UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies on isoniazid-induced hepatotoxicity in rabbits Sarich, Troy Casimir 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1997-251500.pdf [ 11.61MB ]
Metadata
JSON: 831-1.0088330.json
JSON-LD: 831-1.0088330-ld.json
RDF/XML (Pretty): 831-1.0088330-rdf.xml
RDF/JSON: 831-1.0088330-rdf.json
Turtle: 831-1.0088330-turtle.txt
N-Triples: 831-1.0088330-rdf-ntriples.txt
Original Record: 831-1.0088330-source.json
Full Text
831-1.0088330-fulltext.txt
Citation
831-1.0088330.ris

Full Text

STUDIES O N ISONIAZID-INDUCED HEPATOTOXICITY I N RABBITS by T R O Y CASHMIR SARICH B.Sc , The University of British Columbia, 1992 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Pharmacology & Therapeutics) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A September 1997 © Troy Casimir Sarich, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Yy\ts*i-.*s\piad>~ac\>-i t w<£ftA Peuno-s The University of British Columbia Vancouver, Canada Date ^rpre^Sisr^ ^ /^J-DE-6 (2/88) 11 ABSTRACT Approximately one-third of the world's population harbors the tuberculosis mycobacterium and it is estimated that tuberculosis kills over 2.5 million persons per year worldwide. Isoniazid is widely used in the prophylaxis and treatment of tuberculosis in this population. A major risk of isoniazid-therapy is the onset of an unpredictable and potentially fatal hepatotoxicity in 1-2% of individuals. Although the existence of this toxicity has been widely recognized for over 25 years, its mechanism remains unknown. Thus, the hepatotoxicity remains neither preventable nor treatable. Our laboratory has developed a reliable model of isoniazid-induced hepatotoxicity in rabbits which closely resembles the toxicity in humans. This model consists of repeated injections of isoniazid over a two day period along with examination of isoniazid-induced pathological changes before, during and after exposure to isoniazid. The purpose of this thesis was to characterize the biochemical and pathological changes induced by isoniazid and to identify and attempt to alter toxicologically important enzymatic steps in the metabolism of isoniazid. The main findings were as follows: 1. Isoniazid causes hepatic cell damage, hepatic steatosis and hypertriglyceridemia in rabbits. 2. Plasma levels of the isoniazid-metabolite hydrazine, but not acetylhydrazine, correlated significantly with the severity of isoniazid-induced hepatic cell damage. 3. Pre- and co-treatment of isoniazid-treated rabbits with L-thyroxine increases the activity of N A D P H cytochrome P-450 reductase (reductase). I l l 4. An increase in activity of reductase was associated with a decrease, rather than the expected increase, in the severity of isoniazid-induced hepatotoxicity. 5. Isoniazid administration inhibits cytochrome P-450 (CYP) 1A1/2 and CYP2E1 activities in rabbits. 6. Pretreatment of isoniazid-treated rabbits with bis-p-nitrophenyl phosphate (BNPP) (a proposed inhibitor of isoniazid-amidase which hydrolyses isoniazid and its metabolites to hydrazine) effectively decreased plasma levels of isoniazid-derived hydrazine and decreased the severity of isoniazid-induced hepatic cell damage, hepatic steatosis and hypertriglyceridemia. 7. In vitro studies revealed that BNPP acts as an irreversibly acting non-competitive antagonist (IC50 approximately 2 JUM) of isoniazid-amidase. In conclusion, hydrazine is most likely the main causative agent in isoniazid-induced hepatotoxicity in rabbits. The results in this animal model are in accord with recent data on isoniazid-hepatotoxicity in humans. iv TABLE OF CONTENTS A B S T R A C T ii T A B L E OF CONTENTS iv LIST OF FIGURES vii LIST OF TABLES xii LIST OF ABBREVIATIONS xiv A C K N O W L E D G E M E N T xv FOREWORD xvii INTRODUCTION 1 TUBERCULOSIS 1 ISONIAZID 1 Indications 1 Pharmacology 2 Pharmacokinetics 2 Inhibition and Induction of Hepatic Microsomal Enzymes 3 Toxicities 3 HEPATOTOXICITY 4 Historical Perspective 5 Pathology 6 Other Risk Factors 6 HEPATOTOXICITY - A N I M A L MODELS 8 Rats 9 Guinea Pigs 10 Rabbits 10 PROPOSED MECHANISMS 12 Immunogenic Hypothesis 12 Acetylhydrazine Hypothesis 12 Hydrazine Hypothesis 14 THESIS HYPOTHESES 20 Introduction 20 STUDY 1 R A T I O N A L E 20 STUDY 1 HYPOTHESES 22 STUDY 2 R A T I O N A L E 22 S T U D Y 2 HYPOTHESIS 22 STUDY 3 R A T I O N A L E 23 S T U D Y 3 HYPOTHESES 24 STUDY 4 R A T I O N A L E 24 STUDY 4 HYPOTHESIS 25 METHODS 26 A N I M A L S : 26 A N I M A L TREATMENTS 26 INH-Injection Protocol 26 Hypothyroidism 27 Hyperthyroidism 28 Bis-p-Nitrophenyl Phosphate (BNPP) Treatment 29 A N I M A L H A N D L I N G 31 Blood Sampling 31 L I V E R H A N D L I N G 31 Crude Homogenate 31 Microsome Preparation 32 Whole Liver Sample Freezing 32 BIOCHEMICAL A N A L Y S E S 32 Argininosuccinic Acid Lyase (ASAL) Activity 32 Alanine Aminotransferase (ALT) Activity 34 Liver and Plasma Triglyceride Determinations 34 Plasma Free T4 Levels 35 Protein Determination 35 Microsomal P-450 and Enzyme Activity Determination 35 Thiobarbituric Acid Reactive Substances (TBARS) Determination 36 Tissue Glutathione Determination 36 Analytical procedure 36 F L U O R O M E T R Y 38 Resorufin Assay 38 Analytical procedure 39 H I G H P E R F O R M A N C E LIQUID C H R O M A T O G R A P H Y (HPLC) A N A L Y S E S 42 Summary of Hydrazine Assays Used 42 H P L C Analysis of INH, Acetylhydrazine and Hydrazine 42 Determination of Hydrazine in 12 Hour Plasma Samples 46 Rationale 46 Analytical Procedure 47 Validation 48 Determination of Hydrazine in 24 and 32 Hour Plasma Samples 51 Analytical Procedure 51 Hepatic INH-Amidase Activity Determination 54 Rationale 54 Analytical Procedure 55 BNPP IC50 Determination of Hepatic INH-Amidase Activity 58 Rationale 58 Analytical Procedure 58 Determination of the Mechanism of INH-Amidase Inhibition - Lineweaver-Burk and Eadie-Hofstee Plots 59 Rationale •. 59 Analytical Procedure 59 Plasma INH-Amidase Activity Determination 59 Rationale 59 Analytical Procedure 59 Percent Recovery Calculations 62 vi Slope Differences Between 24 & 32 Hour Plasma Hydrazine Procedure and the Amidase Procedures 67 Acetylation Phenotype Assay 68 Rationale 68 Analytical Procedure 68 Statistics 73 RESULTS 95 STUDY 1 95 STUDY 2 106 STUDY 3 115 STUDY 4 138 DISCUSSION 173 STUDY 1 173 DETERMINATION OF HEPATIC C E L L D A M A G E , HEPATIC STEATOSIS A N D HYPERTRIGLYCERIDEMIA, A N D THE EFFECTS OF A C E T Y L A T I O N R A T E A N D GENDER O N INH-INDUCED HEPATOTOXICITY IN RABBITS ; 173 STUDY 2 178 DETERMINATION OF INH, A C E T Y L H Y D R A Z I N E A N D H Y D R A Z I N E CONCENTRATION IN P L A S M A SAMPLES A N D CORRELATION WITH INH-INDUCED HEPATOTOXICITY IN RABBITS 178 STUDY 3 182 THE R O L E OF INCREASED A N D DECREASED HEPATIC R E D U C T A S E ACTIVITY I N INH-INDUCED HEPATOTOXICITY IN RABBITS 182 STUDY 4 192 THE EFFECT OF INHIBITION OF INH-AMIDASE B Y BIS-p-NITROPHENYL PHOSPHATE O N INH-INDUCED HEPATOTOXICITY IN RABBITS 192 FINAL DISCUSSION 210 Interrelationships of INH-Induced Pathological Changes 210 Effects of Food and Water Deprivation 210 The Acetylhydrazine Theory 211 Hydrazine in INH-Induced Hepatotoxicity. 212 Cytochrome P-450 Levels in INH-Induced Hepatotoxicity 213 Reductase Activity and INH-Induced Hepatotoxicity 214 EROD (CYP1A1/2) Activity in INH-Induced Hepatotoxicity 215 /?-Nitrophenol Hydroxylase (CYP2E1) Activity in INH-Induced Hepatotoxicity 215 Amidase Activity and INH-Induced Hepatotoxicity 217 The Rabbit Model in Relation to Humans 219 THESIS CONCLUSIONS 225 REFERENCES 227 APPENDIX 241 APPENDIX 1 241 APPENDIX 2 250 vii LIST OF FIGURES FIGURE 1.1 GLUTATHIONE STANDARD CURVE 75 FIGURE 1.2 RESORUFIN SCANS (EMISSION AND EXCITATION) 76 FIGURE 1.3 RESORUFIN STANDARD CURVE 77 FIGURE 1.4 RESORUFIN STANDARD CURVES AT EXCITATION WAVELENGTHS OF 550 nm AND 575 nm 78 FIGURE 1.5 STANDARD CURVE FOR ISONICOTINYL-3-METHOXYBENZALDHYDRAZONE IN PLASMA 79 FIGURE 1.6 STANDARD CURVE FORiV-ACETYL-3-METHOXYBENZALDHYDRAZONE IN PLASMA 80 FIGURE 1.7 STANDARD CURVE FOR 3 -METHOXYBENZALDAZINE IN PLASMA 81 FIGURE 1.8 CHROMATOGRAMS OF ISONICOTINYL-3-METHOXYBENZALDHYDRAZONE, JV-ACETYL-3-METHOXYBENZALDHYDRAZONE AND i-METHOXYBENZALDAZINE IN PLASMA.... 82 FIGURE 1.9 STANDARD CURVE FOR i-METHOXYBENZALD AZINE IN 12 HOUR PLASMA SAMPLES 83 FIGURE 1.10 CHROMATOGRAMS FOR 3-METHOXYBENZALD AZINE IN 12 HOUR PLASMA SAMPLES 84 FIGURE 1.11 STANDARD CURVE FOR 3-METHOXYBENZALD AZINE IN 24 AND 32 HOUR PLASMA SAMPLES 85 FIGURE 1.12 CHROMATOGRAMS FOR 5-METHOXYBENZALD AZINE IN 24 AND 32 HOUR PLASMA SAMPLES 86 FIGURE 1.13 STANDARD CURVE FOR DETERMINATION OF HEPATIC AMIDASE ACTIVITY 87 FIGURE 1.14 CHROMATOGRAMS FOR HEPATIC AMIDASE ASSAY 88 FIGURE 1.15 STANDARD CURVE FOR DETERMINATION OF PLASMA AMIDASE ACTIVITY 89 FIGURE 1.16 CHROMATOGRAMS FORPLASMA AMIDASE ASSAY 90 viii FIGURE 1.17 STANDARD CURVES FOR H Y D R A Z I N E A S S A Y S USING THE PERCHLORIC ACID A N D ACETONITRTLE D E N A T U R A T I O N STEP 91 FIGURE 1.18 STANDARD C U R V E FOR SULFAMETHAZINE I N P L A S M A 92 FIGURE 1.19 STANDARD C U R V E FOR A C E T Y L S U L F A M E T H A Z I N E I N P L A S M A 93 FIGURE 1.20 C H R O M A T O G R A M S SHOWING SULFAMERAZINE, S U L F A M E T H A Z I N E A N D A C E T Y L S U L F A M E T H A Z I N E I N P L A S M A 94 FIGURE 2.1 CORRELATION OF A C E T Y L A T I O N R A T E WITH P E A K P L A S M A TRIGLYCERIDES 98 FIGURE 2.2 CORRELATION OF A C E T Y L A T I O N R A T E WITH A N I N D E X OF HEPATIC C E L L D A M A G E 99 FIGURE 2.3 CORRELATION OF HEPATIC TRIGLYCERIDE CONTENT WITH P E A K P L A S M A TRIGLYCERIDE CONCENTRATION 100 FIGURE 2.4 CORRELATION OF HEPATIC TRIGLYCERIDE CONTENT WITH A N I N D E X OF HEPATIC C E L L D A M A G E 101 FIGURE 2.5 CORRELATION OF P E A K P L A S M A TRIGLYCERIDE CONCENTRATION WITH A N INDEX OF HEPATIC C E L L D A M A G E 102 FIGURE 2.6 CORRELATION OF TIME TO P E A K P L A S M A TRIGLYCERIDE CONCENTRATION VERSUS TIME TO P E A K HEPATIC C E L L D A M A G E . . . 103 FIGURE 2.7 A C E T Y L H Y D R A Z I N E P L A S M A CONCENTRATION I N 32 H O U R P L A S M A SAMPLES VERSUS A R E A U N D E R THE C U R V E OF A N I N D E X OF HEPATIC C E L L D A M A G E 108 FIGURE 2.8 A C E T Y L H Y D R A Z I N E P L A S M A CONCENTRATION IN 32 HOUR P L A S M A SAMPLES VERSUS A N I N D E X OF HEPATIC C E L L D A M A G E A T 48 HOURS 109 FIGURE 2.9 A C E T Y L H Y D R A Z I N E P L A S M A CONCENTRATION IN 32 H O U R P L A S M A SAMPLES VERSUS A N INDEX OF P E A K HEPATIC C E L L D A M A G E 110 FIGURE 2.10 H Y D R A Z I N E P L A S M A CONCENTRATION I N 32 H O U R P L A S M A SAMPLES VERSUS A R E A U N D E R THE C U R V E OF A N I N D E X OF HEPATIC C E L L D A M A G E I l l ix FIGURE 2.11 HYDRAZINE PLASMA CONCENTRATION IN 32 HOUR PLASMA SAMPLES VERSUS AN INDEX OF HEPATIC CELL DAMAGE AT 48 HOURS 112 FIGURE 2.12 HYDRAZINE PLASMA CONCENTRATION IN 32 HOUR PLASMA SAMPLES VERSUS AN INDEX OF PEAK HEPATIC CELL DAMAGE 113 FIGURE 2.13 PLASMA FREE L-THYROXINE LEVELS AT BASELINE, BEFORE AND AFTER INH IN ANIMALS RECEIVING L-THYROXINE 124 FIGURE 2.14 PLASMA FREE L-THYROXINE LEVELS AT BASELINE, BEFORE AND AFTER INH IN ANIMALS NOT RECEIVING L-THYROXINE 125 FIGURE 2.15 PEAK PLASMA ARGININOSUCCINIC ACID LYASE ACTIVITY IN THE TREATMENT GROUPS 126 FIGURE 2.16 PEAK PLASMA ALANINE AMINOTRANSFERASE ACTIVITY IN THE TREATMENT GROUPS 127 FIGURE 2.17 HEPATIC TRIGLYCERIDE ACCUMULATION IN THE TREATMENT GROUPS 128 FIGURE 2.18 HEPATIC CYTOCHROME P-450 LEVELS IN THE TREATMENT GROUPS.. 129 FIGURE 2.19 HEPATIC REDUCTASE ACTIVITY IN THE TREATMENT GROUPS. 130 FIGURE 2.20 HEPATIC p-MTROPHENOL HYDROXYLASE ACTIVITY (CYP2E1) IN THE TREATMENT GROUPS 131 FIGURE 2.21 HEPATIC CYTOCHROME P-450 LEVELS IN THE PRETREATMENT GROUPS OF THE FOLLOW-UP STUDY.. 132 FIGURE 2.22 HEPATIC REDUCTASE ACTIVITY IN THE PRETREATMENT GROUPS OF THE FOLLOW-UP STUDY 133 FIGURE 2.23 HEPATIC/?-NITROPHENOL HYDROXYLASE ACTIVITY (CYP2E1) IN THE PRETREATMENT GROUPS OF THE FOLLOW-UP STUDY 134 FIGURE 2.24 COMPARISON OF PEAK PLASMA ARGININOSUCCINIC ACID LYASE ACTIVITY BETWEEN TREATMENT GROUPS 147 FIGURE 2.25 COMPARISON OF PEAK PLASMA ALANINE AMINOTRANSFERASE ACTIVITY BETWEEN TREATMENT GROUPS 148 FIGURE 2.26 COMPARISON OF HEPATIC TRIGLYCERIDE ACCUMULATION BETWEEN TREATMENT GROUPS 149 FIGURE 2.27 COMPARISON OF 32 HOUR PLASMA TRIGLYCERIDE LEVELS BETWEEN TREATMENT GROUPS 150 FIGURE 2.28 COMPARISON OF MICROSOMAL HYDRAZINE PRODUCTION BETWEEN TREATMENT GROUPS. 151 FIGURE 2.29 COMPARISON OF PERCENT CHANGE IN MICROSOMAL HYDRAZINE PRODUCTION BETWEEN TREATMENT GROUPS 152 FIGURE 2.30 CONCENTRATION-RESPONSE CURVE FOR BIS-p-NITROPHENYL PHOSPHATE IN MICROSOMAL INCUBATION WITH INH 153 FIGURE 2.31 PERCENT CONCENTRATION-RESPONSE CURVE FOR BlS-p-NITROPHENYL PHOSPHATE IN MICROSOMAL INCUBATION WITH INH 154 FIGURE 2.32 LINEWEAVER-BURK PLOT FOR INH-AMIDASE INHIBITION... 155 FIGURE 2.33 EADIE-HOFSTEE PLOT FOR INH-AMIDASE INHIBITION 156 FIGURE 2.34 COMPARISON OF PLASMA HYDRAZINE CONCENTRATIONS AT 12, 24 AND 32 HOURS 157 FIGURE 2.35 COMPARISON OF HEPATIC GLUTATHIONE LEVELS BETWEEN TREATMENT GROUPS 158 FIGURE 2.36 COMPARISON OF HEPATIC THIOBARBITURIC ACID REACTIVE SUBSTANCES FORMATION BETWEEN TREATMENT GROUPS 159 FIGURE 2.37 COMPARISON OF HEPATIC CYTOCHROME P-450 LEVELS BETWEEN TREATMENT GROUPS 160 FIGURE 2.38 COMPARISON OF HEPATIC CYTOCHROME P-450 REDUCTASE ACTIVITY BETWEEN TREATMENT GROUPS 161 FIGURE 2.39 COMPARISON OF HEPATIC ETHOXYRESORUFIN-O-DEETHYLASE (AN INDEX OF CYP1 Al/2) ACTIVITY BETWEEN TREATMENT GROUPS 162 FIGURE 2.40 COMPARISON OF HEPATIC BENZOYLOXYRESORUFIN-O-DEALKYLASE (AN INDEX OF CYP2B4) ACTIVITY BETWEEN TREATMENT GROUPS... 163 FIGURE 2.41 COMPARISON OF HEPATIC PENTOXYRESORUFIN-O-DEALKYLASE (AN INDEX OF CYP2B4/5) ACTIVITY BETWEEN TREATMENT GROUPS 164 xi F I G U R E 2 .42 C O M P A R I S O N O F H E P A T I C / J - N I T R O P H E N O L H Y D R O X Y L A S E A C T I V I T Y B E T W E E N T R E A T M E N T G R O U P S 1 6 5 F I G U R E 2.43 C O R R E L A T I O N O F A C E T Y L A T I O N R A T E V E R S U S A N I N D E X O F H E P A T I C C E L L D A M A G E 1 6 6 xii LIST OF TABLES TABLE 1.1 DOSING SCHEDULE FOR INH 26 TABLE 1.2 TREATMENT GROUP DEFINITIONS, ABBREVIATIONS AND DOSING TIMES 30 TABLE 1.3 BIS-p-NITROPHENYL PHOSPHATE AND INH INJECTION PROTOCOL 30 TABLE 1.4 VALIDATION SUMMARY - GLUTATHIONE ASSAY 37 TABLE 1.5 PRECISION OF THE PROCEDURE FOR DETERMINATION OF GLUTATHIONE IN LIVER HOMOGENATE. 38 TABLE 1.6 VALIDATION SUMMARY - RESORUFIN ASSAY 40 TABLE 1.7 PRECISION OF THE PROCEDURE FOR DETERMINATION OF RESORUFIN IN MICROSOMES 41 TABLE 1.8 VALIDATION SUMMARY - INH, ACETYLHYDRAZINE AND HYDRAZINE ASSAY 45 TABLE 1.9 INTRA- AND INTER-DAY VARIABILITY (ACCURACY AND PRECISION) : 46 TABLE 1.10 VALIDATION SUMMARY - DETERMINATION OF HYDRAZINE IN 12 HOUR PLASMA SAMPLES 50 TABLE 1.11 PRECISION OF THE PROCEDURE FOR DETERMINATION OF HYDRAZINE IN 12 HOUR PLASMA SAMPLES 51 TABLE 1.12 VALIDATION SUMMARY - DETERMINATION OF HYDRAZINE IN 24 AND 32 HOUR PLASMA SAMPLES 53 TABLE 1.13 PRECISION OF THE ASSAY FOR DETERMINATION OF HYDRAZINE IN 24 AND 32 HOUR PLASMA SAMPLES 54 TABLE 1.14 VALIDATION SUMMARY - HEPATIC AMID ASE ASSAY 57 TABLE 1.15 PRECISION OF THE PROCEDURE FOR DETERMINATION OF HYDRAZINE IN MICROSOMAL INCUBATIONS 58 TABLE 1.16 VALIDATION SUMMARY - PLASMA AMID ASE ASSAY 60 TABLE 1.17 PRECISION OF THE PROCEDURE FOR DETERMINATION OF HYDRAZINE IN PLASMA INCUBATIONS 61 xiii TABLE 1.18 PERCENT RECOVERY FOR THE 12 HOUR PLASMA HYDRAZINE ASSAY... 63 TABLE 1.19 PERCENT RECOVERY FOR THE 24 & 32 HOUR PLASMA HYDRAZINE ASSAY 64 TABLE 1.20 PERCENT RECOVERY FOR THE HEPATIC AMIDASE ASSAY 64 TABLE 1.21 PERCENT RECOVERY FOR THE PLASMA AMIDASE ASSAY 65 TABLE 1.22 VALIDATION SUMMARY - ACETYLATION ASSAY 71 TABLE 1.23 PRECISION OF THE PROCEDURE FOR DETERMINATION OF SULFAMETHAZINE AND ACETYLSULFAMETHAZINE IN PLASMA 72 TABLE 1.24 PHENOTYPING: INTRA- AND INTER-DAY VARIABILITY 72 TABLE 1.25 STANDARD PEAK RATIO DETERMINATION 73 TABLE 2.1 BIOCHEMICAL CHANGES PRE- AND POST-ISONIAZID TREATMENT 104 TABLE 2.2 INDIVIDUAL RABBIT TOXICITY DATA 105 TABLE 2.3 PLASMA CONCENTRATIONS OF INH, ACETYLHYDRAZINE AND HYDRAZINE IN INDIVIDUAL RABBITS 114 TABLE 2.4 COMPARISON BETWEEN TWO PRETREATMENT VEHICLE-CONTROL GROUPS 135 TABLE 2.5 TOXICOLOGICAL MARKERS AND MICROSOMAL ENZYME ACTIVITIES BY GROUP 136 TABLE 2.6 MICROSOMAL ENZYME ACTIVITIES IN THE FOLLOW-UP STUDY 137 TABLE 2.7 TOXICOLOGICAL MARKERS BY GROUP 167 LIST OF ABBREVIATIONS ALT Alanine aminotransferase ANOVA Analysis of variance ASAL Argininosuccinic acid lyase BROD Benzoyloxyresorufin-O-dealkylase BNPP Bis-/?-Nitrophenyl phosphate CYP Cytochrome P-450 EROD Ethoxyresorufin-O-deethylase GSH Glutathione HPLC High performance liquid chromatography INH Isoniazid NADPH Nicotinamide adenine dinucleotide NAPQI iV-acetyl-p-benzoquinone imine PROD Pentoxyresorufin-(9-dealkylase SEM Standard error of the mean TBARS Thiobarbituric acid reactive substances T4 L-Thyroxine TG Triglycerides VEH Vehicle XV ACKNOWLEDGEMENT I would like to express my sincere appreciation to the students, staff and faculty of the Department of Pharmacology & Therapeutics, UBC, all of whom created a positive environment in which to study and many of whom played a direct role in the successful completion of this thesis (notably Dr. Casey van Breemen, Maureen M., Maureen G., Janelle, Christian & George). I owe special thanks to my supervisor, Dr. James Wright, who by encouraging me to take on several projects in addition to my thesis, broadened my experience and provided a continually challenging and rewarding environment in which to study and learn. I am very grateful to my thesis committee of Dr. Gail Bellward, Dr. Thomas Madden and Dr. Michael Walker, who helped to direct this project and to focus my research efforts. I also want to thank Dr. Richard Wall for teaching me valuable analytical techniques and for his much appreciated participation as my interim research supervisor in the final months of my studies. Most of all I owe thanks to my wife Jasia, whose love and support allowed me to successfully complete this thesis. To Jasia, as well as all my family and friends, thanks for believing in me and helping me become successful in my academic pursuits. Special thanks also to my mentor and good friend Dr. David Godin, to whom I will be forever grateful for enlightening me with his limitless energy, inspiration, wisdom and faith. xvi I also want to thank Stephen Adams for his very enthusiastic scientific assistance throughout my studies, as well as Giorgio, Ting, Mohammed, Zuheir, Sharon and Eugene for providing an enjoyable laboratory environment. I would also like to acknowledge funding support from the program project grant GM32165 from the National Institutes of Health, Bethesda, MD, the Pharmaceutical Manufacturers Association of Canada/Medical Research Council and the University Graduate Fellowship Fund. xvii FOREWORD A paper describing this rabbit model of INH-induced hepatotoxicity has been published (Appendix 1): Sarich T.C., Zhou T., Adams S.P., Wall R.A. & Wright J.M. (1996) A Model of Isoniazid-induced Hepatotoxicity in Rabbits. J. Pharmacol. Toxicol. Meth., 34, 109-116. Although I did not perform the in vivo work or in vitro biochemical work in this paper, I did perform the histopathological analysis of the liver slides and the statistical analysis as well as write the first draft and edit the manuscript. A preliminary report of the data in "Study 1" has been published in abstract form: Sarich T.C., Adams S.P., Zhou T. & Wright J.M. (1993) Isoniazid-induced Hepatotoxicity in Rabbits: A Comparison of Necrosis and Steatosis. Clin. Invest. Med., 16 (Abstract), B67. T.C. Sarich and T. Zhou performed the in vivo animal work, T. Zhou performed the phenotyping analysis using HPLC, T.C. Sarich and S.P. Adams performed all in vitro work and T.C. Sarich performed all data analysis (including statistical analyses) and wrote the initial draft and edited the manuscript. A preliminary report of the data in "Study 2" has been published in abstract form: Sarich T.C, Adams S.P., Youssefi M., Zhou T., Wright J.M. (1994) Isoniazid-induced hepatotoxicity in rabbits: role of hydrazine. Can. J. Physiol. Pharmacol., 72(Suppl. 1): 585. The data are also published in a full paper: Sarich T.C, Youssefi M., Zhou T., Adams S.P., Wall R.A. & Wright J.M. (1996) The Role of Hydrazine in the Mechanism of Isoniazid Hepatotoxicity in Rabbits. Arch. Toxicol., 70, 835-840 (Appendix 2). In this study, T.C Sarich and T. Zhou performed the in vivo animal work, T.C. Sarich and S.P. Adams performed all in vitro work, M. Youssefi performed all HPLC analyses and T.C. xviii Sarich performed all data analyses (including statistical analyses), wrote the initial draft and edited the manuscript. 1 INTRODUCTION TUBERCULOSIS It is estimated that tuberculosis kills over 2.5 million persons per year worldwide; more than any other single infectious disease (Murray et al., 1990). Although Koch discovered that this infectious disease is caused b y M tuberculosis in 1882, no effective treatments were available until the discovery of anti-microbials in the 1940's. One of the early anti-tubercular drug discoveries was isoniazid (isonicotinic acid hydrazide; INH), which remains an effective first-line agent in the prevention and treatment of tuberculosis. Despite reports of drug-resistant strains, short-course chemotherapy including INH and other drugs remains effective in over 80% of patients (Murray et al., 1990). ISONIAZID Indications INH is widely used in the prophylaxis and treatment of tuberculosis. INH, rifampin, pyrazinamide and ethambutol are the major first-line agents in the treatment of tuberculosis in North America. The standard therapy for patients with active tuberculosis and sensitive strains of M. tuberculosis includes two months of INH, rifampin and pyrazinamide followed by four months of INH and rifampin (Marrie, 1995). INH alone for 12 months is the standard regimen for the prophylaxis of tuberculosis (Marrie, 1995). INH inhibits y-amino-butyric acid aminotransferase activity in the central nervous system and has been investigated in human neurological disorders where a y-amino-butyric acid deficiency is postulated to contribute to symptoms (Perry et al., 1979; Perry et al., 2 1981; Perry et al., 1982). INH is also indicated for use in severe action tremor in multiple sclerosis patients (Bozekefa/., 1987). Pharmacology The discovery of the inhibitory activity of INH against experimental tuberculosis was reported by Bernstein and coworkers in 1952. The drug is believed to act by inhibiting mycolic acid synthesis in the mycobacterial cell wall (Takayama etal., 1975; Davidson and Takayama, 1979). The specific inhibitory intracellular actions of INH on mycolic acid synthesis are not fully understood, but involve a unique catalase-peroxidase enzyme encoded by a katG gene (Zhang et al., 1992) and a binding target encoded by an inhA gene (Banerjee et al., 1994). Doses of INH used in humans for the prophylaxis and treatment of most tuberculosis infections vary from 0.036 to 0.072 mmoles/kg/day. In order to achieve therapeutic concentrations in the central nervous system, higher doses (up to 0.15 mmoles/kg/day) are used in children with tuberculous meningitis. Pharmacokinetics INH is completely absorbed following oral administration. INH readily gains access to all fluids and cells within the body. INH is inactivated by metabolism in the liver. After a single dose, INH and its metabolites can be almost completely accounted for in the urine within 24 hours. As shown below, a major step in the metabolism of INH (and some of its metabolites) is acetylation. The acetylation of INH by N-acetyltransferase exhibits genetic polymorphism; patients can be phenotyped into rapid (INH half-life of one hour) or slow (INH half-life of three hours) acetylators. The majority of Orientals, Inuits and American 3 Indians are rapid acetylators, whereas in Caucasians, Blacks and East Indians, slow acetylators are slightly more common than rapid acetylators. Inhibition and Induction of Hepatic Microsomal Enzymes INH is a potent inhibitor of hepatic cytochrome P-450s, including cytochrome P-450 (CYP) 2E1 in humans (Epstein et al, 1991). INH is known to inhibit the metabolism of phenytoin (Brennan et al, 1970) and carbamazepine (Wright et al, 1982) thus increasing their plasma concentrations when administered concomitantly with INH. This interaction can lead to significant drug toxicity. Low doses of INH (0.06 mmoles/kg/day for 10 days) have been shown to induce (increase) CYP2E1 activity in rats (Ryan et al, 1986). Two days after therapeutic doses of INH (approximately 0.03 mmoles/kg/day for seven days) CYP2E1 activity is increased in human slow acetylators (Zand et al, 1993; O'Shea et al, 1997) but not in human rapid acetylators (O'Shea et al, 1997). The most likely mechanism for this increased activity is that during the inhibition of CYP2E1 by INH, the enzyme is stabilised from degradation. When INH administration is stopped, the decreased INH concentration unmasks an increased number of active enzymes resulting in increased enzyme activity. Ingestion of acetaminophen during this period of increased CYP2E1 activity may lead to an increased susceptibility to acetaminophen-induced hepatotoxicity (Nolan et al, 1994). Toxicities Isoniazid has a low incidence of toxicity but adverse events include hepatotoxicity, peripheral neuropathy, convulsions, hematologic reactions, drug-induced lupus-erythematosus, mental abnormalities, rash and fever. Peripheral neuropathy can be 4 prevented or treated by administration of pyridoxine (25-100 mg daily). Hepatotoxicity is the most serious and problematic adverse effect of INH therapy. Isoniazid 0 ^ c ^ N H - N H 2 Acetylisoniazid ^NH-NH-C-CH Isonicotinic acid ^ O H N' NH 2 -NH 2 Hydrazine NH 2-NH-C-CH 3 Acetylhydrazine 1 1 CH3-C-NH-NH-C-CH3 Diacetylhydrazine 1 - N-acetyltransferase 2 - Amidohydrolase/Amidase HEPATOTOXICITY One of the most distinctive features of INH-induced hepatotoxicity is that it is not observed in individuals after taking an acute overdose, as, for example, is the case with acetaminophen-induced hepatotoxicity. In patients taking INH daily, a delay in onset, 5 usually a minimum of 3 to 5 days, is a characteristic feature of INH-induced hepatotoxicity. The onset is most frequent in the first two months of therapy, but longer delays have been reported. The reason for this delay is unknown, but suggests that a lag-time exists for the toxicity. Although there have been several theories proposed, the mechanism of INH-induced hepatotoxicity remains unknown. Historical Perspective After introduction of INH, a variety of case reports of liver toxicity associated with INH use appeared in the literature (Randolph and Joseph, 1953; Gellis and Murphy, 1955; Merritt and Fetter, 1959; Haber and Osborne, 1959; Cohen et al, 1961; Sharer and Smith, 1969; Martin and Arthaud, 1970). However, these cases most often involved individuals taking other medications concomitantly and the risk of INH-hepatotoxicity was not appreciated. In 1965, prophylactic use of INH for a duration of one year was recommended in individuals at risk of acquiring tuberculosis (American Thoracic Society, 1965). This program was later expanded to include patients with a positive tuberculin skin test who had not previously received INH (American Thoracic Society, 1967). For the first time, a population of patients taking INH alone existed. In February 1970, an outbreak of tuberculosis at Capital Hill (Washington DC, USA) resulted in the initiation of prophylactic INH therapy in 2321 employees (Garibaldi et al, 1972). Nineteen of these employees developed signs of severe liver toxicity and two died. This incident confirmed that INH can cause serious hepatotoxicity in a small percent (1-2%) of patients and sparked research efforts into studying the mechanism of INH-induced hepatotoxicity. 6 Pathology Daily administration of INH at usual therapeutic doses is associated with mild elevations of liver enzyme activities in plasma in up to 20% of patients (Mitchell et al, 1975a) and severe hepatotoxicity in approximately 1-2% of patients (Barlow et al., 1974). The histopathological manifestations range in severity from patchy focal necrosis of the liver to multilobular, bridging and massive necrosis (Maddrey and Boitnott, 1973). INH-induced liver injury is clinically, biochemically and histologically indistinguishable from viral hepatitis (Black etal, 1975). There have been at least two publications reporting INH-induced hepatotoxicity in which both hepatic necrosis and steatosis (fat accumulation) were documented in patients taking INH and rifampin (Pessayre et al., 1977; Pilheu et al, 1979). However, in the majority of reported cases, the major pathologic feature is hepatic necrosis (as determined histologically) or hepatic cell damage (as determined biochemically). Other Risk Factors The incidence of INH-induced hepatotoxicity increases with age, being rare in individuals.under 20 years of age, 0.3% between the ages of 20-34, 1.2% between the ages of 35-49 and up to 2.3% in persons over the age of 50 (Barlow et al, 1974). Although it has been proposed that in humans, females have a higher susceptibility to INH-induced hepatotoxicity than males (Moulding et al, 1989; Snider and Caras, 1992), this has not been substantiated. Other factors which may increase the risk of hepatotoxicity in patients taking INH include concomitant intake of alcohol or acetaminophen in large quantities. Daily 7 consumption of alcohol during INH therapy was associated with a four-fold increase in risk of INH-induced hepatotoxicity relative to non-drinkers (Kopanoff et al., 1978). Although the results in two additional studies did agree that alcoholics experience greater elevations of plasma liver enzymes (Gronhagen-Riska et al., 1978; Cross et al., 1980), one of the studies reported that this did not result in a greater number of severe adverse reactions in alcoholics (Cross etal., 1980). A single dose of acetaminophen as small as 3.25 grams taken by a patient receiving daily INH therapy caused severe hepatocellular injury (Crippin, 1993). Acetaminophen-induced hepatotoxicity during multiple drug therapy for tuberculosis (including INH) has also been reported to occur at doses of acetaminophen below those normally associated with hepatotoxicity (Nolan et al., 1994). The combination of INH and rifampin, the latter being an inducer of hepatic cytochrome P-450 isozymes of the 2C and 3 A families in human hepatocytes (Morel etal., 1990), has been reported to increase the risk of INH-induced hepatotoxicity (Pessayre et al., 1977; Gronhagen-Riska et al., 1978). A meta-analysis suggested that the incidence of hepatotoxicity increased to 2.6% in individuals receiving the INH/rifampin combination versus 1.6% in individuals receiving INH with other antitubercular drugs, not including rifampin (Steele et al, 1991). However, a randomised controlled trial comparing INH and INH-rifampin regimens showed decreased levels of transaminases in the INH-rifampin group as compared to the INH-only group (Hong Kong, 1992). The role of acetylator phenotype in the risk of INH-induced hepatotoxicity is controversial. Mitchell and coworkers (Mitchell et al., 1975b) originally proposed that 8 rapid acetylators were at significantly increased risk, based on a reportedly higher incidence of INH-induced hepatotoxicity in Orientals and a study showing a higher incidence of rapid acetylators in patients who had recovered from probable severe INH-induced hepatotoxicity. As indicated below, several studies published after 1975 do not indicate an increased risk of INH hepatotoxicity in rapid acetylators. Other studies have shown either no difference in the incidence of INH-induced hepatotoxicity between rapid and slow acetylators (Riska, 1976; Singapore, 1977; Hong Kong, 1977) or an increased incidence of INH-induced hepatotoxicity in slow acetylators (Musch et al., 1982; Eichelbaum et al., 1982). Studies of the metabolism of therapeutic doses of INH in humans show an increased urinary excretion of acetylhydrazine, an INH metabolite and suspected hepatotoxin, in slow acetylators as compared with rapid acetylators (Timbrell et al, 1977; Peretti et al., 1987). Rapid acetylators form more acetylhydrazine in vivo than slow acetylators but rapid acetylators also acetylate acetylhydrazine more rapidly and thus excrete a greater proportion of the parent drug as the non-toxic diacetylhydrazine than do slow acetylators (Timbrell et al., 1977). The lack of a greater exposure to acetylhydrazine in rapid acetylators has been confirmed in the study mentioned above in which plasma concentrations of acetylhydrazine were measured over time in 3 rapid and 3 slow acetylators (Lauterburg etal., 1985). HEPATOTOXICITY - ANIMAL MODELS There have been attempts in several animal species to create an experimental model which resembles INH-induced hepatotoxicity in humans. Some have doubted 9 whether a model can be found, since INH-induced hepatotoxicity has been referred to as a non-dose-dependent or idiosyncratic hepatotoxicity which is characteristically difficult to reproduce in animals. Early pharmacological and toxicological studies indicated that the major acute toxicity of high INH doses in mice, rats, rabbits and dogs was convulsions that preceded death (Benson et al, 1952). Subacute studies of INH resulted in slight hepatic damage in rats and fatty degeneration of the liver in dogs (Rubin et al, 1952). The appearance of liver toxicity in dogs at higher doses was also observed during long-term toxicity studies (Benson etal, 1952). Rats In rats, single doses of INH (2.2 mmoles/kg) do not cause hepatic injury (Mitchell et al, 1976). Although single-cell necrosis in phenobarbital pretreated rats was observed after administration of INH at a dose of 0.73 mmoles/kg every hour for six hours (Mitchell et al, 1976; Timbrell et al, 1980), the severity of the hepatic necrosis was difficult to interpret, since biochemical markers of hepatic cell damage (e.g., elevated liver enzyme activities in plasma) were not used to support the reported minor degree of histologically-eyident hepatic necrosis. There have been no subsequent reports of INH-induced hepatotoxicity in rats with acute, subacute or chronic INH dosing. In addition, unlike humans, rats do not have a genetically inherited acetylation polymorphism, so it is not possible to study the effect of acetylation on susceptibility to INH-induced hepatotoxicity in this species. 10 Guinea Pigs Daily dosing of guinea pigs with INH (0.073 to 0.73 mmoles/kg for up to 11 weeks) resulted in significant elevations of liver enzyme activities in plasma and, in the higher dose groups, histopathological evidence of inflammatory infiltration, focal areas of necrosis and focal and diffuse areas of fatty degeneration of the liver (Heisey et al, 1980). These pathological changes were interpreted by the authors to be mild hepatitis consistent with a hypersensitivity reaction rather than a toxic one. There are no other publications using this guinea pig model. Rabbits Single doses of INH (200 mg/kg, i.p.; 1.5 mmoles/kg) can cause hepatic fat accumulation (steatosis) in rabbits (McKennis et al., 1956). A repeated regimen of INH alone at 50 mg/kg/day (0.36 mmoles/kg/day; i.p.) for 11 days results in significant hepatic triglyceride accumulation (Karthikeyan and Krishnamoorthy, 1991). Although 10 mg/kg/day INH (0.073 mmoles/kg/day; i.p.) for 11 days did not result in increased hepatic triglyceride content, pretreatment with phenobarbital for four days at 60 mg/kg (0.26 mmoles/kg) followed by 10 mg/kg/day INH (0.073 mmoles/kg/day; i.p.) for 11 days did result in significant increases in hepatic triglyceride content (Krishnamoorthy and Karthikeyan, 1991). INH-induced hepatic fat accumulation in rabbits is also potentiated by pretreatment with ethanol (Whitehouse et al, 1978). The mechanism by which ethanol increases INH-induced hepatic fat accumulation is not known, but ethanol does not affect the half-life of INH or the percent excretion of isonicotinic acid, acetyl-INH or isonicotinoyl glycine in rabbits (Thomas etal., 1978). 11 The increased susceptibility of rabbits to INH-induced steatosis, as compared with rats, has been attributed to a greater hepatic amidase activity observed in vitro (Whitehouse et al., 1983) and in vivo (Thomas et al, 1984). A high activity of the amidase enzyme could result in increased conversion of INH to compounds with primary hydrazide functional groups, for example hydrazine and acetylhydrazine, which disrupt hepatocellular biochemistry and lead to fat accumulation in the liver (Whitehouse et al, 1983). There is evidence that hydrazine and various hydrazide compounds (benzhydrazide, phenacethydrazide, iproniazid, l-isonicotinyl-2-acetylhydrazine hydrochloric acid) produce fatty livers in rabbits (Yard and McKennis, 1955; McKennis etal, 1956). The evidence that rabbit liver undergoes fatty changes and fat accumulation when exposed to INH is convincing, but the occurrence of hepatic necrosis/hepatic cell damage in the aforementioned studies was not noted. However, INH does induce both histopathologically and biochemically (elevated marker liver enzymes in plasma) detectable hepatic necrosis/hepatic cell damage and signs of hepatic steatosis in greater than 50% of rabbits after repeated dosing over 33 hours (Sarich et al, 1995; Appendix 1). The hepatic cell damage in these animals is potentiated by pretreatment with phenobarbital. This rabbit model of INH-induced hepatotoxicity demonstrates features of the INH-induced hepatotoxicity observed in humans, including evidence of hepatic cell damage, a requirement for repeated dosing, a delay in onset of toxicity and interindividual variability in the severity of toxicity. As in humans, acetylation rate in rabbits is genetically determined and rabbits can be grouped as fast or slow acetylators of INH (Frymoyer and Jacox, 1963). Although humans and rabbits both demonstrate polymorphic N-12 acetyltransferase activities, the reason for this genetic polymorphism differs between these two species. In humans, slow acetylation is the result of a decrease in the quantity of two structurally and functionally similar JV-acetyltransferase enzymes (Grant et al, 1990; Sim and Hickman, 1991) whereas in rabbits, N-acetyltransferase specific mRNA and protein are missing (Blum et al., 1989). Therefore, although acetylation rates are phenotypically similar between rabbits and humans, the genetic reason for the polymorphism differs in each case. PROPOSED MECHANISMS Immunogenic Hypothesis Some of the earliest observations of INH-induced hepatotoxicity in humans led to the suggestion that the toxicity was due to an immune reaction (Assem et al., 1969; Martin and Arthaud, 1970). Support for this theory included no evidence of increasing hepatocellular damage with increasing dose or duration of drug therapy, as well as an observed recurrence of liver damage after rechallenge with the drug (Maddrey and Boitnott, 1973). However, other than the report of pathologic changes in guinea pigs which-are consistent with a-hypersensitivity-type reaction to INH (Heisey et al,, 1980), there is no further evidence that INH-induced hepatotoxicity is initiated via immune-mediated mechanisms. Acetylhydrazine Hypothesis Acetylhydrazine, a metabolite of INH, has been repeatedly implicated as the INH-derived hepatotoxin. This theory was proposed by Mitchell and coworkers when they showed that while INH itself did not consistently cause hepatic necrosis in rats and mice, 13 acetyl-INH and acetylhydrazine produced diffuse liver necrosis which was potentiated by pretreatment with the drug metabolism inducer phenobarbital and prevented by cobalt chloride, an inhibitor of drug metabolism (Mitchell etal, 1976). It was concluded that the liver damage was due to acetylhydrazine derived from acetyl-INH. Administration of Re-labeled acetylhydrazine to rats at doses of 0.27 mmoles/kg resulted in covalent binding of acetylhydrazine-derived reactive intermediates to hepatic macromolecules (Nelson et al, 1976a). Increases and decreases in the severity of hepatic necrosis were paralleled by increases and decreases, respectively, of covalent binding in the liver (Nelson et al, 1976a). This covalent binding to rat and human liver microsomes was also demonstrated in vitro and was presumably the result of oxidative activation of acetylhydrazine to reactive electrophilic acylating and alkylating species (Nelson et al., 1976b). Thus, a proposed mechanism of INH-induced hepatotoxicity involved acetylation of INH to acetyl-INH by Af-acetyltransferase, hydrolysis of acetyl-INH to isonicotinic acid and acetylhydrazine by an amidase and conversion of acetylhydrazine by the cytochrome P-450 enzyme system to reactive intermediates which covalently bind to intracellular proteins resulting in hepatic necrosis (Timbrell etal, 1980). Histologically detectable hepatic necrosis induced by acetylhydrazine (injected i.p.) has been variously reported to occur in phenobarbital-pretreated Sprague Dawley rats at 0.20 mmoles/kg (Nelson et al, 1975), 0.27 mmoles/kg (Mitchell et al, 1976), 0.7 mmoles/kg (Lauterburg et al, 1979), 0.09 to 0.45 mmoles/kg (Bahri et al, 1981), 0.09 to 0.45 mmoles/kg (Bahri et al, 1982) and 0.3 mmoles/kg (Woodward and Timbrell, 1984). In these studies, the method used to quantify histopathologically evident hepatic necrosis 14 was not detailed and none of the studies measured hepatic enzyme activities in plasma (Nelson et al, 1975; Mitchell et al, 1976; Lauterburg et al, 1979; Bahri et al, 1981; Bahri et al, 1982; Woodward and Timbrell, 1984). Subsequent studies have shown that administration of acetylhydrazine at doses as high as 4.2 mmoles/kg to phenobarbital-pretreated Sprague-Dawley rats does not cause significant elevation of aspartate aminotransferase and/or alanine aminotransferase (ALT) in plasma (Wright et al, 1986). In rats, there is evidence that the metabolism and hepatotoxicity of acetylhydrazine differs when acetylhydrazine is formed as a metabolite of INH versus administration of acetylhydrazine alone (Wright and Timbrell, 1978); that INH actually inhibits the microsomal enzyme-mediated covalent binding of acetylhydrazine to proteins in vitro (Timbrell and Wright, 1979); and that concomitant administration of INH and acetylhydrazine reduces the hepatotoxicity of acetylhydrazine (Bahri et al, 1981). It has also been shown that the covalent "acetyl" attachment to hepatic microsomal macromolecules is not necessarily related to pathological damage, since levels of covalent binding greater than those in the liver were found without the appearance of pathology in many tissues other than the liver (Woodward and Timbrell, 1984). Thus, the covalent binding of reactive metabolites (derived from acetylhydrazine) to cellular macromolecules cannot be used as a measure of pathological damage (Woodward and Timbrell 1984). Hydrazine Hypothesis Although acetylhydrazine has received most attention as the moiety responsible for INH-induced hepatotoxicity, hydrazine has also been implicated. Hydrazine, like INH, produces hepatic steatosis in rabbits on its own (Yard and McKennis, 1955; McKennis et 15 al, 1956; Nbda et al, 1983). Hydrazine also induces hepatic necrosis in rabbits and this hydrazine-induced hepatic necrosis is potentiated by rifampin pretreatment (Noda et al, 1983). Hydrazine-induced histological changes were not observed in rabbits treated with acetylhydrazine or diacetylhydrazine. Unfortunately, in this study it is not clear how histological data were evaluated. In monkeys, hydrazine [at doses ranging from 5 to 20 mg/kg (0.16 to 0.63 mmoles/kg)] has been shown to substantially increase liver enzymes and bilirubin in serum and triglyceride levels in the liver, myocardium, kidney and skeletal muscle, in addition to causing massive hepatic necrosis (Patrick and Back, 1965). Doses of 1.3 to 1.8 mmoles/kg hydrazine in rats cause an approximately 7-fold increase in hepatic triglyceride accumulation which decreased after pretreatment with phenobarbital (a classical hepatic enzyme inducer) and increased after pretreatment with piperonyl butoxide (a classical hepatic enzyme inhibitor) (Scales and Timbrell, 1982; Timbrell et al, 1982; Jenner and Timbrell, 1994a). Inhibition and potentiation of hepatic steatosis after pretreatment with phenobarbital (P-450 inducer) and piperonyl butoxide (P-450 inhibitor), respectively, suggests that hydrazine-induced fatty liver in rats is due to the parent compound rather than to a P-450-generated hydrazine metabolite. However, in rats pretreated with specific CYP2E1 inducers, acetone and INH, hydrazine-induced hepatic triglyceride accumulation is increased (Jenner and Timbrell, 1994a). This finding that CYP2E1 induction increases hydrazine-induced hepatic triglyceride accumulation suggests that a reactive metabolite of hydrazine formed by CYP2E1 causes triglyceride accumulation. Together, these studies suggest that administration of hydrazine produces toxicity in experimental animals with some similarity to that produced by INH. 16 Potentially hepatotoxic free radical intermediates of hydrazine metabolism have been identified. Active radical (•NHNH2) and diimide (NH=NH) metabolites were detected after incubation of rat microsomes with hydrazine (Noda et al., 1985). The conversion of hydrazine to the potentially toxic nitrogen radical, or diimide, has been shown to be catalyzed by NADPH cytochrome P-450 reductase (Noda et al, 1988). However, existence of the putative nitrogen-centered hydrazine radical has been disputed (Sinha, 1987). Instead, formation of a carbon-centered radical, most likely an acetyl radical, has been detected after perfusion of INH, acetylhydrazine and hydrazine through rat livers, suggesting that acetylation is necessary for reactive radical formation. Noda and colleagues reported an equal degree of toxicity in isolated rat hepatocytes after exposure to hydrazine or acetylhydrazine (based on the Trypan Blue viability test) (Noda et al., 1987). However, exposure of isolated rat hepatocytes to acetylhydrazine (3 mM) at 37°C did not result in cell death (measured by Trypan Blue exclusion) and lactate dehydrogenase release, whereas hydrazine (3 mM) caused significant cell death as manifest by an increase in lactate dehydrogenase release from 34.2 ± 2.0 to 48.4 ± 5.4 % (Wright J.M;, personal communication). Although this evidence suggests that both hydrazine and acetylhydrazine produce potentially hepatotoxic free radicals that are cytotoxic to hepatocytes in vitro, their respective roles in INH-induced hepatotoxicity in humans remain to be elucidated. Hydrazine, like acetylhydrazine, is a minor but significant INH metabolite. The 24-hour urinary excretion of hydrazine ranges from 0.4 ± 0.1 (mean ± standard error) to 1.0 ± 0.3 percent of a 300 mg (2.2 mmoles) dose of INH in rapid and poor acetylators, 17 respectively (Peretti etal, 1987). Hydrazine has been detected in the plasma of eight male volunteers taking INH [300 mg (2.2 mmoles)] and rifampin [600 mg (0.73 mmoles)] daily for fifteen days (Beever et al, 1982). In slow acetylators of INH, hydrazine plasma levels were significantly elevated at fifteen days. Sarma and coworkers (1986) have provided evidence of increased hydrolysis of INH to isonicotinic acid and hydrazine during combined INH and rifampin administration to slow acetylator volunteers and cite as supporting evidence that slow acetylators are at greater risk of INH-induced hepatotoxicity, especially when INH is combined with rifampin (Lai et al, 1972; Smith et al, 1972; Dickinson et al, 1977; Gronhagen-Riska et al, 1978; Musch et al, 1982; Parthasarathy et al, 1986). Based on the increased exposure of slow acetylators to hydrazine following administration of INH and rifampin, and the number of studies showing increased incidence of INH-induced hepatotoxicity in slow acetylators, the authors also dispute the proposed theory of an increased incidence of INH-induced hepatotoxicity in rapid acetylators due to increased exposure to acetylhydrazine as proposed by Mitchell and coworkers, 1975b. In fact, one year earlier, Mitchell and coworkers also provided evidence bearing on-the.postulated increase in. susceptibility of slow acetylators to INH-induced hepatotoxicity (Lauterburg et al, 1985). It has been shown that the percent 14CC>2 expired after administration of a 200 mg/kg (1.1 mmoles/kg) dose of 0.15 mCi/mole 14C-(acetyl)-labeled acetyl-INH to rats can be used as a measure of the in vivo exposure of humans to 14C-acetylhydrazine (Timbrell et al, 1980). In humans, increased exhalation of 14CC"2 from slow acetylators administered 12 mg (0.07 mmoles/kg) of 10jj.Ci 14C-(acetyl)-labeled acetyllNH versus rapid acetylators was interpreted as 18 greater in vivo exposure to acetylhydrazine, which agreed with reports of an increased incidence of INH-induced hepatotoxicity in slow acetylators. Mitchell and coworkers claimed that their original observation of increased (rather than decreased) incidence of INH-induced hepatotoxicity in rapid acetylators (Mitchell et al., 1975b) may have been confounded by high alcohol use in their previous study population (Lauterburg et al., 1985). Plasma hydrazine levels 3 hours following oral administration of a 200 mg (1.46 mmoles) dose of INH do not differ between rapid and slow acetylators (Gent etal., 1992), but urinary excretion of hydrazine has been observed to be greater in slow acetylators than in rapid acetylators 24 hours after a 300 mg (2.2 mmoles) dose of INH (Peretti et al., 1987). An important observation is that substantial amounts of urinary hydrazine and acetylhydrazine could be recovered in slow acetylators well beyond 24 hours (up to 36 hours) (Peretti et al., 1987). Incomplete elimination of hydrazine at 24 hours has been shown to contribute to a slowly rising baseline of hydrazine in plasma (Gent et al., 1992). For example, after the first dose of INH, plasma hydrazine levels 24 hours later ranged from undetectable to 9 ng hydrazine/mL (0.28 uM hydrazine) but after six weeks of INH, plasma hydrazine levels 24 hours after a dose of INH ranged from 8 to 33 ng hydrazine/mL (0.3 to 1.0 joM hydrazine) (Gent et al., 1992). The patient with the highest plasma hydrazine levels (33 ng hydrazine/mL; 1.0 uM hydrazine) after a period of six weeks of INH dosing also had elevated plasma transaminases and bilirubin levels (the next closest patient had a concentration of 18 ng hydrazine/mL; (0.6 uM hydrazine) and no reported elevations in transaminases or bilirubin). Evidence for a direct link between 19 plasma hydrazine levels and fatal INH-induced hepatotoxicity emerged from a case report of a 74-year old man with severe INH-induced hepatic cell damage (Woo et al, 1992). After five days of therapy with a standard regimen of antitubercular drugs including INH, the patient showed signs of liver toxicity. The patient died three days later and autopsy revealed severe hepatic necrosis. The patient's peak plasma hydrazine level (25 uM hydrazine) two days before death was greater than the 95th percentile of a group of elderly patients who did not develop hepatotoxicity. Although the accumulation of hydrazine in children resembles the situation in adults, the amount of INH converted to hydrazine and the rate of accumulation in relation to dosage of INH is lower in children (Gent et al, 1992). This is one of the first examples where a measurement in humans plus a mechanistic interpretation could potentially explain the epidemiologic observation that children are less susceptible than adults to INH-induced hepatotoxicity. 20 THESIS HYPOTHESES Introduction The overall purpose of this work was to improve our understanding of the mechanism of INH-induced hepatotoxicity in rabbits in the hope that it will provide insight into the mechanism of INH-induced hepatotoxicity in humans. At the start of my studies, there were many basic unanswered questions about rabbit models of INH-induced hepatotoxicity. Publication of the paper describing the model of INH-induced hepatotoxicity in rabbits (Sarich et al., 1995; Appendix 1) has established this model in the literature. STUDY 1 RATIONALE In our recently described rabbit model (Sarich et al., 1995), INH was shown to induce biochemically detectable hepatic cell damage and histologically detectable hepatic necrosis in rabbits. It has been previously demonstrated that INH-treated rabbits develop fatty changes including hepatic steatosis (increased hepatic triglycerides) and hypertriglyceridemia (increased plasma triglycerides) (McKennis et al., 1956; Karthikeyan and Krishnamoorthy 1991). Histologically detectable hepatic fat accumulation (hepatic steatosis) was observed in our previous study (Sarich et al., 1995); however, it was not quantified. INH-induced hepatotoxicity in humans has been reported to involve both necrosis and steatosis (Pessayre et al., 1977); however, the primary INH-induced insult is hepatic necrosis. Even though INH exposure to rabbits, and possibly humans, causes hepatotoxicity consisting of both hepatic necrosis and hepatic steatosis, the occurrence of 21 INH-induced hepatic necrosis and fatty changes, including hepatic steatosis and hypertriglyceridemia, has not been previously reported in an animal model. It has been proposed that in humans, females have a higher susceptibility to INH-induced hepatotoxicity than males (Moulding et al., 1989; Snider and Caras 1992). The effect of gender on susceptibility has also not been previously studied in an animal model. The rate of acetylation of INH has also been proposed to be a factor in determination of the risk of INH-induced hepatotoxicity; however, data in humans are contradictory, with published evidence indicating either that rapid (Mitchell et al., 1975b) or slow (Dickinson et al., 1981; Musch et al., 1982) acetylators are at increased risk. Using this rabbit model of INH-induced hepatotoxicity, it is possible to study the relationship between INH-induced hepatotoxicity and the rate of acetylation of sulfamethazine (a marker for the rate of acetylation of INH). If acetylation of INH is only a detoxification pathway, then a negative correlation between toxicity and acetylation rate would be expected. However, as seen in the diagram of INH-metabolism (p. 4), acetylation of INH also leads to production of toxic INH-metabolites. The objective in Study 1 was to determine the severity of INH-induced hepatic cell damage, hepatic steatosis and hypertriglyceridemia in male and female rabbits. The rabbits were also phenotyped to determine their rate of acetylation of sulfamethazine and this was compared to the severity of INH-induced pathological changes. 22 STUDY 1 HYPOTHESES • INH administration causes biochemically detectable hepatic cell damage and steatosis in male and female rabbits; female rabbits are more susceptible to these effects. • Rate of acetylation of sulfamethazine correlates with the severity of INH-induced hepatotoxicity. STUDY 2 RATIONALE Most of the previous research on the mechanism of INH-induced hepatotoxicity has focused on rat models with the hypothesis that acetylhydrazine, a metabolite of INH, is the cause of the hepatotoxicity (Nelson et al., 1976a; Mitchell et al., 1976; Timbrell et al., 1980). However, more recent evidence in animals and humans suggests that hydrazine is the metabolite that is predominantly responsible for INH-induced hepatotoxicity (Noda et al., 1983; Woo et al., 1992; Gent et al., 1992). The objective of this study was to describe the relationship between plasma concentrations of INH, acetylhydrazine and hydrazine, and markers of hepatic cell damage, hepatic steatosis and hypertriglyceridemia. This study was done with the hope of finding a clue as to whether any of these three compounds correlate with markers of toxicity. This would provide an avenue for further research into the mechanisms of toxicity. STUDY 2 HYPOTHESIS • Plasma levels of acetylhydrazine and/or hydrazine (suspected INH-derived hepatotoxins) correlate with markers of INH-induced hepatic cell damage. 23 STUDY 3 RATIONALE Study 2 outlined evidence supporting the involvement of hydrazine, or a metabolite(s) of hydrazine, as the INH-derived hepatotoxin(s) responsible for INH-induced hepatic cell damage in rabbits. In the present study, the metabolism of hydrazine was investigated. In rats, there is in vitro evidence that hydrazine is metabolised by NADPH cytochrome P-450 reductase (reductase) to reactive, potentially toxic, intermediates (Noda et al., 1988). Increased reductase activity may therefore be associated with increased conversion of hydrazine to toxic intermediates and thereby increase the severity of INH-induced hepatotoxicity. The state of hyperthyroidism is associated with a generalized increase in hepatic drug metabolic capacity (Shenfield, 1981). More specifically, studies in rats have shown that hepatic reductase activity can be decreased by administration of the anti-thyroid drug methimazole (Ram and Waxman, 1992) and increased by administration of L-thyroxine (Kato and Takahashi, 1968; Waxman et al., 1989). The goal of this study was to alter reductase activity in vivo, and therefore the production of hydrazine-derived reactive metabolic intermediates, by alteration of thyroid functional status in rabbits, and relate this to the severity of INH-induced hepatotoxicity. If the reductase enzyme converts hydrazine to reactive and toxic intermediates in vivo, a decrease in reductase activity (by methimazole pretreatment) should result in a decrease in the severity of INH-induced hepatic cell damage. Conversely, an increase in reductase 24 activity (by L-thyroxine pretreatment) should result in an increase in the severity of INH-induced hepatic cell damage. STUDY 3 HYPOTHESES • Decreased activity of the hepatic reductase enzyme (by methimazole pretreatment) should decrease the conversion of hydrazine to reactive and toxic intermediates in vivo and result in a decrease in the severity of INH-induced hepatic cell damage. • Conversely, increased reductase activity (by L-thyroxine pretreatment) should result in an increase in the conversion of hydrazine to reactive and toxic intermediates in vivo and an increase in the severity of INH-induced hepatic cell damage. STUDY 4 RATIONALE Hydrazine has been implicated as the causative hepatotoxin in INH-induced hepatotoxicity (Noda et al., 1983; Peretti et al., 1987; Woo et al., 1992; Gent et al., 1992). During metabolism of INH, hydrazine can be released by both direct and indirect pathways. The direct pathway involves hydrolysis of the amide bond of INH to produce isonicotinic acid and hydrazine. The indirect pathway involves acetylation of INH to acetyl-INH, hydrolysis of acetyl-INH to isonicotinic acid and acetylhydrazine and hydrolysis/deacetylation of acetylhydrazine to hydrazine. The hydrolysis reactions in both the direct and indirect pathways involve an amidohydrolase (amidase) enzyme. Previous studies in rats have shown that bis-/?-nitrophenyl phosphate (BNPP), an amidase inhibitor, prevented acetyl-INH-induced hepatic necrosis (Mitchell et al., 1976). 25 The investigators suggested this was due to prevention of hydrolysis of acetyl-INH to acetylhydrazine, the suspected INH-derived hepatotoxin in rats. Another possible explanation for their results is by inhibition of release of hydrazine (indirect pathway) from acetylhydrazine (derived from acetyl-INH). If BNPP inhibited INH-induced hepatic cell damage in our rabbit model in vivo, the mechanism is likely due to inhibition of the direct release of hydrazine from INH and the indirect release of hydrazine from acetyl-INH/acetylhydrazine. The primary goal of this study was to determine the involvement of the INH-amidase in the pathogenesis of INH-induced hepatotoxicity in rabbits by administration of the amidase inhibitor BNPP just prior (30 minutes) to each INH injection and by measuring the production of hydrazine in incubates of INH with hepatic microsomes or plasma. STUDY 4 HYPOTHESIS • Inhibition of the activity of INH-amidase using the amidase inhibitor BNPP decreases the conversion of hydrazine from INH and decreases the severity of INH-induced hepatic cell damage. 26 METHODS ANIMALS The use of rabbits for the research of INH-induced hepatotoxicity was reviewed and approved by the University of British Columbia (UBC) Committee on Animal Care. Rabbits were housed in stainless steel cages (one or two animals per cage) with a 12 hour light/dark cycle and free access to food and water. For Study 1, sixteen New Zealand white rabbits (eight male, eight female) weighing 2-3 kg, obtained from the UBC Animal Care Unit, were used. In Study 2, toxicity data and plasma samples (analysed for plasma concentrations of INH, acetylhydrazine and hydrazine) were obtained from the rabbits used in Study 1. For Studies 3 and 4, a total of 149 male New Zealand white rabbits weighing 2-3 kg were obtained from the UBC Animal Care Unit and used. ANIMAL TREATMENTS TABLE 1.1 DOSING SCHEDULE FOR INH TIME DAY 1 DAY 2 DAY 3 DAY 4 DAY 5 08:00 Baseline Blood 24 Hour Blood 48 Hour Blood 72 Hour Blood 96 Hour Blood 09:00 INH injection INH injection Microsomes 12:00 INH injection INH injection 15:00 INH injection INH injection 17:00 32 Hour Blood 56 Hour Blood 80 Hour Blood 18:00 INH injection INH injection INH-Injection Protocol The dosing schedule for INH (Sigma Chemicals) was the same as described previously (Sarich et al, 1995): day one involved a subcutaneous injection (s.c.) of 0.37 27 mmoles/kg (50 mg/kg) followed by three 0.26 mmole/kg (35 mg/kg) injections at three hour intervals. On day 2, the injection protocol from day one was repeated (Table 1.1). Hypothyroidism In Study 3, we attempted to create hypothyroidism in rabbits, using methimazole, to try to decrease hepatic reductase activity. Treatment with 0.025% (w/v) methimazole in the drinking water for 16-24 days (approximately 20 mg/kg/day; 0.18 mmoles/kg/day) has been shown to decrease hepatic reductase activity by 80-85% in rats (Ram and Waxman, 1992). However, there have been no reports of methimazole-induced hypothyroidism in rabbits. The following pilot studies used methimazole (Sigma Chemicals) (dissolved in 0.02 M NH4OH; pH 10.3) injections (i.p.) to decrease plasma free T4 levels and hepatic reductase activities: 10 mg/kg/day (0.09 mmoles/kg/day) for seven days, 2x50 mg/kg/day (2x0.44 mmoles/kg/day; 12 hour intervals) for nine days and 2x150 mg/kg/day (2x1.3 mmoles/kg/day; 12 hour intervals) for nine days. Based on the results of the pilot studies, methimazole at a dose of 2x50 mg/kg/day (2x0.44 mmoles/kg/day; 12 hour intervals) was given for nine days prior to INH. This regimen was chosen as being effective in producing hypothyroidism and was combined with the INH-injection protocol. Methimazole was also added to drinking water at doses of 0.0313% w/v (approximately 25 mg/kg/day; 0.22 mmoles/kg/day) and 0.0626% w/v (approximately 50 mg/kg/day; 0.44 mmoles/kg/day) for 26 days followed by the INH-injection protocol. A total of nine control animals were matched to the methimazole treatment groups by having vehicle injections instead of methimazole injections, or by 28 being provided with regular drinking water instead of methimazole-containing water. These animals received INH and were included as part of the "INH-only" group. Hyperthyroidism A rabbit model of hyperthyroidism involving daily i.p. injections of L-thyroxine (0.4 mg/kg; 5.1 lumoles/kg) for 5 to 7 days has been reported (Shimoni and Banno, 1993). In Study 3, a pilot study using i.p. injections of L-thyroxine (Sigma Chemicals) (dissolved in 0.02 M N H 4 O H ; pH 10.3) at a dose of 0.4 mg/kg/day (5.1 u,moles/kg/day) for seven days to increase plasma free T4 levels and hepatic reductase activities was carried out. The protocol followed for production of hyperthyroidism in 13 rabbits involved i.p. injections of L-thyroxine (dissolved in 0.02 M NH4OH) at 0.4 mg/kg/day (5.1 umoles/kg/day) for 9 days. In these rabbits, since a state of hyperthyroidism was predicted to be reached by day 6, the INH injections were begun on day 6. These animals are referred to as the "T4-INH" group. Thirteen control animals were treated exactly as the L-thyroxine-treated animals except that instead of L-thyroxine, they were injected with vehicle (0.02 M NH4OH). These animals were included with the methimazole controls as part of the "INH-only" group for data analysis: In addition to the hypothyroid and hyperthyroid groups, nine animals were treated as vehicle controls (five as L-thyroxine controls and four as methimazole controls) and did not receive INH. These animals were sacrificed on the morning of the day they would have received the INH protocol and are referred to as "vehicle control" animals. The rabbits were kept and treated in groups of ten. Since the L-thyroxine and methimazole pretreatments were different, each group of L-thyroxine or methimazole-29 treated animals were matched to a control group by randomising groups of ten animals into either L-thyroxine or its vehicle, or methimazole, or its vehicle. Bis-/>Nitrophenyl Phosphate (BNPP) Treatment In study 4, animals were randomised into one of five groups, each involving two treatments. The first treatment involved either a BNPP (Sigma Chemicals) injection of 25 mg/kg (0.074 mmoles/kg) or a BNPP-vehicle injection. BNPP was dissolved in saline (0.9% NaCl) at 7.65 mg/mL after heating to approximately 60°C to make soluble (Heymann and Krisch, 1969) and, after cooling, injected i.p. at a volume of 3.3 mL/kg. The second treatment involved the INH-injection protocol (see above) and animals received subcutaneous injections of either INH or INH-vehicle (0.9% NaCl; saline). The first treatment (BNPP or BNPP-vehicle) was administered 30 minutes prior to each of the INH or INH-vehicle injections. In total there were five groups which included animals receiving BNPP-vehicle and INH-vehicle (VEH-VEH; control group), BNPP-vehicle and INH (VEH-INH; active toxicity group), BNPP and INH (BNPP-INH; BNPP test group), BNPP and INH-vehicle (BNPP-VEH; BNPP control group), and a final group of BNPP-vehicle. and-INH-vehicle (VEH-VEH*; food and water control group) which had treatments delayed one day in order to match their food and water intake to the VEH-INH animals (Table 1.2). During the INH-injection protocol, food intake in the VEH-INH group ranged from 20-40 g/day and water intake ranged from 50-100 mL/day. The VEH-VEH* animals were allowed only the amount of food and water (determined for each animal in the VEH-VEH* group individually based on its weight-matched counterpart in the VEH-INH group) ingested by 30 the animals in the VEH-INH group. The purpose of this group was to control for any pathological or biochemical changes that may occur as a result of decreased food and water intake which we have observed in INH-treated animals. TABLE 1.2 TREATMENT GROUP DEFINITIONS, ABBREVIATIONS AND DOSING TIMES GROUP 1 GROUP 2 GROUP 3 GROUP 4 GROUP 5* Number 12 17 17 12 12 Abbreviations VEH-VEH VEH-INH BNPP-INH BNPP-VEH VEH-VEH* Doses at 08:30, 11:30, 14:30 & 17:30 VEHICLE VEHICLE BNPP BNPP VEHICLE Doses at 09:00, 12:00, 15:00 & 18:00 VEHICLE INH INH VEHICLE VEHICLE * This group was delayed by one day and allowed only the amount of food and water consumed by group 2, the INH-only group. The animals in group 5 were matched by weight to the animals in group 2. Group Definitions: VEH-VEH - Control group; VEH-INH - INH only group; BNPP-INH -animals received BNPP followed by INH; BNPP-VEH - animals received BNPP only; VEH-VEH* -food and water control group. TABLE 1.3 BIS-^-NITROPHENYL PHOSPHATE AND INH INJECTION PROTOCOL TIME DAY 1 DAY 2 DAY 3 08:00 Baseline (0 hour) Blood 24 Hour Blood 48 Hour Blood 08:30 BNPP injection (i.p.) BNPP injection (i.p.) Rabbits killed 09:00 INH injection (s.c.) INH injection (s.c.) Microsomes prepared 11:30 BNPP injection (i.p.) BNPP injection (i.p.) 12:00 INH injection (s.c.) INH injection (s.c.) 14:30 BNPP injection (i.p.) BNPP injection (i.p.) 15:00 INH injection (s.c.) INH injection (s.c.) 17:00 32 Hour Blood 17:30 BNPP injection (i.p.) BNPP injection (i.p.) 18:00 INH injection (s.c.) INH injection (s.c.) 21:00 12 Hour Blood The rabbits were housed and treated in groups of ten animals. The first six groups of ten included two animals for each treatment group. The final group of ten animals 31 (numbers 61-70) included only the VEH-INH and BNPP-INH groups (five of each). A complete treatment and sampling schedule for Study 4 is shown in Table 1.3. ANIMAL HANDLING Blood Sampling Blood samples (1 mL) were taken from the lateral ear vein using a heparinised syringe. Topical administration of xylene (Fisher) was used to dilate the vein during blood sampling. The xylene solution was removed shortly after application by wiping the ear with a moistened swab. Plasma was isolated from the blood samples by centrifugation and frozen (-60 °C) until analysis. For Studies 1 and 2, blood samples were taken before (0 hour), during (32 hours) and at various times (48, 56, 72, 80 and 96 hours) following INH administration (Table 1.1). In study 3, blood samples were taken at 0, 24, 32, 48, 56, 72, 80 and 96 hours (Table 1.1). The rabbits were sacrificed by cervical dislocation, followed by exsanguination after the final (96 hour) blood sample was obtained. In Study 4, blood samples were taken at 0, 12, 24, 32 and 48 hours. The animals were killed at 48 hours by cervical dislocation and exsanguination (Table 1.3). LIVER HANDLING Crude Homogenate In Study 1, immediately after death the livers were removed, weighed, perfused via the portal Vein with 0.25M sucrose in 5 mM Tris buffer (pH 8.0), homogenized (with a glass tube and Teflon pestle) using three volumes of the perfusion buffer per gram of liver and frozen until analysis (-60°C) (Touster et al, 1970). 32 Microsome Preparation Immediately after sacrifice of the rabbits in Study 3 (96 hours) and Study 4 (48 hours), livers were removed, weighed and homogenized (using a glass tube and Teflon pestle) with a homogenizing buffer (1.15% KCL, 10 mM EDTA, 10 mM phosphate, pH 7.4). The crude homogenate was centrifuged at 10,000xg for 20 minutes. The supernatant was then centrifuged at 105,000xg for 60 minutes and the resulting pellet was washed by resuspending in the homogenizing buffer and recentrifuging at 105,000xg for 60 minutes. The final pellet was resuspended in a microsomal storage buffer (20% glycerol, 1.15% KCL, 10 mM EDTA, 10 mM phosphate, pH 7.4) and frozen at -60°C prior to analysis. Whole Liver Sample Freezing In Study 4, samples of liver were placed in vials and quick-frozen in liquid nitrogen for analysis of hepatic glutathione and thiobarbituric acid reactive substances. BIOCHEMICAL ANALYSES Argininosuccinic Acid Lyase (ASAL) Activity ASAL activity in the plasma is a sensitive marker of liver damage (Campanini et al, 1970; Sims and Rautanen 1975) and was used instead of alanine aminotransferase (ALT) as described elsewhere (Sarich et al, 1995). The ASAL enzyme is found specifically in liver parenchyma and acts in the urea cycle by catalyzing the conversion of argininosuccinic acid to arginine and fumaric acid. The quantitation of plasma ASAL activity (expressed as urnoles/100 mL/hour; Takahara units) was assayed according to Campanini et al, (1970) with modifications as described in Sarich et al, (1995) and described as follows. To both unknown and control tubes, 0.1 mL plasma was added 33 followed by addition of 0.3 mL sodium argininosuccinic acid (0.65 mM; substrate) to the unknown tube. The unknown tube was incubated for one hour at 37°C. To both tubes 0.2 mL 20% trichloroacetic acid solution was then added. After standing for 5 minutes, a 0.5 mL aliquot Was taken from each tube and added to separate tubes. Added to these tubes were 0.1 mL of 10% NaOH, 0.25 mL of the dichloronaphthol solution (0.03%) and 0.1 mL of the NaOCl solution (0.63%). The tubes were allowed to stand in an ice bath for 15 minutes and the color read at 515 nm using an LKB 4050 Ultrospec spectrophotometer (LKB Biochrom Ltd. Cambridge). Plasma ASAL activities were logarithmically-transformed for statistical and correlational analyses so as to fit data to a normal distribution. Baseline plasma ASAL activities were derived from pretreatment (0 hour) plasma samples in each animal. In Study 2, hepatic cell damage data is presented as log peak plasma ASAL activity, plasma ASAL activity area under the curve and log plasma ASAL activity at 48 hours. Although these three different presentation methods for plasma ASAL activity are not entirely independent of each other, the use of all three provides greater insight into the pattern and severity of INH-induced hepatotoxicity. Plasma ASAL activity and triglyceride area under the curve values were calculated by weighing (mg paper) the cut out area under the curve from graphs with standard axes on standard paper by an individual who was blinded to the treatment regimen. The area under the curve values were also determined by addition of trapezoidal areas. A high correlation coefficient was obtained when the areas using the trapezoidal technique were correlated with the areas obtained using the paper cut-and-weigh technique (r = 0.96, p < 0.0001, n = 15). 34 Reagents for the (ASAL) assay were obtained as follows: barium argininosuccinate (which was converted to the sodium salt by admixture with sodium sulfate and centrifugation) and 2,4-dichloro-l-naphthol from Sigma; the other reagents required for the ASAL assay were obtained from local chemical suppliers and were all of reagent grade. Alanine Aminotransferase (ALT) Activity A commonly used marker of hepatic necrosis, plasma ALT activity (Units/L), was determined using a kit from Sigma Diagnostics (kit # 59-20). Liver and Plasma Triglyceride Determinations Hepatic triglyceride accumulation (hepatic steatosis) was quantitated by analysis of the triglyceride content of liver crude homogenates using a Folch extraction (Folch et al, 1951) and modified as follows: 0.1 mL of diluted crude homogenate was added to 0.5 mL 2:1 chloroform:methanol in a test tube which was covered, placed on ice and vortexed every 10 minutes for a total of 60 minutes. The mixture was then centrifuged for 5 minutes at l,000xg resulting in the formation of two distinct layers. A 0.3 mL aliquot (from approximately 0.35 mL total) from the bottom layer was removed' and evaporated to dryness using a boiling water bath. The resulting dried layer was analyzed for triglyceride content using a triglyceride kit from Sigma Diagnostics (Kit number 336-10). The triglyceride level in the liver tissue was expressed as mg triglyceride (triolein equivalent) / g liver tissue^  In Study 1, control hepatic triglyceride levels were obtained from the livers of four untreated rabbits. 35 Plasma triglyceride levels (mM triolein equivalent) were quantitated using a triglyceride kit from Sigma Diagnostics (Kit number 336-10). Baseline plasma triglyceride concentrations were determined in pretreatment (0 hour) plasma samples in each animal. Plasma Free T4 Levels Plasma free T4 levels (pmoles/L) in Study 3 were monitored to evaluate the effectiveness of the drug-induced states of experimental hypo- and hyper-thyroidism. The plasma free T4 levels were determined at Vancouver Hospital and Health Sciences Center - UBC Site using a Microparticle Enzyme Immunoassay. A limitation of the microenzyme assay procedure prevented dilution of samples for determination of plasma free T4 levels above 77.2 pmoles/L. Protein Determination Protein quantitation of liver crude homogenate samples and microsomes were carried out using methods described by Bradford (1976). Microsomal P-450 and Enzyme Activity Determination Cytochrome P-450 levels (nmoles P-450/mg protein) were determined using the CO difference spectra with reference wavelengths-of 450 nm and-490 nm, an extinction coefficient of 91 cm"1 mM 1 as outlined in Omura and Sato (1964) and band widths of 2 nm. Reductase activity (nmoles/minute/mg protein) was determined using procedures outlined by Phillips and Langdon (1962), using an extinction coefficient of 19.6 cm'1 mM"1 and a final reaction volume of 1.575 mL. /?-Nitrophenol hydroxylase activity (nmoles/minute/mg protein) (a marker of CYP2E1 activity in rabbit microsomes, Koop et al, 1989) was assayed using an initial substrate concentration (/?-nitrophenol; Sigma 36 Chemicals) of 100 uM, a peak absorbance of 510 nm and an extinction coefficient of 9.53 cm"1 mM"1 (Reinke and Moyer, 1985; Koop, 1986; Jenner and Timbrell, 1994a). Thiobarbituric Acid Reactive Substances (TBARS) Determination Hepatic TBARS were determined using methods as previously described (Tappel and Zalkin 1959; Godin etal., 1989). One gram of thawed liver tissue was homogenised in 10 mL cold 50 mM Tris-0.1 mMEDTA (pH 7.6) buffer. One-half milliliter of homogenate was combined with 0.5 mL /-butyl hydroperoxide (f-BHP; Sigma Chemicals) at 0 mM, 0.1 mM, 0.25 mM, 0.5 mM, 1.0 mM and 5.0 mM concentrations in a solution of 0.9% NaCI -2 mM sodium azide (Sigma Chemical). This mixture was incubated for 30 minutes at 37°C. The reaction was stopped by adding 0.5 mL 0.1 M sodium arsenite (Sigma Chemicals) in 28% TCA (w/v) and the solution was centrifuged (12,700xg) for 5 minutes. Thiobarbituric acid (0.5% in 0.025 M NaOH; Sigma Chemicals) (0.5 mL) was combined with 1.0 mL of the supernatant and the mixture boiled for 15 minutes. Absorbance of the chromophore in the final solution, which is believed to be a malondialdehyde-thiobarbituric acid adduct (Valenzuela, 1991), was measured at 532 nm using an LKB 4050 Ultrospec spectrophotometer (LKB Biochrom Ltd. Cambridge). Tissue Glutathione Determination Analytical procedure Tissue glutathione was determined according to a methods previously described (Ellman, 1959; Moron etal, 1979; Wohaieb and Godin, 1987). One gram of thawed liver tissue was homogenised in 10 mL cold 50 mM Tris-0.1 mM EDTA (pH 7.6) buffer and kept on ice for a maximum of 20 minutes prior to denaturation. One milliliter of 37 homogenate was combined with 0.25 mL cold 25% trichloroacetic acid (TCA; Sigma Chemicals), followed by centrifugation (12,700xg) for 5 minutes. In the sample cuvette, 1.0 mL 0.15 M imidazole buffer (pH 7.4) (Sigma Chemicals), 1.7 mL water, 0.2 mL supernatant and 0.1 mL 3 mM 5,5'-dithiobis(2-nitrobenzoic) acid (Sigma Chemicals), to start the reaction, were combined. In the reference cuvette, 5% TCA was added in place of supernatant. Absorbance (which reflects the amount of TCA soluble non-protein thiols (reduced form) and is due to a 2-nitrobenzoate-5-thiophenolate dianion) was measured at 412 nm 3 minutes after addition of 5,5'-dithiobis(2-nitrobenzoic) acid using an LKB 4050 Ultrospec spectrophotometer (LKB Biochrom Ltd. Cambridge). The absorbances were compared to a standard curve of pure glutathione. All standards and samples were run in triplicate. Units were expressed as u,moles glutathione/gram liver. TABLE 1.4 VALIDATION SUMMARY - GLUTATHIONE ASSAY Limit of Quantitation* 10 nmoles (50 LUM) Standards Range 10 - 300 nmoles (50 - 1500 uM) Sample Range 37 - 156 nmoles (184 - 778 uM) Accuracy Not determined Specificity Non-specific measure of sulfhydryl groups only Linearity: In liver: r = 0.999; slope = 4.4 x 1 0 3 (4.3 x lO-3 to4.5 x 10 3 )* V-int = 1.1 x 10"2 (-6.4 x 1 0 3 to 2.9 x 10"2) Calculation Slope (b = 0) 1 r = 0.999; slope = 4.5 x 10'3 (4.4 x 10"3 to 4.6 x 10"3) * Limit of Quantitation: Defined as the lowest concentration quantified using this procedure. ** Numbers in Brackets are 95% Confidence Intervals. t Calculation slope refers to the slope from the standard curve which was used to calculate the concentration of glutathione in the samples. Using pure glutathione (Sigma Chemicals) a standard curve was constructed (in triplicate) at 10, 50, 100, 200 and 300 nmoles glutathione. The standard curve gave a correlation coefficient (r) of 0.999, a slope of 4.4 x 10"3 (95% confidence intervals 4.3 x 38 10"3 to 4.5 x 10"3) absorbance units/nmole glutathione and a y-intercept of 1.1 x 10'2 (95% confidence intervals -6.4 x 10'3 to 2.9 x 10"2) absorbance units (Figure 1.1; page 75). Using the equation y = mx, the standard curve has a correlation coefficient (r) of 0.999 and a slope of 4.5 x 10"3 (95% confidence intervals 4.4 x 10"3 to 4.6 x 10'3) absorbance units/uM glutathione. A validation summary is given in Table 1.4 and an estimation of precision in Table 1.5. TABLE 1.5 PRECISION OF THE PROCEDURE FOR DETERMINATION OF GLUTATHIONE IN LIVER HOMOGENATE. In Liver Homogenate nmoles (n = 3) CV%(SD/mean) 10 2.1 50 0.9 100 0.6 200 0.9 300 0.4 (mean ± SD) 1.0 ±0 .7% FLUOROMETRY Resorufin Assay The fluorometric determination of resorufin concentrations in microsomal incubates is a common procedure for the determination of specific P-450 microsomal isozyme activities. Incubation of specific alkyloxy-resorufin derivatives with microsomes provides measures of activity of cytochromes P-450 1A1/2, 2B4 and 2B4/5. The rate of dealkylation of these resorufin analogues to resorufin as detected by the increased fluorescence of resorufin in the incubation mixture is an estimate of isozyme activity. 39 Analytical procedure Microsomal ethoxyresorufin deethylase (EROD) activity (CYP1A1/2), benzyloxyresorufin dealkylase (BROD) activity (CYP2B4) and pentoxyresorufin dealkylase (PROD) activity (CYP2B4/5) were determined as described previously (Burke et al., 1985; Grimm et al., 1994). The specificity of benzyloxyresorufin dealkylase (BROD) activity (CYP2B4) and pentoxyresorufin dealkylase (PROD) activity (CYP2B4/5) have been tested in rabbits; however, the specificity of ethoxyresorufin deethylase (EROD) activity (CYP1A1/2) in rabbits has not. An increase in fluorescence produced as a result of de-alkylation of resorufin precursors to resorufin corresponds to activities of resorufin-precursor specific P-450 isozymes. Scans were performed on resorufin standard curve solutions to identify the peak excitation and emission wavelengths of resorufin. A peak excitation wavelength of 575 nm and a peak emission wavelength of 585 nm were identified (Figure 1.2; page 76). A standard curve using resorufin (in triplicate) was constructed at 0.1, 0.25, 0.5, 1, 2 and 5 nmoles resorufin. The standard curve gave a correlation coefficient (r) of 0.999, a slope of 0.322 (95% confidence intervals 0.312 to 0.332) fluorescence units/nmole resorufin.and-a y-intercept of 5.6 x 10*3 (95% confidence intervals -1.6 x 10"2 to 2.7 x 10'2) fluorescence units (Figure 1.3; page 77). Using the equation y = mx, the standard curve has a correlation coefficient (r) of 0.999 and a slope of 0.324 (95% confidence intervals 0.317 to 0.331) fluorescence units/ nmole resorufin. A validation summary (Table 1.6) and an estimation of precision (Table 1.7) are shown. 40 Microsomes (50 uL) were mixed with buffer (1.93 mL of 50 mM HEPES, pH 7.5, 0.1 mM EDTA, 15 mM MgCb) and substrate (10 uL of a 1 mM solution in dimethylsulfoxide) and incubated at 37°C for 2 minutes. Following this pre-incubation, NADPH (10 uL of a 100 mM solution in buffer) was added, the solution was mixed and the fluorescence was recorded at 5 minutes. The analyses were performed on a Shimadzu recording spectrofluorophotometer (model RF-540) with a DR-3 recorder. The slit width for both the excitation and emission beams was 2 nm. Samples were analysed in duplicate. Activity units were expressed as nmole resorufin produced/min/mg protein. TABLE 1.6 VALIDATION SUMMARY - RESORUFIN ASSAY Limit of Quantitation 0.1 nmoles Standards Range 0.1-5 nmoles Sample Range BROD assay 0.1 - 3.5 nmoles; EROD assay 0.1 -1.7 nmoles; PROD assay 0.1 - 0.3 nmoles Accuracy Not determined Specificity Not determined Linearity: In microsomes: r = 0.999; slope = 0.322 (0.312 to 0.332); y-int = 6.0 x 10"3 (-1.6 x 10'2 to 2.7 x 102) Calculation Slope (b = 0)T r = 0.999; slope = 0.324 (0.317 to 0.331) * Calculation slope refers to the slope from the standard curve which was used to calculate the concentration of resorufin in the samples. Although the excitation wavelength of 575 nm used in this study differs from wavelengths of 530 nm (Burke et al, 1985) and 550 nm (Grimm et al, 1994) reported previously in the literature, the scans provided (Figure 1.2; page 76) show a peak excitation of 575 nm and demonstrate no interference of the light-scattering peak which appears from approximately 580-600 nm. Using 2 nm slit widths allowed resolution of the peak at 575 nm. Normally, it would be preferable to use a larger difference between excitation and emission wavelengths. Careful analyses were therefore undertaken to ensure adequate resolution between the excitation of resorufin and the light scattering peak. In this instance, the utilisation of an excitation peak of 575 nm provided a 3-fold increase in sensitivity (as compared to other data generated in the same laboratory using an excitation wavelength of 530 nm) without apparent loss of specificity. In Figure 1.4 (page 78), a standard curve derived using an excitation wavelength of 575 nm is compared to a standard curve derived using an excitation wavelength of 550 nm. The increased slope achieved using an excitation wavelength of 575 nm [r = 0.999; slope = 0.292 (95% confidence intervals 0.25 to 0.335) fluorescence units/nmole resorufin; y-intercept = -0.05 (95% confidence intervals -0.16 to 0.062)] demonstrates the improved sensitivity of the assay as compared to the standard curve using an excitation wavelength of 550 nm [r = 0.999; slope = 0.200 (95% confidence intervals 0.17 to 0.23) fluorescence units/nmole resorufin; y-intercept = -0.036 (95% confidence intervals -0.12 to 0.051)]. TABLE 1.7 PRECISION OF THE PROCEDURE FOR DETERMINATION OF RESORUFIN IN MICROSOMES In Microsomes nmoles (n = 3) CV % (SD/mean) 0.1 5.0 0.25 5.6 0.5 8.4 1.0 0.1 1.5 7.0 2.0 5.9 5.0 1.5 (mean ± SD) 4.8 ± 3.0% 42 HIGH PERFORMANCE LIQUID CHROMATOGRAPHY (HPLC) ANALYSES Summary of Hydrazine Assays Used In Study 2, plasma samples were analysed for INH, acetylhydrazine and hydrazine concentrations using an HPLC gradient mobile phase system with two pumps and a column oven. In Study 4, HPLC analyses of plasma hydrazine were performed using an isocratic system with one pump and no requirement for a column oven. An assay using propanol as the plasma denaturant was used to quantitate hydrazine levels in plasma samples taken at 12 hours. After hydrazine was quantitated in the 12 hour plasma samples, it was thought that HPLC column life could be increased through the use of a denaturation procedure using acetonitrile and perchloric acid instead of propanol. This technique was used for the determination of plasma hydrazine in 24 and 32 hour plasma samples as well as for the determination of hepatic amidase activity. All of these procedures are described in detail as follows. HPLC Analysis of INH, Acetylhydrazine and Hydrazine The analytical procedure used for the quantitation of INH, acetylhydrazine and hydrazine in 32 and 48 hour plasma samples in Study 2 was developed in this laboratory (Youssefi, M.Sc. Thesis, 1992). The derivatising reagent was made up by adding 250 uL of 3-methoxybenzaldehyde and 3.75 mL of formic acid (minimum assay 88% purity) to a 50 mL volumetric flask and making up to 50 mL with propanol. Five mL of 10 mM 9-fluorenone solution (in propanol) was included in this solution. The P-fluorenone served as an internal check on volumetric accuracy. The final solution contained 41.1 mM 3-methoxybenzaldehyde, 1 mM P-fluorenone and 7.5% formic acid (v/v). The solution was 43 transferred to a dark bottle and kept in the dark when not in use. Previous studies have confirmed that the formic acid does not hydrolyse INH (Youssefi, 1992). In order to deproteinize plasma samples were mixed with an equal volume of the solvent (propanol). This mixture was vortexed vigorously and allowed to stand for 2 minutes. The suspension was diluted to twice the original plasma volume with HPLC grade water and allowed to stand for 2 minutes. This solution was then centrifuged in a microcentrifuge (12,700xg) for 2 minutes. The resulting supernatant, containing INH, acetylhydrazine and hydrazine, was filtered to remove any remaining large particles with a 3 mm diameter syringe filter (0.5 um; Gelman). The filtered supernatant was mixed with the derivatising reagent at a ratio of 4:1. This mixture was allowed to stand at room temperature for 2 hours in the dark. The acid present in the derivatising reagent was sufficient to hydrolyze any a-ketoacid hydrazones of INH and acetylhydrazine and a-ketoacid azines of hydrazine present in plasma (Timbrell and Wright, 1984). The INH, acetylhydrazine and hydrazine released in these reactions then reacted with the large excess of 3-methoxybenzaldehyde reagent to form isonicotinyl-3-methoxybenzaldhydrazbne, N-acetyl-3-methoxybenzaldhydrazone and 3-methoxybenzaldazine, respectively. INH, acetylhydrazine and hydrazine were analysed using an HPLC column (4.6x125 mm) packed with Spherisorb ODS-2 (Phase-Separations Ltd.) and a Spectra-Physics SP8000B liquid chromatograph (Santa Clara, CA, USA) fitted with a Spectra-Physics SP8400 uv/vis detector set at 300 nm. The column temperature was 48°C and the flow rate was 1 mL/minute. All samples were determined in triplicate. Of the 30 samples 44 (fifteen of the 32 hour plasma samples and fifteen of the 48 hour plasma samples) analyzed in triplicate for each of INH, acetylhydrazine and hydrazine, the coefficients of determination (mean ± SD) are 7.4 ± 5.9 %, 9.0 ± 6.9 % and 6.4 ± 4.2 %, respectively. Mobile phase A consisted of 20% MeCN/ 10% MeOH and 70 % 5 mM sodium acetate in HPLC water adjusted to pH 5.0 with acetic acid. Mobile phase B consisted of 40% MeCN/ 40% MeOH and 20% 5 mM sodium acetate in HPLC water adjusted to pH 5.0 with acetic acid. All eluents were filtered before use and degassed in the solvent reservoir. A third mobile phase consisting of 20% MeCN/ 10% MeOH and 70% HPLC water (unbuffered) was used for cleaning the column. The gradient elution program (linear) for this separation was as follows: Time (minutes) %A %B 0 100 0 5 60 40 10 0 100 16.7 0 100 17.5 100 0 Retention times for the derivatized hydrazones of acetylhydrazine and INH and the azine of hydrazine were 6.9, 7.8 and 40.4 minutes with the gradient elution program. The calibration curve of INH concentration (uM) versus the ratio of INH to 9-fluorenone had a correlation coefficient (r) of 0.999, a slope of 0.050 (95% confidence intervals 0.049 to 0.052) INH:0-fluorenone/uM INH and a y-intercept of 0.08 (95% confidence intervals -0.27 to 0.44) INH:°-fluorenone (Figure 1.5; page 79). Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.999, a slope of 0.051 (95% confidence intervals 0.050 to 0.051) INH:P-fluorenone. The calibration curve of acetylhydrazine concentration (uM) versus the ratio of acetylhydrazine to P-fluorenone had a correlation coefficient (r) of 0.999, a slope of 0.018 (95% confidence intervals 0.018 to 0.019) acetylhydrazine: P-fluorenone/uM acetylhydrazine and a y-intercept of 0.003 (95% confidence intervals -0.04 to 0.05) acetylhydrazine:P-fluorenone (Figure 1.6; page 80). Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.999, a slope of 0.019 (95% confidence intervals 0.018 to 0.019) acetylhydrazine:P-fluorenone. TABLE 1.8 VALIDATION SUMMARY - INH, ACETYLHYDRAZINE AND HYDRAZINE ASSAY INH Acetylhydrazine Hydrazine Limit of Detection* 3 uM 3 uM 2 uM Limit of Quantitation 20 LIM 7.5 LIM 7.5 LIM Standards Range 20 - 1200 uM 7.5-210 uM 7.5 - 200 jiM Sample Range 20 - 426 LIM 7.5 -170 uM 14 - 104 uM Accuracy See TABLE 1.9 See TABLE 1.9 See TABLE 1.9 Specificity No interfering peaks were observed at the positions of INH, acetylhydrazine, hydrazine or P-fluorenone. Linearity: INH (plasma): Acetylhydrazine (plasma): Hydrazine (plasma): 1) . r = 0.999; slope = 0.050 (0.049 to 0.052); y-int = 0.08 (-0.27 to 0.44) 2) . r = 0.999; slope = 0.018 (0.018 to 0.019); y-int = 0.003 (-0.04 to 0.05) 3) . r = 0.999; slope = 0.074 (0.072 to 0.077); y-int = 0.03 (-0.16 to 0.22) Calculation Slope (b = 0) 1 1) . r = 0.999; slope = 0.051 (0.050 to 0.051) 2) . r = 0.999; slope = 0.019 (0.018 to 0.019) 3) . r = 0.999; slope = 0.075 (0.073 to 0.076) Volumetric Variance of Analytical Procedure Controlled through the use of the internal standard 9-fluorenone and the use of a compound to P-fluorenone ratio to analyze data. * Limit of Detection: Defined as three times the background noise. * Calculation slope refers to the slope from the standard curve which was used to calculate the concentration of INH, acetylhydrazine and hydrazine in the samples. The calibration curve of hydrazine concentration (uM) versus the ratio of hydrazine to P-fluorenone had a correlation coefficient (r) of 0.999, a slope of 0.074 (95% confidence intervals 0.072 to 0.077) hydrazine:P-fluorenone/uM hydrazine and a y-intercept of 0.03 (95% confidence intervals -0.16 to 0.22) hydrazine:P-fluorenone (Figure 46 1.7; page 81). Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.999, a slope of 0.075 (95% confidence intervals 0.073 to 0.076) hydrazine:9-fluorenone. TABLE 1.9 INTRA- AND INTER-DAY VARIABILITY (ACCURACY AND PRECISION) Compound Concentration Added (uM) Concentration Found (uM) Intra-Day (n = 5) (mean ± SD) C V % Inter-Day (n = 10) (mean ± SD) C V % INH 120 123 ± 8 6.9 117 ± 1 3 11 INH 600 606 ± 1 1 1.8 625 ± 26 4.0 INH 1200 1208 ± 6 1 5.0 1237 ± 125 10 Acetylhydrazine 42 45 ± 7 15 44 ± 7 16 Acetylhydrazine 105 106 ± 5 4.5 104 ± 5 4.5 Acetylhydrazine 210 213 ± 7 3.4 214 ± 6 2.9 Hydrazine 20 21 ± 1 6.6 22 ± 3 12 Hydrazine 60 63 ± 4 5.6 62 ± 7 11 Hydrazine 200 204 ± 15 7.4 200 ± 20 10 Sample chromatograms of blank plasma (a) and a 48 hour plasma sample from an animal injected with INH (b) are shown in Figure 1.8 (page 82). A validation summary (Table 1.8) and an estimation of accuracy and precision (Table 1.9) are shown. Determination of Hydrazine in 12 Hour Plasma Samples Rationale In Study 4, plasma hydrazine levels were determined in order to compare the plasma hydrazine levels in animals treated with isoniazid (INH) in the presence and absence of the amidase inhibitor bis-/?-nitrophenyl phosphate (BNPP). This method was used only for the determination of hydrazine and not for determination of INH and acetylhydrazine as has been previously described (see above). The following procedure describes the quantitation of plasma hydrazine levels using an HPLC apparatus equipped 47 with a single pump and no requirements for multiple mobile phases or an oven. Therefore, this new procedure was simpler than the previously described procedure. Since hydrazine readily reacts with aldehyde groups, the majority of hydrazine in plasma is present in the form of pyruvate azine (Ellard and Gammon, 1976) and a-ketoacid azines (Timbrell and Wright, 1984). The basis of the assay was to hydrolyse the azines in a low pH environment and derivatise them in the presence of an excess of the derivatising reagent 3-methoxybenzaldehyde. Detection of conjugated hydrazine (3-methoxybenzaldazine) is then possible at a wavelength of 300 nm. Analytical Procedure The procedure for determination of 12 hour plasma hydrazine levels is described as follows. One hundred uL of plasma and 100 uL of propanol (100%) were combined, vortexed and allowed to stand for 6 minutes. Two hundred uL of HPLC grade water were then added. This mixture was vortexed and centrifuged at 12,700xg for 6 minutes. The supernatant was filtered with a Millipore syringe filter (Millex L C R 4 , 0.5 um, 4 mm, Millipore Corporation, Bedford MA, 01730). Three hundred microlitres of the filtered supernatant was removed and combined with 75 pX of derivatising reagent. The derivatising reagent was made up by adding 125 uL 3-methoxybenzaldehyde and 3.75 mL formic acid (minimum assay purity 88%) to a 50 mL volumetric flask and making up to 50 mL with propanol. Five mL of 10 mM 9-fluorenone solution (internal volumetric standard) was also added to this solution. The sample containing solution was placed in the dark for 2 hours, centrifuged for 6 minutes and filtered again with a Millipore syringe filter. The resulting solution was 48 placed on an HPLC column (4.6x100mm) packed with Spherisorb ODS-2 (Phase-Separations Ltd.). An ISCO (model 2350) liquid chromatograph and an ISCO V 4 UV/vis detector set at 300 nm were used for the analysis. The mobile phase consisted of 35% 5 mM sodium acetate adjusted to pH 5.0 with glacial acetic acid and 65% HPLC grade acetonitrile in HPLC grade water. The flow rate was 1 mL/min. The mobile phase was filtered prior to use (0.45p.m pore, 45 mm Nylon filters, Micron Separations Inc.) and degassed in the solvent reservoir. Room temperature ranged from 23-25°C. Validation Analyses using pure isonicotinyl-3-methoxybenzaldhydrazone, N-acetyl-3-methoxybenzaldhydrazone and 3-methoxybenzaldazine (derivatised INH, acetylhydrazine and hydrazine; synthesized by Dr. R.A. Wall) dissolved in propanol and water were run individually, and together, to determine their retention times. In the isocratic system, the i-methoxybenzaldehyde derivatives of hydrazine, INH and acetylhydrazine had retention times of 6, 2 and 2 minutes, respectively. In plasma spiked with INH, acetylhydrazine and hydrazine individually, and together, it was confirmed that the INH and acetylhydrazine derivatives eluted much earlier than the hydrazine derivative and did not interfere with the azine peak at 6.0 minutes. There were no interfering peaks observed in baseline (zero hour) plasma samples. It was also confirmed that BNPP did not interfere with the hydrazine peak, as there were no peaks present in 12 hour plasma samples from rabbits which received only BNPP (no INH). 49 A standard curve (in triplicate) was prepared using hydrazine dissolved in helium degassed - acidified water at concentrations of 5, 10, 25, 50, 100 and 200 uM hydrazine. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.573 (95% confidence intervals 0.558 to 0.589) mVxseconds/uM hydrazine and a y-intercept of -0.17 (95% confidence intervals -1.6 to 1.3) mVxseconds. Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.999 and a slope of 0.572 (95% confidence intervals 0.563 to 582) mVxseconds/uM hydrazine. A standard curve for hydrazine was prepared using blank rabbit plasma spiked with hydrazine (in triplicate) at concentrations of 10, 25, 50, 100 and 200 pM hydrazine. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.564 (95% confidence intervals 0.553 to 0.575) mVxseconds/pM hydrazine and a y-intercept of 1.5 (95% confidence intervals 0.37 to 2.66) mVxseconds (Figure 1.9; page 83). The reason that the 95% confidence intervals for the y-intercept do not include zero is not known but could be related to an unexplained greater % recovery at the lower concentrations of the standard curve (see percent recovery calculations below). Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.999 and a slope of 0.575 (95% confidence intervals 0.560 to 0.590) mVxseconds/uM hydrazine. The peak area units (mVxseconds) were converted into hydrazine concentrations (pM) using the following equation: mVxseconds / (0.575 mVxseconds/uM hydrazine) = X uM hydrazine Hydrazine concentration is reported as umoles/L (uM). Sample chromatograms of blank plasma (a) and a 12 hour plasma sample from an animal injected with INH (b) are shown in Figure 1.10 (page 84). so TABLE 1.10 VALIDATION SUMMARY - DETERMINATION OF HYDRAZINE IN 12 HOUR PLASMA SAMPLES Limit of Detection* 0.6 uM Limit of Quantitation 2 LLM (14% CV - acceptance criteria 15%) Standards Range 10 - 200 LLM Sample Range 10 - 40 uM % Recovery 62.9 ± 8.9% (mean ± SD) (calculation in Percent Recovery section) Specificity No interference by INH, acetylhydrazine or any other peak at the position of hydrazine. Linearity: In water: In plasma: r = 0.999; slope = 0.573 (0.558 to 0.589); y-int = -0.17 (-1.64 to 1.31) r = 0.999; slope = 0.564 (0.553 to 0.575); y-int = 1.51 (0.37 to 2.66) Calculation Slope (b = 0 ) 1 r = 0.999; slope = 0.575 (0.560 to 0.590) Volumetric Variance of Analytical Procedure 6.4% * Limit of Detection: Defined as three times the background noise. f Calculation slope refers to the slope from the standard curve which was used to calculate the concentration of hydrazine in the samples. Each 12 hour plasma hydrazine sample was run once and concentration was calculated as pmoles/L (uM). In total, 47 samples, standards and blanks were run over two days including two 25 uM and two 50 uM standards, one of which was included every 11 samples. These controls of known concentrations were found to be 24 and 20 uM for the 25 pM standard and 47 and 49 pM for the 50 pM standard. The sample concentrations quantitated ranged from 10 to 40 pM. The 9-fluorenone internal volumetric standard included in all of the 47 runs had a mean peak area of 48.8 mVxseconds and a standard deviation of 3.1 mVxseconds (coefficient of determination 6.4%). Therefore, the internal volumetric standard confirmed the volumetric variance of the analytical procedure to be within 6.4 % (volumetric variance includes potential variance introduced by volumetric, photometric, autoinjector injection volume (35 pL), time and temperature variables potentially introduced at or after 51 the derivatisation step). A validation summary (Table 1.10) and an estimation of precision (Table 1.11) are shown. TABLE 1.11 PRECISION OF THE PROCEDURE FOR DETERMINATION OF HYDRAZINE IN 12 HOUR PLASMA SAMPLES In Water In Plasma uM (n = 3) CV%(SD/mean) MM (n = 3) CV%(SD/mean) 5 3.8 5 N/A 10 1.5 10 10.4 25 2.5 25 5.0 50 1.5 50 4.8 100 2 100 0.4 200 1.9 200 0.9 (mean ± SD) 2.2 ± 0.9% (mean ± SD) 4.3 ± 4.0 % N/A - not available Determination of Hydrazine in 24 and 32 Hour Plasma Samples Analytical Procedure. The procedure for determination of plasma hydrazine levels at 24 and 32 hours was altered slightly from the procedure for the determination of plasma hydrazine levels at 12 hours. Instead of using 100 uL propanol for the denaturation step, 100 |iL acetonitrile and 100 pL perchloric acid (0.6N) were substituted for 100 uL propanol and 100 pL of HPLC-grade water in an attempt to improve the plasma denaturation procedure and to extend the lifetime of the HPLC columns. Formic acid was removed from the derivatising reagent because of the presence of perchloric acid in the denaturation procedure. A different column (Select ODS-2 5 urn; Chromatographic Specialties) was also used. The revised procedure for the determination of plasma hydrazine concentrations (uM) is described as follows. One hundred uL of plasma and 100 uL of acetonitrile (100%) were combined, vortexed and allowed to stand for 6 minutes. One hundred uL of 52 0.6N perchloric acid and 100 uL of HPLC grade water were then added. This mixture was vortexed and centrifuged at 12,700xg for 6 minutes. The supernatant was filtered with a Millipore syringe filter (Millex LCR4 , 0.5 um, 4 mm, Millipore Corporation, Bedford MA, 01730). Three hundred microlitres of the filtered supernatant was removed and combined with 75 uL of derivatising reagent. The derivatising reagent was made up by adding 125 uL 3-methoxybenzaldehyde to a 50 mL volumetric flask and making up to 50 mL with propanol. The sample containing solution was placed for 2 hours in the dark followed by centrifugation for 6 minutes and a second filtration with a Millipore syringe filter. The resulting solution was placed onto an HPLC column (4.6x 100mm) (Select ODS-2 5 um; Chromatographic Specialties). An ISCO (model 2350) liquid chromatograph and an ISCO V 4 UV/vis detector set at 300 nm were used for the analysis. The solvent used consisted of 35% 5 mM sodium acetate adjusted to pH 5.0 with glacial acetic acid and 65% HPLC grade acetonitrile in HPLC grade water. The flow rate was 1 mL/min. All solvents were filtered prior to use (0.45um pore, 45 mm Nylon filters, Micron Separations Inc.) and degassed in the solvent reservoir. Room temperature ranged from 23-25°C. A standard curve was prepared (in triplicate) using hydrazine dissolved in helium degassed - acidified water at concentrations of 5, 10, 25, 50, 100 and 200 uM hydrazine. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.221 (95% confidence intervals 0.207 to 0.236) mVxseconds/uM hydrazine and a y-intercept of 0.12 (95% confidence intervals -1.23 to 1.46) mVxseconds. Using the equation y = mx, the 53 standard curve had a correlation coefficient (r) of 0.999 and a slope of 0.222 (95% confidence intervals 0.213 to 0.231) mVxseconds/uM hydrazine. A standard curve was prepared (in duplicate) using blank rabbit plasma spiked with hydrazine at concentrations of 4, 8, 16, 32, 64 and 128 uM hydrazine. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.235 (95% confidence intervals 0.222 to 0.248) mVxseconds/uM hydrazine and a y-intercept of-0.64 (95% confidence intervals -1.41 to 0.13) mVxseconds (Figure 1.11; page 85). Using the equation y = mx, the i standard curve had a correlation coefficient (r) of 0.998 and a slope of 0.228 (95% confidence intervals 0.216 to 0.239) mVxseconds/uM hydrazine. Sample chromatograms of blank plasma (a) and a 32 hour plasma sample from an animal injected with BNPP and INH (b) are shown in Figure 1.12 (page 86). TABLE 1.12 VALIDATION SUMMARY - DETERMINATION OF HYDRAZINE IN 24 AND 32 HOUR PLASMA SAMPLES Limit of Detection * 0.8 uM Limit of Quantitation 2 uM (14% CV - acceptance criteria 15%) Standards Range 4 - 128 uM Sample Range 4 -74uM % Recovery 19.4 ± 3.4% (mean ± SD) (calculation in Percent Recovery section) Specificity No interference by INH, acetylhydrazine or any other peak at the position of hydrazine. Linearity In water: In plasma: r = 0.999; slope = 0.221 (0.207 to 0.235); y-int = 0.115 (-1.23 to 1.46) r = 0.999; slope = 0.235 (0.222 to 0.248); y-int = -0.64 (-1.41 to 0.128) Calculation Slope (b = 0) 1 r = 0.998; slope = 0.228 (0.216 to 0.239) * Limit of Detection: Defined as three times the background noise. t Calculation slope refers to the slope from the standard curve which was used to calculate the concentration of hydrazine in the samples. 54 All of the 24 and 32 hour plasma samples were run in duplicate. The peak areas of hydrazine (mVxseconds) were converted into hydrazine concentrations using the following equation: mVxseconds / (0.228 mVxseconds/uM hydrazine) = X p M hydrazine Hydrazine concentration is reported as pmoles/L (pM). A validation summary (Table 1.12) and an estimation of precision (Table 1.13) are shown. T A B L E 1.13 PRECISION OF THE A S S A Y FOR DETERMINATION OF H Y D R A Z I N E IN 24 A N D 32 HOUR P L A S M A SAMPLES In Water uM(n = 3) CV%(SD/mean) 5 4.5 10 5.3 25 1.8 50 6.3 100 6.8 200 3.1 (mean ± SD) 4.6 ± 1.9% Hepatic INH-Amidase Activity Determination. Rationale Hepatic INH-amidase is an important enzyme in the metabolism of INH and its metabolites and its activity has previously been identified in hepatic microsomes (Whitehouse etal, 1983; Sendo etal, 1984). A relevant aspect of its activity is its ability to produce hydrazine either directly from INH, or indirectly by conversion of acetyl-INH to acetylhydrazine followed by conversion of acetylhydrazine to hydrazine. The following assay was developed in order to measure the rate of direct conversion of INH to hydrazine 55 by incubation of hepatic microsomes with INH and quantitating the amount of hydrazine produced per mg protein per hour. Analytical Procedure Hepatic amidase activity was determined by incubation of INH with microsomes as previously described (Whitehouse et al., 1983; Sendo et al., 1984) followed by measurement of the production of hydrazine using HPLC. The procedure for the detection of hydrazine produced by incubating microsomes with INH was as follows: One hundred pL of INH in 67 mM KH2P04 buffer (pH 7.0) (3 mM initial concentration) was incubated with 150 uL hepatic microsomes (3.-11 mg protein/mL) and 50 uL 67 mM KH2PO4 buffer (pH 7.0) at 37°C for 30 minutes (300 pL total volume). Hepatic amidase activity has been determined using incubations of INH with microsomes for up to 2 hours (Whitehouse et al., 1983). An incubation time of 30 minutes in the present study was selected following preliminary analyses involving incubations of microsomes with and without INH, for 30, 60 and 120 minutes. Control incubations, consisting of heat-inactivated microsomes, were also included. Incubation of active microsomes with INH showed hydrazine after incubation at 30, 60 and 120 minutes. The heat-inactivated microsomes with INH and the active microsomes without INH did not show any hydrazine. Experiments were carried out in order to determine the time period of a linear reaction rate of INH-amidase. Incubation of INH with active microsomes for 60, 90, 120 and 150 minutes revealed that the reaction rate reached a plateau at around 90 minutes. Incubation of TNH in active microsomes for 15, 30, 60 and 90 minutes showed that 56 hydrazine production was linear up to 30 minutes. Therefore, 30 minutes was chosen as the incubation period of microsomes with INH. Following the 30 minute incubation, the reaction was stopped by the addition of 0.3 mL acetonitrile, followed by mixing, standing for 6 minutes, addition of 0.3 mL 0.6N HCIO4 and centrifugation at 12,700xg for six minutes. The supernatant was filtered through a 0.5 um Millipore syringe filter (Millex LCR4) and 150 uL of the filtrate was combined with 150 uL H P L C grade water and 75 uL derivatising reagent (125 uL 3-methoxybenzaldehyde made up to 50 mL with propanol). This solution was incubated in the dark at room temperature for two hours. The solution was centrifuged 12,700xg for 6 minutes, filtered again with a 0.5 urn syringe filter (Millipore), and placed onto an H P L C column under the same conditions as those described for determination of plasma hydrazine levels in 24 and 32 hour plasma samples. A standard curve using the above procedure was prepared (in triplicate) using active microsomes spiked with hydrazine at concentrations of 5, 12.5, 25, 50, 100 and 250 u M hydrazine in the incubation mixture. The standard curve had a correlation coefficient (r) of 0.998, a slope of 0.152 (95% confidence intervals 0.140 to 0.165) mVxseconds/uM hydrazine and a y-intercept of-1.19 (95% confidence intervals -2.56 to 1.86) mVxseconds (Figure 1.13; page 87). Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.996, a slope of 0.146 (95% confidence intervals 0.134 to 0.157) mVxseconds/uM hydrazine. Sample chromatograms of microsomes with no INH added (a) and microsomes incubated with INH for 30 minutes (b) are shown in Figure 1.14 (page 57 88). A validation summary (Table 1.14) and an estimation of precision (Table 1.15) are shown. Each sample was run in duplicate and the average of these peak areas taken. The sample duplicates of peak area had a variance of (mean ± standard deviation) 3.9 ± 3.7% (n = 58). The peak areas of hydrazine (mVxseconds) were converted into hydrazine concentrations (pM) using the standard curve. Hydrazine production was expressed as nmoles/mg protein/hour using the following formula: mVxseconds / (0.146 mVxseconds/pM hydrazine) = X p M hydrazine X pmoles hydrazine/L x Y mg protein/mL"1 x 0.5 hour'1 x 1000 mL/L' 1 x 1000 nmoles/pmole = nmoles hydrazine produced/mg protein/hour T A B L E 1.14 VALIDATION SUMMARY - HEPATIC AMIDASE ASSAY Limit of Detection* 1.1 uM Limit of Quantitation 2 uM (14% CV - acceptance criteria 15%) Standards Range 5 - 250 uM Sample Range 5-157 LIM % Recovery 5.8 ± 1.0% (mean ± SD) (calculation in Percent Recovery section) Specificity No interference by INH, acetylhydrazine or any other peak at the position of hydrazine. Linearity: In microsomes: r = 0.998; slope = 0.152 (0.140 to 0.165); y-int = -1.19 (-2.26 to 0.19) Calculation Slope (b = 0) f r = 0.996; slope = 0.146 (0.134 to 0.157) * Limit of Detection: Defined as three times the background noise. * Calculation slope refers to the slope from the standard curve which was used to calculate the concentration of hydrazine in the samples. Active control microsomes spiked with hydrazine were used as standards for the standard curve. The standards were not incubated for 30 minutes since the goal was to quantitate the exact amount of hydrazine present at the 30 minute time point in the 58 incubation samples. Incubation of the standards for 30 minutes would have likely altered the known concentration of hydrazine. Therefore, hydrazine concentrations of standards and samples (uM) are related to the concentration of hydrazine in the 300 pL incubation mixture. T A B L E 1.15 PRECISION OF THE PROCEDURE FOR DETERMINATION OF H Y D R A Z I N E IN MICROSOMAL INCUBATIONS In Microsomes* uM (n = 3) CV%(SD/mean) 5 0.8 12.5 1.8 25 8.2 50 1.9 100 0.9 250 4.3 (mean ± SD) 2.9 ± 2.9% * Precision in water was not calculated. BNPP IC50 Determination of Hepatic INH-Amidase Activity. Rationale The potency of the inhibition of INH-amidase by BNPP was investigated. Analytical Procedure Microsomes were incubated with INH (3 mM) for 30 minutes (as described above for the hepatic amidase activity) in the presence of BNPP at concentrations of 0, 31 nM, 63 nM, 125 nM, 250 nM, 500 nM, 1 uM, 2 uM, 4 uM, 8 uM, 16 u M and 32 u M (in duplicate). In the first experiment, no amidase inhibition was observed at BNPP concentrations of 0, 31 nM, 63 nM, 125 n M and 250 nM; therefore, the 31 nM, 63 nM and 125 nM concentrations were not included in the next two experiments. Two further incubations were performed with different microsomal preparations each time. The 59 hydrazine producing activity of the microsomes was calculated as described above for hepatic amidase activity. Results from all three experiments are reported. Determination of the Mechanism of INH-Amidase Inhibition - Lineweaver-Burk and Eadie-Hofstee Plots. Rationale The mechanism of inhibition of INH-amidase by BNPP was investigated using incubations of increasing concentrations of INH in the presence of 0, 1.5 and 3 pM BNPP. Analytical Procedure Microsomes were incubated with INH (at concentrations ranging from 3 pM to 10 mM) for 30 minutes (as described above for the hepatic amidase activity) in the presence of BNPP at concentrations of 0, 1.5 and 3 pM. Incubations were in duplicate. The data are presented as Lineweaver-Burk and Eadie-Hofstee plots. Plasma INH-Amidase Activity Determination Rationale In addition to jTNH-amidase activity in the liver, it has been suggested that INH-amidase activity occurs in plasma (Mitchell et al, 1976). The following assay was developed in order to measure the rate of conversion of INH to hydrazine produced by incubation of plasma with INH. Analytical Procedure Plasma amidase activity was determined by incubating INH with plasma followed by measurement of the hydrazine produced using HPLC. The procedure was as follows: one hundred microlitres of INH in 67 mM KH2P04 buffer (pH 7.0) (3 mM initial 60 concentration) was incubated with 150 uL plasma and 50 uL 67 mM KH2PO4 buffer (pH 7.0) at 37°C for 30 minutes (300 pX total volume). The rest of the procedure was identical to that described above for microsomal hydrazine production from INH using microsomes. A standard curve was prepared (in triplicate) using plasma spiked with hydrazine to concentrations of 2, 4, 8, 16, 32 and 64 uM. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.166 (95% confidence intervals 0.163 to 0.170) mVxseconds/uM hydrazine and a y-intercept of 0.040 (95% confidence intervals -0.062 to 0.142) mVxseconds (Figure 1.15; page 89). Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.999, a slope of 0.167 (95% confidence intervals 0.165 to 0.170) mVxseconds/uM hydrazine. Sample chromatograms of plasma with no added INH (a) and a spiked plasma sample from a rabbit in Study 4 (b) are shown in Figure 1.16 (page 90). A validation summary (Table 1.16) and an estimation of precision (Table 1.17) are shown. TABLE 1.16 VALIDATION SUMMARY - PLASMA AMIDASE ASSAY Limit of Detection* 0.8 nM Limit of Quantitation 2 uM (14% CV - acceptance criteria 15%) Standards Range 2-64 uM Sample Range 2-5 nM % Recovery 8.8 ± 0.6% (mean ± SD) (calculation in Percent Recovery section) Specificity No interference by INH, acetylhydrazine or any other peak at the position of hydrazine. Linearity: In plasma: r = 0.999; slope = 0.166 (0.163 to 0.170); y-int = 0.040 (-0.06 to 0.14) Calculation Slope (b = 0) f r = 0.999; slope = 0.167 (0.165 to 0.170) * Limit of Detection: Defined as three times the background noise. * Calculation slope refers to the slope from the standard curve which was used to calculate the concentration of hydrazine in the samples. 61 All samples were run in duplicate. The peak areas of hydrazine (mVxseconds) were converted into hydrazine concentrations using the standard curve. Hydrazine concentration (pM) was converted into nmoles/mg protein/hour using the following formula: mVxseconds / (0.167 mVxseconds/pM hydrazine) = X pM hydrazine X pmoles hydrazine/L x Y mg protein/mL"1 x 0.5 hour"1 x 1000 mL/L'1 x 1000 nmoles/pmole = nmoles hydrazine produced/mg protein/hour Blank plasma samples spiked with hydrazine were used for the standard curve. The standards were not incubated for 30 minutes. This is because the goal was to quantitate the exact amount of hydrazine present at the 30 minute time point in the samples. Incubation of the standards may have affected the known concentration of hydrazine in them. Hydrazine concentrations of standards and samples (pM) are related to the concentration in the 300 pL incubation mixture. TABLE 1.17 PRECISION OF THE PROCEDURE FOR DETERMINATION OF HYDRAZINE IN PLASMA INCUBATIONS In Plasma* uM (n = 3) CV%(SD/mean) 2 14.2 4 3.6 8 2.2 16 1.3 32 7.5 64 1.8 (mean ± SD) 5.1 ±5.0% * Precision in water was not calculated. 62 Percent Recovery Calculations In order to calculate the percent recovery of hydrazine using the described procedures, known concentrations of pure 3-methoxybenzaldazine (azine) dissolved in propanol/water were injected directly onto the column. A standard curve was prepared (in triplicate) at concentrations of 2.5, 5, 10, 25, 50 and 100 uM 3-methoxybenzaldazine. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.049 (95% confidence intervals 0.048 to 0.051) mVxseconds/pmoles azine and a y-intercept of 1.42 (95% confidence intervals -1.24 to 4.08) mVxseconds. Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.999, a slope of 0.050 (95% confidence intervals 0.049 to 0.051) mVxseconds /pmoles azine. In order to compare the pure 3-methoxybenzaldazine standard curve to derivatised hydrazine, known concentrations of hydrazine dissolved in acidified and helium degassed water were combined with the derivatising reagent, allowed to stand for two hours, and then injected directly onto the column. Using this technique, the standard curve had a correlation coefficient (r) of 0.999, a slope of 0.048 (95% confidence intervals 0.045 to 0.051) mVxseconds /pmoles azine and a y-intercept of 2.70 (95% confidence intervals -1.47 to 6.87) mVxseconds. Using the equation y = mx, the standard curve had a correlation coefficient (r) of 0.998, a slope of 0.049 (95% confidence intervals 0.47 to 0.052) mVxseconds /pmoles azine. These results demonstrate no difference between the slope for pure 3-methoxybenzaldazine and hydrazine put through the derivatising procedure. This supports 6 3 previous data suggesting that derivatisation of hydrazine by 3-methoxybenzaldehyde is complete after 2 hours. The pure 3-methoxybenzaldazine standard curve was used to calculate the percent recovery of hydrazine for the following assays: 12 hour plasma hydrazine, 24/32 hour plasma hydrazine, hepatic amidase activity and plasma amidase activity. The calculation was made by converting peak area (mVxseconds) of each standard curve concentration into pmoles azine injected onto the HPLC column. The amount of azine (pmoles) detected at the column was divided by the theoretical amount of azine put onto the column to calculate percent recovery. The amount of azine put onto the column was calculated by converting the original standard concentration of hydrazine (pM) into pmoles and dividing by the dilution factor before reaching the column. The percent recovery for the 12 hour plasma hydrazine assay was calculated to be 62.9 ± 8.9% (mean ± standard deviation) (Table 1.18). The following equation was used: pM Hydrazine standard x 0.1 mL x 1000/ 5 (dilution factor) = pmoles hydrazine added to column mVxseconds / (0.050 mVxseconds /pmole azine) = pmoles hydrazine found at column TABLE 1.18 PERCENT RECOVERY FOR THE 12 HOUR PLASMA HYDRAZINE ASSAY LIM Hydrazine in Standard (n = 3) pmoles Hydrazine added to column mVxseconds (peak area) pmoles Azine at column Percent Recovery % 10 200 7.81 157.1 78.6 25 500 15.0 301.3 60.3 50 1000 29.7 597.4 59.7 100 2000 57.8 1163.1 58.2 . 200 4000 114.4 2301.8 57.5 62.9 ± 8.9 (SD) 64 The percent recovery for the 24/32 hour plasma hydrazine assay was calculated to be 19.4 ± 3.4% (Table 1.19). The following equation was used: pM Hydrazine standard x 0.1 mL x 1000/ 5 (dilution factor) = pmoles hydrazine added to column mVxseconds / (0.05 mVxseconds/pmole azine) = pmoles hydrazine found at column TABLE 1.19 PERCENT RECOVERY FOR THE 24 & 32 HOUR PLASMA HYDRAZINE ASSAY uM Hydrazine in Standard (n = 2) pmoles Hydrazine added to column mVxseconds (peak area) pmoles Azine at column Percent Recovery % 4 80 0.62 12.4 15.6 8 160 1.38 27.8 17.4 16 320 2.68 53.9 16.8 32 640 6.39 128.7 20.1 64 1280 15.0 301.9 23.6 128 2560 29.3 589.1 23.0 19.4 ± 3 . 4 (SD) The percent recovery of hydrazine in the hepatic amidase activity determinations was calculated to be 5.8 ± 1.0% (Table 1.20). The following equation was used: pM Hydrazine standard x 0.3 mL x 1000/ 7.5 (dilution factor) = pmoles hydrazine added to column mVxseconds / (0.05 mVxseconds/pmole azine) = pmoles hydrazine found at column TABLE 1.20 PERCENT RECOVERY FOR THE HEPATIC AMIDASE ASSAY uM Hydrazine in Standard (n = 3) pmoles Hydrazine added to column mVxseconds (peak area) pmoles Azine at column Percent Recovery % 5 200 0.53 10.6 5.3 10 400 1.23 24.8 5.0 25 1000 2.50 50.2 5.0 50 2000 5.64 113.4 5.7 100 5000 13.0 260.8 6.5 250 10000 37.5 754.3 7.5 5.8 ± 1.0 (SD) 65 The percent recovery of hydrazine in the plasma amidase activity determination was calculated to be 8.8 ± 0.6% (Table 1.21). The following equation was used: uM Hydrazine standard x 0.3 mL x 1000/ 7.5 (dilution factor) = pmoles hydrazine added to column mVxseconds / (0.05 mVxseconds /pmole azine) = pmoles hydrazine found at column TABLE 1.21 PERCENT RECOVERY FOR THE PLASMA AMIDASE ASSAY uM Hydrazine in Standard (n = 3) pmoles Hydrazine added to column mVxseconds (peak area) pmoles Azine at column Percent Recovery % 2 80 0.39 7.9 9.9 4 160 0.74 14.8 9.2 8 320 1.39 28.0 8.7 16 640 2.68 54.0 8.4 32 1280 5.25 105.7 8.3 64 2560 10.74 216.0 8.4 8.8 ± 0.6 (SD) These data show a three-fold difference in the percent recovery between the 12 hour plasma hydrazine assay and the 24/32 hour plasma hydrazine assay. In addition, using the same denaturants as the 24/32 hour plasma hydrazine assay, the percent recovery in the hepatic and plasma amidase assays is decreased even greater than the percent recovery in the 24/32 hour plasma hydrazine assay. The only significant difference between the assay used to determine hydrazine concentration in 12 hour plasma samples versus 24 and 32 hour plasma samples was the substitution of propanol with acetonitrile and perchloric acid in an attempt to improve the denaturation of the samples. Therefore, the difference in percent recovery observed between the two assays is most likely due to the presence of acetonitrile and perchloric 66 acid. A possible reason for the decreased recovery in the presence of perchloric acid is that a step involving the addition of potassium carbonate (K2CO3) to inactivate the perchloric acid after denaturation of proteins was inadvertently omitted from this procedure. Therefore, the difference in recovery between the 12 hour plasma hydrazine assay and the 24/32 hour plasma hydrazine assay is likely that hydrazine is oxidized to an unrecoverable form by perchloric acid. Although perchloric acid is a strong acid and oxidizing agent, moderate concentrations are not known to oxidise hydrazine, even at high temperatures (Koltunov et al., 1976). In the presence of molybdenum, the reaction proceeds at an appreciable rate, which is independent of the concentration of hydrazine and proportional to the concentration of perchloric acid (Koltunov etal., 1976). Since the decrease in recovery is present in the 24/32 hour standard curves in both plasma [19.4 ± 3.4 (SD)] and water [22.6 ± 1.5 (SD)], the decrease is not dependent on an interaction of hydrazine or perchloric acid with plasma. The decreased recovery in the amidase assays as compared to the 24/32 hour plasma hydrazine assay may be explained by the higher original concentration of perchloric acid in the denaturation step (33%) versus the 24/32 hour plasma hydrazine assay (25%). Another possible explanation for difference in recovery of hydrazine is that the perchloric acid interferes with derivatisation of hydrazine to the fully conjugated azine. Perchloric acid (50 mM) in the presence of hydrazine (6 mM) hydrolyses the salicylaldehyde hydrazones back to salicylaldehyde and hydrazine but does not affect the formation or stability of the salicylaldazine (Malone, 1970). It is possible that at the 67 concentrations of perchloric acid used in the present study, perchloric acid inhibited or reversed derivatisation with 3-methoxybenzaldazine. Slope Differences Between 24 & 32 Hour Plasma Hydrazine Procedure and the Amidase Procedures The difference in the slope of the standard curve for the 24 and 32 hour plasma samples versus the hepatic and plasma amidase samples can be attributed to dilution differences between the assays. The actual difference in the slopes is shown in the following calculation of slope ratio: (0.146 + 0.167)/2 = 0.157 (average amidase assay slope) 0.157/0.228 (plasma hydrazine assay slope) = 0.69 Therefore the slope ratio is 0.69 In the plasma hydrazine assay, before reaching the column, the sample is diluted to l/5th of the original concentration. In the amidase activity assays, before reaching the column, the sample is diluted to 1/7.5 of the original concentration. The following calculation demonstrates the dilution ratio: 1/7.5 / 1/5 or 0.133/0.20 = 0.67 Therefore the dilution ratio is 0.67 This calculation therefore accounts for the majority of the difference in slope between the two assays. Figure 1.17 (page 91) compares all three standard curves in one figure. 68 Acetylation Phenotype Assay. Rationale The use of rabbits as a model for genetically dependent acetylation in humans is valuable since it has been established that N-acetyltransferase-2 is the enzyme responsible for polymorphic acetylation of both sulfamethazine and isoniazid in rabbits and humans (Blum et al. 1989; Vatsis etal. 1995). The acetylation characteristics for sulfamethazine in rabbits have been shown to directly parallel the acetylation characteristics of INH in vivo and in vitro, and acetylation phenotyping using sulfamethazine is a reliable technique (Gordon et al 1973). The following procedure was used to phenotype for sulfamethazine acetylation rates. An intravenous injection of 10 mg/kg sulfamethazine was followed by a blood sample at 20 minutes. The ratio of sulfamethazine remaining to the acetylsulfamethazine formed was determined. This ratio gives a reliable estimate of the rate of acetylation of sulfamethazine. Analytical Procedure All rabbits were phenotyped for acetylator status five days prior to INH-injection using a method developed from Fischer and Klotz (1978). A 0.036 mmole/kg (10 mg/kg) dose of sulfamethazine was injected into a lateral ear vein. A blood sample (0.5 mL) was then collected from the opposite ear using a heparinised syringe, 20 minutes after injection of sulfamethazine. Plasma samples were assayed for sulfamethazine and its metabolite acetylsulfamethazine using HPLC following addition of acetonitrile (100 pL to an equal volume of plasma) and sulfamerazine (10 pL of either a 250 pM or a 1 mM solution; internal standard), thorough mixing, incubation for six minutes, addition of and mixing 69 with 200 L IL distilled water, and centrifugation at 12,700xg for six minutes. The supernatant was filtered with a Millipore syringe filter (Millex LCR4 , 0.5 um, 4 mm, Mllipore Corporation, Bedford MA, 01730). The final supernatant (35 uL) was injected onto an HPLC column (4.6x125 mm) packed with Polydex ODS-2. Two separate Polydex ODS-2 columns packed from the same batch of stationary phase by the same person were used throughout these analyses. An ISCO (model 2350) liquid chromatograph and an ISCO V 4 UV/vis detector set at 254 nm were used for the analysis. The solvent consisted of 77% 0.01M sodium acetate (pH 5.0), 15% MeOH and 8% MeCN mixture pumped at a flow rate of 1 ml/minute. All solvents were filtered prior to use (0.45pm pore, 45 mm Nylon filters, Micron Separations Inc.) and degassed in the solvent reservoir. Room temperature ranged from 23-25°C. Standard curves of sulfamethazine and acetylsulfamethazine in plasma were constructed. A standard curve of sulfamethazine was prepared using blank rabbit plasma spiked with sulfamethazine at concentrations of 5, 10, 25, 50, 100, 150 and 200 uM. The standard curve had a correlation coefficient (r) of 0.998, a slope of 0.368 (95% confidence intervals 0.340 to 0.396) mVxseconds/uM sulfamethazine and a y-intercept of-0.63 (95% confidence intervals -3.53 to 2.28) mVxseconds (Figure 1.18; page 92). Using the equation y - mx, the standard curve had a correlation coefficient (r) of 0.997 and a slope of 0.364 (95% confidence intervals 0.347 to 0.380) mVxseconds/uM sulfamethazine. A standard curve of acetylsulfamethazine was prepared using blank rabbit plasma spiked with acetylsulfamethazine at concentrations of 5, 10, 25, 50, 100, 150 and 200 uM. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.526 (95% 70 confidence intervals 0.494 to 0.558) mVxseconds /pM acetylsulfamethazine and a y-intercept of-0.20 (95% confidence intervals -3.55 to 3.15) mVxseconds (Figure 1.19; page 93). Using the equation y = mx, the standard curve has a correlation coefficient (r) of 0.999 and a slope of 0.524 (95% confidence intervals 0.506 to 0.543) mVxseconds/pM acetylsulfamethazine. Sample chromatograms of blank plasma (a) and a plasma sample from a rabbit 20 minutes after injection with 10 mg/kg sulfamethazine (b) are shown in Figure 1.20 (page 94). A validation summary (Table 1.22), an estimation of precision (Table 1.23), an estimation of intra- and inter-day variability (Table 1.24) and peak ratio determination (Table 1.25) are shown. The peak areas for the samples were converted into concentrations (pM) using the following equations: mVxseconds / (0.364 mVxseconds /pM sulfamethazine) = X pM sulfamethazine mVxseconds / (0.524 mVxseconds /pM acetylsulfamethazine) = X pM acetylsulfamethazine The object of this assay was to compare the peak ratios of sulfamethazine and acetylsulfamethazine so as to determine the percent of sulfamethazine acetylated in 20 minutes. After calculation of the concentration of sulfamethazine and acetylsulfamethazine in each plasma sample, the ratio of acetylsulfamethazine/sulfamethazine + acetylsulfamethazine (a measure of the percent of sulfamethazine acetylated in 20 minutes), was calculated. This ratio was used to determine acetylator status. Rabbits were classified as rapid acetylators if more than 50% of the sulfamethazine was acetylated in 20 minutes or slow acetylators if less than 50% of the sulfamethazine was acetylated in 20 minutes. Using this method a clear bimodal distribution of sulfamethazine acetylation, which directly parallels acetylation characteristics for INH, can be demonstrated in rabbits (Gordon etal. 1973). In nine animals, the plasma sulfamethazine concentration was below the quantitation limit (5 uM). In all of these animals the acetylsulfamethazine peak was very large, confirming that the animals did in fact receive sulfamethazine and suggesting that these animals had very fast acetylation rates. The peak ratios for these animals were determined assuming a sulfamethazine concentration of 5 uM, resulting in peak ratios which ranged from 0.93 to 0.98, which corresponds to 93% to 98% acetylation of sulfamethazine in 20 minutes. This number was derived using the following equation: X uM acetylsulfamethazine/(X uM acetylsulfamethazine + 5 pM sulfamethazine) x 100 = Y% TABLE 1.22 VALIDATION SUMMARY - ACETYLATION ASSAY SULFAMET] -IAZINE (SMZ) AND ACETYLSULFAMETHAZINE (AcSMZ) Sulfamethazine Acetylsulfamethazine Limit of Detection* 0.9 uM 0.9 uM Limit of Quantitation 5 uM (9% CV; acceptance criteria defined at 15%) 5 uM (15% CV; acceptance criteria defined at 15%) Standards Range 5 - 200 uM 5 - 200 uM Sample Range 7-118uM 20 - 220 uM Accuracy See TABLE 1.24 See TABLE 1.24 Specificity A small peak present in blank plasma interfered to a small degree with some of the sulfamethazine and acetylsulfamethazine peaks. Linearity SMZ (plasma): AcSMZ (plasma): 1) . r = 0.998; slope = 0.368 (0.340 to 0.396); y-int = -0.63 (-3.53 to 2.28) 2) . r = 0.999; slope = 0.526 (0.494 to 0.558); y-int = -0.20 (-3.55 to 3.15) Calculation Slope (b = 0) 1 1) . r = 0.998; slope = 0.364 (0.347 to 0.380) 2) . r = 0.999; slope = 0.524 (0.506 to 0.543) Volumetric Variance of Analytical Procedure 7.2 to 7.9% (CV % derived from sulfamerazine peak variability over a total of 95 runs) * Limit of Detection: Defined as three times the background noise. t Calculation slope refers to the slope from the standard curve which was used to calculate the concentration of sulfamethazine and acetylsulfamethazine in the samples. 72 The overall variance of the analytical procedure was determined by calculation of the coefficient of variance using the internal standard sulfamerazine. Sulfamerazine was included as either a 10 pL aliquot of a 250 pM solution or a 10 pL aliquot of a 1 mM solution. The concentration of sulfamerazine was changed to increase the peak height after 23 runs. After 23 runs of samples and standards including a 10 pL aliquot of a 250 pM solution, the coefficient of variation of the peak areas was 7.9%. After 72 runs of samples and standards including a 10 pL aliquot of a 1 mM solution, the coefficient of variation of the peak areas was 7.2%. Therefore, the internal standard confirmed the overall variance of the analytical procedure to be between 7.2% to 7.9% (overall variance includes potential variance introduced by volumetric, photometric, autoinjector injection volume (35 pL), time and temperature variables). TABLE 1.23 PRECISION OF THE PROCEDURE FOR DETERMINATION OF SULFAMETHAZINE AND ACETYLSULFAMETHAZINE IN PLASMA . Sulfamethazine Acetylsulfamethazine uM CV%(SD/mean) uM CV%(SD/mean) 5(n = 3) 9.3 5(n = 3) 14.9 10 (n = 3) 1J 10 (n = 3) 5.5 25 (n = 3) 4.7 25 (n = 3) 6.3 50 (n = 4) 4.4 50 (n = 4) 3.5 100 (n = 4) 5.8 100 (n = 4) 3.8 150 (n = 1) N/A 150 (n = 1) N/A 200 (n = 1) N/A 200 (n = 1) N/A (mean ± SD) 5.2 ± 2.8% (mean ± SD) 6.8 ± 4.7% N/A - not available TABLE 1.24 PHENOTYPING: INTRA- AND INTER-DAY VARIABILITY Compound Concentration Added (LLM) Concentration Found (uM) Intra-Day (n = 5) C V % Inter-Day (n = 9) C V % Sulfamethazine 100 108 ± 10 9.5 % 108 ± 13 11.7% Acetylsulfamethazine 100 108 ± 1 1 9.8 % 105 ± 9 8.4 % 73 Control plasma samples spiked with equal concentrations of sulfamethazine and acetylsulfamethazine were run along with actual plasma samples. These control samples (two at 25 uM and 9 at 100 uM) were used as a check that a 50:50 concentration mixture resulted in a 50% acetylation ratio. The standards, which were interspersed among the samples, showed an average peak ratio of 47.7 + 3.6 (SD; n = 11). TABLE 1.25 STANDARD PEAK RATIO DETERMINATION Ratio: Peak Ratio Sulfamethazine Acetylsulfamethazine Peak Ratio SMZ: AcSMZ* added (UM) (MM) Detected** 0 : 200 u M 1.00 N/D*** 197 1.00 50: 150 uM 0.75 45 147 0.77 100: 100 u M 0.50 97 97 0.50 150 : 50 uM 0.25 150 50 0.25 200 :0 uM 0 195 N/D 0 * SMZ: Sulfamethazine; AcSMZ: Acetylsulfamethazine ** Peak Ratio = AcSMZ (uM) / SMZ (uM) + AcSMZ (uM) * * * N/D = not detectable Statistics Plasma ASAL and ALT activities were logarithmically-transformed so as to give normally distributed data suitable for statistical analysis. For the purpose of clarity, plasma ASAL and ALT activities are presented in the un-transformed form (mean ± standard error; calculated from the raw data) in Tables and Figure legends but as the log-transformed data in Figures. Pearson's product moment correlation coefficients were used for correlation analyses of parametric data. Spearman's rank correlation coefficients were used for correlation analyses of non-parametric data. Correlation coefficients given in the results section are Pearson's product moment correlation coefficients unless indicated to be 74 Spearman's rank correlation coefficients. All data are presented as mean ± standard error unless indicated as mean ± standard deviation. Conversion of correlation coefficients (r) to coefficients of determination (r2) was done in order to more accurately determine the biological significance of the correlations (considered significant when r2 > 0.50). T-tests, assuming either equal or unequal variances (as determined by F-tests), were used when comparing two groups (Zar, 1984). Analysis of variance (ANOVA) and the Newman-Keul's multiple comparison test were done for comparison of more than two groups (Zar, 1984). All data are presented as mean ± standard error unless indicated otherwise. Glutathione S tandard C u r v e 0 50 100 150 200 250 300 350 nmoles Glutathione FIGURE 1.1 GLUTATHIONE STANDARD CURVE The glutathione standard curve had a correlation coefficient (r) of 0.999, a slope of 4.4 x 10"3 (95% confidence intervals 4.3 x 10"3 to 4.5 x 10'3) absorbance units/nmole glutathione and a y-intercept of 1.1 x 10~2 (95% confidence intervals -6.4 x 10'3 to 2.9 x 10'2) absorbance units. Standard deviation (n = 3) of each point smaller than symbols. Detailed methodology is described on page 36. 76 l e e . e_ 100. a. 3 S . 0 3 0 . A 4 0 . 0 _ S 0 . 0 _ 6 4 3 . 0 . 2 0 . 0 . 2 « . y _ 0 . Q L 6 0 0 . 0 5 0 0 . 0 100. 0_ 3 0 . 0_ 6 0 . 0 _ 4 0 . 0 _ 6 0 0 . 0 6 0 0 . 0 6 0 0 . 0 FIGURE 1.2 RESORUFIN SCANS (EMISSION AND EXCITATION). Figure A shows an emission scan using wavelengths 560 to 600 nm, an excitation wavelength of 530 nm and in the presence of buffer, microsomes, ethoxyresorufin substrate and 5 nmoles resorufin. Figure B shows an excitation scan using wavelengths from 500 to 600 nm using an emission wavelength of 585 nm and a blank standard curve solution (containing buffer, microsomes, ethoxyresorufin substrate and no resorufin). Figure C is also an excitation scan using wavelengths from 500 to 600 nm, an emission wavelength of 585 nm and in the presence of 5 nmoles resorufin. Detailed methodology is described on page 38. FIGURE 1.3 RESORUFIN STANDARD CURVE The standard curve for resorufin had a correlation coefficient (r) of 0.999, a slope of 0.322 (95% confidence intervals 0.312 to 0.332) absorbance units/nmole resorufin and a y-intercept of 5.6 x 10"3 (95% confidence intervals -1.6 x 10"2 to 2.7 x 10"2) absorbance units. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3. Detailed methodology is described on page 38. FIGURE 1.4 RESORUFIN STANDARD CURVES AT EXCITATION WAVELENGTHS OF 550 nm AND 575 nm. Comparison of standard curves using excitation wavelengths of 550 and 575 nm demonstrates the increased slope achieved using an excitation wavelength of 575 nm [r = 0.999; slope = 0.292 (95% confidence intervals 0.25 to 0.335) fluorescence units/nmole resorufin; y-intercept = -0.05 (95% confidence intervals -0.16 to 0.062)] versus an excitation wavelength of 550 nm [r = 0.999; slope = 0.200 (95% confidence intervals 0.17 to 0.23) fluorescence units/nmole resorufin; y-intercept = -0.036 (95% confidence intervals -0.12 to 0.051)]. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3. Detailed methodology is described on page 38. Derivatised-lsoniazid Standard Curve 70 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 uM Isoniazid FIGURE 1.5 STANDARD CURVE FOR ISONICOTINYL-3-METHOXYBENZALDHYDRAZONE IN PLASMA The calibration curve of isonicotinyl-3-methoxybenzaldhydrazone (derivatized INH) concentration (uM) versus the ratio of isonicotinyl-3-methoxybenzaldhydrazone to P-fluorenone had a correlation coefficient (r) of 0.999, a slope of 0.05 (95% confidence intervals 0.049 to 0.052) isonicotinyl-3-methoxybenzaldhydrazone:P-fluorenone/uM isonicotinyl-3-methoxybenzaldhydrazone and a y-intercept of 0.08 (95% confidence intervals -0.27 to 0.44) isorucotinyl-3-memoxyben2aldhydrazone:P-fluorenone. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3. Detailed methodology is described on page 42. Derivatised-Acetylhydrazine Standard Crv CD C 5 O C CD 0 25 50 75 100 125 150 175 200 225 uM Acetylhydrazine F I G U R E 1.6 S T A N D A R D C U R V E F O R J V - A C E T Y L - 3 -M E T H O X Y B E N Z A L D H Y D R A Z O N E I N P L A S M A The calibration curve of iV-acetyl-3-methoxybenzaldhydrazone (derivatized acetylhydrazine) concentration (uM) versus the ratio of Ar-acetyl-3-methoxybenzaldhydrazone to P-fluorenone had a correlation coefficient (r) of 0.999, a slope of 0.018 (95% confidence intervals 0.018 to 0.019) JV-acetyl-3-methoxybenzalcmydrazone:P-fluorenone/LM and a y-intercept of 0.003 (95% confidence intervals -0.04 to 0.05) A'-acetyl-3-methoxybenzaldhydrazone:P-fluorenone. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3. Detailed methodology is described on page 42. Derivatised-Hydrazine Standard Curve 20 0 25 50 75 100 125 150 175 200 225 uM Hydrazine FIGURE 1.7 STANDARD CURVE FOR 3-METHOXYBENZALDAZINE IN PLASMA The calibration curve of 3-methoxybenzaldazine (derivatized hydrazine) concentration (|iM) versus the ratio of 3-methoxybenzaldazine to P-fluorenone had a correlation coefficient (r) of 0.999, a slope of 0.074 (95% confidence intervals 0.072 to 0.077) 3-methoxybenzaldazine:P-fluorenone/iaM 3-methoxybenzaldazine and a y-intercept of 0.03 (95% confidence intervals -0.16 to 0.22) 3-methoxybenzaldazine:P-fluorenone. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3. Detailed methodology is described on page 42. 82 FIGURE 1.8 CHROMATOGRAMS OF ISOMCOTINYL-5-METHOXYBENZALDHYDRAZONE, JV-ACETYL-3-METHOXYBENZALDHYDRAZONE AND 5-METHOXYBENZALD AZINE IN PLASMA. Chromatograms (gradient) showing A) blank plasma with the excess 5-methoxybenzaldehyde (derivatising reagent) peak (1) and P-fluorenone (internal standard) (2) and an actual sample with B) N-acetyl-3-methoxybenzaldhydrazone (deriatized acetylhydrazine) (3), isonicotinyl-3-methoxybenzaldhydrazone (derivatized INH) (4) and 3-methoxybenzaldazine (derivatized hydrazine) (5) peaks. X-axis is time and the scale is 0.2 ATJFS. Detailed methodology is described on page 42. Derivatised-Hydrazine Standard Curve 0 25 50 75 100 125 150 175 200 225 Plasma Hydrazine Concentration (uM) FIGURE 1.9 STANDARD CURVE FOR 3-METHOXYBENZALDAZINE IN 12 HOUR PLASMA SAMPLES. The standard curve for J-methoxybenzaldazine (derivatized hydrazine) had a correlation coefficient (r) of 0.999, a slope of 0.564 (95% confidence intervals 0.553 to 0.575) mVxseconds/uM 3-methoxybenzaldazine and a y-intercept of 1.5 (95% confidence intervals 0.37 to 2.66) mVxseconds. Error bars (hidden by the symbols) are mean ± standard deviation, n = 3. Detailed methodology is described on page 46. 84 I sco Qwdfaeirch F I G U R E 1.10 C H R O M A T O G R A M S F O R 3 - M E T H O X Y B E N Z A L D A Z I N E IN 12 H O U R P L A S M A S A M P L E S . Chromatograms (isocratic) showing A) blank plasma with a P-fluorenone peak (1) and B) an actual plasma sample with a 5-methoxybenzaldazine (derivatized hydrazine) (2) peak. Detailed methodology is described on page 46. Derivatised-Hydrazine Standard Curve 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Plasma Hydrazine Concentration (uM) FIGURE 1.11 STANDARD CURVE FOR 3-METHOXYBENZALD AZINE IN 24 AND 32 HOUR PLASMA SAMPLES. The standard curve for 5-methoxybenzaldazine (derivatized hydrazine) had a correlation coefficient (r) of 0.999, a slope of 0.235 (95% confidence intervals 0.222 to 0.248) mVxseconds/uM 3-methoxybenzaldazine and a y-intercept of-0.64 (95% confidence intervals -1.41 to 0.13) mVxseconds. Detailed methodology is described on page 51. FIGURE 1.12 CHROMATOGRAMS FOR 3-METHOXYBENZALDAZINE IN 24 AND 32 HOUR PLASMA SAMPLES. Chromatograms (isocratic) for 5-methoxybenzaldazine (derivatized hydrazine) showing A) blank plasma with a P-fluorenone peak (1) and B) an actual plasma sample with a 3-methoxybenzaldazine (2) peak. The chromatograms appears slightly different from determination of the 12 hour samples due to a different sensitivity of the detector. Detailed methodology is described on page 51. Hepatic Amidase Standard Curve (Hydrazine) 0 25 50 75 100 125 150 175 200 225 250 275 Microsomal Hydrazine Concentration (uM) FIGURE 1.13 STANDARD CURVE FOR DETERMINATION OF HEPATIC AMIDASE ACTIVITY. The standard curve had a correlation coefficient (r) of 0.998, a slope of 0.152 (95% confidence intervals 0.140 to 0.165) mVxseconds/uM 3-methoxybenzaldazine (derivatized hydrazine) and a y-intercept of-1.19 (95% confidence intervals -2.56 to 1.86) mVxseconds. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3. Detailed methodology is described on page 54. bcoCiiiiiiiiniiil 88 IxoCfcenltesearch 1BB| x D E T i \ DET 11 riKE(SEC) 428 F I G U R E 1 . 1 4 C H R O M A T O G R A M S F O R H E P A T I C A M I D A S E A S S A Y . Chromatograms (isocratic) showing A) a P-fluorenone peak (1) in microsomes with no added INH and B) an actual microsomal incubation (30 minutes) showing a 5-methoxybenzaldazine (derivatized hydrazine) (2) peak from hydrolysis of INH. Detailed methodology is described on page 54. Plasma Amidase Standard Curve (Hydrazine) 0 10 20 30 40 50 60 70 Plasma Hydrazine Concentration (uM) FIGURE 1.15 STANDARD CURVE FOR DETERMINATION OF PLASMA AMIDASE ACTIVITY. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.166 (95% confidence intervals 0.163 to 0.170) mVxseconds/uM 5-methoxybenzaldazine (derivatized hydrazine) and a y-intercept of 0.040 (95% confidence intervals -0.062 to 0.142) mVxseconds. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3. Detailed methodology is described on page 59. 90 Isco Chenfesearcb n D! T L E > s i. p i DEI 11 "St J L. INE(SEC) 428 FIGURE 1.16 CHROMATOGRAMS FOR PLASMA AMIDASE ASSAY. Chromatograms (isocratic) showing A) a 9-fiuorenone peak (1) in plasma with no added INH and B) an actual plasma incubation (30 minutes) showing a 5-methoxybenzaldazine (derivatized hydrazine) (2) peak (equal to 5 uM hydrazine) from hydrolysis of INH. Detailed methodology is described on page 59. Hydrazine Assay Standard Curves + Hepatic amidase • Plasma Amidase A 24/32 Plasma Hz pmoles Hydrazine FIGURE 1.17 STANDARD CURVES FOR HYDRAZINE ASSAYS USING THE PERCHLORIC ACID AND ACETONITRILE DENATURATION STEP. Three standard curves using the same analytical procedure are shown. There is no difference between the slopes of all three of the standard curves. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3 (except 24/32 hour hydrazine assay, n = 2). Sulfamethazine S tandard C u r v e Sul famethazine Concentrat ion (uM) FIGURE 1.18 STANDARD CURVE FOR SULFAMETHAZINE IN PLASMA. The standard curve had a correlation coefficient (r) of 0.998, a slope of 0.368 (95% confidence intervals 0.340 to 0.396) mVxseconds/uM sulfamethazine and a y-intercept of-0.63 (95% confidence intervals -3.53 to 2.28) peak area units. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3 or 4. Detailed methodology is described on page 68. Acetylsulfamethazine Standard Curve 0 25 50 75 100 125 150 175 200 225 Acetylsulfamethazine Concentration (uM) FIGURE 1.19 STANDARD CURVE FOR ACETYLSULFAMETHAZINE IN PLASMA. The standard curve had a correlation coefficient (r) of 0.999, a slope of 0.526 (95% confidence intervals 0.494 to 0.558) mVxseconds /uM acetylsulfamethazine and a y-intercept of -0.20 (95% confidence intervals -3.55 to 3.15) mVxseconds. Error bars (some hidden by the symbols) are mean ± standard deviation, n = 3 or 4. Detailed methodology is described on page 68. 94 Isco QwnKsaaarcii F I G U R E 1.20 C H R O M A T O G R A M S S H O W I N G S U L F A M E T H A Z I N E , S U L F A M E T H A Z I N E A N D A C E T Y L S U L F A M E T H A Z I N E I N P L A S M A . Chromatograms (isocratic) showing A) blank plasma with a sulfamerazine peak (1) and unknown peak present in blank plasma (2) and B) an actual plasma sample showing sulfamethazine (3) and acetylsulfamethazine (4) peaks. Scale is 0.05 AUFS. Detailed methodology is described on page 68. 95 RESULTS STUDY 1 STUDY 1 HYPOTHESES • INH administration causes biochemically detectable hepatic necrosis and steatosis in male and female rabbits; female rabbits are more susceptible to these effects. • Rate of acetylation of sulfamethazine correlates with the severity of INH-induced hepatotoxicity. One female rabbit died unexpectedly during the acetylation phenotyping (possibly due to an adverse reaction to sulfamethazine); therefore, the following results are from the remaining eight male and seven female rabbits. There were no differences in baseline plasma ASAL activity (4.6 ± 0.4 Takahara units in males vs. 5.4 + 0.7 Takahara units in females; p = 0.28) or plasma triglyceride levels (0.9 ± 0.1 mM vs. 0.9 ± 0.1 mM; p = 1.0) between male and female rabbits, prior to administration of INH. After treatment with INH, plasma ASAL activity, liver triglyceride levels and plasma triglyceride concentrations were all significantly increased with no detectable differences between males and females (Table 2.1). Similarity, there were no significant differences between male and female rabbits for time to peak plasma ASAL activity (65.0 ± 8.2 hours versus 67.4 ± 6.0 hours, not significant) and time to peak plasma triglyceride concentration (61.0 ± 2.8 hours versus 62.9 ± 5.4 hours, not significant), respectively. The data for each rabbit are shown in Table 2.2. Fourteen out of the fifteen rabbits were rapid acetylators (range of 69-98% sulfamethazine acetylated in 20 minutes) and only one was a slow acetylator (30% of the 96 sulfamethazine acetylated in 20 minutes). Not including the slow acetylator, correlational analysis of the percent sulfamethazine acetylated in 20 minutes revealed a significant negative correlation with peak plasma triglyceride concentration (r = -0.56, r2 = 0.31, n = 14, p < 0.05, Figure 2.1; page 98) but not with log peak plasma ASAL activity (r = -0.38, r2 = 0.14, n = 14, p > 0.05, Figure 2.2; page 99) or hepatic triglyceride content (r = -0.44, r2 = 0.19, n = 14, p > 0.05). These correlations explained only 31%, 14% and 19% of the total variance of these measures, respectively. Correlational analyses of INH-induced hepatic necrosis, steatosis and hypertriglyceridemia were attempted. Liver triglyceride accumulation was found to correlate significantly with peak plasma triglyceride concentration (r = +0.69, r2 = 0.48, n = 15, p < 0.005, Figure 2.3; page 100). Log peak plasma ASAL activity correlated significantly with hepatic triglyceride content (r = +0.60, r2 = 0.36, n = 15, p < 0.02, Figure 2.4; page 101) but not with peak plasma triglyceride concentration (r = +0.49, r2 = 0.24, n = 15, p > 0.05, Figure 2.5; page 102). The correlations each accounted for less than 50% of the total variance. Evaluation of the time-course of changes in plasma ASAL activity and plasma triglyceride concentration over a 96 hour time-course was made in order to determine the relationship between these two markers of toxicity. For example, we attempted to determine whether elevation of one variable is followed by, and therefore may be the cause of, elevation of the other. The average time to peak plasma ASAL activity (66.1 ±5.0 hours, n = 15, range from 32 to 96 hours) did not differ from the average time to peak plasma triglyceride concentration (61.9 + 3.3 hours, n = 15, range from 48 to 80 hours, p 97 > 0.05). Time to peak plasma ASAL activity and the corresponding time to peak plasma triglyceride concentrations in individual rabbits did not correlate [r = -0.15 (Spearman's rank correlation), r2 = 0.02, n = 15, p > 0.05, Figure 2.6; page 103]. This lack of correlation is reflected by the fact that plasma ASAL activity became maximal before peak plasma triglyceride concentration in six rabbits, plasma triglyceride concentration became maximal before peak plasma ASAL activity in eight rabbits and in one rabbit, both attained maximal values at the same time. 98 10 20 30 40 50 60 70 80 90 100 110 Acetylation rate % Sulfamethazine Acetylated in 2 0 minutes FIGURE 2.1 CORRELATION OF ACETYLATION RATE WITH PEAK PLASMA TRIGLYCERIDES. Correlation of acetylation rate (percent sulfamethazine acetylated in 20 minutes) with peak plasma triglycerides (mM) is significant (r = -0.56, r 2 = 0.31, n = 14, p < 0.05). The slow acetylator rabbit is shown (•) but was not included in the calculation of the correlation coefficient. 99 F I G U R E 2.2 C O R R E L A T I O N OF A C E T Y L A T I O N R A T E W I T H A N I N D E X OF H E P A T I C C E L L D A M A G E . Correlation of acetylation rate (percent sulfamethazine acetylated in 20 minutes) with peak plasma ASAL activity (log Takahara units) is not significant (r = -0.38, r2 = 0.14, n = 14, p > 0.05). The slow acetylator rabbit is shown (•) but was not included in the calculation of the correlation coefficient. 100 FIGURE 2.3 CORRELATION OF HEPATIC TRIGLYCERIDE CONTENT WITH PEAK PLASMA TRIGLYCERIDE CONCENTRATION. Correlation of hepatic triglyceride content (mg triglyceride/g liver tissue) with peak plasma triglyceride concentration (mM) is significant with a positive slope (r = +0.69, r2 = 0.48, n = 15, p < 0.005). 101 FIGURE 2.4 CORRELATION OF HEPATIC TRIGLYCERIDE CONTENT WITH AN INDEX OF HEPATIC CELL DAMAGE. Correlation of hepatic triglyceride content (mg triglyceride/g liver tissue) with log peak plasma ASAL activity (log Takahara units) is significant with a positive slope (r = +0.60, r2 = 0.36, n = 15, p < 0.02). 102 FIGURE 2.5 CORRELATION OF PEAK PLASMA TRIGLYCERIDE CONCENTRATION WITH AN INDEX OF HEPATIC CELL DAMAGE. Correlation of peak plasma triglyceride concentration (mM) with log peak plasma ASAL activity (log Takahara units) is not significant (r = +0.49, r2 = 0.24, n = 15, p > 0.05). 103 100 *** + * 80 + + + + 60 + + + + 40 - + + 20 n 1 l . J _ i i i i 20 40 60 80 100 Plasma Triglyceride Time to Peak (hours) FIGURE 2.6 CORRELATION OF TIME TO PEAK PLASMA TRIGLYCERIDE CONCENTRATION VERSUS TIME TO PEAK HEPATIC CELL DAMAGE. Correlation of time to peak plasma triglyceride concentration (hours) with time to peak plasma ASAL activity (hours) is not significant (r = -0.15, r2 = 0.02, n = 15, not significant, using Spearman's rank correlation). There are only 11 points visible on the graph because several points overlap. Each asterisk (*) above a point on the graph is used to indicate overlapping points. 104 TABLE 2A BIOCHEMICAL CHANGES PRE- AND POST-ISONIAZID TREATMENT Pre-INH Post-INH Combined Females Males Combined Number of animals (n) 15 7 8 15 Plasma ASAL activity (Takahara Units) 5.0 ± 0 . 4 145 ± 44 127 ± 45 135 ± 301 Plasma triglyceride concentration (mM) 0.9 ± 0 . 1 13.7 ± 2 . 6 12.1 ± 2 . 8 12.9 ± 1.9* Liver triglycerides (mg TG/g liver) 4.1 ± 0 . 6 * 33.0 ± 5 . 4 27.0 ± 3 . 2 29.8 ± 3.2T * n = 4 (non-INH treated controls). 1 p < 0.001 as compared with pre-INH data (T-test). Note: Values are means ± SEM. Initial pre-INH plasma ASAL activity and plasma triglyceride concentrations were measured in plasma samples collected immediately prior to initiating the INH-injection protocol. Liver triglyceride pre-INH values were determined using four control animal livers not treated with INH. The post-INH data for plasma ASAL activity and plasma triglycerides are peak levels. The post-INH data for the liver triglyceride content are from liver samples taken 96 hours after the first dose of INH. 105 TABLE 2.2 INDIVIDUAL RABBIT TOXICITY DATA RABBIT GENDER Peak Plasma ASAL activity (Takahara Units) Liver Triglycerides (mg TG/g liver) Peak Plasma Triglyceride concentration (mM) Acetylator Phenotype (% SMZ acetylated in 20 minutes) 1 Female 93 49 19 84 2 Female 79 31 8.4 94 3 Female 14 4.9 1.9 98 4 Female 153 36 20 30 5 Female 66 24 6.7 77 6 Female 286 48 27 73 7 Female 323 38 7.6 80 8 Male 406 23 6.8 88 9 Male 84 42 15 82 10 Male 29 23 13 73 11 Male 113 22 19 92 12 Male 22 12 2.5 94 13 Male 144 28 21 69 14 Male 184 34 15 91 15 Male 35 32 11 91 mean ± SEM 135 ± 3 0 29.8 ± 3 . 2 12.9 ± 1 . 9 Data for individual rabbits in Study 1 are shown. 106 STUDY 2 STUDY 2 HYPOTHESIS • Plasma levels of acetylhydrazine and/or hydrazine (suspected INH-derived hepatotoxins) correlate with markers of INH-induced hepatic necrosis. Toxicity data for the individual animals in this study are presented in Table 2.2. Plasma concentrations of INH, acetylhydrazine and hydrazine for the individual animals in this study are presented in Table 2.3. Plasma INH concentration at 32 hours did not correlate with plasma ASAL activity (area under the curve) (r = +0.37, r2 = 0.14, n = 15, p > 0.05). Plasma INH concentration at 32 hours correlated significantly with log plasma ASAL activity at 48 hours (r = +0.54, r2 = 0.29, n = 15, p < 0.05) and log peak plasma ASAL activity (r = +0.63, r2 = 0.40, n = 15, p < 0.02); however, these correlations accounted for less than 50% of total variance. Plasma INH at 48 hours did not correlate with any of the three measures of hepatic necrosis e.g. plasma ASAL area under the curve, log plasma ASAL activity at 48 hours or log peak plasma ASAL activity. Plasma acetylhydrazine concentration at 32 hours did not correlate with plasma ASAL activity area under the curve (r = -0.14, r2 = 0.02, n = 15, > 0.05) (Figure 2.7; page 108), log plasma ASAL activity at 48 hours (r = -0.17, r2 = 0.03, n = 15, p > 0.05) (Figure 2.8; page 109) or log peak plasma ASAL activity (r = -0.24, r2 = 0.06, n = 15, p > 0.05) (Figure 2.9; page 110). Plasma acetylhydrazine at 48 hours also did not correlate with any of the markers of hepatic necrosis. Plasma hydrazine concentration at 32 hours correlated significantly with plasma ASAL activity area under the curve (r = +0.73, r2 = 0.54, n = 15, 107 p < 0.002) (Figure 2.10; page 111), log plasma ASAL activity at 48 hours (r = +0.73, r2 = 0.53, n = 15, p < 0.005) (Figure 2.11; page 112) and log peak plasma ASAL activity (r = +0.66, r2 = 0.44, n = 15, p < 0.01) (Figure 2.12; page 113). Plasma hydrazine at 48 hours did not correlate with any of the three measures of hepatic necrosis e.g. plasma ASAL area under the curve, log plasma ASAL activity at 48 hours or log peak plasma ASAL activity. Liver triglyceride accumulation correlated significantly with 48 hour plasma hydrazine concentrations (r = +0.58, r2 = 0.34, n = 15, p < 0.05). However, the correlation did not explain greater than 50% of the total variance. There were no significant correlations between plasma triglycerides (peak plasma triglycerides, plasma triglycerides at 48 hours and plasma triglyceride area under the curve) and 32 or 48 hour plasma INH, acetylhydrazine and hydrazine concentrations (r2 ranging from 0.01 to 0.24). 108 FIGURE 2.7 ACETYLHYDRAZINE PLASMA CONCENTRATION IN 32 HOUR PLASMA SAMPLES VERSUS AREA UNDER THE CURVE OF AN INDEX OF HEPATIC CELL DAMAGE. Correlation of 32 hour plasma acetylhydrazine concentration (LLM) with plasma A S A L activity area under the curve is not significant (r = -0.14, r2 = 0.02, n = 15, p > 0.05). 109 FIGURE 2.8 ACETYLHYDRAZINE PLASMA CONCENTRATION IN 32 HOUR PLASMA SAMPLES VERSUS AN INDEX OF HEPATIC CELL DAMAGE AT 48 HOURS. Correlation of 32 hour plasma acetylhydrazine concentration (uM) with log plasma ASAL activity at 48 hours (log Takahara units) is not significant (r = -0.17, r 2 = 0.03, n = 15, p > 0.05). 110 FIGURE 2.9 ACETYLHYDRAZINE PLASMA CONCENTRATION IN 32 HOUR PLASMA SAMPLES VERSUS AN INDEX OF PEAK HEPATIC CELL DAMAGE. Correlation of 32 hour plasma acetylhydrazine concentration (uM) with log peak plasma ASAL activity (log Takahara units) is not significant (r = -0.24, r 2 = 0.06, n = 15, p > 0.05). I l l FIGURE 2.10 HYDRAZINE PLASMA CONCENTRATION IN 32 HOUR PLASMA SAMPLES VERSUS AREA UNDER THE CURVE OF AN INDEX OF HEPATIC CELL DAMAGE. Correlation of 32 hour plasma hydrazine concentration (uM) with plasma ASAL activity area under the curve is significant with a positive slope (r = +0.73, r2 = 0.54, n = 15, p < 0.002). 112 FIGURE 2.11 HYDRAZINE PLASMA CONCENTRATION IN 32 HOUR PLASMA SAMPLES VERSUS AN INDEX OF HEPATIC CELL DAMAGE AT 48 HOURS. Correlation of 32 hour plasma hydrazine concentration (uM) with log plasma ASAL activity at 48 hours (log Takahara units) is significant with a positive slope (r = +0.73, r 2 = 0.53, n = 15, p < 0.005). 113 FIGURE 2.12 HYDRAZINE PLASMA CONCENTRATION IN 32 HOUR PLASMA SAMPLES VERSUS AN INDEX OF PEAK HEPATIC CELL DAMAGE. Correlation of 32 hour plasma hydrazine concentration (uM) with log peak plasma ASAL activity (log Takahara units) is significant with a positive slope (r = +0.66, r2 = 0.44, n = 15, p < 0.01). 114 TABLE 2.3 PLASMA CONCENTRATIONS OF INH, ACETYLHYDRAZINE AND HYDRAZINE IN INDIVIDUAL RABBITS RABBIT INH (uM) Acetylhydrazine (uM) Hydrazine (uM) 32Hour 48 Hour 32Hour 48 Hour 32 Hour 48 Hour 1 284 30 58 55 63 20 2 217 <20 108 23 65 24 3 143 <20 109 21 51 18 4 426 54 13 7.5 69 21 5 249 <20 89 14 69 16 6 251 33 60 47 50 27 7 293 40 47 28 29 15 8 311 20 117 34 71 23 9 320 21 70 25 60 19 10 346 35 71 37 82 32 11 288 22 91 31 104 28 12 180 <20 123 27 45 14 13 239 <20 168 27 91 17 14 322 27 55 24 44 14 15 147 <20 165 21 62 18 mean ± SEM 268 ± 20 31 ± 3 . 6 90 ± 11 28 ± 3 . 1 64 ± 5.0 20 ± 1.4 n= 15 n = 9 n= 15 n= 15 n=15 n= 15 Data for individual rabbits in Study 2 are shown. 115 STUDY3 STUDY 3 HYPOTHESES • Decreased activity of the hepatic reductase enzyme (by methimazole pretreatment) should decrease the conversion of hydrazine to reactive and toxic intermediates in vivo and result in a decrease in the severity of INH-induced hepatic necrosis. • Conversely, increased reductase activity (by L-thyroxine pretreatment) should result in an increase in the conversion of hydrazine to reactive and toxic intermediates in vivo and an increase in the severity of INH-induced hepatic necrosis. Animals One of the difficulties of using our model of INH-induced hepatotoxicity is the occasional occurrence of central nervous system toxicity (seizures) which can result in premature death of an animal (Sarich et al, 1995). In the present study, 7 rabbits died prematurely and thus their livers were not available for analyses. Since there were no significant differences in any of the relevant toxicological parameters (Table 2.4) between the two vehicle control groups which only received INH (L-thyroxine vehicle n = 13; methimazole vehicle n = 9), the data were pooled to form one single "INH-only" group. Of these twenty-two rabbits, three died during the INH-injection protocol: one at 24 hours (methimazole vehicle), one at 48 hours (L-thyroxine vehicle) and one at 72 hours (L-thyroxine vehicle) (total remaining n = 19). An estimate of hepatic necrosis (plasma liver enzyme activities) was not available from the animal which died at 24 hours after INH administration, but was available from the other two animals which died before the end of the study. 116 Of thirteen rabbits in the group receiving both L-thyroxine and INH (T4-INH group), four died before the end of the study: one at 32 hours, two at 56 hours and one at 72 hours after initiation of the INH-injection protocol. Hypothyroidism Pilot studies to find a methimazole dosing regimen which would decrease both plasma free T4 levels and hepatic reductase activities were performed. The first consisted of injections (i.p.) of methimazole at 10 mg/kg/day (0.09 mmoles/kg/day) for seven days. Animals were sacrificed 45 hours after the last methimazole dose. This timing was followed in order to simulate the state the animals would be in at the time of the last dose of INH during the INH protocol. (It was later decided to continue the administration of methimazole during INH administration). Neither plasma free T4 levels (average 14.7 pmoles/L, n = 3, versus average 12.2 pmoles/L, n = 2) nor hepatic reductase activities (average 145.2 nmoles/min/mg protein, n = 3, versus average 155.0 nmoles/min/mg protein, n = 2) appeared decreased in the methimazole group versus control animals, respectively. The second pilot study involved giving increased doses (i.p.) of methimazole: 2x50 mg/kg/day (2x0.44 mmoles/kg/day) at 12 hour intervals for nine days and 2><150 mg/kg/day (2x1.3 mmoles/kg/day) at 12 hour intervals for nine days. Two of the three animals in the 2x150 mg/kg/day (2x1.3 mmoles/kg/day) group died after four doses of methimazole and the third died after the fifth dose. In the 2x50 mg/kg/day (2x0.44 mmoles/kg/day) animals, neither plasma free T4 levels (average 10 pmoles/L, n = 3, versus average 12 pmoles/L, n = 2) nor hepatic reductase activities [average 135.5 117 nmoles/min/mg protein, n = 3, versus average 200.4 (actual numbers 148.2 and 252.5) nmoles/min/mg protein, n = 2] appeared decreased in the methimazole group versus the control animals, respectively. Animals were sacrificed at 45 hours after the last methimazole dose. Since animals receiving methimazole at 2*50 mg/kg/day (2*0.44 mmoles/kg/day) had been sacrificed 45 hours after the last dose of methimazole, it was thought that the plasma free T4 and hepatic reductase levels may have partially returned to baseline between the last methimazole dose and sacrifice 45 hours later. In addition, it was clear that the maximally tolerated dose of methimazole was 2x150 mg/kg/day (2*1.3 mmoles/kg/day). Therefore, it was decided to continue the 2*50 mg/kg/day (2x0.44 mmoles/kg/day) regimen during the INH injection protocol and use it as the methimazole pretreatment schedule. In the first attempt to combine this model of experimentally-induced hypothyroidism in rabbits with the INH protocol, four animals were randomised to receive methimazole (50 mg/kg; 0.44 mmoles/kg i.p. every 12 hours) for 14 days. One rabbit died after four days of methimazole-therapy: After nine days pretreatment, plasma free T4 levels appeared decreased (average 8.0 pmoles/L, n = 3) compared to baseline levels (average 15.3 pmoles/L). The INH protocol was begun on day 10 in the remaining three methimazole treated rabbits (and in four water-treated controls); however, all three methimazole-treated animals died after experiencing convulsions within 3 hours of the first INH dose. As a result, the dose of methimazole was lowered and the duration of administration lengthened by adding the drug to the drinking water of five rabbits. Plasma 118 free T4 levels in an equal number of controls given regular drinking water were monitored as well. After seven days of methimazole at a dose of 0.0313% w/v (approximately 25 mg methimazole/kg/day; 0.22 mmoles/kg/day) in the drinking water, plasma free T4 levels were not decreased. The dose was increased to 0.0626% w/v (approximately 50 mg/kg/day; 0.44 mmoles/kg/day). On day 27, the plasma free T4 levels in the methimazole group were significantly decreased (5.6 ± 1.0 pmoles/L, n = 5) versus baseline levels (12.4 ± 0.6 pmoles/L, n = 5, p = 0.002) and the plasma free T4 levels in the regular water group were unchanged (13.3 ± 0.4 pmoles/L, n = 5) versus baseline levels (13.2 ± 0.5 pmoles/L, p = 0.86). Two methimazole-treated animals died within 3 hours and one died within 6 hours of initiation of the INH injection protocol. Although two rabbits in the methimazole group survived to the end of the study, the low numbers prevented statistical comparison to the INH-only and control groups. Hyperthyroidism A pilot study to find an L-thyroxine dosing regimen which could increase plasma free T4 levels and hepatic reductase activities was carried out. The regimen consisted of i.p. injections of L-thyroxine (dissolved in 0.02 M NH4OH) at a dose of 0.4 mg/kg/day (5.1 umoles/kg/day) for seven days. The animals were sacrificed 45 hours after the last L -thyroxine dose. Both the plasma free T4 levels (average 77.2 pmoles/L, n = 3, versus 12.2 pmoles/L, n = 2) and the hepatic reductase activities (average 230.7 nmoles/min/mg protein, n = 3 versus 155.0 nmoles/min/mg protein, n = 2) appeared increased in the L -thyroxine group versus the control animals, respectively. 119 When this model of hyperthyroidism was combined with the INH-injection protocol, the L-thyroxine injections were given for five days prior to, two days during and two days after the INH-injection protocol. Plasma free T4 levels in the T4-INH group were significantly elevated after five days of pretreatment with L-thyroxine (the first day of INH injections) (58.4 + 8.1 pmoles/L, n = 13) and after nine days of pretreatment with L-thyroxine (day 10) (70.6 ± 4.2 pmoles/L, n = 9) compared to baseline levels (13.6 ± 0.8 pmoles/L, n = 13) (p < 0.0001, ANOVA; p < 0.05, Newman-Keul's) (Figure 2.13; page 124). In comparison, plasma free T4 levels in the INH-only group were not elevated on the first day of INH injections (13.9 ± 0.9 pmoles/L, n = 22) nor were they elevated at the end of the study (11.1 +0.7 pmoles/L, n = 19) as compared to baseline levels (12.2 + 0.4 pmoles/L, n = 22). However, a small but significant decrease did occur between day 5 and day 10 (p = 0.03, ANOVA; p < 0.05, Newman-Keul's) (Figure 2.14; page 125). Peak plasma ASAL activities in the T4-INH (31.3 ± 20.1 Takahara units, n = 13) and INH-only groups (56.0 ± 19.7 Takahara units, n = 21) were significantly increased versus the vehicle control group (4.3 ± 0.5 Takahara units, n = 9) (p < 0.0001, ANOVA; p < 0.05 Newman-Keul's, statistics on logarithmically-transformed data) (Figure 2.15; page 126; Table 2.5). The peak plasma ASAL activity in the INH-only group was also significantly greater than the T4-INH group. The peak plasma ASAL activities in the two rabbits which survived the methimazole-INH treatment were 19.2 and 20.3 Takahara units, slightly lower than average levels in the T4-INH group. The peak plasma ALT activity in the INH-only group (68.4 ± 18.5 units/L, n = 21) was significantly greater than the activity in the T4-INH (24.8 ± 5.4 units/L, n = 13) and 120 vehicle control groups (27.0 ± 1.9 units/L, n = 9) based on the ANOVA (p = 0.03), but not on the Newman-KeuPs (p > 0.05) (statistics done on logarithmically-transformed data) (Figure 2.16; page 127; Table 2.5). The peak plasma ALT activities in the two rabbits which survived the methimazole-INH treatment were 29.0 and 42.4 units/L. Hepatic triglyceride accumulation was increased significantly in the INH-only group (27.5 ± 3.2 mg triglyceride/ g liver, n = 19) versus the T4-INH (7.4 ± 2.4 mg triglyceride/ g liver, n = 9) and vehicle control groups (9.3 ± 1.8 mg triglyceride/ g liver, n = 9) (p < 0.0001, ANOVA; p < 0.05 Newman-Red's) (Figure 2.17; page 128; Table 2.5). Hepatic cytochrome P-450 levels in both the T4-INH (1.7 ± 0 . 1 nmoles/mg protein, n = 9) and INH-only groups (1.8 ± 0.1 nmoles/mg protein, n = 19) were significantly decreased compared to the vehicle control group (3.5 ± 0.3 nmoles/mg protein, n = 9) (p < 0.0001, ANOVA; p < 0.05 Newman-Keul's) (Figure 2.18; page 129; Table 2.5). Hepatic reductase activity was increased in the T4-INH group (648 ± 44 nmoles/minute/mg protein, n = 9) versus the INH-only (479 ± 23 nmoles/minute/mg protein, n = 19) and vehicle control groups (421. ± 25 nmoles/minute/mg protein, n = 9) (p = 0.0001, ANOVA; p < 0.05 Newman-Keul's) (Figure 2.19; page 130; Table 2.5). Interestingly, hepatic reductase activity in the two rabbits which survived the methimazole-INH treatment were 288 and 379 nmoles/minute/mg protein, lower than average in the other groups. /7-Nitrophenol hydroxylase activity, a measure of CYP2E1 activity, in the INH-only group (1.8 ± 0.2 nmoles/minute/mg protein, n = 19) was significantly decreased as 121 compared to the T4-INH (3.8 ± 0.6 nmoles/minute/mg protein, n = 9) and vehicle control groups (2.9 ± 0.4 nmoles/minute/mg protein, n = 9) (p = 0.0009, ANOVA; p < 0.05 Newman-Keul's) (Figure 2.20; page 131; Table 2.5). Combining both T4-INH and INH-only treated animals (n = 28), correlation of reductase activity with plasma ASAL activity (r = -0.49, p < 0.01) is significant while correlation of reductase activity with plasma ALT activity (r = -0.33, p > 0.05) is not. Nevertheless, the coefficients of determination (r2 = 0.24 and r2 = 0.11, respectively) are both below 0.5, weakening the causative importance of the correlations. Correlations of CYP2E1 activity with plasma ASAL activity (r = -0.62, p < 0.001) and plasma ALT activity (r = -0.48, p < 0.01) are significant. Nevertheless, the coefficients of determination (r2 = 0.38 and r2 = 0.23, respectively) are also both below 0.5. Correlation of reductase activity with hepatic triglyceride accumulation (r = -0.63, n = 28, p < 0.001) indicates a negative association; however, the coefficient of determination (r2 = 0.40) is below 0.5. Correlation of CYP2E1 activity with hepatic triglyceride accumulation (r = -0.45, r2 = 0.20, n = 28, p < 0.02) also indicates a weak negative association. The activity of the reductase and CYP2E1 enzymes as well as the P-450 levels at the time of the first INH injection (0 hours) was measured in a follow-up study in methimazole and L-thyroxine pre-treated animals. Five methimazole-treated animals (26 days of methimazole in the drinking water) and five L-thyroxine-treated animals [5 days of i.p. injections of L-thyroxine at 0.4 mg/kg/day (5.1 pmoles/kg/day)] were compared to the control animals mentioned above (n = 9). 122 In the follow-up study, hepatic cytochrome P-450 levels in the L-thyroxine pretreatment group (2.1 ± 0 . 1 nmoles/mg protein, n = 5), but not in the methimazole pretreatment group (3.0 ± 0.4 nmoles/mg protein, n = 5), were significantly decreased compared to the vehicle control group (3.5 ± 0.3 nmoles/mg protein, n = 9) (p = 0.015, A N O V A ; p < 0.05 Newman-Keul ' s ) (Figure 2.21; page 132; Table 2.6). Hepatic reductase activity was significantly decreased in the methimazole pretreatment group (170 ± 27 nmoles/minute/mg protein, n = 5) versus both the vehicle control group (421 ± 25 nmoles/minute/mg protein, n = 9) and the L-thyroxine pretreatment group (466 ± 27 nmoles/minute/mg protein, n = 5) (p < 0.0001, A N O V A ; p < 0.05 Newman-Keu l ' s ) (Figure 2.22; page 133; Table 2.6). Hepatic C Y P 2 E 1 activities in the L-thyroxine (2.1 ± 0.2 nmoles/minute/mg protein, n = 5) and methimazole (2.7 ± 0.3 nmoles/minute/mg protein, n = 5) pretreatment groups were not significantly different from the control group (2.9 ± 0.4 nmoles/minute/mg protein, n = 9) (p = 0.38, A N O V A ) (Figure 2.23; page 134; Table 2.6). T h e effect o f hyperthyroidism on body and liver weight was monitored from the beginning o f pretreatment with L-thyroxine until the end o f the study (a total o f 9 days). T h e change in body weight (decrease) in the T 4 - I N H animals (-0.54 ± 0.06 kg, n = 9) was significantly greater than in the I N H - o n l y animals (-0.15 ± 0.05 kg , n = 11; p < 0.00001). A t the end o f the study, the liver weight in the T 4 - I N H animals ( 6 5 . 5 + 2 . 8 kg , n = 9) was significantly less than the liver weight in the I N H - o n l y animals (80.6 ± 4.3 g, n = 11; p = 0.01). T h e animals in the follow-up study, which did not receive I N H , were also compared. O n c e again, there was a significantly greater change in body weight in the T 4 -1 2 3 pretreated animals (-0.24 ± 0.06 kg, n = 5) versus the animals receiving T4-vehicle only (+0.08 ± 0.07 kg, n = 5; p = 0.007) which actually gained weight over the 5 day pretreatment. The liver weight in the T4-pretreated animals (65.7 + 2.1 g, n = 5) was again significantly less than the liver weight in the animals receiving T4-vehicle only (94.9 ±5.7g, n = 5;p = 0.001). 124 90 o 80 -E Q. 70 —^^  V) CD 60 — > Le 50 \— 40 -CD CD i _ 30 L L CO 20 -E CO CO 10 DL 0 ANOVA p < 0.0001, Newman Keul's p < 0.05 Baseline < Pre-INH = Post-INH _ * Baseline,n=13 Pre-INH,n=13 Post-INH,n=9 T4-INH GROUP FIGURE 2.13 PLASMA FREE L-THYROXINE LEVELS AT BASELINE, BEFORE AND AFTER INH IN ANIMALS RECEIVING L-THYROXINE. Plasma free T4 levels in the T4-INH group are shown. After five days of pretreatment with L-thyroxine (Pre-INH) (58.4 ±8.1 pmoles/L, n = 13) and after nine days of pretreatment with L-thyroxine (Post-INH; day 10) (70.6 ± 4.2 pmoles/L, n = 9) plasma free T4 levels were significantly elevated as compared to baseline levels (13.6 ± 0.8 pmoles/L, n = 13) (p < 0.0001, ANOVA; p < 0.05, Newman-Keul's). (* significantly different from baseline). 125 20 ANOVA p = 0.03, Newman Keul's p < 0.05 Post-INH < Pre-INH 15 h 10 Baseline,n=22 Pre-INH,n=22 Post-INH,n=19 INH-ONLY GROUP FIGURE 2.14 PLASMA FREE L-THYROXINE LEVELS AT BASELINE, BEFORE AND AFTER INH IN ANIMALS NOT RECEIVING L-THYROXINE. Plasma free T4 levels in the INH-only group are shown. Plasma free T4 levels were not elevated on the first day of INH injections (Pre-INH) (13.9 ± 0.9 pmoles/L, n = 22) nor were they elevated at the end of the study (Post-INH) (11.1 ± 0.7 pmoles/L, n = 19) as compared to baseline levels (12.2 ± 0.4 pmoles/L, n = 22). However, a small but significant decrease did occur between pre-INH and post-INH (p = 0.03, ANOVA; p < 0.05, Newman-Keul's). (* significantly different from Pre-INH time). 126 Controls, n = 9 INH-only,n = 21 T4-INH, n = 13 FIGURE 2.15 PEAK PLASMA ARGINTNOSUCCINIC ACID LYASE ACTIVITY IN THE TREATMENT GROUPS. Peak plasma ASAL activities in the INH-only (56.0 ± 19.7 Takahara units, n = 21) and T4-INH groups (31.3 ± 20.1 Takahara units, n = 13) were significantly increased versus the vehicle control group (4.3 ± 0.5 Takahara units, n = 9) (p < 0.0001, ANOVA; p < 0.05 Newman-Keul's, statistics on logarithmically-transformed data). Peak plasma ASAL activity in the INH-only group was also significantly greater than in the T4-INH group. (* significantly different from controls). 127 FIGURE 2.16 P E A K P L A S M A A L A N I N E AMINOTRANSFERASE ACTIVITY IN T H E T R E A T M E N T GROUPS. No significant difference in peak plasma ALT activity was observed between the INH-only (68.4 ± 18.5 units/L, n = 21), T4-INH (24.8 ± 5.4 units/L, n = 13) or vehicle control groups (27.0 ±1.9 units/L, n = 9) (p = 0.03, ANOVA; p > 0.05 Newman-Keul's, statistics done on logarithrnically-transformed data). 128 FIGURE 2.17 HEPATIC TRIGLYCERIDE ACCUMULATION IN THE TREATMENT GROUPS. Hepatic triglyceride accumulation was significantly increased in the INH-only group (27.5 ± 3.2 mg triglyceride/ g liver, n = 19) versus the T4-INH (7.4 ± 2.4 mg triglyceride/ g liver, n = 9) and vehicle control groups (9.3 ± 1.8 mg triglyceride/ g liver, n = 9) (p < 0.0001, ANOVA; p < 0.05 Newman-Keul's). (* significantly different from controls). 129 FIGURE 2.18 HEPATIC C Y T O C H R O M E P-450 L E V E L S IN T H E T R E A T M E N T GROUPS. Hepatic cytochrome P-450 levels in the T4-INH (1.7 ± 0.1 nmoles/mg protein, n = 9) and INH-only groups (1.8 ± 0 . 1 nmoles/mg protein, n = 19) were significantly decreased versus the vehicle control group (3.5 ± 0.3 nmoles/mg protein, n = 9) (p < 0.0001, ANOVA; p < 0.05 Newman-Keul's). (* significantly different from controls). 130 FIGURE 2.19 HEPATIC REDUCTASE ACTIVITY IN THE TREATMENT GROUPS. Hepatic reductase activity was increased in the T4-INH group (648 ± 44 nmoles/minute/mg protein, n = 9) versus the INH-only (479 ± 23 nmoles/minute/mg protein, n = 19) and vehicle control groups (421 ± 25 nmoles/minute/mg protein, n = 9) (p = 0.0001, ANOVA; p < 0.05 Newman-Keul's). (* significantly different from controls). 131 FIGURE 2.20 HEPATIC ^ -NITROPHENOL HYDROXYLASE ACTIVITY (CYP2E1) IN THE TREATMENT GROUPS. Hepatic CYP2E1 activity in the INH-only group (1.8 ± 0.2 nmoles/minute/mg protein, n = 19) was significantly decreased versus the vehicle control group (2.9 ± 0.4 nmoles/minute/mg protein, n = 9) (p = 0.0009, ANOVA; p < 0.05 Newman-Keul's). Hepatic CYP2E1 activity in the T4-INH group (3.8 ± 0.6 nmoles/minute/mg protein, n = 9) was not significantly different from the control group. (* significantly different from controls). 132 FIGURE 2.21 HEPATIC CYTOCHROME P-450 LEVELS IN THE PRETREATMENT GROUPS OF THE FOLLOW-UP STUDY. Hepatic cytochrome P-450 levels in the L-thyroxine pretreatment group (2.1 ± 0 . 1 nmoles/mg protein, n = 5), but not in the methimazole pretreatment group (3.0 ± 0.4 nmoles/mg protein, n = 5), were significantly decreased compared to the control group (3.5 ± 0.3 nmoles/mg protein, n = 9) (p = 0.015, ANOVA; p < 0.05 Newman-Keul's). (* significantly different from controls). 133 F I G U R E . 2.22 H E P A T I C R E D U C T A S E . A C T I V I T Y I N T H E P R E T R E A T M E N T G R O U P S O F T H E F O L L O W - U P S T U D Y . Hepatic reductase activity was significantly decreased in the methimazole pretreatment group (170 ± 27 nmoles/minute/mg protein, n = 5) versus both the vehicle control group (421 ± 25 nmoles/minute/mg protein, n = 9) and the L-thyroxine pretreatment group (466 ± 27 nmoles/minute/mg protein, n = 5) (p < 0.0001, ANOVA; p < 0.05 Newman-Keul's). (* significantly different from controls). FIGURE 2.23 HEPATIC /J-NITROPHENOL H Y D R O X Y L A S E A C T I V I T Y (CYP2E1) IN THE PRETREATMENT GROUPS OF THE FOLLOW-UP STUDY. Hepatic CYP2E1 activities in the L-thyroxine (2.1 ± 0.2 nmoles/minute/mg protein, n = 5) and methimazole (2.7 ± 0.3 nmoles/minute/mg protein, n = 5) pretreatment groups were not significantly different from the control group (2.9 ± 0.4 nmoles/minute/mg protein, n = 9) (p = 0.38, ANOVA). 135 T A B L E 2.4 C O M P A R I S O N B E T W E E N T W O P R E T R E A T M E N T V E H I C L E - C O N T R O L G R O U P S L-Thyroxine Vehicle Methimazole Vehicle Significance* and INH and INH Peak Plasma ASAL 47.3 ± 19 70.2 ± 43 p = 0.92* Activity; n= 13 n = 8 (Takahara Units) Peak Plasma ALT Activity; 61.0 ± 13 86.7 ± 45 p = 0.85f (Units/L) n= 13 n = 8 Hepatic Triglycerides; 24.0 ± 4 . 1 32.3 ± 4 . 9 p = 0.21 (mg TG/g liver) n = 8 n = 8 Cytochrome P-450; 1.7 ± 0 . 1 1.8 ± 0 . 1 p = 0.51 (nmoles/mg protein) n = 8 n = 8 Reductase Activity; 507 ± 34 440 ± 24 p = 0.15 (nmoles/min/mg protein) n = 8 n = 8 p-Nitrophenol Hydroxylase 2.7 ± 0 . 4 2.2 ± 0 . 4 p = 0.44 (CYP2E1) Activity n=8 n = 8 (nmoles/min/mg protein) * Statistics were done using Mests assuming either equal or unequal variances (as determined by ¥• tests). t Statistics done using logarithmically-transformed data. The data are presented (mean ± standard error) non-transformed for the purpose of clarity. 136 TABLE 2.5 TOXICOLOGICAL MARKERS AND MICROSOMAL ENZYME ACTIVITIES BY GROUP Controls T4-INH INH-Only ANOVA Peak Plasma ASAL Activity; (Takahara Units) 4.3 ±0 .5 n = 9 31.3 ±20* n= 13 56.0 ± 20* n = 21 p< 0.0001* Peak Plasma ALT Activity; (Units/L) 27.0 ± 1.9 n = 9 24.8 ± 5 . 4 n= 13 68.4 ± 19 n = 21 p = 0.03* Hepatic Triglycerides; (mg TG/g liver) 9.3 ± 1 . 8 n = 9 7.4 ± 2.4 n = 9 27.5 ± 3.2* n=19 p < 0.0001 Cytochrome P-450; (nmoles/mg protein) 3.5 ±0 .3 n = 9 1.7 ±0.1* n = 9 1.8 ±0 .1* n=19 p < 0.0001 Reductase Activity; (nmoles/min/mg protein) 421 ± 25 n=9 648 ± 44* n = 9 479 ± 23 n= 19 p = 0.0001 p-Nitrophenol Hydroxylase (CYP2E1) Activity; (nmoles/min/mg protein) 2.9 ± 0 . 4 n = 9 3.8 ±0 .6 n = 9 1.8 ±0.2* n= 19 p = 0.0009 p < 0.05; Newman-Keul's, significantly different from controls. * Statistics done using logarithmically-transformed data. The data are presented (mean ± standard error) non-transformed for the purpose of clarity. 137 TABLE 2.6 MICROSOMAL ENZYME ACTIVITIES IN THE FOLLOW-UP STUDY Pretreatment Controls n = 9 L-Thyroxine n = 5 Methimazole n = 5 ANOVA Cytochrome P-450; (nmoles/mg protein) 3.5 ±0 .3 2.1 ± 0 . ^ 3.0 ± 0 . 4 p< 0.015 Reductase Activity; (nmoles/min/mg protein) 421 ± 2 5 466 ± 27 170 ± 27T p < 0.0001 /7-Nitrophenol Hydroxylase (CYP2E1) Activity; (nmoles/min/mg protein) 2.9 ± 0 . 4 2.1 ± 0 . 2 2.7 ±0 .3 p = 0.38 *p < 0.05; Newman-Keul's, significantly different from controls. 138 S T U D Y 4 S T U D Y 4 H Y P O T H E S I S • Inhibition o f the activity of INH-amidase using the amidase inhibitor B N P P decreases the conversion of hydrazine from I N H and decreases the severity of INH-induced hepatic necrosis. In this study, four out of seventy animals died prematurely. Three were from the " V E H - I N H group (one died between 12 and 24 hours and two between 24 and 32 hours) and one was from the B N P P - I N H group (died between 12 and 24 hours). Plasma samples from all four of these animals prior to death showed highly elevated plasma liver enzymes indicating hepatic necrosis and were included in the toxicity analyses. The liver o f one of the V E H - I N H animals which died between 12 and 24 hours was recovered at the time o f death and immediately frozen in liquid nitrogen and included in the microsomal preparation and analyses. Peak plasma A S A L activity was significantly increased in the V E H - I N H group (239 ± 83 Takahara units, n = 17) but not in the B N P P - I N H (57.6 ± 46.5 Takahara units, n = 17), B N P P - V E H (3.8 ± 0.4 Takahara units, n = 12) or V E H - V E H * (4.6 ± 0.8 Takahara units, n = 12) groups as compared to the V E H - V E H group (3.9 ± 0.4 Takahara units, n = 12) ( A N O V A p < 0.00001, Newman-Keul's p < 0.05) (Figure 2.24; page 147; Table 2.7). Peak plasma A L T activity was significantly increased in the V E H - I N H group (240 ± 81 Units/L, n = 17) but not in the B N P P - I N H (62.8 ± 24.4 Units/L, n = 17), B N P P -V E H (25.2 ± 2.8 Units/L, n = 12) or V E H - V E H * groups (31.3 ± 3.9 Units/L, n = 12) as 139 compared to the VEH-VEH group (24.9 ± 2.3 Units/L, n = 12) (ANOVA p = 0.003, Newman-Keul's p < 0.05) (Figure 2.25; page 148; Table 2.7). Hepatic triglyceride accumulation was significantly increased in the VEH-INH group (29.4 + 3.6 mg TG/g liver, n = 15) but not in the BNPP-INH group (11.9 ± 2.3 mg TG/g liver, n = 16), BNPP-VEH (9.6 ± 2.2 mg TG/g liver, n = 12) or VEH-VEH* groups (10.5 ± 1.6 mg TG/g liver, n = 12) as compared to the VEH-VEH group (7.6 + 1.7 mg TG/g liver, n = 12) (ANOVA p < 0.00001, Newman-Keul's p < 0.05 (Figure 2.26; page 149; Table 2.7). Plasma triglyceride levels at 32 hours was significantly increased in the VEH-INH group (6.5 ± 1.6 mM, n = 14) but not in the BNPP-INH group (2.5 + 1.0 mM, n = 16), BNPP-VEH (1.3 ± 0.2 mM, n = 12) or VEH-VEH* groups (1.0 ± 0.1 mM, n = 12) as compared to the VEH-VEH group (1.7 ± 0.7 mM, n = 12) (ANOVA p = 0.001, Newman-Keul's p < 0.05) (Figure 2.27; page 150; Table 2.7). Hepatic amidase activity was determined by measurement of the amount of hydrazine produced after incubation of INH with microsomes. Hydrazine production in the BNPP-INH (4.0 ± 0.6 nmoles/mg protein/hour, n = 6) and BNPP-VEH groups (4.7 ± 0.7 nmoles/mg protein/hour, n = 6) showed a decrease of approximately 90% versus the VEH-VEH control group (38.8 ± 4.5 nmoles/mg protein/hour, n = 12) (ANOVA p < 0.00001, Newman-Keul's p < 0.05) (Figure 2.28; page 151; Table 2.8). Hydrazine production in 10 of the animals in the BNPP-INH group and 6 in the BNPP-VEH group were below quantitation limits. Values below the quantitation limit were not included in the statistical analysis. A significant decrease in hydrazine production (approximately 140 38%) was also observed in the VEH-INH group (24.1 ± 2.0 nmoles/mg protein/hour, n = 15) but not in the VEH-VEH* group (34.7 ±3.6 nmoles/mg protein/hour, n = 12). Since the most important observation is the relative differences between the groups, the percent differences are shown as well (Figure 2.29; page 152). The correlations of hepatic amidase activity with hepatic necrosis (peak plasma ASAL, r = 0.28, r2 = 0.08, n = 31 p > 0.05; peak plasma ALT activity, r = 0.28, r2 = 0.08, n = 31, p > 0.05), hepatic steatosis (r = 0.49, r2 = 0.24, n = 31, p < 0.01) and hypertriglyceridemia (r = 0.18, r2 = 0.03, n = 30, p > 0.10) are weak. In vitro incubation of BNPP with INH in control microsomes (from VEH-VEH control group rabbits) showed the BNPP IC5o to be approximately 2.0 pM (Figures 2.30 and 2.31; pages 153 & 154; Table 2.9). The results were derived from three separate incubations of microsomes with INH and BNPP. Hydrazine production from microsomes incubated with various concentrations of INH (0.3, 1, 3, 6, 10 and 15 mM) in the presence of 0, 1.5 or 3 pM BNPP was also determined. A Lineweaver-Burk plot of the results showed that at 0 pM BNPP, Vmax was 226 nmoles/min/mg protein and K m was 9.3 mM, at 1.5 uM BNPP, V ^ was 167 nmoles/min/mg protein and K m was 8.3 mM and at 3 pM BNPP, Vmax was 99.4 nmoles/min/mg protein and K m was 12.1 mM (Figure 2.32; page 155; Table 2.10). An Eadie-Hofstee plot of the results showed that at 0 pM BNPP, V ^ was 196 nmoles/min/mg protein and K m was 7.8 mM, at 1.5 pM BNPP, Vmax was 116 nmoles/min/mg protein and K m was 5.3 mM and at 3 pM BNPP, Vmax was 85.5 nmoles/min/mg protein and K m was 10.0 mM (Figure 2.33; page 156; Table 2.10). These 141 data suggest that increasing concentrations of BNPP decreases the V,,^ of INH-amidase but does not significantly change its Km. Plasma amidase activity was determined by measurement of the amount of hydrazine produced after incubation of INH with baseline plasma. Therefore, the effect of BNPP and INH on plasma INH-amidase activity was not evaluated. Plasma hydrazine production was near the limit of quantitation for all samples in all groups. There were 5 animals with plasma amidase activity below the limit of quantitation (1 in the VEH-VEH group, 1 in the VEH-INH group, 1 in the BNPP-INH group and 2 in the VEH-VEH* group). The plasma amidase activities which were quantitated ranged from 0.20 to 0.62 nmoles/hour/mg protein. The plasma hydrazine concentration at 12 hours was 1.9 times higher in the VEH-INH group (25.3 ± 1.4 pM, n - 17) as compared to the BNPP-INH group (13.5 + 1.0 pM, n = 6) (p < 0.00001) (Figure 2.34; page 157; Table 2.11). Plasma hydrazine levels in 11 of the animals in the BNPP-INH group were below the quantitation limit of 10 uM at 12 hours. Values below the quantitation limit were not included in the statistical analysis. The plasma hydrazine concentration at 24 hours was 2 times higher in the VEH-INH group (12.1 ± 4.6 uM, n = 15) as compared to the BNPP-INH group (6.1 ± 0.9 uM, n = 4); however, this difference was not significant (p = 0.22) (Figure 2.34; page 157; Table 2.11). Plasma hydrazine levels in one of the animals in the VEH-INH group and 12 of the animals in the BNPP-INH group were below the quantitation limit of 4 pM at 24 hours. Values below the quantitation limit were not included in the statistical analysis. Of note in the plasma hydrazine concentrations in the VEH-INH group at 24 hours is the high 142 standard error term. This is due to two animals with very high plasma hydrazine levels (25 pM and 74 pM). Both of these animals died within 2 hours of the 24 hour blood sample. At 32 hours, the plasma hydrazine concentration was 3.9 times higher in the VEH-INH group (31.8 ± 2.8 pM, n = 14) as compared to the BNPP-INH group (8.2 + 1.8 pM, n = 13) (p < 0.00001) (Figure 2.34; page 157; Table 2.11). Plasma hydrazine levels in 3 of the animals in the BNPP-INH group were below the quantitation limit of 4 pM at 32 hours. Values below the quantitation limit were not included in the statistical analysis. Plasma hydrazine levels at 12 hours (r = 0.83, r2 = 0.69, n = 31, p < 0.001) and at 32 hours (r = 0.71, r2 = 0.50, n = 27, p < 0.001) correlate with hepatic amidase activity. However, correlations of plasma hydrazine levels at 12 hours (r = 0.43, r2 = 0.18, n = 34, p < 0.02) and 32 hours (r = 0.46, r2 = 0.21, n = 27, p < 0.02) as well as plasma amidase activity (r = 0.10, r2 = 0.01, n = 29, p > 0.05) correlate less strongly with peak plasma ASAL activity. Hepatic glutathione (GSH) content was significantly decreased in the VEH-INH group (4.3 ± 0.3 pmoles GSH/g liver, n = 15) versus the other four groups. Hepatic glutathione content in the VEH-VEH* (6.0 ± 0.3 pmoles GSH/g liver, n = 12) and BNPP-INH groups (6.3 ± 0.3 pmoles GSH/g liver, n = 16) were significantly decreased compared to the VEH-VEH (7.6 + 0.4 pmoles GSH/g liver, n = 12) and BNPP-VEH (7.5 ± 0.4 pmoles GSH/g liver, n = 12) groups (ANOVA p < 0.00001, Newman-Keul's p < 0.05 (Figure 2.35; page 158; Table 2.7). There were no significant correlations between hepatic glutathione and log peak plasma ASAL activity (r = -0.44, r2 = 0.19, n = 31, p > 0.05), log peak plasma ALT activity (r = -0.29, r2 = 0.08, n = 31, p > 0.05) or hepatic 143 triglyceride content (r = -0.33, r2 = 0.11, n = 31, p > 0.05) in animals receiving INH (BNPP-INH and VEH-INH groups). Formation of TBARS, an indirect measure of lipid peroxidation, after challenge of liver tissue with 0.25 mM /-butyl hydroperoxide was significantly increased in the VEH-INH group (0.37 ± 0.03 absorbance units (at 532 nm), n = 15) and VEH-VEH* (0.32 ± 0.01 absorbance units, n = 12) groups compared to the VEH-VEH (0.24 ± 0.03 absorbance units, n = 12) and BNPP-VEH groups (0.24 ± 0.03 absorbance units, n = 12). (ANOVA p < 0.00001, Newman-Keul's p < 0.05 (Figure 2.36; page 159; Table 2.7). The BNPP-INH group (0.30 ± 0.02 absorbance units, n = 16) was not significantly different from any of the other groups. There were no significant correlations between TBARS formation and log peak plasma ASAL activity (r = 0.12, r2 = 0.01, n = 31, p > 0.05), log peak plasma ALT activity (r = 0.02, r2 = 0.01, n = 31, p > 0.05) or hepatic triglyceride content (r = 0.30, r2 = 0.09, n = 31, p > 0.05) in animals receiving INH (BNPP-INH and VEH-INH groups). Hepatic CYP-450 levels in the VEH-INH group (1.1 ± 0.1 nmoles P-450/mg protein, n = 15) were significantly decreased versus the VEH-VEH (2.1 ±0.1 nmoles/mg protein, n = 12), BNPP-INH (1.8 ± 0.1 nmoles/mg protein, n = 16), BNPP-VEH (2.3 ± 0.2 nmoles/mg protein, n = 12) and VEH-VEH* (1.8 ± 0.1 nmoles/mg protein, n = 12), groups (ANOVA p < 0.00001, Newman-Keul's p < 0.05) (Figure 2.37; page 160; Table 2.12). Hepatic reductase activity was not significantly different between the VEH-VEH (361 ± 15 nmoles/min/mg protein, n = 12), VEH-INH (320 ± 17 nmoles/min/mg protein, n 144 = 15), BNPP-INH (343 ± 14 nmoles/min/mg protein, n = 16), BNPP-VEH (360 ± 10 nmoles/min/mg protein, n = 12), or VEH-VEH* groups (362 ± 19 nmoles/min/mg protein, n = 12) (ANOVA p = 0.22) (Figure 2.38; page 161; Table 2.12). Hepatic ethoxyresorufin-O-deethylase (EROD) activity was significantly lower in the VEH-INH (180 ± 20 pmoles/min/mg protein, n = 15), BNPP-INH (354 ± 36 pmoles/min/mg protein, n = 16) and VEH-VEH* groups (490 ± 3 5 pmoles/min/mg protein, n = 12) as compared to the BNPP-VEH (804 ± 68 pmoles/min/mg protein, n = 12) and VEH-VEH groups (804 ± 42 pmoles/min/mg protein, n = 12) (ANOVA p < 0.00001, Newman-Keul's p < 0.05) (Figure 2.39; page 162; Table 2.12). Hepatic benzoyloxyresorufin-O-dealkylase (BROD) activity was not significantly different between the VEH-VEH (1254 ± 142 pmoles/min/mg protein, n = 12), VEH-INH (988 ± 131 pmoles/min/mg protein, n = 15), BNPP-INH (1103 ± 136 pmoles/min/mg protein, n = 16), BNPP-VEH (1209 ± 138 pmoles/min/mg protein, n = 12) or VEH-VEH* groups (780 ± 113 pmoles/min/mg protein, n = 12) (ANOVA p = 0.13) (Figure 2.40; page 163; Table 2.12). Hepatic pentoxyresorufin-0-dealkylase (PROD) activity was not significantly different between the VEH-VEH (105 ± 14 pmoles/min/mg protein, n = 10), VEH-INH (121 ± 15 pmoles/min/mg protein, n = 11), BNPP-INH (112 ± 12 pmoles/min/mg protein, n = 14), BNPP-VEH (112 ± 8 pmoles/min/mg protein, n = 10) or VEH-VEH* groups (82 ± 6 pmoles/min/mg protein, n = 11) (ANOVA p = 0.18) (Figure 2.41; page 164; Table 2.12). PROD activity was below quantitation limits in 2 animals in the VEH-VEH group, 4 animals in the VEH-INH group, 2 animals in the BNPP-INH group, 2 animals in the 145 BNPP-VEH group and 1 animal in the VEH-VEH* group. Values below the quantitation limit were not included in the statistical analysis. Hepatic /?-nitrophenol hydroxylase (CYP2E1) activity was significantly decreased in the VEH-INH group (0.53 ±0.2 nmoles/min/mg protein, n = 7) versus the VEH-VEH (1.8 ± 0.1 nmoles/min/mg protein, n = 12), BNPP-VEH (1.9 ± 0.2 nmoles/min/mg protein, n = 12) and VEH-VEH* (1.4 ±0.1 nmoles/min/mg protein, n = 12) groups. In addition, hepatic /?-nitrophenol hydroxylase activity in the BNPP-INH group (3.2 ± 0.4 nmoles/min/mg protein, n = 15) was significantly greater than all of the other groups (10-fold higher than in the VEH-INH group) (ANOVA p < 0.0001, Newman-Keul's p < 0.05) (Figure 2.42; page 165; Table 2.12). Microsomal /?-nitrophenol hydroxylase activity was undetectable in 8 rabbits in the VEH-INH group and in one rabbit in the BNPP-INH group. The animals with undetectable /?-nitrophenol hydroxylase activity were not included in the statistical analysis. Correlation of EROD activity in the VEH-INH group with hepatic necrosis (peak plasma ASAL, r = -0.66, r2 = 0.44, n = 15 p < 0.01; peak plasma ALT activity, r = -0.70, r2 = 0.49, n = 15, p < 0.005) was negative and significant. In addition, p-nitrophenol hydroxylase activity correlated significantly (negative) with hepatic steatosis (liver triglycerides; r = -0.71, r2 = 0.50, n = 22, p < 0.001). None of the other P-450 enzyme activities correlated with measures of toxicity. Thirty-three out of 34 rabbits were rapid acetylators (range of 58-98% sulfamethazine acetylated in 20 minutes) and only one was a slow acetylator (24% of the sulfamethazine acetylated in 20 minutes). The single slow acetylator rabbit was in the 146 BNPP-INH group and died prior to the end of the study at 48 hours. Correlational analysis of the percent sulfamethazine acetylated in 20 minutes (not including the slow acetylator) revealed non-significant correlations with peak plasma ASAL activity (r = -0.20, r2 = 0.04, n = 33, p > 0.05), peak plasma ALT activity (r = -0.07, r2 = 0.01, n = 33, p > 0.05), liver triglyceride accumulation (r = -0.06, r2 = 0.01, n = 31, p > 0.05) and peak plasma triglyceride levels (r = -0.11, r2 = 0.01, n = 30, p > 0.05). In the animals receiving INH only, correlation of percent sulfamethazine acetylated in 20 minutes (not including the slow acetylator) with peak plasma ASAL activity (r = -0.14, r2 = 0.02, n = 17, p > 0.05), peak plasma ALT activity (r = -0.21, r2 - 0.04, n = 17, p > 0.05), liver triglyceride accumulation (r = -0.20, r2 = 0.04, n = 14, p > 0.05) and peak plasma triglyceride levels (r = -0.25, r2 = 0.06, n = 14, p > 0.05) were also not significant. Including all of the rabbits from Studies 1, 3 and 4 which only received INH, the correlation of acetylator status with hepatic necrosis is not significant (r = -0.23, r2 = 0.06, n = 50, p > 0.05) (Figure 2.43; page 166). 147 < _ j < CO < CO E CO _c0 CL CO 0 CL O) O VEH/VEH VEH/INH BNPP/INH BNPP/VEH VEH/VEH* Treatment Groups FIGURE 2.24 COMPARISON OF PEAK PLASMA ARGININOSUCCINIC ACID LYASE ACTIVITY BETWEEN TREATMENT GROUPS. A comparison of log peak plasma ASAL activity (log Takahara units) between treatment groups (groups defined on page 30). Peak plasma ASAL activity was significantly increased in the VEH-INH group (239 ± 83 Takahara units, n = 17) but not in the BNPP-INH group (57.6 ± 46.5 Takahara units, n = 17), BNPP-VEH (3.8 ± 0.4 Takahara units, n = 12) or VEH-VEH* groups (4.6 ± 0.8 Takahara units, n = 12) as compared to the VEH-VEH group (3.9 ± 0.4 Takahara units, n = 12) (ANOVA p < 0.00001, Newman-Keul's p < 0.05). 148 2.50 ••?= 2.25 h 2.00 1.75 h 1.50 h 1.25 h 1.00 VEH/VEH VEH/INH BNPP/INH BNPP/VEH VEH/VEH* Treatment Groups FIGURE 2.25 COMPARISON OF PEAK PLASMA ALANINE AMINOTRANSFERASE ACTIVITY BETWEEN TREATMENT GROUPS. A comparison of log peak plasma ALT activity (log Units/L) between treatment groups (groups defined on page 30). Peak plasma ALT activity was significantly increased in the VEH-INH group (240 ± 81 Units/L, n = 17) but not in the BNPP-INH group (62.8 ± 24.4 Units/L, n = 17), BNPP-VEH (25.2 ± 2.8 Units/L, n = 12) or VEH-VEH* groups (31.3 ± 3.9 Units/L, n = 12) as compared to the VEH-VEH group (24.9 ± 2.3 Units/L, n = 12) (ANOVA p = 0.003, Newman-Keul's p < 0.05). 149 FIGURE 2.26 COMPARISON OF HEPATIC TRIGLYCERIDE ACCUMULATION BETWEEN TREATMENT GROUPS. A comparison of hepatic triglyceride (TG) accumulation (mg TG/g liver) between treatment groups (groups defined on page 30). Hepatic triglyceride accumulation was significantly increased in the VEH-INH group (29.4 ± 3.6 mg TG/g liver, n = 15) but not in the BNPP-INH group (11.9 ± 2.3 mg TG/g liver, n = 16), BNPP-VEH (9.6 ± 2.2 mg TG/g liver, n = 12) or VEH-VEH* groups (10.5 ± 1.6 mg TG/g liver, n = 12) as compared to the VEH-VEH group (7.6 ± 1.7 mg TG/g liver, n = 12) (ANOVA p < 0.00001, Newman-Keul's p < 0.05. 150 VEH/VEH VEH/INH BNPP/INH BNPP/VEH VEH/VEH* Treatment Groups FIGURE 2.27 COMPARISON OF 32 HOUR PLASMA TRIGLYCERIDE LEVELS BETWEEN TREATMENT GROUPS. A comparison of 32 hour plasma triglyceride (TG) levels (mM) between treatment groups (groups defined on page 30). Plasma triglyceride levels at 32 hours were significantly increased in the VEH-INH group (6.5 ± 1.6 mM, n = 14) but not in the BNPP-INH group (2.5 ± 1.0 mM, n = 16), BNPP-VEH (1.3 ± 0.2 mM, n = 12) or VEH-VEH* groups (1.0 ± 0.1 mM, n = 12) as compared to the VEH-VEH (1.7 ± 0.7 mM, n = 12) (ANOVA p = 0.001, Newman-Keul's p < 0.05). 151 O Q- 50 _>v VEH/VEH VEH/INH BNPP/INH BNPP/VEH VEH/VEH* Treatment Groups FIGURE 2.28 COMPARISON OF MICROSOMAL HYDRAZINE PRODUCTION BETWEEN TREATMENT GROUPS. A comparison of microsomal hydrazine production (nmoles/hour/mg protein) (INH-amidase activity) between treatment groups (groups defined on page 30). Hydrazine production in the BNPP-INH (4.0 ± 0.6 nmoles/mg protein/hour, n = 6) and BNPP-VEH groups (4.7 ± 0.7 nmoles/mg protein/hour, n = 6) showed a decrease of approximately 90% versus the VEH-VEH control group (38.8 ± 4.5 nmoles/mg protein/hour, n = 12) (ANOVA p < 0.00001, Newman-Keul's p < 0.05). A significant decrease in hydrazine production (approximately 38%) was also observed in the VEH-INH group (24.1 ± 2.0 nmoles/mg protein/hour, n = 15) but no decrease occurred in the VEH-VEH* group (34.7 ±3 .6 nmoles/mg protein/hour, n = 12). 152 * ANOVA p < 0.00001; Newman-Keul's p < 0.05 VEH/VEH VEH/INH BNPP/INH BNPP/VEH VEH/VEH* Treatment Groups FIGURE 2.29 COMPARISON OF PERCENT CHANGE IN MICROSOMAL HYDRAZINE PRODUCTION BETWEEN TREATMENT GROUPS. A comparison of microsomal hydrazine production (% INH-amidase activity) between treatment groups (groups defined on page 30). Hydrazine production in the BNPP-INH (10.4 ± 1.6 %, n = 6) and BNPP-VEH groups (12.2 ± 1.9 %, n = 6) showed a decrease of approximately 90% versus the VEH-VEH control group (100 ± 11.5 %, n = 12) (ANOVA p < 0.00001, Newman-Keul's p < 0.05). A significant decrease in hydrazine production (approximately 38%) was also observed in the VEH-INH group (62.2 ± 5.3 %, n = 15) but no significant decrease occurred in the VEH-VEH* group (89.5 ± 9.4 %, n = 12). BNPP IC50 C U R V E S 153 + IC50 2.2uM A IC50 1.5uM O IC50 2.7 uM 95% con f idence intervals 1.8-2.6 0.2 1 10 30 B N P P Concentration (uM) FIGURE 2.30 CONCENTRATION-RESPONSE CURVE FOR BIS-p-NITROPHENYL PHOSPHATE IN MICROSOMAL INCUBATION WITH INH. INH (3 mM) was incubated in the presence of BNPP (0.25, 0.5, 1, 2, 4, 8 and 16 LLM) with control microsomes at 37°C for 30 minutes. By measuring the production of hydrazine in the presence of increasing concentrations of BNPP, the IC 5 0 of BNPP was determined (approximately 2 uM). % BNPP IC50 C U R V E S 154 B N P P Concentration (uM) FIGURE 2.31 PERCENT CONCENTRATION-RESPONSE CURVE FOR BIS-p-NITROPHENYL PHOSPHATE IN MICROSOMAL INCUBATION WITH INH. INH (3 mM) was incubated in the presence of BNPP (0.25, 0.5, 1, 2, 4, 8 and 16 uM) with control microsomes at 37°C for 30 minutes. By measuring the production of hydrazine in the presence of increasing concentrations of BNPP, the IC 5 0 of BNPP was determined (approximately 2 uM). LINEWEAVER-BURK PLOT + OuMBNPP A 1.5 uM BNPP O 3.0 uM BNPP -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 1/ INH Concentration (LVmmoIe) FIGURE 2.32 LINEWEAVER-BURK PLOT FOR INH-AMIDASE INHIBITION. INH was incubated at varying concentrations in control microsomes at 37°C for 30 minutes in the presence of no BNPP (0 uM BNPP), 1.5 uM BNPP or 3 uM BNPP. The data were plotted using a Lineweaver-Burk Plot in order to estimate the and K m and to predict the mechanism of inhibition by BNPP. At 0 uM BNPP, 1.5 uM BNPP and 3 uM BNPP, V , ^ was 226,167 and 99.4 nmoles hydrazine/hour/mg protein, respectively. At 0 yM BNPP, 1.5 nM BNPP and 3 BNPP, K m was 9.3, 8.3 and 12.1 mM INH. EADIE-HOFSTEE PLOT 156 + OuMBNPP A 1.5 uM BNPP O 3 uM BNPP O 0 5 10 15 20 25 30 rr Rate/I NH Concentration (nmoles hydrazine/hour/mg protein)/(mmole/L) FIGURE 2.33 EADIE-HOFSTEE PLOT FOR INH-AMIDASE INHIBITION. INH was incubated at varying concentrations in control microsomes at 37°C for 30 minutes in the presence of no BNPP (0 uM BNPP), 1.5 LIM BNPP or 3 uM BNPP. The data were plotted using an Eadie-Hofstee Plot in order to estimate the V m a x and K m and to predict the mechanism of inhibition by BNPP. At 0 LIM BNPP, 1.5 LIM BNPP and 3 uMBNPP, was 196, 116 and 85.5 nmoles hydrazine/hour/mg protein, respectively. At 0 uM BNPP, 1.5 uM BNPP and 3 uM BNPP, K m was 7.8, 5.3 and 10.0 mM INH. 157 FIGURE 2.34 COMPARISON OF PLASMA HYDRAZINE CONCENTRATIONS AT 12, 24 AND 32 HOURS. Plasma hydrazine concentrations were compared between the BNPP-INH and VEH-INH groups at 12, 24 and 32 hours after initiation of the INH-injection protocol (groups defined on page 30). The plasma hydrazine concentration at 12 hours was 1.9 times higher in the VEH-INH group (25.3 ± 1.4 uM, n = 17) as compared to the BNPP-INH group (13.5 ± 1.0 uM, n = 6) (p < 0.00001). The plasma hydrazine concentration at 24 hours was 2 times higher in the VEH-INH group (12.1 ± 4.6 uM, n = 15) as compared to the BNPP-INH group (6.1 ± 0.9 nM, n = 4); however, this difference was not significant (p = 0.22). At 32 hours, the plasma hydrazine concentration was 3.9 times higher in the VEH-INH group (31.8 ± 2.8 uM, n = 14) as compared to the BNPP-INH group (8.2 ± 1.8 ^M, n = 13) (p < 0.00001). 158 10 9 8 7 6 5 4 3 2 1 0 * ANOVA p < 0.0001, Newman-Keuls p < 0.05 VEH/INH < BNPP/INH - V E H / V E H * < B N P P / V E H = VEH/VEH VEH/VEH VEH/INH BNPP/INH BNPP/VEH VEH/VEH* Treatment Groups FIGURE 2.35 COMPARISON OF HEPATIC GLUTATHIONE LEVELS BETWEEN TREATMENT GROUPS. A comparison of hepatic glutathione (GSH) levels (nmoles GSH/g liver) between treatment groups (groups defined on page 30). The GSH levels (4.3 ± 0.3 umoles GSH/g liver, n = 15) in the VEH-INH group are significantly lower (a 43% decrease from controls) than all of the other groups. The GSH levels in the VEH-VEH* (6.0 ± 0.3 umoles GSH/g liver, n = 12; a 20% decrease from controls) and the BNPP-INH groups (6.3 ± 0.3 umoles GSH/g liver, n = 16; decreased by 18%) groups are significantly lower than the BNPP-VEH (7.5 ± 0.4 umoles GSH/g liver, n = 12) and VEH-VEH (7.6 ± 0.4 umoles GSH/g liver, n = 12) control groups (ANOVA p < 0.0001, Newman-Keul's p < 0.05). 159 FIGURE 2.36 COMPARISON OF HEPATIC THIOBARBITURIC ACID REACTIVE SUBSTANCES FORMATION BETWEEN TREATMENT GROUPS. Formation of TBARS, an indirect measure of lipid peroxidation, after challenge of liver tissue with 0.25 mM /-butyl hydroperoxide was significantly increased in the VEH-INH group (0.37 ± 0.03 absorbance units (at 532 nm), n = 15) and VEH-VEH* (0.32 ± 0.01 absorbance units, n = 12) groups compared to the VEH-VEH (0.24 ± 0.03 absorbance units, n = 12) and BNPP-VEH groups (0.24 ± 0.03 absorbance units, n = 12). (ANOVA p < 0.00001, Newman-Keul's p < 0.05). The BNPP-INH group (0.30 ± 0.02 absorbance units, n = 16) was not significantly different from any of the other groups (groups defined on page 30). 160 FIGURE 2.37 COMPARISON OF HEPATIC CYTOCHROME P-450 LEVELS BETWEEN TREATMENT GROUPS. A comparison of hepatic cytochrome P-450 levels (nmoles/mg protein) between treatment groups (groups defined on page 30). The P-450 levels in the VEH-INH group (1.1 ±0.1 nmoles/mg protein, n = 15) were significantly decreased (by 48%) versus the VEH-VEH controls (2.1 ±0.1 nmoles/mg protein, n = 12). The VEH-VEH* (1.8 ± 0.1 nmoles/mg protein, n = 12), BNPP-INH (1.8 ±0.1 nmoles/mg protein, n = 16) and BNPP-VEH (2.3 ± 0.2 nmoles/mg protein, n = 12) groups were not decreased versus controls (ANOVA p < 0.00001, Newman-Keul's p < 0.05). 161 FIGURE 2.38 COMPARISON OF HEPATIC CYTOCHROME P-450 REDUCTASE ACTIVITY BETWEEN TREATMENT GROUPS. A comparison of hepatic cytochrome P-450 reductase activity (nmoles/minute/mg protein) between treatment groups (groups defined on page 30). There was no difference in reductase activity between the BNPP-INH (343 ± 14 nmoles/min/mg protein, n = 16), BNPP-VEH (360 ± 10 nmoles/min/mg protein, n = 12), VEH-INH (320 ± 17 nmoles/min/mg protein, n = 15), VEH-VEH (361 ± 15 nmoles/min/mg protein, n = 12) or VEH-VEH* groups (362 ± 19 nmoles/min/mg protein, n = 12) (ANOVA p = 0.22). 162 C LU V E H / V E H V E H / I N H B N P P / I N H B N P P / V E H V E H / V E H * Treatment Groups FIGURE 2.39 COMPARISON OF HEPATIC ETHOXYRESORUFrN-0-DEETHYLASE (AN INDEX OF CYP1A1/2) ACTIVITY BETWEEN TREATMENT GROUPS. A comparison of hepatic ethoxyresorufin-O-deethylase (EROD) activity (pmoles/minute/mg protein) between treatment groups (groups defined on page 30). EROD activity was significantly lower in the VEH-INH (180 ± 20 pmoles/min/mg protein, n = 15), BNPP-INH (354 ± 36 pmoles/min/mg protein, n = 16) and VEH-VEH* groups (490 ± 3 5 pmoles/min/mg protein, n = 12) as compared to the BNPP-VEH (804 ± 68 pmoles/min/mg protein, n = 12) and VEH-VEH groups (804 ± 42 pmoles/min/mg protein, n = 12) (ANOVA p < 0.00001, Newman-Keul's p < 0.05). 163 C "O 1500 r Treatment Groups FIGURE 240 COMPARISON OF HEPATIC BENZOYLOXYRESORUFIN-O-DEALKYLASE (AN INDEX OF CYP2B4) ACTIVITY BETWEEN TREATMENT GROUPS. A comparison of hepatic benzoyloxyresorufin-O-dealkylase (BROD) activity (pmoles/minute/mg protein) between treatment groups (groups defined on page 30). Hepatic BROD activity was not significantly different between the VEH-VEH (1254 ± 142 pmoles/min/mg protein, n = 12), VEH-INH (988 ± 131 pmoles/min/mg protein, n = 15), BNPP-INH (1103 ± 136 pmoles/min/mg protein, n = 16), BNPP-VEH (1209 ± 138 pmoles/min/mg protein, n = 12) or VEH-VEH* groups (780 ±113 pmoles/min/mg protein, n = 12) (ANOVA p = 0.13). 164 150 125 h 100 r 75 h 50 r 25 h VEH/VEH VEH/INH BNPP/INH BNPP/VEH VEH/VEH* Treatment Groups FIGURE 2.41 COMPARISON OF HEPATIC PENTOXYRESORUFIN-C?-DEALKYLASE (AN INDEX OF CYP2B4/5) ACTIVITY BETWEEN TREATMENT GROUPS. A comparison of hepatic pentoxyresorufin-O-dealkylase (PROD) activity (pmoles/minute/mg protein) between treatment groups (groups defined on page 30). Hepatic PROD activity was not significantly different between the VEH-VEH (105 ± 14 pmoles/min/mg protein, n = 10), VEH-INH (121 ± 15 pmoles/min/mg protein, n = 11), BNPP-INH (112 ± 12 pmoles/min/mg protein, n = 14), BNPP-VEH (112 ± 8 pmoles/min/mg protein, n = 10) or VEH-VEH* group (82 ± 6 pmoles/min/mg protein) (ANOVA p = 0.18). 16S +5 3.00 2.50 h 2.00 h 1.50 1.00 0.50 h 0.00 * ANOVA p < 0.0001, Newman-Keuls p < 0.05 V E H / I N H < B N P P / V E H = V E H / V E H = V E H / V E H * < B N P P / I N H KXXXXXxl VEH/VEH VEH/INH BNPP/INH BNPP/VEH VEH/VEH* Treatment Groups FIGURE 2.42 COMPARISON OF HEPATIC /?-NITROPHENOL HYDROXYLASE ACTIVITY BETWEEN TREATMENT GROUPS. A comparison of hepatic /?-nitrophenyl hydroxylase (CYP2E1) activity (nmoles/minute/mg protein) between treatment groups (groups defined on page 30). CYP2E1 activity was significantly decreased in the VEH-INH group (0.25 ± 0 . 1 nmoles/min/mg protein, n = 15) versus the VEH-VEH (1.8 ± 0 . 1 nmoles/min/mg protein, n = 12), BNPP-VEH (1.9 ± 0.2 nmoles/min/mg protein, n = 12) and VEH-VEH* (1.4 ± 0 . 1 nmoles/min/mg protein, n = 12) groups (ANOVA p < 0.0001, Newman-Keul's p < 0.05). In addition, hepatic CYP2E1 activity in the BNPP-INH group (3.0 ± 0.4 nmoles/min/mg protein, n = 16) was significantly greater than all of the other groups (10-fold higher than in the VEH-INH group). 166 FIGURE 2.43 CORRELATION OF ACETYLATION RATE VERSUS AN INDEX OF HEPATIC CELL DAMAGE. A correlation of acetylation status (% sulfamethazine acetylated in 20 minutes) versus log peak plasma ASAL activity (log Takahara units) including all rapid acetylator (>50% sulfamethazine acetylated in 20 minutes) animals treated only with INH from Studies 1, 3 and 4 (r = -0.23, r2 = 0.05, n = 50, p > 0.05). 167 TABLE 2.7 TOXICOLOGICAL MARKERS BY GROUP VEH-VEH VEH-INH BNPP-INH BNPP-VEH VEH-VEH* ANOVA Peak plasma ASAL activity (Takahara units) 3.9 ± 0 . 4 n= 12 239 ±83* n= 17 57.6 ± 4 7 n=17 3.8 ± 0 . 4 n=12 4.6 ± 0 . 8 n=12 p < 0.00001 Peak plasma ALT activity (Units/L) 24.9 ± 2.3 n= 12 2 4 0 ± 8 1 f n= 17 62.8 ± 24 n= 17 25.2 ± 2 . 8 n= 12 31.3 ± 3 . 9 n= 12 p = 0.003 Hepatic triglycerides (mg TG/g liver) 7.6 ± 1.7 n= 12 29.4 ± 3.6f n= 15 11.9 ±2 .3 n= 16 9.6 ± 2 . 2 n= 12 10.5 ± 1.6 n= 12 p < 0.00001 32 Hour Plasma Triglycerides (mM) 1.7 ±0 .7 n= 12 6.5 ± 1.6* n= 14 2.5 ± 1.0 n= 16 1.3 ± 0 . 2 n= 12 1.0 ± 0 . 1 n= 12 p = 0.001 Hepatic Glutathione (Limoles/g liver) 7.6 ± 0 . 4 n=12 4.3 ±0.3* n= 15 6.3 ± 0 . 3 f n= 16 7.5 ± 0 . 4 n = 12 6 . 0 ± 0 . 3 r n= 12 p < 0.00001 Hepatic TEARS Formation (absorbance) 0.24 ± 0.03 n= 12 0 . 3 7 ± 0 . 0 3 f n= 15 0.30 ±0 .02 n= 16 0.24 ± 0.03 n= 12 0 . 3 2 ± 0 . 0 1 f n= 12 p < 0.00001 * Food and water restricted control group. * Significantly different from the VEH-VEH group (p < 0.05) using Newman-Keul's multiple comparison test. Groups defined on page 30 168 TABLE 2.8 HYDRAZINE PRODUCTION FROM INCUBATION OF MICROSOMES WITH INH VEH-VEH VEH-INH BNPP-INH BNPP-VEH VEH-VEH* ANOVA Hepatic Amidase Activity (nmoles/mg prot./hour) 38.8 ±4 .5 n= 12 24.1 ±2.0* n= 15 4 . 0 ± 0 . 6 f n = 6 4.7 ±0.7* n = 6 34.7 ±3 .6 n=12 p < 0.00001 Hepatic Amidase Activity (%) 100 ±11 .5 n= 12 62.2 ± 5.3* n= 15 10.4 ± 1.6* n = 6 12.2 ± 1.9* n = 6 89.5± 9.4 n= 12 p < 0.00001 * Food and water restricted control group. * Significantly different from the VEH-VEH group (p < 0.05) using Newman-Keul's multiple comparison test. 169 TABLE 2.9 BNPP IC50 CURVE DATA CURVE 1 CURVE 2 CURVE 3 BNPP nmoles/min/ % nmoles/min/ % nmoles/min/ % Concentration mg protein* mg protein mg protein (uM) 0 40 100 34 100 56 100 0.25 41 103 31 92 55 99 0.5 38 95 30 88 49 87 1 33 82 22 65 43 77 2 23 56 13 38 36 65 4 6.3 16 0 0 14 24 8 0 0 0 0 0 0 16 0 0 0 0 0 0 IC 5 0 (uM)** 2.2 (95% C.I.'s 1.8 to 2.6) 1.5 (95%C.I.'sl.lto2.0) 2.7 (95% C.I.'s 1.5 to 3.8) Values shown for each curve are averages from duplicate samples. ** IC50 derived using the curve fit equation aO + al/(l+(x/a2) ) where aO = minimum, al = maximum, a2 = IC 5 0 and a3 = slope. T A B L E 2 . 1 0 V M A X A N D K M D A T A F R O M L I N E W E A V E R - B U R K A N D E A D I E - H O F S T E E P L O T S LINEWEAVER-BURK PLOT DATA 0 uMBNPP 1.5 uMBNPP 3 uMBNPP VMAX (nmoles/hour/mg prot.) 226 (159 to 293)* 167 (102 to 231) 99 (61 to 138) K M (mM) 9.3 (5.4 to 13) 8.3 (3.7 to 13) 12 (5.3 to 19) EADIE-HOFSTEE PLOT DATA 0 uM BNPP 1.5 uMBNPP 3 uMBNPP VMAX (nmoles/hour/mg prot.) 196 (128 to 418) 116 (60 to 2000) 86 (63 to 137) K M (mM) 7.8 (7.5 to 8.3) 5.3 (4.7 to 6.0) 10 (9.1 to 11) * 95% Confidence Intervals. 171 TABLE 2.11 PLASMA HYDRAZINE LEVELS Time Interval Since Last INH dose VEH-INH BNPP-INH T-test Plasma Hydrazine at 12 hours (uM) 3 hours 25.3 ± 1.4 n= 17 13.5 ± 1 . 0 n = 6* p < 0.00001 Plasma Hydrazine at 24 hours (uM) 15 hours 12.1 ±4 .6 n= 15** 6.1 ± 0 . 9 n = 4** p = 0.22 Plasma Hydrazine at 32 hours (uM) 2 hours 31.8 ± 2 . 8 n= 14 8.2 ± 1 . 8 n= 13** p < 0.00001 * Plasma hydrazine levels were below the quantitation limit (10 uM) in 11 animals in the BNPP-INH group at 12 hours. ** Plasma hydrazine levels were below the quantitation limit (4 uM) in 1 animal in the VEH-INH group and 12 animals in the BNPP-INH group at 24 hours and in 3 of the animals in the BNPP-INH group at 32 hours. Groups defined on page 30 172 TABLE 2.12 MICROSOMAL ENZYME ACTIVITIES BY GROUP VEH-VEH VEH-INH BNPP-INH BNPP-VEH VEH-VEH* ANOVA CYP450 Levels (nmoles/mg protein) 2.1 ± 0 . 1 n=12 1.1 ±0.1* n=15 1.8 ± 0 . 1 n=16 2.3 ± 0 . 2 n= 12 1.8 ± 0 . 1 n=12 p < 0.00001 Reductase Activity (nmoles/min/mg protein) 361 ± 15 n= 12 320 ± 17 n= 15 343 ± 14 n= 16 360 ± 10 n= 12 362 ± 19 n= 12 p = 0.22 EROD Activity (pmoles/min/mg protein) 804 ± 42 n= 12 180 ±20* n= 15 354 ± 36* n= 16 804 ± 68 n= 12 490 ±35* n= 12 p< 0.00001 BROD Activity (pmoles/min/mg protein) 1254 ± 140 n= 12 988 ± 1 3 0 n= 15 1103 ± 140 n = 16 1209 ± 140 n= 12 780 ± 1 1 0 n = 12 p = 0.13 PROD Activity (pmoles/min/mg protein) 105 ± 14 n= 10 121 ± 1 5 n= 11 112 ± 1 2 n= 14 112 ± 8 n= 10 82 ± 6 n= 11 p = 0.18 /7-Nitrophenol Hydroxylase (CYP2E1) Activity (nmoles/min/mg protein) 1.8 ± 0 . 1 n= 12 0.25 ±0.1* n= 15 3.0 ±0.4* n= 16 1.9 ± 0 . 2 n= 12 1.4 ± 0 . 1 n= 12 p < 0.00001 * Food and water restricted control group. * Significantly different from the VEH-VEH group (p < 0.05) using Newman-Keul's multiple comparison test. Groups defined on page 30 173 DISCUSSION STUDY 1 DETERMINATION OF HEPATIC CELL DAMAGE, HEPATIC STEATOSIS AND HYPERTRIGLYCERIDEMIA, AND THE EFFECTS OF ACETYLATION RATE AND GENDER ON INH-INDUCED HEPATOTOXICITY IN RABBITS The rabbits used in the first study were almost entirely rapid acetylators of sulfamethazine (14/15 with greater than 50% of intravenously administered sulfamethazine acetylated in 20 minutes). Thus, interpretation of these results is relevant for rapid acetylator rabbits only. The one slow acetylator rabbit was excluded from the correlational analysis in order to focus on the rapid acetylators; however, it was not more susceptible to INH-induced hepatic cell damage that the rapid acetylators (see Figure 2.2). Since the rabbits used in this study are inbred, these results suggest that the majority of these rabbits are genetically heterozygous dominant or homozygous dominant for the rapid acetylation trait. The use of rabbits as a model for genetically determined acetylation rates in humans is valuable since it has been established that N-acetyltransferase-2 is the enzyme responsible for polymorphic acetylation of both sulfamethazine and isoniazid in rabbits and humans (Blum et al, 1989; Vatsis et al, 1995). The acetylation characteristics for sulfamethazine in rabbits have been shown to directly parallel the acetylation characteristics of INH in vivo and in vitro, and acetylation phenotyping using sulfamethazine is a reliable technique (Gordon etal, 1973). One way to estimate the probable rate of acetylation of INH is to use the rate of acetylation of sulfamethazine as a predictor. In our study in which plasma levels of INH 174 were determined (Study 2 in this thesis; Sarich et al, 1996), the rate of acetylation of sulfamethazine correlated negatively with plasma INH concentrations at 32 hours (r = -0.73, r2 = 0.53, n = 15, p < 0.01) and at 48 hours (r = -0.80, r2 = 0.64, n = 15, p < 0.01) (these data were not included in the previous publication). Therefore, the rate of acetylation of sulfamethazine is a good marker for the rate of metabolism of INH. If the toxic phenomena measured in the present study are related to INH itself, then even within this population of rapid acetylators of INH, percent acetylation of sulfamethazine should negatively correlate with toxicity. In humans it has been suggested that both rapid (Mitchell et al., 1975b) and slow (Dickinson et al., 1981; Musch et al., 1982) acetylator phenotypes are at greater risk of developing INH-induced hepatotoxicity. However, many studies have reported no difference between the incidence of INH-induced hepatotoxicity in rapid and slow acetylators (Riska, 1976; Hong Kong, 1977; Singapore, 1977; Neill et al, 1990; Singh et al, 1995). The findings in the present study suggest that in a population of rapid acetylator rabbits, the rate of acetylation of sulfamethazine, used as a marker for the rate of acetylation of INH, does not correlate with, and thus does not predict the severity of, INH-induced hepatotoxicity. The lack of correlation between acetylation rate of sulfamethazine and INH-induced hepatic cell damage found in this study is consistent with the fact that although acetylation may lead to increased production of toxic metabolites, it also leads to increased detoxification. The lack of correlation of hepatic triglyceride content with the acetylation rate of sulfamethazine and the weak negative correlation between rate of 175 acetylation of sulfamethazine, and peak plasma triglyceride concentration, is also consistent with the complexity of the role of acetylation in this model. Since the acetylation rate of sulfamethazine correlated negatively with plasma INH levels, it confirms that sulfamethazine is a predictor of rate of INH acetylation. If INH itself is toxic, then acetylation should correlate negatively with toxicity. These data agree with our other observations that plasma levels of INH do not correlate with markers of toxicity in this model of INH-induced hepatotoxicity in rabbits (Sarich et al, 1996; Study 2). The pathological changes observed in the present study are in concordance with previously reported data on INH-induced pathological changes in rabbits. Plasma ASAL activities in the present study were increased by 27-fold, which is comparable to a 3 3-fold increase in plasma ASAL activities previously reported using this model of INH-induced hepatotoxicity in rabbits (Sarich et al, 1995). The observed 7.5-fold increase in hepatic triglyceride accumulation and 13-fold increase in plasma triglyceride levels are also very similar to a previously reported 6-fold increase in hepatic triglyceride accumulation and 12-fold increase in plasma triglyceride levels after INH administration to rabbits (0.37 mmoles/kg, 50 mg/kg/day intraperitoneally for 11 days) (Karthekeyan and Krishnamoorthy, 1991). Only male rabbits were used in our original study (Sarich et al, 1995) and so in the present study, the susceptibility of both male and female rabbits to INH-induced pathological changes was assessed. This comparison did not reveal or suggest any differences in susceptibility between sexes. This is in contrast to observational studies in humans which suggested an increased susceptibility in females (Moulding et al, 1989; 176 Snider and Caras 1992). In our animal model of INH-induced hepatotoxicity, the animals were given INH based on body wieght, whereas in humans, a standard dose of 300 mg per day is given regardless of gender or body weight. Since females weigh less on average than males, it is possible that in humans the apparent increase in susceptibility of females is due to a greater dose of INH on a mg/kg basis, suggesting that this toxicity may have a dose related component. High dose INH-therapy is in fact associated with an increased incidence of peripheral neuritis and gastric intolerance in humans (Biehl and Nimitz, 1954). The mechanisms underlying the pathological changes of hepatic cell damage, hepatic steatosis and hypertriglyceridemia are not completely understood. Hepatic cell damage is the primary pathological feature seen in liver biopsies in INH-induced hepatotoxicity in humans (Black et al., 1975). The only report of INH-induced hepatic steatosis in humans was in patients who were also receiving rifampin (Pessayre et al., 1977). Whether INH leads to an increase in liver triglycerides or plasma triglycerides in humans has never been studied. Both INH and hydrazine have been shown to cause alterations in lipid handling in rabbits. In rabbit adipose tissue, INH administration decreases total lipid, triglyceride and cholesterol concentrations and increases free fatty acid concentrations (Karthikeyan and Krishnamoorthy, 1991). In plasma and liver, INH administration substantially increases total lipid, triglyceride, cholesterol and free fatty acid concentrations. These changes are consistent with a mechanism involving mobilisation of free fatty acids from the peripheral adipose tissue and accumulation of lipids in the plasma and liver. Based on previous 177 reports, it is likely that the INH-induced elevation of plasma triglycerides precedes hepatic triglyceride accumulation (Lamb and Banks, 1979; Karthekeyan and Krishnamoorthy, 1991). A significant correlation between hepatic and peak plasma triglycerides (r = +0.69, r2 = 0.48, n = 15, p < 0.005) in the present study supports a link between elevation of hepatic and plasma triglycerides. Similar to the effects of INH-administration in rabbits, administration of hydrazine to rats causes elevated serum free fatty acids, increased uptake of free fatty acid by the liver, increased hepatic triglyceride synthesis and decreased lipoprotein secretion (Lamb and Banks, 1979). Specifically, a 3-5 fold increase in hepatic triglyceride levels occurs after administration of hydrazine (1.6 mmoles/kg, 50 mg/kg, i.p.) to rats (Jenner and Timbrell, 1994a) The correlations between log peak plasma ASAL activity and both hepatic triglyceride content and peak plasma triglyceride concentration, though significant, are weak and account for less than 50% of the total variance in these markers. It is not clear whether hepatic cell damage precedes the onset of hepatic steatosis, or vice versa, since we did not measure the time-course of INH-induced hepatic triglyceride accumulation. However, it was observed that peak severity of hepatic cell damage does not precede peak severity of hypertriglyceridemia in individual rabbits. The lack of a pattern of one marker repeatedly peaking prior to the other argues against the possibility that hepatic cell damage leads to the onset of hypertriglyceridemia or vice versa. In summary, in a population of mostly (14/15) rapid acetylator rabbits, the acetylation rate of sulfamethazine did not correlate with the severity of INH-induced hepatic cell damage or steatosis. INH-administration resulted in a marked increase in 178 plasma ASAL activity, and hepatic and plasma triglyceride levels. Gender had no effect on the susceptibility to INH-induced hepatic cell damage, steatosis or hypertriglyceridemia. INH-induced hepatic triglyceride accumulation and hypertriglyceridemia correlate significantly and these two pathological changes appear to be part of a systemic alteration in lipid disposition which are linked by a common toxicological mechanism. These experiments further our understanding of this rabbit model of INH-induced hepatotoxicity and demonstrate the lack of effect of acetylation rate of sulfamethazine and gender on the magnitude of the toxicities in a population of rapid acetylator rabbits. STUDY 2 DETERMINATION OF INH, ACETYLHYDRAZINE AND HYDRAZINE CONCENTRATION IN PLASMA SAMPLES AND CORRELATION WITH INH-INDUCED HEPATOTOXICITY IN RABBITS. Although plasma INH levels at 32 hours correlated significantly with log plasma ASAL activity at 48 hours and log peak plasma ASAL activity, the correlations were weak and accounted for less than 50% of total variance. It has been reported in humans that plasma INH levels do not correlate with susceptibility to INH-induced hepatotoxicity (Mitchell et al, 1975). Although the concentration range is limited, the data from the present study support such findings. Both acetylhydrazine and hydrazine are proven metabolites of INH in humans (Timbrell et al, 1977; Noda et al, 1978; Blair et al, 1985; Peretti et al, 1987; Gent et al, 1992) and rabbits (Thomas et al, 1981). In human slow acetylators, the potential for accumulation of acetylhydrazine and hydrazine in plasma has been recognized by the 179 presence of these metabolites in urine between 24 and 36 hours after a single 300 mg oral dose of INH (Peretti et al, 1987). It has been suggested that acetylhydrazine is the hepatotoxic INH-metabolite in rats (Snodgrass et al, 1974; Mitchell et al, 1976; Timbrell et al, 1980). If this were also the case in rabbits, one would expect to observe a significant positive correlation between hepatic cell damage and the amount of acetylhydrazine formed, as reflected in the 32 or 48 hour plasma acetylhydrazine concentration. This was not the case in the present study. These results are in concordance with other data showing that acetylation of hydrazine compounds is a detoxification reaction (McQueen etal, 1982). The main positive finding of the present study is the significant correlation between markers of hepatic cell damage and 32 hour plasma hydrazine levels. This suggests that the amount of hydrazine formed from INH and/or acetylhydrazine is an important determinant of INH-induced hepatic cell damage. Hydrazine has been previously implicated as the hepatotoxic metabolite of INH. Hydrazine is a known hepatotoxin in rats (Scales and Timbrell 1982; Timbrell et al, 1982; Jenner and Timbrell 1994a; Jenner and Timbrell 1994b),-rabbits (Yard and McKennis 1955; McKennis etal, 1956; Noda etal, 1983) and monkeys (Patrick and Back 1965). Although much of the research on hydrazine hepatotoxicity has focused on hydrazine-induced fatty changes in the liver, hydrazine has also been shown to cause hepatic necrosis (Patrick and Back 1965; Noda et al, 1983). Prolonged elimination of hydrazine after administration of INH results in a slowly rising plasma baseline of hydrazine which is exaggerated in some patients; the patient with the highest plasma levels of hydrazine in one study also had elevated plasma bilirubin and 180 transaminases (Gent et al, 1992). Progressive accumulation of hydrazine has been shown to occur in INH-treated patients over a period of at least six weeks of treatment. This could explain the delayed appearance of INH hepatotoxicity, which most often occurs within the first 8 weeks of daily INH therapy (Scharer and Smith, 1969) but which can occur 24 weeks or longer into INH therapy (Byrd et al., 1972). In addition, greatly elevated plasma hydrazine levels have been reported in fatal INH-induced hepatotoxicity involving a 74-year old man (Woo et al, 1992). The lack of correlation of INH-induced hepatic cell damage with plasma acetylhydrazine levels is interesting and is in contrast to research suggesting that acetylhydrazine is the INH-derived hepatotoxin in INH-induced hepatotoxicity (Mitchell et al., 1976; Timbrell et al., 1980). Most of the research on acetylhydrazine-induced hepatic necrosis was performed in rats. Although recent evidence that a dose of 1 mmol/kg of acetylhydrazine injected i.p. into phenobarbital-pretreated rats did not produce a significant increase in hepatic necrosis has put the toxicity of acetylhydrazine in rats into question (Ganley et al., 1994), it is possible that different enzymatic pathways present in rabbits, and rats-could explain the lack of correlation of acetylhydrazine to hepatic cell damage in rabbits. For example, higher amidase activity or lower reductase activity, in rabbits versus rats, respectively (Whitehouse et al, 1983) could explain differences between the two species. Experiments by Mtchell and colleagues (1976) showed that inhibition of amidase activity using the amidase inhibitor bis-/?-nitrophenyl phosphate did* not prevent acetylhydrazine-induced hepatic necrosis suggesting that acetylhydrazine itself and not hydrazine was responsible for the toxicity. 181 Whereas the INH-injection protocol causes hepatic cell damage in over 50% of the animals, administration of acetylhydrazine to rabbits at an equimolar dose to the dose of INH given does not cause hepatic cell damage (T. Zhou, M.Sc. Thesis). In comparison, administration of hydrazine at one half of the equimolar dose of INH causes significant hepatic cell damage in 100% of the animals. Higher doses of hydrazine cause almost immediate convulsions and death. Another prominent feature of INH-induced hepatotoxicity in rabbits, unlike in man, is hepatic triglyceride accumulation (hepatic steatosis) (McKennis et al., 1956; Karthikeyan and Krishnamoorthy 1991). Although liver triglyceride accumulation correlated significantly with plasma hydrazine concentrations at 48 hours, less than 50% of total variance was explained by this correlation. Unfortunately, these results do not provide any clues as to whether INH, acetylhydrazine or hydrazine is/are causative in INH-induced hepatic steatosis. One of the limitations of the present study is the lack of time-course data for hepatic triglyceride accumulation. However, plasma triglyceride levels can be used as an indicator of hepatic triglyceride accumulation. Administration, of I N H to rabbits leads to decreased triglyceride levels in adipose tissue and increased triglyceride content of the liver (Karthikeyan and Krishnamoorthy 1991). In rats, there is evidence that steatosis induced by hydrazines involves elevated serum free fatty acids, increased uptake of free fatty acid by the liver, increased hepatic triglyceride synthesis and decreased lipoprotein secretion (Lamb and Banks 1979). Since INH-induced elevation of plasma triglyceride concentrations precedes accumulation of hepatic triglycerides, a correlational analysis of 1 8 2 plasma triglycerides with plasma levels of INH, acetylhydrazine and hydrazine was performed. The lack of correlation of plasma triglyceridejevels with plasma levels of INH, acetylhydrazine and hydrazine suggests that these compounds do not directly cause hypertriglyceridemia. The lack of information concerning the actual time-course of liver triglyceride accumulation prevented more definitive conclusions regarding the role of INH, acetylhydrazine and/or hydrazine in the production of increased liver triglyceride accumulation. In conclusion, neither INH nor acetylhydrazine plasma levels correlated with markers of INH-induced hepatic cell damage. However, a strong correlation exists between hydrazine and markers of INH-induced hepatic cell damage. In general (except for liver triglycerides and plasma hydrazine at 48 hours), this study did not implicate INH, acetylhydrazine or hydrazine in INH-induced fatty changes (hepatic steatosis and/or hypertriglyceridemia). The in vivo metabolic profile of INH, acetylhydrazine and hydrazine during the production of INH-induced hepatotoxicity in rabbits indicates that hydrazine is most likely involved in the mechanism of hepatic cell damage. STUDY 3 THE R O L E OF INCREASED A N D DECREASED HEPATIC R E D U C T A S E ACTIVITY IN INH-INDUCED HEPATOTOXICITY IN RABBITS. This study was designed to test the hypothesis that an induced change in hepatic reductase activity would alter production of reactive and damaging metabolites from hydrazine, a metabolic product of INH, and thereby alter the severity of INH-induced hepatic cell damage. This hypothesis is based on the assumption that the reductase enzyme 183 is the rate limiting step in conversion of hydrazine to reactive and pathologically damaging metabolites. To test this hypothesis, we examined the effects of increases and decreases in the activity of the reductase enzyme on the severity of INH-induced hepatotoxicity. In rats, a decrease in plasma thyroxine levels (induced by the anti-thyroid drug methimazole) is associated with a decrease in hepatic reductase activity (Ram and Waxman, 1992) while an increase in plasma thyroxine levels (achieved by L-thyroxine administration) is associated with an increase in hepatic reductase activity (Kato and Takahashi, 1968; Waxman etal, 1989). Pilot study results led to the identification of a methimazole pretreatment regimen which was below its toxicity threshold and likely to lower plasma free T4 levels and hepatic microsomal reductase activities provided the regimen was continued during the INH-injection protocol. Using this regimen, after the ninth day of methimazole pretreatment [50 mg/kg; 0.44 mmoles/kg methimazole (i.p.) every twelve hours], plasma free T4 levels were lowered. Unfortunately, the unexpected death of these animals, due to an apparent interaction with INH, prevented further evaluation of this pretreatment regimen; In the second methimazole pretreatment regimen [utilising a dose of 0.0313% w/v (approximately 25 mg/kg/day; 0.22 mmoles/kg/day) in the drinking water for 7 days, followed by 0.0626% w/v (approximately 50 mg/kg/day; 0.44 mmoles/kg/day) for 19 days], the plasma free T4 levels were decreased versus baseline levels. However, three out of five animals in this group died after receiving INH and only two remained for observation. Therefore, analysis of the effect of methimazole pretreatment on reductase 184 activity and on the severity of INH-induced hepatotoxicity was not possible due to the low numbers. It is not known whether this fatal interaction was due to interaction of INH with the state of hypothyroidism or with methimazole itself. Hydrazine is known to induce convulsions in mice, rats, rabbits, dogs (Benson et al, 1952) and man (Reid, 1965). It has been proposed that hydrazine induces central nervous system toxicity and convulsions by depleting endogenous pyridoxine (vitamin B6) stores through formation of hydrazine-pyridoxal hydrazones (Williams and Bain, 1961). Possibly, convulsions are a result of inhibition of a pyridoxine-dependent enzyme(s) in the central nervous system. Administration of pyridoxine has been shown to protect rats (Cornish, 1969), rabbits (Hein and Weber, 1984) and humans (reviewed in Holtz and Palm, 1964) from hydrazine/hydrazide-induced convulsions. Since the reductase enzyme may be active in the metabolism of hydrazine, it is possible that the animals which died had increased plasma hydrazine levels due to inhibition of reductase activity by methimazole. This could explain the high incidence of fatal convulsions in this group of methimazole and INH-treated animals (6 of 8). Since hydrazine is also considered to be an hepatotoxic INH metabolite in INH-induced hepatic cell damage, a high level of hepatic cell damage in the two surviving animals would be expected. However, it may be that their apparent protection from INH-induced convulsions and death also protected them from severe INH-induced hepatic cell damage, as they showed only a mild degree of hepatic cell damage. After combination of the model of hyperthyroidism with the INH-injection protocol, it was clear that pretreatment of rabbits with L-thyroxine does not increase the 185 severity of INH-induced hepatic cell damage. Although plasma A S A L activity was significantly elevated in T 4 - I N H and INH-treated animals versus control animals, the INH-only group had significantly higher plasma A S A L activity than the T 4 - I N H animals. This suggests that the T 4 - I N H animals were somewhat protected from INH-induced hepatic cell damage. Although plasma A L T activities were significantly greater in the INH-only group versus the T 4 - I N H and control groups ( A N O V A p = 0.03), statistical difference was not detected using the Newman Keul 's multiple comparisons test. However, plasma A S A L activity is believed to be a better marker of hepatic cell damage than A L T as previously discussed (Sarich et ah, 1995). In fact, this observation is evident in the present study since the range of A S A L activities from baseline to the maximal A S A L value is 86-fold whereas the range from baseline to maximal A L T activity is only 14-fold. This demonstrates that plasma A S A L activity is approximately 6 times more sensitive as a marker for hepatic cell damage than plasma A L T activity; Thus, on the basis of both plasma A L T and A S A L activities, the data suggest that the INH-only animals experienced a significantly greater degree of hepatic cell damage as compared to T 4 - I N H and control animals. It appears that pre- and co-treatment of I N H with L-thyroxine decreases the severity of INH-induced hepatic cell damage. Hepatic cytochrome P-450 levels were decreased by approximately 50%, as shown by the significantly decreased levels in both T 4 - I N H and INH-only groups as compared with the vehicle control animals. Both INH-treated groups experienced this drop in P-450 levels even though the two groups had significantly different degrees o f hepatic cell damage and hepatic steatosis. The follow-up study shows that L-thyroxine pretreatment 186 itself decreases P-450 levels to 60% of control levels. Therefore, the majority of the decrease in P-450 levels in the T4-INH group (to 49% of control) is due to the effects of L-thyroxine. L-thyroxine has been previously reported to decrease hepatic P-450 content by 50% in male, but not female, rats (Kato and Takahashi, 1968). The 51% decrease in P-450 levels in the INH-only group may be due to the INH metabolite acetylhydrazine, since acetylhydrazine has been shown to dose-dependently decrease hepatic cytochrome P-450 levels in rats by approximately 60% (Bahri et al., 1981). However, hydrazine is also a likely candidate for the decreased P-450 levels. It is possible that the decreased P-450 levels caused by L-thyroxine in the T4-INH group played a role in the decreased severity of hepatotoxicity in that group. This suggests that activation of hydrazine by a P-450 may be necessary for toxicity to occur. The hepatic reductase activity in the T4-INH group was significantly higher than in the INH-only and vehicle control animals. The follow-up study showed that hepatic reductase activities appeared to be increasing after 5 days of L-thyroxine administration; however, they were not yet significantly increased as they were in the full length study after 10 days of L-thyroxine administration. Although this induction has been reported in rats, it has not been previously reported in rabbits. It is worth noting that the increase in reductase activity in the T4-INH group occurred in the presence of decreased P-450 levels. In the follow-up study (n = 5), methimazole pretreatment was found to decrease hepatic reductase activity by 60%. In addition, the reductase activities in the methimazole-treated animals which survived INH administration were much lower than in the T4-INH 187 or INH-only groups (n = 2). Therefore, methimazole can decrease hepatic reductase activity in rabbits in addition to in rats (Ram and Waxman, 1992). From the results in this study, it appears likely that the reactive intermediates reported to be formed from the conversion of hydrazine by the reductase enzyme in vitro (Noda et al, 1988) are not involved in the pathogenesis of INH-induced hepatotoxicity. Since the results in the present study suggest that increased reductase activity does not increase the severity of INH-induced hepatotoxicity, these "reactive intermediates", although chemically reactive, may not be pathogenic or have toxicological significance. It is possible that this pathway is actually a detoxification process for hydrazine. It is clear that not all of the metabolites of hydrazine are toxic, because a major product (25%) of hydrazine metabolism in rats is nitrogen gas (N2) (Springer et al, 1981). However, approximately 25% of the dose of hydrazine is also not accounted for after 48 hours. Since the data in the present study suggest that an increase in hepatic reductase activity is associated with a decrease in the severity of INH-induced hepatotoxicity, a negative correlation of INH-induced hepatotoxicity and reductase activity would therefore be expected in the T4-INH animals. However, there was no (p > 0.05) correlation in the T4-INH group of hepatic reductase activity with log peak plasma ASAL activity, log peak plasma ALT activity or hepatic triglyceride accumulation. It is also interesting that rats are relatively resistant to INH-induced hepatotoxicity. As compared with rabbits, rats have approximately double the hepatic reductase activity (Whitehouse et al, 1983). Another possible explanation for the results in the present study is that treatment with L-thyroxine inhibits the activity of an as yet unidentified enzymatic pathway on which 188 reductase is dependent for the production o f INH-hepatotoxicity. F o r example, since the mechanism o f INH- induced hepatic cell damage in rabbits is believed to involve hydrazine [and/or its metabolite(s)] as the hepatotoxic INH-der ived species, reduction o f available free hydrazine would markedly decrease the amount o f hydrazine available to be converted by reductase into reactive and potentially toxic intermediates. One such enzyme is amidase which produces free hydrazine from I N H and acetylhydrazine. However , i f this hypothesis is correct, and prevention o f INH- induced hepatotoxicity is due to the inhibition o f the amidase enzyme by L-thyroxine, reductase activity in the control (non L-thyroxine-treated) animals would be expected to correlate with INH-hepat ic cell damage. However , in the I N H - o n l y animals, there was no statistically significant correlation (p > 0.05) o f hepatic reductase activity with log peak plasma A S A L activity, log peak plasma A L T activity or hepatic triglyceride accumulation. It is also possible that I N H and L-thyroxine compete for an important step in hepatic metabolism. I f a metabolic pathway which is responsible for the conversion o f I N H , or its metabolite(s), into hepatotoxic species does exist, and i f L-thyroxine is also metabolised via this route, this could explain why the T 4 - I N H group experienced less toxicity than the I N H - o n l y group. Al though the metabolism o f L-thyroxine is k n o w n to involve conversion to triiodothyronine and both L-thyroxine and triiodothyronine are conjugated to glucuronides and sulfates and excreted via biliary or fecal routes (Ingbar, 1975), there is no evidence that I N H and L-thyroxine share c o m m o n pathways o f metabolism. In fact, high levels o f plasma L-thyroxine which occur in thyrotoxicosis do not appear to have an effect on I N H metabolism since no significant change in elimination 189 half-life occurred in thyrotoxic patients as compared to the same patients after conversion to euthyroidism (Littley et al, 1988). The activity of CYP2E1 in the INH-only group was significantly decreased (by 38%) as compared with T4-INH and vehicle control groups. Pretreatment with L-thyroxine or methimazole had no effect on CYP2E1 activities in the follow-up study. In humans, INH is known to inhibit CYP2E1 activity and induction of CYP2E1 activity occurs 48 hours after the last dose of INH (Zand et al, 1993). In the present study, CYP2E1 activity in the INH-only group did not return to baseline by 96 hours. It is possible that pathological changes which took place during INH-induced hepatotoxicity slowed the recovery of CYP2E1 back to, or above, baseline activity. This seems to be an indication that whereas a recoverable "inhibition" of CYP2E1 activity occurs at doses of INH given to humans (0.04 mmoles/kg/day for 7 days), "destruction" of CYP2E1 activity occurs at the doses of INH given to these rabbits (1.1 mmoles/kg/day for two days). An interesting finding is that the CYP2E1 activities in the T4-INH group were increased by 30% above controls (not significant). It would be interesting to know if CYP2E1 activity in the T4-INH group was increasing or decreasing at 96 hours. Since the activity was measured 63 hours after the last dose of INH (last dose at 33 hours), it would be predicted to be returning to baseline levels. This would mean that induction of CYP2E1 activity had occurred, which is consistent with the INH-induced CYP2E1 activity changes in humans. Nevertheless, these results at least show that CYP2E1 activity was not destroyed in the T4-INH group, as appears to have occurred in the INH-only group. It is tempting to hypothesize that the apparent prevention of destruction of CYP2E1 activity conferred by 190 L-thyroxine in the T4-INH group is related to the decreased severity of INH-induced hepatotoxicity in the T4-INH group. Possibly, interference of an important step in INH-metabolism by T4 results in both prevention of destruction of CYP2E1 activity and decreased severity of INH-induced hepatotoxicity. CYP2E1 has been suggested to be involved in both acetaminophen- (Nelson, 1990) and ethanol-induced hepatotoxicity (French etal, 1993). Production ofiV-acetyl-/?-benzoquinone imine (NAPQI), the toxic metabolite in acetaminophen-induced hepatotoxicity, in vitro has been reported to be due to oxidative metabolism by rat and human CYP1A2, CYP2E1 and CYP3A4 (Raucy et al, 1989; Patten et al, 1993; Thummel et al, 1993). In addition, acetaminophen-induced hepatotoxicity in mice has been shown to be associated with decreased activity of CYP2E1 and CYP1A2 (Snawder et al, 1994). This suggests that a characteristic of enzymes involved in the production of toxic intermediates is that the enzymes themselves become damaged in addition to the surrounding vital cellular organelles which is the cause of cell death. In relation to INH-induced hepatotoxicity, it is therefore possible that the observed decrease in activity of CYP2E1 in the INH-only group is due to bioactivation of hydrazine which results in damage to both CYP2E1 itself and nearby vital cellular organelles. The inhibition of hepatic triglyceride accumulation in the T4-INH group may also be related to the increased metabolic state in hyperthyroidism. It was observed in this study that the animals receiving L-thyroxine had greater decreases in body weight over the period of the study and smaller livers weights at the end of the study. Thus, the increased metabolic state in hyperthyroidism had measurable effects on body and liver weight and 191 may have been a factor in the amount of hepatic triglyceride accumulation in the T4-INH group. Therefore, the apparent prevention of hepatic triglyceride accumulation in the T4-INH group is likely a combination of the decrease in the toxic potential of INH as observed by the decrease in severity of INH-induced hepatic cell damage and the increased metabolic state of hyperthyroidism. A limitation in the design of this study is that the microsomal reductase and CYP2E1 enzyme activities as well as the cytochrome P-450 levels are measured at 96 hours after the first dose of INH and, in the majority of the animals, after the peak level of hepatic cell damage has occurred (for details on time course of the hepatic cell damage, see Sarich et al, 1995). The last dose of INH administered to the animals was at 33 hours, 63 hours prior to sacrifice. Although trace amounts of INH remain in plasma 24 hours after a single 5 mg/kg (0.04 mmole/kg) dose of INH in rabbits (Walubo et al, 1991), this study would not likely detect all the changes in enzyme activity during this time because the microsomes are not prepared until 63 hours after the last dose of INH in the present study. Therefore, in order to better characterize the activity of potentially important hepatic enzymes, estimation of their activity nearer to the period of exposure to INH is required. The original goal of this study was to observe the effect of alterations in hepatic reductase activity on INH-induced hepatotoxicity. The toxicologic changes in this study (plasma ASAL and ALT activities, hepatic triglyceride accumulation) were monitored over 96 hours. If there had been a significant increase in INH hepatotoxicity in animals treated with L-thyroxine and INH, a follow-up study monitoring hepatic enzyme activities 192 during active hepatotoxic injury (e.g. 48 hours) would have been valuable. In Study 4 of this thesis, a 48 hour protocol was used in order to study more closely the involvement of liver enzymes in this toxicity. Summary A state of experimental hypothyroidism was induced with methimazole which decreased plasma free T4 levels and hepatic reductase activities. Convulsions preceding death occurred in 6 of 8 rabbits administered INH after pretreatment with methimazole. In developing a state of experimental hyperthyroidism, treatment of rabbits with L-thyroxine increased plasma free T4 levels as well as hepatic reductase activities. L-thyroxine treatment significantly decreased cytochrome P-450 levels. Pre- and co-treatment of INH with L-thyroxine decreases the severity of INH-induced hepatic cell damage and prevents INH-induced hepatic steatosis. These findings are contrary to our original hypothesis, which postulated that an increase in hepatic reductase activity would increase the severity of INH-induced hepatotoxicity. STUDY 4 THE EFFECT OF INHIBITION OF INH-AMIDASE BY BIS-/?-NITROPHENYL PHOSPHATE ON INH-INDUCED HEPATOTOXICITY IN RABBITS. The purpose of this study was to determine whether pretreatment with the amidase inhibitor BNPP would decrease the production of hydrazine from INH in vivo and thus decrease the severity of INH-induced hepatic cell damage. BNPP did decrease the severity of INH-induced hepatic cell damage since only the positive control (VEH-INH) group showed a significant degree of hepatic cell damage. In addition, the VEH-INH group was 193 the only group to experience significant hepatic steatosis and hypertriglyceridemia. It is clear from the results that pretreatment with BNPP prevents INH-induced hepatic cell damage, hepatic steatosis and hypertriglyceridemia. It was predicted that BNPP would prevent INH-induced hepatic cell damage but it was not predicted that BNPP would prevent INH-induced hepatic steatosis and hypertriglyceridemia. Interpretation of the above findings results in three obvious possibilities. The first is that hydrazine is involved, directly or indirectly, in the development of all three of these pathological changes rather than just in INH-induced hepatic cell damage. This possibility is supported by previous studies which have shown that hydrazine administration to animals produced fatty changes including hepatic steatosis and hypertriglyceridemia (Yard and McKennis, 1955; Patrick and Back, 1965; Clark et al, 1970). However, in Study 2, INH-induced hepatic steatosis was found not to correlate with plasma INH, acetylhydrazine or hydrazine levels. The findings in the present study indicate that although the findings in Study 2 did not implicate hydrazine in the pathogenesis of INH-induced hepatic steatosis, they did not exclude hydrazine as a possible factor in INH-induced hepatic steatosis. Therefore, the production of hepatic cell damage, hepatic steatosis and hypertriglyceridemia appears dependent upon "BNPP-sensitive" pathways which affect plasma hydrazine levels. The second possible explanation for the inhibition of INH-induced hepatic cell damage, hepatic steatosis and hypertriglyceridemia by BNPP is that BNPP alters pathways in addition to, or other than, the inhibition of hydrolysis of INH to hydrazine. However, the present study confirms that BNPP does inhibit the production of hydrazine from INH. 194 The significantly decreased plasma hydrazine levels at 12 and 32 hours in the BNPP-INH group versus the VEH-INH group suggests that BNPP inhibited formation of hydrazine from INH in vivo, presumably by inhibition of amidase activity. A significant decrease in hepatic microsomal hydrazine production in the BNPP-INH (by 90%) and BNPP-VEH (by 88%) groups confirms that BNPP inhibits the direct conversion of INH to hydrazine in vitro. In addition, it was shown that the direct conversion of INH to hydrazine in incubations of untreated microsomes could be completely inhibited by the addition of BNPP at low (pM) concentrations. Therefore, if BNPP prevented INH-induced hepatic steatosis and hypertriglyceridemia by mechanisms other than by inhibition of hepatic amidase activity, the mechanism is in addition to the ability to inhibit hepatic amidase activity. BNPP could potentially possess properties or actions in addition to INH-amidase inhibition which prevent the onset of the pathological changes in INH-induced hepatotoxicity. Prevention of INH-induced fatty liver in rabbits by co-administration of pyridoxine hydrochloride has been suggested to be via a mechanism involving deactivation of INH-derived primary amines by pyridoxal hydrazone formation (Whitehouse et al, 1983); This demonstrates that it appears possible to inhibit INH-induced fatty changes via mechanisms other than by inhibition of INH-amidase activity. Thirdly, since BNPP likely decreased the production of acetylhydrazine from acetyl-INH (and from hydrazine), it is possible that the decreased production of acetylhydrazine played a role in the decreased severity of INH-induced hepatotoxicity. However, the evidence from Study 2 that hydrazine, and not acetylhydrazine, correlates 195 with INH-induced hepatic cell damage suggests that the decreased production of hydrazine is the most likely mechanism. It would have been useful to measure the rate of production of hydrazine from both acetyl-INH and acetylhydrazine in addition to INH in microsomal incubations in order to estimate whether BNPP also had an effect on these amidase catalysed pathways. In addition, incubation of INH in an S9 fraction containing both microsomal amidase and cytoplasmic N-acetyltransferase, the two major INH metabolising enzymes, would have been valuable in order to more closely estimate the in vivo metabolism of INH and the effect of BNPP on it. However, microsomal amidase activity as measured in this study (direct conversion of INH to hydrazine) correlates significantly with plasma levels of hydrazine (r = +0.83 with 12 hour plasma hydrazine; r = +0.71 with 32 hour plasma hydrazine). BNPP is part of a family of organophosphorus compounds which have been used to study the structure of the active site and mechanism of action of serine hydrolases (Junge and Krisch, 1975). BNPP, however, does not act like most other organophosphates since it does not inhibit acetylcholinesterase, cholinesterase, chymotrypsin, or trypsin. Thus, it is much less toxic than conventional organophosphorus inhibitors. In mice, BNPP has an acute LD 5 0 of 410 mg/kg (Heymann and Krisch, 1967). The mechanism of action of BNPP involves phosphorylation of an hydroxyl group by p-nitrophenyl phosphate and release of a /?-nitrophenyl group: 196 Although some amidases have been studied in detail (Junge and Krisch, 1975), detailed studies on the INH-amidase enzyme specifically have not been reported. In the present study, it was observed that the inhibition of hepatic INH-amidase by BNPP persisted in microsomes in vitro from BNPP-treated animals. This confirms that BNPP inhibits the hepatic INH-amidase irreversibly, as might be expected from the reported phosphorylating mechanism of BNPP. Analyses using Lineweaver-Burk and Eadie-Hofstee plots of data from microsomal incubations of INH in the presence and absence of BNPP revealed that BNPP decreases the Vmax of hepatic INH-amidase. This suggests that BNPP inhibits INH-hydrolysis via a non-competitive mechanism. Therefore, the mechanism of inhibition of hepatic INH-amidase by BNPP likely involves phosphorylation of an hydroxyl group at an allosteric site which results in irreversible and non-competitive enzyme inhibition. Although BNPP pretreatment prevented INH-induced toxicity presumably via inhibition of the release of hydrazine from INH, the correlations of hepatic amidase activity with hepatic cell damage, hepatic steatosis and hypertriglyceridemia are weak. The most likely explanation for the poor correlation is that INH administration affected (decreased) hepatic INH-amidase activity. Therefore, hepatic amidase activity after INH 197 administration is not an accurate measure of amidase activity before INH administration and hepatic amidase activity at baseline may be a better predictor of animal susceptibility to INH-induced hepatotoxicity. It is also possible that the lack of correlation between severity of INH-induced hepatotoxicity and hepatic amidase activity is because BNPP protected from INH-induced hepatotoxicity via another mechanism unrelated to, or in addition to, inhibition of INH-amidase activity. However, this study did not identify any mechanistic possibilities other than INH-amidase inhibition. Still, the mechanism of INH-induced hepatotoxicity is clearly a multifactorial one which depends on several different steps and the lack of correlation of toxicity to amidase activity does not disprove its importance in production of the toxicity. Another likely contributor to the poor correlation of severity of INH-induced hepatotoxicity with INH-amidase activity was that the INH-amidase activity measured in this study represents the activity of direct conversion of INH to hydrazine and does not necessarily reflect other amidase pathways from INH which indirectly lead to hydrazine (acetyl-INH to acetylhydrazine and acetylhydrazine to hydrazine). Amidase activity was determined in baseline plasma samples in order to evaluate whether or not amidase activity in plasma is an important pre-disposing factor for INH-induced hepatotoxicity. It could be hypothesized that a relatively low plasma amidase activity at baseline could protect INH from hydrolysis to hydrazine in the plasma and prevent INH-induced hepatotoxicity. This hypothesis is based on the assumption that plasma amidase activity is quantitatively significant in comparison with hepatic amidase activity. The average liver weight in the VEH-VEH group in this study was 82.5 grams 198 and on average there was 94.5 mg protein/g liver. Using these values, total hepatic INH-amidase activity was estimated to range from 1.3 x 10s to 3.9 x 105 nmoles/hour. Since the average blood volume of a rabbit is approximately 65 mL/kg, the average weight of the rabbits in the VEH-VEH group was 2.7 kg and plasma amidase activity averaged 11.9 nmoles/hour/mL plasma, the estimated total INH-amidase activity in plasma is 2.1 x 103 nmoles/hour. Thus, the hepatic INH-amidase is capable of hydrolysing 62 to 126 times more INH to hydrazine than INH-amidase in plasma and is thus quantitatively more important than plasma amidase activity in the direct conversion of INH to hydrazine in vivo. INH administration decreases hepatic INH-amidase activity by approximately 38%. It is possible that plasma amidase activity is also affected during INH administration. It is assumed, although not demonstrated in this study, that INH-amidase in plasma is inhibited by BNPP like hepatic amidase. An estimation of plasma amidase activities at 48 hours in the INH-treated animals would have been valuable in order to assess the effect of INH, and BNPP, on plasma amidase activity; however, the presence of INH and its metabolites (including hydrazine) in the 48 hour plasma samples would have made it difficult to accurately measure hydrazine production after incubation with INH. An unexplained feature of INH-induced hepatotoxicity in humans is that it occurs only after repeated daily doses of INH. Although the doses are higher and more frequent than those given to humans, repeated doses are also an essential feature of this rabbit model (Sarich et al, 1995). This is thought to be due, in part, to the accumulation of hydrazine in plasma, which has been demonstrated to occur in humans (Gent et al, 1992). 199 In the present study, accumulation of hydrazine also occurred. In the VEH-INH group, on day one, three hours after the fourth dose of INH (12 hours), plasma hydrazine levels averaged 25 uM. At 24 hours, plasma levels averaged 12 uM. This drop is due to the time in between doses of INH at the end of day 1 and the beginning of day 2 (the last dose of INH administered prior to this sample was 15 hours previous) (Table 2.11). At 32 hours, two hours after the third (of four) dose of INH on day 2, plasma hydrazine levels averaged 32 pM. Assuming that the plasma hydrazine levels would increase even higher after the fourth and final dose of INH and that only a small amount of hydrazine would be eliminated between the 32 hour sample and projected plasma hydrazine levels at 36 hours, the levels would be predicted to be substantially higher on the second day as compared to the first day. The hypothesis that BNPP pretreatment would decrease the severity of INH-induced hepatic cell damage by decreasing the production of hydrazine from INH in vivo was confirmed. It follows from this hypothesis that plasma hydrazine levels should correlate to hepatic amidase activity and this was also shown to be true. It would then follow that plasma-hydrazine levels and hepatic amidase activity correlate with the severity of hepatic cell damage. However, this was not shown to be the case. Although a general positive correlation did exist, there appears to be a greater variability in the toxic response to hydrazine as compared to the animals in Study 2. There are some possible explanations for this observation. Three animals in the VEH-INH group and one in the BNPP-INH group died prematurely after having convulsions. The last plasma hydrazine levels obtained prior to death in these animals showed a wide range (10 to 74 pM). Similar 200 variability in response to hydrazine appears to be occurring with respect to hepatic cell damage. The early deaths resulted in the loss of valuable data points since the plasma ASAL activities and/or hydrazine levels which were obtained were likely underestimated as compared to the levels immediately prior to death. Conversely, two animals with high plasma hydrazine levels (51 pM and 55 pM) did not experience hepatic cell damage. One possible explanation for this is that the elevated hydrazine levels, which did not occur until 32 hours in these animals, would have been associated with delayed hepatotoxicity and would not have been detected at 48 hours, the last measured time-point in the study. Delayed elevations in plasma ASAL activities beyond 48 hours have been observed in studies lasting 96 hours (Sarich et al, 1995). Nevertheless, hydrazine does play a role in INH-induced hepatic cell damage. Since some animals show severe hepatic cell damage in the presence of moderate plasma hydrazine levels and some show no hepatic cell damage in the presence of relatively high plasma hydrazine levels, it appears that some other factor(s) in addition to high plasma hydrazine levels must be essential for the development of INH-induced hepatic cell damage. These-factors could include an unidentified hepatic enzyme which converts hydrazine into a toxic or non-toxic intermediate and/or variable susceptibility of liver cells to hydrazine as a toxin. Regardless of the mechanism(s) which leads to hydrazine-induced hepatotoxicity, inhibition of the production of hydrazine from INH appears to be an effective way to prevent INH-induced hepatotoxicity in rabbits. Depletion of hepatic glutathione levels can be an indication of the occurrence of toxic processes. In this model of INH-induced hepatotoxicity, it is clear that toxic 201 processes are occurring; however, it has not been previously established whether or not hepatic glutathione depletion occurs as a result of this toxicity. In acetaminophen-induced hepatotoxicity, depletion of glutathione precedes the onset of hepatic cell damage (Mitchell et al, 1973). Early restoration of hepatic glutathione levels by administration of N-acetylcysteine has been shown to be an effective preventative treatment for acetaminophen overdose causing hepatotoxicity. If INH-induced hepatotoxicity was preceded by depletion of hepatic glutathione stores, a similar treatment would presumably be effective as well. However, in the present study, only a 43% decrease from control levels occurred in the INH-only group. A decrease in glutathione levels such as this is usually not associated with hepatotoxicity (Smith and Mtchell, 1989). In fact, greater than 70% depletion of glutathione stores is required for the onset of irreversible liver damage to occur (Smith and Mitchell, 1989). The lack of complete depletion of glutathione in this study is in concordance with the mechanism of toxicity induced by hydrazines. The chemical nature of reactive metabolites derived from hydrazines (organic free radicals) do not characteristically cause sufficient depletion of hepatic glutathione stores to predispose to hepatotoxicity. (Smith and Mitchell, 1989). The 21% decrease in hepatic glutathione stores observed in the food and water deprived group suggests that at least some of the decrease in glutathione levels in the VEH-INH group was due to the decreased intake of food and water. It has been previously shown that fasting of rats decreases hepatic glutathione levels possibly via protein deprivation (Pessayre et al, 1979; reviewed in MandUra/., 1995). 202 No significant correlations were found between glutathione levels and markers of hepatic cell damage or hepatic steatosis. Thus, the relatively minor decrease in hepatic glutathione stores coupled with the lack of correlation of hepatic glutathione with toxicity markers suggests that hepatic glutathione most likely does not play a significant or direct role in protecting against INH-induced hepatotoxicity and that glutathione depletion is not required prior to development of INH-induced hepatic cell damage or hepatic steatosis. The possible role of oxidative tissue damage in this model of INH-induced hepatotoxicity has not been previously explored. An increase in TBARS formation was observed both in the INH-only group (53%), which did experience significant hepatotoxicity, and in the VEH-VEH* group (32%), which did not experience significant hepatotoxicity. The increase in susceptibility of liver tissue to oxidative challenge in the VEH-INH group (the only group with significant hepatotoxicity) suggests that the pathological processes involved in INH-induced hepatotoxicity alter tissue anti-oxidant capacity. However, the increase in susceptibility of the food and water restricted group to oxidative challenge suggests alteration of antioxidant capacity occurred largely as a result of food.and water, restriction. The.degree of.food and water restriction in these two groups is equal since the food and water received by the VEH-VEH* group equals only that ingested by the VEH-INH group. Therefore, the majority of the decrease in antioxidant capacity observed in the VEH-INH group is most likely due to decreased food and water intake in that group. Thus, this degree of increase in TBARS formation does not seem likely to be the result of hepatotoxic processes because no hepatic cell damage or hepatic steatosis occurred in the VEH-VEH* group despite a 32% increase in TBARS 203 formation. In addition, there were no significant correlations between TBARS formation and markers of hepatic cell damage or hepatic steatosis. Microsomal cytochrome P-450 levels were significantly decreased in the VEH-INH group (to 51% of control levels) as compared to all of the other groups. A decrease in P-450 levels in the INH-only group was also observed in Study 3 and has been previously reported to occur in anatoxin BI-induced hepatotoxicity in rabbits (Guerre et al, 1996). Interestingly, the P-450 levels were not decreased in the BNPP-INH group. If BNPP effectively inhibited the hydrolysis of INH, plasma INH and acetyl-INH levels would be expected to be greater in the BNPP-INH group than in the VEH-INH group. If INH and/or acetyl-INH was/were responsible for the decrease in P-450 levels, the P-450 levels would be expected to be decreased to the level in the VEH-INH group or lower. Since this was not the case, an amidase-dependent INH metabolite is more likely the cause of this decrease. The INH-metabolite acetylhydrazine has been shown to dose-dependently decrease hepatic cytochrome P-450 levels in rats to a maximum of approximately 60% (Bahri et al, 1982). Since BNPP would be expected to decrease plasma acetylhydrazine levels, acetylhydrazine is a likely candidate responsible for the decreased P-450 levels in INH-treated groups. However, it is also very possible that hydrazine is involved. Hepatic reductase activity, the focus of Study 3, was not significantly different in any of the five groups in the present study. Even the VEH-INH group, which showed a significant degree of hepatic cell damage, hepatic reductase activity was not affected. In Study 3, pre- and co-treatment with L-thyroxine was associated with increased hepatic reductase activity and a decreased severity of INH-induced hepatotoxicity. However, in 204 the present study, hepatic reductase activity was not altered and did not seem to be an important factor in INH-induced hepatotoxicity. It is possible that this enzyme does play a role in INH-induced hepatotoxicity as suggested in Study 3, but this role was not detected in the present study. The rationale for studying the involvement of CYP1A2 in INH-induced hepatotoxicity is based on previous studies showing that CYP1A2 is involved in the conversion of acetaminophen to reactive metabolites (Thummel et al, 1993). The rationale for studying the involvement of the phenobarbital inducible CYP2B4/5 isozymes in INH-induced hepatotoxicity was based on a previous study in which phenobarbital pretreatment increased the severity of INH-induced hepatic cell damage in rabbits (Sarich et al, 1995). In addition, Grimm et al. (1994) have shown evidence that CYP2B5 is polymorphically expressed in rabbits. A polymorphically expressed CYP isozyme could potentially explain why only some animals show severe INH-induced hepatotoxicity in this rabbit model, as is the situation in humans. Hepatic EROD activity, a measure of CYP1A1/2 activity (Aix et al, 1994; Rey-Grobellet et al, 1996), was significantly decreased in the VEH-INH (79% decreased), BNPP-INH (58% decreased) and food and water deprived VEH-VEH* groups (42% decreased) as compared to controls. Although BNPP itself had no effect, it appears that INH and food and water deprivation both have effects on CYP1A1/2 activity. About 42% of the decrease in the VEH-INH and BNPP-INH groups is due to decreased food and water intake. Although there are no reports of decreased CYP1A1/2 activity due to decreased food and water intake in the literature, there are reports of decreased 205 CYP1A1/2 activity in acetaminophen-induced hepatotoxicity in mice (Snawder et al, 1994) and anatoxin Bl-induced hepatotoxicity in rabbits (Guerre et al, 1996). The decrease in EROD (CYP1A1/2) activity in mice was apparent 4 hours after the administration of acetaminophen, consistent with toxic metabolite-mediated destruction of the enzyme (Snawder et al, 1994). However, the decrease in EROD (CYP1A1/2) activity in anatoxin Bl-induced hepatotoxicity in rabbits was observed after 5 days of treatment with aflatoxin BI and may be related to a decrease in CYP1A1/2 transcription rates (Guerre etal, 1996). In addition to the effect of food and water deprivation on CYP1A1/2 activity, EROD activity in the VEH-INH group was significantly decreased as compared to the BNPP INH group and this group was significantly decreased compared to the VEH-VEH* group (Figure 2.39). A significant negative correlation of CYP1A1/2 activity with INH-induced hepatic cell damage (versus ASAL and ALT activity) in the two groups receiving INH (VEH-INH and BNPP-INH) shows that more enzyme activity is inhibited as the severity of hepatic cell damage increases. It is possible that toxic intermediates are produced by the CYP1A1/2 isozyme which contribute to or cause INH-induced hepatic cell damage and in the process of producing these reactive intermediates, some of the enzymes are damaged and inactivated. A study looking at CYP1A1/2 activity at time points beyond 48 hours is required to determine whether CYP1A1/2 activity is inhibited reversibly or irreversibly. Given the 48 hour time course of the study, it is possible that the significant decrease in CYP1A1/2 activity was related to decreased RNA or protein synthesis. 206 Neither BROD or PROD activities were induced or inhibited in any of the treatment groups. This is somewhat surprising since, based on previous findings that phenobarbital pretreatment increased the severity of INH-induced hepatic cell damage (Sarich et al, 1995), it was thought that the phenobarbital inducible CYP2B4 isozyme may be involved in the toxicity. However, no significant correlations between hepatic BROD or PROD activities in the VEH-INH group and markers of toxicity were observed. One explanation for the increased severity of INH-induced hepatic cell damage after pretreatment with phenobarbital is that the INH-amidase may have been induced since amidase activity in rats has been shown to be induced by phenobarbital (Whitehouse et al, 1983). However, phenobarbital pre-treatment is known to affect a variety of other hepatic enzymes. p-Nitrophenol hydroxylase (a measure of CYP2E1) activity in the present study was significantly decreased in the VEH-INH group after a dose of INH of 1.1 mmoles/kg/day for two days. This effect was also observed in Study 3 and has been previously reported to occur in acetaminophen-induced hepatotoxicity in mice (Snawder etal, 1994). Aniline hydroxylase activity (another measure of CYP2E1 activity) was also decreased in aflatoxin Bl-induced hepatotoxicity in rabbits (Guerre et al, 1996). Activation of alkylhydrazines into free radical intermediates by CYP2E1 has also been suggested as a mechanism of toxicity of hydrazines (Albano et al, 1995). In acetaminophen-induced hepatotoxicity, CYP1A1/2 and CYP2E1 are involved in the production of reactive and toxic intermediates (Raucy et al, 1989; Patten et al, 1993; Thummel et al, 1993) and are also decreased in activity after the development of toxicity 207 in mice (Snawder et al, 1994). Thus, the inhibition of both CYP1A1/2 and CYP2E1 activities in this model of INH-induced hepatotoxicity in rabbits suggests that these P-450 isozymes may be involved in the production of reactive and toxic intermediates from hydrazine or another potential toxicity-related metabolite. In humans, inhibition of hepatic CYP2E1 activity occurs at doses of (approximately) 0.04 mmoles/kg/day for 7 days and induction of CYP2E1 activity is observed 48 hours after the last dose of INH (Zand et al, 1993). One possible explanation for this observation in humans is that during the inhibition of CYP2E1 activity by INH, the enzyme is stabilised from degradation. When INH administration is stopped, the decreased INH concentration unmasks an increased number of active enzyme molecules, resulting in a transient increase in enzyme activity. Unexpectedly, CYP2E1 activity was increased in the BNPP-INH group (67% greater than the controls) but not in the BNPP-VEH group. In the BNPP-INH animals, plasma INH and acetyl-INH levels would be expected to be higher and the amidase-dependent INH-metabolite (isonicotinic acid, acetylhydrazine and/or hydrazine) levels would be expected to be lower than in the VEH-INH group. The increased CYP2E1 activity in the BNPP-INH group suggests the possibility that INH and/or acetyl-INH is an inducer(s) of CYP2E1, while one or more of the amidase-dependent metabolite(s) are inhibitors of CYP2E1. Therefore, in the absence of amidase-dependent metabolites of INH, the predominant action seen may be induction of CYP2E1. In the VEH-INH group, the decrease in P-450 levels is likely at least partially due to the decreased activity of CYP1A1/2 and CYP2E1. In the BNPP-INH group, CYP1A1/2 activity was significantly decreased and CYP2E1 activity was significantly 208 increased, resulting in no net change in P-450 levels. As discussed previously, if the production of reactive and toxic intermediates by these two enzymes results in damage to the enzymes themselves and vital cellular components, then the lack of a decrease in activity of CYP1A1/2 (72% of the decrease was due to food and water restriction) and CYP2E1 which is comparable to that of the VEH-INH group is likely a reflection of the lack of INH-induced hepatotoxicity in the BNPP-INH group. The acetylator status of the animals in Study 4 does not correlate with any marker of toxicity. The significant negative correlation of plasma triglyceride levels with the rate of acetylation of sulfamethazine in Study 1 may have therefore been a chance occurrence. Overall, there was no clear relationship between the rate of acetylation of sulfamethazine and susceptibility to INH-induced hepatic cell damage, hepatic steatosis or hypertriglyceridemia in the animals studied. As discussed in Study 3, hydrazines are known to induce convulsions in mice, rats, rabbits, dogs (Benson et al, 1952) and man (Reid, 1965). It has been proposed that slow acetylator rabbits are more susceptible to INH-induced convulsions than rapid acetylator rabbits (Hein & Weber, 1984). Four out of a total of 84 rabbits in studies 1, 3 and 4 of this thesis which received INH (not including the methimazole pretreated rabbits, of which 6 of 8 died almost immediately after INH administration) were phenotyped as slow acetylators (< 50% sulfamethazine acetylated in 20 minutes). Two of the four rabbits died prematurely (prior to 48 hours) after experiencing convulsions and the other two did not experience convulsions and survived until the end of the study (an incidence of 50%). In comparison, of 80 rapid acetylator rabbits which received INH in this thesis, only 5 died 209 prematurely (an incidence of 6%). This supports previous observations that animals with a slow acetylation rate are more susceptible to death from INH-induced convulsions than rapid acetylators. In conclusion, lNH-administration decreases but does not deplete hepatic glutathione stores. INH-administration also does not greatly affect hepatic antioxidant capacities. In the presence of BNPP, INH does not decrease P-450 levels or /?-nitrophenol hydroxylase (CYP2E1) activity. In fact, CYP2E1 activity was increased in the presence of BNPP and INH. Food and water deprivation and INH administration both decrease the activity of EROD (CYP1A1/2). The decreased activity of CYP1A1/2 and CYP2E1 in the VTiH-TNH group suggests that these enzymes may play a role in the production of reactive and toxic intermediates. The activities of BROD (CYP2B4) and PROD (CYP2B4/5) are not affected by BNPP or INH and do not appear to be involved in INH-induced hepatotoxicity. BNPP is a potent and irreversible inhibitor of INH-amidase. Administration of BNPP 30 minutes prior to INH in this model of INH-induced hepatotoxicity decreases plasma levels of hydrazine and the severity of INH-induced hepatic cell damage, hepatic steatosis and hypertriglyceridemia. INH-amidase activity was found in significant levels in both the plasma and the liver of rabbits. The mechanism of prevention of INH-associated toxicities most likely involves the decreased production of the amidase dependent INH-metabolite hydrazine. 210 FINAL DISCUSSION Interrelationships of INH-Induced Pathological Changes INH-induced hepatotoxicity in humans is characterised by hepatic necrosis/hepatic cell damage (Maddrey and Boitnott, 1973; Black et al, 1975). The only two studies reporting the occurrence of both INH-induced hepatic necrosis and steatosis were in patients receiving both INH and rifampin (Pessayre et al, 1977; Pilheu et al, 1979). Whether INH leads to an increase in liver triglycerides or plasma triglycerides in humans has never been studied. The pathological changes during INH-induced hepatotoxicity in rabbits include hepatic necrosis/hepatic cell damage, hepatic steatosis and hypertriglyceridemia. Although fatty changes including hepatic steatosis and/or hypertriglyceridemia have been previously reported in INH-treated animals (McKennis et al, 1956; Whitehouse et al, 1983; Karthikeyan and Krishnamoorthy, 1991), this rabbit model has provided the first evidence for the simultaneous occurrence of INH-induced hepatic cell damage, hepatic steatosis and hypertriglyceridemia. Effects of Food and Water Deprivation It is known that animals administered INH greatly decrease food and water intake and lose weight (Amenta and Johnston, 1962; Patrick and Back, 1965). Although the reason for this response is unknown, it likely has to do with the toxic effects of INH which include gastric retention, ileus and central nervous system toxicity (resulting in inability to get to food and water). It has been observed that fasting rats for 24 hours (Amenta and Johnston, 1962) and food and water rationing rabbits for 30 days (Whitehouse et al, 1983) does not account for the increased lipid in plasma and liver which occurs in INH-211 treated animals. In Study 4, food and water deprivation resulted in a significant decrease in hepatic glutathione levels and EROD (CYP1A1/2) activity, and a significant increase in hepatic TBARS formation. Since these changes were also observed in the VEH-INH group, the food and water deprived group acted as a valuable control which explained some of these biochemical changes in the VEH-INH group. The Acetylhydrazine Theory The acetylhydrazine theory proposed in the early 1970's was formulated based on experiments which showed that acetylhydrazine (0.2 - 0.7 mmoles/kg; i.p.) and acetyl-INH caused histologically detectable hepatic necrosis in phenobarbital-induced Sprague-Dawley rats (reviewed in Mitchell et al, 1976; Timbrell et al, 1980) and that there was an increased incidence of INH-induced hepatotoxicity in rapid acetylators of TNH (Mitchell et al, 1975b). Subsequently, however, acetylhydrazine (0.2 - 4.2 mmoles/kg; i.p.) has been shown not to cause significant biochemically detectable hepatic cell damage in phenobarbital-induced Sprague-Dawley rats (Wright et al, 1986). In addition, a 1 mmole/kg dose of acetylhydrazine (i.p.) did not cause significant histologically detectable hepatic necrosis in phenobarbital-pretreated Sprague-Dawley rats (Ganley et al, 1994). These results put in doubt the hepatotoxicity of acetylhydrazine. There is also doubt about the proposed increased incidence of INH-induced hepatotoxicity in rapid acetylators of INH. Some studies have reported no difference in the incidence of INH-induced hepatotoxicity between rapid and slow acetylators (Riska, 1976; Singapore, 1977; Hong Kong, 1977) whereas some studies have shown an increased incidence of INH-induced hepatotoxicity in slow acetylators (Musch et al, 212 1982; Eichelbaum et al, 1982). The lack of any further supporting evidence for the acetylhydrazine theory in the literature, the lack of correlation between acetylhydrazine levels and hepatic cell damage in Study 2 and support for the hydrazine theory, both in the literature and in this thesis, put the acetylhydrazine theory in doubt as an explanation for the mechanism of INH-induced hepatotoxicity. Although there is little convincing evidence implicating acetylhydrazine as the INH-derived hepatotoxin, acetylhydrazine may still play a role in INH-induced hepatotoxicity as a source of hydrazine. In fact, inhibition of INH-induced hepatotoxicity by BNPP (Study 4) could be related to inhibition of production of acetylhydrazine from acetyl-INH. Hydrazine in INH-induced Hepatotoxicity The data in this thesis support the theory that hydrazine is a major factor in the susceptibility of rabbits to INH-induced hepatic cell damage. However, plasma hydrazine levels appear to have significant variability in relation to hepatotoxicity and it was not possible to study all of the factors leading to the variability in this thesis. Continued efforts to relate elevations of INH-derived hydrazine in plasma to hepatotoxicity in animals and humans may improve the understanding of this relationship. Understanding and controlling variations in plasma hydrazine levels may be the key to preventing INH-induced hepatotoxicity in animal models and humans. The greater amidase activity in liver versus plasma suggests that the majority of hydrazine produced from INH is formed in the liver. Even though it is not known whether hydrazine itself, or an activated form of hydrazine, is hepatotoxic, it is likely that the mechanism of INH-induced hepatic cell damage involves formation of hydrazine from INH 213 inside liver cells. The inherent reactivity of hydrazine is exemplified by the majority of hydrazine in plasma being present in the form of pyruvate azine (Ellard and Gammon, 1976) and other ct-ketoacid azines (Timbrell and Wright, 1984). A likely scenario is that the majority of hydrazine present in plasma is the result of diffusion of INH-derived hydrazine out of liver cells, before or after the formation of azines. Therefore, the majority of hydrazine present in plasma is a result of formation in and diffusion out of the liver. One way to obtain actual hydrazine levels in the liver would be to determine hydrazine levels in liver homogenate at the time of death. However, the ease of blood sampling and the ability to obtain repeated samples over time make the use of plasma hydrazine determination a useful marker. Still, hydrazine levels determined in liver may correlate closer to the severity of INH-induced hepatotoxicity as compared to plasma hydrazine levels because plasma hydrazine levels are only a function of hydrazine formed in the liver. In addition, an unknown portion of hydrazine in plasma may represent that produced from other organs and/or tissues. Cytochrome P-450 Levels in INH-Induced Hepatotoxicity In INH-treated animals, a 50% decrease in cytochrome P-450 levels is detected both at 15 and 63 hours after the last dose. Primary (or possibly secondary) activated metabolites of INH may irreversibly inactivate P-450 enzymes. Although the mechanism of this inhibition is unknown, acetylhydrazine has been shown to decrease P-450 levels in rat liver by approximately 60% (Bahri et al, 1982). An interesting observation was that 1-thyroxine decreased P-450 levels by 40%. This decrease in P-450 levels may have played a role in the decrease in INH-induced hepatic cell damage and hepatic steatosis which was 214 observed in these rabbits (Study 3). The decrease in P-450 levels in rabbits treated with INH-only in Studies 3 and 4 reflects, at least partially, the reversible or irreversible damage (decreased activity) of EROD (CYP1A1/2) and p-nitrophenol hydroxylase (CYP2E1). Reductase Activity and INH-induced Hepatotoxicity The role of the hepatic reductase enzyme in INH-induced hepatotoxicity was investigated in Study 3. This enzyme had been previously shown to metabolise hydrazine to chemically reactive and potentially toxic intermediates (Noda et al., 1988). It was thought that a correlation of severity of INH-induced hepatotoxicity with reductase activity may offer insight into the mechanism of INH-induced hepatotoxicity. However, no correlation of reductase activity with markers of liver toxicity were observed in Studies 3 or 4. In order to increase the activity of reductase and thereby hopefully increase the severity of INH-induced hepatotoxicity in Study 3, L-thyroxine was administered to the rabbits. L-thyroxine increased hepatic reductase activity but decreased INH-induced hepatic cell damage and hepatic steatosis. It is possible that the reductase enzyme was directly responsible for the decrease in severity of INH-induced hepatotoxicity or that the increased reductase activity enabled an unknown P-450 enzyme to detoxify toxic intermediates at a greater rate. It is also possible that L-thyroxine prevented the toxicity via a reductase independent mechanism, such as decreased P-450 levels or by affecting hydrazine production from INH. 215 EROD (CYP1A1/2) Activity in INH-induced Hepatotoxicity EROD (CYP1A1/2) activity was significantly decreased in food and water deprived animals and in VEH-INH and BNPP-INH groups. It is clear that although food and water deprivation decreased CYP1A1/2 activity, INH administration also decreases it. The mechanism of the decrease is most likely due to either decreased RNA transcription/protein translation or damage/destruction by reactive and toxic intermediates. A short term study (e.g. 12 hours) might answer whether the decrease could be detected after toxic insult, and before regulatory changes were possible, as in a four hour study in mice which showed significantly decreased CYP1A1/2 activity in acetaminophen treated mice (Snawder et al., 1994). If CYP1A1/2 is actually involved in the production of INH-induced hepatotoxicity, a CYP1A1/2 inhibitor may prevent the toxicity or, conversely, a CYP1A1/2 inducer (e.g. cigarette smoke, omeprazole) may increase the severity of the toxicity. p-Nitrophenol Hydroxylase (CYP2E1) Activity in INH-induced Hepatotoxicity /j-Nitrophenol hydroxylase (CYP2E1) activity is significantly decreased by INH administration. In Study 3, the activity of CYP2E1 was inhibited by 38% as compared to controls at 96 hours, 63 hours after the last dose of INH. In Study 4, the activity of CYP2E1 was inhibited by 86% of controls at 48 hours, 15 hours after the last dose of INH. It appears therefore that the levels in Study 3 were increasing back towards baseline levels. In humans, after an initial inhibition phase, an increase of CYP2E1 activity was detected 48 hours after the last dose of INH (Zand et al., 1993). If the situation was similar in both rabbits and humans, the CYP2E1 activity in the INH-treated animals in 216 Study 3 would be expected to be above or at baseline. However, the activity remained inhibited. This may be an indication that a recoverable "inhibition" of CYP2E1 activity occurs after lower doses of INH in humans (0.04 mmoles/kg/day for 7 days) whereas "destruction" of CYP2E1 occurs after higher doses of INH in rabbits (1.1 mmoles/kg/day for two days), requiring re-synthesis of enzyme to re-establish control activities. Also unexpectedly, although INH was present, CYP2E1 activities were not decreased in either the T4-INH group (Study 3) or the BNPP-INH group (Study 4). In fact, CYP2E1 activities in the BNPP-INH group were significantly above control levels. Neither BNPP nor l-thyroxine alone had any effect on CYP2E1 activity. Given that BNPP alters the metabolism of INH by decreasing the production of amidase dependent metabolites, it is likely that either the increased concentration of INH and/or acetyl-INH or the relative absence of INH metabolites in the BNPP-INH group was responsible for the lack of inhibition of CYP2E1 activity. Interestingly, both l-thyroxine and BNPP decreased the severity of INH-induced hepatotoxicity and prevented the inhibition of CYP2E1 activity by INH. However, if CYP2E1 activity is simply a measure of INH-toxicity, any treatment which prevents the toxicity would be expected to prevent the decrease in CYP2E1 activity. As mentioned above, CYP1A1/2 and CYP2E1 activities have been shown to be decreased in activity in relation to acetaminophen-induced hepatotoxicity in mice, indicating a potential role for the production of reactive and toxic intermediates for CYP1A1/2 and CYP2E1 in INH-induced hepatotoxicity (Snawder etal, 1994). 217 It should be noted that the P-450 levels and CYP2E1 activities in the control animals in Studies 3 and 4 are different by a factor of 1.7 and 1.6, respectively. These differences are most likely the result of seasonal variations in enzyme activities. Amidase Activity and INH-Induced Hepatotoxicity Based on the inhibition of INH-induced hepatotoxicity by BNPP, it was expected that INH-amidase activity would correlate with INH-induced hepatotoxicity; however, this was not the case. It was observed that INH-amidase activity after INH administration is not an accurate measure of amidase activity before INH administration because INH administration was found to decrease amidase activity. Therefore, it is possible that hepatic amidase activity at baseline may be a better predictor of animal susceptibility to INH-induced hepatotoxicity. In addition, the INH-amidase activity measured in this study represents only the activity of direct conversion of INH to hydrazine and does not consider other amidase pathways from INH which indirectly lead to hydrazine (acetyl-INH to acetylhydrazine and acetylhydrazine to hydrazine). The reported amidase activity also does not take into account the competing activity of N-acetyltransferase. The mechanism of INH-induced hepatotoxicity is clearly multifactorial and depends on several distinct steps. Some of these steps are known. Clearly, acetylation and amidase activity are important; however, there are a number of different substrates for both of these enzymes and the steps leading to production of hydrazine are not entirely predictable. For example, a high amidase activity would produce a large amount of free amino groups (hydrazine and acetylhydrazine) but in the presence of a high acetylation rate, the free amino groups would be quickly detoxified by acetylation to 218 diacetylhydrazine. Whitehouse and colleagues (1983) showed that acetyl-INH is a better substrate for amidase than is INH itself. Therefore, the right balance of acetylation and amidase activity may be necessary to produce the maximal amount of hydrazine. For example, since hepatic amidase catalysed hydrolysis of INH and acetyl-INH is 10 to 20 times lower while acetylation rate is only 6 times lower in rats versus rabbits (Whitehouse et al, 1983), a greater proportion of acetylated metabolites over hydrazine metabolites would be expected. According to the hypothesis that hydrazine is an INH-derived hepatotoxin and acetylation is an act of detoxification, a greater proportion of acetylation versus hydrolysis could explain the relative resistance of rats to INH-induced hepatotoxicity (Wright et al, 1986). One of the unanswered aspects of this toxicity is whether hydrazine, assuming for the moment that hydrazine is the INH-derived hepatotoxin, or an activated form of hydrazine is the toxicity producing compound. An interesting observation in Study 4 is that while the activity of some microsomal enzymes are unaltered by INH (reductase, BROD, PROD), amidase activity is decreased by approximately 38%, EROD activity is decreased by approximately 78% and p-nitrophenol hydroxylase activity is decreased by approximately 86%. Could a common link between these enzymes be that they all interact in some way with hydrazine? We know that INH-amidase cleaves hydrazine from INH and it appears that some amidase becomes inactivated in the process. Thus, like INH-amidase, the decreased activities of CYP1A1/2 and CYP2E1 may be due to their affinity for hydrazine. It seems likely that these P-450s activate hydrazine to a damaging intermediate which results in damage to the producing enzyme and surrounding intracellular 219 components. As has been proposed for acetaminophen and its more reactive analogue N-acetyl-7w-aminophenol, if a metabolic intermediate is too reactive, as soon as it is produced it reacts with the producing enzyme as a suicide inactivator and the production of reactive metabolites and toxicity becomes self-limiting (Pumford and Halmes, 1997). In this situation, no toxicity would occur to the cell other than to the producing enzyme since the reactive metabolites would immediately bind to the enzyme before causing damage away from the producing enzyme. This reaction, and its inherent potential for causing toxicity, is thus self-limiting. Therefore, if hydrazine is converted to a reactive intermediate, it appears that it is reactive enough to cause damage to vital intercellular components and cell death, but not so reactive that some hydrazine does not get away from the producing enzyme(s). The lack of correlation of the phenobarbital-inducible BROD and PROD activities with hepatic cell damage was somewhat surprising since phenobarbital has been shown to increase the severity of INH-induced hepatotoxicity in rabbits (Sarich et al, 1995). However, in light of the evidence for a role of INH-amidase in this model of INH-induced hepatotoxicity in rabbits, it may be that the potentiation of INH-induced hepatic cell damage by phenobarbital is related to the ability of phenobarbital to increase the activity of INH-amidase (Whitehouse etal, 1983). The Rabbit Model in Relation to Humans This rabbit model of INH-induced hepatotoxicity has several similarities to the toxicity in humans. Histological characteristics of INH-induced hepatotoxicity in humans range from patchy focal necrosis to multilobular, bridging and massive necrosis (Maddrey and Boitnott, 1973) as well as centrilobular necrosis and inflammatory infiltration (Black 220 et al., 1975). Similarity, in the rabbit model, INH causes focal and centrilobular necrosis and inflammatory infiltration (Sarich et al., 1995). In addition, hepatocellular fatty vacuolization is observed in INH-induced hepatotoxicity in rabbits (Sarich et al., 1995). Fatty accumulation was reported to occur in humans in a study in which liver biopsies from 5 out of 5 patients with elevated plasma liver enzymes demonstrated features of fatty vacuolization (hepatic steatosis) after receiving INH and rifampin (Pessayre et al., 1977). Biochemical changes, or increases in plasma liver enzyme activities, occur in INH-induced hepatotoxicity in humans (Barlow et al., 1974; Mitchell et al., 1975a; Black et al., 1975) and rabbits (Sarich et al., 1995; Sarich et al., 1996; Sarich et al., 1997). Thus, there are close similarities in the histological and biochemical characteristics of INH-induced hepatotoxicity in humans and rabbits. There are no reports of INH-induced hepatotoxicity after single doses of INH in humans. In mice, rats, rabbits and dogs, large single doses of INH do not cause hepatotoxicity but do result in convulsions (Benson et al., 1952). In animals and humans, INH-induced hepatotoxicity occurs only after repeated doses of INH. This delay in onset which is common to rabbits and humans suggests a lag-time for the toxicity which may be related to the accumulation of toxic metabolites of INH. Another interesting comparison between INH-induced hepatotoxicity in animals and humans is the frequency of occurrence of the toxicity. In humans, it is estimated that elevations of plasma liver enzyme activities occur in 10-20% of patients receiving INH (Mitchell et al., 1975a) and severe hepatotoxicity occurs in 1-2% of patients (Barlow et al., 1974). In many of the 10-20% of patients with elevations of plasma liver enzymes, the 221 patients often have no symptoms. In the rabbits in this thesis which only received INH, 34 out of 45 (76%) had significant elevations of plasma liver enzymes (greater than 3 times baseline ASAL activity) and 4 out of 45 (9%) had severe plasma lever enzyme elevations (greater than 100 times baseline ASAL activity). Although the frequency of occurrence of INH-induced hepatotoxicity is greater in the rabbit model, the dose of INH administered is approximately 10 times greater. Even at these high doses, as is the case in humans, not all rabbits exposed to INH have INH-induced hepatotoxicity. As can occur in humans, in the majority of rabbits, elevated plasma liver enzymes return to baseline after INH administration is stopped. This variability which occurs in both humans and rabbits may be, but is not necessarily, related to a similar mechanism of toxicity. Although several studies have reported the accumulation of plasma and liver lipids in rabbits after exposure to INH (McKennis et al., 1956; Whitehouse et al., 1983; Karthikeyan and Krishnamoorthy, 1991), only one study has reported the occurrence of INH-induced hepatic steatosis in humans (Pessayre et al., 1977). However, the patients in this report were receiving rifampin as well as INH (Pessayre et al., 1977). Although fatty accumulation may appear to be a feature of INH-induced hepatotoxicity specific to rabbits, whether INH causes any elevations of plasma or liver lipids in humans has not been properly studied. Although age is an important risk factor for INH-induced hepatotoxicity in humans (Barlow et al., 1974), this aspect of the toxicity has not been studied in an animal model. The influence of gender in INH-induced hepatotoxicity in humans has been debated. Although there has never been a mechanism proposed for it, some evidence 222 suggests that females are at a higher risk of experiencing INH-induced hepatotoxicity than males (Moulding et al, 1989; Snider and Caras, 1992). This was not found in this model of INH-induced hepatotoxicity in rabbits as we discovered that both male and female rabbits are susceptible to a similar degree to both INH-induced hepatic cell damage and hepatic steatosis. One possibility for the suggested increased incidence in females in humans is that, unlike in our animal model where the animals are given INH on a mg/kg basis, a standard dose of 300 mg per day is given to both males and females. Since females weigh less on average than males, it is possible that in humans the apparent increase in susceptibility of females is because they receive a greater dose of INH on a mg/kg basis. This is in concordance with the findings of this thesis and studies in humans, because accumulation of hydrazine appears to be a critical factor and accumulation would be greater with higher doses. Over the years, a controversy concerning acetylation rate as a predictor of risk of INH-induced hepatotoxicity in humans has existed in the literature. Studies in this thesis could not fully address this question, given that the rabbits studied were mostly rapid acetylators.. However, in this population of rapid acetylator rabbits, there was no indication of a linear relationship between the acetylation rate of sulfamethazine, a marker for acetylation of INH, and susceptibility to hepatic cell damage. A general linear model, or multiple regression (used to examine the combined role of acetylation rate of sulfamethazine in 20 minutes and gender versus hepatotoxicity), did not improve the relationship over the two individual parameters. In Study 4, using both amidase activity and acetylation rate versus peak plasma ASAL activity (r = 0.61, n = 15, 223 p < 0.02) improved the correlation over the individual correlations of peak plasma ASAL activity with hepatic amidase activity (r = 0.56, n = 15, p < 0.05) and acetylation rate (r = 0.22, n = 15, p >0.05). However, the largest portion of the variance in plasma ASAL activity was explained by hepatic amidase activity as compared to acetylation rate. In a larger study of animals treated only with INH, and including male and female rabbits with a range of fast and slow acetylation rates, multiple regression of gender, acetylation rate, hepatic amidase activity and possibly baseline CYP2E1 and/or CYP1A1/2 activities with plasma ASAL activity or hepatic triglyceride accumulation may together explain a large part of the variance in INH-induced hepatotoxicity and provide much needed insight into the role of all of these factors in the mechanism of INH-induced hepatotoxicity. There is abundant evidence that hydrazine is a direct hepatotoxin in animals (Yard and McKennis, 1955; McKennis etal, 1956; Patrick and Back, 1965; Scales and Timbrell, 1982; Timbrell et al, 1982; Noda et al, 1983; Jenner and Timbrell, 1994a; Jenner and Timbrell, 1994b). The first studies to suggest that hydrazine is involved in INH-induced hepatotoxicity in humans came in the mid 1980's (Blair et al, 1985; Sarma et al, 1986). Since then, hydrazine has been increasingly implicated and studied as the potential INH-derived hepatotoxin in INH-induced hepatotoxicity. A direct link between high plasma hydrazine levels and fatal INH-induced hepatotoxicity was proposed in a case report of a 74-year old man with severe INH-induced hepatic necrosis (Woo et al, 1992). In another study, incomplete elimination of hydrazine at 24 hours has been shown to contribute to a slowly rising baseline of hydrazine in plasma and elevation of markers of hepatotoxicity were observed in an individual with the highest accumulation of hydrazine (Gent et al, 224 1992). Another study showed that plasma hydrazine levels increased significantly between weeks 1 and 3 of therapy and then slowly returned to week 1 levels at week 26 (Donald et al, 1994). Thus, in rabbits (Sarich et al., 1996) and humans (Woo et al., 1992; Gent et al., 1992, Donald et al, 1994), the accumulation of hydrazine over time may play a role in INH-induced hepatotoxicity. This pattern of accumulation of plasma hydrazine levels fits well with the onset of INH-induced hepatotoxicity in humans given that the incidence of INH-induced hepatotoxicity is characteristically delayed for at least 3-5 days and peak incidence usually occurs in the first month. The recognition of hydrazine as an important factor in INH-induced hepatotoxicity in this rabbit model and humans is relatively recent. The establishment and use of this rabbit model allows theories on the mechanism of INH-induced hepatotoxicity to be tested in a controlled way. Information gained from this rabbit model can lead to hypotheses which can be tested in humans. For example, hydrazine concentrations could be monitored in humans receiving INH to determine whether or not they could be used to predict individuals who are at the greatest risk of developing hepatotoxicity. Further studies in rabbits and humans are essential to determine all of the factors which influence the risk of INH-induced hepatotoxicity in humans, and ultimately to completely understand the mechanism of INH-induced hepatotoxicity. 225 THESIS CONCLUSIONS 1. Three major pathological changes which occur in INH-induced hepatotoxicity in rabbits are hepatic cell damage, hepatic steatosis and hypertriglyceridemia. 2. Both male and female rabbits are susceptible, and to a similar degree, to INH-induced hepatic cell damage, hepatic steatosis and hypertriglyceridemia. 3. Plasma levels of hydrazine, but not INH or acetylhydrazine, correlate with the severity of INH-induced hepatic cell damage in rabbits. 4. L-Thyroxine administration significantly increases hepatic reductase activity in rabbits. 5. Methimazole administration significantly decreases hepatic reductase activity in rabbits. 6. Pre- and co-treatment of rabbits with L-thyroxine decreases the severity of INH-induced hepatic cell damage and prevents INH-induced hepatic steatosis. 7. A decrease, but not a depletion, of hepatic glutathione occurs as a result of INH administration. 8. A biologically significant alteration of hepatic anti-oxidant capacity does not occur as a result of INH administration. 9. A 50% decrease in hepatic cytochrome P-450 levels observed 15 hours after INH administration does not show any recovery 63 hours after the last dose of INH. 10. EROD (CYP1A1/2) activity is inhibited by food and water deprivation and to an even greater extent by INH administration. CYP1A1/2 may play a role in the production of reactive and toxic intermediates in INH-induced hepatotoxicity. 226 11. The activities of hepatic BROD (CYP2B4) and PROD (CYP2B4/5) are not altered by the administration of INH and/or BNPP and these enzymes do not appear to play a role in INH-induced hepatotoxicity. 12. /?-Nitrophenol hydroxylase (CYP2E1) activity is inhibited by 86% at 15 hours and by 38% at 63 hours after the last dose of INH. This inhibition is prevented when TNH is given with L-thyroxine or BNPP. CYP2E1 may play a role in the production of reactive and toxic intermediates in INH-induced hepatotoxicity. 13. In a population of rapid acetylator rabbits, the acetylation rate of sulfamethazine does not appear to be a critical determinant in the susceptibility to INH-induced hepatotoxicity. 14. BNPP is an irreversibly acting non-competitive inhibitor of rabbit hepatic INH-amidase. 15. INH-amidase activity is present in both plasma and liver; however, there is more enzyme, and thus more activity, in the liver. 16. Pretreatment with BNPP significantly decreases the severity of INH-induced hepatic cell damage, hepatic steatosis and. hypertriglyceridemia..as well as plasma.hydrazine levels at 12 at 32 hours. 17. The mechanism of protection from INH-induced hepatotoxicity by BNPP is likely due to a decrease in formation of hydrazine. 18. Hydrazine is most likely the main causative agent in isoniazid-induced hepatotoxicity in rabbits. The results in this animal model are in accord with recent data on isoniazid-hepatotoxicity in humans. 227 REFERENCES Aix L., Rey-Grobellet X., Larrieu G., Lesca P. & Galtier P. (1994) Thiabendazole is an inducer of cytochrome P4501A1 in cultured rabbit hepatocytes. Biochem. Biophys. Res. Commun., 202, 1483-1489. Albano E., Comoglio A., Clot P., Iannone A., Tomasi A. & Ingelman-Sundberg M. (1995) Activation of alkylhydrazines to free radical intermediates by ethanol-inducible cytochromeP-4502E1 (CYP2E1). Biochem. Biophys. Acta., 1243, 414-420. Amenta J.S. & Johnston E.H. (1962) Hydrazine-induced alterations in rat liver. Lab. Invest., 11, 956-962. American Thoracic Society (1965) Untoward reaction to drugs used in the initial chemotherapy of cavitary tuberculosis. Am. Rev. Respir. Dis., 91, 293-296. American Thoracic Society (1967) Chemoprophylaxis for the prevention of tuberculosis. Am. Rev. Respir. Dis., 96, 558-560. Assem E.S.K., Ndoping N., Nicholson H. & Wade J.R. (1969) Liver damage and isoniazid allergy. Clin. Exp. Immunol., 5, 439-442. Bahri A.K., Chiang C.S. & Timbrell J.A. (1981) Acetylhydrazine hepatotoxicity. Toxicol. Appl. Pharmacol., 60, 561-569. Bahri A.K., Chiang C.S. & Timbrell J.A. (1982) Factors affecting acetylhydrazine hepatotoxicity. Adv. Exp. Med. Biol, 136(B), 1055-1065. Banerjee A , Dubnau E., Quemard A., Balsubramanian V., Um K.S., Wilson T., Collins D., de Lisle G. & Jacobs W.R. (1994) inhA, a gene encoding a target for isoniazid and ethionamide in Mycobacterium tuberculosis. Science, 263, 227-230. Barlow, P.B., Black, M., Brummer, D.L., Comstock, G.W., Dubin, I.N., Enterline, P., Gibson, M.L., Hardy, G.E. Jr., Harrel, J.A., Johnston, R.F., Kent, D.C., Marvin, B.A., McCaig, N.C., Mitchell, J.R., Mosley, J.W., Ogasawara, F.R., Popper, H., Reichman, L.B., & Zimmerman, H.J. (1974) Preventive therapy of tuberculous infection. Am. Rev. Resp. Dis., 110,371-374. Beever I.W., Blair LA. & Brodie M.J. (1982) Circulating hydrazine during treatment with isoniazid rifampicin in man. Br. J. Clin. Pharmacol, 13, 599P. Benson W.M., Stefko P.L. & Roe M.D. (1952) Pharmacologic and toxicologic observations on hydrazine derivatives of isonicotinic acid. Am. Rev. Tuberc, 65, 376-391. 228 Bernstein J., Lott W.A., Steinberg B.A. & Yale H.L. (1952) Chemotherapy of experimental tuberculosis. Am. Rev. Tuberc, 65, 357-364. Biehl P.J. & Nimitz H.J. (1954) Studies on the use of a high dose of isoniazid. Am. Rev. Tuberc, 7 0 , 430-441. Black M., Mitchell J.R., Zimmerman H.J., Ishak K G . & Epler G.R. (1975) Isoniazid-associated hepatitis in 114 patients. Gastroenterology., 69, 289-302. Blair I.A., Mansilla Tinoco R, Brodie M.J., Clare R.A., Dollery C.T., Timbrell J.A. & Beever LA. (1985) Plasma hydrazine concentrations in man after isoniazid and hydralazine administration. Human. Toxicol, 4 , 195-202. Blum M., Grant D.M., Demierre A. & Meyer U.A (1989) iV-acetylation pharmacogenetics: A gene deletion causes absence of arylamine iV-acetyltransferase in liver of slow acetylator rabbits. Proc. Natl. Acad. Sci. USA., 86, 9554-9557. Bozek C.B., Kastrukoff L.F., Wright J.M., Perry T.L. & Larsen T.A. (1987) A controlled trial of isoniazid therapy for action tremor in multiple sclerosis. J. Neurol, 2 3 4 , 36-39. Bradford M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye-binding. Ann. Biochem., 7 2 , 248-254. Brennan R.W., Dehejia H., Kutt H., Verebely K. & McDowell F. (1970) Diphenylhydantoin intoxication attendant to slow inactivation of isoniazid. Neurology., 2 0 , 687-693. Burke M.D., Thompson S., Elcombe C.R., Halpert J., Haaparanta T. & Mayer R.T. (1985) Ethoxy-, pentoxy- and benzyloxyphenoxazones and homologues: A series of substrates to distinguish between different induced cytochromes P-450. Biochem. Pharmacol, 3 4 , 3337-3345. Byrd R.B., Nelson R. & Elliot R.C. (1972) Isoniazid toxicity - A prospective study in secondary chemoprophylaxis. J.A.M.A., 2 2 0 , 1471-1473. Campanini R.Z., Tapia R.A., Sarnat W. & Natelson S. (1970) Evaluation of serum argininosuccinate lyase (ASAL) concentrations as an index to parenchymal liver disease. Clin. Chem., 16, 44-53. Clark D.A., Leeder L.G., Foulds E.L. & Trout D.L. (1970) Changes in lipids of rat liver after hydrazine injection. Biochem. Pharmacol, 19, 1743-1752. Cohen R., Kaiser M.H. & Thomson R.V. (1961) Fatal hepatic necrosis secondary to isoniazid therapy. J.AM.A., 116, 877-879. 229 Cornish H.H. (1969) The role of vitamin B 6 in the toxicity of hydrazines. Ann. N. Y. Acad. Sci., 166, 136-145. Crippin J.S. (1993) Acetaminophen hepatotoxicity: Potentiation by isoniazid. Am. J. Gastroenterol., 88, 590-592. Cross F.S., Long M.W., Banner A.S. & Snider D.E. Jr. (1980) Rifampin-isoniazid therapy of alcoholic and nonalcoholic tuberculous patients in a U.S. Public Health Service cooperative therapy trial. Am. Rev. Respir. Dis., 122, 349-353. Davidson L.A. & Takayama K. (1979) Isoniazid inhibition of the synthesis of monounsaturated long-chain fatty acids in Mycobacterium tuberculosis H37Ra. Antimicrob. Agents. Chemother. 16, 104-105. Dickinson D.S., Bailey W.C., Hirschowitz B.I. & Eidus L. (1977) The effect of acetylation status on isoniazid (INH) hepatitis. Am. Rev. Respir. Dis., 115, 395. Dickinson D.S., Bailey W.C., Hirschowitz B.I., Soong S-I, Eidus L. & Hodgkin M.M. (1981) Risk factors for isoniazid (INH)-induced liver dysfunction. J. Clin. Gastroenterol. 3,271-279. Donald P.R., Seifart HI , Parkin D.P. & van Jaarsveld P.P. (1994) Hydrazine production in children receiving isoniazid for the treatment of tuberculous meningitis. Ann. Pharmacother., 28, 1340-1343. Eichelbaum M., Musch E., Castro-Parra M. & von Sassen W. (1982) Isoniazid hepatotoxicity in relation to acetylator phenotype and isoniazid metabolism. Br. J. Clin. Pharmacol, 14, 575-576P. Ellard G.A. & Gammon P.T. (1976) Pharmacokinetics of isoniazid metabolism in man. J. Pharmacokinet. Biopharmaceut., 4, 83-113. Ellman G.L. (1959) Tissue sulfhydryl groups. Arch. Biochem. Biophys., 82, 70-77. Epstein M.M., Nelson S.D., Slattery J.T., Kalhorn T.F., Wall RA. & Wright J.M. (1991) Inhibition of the metabolism of paracetamol by isoniazid. Br. J. Clin. Pharmacol, 31, 139-142. Fischer C. & Klotz U. (1978) Determination of sulfapyridine and its major metabolites in plasma by high-pressure liquid chromatography. J. Chromatogr. 146, 157-162. Folch J., Ascoli I., Lees M., Meath J.A. & LeBaron F.N. (1951) Preparation of lipide extracts from brain tissue. J. Biol. Chem. 191, 833-841. French S.W., Wong K., Jui L., Albano E., Hagbjork A.-L. & Ingelman-Sundberg M., (1993) Effect of ethanol on cytochrome P450 2E1 (CYP2E1), lipid peroxidation, and 230 serum protein adduct formation in relation to liver pathology pathogenesis. Exp. Mol. Pathol, 58, 61-75. Frymoyer J.W. & Jacox R.F. (1963) Investigation of the genetic control of sulfadiazine and isoniazid metabolism in the rabbit. J. Lab. Clin. Med. 62, 891-904. Ganley C.J., Nguyen H.T. & Reidenberg M.M. (1994) The effect of vitamin B6 deficiency on acetylhydrazine hepatic necrosis. Pharmacol. Toxicol, 74, 303-304. Garibaldi R.A., Drusin R.E., Ferebee S.H. & Gregg M.B. (1972) Isoniazid-associated hepatitis. Am. Rev. Respir. Dis., 106, 357-365. Gellis S.N. & Murphy RV. (1955) Hepatitis following isoniazid. Dis. Chest. 28, 462-464. Gent W.L., Seifart HI., Parkin D.P., Donald P.R. & Lamprecht J.H. (1992) Factors in hydrazine formation from isoniazid by paediatric and adult tuberculous patients. Eur. J. Clin. Pharmacol, 43, 131-136. Godin D.V., Ko K.M. & Garnett M.E. (1989) Altered antioxidant status in the ischemic/reperfused rabbit myocardium: effects of allopurinol. Can. J. Cardiol, 5, 365-371. Gordon G.R., Shafizadeh AG. & Peters J.H. (1973) Polymorphic acetylation of drugs in rabbits. Xenobiotica., 3, 133-150. Grant D.M., Morike K., Eichelbaum M. & Meyer U.A. (1990) Acetylation pharmacogenetics: The slow acetylator phenotype is caused by decreased or absent arylamine iV-acetyltransferase in human liver. J. Clin. Invest., 85, 968-972. Grimm S.W., DyrofFM.C, Philpot R.M. & Halpert J.R. (1994) Catalytic selectivity and mechanism-based inactivation of stably expressed and hepatic cytochromes P450 2B4 and 2B5: Implications of the cytochrome P450 2B5 polymorphism. Mol. Pharmacol, 46, 1090-1099. Gronhagen-Riska C , Hellstrom P.-E. & Froseth B. (1978) Predisposing factors in hepatitis induced by isoniazid-rifampin treatment of tuberculosis. Am. Rev. Respir. Dis., 118, 461-466. Guerre P., Eeckhoutte C , Larrieu G., Burgat V. & Galtier P. (1996) Dose-related effect of aflatoxin Bl on liver drug metabolizing enzymes in rabbit. Toxicology, 108, 39-48. Haber E. & Osborne R.K. (1959) Icterus and febrile reactions in response to isonicotinic acid hydrazine. N. Engl. J. Med, 260, 417-420. Heisey G.B., Hughes H.C., Lang CM. & Rozmiarek H. (1980) The guinea pig as a model for isoniazid-induced reactions. Lab. Anim. Sci., 30, 42-50. 231 Hein D.W. & Weber W.W. (1984) Relationship between N-acetylator phenotype and susceptibility toward hydrazine-induced lethal central nervous system toxicity in the rabbit. J. Pharmacol. Exp. Ther., 228, 588-592. Heymann E. & Krisch K. (1967) Phosphorsaure-bis-(/?-nitrophenylester), ein neuer hernmstoffmikrosomaler carboxylesterasen. Z. Physiol. Chem., 348, 609-619. Heymann E. & Krisch K. (1969) Inhibition of phenacetin- and acetanilide-induced methemoglobinemia in the rat by the carboxylesterase inhibitor bis-(p-nitrophenyl) phosphate. Biochem. Pharmacol., 18, 801-811. Holtz P. & Palm D. (1964) Pharmacological aspects of vitamin B 6 . Pharmacol Rev., 16, 113-178. Hong Kong Chest Service/British Medical Research Council. (1977) Controlled trial of 6-month and 9-month regimens of daily and intermittent streptomycin plus isoniazid plus pyrazinamide for pulmonary tuberculosis in Hong Kong. Am. Rev. Respir. Dis., 115, 727-735. Hong Kong Chest Service/Tuberculosis Research Centre, Madras/British Medical Research Council. (1992) A double-blind placebo-controlled trial of three antituberculosis chemoprophylaxis regimens in patients with silicosis in Hong Kong. Am. Rev. Respir. Dis., 145, 36-47. Ingbar S.H. (1975) Drug-induced alterations in the hepatic metabolism of hormones. In Drugs and the Liver, ed W. Gerok and K. Sickinger. F.K. Schattauer Verlag, New York, pp. 213-219. Jenner A.M. & Timbrell J.A. (1994a) Influence of inducers and inhibitors of cytochrome P450 on the hepatotoxicity of hydrazine in vivo. Arch. Toxicol., 68, 349-357. Jenner A.M. & Timbrell J.A. (1994b) Effect of acute and repeated, exposure to low doses of hydrazine on hepatic microsomal enzymes and biochemical parameters in vivo. Arch. Toxicol, 68, 240-245. Junge W. & Krisch K. (1975) The carboxylesterases/amidases of mammalian liver and their possible significance. CRC CritRev. Toxicol, 3, 371-434. Karthikeyan S. & Krishnamoorthy M.S. (1991) Effect of subacute administration of isoniazid and pyridoxine on lipids in plasma, liver, and adipose tissues in the rabbit. Drug. Chem. Toxicol, 14, 293-303. Kato R. & Takahashi A. (1968) Thyroid hormone and activities of drug metabolizing enzymes and electron transport systems of rat liver microsomes. Mol. Pharmacol, 4, 109-20. 232 Koltunov V.S., Tikonov M.F. & Kuperman A.Y. (1976) Molybdenum-catalysed oxidation of hydrazine by perchloric acid. Russ. J. Phys. Chem., 50, 1702-1704. Koop D.R. (1986) Hydroxylation of/7-nitrophenol by rabbit ethanol-inducible cyp P-450 isozyme 3a. Mol. Pharmacol. 29, 399-404. Koop D.R., Laethem C.L. & Tierney D J . (1989) The utility of /j-nitrophenol hydroxylation inP450IIEl analysis. Drug. Metab. Rev., 20, 541-551. Kopanoff D.E. , Snider D.E. Jr. & Caras G.J. (1978) Isoniazid-related hepatitis. Am. Rev. Respir. Dis., 117, 991-1001. Krishnamoorthy M.S. & Karthikeyan S. (1991) Effect of phenobarbitone pretreatment on isoniazid-induced lipid changes in plasma, liver and adipose tissue in the rabbit. Pharmacol. Res., 24, 219-225. Lai S., Singhal S.N., Burley D.M. & Crossley G. (1972) Effect of rifampicin and isoniazid on liver function. Br. Med. J., 1, 148-150. Lamb R.G. & Banks W.L. (1979) Effect of hydrazine exposure on hepatic triacylglycerol biosynthesis. Biochim. Biophys. Acta., 574, 440-447. Lauterburg B.H., Vaishnav Y. & Mitchell J.R. (1979) Rational approach to the prevention of isoniazid hepatitis. Clin. Res., 27, 235(A). Lauterburg B.H., Smith C.V., Todd E.L. & Mitchell J.R. (1985) Oxidation of hydrazine metabolites formed from isoniazid. Clin. Pharmacol. Ther., 38, 566-571. Littley M.D., Hutchings A., Spragg B.P., Routledge P.A. & Lazarus J.H. (1988) The effect of thyrotoxicosis on isoniazid acetylation. Br. J. Clin. Pharmac, 26, 103-106. McKennis H. Jr., Yard A.S. & Pahnelas E.V. (1956) The production of fatty livers in rabbits by isoniazid and other hydrazine derivatives. Am. Rev. Tuberc, 73, 956-959. McQueen C.A., Maslansky C.J., Glowinski I.B., Crescenzi S.B., Weber W.W. & Williams G.M. (1982) Relationship between the genetically determined acetylator phenotype and D N A damage induced by hydralazine and 2-aminofluorine in cultured rabbits hepatocytes. Proc Natl. Acad. Sci. USA., 79, 1269-1272. Maddrey W.C. & Boitnott J.K. (1973) Isoniazid hepatitis. Ann. Intern. Med., 79, 1-12. Malone H.E. (1970) The determination of hydrazino-hydrazide groups, p. 289, Pergammon Press, Toronto. 233 Mandl J., Banhegyi G., Kalapos M.P. & Garzo T. (1995) Increased oxidation and decreased conjugation of drugs in the liver caused by starvation. Altered metabolism of certain aromatic compounds and acetone. Chem. Biol. Interact, 96, 87-101. Marrie T.J. (1995) Infectious Diseases: Tuberculosis. In Therapeutic Choices. Gray J. Ed. Canadian Pharmaceutical Association, pp. 669-677. Martin C.E. & Arthaud J.B. (1970) Hepatitis after isoniazid administration. N. Engl. J. Med, 282, 433-434. Merritt AD. & Fetter B.F. (1959) Toxic hepatic necrosis (hepatitis) due to isoniazid: Report of a case with cirrhosis and death due to hemorrhage from esophageal varices. Ann. Intern. Med., 50, 804-810. Mitchell J.R., Jollow D.J., Potter W.Z., Gillette JR. & Brodie BB. (1973) Acetaminophen-induced hepatic necrosis. IV. Protective role of glutathione. J. Pharmacol. Exp. Ther., 187, 211-217. Mitchell J.R., Long M.W., Thorgeirsson U.P., & Jollow D.J. (1975a) Acetylation rates and monthly liver function tests during one year of isoniazid preventive therapy. Chest., 68, 181-190. Mitchell J.R., Thorgeirsson U.P., Black M., Timbrell J.A., Snodgrass W.R, Potter W.Z., Jollow D.J., & Keiser HR. (1975b) Increased incidence of isoniazid hepatitis in rapid acetylators: possible relation to hydrazine metabolites. Clin. Pharmacol. Ther., 18, 70-79. Mtchell J.R., Zimmerman H.J., Ishak K.G., Thorgeirsson UP., Timbrell J.A., Snodgrass W.R. & Nelson S.D. (1976). Isoniazid liver injury: clinical spectrum, pathology and probable pathogenesis. Ann. Intern. Med., 84, 181-192. Morel F., Beaune P.H., Ratanasavanh D., Flinois J.-P., Yang C.S., Guengerich F.P. & Guillouzo A. (1990) Expression of cytochrome P-450 enzymes in cultured human hepatocytes. Eur. J. Biochem., 191, 437-444. Moron M.S., DePierre J.W. & Mannervik B. (1979) Levels of glutathione, glutathione reductase and glutathione -^transferase activities in rat lung and liver. Biochim. Biophys. Acta., 582, 67-78. Moulding T.S., Redeker AG., & Kanel G.C. (1989) Twenty isoniazid-associated deaths in one state. Am. Rev. Respir. Dis., 140, 700-705. Murray C.J.L., Styblo K. & Rouillon A. (1990) Tuberculosis in developing countries: burden, intervention and cost. Bull. Int. Union. Tuberc. Lung. Dis., 65, 6-24. 234 Musch E., Eichelbaum M., Wang J.K., v. Sassen W., Castro-Parra M., & Dengler H.J. (1982) Incidence of hepatotoxic side effects during antituberculous therapy (INH, RMP, EMB) in relation to acetylator phenotype. Klin. Wochenschr. 60, 513-519. Neill P., Pringle D., Mhonda M., Kusema T., & Nhachi C.F.B. (1990) Effects of two pulmonary tuberculosis drug treatments and acetylator status in a Zimbabwean population. Cent. Afr. J. Med., 36. 104-107. Nelson S.D., Snodgrass W.R. & Mitchell J.R. (1975) Metabolic activation of hydrazines to reactive intermediates: Mechanistic implications for isoniazid and iproniazid hepatitis. Fed. Am. Soc. Exp. Biol, 34, 784. Nelson S.D., Snodgrass W.R. & Mitchell J.R. (1976a) Chemical reaction mechanisms responsible for the tissue injury caused by monosubstituted hydrazines and their hydrazide drug precursors. In In Vitro Metabolic Activiation in Mutagenesis Testing. Editors: F.J. de Serres, J.R. Fouts, J.R. Bend & R.M. Philpot. Elsevier/North Holland Biomedical Press, Amsterdam; 257-276. Nelson S.D., Mitchell J.R., Timbrell J.A., Snodgrass W.R. & Corcoran G.B. (1976b) Isoniazid and iproniazid: activation of metabolites to toxic intermediates in man and rat. Science., 193, 901-903. Nelson S.D. (1990) Molecular mechanisms of the hepatotoxicity caused by acetaminophen. Sem. Liv. Dis., 10, 267-278. Noda A., Goromaru T., Matsuyama K., Sogabe K., Hsu K.-Y. & Iguchi S. (1978) Quantitative determination of hydrazines derived from isoniazid in patients. I. J. Pharm. Dyn., 1, 132-141. Noda A., Hsu K.-Y, Noda H., Yamamoto Y. & Kurozumi T. (1983) Is isoniazid-hepatotoxicity induced by the metabolite hydrazine? J. UOEH., 5, 183-190. Noda A., Noda H., Ohno K., Sendo T., Misaka A., Kanazawa Y., Isobe R.-I. & Hirata M. (1985) Spin trapping of a free radical intermediate formed during microsomal metabolism of hydrazine. Biochem. Biophys. Res. Commun., 133, 1086-1091. Noda A., Sendo T., Ohno K., Noda H. & Goto S. (1987) Metabolism and cytotoxicity of hydrazine in isolated rat hepatocytes. Chem. Pharm. Bull, 35, 2538-2544. Noda A., Noda H., Misaka A., Sumimoto H. & Tatsumi K. (1988) Hydrazine radical formation catalyzed by rat microsomal NADPH-cytochrome P-450 reductase. Biochem. Biophys. Res. Commun., 153, 256-260. Nolan CM., Sandblom R.E., Thummel K.E., Slattery J.T. & Nelson S.D. (1994) Hepatotoxicity associated with acetaminophen usage in patients receiving multiple drug therapy for tuberculosis. Chest., 105, 408-411. 235 O'Shea D., Kim R.B. & Wilkinson G.R. (1997) Modulation of CYP2E1 activity by isoniazid in rapid and slow W-acetylators. Br. J. Clin. Pharmacol, 43 , 99-103. Omura T. & Sato R. (1964) The carbon monoxide binding pigment of liver microsomes. J. Biol. Chem., 239 , 2370-2378. Parthasarathy R, Raghupati Sarma G., Janardhanam B., Ramachandran P., Santha T., Sivasubramanian S., Somasundaram P.R. & Tripathy S.P. (1986) Hepatic toxicity in South Indian patients during treatment of tuberculosis with short-course regimens containing isoniazid, rifampicin and pyrazinamide. Tubercle., 67 , 99-108. Patrick R.L. & Back K.C. (1965) Pathology and toxicology of repeated doses of hydrazine and 1,1-dimethylhydrazine in monkeys and rats. Ind. Med. Surg., 34 , 430-435. Patten C.J., Thomas P.E., Guy R.L., Lee M., Gonzalez F.J., Guengerich F.P. & Yang C.S. (1993) Cytochrome P-450 enzymes involved in acetaminophen activation by rat and human liver microsomes and their kinetics. Chem. Res. Toxicol, 6, 511-518. Peretti E., Karlaganis G. & Lauterburg B.H. (1987) Increased urinary excretion of toxic hydrazino metabolites of isoniazid by slow acetylators. Effect of a slow-release preparation of isoniazid. Eur. J. Clin. Pharmacol, 33 , 283-286. Perry T.L., Wright J.M., Hansen S. & Macleod P.M. (1979) Isoniazid therapy of Huntington Disease. Neurology., 29 , 370-375. Perry T.L., Kish S.J., Hansen S., Wright J.M., Wall R.A., Dunn W.L. & Bellward G.D. (1981) Elevation of brain GAB A content by chronic low-dosage administration of hydrazine, a metabolite of isoniazid. J. Neurochem., 37 , 32-39. Perry T.L., Wright J.M., Hansen S., Baker Thomas S.M., Allan B.M., Baird P.A. & Diewold P.A. (1982) A double-blind clinical trial of isoniazid in Huntington disease. Neurology.,.32, 354-358. Pessayre D., Bentata M., Degott C, Nouel O., Miguet J.P., Rueff B. & Benhamou J.-P. (1977) Isoniazid-rifampin fulminant hepatitis: A possible consequence of the enhancement of isoniazid hepatotoxicity by enzyme induction. Gastoenterology., 72 , 284-289. Pessayre D., Dolder A., Artigou J.-Y., Wandscheer J.-C, Descatoire V., Degott C. & Benhamou J.-P. (1979) Effect of fasting on metabolite-mediated hepatotoxicity in the rat. Gastroenterology., 11, 264-271. Phillips AH. & Langdon R.G. (1962) Hepatic Triphosphopyridine Nucleotide-Cytochrome c Reductase: Isolation, Characterization, and Kinetic studies. J. Biol. Chem., 237 , 2652-2660. 236 Pilheu J.A., De SalvoM.C. & Barcat J.A. (1979) Accion de los esquemas con isoniacida y rifampicina sobre el higado de enfermos tuberculosos. Medicina (Buenos Aires), 39, 298-304. Pumford N.R. and Halmes N.C. (1997) Protein targets of xenobiotic reactive intermediates. Annu. Rev. Pharmacol. Toxicol, 37, 91-117. Ram P.A.& Waxman D.J. (1992) Thyroid hormone stimulation of NADPH-cytochrome P450 reductase expression in liver and extrahepatic tissues - regulation by multiple mechanisms. J. Biol. Chem., 267, 3294-3301. Randolph H. & Joseph S. (1953) Toxic hepatitis with jaundice occurring in a patient treated with isoniazid. JAMA., 152, 38-40. Raucy J.L., Lasker J.M., Lieber CS. & Black M. (1989) Acetaminophen activation by human liver cytochromes P450IJJE1 and P450IA2. Arch. Biochem. Biophys., Ill, 270-283. Reid F.J. (1965) Hydrazine poisoning. Brit. Med. J., 2, 1246. Reinke L.A. & Moyer M.J. (1985) /?-Nitrophenyl hydroxylation: a microsomal oxidation which is highly inducible by ethanol. Drug. Metab. Dis., 13, 548-552. Rey-Grobellet X., Ferre N., Eeckhoutte C, Larrieu G., Pineau T. & Galtier P. (1996) Structural requirements for the induction of cytochromes P450 by benzimidazole anthelmintic derivatives in cultured rabbit hepatocytes. Biochem. Biophys. Res. Commun., 220, 789-794. Riska N. (1976) Hepatitis cases in isoniazid treated groups and in a control group. Bull. Int. Union. Tuberc, 51, 203-208. Rubin B., Hassert GL. Jr., Thomas B.G.H. & Burke J.C. (1952) Pharmacology of isonicotinic acid hydrazide (Nydrazid). Am. Rev. Tuberc, 65, 392-401. Ryan D.E., Koop D.R, Thomas P.E., Coon M.J. & Levin W. (1986) Evidence that isoniazid and ethanol induce the same microsomal cytochrome P-450 in rat liver, an isozyme homologous to rabbit liver cytochrome P-450 isozyme 3a. Arch. Biochem. Biophys., 246, 633-644. Sarich T.C, Adams S.P., Zhou T. & Wright J.M. (1993) Isoniazid-induced hepatotoxicity in rabbits: A comparison of steatosis and necrosis. Clin. Invest. Med., (abstr) 16, B67. Sarich T.C, Adams S.P., Youssefi M., Zhou T. & Wright J.M. (1994) Isoniazid-induced hepatotoxicity in rabbits: role of hydrazine. Can. J. Physiol. Pharmacol, (abstr) 72 (Suppl. 1), 585. 237 Sarich T . C , Zhou T., Adams S.P., Bain A.I., Wall R.A. & Wright J .M. (1995) A model of isoniazid-induced hepatotoxicity in rabbits. J. Pharmacol. Toxicol. Methods. 34, 109-116. Sarich T . C , Youssefi M . , Zhou T., Adams S.P., Wall R.A. & Wright J .M. (1996) The role of hydrazine in the mechanism of isoniazid hepatotoxicity in rabbits. Arch Toxicol, 70, 835-840. Sarma G.R., Chandra I., Kailasam S., Narayana A S . L . & Venkatesan P. (1986) Rifampin-induced release of hydrazine from isoniazid. Am. Rev. Respir. Dis., 133, 1072-1075. Scales M.D.C. & Timbrell J.A. (1982) Studies on hydrazine hepatotoxicity: 1. Pathological findings. J. Toxicol. Environ. Health, 10, 941-953. Scharer L . & Smith J.P. (1969) Serum transaminase elevations and other hepatic abnormalities in patients receiving isoniazid. Ann. Intern. Med., 71, 1113-1120. Sendo T., Noda A , Noda H. , Hsu K . - Y . & Yamamoto Y . (1984) Metabolic hydrolysis of isoniazid by subcellular fractions of rat liver. J.UOEH, 6, 249-255. Shenfield G . M . (1981) Influence of thyroid dysfunction on drug pharmacokinetics. Clin. Pharmacokinet. 6, 275-297. Shimoni Y . & Banno H . (1993) a-Adrenergic modulation of transient outward current in hyperthyroid rabbit myocytes. Am. J. Physiol, 264, H74-H77. Sim E . & Hickman D. (1991) Polymorphism in human iV-acetyltransferase - the case of the missing allele. Trends. Pharmacol. Sci., 12, 211-213. Sims F.H. & Rautanen P. (1975) Serum arginino-succinate lyase: Observations on the sensitivity and specificity of this test in the detection of minimal hepatocellular damage. Clin. Biochem., 8, 213-221. Singapore Tuberculosis Service/British Medical Research Council. (1977) Controlled trial of intermittent regimens of rifampin plus isoniazid for pulmonary tuberculosis in Singapore. Am. Rev. Respir. Dis., 116, 807-820. Singh J., Garg P.K., Thakur V.S. & Tandon R K . (1995) Anti-tubercular treatment induced hepatotoxicity: Does acetylator status matter? Indian. J. Physiol. Pharmacol, 39, 43-46. Sinha B.K. (1987) Activation of hydrazine derivatives to free radicals in the perfused rat liver: A spin-trapping study. Biochim. Biophys. Acta., 924, 261-269. Smith J., Tyrrell W.F., Gow A., Allan G.W. & Lees A.W. (1972) Hepatotoxicity in rifampin-isoniazid treated patients related to their rate of isoniazid inactivation. Chest., 61, 587-588. 238 Smith CV. & Mitchell JR. (1989) Pharmacological aspects of glutathione in drug metabolism. In Coenzymes and Co/actors volume 3: Glutathione, Chemical, Biochemical, and Medical Aspects. Dolphin D., Poulson R., & Avramovic Eds. John Wiley & Sons. Toronto, pp. 1-43. Snawder J.E., Roe A.L., Benson RW & Roberts D.W. (1994) Loss of CYP2E1 and CYP1A2 activity as a function of acetaminophen dose: Relation to toxicity. Biochem. Biophys. Res. Commun., 2 0 3 , 532-539. Snider D.E. & Caras G.J. (1992) Isoniazid-associated hepatitis deaths: A review of available information. Am Rev. Respir. Dis., 145, 494-497. Snodgrass W., Potter W.Z., Timbrell J., Jollow DJ. & Mitchell J.R. (1974) Possible mechanism of isoniazid-related hepatic injury. Clin. Res., (Abstr.) 2 2 , 323A. Springer D.L., Krivak B.M., Broderick D.J., Reed D.J. & Dost F.N. (1981) Metabolic fate of hydrazine. J. Toxicol. Environ. Health., 8 , 21-29. Steele M.A., Burk R.F. & DesPrez R.M. (1991) Toxic hepatitis with isoniazid and rifampin - A meta-analysis. Chest., 99, 465-471. Takayama K., Schnoes H.K., Armstrong E.L. & Boyle R.W. (1975) Site of inhibitory action of isoniazid in the synthesis of mycolic acids in Mycobacterium tuberculosis. J. Lipid. Res., 16, 308-317. Tappel A.L. & Zalkin H. (1959) Lipide peroxidation in isolated mitochondria. Arch Biochem Biophys, 8 0 , 326-332. Thomas B.H., Whitehouse L.W. & Zeitz W. (1978) Isoniazid toxicity alone and in combination with ethanol in the rabbit. II: Pharmacokinetic factors. Toxicol Appl Pharmacol, (abstr.) 45, 351. Thomas B.H., Wong L.T., Zeitz W. & Solomonraj G. (1981) Isoniazid metabolism in the rabbit, and the effect of rifampin pretreatment. Res. Commun. Chem. Pathol. Pharmacol., 3 3 , 235-247. Thomas B.H., Whitehouse L.W. & Zeitz W. (1984) Metabolism of [14C]acetylisoniazid and [14C]acetylhydrazine by the rat and rabbit. Fund. Appl Toxicol, 4, 646-653. Thummel K.E., Lee C.A., Kunze K.L., Nelson S.D. & Slattery J.T. (1993)" Oxidation of acetaminophen to N-acetyl-p-aminobenzoquinone imine by human CYP3A4. Biochem. Pharmacol, 45, 1563-1569. Timbrell J.A., Wright J.M. & Baillie T.A. (1977) Monoacetylhydrazine as a metabolite of isoniazid in man. Clin. Pharm. Ther., 2 2 , 602-608. 239 Timbrell J.A. & Wright J.M. (1979) Studies on the effects of isoniazid on acetylhydrazine metabolism in vivo and in vitro. Drug. Metab. Dispos., 7, 237-240. Timbrell J.A., Mitchell JR., Snodgrass W.R. & Nelson S.D. (1980) Isoniazid hepatotoxicity: The relationship between covalent binding and metabolism in vivo. J. Pharmacol. Exp. Ther., 213, 364-369. Timbrell J.A, Scales M.D.C. & Streeter A.J. (1982) Studies on hydrazine hepatotoxicity: 2. Biochemical findings. J. Toxicol. Environ. Health., 10, 955-968. Timbrell J.A. & Wright J.M. (1984) Urinary metabolic profile of isoniazid in patients who develop isoniazid-related liver damage. Human. Toxicol, 3, 485-495. Touster O., Aronson N.N. Jr., Dulaney J.T. & Hendrickson H. (1970) Isolation of rat liver plasma membranes. J. Cell. Biol, 47, 604-618. Valenzuela A. (1991) The biological significance of malondialdehyde determination in the assessment of tissue oxidative stress. Life Sciences. 48, 301-309. Vatsis K.P., Weber W.W., Bell D.A., Dupret J.-M., Price Evans D.A., Grant D.M., Hein D.W., Lin H.J., Meyer U.A., Relling M.V., Sim E., Suzuki T. & Yamazoe Y. (1995) Nomenclature for Af-acetyltransferases. Pharmacogenetics, 5, 1-17. Walubo A., Chan K. & Wong C.L. (1991) The pharmacokinetics of isoniazid and hydrazine metabolite in plasma and cerebrospinal fluid of rabbits. Meth. Find. Exp. Clin. Pharmacol, 13, 199-204. Waxman D.J., Morrissey J.J. & LeBlanc G.A. (1989) Hypophysectomy differentially alters P-450 protein levels and enzyme activities in rat liver: Pituitary control of hepatic NADPH cytochrome P-450 reductase. Mol. Pharmacol, 35, 519-525. Whitehouse L.W., Tryphonas L., Thomas B.H. & Paul CP. (1978) Isoniazid toxicity alone and in combination with ethanol in the rabbit. I: Pathologic and biochemical factors. Toxicol. Appl. Pharmacol, (abstr) 45, 351. Whitehouse L.W., Tryphonas L., Paul C.J., Solomonraj G., Thomas B.H. & Wong L.T. (1983) Isoniazid-induced hepatic steatosis in rabbits: an explanation for susceptibility and its antagonism by pyridoxine hydrochloride. Can. J. Physiol. Pharmacol, 61, 478-487. Williams H.L. & Bain J.A. (1961) Convulsive effect of hydrazines: Relationship to pyridoxine. Int. Rev. Neurobiol, 3, 319-348. Wohaieb S.A. and Godin D.V. (1987) Starvation-related alterations in free radical tissue defence mechanisms in rats. Diabetes., 36, 169-173. 240 Woo J., Chan C.H.S., Walubo A. & Chan K.K.C. (1992) Hydrazine: a possible cause of isoniazid-induced hepatic necrosis. J. Med., 23, 51-59. Woodward K.N. & Timbrell J.A. (1984) Acetylhydrazine hepatotoxicity: The role of covalent binding. Toxicology., 30, 65-74. Wright J.M. & Timbrell J.A. (1978) Factors affecting the metabolism of [14C]acetylhydrazine in rats. Drug. Metab. Dispos., 6, 561-566. Wright J.M., Stokes E.F. & Sweeney V.P. (1982) Isoniazid-induced carbamazepine toxicity and vice versa. New. Engl. J. Med., 307, 1325-1327. Wright J.M., Ngai H., Adams S., Behm A. & Wall R.A. (1986) Lack of hepatotoxicity of acetylhydrazine in rodents. Acta. Pharmacol. Toxicol, (abstr) 59(Suppl. 5), 221. Wright J.M. (1996) personal communication. Yard A.S. & McKennis H. Jr. (1955) Effect of structure on the ability of hydrazino compounds to produce fatty livers. J. Pharmacol. Exp. Ther., 114, 391-397. Youssefi M. (1992) The effect of acetaminophen on isoniazid metabolism. Masters Thesis, Department of Pharmacology & Therapeutics, University of British Columbia. Zand R., Nelson S.D., Slattery J.T., Thummel K.E., Kalhorn T.F., Adams S.P. & Wright J.M. (1993) Inhibition and induction of cytochrome P-4502El-catalyzed oxidation by isoniazid in humans. Clin. Pharmacol. Ther., 54, 142-149. Zar J.H. (1984) Biostatistical Analysis. Toronto: Prentice-Hall Canada Inc. Zhang Y., Heym B., Allen B., Young D. & Cole S.T. (1992) The catalase-peroxidase gene and isoniazid resistance of Mycobacterium tuberculosis. Nature, 358, 591-593. Zhou T. (1990) Isoniazid Hepatotoxicity. Masters Thesis, Department of Pharmacology & Therapeutics, University of British Columbia. 241 APPENDIX APPENDIX 1 A copy of the method paper " A model of isomazid-INH-induced hepatotoxicity in rabbits", Sarich etal, 1995. 242 E U E V I E R A Model of Isoniazid-induced Hepatotoxicity in Rabbits Troy Casimir Sarich. Ting Zhou, Stephen Paul Adams, Allen Ian Bain, Richard Arthur Wall, and James Morrison Wright Departments of Pharmacology & Therapeutics (T.C.S.. T.Z., S.PA., A.I.B., RA.W., J.M.W.) and Medicine (J.M.I*'.), Faculty of Medicine. The University of British Columbia. Vancouver. Canada Isoniazid (INH) continues to be an effective drug used for chemoprophylaxis and treatment of tuberculosis. Unfortunately. INH is associated with significant hepatotoxicity in up to 2% of individuals exposed, and if this adverse event is not recognized early it can be fatal. Research on INH-induced hepatotoxicity has been hampered by the lack of a suitable animal model that closely resembles the toxicity in humans. The mechanism of INH-induced hepatotoxicity is still unknown. The present study describes the development of a reliable model of INH-induced hepatotoxicity in rabbits. The protocol involves repeated injections of INH over a 2-day period, resulting in significant hepatic necrosis as indicated by elevations of plasma arginino-succinic acid lyase activity. Pretreatment with phenobarbital increased the occurrence of INH-induced hepatic necrosis from approximately 60% (9 out of 15 rabbits) with INH alone to more than 90^ (13 out of 14 rabbits). Morphological indices were used to demonstrate the presence of INH-induced hepatotoxicity. and biochemical indices were used to demonstrate both the presence and severity of INH-induced hepatotoxicity in this model. This model may prove useful for further investigations into the mechanism of INH-induced hepatotoxicity. Key Words: Isoniazid; Drug-induced hepatotoxicity: Argininosuccinic acid lyase: Hepatic necrosis: Cytochrome P-450; Phenobarbital Introduction Isoniazid (INH) continues to be a highly effective drug in the chemoprophylaxis and treatment of tuber-culosis. Unfortunately, daily INH administration is asso-ciated with mild elevations of liver enzyme activities in plasma in up to 20% of patients (Mitchell et al.. 1975a) and significant hepatotoxicity in approximately I9c~2% of patients receiving the drug (Barlow et al.. 1974). If this hepatotoxicity is not recognized early, it can be fatal. It has been reported that INH-induced liver injury is clinically, biochemically, and histologically indistinguish-able from viral hepatitis (Black et al.. 1975). INH has been used clinically since 1952. and although sporadic case reports of suspected INH-induced hepa-totoxicity surfaced shortly alter release of INH. its potential to produce hepatotoxicity was not fully appre-Address reprint requests to James Morrison Wright. M.D.. Ph.D.. F.R.C.P.(C). .Associate Professor. 2176 Health Sciences Mall. Depart-ments of Pharmacology & Therapeutics and Medicine. Faculty of Medicine. The University of British Coiumbia. Vancouver. B.C.. V6T IZJ. Canada. Received February 23. IUI>5: revised and accepted April I1). IWS. Journal of Pharmacological and ToKicologicai Methods 3a. 109-1 If> ( C 1995 Elsevier Science Inc. n55 Avenue of the Americas. New York. N Y :n>10 dated for almost 20 years. In February 1970. some 2,321 Capitol Hill employees were started on a tuberculosis chemoprophylaxis program with INH. Nineteen of the employees developed clinical signs of hepatotoxicity. and two died (Garibaldi et al., 1972). The "Capitol Hill incident" along with previous case reports of INH-induced hepatotoxicity (Randolph and Joseph. 1953; Gellis and Murphy. 1955; Scharer and Smith, 1969: Martin and Arthaud. 1970) confirmed that INH could produce hepatotoxicity. This appreciation initiated re-search efforts to elucidate the mechanism of INH-induced hepatotoxicity. a goal that has yet to be realized. Research on INH-induced hepatotoxicity has been hampered by the lack of a suitable animal model that closely parallels the toxicity in humans. Early studies suggested that the toxicity was due to the conversion of a metabolite of INH to a reactive intermediate (Snodgrass et al.. 1974). Though these initial studies were done using a rat model, there is evidence that INH and its major metabolites are not hepatotoxic in rats (Wright et al., 1986: J.A. Timbrell, personal communi-cation). Hepatotoxic actions of INH in dogs (Rubin et 1056-<J719,95/S9.i0 SSDI 1056-3719(95)00044-! 243 al., 1952). rabbits (McKennis et al., 1956; Whitehouse et al.. 1983). and guinea pigs (Heisey et al.. 1980) have also been investigated. However, none of these experiments yielded an animal model that showed susceptibility to INH-induced hepatotoxicity (hepatic necrosis), a delay in onset of the toxicity and variability in the severity of the toxicity, as is observed in humans. The present report describes a rabbit model of INH-induced hepatotoxicity that demonstrates these features and therefore resem-bles the toxicity in humans. Methods Animals Male New Zealand white rabbits (2-3 kg) were obtained from the Animal Care Unit of the University of British Columbia. The animals were housed in stainless-steel cages with free access to food and water throughout the study. Materials INH and sodium phenobarbital were obtained from British Drug House Chemicals Canada Ltd. Reagents for the argininosuccinic acid lyase (ASAL) assay (Cam-panini et al., 1970) were obtained as follows: Barium argininosuccinate (which was converted to the sodium salt by admixture with sodium sulfate and centrifuga-tion) and 2,4-dichloro-l-naphthol were obtained from Sigma: the other reagents required for the A S A L assay were obtained from local chemical suppliers and were all of reagent grade. The analysis of plasma alanine amino-transferase (.ALT. also known as serum glutamic-pyruvic transaminase, or SGPT) activity was done using a kit from Fisher and a multi-enzyme standard from Sigma. Drug Administration INH was given either orally or subcutaneously. Oral administration was performed with rabbits protected by appropriate restraining frames. Five-percent dextrose was used as the vehicle in preparation of INH (0.76 mol/L) for oral administration. The solution was given by gavage (0.48 mL/kg and 0.33 mL/kg) and was fol-lowed by a 1-mL bolus of 5% dextrose. Subcutaneous injections of INH dissolved in saline were done on the back between the shoulder blades. The administration schedule for INH (for both oral and subcutaneous routes) involved doses of 0.37 mmol/kg (50 mg/kg) followed by three 0.26-mmol/kg (35 mg/kg) doses at 3 hr intervals on day 1. This entire administration protocol was repeated on day 2. Phenobarbital, dissolved in saline (0.69 mol/L). was injected intraperitoneal^' (0.15 mL/kg) during the three mornings preceding the first day of INH injections to induce the microsomal cytochrome P-450 dependent enzymes as described in Whitehouse et al. (1985). except that the dose used was 0.11 mmol/kg (25 mg/kg). Blood Sampling Blood samples (1 mL) were taken from the lateral ear vein using a heparinized syringe. Topical administration of xylene (Fisher) was used to dilate the vein during blood sampling. The xylene solution was removed shortly after application by wiping the ear with an alcohol swab. Plasma was isolated from the blood sam-ples by centrifugation and frozen (-60° C) until analysis. Blood samples were taken before (0 hr). during (30 hr). and after the period of INH administration (48, 56, 72. 80, and 96 hr). The rabbits were killed after the final (96 hr) blood sample was obtained. Biochemical Analyses A S A L activity in the plasma was used as a sensitive marker of liver damage (Campanini et al.. 1970: Sims and Rautanen. 1975). The A S A L enzyme is found specifically in the liver parenchyma and is active in the urea cycle by catalyzing the conversion of argininosuc-cinic acid to arginine and fumaric acid. The quantitation of plasma A S A L activity was done according to Campa-nini et al. (1970) with the following modifications: To both the unknown and control tubes, 0.1-mL plasma was added followed by addition of 0.3-mL sodium arginino-succinic acid to the unknown tube. The unknown tube was incubated for 1 hr at 37° C. To both tubes 0.2-mL trichloroacetic acid solution was added. After letting stand for 5 min, a 0.5-mL aliquot was taken from each tube and added to separate tubes. Added to these tubes were 0.1 mL of 10% NaOH, 0.25 mL of the dichlo-ronaphthol solution, and 0.1 mL of the NaOCl solution. The tubes were allowed to stand in an ice bath for 15 min. and the color was read at an absorbance of 515 nm: A S A L activity is expressed in Takahara units (u.moI/100 mL per hr). .An animal was considered to have '"signifi-cant" hepatic necrosis when two successive blood sam-ples revealed plasma A S A L activity that exceeded the upper 99th log normal percentile (12.8 Takahara units or 1.11 log Takahara units). A L T activity in plasma, expressed as Units/L (u,mol/min per L), was quantitated to compare with plasma A S A L activity as a marker of hepatic necrosis. Even though plasma A S A L activity is thought to be a more sensitive indicator of liver paren-chymal cell damage, plasma A L T (SGPT) activity is more commonly used as a marker of hepatotoxicity (Scharer and Smith. 1969: Nolan et al.. 1994) and was therefore quantitated to verify the usefulness of plasma A S A L activity. 244 T.C. SARICH ET AL. ISONIAZID-INDUCED HEPATOTOXICITY IN RABBITS Slide Preparation Slides of Hematoxylin-Eosin stained liver tissue were prepared as follows: Three 1-mm-thick slices (one from each of the three large lobes) were removed from each liver and placed in Bouin's fixative solution for 30-45 min and then maintained in 70% ethanol until mounting and staining. The Bouin's fixative solution was made up with 75-mL 1.2% saturated aqueous picric acid solution, 25-mL 40% formalin, and 5-mL glacial acetic acid. Histological Analysis Histological examinations were conducted on rabbit liver slices. Liver slices for histological analysis were collected immediately after sacrifice of the rabbits. His-tological examination of morphological changes in the liver indicating hepatic necrosis was done by two of the authors (T.C.S. and J.M.W.), who during the evaluation of the slides were blinded to the treatment groups or the corresponding plasma A S A L activities. The slides were evaluated with the purpose of identifying whether he-patic necrosis was or was not present. Statistical Analysis Plasma A S A L activities were logarithmically trans-formed to fit a normal distribution. Al l statistical proce-dures were performed using the log-transformed values. For clarity of presentation, the plasma A S A L activity values are presented in the nontransformed form (mean - standard error). Analysis of variance (ANOVA). two-sample t tests and F tests (Microsoft Excel 4.0) were used to compare treatment groups. The Tukey test for multiple comparisons was done as outlined in Zar (1984). Graphs (including the correlation) were done using the SlideWrite Plus 2.0 computer program. The points on the time course curves were joined using a "spline" curve (SlideWrite Plus 2.0). Results Initially, a number of experiments were performed in .in effort to find an INH dosage regimen that produced liver damage as measured by significant liver enzyme release in at least 50% of the rabbits. Single doses of INH—for example, 0.15 mmol/kg i.p. (20 mg/kg), 0.44 mmol/"kg i.p. (60 mg/kg)—did not cause hepatic necro-sis. When a dose of 150 mg'Tcg was administered intra-peritoneally. manifestations of acute CNS toxicity and seizures preceding death were observed. Also, several multiple-dose regimens—for example, two 0.37-mmol/kg subcutaneous injections (50 rngkg) 5 hr apart for each of 2 days—of INH over 2 days did not cause significant hepatotoxicity. The regimen consisting of eight doses over 2 days, as outlined in Methods, was the most effective regimen tested in consistently producing measurable hepatic necrosis in greater than 50%- of animals. In order to compare plasma A S A L activity to plasma A L T activity (a plasma liver enzyme used clinically), correlation of plasma A S A L and A L T activities was done using peak plasma A S A L activities versus the corresponding plasma A L T activity in 27 of 29 rabbits in the study (two rabbits without evidence of elevated plasma A S A L activity did not have their plasma A L T activity measured). Plasma A S A L activity positively cor-related with plasma A L T activity, (r = +0.89, p < .0001) (Figure 1). The time course of changes in plasma A S A L and A L T activity, including elevation, peak and decline are shown in Figure 2. Although the activities of plasma A S A L (4.6 ± 0.4 Takahara units) and A L T (55.5 i 3.5 Units/L) associated with INH-induced hepatotoxicity are similar in many ways, the baseline plasma A S A L activities are lower. Examination of liver slides taken from the INH-treated rabbits revealed that the area of hepatic necrosis was low (<10%) for most livers, and many of the rabbit livers exhibited non-uniform histopathological damage. For these reasons, it was not possible to develop a reliable histological grading system for the degree of hepatic necrosis other than its presence or absence. Blinded examination of the liver slides resulted in de-tection of hepatic necrosis in zero out of 2 control rabbits, one out of 7 rabbits that did not experience a significant elevation of plasma A S A L activity and in 19 out of 22 rabbits that did experience a significant elevation of plasma A S A L activity. The 3 rabbits that exhibited no morphological evidence of hepatic necrosis had only mildly elevated peak plasma A S A L activities (23, 45. and 51 Takahara units). An example of a control liver shown in Figure 3 can be compared to an INH-treated liver demonstrating evidence of hepatic necrosis in Figure 4. We have found in our laboratory that plasma A S A L activity does not increase with handling, repeated injec-tions of rabbits with saline solutions, and.or blood sampling. This is shown by comparing pre-lNH plasma ASAL activity in rabbits pre-treated with phenobarbital (4.6 z 0.5 Takahara units, n = 14) and rabbits not pretreated (4.4 ± 0.5 Takahara units, n = 15). In addition, unpublished data from our laboratory has shown that blood withdrawal from 13 rabbits that expe-rienced daily intraperitoneal injections of 0.02-mol N H , O H per liter in saline over 5 days did not affect plasma A S A L activity (4.9 z 1.4 Takahara units) com-pared to baseline levels (3.6 z 0.4 Takahara units). For this reason. post-INH treatment plasma A S A L activities were compared to baseline (pre-lNH) plasma A S A L activity in each rabbit. Peak plasma A S A L activity was used for quantitation 245 0.5 Figure 1. Correlat ion of peak plasma A S A L ac-tivities (log Takahara units) with peak plasma A L T activities (log Units/L) using 27 of the 29 rabbits in the study. Plasma A L T activity was not determined for two rabbits that did not experi-ence elevation of plasma A S A L activity. The correlation is significant (r = +0.89, p < .0001) with a positive slope (+0.59). P e a k P l a s m a A S A L A c t i v i t y of hepatic necrosis in each individual rabbit. Although Figure 2 shows the plasma liver enzyme activities with a peak at 30 hr. the individual animal peak plasma liver enzyme activities varied between 30 and 96 hr. A statistically significant difference (p < .0001, t test) was detected between baseline plasma A S A L activity (4.6 + 0.4 Takahara units, n = 29) and peak plasma A S A L activity in all INH-treated rabbits (328.2 i 113.3 Takahara units, n = 29). In total, 22 out of 29 rabbits demonstrated a significant elevation of plasma A S A L activity (criteria defined in Methods section). There was no difference in peak plasma A S A L activ-ities between rabbits administered INH orally (124.0 £ 76.6 Takahara units, n = 8) or by subcutaneous injection (212.0 ± 122.1 Takahara units, n = 7) (p = .61, t test) or between phenobarbital pre-treated rabbits administered INH orally (548.4 r 315.6 Takahara units, n = 8) or by-subcutaneous injection (442.1 ± 316.3 Takahara units, n - 6) (p = .73, t test). Because the route of INH administration did not affect the hepatotoxic potential of INH, the results for oral and subcutaneous adminis-tration groups were pooled together to form one group of INH-treated rabbits and one group of phenobarbital pre-treated and INH-treated rabbits. Comparison of peak plasma ASAL activity between baseline (4.6 ± 0.4 Takahara units, n = 29). INH-treated (165.1 ± 68.4 Takahara units, n = 15) and phenobarbital pre-treated and INH-treated rabbits (502.9 ± 217.6 400 > 350 < 300 <D E 250 > . N l c LU 200 I. ID > 150 CO 100 E m « 50 -Q_ 0 10 20 30 40 50 60 70 80 90 100 T i m e a f t e r f i r s t I N H d o s e ( h o u r s ) Figure 2. T ime course of plasma A S A L ( + ) (Takahara units) and A L T (•) (Uni ts /L) ac-tivities. The similarity in the rise, peak, and decline of plasma liver enzyme activities after INH-induced hepatic insult are shown. This graph is intended as a representation of the time course of liver enzyme release into plasma and is hot meant to suggest a mecha-nism for the toxicity. The time points are fitted with a "spl ine" curve. 246 T.C. SARICH E T AJL ISONIAZID-rNDUCED HEPATOTOXICITY I N RABBITS Figure 3. A healthy liver slice from an untreated rabbit (4> - central vein). (Heraatoxylin-Eosin stain, magnification 200X) . Takahara units, n = 14) revealed a statistically signifi-cant difference between all three groups (p < .0001. ANOVA;p < .Ol.Tukey's test: Figure 5). Nine out of 15 rabbits (60%) treated with INH only and 13 out of 14 of the phenobarbital pretreated and INH-treated rabbits (90%) demonstrated significant elevations of plasma ASAL activity (as defined in Methods). Discussion The difficulty surrounding the development of an animal model of INH-induced hepatotoxicity has slowed progress in this field. We also found it difficult to establish a dosing regimen in rabbits that reliably pro-duced INH-induced hepatotoxicity For example, when Figure 4. A liver slice from a rab-bit pretreated with phenobarbital and injected with I N H subcutane-ously (=> - central vein). Lympho-cytic infiltration and ballooning de-generation of hepatocytes with individual hepatic necrosis can be seen (—»- hepatocellular necrosis) Marked fatty vacuolization is also detected (>• - fatty vacuoles). Just prior to sacrifice (56 hr after the first I N H dose), this animal had a plasma A S A L activity of 392 Taka-hara units and plasma A L T activity of 850 Uni ts /L . (Hematoxyl in-Eo-sin stain, magnification 200x) . 247 Figure 5. Peak plasma A S A L activity as compared by treatment group. The difference between plasma A S A L ac-tivity (Takahara units) in all three groups including baseline (4.6 r 0.4. n = 29). INH-only- t reated (165.1 = 68.4, n = 15) and phenobarbital pre-treated and INH-treated (502.9 ~ 217.6. n = 14) was statistically signif-icant (*) (p < .0001 using A N O V A and p < .01 using the Tukey test for multiple comparisons). The individ-ual times to peak plasma A S A L ac-tivity varied from 30 to 96 hr between animals. Values shown are mean r standard error. Statistics were done using logarithmically transformed data. o < _J < CO < CO E JO ri-al o 0.5 BASELINE INH ONLY INH+PHENOBARBITAL Treatment G roups INH was administered acutely in a dose of 150 mg/kg. CNS toxicity involving seizures and death precluded the development of INH-induced hepatotoxicity. Only after a protocol in which INH was administered at 50 mg/kg followed by three 35-mg/kg doses at 3-hr intervaJs on day 1 and day 2 did signs of significant hepatic necrosis appear without being confounded by the presence of CNS toxicity (although there remains CNS toxicity and death from INH in approximately one of every 10 to 15 rabbits). Using this model, significant hepatic necrosis develops in more than 50% of the rabbits as quantitated by elevation of plasma liver enzymes and verified by histopathological examination. Isoniazid has been shown to cause a diffuse pattern of hepatic parenchymal injury similar to that of viral hep-atitis (Kaplowitz et al., 1986). Elevations of liver enzyme activity in plasma are associated with acute hepatocellu-lar necrosis and reflect the release of enzymes from the cytoplasm of dying cells (Kaplowitz et al.. 1986). In-creased activity of liver cell enzyme activities is thought to be a more sensitive indicator of cellular damage than are most other biochemical indices and most morpho-logical findings (Schmidt et al.. 1974). Plasma A S A L activity, our marker of hepatic necrosis, has been re-ported to be a sensitive marker for hepatic parenchymal cell injury, even if this is relatively minor (Campanini et al., 1970: Sims and Rautanen. 1975). Therefore, moni-toring plasma A S A L permits reliable detection of mild to severe hepatic necrosis. In order to help validate plasma A S A L activity as a good hepatotoxic marker in INH-induced hepatotoxicity in rabbits, we compared plasma A S A L activity with plasma A L T activity (a common clinically used marker for hepatotoxicity). The peak values of both enzyme activities correlate significantly and the time courses appear quite similar. However, increased A S A L activity in plasma is a more specific indicator of liver damage than is A L T because A L T can be released during injury to other organs and tissues (Campanini et al., 1970). Under conditions of injury to other organs and tissues, this correlation may not be significant. Therefore, by using plasma A S A L activity rather than plasma A L T activity as a marker for liver damage, the potential for interference by damage to organs other than the liver is eliminated. In addition, the plasma A S A L activity assay is especially useful in this model because plasma samples of INH-treated rabbits often become cloudy, which can interfere with the A L T assay but not the A S A L assay. .As a result, although both plasma ASAL and .ALT activities seem to be useful as markers of hepatotoxicity, for practical reasons, plasma A S A L activity is a more useful and convenient marker in this model of INH-induced hepatotoxicity. Significant elevation of plasma A S A L activity oc-curred in the majority of the rabbits treated with INH (22 out of 29). Also, as compared to baseline levels (4.6 units), the elevation of plasma A S A L activity ranged from approximately 3-fold (to 12.8 units to be signifi-cant) to 575-fold (2674 units was highest single peak). This wide range demonstrates the variability of the hepatic necrosis between animals. These observations in the rabbit model are somewhat similar to the observa-tion that up to 20% of humans taking INH develop a mild transient elevation of plasma liver enzymes (Mitch-ell et al., 1975a). and up to 2% develop extreme eleva-tions and severe hepatotoxicity (Barlow et al.. 1974). The 248 TC. SARICH ET .AL. ISONtAZID-IMX'CED HEPATOTOXICITY IN RABBITS rabbit model is characteristic of INH-induced hepato-toxicity in humans because there is large variability in predisposition to INH-induced hepatotoxicity, with some animals resistant, some animals mildly susceptible, and a smaller percentage very susceptible to it. The manifestations of INH-induced hepatotoxicity in humans range in severity from diffuse to multilobular, bridging, and massive necrosis (Maddrey and Boitnott, 1973). In the present study, histological analysis using Hematoxylin-Eosin stained liver slides revealed focal and centrilobular inflammatory infiltration and necrosis in INH-treated rabbits, both with and without pheno-barbital pretreatment. Centrilobular necrosis was present in the majority of the animals that experienced a significant elevation of plasma ASAL activity. Three of the rabbits that did not demonstrate histologically de-tectable centrilobular necrosis did exhibit significant elevations of plasma A S A L activities, though these activity elevations were marginal (23. 45. and 51 Taka-hara units). It is possible that in some cases of marginal release of A S A L into plasma, the characteristic morpho-logical features do not progress to an observable level. Also, because many of the livers exhibited non-uniform histopathological damage, it is possible that the liver slides were taken from a region not damaged by INH. One rabbit showed histological evidence of centri-lobular necrosis without having significantly elevated plasma A S A L activity (as defined in the Methods sec-tion). The reason for this could be that our biochemical criterion for "significant hepatotoxicity" is set too high. Alternatively, it could be that the time that the damage occurs is critical, because the peak plasma A S A L activity of 13 Takahara units in this rabbit occurred just prior to death. Although not a formal part of this study, examination of the liver slides also revealed evidence of varying degrees of hepatocellular vacuolization. These vacuoles are most likely fat-containing because it has been ob-served that INH administration to rabbits can cause hepatic steatosis (Whitehouse et al.. 19S3; Karthikeyan and Krishnamoorthy. 1991: Krishnamoorthy and Karthikeyan. 1991). These other studies evaluated INH-induced hepatic steatosis in rabbits: however, they did not report evidence of INH-induced hepatic necrosis, the predominant lesion observed in humans. Peak plasma A S A L activity was chosen as a marker to estimate the severity of hepatic necrosis. In addition, unpublished data from our laboratory show that the area under the A S A L versus time curve shows a high degree of correlation with peak plasma A S A L activity. The difficulty in using area under the plasma A S A L versus time curve is that some animals were killed early so that the area would not be standard for all rabbits. However, measurement of peak plasma ASAL activity provides a convenient and reliable way to biochemically quantitate hepatic necrosis. Although the numbers of animals in ea:h group are small (ranging from 6-8). the severity of hepatotoxicity produced by INH alone or in the presence of phenobar-bital pretreatment was equivalent after INH administra-tion by the oral or subcutaneous route. The observation that both routes of administration are comparable was interesting because if hepatic metabolism is a crucial component of INH-induced hepatotoxicity, then the first pass metabolism associated with oral administration might have been expected to give rise to greater expo-sure of the drug to the liver than that following subcu-taneous administration. Because this was not the case, data from the oral and subcutaneous administration routes were pooled together, with one group consisting of rabbits treated with INH only and another group that were treated with INH following phenobarbital pretreat-ment. It has been observed that there is no correlation between plasma INH concentration and susceptibility to INH-induced hepatotoxicity (Mitchell et al., 1975a). This supports the role of a metabolite of INH as being responsible for producing hepatotoxicity. It has been suggested that the mechanism of INH-induced hepato-toxicity involves conversion of metabolites of INH by cytochrome P-450 dependent enzymes into hepatotoxic species (Mitchell et al., 1976). It is on this basis that the possible potentiation of INH-induced hepatotoxicity with induction of hepatic P-450 enzymes using pheno-barbital was investigated. Phenobarbital pretreatment of rabbits prior to INH-treatment resulted in a significantly greater increase in plasma A S A L activity and a greater proportion of animals experiencing significant hepatic necrosis as compared with rabbits receiving INH alone. This suggests that induction of P-450 enzymes increases severity of INH-induced hepatotoxicity in rabbits, possi-bly by increasing the conversion of INH or a metabolite to a hepatotoxic intermediate. Acetylation by the N-acetyltransferase enzyme is a major step in the metabolism of INH. In humans, acetylation rate is genetically determined, and humans can be divided into slow and rapid acetylators (Kalow. 1962). Metabolism of INH. and in particular acetylation rate, has been proposed to play an important role in INH-induced hepatotoxicity in humans. Paradoxically, however, it has been suggested that both fast (Mitchell et al.. 1975b; Timbrell. 1979) and slow (Dickinson et al.. 1981: Musch et al.. 1982: Lauterburg et al.. 1985) acetylators are at a high risk of INH-induced hepatotox-icity. Acetylation rate in rabbits, like in humans, is a genetically determined polymorphic trait that follows simple Mendelian inheritance characteristics, and thus rabbits can be grouped as fast or slow acetylators of INH (Frymoyer and Jacox. 1963). Though the rabbit pheno-2 4 9 typing data in the present study was not given, there are in vivo phenotyping techniques available (Gordon et al.. 1973). The large range of NAT activity in rabbits allows for comparison to severity of INH-induced hepatotoxic-ity. Because of this similar metabolic pathway between rabbits and humans, analysis of acetylation rate as a risk of developing INH-induced hepatotoxicity could be car-ried out using this rabbit model. In conclusion, the present study has described a rabbit model of INH-induced hepatotoxicity. We believe that this represents a reproducible, clinically relevant, and useful model of INH-induced hepatotoxicity. The requirement for multiple dosing, the delayed time course of liver enzyme release, the variability in re-sponse, and the histological evidence of both hepatic necrosis and steatosis are all similar to INH-induced hepatotoxicity in humans. Many questions surrounding the mechanism of INH-induced hepatotoxicity can hopefully be answered using this model. The authors would like to thank Dr. David V. Godin and Dr. Gail D. Bellward for their critical reviews and invaluable suggestions to this paper. This, research was funded by the N. l .H. ( C M 32165). References Barlow PB. Black M. Brummer D L . Comstock G W . Dubin IN. Enterline P. Gibson M L . Hardy G E Jr. Harrel J A . Johnston RF. Kent D C . Marvin BA. McCaig NC. Mitchell JR. Mosley JW. Ogasawara FR. Popper H. Reichman LB. Zimmerman HJ (1974) Preventive therapv of tuberculous infection. Am Rev Resp Dis 110:371-374. Black M . Mitchell JR. Zimmerman HJ. Ishak K G . Epler G R (1975) Isoniazid-associated hepatitis in ! 14 patients. Gastroenterology 69 : ;s ii-3o:. Campanini RZ. Tapia RA. Sarnat W. Natelsnn S ( 1 » " 0 ) Evaluation of serum argininosuccinate lyase I ASAL) concentrations as an index to parenchymal liver disease. Clin Chem 16:44-53. Dickinson DS. Bailey WC. Hirschow,u Bl . Soong S-J. Eidus L. Hodgkin M M (1981) Risk factors for isoniazid (INH)-induced liver dvsfunction. / Clin Gastroenterol 3:271-279. Frymoyer JW. Jacox RF (1963) Investigation of the genetic control of sulfadiazine and isoniazid metabolism in the rabbit. / Lab Clin Med h2:S9t-9l)4. Garibaldi RA. Drusin RE. Ferebee SH. Gregg M B ( 19"2) Isoniazid-as>ueiated hepatitis. -\tn Rev Resp Dis I0b:557-3ro. Gelli- SN. Murphy R V I l*>55) Hepalitis following isoniazid. Dis Chest 2S:4o2-4o4. Gordon G R . Shafizadeh A G . Peters JH 11973) Polymorphic acetyla-tion of drugs in rabbits. Xenobiotica 3:133-150. Heisey G B . Hughes H C . Lang C M . Rozmiarek H (1980) The guinea pig as a model for isoniazid-induced reactions. Lab Anim Sci 30:42-50. Kalow WP (I9f)2) Pharmacogenetics: Heredirv and the Response to Drugs. Philadelphia: W.B. Saunders Company, pp. 93-104. Kaplowitz N. Aw T Y . Simon FR. Stolz A ( I'ISh) Drug-induced hepatotoxicity. Ann Intern Med 104:826-839. Karthikeyan S. Krishnamoorthy MS l l ' W I ) Effect of subacute admin-istration of isoniazid and pvndoxine on lipids i n plasma, liver, and adipose [issues in the rabbit. Dnig Client Toxicol 14:293-303. Krishnamoorthy MS. Karthikeyan S (1991) Effect of phenobarbitone pretreatment on isoniazid-induced lipid changes in plasma, liver and adipose tissue in the rabbit. Pharmacol Res 24:219-225. Lauterburg BH. Smith C V . Todd E L . Mitchell JR (1985) Oxidation of hvdrazine metabolites formed from isoniazid. Clin Pharmacol Ther 38:566-571. Maddrey WC, Boitnott JK (1973) Isoniazid hepatitis. Ann Intern Med 79:1-12. Martin C E . Arthaud JB (1970) Hepatitis after isoniazid administra-tion. V Engl J Med 282:433-434. McKennis H Jr. Yard AS, Pahnelas E V (1956) The production of fatty livers in rabbits by isoniazid and other hydrazine derivatives. Am Re\ Tuberc 73:956-959. Mitchell JR. Long MW. Thorgeirsson L P . Jollow DJ (1975aI Acety-lation rates and monthly liver function tests during one year of isoniazid preventive therapy. Chest 68:181-190. Mitchell JR. Thorgeirsson UP, Black M . Timbrell JA, Snodgrass WR, Potter WZ, Jollow DJ, Reiser H R (1975b) Increased incidence of isoniazid hepatitis in rapid acetylators: Possible relation to hydr-azine metabolites. Clin Pharmacol Ther 18:70-79. Mitchell JR, Zimmerman HJ, Ishak K G . Thorgeirsson UP, Timbrell JA. Snodgrass WR. Nelson SD (1976) Isoniazid liver injury Clinical spectrum, pathology and probable pathogenesis. Ann [mem Med 34:181-192. Musch E. Eichelbaum M . Wang JK. v Sassen W. Castro-Parra M . Dengler HJ (1982) Incidence of hepatotoxic side effects during antituberculous therapy (INH. RMP. EMB) in relation to acetyla-tor phenotype. Klin Wochenschr 60:513-519. Nolan C M . Sandblom R E . Thumme! K E . Slattery JT. Nelson SD (1994) Hepatotoxicity associated with acetaminophen usage in patients receiving multiple drug therapv for tuberculosis. Chest 105:408-411. Randolph H . Joseph S ( 1953) Toxic hepatitis with jaundice occurring in a patient treated with isoniazid. JAMA 152:38-40. Rubin B. Hassert G L Jr. Thomas B G H . Burke JC (1952) Pharmacol-oav of isonicotinic acid hydrazide (Nvdrazid). Am Re: Tuberc 65:392-401. Sch.irer L. Smith JP (19f,9) Scrum transaminase elevations ar.d other hepatic abnormalities in patients receivmu isoniazid. A':r. Interr. Ued "1:1113-11211. Schmidt E. Schmidt FW. Mdhr J. Otto P. Vido I. Wrogemann K. Herfarth Ch 11974) Liver morphology and enzyme release: Further studies in the isolated perfused rat liver. In Pathogenesis and Mechanisms of Liver Cell Necrosis. Ed.. D Keppler. Baltimore: University Park Press, pp. 147-162. Sims F H . Rautanen P (1975) Serum arginino-succinate lyase: Obser-vations on the sensitivity and specificity of this test in the detection of minimal hepatocellular damage. Clin Biochem 8:213-221. Snodgrass W. Potter WZ. Timbrell J. Jollow DJ. Mitchell JR il974i Possible mechanism o f isoniazid-related hepatic injury. C'in Res 22:323A. Abstract. Timbrell JA ( 1 9 7 9 ) The role o f metabolism in the hepatotoxicity o f isoniazid and iproniazid. Drug Metab Rev 10:125-147. Whitehouse LW. Iverson F. Wong L T (1985) Effects of nfamfv. pretreatment on hepatic parameters in the rabbit. To.x-.col Lett 24:131-136. Whitehouse LW'. Tryphonas L. Paul CJ. Solomonraj G. Thomas B H . Wong LT (1983) Isoniazid-induced hepatic steatosis in rabbits: .An explanation for susceptibility and its antagonism by pyridoxine hydrochloride. Can J Physiol Pharmacol 61:478-487. Wright JM. Ngui H. Adams S. Behm A. Wall R A (1986) Lack of hepatotoxicity of acetylhydrazine in rodents. Acta Pharmacol Toxi-col 5>)|suppl 5):22I. Abstract. Zar Jerrnld H ( 19X4) Biostatistical Anahsts. Toronto: Pre~:;ce-Hall Canada Inc.. pp. 1X5-205. 250 APPENDIX 2 A copy of Study 2, "Role of hydrazine in the mechanism of isoniazid hepatotoxicity in rabbits", Sarich etal, 1996, Archives of Toxicology. 251 Arch Toxicol (1996) 70: 835-840 ( ) R I ( . I \ \ I Troy C. Sarich • Mohammed Youssefi • Ting Zhou Stephen P. Adams • Richard A. Wall • James M . Wright Role of hydrazine in the mechanism of isoniazid hepatotoxicity in rabbits Received: 21 February 1996 Accepted: 30 Apr i l 1996 Abstract Isoniazid (INH) continues to be a highly ef-fective drug in the chemoprophylaxis and treatment of tuberculosis; however, its use is associated with hepato-toxicity (predominantly hepatic necrosis) in 1-2% of individuals. The INH metabolites, acetylhydrazine and hydrazine, have each been implicated as the causative hepatotoxin in INH-induced hepatotoxicity. Using a model of INH-induced hepatotoxicity in rabbits, in which INH-induced hepatotoxicity manifests as hep-atic necrosis, hepatic steatosis (hepatic fat accumula-tion) and hypertriglyceridaemia (elevated plasma tri-glycerides), we compared the severity of these measures of toxicity with plasma levels of INH, acetylhydrazine and hydrazine. Plasma INH and acetylhydrazine were not correlated with markers of INH-induced hepatic necrosis or fatty changes. Plasma hydrazine at 32 h was correlated significantly with plasma argininosuccinic acid lyase (ASAL. a sensitive marker of hepatic nec-rosis) activity as area under the curve lr~ = 0.54. P < 0.002) and log plasma ASAL activity at 48 h after the first dose of INH (r : = 0.53. p < 0.005), but not with fatty changes. These results show in this model of INH-induced hepatotoxicity in rabbits that hydrazine, and not INH or acetylhydrazine, is most likely involved in the pathogenic mechanism of hepatic necrosis. Key words Isoniazid - Acetylhydrazine • Hydrazine • Hepatotoxicity • ASAL (argininosuccinic acid lyase) Introduction T C . Sarich M. Youssefi T. Zhou S.P. Adams • R.A. Wall v & r t o £ : ^ a a u , i c , • T h S L ' m v m i , y ° f B m ' s h C o l u m b , a -J .M. Wright i a i Departments of Pharmacology and Therapeutics and Medicine The University of British Columbia. 21 "6 Health Sciences Mall Vancouver B.C.. V6T 1Z3 Canada Isoniazid (INH) continues to be a highly effective drug in the chemoprophylaxis and treatment of tuberculosis. Unfortunately, daily INH administration is associated with mild elevations of liver enzyme activities in plasma in up to 20% of patients (Mitchell et al. 1975) and significant hepatotoxicity (predominantly hepatic nec-rosis) in approximately 1-2% of patients receiving the drug (Barlow et al. 1974). If this hepatotoxicity is not recognized early, it can be fatal. It has been reported that INH-induced liver injury is clinically, biochem-ically and histologically indistinguishable from viral hepatitis (Black et al. 1975). The mechanism of INH-induced hepatotoxicity is unknown. Most of the previous research has focused on a rat model with the hypothesis that acetylhydrazine. a metabolite of INH. is the cause of the hepatotoxicity (Nelson et al. 1976; Mitchell et al. 1976; Timbrell et al. 1980). However, there is more recent evidence in an-imals and humans suggesting that hydrazine is the metabolite predominantly responsible for INH-in-duced hepatotoxicity (Noda et al. 1983: Woo et al. 1992; Gent et al. 1992). A model oflNH-induced hepa-totoxicity in rabbits has been developed, which has some similarities to the toxicity in humans iSarich et al. 1995). The present study examines the relationship be-tween markers of INH-induced pathological changes (hepatic necrosis, hepatic steatosis and hypertrig-lyceridaemia) and plasma levels of INH. acetylhyd-razine and hydrazine in the rabbit model of INH-induced hepatotoxicity. Materials and methods Animals Fifteen New Zealand white rabbits leiaht male, seven female, body wt. 2-3 kg) were obtained from the Animal Care Unit at the 252 University of British Columbia and were used in this study. Throughout the experiment, the rabbits were housed individually in stainless steel cages with a 12:12 h light/dark cycle and free access to food and water. This project has been reviewed and approved by the University of British Columbia Committee on Animal Care. Reagents I N H was obtained from Sigma Chemicals. Reagents for the ar-gininosuccinic acid lyase (ASAL) assay were obtained as follows: barium argininosuccinate (which was convened to the sodium salt by admixture with sodium sulphate and centrifugation) and 2.4-dichloro-l-naphlhol from Sigma: the other reagents required for the A S A L assay were obtained from local chemical suppliers and were all of reagent grade. I N H injection protocol The dosing schedule for I N H , as adopted from Sarich et al. (1995). involved a subcutaneous injection of 0.36 mmoles kg (50 mg/kgl followed by three 0.26 mmol/kg (35 mg/kg) injections at 3-h inter-vals. This entire injection protocol was repealed on day 2. Blood sampling Blood samples (1ml) were taken from the lateral ear vein using a heparinized syringe. Topical administration of xylene (Fisher Sci-entific) was used to dilate the vein during blood sampling. The xylene solution was removed shortly after application by wiping the ear with an alcohol containing swab. Plasma was isolated from the blood samples by centrifugation and frozen ( —60'C) prior to analy-sis within 2 weeks. Blood samples were taken before (0 h). during (32 h) and at various times (48. 56, 72. 80 and 96 h) following INH administration. The rabbits were killed by cervical dislocation fol-lowed by exsanguination after the final (96 h) blood sample was obtained. Tissue handling Immediately after sacrifice of the rabbits, the livers were removed, weighed, perfused with 0.25 M sucrose in 5 m M TRIS buffer (pH 8.0), homogenized (using a glass tube and Teflon pestle) in three volumes of the perfusion buffer per gram of liver and frozen at — 6 0 C prior to analysis within 2 weeks (Touster et al. 1970). Biochemical analyses A S A L activity in the plasma has previously been used as a sensitive marker of liver damage (Campanini et al. 1970: Sims and Rautanen 1975). The A S A L enzyme is found specifically in the liver paren-chyma and acts in the urea cycle by catalysing the conversion of argininosuccinic acid to arginine and fumaric acid. The quantitation of plasma A S A L activity (expressed as umol/100 ml per h; Takahara units) was done according to Campanini et al. (1970) with modifica-tions as described in Sarich et al. (1995). Plasma A S A L activities were logarithmically transformed for correlational analysts to fit the data to a normal distribution. Log peak plasma A S A L activity, plasma A S A L activity as area under the curve (AUC) and log plasma A S A L activity at 48 h are used as markers of hepatic nec-rosis. Although these three different measures of plasma ASAL activity are not entirely independent of each other, using all three provides greater insight into the pattern and severity of INH- in-duced hepatotoxicity. Plasma triglyceride levels (mM triolein equivalent) were quant-itated using a triglyceride kit from Sigma Diagnostics (kit number 336-10). Hepatic triglyceride accumulation was quantitated by analysis of the triglyceride content of crude liver homogenates using a Folch extraction (Folch et al. 1951) and modified as follows: 0.1 ml of diluted 33% (w/v) crude homogenate was added to 0.5 ml 2:1 chloroform/methanol in a test tube which was covered, placed on ice and vortexed every 10 min for a total of 60 min. The mixture was then centrifuged for 5 min at 1000 g resulting in the formation of two distinct layers. A 0.3 ml aliquot (from approx. 0.35 ml total) from the bottom layer was removed and evaporated to dryness using a boiling water bath. The resulting dried layer was analyzed for triglyceride content using a triglyceride kit from Sigma Diagnostics (kit number 336-101. The triglyceride level in liver tissue is expressed as mg triglyceride (triolein equivalent! g liver tissue. Values of plasma ASAL activity and triglyceride levels as areas under the curve were calculated by weighing in mg amount of paper representing the area under the curve from graphs with standard axes. Protein quantitation of liver homogenate samples was carried out using methods outlined by Bradford (1976). H P L C analysis of INH. acetylhydrazine and hydrazine The analytical procedure used for the quantitation of I N H . acetyl-hydrazine and hydrazine in the plasma samples obtained at 32 and 48 h after administration was developed in this laboratory and is as follows. The derivatizing reagent was made up by adding 250 uL of 3-methoxybenzaldehyde IMBA). 5 ml of 10 m M 9-fluorenone solution (made up in propan-l-ol) and 3.75 ml of formic acid (minimum assay 88% purity) to a 50 ml volumetric flask and making up to 50 ml with propan-l-ol. The final solution consisted of 41.1 m M M B A . 1 m M 9-fluorenone and 7.5% formic acid (v/v). The solution was transferred to a dark bottle and kept in the dark when not in use. In order to deproteinize the plasma samples, plasma was mixed with an equal volume of the solvent (propan-l-oh. This mixture was vortexed vigorously and left to stand for 2 min. The suspension was diluted to twice the original plasma volume with HPLC-grade water and allowed to stand for 2 min. This solution was then centrifuged in a microcentrifuge for 12 700^ for 2 min. The resulting supernate, containing INH. acetylhydrazine and hydrazine, was filtered to remove any remaining large particles using a 3-mm diameter syringe filter (Gelman). The filtered supernate was mixed with the derivatizing reagent at a ratio of 4:1. This mixture was allowed to stand at room temperature for 2 h in the dark. The acid present in the derivatizing reagent was sufficient to hydrolyze any j-ketoacid hydrazones of INH and acetylhydrazine and j-ketoacid azines of hydrazine present in plasma (Timbrell and Wright 1984). The INH. acetylhydrazine arid hydrazine released in these reactions then reacted with the large excess of M B A reagent to form isonicotinyl-j-methoxybenzaldhydrazone. N-acetyl-J-methoxyben-zaldhydrazone and 3-methoxybenzaldiazine. respectively. The unreactive 9-fluorenone served as an internal check of volumetric accuracy. I N H . acetylhydrazine and hydrazine were analysed using an H P L C column (4.6x125 mm) packed with Spherisorb ODS-2 (Phase-Separations Ltd.) and a Spectra-Physics SP8000B liquid chromatograph (Santa Clara. Calif.. USA) fitted with a Spectra-Physics SP8400 U V vis detector set at 300 nm. The oven temper-ature was 4 S C and the flow rate was 1 ml min. Mobile phase A consisted of 20% M e C N 10% M e O H solution (unbuffered, for cleaning column I. Mobile phase B consisted of 20% M e C N 10% M e O H and 5 mM sodium acetate adjusted to pH 5.0 with acetic acid. Mobile phase C consisted of 40% M e C N 40% M e O H and 5 m M sodium acetate adjusted to pH 5.0 with acetic acid. 253 All eluents were filtered before use and degassed in situ. The gradient elution program for this separation was as follows: Time (min) % B %c 0 100 0 5 60 40 10 0 100 16.66 0 100 17.5 100 0 The calibration curve of I N H concentration (umol/I) versus the ratio of I N H to 9-fluorenone resulted in a correlation coefficient (/•) of 0.997 and a slope of 0.050. The calibration curve of acetylhydrazine concentration (umol/1) versus the ratio of acetylhydrazine to 9-fluorenone resulted in r = 0.999 and a slope of 0.018. The calibration curve of hydrazine concentration (umol.l) versus the ratio of hy-drazine to 9-fluorenone resulted in r = 0.992 and a slope of 0.074. log plasma ASAL 48 hour Fig. 2 Correlation of 32 h plasma acetylhydrazine concentration (uM) with log plasma A S A L activity at 48 h llog Takahara units). The correlation is not significant (r- = 0.03, P > 0.50. n = 15) Statistics Parametric correlations (Pearson's product moment correlation) were used for data analysis. Correlation coefficients were converted to coefficients of determination in order to more accurately deter-mine the biological significance of the correlations (considered sig-nificant when r1 > 0.50). Al l data are presented as mean + standard error. Results Plasma INH concentration at 32 h did not correlate with plasma ASAL activity as A U C (r 2=0.l4, P > 0.10). Plasma INH concentration at 32 h correlated significantly with log plasma ASAL activity at 48 h (r1 = 0.29, P <0.05) and log peak plasma ASAL activ-ity (r- = 0.36, P<0.02); however, these correlations accounted for less than 50% of the total variance. Plasma INH at 48 h did not correlate with any marker of hepatic necrosis. 25 SO 75 100 125 Plasma ASAL area under the curve Fig. 1 Correlation of 32 h plasma acetylhydrazine concentration IULM) with plasma A S A L activity as A U C . ' T h e correlation is not significant (r: = 0.02, P > 0.50. n= 15) (ASAL Argininosuccinic acid lyase. AUC area under the curve I 120 0 25 50 75 100 12S 1 SO Plasma ASAL area under the curve Fig. 3 Correlation of 32 h plasma hydrazine concentration (nM) with plasma A S A L activity as A U C . The correlation is significant with a positive slope tr2 = 0.54. P < 0.002. n = 15l Plasma acetylhydrazine concentration at 32 h did not correlate with plasma ASAL activity as A U C (r : = 0.02, P > 0.50; Fig. 1), log plasma ASAL activity at 48 h (r2 = 0.03, P > 0.50; Fig. 2) or log peak plasma ASAL activity (r2 = 0.06, P > 0.50). Plasma acetylhydr-azine at 48 h also did not correlate with any marker of hepatic necrosis. Plasma hydrazine concentration at 32 h correlated sianificantlv with plasma ASAL activity as AUC (r- = 0.5~4, P < 0.002; Fig. 3), log plasma ASAL activity at 48 h (r : = 0.53, P < 0.005; Fig. 4 and log peak plasma ASAL activity (>:=0.41. P<0.02). Plasma hydrazine at 48 h did not correlate with any marker of hepatic necrosis. Liver triglyceride accumulation correlated signific-antly with 48' h plasma INH (r2 = 0.29, P < 0.05) and plasma hydrazine concentrations (r2 = 0.34, P < 0.05). However, neither of these correlations explain more than 50% of the total variance. There were no signifi-cant correlations between plasma triglycerides (peak plasma triglycerides, plasma triglycerides at 48 h and plasma triglyceride area under the curve) and 32 or 48 h 254 0.50 0.75 100 1 25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 ' log Plasma ASAL 48 hour Fig. 4 Correlation of 32 h plasma hvdrazine concentration (uM) with log plasma A S A L activity at 48 h (log Takahara unitsl. The correlation is significant with a positive slope | r : = 0 5" P < 0 005 n = I5) . . . plasma INH, acetylhydrazine and hydrazine concen-trations (r2 ranging from 0.01 to 0.24). Discussion Although plasma INH levels at 32 h correlated signific-antly with log plasma ASAL activity at 48 h and log peak plasma ASAL activity, the correlations were weak and accounted for much less than 50% of the total variance. It has been reported in humans that, plasma INH levels do not correlate with susceptibility to INH-induced hepatotoxicity (Mitchell et al. 1975). The data from the present study support these previous findings in humans. Both acetylhydrazine and hydrazine are proven metabolites of INH in humans (Timbrell et al. 1977; Noda et al. 1978: Blair et al. 1985; Peretti et al. 1987; Gent et al. 1992) and rabbits (Thomas et al. 1981). In human slow acetylators, the potential for accumulation of acetylhydrazine and hydrazine in plasma was recog-nised, as these metabolites were detected in urine be-tween 24 and 36 h after a single 300 mg oral dose of INH (Peretti et al. 1987). Evidence has suggested that acetylhydrazine is the hepatotoxic INH-metabolite in rats (Snodgrass et al. 1974; Mitchell et al. 1976; Tim-brell et al. 1980). If this were also the case in the rabbit model, one would expect to observe a significant posit-ive correlation between hepatic necrosis and the amount of acetylhydrazine formed as reflected in the 32 or 48 h plasma acetylhydrazine concentration. This was not the case in the present study, however, as there were no significant correlations between the hepatic necrosis indicators and plasma acetylhydrazine. These results are in concordance with other data showing that acetylation of hydrazine compounds is a detoxification reaction (McQueen et al. 1982). The main positive finding of the present study is the significant correlation between markers of hepatic nec-rosis and the 32 h plasma hydrazine concentration. This suggests that the amount of hydrazine formed from INH and or acetylhydrazine is an important de-terminant in the cause of INH-induced hepatic nec-rosis. Hydrazine has been previously implicated as the hepatotoxic metabolite of INH. Hydrazine is a known hepatotoxin in rats (Scales and Timbrell 1982; Timbrell et al. 1982: Jenner and Timbrell 1994), rabbits (Yard and McKennis 1955; McKennis et al. 1956; Noda et al. 1983) and monkeys (Patrick and Back 1965). Although much of the research on hydrazine hepatotoxicity has focused on hydrazine-induced fatty changes in the liver, hydrazine has also been shown to cause hepatic nec-rosis (Patrick and Back 1965: Noda et al. 1983). Pro-longed elimination of hydrazine after administration of INH results in a slowly rising plasma baseline of hydra-zine, which is exaggerated in some patients; one patient with the highest plasma levels of hydrazine had elevated plasma bilirubin and transaminases (Gent et al. 1992). The progressive accumulation of hydrazine has been shown to occur in patients over a period of at least 6 weeks of treatment. This could explain the delayed appearance of INH hepatotoxicity. which most often occurs within the first 8 weeks of daily INH therapy (Scharer and Smith 1969) but which can occur 24 weeks or longer into INH therapy (Byrd et al. 1972). In addition, greatly elevated plasma hydrazine levels were reported in a case of fatal INH-induced hepatotoxicity in a 74-year old man (Woo et al. 1992). Another prominent feature of INH-induced hepa-totoxicity in rabbits, unlike the scenario in man, is hepatic triglyceride accumulation (hepatic steatosis: McKennis et al. 1956: Whitehouse et al. 1983; Kar-thikeyan and Krishnamoorthy 1991). Although liver triglyceride accumulation showed significant correla-tions with plasma INH and hydrazine concentrations at 48 h. less than 50% of the total variance was ex-plained by these correlations. Unfortunately, these re-sults do not provide any clues as to which of INH, acetylhydrazine or hydrazine is/are causative in INH-induced hepatic steatosis. The data support previous findings that INH-induced hepatic necrosis and steato-sis appear to develop independently of each other (Sarich et al. 1993). One of the limitations of the present study is the lack of time-course data for hepatic triglyceride accumula-tion. However, plasma triglyceride levels can be used to estimate the time-course of hepatic triglyceride accu-mulation. Administration of INH to rabbits leads to decreased triglyceride levels in adipose tissue and in-creased triglyceride content of the liver (Karthikeyan and Krishnamoorthy 1991). In rats, there is evidence that steatosis induced by hydrazines involves elevated serum free fatty acids, increased uptake of free fatty acid by the liver, increased hepatic triglyceride synthesis and decreased lipoprotein secretion (Lamb and Banks 1979). Since the INH-induced elevation of plasma 255 triglyceride concentrations precedes accumulation of hepatic triglycerides, a correlational analysis of plasma triglycerides with plasma levels of INH, acetylhydr-azine and hydrazine was performed. The lack of cor-relation of plasma triglyceride levels with plasma levels of INH, acetylhydrazine and hydrazine suggest that these compounds do not directly cause hypertrig-lyceridaemia. The lack of information concerning the actual time-course of liver triglyceride accumulation prevented closer and potentially more definitive con-clusions regarding the role of INH, acetylhydrazine and/or hydrazine in the production of elevated liver triglyceride accumulation. In conclusion, neither INH nor acetylhydrazine plasma levels are correlated with markers of INH-in-duced hepatic necrosis. However, a strong correlation exists between hydrazine and markers of INH-induced hepatic necrosis. This study did not implicate INH, acetylhydrazine or hydrazine in INH-induced fatty changes including hepatic steatosis and/or hypertrig-lyceridaemia. The in vivo metabolic profile of INH, acetylhydrazine and hydrazine during active INH-in-duced hepatotoxicity in rabbits indicates that hydra-zine is most likely involved in the mechanism of hepatic necrosis. Acknowledgements The authors would like to thank Dr. David V. Godin. Professor. Department of Pharmacology & Therapeutics, for helpful discussions. This research was partially supported by pro-gram project grant G M 32165 from the National Institutes of Health, Bethesda. M D . T.C. Sarich was partially funded by a Sum-mer Research Scholarship in Medicine from the Health Research Foundation of the Pharmaceutical Manufacturers Association of Canada and the Medical Research Council of Canada. References Barlow PB. Black M. Brummcr DL. Comstock G W . Dubin IN. Enterline P. Gibson M L . Hardv G E Jr. Harrel JA. Johnston RF. Kent D C . Marvin 8A. McCai'g NC. Mitchell JR. Mosley JW. Ogasawara FR. Popper H. Reichman LB. Zimmerman HJ1197-1 > Preventive therapv of tuberculous infection. Am Rev Resp Dis 110:371-374 Blair l.A. Mansilla Tinoco R. Brodie MJ . Clare RA. Dollery CT. Timbrell JA. Beever (A (1985) Plasma hydrazine concentrations in man after isoniazid and hydralazine administration. Hum Toxicol 4: 195-202 Black M. Mitchell JR. Zimmerman HJ. Ishak K G . EpIerGR 119 o i Isoniazid associated hepatitis in 114 patients. Gastroenteroloa\ 69: 289-302 Bradford M 119761 A rapid and sensitive method for the quantita-tion of microgram quantities of protein utilizing the principle of protein-dye-binding. Anal Biochem 72: 248-254 Byrd RB. Nelson R. Elliot RC 11972) Isoniziad toxicity - a prospect-ive studv in secondary chemoprophvlaxis. 1 Am Med Assoc 220:1471-1473 A preliminary report of this data has been published in abstract form: Sarich TC. Adams SP. Yousseri M. Zhou T. Wriuht J M 11904, Isoniazid-induced hepatotoxicity in rabbits: role of hydrazine Can J Physiol Pharmacol 72 (Suppl H: 5,-55 Campanini RZ. Tapia RA . Sarnat W, Natelson S (1970) Evaluation of serum argininosuccinate (ASAL) concentrations as an index to parenchymal liver disease. C l in Chem 16: 44-53 Folch J. Ascoli 1. Lees M . Meath JA. LeBaron F N (1951) Preparation of lipid extracts from brain tissue. J Biol Chem 191: 333-841 Gent WL . Seifart HI. Parkin D P . Donald PR. Lamprecht J H (1992) Factors in hydrazine formation from isoniazid by paediatric and adult tuberculous patients. Eur J Cl in Pharmacol 43: 131-136 Jenner A M . Timbrell JA 11994) Influence of inducers and inhibitors of cytochrome P450 on the hepatoxicity of hydrazine in vivo. Arch Toxicol 68: 349-357 Karthikeyan S. Krishnamoorthy M S (1991) Effect of subacute ad-ministration of isoniazid and pyridoxine on lipids in plasma, liver, and adipose tissues in the rabbit. Drug Chem Toxicol 14: 293-303 Lamb R G . Banks W L 11979) Effect of hydrazine exposure on hepatic triacylglycerol biosynthesis. Biochim Biophys Acta 574: 440-447 McKennis H Jr. Yard AS. Pahnelas EV (1956) The production of fatty livers in rabbits by isoniazid and other hydrazine deriva-tives. A m Rev Tuberc 73: 956-959 McQueen C A . Maslansky C J . Glowinski IB. Crescenzi SB, Weber WW. Williams G M (1982) Relationship between the genetically determined acetylator phenotype and D N A damage induced by hydralazine and 2-aminofluorene in cultured rabbit hepatocytes. Proc Natl Acad Sci U S A 79: 1269-1272 Mitchell JR. Long M W . Thorgeirsson UP. Jollow DJ 11975) Acetyla-tion rates and monthly liver function tests during one year of isoniazid preventive therapy. Chest 68: 181-190 Mitchell JR. Zimmerman H J . Ishak K G . Thorgeirsson U P , Timbrell J.A. Snodgrass WR. Nelson SD (1976). Isoniazid liver injury-clinical spectrum, pathology and probable pathogenesis. A n n Intern Med 84: 181-192 Nelson SD. Mitchell JR. Timbrell JA. Snodgrass W R . Corcoran G B (1976) Isoniazid and iproniazid: activation of metabolites to toxic intermediates in man and rat. Science 193: 901-903 Noda A. Goromaru T. Matsuyama K. Sogabe K. Hsu K-Y, Iguchi S (1978) Quantitative determination of hydrazines derived from isoniazid in patients. I. 1 Pharmacol Dyn 1: 132-141 Noda A. Hsu K-Y. Noda H. Yamamoto Y. Kurozumi T(19S3l Is isoniazid-hepatotoxicitv induced bv the metabolite hydrazine? J L ' O E H 5: 183-190 Patrick RL . Back 119651 Pathology and toxicology of repeated Joscs of hydrazine and 1.1 -dimethylhydrazine • " monkeys and rats. Ind Med Surg 54: 450-435 Peretti E. Karlagams G. Lauterburg BH M987) Increased urinary excretion of toxic hydrazino metabolites of isoniazid by slow-acetylators. Effect of a slow-release preparation of isoniazid. Eur J Clin Pharmacol 33: 2S3-2S6 Sarich TC. Adams SP. Zhou T. Wright J M 11993) Isoniazid-induced hepatotoxicity in rabbits: a comparison of steatosis and necrosis. Clin Invest Med 16: B67 Sarich TC. Zhou T. Adams SP. Bain A l . Wall RA. Wright J M 11995). A model of isoniazid-induced hepatotoxicity in rabbits. J Pharm-acol Toxicol Methods 34: 109-116 Scales M D C . Timbrell JA 119S2i Studies on hydraz:.-.e hepatotoxic-ity: 1. Pathological rinding*. J Toxicol Environ Health 10: 941-953 Scharer L. Smith J P 11969) Serum transaminase elev auons and other hepatic abnormalities in patients receivins isoniazid. Ann Intem Med 71: 1113-1120 Sims F H . Rautanen P 11975) Serum arginino-succinate lyase: observ-ations on the sensitivity and specificity of this test in the detection of minimal hepatocellular damage. Cl in Biochem 8: 213-221 Snodgrass W. Potter WZ. Timbrell 1 Jollow DJ. Mitchel l JR (1974) Possible mechanism of isoniazid-rclated hepatic injury. Clin Res 1 Abstr 1 22: 323A Thomas B H . Wong LT. Zeitz W. Solomonraj G i !98l) Isoniazid metabolism in the rabbit, and (he effect of rifampin pretreatment. Res Commun Chem Pathol Pharmacol 33: 235-247 256 Timbrell JA. Wright J M (1984) Urinary metabolic pronle of isoniazid in patients who develop isoniazid-related liver damage. Human Toxicol 3: 485-495 Timbrell JA, Wright JM, Baillie TA (1977) Monoacetylhydrazine as a metabolite of isoniazid in man. Clin Pharmacol Ther 22: 602-608 Timbrell JA. Mitchell JR, Snodgrass WR, Nelson SD (1980) Isoniazid hepatotoxicity: the relationship between covalent binding and metabolism in vivo. J Pharmacol Exp Ther 213: 364-369 Timbrell JA, Scales M D C , Streeter AJ (1982) Studies on hydrazine hepatotoxicity: 2. Biochemical findings. J Toxicol Environ Health 10: 955-968 Touster O, Aronson N N Jr, Dulaney JT, Hendrickson H (1970) Isolation of rat liver plasma membranes. J Cell Biol 47: 604-618 Whitehouse LW, Tryphonas L, Paul CJ, Solomonraj G, Thomas BH, Wong LT (1983) Isoniazid-induced hepatic steatosis in rab-bits: an explanation for susceptibility and its antagonism by pyridoxine hydrochloride. Can J Physiol Pharmacol 61:478-487 Woo J, Chan CHS, Walubo A, Chan K.KC (1992) Hydrazine: a possible cause of isoniazid-induced hepatic necrosis. J Med 23: 51-59 Yard AS, McKennis H Jr (1955) Effect of structure on the ability of hydrazino compounds to produce fatty livers. J Pharmacol Exp Ther 114: 391-397 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088330/manifest

Comment

Related Items