroxidation. DNA damage and oxidation of

protein amine groups. Oxidation of a protein may alter its structure, function and integrity.

Oxidative damage to proteins has been shown to resuit in protein turnover and decreased

enzymatic function, and it has been associated with aging and a number of pathological

processes including emphyserna. atherosclerosis and neurological diseases ( S tadtman and

Oliver, 199 1). Alteration of proteins by oxidation rnay also be rnechanistically important in

the toxicity of various chernicals and drugs.

Various functional groups on a protein may be oxidized and form different types of

products: cysteine-cysceine disulfide linkages, tyrosine-tyrosine linkages. methionine

sulfoxide and aldehydes frorn lysine, arginine and proline. In this study we used an

irnrnunoblotting assay to detect protein aldehyde formation, which may occur in vivo as a

result of oxygen activation via redox cycling between the iminoquinone intermediate and its

reduced product 2-hydroxyiminostilbene. Increased protein oxidation was found in the

nuclear and cytosol fractions of the liver from rats treated with carbarnazepine comparing to

that of the controis. Furthemore, in these two fractions, several oxidized polypeptides

with molecular masses of 43, 31. 30 and 28 KDa were oxidized in the sarnples from

125 carbarnazepine-treated rats but not in that of control rats. This susgested that

carbamazepine treatment caused the oxidation of target proteins in rats in vivo. The pattern

of protein oxidation was also different in carbarnazepine-treated rats than that of the control

rats. However, we do not have direct evidence that the protein oxidation wns due to the

formation of the iminoquinone intermediate il2 vivo. It is possible that the parent drug or

other unidentified metabolites may result in the protein oxidation. Agents that can prevent

the formation of the irninoquinone intermediate may prove whether it plays an important

role in protein oxidation and covalent binding. Since inhibition of the hydrolysis of 2-

hydroxycarbamazepine to 2-hydroxyiminostilbene would prevent the formation of the

iminoquinone in vivo, treating the animals with an acyl amidase inhibitor dong with

carbamazepine should be able to inhibit both the covalent binding and protein oxidation.

In surnmary, we have developed an anti-iminoquinone-NAC-KLH antiserurn and

used the antiserum to investigate the covalent binding of the irninoquinone to the bone

rnarrow cells of rats treated with carbamazepine in vivo. In addition. we have dso detected

increased protein oxidation in the nuclear and cytosol fractions of the liver from rats treated

with carbarnazepine compared to the controls.

Chapter 5

Summary and Discussion

5.1 Summary

The aim of the studies reported in this thesis was to investigate whether

bioactivation of indomethacin and carbamazepine could lead to the formation of reactive

metabolites that rnay be responsible for the idiosyncratic adverse reactions caused by these

two dnigs.

The results from the indomethacin study demonstrated that DMBI, a major

metabolite of indornethacin. can be further oxidized by the MPO system in activated

neutrophils to a reactive iminoquinone intermediate. The identity of this reactive

intermediate was supported by several strong pieces of evidence. The diode-cirray UV

spectrophotometric data showed that when DMBI reacted with HOCl, a new species was

formed that showed a UV absorbance at longer wavelengths (suggesting a quinone-like

structure). The iminoquinone intermediate with a MH+ ion at m/z 7W was detected by MS

in a flow system when the products of the reaction between DMBI and HOCl were

analyzed. Furthemore, this irninoquinone intemediate reacted with sulfhydryl-containing

nucleophiles. such as GSH and NAC, to form conjugates, and the structure of the

irnnoquinone-NAC conjugate was confirmed by proton NMR analysis. The NAC was

bound to the aromatic ring orrlzo to the hydroxyl group. The reactivity of the iminoquinone

intermediate as an elecuophile and its formation by MPO in activated neutrophils suggested

that this reactive irninoquinone intermediate may be responsible for indomethacin-induced

idiosyncratic adverse reactions, especinlly agranuloc ytosis.

It has been suggested that idiosyncratic drug reactions associated with

carbamazepine are due ro bioactivation of carbamazepine by CYP450 to form an arene

oxide and deficiency of epoxide hydrolase is a risk factor for these adverse reactions (Shear

and Spielberg, 1988). However, in patients with a history of carbamazepine

hypersensitivity reaction. the decreased epoxide hydrolase genotype was no more prevelant

than in the control group (Gaedigk et al., 1994: Green et al., 1995). The only direct

128 evidence to support the "arene oxide hypothesis" was the finding of a GSH adduct of

carbamazepine in rat bile (Madden et al., 1996). The structure of this GSH adduct is

consistent with an arene oxide precursor and the GSH was substituted ortho-to the

hydroxyl group (Figure 5- 1); however, such an adduct was not found in humans (Maggs et

al., 1997). In addition. the carbarnazepine hypersensitivity reaction involves many organs,

in which CYP450 is not a major oxidizing enzyme, that are unlikely to be able to form

significant amounts of arene oxide.

We have discovered an alternative metabolic pathway of carbamazepine leading to a

reactive irninoquinone (Figure 3- 1). The formation of 2-hydroxycarbamazepine, which is

the first step in this pathway, is known to be a major metabolic pathway in humans

(Eichelbaum et al., 1984). The second step in the formation of the iminoquinone is

hydrolysis of 2-hydroxycarbamazepine to 2-hydroxyiminostilbene, either in the liver or

peripheral tissues. We have found that, in the urine of three patients taking carbamazepine,

2-hydroxyiminostilbene is a major metabolite. Funhermore, we also demonstrated that 2-

hydroxyiminostilbene is readily oxidized to the reactive irninoquinone and this oxidation

even occurs spontaneously in air. This iminoquinone reacts with GSH and NAC and the

substitution is mrtn-to the hydroxyl group, as opposed ta the orrlio substitution of the

putative arene oxide. Another difference between the GSH adduct of the irninoquinone and

that of the arene oxide is the presence of the intact urea group in the adduct formed from the

arene oxide. In the urine of these three patients, we also found a metabolite that appears to

be a thiomethylderivative of the 3--hydroxyiminostilbene, which presurnably came from

further gut metabolism of a GSH conjugate, although its structure has not been

conclusively demonstrated. However, we did not find any sulfur-containing derivatives of

2-hydroxycarbamazepine. which presurnably is the further metabolite of a GSH adduct of

the arene oxide.

To further investigate the role of the iminoquinone intermediate in carbamazepine-

induced adverse reactions. an antibody (anti-iminoquinone-NAC-KLH) against the

129 irninoquinone epitope was developed and used to study the covalent binding of the

iminoquinone to endogenous proteins. Immunoblotting studies showed that the antibody

was able to detect covalent binding in the bone marrow of rats treated with carbamazepine

for four days. Furthemore. the covalent binding w u inhibited by 2-hydroxyiminostilbene

but not by carbamazepine or irninostilbene. These results suggested that the polypeptides

recopnized by the antiserum were likely to be covalently modified by the iminoquinone

reactive intermediate. Furthermore, although the iminoquinone covalently binds to

sulhydryl groups, the dominant reaction is reduction to 2-hydroxyimjnostilbene, and we

have also detected increased protein oxidation in the nuclear and cytosol fractions from the

liver of rats treated with carbamazepine compared to the coiitrols. The protein oxidation is

probably due to reactive oxygen species generated in vivo during the redox cyling between

the irninoquinone and its reduced product 2-hydroxyirninostilbene (Figure 3-41.

We have demonstrated that the major metabolites of indomethacin (DMBI) and

carbamazepine (2-hydroxyiminostilbene) can be oxidized to reactive iminoquinone

intermediates in one target of toxicity, specifically neutrophils. We have dso shown that

these iminoquinone intermediates are reactive electrophiles and that they react with

sulfhydryl-containing nucleophiles (i.e. GSH and NAC). We were able to dctect covalent

binding to bone manow cells in rats treated with carbamazepine using rhe anrisemm raised

against the reactive irninoquinone epitope. However. covalent binding was not detected in

other tvget organs. such as the liver or skin, from rats treated with carbarnazepine. One

possible reason for this discrepancy is that the antiserum was raised against a specific

antigen in which the iminoquinone was covaiently bound to NAC before binding to protein

and this may not be the same way that the irninoquinone binds to proteins in vivo.

Therefore, in the future, i t will be necessary to develop another antibody against the

Unmunogen in which the irninoquinone direcrly binds to proteins. In addition. although we

have detected covalent binding using the antiserum developed against the iminoquinone

epitope, the role of the reactive irninoquinone in the covalent binding needs ro be funher

130 examined. If the major reactive metabolite responsible for the binding is the irninoquinone.

agents that prevent its formation should inhibit the covalent binding. Since inhibition of the

hydrolysis of 2-hydroxycarbamazepine to 2-hydroxyiminostilbene would prevent the

formation of the iminoquinone i t ~ vivo, treating the animals with an acyl amidase inhibitor

at the sarne time as carbamazepine should be able to inhibit the covalent binding. A positive

result from this experirnent, would suggest that the iminoquinone, not the arene oxide, is

responsible for most of the covalent binding. It is important, in the future. to study the

covalent binding in humans because metabolism may be different in animals than in

humans. The cornmon targets of carbarnazepine-induced toxicity are the liver, skin and

blood. Although obtaining a liver biopsy is difficult, it should be much easier to obtain

blood and skin biopsies from patients who are treated with carbamazepine. If we find

binding in these tissues, we should then be able to determine in which cells and where

within the cells the binding is the greatest. Funhermore, we can make an affinity column

with the antibodies to separate proteins that are rnodified by the reactive metabolite and to

detemiine their sequences.

Carbarnazepine Arene oxide

GSH adduct

Figure 5- 1: Proposed pathway of carbarnazepine bioactivation to an arene oxide

intemediate ,and its reaction with GSH.

5.2 Discussion

Idiosyncratic drug reactions pose a signifiant complication in modem therapy

because of their unpredictable nature and potentially fatal outcome. Although the

rnechanisms involved are poorly understood, there is a large amount of circumstantial

evidence to suggest that chemically reactive metaboiites play a pivotal role. Such reactive

rnetabolites may be directly toxic to the cells or they may becorne covalently bound to a

macromolecular carrier, such as a protein, which rnay initiate an immune-mediated reaction.

132 Chernically reactive metabolites are, in general, electrophiles that c m react with nucleophilic

groups on proteins to become haptens. The nucleophilic groups on proteins include lysyl

and cysteinyl residues, and the imidazole and phenol groups present in histidine and

tyrosine respectively. For example. halothane is oxidized in the liver to reactive metabolite

TFA, which reacts preferentially with lysine residues on hepatic proteins (Kenna et al.,

1988). It has also been demonstrated that patients with halothane-induced hepatitis have

antibodies against specific TFA-modified hepatic proteins. Therefore, it is essential to

identiQ the reactive metabolite of a drug in order to understand their interactions with

cellular macromolecules. Having a clear idea of the chemical nature of the reactive

metabolite also may help in developing strategies to prevent drug-induced adverse

reactions. For example, drugs that have an aromatic amino group are usually oxidized to

highly reactive metabolites. such as hydroxylamine and nitroso derivatives. These reactive

metabolites have been associated with adverse reaction caused by dmgs with a free nmino

group, such as sulfonamides. dapsone and procainamide (Coleman rf cd . . 1989; Rieder er

al., 1989; Uetrecht et al.. 198Sb; Cetrecht. 1990). Although the incidence and clinical

outcome of adverse reactions caused by these drugs are variable, in each case the initial

evenr in drug toxicity is N-hydroxylation whch occurs in most individuals taking the drug.

In both carbamazepine and indomethacin studies, we have demonstrated that reactive

iminoquinone intermediates were formed during the oxidation of the major metabolite of

indomethacin (DMBI) and of carbamazepine (2-hydroxyirninostilbene). The common

feature of these two major metabolites (DMBI and 2-hydroxyirninotilbene is that they both

have an amine and a hydroxyl group in para positions to each other on the same aromatic

ring. This allows to bioactivation of the molecules to irninoquinone species. As discussed

in Chapter 1, it has been s h o w that a number of drugs or their metabolites with similar

strucrural features as DiMBI and 3-hydroxyiminostilbene can be bioactivated to reactive

iminoquinone intermediates by various metabolizing enzymes (e.;. CYP450, MPO or

PHs). Studies have also suggested that the formation of reactive iminoquinone

133 intermediates are responsible for the adverse reactions caused by these drugs. For

exarnple, bioactivation of amodiaquine to a reactive iminoquinone and its covalent binding

to various proteins have been demonstrated in several systems: human liver microsomes,

the horseradish peroxidase system and activated neutrophils. Bioactivation of

acetaminophen to a reactive N-acetylbenzoquinone imine by CYP450. MPO and PHs

systems has also been associated with renal and hepatotoxicity in cases of aceraminophen

overdoses. Therefore, a knowledge of the functional groups on a molecule which lead to

bioactivation and formation of a reactive metabolite may dlow the development of analogs

which do not contain such functional groups but still retain their therapeutic propenies.

Although i t is comrnonly accepted that reactive metabolites are involved in

idiosyncratic drug reactions, in most cases, this has not been rigorously proven.

Identification of a reactive metabolite of a dmg does not equal proving its role in cciusing

toxicity. It is possible that other undiscovered reactive metabolites are more imponant;

even if the identified reactive metabolite appears to be responsible for the toxicity, there

could be other risk factors for idiosyncratic reactions besides formation of a reactive

metabolite. ~Most drugs can be metabolized to chemically reactive metabolites: however,

they do not al1 cause a high incidence of idiosyncrütic reactions. This can be explained by

several possibilities. First. the reactive merabolites can be detoxified. in many cases.

before rhey can initiate tissue damage. It has been demonstrated that a number of drugs,

which are associated with occasional irnmunological reactions, may undergo bioactivation

by human liver microsornes to chemically reactive metabolites in vivo (Park and Coleman,

1988). However. parallel iri vivo studies have shown that such bioactivation is normally

precludrd or severely restricted by competing oxidation and conjugation reactions. When

bioactivation does occur, the chemically reactive metabolites are rapidly detoxified by

cellular defense mechanisms such as conjugation by glutathione (or glutathione transferase)

and hydrolysis by epoxide hydrolase. Therefore, the production of haptens. which could

stimulare the immune system, would be prevented (or reduced). However, the normal

134 balance between the bioactivation and detoxication pathways may be disturbed, in which

case the reactive metabolite cm escape detoxication.

Second, the chernical reactivity of the reactive metabolite is a major facror in

determining the degree and character of toxicity (Gillette, 1974a; Gillette. 197Jb: Park and

Kitteringham. 1990). If a reactive metabolite has low reactivity. it can be selecrively

detoxified before binding to some criticai molecules. At the other extreme. if a reactive

metabolite is too reactive. i t rnay react with anything i t encounters, such as water or the

enzyme that formed it, and have little chance to bind to other important molecules

(Uetrecht, 1992). For example, acetaminophen is associared with a low incidence of

idiosyncratic reactions. The reactive metabolite of acetaminophen is the

rtcetylbenzoquinone irnine and the low incidence of xetaminophen-associated idiosy ncratic

reactions is probably due to the selectivity of its reactions. The acetylbenzoquinone imine

metabolite is a soft electrophilr and preferentially r e m s with soft nucleophiles. such as

glutathione. However, its reaction with GSH is relatively slow in the absence of

glutathione transferase (Coles et al., 1988). Therefore. there is little chance for covalent

binding to other important rnolrcules, Le. proteins. Cysteine and lysine are probably the

two most comrnon nucleophilic arnino acids. The reactivity of the reactive metabolite and

whether it is 3 "soft" or "hard" electrophile also determine the amino acid residue to which

it binds (,L'etrecht. 1992). Funhermore, the amount of reactive metabolites binding to

proteins is also important. In jenerai. if a molecule is seing to induce an immune-mediated

reaction by acting as a hapten. the higher the hapten density (nurnber of dnig molecules

conjugated per molecule of protein), the higher intensity of the immune response.

Third. the site of reactive rnetabolite formation and the reactivity of the intermediate

c m also determine the target organ for toxicity. If the reactive metabolite is only formed in

one organ and is too reactive to escape that organ. then the toxicity should be confined CO

that organ. For exampie, halothane appears to be oxidized exclusively by CYP450 in the

liver to form a very reactive metabolite, TFA. As would be expected, the idiosyncratic

135 reactions associated with halothane are limited to the liver. Since liver is the major site of

drug oxidation, it may be considered to be particularly vulnerable to toxic effects of drugs

and their metabolites. However, many idiosyncratic reactions do not involve the liver.

This can be explained by the high capacity of cellular detoxication systems and the large

reserve of glutathione in the liver.

Founh, the irreversible binding of the reactive metabolite to specific proteins in vivo

is presumably important for initiating an immune reaction. For example, captopril can

extensively and covalently bind to cysteine residues in plasma proteins. However,

disposition studies have shown that the disulfide linkage is biochernically labile and can be

readily dissociated by endogenous thiols such as glutathione, which possibly explains why

captopril is only very wenkly immunogenic in both humans and experimental ünimals

(Coleman et al., 1988).

Fifth, the probability of a reactive metabolite of a drug to cause adverse reactions is

dependent on its toxicity to the cells and its efficiency in protein binding. The efficiency in

protein binding of a reacrive metabolite is increased by the therapeutic dose of the drug, the

percentage of the drug that can be bioactivated to the reactive metabolite and the rextivity of

this metabolite to specific proteins.

If al1 idiosyncratic drug reactions involve a reactive rnetabolite. it might be expected

that al1 drugs that can form a similar reactive metabolite would cause the same type of

reactions. However, this is often not the case. For example. i t has been suggested that the

hydroxylarnine rnetabolite of procainamide, which can be nonenzymatically oxidized to a

reactive nitroso-derivative, plays an important role in the mechanisrn of procainamide-

induced lupus. However. dapsone, which is also an arylarnine and can be oxidized to a

hydroxylamine, causes several different types of hypersensitivity reactions. Specifically

they both cause agranulocytosis but procainamide causes lupus and dapsone does not,

while dapsone causes methemoglobinemia and procainarnide does not. In this thesis, we

found bioactivation of carbarnazepine and indornethacin lead to the formation of suucturally

136 similar reactive metabolites (Le. the irninoquinone species); however, the patterns of

idiosyncratic drug reaction causzd by these two dmgs are somewhat different. The major

target of indomethacin-induced idiosyncratic dmg reactions is the blood: but the most

common serious adverse reaction associated with the use of carbamazepine is

"hypersensitivity " reaction, which c m affect any organ. This difference is probably due to

the following reasons. First, indomethacin is given in lower doses (75 mg/day) than

carbarnazepine (700 mg/day). Second, in contrast to carbamazepine. which is usually used

chronically, indomethacin is usually given in acute situations (such as gout) for a short

period of time. Third, Our studies have shown that 2-hydroxyiminostilbene is more readily

oxidized to the iminoquinone intenediate, in fact it undegoes autoxidation when rxposed

in air. Therefore, presumably the iminoquinone of carbamazepine can be formed in any

tissue where 7-hydroxy iminostilbene reaches. Fourth, although both reactil-e metabolites

are irninoquinone species. their chernical reactivities are not identical. We showed that the

irninoquinone of indornethacin rems with both N- and S- containinj nucleophiles. but the

iminoquinone of carbamazepine appears to only react with S-containing nucleophiles. In

addition, we found ihat the iminoquinone of carbamazepine c m dso undergo redox cycling

which could lead to different toxicological consequences than protein binding i~z vivo.

I t is thought that many dmg-induced toxicities are due to the covalent binding of

renctive metabolires to tissue proteins. It has also been suggested that protein bindin; is not

random, but rather selective wirh respect to the proteins targeted (Cohen et al.. 1997). The

selective binding may influence homeostatic or other cellular responses which in turn

contribute to drug toxicity, hypersensitivity or autoimmunity. Modern biochemical,

molecular, and immunochemical approaches have made i t possible to srudy covalent

bindin; of chemically reactive metabolites to cellular proteins and to identifi the specific

protein targets. For example, several protein targets modified by the reactive metabolite of

halothane, TFA. have been isolated frorn liver rnicrosomes of halothane-treated rats and

identified. The 57, 58, 59, 63, 80. 52 and 100 KDa proteins have been identified as

137 protein disulfide isomerase, a disulfide isomerase isoform, a carboxylesterase, calreticulin,

72 KDa endoplasmic reticulum protein, 78 KDa glucose-regulated protein and 94 KDa

ducose-regulated protein, respectively (Amouzadeh and Pohl. 1995). Results from other C

studies indicate that s imilx proteins are present in the livers of humans (Smith et al., 1993)

and they appear to become covalently modified by the TFA moiety after patients have been

administered halothane (Pohl et al.. 1989). The ability to identify individual proreins which

become modified by reactive intermediates may greatly increase our understanding of the

rote of such binding in dmg-induced idiosyncratic reactions. Once these protein targets of

reactive metabolites have been purified. they c m serve as antigens for identifying sensitized

individuals who have experienced a hypersensitivity reaction to the dmg. This information

c m be used to prevent not only an allergic reacrion to the dmg, but possibly cross-reactions

to orher dnigs that are stmcturally related. Another important application of these studies is

the design of safer alternative drugs that will not produce structurally similar reactive

metabolites.

Although protein binding is important. it may not be the only way that a reactive

metabolite can modify proteins and initiate an immune response. Chemically reactive

rnetaboiites c m also cause oxidative stress, which may in turn, initiate an innare immune

response. The innate immune system, althouoh primitive, is an essential part of Our

defense system against infections. An innate immune response involves recognition of

certain chernical features shared by groups of microorganisms but not by the host. such as

the lipopolysaccharide of Gram-negative bacterial ce11 walls; and it d s o appears to be able

to detect cells that have been stressed (Fearon. 1997; Fearon and Locksley, 1996: Poccia et

ai.. 1998). There is also evidrnce that the innote immune response communicates its

recognition to the adaptive system by providing the "danger" signal, without which a new

antigen usually induces tolerance in the adaptive system (Bendeiac and Fearon, 1997).

Some reactive metabolites, which are oxidizing azents, can cause oxidative damage of the

cells either directly or through reactive endogenous oxygen species. Stressed or injured

138 cells can be detected by the innate immune systern, which can destroy the cells or may

provide "danger" signals to the adaptive immune system. This will ultimately lead to tissue

darnage.

In my discussion of the mechanisms involved in idiosyncratic drug reactions, one

question that must be asked is why only certain individuals are exceptionally susceptible to

the toxicity. A number of possibilities exist to explain this (Park et al., 1992; Pohl el al.,

1996; Uetrecht, 1990). Individual susceptibility to toxicity will depend on the balance

between the relative roles of drug bioactivation and detoxication of both parent drug and the

metabolite. Sorne people may have abnormally high levels of an enzyme that bioactivates

the drug to a reactive metabolite or an allelic variant of the enzyme that is catalytically more

active than the normal enzyme. Altematively individuais that develop a particular drug

reaction rnay have abnormally low levels or activities of enzymes that detoxify reactive

metabolites, such as glutathione S transferases, epoxide hydrolase or N-acetyl transferase.

However, abnormal drug metabolism cannot explain why only a srnall number of

parient are susceptible to idiosyncratic drug reaction. For example, halothane is

metaboiized to an acylhalide by CYP450 and the reactive metabolite covalently binds to

proteins to form neoantigens. Studies have shown thnt al1 experimental animals and

patients exposed to the drug generate the antigen (Kenna er ni., 1988); however. the

incidence of halothane hepatitis is between 1 in 35.000 and 1 in 350,000 (Park et di..

1992). Thus it is rhought that the difference in susceptible patients is more likely to be in

the immune response to the antigen rather than in neoantigen formation (Kenna et al.,

1988).

The ability to initiate an immune response on antigen exposure is controlled by the

major hisrocompatibility (MHC) genes, which are the most polyrnorphic genes known in

higher vertebrates (Park et al., 1992). Such a high level of polymorphism could

theoretically result in individuals who are more efficient at recognizing certain epitopes than

others with the result that only these individuais are likely to rnount a vigorous immune

139 response to drug-related antigens. The relationships between idiosyncratic drug reactions

and dmg-related antigens associated with the MHC expressed on accessible cells have not

been established; however. epidemiological studies have shown that certain pol y morphism

of MHC genes are weakly associated with increased susceptibilities to certain antoimmune

diseases such as insulin-dependent diabetes mellitus (Baisch et al.. 1990) and coeliac

disease (Bugawan et al., 1989).

Although initiation of a drug-induced immune response is likely the primary event

in drug hypersensitivity, it is not a pathoiogical process. Many patients with anti-drug

antibodies or drug-induced antoantibodies remain asymptomatic. For example, with

chronic procainarnide therapy, most patients develop antinuclear antibodies and only about

204 develop symptornatic lupus. Therefore, it is likely that there are individual risk factors

associated with the translation of an immune response into tissue damage. The factors in

detennining when sensitization is translated into tissue damage could be the amount and site

of dmg antigen formation. the half-life of the antibodies, the magnitude of the immune

response and the type of the immune response, Le. IgE, IgG or cell-mediated.

Furtherrnore, some individuals with certain diseases or infections may increase their

susceptibility to drug-induced adverse reactions. For examplr. AIDS is associated with an

increase in the incidence in sulfonamide hypersensitivity reactionç (Fischl et al, 1988) and

mononucleosis is associated with an increased incidence of ampicillin rash (Pohl et al.,

1988). Aspirin hepatotoxicity and other types of idiosyncratic drug reactions appear to be

more common in patients with diseases such as lupus (Carneron and Ramsay, 1984;

Konttinen and Tuominen. 197 1; Uetrecht, 1990).

Each of these facrors in itself rnay not be sufficient to produce toxicity, but a

combination of factors acting collectively in a certain individual rnay lead to serious

toxicity. At present. there is no simple screening test with which to predict which dmgs are

likely to cause adverse reactions and to identify those individuals who rnay be susceptible

to drug-induced hypersensitivity. However, studies to identify possible reactive

140 rnetabolites, to identify and purify the protein targets of reactive metabolires and to analyze

the genetic polymorphism, both with respect to immune funcrion and drug metabolism, in

susceptible individuals may provide some insight into the mechanism of idiosyncratic dnig

reactions.

References

Ackroyd JF (1975) Immunological mechanisms in drug hypersensitivity., in Cli>iicol

aspects of irnmunology (Gel1 PGH, Coombs RRA and Lackmann PJ eds), Blackwell

Scientific Publications, Oxford.

Ackrovd JF (1962) The immunological mechanisms in drug hypersensitivity. Proc R Soc

Med 55: 30-36.

Amore BM, Kalhorn TF, Skiles GL, Hunter AP, Bennett GD, Finnell RH, Nelson SD and

Slattery JT (1997) Chvacterization of carbamazepine metabolism in a mouse mode1 of

carbarnazepine teratogenicity . Dnig Mem b Dispos 25: 953-962.

Amouzadeh HR and Pohl LR (1995) Processing o f endoplasmic reticulum luminal antigens

associated with halothane hepatitis in rat hepatocytes. Hepatology 22: 936-943.

Anderson JA ( 1992) Allergic reactions to drugs and biological agents. JAMA 268: 2845-

2857.

Askmark H and Wiholm B (1990) Epidemiology o f adverse reactions to carbamazepine as

seen in a spontaneous reporting system. Acta New01 Scmd 81: 13 1-140.

Baisch JM. Weeks T, Giles R, Hoover M. Stastny P and Capra JD ( 1990) Analysis of

HLX-DQ genotypes and susceptibility in insulin-dependent diabetes mellitus. New

Dig l J bled 322: 1836- 184 1.

Barrett AJ. Weller E, Rozengurt N, Lonzhurst P and Humble JG (1976) Amidopyrine

a_oranulocytosis: drug inhibition of granulocyte coionies in the presence of patient's

serum. Br Med J 2: 850-85 1.

Bendelac A and Fearon DT (1997) Innate pathways that control acquired immunity. C w r

Opin Immttnol 9: 1-3.

Billings RE (1985) Mechanisms of catachol formation from aromatic compounds in

isolated rat hepatocytes. Dnig Metab Disp 13: 287-290.

142 Billings RE and Fischer LJ ( 1985) Oxygen-18 incorporation studies of the rnetabolism of

phenytoin to the catechoi. Drtcg Metab Disp 13: 3 12-3 17.

Boelsterli UA (1993) Specific targets of covalent dmg-prorein interactions in hepatocytes

and their toxicological significance in dmg-induced liver injury. Drug Metab Rev 25:

395-45 1.

Boyd JA and Eling TE (198 1 ) Prostaglandin endoperoxide synthetase-dependent CO-

oxidation of acetaminophen to intermediates which covaienily bind in vitro to rabbit

rend medullary rnicrosomes. J Pharmacol ExpTher 219: 659-664.

Boyum A (1984) Separation of lymphocytes, granulocytes and monocytes from hurnan

blcod using iodinated density gradient media. Methods in Enzyinology 108: 88- 102.

Brodie BB, Reid WD, Cho AK, Sipes G, Krishna G and Gillette JR (1971) Possible

mechanism of liver necrosis caused by aromatic organic compounds. Proc Natl Arad

Sci 68: 160- 164.

Brown LP, Flint OP, Orton TC and Gibson GG (1986) Chemical teratogenesis: testing

methods and the role of metabolism. Dnrg Metab Rev 17: 22 1-760.

Bugawan TL, Angelini G, Larrick J, Auricchio S, Ferrara GB and Erlich HA (1989) A

combination of a particular HLA-DP beta allele and an HLA-DQ heterodimer confers

susceptibility to coeliac disease. Notwe 339: 470-373.

Cameron HA and Ramsay LE (1984) The lupus syndrome induced by hydralazine: ü

common complication with low dose treatment. Br Med J 289: 4 10-412.

Carmela C, Vittorio MD, Jennie J and Muglia MD ( 1995) Anticonvulsant hypersensitivity

syndrome. Arch h tem Med 155: 2285-2290.

Cereghino JJ, Brock JT, van Meter JC, Penry JK. Smith LD and White BG (1974)

Carbarnazepine for epilepsy. Neirrology 24: 40 1-1 10.

Cereghino JJ, van .Meter IC, Brock JT, Penry IK, Smith LD and White BG (1973)

Preliminary observations of serum carbarnazepine concentration in epileptic patients.

Neiirology 23: 357-366.

143 Chang VHT ( 1983) Synthesis of 2-hydroxycarbamazepine: a metabolite of carbamazepine.

J Heterocyclic Chem 20: 237-238.

Christie G, Breckenridge AM and Park BK (1989) Detection of antibodies towards the

antirnalarial amodiaquine and its quinone imine metabolite in man and the rat. Biochem

Ptzarmacol 38: 145 1 - 1458.

Clarke JB, Maggs JL, Kitteringham NR and Park BK (1990) Immunogenecity of

amodiaquine in the rat. Iut Arcli Allergy Appl Immunol 91: 335-342.

Cohen SD, Pumford NU, fiairallah EA, Boekelheide K, Pohl LR, Amouzadeh HR and

Hinson JA (1997) Selective protein covalent binding and target organ toxicity. Toxic

Appl Phannacol 143: 1 - 12.

Coleman J, Foster A, Yeung JHK and Park BK (1988) Dry-protein conjugates XV. A

study of the disposition of D-penicillamine in the rat and its relationship to

imrnunogenicity. Biocliem Phannac 37: 737-742.

Coleman MD, Breckenridge AM and Park BK (1989) Bioactivation of dapsone to a

cytotoxic metabolite by hurnan hepatic microsomal enzymes. Br J ClUi Pltanmcol 28:

389-395.

Coles B, Wilson 1, Wardman P, Hinson JA, Nelson SD and Ketterer B (1988) The

spontaneous and enqrnatic reaction of N-acetyl-p-benzoquinoneimine with glutathionr:

a stopped-flow kinetic study. Arclz Bioclienz Biophys 261: 253-260.

Coombs RRA and Gell PGH (1963) The classification of allergic reactions underlying

disease., in Cli~zical nspecrs of inimtinology (Gell PGH and Coombs RUA eds) pp ,

Blackwell Scientific Publications, Oxford.

Cooper DS (1984) Antithyroid dmgs. New Ejzgl J Med 311: 1353-1362.

Cribb AE, iMiller M. Tesoro A and Spielberg SP (1990a) Peroxidase-dependent oxidation

o f sulfonarnides by monocytes and neutrophils from humans and dogs. Mol Plzarmacol

38: 744-751.

144 Cnbb AE and Spielberg SP (1990b) Hepatic microsomal metabolism of sulfamethoxazole

to the hydroxylamine. Dnig Metab Dispos 18: 784-787.

Cribb AE, Spielberg SP and Griffin GP ( 1995) NJ-hydroxy lation of sulfamethoxazole by

cytochrome P450 of the cytochrome P450 2C subfamily and reduction of

sulfamethoxazole hydroxylamine in human and rat hepatic microsomes. Drrig Metab

Dispos 23: 406-414.

C d 1 WE (1 973) Dmgs 5 years later: carbarnazepine. A m h t e m Med 79: 844-817.

Cuthben MF (1974) Adverse reactions to non-steroidal antirheumatic drugs. Cwr Med Res

Opin 2: 600-6 10.

Dansette PM. Amar C, Smith C, Neau E, Pons C and Mansuy D (1989) I j i vitro. the 5-

hydroxylation of tienilic acid is catalyzed by cytochrome P350 in rat and human

microsomes., in Cyrochroine P-450, Bioche~nisrry cind Biopkysics (Sc huster 1 eds)

pp 75 1-754, Taylor and Francis. New York.

Dansette PM, Amar C, Valadon Pl Pons C, Beaune PH and Mansuy D (1991)

Hydroxylation and formation of electrophilic metabolites of tienilic acid and its isomer

by human liver microsornes. Catalysis by a cytochrome P450 IIC different from that

responsible for mephenytoin hydroxylation. Biochem Pliarr~rcicol 41: 553-560.

Dodrill CB and Troupin AS (1977) Psychotropic effects of carbarnazepine in rpilepsy. A

double-blind comparison with phenytoin. 1Vr~rrology 27: 1023- 1025.

Duzgan DE. Hogans AF, Kwan KC and McMahon FG (1972) The rnetabolisrn of

indomethacin in man. J Phanmzcol Expnier 181: 563-375.

Dunk AA, Walt RP, Jenkins WJ and Sherlock SS (1982) Diclofenac hepatitis. Br Med J

Clin Res 284: 1605- 1606.

Durelli L, Massazza U and Cavallo R (1989) Cnrbamazepine toxicity and poisoning:

incidence. clinical features and management. Merl To.rAdv Drug Ekp 4: 95-107.

145 Eastrnond DA, Smith MT and Irons RD (1987) A n interaction of benzene metcbolites

reproduces the myelotoxicity observed wirh benzene exposure. Toxicol Appl

Plzanrzacol 91: 85-95.

Eastrnond DA, Smith MT. Ruzo LO and Ross D (1986) Metabolic activation of phenol by

human myeloperoxidase and horse-radish peroxidase. Mol Phamacol 30: 674-679.

Eichelbaum M, Jensen C, von Sassen W, Bertilsson L and Tomson T (1984) Ii2 vivo and

iti vitro biotransformation of carbarnazepine in man and rat, in Metabolisin of

anriepileptic dncgs. (Levy RH, Pitlick WH, Eichelbaum M and Meijer J eds) pp 27-33,

Raven Press, New York.

Eling TE, .Mason RP and Sivarajah K (1985) The formation of aminopyrine cation radical

by the peroxidase rtctivity of prostaglandin H synthase and subsequent reactions of the

radical. J Biol Chem 260: 160 1 - 1607.

Ellrodt AG, Murata GH, Reidinger MS, Stewart ME, Mochizuki C and Gray R ( 1984)

Severe neutropenia associated with sustained-release procainamide. Ann Inter Med

100: 197-201.

Epstein M, Vickars L and Stein H (1990) Diclofenac induced immune thrornbocytopenia. J

Rltertrnatol 17: 1303- 1403.

Eustace S. O'Neill T, McHale S and Molony J (1989) Fatal aplristic anaemia following

prolon jed diclofenac use in an elderl y patient. Ir J Med Sci 158: 2 17.

Fearon DT (1997) Seeking wisdom in innate immunity. Nature 388: 323-324.

Fearon DT 'and Locksley RM (1996) The instructive role of innate immunity in the acquired

immune response. Science 272: 50-53.

Fischer V. Haar JA, Greiner L, LLoyd RV and Mason RP (199 1) Possible roIe of free

radical formation in clozapine (clozad)-induced agranulocytosis. Mole Plion?mcol JO:

846-853.

146 Fischl MA, Dickinson GM and La Voie L (1988) Safety and efficacy of sulfamethoxazole

and trimethoprim chemoprophylaxis for pneumocystis carinii pneumonia in AIDS.

JAMA 259: 1 185- 1 189.

Fleming CM, Branch RA, Wilkinson GR and Guengerich FP (1992) Human liver

microsomal N-hydroxylation of dapsone by cytochrome P-450 3A4. Mol Plznrmncol

41: 975-980.

Freudenthal RI and Jones PW ( 1976) Poly?zuclear nronlatic hydroc~rbons: chetnistry,

metabolism, and carcinogenesk Raven Press, New York.

Gaedigk A, Spielberg SP and Grant DM (1994) Characterization of the microsornal

epoxide hydrolase gene in patients with anticonvulsant adverse drug reactions.

Phannacagerzetics 4: 142- 1 5 3.

Gelboin HV ( 1980) Benzo(a)pyrene metabolism, activation and carcinogenesis: role and

regulation of mixed-funtion oxidases and related enzymes. Physiol Rev 60: 1107-

1166.

Gelboin HV and Ts'o POP ( 1978) Polycylic lzydrocarbotrs and cancer. Academic Press,

New York.

Gerson WT, Fine DG. Spielberg SP and Sensenbrennrr LL (1983) Anticonvulsant-

induced aplastic anernia: increased susceptibility to toxic dmg metabolites in vitro.

Blood 61 : 889-893.

Gill HJ, Tingle MD and Park BK (1995) N-hydroxylation of dapsone by multiple enzymes

of cytochrome P450: implicatins for inhibition of haemotoxicity. Br J Clin Pliannucol

40: 53 1-539.

Gillette JR (1974a) h perspective on the role of chemically reactive metabolites of foreign

compounds in toxicity. 1. Correlation of changes in covalent binding of reactivity

metabolites with changes in the incidence and severity of toxicity. Biocliern Plzïirnzocol

23: 2785-2794.

147 Gillette JR ( 1971b) A perspective on the role of chemically reactive metabolites of foreign

compounds in toxicity. II. Alterations in the kinetics of covalent binding. Biochem

Phannacol 23: 2927-2938.

Gilliland BC (1991) Drug-induced autoimmune and hematologic disorders. Imnicnol

Allergy Clin North Am 11: 525-553.

Goldstein RA and Patterson R ( 1 984) Sumrnary. J Allergy Clin Im>nuuol 74: 643-644.

Gram L and Jensen P (1989) Carbarnazepine toxicity., in Antiepileptic h g s . (Levy R,

Mattson R, Meldrum 8, Penry JK and Dreifuss FE eds) pp 555-565, Raven Press.

New York.

Green VJ, Pirmohamed M. Ketteringham NR, Gaedigk A. Grant DM, Boxer M, Burchell

B and Park BK (1995) Gznetic analysis of microsornal epoxide hydrolase in patients

with carbarnazepine hy persensit ivi ty. Biochem Pharnlncof 50: 13 53 - 1 359.

Greenlee WF, Sun JD and Bus JS (1981) A proposed mechanism of benzene toxicity:

formation of reac tive in termediates from pol yphenol metabolites. Toxicol Appl

Plzarmacol 59: 187- 195.

Hansen DK (199 1) The embryotoxicity of phenytoin: an update on possilbe rnechanisms.

Proc Soc Exp Bio Med 197: 36 1-368.

Hrinzlik RP, Hogberg K and Judson CM ( 1984) Microsomal hydroxylation of specifically

deuterated rnonosubstituted benzenes: evidence for direct aromatic hydroxylation.

Bioclternistq) 23: 3015-3055.

Haque KE and Proctor GR ( 1968) Dibenzo[b, f l azepine-Zone. Chem Cornrmm : 1412-

1413.

Haman RE, Meisinger MAP. Davis GE and Kuehl FA (1964) The metabolites of

indomethacin, a new antiinflammatory dmg. J Phannacol Ekp Thrr 143: 2 15-720.

Hatton CSR, Peto TEA. Bunch C and Pasvol G (1986) Frequency of severe neutropenia

associated with arnodiûquine prophylaxis against malaria. Limer 1: 41 1-3 13.

248 Helfgott SM. Sandberg-Cook J. Zakim D and Nestler J (1990) Diclofenac-associated

hepatotoxicity [see comments]. JAMA 264: 2660-2662.

Hinson IA , Pumford NR and Nelson SD (1994) The role of metabolic activation in dmg

toxicity. Dnrg Metab Rrv 26: 395-412.

Hinson JA and Roberts DW ( 1992) Role of covalent and noncovalent interactions in ceil

toxiciiy : effects on proteins. A i n i < Rev Pltommacol To-ricol 32: 47 71 -5 10.

Hoerner M, Behrens UJ, Worner T and Lieber CS (1986) Humoral immune response to

acetaldehyde adducts in alcoholic patients. Res Commun Cheni Porhol Pl~an~zncol 54:

3- 12.

Horowitz S, Patwardhan B and Marcel1 E (1988) Hepatotoxic reactions associated with

carbarnazepine therapy. Epilepsin 29: 149- 154.

Huberman E, Sach SL, Yang SK and Gelboin HV (1976) Identification of mutagenic

rnetabolites of benzo(a)pyrene in mammalian cells. Proc Nari Acnd Sci USA 73: 607-

61 1.

Hucker HB, Zacchei AG, Cox SV, Brodie DA and Cantwell NHR (1966) Studies on the

absorption, distribution and excretion of indomethacin in various species. J Plinnnncol

Exp Ther 153: 237-219.

Isildar M, Abou-Khalil WH, Jimenez JJ, Abou-Khalil A and Yunis AA \ 1988) Aerobic

nitroreduction of dehydrochlorarnphenicol by bone marrow. Tosicol Appl Plinrmcrcol

94: 305.

Islam 1 and Skibo EB (1991) Structure-activity studies of antiturnor agents based on

pyrrolo [ 1.2-31 benzimidazoles: new reduc tive alkylating DNA cleaving agents. J Med

Chern 34: 2954-296 1.

Ivanetich KM, Ziman MR and Bradshaw .JJ ( 1982) 1,1,l-Trichloropropene-2.3-oxide; an

alternate rnechanism for its inhibition of cytochrome p-450. Res Cornm~oz Cliern Patlz

Phannacol 35: 111-119.

149 Jeffrey AM, Weinstein IB, Jennette KW, Grzeskowiak K, Nakanishi K, Harvey RG,

Aurmp H and Horris C (1977) Structures of benzo(a)pyrene nucleic acid adducts

formed in humna and bovine bronchial expiants. Natru-e (Land) 269: 348-350.

Jerina DM and Dalay JW (1971) Arene oxides: a new aspect of drug metabolism. Science

185: 573-582.

Jewell H, Maggs JL, Harrison AC, O'Neill PM. Ruscoe JE and Park BK (1995) Role of

hepatic metabdisrn in the bioactivation and detoxication of amodiaquine. Xmobiotica

25: 199-217.

Kalyanaraman B and Sohnle PG (1985) Generation of free radical intermediates from

foreign compounds by neutrophils-derived oxidants. J Clin Invest 75: 16 18- 1622.

Kapitulnik J, Levin W, Conney AH, Yagi H and Jerina DM (1977) Benzo(a)pyrene 7,8-

dihydrodiol is more carcinogenic than benzo(a)pyrene in newbom mice. Nature (Lorid)

266: 378-3 80.

Kaplowitz N, Aw TY, Simon FR and S tolz A ( 1986) Drug-induced hepatotoxicity. Alitt

Inrem Med 104: 826-839.

Kaufman DW, Kelly JP, Jurgelon JM, Anderson T. Issaragrisil S, Whiholm B-E, Young

NS, Leaverton P, Levy M and Shapiro S (1996) Drugs in the aetiology of

a~ranulocytosis and aplastic anaemia. Eur 3 Hoemnrol 57: 23-30.

Kaufman DW, Kelly JP, Levy M and Shapiro S (199 1) The drtcg etiology of

ugranulocpsis and nplastic anemia. Oxford University Press, New York.

Kelder PP, de Mol NJ, 'T Hart BA and Janssen LH (1991) Metabolic activation of

chlorpromazine by stimulated human polymorphonuclear leukocytes. Induction of

covalent binding of chlorpromazine to nucleic acids and proteins. Chem Bi01 I~zterizct

79: 15-30.

Keller RI, Halrnes NC, Hinson JA and Pumford S R (1993) Immunochernicai detection of

oxidized proteins. Chem Res Toxicol 6: 130433.

150 Kendall PM, Johnson JV and Cook CE (1979) Synthetic route to an aromatic analogue of

siiigol. J Org Clleni 44: 1421-1424.

Kenna JG, Satoh H. Christ DD and Pohl LR (1988) Metabolic b a i s for a drug

hypersensitivity: antibodies in sera from patients with halothane hepatitis recognize liver

neoantigens that contain the tri fluoroacetyl group derived from halotbane. J Pharmacol

Exp Ther 245: 1 103- 1 109.

Kim HL and Kovacs MJ ( 1995) Diclofenac-associated thrombocytopenia and neutropenia.

Ann Pharmacother 29: 7 1 3-7 15.

Klebanoff SJ (1968) Myeloperoxidase-halide-hydrogen peroxide antibacterial system. J

Bacteriof 95: 2131-3138.

Knowles SR, Shapiro L and Shear NH (1996) Serious adverse reactions induced by

minocycline. Arch Dermatol 132: 934-939.

Konttinen YP and Tuominen L (197 1) Reversible procainamide-induced agranulocytosis

twice in one patient. Lancer 2: 925.

Krarner MR, Levene C and Hershko C (1986) Severe reversible antoimmune haemolytic

anaernia and thrombocytopenia associated with diclofenac therapy. Scand J H~zonatol

36: 118-120.

Lam DCC and Lindsay RH (1 9793) Accumulation of 2-[14~]propylthiouracil in human

polymorphonuclear leukocytes. Biochern Phannacol 28: 289-2296.

Lam DCC and Lindsay RH (1979b) Oxidation and binding of [2-14~]propylthiouracil in

hornan polymorphonuclear leukocytes. D m g Metab Dispos 7: 285-289.

Larrey D, Castot A, Pessayre D, Merigot P, Machayekhy JP, Feldarnn G, Lenoir A, Ruelt

B and Benhamoi* .JP (1986) Amodiaquine-induced hepatitis. A m Intern med 104:

80 1-803.

Lazarou J, Pomeranz BH and Corey PN (1998) Incidence of adverse drug reactions in

hospitalized patients: a rneta-analysis of prospective studies. JAMA 279: 1200- 1205.

151 Lee E, Miki Y, Hosokawa M. Sayo H and Kariya K (1988) Oxidative metabolism of

propylthiouracil by peroxidase frorn rat bone marrow. Xenobiorica 18: 1 135- 1 142.

Leeder JS, Gaedigk A, Lu X and Cook VA (1996) Epitope mapping studies with hurnan

anti-cytochrorne P45O 3A antibodies. Mol Pharmacol 49: 234-243.

Leeder JS, Riley RJ, Cook VA and Spielberg SP (1992) Human anti-cytochrome P450

antibodies in aromatic anticonvulsant-induced hypersensitivity reactions. J Pharmacol

Exp Ther 263: 360-367.

Lertratanangkoon K and Homing MG (1982) Metabolism of carbamazepine. Dmg Memb

Dispos 10: 1- 10.

Levine RL, Williams JA, Stadtman ER and Schacter E (1994) Carbonyl assays for

determination of oxidatively modified proteins. Metlzods E~zzynol 233: 346-357.

Lillibridge JH, Arnore BM, Slattery JT. Kalhorn TF, Nelson SD, Finnell RH and Bennett

GD ( 1996) Protein-reactive metabolites of carbamazepine in mouse liver microsornes.

D m g Metab Dispos 24: 509-5 14.

Liu ZC and Uetrecht JP (1995) Clozapine is oxidized by activated humün neutrophils to a

reactive nitrenium ion that irreversibly binds to the cells. J Phanmzcol E.rp Ther 275:

1476-1483.

Lynn RK, Smith RG. Thornpson RM. Deinzer ML. Griffin D and Gerber N (1978)

Characterization of glucuronide metabolites of carbamazepine in human urine by gas

chromatography and mass spectrometry. Dncg Metnb Dispos 6: 494-50 1 .

Madden S. ~Maggs JL and Park BK (1996) Bioactivation of carbamazepine in the rats in

vivo: evidence for the formation of reactive arene oxide(s). Drug Merab Dispos 224:

469-379.

Madden S, Spaldin V, Hayes RN, Woolf TF, Pool WF and Park BK (1995a) Species

variation in the bioactivation of tacrine by hepatic microsomes. Xenobiotica 25: 103-

116.

2 52 Madden S, Spaldin V and Park BK (199%) Clinical pharmacokinetics of tacrine. Clin

Phamacokinef 28: 449-457.

Madden S. Woolf TF, Pool WF and Park BK (1993) An investigation into the formation of

stable, protein-reactive and cytotoxic metabolites from tacrine in vitro. Studies with

h u m a and rat liver microsornes. BiocCrertr Phannacol 46: 13-20.

Madison FW and Squier TL ( 1934) The etiology of primary granulopenia (agranuiocytic

angina). JAMA 102: 755-759.

~Maggs JL, Kitteringham NR, Breckenridge AM and Park BK (1987) Autoxidative

formation of a chernically reactive intermediate from amodiaquine, a myelotoxin and

hepatotoxin in man. Bioclicm Phannucol 36: 206 1-2062.

LMaggs .IL, Pirmohamed M, Kitteringham NR and Park BK (1997) Characterization of the

metabolites of carbarnazepine in patient urine by Iiquid chromatography/mass

spectrometry. Dnig Metab Dispos 25: 275-280.

Maggs L, Tingle NR, Kitteringham NR and Park BK (1988) Mechanisms of formation of

protein-aryiating intermediates from amodiaquine, a myelotoxin and hepatotoxin in

man. Bioclzertt Plrannacol 37: 303-3 1 1.

Magis CC, Barge A and Dausset 1 (1968) Serological study of an ailergic agranulocytosis

due to noramidopyrine. Clin Exp Itnrnirriol 3: 989.

.Maguire IH, Kraus BL. Butler TC and Dubley KH ( 1979) Determination of 5-(3,4-

dihydroxy- 1 $5-cyclohexadien- 1 -yl)-5-phenylhydantoin (dihydrodiol) and quantitative

studies of phenytoin metabolism in man. Ther Dnig ~çianifori~zg 1: 339-370.

Mansuy D, Valadon P, Erdelmeier 1, Lopez-Garcia P, Arnar C1 Girault I and Dansette PM

(199 1) Thiophene S-oxides as new reactive metabolites: formation by cytochrome P450

dependent oxidation and reaction with nucleophiles. J Am Chem Soc 113: 7825-

7826.

153 Marnett LJ and Bienkowski MJ (1980) Hydroperoxide-dependent oxygenation of trans-

7,8-dihydrobenzo(a)pyrene by ram seminal vesicle microsomes. Biochern Biophp Res

Cornrnun 96: 639-647.

Mason RP (1979) Free radical metabolites of foreign compounds and their toxicological

significance.. in Reviervs in biocheniical tosicology. (Hodgson E . Bend JR and

Philpot RM eds) pp 15 1-200. Elsevier-North Holland, New York.

Mathews KP (1984) Clinical spectrum of allergic and pseudoaflergic drug reactions. J

Allergy Clin Immurioi 74: 558-566.

McOmie JFW, Watts ML and West DE (1968) Demethylation of aryl methyl ethers by

boron tribromide. Tetrahedron 24: 2289-2292.

Merk HF and Hertl M (1996) Immunologie mechanisms of cutaneous drug reactions.

Sernin Crtrarz Med Srrrg 15: 228-235.

Miller EC (1978) Some current perspectives on chemical carcinogenesis in humans and

experimental animals: presidential address. Cmcer Res 38: 1479- 1496.

Mitra AK, Thurnmel KE. Kalhom TF, Kharnsch ED. Unadkat JD and Slattery JT ( 1995)

Metabolism of dapsone to its hydroxylamine by CYPZEl in vitro and in vivo. Clin

P/iurmacol Tlier 58: 556-566.

,Miyamoto G, Zahid N and Uetrecht JP (1997) Oxidation of diclofenac to reactive

intermediates by neutrophils, myeloperoxidase, and hypochlorous acid. Clzein Res

Toxicol 10: 4 14-4 19.

Mockenhaupt M and Schopf E (1996) Epiderniology of drug-induced severe skin reactions.

Sernin Crrtan Med Surg 15: 236-243.

~Moeschlin S and Wagner K (1952) Agranulocytosis due to the occurrence of leukocyte-

agglutinins. Acra Huemat 8: 29-40.

Moldeus P and Rahimtula A (1980) Metabolism of paracetamol to a glutathione conjugate

catalysed by prostaglandin synthetase. Bioclzern Biophys Res Comnrun 96: 469-475.

154 Morselli PL and Figerio A ( 1975) Metabolism and pharmacokinetics of carbarnazepine.

Dncg Metab Rev 4: 97- 1 13.

Munns AJ. de Voss JI, Hooper WD, Dickinson RG and Gillam EM (1997) Bioactivation

of phenytoin by human cytochrorne P450: characterization of the mechanism and

targets of covalent adduc t formation. Chem Res Toxicol 10: 1 049- 105 8.

Naisbitt DJ, Ruscoe JE. Williams D, O'Neill PM, Pirrnohamed M and Park BK (1997)

Disposition of amodiaquine and related antirnalarial agents in human neutrophils:

implications for dmg design. J Phannacol Exp Ther 280: 884-893.

Nakajima M, Inoue T. Shirnada N, Tokudome S. Yamamoto T and Kuroiwa Y (1998)

Cytochrome P450 2C9 catalyzes indomethacin O-demethylation in human liver

microsomes. Dricg Metab Dispos 26: 26 1-266.

Neftel KA, Woodtly W, Schrnid M, Frick PG and Fehr 1 (1986) Amodiaquine induced

agranulocytosis and liver damage. Br Med J 292: 72 1-723.

Nelson SD and Pearson PG (1990) Covalent and noncovalent interactions in acute lethal

ce11 injury caused by chemicals. Annu Rev Phamcol Toxicol 30: 169- 195.

Neuberger J and Kenna JG (1987) Halothane hepatitis: a mode1 of immune mediated drug

hepatotoxicity. Chi Sei 72: 263-270.

Nomura F and Lieber CS (1981) Binding of acetaldehyde to rat liver rnicrosomes:

enhancement after chronic alcohol consumption. Biocltem Biophys Res Commun

100: 131-137.

Oesch F ( 1973) Mammalian epoxide hydrases: inducible enzymes catal ysing the

inactivation of carcinogenic and cytotoxic metabolites derived from aromatic and

olefinic compounds. Xet2obiotica 3: 305-340.

Park BK and Coleman JW (1988) The immunological basis of adverse drug reactions. A

report on a Symposium held in Liverpool on 6th April 1988. Br J Clin Phamac 26:

49 1-495.

155 Park BK, Coleman JW and Kitteringham NR (1987) Drug disposition and drug

hypersensitivity. Biochem Phannacol 36: 581-590.

Park BK and Kitteringham NR (1990) Dmg-protein conjugation and its immunoiogical

consequences. Drug Metab Rev 22: 87- 144.

Park BK, Pirmohamed M and Kitteringham NR (1992) Idiosyncratic drug reactions: a

mechanistic evaluation of risk factors. Br J Clin Phannac 34: 377-395.

Park BK, Pirmohamed M and Kitteringham NR (1995) The role of cytochrome P450

enzymes in hepatic and extrahepatic human drug toxicity. Pharmacol Ther 68: 385-

424.

Parker CW (1982) Allergic reactions in man. Plramacol Rev 34: 85- 104.

Piurish DD, Schlosser MJ. Kapeghian JC and Traina VM (1997) Activation of CGS 12094

(Prinomide metabolite) to 1,4-benzoquinone by myeloperoxidase: implications for

human idiosyncratic agranulocytosis. Fitnd Appl Toxicol 35: 197-204.

Pellock IM (1987) Carbamazepine side effects in children and adults. Epilepsia 28: S64-

S70.

Pessayre D, Wandscheer JC, Descatoire V, Artigou JY and Benhamou JP (1979)

Formation and inactivation of a chemically reactive metabolite of vinyl chloride. Toxicol

Appl Pharmacol 49: 505-5 15.

Petzc D and Fudenberg HH (1976) Irnmunologic mechanisms in dnig-induced cytopenias.,

in Progress in Hemarology (Brown EB eds) pp , Grune and Stratton, New York.

Pimohamed M. Kitteringharn NR, Breckenridge AM and Park BK (199%) The effect of

enzyme induction on the cytochrome p45O-mediated bioiictivation of carbamazepine by

mouse Iiver microsomes. Biochern Phannacol 44: 2307-23 13.

Pirmohamed M. Kitteringham NR, Guenthner TM, Breckenridge AM and Park BK

(1992b) An investigation of the formation of cytotoxic, protein-reactive and stable

metabolites from carbamazepine in vitro. Biochern Phannacol 43: 1675- 1682.

156 Pirmohamed M, Kitteringham NR and Park BK (1994) The role of active metabolites in

dnig toxicity. Drug Safery 11: 1 14- 144.

Pirmohamed M. Madden S and Park BK (1996) Idiosyncratic dnig reactions: metabolic

bioactivation as a pathogenic mechanism. Clin Pharmacol 31: 2 15-230.

Pisciotta AV, Ebbe S, Lennon EJ. Metzger GO and Madison FW (1958) Agranulocytosis

following administration of phenothiazine derivatives. Am J Med 25: 2 10-223.

Poccia F, Gougeon ML, Bonneville M, Lopez-Botet M. Moretta A, Battistini L, Wallace

M. Colizzi V and Malkovsky M (1998) Innate T-ce11 irnmunity to nonpeptidic antigens.

Immrrnol Today 19: 253-256.

Pohl LR (1990) Drug-induced aliergic hepatitis. Semin Liver Disense 10: 305-3 15.

Pohl LR, Kenna JG, Satoh H. Christ DD and Martin JL (1989) Neoantigens associated

with halothane hepatitis. Drug Metab Rev 20: 203-2 1 7.

Pohl LR, Pumford NR and Martin L (1996) LMechanisms. chernical structures and drug

metabolism. Eur J Haernatol 57: 98- 104.

Pohl LR, Satoh H, Christ DD and Kenna JG (1988) The irnmunologic and metabolic basis

of drug hypersensitivities. A m Rev Pltannacol 28: 367-3237,

Potter DW and Hinson JA (1987) The 1-electron and 2-electron osidrition of

acetaminophen catalyzed by prostaglandin-H synthase. J Biol Chein 262: 974-980.

Raucy JL, Lasker .TM, Lieber CS and Black M (1989) Acetarninophen activation by human

liver cytochromes P-450IIEI and P-450IAII. Arch biochem Biophp 271: 270-2 53.

Regnaud L. Sirios G and Chakrabarti S (1988) Effect o f four-day treatrnent with

carbarnazepine at different dose levels on microsomal enzyme induction. drug

metabolism and dnig toxicity. Pharmacol Toxicol 62: 3-6.

Reid WD and Krishna G (1973) Centrolobular hepatic necrosis related to covalent binding

of metabolites of halogenated aromatic hydrocarbons. Exp Mol Pathoi 18: 80-99.

Rickert DE, Baker TS, Bus JS, Barrow CS and Irons RD (1979) Benzene disposition in

the rat after exposure by inhalation. Toxicol Appi Pharmacol 49: 417-423.

157 Rieder MJ (1994) Mechanisms of unpredictable adverse drug reactions. Dnrg safew 11:

196-3 12.

Rieder MI. Uetrecht JP, Shear NH, Cannon M, Miller M and Spielberg SP (1989)

Diagnosis of sulfonamide hypersensitivity reactions by in vitro 'rechallenge' with

hydroxylarnine metabolites. Ann h t e m Med 110: 286-289.

Rietjens IM, den Besten C , Hanzlik RP and van Bladern PJ (1997) Cytochrome P450-

catdyzed oxidations of halobenzene. Chem Res Toxicol 10: 629-635.

Riley RJ, Kitteringham NR and Park BK (1989) Structurai requirements for bioactivation

of anticonvulsants to cytotoxic rnetabolism in vitro. Br J Clin Pliamacol 28: 482-487.

Riley RJ, Smith G, Wolf CR, Cook VA and Leeder JS (1993) Human anti-endoplasmic

reticulurn autoantibodies produced in aromatic anticonvulsant hypersensitivity reactions

recognise rodent CYP3A proteins and a similady regulated human P450 enzyme (s).

Biochem Biophys Res Comm 191: 32-40.

Sadler A, Subrahmanyam VV and Ross D (1988) Oxidation of catechol by horseradish

peroxidase and human leukocyte peroxidase: reactions of O-benzoquinone and o-

benzosemiquinone. Toxicol Appl Pharmacol 93: 61-7 1.

Salama A, Gottsche B and Mueller-Eckhardt C (1991) Autoantibodies and drug- or

metabolite-dependent antibodies in patients with diclofenac-induced immune

haernolysis. Br J Haernarol 77: 546-549.

Salama A, Schutz B, Kiefel V, Breithaupt H and Mueller-Eckhardt C (1989) Immune-

mediared agranulocytosis related to drugs and their metabolites: mode of sensitization

and heterogeneity of antibodies. Br J Haentatol 72: 12% 132.

Sawchuk RI, Maloney JA, Cartier LL, Rackley RJ, Chan KK and Lau HS (1995)

Analysis of diclofenac and four of its metabolites in human urine by HPLC. Pham Res

12: 756-762.

Sayo H and Saito M (1990) The mechanism of myeloperoxidase-catalysed oxidation of

aminopyrine. Xenobiotica 20: 957-965.

158 Schwartz MA (1969) Chernical aspects of penicillin allergy. J Phmm Sci 58: 643-66 1.

Shapiro S (1986) Risks of agranulocytosis and aplastic anemia: a first report of their

relation to drug use with special reference to analgesics. J M A 256: 1749- 1757.

Shear NH and Spielberg SP (1988) Anticonvulsant hypersensitivity syndrome. 112 vitro

assessrnent of risk. J Clin hvest 82: 1826- 1832.

Shulman NR (1958) Immunoreaction involving platelets. 1. A steric-kinetic mode1 for

formation of a complex frorn a human antibody. quinidine as a hapten, and platelets;

and for fixation of complement by the complex. J Exp Med 107: 665-690.

Sims P and Grover PL (1974) Epoxides in polycyclic aromatic hydrocarbon metabolism

and carcinogenesis. Adv Camer Res 20: 165- 174.

Smith DB, Mattsen RH. Cramer JA, Collins IF, Novelly RA and Craft B ( 1987) Results

of a nationwide veterans administration cooperative study comparing the side effects

md toxicity of carbarnazepine, phenobarbitol, phenytoin and prirnidone. Epilepsin 28

(Suppl 3): 50-58.

Smith GC, Kenna JG, Harrison DJ, Tew D and Wolf CR ( 1993) Autoantibodies to hepatic

rnicrosomal carboxylrsterase in halothane hepatitis. Lancer 312: 963-964.

Smith MT, Yager JW, Steinmetz YL and Eastmond DA ( 1989) Peroxidase-dependent

metabolism of benzene's phenolic rnetribolites and its potential role in benzene toxicity

and carcinogenicity. Eitviroti Heallh Perspect 82: 23-29.

Snyder R and Kocsis JJ (1975) Current concepts of chronic benzene toxicity. CRC Crit

Rev Toxicol 3: 365-25s.

Spaldin V. Madden S. Pool WF, Woolf TF and Park BK (1994) The effect of enzyme

inhibition on the metabolism and activation of tacrine by human liver rnicrosomes. Br J

Clin Phannacol 38: 15-22.

Spielberg SP, Gordon GB, Blake DA, Melliîs ED and Bross DS (198 1) Anticonvulsant

toxicity in vitro: possible role of arene oxides. J Phannacol Exp Ther 217: 386-389.

159 Stadtman ER and Oliver CN (1991) Metal-catrilyzed oxidation of proteins: Biochemical

mechanism and biological consequences. Free Rndic Bi01 Med 9: 3 15-325.

Stricker BHC and Spoelstra P ( 1985) Dmg-induced Izepatic injury. A compreliensive

slrrvey of the literatrtre on ndcerse drrig reactions L(P to Jnn~iary 1985. Elsevier,

Amsterdam.

Terashima K, Takai S, Usami Y, Adachi T, Sugiyama T, Katagiri Y and Hirano K (1996)

Purification and partial characterization of an indomethacin hydrolyzing enzyme from

pig liver. P h a m Res 13: 132% 1335.

Thornpson PF and Trimple MU (1982) Anticonvukant drugs and cognitive functions.

Epilepsia 23: 53 1-544.

Thummel KE, Lee CA, Kunze KL, Nelson SD and Slattery JT (1993) Oxidation of

acetaminophen to N-acetyl-p-arninobenzoquinone irnine by human CYP3A1. Biocliem

Phamtacol 45: 1563- 1569.

Timbre11 JA (1983) Drug hepatotoxicity. Br J Clif1 Pharnzacol 15: 3- 14.

Uetrecht J, Zahid N and Rubin R (1988a) Metabolism of procainarnide to a hydroxylamine

by human neutrophils and rnononuclear leukocytes. Cllem Res Toxicd 1 : 74-78.

Uetrecht J. Zahid N, Shear NH and Biggar WD (1988b) Metabolism of dapsone to a

hydroxylamine by human neutrophils and mononuclear cells. J Pliannacol E-rp Tlier

245: 274-279.

Uetrecht JP (1990) Drug metabolism by leukocytes and its role in dmg-induced lupus and

other idiosyncratic dmg reactions. CRC Crit Rev Taxicol 20: 313-235.

Uetrecht JP (1996) Reactive metabolites and agranulocytosis. Eur / Hnenrnrol 57: S83-

S88.

Uetrecht JP (1985) Reactivity and possible significance of hydroxylaminr and nitroso

metabolites of procainamide. J PIzamocol Exp Ther 232: 420-425.

Uetrecht SP (1992) The role of leukocyte-generated reactive metabolites in the pathogenesis

of idiosyncratic dmg reactions. Drug Metab Rev 24: 299-366.

160

Uetrecht JP, Ma HM, MacKnight E and McClelland R ( 1995) Oxidation of aminopyrine by

hypochlorite to a reactive dication: possible implications for arninopyrine-induced

agranulocytosis. Chem Res Toxicol 8: 226-233.

Uetrecht JP, Sweetman BJ. Woosley RL and Oates JA (1984) Metaboiism of procainarnide

to a hydroxylamine by rat and human hepatic rnicrosomes. Dmg Metab Dispos 12: 77-

8 1 .

Uetrecht JP and Zahid N ( 199 1 ) N-chlorination and oxidation of procainamide by

myeloperoxidase: toxicological implications. Chern Res Toxicol 4: 2 18-22.

Uetrecht JP, Zahid N and Whitfield D (1994) Metabolism of vesnarinone by activated

neutrophils: implications for vesnarinone-induced agranulocytosis. J Pltnrmzcol E.rp

Ther 270: 565-572.

van Zyl JM, Basson K, Kriegler A and van der Walt BJ (1990) Activation of

chlorpromazine by the myeloperoxidase system of the human neutrophil. Bioclzeni

Phannacol 40: 947-954.

Vermeulen NP, Bessens JGM and van de Straat R (1992) Molecular aspects of

paracerarnol-induced hepatotoxicity and its mechanism-based prevention. Dnrg Merab

Rev 24: 367-407.

Wade L. Kenna JG and Caldwell J (1997) Lmunochemicai identification of mouse hepatic

protein adducts derived from the nonsteroidal anti-intlarnrnatory drugs diclofenac,

sulindac, and ibuprofen. Clzeni Res Tosicol 10: 546-555.

Waldhauser L and Uetrecht J (199 1) Oxidation of propylthiouracil to reactive metabolites

by activared neutrophils: implications for agranulocytosis. D m g Metnb Dispos 19:

354-359.

Weiss SJ (1989) Tissue destruction by neutrophils. New EnglJ Med 320: 365-376.

Winn LM and Wells PG ( 1995) Free radical-mediated mechanisms of anti-convulsant

teratogenicity. Eur J Nerrro 3: 5-29.

161 Wood AW, Chang RL, Levin W, Yagi H, Thakker DR. Jerina DM and Conney AA (1977)

Differences in mutagenicity of the optical enantiomers of the diastereomeric

benzo(a)pyrene 7,8-diol-9.10 epoxides. Biochem Biophys Res Cornmiin 77: 138%

1396.

Zirnmerman HJ (1978) Hepntotoxicity. The adverse e f f e c ~ of dnigs and other chernicals

on the liver. Appleton-Century Crofts. New York.