Drug-Induced Mitochondrial Dysfunction

DRUG-INDUCED MITOCHONDRIAL DYSFUNCTION Edited by JAMES A. DYKENS Pﬁzer, Inc. Sandwich, UK YVONNE WILL Pﬁzer, Inc. Gro...

Author: James A. Dykens | Yvonne Will

254 downloads 2103 Views 8MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

DRUG-INDUCED MITOCHONDRIAL DYSFUNCTION

Edited by JAMES A. DYKENS Pﬁzer, Inc. Sandwich, UK

YVONNE WILL Pﬁzer, Inc. Groton, Connecticut

A JOHN WILEY & SONS, INC., PUBLICATION

DRUG-INDUCED MITOCHONDRIAL DYSFUNCTION

DRUG-INDUCED MITOCHONDRIAL DYSFUNCTION

Edited by JAMES A. DYKENS Pﬁzer, Inc. Sandwich, UK

YVONNE WILL Pﬁzer, Inc. Groton, Connecticut

A JOHN WILEY & SONS, INC., PUBLICATION

Copyright  2008 by John Wiley & Sons, Inc. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and speciﬁcally disclaim any implied warranties of merchantability or ﬁtness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of proﬁt or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Drug-induced mitochondrial dysfunction / [edited by] James A. Dykens, Yvonne Will. p. ; cm. Includes bibliographical references and index. ISBN 978-0-470-11131-4 (cloth) 1. Drugs—Toxicology. 2. Mitochondrial pathology. I. Dykens, James A. II. Will, Yvonne. [DNLM: 1. Mitochondrial Diseases—chemically induced. 2. Mitochondria—drug effects. 3. Mitochondria—physiology. WD 200.5.M6 D794 2008] RA1238.D783 2008 615 .1—dc22 2008002720 Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

CONTENTS

CONTRIBUTORS PREFACE

PART I 1

ix xiii

BASIC CONCEPTS

Basic Mitochondrial Physiology in Cell Viability and Death

1 3

Lech Wojtczak and Krzysztof Zabłocki

2

Basic Molecular Biology of Mitochondrial Replication

37

Immo E. Schefﬂer

3

Drug-Associated Mitochondrial Toxicity

71

Rhea Mehta, Katie Chan, Owen Lee, Shahrzad Tafazoli, and Peter J. O’Brien

4

Pharmacogenetics of Mitochondrial Drug Toxicity

127

Neil Howell and Corinna Howell

PART II

5

ORGAN DRUG TOXICITY: MITOCHONDRIAL ETIOLOGY

Features and Mechanisms of Drug-Induced Liver Injury

141 143

Dominique Pessayre, Alain Berson, and Bernard Fromenty

6

Cardiovascular Toxicity of Mitochondrial Origin

203

Paulo J. Oliveira, Vilma A. Sard˜ao, and Kendall B. Wallace

v

vi

7

CONTENTS

Skeletal Muscle and Mitochondrial Toxicity

235

Timothy E. Johnson

8

Manifestations of Drug Toxicity on Mitochondria in the Nervous System

251

Ian J. Reynolds

9

Lipoatrophy and Other Manifestations of Antiretroviral Therapeutics

273

Ulrich A. Walker

10 Nephrotoxicity

291

Alberto Ortiz, Alberto Tejedor, and Carlos Caramelo

11 Drug Effects in Patients with Mitochondrial Diseases

311

Eric A. Schon, Michio Hirano, and Salvatore DiMauro

PART III

ASSESSMENT OF MITOCHONDRIAL FUNCTION IN VITRO AND IN VIVO

325

12 Polarographic Oxygen Sensors, the Oxygraph, and High-Resolution Respirometry to Assess Mitochondrial Function 327 Erich Gnaiger

13 Use of Oxygen-Sensitive Fluorescent Probes for the Assessment of Mitochondrial Function

353

James Hynes, Tom´as C. O’Riordan, and Dmitri B. Papkovsky

14 Mitochondrial Dysfunction Assessed Quantitatively in Real Time by Measuring the Extracellular Flux of Oxygen and Protons

373

David Ferrick, Min Wu, Amy Swift, and Andy Neilson

15 Assessment of Mitochondrial Respiratory Complex Function In Vitro and In Vivo

383

Mark A. Birch-Machin

16 OXPHOS Complex Activity Assays and Dipstick Immunoassays for Assessment of OXPHOS Protein Levels

397

Sashi Nadanaciva

17 Use of Fluorescent Reporters to Measure Mitochondrial Membrane Potential and the Mitochondrial Permeability Transition 413 Anna-Liisa Nieminen, Venkat K. Ramshesh, and John J. Lemasters

CONTENTS

vii

18 Compartmentation of Redox Signaling and Control: Discrimination of Oxidative Stress in Mitochondria, Cytoplasm, Nuclei, and Endoplasmic Reticulum 433 Patrick J. Halvey, Jason M. Hansen, Lawrence H. Lash, and Dean P. Jones

19 Assessing Mitochondrial Protein Synthesis in Drug Toxicity Screening

463

Edward E. McKee

20 Mitochondrial Toxicity of Antiviral Drugs: A Challenge to Accurate Diagnosis

473

Michel P. de Baar and Anthony de Ronde

21 Clinical Assessment of Mitochondrial Function via [13 C]Methionine Exhalation 493 Laura Milazzo

22 Assessment of Mitochondrial Dysfunction by Microscopy

507

Ingrid Pruimboom-Brees, Germaine Boucher, Amy Jakowski, and Jeanne Wolfgang

23 Development of Animal Models of Drug-Induced Mitochondrial Toxicity

539

Urs A. Boelsterli and Yie Hou Lee

24 Noninvasive Assessment of Mitochondrial Function Using Nuclear Magnetic Resonance Spectroscopy 555 Robert W. Wiseman and J. A. L. Jeneson

25 Targeting Antioxidants to Mitochondria by Conjugation to Lipophilic Cations

575

Michael P. Murphy

INDEX

589

CONTRIBUTORS

Alain Berson, INSERM, Centre de Recherche Biom´edicale Bichat Beaujon, ´ Equipe Mitochondries et Foie, Paris; Facult´e de M´edecine Xavier Bichat, Universit´e Paris 7, Paris, France Mark A. Birch-Machin, Dermatological Sciences, Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne, UK Urs A. Boelsterli, Department of Pharmaceutical Sciences, University of Connecticut School of Pharmacy, Storrs, Connecticut Germaine Boucher, Drug Safety R&D, Pﬁzer, Inc., Groton, Connecticut Carlos Caramelo, Fundaci´on Jimenez Diaz, Madrid, Spain Katie Chan, Graduate Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ontario, Canada Michel P. de Baar, Primagen, Amsterdam, The Netherlands; currently at OctoPlus N. V., Leiden, The Netherlands Anthony de Ronde, Primagen, Amsterdam, The Netherlands Salvatore DiMauro, Columbia University Medical Center, New York, New York David Ferrick, Seahorse Bioscience, Billerica, Massachusetts Bernard Fromenty, INSERM, Centre de Recherche Biom´edicale Bichat Beau´ jon, Equipe Mitochondries et Foie, Paris; Facult´e de M´edecine Xavier Bichat, Universit´e Paris 7, Paris, France Erich Gnaiger, Department of General and Transplant Surgery, D. Swarovski Research Laboratory, Medical University of Innsbruck, Innsbruck, Austria; OROBOROS INSTRUMENTS, Innsbruck, Austria Patrick J. Halvey, Department of Pediatrics, and Division of Pulmonary, Allergy and Critical Care Medicine, Department of Medicine, Emory ix

x

CONTRIBUTORS

University, Atlanta, Georgia; Department of Biochemistry and National Centre for Biomedical Engineering Science, National University of Ireland, Galway, Ireland Jason M. Hansen, Department of Pediatrics, Emory University, Atlanta, Georgia Michio Hirano, Columbia University Medical Center, New York, New York Corinna Howell, Matrilinex LLC, San Diego, California Neil Howell, MIGENIX Corporation, San Diego, California; currently at Matrilinex LLC, San Diego, California James Hynes, Luxcel Biosciences Ltd., BioInnovation Centre, University College–Cork, Cork, Ireland Amy Jakowski, Drug Safety R&D, Pﬁzer, Inc., Groton, Connecticut J. A. L. Jeneson, Biomedical NMR Laboratory, Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands Timothy E. Johnson, Department of Safety Assessment, Merck Research Laboratories, West Point, Pennsylvania Dean P. Jones, Division of Pulmonary, Allergy and Critical Care Medicine, Department of Medicine, Emory University, Atlanta, Georgia Lawrence H. Lash, Department of Pharmacology, Wayne State University School of Medicine, Detroit, Michigan Owen Lee, Graduate Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ontario, Canada Yie Hou Lee, Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore John J. Lemasters, Center for Cell Death, Injury and Regeneration, Charleston, South Carolina; Departments of Pharmaceutical and Biomedical Sciences and Biochemistry and Molecular Biology, Hollings Cancer Center, Medical University of South Carolina, Charleston, South Carolina Edward E. McKee, Indiana University School of Medicine–South Bend, South Bend, Indiana Rhea Mehta, Graduate Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ontario, Canada Laura Milazzo, Institute of Infectious and Tropical Diseases, University of Milan, L. Sacco Hospital, Milan, Italy Michael P. Murphy, MRC Dunn Human Nutrition Unit, Wellcome Trust, Cambridge, UK

CONTRIBUTORS

xi

Sashi Nadanaciva, MitoSciences Inc., Eugene, Oregon; currently at Pﬁzer, Groton, Connecticut Andy Neilson, Seahorse Bioscience, Billerica, Massachusetts Anna-Liisa Nieminen, Center for Cell Death, Injury and Regeneration, Charleston, South Carolina; Department of Pharmaceutical and Biomedical Sciences, Hollings Cancer Center, Medical University of South Carolina, Charleston, South Carolina Peter J. O’Brien, Graduate Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ontario, Canada Paulo J. Oliveira, Center for Neurosciences and Cell Biology, Department of Zoology, University of Coimbra, Coimbra, Portugal Tom´as C. O’Riordan, Luxcel Biosciences Ltd., BioInnovation Centre, University College–Cork, Cork, Ireland Alberto Ortiz, Fundaci´on Jimenez Diaz, Madrid, Spain Dmitri B. Papkovsky, Biochemistry Department, University College–Cork, Cork, Ireland Dominique Pessayre, INSERM, Centre de Recherche Biom´edicale Bichat Beau´ jon, Equipe Mitochondries et Foie, Paris; Facult´e de M´edecine Xavier Bichat, Universit´e Paris 7, Paris, France Ingrid Pruimboom-Brees, Drug Safety, GlaxoSmithKline, Ware, Hertfordshire, UK; currently at Pﬁzer, Inc., Sandwich, UK Venkat K. Ramshesh, Center for Cell Death, Injury and Regeneration, Charleston, South Carolina; Department of Pharmaceutical and Biomedical Sciences, Hollings Cancer Center, Medical University of South Carolina, Charleston, South Carolina Ian J. Reynolds, Neuroscience Drug Discovery, Merck Research Laboratories, West Point, Pennsylvania Vilma A. Sard˜ao, Center for Neurosciences and Cell Biology, Department of Zoology, University of Coimbra, Coimbra, Portugal Immo E. Schefﬂer, Section of Molecular Biology, Division of Biological Sciences, University of California–San Diego, La Jolla, California Eric A. Schon, Columbia University Medical Center, New York, New York Amy Swift, Seahorse Bioscience, Billerica, Massachusetts Shahrzad Tafazoli, Graduate Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ontario, Canada Alberto Tejedor, Hospital Gregorio Mara˜non, Madrid, Spain

xii

CONTRIBUTORS

Ulrich A. Walker, Department of Rheumatology, Basel University, Basel, Switzerland Kendall B. Wallace, Department of Biochemistry and Molecular Biology, University of Minnesota Medical School, Duluth, Minnesota Robert W. Wiseman, Biomedical Imaging Research Center, Departments of Physiology and Radiology, Michigan State University, East Lansing, Michigan Lech Wojtczak, Nencki Institute of Experimental Biology, Warsaw, Poland Jeanne Wolfgang, Drug Safety R&D, Pﬁzer, Inc., Groton, Connecticut Min Wu, Seahorse Bioscience, Billerica, Massachusetts Krzysztof Zabłocki, Nencki Institute of Experimental Biology, Warsaw Poland

PREFACE

According to the U.S. Food and Drug Administration (FDA), there are some 2.2 million adverse drug responses (ADRs) in hospitalized patients in the United States every year [1]. Incidence in outpatient populations is unknown, but ADRs account for approximately 100,000 deaths annually, making ADRs the fourth-leading cause of death, ahead of pulmonary disease, diabetes, AIDS, pneumonia, and accidents, including automobile deaths [2,3]. These numbers suggest that not all sources of iatrogenic drug toxicity have been identiﬁed, and this book summarizes the rapidly expanding literature indicting “off-target” mitochondrial impairment as a major contributor to drug toxicity. Over 75 diseases and metabolic disorders are caused by mitochondrial dysfunction, providing a priori evidence that mitochondrial impairment can yield deleterious consequences. Many of these syndromes arise from mutations or deletions in the mitochondrial genome (mtDNA), and others are caused by mutations in proteins encoded by nuclear DNA (nDNA) but destined for import into mitochondria. Many are characterized by multiorgan involvement, frequently targeting aerobically poised tissues such as central nervous system, cardiovascular, sensory, and motor axes. Moreover, the pharmacopeia of well-known mitochondrial inhibitors, many of which are potent poisons, underscores the importance of mitochondrial integrity. Among these are many natural products, such as oligomycin, antimycin, and cyclosporine, and 60 classes of xenobiotics are known to inhibit the mitochondrial respiratory complex I alone. In this light, it should not be surprising that ethical pharmaceuticals could be capable of inducing mitochondrial dysfunction that leads to cytotoxicity and organ pathology. However, appreciation of drug-induced mitochondrial dysfunction has only recently gained momentum, fostered in large measure by the organ toxicities and metabolic syndromes associated with long-term antiretroviral therapies for HIV infection. Collateral inhibition of mitochondrial replication is widely recognized as the cause of these adverse events, and the pharmaceutical industry now generally screens for this potential effect [4]. xiii

xiv

PREFACE

More recently, direct drug effects on mitochondrial function, such as inhibition of respiration or uncoupling of electron transport from phosphorylation, have been described after acute exposures. Importantly, the extent of mitochondrial impairment is in accord with the clinical disposition of many of these drugs. For example, of the thiazolidinediones, an important class of insulin sensitizers used to treat type 2 diabetes, troglitazone and ciglitazone are among the most potent respiratory inhibitors, and they were withdrawn from the market or discontinued during clinical trials, respectively. Darglitazone and muraglitazar are less potent mitotoxins, and were discontinued prior to release because of organ toxicity. In contrast, pioglitazone and rosiglitazone have much less effect on mitochondrial function, and they are associated with substantially less organ toxicity, although increased risk of myocardial infarction and congestive heart failure caused the FDA to issue a black box warning for both in 2007 [5,6]. Just how important is mitochondrial functional integrity? Human resting metabolism varies with gender but averages 6127 and 7983 kJ/day in women and men, respectively [7]. Under physiological conditions, ATP hydrolysis yields 42 to 50 kJ/mol [8], so women turn over about 133 and men about 173 mol ATP/day. The molecular weight of ATP is 507 g/mol, so women turn over 67,431 g/day (148 lb) and men 87,711 g/day (193 lb). We essentially turn over our body weight in ATP every day, and this is just for resting metabolism; humans have aerobic scope (i.e., can increase activity) of between 10 to 20-fold, so a well-trained marathon runner could turn over more than 1 kg of ATP per minute! Given the selectivity and potency of many of these drugs and the importance of oxidative phosphorylation (OXPHOS) for ATP generation, why are drug-induced mitochondrial dysfunction and organ toxicity not more widespread adverse events? We suspect that the variables are not the drug effects on mitochondrial function, which are likely to be consistent, but rather, the previous organ history and genetics, which together establish a threshold of vulnerability. All cells have physiological scope (i.e., the ability to accelerate metabolic processes), and in all aerobically poised cells, mitochondrial OXPHOS capacity exceeds bioenergetic demand. When stressed, or during exercise, cells can substantially increase ATP turnover with impunity, up to a ﬁnite maximum. The reciprocal also holds; cells require a minimum level of ATP production to maintain functional integrity. However, as mitochondrial capacity is eroded by drug exposure, the scope between ATP demand and production diminishes, and at some point a bioenergetic threshold is crossed where ATP production is insufﬁcient to maintain viability, imperiling the cell and organ. In most mammals, this physiological scope is quite large; as noted above, human metabolism can increase 10 to 20-fold during sustained aerobic exercise. All else being equal, this suggests that mitochondrial capacity, at least in skeletal muscle, needs to be diminished 10-fold before minimal energy requirements are endangered. Mitochondrial complements in organs not subject to conditioning, such as liver, kidney, and the central nervous system, are more ﬁxed, so that bioenergetic thresholds in these tissues are dictated by genetics and by organ

PREFACE

xv

history. A marathon runner has wider aerobic scope than a sedentary scientist, and a liver exposed to years of alcohol abuse is less metabolically resilient than a liver in a younger, drug-naive person. In this way, most adverse drug effects on mitochondria remain latent until sufﬁciently severe to cross the bioenergetic threshold when the cell can no longer fuel metabolism or respond to stress, at which point the cell, and hence the organ, are imperiled. In this light, it is telling that pharmaceuticals are typically evaluated for potential organ toxicity in drug-naive, young, and perfectly healthy animals, precisely the circumstances where mitochondrial impairment is least likely to be detected. Moreover, almost all cell culture systems in contemporary use also fail to reveal mitochondrial dysfunction. It is no wonder that many pharmaceutical scientists remain skeptical about xenobiotic-induced mitochondrial impairment. Several animal and cell models have been developed to detect drug-induced mitochondrial impairment in the short-duration studies typical of preclinical drug evaluations. For example, the hepatotoxicity of troglitazone, which forced its market withdrawal, was not detected in preclinical animal models. But these were healthy animals with robust mitochondrial and antioxidant reserves. However, when the mitochondrial antioxidant Mn-SOD is knocked down by 50% in mice, liver toxicity by troglitazone is readily detected (see Chapter 24). Cells grown in culture are almost uniformly provided with 25 mM glucose, ﬁve times physiological levels. This allows the media to last several days before requiring replacement. However, under these conditions, most cells rely on aerobic glycolysis, producing lactate despite the presence of competent mitochondria. This was described independently in 1929 by Crabtree [9], who noted that respiration is inhibited by glucose, and by Warburg, who reported lactate production despite adequate respiratory capacity [10]. Cells not dependent on OXHOS for ATP are not susceptible to mitotoxicants, and they correspondingly fail to reveal drug-induced mitochondrial dysfunction [10]. To render cultured cells susceptible to mitochondrial impairment, glucose in the media can be replaced by galactose. Galactose requires the investment of 2 ATP equivalents for it to enter glycolysis, and since the latter yields only 2 ATPs, galactose as substrate produces zero net ATP. Under these conditions, to obtain ATP, cells are forced to use OXPHOS, and they correspondingly become susceptible to mitochondrial impairment [11]. To date, analysis of drug-induced mitochondrial dysfunction has been primarily retrospective, examining whether adverse events with unknown etiology could be due to off-target mitochondrial impairment. The picture that is emerging is that not all drug toxicity is via mitochondrial failure, but that drugs with mitochondrial liabilities have disproportionate numbers of potentially serious adverse events attributable directly to mitochondrial impairment. A retrospective analysis of more than 500 pharmaceutically relevant molecules indicates that about 35% directly and acutely impair mitochondrial function via inhibition of respiration and/or uncoupling electron transport from phosphorylation. This is in addition to those drugs that impair mtDNA replication or protein expression that yield long-term mitochondrial dysfunction and depletion. Moreover, mitochondrial ﬁssion and fusion that are required for long-term function are also likely targets

xvi

PREFACE

for disruption by xenobiotics. Clearly, we need to develop preclinical animal and cell models that faithfully predict mitochondrial impairment and resulting organ toxicity in the clinic. This book reﬂects the current understanding of drug-induced mitochondrial impairment as well as recent advances in models designed to detect it preclinically and clinically. We have endeavored to generate a text that provides (1) sufﬁcient basic information about normal mitochondrial function to introduce nonspecialists to the ﬁeld, (2) enough organ pathology to convince the reader that mitochondrial impairment is a legitimate concern for pharmacologists and clinicians, and (3) adequate introductions to techniques used to assess mitochondrial function so that researchers can address drug-induced mitochondrial impairment in their own labs. To that end, the book is organized in three parts: Chapters 1 to 4 cover basic mitochondrial physiology and replication; Chapters 5 to 11 cover various organ toxicities and drug toxicity in patients having mitochondrial diseases; and Chapters 12 to 25 are shorter chapters describing methods to assess mitochondrial dysfunction in vitro and in vivo. None of the chapters is designed to be comprehensive, and authors were encouraged to use examples from their own research. However, all chapters provide context and introduce the reader to the current state of affairs and the literature. We did not include chapters on drug metabolism and detoxiﬁcation, and genetic variation in drug metabolic pathways surely contributes to idiosyncratic drug responses, serving to amplify exposure or generate toxic metabolites in the susceptible individual. However, in many cases, the parent molecule is now recognized as a mitotoxicant, and no doubt there will be many cases where it is the metabolite that undermines mitochondrial function. We therefore focused narrowly on principles of drug-induced mitochondrial toxicity rather than metabolism, and readers interested in the latter are referred to the books by Coleman [12] and Caira [13], where this topic is expertly reviewed. By focusing on organ toxicities rather than toxicities of various drug classes, discussion of some drugs is reiterated in several chapters. But this underscores the notion that like mitochondrial diseases, drug-induced mitochondrial liabilities are expressed in diverse and sometimes unexpected ways. For example, the statins paradoxically yield rhabdomyolysis of anaerobically poised, fast-twitch muscle ﬁbers, sparing mitochondrially enriched, aerobically poised, slow-twitch ﬁbers, and myocardium. This is a likely consequence of the distribution of monocarboxylate transporter isoform 4, which bioaccumulates the statins in susceptible ﬁbers [14,15]. Moreover, the mitochondrial membrane potential, plus the plasma membrane potential, can bioaccumulate permeable molecules 10,000-fold over extracellular levels (see Chapter 17). We thank our colleagues who provided chapters despite other pressing responsibilities. Y.W. thanks her mentor, Don Reed, for introducing her to the ﬁeld of mitochondrial biology, and J.D. thanks his many mentors for their guidance and continuing support. We also thank our editors for their diligence, and Dr. Gregory Stevens at Pﬁzer, who steadfastly encouraged our efforts to make mitochondrial toxicity screening a routine part of discovery toxicology.

PREFACE

xvii

It is our hope that this book will help foster the development of drugs in which the risk-to-beneﬁt ratio will be overwhelmingly biased toward the latter.

REFERENCES 1. http://www.fda.gov/cder/drug/drugReactions/. 2. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA. 1998;279:1200–1205. 3. Gurwitz JH, Field TS, Avorn J, et al. Incidence and preventability of adverse drug events in nursing homes. Am J Med. 2000;109:87–94. 4. Guidance for Industry: Antiviral Product Development: Conducting and Submitting Virology Studies to the Agency. Washington, DC: US Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research; 2006. http://www.fda.gov/cder/guidance/7070fnl.pdf. 5. Nadanaciva S, Dykens JA, Bernal A, Capaldi RA, Will Y. Mitochondrial impairment by PPAR agonists and statins identiﬁed via immunocaptured OXPHOS complex activities and respiration. Toxicol Appl Pharmacol . 2007;223:277–287. 6. Dykens JA, Will Y. The signiﬁcance of mitochondrial toxicity testing in drug development. Drug Discov Today. 2007;12:777–785. 7. De Lorenzo A, Tagliabue A, Andreoli A, Testolin G, Comelli M, Deurenberg P. Measured and predicted resting metabolic rate in Italian males and females, aged 18–59 years. Eur J Clin Nutr. 2001;55:208–214. 8. Campbell NA. Biology, 3rd ed. San Francisco: Benjamin-Cummings; 1993:97–101. 9. Crabtree HG. Observations on the carbohydrate metabolism of tumours. Biochem J. 1929;23:536–545. 10. Warburg O. On the origin of cancer cells. Science. 1956;123:309–315. 11. Marroquin LD, Hynes J, Dykens JA, Jamieson JD, Will Y. Circumventing the Crabtree effect: replacing media glucose with galactose increases susceptibility of HepG2 cells to mitochondrial toxicants. Toxicol Sci. 2007;97:539–547. 12. Coleman M. Human Drug Metabolism: An Introduction. Hoboken, NJ: Wiley; 2005. 13. Caira MR. Drug Metabolism: Current Concepts. New York: Springer; 2006. 14. Westwood FR, Bigley A, Randall K, Marsden AM, Scott RC. Statin-induced muscle necrosis in the rat: distribution, development, and ﬁbre selectivity. Toxicol Pathol . 2005;33:246–257. 15. Nadanaciva S, Dykens JA, Bernal A, Capaldi RA, Will Y. Mitochondrial impairment by PPAR agonists and statins identiﬁed via immunocaptured OXPHOS complex activities and respiration. Toxicol Appl Pharmacol . 2007;223:277–287.

PART I BASIC CONCEPTS

1

1 BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH Lech Wojtczak and Krzysztof Zabłocki Nencki Institute of Experimental Biology, Warsaw, Poland

1. Introduction 1.1. Historical background 1.2. Morphology 1.3. Structure and compartmentation 2. Oxidative phosphorylation 2.1. General principles 2.2. Respiratory chain as a proton pump 2.3. Mitochondrial ATPase/ATP synthase and energy coupling 2.4. Coupling and uncoupling; reversed electron transport 2.5. Mitochondrial carriers 3. Production of reactive oxygen species 4. Calcium signaling 5. Mitochondria and cell death 6 Concluding remarks: mitochondria as a pharmacological target

3 3 4 6 7 7 8 12 15 17 17 23 25 31

1. INTRODUCTION 1.1. Historical Background Mitochondria were described by histologists and cytologists during the second half of the nineteenth century as minute intracellular granules of various sizes and Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

3

4

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

shapes. The discovery by Leonor Michaelis in 1898 that they could be stained with the reduction–oxidation dye Janus green was probably the ﬁrst indication that they were sites of intracellular redox processes. This notion was reinforced by an observation of Otto Warburg, who found in 1913 oxygen consumption by a particulate fraction obtained from tissue dispersions (and he received a Nobel prize in 1931). The name mitochondrion was coined from two Greek terms, µιτoς (mitos, a thread) and χoνδριoν (khondrion, a grain), which best characterized the microscopic appearance of these structures. However, it was not until Albert Claude isolated relatively pure mitochondria in substantial amounts from tissue homogenates by differential centrifugation (1940) that progress in elucidating their importance for cell functions accelerated. In the 1930s, when Hans Krebs showed that the tricarboxylic acid cycle localized to mitochondria, they were recognized as important chemical factories (he received a Nobel prize in 1953). The following years saw the discovery of oxidative phosphorylation, and mitochondria emerged as the main source of cellular ATP [1,2]. However, the modern concepts of oxidative phosphorylation (Nobel prize to Peter Mitchell in 1978) and molecular mechanisms of oxidative phosphorylation (Nobel prize to Paul D. Boyer and John E. Walker in 1997) were formulated later [2–4]. With the discovery of rotating ATPase/ATP synthase [5] (see Section 2), the era of fundamental discoveries related to mitochondria as cellular energy transformers seemed to come to an end. More recently, mitochondriology has undergone a renaissance and is now a focus of intense interest in a wide variety of life sciences. First, the discovery that defects in mitochondrial DNA (mtDNA) are the basis for mitochondrial diseases launched the new ﬁeld of mitochondrial medicine [6]. Second, mitochondria were recognized to play a major role in initiation and execution of programmed cell death (apoptosis) [7,8]. Among other new areas of study are the mitochondrial theory of aging [9] and mechanisms of intracellular signaling [10]. Mitochondria are also targets, either primary or secondary, of numerous therapeutic and toxic xenobiotics [11]. 1.2. Morphology Our knowledge of the inner structure of mitochondria is based primarily on electron microscopic examination of glutaraldehyde- and osmium tetroxide–ﬁxed preparations of whole cells or tissues and of isolated particles (see also Chapter 23). Mitochondria are composed of two membranes that separate two compartments, the intermembrane compartment and the inner compartment, ﬁlled with the mitochondrial matrix (Figure 1). The outer membrane is usually smooth and forms a boundary separating the mitochondrion from the cytosol. The inner membrane forms multiple invaginations into the matrix compartment, the cristae. Depending on the tissue, the density of cristae varies from quite scarce, as in liver mitochondria, to tightly packed, as in muscle mitochondria, where the inner compartments are densely ﬁlled with cristae. This usually correlates with the roles played by mitochondria in different tissues

5

INTRODUCTION

0.2 µm

Figure 1 Electron micrograph of mitochondria in pancreatic centroacinar cell. Bottleneck-like contacts between the intracristal space and the intermembrane space are indicated by arrows. (Reproduced from Tzagoloff [2], with permission from the author and Springer Science and Business Media.)

[i.e., whether they function mainly as energy transductors or as “chemical factories” (e.g., urea production in liver mitochondria)]. The intracristal compartments form a continuum with the intermembrane space. However, due to narrowness and elongation of the cristae and, quite often, a bottlenecklike shape of the connections between the cristae and the intermembrane compartment, free mixing of the contents of both spaces may be hindered (see the discussion of cristae junctions in Chapter 23). In many tissues, separate mitochondria observed by electron microscopy may, in fact, constitute fragments of larger, branched structures. In extreme cases, as in some protozoan or yeast species, it is proposed that the cell contains a single giant mitochondrion whose multiple ﬁngerlike branches may, in thin sections, look like separate mitochondria. In metazoans, mitochondria stained with ﬂuorescent dyes and viewed in a light microscope often form elongated threadlike structures (Figure 2), described by some authors as the mitochondrial network . Moreover, in live cells these structures are not only in constant motion but also split and fuse again. This phenomenon, observed ﬁrst decades ago, has attracted more attention in recent years, as it appeared to be related, among other factors, to the mitochondrial energy state and genetic status and to play a role in the programmed death of cells (apoptosis; see Section 5).

6

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

A

B

Figure 2 Mitochondrial network. Human osteosarcoma cell stained with ﬂuorescent dyes MitoTracker CMXRos (red) for mitochondria, phalloidin-FITC (green) for actin ﬁlaments, and DAPI (blue) for nuclear DNA. (A), Whole cell; (B), higher magniﬁcation of a cell fragment with mitochondria. The bars correspond to 12 µm. (Courtesy of J. Szczepanowska.) (See insert for color representation of ﬁgure.)

1.3. Structure and Compartmentation Despite this dynamic situation, the four intramitochondrial entities described above (i.e., the outer membrane, the intermembrane compartment, the inner membrane, and the matrix compartment) retain their distinct composition and characteristics. Both membranes are formed of phospholipid bilayers with multiple integral and peripheral proteins, many with transporting and enzymatic functions. The outer membrane contains the mitochondrial porin or voltage-dependent anion channel (VDAC) that enables the membrane to function as a molecular sieve by allowing compounds of up to 5000 Da to pass freely while preventing diffusion of larger molecules. It has to be mentioned that despite its name, VDAC is only partly selective toward anions, and it also allows more-or-less free diffusion of cations and uncharged molecules. Thus, the composition of the intermembrane compartment is similar to that of the cytosol as far as low-molecular-weight compounds are concerned. Among its lipidic compounds, the outer membrane contains cholesterol, in contrast to the inner mitochondrial membrane, which is essentially cholesterol-free. This property enables solubilization of the outer membrane by compounds complexing cholesterol, such as digitonin, thus making it possible to obtain mitochondria stripped of the outer membrane, called mitoplasts. Due to the impermeability of the outer membrane to large-molecular-weight compounds, the enzyme composition of the intermembrane compartment differs considerably from that of the cytosol. It is mainly the site of transphosphorylation reactions. An important example is the formation of ADP from AMP and ATP, catalyzed by adenylate kinase: AMP + ATP 2ADP

(1)

OXIDATIVE PHOSPHORYLATION

7

The inner mitochondrial membrane contains the entire respiratory chain plus the ATP synthase complex. Due to operation of the respiratory chain coupled with proton pumping, the inner membrane is also a site of high voltage difference on both sides, ranging up to 180 mV over the membrane thickness of about 100 nm. This capacitance reﬂects the extremely high insulating properties of the phospholipid membrane bilayer. On the other hand, the inner mitochondrial membrane contains several speciﬁc transporters for anionic metabolites, including respiratory substrates, inorganic phosphate, ADP, and ATP. A characteristic feature of most transporters is that they operate as exchange carriers (e.g., transporting dicarboxylic acids in exchange for phosphate or exchanging ATP for ADP, among others; see Section 2.5). The inner compartment encompasses the mitochondrial matrix. This dense solution of enzymatic proteins, coenzymes, metabolites, and inorganic ions is the site of the citric acid cycle (the Krebs cycle), which provides reducing equivalents to the respiratory chain. The matrix also contains the mitochondrial genome responsible for the limited genetic autonomy of the mitochondrion. Mitochondrial DNA (mtDNA), like bacterial DNA, is circular in shape and contains 37 genes. The mtDNA of the prokaryotic type is one of the arguments supporting the endosymbiotic concept of mitochondria origin according to which these organelles developed from prokaryotic organisms that invaded precursors of the present eukaryotes [12].

2. OXIDATIVE PHOSPHORYLATION 2.1. General Principles Approximately 95% of ATP formation in animal cells with aerobic type of metabolism occurs by oxidative phosphorylation (OXPHOS), and mitochondria can be considered as cellular powerhouses converting energy released during substrate oxidation into a form available for cellular processes. Therefore, although mitochondria are a site of many biosynthetic and metabolic processes, OXPHOS is paramount. OXPHOS consists of two functionally independent processes: oxidation of reduced substrates (expressed as respiration or oxygen consumption) and phosphorylation of ADP by inorganic phosphate. The latter, energy-consuming process occurs at the expense of energy released during the former. Thus, the two elements of oxidative phosphorylation are coupled to each other obligatorily. The mechanism of this coupling results from speciﬁc properties of the inner mitochondrial membrane, which is the location of oxidative phosphorylation. From a bioenergetic perspective, the most important feature of the inner mitochondrial membrane is its composition, and as a consequence, its extremely selective permeability to a variety of substances. In comparison to other membranes of an animal cell, the inner mitochondrial membrane contains a much higher proportion of proteins (approximately 80%) and only 20% phospholipids. Among the latter, 10% is cardiolipin, a unique mitochondrial phospholipid with four acyl chains.

8

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

In contrast, the proportion between proteins and lipids in the outer mitochondrial membrane is approximately 1 : 1, which is more typical for most other cellular membranes. In contrast to the outer membrane, the inner mitochondrial membrane can be crossed passively by only a few compounds, such as weak acids (e.g., acetic acid), water-dissolved gases (oxygen, ammonia), and lipophilic compounds. Electrically charged and hydrophilic compounds such as carboxylic anions (including respiratory substrates) and inorganic ions, are unable to pass the inner membrane without participation of specialized transporting proteins (see Section 2.5). High resistance of the inner mitochondrial membrane to protons is crucial for OXPHOS. From a functional point of view, the oxidative phosphorylation machinery consists of two proton-pumping systems capable of proton translocating across the inner membrane from the mitochondrial matrix to the intermembrane compartment, located in a highly H+ -impermeable lipidic core of the membrane. One of these pumps is the respiratory chain as a whole, and the other is the mitochondrial ATPase. As, under physiological conditions, it catalyzes ATP formation at the expense of energy delivered during the respiration, it is also deﬁned as ATP synthase. This name clearly depicts the real function of this enzyme in oxidative phosphorylation. Taken together, ATP formation catalyzed by mitochondrial ATPase is driven by the mitochondrial inner membrane electrochemical proton gradient (mitochondrial protonmotive force, p) built up during respiration. This statement is the most condensed summary of Peter Michell’s principle of the chemiosmotic concept of oxidative phosphorylation (Figure 3; for a comprehensive overview, see [3]). 2.2. Respiratory Chain as a Proton Pump Mitochondrial respiratory chain catalyzes electron transfer from the reduced donors (NADH and FADH2 ) to molecular oxygen (O2 ). The ﬁnal product of this pathway is water. Because of a large redox potential difference between electron donors and the ﬁnal electron acceptor (about 1.10 and 0.90 V for NADH and FADH2 as electron donors, respectively), which is a reﬂection of the displacement of the system from equilibrium, electron ﬂow along the respiratory chain is accompanied by a signiﬁcant decrease in the Gibbs potential (i.e., release of a large amount of free energy). Under physiological conditions, a large proportion of this energy is used to pump protons across the inner mitochondrial membrane from the matrix to the intermembrane space. The remaining energy is dissipated as heat. The proportion of energy utilized to generate the electrochemical proton gradient versus that dissipated as heat depends on the type of tissue and its physiological state. The unequal distribution of protons between the two sides of the inner mitochondrial membrane results in the generation of an electrochemical potential across it consisting of two components: a potential that reﬂects unequal distribution of electrical charges (), and the chemical potential resulting from an unequal distribution of chemical entities, mainly protons (more precisely,

9

OXIDATIVE PHOSPHORYLATION

Respirato r chain y

ATP synthase complex

Proton leak

Figure 3 Schematic representation of the chemiosmotic concept of energy coupling. The inner mitochondrial membrane contains the respiratory chain that operates as a proton pump by translocating protons from the inner to the outer side of the membrane, thus forming the electrochemical proton gradient (the protonmotive force, p) composed of the electric component (, positive outside) and the chemical component (pH, acidic outside). p then drives protons backward through the F1 FO complex (ATP synthase), becoming the driving force of ATP synthesis. The F1 FO complex can also operate in the reverse direction (as mitochondrial ATPase), hydrolyzing ATP and ejecting protons to the outside, thus building p.

hydrated hydrogen ions H3 O+ ) expressed as pH: p = − pH

(2)

In fully energized mitochondria, amounts to 180 to 200 mV, negative inside (the matrix side of the inner membrane is called the N-side, for “negative,” and the external side is designated as the P-side, for “positive”). The hydrogen ion concentration difference in animal mitochondria is usually about 0.5 pH unit, which corresponds to 30 mV. Because, in energized mitochondria, pH is higher inside mitochondria than outside, the pH difference is formally negative. Thus, the total protonmotive force of energized mitochondria may reach a value of 210 to 230 mV. Summing up, the mitochondrial electrochemical membrane potential may be regarded as an intermediate source of energy that is released during respiration and is made available for other, energy-consuming processes, such as ATP synthesis and metabolite transports across the inner membrane.

10

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

As p is a consequence of proton pumping across the inner membrane during electron ﬂow along the respiratory chain, an important question arises concerning the efﬁciency of the mechanisms transforming energy released during chemical reactions into the electrochemical proton gradient. The exact stoichiometry between electron ﬂow and proton pumping is still debated, although the most widely accepted ﬁgure is that 10 H+ are translocated from the mitochondrial matrix to the intermembrane space for each pair of electrons transported from NADH to oxygen. In the case of FADH2 , which donates two electrons to the respiratory chain one step downstream, at complex II, only six protons are extruded from the matrix compartment. Stoichiometry does not, however, deﬁne the efﬁciency of ATP formation in relation to the oxygen consumed. It must be stressed that variable amounts of p can be dispersed as a passive proton leak that bypasses ATP synthesis and other energy-consuming processes driven by p, such as ion and metabolite transport. Therefore, the actual stoichiometry of oxidative phosphorylation is usually expressed as the P/O ratio, the ADP/O ratio, or more generally, the P/2e− ratio. These ratios express the number of molecules of phosphate used for ADP phosphorylation in terms of the number of oxygen atoms consumed or the number of electron pairs transported along the respiratory chain [3]. The mitochondrial respiratory chain is composed of more then 85 proteins assembled in four complexes. Complexes I (NADH–ubiquinone oxidoreductase), III (ubiquinol–cytochrome c oxidoreductase), and IV (cytochrome c oxidase) are located in the inner mitochondrial membrane as integral proteins, whereas complex II, comprising succinate dehydrogenase, which catalyzes one of the steps of the citric acid cycle, is attached to the inner surface of the inner membrane. These enzymatic complexes are connected functionally by diffusible electron acceptors and donors: ubiquinone/ubisemiquinone/ubiquinol and oxidized/reduced cytochrome c. Figure 4 shows the sequence of reactions that comprise the mitochondrial respiratory chain. Complex I (NADH–ubiquinone oxidoreductase) catalyzes the oxidation of reduced nicotinamide nucleotides concomitantly with the reduction of ubiquinone (UQ) to ubiquinol (UQH2 ). This reaction is coupled to the pumping of four protons from the matrix to the intermembrane space per pair of electrons transferred from NADH. Complex I is the largest one within the respiratory chain, consisting of 43 polypeptides. Its redox center contains ﬂavin mononucleotide and a few (six or seven) iron–sulfur centers. Inhibition of this complex, which prevents electron ﬂow to ubiquinone and therefore causes a large accumulation of NADH, may lead to an enhanced formation of reactive oxygen species. As an electron carrier, ubiquinone accepts electrons from both complex I and succinate dehydrogenase, which is the essential part of complex II. Complex II is located at the internal side of the inner membrane and is the only respiratory complex encoded completely by nuclear DNA. It catalyzes oxidation of succinate to fumarate in the tricarboxylic acid cycle, with concomitant reduction of FAD and subsequent reduction of ubiquinone. This enzyme is composed of four subunits. One of them, containing FAD, participates in the oxidation of succinate.

11

OXIDATIVE PHOSPHORYLATION

succinate

fumarate

Figure 4 Overview of the mitochondrial respiratory chain. For each pair of electrons ﬂowing from NADH to oxygen, 10 protons are translocated from the inner to the outer side of the inner mitochondrial membrane. (Drawing by M. R. Wieckowski.)

Another, the extramembrane subunit, contains three Fe–S clusters that transport electrons to the next two subunits, which are internal membrane proteins, and then to ubiquinone. Electron transfer from succinate to UQ is not coupled to H+ translocation across the inner membrane. Ubiquinone can also be reduced in the reaction catalyzed by sn-glycerophosphate dehydrogenase (bound to the outer surface of the inner membrane) and by the electron-transferring ﬂavoprotein (ETF), a soluble enzyme in the matrix that mediates electron transfer in fatty acid oxidation (not shown in Figure 4). The enzymatic activity of these FAD-containing oxidoreductases, similar to that of complex II, is not directly connected to p generation. The next reaction of the respiratory chain is the electron transfer from ubiquinol to cytochrome c, catalyzed by complex III, ubiquinol–cytochrome c oxidoreductase, also known as bc 1 complex. Complex III is a dimer consisting of 11 subunits per monomer. It contains a few redox groups, including a 2Fe–2S center (Rieske protein), and three heme molecules (two b-type cytochromes and one cytochrome c 1 ). The mechanism of complex III activity includes the Q-cycle, in which two-step oxidation of ubiquinol to ubiquinone occurs, with transient formation of ubisemiquinone. This mechanism allows translocation of four protons from the matrix to the intermembrane space. It is noteworthy that ubisemiquinone, being a free radical, may enhance superoxide radical formation via autoxidation. This process is especially efﬁcient under conditions of high that prevent electron ﬂow between heme molecules in the Q-cycle, thereby increasing the half-life of ubisemiquinone (see Section 3). Finally, complex IV (cytochrome c oxidase) catalyzes sequential transfer of four electrons from reduced cytochrome c to molecular oxygen, forming two molecules of water. Complex IV is composed of 13 subunits, but only two of them (subunits I and II) are of high relevance to catalysis. Subunit II has a redox

12

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

center containing two copper atoms (CuA ) clustered with a sulfur atom that undergoes one electron redox process. Subunit I comprises two heme groups (heme a and heme a 3 ) and one copper atom (CuB ). Complex IV is the least efﬁcient p-generating proton pump in the respiratory chain. Calculating based on two electrons transferred to oxygen, only two protons are extruded to the intermembrane space. Other two protons combine with the oxygen atom and two electrons to produce water. Complex IV makes a relatively small contribution to generation of the mitochondrial protonmotive force despite large Gibbs free-energy. In contrast to reactions catalyzed by complexes I, II, and III, oxygen reduction by cytochrome c is irreversible. 2.3. Mitochondrial ATPase/ATP Synthase and Energy Coupling Mitochondrial ATPase (also called complex V, although it is not part of the respiratory chain) is a large protein complex. Negative staining techniques of electron microscopy reveal its structure as “mushroom-like” particles attached to the matrix side of the inner membrane. The head piece of this structure is attached to the membrane by a “stem” embedded in the lipidic phase of the membrane (Figure 5). The head piece, designated by Efraim Racker as coupling factor 1 (abbreviated F1 ), has a molecular mass of 370 kDa, whereas the stem, of 160 kDa molecular mass, is identiﬁed as Racker’s coupling factor O (sensitive to oligomycin, abbreviated FO ). The catalytic mechanism of ATP synthase exploits the mitochondrial p as a source of energy to displace the mass-action ratio for ADP phosphorylation by 7 to 10 orders of magnitude from equilibrium. During oxidative phosphorylation, protons diffuse down the concentration gradient from the intermembrane space

Figure 5 Coupling factor 1 (F1 ) visualized by negative staining. Numerous mushroom-like structures are visible attached to the inner side of inner membrane fragments of a disrupted rat liver mitochondrion. The diameter of the spherical “head” is about 10 nm, and the length of the “stalk” is 5 nm. (Electron micrograph by P. Włodawer.)

OXIDATIVE PHOSPHORYLATION

13

into the matrix through the proton channel formed by the FO subunit. This proton current across the inner membrane is accompanied by a decrease in Gibbs free energy, which drives reversal of ATP hydrolysis. The catalytic activity of the enzyme is associated with the F1 subunit, which hydrolyzes ATP if separated from FO . ATP formation from ADP and inorganic phosphate in a protein-free solution is negligible because of the extremely low equilibrium constant for this reaction. Phosphorylation of ADP bound to F1 , although still very low, is detected even in the absence of p, suggesting a slight increase in the equilibrium constant when reactants are bound to the F1 catalytic center. Such observations indicate that the important energy-consuming step of oxidative phosphorylation is release of ATP from the enzyme active site. In fact, the energy of the proton electrochemical gradient is not utilized directly for combining ADP and inorganic phosphate but, rather, to constrain conformational changes of the catalytic subunits that dictate ADP, Pi , and ATP binding afﬁnity and steric interaction. FO is composed of three types of proteins, called subunits a, b, and c, with the ﬁrst two encoded by mtDNA. The central channel of FO is formed by 10 c subunits organized in a symmetric ring traversing the inner membrane. Subunit a is connected asymmetrically with the external surface of the ring and subunit b, which extends from the membrane, connecting the transmembrane portion of FO with a distant subunit of the F1 particle. Thus, FO forms the H+ -selective, oligomycin-sensitive, channel. F1 consists of ﬁve types of subunits: α, β, γ, δ, and ε, assembled with the stoichiometry of α3 β3 γδε. Subunits α and β are positioned alternately around subunit γ, forming a caplike structure containing three αβ dimers. Subunit γ forms a stalk that connects the cluster of α and β subunits with the c10 ring of FO [5] (Figure 6). Each of the three catalytic centers of ATP synthase located on the three β subunits is able to assume three different conformations, varying in their afﬁnity for substrates (ADP and Pi ) and product (ATP). Conformation O (for “open”) is characterized by low afﬁnity to ATP; conformation L loosely binds ADP and Pi ; and conformation T tightly binds ADP and Pi , leading to ATP formation. In energized mitochondria, protons ﬂowing into the mitochondrial matrix compartment via the membrane-embedded FO sector force subunit γ to rotate, while subunit b forms a stator, holding subunits α and β stationary. For one complete turn of 360◦ , 10 protons must return to the matrix (i.e., one H+ for one subunit c) [3]. Simplifying somewhat, one can assume that rotation by 120◦ results in a switch from one conformation to another. Thus, one subunit β, being at conformation O, changes its structure to conformation L; another subunit β, originally at conformation L, is transformed to conformation T; and a third subunit β, being at conformation T, returns to conformation O (Figure 7). As result, one full revolution of subunit γ results in a complete cycle, in which three molecules of ATP are released. These sequential alterations of subunit β conformation are elicited mechanically by rotation of the asymmetrically oriented subunit γ within the

14

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

Figure 6 Subunit structure of mitochondrial ATPase/ATP synthase. Subunits c of FO are assembled as a ring plunged into the inner membrane. They allow protons to return to the mitochondrial matrix. Transient and sequential protonations of each of the 10 c subunits causes a clockwise rotation (when viewed from the membrane side) of subunit γ, driving a cycle of conformational changes of the α3 β3 assembly of F1 . Full 360◦ rotation requires 10 protons to pass across the inner mitochondrial membrane. This allows for phosphorylation of three molecules of ADP. OSCP (oligomycin sensitivity-conferring protein), together with subunits a and b, comprises a stator that prevents the α3 β3 assembly to rotate together with subunit γ. Note that the OSCP subunit is distant from FO and is not the oligomycin-binding site. However, it makes a link between subunit b and the α3 β3 assembly and prevents the latter from undergoing conformational changes in the presence of oligomycin. (Drawing by M. R. Wieckowski.) (See insert for color representation of ﬁgure.)

α3 β3 head-piece sector (Figure 6). Thus, mitochondrial F1 FO -ATPase, equivalent to ATP synthase, represents an interesting example of a mechanochemical catalytic assembly, a “nanomotor” [14]. Mitochondrial ATP synthesis coupled to inwardly directed proton ﬂux is fully reversible, meaning that ATP hydrolysis, catalyzed by the same enzymatic assembly, results in proton pumping in the reverse direction: from the matrix compartment out to the intermembrane space (and further on to the cytosolic compartment). Thus, under conditions of low p, cytosolic ATP (e.g., formed by glycolysis) is hydrolyzed to ADP and Pi with a concomitant restoration of p. Applying ingenious microtechniques to a ﬂuorescently labeled isolated F1 sector immobilized on a coverslip, it was possible, using ﬂuorescence microscopy, to observe rotation of subunit γ under conditions of ATP hydrolysis [15]. Moreover, by attaching a magnetic bead to subunit γ and applying a rotating magnetic ﬁeld, researchers succeeded in obtaining the formation of minute but detectable amounts of ATP [16].

OXIDATIVE PHOSPHORYLATION

15

Figure 7 Conformational model of oxidative phosphorylation. Alternating binding-site mechanism of mitochondrial ATP synthesis. O, L, and T represent three conformational states (open, loose, and tight, respectively) of the catalytic site, identiﬁed as subunit β. Rotation of asymmetric subunit γ (not shown in this scheme) results in alternating changes of the conformation of subunit β that are characterized by loose binding of ADP and Pi (site L), followed by their tight binding, resulting in the synthesis of ATP (site T) and the ﬁnal release of the ATP formed (site O). Each step (corresponding to a turn by 120◦ ) is executed by transfer of three protons form the inner (N) to the outer (P) side of the membrane. Chemical binding of ADP and Pi with the release of water is believed to result from tight contact of the two molecules due to steric alteration in the catalytic site. Thus, the energy of proton ﬂux down the proton electrochemical potential is transformed into the chemical energy of the high-energy phosphate bond. (From Devlin [13].)

2.4. Coupling and Uncoupling; Reversed Electron Transport As discussed above, most of the reduction–oxidation (redox) reactions of the electron transport system are reversible. Tight coupling of mitochondrial OXPHOS is therefore reﬂected by the reversibility of ATP synthesis/hydrolysis and transmembrane proton ﬂuxes, and at least partial reversibility of electron ﬂow in the respiratory chain and p formation. Indeed, electrons from ubiquinol can be transported “uphill” (i.e., against the redox potential) to complex I and on to NAD+ at the expense of p. The best known example is reduction of NAD+ to NADH by succinate. This process is termed reversed electron transport. Although the reversed electron ﬂow can be observed in energized isolated mitochondria, its role within intact cell under physiological conditions remains unclear.

16

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

Under conditions of excess respiratory substrate and O2 but no ADP and/or Pi , p increases and becomes a limiting factor for electron ﬂow. Such a condition is deﬁned as the resting state or state 4 respiration and is characterized by very low oxygen uptake and is readily obtained in isolated mitochondria. Low O2 uptake in these conditions is limited by a slow dissipation of p. Such proton leak is due to a variety of factors, including weak uncoupling by nonesteriﬁed long-chain fatty acids that cross the inner mitochondrial membrane passively in undissociated form, only to be transported in the opposite direction as anions by the adenine nucleotide carrier and other substrate carriers [17,18]. Futile cation ﬂuxes and ﬂickering of the unspeciﬁc permeability transition pore (see Section 5) also contribute to proton leak. In contrast, in the presence of excess ADP and Pi , mitochondria respire at the maximum rate, which is limited only by the rates of ATP synthesis and ATP/ADP exchange across the inner mitochondrial membrane (assuming that Pi transport is not limiting). Such a metabolic condition is called the active state or state 3. Under these conditions, the energy of the electron ﬂow along the respiratory chain is maximally utilized for ATP synthesis. Under experimental conditions, the inner mitochondrial membrane can be made fully permeable to protons. This can be achieved by disrupting the membrane mechanically or using chemicals that can transfer protons across the phospholipid phase of the membrane. Such protonophores are typically lipophilic weak acids that can cross the lipid bilayer passively in both protonated and deprotonated forms. Most commonly used are 2,4-diniotrophenol (DNP), carbonyl cyanide m-chlorophenylhydrazone (CCCP), and carbonyl cyanide p-triﬂuoromethoxyphenylhydrazone (FCCP). The rate of respiration in the uncoupled state (also described as state U) is essentially limited by the efﬁciency of the respiratory chain and is usually equal to, or somewhat higher than, that in active state 3. Physiological uncoupling is characteristic for some tissues, such as thermogenic brown adipose tissue, also called “brown fat,” present in neonatal mammals, including humans, and in mammals that hibernate. Mitochondria of this tissue may become almost completely uncoupled, due to the presence of a speciﬁc inner membrane protein, the uncoupling protein (UCP1), which enables passage of protons. The mechanism of this intrinsic property of UCP1 is similar to that described above for the adenine nucleotide carrier: namely, cycling of nonesteriﬁed fatty acids [19]. As a result, energy produced by the electron ﬂow in brown adipose tissue mitochondria is not captured in the form of p and utilized for ATP synthesis, but rather, dissipated as heat [20]. Homologs of UCP1 have recently been identiﬁed in other tissues: heart, skeletal muscle, and brain and termed UCP2, UCP3, UCP4, and UCP5. They are present in minute quantities and contribute only slightly to the inner membrane permeability. It is hypothesized that their physiological roles include protection against free-radical formation [19].

PRODUCTION OF REACTIVE OXYGEN SPECIES

17

2.5. Mitochondrial Carriers Although mitochondrial membrane potential is used primarily to drive ATP synthesis, it is also used to drive other processes, including importation of ADP, inorganic phosphate, and respiratory substrates, and to maintain ion gradients. Moreover, ATP and other metabolites formed within mitochondria are exported to the cytosol. For example, liver mitochondria release malate or phosphoenolpyruvate to support cytosolic gluconeogenesis, and citrulline for cytosolic urea synthesis. As mentioned earlier, the inner mitochondrial membrane is impermeable to the majority of substances, including metabolites, phosphate, and inorganic ions. Hence, translocation into and out of the matrix is possible only via speciﬁc transmembrane carriers and channel-forming proteins. Some of these transported substances are accumulated within mitochondria, or released into the cytosol, against a concentration gradient, a process requiring energy obtained from pH or [3]. Mitochondrial transporting mechanisms can be divided into the following four categories: 1. Electroneutral exchange driven by pH. For example, mitochondria accumulate inorganic phosphate, which is exchanged for OH− via the Pi /OH− antiporter, which is equivalent to the Pi /H+ symport. Another example of such transport is the electroneutral exchange of a cation for a proton (e.g., Na+ /H+ ). In energized mitochondria (alkaline inside) this exchange will favor the efﬂux of cations. 2. Electrogenic uniport of cations driven by . Mitochondrial Ca2+ and K+ uptake belongs to this category. 3. Electroneutral exchange of two metabolites (e.g., 2-oxoglutarate2− /malate2− ). Such transport is driven by concentration gradients of each of the substances transported and as such does not dissipate p. 4. Electrogenic exchange of two metabolites (e.g., ATP4− /ADP3− , citrate3− /malate2− ). In this case the direction of exchange is determined by the transmembrane potential. For example, in energized mitochondria the exchange of internal ATP4− against external ADP3− is favored by (negative inside), whereas the reverse exchange is hindered. Such a preference disappears in uncoupled mitochondria.

3. PRODUCTION OF REACTIVE OXYGEN SPECIES The fate of most electrons that enter the respiratory chain is the four-electron reduction of dioxygen (O2 ) to form water at complex IV. However, “electron leak” from other redox sites in the respiratory chain results in small but signiﬁcant one-electron reduction of O2 that yields superoxide anion radical O2 −· . According to a very rough estimation, about 1% of the total oxygen uptake in mammalian tissues is transformed into this free radical. The superoxide anion, or

18

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

its protonated form, HO2 · (pK a ∼ 4.8), can dismutate to form hydrogen peroxide H2 O2 in a reaction catalyzed by superoxide dismutase (SOD): 2O2 −· + 2H+ → H2 O2 + O2

(3)

In the presence of transition metal cations, in particular Fe2+ and Cu+ , hydrogen peroxide reacts nonenzymatically in the Fenton reaction, yielding extremely reactive hydroxyl radical HO· : H2 O2 + Fe2+ → HO· + OH− + Fe3+

(4)

The two free radicals, superoxide anion O2 −· and hydroxyl radical OH· , along with hydrogen peroxide (H2 O2 ), singlet oxygen (1 O2 ) formed in some photochemical reactions, and ozone (O3 ), an air pollutant, constitute a class of reactive oxygen species (ROS) that are far more chemically reactive than the “normal” triplet oxygen molecule, O2 . Among them only O2 −· , H2 O2 , and under speciﬁc conditions, HO· are physiological metabolites. It is now generally agreed that the main sites of O2 −· generation at the level of the mitochondrial electron transport chain are complexes I and III [21,22] (Figure 8). The relative contribution of either complex is not known precisely and may vary among various tissues and depend on metabolic conditions. In complex I the primary source of O2 −· appears to be one of the iron–sulfur clusters. In complex III the likely mechanism of O2 −· generation seems to be coenzyme Q cycling, in which ubisemiquinone, a free radical by itself, functions as a redox intermediate on both sides of the inner mitochondrial membrane. Its one-electron autoxidation in the presence of O2 is likely to generate O2 −· . A general condition that enables O2 −· generation is a highly reduced state of the electron carriers at speciﬁc sites. Only then may a nonenzymatic “leak” of electrons out of the enzymatic electron transport route (respiratory chain) become possible. This can happen if the respiratory chain is blocked downstream from the particular O2 −· generation site [e.g., by the microbial product antimycin A (complex III inhibitor) or the plant poison rotenone (complex I inhibitor)]. Another condition is a temporary decrease in oxygen tension (anoxia or hypoxia) followed by reoxygenation. Under physiological conditions, this occurs during reperfusion following ischemia. In tightly coupled mitochondria, the rate of electron ﬂow through the respiratory chain is limited by the rate of ATP synthesis [i.e., by the availability of ADP (assuming saturating concentration of inorganic phosphate)]. More precisely, it is limited by the rate of proton pumping at the speciﬁc coupling sites (i.e., complexes I, III, and IV). In turn, this is limited by the “counterpressure” of the protonmotive force. The rate of electron ﬂow in the respiratory chain is low in the resting state (state 4) and increases greatly in the active state (state 3). Consequently, the reduction status of respiratory chain carriers, especially those of complexes I and III, is higher in the resting state than in the active state. This explains the observation that generally, mitochondrial production of ROS

19

PRODUCTION OF REACTIVE OXYGEN SPECIES

-. O2

O2

G3P DHAP

C side

-. O2

G3P-DH FeS

O2

I

Qo

.

+ + +

cyt c

III

IV

t c1

cy

FMN

Q-Pool

FeS

∆p

Qi . FeS

II

-. O2

ETF

O2

+

NADH NAD

FAD Acyl-CoA Enoyl-CoA

M side

cyt a/a3 [Cu]

eS RF

-. O2

− − −

O2

O2

H2O

Succ Fum

Figure 8 Sites of ROS generation within the mitochondrial electron transport chain. The superoxide anion (O−· 2 ) is generated mainly at complexes I and III of the respiratory chain and, to a smaller extent, by mitochondrial glycerophosphate dehydrogenase. O−· 2 is released into both the matrix side (M side) and the intermembrane (“cytosolic”) side (C side). O·o and O·i indicate ubisemiquinone radicals at o and i sites of complex III, respectively. Solid lines and arrows show direction of the forward electron transfer; the dotted line indicates the reversed electron transfer driven by the protonmotive force (p). Other abbreviations: Succ, succinate; Fum, fumarate; G3P, sn-glycerol 3-phosphate, DHAP, dihydroxyacetone phosphate; G3P-DH, glycerophosphate dehydrogenase; FMN, ﬂavine mononucleotide; FAD, ﬂavine-adenine dinucleotide; ETF, electron transfer ﬂavoprotein; Q, ubiquinone; FeS, iron-sulfur cluster; R FeS, Rieske iron-sulfur protein; cyt c1, cyt c, cyt a/a3, respective cytochromes. (From Sch¨onfeld, P. and Wojtczak, L. Fatty acids as modulators of the cellular production of reactive oxygen species. Free Radic Biol Med . 2008; 45:231–241; modiﬁed.)

is higher in the resting state than in the active state. Therefore, extrapolating to the whole tissue or even to the whole organism, it is incorrect to conclude that ROS generation is proportional to the rate of oxygen consumption. Hence, the value of about 1% reported for the proportion of oxygen consumed being transformed to ROS (see above) should be regarded as an average and as a very rough approximation, especially given high nonphysiological concentrations of O2 ex vivo. Another factor controlling the rate of ROS generation is the protonmotive force (p). Since its electric component dominates the concentration component (180 to 200 mV for compared to 30 to 60 mV for pH), it can be stated that the second factor regulating the rate of ROS generation is the transmembrane potential. Indeed, a drop of by as little as 30 mV that accompanies the transition from state 4 to state 3 can decrease the rate of ROS generation severalfold. Similarly, chemical protonophores such as DNP or CCCP strongly decrease ROS formation in both isolated mitochondria and intact cells and tissues [23]. In many tissues and organs the role of natural regulators of the mitochondrial

20

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

protonmotive force is played by the uncoupling proteins (UCPs). Discovered originally in the mammalian thermogenic organ, the brown adipose tissue, and designated as uncoupling protein 1 (UCP1), homologous proteins (i.e., UCP2, UCP3, UCP4) have more recently been found in brain, heart, skeletal muscle, liver, and some other tissues. Although solid experimental evidence is lacking, it is hypothesized that one of the functions of these proteins is to provide subtle control of the mitochondrial protonmotive force, and thus of ROS production [19]. It has to be stressed, however, that p affects ROS production primarily by controlling the redox state of respiratory chain components, although some direct effects cannot be excluded. This secondary role is illustrated, for example, by the fact that respiratory inhibitors such as antimycin A and rotenone (in the presence of NAD-linked substrates) increase ROS generation, although they decrease p. A particular case of ROS generation where both redox state and p play a decisive role is reversed electron transport. The reversed electron transfer from ubiquinol uphill to complex I is driven by high p using electrons derived from the oxidation of succinate to fumarate by complex II or, in some tissues, sn-glycerophosphate dehydrogenase. Due to the reversed electron transfer, in tightly coupled mitochondria, succinate oxidation is able to maintain a higher NADH/NAD+ ratio and, consequently, to produce more ROS than in the case of NAD-linked substrates. Both processes are, however, extremely sensitive to p, so they can be halted by even a small decrease in the protonmotive force occurring under transition from state 4 to state 3. It seems highly likely that the widely discussed high sensitivity of ROS production to p or in intact cells may be due to stopping of the reversed electron transport. The extent of reversed electron transfer, or under what conditions it is a physiological process, remains unresolved. O2 −· generated by the respiratory chain appears on both sides of the inner membrane. It seems likely that the superoxide anion produced at the level of complex I is mostly liberated in the matrix compartment, whereas that produced at complex III may appear on both sides (Figure 8). Apart from the two sites of ROS generation in the respiratory chain, there are a few other enzymes that may produce ROS within the mitochondrion [22,24]. They are 2-oxoglutarate dehydrogenase, one of the tricarboxylic acid cycle enzymes present in the matrix; sn-glycerophosphate dehydrogenase, a ﬂavoprotein enzyme present in some tissues at the external side of the inner membrane; cytochrome b 5 reductase; and monoamine oxidase, both present in the outer mitochondrial membrane. The latter enzyme apparently releases hydrogen peroxide directly rather than superoxide. It oxidizes biogenic amines and is highly active in neurons. In addition, signiﬁcant amounts of ROS can be produced outside mitochondria: namely, in the endoplasmic reticulum and during some metabolic transformations of polyunsaturated fatty acids. This is, however, outside the scope of the present chapter. It is also important to note that various forms of ROS can be generated within the cell by the action of ionizing and ultraviolet radiation and xenobiotics, including some pharmaceuticals [11] (e.g., the chemotherapeutic agent doxorubicin, or herbicides such as paraquat).

PRODUCTION OF REACTIVE OXYGEN SPECIES

21

ROS generated during operation of the mitochondrial respiratory chain are generally regarded as by-products of aerobic metabolism. Although they may fulﬁll some signaling functions, they are mostly harmful to the cell. The hydroxyl radical that originates in the cell from hydrogen peroxide only in the presence of transition metals is extremely reactive and can attack almost any compound in its vicinity. The superoxide anion is more stable but can react with lipids, primarily by attacking double bonds of unsaturated fatty acid moieties, and with proteins and nucleic acids, thus producing wide damage in the cell. Moreover, peroxides of fatty acids can initiate chain reactions that propagate from one acyl chain to another, multiplying the initial damage and destabilizing membranes. Several systems decompose ROS and thus protect the cell against its noxious actions (see below). However, if the rate of ROS generation increases and/or the protective systems fail, ROS steady-state concentration increases, resulting in oxidative stress. The ultimate effect of such a situation is cell death, either necrotic or programmed (apoptotic). The protective systems include a number of low-molecular-weight antioxidants and enzymatic systems. The former category includes nutritional products (vitamins) such as ascorbic acid (vitamin C), α-tocopherol (vitamin E), and β-carotene (provitamin A), as well as intrinsic cellular ingredients, including reduced glutathione (GSH) and reduced pyridine nucleotides NADH and NADPH. It remains unclear to what extent these low-molecular-weight compounds function in the intramitochondrial nonenzymatic defense system. This is in contrast to several enzymatic mechanisms, whose function in combating oxidative stress is well established. The chain reaction that aims to detoxify the superoxide anion is initiated by superoxide dismutase, as illustrated in reaction (3). In analogy to prokaryotic superoxide dismutase, the mitochondrial enzyme contains manganese atom in its active center (Mn-SOD). This is in contrast to cytosolic superoxide dismutase, which contains zinc and copper atoms (Cu,Zn-SOD). Mn-SOD is located exclusively in the mitochondrial matrix and transforms the O2 −· generated therein very efﬁciently into H2 O2 . This is underscored by the fact that heterozygous Mn-SOD-knockout mice, containing 50% of the normal activity of the enzyme, appear quite normal, yet homozygous animals, essentially lacking mitochondrial SOD, die during the ﬁrst few weeks after birth [25]. The intermembrane compartment contains Cu,Zn-SOD, which is probably identical or very similar to the cytosolic enzyme. Thus, O2 −· generated at the external side of the inner mitochondrial membrane can be transformed efﬁciently to H2 O2 . It should be noted, however, that the dismutation reaction transforms one reactive oxygen species into another, and although H2 O2 is not a free radical, it is potentially injurious. The danger presented to cellular integrity and viability by hydrogen peroxide is based on two properties of this compound: (1) it crosses biological membranes readily, in contrast to the limited permeability of the superoxide radical; and (2) it generates extremely reactive hydroxyl radical in the presence of ferrous ions [see reaction (4)]. Therefore, the next step in the protective mechanisms against ROS is removal of hydrogen peroxide by catalase,

22

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

a heme enzyme common in various tissues, via the reaction 2H2 O2 → 2H2 O + O2

(5)

Catalase has diffusion-limited kinetics, and one molecule can turn over millions of molecules of hydrogen peroxide per second. Catalase is located mainly in peroxisomes. Its presence in preparations of isolated mitochondria is due to peroxisomal contamination and can even be regarded as a measure of such contamination. Nevertheless, according to some reports, catalase may be intrinsic to heart mitochondria, making this tissue exceptionally capable of opposing oxidative stress [26]. Enzymes that remove hydrogen peroxide by reducing it with electrons derived from organic compounds are classiﬁed as peroxidases. Unspeciﬁc peroxidases are common in the cytoplasm of various animal and plant tissues. Intrinsic to mitochondria is glutathione peroxidase, which reacts with reduced glutathione (GSH) as an electron (or oxygen) acceptor, oxidizing it to glutathione disulﬁde (GSSG), also termed (not quite correctly) oxidized glutathione. Glutathione peroxidase contains selenocysteine in its active center. It is the main, or perhaps the only, enzyme removing H2 O2 from the mitochondrial matrix, and its effectiveness and efﬁciency are enabled by a high, millimolar intramitochondrial concentration of GSH. Another glutathione peroxidase reacts preferentially with phospholipid hydroperoxides, but can also reduce cholesterol peroxides and even H2 O2 to yield GSSG. This enzyme, phospholipid hydroperoxide glutathione peroxidase, is also a selenoenzyme and is thought to be located inside mitochondria. It can play an important role in repairing biological membranes whose phospholipids have already been peroxidized by various types of ROS. It is abundant in some tissues (e.g., testes) but may be absent in many others. GSSG resulting from the reactions catalyzed by glutathione peroxidases must be reduced back to GSH to enable the process to continue. This is catalyzed by glutathione reductase, which is present in the mitochondrial matrix, where it utilizes NADPH selectively as the electron donor. In turn, NADPH can be produced by the transhydrogenation reaction NADH + NADP+ + p → NAD+ + NADPH

(6)

The term p on the left side of this reaction indicates that the reaction running from left to right utilizes energy in the form of the protonmotive force. Hence, maintaining a high intramitochondrial concentration of NADPH is connected with energy expenditure. NADPH can also be generated from isocitrate or malate by the action of the respective dehydrogenases, NADP+ -dependent mitochondrial isocitrate dehydrogenase or, mostly in neurons, decarboxylating malate dehydrogenase (called the malic enzyme). Thus, regeneration of reduced glutathione is an energy-consuming process and can compete with ATP synthesis for the protonmotive force or respiratory substrates, which means that protection against oxidative stress is energetically costly.

CALCIUM SIGNALING

23

There is, however, one ROS-removing process that can, at least theoretically, provide energy to mitochondria instead of utilizing it. This is the oxidation of O2 −· by cytochrome c present in the intermembrane compartment. Since cytochrome c is bound loosely to the outer surface of the inner mitochondrial membrane, it is also present at low, submillimolar concentration in the free form between the inner and the outer membranes. This fraction of free cytochrome c can react nonenzymatically in a one-electron process with the superoxide radical according to the reaction O2 −· + cyt. c(Fe3+ ) → O2 + cyt. c(Fe2+ )

(7)

Reduced cytochrome c can subsequently be reoxidized by complex IV of the inner mitochondrial membrane, thus providing electrons to the ﬁnal step of the respiratory chain that is coupled to proton pumping and p formation [27]. Under normal conditions, all these ROS-metabolizing processes are sufﬁcient to keep intramitochondrial steady-state concentrations of O2 −· and H2 O2 at physiological submicromolar levels. Moreover, it is hypothesized that mitochondria can function as a sink for ROS produced extramitochondrially [24]. Oxidative pathology emerges only after failure of one or more of these scavenging systems and/or substantial elevation of ROS generation, conditions generally termed oxidative stress. The evolutionary adaptations to prevent the noxious effects of oxygen free radicals are the price that organisms living in oxygen-rich environment have to pay for highly efﬁcient aerobic ATP synthesis. Chronic exposure of mitochondria to relatively high ROS concentrations increases the probability of mtDNA mutations, especially because mtDNA is not protected by histones and contains no introns and is therefore more susceptible than nuclear DNA to oxidative damage. Gradual damage of mtDNA during the human life span results in a progressive decrease in the efﬁciency of OXPHOS. This in turn may promote accelerated ROS formation, which further enhances mtDNA mutations. Such a vicious cycle is an unavoidable consequence of aerobic poise and supports the mitochondrial theory of aging [28–32]. Gradual loss of ATP generating capacity undermines many crucial cellular processes and has been implicated in many degenerative diseases. In addition, some pathologies, which may lead to mitochondrial stress, such as inﬂammatory diseases, excessive physical exercise, and ischemic insult followed by reperfusion, can enhance ROS generation and therefore may increase the probability of mtDNA mutation above normal levels.

4. CALCIUM SIGNALING Mitochondrial Ca2+ uptake, extrusion, and accumulation are key to cellular calcium homeostasis. As discussed above, Ca2+ inﬂux through the inner mitochondrial membrane is driven by ; hence, it is sensitive to factors that may affect the mitochondrial energy state, such as uncouplers and respiratory chain

24

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

inhibitors. The bulk cytosolic Ca2+ concentration in the resting cell is about 100 nM, so that given a value of 180 mV (negative inside), mitochondrial Ca2+ concentration at equilibrium should theoretically reach the value of 100 mM (one order of magnitude for each 30 mV of ). In fact, this does not occur since Ca2+ inﬂux is effectively counterbalanced by Ca2+ efﬂux, which occurs via an electrogenic Ca2+ /3Na+ antiporter or, in some cells with low sodium clearance (e.g., hepatocytes), as electroneutral Ca2+ /2H+ exchange. Sodium entering the mitochondrial matrix is removed by the Na+ /H+ exchanger. Thus, the net balance between Ca2+ entry and release is maintained at the expense of the proton circuit [33]. The mitochondrial Ca2+ uniporter, through which cytosolic calcium enters the matrix, displays a low afﬁnity toward Ca2+ , with a K d value greater than 10 µM in the presence of physiological Mg2+ concentrations. This, together with a high activity of the Ca2+ /3Na+ antiporter, makes mitochondrial Ca2+ accumulation inefﬁcient until cytosolic Ca2+ concentration reaches a threshold of about 400 nM. Therefore, mitochondria should not be considered as Ca2+ storage organelles that accumulate calcium in resting cells. In stimulated cells, bulk cytosolic Ca2+ concentration increases up to 1 µM, but locally, in close proximity to Ca2+ channels in the endoplasmic reticulum (ER) and in the plasma membrane, it may reach much higher values. Mitochondria located in such subcompartments accumulate Ca2+ efﬁciently and decrease the local Ca2+ concentration. Such Ca2+ -buffering activity of mitochondria may affect many calcium-dependent processes, including Ca2+ entry through calcium channels and its removal by Ca2+ -ATPases [34,35]. Therefore, mitochondria modulate the intensity of intracellular calcium signals originating from both ER and the extracellular space [36]. In this way, mitochondria can also control calcium waves and oscillations spreading throughout excited cells [37,38]. Matrix Ca2+ is chelated by proteins, nucleotides, and phosphate, and is released readily when cytosolic Ca2+ returns to lower resting levels. It must be emphasized that a slight increase in the mitochondrial Ca2+ concentration stimulates energy metabolism, as calcium activates pyruvate dehydrogenase and two dehydrogenases of the tricarboxylic acid cycle. On the other hand, excessive mitochondrial Ca2+ accumulation fosters activation of the permeability transition pore (PTP) and therefore may be harmful to mitochondria and eventually to the cell [29,39] (see Section 5). However, the PTP can also operate in the low-conductance mode that allows mitochondria to release excess calcium, thus reducing the risk of mitochondrial damage without dissipating p completely [40]. In such a case, the PTP prevents mitochondrial malfunction and supports cell survival. Moreover, a limited opening of the PTP slightly decreases p, which in turn reduces mitochondrial ROS formation. Intracellular organization of the mitochondrial network, which bridges the subplasma membrane space and ER, gives a structural basis for intramitochondrial Ca2+ transfer from the plasma membrane calcium channels to Ca2+ -ATPases that pump Ca2+ into intracellular stores [41]. Apart from calcium channels and

MITOCHONDRIA AND CELL DEATH

25

pumps located in the plasma membrane and in the endoplasmic reticulum, respectively, this phenomenon involves mitochondrial Ca2+ uniporters and mitochondrial Ca2+ /3Na+ antiporters. Calcium released from mitochondria into a limited space makes a local “hot spot” in the proximity of Ca2+ -ATPase. It increases the rate of ﬁlling up the calcium stores and allows for their reloading without an excessive increase in the cytosolic Ca2+ concentration. On the other hand, mitochondria located very close to Ca2+ channels in ER membranes (coupled to IP3 -dependent or ryanodin receptors) may sense local increases in the cytosolic Ca2+ concentration in very discrete junctions between ER and mitochondria, so they may take up Ca2+ almost directly from the intracellular stores [41]. Perturbations in the mitochondrial energy metabolism affecting interfere with cellular calcium homeostasis and may result in serious consequences for the cell. The decrease in oxidative phosphorylation and hence the resulting ATP deﬁciency limit the rate of Ca2+ removal from the cytosol to the extracellular space as well as Ca2+ sequestration in intracellular calcium stores. This leads to a harmful overactivation of numerous calcium-dependent enzymes, such as calpains, phospholipases A2 , and protein kinases C. Moreover, prolonged increases in the poststimulatory cytosolic Ca2+ concentration decrease the excitability of electrically excitable cells. In the case of neurons, it delays recovery of the resting potential that may affect brain plasticity. Such phenomenon is attributed to age-related ROS-induced impairment of OXPHOS [42]. On the other hand, the reduced Ca2+ -buffering capacity of mitochondria strongly affects cellular calcium signaling because of the lowered ability of mitochondria to regulate local and global calcium events such as spikes, sparks, waves, and oscillations [37,43].

5. MITOCHONDRIA AND CELL DEATH It is paradoxical that mitochondria, which are indispensable for cell survival, are also necessary for cell suicidal death. This programmed cell death, also called apoptosis, is a complex sequence of events aimed to eliminate single cells or their assemblies when their natural biological function has come to an end or when a cell has become damaged or mutated to such an extent that its further existence might be deleterious to the whole organism. In particular, apoptosis occurs in embryogenesis, metamorphosis, and in the growth and maturation of individual organs. Apoptosis is also believed to eliminate cells whose metabolism and genomic organization have undergone transformations that may lead to malignancy. Thus, apoptosis is one of the main natural mechanisms protecting against cancer development. On the other hand, the increased propensity of a cell to undergo apoptotic decay may give rise to a series of pathologies, such as neurodegenerative diseases and tissue damage, that develop as a consequence of ischemia, in particular in heart and brain. In general, apoptosis may proceed by two partially interdependent routes, the death receptor pathway and the mitochondrial pathway [8,44,45]. The former is initiated by ligation of death receptors at the cell surface, whereas the latter

26

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

originates in mitochondria. In this case, one of the early events is the release of cytochrome c and some other peptides from mitochondria into the cytosolic compartment. The mechanism by which cytochrome c is liberated from mitochondria to the cytosol is still debated. Earlier hypotheses assumed that mitochondrial swelling causes disruption of the outer membrane. More recent reports indicate, however, that cytochrome c is also released under conditions where the outer membrane retains its integrity. A decisive role in this process is played by the mitochondrial permeability transition pore and proapoptotic proteins of the Bcl-2 family: in particular, Bid and Bax. One of the factors that can initiate this process is oxidative stress and the resulting oxidative attack of reactive oxygen species on phospholipid components of the inner mitochondrial membrane, particularly cardiolipin. Cytochrome c is normally bound to the inner membrane by electrostatic interactions with negatively charged cardiolipin. As cardiolipin is rich in polyunsaturated fatty acids, mostly linolenic acid, it easily undergoes peroxidation, which changes its physicochemical properties drastically and may lead to a partial desorption of bound cytochrome c. As a result, the concentration of free cytochrome c in the intermembrane compartment, normally at low submillimolar levels, may increase sharply, promoting leakage into the cytosol [46] (Figure 9). Apoptosis has often been observed to be accompanied by mitochondrial ﬁssion [48,49]. It remains, however, debatable whether this change in the structure of the mitochondrial network is related to the liberation of cytochrome c and other proapoptotic factors from mitochondria [50]. Cytochrome c released to the cytosol participates in the formation of a multiprotein complex called apoptosome. Together with other components of this complex, and in the presence of dATP or ATP, the apoptosome activates caspase-9. This is a representative of a large class of cysteine proteases that cleave their substrates after the aspartic acid moiety (hence the term caspases). Activation of caspase-9 is, by itself, an autocatalytic proteolysis that transforms procaspase-9 into its active form. Caspase-9 belongs to a class of initiator caspases, as it activates a series of other caspases, called effector caspases, in particular caspase-3 and caspase-7. Activated caspases are mainly responsible for degradation of the cell that is characteristic of the terminal phase of apoptosis. However, accidental activation of one of the initiator caspases might also trigger a chain of reactions eventually leading to cell destruction. To avoid such an inadvertent course of events, cells also contain a protective system in which the central role is played by a family of caspase-inhibitor proteins, IAPs (inhibitors of apoptosis proteins). Thus, to enable programmed cell death to proceed, IAPs are removed or otherwise neutralized concomitant with activation of the caspases. This function is fulﬁlled by another protein, Smac (second mitochondrial activator of caspases; also called Diablo), that is released from the mitochondrial intermembrane space together with cytochrome c and other proapoptotic proteins. As mentioned above, the PTP seems crucial for the release of proapoptotic factors. This pore is located in the contact sites between the outer and inner mitochondrial membranes and in its open state enables free passage of

27

MITOCHONDRIA AND CELL DEATH

(A)

(B)

tBid

Bcl-2 Bcl-XL

Cyt. c

Cyt. c Bax

Apoptotic signal (C) cardiolipin

ROS

Cyt. c

Figure 9 Schematic representation of mechanisms accounting for outer mitochondrial membrane permeabilization and the release of cytochrome c. (A) Induction of permeability transition pore opening, leading to matrix expansion and rupture of the outer membrane. (B) Bax-mediated permeabilization of the outer mitochondrial membrane, involving tBid-induced Bax insertion and homooligomerization that can be inhibited by Bcl-2 or Bcl-XL . (C) Peroxidation of cardiolipin is a key ﬁrst step in mobilizing cytochrome c from the inner mitochondrial membrane prior to Bax-induced (b) permeabilization of the outer membrane. (From Robertson et al. [47] with permission of Macmillan Publishers Ltd., copyright  2003.)

low-molecular-weight compounds, up to 1.5 kDa, between the mitochondrial inner compartment (matrix) and the cytosol [51,52]. It is formed by a complex assembly of several proteins originating from the outer mitochondrial membrane (porin), the inner membrane (adenine nucleotide translocase), and the matrix (cyclophilin D) (Figure 10). Opening of the PTP is favored by factors such as Ca2+ accumulation in mitochondria, reactive oxygen species, and low . The PTP is believed to be a “safety valve” against calcium overload of the mitochondrial inner compartment. Its ﬂickering may also be one of the factors responsible for a limited “proton leak” through the inner membrane of coupled mitochondria. PTP opening results in large-scale mitochondrial swelling. Such swelling, leading to rupture of the outer mitochondrial membrane and liberation of soluble proteins from the intermembrane compartment to the extramitochondrial space, was initially believed to be one of the underlying factors of apoptosis. Subsequent research revealed, however, that it may not be so, because large-scale

28

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

Figure 10 Model of the contact site between the outer and inner mitochondrial membranes that may function as the permeability transition pore. Indications: VDAC, voltage-dependent anion channel (mitochondrial porin); ANT, adenine nucleotide translocase; Cyp D, cyclophilin D; HK, hexokinase; PBR, peripheral benzodiazepine receptor; Bcl-2, antiapoptotic protein Bcl-2. Cytochrome c molecules associated partially with the outer face of the inner membrane and partially free in the intermembrane space are indicated by red circles. (Drawing by M. R. Wieckowski.) (See insert for color representation of ﬁgure.)

swelling, which is due to a difference in the colloidal osmotic pressure inside the matrix compartment and the external medium, is much less pronounced in mitochondria within the cell than in isolated mitochondria suspended in isotonic saline or sucrose media. Also, multiple ultrastructural observations do not show a link between mitochondrial swelling and rupture of the outer membrane with the onset and the progress of apoptosis. The PTP alone is too narrow to allow passage of cytochrome c (13 kDa), apoptosis-inducing factor (AIF), and other proapoptotic proteins. Moreover, the pore connects the cytosolic compartment with the matrix compartment, not the intermembrane space where the aforementioned apoptosis-inducing proteins are located. Therefore, although a relation between PTP opening and the onset of apoptosis has been well documented, connections between these events are complex. It has been shown, however, that a channel permeable to cytochrome c, AIF, and other proteins of the intermembrane space is formed by association of the PTP with the proapoptotic proteins Bax and Bak. In nonapoptotic cells, these two proteins are located in the cytosol or are loosely bound to the outer mitochondrial membrane in monomeric forms. Upon a death stimulus, another

MITOCHONDRIA AND CELL DEATH

29

proapoptotic protein, Bid, undergoes a proteolytic cleavage, and its C-terminal truncated derivative, t-Bid, induces homooligomerization of Bax and Bak, which then associate more ﬁrmly to the outer membrane, making it permeable to cytochrome c. This association and its pore-forming activity are prevented by the antiapoptotic proteins Bcl-2 and Bcl-XL . Thus, a subtle balance between these proapoptotic and antiapoptotic proteins and their interactions with the PTP are decisive for the survival or the apoptotic death of the cell. This balance can be affected by a number of mitochondria-targeted drugs. To make the process even more complex, it has been observed that some heat-shock proteins, in particular HSP70, may also prevent cytochrome c release or can somehow “neutralize” cytochrome c that has already been released. Mitochondria also release endonuclease G, which is involved in DNA degradation. Some other nucleases become activated by caspases. These nucleases are decisive in internucleosomal cleavage of DNA in cells undergoing apoptosis. As mentioned above, one of the apoptosis-promoting factors is reactive oxygen species [53]. Excessive production of ROS in the cell can be induced by a number of xenobiotics, transition metal ions, and ultraviolet and ionizing radiations. ROS action on mitochondria results in both a detachment of cytochrome c from the inner membrane and opening of the PTP, thus promoting liberation of cytochrome c to the cytosol. Ionizing radiation (x-ray and γ radiation), often used in cancer therapy, also acts by inducing apoptosis. Being more energetic than ultraviolet radiation, it also affects DNA and thus initiates both DNA- and mitochondrial-linked apoptosis pathways. Similarly, several anticancer drugs exert their therapeutic effect by inducing the apoptosis of malignant cells. In general, they act by inducing intracellular ROS production (e.g., doxorubicin). The increased level of ROS not only promotes PTP opening but, as mentioned above, also results in the peroxidation of polyunsaturated fatty acid moieties in the phospholipid bilayer of the inner membrane, in particular of cardiolipin, thus promoting desorption of cytochrome c. It is often stressed that massive release of cytochrome c from mitochondria requires not only permeabilization of the outer membrane but also an increased level of free, unbound cytochrome c in the intermembrane compartment. A simpliﬁed scheme of mitochondrial events leading to apoptosis is shown in Figure 11. In contrast to apoptosis, which can be regarded as a controlled process, necrosis is deﬁned as an uncontrolled cell death leading to nonselective cell damage. It usually results from major cell injury and disruption of vital cell functions such as energy production and selective permeability of cell membranes. Necrosis is a pathological rather than a physiological process and is usually followed by inﬂammatory reactions of adjacent cells and tissues. Similar to apoptosis, necrosis can also be induced by extracellular pathological disturbances such as ischemia, trauma, and some neurodegenerative disorders. The most characteristic features of cells dying a necrotic death are mitochondrial permeabilization, disruption of lysosomes, and loss of osmotic balance between intra- and extracellular ﬂuids. This latter event results in an increase in cell volume, eventually leading to plasma

30

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

CuZnSOD Catalase, GPx

O2, H2O

γ- and UVradiation Bax

+ “death signals”

+

OCH3

ROS

+

DNA damage

O

OH

CH2OH OH NH2 O

OH

O OH O

adriamycin (doxorubicin) CH3

pro-caspases ⇒ caspases

−

Bcl-2

O

CELL DEATH PTP Bax

cytochrome c

+ + ROS H+

H+

H+

+ BMD 188

AIF Endo G Smac

Ca2+ M nS G OD Px

porin matrix

O2, H2O

respiratory chain intermembrane compartment

Figure 11 Mitochondrial pathway of apoptosis. The pathway is triggered by various “death signals”, as reactive oxygen species (ROS), DNA damage, and so on, that promote binding of the proapoptotic protein Bax with the outer mitochondrial membrane, probably at the contact sites between the two membranes, and its association with the permeability transition pore (PTP). This enables the release of cytochrome c (circles) and other proapoptotic proteins (the apoptosis-inducing factor AIF, endonuclease G, Smac, etc; squares) from the intermembrane compartment to the cytosol. An elevated intramitochondrial Ca2+ level and ROS production facilitate this process by promoting PTP opening. Once in the cytosol, cytochrome c, in cooperation with a cytosolic factor, Apaf-1 (not indicated), activates caspase-9 and subsequently other members of the caspase family, thus initiating self-digestion of the cell and nuclear DNA fragmentation, eventually leading to apoptotic cell death. Association of Bax with mitochondria is prevented by the antiapoptotic protein Bcl-2. ROS can be decomposed by Mn-containing (mitochondrial) and Cu,Zn-containing (cytosolic) superoxide dismutases (SOD), catalase, and glutathione peroxidase (GPx). Stimulation of ROS production is exempliﬁed here by ultraviolet and ionizing radiation and by two anticancer drugs, adriamycin and BMD188 [cis-1-hydroxy-4-(1-naphthyl)-6-octylpiperidine-2-one]. Activation is indicated by an encircled plus sign, and inhibition by an encircled minus sign. (Modiﬁed from Szewczyk and Wojtczak [11], with permission of the publisher.)

CONCLUDING REMARKS: MITOCHONDRIA AS A PHARMACOLOGICAL TARGET

31

membrane rupture and leakage of the intracellular content (ions, metabolites, and proteins). One of the major biochemical parameters determining the fate of cells challenged by life-threatening stimuli is their energy level. Whereas apoptosis needs a certain ATP content, necrotic cell death does not require energy. Presumably, a switch between apoptotic and necrotic cell death depends on the mitochondrial energy state and the extent of mitochondrial impairment. For example, it has been shown [54,55] that continuous ATP supply in hepatocytes challenged by ischemia or reperfusion stress can switch the cells from the necrotic to the apoptotic mode of dying. This conﬁrms that cell depletion of ATP is a commitment step for necrosis. Apart from apoptosis and necrosis, autophagy is another mode of elimination of unwanted cells [56]. Autophagy enables the removal of long-lived proteins and damaged organelles inside intracellular digesting vesicles. This mechanism offers recycling and salvage of intracellular material as well as delivering essential components during temporary interruption of nutrient supply. In this sense, autophagy is a pro-survival process. On the other hand, autophagy may act as a death mechanism, especially when apoptotic cell death is prevented [57]. Cellular autophagy-related signaling pathways have been studied intensely primarily in yeast, but their mechanisms have not yet been fully clariﬁed. Autophagy is also considered as a mechanism of removal of damaged organelles or those rendered unnecessary because of changing environmental or nutritional conditions [58]. It seems that mitochondrial turnover, which allows replacing of aged or impaired mitochondria, is based on their degradation in intracellular digestive vesicles, autophagosomes. It has been found, at least in yeast, that mitochondrial autophagy needs participation of the outer membrane Uth1p protein [59]. This points to a selective process. Mammalian analogs of Uth1p protein have not yet been found. It is suggested that apart from its role in apoptotic and necrotic cell death, mitochondrial permeability transition may also stimulate autophagy to remove damaged mitochondria in intact cells. This mechanism may protect the whole cell against apoptotic or necrotic death by decreasing the proportion of damaged mitochondria that display extensive ROS production, Ca2+ overload, and activation of the mitochondrial permeability transition pore. Thus, the PTP seems to contribute not only to cell death but may also trigger selective elimination of those mitochondria that may expose the cell to enhanced risk of apoptosis or necrosis [60].

6. CONCLUDING REMARKS: MITOCHONDRIA AS A PHARMACOLOGICAL TARGET The aim of this overview was to introduce multiple aspects of mitochondrial biology with particular attention to the roles these organelles play in cell survival and cell death. The importance of mitochondria in providing the cell with the energy required to maintain integrity and viability, and the mechanisms whereby

32

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

this is accomplished were discussed, as was the mitochondrial contributions to suicidal cell death, a process of vital importance for all multicellular organisms. Their predominant roles as ATP suppliers, as cellular ROS producers, and as important regulators of apoptosis, renders mitochondria promising targets for pharmacological interventions [11]. Currently, most efforts along these lines focus on preventing mitochondrial and cellular oxidative damage that arise from multiple pathological conditions, such as as ischemia and reperfusion-evoked cell damage, diabetes, and neurodegenerative diseases. Examples of such possible pharmacological interventions are (i) induction of mitochondria-dependent apoptosis with prospective importance for cancer therapy, (ii) controlling of mitochondrial permeability transition pore as a possible means for prevention of ischemia and reperfusion-related cell injury, (iii) decreasing mitochondrial membrane potential to increase oxidation of intracellular lipid deposits and reduce ROS production in treatments of obesity and diabetes [61], and (iv) moderate inhibition of the respiratory chain to limit ATP availability for hepatic gluconeogenesis in diabetes [62]. In addition, mitochondria are prospective targets for gene therapy in case of diseases caused by mutations in the mitochondrial genome [63]. Systemic administration of drugs selectively targeting mitochondrial functions presents a number of problems. Therefore, much attention has been paid to mitochondrially-targeted drugs, which may reach these organelles without affecting other intracellular structures and extramitochondrial processes. Such selective drug delivery may be accomplished by using speciﬁc carries that can bind to and enter mitochondria. Among them are delocalized lipophilic cations, which accumulate in the mitochondrial matrix or within the inner mitochondrial membrane at the expense of mitochondrial [64]. Other examples are small peptides that selectively partition to the inner mitochondrial membrane [65], liposomes consisting of self-assembling mitochondriotropic compounds [66], and chimeras composed of mitochondrial signalling peptides combined with other proteins or DNA [67]. Such selective mitochondria-targeted drug delivery seems to be the most promising approach to prevent or treat mitochondrial diseases. However, because of unresolved questions concerning drug delivery to appropriate organs and possible side effects of molecules used as drug carriers, these techniques are still at the stage of experimentation. Many of these issues will be discussed further in the ensuing chapters.

REFERENCES 1. 2. 3. 4.

Lehninger AL. The Mitochondrion. New York: WA Benjamin; 1964. Tzagoloff A. Mitochondria. New York: Plenum; 1982. Nicholls DG, Ferguson SJ. Bioenergetics 3 . London: Academic Press; 2002. Schefﬂer IE. A century of mitochondrial research: achievements and perspectives. Mitochondrion. 2001;1:3–31. ˚ resolution of 5. Abrahams JP, Leslie AGW, Lutter R, Walker JE. Structure at 2.8 A F1 -ATPase from bovine heart mitochondria. Nature. 1994;370:621–628.

REFERENCES

33

6. DiMauro S. Mitochondrial diseases. Biochim Biophys Acta. 2004;1658:80–88. 7. Kroemer G. Mitochondrial control of apoptosis. Biochem Soc Symp. 1999;66:1–15. 8. Kaufmann SH, Hengartner MO. Programmed cell death: alive and well in the new millenium. Trends Cell Biol. 2001;11:526–534. 9. Cortopassi GA, Wong A. Mitochondria in organismal aging and degeneration. Biochim Biophys Acta. 1999;1410:183–193. 10. Jacobson J, Duchen MR. Interplay between mitochondria and cellular calcium signalling. Mol Cell Biochem. 2004;256–257:209–218. 11. Szewczyk A, Wojtczak L. Mitochondria as a pharmacological target. Pharmacol Rev. 2002; 54:101–127. 12. Berry S. Endosymbiosis and the design of eukaryotic electron transport. Biochim Biophys Acta. 2003;1606:57–72. 13. Devlin TM. Textbook of Biochemistry, 5th ed. NewYork: Wiley-Liss; 2002. 14. Rondelez Y, Tresset G, Nakashima T, et al. Highly coupled ATP synthesis by F1 -ATPase single molecules. Nature. 2005;433:773–777. 15. Noji H, Yasuda R, Yoshida M, Kinosita K, Jr. Direct observation of the rotation of F1 -ATPase. Nature. 1997;386:299–302. 16. Itoh H, Takahashi A, Adachi K, et al. Mechanically driven ATP synthesis by F1 -ATPase. Nature. 2004;427:465–468. 17. Skulachev VP. Fatty acid circuit as a physiological mechanism of uncoupling of oxidative phosphorylation. FEBS Lett. 1991;294:158–162. 18. Wojtczak L, Wieckowski MR. The mechanism of fatty acid–induced proton permeability of the inner mitochondrial membrane. J Bioenerg Biomembr. 1999;31:447–455. 19. Jeˇzek P. Possible physiological roles of mitochondrial uncoupling proteins: UCPn. Int J Biochem Cell Biol. 2002;34:1190–1206. 20. Cannon B, Nedergaard J. Brown adipose tissue: function and physiological signiﬁcance. Physiol Rev. 2004;84:277–359. 21. Turrens JF. Mitochondrial formation of reactive oxygen species. J Physiol. 2003;552:335–344. 22. Adam-Vizi V, Chinopoulos C. Bioenergetics and the formation of mitochondrial reactive oxygen species. Trends Pharm Sci. 2006;27:639–645. 23. Skulachev VP. Role of uncoupled and non-coupled oxidations in maintenance of safely low levels of oxygen and its one-electron reductants. Q Rev Biophys. 1996;29:169–202. 24. Andreyev AY, Kushnareva YE, Starkov AA. Mitochondrial metabolism of reactive oxygen species. Biochemistry (Mosc). 2005;70:246–264. 25. Lebovitz RM, Zhang H, Vogel H, et al. Neurodegeneration, myocardial injury, and perinatal death in mitochondrial superoxide dismutase-deﬁcient mice. Proc Natl Acad Sci U S A. 1996;93:9783–9787. 26. Radi R, Turrens JF, Chang LY, Bush KM, Crapo JD, Freeman BA. Detection of catalase in rat heart mitochondria. J Biol Chem. 1991;266:22028–22034. 27. Mailer K. Superoxide radical as electron donor for oxidative phosphorylation of ADP. Biochem Biophys Res Commun. 1990;170:59–64. 28. Lenaz G. Role of mitochondria in oxidative stress and ageing. Biochim Biophys Acta. 1998;1366:53–67.

34

BASIC MITOCHONDRIAL PHYSIOLOGY IN CELL VIABILITY AND DEATH

29. Chan DG. Mitochondria: dynamic organelles in disease, aging, and development. Cell . 2006;125:1241–1252. 30. Sohal RS, Mockett RJ, Orr WC. Mechanisms of aging: an appraisal of the oxidative stress hypothesis. Free Radic Biol Med. 2002;33:575–586. 31. Speakman JR. Body size, energy metabolism and life span. J Exp Biol. 2005; 208:1717–1730. 32. Trifunovic A. Mitochondrial DNA and aging. Biochim Biophys Acta. 2006; 1757:611–617. 33. Gunter TE, Pfeifer DR. Mechanisms by which mitochondria transport calcium. Am J Physiol. 1990;258:C755–C786. 34. Malli R, Frieden M, Osibow K, Graier WF. Mitochondria efﬁciently buffer subplasmalemmal Ca2+ elevation during agonist stimulation, J Biol Chem. 2003;278:10807–10815. 35. Makowska A, Zabłocki K, Duszy´nski J. The role of mitochondria in the regulation of calcium inﬂux into Jurkat cells. Eur J Biochem. 2000;267:877–884. 36. Gilabert JA, Parekh AB. Respiring mitochondria determine the pattern of activation and inactivation of store-operated Ca2+ current I CRAC . EMBO J. 2000;19:6401–6407. 37. Tinel H, Cancela JM, Mogami H, et al. Active mitochondria surrounding the pancreatic acinar granule region prevent spreading of inositoltrisphosphate-evoked local cytosolic Ca2+ signals. EMBO J. 1999;18:4999–5008. 38. Duszy´nski J, Kozieł R, Brutkowski W, Szczepanowska J, Zabłocki K. The regulatory role of mitochondria in capacitative calcium entry. Biochim Biophys Acta. 2006;1757:380–387. 39. Hajn´oczky G, Csord´as G, Das S, et al. Mitochondrial calcium signalling and cell death: approaches for assessing the role of mitochondrial Ca2+ uptake in apoptosis. Cell Calcium. 2006;40:553–560. 40. Bernardi P. Mitochondrial transport of cations: channels, exchangers, and permeability transition. Physiol Rev. 1999;79:1127–1155. 41. Malli R, Frieden M, Osibow K, et al. Sustained Ca2+ transfer across mitochondria is essential for mitochondrial Ca2+ buffering, store-operated Ca2+ entry, and Ca2+ store reﬁlling. J Biol Chem. 2003;278:44769–44779. 42. Toescu EC, Verkhratsky A. Ca2+ and mitochondria as substratum for deﬁcits in synaptic plasticity in normal brain ageing. J Cell Mol Med. 2004;8:181–190. 43. Duchen MR. Mitochondria and calcium: from cell signalling to cell death. J Physiol. 2000;529:57–68. 44. Newmeyer DD, Ferguson-Miller S. Mitochondria: releasing power for life and unleashing the machineries of death. Cell . 2003;112:481–490. 45. Cereghetti GM, Scorrano L. The many shapes of mitochondrial death. Oncogene. 2006;25:4717–4724. 46. Orrenius S, Gogvadze V, Zhivotovsky B. Mitochondrial oxidative stress: implications for cell death. Annu Rev Pharmacol Toxicol. 2007;47:143–183. 47. Robertson JD, Zhivotovsky B, Gogvadze V, Orrenius S. Outer mitochondrial membrane permeabilization: an open-and-shut case ? Cell Death Differ. 2003;10:485–487. 48. Karbowski M, Youle RJ. Dynamics of mitochondrial morphology in healthy cells and during apoptosis. Cell Death Differ. 2003;10:870–880.

REFERENCES

35

49. Perfettini JL, Roumier T, Kroemer G. Mitochondrial fusion and ﬁssion in the control of apoptosis. Trends Cell Biol. 2005;15:179–183. 50. Arnoult D. Mitochondrial fragmentation in apoptosis. Trends Cell Biol. 2007;17:6–12. 51. Bernardi P, Krauskopf A, Basso E, et al. The mitochondrial permeability transition from in vitro artifact to disease target. FEBS J. 2006;273:2077–2099. 52. Grimm S, Brdiczka D. The permeability transition pore in cell death. Apoptosis. 2007;12:841–855. 53. Brookes PS, Yoon Y, Robotham JL, Anders MW, Sheu SS. Calcium, ATP, and ROS: a mitochondrial love–hate triangle. Am J Physiol. 2004;287:C817–C833. 54. Elmore SP, Qian T, Grissom SF, Lemasters JJ. The mitochondrial permeability transition initiates autophagy in rat hepatocytes. FASEB J. 2001;15:2286–2287. 55. Kim J-S, Qian T, Lemasters JJ. Mitochondrial permeability transition in the switch from necrotic to apoptotic cell death in ischemic rat hepatocytes. Gastroenterology. 2003;124:494–503. 56. Tsujimoto Y, Shimizu S. Another way to die: autophagic programmed cell death. Cell Death Differ. 2005;12:1528–1534. 57. Eskelinen E-L. Doctor Jekyll and Mister Hyde: autophagy can promote both cell survival and cell death. Cell Death Differ. 2005;12:1468–1472. 58. Lemasters JJ. Selective mitochondrial autophagy, as a targeted defense against oxidative stress, mitochondrial dysfunction and aging. Rejuvenation Res. 2005;6:3–5. 59. Kiˇssˇ ov´a I, Defﬁeu M, Manon S, Camougrand N. Uth1p is involved in the autophagic degradation of mitochondria. J Biol Chem. 2004;279:39068–39074. 60. Rodriguez-Enriguez S, He L, Lemasters JJ. Role of mitochondrial permeability transition pores in mitochondrial autophagy. Int J Biochem Cell Biol. 2004;36:2463–2472. 61. Armstrong JS. Mitochondrial medicine: pharmacological targeting of mitochondria in disease. Brit J Pharm. 2007;151:1154–1165. 62. Owen MR, Doran E, Halestrap AP. Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochem J. 2000;348:607–614. 63. DiMauro S, Hirano M, Schon EA. Approaches to the treatment of mitochondrial diseases. Muscle Nerve. 2006;34:265–283. 64. Skulachev VP. A biochemical approach to the problem of aging: “Megaproject” on membrane penetrating ions. The ﬁrst results and prospects. Biochemistry (Mosc). 2007;72:1385–1396. 65. Szeto HH. Mitochondria-targeted cytoprotective peptides for ischemia-reperfusion injury. Antioxid Redox Signal. 2008;10:601–619. 66. Weissig V, Lasch J, Erdos G, Meyer HW, Rowe TC, Hughes J. DQAsomes: a novel potential drug and gene delivery system made from Dequalinium. Pharm Res. 1998;15:334–337. 67. Chinnery PF, Taylor RW, Diekert K, Lill R, Turnbull DM, Lightowlers RN. Pepti de nucleic acid delivery to human mitochondria. Gene Ther. 1999;6:1919–1928.

2 BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION Immo E. Schefﬂer Section of Molecular Biology, Division of Biological Sciences, University of California–San Diego, La Jolla, California

1. Introduction 2. The mitochondrial genome 2.1. mtDNA in mammalian mitochondria 2.2. Replication of mtDNA 2.3. Transcription of the mitochondrial genome 2.4. Mitochondrial translation system 3. Import of proteins into mitochondria 3.1. Mitochondrial targeting signal 3.2. Translocation through and into the outer membrane 3.3. Translocation through the inner membrane 3.4. Assembly of the complexes in the inner membrane: supercomplexes 4. Fission of mitochondria and segregation during cell division 5. Control of mitochondrial biogenesis

37 40 40 42 46 48 52 53 54 55 56 58 60

1. INTRODUCTION Mitochondriology is enjoying a revitalization, fostered in large measure by the discoveries that they play central roles not only in cellular bioenergetics, but also in regulating cell death, and even more recently as unanticipated targets of many widely prescribed pharmaceuticals. Indeed, mitochondria have become the Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

37

38

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

subject of nonspecialist books targeting a wide popular audience [1]. Mitochondria are much more than a simple subcellular organelle providing the cell with ATP. Although oxidative phosphorylation is still one of their main functions, their power output is calibrated exquisitely to the needs of the cell. They allow cellular metabolism to be ﬁnely tuned by segregating enzymes and pathways in different compartments of the cell. They are also intimately involved in controlling the cytosolic Ca2+ ion concentration in conjunction with the endoplasmic reticulum (ER). Their positioning in relation to the ER can lead to temporal and spatial ﬂuctuations in the cytosol, allowing Ca2+ to exercise control over a variety of functions [2]. Their role in programmed cell death (apoptosis) has been elaborated in great detail in the past decade, and under certain abnormal/pathological conditions can provide signals [cytochrome c release, reactive oxygen species (ROS)] to terminate the life of a cell and thus protect the organism from cancer, for example. Finally, a constant level of exposure to ROS, although carefully kept in check by various mechanisms under normal circumstances, may nevertheless lead to an accumulation of oxidative damage in the mitochondrial genome to contribute to the senescence and death of an organism. All of the foregoing topics are worthy of much further exploration. Because of the broad spectrum of physiological reactions affected by mitochondria, there is much interest in using pharmacological approaches to manipulate certain activities of mitochondria, and the present volume bears testimony to the high level of activity in this ﬁeld. There is overwhelming evidence and a broad consensus that mitochondria are derived from a prokaryotic ancestor that became involved in a symbiotic relationship with another cell very early in the evolution of life on Earth, and this symbiosis is perhaps the deﬁning event in the evolution of eukaryotic organisms [3–6]. Animals, plants, fungi, and so on, all have mitochondria believed to be derived from a single original event. Over a period of 2 to 3 billion years, the prokaryotic symbiont lost most, but not all, of its genome and the capacity for an independent existence. A signiﬁcant number of its genes were transferred to the nucleus and presumably integrated into the chromosomes of the host. Others were lost, either because they were redundant or because their function was no longer required. At the same time, mechanism(s) evolved for proteins to be imported into the mitochondrion [7]. These include proteins encoded by the prokaryotic genes transferred to the nucleus but still required for functions inside the mitochondrion. In addition, proteins encoded by nuclear genes of the original host cell have also acquired the capacity to be taken up by the mitochondrion to become engaged in novel functions in their new environment [8]. Large databases are being assembled to list and characterize the mitochondrial proteome [9]. High-resolution fractionation techniques coupled with mass spectroscopy are being employed to characterize this proteome, with a further interest in learning how this proteome is differentiated in various tissues of an organism [10–14]. The present discussion focuses on mammalian mitochondria. However, many fundamental insights were gained from molecular genetic studies in

INTRODUCTION

39

microorganisms, especially the yeast Saccharomyces cerevisiae, and reference to such studies will be made frequently. Mammalian mitochondria contain a circular genome (mtDNA) with about 16,500 base pairs (bp); variations between species are minor. This genome encodes two ribosomal RNAs (rRNA) for the large and small ribosomal subunits, 22 tRNAs, and 13 proteins that are subunits of various complexes of the oxidative phosphorylation system. From proteomic studies it is estimated that mitochondria contain more than 1000 distinct proteins, of which about 600 to 800 have been identiﬁed by mass spectroscopy. Thus, the vast majority of proteins in mitochondria are imported. Nevertheless, a highly reduced genome, a transcriptional apparatus, and mitochondrial translation machinery are necessary to synthesize 13 essential proteins. These proteins are essential for oxidative phosphorylation, serving as the active sites of several of the respiratory complexes. Without oxidative phosphorylation, all mammalian organisms (and most metazoans) are not viable. However, in tissue culture, mammalian cells (ﬁbroblasts) can be propagated after mitochondrial respiration and ATP production are abolished by nuclear mutations, as ﬁrst shown by Schefﬂer’s laboratory [15]. Subsequently, mammalian cells without mtDNA have been established in tissue culture in Attardi’s laboratory [16] and by others. More than 100 years ago, mitochondria were ﬁrst seen under the light microscope and misidentiﬁed as bacteria living inside cells. A part of this conclusion was far ahead of its time. With our present understanding we recognize their origin and relationship to prokaryotes, but it is clear that they cannot multiply independently outside a cell. At the same time, their number must increase in cells that are progressing through the cell cycle, to be distributed equally among daughter cells. In this chapter we deal with the biogenesis of mitochondria. Overall and superﬁcially, the process still resembles the proliferation of bacteria: mtDNA is replicated, mitochondrial mass and volume are increased by protein synthesis in the matrix and import of proteins from the cytosol, phospholipids are imported or synthesized to enlarge mitochondrial membranes, and eventually the organelle divides by ﬁssion. Each of these processes is described in more detail below. It might be noted that in contrast to bacteria, mitochondria can also fuse with each other, and for reasons that are not yet entirely clear, fusions and ﬁssions are in a highly dynamic equilibrium during the life of a cell. A simple but perhaps too simpleminded explanation is that these processes serve to homogenize the contents of mitochondria continuously. At the same time, in heteroplasmic organisms (where mitochondria contain two genetically distinguishable populations of mtDNA), a segregation of genotypes (a shift in the ratio of heteroplasmy) can be observed either from one generation to the next, or in somatic cells during the growth and development of the mature organism. This issue is expanded further below. In the following sections, mtDNA replication, the transcription of the mitochondrial genome, the mitochondrial translation machinery, and import of proteins from the cytosol into mitochondria are described in some detail. We also

40

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

discuss brieﬂy the morphological changes associated with the “replication” of mitochondria.

2. THE MITOCHONDRIAL GENOME 2.1. mtDNA in Mammalian Mitochondria Although the phenomenon of cytoplasmic inheritance was discovered much earlier in yeast [17], mitochondrial DNA was deﬁnitely identiﬁed in mitochondria in 1963 by electron microscopy [18]. A few years later it was isolated biochemically [19,20] and characterized in a diverse group of metazoans as a circular DNA with about 16,500 bp. Some authors have referred to it as the “twenty-ﬁfth chromosome” in humans, and it was the ﬁrst human “chromosome” to be completely sequenced, in 1981 [21]. By now, hundreds of metazoan mtDNAs have been sequenced; the information is assembled and made available in various databases (e.g., MmtDB at the University of Bari, Italy) [22]. A human mitochondrial DNA database has been set up [23] and is accessible on the Internet (www.mitomap.org). This site has links to a large number of human sequences from different ethnic groups, and documents polymorphisms, point mutations, and deletions in persons afﬂicted with a mitochondrial disease, genetic maps, related databases, and much more useful information related to the molecular genetics of mtDNA. Publication of the complete sequence for human (and bovine) mtDNA did not immediately reveal all the genes in this genome, because it was not immediately recognized that codon use in mitochondria is different from the universal genetic code. The two rRNA genes and most protein-coding sequences were identiﬁed from homologies with known proteins in other organisms. Finally, two rRNA genes, 22 tRNA genes, and 13 structural genes encoding proteins were resolved [24,25]. The tRNA genes are dispersed around the genome and frequently serve as “punctuation marks” between structural genes. Both heavy and light DNA strands encode genetic information; on one of the strands it is packaged in a highly economical manner, with tightly packed reading frames and overlapping reading frames in some instances. A map of the human mtDNA is presented in Figure 1. As described in more detail in a following section, the mtDNA is transcribed in both directions into large transcripts that are processed to produce the individual rRNAs, tRNAs, and mRNAs. A notable region of about 1120 bp (in humans) is referred to as the control region. It is not expressed, but contains short sequences acting as origins of DNA replication and promoters for transcription. The control region is not very well conserved among mammals, and even in the human population there are many sequence polymorphisms that have been exploited in very interesting forensic and anthropological studies [26]. In vertebrates this control region is ﬂanked by two tRNA genes (tRNAThr and tRNAPhe ; the tRNAPro gene is adjacent to the tRNAThr gene on the other strand). Because of the dynamic behavior of mitochondria, the number of mitochondria per cell is difﬁcult to describe. In extreme cases, or in the presence of speciﬁc

41

THE MITOCHONDRIAL GENOME small rRNA Phe Val large rRNA

Thr cytochrome b

Leu ND1 Glu lle f-Met ND2

ND5 Gln Ala Asn

Trp

Leu Ser His

ND4 CO

Arg

Asp COX2

Lys ATPase8

Gly

ND3

COX3 ATPase6

Figure 1 Map of the human mtDNA. mtDNA is transcribed in both directions into large transcripts that are processed to produce the individual rRNAs, tRNAs, and mRNAs. A notable region of about 1120 bp (in humans) is referred to as the control region. It is not expressed but contains short sequences acting as origins of DNA replication and promoters for transcription.

genetic defects, mitochondria are able to form a continuous reticulum or break up into many small organelles. A more meaningful number is to specify the copy number of mtDNAs per cell (i.e., to normalize the amount of mtDNA relative to the amount of DNA in a diploid nucleus). This index varies widely, from 100 to 10,000, depending on the tissue, and accurate numbers have not been measured for a variety of human tissues. The number of mitochondria that can be distinguished morphologically is lower, and therefore each mitochondrion contains more than one genome. Parenthetically, it should also be noted that the morphology of the inner mitochondrial membrane is highly variable in different tissues [26], and another important parameter is the total surface area of the inner membrane, as reﬂected by the number and shape of the cristae. mtDNA does not exist in the organelle in a “naked” form, but it also is not packaged into nucleosomes, since there are no histones in mitochondria. More recently it has become clear that mtDNA is tightly associated with a number of diverse proteins responsible for packaging, replication, and transcription, forming a structure termed a nucleoid [27–31]. Nucleoids have been visualized by

42

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

staining for DNA, or by using the green ﬂuorescent protein (GFP) as a reporter when fused to nucleoid-associated proteins such as mitochondrial transcription factor A (mtTFA). They typically contain more than one mtDNA and are found in contact with the mitochondrial inner membrane. One can speculate that such an attachment serves in the segregation of mtDNA during ﬁssions, or that transcription and translation are tightly coupled to permit a cotranslational insertion of mtDNA-encoded proteins into the inner membrane. The presence of aconitase in such nucleoids [30] is an intriguing observation. It could represent a link between the metabolic activity of mitochondria (the Krebs cycle) and activity in the nucleod (transcription and translation?). Dominant proteins found in the nucleoid are the transcription factors mtTFA and mtTFB. These factors are described below in a different context, but at present it is noteworthy that they contain segments of about 70 amino acids with a semblance to the bacterial HU protein, which regulates gyrase activity, among other effects, or eukaryotic nuclear high mobility group (HMG) proteins [32,33]. Their relative abundance suggests that they have functions beyond that of a transcription factor binding to two promoters in the control region. 2.2. Replication of mtDNA The mitochondrial genomes have to be replicated during the cell cycle. Initial studies determined that mtDNA replication was “relaxed”: It was not restricted exclusively to the S-phase of the cell cycle when nuclear DNA replication takes place, and it appears that not every mtDNA molecule is replicated during the cell cycle, and some may be replicated more than once. On the other hand, a mechanism must exist to “count” mtDNA and to keep the copy number more or less constant in a given cell type and tissue. How this is achieved remains to be explored in detail. In particular, little is known about the in vivo situation. In one report the human mtTFA was overexpressed in transgenic mice, with the result that the mtDNA copy number was elevated signiﬁcantly [34]. As pointed out by these authors, the human transcription factor is not functioning in transcriptional activation in this heterologous system. It is concluded that the general DNA binding activity of human-mtTFA must play a role in determining the DNA copy number. It was also noted that the increase in copy number did not coincide with elevated mitochondrial mass or a higher rate of oxidative phosphorylation. In cells in culture, there have been experiments in which conditions were manipulated to inﬂuence the copy number. Transient overexpression of mtTFA in tissue culture was reported to stimulate transcription but not to inﬂuence the copy number [35], in contrast to the report on transgenic mice. There have also been experiments investigating the effect of oxidative stress on DNA copy number. A recent review of such studies has been published by Lee and Wei [36]. It appears that mild oxidative stress can stimulate mitochondrial biogenesis but that severe stress causes apoptosis. Various signaling molecules (Ca2+ , cAMP, NO) and signal transduction pathways have been implicated (we discuss the control of mitochondrial biogenesis further in Section 5. It should be kept in mind

THE MITOCHONDRIAL GENOME

43

that there are several potential variables affecting the capacity of mitochondria for oxidative phosphorylation: mtDNA copy number, transcriptional initiation, mitochondrial mass, cristae formation, and potential protein modiﬁcations. In a very elegant series of experiments, Davis and Clayton, using short labeling periods with BrdU, demonstrated that mtDNA was replicated preferentially in mitochondria located around the nuclear periphery [37]. The presence of a nucleus was required, but ongoing nuclear DNA replication was not. Incorporation of the analog into mtDNA was not observed in enucleated cells or in platelets, but it did occur in terminally differentiated cells. It was speculated that factors restricted to the vicinity of the nucleus play an essential role, but their identiﬁcation remains to be achieved. After prolonged labeling, or following a chase, BrdU-labeled mtDNA was also found in more distant mitochondria. One explanation is that mitochondria are moved around the cell with the help of motor proteins (dynein, kinesin) and interactions with the cytoskeleton. The mechanism of mtDNA replication is relatively well understood. Issues addressed here include (1) the enzymes required, (2) the origin(s) of DNA replication, (3) initiation and priming, and (4) the direction of replication. Enzymes There is a major and speciﬁc mitochondrial DNA polymerase belonging to the polymerase γ family [38–40]. It is believed that this enzyme functions not only in DNA replication but also in mtDNA repair and in recombination, although the latter is not well established in mammalian mitochondria. The low abundance of this enzyme proved to be a challenge, and the yeast enzyme was the ﬁrst to be characterized by a combination of genetic and biochemical studies. The human enzyme is now also well characterized [39–44]. The relationship of this polymerase to the various bacterial and eukaryotic DNA polymerases is discussed insightfully by Kornberg [45]. The polymerase proceeds in the usual 5 to 3 direction. In early in vitro studies it was found to prefer ribohomopolymer templates and was therefore thought to resemble the reverse transcriptase of tumor viruses. However, it was found to be antigenically completely different from viral reverse transcriptases. The highly puriﬁed enzyme revealed the existence of two subunits: PolγA (125 to 140 kDa) and PolγB (35 to 54 kDa). The PolγA subunit was established to have both the polymerase and exonuclease activities [46] associated with distinguishable domains that have recognizable homology to domains in the prokaryotic A-type DNA polymerases (e.g., Escherichia coli DNA polymerase I). Korhonen and colleagues have achieved the in vitro reconstitution of a minimal mtDNA replisome [47]. When provided with a synthetic replication fork, three proteins were found to be sufﬁcient to carry out strand elongation at a rate that was close to that reported for the in vivo reaction. These include the polymerase γ, a single-strand binding protein (mtSSB), and Twinkle, a mitochondrial helicase with 5 to 3 directionality. The system and mechanism bears a strong resemblance to the phage T4 and T7 replisomes active in bacteria [48]. Shutt and Gray [49] have surveyed existing databases and identiﬁed Twinkle homologs in a wide spectrum of eukaryotic lineages. These proteins probably evolved from an ancestral protein related to the bifunctional primase–helicase

44

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

of T-odd bacteriophages. Curiously, conserved primase motifs were found in all eukaryotic Twinkle proteins with the exception of metazoans. The authors conclude that Twinkle serves as the primase as well as the helicase for mtDNA replication in most eukaryotes, but clearly a primase function remains to be characterized in metazoans (see below). In humans a mutation in Twinkle has been found to be responsible for autosomal dominant progressive external ophthalmoplegia (adPEO). The defect causes the appearance and accumulation of mtDNA deletions [50]. Origins of mtDNA Replication There is an asymmetric distribution of nucleotides (G + C) in the mtDNA strands leading to a separation of “heavy” and “light” strands on alkaline CsCl gradients. It therefore became possible to show that initiation and elongation of the H-strand preceded the L-strand synthesis. The origin of H-strand synthesis (OH ) is located in the control region. The second origin (OL ) is not an independent origin but must be unmasked by elongation of the H-strand and creation of a displaced, single-strand DNA. It is located at some considerable distance from OH within a region of ﬁve tRNA genes, leading to speculation that a speciﬁc secondary structure must be formed by the single strand, which can now function as an origin [46,51]. As described, the model has been referred to as the strand displacement model . It was challenged by an examination of replication intermediates by two-dimensional gel electrophoresis in which the results could be interpreted to support a synchronous strand-coupled mode of mtDNA replication [52–54]. In response, Brown and others reexamined mtDNA replication intermediates from mouse liver with the help of atomic force microscopy and two-dimensional agarose gel electrophoresis [55,56]. Their observations favor the original strand displacement model, with the modiﬁcation that there are multiple (alternative) origins of lagging strand synthesis. To resolve the conﬂict, it is suggested that the conditions for electrophoretic separation of intermediates appear to favor branch migration of asymmetrically replicating circular mtDNA molecules, which can obscure the analysis. A new twist in this story was recently introduced in a paper from Attardi’s laboratory [57] describing another major origin of DNA replication within the region deﬁned by the D-loop. The D-loop was found originally during the characterization of mtDNA molecules by electron microscopy. Molecules were frequently observed that had a bubble, or D-loop, and such species were interpreted to represent replication intermediates in which heavy-strand DNA elongation from the OH origin was arrested (i.e., the loop represented the displaced single strand). The striking ﬁnding was that the D-loop had a relatively constant size of about 1000 nt. This gave rise to speculations that the arrest at that position was signiﬁcant, perhaps serving a control function. For example, a rate-limiting mechanism could be the release of the replisome from this speciﬁc arrest. Fish and colleagues suggest that the new origin is the true origin under normal conditions, and the previously deﬁned origin OH is perhaps used when increased mtDNA replication is stimulated by novel or abnormal physiological demands. The issue remains to be resolved.

45

THE MITOCHONDRIAL GENOME

Initiation and Priming From a study of RNA transcripts of the mitochondrial genome it became apparent that there are two promoters for transcription in opposite directions. These are referred to as light-strand promoter (LSP) and heavy-strand promoter (HSP), and both are located within the control region about 150 bp apart. A faithful in vitro transcription system has now deﬁned them very precisely (see Section 2.3). It became a good guess, and it was soon conﬁrmed that an RNA polymerase starting at the LSP was responsible for initiating the synthesis of an RNA primer, which was then extended by the DNA polymerase. First, newly initiated DNA strands had a 5 end corresponding to a sequence starting about 200 nt downstream from the LSP. Among the LSP transcripts there were longer molecules to be processed to yield tRNAs and mRNAs, and short forms with 3 ends adjacent to the OH origin. Most convincing was the isolation of LSP transcripts still attached to nascent heavy-strand DNA (see Shadel and Clayton [46] for a summary of these pioneering experiments). A schematic model is shown in Figure 2. Initiation and RNA elongation from the LSP by RNA polymerase (plus factors) create a transcript and a small displacement loop (R-loop) further deﬁned by conserved sequence blocks (CSBs). The RNA is cleaved in two positions, leaving an RNA–DNA

RNA pol

RNA pol

5′

R-loop RNA pol

MRP 5′

5′

D-loop

RNA primer

nascent H-strand (DNA)

Figure 2 Two promoters, light-strand promoter (LSP) and heavy-strand promoter (HSP), are responsible for transcription of mtDNA in opposite directions. Both are located within the control region about 150 bp apart. RNA polymerase starting at the LSP is responsible for initiating the synthesis of an RNA primer which is then extended by DNA polymerase.

46

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

“hybrid” of unusual stability that can be isolated after digestion with restriction enzymes. Physical–chemical experiments in vitro have conﬁrmed its stability and have suggested that the entire structure cannot be represented by Watson–Crick base-pairing schemes in which RNA is base-paired with DNA along its entire length. Instead, hairpin loops within the RNA are thought to interact with distinct sites on the DNA strand containing poly(G) tracts. Further support for such an unconventional structure was obtained by probing with RNase H and RNase T1. The RNA in the R-loop is ultimately cleaved by a speciﬁc endonuclease to create an RNA primer for extension by DNA polymerase. This endonuclease (RNase MRP) was ﬁrst identiﬁed in Clayton’s laboratory [58–60] as a nuclear-encoded enzyme with the unusual requirement for a small RNA “cofactor” that is presumed to aid in the substrate and site selection for endonucleolytic cleavage. This RNA is highly conserved in vertebrates. In yeast a similar enzyme has been localized to mitochondria. In this organism it was technically possible to mutate the nuclear gene encoding this RNA cofactor to demonstrate that it is required for mitochondrial biogenesis. Repair of mtDNA Until about a decade ago it was the “conventional wisdom” that mitochondria had a very low capacity for repairing damage in mtDNA. With the discovery and characterization of several relevant enzymes and repair mechanisms, the view has changed. Mutations induced by reactive oxygen species can be removed by a base excision pathway [61–63]. This repair system has been claimed to be inducible by oxidative stress [63]. Evidence for mismatch repair is sparse [64]. The proofreading capacity of DNA polymerase γ is clearly an important function, as demonstrated most dramatically in mice made homozygous in a repair-deﬁcient DNA polymerase γ by gene knock-in [65,66]. 2.3. Transcription of the Mitochondrial Genome The study of transcription of the mitochondrial genome was pioneered in Attardi’s laboratory with the identiﬁcation of the transcripts, intermediates, and ﬁnal mature rRNAs, tRNAs, and 13 mRNAs from kilogram quantities of HeLa cells [67]. The early history has also been ably reviewed by another key investigator, D. Clayton [60]. As the sequences of mammalian mtDNA became available during the same time frame, it was clear that the genes were very compressed on the DNA, leaving no room for promoter elements unless they were also transcribed to become part of the coding sequences. The problem was solved by demonstrations that mtDNA was transcribed from two promoters located not very far apart (about 150 nt) in the control region. The breakthrough was achieved by development of the ﬁrst in vitro system for initiating transcription from mtDNA [68]. From these promoters (LSP and HSP) the heavy and light strands could be transcribed in opposite directions, yielding long polycistronic transcripts from the entire length of the mtDNA. Endonuclease cleavages and some further posttranscriptional modiﬁcations are required to produce the mature rRNAS, tRNAs, and mRNAs. Thus, transcription from the L-strand yields a transcript that after processing gives rise

THE MITOCHONDRIAL GENOME

47

to the ND6 mRNA and eight tRNAs. The polycistronic RNA encoded by the H-strand is processed to the 12S and 16S rRNAs, the other 14 tRNAs, and 12 mRNAs. Somewhat later it was found that transcription from the HSP starts at two closely spaced initiation sites with differing activities, and the more abundant transcript is terminated at the 3 end of the 16S rRNA. Such a mechanism assures that rRNAs are produced in excess of the mRNAs to produce an abundant supply of ribosomes in the matrix for protein synthesis. It should be noted that these studies were carried out with human or mouse mtDNA. There were many similarities with regard to the size of the control region and the arrangement of relevant sequence elements within it, but there was little nucleotide sequence conservation. Therefore, it was perhaps not too surprising to ﬁnd that human mitochondrial extracts and enzymes failed to transcribe mouse mtDNA, and vice versa. The search was on for the required enzymes and factors, culminating in the identiﬁcation and characterization of the mtRNA polymerase (PolRMT) and a transcription factor now referred to as mtTFA (previously, Tfam). After successful cloning of the gene for a dedicated mitochondrial RNA polymerase from yeast, the human cDNA for the enzyme was identiﬁed and isolated by homology cloning [69]. Sequence comparisons showed that these mitochondrial RNA polymerases were quite similar in lower and higher eukaryotes and clearly related to the RNA polymerases of the bacteriophages T3 and T7. A notable difference was that the mitochondrial enzyme could not bind to DNA in the absence of additional factor(s). Initially, only the single factor mtTFA was thought to be necessary, and its gene in the mouse was cloned and characterized in 1997 by Larsson et al. [70]; the complete human gene was described by Reyes et al. [71]. Somewhat later a second transcription factor, mtTFB1, was identiﬁed [72], followed by the discovery of another related but not redundant factor, mtTFB2 [72,73]. mtTFA is a protein of 25 kDa belonging to the family of high-mobility-group (HMG)-box proteins of the nucleus. There are two segments, HMG box 1 and HMG box 2, comprised of about 70 amino acids each and separated by about 30 residues. Like other HMG-box proteins, the mtTFA protein can, without speciﬁcity, bind DNA, unwind it, and bend it. The mammalian mtTFA has a 25-residue carboxy-terminal tail that is essential for transcriptional activation and for the recognition of a more speciﬁc DNA sequence from the mitochondrial promoters. A detailed analysis of the function of each of these proteins became possible with the development of an in vitro transcription system [73] using highly puriﬁed recombinant proteins. The minimal requirements were a DNA fragment containing the promoter (from the control region of mtDNA), PolRMT, mtTFA, and either mtTFB1 or mtTFB2. The latter turned out to be about 100 times more active than mtTFB1. Either mtTFB1 or mtTFB2 can form a heterodimer with PolRMT. An unexpected discovery was that these factors belong to a family of rRNA methyltransferases that are thought to methylate some speciﬁc residues in the small rRNA of mitochondrial ribosomes, using S -adenosylmethionine (SAM) as methyl donor. The methyltransferase of human mtTFB1 is functional, but the activity can be abolished by speciﬁc mutations without affecting the ability of

48

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

mtTFB1 to stimulate transcription. Thus, mtTFB1 and mtTFB2 must be capable of binding (single-stranded) RNA, and it is inferred that they might also bind single-strand DNA. The most detailed current model has mtTFA binding to mtDNA at or near the HSP or LSP, causing a bend or contortion that unwinds DNA, creating a docking site for the mtTFB1/2-PolRMT heterodimer [28,74]. This model owes much to insights gained from studies of initiation by the phage N4 RNA polymerase II [75]. There are also indications of direct interaction between mtTFA and mtTFB. Transcription termination is another problem remaining to be elucidated in greater detail. On the one hand, the long transcripts from HSP and LSP are made by the progression of the polymerase all the way around the circular genome, and one might even imagine this process continuing while RNA cleavage and maturation proceeds simultaneously. However, as indicated above, there is also an abundant transcript containing the 12S and 16S rRNA sequences that terminates at the 3 end of the 16S rRNA. A mitochondrial transcription termination factor mTERF has been characterized in vitro [76,77], but questions remain about the true in vivo function of this protein and its mechanism. Finally, an interesting problem arises from the recognition that several of the proteins associated with mitochondrial DNA replication and transcription are related to bacteriophage proteins. Speciﬁcally, the PolRMT does not have any relationship to the multisubunit RNA polymerases found in bacteria, but resembles the monomeric phage RNA polymerases, and the helicase Twinkle appears related to the phage T7 gene 4 protein. What happened during the evolution of mitochondria? One hypothesis proposes that the original bacterial RNA polymerase was transferred to the nucleus and subsequently made redundant by acquisition of the phage polymerase. Did a bacteriophage infect the endosymbiont bacteria at an early stage? It has been pointed out that in chloroplasts both monomeric phage-related polymerases and multisubunit bacterial-type polymerases can be found [78]. 2.4. Mitochondrial Translation System The mitochondrial translation system in the mammalian organelle is responsible for the synthesis of 13 proteins that constitute subunits in the complexes (I to IV) of the electron transport chain and ATP synthase (V). Seven subunits are found in complex I (NADH–ubiquinone oxidoreductse), one subunit is found in complex III (the bc1 complex, or ubiquinone–cytochrome c oxidoreductase), three subunits form the core of complex IV (cytochrome c oxidase), and two subunits are part of the ATP synthase (complex V). They are absolutely essential for oxidative phosphorylation. Failure to make any of these proteins is lethal in mammals, and when present at levels below a certain threshold in humans one can observe a variety of symptoms typical of mitochondrial disease. It is beyond the scope of this review to go into more detail on the pathology (see Chapter 11). One can consider these proteins and their coding sequences on mtDNA to be a remnant of the prokaryotic endosymbiont, with the possibility that during

THE MITOCHONDRIAL GENOME

49

continuing evolution more of these genes would be transferred to the nucleus. The four genes for complex II (succinate–ubiquinone oxidoreductase) are all in the nucleus in vertebrates, but in some eukaryotic microorganisms one or more of these genes are still part of the mitochondrial genome [26]. Arguments have been forwarded to rationalize why these 13 genes cannot be transferred to the nucleus; these integral membrane proteins may have too many transmembrane helices, which would make import into the inner membrane impossible. Although critical for functional respiration, the existence and biogenesis of mitochondria in mammalian cells is not dependent on mitochondrial protein synthesis, since near-normal mitochondria can be seen in mutant ﬁbroblasts completely defective in mitochondrial protein synthesis [79–81] or in mammalian cells in culture without mtDNA [16,82]. Although it was recognized and proved experimentally more than 20 years ago that mitochondria synthesize a limited number of proteins with completely autonomous translation machinery, it is worth noting that to this date there has not been a report on a successful in vitro system with exclusively mitochondrial components: mRNA, ribosomes, tRNA, initiation and elongation factors. The initial attempts were made with mixed components from the cytosol and from mitochondria. With the advent of more sequence data it was realized that translation in mitochondria did not conform to the universal triplet codon usage. With only 22 tRNAs encoded by mtDNA, the code must be highly degenerate. (The import of tRNAs into mitochondria was discovered later in other organisms, but it does not occur in mammalian mitochondria.) The translation of codons is somewhat variable in mitochondria from animals, plants, fungi, and insects. The mature mitochondrial tRNAs are produced by a series of posttranscriptional reactions. First, they are cut out from the polycistronic transcripts produced by the PolRMT. The 3 end has to be extended by the addition of some nucleotides to create the typical 3 CCA end for amino acylation by tRNA synthetase. Many of the tRNAs are further modiﬁed by a variety of enzymes to produce, for example, pseudouridine, methylated bases, and N 6 -isopentenyl adenosine [83]. When their primary sequences became known completely, the classical cloverleaf secondary structure frequently has to be modiﬁed by the omission of one of the arms (D-loop or pseudo U-loop) in order to preserve an anticodon loop. Mitochondrial ribosomes are smaller that their cytoplasmic counterparts. The rRNA in the large subunit is about 1600 nt, compared to 4800 nt in the cytosolic subunit, and in the smaller subunit the rRNA has about 950 nt (about 1900 nt in the cytosol). From a determination of buoyant densities in CsCl gradients, their protein/nucleic acid ratio is slightly higher than that of their cytoplasmic versions. About 80 ribosomal proteins are encoded by nuclear genes and imported into mitochondria for assembly into ribosomes. Of these, about half have sequence homology to prokaryotic ribosomal proteins, while the other half are uniquely mitochondrial. Only a detailed comparison of high-resolution structures of the various types of ribosomes will reveal the signiﬁcance of these differences. It appears that a limited number of key residues of mitochondrial rRNAs must be glycosylated based on conserved sequences and on genetic experiments in

50

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

yeast. In Section 2.3 we introduced the factors mtTFB1 and mtTFB2 and their resemblance to methyl transferases. One or both may have dual roles in mammalian mitochondria in transcriptional activation and ribosome maturation. In contrast to cytoplasmic ribosomes, mitochondrial ribosomes (e.g., prokaryotic ribosomes) can be inhibited by the antibiotic chloramphenicol. The opposite is found with the inhibitors emetine or cycloheximide. This ﬁnding was one of the early arguments in favor of the endosymbiont hypothesis. It has been and continues to be of great practical value when one wants to study mitochondrial protein synthesis in intact cells in the absence of cytoplasmic protein synthesis. The ﬁrst cytoplasmic mutations in animal cells in culture (i.e., mutations in mtDNA) were established by selecting chloramphenicol-resistant mouse cells [84,85]. As expected, the mutation(s) could later be shown to occur in rRNA genes. These pioneering studies paved the way for the isolation of other mtDNA mutations in tissue culture [86–88] and suggested that a very signiﬁcant segregation of different alleles in mtDNA can occur under selective pressure, even with more than 1000 copies per cell. About 10 years later the segregation of mutant alleles was demonstrated in human pedigrees, and the concept of inherited “mitochondrial diseases” was formed. Aminoglycosides are antibiotics useful for treatment of bacterial infections by targeting the prokaryotic ribosomes. Extensive use of these antibiotics in patients with hearing loss is associated with worsening of this symptom. A genetic or pedigree analysis of patient families uncovered a mutation in the mitochondrial 12S rRNA gene in a large percentage [89–91]. The expression of the phenotype is also dependent on the combination of the mitochondrial mutation with mutations in nuclear “modiﬁer genes.” One of these hypothesized modiﬁer genes has been identiﬁed as the TRMU gene that encodes an enzyme required for the modiﬁcation or maturation of one or more mitochondrial tRNAs [91]. The processing of the large polycistronic transcripts in the mammalian mitochondrial matrix also gives rise to 13 mRNAs. The mature mRNAs have no 5 cap, but they are polyadenylated. The addition of adenines in some cases is necessary to create the stop codon, UAA. Mitochondrial mRNAs are not only lacking 3 -untranslated regions (3 UTR), but they are also largely devoid of 5 UTRs. The open reading frame (ORF) occupies most of the mRNA with the exception of the poly(A) tail. This situation can be contrasted with that in yeast or plants, where relatively large 5 UTRs are found. Experiments in yeast show that the 5 UTRs do play a signiﬁcant role in initiation of translation of speciﬁc mRNAs [92–94]. Unfortunately, the very elegant and revealing genetic studies in S. cerevisiae have so far not been able to shed much light on the mechanism of initiation of translation in mammalian mitochondria. Attempts to achieve efﬁcient initiation of protein synthesis in vitro with a proper mitochondrial mRNA have failed so far. One can speculate about the reasons, and a plausible explanation is based on the fact that proteins encoded by the mtDNA are all very hydrophobic integral membrane proteins with multiple transmembrane helices. It is probable that their synthesis and translocation into the membrane occurs by a cotranslational

THE MITOCHONDRIAL GENOME

51

mechanism. Initiation of protein synthesis may require the assembly of machinery capable of inserting the nascent peptide into the membrane, something that has not been achieved in vitro. Thus, the active ribosomes must become associated with the inner mitochondrial membrane, probably to proteins already in the membrane or via peripheral membrane proteins. It is not clear whether these mtDNA-encoded proteins are inserted into the membrane independently or in a process that is coupled directly with their assembly into complexes of the electron transport chain. These complexes contain additional subunits imported from the cytosol. The translation of the three mRNAs encoding Cox1p, Cox2p, and Cox3p (of complex IV) in yeast requires three speciﬁc proteins that interact speciﬁcally with the 5 UTRs (which are quite long in this organism). These proteins are membrane associated and they can be cross-linked to each other [94]. Subsequently, speciﬁc translation factors have been demonstrated in ﬁve of seven integral membrane proteins encoded by yeast mtDNA [95,96]. To propose a similar set of proteins and functions in mammalian mitochondria is at this point problematic when the corresponding mRNAs have virtually no 5 UTRs. Another aspect to be considered is the existence of nucleoids on the membrane containing mtDNA and a variety of enzymes responsible for DNA replication and transcription (see above). It is conceivable that initiation of translation begins on nascent mRNAs within this nucleoid, but too little is known at this time. Nevertheless, guided by insights from prokaryotic and cytosolic systems, considerable progress has been achieved in understanding the mitochondrial translation system. In bacteria, three initiation factors (IF1, IF2, and IF3) are needed; initiation in the cytosol of mammalian cells requires eIF1A, eIF2, eIF3, eIF4, and eIF5. Each of these consists of multiple subunits [97,98]. In prokaryotes the small ribosomal subunit is positioned directly over the start codon guided by interactions with the Shine–Dalgarno sequence a short distance upstream from the start codon. In the eukaryotic cytosolic mRNA the 5 cap acts as the initial binding site of eIF4, and an initiation complex including the small ribosome is assembled which then “scans” the 5 UTR to ﬁnd the start codon. Following the addition of the large ribosomal subunit, elongation proceeds by a mechanism that is similar in both prokaryotes and eukaryotes. Which initiation model is followed by the mitochondrial translation system? Starting with sufﬁcient bovine liver to produce about 30 g of mitochondria [99], Schwartzbach et al. have performed pioneering biochemical studies to purify and characterize required factors (see [100] for a review). The present list includes three elongation factors (EF-Tumt , EF-Tsmt , and EF-Gmt ) and two initiation factors (IF-2mt and IF-3mt ). The elongation factors tend to have considerable homology to prokaryotic factors, and the puriﬁed mitochondrial factors have in fact been shown able to replace the corresponding E. coli factors. For example, EF-Tumt can act with bacterial ribosomes and poly(U) to synthesize poly-phenylalanine, EF-Gmt can substitute for the E. coli EF-G in the same system, and EF-Tsmt is active in combination with the E. coli EF-Tu to stimulate the exchange of guanine nucleotides. As pointed out in an authoritative review [100], enough information is now available on primary sequences and the orthologous sequences from other

52

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

organisms to face the challenge of how these factors interact with the often unusual secondary and tertiary structures of the mitochondrial tRNAs. Our understanding is less complete with respect to the initiation factors. IF-2mt (85 kDa) is active in promoting the binding of f-met-tRNA to mitochondrial ribosomes and poly(A,U,G) as template, very reminiscent of the role of the bacterial IF-2. When the cDNA was cloned and sequenced, the protein sequence revealed a number of domains that are functionally conserved from the prokaryotic ancestor. IF-3mt could not be puriﬁed by a biological assay with a mitochondrial extract, but its cDNA was cloned by “cybersearching/homology cloning” using IF-3 sequences from Mycoplasma and Euglena gracilis [100]. The recombinant protein can carry out some steps related to initiation in vitro, but much needs to be learned about its precise role in mitochondria. Finally, no ortholog for the bacterial IF-1 has been found in a mammalian mitochondrial system. In the absence of large-scale genetic screening for mammalian mutant cells defective in mitochondrial protein synthesis, it will be a formidable challenge to further understand mitochondrial protein synthesis, particularly its initiation. There is one mutant Chinese hamster cell line isolated serendipitously that could offer further clues. These cells are respiration deﬁcient and there is no mitochondrial protein synthesis, although mtDNA levels are normal and transcription and processing also appear to be unaffected [79–81]. Preliminary studies suggest that the defect is due to a nuclear mutation that affects the initiation step (I. E. Schefﬂer laboratory, unpublished). A complementation analysis with a cDNA library might offer the prospect of identifying a novel factor in this process.

3. IMPORT OF PROTEINS INTO MITOCHONDRIA From proteomic studies the number of different proteins in mitochondria has been estimated to exceed 1000, with an unknown number of proteins that may be present at very low abundance or in only a select number of tissues yet to be identiﬁed. As discussed above, 13 of these are made in the matrix. Historically, this was a relatively early insight into the biogenesis of mitochondria, and thus studies were begun to elucidate (1) how proteins made in the cytosol were directed to mitochondria, and (2) how these proteins were distributed to the various compartments [matrix, intermembrane space (IMS)] and to the various membranes [i.e., outer membrane (OM), inner membrane (IM)]. Although details of the mechanisms continue to challenge investigators, progress has been substantial. A major problem is to understand how the multisubunit complexes of the oxidative phosphorylation system are assembled and equipped with iron–sulfur centers (Fe–S), heme groups, and other metal ions. Much of our detailed insight and knowledge of the many components comes from a combination of genetic and biochemical studies in yeast. However, many of the relevant genes and proteins have been found in mammalian organisms, and the fundamental processes discussed below are likely to be very similar in all mitochondria.

IMPORT OF PROTEINS INTO MITOCHONDRIA

53

3.1. Mitochondrial Targeting Signal Pioneering studies by the Schatz laboratory began with the discovery that mitochondrial proteins could be synthesized by an in vitro translation system (reticulocyte lysate or wheat germ) and imported in vitro into isolated mitochondria. In other words, import was posttranslational in vitro. The in vivo situation is less clear [101]. Ribosomes have been found to be associated with mitochondria, and from such mitochondria mRNAs have been characterized that encode mitochondrial proteins (e.g., [102]). It seems likely that if protein synthesis is not too fast, import into mitochondria can be initiated before the complete protein is released. Another signiﬁcant ﬁnding was that the mitochondria had to have a membrane potential () to be competent for import. The in vitro system was a powerful tool to establish that there was a strong discrimination between proteins destined for mitochondria (matrix proteins were initially tested) and proteins that must remain in the cytosol [103,104]. At the same time it was shown that proteins synthesized in the absence of mitochondria were slightly longer than the mature proteins found after import. An N-terminal leader sequence was quickly identiﬁed as the signal for mitochondrial import [105–107]. As the number of known mitochondrial targeting sequences increased, it became apparent that there was no simple consensus sequence, as these sequences varied greatly in length as well as in composition. A major distinguishing feature appears to be the capacity to form an amphipathic helix with positive charges on one side when viewed down the long axis. Frequently, a sequence of hydrophobic residues is followed by a sequence containing positively charged side chains. A variety of computer programs can be found on the Internet that will identify mitochondrial proteins from genomic sequence information and rules about the nature of mitochondrial targeting sequences (www.123genomics.com/ﬁles/analysis.html). However, such programs fail to ﬁnd a substantial fraction of mitochondrial proteins. One reason is that many proteins, especially integral membrane proteins, have internal targeting sequences that are difﬁcult to recognize. The targeting sequences at the N-terminal are removed by a metalloprotease (mitochondrial processing peptidase) in the mitochondrial matrix which consists of two subunits, α-MPP and β-MPP [108,109]. Here also a large number of cleavage sites have been analyzed with the goal of identifying precursors (deﬁned by genomic studies). An Arg2 rule has emerged [107], but it is not applicable to all cases. Some proteins are cleaved in two distinct steps, requiring an additional protease, MIP. The genes for these processing peptidases were initially cloned from yeast, but it has become apparent that these enzymes are structurally and functionally conserved across species. A recent review that includes IMP, a peptidase in the inner membrane, has been published by Gakh and colleagues [110]. The nascent peptide in the cytosol has to expose the targeting sequence, and it has to be in the unfolded state for import. A subset of stress proteins, specifically hsp70 proteins, were shown to maintain the unfolded state [111]. As the protein is transferred into the mitochondrion, the hsp70 chaperone is released by a mechanism requiring ATP hydrolysis. Whereas hsp70 is also involved in the translocation of secretory proteins across the ER membrane, another

54

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

ATP-dependent chaperone, referred to as MSF (mitochondrial import stimulating factor), is dedicated exclusively to import into mitochondria. These and perhaps other chaperones may act in distinct import pathways with respect to the receptor(s) on the outer surface and the ﬁnal destination of the protein [112,113]. In summary, mitochondrial protein import can be considered a type of active transport. First, a membrane potential across the inner membrane is required, although the precise mechanism for the utilization of this potential energy is still not known. There has been speculation that it is driving the positively charged targeting sequence across the inner membrane. Second, ATP hydrolysis is required on the outside to release the unfolded peptide into the mitochondria, and as described below, ATP hydrolysis by chaperones in the matrix promotes the transfer, releases the chaperone, and delivers the peptide to other scaffolding proteins for refolding the imported peptide to its tertiary structure. In the following discussion a distinction must be made between various mitochondrial proteins and their ﬁnal destination in the organelle. There are several types of protein in the outer membrane. The simplest topology is found for proteins with a single membrane anchor and a large C-terminal domain extending into the cytosol. A second type is exempliﬁed by the β-barrel structure of VDAC (alias porin) that forms a relatively large channel for the passage of molecules up to 1500 Da in size. There are a large number and variety of integral membrane proteins in the inner membrane, many of which form heteromeric complexes with other integral membrane proteins and peripheral membrane proteins. Many but not all have multiple membrane-spanning helical segments. The largest variety of proteins is probably found in the matrix, and these include many of the well-known metabolic enzymes as well as proteins required for the expression of information from mtDNA. Finally, the number of different proteins in the intermembrane space has grown considerably in the past decade; cytochrome c is a well-known representative. Each of these proteins reaches its ﬁnal location by an import pathway that may share some components with others and also has unique features. 3.2. Translocation Through and into the Outer Membrane Genetic experiments in yeast led the way toward the genetic and biochemical characterization of the import machinery required. The early history can be found in several authoritative reviews [114,115], with expert descriptions of the methodology presented in volume 260 of Methods of Enzymology (1995). Since then the major active laboratories have contributed a steady stream of reviews [7, 116–121]. For the import of matrix proteins, two membranes have to be traversed. Leaders in the ﬁeld agreed on a uniform nomenclature [122] with reference to the TOM complex and the TIM complex(es) [123]. The TOM complex is made up of Tom proteins that are integral or peripheral proteins in the outer membrane, and the TIM complex(es) consist of Tim subunits in the inner membrane. Some small, soluble proteins in the IMS have also been named small Tim proteins. The various subunits of these complexes are distinguished by numbers representing their approximate molecular mass (e.g., Tom20, Tom22, Tom37,

IMPORT OF PROTEINS INTO MITOCHONDRIA

55

Tom70). By homology cloning or cybersearching, orthologous proteins have been identiﬁed in many other organisms, including mammals, and it is safe to assume that the import machinery has largely been conserved in evolution. The TOM complex consists of seven different subunits. Tom20, Tom22, and Tom70 are receptor subunits, Tom40 is the channel-forming subunit [124] and Tom5, Tom6, and Tom7 are small subunits that may be required for stability, but otherwise perform an as yet unknown function [125]. As described above, the unfolded protein to be imported is presented in a complex with chaperones, and the receptor subunits must recognize the targeting sequence and initiate the translocation. Tom40 has a β-barrel structure with a channel that can accommodate an unfolded peptide but not a folded native protein. Thus, the peptide is threaded through the channel. An OM protein with a single transmembrane region must then be released into the lipid bilayer. β-Barrel proteins such as VDAC (and Tom40 itself) are translocated into the IMS, and after a transient association with some small Tim proteins, they are delivered to the SAM (sorting and assembly machinery) complex in the OM. This complex consists of three subunits (Sam37, alias Mas37; Sam50; and Sam35, alias Tom38). Sam 50 is homologous to the bacterial Omp85 protein that serves in β-barrel protein export from the periplasm to the bacterial outer membrane. 3.3. Translocation Through the Inner Membrane Over the years the TIM complex has become distinguishable as two complexes: TIM23 and TIM22. The ﬁrst (Tim23, Tim50, Tim14, Tim17, Tim21, Tim44) is responsible for importing matrix proteins. Tim50 recognizes the targeting sequence emerging into the IMS from the TOM complex and inserts it into the channel formed by Tim23. This step requires a membrane potential. With a very unusual topology, Tim23 actually cross-links the OM and the IM. Its N-terminal domain is inserted into the outer membrane; this is followed by a linker sequence and the C-terminal domain, forming four transmembrane helices in the inner membrane. The C-terminal domain of Tim17 is homologous to that of Tim23. It is a required subunit that cannot take on the function of Tim23. Soon after the targeting sequence appears on the matrix side of TIM23, it is cleaved by the processing peptidase. The following peptide chain then becomes engaged with the import motor [126], also referred to as the PAM complex [119]. This motor is made up of peripheral membrane proteins on the inside face of the IM: Tim44 binds to the membrane proteins Tim23 and Tim17, and with the help of several other Tim proteins (e.g., Tim14) it recruits the mtHSP70 chaperone. The latter binds and releases unfolded proteins, depending on its conformational state, which is regulated by ATP hydrolysis. In one proposed model the power stroke of the mtHsp70 ATPase actively pulls the peptide through the channel into the matrix. A competing model hypothesizes the existence of a Brownian ratchet that biases the back-and-forth oscillations of an unfolded peptide toward a net movement to the inside by virtue of a slightly higher afﬁnity of matrix chaperones. An expert review by Mokranjak and Neupert may be consulted for further details on these models [118].

56

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

The TIM22 complex (Tim22, Tim 54, Tim18, and a subset of small Tims) is employed to insert polytopic integral membrane proteins into the IM. All the mitochondrial carrier proteins fall into this category, and one of the best known is the adenine nucleotide transporter (ANT). Many of these have no N-terminal targeting sequence, and thus their initial recognition by the TOM complex is still a puzzle. In other cases (e.g., the SDHC subunit of complex II) it appears that an N-terminal targeting sequence may initiate recognition by TOM and partial translocation through the TIM23 complex, but further transfer is stopped by a transmembrane “stop-transfer” sequence. The remaining C-terminal domain is then found transiently in the IMS, where a process involving the TIM22 complex leads to the insertion of a hairpin loop (and two additional transmembrane segments) into the inner membrane. These proteins are very hydrophobic, and the question arises how they traverse the aqueous intermembrane space. A strong clue was provided by a convergence of genetic studies of yeast [127] and the identiﬁcation of a defective human X-linked gene (DDP1 ) causing the deafness–dystonia–optic atrophy syndrome [128]. When the human gene was identiﬁed, its function was unknown, but a BLAST search quickly led to the recognition that the protein was homologous to Tim8, a protein found to be defective in a screen of yeast mutants affected in mitochondrial protein import. Subsequently, a family of small Tim proteins (Tim8, Tim9, Tim10, . . . , Tim13) was shown to be localized in the IMS, where they form speciﬁc aggregates (e.g., Tim9/Tim10). They act as chaperones for unfolded proteins in transit, in conjunction with two other components recently discovered: Mia40 and Erv1. All of these IMS proteins have a characteristic motif containing several cysteines, the conserved “twin CX3C” motif [129]. Erv1 functions as a sulfhydryl oxidase with Mia40 as its substrate. One proposal was that these cysteines are the ligands for metal ion binding (Zn2+ ) [130], but another favored hypothesis is that they are involved in a series of transient disulﬁde bond formations with the unfolded peptides as they cross this mitochondrial compartment. An unsolved question is how the binding and release of these chaperones is controlled. 3.4. Assembly of the Complexes in the Inner Membrane: Supercomplexes The assembly of mitochondrial ribosomes and homo- and heterooligomeric active enzyme complexes in the mitochondrial matrix does not confront us in principle with new challenges unique to mitochondria. Subunits have to be folded, aggregated, and equipped with the appropriate cofactors. In contrast, the assembly of the four complexes of the electron transport chain and ATP synthase is a more special problem and a subject of much current investigation. Some common and some unique aspects can be pointed out for each complex. First, all complexes, with the exception of complex II, have integral membrane proteins made in the mitochondrial matrix as well as subunits imported, as described above. How are (if they are) these two processes coordinated? It is likely that there is an assembly pathway, and there is evidence for assembly intermediates, but in some cases one must also consider that intermediates identiﬁed on Blue-Native gels are not

IMPORT OF PROTEINS INTO MITOCHONDRIA

57

assembly intermediates but the result of the dissociation of complexes in which mutations cause weakened interactions between subunits. Complex I is comprised of 45 subunits. These are almost equally split between an integral membrane subcomplex with about 20 subunits (seven encoded by mtDNA) and 60 or more transmembrane helices, and a peripheral membrane subcomplex. It is tempting to suggest that these two subcomplexes are assembled individually and associated in one of the ﬁnal steps, but the evidence is not yet convincing. Complex I contains several small subunits with a single transmembrane domain. Two of these are the MWFE protein (70 amino acids) and the ESSS protein (120 amino acids), and either of these proteins constitutes about 1% of the total mass of the complex (about 950 kDa). They belong to the “accessory” subunits not found in the prokaryotic complex I, which has only 14 subunits. Nevertheless, in the absence of MWFE or ESSS, complex I fails to assemble [131–134]. A model system has been developed with Chinese hamster ﬁbroblasts in which MWFE is expressed exclusively from a transgene with an inducible promoter. When inducer is added (doxycyclin), complex I assembly is initiated. Surprisingly, it takes 48 to 72 hours to restore complex I activity and full respiration in these cells [135], perhaps because the process is slow. Alternatively, one can speculate that the assembled complex has to be moved from the site of assembly into the cristae to become part of a supercomplex named the respirasome [136,137]. Other assembly intermediates and pathways for assembly have been suggested from investigations of complex I in human patients with a mitochondrial disease due to a complex I deﬁciency [138–142]. In such patients one typically ﬁnds reduced levels of complex I relative to the other complexes, rather than complex I with reduced speciﬁc activity. There are indications that complex I assembly requires “assembly factors,” that is, proteins that are required for assembly but are not found in the mature complex. One of these was identiﬁed initially in Neurospora crassa and shown subsequently to be required in humans as well [139]. Experiments from our laboratory suggest the existence of a third X-linked gene (MWFE and ESSS are encoded by the X chromosome [133,143]) required for the formation of a mature complex I [133]. A characteristic of the complexes of the electron transport chain is that they must conduct electrons over a signiﬁcant distance. This is accomplished by the insertion of nonheme iron–sulfur centers (Fe–S) into the protein complexes near the high-potential end of the chain, and heme groups and copper ions in the lower half. Complex I has eight Fe–S clusters in the peripheral membrane subcomplex, and they conduct electrons from the reduced ﬂavin to ubiqinone [144,145]. Thus, after the import of about 25 subunits from the cytosol, several of these subunits must acquire Fe–S clusters as they fold into their tertiary structure and associate with other subunits. Fe–S clusters are also found in complexes II and III. It should be noted that in mammalian mitochondria there are three types: 2Fe–2S, 3Fe–4S, and 4Fe–4S. For some time it was thought that the Fe–S clusters form spontaneously as the corresponding proteins assume their tertiary structure, but a burst of activity in the past decade has revealed that the process is considerably more complicated [146–150]. A total of at least 10 proteins in

58

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

the mitochondrial matrix are required, and it is now clear that all Fe–S clusters are made initially by reactions in the mitochondrial matrix, but subsequently some can be exported into the cytosol for incorporation into cytosolic proteins and even a nuclear protein. It has been argued that synthesis of Fe–S clusters is the only activity for which mitochondria are absolutely required in some organisms, and in some of these the mitochondria have “degenerated” into DNA-less, respiration-incompetent organelles such as hydrogenosomes [151]. Heme groups are present in complexes II, III, and IV and in the mobile carrier cytochrome c. The biosynthesis of heme has been elucidated some time ago and can be found in general biochemistry textbooks, but it is still necessary to discriminate between a, b and c-type hemes and their speciﬁc incorporation into each of these complexes. Finally, complex IV contains hemes a and a3 and two copper centers for the transport of electrons from reduced cytochrome c to oxygen. It is made up of 13 subunits in mammalian mitochondria and 12 in yeast. An exhaustive search for respiration-deﬁcient yeast mutants has uncovered, among others, a large collection of complex IV–deﬁcient mutants [152]. Upon further analysis the absence or reduced cytochrome oxidase (COX) activity was shown to result from a faulty or inefﬁcient assembly of the complex rather than from mutations in the structural genes of the COX subunits. Human patients suffering from Leigh syndrome due to a partial COX deﬁciency have also been analyzed and found to have perfectly normal structural genes for the 13 subunits. It is now recognized that at least 25 additional gene products may be required for the assembly of a fully functional complex IV [153]. Some of these are required (so far shown in yeast only) for translation of the Cox mRNAs in the matrix. Others are speciﬁc for heme a or a3 synthesis and insertion. A third group (the Sco1 and Sco2 proteins) are factors delivering the copper ions to the appropriate centers. The role of several factors is not well understood. In the case of the human patients there is a complete absence of a protein encoded by the SURF-1 gene [154–156]. The SURF-1 protein is not absolutely required for complex IV assembly, but in its absence the assembly (or stability?) is inefﬁcient, and signiﬁcantly reduced levels of activity are found in the patients. Evidently, homozygous (−/−) patients are alive, but not well. The corresponding gene in yeast is Shy1 , and thus gene knockouts can be studied easily in this organism. It is impossible to do justice to a large number of elegant studies in this model system. The interested reader is referred to the original papers and reviews [153, 157–159].

4. FISSION OF MITOCHONDRIA AND SEGREGATION DURING CELL DIVISION When the methodology was developed to stain mitochondria in live cells, several important discoveries were made. The staining was initially accomplished with cationic dyes that were accumulated in mitochondria driven by the membrane potential and because they could traverse the lipid bilayer [160]. In the

FISSION OF MITOCHONDRIA AND SEGREGATION DURING CELL DIVISION

59

following decades a large group of such dyes has been developed commercially (Molecular Probes, Eugene, Oregon), with MitoTracker perhaps the most familiar representative. More recently, reporter proteins such as the green ﬂuorescent protein (GFP) have been targeted to the mitochondrial matrix by the addition of a targeting sequence. Observations in the ﬂuorescent microscope revealed some expected but also some unexpected behavior of mitochondria. First, they move about in the cytosol (or along axons), and it is apparent that they can become engaged with molecular motors such as myosin, dynein, or kinesin to move along cytoskeletal structures (microﬁlaments or microtubules). Such interactions may also be responsible for pulling at a mitochondrion to cause changes in shape (for further discussion, see Schefﬂer [26]). It was anticipated that mitochondria would divide or ﬁssion as their volume increased in cells progressing through the cell cycle, to be distributed equally between daughter cells during cell division. The surprising observation was that ﬁssions take place continuously, but these are balanced by fusions between two or more mitochondria. Mitochondrial fusions had been observed ﬁrst during spermatogenesis in fruit ﬂies, but it was thought to be a relatively exceptional event associated with sperm differentiation. The mutation “fuzzy onion” in Drosophila causes male sterility, and the cause is an arrest in sperm development because of the failure of mitochondria to fuse. When the relevant gene (Fzo) was cloned in Drosophila, a cybersearch revealed that homologous genes exist in other organisms, including yeast (Fzo) and mammals (mitofusin, MFN-1, and MFN-2). Thus, an fzo knockout in yeast cells also prevents fusions, and the continuing ﬁssions lead to a pronounced fragmentation of mitochondria. Experiments such as these stimulated a great number of studies of yeast mutants with abnormal mitochondrial morphologies that are easily revealed by speciﬁc staining and microscopic examination. Mutants were found in which the absence of fusion led to fragmented mitochondria, and in others the absence of ﬁssion resulted in very large, abnormal mitochondria and in some cases a failure to segregate mitochondria into daughter cells. Many, but not all proteins and genes found to be associated with these processes in yeast have been found in mammalian cells. Either they cannot be recognized or there are mechanistic differences yet to be elucidated. The voluminous literature can be accessed through periodic reviews ([161–167]. Okamoto and Shaw have written a particularly noteworthy, up-to-date, comprehensive, and authoritative review [168]. A detailed discussion of the mechanisms and the proteins involved is beyond the scope of this review. Brieﬂy, fusion has recently been achieved with isolated mitochondria in vitro [169,170]. The process requires the Fzo1 protein for physical association of two identical organelles. It also requires GTP and a proton gradient across the inner membrane to achieve fusion between the outer membranes, and higher levels of GTP as well as a membrane potential to complete the fusion of the inner membranes. A GTPase (Mgm1p, OPA1) localized in the IMS is implicated. This protein has also been found to be required for the maintenance of cristae morphology and ATPsynthase assembly [171,172]. The ﬁssion mechanism requires at least three proteins localized in the outer membrane. They recruit a GTPase, Dnm1p, related to the dynamin-related proteins (DRP or DLP1

60

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

in humans) that have been found previously in association with the mechanism of endocytosis. A ringlike structure of dynamin aggregates is thought to constrict a membranous vesicle and ultimately pinch it into halves [173]. Clearly, the mechanism requires the participation of both the outer and inner membranes, as well as a signal for when and where to establish the ﬁssion apparatus.

5. CONTROL OF MITOCHONDRIAL BIOGENESIS In Section 2 we touched on the issue of the establishment, maintenance, and changes in the copy number of mtDNA, and the role of mtTFA has been emphasized. However, mtDNA copy numbers may not control either the number of mitochondria or the various biological functions of mitochondria. The only activity affected directly is the capacity for oxidative phosphorylation, and even this depends on the coordinated expression of 13 mitochondrial genes and dozens of genes encoded by nDNA. It is not completely clear how this coordination is achieved and which genes play a dominant role. One must distinguish between the biogenesis in a given cell or tissue and a response to physiological factors such as oxidative stress, hormones, and muscle activity, and the biogenesis of specialized mitochondria with different proteomes in a large variety of differentiated cells. A detailed discussion of this topic is beyond the scope of this review, but we will summarize brieﬂy and cite recent reviews. With a focus on the control of OXPHOS activity, Scarpulla and colleagues have identiﬁed two transcription factors with special relevance to respiration, and they have been named nuclear respiratory factors (NRF-1 and NRF-2) [174–177]. Unfortunately, NRF also stands for nuclear regulatory factors, including many that are unrelated to mitochondrial functions. Although the initial idea may have been that these factors control the expression of genes encoding subunits for the complexes of the electron transport chain, it is now established that they are involved in the expression of many other genes, particularly those required for replication and transcription of mtDNA (e.g., mtTFA). A second major factor that has emerged from a broader view of fatty acid metabolism, weight control, thermogenesis, and related metabolic activities is the peroxisome proliferators-activated receptor gamma coactivator-1α (PGC-1α). [178–184]. In fact, it stimulates the expression of NRF-1 and NRF-2 and is thus higher up in a hierarchy of factors. The obvious question is: What controls the expression of these factors? [36]. The availability of nutrients, the redox state of the cell (NAD+ /NADH, GSH/GSSG ratios), calcium concentrations, reactive oxygen species under normal conditions as well as under oxidative stress, cAMP, various inhibitors and pharmacological agents, and more factors can be shown or hypothesized to constitute signals for regulatory cascades. The focus is very much on the production, control, and targets of reactive oxygen species [36]. A relationship has also been established between endothelial nitric oxide synthase (eNOS), NO levels, cGMP signaling, and up-regulation of PGC-1α [185–187]. In such

REFERENCES

61

discussions one should clearly distinguish pathological situations (hypoxia, reperfusion) from normal physiological control mechanisms operating in various tissues (e.g., the response of muscle to exercise). REFERENCES 1. Lane N. Power, Sex, Suicide: Mitochondria and the Meaning of Life. Oxford, UK: Oxford University Press; 2005. 2. Rizzuto R., Pozzan T. Microdomains of intracellular Ca2+ : molecular determinants and functional consequences. Physiol Rev. 2006;86:369–408. 3. Embley TM, Martin W. Eukaryotic evolution, changes and challenges. Nature. 2006;440:623–630. 4. Bullerwell CE, Gray MW. Evolution of the mitochondrial genome: protist connections to animals, fungi and plants. Curr Opin Microbiol. 2004;7;528–534. 5. Dyall SD, Brown MT, Johnson PJ. Ancient invasions: from endosymbionts to organelles. Science. 2004;304:253–257. 6. Hirt RP, Horner DS (eds.). Organelles, Genomes, Eukaryotic Phylogeny: An Evolutionary Synthesis in the Age of Genomics. Boca Raton, FL: CRC Press; 2004. 7. Dolezal P, Likic V, Tachezy J, and Lithgow T. Evolution of the molecular machines for protein import into mitochondria. Science. 2006;313:314–318. 8. Gabaldon T., Huynen MA. Shaping the mitochondrial proteome. Biochim Biophys Acta. 2004;1659:212–220. 9. Catalano D, Licciulli F, Turi A, Grillo G, Saccone C, D’Elia D. MitoRes: a resource of nuclear-encoded mitochondrial genes and their products in Metazoa. BMC Bioinformatics. 2006;7:36. 10. Forner F, Foster LJ, Campanaro S, Valle G, Mann M. Quantitative proteomic comparison of rat mitochondria from muscle, heart and liver. Mol Cell Proteom. 2006;5:608–619. 11. Johnson DT, Harris RA, Blair PV, Balaban RS. Functional consequences of mitochondrial proteome heterogeneity. Am J Physiol Cell Physiol. 2006;292:C698–707. 12. Prokisch H, Andreoli C, Ahting U, et al. MitoP2: the mitochondrial proteome database—now including mouse data. Nucleic Acids Res. 2006;34:D705–D711. 13. Vo TD, Palsson B. Building the power house: recent advances in mitochondrial studies through proteomics and systems biology. Am J Physiol Cell Physiol . 2006;292:C164–177. 14. Da Cruz S, Parone PA, Martinou JC. Building the mitochondrial proteome. Expert Rev Proteom. 2005;2:541–551. 15. Schefﬂer IE. Biochemical genetics of respiration-deﬁcient mutants of animal cells. In Carbohydrate Metabolism in Cultured Cells (MJ, Morgan ed). London: Plenum Publishing; 1986:77–109. 16. King MP, Attardi G. Human cells lacking mtDNA: repopulation with exogenous mitochondria by complementation. Science. 1989;246:500–503. 17. Slonimski PP, Ephrussi B. Action de l’acriﬂavine sur les levures: V. Le syst`eme des cytochromes des mutants “petite colonie.” Ann Inst Pasteur . 1949;77:47. 18. Nass S, Nass MMK. Ultramitochondrial ﬁbers with DNA characteristics. J Cell Biol. 1963;19:593–629.

62

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

19. Dawid IB, Wolstenholme DR. Ultracentrifuge and electron microscope studies on the structure of mitochondrial DNA. J Mol Biol. 1967;28:233–245. 20. Hudson B, Clayton DA, Vinograd J. Complex mitochondrial DNA. Cold Spring Harb Symp Quant Biol. 1968;33:435–442. 21. Anderson S, Bankier AT, Barrell BG, et al. Sequence and organization of the human mitochondrial genome. Nature. 1981;290:457–465. 22. Lanave C, Licciulli F, De Robertis M, Marolla A, Attimonelli M. Update of AMmtDB: a database of multi-aligned metazoa mitochondrial DNA sequences. Nucleic Acids Res. 2002;30:174–175. 23. Kogelnik AM, Lott MT, Brown MD, Navathe SB, and Wallace DC. MitoMap: a human mitochondrial genome database—1998 update. Nucleic Acids Res. 1998;26:112–115. 24. Chomyn A, Mariottini P, Cleeter MWJ, et al. Six unidentiﬁed reading frames of human mitochondrial DNA encode components of the respiratory-chain NADH dehydrogenase. Nature. 1985;314:592–597. 25. Chomyn A, Cleeter MW, Ragan CI, Riley M, Doolittle RF, Attardi G. URF6, last unidentiﬁed reading frame of human mtDNA, codes for an NADH dehydrogenase subunit. Science. 1986;234:614–618. 26. Schefﬂer IE. Mitochondria. New York: Wiley; 1999. 27. Bogenhagen DF, Wang Y, Shen EL, Kobayashi R. Protein components of mitochondrial DNA nucleoids in higher eukaryotes. Mol Cell Proteom. 2003;2:1205–1216. 28. Shadel GS. Coupling the mitochondrial transcription machinery to human disease. Trends Genet. 2004;20(Oct):513–519. 29. Chen XJ, Butow RA. The organization and inheritance of the mitochondrial genome. Nat Rev Genet. 2005;6:815–825. 30. Shadel GS. Mitochondrial DNA, aconitase “wraps” it up. Trends Biochem Sci. 2005;30:294–296. 31. Wang Y, Bogenhagen DF. Human mitochondrial DNA nucleoids are linked to protein folding machinery and metabolic enzymes at the mitochondrial inner membrane. J Biol Chem. 2006;281:25791–25802. 32. Megraw TL, Chae C-B. Functional complementarity between the HMG1-like yeast mitochondrial histone HM and the bacterial histone-like protein HU. J Biol Chem. 1993;268:12758–12763. 33. Megraw TL, Kao LR, Chae CB. The mitochondrial histone HM: an evolutionary link between bacterial HU and nuclear HMG1 proteins. Biochimie. 1994;76:909–916. 34. Ekstrand MI, Falkenberg M, Rantanen A, et al. Mitochondrial transcription factor A regulates mtDNA copy number in mammals. Hum Mol Genet. 2004;13:935–944. 35. Maniura-Weber K, Goffart S, Garstka HL, Montoya J, Wiesner RJ. Transient overexpression of mitochondrial transcription factor A (TFAM) is sufﬁcient to stimulate mitochondrial DNA transcription, but not sufﬁcient to increase mtDNA copy number in cultured cells. Nucleic Acids Res. 2004;32:6015–6027. 36. Lee HC, Wei YH. Mitochondrial biogenesis and mitochondrial DNA maintenance of mammalian cells under oxidative stress. Int J Biochem Cell Biol. 2005;37:822–834. 37. Davis AF, Clayton DA. In situ localization of mitochondrial DNA replication in intact mammalian cells. J Cell Biol. 1996;135:883–893.

REFERENCES

63

38. Lecrenier N, Van Der Bruggen P, and Foury F. Mitochondrial DNA polymerases from yeast to man: a new family of polymerases. Gene. 1997;185:147–152. 39. Kaguni LS. DNA polymerase gamma, the mitochondrial replicase 1. Annu Rev Biochem. 2004;73:293–320. 40. Yakubovskaya E, Chen Z, Carrodeguas JA, Kisker C, Bogenhagen DF. Functional human mitochondrial DNA polymerase gamma forms a heterotrimer. J Biol Chem. 2006;281:374–382. 41. Carrodeguas JA, Pinz KG, Bogenhagen DF. DNA binding properties of human Pol gamma B. J Biol Chem. 2002;Au: please complete 42. Ponamarev MV, Longley MJ, Nguyen D, Kunkel TA, Copeland WC. Active site mutation in DNA polymerase gamma associated with progressive external ophthalmoplegia causes error-prone DNA synthesis. J Biol Chem. 2002;277:15225–15228. 43. Lim SE, Ponamarev MV, Longley MJ, Copeland WC. Structural determinants in human DNA polymerase gamma account for mitochondrial toxicity from nucleoside analogs. J Mol Biol. 2003;329:45–57. 44. Longley MJ, Graziewicz MA, Bienstock RJ, Copeland WC. Consequences of mutations in human DNA polymerase [gamma]. Gene. 2005;354:125–131. 45. Kornberg A, Baker T. DNA Replication, 2nd ed. New York: W.H. Freeman; 1992. 46. Shadel GS. Clayton DA. Mitochondrial DNA maintenance in vertebrates. Annu Rev Biochem. 1997;66:409–435. 47. Korhonen JA, Pham XH, Pellegrini M, Falkenberg M. Reconstitution of a minimal mtDNA replisome in vitro. EMBO J. 2004;23:2423–2429. 48. Benkovic SJ, Valentine AM, Salinas F. Replisome-mediated DNA replication. Annu Rev Biochem. 2001;70:181–208. 49. Shutt TE, Gray MW. Twinkle, the mitochondrial replicative DNA helicase, is widespread in the eukaryotic radiation and may also be the mitochondrial DNA primase in most eukaryotes. J Mol Evol. 2006;62:588–599. 50. Spelbrink JN, Li F-Y, Tiranti V, et al. Human mitochondrial DNA deletions associated with mutations in the gene encoding Twinkle, a phage T7 gene 4-like protein localized in mitochondria. Nat Genet. 2001;28:223–231. 51. Clayton DA. Transcription and replication of animal mitochondrial DNAs. Int Rev Cytol. 1992;141:217–232. 52. Holt IJ, Lorimer HE, Jacobs HT. Coupled leading-and-lagging strand synthesis of mammalian mitochondrial DNA. Cell . 2000;100:515–524. 53. Bowmaker M, Yang MY, Yasukawa T, et al. Mammalian mitochondrial DNA replicates bidirectionally from an initiation zone. J Biol Chem. 2003;278:50961–50969. 54. Holt IJ, Jacobs HT. Response: the mitochondrial DNA replication bubble has not burst. Trends Biochem Sci. 2003;28:355–356. 55. Brown TA, Cecconi C, Tkachuk AN, Bustamante C, Clayton DA. Replication of mitochondrial DNA occurs by strand displacement with alternative light-strand origins, not via a strand-coupled mechanism. Genes Dev. 2005;19:2466–2476. 56. Brown TA, Clayton DA. Genesis and wanderings: origins and migrations in asymmetrically replicating mitochondrial DNA. Cell Cycle. 2006;5(9):917–921. 57. Fish J, Raule N, Attardi G. Discovery of a major D-loop replication origin reveals two modes of human mtDNA synthesis. Science. 2004;306:2098–2101.

64

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

58. Chang DD. Clayton DA. A mammalian mitochondrial RNA processing activity contains nucleus-encoded RNA. Science. 1987;235:1178–1184. 59. Chang DD. Clayton DA. Mouse RNase MRP RNA is encoded by a nuclear gene and contains a decamer sequence complementary to a conserved region of mitochondrial RNA substrate. EMBO J. 1987;6:409–417. 60. Clayton DA. Nuclear gadgets in mitochondrial DNA replication and transcription. TIBS 1991;16:107–111. 61. LeDoux SP, Wilson GL. Base excision repair of mitochondrial DNA damage in mammalian cells. Prog Nucleic Acid Res Mol Biol. 2001;68:273–284. 62. Mandavilli BS, Santos JH, Van Houten B. Mitochondrial DNA repair and aging. Mutat Res. 2002;509:127–151. 63. Stuart JA, Hashiguchi K, Wilson DM III, Copeland WC, Souza-Pinto NC, Bohr VA. DNA base excision repair activities and pathway function in mitochondrial and cellular lysates from cells lacking mitochondrial DNA. Nucleic Acids Res. 2004;32:2181–2192. 64. Mason PA, Matheson EC, Hall AG, Lightowlers RN. Mismatch repair activity in mammalian mitochondria. Nucleic Acids Res. 2003;31:1052–1058. 65. Trifunovic A, Wredenberg A, Falkenberg M, et al. Premature ageing in mice expressing defective mitochondrial DNA polymerase. Nature. 2004;429:417–423. 66. Trifunovic A, Hansson A, Wredenberg A, et al. Somatic mtDNA mutations cause aging phenotypes without affecting reactive oxygen species production. Proc Natl Acad Sci U S A. 2005;102:17993–17998. 67. Attardi G. The elucidation of the human mitochondrial genome: a historical perspective. Bioessays.1986;5:34–39. 68. Walberg MW, Clayton DA. In Vitro transcription of human mitochondrial DNA: identiﬁcation of speciﬁc light strand transcripts from the displacement loop region. J Biol Chem. 1983;258:1268–1275. 69. Tiranti V, Savoia A, Forti F, et al. Identiﬁcation of the gene encoding the human mitochondrial RNA polymerase (hmtRPOL) by cyberscreening of the expressed sequence tags database. Hum Mol Genet. 1997;6:615–625. 70. Larsson NG, Barsh GS, Clayton DA. Structure and chromosomal localization of the mouse mitochondrial transcription factor A gene (Tfam). Mamm Genome. 1997;8:139–140. 71. Reyes A, Mezzina M, Gadaleta G. Human mitochondrial transcription factor A (mtTFA): gene structure and characterization of related pseudogenes. Gene. 2002;291:223–232. 72. McCulloch V, Seidel-Rogol BL, Shadel GS. A human mitochondrial transcription factor is related to RNA adenine methyltransferases and binds S -adenosylmethionine. Mol Cell Biol. 2002;22:1116–1125. 73. Falkenberg M, Gaspari M, Rantanen A, Trifunovic A, Larsson NG, Gustafsson CM. Mitochondrial transcription factors B1 and B2 activate transcription of human mtDNA. Nat Genet. 2002;31:289–294. 74. Gaspari M, Larsson NG, Gustafsson CM. The transcription machinery in mammalian mitochondria. Biochim Biophys Acta. 2004;1659:148–152. 75. Carter RH, Demidenko AA, Hattingh-Willis S, Rothman-Denes LB. Phage N4 RNA polymerase II recruitment to DNA by a single-stranded DNA-binding protein. Genes Dev. 2003;17:2334–2345.

REFERENCES

65

76. Christianson TW, Clayton DA. A tridecamer DNA sequence supports human mitochondrial RNA 3 - end formation in vitro. Mol Cell Biol. 1988;8:4502–4509. 77. Fernandez-Silva P, Martinez-Azorin F, Micol V, Attardi G. The human mitochondrial transcription termination factor (mTERF) is a multizipper protein but binds to DNA as a monomer, with evidence pointing to intramolecular leucine zipper interactions. EMBO J. 1997;16(5):1066–1079. 78. Hess WR, Borner T. Organellar RNA polymerases of higher plants. Int Rev Cytol. 1999;190:1–59. 79. Au HC, Schefﬂer IE. A respiration-deﬁcient Chinese hamster cell line with a defect in mitochondrial protein synthesis: rapid turnover of some mitochondrial transcripts. Somat Cell Mol Genet. 1997;23:27–35. 80. Burnett KG, Schefﬂer IE. Integrity of mitochondria in a mammalian cell mutant defective in mitochondrial protein synthesis. J Cell Biol. 1981;90:108–115. 81. Ditta GS, Soderberg K, Schefﬂer IE. Chinese hamster cell mutant with defective mitochondrial protein synthesis. Nature. 1977;268:64–67. 82. Nelson I, Hanna MG, Wood NW, Harding AE. Depletion of mitochondrial DNA by ddC in untransformed human cell lines. Somat Cell Mol Genet. 1997;23;287–290. 83. Wolstenholme DR. Animal mitochondrial DNA: structure and evolution. In Mitochondrial Genomes (DR Wolstenholme, KW Jeong, eds). San Diego, CA: Academic Press; 1992:173–216. 84. Bunn CLD, Wallace DC, Eisenstadt JM. Cytoplasmic inheritance of chloramphenicol resistance in mouse tissue culture cells. Proc Natl Acad Sci U S A. 1974;71:1681–1685. 85. Wallace DC, Eisenstadt JM. Expression of cytoplasmically inherited genes for chloramphenicol resistance in interspeciﬁc somatic cell hybrids and cybrids. Somat Cell Genet. 1979;5:373–396. 86. Breen GAM, Schefﬂer IE. Cytoplasmic inheritance of oligomycin resistance in Chinese hamster ovary cells. J Cell Biol. 1980;86:723–729. 87. Harris M. Cytoplasmic transfer of resistance to antimycin in Chinese hamster cell lines. Proc Natl Acad Sci U S A. 1978;75:5604–5608. 88. Howell N, Sager R. Cytoplasmic genetics of mammalian cells: conditional sensitivity to mitochondrial inhibitors and isolation of new mutant phenotypes. Somat Cell Genet. 1979;5:833–845. 89. Guan MX, Fischel-Ghodsian N, Attardi G. Biochemical evidence for nuclear gene involvement phenotype of non-syndromic deafness associated with mitochondrial 12S rRNA mutation. Hum Mol Genet. 1996;5:963–971. 90. Guan MX. Molecular pathogenetic mechanism of maternally inherited deafness. Ann N Y Acad Sci. 2004;1011:259–271. 91. Guan MX, Yan Q, Li X, et al. Mutation in TRMU related to transfer RNA modiﬁcation modulates the phenotypic expression of the deafness-associated mitochondrial 12S ribosomal RNA mutations. Am J Hum Genet. 2006;79:291–302. 92. Sanchirico ME, Fox TD, Mason TL. Accumulation of mitochondrially synthesized Saccharomyces cerevisiae Cox2p and Cox3p depends on targeting information in untranslated portions of their mRNAs. EMBO J. 1998;17:5796–5804. 93. Dunstan HM, Green-Willms NS, Fox TD. In vivo analysis of Saccharomyces cerevisiae COX2 mRNA 5 - untranslated leader functions in mitochondrial translation initiation and translational activation. Genetics. 1997;147:87–100.

66

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

94. Naithani S, Saracco SA, Butler CA, Fox TD. Interactions among COX1, COX2 , and COX3 mRNA-speciﬁc translational activator proteins on the inner surface of the mitochondrial inner membrane of Saccharomyces cerevisiae. Mol Biol Cell. 2003;14:324–333. 95. Grivell LA. Nucleo-mitochondrial interactions in mitochondrial gene expression. Crit Rev Biochem Mol Biol. 1995;30:121–164. 96. Fox TD. Genetics of mitochondrial translation. In Translational Control (JWB Hershey, MB Mathews, N Sonenberg eds). Cold Spring Harbor, NY: Cold Spring Harbor Press; 1996: 733–758. 97. Richter JD, Sonenberg N. Regulation of cap-dependent translation by eIF4E inhibitory proteins. Nature. 2005;433:477–480. 98. Sonenberg N, Dever TE. Eukaryotic translation initiation factors and regulators. Curr Opin Struct Biol , 2003;13:56–63. 99. Schwartzbach CJ, Farwell M, Liao H-X, Spremulli LL. Bovine mitochondrial initiation and elongation factors. Methods Enzymol. 1996;264:248–261. 100. Spremulli LL, Coursey A, Navratil T, Hunter SE. Initiation and elongation factors in mammalian mitochondrial protein biosynthesis. Prog Nucleic Acid Res Mol Biol. 2004;77:211–261. 101. Schatz G. The protein import system of mitochondria. J Biol Chem. 1996;271: 31763–31766. 102. Suissa M, Suda K, Schatz G. Isolation of the nuclear yeast genes for citrate synthetase and ﬁfteen other mitochondrial proteins by a new screening method. EMBO J. 1984;3:1773–1781. 103. Horwich AL, Kalousek F, Mellman I, Rosenberg LE. A leader peptide is sufﬁcient to direct mitochondrial import of a chimeric protein. EMBO J. 1985;4:1129–1135. 104. Eilers M, Schatz G. Binding of a speciﬁc ligand inhibits import of a puriﬁed precursor protein into mitochondria. Nature. 1986;322:228–232. 105. Allison DS, Schatz G. Artiﬁcial mitochondrial presequences. Proc Natl Acad Sci U S A. 1986;83:9011–9015. 106. Roise D, Schatz G. Mitochondrial presequences. J Biol Chem. 1988;263:4509–4511. 107. Hendrick JP, Hodges PE, Rosenberg LE. Survey of amino-terminal proteolytic cleavage sites in mitochondrial precursor proteins: leader peptides cleaved by two matrix proteases share a three-amino acid motif. Proc Natl Acad Sci U S A. 1989;86:4056–4060. 108. Paces V, Rosenberg LE, Fenton WA, Kalousek F. The β subunit of the mitochondrial processing peptidase from rat liver: cloning and sequencing of a cDNA and comparison with a proposed family of metallopeptidases. Proc Natl Acad Sci U S A. 1993;90:5355–5358. 109. Kalousek F, Neupert W, Omura T, Schatz G, Schmitz UK. Uniform nomenclature for the mitochondrial peptidases cleaving precursors of mitochondrial proteins. Trends Biochem Sci. 1993;18:249. 110. Gakh O, Cavadini P, Isaya G. Mitochondrial processing peptidases. Biochim Biophys Acta Mol Cell Res. 2002;1592;63–77. 111. Deshaies RJ, Koch BD, Werner-Washburne M, Craig EA, Schekman R. A subfamily of stress proteins facilitates translocation of secretory and mitochondrial precursor polypeptides. Nature. 1988;332:800–805.

REFERENCES

67

112. Mihara K., Omura T. Cytoplasmic chaperones in precursor targeting to mitochondria: the role of MSF and hsp70. Trends Cell Biol., 1996;6:104–108. 113. Beddoe T, Lithgow T. Delivery of nascent polypeptides to the mitochondrial surface. Biochim Biophys Acta Mol Cell Res. 2002;1592:35–39. 114. Yaffe MP, Schatz G. Two nuclear mutations that block mitochondrial protein import. Proc Natl Acad Sci U S A. 1984;81:4819–4823. 115. Yaffe MP. Analysis of mitochondrial functions and assembly. Methods Enzymol. 1991;194:627–643. 116. Lister R, Hulett JM, Lithgow T, Whelan J. Protein import into mitochondria: origins and functions today (review). Mol Membr Biol. 2005;22:87–100. 117. Wiedemann N, Frazier AE, Pfanner N. The protein import machinery of mitochondria. J Biol Chem. 2004;279:14473–14476. 118. Mokranjac D, Neupert W. Protein import into mitochondria. Biochem Soc Trans. 2005;33:1019–1023. 119. Stojanovski D, Rissler M, Pfanner N, Meisinger C. Mitochondrial morphology and protein import: a tight connection? Biochim Biophys Acta. 2006;1763:414–421. 120. Lister R, Whelan J. Mitochondrial protein import: convergent solutions for receptor structure. Curr Biol., 2006;16:R197–R199. 121. Perry AJ, Hulett JM, Likic VA, Lithgow T, Gooley PR. Convergent evolution of receptors for protein import into mitochondria. Curr Biol. 2006;16:221–229. 122. Pfanner N, Douglas MG, Endo T, et al. Uniform nomenclature for the protein transport machinery of the mitochondrial membranes. Trends Biochem Sci. 1996;21:51–52. 123. Pfanner N, Meijer M. Mitochondrial biogenesis: the Tom and Tim machine. Curr Biol. 1997;7:R100–R103. 124. Rapaport D. How does the TOM complex mediate insertion of precursor proteins into the mitochondrial outer membrane? J Cell Biol. 2005;171:419–423. 125. Sherman EL, Go NE, Nargang FE. Functions of the small proteins in the TOM complex of Neurospora crasssa. Mol Biol Cell . 2005;16:4172–4182. 126. Milakovic T, Johnson GV. Mitochondrial respiration and ATP production are signiﬁcantly impaired in striatal cells expressing mutant huntingtin. J Biol Chem. 2005;280:30773–30782. 127. Koehler CM, Merchant S, Oppliger W, et al. Tim9p, an essential partner subunit of Tim10p for the import of mitochondrial carrier proteins. EMBO J. 1998;17:6477–6486. 128. Tranebjaerg L, Hamel BC, Gabreels FJ, Renier WO, Van Ghelue M. A de novo missense mutation in a critical domain of the X-linked DDP gene causes the typical deafness–dystonia–optic atrophy syndrome. Eur J Hum Genet. 2000;8:464–467. 129. Koehler C. The samll Tim proteins and the twin Cx3C motif . Trends Biochem Sci. 2004;29:1–4. 130. Lutz T, Neupert W, Herrmann JM. Import of small Tim proteins into the mitochondrial intermembrane space. EMBO J . 2003;22:4400–4408. 131. Au HC, Seo BB, Matsuno-Yagi A, Yagi T, Schefﬂer IE. The NDUFA1 gene product (MWFE protein) is essential for activity of complex I in mammalian mitochondria. Proc Natl Acad Sci U S A. 1999;96:4354–4359.

68

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

132. Potluri P, Yadava N, Schefﬂer IE. The role of the ESSS protein in the assembly of a functional and stable mammalian mitochondrial complex I (NADH–ubiquinone oxidoreductase). Eur J Biochem. 2004;271:3265–3273. 133. Schefﬂer IE, Yadava N, Potluri P. Molecular genetics of complex I–deﬁcient Chinese hamster cell lines. Biochim Biophys Acta. 2004;1659:160–171. 134. Yadava N, Potluri P, Smith E, Bisevac A, Schefﬂer IE. Species-speciﬁc and mutant MWFE proteins: their effect on the assembly of a functional mammalian mitochondrial complex I. J Biol Chem. 2002;277:21221–21230. 135. Yadava N, Houchens T, Potluri P, Schefﬂer IE. Development and characterization of a conditional mitochondrial complex I assembly system. J Biol Chem. 2004;279:12406–12413. 136. Wittig I, Carrozzo R, Santorelli FM, Schagger H. Supercomplexes and subcomplexes of mitochondrial oxidative phosphorylation. Biochim Biophys Acta. 2006; 137. Schagger H, De Coo R, Bauer MF, Hofmann S, Godinot C, Brandt U. Signiﬁcance of respirasomes for the assembly/stability of human respiratory chain complex I. J Biol Chem. 2004;279:36349–36353. 138. Antonicka H, Ogilvie I, Taivassalo T, et al. Identiﬁcation and characterization of a common set of complex I assembly intermediates in mitochondria from patients with complex I deﬁciency. J Biol Chem. 2003;278:43081–43088. 139. Janssen R, Smeitink J, Smeets R, van den Heuvel J. CIA30 complex I assembly factor: a candidate for human complex I deﬁciency. Hum Genet. 2002;110:264–270. 140. Triepels RH, Van den Heuvel LP, Trijbels JM, Smeitink JA. Respiratory chain complex I deﬁciency. Am J Med Genet. 2001;106:37–45. 141. Vogel R, Nijtmans L, Ugalde C, van den HL, Smeitink J. Complex I assembly: a puzzling problem. Curr Opin Neurol. 2004;17:179–186. 142. Vogel RO, Janssen RJ, Ugalde C, et al. Human mitochondrial complex I assembly is mediated by NDUFAF1. FEBS J. 2005;272:5317–5326. 143. Day C, Schefﬂer IE. Mapping of the genes for some components of complex I of the electron transport chain on the X chromosome of mammals. Somat Cell Genet. 1982;8:691–707. 144. Hinchliffe P, Sazanov LA. Organization of iron–sulfur clusters in respiratory complex I. Science. 2005;309:771–774. 145. Sazanov LA, Hinchliffe P. Structure of the hydrophilic domain of respiratory complex I from Thermus thermophilus. Science. 2006;311:1430–1436. 146. Lill R, Muhlenhoff U. Iron–sulfur–protein biogenesis in eukaryotes. Trends Biochem Sci. 2005;30:133–141. 147. Lill R, Dutkiewicz R, Elsasser HP, et al. Mechanisms of iron–sulfur protein maturation in mitochondria, cytosol and nucleus of eukaryotes. Biochim Biophys Acta. 2006;1763:652–667. 148. Lill R, Muhlenhoff U. Iron–sulfur–protein biogenesis in eukaryotes. TIBS . 2005;30:133–141. 149. Gerber J, Lill R. Biogenesis of iron–sulfur proteins in eukaryotes: components, mechanism and pathology. Mitochondrion. 2002;2:71–86. 150. Johnson DC, Dean DR, Smith AD, Johnson MK. Structure, function, and formation of biological iron–sulfur clusters. Annu Rev Biochem. 2005;74:247–281.

REFERENCES

69

151. van der Giezen M, Tovar J, Clark CG. Mitochondrion-derived organelles in protists and fungi. Int Rev Cytol. 2005;244:175–225. 152. Tzagoloff A., Dieckmann CL. PET genes of Saccharomyces cerevisiae. Microbiol Rev. 1990;54:211–225. 153. Herrmann JM, Funes S. Biogenesis of cytochrome oxidase–sophisticated assembly lines in the mitochondrial inner membrane. Gene. 2005;354:43–52. 154. Zhu Z, Yao J, Johns T, et al. SURF-1, encoding a factor involved in the biogenesis of cytochrome c oxidase, is mutated in Leigh syndrome. Nat Genet . 1998;20:337–343. 155. Tiranti V, Galimberti C, Nijtmans LG, Bovolenta S, Perin MP, Zeviani M. Characterization of SURF-1 expression and SURF-1p function in normal and disease conditions. Hum Mol Genet. 1999;8:2533–2540. 156. Tiranti V, Hoertnagel K, Carrozzo R, et al. Mutations in SURF-1 in Leigh disease associated with cytochrome c oxidase deﬁciency. Am J Hum Genet. 2001;63(1609):1621. 157. Barrientos A. Yeast models of human mitochondrial diseases. IUBMB Life. 2003;55:83–95. 158. Richter OM, Ludwig B. Cytochrome c oxidase: structure, function, and physiology of a redox-driven molecular machine. Rev Physiol Biochem Pharmacol . 2003;147:47–74. 159. Robinson BH. Human cytochrome oxidase deﬁciency. Pediatr Res. 2000;48: 581–585. 160. Chen LB. Mitochondrial membrane potential in living cells. Annu Rev Cell Biol. 1988;4:155–181. 161. Yaffe MP. The machinery of mitochondrial inheritance and behavior. Science. 1999;283:1493–1497. 162. Osteryoung KW, Nunnari J. The division of endosymbiotic organelles. Science. 2003;302:1698–1704. 163. Shaw JM, Nunnari J. Mitochondrial dynamics and division in budding yeast. Trends Cell Biol. 2002;12:178. 164. Rube DA, Van der Bliek AM. Mitochondrial morphology is dynamic and varied. Mol Cell Biochem. 2004;256–257:331–339. 165. Meeusen SL, Nunnari J. How mitochondria fuse. Curr Opin Cell Biol. 2005;17:389–394. 166. Altmann K, Westermann B. Role of essential genes in mitochondrial morphogenesis in Saccharomyces cerevisiae. Mol Biol Cell. 2005;16:5410–5417. 167. Anesti V, Scorrano L. The relationship between mitochondrial shape and function and the cytoskeleton. Biochim Biophys Acta. 2006;1757:692–699. 168. Okamoto K, Shaw JM. Mitochondrial morphology and dynamics in yeast and multicellular eukaryotes. Annu Rev Genet. 2005;39:503–536. 169. Meeusen SL, Nunnari J. How mitochondria fuse. Curr Opin Cell Biol. 2005; 17:389–394. 170. Meeusen S, McCaffery JM, Nunnari J. Mitochondrial fusion intermediates revealed in vitro. Science. 2004;305:1747–1752. 171. Amutha B, Gu Y, Pain D. A novel role of Mgm1p, a dynamin-related GTPase, in ATP synthase assembly and cristae formation/maintenance. Biochem J. 2004;381:19–23.

70

BASIC MOLECULAR BIOLOGY OF MITOCHONDRIAL REPLICATION

172. Olichon A, Baricault L, Gas N, et al. Loss of OPA1 perturbates the mitochondrial inner membrane structure and integrity, leading to cytochrome c release and apoptosis. J Biol Chem. 2003;278:7743–7746. 173. Song BD, Schmid SL. A molecular motor or a regulator? Dynamin’s in a class of its own. Biochemistry. 2003;42:1369–1376. 174. Scarpulla RC. Nuclear control of respiratory gene expression in mammalian cells. J Cell Biochem. 2006;97:673–683. 175. Scarpulla RC. Nuclear activators and coactivators in mammalian mitochondrial biogenesis. Biochim Biophys Acta. 2002;1576:1–14. 176. Scarpulla RC. Transcriptional activators and coactivators in the nuclear control of mitochondrial function in mammalian cells. Gene. 2002;286:81–89. 177. Scarpulla RC. Nuclear control of respiratory chain expression in mammalian cells. J Bioenerg Biomembr. 1997;29:109–119. 178. Wu Z, Puigserver P, Andersson U, et al. Mechanisms controlling mitochondrial biogenesis and respiration through the thermogenic coactivator PGC-1. Cell . 1999;98:115–124. 179. Michalik L, Auwerx J, Berger JP, et al. International Union of Pharmacology: LXI. Peroxisome proliferator-activated receptors. Pharmacol Rev. 2006;58:726–741. 180. Handschin C, Spiegelman BM. Peroxisome proliferator-activated receptor {gamma} coactivator 1 coactivators, energy homeostasis, and metabolism. Endocr Rev. 2006;27:728–735. 181. Uldry M, Yang W, St Pierre J, Lin J, Seale P, Spiegelman BM. Complementary action of the PGC-1 coactivators in mitochondrial biogenesis and brown fat differentiation. Cell Metab. 2006;3:333–341. 182. Lin J, Handschin C, Spiegelman BM. Metabolic control through the PGC-1 family of transcription coactivators. Cell Metab. 2005;1:361–370. 183. Rodgers JT, Lerin C, Haas W, Gygi SP, Spiegelman BM, Puigserver P. Nutrient control of glucose homeostasis through a complex of PGC-1alpha and SIRT1. Nature, 2005;434:113–118. 184. Spiegelman BM, Heinrich R. Biological control through regulated transcriptional coactivators. Cell. 2004;119:157–167. 185. Nisoli E, Carruba MO. Nitric oxide and mitochondrial biogenesis. J Cell Sci. 2006;119:2855–2862. 186. Clementi E, Nisoli E. Nitric oxide and mitochondrial biogenesis: a key to long-term regulation of cellular metabolism. Comp Biochem Physiol A Mol Integr Physiol. 2005;142:102–110. 187. Nisoli E, Clementi E, Paolucci C, et al. Mitochondrial biogenesis in mammals: the role of endogenous nitric oxide. Science. 2003;299:896–899.

3 DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY Rhea Mehta, Katie Chan, Owen Lee, Shahrzad Tafazoli, and Peter J. O’Brien Graduate Department of Pharmaceutical Sciences, University of Toronto, Toronto, Ontario, Canada

1. Introduction 2. Drug-induced fatty liver (microvesicular steatosis) and steatohepatitis: endogenous lipotoxins 2.1. Fatty acids as endogenous toxins in NASH 2.2. Drug-induced steatosis or NASH 3. Drug-induced hepatic cholestatic injury 3.1. Endogenous bile acid toxins 3.2. Drug-induced cholestasis 3.3. Prevention and therapy for drug-induced cholestasis 4. Drug-induced oxidative stress and tissue toxicity: endogenous or exogenous reactive oxygen species toxins 4.1. Drugs or xenobiotics that inhibit the electron transport chain 4.2. Drugs that inhibit mitochondrial DNA synthesis 4.3. Drugs that uncouple mitochondrial oxidative phosphorylation (complex V) 4.4. Mitochondrial oxidative stress induced by drugs independent of respiratory inhibition 4.5. Prevention and therapy for drug-induced oxidative stress 5. Structure–activity relationships 5.1. Mitochondrial toxic drugs 5.2. Mitochondrial/lysosomal accumulation by cationic amphiphile drugs 6. Conclusions

72 72 72 75 85 85 86 87 88 88 100 101 105 109 109 109 111 113

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

71

72

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

1. INTRODUCTION In this chapter we survey the literature describing drugs that have been shown to impair mitochondrial function. Mitochondrial toxicity frequently underlies the etiology of adverse events that, in many cases, justiﬁed market withdrawal or regulatory interventions. Cells with mild mitochondrial toxicity that do not affect viability under normal conditions, such as oxidative stress or cytokines, can become more vulnerable to various stressors. We have separated mitochondrially toxic drugs into classes according to the function that is impaired, and when known, the targets and mechanisms of action drugs are described. An attempt has also been made to examine the physicochemical parameters and the structure–activity relationships that contribute to the mitochondrial toxicity of some drugs in the hope that such an endeavor will improve drug discovery and development programs. Drugs often increase endogenous toxins such as reactive oxygen species (ROS), fatty acids, and bile acids. The idiosyncratic toxicity often seen with these drugs suggests that it could also be useful to classify them according to the endogenous toxin that they evoke. In this way it may be possible to improve patients’ resistance to mitochondrial toxicity by decreasing the endogenous toxin exposure via polypharmy or nutritional means, including vitamin or cofactor supplements.

2. DRUG-INDUCED FATTY LIVER (MICROVESICULAR STEATOSIS) AND STEATOHEPATITIS: ENDOGENOUS LIPOTOXINS Although hepatic steatosis is usually nonprogressive, 20 to 25% of nonalcoholic steatohepatitis (NASH) cases may slowly progress to cirrhosis, and recently, cases of hepatocellular carcinoma (HCC) have been identiﬁed in patients undergoing long-term valporate therapy [1]. Major risk factors for primary NASH include obesity, type 2 diabetes mellitus, and hyperlipidemia, whereas secondary NASH is caused by drugs, nutritional factors, as well as metabolic or genetic disorders. Insulin resistance is the most important underlying disorder. Hepatic steatosis was once considered to be benign, but now is believed to have the potential of increasing the cellular free fatty acid pool with cytotoxic consequences. This is because some drugs inhibit the mitochondrial short-, medium-, and long-chain (but not very long-chain) acyl-coenzyme A (CoA) dehydrogenases responsible for the β-oxidation of medium-chain fatty acids in the matrix. Other drugs inhibit carnitine palmitoyltransferase I, which controls access of long-chain fatty acids to the mitochondrial site of β-oxidation [2]. 2.1. Fatty Acids as Endogenous Toxins in NASH Most long-chain fatty acids are bound to speciﬁc or unspeciﬁc fatty acid–binding proteins such as serum albumin. A small fraction is associated with membranes, and only a minute remainder is free. Nonetheless, an elevation of plasma free fatty

DRUG-INDUCED FATTY LIVER AND STEATOHEPATITIS

73

acids is associated with increased insulin resistance in skeletal muscle and release of proinﬂammatory cytokines. Increased fasting plasma free fatty acid levels have also been associated with a higher risk of type 2 diabetes The insulin-sensitizing effects of thiazolidinedione drugs may occur by decreasing plasma free fatty acids as a result of decreased adipose-derived lipolysis [3]. Five activation mechanisms have been proposed to explain the toxicity of free fatty acids. 1. Mitochondrial uncoupling and reactive oxygen species (ROS) formation. Addition to respiring mitochondria of micromolar concentrations of long-chain free fatty acids (e.g., oleate and palmitate) decreased transmembrane potential, increased state 4 respiration with a smaller decrease in state 3 respiration, decreased the respiratory control ratio, and decreased the ADP/O ratio. Fatty acids can thus act as natural uncouplers that decrease reactive oxygen species formation in the resting state by accelerating electron transport and thereby increasing the oxidized state of the respiratory complexes and carriers [4]. However, when calcium-loaded mitochondria are exposed to micromolar free fatty acids, ROS are generated at coupling sites I and II [4]. The impermeability of the inner membrane is undermined by ROS, leading to opening of the permeability transition pore, causing matrix swelling, rupturing the outer membrane, and releasing proapoptotic proteins. 2. Drug-induced inhibition of mitochondrial fatty acid oxidation causing ROS formation. Most of the ATP generated in cells arises from β-oxidation of fatty acids. As shown in Figure 1, the ﬁrst step of β-oxidation in the mitochondrial matrix is catalyzed by the FAD cofactor of acyl-CoA, dehydrogenase that oxidizes fatty acyl CoA, generating FADH2 , which is then reoxidized by the electron transfer protein (ETF) located in the inner membrane. The reduced ETF then transfers its electrons to coenzyme Q (CoQ) and hence, via complexes III and IV, to oxygen, resulting in the formation of water and two ATPs for each pair of carbons removed from the fatty acid. Inhibition of acyl-CoA dehydrogenase by drugs would be expected to slow mitochondrial respiration and thereby increase ROS formation. 3. Mitochondrial toxicity by unsaturated fatty acid autoxidation. One hypothesis is that the accumulation of unsaturated fatty acids in hepatocytes contributes to the progression of steatosis to NASH. These fatty acids could undergo oxidation to form toxic peroxyl radicals and carbonyls that may inhibit mitochondrial respiration, resulting in further ROS formation. In turn, this could promote inﬂammation or induce ﬁbrogenesis. The three enzymes that catalyze oxidation of unsaturated fatty acid toxins include cyclooxygenases, lipoxygenases, and cytochrome P450s. Hepatic CYP2E1, a P450, was also induced in NASH, which generated ROS and increased cellular endogenous ROS formation [5]. CYP2E1 also readily catalyzed unsaturated fatty acid oxidation-induced oxidative stress, which released calcium from intracellular stores and caused mitochondrial toxicity [6].

Intermembrane

74

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

(A) Electron transport chain H+

H+

H+

H+

H+

H+

A- H+

Cyt c

+++++ Q ETF

I Matrix

---

III

IV

Complex I inhibitors:

Complex III inhibitors:

Rotenone, Piericidin, Capsaicin Antimycin A Antihyperlipidemics Stigamatellin Anesthetics Acetaminophen quinoneimine Antidiabetics Isoflurane Anticonvulsants, Idebenone Complex II inhibitors: MPTP, Antipsychotics Malonate Flutamide Oxaloacetate O Isoniazid FADH2 SCoA R

V

Complex IV inhibitors: ADP

(C) Mt protein synthesis and biogenesis

Complex V/ ATP synthase Uncouplers:

H+

Free fatty acids, bile acids Pentamidine NSAIDS. Tamoxifen Tolcapone, Propofol

Inhibitor: CoA

R

AH

H+

O

FAD

ATP

Cephaloridine

II

oligomycin

S OH O

Cytochrome complexes

CoA R

Translation

O2

S

Inhibitors of fatty acid β– oxidation:

O

O

CoA Transcription

R

Antivirals: Zidovudine

MtDNA

O R

CoA S

S

+ R

Tetracycline, NSAIDs, antidepressants, tamoxifen

(D) Oxidative stress caused by:

citric acid cycle

O

Doxorubicin-semiquinone Gentamicin Trovafloxacin

CoA S

(B) Fatty acid β-oxidation

2CO2

Figure 1 Mitochondria–drug interactions: (A) inhibition of mitochondrial electron transport chain complexes; (B) inhibition of fatty acid β-oxidation; (C) inhibition of protein synthesis and biogenesis; (D) formation of mitochondrial oxidative stress. (See insert for color representation of ﬁgure.)

Evidence for this was that arachidonic acid (20 µM) caused apoptosis (mitochondrial cytochome c release) when incubated with rat hepatocytes or HepG2 cells. This did not occur in HepG2 cells, which did not express CYP2E1. This arachidonate cytotoxicity was prevented by antioxidants, e.g., trolox or deferoxamine (a ferric chelator) or diallyldisulﬁde (a CYP2E1 inhibitor), cyclosporine (an inhibitor of mitochondrial permeability transition), or caspase 3 inhibitor. Cytotoxicity (necrosis) was increased by iron, and lipid peroxidation ensued. The CYP2E1 requirement was presumably to generate ROS that catalyzed the arachidonate oxidation. Release of calcium from mitochondrial or endoplasmic reticular calcium stores by oxidative stress was part of the cytotoxic mechanism, and cytotoxicity could be prevented by inhibitors of the calcium-activated phospholipase A2. Calcium activation of phospholipaseA2 would cause arachidonate release and further increase mitochondrial toxicity. The mitochondrial toxicity signaling pathway involved activation of p38 MAPK, whereas the transcription factor Nrf2 pathway prevented toxicity, probably by increasing hepatocyte GSH levels. Calcium-loaded mitochondria are particularly susceptible to arachidonate by a mechanism involving the permeability transition with hepatocyte mitochondria and the ATP/ADP translocator with heart mitochondria [7].

DRUG-INDUCED FATTY LIVER AND STEATOHEPATITIS

75

4. Mitochondrial toxicity by peroxisomal reactive oxygen species. The β-oxidation of fatty acids by fatty acyl-CoA oxidase in peroxisomes forms H2 O2 , unlike the β-oxidation of fatty acids catalyzed by mitochondrial acyl-CoA dehydrogenase. However, much of the H2 O2 formed in peroxisomes is removed by catalase located in the peroxisomes [8]. 5. Mitochondrial and lysosomal lipoapoptosis induced by saturated fatty acids. Saturated fatty acids incubated with mouse hepatocytes were more effective than monounsaturated fatty acids at causing sustained activation of JNK1, which mediated mitochondrial apoptosis involving mitochondrial membrane depolarization, cytochrome c release, and caspase activation. JNK activation may engage the core mitochondrial proapoptotic machinery with Bim-mediated Bax activation. Small interfering RNA targeted knockdown of Bim decreased Bax activation and cell death [9]. JNK inhibitors could prove useful in preventing NASH and end-stage liver disease. Saturated long-chain fatty acids bound to serum albumin when incubated with hepatocytes caused a translocation of cytosolic Bax to lysosomes, which destabilized (permeabilized) their membrane. This released the protease cathepsin B into the cytosol and signaled a TNF-α cascade-induced apoptosis. Genetic or pharmacological inhibition of cathepsin B also prevented fatty liver disease in mice induced by a sucrose diet [10]. Bax inhibitors also prevented fatty acid–induced hepatocyte apoptosis [11]. Lipoapoptosis in β-cells of the islets leading to diabetes or in heart leading to myopathy can be prevented either by caloric restriction, thiazolidinedione treatment, or by iNOS inhibitors [12].

2.2. Drug-Induced Steatosis or NASH (Drug chemical structures are given in Figure 2.) Fatty liver induced by some drugs or xenobiotics usually results from an inhibition of mitochondrial fatty acid β-oxidation. However, in the case of amiodarone, valproic acid, tetracyclines, and demeclocycline, inhibition of hepatocyte triglyceride secretion in very low density lipoprotein (VLDL) also contributes. Ethanol-induced fatty liver results from similar mechanisms but is exacerbated by increased fat mobilization from adipose tissue. Ethionine-induced fatty liver results from inhibition of protein synthesis as well as inhibition of hepatocyte triglyceride secretion in lipoproteins. Fatty liver that progresses into steatohepatitis is also often associated with lipid peroxidation in mice. Administration of ethionine and other compounds to mice caused a sixfold increase in both hepatic triglycerides and lipid peroxidation (measured via ethane exhalation) in all animals after 24 hours. The rank order of potency for hepatic triglyceride elevation and ethane exhalation was ethionine > chlortetracycline > tetracycline, demeclocycline > amiodarone, amineptine > valproate > pirprofen ethanol. A single dose of dexamethasone or doxycycline did not induce fatty liver and lipid peroxidation, although both were apparent after repeated doses. It was concluded that microvesicular steatosis is associated with lipid peroxidation and that “the mere presence of

76

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

Drugs inhibiting mitochondrial fatty acid β-oxidation N

OH

Cl OH O N

NH2

OH OH

O

OH

OH

O

OH

O

Tetracycline

Ibuprofen

Pirprofen O

O S

N

NH

NH2

O NH

O

O OH

OH

Amineptine

Tianeptine

Tamoxifen

Mitochondrial CoA depletion O

OH

Valproic acid

Figure 2 Drugs inhibiting mitochondrial fatty acid β-oxidation, and mitochondrial CoA depletion.

oxidizable fat in the liver triggers lipid peroxidation” [13]. The exact molecular mechanisms for lipid peroxidation are not known, but in the case of ethanol, CYP2E1 and/or ROS catalyzes the oxidation of ethanol to prooxidant α-hydroxyethyl radicals and acetaldehyde. Also, ethanol converts xanthine dehydrogenase to xanthine oxidase, which forms ROS from acetaldehyde. Ethanol also increases hepatocyte NADH levels, which inhibits β-oxidation and later inhibits VLDL secretion [14]. This has been proposed as a mechanism for alcohol-induced steatohepatitis. What contributes to the progression of the steatosis to chronic steatohepatitis (e.g., hepatocyte necrosis, Mallory bodies, hepatic Kupffer cell activation, neutrophil inﬁltration, ﬁbrosis, and cirrhosis) in a few patients is not known. Carbonyls formed from lipid peroxide decomposition products are toxic, to some extent because they covalently bind to proteins and form advanced glycation end products (ALEs). One possibility is that ALEs elicit an inﬂammatory response by binding to ALE receptors. The activated immune cells could then cause hepatocyte necrosis by releasing cytotoxic ROS and cytokines such as TNF.

DRUG-INDUCED FATTY LIVER AND STEATOHEPATITIS

77

a. Inhibition of Mitochondrial Fatty Acid β-Oxidation by Drugs Tetracyclines and Antibiotics Tetracyclines bind to the 30S bacterial ribosomal subunit, which prevents protein synthesis. They are therefore useful as broad-spectrum antibiotics. However, at higher doses they cause fatty liver containing primarily accumulated triglycerides, particularly in female rats and especially if pregnant. This microvesicular steatosis probably results from the rapid and speciﬁc uptake of tetracycline by mitochondria, which inhibited fatty acid β oxidation by mouse hepatic isolated mitochondria and was reversible. Mouse liver triglycerides were increased 2.5-fold in 6 hours and 9-fold in 24 hours by tetracycline in vivo per os (p.o.), and histology showed microvesicular steatosis [15]. Tetracycline at 0.25 mmol/kg also markedly increased hepatic triglycerides and induced lipid peroxidation (ethane exhalation) in mice that was maximal for both 24 hours later [13]. Furthermore, hepatocyte studies showed that tetracycline (1 mM) also impaired the association of triglycerides and apoproteins in the Golgi apparatus but did not cause hepatocyte cytotoxicity [16,17]. However, liver ATP levels in vivo were not affected by the low dose of 50 mg/kg, intravenous (i.v.), rolitetracycline, although fatty liver was induced [16]. Electron microscopy of the liver 3 hours after tetracycline dosing (250 mg/kg, intraperitoneal (i.p.), using Fischer rats) showed that mitochondrial swelling had occurred. Tetracycline also accumulated predominantly in centrilobular hepatocytes and Kupffer cells. Only sparse focal centrilobular necrosis was found at 24 hours. Direct uncoupling of oxidative phosphorylation (OXPHOS) in isolated mitochondria has also been reported [18,19]. Tetracycline-induced hepatic dysfunction is particularly apparent in pregnant women, and increased plasma transaminase levels have been reported in half of the patients taking tetracycline [20]. 2-Arylpropionic Acid Agents (e.g., Pirprofen, Ibuprofen), Nonsteroidal Anti-Inﬂammatory Drugs Pirprofen was an NSAID introduced in 1982 as a treatment for arthritis but caused fulminant hepatitis in a few patients in addition to gastro-intestinal problems. Microvesicular steatosis was also found in their liver biopsy specimens. Pirprofen (2 mol/kg) administered to mice inhibited fatty acid oxidation and decreased plasma ketone bodies and glucose levels in vivo, resulting in microvesicular steatosis. Fatty acid oxidation catalyzed by isolated mouse liver mitochondria was also inhibited by pirprofen [21]. Ibuprofen also inhibited in vivo mitochondrial fatty acid oxidation and inhibited the in vitro β-oxidation of medium- and short-chain fatty acids by mouse liver mitochondria and human lymphocyte mitochondria [22]. Tricyclic Agents Amineptine and Tianeptine, Antidepressant Drugs These drugs have a tricyclic moiety and a heptanoic acid side chain. Amineptine and, to a lesser extent, tianeptine can cause hepatitis associated with microvesicular steatosis. Their heptanoic acid side chain may be responsible for reversibly inhibiting mitochondrial fatty acid oxidation by a competitive mechanism [22,23].

78

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

Tamoxifen, an Antiestrogenic Drug Tamoxifen is a major chemopreventive antiestrogenic drug used against breast cancer. Tamoxifen has also induced fatty liver in some breast cancer patients, resulting in steatohepatitis [24].Transcript and metabolite analysis showed that fatty acid synthase was down-regulated, thereby increasing malonyl CoA, which blocked fatty acid oxidation and caused fatty liver [25]. Tamoxifen has also been shown to uncouple mitochondria (see Section 4.3.c). Combination Antiretroviral Drug Therapy Antiretroviral drug therapy (HAART) has resulted in a major decrease in AIDS morbidity and mortality. However, the major long-term side effect of the drugs used in the primary care of HIV-infected patients is the lipodystrophy syndrome consisting of NASH, insulin resistance, and redistribution of body fat. AZT or stavudine may be partly responsible, as they are mitochondrial toxins [26] (see Chapters 9 and 21). b. Cationic Amphiphilic Drug-Induced Mitochondrial Respiration Inhibition and Phospholipidosis (Drug chemical structures are given in Figure 3.) Most cardiovascular drugs are cationic amphiphilic molecules with a lipophilic aromatic ring system and a hydrophilic amino-substituted side chain, usually protonated at neutral pH—hence the class term cationic amphiphilic drugs (CADs). Lipophilic cations can readily pass through phospholipids bilayers, particularly through membranes with a large transmembrane potential such as the mitochondrial inner membrane. They therefore accumulate readily in the mitochondrial matrix. Many cationinc amphiphilic drugs can also reversibly inhibit the anion channel (IMAC), which mediates the electrophoretic transport of a wide variety of anions and is believed to be an important component of the volume homeostatic mechanism. Lysosomal phospholipidosis is also induced by these drugs, probably because amphiphilic cationic drugs cross the lysosomal membrane and become trapped inside the lysosome because of their acidic intralysosomal milieu. These drugs also form reversible but tight complexes with phospholipids that accumulate in the lysosomes, as the drugs also inhibit the action of intralysosomal phospholipase A1. These phospholipids and drugs accumulate in the lysosome and form lamellar myelinlike bodies in the hepatocyte lysosomes [27]. Phospholipidosis is therefore considered a phospholipid storage disease. Many of these drugs accumulate in the lung and induce lysosomal lipidosis, which increases cell size. Foamy macrophages accumulate within the alveolar spaces of the lung. The lung is probably a common organ target because this organ has the highest phospholipid turnover, due to the synthesis and recycling of pulmonary surfactant phospholipids. 4,4-Diethylaminoethoxyhexestrol, an Antianginal Drug 4,4-Diethylaminoethoxyhexestrol (DEAEH) was used as a coronary dilator (Coralgil) in Japan in 1978 but was recalled after it had caused more than 100 cases of severe liver injury. This was associated with microvesicular and/or macrovacuolar steatosis which after a few months or years progressed to steatohepatitis and liver necrosis,

79

DRUG-INDUCED FATTY LIVER AND STEATOHEPATITIS

Cationic amphiphilic drugs

NH

O

O

NH

N H2 NH3

Cl

N NH

O

F

F F

S

(1) 4,4’-diethylaminoethoxyhexosterol (2) Chlorpromazine

N

(3) Fluoxetine

(4) Tacrine

H2N HO

NH

NH HN

O N N H2

(5) Imipramine

Cl

(7) Perhexiline

(6) Propanolol

N

(8) Chloroquine

N N

H

HO

H OH

O O

HO

N

(9) Quinidine

H

O

(10) Buprenorphine

Benzofurans

NH

I O

O

O

O Br

O

I OH

O

O

O Br

(1) Benzarone

(2) Amiodarone

(3) Benzbromarone

(4) Benzofuran

Figure 3 Drug analogs that inhibit mitochondrial respiration and induces phospholipidosis. Compounds are in order of cytotoxicity, (1) being the most toxic.

80

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

Mallory bodies, neutrophil inﬁltration, and ﬁbrosis. Liver triglycerides and cholesterol were increased 70% in humans. Liver cholesterol esters in rats were increased nine fold, triglycerides were increased fourfold, and phospholipids were increased 2.3-fold [28]. In humans, phospholipidosis is induced by 2.5 mg/kg DEAEH. Lysosomal phospholipidosis in Fischer 344 rats was shown to involve accumulation of DEAEH in the liver lysosomes, reaching a total (free and bound) intralysosomal concentration of 8 mM in 2 hours and 21 mM in 24 hours, followed by complete inhibition of lysosomal phospholipase, as only DEAEH concentrations above 1 mM were required to inhibit phospholipase A1 completely. Phospholipids accumulated six fold in the lysosomes in 12 hours [29]. However, the phospholipidosis was not believed to have an adverse effect on liver function. Mitochondrial studies, on the other hand, suggested that the steatohepatitis could be attributed to DEAEH accumulation in the mitochondria, which inhibited mitochondrial respiration, resulting in ROS formation, and also inhibited fatty acid β-oxidation, resulting in microvesicular lipid deposits (microvesicular steatosis). Unsaturated lipids were then oxidized by the ROS to form lipid peroxides. Rat hepatocytes cultured for 24 hours with 10 µM DEAEH resulted in a doubling of triglyceride levels, drastic ATP depletion, and inhibition of fatty acid β-oxidation. Electron microscopy of the hepatocytes exhibited microvesicular steatosis, elongated or giant mitochondria, and lysosomal phospholipidosis [30]. Benzofurans (e.g., Amiodarone), Antianginal Drugs Amiodarone consists of a benzofuran ring carrying a C4 H9 side chain and a diiodobenzene ring carrying a diethylaminoethoxy side chain. It has class III antiarrhythmic activity and currently is used widely in controlling intractable cardiac arrhythmias that do not respond to other drugs. Besides its use in treating tachyarrhythmias, it is also used for decreasing mortality postmyocardial infarction. However, it can cause several adverse effects to the thyroid, lung pulmonary, and liver. Amiodarone was much less toxic to cardiomyocytes than to hepatocytes, and up to 10 µM protected cardiomyocytes from mitochondrial injury, loss of energy metabolism, and mitochondrial swelling induced by intracellular calcium after transient ischemia reperfusion. However, mitochondrial uncoupling and a permeability transition occurred above 30 µM [31]. About 1 to 2% of patients taking amiodarone suffer from symptomatic liver disease histologically similar to alcoholic steatosis. Both of these hepatic diseases involve fatty liver, microvesicular steatosis, and mitochondrial toxicity. Amiodarone is protonated in the acidic intermembrane space and is electrophoretically transported into the mitochondrial matrix, where a proton is released into the more alkaline matrix. The driving force is probably the membrane potential of the inner membrane and results in a large accumulation of amiodarone in mitochondria, reaching 6.8 nmol/mg protein when 200 µM amiodarone was incubated with isolated mouse liver mitochondria. This caused a collapse of the mitochondrial potential and a protonophoric uncoupling of OXPHOS [32]. Hepatocyte mitochondrial potential was decreased by 33% by 20 µM amiodarone, while state 3 respiration and respiratory control ratios for glutamate were decreased 50% by 13 µM amiodarone. In hepatocyes isolated from Sprague–Dawley (SD) rats,

DRUG-INDUCED FATTY LIVER AND STEATOHEPATITIS

81

mitochondrial respiration involving complexes I to III was found to be more susceptible to inhibition by amiodarone, and hepatocyte ATP levels decreased to 30% of normal after 100 µM amiodarone exposure. ROS formation and opening of the mitochondrial permeability transition pore, leading to cytochrome c release, were involved. This was shown to trigger necrosis and/or apoptosis, depending on the extent of ATP depletion. Ascorbate partly decreased the apoptosis [33,34]. Mitochondrial β-oxidation of palmitic acid was nearly completely inhibited by amiodarone, probably as a result of the inhibition of carnitine palmitoyltransferase 1 [2]. This could explain the 2.5-fold increase in triglycerides that was associated with the appearance of small lipid droplets in the hepatocytes characteristic of microvesicular steatosis. Structure–activity relationships for amiodarone analogs or metabolites suggest that the benzofuran moity, along with its side chains, was responsible for the mitochondrial toxicity, and the presence of iodo moiety was not essential [34]. The N-dealkylated metabolites of amiodarone (LC50 20 µM) were more toxic than amiodarone (LC50 50 µM) toward primary rat hepatocytes in culture [35]. The N-dealkylated metabolite also induced considerally more cytotoxicity of HepG2 cells than did 100 µM amiodarone. Alkyl substitutions at the amino group not only increased hepatocyte cell death but also increased mitochondrial uncoupling activity. The latter effect may be explained by the positive inductive effect of the alkyl groups, rendering the amino group more basic and better as a proton carrier. Furthermore, derivatives with large alkyl substituents at the amino group would have a higher log P value and once deprotonated in the basic mitochondrial matrix, would more readily diffuse out of the matrix and mitochondria. This would explain their lower toxicity toward the electron transport chain and fatty acid β oxidation than that of the amiodarone N-dealkylated metabolites. Some of the less toxic derivatives were also less inhibitory to hERG channels, correspondingly lowering the risk of fatal torsade de pointes [34]. Benzbromarone, used as an uricosuric agent, is structurally similar to amiodarone except that it lacks the cationic diethylaminoethoxy side chain and substitutes a dibromobenzene for diiodobenzene. It was less toxic than amiodarone. However, benzarone, used for the treatment of venous vascular disorders, has a phenol instead of a dibromobenzene moiety and was more toxic than amiodarone [33]. Amiodarone-induced pulmonary toxicity progressing to ﬁbrosis has been diagnosed in 1 to 13% of patients receiving high doses. Prognosis is poor, with 10% fatality if not treated. In hamster lung alveolar macrophages, amiodarone (100 µM) decreased mitochondrial membrane potential, inhibited respiration prior to ATP depletion, and led to cell death. The N-dealkylation metabolite was more toxic than amiodarone [36]. Fibrosis in the hamster model developed after 21 days, as shown by increased hydroxyproline, and was associated with lipid peroxidation. Fibrosis, but not mitochondrial toxicity, was repressed by pirfenidone and vitamin E supplementation. Pirfenidone suppressed pulmonary transforming growth factor (TGF) expression, and its protective effect was attributed to this [37].

82

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

Lysosomal phospholipidosis is also induced by amiodarone, probably because it crosses the lysosomal membrane and becomes trapped inside by the acidic milieu. The drug then forms reversible but tight complexes with phospholipids. Phospholipids accumulate in the lysosomes, as amiodarone inhibits the action of intralysosomal phospholipase A1. Serum phospholipids were also increased. Amiodarone can also cause phospholipidosis in alveolar macrophages, but its contribution to lung toxicity or liver toxicity is not known. Interestingly, the accumulation of amiodarone and phospholipids in the liver and alveolar macrophages of Fischer 344 rats was ﬁvefold greater than that with SD rats [38]. It is likely but not yet shown that Fischer rats are also more susceptible to amiodarone-induced steatohepatitis. Metabonomic studies using SD rats also showed that plasma and urine phenylacetylglycine, a phospholipidosis biomarker, were also increased [39]. Of 30 drugs screened in HepG2 cells, 17 caused phospholipidosis as determined by electron microscopy, with the most potent being ﬂuoxetine > amiodarone, chlorpromazine, perhexiline, sertraline > amitriptyline, chlorcyclizine. Gene expression analysis via DNA microarrays indicated that 12 markers were affected, including genes involved in the inhibition of lysosomal phospholipase, lysosomal enzyme transport (AP1 S1), and promotion of phospholipid and cholesterol biosynthesis [40]. The biological consequences of phospholipidosis to cell function are not clear, and this process could be part of a detoxiﬁcation mechanism in which lysosomes protect the cell by sequestering amphiphilic xenobiotics. Amiodarone-induced liver phospholipidosis in Fischer 344 rats was partly prevented by the antioxidants silymarin or vitamin E, but it should be noted that silymarin also moderated cellular uptake of amiodarone [41]. Dicyclohexyl-2-(-piperidyl) Ethane Agent Perhexiline Maleate, an Antianginal Drug This drug caused fatty liver and was associated with increased serum transaminase in 24 to 50% of the treated patients, often resulting in hepatitis [42]. Experiments with isolated mouse mitochondria showed that perhexiline accumulated in the mitochondria and inhibited complexes I and II, uncoupled oxidative phosphorylation, decreased ATP formation, and inhibited fatty acid β-oxidation. Perhexilene (25 µM) incubated with SD rat hepatocytes for 24 to 72 hours increased triglycerides twofold, inhibited respiration, depleted ATP, and inhibited fatty acid β-oxidation [43]. Phospholipidosis also occurred in patients, but it is not known yet whether perhexiline inhibits phospholipase. Quinuclidine Agent Chloroquine, an Antimalarial, and Quinidine, an Antiarrhythmic Agent Chloroquine is used to treat and prevent malaria, as it can kill the parasite in the red cell and in the hepatocyte. It also induced phospholipidosis fatty liver and accumulated to 6.3 mM in hepatic lysosomes, whereas 4.4 mM chloroquine completely inhibited lysosomal phospholipase [44]. Quinidine is a class I antiarrhythmic agent for the heart. It is a stereoisomer of quinine, originally derived from the bark of the cinchona tree. It also induced hepatic injury and phospholipidosis [40]. Quinidine also inhibited mitochondrial

DRUG-INDUCED FATTY LIVER AND STEATOHEPATITIS

83

ATPase (IC50 of 4.8 mM) and the ATP-dependent mitochondrial potassium channel [45,46]. Amphiphilic Cationic Agent Fluoxetine (Prozac), an Antidepressant Drug This is an amphiphilic cationic antidepressant that in a very small number of patients induces a hypersensitivity pneumonitis associated with pulmonary phospholipidosis. This is also seen in rats [47], and of 12 drugs tested, it induced phospholipidosis in HepG2 cells most potently [40]. Lipophilic Agent Buprenorphine, an Antalgic Drug This opiod is a lipophilic tertiary amine used to manage opiod addicion and chronic pain. It can cause cytolytic hepatitis in some patients and occasionally results in microvesicular steatosis. Buprenorphine is accumulated by mitochondria 14-fold over media, and it impaired fatty acid β-oxidation and ATP formation. It collapsed the membrane potential and uncoupled mitochondrial respiration indicative of a protonophoric mechanism. In hepatocytes it caused moderate GSH depletion, suggesting metabolic activation and/or oxidative stress. At 200 µM it caused early and severe ATP depletion, with marked lactate dehydrogenase release indicative of cell death [48]. Acridinamine Agent Tacrine, an Anti-Alzheimer Drug Tacrine is a reversible cholinesterase inhibitor that induces mild liver dysfunction (increased plasma ALT) in 50% of patients after 4 to 12 weeks of treatment. Hepatocyte studies showed ATP depletion, and mitochondria studies showed loss of mitochondrial potential and stimulated respiration. Like amiodorane, mitochondrial uncoupling by tacrine could result from a protonophoric effect. In this effect there is protonation of the quaternary amine moiety in the intermitochondrial membrane space, transport across the inner membrane by the membrane potential, and deprotonation with the basic pH of the matrix. A ninefold accumulation of tacrine was observed in the mitochondria. The base could then diffuse back across the inner membrane to the acidic intermembrane space without the protons passing across the F0 F1 -ATPase to form ATP. Tacrine also accumulated within the acidic lysosomes and caused phospholipidosis [49]. Tacrine therefore differed from amidorane in uncoupling mitochondria without inhibiting respiration or fatty acid oxidation and thus did not cause steatohepatitis. Such uncoupling explains early toxicity, but in addition, after 28 days in vivo it was found that tacrine inhibited topoisomerases I and II, thereby repressing mtDNA replication. The latter eventually leads to permeability transition and hepatocyte necrosis and/or apoptosis, and is proposed to underlie long-term toxicity in vivo [50]. c. Drug-Induced Mitochondrial CoA Sequestration Short-Chain Fatty Acid Agents (e.g., Valproic Acid), Antiepileptic Drugs Valproic acid is used for the treatment of seizures and is the most widely prescribed antiepileptic drug. Weight gain, the most common side effect, is associated with an increase in insulin resistance and limits its use. Abdominal

84

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

ultrasound results show that characteristics of fatty liver disease were present in 61% of 45 valproate-treated nondiabetic patients [51]. Furthermore, valproate can induce hepatic centrizonal necrosis associated with microvesicular steatosis following an acute overdose that impairs mitochondrial function. Ultrastructural evidence of mitochondrial injury was also found in a rat model [52]. In the clinic, mitochondrial injury by valproate can be determined by assaying α-ketobutyrate decarboxylase, a mitochondrial enzyme, by the [1-13 C]methionine breath test [53] (see Chapter 22). Valproate metabolism requires CoA and carnitine, which results in their depletion. This impairs mitochondrial β-oxidation of long-, medium-, and short-chain fatty acids [54]. Valproate also inhibits the mitochondrial trifunctional protein of fatty acid β-oxidation [55]. Another mechanism involves the formation of a mitochondrial toxic electrophilic metabolite formed by P450 and mitochondria (i.e., 2,4-diene-valproyl-CoA) [54]. Insulin resistance could also enhance adipose tissue lipolysis, increasing the hepatic inﬂux of fatty acids. Valproate also induced hepatic oxidative stress, including increased hydrogen peroxide levels and protein carbonyl formation. Conjugated dienes and malondialdehyde were also increased, indicating lipid peroxidation. Mitochondrial phosphatidylcholine and phosphatidylethanolamine were increased, whereas lysophosphatidyl ethanolamine and phosphatidylethanolamine were decreased [56]. In another study, rat liver and plasma 15-F2t-isoP, an arachidonate oxidation product, increased well before the onset of necrosis, steatosis, and increased serum α-GST reﬂective of hepatotoxicity, which occurred prior to onset of lipid peroxidation [57]. d. Prevention and Therapy for Drug-Induced NASH Cessation of drug administration as soon as NASH is diagnosed would bring some beneﬁt. Dieting and exercise also have beneﬁcial effects by improving insulin sensitivity in muscles, liver, and adipose tissue. Exercise and antidiabetic drugs such as metformin or thiazolidinediones also increase AMP-activated protein kinase (AMPK) activity, which lowers fatty acids by increasing fatty acid oxidation, and in liver, AMPK activation inhibits gluconeogenesis and improves insulin sensitivity. Hypolipidemic drugs such as bezaﬁbrate or gemﬁbrozil (but not cloﬁbrate) also prevented the progression of NASH [58]. Decreasing oxidative stress with superoxide dismutase (SOD) mimetics and natural antioxidants such as vitamin E could also be useful for preventing NASH progression. Increasing fatty acid oxidation via treatment with carnitine or β-aminoisobutyric acid, a thymine catabolite, has also been successful [58]. Recently, high doses of resveratrol, a polyphenolic phytoalexin found in grape skins, was successful in preventing mouse obesity induced by a high-fat diet. This was due to increases in the aerobic capacity of the obese mice, as shown by a doubling of their running endurance and correspondingly increased oxygen consumption in muscle ﬁbers [59]. The mechanism involved activating the protein deacetylase SIRT1, which accelerated mitochondrial biogenesis and thereby increased fatty acid oxidation. In another study, by preventing obesity, resveratrol also extended the life span of mice fed a high-fat diet by increasing insulin

DRUG-INDUCED HEPATIC CHOLESTATIC INJURY

85

sensitivity, decreasing insulinlike growth factor-1, and increasing AMP-activated protein kinase and peroxisome proliferator-activated PGC-1α. This increased mitochondria number and improved motor function [60].

3. DRUG-INDUCED HEPATIC CHOLESTATIC INJURY 3.1. Endogenous Bile Acid Toxins The Greek term cholestasis literally means “a standing still of bile.” Cholestasis is a condition in which bile cannot ﬂow from the liver to the duodenum. Bile formation is a secretory function of the liver, and bile is responsible for emulsiﬁng triacylglycerols in the duodenum, thereby rendering fats accessible to pancreatic lipases. Bile is also important for the elimination of xenobiotics (and their metabolites), cholesterol, bilirubin, and hormones. Bile acid accumulation can result from mechanical blockage in the duct system such as can occur from a gallstone or malignancy. It can also occur from disturbances in bile formation because of genetic defects or acquired as a side effect of many drugs. Because of negative feedback inhibition, bile acid accumulation in the hepatocyte represses bile acid synthesis from cholesterol. Thus, failure to excrete bile acids into the hepatic canaliculus resulted in elevated bile acid concentrations in hepatocytes and serum. Hydrophobic bile acids are hepatotoxic and can cause cirrhosis and liver failure. Normally, the concentration of total bile acids in the portal vein is approximately 20 µM, but during cholestasis it can reach 300 µM. Hepatocyte cytotoxicity occurs at concentrations less than that required for micellar detergentlike activity. The hydrophobic bile acids are sterol-derived molecules (e.g., chenodeoxycholic acid or glycochenodeoxycholic acid) and were previously thought to be toxic as a result of their membrane-disrupting detergent properties. However, at much lower and pathophysiologically relevant concentrations (20 to 100 µM) they induced hepatocyte apoptosis and oncotic necrosis, which could be prevented by an antioxidant lazeroid [61]. Rat hepatocyte studies showed that the relative cytotoxicity of bile acids after 24 hours was lithocholic acid (LD50 ∼ 75 µM)> chenodeoxycholic acid > glycochenodeoxycholic acid > ursodeoxycholic acid, taurochenodeoxycholic acid. Glycoursodeoxycholic acid and tauroursodeoxycholic acid were not toxic at 1 mM [62]. Bile acid toxicity involved mitochondrial toxicity because cytotoxicity was preceded by a decrease in mitochondrial membrane potential and ATP depletion. Furthermore, cytotoxicity and ATP depletion were prevented by fructose. However, it is not known whether ROS was involved. Studies with isolated rat liver mitochondria showed that bile acids decreased the mitochondrial membrane potential, increased state 4 respiration, and decreased state 3 respiration in a dose-dependent manner. With chenodeoxycholic acid, this occurred at 50 µM. Inhibition of state 3 respiration reﬂects OXPHOS impairment, while increased state 4 respiration indicates uncoupling.

86

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

The relative potency of bile acids at inducing mitochondrial permeability transition (MPT) in isolated mitochondria preloaded with calcium was litocholic acid (less than 15 µM) > chenodeoxycholic (less than 50 µM), deoxycholic acid (less than 100 µM) > ursodeoxycholic acid (∼300 µM). Ursodeoxycholic acid, taurochenodeoxycholic acid, and glycochenodeoxycholic acid were not toxic at 200 µM. This probably correlated with hydrophobicity [63]. Cyclosporine A prevented MPT induction and when induced by chenodeoxycholic acid, MPT resulted in cytochrome c release and apoptosis of HepG2 cells. In these cells, caspase 9 was also activated, and extensive PARP cleavage and DNA fragmentation occurred. By contrast, hydrophilic bile acids such as ursodeoxycholic acid and its taurine derivative stabilized the mitochondrial membrane through pathways independent of the MPT. Indeed, ursodeoxycholic acid has been used as therapy for cholestasis [64]. Human hepatic mitochondrial studies showed that glycochenodeoxycholic acid induced ROS formation and the permeability transition, which was prevented by cyclosporin and antioxidants. Cytochrome c and apoptosis-inducing factor were also released. This mitochondrial toxicity was prevented by cyclosporin A (a permeability transition inhibitor) or a caspase 9 inhibitor or ursodeoxycholic acid (an antiapoptotic hydrophilic bile acid) or the antioxidants tocopherol and glycyrrhetinic acid [65,66]. Hepatocytes incubated with this bile acid also induced ROS formation, MPT, and cytochrome c release, which was also prevented by antioxidants [67]. Besides the intrinsic mitochondrial cell death pathway, a cell surface extrinsic apoptosis pathway involving Fas receptors was also involved, as bile acids promoted the rapid transport of cytoplasmic vesicular Fas to the plasma membrane [68]. Fas-deﬁcient mice also have less hepatic injury and ﬁbrosis following bile duct ligation [69]. Tumor necrosis factor–induced apoptosis is another death-receptor signaling pathway involved and would be another therapeutic goal for treating cholestasis [70]. 3.2. Drug-Induced Cholestasis (Drug chemical structures are given in Figure 4.) Bile acids are synthesized from cholesterol in the hepatocyte primarily via two pathways: the “classic” pathway, initiated by microsomal cholesterol 7a-hydroxylase (CYP7A1), and an “alternative” (acidic) pathway initiated by sterol 27-hydroxylase (CYP27) in the inner mitochondrial membrane. The bile acids formed are then released into the bile canaliculus via an ATP-dependent bile salt export pump (BSEP) of the canalicular membrane. The intrinsic transport activity by BSEP of bile salts is taurochendeoxycholate > taurocholate > tauroursodeoxycholate > glycocholate. The sodium-dependent taurocholate transporter (NTCP) of the hepatocyte basolateral (sinusoidal) membrane mediates the removal of bile salts from the sinusoidal blood. Cholestasis in children has been attributed to a mutation in the bile salt excretory pump (BSEP) which if not diagnosed can result in liver cirrhosis and death from liver failure. This pump can also be inhibited by some drugs, resulting in a buildup of toxic bile salts in the hepatocyte, causing hepatitis. Drugs inhibiting BSEP include chlorpromazine, ketoconazole, cyclosporine

87

DRUG-INDUCED HEPATIC CHOLESTATIC INJURY OH OH O O

O

H H

H HO

H H H

HO

H

H HO

H

(1) Lithocholic acid

H

OH H OH

(3) Chenodeoxycholic acid

(2) Deoxycholic acid O OH

HN

OH O

O H H

HO

H H OH

H

HO

(4) Glycochenodeoxycholic acid

H

OH

(5) Ursodeoxycholic acid O S

HN

OH O

O H

HO

H

H H OH

(6) Taurochenodeoxycholic acid

Figure 4 Endogenous bile acid toxins that cause mitochondrial uncoupling and hepatic cholestatic injury. Compounds are in order of cytotoxicity, (1) being the most toxic.

A, rifampicin, glibenclamide, glyburide, nafazodone, and troglitazone [71], and most of them can induce cholestasis. Hepatitis is not induced by steroids (e.g., contraceptive steroids or 17α-alkylated androgenic steroids, which also inhibit BSEP). 3.3. Prevention and Therapy for Drug-Induced Cholestasis Cessation of cholestatic drug administration and decreasing fat consumption to no greater than 0.5 g/kg per day would be the best approach. Carnitine supplements could be useful if the cholestasis had caused carnitine deﬁciency. High doses of ursodeoxycholate (600 mg/day), a natural hydrophilic bile, acid that is present at low concentrations in human bile, has been a useful treatment for patients with cholestasis. This bile acid has antioxidant and cytoprotective properties and increases mitochondrial GSH, possibly by up-regulating glutamyl cysteine synthetase [58]. Another approach could be to use agents that induce hepatic P4503A,

88

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

which hydroxylates the toxic taurohydrodeoxycholic acid causing cholestasis, thereby decreasing its hepatotoxicity and facilitating its biliary excretion. 4. DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY: ENDOGENOUS OR EXOGENOUS REACTIVE OXYGEN SPECIES TOXINS Endogenous superoxide radicals are formed when the mitochondrial respiratory chain is inhibited and are also formed by cytochrome P450 of the endoplasmic reticulum electron transport system [72]. Immune cells also form superoxide radicals when a plasma membrane ecto NADPH oxidase is activated. Exogenous superoxide radicals can also be formed from the metabolism of environmental xenobiotics or drugs. Endogenous hydrogen peroxide is also formed by peroxisomal oxidases and is normally detoxiﬁed by peroxisomal catalase. Endogenous reactive nitrogen species are formed when superoxide radicals or oxygen react with nitric oxide formed by iNOS to form peroxynitrite or nitrogen dioxide, respectively. While reactive oxygen species react with protein amino acids to form protein carbonyls, reactive nitrogen species nitrosylate protein amino acids. These reactive oxygen and nitrogen species are readily detoxiﬁed by speciﬁc enzymes or cellular antioxidants because at low levels they can cause signaling, resulting in nonphysiological cell death (necrosis) or gene-regulated programmed cell death involving speciﬁc proteins (apoptosis). Necrosis is prominent in the core of a lesion induced by ischemia–reoxygenation, mitochondrial respiratory toxins, or uncouplers (resulting in mitochondrial swelling), whereas apoptosis dominates in the penumbra. Depletion of ATP favors a switch from apoptotic to necrotic cell death in part because ATP is required to fuel apoptosis. ROS plays a critical role by oxidizing the thiols of thioredoxin and glutaredoxin that are part of the apoptosis signal-regulating kinase (ASK1). This results in ASK1 activation, which activates MKK4 and causes sustained JNK activation, resulting in apoptosis. Other pathways activated by ROS that cause JNK activation include the Src-Gabl, GSTπ, and RIP-TRAF2 pathways. Tumor necrosis factor (TNF) binds to a surface membrane receptor, resulting in the formation of superoxide and caspase 8 activation, resulting in mitochondrial ROS formation. ROS or JNK also cause cytochrome c release, which activates caspases [73]. Hydrogen peroxide, transition metals, lysosomotropic toxins, and drugs that form ROS via redox cycling cause lysosomal membrane permeabilization, releasing cathepsins B and D, which lead to mitochondrial failure and cell death [74,75]. Furthermore, ROS can increase the expression of cytokines, including transforming growth factor (TGF-β), interleukin (IL-8), TNF-α, and Fas ligand [24]. 4.1. Drugs or Xenobiotics That Inhibit the Electron Transport Chain Many drugs or their electrophilic metabolites and xenobiotics inhibit mitochondrial electron transport chain, thereby increasing ROS formation, which can trigger MPT and cell death [76]. ROS can also be formed when prooxidant drug radicals are formed by peroxidase-catalyzed drug metabolism.

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

89

Inhibition of any of the four respiratory complexes in the respiratory steps can result in the generation of reactive oxygen species, as only cytochrome oxidase (complex IV) can reduce oxygen with four electrons to form water. Figure 1 illustrates the interactions of drugs with the mitochondrial electron transport system, fatty acid β-oxidation, protein synthesis, and biogenesis. a. Complex I Inhibitors (Drug chemical structures are given in Figure 5.) Complex I inhibitors include rotenoids, piericidins, capsaicins, and pyridinium-type inhibitors which act at or close to the ubiquinone reduction site. Many of these inhibitors have been synthesized for use as insecticide and acaricide agrochemicals. Barbiturates such as amytal were the ﬁrst drugs found to inhibit mitochondrial respiration by inhibiting NADH dehydrogenase. Other drugs that are also complex I inhibitors include meperidine (demerol), haloperidol, dequalinium chloride, cinnarizine, ranolazine, and idebenone [77]. The structure of complex 1 so far has been too complex to investigate by x-ray crystallographic studies, so the binding sites of these inhibitors to NADH dehydrogenase have not been identiﬁed. Radio- and ﬂuorescent-ligand binding experiments show that most inhibitors share a common large binding domain with partially overlapping sites at the terminal electron transfer step that reduces CoQ [78]. Rotenone, found in plants and still used as an insectide, was much more potent and more hydrophobic than amytal. Rotenoids binds irreversibly to mitochondrial NADH dehydrogenase (complex I) at two sites, one buried in the hydrophobic part of the inner membrane and the other at an external site on the matrix face. This binding blocked electron transport from the dehydrogenase iron–sulfur cluster to CoQ and results in ROS formation. Rotenone has been classiﬁed as a semiquinone antagonist [77]. Complex I ROS formation has been attributed to autoxidation of the ﬂavin semiquinone of NADH dehydrogenase. Rotenoids markedly inhibited rat brain complex I and caused Parkinson’s disease in rats, thereby providing a useful animal model for Parkinson’s disease [79]. Rotenone also increased mitochondrial ROS formation, DNA fragmentation, cytochrome c release, caspase 3 activation, and apoptosis in HL-60 cells with only a 36% decrease in ATP [80]. Piericidin A is produced by Streptomyces strains and contains a free pyridinol hydroxyl group which resembles the quinone ring of ubiquinone and is essential for inhibitory activity. Piericidin acts as a quinone antagonist at the ﬁrst site and as a semiquinone antagonist at the second site [77]. Capsaicin, the pungent active agent in hot peppers, acts as a competitive inhibitor for ubiquinone with complex I. Capsaicins with an acyl group of 10 to 12 carbons are much more potent inhibitors than capsaicin, indicating that hydrophobicity is important. The phenolic group was not essential for activity. Idebenone is a synthetic analog of CoQ10 , a component of the mitochondrial electron transport chain. It is used therapeutically to treat mitochondrial diseases [e.g., Friedreich’s ataxia, Leber’s hereditary optic neuropathy, mitochondrial encephalopathy with lactic acidosis and stroke-like episodes (MELAs)] and as therapy for cardiomyopathy. This is largely because it is an effective substrate for succinate–Q reductase and ubiquinol– cytochrome c reductase. However,

90

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY H

Classical inhibitors

Aromatic N-heterocyclic agent

HO H H H O

H

O

H N

O HN

O H

NH

O

O

O

O

O N

O

O

Amytal

Piericidin A

Rotenone

MPTP

Coenzyme Q analogs

O

O

H

O

OH

O

N H

O

H

HO

O

Capsaicin

Idebenone

Antipsychotic neuroleptic drugs Cl HO OH N

N

N

O

N

N

N

O F

F

Haloperidol

N

N

S

S

Chlorpromazine

N

F F

Fluphenazine

O

N

Risperidone

N N NH N

Cl

Clozapine

Figure 5

Respiratory complex I inhibitors.

it also acts as a quinone antagonist and effectively inhibits complex I [77]. Although used as an antioxidant, reduced idebenone can autoxidize to form ROS. Aromatic N-Heterocyclic Agents, MPTP Toxin and Haloperidol, Antipsychotic Neuroleptic Drugs Iminium metabolites, formed by the oxidation of tert-aliphatic amines, inhibit complex I. Examples include aromatic

91

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY Thiazolidinedione drugs O HO NH

O

O

NH

O

S

N O

O

O

N

S

Troglitazone

Rosiglitazone H N

O

O

O

S

Pioglitazone Antidiabetic biguanide drugs

NH2

H N

N NH

NH2

N

H2N

N N

NH2

N NH2

NH

Metformin

Buformin

NH2 NH2

Phenformin

Anesthetics F N

F

N HN

Br

O HN

F

O2N

O

NH

F

Cl

F

O F

Bupivacaine

Lidocaine

Halothane

Flutamide

Figure 5 (Continued)

N-heterocycles such as pyridinium, the well-known herbicide paraquat, and MPTP. 1. MPTP (1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine). In 1983 an outbreak of Parkinson’s disease in young illicit drug users in San Francisco was diagnosed as resulting from MPTP, a contaminant of an illicit synthesized narcotic. Further research showed that MPTP selectively killed dopaminergic neurons of the nigostriatal pathway after undergoing a MAO-catalyzed oxidation to MPP+ (1-methyl-4-phenylpyridine). The cation MPP+ is accumulated by mitochondria, where it inhibits complex I by oxidizing protein thiol groups, accelerating ROS formation and leading to neuronal death [81].

92

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

Hydantoins H N

O

O N N

H H N

O O

O NH

NO2

Dantrolene

Phenytoin

Antihyperlipidemic drugs

O

O O O

OH O

O

O

O

O

Cl Cl

Clofibrate

Cl

Fenofibrate

Cl

Ciprofibrate

Figure 5 (Continued)

2. Antipsychotic neuroleptic drugs haloperidol, chlorpromazine, ﬂuphenazine, risperidone, and clozapine (Figure 2). These drugs are used primarily as D2 dopamine receptor antagonists in the management of schizophrenia and bipolar disorders. However, they can result in severe extrapyramidal tract side effects, such as Parkinsonism and tardive dyskinesia. In 1964 it was ﬁrst shown that chlorpromazine inhibits brain and liver mitochondrial respiration [82]. Recently, haloperidol and ﬂuphenazine (<100 nM) were shown to deplete glutathione and inhibit mitochondrial respiration in brain slices. Respiration was restored by the thiol reductant dithiothreitol, suggesting that inhibition was due to oxidative complex I inactivation. The rank order of potency for inhibiting rat brain cortex mitochondrial complex I was haloperidol > chlorpromazine > ﬂuphenazine > risperidone [83]. Similar results were found in human brain cortex, although clozapine was less effective [84]. This order potency parallels the extrapyrimidal toxicity. Other studies showed that complex I and II inhibition by haloperidol may have similarities to MPTP in being caused by the HPP+ pyridinium metabolites formed by haloperidol oxidation catalysed by microsomal CYP3A

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

93

[85,86]. Cinnarizine and ﬂunarizine are piperazine derivatives with anticonvulsant and calcium antagonist properties, and both have been implicated in induction of Parkinsonism in elderly patients. Flunarizine (K i 1 to 10 µM) was more effective than cinnarizine at inhibiting mitochondrial complexes I and II [87]. Lipophilic Tertiary Amine Agents Bupivacaine and Lidocaine, Local Anesthetics 1. Bupivacaine. Local anesthetics such as bupivacaine are highly lipophilic tertiary amines which are associated with cardiotoxicity and myotoxicity. This could result partially from their ability to impair mitochondrial energy metabolism by cycling protons through the mitochondrial inner membrane (described below). Complex I was the most sensitive to inhibition by bupivacaine (involving a binding site on the cytosolic side). Bupivacaine induced myotoxicity and myopathies, probably as a result of its mitochondrial toxicity, resulting in PTP opening. Ropivacaine was less effective at inhibiting complex 1, probably because of its lower lipophilicity [88]. 2. Lidocaine. Lidocaine is a Na+ channel blocker and local anesthetic used for spinal anesthesia. It has now been associated with neurotoxicity initiated via inhibition of neuronal mitochondrial respiration, mitochondrial depolarization, cytochrome c release, and caspase activation [89]. Halogenated Hydrocarbon Agents Halothane, Enﬂurane, Isoﬂurane, and Sevoﬂurane, Volatile Anesthetics Type I hepatotoxicity (mild) is relatively common, occurring in 25 to 30% of patients after halothane exposure, and is attributed to reductive (anaerobic) metabolism resulting in lipid peroxidation. Administration of isoﬂurane or sevoﬂurane is less toxic, as they are metabolized more poorly. Type II hepatotoxicity (fulminant hepatitis) incidence is <1 : 6000 and is associated with massive centrilobular liver necrosis. Hepatotoxicity is probably immune-initiated, as a result of a P450-catalyzed oxidative metabolism (about 20%) to triﬂuoroacetyl chloride, which binds covalently to hepatic endoplasmic reticular proteins, which act as haptens. Type II halothane hepatotoxicity has a 50% fatality rate. Enﬂurane and isoﬂurane, on the other hand, undergo metabolism of about 2% and 0.2%, respectively, and are therefore much less hepatotoxic. Mitochondrial toxicity may contribute to the hepatitis, as halothane (1.5 mM) increases NADH levels of cardiomyocytes, probably by inhibiting complex I. Isoﬂurane and sevoﬂurane were less inhibitory. However, only halothane also inhibited succinate dehydrogenase, complex II [90]. Hydantoin Agents Dantrolene, a Muscle Relaxant, and Phenytoin, an Anticonvulsant 1. Dantrolene. Dantrolene, a muscle relaxant, is used to treat spasticity or muscle spasms associated with spinal cord injuries, stroke, multiple sclerosis, cerebral palsy, or other conditions. However, serious hepatic injury

94

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

is associated with this drug. Dantrolene inhibited hepatic mitochondrial glutamate oxidation and accelerated Ca-induced mitochondrial swelling. Complex I of submitochondrial particles was inhibited by dantrolene [91]. 2. Phenytoin. Phenytoin is an anticonvulsant drug which can be useful in the treatment of epilepsy. However, it is associated with liver injury. Mitochondrial studies show that there are two sites of inhibition by phenytoin; one site within the respiratory chain proximal to cytochrome b, and a second site within site I responsible for proton pumping and hence for ATP formation [92]. Fibrate Agents Cloﬁbrate and Fenoﬁbrate Antihyperlipidemic Drugs The ﬁbrate drugs (Figure 5), cloﬁbrate, ciproﬁbrate, and fenoﬁbrate are prescribed to decrease hyperlipidemia, thereby forestalling heart disease. They are agonists of the peroxisome proliferators receptor and induce peroxisomes containing high palmitoyl-CoA ligase activity, which metabolizes fatty acids and lowers serum triglycerides and LDL cholesterol while increasing HDL. However, in some patients they can cause liver toxicity, hepatomegaly, and muscle toxicity. Cloﬁbrate administered in high doses to rodents can induce hepatocarcinogenesis, which is associated with oxidative damage to mitochondrial DNA and proteins. However, no association with cancer has been reported for humans taking hypolipidemic drugs. Several oxidative stress mechanisms have been suggested [93]. The ligase also generates H2 O2 , and the H2 O2 could contribute to the lipid peroxidation and mitochondrial toxicity that is observed in rodent hepatocyte studies. Studies using cultured hepatocytes (uninduced) also showed that another ROS source was mitochondrial, caused by a collapse of the membrane potential due to uncoupling of oxidative phosphorylation at complex II or III [94]. However, earlier perfused liver studies showed that the ROS arose from an inhibition of state 3 mitochondrial respiration by cloﬁbrate or ciproﬁbrate at complex I [95]. Mitochondrial ROS overproduction results in the release of apoptogenic factors, which causes apoptosis [93]. Fenoﬁbrate inhibits state 3 respiration with glutamate/malate but not succinate, revealing complex I inhibition [96,97]. Biguanide Agents (e.g., Metformin), Antidiabetic Drugs This class of drugs includes metformin (N 1 N 1 -dimethylbiguanide) currently in use (Figure 3), and the older phenformin (phenethylbiguanide) and buformin (butylbiguanide), which were used for management of hyperglycemia in type 2 diabetes mellitus. Metformin mainly inhibits hepatic glucose release, gluconeogenesis, and β-oxidation of fatty acids. The side effects of these drugs include gastrointestinal symptoms and type A lactic acidosis. Buformin was withdrawn in 1978 because of lactic acidosis. Phenformin was widely used in the 1960s and 1970s, but the lactic acidosis–associated mortality (60 cases per 100,000 patient-years) led to its withdrawal in 1977 in the United States and several other countries, although it is still available elsewhere. Metformin became available in Canada in 1972 and in the United States in 1995, as it was found to cause 20-fold less acidosis than phenformin. Adherence to known exclusion criteria in prescribing

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

95

metformin has minimized the incidence of lactic acidosis [98]. Interestingly, the antidiabetic effect of these biguanides as well as the lactic acidosis have been attributed to the inhibition of the mitochondrial respiratory complex I. Metformin irreversibly decreased hepatocyte or perfused liver respiration with complex I substrates (glutamate, malate) but not complex II substrates (succinate). Metformin also inhibited respiration of complex I substrates incubated with isolated mitochondria or complex I preparations. No metformin metabolism occurred and the investigators proposed a cell complex I signaling mechanism [99]. Furthermore, other investigators found that this inhibition of hepatocyte complex I caused a complete inhibition of gluconeogenesis. Also, the respiration of submitochondrial particles was much more resistant to metformin than were mitochondria or intact cells, suggesting that a slow membrane potential–driven accumulation of metformin across the inner membrane into the mitochondrial matrix is required for inhibiting complex I. With a membrane potential of −180 mV, thermodynamic considerations would predict a 1000-fold accumulation of metformin within the mitochondria. However, the complex I inhibition would be self-limiting, as the accumulating metformin would decrease the mitochondrial potential and prevent further metformin uptake [100]. The mechanism of NADH dehydrogenase (complex I) inhibition is not known but could involve the hydrocarbon part of the biguanide molecule binding the hydrocarbon chain of the membrane phospholipids with the positive charge of the protonated guanidinium group interacting with the phospholipid phosphate group. Thiazolidinediones as Antihyperglycemic Drugs This class includes troglitazone, rosiglitazone (Figure 3), and pioglitazone, which are used to ameliorate hyperglycemia by increasing insulin-stimulated glucose uptake by skeletal muscle. These drugs are known to bind to and activate the nuclear peroxisome proliferation receptor γ (PARPγ), although this may not be the dominant mechanism of action because these drugs also inhibit mitochondrial complex I. As a compensatory response to loss of aerobically generated ATP, glycolytic ﬂux accelerates, leading to lactic acidosis, but this increased glucose utilization lowers serum glucose, which is the desired effect. The efﬁcacy of thiazolidinediones inhibiting complex I or causing lactate release in skeletal muscle or rat liver homogenates is: troglitazone > rosiglitazone > metformin [101]. Troglitazone incubated with HepG2 cells decreased cellular ATP levels and mitochondrial membrane potential [102]. The order of effectiveness for thiazolidinediones (50 µM) opening the mitochondrial permeability transition pore of mouse liver mitochondria was troglitazone > ciglitazone; rosiglitazone and pioglitazone were much less effective. This ranking correlated with their hepatotoxicity [103]. Troglitazone was introduced in 1997 but was withdrawn in 2000 following an U.S. Food and Drug Administion request after 90 severe hepatotoxicity cases had been reported in the 1.9 million patients taking the drug. A black box warning had been issued earlier. Unlike the other thiazolidinediones, troglitazone contains a vitamin E phenolic moiety and is thus a potent antioxidant. However, vitamin E can readily be oxidized to a cytotoxic prooxidant phenoxyl radical metabolite catalyzed by a peroxidase which is probably an intermediate formed

96

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

during CYP3A-catalyzed quinone metabolite formation [104]. Further evidence for this is that multiparameter ﬂow cytometry was used to show that rat hepatoma cells incubated with troglitazone underwent oxidative stress cytotoxicity involving membrane peroxidation, a collapse in mitochondrial membrane potential, and superoxide radical formation. The superoxide radical formation was prevented through the use of cyclosporin, a mitochondrial permeability transition inhibitor [105]. Whether troglitazone-induced hepatotoxicity was a consequence of mitochondrial damage and oxidative stress [106] and/or hypersensitivity requires more research. Nitroaromatic Agent Flutamide, an Antiandrogen/Antiprostate Cancer Drug The efﬁcacy of this drug is marred by rare occurrences of hepatitis. Hepatocytes exposed to ﬂutamide show ATP depletion, GSH depletion, and GSH oxidation but no lipid peroxidation. CYP3A-inhibited hepatocytes were more resistant, and hepatocytes previously depleted of GSH were much more susceptible. Furthermore, ﬂutamide (50 µM) decreased the respiratory control ratio of isolated mitochondria and markedly inhibited state 3 respiration with both glutamate/malate and succinate. Flutamide (50 µM) markedly inhibited the respiration of isolated rat liver mitochondria at the level of complex I, probably mediated by a metabolite, but complex II or III is also implicated given inhibition using succinate as fuel [107]. It is not known whether ROS formation was involved. b. Complex II Inhibitors (Chemical structures are shown in Figure 6.) X-ray crystallographic studies show the structure of succinate–ubiquinone oxidoreductase (complex II) with oxaloacetate, a classical competitive inhibitor bound to the dicarboxylate site similar to malate or fumarate. Other inhibitors include the competitive inhibitor malonate. Cyclophosphamide, ketoconazole, and hydrazine also inhibit succinate dehydrogenase [108]. c. Complex III Inhibitors (Chemical structures are shown in Figure 6.) X-ray crystallographic studies showed that complex III binds antimycin A, an antibiotic, to a domain of cytochrome b H that inhibits cytochrome bc 1 (complex III) and blocks electron transport from the heme b H center to ubiquinone [78]. This inhibition of mitochondrial electron transport also markedly increased hepatocyte mitochondrial ROS formation [72], which could be prevented with ubiquinone [109]. Complex III ROS formation has been attributed to autoxidation of ubisemiquinone radicals formed by the Rieske iron–sulfur cluster of the cytochrome bc 1 complex when the complex was inhibited by antimycin A. The inhibitor stigmatellin is useful in cell studies to identify the site of ROS formation, as it binds to the ubiquinol oxidation site in the bc 1 complex and prevents ROS formation [110]. This inhibition of mitochondrial electron transport increased cellular NADH, which contributed to ROS formation and cell death by releasing iron from ferritin. This cytotoxicity mechanism is called reductive stress, as added NADH generators increased cytotoxicity, whereas NADH oxidants protected the cell from respiratory inhibitors [111,112].

97

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

Respiratory complex II inhibitors O

O HO

HO

OH

HO

OH

OH O

O

Succinate (respiratory, substrate)

O

O

Malonate (inhibitor)

O

Oxaloacetate (inhibitor)

Respiratory complex III inhibitors H O

O

H

O

O

H H

H HN

O

O

O

H

O

O

O

H

H H

OH

OH

N

O

O

HN O

O

F

F

F

O

F

F

F O

F

F

Acetaminophen quinoneimine

F

F F

O

Stigmatellin

Antimycin A

Cl

O

F

Isoflurane

Sevoflurane

Respiratory complex IV inhibitors

Respiratory complex V inhibitor HO O OH

N

H

−O

O

OH H

O O

+ N

N S

Tamoxifen

OH

O

H O

S

NH

Cephaloridine

Figure 6 Other 7 inhibitors.

O O

O O OH

H

Oligomycin

98

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

Drugs inhibiting mitochondrial DNA synthesis that decrease complexes O

NH2

OH

O HN HN

N O

N

O

H N

N

O

O O

− + N N

N OH

N

Zidovudine

HO

N

O N

O

N O

Stavudine

HO

Didanosine

Zalcitabine

Drugs uncoupling oxidative phosphorylation (complex V) HN

NH2

O O

O

HO

O Cl

S HN O

N

O

HN Cl

HN

O O

Cl

O NO2

NH2

Pentamidine

Diclofenac F

F

OH

Nimesulide

Indomethacin

F O 2N

N HO HN

O

Bupivacaine

O

H N

OH

O

Fluoxetine

Propofol

N

H N OH

OH OH

O2N OH

Entacapone

O

β-thujaplicin

O

Tropolone

Figure 6 (Continued)

O

HO

Tolcapone

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

99

Acetyl-p-Aminophenol Agent Acetaminophen, an Analgesic and Antipyretic Although safe at therapeutic doses, accidental acetaminophen overdose is the most common cause of acute drug-induced liver failure in the United States. Acetaminophen is metabolized in the mouse liver by CYP450 to the reactive metabolite, N -acetyl-p-benzouquinone imine (NAPQI). NAPQI reacts with sulfhydryl groups of glutathione and as a result, depletes the primary antioxidant defense in hepatocytes. Subsequently, it forms covalent adducts with intracellular proteins [9,113]. These events lead to ATP depletion, the onset of the MPT, and mitochondrial oxidative stress via formation of reactive oxygen and peroxynitrite [113]. Mitochondria are therefore the primary target of the reactive metabolite, and they play a central role in the mechanism by which acetaminophen induces liver injury and parenchymal cell death.The mitochondrial target for acetaminophen quinoneimine is complex III [114]. Halogenated Ether Agents, Isoﬂurane and Sevoﬂurane, Local Anesthetics A brief exposure to the volatile anesthetic isoﬂurane, termed preconditioning, induces ischemic tolerance in rat brain or heart and improves long-term neurological outcome after brain or heart ischemia. The mechanisms for this neuroor cardioprotection are unknown, but ROS scavengers abolish preconditioning. Furthermore, isoﬂurane induces mitochondrial ROS formation as a result of inhibiting respiratory complex III, probably via ubisemiquinone autoxidation, because it was prevented by myxothiazol, a complex III inhibitor, but not by diphenyleneiodonium, a complex I inhibitor [115]. The cardioprotection is probably mediated by ROS signal transduction. Anesthetic conditioning by sevoﬂurane also involves mitochondrial ROS [116]. d. Complex IV (Cytochrome Oxidase) Inhibitors (Chemical structures are ˚ resolution x-ray structure of the O2 reduction shown in Figure 6.) The 1.9-A site of bovine heart cytochrome c oxidase in the fully reduced state indicates trigonal planar coordination of CuB by three histidine residues. One of the three histidine residues has a covalent link to a tyrosine residue to ensure retention of the tyrosine at the O2 reduction site. These moieties facilitate a four-electron reduction of O2 and prevent formation of active oxygen species. Reduction of the oxidase causes deprotonation of Asp51 [117]. Cytochrome oxidase inhibitors such as cyanide, azide, and hydrogen sulﬁde complex the heme and Cu moieties of oxidized cytochrome oxidase. Cephalosporin Agent, Cephaloridine, an Antibiotic Cephaloridine induces nephrotoxicity and causes acute renal failure in humans and animals. It is characterized by acute proximal tubular necrosis lesions, particularly in the S2 segment of the tubules, where the drug is transported from the blood to the proximal tubular cell by the organic anion transporter. This accumulation in proximal tubular epithelial cells inactivated mitochondrial cytochrome oxidase by an unknown mechanism, and this inhibition has been proposed as the nephrotoxic mechanism associated with this drug [118]. Cytochrome oxidase

100

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

inhibition was also shown for cefazolin and cefalotin, two other cephalosporin antibiotics. The lipid peroxidation induced probably resulted from ROS formation formed by the inhibition of mitochondrial electron transport [119]. Pretreatment with rottlerin, an inhibitor of mitochondrial protein kinase C, prevented the early translocation of protein kinase into the mitochondria and prevented cephaloridine-induced renal dysfunction [120]. Cephaloridine also acylates and inactivates the transporter that transports succinate into the mitochondria. Furthermore, cephaloridine caused renal cortex GSH oxidation and inactivated GSH reductase [121]. 4.2. Drugs That Inhibit Mitochondrial DNA Synthesis (Chemical structures are shown in Figure 6.) Nucleoside Reverse Transcriptase Inhibitors AZT (Zidovudidine), Stavudine (d4T), Didanoside (ddI), and Zalcitabine (ddC), Anti-HIV Drugs These drugs require endogenous phosphorylation to be effective as antiretroviral drugs. However, as a class, and in particular AZT, they are also potent inhibitors of mitochondrial DNA synthesis by inhibiting DNA polymerase γ and mtDNA replication. Impaired synthesis of mtDNA-encoded respiratory chain polypeptides in cultured human muscle cells decreases respiration, resulting in increased lipid droplet accumulation and lactate production [122]. This would slow electron transport, resulting in ROS formation. Secondarily, ROS would also inhibit aconitase and α-ketoglutarate dehydrogenase of the Krebs cycle, whereas ATP depletion would inhibit fatty acid oxidation, causing stenosis leading to NASH and lactic acidosis. Mitochondrial oxidative stress probably underlies the deleterious side effects of skeletal myopathy, cardiomyopathy, pancreatitis, bone marrow suppression, and peripheral neuropathy. Dilated cardiomyopathy is probably caused by mitochondrial ROS formation, oxidation of mtDNA and GSH, lipid peroxidation, uncoupling of OXPHOS, and mitochondrial dysfunction. AZT, ddI, and ddC were similar in their effectiveness at inhibiting myogenic cell proliferation and mitochondrial toxicity. However, the in vivo plasma concentration of AZT was 20-fold higher than that of zalcitabinegiven, which surely contributed to the much higher muscle mitochondrial toxicity observed for AZT. Rats treated with AZT i.p. lost weight and had a 100-fold increase in serum creatine kinase and increased lactate with the high-dose AZT. Mitochondria isolated from heart and skeletal muscle showed decreased respiratory control ratios [123]. Rat heart mitochondrial ROC production was moderated by dietary supplementation with the antioxidant vitamin E or C [124]. Furthermore, the intracellular concentrations of these drugs are increased by the protease inhibitor ritonavir (part of HAART therapy) because it potently inhibits P-gp-mediated extrusion of the antiretrovirals. Interestingly, the CD4 T-cell depletion in AIDS resulting from HIV infection has also been attributed to mitochondrial-dependent cell death [26] (see also Chapters 9 and 21).

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

101

4.3. Drugs That Uncouple Mitochondrial Oxidative Phosphorylation (Complex V) (Chemical structures are shown in Figure 6.) These drugs are weak acids, and the weak acid group interacts with the phospholipids in the target membrane, making the membrane permeable to protons [i.e., they uncouple oxidative phosphorylation by their protonophoric properties (e.g., phenols, salicylic acids, diphenylamine NSAIDs, and aromatic amine local anesthetics)]. The most potent uncouplers have an acid-dissociable group, an electron-withdrawing moiety, and a bulky hydrophobic group. The unprotonated and protonated drug can cycle back and forth across the inner mitochondrial membrane, causing the reentry of protons into the mitochondrial matrix. This is measured using isolated mitochondria by determining the minimal dose required to maximize mitochondrial state 4 respiration when ADP is completely phosphorylated. a. Cationic Drug Uncouplers Lipophilic cationic drugs (i.e., those containing a positive charge) are accumulated in the mitochondria, as they readily pass through the inner membrane driven by its membrane potential. This accumulation in the matrix often causes a type of uncoupling that requires inorganic phosphate. Pentamidine is used in the therapy and prophylaxis of African trypanosomiasis and leishmaniasis and in treating Pneumocystis infections in AIDS patients. At 200 µM, pentamidine uncouples isolated mitochondria, causes Ca2+ efﬂux, inhibits respiratory control, and increases latent ATPase. Mitochondrial respiration is not inhibited. Pentamidine accumulates in the mitochondria through an electrophoretic mechanism which partially collapses the inner membrane potential, causing permeabilization and releasing Ca2+ [125]. The uncoupling order of effectiveness for chlorpromazine and amine local anesthetics were chlorpromazine quinine, dibucaine > quinidine > butacaine > propanolol, tetracaine. It was concluded that they were not protonophores, but uncoupled as a result of the protonated cationic amine forming a lipophilic ion pair with an appropriate anion which underwent transmembrane cycling [126]. Other cationic drugs include diltiazem, imipramine, and the cationic amphiphile drugs (described previously). b. Nonsteroidal Anti-inﬂammatory Drugs (NSAIDs) NSAIDs need a weak acid group (carboxyl) in order to inhibit cyclooxygenase via binding to the arachidonate binding site. Unfortunately, this acid group and their lipid soluble nature also allows NSAIDs to interact with the enterocyte mitochondrial inner membrane phospholipids, acting as a protonophore, thereby depleting ATP and causing cytotoxicity. Enterocyte cell death undermines intestinal impermeability, which elicits a low-grade inﬂammatory response. The lack of membrane-protective prostaglandins also contributes to ulceration of the small bowel. Such ulceration could therefore be induced in 20 hours with a low dose of aspirin followed an hour later by DNP, an uncoupler [127]. One study examined the respiratory

102

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

control ratio of isolated mitochondria following treatment with NSAIDs containing a diphenylamine moiety. NSAID potency in increasing state 4 and inhibiting state 3 respiration was: ﬂufenamic acid, diﬂunisal > tolfenamic acid > mefenamic acid > diclofenac > indomethacin > benoxaprofen, naproxen, fenoprofen [128]. Diphenylamine Moiety Containing Agents, Diclofenac, an NSAID Drug Diclofenac taken daily can also cause rare but signiﬁcant hepatotoxicity after more than 1 to 3 months of administration. This has been attributed to uptake via hepatocyte mitochondrial anion carriers, resulting in uncoupling and induction of the MPT, causing cell death [129]. However, mitochondrial ROS formation caused by an inhibition of mitochondrial respiration at lower diclofenac concentrations may be more important than mitochondrial uncoupling. ROS could, however, also be formed by inhibition of mitochondria respiration by the diphenylamine radical or the quinoneimine metabolite formed by P450 catalysis. ROS scavengers prevented diclofenac-induced MPT induction and caspase cascade induction in hepatocytes [45]. A similar ROS-mediated mechanism was shown in HL-60 cells incubated with diclofenac, where ROS suppressed Akt. This activated caspase 8 and Bid cleavage, cytochrome c release, and caspase 9 and 3 activation. This sequence of events was prevented by ROS scavengers [130]. Nitroaromatic Agent Nimesulide, an NSAID Cox2 Inhibitor Nimesulide is a sulfoanilide NSAID containing a nitroaromatic moiety. It has less GI toxicity than other NSAIDs, but it can cause rare but serious hepatic dysfunction and injury weeks and months after exposure. Hepatocyte and mitochondrial studies with nimesulide revealed uncoupling and excessive NAD(P)H oxidation. At low micromolar concentrations, nimesulide induced a sudden increase in the permeability of mitochondria, which led to a collapse of the mitochondrial potential. Whether oxidative stress is involved remains unclear. The MPT induction caused uncoupling and matrix expansion, thereby releasing intermembrane proapoptotic factors, matrix solutes, and antioxidants such as GSH and CoA. The MPT opening caused ROS formation, which probably further promoted MPT opening and shifted the mitochondrial redox state to a more prooxidant state. Another source of ROS could be the redox cycling of the nitroanion radical and hydroxylamine metabolites [131]. However, no evidence of the nitroanion radical was obtained when nimesulide was reduced anaerobically with rat liver microsomes and NADPH. Mitochondrial toxicity and increased ROS levels also contributed to human hepatoma cell death induced by nimesulide [132]. Diphenylamine-Containing Agent Indomethacin, an NSAID Drug Indomethacin contains a diphenylamine moiety and can cause toxicity to the GI tract and the kidneys. It inhibits cyclooxygenase and therefore slows the formation of protective prostaglandins, but it also accelerates mitochondrial ROS formation and elicits neutrophil inﬁltrating, thereby exacerbating oxidative stress. Kidney mitochondrial phosphatidic acid and cholesterol were also increased [133].

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

103

c. Antiestrogens (Dimethylaminoethoxyphenyl)diphenylbutene Agent Tamoxifen, a Breast Cancer Chemotherapy Drug Tamoxifen is currently the most widely used chemotherapeutic agent of breast cancer. Formation of the major phenolic metabolite, 4-hydroxytamoxifen, is catalyzed by P450, and it has a two-order-of-magnitude higher afﬁnity for the estrogen receptor (ER). However, it is not clear whether the metabolite contributes to the chemothereupic efﬁcacy activity because tumor cells lacking ERs were also susceptible to tamoxifen. Added to mitochondria at 10 nmol/mg protein, tamoxifen undermined the respiratory control and the ADP/O ratios in a dose dependent manner. The mitochondrial membrane potential was also decreased, and state 4 respiration and ATPase activity increased. At a higher concentration (> 40 nmol/mg protein), state 3 respiration was inhibited [134]. Tamoxifen was more effective than 4-hydroxytamoxifen at depleting mitochondrial ATP levels or inhibiting the adenine nucleotide translocase and phosphate carrier, underscoring the fact that mitochondrial effects are independent of ER [135]. Tamoxifen has also been shown to inhibit complexes III, IV, and V [136], so ROS formation is likely. Mechanistic cytotoxicity studies using intact cells are needed to understand how mitochondrial toxicity and/or ROS formation contributes to the pathways resulting in apoptosis and necrosis. Tamoxifen may also induce fatty liver and NASH in breast cancer patients taking tamoxifen (see Section 2.2a). d. Dimethylphenylacetamide Agents Bupivacaine and Etidocaine, Local Anesthetics Severe cardiotoxicity and myotoxicity of the potent long-lasting anesthetics bupivacaine (Figure 3) and etidocaine limit their utility and have been attributed to their mitochondrial uncoupling activity. As highly lipophilic amphiphilic amines they can shuttle protons across mitochondrial membranes in a true protonophoretic mechanism [137]. Bupivacaine uncoupled heart cell mitochondria more effectively than did ropivacaine, and both inhibited complex I, whereas lidocaine was a much less potent uncoupler [138]. Such mitochondrial impairment has been poposed to underlie bupivacaine-induced myopathies [139]. Ropivacaine is a promising replacement anesthetic, being less lipophilic and having minimal effects on mitochondrial function [140]. e. Tricyclic Agent Fluoxetine, an Antidepressant Drug Fluoxetine (Prozac) is a widely used antidepressant. Its demethylated metabolite, norﬂuoxetine, inhibits neuronal serotonin reuptake. Although it has a high therapeutic index, some cardiovascular and extrapyramidal side effects, plus drug–drug interactions, have been reported. These may be due to drug interference with the lipid bilayer of the inner mitochondrial membrane, particularly at high doses. Addition of ﬂuoxetine or norﬂuoxetine to isolated rat brain mitochondria uncoupled OXPHOS and inhibited F1 F0 ATPase activity with an IC50 value of 80 µM [141]. f. Alkylphenol Agent Propofol, an Anesthetic Propofol (2,6-diisopropylphenol) (Figure 4) is an intravenous anaesthetic that interacts with the GABA receptor.

104

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

However, prolonged infusions of propofol can induce metabolic lactic acidosis in patients, especially children, and several cases of hepatocellular damage have been reported. Lactic acidosis could indicate mitochondrial toxicity, and this was veriﬁed when propofol, perfused through the liver at 25 µM, increased oxygen uptake and glycolysis but decreased glucose synthesis [142]. Impairment of mitochondrial energy metabolism was attributed to its mild protonophore activity, and it dissipated membrane potential when added to isolated mitochondria, Another adverse effect of propofol is the suppression of macrophage function, another effect attributed to mitochondrial toxicity [143]. Other alkyl phenols are also mild protonophores when added to isolated mitochondria, but their effectiveness as uncouplers was limited by the low fraction of phenol dissociated at near-physiological pH. g. Nitrocatechol Agents Tolcapone and Entacapone, COMT Inhibitor Drugs for Parkinson’s Therapy The nitrocatechol drugs tolcapone and entacapone (Figure 4) are catechol-O-methyltransferase inhibitors used as adjunct therapies to forestall the metabolism of levodopa in Parkinson’s disease patients. The nitrophenol drugs enhanced plasma levodopa levels, which improved the symptoms and extended the overall quality of life of Parkinson’s patients. Tolcapone was introduced late in 1997, but within six months of use, three patients (from 40,000 patient-years) died from fulminant hepatic failure. Because of this, tolcapone was withdrawn in the European Union and Canada late in 1998, whereas in the United States the FDA issued restrictive liver enzyme monitoring measures which severely limited its use [144]. Previously, it had been shown that tolcapone at 10 µM uncoupled isolated mitochondria to the same extent as 50 µM 2,4-dinitrophenol or 200 µM entacapone [145]. Later, the same laboratory showed that rat rectal body temperature was increased following the administration of tolcapone (50 mg/kg), but not entacapone (400 mg/kg), with or without levodopa and carbidopa. This was attributed to mitochondrial uncoupling of OXPHOS [146]. h. Nitrophenol Agent, Dinitrophenol, a withdrawn Diet Drug 2,4-Dinitrophenol (DNP) was used in diet pills in the 1930s, sold under several trade names, including Caswell No. 392, Sulfo Black B, and Nitro Kleenup. However, deaths from hyperthermia resulted in its ban in the United States under the 1938 Federal Food, Drug, and Cosmetic Act. Before this, drugmakers were not required to prove that their products were safe before marketing. DNP is a classical protonophoretic OXPHOS uncoupler and causes rapid and dramatic weight loss. Today, DNP is used by bodybuilders, often illegally, to lose body fat rapidly before contests. However, the body has no negative feedback system, so the upper limit of body temperature with overdose is likely to be lethal. 1. Aminophenol and Chlorophenol Drug Metabolites. Chlorophenols (Figure 4) are widely used as bactericides, fungicides, and herbicides, and pentachlorophenol is a major wood preservative. Halophenols are also major metabolites of halobenzene-containing agents and probably contribute to the

105

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

mitochondrial toxicity of these agents. The mitochondrial toxicity of diclofenac and amiodarone have already been described. However, 4 -hydroxydiclofenac formed by a CYP2C-catalyzed ring hydroxylation of diclofenac is a toxic reactive quinoneimine that probably also contributes to the mitochondrial toxicity of diclofenac [130,147,148]. The phenolic metabolites formed by O-dealkylation of the benzofuran drug amiodarone could also contribute to the mitochondrial toxicity of amiodarone [31–35,130,147,149]. 2. Tropolone Agent β-Thujaplicin, an Antiviral, Antitumor, Antifungal Drug. Tropolones are constituents of woody essential oil, of considerable pharmaceutical signiﬁcance, as they have powerful antibacterial, antiviral, antitumor, and antifungal properties. β-Thujaplicin (4-isopropyl tropolone) at 1 mM is toxic to hepatocytes, where it depleted ATP and GSH. No lipid peroxidation was induced. Isolated rat liver mitochondria were uncoupled by thujaplicin, although state 3 respiration was also substantially repressed. The rank order of potency of tropolones for inhibiting state 3 respiration was thujaplicin > tropolone > tropone [150]. 4.4. Mitochondrial Oxidative Stress Induced by Drugs Independent of Respiratory Inhibition (Chemical structures are shown in Figure 7.) a. Redox-Cycling Oxygen Activation Quinone Agents, Doxorubicin (Adriamycin) and Mitomycin C, Anticancer Drugs Doxorubicin was introduced in the 1970s as a potent anticancer agent to treat a broad range of malignancies. It is a planar anthracycline antibiotic that intercalates into the DNA double helix and arrests cell proliferation. However, it was NH2 HO NH2 O

O

OH

O

OH

O

O

O

NH

O HO HO

N

Doxorubicin

Isoniazid O

NH

O HO

O F

H2N O

HO

HO N

OH

NH2

Gentamicin

Figure 7

N

H

H

O

HN

N

F

O

NH2

F

Trovafloxacin

Drugs that cause mitochondrial oxidative stress.

106

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

soon discovered that therapy is limited because of the increased risk of cardiotoxicity at higher doses. This is due to redox cycling of the drug and consequent loss of mitochondrial function in this particularly aerobically poised tissue (see Chapter 6). Doxorubicin added to isolated mitochondria increased state 4 oxygen uptake and inhibited state 3 respiration. Mitochondrial dysfunction is caused by redox cycling of semiquinone formed via reduction by the outer membrane NADH–b 5 reductase and inner membrane complex I, which reacts with oxygen to form superoxide. The oxygen consumed by redox cycling impairs oxidative phosphorylation by diverting electrons from the respiratory chain, eliciting a compensatory increase in glycolysis [151]. Doxorubicin also readily forms a stable complex with Fe3+ (stability constant of 1033 ), which is reduced to a Fe2+ complex by semiquinone radicals. The ferrous complex then reacts with hydrogen peroxide to form extremely reactive hydroxyl radicals. Indeed, doxorubicin cytotoxicity and mitochondrial dysfunction are prevented by dexrazoxane, a ferric chelator [152], and by antioxidants such as cardevilol [153] or the overexpression of antioxidant enzymes [154,155]. ROS are also probably responsible for inhibition of the adenine nucleotide translocase found in mitochondria isolated from rats treated for 4 to 8 weeks with therapeutic levels of doxorubicin. The inhibited translocase alters the calcium regulation of the permeability transition pore and decreases the mitochondrial calcium loading capacity or calcium signaling pathways [156]. Another factor contributing to mitochondrial dysfunction is that doxorubicin is positively charged and forms a strong complex with cardiolipin, an acidic phospholipid indigenous to the inner mitochondrial membrane [157]. In vivo, doxorubicin decreased complex I activity and induced ultrastructural mitochondrial injury in the heart, but this did not occur in transgenic mice expressing twofold more of the mitochondrial form of superoxide dismutase (MnSOD), suggesting that superoxide radicals caused the inhibition of complex I [155]. Transgenic mice overexpressing catalase were also resistant [154]. Mitomycin C and porﬁromycin are pro-drugs that are reduced in the hypoxic region of solid tumors by mitochondrial outer membrane NADH–cytochrome b 5 reductase. The hydroquinone electrophile formed cross-links genomic DNA and mono- and dialkylates DNA, thereby preventing tumor cell proliferation. However, this anticancer therapy is limited because of toxicity to lung epithelial cells, resulting in pneumonitis that can progress to interstitial lung ﬁbrosis. Under these aerobic conditions, mitomycin C caused mitochondrial dysfunction when its semiquinone intermediates redox cycle to form ROS. These oxidizes biomolecules such as DNA, GSH, and dithiols of the mitochondrial permeability transition pore opening, thereby initiating apoptotosis [158,159]. b. Drug Radical–Mediated Oxygen Activation Hydrazine Agents (e.g., Isoniazid), Antituberculosis Drugs Hydrazine is a reactive metabolite of isoniazid that causes idiosyncratic hepatotoxicity in some patients. Hydrazine autoxidizes and causes ROS formation. Hydrazine also induces a phosphate-dependent transitory mitochondrial uncoupling that yields

DRUG-INDUCED OXIDATIVE STRESS AND TISSUE TOXICITY

107

inhibition of state 3 respiration [160]. Hepatocyte cytotoxicity studies showed that hydrazine readily caused ATP depletion and decreased the mitochondrial membrane potential. Succinate dehydrogenase (complex II) was also inhibited and could explain the lactic acidosis. The Krebs cycle was also inhibited as a result of α-ketoglutarate depletion by hydrazone formation with hydrazine. Oxidative stress was prominent as ROS formation and lipid peroxidation were increased and GSH was oxidized (including mitochondrial GSH). Hepatocyte catalase was inhibited, whereas GSH reductase or GSH reductase activity was not affected. Hydrazine toxicity to hepatocytes was also increased when catalase is inhibited or GSH is depleted [161]. Hepatocyte protein synthesis or liver protein synthesis in vivo were particularly sensitive to inhibition by hydrazine, probably because of ATP depletion. ATP depletion could also explain the triglyceride accumulation, inhibition of the urea cycle, and depletion of NADPH noted with this drug [108]. c. Iron/Polyunsaturated Lipid–Mediated Oxygen Activation Aminoglycoside Agents (e.g., Gentamicin), Antibiotics Gentamicin-induced nephrotoxicity accounts for 10 to 15% of all cases of acute renal failure. It also damages sensory cells of the inner ear, causing ototoxicity. It is believed to induce nephrotoxicity by releasing iron from renal cortical mitochondria and by inducing mitochondrial hydrogen peroxide and ROS formation. Gentamicin nephrotoxicity was also prevented by the antioxidants vitamin E and selenium [162]. Kidney cells incubated with gentamicin accumulate it in lysosomes, slowly permeabilizing them. This then causes a loss of mitochondrial membrane potential, the release of cytochrome c, and the activation of apoptosis. It was not clear whether the mitochondrial apoptosis pathway was due to a direct effect of gentamicin [163]. How gentamicin causes ROS formation is also not clear. Recently, a ternary complex of gentamicin with iron and polyunsaturated lipids was found to react with oxygen to form ROS [164]. d. Fluoroquinolone Agent Trovaﬂoxacin, an Antibiotic Drug Trovaﬂoxacin is a new broad-spectrum antibiotic. However, in a relatively small number of patients it induces hepatitis involving centrilobular necrosis. A gene expression analysis of human hepatocytes treated with six different quinolone agents (i.e., trovaﬂoxacin, grepaﬂoxacin, ciproﬂoxacin, clinaﬂoxacin, gatiﬂoxacin, levoﬂoxacin) showed that trovafoxacin induced far more gene expression changes than the others. The genes expressed included a number of mitochondrial genes that were not altered by the other quinolones. Mitochondrial genes down-regulated by trovaﬂoxacin include ribosome proteins, mitofusin-1, bax , and the oxidative stress genes heme oxygenase and thioredoxin reductase 1. When trovaﬂoxacin was incubated with HepG2 cells, total hepatocyte GSH was decreased and ROS formation increased. The order of quinolone potency in depleting GSH was trovaﬂoxacin grepaﬂoxacin ciproﬂoxacin > gatiﬂoxacin > clinaﬂoxacin > levoﬂoxacin [165]. However, no mitochondrial studies have been published, so which proteins are targeted by trovaﬂoxacin or whether respiration is impaired remain unknown.

108

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

e. Drug Radical–Induced Megamitochondria Formation Megamitochondria are deﬁned as mitochondria that are more than three times larger in diameter than normal (see also Chapter 23). Megamitochondria develop in cells continuously exposed for 22 to 24 hours to oxidative stressors, including H2 O2 , ethanol, hydrazine, chloramphenicol, methylglyoxal bis(guanylhydrazone), cuprizone, or iron [166]. Troglitazone incubated with a human hepatocyte cell line induced H2 O2 formation at 1 hour, and at 15 hours megamitochondria with a lower membrane potential were apparent [167]. Hepatocyte megamitochondria were also induced in vivo in rodents administered ethanol or hydrazine, and they are also increased in patients with NASH [168]. In this phenomenon, mitochondria fuse to form megamitochondria, probably as an adaptive change because they form less ROS and maintain ATP production in this state. Megamitochondria formation is prevented, or the megamitochondria return to normal, if free-radical exposure decreases or if ROS scavengers such as tocopherols, coenzyme Q, and 4-OH-TEMPO are added. However, additional oxidative stress results in megamitochondria swelling, loss of membrane potential, cytochrome c release, caspase activation, and apoptosis [166]. f. Halohydrocarbon Agents Chloroform and Carbon Tetrachloride, Toxic Xenobiotics Chloroform was used as an anesthetic from the mid-1800s to 1900, but it had toxic side effects, including hepatotoxicity, and is now considered a carcinogen. Chloroform-induced hepatotoxicity involves reductive activation to form trichloromethyl radical anions, which form covalent adducts with unsaturated fatty acids, and/or oxidative activation catalyzed by P450 to form phosgene. The latter also bound covalently to two unsaturated fatty acids of phosphatidylethanolamine in the mitochondria. These adducts were associated with mitochondrial swelling and megamitochondria formation.The megamitochondra formed often contained membranous whorls in the matrix [169]. Carbon tetrachloride was used for a brief period as an anesthetic in the 1950s until its potent hepatotoxicity was realized. The hepatotoxic mechanism was shown to involve reductive metabolism to form trichloromethyl radicals, which form adducts with unsaturated fatty acids or react with oxygen to form peroxyl radicals. These radicals initiate lipid peroxidation, and antioxidants can prevent hepatocyte cytotoxicity. Carbon tetrachloride induces fatty liver in rats, and mitochondria isolated from rats exposed to it showed decreased respiratory control, inhibited ATP synthesis, and increased phosphate levels [170]. However, it is not clear whether mitochondrial toxicity contributed to carbon tetrachloride hepatotoxicity, or whether the hepatotoxicity caused mitochondrial damage. There have been no reports of mitochondrial damage in hepatocyte studies or of ROS formation before cytotoxicity was apparent, so it is not known if the mitochondria are susceptible to the radical metabolites of carbon tetrachloride. g. Nitroaromatic Drug Radicals Flutamide, nimeslide, nilutamide, tolcapone, and nitrofurantoin (described above) all have a nitroaromatic moiety and have all

STRUCTURE–ACTIVITY RELATIONSHIPS

109

been associated with rare cases of idiosyncratic liver injury. As described above, these drugs are also mitochondrial toxins and have the potential for oxidative stress, as their nitro group can be reduced or catalyzed, by reductases to form reactive nitroso and N -hydroxy derivatives that can oxidize biomolecules and form ROS. It is not yet known whether these drugs induce megamitochondria. 4.5. Prevention and Therapy for Drug-Induced Oxidative Stress Antioxidant therapy designed to moderate drug-induced toxicity includes the use of iron chelation as well as natural antioxidants or derivatives such as vitamins E and C or trolox. Superoxide dismutase and catalase mimics can also be used to scavenge superoxide anions, hydrogen peroxide, and peroxynitrite to moderate oxidative stress [58]. Recently, mitochondrial targeted antioxidants conjugated to lipophilic cations have become available [171] (see Chapter 26). Carbonyl traps such as metformin and pyridoxamine are also useful for preventing oxidative stress in type 2 diabetes. 5. STRUCTURE–ACTIVITY RELATIONSHIPS 5.1. Mitochondrial Toxic Drugs Structure–activity relationships attempt to correlate biological activity with structural features of molecules. This method is based on the general premise that the molecular properties characteristic of all active compounds must in some way be essential for activity. The basic goal of quantitative structure–activity relationship (QSAR) studies is to explain the observed variation in biological activities of a series of compounds in terms of variations in the chemical structures. Positive associations within the data set can foster predictive extrapolations. Examples of the physiochemical variables used are the acid dissociation constant (pK a ), lipophilicity (log P or c log P; π), inﬂuence of substituents on charge distribution (σ), steric effects of ortho moities (Taft–Kutter–Hansch constant), and indices of molecular excitability, such as HOMO–LUMO (highest occupied molecular orbital–lowest unoccupried molecular orbital). a. Nitrophenols All nitrophenols uncouple oxygen consumption from ATP synthesis and are associated with pK a values in the range 3.8 to 8.5. The decreasing order of toxicity to isolated rat liver mitochondria, by either uncoupling or inhibiting respiration, is 2,4-(NO2 )2 > 2,5-(NO2 )2 > 2,6-(NO2 )2 > 4-NO2 > 3-NO2 > 2-NO2 and is related to their log P (π) and σ, whereas pK a was a poor measure [172]. Although dinitrophenol cytotoxicity can be attributed to its protonophoric uncoupling activity, QSARs for the cytotoxicity of mononitrophenols involve log P plus bond dissociation energies or E LUMO [173–175], which suggests that metabolism to cytotoxic radicals and/or electrophiles also contributes to their cytotoxicity.

110

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

b. Chlorophenols and Drug Metabolites The potency of chlorophenols at inhibiting ATP production and succinate-induced reversed electron ﬂow in beef heart submitochondrial particles was determined. The decreasing order of toxicity was pentachlorophenol > 2,3,5,6-Cl4 > 2,3,5-Cl3 > 2,3,4-Cl3 > 2,4,5-Cl3 > 2,4,6Cl3 > 2,3,6-Cl3 > 3,5Cl2 > 3,4-Cl2 > 2,3-Cl2 > 2,4-Cl2 > 2,6-Cl2 > 4-Cl > 2-Cl [177]. QSAR analysis showed that log P was the best descriptor, indicating that chlorophenols need to partition into the lipid bilayer of the mitochondrial membrane to cause toxicity rather than bind to the inner surface of the inner membrane. The electronic parameter σ (the sum of substituent σ values) was also important and reﬂects the inﬂuence of substituents on charge distribution within the molecule that increase stability of the phenolate anion and so make it a more efﬁcient H+ exchanger. The acid dissociation constant (pK a ) was a poor descriptor, which was surprising since σ and pK a are usually correlated. Analysis via a correlation matrix between the various parameters indicate that log K ow and σ are highly correlated (r 2 = 0.95), making it difﬁcult to distinguish the importance of hydrophobic versus electronic effects. The second parameter, pK a , was a poor descriptor. Ortho-substituted phenols were the most acidic and also the least toxic, probably because localization of the phenolate charge lowers lipophilicity. These results are consistent with the molecular uncoupling mechanism based on the chemiosmotic theory and on the protonophoric properties of chlorophenols [176]. However QSARs for the cytotoxicity of halophenols suggest that cytotoxic prooxidant phenoxyl radicals, and/or quinone electrophiles, also contribute to the cytotoxic mechanism [173–175]. c. Alkylphenols The relative order of alkyl phenol uncoupling activity toward isolated rat liver mitochondria was 4-t-Pent > 4t-Bu > 2t- > 4n-Pr > 4-Et > 4Me > H. The QSAR equation involved log P , pK , and a Taft–Kutter–Hansch steric constant for ortho substituents [177]. QSARs for the cytotoxicity of alkyl phenols to hepatocytes or other cells [173–175], suggest that cytotoxic prooxidant phenoxyl radicals and/or quinone methide electrophiles contribute to their cytotoxic mechanisms. d. NSAIDs Although the order of NSAID inhibition of mitochondrial respiration is not available, the order of NSAID uncoupling effectiveness found was ﬂufenamic acid > diﬂunisal > tolfenamic acid > mefenamic acid > diclofenac > indomethacin > naproxen, fenoprofen > salicylic acid [128]. Although a structure–activity analysis was not performed by the authors, a signiﬁcant correlation (r 2 = 0.92) could be found when the NSAID uncoupling activity was related to their respective log P values. If naproxen and fenoprofen (propionic acid derivatives) were excluded, the correlation was even more signiﬁcant (r 2 = 0.98). Recently, we have also performed a structure–toxicity study for 20 NSAIDs toward isolated rat hepatocytes and found that if both propionic acids and benzoic acids (otherwise known as diphenylamine NSAIDs) were analyzed as one group, then cytotoxicity correlated with

STRUCTURE–ACTIVITY RELATIONSHIPS

111

log P . However, if the analysis was carried out on the diphenylamine moiety containing NSAIDs, the cytotoxicity correlated with the HOMO–LUMO gap and the ﬁrst-order molecular connectivity index, whereas the cytotoxicity of the propionic acid NSAIDs was still dependent on log P. These differences suggest a different cytotoxic mechanism for the two NSAIDs [178]. The HOMO–LUMO parameter in the QSAR obtained for the benzoic acid NSAIDs (which are also diphenylamine NSAIDs) suggested that they undergo a redox metabolic activation. A one-electron metabolic oxidation to form a prooxidant diphenylamine N-cation radical could therefore be part of the mitochondrial toxicity mechanism [179]. e. Local Anesthetics Cytochrome c oxidase (complex IV) is also inhibited by local anesthetics, with the following order of potency: quinisocaine > butacaine > pramocaine > bupivacaine > carticaine > lidocaine, procaine > prilocaine. Like their anesthetic activity, respiratory inhibition is also correlated to some extent with their lipophilicity (log P), suggesting that lipophilic interactions are involved in cytochrome oxidase–anesthetic binding [180,181]. 5.2. Mitochondrial/Lysosomal Accumulation by Cationic Amphiphile Drugs Understanding the physicochemical features of cationic amphiphile drugs (CADs) can provide insight into how they induce phospholipidosis or alter cellular functions. Many efforts have been made to identify the principles of drug–phospholipid interaction on the molecular level. However, a drawback to applying CAD therapeutics as test compounds is their heterogeneity in chemical structure and physicochemical properties. Only a few structure–activity studies for CADs have been reported to date [183]. An early study by Bandyopadhyay et al. [183] determined that the induction of multilamellar inclusions were produced by drugs with high partition coefﬁcients, whereas those with low partition coefﬁcients did not. It has been observed that drugs with low partition coefﬁcients diffuse passively inside the cells and failed to induce multilamellar inclusions because of their inherent limitation of reaching high enough concentrations inside the hepatocytes. Drug accumulation is therefore important to the effect of CADs, and penetration probably depends on aspects of the drug such as the pK a of the amine groups and hydrophobicity [184]. Ploemen et al. [185] used physicochemical calculations to study the molecular properties of compounds suspected of inducing phospholipidosis. Five cationic amphiphilic drugs (chlorpromazine, amiodarone, imipramine, propranolol, and ﬂuoxetine) were studied and compared to a set of structurally related compounds (gepirone, 1-phenylpiperzine, its major metabolites, 3-OH-gepirone, and 1-pyrimidinylpiperzine, and buspirone). The order of cytotoxicity found for the the drug set was chlorpromazine > ﬂuoxetine > amiodarone > imipramine > propanolol, but the degree of lysosomal phospholipase A2 inhibition was not measured. The second set of compounds were not CADs; however, gepirone given

112

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

to rats in a one-year toxicity study caused some phospholipidosis. Ploemen’s group showed that calculation of the parameters c log P (partition coefﬁcient octanol/water) and pK a can help predict whether a compound may have the potential to induce phospholipidosis. Positive inducers of phospholipidosis had a relatively high c log P (apolar region, hydrophobic), accompanied by a relatively high calculated pK a (polar region, highly ionized amine). Phospholipidosis was conﬁrmed in vitro using a human monoblastoid cell line by detection of lamellar inclusion bodies via electron microscopy. The chemically related series of gepirone compounds did not have prominent amphiphilic cationic properties and were not capable of inducing lamellar inclusion bodies in the in vitro system. The afﬁnity of cationic amphiphilic drugs for phospholipids appears to involve both electrostatic and hydrophobic binding forces. CADs contain in close proximity a lipophilic aromatic ring system and a side chain with a nitrogen that is protonized at physiological pH. Inherently, these drugs are prone to interaction with membrane phospholipids. The cationic nitrogen is attracted to the negatively charged phosphate of the phospholipid headgroup, and the aromatic ring system is directed toward the hydrophobic interior of the phospholipid layer. However, the cationic amphiphilic nature may have an impact on drug pharmacokinetics and pharmacodynamics [182]. Klein et al. [182] tested and characterized a set of cationic amphiphilic model compounds and analyzed their membrane interactions using QSAR models. They tested a series of phenylpropylamine derivatives where modiﬁcations were incorporated at the aromatic part of the molecule. A propyl group was linked between the aromatic ring and the nitrogen, thus leaving the alkaline character relatively independent of the aromatic variations. Three different model systems of biologically activity were investigated to form QSAR models: (1) the inhibition of Ca2+ adsorption to phosphatidylserine monolayers to measure the interaction of compounds with the phospholipid surface charge and monitor drug binding; (2) the inﬂuence in the phase-transition temperature of lipososmes of dipalmitoylphosphatidic acid (DPPA) to assess the perturbing action of the drugs on the structural organization of phospholipid assemblies; and (3) the antiarrhythmic activity of compounds in isolated guinea pig hearts to assess membrane-stabilizing potency. Classical intramolecular and novel intermolecular descriptors were used to build QSARs. The intermolecular modeling QSAR approach was attempted based on molecular dynamic simulations of compounds in a phospholipid environment. The compounds selected were suitable for performing this proposed membrane-interaction QSAR analysis, because there is minimal uncertainty about the orientation of these compounds within the phospholipid membrane–water interphase. The cationic amino group is presumed to be anchored near the phospholipid head group, while the aromatic hydrocarbon part is located within the core region of the membrane. Intermolecular membrane descriptions were important to the phase-transition temperature of DPPA liposomes. Suggesting that the behavior for the drug-induced phase-transition temperature of DPPA

CONCLUSIONS

113

liposome–drug mixtures involves the incorporation of compounds into the hydrocarbon chain region of the membrane bilayer. For antiarrhythmic and calcium-displacing activity, intramolecular descriptors were adequate in describing these endpoints. For antiarrhythmic activity, lipophilicity and molecular size were the major factors determining antiarrhythmic activity of CADs. The lack of intermolecular descriptors indicates that spatial requirements of the receptor are comparably unspeciﬁc, and lipophilic interactions play a major role in the inactivation process. For calcium ion replacement, the QSAR descriptors of importance were electrostatic/charge properties in nature and were consistent with electrostatic interactions at the surface of a membrane, probably involving headgroups of phospholipids. Calcium ions and CADs bind to the negatively charged head of phosphatidylserine, which is exposed to the aqueous phase in the assay system. QSAR models, determined by Klein’s group, provide a physicochemical rationale for the failure of these endpoints to correlate. The set of descriptors in each of the three QSAR models is quite different, which suggests that different mechanisms and/or sites of action result from each of these properties. Tomizawa et al. [186] recently measured the phospholipidosis-inducing potency of a test set of 33 compounds in isolated rat hepatocytes. Lipid accumulation was measured via a ﬂuorescence-labeled lipid assay and veriﬁed by electron microscopy. They report that the high net charge of a given molecule, which corresponds directly to the ionization state of compounds in organelles, and high c log P (>1) best describe the potential to cause phospholipidosis. Overall, these studies suggest that the prediction of the capacity of a drug to induce phospholipidosis can be made by calculation or measurement of their cationic amphiphilic properties. Also, the general structure of the CADs can be illustrated by molecular calculations. Phospholipidosis induced by CADs is probably a defense mechanism that involves storing the CADs. Whether phospholipidosis interferes with cell function is not clear. Future QSAR studies would be worthwhile, as the only endpoints studied are related to phospholipidosis, which is merely a symptom of excessive storage of phospholipids. The responses of phospholipidosis-containing cells to stressors such as oxidative stress still need to be determined.

6. CONCLUSIONS There is growing evidence that mitochondrial dysfunction is critical to the progression of steatosis to liver injury observed in NASH, regardless of whether it is induced by xenobiotics or by a Western sucrose/fat-rich diet and lack of exercise. Mitochondrial dysfunction, probably initiated and exacerbated by increased levels of endogenous free fatty acids, results in mitochondrial ROS production, depletion of cellular antioxidants, and induction of inﬂammatory and/or ﬁbrosis cytokine release.

114

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

Drugs that are lipophilic and cationic (e.g., cationic amphiphiles) are readily accumulated by the mitochondria (often, manyfold), so it is not surprising that they are more likely to cause mitochondrial toxicity than those drugs that are only lipophilic. Although the cationic amphiphile moiety of these drugs is not necessarily toxicophores, this moiety plays a major role in contributing to drug accumulation by the mitochondria. The toxicophore is largely the moiety that contributes to the lipophilicity and aromaticitiy of the drug, thereby enabling the drug access, and bind to target proteins that cause mitochondrial dysfunction and hepatotoxicity. This has largely been overshadowed by too much attention to the accumulation of these drugs and phospholipids by lysosomes (lysosomal phospholidosis), even though this has not yet been shown to affect cell viability. From this perspective, phospholipidosis could be regarded as a cellular defence mechanism, with the binding of the drugs to phospholipids acting as a detoxiﬁcation mechanism. Additionally, the inactivation by the drug of lysosomal phospholipase A2 would increase lysosomal phospholipids and could prevent the release of free fatty acids, and particularly the release of toxic unsaturated fatty acids. Intracellular free fatty acids are normally kept at low concentrations by undergoing β-oxidation in the mitochondrial matrix. However, fatty acid detoxiﬁcation would be prevented by cation drug accumulation in mitochondria, as it would inhibit fatty acid oxidation, respiratory inhibition, and/or uncoupling. Most of the drugs withdrawn from the market because of hepatotoxicity are mitochondrial toxins. These drugs include phenformin, buformin, troglitazone (rezulin), tolcapone (tasmar), and cerivastatin (baycol). Pernoline, used to treat attention-deﬁcit syndrome, was also withdrawn because of hepatotoxicity concerns in 2005, and it has a structure similar to that of some mitochondrial toxins. Drugs receiving black box warnings for hepatotoxicity or cardiotoxicity include antivirals and HIV protease inhibitors, and NSAIDs such as ketorolac, celecoxib, and naproxen. Of 70 drugs receiving this warning, mitochondrial liabilities have already been described for just over 50% [187]. Drugs or xenobiotics can impair mitochondrial function in many ways. Long-term inhibition of mitochopndria replication by antivirals and antibiotics is responsible for a host of pathologies (see Chapters 9 and 21). More acute inhibition of respiration undermines ATP production and accelerates ROS production. Uncoupling electron transport from phosphorylation also represses ATP generation, and both processes increase the probability of irreversible mitochnorial failure via MPT. In cells capable of responding, loss of mitochondrial ATP production results in compensatory increases in glycolysis and hence in lactate formation. Increased serum lactate and the symptoms of acidosis are biomarkers of mitochondrial toxicity. Presently, it is possible to predict potential mitochondrial toxicity of a drug from its structure only if it is structurally similar to electron donors or acceptors in the respiratory complex and so inhibits competitively. QSAR has the potential to estimate mitochondrial accumulation of cationic amphiphiles or the uncoupling protonophoric activity of simple phenolic xenobiotics. However, drugs can also

REFERENCES

115

inhibit electron transport via allosteric interactions that are not yet resolved sufﬁciently to enable QSAR approaches. Nevertheless, high-throughput assays are currently available to screen new chemical entities or candidate drugs for mitochondrial toxicity based on redox or potentiometric dyes, and respiration. Effects of drugs on MPT are readily monitored by following mitochondrial swelling at 540 nm, and antibodies for the respiratory complexes are available to identify drug targets (see Chapters 12 to 25). Finally, once the etiology and form of drug-induced mitochondrial dysfunction are identiﬁed, it can be addressed preclinically by SAR efforts to minimize or eliminate it. In cases where mitochondrial toxicity is an unavoidable consequence of on-target efﬁcacy, pharmaceutical or dietary interventions might be able to moderate tissue toxicity and adverse events. Such efforts are, however, increasingly justiﬁed by the growing appreciation that mitochondrial dysfunction contributes to drug toxicity and adverse side effects, but in some cases, also to the desired therapeutic effects of many beneﬁcial drugs. REFERENCES 1. Sato K, Ueda Y, Ueno K, Okamoto K, Iizuka H, Katsuda S. Hepatocellular carcinoma and nonalcoholic steatohepatitis developing during long-term administration of valproic acid. Virchows Arch. 2005;447:996–999. 2. Kennedy JA, Unger SA, Horowitz JD. Inhibition of carnitine palmitoyltransferase-1 in rat heart and liver by perhexiline and amiodarone. Biochem Pharmacol. 1996;52:273–280. 3. Pankow JS, Duncan BB, Schmidt MI, et al. Fasting plasma free fatty acids and risk of type 2 diabetes: the atherosclerosis risk in communities study. Diabetes Care. 2004;27:77–82. 4. Di Paola M, Lorusso M. Interaction of free fatty acids with mitochondria: coupling, uncoupling and permeability transition. Biochim Biophys Acta. 2006;1757:1330–1337. 5. Weltman MD, Farrell GC, Hall P, Ingelman-Sundberg M, Liddle C. Hepatic cytochrome P450 2E1 is increased in patients with nonalcoholic steatohepatitis. Hepatology. 1998;27:128–133. 6. Caro AA, Cederbaum AI. Role of intracellular calcium and phospholipase A2 in arachidonic acid-induced toxicity in liver cells overexpressing CYP2E1. Arch Biochem Biophys. 2007;457:252–263. 7. Di Paola M, Zaccagnino P, Oliveros-Celis C, Lorusso M. Arachidonic acid induces speciﬁc membrane permeability increase in heart mitochondria. FEBS Lett. 2006;580:775–781. 8. Wanders RJ, Tager JM. Lipid metabolism in peroxisomes in relation to human disease. Mol Aspects Med. 1998;19:69–154. 9. Malhi H, Gores GJ, Lemasters JJ. Apoptosis and necrosis in the liver: a tale of two deaths? Hepatology. 2006;43:S31–S44. 10. Feldstein AE, Werneburg NW, Canbay A, et al. Free fatty acids promote hepatic lipotoxicity by stimulating TNF-alpha expression via a lysosomal pathway. Hepatology. 2004;40:185–194.

116

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

11. Feldstein AE, Werneburg NW, Li Z, Bronk SF, Gores GJ. Bax inhibition protects against free fatty acid-induced lysosomal permeabilization. Am J Physiol Gastrointest Liver Physiol. 2006;290:G1339–G1346. 12. Unger RH. Lipotoxic diseases. Annu Rev Med. 2002;53:319–336. 13. Letteron P, Fromenty B, Terris B, Degott C, Pessayre D. Acute and chronic hepatic steatosis lead to invivo lipid peroxidation in mice. J Hepatol. 1996;24:200–208. 14. Eaton S, Zaitoun AM, Record CO, Bartlett K. Beta-oxidation in human alcoholic and non-alcoholic hepatic steatosis. Clin Sci (Lond). 1996;90:307–313. 15. Freneaux E, Labbe G, Letteron P, et al. Inhibition of the mitochondrial oxidation of fatty acids by tetracycline in mice and in man: possible role in microvesicular steatosis induced by this antibiotic. Hepatology. 1988;8:1056–1062. 16. Deboyser D, Goethals F, Krack G, Roberfroid M. Investigation into the mechanism of tetracycline-induced steatosis: study in isolated hepatocytes. Toxicol Appl Pharmacol. 1989;97:473–479. 17. Labbe G, Fromenty B, Freneaux E, et al. Effects of various tetracycline derivatives on in vitro and in vivo beta-oxidation of fatty acids, egress of triglycerides from the liver, accumulation of hepatic triglycerides, and mortality in mice. Biochem Pharmacol. 1991;41:638–641. 18. Damjanov I, Solter D. Ultrastructure of acute tetracycline induced liver change. Experientia. 1971;27:1204–1205. 19. Du Buy HG, Showacre JL. Selective localization of tetracycline in mitochondria of living cells. Science. 1961;133:196–197. 20. Combes B, Whalley PJ, Adams RH. Tetracycline and the liver. Prog Liver Dis. 1972;4:589–596. 21. Geneve J, Hayat-Bonan B, Labbe G, et al. Inhibition of mitochondrial beta-oxidation of fatty acids by pirprofen: role in microvesicular steatosis due to this nonsteroidal anti-inﬂammatory drug. J Pharmacol Exp Ther. 1987;242:1133–1137. 22. Fromenty B, Freneaux E, Labbe G, et al. Tianeptine, a new tricyclic antidepressant metabolized by beta-oxidation of its heptanoic side chain, inhibits the mitochondrial oxidation of medium and short chain fatty acids in mice. Biochem Pharmacol. 1989;38:3743–3751. 23. Le Dinh T, Freneaux E, Labbe G, et al. Amineptine, a tricyclic antidepressant, inhibits the mitochondrial oxidation of fatty acids and produces microvesicular steatosis of the liver in mice. J Pharmacol Exp Ther. 1988;247:745–750. 24. Pessayre D, Fromenty B. NASH: a mitochondrial disease. J Hepatol . 2005;42:928–940. 25. Lelliott CJ, Lopez M, Curtis RK, et al. Transcript and metabolite analysis of the effects of tamoxifen in rat liver reveals inhibition of fatty acid synthesis in the presence of hepatic steatosis. FASEB J. 2005;19:1108–1119. 26. Petit F, Fromenty B, Owen A, Estaquier J. Mitochondria are sensors for HIV drugs. Trends Pharmacol Sci. 2005;26:258–264. 27. Anderson N, Borlak J. Drug-induced phospholipidosis. FEBS Lett. 2006;580:5533–5540. 28. Yamamoto A, Adachi S, Matsuzawa Y, Kitani T, Hiraoka A, Seki K. Studies on drug-induced lipidosis: VII. Effects of bis-beta-diethyl-aminoethylether of hexestrol, chloroquine, homochlorocyclizine, prenylamine, and diazacholesterol on the lipid composition of rat liver and kidney. Lipids. 1976;11:616–622.

REFERENCES

117

29. Kubo M, Hostetler KY. Metabolic basis of diethylaminoethoxyhexestrol-induced phospholipid fatty liver. Am J Physiol. 1987;252:E375–E379. 30. Berson A, De B, V, Letteron P, et al. Steatohepatitis-inducing drugs cause mitochondrial dysfunction and lipid peroxidation in rat hepatocytes. Gastroenterology. 1998;114:764–774. 31. Varbiro G, Toth A, Tapodi A, Veres B, Sumegi B, Gallyas F, Jr. Concentration dependent mitochondrial effect of amiodarone. Biochem Pharmacol. 2003;65:1115–1128. 32. Fromenty B, Fisch C, Berson A, Letteron P, Larrey D, Pessayre D. Dual effect of amiodarone on mitochondrial respiration: initial protonophoric uncoupling effect followed by inhibition of the respiratory chain at the levels of complex I and complex II. J Pharmacol Exp Ther. 1990;255:1377–1384. 33. Kaufmann P, Torok M, Hanni A, Roberts P, Gasser R, Krahenbuhl S. Mechanisms of benzarone and benzbromarone-induced hepatic toxicity. Hepatology. 2005;41:925–935. 34. Waldhauser KM, Torok M, Ha HR, et al. Hepatocellular toxicity and pharmacological effect of amiodarone and amiodarone derivatives. J Pharmacol Exp Ther. 2006;319:1413–1423. 35. Bandyopadhyay S, Klaunig JE, Somani P. Cytotoxic interactions of cardioactive cationic amphiphilic compounds in primary rat hepatocytes in culture. Hepatology. 1990;12:48–58. 36. Christopoulos A, Grant MK, Ayoubzadeh N, et al. Synthesis and pharmacological evaluation of dimeric muscarinic acetylcholine receptor agonists. J Pharmacol Exp Ther. 2001;298:1260–1268. 37. Card JW, Racz WJ, Brien JF, Margolin SB, Massey TE. Differential effects of pirfenidone on acute pulmonary injury and ensuing ﬁbrosis in the hamster model of amiodarone-induced pulmonary toxicity. Toxicol Sci. 2003;75:169–180. 38. Reasor MJ, McCloud CM, Beard TL, et al. Hostetler KY. Comparative evaluation of amiodarone-induced phospholipidosis and drug accumulation in Fischer-344 and Sprague–Dawley rats. Toxicology. 1996;106:139–147. 39. Dieterle F, Ross A, Schlotterbeck G, Senn H. Metabolite projection analysis for fast identiﬁcation of metabolites in metabonomics: application in an amiodarone study. Anal Chem. 2006;78:3551–3561. 40. Sawada H, Takami K, Asahi S. A toxicogenomic approach to drug-induced phospholipidosis: analysis of its induction mechanism and establishment of a novel invitro screening system. Toxicol Sci. 2005;83:282–292. 41. Agoston M, Orsi F, Feher E, et al. Silymarin and vitamin E reduce amiodarone-induced lysosomal phospholipidosis in rats. Toxicology. 2003;190:231–241. 42. Pessayre D, Bichara M, Degott C, Potet F, Benhamou JP, Feldmann G. Perhexiline maleate–induced cirrhosis. Gastroenterology. 1979;76:170–177. 43. Deschamps D, DeBeco V, Fisch C, Fromenty B, Guillouzo A, Pessayre D. Inhibition by perhexiline of oxidative phosphorylation and the beta-oxidation of fatty acids: possible role in pseudoalcoholic liver lesions. Hepatology. 1994;19:948–961. 44. Hostetler KY, Reasor M, Yazaki PJ. Chloroquine-induced phospholipid fatty liver: measurement of drug and lipid concentrations in rat liver lysosomes. J Biol Chem. 1985;260:215–219.

118

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

45. Almotreﬁ AA. Effects of class I antiarrhythmic drugs on mitochondrial ATPase activity in guinea pig heart preparations. Gen Pharmacol. 1993;24:233–237. 46. Mironova GD, Grigoriev SM, Skarga Y, Negoda AE, Kolomytkin OV. ATP-dependent potassium channel from rat liver mitochondria: inhibitory analysis, channel clusterization. Membr Cell Biol. 1997;10:583–591. 47. Wold JS, Joost RR, Grifﬁng WJ. Phospholipid accumulation in rats produced by ﬂuoxetine and chlorphentermine. Toxicol Appl Pharmacol. 1976;37:118–119. 48. Berson A, Fau D, Fornacciari R, et al. Mechanisms for experimental buprenorphine hepatotoxicity: major role of mitochondrial dysfunction versus metabolic activation. J Hepatol. 2001;34:261–269. 49. Berson A, Renault S, Letteron P, et al. Uncoupling of rat and human mitochondria: a possible explanation for tacrine-induced liver dysfunction. Gastroenterology. 1996;110:1878–1890. 50. Mansouri A, Haouzi D, Descatoire V, et al. Tacrine inhibits topoisomerases and DNA synthesis to cause mitochondrial DNA depletion and apoptosis in mouse liver. Hepatology. 2003;38:715–725. 51. Luef GJ, Waldmann M, Sturm W, et al. Valproate therapy and nonalcoholic fatty liver disease. Ann Neurol. 2004;55:729–732. 52. Sugimoto T, Woo M, Nishida N, Takeuchi T, Sakane Y, Kobayashi Y. Hepatotoxicity in rat following administration of valproic acid. Epilepsia. 1987;28:142–146. 53. Spahr L, Negro F, Rubbia-Brandt L, et al. Acute valproate-associated microvesicular steatosis: Could the [13 C]methionine breath test be useful to assess liver mitochondrial function? Dig Dis Sci. 2001;46:2758–2761. 54. Fromenty B, Pessayre D. Inhibition of mitochondrial beta-oxidation as a mechanism of hepatotoxicity. Pharmacol Ther. 1995;67:101–154. 55. Baldwin GS, Abbott FS, Nau H. Binding of a valproate metabolite to the trifunctional protein of fatty acid oxidation. FEBS Lett. 1996;384:58–60. 56. Natarajan SK, Eapen CE, Pullimood AB, Balasubramanian KA. Oxidative stress in experimental liver microvesicular steatosis: role of mitochondria and peroxisomes. J Gastroenterol Hepatol. 2006;21:1240–1249. 57. Tong V, Teng XW, Chang TK, Abbott FS. Valproic acid I: time course of lipid peroxidation biomarkers, liver toxicity, and valproic acid metabolite levels in rats. Toxicol Sci. 2005;86:427–435. 58. Begriche K, Igoudjil A, Pessayre D, Fromenty B. Mitochondrial dysfunction in NASH: causes, consequences and possible means to prevent it. Mitochondrion. 2006;6:1–28. 59. Lagouge M, Argmann C, Gerhart-Hines Z, et al. Resveratrol improves mitochondrial function and protects against metabolic disease by activating SIRT1 and PGC-1alpha. Cell . 2006;127:1109–1122. 60. Baur JA, Pearson KJ, Price NL, et al. Lakatta EG, Le Couteur D, Shaw RJ, et al. Resveratrol improves health and survival of mice on a high-calorie diet. Nature. 2006;444:337–342. 61. Patel T, Gores GJ. Inhibition of bile-salt-induced hepatocyte apoptosis by the antioxidant lazaroid U83836E. Toxicol Appl Pharmacol. 1997;142:116–122. 62. Rolo AP, Palmeira CM, Wallace KB. Interactions of combined bile acids on hepatocyte viability: cytoprotection or synergism. Toxicol Lett. 2002;126:197–203.

REFERENCES

119

63. Rolo AP, Oliveira PJ, Moreno AJ, Palmeira CM. Bile acids affect liver mitochondrial bioenergetics: possible relevance for cholestasis therapy. Toxicol Sci. 2000;57:177–185. 64. Palmeira CM, Rolo AP. Mitochondrially-mediated toxicity of bile acids. Toxicology. 2004;203:1–15. 65. Gumpricht E, Dahl R, Devereaux MW, Sokol RJ. Licorice compounds glycyrrhizin and 18beta-glycyrrhetinic acid are potent modulators of bile acid–induced cytotoxicity in rat hepatocytes. J Biol Chem. 2005;280:10556–10563. 66. Sokol RJ, Dahl R, Devereaux MW, Yerushalmi B, Kobak GE, Gumpricht E. Human hepatic mitochondria generate reactive oxygen species and undergo the permeability transition in response to hydrophobic bile acids. J Pediatr Gastroenterol Nutr. 2005;41:235–243. 67. Yerushalmi B, Dahl R, Devereaux MW, Gumpricht E, Sokol RJ. Bile acid–induced rat hepatocyte apoptosis is inhibited by antioxidants and blockers of the mitochondrial permeability transition. Hepatology. 2001;33:616–626. 68. Sodeman T, Bronk SF, Roberts PJ, Miyoshi H, Gores GJ. Bile salts mediate hepatocyte apoptosis by increasing cell surface trafﬁcking of Fas. Am J Physiol Gastrointest Liver Physiol. 2000;278:G992–G999. 69. Guicciardi ME, Gores GJ. Cholestatic hepatocellular injury: What do we know and how should we proceed? J Hepatol. 2005;42:297–300. 70. Higuchi H, Gores GJ. Bile acid regulation of hepatic physiology: IV. Bile acids and death receptors. Am J Physiol Gastrointest Liver Physiol. 2003;284:G734–G738. 71. Byrne JA, Strautnieks SS, Mieli-Vergani G, Higgins CF, Linton KJ, Thompson RJ. The human bile salt export pump: characterization of substrate speciﬁcity and identiﬁcation of inhibitors. Gastroenterology. 2002;123:1649–1658. 72. Siraki AG, Pourahmad J, Chan TS, Khan S, O’Brien PJ. Endogenous and endobiotic induced reactive oxygen species formation by isolated hepatocytes. Free Radic Biol Med. 2002;32:2–10. 73. Shen HM, Liu ZG. JNK signaling pathway is a key modulator in cell death mediated by reactive oxygen and nitrogen species. Free Radic Biol Med. 2006;40:928–939. 74. Pourahmad J, Khan S, O’Brien PJ. Lysosomal oxidative stress cytotoxicity induced by nitrofurantoin redox cycling in hepatocytes. Adv Exp Med Biol. 2001;500:261–265. 75. Terman A, Gustafsson B, Brunk UT. The lysosomal–mitochondrial axis theory of postmitotic aging and cell death. Chem Biol Interact. 2006;163:29–37. 76. Chan K, Truong D, Shangari N, O’Brien PJ. Drug-induced mitochondrial toxicity. Expert Opin Drug Metab Toxicol. 2005;1:655–669. 77. Degli EM. Inhibitors of NADH–ubiquinone reductase: an overview. Biochim Biophys Acta. 1998;1364:222–235. 78. Miyoshi H. Inhibitors of mitochondrial respiratory enzymes. J Pestic Sci. 2005;30:120–121. 79. Caboni P, Sherer TB, Zhang N, et al. Rotenone, deguelin, their metabolites, and the rat model of Parkinson’s disease. Chem Res Toxicol. 2004;17:1540–1548. 80. Li N, Ragheb K, Lawler G, et al. Mitochondrial complex I inhibitor rotenone induces apoptosis through enhancing mitochondrial reactive oxygen species production. J Biol Chem. 2003;278:8516–8525.

120

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

81. Annepu J, Ravindranath V. 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine–induced complex I inhibition is reversed by disulﬁde reductant, dithiothreitol in mouse brain. Neurosci Lett. 2000;289:209–212. 82. Guth PS, Spirtes MA. The phenothiazine tranquilizers: biochemical and biophysical actions. Int Rev Neurobiol. 1964;12:231–278. 83. Balijepalli S, Boyd MR, Ravindranath V. Inhibition of mitochondrial complex I by haloperidol: the role of thiol oxidation. Neuropharmacology. 1999;38:567–577. 84. Maurer I, Moller HJ. Inhibition of complex I by neuroleptics in normal human brain cortex parallels the extrapyramidal toxicity of neuroleptics. Mol Cell Biochem. 1997;174:255–259. 85. Igarashi K, Kasuya F, Fukui M, Usuki E, Castagnoli N, Jr.. Studies on the metabolism of haloperidol (HP): the role of CYP3A in the production of the neurotoxic pyridinium metabolite HPP+ found in rat brain following ip administration of HP. Life Sci. 1995;57:2439–2446. 86. Rollema H, Skolnik M, D’Engelbronner J, Igarashi K, Usuki E, Castagnoli N, Jr.. MPP( + )-like neurotoxicity of a pyridinium metabolite derived from haloperidol: in vivo microdialysis and in vitro mitochondrial studies. J Pharmacol Exp Ther. 1994;268:380–387. 87. Veitch K, Hue L. Flunarizine and cinnarizine inhibit mitochondrial complexes I and II: possible implication for parkinsonism. Mol Pharmacol. 1994;45:158–163. 88. Sztark F, Nouette-Gaulain K, Malgat M, Dabadie P, Mazat JP. Absence of stereospeciﬁc effects of bupivacaine isomers on heart mitochondrial bioenergetics. Anesthesiology. 2000;93:456–462. 89. Johnson ME, Uhl CB, Spittler KH, Wang H, Gores GJ. Mitochondrial injury and caspase activation by the local anesthetic lidocaine. Anesthesiology. 2004;101:1184–1194. 90. Hanley PJ, Ray J, Brandt U, Daut J. Halothane, isoﬂurane and sevoﬂurane inhibit NADH:ubiquinone oxidoreductase (complex I) of cardiac mitochondria. J Physiol. 2002;544:687–693. 91. Tanaka S, Ohta S, Shimonaka H, et al. The effects of dantrolene on the respiratory function in isolated rat liver mitochondria. Masui (Japanese Journal of Anesthesiology). 1986;35:755–761. 92. Haettinger JR, Mackerer CR, Colton DG, Mehlman MA. Effects of 5,5 -diphenyl-2-thiohydantoin on respiration and oxidative phosphorylation of rat liver mitochondria. Chem Biol Interact. 1975;10:255–264. 93. Becuwe P, Dauca M. Comparison of cytotoxicity induced by hypolipidemic drugs via reactive oxygen species in human and rodent liver cells. Int J Mol Med. 2005;16:483–492. 94. Qu B, Li QT, Wong KP, Tan TM, Halliwell B. Mechanism of cloﬁbrate hepatotoxicity: mitochondrial damage and oxidative stress in hepatocytes. Free Radic Biol Med. 2001;31:659–669. 95. Keller BJ, Yamanaka H, Thurman RG. Inhibition of mitochondrial respiration and oxygen-dependent hepatotoxicity by six structurally dissimilar peroxisomal proliferating agents. Toxicology. 1992;71:49–61. 96. Brunmair B, Lest A, Staniek K, et al. Fenoﬁbrate impairs rat mitochondrial function by inhibition of respiratory complex I. J Pharmacol Exp Ther. 2004;311:109–114. 97. Zhou S, Wallace KB. The effect of peroxisome proliferators on mitochondrial bioenergetics. Toxicol Sci. 1999;48:82–89.

REFERENCES

121

98. Chan NN, Brain HP, Feher MD. Metformin-associated lactic acidosis: a rare or very rare clinical entity? Diabet Med. 1999;16:273–281. 99. El Mir MY, Nogueira V, Fontaine E, Averet N, Rigoulet M, Leverve X. Dimethylbiguanide inhibits cell respiration via an indirect effect targeted on the respiratory chain complex I. J Biol Chem. 2000;275:223–228. 100. Owen MR, Doran E, Halestrap AP. Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochem J. 2000;348 (pt 3): 607–614. 101. Brunmair B, Staniek K, Gras F, et al. Thiazolidinediones, like metformin, inhibit respiratory complex I: a common mechanism contributing to their antidiabetic actions? Diabetes. 2004;53:1052–1059. 102. Tirmenstein MA, Hu CX, Gales TL, et al. Effects of troglitazone on HepG2 viability and mitochondrial function. Toxicol Sci. 2002;69:131–138. 103. Masubuchi Y, Kano S, Horie T. Mitochondrial permeability transition as a potential determinant of hepatotoxicity of antidiabetic thiazolidinediones. Toxicology. 2006;222:233–239. 104. Tafazoli S, O’Brien PJ. Peroxidases: a role in the metabolism and side effects of drugs. Drug Discov Today. 2005;10:617–625. 105. Narayanan PK, Hart T, Elcock F, et al. Troglitazone-induced intracellular oxidative stress in rat hepatoma cells: a ﬂow cytometric assessment. Cytometry A. 2003;52:28–35. 106. Smith MT. Mechanisms of troglitazone hepatotoxicity. Chem Res Toxicol. 2003;16:679–687. 107. Fau D, Eugene D, Berson A, et al. Toxicity of the antiandrogen ﬂutamide in isolated rat hepatocytes. J Pharmacol Exp Ther. 1994;269:954–962. 108. Waterﬁeld CJ, Delaney J, Kerai DJ, Timbrell JA. Correlations between invivo and invitro effects of toxic compounds: studies with hydrazine. Toxicol invitro. 1997;11:217–227. 109. Chan TS, Teng S, Wilson JX, Galati G, Khan S, O’Brien PJ. Coenzyme Q cytoprotective mechanisms for mitochondrial complex I cytopathies involves NAD(P)H:quinone oxidoreductase 1(NQO1). Free Radic Res. 2002;36:421–427. 110. Sun J, Trumpower BL. Superoxide anion generation by the cytochrome bc1 complex. Arch Biochem Biophys. 2003;419:198–206. 111. Khan S, O’Brien PJ. Modulating hypoxia-induced hepatocyte injury by affecting intracellular redox state. Biochim Biophys Acta. 1995;1269:153–161. 112. Niknahad H, Khan S, O’Brien PJ. Hepatocyte injury resulting from the inhibition of mitochondrial respiration at low oxygen concentrations involves reductive stress and oxygen activation. Chem Biol Interact. 1995;98:27–44. 113. Masubuchi Y, Suda C, Horie T. Involvement of mitochondrial permeability transition in acetaminophen-induced liver injury in mice. J Hepatol. 2005;42:110–116. 114. Ramsay RR, Rashed MS, Nelson SD. In vitro effects of acetaminophen metabolites and analogs on the respiration of mouse liver mitochondria. Arch Biochem Biophys. 1989;273:449–457. 115. Ludwig LM, Tanaka K, Eells JT, et al. Preconditioning by isoﬂurane is mediated by reactive oxygen species generated from mitochondrial electron transport chain complex III. Anesth Analg. 2004;99:1308–1315.

122

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

116. Riess ML, Kevin LG, McCormick J, Jiang MT, Rhodes SS, Stowe DF. Anesthetic preconditioning: the role of free radicals in sevoﬂurane-induced attenuation of mitochondrial electron transport in guinea pig isolated hearts. Anesth Analg. 2005;100:46–53. 117. Yoshikawa S, Muramoto K, Shinzawa-Itoh K, et al. Reaction mechanism of bovine heart cytochrome c oxidase. Biochim Biophys Acta. 2006;1757:395–400. 118. Kiyomiya K, Matsushita N, Matsuo S, Kurebe M. Cephaloridine-induced inhibition of cytochrome c oxidase activity in the mitochondria of cultured renal epithelial cells (LLC-PK(1)) as a possible mechanism of its nephrotoxicity. Toxicol Appl Pharmacol. 2000;167:151–156. 119. Kiyomiya K, Matsushita N, Kurebe M, Nakagawa H, Matsuo S. Mitochondrial cytochrome c oxidase as a target site for cephalosporin antibiotics in renal epithelial cells (LLC-PK(1)) and renal cortex. Life Sci. 2002;72:49–57. 120. Kohda Y, Gemba M. Cephaloridine induces translocation of protein kinase C delta into mitochondria and enhances mitochondrial generation of free radicals in the kidney cortex of rats causing renal dysfunction. J Pharmacol Sci. 2005;98:49–57. 121. Tune BM, Hsu CY. The renal mitochondrial toxicity of cephalosporins: speciﬁcity of the effect on anionic substrate uptake. J Pharmacol Exp Ther. 1990;252:65–69. 122. Benbrik E, Chariot P, Bonavaud S, et al. Cellular and mitochondrial toxicity of zidovudine (AZT), didanosine (ddI) and zalcitabine (ddC) on cultured human muscle cells. J Neurol Sci. 1997;149:19–25. 123. Lamperth L, Dalakas MC, Dagani F, Anderson J, Ferrari R. Abnormal skeletal and cardiac muscle mitochondria induced by zidovudine (AZT) in human muscle in vitro and in an animal model. Lab Invest. 1991;65:742–751. 124. de la Asuncion JG, Del Olmo ML, Gomez-Cambronero LG, Sastre J, Pallardo FV, Vina J. AZT induces oxidative damage to cardiac mitochondria: protective effect of vitamins C and E. Life Sci. 2004;76:47–56. 125. Moreno SN. Pentamidine is an uncoupler of oxidative phosphorylation in rat liver mitochondria. Arch Biochem Biophys. 1996;326:15–20. 126. Garlid KD, Nakashima RA. Studies on the mechanism of uncoupling by amine local anesthetics: evidence for mitochondrial proton transport mediated by lipophilic ion pairs. J Biol Chem. 1983;258:7974–7980. 127. Somasundaram S, Sigthorsson G, Simpson RJ, et al. Uncoupling of intestinal mitochondrial oxidative phosphorylation and inhibition of cyclooxygenase are required for the development of NSAID-enteropathy in the rat. Aliment Pharmacol Ther. 2000;14:639–650. 128. Masubuchi Y, Yamada S, Horie T. Diphenylamine as an important structure of nonsteroidal anti-inﬂammatory drugs to uncouple mitochondrial oxidative phosphorylation. Biochem Pharmacol. 1999;58:861–865. 129. Boelsterli UA. Diclofenac-induced liver injury: a paradigm of idiosyncratic drug toxicity. Toxicol. Appl Pharmacol. 2003;192:307–322. 130. Inoue A, Muranaka S, Fujita H, Kanno T, Tamai H, Utsumi K. Molecular mechanism of diclofenac-induced apoptosis of promyelocytic leukemia: dependency on reactive oxygen species, Akt, Bid, cytochrome and caspase pathway. Free Radic Biol Med. 2004;37:1290–1299.

REFERENCES

123

131. Tay VK, Wang AS, Leow KY, Ong MM, Wong KP, Boelsterli UA. Mitochondrial permeability transition as a source of superoxide anion induced by the nitroaromatic drug nimesulide in vitro. Free Radic Biol Med. 2005;39:949–959. 132. Berson A, Cazanave S, Descatoire V, et al. The anti-inﬂammatory drug, nimesulide (4-nitro-2-phenoxymethane-sulfoanilide), uncouples mitochondria and induces mitochondrial permeability, transition in human hepatoma cells: protection by albumin. J Pharmacol Exp Ther. 2006;318:444–454. 133. Basivireddy J, Jacob M, Pulimood AB, Balasubramanian KA. Indomethacin-induced renal damage: role of oxygen free radicals. Biochem Pharmacol. 2004;67:587–599. 134. Cardoso CM, Custodio JB, Almeida LM, Moreno AJ. Mechanisms of the deleterious effects of tamoxifen on mitochondrial respiration rate and phosphorylation efﬁciency. Toxicol Appl Pharmacol. 2001;176:145–152. 135. Cardoso CM, Moreno AJ, Almeida LM, Custodio JB. Comparison of the changes in adenine nucleotides of rat liver mitochondria induced by tamoxifen and 4-hydroxytamoxifen. Toxicol in vitro. 2003;17:663–670. 136. Hynes J, Marroquin LD, Ogurtsov VI, et al. Investigation of drug-induced mitochondrial toxicity using ﬂuorescence-based oxygen-sensitive probes. Toxicol Sci. 2006;92:186–200. 137. Sun X, Garlid KD. On the mechanism by which bupivacaine conducts protons across the membranes of mitochondria and liposomes. J Biol Chem. 1992;267:19147–19154. 138. Sztark F, Malgat M, Dabadie P, Mazat JP. Comparison of the effects of bupivacaine and ropivacaine on heart cell mitochondrial bioenergetics. Anesthesiology. 1998;88:1340–1349. 139. Irwin W, Fontaine E, Agnolucci L, et al. Bupivacaine myotoxicity is mediated by mitochondria. J Biol Chem. 2002;277:12221–12227. 140. Floridi A, Di Padova M, Barbieri R, Arcuri E. Effect of local anesthetic ropivacaine on isolated rat liver mitochondria. Biochem Pharmacol. 1999;58:1009–1016. 141. Curti C, Mingatto FE, Polizello AC, Galastri LO, Uyemura SA, Santos AC. Fluoxetine interacts with the lipid bilayer of the inner membrane in isolated rat brain mitochondria, inhibiting electron transport and F1 FO -ATPase activity. Mol Cell Biochem. 1999;199:103–109. 142. Acco A, Comar JF, Bracht A. Metabolic effects of propofol in the isolated perfused rat liver. Basic Clin Pharmacol Toxicol. 2004;95:166–174. 143. Chen RM, Wu CH, Chang HC, et al. Propofol suppresses macrophage functions and modulates mitochondrial membrane potential and cellular adenosine triphosphate synthesis. Anesthesiology. 2003;98:1178–1185. 144. Borges N. Tolcapone-related liver dysfunction: implications for use in Parkinson’s disease therapy. Drug Saf. 2003;26:743–747. 145. Nissinen E, Kaheinen P, Penttila KE, Kaivola J, Linden IB. Entacapone, a novel catechol-O-methyltransferase inhibitor for Parkinson’s disease, does not impair mitochondrial energy production. Eur J Pharmacol. 1997;340:287–294. 146. Haasio K, Koponen A, Penttila KE, Nissinen E. Effects of entacapone and tolcapone on mitochondrial membrane potential. Eur J Pharmacol. 2002;453:21–26. 147. Masubuchi Y, Yamada S, Horie T. Diphenylamine as an important structure of nonsteroidal anti-inﬂammatory drugs to uncouple mitochondrial oxidative phosphorylation. Biochem Pharmacol. 1999;58:861–865.

124

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

148. Boelsterli UA. Diclofenac-induced liver injury: a paradigm of idiosyncratic drug toxicity. Toxicol Appl Pharmacol. 2003;192:307–322. 149. Spaniol M, Bracher R, Ha HR, Follath F, Krahenbuhl S. Toxicity of amiodarone and amiodarone analogues on isolated rat liver mitochondria. J Hepatol. 2001;35:628–636. 150. Nakagawa Y, Tayama K. Mechanism of mitochondrial dysfunction and cytotoxicity induced by tropolones in isolated rat hepatocytes. Chem Biol Interact. 1998;116:45–60. 151. Berthiaume JM, Wallace KB. Adriamycin-induced oxidative mitochondrial cardiotoxicity. Cell Biol Toxicol. 2007;23:15–25. 152. Hasinoff BB, Schnabl KL, Marusak RA, Patel D, Huebner E. Dexrazoxane (ICRF-187) protects cardiac myocytes against doxorubicin by preventing damage to mitochondria. Cardiovasc Toxicol. 2003;3:89–99. 153. Santos DL, Moreno AJ, Leino RL, Froberg MK, Wallace KB. Carvedilol protects against doxorubicin-induced mitochondrial cardiomyopathy. Toxicol Appl Pharmacol. 2002;185:218–227. 154. Kang YJ, Sun X, Chen Y, Zhou Z. Inhibition of doxorubicin chronic toxicity in catalase-overexpressing transgenic mouse hearts. Chem Res Toxicol. 2002;15:1–6. 155. Yen HC, Oberley TD, Gairola CG, Szweda LI, St Clair DK. Manganese superoxide dismutase protects mitochondrial complex I against adriamycin-induced cardiomyopathy in transgenic mice. Arch Biochem Biophys. 1999;362:59–66. 156. Zhou S, Starkov A, Froberg MK, Leino RL, Wallace KB. Cumulative and irreversible cardiac mitochondrial dysfunction induced by doxorubicin. Cancer Res. 2001;61:771–777. 157. Goormaghtigh E, Huart P, Praet M, Brasseur R, Ruysschaert JM. Structure of the adriamycin–cardiolipin complex: role in mitochondrial toxicity. Biophys Chem. 1990;35:247–257. 158. Lee CS, Park SY, Ko HH, Han ES. Effect of change in cellular GSH levels on mitochondrial damage and cell viability loss due to mitomycin C in small cell lung cancer cells. Biochem Pharmacol. 2004;68:1857–1867. 159. Park SY, Ko HH, Song JH, Han ES, Lee CS. Differential effect of nitrogen species on changes in mitochondrial membrane permeability due to mitomycin C in lung epithelial cells. Naunyn Schmiedebergs Arch Pharmacol. 2004;369:312–321. 160. Hadler HI, Cook GL. A requirement of Pi for the transitory uncoupling of rat liver mitochondria by hydrazine, when beta-hydroxybutyrate is the substrate. J Environ Pathol Toxicol. 1978;1:419–432. 161. Hussain SM, Frazier JM. Cellular toxicity of hydrazine in primary rat hepatocytes. Toxicol Sci. 2002;69:424–432. 162. Baliga R, Ueda N, Walker PD, Shah SV. Oxidant mechanisms in toxic acute renal failure. Drug Metab Rev. 1999;31:971–997. 163. Servais H, Van Der SP, Thirion G, Van der EG, Van Bambeke F, Tulkens PM, Mingeot-Leclercq MP. Gentamicin-induced apoptosis in LLC-PK1 cells: involvement of lysosomes and mitochondria. Toxicol Appl Pharmacol. 2005;206:321–333. 164. Lesniak W, Pecoraro VL, Schacht J. Ternary complexes of gentamicin with iron and lipid catalyze formation of reactive oxygen species. Chem Res Toxicol. 2005;18:357–364.

REFERENCES

125

165. Liguori MJ, Anderson MG, Bukofzer S, et al. Microarray analysis in human hepatocytes suggests a mechanism for hepatotoxicity induced by trovaﬂoxacin. Hepatology. 2005;41:177–186. 166. Wakabayashi T. Megamitochondria formation: physiology and pathology. J Cell Mol Med. 2002;6:497–538. 167. Shishido S, Koga H, Harada M, Kumemura H, et al. Hydrogen peroxide overproduction in megamitochondria of troglitazone-treated human hepatocytes. Hepatology. 2003;37:136–147. 168. Caldwell SH, Swerdlow RH, Khan EM, et al. Mitochondrial abnormalities in non-alcoholic steatohepatitis. J Hepatol . 1999;31:430–434. 169. Guastadisegni C, Balduzzi M, Mancuso MT, Di Consiglio E. Liver mitochondria alterations in chloroform-treated Sprague–Dawley rats. J Toxicol Environ Health A. 1999;57:415–429. 170. Hernandez-Munoz R, Sanchez-Sevilla L, Martinez-Gomez A, Dent MA. Changes in mitochondrial adenine nucleotides and in permeability transition in two models of rat liver regeneration. Hepatology. 2003;37:842–851. 171. Murphy MP, Smith RA. Targeting antioxidants to mitochondria by conjugation to lipophilic cations. Annu Rev Pharmacol Toxicol. 2007;47:629–656. 172. Stockdale M, Selwyn MJ. Effects of ring substituents on the activity of phenols as inhibitors and uncouplers of mitochondrial respiration. Eur J Biochem. 1971;21:565–574. 173. Cronin MT, Manga N, Seward JR, Sinks GD, Schultz TW. Parametrization of electrophilicity for the prediction of the toxicity of aromatic compounds. Chem Res Toxicol. 2001;14:1498–1505. 174. Hansch C, McKarns SC, Smith CJ, Doolittle DJ. Comparative QSAR evidence for a free-radical mechanism of phenol-induced toxicity. Chem Biol Interact. 2000;127:61–72. 175. Moridani MY, Siraki A, O’Brien PJ. Quantitative structure toxicity relationships for phenols in isolated rat hepatocytes. Chem Biol Interact. 2003;145:213–223. 176. Argese E, Bettiol C, Giurin G, Miana P. Quantitative structure–activity relationships for the toxicity of chlorophenols to mammalian submitochondrial particles. Chemosphere. 1999;38:2281–2292. 177. Miyoshi H, Tsujishita H, Tokutake N, Fujita T. Quantitative analysis of uncoupling activity of substituted phenols with a physicochemical substituent and molecular parameters. Biochim Biophys Acta. 1990;1016:99–106. 178. Siraki AG, Chevaldina T, O’Brien PJ. Application of quantitative structure-toxicity relationships for acute NSAID cytotoxicity in rat hepatocytes. Chem Biol Interact. 2005;151:177–191. 179. Galati G, Tafazoli S, Sabzevari O, Chan TS, O’Brien PJ. Idiosyncratic NSAID drug induced oxidative stress. Chem Biol Interact. 2002;142:25–41. 180. Casanovas AM, Malmary Nebot MF, Courriere P, Oustrin J. Inhibition of cytochrome oxidase activity by local anaesthetics. Biochem Pharmacol. 1983;32:2715–2719. 181. Stringer BK, Harmon HJ. Inhibition of cytochrome oxidase by dibucaine. Biochem Pharmacol. 1990;40:1077–1081.

126

DRUG-ASSOCIATED MITOCHONDRIAL TOXICITY

182. Klein CD, Klingmuller M, Schellinski C, et al. Synthesis, pharmacological and biophysical characterization, and membrane-interaction QSAR analysis of cationic amphiphilic model compounds. J Med Chem. 1999;42:3874–3888. 183. Bandyopadhyay S, Klaunig JE, Somani P. Cytotoxic interactions of cardioactive cationic amphiphilic compounds in primary rat hepatocytes in culture. Hepatology. 1990;12:48–58. 184. Lullmann H, Lullmann-Rauch R, Wassermann O. Lipidosis induced by amphiphilic cationic drugs. Biochem Pharmacol. 1978;27:1103–1108. 185. Ploemen JP, Kelder J, Hafmans T, et al. Use of physicochemical calculation of pK a and C Log P to predict phospholipidosis-inducing potential: a case study with structurally related piperazines. Exp Toxicol Pathol. 2004;55:347–355. 186. Tomizawa K, Sugano K, Yamada H, Horii I. Physicochemical and cell-based approach for early screening of phospholipidosis-inducing potential. J Toxicol Sci. 2006;31:315–324. 187. Dykens JA, Will Y. Signiﬁcance of mitochondrial toxicity testing in drug development. Drug Dev Today. 2007;12:777–785.

4 PHARMACOGENETICS OF MITOCHONDRIAL DRUG TOXICITY Neil Howell MIGENIX Corporation, San Diego, California

Corinna Howell Matrilinex LLC, San Diego, California

1. 2. 3 4.

Introduction Mitochondrial DNA mutations, aminoglycosides, and deafness Disputed Role of 16189 mtDNA polymorphism in type 2 diabetes Conclusions

127 128 130 135

1. INTRODUCTION As discussed throughout this volume, mitochondrial drug toxicity is a serious problem that limits drug development and has severe medical consequences. Fortunately, there is increasing awareness of this problem [1–3]. As one sign of progress, microarray analysis, gene expression studies, and other new and sophisticated technologies are being used to analyze mitochondrial drug toxicity [4–8]. There will be an increasing use of such techniques earlier during the drug development process to identify mitochondrial toxicity. Our focus here is not on mitochondrial drug toxicity itself but on the potential utility and value of mitochondrial pharmacogenetics. Framing the issue as a question: Will the mitochondrial DNA (mtDNA) genotype determine the response to Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

127

128

PHARMACOGENETICS OF MITOCHONDRIAL DRUG TOXICITY

a drug? At the present time, there is little work in this area. However, recognizing that dozens of disorders involve mitochondrial dysfunction, and given our increasing understanding of mitochondrial dug toxicity as a frequent and serious problem, we believe that this is an opportune moment to lay some groundwork. In this chapter we ﬁrst review and critically summarize the one system in which a mitochondrial genotype has profound consequences for drug-induced toxicity. Second, we analyze the purported role of an mtDNA polymorphism in a major disease. This second topic, the possible pathogenic role of mtDNA sequence changes in complex disorders, is germane because it offers valuable cautionary lessons for mitochondrial pharmacogenetics (see also [9] and [10]). Mitochondrial pharmacogenetics is an important component of the drug toxicity/drug development effort. By recognizing the complexities of human mitochondrial genetics and by establishing appropriate quality control measures, the inevitable “teething problems” can be reduced.

2. MITOCHONDRIAL DNA MUTATIONS, AMINOGLYCOSIDES, AND DEAFNESS A large number of mutations, ranging from single base pair substitutions to large deletions, in the human mitochondrial genome are pathogenic. Given the key role of these organelles in cellular energy production (see Chapter 1), mitochondrial genetic diseases often affect multiple tissues, especially those with high energy or metabolic demands. Typically, patients present with an array of clinical abnormalities, and there is marked syndromic heterogeneity among maternal relatives. Syndromic deafness is one frequent deﬁcit in such patients, a pathology that ﬁts with the high energy demand of the auditory nerves (reviewed in [11]). For example, about 40% of patients with either MERRF (myoclonic epilepsy with ragged-red ﬁber disease) or MELAS (mitochondrial encephalopathy with lactic acidosis and strokelike episodes), caused by the A8344G and A3243G mutations, respectively, showed clinically signiﬁcant deafness [12]. In addition to these cases of syndromic deafness, an mtDNA mutation at nucleotide 1555 within the gene encoding the 12S ribosomal RNA (rRNA) was associated with nonsyndromic deafness in the early 1990s ([13]; reviewed in 11,14,15). The 1555 mutation, as it is termed, has now been identiﬁed in a number of maternal pedigrees from around the world, and it occurs in peoples of European, Asian, and African ethnicity. It thus appears that this mutation has arisen many times in the mitochondrial gene pool during human evolution. Furthermore, this mutation is not especially rare. For example, in a random and anonymous screening of newborns from the United States, one instance of the 1555 mutation was found in slightly fewer than the 1200 people tested [16]. The deafness in 1555 families is not completely penetrant, and some family members have clinically normal hearing throughout life. Also, the severity of the hearing loss varies, ranging from profound sensorineural deafness developing during infancy to mild hearing loss during late adulthood. Fischel-Ghodsian [11]

MITOCHONDRIAL DNA MUTATIONS, AMINOGLYCOSIDES, AND DEAFNESS

129

has reviewed the audiological and other clinical studies and concludes that the loss of hearing in these individuals is due primarily to cochlear dysfunction, but that the vestibular system is essentially spared. The initial studies of multiple maternal pedigrees that carry the 1555 mutation showed that the deafness occurred after exposure to aminoglycoside antibiotics such as streptomycin, gentamycin, and kanamycin. Hutchin [14] cites the cumulative ﬁndings that of the more than 120 people who carried the 1555 mutation and received normal therapeutic doses of aminoglycoside, all of them suffered hearing loss. So here, clearly, is an example of a mtDNA genotype that profoundly affects the response to antibiotics. Beyond that simple and important conclusion, however, subsequent studies have shown that the situation is complicated. One complication is that there are a signiﬁcant number of 1555 family members who suffer hearing loss but who have not received aminoglycoside antibiotics [11,14,15]. At this point, before further discussion, we need to review brieﬂy the pathophysiology of the 1555 mtDNA mutation. One question is why systemic administration of a drug to those who carry an mtDNA mutation in all of their cells and tissues develop a speciﬁc pathology, loss of hearing. Pharmacokinetic studies provide part of the answer in that aminoglycoside antibiotics are cleared rapidly from the body, except for the perilymph and endolymph of the inner ear, where they persist for long periods of time [17]. Second, aminoglycoside antibiotics work by binding to bacterial ribosomes, thereby inhibiting protein synthesis. Mitochondrial ribosomes (but not those in the cytoplasm, which have nuclear gene-encoded rRNAs) become more sensitive to aminoglycosides as a result of the 1555 mutation. The nucleotide at position 1555 is part of the aminoacyl-tRNA decoding region of the mitochondrial ribosome, which is the site of codon–anticodon interaction and the site where aminoglycoside antibiotics bind to bacterial ribosomes [11,14,15]. It is thus proposed that in those persons who carry the 1555 mutation, the antibiotics inhibit mitochondrial protein synthesis in the cochlea, and this inhibition leads to tissue damage and deafness. It is still not yet understood why it is the hair cells and cochlear neurons of the inner ear that are so sensitive to antibiotic inhibition [11,15]. The 1555 story is far from complete, however. Some 1555 family members never manifest clinical deafness, while others lose hearing without exposure to aminoglycosides. The current view [11,14,15] is that the 1555 mutation is necessary, but not sufﬁcient, for hearing loss. It is a major risk factor, but other, secondary pathogenic factors are required for cochlear damage and hearing loss. Clearly, environmental factors are important, and there might well be others besides aminoglycoside antibiotic treatment. Substantial effort has gone into the identiﬁcation of other genetic risk factors, both nuclear and mitochondrial. As with most complex diseases, there have been numerous studies, but many of the putative risk factors reported have not been validated subsequently. One example of a genetic modiﬁer system will sufﬁce to highlight the challenges. Guan et al. [18] have reported on the possible role of a mutation in the nuclear gene TRMU , which encodes an RNA-modifying enzyme that is involved in

130

PHARMACOGENETICS OF MITOCHONDRIAL DRUG TOXICITY

modifying the wobble base in certain mitochondrial tRNAs. These investigators showed that this mutation, when homozygous, reduced both the steady-state levels of those mitochondrial tRNAs and the level of mitochondrial protein synthesis in lymphoblastoid cell lines. They concluded that a combination of the homozygous TRMU mutation and the mitochondrial 1555 mutation was sufﬁcient to lead to hearing loss. The genetic screening results, however, were less compelling than the biochemical studies. Studies of maternal pedigrees with the 1555 mutation provided support for their model in the case of TRMU homozygotes, but many affected family members did not carry the TRMU mutation or were heterozygous [18]. Furthermore, the TRMU mutation did not occur at detectable frequencies in the Chinese 1555 families and controls. As a result, it was suggested that other nuclear genetic modiﬁer loci, not yet identiﬁed, were also involved. Only a fraction of all cases of aminoglycoside-induced ototoxicity are associated with the mitochondrial 1555 mutation [11,14,15], thus leading to the search for additional mtDNA mutations that might be involved. For example, Zhao et al. [19] identiﬁed a pathogenic mutation at mtDNA site 1494 in a large Chinese family, and modeling studies suggested that the nucleotide at site 1494 in the folded mitochondrial 12S rRNA molecule base pairs with the nucleotide at position 1555. A number of other candidate mtDNA mutations have been reported (reviewed in [11] and [15]). However, and this is extremely important, Yao et al. [20] have undertaken a careful analysis of several of these studies and shown that the quality of the mtDNA sequences is not of the highest standard and that, as a result, many of these candidate mutations are questionable. It is thus clear from the studies of the 1555 mtDNA mutation that a mitochondrial genotype can affect drug sensitivity. Mitochondrial pharmacogenetics, as a starting point, would involve comparative sequence analysis of cohorts of different drug responders to identify candidate mtDNA sites that affect response. Such studies will not be simple, but they are likely to become increasingly necessary, especially for drugs with known mitochondrial mechanisms of action. 3. DISPUTED ROLE OF 16189 mtDNA POLYMORPHISM IN TYPE 2 DIABETES Insulin production in pancreatic beta cells depends on mitochondrial energy production [21–23]. As a corollary, abnormal mitochondrial energy metabolism appears to play a key role both in type 2 diabetes mellitus (T2D), which affects more than 100 million people on a worldwide basis, and in insulin resistance, an early stage in T2D pathogenesis [24–28]. T2D is a complex disorder, but it is known that a speciﬁc pathogenic mtDNA mutation at nucleotide position 3243 is causative in a subset of T2D patients that totals 1 to 2% of all cases in populations of European descent (reviewed in [29]). Important but unanswered questions about the etiology of T2D are (1) whether other mtDNA mutations are pathogenic and (2) whether, in aggregate, mtDNA mutations make a major contribution to T2D etiology and pathogenesis.

DISPUTED ROLE OF 16189 mtDNA POLYMORPHISM IN TYPE 2 DIABETES

131

For several years, Poulton and co-workers have investigated the pathogenic role of a sequence change in the mtDNA noncoding control region. One of the most common substitutions in the human mtDNA is a T:C transition at nucleotide position 16189, which generates a homopolymeric C-repeat that is 10 residues in length (termed the 16189 variant). This repeat undergoes frequent expansion and contraction, and as a result, individuals carry mtDNAs with differing length variants, a condition known as heteroplasmy [30,31]. The 16189 variant occurs in about 10% of persons of European descent (e.g., [32]). A brief recap of the main results from Poulton and co-workers illustrates both the importance and challenges of unraveling the pathogenic effects of the 16189 variant as a risk factor in complex human diseases. 1. Poulton et al. [33] reported that the 16189 variant was associated with higher fasting insulin levels in persons drawn from a single population in the UK. An association of the variant with T2D or with impaired glucose tolerance (IGT) did not reach statistical signiﬁcance, and there was no association found with insulin secretion, birth weight, or infant weight (as single variables). A signiﬁcant association with higher fasting insulin levels, after controlling for age, gender, and body mass index (BMI), was obtained subsequently in another of their studies [34]. 2. In analyses of subpopulations from Europe, Asia, and Africa, a signiﬁcant association between the 16189 variant and T2D was obtained [35,36]. For example, in an analysis of T2D patients (n = 463) and nondiabetic controls (n = 469) from Cambridgeshire, UK, the 16189 variant was carried by 9.9% and 6.4%, respectively, of the members of these two groups [36]. This difference just reaches statistical signiﬁcance (the authors cite a p value of 0.048 for the association). 3. Casteels et al. [37] reported that the 16189 variant was slightly, but significantly, associated with a lower ponderal index, a measure of adiposity at birth, although the association was lost by the age of 2 years. The authors suggested that the variant might increase maternal restraint on fetal growth and thereby increase the risk of developing insulin resistance and T2D. More recently, Parker et al. [38] reported that the 16189 variant in an Australian cohort showed a signiﬁcant association with a lower maternal BMI value, a lower BMI value at the age of 20 in their offspring, and increased placental weight, placental/birth weight ratio, and the length of gestation. In this study, however, there was no association with the ponderal index. 4. Khogali et al. [39] reported that the 16189 variant was associated with sporadic dilated cardiomyopathy. We summarize here our preliminary case–control analysis of the association of the 16189 length variant with T2D and IGT in control and patient cohorts from the United States. These experiments were carried out at MitoKor Inc. (San Diego, California) with Christen Anderson and in collaboration with Jerrold Olefsky

132

PHARMACOGENETICS OF MITOCHONDRIAL DRUG TOXICITY

(Veterans’ Administration Medical Center, University of California at San Diego). A more detailed report is in preparation and will be published elsewhere. A total of 378 subjects were analyzed: 1. Lean controls normoglycemic with a BMI value below or equal to 27 (n = 98) 2. Obese controls that were normoglycemic but with a BMI value above 27 (n = 99) 3. T2D patients (n = 106) 4. Persons with impaired glucose tolerance (IGT; n = 75). The complete sequence of the mtDNA from these subjects was determined as reported previously [40]. The haplogroup for each mtDNA was determined using standard criteria and characteristic polymorphisms, and the mtDNAs were classiﬁed as European, Asian/Native American, and African, irrespective of the ethnic self-identity of the subject. All subjects were characterized as well for their age at the time that blood was drawn for DNA preparation, BMI, and gender. For these studies we analyzed the frequencies of the T16189C polymorphism itself and the 16189 sequence variant. The former involves only the nucleotide sequence at nucleotide position 16189, as there can be other nucleotide variants in this region, whereas the latter is limited to mtDNAs that carry this polymorphism and no other polymorphisms that disrupt the T16189C-generated simple repeat. Poulton and co-workers focused on the 16189 sequence variant because it undergoes contraction or expansion. The point is that there are a substantial proportion of 16189C mtDNAs that do not undergo expansion or contraction because of these proximate sequence variants. As a result, we can, in theory at least, distinguish between pathogenic effects due speciﬁcally to the allele state at nucleotide position 16189 and those due instead to an unstable simple repeat sequence that undergoes expansion or contraction. The frequencies of the T16189C polymorphism and the 16189 sequence variant in the study groups are shown in Table 1. As found by other investigators, the T16189C polymorphism is very common, a point that is discussed below. The frequency of the T16189C variant is highest in African mtDNAs, slightly lower in Asian mtDNAs, and lowest in European mtDNAs. However, a relatively high proportion of the African T16189C mtDNAs carry a second mutation that disrupts the simple repeat sequence and thereby reduces expansion or contraction. As a result, the frequency of the 16189 variant (to use the terminology of Poulton and co-workers) is highest in Asian mtDNAs (Table 1). We next subjected these frequencies to statistical tests to determine if there was an association with mtDNA genotype in these groups. There was no association of either the T16189C polymorphism or the 16189 sequence variant with T2D, IGT, or with the combined cohort of T2D and IGT patients (Table 2). We also observed that there were no differences between the lean and obese control groups, as well as no signiﬁcant associations with the pooled control group. Because our groups included individuals with European, Asian/Native

133

DISPUTED ROLE OF 16189 mtDNA POLYMORPHISM IN TYPE 2 DIABETES

TABLE 1 Group

Distributions of the T16189C Polymorphism and the 16189 Variant

a

T16189C

Lean controls Europeans Asians/Native Africans Obese controls Europeans Asians/Native Africans T2D Europeans Asians/Native Africans IGT Europeans Asians/Native Africans All controls Europeans Asians/Native Africans T2D + IGT Europeans Asians/Native Africans

Americans

Americans

Americans

Americans

Americans

Americans

25/96 10/49 7/25 8/21 30/101 5/38 16/37 9/20 30/106 11/57 10/29 9/20 19/75 10/56 6/13 3/5 55/197 15/87 23/62 17/41 49/181 21/113 16/42 12/25

(26%) (20%) (28%) (38%) (30%) (12%) (43%) (45%) (28%) (19%) (34%) (45%) (25%) (18%) (46%) (60%) (28%) (17%) (37%) (41%) (27%) (19%) (38%) (48%)

16189 Variant 16/96 8/49 7/25 1/21 23/101 5/38 16/37 2/20 18/106 6/57 10/29 2/20 18/75 10/56 6/13 2/5 39/197 13/87 23/62 3/41 36/181 16/113 16/42 4/25

(17%) (16%) (28%) (5%) (23%) (12%) (43%) (10%) (17%) (11%) (34%) (10%) (24%) (18%) (46%) (40%) (20%) (15%) (37%) (7%) (20%) (14%) (38%) (16%)

a The mtDNAs were classiﬁed as European, Asian/Native American, or African based on characteristic polymorphisms in the coding region [40]. There were a few “other” mtDNAs that could not be classiﬁed, thus explaining why, depending on the group, the sum of the numbers in the three major ethnic groups is less than the total number.

TABLE 2

Statistical Tests of Association with T2D or IGT

Comparison Lean/obese T2D/IGT Controls/T2D Controls/IGT Lean/T2D Obese/T2D Lean/IGT Obese/IGT Controls/T2D + IGT Controls/T2D + IGT (Europeans)b Controls/T2D + IGT (non-Europeans)b a b

T16189Ca 0.50 < p < 0.75 0.75 < p < 0.90 0.975 < p < 0.99 0.50 < p < 0.75 p ∼ 0.95 p ∼ 0.95 0.995 < p 0.50 < p < 0.75 0.975 < p < 0.99 p ∼ 0.95 0.95 < p < 0.975

16189 Varianta 0.10 < p < 0.25 0.25 < p < 0.50 0.50 < p < 0.75 0.50 < p < 0.75 p ∼ 0.95 0.25 < p < 0.50 0.25 < p < 0.50 p ∼ 0.99 p ∼ 0.90 0.95 < p < 0.975 0.50 < p < 0.75

The p values were determined by 2 × 2 χ2 tests (corrected for continuity). These association tests did not include the mtDNAs that fell into the “other” haplogroup category.

134

PHARMACOGENETICS OF MITOCHONDRIAL DRUG TOXICITY

American, and African mtDNAs, we also carried out analyses with European and non-European subgroups. Again, we found no signiﬁcant associations. Although these results do not agree with those of Poulton et al. [33,35,36], they do agree with other recent studies. Mohlke et al. [41] analyzed groups drawn from the Finnish population. Neither the 16189 polymorphism nor the sequence variant was associated with T2D, although associations were obtained with reduced ponderal index at birth, higher fasting insulin levels, and with reduced birth weight. These authors also showed that there was no preferential maternal inheritance of T2D in this population. Chinnery et al. [32] have recently analyzed the Warren 2 cohort of T2D cases in the UK, and they also found no association with the 16189 variant. Moreover, they also carried out a metaanalysis of more than 1400 T2D patients and more than 3100 controls and again found no association. Saxena et al. [42] analyzed a sample cohort from more than 6000 T2D patients and controls collected from populations of European origin and also found no association of the 16189 variant with T2D. Moreover, they found no association of T2D with any of the more than 60 mtDNA coding region single nucleotide polymorphisms that they used for screening. On the other hand, some groups continue to report a pathogenic role for the 16189 variant. For example, Weng et al. [43] reported a signiﬁcant association of the T16189C polymorphism with T2D and metabolic syndrome in populations of Chinese descent, as has Liou et al. [44] for a Taiwanese population of Chinese descent. Finally, Bhat et al. [45] studied two North Indian populations and concluded that the 16189 variant was signiﬁcantly associated with T2D in one of the two populations as well as in the pooled sample set. They also reported a signiﬁcant association with a polymorphism in the mtDNA coding region. There are two general explanations for the discrepancies among these studies. The ﬁrst possibility is that the 16189 variant has no pathogenic role and that the signiﬁcant associations obtained were a consequence of the small samples analyzed. This is a common limitation of genetic association studies [46,47], and it is telling that the larger analyses and metaanalyses ﬁnd no association between T2D and the 16189 variant. One can extrapolate from power studies such as that of Samuels et al. [48] to conclude that analyses of mtDNA sequence changes will need to be very large to detect signiﬁcant associations with a disease (or with differences in response to a drug). The second possibility is that the 16189 variant is a risk factor for T2D but that it acts in concert with other environmental or genetic risk factors. According to this scenario, each study, both positive and negative, is true and reﬂects the particular combination of risk factors in the study population analyzed. The pathogenic role of the 16189 variant, however, is “washed out” when different populations are pooled (e.g., [32] and [42]). One need only point to the complexities of the 1555 mutation and mitochondrial deafness to understand why this possibility merits consideration. It might also be germane that it seems that a signiﬁcant association for the 16189 variant has been reported more consistently in populations of Asian ethnicity than for European populations. The obvious

CONCLUSIONS

135

problem is that at least at present, such a disease model cannot be falsiﬁed by standard methods of study. We tend to favor the ﬁrst possibility and do not see that the 16189 polymorphism is a risk factor in T2D. That conclusion is based in part on some additional considerations that have not been discussed adequately previously. In the ﬁrst place it is very difﬁcult to understand how a sequence change in the noncoding control region can lead to a speciﬁc pathogenic condition, T2D. Second, there is the 16189 mutation itself. Brieﬂy summarizing our work from other studies [49], we have shown that this mutation occurs very rapidly during human evolution and that it shows no evidence of being affected by positive or negative selection. This is not the “genetic signature” we would expect from a pathogenic mutation, but its unusual evolutionary behavior might cause some complex effects within populations and thereby contribute to the discrepant association studies.

4. CONCLUSIONS The two areas reviewed here provide some important guidelines for the emerging ﬁeld of mitochondrial pharmacogenetics. 1. It all starts with the quality of the experimental data and the critical analysis of the results. Nucleotide sequencing of mtDNA is now a straightforward and technically simple operation. It is also key to identifying mtDNA sites that are primarily pathogenic, secondary risk factors, drug-response loci, and so on. Yet there are many studies in which the results are questionable because of rather obvious ﬂaws in the mtDNA sequences reported (e.g., [20]). A low frequency of sequence errors is probably unavoidable, at least to some extent in large sequencing efforts, but quality control measures must be in place for each study. 2. Power your mitochondrial pharmacogenetic studies appropriately. Although we cannot rule out very complex disease models, the association of the 16189 variant with T2D is just one example of a larger problem whereby small studies jump to a conclusion that does not hold up to subsequent investigation (see [50] for another recent example). This is not a problem that is helped by the pressure to publish exciting results or by the publication bias against “negative” results. 3. Avoid post hoc analysis. There is a tendency for studies to characterize study groups for multiple parameters and then focus on the signiﬁcant associations found, typically with no correction for what are actually multiple statistical comparisons. For example, shouldn’t it have raised a concern when the same investigators found a signiﬁcant association of ponderal index with the 16189 variant in one study population [37] but not another [38]? Although we focused our comments on the association with T2D, the studies cited reported associations on several different clinical parameters.

136

PHARMACOGENETICS OF MITOCHONDRIAL DRUG TOXICITY

4. Finally, do not forget that mitochondria are complex whether one considers structure, function, genetics, or their role in disease. We were able here only to brieﬂy summarize two examples of complex human mitochondrial genetics, but they are the norm and not the exception. Mitochondrial pharmacogenetics is just beginning, but it is safe to predict that much will be learned and that eventually, those efforts will lead to better drugs with greater speciﬁcity and less toxicity.

REFERENCES 1. Amacher DE. Drug-associated mitochondrial toxicity and its detection. Curr Med Chem. 2005;12:1829–1839. 2. Scatena R, Bottoni P, Botta G, Martorana GE, Giardina B. The role of mitochondria in pharmacotoxicology: a reevaluation of an old, newly emerging topic. Am J Physiol Cell Physiol. 2007;293:C12–C21. 3. Dykens JA, Will Y. The signiﬁcance of mitochondrial toxicity testing in drug development. Drug Discov Today. 2007;12:777–785. 4. Liguori MJ, Anderson MG, Bukofzer S, et al. Microarray analysis in human hepatocytes suggests a mechanism for hepatotoxicity induced by trovaﬂoxacin. Hepatology. 2005;41:177–186. 5. Alesci S, Manoli I, Michopoulos VJ, et al. Development of a human mitochondriafocused cDNA microarray (hMitChip) and validation in skeletal muscle cells: implications for pharamco- and mitogenomics. Pharmacogenom J. 2006;6:333–342. 6. Cornwell PD, Ulrich RG. Investigating the mechanistic basis for hepatic toxicity induced by an experimental chemokine receptor 5 (CCR5) antagonist using a compendium of gene expression proﬁles. Toxicol Pathol. 2007;35:576–588. 7. Will Y, Hynes J, Ogurtsov VI, Papkovsky DB. Analysis of mitochondrial function using phosphorescent oxygen-sensitive probes. Nat Protocols. 2006;1:2563–2572. 8. Nadanaciva S, Bernal A, Aggeler R, Capaldi R, Will Y. Target identiﬁcation of drug induced mitochondrial toxicity using immunocapture based OXPHOS activity assays. Toxicol in Vitro. 2007;21:902–911. 9. Herrnstadt C, Howell N. An evolutionary perspective on pathogenic mtDNA mutations: haplogroup associations of clinical disorders. Mitochondrion. 2004;4:791–798. 10. Howell N, Elson JL, Chinnery PF, Turnbull DM. mtDNA mutations and common neurodegenerative disorders. Trends Genet. 2005;21:583–586. 11. Fischel-Ghodsian N. Mitochondrial deafness. Ear Hear. 2003;24:303–313. 12. Chinnery PF, Howell N, Lightowlers RN, Turnbull DM. Molecular pathology of MELAS and MERRF: the relationship between mutation load and clinical phenotypes. Brain. 1997;120:1713–1721. 13. Prezant TR, Agapian JV, Bohlman MC, et al. Mitochondrial ribosomal RNA mutation associated with both antibiotic-induced and non-syndromic deafness. Nat. Genet. 1993;4:289–294.

REFERENCES

137

14. Hutchin T. Sensorineural hearing loss and the 1555G mitochondrial DNA mutation. Acta Otolaryngol. 1999;119:48–52. 15. Guan M-X. Mitochondrial DNA mutations associated with aminoglycoside ototoxicity. Audiolog Med. 2006;4:170–178. 16. Tang H-Y, Hutcheson E, Neill S, Drummond-Borg M, Speer M, Alford RL. Genetic susceptibility to aminoglycoside ototoxicity: How many are at risk? Genet Med. 2002;4:336–345. 17. Henley CM, Schacht J. Pharmacokinetics of aminoglycoside antibiotics in blood, inner-ear ﬂuids and tissues and their relationship to ototoxicity. Audiology. 1988;27:137–146. 18. Guan M-X, Yan Q, Li X, Bykhovskaya Y, et al. Mutation in TRMU related to transfer RNA modiﬁcation modulates the phenotypic expression of the deafness-associated mitochondrial 12S ribosomal RNA mutations. Am J Hum Genet. 2006;79:291–302. 19. Zhao H, Li R, Wang Q, et al. Maternally inherited aminoglycoside-induced and nonsyndromic deafness is associated with the novel C1494T mutation in the mitochondrial 12S rRNA gene in a large Chinese family. Am J Hum Genet. 2004;74:139–152. 20. Yao Y-G, Salas A, Bravi CM, Bandelt H-J. A reappraisal of complete mtDNA variation in East Asian families with hearing impairment. Hum Genet. 2006;119:505–515. 21. Maechler P, Wollheim CB. Mitochondrial function in normal and diabetic beta-cells. Nature. 2001;414:807–812. 22. Leibowitz G, Khaldi MZ, Shauer A, et al. Mitochondrial regulation of insulin production in rat pancreatic cells. Diabetologia. 2005;48:1549–1559. 23. Pagliarini DJ, Wiley SE, Kimple ME, et al. Involvement of a mitochondrial phosphatase in the regulation of ATP production and insulin secretion in pancreatic β cells. Mol Cell. 2005;19:197–207. 24. Mootha VK, Lindgren CM, Eriksson K-F, et al. PGC-1α-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003;34:267–273. 25. Patti ME, Butte AJ, Crunkhorn S, et al. Coordinated reduction of genes of oxidative metabolism in humans with insulin resistance and diabetes: Potential role of PGC1 and NRF1 . Proc Natl Acad Sci U S A 2003;100:8466–8471. 26. Petersen KF, Dufour S, Befroy D, Garcia R, Shulman GI. Impaired mitochondrial activity in the insulin-resistant offspring of patients with type 2 diabetes. N Engl J Med. 2004;350:664–671. 27. Morino K, Petersen KF, Dufour S, et al. Reduced mitochondrial density and increased IRS-1 serine phosphorylation in muscle of insulin-resistant offspring of type 2 diabetic parents. J Clin Invest. 2005;115:3587–3593. 28. Petersen KF, Dufour S, Shulman GI. Decreased insulin-stimulated ATP synthesis and phosphate transport in muscle of insulin-resistant offspring of type 2 diabetic patients. PLoS Med. 2005;2:e233. 29. Barrett TG. Mitochondrial diabetes, DIDMOAD and other inherited diabetes syndromes. Best Pract Res Clin Endocrinol Metab. 2001;15:325–343. 30. Bendall KE, Sykes BC. Length heteroplasmy in the ﬁrst hypervariable segment of the human mtDNA control region. Am J Hum Genet. 1995;57:248–25. 31. Howell N, Bogolin Smejkal C. Persistent heteroplasmy of a mutation in the human mtDNA control region: hypermutation as an apparent consequence of simple-repeat expansion/contraction. Am J Hum Genet. 2000;66:1589–1598.

138

PHARMACOGENETICS OF MITOCHONDRIAL DRUG TOXICITY

32. Chinnery PF, Elliott HR, Patel S, et al. Role of the mitochondrial DNA 16184– 16193 poly-C tract in type 2 diabetes. Lancet. 2005;366:1650–1651. 33. Poulton J, Scott Brown M, Cooper A, Marchington DR, Phillips DIW. A common mitochondrial DNA variant is associated with insulin resistance in adult life. Diabetologia. 1998;41:54–58. 34. Poulton J, Bednarz AL, Scott-Brown M, Thompson C, Macaulay VA, Simmons D. The presence of a common mitochondrial DNA variant with fasting insulin levels in Europeans in Auckland. Diabetes Med. 2002;19:969–971. 35. Poulton J. Does a common mitochondrial DNA polymorphism underlie susceptibility to diabetes and the thrifty genotype? Trends Genet. 1998;14:387–389. 36. Poulton J, Luan J, Macaulay V, Hennings S, Mitchell J, Wareham NJ. Type 2 diabetes is associated with a common mitochondrial variant: evidence from a population-based case–control study. Hum Mol Genet. 2002;11:1581–1583. 37. Casteels K, Ong K, Phillips D, Bendall H, Pembrey M, the ALSPAC study team, Poulton J, Dunger D. Mitochondrial 16189 variant, thinness at birth, and type-2 diabetes. Lancet. 1999;353:1499–1500. 38. Parker E, Phillips DIW, Cockington RA, Cull C, Poulton J. A common mitochondrial DNA variant is associated with thinness in mothers and their 20 year old offspring. Am J Physiol Endocrinol Metab. 2005;289:E1110–E1114. 39. Khogali SS, Mayosi BM, Beattie JM, McKenna WJ, Waitkins H, Poulton J. A common mitochondrial variant associated with susceptibility to dilated cardiomyopathy in two different populations. Lancet. 2001;357:1265–1267. 40. Herrnstadt C, Elson JL, Fahy E, et al. Reduced-median-network analysis of complete mitochondrial DNA coding-region sequences for the major African, Asian, and European haplogroups. Am J Hum Genet. 2002;70:1152–1171. 41. Mohlke KL, Jackson AU, Scott LJ, et al. Mitochondrial polymorphisms and susceptibility to type 2 diabetes-related traits in Finns. Hum Genet. 2005;118:245–254. 42. Saxena R, de Bakker PIW, Singer K, et al. Comprehensive association testing of common mitochondrial DNA variation in metabolic disease. Am J Hum Genet. 2006;79:54–61. 43. Weng S-W, Liou C-W, Lin T-K, et al. Association of mitochondrial deoxyribonucleic acid 16189 variant (T → C transition) with metabolic syndrome in Chinese adults. J Clin Endocrinol Metab. 2005;90:5037–5040. 44. Liou C-W, Lin T-K, Weng HH, et al. A common mitochondrial DNA variant and increased body mass index as associated factors for development of type 2 diabetes: additive effects of genetic and environmental factors. J Clin Endocrinol. Metab. 2007;92:235–239. 45. Bhat A, Koul A, Sharma S, et al. The possible role of 10398A and 16189C mtDNA variants in providing susceptibility to T2DM in two North Indian populations: a replicative study. Hum Genet. 2007;120:821–826. 46. Ioannidis JPA, Trikalinos TA, Ntzani EE, Contopoulos-Ioannidis DG. Genetic associations in large versus small studies: an empirical assessment. Lancet. 2003;361:567–571. 47. Trikalinos TA, Ntzani EE, Contopoulos-Ioannidis DG, Ioannidis JPA. Establishment of genetic associations for complex diseases in independent of early study ﬁndings. Eur J Hum. Genet. 2004;12:762–769.

REFERENCES

139

48. Samuels DC, Carothers AD, Horton R, Chinnery PF. The power to detect disease associations with mitochondrial DNA haplogroups. Am J Hum Genet. 2006;78:713–720. 49. Howell N, Elson JL, Howell C, Turnbull DM. Relative rates of evolution in the coding and control regions of African mtDNAs. Mol Biol Evol. 2007;24:2213–2221. 50. Pereira L, Goncalves J, Franco-Duarte R, et al. No evidence for an mtDNA role in sperm motility: data from complete sequencing of asthenozoospermic males. Mol Biol Evol. 2007;24:868–874.

PART II ORGAN DRUG TOXICITY: MITOCHONDRLAL ETIOLOGY

141

5 FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY Dominique Pessayre, Alain Berson, and Bernard Fromenty ´ INSERM, Centre de Recherche Biom´edicale Bichat Beaujon, Equipe Mitochondries et Foie, Paris; Facult´e de M´edecine Xavier Bichat, Universit´e Paris 7, Paris, France

1. Introduction 2. General features of DILI 2.1. Frequency of DILI 2.2. Legal and financial implications of DILI 2.3. Difﬁcult prediction of DILI before marketing 2.4. Difﬁcult avoidance of severe DILI after marketing 2.5. Diversity of DILI 2.6. Diagnosis of DILI 2.7. Avoidance of inadvertent rechallenges 2.8. Resumption of treatment 2.9. Two main mechanisms of DILI 3. Reactive metabolite–mediated mitochondrial disruption 3.1. Direct toxicity 3.2. Immune reactions 3.3. Tolerance 4. Parent drug–mediated permeability transition 4.1. Anionic uncouplers 4.2. Other drugs triggering permeability transition 5. Primary impairment of mitochondrial β-oxidation 5.1. Fat removal from the liver 5.2. Drug-induced steatosis 5.3. Tetracyclines 5.4. Valproic acid 5.5. Aspirin and Reye’s syndrome

144 145 145 145 146 147 147 148 149 149 150 150 150 154 157 157 158 162 163 163 163 165 166 167

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

143

144

6.

7. 8. 9.

10.

11.

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

5.6. Female sex hormones and the acute fatty liver of pregnancy 5.7. NSAIDs having a 2-arylpropionate structure 5.8. Glucocorticoids 5.9. Amineptine and tianeptine 5.10. Calcium hopantenate, panadiplon, and pivampicillin Primary impairment of both β-oxidation and respiration 6.1. Amiodarone, 4,4 -diethyaminoethoxyhexestrol, and perhexiline 6.2. Tamoxifen 6.3. Buprenorphine 6.4. Antimalarial drugs 6.5. Benzarone and benzbromarone Inhibition of ATP synthase Inhibition of the adenine nucleotide translocator Interference with mitochondrial DNA and/or mitochondrial transcripts 9.1. Degradation of mtDNA by alcohol 9.2. Degradation of mtDNA by acetaminophen (paracetamol) 9.3. Impairment of mtDNA replication by drugs inhibiting topoisomerases and/or binding to DNA 9.4. Impairment of mtDNA replication by 2 ,3 -dideoxynucleosides and abacavir 9.5. Impairment of mtDNA replication by fialuridine and ganciclovir 9.6. Decreased synthesis and stability of mitochondrial transcripts in cells treated with interferon-α 9.7. Decreased translation of mitochondrial transcripts into proteins Mechanisms behind idiosyncrasy 10.1. Metabolic factors 10.2. Co-morbidity factors Conclusions

169 169 170 170 170 170 171 172 173 173 173 173 174 174 176 177 178 180 183 184 185 185 185 186 186

1. INTRODUCTION More than 1000 drugs are hepatotoxic, and these drugs can trigger diverse types of liver disease, reproducing the entire spectrum of liver pathology [1]. Drug-induced liver injury (DILI) is therefore a major concern for both the pharmaceutical industry and for physicians [1]. The two most frequent mechanisms responsible for DILI are the formation of reactive metabolites [2,3] and drug-induced mitochondrial dysfunction [4,5]. In this chapter we ﬁrst recall some general features of DILI. We then consider how reactive metabolites can cause either toxic hepatitis or immunoallergic hepatitis, with special emphasis on the role of mitochondrial disruption as a ﬁnal mechanism of cell death. Finally, we describe several ways whereby the parent drug can disturb mitochondrial function and trigger diverse forms of DILI.

GENERAL FEATURES OF DILI

145

2. GENERAL FEATURES OF DILI 2.1. Frequency of DILI Considering all causes of liver diseases, DILI is relatively uncommon, coming well after obesity/diabetes, viral hepatitis, or alcohol abuse. It is estimated that DILI may be responsible for about 9% of the cases of liver test abnormalities [6]. In a population-based study in France, the incidence rate of DILI was 14 cases for 100,000 inhabitants per year [7]. Interestingly, the number of cases reported to the French pharmacovigilance agency was 16 times lower, emphasizing the magnitude of underreporting [7]. Although DILI is relatively uncommon in young patients, its prevalence increases markedly in old age, due primarily to increased use of medications in elderly patients and polypharmacy [1]. Older patients often suffer from different ailments and therefore consult diverse specialists, each prescribing several medications. Whatever the age, DILI has a disproportionate etiological role in fulminant hepatitis [8]. In this severe but fortunately rare condition, the acute destruction of a large proportion of the hepatocytes leads to jaundice and hepatic encephalopathy, and can lead to death unless liver transplantation is performed. The high frequency of DILI (52%) as a cause of fulminant hepatitis is due primarily to a large contribution of acetaminophen intoxication (40%), whereas all the other drugs together cause only 12% of the cases of fulminant hepatitis [8]. In the United States about half of acetaminophen-related cases of fulminant hepatitis involve massive overdoses taken in an attempt at suicide. In the other half, excessive therapeutic doses are taken for pain relief by patients who are unaware of the safe upper limit of acetaminophen tablets, disregard this limit, or inadvertently associate a variety of over-the-counter analgesics, which, unbeknown to them, all contain acetaminophen [8]. 2.2. Legal and Financial Implications of DILI DILI is a major problem for the pharmaceutical industry, because it is a frequent cause of the failure of drug molecules to get approved for human use, and a frequent cause for court litigations and/or drug withdrawal after marketing. DILI is also a major concern for physicians, who can cure their patients simply by withdrawing the offending drug, whereas failure to make the diagnosis and continuation of the treatment can lead to either one of three alternatives [1]: 1. An asymptomatic patient with a mild increase in serum transaminases can spontaneously adapt to the drug and show ﬁrst improving, and then normal, liver tests. 2. However, another patient, initially with the same liver test proﬁle, can quickly deteriorate and develop fulminant acute hepatitis. 3. Finally, a third patient with the same initial liver test proﬁle can sustain prolonged, asymptomatic liver injury as long as the administration of the

146

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

offending drug is being continued. Although this protracted infraclinical liver injury may have no serious health consequences in most patients, it can lead to chronic liver disease, such as chronic hepatitis or steatohepatitis in a few patients. 2.3. Difﬁcult Prediction of DILI Before Marketing Preclinical Preclinical studies could easily look for extensive formation of reactive metabolites or for mitochondrial toxicity of drugs used at the anticipated therapeutic concentrations. However, these types of studies are not demanded by the various regulatory agencies, and therefore are not performed routinely by drug companies. Clinical Despite these difﬁculties, most drugs which have gone successfully through the preclinical study screen are indeed essentially safe for most recipients. Although they can cause mild liver test abnormalities in a small percentage of recipients, they do not trigger clinically patent DILI, except in a few recipients unlucky enough to have uncommon predisposition factor(s) (either acquired or genetic) that render them susceptible to drug-induced toxicity or immune reactions. However, these predisposing factors are incompletely known and vary with the drug, its metabolism and the mechanism of hepatotoxicity. With rare exceptions, it is therefore difﬁcult to devise trials that would speciﬁcally target susceptible patients. As for the standard clinical trials, they are extremely expensive and are necessarily limited in size. Therefore, standard clinical trials may not include susceptible patients in sufﬁcient numbers. Although they easily disclose asymptomatic DILI, which is relatively frequent, they can fail to show severe hepatotoxic reactions and can therefore miss the actual hepatotoxic potential of some drugs. Hy’s Rule On the other hand, if clinical trials reveal patients with an alanine aminotransferase (ALT) activity more than three times the upper limit of normal (ULN) and a conjugated bilirubin level more than two times the ULN, whereas no such cases have been observed with comparator drugs, it can safely be predicted that the drug will cause severe hepatitis in some patients after being marketed. The above-mentioned criteria correspond mostly, if not exclusively, to cases of cytolytic hepatitis (as suggested by marked ALT elevation), which are severe enough to destroy a large number of hepatocytes, thus causing jaundice (as indicated by conjugated bilirubin at 2 ULN or more). Some 30 years ago, the late and regretted Hyman Zimmerman noticed that the mortality of jaundiced patients with drug-induced cytolytic hepatitis was usually around 10% (with, however, some notable exceptions, such as the much poorer prognosis of halothane- or iproniazid-induced jaundice) [9]. This rule has become affectionately known as “Hy’s rule.” If cases ﬁtting Hy’s rule are observed with a frequency of perhaps 1 in 1000 during clinical trials, one may

GENERAL FEATURES OF DILI

147

expect that about 1 in 10,000 patients may die with liver disease once the drug is marketed. 2.4. Difﬁcult Avoidance of Severe DILI After Marketing Frequent Liver Test Monitoring Even markedly hepatotoxic drugs can sometimes be used, especially when they treat a severe disease for which there are no better treatments. With markedly hepatotoxic drugs, frequent serum liver testing can be recommended and may efﬁciently prevent severe liver injury. For example, mandatory liver tests every 2 weeks for the ﬁrst 6 weeks, and then every 4 weeks, with discontinuation of the treatment whenever serum ALT activity increased to more than 5 ULN, may have contributed to the rarity of tacrine-induced jaundice, despite a high incidence of asymptomatic liver dysfunction with this cholinesterase inhibitor. Infrequent Monitoring For drugs that rarely cause severe liver injury, such as statins or most nonsteroidal anti-inﬂammatory drugs (NSAIDs), frequent liver test monitoring represents an unreasonable imposition on the patient. In this case, infrequent liver test monitoring has often been recommended in the past. However, unlike frequent liver test monitoring, infrequent liver test monitoring is probably useless. Obviously, severe liver injury can develop in the interim. Furthermore, infrequent liver test monitoring tends to be forgotten by patients and physicians alike, and is frequently postponed, sometimes indeﬁnitely. Warning Rather than relying on systematic but infrequent liver test monitoring, it may be best to warn patients of possible, albeit uncommon, adverse effects to the liver. Patients should be advised to quickly consult and undergo liver tests if they feel unwell, and to cease taking the drug immediately if they become jaundiced. 2.5. Diversity of DILI More than 1000 drugs are hepatotoxic [10], and these drugs can damage the liver through various mechanisms to cause a variety of different liver diseases, reproducing the entire spectrum of liver pathology. Hepatitis The most frequent drug-induced liver lesion is acute hepatitis [1]. In all cases of acute hepatitis, there is some hepatic inﬂammation, but ﬁbrosis is absent in this acute form. If performed, a liver biopsy can help classify acute hepatitis into three main types: 1. Cytolytic acute hepatitis associates hepatic inﬂammation with signs of cell death, such as necrosis, apoptosis, and/or cell dropout.

148

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

2. Cholestatic acute hepatitis is characterized by inﬂammation and by brownish bile deposits in the cytoplasm of hepatocytes and within bile canaliculi (a space between two adjacent hepatocytes). In addition, cholangiolitis is sometimes associated. 3. Mixed acute hepatitis associates necrosis and cholestasis, in addition to inﬂammation. Other Liver Lesions DILI can also cause isolated “bland” hepatic cholestasis (without necrosis or inﬂammation), hepatic steatosis (fatty liver), steatohepatitis (with both steatosis and necroinﬂammation), and also subacute hepatitis, chronic hepatitis, cirrhosis, sinusoidal dilation, peliosis, venoocclusive liver disease, Budd–Chiari syndrome, hepatic adenoma, and very rarely, malignant liver tumors [1]. Biochemical Classiﬁcation Although a liver biopsy is often required to ascertain the diagnosis in chronic forms of DILI, this invasive procedure entails some risk and is not required for the diagnosis of most cases of acute DILI, particularly in mild or moderate cases. Although liver histology is typically lacking, it is nevertheless useful to be able to classify these cases from the maximal increase in serum ALT, a marker of hepatocyte damage or cytolysis, and the maximal increase in alkaline phosphatase (AP), a marker of cholestasis, and from the ALT/AP ratio, each activity being expressed in multiples of its own ULN [11]: •

• •

The liver injury is termed cholestatic if only alkaline phosphatase is increased (>2 ULN) or, when both ALT and AP are increased, if the ALT/AP ratio is 2 or less. The liver injury is designated as mixed when both ALT and AP are increased and the ALT/AP ratio is between 2 and 5. The liver injury is classiﬁed as hepatocellular if only ALT is increased (>2ULN) or, when both activities are increased, if the ALT/AP ratio is 5 or more. In this case, however, one cannot fully equate this hepatocellular injury with cytolytic hepatitis, because several other liver lesions, such as steatosis, Budd–Chiari syndrome, venoocclusive disease, low cardiac output, and passage of a gallbladder stone through the common bile duct, among others, can also yield this liver proﬁle.

2.6. Diagnosis of DILI Due to the many different possible aspects of DILI, an adverse reaction should be considered systematically when faced with almost any kind of liver disease. Insistent questioning may disclose use of other drugs, which were not mentioned initially for diverse reasons. The patient may consider these drugs as safe (e.g., herbal remedies, over-the-counter medications) or may feel uneasy about acknowledging their use (e.g., antipsychotic drugs, antidepressant drugs, analgesic drugs, NSAIDs, hypnotics, or illicit drugs such as ecstasy). Finally, elderly

GENERAL FEATURES OF DILI

149

patients with faltering memories may have difﬁculty remembering multiple, oddly named, and frequently changing medicines. A diagnosis of DILI is suspected in the absence of other causes of liver disease, as indicated by the patient’s history, viral serologies, blood chemistry, and ultrasonography. Compatible chronology is supportive, as is similarity of the patient’s presentation with the cases of liver injury previously reported with the suspected drug. However, new drugs that are not known to be hepatotoxic should not be considered risk-free because the hepatotoxic potential of some drugs is only recognized several years after their marketing. When present, rash and/or blood eosinophilia may reﬂect immunoallergic DILI. A liver biopsy is performed in severe and difﬁcult cases, and is often the only way to ascertain the diagnosis when chronic liver disease is suspected. However, a liver biopsy entails some risk and is rarely performed in mild, acute cases of DILI. Withdrawing the suspected drug(s) usually improves or cures the liver disease, thus providing a likely diagnosis. Rarely, a speciﬁc test is available to conﬁrm the diagnosis, such as a positive lymphocyte transformation test with the suspected drug [12], or the presence of speciﬁc autoantibodies or antiadduct antibodies [13]. Indeed, relatively speciﬁc antitriﬂuoroacetylated protein antibodies are present in halothane-induced hepatitis, anti-M6 mitochondrial autoantibodies (anti-M6) in iproniazid-induced hepatitis, anti-liver kidney microsomes type 2 (anti-LKM2) autoantibodies in tienilic acid–induced hepatitis, and anti-liver microsomes (anti-LM) autoantibodies in dihydralazine-induced hepatitis [13]. 2.7. Avoidance of Inadvertent Rechallenges Reintroducing the drug for the sake of diagnosis is usually unethical. Instead, when the diagnosis is likely, the patient should be given a note of all pharmaceuticals containing the suspected drug and should be advised against taking these medicines again. The primary physician should be notiﬁed of the DILI diagnosis. However, these precautions are not always effective, and some patients take the offending drug again. Such inadvertent rechallenges can conﬁrm the diagnosis if liver test abnormalities recur after the rechallenge. In contrast, a negative inadvertent rechallenge does not necessarily exclude DILI, because the absence of recurrence may be due to a lower dosage of the drug, a shorter duration of treatment, the use of different co-medications, or differences in the medical condition of the patient, all of which can modulate hepatotoxicity. 2.8. Resumption of Treatment Readministration of a possibly implicated drug is licit only in rare circumstances. Reintroduction may be considered if the responsibility of the drug in the episode of DILI appears unlikely and/or if the drug, albeit possibly involved in DILI, is required to cure a severe disease for which alternative treatments are insufﬁciently active. Even in these cases, rechallenge is best avoided when the drug is thought to cause immune-mediated cytolytic hepatitis, because reintroducing the

150

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

allergen may trigger rapid liver failure in a few patients [1]. In contrast, careful reintroduction may be attempted if DILI is thought to be due to a toxic mechanism, and one can hope that using the drug in lower daily doses or in different circumstances can improve tolerance. When rechallenge is undertaken, liver tests must be performed frequently, thus enabling prompt interruption of the treatment should DILI recur. 2.9. Two Main Mechanisms of DILI Drugs can damage the liver through a variety of mechanisms, the two most frequent of which are the formation of reactive metabolites [2,3] and mitochondrial dysfunction [4,5]. However, even in the frequent case when toxic or immunoallergic hepatitis is due initially to the formation of reactive metabolites, mitochondrial disruption is often involved as a ﬁnal mechanism of cell death. 3. REACTIVE METABOLITE–MEDIATED MITOCHONDRIAL DISRUPTION Several hepatotoxic drugs are transformed by cytochrome P450 into chemically reactive electrophilic metabolites which react spontaneously with, and covalently bind to, hepatic proteins and hepatic glutathione. These reactive metabolites can cause direct toxicity or can lead to immune reactions or to tolerance, depending, in part, on the reactivity and formation rate of the reactive metabolites. 3.1. Direct Toxicity Mechanisms When they are formed in large amounts (e.g., after the ingestion of excessive doses of acetaminophen), reactive metabolites can kill hepatocytes directly through toxic mechanisms [2,3]. This toxicity involves a number of different events [14] (Figure 1): 1. The extensive formation of reactive metabolites can lead to DNA damage, stabilization of p53, sequestration of the antiapoptotic Bcl-XL by p53, p53-mediated induction of the proapoptotic proteins, Bim, PUMA, NOXA, and Bax, and migration of p53 to the mitochondria. 2. Reactive metabolites also deplete hepatic glutathione, and they covalently bind to protein thiols, thus decreasing hepatic protein thiols and inactivating plasma membrane Ca2+ -ATPases. 3. The decreased extrusion of cell calcium from the cell increases the cytosolic concentration of free Ca2+ . 4. Calcium activates several calcium-dependent enzymes (Figure 1), including tissue transglutaminase, which forms a cross-linked protein scaffold, calpain, which severs several proteins involved in the organization and attachment of microﬁlaments, endonuclease, which contributes to DNA fragmentation, and ﬁnally, phospholipase A2, which releases arachidonic acid from membranes.

151

REACTIVE METABOLITE–MEDIATED MITOCHONDRIAL DISRUPTION

HEPATOCYTE

Drug Cytochrome P-450 Reactive metabolite

DNA damage

p53 stabilization

Induced Bim, PUMA, NOXA and Bax Covalent binding

GSH depletion MPT

Decreased protein thiols Ca2+ Inactivated Ca2+-ATPases Transglutaminase Protein cross-links Prot. -NH-CO- Prot.

Increased Ca2+ Calpain

Released cytochrome c Decreased ATP Liberated CAD arachidonic acid

Phospholipase A2 Endonuclease

Disruption of microfilamentassociated proteins

Caspase-9 Caspase-3 Cuts ICAD and other proteins

ICAD

DNA fragmentation Nucleosome NUCLEUS

Figure 1 The direct toxicity of reactive metabolites involves mitochondrial permeability transition (MPT) as a ﬁnal mechanism of cell death. The extensive formation of reactive metabolites by cytochrome P450 may cause glutathione (GSH) depletion, covalent binding to protein thiols, and DNA damage, leading to p53 stabilization and induction of the pro-apoptotic proteins Bim, PUMA, NOXA, and Bax. Furthermore, GSH depletion and covalent binding decrease protein thiols and inactivate plasma membrane Ca2+ -ATPases, therefore increasing cell Ca2+ . The increased cell calcium activates Ca2+ -dependent enzymes, including tissue transglutaminase (forming a cross-linked protein scaffold), calpain (severing proteins involved in the formation and attachments of the microﬁlament network), endonucleases (contributing to DNA fragmentation), and phospholipase A2 (releasing arachidonic acid). The overexpression of pro-apoptotic proteins, the oxidation of protein thiols causing disulﬁde bond formation in the protein structure of the MPT pore, the increase in intramitochondrial Ca2+ , and the released arachidonic acid may all act together to open the MPT pore in some mitochondria. Whereas unaffected mitochondria keep synthesizing ATP, the permeabilized mitochondria release cytochrome c, which activates caspase-9, which in turn activates caspase-3. The latter cuts diverse proteins, including the inhibitor of caspase-activated deoxyribonuclease, thus allowing this nuclease (CAD) to enter the nucleus and fragment DNA.

5. Another major consequence of the increase in cytosolic calcium is to trigger the entry of calcium into the mitochondrial matrix. The entry of calcium into the matrix, the induction of the proapoptotic proteins Bim, PUMA, NOXA, and Bax, the decreased cellular levels of glutathione, and the generation of arachidonic acid may all act together to open the mitochondrial permeability transition pore (MPT).

152

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

6. The opening of the MPT pore, which is located at contact sites between the outer and inner mitochondrial membranes, has two main consequences (Figure 2). (a) Pore opening allows a massive reentry of protons into the mitochondrial matrix, thus bypassing ATP synthase and preventing ATP synthesis. If the pore opens quickly in all mitochondria, a major drop in cell ATP prevents apoptosis, which requires energy. Instead, low ATP levels prevent active ion transports, thus triggering cell swelling, rupture of the plasma membrane and necrotic cell death (Figure 3). (b) Due to the oncotic pressure of matrix proteins, pore opening triggers the inﬂux of water, thus causing mitochondrial matrix swelling (Figure 2). The inner mitochondrial membrane has many folds and can easily accommodate an increased matrix volume without bursting. In contrast, the spherical outer membrane bursts when the mitochondrial

NORMAL Cristae

Respiratory chain

MPT Matrix swelling

ATP synthase

e−

e− H+

H+

H+

Matrix

Closed MPT pore

H+ H+, H2O

Intermembrane space Open MPT pore

Outer membrane rupture

Bypassed ATP synthase (no ATP formation)

Figure 2 Mitochondrial permeability transition prevents ATP formation and induces swelling and outer mitochondrial membrane rupture. Normally, the transfer of electrons along the respiratory chain is associated with the extrusion of protons from the mitochondrial matrix into the intermembrane space. When cells need energy, the reentry of protons into the matrix through ATP synthase then transforms ADP into ATP. However, the opening of the MPT pore has two major consequences. First, this opening causes a massive reentry of protons through the pore, thus bypassing ATP synthase and preventing ATP formation. Second, pore opening allows an inﬂux of water into the matrix driven by the osmotic pressure of matrix proteins. This inﬂux triggers matrix expansion and rupture of the spherical outer membrane. In contrast, the inner membrane, with its many folds, can accommodate the increased matrix volume without bursting.

153

REACTIVE METABOLITE–MEDIATED MITOCHONDRIAL DISRUPTION

NORMAL MITOCHONDRIA

MITO

ATP

MITO

ATP

MPT IN ALL MITOCHONDRIA MPT H+ water

No ATP

MPT H+ water

No ATP

Cell swelling

MPT IN SOME MITOCHONDRIA

MITO

ATP

MPT H+ water

No ATP

Cytochrome c Caspases

Plasma membrane rupture LIVING CELL

NECROSIS

APOPTOSIS

Figure 3 Regulation of life and death by the mitochondrial permeability transition pore. Maintaining the pore in a close state permits cell survival, whereas pore opening can trigger either apoptosis or necrosis. If the MPT pore opens in all mitochondria, severe ATP depletion prevents apoptosis, which is an energy-requiring process. Instead, the lack of ATP prevents active ion transports, causing cell swelling, plasma membrane rupture, and cell death from necrosis. In contrast, if the pore opens in only some mitochondria, the unaffected mitochondria keep synthesizing ATP (thus avoiding necrosis), while the permeabilized mitochondria release cytochrome c, which activates caspases in the cytosol to cause apoptosis.

membrane swells. Rupture of the outer membrane allows the translocation of cytochrome c from the intermembrane space of mitochondria to the cytosol. If the pore opens only in some mitochondria, the unaffected mitochondria continue to generate ATP, which prevents necrosis while the affected mitochondria release cytochrome c. The latter activates caspase-9 in the cytosol, which then activates effector caspases, such as caspase-3, to cause apoptosis (Figure 3). Thus, the extensive formation of reactive metabolites may cause necrosis, apoptosis, or both types of cell death in different liver cells under different circumstances. Clinical Features Clinically, hepatitis due to the direct toxicity of reactive metabolites has the following characteristics [1]: The incidence of hepatitis and its severity are related to the dose; hepatitis is not associated with hypersensitivity manifestations; and after recovery, the readministration of a small dose does not lead to the recurrence of DILI.

154

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

3.2. Immune Reactions Mild Direct Toxicity Formed in intermediate amounts, reactive metabolites are unable to trigger severe toxic hepatitis. However, mild cellular toxicity can still occur in a few susceptible patients, as shown by a mild, clinically silent increase in serum transaminase activity. This mild toxicity allows the release of hepatic proteins, which have been modiﬁed by the covalent binding of reactive metabolites [1] (Figure 4). Antigen-presenting cells may then take up the modiﬁed hepatic proteins, and process these proteins into small peptides, some of which are then presented within the groove of major histocompatibility (MHC) class II molecules expressed on the surface of the antigen-presenting cells. Because some of the amino acids of the initial protein were modiﬁed by the presence of a covalently bound metabolite, the antigen-presenting cell can present not only normal peptides, but also covalently modiﬁed peptides. In most subjects, nothing else may happen. Immunization In a few subjects, however, the haptenized peptide (“modiﬁed self”) presented by the antigen-presenting cell may be recognized by helper T cells (Figure 4) [4,5]. Concomitantly, the mild cell necrosis due to the mild direct

MHC class II Metabolite

Metabolite

Protein

SEC

CD4 + helper T cell

Peptide APC SEC

Help

TCR SEC

CD8 + cytotoxic T cell

TCR MHC class I

KILL

Covalently bound metabolite Hepatocyte destroyed by mild direct toxicity

Still living but doomed hepatocyte

Figure 4 Possible mechanisms for the immune response triggered by reactive metabolites. The mild direct toxicity of reactive metabolites may cause the death of a few hepatocytes. This may allow haptenized hepatic proteins to enter the sinusoid through the fenestrae of sinusoidal endothelial cells (SECs). The uptake of haptenized proteins by Kupffer cells can lead to the presentation of haptenized peptides on the major histocompatibility (MHC) class II molecules of these antigen-presenting cells (APCs). The modiﬁed peptide may then be recognized by the T-cell receptor (TCR) of a helper T cell. The latter may provide help to a cytotoxic T lymphocyte recognizing haptenized peptides presented by MHC class I molecules on the surface of a hepatocyte.

REACTIVE METABOLITE–MEDIATED MITOCHONDRIAL DISRUPTION

155

toxicity of the reactive metabolite may trigger a mild inﬂammatory reaction, which may give “co-stimulatory signals” to the immune system, thus orienting its response toward immunization rather than toleration. Thus, detection of the modiﬁed self combined with the concomitant presence of costimulatory signals may lead to cellular and humoral immune reactions directed against both the modiﬁed parts of proteins or peptides (the neoantigens), and also against unmodiﬁed protein epitopes (autoimmunity) [4,5]. The effector cells can include both B cells, which maturate into antibody-secreting plasmocytes, and cytotoxic T lymphocytes. Lymphocyte-Mediated Mitochondrial Disruption and Apoptosis Cytotoxic T lymphocytes may bind to hepatocytes to destroy them as depicted in Figure 5 [15]. •

In the hepatocytes, the reactive metabolites covalently bind to hepatic proteins, thus modifying these proteins. Like normal cell proteins, the modiﬁed hepatic proteins undergo proteolytic processing, which releases both normal peptides and peptides modiﬁed by the covalent attachment of the metabolite.

Cytotoxic T lymphocyte

TCR

Fas Ligand

TNF-α

TRAIL

Granzyme B

MHC class I Outer membrane permeabilization and MPT

Caspase-8

Peptide Bid Protein

tBid Bax

Bax

MITO

Covalently bound metabolite

HEPATOCYTE

APOPTOSIS

Cytochrome c Caspase-9 Caspase-3

Figure 5 Cytotoxic T lymphocytes kill their targets by inducing outer mitochondrial membrane permeabilization and rupture. The covalent binding of reactive metabolites to hepatic proteins may lead to the presentation of metabolite-bound peptides on major histocompatibility class I molecules on the surface of hepatocytes. These modiﬁed peptides may be recognized by the T-cell receptor of cytotoxic T lymphocytes. The latter kill target cells by expressing Fas ligand, tumor necrosis factor-α (TNF-α), TNF-α-related apoptosis-inducing ligand (TRAIL), and granzyme B at contact sites. All four substances trigger Bid truncation indirectly or directly. Truncated Bid (tBid) causes a conformational change in Bax, which translocates to the mitochondria to trigger permeabilization, and sometimes also rupture, of the outer mitochondrial membrane. The release of cytochrome c into the cytosol activates caspase-9, which activates caspase-3 that triggers apoptosis.

156 •

•

•

•

•

•

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

Peptides are then transported into the lumen of the endoplasmic reticulum, where some of them can bind to the groove of a MHC class I molecule. After vesicular transport and then fusion of the transport vesicle with the plasma membrane, MHC class I molecules are distributed to the outer surface of hepatocytes, where they display both normal peptides and peptides haptenized by the reactive metabolite. The haptenized peptides differ from the normal self of the individual and therefore can be recognized by the T-cell receptor (TCR) of some cytotoxic T lymphocytes. Cytotoxic T lymphocytes kill target cells through four main mechanisms (Figure 5). They express Fas ligand on their surface. They express tumor necrosis factor-α (TNF-α) on their surface and release it at contact sites. They express TNF-α-related apoptosis-inducing ligand (TRAIL). Finally, they release granzyme B and perforin. The latter forms holes in the plasma membrane and also in the membrane of endocytic vesicles, thus allowing the entry of endocytosed granzyme B into the cytoplasm. The interaction of Fas ligand with Fas, that of TNF-α with the TNF-α receptor 1, or the interaction of TRAIL with its active receptors all activate caspase-8, which then cuts Bid into truncated Bid (tBid). Granzyme B also cuts Bid into tBid. The latter then causes a conformational change in Bax. The modiﬁed Bax migrates to the mitochondria and inserts into the mitochondrial outer membrane. Large Bax aggregates are formed which permeabilize the outer mitochondrial membrane and allow the egress of cytochrome c and other pro-apoptotic proteins from the intermembrane space of mitochondria into the cytosol (Figure 5). The loss of cytochrome c impairs electron ﬂow between complexes III and IV of the mitochondrial respiratory chain, thus causing the accumulation of electrons within complexes I and III of the respiratory chain. The accumulated electrons react increasingly with oxygen to form the superoxide anion radical, which is transformed by manganese superoxide dismutase into hydrogen peroxide. The superoxide radical reacts with nitric oxide to form peroxynitrite, while hydrogen peroxide reacts with ferrous ion to form the hydroxyl radical. These highly reactive nitrogen and/or oxygen species may then trigger the opening of the MPT pore in some mitochondria, thus causing outer membrane rupture, and further, releasing cytochrome c, which activates caspases in the cytosol to trigger apoptosis (Figure 5).

Clinical Features Clinically, immune hepatitis has the following characteristics [1]: •

Although a mild increase in serum transaminase activity is relatively frequent (e.g., 2 to 10%) and is probably due to mild direct toxicity, clinical

PARENT DRUG–MEDIATED PERMEABILITY TRANSITION

• •

•

157

hepatitis due to immune reaction has a low frequency (e.g., 1 : 10,000 or 1 : 100,000). However, clinical hepatitis can occasionally be severe and can lead in some patients to fulminant hepatitis, which may require a liver transplant. This immune form of drug-induced hepatitis is often associated with other hypersensitivity manifestations, such as fever, rash, blood eosinophilia, or the presence of eosinophils and/or granulomas on histological slices. Antiadduct antibodies and/or autoantibodies can be detected in some instances, and the lymphocyte transformation test may be positive in the presence of the drug. However, all of these hypersensitivity hallmarks are inconstant, and their absence in an individual patient does not eliminate an immune mechanism. After termination of drug treatment and recovery of the patient, an inadvertent drug rechallenge can lead to a prompt recurrence of hepatitis, sometimes within the ﬁrst day of the resumption of treatment. Not only can the second bout of hepatitis occur sooner but it can be more severe than the ﬁrst episode of DILI [1].

3.3. Tolerance When they are formed in very low amounts (e.g., after the administration of a very low daily dose of the parent drug), reactive metabolites cause no direct toxicity even in genetically susceptible patients, and do not trigger immune reactions. Three reasons may explain this immune tolerance. 1. At these very low doses, covalent binding to hepatic proteins is minimal. 2. The lack of any direct toxicity even in susceptible patients prevents the few modiﬁed hepatic proteins present within the hepatocytes from leaving these cells and reaching the immune system. 3. The spontaneous apoptosis rate is very low in the hepatocytes of a normal liver. Although a few modiﬁed proteins could still reach Kupffer cells through the spontaneous apoptosis of a few hepatocytes, tolerance rather than immunization might be induced, because the absence of concomitant inﬂammation may fail to provide co-stimulatory signals for an immune response. These reasons may explain the empirical observation that drugs which are used at doses of 10 mg daily or less are seldom hepatotoxic [16], even when they form reactive metabolites. 4. PARENT DRUG–MEDIATED PERMEABILITY TRANSITION In other instances, the parent drug itself can trigger MPT directly and/or can sensitize mitochondria to the MPT-inducing effects of calcium or cytokines. This

158

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

effect is seen with various compounds, including several anionic uncouplers of mitochondrial respiration. 4.1. Anionic Uncouplers Effects on Respiration and ATP Formation Under normal circumstances, the ﬂow of electrons along the respiratory chain is coupled with the extrusion of protons from the mitochondrial matrix into the intermembrane space (Figure 6) [4]. Once a high membrane potential is achieved, this high potential then slows NORMAL Respiratory chain

ATP synthase

e− H+

H+

O2

Mitochondrial matrix

− +

Intermembrane space

UNCOUPLING Respiration

ATP formation

Re-entry of H+

e− H+

H+

H+

O2 − − R-COO

R-COOH

+ R-COO−

R-COOH

∆Ψm

H+

Figure 6 Anionic uncouplers increase mitochondrial respiration but decrease ATP formation. Anionic uncouplers, such as drugs with a carboxylic group (R–COOH) can translocate protons across the inner membrane, thus forming the anionic form (R–COO− ), which is then pushed back into the intermembrane space by the mitochondrial membrane potential ( m ). The reentry of protons into the mitochondrial matrix decreases the m , thus unleashing the ﬂow of electrons in the respiratory chain and increasing mitochondrial respiration. However, ATP synthase is bypassed, so that the increased respiration produces heat instead of ATP.

PARENT DRUG–MEDIATED PERMEABILITY TRANSITION

159

the ﬂow of electrons in the respiratory chain, thus decreasing the rate of respiration. However, carboxylic and other acidic compounds, including natural free fatty acids and several NSAIDs, can cause the reentry of protons from the intermembrane space of mitochondria into the mitochondrial matrix and can increase the respiration rate as follows (Figure 6). •

• •

•

• •

•

In the acidic intermembrane space of mitochondria, these drugs are present as the uncharged species (e.g., R–COOH) [17,18]. In this uncharged form, the drug can freely cross the lipid bilayer of the inner mitochondrial membrane. Once in the more alkaline matrix, the uncharged molecule then dissociates into the anionic form (R–COO− ) and a proton. The anionic form is then electrophoretically translocated from the mitochondrial matrix into the intermembrane space by the mitochondrial membrane potential. This second crossing of the inner membrane probably occurs through diverse anion transporters [19,20]. Once the drug is again located within the acidic intermembrane space, the uncharged molecule (R–COOH) is formed again, ready for another cycle of proton translocation. By causing the reentry of protons into the mitochondrial matrix, anionic uncouplers decrease the mitochondrial membrane potential. This decreased potential allows more electrons to rush through the respiratory chain and to end up in cytochrome c oxidase, thus increasing oxygen consumption (Figure 6). However, because ATP synthase is bypassed, this increased respiration occurs in vain, to produce heat instead of ATP. Severe uncoupling can therefore decrease cell ATP, which can cause cell dysfunction and even cell death.

Mitochondrial Permeability Transition (MPT) A mild ATP depletion due to anionic uncouplers can be secondarily aggravated by the occurrence of MPT. The involvement of permeability transition as a ﬁnal mechanism of cell death has been demonstrated with the prototypical uncoupler, FCCP [21,22], and with the NSAID drugs diclofenac [23] and nimesulide, at least when the latter was incubated in the absence of albumin [24]. Albumin sequesters nimesulide in the medium, minimizing its bioavailability [24]. Similarly, the anionic uncouplers salicylic acid and valproic acid have been shown to facilitate calcium-induced permeability transition in isolated mitochondria [25]. To appreciate how anionic uncouplers can help trigger MPT, let us review how the superoxide anion, which is formed by the respiratory chain, is then detoxiﬁed within the mitochondria (Figure 7) [26]. The ﬁrst step in this process is the dismutation of the superoxide anion into hydrogen peroxide by manganese superoxide dismutase. The formed hydrogen peroxide is then detoxiﬁed

160

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

HEPATOCYTE Intermembrane space 2 H2O

Mitochondrial matrix 2 O2.- + 2 H+

GPx1

MnSOD

NADPH NAD+

GSSG

GR

H2O2 2 GSH O2 O2.O2 O2.-

O2

e−

I

e−

III

TH

H+

NADP+ NADP e−− IV ee− e−

O2

H2O

Figure 7 Formation and inactivation of the superoxide anion radical in mitochondria. Most of the electrons entering the respiratory chain ﬁnish in cytochrome c oxidase (complex IV of the respiratory chain), where four electrons are added quickly and in a tight cage to oxygen, so that reactive oxygen species are not released but only water is safely formed. However, a few electrons can react with oxygen within complexes I and III to form the superoxide anion radical (O2 −· ). In the mitochondrial matrix, manganese superoxide dismutase (MnSOD) then dismutates two molecules of the superoxide anion into one oxygen molecule and one molecule of hydrogen peroxide (H2 O2 ). Glutathione peroxidase 1 (GPx1) then reduces H2 O2 into water while oxidizing two reduced glutathione molecules into one glutathione disulﬁde (GSSG). Glutathione reductase (GR) then regenerates GSH, at the expense of NADPH. Finally, an energy-linked NAD(P)+ transhydrogenase (TH) uses both NADH and the mitochondrial membrane potential to regenerate NADPH from NADP+ .

into water by the mitochondrial glutathione peroxidase 1, which concomitantly converts glutathione (GSH) into glutathione disulﬁde (GSSG). The GSSG formed is then reduced back into GSH by glutathione reductase coupled to NADPH oxidation. Finally, mitochondrial NADP transhydrogenase regenerates NADPH from NADP+ by consuming both NADH (which is oxidized to NAD+ ) and the mitochondrial membrane potential (which partially decreases) [26,27]. Uncouplers can impair the ﬁnal detoxiﬁcation of hydrogen peroxide in the following way (Figure 8) [24]: • •

•

Uncouplers increase mitochondrial respiration and therefore increase mitochondrial consumption of NADH. The resulting NADH depletion, as well as a marked decrease in the mitochondrial membrane potential, may impair the activity of NADP transdehydrogenase, thus retro-inhibiting the successive enzymatic loops required for the detoxiﬁcation of hydrogen peroxide. The resulting increase in the steady levels of reactive oxygen species (ROS) and the decreased GSH/GSSG ratio both tend to trigger MPT [22,26,28].

161

PARENT DRUG–MEDIATED PERMEABILITY TRANSITION Nimesulide without albumin ROS

H+ translocation across the inner mitochondrial membrane

Detoxification of ROS

∆ψm

MPT

Respiration NADH oxidation

HEPATOCYTE

GSH ATP

Ca2+

NADH NADPH

GSSG

NECROSIS

GSSG Efflux

Figure 8 Effects of anionic uncouplers on mitochondrial permeability transition. In the absence of albumin, which otherwise sequesters nimesulide in the medium, nimesulide enters hepatocytes and mitochondria, where this weak acid translocates protons into the mitochondrial matrix, thus decreasing the mitochondrial membrane potential ( m ). The decreased m unleashes the ﬂow of electrons in the respiratory chain and increases mitochondrial respiration and the reoxidation of NADH. The resulting NAD(P)H depletion decreases the reduction of GSSH into GSH, thus increasing the cellular levels of GSSG and its efﬂux from cells, which progressively decreases cell GSH. The decreased GSH contributes to the increase in cellular ROS, by decreasing their inactivation. Together with the decreased m , the increase in the GSSG/GSH ratio and the increase in ROS may start to trigger MPT in some mitochondria. As the percentage of damaged mitochondria increases, severe ATP depletion ﬁnally occurs and inhibits Ca2+ -ATPases, thus increasing cell calcium, which may trigger MPT in still undamaged mitochondria. Severe ATP depletion inhibits apoptosis (an energy-requiring process) and triggers necrosis.

•

The permeability transition pore may then open in a few mitochondria, thus allowing the egress of NADH, NAD+ , NADPH, and NADP+ , which further impairs ROS inactivation [28]. Pore opening also causes matrix swelling and outer membrane rupture, which releases cytochrome c from the damaged mitochondria, thus increasing succinate-supported ROS formation by the affected mitochondria [29]. Finally, MPT pore opening also releases calcium from the affected mitochondria. The increase in the concentration of cytosolic calcium causes its entry in still unaffected mitochondria. Both calcium and ROS formation can then trigger MPT in other mitochondria, resulting in a self-amplifying, propagating wave of mitochondrial disruption [24].

Redox Cycling In addition to the NADH depletion due to uncoupling and increased respiration, yet another mechanism might contribute to NADPH

162

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

depletion with diclofenac. It has been suggested that redox cycling could occur between 5-hydroxydiclofenac and N ,5-dihydroxydiclofenac, and that this cycling may consume NADPH both during the cytochrome p450 (CYP)–mediated oxidation of 5-hydroxydiclofenac into N ,5-dihydroxydiclofenac and during the subsequent reduction of the latter back into 5-hydroxydiclofenac [30]. 4.2. Other Drugs Triggering Permeability Transition Betulinic Acid and Lonidamide Betulinic acid is a pentacyclic triterpene, which is investigated as a potential anticancer drug. Betulinic acid triggers MPT in isolated mitochondria, even without added calcium, and it causes apoptosis in treated cells [31]. Betulinic acid also inhibits topoisomerase I and II, and this inhibition may contribute to its pro-apoptotic effects [32]. The investigational antineoplastic agent lonidamide targets the adenine nucleotide translocator. Lonidamide can trigger MPT in the absence of added calcium, and can cause apoptosis [33,34]. Troglitazone The peroxisome proliferator–associated receptor-γ agonist, troglitazone, was removed from the market because of its hepatotoxic potential [35]. Although troglitazone caused mixed hepatitis in some patients, it mostly triggered hepatocellular, sometimes severe, liver injury [36]. One effect of troglitazone is to inhibit the canalicular bile salt export pump [37]. Although this inhibition may contribute to cholestasis in patients with mixed hepatitis, it cannot account solely for the severe, life-threatening hepatocellular injury observed in a few patients. A possible mechanism for hepatocellular injury is CYP-mediated metabolic activation, which can occur on either the α-tocopherol moiety or the thiazolidinedione moiety of troglitazone [38]. Another possible mechanism involves mitochondrial proapoptotic effects. Troglitazone triggers c-Jun N-terminal protein kinase activation, Bid truncation, MPT, mitochondrial membrane potential collapse, mitochondrial cytochrome c release, ROS formation, and apoptosis in hepatic cell lines [39,40]. These effects occur even in hepatic human cell lines that have negligible cytochrome P450 expression, thus excluding a role for metabolic activation in these mitochondrial, proapoptotic effects [39]. In addition, troglitazone potently inhibits respiration and uncouples oxidative phosphorylation (OXPHOS) in isolated rat liver mitochondria (see Chapter 16). Which of these diverse cellular effects actually cause hepatitis in humans remains unknown. Conceivably, hepatitis could be due to different mechanisms in different subjects. Alternatively, several effects of troglitazone may act together to trigger hepatitis in one patient. For example, increased bile acid levels, mitochondrial effects, and/or the direct toxicity of reactive metabolites may kill a few hepatocytes, thus permitting the immunization of some patients against hepatic proteins modiﬁed by the covalent binding of reactive metabolites. A positive lymphocyte stimulation test and/or the presence of hepatic eosinophils or granulomas have been observed in some cases, suggesting an immunoallergic mechanism, at least in some patients [41].

PRIMARY IMPAIRMENT OF MITOCHONDRIAL β-OXIDATION

163

Hydrochloroquine Mitochondrial membrane disruption can also occur as a consequence of primary lysosomal effects. The lysosomotropic antimalarial drug, hydrochloroquine, releases cathepsin B from lysosomes, which may cause the translocation of Bax from the cytosol to mitochondria, thus triggering outer mitochondrial membrane permeabilization and/or permeability transition, and apoptosis [42]. Peripheral Benzodiazepine Receptor Ligands The peripheral benzodiazepine receptor (PBR) is located on the outer mitochondrial membrane and interacts with the MPT pore. In different experimental conditions, PBR ligands have been shown to either inhibit or augment apoptosis by modulating MPT. At a low concentration, the PBR ligand, alpidem was not toxic alone to hepatocytes, but increased TNF-α-mediated toxicity [43]. In ﬁbroblasts, PBR ligands, although not toxic by themselves, increased the MPT and cell death caused by proapoptotic substances, such as TNF-α [44]. In hepatic stellate cells, 4 -chlorodiazepam and another selective PBR ligands decrease the mitochondrial membrane potential and triggered apoptosis [45]. 5. PRIMARY IMPAIRMENT OF MITOCHONDRIAL β-OXIDATION In many other instances, the parent drug itself directly impairs the mitochondrial uptake of fatty acids and/or inhibit their mitochondrial β-oxidation directly, thus impairing the oxidation of fat in the liver. 5.1. Fat Removal from the Liver Two main pathways remove fat from the liver (Figure 9). A ﬁrst pathway is fat oxidation [4]. In this process, a long-chain fatty acid ﬁrst forms the acyl-CoA derivative, followed by the transient formation of an acyl-carnitine derivative, which enters into the mitochondria, where the long-chain fatty acyl-CoA is regenerated [4]. Inside the mitochondria, the acyl-CoA is then cut by the β-oxidation cycles into acetyl-CoA subunits, which are ﬁnally degraded into CO2 and water by the tricarboxylic acid cycle and the oxidative phosphorylation process [4]. The second major pathway removing fat from the liver is the hepatic secretion of triglycerides. In the lumen of the endoplasmic reticulum, microsomal triglyceride transfer protein (MTP) lipidates apolipoprotein B into very low density lipoproteins (VLDLs), which follow vesicular ﬂow to the plasma membrane to be secreted [46]. 5.2. Drug-Induced Steatosis Hepatic steatosis (“fatty liver”) is deﬁned as the excessive accumulation of fat, mainly triglycerides, in the liver. Drugs can cause hepatic steatosis by impairing

164

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

Fat droplet FFA HEPATOCYTE FFA synthesis uptake FFA FA-CoA

TG

TG TG MTP

ER

MITO

TG

Apo B Vesicular flow

CO2 + H2O Secretory vesicle

TG-rich VLDL

Figure 9 Fat metabolism in hepatocytes. Free fatty acids (FFAs) are synthesized within hepatocytes or are taken up from the plasma coming from the adipose tissue. Long-chain FFAs form fatty acyl-CoA thioesters (FA-CoA), which either enter mitochondria to be oxidized into CO2 and water, or undergo esteriﬁcation into triglycerides (TGs) that are either stored in the cytoplasm or secreted. In the lumen of the endoplasmic reticulum (ER), microsomal TG transfer protein (MTP) lipidates apolipoprotein B (Apo B) into TG-rich very low density lipoproteins (VLDLs), which follow vesicular ﬂow to the plasma membrane to be secreted.

the mitochondrial β-oxidation of fatty acids by inhibiting MTP activity and hepatic VLDL secretion, or, quite frequently, by inhibiting both mitochondrial β-oxidation and MTP activity concomitantly [46]. Mild inhibition of mitochondrial β-oxidation alone is not enough to cause steatosis. Severe impairment is required [4]. In the latter case, the free fatty acids, which are taken up by the liver or are synthesized within the liver, are insufﬁciently oxidized by the deﬁcient mitochondria, and are instead esteriﬁed into triglycerides, which accumulate within the cytoplasm of hepatocytes, thus causing steatosis [4]. Acute impairment of fatty acid β-oxidation typically causes microvesicular steatosis [4]. In this peculiar form of steatosis, numerous tiny lipid vesicles displace the nucleus to the center of the cell and give the hepatocyte a “foamy,” “spongiocytic” appearance (see Chapter 21). However, when β-oxidation is more chronically impaired, mixed forms of steatosis can also occur. Some hepatocytes are then ﬁlled with tiny lipid vesicles, while other hepatocytes exhibit large fat vacuoles or exhibit both small vesicles and large vacuoles. These associations and transitions suggest that tiny lipid vesicles can progressively coalesce into larger vacuoles. Indeed, prolonged causes of steatosis tend, rather, to cause

PRIMARY IMPAIRMENT OF MITOCHONDRIAL β-OXIDATION

165

macrovacuolar steatosis [4]. In this other form of steatosis, hepatocytes are distended by a single large vacuole of fat, displacing the nucleus to the periphery of the cell. The primary impairment of fatty acid oxidation can be a serious medical condition and can cause death in particularly severe cases. It has been suggested that a major impairment of fatty acid oxidation hampers cell energy production during fasting episodes [4]. This could occur through four mechanisms [4]. 1. First, although fatty acid oxidation represents the main cellular source of energy between meals, subjects whose mitochondrial β-oxidation is severely impaired cannot use this major source of energy when they fast [4]. 2. These subjects cannot derive energy from fatty acids, and they also have difﬁculty in getting energy from other energetic fuels because the inhibition of β-oxidation secondarily inhibits hepatic gluconeogenesis [4]. Hypoglycemia frequently occurs when these patients fast, thus hampering energy production from glucose [4]. 3. Moreover, fasting triggers massive adipose tissue lipolysis, thus ﬂooding the liver with free fatty acids. The latter are not oxidized by the deﬁcient mitochondria and therefore accumulate within hepatocytes [4]. Free fatty acids and their dicarboxylic acid derivatives inhibit and uncouple mitochondrial respiration, thus further decreasing energy production [4]. 4. Finally, steatosis leads to lipid peroxidation, whose reactive products damage the respiratory chain and mtDNA [47]. The combination of these four effects may hamper cell energy production sufﬁciently to cause cell dysfunction in some organs. Patients whose mitochondrial β-oxidation is severely impaired do not tolerate fasting. If fasted, they can develop mild liver failure, renal failure, pancreatitis, and severe brain dysfunction, leading to coma and death [4]. These severe complications were ﬁrst reported after the administration of tetracyclines at high intravenous doses. 5.3. Tetracyclines At currently administered oral doses, tetracycline and its derivatives produce only minor degrees of hepatic steatosis of no clinical severity in humans. However, severe, often fatal microvesicular steatosis has occurred in the past following intravenous administration of high doses of tetracycline [9]. Predisposing factors included impaired renal function (which may decrease tetracycline elimination) and pregnancy [which may impair mitochondrial function (discussed below)] [9]. The syndrome usually appeared after 4 to 10 days of tetracycline infusion. Microvesicular steatosis has also been observed after the intravenous administration of several other tetracycline derivatives [1,9]. Tetracycline itself and the various tetracycline derivatives produce extensive microvesicular steatosis of the liver in experimental animals [48,49]. This is due to the dual effect of these antibiotics, inhibiting both the mitochondrial

166

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

β-oxidation of fatty acids [48,49] and the hepatic secretion of VLDLs [49]. The latter effect occurs at doses that do not inhibit protein synthesis [50], and is due to the inhibition of MTP activity [46]. 5.4. Valproic Acid Clinical Effects Another drug, that can cause microvesicular steatosis is valproic acid. This branched-chain fatty acid is used to treat several seizure disorders. An asymptomatic increase in serum aminotransferase activity, which normalizes with either dose reduction or drug discontinuation, is frequent in patients treated with this antiepileptic agent [51]. A much less frequent side effect is a Reye’s-like syndrome, which occurs mainly (but not exclusively) in very young children and between the ﬁrst and fourth month of treatment. Centrizonal and midzonal microvesicular steatosis are associated with centrizonal necrosis, and sometimes, cirrhosis [52]. The combination of microvesicular fat, liver cell death and scarring may be related to the dual effect of valproic acid, which both inhibits mitochondrial β-oxidation, thus causing steatosis, and triggers mitochondrial permeability transition, thus causing cell death [4]. Sequestration of Coenzyme A Like natural medium and short-chain fatty acids, valproic acid enters mitochondria without requiring previous activation to the acyl-coenzyme A (CoA) and acylcarnitine derivatives. Inside mitochondria, valproic acid is then transformed extensively into valproyl-CoA [53]. The extensive formation of this derivative depletes intramitochondrial CoA. Fatty acids must be in the form of an acyl-CoA derivative to be able to undergo the β-oxidation process. Therefore, the sequestration of CoA by valproic acid inhibits the β-oxidation of long-, medium-, and short-chain fatty acids (Figure 10) [53,54]. The lack of CoA may also inhibit pyruvate dehydrogenase, which requires CoA as a necessary cofactor. This may explain why valproate markedly decreases mitochondrial respiration from pyruvate, although it has little effect on the respiration supported by malate and glutamate [55]. The inhibition of both fatty acid β-oxidation and pyruvate-supported respiration by valproate may explain why this drug can aggravate both inborn β-oxidation defects [56,57] and mitochondrial cytopathies [58–60] (see Chapter 11). Mitochondrial Permeability Transition As already mentioned above, valproic acid has an uncoupling effect, which favors MPT (Figure 10) [25]. Metabolic Activation Finally, the cytochrome P450s CYPs 2C9 and 2A6 can form a double bond between the two outer carbons of valproate, thus forming 4-ene-valproate (Figure 10) [61]. This metabolite is then activated into 4-ene-valproyl-CoA inside mitochondria [62,63]. The ﬁrst dehydrogenation step of the β-oxidation cycle then forms 2,4-diene-valproyl-CoA, a chemically reactive metabolite that can inactivate β-oxidation enzymes [63,64]. This CYP-involving pathway could explain why the hepatotoxicity of valproate can be enhanced by the concomitant administration of the CYP-inducing

PRIMARY IMPAIRMENT OF MITOCHONDRIAL β-OXIDATION

4-Enevalproic acid

CYP

CoA

Valproic acid

Valproate + H+ CoA

4-Enevalproyl-CoA

Valproyl-CoA

MITOCHONDRIA Reactive 2,4-dienevalproyl-CoA

167

Uncoupling ± MPT Sequestration of CoA β-Oxidation Pyruvate oxidation Inactivation of β-oxidation enzymes?

Figure 10 Mitochondrial effects of valproic acid. Valproic acid freely enters mitochondria, and thus translocates protons into the mitochondrial matrix. This protonophoric effect can slightly uncouple mitochondrial respiration, and can help trigger mitochondrial permeability transition (MPT). Inside the matrix, valproate is transformed extensively into valproyl-CoA, thus sequestering intramitochondrial CoA. The lack of CoA impairs both mitochondrial fatty acid β-oxidation and pyruvate oxidation. Valproate is also dehydrogenated by cytochrome P450 (CYP) into 4-ene-valproate, which then forms 4-ene-valproyl-CoA and 2,4-diene-valproyl-CoA within mitochondria. The latter is an electrophilic metabolite, which may inactivate β-oxidation enzymes.

antiepileptic drugs, phenytoin and carbamazepine, which increase the formation of 4-ene-valproate [65]. 5.5. Aspirin and Reye’s Syndrome Mechanisms Aspirin is quickly hydrolyzed into salicylic acid, which is activated into salicylyl-CoA on the outer mitochondrial membrane [66]. Extensive salicylyl-CoA formation ties up extramitochondrial CoA, leaving insufﬁcient CoA to activate long-chain fatty acids, which prevents their entry into mitochondria and their β-oxidation [67]. Yet another effect of salicylate is to uncouple mitochondrial respiration slightly [67] and favor opening of the MPT pore [68], as mentioned above. The latter effect could be involved in the spotty liver cell death observed in patients receiving high therapeutic doses of aspirin [69], and could also contribute to the occurrence of Reye’s syndrome. Reye’s Syndrome Even though lethal overdoses of aspirin frequently cause microvesicular steatosis [70], therapeutic doses do not, although they can trigger Reye’s syndrome in children with viral infections (Figure 11). In some children with an initially benign viral infection such as varicella or inﬂuenza, there may

168

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

± ASPIRIN

VIRAL INFECTION 1. NO, TNF-α, IFN-α 2. Fever increasing energy demands 3. Anorexia increasing adipose tissue lipolysis and FFA release

± INBORN β-OXIDATION DEFECT

REYE'S SYNDROME

Figure 11 Multifactorial origin of Reye’s syndrome. Viral infections release nitric oxide (NO), tumor necrosis factor-α (TNF-α) and interferon-α (IFN-α), which all hamper mitochondrial function. Furthermore, viral infections increase energy demands through fever, and they can cause anorexia and fasting, thus ﬂooding the liver with free fatty acids coming from the adipose tissue. However, these effects are rarely sufﬁcient alone to trigger Reye’s syndrome, unless mitochondrial function is additionally impaired by salicylic acid and/or by a previously latent genetic defect in mitochondrial β-oxidation enzymes.

suddenly occur protracted vomiting; obnubilation; elevated serum liver enzymes; hyperammonemia; a hyperechogenic liver on ultrasonography, indicating the presence of steatosis; and ﬁnally, coma and death. This serious postinfectious disease, known as Reye’s syndrome, is thought to be due to an acquired mitochondrial dysfunction. Interferon-α, TNF-α, and nitric oxide are released during viral infections, and all can impair mitochondrial function. As discussed later, interferon-α decreases the synthesis and stability of mitochondrial transcripts. Nitric oxide reversibly inhibits mitochondrial respiration [71] and may open the MPT pore [72]. TNF-α can also inhibit respiration and open the MPT pore [73]. Nevertheless, viral infections rarely cause Reye’s syndrome, suggesting that these endogenous substances usually do not impair mitochondrial function sufﬁciently to trigger the disease. However, if children take aspirin during a viral illness, the added effects of salicylate on mitochondrial function may then sufﬁciently impair mitochondrial function to trigger the syndrome in some children. The potentiating effect of aspirin on the occurrence of Reye’s syndrome is supported by the following evidence: • •

In the past, 93% of children with Reye’s syndrome had received aspirin during an acute viral illness [74]. Children with Reye’s syndrome had received aspirin more frequently than those with similar viral diseases not followed by Reye’s syndrome [75].

PRIMARY IMPAIRMENT OF MITOCHONDRIAL β-OXIDATION •

169

When aspirin use was advised against in feverish children, there was a parallel decline in the use of aspirin and the incidence of Reye’s syndrome in the Unhited States [76].

Now that the use of aspirin has been curtailed, the few residual cases of Reye’s syndrome seen nowadays occur mainly in children with another potentiating factor, such as a previously latent genetic defect in mitochondrial β-oxidation enzymes or other mitochondrial disorders (Figure 11) [77] (see Chapter 11). The deﬁcit becomes patent when the viral infection further damages mitochondria while fever is increasing energy demands and anorexia and insufﬁcient nutrition trigger adipose tissue lipolysis, thus ﬂooding the liver with free fatty acids which are not oxidized by the deﬁcient mitochondria. 5.6. Female Sex Hormones and the Acute Fatty Liver of Pregnancy About 1 in 13,000 pregnant women develop microvesicular steatosis during the last trimester of pregnancy [78]. Untreated, the disease progresses to coma, kidney failure, and hemorrhage, and leads to the death of the mother and child in 75 to 85% of cases. In contrast, in most cases a rapid termination of pregnancy usually allows delivery of a healthy child followed by rapid resolution of the mother’s disease [79]. Both pregnancy itself [80] and the administration of estradiol and progesterone [81] alter mitochondrial ultrastructure and function in mice. However, these effects are mild. The mitochondrial β-oxidation of fatty acids is only slightly impaired, and microvesicular steatosis does not develop in these mice [80,81]. Similarly, most human pregnancies do not cause fatty liver. Therefore, additional factors are probably required to trigger this syndrome in a few pregnant women. Partial deﬁciency of long-chain 3-hydroxyacyl-CoA dehydrogenase (LCHAD), which is part of the trifunctional membrane-bound β-oxidation enzyme, has been reported in some women with acute fatty liver of pregnancy [82]. Mothers with a single defective LCHAD allele who marry a heterozygous carrier and conceive a fetus with two defective alleles develop the disease [82], whereas those who bear an unaffected child usually have uncomplicated pregnancies. Although the fetus itself may not use fatty acids for energy production, the fetal placenta metabolizes fatty acids [83]. During LCHAD-negative conceptions, the fetal placenta may therefore release toxic 3-hydroxy fatty acids in the maternal circulation. Hypothetically, these toxic fatty acids could then trigger fatty liver in the mother [83]. 5.7. NSAIDs Having a 2-Arylpropionate Structure Several NSAIDs are 2-arylpropionate derivatives and can cause either hepatitis or microvesicular steatosis of the liver. The latter condition has been observed in a few patients treated with pirprofen, naproxen, ibuprofen, or ketoprofen [84–87]. 2-Arylpropionates have an asymmetric carbon and exist as either the S(+)- or the R(-)-enantiomers. Only the S(+)-enantiomer inhibits prostaglandin synthesis,

170

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

whereas only the R(-)-enantiomer is converted into the acyl-CoA derivative. Nevertheless, both the S(+)- and the R(-)-enantiomer of ibuprofen inhibit the β-oxidation of medium- and short-chain fatty acids [88]. Pirprofen, tiaprofenic acid, and ﬂurbiprofen also inhibit mitochondrial β-oxidation [89]. 5.8. Glucocorticoids Glucocorticoids impair mitochondrial β-oxidation by inhibiting acyl-CoA dehydrogenases [90]. Glucocorticoids can cause steatosis [9] and, rarely, steatohepatitis [91] in humans. 5.9. Amineptine and Tianeptine These French antidepressant drugs have both a tricyclic moiety and a heptanoic side chain. The tricyclic moiety undergoes metabolic activation by CYPs [92,93], explaining why amineptine and tianeptine rarely cause immunoallergic hepatitis. The heptanoic side chain undergoes the mitochondrial β-oxidation process, which shortens it to the 5- and 3-carbon derivatives [94,95]. In patients treated with these drugs, mitochondria are thus exposed to C7, C5, and C3 analogs of natural fatty acids. These analogs reversibly inhibit the β-oxidation of medium- and short-chain fatty acids [96,97], explaining why amineptine or tianeptine can, in rare cases, cause mild hepatic steatosis due to impaired β-oxidation [1]. 5.10. Calcium Hopantenate, Panadiplon, and Pivampicillin Calcium hopantenate (also called calcium homopantothenate) has been marketed as a cerebral activator. Calcium hopantenate is a homolog of pantothenic acid, which is a constituent of CoA. The administration of calcium hopantenate can decrease CoA and can inhibit mitochondrial β-oxidation [98]. The drug has caused several cases of Reye’s-like syndrome in Japan [98]. Panadiplon was developed as an anxiolytic drug, but its development was terminated due to several instances of transaminase elevations during clinical trials [99]. Panadiplon is converted into cyclopropane carboxylic acid, which sequesters both CoA and carnitine, and inhibits the mitochondrial β-oxidation of fatty acids [99]. The administration of pivampicillin results in the extensive formation of pivaloylcarnitine, thus depleting free carnitine and inhibiting fatty acid oxidation [100]. 6. PRIMARY IMPAIRMENT OF BOTH β-OXIDATION AND RESPIRATION In other instances, the parent drug itself directly impairs mitochondrial β-oxidation, thus causing steatosis; it can also impair respiration directly, thus increasing mitochondrial ROS formation. This combination of events may cause steatohepatitis, which is characterized by the combination of steatosis with necrosis, apoptosis, inﬂammation, and sometimes ﬁbrosis.

PRIMARY IMPAIRMENT OF BOTH β-OXIDATION AND RESPIRATION

171

6.1. Amiodarone, 4,4 -Diethyaminoethoxyhexestrol, and Perhexiline Amiodarone, 4,4 -diethylaminoethoxyhexestrol, and perhexiline are cationic amphiphilic drugs (see Chapter 2). These drugs have a lipophilic moiety and also an amine function, which can become protonated and thus positively charged. The cationic amphiphilic structure of these drugs can interfere with both lysosomal and mitochondrial function, explaining their propensity to trigger both lysosomal phospholipidosis and steatohepatitis [101]. Lysosomes The uncharged, lipophilic form of amiodarone, 4,4 -diethylaminoethoxyhexestrol, and perhexiline crosses the lysosomal membrane [102]. In the acidic lysosomal milieu, the unprotonated drug molecule is protonated and is trapped inside, since the charged species can no longer cross the lysosomal membrane. The protonated drugs therefore reach much higher concentrations inside lysosomes than their concentrations in the surrounding cytosol. Once inside the lysosomes, the cationic amphiphilic drug forms noncovalent but tight complexes with phospholipids, thus hampering the action of intralysosomal phospholipases [102]. Phospholipids are not degraded, and the phospholipid– drug complexes progressively accumulate as myelinlike ﬁgures in progressively enlarging lysosomes [102]. Although phospholipidosis is frequent and perhaps constant in patients receiving these drugs, phospholipidosis appears to have limited clinical consequence, since it often occurs without clinical symptoms and without marked biochemical disturbances [103]. Mitochondria The cationic amphiphilic structure of these drugs also impairs mitochondrial function as follows (Figure 12) [104–108]: •

•

•

•

The unprotonated, lipophilic form of the drug easily crosses the mitochondrial outer membrane and is then protonated in the acidic intermembrane space of the mitochondria [104–108]. The positively charged, protonated form is then electrophoretically “pushed” inside the mitochondrial matrix by the high electrochemical potential existing across the inner mitochondrial membrane. It remains unknown whether the protonated drug crosses the inner membrane through the aqueous channels of some transporter(s), or, more probably, crosses the lipid bilayer directly, thanks to charge delocalization. Whatever the route, this active electrophoretic uptake leads to high intramitochondrial drug concentrations [104–108]. At these high concentrations, the drugs inhibit β-oxidation, thus causing steatosis, and also partially hamper the transfer of electrons along the respiratory chain [104–108]. Upstream respiratory chain components therefore become overly reduced, and they transfer their electrons directly to oxygen to form the superoxide anion radical and other ROSs [108]. The increased mitochondrial ROS formation causes lipid peroxidation [108], which together with cytokines could trigger steatohepatitis lesions.

172

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

A (Amiodarone, perhexiline, DEAEH)

Respiratory chain

A

H+

AH+ +

CYTOSOL

MITOCHONDRIAL INTERMEMBRANE SPACE

−

MATRIX

High concentrations Respiratory chain

e−

e− O2−. e−− e O

Electorn flow

β-Oxidation

2

ROS

Steatosis

Figure 12 Effects of amphiphilic cationic drugs on mitochondrial function. After crossing the outer membrane, the uncharged tertiary or secondary amine (A) of amiodarone, perhexiline, or diethylaminoethoxyhexestrol (DEAEH) is protonated in the acidic intermembrane space. The positively charged molecule (AH+ ) is then electrophoretically “pushed” into the matrix by the mitochondrial membrane potential. High intramitochondrial concentrations inhibit β-oxidation, thus causing steatosis, and also hamper the ﬂow of electrons within the respiratory chain, thus increasing reactive oxygen species (ROS) formation. ROS may oxidize fat deposits, causing lipid peroxidation, which together with possible ROS-induced cytokine production could cause steatohepatitis.

Indeed, the prolonged administration of 4,4 -diethylaminoethoxyhexestrol, perhexiline, or amiodarone can cause steatosis, necrosis, Mallory bodies, a mixed inﬂammatory cell inﬁltrate (containing neutrophils), ﬁbrosis, and even cirrhosis [109–114]. 6.2. Tamoxifen The antiestrogen tamoxifen can cause steatohepatitis [115,116], particularly in overweight women [117]. Tamoxifen is a cationic amphiphilic drug which is transported electrophoretically into the mitochondrial matrix, where it achieves high concentrations that inhibit both mitochondrial β-oxidation and mitochondrial respiration directly [118]. In addition, tamoxifen intercalates between DNA bases, inhibiting mtDNA synthesis, and depleting mtDNA in mice (discussed below) [118]. Although tamoxifen has been shown to impair lysosomal acidiﬁcation [119] and to cause intralysosomal storage of polar lipids after administration of high doses to animals [120], apparently phospholipidosis has not been reported in human livers.

INHIBITION OF ATP SYNTHASE

173

6.3. Buprenorphine The morphine analog buprenorphine is used as a substitution drug in heroin addicts. The sublingual route is used, to avoid extensive ﬁrst-pass metabolism in the liver. At high concentrations, buprenorphine inhibits both mitochondrial β-oxidation and respiration in rat hepatocyte mitochondria [121]. Much lower concentrations are typical for humans, and the drug is usually well tolerated. However, cytolytic hepatitis and steatosis have been observed in a few patients [122]. Predisposing factors could include intravenous buprenorphine misuse (resulting in higher concentrations) and concomitant exposure to viruses, other drugs, or ethanol, all of which could additively impair mitochondrial function [122]. 6.4. Antimalarial Drugs Primaquine [123] and amodiaquine [124] form reactive metabolites, and this metabolic activation may play an important role in amodiaquine-induced agranulocytosis and hepatitis [124,125]. Antimalarial drugs can also interfere with lysosomal and mitochondrial functions. Chloroquine and most of the other antimalarial drugs are cationic compounds which accumulate in the acidic vacuole of the malaria parasite to disrupt its function due to alkalinization of the vacuole [126]. The accumulation of chloroquine at high concentrations into the lysosomes of the host can cause phospholipidosis [127]. These drugs can also interfere with mitochondrial function. Indeed, chloroquine, primaquine, and quinine impair respiration in rat liver mitochondria [128]. 6.5. Benzarone and Benzbromarone Despite their structural analogy with amiodarone, benzarone and benzbromarone are not cationic drugs but rather, phenolic compounds. Benzarone and benzbromarone both uncouple and inhibit respiration at low concentrations [129]. Although these two drugs also impair mitochondrial β-oxidation, this effect requires higher concentrations [129]. In humans, benzarone and benzbromarone can cause hepatocellular liver injury [130,131].

7. INHIBITION OF ATP SYNTHASE Organotin compounds [132] and several natural toxins, such as apoptolidin [133], aurovertin [134], citreoviridin [135], efrapeptins [136], oligomycin, and venturicidin [137], are potent inhibitors of ATP synthase. These toxins block aerobic ATP formation by mitochondria and can damage aerobically poised cells that cannot synthesize enough ATP glycolytically. ATP synthase activity is also inhibited by supraphysiological concentrations of estrogens [138], and by several phenolic phytochemicals present in human diet, such as resveratrol, curcumin, genistein, or quercetin [139].

174

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

8. INHIBITION OF THE ADENINE NUCLEOTIDE TRANSLOCATOR Adenine nucleotide translocator (ANT) exchanges ATP for ADP across the inner membrane, and also modulates MPT [140]. ANT achieves two different conformational states, with the binding site facing the cytosol, the c-state, or facing the matrix, the m-state [141]. Two different inhibitors speciﬁcally stabilize one or the other of these two conformation states. Atractyloside and the related carboxyatractyloside bind to, and stabilize, the cytosolic c-state of the transporter [141] and can trigger MPT [142]. In contrast, bongkrekic acid binds to, and stabilizes, the matrix m-state of the transporter [141], and inhibits MPT [142]. However, both of these ANT inhibitors block the transport of adenine nucleotides and decrease cell ATP. Like bongkrekic acid, the HIV protease inhibitor nelﬁnavir binds to ANT on its matrix side, and inhibits MPT and apoptosis [143]. Long-chain acyl-CoA esters bind and inhibit ANT from both sides of the inner membrane [144], and can trigger MPT [145]. Impairment of the ANT also increases the membrane potential, blocks electron ﬂow, causes the overreduction of respiratory chain complexes, and increases mitochondrial ROS formation, thus causing mtDNA lesions [146]. Thus, as discussed below, inhibition of ANT by zidovudine could contribute to the increased ROS formation observed in animals or patients treated with this nucleoside analog.

9. INTERFERENCE WITH MITOCHONDRIAL DNA AND/OR MITOCHONDRIAL TRANSCRIPTS Drugs can have a variety of effects on mtDNA and mitochondrial transcripts [5]. They can degrade mtDNA, inhibit or terminate mtDNA replication, inhibit mtDNA transcription and impair the stability of mtDNA transcripts, or inhibit the translation of mtDNA transcripts into proteins [5]. These various effects can decrease the synthesis of mtDNA-encoded respiratory chain polypeptides, with different consequences depending on whether the ﬂow of electrons in the respiratory chain is either mildly or severely restricted (Figure 13) [147]. 1. A mild constraint in the ﬂow of electrons can be compensated by the accumulation of electrons upstream of the block. This accumulation can ensure a normal ﬁnal ﬂow of electrons, and thus a normal respiratory rate, at the expense, however, of increased formation of ROS by the overly reduced respiratory chain (Figure 13) [147]. 2. In contrast, a very narrow bottleneck in the ﬂow of electrons can have three major consequences (Figure 13): a. Despite the accumulation of electrons, a major impediment to the ﬂow of electrons can decrease the ﬁnal ﬂow rate of electrons, thus decreasing mitochondrial respiration and ATP formation, which may cause cell dysfunction and cell death.

175

INTERFERENCE WITH MITOCHONDRIAL DNA NORMAL Slight, basal ROS formation O2 O2-. e− − e−− e− e e e− e− e−

WIDE BOTTLENECK

e− buildup

e− e− e− e−

e− e−

O2 Electron flow and repiration H2O

O2

e− e− e−− e− e e− e− e− e− e−e−

O2-. O2

NARROW BOTTLENECK H 2O O2 e− − − e− e e − e O2-. e− e− − buildup − e e− e − e− − e O2 e− − e e− e e− H2O

ROS formation Normal e− flow Normal respiration

ROS formation e− Flow Respiration β-Oxidation Pyruvate oxidation

Figure 13 Different consequences of either a mild or a severe restriction to the ﬂow of electrons within the respiratory chain. A moderate restriction to the ﬂow of electrons (e− ) within the respiratory chain (“wide bottleneck”) can be compensated by the accumulation of electrons within the chain. Up to a certain point, this accumulation can help maintain a normal ﬂow rate of electrons at the level of cytochrome c oxidase, and thus a normal respiratory rate. However, the buildup of electrons within complexes I and III can increase the formation of the superoxide anion radical (O2 − ) and other reactive oxygen species. In contrast, a severe restriction on the ﬂow of electrons within the respiratory chain (“narrow bottleneck”) can decrease the ﬁnal ﬂow rate of electrons through cytochrome c oxidase, and thus the respiratory rate. The decreased reoxidation of NADH into NAD+ can then secondarily hamper both the β-oxidation of fatty acids and the mitochondrial oxidation of pyruvate.

b. Impairment of respiration can also secondarily impair the mitochondrial β-oxidation of fatty acids [148]. Under normal conditions, NADH formed by β-oxidation is reoxidized by the mitochondrial respiratory chain, thus regenerating NAD+ required to continue β-oxidation. When respiration is severely impaired, not enough NAD+ is regenerated to sustain β-oxidation [148], which can yield microvesicular steatosis [4]. c. Yet another possible consequence of a severe decrease in mitochondrial respiration may be lactic acidosis. The lack of NAD+ inhibits

176

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

the oxidation of pyruvate into acetyl-CoA by pyruvate dehydrogenase, which requires both NAD+ and CoA as necessary cofactors. Therefore, pyruvate is not oxidized. Instead, due to the high NADH/NAD+ ratio, pyruvate is reduced excessively into lactate, whose accumulation can trigger lactic acidosis. 9.1. Degradation of mtDNA by Alcohol Ethanol is probably the most frequently used psychoactive drug, and therefore merits a place in this chapter. Effects on ROS Alcohol abuse increases ROS formation and decreases ROS inactivation through several mechanisms [149]. •

• •

•

The metabolism of ethanol into acetaldehyde and acetate, by alcohol dehydrogenase and aldehyde dehydrogenase, respectively, reduces NAD+ to NADH, thus increasing the NADH/NAD+ ratio, which in turn increases the NADPH/NADP+ ratio [150]. The high NADPH/NADP+ ratio can then reduce ferric iron into ferrous state, a potent catalyst for hydroxyl radical generation [150]. Ethanol administration stabilizes and, to some extent, can also induce the ROS-generating cytochrome P450 2E1 [151]. Ethanol administration increases the formation of ROS by mitochondria [152], and it decreases the detoxiﬁcation of H2 O2 in mitochondria by decreasing the mitochondrial levels of glutathione [153] and the activity of glutathione peroxidase [154]. Finally, the ingestion of ethanol increases the permeability of the gut to endotoxin [155]. High blood endotoxin levels stimulate Kupffer cells, leading to the plasma membrane assembly of NADPH oxidases, which transfer one electron from NADPH to oxygen to form the superoxide anion radical [155].

In intoxicated animals, high hepatic ROS levels can cause oxidative damage to mitochondrial lipids [156], proteins [157], and mtDNA [157,158]. Effects on mtDNA The intragastric administration of a single high dose of ethanol to mice causes extensive degradation of hepatic mtDNA, which is maximally depleted 2 hours after ethanol administration [158,159]. Interestingly, mtDNA depletion also occurs in skeletal muscle, heart, and brain in inebriated mice [159]. The depletion of mtDNA can be prevented by 4-methylpyrazole, which blocks ethanol metabolism, or by melatonin, vitamin E, or ubiquinone, three antioxidant drugs [158,159]. The outcome of the hepatic mtDNA depletion differs after acute or repeated treatments.

INTERFERENCE WITH MITOCHONDRIAL DNA •

•

177

After a single alcohol binge, the damaged mtDNA molecules are quickly repaired and resynthesized de novo, so that mtDNA levels are quickly restored, with even an overshoot phenomenon in hepatic mtDNA levels at 24 hours [158,159]. In contrast, after 4 days of daily binges, the accumulation of nonrepaired bulky lesions (possibly due to lipid peroxidation products) on mtDNA limits the number of intact mtDNA templates and inhibits the resynthesis of mtDNA. The depletion of mtDNA therefore lasts for several days after interruption of the alcohol intoxication [160]. The repetition of mtDNA strand breaks during chronic alcoholism can also cause mtDNA deletions. Indeed, the prevalence of hepatic mtDNA deletions is increased in alcoholics with microvesicular steatosis, but not in patients with alcoholic hepatitis or cirrhosis [161,162]. The latter conditions increase liver cell turnover [163,164], which could eliminate mutated mtDNA genomes if cells with a high proportion of mutated genomes either fail to replicate and/or are selectively eliminated through apoptosis. When ethanol ingestion is stopped, alcohol-induced mtDNA deletions disappear quickly in white blood cells [165], which have a quick cell turnover.

Mechanisms of Steatosis Alcohol-induced steatosis seems to be due to a combination of several mechanisms [166]. •

•

• •

A ﬁrst mechanism is an increased hepatic expression of sterol regulatory element-binding protein-1, which increases hepatic fatty acid synthesis [167]. The excessive reduction of NAD+ into NADH during the metabolism of ethanol can decrease NAD+ levels. The lack of NAD+ slightly impairs mitochondrial β-oxidation and markedly inhibits the tricarboxylic acid cycle [168]. ROS-dependent damage to mitochondrial lipids, proteins, and DNA may further impair mitochondrial function [161]. Finally, a decreased hepatic expression of MTP may partially limit the adaptive increase in hepatic lipoprotein secretion [169].

9.2. Degradation of mtDNA by Acetaminophen (Paracetamol) The inadvertent or deliberate intake of large doses of acetaminophen leads to extensive formation of N -acetyl-p-benzoquinonimine. This electrophilic metabolite depletes hepatic glutathione and protein thiols; increases cell calcium; damages mitochondria; increases the formation of ROS, including peroxynitrite; activates c-Jun N-terminal kinase; and causes MPT and liver cell necrosis [170–173]. Hepatic mtDNA was rapidly depleted after an acetaminophen overdose in mice, whereas nuclear DNA, albeit partially fragmented, was not

178

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

signiﬁcantly depleted [174]. The rapid depletion of mtDNA may be due to the DNA damage caused by peroxynitrite and other ROSs [174]. The absence of a rapid rebound in mtDNA levels may suggest impaired mtDNA resynthesis. Possible reasons could be the development of cell necrosis and the poisoning of topoisomerase II by N -acetyl-p-benzoquinoneimine [175]. 9.3. Impairment of mtDNA Replication by Drugs Inhibiting Topoisomerases and/or Binding to DNA Drugs that intercalate between DNA bases or bind strongly to a DNA groove can directly inhibit the replication of mtDNA [176]. Moreover, intercalating drugs also frequently inhibit DNA topoisomerases, which can further impair mtDNA replication [176]. Topoisomerases play an important role in DNA replication and transcription [177]. These enzymes cut the phosphodiester DNA backbone by forming a covalent bond between the liberated phosphorus of DNA and a tyrosine of the enzyme. Type I topoisomerases act as monomers and cut only one strand of DNA, whereas type II topoisomerases act as dimers or multimers and cut both strands of DNA, thus allowing DNA strand(s) to cross the gap. Normally, the gap in DNA strand(s) formed by the topoisomerases is only transient. After cutting DNA, topoisomerases then promptly reseal the DNA gap. Several antibacterial drugs (e.g., 4-quinolones, novobiocine) and anticancer drugs (e.g., amsacrine, etoposide, anthracyclines, ellipticines, actinomycins) are topoisomerase inhibitors [177]. Although a few inhibitors impair the initial cutting of DNA by topoisomerases, most topoisomerase-interfering drugs inhibit mostly the resealing of DNA. These drugs, which are said to “poison” topoisomerases, increase the number of enzyme-bound DNA complexes. These complexes are called cleavable complexes or cleavage complexes, because detergents can destabilize the complexes so that the DNA ends are no longer held together, and the protein-linked DNA breaks are thus revealed. Both the inhibition and the poisoning of topoisomerases are deleterious to cells. The collision of a transcription complex or a replication fork against a topoisomerase-associated DNA break interrupts RNA or DNA synthesis and can lead to real (nontopoisomerase-bound) double-strand breaks and to gene translocations that can trigger apoptosis and/or cancer [178]. Because mitochondria contain both a type I topoisomerase [179] and a bacterial-like type II topoisomerase [180], topoisomerase inhibitors or poisons can affect the replication of mtDNA. Indeed, mtDNA rather than nuclear DNA can be selectively depleted by drugs that are electrophoretically concentrated in the mitochondrial matrix, such as tacrine. Tacrine The reversible cholinesterase inhibitor tacrine has been used for the symptomatic treatment of Alzheimer’s disease. Monitoring of serum ALT activity and tolerance-dependent stepwise escalation of the doses were recommended, because the drug increased ALT activity, usually after about 6 weeks of treatment,

INTERFERENCE WITH MITOCHONDRIAL DNA

179

in 50% of recipients [181]. The weak base tacrine is taken up by mitochondria, where it can cycle back and forth across the mitochondrial inner membrane, uncoupling respiration and dissipating potential energy as heat without ATP production [182]. First-pass metabolism in the liver lowers exposure in other organs and explains why the liver is injured preferentially [182]. The electrophoretic accumulation of tacrine within the mitochondrial matrix also explains why this organelle is a selective target [182,183]. Tacrine intercalates between mtDNA bases, poisons topoisomerases, and decreases the synthesis of mtDNA in mice [183]. This leads to a progressive depletion of hepatic mtDNA in mice, eventually followed by the death of a few hepatocytes by necrosis or apoptosis [183]. Tamoxifen The antiestrogenic drug tamoxifen is used in the treatment of advanced breast cancer but can trigger steatohepatitis in overweight women [117]. This cationic amphiphilic drug accumulates electrophoretically within mitochondria, where it directly inhibits mitochondrial respiration and mitochondrial β-oxidation, thus causing steatosis [118]. In addition, tamoxifen intercalates between DNA bases and inhibits topoisomerases and mtDNA synthesis [118]. Tamoxifen progressively depletes hepatic mtDNA in mice [118]. Quinolone Antibiotics The 4-quinolone antibiotics inhibit gyrase (a bacterial type II topoisomerase) and also inhibit the mitochondrial type II topoisomerase [184]. Ciproﬂoxacin blocks the resealing of mtDNA breaks, causing the accumulation of protein-linked double-strand mtDNA breaks [185]. Ciproﬂoxacin and nalidixic acid progressively decrease mtDNA in cultured mammalian cells and impair mitochondrial respiration and cell growth [184]. 4-Quinolone antibiotics can cause cholestasis, steatosis, and necrosis in treated patients [186,187], and both trovaﬂoxacin and alatroﬂoxacin were taken off the market because of an unacceptable risk of fulminant liver failure. However, it remains unknown whether mtDNA depletion actually occurs in humans or experimental animals treated with quinolone antibiotics. Alternative mechanisms for quinolone-induced hepatitis could include lysosomal membrane permeabilization and MPT [188], altered expression of mitochondrial proteins [189], and the occurrence of immune reactions in some patients [190]. Pentamidine Pentamidine is used in the prevention and treatment of Pneumocystis carinii infections. Although pentamidine inhibits the topoisomerases of P. carinii and trypanosomes, it has little effect on mammalian topoisomerases [176]. However, pentamidine binds to the minor groove of duplex DNA and can deplete the mtDNA of mammalian cells in vitro [176]. Pentamidine may also inhibit mitochondrial translation [191]. Other Drugs Methylglyoxal bis(guanine hydrazone) and several polyamine analogs, including N1 ,N12 -bis(ethyl)spermine and N1 ,N8 -bis(ethyl)spermidine, have been shown to decrease mtDNA levels progressively in cultured cell lines

180

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

[192,193]. The polycationic nature of the polyamines causes their electrophoretic accumulation within the mitochondrial matrix [176]. Moreover, this polycationic nature also leads to their strong interaction with the successive phosphate groups of the DNA backbone on the major groove of DNA, thus causing conformational changes in DNA [176]. Ethidium bromide, ditercalinium, and dequalinium are also cationic drugs that bioaccumulate electrophoretically into the matrix, where they intercalate between mtDNA bases, inhibit mtDNA synthesis, and cause progressive mtDNA depletion [194–196]. 1-Methyl-4-phenylpyridinium ion (MPP+ ) is an oxidative metabolite of a mitochondrial toxin that causes a Parkinsonian syndrome in humans and other species. This positively charged metabolite accumulates in mitochondria, where it inhibits mtDNA synthesis and depletes mtDNA [197], in addition to its more acute effect of inhibiting complex I. 9.4. Impairment of mtDNA Replication by 2 ,3 -Dideoxynucleosides and Abacavir Several 2 ,3 -dideoxynucleosides are used in patients infected by the human immunodeﬁciency virus (HIV). These analogs include 3 -azido-2 ,3 -dideoxythymidine (zidovudine, AZT), 2 ,3 -dideoxycytidine (zalcitabine, ddC), 2 ,3 dideoxyinosine (didanosine, ddI), 2 ,3 -didehydro-3 -deoxythymidine (stavudine, d4T), and (-)-2 -deoxy-3 -thiacytidine (lamivudine, 3TC). A related molecule is abacavir, which contains a cyclopentene–methanol moiety instead of the dideoxyribose moiety of the above-mentioned drugs. These analogs can impair mtDNA replication through several mechanisms, including their incorporation into mtDNA. Incorporation into mtDNA and Termination of mtDNA Replication The normal 5 -hydroxyl group found in deoxyribose is present in the sugar analog of these diverse analogs, thus allowing the formation of the triphosphate derivative and then the incorporation of the nucleotide analog into a growing chain of replicating DNA. In contrast, the normal 3 -hydroxyl group of deoxyribose is absent in these analogs. Once a single molecule of the analog has been incorporated, the DNA molecule now lacks a 3 -hydroxyl group. Unless the analog can be removed, no other nucleotide can be incorporated, therefore interrupting DNA replication (Figure 14) [198,199]. The effects of the nucleoside analog therefore depend on the ability of various DNA polymerases to incorporate the analog into DNA. • • •

HIV reverse transcriptase can perform this incorporation, thus impairing the reverse transcription of the HIV RNA [198]. In contrast, the DNA polymerases, which act in the nucleus, barely perform this incorporation, thus allowing the therapeutic use of these analogs [199]. However, DNA polymerase γ, which acts in the mitochondria, also incorporates nucleotide analogs into growing chains of mtDNA. This incorporation

181

INTERFERENCE WITH MITOCHONDRIAL DNA

MITOCHONDRIA T

CONSEQUENCES TTP

TMP

Inhibition of mtDNA replication

TK AZT

AZT-TP

AZT-MP

ATP ADP

AZT

Incorporation into, and termination of, mtDNA replication

Respiration

ANT Accumulation of electrons ROS formation mtDNA mutations

Figure 14 Possible mechanisms for the mitochondrial effects of azidothymidine (AZT). Because AZT and thymidine compete for phosphorylation by thymidine kinase (TK), high AZT concentrations can decrease the formation of thymidine monophosphate (TMP) and thymidine triphosphate (TTP). The lack of TTP may then slow down the synthesis of mitochondrial DNA (mtDNA). Furthermore, a small fraction of the AZT monophosphate (AZT-MP) formed by TK may be further phosphorylated to the AZT triphosphate (AZT-TP), thus possibly leading to the incorporation of an AZT-MP pseudonucleotide into mtDNA. Unless this wrong nucleotide can be removed by the proofreading activity of DNA polymerase γ, it terminates mtDNA replication due to the absence of a 3 -hydroxyl group on the AZT-ended DNA chain. The inhibition or termination of mtDNA synthesis may then decrease the synthesis of mtDNA-encoded respiratory chain polypeptides, thus impairing mitochondrial respiration. Furthermore, AZT also inhibits the adenine nucleotide translocator (ANT). Under normal circumstances, the ANT exchanges ATP for ADP, thus stimulating the reentry of proton through ATP synthase and the associated ﬂow of electrons in the respiratory chain. In contrast, inhibition of the ANT can decrease this ﬂow of electrons and mitochondrial respiration. The partial block in the ﬂow of electrons can cause overreduction of respiratory chain complex I and complex III, thus increasing the mitochondrial formation of reactive oxygen species, which can trigger mtDNA mutations.

terminates mtDNA replication (Figure 14) [200,201] unless the nucleotide analog can be removed by the proofreading, 3 -5 -exonuclease activity of polymerase γ [202]. Proofreading Polymerase γ is slow at removing dideoxynucleotides when they are correctly base-paired, even though the sugar–analog–phosphate backbone is abnormal. Thus, the rate for the incorporation of nucleotide analogs into mtDNA is much faster than the rate for their removal [202]. Furthermore, once one analog

182

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

molecule has been removed and elongation has been able to resume transiently, another molecule of the analog is likely to be incorporated a little further on. This new blocking molecule will have to be removed again, and so on. The end result is to considerably slow down the rate of efﬁcient, complete mtDNA replication [202]. Impaired Thymidine–Triphosphate Formation Another mechanism contributing to mtDNA depletion, at least with zidovudine, is that zidovudine (AZT) and thymidine compete for phosphorylation by thymidine kinase into AZT-monophosphate (AZT-MP) and thymidine monophosphate (TMP), respectively (Figure 14) [203]. Due to this competition, AZT decreases the formation of TMP and thus eventually decreases the formation of thymidine triphosphate (TTP). The lack of enough TTP may then impair mtDNA replication. Interestingly, it has been shown that uridine administration can prevent the toxicity of zidovudine, zalcitabine, and stavudine in experimental animals and possibly also in humans [204,205]. One hypothesis is that uridine could eventually lead to the synthesis of TTP [203]. By preventing TTP depletion, uridine could prevent mtDNA depletion [203] (see Chapter 9). mtDNA Depletion When mtDNA replication is markedly slowed, mtDNA levels may decrease progressively (Figure 14). For reasons that are not yet fully understood, different dideoxynucleosides tend to have differential effects on mtDNA in diverse organs. Although zidovudine can occasionally cause mtDNA depletion in the liver [206], the “d-drugs” [i.e., ddC (zalcitabine), ddI (didanosine), and d4T (stavudine)] may be more likely to cause hepatic mtDNA depletion than zidovudine, lamivudine, or abacavir [207]. In a patient with lactic acidosis after treatment with both didanosine and stavudine, all mitochondrial complexes were markedly decreased except for complex II, which is the only respiratory complex encoded completely by nuclear DNA [208]. ROS Formation The impaired synthesis of mtDNA-encoded respiratory chain polypeptides can partially hamper the ﬂow of electrons in the respiratory chain [209]. Whenever the ﬂow of electrons in the respiratory chain is partially hampered, electrons may accumulate within complexes III and I, and may react increasingly with oxygen to form the superoxide anion. In the case of zidovudine, this effect may be aggravated further by the inhibitory effect of this analog on the ANT (Figure 14) [210]. The lack of ANT in knockout mice blocks the exchange of mitochondrial ATP for cytosolic ADP [146]. The impaired entry of ADP into the mitochondrial matrix prevents the reentry of protons through ATP synthase and causes a high mitochondrial potential [146]. This high potential blocks the ﬂow of electrons in the respiratory chain and causes the overreduction of respiratory chain complexes, thus increasing mitochondrial ROS formation and triggering mtDNA deletions [146]. Zidovudine administration increases peroxide formation by hepatic mitochondria and causes the oxidation of guanosine into 8-hydroxydeoxyguanosine in the mtDNA of mouse liver [211].

INTERFERENCE WITH MITOCHONDRIAL DNA

183

Like zidovudine-treated mice, asymptomatic HIV-infected patients treated with zidovudine have higher urinary excretion of 8-hydroxydeoxyguanonise than do untreated patients [212]. mtDNA Mutations Oxidative damage to mtDNA can occasionally cause both point mutations and deletions (Figure 14). In patients treated with diverse nucleoside reverse transcriptase inhibitors, heteroplasmic point mutations were shown to accumulate in peripheral blood cell mtDNA [213]. Similarly, mtDNA deletions were more prevalent in the sperm of patients treated with various nucleoside analogs for 12 months or more than in those with shorter treatments [214]. In one of these patients, multiple deletions were present after 6 months of treatment, whereas none had been found before treatment [214]. In another patient treated with lamivudine, stavudine, and saquinavir, most of the hepatic mtDNA existed as six distinct deleted forms [215]. This patient developed lactic acidosis, although hepatic mtDNA was not depleted [215]. Thus, once mtDNA deletions and point mutations have developed, severe mitochondrial dysfunction may be present even though the tissue mtDNA level (which is the sum of normal and mutated mtDNA) may be normal (see Chapter 23). Mitochondrial Biogenesis Tissues whose mtDNA is decreased or abnormal can attempt to increase both mtDNA replication and transcription by inducing compensatory increases in mitochondrial biogenesis. In human volunteers treated for 2 weeks with either stavudine/lamivudine or zidovudine/lamivudine, a decrease in mtDNA-encoded messenger RNAs was associated with an increased expression of peroxisome proliferator receptor gamma coactivator 1 (PGC1), nuclear respiratory factor 1, and mitochondrial transcription factor A in adipose tissue [216]. All three factors are master regulators of mitochondrial biogenesis [217] and may help attenuate the adverse effects of nucleoside reverse transcriptase inhibitors. Indeed, an increase in the number of muscle mitochondria is typically observed in patients with zidovudine-induced myopathy [201]. Although mitochondrial proliferation is much less conspicuous in the liver, it can sometimes occur [218]. Indeed, in one patient with stavudine-induced lactic acidosis, extensive hepatic mitochondrial proliferation led to tightly packed mitochondria on electron microscopy, and to a pink granular cytoplasm on light microscopy, thus giving to the hepatocytes an “oncocytic” appearance [218]. 9.5. Impairment of mtDNA Replication by Fialuridine and Ganciclovir Fialuridine Fialuridine was tried for treatment of patients with chronic hepatitis B. However, these clinical trials had to be interrupted after several patients developed microvesicular steatosis and unmanageable lactic acidosis, sometimes associated with pancreatitis, neuropathy, or myopathy [219]. These complications were unexpected because ﬁaluridine possesses both a 5 -hydroxyl group and a 3 -hydroxyl group, so the incorporation of a single molecule of ﬁaluridine into DNA should not immediately terminate mtDNA replication. However,

184

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

when several adjacent molecules of ﬁaluridine are incorporated successively, DNA polymerase γ activity is then inhibited, substantially decreasing mtDNA replication and mtDNA levels [220]. Ganciclovir Ganciclovir is used primarily for treatment of cytomegalovirus infections [221]. The viral kinases, which are encoded by cytomegalovirus, the varicella/zoster virus, or the herpes simplex viruses, convert ganciclovir into ganciclovir monophosphate. The latter is then activated by cellular kinases into ganciclovir triphosphate, which can lead to the incorporation of a ganciclovir nucleotide into a growing DNA molecule [221]. The acyclic pseudosugar analog which is present in ganciclovir has two hydroxyl groups, so incorporation of a ganciclovir nucleotide into DNA does not terminate DNA replication. However, the ganciclovir molecules incorporated may distort the DNA helix, which may block the next DNA replication cycle when the ganciclovir-modiﬁed strand serves as the replication template [221]. Ganciclovir is also incorporated in mtDNA where it triggers mtDNA depletion and ultrastructural mitochondrial lesions, resulting in steatosis and apoptosis [222]. 9.6. Decreased Synthesis and Stability of Mitochondrial Transcripts in Cells Treated with Interferon-α Interferon-α is used in patients with chronic viral hepatitis B or C and some forms of cancer. Interferons induce 2 ,5 -oligoadenylate synthases, which synthesize 2 ,5 -oligoadenylates in the presence of double-stranded RNAs [223]. The 2 ,5 -oligoadenylates formed then activate RNase L [223]. Not only is this RNA-degrading enzyme activated by interferon, it is also induced by interferon-α and β [223]. The activation and induction of RNAse L by interferon may affect mitochondrial transcripts both indirectly and directly. •

•

RNAse L can act indirectly by cleaving ﬁrst the nuclear DNA-encoded mRNA of mitochondrial transcription factor A (mtTFA) [224]. In the mitochondrial matrix, mtTFA binds to enhancer sequences located upstream of the origins of transcription of both the light and heavy strands of mtDNA to increase mtDNA transcription. By decreasing mtTFA, RNAse L may therefore decrease the synthesis of mitochondrial mRNAs [225]. RNase L is also present in mitochondria, where it can degrade mitochondrial mRNAs [226]. Thus, the treatment of cells with interferon-α can decrease both the synthesis and the stability of mitochondrial transcripts [225,226].

In cultured cells, these dual effects of interferon-α can eventually decrease mtDNA-encoded respiratory chain polypeptides and mitochondrial respiration [227]. Whether similar effects also occur in treated patients is unknown. However, it is noteworthy that some of the adverse effects of interferon-α, such as hepatic steatosis [228] or minor blood dyscrasias, myalgias, paresthesias, convulsions, and depression [229], resemble the mild clinical manifestations that can occur in the mild forms of the inborn mitochondrial cytopathies.

MECHANISMS BEHIND IDIOSYNCRASY

185

9.7. Decreased Translation of Mitochondrial Transcripts into Proteins Erythromycins Erythromycins are amphiphilic cationic drugs that accumulate in acidic compartments, including lysosomes, where they can inhibit phospholipases to cause phospholipidosis [230]. These antibiotics also bind to the 50S ribosomal subunit of bacteria and inhibit the transfer of amino acids from the aminoacyl-tRNA to the peptide chain [231]. Erythromycins can also inhibit mitochondrial protein synthesis, at least to some extent [232]. They can cause megamitochondria [233] and can trigger sensorineural hearing loss [234]. Finally, erythromycins are transformed into reactive metabolites, which may covalently bind to proteins, thus forming neoantigens, which may trigger immunization and hepatitis [235,236]. It is not known whether mitochondrial effects can contribute to erythromycin-induced cholestatic or mixed hepatitis (possibly by causing bile duct lesions and the release of immunizing neoantigens). Chloramphenicol and Thiamphenicol Chloramphenicol and thiamphenicol also bind to the 50S ribosomal subunit to inhibit protein synthesis in both bacteria and mitochondria [237]. Mitochondrial dysfunction is probably involved in the reversible bone-marrow suppression induced by chloramphenicol [238]. However, it remains unknown whether other effects of chloramphenicol, including aplastic anemia and cholestatic hepatitis, are due to mitochondrial dysfunction and/or to reactive metabolite formation [238]. Linezolid The oxazolidinone antibiotic linezolid is used to treat drug-resistant gram-positive pathogens [239]. It acts by inhibiting bacterial protein synthesis [239]. Linezolid also inhibits the synthesis of mitochondrial proteins [239]. Although it does not change mtDNA levels, linezolid decreases the activity of respiratory chain complexes containing mtDNA-encoded proteins. In humans the drug can trigger lactic acidosis and neuropathy [239] (see Chapter 20).

10. MECHANISMS BEHIND IDIOSYNCRASY If an investigational drug molecule is shown to cause frequent adverse effects in humans, it is rarely released to the market. A corollary of this exclusion rule is that drugs given at therapeutic doses cause DILI in only a few recipients. With the exception of large overdoses, all cases of DILI can therefore be considered as idiosyncratic. The reasons for the unique susceptibility of these few subjects are incompletely understood. However, a few examples show that metabolic factors and co-morbidity diseases can modulate the hepatotoxicity of drugs that impair mitochondrial function. 10.1. Metabolic Factors When the parent drug, rather than a metabolite, impairs mitochondrial function, any factor decreasing drug elimination may enhance toxicity.

186 •

•

•

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

For example, renal insufﬁciency, which decreases tetracycline elimination, was a risk factor for severe microvesicular steatosis after high intravenous doses of tetracycline [9]. Although chloramphenicol inhibits mitochondrial protein synthesis, it is detoxiﬁed by glucuronide formation. The mitochondrial toxicity of high doses of chloramphenicol was increased in premature or newborn babies, whose capacity for drug glucuronide formation was still immature [240]. Perhexiline maleate inhibits mitochondrial fat oxidation and energy production but is detoxiﬁed through the formation of water-soluble metabolites by CYP2D6 [241]. Patients genetically deﬁcient in CYP2D6 were at increased risk of developing perhexiline-induced liver injury [241].

10.2. Co-morbidity Factors Several different causes additively impair mitochondrial function and damage the liver in a single patient. •

•

•

•

The onset of Reye’s syndrome can be triggered by the combination of a viral infection and aspirin use, or the combination of a previously latent genetic defect in β-oxidation enzymes and a viral infection [76,77]. The prevalence of microvesicular steatosis after large intravenous doses of tetracyclines seemed to be increased by pregnancy [9], which impairs fatty acid oxidation in mice [80]. Valproate, which inhibits both fatty acid oxidation and pyruvate oxidation, can unravel both inborn β-oxidation defects and inborn mitochondrial cytopathies [56–60]. Obesity can cause insulin resistance, hepatic steatosis, and mitochondrial lesions [217]. Obesity increases the risk of tamoxifen-induced steatosis and steatohepatitis in women [117].

11. CONCLUSIONS A frequent mechanism for DILI is the formation of reactive metabolites, which can trigger hepatitis via direct toxicity or immune reactions. In both instances, however, mitochondrial membrane permeabilization often occurs as a ﬁnal mechanism of cell death. In other instances the parent drug itself can trigger MPT (as with acidic NSAIDs, valproic acid, salicylate) or it can impair mitochondrial function through a variety of mechanisms. Drugs can sequester coenzyme A (e.g., aspirin, valproic acid), inhibit mitochondrial β-oxidation enzymes (e.g., tetracyclines, 2-arylpropionate anti-inﬂammatory drugs, amineptine, tianeptine, glucocorticoids, amiodarone, perhexiline, tamoxifen), inhibit both β-oxidation enzymes and the transfer of electrons within the respiratory chain (e.g., perhexiline, amiodarone), impair mitochondrial structure and function (e.g., female sex hormones), or inhibit the ANT or ATP synthase. Drugs can also destroy mtDNA

REFERENCES

187

(e.g., alcohol, paracetamol) or inhibit mtDNA replication (e.g., dideoxynucleosides, abacavir, ﬁaluridine, ganciclovir, tacrine, tamoxifen). Drugs can also decrease the synthesis and stability of mtDNA transcripts (e.g., interferon-α) or can impair mitochondrial protein synthesis (e.g., erythromycins, chloramphenicol). Quite often, a single drug (e.g., valproic acid) has many different effects on mitochondrial function. A severe impairment of β-oxidation can cause a fatty liver. When β-oxidation is severely impaired, fatty acids are poorly oxidized by mitochondria and are instead esteriﬁed into triglycerides, which initially accumulate as small lipid vesicles that can progressively coalesce with time into larger vacuoles. Importantly, the primary impairment of one mitochondrial function can secondarily impair another. Thus, severe primary impairment of β-oxidation may also decrease energy formation during periods of fasting when the β-oxidation normally becomes the main source of energy. In patients with severe β-oxidation impairment, decreased gluconeogenesis and the mitochondrial toxicity of free fatty acids, dicarboxylic acids, and lipid peroxidation products can also impair energy production in other organs. This could explain the severity of microvesicular steatosis and its propensity to cause liver failure, coma, and death. Conversely, a severe primary impairment of OXPHOS can secondarily inhibit β-oxidation, thus causing steatosis, and can also inhibit the mitochondrial catabolism of pyruvate, increasing the likelihood of lactic acidosis. DILI due to mitochondrial dysfunction occurs only in some recipients. Otherwise, the drug would not have been marketed. Both metabolic factors and/or co-morbidity factors play a role in the idiosyncratic occurrence of these adverse effects. Metabolic factors can impair the removal of a toxic parent compound and increase toxicity in some recipients. Furthermore, several different medical conditions may each add their own deleterious effects on mitochondrial function (e.g., aspirin and viral infections; tamoxifen and obesity; valproate and inborn mitochondrial cytopathies; valproate and inborn β-oxidation defects). The mitochondrial mechanisms of DILI have been described only recently and are still not investigated routinely during the preclinical development of new drug molecules. However, cases of microvesicular steatosis have led to the recall of diethylaminoethoxyhexestrol, the discontinuation of clinical trials with ﬁaluridine, limited use of perhexiline or tacrine, as well as early therapeutic misadventures with tetracyclines and valproic acid. Mitochondrial dysfunction also appears to be involved in the toxicity of troglitazone, trovaﬂoxacin, and alatroﬂoxacin, all of which had to be removed from clinical use. We therefore suggest that new drug molecules be screened for possible mitochondrial effects before they are released on the market. REFERENCES 1. Pessayre D, Larrey D, Drug-induced liver injury. In Textbook of Hepatology: From Basic Science to Clinical Practice (J Rodes, JP Benhamou, AT Blei, J Reichen, M Rizzetto, eds). Oxford, UK: Blackwell Publishing; 2007:1211–1268.

188

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

2. Pessayre D. Role of reactive metabolites in drug-induced hepatitis. J Hepatol. 1995; 23:S16–S24. 3. Robin MA, Le Roy M, Descatoire V, Pessayre D. Plasma membrane cytochromes P450 as neoantigens and autoimmune targets in drug-induced hepatitis. J Hepatol. 1997;26(suppl 1):23–30. 4. Fromenty B, Pessayre D. Inhibition of mitochondrial β-oxidation as a mechanism of hepatotoxicity. Pharmacol Ther. 1995;67:101–154. 5. Pessayre D, Fromenty B, Mansouri A, Berson A. Hepatotoxicity due to mitochondrial dysfunction. In Drug-Induced Liver Disease, 2nd ed. (N Kaplowitz, LD DeLeve, eds). New York: Informa Healthcare USA; 2007:49–84. 6. Bagheri H, Michel F, Lapeyre-Mestre M, Lagier E, Cambus JP, Valdigui´e P, Montastruc JL. Detection and incidence of drug-induced liver injuries in hospital: a prospective analysis from laboratory signals. Br J Clin Pharmacol. 2000;50:479–484. 7. Sgro C, Clinard F, Ouazir K, et al. Incidence of drug-induced hepatic injuries: a French population-based study. Hepatology. 2002;36:451–455. 8. Lee WM. Acute liver failure in the United States. Semin Liver Dis. 2003;23:217–226. 9. Zimmerman HJ. Hepatotoxicity: The Adverse Effects of Drugs and Other Chemicals on the Liver . New York: Appleton-Century-Crofts; 1978. 10. Stricker BHC. Drug-Induced Hepatic Injury. 2nd ed. Amsterdam: Elsevier; 1992. 11. Benichou C. Criteria for drug-induced hepatitis: report of an international consensus meeting. J Hepatol. 1990;11:272–276. 12. Maria VAJ, Victorino RMM. Diagnostic value of speciﬁc T cell reactivity to drugs in 95 cases of drug induced liver injury. Gut. 1997;41:534–540. 13. Beaune P, Pessayre D, Dansette P, Mansuy D, Manns M. Autoantibodies against cytochromes P450: role in human diseases. Adv Pharmacol. 1994;30:199–245. 14. Haouzi D, Lekehal M, Moreau A, et al. Cytochrome P450–generated reactive metabolites cause mitochondrial permeability transition, caspase activation and apoptosis in rat hepatocytes. Hepatology. 2000;32:303–311. 15. Pessayre D, Haouzi D, Fau D, Robin MA, Mansouri A, Berson A. Withdrawal of life support, altruistic suicide, fratricidal killing and euthanasia by lymphocytes: different forms of drug-induced hepatic apoptosis. J Hepatol. 1999;31:760–770. 16. Uetrecht JP. New concepts in immunology relevant to idiosyncratic drug reactions: the danger hypothesis and innate immune system. Chem Res Toxicol. 1999;12:387–395. 17. Petrescu I, Tarba C. Uncoupling effects of diclofenac and aspirin in the perfused liver and isolated hepatic mitochondria of rat. Biochem Biophys Acta. 1997;1318:385–394. 18. Moreno-S´anchez R, Bravo C, V´asquez C, Ayala G, Silveira LH, M´artinez-Lav´ın M. Inhibition and uncoupling of oxidative phosphorylation by nonsteroidal anti-inﬂammatory drugs: study in mitochondria, submitochondrial particles, cells and whole heart. Biochem Pharmacol. 1999;57:743–752. 19. Brustovetsky N, Klingenberg M. The reconstituted ADP/ATP carrier can mediate H+ transport by free fatty acids, which is further stimulated by Mersalyl. J Biol Chem. 1994;269:27329–27336. 20. Wieckowski MR, Wojtczak L. Involvement of the dicarboxylate carrier in the protonophoric action of long-chain fatty acids in mitochondria. Biochem Biophys Res Commun. 1997;232:414–417.

REFERENCES

189

21. Costantini P, Chernyak BV, Petronilli V, Bernardi P. Modulation of the mitochondrial permeability transition pore by pyridine nucleotides and dithiol oxidation at two separate sites. J Biol Chem. 1996;271:6746–6751. 22. Catisti R, Vercesi A. The participation of pyridine nucleotides redox state and reactive oxygen in the fatty acid–induced permeability transition in rat liver mitochondria. FEBS Lett. 1999;464:97–101. 23. Masubuchi Y, Nakayama S, Horie T. Role of mitochondrial permeability transition in diclofenac-induced hepatocyte injury in rats. Hepatology. 2002;35:544–551. 24. Berson A, Cazanave S, Descatoire V, et al. The anti-inﬂammatory drug, nimesulide (4-nitro-2-phenoxymethane-sulfoanilide), uncouples mitochondria and induces mitochondrial permeability transition in human hepatoma cells: protection by albumin. J Pharmacol Exp Ther. 2006;318:444–454. 25. Trost LC, Lemasters JJ. The mitochondrial permeability transition: a new pathophysiological mechanism for Reye’s syndrome and toxic liver injury. J Pharmacol Exp Ther. 1996;278:1000–1005. 26. Kowaltowski AJ, Vercesi AE. Reactive oxygen generation by mitochondria. In Mitochondria in Pathogenesis (JJ Lemasters, AL Nieminen, eds). New York: Kluwer Academic/Plenum Publishers; 2001:281–300. 27. Kowaltowski AJ, Castilho RF, Vercesi AE. Mitochondrial permeability transition and oxidative stress. FEBS Lett. 2001;495:12–15. 28. Brunner M, Moeslinger T, Spieckermann PG. Regulation of cyclosporin A sensitive mitochondrial permeability transition by the redox state of pyridine nucleotides. Comp Biochem Physiol B Biochem Mol Biol. 2001;128:31–41. 29. Cai J, Jones DP. Superoxide in apoptosis. Mitochondrial generation triggered by cytochrome c loss. J Biol Chem. 1998;273:11401–11404. 30. Bort R, Ponsoda X, Jover R, G´omez-Lech´on J, Castell JV. Diclofenac toxicity to hepatocytes: a role for drug metabolism in cell toxicity. J Pharmacol Exp Ther. 1999;288:65–72. 31. Fulda S, Scafﬁdi C, Susin SA, et al. Activation of mitochondria and release of mitochondrial apoptogenic factors by betulinic acid. J Biol Chem. 1998;273:33942–33948. 32. Chowdhury AR, Mandal S, Goswami A, et al. Dihydrobetulinic acid induces apoptosis in Leishmania donovani by targeting DNA topoisomerases I and II: implications in antileishmanial therapy. Mol Med. 2003;9:26–36. 33. Ravagnan L, Marzo I, Costantini P, et al. Lonidamide triggers apoptosis via a direct, Bcl-2-inhibited effect on the mitochondrial permeability transition pore. Oncogene. 1999;18:2537–2546. 34. Belzacq AS, El Hamel C, Vieira HL, et al. Adenine nucleotide translocator mediates the mitochondrial membrane permeabilization induced by lonidamide, arsenic acid and CD437. Oncogene. 2001;15:7579–7587. 35. Watkins PB. Insight into hepatotoxicity: the troglitazone experience. Hepatology. 2005;41:229–230. 36. Murphy, Davern TJ, Shakil AO, Shick L, et al. (Acute Liver Failure Study Group). Troglitazone-induced fulminant hepatic failure. Dig Dis Sci. 2000;45:549–553. 37. Funk C, Ponelle C, Scheuermann G, Pantze M. Cholestatic potential of troglitazone as a possible factor contributing to troglitazone-induced hepatotoxicity: in vivo and

190

38.

39.

40.

41.

42.

43.

44.

45.

46.

47. 48.

49.

50.

51. 52.

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

in vitro interaction at the canalicular bile salt export pump (Bsep) in the rat. Mol Pharmacol. 2001;59:627–635. Kassahun K, Pearson PG, Tang W, et al. Studies on the metabolism of troglitazone to reactive intermediates in vitro and in vivo: evidence for novel pathways involving quinone methide formation and thiazolidinedione ring scission. Chem Res Toxicol. 2001;14:62–70. Bae MA, Song BY. Critical role of c-Jun N-terminal protein kinase activation in troglitazone-induced apoptosis of human HepG2 hepatoma cells. Mol Pharmacol. 2003;63:401–408. Shishido S, Koga H, Harada M, Kumemura H, Hanada S, Taniguchi E, Kumashiro R, Ohira H, Sato Y, Namba M, Ueno T, Sata M. Hydrogen peroxide overproduction in megamitochondria of troglitazone-treated human hepatocytes. Hepatology. 2003;37:136–147. Kohlroser J, Mathai J, Reichheld J, Banner BF, Bonkovsky HL. Hepatotoxicity due to troglitazone: report of two cases and review of adverse events reported to the United States Food and Drug Administration. Am J Gastroenterol. 2000;95:272–276. Boya P, Gonzalez-Polo RA, Poncet D, et al. Mitochondrial membrane permeabilization is a critical step of lysosome-initiated apoptosis induced by hydroxychloroquine. Oncogene. 2003;22:3927–3936. Berson A, Descatoire V, Sutton A, et al. Toxicity of alpidem, a peripheral benzodiazepine receptor ligand, but not zolpidem in rat hepatocytes: role of mitochondrial permeability transition and metabolic activation. J Pharmacol Exp Ther. 2001;299:793–800. Pastorino JG, Simbula G, Yamamoto K, Glascott PA, Jr, Rothman RJ, Farber JL. The cytotoxicity of tumor necrosis factor depends on induction of the mitochondrial permeability transition. J Biol Chem. 1996;271:29792–29798. Fisher R, Schmitt M, Bode JG, H¨aussinger D. Expression of the peripheral-type benzodiazepine receptor and apoptosis induction in hepatic stellate cells. Gastroenterology. 2001;120:1212–1226. Lett´eron P, Sutton A, Mansouri A, Fromenty B, Pessayre D. Inhibition of microsomal triglyceride transfer protein: another mechanism for drug-induced steatosis in mice. Hepatology. 2003;38:133–140. Pessayre D, Berson A, Fromenty B, Mansouri A. Mitochondria in steatohepatitis. Semin Liver Dis. 2001;22:57–69. Fr´eneaux E, Labbe G, Lett´eron P, et al. Inhibition of the mitochondrial oxidation of fatty acids by tetracycline in mice and in man: possible role in microvesicular steatosis induced by this antibiotic. Hepatology. 1988;8:1056–1062. Labbe G, Fromenty B, Fr´eneaux E, et al. Effects of various tetracycline derivatives on in vitro and in vivo β-oxidation of fatty acids, egress of triglycerides from the liver, accumulation of hepatic triglycerides, and mortality in mice. Biochem Pharmacol. 1991;41:638–641. Deboyser D, Goethals F, Krack G, Roberfroid M. Investigation into the mechanism of tetracycline-induced steatosis: study in isolated hepatocytes. Toxicol Appl Pharmacol. 1989;97:473–479. Farrell G. Drug-Induced Liver Disease. London: Churchill-Livingstone; 1994. Zimmerman HJ, Ishak KG. Valproate-induced hepatic injury: analyses of 23 fatal cases. Hepatology. 1982;2:591–597.

REFERENCES

191

53. Ponchaut S, Van Hoof F, Veitch K. In Vitro effects of valproate and valproate metabolites on mitochondrial oxidations: relevance of CoA sequestration to the observed inhibitions. Biochem Pharmacol. 1992;43:2435–2442. 54. Turnbull DM, Bone AJ, Bartlett K, Koundakjian PP, Sherratt HSA. The effects of valproate on intermediary metabolism in isolated rat hepatocytes and intact rats. Biochem Pharmacol. 1983;32:1887–1892. 55. Silva MFB, Ruiter JPN, Ijlst L, et al. Valproate inhibits the mitochondrial pyruvate-driven oxidative phosphorylation. J Inherited Metab Dis. 1997;20:397–400. 56. Njølstad PR, Skjeldal OH, Agsteribbe E, et al. Medium chain acyl-CoA dehydrogenase deﬁciency and fatal valproate toxicity. Pediatr Neurol. 1997;16:160–162. 57. Kottlors M, Jaksch M, Ketelsen UP, Weiner S, Glocker FX, L¨ucking CH. Valproic acid triggers acute rhabdomyolysis in a patient with carnitine palmitoyltransferase type II deﬁciency. Neuromuscular Disord. 2001;11:757–759. 58. Chabrol B, Mancini J, Chretien D, Rustin P, Munnich A, Pinsard N. Valproate-induced hepatic failure in a case of cytochrome c oxidase deﬁciency. Eur J Pediatr. 1994;153:133–135. 59. Lam CW, Lau CH, Williams JC, Cahn YW, Wong LJC. Mitochondrial myopathy, encephalopathy, lactic acidosis and stroke-like episodes (MELAS) triggered by valproate therapy. Eur J Pediatr. 1997;156:562–564. 60. Kr¨ahenb¨uhl S, Brandner S, Kleinle S, Liechti S, Straussmann D. Mitochondrial diseases represent a risk factor for valproate-induced fulminant liver failure. Liver. 2000;20:346–348. 61. Sadeque AJM, Fisher MB, Korzekwa KR, Gonzalez FJ, Rettie AE. Human CYP2C9 and CYP2A6 mediate formation of the hepatotoxin 4-ene-valproic acid. J Pharmacol Exp Ther. 1997;283:698–703. 62. Bjorge SM, Baillie TA. Inhibition of medium-chain fatty acid β-oxidation in vitro by valproic acid and its unsaturated metabolite, 2-n-propyl-4-pentenoic acid. Biochem Biophys Res Commun. 1985;132:245–252. 63. Kassahun K, Farrell K, Abbott F. Identiﬁcation and characterization of the glutathione and N -acetylcysteine conjugates of (E )-2-propyl-2,4-pentadienoic acid, a toxic metabolite of valproic acid, in rats and humans. Drug Metab Dispos. 1991;19:525–535. 64. Kassahun K, Hu P, Grillo MP, Davis MR, Jin L, Baillie TA. Metabolic activation of unsaturated derivatives of valproic acid. Identiﬁcation of novel glutathione adducts formed through coenzyme A–dependent and –independent mechanisms. Chem Biol Interact. 1994;90:253–275. 65. Levy RH, Rettenmeier AW, Anderson GD, et al. Effects of polytherapy with phenytoin, carbamazepine and stiripentol on formation of 4-ene-valproate, a hepatotoxic metabolite of valproic acid. Clin Pharmacol Ther. 1990;48:225–235. 66. Killenberg PG, Davidson ED, Webster LT. Evidence for a medium-chain fatty acid coenzyme A ligase (adenosine monophosphate) that activates salicylate. Mol Pharmacol. 1971;7:260–268. 67. Deschamps D, Fisch C, Fromenty B, Berson A, Degott C, Pessayre D. Inhibition by salicylic acid of the activation and thus oxidation of long-chain fatty acids. Possible role in the development of Reye’s syndrome. J Pharmacol Exp Ther. 1991;259:894–904.

192

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

68. Oh KW, Qian T, Brenner DA, Lemasters JJ. Salicylate enhances necrosis and apoptosis mediated by the mitochondrial permeability transition. Toxicol Sci. 2003;73:44–52. 69. Zimmerman HJ. Effects of aspirin and acetaminophen on the liver. Arch Intern Med. 1981;141:333–342. 70. Partin JS, Daugherty CC, McAdams AJ, Partin JC, Schubert WK. A comparison of liver ultrastructure in salicylate intoxication and Reye’s syndrome. Hepatology. 1984;4:687–690. 71. Borutait´e V, Brown GC. Rapid reduction of nitric oxide by mitochondria, and reversible inhibition of mitochondrial respiration by nitric oxide. Biochem J. 1996;315:295–299. 72. Susin SA, Zamzami N, Kroemer G. Mitochondria as regulators of apoptosis: doubt no more. Biochim Biophys Acta. 1998;1366:151–165. 73. Pessayre D, Feldmann G, Haouzi D, Fau D, Moreau A, Neuman M. Hepatocyte apoptosis triggered by natural substances (cytokines, other endogenous substances and foreign toxins). In Apoptosis and Its Modulation by Drugs (RG Cameron, G Feuer, eds). Heidelberg, Germany: Springer-Verlag; Handb Exp Pharmacol. 2000;142:59–108. 74. Hurwitz ES, Barrett MJ, Bregman D, et al. Public health service study on Reye’s syndrome and medications. Report of the pilot phase. N Engl J Med. 1985;313:849–857. 75. Forsyth BW, Horwitz RI, Acampora D, et al. New epidemiologic evidence conﬁrming that bias does not explain the aspirin/Reye’s syndrome association. JAMA. 1989;261:2517–2524. 76. Remington PL, Rowley D, McGee H, Hall WN, Monto AS. Decreasing trends in Reye syndrome and aspirin use in Michigan, 1979 to 1984. Pediatrics. 1986;77:93–98. 77. Rowe PC, Valle D, Brusilow SW. Inborn errors of metabolism in children referred with Reye’s syndrome: a changing pattern. JAMA. 1988;260:3167–3170. 78. Kaplan MM. Acute fatty liver of pregnancy. N Engl J Med. 1985;313:367–370. 79. Ebert EC, Sun EA, Wright SH, et al. Does early diagnosis and delivery in acute fatty liver of pregnancy lead to improvement in maternal and infant survival? Dig Dis Sci. 1984;29:453–455. 80. Grimbert S, Fromenty B, Fisch C, et al. Decreased mitochondrial oxidation of fatty acids in pregnant mice: possible relevance to development of acute fatty liver of pregnancy. Hepatology. 1993;17:628–637. 81. Grimbert S, Fisch C, Deschamps D, et al. Effects of female sex hormones on liver mitochondria in non-pregnant female mice: possible role in acute fatty liver of pregnancy. Am J Physiol 1995;268 (Gastrointest Liver Physiol. 31):G107–G115. 82. Ibdah JA, Bennet MJ, Rinaldo P, et al. A fetal fatty-acid oxidation disorder as a cause of liver disease in pregnant women. N Engl J Med. 1999;340:1723–1731. 83. Shekhawat P, Bennett MJ, Sadovsky Y, Nelson DM, Rakheja D, Strauss AW. Human placenta metabolizes fatty acids: implications for fetal fatty acid oxidation disorders and maternal liver diseases. Am J Physiol Endocrinol Metab. 2003;284:E1098–E1105. 84. Bravo JF, Jacobson MP, Mertens BF. Fatty liver and pleural effusion with ibuprofen therapy. Ann Intern Med. 1997;87:200–201.

REFERENCES

193

85. Victorino RM, Silveira JC, Baptista A, de Moura MC. Jaundice associated with naproxen. Postgrad Med J. 1980;56:368–370. 86. Danan G, Trunet P, Bernuau J, et al. Pirprofen-induced fulminant hepatitis. Gastroenterology. 1985;89:210–213. 87. Dutertre JP, Bastides F, Jonville AP, et al. Microvesicular steatosis after ketoprofen administration. Eur J Gastroenterol Hepatol. 1991;3:953–954. 88. Fr´eneaux E, Fromenty B, Berson A, et al. Stereoselective and nonstereoselective effects of ibuprofen enantiomers on mitochondrial β-oxidation of fatty acids. J Pharmacol Exp Ther. 1990;255:529–535. 89. Gen`eve J, Hayat-Bonan B, Labbe G, et al. Inhibition of mitochondrial β-oxidation of fatty acids by pirprofen. Role in microvesicular steatosis due to this nonsteroidal anti-inﬂammatory drug. J Pharmacol Exp Ther. 1987;242:1133–1137. 90. Lett´eron P, Brahimi-Bourouina N, Robin MA, Moreau A, Feldmann G, Pessayre D. Glucocorticoids inhibit mitochondrial matrix acyl-CoA dehydrogenases and fatty acid β-oxidation. Am J Physiol. 1997;272 (Gastrointest Liver Physiol . 35):G1141–G1150. 91. Itoh S, Igarashi M, Tsukada Y, Ichinoe A. Non-alcoholic fatty liver with alcoholic hyalin after long-term glucocorticoid therapy. Acta Hepato-Gastroenterol. 1977;24:415–418. 92. Gen`eve J, Larrey D, Amouyal G, Belghiti J, Pessayre D. Metabolic activation of the tricyclic antidepressant amineptine by human liver cytochrome P-450. Biochem Pharmacol. 1987;36:2421–2424. 93. Larrey D, Tinel M, Lett´eron P, Maurel P, Loeper J, Belghiti J, Pessayre D. Metabolic activation of the new tricyclic antidepressant tianeptine by human liver cytochrome P-450. Biochem Pharmacol. 1990;40:545–550. 94. Sbarra C, Castelli MG, Noseda A, Fanelli R. Pharmacokinetics of amineptine in man. Eur J Drug Metab Pharmacokinet. 1981;6:123–126. 95. Grislain L, Gel´e P, Bertrand M, et al. The metabolic pathways of tianeptine, a new antidepressant, in healthy volunteers. Drug Metab Dispos. 1990;18:804–808. 96. Le Dinh T, Fr´eneaux E, Labbe G, et al. Amineptine, a tricyclic antidepressant, inhibits the mitochondrial oxidation of fatty acids and produces microvesicular steatosis of the liver in mice. J Pharmacol Exp Ther. 1988;247:745–750. 97. Fromenty B, Fr´eneaux E, Labbe G, et al. Tianeptine, a new tricyclic antidepressant metabolized by β-oxidation of its heptanoic side chain, inhibits the mitochondrial oxidation of medium and short chain fatty acids in mice. Biochem Pharmacol. 1989;38:3743–3751. 98. Noda S, Umezaki H, Yamamoto K, Araki T, Murakami T, Ishii N. Reye-like syndrome following treatment with the pantothenic acid antagonist, calcium hopantenate. J Neurol Neurosurg Psychiatry. 1988;51:582–585. 99. Ulrich RG, Bacon JA, Brass EP, Cramer CT, Petrella DK, Sun EL. Metabolic, idiosyncratic toxicity of drugs: overview of the hepatic toxicity induced by the anxiolytic, panadiplon. Chem Biol Interact. 2001;134:251–270. 100. Melegh B, Pap M, Moln´ar D, Masszi G, Kopcs´anyi G. Carnitine administration ameliorates the changes in energy metabolism caused by short-term pivampicillin medication. Eur J Pediatr. 1997;156:795–799.

194

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

101. Pessayre D, Fromenty B, Mansouri A. Drug-induced steatosis and steatohepatitis. In Mitochondria in Pathogenesis (JJ Lemasters, AL Niemenen, eds). New York: Plenum Press; 2001:489–517. 102. Kodavanti UP, Mehendale HM. Cationic amphiphilic drugs and phospholipid storage disorder. Pharmacol Rev. 1990;42:327–354. 103. Guigui B, Perrot S, Berry JP, et al. Amiodarone-induced hepatic phospholipidosis: a morphological alteration independent of pseudoalcoholic liver disease. Hepatology. 1988;8:1063–1068. 104. Fromenty B, Fisch C, Labbe G, et al. Amiodarone inhibits the mitochondrial β-oxidation of fatty acids and produces microvesicular steatosis of the liver in mice. J Pharmacol Exp Ther. 1990;255:1371–1376. 105. Fromenty B, Fisch C, Berson A, Lett´eron P, Larrey D, Pessayre, D. Dual effect of amiodarone on mitochondrial respiration. Initial protonophoric uncoupling effect followed by inhibition of the respiratory chain at the levels of complex I and complex II. J Pharmacol Exp Ther. 1990;255:1377–1384. 106. Fromenty B, Lett´eron P, Fisch C, Berson A, Deschamps D, Pessayre D. Evaluation of human blood lymphocytes as a model to study the effects of drugs on human mitochondria: effects of low concentrations of amiodarone on fatty acid oxidation, ATP levels and cell survival. Biochem Pharmacol. 1993;46:421–432. 107. Deschamps D, De Beco V, Fisch C, Fromenty B, Guillouzo A, Pessayre D. Inhibition by perhexiline of oxidative phosphorylation and the β-oxidation of fatty acids: possible role in pseudoalcoholic liver lesions. Hepatology. 1994;19:948–961. 108. Berson A, De Beco V, Lett´eron P, et al. Steatohepatitis-inducing drugs cause mitochondrial dysfunction and lipid peroxidation in rat hepatocytes. Gastroenterology. 1998;114:764–774. 109. De la Iglesia FA, Feuer G, Takada A, Matsuda Y. Morphologic studies on secondary phospholipidosis in humans. Lab Invest. 1974;4:539–549. 110. Shikata T, Kanetaka T, Endo Y, Nagashima K. Drug-induced generalized phospholipidosis. Acta Pathol Jpn. 1972;22:517–531. 111. Pessayre D, Bichara M, Feldmann G, Degott C, Potet F, Benhamou JP. Perhexiline maleate-induced cirrhosis. Gastroenterology. 1979;76:170–177. 112. Simon JB, Manley PN, Brien JF, Armstrong PW. Amiodarone hepatotoxicity simulating alcoholic liver disease. N Engl J Med. 1984;311:167–172. 113. Poucell S, Ireton J, Valencia-Mayoral P, et al. Amiodarone-associated phospholipidosis and ﬁbrosis of the liver: light, immunohistochemical, and electron microscopic studies. Gastroenterology. 1984;86:926–936. 114. Lewis JH, Ranard RC, Caruso A, et al. Amiodarone hepatotoxicity: prevalence and clinicopathologic correlations among 104 patients. Hepatology. 1989;9:679–685. 115. Cortez-Pinto H, Baptista A, Camilo MA, de Costa EB, Valente A, de Moura MC. Tamoxifen-associated hepatitis: report of three cases. J Hepatol. 1995;23:95–97. 116. Oien KA, Moffat D, Curry GW, et al. Cirrhosis with steatohepatitis after adjuvant tamoxifen. Lancet . 1999;353:36–37. 117. Bruno S, Maisonneuve P, Castellana P, Rotmensz N, et al. Incidence and risk factors for non-alcoholic steatohepatitis: prospective study of 5408 women enrolled in Italian tamoxifen chemoprevention trial. Br Med J. 2005;330:932.

REFERENCES

195

118. Larosche I, Lett´eron P, Fromenty B, et al. Tamoxifen inhibits topoisomerases, depletes mitochondrial DNA, and triggers steatosis in mouse liver. J Pharmacol Exp Ther. 2007;321:526–535. 119. Altan N, Chen Y, Schindler M, Simon SM. Tamoxifen inhibits acidiﬁcation in cells independent of the estrogen receptor. Proc Natl Acad Sci U.S.A 1999;96:4432–4437. 120. Ioannidis N. Lipidosis induced in rat uteri by high doses of tamoxifen. Ann Anat. 1998;180:315–319. 121. Berson A, Fau D, Fornacciari R, Degove-Goddard P, Sutton A, Descatoire V, Haouzi D, Lett´eron P, Moreau A, Feldmann G, Pessayre D. Mechanism for experimental buprenorphine hepatotoxicity: major role of mitochondrial dysfunction versus metabolic activation. J Hepatol. 2001;34:261–269. 122. Berson A, Gervais A, Cazals D, Boyer N, Durand F, Bernuau J, Marcellin P, Degott C, Valla D, Pessayre D. Hepatitis after buprenorphine misuse in heroin addicts. J Hepatol. 2001;34:346–350. 123. V´asquez-Vivar J, Augusto O. Hydroxylated metabolites of the antimalarial drug primaquine: oxidation and redox cycling. J Biol Chem. 1992;267:6848–6854. 124. Maggs JL, Kitteringham NR, Breckenridge AM, Park BK. Autoxidative formation of a chemically reactive intermediate from amodiaquine, a myelotoxin and hepatotoxin in man. Biochem Pharmacol. 1987;36:2061–2062. 125. Larrey D, Castot A, Pessayre D, M´erigot P, Machayekhy JP, Feldmann G, Lenoir A, Rueff B, Benhamou JP. Amodiaquine-induced hepatitis: a report of seven cases. Ann Intern Med. 1986;104:801–803. 126. Ginsburg H, Geary TG. Current concepts and new ideas on the mechanism of action of quinoline-containing antimalarials. Biochem Pharmacol. 1987;36:1567–1576. 127. Hostetler KY, Reasor M, Yazaki PJ. Chloroquine-induced phospholipid fatty liver: measurement of drug and lipid concentrations in rat liver lysosomes. J Biol Chem. 1985;260:215–219. 128. Katewa SD, Katyare SS. Treatment with antimalarials adversely affects the oxidative energy metabolism in rat liver mitochondria. Drug Chem Toxicol. 2004;27:41–53. 129. Kaufmann P, Torok M, Hanni A, Roberts P, Gasser R, Kr¨ahenb¨uhl S. Mechanisms of benzarone and benzbromarone-induced hepatic toxicity. Hepatology. 2005;41:925–935. 130. Babany G, Larrey D, Pessayre D, Degott C, Rueff B, Benhamou JP. Chronic active hepatitis caused by benzarone. J Hepatol. 1987;5:332–335. 131. van der Klauw MM, Houtman PM, Stricker BH, Spoelstra P. Hepatic injury caused by benzbromarone. J Hepatol. 1994;20:376–379. 132. Matsuno-Yagi A, Hateﬁ Y. Studies on the mechanism of oxidative phosphorylation: different effects of F0 inhibitors on unisite and multisite ATP hydrolysis by bovine submitochondrial particles. J Biol Chem. 1993;268:1539–1545. 133. Salomon AR, Voehringer DW, Herzenberg LA, Khosla C. Apoptolidin, a selective cytotoxic agent, is an inhibitor of F0 F1 -ATPase. Chem Biol . 2001;8:71–80. 134. Ebel RE, Lardy HA. Inﬂuence of aurovertin on mitochondrial ATPase activity. J Biol Chem. 1975;250:4992–4995. 135. Linnett PE, Mitchell AD, Osselton MD, Muhleirn LJ, Beechey RB. Citreoviridin, a speciﬁc inhibitor of the mitochondrial adenosine triphosphatase. Biochem J . 1978;170:503–510.

196

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

136. Cross RL, Kohlbrenner WE. Efrapeptin prevents modiﬁcation by phenylglyoxal of an essential arginyl residue in mitochondrial adenosine triphosphatase. J Biol Chem. 1978;253:4865–4873. 137. Qian T, Herman B, Lemasters J. The mitochondrial permeability transition mediates both necrotic and apoptotic death of hepatocytes exposed to Br-A23187. Toxicol Appl Pharmacol . 1999;154:117–125. 138. Zheng J, Ramirez VD. Rapid inhibition of rat brain mitochondrial proton F0F1-ATPase activity by estrogens: comparison with Na+ , K+ -ATPase of porcine cortex. Eur J Pharmacol . 1999;368:95–102. 139. Zheng J, Ramirez VD. Inhibition of mitochondrial proton F0F1-ATPase/ATP synthase by polyphenolic phytochemicals. Br J Pharmacol . 2000;130:1115–1123. 140. Kokoszka JE, Waymire KG, Levy SE, et al. The ADP/ATP translocator is not essential for the mitochondrial permeability transition pore. Nature. 2004;427:461–465. 141. Buchanan BB, Eiermann W, Riccio P, Aquila H, Klingenberg M. Antibody evidence for different conformational states of ADP/ATP translocator protein isolated from mitochondria. Proc Natl Acad Sci USA. 1976;73:2280–2284. 142. Green DR, Reed JC. Mitochondria and apoptosis. Science. 1998;281:1309–1312. 143. Weaver JG, Tarze A, Moffat TC, et al. Inhibition of adenine nucleotide translocator pore function and protection against apoptosis in vivo by an HIV protease inhibitor. J Clin Invest. 2005;115:1828–1838. 144. Chua BH, Shrago E. Reversible inhibition of adenine nucleotide translocation by long chain acyl-CoA esters in bovine heart mitochondria and inverted submitochondrial particles. J Biol Chem. 1977;252:6711–6714. 145. Furuno T, Kanno T, Arita K, Asami M, Utsumi T, Doi Y, Inoue M, Utsumi K. Roles of long chain fatty acids and carnitine in mitochondrial membrane permeability transition. Biochem Pharmacol . 2001;62:1037–1046. 146. Esposito LA, Melov S, Panov A, Cottrell BA, Wallace DC. Mitochondrial disease in mouse results in increased oxidative stress. Proc Natl Acad Sci USA. 1999;96:4820–4825. 147. Mazat JP, Rossignol R, Malgat M, Rocher C, Faustin B, Letellier T. What do mitochondrial diseases teach us about mitochondrial function. . . that we already knew: threshold expression of mitochondrial defects. Biochim Biophys Acta. 2001;1504:20–30. 148. Watmough NJ, Bindoff LA, Birch-Machin MA, et al. Impaired mitochondrial β-oxidation in a patient with an abnormality of the respiratory chain: studies in skeletal muscle mitochondria. J Clin Invest. 1990;85:177–184. 149. Lett´eron P, Duchatelle V, Berson A, et al. Increased ethane exhalation, an in vivo index of lipid peroxidation, in alcohol abusers. Gut . 1993;34:409–414. 150. Kukielka E, Cederbaum AI. NADH-dependent microsomal interaction with ferric complexes and production of reactive oxygen species. Arch Biochem Biophys. 1989;275:540–550. 151. Dai Y, Rashba-Step J, Cederbaum AI. Stable expression of human cytochrome P450 2E1 in HepG2 cells: characterization of catalytic activities and production of reactive oxygen intermediates. Biochemistry. 1993;32:6928–6937. 152. Kukielka E, Dicker E, Cederbaum AI. Increased production of reactive oxygen species by rat liver mitochondria after chronic ethanol treatment. Arch Biochem Biophys. 1994;309:377–386.

REFERENCES

197

153. Fernandez-Checa JC, Garcia-Ruiz C, Ookhtens M, Kaplowitz N. Impaired uptake of glutathione by hepatic mitochondria from chronic ethanol-fed rats: tracer kinetic studies in vitro and in vivo and susceptibility to oxidant stress. J Clin Invest. 1991;87:397–405. 154. Polavarapu R, Spitz DR, Sim JE, et al. Increased lipid peroxidation and impaired antioxidant enzyme function is associated with pathological liver injury in experimental alcoholic liver disease in rats fed diets high in corn oil and ﬁsh oil. Hepatology. 1998;27:1317–1323. 155. Uesugi T, Froh M, Arteel GE, Bradford BU, Thurman RG. Toll-like receptor 4 is involved in the mechanism of early alcohol-induced liver injury in mice. Hepatology. 2001;34:101–108. 156. Nordmann R, Ribi`ere C, Rouach H. Implication of free radical mechanisms in ethanol-induced cellular injury. Free Radic Biol Med . 1992;12:219–240. 157. Wieland P, Lauterburg BH. Oxidation of mitochondrial proteins and DNA following administration of ethanol. Biochem Biophys Res Commun. 1995;213:815–819. 158. Mansouri A, Gaou I, De Kerguenec C, et al. An alcoholic binge causes massive degradation of hepatic mitochondrial DNA in mice. Gastroenterology. 1999;117:181–190. 159. Mansouri A, Demeilliers C, Amsellem S, Pessayre D, Fromenty B. Acute ethanol administration oxidatively damages and depletes mitochondrial DNA in mouse liver, brain, heart and skeletal muscles: protective effects of antioxidants. J Pharmacol Exp Ther . 2001;298:737–743. 160. Demeilliers C, Maisonneuve C, Grodet A, et al. Impaired adaptive resynthesis and prolonged depletion of hepatic mitochondrial DNA after repeated alcohol binges in mice. Gastroenterology. 2002;123:1278–1290. 161. Fromenty B, Grimbert S, Mansouri A, et al. Hepatic mitochondrial DNA deletion in alcoholics: association with microvesicular steatosis. Gastroenterology. 1995;108:193–200. 162. Mansouri A, Fromenty B, Berson A, et al. Multiple hepatic mitochondrial DNA deletions suggest premature oxidative aging in alcoholic patients. J Hepatol . 1997;27:96–102. 163. Hillan KJ, Logan MC, Ferrier RK, et al. Hepatocyte proliferation and serum hepatocyte growth factor levels in patients with alcoholic hepatitis. J Hepatol . 1996;24:385–390. 164. Ballardini G, Groff P, Zoli M, et al. Increased risk of hepatocellular carcinoma development in patients with cirrhosis and with high hepatocellular proliferation. J Hepatol . 1994;20:218–222. 165. Tsuchishima M, Tsutsumi M, Shiroeda H, et al. Study of mitochondrial DNA deletion in alcoholics. Alcohol Clin Exp Res. 2000;24:12S–15S. 166. Lieber CS. Hepatic, metabolic and toxic effects of ethanol: 1991 update. Alcohol Clin Exp Res. 1991;15:573–592. 167. You M, Fischer M, Deeg MA, Crabb DW. Ethanol induces fatty acid synthesis pathways by activation of sterol regulatory element-binding protein (SREBP). J Biol Chem. 2002;277:29342–29347. 168. Grunnet N, Kondrup J. The effect of ethanol on the β-oxidation of fatty acids. Alcohol Clin Exp Res. 1986;10:64S–68S.

198

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

169. Lin MCM, Li JJ, Wang EJ, Princler GL, Kauffman FC, Kung HF. Ethanol down-regulates the transcription of microsomal triglyceride transfer protein gene. FASEB J . 1997;11:1145–1152. 170. Bajt ML, Knight TR, Farhood A, Jaeschke H. Scavenging peroxynitrite with glutathione promotes regeneration and enhances survival during acetaminophen-induced liver injury in mice. J Pharmacol Exp Ther . 2003;307:67–73. 171. Gunawan BK, Liu ZX, Han D, Hanawa N, Gaarde WA, Kaplowitz N. c-Jun N-terminal kinase plays a major role in murine acetaminophen hepatotoxicity. Gastroenterology. 2006;131:165–178. 172. Masubuchi Y, Suda C, Horie T. Involvement of mitochondrial permeability transition in acetaminophen-induced liver injury in mice. J Hepatol . 2005;42:110–116. 173. Reid AB, Kurten RC, McCullough SS, Brock RW, Hinson JA. Mechanisms of acetaminophen-induced hepatotoxicity: role of oxidative stress and mitochondrial permeability transition in freshly isolated mouse hepatocytes. J Pharmacol Exp Ther . 2005;312:509–516. 174. Cover C, Mansouri A, Knight TR, Bajt ML, Lemasters JJ, Pessayre D, Jaeschke H. Peroxynitrite-induced mitochondrial and endonuclease-mediated nuclear DNA damage in acetaminophen hepatotoxicity. J Pharmacol Exp Ther . 2005;315:879–887. 175. Bender RP, Lindsey RH, Burden DA Osheroff N. N -Acetyl-benzoquinone imine, the toxic metabolite of acetaminophen, is a topoisomerase II poison. Biochemistry. 2004;43:3731–3739. 176. Rowe TC, Weissig V, Lawrence JW. Mitochondrial DNA metabolism targeting drugs. Adv Drug Deliv Rev . 2001;49:175–187. 177. Schneider E, Hsiang YH, Liu LF. DNA topoisomerases as anticancer drug targets. Adv Pharmacol . 1990;21:149–183. 178. Li TK, Liu LF. Tumor cell death induced by topoisomerase-targeting drugs. Annu Rev Pharmacol Toxicol . 2001;41:53–77. 179. Zhang H, Barcel´o JM, Lee B, et al. Human mitochondrial topoisomerase I. Proc Natl Acad Sci USA. 2001;98:10608–10613. 180. Lin JH, Castora FJ. DNA topoisomerase II from mammalian mitochondria is inhibited by the antitumor drugs, m-ASA and VM-26. Biochem Biophys Res Commun. 1991;176:690–697. 181. Watkins PB, Zimmermann HJ, Knapp MJ, Gracon SI, Lewis KW. Hepatotoxic effects of tacrine administration in patients with Alzheimer’s disease. JAMA. 1994;271:992–998. 182. Berson A, Renault S, Lett´eron P, et al. Uncoupling of rat and human mitochondria: a possible explanation for tacrine-induced liver dysfunction. Gastroenterology. 1996;110:1878–1890. 183. Mansouri A, Haouzi D, Descatoire V, et al. Tacrine inhibits topoisomerases and DNA synthesis to cause mitochondrial DNA depletion and apoptosis in mouse liver. Hepatology. 2003;38:715–725. 184. Lawrence JW, Darkin-Rattray S, Xie F, Neims AH, Rowe TC. 4-Quinolones cause a selective loss of mitochondrial DNA from mouse L1210 leukemia cells. J Cell Biochem. 1993;51:165–174. 185. Lawrence JW, Claire DC, Weissig V, Rowe TC. Delayed cytotoxicity and cleavage of mitochondrial DNA in ciproﬂoxacin-treated mammalian cells. Mol Pharmacol . 1996;50:1178–1188.

REFERENCES

199

186. Labowitz JK, Silverman WB. Cholestatic jaundice induced by ciproﬂoxacin. Dig Dis Sci . 1997;42:192–194. 187. Lopez-Navidad A, Domingo P, Cadafalch J, Farrerons J. Norﬂoxacin-induced hepatotoxicity. J Hepatol . 1990;11:277–278. 188. Boya P, Andreau K, Poncet D, et al. Lysosomal membrane permeabilization induces cell death in a mitochondria-dependent fashion. J Exp Med . 2003;197:1323–1334. 189. Liguori MJ, Anderson MG, Bukofzer S, et al. Microarray analysis in human hepatocytes suggests a mechanism for hepatotoxicity induced by trovaﬂoxacin. Hepatology. 2005;41:177–186. 190. Lazarczyk DA, Goldstein NS, Gordon SC. Trovaﬂoxacin hepatotoxicity. Dig Dis Sci . 2001;46:925–926. 191. Zhang Y, Bell A, Perlman PS, Leibowitz MJ. Pentamidine inhibits mitochondrial intron splicing and translation in Saccharomyces cerevisiae. RNA. 2000;6:937–951. 192. Nass MM. Analysis of methylglyoxal bis(guanylhydrazone)-induced alterations of hamster tumor mitochondria by correlated studies of selective rhodamine binding, ultrastructural damage, DNA replication and reversibility. Cancer Res. 1984;44:2677–2688. 193. Bergeron RJ, Neims AH, McManis JS, et al. Synthetic polyamine as antineoplastics. J Med Chem. 1988;31:1183–1190. 194. Wiseman A, Attardi G. Reversible tenfold reduction in mitochondria DNA content of human cells treated with ethidium bromide. Mol Gen Genet. 1978;167:51–63. 195. Esnault C, Brown SC, Segal-Bendirdjian E, et al. Selective alteration of mitochondrial function by dirtercalinium (NSC 335153), a DNA bisintercalating agent. Biochem Pharmacol . 1990;39:109–122. 196. Schneider Berlin KR, Ammini CV, Rowe TC. Dequalinium induces a selective depletion of mitochondrial DNA from HeLa human cervical carcinoma cells. Exp Cell Res. 1998;245:137–145. 197. Miyako K, Kai Y, Irie T, Takeshige K, Kang D. The content of intracellular mitochondrial DNA is decreased by 1-methyl-4-phenylpyridium ion (MPP+ ). J Biol Chem. 1997;272:9605–9608. 198. Mitsuya H, Broder S. Inhibition of the in vitro infectivity and cytopathic effect of human T-lymphotrophic virus type III/lymphadenopathy-associated virus (HTLV-III/LAV) by 2 ,3 -dideoxynucleosides. Proc Natl Acad Sci USA. 1986;83:1911–1915. 199. Yarchoan R, Mitsuya H, Myers CE, Broder S. Clinical pharmacology of 3 -azido-2 ,3 -dideoxythymidine (zidovudine) and related dideoxynucleosides. N Engl J Med . 1989;321:726–738. 200. Chen CH, Cheng YC. Delayed cytotoxicity and selective loss of mitochondrial DNA in cells treated with the anti-human immunodeﬁciency virus compound 2 ,3 -dideoxycytidine. J Biol Chem. 1989;264:11934–11937. 201. Lewis W, Dalakas MC. Mitochondrial toxicity of antiviral drugs. Nat Med . 1995;1:417–422. 202. Johnson AA, Ray AS, Hanes J, et al. Toxicity of antiviral nucleoside analogs and the human mitochondrial DNA polymerase. J Biol Chem. 2001;276:40847–40857. 203. Lynx MD, McKee EE. 3 -Azido-3 -deoxythymidine (AZT) is a competitive inhibitor of thymidine phosphorylation in isolated rat heart and liver mitochondria. Biochem Pharmacol . 2006;72:239–243.

200

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

204. Walker UA, Venhoff N. Uridine in the prevention and treatment of NRTI-related mitochondrial toxicity. Antivir Ther. 2005;10(suppl 2):M117–M123. 205. Banasch M, Goetze O, Knyhala K, et al. Uridine supplementation enhances hepatic mitochondrial function in thymidine-analogue treated HIV-infected patients. AIDS . 2006;20:1554–1556. 206. Chariot P, Drogou I, de Lacroix-Szmania I, et al. Zidovudine-induced mitochondrial disorder with massive liver steatosis, myopathy, lactic acidosis and mitochondrial DNA depletion. J Hepatol . 1999;30:156–160. 207. Walker UA, B¨auerle J, Laguno M, et al. Depletion of mitochondrial DNA in liver under antiretroviral therapy with didanosine, stavudine or zalcitabine. Hepatology. 2004;39:311–317. 208. Brivet FG, Nion I, M´egarbane B, Slama A, Brivet P, Rustin P, Munnich A. Fatal lactic acidosis and liver steatosis associated with didanosine and stavudine treatment: a respiratory chain dysfunction? J Hepatol . 1999;30:364–365. 209. Lewis W, Copeland WC, Day BJ. Mitochondrial DNA depletion, oxidative stress and mutation: mechanisms of dysfunction from nucleoside reverse transcriptase inhibitors. Lab Invest. 2001;81:777–790. 210. Barile M, Valenti D, Passarella S, Quagliariello E. 3 -Azido-3 -deoxythymidine uptake into isolated rat liver mitochondria and impairment of ADP/ATP translocator. Biochem Pharmacol . 1997;53:913–920. 211. de la Asuncion JG, del Olmo ML, Sastre J, Pallardo FV, Vina J. Zidovudine (AZT) causes an oxidation of mitochondrial DNA in mouse liver. Hepatology. 1999;29:985–987. 212. de la Asuncion JG, del Olmo ML, Sastre J, Millan A, Pellin A, Pallardo FV. AZT treatment induces molecular and ultrastructural oxidative damage to muscle mitochondria: prevention by antioxidant vitamins. J Clin Invest. 1998;102:4–9. 213. Martin AM, Hammond E, Nolan D, et al. Accumulations of mitochondrial DNA mutations in human immunodeﬁciency virus-infected patients treated with nucleoside-analogue reverse transcriptase inhibitors. Am J Hum Genet. 2003;72:549–560. 214. White DJ, Mital D, Taylor S, St John JC. Sperm mitochondrial DNA deletions as a consequence of long-term highly active antiretroviral therapy. AIDS . 2001;15:1061–1062. 215. Bartley PB, Westacott L, Boots RJ, et al. Large hepatic mitochondrial DNA deletions associated with l-lactic acidosis and highly active antiretroviral therapy. AIDS . 2001;15:419–420. 216. Mallon PWG, Unemori P, Sedwell R, et al. In Vivo, nucleoside reverse transcriptase inhibitors alter expression of both mitochondrial and lipid metabolism genes in the absence of depletion of mitochondrial DNA. J Infect Dis. 2005;191:1686–1696. 217. Pessayre D, Fromenty B. NASH: a mitochondrial disease. J Hepatol . 2005;42: 928–940. 218. Kamal MA, French SW. Drug-induced mitochondrial biogenesis in a liver biopsy. Exp Mol Pathol . 2004;77:201–204. 219. McKenzie R, Fried MW, Sallie R, et al. Hepatic failure and lactic acidosis due to ﬁaluridine (FIAU), an investigational nucleoside analogue for chronic hepatitis B. N Engl J Med . 1995;333:1099–1105.

REFERENCES

201

220. Lewis W, Levine ES, Griniuviene B, Tankersley KO, et al. Fialuridine and its metabolites inhibit DNA polymerase γ at sites of multiple adjacent analog incorporation, decrease mtDNA abundance, and cause mitochondrial structural defects in cultured hepatoblasts. Proc Natl Acad Sci USA. 1996;93:3592–3597. 221. Thust R, Tomicic M, Kl¨ocking R, Voutilainen N, Wutzler P, Kaina B. Comparison of the genotoxic and apoptosis-inducing properties of ganciclovir and penciclovir in Chinese hamster ovary cells transfected with the thymidine kinase gene of herpes simplex virus-1: implications for gene therapeutic approaches. Cancer Gene Ther. 2000;7:107–117. 222. Herraiz M, Beraza N, Solano A, et al. Liver failure caused by herpes simplex virus thymidine kinase plus ganciclovir therapy is associated with mitochondrial dysfunction and mitochondrial DNA depletion. Hum Gene Ther. 2003;14:463–472. 223. Samuel CE. Antiviral actions of interferons. Clin Microbiol Rev . 2001;14:778–809. 224. Inagaki H, Matsushima Y, Ohshima M, Kitagawa Y. Interferons suppress mitochondrial gene transcription by depleting mitochondrial transcription factor A (mtTFA). J Interferon Cytokine Res. 1997;17:263–269. 225. Shan B, Vazquez E, Lewis JA. Interferon selectively inhibits the expression of mitochondrial genes: a novel pathway for interferon-mediated responses. EMBO J . 1990;9:4307–4314. 226. Le Roy F, Bisbal C, Silhol M, Martinand C, Lebleu B, Salehzada T. The 2-5A/RNase L/RNase inhibitor (RLI) pathway regulates mitochondrial mRNAs stability in interferon-α-treated H9 cells. J Biol Chem. 2001;276:48473–48482. 227. Lewis JA, Huq A, Najarro P. Inhibition of mitochondrial function by interferon. J Biol Chem. 1996;22:13184–13190. 228. Cast´era L, Kalinsky E, Bedossa P, Tertian G, Buffet C. Macrovacuolar steatosis induced by interferon alfa therapy for chronic myelogenous leukemia. Liver. 1999;19:259–260. 229. Okanoue T, Sakamoto S, Itoh Y, et al. Side effects of high-dose interferon therapy for chronic hepatitis C. J Hepatol . 1996;25:283–291. 230. Montenez JP, Van Bambeke F, Piret J, Brasseur R, Tulkens PM, Mingeot-Leclercq MP. Interactions of macrolide antibiotics (erythromycin A, roxithromycin, erythromycylamine [Dirithromycin] and azithromycin) with phospholipids: computer-aided conformation analysis and studies on acellular and cell culture models. Toxicol Appl Pharmacol . 1999;156:129–140. 231. Oleinick NL, Corcoran JW. Two types of binding of erythromycin to ribosomes from antibiotic-sensitive and resistant Bacillus subtilis 168. J Biol Chem. 1969;244:727–735. 232. Anandatheerthavarada HK, Vijayasarathy C, Bhagwat SV, Biswas G, Mullick J, Avadhani NG. Physiological role of the N-terminal processed P4501A1 targeted to mitochondria in erythromycin metabolism and reversal of erythromycin-mediated inhibition of mitochondrial protein synthesis. J Biol Chem. 1999;274:6617–6625. 233. Karbowski M, Kurono C, Wozniak M, et al. Free radical–induced megamitochondria formation and apoptosis. Free Radic Biol Med . 1999;26:396–409. 234. Ress BD, Gross EM. Irreversible sensorineural hearing loss as a result of azithromycin ototoxicity. Ann Otol Rhinol Laryngol . 2000;109:435–437.

202

FEATURES AND MECHANISMS OF DRUG-INDUCED LIVER INJURY

235. Danan G, Descatoire V, Pessayre D. Self-induction by erythromycin of its own transformation into a metabolite forming an inactive complex with reduced cytochrome P-450. J Pharmacol Exp Ther . 1981;218:509–514. 236. Pessayre D, Larrey D, Funck-Brentano C, Benhamou JP. Drug interactions and hepatitis produced by some macrolide antibiotics. J Antimicrob Chemother . 1985;16(suppl A):181–194. 237. Kroon AM, de Vries H. The effects of chloramphenicol on the biogenesis of mitochondria of rat liver in vivo. FEBS Lett. 1969;3:208–210. 238. Yunis AA. Chloramphenicol toxicity: 25 years of research. Am J Med . 1989;87: 44N–48N. 239. De Vriese AS, Coster RV, Smet J, et al. Linezolid-induced inhibition of mitochondrial protein synthesis. Clin Infect Dis. 2006;42:1111–1117. 240. Weiss CF, Glazko AJ, Weston JK. Chloramphenicol in the newborn infant: a physiologic explanation for its toxicity when given in excessive doses. N Engl J Med . 1960;262:787–794. 241. Morgan MY, Reshef R, Shah RR, Oates NS, Smith RL, Sherlock S. Impaired oxidation of debrisoquine in patients with perhexiline liver injury. Gut. 1984;10:1057–1064.

6 CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN ˜ Paulo J. Oliveira and Vilma A. Sardao Center for Neurosciences and Cell Biology, Department of Zoology, University of Coimbra, Coimbra, Portugal

Kendall B. Wallace Department of Biochemistry and Molecular Biology, University of Minnesota Medical School, Duluth, Minnesota

1. Introduction 1.1. Mitochondria: furnace of the heart 1.2. Heart mitochondria in cell calcium homeostasis 1.3. Cardiac oxidative stress and mitochondrial permeability transition 1.4. Mitochondrial dysfunction in cardiac ischemia and reperfusion 2. Induction of cardiovascular mitochondrial toxicity by xenobiotics 2.1. Antineoplastic therapy: the case for doxorubicin 2.2. Toxicity of nucleoside reverse transcriptase inhibitors on cardiac mitochondria 3. Conclusions

203 203 205 206 209 212 212 219 221

1. INTRODUCTION 1.1. Mitochondria: Furnace of the Heart During normal performance of the adult heart, fatty acids (60 to 90%) and lactate or glucose (10 to 40%) are the preferred substrates to fuel cardiac energy Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

203

204

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

metabolism [1]. Cardiac ATP is required for maintenance of ionic gradients (e.g., energy for sarcolemmal and sarcoplasmic reticulum calcium-ATPases) and, most important, for supporting muscle contraction and relaxation. In such an energy-demanding organ, mitochondrial oxidative phosphorylation is the primary source of the requisite ATP. The remaining ATP derives from glycolysis. Myocardial energy storage capacity, either as lipids or glycogen, is limited, so that uptake and oxidation of fatty acids is tightly coupled; transport proteins exist in both the sarcolemma and mitochondrial membranes to assure a steady supply of fatty acids to mitochondrial β-oxidation. Coordinating increases in ATP demand with enhanced mitochondrial ATP supply is essential during episodes of increased cardiac workload, which underscores the notion that mitochondrial structural and functional integrity are essential for survival of the myocardium. In terms of spatial organization, cardiac cells present two distinct mitochondrial populations, one beneath the plasma membrane, the other distributed throughout the contractile myoﬁbrils (Figure 1). It has been suggested that subsarcolemmal mitochondria (SSM) are responsible for generating ATP for membrane ion pumps, while interﬁbrillar mitochondria are responsible for supplying ATP to the contractile apparatus [1–3]. In fact, functional differences

Subsarcolemmal mitochondria

Myofibrils

Sarcoplasmic reticulum

Interfibrillar mitochondria

Nucleus

Figure 1 Basic scheme of the mitochondrial network inside myocytes. Subsarcolemmal mitochondria are located underneath the plasma membrane (or sarcolemma). Their ATP production is used mainly by ATP-dependent pumps (ATPases) in the cell membrane. Interﬁbrillar mitochondria located among the myoﬁbrils supply ATP for the contractile processes of the myocyte. Mitochondria located in the close vicinity of the sarcoplasmic reticulum may module spatial and temporal amplitude of calcium signaling and also supply ATP for calcium pumps located in the sarcoplasmic reticulum membrane. The insert shows one H9c2 myoblast labeled with the mitochondrial-speciﬁc tetramethylrhodamine methyl ester (TMRM). Although not a “true” adult cardiomyocyte, the ﬁgure is illustrative of the complex arrangement of the mitochondrial network.

INTRODUCTION

205

between the two types of mitochondria in terms of calcium accumulation and respiratory activities have been demonstrated [2–5]. The energy-transducing system in the myocardium consists of several components that link ATP production in the mitochondrial matrix to ATP hydrolysis by the membrane pumps and contractile proteins in the cytosol. The energy-transducing network includes the mitochondrial electron transport chain of the inner mitochondrial membrane, the protein transporters spanning both the inner and outer mitochondrial membranes to exchange substrates and products, and also the creatine shuttle, which helps deliver energy in the form of creatine phosphate to the cytoplasm. Creatine kinase, which transfers the high-energy phosphate group from creatine to ADP, is localized in both the mitochondrial and myoﬁbril membranes [6]. It warrants emphasizing that the creatine shuttle is essential to both the storage and transfer of energy within the cardiac myocardium [7–9]. 1.2. Heart Mitochondria in Cell Calcium Homeostasis Heart mitochondria are able to accumulate large amounts of calcium in the matrix without compromising function, the inward driving force being the large transmembrane electrochemical gradient. The protein responsible for this inward calcium current is the mitochondrial calcium uniporter. Although it is documented that the protein transports calcium according to the transmembrane electric potential (negative inside), its low-afﬁnity constant in isolated mitochondria [10,11] suggests that mitochondria do not have an active role in cellular calcium homeostasis (see Chapter 1). Nevertheless, an increase in intramitochondrial calcium concentration does occur in response to numerous stimuli. In fact, energized mitochondria are required for spatial and temporal modulation of calcium signaling across the cell [12–18]. Intramitochondrial calcium spikes track closely with cytosolic spikes in heart cells [19,20]. Also, mitochondrial calcium transients that occur during the contractile cycle are translated into a time-averaged increase in mitochondrial ATP production (see below) that keeps pace with increased cytosolic demand [19,20]. Similarly, a decrease in mitochondrial calcium accumulation during the contractile cycle is associated with a decrease in ATP generation and contributes to energy transitions in the myocardium [20]. The apparent inconsistency between the low afﬁnity of the calcium uniporter in isolated mitochondria and the strict synchronization observed in intact cells can be explained by the close proximity of the mitochondrial network to the sarcoplasmic reticulum [21,22]. Such colocalization allows for the creation of microdomains where calcium concentrations can greatly exceed that in the bulk cytosplasm [23]. In such microdomains, calcium concentrations within the working range of the mitochondrial calcium uniporter are often achieved routinely. Within the matrix of heart mitochondria, calcium stimulates mitochondrial NADH generation by the activation of selected dehydrogenases (pyruvate dehydrogenase, isocitrate dehydrogenase, and 2-oxoglutarate dehydrogenase) through stimulation of calcium-dependent protein kinase–catalyzed protein phosphorylation [24–28]. Calcium also regulates the mitochondrial ATP synthase [29,30].

206

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

Figure 2 Cardiac mitochondrial calcium interactions. In physiological matrix concentrations, calcium activates the Krebs cycle, through stimulation of speciﬁc dehydrogenases, and also stimulate the ATP synthase. The end result is a net increase in ATP synthesis. Calcium enters mitochondria electrophoretically through a speciﬁc calcium uniporter and leaves the mitochondria through a sodium–calcium exchanger. Mitochondrial calcium overload can lead to induction of the mitochondrial permeability transition pore through interaction with protein and membrane calcium-binding sites.

The fact that very little calcium uptake occurs during resting conditions suggests that calcium may be an intramitochondrial signal for increased ATP generation. Several calcium efﬂux pathways exist in mitochondria [31]. A calcium-dependent calcium extrusion mechanism that is abundant in the heart is the Na+ /Ca2+ exchanger. Figure 2 summarizes the pathways involved in cardiac mitochondrial calcium homeostasis and the intramitochondrial roles of calcium. In conclusion, although mitochondria may not play a major role in cellular calcium homeostasis under resting conditions, they may be a critical factor in determining the shape and duration of cytosolic calcium spikes during cell stimulation. Intracellular localization in close proximity to the sarcoplasmic reticulum may be a critical feature for the mitochondrial calcium buffer capacity. Calcium overload is deleterious for mitochondrial and cellular function, through induction of the mitochondrial permeability transition (MPT). 1.3. Cardiac Oxidative Stress and Mitochondrial Permeability Transition A disturbance between pro-oxidants and antioxidant systems in favor of the former is one form of oxidative stress. Oxidative stress can also result when reactive nitrogen species are formed. For example, nitric oxide (· NO) can be formed by the enzymatic oxidation of citrulline by nitric oxide synthases (NOSs), which exist in several locations, including the mitochondrial matrix. Nitric oxide can generate peroxynitrite (· NOOO− ) from the reaction with superoxide anion (Figure 3). Nitric oxide itself is very important in the context of the cardiovascular system, in particular by acting as a vasorelaxing factor and by regulating myocardial contractility (for a extensive review, see [32–34]). Nitric oxide can also regulate

INTRODUCTION

207

Figure 3 Mitochondrial reactions of free radicals and derivatives. The ﬁgure also depicts some of the components of the antioxidant network. Nitric oxide (NO) produced by the mitochondrial nitric oxide synthase can react with superoxide anion to create peroxynitrite. The image also highlights the importance of glutathione in maintaining the redox equilibrium in the matrix. GPx, glutathione peroxidase; GRed, glutathione reductase; MTDH, energy-linked transhydrogenase; MnSOD, mitochondrial manganese superoxide dismutase; GSH, reduced glutathione; GSSG, oxidized glutathione.

mitochondrial respiration, by inhibiting cytochrome c oxidase [35–37], which is important in a context of ischemia where oxygen concentration is limiting. Although it is rather inactive under resting conditions [38], the mitochondrial nitric oxide synthase (NOS) appears to be up-regulated under conditions of tissue hypoxia [39]. Nitric oxide and peroxynitrite also have toxic effects on proteins of the mitochondrial respiratory chain. For example, nitric oxide can directly nitrosylate thiol groups [40] and decrease mitochondrial respiration via a direct effect on complex I [41]. Alternatively, it may have an indirect effect through the generation of peroxynitrite [42,43]. Nitric oxide was also shown to react with ubiquinol [44]. Oxidative modiﬁcation of mitochondrial proteins may lead to inactivation and inhibition of mitochondrial oxidative phosphorylation. Aconitase, a Krebs cycle enzyme, is one of the preferred targets for free radicals because of its high iron content, and depending on the speciﬁc reaction, can produce hydroxyl radical [45]. Mitochondrial calcium overload and oxidative stress are also known to induce MPT (Figure 4), which, depending on the extent of mitochondrial collapse, undermines myocyte bioenergetics and function. The MPT is characterized by a loss of the permeability barriers that normally characterize the inner mitochondrial membrane, with the consequent loss of membrane potential. It is believed to originate

208

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

A

B

Figure 4 (A) MPT pore complex as it is proposed conventionally. Some of the pictured pore components are the adenine nucleotide translocator (ANT), the voltage-dependent anion channel (VDAC), mitochondrial kinases such as creatine kinase (mtCK) and hexokinase (HK), and cyclophilin D (Cycl.D). Also indicated are some known pore inducers (arrows) and inhibitors (dashed lines). Pro-apoptotic proteins as Bax or antiapoptotic proteins (as Bcl-2) were found interacting with the pore complex. A key point for formation and/or opening of the MPT pores is the oxidation of protein thiol groups. The pore appears to be formed on contact points between the outer and inner mitochondrial membranes. (B) Calcium-induced MPT pore opening leads to solute entry in the mitochondrial matrix that is followed osmotically by the entry of water, leading to increased mitochondrial volume, expansion of the inner mitochondrial membrane, and rupture of the outer membrane, which possesses a smaller area. Rupture of the outer membrane can lead to the release of pro-apoptotic proteins such as cytochrome c or the apoptosis-inducing factor (represented by small dots and squares). The panel also shows typical micrographs of isolated heart mitochondria before and after suffering the MPT.

INTRODUCTION

209

from the formation of large proteinaceous pores within the inner mitochondrial membrane that traverse or interdigitate with the outer mitochondrial membrane [46–48]. Induction of the MPT typically requires excess intramitochondrial calcium, and the threshold varies for other inducing factors. It has also become clear that MPT pores are formed when calcium and oxidative stress are present [49–51] or when membrane depolarization occurs [52–55]. A critical factor in MPT pore formation and opening is the oxidation of protein thiol groups and generation of dithiol cross-links [55–57]. The identity of individual components of the mitochondrial permeability transition pore remains controversial (Figure 4). Consensus is that the adenine nucleotide translocator (ANT) is a critical element of the pore [58–60]. Other proteins have also been proposed to be intimately involved, such as cyclophilin D, a matricial protein [55,61], the voltage-dependent anion channel (VDAC) [62–64], several kinases [65–67], and even mitochondrial respiratory complex I [68,69], albeit in a regulatory role. The true nature and composition of the pore complex remains elusive, although in vivo knockout of some proposed pore components, including ANT and cyclophilin D, yields unexpected results [70,71], such as a regulatory, not an essential, role for ANT [72]. Consequences of MPT in isolated heart mitochondria include mitochondrial swelling due to water entry (Figure 4B), membrane depolarization, and equilibration of all transmembrane solute gradients. Some investigators have described an in situ physiologic role for the MPT. The low-conductance state of the MPT pore has been suggested to be both useful in releasing calcium accumulated in the mitochondrial matrix and in keeping the mitochondrial electric potential below a certain threshold, thereby avoiding membrane hyperpolarization [73–75]. The transition to a high-conductance state depends on the saturation of calcium-binding sites and is proposed to be responsible for permanent ATP deﬁcit that leads to cell necrosis [75]. The best known inhibitor of the MPT is the immunosuppressive peptide cyclosporin A, which is believed to bind cyclophilin D, a peptidylprolyl cis-trans isomerase that is believed to interact with the pore complex [76–78]. Antioxidants can also inhibit MPT induction in isolated mitochondria [50,79,80] by decreasing the oxidation of critical thiol groups and/or the oxidation of pyridine nucleotides needed for formation of the pore complex [81]. It is known that cytochrome c and other pro-apoptotic proteins can be released from mitochondria because of outer membrane rupture during the MPT [82–84] (Figure 4, lower panel). Nevertheless, some reports show that induction of the MPT is not required for cytochrome c release from mitochondria [85–87]. The effect of MPT inhibitors in the apoptotic process is not yet fully developed, despite some promising isolated results [88–90]. 1.4. Mitochondrial Dysfunction in Cardiac Ischemia and Reperfusion During cardiac ischemia and reperfusion (IR), the myocyte is challenged by heightened intracellular concentrations of phosphate and calcium, which occur as

210

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

Figure 5 Some relevant biochemical alterations that occur during ischemia and reperfusion in the cardiomyocyte. Among the most important is the interplay between the Na+ /Ca2+ and H+ /Na2+ exchangers, which leads to an increased cytosolic calcium upon reperfusion. Sudden production of oxygen free radicals during the reperfusion phase (most of them originating from the respiratory chain) also causes mitochondrial dysfunction, due primarily to the oxidation of proteins and lipids, as well as the induction of the mitochondrial permeability transition. Mitochondrial DNA can also suffer oxidation upon increased oxidative stress.

a result of inhibition of ATP-dependent ion homeostasis (Figure 5). It is now clear that increased generation of oxygen free radicals occurs during the reperfusion phase [91], most of which originates from the mitochondrial respiratory chain [92]. In fact, reperfusion under hypoxic conditions is much less deleterious than under normoxic conditions [93]. Similarly, targeting antioxidants to mitochondria decreases cardiac IR-induced damage [94]. As described above, the association of oxidative stress with increased mitochondrial calcium and free phosphate are conditions favorable to MPT induction in vitro. Interestingly, it has been demonstrated that MPT pores remains closed during ischemia and open upon reperfusion, mainly because of low pH levels during ischemia (cellular acidosis due to glycolysis-originated lactate production) [55,95]. In fact, MPT pore opening is now well established to occur in situ during reperfusion following ischemia or anoxia [96] (Figure 5). MPT pore opening also causes mitochondrial and cytosolic NAD+ depletion, which are characteristics of myocyte death due to postischemic reperfusion of the heart. In agreement, inhibitors of the MPT pore also inhibit NAD+ efﬂux from mitochondria [97]. Additional evidence for a primary role of MPT during IR is demonstrated in the work of Nadtochiy et al. [98], who showed that mitochondrial proton leak increases signiﬁcantly after IR. The increase was sensitive to cyclosporin-A and carboxyatractyloside, implicating both the MPT and the adenine nucleotide translocator in the process. In the same context, Bosetti et al. [99] reported an increased resting-state respiration rate for mitochondria isolated from IR-subjected rat hearts.

INTRODUCTION

211

The duration and reversibility of MPT pore openings are critical factors for cell recovery after reperfusion. After recuperation of cytosolic calcium levels during the reperfusion phase, certain subpopulations of the MPT pores close, the mitochondria recover their membrane potential, and that particular cell is likely to survive. However, MPT is often irreversible, and in mitochondria where MPT remains open, ATP production ceases, and the cell will eventually die by necrosis or apoptosis, depending on the severity of mitochondrial failure [55]. It is well established that cardiac IR causes myocyte cell death [100]. As described above, calcium overload, a condition known to induce the MPT, also causes cytochrome c release in the heart [101]. Pyruvate, an antioxidant and regulator of a high mitochondrial NAD(P)H/NAD(P) ratio, has been shown to ameliorate cardiac performance after IR [102], which was attributed to MPT inhibition during the reperfusion phase [103]. Although it is suspected that the majority of oxidative stress-associated cell death occurs during reperfusion, there are reports that a substantial release of cytochrome c and activation of cytosolic caspases happen during the ischemic period itself [37]. Despite the strong connections of the MPT and mitochondrial dysfunction induced by IR, questions remain regarding the manifestation of MPT pore opening in vivo, especially because the concentration of adenine nucleotides in the cell are well above those known to inhibit pore opening. One particular and important target of oxygen free radicals during the reperfusion phase is cardiolipin, a phospholipid in the inner mitochondrial membrane that plays an important role in the regulation of cytochrome c release, and thus in mitochondrial-dependent cell death [104,105]. In fact, several investigations implicate cardiolipin directly in the decreased activity of respiratory chain complexes [106,107] and cytochrome c release during reperfusion [108]. Inhibition of the mitochondrial respiratory chain due to cardiolipin alterations occurs mostly during the ischemic phase itself [109]. Besides cardiolipin, increased oxidative stress during reperfusion can oxidize other mitochondrial lipids, causing alterations in membrane ﬂuidity and increased permeability [110]. Alterations of protein activity during ischemia and reperfusion have recently been reviewed [111]. The main question is whether such alterations are due to direct effects of oxidative stress on the protein itself, or altered signaling pathways originating from the ischemic and/or reperfusion process. One of the complexes most affected during IR is complex I [112,113]. Complexes III and IV have also been shown to be affected, but to a lesser extent [107,110,114]. Whatever the cause, the major consequence is that critical mitochondrial proteins are oxidatively modiﬁed [115,116], which can hinder their normal physiological function. Alternatively, protein oxidation may reverse the activity of the ATP synthase, which in a futile attempt to pump protons and generate membrane potential hydrolyzes the last of the severely limited ATP reserves [117]. Fortunately, ATP synthase has an inhibitory subunit that normally shuts down the ATPase hydrolysis and so blocks complete ATP depletion [118]. Mitochondrial DNA is also vulnerable during cardiac IR [119]. Interestingly, the two cardiac mitochondrial subpopulations have different susceptibility to

212

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

IR in the aged heart, with interﬁbrillar mitochondria being more sensitive to IR-induced oxidative damage [120,121]. Diabetes is another condition that can inﬂuence the resistance of the heart to IR. Mortality in the diabetic population is due primarily to cardiovascular disease, which is not dissimilar from the nondiabetic population. However, the diabetic population has a greater probability of suffering cardiovascular problems and heart failure than does the general population [122]. Oliveira et al. [123,124] demonstrated that heart mitochondria isolated from two different experimental models for diabetes, the Goto–Kakizaki and the streptozotocin-injected rat, have opposite susceptibilities to calcium-induced MPT. Heart mitochondria isolated from the Goto–Kakizaki rat were more resistant than the control to calcium-induced MPT [123], whereas those from streptozotocin-treated rats were more susceptible [124]. Nevertheless, it must be stressed that the average blood glucose levels are lower in the Goto–Kakizaki rat than in the streptozotocin-treated rat. The relationship between MPT induction and the severity of diabetes may explain some of the contradictory results concerning the resistance of the diabetic heart to IR damage [125–127].

2. INDUCTION OF CARDIOVASCULAR MITOCHONDRIAL TOXICITY BY XENOBIOTICS We deal next with some well-known cardiac mitochondrial toxicants with clinical relevance. Among the wide variety of drug toxicities involving cardiac mitochondrial dysfunction, the best documented are nucleoside reverse transcriptase inhibitor (NRTI)–induced mitochondrial DNA depletion and doxorubicin-induced redox cycling and cardiac mitochondrial dysfunction [128]. The primary focus here is on doxorubicin, on which the literature is extensive. However, over 50% of the drugs receiving a black box warning from the U.S. Food and Drug Administration for cardiovascular toxicity are now known to have mitochondrial liabilities (see Chapter 26). Not all such adverse cardiovascular effects arise from cytotoxicity, and studies on myocyte response to drugs now known to undermine mitochondrial function, perhaps fostered by this book, will help illuminate this issue. 2.1. Antineoplastic Therapy: The Case for Doxorubicin Doxorubicin (DOX; Adriamycin) is an anthracycline antibiotic originally isolated by the aerobic fermentation of Streptomyces peucetius caesius [129]. DOX is one of the most potent antineoplastic drugs, prescribed alone or in combination with other agents, to treat various types of tumors, including endometrium, ovary, testicle, thyroid, and lung carcinomas and in sarcomas such as neuroblastoma, Ewing’s sarcoma, osteosarcoma, or rhabdomyosarcoma. DOX is also effective in the treatment of hematological cancers, including acute leukemia, multiple myeloma, Hodgkin’s disease, and the diffuse non-Hodgkin lymphomas.

INDUCTION OF CARDIOVASCULAR MITOCHONDRIAL TOXICITY

213

The mechanism proposed for the antineoplastic effect of DOX is based on its ability to interfere with DNA replication and transcription via intercalation between adjacent base pairs of the double helical structure [130]. DOX intercalation in DNA causes stereochemical template disordering, inhibiting enzymes involved in DNA replication (such as topoisomerase II) and transcription (such as RNA polymerases) [131]. Although possessing a superior antineoplastic activity [132], broader clinical use of DOX is restricted by the high incidence of life-threatening cumulative cardiomyopathy. Acute cardiovascular effects occur within a few minutes of DOX administration and include hypotension, tachycardia, and arrhythmias [133,134]. These are all fairly easy to manage, and resolve spontaneously when the treatment is discontinued [134]. On the other hand, chronic DOX cardiotoxicity is rarely reversible and is much more complicated to manage. One particularly challenging aspect of DOX-induced chronic cardiotoxicity is its cumulative nature. In fact, DOX-induced cardiac failure can appear as late as 20 years after the last exposure [136,137]. Doxorubicin-induced cardiomyopathy is characterized by several forms of tachycardia [138] and altered left ventricular function [139]. Also, severe histological changes can occur, including loss of myoﬁbrils, altered sarcoplasmic reticulum, deposition of lipid droplets, vacuolization of the cytoplasm, and mitochondrial swelling [134,140–144]. Several hypotheses have been proposed to explain cardiac DOX toxicity [145], and oxidative stress is among the most widely accepted. DOX was initially shown to be metabolized by cardiac microsomes [146,147]. However, more pathogenically, DOX readily undergoes a univalent reduction of the quinone moiety to form a semiquinone free radical. The resulting unstable intermediate can rapidly transfer the unpaired electron to a suitable electron acceptor, such as molecular oxygen, generating a reactive free radical while regenerating the parent quinone molecule, to complete a vicious redox cycle. DOX can also accept a second electron to form a stable hydroquinone derivative or form covalent adducts with DNA or proteins, which at least in vitro occurs mostly in the absence of suitable electron acceptors [135,148,149]. Although DOX can accept electrons from several electron donors (for a review, see [150]), a particularly important role has been attributed to mitochondrial oxidoreductases. In fact, oxidative stress in the cardiac tissue has been related to metabolic activation at the mitochondrial level. DOX can redox-cycle at mitochondrial complex I, generating ROS in the process (Figure 6) [151–154]. Several authors have documented DOX-induced oxidative damage to mitochondrial membranes [155], proteins [156], and nucleic acids [157], accompanied by a decrease in cellular energy charge [158] following activation by mitochondrial complex I. A growing amount of evidence suggests that mitochondria are principal targets in the development of DOX-induced cardiomyopathy [145,159,160]. In fact, early stages of DOX-induced cardiomyopathy are characterized by changes in both morphology and function of heart mitochondria [161–164], including

214

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN O

OH

O CCH2OH

NAD+ NADP+

OH

O2

OCH3 O

H

OH O

O

CH3 OH NH2 O

•

OH

O CCH2OH OH

OCH3OH

O2•

NADH NADPH

H

OH O CH3 OH NH2

O

Oxidative stress

mtDNA

−SH HS−

+ + + + H H H H

H+

−S −S−

H+

H+

H+

H+

H+

H+

Figure 6 Doxorubicin redox cycle. The molecule can suffer univalent reduction by several electron donors (univalent reduction potential is around −320 mV). When oxygen is not limiting, DOX undergoes redox cycling, producing oxygen free radicals in the process. No net consumption of DOX occurs in the reaction. Several cell oxidoreductases can reduce DOX, including microsomal NADPH reductases and mitochondrial NADH dehydrogenases (such as complex I). DOX redox cycling at complex I generates oxygen free radicals: superoxide anion, hydrogen peroxide (H2 O2 ), and ultimately, in the presence of iron, hydroxyl radical (HO−· ). Resulting oxidative stress can damage mitochondrial proteins, lipids, and DNA. Of particular interest is the increased oxidative damage to proteins that are part of the MPT pore complex. Oxidative stress resulting from DOX redox cycling may directly act on thiol groups of proteins that constitute the MTP pore complex. Thiol cross-linking can cause the formation of active MPT pores and mitochondrial swelling, as shown in the micrographs.

interference with mitochondrial calcium homeostasis at subclinical cumulative doses [144,165]. Chronic DOX administration to Sprague–Dawley rats compromises the respiratory function of cardiac mitochondria, especially during ATP synthesis

INDUCTION OF CARDIOVASCULAR MITOCHONDRIAL TOXICITY

215

(state 3 respiration) [143,166,167]. Interestingly, the decrease in state 3 respiration observed in DOX-treated rats can be reversed by dithiothreitol (DTT), which suggests speciﬁc alterations in the redox status of protein thiol groups [168]. On the other hand, state 4 respiration is altered after in vivo treatment [143,167], which is in agreement with previous observations [162]. However, heart mitochondria from DOX-treated rats maintain the same efﬁciency of ATP synthesis, as indicated by the ADP/O ratio [143]; in other words, heart mitochondria from DOX-treated rats phosphorylate the same amount of ADP per amount of oxygen consumed, although the rate of catalysis is lower. This is in accord with there being no increase in state 4 respiration, which would have been indicative of uncoupling. One possible mechanism of inhibition of state 3 respiration by DOX is inactivation of mitochondrial enzymes, as has been reported by several groups [143,169,170]. NADH–dehydrogenase activity is decreased in cardiac mitochondria from DOX-treated rats [143] and mitochondrial sucinate dehydrogenase activity was impaired in submitochondrial particles treated in vitro with DOX [170]. In contrast, the activity of cytochrome c oxidase is unaffected after in vivo DOX treatment [143], but in vitro studies reported a decreased enzyme activity [171]. As with IR-induced damage (see above), mitochondrial enzyme inactivation can be a consequence of lipid peroxidation or can occur as a direct effect of reactive oxygen species on susceptible thiols and Fe–S complexes. Another mechanism for decreased protein activity is an altered expression, which is likely to occur during in vivo DOX treatment. In fact, altered expression of proteins such as the adenine nucleotide translocator or the Reiske iron–sulfur protein, a ubiquitously expressed electron transport chain component, has been demonstrated [172]. Interestingly, heart mitochondria are far more sensitive than mitochondria from liver or kidney to in vitro DOX-related mitochondrial injury [143,173]. Controversial results [143,174] regarding effects of DOX treatment on oxygen consumption may reﬂect that measurement of respiration is not a sensitive and deﬁnitive indicator of mitochondrial dysfunction. Instead, it is the decrease in mitochondrial calcium-loading capacity and loss of mitochondrial calcium homeostasis that can be considered as an early and more sensitive and quantiﬁable indication of graded mitochondrial dysfunction caused by drug treatment [144,166,167,174]. Several in vitro experiments demonstrate that mitochondrial calcium homeostasis is affected by exposure of isolated heart mitochondria or cells to DOX. Moore et al. [175] were ﬁrst to demonstrate that DOX inhibits the “cardiac mitochondrial calcium pump.” Conﬁrming this, Revis and Marusic [176] observed a lower calcium uptake of heart mitochondria in the presence of DOX or its aglycone metabolite, although the mechanism was still unknown at the time. Chacon and Acosta [177] reported a disruption in mitochondrial calcium homeostasis by DOX which could be involved in the production of ROS and its cardiotoxicity. The authors demonstrated that ruthenium red, an inhibitor of mitochondrial calcium uptake, not only attenuated the enhanced formation of intracellular ROS

216

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

but also increased cell viability upon exposure to DOX [177]. These results are key evidence that calcium is a major player in the generation of mitochondrial oxidative stress and in the killing of cells cultured in vitro with DOX. Chacon et al. [178] provided another important piece of evidence by showing that DOX induced more than a twofold increase in mitochondrial calcium levels before actual changes in cytosolic calcium could be detected. The increase in mitochondrial calcium was simultaneous with a dissipation of mitochondrial membrane potential and a decrease in cellular ATP levels. DOX-induced mitochondrial calcium deregulation was subsequently associated with induction of the MPT. Solem et al. [167] demonstrated that incubating cardiac mitochondria with DOX caused a decreased accumulation and a delayed spontaneous release of calcium from cardiac mitochondria. Cyclosporin A (a speciﬁc inhibitor of the MPT in isolated mitochondria [76]), but not diltiazem (an inhibitor of the mitochondrial sodium–calcium exchanger), completely inhibited DOX-induced calcium release. The results were conﬁrmed in vitro [179]; enhanced MPT in the presence of DOX and calcium was prevented by carvedilol, a β-adrenergic antagonist with antioxidant properties. The increased susceptibility of heart mitochondria from DOX-treated rats to calcium-induced MPT was further explored by Wallace et al. [143,144,166–168,174,180,181]. Heart mitochondria isolated from DOX-treated rats (13 weekly injections, 2 mg/kg per week subcutaneously) exhibited a lower respiratory control ratio than that in saline-treated rats, and an enhanced cyclosporin A–sensitive calcium-induced calcium release [167]. Interestingly, mitochondrial calcium deregulation was still persistent 4 to 7 days after the last drug treatment, reminiscent of the cumulative cardiotoxicity associated with DOX therapy. Solem et al. [166] demonstrated that cardiomyocytes isolated from rats treated for 6 weeks with DOX were more sensitive to calcium-induced cell killing. Key to this experiment was the protection afforded by cyclosporin A and ruthenium red, an inhibitor of calcium uptake via the uniporter. Taken in toto, the data indicate that calcium-dependent MPT is enhanced in heart mitochondria from DOX-treated rats and is thus an important factor in DOX-induced cardiotoxicity (Figure 6). Al-Nasser [182] demonstrated that cyclosporin A and FK506, when administered simultaneously with a single dose of DOX to rats, prevented the increased calcium-dependent MPT in the DOX-treated group. The author concluded that both cyclosporin A and FK506 prevent DOX-induced mitochondrial dysfunction by preventing induction of the calcium-dependent MPT. Although the idea is attractive, both FK506 and cyclosporin A are also calcineurin inhibitors, which can confound the interpretation [183]. The manipulation of intracellular calcium concentration reveals differences in the response of myocytes from saline- and DOX-treated animals to cytosolic calcium overload. Zhou et al. [184] demonstrated that cardiomyocytes isolated from rats treated with six weekly injections of DOX show an enhanced susceptibility to A23184, ouabain, and caffeine (compounds that modify cytosolic calcium concentrations) induced cell injury. The authors concluded that by interfering

INDUCTION OF CARDIOVASCULAR MITOCHONDRIAL TOXICITY

217

with mitochondrial calcium regulation, chronic treatment with DOX renders myocytes more susceptible to cytosolic calcium overload. Again, interfering with mitochondrial calcium handling appears as one possible relevant explanation for DOX-induced cardiotoxicity. Another confounding factor is the persistence of mitochondrial changes after DOX treatment. In one particular study, Sprague–Dawley rats were treated with DOX for 4 to 8 weeks [144]. Cardiac mitochondria isolated from rats after 4 weeks of treatment with DOX had a lower calcium loading capacity compared to mitochondria isolated from saline-treated rats. The decrease in calcium loading capacity was more pronounced with successive doses up to 8 weeks of treatment. The key result was that changes in mitochondrial calcium loading capacity persisted for 5 weeks after the last DOX injection [144]. The observations indicate that DOX causes persistent alterations to heart mitochondrial function, calcium regulation speciﬁcally. Other alterations observed in rats treated chronically with DOX include a decrease in the activity and protein concentration of the adenine nucleotide translocator [144,180] and enhanced susceptibility to the MPT induced by thiol-oxidizing agents [168]. Such alterations occur in conjunction with increased oxidative stress in isolated mitochondria [168] and myocytes [185] from DOX-treated rats, again supporting a possible role of free radicals and oxidative stress in the pathogenesis of DOX-induced mitochondrial toxcity. A single DOX injection (acute DOX toxicity) is enough to trigger differential calcium handling by heart mitochondria from treated animals. By using an oxygen electrode, Ascenc¸a˜ o et al. [186] demonstrated that heart mitochondria from DOX-treated rats did not recover from the stimulation of respiration caused by the addition of calcium. A single acute DOX administration increased the extent of protein oxidation, some of which were integral components of the mitochondrial respiratory chain [187]. Work discussed thus far allows us to conclude that (1) in vivo treatment with DOX increases the susceptibility of cardiac mitochondria to calcium-induced MPT, (2) such alterations can persist several weeks after the last DOX administration, and (3) imbalanced mitochondrial calcium control can result in myocyte death. Nevertheless, a cautionary note is warranted. To date, there is no deﬁnitive evidence that induction of the MPT is a major cause of DOX-induced cardiomyopathy. The increased MPT can be a cause or a consequence of the increased and persistent oxidative stress observed in myocytes of DOX-treated animals [185]. Also, the altered MPT can be the result of altered expression of proteins that form the MPT pore complex, such as the adenine nucleotide translocator (ANT) [144,180] (Figure 7). Interestingly, a causal relation between increased MPT induction and the inhibition of state 3 respiration in DOX-treated rats has been proposed. Oliveira et al. [168] demonstrated that the inhibition of mitochondrial respiration could partially be prevented by adding cyclosporin A to a suspension of heart mitochondria isolated from DOX-treated rats. The authors’ conclusion was that the binding of cyclosporin A to cyclophylin D causes the desegregation of preformed pores in heart mitochondria from DOX-treated rats, thus allowing a

218

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

Doxorubicin

Redox cycling

Interaction with Nuclear and mt DNA

Altered expression/activity of mitochondrial proteins -ANT -Complex I -...

Mitochondrial Oxidative stress

Mitochondrial Effects

Inhibition of Mitochondrial Respiration

Increased Mitochondrial Permeability Transition

Decreased ATP production Cytochrome c release

CARDIOTOXICITY

Cell death

Figure 7 Simpliﬁed scheme of probable mechanisms by which DOX interferes with cardiac mitochondrial bioenergetics, causing cardiotoxicity as a consequence. The scheme should be taken carefully, as the DOX effect on cardiac cells is certainly multifactorial. Interaction or oxidative damage of DOX to nuclear or mitochondrial DNA may end up affecting the expression (and activity) of several important mitochondrial proteins, such as complex I subunits or the adenine nucleotide translocator (among others). Consequences can include increased mitochondrial oxidative stress, inhibition of respiration, and increased mitochondrial permeability transition induction. Oxidative stress can also be caused directly by DOX redox cycling. Oxidative damage to mitochondrial proteins (including the oxidation of protein thiol groups) can promote alterations on the activities of several proteins (either due to direct oxidative damage to proteins or through increased lipid peroxidation). Alterations in protein function can also inhibit mitochondrial oxygen consumption and decrease calcium loading capacity through increased MPT pore openings. Whatever the precise mechanisms are, the end result is decreased ATP production, cytochrome c release (which will feedback to contribute to inhibit respiration and increase oxidative stress), and ultimately, cell death, which will end up by causing myocytes loss and cardiotoxicity.

crucial component, most likely the ANT (as described in [188]), to participate in the oxidative phosphorylation process. Figure 6 demonstrates the possible targets and mechanism of DOX-induced cardiac mitochondrial dysfunction. In view of the implication of oxygen free radicals in the mechanism of DOX-induced damage to the myocardium, one of the approaches to minimize DOX cardiotoxicity has been through the use of free-radical scavengers and other antioxidants [159,189–191]. A good example is carvedilol [143,174]. Carvedilol competitively blocks β1 -, β2 -, and α1 -adrenergic receptors while displaying vasodilating properties. A distinctive characteristic of carvedilol is its potent antioxidant properties, which are not shared by other β-adrenergic receptor antagonists [192,193]. It has been shown [143,174] that when coadministered with DOX, carvedilol prevented the inhibition of the state 3

INDUCTION OF CARDIOVASCULAR MITOCHONDRIAL TOXICITY

219

respiration rate and restored the RCR of cardiac mitochondria to control values [143]. Coadministration of carvedilol to DOX-treated rats decreased the extent of cellular vacuolization in cardiac myocytes and also prevented the inhibitory effect of DOX on the cardiac mitochondrial calcium loading capacity [143,174]. Atenolol was not able to mimic the effects of carvedilol, despite sharing similar β-adrenergic receptor antagonism, which again suggests that the antioxidant properties of carvedilol are important in the prevention of DOX cardiotoxicity [23]. Highlighting the role of oxidative stress in the pathogenesis of DOX cardiotoxicity, Xiong et al. [194] demonstrated that cardiac glutathione peroxidase overexpression prevents most mitochondrial alterations observed after DOX treatment. An emerging hypothesis is that the development of cardiomyopathy induced by DOX treatment involves the apoptosis of cardiomyocytes [195]. Childs et al. [196] postulated that DOX causes cytochrome c release from mitochondria to the cytoplasm, which is accompanied by caspase 3 activation and DNA fragmentation in the hearts of treated rats. The same work also measured an increase in the respiratory ADP/O ratio and the Bcl-2/Bax ratio, which can be an adaptive response to DOX injury. Furthermore, Childs et al. [196] measured an increase in the activity of the cytosolic copper–zinc superoxide dismutase, supporting the notion that DOX increases the production of ROS. Different pathways have been proposed to explain the ﬁnal outcome, the majority involving the production of ROS after DOX activation by mitochondria [197,198]. According to some authors, increased ROS generation induced by DOX exposure can be the trigger for the apoptotic pathway, as apoptosis is partly prevented by antioxidants [199–201]. In conclusion, doxorubicin is an extremely important anthracycline used in clinical practice and is a well-studied prototype for a mitochondrial poison. In fact, evidence shows that DOX can cause mitochondrial dysfunction and loss of myocytes through direct mechanisms that involve oxidative stress in mitochondria, and indirectly by affecting the expression of a variety of mitochondrially relevant genes affecting cellular bioenergetics (Figure 7). It remains to be seen whether the late-onset cardiotoxicity, sometimes occurring years after exposure, could result from oxidative damage to mtDNA during treatment. Over the fullness of time, such damage could gradually erode mitochondrial capacity until a pathogenic bioenergetic threshold is reached and drug-induced cardiotoxicity ﬁnally emerges. 2.2. Toxicity of Nucleoside Reverse Transcriptase Inhibitors on Cardiac Mitochondria Nucleoside reverse transcriptase inhibitor (NRTI)-induced cardiomyopathy has also been suggested to involve important mitochondrial targets (see Chapters 2, 9, and 21). The most broadly professed mechanism of NRTI-induced mitochondrial toxicity involves inhibition of DNA polymerase gamma and interference with mitochondrial DNA replication [202,203]. There is evidence that NRTIs are

220

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

incorporated into replicating DNA, possibly because of the strong similarities with substrates for mitochondrial DNA polymerase gamma [204]. This is then proposed to lead to truncation of the mtDNA and inhibition of mtDNA replication. The decrease in mitochondrial DNA content has been suggested to account for some aspects of metabolic failure observed in patients undergoing therapy. Nevertheless, the evidence is compelling that NRTIs undermine mitochondrial bioenergetics, both directly and indirectly, by mechanisms independent of DNA polymerase γ. Lund and Wallace [205] demonstrated that several NRTIs and their corresponding nucleotides directly inhibit mitochondrial respiration, membrane potential development, and calcium accumulation by isolated heart mitochondria. AZT was thoroughly characterized as a mitochondrial poison whose effects occur independent of mitochondrial DNA polymerase γ. In fact, AZT appears to have different effects on isolated heart mitochondria, including competitive inhibition of thymidine phosphorylation [206,207], induction of superoxide anion formation

Figure 8 Interaction of NRTIs with cardiac mitochondria. Popular thoughts imply the interaction of NRTIs with mitochondrial DNA polymerase γ to explain long-term decrease in mtDNA copy number. Ultimately, a decrease in mtDNA copy number will disturb transcription and replication, with severe consequences for mitochondrial oxidative phosphorylation or mitochondrial replication. On the other hand, some particular aspects of NRTI-induced toxicity cannot be explained solely by inhibition of mitochondrial DNA polymerase γ. In fact, direct inhibition of mitochondrial respiration has been demonstrated, including increased generation of oxygen free radicals. Both immediate (inhibition of the oxygen respiratory chain) and long-term (based on interference with mitochondrial genetics) effects contribute to mitochondrial dysfunction and cardiotoxicity.

REFERENCES

221

[208,209], or inhibition of the adenine nucleotide translocator [210]. In vivo AZT treatment also induced an increase in oxidative stress in the heart mitochondria of treated rats, including oxidation of mitochondrial DNA [208]. During in vivo studies, AZT treatment for up to 49 days disrupts cardiac mitochondrial ultrastructure and inhibits the expression of mitochondrial cytochrome b mRNA in a dose- and time-dependent fashion [211]. Although there remains much to be accomplished in characterizing the effects of the individual NRTIs on mitochondrial function, including distinguishing direct effects from those that are dependent on effects at the mitochondrial genome level, there is little doubt that the ultimate source for cardiac cell injury is based on the fact that mitochondria are a critical target in the mechanism of drug-induced bioenergetic failure (Figure 8).

3. CONCLUSIONS Mitochondria are the targets of a number of xenobiotic-induced organopathies. Although a factor in most tissues, tissues with a high metabolic demand that are most reliant on aerobic metabolism and oxidative phosphorylation are exquisitely sensitive to drug-induced mitochondrial dysfunction and organ failure. The highly demanding and aerobically poised cardiac tissue is perhaps an extreme example of an organ most sensitive to agents that interfere with mitochondrial oxidative phosphorylation. In this chapter we describe two distinct mechanisms by which drugs may inhibit mitochondrial function: direct interference with mitochondrial respiration (inhibition of respiratory complex activity or redox cycling) and inhibition of mitochondrial gene replication and expression. Regardless of the initial event, the ultimate result is fundamentally the same: inhibition of ATP synthesis and bioenergetic failure of the tissue. It is only through researching the underlying cause that one can develop therapeutic interventions designed to prevent or circumvent such drug-induced mitochondrial cardiomyopathies.

REFERENCES 1. Eaton S, Pourfarzam M, Bartlett K. The effect of respiratory chain impairment of beta-oxidation in rat heart mitochondria. Biochem J. 1996;319(pt 2):633–640. 2. McMillin-Wood J, Wolkowicz PE, Chu A, Tate CA, Goldstein MA, Entman ML. Calcium uptake by two preparations of mitochondria from heart. Biochim Biophys Acta. 1980;591(2):251–265. 3. Palmer JW, Tandler B, Hoppel CL. Heterogeneous response of subsarcolemmal heart mitochondria to calcium. Am J Physiol. 1986;250(5 pt 2):H741–H748. 4. Palmer JW, Tandler B, Hoppel CL. Biochemical properties of subsarcolemmal and interﬁbrillar mitochondria isolated from rat cardiac muscle. J Biol Chem. 1977;252(23):8731–8739.

222

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

5. Palmer JW, Tandler B, Hoppel CL. Biochemical differences between subsarcolemmal and interﬁbrillar mitochondria from rat cardiac muscle: effects of procedural manipulations. Arch Biochem Biophys. 1985;236(2):691–702. 6. Menin L, Panchichkina M, Keriel C, et al. Macrocompartmentation of total creatine in cardiomyocytes revisited. Mol Cell Biochem. 2001;220(1–2):149–159. 7. Joubert F, Hoerter JA, Mazet JL. Discrimination of cardiac subcellular creatine kinase ﬂuxes by NMR spectroscopy: a new method of analysis. Biophys J. 2001; 81(6):2995–3004. 8. Joubert F, Mazet JL, Mateo P, Hoerter JA. Identiﬁcation of subcellular energy ﬂuxes by P NMR spectroscopy in the perfused heart: contractility induced modiﬁcations of energy transfer pathways. Mol Biol Rep. 2002;29(1–2):171–176. 9. Joubert F, Hoerter JA, Mazet JL. Modeling the energy transfer pathways: creatine kinase activities and heterogeneous distribution of ADP in the perfused heart. Mol Biol Rep. 2002;29(1–2):177–182. 10. Gunter TE, Buntinas L, Sparagna G, Eliseev R, Gunter K. Mitochondrial calcium transport: mechanisms and functions. Cell Calcium. 2000;28(5–6):285–296. 11. Gunter TE, Gunter KK. Uptake of calcium by mitochondria: transport and possible function. IUBMB Life. 2001;52(3–5):197–204. 12. Gilabert JA, Bakowski D, Parekh AB. Energized mitochondria increase the dynamic range over which inositol 1,4,5-trisphosphate activates store-operated calcium inﬂux. EMBO J. 2001;20(11):2672–2679. 13. Gilabert JA, Parekh AB. Respiring mitochondria determine the pattern of activation and inactivation of the store-operated Ca(2+) current I(CRAC). EMBO J. 2000;19(23):6401–6407. 14. Hajnoczky G, Csordas G, Madesh M, Pacher P. The machinery of local Ca2+ signalling between sarco-endoplasmic reticulum and mitochondria. J Physiol. 2000; 529(pt 1):69–81. 15. Pozzan T, Magalhaes P, Rizzuto R. The comeback of mitochondria to calcium signalling. Cell Calcium. 2000;28(5–6):279–283. 16. Pozzan T, Rizzuto R. High tide of calcium in mitochondria. Nat Cell Biol. 2000; 2(2):E25–E27. 17. Pozzan T, Rizzuto R. The renaissance of mitochondrial calcium transport. Eur J Biochem. 2000;267(17):5269–5273. 18. Makowska A, Zablocki K, Duszynski J. The role of mitochondria in the regulation of calcium inﬂux into Jurkat cells. Eur J Biochem. 2000;267(3):877–884. 19. Huser J, Blatter LA, Sheu SS. Mitochondrial calcium in heart cells: beat-to-beat oscillations or slow integration of cytosolic transients? J Bioenerg Biomembr. 2000; 32(1):27–33. 20. Maack C, Cortassa S, Aon MA, Ganesan AN, Liu T, O’Rourke B. Elevated cytosolic Na+ decreases mitochondrial Ca2+ uptake during excitation-contraction coupling and impairs energetic adaptation in cardiac myocytes. Circ Res. 2006;99(2):172–182. 21. Mannella CA. Introduction: our changing views of mitochondria. J Bioenerg Biomembr. 2000;32(1):1–4. 22. Wang HJ, Guay G, Pogan L, Sauve R, Nabi IR. Calcium regulates the association between mitochondria and a smooth subdomain of the endoplasmic reticulum. J Cell Biol. 2000;150(6):1489–1498.

REFERENCES

223

23. Arnaudeau S, Kelley WL, Walsh JV, Jr., Demaurex N. Mitochondria recycle Ca(2+) to the endoplasmic reticulum and prevent the depletion of neighboring endoplasmic reticulum regions. J Biol Chem. 2001;276(31):29430–29439. 24. McCormack JG, Denton RM. The effects of calcium ions and adenine nucleotides on the activity of pig heart 2-oxoglutarate dehydrogenase complex. Biochem J. 1979; 180(3):533–544. 25. McCormack JG, Denton RM. Hormonal control of intramitochondrial Ca2+ -sensitive enzymes in heart, liver and adipose tissue. Biochem Soc Trans. 1985;13(4):664–667. 26. McCormack JG, Denton RM. The role of Ca2+ in the regulation of intramitochondrial energy production in heart. Biomed Biochim Acta. 1987;46(8–9):S487–S492. 27. McCormack JG, Denton RM. Role of Ca2+ ions in the regulation of intramitochondrial metabolism in rat heart: evidence from studies with isolated mitochondria that adrenaline activates the pyruvate dehydrogenase and 2-oxoglutarate dehydrogenase complexes by increasing the intramitochondrial concentration of Ca2+ . Biochem J. 1984;218(1):235–247. 28. Hopper RK, Carroll S, Aponte AM, et al. Mitochondrial matrix phosphoproteome: effect of extra mitochondrial calcium. Biochemistry. 2006;45(8):2524–2536. 29. Harris DA. Regulation of the mitochondrial ATP synthase in rat heart. Biochem Soc Trans. 1993;21(pt 3)(3):778–781. 30. Harris DA, Das AM. Control of mitochondrial ATP synthesis in the heart. Biochem J. 1991;280(pt 3):561–573. 31. Gunter TE, Pfeiffer DR. Mechanisms by which mitochondria transport calcium. Am J Physiol. 1990;258(5 pt 1):C755–C786. 32. Massion PB, Pelat M, Belge C, Balligand JL. Regulation of the mammalian heart function by nitric oxide. Comp Biochem Physiol A Mol Integr Physiol. 2005;142(2): 144–150. 33. Schulz R, Rassaf T, Massion PB, Kelm M, Balligand JL. Recent advances in the understanding of the role of nitric oxide in cardiovascular homeostasis. Pharmacol Ther. 2005;108(3):225–256. 34. Belge C, Massion PB, Pelat M, Balligand JL. Nitric oxide and the heart: update on new paradigms. Ann N Y Acad Sci. 2005;1047:173–182. 35. Borutaite V, Budriunaite A, Brown GC. Reversal of nitric oxide-, peroxynitrite- and S -nitrosothiol-induced inhibition of mitochondrial respiration or complex I activity by light and thiols. Biochim Biophys Acta. 2000;1459(2–3):405–412. 36. Brown GC. Regulation of mitochondrial respiration by nitric oxide inhibition of cytochrome c oxidase. Biochim Biophys Acta. 2001;1504(1):46–57. 37. Brown GC, Borutaite V. Nitric oxide, mitochondria, and cell death. IUBMB Life. 2001;52(3–5):189–195. 38. French S, Giulivi C, Balaban RS. Nitric oxide synthase in porcine heart mitochondria: evidence for low physiological activity. Am J Physiol Heart Circ Physiol. 2001;280(6):H2863–H2867. 39. Lacza Z, Puskar M, Figueroa JP, Zhang J, Rajapakse N, Busija DW. Mitochondrial nitric oxide synthase is constitutively active and is functionally upregulated in hypoxia. Free Radic Biol Med. 2001;31(12):1609–1615. 40. Stamler JS, Simon DI, Osborne JA, et al. S-Nitrosylation of proteins with nitric oxide: synthesis and characterization of biologically active compounds. Proc Natl Acad Sci U S A. 1992;89(1):444–448.

224

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

41. Clementi E, Brown GC, Feelisch M, Moncada S. Persistent inhibition of cell respiration by nitric oxide: crucial role of S-nitrosylation of mitochondrial complex I and protective action of glutathione. Proc Natl Acad Sci U S A. 1998;95(13):7631–7636. 42. Boveris A, Costa LE, Poderoso JJ, Carreras MC, Cadenas E. Regulation of mitochondrial respiration by oxygen and nitric oxide. Ann N Y Acad Sci. 2000;899: 121–135. 43. Brown GC, Borutaite V. Nitric oxide, cytochrome c and mitochondria. Biochem Soc Symp. 1999;66:17–25. 44. Poderoso JJ, Lisdero C, Schopfer F, et al. The regulation of mitochondrial oxygen uptake by redox reactions involving nitric oxide and ubiquinol. J Biol Chem. 1999;274(53):37709–37716. 45. Vasquez-Vivar J, Kalyanaraman B. Generation of superoxide from nitric oxide synthase. FEBS Lett. 2000;481(3):305–306. 46. Zoratti M, Szabo I. The mitochondrial permeability transition. Biochim Biophys Acta. 1995;1241(2):139–176. 47. Vercesi AE, Kowaltowski AJ, Grijalba MT, Meinicke AR, Castilho RF. The role of reactive oxygen species in mitochondrial permeability transition. Biosci Rep. 1997;17(1):43–52. 48. Bernardi P. Mitochondrial transport of cations: channels, exchangers, and permeability transition. Physiol Rev. 1999;79(4):1127–1155. 49. Castilho RF, Kowaltowski AJ, Meinicke AR, Vercesi AE. Oxidative damage of mitochondria induced by Fe(II)citrate or t-butyl hydroperoxide in the presence of Ca2+ : effect of coenzyme Q redox state. Free Radic Biol Med. 1995;18(1):55–59. 50. Kowaltowski AJ, Castilho RF, Grijalba MT, Bechara EJ, Vercesi AE. Effect of inorganic phosphate concentration on the nature of inner mitochondrial membrane alterations mediated by Ca2+ ions: a proposed model for phosphate-stimulated lipid peroxidation. J Biol Chem. 1996;271(6):2929–2934. 51. Kowaltowski AJ, Castilho RF, Vercesi AE. Opening of the mitochondrial permeability transition pore by uncoupling or inorganic phosphate in the presence of Ca2+ is dependent on mitochondrial-generated reactive oxygen species. FEBS Lett. 1996;378(2):150–152. 52. Bernardi P. Modulation of the mitochondrial cyclosporin A–sensitive permeability transition pore by the proton electrochemical gradient: evidence that the pore can be opened by membrane depolarization. J Biol Chem. 1992;267(13):8834–8839. 53. Nicolli A, Petronilli V, Bernardi P. Modulation of the mitochondrial cyclosporin A–sensitive permeability transition pore by matrix pH: evidence that the pore open–closed probability is regulated by reversible histidine protonation. Biochemistry. 1993;32(16):4461–4465. 54. Petronilli V, Cola C, Massari S, Colonna R, Bernardi P. Physiological effectors modify voltage sensing by the cyclosporin A–sensitive permeability transition pore of mitochondria. J Biol Chem. 1993;268(29):21939–21945. 55. Das M, Parker JE, Halestrap AP. Matrix volume measurements challenge the existence of diazoxide/glibencamide-sensitive KATP channels in rat mitochondria. J Physiol. 2003;547(pt 3):893–902. 56. Costantini P, Chernyak BV, Petronilli V, Bernardi P. Modulation of the mitochondrial permeability transition pore by pyridine nucleotides and dithiol oxidation at two separate sites. J Biol Chem. 1996;271(12):6746–6751.

REFERENCES

225

57. Costantini P, Colonna R, Bernardi P. Induction of the mitochondrial permeability transition by N -ethylmaleimide depends on secondary oxidation of critical thiol groups: potentiation by copper-ortho-phenanthroline without dimerization of the adenine nucleotide translocase. Biochim Biophys Acta. 1998;1365(3):385–392. 58. de Macedo DV, Nepomuceno ME, Pereira-da-Silva L. Involvement of the ADP/ATP carrier in permeabilization processes of the inner mitochondrial membrane. Eur J Biochem. 1993;215(3):595–600. 59. Brustovetsky N, Klingenberg M. Mitochondrial ADP/ATP carrier can be reversibly converted into a large channel by Ca2+ . Biochemistry. 1996;35(26):8483–8488. 60. Ruck A, Dolder M, Wallimann T, Brdiczka D. Reconstituted adenine nucleotide translocase forms a channel for small molecules comparable to the mitochondrial permeability transition pore. FEBS Lett. 1998;426(1):97–101. 61. Connern CP, Halestrap AP. Recruitment of mitochondrial cyclophilin to the mitochondrial inner membrane under conditions of oxidative stress that enhance the opening of a calcium-sensitive non-speciﬁc channel. Biochem J. 1994;302(pt 2):321–324. 62. Szabo I, De Pinto V, Zoratti M. The mitochondrial permeability transition pore may comprise VDAC molecules: II. The electrophysiological properties of VDAC are compatible with those of the mitochondrial megachannel. FEBS Lett. 1993; 330(2):206–210. 63. Szabo I, Zoratti M. The mitochondrial permeability transition pore may comprise VDAC molecules: I. Binary structure and voltage dependence of the pore. FEBS Lett. 1993;330(2):201–205. 64. Zheng Y, Shi Y, Tian C, Jiang C, et al. Essential role of the voltage-dependent anion channel (VDAC) in mitochondrial permeability transition pore opening and cytochrome c release induced by arsenic trioxide. Oncogene. 2004;23(6): 1239–1247. 65. Beutner G, Ruck A, Riede B, Brdiczka D. Complexes between hexokinase, mitochondrial porin and adenylate translocator in brain: regulation of hexokinase, oxidative phosphorylation and permeability transition pore. Biochem Soc Trans. 1997;25(1):151–157. 66. Beutner G, Ruck A, Riede B, Brdiczka D. Complexes between porin, hexokinase, mitochondrial creatine kinase and adenylate translocator display properties of the permeability transition pore: implication for regulation of permeability transition by the kinases. Biochim Biophys Acta. 1998;1368(1):7–18. 67. Beutner G, Ruck A, Riede B, Welte W, Brdiczka D. Complexes between kinases, mitochondrial porin and adenylate translocator in rat brain resemble the permeability transition pore. FEBS Lett. 1996;396(2–3):189–195. 68. Fontaine E, Bernardi P. Progress on the mitochondrial permeability transition pore: regulation by complex I and ubiquinone analogs. J Bioenerg Biomembr. 1999;31(4): 335–345. 69. Fontaine E, Eriksson O, Ichas F, Bernardi P. Regulation of the permeability transition pore in skeletal muscle mitochondria: modulation by electron ﬂow through the respiratory chain complex i. J Biol Chem. 1998;273(20):12662–12668. 70. Blachly-Dyson E, Forte M. VDAC channels. IUBMB Life. 2001;52(3–5):113–118. 71. Krauskopf A, Eriksson O, Craigen WJ, Forte MA, Bernardi P. Properties of the permeability transition in VDAC1(−/−) mitochondria. Biochim Biophys Acta. 2006; 1757(5–6):590–595.

226

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

72. Kokoszka JE, Waymire KG, Levy SE, et al. The ADP/ATP translocator is not essential for the mitochondrial permeability transition pore. Nature. 2004;427(6973): 461–465. 73. Ichas F, Jouaville LS, Mazat JP. Mitochondria are excitable organelles capable of generating and conveying electrical and calcium signals. Cell. 1997;89(7): 1145–1153. 74. Ichas F, Jouaville LS, Sidash SS, Mazat JP, Holmuhamedov EL. Mitochondrial calcium spiking: a transduction mechanism based on calcium-induced permeability transition involved in cell calcium signalling. FEBS Lett. 1994;348(2):211–215. 75. Ichas F, Mazat JP. From calcium signaling to cell death: two conformations for the mitochondrial permeability transition pore: switching from low- to high-conductance state. Biochim Biophys Acta. 1998;1366(1–2):33–50. 76. Broekemeier KM, Dempsey ME, Pfeiffer DR. Cyclosporin A is a potent inhibitor of the inner membrane permeability transition in liver mitochondria. J Biol Chem. 1989;264(14):7826–7830. 77. Szabo I, Zoratti M. The giant channel of the inner mitochondrial membrane is inhibited by cyclosporin A. J Biol Chem. 1991;266(6):3376–3379. 78. Nicolli A, Basso E, Petronilli V, Wenger RM, Bernardi P. Interactions of cyclophilin with the mitochondrial inner membrane and regulation of the permeability transition pore, and cyclosporin A–sensitive channel. J Biol Chem. 1996;271(4):2185–2192. 79. Kowaltowski AJ, Castilho RF, Vercesi AE. Ca(2+)-induced mitochondrial membrane permeabilization: role of coenzyme Q redox state. Am J Physiol. 1995; 269(1 pt 1):C141–C147. 80. Kowaltowski AJ, Netto LE, Vercesi AE. The thiol-speciﬁc antioxidant enzyme prevents mitochondrial permeability transition: evidence for the participation of reactive oxygen species in this mechanism. J Biol Chem. 1998;273(21):12766–12769. 81. Petronilli V, Costantini P, Scorrano L, Colonna R, Passamonti S, Bernardi P. The voltage sensor of the mitochondrial permeability transition pore is tuned by the oxidation–reduction state of vicinal thiols: increase of the gating potential by oxidants and its reversal by reducing agents. J Biol Chem. 1994;269(24):16638–16642. 82. Yang JC, Cortopassi GA. Induction of the mitochondrial permeability transition causes release of the apoptogenic factor cytochrome c. Free Radic Biol Med. 1998; 24(4):624–631. 83. Hirsch T, Susin SA, Marzo I, Marchetti P, Zamzami N, Kroemer G. Mitochondrial permeability transition in apoptosis and necrosis. Cell Biol Toxicol. 1998; 14(2):141–145. 84. Feldmann G, Haouzi D, Moreau A, et al. Opening of the mitochondrial permeability transition pore causes matrix expansion and outer membrane rupture in Fas-mediated hepatic apoptosis in mice. Hepatology. 2000;31(3):674–683. 85. Doran E, Halestrap AP. Cytochrome c release from isolated rat liver mitochondria can occur independently of outer-membrane rupture: possible role of contact sites. Biochem J. 2000;348(pt 2):343–350. 86. Andreyev AY, Fahy B, Fiskum G. Cytochrome c release from brain mitochondria is independent of the mitochondrial permeability transition. FEBS Lett. 1998; 439(3):373–376.

REFERENCES

227

87. Eskes R, Antonsson B, Osen-Sand A, et al. Bax-induced cytochrome c release from mitochondria is independent of the permeability transition pore but highly dependent on Mg2+ ions. J Cell Biol. 1998;143(1):217–224. 88. Zamzami N, Marchetti P, Castedo M, et al. Inhibitors of permeability transition interfere with the disruption of the mitochondrial transmembrane potential during apoptosis. FEBS Lett. 1996;384(1):53–57. 89. Teplova V, Evtodienko Y, Odinokova I, Kruglov A, Kudrjavtsev A. Suppression of mitochondrial permeability transition pore and induction of lymphoma P388 cell death by cyclosporin A. IUBMB Life. 2000;50(1):75–80. 90. Fauvel H, Marchetti P, Obert G, Joulain O, Chopin C, Formstecher P, Neviere R. Protective effects of cyclosporin A from endotoxin-induced myocardial dysfunction and apoptosis in rats. Am J Respir Crit Care Med. 2002;165(4):449–455. 91. Makazan Z, Saini HK, Dhalla NS. Role of oxidative stress in alterations of mitochondrial function in the ischemic reperfused hearts. Am J Physiol Heart Circ Physiol. 2006. 92. Ambrosio G, Zweier JL, Duilio C, et al. Evidence that mitochondrial respiration is a source of potentially toxic oxygen free radicals in intact rabbit hearts subjected to ischemia and reﬂow. J Biol Chem. 1993;268(25):18532–18541. 93. Petrosillo G, Di Venosa N, Ruggiero FM, et al. Mitochondrial dysfunction associated with cardiac ischemia/reperfusion can be attenuated by oxygen tension control: role of oxygen-free radicals and cardiolipin. Biochim Biophys Acta. 2005; 1710(2–3):78–86. 94. Adlam VJ, Harrison JC, Porteous CM, et al. Targeting an antioxidant to mitochondria decreases cardiac ischemia-reperfusion injury. FASEB J. 2005;19(9):1088–1095. 95. Grifﬁths EJ, Halestrap AP. Mitochondrial non-speciﬁc pores remain closed during cardiac ischaemia, but open upon reperfusion. Biochem J. 1995;307(pt 1):93–98. 96. Korge P, Honda HM, Weiss JN. Regulation of the mitochondrial permeability transition by matrix Ca(2+) and voltage during anoxia/reoxygenation. Am J Physiol Cell Physiol. 2001;280(3):C517–C526. 97. Di Lisa F, Menabo R, Canton M, Barile M, Bernardi P. Opening of the mitochondrial permeability transition pore causes depletion of mitochondrial and cytosolic NAD+ and is a causative event in the death of myocytes in postischemic reperfusion of the heart. J Biol Chem. 2001;276(4):2571–2575. 98. Nadtochiy SM, Tompkins AJ, Brookes PS. Different mechanisms of mitochondrial proton leak in ischaemia/reperfusion injury and preconditioning: implications for pathology and cardioprotection. Biochem J. 2006;395(3):611–618. 99. Bosetti F, Baracca A, Lenaz G, Solaini G. Increased state 4 mitochondrial respiration and swelling in early post-ischemic reperfusion of rat heart. FEBS Lett. 2004;563(1–3):161–164. 100. Maroko PR, Libby P, Ginks WR, et al. Coronary artery reperfusion: I. Early effects on local myocardial function and the extent of myocardial necrosis. J Clin Invest. 1972;51(10):2710–2716. 101. Borutaite V, Morkuniene R, Brown GC. Release of cytochrome c from heart mitochondria is induced by high Ca(2+) and peroxynitrite and is responsible for Ca(2+)-induced inhibition of substrate oxidation. Biochim Biophys Acta. 1999;1453(1):41–48.

228

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

102. Dobsak P, Courderot-Masuyer C, Zeller M, et al. Antioxidative properties of pyruvate and protection of the ischemic rat heart during cardioplegia. J Cardiovasc Pharmacol. 1999;34(5):651–659. 103. Kerr PM, Suleiman MS, Halestrap AP. Reversal of permeability transition during recovery of hearts from ischemia and its enhancement by pyruvate. Am J Physiol. 1999;276(2 pt 2):H496–H502. 104. Hatch GM. Cell biology of cardiac mitochondrial phospholipids. Biochem Cell Biol. 2004;82(1):99–112. 105. Kagan VE, Tyurina YY, Bayir H, et al. The “pro-apoptotic genies” get out of mitochondria: oxidative lipidomics and redox activity of cytochrome c/cardiolipin complexes. Chem Biol Interact. 2006;163(1–2):15–28. 106. Paradies G, Petrosillo G, Pistolese M, Ruggiero FM. Reactive oxygen species generated by the mitochondrial respiratory chain affect the complex III activity via cardiolipin peroxidation in beef-heart submitochondrial particles. Mitochondrion. 2001;1(2):151–159. 107. Petrosillo G, Ruggiero FM, Di Venosa N, Paradies G. Decreased complex III activity in mitochondria isolated from rat heart subjected to ischemia and reperfusion: role of reactive oxygen species and cardiolipin. FASEB J. 2003;17(6):714–716. 108. Petrosillo G, Ruggiero FM, Pistolese M, Paradies G. Reactive oxygen species generated from the mitochondrial electron transport chain induce cytochrome c dissociation from beef-heart submitochondrial particles via cardiolipin peroxidation. Possible role in the apoptosis. FEBS Lett. 2001;509(3):435–438. 109. Lesnefsky EJ, Chen Q, Slabe TJ, et al. Ischemia, rather than reperfusion, inhibits respiration through cytochrome oxidase in the isolated, perfused rabbit heart: role of cardiolipin. Am J Physiol Heart Circ Physiol. 2004;287(1):H258–H267. 110. Paradies G, Petrosillo G, Pistolese M, Di Venosa N, Serena D, Ruggiero FM. Lipid peroxidation and alterations to oxidative metabolism in mitochondria isolated from rat heart subjected to ischemia and reperfusion. Free Radic Biol Med. 1999;27(1–2):42–50. 111. Solaini G, Harris DA. Biochemical dysfunction in heart mitochondria exposed to ischaemia and reperfusion. Biochem J. 2005;390(pt 2):377–394. 112. Paradies G, Petrosillo G, Pistolese M, Di Venosa N, Federici A, Ruggiero FM. Decrease in mitochondrial complex I activity in ischemic/reperfused rat heart: involvement of reactive oxygen species and cardiolipin. Circ Res. 2004;94(1): 53–59. 113. Tompkins AJ, Burwell LS, Digerness SB, Zaragoza C, Holman WL, Brookes PS. Mitochondrial dysfunction in cardiac ischemia-reperfusion injury: ROS from complex I, without inhibition. Biochim Biophys Acta. 2006;1762(2):223–231. 114. Paradies G, Petrosillo G, Pistolese M, Ruggiero FM. The effect of reactive oxygen species generated from the mitochondrial electron transport chain on the cytochrome c oxidase activity and on the cardiolipin content in bovine heart submitochondrial particles. FEBS Lett. 2000;466(2–3):323–326. 115. Canton M, Neverova I, Menabo R, Van Eyk J, Di Lisa F. Evidence of myoﬁbrillar protein oxidation induced by postischemic reperfusion in isolated rat hearts. Am J Physiol Heart Circ Physiol. 2004;286(3):H870–H877.

REFERENCES

229

116. Tatarkova Z, Aplan P, Matejovicova M, Lehotsky J, Dobrota D, Flameng W. Effect of ischemia and reperfusion on protein oxidation in isolated rabbit hearts. Physiol Res. 2005;54(2):185–191. 117. Grover GJ, Atwal KS, Sleph PG, et al. Excessive ATP hydrolysis in ischemic myocardium by mitochondrial F1F0-ATPase: effect of selective pharmacological inhibition of mitochondrial ATPase hydrolase activity. Am J Physiol Heart Circ Physiol. 2004;287(4):H1747–H1755. 118. Ylitalo K, Ala-Rami A, Vuorinen K, et al. Reversible ischemic inhibition of F(1)F(0)-ATPase in rat and human myocardium. Biochim Biophys Acta. 2001; 1504(2–3):329–339. 119. Kang D, Hamasaki N. Mitochondrial oxidative stress and mitochondrial DNA. Clin Chem Lab Med. 2003;41(10):1281–1288. 120. Lesnefsky EJ, Hoppel CL. Ischemia-reperfusion injury in the aged heart: role of mitochondria. Arch Biochem Biophys. 2003;420(2):287–297. 121. Lesnefsky EJ, Gudz TI, Moghaddas S, et al. Aging decreases electron transport complex III activity in heart interﬁbrillar mitochondria by alteration of the cytochrome c binding site. J Mol Cell Cardiol. 2001;33(1):37–47. 122. Lawrenson R, Pierce M, Tyrer F. Mortality rates in diabetic patients from a community-based population. Diabetic Med. 1997;14(10):895. 123. Oliveira PJ, Rolo AP, Seica R, Palmeira CM, Santos MS, Moreno AJ. Decreased susceptibility of heart mitochondria from diabetic GK rats to mitochondrial permeability transition induced by calcium phosphate. Biosci Rep. 2001;21(1):45–53. 124. Oliveira PJ, Seica R, Coxito PM, et al. Enhanced permeability transition explains the reduced calcium uptake in cardiac mitochondria from streptozotocin-induced diabetic rats. FEBS Lett. 2003;554(3):511–514. 125. Feuvray D, Lopaschuk GD. Controversies on the sensitivity of the diabetic heart to ischemic injury: the sensitivity of the diabetic heart to ischemic injury is decreased. Cardiovasc Res. 1997;34(1):113–120. 126. Paulson DJ. The diabetic heart is more sensitive to ischemic injury. Cardiovasc Res. 1997;34(1):104–112. 127. Lashin O, Romani A. Mitochondria respiration and susceptibility to ischemiareperfusion injury in diabetic hearts. Arch Biochem Biophys. 2003;420(2):298–304. 128. Deglin SM, Deglin JM, Chung EK. Drug-induced cardiovascular diseases. Drugs. 1977;14(1):29–40. 129. Arcamone F, Cassinelli G, Fantini G, et al. Adriamycin, 14-hydroxydaunomycin, a new antitumor antibiotic from S. peucetius var. caesius. Reprinted from Biotechnology and Bioengineering, Vol. XI, Issue 6, Pages 1101–1110 (1969). Biotechnol Bioeng. 2000;67(6):704–713. 130. Calendi E, Dimarco A, Reggiani M, Scarpinato B, Valentini L. On physico-chemical interactions between daunomycin and nucleic acids. Biochim Biophys Acta. 1965; 103:25–49. 131. Cutts SM, Parsons PG, Sturm RA, Phillips DR. Adriamycin-induced DNA adducts inhibit the DNA interactions of transcription factors and RNA polymerase. J Biol Chem. 1996;271(10):5422–5429. 132. Blum RH, Carter SK. Adriamycin. A new anticancer drug with signiﬁcant clinical activity. Ann Intern Med. 1974;80(2):249–259.

230

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

133. Herman E, Mhatre R, Lee IP, Vick J, Waravdekar VS. A comparison of the cardiovascular actions of daunomycin, adriamycin and N -acetyldaunomycin in hamsters and monkeys. Pharmacology. 1971;6(4):230–241. 134. Lefrak EA, Pitha J, Rosenheim S, Gottlieb JA. A clinicopathologic analysis of adriamycin cardiotoxicity. Cancer. 1973;32(2):302–314. 135. Olson RD, Mushlin PS, Brenner DE, et al. Doxorubicin cardiotoxicity may be caused by its metabolite, doxorubicinol. Proc Natl Acad Sci U S A. 1988;85(10):3585–3589. 136. Steinherz L, Steinherz P. Delayed cardiac toxicity from anthracycline therapy. Pediatrician. 1991;18(1):49–52. 137. Steinherz LJ, Steinherz PG, Tan CT, Heller G, Murphy ML. Cardiac toxicity 4 to 20 years after completing anthracycline therapy. JAMA. 1991;266(12):1672–1677. 138. Bristow MR, Minobe WA, Billingham ME, et al. Anthracycline-associated cardiac and renal damage in rabbits: evidence for mediation by vasoactive substances. Lab Invest. 1981;45(2):157–168. 139. Hrdina R, Gersl V, Klimtova I, Simunek T, Machackova J, Adamcova M. Anthracycline-induced cardiotoxicity. Acta Med (Hradec Kralove). 2000;43(3): 75–82. 140. Olson HM, Capen CC. Chemoresponsiveness of Moloney sarcoma virus–induced osteosarcoma to adriamycin in the rat. Cancer Res. 1978;38(6):1561–1567. 141. Iwasaki T, Suzuki T. Ultrastructural alterations of the myocardium induced by doxorubicin. A scanning electron microscopic study. Virchows Arch B Cell Pathol Incl Mol Pathol. 1991;60(1):35–39. 142. Papadopoulou LC, Theophilidis G, Thomopoulos GN, Tsiftsoglou AS. Structural and functional impairment of mitochondria in adriamycin-induced cardiomyopathy in mice: suppression of cytochrome c oxidase II gene expression. Biochem Pharmacol. 1999;57(5):481–489. 143. Santos DL, Moreno AJM, Leino RL, Froberg MK, Wallace KB. Carvedilol protects against doxorubicin-induced mitochondrial cardiomyopathy. Toxicol App Pharmacol . 2002;185:218–227. 144. Zhou S, Starkov A, Froberg MK, Leino RL, Wallace KB. Cumulative and irreversible cardiac mitochondrial dysfunction induced by doxorubicin. Cancer Res. 2001;61(2):771–777. 145. Olson RD, Mushlin PS. Doxorubicin cardiotoxicity: analysis of prevailing hypotheses. FASEB J. 1990;4(13):3076–3086. 146. Bachur NR, Gordon SL, Gee MV. Anthracycline antibiotic augmentation of microsomal electron transport and free radical formation. Mol Pharmacol. 1977;13(5): 901–910. 147. Wallace KB. Aglycosylation and disposition of doxorubicin in isolated rat liver nuclei and microsomes. Drug Metab Dispos. 1986;14(4):399–404. 148. Scheulen ME, Kappus H. Metabolic activation of adriamycin by NADPH– cytochrome P-450 reductase, rat liver and heart microsomes and covalent protein binding of metabolites. Adv Exp Med Biol. 1981;136(pt A):471–485. 149. Scheulen ME, Kappus H, Nienhaus A, Schmidt CG. Covalent protein binding of reactive adriamycin metabolites in rat liver and rat heart microsomes. J Cancer Res Clin Oncol. 1982;103(1):39–48. 150. Wallace KB. Doxorubicin-induced cardiac mitochondrionopathy. Pharmacol Toxicol. 2003;93(3):105–115.

REFERENCES

231

151. Zhang Y, Marcillat O, Giulivi C, Ernster L, Davies KJ. The oxidative inactivation of mitochondrial electron transport chain components and ATPase. J Biol Chem. 1990;265(27):16330–16336. 152. Doroshow JH. Anthracycline antibiotic-stimulated superoxide, hydrogen peroxide, and hydroxyl radical production by NADH dehydrogenase. Cancer Res. 1983; 43(10):4543–4551. 153. Doroshow JH. Effect of anthracycline antibiotics on oxygen radical formation in rat heart. Cancer Res. 1983;43(2):460–472. 154. Doroshow JH, Davies KJ. Redox cycling of anthracyclines by cardiac mitochondria: II. Formation of superoxide anion, hydrogen peroxide, and hydroxyl radical. J Biol Chem. 1986;261(7):3068–3074. 155. Goodman J, Hochstein P. Generation of free radicals and lipid peroxidation by redox cycling of adriamycin and daunomycin. Biochem Biophys Res Commun. 1977; 77(2):797–803. 156. Demant EJ. Inactivation of cytochrome c oxidase activity in mitochondrial membranes during redox cycling of doxorubicin. Biochem Pharmacol. 1991;41(4): 543–552. 157. Palmeira CM, Serrano J, Kuehl DW, Wallace KB. Preferential oxidation of cardiac mitochondrial DNA following acute intoxication with doxorubicin. Biochim Biophys Acta. 1997;1321(2):101–106. 158. Vidal RF, Eksborg S, Sundberg M, Carlberg M, Elfsson B, Andersson BS. Doxorubicin- and daunorubicin-induced energy deprivation and nucleotide degradation in isolated cardiomyocytes. Toxicology. 1996;114(1):1–10. 159. Wallace KB, Eells JT, Madeira VM, Cortopassi G, Jones DP. Mitochondria-mediated cell injury: symposium overview. Fundam Appl Toxicol. 1997;38(1):23–37. 160. Yoon SB, Kajiyama K, Hino Y, Sugiyama M, Ogura R. Effect of adriamycin on lipid peroxide, glutathione peroxidase and respiratory responses of mitochondria from the heart, liver and kidney. Kurume Med J. 1983;30(1):1–4. 161. Singal PK, Iliskovic N. Doxorubicin-induced cardiomyopathy. N Engl J Med. 1998;339(13):900–905. 162. Singal PK, Iliskovic N, Li T, Kumar D. Adriamycin cardiomyopathy: pathophysiology and prevention. FASEB J. 1997;11(12):931–936. 163. Ferrero ME, Ferrero E, Gaja G, Bernelli-Zazzera A. Adriamycin: energy metabolism and mitochondrial oxidations in the heart of treated rabbits. Biochem Pharmacol. 1976;25(2):125–130. 164. Mailer K, Petering DH. Inhibition of oxidative phosphorylation in tumor cells and mitochondria by daunomycin and adriamycin. Biochem Pharmacol. 1976;25(18):2085–2089. 165. Bachmann E, Weber E, Zbinden G. Effects of mitoxantrone and doxorubicin on energy metabolism of the rat heart. Cancer Treat Rep. 1987;71(4):361–366. 166. Solem LE, Heller LJ, Wallace KB. Dose-dependent increase in sensitivity to calcium-induced mitochondrial dysfunction and cardiomyocyte cell injury by doxorubicin. J Mol Cell Cardiol. 1996;28(5):1023–1032. 167. Solem LE, Henry TR, Wallace KB. Disruption of mitochondrial calcium homeostasis following chronic doxorubicin administration. Toxicol Appl Pharmacol. 1994;129(2):214–222.

232

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

168. Oliveira PJ, Santos MS, Wallace KB. Doxorubicin-induced thiol-dependent alteration of cardiac mitochondrial permeability transition and respiration. Biochemistry (Mosc). 2006;71(2):194–199. 169. Gosalvez M, Blanco M, Hunter J, Miko M, Chance B. Effects of anticancer agents on the respiration of isolated mitochondria and tumor cells. Eur J Cancer. 1974;10(9):567–574. 170. Demant EJ. NADH oxidation in submitochondrial particles protects respiratory chain activity against damage by adriamycin-Fe3+ . Eur J Biochem. 1983;137(1–2): 113–118. 171. Muraoka S, Miura T. Inactivation of mitochondrial succinate dehydrogenase by adriamycin activated by horseradish peroxidase and hydrogen peroxide. Chem Biol Interact. 2003;145(1):67–75. 172. Jeyaseelan R, Poizat C, Wu HY, Kedes L. Molecular mechanisms of doxorubicininduced cardiomyopathy: selective suppression of Reiske iron–sulfur protein, ADP/ATP translocase, and phosphofructokinase genes is associated with ATP depletion in rat cardiomyocytes. J Biol Chem. 1997;272(9):5828–5832. 173. Nicolay K, de Kruijff B. Effects of adriamycin on respiratory chain activities in mitochondria from rat liver, rat heart and bovine heart: evidence for a preferential inhibition of complex III and IV. Biochim Biophys Acta. 1987;892(3):320–330. 174. Oliveira PJ, Bjork JA, Santos MS, et al. Carvedilol-mediated antioxidant protection against doxorubicin-induced cardiac mitochondrial toxicity. Toxicol Appl Pharmacol. 2004;200(2):159–168. 175. Moore L, Landon EJ, Cooney DA. Inhibition of the cardiac mitochondrial calcium pump by adriamycin in vitro. Biochem Med. 1977;18(2):131–138. 176. Revis NW, Marusic N. Effects of doxorubicin and its aglycone metabolite on calcium sequestration by rabbit heart, liver, and kidney mitochondria. Life Sci. 1979;25(12):1055–1063. 177. Chacon E, Acosta D. Mitochondrial regulation of superoxide by Ca2+ : an alternate mechanism for the cardiotoxicity of doxorubicin. Toxicol Appl Pharmacol. 1991;107(1):117–128. 178. Chacon E, Ulrich R, Acosta D. A digitized-ﬂuorescence-imaging study of mitochondrial Ca2+ increase by doxorubicin in cardiac myocytes. Biochem J. 1992; 281(pt 3):871–878. 179. Oliveira PJ, Rolo AP, Sardao VA, et al. Advantages in the use of carvedilol versus propranolol for the protection of cardiac mitochondrial function. Rev Port Cardiol. 2004;23(10):1291–1298. 180. Oliveira PJ, Wallace KB. Depletion of adenine nucleotide translocator protein in heart mitochondria from doxorubicin-treated rats: relevance for mitochondrial dysfunction. Toxicology. 2006;220(2–3):160–168. 181. Solem LE, Wallace KB. Selective activation of the sodium-independent, cyclosporin A–sensitive calcium pore of cardiac mitochondria by doxorubicin. Toxicol Appl Pharmacol. 1993;121(1):50–57. 182. Al-Nasser IA. In vivo prevention of adriamycin cardiotoxicity by cyclosporin A or FK506. Toxicology. 1998;131(2–3):175–181. 183. Halloran PF, Kung L, Noujaim J. Calcineurin and the biological effect of cyclosporine and tacrolimus. Transplant Proc. 1998;30(5):2167–2170.

REFERENCES

233

184. Zhou S, Heller LJ, Wallace KB. Interference with calcium-dependent mitochondrial bioenergetics in cardiac myocytes isolated from doxorubicin-treated rats. Toxicol Appl Pharmacol. 2001;175(1):60–67. 185. Zhou S, Palmeira CM, Wallace KB. Doxorubicin-induced persistent oxidative stress to cardiac myocytes. Toxicol Lett. 2001;121(3):151–157. 186. Ascenc¸a˜ o A, Magalhaes J, Soares JM, et al. Moderate endurance training prevents doxorubicin-induced in vivo mitochondriopathy and reduces the development of cardiac apoptosis. Am J Physiol Heart Circ Physiol. 2005;289(2):H722–H731. 187. Chen Y, Daosukho C, Opii WO, et al. Redox proteomic identiﬁcation of oxidized cardiac proteins in adriamycin-treated mice. Free Radic Biol Med. 2006;41(9):1470–1477. 188. Faustin B, Rossignol R, Rocher C, Benard G, Malgat M, Letellier T. Mobilization of adenine nucleotide translocators as molecular bases of the biochemical threshold effect observed in mitochondrial diseases. J Biol Chem. 2004;279(19):20411–20421. 189. Praet M, Calderon PB, Pollakis G, Roberfroid M, Ruysschaert JM. A new class of free radical scavengers reducing adriamycin mitochondrial toxicity. Biochem Pharmacol. 1988;37(24):4617–4622. 190. Siveski-Iliskovic N, Hill M, Chow DA, Singal PK. Probucol protects against adriamycin cardiomyopathy without interfering with its antitumor effect. Circulation. 1995;91(1):10–15. 191. Singal PK, Khaper N, Palace V, Kumar D. The role of oxidative stress in the genesis of heart disease. Cardiovasc Res. 1998;40(3):426–432. 192. Yue TL, Cheng HY, Lysko PG, et al. Carvedilol, a new vasodilator and beta adrenoceptor antagonist, is an antioxidant and free radical scavenger. J Pharmacol Exp Ther. 1992;263(1):92–98. 193. Yue TL, McKenna PJ, Ruffolo RR, Jr., Feuerstein G. Carvedilol, a new beta-adrenoceptor antagonist and vasodilator antihypertensive drug, inhibits superoxide release from human neutrophils. Eur J Pharmacol. 1992;214(2–3):277–280. 194. Xiong Y, Liu X, Lee CP, Chua BH, Ho YS. Attenuation of doxorubicin-induced contractile and mitochondrial dysfunction in mouse heart by cellular glutathione peroxidase. Free Radic Biol Med. 2006;41(1):46–55. 195. Arola OJ, Saraste A, Pulkki K, Kallajoki M, Parvinen M, Voipio-Pulkki LM. Acute doxorubicin cardiotoxicity involves cardiomyocyte apoptosis. Cancer Res. 2000;60(7):1789–1792. 196. Childs AC, Phaneuf SL, Dirks AJ, Phillips T, Leeuwenburgh C. Doxorubicin treatment in vivo causes cytochrome c release and cardiomyocyte apoptosis, as well as increased mitochondrial efﬁciency, superoxide dismutase activity, and Bcl-2:Bax ratio. Cancer Res. 2002;62(16):4592–4598. 197. Kotamraju S, Konorev EA, Joseph J, Kalyanaraman B. Doxorubicin-induced apoptosis in endothelial cells and cardiomyocytes is ameliorated by nitrone spin traps and ebselen: role of reactive oxygen and nitrogen species. J Biol Chem. 2000;275(43):33585–33592. 198. Kumar D, Jugdutt BI. Apoptosis and oxidants in the heart. J Lab Clin Med. 2003;142(5):288–297. 199. Boucek RJ, Jr., Olson RD, Brenner DE, Ogunbunmi EM, Inui M, Fleischer S. The major metabolite of doxorubicin is a potent inhibitor of membrane-associated ion

234

200.

201. 202.

203.

204.

205. 206.

207.

208.

209.

210. 211.

CARDIOVASCULAR TOXICITY OF MITOCHONDRIAL ORIGIN

pumps: a correlative study of cardiac muscle with isolated membrane fractions. J Biol Chem. 1987;262(33):15851–15856. Kim DS, Kim HR, Woo ER, Hong ST, Chae HJ, Chae SW. Inhibitory effects of rosmarinic acid on adriamycin-induced apoptosis in H9c2 cardiac muscle cells by inhibiting reactive oxygen species and the activations of c-Jun N-terminal kinase and extracellular signal-regulated kinase. Biochem Pharmacol. 2005;70(7):1066–1078. Kim DS, Kim HR, Woo ER, et al. Protective effect of calceolarioside on adriamycininduced cardiomyocyte toxicity. Eur J Pharmacol. 2006;541(1–2):24–32. Lewis W, Kohler JJ, Hosseini SH, et al. Antiretroviral nucleosides, deoxynucleotide carrier and mitochondrial DNA: evidence supporting the DNA pol gamma hypothesis. AIDS 2006;20(5):675–684. Lund KC, Wallace KB. Direct, DNA pol-gamma-independent effects of nucleoside reverse transcriptase inhibitors on mitochondrial bioenergetics. Cardiovasc Toxicol. 2004;4(3):217–228. Lewis W, Simpson JF, Meyer RR. Cardiac mitochondrial DNA polymerase-gamma is inhibited competitively and noncompetitively by phosphorylated zidovudine. Circ Res. 1994;74(2):344–348. Lund KC, Wallace KB. Direct effects of nucleoside reverse transcriptase inhibitors on rat cardiac mitochondrial bioenergetics. Mitochondrion. 2004;4(2–3):193–202. Lynx MD, Bentley AT, McKee EE. 3 -Azido-3 -deoxythymidine (AZT) inhibits thymidine phosphorylation in isolated rat liver mitochondria: a possible mechanism of AZT hepatotoxicity. Biochem Pharmacol. 2006;71(9):1342–1348. Lynx MD, McKee EE. 3 -Azido-3 -deoxythymidine (AZT) is a competitive inhibitor of thymidine phosphorylation in isolated rat heart and liver mitochondria. Biochem Pharmacol. 2006;72(2):239–243. de la Asuncion JG, Del Olmo ML, Gomez-Cambronero LG, Sastre J, Pallardo FV, Vina J. AZT induces oxidative damage to cardiac mitochondria: protective effect of vitamins C and E. Life Sci. 2004;76(1):47–56. Szabados E, Fischer GM, Toth K, et al. Role of reactive oxygen species and poly-ADP-ribose polymerase in the development of AZT-induced cardiomyopathy in rat. Free Radic Biol Med. 1999;26(3–4):309–317. Valenti D, Barile M, Passarella S. AZT inhibition of the ADP/ATP antiport in isolated rat heart mitochondria. Int J Mol Med. 2000;6(1):93–96. Lewis W, Papoian T, Gonzalez B, et al. Mitochondrial ultrastructural and molecular changes induced by zidovudine in rat hearts. Lab Invest. 1991;65(2):228–236.

7 SKELETAL MUSCLE AND MITOCHONDRIAL TOXICITY Timothy E. Johnson Department of Safety Assessment, Merck Research Laboratories, West Point, Pennsylvania

1. Introduction 2. Drug- and other xenobiotic-induced skeletal muscle myopathy and rhabdomyolysis 2.1. Statins 2.2. Fibrates 2.3. Statin and fibrate synergy 2.4. Nucleoside reverse transcriptase inhibitors 2.5. Other xenobiotics associated with skeletal muscle injury 3. Conclusions

235 236 236 240 242 243 244 245

1. INTRODUCTION A variety of drugs and other xenobiotics cause skeletal muscle myopathy and there is accumulating evidence that the toxicity induced by at least some of these agents is mediated by impairing mitochondrial function. The clearest examples are the antiretroviral nucleoside reverse transcriptase inhibitors, which inhibit mitochondrial DNA synthesis and subsequently deplete the number of mitochondria. Other xenobiotics, including bupivacaine, veratum alkaloids, methylenedioxymethamphetamine, and organophosphate pesticides, also damage skeletal muscle by disrupting mitochondrial function. The role of mitochondria in statin-or Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

235

236

SKELETAL MUSCLE AND MITOCHONDRIAL TOXICITY

ﬁbrate-induced myopathy is less clear. There are reports supporting mitochondrial involvement in high-dose acute exposures to these drugs, but little to no evidence at pharmacological concentrations, suggesting that other pharmacokinetic considerations may be involved. Statins and ﬁbrates also cause myotoxicity through different pathways and target different ﬁber types in skeletal muscle, making it unlikely that they act synergistically to induce myopathy.

2. DRUG- AND OTHER XENOBIOTIC-INDUCED SKELETAL MUSCLE MYOPATHY AND RHABDOMYOLYSIS The classical deﬁnition of skeletal myopathy is an unexplained muscle pain or weakness and at least a 10-fold increase in creatine phosphokinase (CK) levels in the plasma. However, it has been recognized recently that not all skeletal myopathy results in a rise in CK. For example, Phillips et al. identiﬁed a subset of patients taking statins who had normal CK levels but exhibited ragged-red ﬁbers, a hallmark of muscle degeneration, upon biopsy [1]. Thus, CK measurements are not always a reliable clinical tool to predict myopathy. As myopathy progresses, the muscle continues to break down, resulting in the release of large amounts of myoglobin into the blood, which can compromise kidney function and in some cases can progress to kidney failure and even death. Clinically, rhabdomyolysis is diagnosed when the patient exhibits signs of myopathy, has brown-colored urine, and/or has extreme elevations of CK, usually greater than 10,000 units [2]. For most drug-induced myopathies, the symptoms can usually be reversed when treatment is stopped, as long as permanent kidney damage has not occurred. A wide diversity of drugs and xenobiotics can cause myopathy. The best studied are the statins, ﬁbrates, and nucleoside reverse transcriptase inhibitors. However, skeletal myotoxicity has also been observed with glucocorticoids, niacin, veratum alkaloids, bupivacaine, organophosphate insecticides, and methylenedioxymethamphetamine. Compared with other tissues, skeletal muscle contains numerous mitochondria, a reﬂection of its high energy demands. Drug-induced mitochondrial damage can result in muscle cell death and compromise the function of the myoﬁber and the muscle bundle. Thus, not surprisingly, many drugs that damage skeletal muscle have been postulated to cause a mitochondrial insult. In the remainder of this chapter I discuss the evidence supporting and refuting this hypothesis. 2.1. Statins Statins inhibit hydroxymethylglutaryl (HMG)–coenzyme A (CoA) reductase and subsequently decrease plasma cholesterol. They are commonly prescribed for hypercholesteremia, but recent evidence suggests that they might have additional activities that provide protection against cardiovascular disease. Statins are safe and welltolerated, but the major adverse effect, occurring in less than 1% of

DRUG- AND OTHER XENOBIOTIC-INDUCED SKELETAL MUSCLE MYOPATHY

237

patients, is skeletal myopathy [3]. All marketed statins have a comparable incidence of producing myopathy, with the exception of cerivastatin, which was withdrawn from clinical use due to an unacceptable number of rhabdomyolysis cases. Myopathy risk increases with statin concentration and with coadministration of other xenobiotics, including ﬁbrates, niacin, grapefruit juice, and calcium channel blockers [4]. Although some of these drug–drug interactions act to raise statin plasma drug levels, the mechanism of statin-induced myopathy and its risk with combination treatments are still poorly understood. There also appears to be an unknown genetic predisposition in certain individuals, resulting in an increased susceptibility to statin and other drug-associated myopathies. Interestingly, many congenital myopathies are associated with defects in mitochondrial enzymes [5,6] (see Chapters 1 and 11). Mitochondrial Involvement in Statin-Induced Myopathy A number of reports describe acute effects of statins on skeletal muscle mitochondria. In L6 myoblasts, the mitochondrial membrane potential decreased more than 50% by 100 µM cerivastatin, ﬂuvastatin, and atorvastatin. In mitochondria isolated from rat skeletal muscle the glutamate-driven stage 3 respiration respiratory control ratio was also affected [7]. Lovastatin and simvastatin (10 to 80 µM) were reported in vitro to induce mitochondrial permeability transition (MPT), and they also decreased the content of total mitochondrial membrane thiol groups in mitochondria isolated from mouse hindlimb [8]. This same group also found that mitochondria isolated from LDL receptor knockout mice, treated with 100 mg/kg lovastatin, had a higher incidence of developing MPT. Ex vivo treatment with simvastatin caused mitochondrial membrane depolarization (EC50 = 1.96 µM) and triggered release of cytoplasmic calcium (EC50 = 7.8 µM) in isolated muscle ﬁbers from human skeletal muscle biopsies [9]. Degenerate mitochondria were also observed in rat skeletal muscle ﬁbers treated with 0.5 mg/kg cerivastatin [10]. Although these ﬁndings point to mitochondrial involvement in statin-induced myotoxicty, the micromolar concentrations used in these studies greatly exceed normal pharmacological doses, undermining the justiﬁed inference of cause and effect and suggesting that other pharmacokinetic factors may also play a role (see section 3). Ubiquinone Depletion as a Factor in Statin-Induced Myopathy Another favored hypothesis of statin-induced myopathy centers around ubiquinone levels [11,12]. Ubiquinone is synthesized from farnesylpyrophosphate, an intermediate in the mevalonate pathway. This long-chain molecule plays a critical role in oxidative phosphorylation. Although this hypothesis is plausible given the ability of statins to inhibit mevalonate synthesis and thus the production of isoprenoids, including farnesylpyrophosphate, there is no strong evidence to support it. Ubiquinone levels have been observed to be decreased in plasma in rats and humans treated with simvastatin or lovastatin, but there was no apparent effect in skeletal muscle [13,14].

238

SKELETAL MUSCLE AND MITOCHONDRIAL TOXICITY

In Vitro Studies in Rat and Human Myotube Cultures To further test these hypotheses of statin-induced myopathy, we optimized a rat and human myotube culture system using L6 cells and fetal-derived human skeletal muscle cells [15]. Rat L6 and human myoblasts were grown on matrigel- or laminin-coated plates and differentiated for 6 or 3 days, respectively, in a differentiation cocktail that included Opti-MEM, 2% horse serum, and 5 µg/mL insulin. We characterized each system for morphology, intracellular CK activity, and mRNA expression of muscle markers (e.g., CK, myoD, and myogenin). We found that statins, including cerivastatin and a Merck experimental statin that is structurally similar to lovastatin and simvastatin (compound A), induced myotoxicity in rat myotubes in a dose- and time-dependent manner. Interestingly, we observed that concentrations as low as 50 nM of compound A caused a signiﬁcant increase in TUNEL-positive nuclei at 72 hours (Figure 1). In addition, the highly potent HMG-CoA reductase inhibitor cerivastatin caused myotoxicity at exposures as low as 50 nM after 7 hours, consistent with the hypothesis of a HMG-CoA reductase mechanism–based toxicity. Thus, statins induced myotoxicity in our myotube culture model at more relevant pharmacological concentrations. Furthermore, we found that statins cause cell death through a caspase-3-dependent apoptosis, and the ability of geranylgeraniol, but not farnesol, to prevent cell death suggests that the inhibition of the isoprenylation of geranylgeranylated proteins is in part responsible for the toxicity. Supporting this hypothesis, we found that several GGTase inhibitors also caused myotoxicity in this model, while an FTase inhibitor had little or no effect. Similar results were seen in human myotube cultures. Although we did not examine the effect of

% TUNEL Positive Nuclei

40 ∗ ∗

30 ∗ 20 ∗ 10

0

DMSO

50

100 500 Cmpd A (nM)

1000

Figure 1 Effect of statin (compound A) to induce apoptosis in rat myotube cultures. L6 myoblasts were cultured and differentiated into myotubes. On day 6 of differentiation, myotube cultures were treated with compound A (Cmpd A) at the concentrations indicated for 72 hours. Shown is the mean percentage of TUNEL-positive nuclei, ±2 SE, of duplicate wells. ∗ , Statistically signiﬁcant (p < 0.05) compared with DMSO treatment.

239

DRUG- AND OTHER XENOBIOTIC-INDUCED SKELETAL MUSCLE MYOPATHY

statins on mitochondrial function in these studies, the activation of caspase-3 is consistent with cytochrome c release from the mitochondria [16]. We also tested the ability of cerivastatin to decrease ubiquinone concentration in our rat and human myotube cultures. We observed that the levels of coenzyme Q (CoQ9 ) (rat) and CoQ10 (human), although variable between treatments, did not exhibit a clear dose response or correlate with apoptosis (Figure 2). Furthermore, in cultures treated with cerivastain in the presence of mevalonate, the ubiquinone levels did not change, despite the complete prevention of toxicity. Our ﬁndings are supported from in vivo observations by Schaefer et al. [17], who studied the effect of a 15-day treatment with up to 1 mg/kg cerivastatin on CoQ9

150

50

CoQ10 (ng/mg)

CoQ9 (ng/mg)

60

40 30 20

100

∗ ∗

50

∗

10 0

∗

10

∗

8 6

∗ ∗

4 2 0 Medium

Ceriva (nM)

∗∗

Mev

500

100

∗

50

10

∗

2

Mev + Ceriva (1000)

∗

4

0

Complete Toxicity

6

1000

8

TUNEL (Fold Increase)

∗

Medium

TUNEL (Fold Increase)

0

10

50

100 500 1000

Ceriva (nM)

Figure 2 Statin-induced myotoxicity in rat and human myotube cultures is not associated with ubiquinone depletion. Differentiated rat and human myotube cultures were treated for 96 hours with cerivastatin (Ceriva) at the concentrations indicated. For rat myotubes, the 1 µM Ceriva concentration was also assessed in the presence of mevalonic acid lactone (Mev, 100 µM). Cultures were ﬁxed and TUNEL-stained and a parallel set was lysed and assayed for CoQ9 (rat) or CoQ10 (human) concentration. Shown is the mean percentage of TUNEL-positive nuclei, expressed as an x -fold increase over the medium control, ±2 SE, of triplicate wells. Ubiquinone concentration was normalized to total protein and was expressed as the mean concentration ±2 SE of triplicate samples. All compounds were dissolved directly in culture medium instead of DMSO. ∗ , Statistically signiﬁcant (p < 0.05) compared with the medium treatment. ∗∗ , Statistically signiﬁcant (p < 0.05) compared with Ceriva treatment.

240

SKELETAL MUSCLE AND MITOCHONDRIAL TOXICITY

levels and mitochondrial function in rat quadriceps. Although ultrastructurally, some degenerating myoﬁbers were seen that contained a few swollen mitochondria with disorganized cristae, there was no signiﬁcant decrease in CoQ9 levels over the course of the study. Furthermore, mitochondrial function, including respiration, membrane potential, and substrate-linked ATP, was evaluated in skeletal muscle homogenates and there were no effects on these parameters at any dose level at day 5 or 10 of the study. Finally, the direct effects of 10 µM cerivastatin on mitochondrial function were examined. No effect on state 3 or state 4 respiration or on other mitochondrial endpoints was found. Taken together, these in vitro myotube and in vivo studies suggest that other than perhaps cytochrome c release, statins do not disrupt mitochondrial function or decrease the levels of ubiquinone. 2.2. Fibrates Fibrates comprise another class of compounds that cause myopathy. These drugs are an effective therapy for hypertriglyceridemia and have been used safely for decades. All ﬁbrates, including gemﬁbrozil, fenoﬁbrate, and bezaﬁbrate, can cause skeletal muscle myopathy, and the incidence ranges from 0.1 to 0.5% [18,19]. It is now widely accepted that ﬁbrates bind to a family of nuclear receptors, peroxisome proliferator-activated receptors (PPARs). There are three PPAR subtypes: alpha, gamma, and delta (also known as beta and hNUC1). Most ﬁbrates, including fenoﬁbrate, gemﬁbrozil, and WY-14643, bind preferentially to PPARα, but bezaﬁbrate, the most commonly used ﬁbrate in Europe, was reported to be a pan agonist for all three PPAR subtypes in human and mouse transactivation assays [20]. The mechanism of ﬁbrate-induced myopathy is unknown, but it has been hypothesized to be mediated pharmacologically through PPARα. However, other PPAR subtypes have also been shown to be expressed and/or act in skeletal muscle [21–23]. Mitochondrial Involvement in Fibrate-Induced Myotoxicity Fibrates have been shown to impair mitochondrial function in several studies. For example, in HL-60 cells, high micromolar concentrations of cloﬁbric acid, bezaﬁbrate, gemﬁbrozil, and ciglitizone were reported to increase lactate and acetate levels signiﬁcantly and increase glucose consumption, suggesting a compensatory increase in glycolysis [24]. A similar proﬁle was seen in the ability of the compounds to inhibit NADH–cytochrome c reductase activity, and interestingly, there was a correlation between the ability of these compounds to induce mitochondrial toxicity and inhibit cell growth. In rat skeletal muscle homogenates and in isolated mitochondria, ex vivo treatment with fenoﬁbrate dose inhibited complex I activity dependently [25]. Other studies have found that WY-14643 acts a metabolic uncoupler [26] and that peroxisome proliferators induce MPT [27]. Thus, a hypothesis has been put forth that ﬁbrates (PPARα ligands) and thiazolidinediones (PPARγ agonists) induce toxicity by disrupting mitochondrial function through a mechanism that is not entirely dependent on PPARs [28].

241

DRUG- AND OTHER XENOBIOTIC-INDUCED SKELETAL MUSCLE MYOPATHY

6 5

TUNEL (Fold Increase)

24 hr 48 hr

4 3 ∗

2 1 0

DMSO 1 10 100 1000 10000 Rosiglitazone (nM)

A

C

TUNEL (Fold Increase)

TUNEL (Fold Increase)

Evaluation of PPAR Agonist–Induced Myotoxicity in Rat Myotubes Using our rat myotube model, we examined the ability of ﬁbrates and PPAR subtype selective agonists to induce myotoxicity [29]. We found that ﬁbrates, including WY-14643, gemﬁbrozil, and bezaﬁbrate, increased the number of TUNEL-positive nuclei in a dose-and time-dependent manner. We also observed that a potent Merck PPARα agonist induced myotoxicity after 24 hours in these cultures at exposures as low as 1 nM. In contrast, no toxicity was observed with the PPARγ selective compound Rosiglitazone at up to 48 hours, and comparatively less cell killing was noted for the PPARδ selective agonist GW-501516 (Figure 3). The data support the hypothesis that PPARα is mediating part of the myotoxicity induced by ﬁbrates. Because an increase in TUNEL staining can be caused by apoptosis or necrosis, we examined the ability of the Merck PPARα agonist to induce caspase 3/7 activation. Interestingly, at 10 µM, a concentration that caused about a sixfold increase in TUNEL staining, we found no evidence for caspase 3/7 activation at 7 or 24 hours. In contrast, staurosporine and a Merck experimental statin induced a signiﬁcant increase in apoptosis. We also examined the effect of the Merck PPARα agonist on

5

24 hr 48 hr

∗

4 3 2

∗∗ ∗∗

∗∗

∗

∗∗

1 0

DMSO 1 10 100 100010000 GW-501516 (nM)

∗

8 6 4

B

6

∗

∗

∗

∗

2 0 DMSO 1 10 100 100010000 Merck PPARa agonist (nM)

Figure 3 Selectivity of myotoxicity response induced by PPAR agonists. Differentiated cultures were treated with the indicated compounds and analyzed for TUNEL staining. (A) Cells were treated with rosiglitazone for 24 or 48 hours at the concentrations indicated. (B) Cultures were treated with GW-501516 for 24 or 48 hours at the concentrations indicated. (C) Cells were treated with compound A for 24 hours at the concentrations indicated. Shown is the mean percentage of TUNEL-positive nuclei, expressed as an x -fold increase over the DMSO control, ±2 SE, of duplicate or triplicate wells. ∗ , Statistically signiﬁcant (p < 0.05) compared with the concurrent DMSO treatment.

242

SKELETAL MUSCLE AND MITOCHONDRIAL TOXICITY

ATP levels in the myotube cultures. Compared with lithocolic acid, a known mitochondrial uncoupler, or a statin, we did not see a signiﬁcant decrease in ATP levels with the Merck PPARα agonist (10 µM) at up to 48 hours. Thus, in our rat myotube model, ﬁbrates and other PPARα agonists appear to cause myotoxicity through a mechanism that is different from that seen with statins. Although we did not examine mitochondrial function per se, there was no signiﬁcant decrease in ATP levels at a clearly myotoxic exposure, suggesting that the myotoxicity was not related to mitochondrial perturbation, or that glycolytic capacity was sufﬁcient to maintain the adenylate charge. 2.3. Statin and Fibrate Synergy An increasing number of patients exhibit mixed hyperlipidemia, resulting in the need for statin–ﬁbrate combination therapy. It is clear that there is an inherent myopathy risk with statin–ﬁbrate combination treatment, but the estimates vary widely, depending on the reporting conditions, ranging from 0.2% to as high as 1 to 5% [30–32]. Furthermore, the majority of deaths attributed to cerivastatin occurred in patients also taking ﬁbrates, suggesting that combination therapy with certain statins poses a signiﬁcant health risk [33,34]. There is also a concern that statins and ﬁbrates may act synergistically to promote skeletal muscle myopathy [35]. Some genes are induced synergistically by statins and PPAR agonists [36–39], and we reported previously that peroxisome proliferators (PPARα ligands) and fatty acids regulate the mevalonate pathway negatively in liver cells [40]. In addition, pharmacokinetic interactions between statins and ﬁbrates have been observed [41,42]. Accordingly, we also examined the effect of combination ﬁbrate/statin treatment on myotoxicity in our rat myotube culture system [29]. Although we found at most a roughly additive effect with a Merck statin (compound B) and WY-14643, and with atorvastatin and gemﬁbrozil, there was no evidence for a synergistic effect with two different statins and three ﬁbrates (Figure 4). Fiber Type Selectiveness in Statin- and Fibrate-Induced Myopathy Other studies provide evidence that statins and ﬁbrates preferentially act in a ﬁber-type speciﬁc manner. For example, statin-induced lesions in rat skeletal muscle occurred primarily in type II (mitochondrial poor, glycolytic) ﬁbers that were devoid of glycogen [43]. Using the same technique, supplemented by immnunohistochemistry for fast and slow myosin, Westwood et al. found that type II ﬁbers showed a low-grade necrosis in response to simvastatin and cerivastatin treatment. Interestingly, the adjacent type I ﬁbers present in the same muscle bundle were largely unaffected [44]. This group also found that the most sensitive ﬁber to statin-induced necrosis was type IIb, the most glycolytic. In contrast, there is some evidence that ﬁbrates and other PPARα agonists primarily affect type I (mitochondrial rich, oxidative) ﬁbers. For example, gene expression proﬁling in rat skeletal muscle revealed a PPARα signature in soleus (type I) but not quadriceps (type II), suggesting that ﬁbrates regulate genes in

40

None + Cmpd B Difference

30

∗

20 ∗

10 0

DMSO

% TUNEL Positive Nuclei

A

B

∗

∗

1 10 100 WY-14643 (mM) None + Cmpd B Difference ∗ ∗

20 10

500

∗ ∗

−10 1 10 100 Bezafibrate (mM)

500

D

243

25 None + Atorva Difference ∗

20 15

∗

10

∗

∗∗

∗

∗

5 0

C

0

DMSO

% TUNEL Positive Nuclei

50

% TUNEL Positive Nuclei

% TUNEL Positive Nuclei

DRUG- AND OTHER XENOBIOTIC-INDUCED SKELETAL MUSCLE MYOPATHY

DMSO 1 10 100 500 Gemfibrozil (mM)

20

None + Atorva Difference ∗

15 10

∗

∗

5 0 −5 DMSO 1 10 100 500 Bezafibrate (mM)

Figure 4 Effects of compound B or atorvastatin alone and in combination with PPARα agonists to induce myotoxicity. Differentiated cultures were treated with the compounds indicated and analyzed for TUNEL staining. (A) Cells were treated with 1 µM compound B for 24 hours in the presence or absence of the concentrations of WY-14643 indicated. (B) Cultures were exposed to 1 µM compound B for 24 hours in the presence or absence of the concentrations of bezaﬁbrate indicated. (C) Cells were treated with 0.5 µM atorvastatin (Atorva) for 24 hours in the presence or absence of the concentrations of gemﬁbrozil indicated. (D) Cultures were exposed to 0.5 µM atorvastatin (Atorva) for 24 hours in the presence or absence of the concentrations of Bezaﬁbrate indicated. Shown is the mean percentage of TUNEL-positive nuclei, ±2 SE, of duplicate or triplicate wells. The net difference (hatched bars) between the statin (compound B or atorva alone) and statin (compound B or atorva)/PPAR agonist treatment is shown. None of the differences were statistically signiﬁcant. ∗ , Statistically signiﬁcant (p < 0.05) as compared with DMSO treatment in the absence of statin (compound B or atorva) and PPAR agonists.

oxidatively rich ﬁbers, which is consistent with their known involvement in fatty acid β-oxidation pathways [45]. Thus, in rats there is no evidence to support the hypothesis that statins and ﬁbrates would synergize to enhance myopathy, since they appear to cause toxicity through independent pathways and in different ﬁber types. However, an additive toxicity is possible because in some skeletal muscles there is a mix of ﬁber types. This is especially true in humans. 2.4. Nucleoside Reverse Transcriptase Inhibitors One class of drugs where there is a clear association with mitochondrial toxicity are nucleoside reverse transcriptase inhibitors (NRTIs). These drugs include 3 -azido-3 -deoxythymidine (AZT), ﬁaluridine (FIAU), zidovudine (ZDV),

244

SKELETAL MUSCLE AND MITOCHONDRIAL TOXICITY

didanosine (ddI), zalcitabine (ddC), stavudine (d4T), lamivudine (3TC), abacavir (ABC), and emtricitabine (FTC), which are used as antiretroviral therapies in HIV and other retroviral diseases. Clinical manifestations of NRTI toxicity include lactic acidosis, elevated increases in serum liver enzymes (indicating cell death), and skeletal and cardio myopathies [46]. NRTIs are chemically modiﬁed structural analogs of purines and pyrimidines and can serve as substrates for the retroviral polymerases. They act by preventing nucleotide addition, resulting in premature termination of the viral DNA strand and thus block retroviral replication [47]. Mitochondrial Impairment in NRTI-Associated Myopathy NRTI toxicity is thought to be mediated through the inhibition of mitochondrial DNA polymerase-gamma, which results in suppression of mitochondrial DNA synthesis and subsequent mitochondrial depletion [48]. Recently, however, NRTI inhibition of mitochondrial thymidine kinase has also been proposed [49]. Most of the mechanistic evidence indicating an effect on mitochondria comes from in vitro experiments conducted in human and rodent cells. Studies in T-lymphoblastoid cells found that NRTIs reduced mitochondrial (mt) DNA content and increased lactate production, which correlated with a decrease in cell number [50–52]. In skeletal muscle derived from human biopsies, ddI, ddC, and ZDV increased lactate levels and inhibited the activities of respiratory chain complexes II and IV [53]. Supporting these observations, ddC and ddI reduced the level of cytochrome c oxidase expression in skeletal muscle cells [54]. There is also some in vivo evidence suggesting that NRTIs affect mitochondrial function. Using a stable isotope mass spectrometric method to measure mtDNA synthesis, Collins et al. demonstrated in rat cardiac and hindlimb muscle that AZT treatment for up to 8 weeks caused about a twofold reduction in mtDNA fractional synthesis [55]. In addition, cytochrome c oxidase content was decreased by 50% in the hindlimb after 4 weeks of treatment. Furthermore, in human muscle biopsies from patients treated with NRTIs, mtDNA was markedly reduced, which correlated with a decrease in whole-body oxidative capacity in the subjects [56]. Collectively, these in vitro and in vivo studies strongly support the hypothesis that NRTIs cause myopathy through a mechanism that impairs mitochondrial function and leads to a depletion in mitochondrial number. 2.5. Other Xenobiotics Associated with Skeletal Muscle Injury A number of other agents appear to cause myopathy by affecting mitochondrial function. For example, the local anesthetic bupivacaine causes muscle degeneration, and this has been linked to mitochondrial depolarization and an opening of the MPT [57]. 3,4-Methylenedioxymethamphetamine (MDMA), a substance of abuse popularly known as ecstasy, can cause skeletal myopathy and rhabdomyolysis that can lead to death. In a rat study, MDMA treatment caused elevated increases in plasma CK which correlated with its ability to act as an oxidative uncoupler in skeletal muscle [58]. Other examples include veratine, a mixture

CONCLUSIONS

245

of alkaloids extracted from certain plants that causes skeletal muscle toxicity. In isolated rat skeletal muscle mitochondria, veratrine, and a component of this mixture, veratridine, caused a concentration-dependent inhibition of state III respiration and a decrease in mitochondrial membrane potential [59]. Myopathy can also be caused by environmental toxins. High exposures to organophosphate insecticides (e.g., dimethoate) has been associated with the development of intermediate myasthenia syndrome. Interestingly, dimethoate was shown to alter energy metabolism, including effects on sodium, potassium, and calcium ATPase in the mitochondria and in the cytoplasm of primary rat muscle cells [60]. Thus, a number of structurally diverse myotoxic agents appear to exert their effects through mechanisms that impair mitochondria. This is not surprising given the critical role that mitochondria play in skeletal muscle biology.

3. CONCLUSIONS There is ample evidence that a number of diverse classes of drugs and other xenobiotics cause muscle toxicity by disrupting mitochondrial function. The clearest examples are the NRTIs, where there is compelling evidence showing that these drugs inhibit mitochondrial DNA synthesis, leading to mitochondrial depletion, and ultimately, to the death of muscle ﬁbers. With statins and ﬁbrates the connections between mitochondrial failure and tissue injury are less clear, and in some cases the data are contradictory. For example, acute high-dose treatment with these drugs can cause mitochondrial uncoupling and induce MPT. However, using more relevant pharmacological concentrations, we found no effect on mitochondrial parameters, with the exception of possible cytochrome c release in the case of statins. The question remains whether long-term exposure at pharmacological concentrations over several months or years, or some other pharmacokinetic factor, would result in an accumulation of mitochondrial damage which over time could cause myopathy. For example, inhibition of the monocarboxylate transporter isoform 4 (MCT-4) prevents pathogenic Ca2+ dysregulation induced by simvastatin in intact rat skeletal muscle ﬁbers [9]. MCT-4 is localized to anaerobically poised fast-twitch ﬁbers, which deteriorate preferentially during statin-induced rhabdomyolysis, whereas the myocardium and slow-twitch skeletal ﬁbers that are spared express MCT-1, an isoform with poor afﬁnity for the statins [9,44,61]. In this way, statins may bioaccumulate to concentrations higher than those in the serum. Regardless, muscle ﬁber type selectivity between statins and ﬁbrates, plus our data showing that these drugs kill different muscle cells via distinct pathways, strongly suggest that combination therapy is unlikely to result in synergistically enhanced myopathy. However, the mixed ﬁber types seen in human skeletal muscle compared to rat could result in some muscles being more severely affected by additive toxicity responses. It is also unclear whether statins and ﬁbrates might cause mitochondrial dysfunction in certain individuals genetically predisposed to myopathy, especially

246

SKELETAL MUSCLE AND MITOCHONDRIAL TOXICITY

given the well-established links between congenital myopathies and mutations in mtDNA. These issues await future research, but it is increasingly clear that the risk–beneﬁt ratios of these important drugs, dosed alone or in combination, need to consider the possibility of untoward collateral impairment of mitochondrial function. REFERENCES 1. Phillips PS, Haas RH, Bannykh S, et al. Statin-associated myopathy with normal creatine kinase levels. Ann Intern Med . 2002;137(7): 581–585. 2. Dayer-Berenson L. Rhabdomyolysis: a comprehensive guide. Anna J . 1994;21(1): 15–18. 3. Davidson MH. Safety proﬁles for the HMG-CoA reductase inhibitors: treatment and trust. Drugs. 2001;61(2): 197–206. 4. White CM. An evaluation of CYP3A4 drug interactions with HMG-CoA reductase inhibitors. Formulary. 2000;35(4): 343+. 5. Cornelio F, Di Donato S. Myopathies due to enzyme deﬁciencies. J Neurol . 1985;232(6): 329–340. 6. Wallace DC. Mitochondrial defects in cardiomyopathy and neuromuscular disease. Am Heart J . 2000;139(2 suppl 2): S70–S85. 7. Kaufmann P, Torok M, Zahno A, Waldhauser KM, Brecht K, Krahenbuhl S. Toxicity of statins on rat skeletal muscle mitochondria. Cell Mol Life Sci . 2006;63(19–20): 2415–2425. 8. Velho JA, Okanobo H, Degasperi GR, et al. Statins induce calcium-dependent mitochondrial permeability transition. Toxicology. 2006;219(1–3): 124–132. 9. Sirvent P, Mercier J, Vassort G, Lacampagne A. Simvastatin triggers mitochondriainduced Ca2+ signaling alteration in skeletal muscle. Biochem Biophys Res Commun. 2005;329(3): 1067–1075. 10. Seachrist JL, Loi CM, Evans MG, Criswell KA, Rothwell CE. Roles of exercise and pharmacokinetics in cerivastatin-induced skeletal muscle toxicity. Toxicol Sci . 2005;88(2): 551–561. 11. Bliznakov EG, Wilkins DJ. Biochemical and clinical consequences of inhibiting coenzyme Q(10) biosynthesis by lipid-lowering HMG-CoA reductase inhibitors (statins): a critical overview. Adv Ther. 1998;15(4): 218–228. 12. Folkers K, Langsjoen P, Willis R, et al. Lovastatin decreases coenzyme Q levels in humans. Proc Natl Acad Sci U S A. 1990;87(22): 8931–8934. 13. Laaksonen R, Jokelainen K, Sahi T, Tikkanen MJ, Himberg JJ. Decreases in serum ubiquinone concentrations do not result in reduced levels in muscle tissue during short-term simvastatin treatment in humans. Clin Pharmacol Ther. 1995;57(1): 62–66. 14. Willis RA, Folkers K, Tucker JL, Ye CQ, Xia LJ, Tamagawa H. Lovastatin decreases coenzyme Q levels in rats. Proc Natl Acad Sci U S A. 1990;87(22): 8928–8930. 15. Johnson TE, Zhang XH, Bleicher KB, et al. Statins induce apoptosis in rat and human myotube cultures by inhibiting protein geranylgeranylation but not ubiquinone. Toxicol Appl Pharmacol . 2004;200(3): 237–250.

REFERENCES

247

16. Dirks AJ, Jones KM. Statin-induced apoptosis and skeletal myopathy. Am J Physiol Cell Physiol . 2006;291(6): C1208–C1212. 17. Schaefer WH, Lawrence JW, Loughlin AF, et al. Evaluation of ubiquinone concentration and mitochondrial function relative to cerivastatin-induced skeletal myopathy in rats. Toxicol Appl Pharmacol . 2004;194(1): 10–23. 18. Hodel C. Myopathy and rhabdomyolysis with lipid-lowering drugs. Toxicol Lett . 2002;128(1–3): 159–168. 19. Ballantyne CM, Corsini A, Davidson MH, et al. Risk for myopathy with statin therapy in high-risk patients. Arch Intern Medi . 2003;163(5): 553–564. 20. Willson TM, Brown PJ, Sternbach DD, Henke BR. The PPARs: from orphan receptors to drug discovery. J Medi Chem. 2000;43(4): 527–550. 21. Dressel U, Allen TL, Pippal JB, Rohde PR, Lau P, Muscat GEO. The peroxisome proliferator-activated receptor beta/delta agonist, GW501516, regulates the expression of genes involved in lipid catabolism and energy uncoupling in skeletal muscle cells. Mol Endocrinol . 2003;17(12): 2477–2493. 22. Son C, Hosoda K, Matsuda J, et al. Up-regulation of uncoupling protein 3 gene expression by fatty acids and agonists for PPARs in L6 myotubes. Endocrinology. 2001;142(10): 4189–4194. 23. Tanaka T, Yamamoto J, Iwasaki S, et al. Activation of peroxisome proliferatoractivated receptor delta induces fatty acid beta-oxidation in skeletal muscle and attenuates metabolic syndrome. Proc Natl Acad Sci U S A. 2003;100(26): 15924–15929. 24. Scatena R, Martorana GE, Bottoni P, Giardina B. Mitochondrial dysfunction by synthetic ligands of peroxisome proliferator activated receptors (PPARs). IUBMB Life. 2004;56(8): 477–482. 25. Brunmair B, Lest A, Staniek K, et al. Fenoﬁbrate impairs rat mitochondrial function by inhibition of respiratory complex I. J Pharmacol Exp Ther . 2004;311(1): 109–114. 26. Keller BJ, Bradford BU, Marsman DS, et al. The nongenotoxic hepatocarcinogen Wy-14,643 is an uncoupler of oxidative-phosphorylation in vivo. Toxicol Appl Pharmacol . 1993;119(1): 52–58. 27. Masubuchi Y, Suda C, Horie T. Involvement of mitochondrial permeability transition in acetaminophen-induced liver injury in mice. J Hepatol . 2005;42(1): 110–116. 28. Scatena R, Bottoni P, Martorana GE, et al. Mitochondrial respiratory chain dysfunction, a non-receptor-mediated effect of synthetic PPAR-ligands: biochemical and pharmacological implications. Biochem Biophys Res Commun. 2004;319(3): 967–973. 29. Johnson TE, Zhang XH, Shi S, Umbenhauer DR. Statins and PPAR alpha agonists induce myotoxicity in differentiated rat skeletal muscle cultures but do not exhibit synergy with co-treatment. Toxicol Appl Pharmacol . 2005;208(3): 210–221. 30. Athyros VG, Papageorgiou AA, Hatzikonstandinou HA, et al. Safety and efﬁcacy of long-term statin–ﬁbrate combinations in patients with refractory familial combined hyperlipidemia. Am J Cardiol . 1997;80(5): 608–613. 31. Gavish D, Leibovitz E, Shapira I, Rubinstein A. Bezaﬁbrate and simvastatin combination therapy for diabetic dyslipidaemia: efﬁcacy and safety. J Intern Med . 2000;247(5): 563–569. 32. Shek A, Ferrill MJ. Statin–ﬁbrate combination therapy. Ann Pharmacother . 2001;35(7–8): 908–917. 33. Evans M, Rees A. Effects of HMG-CoA reductase inhibitors on skeletal muscle: Are all statins the same? Drug Saf . 2002;25(9): 649–663.

248

SKELETAL MUSCLE AND MITOCHONDRIAL TOXICITY

34. Muscari A, Puddu GM, Puddu P. Lipid-lowering drugs: Are adverse effects predictable and reversible? Cardiology. 2002;97(3): 115–121. 35. Matzno S, Tazuya-Murayama K, Tanaka H, et al. Evaluation of the synergistic adverse effects of concomitant therapy with statins and ﬁbrates on rhabdomyolysis. J Pharm Pharmacol . 2003;55(6): 795–802. 36. Inoue I, Itoh F, Aoyagi S, et al. Fibrate and statin synergistically increase the transcriptional activities of PPAR alpha/RXR alpha and decrease the transactivation of NF kappa B. Biochem Biophys Res Commun. 2002;290(1): 131–139. 37. Martin G, Duez H, Blanquart C, et al. Statin-induced inhibition of the rho-signaling pathway activates PPAR alpha and induces HDL apoA-I. J Clin Invest. 2001;107(11): 1423–1432. 38. Motojima K, Seto K. Fibrates and statins rapidly and synergistically induce pyruvate dehydrogenase kinase 4 mRNA in the liver and muscles of mice. Biol Pharm Bull 2003;26(7): 954–958. 39. Ruiz-Velasco N, Dominguez A, Vega MA. Statins upregulate CD36 expression in human monocytes, an effect strengthened when combined with PPAR-gamma ligands: putative contribution of rho GTPases in statin-induced CD36 expression. Biochem Pharmacol . 2004;67(2): 303–313. 40. Johnson TE, Ledwith BJ. Peroxisome proliferators and fatty acids negatively regulate liver X receptor–mediated activity and sterol biosynthesis. J Steroid Biochem Mol Biol . 2001;77(1): 59–71. 41. Backman JT, Kyrklund C, Kivisto KT, Wang JS, Neuvonen PJ. Plasma concentrations of active simvastatin acid are increased by gemﬁbrozil. Clin Pharmacol Ther. 2000;68(2): 122–129. 42. Prueksaritanont T, Subramanian R, Fang XJ, et al. Glucuronidation of statins in animals and humans: a novel mechanism of statin lactonization. Drug Metab Dispos. 2002;30(5): 505–512. 43. Smith PF, Eydelloth RS, Grossman SJ, et al. HMG-CoA reductase inhibitor-induced myopathy in the rat: cyclosporine A interaction and mechanism studies. J Pharmacol Exp Ther. 1991;257(3): 1225–1235. 44. Westwood FR, Bigley A, Randall K, Marsden AM, Scott RC. Statin-induced muscle necrosis in the rat: distribution, development, and ﬁbre selectivity. Toxicol Pathol . 2005;33(2): 246–257. 45. De Souza AT, Cornwell PD, Dai XD, Caguyong MJ, Ulrich RG. Agonists of the peroxisome proliferator-activated receptor alpha induce a ﬁber-type-selective transcriptional response in rat skeletal muscle. Toxicol Sci . 2006;92(2): 578–586. 46. McComsey G, Lonergan JT. Mitochondrial dysfunction: patient monitoring and toxicity management. J Acquir Immune Deﬁc Syndr. 2004;37:S30-S35. 47. Hoschele D. Cell culture models for the investigation of NRTI-induced mitochondrial toxicity: relevance for the prediction of clinical toxicity. Toxicol in Vitro. 2006;20(5): 535–546. 48. Walker UA. Update on mitochondrial toxicity: Where are we now? J HIV Ther . 2003;8(2): 32–35. 49. Perez-Perez MJ, Hernandez AI, Priego EM, et al. Mitochondrial thymidine kinase inhibitors. Curr Top Medi Chem. 2005;5(13): 1205–1219.

REFERENCES

249

50. Chen CH, Vazquezpadua M, Cheng YC. Effect of anti-human-immunodeﬁciency-virus nucleoside analogs on mitochondrial-DNA and its implication for delayed toxicity. Mol Pharmacol . 1991;39(5): 625–628. 51. Medina DJ, Tsai CH, Hsiung GD, Cheng YC. Comparison of mitochondrial morphology, mitochondrial-DNA content, and cell viability in cultured-cells treated with 3 anti-human-immunodeﬁciency-virus deoxynucleosides. Antimicrobi Agents Chemother. 1994;38(8): 1824–1828. 52. Martin JL, Brown CE, Matthewsdavis N, Reardon JE. Effects of antiviral nucleoside analogs on human DNA-polymerases and mitochondrial-DNA synthesis. Antimicrob Agents Chemother. 1994;38(12): 2743–2749. 53. Benbrik E, Chariot P, Bonavaud S, et al. Cellular and mitochondrial toxicity of zidovudine (AZT), didanosine (ddI) and zalcitabine (ddC) on cultured human muscle cells. J Neurol Sci . 1997;149(1): 19–25. 54. Birkus G, Hitchcock MJM, Cihlar T. Assessment of mitochondrial toxicity in human cells treated with tenofovir: comparison with other nucleoside reverse transcriptase inhibitors. Antimicrob Agents Chemother . 2002;46(3): 716–723. 55. Collins ML, Sondel N, Cesar D, Hellerstein MK. Effect of nucleoside reverse transcriptase inhibitors on mitochondrial DNA synthesis in rats and humans. J Acquir Immune Deﬁc Syndr. 2004;37(1): 1132–1139. 56. Haugaard SB, Andersen O, Pedersen SB, et al. Depleted skeletal muscle mitochondrial DNA, hyperlactatemia, and decreased oxidative capacity in HIV-infected patients on highly active antiretroviral therapy. J Med Virol . 2005;77(1): 29–38. 57. Irwin W, Fontaine E, Agnolucci L, et al. Bupivacaine myotoxicity is mediated by mitochondria. J Biol Chem. 2002;277(14): 12221–12227. 58. Rusyniak DE, Tandy SL, Hekmatyar SK, et al. The role of mitochondrial uncoupling in 3,4-methylenedioxymethamphetamine–mediated skeletal muscle hyperthermia and rhabdomyolysis. J Pharmacol Exp Ther . 2005;313(2): 629–639. 59. Freitas EMS, Fagian MM, Hoﬂing MAD. Effects of veratrine and veratridine on oxygen consumption and electrical membrane potential of isolated rat skeletal muscle and liver mitochondria. Toxicon. 2006;47(7): 780–787. 60. Yang DR, Lu XF, Zhang WG, He FS. Biochemical changes in primary culture of skeletal muscle cells following dimethoate exposure. Toxicology 2002;174(2): 79–85. 61. Halestrap AP, Price NT. The proton-linked monocarboxylate transporter (MCT) family: structure, function and regulation. Biochem J . 1999;343:281–299.

8 MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA IN THE NERVOUS SYSTEM Ian J. Reynolds Neuroscience Drug Discovery, Merck Research Laboratories, West Point, Pennsylvania

1. Introduction 2. Mitochondrial mechanisms of peripheral neuropathy 2.1. Reverse transcriptase inhibitors 2.2. Microtubule-modifying agents and mitochondria 2.3. Statins and peripheral neuropathy 3. Mitochondria and retinal drug toxicity 3.1. Chlorophenicol 3.2. Ethambutol 3.3. Methanol 4. Mitochondria and ototoxicity 5. Mitochondrial mechanisms of CNS injury 5.1. Mitochondrial mechanisms of neuronal injury 5.2. Potential manifestations of drug-induced mitochondrial dysfunction in the CNS 6. Conclusions

251 253 253 255 256 257 257 258 258 259 261 261 263 264

1. INTRODUCTION Although mitochondria are critical for the function of most tissues, they are particularly important for the maintenance and integrity of the nervous system. Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

251

252

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

It is typically estimated that whereas the central nervous system (CNS) accounts for approximately 2% of body mass, the CNS accounts for some 20% of oxygen utilization, the vast majority of which is consumed by oxidative phosphorylation [1]. In relative terms, brain mitochondria are also more active than mitochondria in other tissues [2]. The potential for injury as a consequence of inhibition of mitochondrial function in neurons is illustrated by a number of different examples of disease states that can be modeled using electron transport chain inhibitors, including Parkinson’s and Huntington’s disease [3,4], and the links between impaired mitochondrial function and CNS disease are further emphasized by the prominently neurological phenotype of syndromes associated with such mtDNA mutations as Leber’s hereditary optic neuropathy (LHON) and mitochondrial encephalopathy, lactic acidosis, and stroke (MELAS; [5]). These disorders also illustrate another important principle, which is that uniformly applied mitochondrial inhibition can result in perhaps surprisingly speciﬁc injury phenotypes in the CNS. For example, systemic administration of low doses of rotenone produces a Parkinson-like disorder associated with the loss of dopaminergic neurons, even though complex I is inhibited to a similar extent across the brain [3], while similar utilization of the complex II inhibitor 3-nitroproprionic acid produces striatal injury resembling Huntington’s disease [4]. The basis for this selective vulnerability remains unclear, but the examples illustrate the potentially profound consequences of the toxic actions of drugs on mitochondrial function. The discussion that follows will highlight speciﬁc examples of drug toxicity associated with altered mitochondrial function in neurons. However, it may also be informative to consider some of the special circumstances of neurons that may be affected when mitochondria are impaired. More than half of the ATP generated by oxidative phosphorylation in the brain is used for maintaining ionic gradients necessary for electrical signaling of neurons and glia, principally the Na+ /K+ ATPase [6]. Under circumstances of signiﬁcant impairment of ATP generation resulting from drug toxicity, altered electrical signaling (e.g., seizure activity) could result. This is not a common ﬁnding as a mechanism of toxicity of approved drugs, perhaps because it is a gross form of toxicity that would have been readily apparent preclinically. A more subtle form of injury would be associated with the generation of reactive oxygen species (ROS) by brain mitochondria. This can occur with relatively low levels of impairment of complex I (e.g., [7]) that are not associated with profound decreases in oxygen consumption. Progressive accumulation of oxidant damage in the lipid-rich environment of the nervous system is associated with many neurodegenerative diseases [8]. This form of injury would be more difﬁcult to detect, especially at threshold levels of inhibition, as discussed further below. A unique challenge to neurons is the targeting, delivery, and retrieval of mitochondria to appropriate locations within neurons. Mitochondria are highly motile organelles and move at speeds up to 2 to 3 µm/s in neuronal processes in culture [9]. It is generally assumed that new mitochondria are generated at locations remote from the site within the neuron where the mitochondria will be deployed. As the longest neurons in a human exceed 1 m in length, the distance that a

MITOCHONDRIAL MECHANISMS OF PERIPHERAL NEUROPATHY

253

mitochondrion must travel is potentially considerable. It has also been proposed that damaged mitochondria are retrieved to the cell body prior to disposal such that retrograde trafﬁcking is also an important consideration [10]. Trafﬁcking of mitochondria is an energy-dependent process, and a number of neurotoxins that impair ATP generation also inhibit mitochondrial movement [9]. Trafﬁcking also depends on the integrity of the cytoskeleton and is modulated by a number of other signals, including growth factors and neurotoxins that do not impair ATP synthesis [11–13]. If it is reasonable to assume that impaired mitochondrial trafﬁcking will disrupt neuronal function, trafﬁcking is an important potential target for drug toxicity. This is discussed further below. While understanding the nature of drug interactions with mitochondria can provide the basis for insights into potential mechanisms of toxicity, it remains difﬁcult to predict the manifestations of that toxicity in the intact organism. Typically, adverse effects of drugs that are toxic to nervous system mitochondria are reported as alterations in the function of a sensory or motor modality. Some of the best established examples of mitochondrial toxicity are manifested as peripheral neuropathy, for example. In this chapter we also consider examples of drugs that produce retinal toxicity, or ototoxicity, and have the potential for effects in the CNS. It is not always clear why selectivity exists for a particular subset of neurons. In some cases the presence of the blood–brain barrier or effective drug efﬂux pathways can protect central neurons from drug exposures that would otherwise prove toxic and that could account for a preferential toxic action on peripheral versus central neurons. However, this is clearly not an issue when CNS penetration can be established clearly. Whether differential toxicity results from differential exposure, distinct properties of mitochondria within different types of neurons, or increased vulnerability of neurons arising from a speciﬁc feature of the interaction between mitochondria and their cellular environment is often unclear. For this reason, the strategy taken in this chapter is to review toxicities based on phenotype and target tissues and then to assess the evidence for a mitochondrial mechanism for the toxicity, as opposed to identifying drugs that are known to affect mitochondrial function at some concentration and then searching for evidence for mitochondrially mediated injury. Given the considerable range of CNS active compounds for which mitochondrial activity has been documented [14], the potential for nervous system injury remains substantial. We conclude with a discussion of the potential for CNS injury in the context of already established neurotoxic mechanisms.

2. MITOCHONDRIAL MECHANISMS OF PERIPHERAL NEUROPATHY 2.1. Reverse Transcriptase Inhibitors Nucleoside and nonnucleoside reverse transcriptase inhibitors (NNRTIs) are effective antiviral agents that are used in the treatment of herpes virus, hepatitis

254

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

B virus and, importantly, human immunodeﬁciency virus. While generally considered safe, extending the duration of therapy with this diverse group of drugs reveals a collection of toxicities in various tissues that include peripheral neuropathy, skeletal and cardiac myopathy, lactic acidosis, lipoatrophy and hepatic failure. While the therapeutic target for these agents is viral DNA replication, it is generally believed that the diverse collection of toxicities arises from interaction with mammalian DNA polymerases, and DNA polymerase γ in particular [15]. The mechanisms of NNRTI toxicity and the mitochondrial basis for these effects is covered in Chapter 9. Here the focus is on the peripheral neuropathy associated with a subset of these drugs. Peripheral neuropathy is a characteristic of untreated HIV infections in addition to NNRTI inhibitors. The neuropathy is characterized by symmetrical distal parasthesias that are commonly painful. These effects are typically in the lower limbs and may progress centrally over time. The key distinguishing features of neuropathy associated with drug treatment rather than the disease itself are a more abrupt onset of the neuropathy and also an onset associated with the initiation of therapy [16]. Onset is typically delayed from the start of treatment and occurs after two weeks or, more typically, longer periods of drug exposure. The neuropathy is also reversible after drug removal, at least to a limited extent. The potential for NNRTIs to cause neuropathy was ﬁrst encountered with vidarabine in the treatment of hepatitis B, and has subsequently been associated with the use of zalcitabine, didanosine, and stavudine (reviewed by [16,17]). The ability of these agents to trigger neuropathies is ostensibly associated with their ability to inhibit mtDNA replication in lymphoblastoid cells [18]. The delay in onset of neuropathy presumably then reﬂects the time required to deplete mtDNA to an extent that cell function is impaired. The basis for the selective effect that results in neuropathy versus the other forms of toxicity associated with NNRTIs is not clear, although there are a number of potential explanations. The action of NNRTIs depends on cell accumulation, transport into the mitochondria, phosphorylation by thymidine kinase, and then ultimately either inhibition of polymerase γ or incorporation into mtDNA, where the abnormal nucleoside structure results in chain termination [17]. It is also clear that deoxyribonuleotides synthesized in the cytoplasm can be transported into mitochondria [19]. At the site of action, the phosphorylated drugs compete with endogenous deoxyribonucleotide pools for incorporation into mtDNA. This allows the possibility that cell types have differential pools of mtDNA precursors that would result in varying sensitivity to inhibition [19]. It is not clear whether neurons or Schwann cells are the target for injury caused by these agents. However, Anderson and colleagues [20] showed that dideoxycytidine administration to rabbits produced a neuropathy that was associated with dysfunctional mitochondria in Schwann cells but not neurons. This suggests that disruption of neuronal support cells is critical to the neuropathy produced by NNRTIs. The situation in patients can be complicated by the presence of other factors that predispose to the development of neuropathy, which includes alcohol toxicity, diabetes, dietary deﬁciencies, and a host of other pharmacological agents [16]. The presence of

MITOCHONDRIAL MECHANISMS OF PERIPHERAL NEUROPATHY

255

these other risk factors may effectively amplify the burdens imposed by NNRTIs, although it is not clear that they share the same mechanism of action. 2.2. Microtubule-Modifying Agents and Mitochondria Peripheral neuropathy is a common adverse effect of treatment with antineoplastic agents. In particular, paclitaxel, cisplatin, vincristine, and suramin commonly produce sensory neuropathies and, less often, motor symptoms [21,22]. Several of these agents owe their primary mechanism of action to either destabilizing (vincristine) or stabilizing (paclitaxel) microtubules. Microtubules have an essential role in trafﬁcking of mitochondria in neurons [23], so this raises the possibility that the neuropathy associated with microtubule modifying drugs results from an impact on mitochondrial function. A number of studies have demonstrated mitochondrial effects of paclitaxel, colchicine, and vincristine, although not all of these are necessarily associated directly with altered mitochondrial trafﬁcking. For example, several studies have shown that microtubule agents alter cellular calcium homeostasis mediated by mitochondria. This could be the consequence of either opening or alternatively delaying the closure of the mitochondrial permeability transition pore (mPTP) in neurons [24,25]. Mironov and colleagues [26] suggested that activation of mPTP by paclitaxel was inﬂuenced by the association of microtubules, endoplasmic reticulum, and mitochondria, and that the drug-induced disruption of this relationship resulted in mPTP activation. In central neurons, vinblastine causes the redistribution of mitochondria in neuronal cell bodies, and this is associated with impaired mitochondrial calcium homeostasis but not with altered bioenergetic function [27]. Paclitaxel also triggers the release of cytochrome c from mitochondria isolated from neuroblastoma cells [28], an effect that could be the consequence of calcium-induced mPTP activation. The suggestion that mPTP is involved in these phenomena is supported by the observation that cyclosporine A effectively reverses several of the effects on mitochondrial calcium homeostasis and function triggered by these drugs [24–26,28]. It is noteworthy that recent studies have suggested a preferential effect of submicromolar concentrations of paclitaxel on the IP3 receptor from brain, so the conclusion that modiﬁed cellular calcium handling must be mediated by an impact on mitochondria should be approached cautiously. The neuropathy caused by paclitaxel and vincristine can effectively be recapitulated in animal models. Interestingly, in a rat model of paclitaxel toxicity, hyperalgesia following four doses of the drug was associated with an increase in the frequency of swollen mitochondria, while no evidence was found for gross axonal degeneration [29]. This would be consistent with activation of mPTP, although it is difﬁcult to establish the sequence of events that results in altered mitochondrial morphology in experiments of this nature. The involvement of calcium in this pathology is suggested by the ﬁnding that intrathecal administration of calcium chelators blocked the expression of hyperalgesia and mechanoallodynia in rats treated with paclitaxel [30]. In contrast, studies with vincristine provided no evidence for altered mitochondrial morphology [31].

256

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

The interference with axonal transport that should arise from disruption of microtubules in the extremely long axons of sensory and motor nerves ought to provide a clear example of the impact of interfering with mitochondrial trafﬁcking on neuronal function. The studies reported here document a number of different effects on mitochondria that include altered calcium handling and morphological changes that appear to produce results that are consistent between in vitro and in in vivo experiments. However, neither altered morphology nor differences in calcium handling are obvious sequelae of a selective effect on inhibition of mitochondrial trafﬁcking. Whether altered trafﬁcking contributes to the neuropathy associated with microtubule disrupting drugs remains to be determined. 2.3. Statins and Peripheral Neuropathy Hydroxymethylglutaryl (HMG)-coenzyme A (CoA) reductase inhibitors, more commonly known as statins, have been reported to have a number of adverse effects on the sensorimotor axis. Although these effects are relatively uncommon (less than 1% of patients), the widespread use of statins in cholesterol management makes this a potentially important issue. Statins have been associated with skeletal muscle myopathy [32], which is discussed in Chapter 7. There is evidence that this myopathy results from myocyte apoptosis mediated by the mitochondrial pathway, but that the apoptosis is triggered by inhibition of geranylgeranyl transferase inhibition [33], which makes a rather indirect case for a mitochondrial target for the adverse effects of statins. Separate from the effects on skeletal muscle, there is clearly evidence of neuropathy induced by statins [34,35]. This has been associated with most of the widely used statins, including lovastatin, simvastatin, ﬂuvastatin, and rosuvastatin [34]. Mechanistically, the basis for the neuropathy remains unclear, although two interesting links to mitochondrial function have been documented. First, lovastatin and simvastatin inhibit mitochondrial respiration in isolated liver mitochondria by a direct effect on electron transport at complex II/III, complex IV, and/or complex V [36]. In addition, Kaufmann and colleagues [37] reported that a range of statins decreased mitochondrial membrane potential, inhibited β-oxidation, and triggered swelling of isolated skeletal muscle mitochondria. If such effects are recapitulated in neuronal mitochondria, the resulting impairment in mitochondrial function could contribute to neuropathic injury. Second, it has been reported that statins deplete the ubiquinone coenzyme Q10 (CoQ10 ) [38], although not all studies have been able to detect this signal [39]. This is presumably the consequence of the inhibition of the formation of mevalonic acid, which is a key product of HMG-CoA reductase, because mevalonic acid is a precursor of CoQ10 . CoQ10 serves to shuttle electrons from complexes I and II in the electron transport chain to complex III, and may also act as an antioxidant. Although it is not clear that CoQ10 is depleted sufﬁciently to impair electron transport, it has been reported that CoQ10 supplementation is beneﬁcial in Parkinson’s disease [40]. It is an obvious extrapolation to suggest that CoQ10 depletion could contribute to mitochondrial dysfunction that would contribute to statin-induced neuropathy.

MITOCHONDRIA AND RETINAL DRUG TOXICITY

257

No evidence has emerged to suggest that statins increase the risk or severity of Parkinson’s disease [41], but this type of mechanism would result in a fairly subtle form of dysfunction that would be quite difﬁcult to detect, yet could be an important cause of adverse effects. 3. MITOCHONDRIA AND RETINAL DRUG TOXICITY Toxic actions of drugs in the eye may have a number of distinct manifestations, and many different effects have been documented (reviewed in [42]). The focus of this section will be on injury to the retina and optic nerve. The link between drug-induced mitochondrial dysfunction and retinal damage is anticipated by ﬁndings of inherited diseases associated with mitochondrial genes and optic neuropathy. For example, LHON results from maternally inherited mutations in one of several mitochondrial genes and results in the loss of retinal ganglion cells and thinning of the nerve ﬁber layer [43]. Vision loss is also associated with Kjer disease and has been attributed to mutations in the protein OPA1, which regulates mitochondrial morphology and ﬁssion [44], as well as Leigh’s syndrome, which arises from mutations in one of several mitochondrial or nuclear-encoded genes [43]. These observations from inherited diseases validate the concept that retinal injury can result from impairment of mitochondrial function and provide a framework for understanding the effects of xenobiotics. The potential impact of toxins on the retina is compounded by the fact that the retina is very active in metabolic terms and has a low energy reserve under normal conditions [45]. In this section the actions of three well-known retinal toxins are discussed, along with the evidence for a mitochondrial target for the toxicity. 3.1. Chlorophenicol This widely used antibiotic owes its mechanism of action to inhibition of bacterial protein synthesis. Similar to gentamycin, discussed in the next section, it is widely believed that the toxicity associated with chloramphenicol arises from inhibition of mitochondrial protein synthesis because of the similarity in structure of bacterial and mitochondrial ribosomes [46,47]. A number of adverse consequences arise from the clinical use of chloramphenicol, including bone marrow suppression. Notably, though, is the optic neuropathy that was documented following the treatment of children with cystic ﬁbrosis [48]. The presumption is that this neuropathy is due to the inhibition of mitochondrial protein synthesis, although this is obviously difﬁcult to document directly in patients. In experimental models involving the study of neuronal function, additional evidence emerges suggesting an interaction between chloramphenicol and mitochondrial function. For example, injury to the auditory nerve induced by high noise levels or gentamycin produces an injury that is associated with (and perhaps ameliorated by) increased biogenesis of mitochondria. Treatment of animals with chloramphenicol exacerbates this form of injury [49–51]. Chloramphenicol also decreases glucose utilization and inhibits NADH oxidation in rat brain [52,53]. Mitochondrial

258

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

distribution in the optic nerve is driven by the subcellular metabolic needs of the neuron, such that mitochondria tend to cluster around nonmyelinated sections of the axon where the density of sodium channels is greatest (and thus the need for ion pumping) [54]. As trafﬁcking of mitochondria is an energy-dependent process, the effects of disruption of ATP generation could be at least partially caused by the inability to deliver mitochondria to appropriate cellular locations, thus contributing to the neuropathy [55]. 3.2. Ethambutol Ethambutol remains a drug of choice for the treatment of tuberculosis, particularly with the emergence of drug-resistant forms of this disease. It is generally considered that optic neuropathy is the dose-limiting toxicity with ethambutol treatment [56,57]. The mechanism by which ethambutol produces a selective effect on optic nerve function is not clear, but studies have suggested a speciﬁc involvement of mitochondria. Thus, ethambutol and its metabolites effectively bind copper, even though copper binding does not appear to be associated with the mechanism of action of this drug [58,59]. Separately, it has been shown that brain copper deﬁciency results in mitochondrial dysfunction mediated by a decrease in the activity of complex IV (cytochrome c oxidase) of the electron transport chain [60]. This is a reﬂection of impaired assembly of the copper-dependent cytochrome oxidase complex. The toxicity of ethambutol can be recapitulated in rat models, where drug exposure results in a selective loss of retinal ganglion cells in the eye [61]. This study also concluded that excitotoxic activation of NMDA receptors played a key role in ethambutol-mediated injury, because memantine protected against injury. Ampliﬁcation of excitotoxic injury by impaired mitochondrial function is a well-recognized phenomenon and reﬂects the inability of neurons to increase energy production to meet increased metabolic demand imposed by glutamate receptor activation [62]. Yoon and colleagues [63] have proposed a distinct mechanism of action for ethambutol based on their observations of zinc dependence of ethambutol injury to retinal ganglion cells in culture. Thus, unlike the metal depletion mechanism of injury, these investigators suggest that ethambutol injury is associated with enhanced zinc accumulation in neurons, perhaps indicating a mechanism of facilitated zinc entry into neurons as is observed with agents such as pyrithione [64]. Neurons are highly sensitive to injury by zinc, and interestingly, the trafﬁcking of mitochondria in neurons is highly sensitive to inhibition by zinc [12]. Thus, the proposed effects of ethambutol on mitochondrial function recognize several different mechanisms that could contribute to neuronal injury. 3.3. Methanol There is a long history of methanol toxicity that arises from its use as an industrial solvent, as a fuel source, and as a contaminant in ethanol-containing beverages (reviewed in [65]). Characteristically, after a short period of CNS depression there is a latent period of 12 to 24 hours before the acute toxic effects of methanol

MITOCHONDRIA AND OTOTOXICITY

259

appear [66]. The toxicity arises as a result of metabolism of methanol to formic acid and the accumulation of this metabolite. Formic acid, in turn, is an effective inhibitor of complex IV of the electron transport chain, so that toxicity is associated with acute metabolic failure [67]. Formate is not a particularly potent inhibitor of cytochrome oxidase, with K i values reported between 1 and 30 mM, depending on the oxidation state of the cytochrome [67]. However, millimolar concentrations of alcohols in plasma are typical following normal consumption. The failure to oxidize formic acid, is a critical element in the toxicity of methanol in humans. As most animal species have the ability to detoxify formic acid, they do not normally suffer methanol toxicity. However, it is possible to model methanol toxicity in rodents by inhibiting formic acid oxidation, and this produces an injury pattern that is similar in scope and time course to that of methanol poisoning in humans [68]. In both humans and rats, the retina is an early target of the toxic effects of methanol, and the retinal ganglion cells in particular [66,68]. Interestingly, several other complex IV toxins, including cyanide and carbon monoxide, have also been characterized as producing optic neuropathies [43], suggesting that the retina may be particularly sensitive to this form of metabolic inhibition.

4. MITOCHONDRIA AND OTOTOXICITY Ototoxicity is a term that refers to injury to either the cochlea, which results in hearing impairment (HI), or to the vestibular system, which causes problems with balance and spatial orientation. A number of drugs have well-characterized ototoxic effects, including aminoglycoside antibiotics, cisplatin, erythromycin, quinine, loop diuretics, and nonsteriodal anti-inﬂammatory drugs, including aspirin and naproxen [69]. In general, although the associations between the use of these drugs and ototoxic adverse effects are well documented, the mechanisms underlying damage to the ear are not well understood. In this section the focus will be on aminoglycoside ototoxicity, where a link to mitochondrial function has been established. The propensity of aminoglycoside antibiotics to cause HI is well recognized. This is associated with the use of streptomycin, neomycin, and kanamycin, as well as the newer agents gentamicin, tobramycin, and amikacin [69]. These drugs produce antibacterial effects by interacting with bacterial ribosomes and inhibiting bacterial protein synthesis, and it has been suggested that mitochondrial toxicity could arise as the result of aminoglycoside interactions with mitochondrial ribosomes and the subsequent impairment of mitochondrial protein synthesis [70] (see Chapter 20). Injury to the inner ear with aminoglycosides occurs with exposure levels that are within the therapeutic range, and affect approximately 10% of persons who are administered the drugs. These agents damage both the cochlea, where the main manifestation is a loss of sensitivity to high-frequency sound, and the vestibular system, although the balance of cochlear to vestibular injury varies between drugs [69,71]. The HI is a consequence of damage ﬁrst to the

260

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

outer hair cells, which appear to be most sensitive, followed by the inner hair cells. Selectivity of injury to the inner ear over other tissues may arise from the accumulation and slow clearance of these drugs from the ear after extended treatment [71]. The mechanism by which aminoglycosides produce injury is not fully resolved. However, it has been suggested that gentamycin can chelate iron and that the iron–aminoglycoside complex is redox active such that an excess of superoxide is produced [72]. Consistent with this hypothesis, certain antioxidants protect against aminoglycoside toxicity in vivo, and iron chelators such as desferroxamine are particularly effective [73]. An alternative hypothesis suggests that aminoglycosides promote activation of the NMDA receptor via stimulation of the polyamine site, a proposal supported by the observation that concurrent treatment with NMDA receptor antagonists limits aminoglycoside ototoxicity [74]. Notably, this mechanism would also involve the excess production of oxidants following toxic NMDA receptor activation and mitochondrial calcium accumulation [75,76] such that oxidative stress is a common feature of the two mechanisms proposed. The function of the inner ear is particularly sensitive to impairment of mitochondrial function. This is reﬂected by the relatively frequent association of maternally inherited mitochondrial dysfunction with HI, both as a part of syndromes such as MELAS and Kearns–Sayre syndrome, as well as nonsyndromic HI that follows identiﬁed mtDNA mutations [77]. Intriguingly, several studies have associated mtDNA mutations with enhanced sensitivity to aminoglycoside ototoxicity [78,79]. The ﬁrst mutation identiﬁed, 1555G , is located in a mitochondrial 12S rRNA gene, which represents a mitochondrial homolog of the aminoglycoside target in bacteria. This mutation is associated with a greatly enhanced susceptibility to aminoglycoside ototoxicity. Subsequently, several other mtDNA alterations have been identiﬁed that are located in the tRNAser(UCN) gene [77] and result in nonsyndromic deafness and increased aminoglycoside sensitivity. Although a recent study concluded that screening for 12S rRNA mutations prior to initiating therapy with aminoglycosides was not warranted based on existing data [80], the risk in patients carrying these mutations is clearly elevated substantially (see Chapter 11). These interesting ﬁndings do not unequivocally establish that aminoglycoside toxicity is based primarily on a mitochondrial mechanism, although minimally they suggest that the toxicity is ampliﬁed if mitochondrial function is altered. It has been suggested that the effect of the mtDNA mutation is decreased synthesis of mitochondrially encoded complex I subunits [70], and low levels of complex I impairment result in excess superoxide production in neural tissues [7]. Thus, if the generation of neactive oxygen species (ROS) is the principal initiator of ototoxicity, an additive oxidative burden would be imposed by concomitant mitochondrial impairment, which would amplify the injury.

MITOCHONDRIAL MECHANISMS OF CNS INJURY

261

5. MITOCHONDRIAL MECHANISMS OF CNS INJURY The preceding sections have highlighted well-established examples of drugs that produce characteristic adverse effects in the nervous system that are most likely the result of impairment of mitochondrial function. Attribution of the adverse effects to mitochondrial impairment is facilitated by the ability to associate speciﬁc and, typically, sensory deﬁcits with drug exposure and then ultimately to link pathophysiological sequelae to one or more mitochondrial targets. This association proves to be much more difﬁcult to establish in the context of potential toxic effects in the CNS that arise from mitochondrial inhibition. Although there are numerous drugs that act on the nervous system that also affect mitochondrial function (reviewed recently by Chan and colleagues [14]), it proves to be much more difﬁcult to establish explicit pathological consequences for the CNS that arise from mitochondrial inhibition. This is because the injury phenotypes would be expected to develop relatively slowly and would be manifested by altered behaviors or perhaps subtle neurological symptoms rather than alterations in sensory modalities, which are easier to detect and isolate. Rather than attempting to create speculative links between drugs and potential CNS toxicity, it is perhaps more useful to review mitochondrial mechanisms underlying neuronal injury based on known mechanisms derived from established neurotoxins. This will provide the context for understanding the potential manifestations of drug-induced injury mediated by mitochondria and will provide an appreciation of the nature of the challenge of studying this kind of toxicity. 5.1. Mitochondrial Mechanisms of Neuronal Injury There are several well-established mechanisms by which mitochondrial mechanisms mediate neuronal injury. Commonly, elevated intracellular calcium, ROS, and energy depletion serve as a lethal triumvirate, and collectively these represent the major known pathophysiologic mechanisms affecting neurons. There is also emerging evidence linking effects on mitochondrial morphology and trafﬁcking with neuronal viability, which we discuss brieﬂy. Mitochondria also serve as a focal point for the intrinsic pathways of apoptosis, where they release apoptogens. However, it would be unusual for a drug to trigger apoptogen release directly from mitochondria, as this would result in acute toxicity that would readily be detected and would serve as the basis for stopping compound development. Accordingly, this is considered beyond the scope of the present discussion. It has long been appreciated that neurons are injured by large elevations in intracellular calcium. Toxic intracellular calcium loads are imposed by the excessive activation of NMDA receptors in a process known as excitotoxicity [81,82]. The target of the elevated calcium has been the subject of much study, but it has become clear that an impairment of mitochondrial function is a key consequence of intracellular calcium loading. Mitochondria depolarize, stop moving, and change their morphology when exposed to toxic intracellular calcium concentrations [83–86], effects that appear to be largely a consequence of mitochondrial

262

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

calcium accumulation. Perhaps surprisingly, inhibition of mitochondrial calcium loading protects neurons against excitotoxic injury, even though this is associated with very high cytoplasmic calcium concentrations and the ongoing generation of potentially toxic mediators such as nitric oxide [76,87,88]. The precise mechanisms that cause mitochondrial depolarization are not clear. Calcium cycling by mitochondria comes at the expense of the proton gradient, and this could certainly contribute to the depolarization [89]. Glutamate stimulation increases oxygen consumption of neurons in culture [90], so the toxic effect of calcium loading is not likely to be due to inhibition of electron transport. Activation of mitochondrial permeability transition remains a possibility, although it is still difﬁcult to deﬁne this phenomenon unequivocally in intact neurons [91]. Nevertheless, it is clear that at some level of calcium loading, mitochondria fail to maintain a membrane potential and lose accumulated calcium stores, and that this causes a failure in ATP generation. Impaired energy production enhances the toxicity of glutamate [92], so that the impairment of mitochondrial function serves as a feedforward mechanism that results in the ampliﬁcation of neuronal injury. The second key mitochondrial mechanism that contributes to neuronal injury is the generation of ROS. Much has been written about mitochondrial ROS generation in the context of neurodegenerative disease (see [8,93] for recent reviews). In general, the view is that ROS generation in the form of superoxide is an inevitable consequence of the consumption of oxygen by mitochondria. Mitochondrial ROS generation is stimulated by inhibitors of electron transport [94], presumably due to the leakage of electrons from complexes I and III. In the context of drug toxicity to neurons, the inhibition of electron transport and subsequent production of ROS may be one of the more pathophysiologically relevant events. In isolated brain mitochondria, ROS signals can be observed with relatively modest levels of inhibition of complex I that are not associated with signiﬁcant decreases in oxygen consumption [7]. Interestingly, concentrations of rotenone sufﬁcient to stimulate ROS production by brain mitochondria are similar to those associated with Parkinsonian-like lesions in animal models, suggesting that mitochondrial ROS generation could be associated with the emergence of speciﬁc neurological disease phenotypes [3,95]. Toxic stimulation of neurons by glutamate results in a calcium-dependent increase in ROS signal [75,96–98], although it is also clear that mitochondria are not the only source of ROS in neurons challenged by ischemic injury [99]. The third component of the toxic triumvirate is energy depletion. As noted above, the brain is highly dependent on oxidative phosphorylation to support ion transport to maintain normal neuronal activity. At a gross level, ischemic brain injury is the consequence of the failure of mitochondrial ATP production due to the interruption of a supply of key substrates. Limiting mitochondrial ATP generation also ampliﬁes injury triggered by toxins such as glutamate [62,92]. Davey and colleagues [100] showed that the spare capacity of electron transport chain complexes varies considerably, with complex I having a spare capacity of about 25%, and complexes III and IV, more than 75%. This suggests that

MITOCHONDRIAL MECHANISMS OF CNS INJURY

263

inhibition of complex I represents a greater risk for energy impairment than that of the other complexes. It is often assumed that partial inhibition of complex I ampliﬁes injury as a consequence of oxidative stress, given the low levels of complex I inhibition necessary to increase ROS generation. However, it is interesting to note that a recent test of this assumption concluded that impaired energy generation had a greater impact than oxidative stress on cell viability [101]. It is not yet clear whether spare respiratory capacity varies between speciﬁc neuronal populations, which would be an interesting way to account for the differential toxicity of drugs on the nervous system described in this chapter. More recent mechanisms to emerge in studies of mitochondrially mediated neuronal injury include alterations in mitochondrial trafﬁcking and morphology. For example, agents that inhibit mitochondrial energy production, including glutamate, oligomycin, and nitric oxide, also acutely impair mitochondrial movement in neurons [102–105]. Mitochondrial trafﬁcking is also impaired by toxic proteins, including expanded poly(Q) huntingtin [106,107], and by low but toxic concentrations of intracellular zinc [12]. However, there are also signals that regulate mitochondrial movement that are clearly not toxic to neurons, including growth factors [11], and signals that stop mitochondrial movement in the vicinity of synapses may be important to ensure correct localization of mitochondria at sites of high energy demand (reviewed in [9]). Mitochondrial morphology might also be an important regulator of neuronal viability given that mitochondrial ﬁssion appears to be associated with a greater vulnerability to cellular injury [86,108,109] and that toxins such as glutamate decrease mitochondrial size [102]. However, it remains difﬁcult to deﬁne a precise cause-and-effect relationship that would link either altered trafﬁcking or morphology to neuronal injury, even though hints of such a relationship exist. 5.2. Potential Manifestations of Drug-Induced Mitochondrial Dysfunction in the CNS In Section 5.1 we detailed some of the neuronal injury mechanisms that are associated with mitochondrial dysfunction. In many cases, these mechanisms have been established experimentally by the acute application of high concentrations of toxins in short-term experiments in vitro. It can be challenging to extrapolate from these ﬁndings to account for potential effects of mitochondrial toxins that are the result of long-term exposures of low concentrations of toxins in vivo. However, even assuming that drug-mediated neuronal injury will be the consequence of relatively modest and potentially long-term exposures, it is possible to identify scenarios that could be indicative of the effects of a mitochondrial toxin in the CNS. Effects on a Selectively Vulnerable Neuronal Population The experiments that expose rats to systemic low doses of rotenone and produce a Parkinson-like syndrome [3] raise the possibility that toxin exposure at relatively low concentrations at a mitochondrial target within a selectively vulnerable neuronal population

264

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

could promote neuronal loss and trigger a disease like Parkinson’s. The basis for selective cell vulnerability is rarely understood, although selective vulnerability is implied in most of the established examples of drug toxicity described in this chapter. Understanding the primary target for this form of injury (i.e., complex I) and the dose and time relationships between drug exposure and injury would allow the prediction of toxicity in this condition. Additivity with Other Toxic Burdens As already documented, relatively modest inhibition of mitochondrial function will amplify neuronal injury triggered by other mechanisms. Perhaps the most obvious scenario would be the addition of a low-level drug-mediated impairment of mitochondrial function with a second form of injury, such as a stroke. The consequence in this case would be a worse outcome from the stroke. This would be a more difﬁcult form of drug effect to study, because the conclusions would be highly dependent on the choice of injury mechanism to use. It would also be difﬁcult to detect following drug exposure in humans, due to the limited number of comparable injury-producing events in drug-taking patients. Slowly Accumulated Injury A surprising observation is the time that is sometimes required to see an expression of injury following mitochondrial impairment. Examples of slowly developing injury that is unequivocally attributable to altered mitochondrial function include the mitochondrial late-onset degeneration (MILON) mouse, which requires several months following interruption of mtDNA replication and gene expression in the forebrain prior to manifestations of injury [110], and the MitoPark mouse, in which dopaminergic neurons die over a year after the elimination of mitochondrial transcription factor A [111]. Injury that requires long periods (perhaps several years) of low-level drug exposure prior to development of a phenotype are very difﬁcult to detect in preclinical models, and typically require extrapolation from the consequences of substantially higher drug doses. These mechanisms are clearly not mutually exclusive, and an anticipated paradigm would reasonably reﬂect both the slow accumulation of injury and selective tissue vulnerability. Individual pharmacogenomic variation could also be important through an impact on drug absorption, metabolism, or excretion that changes either the drug exposure levels or the formation of active metabolites. The consequence of this combination of variables could plausibly be CNS injury, which can only be detected with careful retrospective analysis of data from very large databases of reported adverse reactions, such as a recent report of an amyotrophic lateral sclerosislike syndrome in patients taking statins [112] (see Chapter 11). 6. CONCLUSIONS In this chapter we have documented several relatively well-established cases where drug adverse effects are likely to be due to toxic effects on mitochondrial

REFERENCES

265

function. The manifestations of these established toxic effects are largely impairments of sensory or motor function that arise from the inhibition of mitochondrial DNA replication, or transcription and translation of the mitochondrial genome. While the mechanisms that link drug action to injury are fairly compelling, it is rare that the basis for the selectivity of the injury is understood. In addition to these examples, there are numerous cases where drugs have measurable effects on electron transport and ATP generation. As several inhibitors of this type are known to produce selective neuronal injury, it raises the possibility that these drugs could also produce toxic effects in the CNS. However, to a large extent, these toxicities have not yet been identiﬁed. A further understanding of the consequences of modest mitochondrial impairment as well as the mechanisms underlying the injury produced by known neurotoxins should provide important insights that will reﬁne predictive approaches to understanding drug toxicity in the nervous system.

REFERENCES 1. Silver I, Erecinska M. Oxygen and ion concentrations in normoxic and hypoxic brain cells. Adv Exp Med Biol. 1998;454:7–16. 2. Benard G, Faustin B, Passerieux E, et al. Physiological diversity of mitochondrial oxidative phosphorylation. Am J Physiol Cell Physiol. 2006;291:1172–1182. 3. Betarbet R, Sherer TB, MacKenzie G, Garcia-Osuna M, Panov AV, Greenamyre JT. Chronic systemic pesticide exposure reproduces features of Parkinson’s disease. Nat Neurosci. 2000;3:1301–1306. 4. Beal MF, Brouillet E, Jenkins BG, et al. Neurochemical and histologic characterization of striatal excitotoxic lesions produced by the mitochondrial toxin 3-nitropropionic acid. J Neurosci. 1993;13:4181–4192. 5. Wallace DC. Diseases of the mitochondrial DNA. Annu Rev Biochem. 1992;61:1175–1212. 6. Erecinska M, Silver IA. Ions and energy in mammalian brain. Prog Neurobiol. 1994;43:37–71. 7. Votyakova TV, Reynolds IJ. Delta psi(m)–dependent and –independent production of reactive oxygen species by rat brain mitochondria. J Neurochem. 2001;79:266–277. 8. Lin MT, Beal MF. Mitochondrial dysfunction and oxidative stress in neurodegenerative diseases. Nature. 2006;443:787–795. 9. Chang DTW, Reynolds IJ. Mitochondrial trafﬁcking and morphology in healthy and injured neurons. Prog Neurobiol. 2006;80:241–268. 10. Miller KE, Sheetz MP. Axonal mitochondrial transport and potential are correlated. J Cell Sci. 2004;117:2791–2804. 11. Chada SR, Hollenbeck PJ. Nerve growth factor signaling regulates motility and docking of axonal mitochondria. Curr Biol. 2004;14:1272–1276. 12. Malaiyandi LM, Honick AS, Rintoul GL, Wang QMJ, Reynolds IJ. Zn2+ inhibits mitochondrial movement in neurons by phosphatidylinositol 3-kinase activation. J Neurosci. 2005;25:9507–9514.

266

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

13. Reynolds IJ, Malaiyandi LM, Coash M, Rintoul GL. Mitochondrial trafﬁcking in neurons: a key variable in neurodegeneration? J Bioenerg. Biomembr. 2004;36:283–286. 14. Chan K, Truong D, Shangari N, O’Brien PJ. Drug-induced mitochondrial toxicity. Expert Opin Drug Metab Toxicol. 2005;1:655–669. 15. Lewis W, Dalakas MC. Mitochondrial toxicity of antiviral drugs. Nat Med. 1995;1:417–422. 16. Moyle GJ, Sadler M. Peripheral neuropathy with nucleoside antiretrovirals: risk factors incidence and management. Drug Saf. 1998;19:481–494. 17. Kakuda TN. Pharmacology of nucleoside and nucleotide reverse transcriptase inhibitor–induced mitochondrial toxicity. Clin Ther. 2000;22:685–708. 18. Chen CH, Vazquez-Padua M, Cheng YC. Effect of anti-human immunodeﬁciency virus nucleoside analogs on mitochondrial DNA and its implication for delayed toxicity. Mol Pharmacol. 1991;39:625–628. 19. Mathews CK, Song S. Maintaining precursor pools for mitochondrial DNA replication. FASEB J . 2007;10:2294–2303. 20. Anderson TD, Davidovich A, Feldman D, et al. Mitochondrial schwannopathy and peripheral myelinopathy in a rabbit model of dideoxycytidine neurotoxicity. Lab Invest. 1994;70:724–739. 21. Quasthoff S, Hartung HP. Chemotherapy-induced peripheral neuropathy. J Neurol. 2002;249:9–17. 22. Rowinsky EK, Chaudhry V, Cornblath DR, Donehower RC. Neurotoxicity of Taxol. J Natl Cancer Inst Monogr. 1993:107–115. 23. Morris RL, Hollenbeck PJ. Axonal transport of mitochondria along microtubules and F-actin in living vertebrate neurons. J Cell Biol. 1995;131:1315–1326. 24. Evtodienko YV, Teplova VV, Sidash SS, Ichas F, Mazat JP. Microtubule-active drugs suppress the closure of the permeability transition pore in tumour mitochondria. FEBS Lett. 1996;393:86–88. 25. Kidd JF, Pilkington MF, Schell MJ, et al. Paclitaxel affects cytosolic calcium signals by opening the mitochondrial permeability transition pore. J Biol Chem. 2002;277:6504–6510. 26. Mironov SL, Ivannikov MV, Johansson M. [Ca2+ ](i) signaling between mitochondria and endoplasmic reticulum in neurons is regulated by microtubules: from mitochondrial permeability transition pore to Ca2+ -induced Ca2+ release. J Biol Chem. 2005;280:715–721. 27. Wang GJ, Jackson JG, Thayer SA. Altered distribution of mitochondria impairs calcium homeostasis in rat hippocampal neurons in culture. J Neurochem. 2003;87:85–94. 28. Andre N, Braguer D, Brasseur G, et al. Paclitaxel induces release of cytochrome c from mitochondria isolated from human neuroblastoma cells. Cancer Res. 2000;60:5349–5353. 29. Flatters SJL, Bennett GJ. Studies of peripheral sensory nerves in paclitaxel-induced painful peripheral neuropathy: evidence for mitochondrial dysfunction. Pain. 2006;122:245–257. 30. Siau C, Bennett GJ. Dysregulation of cellular calcium homeostasis in chemotherapy-evoked painful peripheral neuropathy. Anesth Analg. 2006;102:1485–1490.

REFERENCES

267

31. Topp KS, Tanner KD, Levine JD. Damage to the cytoskeleton of large diameter sensory neurons and myelinated axons in vincristine-induced painful peripheral neuropathy in the rat. J Comp Neurol. 2000;424:563–576. 32. Dirks AJ, Jones KM. Statin-induced apoptosis and skeletal myopathy. Am J Physiol Cell Physiol. 2006;291:1208–1212. 33. Johnson TE, Zhang X, Bleicher KB, et al. Statins induce apoptosis in rat and human myotube cultures by inhibiting protein geranylgeranylation but not ubiquinone. Toxicol Appl Pharmacol. 2004;200:237–250. 34. Backes JM, Howard PA. Association of HMG-CoA reductase inhibitors with neuropathy. Ann Pharmacother. 2003;37:274–278. 35. Peltier AC, Russell JW. Advances in understanding drug-induced neuropathies. Drug Saf. 2006;29:23–30. 36. Nadanaciva S, Dykens JA, Bernal A, Capaldi RA, Will Y. Mitochondrial impairment by PPAR agonists and statins identiﬁed via immunocaptured OXPHOS complex activities and respiration. Toxicol Appl Pharmacol. 2007;223:277–287. 37. Kaufmann P, Torok M, Zahno A, Waldhauser KM, Brecht K, Krahenbuhl S. Toxicity of statins on rat skeletal muscle mitochondria. Cell Mol Life Sci. 2006;63:2415–2425. 38. Langsjoen PH, Langsjoen AM. The clinical use of HMG CoA-reductase inhibitors and the associated depletion of coenzyme Q10 : a review of animal and human publications. Biofactors. 2003;18:101–111. 39. Bleske BE, Willis RA, Anthony M, et al. The effect of pravastatin and atorvastatin on coenzyme Q10 . Am Heart J. 2001;142:E2. 40. Shults CW, Oakes D, Kieburtz K, et al. Effects of coenzyme Q10 in early Parkinson disease: evidence of slowing of the functional decline. Arch Neurol. 2002;59:1541–1550. 41. Lieberman A, Lyons K, Levine J, Myerburg R. Statins, cholesterol, co-enzyme Q10, and Parkinson’s disease. Parkinsonism Relat Disord. 2005;11:81–84. 42. Santaella RM, Fraunfelder FW. Ocular adverse effects associated with systemic medications: recognition and management. Drugs. 2007;67:75–93. 43. Carelli V, Ross-Cisneros FN, Sadun AA. Optic nerve degeneration and mitochondrial dysfunction: genetic and acquired optic neuropathies. Neurochem Int. 2002;40:573–584. 44. Davies VJ, Hollins AJ, Piechota MJ, et al. Opa1 deﬁciency in a mouse model of autosomal dominant optic atrophy impairs mitochondrial morphology, optic nerve structure and visual function. Hum Mol Genet. 2007;16:1307–1318. 45. Ames A, Li YY, Heher EC, Kimble CR. Energy-metabolism of rabbit retina as related to function: high cost of Na+ transport. J Neurosci. 1992;12:840–853. 46. Holt D, Harvey D, Hurley R. Chloramphenicol toxicity. Adverse Drug React Toxicol Rev. 1993;12:83–95. 47. Yunis AA. Chloramphenicol: relation of structure to activity and toxicity. Annu Rev Pharmacol Toxicol. 1988;2883–100. 48. Harley RD, Huang NN, Macri CH, Green WR. Optic neuritis and optic atrophy following chloramphenicol in cystic ﬁbrosis patients. Trans Am Acad Ophthalmol Otolaryngol. 1970;74:1011–1031.

268

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

49. Hyde GE, Durham D. Increased deafferentation-induced cell-death in chick brain-stem auditory neurons following blockade of mitochondrial protein-synthesis with chloramphenicol. J Neurosci. 1994;14:291–300. 50. Hyde GE, Rubel EW. Mitochondrial role in hair cell-survival after injury. Otolaryngol Head Neck Surg. 1995;113:530–540. 51. Hartlage-Rubsamen M, Rubel EW. Inﬂuence of mitochondrial protein synthesis inhibition on deafferentation-induced ultrastructural changes in nucleus magnocellularis of developing chicks. J Comp Neurol. 1996;371:448–460. 52. Mottin S, Laporte P, Cespuglio R. Inhibition of NADH oxidation by chloramphenicol in the freely moving rat measured by picosecond time-resolved emission spectroscopy. J Neurochem. 2003;84:633–642. 53. Moulin-Sallanon M, Millet P, Rousset C, et al. Chloramphenicol decreases brain glucose utilization and modiﬁes the sleep–wake cycle architecture in rats. J Neurochem. 2005;93:1623–1632. 54. Barron MJ, Grifﬁths P, Turnbull DM, Bates D, Nichols P. The distributions of mitochondria and sodium channels reﬂect the speciﬁc energy requirements and conduction properties of the human optic nerve head. Br J Ophthalmol. 2004;88:286–290. 55. Rizzo JF. Adenosine triphosphate deﬁciency: a genre of optic neuropathy. Neurology. 1995;45:11–16. 56. Barron GJ, Tepper L, Iovine G. Ocular toxicity from ethambutol. Am J Ophthalmol. 1974;77:256–260. 57. Fraunfelder FW, Sadun AA, Wood T. Update on ethambutol optic neuropathy. Expert Opin Drug Saf. 2006;5:615–618. 58. Kozak SF, Inderlied CB, Hsu HY, Heller KB, Sadun AA. The role of copper on ethambutol’s antimicrobial action and implications for ethambutol-induced optic neuropathy. Diagn Microbiol Infect Dis. 1998;30:83–87. 59. Cole A, May PM, Williams DR. Metal-binding by pharmaceuticals: 1. Copper(Ii) and zinc(Ii) interactions following ethambutol administration. Agents Actions. 1981;11:296–305. 60. Prohaska JR, Wells WW. Copper deﬁciency in the developing rat brain: evidence for abnormal mitochondria. J Neurochem. 1975;25:221–228. 61. Heng JE, Vorwerk CK, Lessell E, Zurakowski D, Levin LA, Dreyer EB. Ethambutol is toxic to retinal ganglion cells via an excitotoxic pathway. Invest Ophthalmol Vis Sci. 1999;40:190–196. 62. Greene JG, Greenamyre JT. Bioenergetics and glutamate excitotoxicity. Prog Neurobiol. 1996;48:613–634. 63. Yoon YH, Jung KH, Sadun AA, Shin HC, Koh JY. Ethambutol-induced vacuolar changes and neuronal loss in rat retinal cell culture: mediation by endogenous zinc. Toxicol Appl Pharmacol. 2000;162:107–114. 64. Dineley KE, Scanlon JM, Kress GJ, Stout AK, Reynolds IJ. Astrocytes are more resistant than neurons to the cytotoxic effects of increased [Zn(2 + )](i). Neurobiol Dis. 2000;7:310–320. 65. Tephly TR. The Toxicity of methanol. Life Sci. 1991;48:1031–1041. 66. Roe O. The metabolism and toxicity of methanol. Pharmacol Rev. 1955;7:399–412. 67. Nicholls P. Effect of formate on cytochrome-Aa3 and on electron-transport in intact respiratory-chain. Biochim Biophys Acta. 1976;430:13–29.

REFERENCES

269

68. Eells JT, Henry MM, Lewandowski MF, Seme MT, Murray TG. Development and characterization of a rodent model of methanol-induced retinal and optic nerve toxicity. Neurotoxicology. 2000;21:321–330. 69. Yorgason JG, Fayad JN, Kalinec F. Understanding drug ototoxicity: molecular insights for prevention and clinical management. Expert Opin Drug Saf. 2006;5:383–399. 70. Hutchin T, Cortopassi G. Proposed molecular and cellular mechanism for aminoglycoside ototoxicity. Antimicrob Agents Chemother. 1994;38:2517–2520. 71. Bates DE. Aminoglycoside ototoxicity. Drugs Today (Barc). 2003;39:277–285. 72. Priuska EM, Schacht J. Formation of free radicals by gentamicin and iron and evidence for an iron/gentamicin complex. Biochem Pharmacol. 1995;50:1749–1752. 73. Song BB, Schacht J. Variable efﬁcacy of radical scavengers and iron chelators to attenuate gentamicin ototoxicity in guinea pig in vivo. Hear Res. 1996;94:87–93. 74. Basile AS, Huang JM, Xie C, Webster D, Berlin C, Skolnick P. N -Methyld-aspartate antagonists limit aminoglycoside antibiotic-induced hearing loss. Nat Med. 1996;2:1338–1343. 75. Reynolds IJ, Hastings TG. Glutamate induces the production of reactive oxygen species in cultured forebrain neurons following NMDA receptor activation. J Neurosci. 1995;15:3318–3327. 76. Stout AK, Raphael HM, Kanterewicz BI, Klann E, Reynolds IJ. Glutamate-induced neuron death requires mitochondrial calcium uptake. Nat Neurosci. 1998;1:366–373. 77. Van Camp G, Smith RJ. Maternally inherited hearing impairment. Clin Genet. 2000;57:409–414. 78. Higashi K. Unique inheritance of streptomycin-induced deafness. Clin Genet. 1989;35:433–436. 79. Prezant TR, Agapian JV, Bohlman MC, et al. Mitochondrial ribosomal RNA mutation associated with both antibiotic-induced and non-syndromic deafness. Nat Genet. 1993;4:289–294. 80. Gurtler N, Schmuziger N, Kim Y, Mhatre AN, Jungi M, Lalwani AK. Audiologic testing and molecular analysis of 12S rRNA in patients receiving aminoglycosides. Laryngoscope. 2005;155:640–644. 81. Choi DW, Rothman SM. The role of glutamate neurotoxicity in hypoxic–ischemic neuronal death. Annu Rev Neurosci. 1990;13:171–182. 82. Choi DW. Calcium-mediated neurotoxicity: relationship to speciﬁc channel types and role in ischemic damage. Trends Neurosci. 1988;11:465–469. 83. White RJ, Reynolds IJ. Mitochondrial depolarization in glutamate-stimulated neurons: an early signal speciﬁc to excitotoxin exposure. J Neurosci. 1996;16:5688–5697. 84. Schinder AF, Olson EC, Spitzer NC, Montal M. Mitochondrial dysfunction is a primary event in glutamate neurotoxicity. J Neurosci. 1996;16:6125–6133. 85. Vergun O, Keelan J, Khodorov BI, Duchen MR. Glutamate-induced mitochondrial depolarisation and perturbation of calcium homeostasis in cultured rat hippocampal neurones. J Physiol (Lond). 1999;519:451–466. 86. Brocard JB, Rintoul GL, Reynolds IJ. New perspectives on mitochondrial morphology in cell function. Biol Cell. 2003;95:239–242.

270

MANIFESTATIONS OF DRUG TOXICITY ON MITOCHONDRIA

87. Budd SL, Nicholls DG. Mitochondria, calcium regulation, and acute glutamate excitotoxicity in cultured cerebellar granule cells. J Neurochem. 1996;67:2282–2291. 88. Pivovarova NB, Nguyen HV, Winters CA, Brantner CA, Smith CL, Andrews SB. Excitotoxic calcium overload in a subpopulation of mitochondria triggers delayed death in hippocampal neurons. J Neurosci. 2004;24:5611–5622. 89. Nicholls D, Akerman K. Mitochondrial calcium transport. Biochim Biophys Acta. 1982;683:57–88. 90. Jekabsons MB, Nicholls DG. In situ respiration and bioenergetic status of mitochondria in primary cerebellar granule neuronal cultures exposed continuously to glutamate. J Biol Chem. 2004;279:32989–33000. 91. Reynolds IJ. Mitochondrial membrane potential and the permeability transition in excitotoxicity. Oxidat/Energy Metab Neurodegen Disord. 1999;893:33–41. 92. Novelli A, Reilly JA, Lysko PG, Henneberry RC. Glutamate becomes neurotoxic via the N -methyl- d-aspartate receptor when intracellular energy-levels are reduced. Brain Res. 1988;451:205–212. 93. Orrenius S, Gogvadze V, Zhivotovsky B. Mitochondrial oxidative stress: implications for cell death. Annu Rev Pharmacol Toxicol. 2007;47:143–183. 94. Turrens JF. Superoxide production by the mitochondrial respiratory chain. Biosci Rep. 1997;17:3–8. 95. Sherer TB, Betarbet R, Testa CM, et al. Mechanism of toxicity in rotenone models of Parkinson’s disease. J Neurosci. 2003;23:10756–10764. 96. Lafoncazal M, Pietri S, Culcasi M, Bockaert J. NMDA-dependent superoxide production and neurotoxicity. Nature. 1993;364:535–537. 97. Dugan LL, Sensi SL, Canzoniero LMT, et al. Mitochondrial production of reactive oxygen species in cortical-neurons following exposure to N -methyl- d-aspartate. J Neurosci. 1995;15:6377–6388. 98. Bindokas VP, Jordan J, Lee CC, Miller RJ. Superoxide production in rat hippocampal neurons: selective imaging with hydroethidine. J Neurosci. 1996;16:1324–1336. 99. Abramov AY, Scorziello A, Duchen MR. Three distinct mechanisms generate oxygen free radicals in neurons and contribute to cell death during anoxia and reoxygenation. J Neurosci. 2007;27:1129–1138. 100. Davey GP, Peuchen S, Clark JB. Energy thresholds in brain mitochondria: potential involvement in neurodegeneration. J Biol Chem. 1998;273:12753–12757. 101. Yadava N, Nicholls DG. Spare respiratory capacity rather than oxidative stress regulates glutamate excitotoxicity after partial respiratory inhibition of mitochondrial complex I with rotenone. J Neurosci. 2007;27:7310–7317. 102. Rintoul GL, Filiano AJ, Brocard JB, Kress GJ, Reynolds IJ. Glutamate decreases mitochondrial size and movement in primary forebrain neurons. J Neurosci. 2003;23:7881–7888. 103. Rintoul GL, Bennett VJ, Papaconstandinou NA, Reynolds IJ. Nitric oxide inhibits mitochondrial movement in forebrain neurons associated with disruption of mitochondrial membrane potential. J Neurochem. 2006;97:800–806. 104. Brown GC. Nitric oxide and mitochondrial respiration. Biochim Biophys Acta Bioenerg. 1999;1411:351–369. 105. Zanelli SA, Trimmer PA, Solenski NJ. Nitric oxide impairs mitochondrial movement in cortical neurons during hypoxia. J Neurochem. 2006;97:724–736.

REFERENCES

271

106. Chang DTW, Rintoula GL, Pandipati S, Reynolds IJ. Mutant huntingtin aggregates impair mitochondrial movement and trafﬁcking in cortical neurons. Neurobiol Dis. 2006;22:388–400. 107. Trushina E, Dyer RB, Badger JD, et al. Mutant huntingtin impairs axonal trafﬁcking in mammalian neurons in vivo and in vitro. Mol Cell Biol. 2004;24:8195–8209. 108. Frank S, Gaume B, Bergmann-Leitner ES, et al. The role of dynamin-related protein 1, a mediator of mitochondrial ﬁssion, in apoptosis. Dev Cell. 2001;1:515–525. 109. Bossy-Wetzel E, Barsoum MJ, Godzik A, Schwarzenbacher R, Lipton SA. Mitochondrial ﬁssion in apoptosis, neurodegeneration and aging. Curr Opin Cell Biol. 2003;15:706–716. 110. Sorensen L, Ekstrand M, Silva JP, et al. Late-onset corticohippocampal neurodepletion attributable to catastrophic failure of oxidative phosphorylation in MILON mice. J Neurosci. 2001;21:8082–8090. 111. Ekstrand MI, Terzioglu M, Galter D, et al. Progressive parkinsonism in mice with respiratory-chain-deﬁcient dopamine neurons. Proc Nat Acad Sci USA. 2007;104:1325–1330. 112. Edwards IR, Star K, Kiuru A. Statins, neuromuscular degenerative disease and an amyotrophic lateral sclerosis-like syndrome: an analysis of individual case safety reports from vigibase. Drug Saf. 2007;30:515–525.

9 LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRAL THERAPEUTICS Ulrich A. Walker Department of Rheumatology, Basel University, Basel, Switzerland

1. Introduction 2. Pathogenesis of NRTI toxicity 3. Clinical spectrum of mitochondrial toxicity 3.1. Mitochondrial myopathy 3.2. Lipodystrophy 4. Monitoring and predicting mitochondrial toxicity 5. Drug interactions 6. Therapy

273 274 276 277 277 280 281 281

1. INTRODUCTION In anti-HIV therapy, at least two nucleoside analog reverse transcriptase inhibitors (NRTIs) are usually combined with a protease inhibitor, or with a nonnucleoside analog reverse transcriptase inhibitor. This triple combination of antiretrovirals is referred to as highly active retroviral therapy (HAART) and has saved thousands of lives since its widespread clinical implementation in 1996. Unfortunately, most people experience some side effects during long-term antiretroviral treatment, the best known being lipodystrophy. The term lipodystrophy was coined to

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

273

274

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

describe a syndrome of abnormal distribution of subcutaneous adipose tissue in its more narrow sense, but also a metabolic syndrome consisting of dyslipidemia and insulin resistance in a broader sense. Today, more than 10 years after the introduction of HAART, the lipodystrophy syndrome is known to result from overlapping but distinct effects of the various drugs within the HAART cocktail. The primary pathogenetic mechanism through which NRTIs precipitate metabolic changes and organ toxicities is via mitochondrial toxicity [1,2]. 2. PATHOGENESIS OF NRTI TOXICITY NRTIs (Table 1) are prodrugs because they require activation in the cell through phosphorylation before they are able to inhibit HIV reverse transcriptase [3,4]. In addition to impairing the HIV replication machinery, the NRTI triphosphates also inhibit polymerase γ, which is responsible for the replication of mitochondrial DNA (mtDNA). Polymerase γ inhibition actually results from several distinct steps [4]. The ﬁrst step involves competition of NRTI triphosphates with the natural nucleotide triphosphate. If this competition is successful, the NRTIs are incorporated into the nascent mtDNA strand. This second step causes chain termination because the NRTIs lack a second hydroxyl group to which new DNA building blocks can be attached. The result of polymerase γ impairment is mtDNA depletion, a quantitative reduction of the mtDNA copy number. The respiratory chain is not only responsible for the synthesis of ATP through oxidative phosphorylation, but by consuming NADH and FADH as end products of fatty acid oxidation, it is also an important regulator of β-oxidation [5]. This explains the micro- or macrovesicular accumulation of intracellular triglycerides which often accompanies mitochondrial toxicity. Normal oxidative phosphorylation is also essential for the de novo synthesis of all intracellular pyrimidine nucleosides by dihydroorotate dehydrogenase (DHODH) [6,7]. Therefore, TABLE 1

Clinically Licensed Nucleoside Analog Reverse Transcriptase Inhibitors

NRTI Name Zidovudine (3 -azido-3 -deoxythymidine) Stavudine (2 ,3 -didehydro-2 ,3 deoxythymidine) Zalcitabine (2 ,3 -dideoxycytidine) Lamivudine (2 ,3 -dideoxy-3 -thiacytidine) Emtricitabine (2 -Deoxy-5-ﬂuoro-3 thiacytidine) Abacavir (4 -[2 -amino-6 (cyclopropylamino)-9H -purine-9-yl]2-cyclopentene-1-methanol) Didanosine (2 ,3 -dideoxyinosine) Tenofovir (R-9-[2-phosphonomethoxypropyl] adenine)

Abbreviation

ddC 3TC FTC

Analog of:   Thymidine        Pyrimidines       Cytidine     

ABC

Guanosine

ddI TDF

Inosine Adenosine

AZT d4T

              

Purines

275

PATHOGENESIS OF NRTI TOXICITY NRTI-incorporation into mtDNA by gamma polymerase

mtDNAdepletion

respiratory chain dysfunction

relative NRTI-excess

pyrimidine pool↓

DHODHinhibition

uridine

Figure 1 A vicious cycle probably contributes to the mitochondrial toxicity of antiretroviral pyrimidine NRTIs. The beneﬁcial effects of therapeutic uridine supplementation are thought to be mediated by replenishing intracellular pyrimidine pools. DHODH, dihydroorotate dehydrogenase.

respiratory impairment that results in pyrimidine depletion increases the efﬁciency of competition by pyrimidine NRTIs at polymerase γ [8,9], establishing a vicious circle leading to mtDNA depletion (Figure 1). Support for this cycle stems from the observation that experimental pyrimidine depletion with redoxal, a direct DHODH inhibitor, exacerbates NRTI-mediated toxicity (Setzer B, Lebrecht D, Walker UA, Am J Pathol. 2008; 172:681–690). Research with leﬂunomide [10,11], another direct DHODH inhibitor and a licensed immunosuppressive drug, has shown that pyrimidine depletion activates p53 and its immediate transcriptional target p21 [10,12]. p53 also regulates activation of Rb protein and of cyclins [13] that arrests the cell cycle in the G1 phase. p53 activation can also promote apoptosis [14]. These molecular mechanisms provide a tenable explanation of why cells with mtDNA depletion cease dividing and then die. The importance of the intracellular pyrimidine pools for the survival of cells without a functional respiratory chain is also supported by the fact that cells lacking mtDNA (rho0 cells) are rescued from cell death and grow normally if the intracellular pyrimidine pools are replenished by pyrimidine precursors such as uridine, which can restore pyrimidine pools distal from DHODH [15]. The mitochondrial toxicity of NRTIs follows certain principles: 1. Not all NRTIs are equipotent inhibitors of polymerase γ. The “d-drugs”— namely, the dideoxynucleosides zalcitabine, didanosine, and stavudine— are relatively strong inhibitors of isolated polymerase γ [3,16], whereas the other NRTIs have a weaker effect. The hierarchy of polymerase γ inhibition for the active NRTI metabolites has been determined as follows: zalcitabine > didanosine > stavudine > lamivudine ≥ abacavir ≥ tenofovir ≥ emtricitabine. 2. Zidovudine is also a NRTI but is atypical because the activated zidovudine triphosphate is a weak inhibitor of polymerase γ [3]. Zidovudine is also an inhibitor of thymidine kinases, and as such interferes with the formation of

276

3.

4.

5.

6. 7.

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

this nucleotide [17]. Limited nucleotide supply may impair mtDNA synthesis. This is demonstrated by inborn defects of mitochondrial thymidine and deoxyguanosine kinase, both of which cause mtDNA depletion in human muscle [18,19]. Zidovudine also binds to adenylate kinase and inhibits the mitochondrial ADP/ATP translocator [20–22]. The latter mechanisms may explain why zidovudine impairs mitochondrial function more rapidly than do other NRTIs [21,23]. Mitochondrial toxicity is concentration dependent; high NRTI concentrations induce more profound mtDNA depletion. The clinical dosing of some nucleoside analogs is close to the limit of tolerability with respect to mitochondrial toxicity [24]. The onset of mitochondrial toxicity requires prolonged time. Changes in mitochondrial metabolism are observed only when mtDNA depletion exceeds a certain threshold that varies in different tissues. As a consequence of this effect, the onset of mitochondrial toxicity is not observed clinically in the ﬁrst few months of HAART [25]. Mitochondrial toxicity is tissue speciﬁc. Tissue speciﬁcity is explained by the fact that the uptake of the NRTI prodrugs into mitochondria and their activation by phosphorylation vary between individual cell types. Differences of organ susceptibility may also result from different intracellular concentrations of the normal nucleotides with which the NRTI compete at various steps of transport and activation and at polymerase γ itself. Didanosine, for example, has relatively strong hepatotoxicity [23,26] but does not appear to affect adipocytes [27]. Zidovudine, in contrast, appears to impair the metabolism of adipocytes but not of hepatocytes [23,26,27]. The combination of different NRTIs can result in unpredictable additive or synergistic toxicity, even if one drug is not particularly toxic alone [23,26]. Recent data suggest that transcription of mtDNA can be impaired even in the absence of mtDNA depletion [28,29]. The mechanism and the clinical signiﬁcance of this observation are not yet understood.

3. CLINICAL SPECTRUM OF MITOCHONDRIAL TOXICITY MtDNA depletion may manifest clinically in one or several main target tissues. In the liver, mitochondrial toxicity is associated with hepatomegaly and increased lipid deposits, resulting in micro- or macrovesicular steatosis [30]. Steatosis may be accompanied by elevated serum liver transaminases. Such steatohepatitis may progress to liver failure and lactic acidosis, a potentially fatal but fortunately rare complication [24,31–35]. Steatohepatitis and lactic acidosis were ﬁrst described in the early 1990s in patients receiving didanosine monotherapy [24,31]. Mitochondrial liver complications have been observed primarily with dideoxynucleosides (e.g., with didanosine, stavudine, and zalcitabine), but also with other NRTIs. MtDNA depletion has been demonstrated in the liver of HIV

CLINICAL SPECTRUM OF MITOCHONDRIAL TOXICITY

277

patients, with each of the dideoxynucleosides inducing a time-dependent mtDNA depletion [26]. Morphologically abnormal mitochondria were observed via electron microscopy. A classical complication of mitochondrial toxicity is elevation of serum lactate. The liver is a major site for oxidation of lactate generated by glycolysis in skeletal muscle. Circulating lactate, then, reﬂects both production and oxidation, so that mitochondrial impairment in liver will contribute to hyperlactatemia by decreasing lactate utilization, whether it arises systemically or endogenously from a metabolic compensation by the hepatocyte [31,35]. Symptoms associated with hyperlactatemia are nonspeciﬁc and consist of nausea, right upper quadrant abdominal tenderness, or myalgias. In the majority of cases, bicarbonate levels and the anion gap (Na+ –[HCO3− + Cl− ]) are normal. 3.1. Mitochondrial myopathy Mitochondrial myopathy in antiretrovirally treated HIV patients was ﬁrst described with high-dose zidovudine therapy [36] (see Chapter 7). Skeletal muscle weakness may manifest under dynamic or static exercise, and serum creatine kinase (CK) is often normal or only minimally elevated. Muscle histology can distinguish this form of NRTI toxicity from HIV myopathy, which may occur concurrently. On histochemical examination, the muscle ﬁbers of the former are frequently negative for cytochrome c oxidase and carry ultrastructurally abnormal mitochondria, whereas those of the latter are typically inﬁltrated by CD8-positive T lymphocytes. Exercise challenge may reveal a low lactate threshold and reduced lactate clearance [37], but in clinical practice these changes are difﬁcult to distinguish from lack of aerobic exercise (detraining). Prolonged treatment with dideoxynucleosides frequently leads to a predominantly symmetrical, sensory, and distal polyneuropathy of the lower extremities [24,38–42]. Elevated serum lactate levels may help to distinguish this axonal neuropathy from its HIV-associated phenocopy, although in most cases lactate levels are normal [43]. The differential diagnosis should also take into account the fact that mitochondrial polyneuropathy generally occurs weeks or months after beginning didexoxynucleoside treatment. In contrast, HIV-associated polyneuropathy generally does not worsen, and may even improve, with prolonged antiretroviral treatment. NRTI-related polyneuropathy may slowly reverse upon drug cessation [43], and animal data suggest that its occurrence can be prevented by oral uridine supplementation (D. Lebrecht, personal communication). 3.2. Lipodystrophy One of the most debilitating side effects of HAART is a physical alteration of body composition called lipodystrophy. Some subjects affected with lipodystrophy may experience abnormal fat accumulation intraabdominally and in the dorsocervical region, whereas others may develop subcutaneous fat wasting at the Bichat’s fat pad in the cheeks, of temporal fat, or at the buttocks and

278

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

extremities (see Chapter 21). The two opposite components of lipodystrophy (fat accumulation and fat loss) may manifest either independent of each other or simultaneously in the same person. Fat wasting (also called lipoatrophy) generally begins no sooner than one year after the initiation of HAART [25]. The manifestation of lipoatrophy is inﬂuenced primarily by the choice of NRTIs, and randomized trials have identiﬁed stavudine and zidovudine in particular [44,45]. Numerous groups have described mtDNA depletion in cultured adipocytes and subcutaneous adipose tissue of lipoatrophic subjects [46,47]. When stavudine or zidovudine is replaced by another NRTI, mtDNA levels and apoptotic indices in fat improve along with an objectively measurable, albeit small increase in subcutaneous adipose tissue [48]. Electron microscopy of subcutaneous fat shows that the mitochondria contain vacuoles and inclusions [46]. The abnormal fat accumulation of lipodystrophy has generally been associated with protease inhibitors and frequently is accompanied by insulin resistance and dyslipidemia. In murine adipose cell lines, protease inhibitors were found to promote defects in the maturation of lamins that are important for the organization and stability of the nucleus within cells. Altered lamins may be responsible for an impaired nuclear localization of the proadipogenic transcription factor SREBP-1 [49]. In contrast to trials that found a beneﬁcial effect on lipoatrophy if HAART was switched away from NRTIs, deleting protease inhibitors did not ameliorate lipoatrophy or adipocyte apoptosis [50–53]. Taken together, the data demonstrate a predominant role of NRTI-induced mitochondrial dysfunction in the pathogenesis of lipoatrophy. Some studies have suggested an effect of NRTIs on the mtDNA levels in peripheral blood mononuclear cells [54,55]. Functionally, mitochondrial toxicity on lymphocytes may result in lymphopenia. CD4 and CD8 lymphocytopenia of delayed onset was observed when didanosine plasma levels were slightly increased through pharmacokinetic interactions or a low body weight of patients [56]. Exposure of mitotically stimulated T lymphocytes to didanosine, stavudine, zalcitabine, and zidovudine also resulted in a substantial mtDNA depletion, a late-onset decline of lymphocyte proliferation and increased apoptosis [57,58]. Taken together, the data suggest that NRTIs can act as immunosuppressive drugs. Zidovudine also induces anemia and neutropenia by inhibiting hematopoietic progenitor cells [59] (see Chapter 21). In 1994 a trial showed that zidovudine (given ante and intra partum to the mother and to the newborn for 6 weeks) reduced the risk of vertical HIV transmission by approximately two-thirds [60]. Since then, perinatal antiretroviral therapy is instituted in pregnant HIV-positive women as the standard of care even if the mother herself does not require HAART. Zidovudine is the agent most widely used. However, studies of pregnant women, newborns, and animals have raised concerns regarding the safety of NRTIs in the prevention of perinatal HIV transmission. MtDNA depletion and moderate mitochondrial dysfunction in skeletal muscle, heart, and placenta were detected in a simian model of perinatal AZT treatment [61,62]. Low levels of mtDNA were also measured in human placenta

CLINICAL SPECTRUM OF MITOCHONDRIAL TOXICITY

279

and cord blood of neonates [62–64]. In zidovudine-exposed but HIV-uninfected infants, transient anemia and additional long-term blood abnormalities (neutropenia, thrombopenia, and lymphopenia) have been documented. The overall risk of mortality and congenital abnormalities does not appear to be increased [65,66], but rare mitochondrial events cannot be excluded for lack of statistical power. In children born to mothers infected with HIV and perinatally treated with AZT with or without additional lamivudine, an unexpected increase in symptomatic and biochemically conﬁrmed mitochondrial impairment was observed [67]. Other cohorts did not conﬁrm an elevated risk for mitochondrial syndromes after perinatal exposure to antiretroviral therapy in humans, but mitochondrial dysfunction was not speciﬁcally evaluated [68,69]. The incidence of asymptomatic hyperlactatemia in HIV-uninfected children after transplacental NRTI exposure seems to be signiﬁcant [70–72]. Hyperlactatemia was found in 50% of infants, with 30% still showing elevated lactate levels at 1 year of age [70]. In another cohort, in utero exposure to zidovudine or dual NRTI therapy has resulted in at least one abnormally high plasma lactate reading (>2.1 mmol/L, median three samples per infant) in 35 of 38 infants up to 6 months of age [72]. Hyperlactatemia appears to be higher in HIV-uninfected infants who were exposed perinatally to antiretrovirals than in older, HIV-infected pediatric patients on chronic long-term HAART. These ﬁndings may be explained by higher energy requirements of the younger, developing neonate, and by the possibility that mtDNA depletion is faster in rapidly dividing cells where mtDNA segregates to the daughter cells, and thus diminishes with each mitosis. Current guidelines discourage the use of ddI and d4T in HIV-infected pregnant women because of increased mtDNA depletion [62] and increased maternal mortality secondary to lactic acidosis and hepatic steatosis [73,74]. Preclinical data demonstrate zidovudine to be incorporated into nuclear and mitochondrial DNA [75,76] and also to be a carcinogen [75,77]. Long-term follow-up data of perinatal exposed babies in large cohorts are needed. Some NRTIs are also known to cause hyperuricemia [24,78]. Urate may be increased by mitochondrial dysfunction that enhances the formation of lactate because the latter competes with urate for tubular secretion in the kidney [79]. ATP depletion may also increase urate production in the purine nucleotide cycle [79,80]. This mechanism could be the basis for the hyperuricemia in some metabolic myopathies and may provide an explanation for the association between “d-drugs” (e.g., stavudine, didanosine, zalcitabine) and elevated urate [81]. The existence of mitochondrial damage to the kidney is controversial. Supratherapeutic doses of the nucleotide analog reverse transcriptase inhibitor tenofovir induced a Fanconi syndrome (increased urine output resulting from repressed solute uptake by proximate tubule cells) with tubular phosphate loss and osteomalacia in animals. Similar observations were also made with high doses of adefovir, a nucleotide analog now used in lower doses for the treatment of chronic hepatitis B [82]. Nucleotide analogs are imported into renal tubules by means of an anion transporter [83], and it cannot be ruled out that

280

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

excessively high intracellular drug concentrations lead to clinically relevant polymerase-gamma inhibition and mtDNA depletion. Isolated cases of renal failure and tubular damage have been reported with tenofovir [84], and decreased mtDNA levels were found in renal biopsies from patients exposed to tenofovir plus didanosine [85]. However, the data do not allow conclusions about an effect of tenofovir alone. Most trials have concentrated on measuring creatinine clearance and serum phosphate [86], despite the fact that renal dysfunction is mostly preserved in Fanconi’s syndrome. Serum phosphate levels may also be maintained, despite increased phosphate mobilization from bone, thus masking increased renal loss. A review of the data from randomized tenofovir trials suggests no signiﬁcant alterations in phosphate, calcium, or bone mineral density over time, but the number of people studied systematically may be too small. More sensitive methods have recently revealed a diminished renal phosphate reabsorption and an elevated alkaline phosphatase in patients treated with tenofovir [87].

4. MONITORING AND PREDICTING MITOCHONDRIAL TOXICITY There is currently no method to reliably predict the mitochondrial risk of an individual patient. Routine screening of lactate levels in asymptomatic NRTI-treated subjects is not warranted since lactate levels are not predictive of clinical mitochondrial toxicity in these patients [88]. In contrast, lactate levels should be checked promptly in subjects who experience symptoms consistent with mitochondrial toxicity. The predictive value of mtDNA measurements in peripheral blood mononuclear cells (PBMCs) is confounded by technical issues such as platelet contamination, different proportions of different cell populations in the PBMC mixture and the activation status of lymphocytes [89] (but see Chapter 21). Any process that causes lymphocytes to proliferate triples their mtDNA content [58]. Quantifying mtDNA within affected tissues is likely to be more accurate; but this form of monitoring is invasive, not standardized internationally and not prospectively evaluated with regard to clinical endpoints. Once symptoms are established, histological examination of a biopsy may contribute to the correct diagnosis. The following ﬁndings in tissue biopsies point toward a mitochondrial etiology: ultrastructural abnormalities of mitochondria, diminished histochemical activities of cytochrome c oxidase, the detection of intracellular and more speciﬁcally microvesicular steatosis, and ragged-red ﬁbers in skeletal muscle (see Chapter 23). Tools are being developed to assess metabolic toxicity noninvasively within an individual patient and to predict an individual’s metabolic risk adequately prior to its onset (see Chapters 16, 21, and 22). In the interim, clinical vigilance and caution remain critical in the routine care of subjects under antiretroviral treatment.

THERAPY

281

5. DRUG INTERACTIONS Known mitochondrial toxins should be avoided. Valproate and acetylsalicylic acid were reported to trigger lactic acidosis in patients with HIV [90] and in persons with genetic mitochondrial syndromes [91,92] (see Chapter 11). Coadministering didanosine with allopurinol, ganciclovir, or tenofovir is not advised because the latter substances interfere with the breakdown of purines and therefore increase didanosine levels [93,94]. Hydroxyurea [95] and ribavirin [96] increase the active metabolite of didanosine in relation to its natural nucleoside competitor at polymerase γ and augment mitochondrial side effects. Adefovir and cidofovir are also inhibitors of polymerase γ and should be avoided in established mitochondrial toxicity [83,97,98]. Brivudin is a thymidine analog herpes virostatic that could exacerbate NRTI-related mitochondrial toxicity because one of its metabolites interferes with pyrimidine metabolism [99]. This interaction may compromise the antiretroviral activity of NRTIs, and until more data become available, combined use of brivudin with antiretroviral pyrimidine analogs should be avoided. Statins may compromise the respiratory chain by interfering with the isoprenylation of coenzyme Q (but see Chapter 7), but there appear to be no data on potential synergistic toxicity with polymerase-gamma inhibitors. Surprisingly, in one study pravastatin resulted in moderate increases in subcutaneous fat as a secondary endpoint despite limited effects on cholesterol as the primary endpoint [100]. Metformin improved insulin sensitivity in HIV-infected subjects [101]. Although lactic acidosis was described as a rare but serious adverse event of metformin, increased risk has not been demonstrated in HIV-infected patients [101,102].

6. THERAPY The best intervention for mitochondrial toxicity is avoiding the responsible NRTI. Switching away from stavudine has improved lipoatrophy in randomized studies, but the fat gain was very small [103]. On the other hand, eliminating protease inhibitors has not led to objective improvement of lipoatrophy [50–53] consistent with a predominant effect of NRTIs in the pathogenesis of this condition. Supplementation of thiamine, riboﬂavin, and l-carnitine has been recommended as a treatment of NRTI-induced lactic acidosis [104]. Clinical improvement with riboﬂavin [105] and thiamine [106] was reported in individual cases of NRTI-induced lactic acidosis. These interventions were also recommended in patients with inherited mutations in mtDNA but failed to show effectiveness in vitro and in clinical studies [107,108]. In one study, l-acetylcarnitine substitution was found to improve subcutaneous neurites in NRTI-induced distal symmetric polyneuropathy [109]. Rosiglitazone did not increase subcutaneous

282

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

adipose tissue in most studies of HIV-infected persons with lipoatrophy, but unexpectedly increased serum cholesterol [110,111]. In keeping with its mechanism, rosiglitazone improved insulin sensitivity [110,111]. The relationship between respiratory chain dysfunction and pyrimidine metabolism makes pyrimidine precursors an attractive candidate to prevent and treat NRTI-related mitochondrial toxicity. A pyrimidine precursor that can replenish pyrimidine pools distal from DHODH is uridine (Figure 1). Indeed, neuronal cells exposed to zalcitabine are rescued from death and improve in proliferation and neurite outgrowth if uridine is provided [112]. Uridine also completely reversed the hematopoietic toxicity of zidovudine on human blood progenitor cells [113]. Similar strategies in mouse models of zidovudine-induced bone marrow suppression reversed anemia and leucopoenia and increased peripheral reticulocytes and bone marrow cellularity [113,114]. In models of mitochondrial steatohepatitis, uridine was not only able to prevent cell death, but also to prevent the onset of mtDNA depletion, thereby improving the expression of mtDNA-encoded respiratory chain subunits and moderating lactate production and hepatic steatosis [8,30]. A recent trial in HIV-infected subjects showed that NucleomaxX, a dietary uridine supplement with high bioavailability of uridine [115], rapidly normalizes hepatic functions despite continued zidovudine or stavudine-containing HAART, as measured noninvasively with a [13 C]methionine breath test [116] (see Chapter 22). Interestingly, in the absence of uridine it takes considerably longer for mtDNA depletion to develop (weeks) than for uridine to revert such mitochondrial toxicity (days) [8,27]. This relatively quick therapeutic effect of uridine relative to the more prolonged development of mitochondrial dysfunction may allow for intermittent uridine dosing in order to “reset the mitochondrial clock.” Importantly, the effect of uridine was dose dependent and only improved mitochondrial toxicity caused by pyrimidine NRTIs, consistent with a competitive mechanism of action. Uridine was also shown to prevent and reverse the lipoatrophic phenotype induced by NRTIs to adipocytes, as measured by apoptosis, loss of lipids, mtDNA depletion, and loss of mtDNA-encoded respiratory chain subunits and mitochondrial membrane potential [27]. A recent randomized placebo-controlled double-blind trial has evaluated oral uridine in the form of NucleomaxX in the treatment of lipoatrophy. NucleomaxX rapidly improved subcutaneous fat despite the fact that HAART containing zidovudine or stavudine was unchanged [117]. Several phase I and phase II trials show that oral and intravenous uridine is well tolerated by humans [9]. Theoretically, high-dose uridine could antagonize the antiretroviral activity of NRTI by competing with nucleoside analogs at HIV reverse transcriptase. However, this was not observed in phenotypic HIV-resistance assays and animals [113,114,118]. Human trials have not found uridine supplementation in the form of NucleomaxX to interfere with the efﬁcacy of HAAT [117,119]. This indicates a differential action of uridine on the mitochondrial and antiretroviral replication enzymes, possibly due to differences

REFERENCES

283

between the HIV reverse transcriptase and polymerase γ in selecting the natural nucleotide over the activated NRTI. Alternative explanations include separate regulation of mitochondrial and cytoplasmic dNTP pools either at the level of mitochondrial transport [120] or by the presence of disparate kinases in the cytoplasmic and mitochondrial compartments [121]. The current safety and efﬁcacy data justify the dietary supplementation of uridine in the form of NucleomaxX in patients who have mitochondrial side effects and cannot be switched to a less toxic antiretroviral regimen.

REFERENCES 1. Lewis W, Dalakas MC. Mitochondrial toxicity of antiviral drugs. Nat Med. 1995; 1:417–422. 2. Brinkman K, Smeitink JA, Romijn JA, Reiss P. Mitochondrial toxicity induced by nucleoside-analogue reverse-transcriptase inhibitors is a key factor in the pathogenesis of antiretroviral-therapy-related lipodystrophy. Lancet . 1999; 354:1112–1115. 3. Kakuda TN. Pharmacology of nucleoside and nucleotide reverse transcriptase inhibitor-induced mitochondrial toxicity. Clin Ther. 2000; 22:685–708. 4. Lewis W, Day BJ, Copeland WC. Mitochondrial toxicity of NRTI antiviral drugs: an integrated cellular perspective. Nat Rev Drug Discov. 2003; 2:812–822. 5. Pessayre D, Berson A, Fromenty B, Mansouri A. Mitochondria in steatohepatitis. Semin Liver Dis. 2001; 21:57–69. 6. L¨ofﬂer M, J¨ockel J, Schuster G, Becker C. Dihydroorotate–ubiquinone oxidoreductase links mitochondria in the biosynthesis of pyrimidine nucleotides. Mol Cell Biochem. 1997; 174:125–129. 7. Gattermann N, Dadak M, Hofhaus G, et al. Severe impairment of nucleotide synthesis through inhibition of mitochondrial respiration. Nucleosides Nucleotides Nucleic Acids. 2004; 23:1275–1279. 8. Walker UA, Venhoff N, Koch E, Olschweski M, Schneider J, Setzer B. Uridine abrogates mitochondrial toxicity related to nucleoside analogue reverse transcriptase inhibitors in HepG2 cells. Antivir Ther. 2003; 8:463–470. 9. Walker UA, Venhoff N. Uridine in the prevention and treatment of NRTI-related mitochondrial toxicity. Antivir Ther. 2005; 10:M117–M123. 10. Fox RI. Mechanism of action of leﬂunomide in rheumatoid arthritis. J Rheumatol . 1998; 53(suppl): 20–26. 11. Cherwinski HM, Cohn RG, Cheung P, et al. The immunosuppressant leﬂunomide inhibits lymphocyte proliferation by inhibiting pyrimidine biosynthesis. J Pharmacol Exp Ther. 1995; 275:1043–1049. 12. Oren M. Regulation of the p53 tumor suppressor protein. J Biol Chem. 1999; 274:36031–36034. 13. Laliberte J, Yee A, Xiong Y, Mitchell BS. Effects of guanine nucleotide depletion on cell cycle progression in human T lymphocytes. Blood . 1998; 91:2896–2904. 14. Schuler M, Green DR. Mechanisms of p53-dependent apoptosis. Biochem Soc Trans. 2001; 29:684–688.

284

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

15. Bodnar AG, Cooper JM, Leonard JV, Schapira AH. Respiratory-deﬁcient human ﬁbroblasts exhibiting defective mitochondrial DNA replication. Biochem J . 1995; 305:817–822. 16. Martin JL, Brown CE, Matthews-Davis N, Reardon JE. Effects of antiviral nucleoside analogs on human DNA polymerases and mitochondrial DNA synthesis. Antimicrob Agents Chemother . 1994; 38: 2743–2749. 17. Lynx MD, McKee EE. 3 -Azido-3 -deoxythymidine (AZT) is a competitive inhibitor of thymidine phosphorylation in isolated rat heart and liver mitochondria. Biochem Pharmacol . 2006; 72:239–243. 18. Mandel H, Szargel R, Labay V, et al. The deoxyguanosine kinase gene is mutated in individuals with depleted hepatocerebral mitochondrial DNA. Nat Genet . 2001; 29:337–341. 19. Saada A, Shaag A, Mandel H, Nevo Y, Eriksson S, Elpeleg O. Mutant mitochondrial thymidine kinase in mitochondrial DNA depletion myopathy. Nat Genet . 2001; 29:342–344. 20. Barile M, Valenti D, Passarella S, Quagliariello E. 3 -Azido-3 -deoxythmidine uptake into isolated rat liver mitochondria and impairment of ADP/ATP translocator. Biochem Pharmacol. 1997; 53:913–920. 21. Hobbs GA, Keilbaugh SA, Rief PM, Simpson MV. Cellular targets of 3 -azido-3 deoxythymidine: an early (non-delayed) effect on oxidative phosphorylation. Biochem Pharmacol. 1995; 50:381–390. 22. Barile M, Valenti D, Hobbs GA, et al. Mechanisms of toxicity of 3 -azido-3 deoxythymidine: its interaction with adenylate kinase. Biochem Pharmacol. 1994; 48:1405–1412. 23. Walker UA, Setzer B, Venhoff N. Increased long-term mitochondrial toxicity in combinations of nucleoside analogue reverse-transcriptase inhibitors. AIDS . 2002; 16:2165–2173. 24. Lambert JS, Seidlin M, Reichman RC, et al. 2 ,3 -Dideoxyinosine (ddI) in patients with the acquired immunodeﬁciency syndrome or AIDS-related complex: a phase I trial. N Engl J Med. 1990; 322:1333–1340. 25. Mallon PW, Miller J, Cooper DA, Carr A. Prospective evaluation of the effects of antiretroviral therapy on body composition in HIV-1-infected men starting therapy. AIDS . 2003; 17:971–979. 26. Walker UA, B¨auerle J, Laguno M, et al. Depletion of mitochondrial DNA in liver under antiretroviral therapy with didanosine, stavudine, or zalcitabine. Hepatology. 2004; 39:311–317. 27. Walker UA, Auclair M, Lebrecht D, Kornprobst M, Capeau J, Caron M. Uridine abrogates the adverse effects of antiretroviral pyrimidine analogues on adipose cell functions. Antivir Ther. 2006; 11:25–34. 28. Mallon PW, Unemori P, Sedwell R, et al. In vivo, nucleoside reverse-transcriptase inhibitors alter expression of both mitochondrial and lipid metabolism genes in the absence of depletion of mitochondrial DNA. J Infect Dis. 2005; 191:1686–1696. 29. Galluzzi L, Pinti M, Troiano L, et al. Changes in mitochondrial RNA production in cells treated with nucleoside analogues. Antivir Ther. 2005; 10:191–195. 30. Lebrecht D, Vargas Infante YA, Setzer B, Kirschner J, Walker UA. Uridine supplementation antagonizes zalcitabine-induced microvesicular steatohepatitis in mice. Hepatology. 2007; 15:72–79.

REFERENCES

285

31. Lai KK, Gang DL, Zawacki JK, Cooley TP. Fulminant hepatic failure associated with 2 ,3 -Dideoxyinosine (ddI). Ann Intern Med. 1991; 115:283–284. 32. Gisolf EH, Dreezen C, Danner SA, Weel JL, Weverling GJ (Prometheus Study Group). Risk factors for hepatotoxicity in HIV-1-infected patients receiving ritonavir and saquinavir with or without stavudine. Clin Infect Dis. 2000; 31:1234–1239. 33. Fortgang IS, Belitsos PC, Chaisson RE, Moore RD. Hepatomegaly and steatosis in HIV-infected patients receiving nucleoside analog antiretroviral therapy. Am J Gastroenterol . 1995; 90:1433–1436. 34. Morris AA, Taanman JW, Blake J, et al. Liver failure associated with mitochondrial DNA depletion. J Hepatol . 1998; 28:556–563. 35. Chariot P, Drogou I, de Lacroix-Szmania I, et al. Zidovudine-induced mitochondrial disorder with massive liver steatosis, myopathy, lactic acidosis, and mitochondrial DNA depletion. J Hepatol . 1999; 30:156–160. 36. Arnaudo E, Dalakas M, Shanske S, Moraes CT, DiMauro S, Schon EA. Depletion of muscle mitochondrial DNA in AIDS patients with zidovudine-induced myopathy. Lancet . 1991; 337:508–510. 37. Bauer AM, Sternfeld T, Horster S, Schunk M, Goebel FD, Bogner JR. Kinetics of lactate metabolism after submaximal ergometric exercise in HIV-infected patients. HIV Med. 2004; 5:371–376. 38. Moore RD, Wong WM, Keruly JC, McArthur JC. Incidence of neuropathy in HIV-infected patients on monotherapy versus those on combination therapy with didanosine, stavudine and hydroxyurea. AIDS . 2000; 14:273–278. 39. Feldman D, Anderson TD. Schwann cell mitochondrial alterations in peripheral nerves of rabbits treated with 2 ,3 -dideoxycytidine. Acta Neuropathol . 1994; 87:71–80. 40. Cui L, Locatelli L, Xie MY, Sommadossi JP. Effect of nucleoside analogs on neurite regeneration and mitochondrial DNA synthesis in PC-12 cells. J Pharmacol Exp Ther. 1997; 280:1228–1234. 41. Simpson DM, Tagliati M. Nucleoside analogue– associated peripheral neuropathy in human immunodeﬁciency virus infection. J Acquir Immune Deﬁc Syndr. 1995; 9:153–161. 42. Moyle GJ, Sadler M. Peripheral neuropathy with nucleoside antiretrovirals: risk factors, incidence and management. Drug Saf. 1998; 19:481–494. 43. Brew BJ, Tisch S, Law M. Lactate concentrations distinguish between nucleoside neuropathy and HIV neuropathy. AIDS . 2003; 17:1094–1096. 44. Joly V, Flandre P, Meiffredy V, et al. Increased risk of lipoatrophy under stavudine in HIV-1-infected patients: results of a substudy from a comparative trial. AIDS . 2002; 16(18): 2447–2454. 45. Chene G, Angelini E, Cotte L, et al. Role of long-term nucleoside-analogue therapy in lipodystrophy and metabolic disorders in human immunodeﬁciency virus-infected patients. Clin Infect Dis. 2002; 34:649–657. 46. Walker UA, Bickel M, L¨utke Volksbeck SI, et al. Evidence of nucleoside analogue reverse transcriptase inhibitor-associated genetic and structural defects of mitochondria in adipose tissue of HIV-infected patients. J Acquir Immune Deﬁc Syndr. 2002; 29:117–121.

286

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

47. Cherry CL, Gahan ME, McArthur JC, Lewin SR, Hoy JF, Wesselingh SL. Exposure to dideoxynucleosides is reﬂected in lowered mitochondrial DNA in subcutaneous fat. J Acquir Immune Deﬁc Syndr. 2002; 30:271–277. 48. McComsey GA, Paulsen DM, Lonergan JT, et al. Improvements in lipoatrophy, mitochondrial DNA levels and fat apoptosis after replacing stavudine with abacavir or zidovudine. AIDS . 2005; 19:15–23. 49. Caron M, Auclair M, Sterlingot H, Kornprobst M, Capeau J. Some HIV protease inhibitors alter lamin A/C maturation and stability, SREBP-1 nuclear localization and adipocyte differentiation. AIDS . 2003; 17:2437–2444. 50. Martinez E, Garcia-Viejo MA, Blanco JL, et al. Impact of switching from human immunodeﬁciency virus type 1 protease inhibitors to efavirenz in successfully treated adults with lipodystrophy. Clin Infect Dis. 2000; 31:1266–1273. 51. Domingo P, Matias-Guiu X, Pujol RM, et al. Switching to nevirapine decreases insulin levels but does not improve subcutaneous adipocyte apoptosis in patients with highly active antiretroviral therapy-associated lipodystrophy. J Infect Dis. 2001; 184:1197–1201. 52. Negredo E, Cruz L, Paredes R, et al. Virological, immunological, and clinical impact of switching from protease inhibitors to nevirapine or to efavirenz in patients with human immunodeﬁciency virus infection and long-lasting viral suppression. Clin Infect Dis. 2002; 34:504–510. 53. Moyle GJ, Baldwin C, Langroudi B, Mandalia S, Gazzard BG. A 48-week, randomized, open-label comparison of three abacavir-based substitution approaches in the management of dyslipidemia and peripheral lipoatrophy. J Acquir Immune Deﬁc Syndr. 2003; 33:22–28. 54. Cot´e HC, Brumme ZL, Craib KJ, et al. Changes in mitochondrial DNA as a marker of nucleoside toxicity in HIV-infected patients. N Engl J Med. 2002; 346:811–820. 55. Miro O, Lopez S, Pedrol E, et al. Mitochondrial DNA depletion and respiratory chain enzyme deﬁciencies are present in peripheral blood mononuclear cells of HIV-infected patients with HAART-related lipodystrophy. Antivir Ther. 2003; 8:333–338. 56. Negredo E, Molt´o J, Burger D, et al. Unexpected CD4 cell count decline in patients receiving didanosine and tenofovir-based regimens despite undetectable viral load. AIDS . 2004; 18:459–463. 57. Setzer B, Schlesier M, Thomas AK, Walker UA. Mitochondrial toxicity of nucleoside analogues in primary human lymphocytes. Antivir Ther. 2005; 10:327–334. 58. Setzer B, Schlesier M, Walker UA. Effects of of didanosine-related depletion of mtDNA in human T lymphocytes. J Infect Dis. 2005; 191:848–855. 59. Sommadossi JP, Carlisle R. Toxicity of 3 -azido-3 -deoxythymidine and 9-(1,3dihydroxy-2-propoxymethyl)guanine for normal human hematopoietic progenitor cells in vitro. Antimicrob Agents Chemother . 1987; 31:452–454. 60. Connor EM, Sperling RS, Gelber R, et al. Reduction of maternal– infant transmission of human immunodeﬁciency virus type 1 with zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group. N Engl J Med. 1994; 331:1173–1180. 61. Gerschenson M, Poirier MC. Fetal patas monkeys sustain mitochondrial toxicity as a result of in utero zidovudine exposure. Ann N Y Acad Sci. 2000; 918:269–281.

REFERENCES

287

62. Divi RL, Leonard SL, Kuo MM, et al. Transplacentally exposed human and monkey newborn infants show similar evidence of nucleoside reverse transcriptase inhibitor-induced mitochondrial toxicity. Environ Mol Mutagen. 2007; 48:201–209. 63. Shiramizu B, Shikuma KM, Kamemoto L, et al. Placenta and cord blood mitochondrial DNA toxicity in HIV-infected women receiving nucleoside reverse transcriptase inhibitors during pregnancy. J Acquir Immune Deﬁc Syndr. 2003; 32:370–374. 64. Divi RL, Walker VE, Wade NA, et al. Mitochondrial damage and DNA depletion in cord blood and umbilical cord from infants exposed in utero to Combivir. AIDS . 2004; 18:1013–1021. 65. The Perinatal Safety Review Working Group. Nucleoside exposure in the children of HIV-infected women receiving antiretroviral drugs: absence of clear evidence for mitochondrial disease in children who died before 5 years of age in ﬁve United States cohorts. J Acquir Immune Deﬁc Syndr. 2000; 25:261–268. 66. European Collaborative Study. Exposure to antiretroviral therapy in utero or early life: the health of uninfected children born to HIV-infected women. J Acquir Immune Deﬁc Syndr. 2003; 32:380–387. 67. Blanche S, Tardieu M, Rustin P, et al. Persistent mitochondrial dysfunction and perinatal exposure to antiretroviral nucleoside analogues. Lancet . 1999; 354:1084–1089. 68. Culnane M, Fowler M, Lee SS, et al. (Pediatric AIDS Clinical Trials Group Protocol 219/076 Teams). Lack of long-term effects of in utero exposure to zidovudine among uninfected children born to HIV-infected women. JAMA. 1999; 281:151–157. 69. Lipshultz SE, Easley KA, Orav EJ, et al. Absence of cardiac toxicity of zidovudine in infants: pediatric pulmonary and cardiac complications of vertically transmitted HIV infection study group. N Engl J Med. 2000; 343:759–766. 70. Noguera A, Fortuny C, Munoz-Almagro C, et al. Hyperlactatemia in human immunodeﬁciency virus-uninfected infants who are exposed to antiretrovirals. Pediatrics. 2004; 114:e598–e603. 71. Giaquinto C, De Romeo A, Giacomet V, et al. Lactic acid levels in children perinatally treated with antiretroviral agents to prevent HIV transmission. AIDS . 2001; 15:1074–1075. 72. Alimenti A, Burdge DR, Ogilvie GS, Money DM, Forbes JC. Lactic acidemia in human immunodeﬁciency virus-uninfected infants exposed to perinatal antiretroviral therapy. Pediatr Infect Dis J . 2003; 22:782–789. 73. Sarner L, Fakoya A. Acute onset lactic acidosis and pancreatitis in the third trimester of pregnancy in HIV-1 positive women taking antiretroviral medication. Sex Transm Infect . 2002; 78:58–59. 74. Mandelbrot L, Kermarrec N, Marcollet A, et al. Case report: nucleoside analogue– induced lactic acidosis in the third trimester of pregnancy. AIDS . 2003; 17:272–273. 75. Olivero OA, Anderson LM, Diwan BA, et al. Transplacental effects of 3 -azido-2 , 3 -dideoxythymidine (AZT): tumorigenicity in mice and genotoxicity in mice and monkeys. J Natl Cancer Inst. 1997; 89:1602–1608. 76. Olivero OA, Shearer GM, Chougnet CA, et al. Incorporation of zidovudine into leukocyte DNA from HIV-1-positive adults and pregnant women, and cord blood from infants exposed in utero. AIDS . 1999; 13:919–925. 77. Diwan BA, Riggs CW, Logsdon D, et al. Multiorgan transplacental and neonatal carcinogenicity of 3 -azido-3 -deoxythymidine in mice. Toxicol Appl Pharmacol . 1999; 161:82–99.

288

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

78. Richardson D, Liou SH, Kahn JO. Uric acid and didanosine compliance in AIDS clinical trials: an analysis of AIDS Clinical Trials Group protocols 116A and 116B/117. J Acquir Immune Deﬁc Syndr. 1993; 6:1212–1223. 79. Becker MA, Roessler BJ. Hyperuricemia and gout. In The Metabolic and Molecular Basis of Inherited Disease (CR Scriver, AL Beaudet, WS Sly, D Valle, eds). New York: McGraw-Hill; 1995. 80. Mineo I, Tarui S. Myogenic hyperuricemia: What can we learn from metabolic myopathies ? Muscle Nerve. 1995; 3:S75–S81. 81. Walker UA, Hoffmann C, Enters M, Thoden J, Behrens G, Mitzel SL. High serum urate in HIV-infected persons: the choice of the antiretroviral drug matters. AIDS . 2006; 20:1556–1558. 82. Fisher EJ, Chaloner K, Cohn DL, et al. The safety and efﬁcacy of adefovir dipivoxil in patients with advanced HIV disease: a randomized, placebo-controlled trial. AIDS . 2001; 15:1695–1700. 83. Cihlar T, Ho ES, Lin DC, Mulato AS. Human renal organic anion transporter 1 (hOAT1) and its role in the nephrotoxicity of antiviral nucleotide analogs. Nucleosides Nucleotides Nucleic Acids. 2001; 20:641–648. 84. Verhelst D, Monge M, Meynard JL, et al. Fanconi syndrome and renal failure induced by tenofovir: a ﬁrst case report. Am J Kidney Dis. 2002; 40:1331–1333. 85. Cote HC, Magil AB, Harris M, et al. Exploring mitochondrial nephrotoxicity as a potential mechanism of kidney dysfunction among HIV-infected patients on highly active antiretroviral therapy. Antivir Ther. 2006; 11:79–86. 86. Izzedine H, Hulot JS, Vittecoq D, et al. Long-term renal safety of tenofovir disoproxil fumarate in antiretroviral-naive HIV-1-infected patients. Data from a double-blind randomized active-controlled multicentre study. Nephrol Dial Transplant. 2005; 20:743–746. 87. Kinai E, Hanabusa H. Renal tubular toxicity associated with tenofovir assessed using urine-beta 2 microglobulin, percentage of tubular reabsorption of phosphate and alkaline phosphatase levels. AIDS . 2005; 19:2031–2033. 88. McComsey GA, Yau L. Asymptomatic hyperlactataemia: predictive value, natural history and correlates. Antivir Ther. 2004; 9:205–212. 89. Cossarizza A. Tests for mitochondrial function and DNA: potentials and pitfalls. Curr Opin Infect Dis. 2003; 16:5–10. 90. Bartley PB, Westacott L, Boots RJ, et al. Large hepatic mitochondrial DNA deletions associated with l-lactic acidosis and highly active antiretroviral therapy. AIDS . 2001; 15:419–420. 91. Lam CW, Lau CH, Williams JC, Chan YW, Wong LJ. Mitochondrial myopathy, encephalopathy, lactic acidosis and stroke-like episodes (MELAS) triggered by valproate therapy. Eur J Pediatr. 1997; 156:562–564. 92. Krahenbuhl S, Brandner S, Kleinle S, Liechti S, Straumann D. Mitochondrial diseases represent a risk factor for valproate-induced fulminant liver failure. Liver. 2000; 20:346–348. 93. Ray AS, Olson L, Fridland A. Role of purine nucleoside phosphorylase in interactions between 2 ,3 -dideoxyinosine and allopurinol, ganciclovir, or tenofovir. Antimicrob Agents Chemother. 2004; 48:1089–1095.

REFERENCES

289

94. Boelaert JR, Dom GM, Huitema AD, Beijnen JH, Lange JM. The boosting of didanosine by allopurinol permits a halving of the didanosine dosage. AIDS . 2002; 16:2221–2223. 95. Harris M, Tesiorowski A, Chan K, et al. Lactic acidosis complicating antiretroviral therapy: frequency and correlates. Antivir Ther. 2000; 5 (suppl 2): 31. 96. Harvie P, Omar RF, Dusserre N, et al. Ribavirin potentiates the efﬁcacy and toxicity of 2 ,3 - dideoxyinosine in the murine acquired immunodeﬁciency syndrome model. J Pharmacol Exp Ther. 1996; 279:1009–1017. 97. Ho ES, Lin DC, Mendel DB, Cihlar T. Cytotoxicity of antiviral nucleotides adefovir and cidofovir is induced by the expression of human renal organic anion transporter 1. J Am Soc Nephrol . 2000; 11:383–393. 98. Tanji N, Tanji K, Kambham N, Markowitz GS, Bell A, D’agati VD. Adefovir nephrotoxicity: possible role of mitochondrial DNA depletion. Hum Pathol . 2001; 32:734–740. 99. De Clercq E. Discovery and development of BVDU (brivudin) as a therapeutic for the treatment of herpes zoster. Biochem Pharmacol . 2004; 68:2301–2315. 100. Mallon PW, Miller J, Kovacic JC, et al. Effect of pravastatin on body composition and markers of cardiovascular disease in HIV-infected men: a randomized, placebo-controlled study. AIDS . 2006; 20:1003–1010. 101. Hadigan C, Rabe J, Grinspoon S. Sustained beneﬁts of metformin therapy on markers of cardiovascular risk in human immunodeﬁciency virus-infected patients with fat redistribution and insulin resistance. J Clin Endocrinol Metab. 2002; 87:4611–4615. 102. Martinez E, Domingo P, Ribera E, et al. Effects of metformin or gemﬁbrozil on the lipodystrophy of HIV-infected patients receiving protease inhibitors. Antivir Ther. 2003; 8:403–410. 103. Carr A, Workman C, Smith DE, et al. Abacavir substitution for nucleoside analogs in patients with HIV lipoatrophy: a randomized trial. JAMA. 2002; 288(2): 207–215. 104. Brinkman K, Vrouenraets S, Kauffman R, Weigel H, Frissen J. Treatment of nucleoside reverse transcriptase inhibitor-induced lactic acidosis. AIDS . 2000; 14:2801–2802. 105. Fouty B, Frerman F, Reves R. Riboﬂavin to treat nucleoside analogue-induced lactic acidosis. Lancet . 1998; 352(9124): 291–292. 106. Schramm C, Wanitschke R, Galle PR. Thiamine for the treatment of nucleoside analogue-induced severe lactic acidosis. Eur J Anaesthesiol . 1999; 16(10): 733–735. 107. Venhoff N, Setzer B, Lebrecht D, Walker UA. Dietary supplements in the treatment of NRTI-related mitochondrial toxicity. AIDS . 2002; 16:800–802. 108. Walker UA, Byrne E. The therapy of respiratory chain encephalomyopathy: a critical review of the past and current perspective. Acta Neurol Scand . 1995; 92:273–280. 109. Hart AM, Wilson AD, Montovani C, et al. Acetyl-l-carnitine: a pathogenesis based treatment for HIV-associated antiretroviral toxic neuropathy. AIDS . 2004; 18:1549–1560. 110. Sutinen J, Hakkinen AM, Westerbacka J, et al. Rosiglitazone in the treatment of HAART-associated lipodystrophy: a randomized double-blind placebo-controlled study. Antivir Ther. 2003; 8:199–207. 111. Carr A, Workman C, Carey D, et al. No effect of rosiglitazone for treatment of HIV-1 lipoatrophy: randomised, double-blind, placebo-controlled trial. Lancet . 2004; 363:429–438.

290

LIPOATROPHY AND OTHER MANIFESTATIONS OF ANTIRETROVIRALS

112. Keilbaugh SA, Hobbs GA, Simpson MV. Anti-human immunodeﬁciency virus type 1 therapy and peripheral neuropathy: prevention of 2 ,3 -dideoxycytidine toxicity in PC12 cells, a neuronal model, by uridine and pyruvate. Mol Pharmacol . 1993; 44:702–706. 113. Sommadossi JP, Carlisle R, Schinazi RF, Zhou Z. Uridine reverses the toxicity of 3 -azido-3 -deoxythymidine in normal human granulocyte– macrophage progenitor cells in vitro without impairment of antiretroviral activity. Antimicrob Agents Chemother. 1988; 32:997–1001. 114. Calabresi P, Falcone A, St Clair MH, Wiemann MC, Chu SH, Darnowski JW. Benzylacyclouridine reverses azidothymidine-induced marrow suppression without impairment of anti-human immunodeﬁciency virus activity. Blood . 1990; 76:2210–2215. 115. Venhoff N, Zilly M, Lebrecht D, et al. Uridine pharmacokinetics of Mitocnol, a sugar cane extract. AIDS . 2005; 19:739–740. 116. Banasch M, Goetze O, Knyhala K, et al. Uridine supplementation enhances hepatic mitochondrial function in thymidine-analogue treated HIV-infected patients. AIDS . 2006; 20:1554–1556. 117. Sutinen J, Walker UA, Sevastianova K, Klinkes H, H¨akkinen AM, Ristola M, Yki-Jarvinen H. Uridine for the treatment of HAART-associated lipodystrophy: a randomized, double-blind, placebo-controlled trial. Antivir Ther. 2007; 12:97–105. 118. Koch EC, Schneider J, Weiss R, Penning B, Walker UA. Uridine excess does not interfere with the antiretroviral efﬁcacy of nucleoside analogue reverse transcriptase inhibitors. Antivir Ther. 2003; 8:485–487. 119. McComsey GA, O’Riordan M, Setzer B, Lebrecht D, Baron E, Walker UA. Effect of NucleomaxX on fat and blood mitochondrial DNA in d4T-treated subjects with clinical lipoatrophy. Eur J Clin Nutr. 2007;dois10.1038/sj.ejcn.1602793. 120. Rossi L, Seraﬁni S, Schiavano GF, et al. Metabolism, mitochondrial uptake and toxicity of 2 ,3 -dideoxycytidine. Biochem J . 1999; 344:915–920. 121. Gallinaro L, Crovatto K, Rampazzo C, et al. Human mitochondrial 5 -deoxyribonucleotidase: overproduction in cultured cells and functional aspects. J Biol Chem. 2002; 277:35080–35087.

10 NEPHROTOXICITY Alberto Ortiz Fundaci´on Jimenez Diaz, Madrid, Spain

Alberto Tejedor Hospital Gregorio Mara˜non, Madrid, Spain

Carlos Caramelo Fundaci´on Jimenez Diaz, Madrid, Spain

1. Introduction 2. Peculiarities of tubular cells 3. Nephrotoxicity and mitochondria 3.1. Respiratory chain: reactive oxygen species formation 4. Calcineurin inhibitor nephrotoxicity 4.1. Calcineurin inhibitors: mitochondrial dysfunction 4.2. Apoptosis in CsA nephrotoxicity 5. HAART and nephrotoxicity 5.1. Transporters 5.2. HAART and mitochondrial dysfunction 5.3. Nucleotide antiviral drugs and tubular cell apoptosis 6. Other nephrotoxic drugs and mitochondria 7. Future perspectives

292 292 293 293 295 296 297 300 302 303 304 304 305

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

291

292

NEPHROTOXICITY

1. INTRODUCTION Acute kidney injury can be caused by hundreds of drugs, and it is estimated that one-fourth of all acute renal failures result from drug-induced nephrotoxicity, even though it is frequently underdiagnosed [1]. Tubular and endothelial cells are critical targets of nephrotoxic drugs because abundant ion and other solute transporters can bioaccumulate potentially injurious drugs. Many nephrotoxic drugs target mitochondrial function, commonly leading to acute renal failure. In some cases, kidney function may fully recover upon withdrawal of the drug, but in many instances it can evolve into chronic kidney disease requiring renal replacement, either as a result of continuing exposure to the drug (cyclosporin A is an example discussed below) or as a result of a severe acute insult that interferes with recovery mechanisms. Drug-induced acute tubular necrosis is characterized by tubular cell death (reviewed in [2]) and is the primary cause of acute renal failure. The term acute tubular necrosis predates the term apoptosis and does not describe a speciﬁc form of cell death [2]. Rather, the death of tubular cells can proceed through apoptosis or necrosis, although it is often difﬁcult to determine the predominant mode of cell death because histological sections recognize only the absence of tubular cells. The relative contribution of the two mechanisms to the initial tubular cell loss depends on the severity of the insult. Additional mechanisms of drug-induced renal injury include idiosyncratic immune-mediated renal injury leading to acute tubulointerstitial nephritis. If a drug or a metabolite is a substrate for one of the many plasma membrane transporters, it can be accumulated to such an extent that intratubular precipitation can lead to crystalluria, nephrolithiasis, and obstructive renal disease.

2. PECULIARITIES OF TUBULAR CELLS Daily function of the kidneys implies the glomerular production of 150 to 200 L of a plasma ultraﬁltrate essentially free of proteins. Tubular cells modify the composition of this ﬂuid by reabsorbing the bulk of it and by secreting certain molecules. Proximal tubular epithelial cells account for most of the tubular transport of molecules. As an example, over 200 g of NaCl and 1 kg of glucose are reabsorbed from the tubular lumen every day. Proximal tubular cells are enriched in mitochondria, which provide the energy for transport by the many cell membrane transporters (Figure 1). As such, they are highly susceptible to drugs that undermine mitochondrial function. The transporters include several for organic anions, such as numerous drugs, thereby increasing the intracellular concentrations of potentially cytotoxic drugs over the concentration found in other organs, facilitating proximal tubular cell-speciﬁc toxicity, as is the case for certain antiviral drugs.

293

NEPHROTOXICITY AND MITOCHONDRIA Tubular lumen: urinary space MRP2

MRP4 Microvilli

Mitochondria OAT1

Drug fluxes

Figure 1 Schematic representation of the proximal tubular cell. Note the increased cell surface at both the luminal (apical) face, created by microvilli, and the basolateral face, created by membrane invaginations, which serves to increase the transport capacity. Several transporter molecules that facilitate antiviral drug ﬂuxes and nephrotoxicity are indicated. An unusually high number of mitochondria provide the energy required for such intense transmembrane transport. OAT, organic acid transporter; MRP, multidrug-resistant protein. (Courtesy of Alejandro Ortiz.)

3. NEPHROTOXICITY AND MITOCHONDRIA As is the case in other tissues (see Chapters 6 to 9), renal mitochondria are involved in substrate oxidation, ATP generation, and cellular calcium homeostasis. In the kidney, in addition to these general functions, mitochondria in different parts of the nephron have other speciﬁc functions. For example, mitochondria of the proximal tubule activate 25-dihydroxycholecalciferol by 1α hydroxylation, thereby yielding the active metabolite of vitamin D in these cells. These mitochondria also release ammonia required by distal segments of the nephron to secrete protons into the urine. This pivotal position of mitochondria in bioenergetics plus interrelated processes key to renal function explains why different drugs affect nephrotoxicity induction, mediation, or protection. 3.1. Respiratory Chain: Reactive Oxygen Species Formation Electrons leaking from the respiratory chain at the NADH ubiquinone reductase and the ubiquinone cytochrome c reductase levels univalently reduce oxygen

294

NEPHROTOXICITY

to reactive superoxide radicals, which eventually results in H2 O2 , which can reach intracellular concentrations of 10−9 to 10−7 M [3]. Respiration is regulated by ADP availability, which in turn depends on cellular ATP turnover: Increased ATP demand in highly active, aerobically poised cells typically increases reactive oxygen species (ROS) production. Under certain circumstances, ADP availability does not regulate mitochondrial oxidation, and respiration is uncoupled from ADP phosphorylation. For example, uncoupling nitrophenols are nephrotoxins [4]. Respiration is also regulated by local production of nitric oxide (NO) by the mitochondrial isoform of nNOS (mtNOS) [5], which has speciﬁc postransductional modiﬁcations. Under physiological conditions, low NO concentrations (<100 nM) inhibit cytochrome c oxidase activity, thereby decreasing respiration [6]. Higher concentrations of NO, however, are associated with an increase in free radical production, including various combinations of NO with ROS that generate nitrogen radicals such as peroxynitrite, which impairs mitochondrial function (see below) and can trigger apoptosis [7]. The chemiosmotic gradient created by the respiratory chain may be used to generate ATP or to import calcium when cytosolic concentration is elevated. Excessive mtCa2+ , particularly in the presence of oxidative or nitrosic stress, induces assembly and transient opening of a high-permeability transition pore (MPT) by interaction of the nucleotide traslocase (ANT) with other proteins, including voltage-dependent anion channel (VDAC) and cyclophilin D (CyD) [8] (Figure 2; see Chapters 1 and 19). The MPT opening can be transient and not associated with loss of membrane potential. However, under excessive mtCa2+

H++ H++ H++ cc II Q II III III SUST RA TO SUBSTRATE

O2 CO22+H +H22O

ADP ADP CsA ++ Ca++

H22O H O

CyD ci fD

ATP

++ ++

Ca Ca

CyD ci fD

H H++ Ca++ UniPorter VDAC

H2O

ADP

ATP

MPT PTM

ANTc

ANTm

Figure 2 The transient mitochondrial permeability pore (MPT) is formed by the action of calcium when type 1 amine nucleotide translocase (ANT) in its c conformation combines with the voltage-dependent anion channel (VDAC). Interaction of intramitochondrial cyclophilin D (CyD) with ANT facilitates such binding, although this is not essential. CsA binding to CyD prevents its binding to the ANT and MPT pore activation. (See insert for color representation of ﬁgure.)

CALCINEURIN INHIBITOR NEPHROTOXICITY

295

concentrations the pore opens irreversibly, dissipating membrane potential and releasing proapoptotic factors such as cytochrome c. Some ionophores used as antifungal agents (e.g., nystatin) have a circuitous effect on renal tubular epithelial cells [9] where both Na+ and Ca2+ enter the cell via the drug. The plasma membrane sodium pump becomes stimulated, increasing ATP hydrolysis and inducing a sharp increase in oxygen consumption. However, calcium uptake by mitochondria dissipates membrane potential, and ATP availability declines, starving the ion-dependent transmembrane transporters, and the cell dies via osmotic crisis. In addition to an excess of calcium, other xenobiotics that can induce MPT irreversibly include nephrotoxic agents such as high concentrations of nitric oxide, cisplatin, cyclosporine A, tacrolimus, and vancomycin.

4. CALCINEURIN INHIBITOR NEPHROTOXICITY Cyclosporine A (CsA) and tacrolimus (FK506) are two calcineurin inhibitors currently in use as immunosuppresants, although toxic effects limit the therapeutic application of these drugs. In practical terms, CsA can be examined as the archetype of this group of drugs. Although vascular injury has generally been considered a common factor in CsA-induced organ damage [10,11], renal tubular and vascular damage, plus hypertension, are the most relevant side effects limiting utility. CsA causes acute renal damage as well as a chronic tubulointerstitial nephropathy. The occurrence of end-stage renal disease (ESRD) requiring chronic dialysis has been reported in up to 20% of nonrenal transplant recipients. The use of calcineurin inhibitors is a risk factor [12]. Acutely, CsA induces renal vasoconstriction, mediated by an imbalance of vasoconstrictors and vasodilators, which is reversible upon dose reduction [10,11]. Chronic nephrotoxicity due to CsA is characterized by tubular atrophy, loss of tubular cells, obliterative arteriolopathy, and interstitial ﬁbrosis (reviewed in [13]). Histopathological studies have suggested a toxic effect of the drug on both afferent arterioles and tubular epithelial cells. In vessels, vascular smooth muscle cells initially appeared to be the main CsA target, but the later ﬁnding of multiple endothelial effects has suggested a role for endothelial cell toxicity [7,14–16]. More recently, widespread damage to the peritubular capillaries is deemed to be critical in both acute and chronic CsA nephropathy. In addition to vascular effects, tubular cells are also a prominent target of CsA, and the drug kills tubular epithelial cells, leading to tubular atrophy. CsA may also cause epithelial–mesenchymal transition in cultured cells [17]. This could be relevant for the conspicuous interstitial ﬁbrosis characteristic of CsA nephrotoxicity. A considerable literature deals with pathways involved in CsA-induced cellular injury. The toxic mechanisms appear to be well understood in the lymphocyte, but remain less clear in other cell types. In lymphocytes, CsA and tacrolimus bind

296

NEPHROTOXICITY

cytosolic peptidylprolyl cis-trans isomerases (PPIases), designated as cyclophilin and FK-BP, respectively. The resulting CsA-cyclophilin, or tacrolimus-FK-BP, complex binds to and inhibits a calcium/calmodulin-dependent type 2B protein phosphatase, calcineurin (reviewed in [10]). Calcineurin regulates the dephosphorylation and nuclear import of the transcription factor nuclear factor of activated T cells (NFAT). Providing support to the possible existence of a more generalized mechanism, CsA has been shown to inhibit NFAT DNA binding in other types of cells (e.g., endothelial cells) [14]. Although most of the toxic effects appear to be mediated via calcineurin inhibition, recent literature advocates a reevaluation of the role of cyclophilin blockade, suggesting that cyclophilins may be more central to mode of toxicity, not just a signaling step. CsA binds to several cyclophilins, and mitochondrial CyD has been strongly implicated in apoptosis induced by CsA and FK506 [18]. The possibility exists that the relative contribution of cyclophilin and calcineurin would differ in diverse tissues. The precise downstream sequence of toxic signaling has not yet been described completely, but the role of ROS and peroxynitrite appears to be critical [7,10]. Other intracellular targets of CsA include mitochondrial respiration, cellular calcium signaling, protein kinase C, and protein synthesis. However, the signiﬁcance of these intracellular effects for CsA-induced nephrotoxicity remains to be demonstrated. More information on apoptosis and mitochondrial effects of the calcineurin inhibitors is provided in the following sections. Numerous treatments and approaches have been tried in an attempt to block calcineurin-inhibitor toxicity. Despite encouraging preclinical data, however, none of them has proved effective on a long-term basis in humans. Recent promising investigations in experimental models have targeted the protection and regeneration of the peritubular capillaries, but no clinical evidence is yet available (reviewed in [10]). 4.1. Calcineurin Inhibitors: Mitochondrial Dysfunction Early magnetic resonance spectroscopic studies indicated that renal toxicity is precipitated by mitochondrial failure [19]. This technique also showed that renal mitochondria in vivo are more sensitive to injury by CsA than those in other organs [20]. CsA causes direct, acute, and dose-dependent toxicity to proximal tubule mitochondria. Although some authors ﬁnd that there is a detrimental effect on respiratory complex II [21], CsA has no effect on oxygen consumption, substrate oxidation, or ATP synthase activity, although the respiratory quotient is signiﬁcantly reduced. CsA decreases constitutive NO synthesis as well as induced NO synthesis in proximal tubule cells [18], resulting in increased H2 O2 production. The main effect of CsA on proximal tubule mitochondria is inhibition of Ca2+ activation of CyD by the formation of a CsA–CyD complex (Figure 2). CyD, together with ANT, VDAC, and other proteins, forms MPT, resulting in loss of mitochondrial membrane potential. CsA binding to CyD prevents its binding to

CALCINEURIN INHIBITOR NEPHROTOXICITY

297

the ANT and hence prevents MPT activation [22] by increased mtCa2+ . In the short term, this can protect against other nephrotoxic agents whose toxicity is mediated by direct and long-lasting opening of the MPT pore and explains some of the contradictory observations about the protective effect of CsA on different cellular types and under different circumstances [18,23]. However, in the long term, the lack of MPT pore assembly escalates mtCa2+ uptake while maintaining oxidative phosphorylation (Figure 3). The progressive increases in mtCa2+ ﬁnally overwhelms CsA binding to CyD, and the resulting MPT pore opening becomes irreversible [18] with loss of mitochondrial potential and release of cytochrome c and other proapoptotic factors into the cytosol. The initial inhibition of MPT pore assembly by CsA is also observed with analogs able to bind cyclophilin but that lack anticalcineurinic activity. 4.2. Apoptosis in CsA Nephrotoxicity An increased rate of tubular cell apoptosis is observed in humans and animals with CsA nephrotoxicity [11,24]. In addition, several independent groups have shown that CsA induces apoptosis in cultured tubular cells in a dose- and time-dependent manner [18,25–28]. Tacrolimus also induces renal proximal tubule cell apoptosis [18]. It was suggested initially that death receptors might be involved in CsA-induced apoptosis and that CsA increases Fas expression in cultured tubular cells, and increased FasL and Fas expression has been reported in chronic CsA nephrotoxicity [25,29]. However, neutralizing anti-FasL antibodies do not decrease apoptosis induced by CsA, and CsA did not exacerbate FasL-induced cell death [25]. In addition, no activation of caspase-8 was detected and the caspase-8 inhibitor IETD did not prevent CsA-induced apoptosis [25]. Caspase 12 was not processed, arguing against involvement of the endoplasmic reticulum [25]. In contrast, caspase-12 was activated in tubular cells exposed to the nephrotoxin acetaminophen, indicating heterogeneity in the mechanisms that activate lethal pathways in nephrotoxicity. Mitochondrial injury is a key event in CsA-induced apoptosis of renal tubular cells [18,25]. CsA-induced Bax translocation and oligomerization at the mitochondria permeabilized the outer membrane, released cytochrome c and Smac/Diablo into the cytosol [25], and collapsed the transmembrane membrane potential. Bax is a critical mediator of mitochondrial dysfunction in tubular cells induced by CsA, and Bax antisense oligodeoxynucleotides protected against CsA-induced apoptosis and cell death. Caspases-2, -9, and -3 also participate in CsA nephrotoxicity [25], with initiator caspases-2 and -9 activated upstream of caspase-3, but the hierarchy between them in CsA-induced apoptosis is unclear. Since caspase-2 neither processes nor activates executioner caspases directly [30], caspase-2 is probably the initial step in the cascade. The pan-caspase inhibitor zVAD did not prevent Bax traslocation or cytochrome c release, thus placing these events upstream of the activation of most caspases. As zVAD does not inhibit caspase-2 [31], a role for caspase-2 in promoting cytochrome c release was not explored [25]. In other models of

298

NEPHROTOXICITY A d[Ca]/dTIME (pmol/mL.seg)

0.015

0.005 0.000

30 60 90 120150180 TIME (s)

1

B dVOLUME / dTIME (µL/mg.min)

Succ 5 mM

0.010

0.8 0.6

Succ 5 mM

0.4

ADP 2 mM

0.2

CsA

0 0 10 s

1

2

3 4 TIME (min)

25 s

5

6

7

MITOCHONDRIAL CALCIUM

C

0.1 µM

CSA 1 µg/mL

CONTROL

Succinate 5 mM 0

50

100 150 TIME (sec)

200

250

Figure 3 Cyclosporin and mitochondrial calcium ﬂuxes. (A, B) Temporary relationship between changes in mitochondrial calcium inﬂux, expressed as the ﬁrst derivative of mitochondrial calcium concentration against time, and the change in mitochondrial volume, expressed as the ﬁrst derivative of mitochondrial volume, versus time. Addition of 5 mM succinate causes a net calcium inﬂux that reaches a maximum at 30 seconds, subsequently decreasing until it reaches a new steady state by 60 seconds (A). Simultaneously, a gradual change in mitochondrial volume occurs (B). Cyclosporin (CsA) greatly reduces volume inﬂux via mitochondrial pore in response to succinate addition (B). Mitochondrial swelling observed in the presence of CsA is approximately half of that observed in the control condition. (C) CsA increases intramitochondrial calcium by inhibiting the opening of the MPT pore and reducing the acute inﬂux of ﬂuid to the mitochondria.

299

CALCINEURIN INHIBITOR NEPHROTOXICITY

cell death, caspase-2 may induce apoptosis via cleavage of Bid and the subsequent engagement of the mitochondrial pathway [32]. In contrast to the lack of effect on the release of cytochrome c, zVAD prevented the loss of mitochondrial membrane potential [25]. Speciﬁc inhibition of any of the three caspases prevented apoptosis and prolonged cell survival in long-term (7-day) assays [25]. The increase in long-term cell survival indicates that in this model, caspase inhibitors rescue cells from apoptosis and from other eventual forms of cell death. But this is not always the case, and other forms of renal tubular cell apoptosis are not prevented by caspase inhibition; rather, effective caspase inhibition shifts the pathway from apoptosis to necrosis [33,34]. Thus, a picture emerges in which CsA causes MPT-independent mitochondrial outer membrane permeabilization mediated by Bax (with the role of caspase-2 not yet fully characterized), leading to cytochrome c release, which, in turn, activates caspases that further damage the mitochondria and lead to the loss of mitochondrial transmembrane potential (Figure 4). The feedback loop is essential for apoptosis and cell death to proceed. This is one of several models for the participation of mitochondrial injury in apoptosis [35]. This model is consistent with the known effects of CsA on mitochondria via binding to CyD and is not shared by other nephrotoxins, such as acetaminophen, or by drugs that directly undermine mitochondrial function via respiratory inhibition or uncoupling. Although not speciﬁcally studied in CsA-induced apoptosis, caspases may cause mitochondrial respiratory dysfunction. Caspase-3 cleaves components of the electron transport chain, leading to loss of mitochondrial transmembrane potential and increased production of ROS [36]. This results in lipid peroxidation

CsA Casp2??

Bax ODN

Bax translocation and mitochondrial injury

Further mitochondrial injury and loss of MMP

Cytochrome c release Casp 3, Casp 9 inhibitor

Casp9 and casp3 activation

Apoptosis

Figure 4 Intracellular pathways for CsA-induced renal tubular cell apoptosis. (After Justo et al. [25].)

300

NEPHROTOXICITY

and organellar swelling that might facilitate release of apoptosis-inducing factor (AIF) and other pro-apoptotic factors from the mitochondrial intermembrane space [37]. Induction of apoptosis in tubular cells contrasts with the activity of CsA as a potent inhibitor of various forms of apoptosis. In fact, CsA has been shown to be cardioprotective and neuroprotective [35]. Several potential mechanisms of action may account for this effect. CsA binds several cyclophilin family members and other proteins, which could be involved in its pro- and antiapoptotic actions. Indeed, CsA and FK506 protect macrophages against apoptosis through the inhibition of iNOS expression [23]. In contrast to macrophages, CsA/FK506 and NO exert a synergistic pro-apoptotic action in tubular epithelial cells, probably through an increase in the release of mitochondrial apoptotic mediators [18]. This would be deleterious under conditions of renal inﬂammation, such as during the immediate postransplant period. In cultured endothelial cells, CsA was cytotoxic/pro-apoptotic or cytoprotective/antiapoptotic at high or low concentrations, respectively [10]. The biphasic responses to CsA in endothelial cells depends on the interaction of CsA with cyclophilin rather than with calcineurin. Proximal tubules and endothelial cells adapt to the presence of CsA, minimizing the expected damage. Chronic CsA treatment increases mitochondrial concentration of CyD as well as synthesis of chaperon molecules involved in CyD exportation to mitochondria (HsP70). Inhibition of HsP70 with geldanamycin increases endothelial toxicity of CsA. VEGF is also a critical factor in the cytoprotective effect of CsA [14]. Endogenous VEGF protects renal tubular cells against CsA toxicity in cell culture and in vivo [17]. Proximal cells preconditioned with CsA are resistant to tacrolimus toxicity, and vice versa. Cilastatin, an inhibitor of brush border dipeptidases, protected against CsA-induced apoptosis by inhibiting CsA uptake, thereby reducing CsA in mitochondria [38]. It is used clinically because it decreases renal degradation, transport, and toxicity of the intravenous β-lactam antibiotic imipenem in proximal cells. Clinical and experimental studies have reported that administration of the combination imipenem/cilastatin represses CsA-associated nephrotoxicity as well as the nephrotoxicity of vancomycin, cisplatin, and FK-506 (reviewed in [38]). The basis for such broad protection is unclear.

5. HAART AND NEPHROTOXICITY Highly active antiretroviral therapy (HAART) for the treatment of HIV infection was introduced over a decade ago. It is designed to target multiple critical steps in the life cycle of HIV. It usually includes two nucleoside analog HIV reverse transcriptase inhibitors and at least one HIV protease inhibitor or nonnucleoside inhibitor of HIV reverse transcriptase. The reverse transcriptase inhibitor prevents transcription of viral RNA to proviral DNA, and the protease inhibitor prevents the cleavage of viral precursor polypeptides.

301

HAART AND NEPHROTOXICITY

TABLE 1

HAART Drugs and Nephrotoxicitya

Drug Family Acyclic nucleotide phosphonates (transcriptase inhibitors) Protease inhibitors

Drug

Nephrotoxic Effect

Adefovirb Tenofovirc Cidofovird Indinavir

Proximal tubular cell toxicity: Fanconi syndrome, acute renal failure (acute tubular necrosis) Crystalluria and nephrolithiasis may lead to acute renal failure and chronic kidney disease Case reports of acute renal failure (acute tubular necrosis)

Ritonavirc

Source: Roling et al. [39]. a Nucleoside transcriptase inhibitors are generally safe, although there are case reports of tubular dysfunction with didadenosine and lamivudine–stavudine. Nonnucleoside transcriptase inhibitors and fusion inhibitors are generally safe. b No longer used because of nephrotoxicity. c Displays nephrotoxicity mainly when in combination with other potentially nephrotoxic drugs. d Used for treatment of CMV infection in HIV patients.

Renal damage caused by antiretroviral drugs can result in a variety of toxic effects presenting as acute renal failure, tubular dysfunction, kidney stones, or chronic renal disease (Table 1) (reviewed in [39,40]). An incidence of 5.9 cases of acute renal failure per 100 patient-years has been reported. Long-term survival favors an increase in HAART-induced metabolic alterations, including elevations in serum lipid levels, diabetes, and hypertension, which are likely to be associated with an increase in secondary renal damage, such as hypertensive nephrosclerosis and diabetic glomerulopathy. Acyclic nucleotide phosphonates, including adefovir, cidofovir, and tenofovir, are eliminated predominantly via the urine. Renal failure is their dose-limiting toxicity, particularly for adefovir and cidofovir. They induce proximal tubular cell toxicity that can lead to Fanconi syndrome, consisting of glucosuria, tubular proteinuria, inappropriate phosphaturia, aminoaciduria, and bicarbonaturia, alone or in combination. Ensuing acute renal failure has the histological appearance of acute tubular necrosis [41–45]. The high incidence of toxicity prompted discontinuation of clinical trials for adefovir for treatment of HIV infection. Tenofovir was initially reported to be nonnephrotoxic in clinical trials [46] and in cultured proximal tubular epithelial cells [47]. However, cases of tenofovir-induced acute renal failure have been published [42]. These patients frequently had other risk factors, such as preexisting chronic renal disease, cirrhosis, or the use of other potentially nephrotoxic drugs. Renal function usually recovers following drug withdrawal [42]. Recent reports have linked HAART regimens that contain tenofovir to a mild, time-dependent elevation in serum creatinine. The combination of ritonavir and tenofovir may potentiate nephrotoxicity. Cidofovir has activity against a wide array of DNA viruses, including poxvirus (smallpox) and cytomegalovirus (CMV), the latter being germane to HIV patients.

302

NEPHROTOXICITY

Nephrotoxicity is a dose-limiting side effect and in clinical trials led to dose reductions in 25 to 30% of patients [43]. Cidofovir nephrotoxicity is observed primarily in proximal tubular epithelium, and leads to Fanconi’s syndrome and acute renal failure, which may be irreversible [44]. Treatment with cidofovir requires the routine use of prophylactic measures to prevent nephrotoxicity, including hydration, the use of probenecid, and avoidance of other nephrotoxic agents. Renal dysfunction peaks one week following cidofovir administration [45] and is usually at least partially reversible. As discussed below, this time course is consistent with that of cidofovir-induced apoptosis in cultured tubular epithelium. Renal toxicity associated with the use of nucleoside reverse transcriptase inhibitors is uncommon, although they may potentiate the toxicity of nucleotides [48]. However, two protease inhibitors, indinavir and ritonavir, have been associated with nephrotoxicity [39,40]. Indinavir is the most nephrotoxic protease inhibitor. Crystalluria and nephrolithiasis are the most frequent side effects and may lead to acute renal failure in 14 to 33% of patients [39,49]. Interstitial nephritis was demonstrated in renal biopsy specimens; this can have a self-limited course or lead to interstitial ﬁbrosis and renal atrophy [49]. Papillary necrosis is seen occasionally. Risk factors for indinavir stone formation are a urine pH above 6 and concomitant treatment with trimethoprim–sulfamethoxazole, acyclovir, or ritonavir. Maintaining a urinary output of at least 1500 mL/day is recommended as prophylaxis. Other protease inhibitors are relatively safe for the kidney, although one case of nelﬁnavir lithiasis has been reported. Case reports have linked ritonavir use to reversible renal failure [50]. The majority of reported patients had received concomitant medication with potentially nephrotoxic drugs, such as tenofovir or indinavir, or had other underlying renal pathology. Ritonavir nephrotoxicity usually occurs within 3 to 21 days following introduction of the drug. In HIV patients, drug-related toxicity is not always easy to differentiate from the alterations produced by HIV infection itself. In addition to various antiretroviral drugs, which may themselves interact, HIV patients are frequently exposed to other potentially nephrotoxic drugs, including acyclovir, aminoglycoside antibiotics, amphotericin, foscarnet, pentamidine, sulfadiazine, and trimethoprim–sulfamethoxazole [51]. In this regard, incubation of cells in combinations of tenofovir with other antiretroviral drugs increased toxicity [52]. As a concluding remark, in addition to the putative toxic potential of each individual drug, the complex array of interactions between the multiple drugs being used creates new sources of toxicity in the fast-changing ﬁeld of HIV therapy. 5.1. Transporters A critical aspect of proximal tubular epithelial cell toxicity is drug ﬂuxes in and out of the cell (Figure 1). Intracellular drug accumulation as a consequence of such ﬂuxes is involved in the selective cytotoxicity of certain antivirals in these cells. Tubular secretion of the drugs requires uptake into the cell at the

HAART AND NEPHROTOXICITY

303

basal lamina, followed by secretion to the tubular lumen. The renal organic anion transporter 1 (OAT1) in the basolateral membrane mediates the uptake of tenofovir and cidofovir from blood into proximal tubule cells, leading to selective accumulation and toxicity [53,54]. OAT3 also contributes to this transport. Probenecid inhibits OAT1, prevents the uptake of cidofovir by proximal tubular epithelium, and is therefore used to decrease the incidence of clinical nephrotoxicity [53]. Comparable clinical utility of other drugs should be explored because 56% of patients receiving probenecid have side effects that are dose limiting in 7% of these patients [43]. Nonsteroidal anti-inﬂammatory drugs (NSAIDs) efﬁciently inhibit OAT1-speciﬁc transport of adefovir at clinically relevant concentrations, reducing adefovir cytotoxicity without interfering with its anti-HIV activity [55]. However, NSAIDS may promote acute renal failure in patients with risk factors. Multidrug resistance proteins (MRP) 2 and 4 are the only members of the family expressed at the apical membrane of kidney proximal tubules. MRP4 extracts both adefovir and tenofovir from the proximal tubule cells and secretes them into the tubular lumen, but it makes only a limited contribution to the urinary excretion of cidofovir [56]. The luminal efﬂux mechanism of cidofovir remains unclear. Probenecid, dipyridamole, NSAIDS, and nucleoside drugs such as acyclovir are potent MRP4 inhibitors [57,58]. MRP2 efﬁciently transport ritonavir and indinavir, the former behaving as an inhibitor [59]. This effect may facilitate interaction with other nephrotoxic drugs. 5.2. HAART and Mitochondrial Dysfunction HAART nephrotoxicity may result from damage to mitochondrial DNA (mtDNA), as demonstrated for adefovir nephrotoxicity [41]. Adefovir serves as a substrate for DNA polymerase γ, which is responsible for the replication of mtDNA and inhibits mtDNA replication. A similar mechanism of cellular toxicity has been implicated in myopathy caused by a related nucleoside analog, zidovudine. Renal biopsies from patients with acute adefovir-induced renal failure revealed proximal tubule necrosis with dysmorphic and enlarged mitochondria. Enzymatic studies revealed a deﬁciency of mtDNA-encoded enzymes. Single-tubule PCR showed that the amount of mtDNA in the affected tubules was 35 to 64% of that in unaffected tubules or tubules of normal kidney control. However, there were no abnormalities in nuclear DNA–encoded proteins. Mitochondrial membrane potential, oxidative phosphorylation coupling, and ATP synthesis were not modiﬁed by HAART, at least at this time point [60]. Biochemical consequences of mtDNA depletion are fatty acid accumulation, dicarboxylic acid accumulation, lactic acidosis, and ROS damage [61,62]. Mitochondrial dysfunction is also involved in hepatic steatosis, lipodistrophy, myopathy, and peripheral neuropathy observed in patients under HAART [63]. However, the in vitro toxicity index, based on the relative amounts of mtDNA and nuclear DNA, underpredicts the toxicity of some drugs, suggesting that factors other than inhibition of DNA polimerase γ may be responsible for

304

NEPHROTOXICITY

nephrotoxicity [64]. Mitochondrial nephrotoxicity may involve more than one drug and/or pathogenesis. Indeed, kidney mtDNA depletion was associated with HIV infection and concurrent tenofovir/didanosine therapy, but not tenofovir use alone, while kidney mitochondrial ultrastructural abnormalities were seen with tenofovir use [65]. In this regard, tenofovir did not produce signiﬁcant changes in mtDNA levels in cultured human renal proximal tubule epithelial cells, contrary to observations with less nephrotoxic reverse transcriptase inhibitors [66]. 5.3. Nucleotide Antiviral Drugs and Tubular Cell Apoptosis Whereas the cytotoxicity of nucleotides in a variety of cell lines, including proximal renal tubular cells, is well established, the molecular mechanisms of cell death have been evaluated primarily in the case of cidofovir. Cidofovir was found to induce apoptosis speciﬁcally in human proximal tubular epithelial cells in an OAT1-dependent, probenecid-sensitive manner [44]. Apoptosis was evident in the human proximal tubular cell line HK-2 as well as in primary tubular cell cultures. By contrast, human renal ﬁbroblasts, devoid of OAT-1, were resistant. The time course, 5 to 7 days of exposure to reach the peak apoptotic effect, was consistent with that observed in the clinic, and the lethal concentration is within the therapeutic range. Cell death is caspase dependent and can be inhibited by caspase-3 inhibitors and by survival cytokines such as IGF-1 and HGF. Protection by survival factors that increase BclxL expression and by caspase-3 inhibitors points to a mitochondrial mechanism similar to that of CsA. However, the involvement of mitochondria was not speciﬁcally evaluated in this study. The exquisite selectivity of cidofovir for proximal tubular epithelium, and the fact that it is administered in a programmed fashion every 2 to 3 weeks, make it an ideal candidate for proof-of-concept studies of manipulating tubular cell apoptosis in the clinic.

6. OTHER NEPHROTOXIC DRUGS AND MITOCHONDRIA Mitochondrial injury has been implicated in nephrotoxicity induced by other drugs. Cisplatin causes acute renal failure by acute tubular necrosis and may lead to chronic renal failure requiring long-term renal replacement therapy [67]. Proximal tubular epithelial cells accumulate cisplatin to a higher degree than do other cells, explaining the high susceptibility to cisplatin-induced apoptosis. Activation of mitochondrial pathways is important in apoptosis induced by cisplatin. Nephrotoxicity induced by aminoglycoside antibiotics manifests clinically as nonoliguric renal failure, with a slow rise in serum creatinine after several days of treatment [68]. The time course reﬂects the requirement that drugs accumulate in tubular cells. Lysosomes are key sites of aminoglycoside accumulation, and it has been proposed that excessive intralysosomal accumulation of the drug leads to permeabilization of lysosomes and the release of aminoglycosides and

FUTURE PERSPECTIVES

305

lysosomal enzymes. Localization of aminoglycosides in the cytosol would lead to mitochondrial injury, with loss of mitochondrial membrane potential and release of cytochrome c, caspase activation, and apoptosis [69]. In some cases, endogenous products are nephrotoxic (e.g., high concentrations of glucose promote tubular epithelial cell death) [70]. This may account for the increased susceptibility of diabetic patients to acute tubular necrosis in a host of clinical situations [71]. High glucose-induced apoptosis is Bax dependent, suggesting a role for mitochondrial injury [72]. More recently, glucose degradation products have been shown to promote Bax-dependent mitochondrial injury and apoptosis in tubular epithelial cells [73]. 7. FUTURE PERSPECTIVES Tubular epithelial cells, and especially proximal tubular epithelial cells, are targets for nephrotoxic drugs that damage mitochondria. The abundance of mitochondria and the existence of transport mechanisms that facilitate drug accumulation inside these cells are responsible for cytotoxicity. A great deal has been learned about cell damage mechanisms as well as about the intracellular molecular pathways of cell death following injury. The challenge for the future is to design new prophylactic and therapeutic strategies that limit renal injury. The ﬁrst opportunity relates to the use of drugs that decrease the renal accumulation of nephrotoxins. Probenecid is already used to moderate cidofovir nephrotoxicity, but it has frequent side effects, and alternatives need to be developed. Drug combinations that limit the entry of toxic compounds into the cell and facilitate their efﬂux should be studied. Cilastatin remains an enigmatic drug that merits further study. A second opportunity relates to interfering with mechanisms of injury. The acute use of renal survival factors is feasible for drugs such as cidofovir, which is administered once every 2 to 3 weeks. However, speciﬁc delivery of survival factors to renal cells may be required if they are going to be used chronically, in order not to interfere with other physiological processes. Taking advantage of the existence of transporters at the cell membrane to increase the local tubular cell concentration of such drugs is a possibility. An alternative would be to administer tubular localizing drugs that modulate the local expression of intracellular or extracellular survival factors. Acknowledgments The authors’ research has been suported by grants from Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III, Red REDinREN RD 06/0016, FIS 06/0046, Ministerio de Educaci´on y Ciencia SAF 03/884, and Comunidad de Madrid LISCNIC S-BIO 0283/2006, Sociedad Espa˜nola de Nefrologia A.O. was supported by the Programa de Intensiﬁcaci´on de la Actividad Investigadora in the Sistema Nacional de Salud of the Instituto de Salud Carlos III and the Agencia “Pedro La´ın Entralgo” of the Comunidad de Madrid.

306

NEPHROTOXICITY

REFERENCES 1. Fillastre JP, Godin M. Drug-induced nephropathies. In Oxford Textbook of Clinical Nephrology (AM Davison et al., eds). Oxford, UK: Oxford University Press; 1998: 2645–2657. 2. Ortiz A, Justo P, Sanz A, Lorz C, Egido J. Targeting apoptosis in acute tubular necrosis. Biochem Pharmacol. 2003; 66:1589–1594. 3. http://www.cyberlipid.org/perox/oxid0003.htm (accessed January 20, 2007). 4. Koizumi M, Yamamoto Y, Ito Y, et al. Comparative study of toxicity of 4-nitrophenol and 2,4-dinitrophenol in newborn and young rats. J Toxicol Sci. 2001; 26:299–311. 5. Kanai A, Epperly M, Pearce L, et al. Differing roles of mitochondrial nitric oxide synthase in cardiomyocytes and urothelial cells. Am J Physiol Heart Circ Physiol. 2004; 286:H13–H21. 6. Poderoso JJ, Carreras MC, Lisdero C, Riobo N, Schopfer F, Boveris A. Nitric oxide inhibits electron transfer and increases superoxide radical production in rat heart production in rat heart mitochondria nad submitochondrial particles. Arch Biochem Biophys. 1996; 328:85–92. 7. Navarro-Antolin J, Redondo-Horcajo M, Zaragoza C, et al. Role of peroxynitrite in endothelial damage mediated by cyclosporine A. Free Radic Biol Med. 2007; 42:394–403. 8. Vyssokikh MY, Katz A, Rueck A, et al. Adenine nucleotide translocator isoforms 1 and 2 are differently distributed in the mitochondrial inner membrane and have distinct afﬁnities to cyclophilin D. Biochem J. 2001; 358:349–358. 9. Tejedor A, Noel J, Vinay P, Boulanger Y, Gougoux A. La Respiration r´enale est r´egl´e a` la fois por la disponibilit´e d’ADP et de NADH. Med Sci. 1987; 9(suppl): 33(a). 10. Caramelo C, Alvarez-Arroyo MV, Yague S, et al. Cyclosporin A toxicity, and more: vascular endothelial growth factor (VEGF) steps forward. Nephrol Dial Transplant. 2004; 19:285–288. 11. Shihab FS. Cyclosporine nephropathy: pathophysiology and clinical impact. Semin Nephrol. 1996; 16:536–547. 12. Stratta P, Canavese C, Quaglia M, et al. Posttransplantation chronic renal damage in nonrenal transplant recipients. Kidney Int. 2005; 68:1453–1463. 13. Mihatsch MJ, Thiel G, Ryffel B. Histopathology of cyclosporine nephrotoxicity. Transplant Proc. 1988; 20(3 suppl 3): 759–771. 14. Alvarez-Arroyo MV, Yague S, Wenger RM, et al. Cyclophilin-mediated pathways in the effect of cyclosporin A on endothelial cells: role of vascular endothelial growth factor. Circ Res. 2002; 91:202–209. 15. Gallego MJ, Garcia Villalon AL, Lopez Farre AJ, et al. Mechanisms of the endothelial toxicity of cyclosporin A: role of nitric oxide, cGMP, and Ca2+ . Circ Res. 1994; 74:477–484. 16. Navarro-Antolin J, Lopez-Munoz MJ, Klatt P, Soria J, Michel T, Lamas S. Formation of peroxynitrite in vascular endothelial cells exposed to cyclosporine A. FASEB J. 2001; 15:1291–1293. 17. Slattery C, Campbell E, McMorrow T, Ryan MP. Cyclosporine A–induced renal ﬁbrosis: a role for epithelial–mesenchymal transition. Am J Pathol. 2005; 167:395–407.

REFERENCES

307

18. Hortelano S, Castilla M, Torres AM, Tejedor A, Bosca L. Potentiation by nitric oxide of cyclosporin A and FK506-induced apoptosis in renal proximal tubule cells. J Am Soc Nephrol. 2000; 11:2315–2323. 19. Ruiz-Cabello J, Buss WC, Collier SW, Glazer RI, Cohen JS. Changes in ATP after cyclosporin A treatment in a renal epithelial cell line in the rat studied by 31P-NMR spectroscopy. Res Commun Mol Pathol Pharmacol. 1994; 86:3–13. 20. Serkova N, Klawitter J, Niemann CU. Organ-speciﬁc response to inhibition of mitochondrial metabolism by cyclosporine in the rat. Transpl Int. 2003; 16:748–755. 21. Henke W, Jung K. Comparison of the effects of the immunosuppressive agents FK506 and cyclosporin A on rat kidney mitochondria. Biochem Pharmacol. 1993; 46:829–832. 22. Nakagawa T, Shimizu S, Watanabe T, et al. Cyclophilin D–dependent mitochondrial permeability transition regulates some necrotic but not apoptotic cell death. Nature. 2005; 434:652–658. 23. Alvarez Arroyo MV, Suzuki Y, Yague S, et al. Role of endogenous vascular endothelial growth factor in tubular cell protection against acute cyclosporine toxicity. Transplantation. 2002; 74:1618–1624. 24. Hortelano S, Lopez-Collazo E, Bosca L. Protective effect of cyclosporin A and FK506 from nitric oxide–dependent apoptosis in activated macrophages. Br J Pharmacol. 1999; 126:1139–1146. 25. Justo P, Lorz C, Sanz A, Egido J, Ortiz A. Intracellular mechanisms of cyclosporin A–induced tubular cell apoptosis. J Am Soc Nephrol. 2003; 14:3072–3080. 26. Amore A, Emancipator SN, Cirina P. Nitric oxide mediates cyclosporine-induced apoptosis in cultured renal cells. Kidney Int. 2000; 57:1549–1599. 27. Ortiz A, Lorz C, Catal'an MP, Ortiz A, Coca S, Egido J. Cyclosporine A induces apoptosis in murine tubular epithelial cells: role of caspases. Kidney Int Suppl. 1998; 68:S25–S29. 28. Healy E, Dempsey M, Lally C, Ryan MP: Apoptosis and necrosis: mechanisms of cell death induced by cyclosporine A in a renal proximal tubular cell line. Kidney Int. 1998; 54:1955–1966. 29. Yang CW, Faulkner GR, Wahba IM, et al. Expression of apoptosis-related genes in chronic cyclosporine nephrotoxicity in mice. Am J Transplant. 2002; 2:391–399. 30. Guo Y, Srinivasula SM, Druilhe A, Fernandes-Alnemri T, Alnemri ES. Caspase-2 induces apoptosis by releasing proapoptotic proteins from mitochondria, J Biol Chem. 2002; 277:13430–13437. 31. Lassus P, Opitz-Araya X, Lazebnik Y. Requirement for caspase-2 in stress-induced apoptosis before mitochondrial permeabilization. Science. 2002; 297:1352–1354. 32. Bonzon C, Bouchier-Hayes L, Pagliari LJ, Green DR, Newmeyer DD. Caspase2-induced apoptosis requires bid cleavage: a physiological role for bid in heat shock-induced death. Mol Biol Cell. 2006; 17:2150–2157. 33. Lorz C, Justo P, Sanz AB, Egido J, Ortiz A. Role of Bcl-xL in paracetamol-induced tubular epithelial cell death. Kidney Int. 2005; 67:592–601. 34. Justo P, Sanz AB, Sanchez-Ni˜no MD, Winkles JA, Lorz C, Egido J, Ortiz A. Cytokine cooperation in renal tubular cell injury: the role of TWEAK. Kidney Int. 2006; 70:1750–1758. 35. Green DR, Kroemer G. The pathophysiology of mitochondrial cell death. Science. 2004; 305:626–629.

308

NEPHROTOXICITY

36. Ricci JE, Munoz-Pinedo C, Fitzgerald P, et al. Disruption of mitochondrial function during apoptosis is mediated by caspase cleavage of the p75 subunit of complex I of the electron transport chain. Cell. 2004; 117:773–786. 37. Breckenridge DG, Xue D. Regulation of mitochondrial membrane permeabilization by BCL-2 family proteins and caspases. Curr Opin Cell Biol. 2004; 16:647–652. 38. Perez M, Castilla M, Torres AM, Lazaro JA, Sarmiento E, Tejedor A. Inhibition of brush border dipeptidase with cilastatin reduces toxic accumulation of cyclosporin A in kidney proximal tubule epithelial cells. Nephrol Dial Transplant. 2004; 19:2445–2455. 39. Roling J, Schmid H, Fischereder M, Draenert R, Goebel FD. HIV-associated renal diseases and highly active antiretroviral therapy-induced nephropathy. Clin Infect Dis. 2006; 42:1488–1495. 40. Valle R, Haragsim L. Nephrotoxicity as a complication of antiretroviral therapy. Adv Chronic Kidney Dis. 2006; 13:314–319. 41. Tanji N, Tanji K, Kambham N, Markowitz GS, Bell A, D’agati VD. Adefovir nephrotoxicity: possible role of mitochondrial DNA depletion. Hum Pathol. 2001; 32:734–740. 42. Verhelst D, Monge M, Meynard JL, et al. Fanconi syndrome and renal failure induced by tenofovir: a ﬁrst case report. Am J Kidney Dis. 2002; 40:1331–1333. 43. Lalezari JP, Stagg RJ, Kuppermann BD, et al. Intravenous cidofovir for peripheral cytomegalovirus retinitis in patients with AIDS: a randomized, controlled trial. Ann Intern Med. 1997; 126:257–263. 44. Ortiz A, Justo J, Sanz A, et al. Tubular cell apoptosis and cidofovir-induced acute renal failure. Antiviral Ther. 2005; 10:185–190. 45. Vandercam B, Moreau M, Gofﬁn E, Marot JC, Cosyns JP, Jadoul M. Cidofovir-induced endstage renal failure. Clin Infect Dis. 1999; 29:948–949. 46. Schooley RT, Ruane P, Myers RA, et al. Study 902 Team. Tenofovir DF in antiretroviral-experienced patients: results from a 48-week, randomized, double-blind study. AIDS. 2002; 16:1257–1263. 47. Cihlar T, Birkus G, Greenwalt DE, Hitchcock MJM. Tenofovir exhibits low cytotoxicity in various human cell types: comparison with other nucleoside reverse transcriptase inhibitors. Antiviral Res. 2002; 54:37–45. 48. Murphy MD, O’Ilearn M, Chou S. Fatal lactic acidosis and acute renal failure after addition of tenofovir to an antiretroviral regimen containing didanosine. Clin Infect Dis. 2003; 36:1082–1085. 49. Tashima KT, Horowitz JD, Rosen S. Indinavir nephropathy. N Engl J Med. 1997; 336:138–140. 50. Chugh S, Bird R, Alexander EA. Ritonavir and renal failure. N Engl J Med. 1997; 336:138. 51. de Sequera P, Albalate M, Hernandez J, et al. Acute renal failure due to sulphadiazine crystalluria in AIDS patients. Postgrad Med J. 1996; 72:557–558. 52. Vidal F, Domingo JC, Guallar J, et al. In vitro cytotoxicity and mitochondrial toxicity of tenofovir alone and in combination with other antiretrovirals in human renal proximal tubule cells. Antimicrob Agents Chemother. 2006; 50:3824–3824. 53. Cihlar T, Lin DC, Pritchard JB, Fuller MD, Mendel DB, Sweet DH. The antiviral nucleotide analogs cidofovir and adefovir are novel substrates for human and rat renal organic anion transporter 1. Mol Pharmacol. 1999; 56:570–580.

REFERENCES

309

54. Ho ES, Lin DC, Mendel DB, Cihlar T. Cytotoxicity of antiviral nucleotides adefovir and cidofovir is induced by the expression of human renal organic anion transporter 1. J Am Soc Nephrol. 2000; 11:383–393. 55. Mulato AS, Ho ES, Cihlar T. Nonsteroidal anti-inﬂammatory drugs efﬁciently reduce the transport and cytotoxicity of adefovir mediated by the human renal organic anion transporter 1. J Pharmacol Exp Ther. 2000; 295:10–5. 56. Imaoka T, Kusuhara H, Adachi M, Schuetz JD, Takeuchi K, Sugiyama Y. Functional involvement of multidrug resistance associated protein 4 (MRP4/ABCC4) in the renal elimination of the anti-viral drugs, adefovir and tenofovir. Mol Pharmacol. 2006(Nov 16); Epub ahead of print. 57. El-Sheikh AA, van den Heuvel JJ, Koenderink JB, Russel FG. Interaction of nonsteroidal anti-inﬂammatory drugs with multidrug resistance protein (MRP) 2/ABCC2and MRP4/ABCC4-mediated methotrexate transport. J Pharmacol Exp Ther. 2007; 320:229–235. 58. Smeets PH, van Aubel RA, Wouterse AC, van den Heuvel JJ, Russel FG. Contribution of multidrug resistance protein 2 (MRP2/ABCC2) to the renal excretion of p-aminohippurate (PAH) and identiﬁcation of MRP4 (ABCC4) as a novel PAH transporter. J Am Soc Nephrol. 2004; 15:2828–2835. 59. Huisman MT, Smit JW, Crommentuyn KM, et al. Multidrug resistance protein 2 (MRP2) transports HIV protease inhibitors, and transport can be enhanced by other drugs. AIDS. 2002; 16:2295–2301. 60. Nerurkar PV, Pearson L, Frank JE, Yanagihara R, Nerurkar VR. Highly active antiretroviral therapy (HAART)–associated lactic acidosis: in vitro effects of combination of nucleoside analogues and protease inhibitors on mitochondrial funcion and lactic acid production. Cell Mol Biol. 2003; 49:1205–1211. 61. Kakuda TN. Pharmacology of nucleoside and nucleotide reverse transcriptase inhibitor–induced mitochondrial toxicity. Clin Ther. 2000; 22:685–708. 62. Moyle G. Mechanisms of HIV and nucleoside reverse transcriptase inhibitor injury to mitochondria. Antiviral Ther. 2005; 10(suppl 2): M47–M52. 63. Hoschele D. Cell culture models for the investigation of NRTI-induced mitochondrial toxicity: relevance for the prediction of clinical toxicity. Toxicol in Vitro. 2006; 20:535–546. 64. Lee H., Hanes J, Johnoson K.A. Toxicity of nucleoside analogues used to treat AIDS and the selectivity of the mitochondrial DNA polymerase. Biochemistry. 2003; 42:14711–14719. 65. Cote HC, Magil AB, Harris M, et al. Exploring mitochondrial nephrotoxicity as a potential mechanism of kidney dysfunction among HIV-infected patients on highly active antiretroviral therapy. Antiviral Ther. 2006; 11:79–86. 66. Birkus G, Hitchcock MJ, Cihlar T. Assessment of mitochondrial toxicity in human cells treated with tenofovir: comparison with other nucleoside reverse transcriptase inhibitors. Antimicrob Agents Chemother. 2002; 46:716–723. 67. Arany I, Saﬁrstein RL. Cisplatin nephrotoxicity. Semin Nephrol. 2003; 23:460–464. 68. Mingeot-Leclercq MP, Tulkens PM. Aminoglycosides: nephrotoxicity. Antimicrob Agents Chemother. 1999; 43:1003–1012. 69. Servais H, Jossin Y, Van Bambeke F, Tulkens PM, Mingeot-Leclercq MP. Gentamicin causes apoptosis at low concentrations in renal LLC-PK1 cells subjected to electroporation. Antimicrob Agents Chemother. 2006; 50:1213–1221.

310

NEPHROTOXICITY

70. Ortiz A, Ziyadeh FN, Neilson EG. Expression of apoptosis-regulatory genes in renal proximal tubular epithelial cells exposed to high ambient glucose and in diabetic kidneys. J Invest Med. 1997; 45:50–56. 71. Mangano CM, Diamondstone LS, Ramsay JG, Aggarwal A, Herskowitz A, Mangano DT (The Multicenter Study of Perioperative Ischemia Research Group). Renal dysfunction after myocardial revascularization: risk factors, adverse outcomes, and hospital resource utilization. Ann Intern Med. 1998; 128:194–203. 72. Moley KH, Chi MM, Knudson CM, Korsmeyer SJ, Mueckler MM. Hyperglycemia induces apoptosis in pre-implantation embryos through cell death effector pathways. Nat Med. 1998; 4:1421–1424. 73. Justo P, Sanz AB, Egido J, Ortiz A. 3,4-Di-deoxyglucosone-3-ene induces apoptosis in renal tubular epithelial cells. Diabetes. 2005; 54:2424–2429.

11 DRUG EFFECTS IN PATIENTS WITH MITOCHONDRIAL DISEASES Eric A. Schon, Michio Hirano, and Salvatore DiMauro Columbia University Medical Center, New York, New York

1. Mitochondrial diseases: a brief overview 1.1. Diseases due to mutations in mtDNA 1.2. Diseases due to mutations in nDNA 2. Drug toxicity in mitochondrial diseases: direct effects 2.1. Interaction of antibiotics with mitochondria 2.2. Toxicity of aminoglycosides 2.3. Toxicity of linezolid 3. Drug toxicity in mitochondrial diseases: indirect effects 3.1. Dichloroacetate 4. Drug toxicity in mitochondrial diseases: hypothetical effects 4.1. Metformin 4.2. Growth hormone 4.3. Valproic acid and other anticonvulsants 4.4. Statins 5. Potential therapies 5.1. l-arginine 5.2. Ketogenic diet 6. Conclusions

311 312 313 316 316 316 317 318 318 319 319 319 319 320 320 320 320 321

1. MITOCHONDRIAL DISEASES: A BRIEF OVERVIEW For the purposes of this discussion, mitochondrial diseases will be conﬁned to disorders that disrupt the mitochondrial respiratory chain and oxidative Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

311

312

DRUG EFFECTS IN PATIENTS WITH MITOCHONDRIAL DISEASES

phosphorylation [1]. Most commonly, these disorders are divided into two broad groups: those due to mutations in mitochondrial DNA (mtDNA), which are generally inherited maternally, and those due to mutations in nuclear DNA (nDNA), which are inherited in a Mendelian fashion. Because numerous articles and books have summarized the genetic and clinical features of mitochondrial diseases [1], they are described here only brieﬂy, with a focus on disorders for which a toxic drug interaction has been identiﬁed or surmised. 1.1. Diseases Due to Mutations in mtDNA The human mitochondrial genome contains 37 genes that encode two ribosomal RNAs (rRNAs), 22 transfer RNAs (tRNAs), and 13 polypeptides, which are essential components of the respiratory chain and oxidative phosphorylation (see Chapters 1 and 2). Mutations have now been identiﬁed in all 37 genes. Curiously, of the 239 known mutations, 118 (49%) are located in tRNA genes, even though the 22 tRNAs comprise only about 10% of the coding capacity of the genome (Table 1). Point mutations and microdeletions (i.e., deletions of one or two bases) of mtDNA cause a variety of disorders, exacerbated by lack of introns in mtDNA. As expected, mutations in subunits of the respiratory chain/oxidative phosphorylation system cause deﬁciencies conﬁned to individual enzyme complexes. These

TABLE 1 Pathogenic Mutations in Human mtDNA Causing Respiratory Disfunction Gene Product tRNA-Phe 12S rRNA tRNA-Val 16S rRNA tRNA-Leu(UUR) ND1 tRNA-Ile tRNA-Gln tRNA-Met ND2 tRNA-Trp tRNA-Ala tRNA-Asn tRNA-Cys tRNA-Tyr COX I tRNA-Ser(UCN) tRNA-Asp COX II

Number of Mutations 6 5 5 1 23 10 11 2 2 3 4 3 4 3 4 7 10 2 8

Gene Product tRNA-Lys ATPase 8 ATPase 6 COX III tRNA-Gly ND3 tRNA-Arg ND4L ND4 tRNA-His tRNA-Ser(AGY) tRNA-Leu(CUN) ND5 ND6 tRNA-Glu Cyt b tRNA-Thr tRNA-Pro Total

Number of Mutations 12 1 11 6 3 4 1 1 5 3 2 7 10 9 6 20 2 3 239

MITOCHONDRIAL DISEASES: A BRIEF OVERVIEW

313

polypeptide-speciﬁc mutations often cause Leigh syndrome (LS), a devastating encephalopathy due to a failure in oxidative energy production, either because of a defect in proton pumping/electron transport through complexes I to IV of the respiratory chain, or of proton ﬂow back through complex V to drive the synthesis of ATP [2]. Thus, LS can arise from mutations in subunits of complex I (ND1 to ND6), complex IV (COX I to III), and complex V (ATPase 6). Besides LS, mutations in complex I subunits frequently cause Leber’s hereditary optic neuropathy (LHON) [3], a maternally inherited optic neuropathy typically affecting young men. In LHON, the complex I deﬁciency is relatively mild. In contrast, while many mutations have been identiﬁed in cytochrome b (cyt b), the solitary mtDNA-encoded subunit of complex III, few cases of LS due to complex III deﬁciency have been reported, as many patients with cyt b mutations have had pure myopathies due to spontaneous mutations restricted to muscle [4]. Mutations in the 13 mtDNA-encoded polypeptides can have other consequences, giving rise to a spectrum of clinical phenotypes, such as diabetes, dystonia, cardiomyopathy, motor neuron disease, epilepsy, and sideroblastic anemia. One disorder, MELAS (mitochondrial encephalomyopathy, lactic acidosis, and strokelike episodes), deserves special mention. MELAS, a multisystem and ultimately fatal disorder, is a common maternally inherited mitochondrial disease, usually due to mutations in tRNAs (see below) or, less frequently, to mutations in complex I genes, most often in ND5. One might expect that mutations in rRNAs and tRNAs, which are required for the synthesis of all 13 mtDNA-encoded polypeptides, would cause a generalized disruption in overall respiratory chain function, and indeed that is true in most cases. However, even when mitochondrial protein synthesis is compromised, mutations in different protein-synthetic genes can result in completely different disorders. For example, mutations in tRNALeu(UUR) typically cause MELAS, whereas mutations in tRNALys characteristically cause MERRF (myoclonus epilepsy with ragged-red ﬁbers (a hallmark of mitochondrial proliferation in muscle; see Chapter 23) and mutations in tRNAlle often cause cardiopathy. Similarly, while mutations in rRNA genes would be expected to have multisystemic effects due to a reduction in translational capacity, most mutations cause either cardiomyopathy (typically, mutations in 16S rRNA) or aminoglycoside-induced deafness (typically, mutations in 12S rRNA) [5]. 1.2. Diseases Due to Mutations in nDNA Whereas the mitochondrial genome contains only 37 genes, the nuclear genome speciﬁes more than 1300 gene products destined for the mitochondria (Table 2). These include proteins required for general maintenance of the organelle (e.g., organellar morphology and inheritance; protein import, sorting, and stability; mtDNA replication, transcription, and translation) as well as the specialized functions unique to mitochondria (e.g., amino acid, lipid, and intermediate metabolism; respiratory chain and oxidative phosphorylation; apoptosis). Of these 1300+ genes, approximately 100 to 200 are required for the respiratory health of the organelle, and mutations in any one of them could conceivably

314

TABLE 2

DRUG EFFECTS IN PATIENTS WITH MITOCHONDRIAL DISEASES

Gene Products Present in Mammalian Mitochondria

Category

Function

Maintenance functions (644) Protein translation and stability Carriers and transporters Organellar morphology and inheritance Nucleic acid metabolism Protein import and sorting Stress response Specialized functions (783) Respiratory chain and oxidative phosphorylation Lipid metabolism Signal transduction Intermediate metabolism Apoptosis Amino acid and nitrogen metabolism Miscellaneous/unknown Totala a

Number 227 126 104 86 57 44 175 126 118 105 105 62 92 1369

Some gene products have more than one function.

cause a disorder in mitochondrial oxidative energy metabolism. To date, of this potentially large group of genes, mutations have been identiﬁed in fewer than ﬁfty (Table 3), falling into three groups: 1. Mutations in genes encoding structural subunits of the respiratory chain (14 genes). These include mutations in 11 of the 39 nDNA-encoded TABLE 3 Pathogenic Mutations in Human nDNA Causing Respiratory Chain Dysfunction Function Cardiolipin remodeling CoQ function CoQ synthesis Complex I assembly Complex I structure

Complex II structure Complex III assembly Complex III structure Complex IV assembly Complex V assembly Maintenance of mtDNA Translation of mtRNA

Genes TAZ ETFDH APTX, COQ2, PDSS1, PDSS2 NDUFA12L NDUFA1, NDUFA11, NDUFS1, NDUFS2, NDUFS3, NDUFS4, NDUFS6, NDUFS7, NDUFS8, NDUFV1, NDUFV2 SDHA BCS1L UQCRB, UQCRQ COX10, COX15, LRPPRC, SCO1, SCO2, SURF1 ATPAF2 ANT1, DGUOK, MPV17, PEO1, POLG, POLG2, RRM2B, SUCLA2, SUCLG1, TK2, TP DARS2, GFM1, MRPS16, PUS1, TSFM, TUFM

MITOCHONDRIAL DISEASES: A BRIEF OVERVIEW

315

subunits of complex I [6], essentially all of which cause LS; a mutation in two subunits of complex III; and mutations in one of the four subunits of complex II (SDHA), the only respiratory complex that contains no mtDNA-encoded subunits. Interestingly, mutations in SDH subunits B, C, and D have also been found, but rather than causing respiratory chain deﬁciency, all cause tumors [7]. Mutations in three enzymes in the coenzyme Q10 (CoQ10 or ubiquinone) biosynthetic pathway—PDSS1 [8], PDSS2 [9], and COQ2 [8]—have been identiﬁed as causes of primary CoQ10 deﬁciency. In addition, secondary CoQ10 deﬁciency has been linked to mutations in the electron transfer ﬂavoprotein dehydrogenase (ETFDH), which transfers electrons from the β-oxidation pathway to CoQ10 [10] and to mutations in aprataxin (APTX), a protein required for the repair of single-stranded breaks in DNA [11]. 2. Mutations in genes encoding “ancillary” proteins required for the proper assembly and/or functioning of the respiratory chain (10 genes). While mutations have been found in proteins required to assemble complexes I, III, IV, and V, complex IV has suffered the most “hits” to date, with mutations in six assembly genes. Also included in this group is tafazzin (TAZ ), a protein required for the remodeling of cardiolipin, an atypical phospholipid with four acyl tails found in the inner membrane, which is required for proper respiratory chain function [12]. 3. Mutations in genes encoding proteins required for the synthesis of respiratory chain components (15 genes). Most of these mutations cause multiple deletions and/or depletion of mtDNA. This includes mutations in proteins involved in replicating mtDNA (three genes), in providing nucleotides for mtDNA synthesis (seven genes), and in translating the 13 mtDNA-encoded mRNAs on mitoribosomes (ﬁve genes). One gene, MPV17 , which when mutated causes mtDNA depletion [13], encodes a mitochondrial protein unknown function. Mutations in DARS2 , which encodes the mitochondrial aspartyl-tRNA synthetase, were recently found to cause leukoencephalopathy with brain stem and spinal cord involvement and lactate elevation (LBSL) [14]. Finally, respiratory chain function can be compromised as a consequence of mutations in genes that one normally would not associate speciﬁcally with oxidative energy metabolism. An interesting example in this group is a fatal disorder reported recently in which an infant was born with microcephaly, abnormal brain development, optic atrophy, and elevated very-long-chain fatty acids in the plasma (indicative of a peroxisomal disorder), and persistent lactic acidemia (indicative of a mitochondrial disorder). Surprisingly, the patient harbored a mutation in DNM1L, the gene encoding dynamin-like protein 1 (DLP1), a protein required for the ﬁssion of both mitochondria and peroxisomes [15] (see Chapter 1). A second such recent example is a fatal disorder in two siblings with hypertrophic cardiomyopathy, muscular hypotonia, lactic acidosis, and a defect in ATP synthesis in muscle; both infants harbored a homozygous mutation in the muscle-speciﬁc alternatively spliced isoform of the mitochondrial phosphate carrier (SLC25A3 ) that transports inorganic phosphate from the cytosol into the matrix [16].

316

DRUG EFFECTS IN PATIENTS WITH MITOCHONDRIAL DISEASES

2. DRUG TOXICITY IN MITOCHONDRIAL DISEASES: DIRECT EFFECTS 2.1. Interaction of Antibiotics with Mitochondria The widely accepted endosymbiont hypothesis states that eukaryotc mitochondria have a prokaryotic origin [17]. Besides genetic, physiological, and morphological support for this concept, there are also corroborating data from the ﬁeld of toxicology. In particular, it has long been known that antibiotics that affect bacteria can also disrupt mitochondrial function, while leaving the rest of the eukaryotic cell relatively untouched. One of the earliest examples of this “bacterial–mitochondrial” connection involved chloramphenicol (Cm). Blood dyscrasia was recognized as a side effect of chloramphenicol treatment more than 40 years ago [18], although the mechanism of toxicity was unknown at that time. It turns out that Cm inhibits bacterial translation while also inhibiting mitochondrial translation, and for the same underlying reason: Cm binds to a speciﬁc region of the bacterial large subunit (23S) ribosomal RNA. Human mitochondrial 16S rRNA is the homolog of the bacterial 23S rRNA [19], and Cm binds to the region in 16S rRNA that is analogous to the Cm-binding site in bacterial 23S rRNA. A similar state of affairs exists for eperezolid and linezolid, members of a new class of antibiotics called oxazolidinones. These compounds inhibit protein synthesis by binding to the bacterial 23S rRNA in domain V [20], which is the center for peptidyl transfer where the peptide bond is made (it is homologous to the domain V region of mitochondrial 16S rRNA [19]) of the 50S subunit, thereby preventing its association with the 30S subunit and formation of the initiation complex [21], Thus, oxazolidinones inhibit bacterial and mitochondrial translation at the step of initiation by preventing formation of the N -formylmethionyl– tRNA–ribosome–mRNA ternary complex [22]. Aminoglycoside antibiotics such as streptomycin, gentamicin, neomycin, and kanamycin appear to interfere with translation via binding of the drug to the large subunit rRNA [23]. Macrolide antibiotics, such as erythromycin, also bind to the ribosome to inhibit translation, but have multiple mechanisms of action, including interaction with the 23S rRNA [24]. 2.2. Toxicity of Aminoglycosides Given the plethora of bacterial antibiotics that interfere with bacterial translation, and the similarity of bacterial and mitochondrial ribosomes, it stands to reason that some mutations in mitochondrial DNA (especially in the 12S and 16S rRNAs) might confer a particular susceptibility to these types of antibiotics. Indeed, there are now at least two examples of a hypersusceptibility to antibiotics in patients harboring pathogenic mutations in mtDNA. The best known example is the A → G mutation at position 1555 in the 12S rRNA gene that renders patients particularly susceptible to aminoglycoside-induced deafness (AID) [5,23,25]. Besides A1555G, other 12S rRNA mutations associated with AID

DRUG TOXICITY IN MITOCHONDRIAL DISEASES: DIRECT EFFECTS

317

include A827G, T1095C, C1494T, and various mutations in a C-rich region around nt-961. Interestingly, AID can also be caused by mutations in the tRNASer(UCN) gene, most notably A7445G [26]. The second example involves LHON. A few weeks after being treated with erythromycin, a patient harboring the G11778A mutation in the ND4 subunit of complex I (among the most common of the LHON mutations) suffered bilateral vision loss with atrophy of the optic nerves [27]. While late-onset vision loss is the deﬁning feature of LHON, the rapid onset following administration of the drug, and the unusual loss of vision in both eyes simultaneously, implicated a synergistic effect of the drug in the setting of the G11778A mutation as the precipitating cause of blindness. It was determined subsequently that erythromycin can indeed inhibit mitochondrial protein synthesis, and probably exacerbated an underlying predisposition to a bioenergetic crisis [27]. The fact that aminoglycosides unmask the deafness phenotype conferred by the AID-1555 mutation, while erythromycin exacerbates the blindness phenotype conferred by the LHON-11778 mutation, may be teaching us something about the tissue speciﬁcity of these two disorders. If these drugs inhibit mitochondrial translation, one might expect the effect to be general and systemic, yet both drugs appear to reduce the threshold for bioenergetic susceptibility in the respective target tissues. This result implies that mitochondrial toxicity as revealed by antibiotics might be used as a probe of tissue-speciﬁc physiological effects of mitochondria. Work by Moraes and colleagues showing that erythromycin affected mitochondrial function in cybrids harboring the LHON-11778 mutation [27] implies that the effect of the drug is general: what is speciﬁc appears to be the threshold for that effect. It is not clear whether the half-life or the binding afﬁnity to 16S rRNA of, for example, gentamicin, varies in different tissues. Perhaps treatment of cells derived from “susceptible” versus “resistant” tissues with gentamicin or erythromycin, or high-throughput screening of a panel of such cells with a large number of antibiotics, might reveal tissues that are particularly prone to mitochondrial dysfunction in the context of different mtDNA haplotypes or known pathogenic mtDNA mutations. Such an approach might even reveal why certain tissues are particularly affected in mitochondrial disease, such as the high levels of mutation found in the choroid plexus in patients with Kearns–Sayre syndrome (presumably causing the elevated CSF protein found in this disorder) and the frequent occurrence of cervical lipomas in patients with MERRF. 2.3. Toxicity of Linezolid Interestingly, there are at present no examples of an interaction between chloramphenicol and an underlying mtDNA susceptibility mutation. The failure to observe an adverse interaction may be due in part to the fact that Cm is no longer used in many countries, due to its toxicity. In fact, the only interaction between Cm and mitochondria is the observation that a number of mutations in mtDNA confer resistance to Cm rather than susceptibility [28].

318

DRUG EFFECTS IN PATIENTS WITH MITOCHONDRIAL DISEASES

In the case of linezolid, which is used as a drug of last resort against vancomycin-resistant bacteria and has been on the market only since 2000, two mtDNA polymorphisms in the 16S rRNA—G3010A and A2706G [29,30]—have been identiﬁed. However, the pathogenicity of these mutations has yet to be demostrated. The relatively rapid discovery of these potential linezolid-susceptibility mutations in mtDNA is consistent with the recent ﬁnding that linezolid resistance is associated with mutations in bacterial 23S rRNA genes [31].

3. DRUG TOXICITY IN MITOCHONDRIAL DISEASES: INDIRECT EFFECTS 3.1. Dichloroacetate Perhaps the best known indirect effects of a drug used to treat mitochondrial disease are those that develop in patients with MELAS who are treated with dichloroacetate (DCA). DCA is an activator of the mitochondrial pyruvate dehydrogenase complex (PDHC), which catalyzes the production of acetyl-CoA via the irreversible decarboxylation of pyruvate. The activity of PDHC is modulated by the reversible phosphorylation of its E1 subunit by pyruvate dehydrogenase kinase (PDK). DCA binds to, and inhibits PDK, thereby preventing the phosphorylation of E1 and locking PDHC in its active form (i.e., continually decarboxylating pyruvate to acetyl-CoA). In this way, DCA can be used as a lactate-lowering agent, by essentially causing pyruvate to be oxidized within the mitochondria rather than being converted to lactate in the cytosol. In fact, DCA has been shown to reduce lactic acidosis and improve symptoms in patients with mitochondrial disease [32–34]. Although DCA can indeed reduce lactate levels, a recent clinical trial of DCA in MELAS patients has demonstrated that a major side effect of DCA treatment is the development of peripheral neuropathy [35]. DCA helped reduce some symptoms of MELAS, such as the amelioration of persistent headaches and muscle weakness, but all patients on DCA developed mild liver dysfunction, and some had hypocalcemia and peripheral neuropathy [35]. Interestingly, DCA has recently been shown to promote apoptosis in tumors [36]. Since most tumors are glycolytic (the well-known Warburg effect), it had been hypothesized that DCA could shift metabolism in these cells from glycolysis to glucose oxidation. In fact, DCA induced apoptosis in tumor cells, decreased cellular proliferation, up-regulated the potassium voltage-gated channel Kv1.5, and inhibited tumor growth in mice [36]. Notably, normal cells were essentially unaffected by DCA, implying that this compound could be of potential therapeutic value in cancer. Clearly, the potential side effects of DCA, or of other compounds like it that interdict glycolysis, need to be addressed if and when such agents are used therapeutically.

DRUG TOXICITY IN MITOCHONDRIAL DISEASES: HYPOTHETICAL EFFECTS

319

4. DRUG TOXICITY IN MITOCHONDRIAL DISEASES: HYPOTHETICAL EFFECTS Although the literature regarding the untoward effects of drugs and toxins in patients with mitochondrial disease is small, one can envision situations in which such agents could exacerbate the underlying symptoms of mitochondrial disease. 4.1. Metformin Endocrinopathies are common in patients with mitochondrial disease, and include diabetes [usually non-insulin-dependent diabetes mellitus (NIDDM)], growth hormone deﬁciency, hypothyroidism, hypogonadism, and hypoparathyroidism [37]. One of the treatments for hyperglycemia in NIDDM is metformin, an insulin-sensitizing biguanide [38]. Although the exact mechanism of metformin-mediated hepatic glucose reduction is unclear, it may well involve the inhibition of mitochondrial respiratory chain function [38,39]. In fact, lactic acidosis is a well-recognized rare complication of metformin treatment [39]. Thus, in patients with mitochondrial disease and NIDDM, where respiratory chain function is already compromised, treatment of the diabetes with metformin may cause a mitochondrial toxicity similar conceptually to the effects of antibiotics in AID and LHON alluded to above. 4.2. Growth Hormone In a similar vein, one might consider the use of growth hormone (GH) to treat growth hormone deﬁciency, especially in patients with Kearns–Sayre syndrome (KSS), who are typically quite short. However, the bioenergetic demands imposed by GH need to be weighed against the obvious compromised bioenergetic status in such patients. Moreover, it has been reported that treatment of a KSS patient with GH had little effect on the patient’s growth velocity [40]. 4.3. Valproic Acid and Other Anticonvulsants It has been observed repeatedly that children with Alpers’ syndrome (almost invariably due to mutations in mitochondrial DNA polymerase [POLG]) develop acute liver failure when treated with valproic acid [41]. Clinicians should be aware of this frequent complication, because Alpers’ syndrome is characterized by drug-resistant seizures, and valproic acid is an effective and widely used antiepileptic drug that undermines mitochondrial function. Valproic acid should also be avoided in children with carnitine deﬁciency (be it primary or secondary), because it, and other anticonvulsants, such as gabapentin, lamotrigine, phenytoin, tiagabine, topiramate, and vigabatrin, impair carnitine uptake [42]. Consequently, it is advisable that all children on valproic acid be given supplemental amounts of l-carnitine to protect them from developing secondary carnitine deﬁciency due to their medication.

320

DRUG EFFECTS IN PATIENTS WITH MITOCHONDRIAL DISEASES

4.4. Statins Statins reduce cholesterol synthesis by inhibiting the activity of 3-hydroxy-3methylglutaryl coenzyme A (HMG-CoA) reductase. However, HMG-CoA reductase is also required for the synthesis of CoQ10 . For this reason, statins can also inhibit CoQ10 biosynthesis, resulting in CoQ10 -related side effects in a minority of otherwise-normal subjects who take anticholesterol statins prophylactically [43]. Although the mechanism (or mechanisms) leading to “statin myopathy” (exercise intolerance, myalgia, elevated serum CK, rarely acute myoglobinuria) are unclear, and CoQ10 deﬁciency does not seem to play a major role (see Chapter 7), it seems prudent to take supplemental CoQ10 when on chronic treatment with statins, especially at high doses [44]. More germane, patients with primary or secondary CoQ10 deﬁciency are at risk of exacerbating their CoQ10 deﬁciency when taking statins and should be monitored with particular care. 5. POTENTIAL THERAPIES 5.1. l-Arginine A treatment strategy that is unique to MELAS patients is the attempt to mitigate the onset and/or duration of the strokelike episodes in this disorder. It has been suggested that the strokes in MELAS are due to impaired vascular autoregulation in the brain. MELAS is the only mitochondrial disorder in which most ragged-red ﬁbers (indicative of massive mitochondrial proliferation) are positive for cytochrome c oxidase (COX) activity (see Chapter 23). Moreover, the same pattern is also seen in blood vessels in these patients: the COX-positive SSVs (strongly succinate dehydrogenase-positive blood vessels) [45]. The unique pattern of COX-positive RRFs and SSVs in MELAS, coupled with the fact that the heme moieties in COX can bind nitric oxide (NO), led to a hypothesis that compromised vasodilation in this disorder might be due to a “titration” of NO, and that strategies to increase circulating NO might be of value in mitigating the onset and duration of the strokelike episodes [46]. In fact, Koga’s group in Japan has now shown that l-arginine may indeed be of therapeutic value [47–49]. One obvious potential side effect of l-arginine (or any other agent that increases circulating NO levels, such as nitroglycerin) in MELAS is the possible effect of increased vasodilation. Many patients with MELAS, and many asymptomatic and oligosymptomatic individuals harboring mtDNA mutations associated with MELAS (especially the A3243G mutation in tRNALeu(UUR) ), are susceptible to migraine headaches. Thus, treatment with l-arginine, while helpful with regard to strokes, may exacerbate the prevalence, frequency, or course of the migraine episodes. 5.2. Ketogenic Diet In cytoplasmic hybrid (cybrid) cell lines harboring a heteroplasmic population of wild-type and deleted mtDNAs from a patient with Kearns–Sayre syndrome,

REFERENCES

321

the proportion of mutated genomes was reduced signiﬁcantly merely by growing the cells in a medium in which glucose was replaced by ketone precursors (e.g., β-hydroxybutyrate or acetoacetate) [50]. This result implies that a ketogenic diet might be of therapeutic value in “shifting heteroplasmy” in patients with mtDNA deletions (and perhaps some point mutations as well). However, while cybrids containing exclusively wild-type DNA grew well in ketogenic medium, those containing exclusively deleted mtDNA failed to grow at all. Thus, there is a possibility that treatment of patients with a ketogenic diet could have harmful effects, especially in cells containing extremely high amounts of mutation. In the case of KSS, where some cells in the brain appear to have an extremely high mutant load (e.g., the choroid plexus [51] and the cerebellar dentate nucleus [52,53]), a ketogenic diet might actually kill these cells and make the treatment worse than the disease. For this reason, a reasonable approach to a clinical trial with ketogenic medium might involve patients with a tissue-delimited disorder, such as patients with isolated myopathy (e.g., sporadic progressive external ophthalmoplegia).

6. CONCLUSIONS Growing awareness of drug-induced mitochondrial dysfunction should engender increased clinical vigilance in all patients, but especially in patients with mitochondrial diseases. Although the data are relatively sparse, taken in toto they justify the inference that such dysfunction can exacerbate preexisting mitochondrial impairment. Such knowledge helps inform risk-to-beneﬁt considerations inherent in all treatment decisions, and should also foster increased preclinical mitochondrial assessments of nascent drugs during development.

REFERENCES 1. DiMauro S, Schon EA. Mitochondrial respiratory-chain diseases. N Engl J Med. 2003;348:2656–2668. 2. Smeitink JA, van den Heuvel LW, Koopman WJ, Nijtmans LG, Ugalde C, Willems PH. Cell biological consequences of mitochondrial NADH:ubiquinone oxidoreductase deﬁciency. Curr Neurovasc Res. 2004;1:29–40. 3. Carelli V, Ross-Cisneros FN, Sadun AA. Mitochondrial dysfunction as a cause of optic neuropathies. Prog Retin Eye Res. 2004;23:53–89. 4. Andreu AL, Hanna MG, Reichmann H, et al. Exercise intolerance due to mutations in the cytochrome b gene of mitochondrial DNA. N Engl J Med. 1999;341:1037–1044. 5. Fischel-Ghodsian N. Mitochondrial deafness. Ear Hear. 2003;24:303–313. 6. Janssen RJ, Nijtmans LG, van den Heuvel LP, Smeitink JA. Mitochondrial complex: I. Structure, function and pathology. J Inherit Metab Dis. 2006;29:499–515. 7. Favier J, Briere JJ, Strompf L, et al. Network PN: hereditary paraganglioma/pheochromocytoma and inherited succinate dehydrogenase deﬁciency. Hormone Res. 2005;63:171–179.

322

DRUG EFFECTS IN PATIENTS WITH MITOCHONDRIAL DISEASES

8. Mollet J, Giurgea I, Schlemmer D, et al. Prenyldiphosphate synthase, subunit 1 (PDSS1) and OH-benzoate polyprenyltransferase (COQ2) mutations in ubiquinone deﬁciency and oxidative phosphorylation disorders. J Clin Invest. 2007;117 :765–772. 9. Lopez LC, Schuelke M, Quinzii CM, et al. Leigh syndrome with nephropathy and CoQ10 deﬁciency due to decaprenyl diphosphate synthase subunit 2 (PDSS2) mutations. Am J Hum Genet. 2006;79:1125–1129. 10. Gempel K, Topaloglu H, Talim B, et al. The myopathic form of coenzyme Q10 deﬁciency is caused by mutations in the electron-transferring-ﬂavoprotein dehydrogenase (ETFDH) gene. Brain. 2007;130:2037–2044. 11. Quinzii CM, Kattah AG, Naini A, Akman HO, Mootha VK, DiMauro S, Hirano M. Coenzyme Q deﬁciency and cerebellar ataxia associated with an aprataxin mutation. Neurology. 2005;64:539–541. 12. Schlame M, Ren M. Barth syndrome, a human disorder of cardiolipin metabolism. FEBS Lett. 2006;580:5450–5455. 13. Spinazzola A, Viscomi C, Fernandez-Vizarra E, et al. MPV17 encodes an inner mitochondrial membrane protein and is mutated in infantile hepatic mitochondrial DNA depletion. Nat Genet. 2006;38:570–575. 14. Scheper GC, van der Klok T, van Andel RJ, et al. Mitochondrial aspartyl-tRNA synthetase deﬁciency causes leukoencephalopathy with brain stem and spinal cord involvement and lactate elevation. Nat Genet. 2007;39:534–539. 15. Waterham HR, Koster J, van Roermund CW, Mooyer PA, Wanders RJ, JV. L. A lethal defect of mitochondrial and peroxisomal ﬁssion. N Engl J Med. 2007;356:1746–1741. 16. Mayr JA, Merkel O, Kohlwein SD, et al. Mitochondrial phosphate-carrier deﬁciency: a novel disorder of oxidative phosphorylation. Am J Hum Genet. 2007;80:478–484. 17. Martin W, Hoffmeister M, Rotte C, Henze K. An overview of endosymbiotic models for the origins of eukaryotes, their ATP-producing organelles (mitochondria and hydrogenosomes), and their heterotrophic lifestyle. Biol Chem. 2001;382:1521–1539. 18. von Rechenberg HK, Batz K. Drei Falle von Blutdyskrasien nach Chloramphenicolbehandlung. Praxis. 1966;55:214–216. 19. Burk A, Douzery EJP, Springer MS. The secondary structure of mammalian mitochondrial 16S rRNA molecules: reﬁnements based upon a comparative phylogenetic approach. J Mammal Evol. 2002;9:225–243. 20. Matassova NB, Rodnina MV, Endermann R, et al. Ribosomal RNA is the target for oxazolidinones, a novel class of translational inhibitors. RNA. 1999;5:939–946. 21. Colca JR, McDonald WG, Waldon DJ, et al. Cross-linking in the living cell locates the site of action of oxazolidinone antibiotics. J Biol Chem. 2003;278:21972–21979. 22. Swaney SM, Aoki H, Ganoza MC, Shinabarger DL. The oxazolidinone linezolid inhibits initiation of protein synthesis in bacteria. Antimicrob Agents Chemother . 1998;42:3251–3255. 23. Hutchin T, Cortopassi G. Proposed molecular and cellular mechanism for aminoglycoside ototoxicity. Antimicrob Agents Chemother . 1994;38:2517–2520. 24. Karahalios P, Kalpaxis DL, Fu H, Katz L, Wilson DN, Dinos GP. On the mechanism of action of 9-O-arylalkyloxime derivatives of 6-O-mycaminosyltylonolide, a new class of 16-membered macrolide antibiotics. Mol Pharmacol. 2006;70:1271–1280. 25. Selimoglu E. Aminoglycoside-induced ototoxicity. Curr Pharm Des. 2007;13:119– 126.

REFERENCES

323

26. Xing G, Chen Z, Cao X. Mitochondrial rRNA and tRNA and hearing function. Cell Res. 2007;17:227–239. 27. Luca CC, Lam BL, Moraes CT. Erythromycin as a potential precipitating agent in the onset of Leber’s hereditary optic neuropathy. Mitochondrion. 2004;4:31–36. 28. Howell N, Lee A. Sequence analysis of mouse mitochondrial chloramphenicolresistant mutants. Somat Cell Mol Genet. 1989;15:237–244. 29. Palenzuela L, Hahn NM, Nelson RP, Jr., et al. Does linezolid cause lactic acidosis by inhibiting mitochondrial protein synthesis ? Clin Infect Dis. 2005;40:e113–6. 30. Carson J, Cerda J, Chae JH, Hirano M, Maggiore P. Severe lactic acidosis associated with linezolid use in a patient with the mitochondrial DNA A2706G polymorphism. Pharmacotherapy. 2007;27:771–774. 31. Bonora MG, Solbiati M, Stepan E, et al. Emergence of linezolid resistance in the vancomycin-resistant Enterococcus faecium multilocus sequence typing C1 epidemic lineage. J Clin Microbiol. 2006;44:1153–1155. 32. De Stefano N, Matthews PM, Ford B, Genge A, Karpati G, Arnold DL. Short-term dichloroacetate treatment improves indices of cerebral metabolism in patients with mitochondrial disorders. Neurology. 1995;45:1193–1198. 33. Taivassalo T, Matthews PM, De Stefano N, et al. Combined aerobic training and dichloroacetate improve exercise capacity and indices of aerobic metabolism in muscle cytochrome oxidase deﬁciency. Neurology. 1996;47:529–534. 34. Mori M, Yamagata T, Goto T, Saito S, Momoi MY. Dichloroacetate treatment for mitochondrial cytopathy: long-term effects in MELAS. Brain Dev. 2004;26:453–458. 35. Kaufmann P, Engelstad K, Wei Y, et al. Dichloroacetate causes toxic neuropathy in MELAS: a randomized, controlled clinical trial. Neurology. 2006;66:324–330. 36. Bonnet S, Archer SL, Allalunis-Turner J, et al. A mitochondria-K+ channel axis is suppressed in cancer and its normalization promotes apoptosis and inhibits cancer growth. Cancer Cell . 2007;11:37–51. 37. Quade A, Zierz S, Klingmuller D. Endocrine abnormalities in mitochondrial myopathy with external ophthalmoplegia. Clin Invest. 1992;70:396–402. 38. Kirpichnikov D, McFarlane SI, Sowers JR. Metformin: an update. Ann Intern Med. 2002;137:25–33. 39. Owen MR, Doran E, Halestrap AP. Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochem J. 2000;348(pt 3): 607–614. 40. De Sanctis V, Calzolari F, Piazzi A, Berio A. Sindrome di Kearns–Sayre e dif`etto di GH ipoﬁsario [Kearns–Sayre syndrome associated with growth hormone deﬁciency]. Pediatr Med Chir. 2003;25:447–451. 41. Schwabe MJ, Dobyns WB, Burke B, Armstrong DL. Valproate-induced liver failure in one of two siblings with Alpers’ disease. Pediatr Neurol. 1997;16:337–343. 42. Wu SP, Shyu MK, Liou HH, Gau CS, Lin CJ. Interaction between anticonvulsants and human placental carnitine transporter. Epilepsia. 2004;45:204–210. 43. Hargreaves IP, Duncan AJ, Heales SJ, Land JM. The effect of HMG-CoA reductase inhibitors on coenzyme Q10 : possible biochemical/clinical implications. Drug Saf. 2005;28:659–676. 44. Bargossi AM, Grossi G, Fiorella PL, Gaddi A, Di Giulio R, Battino M. Exogenous CoQ10 supplementation prevents plasma ubiquinone reduction induced by HMG-CoA reductase inhibitors. Mol Aspects Med. 1994;15(suppl): S187–S193.

324

DRUG EFFECTS IN PATIENTS WITH MITOCHONDRIAL DISEASES

45. Hasegawa H, Matsuoka T, Goto Y, Nonaka I. Strongly succinate dehydrogenasereactive blood vessels in muscles from patients with mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes. Ann Neurol. 1991;29:601–605. 46. Naini A, Kaufmann P, Shanske S, Engelstad K, De Vivo DC, Schon EA. Hypocitrullinemia in patients with MELAS: an insight into the “MELAS paradox.” J Neurol Sci. 2005;229–230:187–193. 47. Koga Y, Ishibashi M, Ueki I, et al. Effects of l-arginine on the acute phase of strokes in three patients with MELAS. Neurology. 2002;58:827–828. 48. Koga Y, Akita Y, Nishioka J, et al. l-Arginine improves the symptoms of strokelike episodes in MELAS. Neurology. 2005;64:710–712. 49. Koga Y, Akita Y, Junko N, et al. Endothelial dysfunction in MELAS improved by l-arginine supplementation. Neurology. 2006;66:1766–1769. 50. Santra S, Gilkerson RW, Davidson M, Schon EA. Ketogenic treatment reduces deleted mitochondrial DNAs in cultured human cells. Ann Neurol. 2004;56:662–669. 51. Tanji K, Schon EA, DiMauro S, Bonilla E. Kearns–Sayre syndrome: oncocytic transformation of choroid plexus epithelium. J Neurol Sci. 2000;178:29–36. 52. Tanji K, Vu TH, Schon EA, DiMauro S, Bonilla E. Kearns–Sayre syndrome: unusual pattern of expression of subunits of the respiratory chain in the cerebellar system. Ann Neurol. 1999;45:377–383. 53. Tanji K, DiMauro S, Bonilla E. Disconnection of cerebellar Purkinje cells in Kearns–Sayre syndrome. J Neurol Sci. 1999;166:64–70.

PART III ASSESSMENT OF MITOCHONDRIAL FUNCTION IN VITRO AND IN VIVO

325

12 POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH, AND HIGH-RESOLUTION RESPIROMETRY TO ASSESS MITOCHONDRIAL FUNCTION Erich Gnaiger Department of General and Transplant Surgery, D. Swarovski Research Laboratory, Medical University of Innsbruck, Innsbruck, Austria; OROBOROS INSTRUMENTS, Innsbruck, Austria

1. Introduction 2. Polarographic oxygen sensor and traditional oxygraphy 3. High-resolution respirometry: the Oxygraph-2k 3.1. Calibration of polarographic oxygen sensors and oxygen concentration in respiration media at air saturation 3.2. From oxygraph slopes to respiratory flux corrected for background effects 4. Phosphorylation control protocol with intact cells 4.1. Titration steps of the PC protocol 4.2. Experimental example for the PC protocol 4.3. Flux control ratios from the PC protocol 5. Intact cells, permeabilized cells and tissue, or isolated mitochondria 5.1. Intact cells 5.2. Permeabilized cells and tissue 5.3. Isolated mitochondria 6. Titration protocols in permeabilized cells, permeabilized tissue preparations, and isolated mitochondria 7. Multisensor applications

328 329 333 334 337 340 341 342 343 344 344 345 346 346 347

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

327

328

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

1. INTRODUCTION Recent developments in high-resolution respirometry, combining advanced instrumentation with multiple substrate–uncoupler–inhibitor titration protocols, provide standardized and routine analyses of metabolic ﬂux through various mitochondrial pathways. This technology allows assessments of membrane integrity (coupling of oxidative phosphorylation; cytochrome c release), of respiratory inhibition resulting from effects on the phosphorylation or on the electron transport systems, as well as activities of dehydrogenases, and metabolite transport across the inner mitochondrial membrane. These assessments are done via selective permeabilization of the cell membrane while leaving the mitochondrial membranes intact, or via isolation of mitochondria, and can reveal acute drug effects on mitochondria. The use of intact cells limits the application of substrates, inhibitors, and uncouplers to those that are permeable through the plasma membrane. On the other hand, use of intact cells offers the advantage of studying the control of oxidative phosphorylation under more physiological conditions and evaluating drug effects on mitochondrial function that might be mediated through complex cellular signaling pathways. Most applications of high-resolution respirometry use the measurement of oxygen in a closed system, which allows continuous monitoring of respiration in response to titrated compounds. In this chapter, the new methodological standards set by high-resolution respirometry are discussed and compared with traditional polarographic methods. Also covered is potential multiplexing respirometry with sensors for simultaneous measurement of pH (proton ﬂux, related to H+ /O ratios in isolated mitochondria, or acid production by glycolysis in intact cells), mitochondrial membrane potential (via cationic phenylphosphonium probes), nitric oxide (inhibition of respiratory complexes), and cytochrome spectra (redox state). Incorporating an electronically controlled titration-injection micropump provides the beneﬁts of an open system, such as measurements at steady state and feedback control of experimental variables, while retaining the advantages of a closed system. The interested reader is encouraged to consult www.oroboros.at, where details of most of the protocols and assays described here are provided. Mitochondrial dysfunction is implicated in a wide range of pathologies, including drug-induced organ toxicity and other adverse events. This sets new demands on standardization and routine screens for diagnosis of mitochondrial dysfunction, since limitations of traditional methods have frequently discouraged routine evaluation of mitochondrial liabilities within the drug development process [1] or in the clinical laboratory in general [2]. Defects in oxidative phosphorylation are generally studied by the measurement of mitochondrial oxygen consumption, where use of the polarographic oxygen sensor [3] has long replaced the classical manometric (Warburg) apparatus. High-resolution respirometry is, however, a comparatively recent development [4,5], which now provides a widely applicable tool for routine and speciﬁc analyses of mitochondrial function/dysfunction where (1) reliability and quality control are important (clinical studies, drug toxicity),

POLAROGRAPHIC OXYGEN SENSOR AND TRADITIONAL OXYGRAPHY

329

(2) the amount of biological material is limited (cultured cells, tissue biopsies), (3) pathological effects result in reduced respiration, and (4) effects need to be tested at physiological, low intracellular oxygen levels [4–9]. Apart from the oxygraphic assay of cytochrome c oxidase activity [7] and activity measurements of other oxygen-consuming enzymes [10], diagnostic substrate–uncoupler–inhibitor titration protocols in mitochondrial respiratory studies [11–13] yield information on the capacities of metabolic pathways rather than individual enzymes, and thus provide insight into pathological effects on integrated mitochondrial function. It is well established that various defects in oxidative phosphorylation (OXPHOS) can be identiﬁed by the measurement of oxygen consumption (respirometric OXPHOS analysis) in fresh sample preparations that are not detected via more focused enzymatic analysis, via measurement of mitochondrial membrane potential, or via electron microscopic analyses of mitochondrial structure [11]. This is not surprising, since respirometric OXPHOS analysis provides a screening approach that integrates structural and biochemical injuries to mitochondria that reﬂect not only derangement of membrane structure, but also defects of enzyme systems in membrane transport, dehydrogenases, electron transport, and coupled ADP phosphorylation. Although some drugs inhibit speciﬁc mitochondrial enzymes directly and hence yield mitochondrial cytopathies comparable to speciﬁc genetic defects [11,14], many drug-induced mitochondrial toxicities result from multiple impairments that may be more comparable to the complex pathological patterns observed in degenerative diseases [15], aging [16,17], oxidative stress [18,19], ischemia–reperfusion injury [13], and apoptosis [20]. Small changes in cellular respiration, minor alterations in respiratory ﬂux control ratios, and subtle differences in the effects of drugs on respiration, particularly when the response is immediate, may indicate signiﬁcant mitochondrial defects that reﬂect injuries of mitochondrial proteins or membranes. On longer time scales, such perturbations may reﬂect defects of mtDNA or alterations in mitochondrial and/or other cellular signaling cascades. Assessment of OXPHOS in living and permeabilized cells, small amounts of tissue from transgenic animal models, human needle biopsies, and small numbers of isolated mitochondria requires high-resolution respirometry for accurate results. For example, high-resolution respirometry can be performed with less than 1 mg of fresh muscle tissue, fewer than 500.000 cells, or less than 0.05 mg of mitochondrial protein, which is 10-fold less than is required using conventional oxygraphic instruments.

2. POLAROGRAPHIC OXYGEN SENSOR AND TRADITIONAL OXYGRAPHY The principle of respirometry in a closed chamber is based on monitoring oxygen concentration, which declines as the biological sample consumes oxygen. Plotting oxygen concentration over time (Figure 1) yields the oxygraph slope,

330

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k Mito ADP a 10 60

Atr

Mito

DNP 13

b

10

ADP 28

TMPD

60 Atr DNP 19 9

KCN 14

10 µM O2

164 140

2 min

Figure 1 Traces of oxygen concentration over time with traditional oxygraphy in an experiment with mitochondria isolated from mouse liver at 22◦ C: curve a, freshly prepared; curve b, trehalose frozen. Addition of mitochondria, Mito (protein concentration 1 mg/mL); ADP, Atr (atractyloside), DNP, TMPD, and KCN indicate various additions (arrows). Numbers indicate the negative slope (µM/min). (From Yamaguchi et al. [22], with permission from Macmillan. Copyright  2007.)

which is in large measure related to oxygen consumption by the sample but is also confounded by undesired artifacts. The latter include instrument background and are typically minimized by the hardware and/or corrected for by software according to standards established more than 10 years ago [4,5]. The oxygen concentration is measured by a Clark electrode (named after its inventor Leland Clark, a professor of chemistry at Antioch College), which contains a gold or platinum cathode and a Ag/AgCl anode separated by concentrated KCl aqueous solution. Voltage is applied (∼0.6 to 0.8 V) and these two half-cells are separated from the solution being monitored by a O2 -permeant membrane, often Teﬂon, that excludes ions and other potential reductants. Dissolved O2 diffuses from the solution through the membrane and is reduced to water by electrons at the cathode, yielding a hydroxide that forms KOH. The resulting Cl is drawn to the Ag/AgCl anode, where it precipitates, providing a current

POLAROGRAPHIC OXYGEN SENSOR AND TRADITIONAL OXYGRAPHY

331

that is proportional to O2 partial pressure, pO2 , in the experimental solution. These sensors are typically calibrated using air-saturated media, and dithionite for zero pO2 . To convert pO2 to oxygen concentration, the oxygen solubility of the medium must be known, which is a function of temperature and salt concentration. Although the physical conﬁguration varies between various instruments, all require some means of regulating and maintaining constant temperature [3]. Traces of oxygen concentration as a function of time [21–24] (Figure 1) yield information that can be optimized when time derivatives and corrections for side effects are included, although respiratory ﬂux techniques improve resolution substantially (Figure 2). Issues typical of traditional polarographic assessments can be discerned in most such traces and are described, along with their solutions, below. 1. Signal stability. Before addition of mitochondria into the chamber, instability of the oxygen signal indicates the possible magnitude of variation (Figure 1, trace a), which may be due to an initial equilibration process or to irregular sensor behavior. This would then exert a stronger inﬂuence on the initial portions of the experiment, and less so on later sections, when the system becomes pregressively stable. 2. Oxygen consumption of the sensor. The contribution of oxygen-dependent oxygen consumption of the polarographic oxygen sensor is not quantiﬁed or may be within the limits of sensor instability. However, this signal is typically small compared to respiration and can be determined empirically. 3. Linearity of the slope. Within each metabolic state, linearity of the slope of oxygen over time is assumed (e.g., before addition of ADP in Figure 1, trace b), whereas observation of the time course of changes in respiration (resulting in nonlinear slopes) reveals important details of mitochondrial function (continuous loss of respiratory capacity in trace b, Figure 1). 4. Time resolution. Time-dependent responses to addition of substrates or drugs cannot be resolved in the time frame of experiments with high mitochondrial concentrations, due to rapid oxygen depletion of the medium. 5. Limited time for titrations. Oxygen concentrations drops by about 100 µM or about 40% air saturation within about 6 minutes after addition of mitochondria at high concentration (Figure 1), leaving little or no scope for experimental controls, such as stepwise titration of dinitrophenol (DNP) to evaluate the optimum uncoupler concentration at which ﬂux is maximum (Figure 2; optimum uncoupler concentration may be different in mitochondria incubated with various drugs or after pretreatment of mitochondria). 6. Oxygen back-diffusion. After KCN titration, oxygen concentration increases steeply, due to back-diffusion of oxygen into the chamber at about 60% air saturation. In the absence of autoxidation of N ,N ,N ,N tetramethyl-p-phenylenediamine dihydrochloride (TMPD) and ascorbate, back-diffusion would be even higher, indicating that this is a signiﬁcant source of error for measuring respiration (e.g., after inhibition by atractyloside), even at

332

100

[O2] ETS uncoupled

60

yc

90

CP

40

FC

60

80

in

120

RO UT IN Ol E igo m

30

20

LEAK 0 0:10

0 0:20

0:30

0:40

0:50

1:00

1:10

Time [h:min]

A 150

100 ETS uncoupled

60

CP

40

FC

60

UT Ol INE igo my

90

30

20

LEAK

0 0:20

B

80

cin

120

RO

O2 concentration [µM]

[O2]

0:30

0:40

0:50

1:00

0 1:10

Respiration, JV,O2 [pmol·s−1·cm−3]

O2 concentration [µM]

150

Respiration, JV,O [pmol·s−1·cm−3] 2

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

Time [h:min]

Figure 2 High-resolution respirometry and phosphorylation control titration in intact cells. Superimposed plots of oxygen concentration [O2 ] and respiration, calculated as the negative time derivative of oxygen concentration. Simultaneous replicate measurements in the left and right chamber (2 cm3 ) of the Oroboros Oxygraph-2k are shown in each graph. Parental hematopoietic 32D cells were used at 1.1 × 106 to 1.5 × 106 cells/cm3 , suspended in culture medium RPMI at 37◦ C (ROUTINE respiration, R). Oligomycin (2 µg/ml; resting or LEAK state, L) is titrated manually (1 µl). After inhibition of ATP synthase, nonphosphorylating ﬂux declines to a new steady state, although LEAK respiration tends to increase gradually with time. Uncoupler (FCCP) was titrated automatically with the TIP-2k, in steps of 0.1 (1.0) µl, adding 10 (1.0) mM FCCP, corresponding to an increase in the ﬁnal concentration of 0.5 µM FCCP in the Oxygraph-2k chamber. Downward deﬂections of apparent ﬂux are caused by the high oxygen concentration in ethanol and are stronger as the titrated volume is increased. The titration is fast (20 and 30 µl/s), at intervals of 120 (180) s. Maximum uncoupled ﬂux (capacity of the electron transport system, ETS; state E ) was reached at 5.5 and 4.5 µM FCCP, and maximum sample dilution was <1% when respiration was inhibited by higher FCCP concentrations. O2k experiments 2005-04-09 EF-03 (A) and 2005-09-14 EF-01 (B), carried out by participants of Oxygraph-2k courses.

relatively high oxygen levels and at high protein concentrations of mitochondria in the chamber (1 mg/mL). 7. Chemical oxygen consumption. Ascorbate and TMPD in mitochondrial respiration media show a high reactivity with oxygen due to autoxidation, which varies as a function of oxygen concentration. Correction of respiration for autoxidation is required to evaluate cytochrome c oxidase (CcOX) activity [7,25].

HIGH-RESOLUTION RESPIROMETRY: THE OXYGRAPH-2k

333

Attempts to solve these problems by producing microchambers were largely counterproductive: Although the higher sample concentration in a smaller volume yields a higher volume-speciﬁc respiration, thus addressing problems 1 and 7 by increasing the ratio of mitochondrial respiration to signal drift and chemical autoxidation rate, the side effect (problem 2) of sensor consumption per volume increases as much as mitochondrial respiration per volume. Reduced sensor consumption can be achieved by application of microsensors, the stability of which declines with the reduction of cathode area [3]. Limitations 3 to 5 become more severe as sample concentration increases. Problem 6 relative to respiratory activity increases in microchambers due to the unfavorable surface-to-volume ratio. Since applications of microchambers [26–29] are restricted and problematic, or difﬁcult to apply [30], a paradigm shift to a large chamber volume (2 cm3 ) was involved in the development of high-resolution respirometry for routine application with small amounts of sample, taking care to eliminate artifacts. 3. HIGH-RESOLUTION RESPIROMETRY: THE OXYGRAPH-2k Several features distinguish high-resolution respirometry from traditional oxygraphs [5], combined in the new Oxygraph-2k (O2k, OROBOROS INSTRUMENTS, Innsbruck, Austria; Figure 3). The speciﬁcations are unique: the limit of detection of respiratory ﬂux is 1 pmol·s−1·cm−3 (0.001 µM/s) and the limit of detection of oxygen concentration extends to 0.005 µM O2 . For the nonspecialist, the O2k provides robustness and reliability of instrumental performance. With small amounts of sample and correspondingly low respiratory ﬂux per volume, the oxygen capacity of the system is exhausted slowly, allowing sufﬁcient time to evaluate the stability of respiratory activity in each metabolic state and to permit complex titration regimes in intact cells (Figure 2) or in permeabilized cells and tissues [13,16–20]. To increase throughput in research with cell cultures and in the pharmacological arena, user-friendly features make it possible to apply several instruments in parallel, each O2k with two independent chambers (Figure 3). The chambers, sensors, and electronics are shielded by a copper block and stainless steel housing (Figure 3). High long-term signal stability and low noise of the oxygen signal are a basis for online calculation of oxygen ﬂux taken as the negative time derivative of oxygen concentration with high time resolution (Figure 2). Angular insertion of the oxygen sensor into the cylindrical chamber places the membrane-covered cathode of the polarographic oxygen sensor into an optimum position for stirring (Figure 3, inset). All materials in contact with the respiration medium are diffusion tight, with the glass chamber and PVDF or titanium stoppers (vs. Perspex), Viton O-rings, and magnetic stirring bars coated by PVDF or PEEK (vs. Teﬂon). In particular, Teﬂon, which has an oxygen solubility >10-fold that of aqueous media, is not feasible for high-resolution assays, due to long delays of oxygen back-diffusion from the Teﬂon material when oxygen declines in the closed chamber. Fully integrated instrumental control includes

334

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

Peltier temperature control, ±0.001°C Stopper A

O2 Sensor POS B

Window chamber A

RS232 to PC and DatLab 4

O2 sensor POS A

A

Stainless steel housing

B

Preamplifier POS B

Stopper Insulated copper block POS

PVDF stirrer Glass chamber

Stirrer control light B Ion-sensitive electrode plug B

Control light for connection to PC and DatLab Barometric pressure transducer

Figure 3 Oxygraph-2k for high-resolution respirometry. Two identical and independent 2-mL glass chambers (A and B) are housed in an insulated copper block which is maintained at constant temperature by electronic Peltier temperature regulation. Oxygen concentration is recorded continuously by polarographic oxygen sensors (POS) in each chamber, which is sealed by a stopper containing a capillary for extrusion of gas bubbles and insertion of a needle for titrations. Additional holes through the stopper (PVDF) are made for insertion of various electrodes, the signals of which are simultaneously recorded by the DatLab software. The PVDF (or PEEK) stirrers are powered by electically pulsed magnets inserted in the copper block. Data are processed online to calculate oxygen consumption of intact cells, permeabilized tissues, or isolated mitochondria. (Copyright  2005–2008, OROBOROS INSTRUMENTS. Reproduced with permission; www.oroboros.at.)

electronic Peltier temperature regulation (2 to 45◦ C, stability at ±0.001◦ C), stirrer control, and the automatic titration-injection micropump TIP-2k (not shown). Standardized calibration procedures of the oxygen signal, response time of the sensor, and instrumental or chemical background effects provide an experimental basis for high accuracy [5]. 3.1. Calibration of Polarographic Oxygen Sensors and Oxygen Concentration in Respiration Media at Air Saturation The polarographic oxygen sensors (OROBoPOS) are stable for several months without exchange of membrane or electrolyte (Figure 4). This long-term stability of the polarographic oxygen sensors ensures that the O2k is ready to use. The amperometric measurement of oxygen by Clark-type polarographic oxygen sensors yields a current that is converted to a voltage and is strongly

335

HIGH-RESOLUTION RESPIROMETRY: THE OXYGRAPH-2k

R1, relative deviation

0.10

(A)

0.05 0.00 −0.05 −0.10 0

10

A

20 30 Time [days]

40

R0, relative deviation

0.03

50

(B)

0.02 0.01 0.00 −0.01 −0.02 −0.03 0

B

10

20 30 Time [days]

40

50

Figure 4 Stability of the signals of six polarographic oxygen sensors, OROBoPOS, measured in separate chambers at air calibration, R 1 , and zero calibration, R 0 , over a period of >1 month at constant temperature (25◦ C) but in different aqueous media, at stirrer speeds of 750 rpm or 300 rpm and with different stirrers (PEEK, PVDF, and small Viton-covered Teﬂon stirrer bars). Between experiments (isolated mitochondria and cell homogenates with typical substrate + uncoupler + inhibitor titrations), membranes were never exchanged and the sensors were left mounted to the O2k chambers, which were ﬁlled with 70% ethanol. Under such variable experimental conditions, daily air calibration improves the accuracy (A), whereas zero calibrations are not required at a regular basis for routine experiments (B). (A) Relative deviation of R 1 at time t , relative to day 1, is R 1 (t)/R 1 (1) − 1; (B) relative deviation of R 0 is R 0 (t)/R 1 (t) − R 0 (1)/R 1 (1). R 0 (1)/R 1 (1) ranged from 0.02 to 0.14.

inﬂuenced by temperature [3]. Nevertheless, in the range from zero oxygen to pure oxygen at about 1 mM dissolved O2 , modern polarographic instruments are superior to other technologies, such as optical sensors. This imposes high demands on the electronics; the digital resolution is 2 nM, yielding a 500,000-fold dynamic range. Air calibration is conveniently performed in an experimental medium at experimental temperature in the O2k chamber, providing a small gas phase of air and observing the stabilization of the sensor signal as equilibration is reached between gas and the well-stirred aqueous phase

336

Backgr., J °V,O [pmol·s−1·cm−3]

4

J °O2 = 0.027 cO2 − 1.20

3 2

3

0 −1 −2 0

50

100

150

J°O2 = 0.025 cO2 − 1.68

2 1 0

2

1

2

Backgr., J °V,O [pmol·s−1·cm−3]

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

−1 −2

J°O2 = 0.030 cO2 − 2.56

−3 0

200

50

150

200

8

J1°

R1

J2°

150

4

J3°

0

100 −4 50 0 0:00

−8 0:20

0:40

1:00

1:20

J°2

J°1

1:40

2:00

Close c As Reox J°4 J°3

Range [h:min]: 2:00

Close

250 200

8

J1°

R1

J2°

150

4

J° 3

0

100 −4 50 0 0:00

−8 0:20

0:40

1:00

1:20

1:40

2:00

3 2 1

2

200

Residuals, J°V,O [pmol·s−1·cm−3]

250

O2 Slope uncorr. (E) [pmol/(s*ml)]

C

O2 Slope uncorr. (F) [pmol/(s*ml)]

O2 Concentration (E) [nmol/ml]

A

O2 Concentration (F) [nmol/ml]

100

O2 concentration [µM]

O2 concentration [µM]

0

−1 −2 −3 0

50

100

150

200

O2 concentration [µM]

Range [h:min]: 2:00

B

Close J˚1

J˚2

Close J˚3

c As

D

Figure 5 Instrumental background oxygen ﬂux measured in culture medium without cells at 37◦ C in the Oxygraph-2k with a 2-cm3 chamber volume. (A) Background test in four chambers under routine laboratory conditions (compare [12]). The regression line is calculated for all data points in the different chambers. (B) Background test performed by students during an O2k course. (C) Plot of instrumental background ﬂux as a linear function of oxygen concentration, from traces shown in (B). (D) Deviation from the linear background regression [residuals from (A) and (C)], indicating the limit of detection of biological respiration at ±1 pmol·s−1·cm−3 , when the linear parameters are applied for automatic online correction of respiration.

(R 1 ; Figure 5B). Oxygen calibration is fully supported by the software (DatLab, OROBOROS INSTRUMENTS, Innsbruck, Austria) and combines the following information: 1. The raw signal, R 1 , obtained at air saturation of the medium. 2. The experimental temperature, T [◦ C], measured in the thermoregulated copper block encasing the glass chambers. 3. The barometric pressure, p b [kPa], measured by an electronic pressure transducer.

HIGH-RESOLUTION RESPIROMETRY: THE OXYGRAPH-2k

337

4. The oxygen partial pressure, pO2 (kPa) in air saturated with water vapor, as a function of barometric pressure and temperature and the oxygen solubility S O2 [µM/kPa] in pure water as a function of temperature, as calculated by the software [3]. 5. The oxygen solubility factor of the incubation medium, F M , which expresses the effect of the salt concentration on oxygen solubility relative to pure water, which must be known for accurate calibration. In mitochondrial respiration medium MiR05 (OROBOROS INSTRUMENTS; [31]), F M is 0.92 determined at 30 and 37◦ C [32], and F M is 0.89 in serum at 37◦ C [33]; we use the same factor for culture media such as RPMI, endothelial growth medium, or Dulbecco’s modiﬁed Eagle’s medium. Air calibrations are best performed daily before starting an experiment. 6. The most convenient second calibration point, R 0 , chosen at zero oxygen concentration. Occasional checks over a period of months are sufﬁcient (Figure 4B), except in studies of oxygen kinetics, when short-term zero drift must be accounted for by internal zero calibration for resolution in the nM oxygen range [4,5]. Zero oxygen is obtained when oxygen is depleted by mitochondrial respiration (Figure 2A) or when dithionite, Na2 S2 O4 , is added for fast zero calibration. At standard barometric pressure (100 kPa), the oxygen concentration at air saturation is 207.3 µ at 37◦ C (19.6 kPa partial oxygen pressure). In MiR05 and serum, the corresponding saturation concentrations are only 191 and 184 µM. In bioenergetics, mitochondrial respiration can be given in the units natom O·s−1·cm−3 and the dioxygen concentrations have to be multiplied by 2 to obtain µM O instead of µM O2 . Errors of 15%, due to inaccurate oxygen solubility values, appear in the literature. 3.2. From Oxygraph Slopes to Respiratory Flux Corrected for Background Effects Some sources of error in respiratory measurements with an oxygraph are due to the oxygen sensor. Linear sensor drift by 10% per day, at 190 µM at air saturation (37◦ C), would amount to a slope of 0.22 pmol·s−1·cm−3 (0.013 µM/min). The long-term stability of the OROBoPOS (Figure 4A) presents no limitation on the accuracy of measurement of the ﬂux. Thermal ﬂuctuations at a temperature dependence of the signal of the POS of 3% per◦ C [3] present a considerable problem: With thermal oscillations amounting to changes of 0.01◦ C per minure, ﬂux would ﬂuctuate at ±1 pmol·s−1·cm−3 as a function of temperature. Improved temperature stability is therefore required for high-resolution respirometry (Figure 3) and for continuous display of smooth traces of respiratory ﬂux (Figure 2). Since the signal of the polarographic oxygen sensor is sensitive to stirring of the aqueous medium, any irregular movements of the stirrer cause noise proportional to oxygen concentration. At low oxygen concentration, therefore, smaller absolute deviations of oxygen concentration per unit of time are observed, and the plot of oxygen ﬂux becomes smoother (Figure 5B).

338

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

The lower oxygen consumption of microsensors is related to the low stirring sensitivity, but the generally lower signal-to-noise ratio renders oxygen microsensors unsuitable for high-resolution respirometry. In addition to the function of the oxygen sensor, the properties of the oxygraph chamber inﬂuence the respirometric results. Standardized protocols for chamber calibration (instrumental background test) constitute an essential component of high-resolution respirometry, as they reduce instrumental artifacts. Consideration of oxygen back-diffusion is of major importance compared to correction for the oxygen consumption of the sensor [4,5,17]. The key study paving the way to measuring oxidative phosphorylation in isolated mitochondria [34] used a vibrating platinum microelectrode in a 1 cm3 cuvette that was not sealed against the air, with oxygen back-diffusion amounting to 100 × 103 pmol·s−1·cm−3 after depletion of half of the oxygen dissolved at air saturation. The principle of a closed chamber could be applied when using high mitochondrial concentrations that lead to oxygen depletion within 120 to 200 s, corresponding to respiratory ﬂuxes of 800 × 103 to 2,000 × 103 pmol·s−1·cm−3 (20,000- to 50,000-fold above the average ﬂuxes in Figure 2). If no correction for back-diffusion is applied, respiratory ﬂuxes would include systematic errors of 5 to 12% at 50% air saturation under these conditions. Rates of back-diffusion in closed chambers are ideally zero, but this is difﬁcult to achieve in practice. Back-diffusion at zero oxygen concentration in the 2 cm3 chamber is 2 ± 1 pmol·s−1·cm−3 with high-resolution respirometry [5], or 4 pmol/s into the chamber (Figure 5). This speciﬁcation can be compared with few determinations of oxygen back-diffusion ranging from 10 to 25 pmol/s when extrapolated to zero oxygen concentration, in oxygraphs with volumes in the range 1 to 8 cm3 , speciﬁcally designed for accurate measurements of P/O ratios or for studies at low oxygen levels [35–38]. With a progressive decline of oxygen concentration in the chamber, diffusion gradients increase and uncorrected back-diffusion of oxygen into the medium distorts the results. In high-resolution respirometry, oxygen ﬂux is background-corrected online as a continuous function of oxygen concentration [4,5]. Instrumental background is determined as a function of experimental oxygen concentration (Figure 5), and numerically calculated slopes are corrected on the ﬂy for instrumental background by DatLab (Figure 2). A typical instrumental background experiment is shown in Figure 5B, starting with the standard protocol for air calibration of the oxygen sensor in an experimental medium. Subsequent to testing for O2 sensor performance, the instrumental background test yields a calibration of the O2k chamber performance. When closing the chamber after equilibration at air saturation, oxygen diffusion into or out of the chamber is zero, and the oxygen consumption by the polarographic oxygen sensor can be measured (Figure 5B; ﬁrst mark: J◦ 1; 3 pmol·s−1·cm−3 , owing to electrochemical oxygen reduction at the cathode). Oxygen consumption by the polarographic oxygen sensor increases linearly with oxygen pressure, whereas back-diffusion is maximum at zero oxygen and after rapid aerobic–anoxic transitions (Figure 5B). For reducing oxygen concentration rapidly, the stopper is lifted into a reproducible stopper position deﬁned by a

HIGH-RESOLUTION RESPIROMETRY: THE OXYGRAPH-2k

339

spacer, to obtain a gas phase above the stirred medium. After injecting a small volume of argon or nitrogen into this gas phase, the oxygen concentration in the medium drops quickly, and the stoppers are pushed gently into the chamber to extrude the entire gas phase. Flux stabilizes after an undershoot (Figure 5B), and the second mark, J◦ 2, is set on the section of stable ﬂux. This is continued at one or two more reduced oxygen levels (Figure 5B; third mark: J◦ 3). In this instrumental background test, oxygen back-diffusion is evaluated by following an overall time course of oxygen depletion (Figure 5B) which matches the time course of the decline of oxygen concentration in the actual experiment (Figure 2). Plotting background oxygen ﬂux as a function of oxygen concentration yields a fairly linear relation with intercept a ◦ and slope b ◦ (in Figure 5A, −1.2 pmol·s−1·cm−3 and 0.027, respectively) [39]. These values are used (1) to conﬁrm proper function of the respirometer (results are close to the default values of −2 and 0.025), (2) to monitor the instrumental characteristics over time (a ◦ may become more negative suddenly or gradually over weeks of experiments, indicating an increasing leak, due possibly to a defective O-ring on the stopper that must be replaced), and (3) for online instrumental background correction of ﬂux during respirometric experiments in the corresponding O2k chambers. Background-corrected oxygen consumption, JV ,O2 . [pmol·s−1·cm−3 ] is calculated as JV ,O2 = −1000 · dcO2 /dt − (a o + bo · cO2 )

(1)

where cO2 [µM or nmol/cm3 ] is oxygen concentration measured at time t [equation (1)], dcO2 /dt is the time derivative of oxygen concentration, and the expression in parentheses is the background oxygen ﬂux. Using high-resolution respirometry, experimental respiratory ﬂuxes in resting states are typically about 10 pmol·s−1·cm−3 (Figure 2), and corresponding background corrections amount to 20% under these conditions (Figure 6). By comparison with traditional Clark sensor technology, atractyloside inhibited respiration is 9 µM/min in Figure 1 (150 pmol·s−1·cm−3 ). Traditional oxygraphs using chambers or stoppers made of Perspex, Teﬂon stirrers, or sealings that are not diffusion-tight therefore require >10 times higher amounts of cells or mitochondria, and the problem of oxygen diffusion is further aggravated when the chamber volume is reduced. Due to the low residual oxygen consumption after inhibition by rotenone + antimycin A, the relative effect of instrumental background correction is large and highly oxygen dependent (Figure 6). In polarographic determination of cytochrome c oxidase activity, ascorbate, TMPD, and cytochrome c are used as substrates. Chemical autoxidation of these substrates is a function of substrate concentration and is strongly oxygen dependent. Chemical background oxygen ﬂux is a linear function of oxygen concentration above 40 to 50 µM, and corrections in the form of equation (1) can be applied online. The linear parameters a and b (chemical background, after correction for instrumental background) are characteristic for the chemical process in the particular medium. The mean ± SD

340 80 60 40 22.7 22.6

20 4.8 5.0 3.1 2.4

O2 Concentration [µM]

100

ETS

ROX 70

Rotenone + Antimycin A

80

[O2]left

60

56 42

40

[O2]right

JV,O2 (left)

20

1:26

uncoupled

1:28

1:30

28 14

JV,O2 (right)

0 1:24

1:32

Time [h:min]

1:34 ROX

10

0 1:36

6

18

16

6 0

0

−1

−10 −20

C

−28 ROUTINE LEAK

ETS

ROX

150 O2 Concentration [µM]

ROUTINE LEAK

18

20

−30

0

A

B

Background effect [%]

30 59.6 59.4

Respiration, JV,O2 [pmol·s−1·cm−3]

Respiration, JV,O2 [pmol·s−1·cm−3]

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

128 132

125 128

100

87

56

61 52

50

0

ROUTINE LEAK

ETS

ROX

D

Figure 6 Analysis of phosphorylation control by high-resolution respirometry. (A) Bar graph obtained online from marked sections of the experiment (solid bars with values, from the two chambers in Figure 2A), and means ±SD (hatched bars) from six parallel runs. Respiration of intact cells in state R (ROUTINE), L (LEAK; oligomycin-inhibited) and E (ETS; uncoupled, at optimum FCCP concentration for maximum ﬂux). In the inhibited state (rotenone + antimycin A), residual oxygen consumption, ROX, is in large part due to cellular, nonmitochondrial oxygen consumption. R, L, and E are corrected for ROX. (B) Measurement of ROX (continuation of experiments of Figure 2A) after reoxygenation of the medium to different levels in the two chambers ([O2 ]left and [O2 ]right ). Flux was independent of oxygen concentration and declined progressively with time due to high inhibitory FCCP concentrations, and was inhibited to a constant level after addition of rotenone + antimycin A. All results on respiration are corrected for instrumental background. (C) Relative effect of instrumental background on respiratory ﬂux (solid bars in A), as a function of average oxygen concentration measured during the respective time intervals (D).

from six Oxygraph-2k chambers with MiR05 (three instruments operated in parallel by participants of an O2k teaching course) were a = 10.7 ± 1.4 and b = 0.24 ± 0.07 [12]. 4. PHOSPHORYLATION CONTROL PROTOCOL WITH INTACT CELLS A simple phosphorylation control protocol (PC protocol) is described and interpreted for evaluation of the physiological respiratory control state of the

PHOSPHORYLATION CONTROL PROTOCOL WITH INTACT CELLS

341

intact cells, the mitochondrial coupling state, uncoupled respiratory capacity, and rotenone + antimycin A–sensitive respiration. Respiratory control states are induced in intact cells by application of speciﬁc membrane-permeable inhibitors and uncouplers (Figure 2). The initial incubation state is cellular ROUTINE respiration, C R , reﬂecting physiological respiratory control. Cells may be suspended in a culture medium supporting ROUTINE respiration and growth by exogenome substrates (C gR ), whereas a crystalloid medium without energy substrate (e.g., mitochondrial respiration medium, MiR05 [31]) yields the state of ROUTINE respiration with endogenome substrates (C eR ). In the latter case, the effect of an intracellular formulation of ion composition on cell respiration must be evaluated. No differences in ROUTINE respiration of intact endothelial cells are observed in culture medium and mitochondrial medium [18]. In mitochondrial medium, the PC protocol can be extended to obtain a measure of enzyme activity of cytochrome c oxidase in the presence of TMPD + ascorbate, which increases after permeabilization of the plasma membrane [20]. Application of cell culture medium for respiratory measurements is advantageous when aiming at near-physiological conditions of intact cells. All inhibitors and the uncoupler applied in this protocol are freely permeable through the intact plasma membrane and therefore do not require plasma membrane permeabilization [16]. The PC protocol takes about 90 minutes (Figures 2 and 6B). 4.1. Titration Steps of the PC Protocol 1. A 10-minute period of routine respiration, reﬂecting the aerobic metabolic activity in the physiological ROUTINE state, R. 2. Nonphosphorylating (oligomycin-inhibited) LEAK rate of respiration, caused mainly by compensation for the proton leak after inhibition of ATP synthase (state L). Analogous to ADP limitation of respiration in state 4 [34], inhibition of ATP synthase (complex V) by oligomycin (1 µg/mL), or inhibition of adenylate translocase by atractyloside, arrests mitochondrial respiration at a resting level. Oxygen ﬂux measured in this LEAK state reﬂects (a) proton leak or futile respiration at maximum mitochondrial membrane potential, which is the main component, (b) proton or electron slip [decoupled respiration which includes electrons diverted away toward reactive oxygen species (ROS) production], (c) cation cycling (Ca2+ , K+ ), and (d) correction should be made for residual oxygen consumption (ROX), including peroxidase and oxidase activities which partially contribute to ROS production. 3. Uncoupler titration with the titration–injection micropump, which yields the maximum stimulated respiration as a measure of the capacity of the electron transport system (ETS) in nonpermeabilized cells (state E ), and quantitatively describes the dependence of respiration on uncoupler concentration (Figure 2). The addition of uncouplers, such as the protonophores carbonyl

342

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

cyanide p-triﬂuoromethoxyphenylhydrazone (FCCP) or DNP induces a state of maximum uncoupled respiration. Uncouplers dissipate the mitochondrial membrane potential and so maximally activate the electron transport system. Uncoupler titrations must be performed carefully, since optimum uncoupler concentrations have to be applied to achieve maximum stimulation of ﬂux, avoiding overtitration, which paradoxically, inhibits respiration [40]. Optimum uncoupler concentrations depend on the cell type, cell concentration, medium, and are different in permeabilized versus intact cells. The release of mitochondrial respiratory control by the phosphorylation system in the uncoupled state compared to the maximum inhibition of respiration achieved through blocking ATP synthesis by oligomycin leads to information on potential respiratory control by coupling as expressed by the respiratory control ratio, RCR (= ratio of respiration in the uncoupled state over respiration in the presence of oligomycin [12]). A possible inﬂuence on uncoupled ﬂux by prior inhibition of phosphorylation should be checked by controls in the absence of oligomycin. 4. Rotenone + antimycin A–inhibited respiration after inhibition of complexes I and III (residual oxygen consumption, ROX; Figure 6). Mitochondria contribute to residual oxygen consumption (particularly related to ROS production) after inhibition of complexes I and III, which argues against correcting respiration in states R, L, and E for the residual observed after inhibition with rotenone and antimycin A [16–20]. Uncoupling prior to inhibition by rotenone and antimycin A, however, prevents the large increase in mitochondrial ROS production known to occur in the presence of rotenone, and particularly antimycin A, in isolated coupled mitochondria [41,42]. Further inhibition of residual respiration by cyanide may be related to speciﬁc inhibition of cytochrome c oxidase, but may also be due to inhibition of cyanide sensitive oxygen consuming enzymes. 4.2. Experimental Example for the PC Protocol To illustrate the precision of high-resolution respirometry, superimposed plots of oxygen consumption and oxygen concentration from the two O2k chambers with identical cell densities are shown in Figure 2. The low standard deviation of the results (Figure 6A; Table 1) is a measure of methodological variability using subsamples from the same culture ﬂask, whereas cell physiological variability is larger between cultures grown on different days. Highest accuracy is achieved by step titrations of small volumes of uncoupler. The titration is terminated when a small increase in uncoupler concentration does not yield a further stimulation of oxygen ﬂux. The Oroboros titration-injection micropump TIP-2k provides an accurate and convenient tool for automatic performance of such step titrations (Figure 2). Two Hamilton syringes with 27-mm needle length and 0.09-mm needle inner diameter are mounted on the TIP-2k for simultaneous titrations into the two O2k chambers. After an aerobic–anoxic transition (Figure 2A), the two chambers were reoxygenated to different levels, and recording of respiration was continued with manual titrations of rotenone (0.5 µM) and antimycin A (2.5 µM; Figure 6B).

343

PHOSPHORYLATION CONTROL PROTOCOL WITH INTACT CELLS

TABLE 1 Metabolic States and Flux Control Ratios in the Phosphorylation Control Protocol with Intact Cellsa Metabolic State ETS, E ROUTINE, R LEAK, L

Additions FCCP None Oligomycin Atractyloside

Net ROUTINE, netR/E Residual oxygen Rotenone + consumption, Antimycin A ROX

Flux Control Ratio

Deﬁnition

Mean ± SD

Reference State 0.39 ± 0.02 0.10 ± 0.02

R/E L/E netR/E

= (R − L)/E

ROX/E

Uncoupling control ratio Respiratory control ratio

0.29 ± 0.02 0.03 ± 0.01

UCR = (R/E)−1

2.6 ± 0.1

RCR = (L/E)−1

10.1 ± 1.8

a Capacity for electron transport is the reference for normalization, E = E − ROX, where E is the apparent (uncorrected) electron transport capacity. Mean ± standard deviation of six replicate O2k measurements with 32D cells (from Figure 6).

This residual oxygen consumption (ROX) is 9% of ROUTINE respiration, but 34% and 3% of states LEAK and ETS (Figure 6A). Residual respiration after inhibition by rotenone and antimycin A is signiﬁcantly lower in permeabilized cells, suggesting that the major contribution to residual respiration is not due to mitochondria (which remain intact after cell membrane permeabilization), but rather to nonmitochondrial, cellular oxygen-consuming processes [8]. Cell respiration in various states should be corrected for ROX (Table 1). 4.3. Flux Control Ratios from the PC Protocol Normalized ﬂuxes are expressed as ratios relative to a common reference state. When the capacity of the electron transport system (ETS) in uncoupled respiration, E , is chosen as the reference state, normalized ﬂuxes in the PC protocol have the boundaries from 0.0 to 1.0 (Table 1). If the protocol is extended by measurement of cytochrome c oxidase, then the ratio of CcOX activity and uncoupled respiration is an index of the apparent excess capacity of this enzyme step in the pathway [20]. Routine respiration of 32D cells [12,43] operates at 0.39 of ETS capacity, as expressed by the R/E ratio of 0.39. 0.29 ETS capacity is used for oxidative phosphorylation under routine conditions (net R/E ; Table 1). If mild uncoupling leads to a parallel increase of R/E and L/E , the normalized net routine respiration, net R/E , remains unchanged (e.g., in senescent ﬁbroblasts at 0.2 [16]). The R/E ratio is <1.0 when cells respire below their ETS capacity. But even at full activation of OXPHOS, the R/E ratio remains <10, if the activity of the

344

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

phosphorylation system (adenine nucleotide translocator, phosphate transporter, and ATP synthase) limits maximum coupled ﬂux relative to the maximum ETS capacity. This requires evaluation of the uncoupling effect on ADP-activated respiration (state 3) in isolated mitochondria or permeabilized cells [12,44]. The inverse values of the ROUTINE and LEAK ﬂux control ratios are the uncoupling control ratio, UCR (2.6), and respiratory control ratio, RCR (10.1; Table 1). These conventional inverse ratios are mathematically less convenient, since the RCR has awkward boundaries from 1.0 (fully uncoupled) to inﬁnity (zero LEAK), compared to the L/E ratio with boundaries from 0.0 (zero LEAK) to 1.0 (fully uncoupled). The low R/E ﬂux control ratio indicates a high apparent excess capacity of respiratory complexes (including CcOX) in the parental hematopoietic 32D cells. These results from cells grown and measured in suspension are in agreement with respiratory control in lymphoblastoma cells [20], and with various attached cell types grown in monolayers but evaluated using high-resolution respirometry in suspension, including human umbilical vein endothelial cells [17,40,45], transformed endothelial EA.hyb 926 cells [18], human ﬁbroblasts [9,16], and mesenchymal cells [46]. The reason for lower uncoupling control ratios reported by Villani and Attardi [47,48] is not clear. Interpretation of respiratory ﬂux control ratios is complicated if drug toxicity causes multiple mitochondrial defects. An increase in the L/E ﬂux control ratio can be caused not only by uncoupling, but by a decrease of respiratory capacity. Diagnostic protocols for separating these effects in permeabilized muscle ﬁbers are discussed elsewhere [13]. 5. INTACT CELLS, PERMEABILIZED CELLS AND TISSUE, OR ISOLATED MITOCHONDRIA Depending on the experimental sample, the rate of oxygen consumption is frequently expressed per million cells, per milligram of tissue (wet or dry weight), per milligram of mitochondrial protein, or per unit of a mitochondrial marker. It is important to note that interpretation of changes in respiratory ﬂux is very different when expressed as per million cells, per mass of cells or tissue, or per mitochondrial marker [20]. For direct comparison of results, a common marker has to be quantiﬁed, such as mitochondrial DNA [15], citrate synthase activity [49], CcOX activity [20], or cytochrome aa 3 content [50]. Subsamples or the entire contents can be collected from the O2k chamber for analysis of cell count and viability, protein concentration, and enzyme assays (e.g., complex I, CS, and LDH). Normalized ﬂux ratios (Table 1) are, however, independent of the choice of normalization mode and of errors in the quantities chosen as a basis of expressing respiratory ﬂux. 5.1. Intact Cells Taken together, the phosphorylation control assay provides a standardized protocol to assess mitochondrial function in cell-culture-based model systems suitable

INTACT CELLS, PERMEABILIZED CELLS AND TISSUE

345

for drug testing (Figure 6). A particular advantage of studying intact cells is the quantiﬁcation of respiration in the physiologically controlled state (ROUTINE) in response to intracellular inhibitors or activators of respiration, the availability of substrates and ions (e.g., Ca2+ ) at physiological combinations, concentrations and spatial distributions, and the avoidance of potential artifacts caused by mitochondrial isolation or cell permeabilization. For analysis of the mechanism of a mitochondrial defect, however, the intact cell imposes limitations on assessment of maximum OXPHOS capacity in the coupled state, because ADP does not pass through the cell membrane, and on functional assays of various components of the respiratory system because several substrates do not enter the cell. Extended analyses require either isolation of mitochondria, or selective cell membrane permeabilization by mild detergents, such as digitonin or saponin [13,15,18,51]. Whereas isolated mitochondria remain one of the gold standards in studies of bioenergetics and mitochondrial physiology, permeabilized tissues and cells are an established alternative offering several advantages. However, some disadvantages have to be considered for optimum experimental design and critical evaluation of results. 5.2. Permeabilized Cells and Tissue When comparing results with high speciﬁc ﬂuxes and correspondingly high quality control, respiratory activity in permeabilized human skeletal muscle ﬁbers is in accord with isolated mitochondria [52]. With cell membrane permeabilization, fewer cells or less tissue are required than with isolated mitochondria. Using the Oroboros Oxygraph-2k, 1 mg wet weight of cardiac ﬁbers or 0.3 million ﬁbroblasts or endothelial cells per experimental test are sufﬁcient using a 2-mL chamber at 37◦ C. Optimization of mitochondrial isolation is more time-consuming than the optimization of cell membrane permeabilization that is done via standardized protocols. The degree of mechanical tissue separation may be evaluated by observing a change in the pale coloring of the separated ﬁber bundles (similar for liver). This is best observed when placing the petri dish onto a dark background. Appropriate forceps have to be used. Initially, the main difﬁculty is application of excess tissue, which makes mechanical separation of small amounts of tissue tedious (for isolated mitochondria, cf. [52]). A practical quantity for routine experiments is 10 to 12 mg wet weight of tissue, subsequently separated into 2-mg samples for parallel experiments. The mechanical tissue preparation leads to partial (skeletal muscle) or full permeabilization of the cell membrane (heart muscle [13]; liver tissue [49]). A preparation of fully intact cells for the study of routine respiration cannot be obtained by this approach. Partially permeabilized preparations need additional chemical permeabilization by saponin or digitonin [13,51] and via standardized incubation conditions that leave the outer and inner mitochondrial membranes intact. In merely mechanically permeabilized tissue, full permeabilization must be checked by addition of saponin or digitonin to the respirometer during state 3 with succinate + rotenone. Under these conditions,

346

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

no stimulation of respiration is expected in fully permeabilized cells, whereas partial permeabilization is indicated by the stimulatory effect of added detergent [53]. ADP has to be added to permeabilized cells at high concentrations to achieve maximum state 3 respiration, due to diffusion restrictions and because the outer mitochondrial membrane may exert a barrier different from isolated mitochondria [5,51]. Low oxygen levels have to be strictly avoided in studies with muscle ﬁbers, due to a 100-fold higher oxygen sensitivity in ﬁbers compared to isolated mitochondria [8]. Stability of the tissue preparation in the respirometer at 37◦ C depends on the application of a high-quality mitochondrial respiration medium [31], and in isolated mitochondria the isolation and preservation media need to be considered carefully [31,54]. High stability allows for application of complex and extended substrate-uncoupler-inhibitor titration protocols [13,15,55]. All types of mitochondria are accessible experimentally in permeabilized cells and tissues, whereas preparation of isolated mitochondria offers the advantage of separation of different mitochondrial populations. It is generally held, but not well documented, that isolation of mitochondria may involve the selective loss of damaged mitochondria, thereby confounding extension of ex vivo data with in vivo circumstances. 5.3. Isolated Mitochondria Isolated mitochondria are required for separation and respirometric study of different mitochondrial subpopulations [56,57]. The homogeneous suspension of isolated mitochondria yields a representative average for the tissue sample, and fewer replica are required for averaging over heterogeneous subsamples of ﬁbers. The oxygen dependence of respiration in permeabilized muscle ﬁbers is increased by two orders of magnitude, due to oxygen diffusion to the mitochondria in the small unperfused ﬁber bundle. Isolated mitochondria are, therefore, a better choice for the study of mitochondrial oxygen kinetics, although small isolated cells are a good model as well [4–9]. 6. TITRATION PROTOCOLS IN PERMEABILIZED CELLS, PERMEABILIZED TISSUE PREPARATIONS, AND ISOLATED MITOCHONDRIA By using either NADH-linked substrates (pyruvate + malate; glutamate + malate) or the classical succinate + rotenone combination, different segments of the electron transport system can be interrogated [12,34]. Mitochondrial respiration depends on the continuous ﬂow of substrates across the mitochondrial membranes into the matrix space. Many substrates are strong anions that cannot permeate lipid membranes and hence require carriers. Various anion carriers in the inner mitochondrial membrane are involved in the transport of mitochondrial metabolites. Their distribution across the mitochondrial membrane varies mainly with

MULTISENSOR APPLICATIONS

347

pH and not , since most carriers (but not the glutamate–aspartate carrier) operate nonelectrogenically by anion exchange or cotransport of protons. Depending on the concentration gradients, these carriers also allow for the transport of mitochondrial metabolites from the mitochondria into the cytosol and for the loss of intermediary metabolites into the incubation medium. Export of intermediates of the tricarboxylic acid (TCA) cycle plays an important metabolic role in the intact cell. This must be considered when interpreting the effect on respiration of speciﬁc substrates used in studies of isolated mitochondria. Substrate combinations of pyruvate + malate (PM) and glutamate + malate (GM) activate dehydrogenases yielding reduced nicotinamide adenine dinucleotide (NADH), which feeds electrons into complex I (NADH–ubiquinone oxidoreductase) and hence down the thermodynamic cascade through the Q-cycle and complex III to complex IV and ultimately to O2 . Complex II is the only membrane-bound enzyme in the tricarboxylic acid cycle and is part of the mitochondrial electron transport chain. The ﬂavoprotein succinate dehydrogenase is the largest polypeptide of complex II, located on the matrix face of the inner mitochondrial membrane. Following succinate oxidation, the enzyme transfers electrons directly to the quinone pool [58]. Whereas complex I is NADH-linked upstream of the dehydrogenases of the tricarboxylic acid cycle, complex II is FADH2 -linked downstream with subsequent electron ﬂow to CoQ [12]. Electrons ﬂow to oxygen from either complex I, with a total of three proton pumps in series, or from complex II and other ﬂavoproteins, providing multiple entries into the Q-cycle with only two proton pumps downstream. Substrate combinations that match physiological intracellular conditions are required for evaluation of the maximum capacity of oxidative phosphorylation. A novel perspective of mitochondrial physiology and respiratory control by simultaneous supply of various substrates emerged from a series of studies based on high-resolution respirometry [12,15,19,59]. Application of substrate combinations in multiple substrate-uncoupler-inhibitor titration protocols extends conventional bioenergetic studies to the level of mitochondrial physiology suitable for full characterization of respiratory control in health and disease.

7. MULTISENSOR APPLICATIONS The large chamber of the Oxygraph-2k (16 mm in diameter) offers the possibility of inserting additional sensors into the chamber through the stopper. The signal can be fed into the multichannel electronics of the O2k, providing simultaneous recordings in the DatLab software. Of particular diagnostic value is the simultaneous measurement of respiration and mitochondrial membrane potential [60] using an ion-selective electrode that responds to TPP+ or TPMP+ [61,62]). The simultaneous measurement of oxygen consumption and cellular proton production offers a potential for real-time monitoring of aerobic and anaerobic metabolism in intact cells comparable to combined respirometry

348

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

and microcalorimetry (calorespirometry [6,63,64]). Proton production in conjunction with the enthalpy of neutralization requires consideration as a correction term in calorimetric determination of anaerobic metabolism [64]. Beyond a mere correction term, the ratio of proton production from acidic end products to ATP turnover [65] yields a sensitive estimate of anaerobic metabolism. Combining high-resolution respirometry and measurement of nitric oxide under physiological low oxygen levels yields new insights into the competitive inhibition of CcOX [66]. A variety of multisensor applications, such as ion-sensitive electrodes or light guides for spectroscopy or ﬂuorescence protocols, render high-resolution respirometry a powerful technology for detection and parsing drug-induced mitochondrial dysfunction. REFERENCES 1. Dykens JA, Will Y. The signiﬁcance of mitochondrial toxicity testing in drug development. Drug Discov Today. 2007;12:777–785. 2. Gellerich FN, Mayr JA, Reuter S, Sperl W, Zierz S. The problem of interlab variation in methods for mitochondrial disease diagnosis: enzymatic measurement of respiratory chain complexes. Mitochondrion. 2004;4:427–439. 3. Gnaiger E, Forstner H, eds. Polarographic Oxygen Sensors: Aquatic and Physiological Applications. New York: Springer-Verlag; 1983. 4. Gnaiger E, Steinlechner-Maran R, M´endez G, Eberl T, Margreiter R. Control of mitochondrial and cellular respiration by oxygen. J Bioenerg Biomembr. 1995;27:583–596. 5. Gnaiger E. Bioenergetics at low oxygen: dependence of respiration and phosphorylation on oxygen and adenosine diphosphate supply. Respir Physiol. 2001;128:277–297. 6. Gnaiger E, M´endez G, Hand SC. High phosphorylation efﬁciency and depression of uncoupled respiration in mitochondria under hypoxia. Proc Natl Acad Sci U S A. 2000;97:11080–11085. 7. Gnaiger E, Kuznetsov AV. Mitochondrial respiration at low levels of oxygen and cytochrome c. Biochem Soc Trans. 2002;30:252–258. 8. Gnaiger E. Oxygen conformance of cellular respiration: a perspective of mitochondrial physiology. Adv Exp Med Biol. 2003;543:39–56. 9. Pecina P, Gnaiger E, Zeman J, Pronicka E, Houˇstˇek J. Decreased afﬁnity to oxygen of cytochrome c oxidase in Leigh syndrome caused by SURF1 mutations. Am J Physiol Cell Physiol. 2004;287:C1384–C1388. 10. Rostrup M, Fossbakk A, Hauge A, Kleppe R, Gnaiger E, Haavik J. Oxygen dependence of tyrosine hydroxylase. Amino Acids. 2008;34:455–464. 11. Puchowicz MA, Varnes ME, Cohen BH, Frieman NR, Kerr DS, Hoppel CL. Oxidative phosphorylation analysis: assessing the integrated functional activity of human skeletal muscle mitochondria: case studies. Mitochondrion. 2004;4:377–385. 12. Gnaiger E, editor. Mitochondrial Pathways and Respiratory Control . Innsbruck, Austria: Oroboros MiPNet Publications; 2007. http://www.oroboros.at. 13. Kuznetsov AV, Schneeberger S, Seiler R, et al. Mitochondrial defects and heterogeneous cytochrome c release after cardiac cold ischemia and reperfusion. Am J Physiol Heart Circ Physiol. 2004;286:H1633–H1641.

REFERENCES

349

14. Wenchich L, Drahota1 Z, Honz´ık T, et al. Polarographic evaluation of mitochondrial enzymes activity in isolated mitochondria and in permeabilized human muscle cells with inherited mitochondrial defects. Physiol Res. 2003;52:781–788. 15. Boushel R, Gnaiger E, Schjerling P, Skovbro M, Kraunsøe R, Dela F. Patients with type 2 diabetes have normal mitochondrial function in skeletal muscle. Diabetologia. 2007;50:790–796. 16. H¨utter E, Renner K, Pﬁster G, St¨ockl P, Jansen-D¨urr P, Gnaiger E. Senescenceassociated changes in respiration and oxidative phosphorylation in primary human ﬁbroblasts. Biochem J. 2004;380:919–928. 17. H¨utter E, Unterluggauer H, Garedew A, Jansen-D¨urr P, Gnaiger E. High-resolution respirometry: a modern tool in aging research. Exp Gerontol. 2006;41:103–109. 18. Stadlmann S, Rieger G, Amberger A, Kuznetsov AV, Margreiter R, Gnaiger E. H2 O2 -mediated oxidative stress versus cold ischemia-reperfusion: mitochondrial respiratory defects in cultured human endothelial cells. Transplantation. 2002;74:1800–1803. 19. Aragon´es J, Schneider M, Van Geyte K, et al. Deﬁciency or inhibition of oxygen sensor Phd1 induces hypoxia tolerance by reprogramming basal metabolism. Nature Genetics 2008;40:170–180. 20. Renner K, Amberger A, Konwalinka G, Koﬂer R, Gnaiger E. Changes of mitochondrial respiration, mitochondrial content and cell size after induction of apoptosis in leukemia cells. Biochim Biophys Acta. 2003;1642:115–123. 21. Hofhaus G, Shakeley RM, Attardi G. Use of polarography to detect respiration defects in cell cultures. Methods Enzymol. 1996;264:476–483. 22. Yamaguchi R, Andreyev A, Murphy AN, Perkins GA, Ellisman MH, Newmeyer DD. Mitochondria frozen with trehalose retain a number of biological functions and preserve outer membrane integrity. Cell Death Differ. 2006;14:616–624. 23. Frezza C, Cipolat S, Scorrano L. Organelle isolation: functional mitochondria from mouse liver, muscle and cultured ﬁbroblasts. Nat Protocols. 2007;2:287–295. 24. Villani G, Attardi G. Polarographic assays of respiratory chain complex activity. Methods Cell Biol. 2007;80:121–133. 25. Renner K, Koﬂer R, Gnaiger E. Mitochondrial function in glucocorticoid triggered T-ALL cells with transgenic Bcl-2 expression. Mol. Biol Rep. 2002;29:97–101. 26. Holtzmann D, Moore CL A micro-method for the study of oxidative phosphorylation. Biochim Biophys Acta. 1971;234:1–8. 27. Atkinson HJ, Smith L. An oxygen electrode microrespirometer. J Exp Biol. 1973;59:247–253. 28. Suchy J, Cooper C. Isolation and respiratory measurement on a single large mitochondrion. Exp Cell Res. 1974;88:198–202. 29. Justice RE, Utsunomiya T, Krausz MM, Valeri CR, Shepro D, Hechtman HB. A miniaturized chamber for the measure of oxygen consumption. J Appl Physiol. 1982;52:488–490. 30. Rasmussen HN, Rasmussen UF. Respiration measurements in small scale. Anal Biochem. 1993;208:244–248. 31. Gnaiger E, Kuznetsov AV, Schneeberger S, Seiler R, Brandacher G, Steurer W, Margreiter R. Mitochondria in the cold. In Life in the Cold (G Heldmaier, M Klingenspor, eds). New York: Springer-Verlag; 2000: 431–442.

350

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

32. Rasmussen HN, Rasmussen UF. Oxygen solubilities of media used in electrochemical respiration measurements. Anal Biochem. 2003;319:105–113. 33. Baumg¨artl H, L¨ubbers DW. Microaxial needle sensor for polarographic measurement of local O2 pressure in the cellular range of living tissue: its construction and properties. In Polarographic Oxygen Sensors (E Gnaiger, H Forstner, eds). New York: Springer-Verlag; 1983: 37–65. 34. Chance B, Williams GR. Respiratory enzymes in oxidative phosphorylation. I. Kinetics of oxygen utilization. J Biol Chem. 1955;217:383–393. 35. Lemasters JJ. The ATP-to-oxygen stoichiometries of oxidative phosphorylation by rat liver mitochondria. J Biol Chem. 1984;259:13123–13130. 36. Stoner CD. Determination of the P/2e− stoichiometries at the individual coupling sites in mitochondrial oxidative phosphorylation. J Biol Chem. 1987;262,11445–11453. 37. Robiolio M, Rumsey WL, Wilson DF. Oxygen diffusion and mitochondrial respiration in neuroblastoma cells. Am J Physiol. 1989;256:C1207–C1213. 38. Hollis VS, Palacios-Callender M, Springett RJ, Delpy DT, Moncada S. Monitoring cytochrome redox changes in the mitochondria of intact cells using multi-wavelength visible light spectroscopy. Biochim Biophys Acta. 2003;1607:191–202. 39. Garedew A, H¨utter E, Haffner B, Gradl P, Gradl L, Jansen-D¨urr P, Gnaiger E. High-resolution respirometry for the study of mitochondrial function in health and disease. The OROBOROS Oxygraph-2k. Proceedings of the 11th Congress of the European Shock Society, Vienna, Austria (H Redl, ed) Bologna, Italy: Medimond International Proceedings; 2005:107–111. 40. Steinlechner-Maran R, Eberl T, Kunc M, Margreiter R, Gnaiger E. Oxygen dependence of respiration in coupled and uncoupled endothelial cells. Am J Physiol. 1996;271:C2053–C2061. 41. Boveris A, Chance B. The mitochondrial generation of hydrogen peroxide: general properties and effect of hyperbaric oxygen. Biochem J. 1973;134:707–716. 42. Garait B, Couturier K, Servais S, et al. Fat intake reverses the beneﬁcial effects of low caloric intake on skeletal muscle mitochondrial H2 O2 production. Free Radic Biol Med. 2005;39:1249–1261. 43. Troppmair J, Rapp UR. Raf and the road to cell survival: a tale of bad spells, ring bearers and detours. Biochem Pharmacol. 2003;66:1341–1345. 44. Rasmussen UF, Rasmussen HN, Krustrup P, Quistorff B, Saltin B, Bangsbo J. Aerobic metabolism of human quadriceps muscle: in vivo data parallel measurements on isolated mitochondria. Am J Physiol Endocrinol Metab. 2001;280:E301–E307. 45. Steinlechner-Maran R, Eberl T, Kunc M, Schr¨ocksnadel H, Margreiter R, Gnaiger E. Respiratory defect as an early event in preservation/reoxygenation injury in endothelial cells. Transplantation. 1997;63:136–142. 46. Stadlmann S, Renner K, Pollheimer J, et al. Preserved coupling of oxydative phosphorylation but decreased mitochondrial respiratory capacity in IL-1β treated human peritoneal mesothelial cells. Cell Biochem Biophys. 2006;44:179–186. 47. Villani G, Attardi G. In vivo control of respiration by cytochrome c oxidase in wildtype and mitochondrial DNA mutation-carrying human cells. Proc Natl Acad Sci U S A. 1997;94:1166–1171. 48. Villani G, Greco M, Papa S, Attardi G. Low reserve capacity of cytochrome c oxidase capacity in vivo in the respiratory chain of a variety of human cell types. J Biol Chem. 1998;273:31829–31836.

REFERENCES

351

49. Kuznetsov AV, Strobl D, Ruttmann E, K¨onigsrainer A, Margreiter R, Gnaiger E. Evaluation of mitochondrial respiratory function in small biopsies of liver. Anal Biochem. 2002;305:186–194. 50. Mootha VK, Arai AE, Balaban RS. Maximum oxidative phosphorylation capacity of the mammalian heart. Am J Physiol. 1997;272:H769–H775. 51. Saks VA, Veksler VI, Kuznetsov AV, et al. Permeabilised cell and skinned ﬁber techniques in studies of mitochondrial function in vivo. Mol Cell Biochem. 1998;184:81–100. 52. Rasmussen UF, Rasmussen HN. Human quadriceps muscle mitochondria: a functional characterization. Mol Cell Biochem. 2000;208:37–44. 53. Gnaiger E, Kuznetsov AV, Lassnig B, et al. High-resolution respirometry. Optimum permeabilization of the cell membrane by digitonin. In BioThermoKinetics in the Post Genomic Era (C Larsson, I-L P˚ahlman, L Gustafsson, eds). G¨oteborg, Sweden: Chalmers Reproservice; 1998: 89–95. 54. Brewer GJ, Jones TT, Wallimann T, Schlattner U. Higher respiratory rates and improved creatine stimulation in brain mitochondria isolated with anti-oxidants. Mitochondrion. 2004;4:49–57. 55. Gnaiger E, Wright-Paradis C, Sondergaard H, et al. High-resolution respirometry in small biopsies of human muscle: correlations with body mass index and age. Mitochondr Physiol Network . 2005;10(9):14–15. http://www.mitophysiology.org/ index.php?gnaigere. 56. Palmer JW, Tandler B, Hoppel CL. Biochemical properties of subsarcolemmal and interﬁbrillar mitochondria isolated from rat cardiac muscle. J Biol Chem. 1977;252:8731–8739. 57. Riva A, Tandler B, Loffredo F, Vazquez E, Hoppel CL. Structural differences in twobiochemically deﬁned populations of cardiac mitochondria. Am J Physiol Heart Circ Physiol. 2005;289:H868-H872. 58. Sun F, Huo X, Zhai Y, et al. Crystal structure of mitochondrial respiratory membrane protein complex II. Cell. 2005;121:1043–1057. 59. Lemieux H, Garedew A, Blier PU, Tardif J-C, Gnaiger E. Temperature effects on the control and capacity of mitochondrial respiration in permeabilized ﬁbers of the mouse heart. Biochim Biophys Acta (EBEC Short Rep Suppl). 2006;14:201–202. 60. Jekabsons MB, Nicholls DG. In situ respiration and bioenergetic status of mitochondria in primary cerebellar granule neuronal cultures exposed continuously to glutamate. J Biol Chem. 2004;279:32989–33000. 61. Brand MD. Measurement of mitochondrial protonmotive force. In Bioenergetics: A Practical Approach (GC Brown, CE Cooper, eds). New York: Oxford University Press; 1995. 62. Labajova A, Vojtiskova A, Krivakova P, Kofranek J, Drahota Z, Houstek J. Evaluation of mitochondrial membrane potential using a computerized device with a tetraphenylphosphonium-selective electrode. Anal Biochem. 2006;353:37–42. 63. Gnaiger E. The twin-ﬂow microrespirometer and simultaneous calorimetry. In Polarographic Oxygen Sensors: Aquatic and Physiological Applications (E Gnaiger, H Forstner, eds). New York: Springer-Verlag; 1983:134–166. 64. Gnaiger E, Kemp RB. Anaerobic metabolism in aerobic mammalian cells: information from the ratio of calorimetric heat ﬂux and respirometric oxygen ﬂux. Biochim Biophys Acta. 1990;1016:328–332.

352

POLAROGRAPHIC OXYGEN SENSORS, THE OXYGRAPH-2k

65. Gnaiger E. Heat dissipation and energetic efﬁciency in animal anoxibiosis: economy contra power. J Exp Zool. 1983;228:471–490. 66. Aguirre E, Rodriguez-Juarez F, Gnaiger E, Cadenas S. Measurement of the control of cellular respiration by nitric oxide under normoxia and hypoxia: instrumental comparison including high-resolution respirometry. Biochim Biophys Acta. (EBEC Short Rep Suppl). 2006;14:136–137.

13 USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES FOR THE ASSESSMENT OF MITOCHONDRIAL FUNCTION ´ C. O’Riordan James Hynes and Tomas Luxcel Biosciences Ltd., BioInnovation Centre, University College–Cork, Cork, Ireland

Dmitri B. Papkovsky Biochemistry Department, University College–Cork, Cork, Ireland

1. Introduction 2. Quenched-ﬂuorescence oxygen sensing 2.1. Principles of quenched-fluorescence oxygen sensing 2.2. Data output 2.3. Probes for quenched-fluorescence oxygen sensing 3. Applications of oxygen-sensing probes 3.1. Analysis of isolated mitochondria 3.2. Extracellular assessment of oxygen consumption 3.3. Intracellular assessment of oxygen concentration 3.4. Alternative approaches 4. Conclusions and future perspectives

353 355 355 357 358 360 361 362 363 366 367

1. INTRODUCTION An increasing appreciation of the importance of mitochondrial function to cellular homoeostasis has made analysis of the mitochondrion and the complex Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

353

354

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES

interrelationships with which it is involved relevant to ever-increasing ﬁelds of biological and toxological research. As outlined in previous chapters, mitochondrial dysfunction has been shown to play a central role in a wide array of diseases and toxicities, including the progression and regulation of apoptosis [1], neurological, and cardiovascular disorders [2,3] and drug-induced cytotoxicity [4–6]. Methods capable of assessing mitochondrial function have therefore become relevant to a broad community of researchers. A comprehensive understanding of the metabolic and mitochondrial processes involved in these complex interrelationships requires analysis across a wide range of relevant cellular parameters. Some of the most germane and widely measured of these include cellular ATP, mitochondrial membrane potential ( m ), reactive oxygen species (ROS) production, extracellular acidiﬁcation, cellular calcium ﬂuctuations, and oxygen consumption. Parameters such as m , cellular ATP, and cellular NADH provide valuable information on the energy balance of the cell. However, difﬁculties can arise when interpreting such data without knowledge of the associated turnover rates. Oxygen consumption is valuable in this regard, as it provides speciﬁc information on the rate of electron ﬂux though the electron transport chain (ETC), supplying direct information on the generation of m . Any perturbance of ETC activity will be detected as an alteration in oxygen consumption. Such perturbations can occur via direct mechanisms such as speciﬁc inhibition of complexes of the ETC or the F1 FO -ATPase, or uncoupling of ETC activity from ADP phosphorylation. More indirect mechanisms, such as a disruption of the supply of reducing equivalents through the inhibition of supply pathways can also perturb ETC activity. Other factors, such as calcium-mediated alterations of regulatory mechanisms, changes in glutathione or redox status, or inhibition of speciﬁc transporters can also contribute to altered oxygen uptake. This sensitivity makes oxygen consumption analysis an extremely effective indicator of mitochondrial dysfunction. In addition, recent advances in our understanding of cellular responses to hypoxia [7–10] and the role of both calcium and nitric oxide as regulators of metabolism [11–13] demonstrate the importance that such measurements will have in these developing ﬁelds of research. The power of oxygen consumption analysis as an indicator of mitochondrial dysfunction is generally accepted; however, the difﬁculty has been in harnessing this potential in an appropriate measurement format. Traditionally, the mainstay of biological oxygen consumption analysis has been potentiometric systems such as those based on the Clark electrode [14]. This approach, particularly when using more modern systems [15], is relatively simple, robust, and sensitive and is well suited to situations where detailed analysis of low numbers of samples is required (see Chapter 12). Difﬁculties arise, however, when trying to adapt this approach to larger numbers of samples in drug development arenas, as throughput is extremely limited. Intracellular application of this technology is also highly problematic, due to the invasive and consumptive nature of the measurement system. Such approaches are also blind to the spatial distribution of oxygen within the cell.

QUENCHED-FLUORESCENCE OXYGEN SENSING

355

The challenge in recent years has therefore been the development of noninvasive measurement methodologies that allow assessment of oxygen consumption in a format compatible with current assay paradigms, thereby combining the high information content of oxygen consumption measurements with the throughput and ﬂexibility of microtiter plate-based assays. Ideally, such methodologies would be capable of measuring both intra- and extracellularly. Extracellular measurements allow simple assessments of sustained changes in levels of biological oxygen consumption, while intracellular measurements allow the assessment of rapid and transient changes in oxygen consumption and can also inform as to oxygen distribution within the cell. In addressing these challenges, an alternative technology has come to the fore, quenched ﬂuorescence oxygen sensing, a technology that allows quantiﬁcation of dissolved oxygen based on the ability of dioxygen molecules to quench certain long-decay luminophores [16]. When conﬁgured correctly, levels of sensitivity are comparable to potentiometric systems, with superior performance at lower oxygen concentrations. This approach was originally applied to the assessment of tissue and tumor oxygenation using a range of water-soluble metalloporphyrins and lifetime-based ﬂuorescence imaging [17–20]. However, the noninvasive nature of the optical approach and compatibility with standard luminescence detection technologies has facilitated application of this technology to plate-based systems. Remote probe interrogation also facilitates intracellular analysis as it circumvents the invasive nature of microelectrode-based assessments of intracellular oxygen [21,22]. In this chapter we focus on the intracellular and extracellular use of phosphorescent oxygen-sensitive probes for the assessment of cellular oxygen consumption and distribution as well as outlining how these approaches may be used to help answer speciﬁc research questions.

2. QUENCHED-FLUORESCENCE OXYGEN SENSING 2.1. Principles of Quenched-Fluorescence Oxygen Sensing Recent years have seen the adaptation of quenched-ﬂuorescence oxygen-sensing technology to the requirements of life science research [22–24], with a general trend toward assay designs that are compatible with current measurement technologies. This is a ﬂuorescence-based method of measuring molecular oxygen, achieved by harnessing the ability of molecular oxygen to dynamically quench the excited state of certain ﬂuorophores via a nonchemical (collisional) mechanism. The ﬂuorophores normally employed are either phosphorescent metalloporphyrins [25,26] or ruthenium(II) metal–ligand complexes [27–30], as these two classes of dyes display both suitable photophysical characteristics and the requisite sensitivity to oxygen. The relationship between oxygen concentration and the emissive properties of the ﬂuorophore is inversely proportional; increasing concentration/partial pressure of dissolved oxygen represses probe emission [16] (Figure 1A). However, this quenching is sensitive to temperature; warming the system increases the

356

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES

A

B (lo/l)-1 KD Fluorescent Signal

Fluorescent Signal

[O2] =

[O2]

Deoxygenation Reoxygenation 0% O2

100% O2

100% O2 Time

Figure 1 (A) The emission intensity/lifetime of an oxygen-sensitive dye is quenched by molecular oxygen. Increasing concentrations of molecular oxygen therefore cause a decrease in emission intensity or lifetime. Appropriate dyes give a large signal change over the oxygen concentration of interest (solid curve). Unsuitable dyes display a small signal change over this range (dashed curve). (B) Samples exposed to the air contain oxygen at air-saturated levels (21%). Deoxygenation of a sample causes an increase in signal due to the removal of the quenching effect of oxygen. This effect is fully reversible, as on deoxygenation, the signal returns to its original level.

frequency of collisions between ﬂuorophore and oxygen while affecting the emissive properties of the ﬂuorophore. Therefore, precise temperature control and calibration data at different temperatures are paramount in designing a precise oxygen-sensing system. The overall response of the ﬂuorophore over the oxygen concentration range of interest is critical. Certain dyes are heavily quenched, meaning that although they are very suitable for analysis at low oxygen concentrations, they are not suitable for analysis over the physiological range (dashed line in Figure 1A). Other ﬂuorophores display more moderate quenching and, due to the large signal changes that occur over the physiological range, are well suited to such analysis (solid line in Figure 1A). Under ambient conditions ([O2 ] = 235 µM; 21% or 21 kPa; 160 torr), probe emission is quenched and deoxygenation of the sample dequenches the probe, thereby rapidly increasing the signal. This increase is completely reversible, and upon reoxygenation of the sample, probe signal returns to a baseline (Figure 1B). Measuring probe signals therefore allows quantiﬁcation of dissolved oxygen, with changes in this signal reﬂecting changes in oxygen concentration within the sample [16]. Quantiﬁcation may be achieved by measuring the emissive properties of a probe, such as intensity or lifetime, and relating these to oxygen concentrations via a known calibration function. Such measurements can potentially be taken from the extracellular medium, reﬂecting extracellular oxygen

QUENCHED-FLUORESCENCE OXYGEN SENSING

357

concentrations, or from within the cell, reﬂecting intracellular oxygen concentrations. There are a number of limitations associated with this approach, including temperature dependence of probe response to oxygen, possible optical interferences from ﬂuorescent samples, and photobleaching of oxygen-sensitive materials under intense or continuous illumination. However, most of these problems can be solved effectively by the proper selection of oxygen-sensitive material (sensor/probe), measurement format, and detection instrumentation. 2.2. Data Output Oxygen-sensitive probes may be interrogated by analyzing either probe emission intensity or lifetime (i.e., the mean time spent in the excited state). In many cases, intensity-based measurements are employed with success; however, there are instances where lifetime-based measurements are necessary to report more reliably on absolute concentrations of oxygen. Extracellular Measurement When assessing cellular oxygen consumption in an open system such as a microtiter plate, consideration must be given to the competing processes taking place. Dissolved oxygen is being consumed by the activity of the electron transport chain, resulting in the generation of an oxygen gradient. This in turn drives the diffusion of ambient oxygen into the sample at both the liquid–air interface and through the polystyrene body of the microtiter plate. Eventually, after an equilibration phase, either all oxygen is depleted or a steady state is reached. Sealing the plate limits this diffusion, thereby increasing assay sensitivity [31,32]. This may be achieved in a variety of ways, including through the addition of a layer of mineral oil. Increased sensitivity may also be achieved by using oxygen-impermeable materials [33]. When an oxygen-sensitive probe is used to report changes in oxygen concentration, the raw data output appears as presented in Figure 2A. Higher rates of oxygen consumption give more rapid signal changes, with a steady state being reached at lower oxygen concentrations, as reﬂected by higher probe intensities. A reduction in biological activity slows oxygen consumption, which is seen as a slower rate of signal change, and steady state is reached at a higher concentration of oxygen. Successful analysis therefore requires rates of consumption that produce a measurable equilibration phase. The raw data may be used semiquantitatively to determine increased or decreased oxygen consumption. For example, in Figure 2A, the oxygen consumption rate is decreasing from blue to green, with the black proﬁle showing no detectable change in dissolved oxygen concentration. In the context of drug treatment, the upper proﬁle would reﬂect uncoupling, with the lower proﬁle reﬂecting inhibition. For more quantitative data, these raw proﬁles may be linearized [34] or converted to oxygen concentrations [35]. Intracellular Measurement When sensing intracellular oxygen, the processes involved are slightly different [21]. Intracellular oxygen concentrations vary depending on cell type [36], in many cases are lower than those of the

358

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES A

B Uncoupling

Untreated

Inhibition Control

Time

Fluorescent Signal

Fluorescent Signal

Activation Treatment

Resting

Inhibition Time

Figure 2 Diagrammatic representation of data output from extracellular (A) and intracellular (B) analysis of cellular oxygen consumption. (A) Extracellular measurement reﬂects the concentration of dissolved oxygen in the extracellular medium. Measurement begins at air-saturated concentrations of oxygen. The activity of the ETC causes a reduction in the concentration of dissolved oxygen, resulting in an increase in probe emission. Higher levels of ETC activity cause more rapid oxygen depletion and greater signal changes. The rate of these signal changes is analyzed, allowing the identiﬁcation of both inhibitors and uncouplers. (B) Intracellular measurements reﬂect the intracellular concentration of dissolved oxygen. Due to the activity of the ETC, measurement usually begins at oxygen concentrations that are below air-saturated levels. Reduced ETC activity results in an increase in intracellular oxygen concentration, resulting in a decrease in probe signal, while increased ETC activity results in a decrease in intracellular oxygen concentration, resulting in an increase in probe signal. Increases may be prolonged or more transient, depending on the processes involved. This allows the analysis of inhibited and activated ETC activity within the cell.

extracellular environment due to cellular oxygen consumption. In simple terms, an increase or decrease in the rate of this consumption will alter the balance between oxygen consumption and diffusion of oxygen across the cell membrane, resulting in an altered intracellular oxygen concentration. Endpoint analysis allows comparison between treated and untreated cells and reveals any changes in the concentration of intracellular oxygen as a result of treatment at the time of measurement. However, as presented in Figure 2B, kinetic analysis allows more sensitive detection, facilitating analysis of the kinetics and magnitude of these changes in real time. A reduction in the rate of consumption causes an increase in the concentration of intracellular oxygen, resulting in a decrease in signal, with the opposite being true for an increase in the rate of consumption. 2.3. Probes for Quenched-Fluorescence Oxygen Sensing As outlined in Section 2.1, probes used for the assessment of biological oxygen consumption require sensitivity to oxygen over the entire physiological range,

359

QUENCHED-FLUORESCENCE OXYGEN SENSING

ideally displaying a linear relationship between the concentration of oxygen and the probe response (termed the Stern–Volmer calibration function [37]) and high quantum yields. In addition, they should display the requisite photostability, rapid response times, and low cyto- and phototoxicity and should be compatible with existing instrumentation and measurement paradigms. Compatibility with simple pipetting provides an additional advantage. Hydrophilic metalloporphyrin-based probes fulﬁll the foregoing requirements and are therefore particularly well suited for the assessment of biological oxygen consumption. Such probes have distinct photophysical properties. They are excitable at either 380 to 400 nm or 525 to 600 nm (Figure 3A), and have an emission maximum in the near infrared (650 to 800 nm) (Figure 3B). Deoxygenation provides positive signal response (Figure 3B) which can be measured on standard ﬂuorescent plate readers and imagers. The long emission lifetimes of these probes (microsecond durations) facilitate the use of time-resolved ﬂuorescence detection, thereby increasing sensitivity to a probe by increasing the signal-to-blank ratio. The increased sensitivity conferred by this measurement mode allows the use of a probe at nanomolar concentrations while also signiﬁcantly reducing optical interferences by allowing autoﬂuorescence to decay prior to the beginning of measurement. The long lifetimes also facilitate lifetime-based sensing, which is particularly well suited to oxygen measurements, as it is independent of probe concentration and sample geometry and is highly resistant to optical interferences, resulting in even more robust measurement. In the microsecond time scale, this can be achieved relatively simply using time-resolved ﬂuorometers (see below). When analyzing extracellularly, these probes can be dispensed directly into the test sample, allowing the use of standard microtiter plates and detection

A

1

B

0.8 0.6 0.4 0.2 0 300

2

0.8 Fold Increase

Normalized Intensity

1

0.6 0.4 0.2

500 400 Wavelength (nm)

600

0 600

1

650 Wavelength (nm)

700

Figure 3 (A) Normalized excitation spectrum of an oxygen probe, with emission measured at 650 nm. Excitation maxima are observed at 340 to 400 or 525 to 545 nm. (B) Emission spectrum of oxygen probe in oxygenated (1) and deoxygenated (2) conditions, with excitation at 380 nm. Under the conditions of measurements, signals increased ﬁvefold on deoxygenation.

360

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES

instrumentation [20,22,38]. The probes also confer a high degree of ﬂexibility, facilitating measurement in a diverse array of measurement formats from standard microplates [22], to high-sensitivity glass capillaries [39] and specialized high-sensitivity biochip platforms [40,41]. The disadvantage associated with the use of such probes is the potential susceptibility to interference from compounds in the measurement medium. The use of approaches such as time-resolved ﬂuorescence and data normalization can minimize such limitations [34,42]. Care must also be taken to ensure that components are mixed sufﬁciently [43]. Measuring intracellularly imposes an additional set of constraints. Probes must be compatible with minimally invasive delivery systems and must be photostable under extended high-intensity illumination. Additionally, photosensitized generation of singlet oxygen during measurement must be dealt with to limit associated cytotoxicity. This can be achieved through the introduction of scavengers into the medium or by linking absorbers (“sinks”) to the probe. Hydrophilic metalloporphyrin-based probes are again well suited to such analysis, as they can also be loaded passively using standard liposomal transfer techniques, thereby avoiding the necessity for harsh approaches such as microprojectile delivery [21]. A number of dedicated intracellular probes have the requisite sensitivity and photostability and low phototoxicity, due primarily to the protein component of the probe, which scavenges reactive oxygen species [38].

3. APPLICATIONS OF OXYGEN-SENSING PROBES The oxygen-sensitive probes described in Section 2 can be used to assess changes in cellular oxygen consumption, and in relation to the assessment of mitochondrial function, this is generally performed with either isolated mitochondria or whole cells. The speciﬁc biological parameters being investigated will determine whether it is more appropriate to use these probes extracellularly or intracellularly. Assessment of changes in the extracellular concentration of dissolved oxygen using water-soluble oxygen probes allows detailed, high-throughput interrogation of mitochondrial function. This type of bulk analysis may be applied to both isolated mitochondria and whole cells and is appropriate for the measurement of prolonged, reasonably synchronous alterations in oxygen consumption typical of cytotoxicity [44], drug-induced mitochondrial dysfunction [35], and even apoptosis induction [45]. Intracellular analysis is more invasive and laborious than extracellular measurements, in that cells must be preloaded with an oxygen-sensitive dye. However, it facilitates certain analyses that extracellular measurements of cell populations cannot achieve. For example, it is particularly well suited to the analysis of more subtle asynchronous or transient alterations in cellular oxygen consumption that are beyond the sensitivity of extracellular analysis. Fluorescent imaging also facilitates the measurement of changes in oxygen within individual cells. In addition, the kinetics of local changes in intracellular oxygen are much

APPLICATIONS OF OXYGEN-SENSING PROBES

361

faster than the formation of oxygen gradients in the extracellular medium, which eliminates the need for exclusion of ambient oxygen and sealing test samples. 3.1. Analysis of Isolated Mitochondria Measurement of isolated mitochondria with oxygen-sensitive probes is generally carried out on standard 96- or 383-well plates. To perform analysis, mitochondria are added to the wells in an appropriate measurement buffer at an appropriate measurement concentration [35]. Oxidizible substrates and probe are then added and the wells are covered with a layer of heavy mineral oil. The plate is then monitored kinetically on a standard ﬂuorescence plate reader. ADP is added to the measurement buffer if respiratory control ratios or data on mitochondria response in state 3 are required [34,35]. An example of the raw data output of such an assay is presented in Figure 4. As outlined above, increasing the biomass of mitochondria increases oxygen consumption correspondingly. As can be seen in Figure 4, results can be visualized in real time and analyzed either semiquantitatively using standard plate reader software or processed further for

Figure 4 Raw data output of mitochondrial oxygen consumption measurements using oxygen-sensitive probes. The plate map (left window) shows the raw intensity proﬁles at decreasing concentrations of mitochondrial protein. The top proﬁles are measured in state 3 (in the presence of ADP), while the bottom proﬁles are measured in state 2 (in the absence of ADP). Five state 3 sample proﬁles are displayed in the left windrow. Higher concentrations of mitochondrial protein cause more rapid deoxygenation, resulting in rapid signal changes. Lower concentrations result in lower rates of consumption, resulting in slower signal changes. Background signal is also displayed (bottom proﬁle).

362

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES

quantitative analysis [34]. This approach allows the identiﬁcation of both mitochondrial inhibitors and uncouplers, while the excellent reproducibility of the method (coefﬁcient of variation values of less than 3% being typical) allows low numbers of replicates to be used during analysis and, where appropriate, can facilitate single-point measurements. As is invariably the case when carrying out functional analysis on isolated mitochondria, success hinges on the quality of the mitochondrial preparation. Mitochondria should be characterized postisolation to ensure that they are well coupled and produce the expected levels of activity. These issues are dealt with in detail elsewhere [34,35]. Due to the sensitivity of oxygen consumption to altered mitochondrial function and the high-throughput nature of measurement, the approach is particularly well suited to the analysis of drug-induced mitochondrial dysfunction. It facilitates the rapid identiﬁcation of compounds that perturb mitochondrial function and contributes to the elucidation of the mechanisms involved. For example, inhibition of ATP synthase may be assessed using ADP-activated mitochondria using a single concentration in triplicate that would not be apparent during basal respiration. Testing for inhibitors of complexes I and II requires analysis of both glutamate/malate- and succinate-driven respiratory activity, respectively. In addition, uncoupling compounds may be missed or misidentiﬁed as inhibitors if tested at a single concentration [35]. Further delineation is possible using tetramethylphenylenediamine (TMPD) and ascorbate, allowing the introduction of electrons to complex IV via cytochrome c, thereby bypassing inhibition at an earlier point. Mechanistic insight therefore requires that compounds be tested at multiple concentrations on both basal and ADP-activated succinate and glutamate/ malate-driven respiration. The use of oxygen-sensitive probes provides the throughput needed for such a comprehensive analysis of mechanism while facilitating the generation of dose–response information. The method is amenable to mitochondria from a variety of tissues, including liver, skeletal muscle, and cardiac muscle [35], which is important because drug-induced toxicity can exhibit tissue speciﬁcity both in vitro and in vivo. This type of analysis can also prove useful when evaluating the function of mitochondria isolated from animals exposed to a test substance, as assessment of RCRs can provide an indication as to whether organ-speciﬁc mitochondrial function is impaired. This approach could be useful in drug development programs where such toxicities may be problematic, such as HIV or diabetes. In addition, as these probes facilitate low volume analysis (down to 15 µL) [39], it may also be applied in instances where the amount of biomaterial is extremely limited, such as from biopsy sample or in vivo studies examining mitochondrial dysfunction of mouse soleus muscle. 3.2. Extracellular Assessment of Oxygen Consumption Although analysis of isolated mitochondria yields high-quality informative data, it is blind to factors such as compound permeability, biotransformation, alter+ ations via cytosolic signaling molecules such as NO and Ca2 , and secondary

APPLICATIONS OF OXYGEN-SENSING PROBES

363

mitochondria effects. To address these, one must assess the effect of the treatment of interest on a relevant whole-cell model. Such analysis may be carried out on primary cells, such as hepatocytes, or on immortalized cell lines in a manner similar to that outlined for the assessment of isolated mitochondria. Oxidative cell lines such as HepG2 may be cultured and treated in a standard 96-well plate [44,46]. However, certain glycolytic cell lines do not provide a detectable level of oxygen consumption at or below 100% conﬂuence and must be trypsinized and concentrated prior to measurement [44]. Researchers should also be cognizant of inadequacies of certain immortalized cell-line models for assessing mitochondrial function [46,47] The data output of such measurements are similar to those outlined above for isolated mitochondrial assessments; however, the signals are generally lower, due to slower rates of oxygen consumption. Processing these proﬁles again allows a comparison between treated and untreated samples. Such analysis overcomes the limitation of many conventional mitochondrial toxicity assays such as ATP measurements, in that impaired mitochondrial function is detectable speciﬁcally and immediately, and uncoupling can be distinguished from inhibition in real time. In addition, using digitonin or saponin permeabilization, the cell membrane can be bypassed, allowing the mitochondria to be interrogated without disturbing relevant intracellular networks such as those between the mitochondrial network and the endoplasmic reticulum [48]. While assessment of oxygen consumption in isolation gives highly informative data, such measurements are also useful as part of a multiparametric approach. For example, parallel analysis of oxygen consumption, cellular ATP concentration, reactive oxygen production, and cell membrane integrity allow direct mitochondrial effects to be distinguished from nonspeciﬁc toxicity, such as membrane damage [45]. 3.3. Intracellular Assessment of Oxygen Concentration Water-soluble metalloporphyrin probes have proven to be highly applicable to intracellular oxygen analysis, demonstrating a ﬂexibility of probe design and efﬁcient probe loading [21]. The use of hydrophilic dyes such as Pt-coproporphyrin facilitates the use of macromolecular carriers such as proteins with a variety of molecular weights. These carriers limit probe aggregation within the cell, while also acting as a sink for damaging singlet oxygen. The size or type of this carrier may be changed to optimize signal stability, sensitivity to oxygen, photostability, calibration functions, loading efﬁciency, and the effects of reactive oxygen species. The use of water-soluble probes also facilitates simple probe loading using a number of simple techniques, such as electroporation, osmotic shock, liposomal transfer, simple diffusion, or incorporation into peptide vesicles. The complexity of an intracellular sensing scheme depends on throughput requirements and on the level of detail required; however, in all cases the principle remains the same. The delivery of a probe into the cell is critical to providing sufﬁcient signal. Once this is achieved, there are two main measurement options:

364

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES

ﬂuorescence imaging techniques, or more straightforward lifetime-based spectroﬂuorometry. The former allows low-throughput, high-resolution analysis of intracellular oxygen at the level of a single cell or a small number of cells, while the latter allows simpler measurement of cell populations (103 to 106 cells). The combination of live-cell ﬂuorescence imaging techniques and intracellular oxygen-sensing methodologies facilitates the measurement of small and transient changes in oxygen concentration at the level of the single cell or even to the subcellular level (see Figure 5) [21]. Higher resolution may be achieved

A

Normalised Intensity

C 300

Normalised Intensity

B 190 180 170 160 150 140 130 120 110 100 90 80 0

200

400 600 Time (sec)

800

250 200 150 100 50 0 100 200 300 400 500 600 700 800 Time (sec)

Figure 5 Sensing intracellular oxygen using ﬂuorescence imaging microscopy. (A) Confocal laser scanning image of ﬁxed A549 cells loaded with intracellular probe and co-stained with the nuclear stain DAPI. (B) Analyzed live cell imaging data of HeLa cells loaded with either PdCPK (dashed curve) or PtCPK (solid curve) probes and treated with 1 µM valinomycin. Uncoupling results in increased oxygen uptake, which in turn reduces the concentration of intracellular oxygen, resulting in increased probe intensity. The higher rate of intensity increase by the more sensitive PdCPK probe (0 to 10 kPa) indicates a low basal level of intracellular oxygen in HeLa cells. (C) Analyzed live cell imaging data of A549 cells loaded with PtCPK and treated with 500 µM of the ryanadoine receptor (RyR1) agonist 4-CMC in the presence of (dashed curve) or absence (solid curve) of 500-µM RyR1 channel blocker procaine. Treatment causes an increased ﬂux of Ca2+ through RyR1, which causes a parallel increase in oxygen uptake, demonstrating the links between calcium signaling and metabolism. The effect is blocked by procaine.

365

APPLICATIONS OF OXYGEN-SENSING PROBES

by the use of confocal laser scanning microscopy (Figure 5A); however, the speed of measurement may be too slow to reveal transient cellular effects. Such measurements require highly photostable probes, due to the high intensity of continuous illumination employed in ﬂuorescence microscopy. Recently, a family of probes based on water-soluble Pt(II) and Pd(II) coproporphyrin ketones have been shown to have the required photostability and sensitivity to measure intracellular oxygen over the entire physiological range. The loading efﬁciency by passive transfer techniques is sufﬁcient (see Figure 5A) to detect the effects of different compounds, such as valinomycin (Figure 5B), and Ca2+ inﬂux caused by the ryanodine receptor agonist 4-chloro-m-cresol (Figure 5C). Thus, these probes allow simple high-resolution measurement of oxygen using widely available cell imaging instrumentation [21]. The measurement of intracellular oxygen by spectroﬂuorometry can also detect changes induced by effectors of oxidative phosphorylation. Prompt ﬂuorescence intensity measurements are not favored for such intracellular measurement, due to a lack of both sensitivity and signal stability. Lifetime-based sensing is therefore the preferred mode, as it gives a direct correlation with absolute oxygen concentrations and is relatively easily achieved using spectroﬂuorometrs or time-resolved plate ﬂuorometers. Lifetime-based intracellular measurements are also more robust, as they are independent of ﬂuctuations in intensity and scattering and are more resistant to compound interferences. For measurement, cells are loaded with probe by techniques, such as liposomal transfer or peptide vesicles, followed by measurement on a time-resolved plate ﬂuorometer. The lifetime of the probe within the cell is calculated either by measuring probe intensity at two delay times following successive intensity pulses, thereby facilitating the calculation of probe lifetime using a simple equation [49] (Figure 6A), or by measurment of the decay curve of the probe using

A 42

FCCP Rotenone

40

B

Lifetime (µs)

Lifetime (µs)

38 36 34 32 30 28 26 24 0

500

1000 1500 2000 2500 3000 Time (sec)

33 32 31 30 29 28 27 26 25 24 23 22 21

1 µM Rotenone 100 nM Rotenone 10 nM Rotenone 1 nM Rotenone

0

200 400 600 800 1000 1200 1400 1600 Time (sec)

Figure 6 Lifetime-based sensing of intracellular oxygen in Jurkat T cells measured by time-resolved ﬂuorometry on a Victor plate ﬂuorometer (Perkin–Elmer). (A) Jurkat T cells uncoupled with 1 µM FCCP or inhibited with 1 µM rotenone, with both effectors added at t 1500 . (B) Effect of different concentrations of rotenone on intracellular oxygen in Jurkat T cells with effector addition at t 700 following measurement of basal lifetime.

366

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES

time-correlated single-photon counting. Basal levels of intracellular probe lifetime in Jurkat T cells (Figure 5B and C) (28 to 29 µs) suggest substantial oxygen gradients between the intracellular and extracellular environments [21]. Addition of uncouplers such as carbonyl cyanide p-triﬂuoro methoxy phenylhydrazone (FCCP) causes an immediate and substantial decrease in the concentration of intracellular oxygen. This is due to an increased rate of oxygen consumption and is reﬂected by an increase in measured probe lifetime (Figure 6B). Alternatively, inhibitors of oxidative phosphorylation such as rotenone (Figures 6B and 5C) result in intracellular probe lifetimes returning to basal levels (23 µs), representing an intracellular oxygen concentration of approximately 21 kPa [21]. This technique therefore facilitates simple, lifetime-based sensing of intracellular oxygen on a medium- to high-throughput scale. 3.4. Alternative Approaches The main alternative to the use of water-soluble probes is to use a solid-state oxygen-sensitive coating, with the oxygen-sensitive dye embedded in a polymer matrix [23,50] (see Chapter 14). This approach has been applied successfully to the assessment of bacterial [24], yeast [31], and mammalian cells [51,52]. Its main advantage is the protection afforded the dye from potentially interfering compounds present in the test sample, although the porosity of the matrix may provide access to low-molecular-weight compounds. There is also a larger choice of compatible oxygen-sensitive dyes for such detection schemes. There are, however, a number of associated disadvantages. The high concentration of dye required can result in the generation of signiﬁcant quantities of singlet oxygen during operation of the sensor, thereby imposing signiﬁcant phototoxicity. This is particularly true when the biological entity of interest is in direct contact with the sensor. There are also issues of biocompatibility whereby the hydrophobic nature of these coatings can prevent normal adherent cell growth. Sensor response time can also be a limiting factor [32]; however, by far the most signiﬁcant disadvantage of the solid-state approach is the lack of measurement ﬂexibility. The necessity for solid-state coatings limits the measurement format, which in turn limits the sample volume that can be measured. Intracellular analysis has also been attempted using solid-state oxygensensitive probes with the oxygen-sensitive dye embedded in a polymer matrix [23,50]. These probes have been introduced into cells in the form of solid-state particles, whereby the particles are loaded by either microprojectile delivery or in the case of macrophage analysis, by phagocytosis [53–55]. While particulate probes can provide the requisite photostability, loading by projectile delivery or endocytosis is rather inefﬁcient and can cause irreparable damage to the cell. In addition, random distribution of the small number of particulate sensors within the cell may yield a dyshomogeneous representation of intracellular oxygen distribution. The generation on singlet oxygen as a by-product of probe excitation can also be problematic. Lifetime-based imaging has also been

CONCLUSIONS AND FUTURE PERSPECTIVES

367

described for sensing intracellular oxygen using either intensity-based [56] or frequency-based [57] sensing. Termed ﬂuorescence lifetime imaging microscopy (FLIM), the former relies on measurement at different delay times, as described above (Figure 5A), while the latter relies on changes in frequency of emitted light following modulated excitation. These are more complex techniques that can give highly detailed information of the absolute concentrations of oxygen at different parts of a single cell. However, they rely on custom-designed instrumentation.

4. CONCLUSIONS AND FUTURE PERSPECTIVES Recent years have seen a growing appreciation of the role played by the mitochondrion in maintaining cellular homeostasis. It is no longer seen as an isolated organelle but as a complex network integrated into a wide variety of fundamental cellular signaling pathways. Knowledge of altered cellular oxygen distribution and consumption, and by extension methods for such analysis, has therefore become relevant to an ever-broadening community of researchers. As outlined here, such analysis may be achieved using water-soluble oxygen probes, as they allow detailed investigation of the involvement of mitochondria in speciﬁc signaling pathways as well as more applied tasks, such as detection of mitochondrial toxicity by putative drugs. The information content associated with oxygen measurements, when allied to the throughput of ﬂuorescence-based measurement approaches, offers a powerful tool for the investigation of mitochondrial function. In addition, the ability to measure intracellularly provides a means to investigate the signal transduction involved in speciﬁc mitochondria-linked signaling pathways, such as those involved in the cellular response to hypoxia and calcium signaling. There are a number of foreseeable avenues through which to realize the full potential of this technology. The imaging scheme may be enhanced by employing time-resolved microscopy, which functions to increase sensitivity to oxygen while increasing the resolution and contrast of the images. Furthermore, these probes are particularly well suited to multiplexed time-resolved imaging in parallel with other long-lived probes, such as europium chelates. Multiplexed intracellular and extracellular oxygen sensing may be achieved using spectrally and temporally distinct metalloporphyrin probes. Lifetime-based sensing using FLIM is also a distinctly achievable goal that would give access to absolute concentrations of intracellular oxygen. The plate-based intracellular sensing schemes offer further scope to increase the throughput of this technique. The generation of data on the diversity of response across cell populations is also achievable. Overall, these probes and sensing schemes offer a new and exciting means to assess a range of parameters governing the physiological status of the cell from overall viability, to mitochondrial function, to speciﬁc pathways of signal transduction and cell death.

368

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES

REFERENCES 1. Waterhouse NJ, Ricci J-E, Green DR. And all of a sudden it’s over: mitochondrial outer-membrane permeabilization in apoptosis. Biochimie. 2002;84(2–3):113–121. 2. Mattson MP, Kroemer G. Mitochondria in cell death: novel targets for neuroprotection and cardioprotection. Trends Mol Med. 2003;9(5):196–205. 3. Shults CW. Mitochondrial dysfunction and possible treatments in Parkinson’s disease: a review. Mitochondrion. 2004;4(5–6):641–648. 4. Lewis W, Day BJ, Copeland WC. Mitochondrial toxicity of NRTI antiviral drugs: an integrated cellular perspective. Nat Rev Drug Disc. 2003;2(10):812–822. 5. Duchen MR. Mitochondria in health and disease: perspectives on a new mitochondrial biology. Mol Aspects Med. 2004;25(4):365–451. 6. Amacher DE. Drug-associated mitochondrial toxicity and its detection. Curr Med Chem. 2005;12(16):1829–1839. 7. Hagen T, Taylor CT, Lam F, Moncada S. Redistribution of intracellular oxygen in hypoxia by nitric oxide: effect on HIF1 {alpha}. Science, 2003;1975–1978. 8. Chandel NS, Budinger GRS. The cellular basis for diverse responses to oxygen. Free Radic Biol Med. 2007;42(2):165–174. 9. Fandrey J, Gorr TA, Gassmann M. Regulating cellular oxygen sensing by hydroxylation. Cardiovasc Res. 2006;71(4):642–651. 10. Schoﬁeld CJ, Ratcliffe PJ. Signalling hypoxia by HIF hydroxylases. Biochem Biophys Res Commun. 2005;338(1):617–626. 11. Ghafourifar P, Cadenas E. Mitochondrial nitric oxide synthase. Trends Pharmacol Sci. 2005;26(4):190–195. 12. Nicholls DG. Mitochondria and calcium signaling. Cell Calcium. 2005;38 (3–4):311–317. 13. Hayakawa Y, Nemoto T, Iino M, Kasai H. Rapid Ca2+ -dependent increase in oxygen consumption by mitochondria in single mammalian central neurons. Cell Calcium. 2005;37(4):359–370. 14. Clark LC. Electrochemical device for chemical analysis. US patent 2,913,386; 1959. 15. Hutter E, Unterluggauer H, Garedew A, Jansen-Durr P, Gnaiger E. High-resolution respirometry: a modern tool in aging research. Exp Gerontol. 2006;41:103–109. 16. Lakowicz JRe. Principles of Fluorescence Spectroscopy. New York: Plenum Press; 1999. 17. Wilson DF, Rumsey WL, Vanderkooi JM. Oxygen distribution in isolated perfused liver observed by phosphorescence imaging. Adv Exp Med Biol. 1989;248:109–115. 18. Wilson DF, Vanderkooi JM, Green TJ, Maniara G, DeFeo SP, Bloomgarden DC. A versatile and sensitive method for measuring oxygen. Adv Exp Med Biol. 1987;215:71–77. 19. Wilson DF, Vinogradov SA, Rozhkov V, Creed J, Rietveld I, Pastuszko A. Monitoring the dynamics of tissue oxygenation in vivo by phosphorescence quenching. Adv Exp Med Biol. 2003;540:1–5. 20. Dunphy I, Vinogradov SA, Wilson DF. Oxyphor R2 and G2: phosphors for measuring oxygen by oxygen-dependent quenching of phosphorescence. Anal Biochem. 2002;310(2):191–198.

REFERENCES

369

21. O’Riordan TC, Fitzgerald K, Ponomarev GV, et al. Sensing intracellular oxygen using near infrared phosphorescent probes and live-cell ﬂuorescence imaging. Am J Physiol Regul Integr Comp Physiol. 2007;292:1613–1620. 22. Hynes J, Floyd S, Soini AE, O’Connor R, Papkovsky DB. Fluorescence-based cell viability screening assays using water-soluble oxygen probes. J Biomol Screening. 2003;8(3):264–272. 23. Papkovsky DB. Methods in optical oxygen sensing: protocols and critical analyses. Meth Enzymol. 2004;381:715–735. 24. John GT, Klimant I, Wittmann C, Heinzle E. Integrated optical sensing of dissolved oxygen in microtiter plates: a novel tool for microbial cultivation. Biotechnol Bioeng. 2003;81(7):829–836. 25. Papkovsky DB, Ponomarev GV, Wolfbeis OS. Longwave luminescent porphyrin probes. Spectrochim Acta A. 1996;52(12):1629–1638. 26. Atwater BW. Substituent effects on the excited-state properties of platinum meso-tetraphenylporphyrins. J Fluorescence. 1992;2(4):237–246. 27. Singer E, Duveneck GL, Ehrat M, Widmer HM. Fiber optic sensor for oxygen determination in liquids. Sensors Actuators A Phys. 1994;42(1–3):542–546. 28. Carraway ER, Demas JN, DeGraff BA. Luminescence quenching mechanism for microheterogeneous systems. Anal. Chem. 1991;65:332–336. 29. Bacon JR, Demas JN. Determination of oxygen concentrations by luminescence quenching of a polymer-immobilized transition-metal complex. Anal Chem. 1987;59(23):2780–2785. 30. Hartmann P, Leiner MJP, Lippitsch ME. Luminescence quenching behavior of an oxygen sensor based on a Ru(II) complex dissolved in polystyrene. Anal Chem. 1995;67(1):88–93. 31. O’Riordan TC, Buckley D, Ogurtsov V, O’Connor R, Papkovsky DB. A cell viability assay based on monitoring respiration by optical oxygen sensing. Anal Biochem. 2000;278(2):221–227. 32. Arain S, Weiss S, Heinzle E, John GT, Krause C, Klimant I. Gas sensing in microplates with optodes: inﬂuence of oxygen exchange between sample, air, and plate material. Biotechnol Bioeng. 2005;90(3):271–280. 33. O’Mahony FC, O’Donovan C, Hynes J, Moore T, Davenport J, Papkovsky DP. Optical oxygen micro-respirometry as a platform for environmental toxicology and animal model studies. Environ Sci Technol. 2005;39:5010–5014. 34. Will Y, Hynes J, Ogurtsov VI, Papkovsky DB. Analysis of mitochondrial function using phosphorescent oxygen-sensitive probes. Nat Protocols. 2007;1(6):2563–2572. 35. Hynes J, Marroquin LD, Ogurtsov VI, et al. Investigation of drug-induced mitochondrial toxicity using ﬂuorescence-based oxygen-sensitive probes. Toxicol Sci. 2006;92(1):186–200. 36. Papandreou I, Powell A, Lim AL, Denko N. Cellular reaction to hypoxia: sensing and responding to an adverse environment. Mutat Res. 2005 (Jan 6);569(1–2):87–100. 37. Stern O, Volmer M. Uber die ablingungszeit der ﬂuoreszenz. Phys Z. 1919;20: 183–188. 38. Vanderkooi JM, Maniara G, Green TJ, Wilson DF. An optical method for measurement of dioxygen concentration based upon quenching of phosphorescence. J Biol Chem. 1987;262(12):5476–5482.

370

USE OF OXYGEN-SENSITIVE FLUORESCENT PROBES

39. O’Mahony FC, O’Donovan C, Hynes J, Moore T, Davenport J, Papkovsky DB. Optical oxygen microrespirometry as a platform for environmental toxicology and animal model studies. Environ Sci Technol. 2005;39(13):5010–5014. 40. O’Donovan C, Twomey E, Alderman J, Moore T, Papkovsky D. Development of a respirometric biochip for embryo assessment. Lab on a Chip. 2006;6(11):1438–1444. 41. Alderman J, Hynes J, Floyd SM, Kruger J, O’Connor R, Papkovsky DB. A lowvolume platform for cell-respirometric screening based on quenched-luminescence oxygen sensing. Biosens Bioelectron. 2004;19(11):1529–1535. 42. O’Donovan C, Hynes J, Yashunski D, Papkovsky DB. Phosphorescent oxygensensitive materials for biological applications. J Mater Chem. 2005;15(27–28):2946– 2951. 43. Weiss S, John GT, Klimant I, Heinzle E. Modeling of mixing in 96-well microplates observed with ﬂuorescence indicators. Biotechnol Prog., 2002;4:821–830. 44. Hynes J, Hill R, Papkovsky DB. The use of a ﬂuorescence-based oxygen uptake assay in the analysis of cytotoxicity. Toxicol In Vitro. 2006;20(5):785–792. 45. Hynes J, O’Riordan TC, Curtin J, Cotter TG, Papkovsky DB. Fluorescence based oxygen uptake analysis in the study of metabolic responses to apoptosis induction. J Immunol Methods. 2005;306:193–201. 46. Marroquin LD, Hynes J, Dykens JA, Jamieson JD, Will Y. Circumventing the Crabtree effect: replacing media glucose with galactose increases susceptibility of HepG2 cells to mitochondrial toxicants. Toxicol Sci. 2007; Epub ahead of print. 47. Rossignol R, Gilkerson R, Aggeler R, Yamagata K, Remington SJ, Capaldi RA. Energy substrate modulates mitochondrial structure and oxidative capacity in cancer cells. 2004;64:985–993. 48. Rizzuto R, Duchen MR, Pozzan T. Flirting in Little Space: The ER/Mitochondria Ca2+ Liaison. Sci STKE , 2004;215:Re 1. 49. O’Riordan TC, Zhdanov AV, Ponomarev GV, Papkovsky DB. Analysis of intracellular oxygen and metabolic responses of mammalian cells by time-resolved ﬂuorometry. Anal. Chem. 2007;79(24):9414–9419. 50. Mills A. Controlling the sensitivity of optical oxygen sensors. Sensors Actuators B Chem. 1998;51(1–3):60–68. 51. Wodnicka M, Guarino RD, Hemperly JJ, Timmins MR, Stitt D, Pitner JB. Novel ﬂuorescent technology platform for high throughput cytotoxicity and proliferation assays. J Biomol Screening. 2000;5(3):141–152. 52. Deshpande RR, Koch-Kirsch Y, Maas R, John GT, Krause C, Heinzle E. Microplates with integrated oxygen sensors for kinetic cell respiration measurement and cytotoxicity testing in primary and secondary cell lines. Assay Drug Dev Technol. 2005 (Jun);3(3):299–307. 53. Ji J, Rosenzweig N, Jones I, Rosenzweig Z. Molecular oxygen-sensitive ﬂuorescent lipobeads for intracellular oxygen measurements in murine macropha ges. Anal Chem. 2001;73(15):3521–3527. 54. Koo YE, Cao Y, Kopelman R, Koo SM, Brasuel M, Philbert MA. Real-time measurements of dissolved oxygen inside live cells by organically modiﬁed silicate ﬂuorescent nanosensors. Anal Chem. 2004;76(9):2498–2505. 55. Xu H, Aylott JW, Kopelman R, Miller TJ, Philbert MA. A real-time ratiometric method for the determination of molecular oxygen inside living cells using

REFERENCES

371

sol-gel-based spherical optical nanosensors with applications to rat C6 glioma. Anal Chem. 2001;73(17):4124–4133. 56. Zhong W, Urayama P, Mycek MA. Imaging ﬂuorescence lifetime modulation of a ruthenium-based dye in living cells: the potential for oxygen sensing. J Phys D Appl Phys. 2003;36(14):1689–1695. 57. Schmalzlin E, van Dongen JT, Klimant I, et al. An optical multifrequency phase-modulation method using microbeads for measuring intracellular oxygen concentrations in plants. Biophys J. 2005;89(2):1339–1345.

14 MITOCHONDRIAL DYSFUNCTION ASSESSED QUANTITATIVELY IN REAL TIME BY MEASURING THE EXTRACELLULAR FLUX OF OXYGEN AND PROTONS David Ferrick, Min Wu, Amy Swift, and Andy Neilson Seahorse Bioscience, Billerica, Massachusetts

1. Introduction 2. The XF24 analyzer 3. Assessing mitochondrial function and cellular bioenergetics 3.1. Uncouplers of mitochondrial respiration 3.2. Inhibitors of mitochondrial respiration 4. Conclusions

373 374 376 377 378 380

1. INTRODUCTION Mitochondria generate 95% of cellular ATP via oxidative phosphorylation and are central to intermediary metabolism, free-radical generation, and regulating apoptosis [1,2]. Loss of mitochondrial function is tolerated by most cells until a threshold is reached when lack of ATP generation potential endangers the cell. Glycolytic ﬂux accelerates to compensate, but this is ﬁnite, and at some point, which is different for different cells, the cell will die via necrosis or apoptosis. Their crucial role in maintaining cell viability and in many other Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

373

374

MITOCHONDRIAL ASSESSMENT IN LIVING CELLS

metabolic pathways renders mitochondrial integrity a key target for compound toxicity. Assessing toxicity in isolated mitochondria is a common and useful method, but it suffers from (1) the inability to identify or predict chronic toxicity, (2) less physiological relevance since they are devoid of cellular constituents, and (3) the inability to identify toxic effects of compounds requiring bioactivation or where the metabolite is the toxin. Valuable insight into the physiological state of living cells, and the changes in these cells in response to experimental intervention, has been gained via independent monitoring of oxygen consumption using Clark electrodes [13] or soluble probes (see Chapter 13) and/or the rate of extracellular acidiﬁcation using a microphysiometer [14,15]. Under typical in vitro cell culture conditions, the oxygen consumption rate (OCR) is a direct measure of mitochondrial respiration, and the extracellular acidiﬁcation rate (ECAR) is dominated by lactic acid production from increased glycolytic carbon ﬂux. Measuring both parameters simultaneously enables a more comprehensive assessment of cellular energetics and provides insight into the dynamic interplay between these two energy-yielding pathways. Recognition of the value of simultaneous monitoring of OCR and ECAR is underscored by the growing number of investigators using a variety of techniques to achieve this measurement [16–20]. The data presented in this chapter were generated using the Seahorse XF24 extracellular ﬂux analyzer. This instrument measures the ﬂux of oxygen and protons of adherent cells simultaneously in a microplate [21]. Extracellular ﬂux (XF) technology enables rapid real-time detection of metabolic changes in cells in vitro, speciﬁcally changes in mitochondrial respiration and glycolysis [21–24]. Thus, XF technology is well suited not only to identifying compounds with mitochondrial liabilities, but also in elucidating their mechanisms of action.

2. THE XF24 ANALYZER The XF24 is a fully integrated 24-well instrument that utilizes dedicated assay kits to measure rates of oxygen consumption and efﬂux of protons by cells in culture. Each assay kit contains a disposable sensor cartridge embedded with 24 pairs of ﬂuorescent biosensors (oxygen and pH), which are coupled to a ﬁber-optic waveguide. The waveguide delivers light at various excitation wavelengths (oxygen = 532 nm, pH = 470 nm) and transmits the resulting ﬂuorescent signal through optical ﬁlters (oxygen = 650 nm, pH = 530 nm) to a set of highly sensitive photodetectors. The instrument currently supports the simultaneous measurement of pH and oxygen and has the capacity for two additional sensors (in development), allowing the user to select an array of up to four sensors, speciﬁc for a particular application. In the measurement position, the sensor cartridge is lowered to a stop point in all 24 wells of the culture plate, thereby forming a small microchamber (7 µL in volume), so that changes in oxygen tension and pH are obtained within 1.5 to 4 minutes (Figure 1). During this time, analyte levels are measured every

375

THE XF24 ANALYZER 7.45

150

7.40

OCR, pmol/min

140

7.35

ECAR, mpH/min

130

7.30 7.25

120 100

pH

Oxygen (mm Hg)

160

1

2

3

4 5 Time (min)

6

7

7.20

Well 1

Compound Sensors Analyte (O2) Cell

Well 2

Figure 1 XF24 analyzer: how real-time measurements of OCR and ECAR are made. (See insert for color representation of ﬁgure.)

22 seconds until the oxygen concentration drops approximately 30 mmHg and the media pH declines up to 0.4 pH unit. It is important that the measurement time be long enough to calculate accurate rates of oxygen consumption and extracellular acidiﬁcation but not enough to affect the biology of the cells via hypoxia and/or acidic conditions. Because the sensors that are positioned over the cells are illuminated and monitored from above (Figure 1), autoﬂuorescence due to cells, culture components, or test compounds are effectively eliminated. Each sensor cartridge is also equipped with four reagent delivery chambers per well for injecting test agents during an assay. These are commonly used to generate dose curves and/or to add agonists/antagonists automatically without having to interrupt the assay. All assay parameters are controlled by a dedicated computer and software. Over 50 primary and immortalized cells have been evaluated in the XF24. The only constraint is that the cells must be on the bottom of the well. Suspension cells can be assayed by immobilizing them using any number of adherent matrixes, such as collagen, Cell Tak, or Matrigel. Some of the most commonly used cell types are C2C12, HepG2, SK-Hep1, 3T3L1, L6, CHO, SH-SY5Y, LN18, MCF-7, H460, A549, PC-3, PC-12, HeLa, and primary rat, mouse, and human skeletal muscle cells, adipocytes, hepatocytes, and cardiomyocytes. Prior to the start of the assay, all biosensors are calibrated independently using an automated routine that determines a unique sensor gain based on output in a calibration reagent of known pH and oxygen concentration. During the assay, baseline rates are measured at least twice. OCR is reported as pmol/min and ECAR as mpH/min. Testing chemical(s) or biological

376

MITOCHONDRIAL ASSESSMENT IN LIVING CELLS

solution(s) are preloaded in the reagent delivery chambers of the sensor probe and are subsequently injected pneumatically into the media. After 3 to 5 minutes of mixing and waiting, postexposure OCR and ECAR measurements are made multiple times. Because XF measurements are nondestructive, the metabolic rate of the same cell population can be monitored repeatedly over time, while up to four different testing agents can be injected sequentially or simultaneously into each well. Upon completion of an assay, other types of biological assays, including normalization protocols, such as cell viability, can be performed on the same cells. Importantly, this experimental design permits use of paired-comparison statistics, which are among the most powerful for revealing subtle changes in dynamic systems, such as cells exposed to xenobiotics.

3. ASSESSING MITOCHONDRIAL FUNCTION AND CELLULAR BIOENERGETICS To demonstrate the linkage between OCR and mitochondrial respiration and between ECAR and glycolysis, A549 cells were exposed sequentially to each of three well-deﬁned small molecule modulators of mitochondrial and glycolytic energy metabolism: the respiratory uncoupler 2,4-dinitrophenol (2,4-DNP); the hexokinase inhibitor 2-deoxyglucose (2-DG), which blocks glycolysis; and rotenone, which speciﬁcally inhibits mitochondrial NADH dehydrogenase/complex I (Figure 2A) [21]. As anticipated, 2,4-DNP administration increased OCR and evoked a concomitant increase in ECAR. Subsequent addition of 2-DG to the same wells elicited a rapid decrease of the 2,4-DNP-stimulated ECAR to well below the baseline rate while OCR remained unaffected. Finally, addition of the mitochondrial respiration inhibitor rotenone abolished OCR and further reduced ECAR. These data demonstrate that the technology can readily detect mitochondrial uncoupling and respiratory inhibition, as well as both acceleration and inhibition of glycolysis. To correlate these changes observed in OCR and ECAR directly to ATP production, we measured cellular ATP concentration in equivalently treated cells exposed for 45 minutes to (a) vehicle, (b) 2,4-DNP, (c) 2,4-DNP plus 2-DG, and (d) 2,4-DNP plus 2-DG plus rotenone (Figure 2B). We observed that the 2,4-DNP treatment reduced the ATP content by approximately 40% compared to vehicle-treated cells, due to the uncoupling of ADP phosphorylation from respiration. 2-DG in combination with 2,4-DNP caused a dramatic decrease in ATP, to less than 10% of the control cells, as glycolytic ATP production was now also blocked. Finally, the combination of the mitochondrial respiration inhibitor, rotenone, along with the glycolysis inhibitor, 2-DG, and uncoupler, 2,4-DNP, effectively depletes all the ATP. Since the ATP concentration of a cell population is determined by both the number of viable cells present and their relative metabolic rates, the cell viability of each population must be determined to ensure similar viability of control and treated cells. Therefore, cell viability was measured in a parallel experiment with

ASSESSING MITOCHONDRIAL FUNCTION AND CELLULAR BIOENERGETICS

400 350 300 250 200 150 100 50 0

Rotenone injection

B 100 80 60 40

OCR ECAR

−16 −8 14 28 42 56 70 84 Time (minute)

20 0

% relative to control

2,4-DNP 2-DG injection injection

ECAR (mpH/min)

OCR (pmoles/min)

A

377

150% 100% 50%

0% 2,4-DNP 2-DG Rotentone

− − −

+ − −

+ + −

+ + +

Figure 2 Real-time analysis of bioenergetic pathways in A549 cells by perturbing them with small molecule metabolic modulators: (A) mitochondria uncoupler 2,4-DNP (100 µM), glycolysis inhibitor 2-DG (100 mM), and mitochondria complex I inhibitor, rotenone (1 µM), were injected sequentially at the into each well containing A549 cells time points indicated after baseline rate measurement; (B) intracellular ATP level (bars) in response to 2,4-DNP (100 µM), 2-DG (100 mM), and rotenone (1 µM) treatment (45 minutes) individually or in combinations as indicated. The viability of the treated cells remained similar to that of control cells, as shown by Calcein AM stain (line). 10,000 cells per well were seeded 24 hours prior to the assay.

identically treated cells using calcein AM staining. The cell viability remained essentially constant among all the experimental samples (Figure 2B), thus indicating that the reduced ATP levels in the A549 cells were due to metabolic changes and not cell death, as a result of compound toxicity. No doubt, had the cells been exposed for longer periods of time, their viability would have been reduced signiﬁcantly. But this highlights the ability of this technology to make subacute, physiological measurements prior to overt viability changes. 3.1. Uncouplers of Mitochondrial Respiration Compounds that dissipate the mitochondrial membrane potential uncouple the electron transfer system from adenylate phosphorylation. Since the ETS is now freed from doing the work required to pump protons against the normal electrochemical gradient, electron ﬂux, and hence oxygen consumption, increase. This is demonstrated in Figure 3 by exposing the human liver cell line, HepG2, to the uncoupler, carbonylcyanide-p-triﬂuoromethoxyphenylhydrazone (FCCP). In Figure 3A, cells were exposed to increasing concentrations of FCCP that were injected automatically during the experiment. Plotting the OCR and ECAR responses together produced a bioenergetic chart indicative of both mitochondrial respiration and glycolysis, respectively. By injecting sequentially three dose series containing four escalating concentrations of FCCP, low (0.01, 0.03, 0.05, and 0.1 µM), medium (0.2, 0.4, 0.8, and 1.6 µM) and high (3.2, 6.4, 12.8, and 25.6 µM), the minimum, maximal, and toxic responses were readily determined.

378

MITOCHONDRIAL ASSESSMENT IN LIVING CELLS

A

300

B

250 Basal

A

B

C

D

Low Dose Series

0mM

0.01mM

0.03mM

0.05mM

0.10mM

Medium Dose Series

0mM

0.20mM

0.40mM

0.80mM

1.60mM

High Dose Series

0mM

3.20mM

6.40mM

12.80mM

25.60mM

200 150 100

OCR %

Injection Port

50 0 0.01

OCR vs ESCAR (% of Base Line) (Avg)

0.1

1

10

−50 100

FCCP (uM)

350 300

C

200 150 100 50 0 −50 −25 −12.5

110 Percent Maximal OCR

OCR %

250

0

12.5 25 37.5 50 62.5 75 ECAR %

90

EC50 = 0.58 µM

70 50 30 10 −10 Vehicle 10−7.0 10−6.5 10−6.0 10−5.5 [FCCP], M

Figure 3 Multiparameter analysis of the dose-dependent response of HepG2 to FCCP. Escalating doses of FCCP were injected sequentially every 9 minutes into wells containing overnight cultures of the human liver cell line, HepG2, seeded at 30,000 cells/well. (A) OCR data (y-axis) are plotted against ECAR (x -axis). Three triplicate wells, each injected sequentially four times with increasing amounts of a low-, medium-, and high-dose series of FCCP, were measured simultaneously for OCR and ECAR after each injection. The data are represented as the average percent above baseline ± STD. (B) Graphing only OCR versus the concentration of FCCP shows the minimum, maximum, and toxic dose responses. (C) The dose–response curve and EC50 were calculated using GraphPad Prism 4.

Figure 3B shows only the mitochondrial component of the data, OCR, versus FCCP concentration. After the maximal response is reached, higher concentrations become acutely toxic to the cells, resulting in decreased OCR, in accordance with other studies on FCCP [25]. The ability to inject up to four compounds into each well facilitates generation of EC50 values; the data for Figure 3C are derived from 12 wells and took 1.5 hours. 3.2. Inhibitors of Mitochondrial Respiration The four redox active complexes (I to IV) of the mitochondria, plus the ATP-synthase, complex V, make up the ETS responsible for oxidative phosphoration (OXPHOS). Compounds that impair redox reactions of the mitochondria, such as in the TCA cycle or ETS, will repress oxygen consumption. In Figure 4 we demonstrate the sensitivity and simplicity of identifying and determining the potency of compounds that inhibit mitochondrial respiration using three well-characterized drugs with mitochondrial liabilities: rotenone, which blocks

ASSESSING MITOCHONDRIAL FUNCTION AND CELLULAR BIOENERGETICS

% of baseline rate

ECAR

100 80 60

OCR

40 20

IC50 = 38 nM 0 1.00E-12 1.00E-09 1.00E-06 Rotenone

% of baseline rate

B 140

A 140 120

379

120

ECAR

100 80 60

OCR

40 20

IC50 = 14 nM 0 1.00E-12 1.00E-09 1.00E-06 Antimycin

% of baseline rate

C 140 120

ECAR

100 80

OCR

60 40 20

IC50 = 400 nM 0 1.00E-12 1.00E-09 1.00E-06 Oligomycin

Figure 4 Dose-dependent inhibition of mitochondrial function in HepG2 cells. The human liver cell line, HepG2, was seeded in XF24 microtiter plates at 30,000 cells/well and incubated overnight. Escalating doses of three compounds known to impair mitochondrial respiration were injected into wells automatically, and the OCR and ECAR were subsequently measured. OCR response curve and EC50 for (A) rotenone, (B) antimycin, and (C) oligomycin. These data were calculated using GraphPad Prism 4.

the NADH–ubiquinone oxidoreductase of complex I; antimycin, which inhibits cytochrome c reductase of complex III; and oligomycin, an inhibitor of complex V, F1 FO -ATPase. In Figure 4, OCR reﬂects the decrease in mitochondrial respiration. EC50 values are indicated at the lower left of the graph. ECAR shows the ability of the cells to counter the loss of aerobic ATP productin. A decrease in ECAR under these circumstances would suggest another mechanism of toxicity, and resulting cell death. Figure 4A shows the dose-dependent inhibition of mitochondrial respiration in HepG2 cells by the complex I inhibitor, rotenone. Inhibition of complex III by antimycin is illustrated in Figure 4B, and inhibition of complex V by oligomycin, in Figure 4C. The six-dose concentration curve for each compound was generated in a single XF24 microtiter plate, and the data are in accord with a similar study using other parameters to assess mitochondrial function [25]. Extracellular ﬂux assays provide comparable performance to biochemical and radioactive methods, with higher throughput and without requiring parallel assays,

380

MITOCHONDRIAL ASSESSMENT IN LIVING CELLS

the preparation of complicated labels, or radioactive materials. Fluorescent interference is negated because detection of the ﬂuorescent signal occurs from above the cell monolayer and not via a signal traversing the cells and media. This is particularly advantageous in the drug screening arena, where compound autoﬂuorescence is common. Moreover, this technique also monitors the other major source of energy production, glycolysis. This can be very informative and facilitates mechanistic understanding, as subtle changes in mitochondrial function are almost always reﬂected in compensatory responses in glycolysis as the cells strive to maintain their bioenergetic budget [21,22]. Compounds that affect aerobic respiration may also indirectly or directly inﬂuence glycolysis. It should be noted in this context that XF assays are also useful for the identiﬁcation of agents that stabilize mitochondrial integrity under pathogenic conditions. This can be exploited to discover therapeutics that target mitochondrial diseases, the number of which is growing rapidly as more is learned about the mitochondria [1–3] (see Chapter 11). 4. CONCLUSIONS Assessing mitochondrial function earlier and more comprehensively in the drug development process will help avoid costly late-stage attrition, and more important, improve drug safety. As we develop novel therapeutics aimed at restoring equilibrium in diseases that are metabolic in nature, the likelihood of concomitant mitochondrial dysfunction will increase and is likely to become more insidious, thereby necessitating more sensitive and frequent screening. To that end, we describe here a novel technology for noninvasively quantifying cellular bioenergetics via measuring mitochondrial respiration and proton efﬂux simultaneously. The current instrument is capable of revealing compounds that uncouple or inhibit mitochondrial function while monitoring compensatory changes in glycolytic ﬂux. The technology lends itself to powerful paired-comparison statistics that further increase its sensitivity. Potentially deleterious effects of a given compound can be detected in minutes, and the capability of sequential injections expedites generation of EC50 values and kinetic analysis [21–24]. REFERENCES 1. Dykens JA. RedOx targets: enzyme systems and drug development strategies for mitochondrial dysfunction. In Comprehensive Medicinal Chemistry II , DV Triggle and JB Taylor, eds. Elsevier, Oxford; pp.1053–1087. 2. Schefﬂer, I. Mitochondria. New York: Wiley; 1999. 3. Saxena R, de Bakker PI, Singer K, et al. Comprehensive association testing of common mitochondrial DNA variation in metabolic disease. Am J Hum Genet. 2006;79:54–61.

REFERENCES

381

4. Wallace KB, Starkov A. Mitochondrial targets of drug toxicity. Anna Rev Pharmacol Toxicol . 2000;40:353–388. 5. Szewczyk A, Wojtczak L. Mitochondria as a pharmacological target. Pharmacol Rev. 2002;54:101–127. 6. Zhou S, Wallace KB. The effect of peroxisome proliferators on mitochondrial bioenergetics. Toxicol Sci. 1999;48:82–89. 7. Brunmair B, Lest A, Staniek K, et al. Fenoﬁbrate impairs rat mitochondrial function by inhibition of respiratory control complex I. J. Pharmacol Exp Ther. 2004;311:109–114. 8. Brumair B, Staniek K, Gras F, et al. Thiazolidinediones like metformin inhibit respiratory complex I. Diabetes. 2004;53:1052–1059. 9. McKee EE, Ferguson M, Bentley AT, Marks TA. Inhibition of mammalian mitochondrial protein synthesis by oxazolidinones. Antimicrob Agents Chemother . 2006;50:2042–2049. 10. Pinti M, Salomoni P, Coassarizza A. Anti-HIV drugs and the mitochondria. Biochim Biophys Acta. 2006;1757:700–707. 11. Hynes J, Marroquin LD, Ogurtsov VI, et al. Investigation of drug-induced mitochondrial toxicity using ﬂuorescence-based oxygen-sensitive probes. Toxicol Sci. 2006;92:186–200. 12. Kraus-Berthier L, Jan M, Guilbaud N, Naze M, Pierre A, Atassi G. Histology and sensitivity to anticancer drugs of two human non-small cell lung carcinomas implanted in the pleural cavity of nudemice. Clin Cancer Res. 2000;6:297–304. 13. Clark LC, Jr. Monitor and control of blood and tissue oxygen tension. Trans Am Soc Artif Intern Organs. 1956;2:41–48. 14. Owicki JC, Parce JW. Biosensors based on the energy metabolism of living cells: the physical chemistry and cell biology of extracellular acidiﬁcation. Biosens Bioelectron. 1992;7:255–272. 15. Parce JW, Owicki JC, Kercso KM, et al. Detection of cell-affecting agents with a silicon biosensor Science. 1989;246:243–247. 16. Eklund SE, Cliffel DE, Kozlov E, Prokop A, Wikswo JP, Jr., Baudenbacher FJ. Modiﬁcation of the Cytosensor microphysiometer to simultaneously measure extracellular acidiﬁcation and oxygen consumption rates. Anal Chim Acta. 2003;496:93–101. 17. Eklund SE, Snider RM, Wikswo J, Baudenbacher F, Prokop A, Cliffel DE. Multianalyte microphysiometry as a tool in metabolomics and systems biology. J Electroanal Chem. 2006;587:333–339. 18. Wiley C, Beeson C. Continuous measurement of glucose utilization in heart myoblasts. Anal Biochem. 2002;304:139–146. 19. Henning T, Brischwein M, Baumann W, et al. Approach to a multiparametric sensor-chip-based tumor chemosensitivity assay. Anticancer Drugs. 2001;12:21–32. 20. Mestres P, Morguet A, Schmidt W, Kob A, Thedinga E. A new method to assess drug sensitivity on breast tumor acute slices preparation. Ann N Y Acad Sci. 2006;1091:460–469. 21. Wu M, Neilson A, Swift A, Moran R, et al. Multiparameter metabolic analysis reveals a close link between attenuated mitochondrial bioenergetic function and enhanced glycolysis dependency in human tumor cells. Am J Physiol Cell Physiol. 2006;292:125–136.

382

MITOCHONDRIAL ASSESSMENT IN LIVING CELLS

22. Sridharan V, Guichard J, Bailey RM, Kasiganesan H, Beeson C, Wright GL. The prolyl hydroxylase oxygen-sensing pathway is cytoprotective and allows maintenance of mitochondrial membrane potential during metabolic inhibition. Am J Physiol Cell Physiol. 2007;292:719–728. 23. Watanabe M, Houten SM, Mataki C, et al. Bile acids induce energy expenditure by promoting intracellular thyroid hormone activation. Nature. 2006;439:484–489. 24. da-Silva WS, Harney JW, Kim BW, et al. The small polyphenolic molecule kaempferol increases cellular energy expenditure and thyroid hormone activation. Diabetes. 2007;56:767–776. 25. Marroquin LD, Hynes J, Dykens JA, Jamieson JD, Will Y. Circumventing the Crabtree effect: replacing media glucose with galactose increases susceptibility of HepG2 Cells to mitochondrial toxicants. Toxicol Sci. 2007; Epub ahead of print.

15 ASSESSMENT OF MITOCHONDRIAL RESPIRATORY COMPLEX FUNCTION IN VITRO AND IN VIVO Mark A. Birch-Machin Dermatological Sciences, Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne, UK

1. Spectrophotometric measurement of the activities of individual complexes I to IV 1.1. Complex I: NADH–ubiquinone oxidoreductase activity 1.2. Complex II: succinate–ubiquinone oxidoreductase activity 1.3. Complex III: ubiquinol–ferricytochrome c oxidoreductase 1.4. Complex IV: cytochrome c oxidase 2. Spectrophotometric measurement of the activities of linked assays 2.1. Succinate–hexacyanoferrate reductase 2.2. Succinate–cytochrome c reductase 3. Limitations of the assays 4. Drug-mediated toxicity 4.1. Photodynamic therapy 4.2. Anthralin 4.3. Imiquimod 4.4. Ultraviolet radiation 4.5. Resveratrol 4.6. MitoQ

384 384 385 385 386 387 387 388 388 389 390 390 391 392 392 393

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

383

384

ASSESSMENT OF MITOCHONDRIAL RESPIRATORY COMPLEX FUNCTION

1. SPECTROPHOTOMETRIC MEASUREMENT OF THE ACTIVITIES OF INDIVIDUAL COMPLEXES I TO IV All assays are performed at 30◦ C using an appropriate spectrophotometer, and the detailed protocol, including the preparation of mitochondrial fractions from tissue and cultured cells, is described elsewhere [1]. 1.1. Complex I: NADH–Ubiquinone Oxidoreductase Activity Brief Protocol Complex I speciﬁc activity is measured by following the decrease in absorbance due to the oxidation of NADH at 340 nm (with 425 nm as the reference wavelength if the spectrophotometer is dual wavelength). NADH, ubiquinone1 , antimycin A, and mitochondrial protein are present in the assay. Complex I activity is the rotenone-sensitive NADH–ubiquinone oxidoreductase activity. Comments Accurate measurement of complex I activity in a mitochondrial fraction depends on access of NADH to its binding site on the inner aspect of the inner mitochondrial membrane. To establish the optimal conditions for release of complex I activity, we compared several established methods of membrane disruption, such as freeze–thawing mitochondria in hypotonic or isotonic media, sonication, hypotonic shock, or addition of detergent, including the nonionic detergent dodecyl maltoside (n-dodecyl-β-d-maltoside). Maximum complex I activity was obtained by freeze–thawing the mitochondria three times in hypotonic media (25 mM potassium phosphate, 5 mM MgCl2 , pH 7.2). In addition, the freeze–thaw technique is simpler and easier to perform with small samples [1]; sonication requires extra equipment and has the added risk of microdroplet formation, always a concern with human samples. Both ubiquinone1 and ubiquinone2 can be used as electron acceptors, but the rate with ubiquinone1 is faster. Complex I activity is measured as the rotenone-sensitive NADH–ubiquinone oxidoreductase activity, so it is important to establish that maximum rotenone binding has been achieved. Addition of 2.5 mg/mL of bovine serum albumin to the assay medium increases rotenone sensitivity by at least 1.5-fold. The proportion of the total NADH–ubiquinone oxidoreductase activity that is rotenone sensitive varies considerably in different tissues [1]. This, in part, is related to the relative purity of the mitochondrial preparation, and therefore the contamination by other enzymes with NADH–ubiquinone oxidoreductase activity [2]. A potentially important enzyme contributing to the rotenone-insensitive NADH–ubiquinone oxidoreductase is NADH–cytochrome b 5 reductase. This enzyme is involved in the oxidation of NADH generated in the cytosol and is situated on the outer mitochondrial membrane and on the endoplasmic reticulum [3]. Using an antibody raised against puriﬁed NADH–cytochrome b 5 reductase, we have shown that this enzyme makes an important conribution to the rotenone-insensitive NADH–ubiquinone oxidoreductase activity [1]. This has

SPECTROPHOTOMETRIC MEASUREMENT OF COMPLEXES I TO IV

385

important implications in mitochondria prepared from tissues other than muscle, which have a high percentage of rotenone-insensitive activity, as under these circumstances it is difﬁcult to detect a partial decrease in rotenone-sensitive NADH oxidation (e.g., lymphocytes) [2]. Mitochondria isolated from ﬁbroblasts have considerable rotenone-insensitive NADH–ubiquinone oxidoreductase activity. Washing in hypotonic buffer appears to partially remove this contaminating enzyme activity and results in an increase in the percentage of activity due to complex I (the rotenone-sensitive component) by 1.4-fold. 1.2. Complex II: Succinate–Ubiquinone Oxidoreductase Activity Brief Protocol Complex II speciﬁc activity is measured by following the reduction of 2,6-dichlorophenolindophenol at 600 nm. Mitochondria are preincubated in the assay medium plus succinate at 30◦ C for 10 minutes to achieve fully activated enzyme. Antimycin A, rotenone, KCN, and dichlorophenolindophenol are added and the reaction is started with ubiquinone. Comments In isolated mitochondria a variable proportion of complex II is deactivated by tight binding of the competitive inhibitor oxaloacetate. Fully activated enzyme is achieved by preincubation with succinate (at least 1.6-fold activation). There is no increase in activity by preincubation at a higher temperature or by incubation for a longer period of time or with a higher concentration of succinate. Complex II activity is also dependent on disruption of the inner mitochondrial membrane. The optimal disruption method for mitochondrial fractions is freeze–thawed three times in hypotonic medium (25 mM potassium phosphate, 5 mM MgCl2 , pH 7.2). No further increase in activity is achieved with dodecyl maltoside or Triton X100. In contrast to complex I, the activity of complex II, when measured with dicholorophenol as electron acceptor, is the same whether ubiquinone1 or ubiquinone2 is used in the assay. 1.3. Complex III: Ubiquinol–Ferricytochrome c Oxidoreductase Brief Protocol Complex III speciﬁc activity is measured by monitoring the reduction of cytochrome c (III) at 550 nm with 580 nm as the reference wavelength (where a dual-wavelength spectrophotometer is available). Cytochrome c (III), rotenone, dodecyl-β-d-maltoside, and ubiquinol2 are added to the assay and the reaction is started by addition of mitochondria. The increase in absorbance rapidly becomes nonlinear, and activity is expressed as an apparent ﬁrst-order rate constant after reduction of the remaining cytochrome c (III) with a few grains of ascorbate. Ubiquinol is prepared as described previously [1]. Comments The inclusion of KCN in the assay media prevents reoxidation of the product, cytochrome c (II), by inhibition of complex IV of the respiratory chain (cytochrome c oxidase). The assay medium also contains rotenone,

386

ASSESSMENT OF MITOCHONDRIAL RESPIRATORY COMPLEX FUNCTION

which prevents any nonspeciﬁc changes in the ubiquinol concentration through inhibition of complex I activity. Since the respiratory chain enzymes are embedded in the inner mitochondrial membrane, it is important to ensure access of both ubiquinol and cytochrome c to the enzyme. The hypotonic freeze–thawing technique used to measure complex I and II activities is not appropriate for the measurement of complex III. In fact, this technique consistently lowers complex III values by 40% compared to non-freeze-thawed mitochondria. We and others have evaluated the effect of dodecylmaltoside on complex III activity and found that the activity of complex III increases twofold, but only with concentrations of dodecylmaltoside greater than 0.44 mM (i.e., in excess of the critical micellar concentration of 0.16 mM) [1,4]. This increase in complex III activity is independent of the mitochondrial protein concentration, providing a measurable rate. Preincubation of the mitochondria with detergent is not required, and the detergent is simply added to the assay buffer. The inclusion of bovine serum albumin (BSA) is necessary because in its absence, the reliability of the assay decreases. The mechanism of the effect of dodecylmaltoside on complex III is uncertain; the detergent may affect the solubility of the ubiquinol, alter the conformation of the enzyme, causing a monomer-to-dimer transition, or allow greater access of cytochrome c. The natural substrate for complex III is ubiquinol10 , but this compound is insoluble in aqueous solution and the use of low-molecular-weight ubiquinols such as ubiquinol1 and ubiquinol2 is a reasonable compromise. Some experiments [5] have used custom-synthesized n-decylcoenzyme Q-ol. Measured complex III activity increases with the length of the isoprenoid chain of the ubiquinol homolog, duroquinol < ubiquinol1 < ubiquinol2 . Complex III activity is inhibited 95% by antimycin A (1µg/mL), and the 5% electron leak through the Qo center can be prevented by addition of myxothiazol. Chretien et al. [4] have suggested the inclusion of a metal cation chelator such as EDTA, which reduces superoxide production through quinol autoxidation, although pentaorbital chelators may be more efﬁcient in reducing Fenton chemistry. 1.4. Complex IV: Cytochrome c Oxidase Brief Protocol Complex IV speciﬁc activity is measured by following the oxidation of cytochrome c (II) at 550 nm (with 580 nm as the reference wavelength for dual-wavelength spectrophotometers). Cytochrome c (II) and dodecylmaltoside are added to the assay and the reaction is started with mitochondrial protein. Complex IV activity is measured as the initial rate or as the apparent ﬁrst-order rate constant after fully oxidizing cytochrome c (II) by the addition of a few grains of potassium hexacyanoferrate. Cytochrome c (II) is prepared as described previously [1]. Comments Similar to complex III, disruption of the inner mitochondrial membrane by freeze–thawing in hypotonic solution also lowers complex IV activity,

SPECTROPHOTOMETRIC MEASUREMENT OF LINKED ASSAYS

387

but only by 15 to 20%. We have evaluated the effect of dodecylmaltoside on complex IV activity in human skeletal muscle mitochondria and observed about a fourfold increase when the concentration of dodecylmaltoside in the assay buffer was greater than the critical micellar concentration. Like its effect on complex III, the effect of dodecylmaltoside on complex IV activity is independent of protein concentration. Since other detergents, such as Triton X100 at concentrations of 0.07 to 3 mM, did not increase activity, it is thought that dodecylmaltoside stimulates activity by conferring active conformation of the enzyme [1,6] rather than simply allowing better access of cytocrome c.

2. SPECTROPHOTOMETRIC MEASUREMENT OF THE ACTIVITIES OF LINKED ASSAYS Electron ﬂow through complex III can be assessed by two linked assays which measure a segment of the respiratory chain linking the activity of complex II with that of complex III. In these linked assays, succinate is the reductant and the terminal electron acceptor is either hexacyanoferrate ([Fe(CN)6]3-) or cytochrome c (III). 2.1. Succinate–Hexacyanoferrate Reductase Brief Protocol Enzyme activity is measured by following the reduction of the hexacyanoferrate (III) at 420 nm with 475 nm as the reference wavelength. Intact mitochondria, succinate, and rotenone are incubated for 10 minutes at 30o C. The reaction is started by the addition of potassium hexacyanoferrate (0.5 mM), and the linear decrease in absorbance is recorded. Comments The artiﬁcial electron acceptor hexacyanoferrate (III) is able to accept electrons derived from oxidation of succinate in the mitochondrial matrix by reacting directly with the respiratory chain at the level of cytochrome c, which is located on the outer face of the mitochondrial inner membrane. Hexacyanoferrate does not cross the inner membrane and therefore is unable to react nonspeciﬁcally with reducing equivalents in the matrix. The reaction buffer is isotonic to ensure that the mitochondrial membranes remain intact. Therefore, transfer of electrons from succinate is blocked almost completely by addition of antimycin, since the succinate dehydrogenase component of complex II is located on the inner face of the inner membrane, and electrons derived from succinate can be conducted to an exogenous acceptor only via complex III. ADP is included in the reaction buffer to maintain mitochondria in state 3 respiration, although mitochondria become progressively uncoupled in the presence of hexacyanoferrate (III). Inclusion of KCN and rotenone ensures inhibition of complex IV and complex I activities, respectively. As described above, preincubation of mitochondria with succinate ensures that complex II is fully activated.

388

ASSESSMENT OF MITOCHONDRIAL RESPIRATORY COMPLEX FUNCTION

2.2. Succinate–Cytochrome c Reductase Brief Protocol Enzyme activity is measured by following the reduction of cytochrome c (III) at 550 nm, with 580 nm as the reference wavelength. Mitochondria, succinate, and rotenone are incubated for 10 minutes at 30o C. The reaction is started by the addition of cytochrome c (III), which gives a linear increase in absorbance. Comments The succinate–cytochrome c reductase (SCR) assay is used widely, and all variations are based on the method of Sottocasa et al. [3]. Some groups [7] include BSA (1 mg/mL) in the reaction, but we have not found this necessary. To allow access of the exogenously added cytochrome c to complex III, the mitochondrial fraction is disrupted by freeze–thawing in hypotonic media (25 mM potassium phosphate, pH 7.2). In contrast to its effect on complex III activity, addition of dodecyl maltoside abolishes SCR activity; furthermore, sonication of mitochondria also decreases activity [5]. It is thought that these treatments result in separation of the two respiratory chain components of the linked assay.

3. LIMITATIONS OF THE ASSAYS The assessment of mitochondrial respiratory chain enzyme activity in human samples is difﬁcult due to both the small amount of tissue generally available and the frequent need to perform enzyme activity measurement in crude mitochondrial fractions. This is particularly true for the measurement of complex III activity in which a partial deﬁciency can easily be overlooked [4,6,8,9]. The study by Chretien [4] showed that use of the detergent, dodecyl maltoside (or lauryl maltoside), allows (1) better detection of an inherited partial defect affecting cytochrome b, a catalytic subunit of the mitochondrial complex III; and (2) to possibly, discriminate decreased complex III activity resulting from abnormal complex III assembly (BCS1 gene mutation) from hampered catalytic activity originating from a cytochrome b mutation. One of the most commonly used assays in the investigation of patients with defects of the respiratory chain is the measurement of the linked assay succinate–cytochrome c reductase (SCR). However, as with all linked or multicomponent assays, it measures several components of the respiratory chain, and therefore the ability to detect a partial defect in complex III activity will depend on the amount of control exerted by that enzyme step on overall electron ﬂux on the SCR assay. We ﬁrst showed in rat muscle mitochondria that complex III exerts a low degree of control on electron ﬂux through SCR [10]. Second, in human muscle mitochondria, titration of electron ﬂow through complex III with speciﬁc inhibitors (myxothiazol and antimycin A) have shown that signiﬁcant decreases in individual complex III activity (measured as ubiquinol– cytochrome c reductase) are not reﬂected in the SCR assay [8]. For example, inhibition of complex III by 50% (with either inhibitor) gives no change in ﬂux through SCR.

DRUG-MEDIATED TOXICITY

389

This has important diagnostic implications, as it suggests that a partial decrease in complex III activity would not be detected by measuring SCR activity. This scenario was conﬁrmed in our study in patients with partial defects of complex III [8]. In addition, we have investigated the control exerted by complex II on SCR activity by titrating complex II activity with malonate and have found that this complex represents a major point of control of ﬂux through SCR, and thus it is likely that measurement of SCR activity will detect defects of complex II. 4. DRUG-MEDIATED TOXICITY Growing evidence suggests that cancer cells exhibit increased intrinsic ROS (reactive oxygen species) generation, due in part to oncogenic stimulation, increased metabolic activity, and mitochondrial malfunction. Since the mitochondrial respiratory chain is a major source of ROS generation, vulnerability of mitochondrial DNA (mtDNA) to ROS-mediated damage could serve to amplify oxidative stress in cancer cells by promoting genetic instability and development of drug resistance. Mitochondrial dysfunction also alters cellular apoptotic response to anticancer agents. In this context, it might be possible to exploit this biochemical feature and develop novel anticancer therapeutic strategies that preferentially kill cancer cells through ROS-mediated mechanisms. Disruption of mitochondrial function by drugs can result in cell death by necrosis or signal cell death by apoptosis, depending on the severity of mitochondrial failure. Drugs that damage mitochondria usually do so by inhibiting the respiration, inhibiting/uncoupling oxidative phosphorylation, inducing mitochondrial oxidative stress, or inhibiting DNA replication, transcription, or translation. It is therefore crucial early in drug development to test for mitochondrial toxicity, such as screening for mitochondrial function by the methyl thiazol tetrazolium (MTT) assay. On the other hand, mitochondrial toxicity induced by drugs could be part of the therapeutic mechanism of action: for example targeting tumor mitochondrial toxicity. Here we consider brieﬂy the various drug categories in terms of their action on mitochondria (reviewed in [11]) in the context of speciﬁc examples. 1. Drugs that Inhibit the Electron Transport Chain. There are the classical respiratory inhibitors described above. For example, rotenone, found in plants, is a potent inhibitor of complex I and is still used as an insecticide. Antimycin A, from fungi, is an antibiotic that inhibits complex III by blocking electron transport from the high-potential cytochrome b heme to ubiquinone. By contrast, the inhibitor stigmatellin binds to the ubiquinol oxidation site in the bc 1 complex and prevents ROS formation. 2. Drugs that Uncouple Mitochondrial Oxidative Phosphorylation. These are often weak acids that interact with the phospholipids in the mitochondrial membrane, making it permeable to protons. In addition, the unprotonated and protonated drug can cycle back and forth across the inner mitochondrial membrane,

390

ASSESSMENT OF MITOCHONDRIAL RESPIRATORY COMPLEX FUNCTION

causing the reentry of protons into the matrix and collapse of the membrane potential. Examples include nitrophenols such as 2,4-dinitrophenol (DNP), which was used as a diet pill in the 1930s. 3. Drugs that Cause Mitochondrial Stress. These include anticancer drugs such as doxorubicin. However, doxorubicin therapy is limited, due to the risk of cardiotoxicity at high doses, which cause mitochondrial dysfunction (see Chapter 6). 4. Drugs that Inhibit Mitochondrial DNA (mtDNA) Synthesis. Zidovudine (AZT) or stavudine (d4T) as thymine analogs will also inhibit mtDNA replication in treated patients through its action on mtDNA polymerase γ (see Chapters 9 and 21). 4.1. Photodynamic Therapy As a treatment modality for malignant and certain nonmalignant diseases, photodynamic therapy (PDT) involves a two-step protocol consisting of the selective uptake and accumulation of a photosensitizing agent in target cells, followed by irradiation with visible light. ROS produced during this process cause cellular damage leading to repair and survival, apoptotic cell death, or necrosis. PDT-induced apoptosis has received increasing attention because of the intimate connection between ROS generation, mitochondrial failure, and apoptosis, which differ from those generally elicited by radio- and chemotherapeutics, fueling hope of inducing apoptosis in cells resistant to conventional treatments. In living cells, PDT-induced ROS is accelerated by inhibitors of the respiratory chain, such as piericidin, rotenone, and myxothiazol, and inhibited by diphenyleneiodonium, an inhibitor of ﬂavin enzymes, indicating that the ﬂavin of complex I is involved in ROS production [12]. 4.2. Anthralin The antipsoriatic drug anthralin accumulates in skin mitochondria, dissipates mitochondrial membrane potential, and induces apoptosis through a pathway dependent on respiratory competent mitochondria Many therapeutic drugs are derived from compounds found in nature, as are almost all of the canonical mitochondrial inhibitors and uncouplers, and elucidating their mechanism of action has provided important insights into diverse areas of biology. Anthralin [1,8-dihydroxy-9(10H )-anthracenone, dithranol] was ﬁrst synthesized as a derivative of chrysarobin, isolated from the araroba tree. It is an established, safe, and effective topical treatment for psoriasis [13], a hyperproliferative skin disease affecting about 2% of the population in Western countries. We have shown that anthralin disrupts mitochondrial membrane potential and causes endogenous cytochrome c release, resulting in the activation of caspase-3 and characteristic morphological changes of apoptosis [14]. In contrast isogenic human rho0 positive cells, human rho0 cells, which lack mitochondrial DNA, were

391

DRUG-MEDIATED TOXICITY Drug (e.g. anthralin, mitoQ) ?Activation of Intrinsic Apoptotic markers

Complex II

Drug (Interaction the Q pool) e− e− − e e− e− Complex III Complex IV Q Pool e−

Complex I

Figure 1 pool.

Proposed site of drug (e.g., anthralin, mitoQ) interaction with the ubiquinone

resistant to anthralin-induced cell death and disruption of mitochondrial membrane potential and cytochrome c release. Anthralin also accumulated rapidly within mitochondria. Using assays that measure individual respiratory chain complexes as described in this chapter, we showed that anthralin speciﬁcally interacts with the ubiquinone pool. Interestingly, complex II activity was increased by anthralin even in the presence of the inhibitor malonate. Malonate completely inhibited succinate dehydrogenase activity in the presence of anthralin or vehicle. These observations indicate that anthralin is able to exchange electrons with endogenous ubiquinone/semiquinone in the mitochondria, which in turn will reduce the terminal exogenous electron acceptor in the assay, DCPIP. This model is supported by the observation that the reduction potential of the ubiquinone–ubiquinol redox couple (E o = +0.045 V) favors the transfer of electrons from anthralin (E o = −0.75 V), which would more fully reduce the redox status of the quinone pool, in accord with a study performed in rat liver [15]. Our data also suggest that anthralin may be modulating the redox status of the endogenous ubiquinone pool at the level of the ubisemiquinone anion (see Figure 1). The resulting ubiquinol species would then be able to reduce cytochrome c via the Fe–S protein and cytochrome c 1 . In turn, these events in the ubiquinone pool, particularly at the level of the ubisemiquinone anion, may lead to increased generation of ROS and apoptosis induction [16], thereby resulting in the preferential death of highly proliferative psoriatic keratinocytes, which demand a high energetic requirement from mitochondria. 4.3. Imiquimod Ultraviolet radiation (UVR) has an important role in the development of nonmelanoma skin cancer (NMSC). Solar radiation damages or depletes antigen-presenting Langerhan cells and causes keratinocytes to secrete cytokines that lead to dermal immunosuppression. Aberrant cells therefore cannot be

392

ASSESSMENT OF MITOCHONDRIAL RESPIRATORY COMPLEX FUNCTION

identiﬁed and removed by the immune system. The role of immunosuppression in the development of NMSC suggests a role for therapeutic agents that stimulate the immune system. The imidazoquinoline immune response modiﬁer (IRM) imiquimod has demonstrated potent antitumor and antiviral activity in vivo and in vitro. Imiquimod activates cells of the immune system via Toll-like receptor 7 (TLR-7) and stimulates innate and adaptive immune responses. Imiquimod also appears to reactivate the apoptosis pathways that are usually suppressed in transformed epithelial cells [17–19]. Two mechanisms for this activity have so far been described: increased expression of the cell surface death receptor FasR (CD95-R), and direct activation of the caspase cascade originating in the mitochondria [20,21]. The pro-apoptotic activity of imiquimod is supported by the observation of reduced expression of anti-apoptotic Bcl-2 protein and down-regulation of anti-apoptotic genes following imiquimod treatment [22–24]. 4.4. Ultraviolet Radiation The skin is exposed continuously to UVR, which is known to stimulate the intracellular production of reactive oxygen species and reactive nitrogen species (RNS; e.g., nitric oxide, peroxinitrite). This has implications in mutagenesis, carcinogenesis, and photoaging. An imbalance in the production of ROS/RNS and the antioxidant defense systems leads to oxidative and nitrosive stress. As mitochondria play an important role in ROS production, it is important to determine their role in UVR-induced oxidative stress. The production of nitric oxide (NO) and superoxide (O2 −· ) was measured directly and in real time by electrochemical detection in an immortalized keratinocyte cell line following exposure to a variety of UVR sources. UVR caused an increase in the rate of production of both NO and O2 −· [25]. Speciﬁc inhibitors of the mitochondrial electron transport chain complexes I to IV were used to identify the contribution of individual complexes in UVR-induced oxidative stress [26]. The largest increase in ﬂuorescence followed the addition of TTFA (thenoyl triﬂuoroacetate), a complex II inhibitor. Interestingly, these doses of UVR can also induce large (3 to 5 kilobase pairs) mtDNA deletions in human skin and cultured skin cells (for a review, see [27]). The major determinant of NMSC is the UVR in sunlight, which induces nuclear and mitochondrial DNA damage. To determine a reliable marker of cumulative UVR exposure in human skin, we have proposed the use of mtDNA, rather than nuclear DNA, as a more reliable biomarker of UV-induced DNA damage [27,28]. 4.5. Resveratrol Resveratrol is a nontoxic natural plant compound particularly well known for its occurrence in the skin of grapes. Studies on the effects of resveratrol on rat brain respiratory chain have shown a decrease in complex III activity by competition with coenzyme Q [29]. This property is especially interesting, as this complex is one of the sites where ROS are generated. By decreasing the activity of complex

REFERENCES

393

III, resveratrol cannot only oppose the production of ROS but can also scavenge them and thereby fulﬁll some of the requirements of an antioxidant. Diminished mitochondrial oxidative phosphorylation and aerobic capacity are associated with reduced longevity. Resveratrol is known to extend lifespan, and affect mitochondrial function and metabolic homeostasis. Treatment of mice with resveratrol signiﬁcantly increased their aerobic capacity, as evidenced by their increased aerobic endurance and oxygen consumption in muscle ﬁbers. Resveratrol’s effects were associated with an induction of genes for oxidative phosphorylation and mitochondrial biogenesis [30]. However, resveratrol also induces mitochondrial-mediated apoptosis [31]. 4.6. MitoQ Mitochondrial ROSs are thought to play an important pathogenic role in an increasing number of denerative diseases, and this has led to efforts to develop antioxidant compounds targeted to mitochondria [32] (see Chapter 26). These are a range of mitochondria-targeted probe molecules that comprise a lipophilic cation covalently attached to an active moiety. The lipophilic cation causes the accumulation of these molecules into mitochondria, driven by the mitochondrial membrane potential. One such compound, an alkyltriphenylphosphonium cation, termed mitoQ (mitoquinol, or mitoquinone, or a mixture of these redox forms), consists of compound Q bound covalently to the cation triphenylphosphonium. By virtue of its delocalized positive charge, mitoQ accumulates several hundredfold in mitochondria and has been used to modulate ROS in the mitochondrial matrix [33]. The major sites of superoxide production in the mitochondria have been somewhat controversial, but there is evidence that most derive from complexes I and III (although complex II has not been studied extensively) (see Figure 1). There is also evidence that complex I superoxide is released exclusively to the matrix side of the inner membrane, whereas complex III probably generates superoxide to both the matrix and intermembrane space [34]. A recent study of mitoQ administration in endothelial cell mitochondria has shown that superoxide production at complex I results from reverse electron transport and, to a lesser extent, by forward transport [35]. Superoxide is also generated at the Q-cycle in complex III. MitoQ acts in complex I to block ROS generated by reverse transport, but markedly enhance superoxide production derived from forward electron transport. These effects are likely to occur at one or more Q binding sites in complex I. REFERENCES 1. Birch-Machin MA, Briggs H, Saborido AA, Bindoff LA, Turnbull DM. An evaluation of the measurement of the activities of complexes I–IV in the respiratory chain of human skeletal muscle mitochondria. Biochem Med Metab Biol. 1994;51:35–42. 2. Rustin, P, Chretien D, Bourgeron T, et al. Biochemical and molecular investigations in respiratory chain deﬁciencies. Clin Chim Acta. 1994;228(1):35–51.

394

ASSESSMENT OF MITOCHONDRIAL RESPIRATORY COMPLEX FUNCTION

3. Sottocasa GL, Kuylenstierna B, Ernster L, Bergstrand A. An electron transport system associated with the outer membrane of liver mitochondria. J Cell Biol. 1967;32:415–438. 4. Chretien D, Slama A, Munnich A, Rotig A, Rustin P. Revisiting pitfalls, problems and tentative solutions for assaying mitochondrial respiratory chain complex III in human samples. Curr Med Chem. 2004;11:233–239. 5. Zheng X, Shoffner JM, Voljavec AS, Wallace DC. Evaluation of procedures for assaying oxidative phosphorylation: enzyme activities in mitochondrial myopathy muscle biopsies. Biochim Biophys Acta. 1990;1019:1–10. 6. Birch-Machin MA, Howell N, Turnbull DM. Defects at coupling site II. Methods Toxicol. 1993;2:324–336. 7. Chretien D, Gallego J, Barrientos A, et al. Biochemical parameters for the diagnosis of mitochondrial respiratory chain deﬁciency in humans, and their lack of age-related changes. Biochem J. 1998;329:249–254. 8. Taylor RW, Birch-Machin MA, Bartlett K, Turnbull DM. Succinate–cytochrome c reductase: assesssment of its value in the investigation of defects of the respiratory chain. Biochim Biophys Acta. 1993;1181:261–265. 9. Birch-Machin MA, Singh-Kler R, Turnbull DM. Study of skeletal muscle dysfunction. In Mitochondrial Dysfunction (LH Lash, DP Jones, eds). San Diego, CA: Academic Press; 1993:51–69. 10. Taylor RW, Birch-Machin MA, Bartlett K, Lowerson SA, Turnbull DM. The control of mitochondrial oxidations by complex III in rat muscle and liver mitochondria: implications for our understanding of mitochondrial cytopathies in man. J Biol Chem. 1994;269(5):3523–3528. 11. Chan K, Truong D, Shangari N, O’Brien PJ. Drug-induced mitochondrial toxicity. Expert Opin Drug Metab Toxicol. 2005;1(4):655–669. 12. Chernyak BV, Izyumov DS, Lyamzaev KG, et al. Production of reactive oxygen species in mitochondria of HeLa cells under oxidative stress. Biochim Biophys Acta. 2006;1757(5–6):525–534. 13. Marsden JR, Coburn PR, Marks J, Shuster S. Measurement of the response of psoriasis to short-term application of anthralin. Br J Dermatol. 1983;109:209–218. 14. McGill A, Frank A, Emmett N, et al. The antipsoriatic drug anthralin accumulates in keratinocyte mitochondria, dissipates mitochondrial membrane potential, and induces apoptosis through a pathway dependent on respiratory competent mitochondria. FASEB J. 2005;19:1012–1014. 15. Fuchs J, Nitschmann WH, Packer L. The antipsoriatic compound anthralin inﬂuences bioenergetic parameters and redox properties of energy transducing membranes. J Invest Dermatol. 1990;94:71–76. 16. Pelicano H, Feng L, Zhou Y, Hileman EO, Keating MJ, Huang P. Inhibition of mitochondrial respiration: a novel strategy to enhance drug-induced apoptosis in human leukemia cells by a reactive oxygen species–mediated mechanism. J Biol Chem. 2003;278:37832–37839. 17. Meyer T, Nindl I, Schmook T, et al. Induction of apoptosis by Toll-like receptor-7 agonist in tissue cultures. Br J Dermatol. 2003;149(suppl 66):9–14. 18. Schon M, Bong AB, Drewniok C, et al. Tumor-selective induction of apoptosis and the small-molecule immune response modiﬁer imiquimod. J Natl Cancer Inst. 2003;95(15):1138–1149.

REFERENCES

395

19. Dummer R, Urosevic M, Kempf W, et al. Imiquimod in basal cell carcinoma: how does it work? Br J Dermatol. 2003;149(suppl 66):57–58. 20. Schon MP, Wienrich BG, Drewniok C, et al. Death receptor–independent apoptosis in malignant melanoma induced by the small-molecule immune response modiﬁer imiquimod. J Invest Dermatol. 2004;122(5):1266–1276. 21. Berman B, Sullivan T, De Araujo T, et al. Expression of Fas-receptor on basal cell carcinomas after treatment with imiquimod 5% cream or vehicle. Br J Dermatol. 2003;149(suppl 66:59–61. 22. Vidal D, Matias-Guiu X, Alomar A. Efﬁcacy of imiquimod for the expression of Bcl-2, Ki67, p53 and basal cell carcinoma apoptosis. Br J Dermatol. 2004;151(3):656–662. 23. Urosevic M, Maier T, Benninghoff B, et al. Mechanisms underlying imiquimodinduced regression of basal cell carcinoma in vivo. Arch Dermatol. 2003;139(10): 1325–1332. 24. Lysa B, Tartler U, Wolf R, et al. Gene expression in actinic keratoses: pharmacological modulation by imiquimod. Br J Dermatol. 2004;151(6):1150–1159. 25. Aitken G, Henderson JR, Chang S-C, McNeil C, Birch-Machin MA. Direct monitoring of UV-induced free radical generation in HaCaT keratinocytes. Clin Exp Dermatol . 2007;32:722–727. 26. Aitken G, Henderson JR, Chang S-C, McNeil C, Birch-Machin MA. Direct real-time electrochemical measurement and ﬂow cytometry of UVA induced oxidative stress and its perturbation by mitochondrial electron transport chain inhibitors. Br J Dermatol. 2005;152:830. 27. Birch-Machin MA. The role of mitochondria in ageing and carcinogenesis. Clin Exp Dermatol. 2006 Jun;31(4):548–552. 28. Durham SE, Krishnan KJ, Betts J, Birch-Machin MA. Mitochondrial DNA damage in non-melanoma skin cancer. Br J Cancer. 2003;88:90–95. 29. Zini R, Morin C, Bertelli A, Bertelli AA, Tilement JP. Effects of resveratrol on the rat brain respiratory chain. Drugs Exp Clin Res. 1999;25(2–3):87–97. 30. Lagouge M, Argmann C, Gerhart-Hines Z, et al. Resveratrol improves mitochondrial function and protects against metabolic disease by activating SIRT1 and PGC-1alpha. Cell. 2006;127(6):1109–1122. 31. Sareen D, van Ginkel PR, Takach JC, et al. Mitochondria as the primary target of resveratrol-induced apoptosis in human retinoblastoma cells. Invest Ophthalmol Vis Sci. 2006;47(9):3708–3716. 32. Murphy MP. Investigating mitochondrial radical production using targeted probes. Biochem Soc Trans. 2004;32:1011–1014. 33. Echtay KS, Murphy MP, Smith RA, Talbot DA, Brand MD. Superoxide activates mitochondrial uncoupling protein 2 from the matrix side. J Biol Chem. 2002;277:47129–47135. 34. St-Pierre J, Buckingham JA, Roebuck SJ, Brand MD. Topology of superoxide production from different sites in the mitochondrial electron transport chain. J Biol Chem 2002;277:44784–44790. 35. O’Malley Y, Fink BD, Ross NC, Prisinzano TE, Sivitz WI. Reactive oxygen and targeted antioxidant administration in endothelial cell mitochondria. J Biol Chem. 2006;281(52):39766–39775.

16 OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS FOR ASSESSMENT OF OXPHOS PROTEIN LEVELS Sashi Nadanaciva MitoSciences Inc., Eugene, Oregon

1. Introduction 2. Immunocapture-based OXPHOS activity assays 2.1. Immunocapture-based complex I activity (NADH–ubiquinone oxidoreductase) assay 2.2. Immunocapture-based complex II activity (succinate–ubiquinone oxidoreductase) assay 2.3. Immunocapture-based complex IV activity (cytochrome c oxidase) assay 2.4. Immunocapture-based complex V activity (F1 FO -ATPase) assay 2.5. Data Analysis 3. Speciﬁcity of the immunocapture-based OXPHOS activity assays: effect of classical inhibitors 4. Screening of therapeutic drugs using the immunocapture-based OXPHOS activity assays 5. Lateral ﬂow dipstick immunoassays 6. Future perspectives

398 398 399 399 400 400 400 401 401 404 409

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

397

398

OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS

1. INTRODUCTION Mitochondria make approximately 95% of a cell’s ATP via oxidative phosphorylation (OXPHOS), a process carried out by ﬁve protein complexes. Mitochondria also play a crucial role in apoptosis (programmed cell death), fatty acid oxidation, and heme and steroid synthesis. These organelles contain at least 600 different proteins, 13 of which are encoded by mitochondrial DNA (mtDNA) and made using the mitochondrion’s own replication machinery. Mitochondria are often the target of drug-induced toxicity. This can occur through several mechanisms, such as inhibition of the oxidative phosphorylation complexes, uncoupling of electron transport from ATP synthesis, redox cycling of the xenobiotic, inhibition of the Krebs cycle, inhibition of β-fatty acid oxidation, irreversible opening of the mitochondrial permeability transition pore, inhibition of transporters within the mitochondrial membranes, and impairment of either mtDNA synthesis or mtDNA-encoded protein synthesis [1–3] (see Chapters 2, 9, 20, and 21). Although there are assays for identifying drug-induced inhibition of oxidative phosphorylation, such as determination of ATP production by the luciferin–luciferase system [4] and oxygen consumption by polarography [5] or ﬂuorescence (see Chapters 12 and 13), there is a need for high-throughput assays that identify the target of inhibition of oxidative phosphorylation. In this chapter we describe very sensitive, highly selective, high-throughput immunocapture-based assays of the activities of the oxidative phosphorylation complexes that readily identify drug-induced inhibition. We also describe lateral ﬂow dipstick immunoassays for identifying drug-induced impairment of mtDNA-encoded protein expression. 2. IMMUNOCAPTURE-BASED OXPHOS ACTIVITY ASSAYS In each of these assays, 96-well plates coated with a monoclonal antibody raised against complex I, II, IV, or V are used to immunocapture a functionally active complex from small amounts of biological material [6]. Stock solutions of the compounds to be tested are prepared in dimethyl sulfoxide (DMSO) added to multichannel Dilux dilution reservoirs (ISC BioExpress, Kaysville, UT) containing the appropriate assay solution and then dispensed into each 96-well plate in quadruplicate wells. Measurements for a compound at a given concentration are made in triplicate wells coated with the appropriate immunocapture antibody and a single well containing a null capture antibody used as a negative control. The ﬁnal DMSO concentration in all assays is 1.5% v/v. Each assay is read in a plate reader (such as SpectraMax Plus) immediately after addition of the assay solution (containing the compounds) to the 96-well plates. The immunocapture assays can be used to screen drugs rapidly for their effect on the OXPHOS complexes in cell and tissue extracts without the need to isolate pure mitochondria. The tissue or cell sample to be analyzed should ﬁrst be homogenized with a Dounce or Ultra-Turrax homogenizer in isotonic medium (250 mM sucrose,

IMMUNOCAPTURE-BASED OXPHOS ACTIVITY ASSAYS

399

10 mM Tris-HCl, 0.2 mM EDTA, pH 7.8) with a protease inhibitor cocktail, and centrifuged at 1000g (∼10 minutes at 4◦ C) to sediment the debris. The protein concentration of the supernatant is determined and the sample is solubilized at 5.5 mg/mL with a 1/10 volume of 10% w/v n-dodecyl-β-d-maltopyranoside (Anatrace) such that the ﬁnal detergent concentration is 1%. The sample is left on ice for 20 minutes and then centrifuged at 25,000g for 20 minutes at 4◦ C. The supernatant is then diluted to the appropriate concentration for each assay. 2.1. Immunocapture-Based Complex I Activity (NADH–Ubiquinone Oxidoreductase) Assay Protein G plates (Pierce) are coated with an anti-complex I mAb (MitoSciences Inc., Eugene, Oregon) at 40 µg/mL in PBS, incubated overnight at 4◦ C, washed three times with 50 mM Tris-HCl, pH 7.5, incubated for 2 hours at room temperature with the sample of interest (e.g., 15 µg/well solubilized bovine heart mitochondria), washed three times with 20 mM Tris-HCl, 50 mM KCl, 0.015% w/v n-dodecyl-β-d-maltopyranoside pH 7.5, and incubated for 45 minutes at 4◦ C with 56 µg/mL phosphatidylcholine that has been dissolved in 20 mM Tris-HCl, 50 mM KCl, 0.015% w/v n-dodecyl-β-d maltopyranoside pH 7.5 [6]. The assay can also be done using Nunc maxisorp plates instead of Protein G plates, followed by a 2-hour incubation with a blocking solution such as 1X Sigma Block, prior to adding the detergent-solubilized sample (Autumn Bernal, personal communication). Complex I activity is measured by adding an assay solution containing 25 mM KH2 PO4 pH 7.2, 5 mM MgCl2 , 1.2% w/v bovine serum albumin (BSA), 0.15 mM coenzyme Q1 , and 0.26 mM NADH. The oxidation of NADH is monitored by measuring its decrease in absorbance at 340 nm in a plate reader in kinetic mode at 30◦ C. 2.2. Immunocapture-Based Complex II Activity (Succinate–Ubiquinone Oxidoreductase) Assay Nunc maxisorp plates are coated with an anti-complex II mAb (MitoSciences Inc.) at 5 µg/mL in 50 mM KH2 PO4 , pH 7.2, incubated overnight at 4◦ C, aspirated, blocked with 5% nonfat dry milk (dissolved in 50 mM KH2 PO4 , pH 7.2) for 2 hours at room temperature, washed with 50 mM KH2 PO4 , pH 7.2, incubated with the sample of interest (e.g., 25 µg/well of solubilized bovine heart mitochondria in 20 mM KH2 PO4 , pH 7.2, 0.015% w/v n-dodecyl-β-d-maltopyranoside) for 2 hours at room temperature, and washed twice with 20 mM KH2 PO4 , pH 7.2, 0.015% w/v n-dodecyl-β-d-maltopyranoside [6]. Complex II activity is measured by adding an assay solution containing 25 mM KH2 PO4 , pH 7.2, 20 mM sodium succinate, 65 µM coenzyme Q2 , 50 µM dichlorophenolindophenol, and 0.115% w/v n-dodecyl-β-d-maltopyranoside. The decrease in absorbance of dichlorophenolindophenol is measured at 600 nm in a plate reader in kinetic mode at room temperature while the rate is linear.

400

OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS

2.3. Immunocapture-Based Complex IV Activity (Cytochrome c Oxidase) Assay Nunc maxisorp plates are coated with an anti-complex IV mAb (MitoSciences Inc.) at 5 µg/mL in 50 mM KH2 PO4 , pH 7.2, incubated overnight at 4◦ C, aspirated, blocked with 5% nonfat dry milk (dissolved in 50 mM KH2 PO4 , pH 7.2) for 2 hours at room temperature, washed with 50 mM KH2 PO4 , pH 7.2, incubated for 3 hours with the sample of interest (e.g., 0.1 µg/well solubilized bovine heart mitochondria in 20 mM KH2 PO4 , pH 7.2, 0.015% w/v n-dodecyl-β-d-maltopyranoside) at room temperature, and washed twice with 20 mM KH2 PO4 , pH 7.2, 0.015% w/v n-dodecyl-β-d-maltopyranoside [6,7]. Complex IV activity is measured by adding an assay solution containing 60 µM reduced cytochrome c, 20 mM KH2 PO4 , pH 7.2, and 0.015% w/v n-dodecyl-β-d-maltopyranoside. The oxidation of reduced cytochrome c is monitored by measuring its decrease in absorbance at 550 nm in a plate reader in kinetic mode either at room temperature or at 30◦ C while the rate is linear.

2.4. Immunocapture-Based Complex V Activity (F1 FO -ATPase) Assay Nunc maxisorp plates are coated with goat anti-mouse IgG-Fc (Fcγ for all subclasses) at 5 µg/mL in TBS (50 mM Tris-HCl, 0.5 M NaCl, pH 7.5), incubated overnight at 4◦ C, washed three times with TBS, coated with an anti-complex V mAb (MitoSciences Inc.) at 5 µg/mL in TBS containing 2.5% w/v BSA, incubated overnight at 4◦ C, washed three times in TBS, and incubated for 2 hours at room temperature with the sample of interest (e.g., 20 µg/well solubilized bovine heart mitochondria in TBS containing 2.5% BSA), and then washed three times with TBS [6,8]. Complex V activity is measured by addition of an ADP-coupled assay solution containing 25 mM HEPES, 25 mM KCl, 2 mM MgCl2 , 2 mM phosphoenolpyruvate, 2 mM ATP, 0.5 mM NADH, 30 units/mL pyruvate kinase and 30 units/mL l-lactic dehydrogenase, pH 7.5. The decrease in absorbance of NADH is measured at 340 nm in a plate reader in kinetic mode at 30◦ C while the rate is linear.

2.5. Data Analysis Absorbance values obtained during all activity assays on the plate reader are imported into a graph-plotting software program such as SigmaPlot or GraphPad Prism. One hundred percent activity (i.e., no inhibition) for each complex is determined as the mean of the triplicate measurement in the absence of compound minus the negative control value in the absence of compound. IC50 values are generated using a four-parameter logistic equation. Table 1 summarizes the intraand interassay variation of the immunocapture-based OXPHOS activity assays performed with bovine heart mitochondria.

401

SCREENING OF THERAPEUTIC DRUGS

TABLE 1 Intra- and Interassay Coefﬁcients of Variation of the Immunocapture OXPHOS Activity Assaysa Complex I Activity Day

Mean ± SD CV n (mOD/min) (%)

1 3 2 3 3 3 4 3 Inter assay, days 1 to 4 a Data

Complex II Activity Mean ± SD (mOD/min)

3.46 ± 0.28 8.1 20.31 ± 2.40 3.94 ± 0.41 10.4 19.97 ± 2.36 3.94 ± 0.30 7.6 21.52 ± 2.73 3.43 ± 0.28 8.2 18.46 ± 1.98 3.69 ± 0.32 8.7 20.06 ± 2.38

Complex IV Activity

CV Mean ± SD CV (%) (mOD/min) (%) 11.8 11.8 12.7 10.7 11.9

Complex V Activity Mean ± SD CV (mOD/min) (%)

4.57 ± 0.51 11.2 13.32 ± 0.67 5.0 4.91 ± 0.36 7.3 13.50 ± 1.16 8.6 4.71 ± 0.49 10.4 12.83 ± 1.08 8.4 4.35 ± 0.42 9.7 12.26 ± 1.17 9.5 4.63 ± 0.45 9.7 12.97 ± 1.04 8.0

are described as mean ± SD (standard deviation) and coefﬁcient of variation (CV).

3. SPECIFICITY OF THE IMMUNOCAPTURE-BASED OXPHOS ACTIVITY ASSAYS: EFFECT OF CLASSICAL INHIBITORS We have evaluated the immunocapture-based OXPHOS activity assays with classical inhibitors of OXPHOS complexes to test their speciﬁcity and to provide a baseline toxicity against which other compounds can be compared [6]. These assays were performed using bovine heart mitochondria. Rotenone, a classical inhibitor of complex I, blocks electron transfer from Fe–S centers to the ubiquinone-binding site within this complex [9]. Figure 1A shows that rotenone is a potent inhibitor of the activity of complex I immunocaptured from bovine heart mitochondria. Thus, complex I activity is 85% sensitive to rotenone, with an IC50 of 17.3 nM. 2-Thenoyltriﬂuoroacetone (TTFA) inhibits complex II by binding to this complex’s ubiquinone-binding sites [10]. Inhibition of immunocaptured complex II by this compound occurred with an IC50 of 30 µM (Figure 1B). KCN, a noncompetitive inhibitor of complex IV [11], binds to the heme a 3 site within this complex and blocks electron transfer to oxygen. Figure 1C demonstrates that complex IV immunocaptured from bovine heart mitochondria is inhibited by KCN with an IC50 of 3.2 µM. Oligomycin inhibits complex V when the catalytic F1 sector of this complex is bound to the proton-conducting Fo sector [12]. Ninety percent of complex V activity immunocaptured from bovine heart mitochondria is inhibited by oligomycin with an IC50 of 8 nM (Figure 1D). 4. SCREENING OF THERAPEUTIC DRUGS USING THE IMMUNOCAPTURE-BASED OXPHOS ACTIVITY ASSAYS The immunocapture-based assays described here can be used to rapidly screen new chemical entities for direct inhibition of the OXPHOS complexes. Since

402

OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS

Figure 1 Effect of rotenone on complex I activity (A), TTFA on complex II activity (B), KCN on complex IV activity (C), and oligomycin on complex V activity (D). 100% represents the activity of uninhibited enzyme. Each data point is the mean of triplicate measurements. The error bars represent calculated standard deviations. The line is a best ﬁt using a four-parameter logistic equation.

currently no high-afﬁnity antibody is available that immunocaptures complex III in a functionally active form, we have used an assay for measuring complex II + III activity, with whole mitochondria, based on a method of Kramer et al. [13]. The immunocapture-based OXPHOS activity assays and the conventional complex II + III assay were used to test a selection of drugs [6], many with previously reported mitochondrial toxicity. Each drug was tested initially at 50 µM concentration in triplicates (Figure 2). Complex I activity was strongly inhibited (>70%) by nefazodone (Figure 2A). Approximately 25% inhibition was observed with simvastatin and paroxetine. Slight (<25%) but statistically signiﬁcant inhibition was observed with amiodarone (Figure 2A). Slight inhibition was also observed with chlorpromazine, metformin, rosiglitazone, and tamoxifen, but the differences in comparison with the control were not statistically signiﬁcant. At the concentration tested, no inhibition was observed for all other compounds. None of the compounds showed any inhibition of complex II activity (data not shown). Simvastatin and tamoxifen inhibited complex II + III strongly, and weaker inhibition was observed with amiodarone, chlorpromazine, and paroxetine (Figure 2B). A small amount of inhibition (15%) was also observed with

SCREENING OF THERAPEUTIC DRUGS

403

Figure 2 Effect of a selection of drugs (50 µM) on complex I activity (A), complex II + III activity (B), complex IV activity (C), and complex V activity (D). 100% represents the activity of uninhibited enzyme. Data are expressed as mean ± SD (n = 3). For establishment of signiﬁcance, one-way ANOVA was performed followed by Tukey’s post hoc test. Statistically signiﬁcant values compared with the control are reported as follows: *, p < 0.05; **, p < 0.01; ***, p < 0.001.

nefazodone. All other compounds tested showed no signiﬁcant inhibition in the complex II + III assay (Figure 2B). Complex IV activity was inhibited the most (80%) by tamoxifen (Figure 2C). With simvastatin and nefazodone, 25 to 50% inhibition was found. A slight (∼25%) but statistically signiﬁcant inhibition was observed with amiodarone. Complex V activity was strongly (>75%) inhibited with simvastatin, paroxetine, and tamoxifen (Figure 2D). Seventy percent inhibition was observed with chlorpromazine and approximately 50% inhibition with amiodarone. Less than 50% inhibition was observed with diclofenac and nefazodone. All other compounds had no statistically signiﬁcant effect on complex V activity. Some of the drugs found to impair OXPHOS complexes in the immunocapture-based OXPHOS activity assays had previously been reported to target mitochondria. Amiodarone has been reported to have multiple effects on mitochondria, including inhibition of OXPHOS complexes, uncoupling of

404

OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS

OXPHOS, and inhibition of fatty acid oxidation [14–16]. Chlorpromazine has been shown to inhibit multiple OXPHOS complexes [17], and Bullough et al. [18] reported 50% inhibition of F1 -ATPase activity from bovine mitochondria with 50 µM chlorpromazine. Diclofenac has been shown to inhibit complex V as well as act as an uncoupler [19]. Tuquet et al. [20] reported strong inhibition of complex III and slightly weaker inhibition of complex IV by tamoxifen in rat liver mitochondria, and data by Marroquin et al. [21] indicate that tamoxifen also strongly inhibits complex V. The immunocapture-based assays described here have advantages over previous methods. Complex V is not easily measured in concert with the electron transfer complexes in intact mitochondria. When physically isolated in an immunocapture assay, complex V activity is not confounded by other processes using ATP/ADP. Similarly, a wide variety of NADH/NAD-utilizing enzymes are present in mitochondria, but isolating complex I by an immunocapture technique removes confounding activities. Moreover, the immunocapture assays do not require prior isolation of mitochondria since the enzymes can be immunocaptured directly from cell culture material or tissue extract.

5. LATERAL FLOW DIPSTICK IMMUNOASSAYS Drug-induced mitochondrial depletion and dysfunction can occur as a result of impairment of mtDNA replication (see Chapter 2). Nucleoside reverse transcriptase inhibitors (NRTIs) inhibit HIV reverse transcriptase, but some, such as dideoxycytidine (Zalcitabine) and stavudine, also inhibit the enzyme responsible for replicating mtDNA, DNA polymerase γ. This leads to depletion of mtDNA and the 13 proteins encoded by it, all of which are subunits of the OXPHOS complexes. Mitochondrial dysfunction can also occur as a result of direct inhibition of mtDNA-encoded protein synthesis. Some antibiotics, such as chloramphenicol and linezolid, which inhibit bacterial protein synthesis, also inhibit mtDNA-encoded protein synthesis, due to the similarity between the bacterial protein replication machinery and the mitochondrial protein replication machinery [22,23] (see Chapters 1 and 20). There have been recent reports of lactic acidosis, a hallmark of mitochondrial dysfunction, in patients on a long-term schedule (>1 month) of linezolid, as well as decreased activities of complexes I and IV in tissue biopsies and peripheral blood mononuclear cells of these patients [24,25] (see Chapter 21). Inhibition of mtDNA-encoded protein synthesis has previously been analyzed by in vitro assays based on incorporation of [35 S]methionine, [3 H]leucine, or other radioactive amino acids [22,23,26]. Lateral ﬂow dipstick immunoassays, which measure the level of a mtDNA-encoded protein as well as a nuclear DNA-encoded protein, enable simple and rapid determination of the ratio between these two proteins and hence identify impaired mtDNA-encoded protein synthesis. These dipsticks can be used to analyze cell extracts, tissue extracts, and patient samples such as blood. We have analyzed mitochondrial depletion caused by antibiotics

405

LATERAL FLOW DIPSTICK IMMUNOASSAYS

and NRTIs in cells, using dipsticks that compare the levels of complex IV (cytochrome c oxidase), a mitochondrial protein that has three subunits encoded by mitochondrial DNA (mtDNA) and made by mitochondrial ribosomes, with that of frataxin, a mitochondrial protein encoded by nuclear DNA and made by cytosolic ribosomes. The ratio of complex IV to frataxin decreases when a drug inhibits mtDNA replication or mtDNA-encoded protein expression. A schematic representation of the complex IV + frataxin PQuant (protein quantity) dipstick (MitoSciences Inc.) is shown in Figure 3A. Each dipstick consists of a nitrocellulose membrane laminated at the lower portion of an adhesive support card and a cellulose wicking pad laminated at the upper portion of the card. Goat anti-mouse IgG-Fc antibody, an anti-frataxin capture mAb, and an anti-complex IV capture mAb are applied in a narrow upper zone, middle zone, and lower zone, respectively, on the nitrocellulose membrane. The sample to be analyzed is solubilized with detergent and applied to the dipstick with a gold-conjugated anti-frataxin detector mAb and a gold-conjugated anti-complex IV detector mAb. Frataxin and complex IV in the sample are captured by their respective antibodies on the dipstick and visualized by the two gold-conjugated detector mAbs. To avoid including partial assemblies of the oligomeric complex IV in the analysis, the anti-complex IV capture mAb and anti-complex IV detector mAb are selected such that they form a sandwich only with fully assembled complex B 0.7 0.6 Absorbance

A

Wicking pad

0.5 0.4 0.3

Complex IV Frataxin

0.2 0.1 0.0

0 1 2 3 4 5 6 7 8 9 10 Solubilized protein from HepG2 cells (µg)

Y YY

Anti-frataxin capture mAb

YY Y

Goat anti-mouse IgG-Fc Ab

Gold-conjugated anti-frataxin detector mAb

Frataxin

Anti-complex IV capture mAb Nitrocellulose membrane

Gold-conjugated anti-complex IV detector mAb

Complex IV Sample, blocking buffer, and gold-conjugated detector mAbs

Figure 3 (A) Schematic representation of the complex IV + frataxin PQuant (protein quantity) dipstick; (B) working range of the complex IV + frataxin PQuant dipstick with HepG2 cell extracts.

406

OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS

IV. Hence cells such as rho0 cells, which lack mtDNA-encoded proteins, do not produce a complex IV signal with the complex IV + frataxin PQuant dipsticks. The uppermost zone of goat anti-mouse IgG-Fc antibody on the dipstick veriﬁes that the entire sample has wicked up through the nitrocellulose membrane. Figure 3B shows that the complex IV and frataxin protein levels in a HepG2 cell extract can be measured using as little as 1 to 2 µg total protein. In contrast, Western blot analysis of cell extracts requires at least 20 µg of total protein. Moreover, in comparison with Western blot analysis, measurements via dipsticks are more rapid. Once the samples have been solubilized in detergent, the result is visualized within 20 minutes and quantiﬁable measurements are obtained within 90 minutes. The interassay variation with the dipsticks is <10%. To identify mitochondrial depletion in cells exposed to drugs, we have grown HepG2 cells in the presence of various drugs for at least three cell population doublings. The cells are then detached from the surface by trypsinization, centrifuged, washed once with phosphate-buffered saline, solubilized in 10 volumes of solubilization buffer (25 mM HEPES, 100 mM NaCl, 1.5% w/v n-dodecyl-β-d-maltopyranoside) with 1% v/v protease inhibitor cocktail, pH 7.4, left on ice for 20 minutes, and centrifuged in a microfuge centrifuge at 25,000g for 20 minutes at 4◦ C. The protein concentration of the supernatant is determined by the bicinchoninic acid method (Pierce) and the sample diluted in solubilization buffer to a ﬁnal concentration of 0.08 mg/mL. Twenty-ﬁve microliters (2 µg of total protein) of each sample is mixed with 25 µL of 2X blocking buffer (Sigma), 5 µL of gold-conjugated anti-frataxin mAb, and 5 µL of gold-conjugated anti-complex IV mAb, and added to a well of a 96-well plate (Costar). A complex IV + frataxin PQuant dipstick is dipped into this mixture for 20 minutes so that the entire sample is able to wick up onto the dipstick. Forty microliters of wash buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.4) is then allowed to wick up onto each dipstick. Twenty minutes later, the dipsticks are air-dried and the signal intensities of the levels of complex IV and frataxin are measured with an immunochromatographic dipstick reader (Hamamatsu ICA-1000). Figure 4A shows the complex IV/frataxin ratio in HepG2 cells exposed to 40 µM chloramphenicol for up to ﬁve cell population doublings. The 100% value in the bar graph represents the complex IV/frataxin ratio observed in vehicle-treated (0.1% DMSO) cells. The results show that complex IV is decreased in chloramphenicol-treated cells, revealing chloramphenicol impairment of mtDNA-encoded protein synthesis. A Western blot (Figure 4B) of these cells conﬁrms the reduced level of the mtDNA-encoded subunit 2 of complex IV and also shows a reduced level of the mtDNA-encoded 20-kDa subunit of complex I. In contrast to the mtDNA-encoded proteins, the levels of nuclear DNA–encoded mitochondrial proteins (complex V α subunit, porin, and complex II 30-kDa subunit) are not affected by chloramphenicol (Figure 4B). The effect of the antibiotic linezolid on mtDNA-encoded protein synthesis in HepG2 cells is shown in Figure 5. Dipstick analysis shows that 40 µM linezolid impairs mtDNA-encoded protein synthesis, as seen by the decreased complex IV/frataxin ratio (Figure 5A). A Western blot of the cells grown in this antibiotic

407

LATERAL FLOW DIPSTICK IMMUNOASSAYS

Complex IV / frataxin (%)

A 110 100 90 80 70 60 50 40 30 20 10 0 0 1 2 5 Cell population doublings after addition of 40 µM chloramphenicol

B

1

2

3

4 Complex V α subunit Porin

Complex II-30 kDa subunit Complex IV-subunit 2 Complex I-20 kDa subunit

Figure 4 Effect of 40 µM chloramphenicol on HepG2 cells. HepG2 cells were treated with the vehicle (0.1% DMSO) or 40 µM chloramphenicol for ﬁve cell population doublings. (A) 2 µg protein from detergent-solubilized cells was analyzed with complex IV + frataxin PQuant dipsticks. 100% represents the complex IV/frataxin ratio in vehicle-treated cells. Data are expressed as mean ± SD (n = 2 for cell cultures; n = 2 for dipsticks). (B) 20 µg protein from detergent-solubilized cells was loaded on a 10 to 20% acrylamide gel, subjected to SDS polyacrylamide gel electrophoresis, transferred to a polyvinylidine diﬂuoride membrane, and probed with a cocktail of mAbs (anti-complex V α subunit mAb, anti-Porin mAb, anti-complex II 30-kDa subunit mAb, anti-complex IV subunit 2 mAb, anti-complex I 20-kDa subunit mAb). Lane 1: cells grown in the vehicle, 0.1% DMSO; lanes 2, 3, 4: cells grown in 40 µM chloramphenicol for one, three, and ﬁve cell population doublings, respectively.

conﬁrms the dipstick analysis, showing a decreased level of both complex IV and complex I (Figure 5B) compared to vehicle-treated cells. Immunocytochemical analysis of cells stained with an anti-complex IV mAb and an anti-porin mAb also demonstrates reduced levels of complex IV in linezolid-treated cells compared to vehicle-treated cells, whereas the level of porin is similar in both cell groups (Figure 5C).

408

OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS

Complex IV/ frataxin (%)

A

110 100 90 80 70 60 50 40 30 20 10 0

0 1 3 5 Cell population doublings after addition of 40 µM linezolid B

1

2

3

4 Complex V α subunit Porin

Complex II-30 kDa subunit Complex IV-subunit Complex I-20 kDa subunit

Figure 5 Effect of 40 µM linezolid on HepG2 cells. HepG2 cells were treated with the vehicle (0.1% DMSO) or 40 µM linezolid for ﬁve cell population doublings. (A) 2 µg protein from detergent-solubilized cells was analyzed with complex IV + frataxin PQuant dipsticks. 100% represents the complex IV/frataxin ratio in vehicle-treated cells. Data are expressed as mean ± SD (cell cultures were grown in duplicates and each culture was analyzed with duplicate dipsticks). (B) 20 µg protein from detergent-solubilized cells was loaded on a 10 to 20% acrylamide gel, subjected to SDS polyacrylamide gel electrophoresis, transferred to a polyvinylidine diﬂuoride membrane, and probed with a cocktail of mAbs (anti-complex V α subunit mAb, anti-porin mAb, anti-complex II 30-kDa subunit mAb, anti-complex IV subunit 2 mAb, anti-complex I 20-kDa subunit mAb). Lanes 1, 2, 3: cells grown in 40 µM linezolid for one, three, and ﬁve cell population doublings, respectively. Lane 4, cells grown in the vehicle (0.1% DMSO). (C) Immunocytochemistry analysis of HepG2 cells treated with either 40 µM linezolid for ﬁve cell population doublings (images 1 to 3) or the vehicle, 0.1% DMSO (images 4 to 6). Cells were ﬁxed with 4% paraformaldehyde, permeabilized with 0.1% Triton X100 in PBS, and stained with an anti-complex IV mAb and Alexa Fluor 594 goat anti-mouse IgG2a antibody (images 1 and 4), an anti-porin mAb and Alexa Fluor 488 goat anti-mouse IgG2b antibody (images 2 and 5), and the nucleic acid stain DAPI (4 ,6-diamidino-2-phenylindole). The vehicle-treated cells shown in image 6 (the merged image of images 4 and 5) appear yellow and show that complex IV (image 4, red) and porin (image 5, green) co-localize. In contrast, the linezolid-treated cells shown in image 3 (the merged image of 1 and 2) appear green and emphasize the reduced level of complex IV (image 1) and normal level of porin (image 2). (See insert for color representation of ﬁgure.)

409

FUTURE PERSPECTIVES C

1

2

3

4

5

6

Figure 5 (Continued )

6. FUTURE PERSPECTIVES The results described in this chapter were obtained from in vitro- and cell-based assays. How does one extrapolate these results to in vivo situations? Some of the drugs described in this chapter, such as pravastatin, have no effect on the OXPHOS complexes in vitro, which correlates well with their safety proﬁle. In contrast, some drugs, such as simvastatin, show considerable toxicity in the in vitro OXPHOS assays, although there are few reports of their adverse effects in the clinic. One probable explanation is that mitotoxic concentrations are considerably higher than the plasma concentration, C max (∼20 nM for simvastatin), and that the drug is bioaccumulated in susceptible muscle cells by monocarboxylate transporter isoform 4 (see Chapters 7 and 18). Second, the phenomenon of mitochondrial energy thresholds in different tissues has to be borne in mind when extrapolating in vitro OXPHOS inhibition results to in vivo scenarios. OXPHOS complexes have to be inhibited above a certain percentage before bioenergetic capacity is severely impaired in vivo, and this metabolic threshold varies for each complex and among different tissues. There is evidence that complexes III and IV of various tissues have to be inhibited by 70 to 80% and 60 to 70%, respectively, before respiration and ATP synthesis are compromised [27,28]. In contrast, inhibition of complex I by as little as 25% in synaptic mitochondria [27] results in impairment of respiration, whereas a 72% inhibition of this complex is necessary for respiration to be affected in nonsynaptic mitochondria [29]. Third, while many of the drugs described here are well tolerated by most patients, some drugs cause an idiosyncratic response, a feature that often arises as a result of genetic diversity. Given that there is considerable variation in mtDNA within humans and that some polymorphisms are known to predispose individuals to Alzheimer’s disease, Parkinson’s disease, and type II diabetes [30], it is

410

OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS

conceivable that polymorphic variants in the sequences of mtDNA can contribute to altered sensitivity to drugs. There are already reports to this effect, such as a case in which the rare serious adverse effect of statins unmasked underlying undiagnosed, mild mitochondrial myopathy due to mtDNA mutations [31] and cases where mtDNA polymorphisms predisposed individuals to linezolid-associated lactic acidosis [32,33] (see Chapters 4 and 11). As we proceed to assess mitochondrial dysfunction in the drug discovery and development arena, it is hoped that the immunocapture-based OXPHOS activity assays can, not only provide mechanistic information to facilitate structure–activity relationship studies for safer compounds to be synthesized, but also serve as a biomarker for mitochondrial toxicity in subsequent animal and clinical studies. When screening certain classes of compounds that have a history of causing mitochondrial toxicity, such as nucleoside reverse transcriptase inhibitors and antibiotics that inhibit bacterial protein synthesis, lateral ﬂow dipstick immunoassays will enable rapid, simple, and robust assessments of mitochondrial dysfunction both during drug development and in the clinic. Acknowledgments My sincere thanks to Rod Capaldi, Mike Marusich, James Murray, John Willis, and Autumn Bernal for valuable discussions. This work has been supported by a Pﬁzer DSRD grant and a National Institutes of Health grant 5R42-GM071052-03.

REFERENCES 1. Dykens JA, Marroquin LD, Will Y. Strategies to reduce late-stage drug attrition due to mitochondrial toxicity. Expert Rev Mol Diagn. 2007;7:161–175. 2. Chan K, Truong D, Shangari N, O’Brien PJ. Drug-induced mitochondrial toxicity. Expert Opin Drug Metab Toxicol. 2005;1:655–669. 3. Wallace KB, Starkov AA. Mitochondrial targets of drug toxicity. Annu Rev Pharmacol Toxicol. 2000;40:353–388. 4. Strehler BL, Trotter Jr. Determination of ATP and related compounds: ﬁreﬂy luminescence and other methods. Methods Biochem Anal. 1954;1:341–356. 5. Clark LC, Jr. Intravascular polarographic and potentiometric electrodes for the study of circulation. Trans Am Soc Artif Intern Org. 1960;6:348–354. 6. Nadanaciva S, Bernal A, Aggeler R, Capaldi R, Will Y. Target identiﬁcation of drug induced mitochondrial toxicity using immunocapture based OXPHOS activity assays. Toxicol in Vitro. 2007;21:902–911. 7. Murray J, Schilling B, Row RH, et al. Small scale immunopuriﬁcation of cytochrome c oxidase for a high throughput, multiplexing analysis of enzyme activity, and amount. Biotechnol Appl Biochemi 2007 May 18. 8. Aggeler R, Coons J, Taylor SW, et al. A functionally active human F1 FO ATPase can be puriﬁed by immunocapture from heart tissue and ﬁbroblast cell lines: subunit structure and activity studies. J Biol Chem. 2002;277:33906–33912.

REFERENCES

411

9. Horgan DJ, Ohno H, Singer TP, Casida JE. Studies on the respiratory chain-linked reduced nicotinamide adenine dinucleotide dehydrogenase: XV. Interactions of piericidin with the mitochondrial respiratory chain. J Biol Chem. 1968;243:5967–5976. 10. Sun F, Huo X, Zhai Y, et al. Crystal structure of mitochondrial respiratory membrane protein complex II. Cell. 2005;121:1043–1057. 11. Isom GE, Way JL. Effects of oxygen on the antagonism of cyanide intoxication: cytochrome oxidase, in vitro. Toxicol Appl Pharmacol. 1984;74:57–62. 12. Walker JE, Collinson IR, Van Raaij MJ, Runswick MJ. Structural analysis of ATP synthase from bovine heart mitochondria. Methods Enzymol. 1995;260:163–190. 13. Kramer KA, Oglesbee D, Hartman SJ, et al. Automated spectrophotometric analysis of mitochondrial respiratory chain complex enzy me activities in cultured skin ﬁbroblasts. Clin Chem. 2005;51:2110–2116. 14. Fromenty B, Fisch C, Labbe G, et al. Amiodarone inhibits the mitochondrial beta-oxidation of fatty acids and produces microvesicular steatosis of the liver in mice. J Pharmacol Exp Ther. 1990;255:1371–1376. 15. Fromenty B, Fisch C, Berson A, Letteron P, Larrey D, Pessayre, D. Dual effect of amiodarone on mitochondrial respiration: initial protonophoric uncoupling effect followed by inhibition of the respiratory chain at the levels of complex I and complex II. J Pharmacol Exp Ther. 1990;255:1377–1384. 16. Prasada Rao KS, Rao SB, Camus PH, Mehendale HM. Effect of amiodarone on Na+ -, K+ -ATPase and Mg2+ -ATPase activities in rat brain synaptosomes. Cell Biochem Funct. 1986;4:143–151. 17. Chazotte B, Vanderkooi, G. Multiple sites of inhibition of mitochondrial electron transport by local anesthetics. Biochim Biophys Acta. 1981;636:153–161. 18. Bullough DA, Kwan M, Laikind PK, Yoshida M, Allison WS. The varied responses of different F1 -ATPases to chlorpromazine. Arch Biochem Biophys. 1985;236:567–575. 19. Moreno-Sanchez R, Bravo C, Vasquez C, Ayala G., Silveira LH, Martinez-Lavin M. Inhibition and uncoupling of oxidative phosphorylation by nonsteroidal anti-inﬂammatory drugs: study in mitochondria, submitochondrial particles, cells, and whole heart. Biochem Pharmacol. 1999;57:743–752. 20. Tuquet C, Dupont J, Mesneau A, Roussaux J. Effects of tamoxifen on the electron transport chain of isolated rat liver mitochondria. Cell Biol Toxicol. 2000;16:207–219. 21. Marroquin LD, Stevens G, Will Y. Testing cascade to predict the potential of drug discovery compounds to induce mitochondrial dysfunction. Toxicologist. 2005;84:99. 22. McKee EE, Ferguson M, Bentley AT, Marks TA. Inhibition of mammalian mitochondrial protein synthesis by oxazolidinones. Antimicrob Agents Chemother. 2006;50:2042–2049. 23. Nagiec EE, Wu L, Swaney SM, et al. Oxazolidinones inhibit cellular proliferation via inhibition of mitochondrial protein synthesis. Antimicrob Agents Chemother. 2005;49:3896–3902. 24. De Vriese AS, Coster RV, Smet J, et al. Linezolid-induced inhibition of mitochondrial protein synthesis. Clin Infect Dis. 2006;42:1111–1117. 25. Garrabou G, Soriano A, Lopez S, et al. Reversible inhibition of mitochondrial protein synthesis during linezolid-related hyperlactatemia. Antimicrob Agents Chemother. 2007;51:962–967. 26. Riesbeck K, Bredberg A, Forsgren A. Ciproﬂoxacin does not inhibit mitochondrial functions but other antibiotics do. Antimicrob Agents Chemother. 1990;34:167–169.

412

OXPHOS COMPLEX ACTIVITY ASSAYS AND DIPSTICK IMMUNOASSAYS

27. Davey GP, Peuchen S, Clark JB. Energy thresholds in brain mitochondria: potential involvement in neurodegeneration. J Biol Chem. 1988;273:12753–12757. 28. Letellier T, Heinrich R, Malgat M, Mazat JP, 1994. The kinetic basis of threshold effects observed in mitochondrial diseases: a systemic approach. Biochemi J. 1994;302:171–174. 29. Davey GP, Clark JB. Threshold effects and control of oxidative phosphorylation in nonsynaptic rat brain mitochondria. J Neurochem. 1996;66:1617–1624. 30. Wallace DC. Mitochondrial DNA sequence variation in human evolution and disease. Proc Nat Acad Sci U S A. 1994;91:8739–8746. 31. Diaczok BJ, Shali R. Statins unmasking a mitochondrial myopathy: a case report and proposed mechanism of disease. South Med J. 2003;96:318–320. 32. Palenzuela L, Hahn NM, Nelson RP Jr, et al. Does linezolid cause lactic acidosis by inhibiting mitochondrial protein synthesis? Clin Infect Dis. 2005;40:113–116. 33. Carson J, Cerda J, Chae JH, Hirano M, Maggiore P. Severe lactic acidosis associated with linezolid use in a patient with the mitochondrial DNA A2706G polymorphism. Pharmacotherapy. 2007;27:771–774.

17 USE OF FLUORESCENT REPORTERS TO MEASURE MITOCHONDRIAL MEMBRANE POTENTIAL AND THE MITOCHONDRIAL PERMEABILITY TRANSITION Anna-Liisa Nieminen and Venkat K. Ramshesh Center for Cell Death, Injury and Regeneration, Charleston, South Carolina; Department of Pharmaceutical and Biomedical Sciences, Hollings Cancer Center, Medical University of South Carolina, Charleston, South Carolina

John J. Lemasters Center for Cell Death, Injury and Regeneration, Charleston, South Carolina; Departments of Pharmaceutical and Biomedical Sciences and Biochemistry and Molecular Biology, Hollings Cancer Center, Medical University of South Carolina, Charleston, South Carolina

1. 2. 3. 4.

Introduction Isolated mitochondria Nernstian distribution of ﬂuorescent probes Monitoring membrane potential in isolated mitochondria 4.1. Pitfalls with potential-indicating fluorophores 5. Imaging in single intact cells 5.1. Image acquisition and processing 5.2. Background subtraction 5.3. Fluorescence of the extracellular space 5.4. Pixel-by-pixel calculation of 6. Mitochondrial permeability transition 6.1. Swelling assay of the MPT 6.2. Release of carboxydichloroﬂuorescein

414 414 416 417 417 420 420 421 421 422 423 424 424

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

413

414

7. Visualizing the MPT in intact cells 8. Plasma membrane permeability 9. Conclusions

USE OF FLUORESCENT REPORTERS

425 428 428

1. INTRODUCTION A protonmotive force (p) generated by respiratory chain activity in the mitochondrial inner membrane drives mitochondrial ATP production. p is made up primarily of a negative electrical potential difference ( m ) across the mitochondrial inner membrane. Thus, m is the principal driving force for mitochondrial ATP synthesis to meet cellular energy demands. Inhibition of mitochondrial ATP production due to perturbations of m frequently occurs after injurious stresses and can have catastrophic consequences on cellular function. Consequently, m is widely measured in studies of the pathogenesis of cellular injury and death. Two approaches are frequently employed to monitor mitochondrial after various treatments. In the ﬁrst, mitochondria are isolated from tissues and assessed for their ability to generate m either after direct exposure to toxicants or after tissue exposure to injurious stresses. In the second approach, m is measured in situ in cultured cells or the living animal.

2. ISOLATED MITOCHONDRIA Before the development of cellular imaging technologies, m measurements were performed primarily in isolated mitochondria puriﬁed from tissues by homogenization and differential centrifugation in isoosmotic sucrose or sucrose/mannitol solution (Table 1). After Teﬂon-on-glass homogenization, the crude homogenate is centrifuged at about 600g to remove nuclei and incompletely disrupted cells. The supernatant is then centrifuged at about 10,000g to sediment the mitochondria, which are resuspended and repelleted two or three more times to yield puriﬁed mitochondria. Mitochondria isolated from liver originate mostly from hepatocytes, and contamination with mitochondria from endothelial cells or Kupffer cells is minor. Since liver mitochondria are quite homogeneous and can be obtained in large amounts, isolated liver mitochondria preparations are commonly used to study the effects of drugs or toxins on mitochondrial function, especially since the liver is a frequent target of drug toxicity. Similarly, left ventricular heart mitochondria can be isolated in large numbers and high purity from a variety of species, including beef and common laboratory animals. However, many heart mitochondria are enmeshed in the myoﬁbrilar apparatus and are difﬁcult to release without digestion of the myoﬁbrils with enzyme preparations such as Nagarse.

415

ISOLATED MITOCHONDRIA

TABLE 1

Isolation of Rat Liver Mitochondria

Step Preparation Isolation medium Homogenization

Centrifuge 600g × 10 min 10,000g × 15 min

10,000g × 10 min

Comments Wash all glassware and utensils in deionized distilled water to remove detergent residues. Prechill. Ice-cold 0.25 M sucrose (2 mM K-HEPES buffer, pH 7.4, and 0.57 mM EGTA is optional). Homogenize freshly excised rat liver cut in small pieces at 30% w/v with three or four strokes in a Potter–Elvehjem tissue grinder with a Teﬂon pestle. Use a loose-ﬁtting pestle for the ﬁrst stroke, to avoid excess pressure. Keep ice cold at all times. Carefully remove supernatant for the next step, taking care not to disturb the loose pellet of nuclei, erthryocytes, and partially broken cells. Aspirate; muddle the pellet with a precooled glass or Teﬂon rod in a few drops of medium and resuspend in 40 mL. Use EGTA-free medium for this and subsequent steps. Muddle around any erythrocytes at the bottom of the pellet so that they are not resuspended. Repeat twice. Aspirate the supernatant and the ﬂuffy surface layer of the pellet. Resuspend the ﬁnal pellet in ∼1 mL of medium. Measure the protein and adjust the concentration to 50 or 100 mg protein/mL.

Source: Data from Schneider [49] and Lemasters and Hackenbrock [50].

Isolation of mitochondria from brain tissue is more problematic because of the intrinsic cellular heterogeneity of the nervous system. Isolation of mitochondria from whole brain results in a heterogeneous population of mitochondria that originate from various regions of brain, such as cortex, hippocampus, cerebellum, striatum, and substantia nigra. These mitochondria differ in size, enzyme content, and apparent susceptibility to toxicants. For example, systemic administration of complex II (cytochrome bc 1 ) inhibitors such as 3-nitropropionate induces selective degeneration of striatal neurons that resembles Huntington disease both pathologically and behaviorally [1,2]. Similarly, complex I (NADH–ubiquinone oxidoreductase) inhibition with rotenone targets dopaminergic neurons in substantia niagra, producing a Parkinson-like syndrome in rats [3]. Therefore, mitochondrial preparations pooling different brain regions are not ideal to study the effects of neurotoxins on mitochondria. Isolation of mitochondria from speciﬁc regions of brain, such as cerebral cortex, hippocampus, and striatum, is possible, but yields will be very low from laboratory animals such as rats and mice [4,5]. Mitochondrial fractions from brain are typically contaminated with synaptosomes. Synaptosomes have a negative membrane potential and therefore accumulate positively charged membrane potential indicators, which can lead to incorrect

416

USE OF FLUORESCENT REPORTERS

inferences regarding mitochondrial membrane potential. Different strategies are employed to remove synaptosomes from mitochondrial preparations. Percoll density gradient centrifugation removes synaptosomes based on density differences of mitochondria and synaptosomes to yield metabolically active and well-coupled mitochondria [6]. Another well-established method to remove synaptosomes is to permeabilize them with a small amount of digitonin. Digitonin preferably binds to cholesterol containing membranes lysing synaptosomal membranes but sparing mitochondria [7]. A disadvantage of the digitonin method is that residual amounts of digitonin may alter the permeability of the outer mitochondrial membrane, promoting mitochondrial swelling and dysfunction [8]. In that respect, density gradient centrifugation may be the method preferred to isolate synaptosome-free mitochondria. Methods of mitochondrial isolation are optimized to select healthy mitochondria over damaged mitochondria, which becomes a problem when studying the effects of drugs and toxicants. Damaged mitochondria are often swollen, sediment differently from healthy mitochondria, and are not recovered in differential centrifugation and density gradient procedures. Thus, mitochondrial damage after in vivo drug–toxin exposure may be underestimated in isolated mitochondrial preparations. Measurement of marker enzymes of the inner membrane, such as succinate dehydrogenase, is useful to monitor mitochondrial recovery.

3. NERNSTIAN DISTRIBUTION OF FLUORESCENT PROBES Membrane-permeant cationic ﬂuorescent dyes are commonly used to measure mitochondrial membrane potential () in isolated mitochondria and intact cells. These dyes have a delocalized positive charge and accumulate electrophoretically in response to the electrical potential [9–11]. At equilibrium, distribution of the permeant dye is dictated by the Nernst equation: = −59log

Fin Fout

(1)

where is the electrical potential difference in millivolts, F in the cationic ﬂuorophore concentration inside mitochondria, and F out the ﬂuorophore concentration outside mitochondria. Because mitochondrial can be as great as −180 mV, as much as a 1000-fold concentration gradient can form across the mitochondrial inner membrane. These ﬂuorophores appear to be speciﬁc for mitochondria because mitochondrial is so much more negative than other cellular compartments, but potential-indicating ﬂuorophores will be taken up into any negatively charged compartment. Typically, plasmalemmal is negative. Thus, cationic ﬂuorophores are ﬁrst taken up by plasmalemmal and then by mitochondrial . In this way, plasmalemmal magniﬁes mitochondrial uptake, and if plasmalemmal decreases, mitochondria will lose ﬂuorescence even if mitochondrial is unchanged.

MONITORING MEMBRANE POTENTIAL IN ISOLATED MITOCHONDRIA

417

4. MONITORING MEMBRANE POTENTIAL IN ISOLATED MITOCHONDRIA In isolated mitochondria, uptake of membrane-potential-indicating ﬂuorophores into the matrix space leads to self-quenching of the ﬂuorophores and a red shift of their absorbance spectrum that varies linearly with . Thus, for ﬂuorophores such as rhodamine 123, safranine, tetramethylrhodamine methyl ester mitochondrial, and others, a decrease in total ﬂuorescence or a change in absorbance of mitochondria suspended in a cuvette signiﬁes mitochondrial polarization [10,12,13]. These cuvette-based techniques have also been adapted to multiwell plate readers to monitor mitochondrial in all the wells of 24and 96-well microtiter plates virtually simultaneously [14]. 4.1. Pitfalls with Potential-Indicating Fluorophores A variety of potential-indicating ﬂuorophores are available for membrane potential measurements. The recognition of the possible drawbacks related to the use of different ﬂuorophores is critical in achieving meaningful results. The most common problems associated with the probes are interference with cell and mitochondrial metabolism, phototoxicity, especially when imaging is performed by laser scanning confocal microscopy, and probe binding to cellular proteins. Inhibition of mitochondrial respiration increases in the order tetramethylrhodamine methyl ester (TMRM) < rhodamine 123 < tetramethylrhodamine ethyl ester << DiOC6 [13,15]. Rhodamine 123 at high matrix concentrations inhibits the mitochondrial ATP synthase of oxidative phosphorylation [10]. DiOC6 , which is frequently used in ﬂow cytometric studies, strongly inhibits mitochondrial respiration. Therefore, to avoid this effect, DiOC6 should be used, at less than 1 nM loading concentration [16]. Of the commonly used probes, TMRM appears to be the least toxic to mitochondrial metabolism [13]. Rhodamine 123 is used principally in the matrix quenching mode in isolated mitochondria and intact cells. Rhodamine 123 as well as TMRM can accumulate in mitochondria to levels that exceed those predicted by the Nernst equation [10]. Excess accumulation may be due to ﬂuorophore stacking and formation of aggregates. Formation of aggregates causes ﬂuorescence quenching and a red shift of absorbance. In cuvette and multiwell assays, decreased ﬂuorescence due to quenching signiﬁes an increase of mitochondrial [10,13,14]. Similarly, increased ﬂuorescence signiﬁes depolarization, and a 50% increase in total rhodamine 123 ﬂuorescence occurs in cultured hepatocytes after mitochondrial depolarization with HgCl2 [17]. The increase in total cellular rhodamine 123 ﬂuorescence represents unquenching of ﬂuorescence when the probe moves out from mitochondria into the cytosol. Aggregate formation and quenching are dependent on ﬂuorophore concentration and can be minimized by using lower loading concentrations. Cationic ﬂuorophores used to measure mitochondrial membrane potential are redistribution probes, since the probes must physically redistribute between

418

USE OF FLUORESCENT REPORTERS

intra- and extramitochondrial compartments as changes. Such redistribution takes time, and temporal resolution of changes of from simple intensity measurements is not better than several seconds. Nicholls, especially, has developed kinetic models that utilize the rate as well as the magnitude of ﬂuorescence changes to improve temporal resolution [18]. This approach requires characterization of the kinetics of probe redistribution across both the mitochondrial and plasma membranes, which are, in general, different for different cell types. Measurement of total cellular ﬂuorescence of potential-indicating ﬂuorophores has its drawbacks because total ﬂuorescence does not detect changes in individual mitochondria but, instead, averages all mitochondria. Moreover, for intact cells, total ﬂuorescence will be inﬂuenced by changes in both mitochondrial and plasmalemmal . Therefore, ﬂuorescence microscopy has become a more attractive method, as shown originally for rhodamine 123 [19]. Fluorescence microscopy, especially confocal and multiphoton microscopy, permit observations of the heterogeneity of mitochondrial responses. For example, after exposure of rat phaeocromocytoma 6 cells to staurosporine, mitochondrial release TMRM in an asynchronous manner [20]. Confocal images reveal that some cellular mitochondria depolarize abruptly, whereas other mitochondria remain fully polarized for a longer period. In this example, the heterogeneity of the mitochondrial response within individual cells would not have been detected without high-resolution confocal imaging. Like rhodamine 123 and TMRM, JC-1 accumulates into mitochondria electrophoretically. Inside mitochondria, JC-1 forms precipitates, called J-aggregates, in a concentration-dependent fashion [21,22]. Other probes may also form aggregates, but in contrast to rhodamine 123 and TMRM, J- aggregation formation by JC-1 leads to a shift from green to red ﬂuorescence emission instead of simple ﬂuorescence quenching. Thus, mitochondrial hyperpolarization increases red ﬂuorescence. Agents that abolish (e.g., protonophores) then cause loss of red ﬂuorescence, and the ratio of red to green ﬂuorescence reﬂects the magnitude of mitochondrial polarization [21]. Many studies have used a ratio of red and green ﬂuorescence as a semiquantitative measure of in intact cells [23–25]. One of the disadvantages of JC-1 is that the probe equilibrates very slowly across the plasma membrane. Mitochondria near the plasma membrane surface may accumulate more JC-1 and therefore appear to have higher membrane potential than those farther away from the plasma membrane. Therefore, heterogeneity of the mitochondrial membrane potential between individual mitochondria within a cell may be due simply to uneven distribution of JC-1 and not to variations. Another problem associated with JC-1 is that its red-ﬂuorescing microprecipitates within a single mitochondrion give the impression that mitochondrial is heterogeneous even within an individual mitochondrion. Confocal images of JC-1 loaded hepatocyte show these JC-1 precipitates within the matrix of individual mitochondria ﬂuorescing with the green JC-1 monomer (Figure 1). The fact that the red-ﬂuorescing J-aggregates do not ﬁll the entire mitochondrial matrix

MONITORING MEMBRANE POTENTIAL IN ISOLATED MITOCHONDRIA

419

5 µm Figure 1 Red and green mitochondrial ﬂuorescence after loading with JC-1. A cultured mouse hepatocyte was loaded with 100 nM JC-1 for 30 minutes in Krebs–Ringer–HEPES buffer (KRH) at 37◦ C, and green and red ﬂuorescence were imaged by multitrack confocal microscopy using 488- and 543-nm excitation light, respectively. Red inclusions within green ﬂuorescing mitochondria are JC-1 J-aggregates. (See insert for color representation of ﬁgure.)

indicates that the aggregates are indeed microprecipitates. In Figure 1, aggregates are seen to be localized to the lateral margins of mitochondria, presumably in association with mitochondrial membranes. Heterogeneity of red and green ﬂuorescence in single JC-1-loaded mitochondria thus represents the physical distribution of microprecipitates, and such heterogeneity does not indicate variation of within the single mitochondria. Formation of physical aggregates of JC-1 within mitochondria limits the use of the probe for high-resolution imaging of cellular and mitochondrial . During mitochondrial depolarization, most potential-indicating ﬂuorophores are released from mitochondria. However, a group of cationic ﬂuorophores, called MitoTracker probes, accumulate electrophoretically into mitochondria but are retained within mitochondria even after depolarization. Inside mitochondria, chloromethyl groups of MitoTracker probes form covalent adducts with protein sulfhydryls [26]. Once the probes have formed adducts, ﬂuorescence of the probes becomes independent of mitochondrial . Since MitoTracker ﬂuorescence is retained even after ﬁxation, they are commonly used also to co-localize mitochondria in immunocytochemical preparations. However, the intensity of MitoTracker ﬂuorescence will depend on many factors, including the time, typically many minutes, required to form covalent bonds with protein thiols. In general, MitoTracker ﬂuorescence should not be used as a measure of changes

420

USE OF FLUORESCENT REPORTERS

of . However, MitoTracker Green used in combination with TMRM can be used to visualize selectively mitochondria undergoing depolarization [27].

5. IMAGING IN SINGLE INTACT CELLS Confocal and multiphoton microscopy create submicron optical sections through single cells. With this resolution, ﬂuorophore accumulation of individual mitochondria can be quantiﬁed in comparison to extramitochondrial ﬂuorophore concentration. Using confocal microscopy to measure the intracellular distribution of membrane potential–indicating ﬂuorophores, the distribution of intracellular ( relative to the extracellular space) can then be calculated using equation (1). From differences of between subcellular compartments, individual membrane potentials can be determined, such as plasmalemmal (cytosolic minus extracellular ) and mitochondrial (mitochondrial minus cytosolic ). In most cells, plasmalemmal ranges from −30 to −100 mV and mitochondrial ranges from −120 to −160 mV. Because these ’s are additive, mitochondria are 150 to 260 mV more negative than the extracellular space. Such large s correspond to cation concentration ratios between mitochondria and the extracellular space that can exceed 10,000 to 1 [equation (1)]. Such large gradients cannot be imaged using a conventional linear scale of 256 gray levels per pixel (8-bit pixel memory). Rather, 12- to 16-bit memory, sequential imaging at different laser powers, or a nonlinear logarithmic (gamma) scale must be used instead [28]. Gamma scales, which are commonly used in scanning electron microscopy, logarithmically compress input signals into the available 256 gray levels of pixel memory. Earlier models of confocal microscopes, such as the Bio-Rad MRC-600, included gamma imaging circuits, but gamma circuitry unfortunately is no longer available in most current laser scanning confocal/multiphoton microscope systems. 5.1. Image Acquisition and Processing After cells have been loaded with a ﬂuorophore, confocal/multiphoton images may then be collected. At least two images are required. The ﬁrst image is an optical section through the specimen of interest. The second image is that obtained after refocusing inside the glass coverslip. The latter image serves as a background image. Both images must be collected using identical instrumental settings of laser power, gain, and brightness. Image oversaturation (pixels at the highest gray level) and undersaturation (pixels with a zero gray level) must be avoided. Care must be taken to avoid laser-induced photodamage. A low laser power setting (≤1%) should be used to minimize photodamage, especially if serial imaging over time of the same area of the specimen will be performed. If the microscope can collect images with a pixel depth of 12 or 16 bits (4096 and 65,536 gray levels, respectively), additional images may not be needed. However, if the pixel depth is only 8 bits (256 gray levels) or if the ratio of ﬂuorescence

IMAGING IN SINGLE INTACT CELLS

421

intensity between mitochondria and the extracellular space exceeds the gray-level range, additional images at about 10 times greater laser power must be collected both in the cell and within the glass coverslip. With confocal microscopes equipped with multitracking systems, lower and higher laser power images can be collected simultaneously one line at a time. Each line (row of pixels) in the images is collected sequentially. A ﬁrst line scan is performed at lower laser power, followed immediately by a second line scan at higher laser power, a process that is repeated with each succeeding line in the image. Multitracking capability permits the two images to be acquired simultaneously and eliminates problems associated with specimen movement and focus drift. When images at higher laser power are collected, oversaturation of mitochondria will be evident, but oversaturation in higher power images does not pose a problem, since these images will only be used to quantify areas of weak ﬂuorescence intensity in the extracellular space. After processing images collected through the sample and within the glass coverslip, maps of the intracellular distribution of can be generated by a three-step procedure: (1) background subtraction, (2) quantitation of extracellular ﬂuorescence, and (3) calculation of for each pixel gray level to create a pseudocolor map of .

5.2. Background Subtraction Light detectors such as the photomultiplier tubes commonly used in confocal microscopes generate signals even in the absence of light. This background signal must be subtracted from the signal collected in the presence of light in order to obtain an output truly proportional to ﬂuorescence intensity. In confocal microscopy, images collected in the plane of the coverslip represent this background, since added ﬂuorophore cannot penetrate the glass. A mean pixel value is then calculated for all the pixels of each background image using image analysis software (e.g., ImageJ, Adobe Photoshop, MetaMorph). This average background value is then subtracted from every pixel of the corresponding specimen image. The background-corrected images represent the true relative distribution of ﬂuorescence intensity within the images.

5.3. Fluorescence of the Extracellular Space Since ﬂuorescence should be the same everywhere in the extracellular space, all areas of the extracellular space in the background-subtracted specimen image are selected, and average pixel intensity is determined. In case extracellular ﬂuorescence is too weak, extracellular ﬂuorescence can be measured in the same way at higher laser power. Division (after background subtraction) of mean extracellular ﬂuorescence from the higher laser power image by the power ratio between the higher and lower power images then yields an estimate of extracellular ﬂuorescence for the lower power image.

422

USE OF FLUORESCENT REPORTERS

5.4. Pixel-by-Pixel Calculation of Using equation (1), a value for ( relative to the extracellular space) can be calculated for every pixel of the background-corrected image on the assumption that ﬂuorescence intensity is proportional to monovalent cationic ﬂuorophore concentration. To display the intracellular distribution of , the images can be pseudocolored by assigning different colors to speciﬁc millivolt ranges of . To determine the pixel value corresponding to a speciﬁc millivolt value of , equation (1) is rearranged: 59 log Pout − Pi = antilog (2) 59 where P out is the average background-subtracted pixel intensity in the extracellular space and P i is the pixel value representing a particular millivolt value of . Ranges of gray-level values corresponding to speciﬁc ranges of can be calculated from equation (2). Individual colors can be assigned to different gray levels. Figure 2 illustrates a pseudocolored image of a TMRM-loaded adult feline cardiac myocyte imaged by confocal microscopy. The difference in between the extracellular space (where is zero) and the cytosol/nucleus represents plasmalemmal , whereas the difference between the cytosol/nucleus and mitochondria represents mitochondrial . In the cardiac myocyte, the cytoplasm

Figure 2 Distribution of electrical potential in a cardiac myocyte. An adult feline cardiac myocyte was loaded with 200 nM TMRM for 20 minutes at 37◦ C, and TMRM ﬂuorescence was imaged by confocal microscopy using 543-nm excitation and a 565 to 615-nm emission ﬁlter. The distribution of is displayed in pseudocolor, as described in the text. (See insert for color representation of ﬁgure.)

MITOCHONDRIAL PERMEABILITY TRANSITION

423

was so densely packed with mitochondria that cytosolic needs to be estimated from the nucleus, which has the same potential as the cytosol. Thus, the plasmalemmal (sarcolemmal) was estimated to be about −100 mV. This plasmalemmal is consistent with the plasmalemmal polarization of excitable cells. Since mitochondria is up to −220 mV, mitochondrial is in the range of −120 mV. In hepatocytes, plasmalemmal is about −30 mV, but mitochondrial remains −120 mV (not shown). Mitochondrial diameter in hepatocytes and cardiac myocytes is about 1 µm. Since the thickness of confocal optical sections is slightly less than 1 µm, some mitochondria occupy the entire thickness of the optical slices, but some mitochondria occupy only part of the thickness of the section. Mitochondria that occupy the confocal optical sections only partially contribute to apparent heterogeneity and underestimation of mitochondrial . Thus, the highest values of mitochondrial probably reﬂect true mitochondrial . In other cell types, such as cancer cells, mitochondrial diameter is much less than 1 µm, which makes it more difﬁcult to estimate mitochondrial correctly.

6. MITOCHONDRIAL PERMEABILITY TRANSITION In the mitochondrial permeability transition (MPT), opening of nonselective, highly conductive permeability transition (PT) pores causes the inner membrane of mitochondria to become permeable to molecules of up to 1500 Da, which leads to mitochondrial depolarization, uncoupling of oxidative phosphorylation, and large-amplitude mitochondrial swelling [29]. Cyclosporin A (CsA) and various of its analogs inhibit PT pore opening. The molecular composition of PT pores remains uncertain. In one model, PT pores form from the adenine nucleotide translocator (ANT) in the inner membrane, the voltage-dependent anion channel (VDAC) in the outer membrane, and cyclophilin D (CypD), a CsA-binding protein from the matrix (reviewed in [30,31]). Recent ﬁndings that the MPT still occur in mitochondria that are deﬁcient in ANT, VDAC, and even CypD challenge the validity of this model [32–34]. In another model, PT pores form from damaged, misfolded membrane proteins that aggregate at hydrophilic surfaces facing the bilayer to create aqueous transmembrane channels [35]. CypD and other molecular chaperones close these nascent PT pores to protect mitochondria from depolarization and swelling. CypD confers sensitivity to Ca2+ by which increased matrix Ca2+ opens the PT pores, an action antagonized by CsA. When nascent pores formed by misfolded protein aggregates exceed the chaperones available to regulate their conductance, unregulated PT pore opening occurs. Such unregulated PT pores are insensitive to CsA inhibition and no longer require Ca2+ for opening. The protein misfolding model accounts for why the MPT occurs in ANT-deﬁcient mitochondria [32], since misfolding of other mitochondrial membrane proteins causes PT pore formation in the absence of ANT. Recently, strains of CypD knockout mice have been created. Mitochondria from these mice are desensitized

424

USE OF FLUORESCENT REPORTERS

to onset of the Ca2+ -induced MPT, and higher Ca2+ is needed to induce the MPT [33], which is consistent with a role of CypD as the Ca2+ sensor of the MPT. Growing evidence supports a critical role of the MPT in cell necrosis and apoptosis from oxidant stress, Ca2+ toxicity, warm ischemia/reperfusion, cytokines, Fas ligation, bile acid toxicity, ethanol, acetaminophen, pharmaceutical xenobiotics, and a variety of other stresses to hepatocytes and other cell types [36–39]. 6.1. Swelling Assay of the MPT When the MPT occurs, mitochondrial membrane potential collapses, and onset of the MPT can be monitored via potential-indicating ﬂuorophores. However, the MPT is not the only cause of mitochondrial depolarization. A more speciﬁc assay is to follow mitochondrial swelling as assayed by a decrease in absorbance of visible light (680 nM). At the onset of the MPT, large-amplitude swelling occurs driven by the colloid osmotic force of protein in the matrix. For a dilute mitochondrial suspension (0.2 to 0.5 mg protein/ml), this swelling leads a decrease of absorbance of 0.2 to 0.4 unit, which is discernible even to the naked eye (Figure 3). The requirements for this in vitro assay in isolated mitochondria are (1) an absence of Mg2+ , since Mg2+ strongly inhibits the MPT); (2) a source of energy, such as respiratory substrate (e.g., succinate plus rotenone); (3) 1 to 3 mM P i , which promotes Ca2+ uptake as well as independently promoting the MPT; and (4) boluses of CaCl2 to induce MPT onset. In such an assay, addition of Ca2+ causes large-amplitude swelling (decrease of absorbance) beginning after a delay of several seconds to even a few minutes (Figure 3). A variety of reactive chemicals (e.g., phenylarsine oxide, diamide) act to lower the threshold of the Ca2+ -induced MPT, whereas MPT inhibitors (e.g., CsA) increase the threshold for Ca2+ . Importantly, at the onset of the MPT, all Ca2+ that has been taken up into the matrix is released abruptly. Thus, MPT onset can also be monitored by Ca2+ release using Ca2+ -ﬂuorophores such as Fluo-5 N added to the extracellular medium [14]. 6.2. Release of Carboxydichloroﬂuorescein At the onset of the MPT, trapped solutes inside mitochondria are released into the medium. To assess this aspect of PT pore opening, mitochondria can be ester loaded with carboxydichloroﬂuorescein diacetate during isolation. Matrix esterases release carboxydichloroﬂuorescein free acid (carboxy-DCF) which remains trapped in the matrix space. Like other ﬂuorophores in the matrix, the ﬂuorescence of carboxy-DCF is partially quenched. At onset of the MPT, carboxy-DCF is released into the medium with unquenching of ﬂuorescence. Hence, an increase in carboxy-DCF ﬂuorescence indicates increased nonspeciﬁc permeability of the mitochondrial inner membrane to carboxy-DCF (Figure 3).

VISUALIZING THE MPT IN INTACT CELLS

425

A

B

Figure 3 Mitochondrial swelling and release of carboxy-DCF after induction of the mitochondrial permeability transition. Rat liver mitochondria were loaded with 5 mM carboxy-DCF diacetate during isolation. Carboxy-DCF-loaded mitochondria (0.5 mg/mL) were added to incubation buffer containing 100 mM KCl, 20 mM Tris, 20 mM K-HEPES, 10 mM NaCl, 1 mM KH2 PO4 , 5 mM succinate, 20 µM EGTA, 2 µM rotenone, and 1 µg/ml oligomycin. Absorbance change at 620 nm (A) and carboxy-DCF ﬂuorescence excited at 485 nm (B) were determined using a multiwell ﬂuorescence plate reader. CaCl2 (100 µM) was added after 1 minute (ﬁrst arrow). Alamethicin (5 µM, second arrow) was added to permeabilize and induce swelling of all mitochondria. Dashed lines indicate gaps in measurements as additions were made.

7. VISUALIZING THE MPT IN INTACT CELLS Many of the ion-indicating ﬂuorophores used for optical microscopy of living cells are multivalent organic anions. To load these molecules into cells, the charged acids are neutralized by forming acetate or acetoxymethyl esters, which are membrane permeable. After incubation with the esters, cytosolic esterases regenerate the free acid form of the ﬂuorophores, which become trapped in the cytosol. The temperature of loading strongly affects intracellular distribution [40]. In cells like cultured hepatocytes and cardiac myocytes, warm loading at 37◦ C promotes nearly exclusive loading into the cytosol and nucleus, whereas cold loading leads to loading of mitochondria, lysosomes, and possibly other intracellular organelles. Calcein is a pentacarboxylic acid ﬂuorophore whose ﬂuorescence is independent of phyisological changes of intracellular ions. Calcein acetoxymethyl ester (AM) loaded at 37◦ C localizes almost exclusively to the cytosol and nucleus

426

USE OF FLUORESCENT REPORTERS

(Figure 4). After warm ester loading, mitochondria exclude calcein and appear instead as small round dark voids in the green calcein ﬂuorescence. These voids correspond exactly to mitochondria, as shown by simultaneous loading with TMRM (Figure 4). Mitochondria exclude calcein because the PT pores are closed and the inner membrane is otherwise impermeable to this very polar 623-Da solute. However, at the onset of the MPT, PT pores open and permit calcein to enter the matrix, as illustrated in Figure 4 for a rat hepatocyte exposed to the oxidant chemical, t-butyl hydroperoxide. As a consequence, the dark voids disappear. Simultaneously, TMRM ﬂuorescence disappears, indicating mitochondrial depolarization. Inhibition of these changes by CsA then supports the conclusion that the inner membrane permeabilization and depolarization is speciﬁcally due to the MPT.

Figure 4 Increased mitochondrial inner membrane permeability in a rat hepatocyte induced by t -butyl hydroperoxide. A cultured rat hepatocyte was loaded with TMRM (left panel) and calcein (right panel). Note that dark round voids in the green calcein ﬂuorescence coincide with red TMRM labeling of mitochondria. After 9 minutes exposure to 100 µM t-butyl hydroperoxide, dark mitochondrial voids ﬁlled with green calcein ﬂuorescence. Simultaneously, mitochondria release red TMRM ﬂuorescence. These events signiﬁed the onset of the MPT. (See insert for color representation of ﬁgure.)

VISUALIZING THE MPT IN INTACT CELLS

427

In contrast to loading at 37◦ C, calcein-AM loading at 4◦ C causes entry of calcein into both cytosol and mitochondria [40]. Presumably during warm loading, cytosolic esterases are so active that calcein AM is hydrolyzed before it can enter mitochondria. At 4◦ C, esterase activity is slowed, allowing movement of unhydrolyzed calcein AM into mitochondria where mitochondrial esterases cleave the ester to entrap calcein free acid. The temperature dependence of ester loading is both probe and cell-type speciﬁc, but the loading technique works well in both cultured hepatocytes and adult heart cardiac myocytes [41,42]. The plasma membrane gradually pumps ﬂuorophores out from the cytosol through an organic anion carrier [43]. Fluorophores trapped in mitochondria, however, are retained. Thus, when cold calcein-AM loading is by warm incubation of several hours, calcein localization becomes almost exclusively mitochondrial (Figure 5) [44]. Now at onset of the MPT, instead of calcein moving into the mitochondria, calcein moves out, as illustrated in Figure 5 for a cardiac myocyte after ischemia–reperfusion [45]. Due to unquenching, overall calcein ﬂuorescence also becomes brighter after release. Due to differences in esterase activity and intracellular distribution in different cells types, speciﬁc localization of calcein to cytosol or mitochondria may not be possible. In this instance, calcein AM may be loaded in the presence of 1 mM CoCl2 [46]. Co2+ strongly quenches calcein but enters only the cytosol. Hence, mitochondria, but not cytosol, ﬂuoresce after calcein ester loading with CoCl2 . Then after onset of the MPT, release of calcein into the cytosol and quenching by Co2+ lead to the loss of calcein ﬂuorescence. This technique for visualizing the MPT in living cells has proven useful in various cancer cell lines [46,47].

Figure 5 Inner membrane permeabilization after ischemia/reperfusion in rat myocytes visualized by mitochondrial calcein release after cold ester loading/warm incubation. An adult rat cardiac myocyte was cold-loaded with calcein AM and subjected to 3 hours of simulated ischemia at 37◦ C at pH 6.2 followed by reperfusion at pH 7.4 for 10 and 20 minutes. Green calcein ﬂuorescence was retained by mitochondria at the end of ischemia. After reperfusion, mitochondria began to release calcein, signifying inner membrane permeabilization. For experimental details, see [45]. (See insert for color representation of ﬁgure.)

428

USE OF FLUORESCENT REPORTERS

8. PLASMA MEMBRANE PERMEABILITY Directed compartmental loading of calcein also allows direct observation in cells of changes of permeability of the plasma membrane. For example, at the onset of cell death, after toxic and hypoxic injury, trapped cytosolic calcein is lost almost instantaneously [48]. Similarly, when calcein is in the extracellular space, the ﬂuorophore enters the cell interior at the onset of cell death. Other extracellular ﬂuorophores also enter cells at the onset of cell death, such as propidium iodide, which binds to DNA with an enhancement of ﬂuorescence.

9. CONCLUSIONS Quantitative confocal/multiphoton microscopy of the intracellular distribution of membrane-permeant cationic ﬂuorophores provides a minimally perturbing means to measure dynamically both mitochondrial and plasmalemmal in living cells. The method eliminates artifacts associated with measurement of by ﬂow cytometry and related nonmicroscopic techniques. A limiting factor is time resolution, since several seconds are required for cationic ﬂuorophores to reestablish a new steady-state equilibrium after a change of . Unlike virtually any other technique, confocal microscopy permits nondestructive serial observation of the values of populations of cells and their individual mitochondria. Similarly, confocal microscopy can be adapted to visualize directly the onset of inner membrane permeabilization due to the MPT in single mitochondria of living cells. Live-cell, three dimensionally–resolved confocal (and multiphoton) microscopy is now an indispensable tool for studying mitochondrial function and pathophysiology. Acknowledgments We thank Insil Kim for mouse hepatocytes and Donald R. Menick for adult feline cardiac myocytes.

REFERENCES 1. Beal MF, Hyman BT, Koroshetz W. Do defects in mitochondrial energy metabolism underlie the pathology of neurodegenerative diseases? Trends Neurosci. 1993;16:125–131. 2. Brouillet E, Jacquard C, Bizat N, Blum D. 3-Nitropropionic acid: a mitochondrial toxin to uncover physiopathological mechanisms underlying striatal degeneration in Huntington’s disease. J Neurochem. 2005;95:1521–1540. 3. Panov A, Dikalov S, Shalbuyeva N, Taylor G, Sherer T, Greenamyre JT. Rotenone model of Parkinson disease: multiple brain mitochondria dysfunctions after short term systemic rotenone intoxication. J Biol Chem. 2005;280:42026–42035.

REFERENCES

429

4. Genova ML, Bovina C, Marchetti M, et al. Decrease of rotenone inhibition is a sensitive parameter of complex I damage in brain non-synaptic mitochondria of aged rats. FEBS Lett. 1997;410:467–469. 5. Battino M, Gorini A, Villa RF, et al. Coenzyme Q content in synaptic and non-synaptic mitochondria from different brain regions in the ageing rat. Mech Ageing Dev. 1995;78:173–187. 6. Anderson MF, Sims NR. Improved recovery of highly enriched mitochondrial fractions from small brain tissue samples. Brain Res Brain Res Protocols, 2000;5:95–101. 7. Rosenthal RE, Hamud F, Fiskum G, Varghese PJ, Sharpe S. Cerebral ischemia and reperfusion: prevention of brain mitochondrial injury by lidoﬂazine. J Cereb Blood Flow Metab. 1987;7:752–758. 8. Brustovetsky N, Jemmerson R, Dubinsky JM. Calcium-induced cytochrome c release from rat brain mitochondria is altered by digitonin. Neurosci Lett. 2002;332:91–94. 9. Ehrenberg B, Montana V, Wei MD, Wuskell JP, Loew LM. Membrane potential can be determined in individual cells from the nernstian distribution of cationic dyes. Biophys J. 1988;53:785–794. 10. Emaus RK, Grunwald R, Lemasters JJ. Rhodamine 123 as a probe of transmembrane potential in isolated rat-liver mitochondria: spectral and metabolic properties. Biochim Biophys Acta. 1986;850:436–448. 11. Johnson LV, Walsh ML, Bockus BJ, Chen LB. Monitoring of relative mitochondrial membrane potential in living cells by ﬂuorescence microscopy. J Cell Biol. 1981;88:526–535. 12. Akerman KE, Wikstrom MK. Safranine as a probe of the mitochondrial membrane potential. FEBS Lett. 1976;68:191–197. 13. Scaduto RC, Jr.. Grotyohann LW. Measurement of mitochondrial membrane potential using ﬂuorescent rhodamine derivatives. Biophys J. 1999;76:469–477. 14. Blattner JR, He L, Lemasters JJ. Screening assays for the mitochondrial permeability transition using a ﬂuorescence multiwell plate reader. Anal Biochem. 2001;295:220–226. 15. Rottenberg H, Wu S. Quantitative assay by ﬂow cytometry of the mitochondrial membrane potential in intact cells. Biochim Biophys Acta. 1998;1404:393–404. 16. Rottenberg H, Wu S. Mitochondrial dysfunction in lymphocytes from old mice: enhanced activation of the permeability transition. Biochem Biophys Res Commun. 1997;240:68–74. 17. Nieminen AL, Gores GJ, Dawson TL, Herman B, Lemasters JJ. Toxic injury from mercuric chloride in rat hepatocytes. J Biol Chem. 1990;265:2399–2408. 18. Nicholls DG. Simultaneous monitoring of ionophore- and inhibitor-mediated plasma and mitochondrial membrane potential changes in cultured neurons. J Biol Chem. 2006;281:14864–14874. 19. Johnson LV, Walsh ML, Chen LB. Localization of mitochondria in living cells with rhodamine 123. Proc Natl Acad Sci U S A. 1980;77:990–994. 20. Heiskanen KM, Bhat MB, Wang HW, Ma J, Nieminen AL. Mitochondrial depolarization accompanies cytochrome c release during apoptosis in PC6 cells. J Biol Chem. 1999;274:5654–5658. 21. Reers M, Smith TW, Chen LB. J-aggregate formation of a carbocyanine as a quantitative ﬂuorescent indicator of membrane potential. Biochemistry. 1991;30:4480–4486.

430

USE OF FLUORESCENT REPORTERS

22. Reers M, Smiley ST, Mottola-Hartshorn C, Chen A, Lin M, Chen LB. Mitochondrial membrane potential monitored by JC-1 dye. Methods Enzymol. 1995;260:406–417. 23. Ankarcrona M, Dypbukt JM, Bonfoco E, et al. Glutamate-induced neuronal death: a succession of necrosis or apoptosis depending on mitochondrial function. Neuron. 1995;15:961–973. 24. White RJ, Reynolds IJ. Mitochondrial depolarization in glutamate-stimulated neurons: an early signal speciﬁc to excitotoxin exposure. J Neurosci. 1996;16:5688–5697. 25. Scanlon JM, Reynolds IJ. Effects of oxidants and glutamate receptor activation on mitochondrial membrane potential in rat forebrain neurons. J Neurochem. 1998;71:2392–2400. 26. Macho A, Decaudin D, Castedo M, et al. Chloromethyl-X-Rosamine is an aldehydeﬁxable potential-sensitive ﬂuorochrome for the detection of early apoptosis. Cytometry. 1996;25:333–340. 27. Elmore SP, Qian T, Grissom SF, Lemasters JJ. The mitochondrial permeability transition initiates autophagy in rat hepatocytes. FASEB J. 2001;15:2286–2287. 28. Chacon E, Reece JM, Nieminen AL, Zahrebelski G, Herman B, Lemasters JJ. Distribution of electrical potential, pH, free Ca2+ , and volume inside cultured adult rabbit cardiac myocytes during chemical hypoxia: a multiparameter digitized confocal microscopic study. Biophys J. 1994;66:942–952. 29. Hunter DR, Haworth RA, Southard JH. Relationship between conﬁguration, function, and permeability in calcium-treated mitochondria. J Biol Chem. 1976;251:5069–5077. 30. Halestrap AP, Brennerb C. The adenine nucleotide translocase: a central component of the mitochondrial permeability transition pore and key player in cell death. Curr Med Chem. 2003;10:1507–1525. 31. Crompton M, Virji S, Doyle V, Johnson N, Ward JM. The mitochondrial permeability transition pore. Biochem Soc Symp. 1999;66:167–179. 32. Kokoszka JE, Waymire KG, Levy SE, et al. The ADP/ATP translocator is not essential for the mitochondrial permeability transition pore. Nature. 2004;427:461–465. 33. Basso E, Fante L, Fowlkes J, Petronilli V, Forte MA, Bernardi P. Properties of the permeability transition pore in mitochondria devoid of cyclophilin D. J Biol Chem. 2005;280:18558–18561. 34. Baines CP, Kaiser RA, Sheiko T, Craigen WJ, Molkentin JD. Voltage-dependent anion channels are dispensable for mitochondrial-dependent cell death. Nat Cell Biol. 2007;9:550–555. 35. He L, Lemasters JJ. Regulated and unregulated mitochondrial permeability transition pores: a new paradigm of pore structure and function? FEBS Lett. 2002;512:1–7. 36. Kantrow SP, Tatro LG, Piantadosi CA. Oxidative stress and adenine nucleotide control of mitochondrial permeability transition. Free Radic Biol Med. 2000;28:251–260. 37. Zamzami N, Susin SA, Marchetti P, et al. Mitochondrial control of nuclear apoptosis. J Exp Med. 1996;183:1533–1544. 38. Kim JS, He L, Lemasters JJ. Mitochondrial permeability transition: a common pathway to necrosis and apoptosis. Biochem Biophys Res Commun. 2003;304:463–470. 39. Lemasters JJ. Dying a thousand deaths: redundant pathways from different organelles to apoptosis and necrosis. Gastroenterology. 2005;129:351–360. 40. Nieminen AL, Saylor AK, Tesfai SA, Herman B, Lemasters JJ. Contribution of the mitochondrial permeability transition to lethal injury after exposure of hepatocytes to t -butylhydroperoxide. Biochem J. 1995;307 (pt 1): 99–106.

REFERENCES

431

41. Ohata H, Chacon E, Tesfai SA, Harper IS, Herman B, Lemasters JJ. Mitochondrial Ca2+ transients in cardiac myocytes during the excitation-contraction cycle: effects of pacing and hormonal stimulation. J Bioenerg Biomembr. 1998;30:207–222. 42. Lemasters JJ, Nieminen AL, Qian T, et al. The mitochondrial permeability transition in cell death: a common mechanism in necrosis, apoptosis and autophagy. Biochim Biophys Acta. 1998;1366:177–196. 43. Wieder ED, Hang H, Fox MH. Measurement of intracellular pH using ﬂow cytometry with carboxy-SNARF-1. Cytometry. 1993;14:916–921. 44. Trollinger DR, Cascio WE, Lemasters JJ. Selective loading of Rhod 2 into mitochondria shows mitochondrial Ca2+ transients during the contractile cycle in adult rabbit cardiac myocytes. Biochem Biophys Res Commun. 1997;236:738–742. 45. Kim JS, Jin Y, Lemasters JJ. Reactive oxygen species, but not Ca2+ overloading, trigger pH- and mitochondrial permeability transition-dependent death of adult rat myocytes after ischemia-reperfusion. Am J Physiol Heart Circ Physiol. 2006;290:H2024–H2034. 46. Petronilli V, Miotto G, Canton M, et al. Transient and long-lasting openings of the mitochondrial permeability transition pore can be monitored directly in intact cells by changes in mitochondrial calcein ﬂuorescence. Biophys J. 1999;76:725–734. 47. Lam M, Oleinick NL, Nieminen AL. Photodynamic therapy-induced apoptosis in epidermoid carcinoma cells: reactive oxygen species and mitochondrial inner membrane permeabilization. J Biol Chem. 2001;276:47379–47386. 48. Zahrebelski G, Nieminen AL, al Ghoul K, Qian T, Herman B, Lemasters JJ. Progression of subcellular changes during chemical hypoxia to cultured rat hepatocytes: a laser scanning confocal microscopic study. Hepatology. 1995;21:1361–1372. 49. Schneider WC. Intracellular distribution of enzymes: III. The oxidation of octanoic acid by rat liver fractions. J Biol Chem. 1948;176:259–266. 50. Lemasters JJ, Hackenbrock CR. The energized state of rat liver mitochondria. ATP equivalence, uncoupler sensitivity, and decay kinetics. J Biol Chem. 1980;255:5674–5680.

18 COMPARTMENTATION OF REDOX SIGNALING AND CONTROL: DISCRIMINATION OF OXIDATIVE STRESS IN MITOCHONDRIA, CYTOPLASM, NUCLEI, AND ENDOPLASMIC RETICULUM Patrick J. Halvey Department of Pediatrics, and Division of Pulmonary, Allergy and Critical Care Medicine, Department of Medicine, Emory University, Atlanta, Georgia; Department of Biochemistry and National Centre for Biomedical Engineering Science, National University of Ireland, Galway, Ireland

Jason M. Hansen Department of Pediatrics, Emory University, Atlanta, Georgia

Lawrence H. Lash Department of Pharmacology, Wayne State University School of Medicine, Detroit, Michigan

Dean P. Jones Division of Pulmonary, Allergy and Critical Care Medicine, Department of Medicine, Emory University, Atlanta, Georgia

1. 2. 3. 4. 5.

Introduction Mitochondrial oxidative stress and protective thiol–disulﬁde systems Sources of ROS in the mitochondria Mitochondrial GSH and GSH-dependent systems Mitochondrial thioredoxin and peroxiredoxin systems

434 436 436 439 441

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

433

434

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

6. Non-thiol-based mitochondrial antioxidants 7. Proteomic techniques for identiﬁcation of oxidatively modiﬁed proteins 8. Fluorescent probes for detection of mitochondrial ROS 9. Compartmentation in redox signaling and control in the nucleus and cytoplasm 10. Oxidative protein folding and redox signaling events in the endoplasmic reticulum 11. Conclusions and future perspectives

443 444 445 447 450 452

1. INTRODUCTION Mitochondria are a major source of reactive oxygen species (ROS), formed via autoxidation of components in the electron transport system in the inner mitochondrial membrane. Increased mitochondrial ROS, caused by electron leakage or disruption of electron transport, results in oxidative stress and is a salient feature of numerous pathological conditions [1–3]. A large number of mitochondrial enzymes, including dehydrogenases and transport ATPases, contain redox-sensitive sulfhydryl groups that must be maintained in the reduced state for proper function [4,5]. Furthermore, redox-sensitive components in the electron transport chain, such as iron in heme-prosthetic groups and iron–sulfur centers, are susceptible to oxidation which can lead to mitochondrial dysfunction [6]. Given the capacity of mitochondrial ROS to interfere with these components and to initiate degradative processes, such as DNA oxidation and lipid peroxidation, it is not surprising that mitochondria possess antioxidant systems that act as a counterbalance. Several NADPH-dependent thiol–disulﬁde redox couples, including glutathione and glutathione disulﬁde (GSH/GSSG), thioredoxins (Trx), and peroxiredoxins (Prx), play a crucial part in elimination of mitochondrial ROS [7]. An increase in the relative amounts of these antioxidants confers protection against both xenobiotically and pathologically induced oxidative stress [8,9]. In addition to their role in free-radical scavenging, there is increasing evidence that thiol–disulﬁde redox couples are key elements of low-ﬂux redox circuits that function in cell signaling and metabolic control [7]. These “sulfur switches” consist of cysteinyl groups that undergo S-thiylation with GSH, cysteine, or other thiol-containing proteins, as well as proximal cysteinyl groups capable of undergoing intraprotein disulﬁde formation. Recent investigations of the speciﬁcity of redox signaling have focused on the oxidation of thiols, presumed targets and mediators of ROS signals. Accumulating evidence suggests that differential responses in the nucleus, cytoplasm, and mitochondria form a basis for selectivity in redox signaling mechanisms,

435

INTRODUCTION

reﬂecting the distinct functions and properties of each organelle [10–12]. Methods are now available for the simultaneous measurement of the redox state of subcellular thiol–disulﬁde redox pools, such as thioredoxin 1 (Trx1) (nuclear and cytoplasmic), thioredoxin-2 (Trx2) (mitochondrial), and cellular glutathione (GSH/GSSG), and can be used to map redox signaling events at a subcellular level [10,13] (see Figure 1). Thioredoxins are a family of oxidoreductases that function primarily as protein reductants. GSH is the most abundant intracellular, nonprotein biothiol, acting as the principal redox buffer in numerous redox-dependent reactions such as ROS scavenging and xenobiotic detoxiﬁcation [14]. Both Trx and GSH are ultimately dependent on the reducing power of NADPH/NADP+ , which is maintained metabolically. Studies of cellular ROS and thiol–disulﬁde components have focused largely on whole cell measurements. However, increasing evidence points to the importance of compartmentation of redox signaling and oxidative stress. For instance, the unfolded protein response (UPR) in the endoplasmic reticulum (ER) results in ROS-dependent Ca2+ release from the ER, which causes mitochondrial toxicity without exposure of mitochondria to the ROS. Tumor necrosis factor–alpha (TNF-α) binds to cell-surface receptors, inducing rapid mitochondrial ROS production followed by oxidation of mitochondrial Trx2 without oxidation of cytoplasmic Trx1 (see Figure 1). NF-κB is activated in cytoplasm by ROS, but DNA binding in the nucleus is blocked by ROS [15] (see Figure 4). Both GSH and Trx1 systems block activation in the cytoplasm but support transcriptional activation in the nucleus [7]. Here we review and discuss progress in understanding mitochondrial thiol– disulﬁde systems in the context of subcellular compartmentation. Although knowledge is incomplete, newly available and emerging technologies create

Oxidized Reduced

Trx1 −

+ EGF

−

+

TNF-a Reduced Oxidized

Trx2 −

+ EGF

−

+

TNF-a

Figure 1 Differential redox signaling patterns in mitochondria and cytoplasm. Redox Western blot showing changes in thioredoxin-1 (Trx1) (cytoplasmic) (top) and thioredoxin-2 (Trx2) (mitochondrial) (bottom) redox states in response to epidermal growth factor (EGF) (5 minutes) treatment as described by Halvey et al. [10] and TNF-α (5 minutes) treatment as described by Hansen et al. [12]. For Trx1 the top band represents the oxidized one-disulﬁde form, while the lower band represents the reduced dithiol form. For Trx2 the top band represents the reduced dithiol form and the lower band represents the oxidized one-disulﬁde form. (Reproduced in modiﬁed form from Halvey et al. [10] and Hansen et al. [12], with permission from the publisher.)

436

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

the possibility of developing a detailed compartmental map of redox signaling and control pathways, thereby clarifying the mechanisms of disruption of mitochondrial redox circuitry during oxidative stress. Importantly, a picture is beginning to emerge that antioxidant systems have coevolved with organellar functions, with mitochondria being the most highly reducing and most susceptible to oxidative damage, nuclei being next most reducing and most resistant to oxidative changes, cytoplasm undergoing variation in redox state depending on cell function, and endoplasmic reticulum being relatively oxidizing to aid in the production of secreted proteins.

2. MITOCHONDRIAL OXIDATIVE STRESS AND PROTECTIVE THIOL–DISULFIDE SYSTEMS Mitochondria evolved as a speciﬁc subcellular compartment with very high oxidative activity. Their specialized role in electron transport and oxidative phosphorylation requires a specialized redox environment and redox circuitry distinct from other subcellular compartments, as outlined in Figure 2. Mitochondrial pH is relatively more alkaline (∼pH 7.8), thereby contributing to a more reducing microenvironment from increased electron-donating tendency of hydroxylates and thiolates relative to hydroxyl and thiol groups. In the presence of respiratory inhibitors, or defects in the mitochondrial electron transfer chain, increased superoxide anion radical and hydrogen peroxide are generated. In the mitochondria, these reactive species are detoxiﬁed by the antoxidant systems, which also control signaling of apoptosis and necrosis. Thus, there is a close communication between the control of oxidative processes that produce ATP and those that signal the oxidative stress response and activation of cell death mechanisms.

3. SOURCES OF ROS IN THE MITOCHONDRIA Mitochondrial ROS (principally, O2 −· ) are derived from several sources. Complex I (NADH–quinone oxidoreductase) and complex III (coenzyme Q–cytochrome c oxidoreuctase) are the main sites of electron leakage in mitochondrial electron transport chains. Inhibitors of complex I (rotenone) and complex III (antimycin A) cause increased ROS generation at these sites [16,17]. In the case of complex I, both ﬂavin mononucleotide (FMN) and Fe–S cluster components are involved in ROS generation [18,19]. The coenzyme Q analog idebenone, which is used to treat mitochondrial cytopathies, enhances Fe–S cluster-mediated O2 −· production at complex I [20]. In the case of complex III, ubisemiquinone is the primary electron donor. O2 −· can be released into the intermembrane space and the mitochondrial matrix via Qo (oriented toward the intermembrenae space) and Qi (oriented toward the mitochondrial matrix), respectively [21,22]. Qo -derived ROS is not subject to mitochondrial antioxidant systems. Electrons can be transferred to complex III

437

SOURCES OF ROS IN THE MITOCHONDRIA Electron Flow H+

Mitochondrial Membrane III

I NADH

CoQ

III

FADH2 FAD NAD+ O2•H2O2

Prx3-(SH)2 MPT

SOD2

Prx3(-S-S-)

Trx2-(SH)2

Trx2(-S-S-)

ASK1 TR2 NADPH NADP+

CytC

IV

ATP Synthase O2

H+ ADP Disruption of ETC Electron Leakage ATP DIC H2O2 Gpx1,4 GSH GSSG

Pi2−

GS− OGC Grx2 Pr-SSG

2-OG2−

GR

Complex I Cardiolipin NADPH NADP+ Isocitrate dehydrogenase

Figure 2 Detoxiﬁcation of mitochondrial ROS by thiol–disulﬁde redox-dependent systems. NADH/FADH2 -dependent electron transport chains (ETC) in the mitochondrial inner membrane generate a proton (H+ ) gradient across the membrane which drives ATP synthesis. Superoxide (O2 −· ) is produced as a result of disruption to ETC or electron leakage and is converted to H2 O2 via superoxide dismutase 2 (SOD2). Mitochondrial H2 O2 is eliminated by NADPH-dependent systems, including peroxiredoxin-3 (Prx3), glutathione peroxidase-1 (Gpx1), and glutathione peroxidase-4 (Gpx4). Prx3 is reduced to its functional from by thioredoxin 2 (Trx2), which is itself reduced by thioredoxin reductase 2 (TR2). Oxidation of Trx2 results in activation of membrane permeability transition (MPT) pore and the apoptosis signal-regulating kinase-1 (ASK1). Glutathione (GSH) acts as reductant source for Gpx3 and is transported into mitochonrdria by the dicarboxylate carrier (DIC) or the 2-oxoglutarate carrier (OGC) as a thiolate anion (GS− ). Glutathione disulﬁde (GSSG) is reduced to GSH by the enzyme glutathione reductase (GR) or may participate in protein glutathionylation (Pr-SSG) via glutaredoxin 2 (Grx2).

from the β-oxidation pathway via electron transfer ﬂavoprotein (ETF)/electron transfer ﬂavoprotein–quinone oxidoreductase (ETF-QOR)[23]. ETC/ETF-QOR can accept electrons from several dehydrogenases in the mitochondrial matrix, including acyl-CoA dehydrogenase. Electrons are transferred from the FADH2 prosthetic group of the reduced acyl-CoA dehydrogenase to ETF, and are in turn transferred to ETF-QOR. Ubiquinone is thereby reduced to ubiquinol, and electrons are donated to complex III [23]. Other mitochondrial dehydrogenases, including glycerol-3-phosphate dehydrogenase (G3PDH) and α-ketoglutarate dehydrogenase (KGDH), have been shown to be involved in mitochondrial ROS generation [24,25]. G3PDH donates electrons to ubiquinone, leading to O2 −· release into the mitochondrial matrix and subsequent H2 O2 production via MnSOD [24]. The growth factor adaptor p66Shc is an important mediator of oxidative stress–induced apoptosis and induces mitochondrial H2 O2 production via interaction with cytochrome, thereby participating

438

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

in apoptotic redox signaling [26]. Others have shown that monoamine oxidase contributes to steady-state concentrations of H2 O2 in mitochondria [27]. The biochemistry and pathology of ROS is confounded by production of nitric oxide (NO) by an endogenous mitochondrial isform of NO synthase (mNOS). Local NO modulates complex IV activity by binding reversibly to the heme in competition with O2 . Superoxide, derived from mitochondrial electron transport chains (ETCs), and NO, derived from mNOS, may act as signaling molecules, indicating that ETC is competent for the generation of protonmotive force [28]. However, disruption of ETC by toxicological and pharmacological agents is more likely to result in oxidative damage. O2 −· reacts readily with NO· to from the peroxynitrite anion (ONOO− ), a more reactive species than either NO· or O2 −· , and a potent oxidizing agent that can cause DNA damage and lipid oxidation [29,30]. Several drugs cause mitochondrial toxicity by oxidative mechanisms. The widely used anticancer drug doxorubicin has been studied extensively [31] and causes a dose-limiting cardiotoxicity and mitochondrial DNA damage due to enhanced ROS generation (see Chapter 6). Doxorubicin, or one of its metabolites, is reduced by complex I, which can either react with DNA or protein or activate ROS generation, causing oxidative damage [32]. Toxic overdose concentrations of acetaminophen also cause mitochondrial toxicity, apparently mediated through the reactive intermediate N -acetyl-p-benzoquinonimine (NAPQI) [33]. NAPQI binds to proteins of the mitochondria electron transport chain, resulting in generation of superoxide and peroxynitrite [34]. Recent genomic and proteomic analyses have revealed rapid modiﬁcation of a number of mitochondrial proteins preceding functional changes that affect β-oxidation and ATP production [35]. Considerable recent interest has focused on thiazolidinediones (glitazones), synthetic agonists of the transcription factor peroxisome proliferator-activated receptor-γ (PPARγ). Rosiglitazone and pioglitazone are used for treatment of type 2 diabetes because they have insulin-sensitizing effects; glitazones have a range of effects on vascular functions, inﬂammation, cell proliferation, and cell death. Thus, individual chemicals in this class differ in characteristic responses and are often cell speciﬁc. Cigliazone and rosiglitazone were found to be preferentially toxic to astoglioma cells compared to primary rat astrocytes [36]. Toxicity was linked to a rapid production of ROS and mitochondrial membrane depolarization. Recent research shows that the glitazones induce superoxide anion, hydrogen peroxide and peroxynitrite because of inhibition of complex I [37]. Ciglitazone, rosiglitazone, and pioglitazone each had this effect, but ciglitazone was most active. Glitazones have also been found to affect mitochondria in Jurkat [38,39], Raji [38], multiple myeloma [40], and hepatoma cells [41]. Thus, enhanced mitochondrial ROS generation and related toxicity appear to be common features of this important class of drugs. Glitazones can also elicit toxic responses by PPARγ-related activities, such as induction of pro-apoptotic praline oxidase, which localizes to the mitochondria by troglitazone, while subtoxic doses can elicit cytoprotective responses [42]. In contrast, rosiglitazone protects human neuroblastoma SH-SY5Y cells by inducing

MITOCHONDRIAL GSH AND GSH-DEPENDENT SYSTEMS

439

expression of superoxide dismutase and catalase and regulating the expression of Bcl-2 and Bax [43]. These compensatory adaptations can occur as a consequence of the same inhibitory processes that cause toxicity because at low levels of ROS generation, ROS signals increased expression of protective enzymes. An understanding of the controls of these opposing beneﬁcial and toxic responses will provide an important clariﬁcation of the pathophysiology of these drugs [44].

4. MITOCHONDRIAL GSH AND GSH-DEPENDENT SYSTEMS Several lines of evidence have established that mitochondrial GSH/GSSG is a distinct subcellular redox pool. Cytoplasmic GSH is selectively depleted by diethylmaleate (DEM), whereas mitochondrial GSH is depleted selectively by (R,S )-3-hydroxy-4-pentenoate (3-HP) [45,46]. In hepatocytes, 15% of cell volume is occupied by mitochondria, and approximately 15% of the total GSH is in the mitochondrial pool. Estimation of mitochondrial GSH/GSSG redox state is technically challenging due to loss and/or oxidation of GSH/GSSG during isolation of mitochondria. Based on available data and assuming a pH of 7.8, mitochondrial GSH/GSSG is about −280 mV and becomes oxidized to −235 mV during apoptosis [47,48]. Mitochondria lack the enzymes needed to synthesize GSH (i.e., glutamate– cysteine ligase and glutathione synthase) and must rely on de novo GSH synthesis in the cytosol and subsequent transport into the mitochondria [49–53]. However, mitochondria do contain several GSH-dependent enzymes, including glutathione peroxidase-1 and -4 (Gpx1, Gpx4), glutathione reductase (GR), and glutaredoxin-2 (Grx2) [54,55]. Studies involving Gpx1 knockout mice demonstrate increased susceptibility to mitochondrial oxidative stress [56]. Gpx4 (−/−) knockouts are embryonically lethal at gestational day 7.5 [57]. Cells isolated from Gpx4-deﬁcient mice (+/−) showed an increased susceptibility to inducers of oxidative stress, including peroxides, gamma radiation, and paraquat [57]. Speciﬁc to mitochondrial function, oxidative stress has been shown to decrease ATP production, whereas GSH-Px4 transgenic mice treated with diquat showed a resistance to mitochondrial dysfunction and maintenance of the membrane potential [58]. The mitochondrial form of glutaredoxin, Grx2, is a very efﬁcient catalyst of monothiol reactions because of its high afﬁnity for protein glutathione-mixed disulﬁdes. Unlike glutaredoxin-1(Grx1), its cytoplasmic homolog, Grx2 is not inhibited by oxidation of structural cysteine residues [55]. Grx2 is an efﬁcient catalyst of both glutathionylation and deglutathionylation of mitochondrial membrane proteins [59]. Complex I is glutathionylated by GSSG in the presence of Grx2, leading to loss of activity [60]. Other studies have found that isocitrate dehydrogenase, an enzyme of the citric acid cycle, undergoes Grx2-dependent deglutathionylation, thereby regulating its enzymatic activity during oxidative stress [61]. Overexpression of Grx2 in HeLa cells inhibited cardiolipin loss

440

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

and prevented apoptosis induced by doxorubicin [62]. Thus, Grx2 has a central role in GSH-dependent redox regulation in mitochondria and suggests that S-glutathionylation may provide a marker for mitochondrial oxidative stress. The mitochondrial GSH/GSSG couple is involved in several pathological and toxicological conditions. Alcoholic liver disease (ALD) is associated with the loss of mitochondrial GSH/GSSG pool by up to 60%, as demonstrated in isolated hepatocytes from ethanol-fed rats [63]. Radiolabeling experiments show that cytosolic transport of GSH into the mitochondria is decreased by approximately 35%, an effect that mostly accounts for mitochondrial GSH depletion with ethanol exposure [64]. The neurotoxin and Parkinson’s disease mimetic, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), selectively inhibits complex I of the mitochondrial electron transport chain, leading to ROS production [65]. Studies reveal that depleting GSH with drugs, such as buthionine sulfoximine (BSO), exacerbates MPTP toxicity in the mouse [66]. Conversely, increasing brain GSH attenuates MPTP neurotoxicity [67]. Furthermore, MPTP treatment depletes brainstem GSH, which is prevented by antioxidant pretreatment [68]. Mice deﬁcient in glutathione peroxidase similarly exhibit increased vulnerability to MPTP. Similarly, the effects of the complex I inhibitor rotenone in cultured cells can be attenuated by addition of exogenous GSH [69]. Acetaminophen, a common analgesic and antipyretic drug, preferentially depletes mitochondrial GSH compared to that in the cytosol, suggesting a possible role for mitochondrial GSH depletion in acetaminophen toxicity [70]. Others have demonstrated that chronic ethanol exposure in rodents results in a selective depletion of mitochondrial GSH, which increases acetaminophen-induced hepatotoxicity [71]. Much attention has focused on the role of GSH/GSSG in mitochondriamediated apoptosis. Oxidation of GSH/GSSG is associated with apoptosis, although oxidation of mitochondrial GSH/GSSG occurs secondary to cytochrome c release following activation of the intrinsic pathway [72]. Nonetheless, a variety of studies support the crucial role played by GSH in mitochondria as an activating event in apoptotic cell death. GSH depletion by BSO or DEM resulted in overproduction of endogenous ROS from complex III and apoptosis of HL-60 cells [13]. Interestingly, this event was mediated by ROS-induced mitochondrial membrane permeabilization, suggesting a role for GSH in controlling the endogenous generation of ROS. Resistance to peroxide-induced apoptosis in mutant cystic ﬁbrosis transmembrane conductance regulator (CFTR)–expressing cells was found to be dependent on GSH-mediated activation of BAX [73]. Furthermore, mitochondrial GSH/GSSG may have an important role in other types of toxicant-induced apoptosis. For instance, treatment with the cross-linking agent cisplatin or the alkylating agent mephalan caused peroxide accumulation and apoptosis in U-937 human promonocytic cells, both of which were potentiated by GSH depletion [74]. GSH transport from the cytoplasm into the mitochondria is a key determinant of mitochondrial GSH status [75,76,50] and is of importance in pathologies associated with mitochondrial GSH depletion. Due to mitochondrial membrane potential, the matrix space is negative relative to the cytoplasm, meaning that

MITOCHONDRIAL THIOREDOXIN AND PEROXIREDOXIN SYSTEMS

441

negatively charged GSH must be transported actively or in exchange for another anion. The relatively alkaline environment of the mitochondria (pH 7.8) effectively lowers the pK a value of the thiol group so that a much higher proportion of GSH molecules are in the thiolate form. Transport is stimulated by energizing mitochondria and inhibited by excessive glutamate or disruption of the protonmotive force with proton ionophores. Transport of GSH into mitochondria is mediated primarily by two carriers, the dicarboxylate carrier (DIC; Slc25a10 ) and 2-oxoglutarate carrier (OGC; Slc25a11 ). These anion-exchange systems indirectly use the energy of the membrane potential to drive the uptake of the anionic GSH against the prevailing potential (negative inside). While DIC and OGC appear to account for virtually all of the activity of mitochondrial GSH transport in renal cortical mitochondria [75,76], the situation appears to be somewhat different in hepatic mitochondria, where at least one additional carrier besides the DIC and OGC plays a quantitatively signiﬁcant role in the transport of GSH into mitochondria [76a]. Overexpression of DIC in NRK-52E cells caused increased mitochondrial uptake and accumulation of GSH and protection from chemically induced apoptosis [51]. Similarly, OGC overexpression resulted in a 10- to 20-fold increase in GSH uptake activity and protection from oxidant-induced apoptosis [77]. The ability to modulate the level and activity of mitochondrial GSH carriers has two important implications. First, alterations in levels of dicarboxylates (e.g., 2-oxoglutarate, malate) due to nutritional variations can affect transport function. Second, the DIC and OGC may be potential therapeutic targets to treat some of the numerous diseases and pathological states that involve mitochondrial oxidative stress as part of their mechanism of action. The presence of speciﬁc carriers for GSH on the mitochondrial inner membrane and the prevailing membrane potential and pH gradient from cytoplasm to mitochondrial matrix also serve to physically and functionally segregate the two pools of GSH. There are numerous examples of chemical agents that selectively deplete or oxidize one GSH pool, leaving the other pool unaffected. For example, the nephrotoxic metabolite of the environmental contaminant trichloroethylene, S -(1,2-dichlorovinyl)-l-cysteine (DCVC), selectively oxidizes mitochondrial GSH and not cytoplasmic GSH [78].

5. MITOCHONDRIAL THIOREDOXIN AND PEROXIREDOXIN SYSTEMS The other major thiol–disulﬁde antioxidant system in the mitochondria is dependent on thioredoxin-2 (Trx2). Trx2 is maintained in the reduced form by electron transfer from NADPH, catalyzed by thioredoxin reductase-2 (TrxR2) (analogous to the cytosolic Trx1 system). Trx2 supplies electrons to peroxiredoxin-3 (Prx3), a mitochondrial peroxidase, which reduces H2 O2 to water [79]. Indeed, a primary function of Trx2 appears to be the detoxiﬁcation of ROS through Prx3. Several studies have outlined the important role played by Trx2 during oxidative stress.

442

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

Stable overexpression of human Trx2 in 143B osteosarcoma cells and SH-SY5Y human neuroblastoma cells protected cells from oxidant-induced apoptosis [8,80]. Knocking out Trx2 in DT40 chicken B cells caused spontaneous apoptosis and increased production of ROS [81]. Cytochrome c release from mitochondria and caspase-9 and caspase-3 activation were observed in Trx2 knockout cells. Overexpression of Trx2 or an active site mutant of Trx2 reduced these effects, suggesting that Trx2 regulates mitochondrial outer membrane permeabilization and apoptosis by redox-active site cysteine-independent mechanisms [82,83]. Trx2 has also been shown to associate with apoptosis-signaling kinase-1 (ASK1) [84]. Oxidation of Trx2 stimulates release of ASK1, allowing ASK1-dependent apoptosis to occur. Similarly, Prx3 has been implicated as a mitochondrion-speciﬁc component of the oxidative stress response. Depletion of Prx3 in HeLa cells by RNA interference, resulted in increased intracellular levels of H2 O2 and sensitized cells to induction of apoptosis by staurosporine or TNF-α [85]. Apoptotic events such as mitochondrial membrane potential collapse, cytochrome c release, and caspase activation were increased in Prx3-depleted cells, and these effects were reversed by ectopic expression of Prx3 [85]. Stable overexpression of Prx3 protected cells against apoptosis caused by H2 O2 or t-butyl hydroperoxide (tBH) [9]. Since Trx2 is speciﬁc to the mitochondria and is regulated independently, it is an ideal candidate for the selective determination of redox state within the mitochondria. The use of antibodies speciﬁc to Trx2 allows measurement of mitochondrial redox state by the redox Western blot method and eliminates the need for subcellular fractionation [7]. Using the thiol-reactive agent 4-acetamido-4 -maleimidylstilbene-2,2 -disulfonic acid (AMS), one can distinguish between reduced and oxidized forms of Trx2 [83]. Samples are separated by nonreducing SDS-PAGE. Trx2 derivatized with AMS has a greater molecular weight (approximately 0.5 kDa/thiol) and migrates through the gel more slowly than does oxidized/underivatized Trx2 [10,83]. This provides a means to compare redox state in the mitochondria with other subcellular compartments. Results with the redox Western blot for Trx2 show that tBH oxidizes Trx2 preferentially compared to Trx1 [80]. Chemical toxicants capable of inducing Parkinson-like symptoms via inhibition of complex I show different effects on Trx1 and Trx2; paraquat oxidizes cytosolic Trx1, whereas MPTP and rotenone oxidize Trx2 in neuroblastoma cells [86]. Similarly, the toxic metals arsenic, cadmium, and mercury showed little oxidation of GSH, but oxidized both Trx1 and Trx2 signiﬁcantly [87]. Interestingly, these toxicants caused greater oxidation of Trx2 (>60 mV) than did the cytoplasmic Trx1 (20 to 40 mV). ASK1 activation and cell death were observed only with metals that oxidized thioredoxins but not with metals that oxidized GSH, suggesting that thioredoxins may convey different levels of control in apoptotic and toxic signaling pathways. In addition, studies on compartmental redox signaling revealed that Trx2 is not oxidized by epidermal growth factor (EGF) signaling [10] but is oxidized during TNF-α treatment [12]. In the case of TNF-α signaling, overexpression of Trx2, but not Trx1, decreased

NON-THIOL-BASED MITOCHONDRIAL ANTIOXIDANTS

443

ROS generation, suggesting mitochondrial compartmentation of ROS production and subsequent speciﬁc detoxiﬁcation by Trx2, not Trx1 [12]. Evaluation of Trx2 redox state by redox Western blot may provide new insights into drug-induced alterations to mitochondrial homeostasis. Statins (3-hydroxy-3-methylglutaryl-CoA reductase inhibitors) are a family of compounds used in the treatment of hypercholesterolemia and cardiovascular disease that function in part by exerting mitochondria-mediated cardioprotective effects [88]. Simvastatin attenuates oxidant-mediated mitochondrial membrane depolarization [89], which has shown to be dependent on activation of mitochondrial KATP channels. The opening of these channels serves to depolarize the mitochondrial membrane, maintain the integrity of the mitochondrial matrix, and decrease ROS generation by the mitochondria following ischemia and reperfusion [90]. However, adverse reactions to statins involving skeletal muscle have also been reported [91]. Simvastatin can trigger a large release of Ca2+ from the sarcoplasmic reticulum (SR), resulting in alterations in mitochondrial function in human skeletal muscle cells [92] (see Chapter 7). Complex I of respiratory chain appears to be targeted speciﬁcally by simvastatin, representing a critical component of statin-induced toxicity in muscle cells [93]. This effect appears preferentially in fast-type skeletal muscle ﬁbers but not in intact heart cells that lack the monocarboxylate transporter (MCT4), necessary for statin uptake [93]. It is likely that the interaction of simvastatin with mitochondrial electron transport results in alterations to mitochondrial redox homeostasis. Therefore, assessment of mitochondrial redox state by Trx2 redox Western will enhance our understanding of the divergent effects of statins on mitochondrial function in heart and muscle cells.

6. NON-THIOL-BASED MITOCHONDRIAL ANTIOXIDANTS While there has been intense study of thiol-based antioxidant systems in the mitochondria, other important antioxidant systems aid in ROS elimination. Superoxide dismutase (SOD) is an enzyme that catalyzes the conversion of superoxide anion to hydrogen peroxide. Interestingly, there are two isoforms of SOD, a copper- and zinc-containing form found primarily in the cytosol (CuZnSOD or SOD1) and a manganese-containing form found in the mitochondria (MnSOD or SOD2). Despite the modest reactivity of superoxide anion in aqueous environments, excessive superoxide anion can give rise to more reactive, toxic ROS. MnSOD is an enzyme that begins mitochondrial ROS removal by converting superoxide anion to hydrogen peroxide. The importance of MnSOD has been demonstrated in a knockout mouse model where homozygous animals die within 2 weeks of birth [94,95]. Observed injuries include an increase in mtDNA oxidation and a general decrease in mitochondrial enzyme function [96,95] (see Chapter 24). SOD mimetics targeted to the mitochondria show an enhanced resistance to peroxide-induced cytochrome c release and caspase-3 activation [97]. While it appears that the primary function of MnSOD is ROS removal, there

444

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

may be other physiological functions. Studies show that mitochondrial oxidants are responsible for the activation of JNK, which can lead to the inhibition of cytosolic enzymes such as glycogen synthase kinase 3b and glycogen synthase [98]. Therefore, MnSOD may play a role in regulating extramitochondrial events. As a consequence of toxic damage, MnSOD dysregulation may have multiple effects in cellular function. Catalase (CAT) is a peroxidase that also catalyzes the reduction of hydrogen peroxide to water. Most research focuses on CAT function in extramitochondrial compartments: the peroxisomes. There is some evidence that CAT is present in heart mitochondria, supported by immunocytological and immunoblotting techniques [99]. However, it is generally believed that CAT is absent from mitochondria in other tissues, and most studies have not demonstrated a major role for CAT in mitochondria. Still, targeting of CAT to the mitochondria in Drosophila conferred a resistance to oxidative stress as induced by the addition of exogenous hydrogen peroxide or paraquat, but made no contribution to the extension of lifespan [100]. Conversely, overexpression of mitochondria-targeted CAT in mice increased lifespan signiﬁcantly (>5 months), reduced hydrogen peroxide production, diminished mitochondrial DNA deletions, and also delayed the onset of cardiac and cataract pathology [101]. 7. PROTEOMIC TECHNIQUES FOR IDENTIFICATION OF OXIDATIVELY MODIFIED PROTEINS Redox proteomic techniques have the potential to provide new insights into the components of mitochondrial redox circuitry. Such methods can address the function of electron transfer systems in terms of redox states of individual protein thiols and patterns of intermediary metabolites, thereby facilitating a systems biology–based approach to the evaluation of oxidative stress. One strategy to identify redox-sensitive thiols in complex protein mixtures involves the use of isotope-coded afﬁnity tag (ICAT) labeling. Only free cysteinyl thiols are susceptible to labeling by iodoacetamide-based ICAT, and liquid chromatography–mass spectrometry (LC-MS) can be used to quantify the relative labeling of free thiols. Proof-of-principle studies have shown a selective oxidant-induced decrease in ICAT labeling at cys-283 in creatine kinase, a previously identiﬁed redox-sensitive site [102]. The same approach has been used to identify protein thiols in a rabbit heart membrane fraction that are sensitive to a high concentration of H2 O2 . Of the many protein thiols labeled by ICAT, a relatively small number were oxidized more than 50%, suggesting that redox-sensitive thiols are comparatively rare [103]. A modiﬁcation of the redox ICAT methodology uses sequential labeling by ICAT-H (heavy), reduction, and labeling by ICAT-L (light) to obtain reduced fractions of speciﬁc protein thiols. This approach offers the possibility of further understanding redox changes in terms of the redox potentials of speciﬁc protein thiols [104]. Redox proteomic techniques have also been used to identify oxidatively modiﬁed proteins in toxicological and pathological conditions. Oxidized proteins

FLUORESCENT PROBES FOR DETECTION OF MITOCHONDRIAL ROS

445

from the hippocampus of patients with Alzheimer’s disease (AD) were modiﬁed by reaction with dinitrophenylhydrazine (DNPH), separated by two-dimensional gel electrophoresis and identiﬁed by MALDI-TOF analysis. Targets of oxidation included proteins involved in ATP synthesis, protein degradation, axonal growth, pH regulation, and vesicular transport [105]. A similar redox proteomic strategy was applied to a murine model of doxorubicin-induced cardiotoxicity. Carbonyl levels were increased in triose phosphate isomerase (TPI) and β-enolase, key enzymes of the glycolytic pathway, and in electron transfer ﬂavoprotein–ubiquinone oxidoreductase (ETF-QO), which acts as a transporter for electrons to the mitochondrial respiratory chain [106]. These studies demonstrate the potential of redox proteomics as a tool for the identiﬁcation of key mitochondrial components of toxicologically and pathologically induced mitochondrial oxidative stress. In particular, redox ICAT techniques will facilitate discovery of novel redox-sensitive protein thiols, thus providing the opportunity to map mitochondrial redox circuits and identify sites of disruption during oxidative stress. 8. FLUORESCENT PROBES FOR DETECTION OF MITOCHONDRIAL ROS A relatively new ﬂuorescent probe to detect mitochondrial ROS is the indicator dye MitoSOX, a derivative of dihydroethidine coupled to the mitochondria-targeting triphenylphosphonium cation moiety (Invitrogen Corporation, Carlsbad, CA). MitoSOX is cell permeable and accumulates in the mitochondria. Unlike other popular ﬂuorescent probes for ROS assessment that respond primarily to hydrogen peroxide (e.g., dichloroﬂuoroscein), MitoSOX interacts speciﬁcally with superoxide anion and upon oxidation, exhibits red ﬂuorescence (see Figure 3). MitoSOX used to detect ROS in cells treated with TNF-α revealed a dose-dependent increase in superoxide generation with this cytokine [12]. MitoSOX has been used to evaluate drug-induced mitochondrial ROS generation. In heterozygous SOD2+/− mice, troglitazone treatment resulted in a twofold increases in MitoSOX ﬂuorescence [107] (see Chapter 24). Similarly, MitoSOX has been used to show that the cancer-preventive agent 2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid and its derivatives generate signiﬁcant amounts of mitochondrial ROS [108]. MitoSOX has been used in conjunction with other toxicological agents, including the adenine nucleotide transporter inhibitor atractyloside and the sarco(endo)plasmic reticulum Ca2+ ATPase inhibitor thapsigargin [109]. Other mitochondrial ﬂuorescent probes, such as lucigenin and hydroethidine, also have been shown to detect superoxide anion [110,111]. Other ﬂuorescent probes have increased sensitivity to hydrogen peroxide as compared to superoxide. Dihydrorhodamine 123 (DH123) is a nonﬂuorescent compound that also accumulates in the mitochondria and exhibits red ﬂuorescence upon oxidation by hydrogen peroxide. This dye has been utilized for the measurement of ROS generation in the mitochondria of cells treated with staurosporine [48].

446

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL A

Mitochondrial-targeted

Nuclear-targeted

− D-amino acid

+ D-amino acid

Wildtype

Trx2 overexpression

Redox-sensitive GFP (roGFP)

B

CMF-DA stained NLS-DAAOtransfected

C

MitoSox Stained

Figure 3 Visualization of compartmental redox pools by redox-sensitive ﬂuorescent probes. (A) HeLa cells transfected with redox-sensitive GFP (roGFP) targeted to the mitochondria (green) and nucleus (white), providing mitochondria and nuclear-speciﬁc assessment of the redox state, as described by Hanson et al. [112] and Dooley et al. [113], respectively. (B) HeLa cells transfected with nuclear targeted d-amino acid oxidase (NLS-DAAO), treated with or without d-amino acids and stained with the thiol-speciﬁc ﬂuorescent dye 5-methylchloroﬂuorescein (CMF-DA), as described by Halvey et al. [15]. A selective decrease in nuclear thiol staining occurs during nuclear oxidative stress. (C) HeLa cells were transfected with an empty vector or a thioredoxin-2 (Trx2) overexpressing plasmid and stained with the mitochondrial ROS-speciﬁc indicator dye MitoSOX (Invitrogen Corporation) (unpublished data). [(A) Reproduced from refs. [112] and [113] with permission of the publisher. (B) Reproduced in modiﬁed form from Halvey et al. [15], with permission of the publisher.] (See insert for color presentation of ﬁgure.)

A complementary approach to determining changes in redox environments in mitochondria is the redox-sensitive green ﬂuorescent protein (roGFP) [112,113]. Oxidized and reduced forms of roGFP2 have different excitation wavelengths (reduced at 490 nm and oxidized at 400 nm) for ﬂuorescence. With the assumption that ﬂuorescence characteristics are unaffected by other components in mitochondria, ratiometric redox calibrations of the isolated protein can be used along with measurements in cells to calculate the mitochondrial redox potential, which has

COMPARTMENTATION IN REDOX SIGNALING AND CONTROL

447

been estimated to be −360 mV [112]. Using a similar nuclear-targeted probe, addition of hydrogen peroxide caused an oxidation of nuclear roGFP, as indicated by changes in the ratios of the intensities of reduced and oxidized roGFP. The engineered forms of roGFP show the feasibility of using a redox reporter to measure thiol–disulﬁde redox changes in the mitochondria, but additional isoforms are needed to obtain a form with suitable sensitivity to monitor signaling events.

9. COMPARTMENTATION IN REDOX SIGNALING AND CONTROL IN THE NUCLEUS AND CYTOPLASM In parallel with the methods to measure mitochondrial redox changes and oxidative stress selectively, methods have been developed to measure compartment-speciﬁc redox changes in cytoplasm and nuclei. The availability of these methods permit studies to determine the interaction of the mitochondria with redox events emanating from other speciﬁc compartments and whether mitochondrial oxidative stress affects other compartments. Results show that under conditions of physiologic signaling, redox processes can be highly compartmentalized. Distinct redox regulation in response to EGF-receptor tyrosine kinase activation was shown in the cytoplasm of keratinocytes exposed to EGF [10]. Oxidation of cytoplasmic Trx1 occurred without oxidation of nuclear Trx1, mitochondrial Trx2, or cellular GSH (see Figure 1). Similarly, the mechanistic response to oxidative stress by transcriptional activation of antioxidant systems is compartmentalized. Several transcription factors, including NF-κB, p53, and AP-1, possess reversibly oxidizable cysteines in their DNA-binding domains that are subject to modulation via interaction with a redox sensor protein, redox factor 1 (Ref1) [114–117]. NF-E2-related factor 2 (Nrf2) also undergoes compartmental redox regulation. Cytoplasmic activation of Nrf-2 (release from its inhibitor protein Keap1) is regulated by GSH, and in the nucleus, binding of Nrf2 to its DNA-binding site is regulated primarily by nuclear Trx1 [11]. Similarly, Trx1 has a dual role in the regulation of NF-κB activation and DNA binding. In the cytoplasm, Trx1 has an inhibitory effect on I-κB kinase, thus inhibiting NF-κB translocation to the nucleus [118]. In the nucleus, Trx1 promotes DNA binding of NF-κB by maintaining redox-active cysteines of the p50 subunit in a reduced state via its interaction with Ref1 [118] (see Figure 4). Overexpression of nuclear Trx1 promotes DNA binding of p50, and redox modiﬁcation of Cys62 in p50 results in decreased DNA binding [118–120]. While bolus addition of tBH caused oxidation of both nuclear and cytoplasmic Trx1 as well as cellular GSH [121], thiol–disulﬁde redox couples have compartmentalized responses to other types of oxidative stress. Redox western blot can be used to determine individual protein redox states in nuclear and cytoplasmic fractions. However, the ability to measure nuclear GSH/GSSG redox state is confounded by loss of GSH from the nuclei during fractionation procedures. Previous

448

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

estimates of nuclear GSH have given contradictory results [122–124]. Therefore, an indirect estimation of nuclear GSH/GSSG can be obtained by measuring nuclear S -glutathionylated protein (Pr-SSG). Pr-SSG can form from reaction of GSSG with protein thiols (PrSH) or reaction of oxidized proteins with GSH. Pr-SSG concentrations in nuclear and cytoplasmic components can be assayed following reduction of nuclear and cytoplasmic protein fractions with dithiothreitol, releasing GSH for measurement by high-performance liquid chromatography [10,125]. These methods have been applied to the investigation of nutrient deprivation and oxidative stress in cultured cells. Glutamine- and glucose-deﬁcient cells show increased production of ROS and undergo differential subcellular redox changes [125]. Cytoplasmic Trx1 and Ref1, measured by redox Western immunoblot and biotin–iodoactamide labeling, respectively, were oxidized,

Glucose, Glutamine Deficiency; TNF-a Cytoplasm

Cytoplasmic Oxidative Stress ROS NF-KB p50 p65

Trx1Red GSH

I-KB Kinase I-KB I-KB P

p50 p65

Degradation

p50 p65 Nucleus

p50 p65

Normal Nuclear Redox

Red

Trx1

Transcriptional Activation

p50 p65 Ref1

A

Figure 4 Effects of cytoplasmic- and nuclear-localized oxidative stress on redoxdependent transcription factor activity. NF-κB (p50 and p65 subunits) localizes to the cytoplasm, where it is bound to its inhibitory protein, I-κB. Oxidative stress promotes dissociation of NF-κB and I-κB by stimulating I-κB kinase, which initiates the degradation of I-κB. A redox-sensitive cysteine (Cys-62) in the p50 subunit can occur as reduced (circle), oxidized (square), or unspeciﬁed oxidized/reduced (triangle) forms. Cys-62 must be in reduced form to facilitate DNA binding. (A) Under conditions of cytoplasmic oxidative stress, Trx1 and GSH cannot inhibit IKK activity, leading to degradation of I-κB and translocation of NF-κB to the nucleus. In the nucleus, Trx1 reduces Ref1 and Ref1 reduces p50 to allow DNA binding and transcriptional activation. (B) Under conditions of nuclear oxidative stress generated by nuclear localized d-amino acid oxidase (NLS-DAO) the Trx1/Ref1 system is prevented from maintaining Cys-62 in a reduced form, and p50 DNA binding is inhibited.

449

COMPARTMENTATION IN REDOX SIGNALING AND CONTROL NF-KB Stimulus Normal Cytoplasmic Redox NF-KB p50 p65

Trx1 GSH

I-KB Kinase Cytoplasm

I-KB

p50 p65

I-KB P

Degradation

NLS-D-amino Acid Oxidase p50 p65 Nuclear Oxidative Stress H2O2

Ref1 Trx1

p50 p65

Nucleus

Oxidized No DNA binding No Transcriptional Activation

B

Figure 4 (Continued)

whereas nuclear Trx1 and Ref1 were unaffected. Pr-SSG increased in the cytoplasm but not in the nucleus, suggesting that nuclei contain enhanced antioxidant capacity under these stress conditions [125]. The effects of selective nuclear oxidative stress have been examined by targeting a ROS-generating enzyme, d-amino acid oxidase, to the nucleus [15]. Transient transfection of nuclear-targeted d-amino acid oxidase (NLS-DAAO) in HeLa cells, followed by exogenous supply of N -acetyl-d-alanine (NADA) substrate, caused a signiﬁcant increase in ROS formation. Selective oxidation of the nuclear thiol pool was conﬁrmed using the thiol-binding dye 5-methylchloroﬂuorescein. Increases in nuclear Pr-SSG were observed in NLS-DAAO/NADA-treated cells, but nuclear Trx1 was not oxidized under the same conditions, indicating that the GSH/GSSG pool responds more readily to localized nuclear oxidative stress. Interestingly, stimulation of NF-κB reporter activity by TNF-α was inhibited by NLS-DAAO/NADA, suggesting a major role for nuclear protein S-glutathionylation in controlling nuclear transcription during nuclear oxidative stress [15] (see Figures 3 and 4). These new tools to measure oxidative stress selectively in the mitochondria, cytoplasm and nuclei, create the possibility of determining sites of toxicity due to agents that cause oxidative stress. For instance, doxorubicin intercalates into DNA and can cause either mitochondrial or nuclear toxicity. Redox cycling agents can generate ROS in the mitochondria, cytoplasm, or endoplasmic reticulum. Acetaminophen and ethanol can cause oxidative reactions in the endoplasmic

450

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

reticulum and mitochondria. Thus, application of these methods should considerably extend the knowledge of toxicologic mechanisms. 10. OXIDATIVE PROTEIN FOLDING AND REDOX SIGNALING EVENTS IN THE ENDOPLASMIC RETICULUM Convenient methods are not yet available to study oxidative stress in the endoplasmic reticulum (ER), but much information is available concerning the redox systems, and their failure, during ER stress. Redox changes in the ER, associated with protein misfolding and disruption of calcium homeostasis, have downstream effects on redox signaling processes. The relatively oxidizing environment of the ER (−180 mV) facilitates protein disulﬁde formation [126]. GSH/GSSG ratios in the ER vary between 3:1 and 1:1, considerably more oxidized than the cytoplasmic ratio of 30:1 to 100:1, accounting in part for the oxidizing microenvironment. A signiﬁcant portion of ER GSH occurs as protein mixed disulﬁdes, which may regulate the activity of redox-active thiol-containing proteins and acts as a buffer against hyperoxidizing conditions in the ER [127] [128]. Interand intraprotein disulﬁde formation in the ER is catalyzed principally by protein disulﬁde isomerase (PDI) and several related oxidoreductases. PDI contains two thioredoxin-like CGHC motifs, which promote disulﬁde bond formation and rearrangement. Endoplasmic reticulum oxidase 1 (ERO1) recycles PDI back to its active disulﬁde form by supplying oxidizing equivalents. ERO1 is itself oxidized by molecular oxygen (O2 ), transferring electrons to O2 and generating O2 ·− in the process [129] (see Figure 5). A variety of pathological and toxicological stresses cause accumulation of misfolded proteins in the ER, leading to up-regulation of ER protein folding Mitochondria

Endoplasmic Reticulum Electron Flow

Unfolded protein Pr Pr

SH SH S S

ERO1

PDI S

S

SH

PDI SH

SH

SH

ERO1 S

S

Redox sensitive Ca2+ release InsP3R O2 Ca2+ ? -• RyR O2 ROS

Ca2+ uptake

MPT Induction

Apoptosis

Figure 5 Effects of oxidative protein folding–derived ROS in the ER on mitochondria. Correct folding of proteins in the endoplasmic reticulum is catalyzed by protein disulﬁde isomerase (PDI), which is recycled back to its active oxidized from by ER oxidase 1 (ERO1). ERO1 is itself oxidized by molecular oxygen (O2 ), generating superoxide as a by-product of oxidative protein folding. Under stress conditions, excess superoxide or components of protein folding machinery may activate redox-sensitive Ca2+ release from the ryanodine receptor (RyR) and the inositol 1,4,5-trisphosphate receptor (InsP3 R). Uptake of released Ca2+ release by mitochondria can stimulate opening of the mitochondrial permeablization transition pore (MPTP), leading to apoptosis.

OXIDATIVE PROTEIN FOLDING AND REDOX SIGNALING

451

chaperones and attenuation of translation initiation, a process termed the unfolded protein response (UPR) [130,131]. Prolonged UPR activation is associated with oxidative stress and cell death [132]. UPR-derived ROS come from two possible sources: oxidative protein folding machinery in the ER and mitochondrial redox cycling. Increased ERO1 activity, due to an abundance of misfolded proteins, leads to electron transfer to O2 , thereby generating O2 ·− [129]. Furthermore, GSH reduces incorrect protein disulﬁdes, leading to GSH depletion and diminishing the capacity of the ER to counteract ROS [127]. Overexpression of Ero1p in yeast resulted in both enhanced ROS levels and a decrease in GSH under nonstress conditions. Therefore, during UPR activation, similar up-regulation of Ero1p and Pdi1p (yeast PDI) would greatly exacerbate oxidative stress in the ER by further increasing the rates of ROS production and GSH depletion to reduce improper disulﬁdes [132]. Other studies have demonstrated a link between the UPR pathway and Nrf2 signaling [133]. The UPR-activated kinase, PERK, triggers the dissociation of Nrf2/Keap1 complexes and inhibits reassociation of Nrf2/Keap1. UPR-derived ROS may stimulate Ca2+ release from the ER, thereby activating mitochondria and generating more ROS. Ca2+ release and uptake channels in the ER membrane contain redox-sensitive cysteines that, in part, regulate their activities. The RyR (ryanodine receptor) is an intracellular ion channel mediating Ca2+ release from the ER. Although RyR channel activity is regulated by Ca2+ , Mg2+ , and ATP, redox-sensitive thiol groups play a critical role in channel opening. Oxidizing conditions promote channel opening, while reducing conditions have the opposite effect. Pharmacological thiol-reactive reagents, including dithiodipyridines, N -ethylmaleimide, and diamide, activate RyR Ca2+ release from skeletal and cardiac muscle. The reducing reagents dithiothreitol and GSH reversed this effect [134,135]. Cys-3635 of RyR has been identiﬁed as being functionally relevant to the redox-sensing properties of the channel. Expression of the C3635A-RyR1 mutant altered the sensitivity of the ryanodine receptor activation. Similarly, regulation of the ER calcium release channel inositol 1,4,5-trisphosphate receptor (InsP3 R) has a redox-dependent component [136]. InsP3 R interacts with an ER resident protein ERp44, a PDI-like chaperone that is up-regulated during ER stress [136]. When active-site cysteines of ERp44 are in a reduced state, the protein binds to InsP3 R, inhibiting Ca2+ release. Interestingly, calcium reuptake into the ER is also regulated by redox conditions. ERp57, a PDI-like oxidoreductase, modulates sarco/endoplasmic reticulum Ca2+ ATPase 2b (SERCA2b) activity [137]. Under conditions of high lumenal calcium, ERp57 and its interacting partner calreticulin associate with SERCA2b, inhibiting its activity [137]. Ca2+ depletion results in ERp57 dissociation and abolishes SERCA2b inhibition. These interactions suggest an important link between oxidative protein folding machinery in the ER and Ca2+ signaling (see Figure 5). In the context of ER- mitochondria communication, ROS-induced Ca2+ release from the ER may propagate UPR-derived redox signaling events. This has signiﬁcant implications for the ER stress response and ER stress-induced apoptosis. When large quantities of Ca2+ are accumulated in the mitochondrial matrix, Ca2+

452

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

interacts with cyclophilin D to induce opening of the mitochondrial permeability transition pore (MPTP), a key component of the apoptotic pathway [138]. Furthermore, the rise in mitochondrial Ca2+ stimulates the generation of ROS, which also promotes the opening of the MPTP [139]. Nitric oxide (NO) stimulates Ca2+ release from the mitochondria in a redox-dependent manner, which can activate ATF6, an ER stress-activated transcription factor [140]. Additional factors regulate ER-mitochondrial Ca2+ signaling. The proapoptotic proteins Bax and Bak localize to the ER membrane and are required to optimize the storage and mobilization of Ca2+ [141]. Cells deﬁcient in Bax and Bak were found to have a reduced resting concentration of Ca2+ in the ER, resulting in decreased uptake of Ca2+ by mitochondria [141]. The effect of Bax/Bak may be due to suppression of the interaction between the InsP3 R and the anti-apoptotic Bcl-2 and Bcl-xL, which controls Ca2+ leakage from the ER [142]. It is likely that components of the oxidative folding machinery in the ER interact with pro- and anti-apoptotic factors, linking oxidative stress, Ca2+ signaling, and mitochondrially mediated apoptosis (see Figure 5).

11. CONCLUSIONS AND FUTURE PERSPECTIVES The concept of oxidative stress as an overproduction of reactive oxygen and nitrogen species, resulting in macromolecular damage, has existed for some time. More recently, ROS and RNS have emerged as important cell signal mediators. Selective shifts in redox balance of speciﬁc proteins, associated with subtoxic levels of ROS and RNS, provide molecular signals that regulate a variety of processes, including transcription factor activity, kinase signaling, and metabolic control. Protein and nonprotein thiols are both vital components of redox signaling machinery, due to their broad range of redox-dependent reactions. These reactions include structural changes in proteins, which are mostly due to interand intraprotein disulﬁde bridges, and changes in surface properties due to cysteinylation or glutathionylation. GSH/GSSG, Trx, and Prx are key participants in redox signaling processes, acting as distinct nodes in thiol–disulﬁde redox circuits. Here we have examined the concept of subcellular compartmentation of thiol–disulﬁde circuitry in redox signaling control and the current methods available for assessment of redox circuitry. Localized peroxide signals can provide selective oxidation of speciﬁc protein thiols in organelles where the ratio of the metabolic or transport ﬂux relative to the mean concentration is high [143]. Therefore, methods for detection of the redox state of individual components in subcellular compartments will yield more sensitive and speciﬁc indicators of oxidative stress. New approaches such as the redox Western blot, redox proteomics, and redox-sensitive ﬂuorescent indicators will allow for the delineation of redox signals in terms of subcellular compartmental redox states. Localization of cell signaling events is not limited to redox signaling. Ca2+ microdomains fulﬁll the need to spatially separate Ca2+ regulation of different

REFERENCES

453

cellular processes. However, due to the heterogeneity of both ROS/RNS signaling molecules and the protein modiﬁcations they induce, as well as the relative instability of redox modiﬁcations, assessment of redox microdomains remains technically challenging. Therefore, quantitative proteomic approaches, coupled with tools for the validation of speciﬁc redox modiﬁcations, are necessary to detect the subtle redox changes that occur under physiologic conditions. Furthermore, identiﬁcation of compartment-speciﬁc redox changes will necessitate more sensitive organelle-targeted indicators, as has been demonstrated by the use of mitochondrial-targeted redox-sensitive GFP. Future studies of compartmental redox signaling will shed new light on toxicological and pathological conditions where oxidative stress has been demonstrated previously. Acknowledgments The research on which this chapter was based was supported by National Institutes of Health grants ES009047, ES011195, and DK040725.

REFERENCES 1. Puddu P, Puddu GM, Galletti L, Cravero E, Muscari A. Mitochondrial dysfunction as an initiating event in atherogenesis: a plausible hypothesis. Cardiology. 2005;103:137–141. 2. Lin MT, Beal MF. Mitochondrial dysfunction and oxidative stress in neurodegenerative diseases. Nature. 2006;443:787–795. 3. Klaunig JE, Kamendulis LM. The role of oxidative stress in carcinogenesis. Annu Rev Pharmacol Toxicol . 2004;44:239–267. 4. Chen YR, Chen CL, Zhang L, Green-Church KB, Zweier JL. Superoxide generation from mitochondrial NADH dehydrogenase induces self-inactivation with speciﬁc protein radical formation. J Biol Chem. 2005;280:37339–37348. 5. McStay GP, Clarke SJ, Halestrap AP. Role of critical thiol groups on the matrix surface of the adenine nucleotide translocase in the mechanism of the mitochondrial permeability transition pore. Biochem J . 2002;367:541–548. 6. Taylor ER, Hurrell F, Shannon RJ, Lin TK, Hirst J, Murphy MP. Reversible glutathionylation of complex I increases mitochondrial superoxide formation. J Biol Chem. 2003;278:19603–19610. 7. Hansen JM, Go YM, Jones DP. Nuclear and mitochondrial compartmentation of oxidative stress and redox signaling. Annu Rev Pharmacol Toxicol . 2006;46:215–234. 8. Chen Y, Cai J, Murphy TJ, Jones DP. Overexpressed human mitochondrial thioredoxin confers resistance to oxidant-induced apoptosis in human osteosarcoma cells. J Biol Chem. 2002;277:33242–33248. 9. Nonn L, Berggren M, Powis G. Increased expression of mitochondrial peroxiredoxin-3 (thioredoxin peroxidase-2) protects cancer cells against hypo xia and drug-induced hydrogen peroxide-dependent apoptosis. Mol Cancer Res. 2003;1:682–689.

454

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

10. Halvey PJ, Watson WH, Hansen JM, Go YM, Samali A, Jones DP. Compartmental oxidation of thiol-disulphide redox couples during epidermal growth factor signalling. Biochem J . 2005;386:215–219. 11. Hansen JM, Watson WH, Jones DP. Compartmentation of Nrf-2 redox control: regulation of cytoplasmic activation by glutathione and DNA binding by thioredoxin-1. Toxicol Sci. 2004;82:308–317. 12. Hansen JM, Zhang H, Jones DP. Mitochondrial thioredoxin-2 has a key role in determining tumor necrosis factor-alpha-induced reactive oxygen species generation, NF-kappaB activation, and apoptosis. Toxicol Sci. 2006;91:643–650. 13. Armstrong JS, Jones DP. Glutathione depletion enforces the mitochondrial permeability transition and causes cell death in Bcl-2 overexpressing HL60 cells. FASEB J . 2002;16:1263–1265. 14. Powis G, Montfort WR. Properties and biological activities of thioredoxins. Annu Rev Biophys Biomol Struct. 2001;30:421–455. 15. Halvey PJ, Hansen JM, Johnson JM, Go YM, Samali A, Jones DP. Selective oxidative stress in cell nuclei by nuclear-targeted d-amino acid oxidase. Antioxid Redox Signal . 2007;9:807–816. 16. Kushnareva Y, Murphy AN, Andreyev A. Complex I–mediated reactive oxygen species generation: modulation by cytochrome c and NAD(P)+ oxidation–reduction state. Biochem J . 2002;368:545–553. 17. Chen Q, Vazquez EJ, Moghaddas S, Hoppel CL, Lesnefsky EJ. Production of reactive oxygen species by mitochondria: central role of complex III. J Biol Chem. 2003;278:36027–36031. 18. Genova ML, Ventura B, Giuliano G, et al. The site of production of superoxide radical in mitochondrial complex I is not a bound ubisemiquinone but presumably iron–sulfur cluster N2. FEBS Lett. 2001;505:364–368. 19. Liu Y, Fiskum G, Schubert D. Generation of reactive oxygen species by the mitochondrial electron transport chain. J Neurochem. 2002;80:780–787. 20. Genova ML, Pich MM, Biondi A, et al. Mitochondrial production of oxygen radical species and the role of coenzyme Q as an antioxidant. Exp Biol Med (Maywood). 2003;228:506–513. 21. Aguilaniu H, Gustafsson L, Rigoulet M, Nystrom T. Asymmetric inheritance of oxidatively damaged proteins during cytokinesis. Science. 2003;299:1751–1753. 22. Turrens JF, Alexandre A, Lehninger AL. Ubisemiquinone is the electron donor for superoxide formation by complex III of heart mitochondria. Arch Biochem Biophys. 1985;237:408–414. 23. St-Pierre J, Buckingham JA, Roebuck SJ, Brand MD. Topology of superoxide production from different sites in the mitochondrial electron transport chain. J Biol Chem. 2002;277:44784–44790. 24. Miwa S, St-Pierre J, Partridge L, Brand MD. Superoxide and hydrogen peroxide production by Drosophila mitochondria. Free Radic Biol Med. 2003;35:938–948. 25. Starkov AA, Fiskum G, Chinopoulos C, et al. Mitochondrial alpha-ketoglutarate dehydrogenase complex generates reactive oxygen species. J Neurosci . 2004;24:7779–7788. 26. Giorgio M, Migliaccio E, Orsini F, et al. Electron transfer between cytochrome c and p66Shc generates reactive oxygen species that trigger mitochondrial apoptosis. Cell. 2005;122:221–233.

REFERENCES

455

27. Sandri G, Panﬁli E, Ernster L. Hydrogen peroxide production by monoamine oxidase in isolated rat-brain mitochondria: its effect on glutathione levels and Ca2+ efﬂux. Biochim Biophys Acta. 1990;1035:300–305. 28. Jones DP. Disruption of mitochondrial redox circuitry in oxidative stress. Chem Biol Interact . 2006;163:38–53. 29. Beckman JS, Beckman TW, Chen J, Marshall PA, Freeman BA. Apparent hydroxyl radical production by peroxynitrite: implications for endothelial injury from nitric oxide and superoxide. Proc Natl Acad Sci U S A. 1990;87:1620–1624. 30. Pacher P, Beckman JS, Liaudet L. Nitric oxide and peroxynitrite in health and disease. Physiol Rev. 2007;87:315–424. 31. Wallace KB. Doxorubicin-induced cardiac mitochondrionopathy. Pharmacol Toxicol . 2003;93:105–115. 32. Scatena R, Bottoni P, Botta G, Martorana GE, Giardina B. The role of mitochondria in pharmacotoxicology: a re-evaluation of an old, newly emerging topic. Am J Physiol Cell Physiol. 2007;293:C12–21. 33. Burcham PC, Harman AW. Acetaminophen toxicity results in site-speciﬁc mitochondrial damage in isolated mouse hepatocytes. J Biol Chem. 1991;266:5049–5054. 34. Knight TR, Kurtz A, Bajt ML, Hinson JA, Jaeschke H. Vascular and hepatocellular peroxynitrite formation during acetaminophen toxicity: role of mitochondrial oxidant stress. Toxicol Sci. 2001;62:212–220. 35. Ruepp SU, Tonge RP, Shaw J, Wallis N, Pognan F. Genomics and proteomics analysis of acetaminophen toxicity in mouse liver. Toxicol Sci. 2002;65:135–150. 36. Perez-Ortiz JM, Tranque P, Vaquero CF, et al. J. Glitazones differentially regulate primary astrocyte and glioma cell survival. Involvement of reactive oxygen species and peroxisome proliferator-activated receptor-gamma. J Biol Chem. 2004;279:8976–8985. 37. Perez-Ortiz JM, Tranque P, Burgos M, Vaquero CF, Llopis J. Glitazones induce astroglioma cell death by releasing reactive oxygen species from mitochondria: modulation of cytotoxicity by nitric oxide. Mol Pharmacol. 2007;72:407–417. 38. Kanunfre CC, da Silva Freitas JJ, Pompeia C, et al. Ciglitizone and 15d PGJ2 induce apoptosis in Jurkat and Raji cells. Int Immunopharmacol . 2004;4:1171–1185. 39. Soller M, Drose S, Brandt U, Brune B, von Knethen A. Mechanism of thiazolidinedione-dependent cell death in Jurkat T cells. Mol Pharmacol . 2007;71:1535–1544. 40. Ray DM, Bernstein SH, Phipps RP. Human multiple myeloma cells express peroxisome proliferator-activated receptor gamma and undergo apoptosis upon exposure to PPARgamma ligands. Clin Immunol . 2004;113:203–213. 41. Masubuchi Y. Metabolic and non-metabolic factors determining troglitazone hepatotoxicity: a review. Drug Metab Pharmacokinet. 2006;21:347–356. 42. Pandhare J, Cooper SK, Phang JM. Proline oxidase, a proapoptotic gene, is induced by troglitazone: evidence for both peroxisome proliferator-activated receptor gamma-dependent and -independent mechanisms. J Biol Chem. 2006;281:2044–2052. 43. Jung TW, Lee JY, Shim WS, et al. Rosiglitazone protects human neuroblastoma SH-SY5Y cells against MPP+ induced cytotoxicity via inhibition of mitochondrial dysfunction and ROS production. J Neurol Sci. 2007;253:53–60.

456

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

44. Scatena R, Bottoni P, Martorana GE, et al. Mitochondria, ciglitazone and liver: a neglected interaction in biochemical pharmacology. Eur J Pharmacol . 2007;567:50–58. 45. Shan X, Jones DP, Hashmi M, Anders MW. Selective depletion of mitochondrial glutathione concentrations by (R,S )-3-hydroxy-4-pentenoate potentiates oxidative cell death. Chem Res Toxicol . 1993;6:75–81. 46. Meredith MJ, Reed DJ. Status of the mitochondrial pool of glutathione in the isolated hepatocyte. J Biol Chem. 1982;257:3747–3753. 47. Rebrin I, Sohal RS. Comparison of thiol redox state of mitochondria and homogenates of various tissues between two strains of mice with different longevities. Exp Gerontol . 2004;39:1513–1519. 48. Cai J, Jones DP. Superoxide in apoptosis: mitochondrial generation triggered by cytochrome c loss. J Biol Chem. 1998;273:11401–11404. 49. Grifﬁth OW, Meister A. Origin and turnover of mitochondrial glutathione. Proc Natl Acad Sci U S A. 1985;82:4668–4672. 50. Lash LH. Mitochondrial glutathione transport: physiological, pathological and toxicological implications. Chem Biol Interact . 2006;163:54–67. 51. Lash LH, Putt DA, Matherly LH. Protection of NRK-52E cells, a rat renal proximal tubular cell line, from chemical-induced apoptosis by overexpression of a mitochondrial glutathione transporter. J Pharmacol Exp Ther. 2002;303:476–486. 52. McKernan TB, Woods EB, Lash LH. Uptake of glutathione by renal cortical mitochondria. Arch Biochem Biophys. 1991;288:653–663. 53. Kurosawa K, Hayashi N, Sato N, Kamada T, Tagawa K. Transport of glutathione across the mitochondrial membranes. Biochem Biophys Res Commun. 1990;167:367–372. 54. Arai M, Imai H, Koumura T, et al. Mitochondrial phospholipid hydroperoxide glutathione peroxidase plays a major role in preventing oxidative injury to cells. J Biol Chem. 1999;274:4924–4933. 55. Lundberg M, Johansson C, Chandra J, et al. Cloning and expression of a novel human glutaredoxin (Grx2) with mitochondrial and nuclear isoforms. J Biol Chem. 2001;276:26269–26275. 56. Esposito LA, Kokoszka JE, Waymire KG, Cottrell B, MacGregor GR, Wallace DC. Mitochondrial oxidative stress in mice lacking the glutathione peroxidase-1 gene. Free Radic Biol Med. 2000;28:754–766. 57. Yant LJ, Ran Q, Rao L, et al. The selenoprotein GPX4 is essential for mouse development and protects from radiation and oxidative damage insults. Free Radic Biol Med. 2003;34:496–502. 58. Liang H, Van Remmen H, Frohlich V, Lechleiter J, Richardson A, Ran Q. Gpx4 protects mitochondrial ATP generation against oxidative damage. Biochem Biophys Res Commun. 2007;356:893–898. 59. Holmgren A. Antioxidant function of thioredoxin and glutaredoxin systems. Antioxid Redox Signal . 2000;2:811–820. 60. Beer SM, Taylor ER, Brown SE, et al. Glutaredoxin 2 catalyzes the reversible oxidation and glutathionylation of mitochondrial membrane thiol proteins: implications for mitochondrial redox regulation and antioxidant DEFENSE. J Biol Chem. 2004;279:47939–47951.

REFERENCES

457

61. Kil IS, Park JW. Regulation of mitochondrial NADP+ -dependent isocitrate dehydrogenase activity by glutathionylation. J Biol Chem. 2005;280:10846–10854. 62. Enoksson M, Fernandes AP, Prast S, Lillig CH, Holmgren A, Orrenius S. Overexpression of glutaredoxin 2 attenuates apoptosis by preventing cytochrome c release. Biochem Biophys Res Commun. 2005;327:774–779. 63. Fernandez-Checa JC, Hirano T, Tsukamoto H, Kaplowitz N. Mitochondrial glutathione depletion in alcoholic liver disease. Alcohol . 1993;10:469–475. 64. Fernandez-Checa JC, Garcia-Ruiz C, Ookhtens M, Kaplowitz N. Impaired uptake of glutathione by hepatic mitochondria from chronic ethanol-fed rats. Tracer kinetic studies in vitro and in vivo and susceptibility to oxidant stress. J Clin Invest. 1991;87:397–405. 65. Przedborski S, Jackson-Lewis V. Mechanisms of MPTP toxicity. Mov Disord . 1998;13 (suppl 1): 35–38. 66. Wullner U, Loschmann PA, Schulz JB, et al. Glutathione depletion potentiates MPTP and MPP+ toxicity in nigral dopaminergic neurones. Neuroreport. 1996;7:921–923. 67. Chinta SJ, Rajagopalan S, Butterﬁeld DA, Andersen JK. In vitro and in vivo neuroprotection by gamma-glutamylcysteine ethyl ester against MPTP: relevance to the role of glutathione in Parkinson’s disease. Neurosci Lett. 2006;402:137–141. 68. Perry TL, Yong VW, Jones K, Wright JM. Manipulation of glutathione contents fails to alter dopaminergic nigrostriatal neurotoxicity of N -methyl-4-phenyl-1,2,3,6tetrahydropyridine (MPTP) in the mouse. Neurosci Lett. 1986;70:261–265. 69. Li N, Ragheb K, Lawler G, Sturgis J, Rajwa B, Melendez JA, Robinson JP. Mitochondrial complex I inhibitor rotenone induces apoptosis through enhancing mitochondrial reactive oxygen species production. J Biol Chem. 2003;278:8516–8525. 70. Vendemiale G, Grattagliano I, Altomare E, Turturro N, Guerrieri F. Effect of acetaminophen administration on hepatic glutathione compartmentation and mitochondrial energy metabolism in the rat. Biochem Pharmacol . 1996;52:1147–1154. 71. Zhao P, Kalhorn TF, Slattery JT. Selective mitochondrial glutathione depletion by ethanol enhances acetaminophen toxicity in rat liver. Hepatology. 2002;36:326–335. 72. Ghibelli L, Coppola S, Fanelli C, et al. Glutathione depletion causes cytochrome c release even in the absence of cell commitment to apoptosis. FASEB J. 1999;13:2031–2036. 73. Jungas T, Motta I, Dufﬁeux F, Fanen P, Stoven V, Ojcius DM. Glutathione levels and BAX activation during apoptosis due to oxidative stress in cells expressing wild-type and mutant cystic ﬁbrosis transmembrane conductance regulator. J Biol Chem. 2002;277:27912–27918. 74. Troyano A, Fernandez C, Sancho P, de Blas E, Aller P. Effect of glutathione depletion on antitumor drug toxicity (apoptosis and necrosis) in U-937 human promonocytic cells. The role of intracellular oxidation. J Biol Chem. 2001;276:47107–47115. 75. Chen Z, Lash LH. Evidence for mitochondrial uptake of glutathione by dicarboxylate and 2-oxoglutarate carriers. J Pharmacol Exp Ther. 1998;285:608–618. 76. Chen Z, Putt DA, Lash LH. Enrichment and functional reconstitution of glutathione transport activity from rabbit kidney mitochondria: further evidence for the role of the dicarboxylate and 2-oxoglutarate carriers in mitochondrial glutathione transport. Arch Biochem Biophys. 2000;373:193–202. 76a. Zhong Q, Putt DA, Xu F, Lash LH. Hepatic mitochondrial transport of glutathione. Arch Biochem Biophys. 2008, in press.

458

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

77. Xu F, Putt DA, Matherly LH, Lash LH. Modulation of expression of rat mitochondrial 2-oxoglutarate carrier in NRK-52E cells alters mitochondrial transport and accumulation of glutathione and susceptibility to chemically induced apoptosis. J Pharmacol Exp Ther. 2006;316:1175–1186. 78. Lash LH, Anders MW. Mechanism of S -(1,2-dichlorovinyl)-l-cysteine- and S -(1,2-dichlorovinyl)-l-homocysteine-induced renal mitochondrial toxicity. Mol Pharmacol . 1987;32:549–556. 79. Watabe S, Hiroi T, Yamamoto Y, et al. SP-22 is a thioredoxin-dependent peroxide reductase in mitochondria. Eur J Biochem. 1997;249:52–60. 80. Chen Y, Yu M, Jones DP, Greenamyre JT, Cai J. Protection against oxidant-induced apoptosis by mitochondrial thioredoxin in SH-SY5Y neuroblastoma cells. Toxicol Appl Pharmacol . 2006;216:256–262. 81. Tanaka T, Hosoi F, Yamaguchi-Iwai Y, et al. Thioredoxin-2 (TRX-2) is an essential gene regulating mitochondria-dependent apoptosis. EMBO J. 2002;21:1695–1703. 82. Wang D, Masutani H, Oka S, et al. Control of mitochondrial outer membrane permeabilization and Bcl-xL levels by thioredoxin 2 in DT40 cells. J Biol Chem. 2006;281:7384–7391. 83. Damdimopoulos AE, Miranda-Vizuete A, Pelto-Huikko M, Gustafsson JA, Spyrou G. Human mitochondrial thioredoxin: involvement in mitochondrial membrane potential and cell death. J Biol Chem. 2002;277:33249–33257. 84. Zhang R, Al-Lamki R, Bai L, Streb JW, Miano JM, Bradley J, Min W. Thioredoxin-2 inhibits mitochondria-located ASK1-mediated apoptosis in a JNK-independent manner. Circ Res. 2004;94:1483–1491. 85. Chang TS, Cho CS, Park S, Yu S, Kang SW, Rhee SG. Peroxiredoxin III, a mitochondrion-speciﬁc peroxidase, regulates apoptotic signaling by mitochondria. J Biol Chem. 2004;279:41975–41984. 86. Ramachandiran S, Hansen JM, Jones DP, Richardson JR, Miller GW. Divergent mechanisms of paraquat, MPP+ , and rotenone toxicity: oxidation of thioredoxin and caspase-3 activation. Toxicol Sci. 2007;95:163–171. 87. Hansen JM, Zhang H, Jones DP. Differential oxidation of thioredoxin-1, thioredoxin-2, and glutathione by metal ions. Free Radic Biol Med. 2006;40: 138–145. 88. Calvert JW, Lefer DJ. Statin therapy and myocardial no-reﬂow. Br J Pharmacol . 2006;149:229–231. 89. Jones SP, Teshima Y, Akao M, Marban E. Simvastatin attenuates oxidant-induced mitochondrial dysfunction in cardiac myocytes. Circ Res. 2003;93:697–699. 90. Zhao JL, Yang YJ, Cui CJ, You SJ, Gao RL. Pretreatment with simvastatin reduces myocardial no-reﬂow by opening mitochondrial K(ATP) channel. Br J Pharmacol . 2006;149:243–249. 91. Ucar M, Mjorndal T, Dahlqvist R. HMG-CoA reductase inhibitors and myotoxicity. Drug Saf. 2000;22:441–457. 92. Sirvent P, Mercier J, Vassort G, Lacampagne A. Simvastatin triggers mitocho ndria-induced Ca2+ signaling alteration in skeletal muscle. Biochem Biophys Res Commun. 2005;329:1067–1075. 93. Sirvent P, Bordenave S, Vermaelen M, et al. Simvastatin induces impairment in skeletal muscle while heart is protected. Biochem Biophys Res Commun. 2005;338:1426–1434.

REFERENCES

459

94. Li Y, Huang TT, Carlson EJ, Melov S, et al. Dilated cardiomyopathy and neonatal lethality in mutant mice lacking manganese superoxide dismutase. Nat Genet. 1995;11:376–381. 95. Melov S, Schneider JA, Day BJ, et al. A novel neurological phenotype in mice lacking mitochondrial manganese superoxide dismutase. Nat Genet. 1998;18:159–163. 96. Melov S, Coskun P, Patel M, et al. Mitochondrial disease in superoxide dismutase 2 mutant mice. Proc Natl Acad Sci U S A. 1999;96:846–851. 97. Dhanasekaran A, Kotamraju S, Karunakaran C, et al. Mitochondria superoxide dismutase mimetic inhibits peroxide-induced oxidative damage and apoptosis: role of mitochondrial superoxide. Free Radic Biol Med. 2005;39:567–583. 98. Nemoto S, Takeda K, Yu ZX, Ferrans VJ, Finkel T. Role for mitochondrial oxidants as regulators of cellular metabolism. Mol Cell Biol. 2000;20:7311–7318. 99. Radi R, Turrens JF, Chang LY, Bush KM, Crapo JD, Freeman BA. Detection of catalase in rat heart mitochondria. J Biol Chem. 1991;266:22028–22034. 100. Mockett RJ, Bayne AC, Sohal BH, Sohal RS. Biochemical assay of superoxide dismutase activity in Drosophila. Methods Enzymol . 2002;349:287–292. 101. Schriner SE, Linford NJ, Martin GM, et al. Extension of murine life span by overexpression of catalase targeted to mitochondria. Science. 2005;308:1909–1911. 102. Sethuraman M, McComb ME, Heibeck T, Costello CE, Cohen RA. Isotope-coded afﬁnity tag approach to identify and quantify oxidant-sensitive protein thiols. Mol Cell Proteom. 2004;3:273–278. 103. Sethuraman M, McComb ME, Huang H, et al. Isotope-coded afﬁnity tag (ICAT) approach to redox proteomics: identiﬁcation and quantitation of oxidant-sensitive cysteine thiols in complex protein mixtures. J Proteome Res. 2004;3:1228–1233. 104. Go YM, Pohl J, Jones DP. Quantiﬁcation of redox conditions in the nucleus. Methods Mol. Biol. 2008; in press. 105. Sultana R, Perluigi M, Butterﬁeld DA. Redox proteomics identiﬁcation of oxidatively modiﬁed proteins in Alzheimer’s disease brain and in vivo and in vitro models of AD centered around Abeta(1–42). J Chromatogr B Anal Technol Biomed Life Sci. 2006;833:3–11. 106. Chen Y, Daosukho C, Opii WO, et al. Redox proteomic identiﬁcation of oxidized cardiac proteins in adriamycin-treated mice. Free Radic Biol Med. 2006;41:1470–1477. 107. Ong MM, Latchoumycandane C, Boelsterli UA. Troglitazone-induced hepatic necrosis in an animal model of silent genetic mitochondrial abnormalities. Toxicol Sci. 2007;97:205–213. 108. Brookes PS, Morse K, Ray D, et al. The triterpenoid 2-cyano-3,12-dioxooleana-1,9dien-28-oic acid and its derivatives elicit human lymphoid cell apoptosis through a novel pathway involving the unregulated mitochondrial permeability transition pore. Cancer Res. 2007;67:1793–1802. 109. Zhang Y, Soboloff J, Zhu Z, Berger SA. Inhibition of Ca2+ inﬂux is required for mitochondrial reactive oxygen species-induced endoplasmic reticulum Ca2+ depletion and cell death in leukemia cells. Mol Pharmacol . 2006;70:1424–1434. 110. Rembish SJ, Trush MA. Further evidence that lucigenin-derived chemiluminescence monitors mitochondrial superoxide generation in rat alveolar macrophages. Free Radic Biol Med. 1994;17:117–126.

460

COMPARTMENTATION OF REDOX SIGNALING AND CONTROL

111. Meany DL, Poe BG, Navratil M, Moraes CT, Arriaga EA. Superoxide released into the mitochondrial matrix. Free Radic Biol Med. 2006;41:950–959. 112. Hanson GT, Aggeler R, Oglesbee D, et al. Investigating mitochondrial redox potential with redox-sensitive green ﬂuorescent protein indicators. J Biol Chem. 2004;279:13044–13053. 113. Dooley CT, Dore TM, Hanson GT, Jackson WC, Remington SJ, Tsien RY. Imaging dynamic redox changes in mammalian cells with green ﬂuorescent protein indicators. J Biol Chem. 2004;279:22284–22293. 114. Haddad JJ, Olver RE, Land SC. Antioxidant/pro-oxidant equilibrium regulates HIF-1alpha and NF-kappa B redox sensitivity. Evidence for inhibition by glutathione oxidation in alveolar epithelial cells. J Biol Chem. 2000;275:21130–21139. 115. Hirota K, Matsui M, Iwata S, Nishiyama A, Mori K, Yodoi J. AP-1 transcriptional activity is regulated by a direct association between thioredoxin and Ref-1. Proc Natl Acad Sci U S A. 1997;94:3633–3638. 116. Matthews JR, Kaszubska W, Turcatti G, Wells TN, Hay RT. Role of cysteine62 in DNA recognition by the P50 subunit of NF-kappa B. Nucleic Acids Res. 1993;21:1727–1734. 117. Jayaraman L, Murthy KG, Zhu C, Curran T, Xanthoudakis S, Prives C. Identiﬁcation of redox/repair protein Ref-1 as a potent activator of p53. Genes Dev. 1997;11:558–570. 118. Hirota K, Murata M, Sachi Y, Nakamura H, Takeuchi J, Mori K, Yodoi J. Distinct roles of thioredoxin in the cytoplasm and in the nucleus: a two-step mechanism of redox regulation of transcription factor NF-kappaB. J Biol Chem. 1999;274:27891–27897. 119. Marshall HE, Stamler JS. Inhibition of NF-kappa B by S-nitrosylation. Biochemistry. 2001;40:1688–1693. 120. Pineda-Molina E, Klatt P, Vazquez J, et al. Glutathionylation of the p50 subunit of NF-kappaB: a mechanism for redox-induced inhibition of DNA binding. Biochemistry. 2001;40:14134–14142. 121. Watson WH, Jones DP. Oxidation of nuclear thioredoxin during oxidative stress. FEBS Lett. 2003;543:144–147. 122. Bellomo G, Vairetti M, Stivala L, Mirabelli F, Richelmi P, Orrenius S. Demonstration of nuclear compartmentalization of glutathione in hepatocytes. Proc Natl Acad Sci U S A. 1992;89:4412–4416. 123. Voehringer DW, McConkey DJ, McDonnell TJ, Brisbay S, Meyn RE. Bcl-2 expression causes redistribution of glutathione to the nucleus. Proc Natl Acad Sci U S A. 1998;95:2956–2960. 124. Cotgreave IA. Analytical developments in the assay of intra- and extracellular GSH homeostasis: speciﬁc protein S-glutathionylation, cellular GSH and mixed disulphide compartmentalisation and interstitial GSH redox balance. Biofactors. 2003;17:269–277. 125. Go YM, Ziegler TR, Johnson JM, Gu L, Hansen JM, Jones DP. Selective protection of nuclear thioredoxin-1 and glutathione redox systems against oxidation during glucose and glutamine deﬁciency in human colonic epithelial cells. Free Radic Biol Med. 2007;42:363–370. 126. Hwang C, Sinskey AJ, Lodish HF. Oxidized redox state of glutathione in the endoplasmic reticulum. Science. 1992;257:1496–1502.

REFERENCES

461

127. Cuozzo JW, Kaiser CA. Competition between glutathione and protein thiols for disulphide-bond formation. Nat Cell Biol. 1999;1:130–135. 128. Bass R, Ruddock LW, Klappa P, Freedman RB. A major fraction of endoplasmic reticulum-located glutathione is present as mixed disulﬁdes with protein. J Biol Chem. 2004;279:5257–5262. 129. Tu BP, Ho-Schleyer SC, Travers KJ, Weissman JS. Biochemical basis of oxidative protein folding in the endoplasmic reticulum. Science. 2000;290:1571–1574. 130. Marciniak SJ, Ron D. Endoplasmic reticulum stress signaling in disease. Physiol Rev. 2006;86:1133–1149. 131. Szegezdi E, Logue SE, Gorman AM, Samali A. Mediators of endoplasmic reticulum stress-induced apoptosis. EMBO Rep. 2006;7:880–885. 132. Haynes CM, Titus EA, Cooper AA. Degradation of misfolded proteins prevents ER-derived oxidative stress and cell death. Mol Cell. 2004;15:767–776. 133. Cullinan SB, Zhang D, Hannink M, Arvisais E, Kaufman RJ, Diehl JA. Nrf2 is a direct PERK substrate and effector of PERK-dependent cell survival. Mol Cell Biol. 2003;23:7198–7209. 134. Zable AC, Favero TG, Abramson JJ. Glutathione modulates ryanodine receptor from skeletal muscle sarcoplasmic reticulum: evidence for redox regulation of the Ca2+ release mechanism. J Biol Chem. 1997;272:7069–7077. 135. Aghdasi B, Zhang JZ, Wu Y, Reid MB, Hamilton SL. Multiple classes of sulfhydryls modulate the skeletal muscle Ca2+ release channel. J Biol Chem. 1997;272:3739–3748. 136. Higo T, Hattori M, Nakamura T, Natsume T, Michikawa T, Mikoshiba K. Subtype-speciﬁc and ER lumenal environment-dependent regulation of inositol 1,4,5-trisphosphate receptor type 1 by ERp44. Cell. 2005;120:85–98. 137. Li Y, Camacho P. Ca2 + -dependent redox modulation of SERCA 2b by ERp57. J Cell Biol. 2004;164:35–46. 138. Basso E, Fante L, Fowlkes J, Petronilli V, Forte MA, Bernardi P. Properties of the permeability transition pore in mitochondria devoid of cyclophilin D. J Biol Chem. 2005;280:18558–18561. 139. Starkov AA, Polster BM, Fiskum G. Regulation of hydrogen peroxide production by brain mitochondria by calcium and Bax. J Neurochem. 2002;83:220–228. 140. Xu W, Liu L, Charles IG, Moncada S. Nitric oxide induces coupling of mitochondrial signalling with the endoplasmic reticulum stress response. Nat Cell Biol. 2004;6:1129–1134. 141. Scorrano L, Oakes SA, Opferman JT, et al. BAX and BAK regulation of endoplasmic reticulum Ca2+ : a control point for apoptosis. Science. 2003;300:135–139. 142. Oakes SA, Opferman JT, Pozzan T, Korsmeyer SJ, Scorrano L. Regulation of endoplasmic reticulum Ca2+ dynamics by proapoptotic BCL-2 family members. Biochem Pharmacol. 2003;66:1335–1340. 143. Jiang S, Moriarty-Craige SE, Orr M, Cai J, Sternberg P, Jr., Jones DP. Oxidant-induced apoptosis in human retinal pigment epithelial cells: dependence on extracellular redox state. Invest Ophthalmol Vis Sci. 2005;46:1054–1061.

19 ASSESSING MITOCHONDRIAL PROTEIN SYNTHESIS IN DRUG TOXICITY SCREENING Edward E. McKee Indiana University School of Medicine–South Bend, South Bend, Indiana

1. 2. 3. 4.

Introduction Antibiotics and mitochondrial protein synthesis Methods to assess the effects of drugs on mitochondrial protein synthesis Use of isolated intact mitochondria to assess drug susceptibility of mitochondrial protein synthesis 5. Mitochondrial isolation: quality and the respiratory control ratio 6. Mitochondrial isolation: choice of tissue

463 464 467 467 468 469

1. INTRODUCTION Mitochondrial protein synthesis is carried out in the matrix by a system unique from the cytoplasmic system. The mitochondrial proteome reﬂects a highly regulated and cooperative effort between expression of the mitochondrial and nuclear genomes (for a review, see [1]). The products of the mammalian mitochondrial genome include the small and large ribosomal RNAs, 22 tRNAs and 13 mRNAs, that include transcripts for seven subunits of complex I (NADH reductase), cytochrome b of complex II (bc1 complex), three subunits of complex IV (cytochrome c oxidase), and two membrane subunits of complex V (F1 FO -ATPAse) [2]. The nuclear genome encodes all of the enzymes and Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

463

464

ASSESSING MITOCHONDRIAL PROTEIN SYNTHESIS

protein cofactors required to carry out protein synthesis, including all of the tRNA synthetases, ribosomal proteins, and initiation and elongation factors. The nuclear genes of this system are transcribed in the nucleus, the mRNAs translated in the cytoplasm, and the proteins transported into the mitochondrial matrix. The mitochondrial protein synthesis system is unique in many respects (for a review, see [1]). Both strands of the genome are transcribed completely, from promoters in a region referred to as the D-loop, into two large polycistronic transcripts. These transcripts are subsequently cut at the 5 and 3 boundaries of tRNA sequences that nicely punctuate the genome and yield all of the products noted above. Outside a short poly(A) tail that is added to the mRNAs, there is no further processing of mitochondrial mRNAs. There are no introns and there are no intergene sequences. Unlike bacterial messages, mitochondrial mRNAs have no ribosomal binding site, but rather, start immediately with an AUG start and end with a termination codon. Mitochondria have also evolved several important changes in tRNA use, in which some tRNAs can read all four codons in a codon family, allowing the mitochondrial system to use just 22 tRNAs rather than the 30 or so required for other systems. As part of the alteration in tRNA use, mitochondria have evolved changes in the universal triplet code. The triplet codon UGA, which is a stop codon in all other systems, encodes tryptophan in all animal mitochondria. AUA encodes methionine rather than isoleucine, while AGA/AGG are stop codons in mammalian mitochondria rather than arginine. These changes support the reduced numbers of tRNAs by simplifying the reading of these codon boxes. However, because of these changes, it is not possible to translate mitochondrial mRNAs in nonmitochondrial protein-synthesizing systems.

2. ANTIBIOTICS AND MITOCHONDRIAL PROTEIN SYNTHESIS The mitochondrial protein synthesis system is more closely related to the bacterial endosymbiont from which mitochondria are thought to have evolved than to the cytoplasmic protein synthesis system [3–6]. As a result, antibiotics that bind to the bacterial ribosome and target bacterial protein synthesis may also bind to the mitochondrial ribosome and inhibit mitochondrial protein synthesis [5,7] (see Chapter 2). Toxicity caused by inhibition of mitochondrial protein synthesis is not immediately observed. Mitochondria turn over slowly in many tissues. Depending on the degree of inhibition, time is required for the amount of mitochondrial machinery and the ability to synthesize ATP to fall below a pathogenic threshold value. As a result, tissue toxicity caused by inhibition of mitochondrial protein synthesis can be difﬁcult to assess in typical preclinical in vivo toxicity studies. Classes of antibiotics that bind to the bacterial ribosome and inhibit bacterial protein synthesis include chloramphenicol, tetracyclines, aminoglycosides, macrolides, lincosamides, and most recently, the oxazolidinones.

ANTIBIOTICS AND MITOCHONDRIAL PROTEIN SYNTHESIS

465

Chloramphenicol Chloramphenicol is the oldest member of this group. It binds to the 50S bacterial ribosome as well as to a similar conserved site on the 39S mammalian mitochondrial ribosome. Chloramphenicol inhibition of mitochondrial protein synthesis has been well documented [8] with an IC50 in isolated rat liver and heart mitochondria of 9.8 ± 0.5 and 11.8 ± 0.6 µM, respectively [9], well within the therapeutic dose range (46 to 77 µM) [9]. Chloramphenicol has long been associated with three well-established toxicities: (1) a dose-dependent and reversible bone marrow depression [8]; (2) “gray baby syndrome” observed in infants given high doses of chloramphenicol (100 mg/kg per day); and (3) fatal aplastic anemia in certain genetically sensitive persons (1 in 25,000 to 40,000) [8,10] (see Chapter 11). Signiﬁcant evidence has established that the bone marrow depression and the gray baby syndrome are dose-dependent reversible toxic side effects of chloramphenicol therapy, caused by inhibition of mitochondrial protein synthesis (for a review, see [8,11]). The mechanism of the irreversible aplastic anemia has not been established. A unique genotoxic role for the nitroso group on chloramphenicol was suggested based on evidence that both chloramphenicol and the closely related thiamphenicol equally inhibited mitochondrial translation, but thiamphenicol was not initially associated with aplastic anemia [8,11]. However, the emergence of aplastic anemia in patients taking thiamphenicol has called this into question [11], and mitochondrial toxicity remains a possibility. As a result of these toxicities and the availability of other antibiotics, chloramphenicol is now used only rarely in life-threatening situations and as a topical antibiotic for treatment of eye infections. Tetracyclines The tetracyclines bind to the small subunits of both prokaryotic and mitochondrial ribosomes, and members of this class of antibiotics are also known to be potent inhibitors of mitochondrial protein synthesis [12–16]. The IC50 value of tetracycline inhibition of mitochondrial protein synthesis is 2.1 ± 0.5 µM in both isolated rat heart and liver mitochondria, again well within the human dosage range for this antibiotic (2.2 to 11 µM) [9]. Signiﬁcant toxicities have been noted for the tetracyclines [17], including anemia, thrombocytopenia, central nervous system effects, endocrine and metabolic effects, tooth discoloration, skin photosensitivity, renal and liver toxicity (especially liver steatosis), and myopathy. The etiology of many of these toxicities is poorly understood, and the degree that any of the tetracycline toxicities are caused by mitochondrial dysfunction is unknown. However, many of the reported toxicities are consistent with mitochondrial dysfunction. For example, the tetracycline-associated multisystem disease described in a case report by Fox et al. [18] is analogous to the multisystem disorders associated with known mitochondrial diseases [19] (see Chapter 11). Aminoglycosides The aminoglycosides bind to the small bacterial ribosome and potentially to a similar site on the small mitochondrial ribosome [7,20]. Members of this class of antibiotic are known to concentrate and persist in the endolymph and perilymph of the inner ear and to cause ototoxicity when given at high

466

ASSESSING MITOCHONDRIAL PROTEIN SYNTHESIS

concentrations. Interestingly, a single polymorphic base pair change in the small mitochondrial ribosomal RNA has been implicated in a form of maternally inherited ototoxicity and nerve deafness associated with streptomycin treatment [20,21] (see also Chapters 4 and 11). Evidence has shown that mitochondrial rRNA containing this polymorphic base pair change binds streptomycin with high afﬁnity in vitro, whereas mitochondrial rRNA with the typical base-pair sequence does not [22]. Analogously, the ototoxicity observed in the more general population may be related either to this polymorphism or to other polymorphisms in human mitochondrial rRNA, causing inhibition of mitochondrial protein synthesis within the hair cells of the ear. Treatment of isolated rat heart and liver mitochondria with streptomycin had no effect on mitochondrial protein synthesis [9], suggesting that rat mitochondrial ribosomes may not bind streptomycin as avidly as the human ribosome. Oxazolidinones The most recent addition to the antibiotic drugs are the oxazolidinones, with linezolid approved for use in 2000 [23]. These compounds bind to the large bacterial ribosome at a site within the peptidyl transferase center that overlaps the chloramphenicol binding site and inhibits bacterial protein synthesis [24–26]. The oxazolidinones have also been shown to bind to the same site on human mitochondrial ribosomes [25] and to inhibit mitochondrial protein synthesis in human cells [27] and in mitochondria isolated from rat heart and liver and from rabbit heart and bone marrow [9]. The IC50 noted for linezolid in isolated rat heart and liver mitochondrial studies was 12.8 ± 2.8 µM, well within the human therapeutic dose range of 18 to 60 µM. Dose-dependent and reversible bone marrow suppression was noted early as a side effect of treatment with linezolid [28]. In a recent study of 44 patients on linezolid therapy for serious infections, the authors reported good clinical outcomes (73% cure rate) but a high number of adverse reactions, including thrombocytopenia (30%), anemia (16%), pancytopenia (5%), and single instances of peripheral neuropathy and lactic acidosis [29]. Many of these 32 patients were on linezolid therapy for a period of >2 weeks (8 to 185 day range, 29 ± 28 day mean). All of these toxicities are consistent with inhibition of mitochondrial protein synthesis. Finally, our laboratory demonstrated that oxazolidinones that were much more potent than linezolid as an antibiotic were also much more potent at inhibiting mitochondrial protein synthesis and were much more toxic in rat studies [9]. Lincosamides and Macrolides The remaining two classes of antibiotics that function by inhibiting bacterial protein synthesis, the lincosamides and macrolides, have never been shown to have an effect on mitochondrial protein synthesis, even at very high concentrations [9,13]. They are also not associated with any discernible toxicity. These antibiotics bind to the large bacterial ribosomal subunit at a site that requires an A at position 2058 [7,30]; a G in this position confers bacterial resistance. Since the equivalent base on the 16S rRNA in mammalian mitochondria is G [7], the lincosamides and macrolides do not inhibit mitochondrial protein synthesis because they do not bind to the mitochondrial ribosome.

USE OF ISOLATED INTACT MITOCHONDRIA

467

3. METHODS TO ASSESS THE EFFECTS OF DRUGS ON MITOCHONDRIAL PROTEIN SYNTHESIS A variety of methods have been used to demonstrate antibiotic toxicity to mitochondrial protein synthesis. A number of in vivo studies have shown that the activity and amount of complexes I, III, and IV, which contain mitochondrially encoded subunits, are signiﬁcantly reduced in linezolid-treated cultured human cells [27], in cells and tissues from treated patients [31,32], and in treated rats [32] (see also Chapter 16). Mitochondrial DNA and other mitochondrial protein complexes were unaffected by linezolid in these studies, suggesting that synthesis of mitochondrial products was affected speciﬁcally. Another method used to assess toxicity is to quantitate the level of mitochondrial translation products by in vivo labeling using subunit-speciﬁc antibodies [33]. A reduction in labeling of mitochondrially translated subunits compared to controls, while labeling of cytoplasmically translated subunits go up or remain unchanged indicates an effect speciﬁcally on the mitochondrial system (see Chapter 16). Further, coupling the quantitation of protein subunits with quantitation of their mitochondrial mRNAs will determine whether or not the effect is at the level of translation. One study in patients with linezolid-associated hyperlactatemia demonstrated a signiﬁcant reduction in complex IV activity, together with a highly reduced amount of the mitochondrially synthesized subunit II [31]. The levels of subunit II mRNA were increased substantially in this study, showing that inhibition was clearly at the level of mitochondrial translation. The increase in subunit II mRNA was observed without any change in the level of mtDNA, and may represent a compensatory mechanism of the cell acting at the level of transcription [31]. Although these data clearly demonstrate that linezolid inhibited mitochondrial translation, these methods are too costly and tedious to employ in routine drug toxicity assessment. In contrast to in vivo methods, others have studied antibiotic susceptibility of mitochondrial translation using in vitro methods that employ puriﬁed components of the mitochondrial translation system, such as mitochondrial ribosomes or rRNA constructs [22]. Several groups have studied mitochondrial antibiotic susceptibility using a membrane-free mitochondrial protein synthesizing system programmed with polyU RNA [5,13]. These methods have the advantage of studying drug effects in the absence of the highly impermeable inner mitochondrial membrane and mitochondrial metabolic pathways, and are particularly useful in binding studies. However, these methods are inappropriate for routine drug toxicity assessment precisely because they lack an intact inner membrane and the other mitochondrial pathways, either of which could have a major impact on the matrix availability and toxicity of a drug. 4. USE OF ISOLATED INTACT MITOCHONDRIA TO ASSESS DRUG SUSCEPTIBILITY OF MITOCHONDRIAL PROTEIN SYNTHESIS The best method for the routine assessment of drug effects on mitochondrial protein synthesis is the use of isolated, high-quality, intact mitochondria. When

468

ASSESSING MITOCHONDRIAL PROTEIN SYNTHESIS

placed in an appropriate medium, these organelles are capable of all aspects of mitochondrial metabolism, including membrane transport, substrate and oxygen consumption, electron transport, and oxidative phosphorylation. Isolated mitochondria can also carry out processes of biogenesis, including DNA replication, mRNA transcription, and protein translation. A drug added to the medium of such a system should interact with the mitochondria in a manner essentially identical to its interaction in an intact tissue. For example, if a drug is not transported across the inner membrane in an intact tissue, it is unlikely to cross the inner membrane in the isolated mitochondrial system. An isolated intact mammalian mitochondrial protein-synthesizing system was characterized and optimized in our laboratory [34,35] and has subsequently been used to study antibiotic toxicity [9,27]. The details of this method have been previously described [9]. Brieﬂy, mitochondria are isolated from a tissue such as rat heart and incubated in a protein synthesizing medium with [35 S]methionine and with various concentrations of the drug to be screened. Aliquots of the incubation are taken over a time course of 1 to 2 hours and spotted onto ﬁlter paper disks that are dropped into 5% trichloroacetic acid +5 mM methionine. The disks are washed in a batchwise process and the amounts of radioactivity on the disks are quantitated by liquid scintillation counting. To account for variations in the rate of mitochondrial protein synthesis in different mitochondrial preparations, the rate of [35 S]methionine incorporation in the drug-treated samples is expressed as a percent of the rate for each preparation in the vehicle control samples. Dose–response curves are prepared by plotting the percent of vehicle control of the drug-treated mitochondria as a function of drug concentration. The day-to-day reproducibility of the results is remarkably high [9]. This system is typically capable of linear rates of protein synthesis for up to 2 hours of incubation. Incubation volumes can be kept quite low (20 to 75 µL), so very little drug is needed for this assay. The ﬁlter paper disk technique is amenable to a high number of samples. Reversibility of inhibition can be demonstrated with this system by removing the drug by centrifugation, followed by resuspension of the washed mitochondria and measurement of subsequent protein synthesis. 5. MITOCHONDRIAL ISOLATION: QUALITY AND THE RESPIRATORY CONTROL RATIO The quality of the mitochondrial preparation and the resulting rate of mitochondrial protein synthesis are related. An excellent measure of mitochondrial quality is the respiratory control ratio (RCR) [36]. The respiratory control ratio is a measure of the rate of oxygen consumption during active ATP synthesis (state III) divided by the slower rate observed in the absence of ATP synthesis in resting mitochondria (state IV). This ratio is an index of the intactness of the mitochondrial preparation. A low ratio is indicative of “leaky” mitochondria that have difﬁculty in maintaining proton gradients when in the resting state and increase electron transport and oxygen consumption to compensate for the leak. For isolated heart mitochondria, a ratio >6 is accepted as a highly intact and coupled

MITOCHONDRIAL ISOLATION: CHOICE OF TISSUE

469

preparation. In previous work we have shown that rates of isolated mitochondrial protein synthesis are highest in mitochondria with high respiratory control ratios [34]. However, high-quality mitochondria are difﬁcult to isolate from some tissues. In our hands the respiratory control ratio observed for mitochondria isolated from rabbit bone marrow was typically around 3 [9]. Associated with this lower respiratory ratio was a rate of mitochondrial protein synthesis that was linear for only 30 minutes and reached only half the rate observed in mitochondria from heart and liver [9]. In other work, we [34] and others [37] have demonstrated that mitochondrial protein synthesis is regulated by the energy charge of the mitochondria and is dependent on the mitochondrial membrane potential. A variety of inhibitors that are known to affect mitochondrial pathways, such as uncouplers (dinitrophenol, DNP), tricarboxylic acid cycle inhibitors (arsenate), ADP/ATP translocase inhibitor (atractyloside), and ATP synthase inhibitor (oligomycin), are all associated with decreased mitochondrial protein synthesis ([9,37] and unpublished observations). Thus, if a drug screened in this system has no effect on mitochondrial protein synthesis, it is also unlikely to affect a variety of other mitochondrial pathways. Alternatively, drugs that have even modest effects on these mitochondrial pathways, when given over long periods of time, may eventually display profound mitochondria toxicity in tissues with high ATP demand.

6. MITOCHONDRIAL ISOLATION: CHOICE OF TISSUE A question that arises in studies with isolated mitochondria is the appropriate choice of tissue for mitochondrial isolation [9]. The rRNA of the mitochondrial ribosome is the same in all tissues of a species, and mitochondrial ribosomal protein genes appear to be single-copy genes that do not display tissue speciﬁcity, suggesting that the basic mitochondrial ribosome is identical in all tissues. This implies that antibiotics that bind to the mitochondrial ribosome and inhibit mitochondrial protein synthesis would inhibit synthesis more or less equally in all cells and could cause pathology in many tissues. However, the side effects noted for chloramphenicol and linezolid appeared ﬁrst in the bone marrow compartment. We addressed this question in recent work by studying the effects of antibiotics on mitochondrial protein synthesis in a variety of tissues, including rat heart and liver, and rabbit heart and bone marrow, and demonstrated that antibiotic effects were remarkably similar in all of the tissues studied. It seems likely that the preferential toxicity noted for chloramphenicol and linezolid reﬂects the increased role of mitochondrial protein synthesis in actively multiplying cells, such that pathology is observed there ﬁrst. While mitochondria from a variety of tissues can be used to study drug susceptibility of mitochondrial protein synthesis, we have routinely chosen rat heart as a convenient and reproducible source of high-quality mitochondria. A remaining question in the choice of mitochondria used for drug assays concerns the potential differences between humans and other mammals. While

470

ASSESSING MITOCHONDRIAL PROTEIN SYNTHESIS

the mammalian mitochondrial ribosome has some differences compared to the lower eukaryotic mitochondrial ribosomes, they are highly similar among mammals [38], and effects of chloramphenicol and linezolid in rat, rabbit [9], and human cell mitochondria [27] have been identical. Thus, the rat mitochondrial ribosome appears to be an excellent substitute for the human mitochondrial ribosome for general antibiotic susceptivity studies. However, an important problem that has not been addressed adequately is the presence of polymorphisms within the mitochondrial rRNAs of human populations, such as the one described above associated with maternally inherited streptomycin-induced deafness [20–22]. The association of polymorphisms in the mitochondrial 16S rRNA has also been described in several patients with linezolid-induced lactic acidosis [39,40] (see also Chapters 4 and 11). Whereas pharmacogenomic techniques are available for screening human populations for mitochondrial rRNA polymorphisms from blood samples, the ability to isolate mitochondria and test antibiotic susceptibility in blood samples remains to be established and would be an important contribution to solving this problem.

REFERENCES 1. Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P. Molecular Biology of the Cell . New York: Garland Science; 2002. 2. Chomyn A, Cleeter MW, Ragan CI, Riley M, Doolittle RF, Attardi G. URF6, last unidentiﬁed reading frame of human mtDNA, codes for an NADH dehydrogenase subunit. Science. 1986;234:614–618. 3. Brien T. Properties of human mitochondrial ribosomes. IUBMB Life. 2003; 55:505–513. 4. Spremulli LL, Coursey A, Navratil T, Hunter SE. Initiation and elongation factors in mammalian mitochondrial protein biosynthesis. Prog Nucleic Acid Res Mol Biol. 2004;77:211–261. 5. Zhang L, Ging NC, Komoda T, Hanada T, Suzuki T, Watanabe K. Antibiotic susceptibility of mammalian mitochondrial translation. FEBS Lett. 2005;579:6423–6427. 6. Mears JA, Sharma MR, Gutell RR, et al. A structural model for the large subunit of the mammalian mitochondrial ribosome. J Mol Biol. 2006;358:193–212. 7. Bottger EC. Antimicrobial agents targeting the ribosome: the issue of selectivity and toxicity—lessons to be learned. Cell Mol Life Sci. 2007;64:791–795. 8. Yunis AA. Chloramphenicol toxicity: 25 years of research. Am J Med. 1989; 87:44N–48N. 9. McKee EE, Ferguson M, Bentley AT, Marks TA. Inhibition of mammalian mitochondrial protein synthesis by oxazolidinones. Antimicrob Agents Chemother. 2006;50:2042–2049. 10. Feder HM, Jr. Chloramphenicol: what we have learned in the last decade. South Med J. 1986;79:1129–1134. 11. Turton JA, Andrews CM, Havard AC, Robinson S, York M, Williams TC, Gibson FM. Haemotoxicity of thiamphenicol in the BALB/c mouse and Wistar Hanover rat. Food Chem Toxicol. 2002;40:1849–1861.

REFERENCES

471

12. Riesbeck K, Bredberg A, Forsgren A. Ciproﬂoxacin does not inhibit mitochondrial functions but other antibiotics do. Antimicrob Agents Chemother. 1990;34:167–169. 13. Ibrahim NG, Burke JP, Beattie DS. The sensitivity of rat liver and yeast mitochondrial ribosomes to inhibitors of protein synthesis. J Biol Chem. 1974;249:6806–6811. 14. van den Bogert C, Holtrop M, Melis TE, Roefsema PR, Kroon AM. Different effects of oxytetracycline and doxycycline on mitochondrial protein synthesis in rat liver after long-term treatment. Biochem Pharmacol. 1987;36:1555–1559. 15. Kroon AM, Dontje BH, Holtrop M, van den Bogert C. The mitochondrial genetic system as a target for chemotherapy: tetracyclines as cytostatics. Cancer Lett. 1984;25:33–40. 16. van den Bogert C, Kroon AM. Tissue distribution and effects on mitochondrial protein synthesis of tetracyclines after prolonged continuous intravenous administration to rats. Biochem Pharmacol. 1981;30:1706–1709. 17. Shapiro LE, Knowles SR, Shear NH. Comparative safety of tetracycline, minocycline, and doxycycline. Arch Dermatol. 1997;133:1224–1230. 18. Fox SA, Berenyi MR, Straus B. Tetracycline toxicity presenting as a multisystem disease. Mt Sinai J Med. 1976;43:129–135. 19. Finsterer J. Mitochondriopathies. Eur J Neurol. 2004;11:163–186. 20. Cortopassi G, Hutchin T. A molecular and cellular hypothesis for aminoglycosideinduced deafness. Hear Res. 1994;78:27–30. 21. Prezant TR, Agapian JV, Bohlman MC, et al. Mitochondrial ribosomal RNA mutation associated with both antibiotic-induced and non-syndromic deafness. Nat Genet. 1993;4:289–294. 22. Hamasaki K, Rando RR. Speciﬁc binding of aminoglycosides to a human rRNA construct based on a DNA polymorphism which causes aminoglycoside-induced deafness. Biochemistry. 1997;36:12323–12328. 23. Ford CW, Zurenko GE, Barbachyn MR. The discovery of linezolid, the ﬁrst oxazolidinone antibacterial agent. Curr Drug Targets Infect Disord. 2001;1:181–199. 24. Colca JR, McDonald WG, Waldon DJ, et al. Cross-linking in the living cell locates the site of action of oxazolidinone antibiotics. J Biol Chem. 2003;278:21972–21979. 25. Leach KL, Swaney SM, Colca JR, et al. The site of action of oxazolidinone antibiotics in living bacteria and in human mitochondria. Mol Cell. 2007;26:393–402. 26. Lin AH, Murray RW, Vidmar TJ, Marotti KR. The oxazolidinone eperezolid binds to the 50S ribosomal subunit and competes with binding of chloramphenicol and lincomycin. Antimicrob Agents Chemother. 1997;41:2127–2131. 27. Nagiec EE, Wu L, Swaney SM, Chosay JG, Ross DE, Brieland JK, Leach KL. Oxazolidinones inhibit cellular proliferation via inhibition of mitochondrial protein synthesis. Antimicrob Agents Chemother. 2005;49:3896–3902. 28. Gerson SL, Kaplan SL, Bruss JB, Le V, Arellano FM, Hafkin B, Kuter DJ. Hematologic effects of linezolid: summary of clinical experience. Antimicrob Agents Chemother. 2002;46:2723–2726. 29. Bishop E, Melvani S, Howden BP, Charles PG, Grayson ML. Good clinical outcomes but high rates of adverse reactions during linezolid therapy for serious infections: a proposed protocol for monitoring therapy in complex patients. Antimicrob Agents Chemother. 2006;50:1599–1602.

472

ASSESSING MITOCHONDRIAL PROTEIN SYNTHESIS

30. Pﬁster P, Corti N, Hobbie S, Bruell C, Zarivach R, Yonath A, Bottger EC. 23S rRNA base pair 2057–2611 determines ketolide susceptibility and ﬁtness cost of the macrolide resistance mutation 2058A → G. Proc Natl Acad Sci U S A. 2005;102:5180–5185. 31. Garrabou G, Soriano A, Lopez S, et al. Reversible inhibition of mitochondrial protein synthesis during linezolid-related hyperlactatemia. Antimicrob Agents Chemother. 2007;51:962–967. 32. De Vriese AS, Coster RV, Smet J, et al. Linezolid-induced inhibition of mitochondrial protein synthesis. Clin Infect Dis. 2006;42:1111–1117. 33. Chomyn A. In vivo labeling and analysis of human mitochondrial translation products. Methods Enzymol. 1996;264:197–211. 34. McKee EE, Grier BL, Thompson GS, Leung AC, McCourt JD. Coupling of mitochondrial metabolism and protein synthesis in heart mitochondria. Am J Physiol. 1990;258:E503–E510. 35. McKee EE, Grier BL, Thompson GS, McCourt JD. Isolation and incubation conditions to study heart mitochondrial protein synthesis. Am J Physiol. 1990;258:E492–E502. 36. Chance B, Williams GR. Respiratory enzymes in oxidative phosphorylation: I. Kinetics of oxygen Utilization. J Biol Chem. 1955;217:383–394. 37. Cote C, Boulet D, Poirier J. Expression of the mammalian mitochondrial genome: role for membrane potential in the production of mature translation products. J Biol Chem. 1990;265:7532–7538. 38. O’Brien TW. Properties of human mitochondrial ribosomes. IUBMB Life. 2003;55:505–513. 39. Carson J, Cerda J, Chae JH, Hirano M, Maggiore P. Severe lactic acidosis associated with linezolid use in a patient with the mitochondrial DNA A2706G polymorphism. Pharmacotherapy. 2007;27:771–774. 40. Palenzuela L, Hahn NM, Nelson RP, Jr, et al. Does linezolid cause lactic acidosis by inhibiting mitochondrial protein synthesis? Clin Infect Dis. 2005;40:E113–E116.

20 MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS: A CHALLENGE TO ACCURATE DIAGNOSIS Michel P. de Baar and Anthony de Ronde Primagen, Amsterdam, The Netherlands

1. 2. 3. 4. 5.

Introduction Current treatments Lactic acidosis Lipodystrophies Mitochondrial dysfunction and antiviral therapy: the DNA polymerase γ hypothesis 6. In vitro assessment of mtDNA depletion 7. Technologies for measuring mtDNA 8. Conclusions

473 474 476 476 478 480 481 483

1. INTRODUCTION The introduction of highly active antiretroviral therapy (HAART) in 1996 dramatically decreased morbidity and mortality, transforming HIV-1 infection into a chronic disease requiring life-long treatment. However, soon after its introduction, adverse effects of HAART were recognized, forcing changes in treatment regimes in more than 50% of patients [1]. The diversity and range of

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

473

474

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

adverse effects is large and includes peripheral neuropathy [2–5], cardiomyopathy [6–10], hepatic steatosis [11–16] or hepatotoxicity [17–19], lactic acidosis [14,20–25], type 2 diabetes [26–28], and lipodystrophy [29–36]. The evidence is compelling that depletion of mitochondrial DNA (mtDNA) due to coincident inhibition of polymerase γ (pol-g), is the major underlying cause of these toxicities [37,38]. MtDNA depletion has been demonstrated in cells derived from the blood as well as from affected tissues of treated patients, supporting this contention [6,39–51]. Recently introduced drugs have been screened during preclinical development for their effects on mtDNA depletion. Some of them are currently being evaluated in patients [44,52–56], so we will soon know whether circumventing mtDNA depletion will moderate deleterious side effects. In this review we focus on the mitochondrial toxicity of antiviral drugs, their adverse effects, and the challenges we face in monitoring mtDNA effects in order to optimize therapy for each patient.

2. CURRENT TREATMENTS Human immunodeﬁcieny virus type 1 (HIV-1), which was identiﬁed only 25 years ago, has infected more than 60 million people, and an estimated 40 million people are currently living with HIV/AIDS [57]. The usual HAART regimen targets the HIV-1 reverse transcriptase (RT) and HIV-1 protease by combining three or more different drugs from different classes, such as two nucleoside reverse transcriptase inhibitors (NRTIs) and a protease inhibitor (PI), or two NRTIs and a nonnucleoside reverse transcriptase inhibitor (NNRTI), or other combinations. Zidovudine, known as AZT, was the ﬁrst NRTI approved for the treatment of HIV by the U.S. Food and Drug Administration (FDA) in 1985 and by the European Medicines Agency (EMEA) in 1986. It was soon followed by didanosine (ddII), zalcitabine (ddC), stavudine (d4T), and lamivudine (3TC). The number of NRTIs available has increased with tenofovir (TDV), a nucleotide analog, and emtricitabine (FTC) is the most recently approved NRTI. Development of novel NRTIs continues, prompted by growing resistance of the virus, and the high-incidence toxicity for the approved NRTIs. For example, some NRTIs currently in development are AVX954 (Avexa Pharmaceuticals, Australia), 204937 (GlaxoSmithKline, UK), elvucitabine (Achillion Pharmaceuticals, United States), and MIV310 (Medivir, Sweden), among others. In 1996, the NNRTIs were introduced, which were powerful drugs inhibiting the reverse transcriptase but via a mechanism different from the NRTIs. Unlike the NRTIs, which are nucleoside analogs that bind to the active site of the polymerase, the NNRTIs bind to a hydrophobic pocket near the active site, and are therefore noncompetitive inhibitors. The ﬁrst NNRTI introduced was nevirapine, soon followed by efavirenz. Several new NNRTIs are currently in development, such as TMC125 (Tibotec, Belgium) and 695634 (GlaxoSmithKline, UK).

CURRENT TREATMENTS

475

The NNRTIs demonstrated relatively acute adverse effects, occurring mainly during the ﬁrst 12 to16 weeks of therapy, such as skin hypersensitivity, including the Stevens–Johnson syndrome (a milder form of toxic epidermal necrolysis), and hepatotoxicity, which occasionally progresses to liver failure. Efavirenz has been associated with persistent neuropsychiatric disorders [59–61]. A positive side effect of the NNRTI nevirapine is the elevation of HDL (high-density lipoprotein), which can moderate the development of heart disease [62]. Rather than blocking viral RNA transcription, the protease inhibitors (PIs) inhibit HIV-1 protease, the viral enzyme that cleaves nascent proteins needed for the assembly of new virons. Saquinavir, approved in 1995, was the ﬁrst in the class that now also includes ritonavir, indinavir, atazanavir, amprenavir, and several others. The PIs were introduced later than the NRTIs, which fostered the perception that most of the adverse effects of HAART were due to PIs [58]. Although the PIs do have their own adverse effects, the majority of the adverse effects observed during antiretroviral therapy are due to NRTIs. Indeed, the PIs prolonged patient lives so successfully that the long-term adverse effects of the NRTIs were ﬁnally revealed. Recently, several new classes of antiviral drugs have been introduced, such as inhibitors that block the binding and fusion required for the virus to enter the host cell. These include enfuvirtide (T20, Roche, Switzerland) which speciﬁcally inhibits the fusion between virus and host cell, and maraviroc (Pﬁzer, United States), which inhibits the binding of the virus with one of its receptors on the host cell, CCR5. The newest class of drugs in development comprise the integrase inhibitors, which are currently in late stage of clinical development (e.g., MK-0518, Merck USA). Potential adverse effects of these newer classes of drugs are not yet known. Patients are now commonly treated with a combination of three or more chemotherapeutic drugs (HAART) from several classes of drugs. Typically, two or more NRTIs are given in combination with NNRTI or with a PI, but other combinations are also used. After the introduction of PIs, HIV could effectively be suppressed and the lives of the patients extended so that adverse effects frequently became noticed. Since effective therapies have become available, the number of patients being treated has increased enormously, so one might expect more frequent adverse events based on the numbers alone. NRTIs and PIs are associated with a wide range of toxicities, many of which are also caused or exacerbated by HIV-1 itself, making it difﬁcult to separate the effect of the drug from the progression of the disease. The adverse effects that are attributed predominantly to NRTIs include peripheral neuropathy, cardiomyopathy, hepatic steatosis and hepatotoxicity, hyperlactatemia and lactic acidosis, type 2 diabetes or insulin resistance, and lipodystrophy [63]. Peripheral neuropathy has been observed in 10 to 30% of patients receiving ddC, ddI, or d4T treatment [29]. Myopathy has been reported in approximately 17% of patients receiving AZT, with cardiomyopathy occurring only in rare cases [64,65].

476

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

500 450 400 350 300 250 200 150 100 50 0

11 10 9 8 7 6 5 4 3 2 1 0

: Therapy with ddI/d4T/NFV

Lactate (mmol/L)

Copies mtDNA/cell

Stop therapy

360

330

300

270

240

210

180

150

120

90

60

30

0

−30

−60

Time after start of therapy (days)

Figure 1 Results of combination therapy. Graph presented by Anthony de Ronde at the International AIDS Conference, July 2002, Barcelona, Spain. (See insert for color representation of ﬁgure.)

3. LACTIC ACIDOSIS One of the most severe and potentially life-threatening side effects observed is severe lactic acidosis accompanied by massive hepatic steatosis, which fortunately, is a rare occurrence. A hallmark of mitochondrial dysfunction is hyperlactatemia, either symptomatic or asymptomatic, caused by increased lactate/pyruvate ratios [12,20,22]. This is due in part to impairment of mitochondrial ability to oxidize lactate, but is also exacerbated by increased glycolytic ﬂux as a metabolic compensation for the loss of oxidative phosphorylation (OXPHOS). This is a more common side effect, occurring in about 15% of patients receiving ddI, d4T, or AZT. Although these NRTIs are no longer among the ﬁrst-line therapies in developed countries, they are in resource-restricted countries, as recommended by the World Health Organization [66]. The current diagnosis of lactic acidosis is at least two consecutive measurements of plasma lactate levels elevated above 2.2 mM (Figure 1) [67]. Assessments of lactate alone, however, are not highly predictive of NRTI toxicity because lactate accumulates as a function of exercise prior to venipuncture, causing false positives. The requirements for cooled collection tubes, expeditious delivery to the chemistry laboratory, and prompt completion of the assay tend to increase variability, which undermines the utility of lactate as a diagnostic or prognostic marker [68]. 4. LIPODYSTROPHIES The occurrence of fat redistribution lipodystrophy has been reported in many HIV patients, but the incidence is dramatically higher in patients receiving HAART

LIPODYSTROPHIES

477

versus naive patients, linking it strongly to drug exposure [29,37,69–72]. The lipodystrophy syndrome is characterized by a loss of subcutaneous fat from the extremities and the face, coupled with an excess fat deposition in the breasts, abdominal cavity, and trunk, and in rarer cases, the neck (“buffalo hump”) (Figure 2). Diagnosis of the lipodystrophy syndrome has been approached in various ways, the easiest being physicial examination and self-monitoring, but a more objective and quantitative approach that uses a number of parameters to calculate lipodystrophy severity is gaining acceptance [73]. It is increasingly recognized

Figure 2 Changes in appearance associated with lipodystrophy include central abdominal obesity, gynecomastia, fat accumulation behind the neck, and fat loss in the face, arms, legs, or buttocks.

478

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

that lipodystrophy is actually the net result of two syndromes, lipoatrophy and lipohypertrophy, each probably having a different aetiology [74]. Based on rat studies, in 2003 a hypothesis was proposed where the sympathetic and parasympathetic nervous systems control the balance between lipoatrophy and lipohypertrophy [75]. The state of neural stimulation of the adipose tissue could change in response to NRTI therapy, resulting in lipoatrophy or lipohypertrophy. If such neural changes are inﬂuencing the lipodystrophy syndrome, its pathogenesis would be as similar as the aetiology behind peripheral neuropathy, a frequently observed adverse effect of NRTI therapy [75]. The current model of lipodystrophies holds that prolonged NRTI use (adverse effects correlate with duration of exposure) leads to a decline of mtDNA that eventually yields mitochondrial dysfunction and adipocyte apoptosis. Macrophages remove the apoptotic adipocytes, with the net result of a decline in subcutaneous fat and lipoatropy of the extremities and the face (Figure 2). The mechanisms that govern the lipohypertrophy of the abdomen and neck (buffalo hump) (Figure 2) are less well understood. It seems likely, however, that degradation of the subcutaneous fat tissue and other disorders of energy metabolism due to mitochondrial dysfunction result in increased circulating levels of fatty acids, which are then stored in the visceral fat compartments.

5. MITOCHONDRIAL DYSFUNCTION AND ANTIVIRAL THERAPY: THE DNA POLYMERASE γ HYPOTHESIS These diverse adverse effects are thought to result from chronic drug exposure that impairs mitochondria replication. This notion is based on in vitro and in vivo studies demonstrating impaired mitochondrial function after NRTI exposure and the fact that many of the toxicities reported are also symptoms of inherited and/or acquired mitochondrial diseases (see Chapters 4 and 11). The evidence for drug-induced mitochondrial dysfunction in vitro and in vivo is accumulating, but is already compelling. The effect of NRTIs on mitochondria is in agreement with the DNA polymerase γ hypothesis of Lewis [38], which was extended by Brinkman and colleagues [37] (Figure 3). This model proposes that in addition to inhibiting the HIV-1 reverse transcriptase, to various extents the NRTIs also inhibit mitochondrial DNA polymerase γ required for mitochondrial replication [76–78]. Depletion of mtDNA and/or accumulation of mutated/truncated mtDNA would result in impaired synthesis of respiratory chain complexes involved in ATP production, thereby impairing energy production. Such a loss of physiological scope has deleterious consequences on cellular structure and function and is considered responsible for many of the long-term toxic side effects of NRTIs (see Chapter 2). Tissue-speciﬁc adverse events from drug treatment reﬂects differential afﬁnity and off-rates of the various NRTIs for DNA polymerase γ, combined with organelle-speciﬁc nucleotide pools, plus kinase activities and bioenergetic demand of the various cell types [38,76,78,79]. The individuals’ genotype also

479

MITOCHONDRIAL DYSFUNCTION AND ANTIVIRAL THERAPY

Mitochondrial DNA-pol-g

NRTIs

Glucose

free fatty acids

Energy production

Fat metabolism

ATP Glucose

Pyruvate

ATP Pyruvate

NADH NAD+

Krebs cycle

Acetyl-CoA

lactate

Mitochondrion

serum lactate

lipid droplets

Krebs cycle b-oxidation Acetyl-CoA of fatty acids

Mitochondrion

free fatty acids

Figure 3 Relation between mitochondrial DNA and mitochondrial function: Due to NRTI treatment, DNA polymerase γ inhibition leads to depletion of mtDNA and mtDNA-encoded proteins, thereby impairing mitochondrial function. Ultimately, this leads to changes in energy production and fat metabolism with increased serum lactate levels and increased free fatty acids. (See insert for color representation of ﬁgure.)

modulates sensitivity to the toxic effects of the compounds, and this remains a focus of ongoing study. For example, variations in the TNF-α promoter are associated with increased risk for lipoatrophy [80,81]. As might be expected, variations in the DNA polymerase γ also contribute to susceptibility to NRTI mitochondrial toxicity [82]. Changes in mtDNA in affected cells or tissues can be monitored as a marker of toxicity [47,48,83–85], but these changes can also be detected at the level of peripheral blood mononuclear cells (PBMCs) [46,50,68,86–89], especially with continuous exposure to the drugs (see below). The effect of PIs on mitochondrial function is less straightforward than the inhibition of mtDNA synthesis by NRTIs. A possible, but as yet not completely elucidated mechanism of PIs may be via inhibition of the coordinated synthesis of the respiratory chain complexes [90,91]. PIs are also thought to affect the action of the glucose transporter GLUT4, sterol regulatory element binding protein SREBP1, and peroxisome proliferator-activated receptor (PPAR) γ, among others, precipitating in non-insulin-dependent diabetes [26,90,92,93,93–99]. Oxidative stress probably plays a role in drug-induced mitochondrial dysfunction. Reactive oxygen species (ROS) are produced during oxidative phosphorylation, and an improperly functioning respiratory chain is thought to increase ROS production [100–103]. Indeed, prolonged NRTI treatment of mice and rats leads

480

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

to increased ROS production and oxidative damage to mtDNA and lipids, plus glutathione depletion in skeletal and cardiac muscle. Inhibitory effects of NRTIs on the adenine nucleotide translocator NADH–cytochrome reductase, NADH oxidase, adenylate kinase, and the succinate transport system have been observed. In addition, HIV-1 itself may lead to increased ROS production [104–106]. It has also been suggested that NRTIs directly decrease the activity of Krebs cycle enzymes and play a role in triggering mitochondrial-induced apoptosis. Mitochondrial function serves as a sensor for the viability of the cell and mitochondrial proteins (e.g., cytochrome c) are involved in the fas–bcl2 cascade of inducing apoptosis [104,107–114]. Loss of mitochondrial function may lead to programmed cell death, and this may contribute to the loss of fat tissue in lipoatrophy. Such increased apoptosis, together with the decreases in mtDNA in the remaining cells, will together affect the adipose tissue even more severely than either separately. 6. IN VITRO ASSESSMENT OF mtDNA DEPLETION Changes in mtDNA reﬂect toxicity in cells and organs [47,48,83–85], but can also be detected in peripheral blood mononuclear cells (PBMCs) [46,50,68,86–89] For example, using the technology of mtDNA quantiﬁcation [87,115], the effects of the various NRTIs have been studied in in vitro cell culture experiments [77,78,116]. In a typical experiment, we tested the effects of ddC, one of the most toxic NRTIs, on a culture of ﬁbroblasts that contain a relatively high number of mtDNA copies, thereby increasing assay resolution. For 11 weeks, cells were cultured in both the absence and presence of 30 µM ddC. Samples were taken to be tested for both the mtDNA copy numbers per cell as well as the lactate: pyruvate ratio. The lactate/pyruvate ratio is a measure of the mitochondrial function of the cell; if this ratio exceeds 50, the cells have poor mitochondrial function that jeopardizes viability. Within 3 weeks after the start of the culture, the lactate/pyruvate ratio reached 50, and it continued to increase, after 4 weeks reaching about 120, where it remained for the duration of the treatment. When the ddC was washed out, the ratio decreased to under 50 after 4 weeks, and to 40 at week 11. The mtDNA copy number per cell dropped more than 80% within the ﬁrst week, and at week 3 to <1%, where it remained. As above, washing out the ddC at week 4 caused a rapid increase in mtDNA copy numbers to 10% of the start value after 3 weeks, and restoration to 100% after 10 weeks [88]. These experiments show that loss of mtDNA precedes the rise in lactate/pyruvate ratio above critical levels. It also shows that mtDNA returns to baseline levels far sooner than the lactate/pyruvate ratio decreases to levels in the normal range. This in vitro experiment shows that the mtDNA copy number responds dynamically to antiviral drugs, returning to normal levels if the antiviral drug is withdrawn [88]. Since several of the clinical adverse events resulting from the use of NRTIs are related to reduced liver function, such as hyperlactatemia, many in vitro

TECHNOLOGIES FOR MEASURING mtDNA

481

experiments have been done using HepG2 cells, an immortalized liver cell line. Like ﬁbroblasts, HepG2 cells also have a high number of mtDNA copies per cell (∼1000 copies/cell), allowing plenty of range to detect inhibition of DNA polymerase γ and/or chain termination. In agreement with the results of the ﬁbroblast experiment, culturing HepG2 cells in the presence of ddC showed a decline in mtDNA levels. Other NRTIs were compared for their relative potency in HepG2 cells. Several groups have reported that the rank order of toxicity is ddC > ddI > d4T > AZT, which concurs with the in vitro inhibition by these drugs of the isolated DNA polymerase γ enzyme [76,78,116,117]. Later experiments in our labs included the other NRTIs in this ranking, and although the difference between ddI and d4T varied between experiments, our ﬁndings are in accord with the literature, with inhibition of mtDNA transcription inhibited by ddC > ddI > d4T > AZT > ABC = 3TC = TDF [77,78]. The in vitro ﬁndings correlate well with the observed clinical adverse events related to mitochondrial toxicity. Speciﬁcally, ddC is known as highly toxic, and d4T is renowned for its relatively high number of toxic events. However, it should be noted that for many years, d4T has been the most widely prescribed NRTI because of its high antiviral potency, and as such, the absolute number of adverse events is expected to be higher. Regardless, the ﬁndings support the contention that in vitro assays can serve as a predictive tool for anticipated adverse effects in vivo.

7. TECHNOLOGIES FOR MEASURING mtDNA Our hypothesis is that mtDNA depletion in PBMCs is a reﬂection of loss in other organs. The limitation of such a general toxicity marker is that it does not indicate which organ or tissue is at highest risk of clinical dysfunction or injury, but it does alert the health care provider that increased surveillance is warranted [87]. Substantial depletion of mtDNA has been observed in PBMCs and in affected tissues (subcutaneous fat, liver, muscle) after the initiation of antiviral therapy with NRTIs (Figure 4), with d4T and ddI being the most toxic. Since the early 1990s, scientist have been interested in measuring the quantity of mtDNA in cells. In the early days, quantiﬁcation was done via Northern blotting, a technique that provides a semiquantitative index useful for ranking responses in a given experiment. More recently, modern ampliﬁcation technologies provide highly accurate measurements. For example, a number of labs have developed real-time PCR-based quantiﬁcation assays [49,86,118–123], where a primer pair for the detection and quantiﬁcation of both mtDNA and nuclear DNA (nDNA) is used to quantify the number of mtDNA copies per two nDNA copies. The number of nDNA copies per mammalian cell in general is two, but this can vary depending on the cell type, particularly transformed cells used in culture systems. The exceptions for such an approach are when measuring haploid gametes (spermatozoa or oocytes) or the number of nDNA copies in muscle cells which are normally multinucleate.

482

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

Percentage copies DNA from baseline

150 AZT mono AZT/ddl AZT/ddc 100

50

0 0

4

24 weeks since start of treatment

48

Figure 4 Rapid decline of mtDNA in the groups treated with dual therapy (AZT + ddI and AZT + ddC) but not in the group with AZT monotherapy.

All of the PCR-based quantiﬁcation assays use a calibration curve based on plasmids containing the gene that is ampliﬁed and represents either the mtDNA or nDNA molecules. Plotting the measured real-time PCR values on the calibration curve provides an assessment of the quantity of each of the two genes, representing the mtDNA and nDNA, respectively. The relative measure of the number of mtDNA copies per cell is obtained by dividing the number of mtDNA copies per two nDNA copies. Variation in this type of assay can be inﬂuenced by a wide assortment of factors, including differences in calibration and standardization of the standardization curve, different ampliﬁcation efﬁciencies between the various targeted ampliﬁcation regions but also between the two separately targeted genes (unless the genes are ampliﬁed in a multiplex or duplex assay), difference in nuclear extraction efﬁciency between methods and/or cells, variation in detection probe hybridization, and a number of other factors [124]. Given the range of possible sources of variability, it should not be surprising that intralaboratory comparisons have confounded interpretation of NFTI toxicity. For example, the threshold of mtDNA depletion and association with lipodystrophy varies depending on the primers and laboratory protocols used. To address this, a metastudy to determine the accuracy of two techniques, polymerase chain reaction and a newer nucleic acid sequence base ampliﬁcation (NASBA) protocol, has shown reasonable concordance between the assays and across several labs [115]. In addition to clinical or academic labs, one NASBA-based assay is commercially available (www.primagen.com). In contrast to the other assays, NASBA is not a cyclic process but a continuous isothermal ampliﬁcation technology, which can be performed using relatively simple and less expensive equipment. NASBA can also be used for duplex or multiplex assays, which makes it a ﬂexible platform. The assay makes use of a primer set to amplify mtDNA that spans a RNA splice site in the 3 end of 16S rRNA with one primer and the other in the

REFERENCES

483

tRNAL just downstream of the mitochondrial transcription terminator, thereby avoiding coampliﬁcation of mitochondrial RNA. For the ampliﬁcation of nuclear gene snRNP U1A DNA, a primer set was developed with one of the primers in an intron and the other in the adjacent exon sequence [87]. The assay has been tested extensively using various tissue cell types and cell lines. The excellent performance of this assay has allowed better resolution of the “safety margin” in the amount of mtDNA depletion by tenofovir [125].

8. CONCLUSIONS Combination antiretroviral therapy has changed HIV infection from an immediately life-threating disease into a chronic disease requiring life-long therapy. However, prolonged treatment has resulted in adverse effects that can be linked to mitochondrial dysfunction, including cardiomyopathy, hepatic steatosis and hyperlactatemia, lipoatrophy and lipohypotrophy, kidney malfunction (i.e., Fanconi syndrome), and peripheral neuropathy. Development of antivirals with less toxic effects on mitochondria, plus secondary treatments for iatrogenic mitochondrial dysfunction (e.g., uridine; see Chapter 9), together with improved monitoring of mitochondrial dysfunction by measuring changes in mtDNA has started to address adverse effects [77,116,125]. The utility of mtDNA as a biomarker to predict adverse effects has proven to be difﬁcult to demonstrate, due to contradictory results between methods and studies. Working groups consisting of academic and industry researchers are establishing international standards to calibrate assays to decrease the intralaboratory variability. Meanwhile, an increasing number of well-designed studies have fostered acceptance of mtDNA assessment as an index of NRTI toxicity in vitro [37,38,46,49,76,122,126,127]. Minimizing intralaboratory variability will expedite acceptance of its value for in vivo studies, and ultimately, utility for therapeutic management of HIV-infected patients.

REFERENCES 1. Dieleman JP, Jambroes M, Gyssens IC, et al. Determinants of recurrent toxicity-driven switches of highly active antiretroviral therapy. The ATHENA cohort. AIDS. 2002;16(5):737–745. 2. Cherry CL, McArthur JC, Hoy JF, Wesselingh SL. Nucleoside analogues and neuropathy in the era of HAART. J Clin Virol. 2003;26(2):195–207. 3. Dalakas MC. Peripheral neuropathy and antiretroviral drugs. J Peripheral Nerv Syst. 2001;6(1):14–20. 4. Keilbaugh SA, Hobbs GA, Simpson MV. Anti-human immunodeﬁciency virus type 1 therapy and peripheral neuropathy: prevention of 2 ,3 -dideoxycy tidinetoxicity in PC12 cells, a neuronal model, by uridine and pyruvate. Mol Pharmacol. 1993;44(4):702–706.

484

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

5. Rabing CE, Stegger M, Jensen-Fangel S, Laursen AL, Ostergaard L. Mitochondrial DNA levels in fat and blood cells from patients with lipodystrophy or peripheral neuropathy and the effect of 90 days of high-dose coenzyme Q treatment: a randomized, double-blind, placebo-controlled pilot study. Clin Infect Dis. 2004;39(9):1371–1379. 6. Arnaudo E, Dalakas M, Shanske S, Moraes CT, Dimauro S, Schon EA. Depletion of muscle mitochondrial DNA in AIDS patients with zidovudine-induced myopathy. Lancet. 1991;337(8740):508–510. 7. Chariot P, Drogou I, Lacroix-Szmania I, et al. Zidovudine-induced mitochondrial disorder with massive liver steatosis, myopathy, lactic acidosis, and mitochondrial DNA depletion. J Hepatol. 1999;30(1):156–160. 8. Grau JM, Casademont J. HIV or zidovudine myopathy? Neurology. 1994;44(2): 361–364. 9. Lewis W. Mitochondrial DNA replication, nucleoside reverse-transcriptase inhibitors, and AIDS cardiomyopathy. Prog Cardiovasc Dis. 2003;45(4):305–318. 10. Masanes F, Barrientos A, Cebrian M, et al. Clinical, histological and molecular reversibility of zidovudine myopathy. J Neurol Sci. 1998;159(2):226–228. 11. Bleeker-Rovers CP, Kadir SW, van Leusen R, Richter C. Hepatic steatosis and lactic acidosis caused by stavudine in an HIV-infected patient. Neth J Med. 2000;57(5):190–193. 12. Brinkman K. Editorial response: hyperlactatemia and hepatic steatosis as features of mitochondrial toxicity of nucleoside analogue reverse transcriptase inhibitors. Clin Infect Dis. 2000;31(1):167–169. 13. McGovern BH, Ditelberg JS, Taylor LE, et al. Hepatic steatosis is associated with ﬁbrosis, nucleoside analogue use, and hepatitis C virus genotype 3 infection in HIV-seropositive patients. Clin Infect Dis. 2006;43(3):365–372. 14. Olano JP, Borucki MJ, Wen JW, Haque AK. Massive hepatic steatosis and lactic acidosis in a patient with AIDS who was receiving zidovudine. Clin Infect Dis. 1995;21(4):973–976. 15. Pao D, Watson C, Peters B, Lucas SB, Miller RF. Hyperlactataemia and hepatic steatosis: mitochondrial toxicity of nucleoside reverse transcriptase inhibitors. Sex Transm Infect. 2001;77(5):381–384. 16. Ristig M, Drechsler H, Powderly WG. Hepatic steatosis and HIV infection. AIDS Patient Care STDS. 2005;19(6):356–365. 17. Kontorinis N, Dieterich D. Hepatotoxicity of antiretroviral therapy. AIDS Rev. 2003;5(1):36–43. 18. Nunez M. Hepatotoxicity of antiretrovirals: incidence, mechanisms and management. J Hepatol. 2006;44(1 suppl):S132–S139. 19. Montessori V, Harris M, Montaner JS. Hepatotoxicity of nucleoside reverse transcriptase inhibitors. Semin Liver Dis. 2003;23(2):167–172. 20. Boubaker K, Flepp M, Sudre P, et al. Hyperlactatemia and antiretroviral therapy: the Swiss HIV Cohort Study. Clin Infect Dis. 2001;33(11):1931–1937. 21. Brinkman K. Evidence for mitochondrial toxicity: lactic acidosis as proof of concept. J HIV Ther. 2001;6(1):13–16. 22. John M, Mallal S. Hyperlactatemia syndromes in people with HIV infection. Curr Opin Infect Dis. 2002;15(1):23–29. 23. Khouri S, Cushing H. Lactic acidosis secondary to nucleoside analogue antiretroviral therapy. AIDS Read. 2000;10(10):602–606.

REFERENCES

485

24. Mokrzycki MH, Harris C, May H, Laut J, Palmisano J. Lactic acidosis associated with stavudine administration: a report of ﬁve cases. Clin Infect Dis. 2000;30(1): 198–200. 25. ter Hofstede HJ, de Marie S, Foudraine NA, Danner SA, Brinkman K. Clinical features and risk factors of lactic acidosis following long-term antiretroviral therapy: 4 fatal cases. Int J STD AIDS. 2000;11(9):611–616. 26. Baker WC. Impaired glucose metabolism and HIV: an overview. Adv Nurse Pract. 2003;11(4):47–49. 27. Ledergerber B, Furrer H, Rickenbach M, et al. Factors associated with the incidence of type 2 diabetes mellitus in HIV-infected participants in the Swiss HIV Cohort Study. Clin Infect Dis. 2007;45(1):111–119. 28. Fleischman A, Johnsen S, Systrom DM, et al. Effects of a nucleoside reverse transcriptase inhibitor, stavudine, on glucose disposal and mitochondrial function in muscle of healthy adults. Am J Physiol Endocrinol Metab. 2007;292(6):E1666–E1673. 29. Carr A, Cooper DA. Adverse effects of antiretroviral therapy. Lancet. 2000; 356(9239):1423–1430. 30. Herman JS, Easterbrook PJ. The metabolic toxicities of antiretroviral therapy. Int J STD AIDS. 2001;12(9):555–562. 31. Mallon PW, Cooper DA, Carr A. HIV-associated lipodystrophy. HIV Med. 2001;2(3):166–173. 32. Moyle G. Lipodystrophy: lack of agreement on deﬁnition and etiology presents a challenge to research and therapy. AIDS Read. 2002;12(10):438, 440–438, 442. 33. ter Hofstede HJ, Burger DM, Koopmans PP. Antiretroviral therapy in HIV patients: aspects of metabolic complications and mitochondrial toxicity. Neth J Med. 2003;61(12):393–403. 34. Tozser J. HIV inhibitors: problems and reality. Ann N Y Acad Sci. 2001;946:145–159. 35. Walker UA, Brinkman K. NRTI induced mitochondrial toxicity as a mechanism for HAART related lipodystrophy: fact or ﬁction? HIV Med. 2001;2(3):163–165. 36. Wanke C. Pathogenesis of HIV-associated lipodystrophy. AIDS Clin Care. 2000; 12(12):100. 37. Brinkman K, Smeitink JA, Romijn JA, Reiss P. Mitochondrial toxicity induced by nucleoside-analogue reverse-transcriptase inhibitors is a key factor in the pathogenesis of antiretroviral-therapy-related lipodystrophy. Lancet. 1999;354(9184): 1112–1115. 38. Lewis W, Dalakas MC. Mitochondrial toxicity of antiviral drugs. Nat Med. 1995; 1(5):417–422. 39. Bendele RA, Richardson FC. Adefovir nephrotoxicity and mitochondrial DNA depletion. Hum Pathol. 2002;33(5):574. 40. Buffet M, Schwarzinger M, Amellal B, et al. Mitochondrial DNA depletion in adipose tissue of HIV-infected patients with peripheral lipoatrophy. J Clin Virol. 2005;33(1):60–64. 41. Chabi B, Mousson dC, Duborjal H, Issartel JP, Stepien G. Quantiﬁcation of mitochondrial DNA deletion, depletion, and overreplication: application to diagnosis. Clin Chem. 2003;49(8):1309–1317. 42. Church JA, Mitchell WG, Gonzalez-Gomez I, et al. Mitochondrial DNA depletion, near-fatal metabolic acidosis, and liver failure in an HIV-infected child treated with combination antiretroviral therapy. J Pediatr. 2001;138(5):748–751.

486

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

43. Dalakas MC, Semino-Mora C, Leon-Monzon M. Mitochondrial alterations with mitochondrial DNA depletion in the nerves of AIDS patients with peripheral neuropathy induced by 2 3 -dideoxycytidine (ddC). Lab Invest. 2001;81(11):1537–1544. 44. Mancuso M, Filosto M, Bonilla E, et al. Mitochondrial myopathy of childhood associated with mitochondrial DNA depletion and a homozygous mutation (T77M) in the TK2 gene. Arch Neurol. 2003;60(7):1007–1009. 45. Mancuso M, Salviati L, Sacconi S, et al. Mitochondrial DNA depletion: mutations in thymidine kinase gene with myopathy and SMA. Neurology. 2002;59(8):1197–1202. 46. Miura T, Goto M, Hosoya N, et al. Depletion of mitochondrial DNA in HIV-1infected patients and its amelioration by antiretroviral therapy. J Med Virol. 2003; 70(4):497–505. 47. Nolan D, Hammond E, Martin A, et al. Mitochondrial DNA depletion and morphologic changes in adipocytes associated with nucleoside reverse transcriptase inhibitor therapy. AIDS. 2003;17(9):1329–1338. 48. Pace CS, Martin AM, Hammond EL, Mamotte CD, Nolan DA, Mallal SA. Mitochondrial proliferation, DNA depletion and adipocyte differentiation in subcutaneous adipose tissue of HIV-positive HAART recipients. Antivir Ther. 2003;8(4):323–331. 49. Petit C, Mathez D, Barthelemy C, et al. Quantitation of blood lymphocyte mitochondrial DNA for the monitoring of antiretroviral drug-induced mitochondrial DNA depletion. J Acquir Immune Deﬁc Syndr. 2003;33(4):461–469. 50. Reiss P, Casula M, de Ronde A, Weverling GJ, Goudsmit J, Lange JM. Greater and more rapid depletion of mitochondrial DNA in blood of patients treated with dual (zidovudine + didanosine or zidovudine + zalcitabine) vs. single (zidovudine) nucleoside reverse transcriptase inhibitors. HIV Med. 2004;5(1):11–14. 51. Tanji N, Tanji K, Kambham N, Markowitz GS, Bell A, D’agati VD. Adefovir nephrotoxicity: possible role of mitochondrial DNA depletion. Hum Pathol. 2001; 32(7):734–740. 52. Walker UA, Auclair M, Lebrecht D, Kornprobst M, Capeau J, Caron M. Uridine abrogates the adverse effects of antiretroviral pyrimidine analogues on adipose cell functions. Antivir Ther. 2006;11(1):25–34. 53. Walker UA, Venhoff N, Koch EC, Olschewski M, Schneider J, Setzer B. Uridine abrogates mitochondrial toxicity related to nucleoside analogue reverse transcriptase inhibitors in HepG2 cells. Antivir Ther. 2003;8(5):463–470. 54. Banasch M, Goetze O, Knyhala K, et al. Uridine supplementation enhances hepatic mitochondrial function in thymidine-analogue treated HIV-infected patients. AIDS. 2006;20(11):1554–1556. 55. McComsey GA, O’Riordan M, Setzer B, Lebrecht D, Baron E, Walker UA. Uridine supplementation in HIV lipoatrophy: pilot trial on safety and effect on mitochondrial indices. Eur J Clin Nutr. 2007; in press (advanced online publication 30 May). 56. Sutinen J, Walker UA, Sevastianov K, et al. Uridine supplementation for the treatment of antiretroviral therapy-associated lipoatrophy: a randomized, double-blind, placebo-controlled trial. Antivir Ther. 2007;12(1):97–105. 57. UNAIDS. AIDS epidemic update December 2006. WHO Library Cataloguing-in Publication Data. 2006. 58. van Leth F, Phanuphak P, Ruxrungtham K, et al. Comparison of ﬁrst-line antiretroviral therapy with regimens including nevirapine, efavirenz, or both drugs, plus

REFERENCES

59. 60. 61.

62.

63. 64. 65. 66. 67.

68.

69.

70.

71.

72.

73.

74.

487

stavudine and lamivudine: a randomised open-label trial, the 2NN Study. Lancet. 2004;363(9417):1253–1263. Kontorinis N, Dieterich DT. Toxicity of non-nucleoside analogue reverse transcriptase inhibitors. Semin Liver Dis. 2003;23(2):173–182. Rivero A, Mira JA, Pineda JA. Liver toxicity induced by non-nucleoside reverse transcriptase inhibitors. J Antimicrob Chemother. 2007;59(3):342–346. van der Valk M, Kastelein JJ, Murphy RL, et al. Nevirapine-containing antiretroviral therapy in HIV-1 infected patients results in an anti-atherogenic lipid proﬁle. AIDS. 2001;15(18):2407–2414. Carr A, Miller J, Law M, Cooper DA. A syndrome of lipoatrophy, lactic acidaemia and liver dysfunction associated with HIV nucleoside analogue therapy: contribution to protease inhibitor-related lipodystrophy syndrome. AIDS. 2000;14(3):F25–F32. McComsey G. Update on mitochondrial toxicity of antiretrovirals and its link to lipodystrophy. AIDS Rev. 2002;4(3):140–147. Chalmers AC, Greco CM, Miller RG. Prognosis in AZT myopathy. Neurology. 1991;41(8):1181–1184. Miro O, Pedrol E, Cebrian M, et al. Skeletal muscle studies in patients with HIV-related wasting syndrome. J Neurol Sci. 1997;150(2):153–159. UNAIDS. Recommendations for ﬁrst line therapy. http://www.who.int/hiv/topics/arv/ en/. 2007. Moyle GJ, Datta D, Mandalia S, Morlese J, Asboe D, Gazzard BG. Hyperlactataemia and lactic acidosis during antiretroviral therapy: relevance, reproducibility and possible risk factors. AIDS. 2002;16(10):1341–1349. Maagaard A, Holberg-Petersen M, Kvittingen EA, Sandvik L, Bruun JN. Depletion of mitochondrial DNA copies/cell in peripheral blood mononuclear cells in HIV-1-infected treatment-naive patients. HIV Med. 2006;7(1):53–58. Gerschenson M, Shiramizu B, LiButti DE, Shikuma CM. Mitochondrial DNA levels of peripheral blood mononuclear cells and subcutaneous adipose tissue from thigh, fat and abdomen of HIV-1 seropositive and negative individuals. Antivir Ther. 2005;10 (Suppl 2):M83–M89. Heath KV, Montaner JS, Bondy G, Singer J, O’Shaughnessy MV, Hogg RS. Emerging drug toxicities of highly active antiretroviral therapy for human immunodeﬁciency virus (HIV) infection. Curr Drug Targets. 2003;4(1):13–22. Heath KV, Hogg RS, Singer J, Chan KJ, O’Shaughnessy MV, Montaner JS. Antiretroviral treatment patterns and incident HIV-associated morphologic and lipid abnormalities in a population-based chort. J Acquir Immune Deﬁc Syndr. 2002;30(4): 440–447. Giralt M, Domingo P, Guallar JP, et al. HIV-1 infection alters gene expression in adipose tissue, which contributes to HIV- 1/HAART-associated lipodystrophy. Antivir Ther. 2006;11(6):729–740. Carr A, Emery S, Law M, Puls R, Lundgren JD, Powderly WG. An objective case deﬁnition of lipodystrophy in HIV-infected adults: a case–control study. Lancet. 2003;361(9359):726–735. Polo R, Verdejo J, Martinez-Rodriguez S, Madrigal P, Gonzalez-Munoz M. Lipoatrophy, fat accumulation, and mixed syndrome in protease inhibitor-naive HIV-infected patients. J Acquir Immune Deﬁc Syndr. 2000;25(3):284–286.

488

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

75. Fliers E, Sauerwein HP, Romijn JA, et al. HIV-associated adipose redistribution syndrome as a selective autonomic neuropathy. Lancet. 2003;362(9397):1758–1760. 76. Lewis W, Kohler JJ, Hosseini SH, et al. Antiretroviral nucleosides, deoxynucleotide carrier and mitochondrial DNA: evidence supporting the DNA pol gamma hypothesis. AIDS. 2006;20(5):675–684. 77. Cihlar T, Birkus G, Greenwalt DE, Hitchcock MJ. Tenofovir exhibits low cytotoxicity in various human cell types: comparison with other nucleoside reverse transcriptase inhibitors. Antivir Res. 2002;54(1):37–45. 78. de Baar MP, de Rooij ER, Smolders KGM, van Schijndel HB, Timmermans EC, Bethell RC. Effects of apricitabine and other nucleoside reverse transcriptase inhibitors on replication of mitochondrial DNA in HepG2 cells. Antivir. Res. 2007;76(1):68–74. 79. Lewis W, Gonzalez B, Chomyn A, Papoian T. Zidovudine induces molecular, biochemical, and ultrastructural changes in rat skeletal muscle mitochondria. J Clin Invest. 1992;89(4):1354–1360. 80. Maher B, Alﬁrevic A, Vilar FJ, Wilkins EG, Park BK, Pirmohamed M. TNF-alpha promoter region gene polymorphisms in HIV-positive patients with lipodystrophy. AIDS. 2002;16(15):2013–2018. 81. Nolan D, Moore C, Castley A, et al. Tumour necrosis factor-alpha gene-238G/A promoter polymorphism associated with a more rapid onset of lipodystrophy. AIDS. 2003;17(1):121–123. 82. Yamanaka H, Gatanaga H, Kosalaraksa P, et al. Novel mutation of human DNA polymerase gamma associated with mitochondrial toxicity induced by anti-HIV treatment. J Infect Dis. 2007;195(10):1419–1425. 83. Cherry CL, Gahan ME, McArthur JC, Lewin SR, Hoy JF, Wesselingh SL. Exposure to dideoxynucleosides is reﬂected in lowered mitochondrial DNA in subcutaneous fat. J Acquir Immune Deﬁc Syndr. 2002;30(3):271–277. 84. Cherry CL, Nolan D, James IR, et al. Tissue-speciﬁc associations between mitochondrial DNA levels and current treatment status in HIV-infected individuals. J Acquir Immune Deﬁc Syndr. 2006;42(4):435–440. 85. McComsey GA, Paulsen DM, Lonergan JT, et al. Improvements in lipoatrophy, mitochondrial DNA levels and fat apoptosis after replacing stavudine with abacavir or zidovudine. AIDS. 2005;19(1):15–23. 86. Cote HC, Brumme ZL, Craib KJ, et al. Changes in mitochondrial DNA as a marker of nucleoside toxicity in HIV-infected patients. N Engl J Med. 2002;346(11):811–820. 87. Timmermans EC, Tebas P, Ruiter JP, Wanders RJ, de Ronde A, de Baar MP. Real-time nucleic acid sequence-based ampliﬁcation assay to quantify changes in mitochondrial DNA concentrations in cell cultures and blood cells from HIV-infected patients receiving antiviral therapy. Clin Chem. 2006;52(6):979–987. 88. Casula M, Weverling GJ, Wit FW, et al. Mitochondrial DNA and RNA increase in peripheral blood mononuclear cells from HIV-1-infected patients randomized to receive stavudine-containing or stavudine-sparing combination therapy. J Infect Dis. 2005;192(10):1794–1800. 89. Casula M, van d, V, Wit FW, Nievaard MA, Reiss P. Mitochondrial DNA assessment in adipocytes and peripheral blood mononuclear cells of HIV-infected patients with lipodystrophy according to a validated case deﬁnition. HIV Med. 2007;8(1):32–37.

REFERENCES

489

90. Ben Romano R, Rudich A, Etzion S, et al. Nelﬁnavir induces adipocyte insulin resistance through the induction of oxidative stress: differential protective effect of antioxidant agents. Antivir Ther. 2006;11(8):1051–1060. 91. Nerurkar PV, Pearson L, Frank JE, Yanagihara R, Nerurkar VR. Highly active antiretroviral therapy (HAART)-associated lactic acidosis: in vitro effects of combination of nucleoside analogues and protease inhibitors on mitochondrial function and lactic acid production. Cell Mol Biol (Noisy-Le-Grand). 2003;49(8):1205–1211. 92. Hertel J, Struthers H, Horj CB, Hruz PW. A structural basis for the acute effects of HIV protease inhibitors on GLUT4 intrinsic activity. J Biol Chem. 2004;279(53): 55147–55152. 93. Murata H, Hruz PW, Mueckler M. Indinavir inhibits the glucose transporter isoform Glut4 at physiologic concentrations. AIDS. 2002;16(6):859–863. 94. Bastard JP, Caron M, Vidal H, et al. Association between altered expression of adipogenic factor SREBP1 in lipoatrophic adipose tissue from HIV-1-infected patients and abnormal adipocyte differentiation and insulin resistance. Lancet. 2002; 359(9311):1026–1031. 95. Caron M, Auclair M, Vigouroux C, Glorian M, Forest C, Capeau J. The HIV protease inhibitor indinavir impairs sterol regulatory element-binding protein-1 intranuclear localization, inhibits preadipocyte differentiation, and induces insulin resistance. Diabetes. 2001;50(6):1378–1388. 96. Murata H, Hruz PW, Mueckler M. The mechanism of insulin resistance caused by HIV protease inhibitor therapy. J Biol Chem. 2000;275(27):20251–20254. 97. Nolan D, John M, Mallal S. Antiretoviral therapy and the lipodystrophy syndrome, part 2: concepts in aetiopathogenesis. Antivir Ther. 2001;6(3):145–160. 98. Rudich A, Vanounou S, Riesenberg K, et al. The HIV protease inhibitor nelﬁnavir induces insulin resistance and increases basal lipolysis in 3T3-L1 adipocytes. Diabetes. 2001;50(6):1425–1431. 99. Kuenzli S, Saurat JH. Peroxisome proliferator-activated receptors in cutaneous biology. Br J Dermatol. 2003;149(2):229–236. 100. de la Asuncion JG, del Olmo ML, Sastre J, et al. AZT treatment induces molecular and ultrastructural oxidative damage to muscle mitochondria: prevention by antioxidant vitamins. J Clin Invest. 1998;102(1):4–9. 101. de la Asuncion JG, del Olmo ML, Gomez-Cambronero LG, Sastre J, Pallardo FV, Vina J. AZT induces oxidative damage to cardiac mitochondria: protective effect of vitamins C and E. Life Sci. 2004;76(1):47–56. 102. Skuta G, Fischer GM, Janaky T, et al. Molecular mechanism of the short-term cardiotoxicity caused by 2 ,3 -dideoxycytidine (ddC): modulation of reactive oxygen species levels and ADP-ribosylation reactions. Biochem Pharmacol. 1999;58(12): 1915–1925. 103. Szabados E, Fischer GM, Toth K, et al. Role of reactive oxygen species and poly-ADP-ribose polymerase in the development of AZT-induced cardiomyopathy in rat. Free Radic Biol Med. 1999;26(3–4):309–317. 104. Macho A, Calzado MA, Jimenez-Reina L, Ceballos E, Leon J, Munoz E. Susceptibility of HIV-1-TAT transfected cells to undergo apoptosis: biochemical mechanisms. Oncogene. 1999;18(52):7543–7551.

490

MITOCHONDRIAL TOXICITY OF ANTIVIRAL DRUGS

105. Olivetta E, Pietraforte D, Schiavoni I, Minetti M, Federico M, Sanchez M. HIV-1 Nef regulates the release of superoxide anions from human macrophages. Biochem J. 2005;390(pt 2):591–602. 106. Gulow K, Kaminski M, Darvas K, Suss D, Li-Weber M, Krammer PH. HIV-1 trans-activator of transcription substitutes for oxidative signaling in activation-induced T cell death. J Immunol. 2005;174(9):5249–5260. 107. Delchev SD, Georgieva KN, Koeva YA, Atanassova PK. Bcl-2/Bax ratio, mitochondrial membranes and aerobic enzyme activity in cardiomyocytes of rats after submaximal training. Folia Med (Plovdiv.) 2006;48(2):50–56. 108. Eissing T, Waldherr S, Allgower F, Scheurich P, Bullinger E. Response to bistability in apoptosis: roles of Bax, Bcl-2, and mitochondrial permeability transition pores. Biophys J. 2007;92(9):3332–3334. 109. Kim H, Raﬁuddin-Shah M, Tu HC, et al. Hierarchical regulation of mitochondriondependent apoptosis by Bcl-2 subfamilies. Nat Cell Biol. 2006;8(12):1348–1358. 110. Matarrese P, Gambardella L, Cassone A, Vella S, Cauda R, Malorni W. Mitochondrial membrane hyperpolarization hijacks activated T lymphocytes toward the apoptotic-prone phenotype: homeostatic mechanisms of HIV protease inhibitors. J Immunol. 2003;170(12):6006–6015. 111. Metrailler-Ruchonnet I, Pagano A, Carnesecchi S, Ody C, Donati Y, Barazzone AC. Bcl-2 protects against hyperoxia-induced apoptosis through inhibition of the mitochondria-dependent pathway. Free Radic Biol Med. 2007;42(7):1062–1074. 112. Shankar S, Srivastava RK. Involvement of Bcl-2 family members, phosphatidylinositol 3 -kinase/AKT and mitochondrial p53 in curcumin (diferulo lylmethane)-induced apoptosis in prostate cancer. Int J Oncol. 2007;30(4):905–918. 113. Simbula G, Columbano A, Ledda-Columbano GM, et al. Increased ROS generation and p53 activation in alpha-lipoic acid–induced apoptosis of hepatoma cells. Apoptosis. 2007;12(1):113–123. 114. Somma F, Tuosto L, Montani MS, Di Somma MM, Cundari E, Piccolella E. Engagement of CD4 before TCR triggering regulates both Bax- and Fas (CD95)-mediated apoptosis. J Immunol. 2000;164(10):5078–5087. 115. Hammond EL, Sayer D, Nolan D, et al. Assessment of precision and concordance of quantitative mitochondrial DNA assays: a collaborative international quality assurance study. J Clin Virol. 2003;27(1):97–110. 116. Birkus G, Hitchcock MJ, Cihlar T. Assessment of mitochondrial toxicity in human cells treated with tenofovir: comparison with other nucleoside reverse transcriptase inhibitors. Antimicrob Agents Chemother. 2002;46(3):716–723. 117. Chen CH, Vazquez-Padua M, Cheng YC. Effect of anti-human immunodeﬁciency virus nucleoside analogs on mitochondrial DNA and its implication for delayed toxicity. Mol Pharmacol. 1991;39(5):625–628. 118. Gahan ME, Miller F, Lewin SR, et al. Quantiﬁcation of mitochondrial DNA in peripheral blood mononuclear cells and subcutaneous fat using real-time polymerase chain reaction. J Clin Virol. 2001;22(3):241–247. 119. Cossarizza A, Pinti M, Moretti L, et al. Mitochondrial functionality and mitochondrial DNA content in lymphocytes of vertically infected human immunodeﬁciency virus-positive children with highly active antiretroviral therapy-related lipodystrophy. J Infect Dis. 2002;185(3):299–305.

REFERENCES

491

120. Chene G, Amellal B, Pedrono G, et al. Changes in the peripheral blood mtDNA levels in naive patients treated by different nucleoside reverse transcriptase inhibitor combinations and their association with subsequent lipodystrophy. AIDS Res Hum Retroviruses. 2007;23(1):54–61. 121. Maagaard A, Holberg-Petersen M, Kollberg G, Oldfors A, Sandvik L, Bruun JN. Mitochondrial (mt)DNA changes in tissue may not be reﬂected by depletion of mtDNA in peripheral blood mononuclear cells in HIV-infected patients. Antivir Ther. 2006;11(5):601–608. 122. Gourlain K, Amellal B, Ait AZ, Dupin N, Katlama C, Calvez V. Quantitative analysis of human mitochondrial DNA using a real-time PCR assay. HIV Med. 2003; 4(3):287–292. 123. Cossarizza A. Tests for mitochondrial function and DNA: potentials and pitfalls. Curr Opin Infect Dis. 2003;16(1):5–10. 124. Gallant J, Staszewski S, Pozniak A, et al. Favorable lipid and mitochondrial (mt) DNA proﬁle for tenofovir disoproxil fumarate (TDF) compared to stavudine (d4T) in combination with lamivudine (3TC) and efavirenz (EFV) in antiretroviral therapy (ART) naive patients: a 48 week interim analysis. Abstract. 42nd Interscience Conference on Antimicrobial Agents and Chemotherapy Sept. 27–30, 2002. 125. Casula M, Vrisekoop N, Wit FW, et al. Mitochondrial DNA decline in T cells of HIV-1 seroconverters may be dependent on immune activation. J Infect Dis. 2007; 196(3):371–376. 126. Casula M, Bosboom-Dobbelaer I, Smolders K, et al. Infection with HIV-1 induces a decrease in mtDNA. J Infect Dis. 2005;191(9):1468–1471. 127. Polo R, Martinez S, Madrigal P, Gonzalez-Munoz M. Factors associated with mitochondrial dysfunction in circulating peripheral blood lymphocytes from HIV-infected people. J Acquir Immune Deﬁc Syndr. 2003;34(1):32–36.

21 CLINICAL ASSESSMENT OF MITOCHONDRIAL FUNCTION VIA [13C]METHIONINE EXHALATION Laura Milazzo Institute of Infectious and Tropical Diseases, University of Milan, L. Sacco Hospital, Milan, Italy

1. 2. 3. 4.

Introduction Development of exhalation assays Methionine breath test [13 C]Methionine breath test in the study of drug-induced mitochondrial toxicity 5. Future perspectives

493 494 495 498 502

1. INTRODUCTION It is increasingly evident that a wide variety of diseases and drug-induced adverse events entail impairment of mitochondrial function. As described in several chapters, noninvasive techniques to detect and monitor such dysfunction in vivo are being developed. We have been developing a simple, noninvasive protocol for measuring mitochondrial function in vivo based on assessments of 13 C exhalation. Substrate selection and pattern of 13 C labeling allows this technology to interrogate several cellular processes selectively in different tissues. For example, [13 C]methionionine is preferentially metabolized via a transmethylation pathway in the liver that is not found, or has very low activity, in other tissues. In this Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

493

494

CLINICAL ASSESSMENT OF MITOCHONDRIAL FUNCTION

chapter we brieﬂy review the development of exhalation techniques and describe recent use of [13 C]methionine as an indicator of hepatic mitochondrial status in HIV patients receiving antiretroviral therapy. 2. DEVELOPMENT OF EXHALATION ASSAYS Tests available for the in vivo assessment of hepatic mitochondrial function include (modiﬁed from Candelli et al. [1]): • • • • • •

The acetoacetate/α-hydroxybutyrate ratio in arterial blood The pyruvate/lactate ratio in arterial blood The metabolism of benzoic acid Hepatic nitrogen clearance determination 31 P-nuclear magnetic resonance spectroscopy Breath tests

Such assessments of hepatic mitochondrial function, although invasive, have been used in a host of pathologies, including Reye’s syndrome, vitamin deﬁciency, acute fatty liver, alcoholic liver injury, liver cirrhosis, and drug toxicity. An exhalation protocol would clearly be a useful clinical probe, and several have been developed. For example, the ﬁrst application of breath testing for the evaluation of liver metabolism was described by Hepner and Vesell in 1974 [2] using [14 C]aminopyrine as substrate. The rationale of using breath test is to recover labeled CO2 in expired air after the administration of a test compound in which the 12 C atom of a functional group has been replaced by a radioactive 14 C, or a stable 13 C, atom. The functional group is removed during the metabolic process that is being interrogated, with ﬁnal production of labeled CO2 that is expired with the systemic bicarbonate pool. In the past these tests were performed with radioactive 14 C-labeled compounds. To circumvent potential radiation hazards, stable, nonradioactive 13 C-labeled substrates are now used, and 13 C enrichment of expired CO2 is analyzed via isotope ratio mass spectrometry [3–6]. Depending on the location of the limiting metabolic step, a variety of physiological and pathological metabolic pathways, such as microsomal, cytoplasmic, and mitochondrial function metabolism, can be studied [7–13]. The introduction of breath tests provided a noninvasive tool for the assessment of dynamic hepatic function, and several 13 C-labeled compounds have been investigated. For example, aminopyrine, phenacetin, methacetin, caffeine, and erythromycin interrogate microsomal cytochrome P450 enzymatic pathways; phenylalanine and galactose are used for assessment of liver cytosolic metabolism; and α-ketoisocaproic acid (KIKA), octanoate, and methionine have been proposed for the evaluation of hepatic mitochondrial function in vivo [6,14]. Ketoisocaproic acid metabolism follows two main enzymatic pathways: (1) oxidative decarboxylation through a branched-chain α-chetoacid dehydrogenase

METHIONINE BREATH TEST

495

complex located exclusively in mitochondria and activated only in the liver, or (2) transamination into the amino acid leucine [15]. Accordingly, hepatic mitochondrial function has been assessed using carbon-labeled ketoisocaproic acid, with a leucine load to inhibit KIKA transamination, by measuring 13 C (or14 C) enrichment of expired CO2 . Conﬂicting data have been reported using the clinical application of the KIKA breath test. Lauterburg et al. [12] reported a signiﬁcant decrease of 14 C exhalation in patients with alcoholic liver disease compared to nonalcoholic liver diseases and healthy controls. However, these results were not corroborated by Bendtsen et al. [16], who found no difference between alcoholic patients and healthy volunteers. The KIKA breath test was also used to assess mitochondrial function in patients having chronic hepatitis B treated with the NRTI lamivudine [17], and the results, no difference between treated and untreated, were in accord with the absence of drug toxicity by histological and ultrastructural evaluation and assessment of mitochondrial function. Octanoic acid, a medium-chain fatty acid metabolized to CO2 through β-oxidation and acetyl-coenzyme A production, has been proposed by Miele et al. [14] as a reliable substrate for the study of hepatic mitochondrial β-oxidation in vivo by means of a breath test. Its 13 C-labeled isotope has been investigated in patients affected by nonalcoholic steatohepatitis (NASH), showing a signiﬁcant increase in hepatic mitochondrial β-oxidation [14].

3. METHIONINE BREATH TEST Methionine is an essential amino acid, predominantly metabolized in the liver via two major pathways (Figure 1). Transamination to α-keto-γ-methiolbutyric acid occurs mainly in the liver, but not under normal metabolic conditions [18–20]. Under physiological conditions, transmethylation by methionine adenosyltranferase into homocysteine is the major metabolic pathway, and this occurs only in the liver, as most other tissues lack one or more required enzymes [21]. Assessment of hepatic mitochondrial function via transmethylation of methionine allows interrogation of different metabolic pathways, depending on which carbon in methionine is labeled. Using three- or four-carbon-labeled methionine it is possible to assess the trans-sulfuration pathway via release of the carbon as α-ketobutyrate, further metabolized into CO2 via the tricarboxylic acid cycle. When a one-carbon-labeled methionine is used, the activity of α-ketobutyrate decarboxylase is interrogated, since this liver mitochondrial enzyme is a limiting step of the substrate oxidation into CO2 [22,23]. Finally, liver mitochondrial function is assessed through the administration of methyl-13 C-labeled methionine. S -Adenosyl-l-methionine is converted to S -adenosylhomocysteine by N -methyltransferase, which is mainly a hepatic enzyme. The function of this enzyme is to remove methyl groups leading to different products, and the major pathway to remove methionine methyl groups is via sarcosine production (Figure 1) [24,25]. Sarcosine is oxidized by sarcosine–dehydrogenase to produce a one-carbon fragment at the oxidation

496

CLINICAL ASSESSMENT OF MITOCHONDRIAL FUNCTION

Figure 1 Schematic representation of the metabolism of the essential amino acid methionine via transamination and transmethylation pathways.

level of formaldehyde which can subsequently be converted into CO2 . Sarcosine dehydrogenase is oxidized by a mitochondrial oxidation system [23,26,27], and it has been shown that the sarcosine oxidase system of rat liver is present exclusively in the mitochondria [28]. Therefore, methyl-13 C-labeled methionine could be used to evaluate the oxidative capacity of liver mitochondria [1]. The [13 C]methionine breath test was ﬁrst described by Armuzzi et al. [13] as a simple, noninvasive, safe, and nonradioactive technique to assess mitochondrial function in vivo. They used methyl-13 C-labeled methionine as an oral tracer in healthy subjects before and after the administration of ethanol. After an overnight fast and a 30-minute rest before and during the test, patients drank 200 mL of orange juice, to normalize gastric pH, and 30 minutes later received 2 mg/kg body weight methyl-13 C-labeled methionine dissolved in 100 mL of water. Exhalation samples were obtained before administration of the labeled molecule, and then every 15 minutes for 3 hours thereafter. Two days later, the test was repeated except that ingestion orange juice contained 0.3 g ethanol/kg body weight. Enrichment of 13 CO2 in breath was analyzed with a gas isotope ratio mass spectrometer. Results were expressed as the percentage of the dose of 13 C administered recovered per hour (percentage 13 C dose/h), the peak percentage of the dose of 13 C administered, and the cumulative percentage

497

120

80

*

40

0 Healthy controls Sbeabosis

Cirrhosis

dose of 13CO2 revocered per hour (%)

DOB (% excess 13C over basal 13C level)

METHIONINE BREATH TEST

50 40

** *

30 20

*

10 0 Healthy controls

Sbeabosis

Cirrhosis

cumulative dose (%)

16

13C

18

4

14 12 10 8 6

* *

2 0 Healthy controls Sbeabosis

Cirrhosis

13

Figure 2 Values of C expressed as delta over base (DOB), dose recovered per hour, and cumulative dose measured by a breath test in healthy subjects, patients with macrovesicular steatosis, and patients with cirrhosis. ∗ p < 0.05 from values in healthy subjects; ∗∗ p < 0.05 from values in patients with cirrhosis. (Modiﬁed from Spahr et al. [30].)

of the dose of 13 C administered recovered over the study period (percentage 13 C cumulative dose). These data provide an indirect estimate of the oxidative capacity of liver mitochondria in healthy subjects and show that the protocol has sufﬁcient sensitivity to detect impaired hepatic mitochondrial function after mild acute ethanol exposure [13]. A differently labeled isomer, 1-[13 C]methionine, has been investigated as substrate by Spahr et al. [29] to assess liver mitochondrial function in acute microvesicular steatosis caused by valproate toxicity. The noninvasive [13 C] methionine breath test revealed mitochondrial impairment in accord with histopathology and ultrastructural abnormalities. Further investigation with the 1-[13 C]methionine breath test has been performed on patients with nonalcoholic macrovesicular steatosis and cirrhosis [30]. Compared to healthy controls, the [13 C]methionine breath test showed impaired hepatic mitochondrial oxidation in both groups of patients. Patients with cirrhosis had the poorest mitochondrial function and also had the lowest 13 C exhalation (Figure 2). Taken in toto, these data support the utility of this breath test for detection of hepatic mitochondrial abnormalities in liver diseases.

498

CLINICAL ASSESSMENT OF MITOCHONDRIAL FUNCTION

4. [13 C]METHIONINE BREATH TEST IN THE STUDY OF DRUG-INDUCED MITOCHONDRIAL TOXICITY We have recently used the [13 C]methionine breath test to assess liver mitochondrial toxicity induced by antiretroviral drugs in HIV patients. A reduction of mtDNA in several tissues, mainly muscle, nerve, and adipose tissue biopsies, has been reported from untreated HIV-1 patients, raising the possibility that HIV may injure mitochondria directly. HIV may induce mitochondrial damage directly through viral gene products as tat and viral protein R (vpr) or through the mediation of cytokines as tumor necrosis factor (TNF)-α and interferon (IFN)-γ. [31–34]. Therefore, mitochondrial dysfunction and/or reduction of mtDNA content may exist before initiation of NRTI treatment, and it is likely to be exacerbated by antiretroviral therapy. The introduction of highly active antiretroviral therapy (HAART) has dramatically changed the prognosis of HIV infection, decreasing morbidity and mortality rates, transforming a highly lethal disease into a chronic illness. However, the long-term use of HAART has been associated with the emergence of several drug-related adverse effects, such as hepatic steatosis, myopathy, cardiomyopathy, peripheral neuropathy, pancreatitis, diabetes mellitus, lipid metabolic dysfunction, lypodystrophy syndrome, and lactic acidosis [35–37] (see Chapters 9 and 21). NRTIs may act as competitive inhibitors of the human DNA polymerase γ, leading to depletion of mitochondrial DNA and damage of the respiratory chain [38,39]. D-NRTIs, including zalcitabine (ddC), didanosine (ddI), and stavudine (d4T), have the greatest afﬁnity for this enzyme [40]. Further mechanisms of NRTI-related mitochondrial toxicity have been elucidated for ZDV, such as inhibition of succinate transport, ADP/ATP exchange, adenylate kinase and cytochrome c oxidase activities, reduction in carnitine, and increased oxidative damage [41–43]. Painful peripheral neuropathy, observed in patients treated with d-NRTIs, is attributed to mitochondrial dysfunction arising from inhibition of mtDNA replication. Nerve biopses from patients with NRTI-associated peripheral neuropathy show abnormal mitochondria with excessive vacuolization, electron-dense inclusions, distorted cristae, and decreased mtDNA [44] (see Chapter 22). Moreover, ZDV therapy may affect both skeletal and cardiac muscle and kidney, often producing abnormal mitochondria with paracrystalline inclusions [45,46], although some mitochondrial abnormalities have also been reported in untreated individuals with HIV-related myopathy [47]. Because ZDV toxicity results from both its impact on mitochondrial enzymes and on mtDNA replication [48], measuring mtDNA content only is likely to understimate the mitochondrial dysfunction induced by ZDV. A rare adverse effect of NRTIs, also related to mitochondrial impairment, is lactic acidosis, a dramatic and potentially fatal condition associated with severe hepatic steatosis [35,49]. Moreover, a chronic, asymptomatic hyperlactataemia

[13 C]METHIONINE BREATH TEST

499

has been observed in patients receiving antiretroviral therapy in association with mtDNA depletion [36]. Finally, lipodystrophy following treatment with NRTI-containing HAART is associated with a decrease in peripheral blood lymphocytes and adipose tissue mtDNA and mtRNA content [50–53] (see Chapter 20). Given the important role of mitochondria in the pathogenesis of these HAARTrelated side effects, different approaches to study mitochondrial function have been tried. Among the available techniques for in vivo detection of drug-related mitochondrial toxicity, the measurement of ketone body ratios, as the ratio of the redox couples acetoacetate/β-hydroxybutyrate or pyruvate/lactate, has been evaluated in children with perinatal HIV infection treated with ZDV, without signiﬁcant results [54]. The quantiﬁcation of mtDNA and mtRNA in peripheral blood cells or tissues has been widely studied in HAART-treated, HIV-positive patients, but the choice of the best cell type for the measure is still under debate. Puriﬁed cell populations such as isolated CD4+ or CD8+ lymphocytes or platelet-depleted lymphocytes are probably the best option, but further investigation is needed [55]. Moreover, the direct monitoring of mitochondrial DNA in target tissue (adipose, hepatic, renal) could provide more appropriate information on the pathogenesis underlying a speciﬁc drug toxicity, but this requires repeated biopsies (see Chapter 21). An indirect and nonspeciﬁc tool for the diagnosis of mitochondrial toxicity is the measurement of lactatemia, which has frequently been found above the normal range in NRTI-treated subjects [56], even in the absence of any symptoms. We have explored the feasibility of the [13 C]methionine breath test for detection of NRTI-related liver mitochondrial impairment [58]. We evaluated four HIV-infected patients with hyperlactatemia but without lactic acidosis or elevated liver enzymes [57]. They were not affected by viral hepatitis and had started HAART approximately one year before the acute onset of symptoms of hyperlactatemia, such as nausea, vomiting, and peripheral neuropathy. An initial breath test was performed upon onset of symptoms, when plasma lactate levels were between 4.7 and 8.38 mM (normal range = 0.7 to 2.47 mM), and results were compared with healthy controls. Hyperlactatemic patients showed dramatically decreased 13 CO2 exhalation compared to controls. Interestingly, drug cessation or substitution of a “d-drug” with another NRTI (tenofovir) induced a rapid amelioration of both the clinical picture and the mitochondrial respiratory function, as shown by improved serum lactate levels and 13 C exhalation rates, although neither reached the levels of healthy controls (Figure 3). In a subsequent study, we evaluated liver mitochondrial function via the [13 C] methionine breath test by comparing patients who were (1) HIV infected but naive for antiretroviral treatment, (2) NRTI-multiexperienced HIV positive, (3) NRTItreated patients with hyperlactatemia, and (4) healthy subjects [58]. We enrolled 15 ART multiexperienced HIV-infected patients, six NRTI-treated patients with symptomatic hyperlactatemia, 11 HIV-infected patients naive for ART, and 10 healthy controls.

500

CLINICAL ASSESSMENT OF MITOCHONDRIAL FUNCTION Patients on treatment Patient 1

10

8

8

6

4

2

2 0

12

Healthy controls

Patient 2

6

4 0

0 30 45 60 75 90 105 120 0 30 45 60 75 90 105 120 Time from methionine administration (min)

Patient 3

12

10

10

8

8

DOB

DOB

12

10 DOB

DOB

12

Drug modification or suspension

6

Patient 4

6

4

4

2

2 0

0 0

30 45 60 75 90 105 120

0

30 45 60 75 90 105 120

Time from methionine administration (min)

Figure 3 13 CO2 breath exhalation curves of four patients with drug-related hyperlactatemia during therapy and after therapy modiﬁcation compared with healthy controls. (From Milazzo et al. [57].)

Similar 13 CO2 breath exhalation data, expressed as DOB, were obtained from healthy controls and HIV-infected subjects naive for ART. Asymptomatic HIV-positive patients on long-term antiretroviral treatment, and with normal blood lactate values, had repressed 13 CO2 exhalation values compared to controls. HIV-treated patients presenting symptoms of hyperlactatemia had even lower, and delayed, excretion proﬁles, which are reﬂected in the cumulative excretion dose per hour and cumulative dose values (Figures 4 and 5). Importantly, none of the patients had chronic viral liver disease, reported alcohol abuse, or was receiving any drug other than antiretrovirals. No correlation was found between serum liver enzymes and 13 CO2 excretion. Moreover, comparable 13 CO2 excretion patterns in untreated HIV-infected patients and healthy controls suggest that mitochondrial impairment is drug related rather than HIV related, at least in patients without advanced disease. These results indicate that the [13 C]methionine breath test is sensitive enough to detect early NRTI-related mitochondrial toxicity in HIV-positive patients, even before the appearance of liver biochemical abnormalities, alterations of serum lactate, or symtoms of hyperlactatemia.

[13 C]METHIONINE BREATH TEST

501

%13C cumulative dose

% dose of 13CO2 recovered per hour

Figure 4 Mean values of 13 CO2 breath excretion rate after administration of [13 C] methionine (DOB) in antiretroviral-naive HIV-infected patients (pts), patients on treatment with normal serum lactate, patients on treatment with hyperlactatemia, and healthy controls. p = 0.001 by ANOVA test of the four curves from 30 to 90 minutes; p < 0.05 for DOB at 60 minutes of (*) healthy controls and naive patients versus asymptomatic patients and (**) asymptomatic patients versus hyperlactatemic patients. (From Milazzo et al. [58].) Healthy controls Asymptomatic pts 9 8 7 6 5 4 3 2 1 0

4 3

Naive pts Hyperlactataemic pts

p = 0.03 p = 0.05

p = 0.05 p = 0.04

2 1 0

Figure 5 13 CO2 dose per hour and cumulative dose measured by breath test at 60 minutes in healthy controls, antiretroviral-naive HIV-infected patients (pts), patients on treatment with normal serum lactate, and patients on treatment with hyperlactatemia. p = 0.001 (% dose of 13 CO2 recovered per hour) by ANOVA; p = 0.02 (% 13 C cumulative dose). A comparison by Kruskall–Wallis test is reported. (From Milazzo et al. [58].)

502

CLINICAL ASSESSMENT OF MITOCHONDRIAL FUNCTION

Others have conﬁrmed these ﬁndings and shown that this test can detect chronic mitochondrial toxicity related to long-term exposure to antiretroviral therapy in patients with clinical evidence of lipoatrophy [59]. However, in contrast to our ﬁndings, Banasch et al. [59] did not ﬁnd a difference in [13 C]methionine excretion between asymptomatic ART-treated patients and healthy controls, but they did detect a signiﬁcant difference between untreated HIV-infected subjects and healthy controls. Interestingly, they also found a correlation between methionine breath test results and mtDNA content in PBMCs in ART-treated patients (see Chapter 21). These discrepancies are probably due to the varying characteristics of the patients. HAART-treated HIV-infected patients in our study [58] had been pretreated more aggressively, particularly with thymidine analogs and d-drugs. Moreover, Banasch et al. [59] described a signiﬁcant improvement of hepatic mitochondrial function in a small group of patients three months after starting antiretroviral therapy, whereas a longitudinal evaluation of 19 HIV-positive subjects in our study 12 months from the introduction of HAART did not reveal any improvement in the [13 C]methionine breath test [60]. 5. FUTURE PERSPECTIVES The observations described here indicate the need for further validation of the [13 C]methionine breath test for the routine diagnosis of drug-related mitochondrial toxicity: by performing studies on larger populations, with longer follow-up periods, and possibly by including other techniques, such as the mtDNA content of circulating PBMCs, tissue samples, and histological and ultrastructural examinations. To our knowledge, aside from a case report by Spahr et al. [29], no other drugs have been investigated for their potential mitochondrial toxicity via a [13 C]methionine breath test, even though it can provide a useful noninvasive method to discover such toxicities. Other ﬁelds of research for future applications of [13 C]methionine breath test are drug–drug interactions in patients coinfected by HIV and HCV undergoing HCV treatment. Among the most important drug interactions are those of ribavirin with didanosine and stavudine, where pathogenic mechanisms are via enhanced mitochondrial toxicity, possibly resulting in pancreatitis and lactic acidosis [61]. Another possible use of breath tests is monitoring of new therapeutic approaches to HAART-related mitochondrial impairment. Recently, supplementation with antioxidant or mitochondrial protective compounds have been investigated in an NRTI-treated population. The efﬁcacy of uridine supplementation in the treatment of mitochondrial toxicity due to pyrimidine d-drugs, by a competitive process, has been observed both in vitro and in vivo [62,63], and has been assessed in lipoatrophic HIV-infected patients using the [13 C]methionine breath test to quantify hepatic mitochondrial function [64] (see Chapter 9). Acetyl-l-carnitine has been proposed as a therapeutic agent for distal symmetrical polineuropathy in HIV-infected patients. It was demonstrated to increase

REFERENCES

503

nerve ﬁber density and to improve neuropathic pain in a high percentage of patients [65] (see Chapter 11). Many antioxidant compounds are under investigation to reduce mitochondrial DNA damage of NRTIs, and the [13 C]methionine breath test may offer a simple and noninvasive tool for the in vivo assessment of drug efﬁcacy and potential drug-induced mitochondrial dysfunction.

REFERENCES 1. Candelli M, Cazzato IA, Zocco MA, et al. 13 C-Breath tests in the study of mitochondrial liver function. Eur Rev Med Pharmacol Sci. 2004;8:23–31. 2. Hepner GW, Vesell ES. Assessment of aminopyrine metabolism in man after oral administration of 14 C-aminopyrine: effects of phenobarbital, disulﬁram and portal cirrhosis. N Engl J Med. 1974;291:1384–1388. 3. Schwabe A, Cozzetto F, Bennet L, Mellinkoff S. Estimation of fat absorption by monitoring of expired radio-active carbon dioxide after feeding a radioactive fat. Gastroenterology. 1962;42:285–321. 4. Schoeller DA, Schneider JF, Solomons NW, Watkins JB, Klein PD. Clinical diagnosis with the stable isotope 13 C in CO2 breath tests: methodology and fundamental considerations. J Lab Clin Med. 1977;90:412–421. 5. Henry DA, Sharpe G, Chaplain S, et al. The [14 C]-aminopyrine breath test: a comparison of different forms of analysis. Br J Clin Pharmacol. 1979;8:539–545. 6. Armuzzi A, Candelli M, Zocco MA, et al. Review article: breath testing for human liver function assessment. Aliment Pharmacol Ther. 2002;16:1977–1996. 7. Breen KJ, Bury RW, Calder IV, Desmond PV, Peters M, Mashford ML. A [14 C]phenacetin breath test to measure hepatic function in man. Hepatology. 1984;4:47–52. 8. Matsumoto K, Suehiro M, Iio M, et al. [13 C]methacetin breath test for evaluation of liver damage. Dig Dis Sci. 1987;32:344–348. 9. Watkins PB, Murray SA, Winkelman LG, Heuman DM, Wrighton SA, Guzelian PS. Erythromycin breath test as an assay of glucocorticoid-inducible liver cytochromes P-450. Studies in rats and patients. J Clin Invest. 1989;83:688–697. 10. Burke PA, Stack JA, Wagner D, Lewis DW, Jenkins RL, Forse RA. L-[1-(13)C] Phenylalanine oxidation as a measure of hepatocyte functional capacity in end-stage liver disease. Am J Surg. 1997;173:270–274. 11. Shreeve WW, Shoop JD, Ott DG, McInteer BB. Test for alcoholic cirrhosis by conversion of [14 C]- or [13 C]galactose to expired CO2. Gastroenterology. 1976;72:98–101. 12. Lauterburg BH, Liang D, Schwarzenbach FA, Breen KJ. Mitochondrial dysfunction in alcoholic patients as assessed by breath analysis. Hepatology. 1993;17:418–422. 13. Armuzzi A, Marcoccia S, Zocco MA, et al. Non-invasive assessment of human hepatic mitochondrial function through the 13 C-methionine breath test. Scand J Gastroenterol. 2000;35:650–653. 14. Miele L, Grieco A, Armuzzi A, et al. Hepatic mitochondrial beta-oxidation in patients with nonalcoholic steatohepatitis assessed by 13 C-octanoate breath test. Am J Gastroenterol. 2003;98(10): 2335–2336.

504

CLINICAL ASSESSMENT OF MITOCHONDRIAL FUNCTION

15. Harper AE, Miller RH, Block KP. Branched-chain amino acid metabolism. Annu Rev Nutr. 1984;4:409–454. 16. Bendtsen P, Hannestad U, Pahlsson P. Evaluation of the carbon 13–labeled ketoisocaproate breath test to assess mitochondrial dysfunction in patients with high alcohol consumption. Alcohol Clin Exp Res. 1998;22:1792–1795. 17. Honkoop P, De Man RA, Scholte HR, et al. Effect of lamivudine on morphology and function of mitochondria in patients with chronic hepatitis B. Hepatology. 1997;26:211–215. 18. Steele RD, Benevenga NJ. Identiﬁcation of 3-methylthiopropionic acid as an intermediate in mammalian methionine metabolism in vitro. J Biol Chem. 1978;253: 7844–7850. 19. Cooper AJ. Methionine transamination in vivo. Biochem J. 1989;262:689–691. 20. Wu GY, Thompson JR. Is methionine transaminated in skeletal muscle? Biochem J. 1989;257:281–284. 21. Finkelstein JD. Methionine metabolism in mammals. J Nutr Biochem. 1990;1:228– 237. 22. Harper AE. Some recent developments in the study of amino acid metabolism. Proc Nutr Soc. 1983;42:437–449. 23. Mato JM, Alvarez L, Ortiz P, Pajares MA. S -Adenosylmethionine synthesis: molecular mechanisms and clinical implications. Pharmacol Ther. 1997;73:265–280. 24. Mudd SH, Poole JR. Labile methyl balances for normal humans on various dietary regimens. Metabolism. 1975;24:721–735. 25. Stipanuk MH. Metabolism of sulfur-containing amino acids. Annu Rev Nutr. 1986;6:179–209. 26. Stipanuk MH. Sulfur amino acid metabolism: Pathways for production and removal of homocysteine and cysteine. Annu Rev Nutr. 2004;24:539–577. 27. Mudd SH, Ebert MH, Scriver CR. Labile methyl balances in the human: the role of sarcosine. Metabolism. 1980;29:707–720. 28. Frisell WR, Cronin JR, Mackenzie CG. Coupled ﬂavoenzymes in mitochondrial oxidation of n-methyl groups. J Biol Chem. 1962;237:2975–2980. 29. Spahr L, Negro F, Rubbia-Brandt L, et al. Acute valproate-associated microvesicular steatosis. Could the 13 C-methionine breath test be useful to assess liver mitochondrial function? Dig Dis Sci. 2001;46:2758–2761. 30. Spahr L, Negro F, Leandro G, et al. Impaired hepatic mitochondrial oxidation using the 13 C-methionine breath test in patients with macrovesicular steatosis and patients with cirrhosis. Med Sci Monit. 2003;9:CR6–CR11. 31. Raidel SM, Haase C, Jansen NR, et al. Targeted myocardial transgenic expression of HIV Tat causes cardiomyopathy and mitochondrial damage. Am J Physiol Heart Circ Physiol. 2002;282:H1672–H1678. 32. Jacotot E, Ravagnan L, Loefﬂer M, et al. The HIV-1 viral protein R induces apoptosis via a direct effect on the mitochondrial permeability transition pore. J Exp Med. 2000;191:33–46. 33. Li YY, Chen D, Watkins SC, Feldman AM. Mitochondrial abnormalities in tumor necrosis factor-alpha-induced heart failure are associated with impaired DNA repair activity. Circulation. 2001;104:2492–2497.

REFERENCES

505

34. Geng Y, Hansson GK, Holme E. Interferon-gamma and tumor necrosis factor synergize to induce nitric oxide production and inhibit mitochondrial respiration in vascular smooth muscle cells. Circ Res. 1992;71:1268–1276. 35. Boubaker K, Flepp M, Sudre P. et al. Hyperlactatemia and antiretroviral therapy: the Swiss HIV Cohort Study. Clin Infect Dis. 2001;33:1931–1937. 36. Cote HC, Brumme ZL, Craib KJ, et al. Changes in mitochondrial DNA as a marker of nucleoside toxicity in HIV-infected patients. N Engl J Med. 2002;346:811–820. 37. Lewis W. Mitochondrial dysfunction and nucleoside reverse transcriptase inhibitor therapy: experimental clariﬁcations and persistent clinical questions. Antivir Res. 2003;58:189–197. 38. Johnson AA, Ray AS, Hanes J, et al. Toxicity of antiviral nucleoside analogs and the human mitochondrial DNA polymerase. J Biol Chem. 2001;276:40847–40857. 39. White EL, Parker WB, Macy LJ, et al. Comparison of the effect of Carbovir, AZT, and dideoxynucleoside triphosphates on the activity of human immunodeﬁciency virus reverse transcriptase and selected human polymerases. Biochem Biophys Res Commun. 1989;161:393–398. 40. Martin JL, Brown CE, Matthews-Davis N, Reardon JE. Effects of antiviral nucleoside analogs on human DNA polymerases and mitochondrial DNA synthesis. Antimicrob Agents Chemother. 1994;38:2743–2749. 41. Barile M, Valenti D, Hobbs GA, et al. Mechanisms of toxicity of 3 -azido-3 -deoxythymidine: its interaction with adenylate kinase. Biochem Pharmacol. 1994;48:1405– 1412. 42. Valenti D, Barile M, Passerella S. AZT inhibition of the ADP/ATP antiport in isolated rat heart mitochondria. Int J Mol Med. 2000;6:93–96. 43. Valenti D, Atlante A, Barile M, Passarella S. Inhibition of phosphate transport in rat heart mitochondria by 3 -azido-3 -deoxythymidine due to stimulation of superoxide anion mitochondrial production. Biochem Pharmacol. 2002;64(2): 201–206. 44. Dalakas MC, Semino-Mora C, Leon-Monzon M. Mitochondrial alterations with mitochondrial DNA depletion in the nerves of AIDS patients with peripheral neuropathy induced by 2 3 -dideoxycytidine (ddC). Lab Invest. 2001;81:1537–1544. 45. McCurdy DT, III Kennedy JM. AZT decreases rat myocardial cytochrome oxidase activity and increases beta-myosin heavy chain content. J Mol Cell Cardiol. 1998;30:1979–1989. 46. Dalakas MC, Illa I, Pezeshkpour GH, Laukaitis JP, Cohen B, Grifﬁn JL. Mitochondrial myopathy caused by long-term zidovudine therapy. N Engl J Med. 1990;322:1098–1105. 47. Morgello S, Wolfe D, Gadfrey E, Feinstein R, Tagliati M, Simpson DM. Mitochondrial abnormalities in human immunodeﬁciency virus-associated myopathy. Acta Neuropathol. 1995;90:366–374. 48. Lewis W, Kohler JJ, Hosseini SH, Haase CP, et al. Antiretroviral nucleosides, deoxynucleotide carrier and mitochondrial DNA: evidence supporting the DNA Pol gamma hypothesis. AIDS. 2006;20(5): 675–684. 49. Chattha G, Arieff AI, Cummings C, Tierney LM, Jr. Lactic acidosis complicating the acquired immunodeﬁciency syndrome. Ann Intern Med. 1993;118:37–39. 50. Cossarizza A, Riva A, Pinti M, et al. Increased mitochondrial DNA content in peripheral blood lymphocytes from HIV-infected patients with lipodystrophy. Antivir Ther. 2003;8:315–321.

506

CLINICAL ASSESSMENT OF MITOCHONDRIAL FUNCTION

51. Hoy JF, Gahan ME, Carr A, et al. Changes in mitochondrial DNA in peripheral blood mononuclear cells from HIV-infected patients with lipoatrophy randomized to receive abacavir. J Infect Dis. 2004;190:688–692. 52. Shikuma CM, Hu N, Milne C, et al. Mitochondrial DNA decrease in subcutaneous adipose tissue of HIV-infected individuals with peripheral lipoatrophy. AIDS. 2001;15:1801–1809. 53. Galluzzi L, Pinti M, Guaraldi G, et al. Altered mitochondrial RNA production in adipocytes from HIV-infected individuals with lipodystrophy. Antivir Ther. 2005;10S2:M91–M99. 54. Abe K, Matsuo Y, Kadekawa J, Inoue S, Yanagihara T. Measurement of tissue oxygen consumption in patients with mitochondrial myopathy by noninvasive tissue oximetry. Neurology. 1997;49:837–841. 55. Cossarizza A, Reiss P. Mitochondria, HIV infection and its treatment: Where do we go from here? Antivir Ther. 2005;10S2:M1–M2. 56. Harris M, Chan KJ, Tesiorowski AM, et al. Random venous lactate levels among HIV-positive patients on antiretroviral therapy. J Acquir Immune Deﬁc Syndr. 2002;31:448–450. 57. Milazzo L, Riva A, Sangaletti O, Piazza M, Antinori S, Moroni M. 13 C-Methionine breath test detects liver mitochondrial impairment in HIV-infected patients with antiretroviral drug-related hyperlactatemia. J Acquir Immune Deﬁc Syndr. 2004;35:429–432. 58. Milazzo L, Piazza M, Sangaletti O, et al. [13 C]Methionine breath test: a novel method to detect antiretroviral drug-related mitochondrial toxicity. J Antimicrob Chemother. 2005;55:84–89. 59. Banasch M, Goetze O, Hollborn I, et al. 13 C-methionine breath test detects distinct hepatic mitochondrial dysfunction in HIV-infected patients with normal serum lactate. J Acquir Immune Deﬁc Syndr. 2005;40:149–154. 60. Milazzo L, Menzaghi B, Massetto B, Sangaletti O, Riva A. 13 C-Methionine breath test detects drug-related hepatic mitochondrial dysfunction in HIV-infected patients. J Acquir Immune Deﬁc Syndr. 2006;41(2): 252–253. 61. Garcia-Benayas T, Blanco F, Soriano V. Weight loss in HIV-infected patients. N Engl J Med. 2002;347:1287–1288. 62. Sommadossi JP, Carlisle R, Schinazi RF, Zhou Z. Uridine reverses the toxicity of 3 -azido-3 -deoxythymidine in normal human granulocyte-macrophage progenitor cells in vitro without impairment of antiretroviral activity. Antimicrob Agents Chemother. 1988;32:997–1001. 63. Walker UA, Venhoff N, Koch EC, Olschewski M, Schneider J, Setzer B. Uridine abrogates mitochondrial toxicity related to nucleoside analogue reverse transcriptase inhibitors in HepG2 cells. Antivir Ther. 2003;8:463–470. 64. Banasch M, Goetze O, Knyhala K, et al. Uridine supplementation enhances hepatic mitochondrial function in thymidine-analogue treated HIV-infected patients. AIDS. 2006;20:1554–1556. 65. Hart AM, Wilson AD, Montovani C, et al. M. Acetyl-l-carnitine: a pathogenesis based treatment for HIV-associated antiretroviral toxic neuropathy. AIDS. 2004;18:1549–1560.

22 ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY Ingrid Pruimboom-Brees Drug Safety, GlaxoSmithKline, Ware, Hertfordshire, UK

Germaine Boucher, Amy Jakowski, and Jeanne Wolfgang Drug Safety R&D, Pﬁzer, Inc., Groton, Connecticut

1. Introduction 1.1. Normal mitochondrial morphology 1.2. Mitochondrial fusion and fission 1.3. Mitochondrial OXPHOS and assays 2. Mitochondria and cell death 2.1. Necrosis 2.2. Apoptosis 3. Mitochondriopathies 3.1. Oxidative stress 3.2. Autophagy and mitophagy 3.3. Calcium densities 3.4. Glycogen and fat deposition 3.5. Mitochondrial response in metabolic diseases 3.6. Drug-induced mitochondriopathies 3.7. Morphology of mitochondria in cancer 3.8. Mitochondrial DNA mutations and diseases 4. Conclusions

508 508 511 512 513 513 516 518 518 520 520 521 522 523 530 531 534

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

507

508

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

1. INTRODUCTION Detection and interpretation of microscopic injury requires a systematic analysis of all cellular compartments discernible by either light or ultrastructural microscopy, in contrast to normal histology. Although resolution is ﬁner in ultrastructural studies, the small size of the tissue sample and the thin sections required limit its spatial coverage. Therefore, light microscopy is critical to determine location, extension, and potentially timing of a pathologic process within a tissue. Combined with special techniques such as staining, immunohistochemistry, and histochemistry, an overall characterization of the cell/organ morphology can also be achieved, providing insight into function and pathogenesis. During ultrastructural examination, shape, size, relative number, and location of each cellular component must be ascertained precisely before initiating labor-intensive techniques such as quantitative immunoelectron microscopy capable of illuminating organelle integrity and function. A rough determination of size can be achieved by comparing the size of the structure of interest with ribosomes and β-glycogen monoparticles, which are approximately 22 and 29 nm in diameter, respectively, and are in nearly every cell [7]. The focus of this chapter is twofold: to review the normal mitochondrial structure and to describe how microscopic techniques can be used to diagnose mitochondriopathies and investigate their pathogenic mechanisms, including drug-induced mitochondrial dysfunction. 1.1. Normal Mitochondrial Morphology Since the advent of transmission electron microscopy (TEM) in 1931, the structure of mitochondria has been investigated extensively. Insight into three-dimensional structure was hampered by the requirement of very thin sections, but organelle and cellular anatomy was soon resolved via serial sections and stereological methods. More recently, advances in high-voltage TEM instruments with sufﬁcient power to penetrate thick sections and massively parallel computing systems have enabled high-resolution three-dimensional imaging of complex structures, notably mitochondria. Analysis of tilt images of thick sections via electron tomography has revealed previously undetected details that have improved current models of mitochondrial structure and associated function (Figure 1). For example, the discovery of crista junctions, where the inner membrane invaginates to form cristae, suggests further compartmentalization of inner membrane function than appreciated previously [1–3]. Regardless of the microscopic technique at hand, careful consideration must be given to choose the chemical ﬁxation procedure that best preserves the mitochondrial component of interest. Strict glutaraldehyde ﬁxation results in a well-ﬁxed matrix that upon staining will be too dark and will obscure mitochondrial membranes [4]. Some authors have suggested the use of permanganate to reduce the number of artifacts linked to osmium ﬁxation and to increase the deﬁnition of membranes such as myelin sheaths and nuclear and mitochondrial membranes [5].

INTRODUCTION

509

Figure 1 Dendritic mitochondrion from chick cerebellum. Electron tomography reveals unprecedented details of mitochondrial structure, including ﬁne structure of junctions between cristae and the inner membrane. The coarse texture apparent on the cristae is caused by the respiratory complexes. The outer membrane is shown in purple, the inner boundary membrane in aqua, and the cristae in yellow, green-gray, and red to demonstrate variety of cristal structure. [From G. A. Perkins (University of California–San Diego) and T. G. Frey (San Diego State University), with permission.] (See insert for color representation of ﬁgure.)

We found that a buffered solution of 2.5% glutaraldehyde/2% paraformaldehyde, followed by a secondary ﬁxation with 0.05 M potassium ferrocyanide in 1% osmium tetroxide, increases the contrast of membranes in comparison to the matrix and facilitates glycogen retention in tissues [6]. With a cryopreserved specimen, ﬁxation with 1% osmium (in acetone) prior to processing often results in more closely apposed outer and inner mitochondrial membranes than when using glutaraldehyde/paraformaldehyde. The model of mitochondrial structure that emerges from these techniques is that the organelle is delineated by two morphologically and functionally distinct membranes. The outer membrane (OM) represents a macromolecular barrier freely permeable to solutes up to few thousand daltons in molecular weight through pore-forming proteins. The inner mitochondrial membrane (IM) is impermeable, even for small solutes such as protons and metabolic substrates that must be transported through protein carriers. The IM is also the most protein-rich lipid bilayer in all biological systems, most of these proteins being involved in oxidative phosphorylation (OXPHOS) and ATP/ADP trafﬁcking. The IM is organized in an inner boundary membrane (IBM) and a cristae membrane (CM). The IBM interacts with the OM through punctuate contact sites and with the CM through cristae junctions, which are tubular openings uniformly about 30 nm in diameter [1,2]. The CM represents invaginations of the IBM by juxtaposition of two inner membrane leaﬂets forming narrow tubular or lamellar cristae, the density of which reﬂects the aerobic poise of the cell, while the shape and arrangement (lamellar or tubular) are often cell-type speciﬁc. The shape of the IM can inﬂuence OXPHOS function: narrow cristae junctions restrict diffusion between

510

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

intracristal and external compartments, causing depletion of ADP and decreased ATP output inside the cristae [8]. The IM delineates the mitochondrial matrix, which contains multiple copies of the mitochondrial DNA (mtDNA) and the proteins needed for its expression and replication, the Krebs cycle enzymes, and a host of other proteins. Ultrastructurally, mtDNA appears as tiny ﬁlaments in the electron–lucent matrix. Mitochondrial morphology and function differ between tissues and cellular locations. For example, there is a direct correlation between number and length of mitochondria cristae, mitochondrial respiration rate, and the aerobic demand of the cell. Brown adipocytes, which rely heavily on mitochondrial OXPHOS for thermogenesis, are densely packed with mitochondria characterized by numerous stacked lamellar cristae (Figure 2). Cardiomyocytes (cardiac myoﬁbers) that rely almost exclusively on aerobic respiration correspondingly have a high density of mitochondria containing long cristae (Figure 3). Osmium extraction of cardiac tissue combined with high-resolution scanning electron microscopy revealed that within interﬁbrillar mitochondria and compared to subsarcolemmal mitochondria, cristae are primarily tubular, and this feature correlates biochemically with higher

Figure 2 Ultrastructure of rat hibernoma. The cytoplasm of brown adipocytes contain numerous round mitochondria of variable size and few lipid vacuoles. These mitochondria characteristically contain lamellar cristae encompassing the entire width of the organelle. Note differences between lamellar cristae here and tubular cristae, as found in mammalian heart in Figure 3. Bar = 500 nm.

511

INTRODUCTION

1µ

Figure 3 Ultrastructure of mammalian (bat) heart mitochondria. Note the abundance of cristae, which reﬂects aerobic poise and high metabolic demand, and the proliferation of sarcolemmal tubules, which reﬂects Ca2+ uptake capacity, and hence is a determinant of maximal heart rate. (Image by Keith Porter; from ASCB Image Library, with permission.)

succinate dehydrogenase (SDH) and citrate synthase (Krebs cycle) activities and respiration rates. In contrast, subsarcolemmal mitochondria contain mainly lamelliform cristae [9–11]. Skeletal muscles are composed of two major types of myoﬁbers, which contract with slow or fast velocities. The fast-twitch myoﬁbers contain fewer mitochondria, are anaerobically poised, and fatigue relatively fast, whereas the aerobically poised slow-twitch myoﬁbers contain many mitochondria and can sustain long periods of moderate activity (see Chapter 7). All of these structural and biochemical differences are plastic and respond to physiological and pathological conditions. For example, there is a signiﬁcant increase in the subsarcolemmal mitochondrial volume in slow-twitch oxidative myoﬁbers after endurance training in rats and this change also correlates with increased SDH activity. This increase is probably required to supply the energy for the active transport of metabolites through the sarcolemmal membrane. [12,13]. 1.2. Mitochondrial Fusion and Fission Mitochondria are typically pictured as bean-shaped organelles, but in many cells they also fuse to form an extended reticulum. Overall mitochondrial morphology is a balance between opposing processes of fusion and ﬁssion [14,15] that

512

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

affect the interpretation of ultrastructural changes signiﬁcantly during a pathologic process. Large, conserved GTPases, mitofusin 1 and 2 (Mfn1 and Mfn2) and OPA1, the disease gene in autosomal dominant optic atrophy, are the main proteins involved in the fusion machinery; other proteins are still being described. Mfn 1 and 2 are localized to the OM, with most of the protein exposed to the cytosol. Homotypic interactions between Mfn1 molecules spanning adjacent mitochondria result in tethering of the OM, while the actual fusion involves interaction of the GTPase domain with other proteins. Fusion also involves OPA1 located in the intermembrane space in close association with the inner membrane. For ﬁssion, dynamin-related protein 1 and Fis1 are key players. Dynamin-related protein 1 (Drp1) is mainly a cytosolic GTPase with a subpool of punctuate spots on mitochondrial tubules that seem to be critical for the constriction process during ﬁssion. Recruitment of Dnm1 (the yeast equivalent to the mammalian Drp1) to the mitochondrial membrane seems to depend on Fis1, a small OM protein comprising a helical domain facing the cytosol that could act as a Fis1 binding site [14,15]. The biological signiﬁcance of fusion/ﬁssion is multifaceted, being critical for mitochondrial function, organismal development, organelle interaction, cell signaling, and viability. Fusion tends to promote intermitochondrial cooperation where by substrates and ATP generation are more uniformly distributed; ﬁssion enables compartmentalization, delivery of isolated mitochondria to distant parts of the cell such as in axonal synapses, and equitable distribution of mitochondria to the daughter cells during cell division. The equilibrium between ﬁssion and fusion shifts in response to stressors. For example, lack of fusion causes mitochondrial fragmentation, heterogeneity in membrane potential, and compromised oxygen consumption. In such cells, mitochondria become autonomous organelles in which dysfunction such as damage to mitochondrial DNA (mtDNA) or depletion of metabolites or substrates cannot be remedied via fusion with healthy mitochondria [14,15]. 1.3. Mitochondrial OXPHOS and Assays The molecular composition of mitochondrial membranes, together with their dynamics, biogenesis, and function, was studied extensively by quantitative electron microscopy using immunogold labeling. Vogel. et al found that the distribution of proteins involved in OXPHOS proteins, protein translocation, metabolite exchange, and mitomorphology is dynamic (subcompartmentalization of proteins is physiologically responsive) and uneven between IBM and cristae [3]. Protein subcompartmentalization is also physiologically responsive. For example, the IBM is enriched in components of the protein import and mitochondrial fusion/ﬁssion machinery, where they can interact with similar components of the facing OM. The stable molecular organization in the CM is achieved because the F1 F0 -ATP synthase serves as a scaffold for other respiratory complexes and prevents these complexes from diffusing to the IBM through the narrow cristae junctions [1–3].

MITOCHONDRIA AND CELL DEATH

513

Histochemistry is considered the gold standard for the investigation of mitochondrial OXPHOS in frozen tissues where protein cross-linking has not been caused by ﬁxation. Protocols are available for respiratory complexes I (nicotinamide adenine dinucleotide reductase) [16], II (succinate dehydrogenase) [17], and IV (cytochrome c oxidase) [18,19] (Figure 4). In tissues, however, complex I activity is rotenone insensitive, due to the nature of the chemical reaction coupled to the dye. Histochemical assessment of glycerol-3-phosphate dehydrogenase is used occasionally as a complementary marker of glycolysis which accelerates in compensation for decreases in mitochondrial β-oxidation [71,72]. In the last few years, however, antibodies that recognize subunits of all of the OXPHOS complexes have become available [20], which permits immunohistochemical detection of OXPHOS proteins in primary cells and in formalin-ﬁxed, parafﬁn-embedded sections and quantiﬁcation by various image analysis methods, such as laser scanning cytometry (Pruimboom-Brees, unpublished data). An example of the former is the staining of ﬁbroblasts from three patients with Leigh’s syndrome (Figure 5), a mitochondrial disease caused by mutations in nuclear DNA (see Chapter 11). The patients had three different mutations. The antibodies used were NDUFS3 for Complex I, the 30-kDa succinate dehydrogenase subunit of complex II, core protein 2 of complex III, subunit I of complex IV, oligomycin sensitivity–conferring protein for ATP synthase, and the E1α subunit of PDH [20]. This technique identiﬁed a lack of full assembly of complex I in patient 1, PDH impairment in patient 2 and cytochrome c oxidase defect in patient 3 [16].

2. MITOCHONDRIA AND CELL DEATH Mitochondria participate in cell death via both necrosis (from Greek nekrosis, the death of tissue or cells in a living animal) and apoptosis (Greek apo, from ptosis, falling). 2.1. Necrosis During hypoxia-related necrosis, a number of ultrastructural changes are apparent and well documented by Cheville [7]. Within seconds of anoxia, mitochondria swell and there is ﬂocculation of matrix proteins, membrane disintegration, and cristaelysis. This loss of impermeability of the IM signiﬁes irreversible mitochondrial morphological and functional (ATP generation) failure. Mitochondria ﬁnally appear as vacuoles outlined by a dense cytoplasmic mass. Calcium ﬂuxes contribute to mitochondrial failure, in that increasing cytosolic Ca2+ , a reﬂection of Ca2+ -ATPase failure, enters the mitochondria down the electrochemical gradient via an extraordinarily fast transporter in the IM. The efﬂux pumps are slower and inefﬁcient since they require energy (no longer available). The net effect is acute mitochondrial Ca2+ accumulation, which accelerates free-radical production and induces irreversible permeability transition (details in Chapters 2 and 21).

514

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

A

B

C

D

E

F

Figure 4 Histochemistry assessment of myoﬁber type in fast-twitch rat extensor digitorum longus (EDL; panels A, C, and E) and slow-twitch postural soleus muscle (panels B, D, and F), stained for complex 1 activity (NADH stain; top pair), complex II (succinate dehydrogenase; middle pair), and Complex IV (cytochrome c oxidase; bottom pair). Note the heterogeneous distribution of ﬁber type in EDL, with larger, anaerobically poised, fast-twitch ﬁbers appearing fainter than the aerobically poised, mitochondrially enriched slow ﬁbers. Note also the relatively more homogeneous ﬁber population in soleus, consisting of ﬁbers containing intermediate mitochondrial levels. Magniﬁcation in all is 200x. (See insert for color representation of ﬁgure.)

515

MITOCHONDRIA AND CELL DEATH Patient 1

Patient 2

Patient 3

CI

CII

CIII

CIV

CV

PDH

Figure 5 Immunohistochemical analysis of mitochondrial dysfunction. Fibroblasts from three patients with Leigh’s syndrome, due to three different nuclear DNA mutations, are labeled with a porin mAb (red) as a mitochondrial marker, and a second mAb (green) against the OXPHOS complex I, II, III, or IV, ATP synthase, or PDH indicated. Nuclei are stained with DAPI (blue). The merged red, green, and blue images are shown. Cells with a reduced labeling of a particular mitochondrial complex appear red, while normal levels of a particular complex appear yellow. (From Capaldi et al. [20], with permission.) (See insert for color representation of ﬁgure.)

516

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

At the end, calcium phosphates precipitate out in the form of electron-dense granular inclusions in the mitochondrial matrix, a hallmark of necrosis, and/or as electron-dense radiating structures on many cytoplasmic organelles or elements. Concomitantly, glycogen particles disappear from the cytoplasm as the mitochondrial ATP production decreases, and the cell metabolism shifts toward less efﬁcient glycolysis. Another hallmark of the transition from reversible to irreversible cellular injury is the lysis and detachment of the cytoskeleton from the plasma membrane, which is followed by cytoplasmic blebbing and loss of cilia and microvilli [22]. Cytoplasm blebbing is a reﬂection of membrane lipid peroxidation engendered by free-radical generation, activation of membrane phospholipidases, uncoupling of gap junctions, and depolymerization of microﬁlaments by elevated intracellular calcium levels. A disorganized cytoskeleton, broken intercellular junctions, decreased cytoplasmic volume, and protein aggregation give to the cell a more rounded, darker, and denser aspect on light microscopy. Cells also become progressively isolated from adjacent cells and dislodge from their normal positions. Adequate attention to the underlying basal laminae of cells, which are often damaged during degeneration, is important, as it is an indication of how well the tissue might recover. If the basal lamina persists, it provides a scaffold for migration, positioning, and attachment of new cells. If it is destroyed, tissue regeneration can be delayed or prevented and the tissue architecture may never recover. Nuclear changes vary considerably based on the type of injury. During a hypoxemic event, nuclei appear dark and shrunken, with irregular or ruptured nuclear membranes (pyknosis or karryorrhexis), and ﬁnally, dissolution of nucleus and loss of chromatin (karryolysis). In contrast, when apoptosis is triggered by nuclear toxins and viruses, the nuclear volume may persist and there is a speciﬁc pattern of chromatin aggregation and nuclear appearance called chromatin margination. On the other end, if the nucleus lyses during apoptosis, the remaining cytoplasmic mass appears as a darkly stained, anucleate, eosinophilic globule, sometimes referred to as a Councilman hyaline body. 2.2. Apoptosis Compared to necrosis, the cellular changes accompanying apoptosis evolve in an orderly and reproducible sequence because they follow a genetically deﬁned succession of events leading to programmed cell death. Ultrastructural changes may not themselves be pathognomonic of apoptosis (e.g., some features are shared with necrosis) [21–23]. Membrane blebbing is exaggerated so that the cell surface presents protuberances that pinch off to form extracellular membrane-bound apoptotic bodies (Figure 6), which are phagocytized by neighboring cells and macrophages, thereby avoiding release of cytoplasmic enzymes that would elicit an inﬂammatory reaction. There is loss of intercellular junctions, microvilli, and cilia as well, and a progressive rounding of the cell. The nuclei become convoluted so that the nuclear membrane invaginates into the nucleoplasm. Massive

517

MITOCHONDRIA AND CELL DEATH

A

B

Figure 6 Drug-induced apoptosis in the liver of a Sprague–Dawley rat via transmission electron microscopy: (A) membrane blebbing is exaggerated, resulting in the formation of apoptotic bodies; (B) membrane blebs contain various organelles such as ribosome and fragmented mitochondria. The scale is indicated; ﬁxation as in Figure 2. (Courtesy of Germaine Boucher, Pﬁzer Inc.)

deposits of chromatin line the inner face of the nuclear envelope, a process referred to as chromatin margination. The nuclear membrane remains intact until the cell lyses. Compared to necrosis, cytoplasmic changes are relatively slow because organelles maintain their integrity longer. Mitochondria are small, fragmented, but still distinct until the cell is lysed. Autophagosomes form and contain membranous whorls. Although mitochondrial fragmentation can occur without activation of apoptosis, apoptosis requires activation of the mitochondrial ﬁssion machinery. In healthy cells, mitochondria exist as a tissue-speciﬁc network of interconnected organelles, but early during apoptosis, this network can undergo fragmentation and perinuclear clustering (Figure 7) [20–24]. The fusion protein Opa1 participates by controlling the diameter of the cristae junction, and its release from the mitochondrial membrane inhibits mitochondrial fusion in cells committed to die. During apoptotic ﬁssion, opening of the narrow tubular cristae junction and fusion of individual cristae results in the release of cristae-associated cytochrome c with subsequent activation of caspases and other apoptotic mediators, including Bcl-2 family members such as Bid, Bim, Bik, Bak, and Bax, among others [21–23]. Colocalization studies place Bax, Drp-1, and Mfn-2 at ﬁssion sites where Bax activates Drp-1-dependent ﬁssion and inhibits fusion through its interaction with Mfn-2. Finally, migration of the ﬁssion protein Drp-1 to the mitochondria at ﬁssion sites involves activation of the cytosolic calcineurin, which results in Drp-1 dephosphorylation and dissociation from the cytosolic calcineurin-Drp-1 complex with subsequent mitochondrial translocation [23].

518

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

Figure 7 Perinuclear aggregation of mitochondria in WI38VA cells (transformed human lung ﬁbroblasts) undergoing apoptosis induced by a tetramethoxystilbene (MR-4) analog of resveratrol (R-3), which is a trihydroxystilbene. The IC50 value of MR-4 on the growth inhibition of transformed cells was 0.5 µM, compared to the value of greater than 50 µM for normal WI38 cells. Such rapid appearance of perinuclear mitochondrial clustering mitochondria suggests that this signal could serve as an early target of MR-4. (From Gosslau et al. [24].)

3. MITOCHONDRIOPATHIES 3.1. Oxidative Stress In addition to serving as the main intracellular energy source, mitochondria play an important role in the maintenance of the cellular redox status by being an

MITOCHONDRIOPATHIES

519

important source of reactive oxygen species (ROS). Oxygen is normally tetravalently reduced to form H2 O by complex IV. However, complex I (iron-sulfur clusters) and complex III (Q0 semiquinone) can autoxidize and univalently reduce O2 to the superoxide radical O2 −· . This superoxide is released at both sides of the inner mitochondrial membrane via ubiquinone–complex III interactions [25]. Estimates vary, but between 1 and 6% of the O2 consumed during normal respiration is univalently reduced to form superoxide radical O2 −· . Superoxide enzymatically or spontaneously dismutates to form hydrogen peroxide, H2 O2 , which in turn can form hydroxyl radical, the most reactive and damaging species known in biological systems. Several of the redox active centers of the respiratory complexes are inactivated by free-radical exposure, so that radical production inhibits respiration, which in turn increases subsequent radical production in a deleterious feedforward process [26] (see Chapter 18). It bears reiteration that mitochondrial O2 −· production and detoxiﬁcation differ among tissues and species. Using Amplex Red assay for H2 O2 , Anderson and Neufer showed that slow-twitch myoﬁbers (soleus) have relatively higher glutathione peroxidase activity and correspondingly higher oxidative exposures than those of glycolytically poised fast-twitch myoﬁbers [27]. This could potentially exacerbate statin-induced mitochondrial failure in fast-twich myoﬁbero due to bioaccumulation via monocarboxylate transporter isoform 4 activity [28] (see Chapter 7). The superoxide dismutase (SOD) family of enzymes, found in all aerotolerant organisms, catalyze the reaction of O2 into water and oxygen. The family includes several forms of copper/zinc-containing SOD (Cu/Zn-SOD) found in the cytosol and intercellular space of metazoans, iron-containing SOD of bacteria and chloroplasts, and manganese-dependent SOD (MnSOD) in all mitochondria and some bacteria. MnSOD is nuclear encoded and hence imported into the mitochondria. MnSOD knockout mice die 5 to 21 days after birth, which is not the case for Cu/Zn-SOD and EC-SOD knockout mice, underscoring the importance of MnSOD and mitochondrial function for survival [29]. To localize H2 O2 in vivo by TEM or confocal laser scanning microscopy, laboratory animals can be perfused, or tissues immersed, in a solution of 2 mM cerium chloride. This substance, in the presence of endogenous H2 O2 , results in the generation of cerium perhydroxide, which appears as small electron-dense particles by TEM. Immunohistochemistry or immunoelectron microscopy targeting MnSOD can provide complementary information on the oxidative stress status of mitochondria during a pathologic process [30,31]. These methods were used to demonstrate that the suppression of mitochondrial oxidative stress in aerobically poised retinal ganglion cells and axons constituting the optic nerve provides long-term neuroprotection in experimental optic neuritis [32]. Mitochondrial responses to reactive nitrogen species (RNS) are complicated by endogenous NO production that normally regulates respiration rate by reversibly modulating complex IV electron ﬂux. NO binds to the binuclear site competitively with O2 and cyanide, with afﬁnities that depend on the redox status of the Cu and Fe [33]. However, excessive cellular generation of nitric oxide (NO) and other RNS reactive nitrogen species will lead to mitochondrial dysfunction.

520

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

For example, peroxynitrite, the product of reaction of NO with O2 − , can break down to produce hydroxyl radical that inhibits mitochondrial complexes I, II, and V and cytochrome c through the nitration of critical tyrosine residues. The presence of these 3-tyrosine nitrated proteins can be assessed by immunohistochemistry on formalin-ﬁxed tissues or immunoelectronmicroscopy, thereby providing evidence of oxidative stress [34]. Similar histological assessments reveal oxidative stress in human acute pancreatitis and in alveolar type II cells in lungs of rats given IL-1 intratracheally [35,36]. 3.2. Autophagy and Mitophagy Autophagy is a closely regulated process of degrading and recycling cellular constituents that not only removes excess or damaged organelles during physiological turnover and pathological processes, but also liberates free amino acids and other nutrients during starvation. Fasting is a potent inducer of autophagy in hepatocytes and is partially regulated by the glucagon status. Autophagy is also observed in nerve cells following treatment with amphetamines, tryptamine derivatives, or exposure to the mitochondrial neurotoxin 1-methyl-4-phenylpyridinium [37]. The mechanisms underlying autophagy are not fully understood, and many studies have been performed in vitro with speciﬁc cellular dyes and techniques such as cytochemistry for autophagy genes Atg12, Atg5, and Atg8/LC3 [cite] or monodansylcadaverine staining for lysosomes [38,39]. The best investigative tool available to study autophagy in vivo remains TEM. Ultrastructurally, autophagy is subdivided into macrophagy and microphagy. Macrophagy evolves from the formation of double-membraned autophagic vacuoles called autophagosomes, which contain cytoplasmic components such as swollen mitochondria and endoplasmic reticulum. Fusion of lysosomes with autophagosomes forms autophagolysosomes, where digestion of the cytoplasmic detritus occurs. Microphagy, on the other hand, refers to the direct lysosomal delivery of cytoplasmic proteins through chaperone-mediated autophagy or lysosomal membrane invagination. At the end, the digests appear as electron-dense structures called residual bodies [37,38,40,41]. The term mitophagy was introduced because mitochondria can occupy 20% of the cytoplasmic hepatocellular volume and are often selectively targeted for autophagy [42]. Mitochondrial replication is asynchronous with cell division, and in terminally differentiated cells, mitochondria have a half-life of days to weeks, depending on metabolic poise and activity. Accordingly, mitophagy is required for normal mitochondrial turnover, which serves to clear the cells of dysfunctional mitochondria. One model of aging proposes that compromised mitophagy leads to a gradual accumulation of mitochondria harboring mutations in mtDNA, which contributes to cellular senescence [42]. 3.3. Calcium Densities As discussed above, the bioenergetic failure of ion-dependent membrane pumps leads to increased cytosolic Ca2+ , which accumulates rapidly in the mitochondrial

MITOCHONDRIOPATHIES

521

matrix because of the Ca2+ uniporter and . Physiological Ca2+ accumulation normally occurs when ionotrophic receptors, including neuronal NMDA, kainate, and ibotenate receptors, are activated by excitatory dicarboxylic acids such as glutamate. However, excessive glutamate excitation translates into pathogenic Ca2+ concentrations that, above a threshold, induce mitochondrial failure and excitotoxic cell death [43]. During the terminal degenerative stages of excitotoxicity, mitochondria accumulate calcium as dense granular bodies in the matrix which are precipitated as calcium phosphates, as discussed above. In very late necrosis, calcium may also form salts that precipitate as spicules around degenerating mitochondria or even other cytoplasmic organelles.

3.4. Glycogen and Fat Deposition Intracytoplasmic deposition of glycogen and fats is also suggestive of mitochondrial dysfunction and illustrates that ultrastructural mitochondrial changes should be interpreted in conjunction with other cellular changes and not independently. For example, chronic progressive external ophthalmoplegia (CPEO) due to mtDNA mutations is one manifestation of several mitochondrial encephalomyopathies (see Chapter 11). In some CPEO patients there is signiﬁcant reduction in OXPHOS activity in the heart, especially complexes III and IV, which have components encoded by the mitochondrial genome. This results in mitochondrial proliferation or abnormalities, that can be assessed microscopically and by the levels of citrate synthase activity. In such cardiomyocytes, the OXPHOS defects also impair glucose and fat metabolism since both the citric acid cycle and β-oxidation of fatty acids require NAD and FAD to progress. When the electron transport system is impaired, the reduced ﬂavoprotins accumulate and correspondingly, constrict Kreb’s cycle and β-oxidation. As a result, glycogen and fats can accumulate, and this is apparent microscopically as cytoplasmic vacuolation due to glycogen (PAS positive, diastase sensitive) or fat (oil red O stain positive) accumulation (see the discussion of microvescicular steatosis in Section 3.6). Under such circumstances, glycogen deposition suggests that the heart continues to use fatty acids as a main fuel source, with relative preservation of function, whereas fat accumulation suggests failure of β-oxidation and loss of the preferred myocardial fuel [44,45]. Glycogen may also accumulate within mitochondria. Intramitochondrial glycogen deposition was observed by Buja et al. in dogs after anoxic cardiac arrest [46]. The intramitochondrial glycogen deposits occurred in monoparticulate (β) form and were located within dilated cristae and therefore lined by single membranes (CM); mitochondrial membranes were intact. The postulated sequence of events leading to intermembranar glycogen deposition included (1) solubilization of the enzymes of glycogen metabolism during the anoxic period, (2) increased permeability due to anoxic damage of the outer (macromolecular barrier) mitochondrial membrane, and (3) diffusion of the solubilized enzymes of glycogen synthesis into the mitochondria, with subsequent formation of glycogen [46,47].

522

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

3.5. Mitochondrial Response in Metabolic Diseases Mitochondrial swelling is frequently seen after pathogenic stressors such as hypoxia and excitotoxicity. In hypoosmotic media, mitochondrial can swell to two to three times normal size before losing membrane integrity [48]. Megamitochondria (MGs) are organelles that exceed this osmotic threshold and are characterized by a markedly enlarged matrix, reduced number and length of cristae, and impaired function. A distinction must be made between a normally functioning extended mitochondrial reticulum, composed via extensive mitochondrial fusion, which is the norm for some cell types, versus MGs, where function is impaired. MGs are found in a wide variety of tissues and in many diseases and syndromes, including metabolic alcohol hepatitis, diabetes, vitamin E deﬁciencies, and muscular dystrophies (Figure 8). MGs are also induced by toxicicants and drugs such as thiazolidinediones (see Section 3.6), ethanol, hydrazine, chloramphenicol, erythromycin, and H2 O2 [48,49]. For example, exposure to ethanol and other alkyl alcohols causes an increase in the number of megamitochondria in hepatocytes concomitant with a reduction in number and size of cristae in normal mitochondria. Studies with the ammonia derivatives cuprizone and hydrazine have highlighted the critical role of electron-releasing substituting groups in the formation of megamitochondria: cells exposed to these compounds may be forced to consume more oxygen to deal with extra electrons released from the substituting groups [48,49]. Consequently, relative cellular hypoxia is accompanied by the formation of megamitochondria, as seen in the hypoxic heart. Megamitochondria occasionally contain paracrystalline inclusions, as is the case in nonalcoholic steatohepatitis [49]. Paracrystalline inclusions are also

Figure 8 Megamitochondria in an hepatocyte of a patient with alcoholic hepatic disease (A) and in skeletal muscle of a patient with a mitochondtrial myopathy (B). (From Wakabayashi [48], with permission.)

MITOCHONDRIOPATHIES

523

described in Escherichia coli under conditions of oxidative stress. In E. coli , the crystals are formed by copolymerization of ferritinlike molecules and bacterial DNA under conditions where the DNA would be damaged. Because of the evolutionary origin of mitochondria from symbiotic prokaryotes and the important role of iron metabolites in oxidative stress, one could speculate that the paracrystalline structures in the megamitochondria of nonalcoholic steatohepatitis have a similar function. The evidence indicates that free radicals are proximate mediators of MGs formation, and treatment with free-radical scavengers such as α-tocopherol, coenzyme Q10 , and 4-OH-TEMPO prevent, or reverse, MGs formation [48]. The current model proposes that MGs form in response to an oxidative stress imposed either exogenously or endogenously via autoxidation of OXPHOS components. Repressing respiration via MGs formation concomitantly decreases ROS generation. In accord with this model, if intracellular ROS levels return to normal, MGs shrink and resume normal ATP production. Conversely, if elevated oxidative stress continues, MGs swell even more, declines, and cytochrome c is released, leading to apoptosis. It is interesting to speculate that simple surface area (SA)/volume considerations foster enlargement as a mechanism to moderate function; a grossly enlarged organelle might become substrate limited, slowing respiration, when availability is dictated by SA constraints. In any event, it appears that megamitochondria formation may be an adaptative response to stressors at the level of intracellular orgranelles [48–50]. 3.6. Drug-Induced Mitochondriopathies Xenobiotics that interact with mitochondria can be divided into several classes, such as (1) drugs speciﬁcally designed to affect mitochondrial functions, such as those acting on mitochondrial β-oxidation for the treatment of lipid disorders; (2) drugs for which mitochondrial effects result from secondary “off-target” interactions, often resulting in deleterious side effects, such as some antivirals and PPAR agonists; and (3) toxins that directly undermine mitochondrial function by inhibiting OXPHOS, such as rotenone, antimycin A, and oligomycin. For compounds interacting with mitochondria as a secondary target, a review of the toxicologic studies during which the drug is used at concentrations exceeding therapeutic levels is required to understand the mechanism of potential side effects of chronic administration. The length of these studies is also very important since some drugs can accumulate in particular tissues or organs and attain concentrations higher than those calculated for the whole body. Organs that are very active metabolically, such as cardiac and skeletal muscles, liver, brain, endocrine glands, and kidney, among others, contain the largest number of mitochondria and are therefore more susceptible to drug-induced mitochondriopathies. Additionally, it bears repeating that mitochondria maintain a negatively charged interior responsible for the ψm conditions, which promote accumulation of some compounds, such as weak acids, in their anionic form (see Chapters 17 and 25). Mitochondrial transporters located in the OM, such as the carboxylic acid transporters, can also promote speciﬁc uptake of drugs.

524

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

Anticancer Drugs A large number of anticancer drugs exert their therapeutic action by inducing apoptosis in rapidly dividing malignant cells, including etoposide, doxorubicin, lonidamide, betulinic acid, arsenite, and amphiphilic cationic α-helical peptides, among others. These drugs exert their proapotptotic action partially by inducing mitochondrial transition pore formation (e.g., lonidamide) and through the generation of ROS (e.g., adriamycin) (see Chapter 6). As described in Chapter 25, drugs can be designed to contain cationic charge to facilitate mitochondrial uptake into the matrix and/or to bind to the IM but not the cytoplasmic membrane, based on potential and phospholipid composition, which induces mitochondrial swelling and apoptosis [51]. Alcohol In many species, alcohol administration is followed by the development of fatty liver (macrovesicular steatosis) and secondary inﬂammation with the elevation of serum liver enzymes, which reﬂects hepatocyte death. One of the earliest and most consistent microscopic alterations in the liver of the chronic alcohol consumer is a change in the structure and function of the mitochondria. Mitochondria are enlarged, often misshapen, and demonstrate disrupted and aggregated cristae [7]. These ultrastructural changes are accompanied by a decreased capacity for energy conservation and repressed NADH-linked respiration. The changes were also associated with signiﬁcant oxidative stress [52] (see Chapter 5). Drug-Induced Microvesicular Steatosis In the liver, macrovesicular and drug-induced microvesicular steatosis have different etiologies and clinical outcomes. Macrovesicular steatosis is the more common form and is characterized by the presence of a few, large triglyceride-containing vacuoles which displace the hepatocellular nucleus to the cell periphery. Ultrastructurally, these vacuoles lack an apparent membrane, internal differentiation, and are electron-lucent. In the absence of other liver lesions, macrovesicular steatosis is recognized as a relatively benign condition reﬂecting increased mobilization of fatty acids from adipose tissue, increased hepatic synthesis of fatty acids, and decreased esteriﬁcation of fatty acids into triglycerides or decreased egress of triglycerides from the liver [7]. In contrast, in microvesicular steatosis, mitochondrial impairment leads to the accumulation of numerous small lipid vacuoles in the cytoplasm around the centrally localized nuclei, enlarging the hepatocyte (Figures 9 and 10). Mirovesicular steatosis can be lobular, or limited to centrilobular, midzonal, or periportal areas. Impaired mitochondrial β-oxidation can be secondary to a paucity of cofactors (CoA, l-carnitine), impaired transporter activity, or inhibition of OXPHOS. Resulting adverse clinical signs appear because mitochondrial β-oxidation is the source of ATP in heart and in the liver during fasting, and high concentrations of nonesteriﬁed fatty acids undermine gluconeogenesis, ureagenesis, tricarboxylic acid cycle, and OXPHOS by uncoupling electron transport from phosphorylation (directly and indirectly through the metabolization of nonesteriﬁed fatty acids into dicarboxylic acids). Nonesteriﬁed fatty acids also have detergent properties,

525

MITOCHONDRIOPATHIES

A

B

Figure 9 Oil red O stain for lipid in normal rat liver (A), and showing the lipid accumulation characteristic of microvescular steatosis (B). In this case the steatosis was induced by a drug in development that potently inhibited OXPHOS complex V at submicrometer concentrations. (From Pruimboom et al. [53].) (See insert for color representation of ﬁgure.)

A

B

Figure 10 Ultrastructural appearance of hepatocytes from the preclinical safety study shown above. Note the accumulation of lipid vacuoles < 1 nM in diameter that is characteristic of microvesicular steatosis in the treated dog (B) compared to control dog (A). (From Pruimboom et al. [53].)

which can cause the disorganization of membrane structures (phospholipids bilayers). For these reasons, a ﬁnding of microvesicular steatosis in the liver during preclinical toxicology studies should be regarded as a potential indication of mitochondrial dysfunction, and its human relevance should be discussed. The preclinical compound shown here was not developed further.

526

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

HAART The beneﬁts of antiretroviral combined therapy in human immunodeﬁciency virus-infected patients in terms of mortality and morbidity have been established clearly. However, liver mitochondriopathies have been described with nucleoside reverse transcriptase inhibitors secondary to defective mtDNA replication, due to the NRTI inhibition of mitochondrial gamma polymerase (see Chapters 2, 9, 21, and 22). Van Huyen et al. [70] have evaluated the expression of cytochrome oxidase subunits I and IV (encoded by mitochondrial and nuclear DNA, respectively) by immunohistochemistry in NRTI hepatotoxicity and the resulting ultrastructural alterations. They discovered that decreased COX subunit 1 expression is associated with severe ultrastructural mitochondrial alterations and may represent overt NRTI-induced mitochondrial cytopathy. In the liver, the resulting mitochondrial dysfunction causes defects in fatty acid β-oxidation leading to lactic acidosis and macrovesicular or microvesicular steatosis. Microscopically, NRTI-induced mitochondrial impairment in the liver is characterized by macrovesicular, and occasionally microvesicular, steatosis with variable grades of ﬁbrosis and inﬂammation. Other features include foci of ballooning degeneration, Mallory bodies, and globular eosinophilic structures consistent with megamitochondria in TEM. Ultrastructural mitochondrial alterations included effacement to complete loss of mitochondrial cristae, paracrystalline inclusions in the mitochondria, and variation in mitochondrial size and shape (including megamitochondria), leading to the formation of autophagocytic vacuoles containing remnants of degenerative mitochondria [54] (Figure 11) (see Chapter 9). By immunohistochemistry, decreased COX subunit I labeling was observed in nine mono-infected and ﬁve co-infected treated patients but not in untreated patients. Subunit IV labeling was unaffected. Furthermore, reduction in treated patients was associated with an increased frequency of metabolic and microscopic disturbances [54]. Statins In humans, some lipid-lowering drugs have been associated with rare cases of skeletal muscle pain, myoﬁber degeneration, and rhabdomyolysis. Inhibitors of 3-hydroxy-3-methylglutaryl–coenzyme A (HMG-CoA) reductase, such as the statins, uncouple mitochondrial OXPHOS. For example, pravastatin treatment is associated with a decline in mitochondrial complex I and IV activities in the diaphragm and psoai major (but not heart and liver mitochondrial respiratory function) in 35-to 55-week-old rats and 55-week-old rats, respectively [55]. HMG CoA reductase inhibitor accelerates the aging effect on diaphragm mitochondrial respiratory function in rats. According to Westwood et al. [28], fast-twitch myoﬁbers are sensitive to this effect because expression of the monocarboxylate transporter isoform 4 in these cells serves to accumulate statins. This is generally a rare adverse event, but its occurrence increases when patients are co-treated with other hypolipodemic drugs, such as gemﬁbrozil [56], which also undermines mitochondrial function directly by uncoupling electron ﬂux from phosphorylation [57] (see Chapter 7). As might be anticipated by the effects on respiration, statins cause mitochondrial swelling and loss of cristae, among other ultrastructural ﬁndings (Figures 12

527

MITOCHONDRIOPATHIES

A

B

C

D

Figure 11 Disruption of normal mitochondrial ultrastructure after a 4-day exposure to various dideoxynucleosides: 2 ,3 -dideoxycytidine at 10 µM (A); 2 ,3 -didehydro-3-deoxythymidine at 54 µM (B); and 2 ,3 - dideoxyinosine at 750 µM (C, D). Ultrastructural changes included swollen or hydropic mitochondria with an electron-lucent matrix, and loss and distortion of cristae, which sometimes formed concentrically arranged rings. Similar ultrastructural changes are also described in cell cultures in the presence of either ethidium bromide or chloramphenicol, which are known to be inhibitors of mtDNA synthesis and protein synthesis, respectively. Magniﬁcation, 25,000 for A,B,C, and 8000 for D. (From Medina et al. [54], with permission.)

528

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

and 13). For example, rabbits dosed orally with simvastatin (50 mg/kg daily for 4 weeks) showed signiﬁcantly increased myotonic discharges skeletal muscle degeneration, and resulting elevations in serum creatine kinase levels [58]. Ultrastructurally, there is mitochondrial swelling, loss of cristae, and autophagy (Figures 12 and 13). Indeed, mitochondrial disruption precedes frank histological toxicity. For example, Westwood et al. [28] showed that in rats dosed with simvastatin, even in the absence of obvious evidence of necrosis by histology, there was an increased incidence of mitochondrial vacuolation and degeneration, leading to the formation of subsarcolemal mitochondrial myelinoid bodies. PPARγ Agonists Thiazolidinediones are a class of PPARγ agonists used effectively for the treatment of type 2 diabetes. The development of some of these PPARγ agonists has been impeded by their association with adverse preclinical or clinical events, not always related to mitothondiral toxicity. Troglitazone, the ﬁrst thiazolidinedione marketed, was withdrawn because of idiosyncratic liver reaction, the mechanism of which is currently speculative. However, mitochondrial OXPHOS dysfunction may play a role in this toxicity, as troglitazone is a strong inhibitor of complex V [57] and as ultrastructurally, troglitazone induces MGs formation of MGs in the human hepatocyte cell line (OUMS-29 cells treated with <50 µM for 5 days) and intramitochondrial myelinlike deposits (Figure 14) [60]. Secondary H2 O2 production and loss of is also observed, but interestingly, widespread cell death is not apparent [60]. In contrast, the marketed drug pioglitazone induced much less mitochondrial OXPHOS or ultrastructural disturbances in vitro (Figure 14) [61]. The development of ciglitazone was stopped preclinically because of its association with a dose-dependent increase in the incidence and severity of nuclear cataract formation in Sprague–Dawley rats. Lenticular opacities are generally caused by compounds that penetrate the lens directly and disrupt the cellular processes required to maintain lens transparency or interact indirectly with the lens surface through the generation of reactive oxygen species ([62]. In the case of ciglitazone, direct interaction with the lens is indicated by the fact that the compound distributes to the rat eye after oral administration [63]. Subsequently, Aleo et al. [64] demonstrated that ciglitazone causes concentration- and time-dependent declines in lens ATP content, mitochondrial reduction of tetrazolium dye–reduced glutathione content, and morphometric lens wet weight and clarity, all indices of cytotoxicity. Pretreatment of lenses with ruthenium red, an inhibitor of the mitochondrial calcium uniporter, protected the lenses from ciglitazone-mediated toxicity, further implicating mitochondrial dysfunction as the etiology of ciglitazone-mediated cataract formation [64]. In the liver of nonalcoholic steatohepatitis (NASH) patients, approximately 5% of mitochondria in hepatocytes harbor linear crystalline inclusions similar to those found in alcohol-related liver disease and Wilson’s disease (Figure 15) [65]. In NASH, the presence of these structures correlates with measures of oxidative stress, and optical diffraction studies indicate a predominantly lipid composition

529

MITOCHONDRIOPATHIES A

B

C

D

E

F

G

H

Figure 12 Ultrastructural changes induced in rat EDL muscle by cerivastatin are exacerbated by exercise [59]. At a dose of 0.5 mg/kg (rat, 14-day duration, once a day via gavage), mitochondrial swelling and cristaelysis are evident, and are exacerbated by daily treadmill exercise (D) (arrows in C and D). Note that not all mitochondrial are affected equally, as normal-appearing organelles occur within the same ﬁeld. Mitochondrial damage is worsened by oral dose of 1 mg/kg (E, F), as illustrated by the presence of concentric membranous whorls around or within phagocytic vacuoles containing cytoplasmic components such as ribosomes and degenerated mitochondria. Also note inﬁltration of inﬂammatory cells (arrow in G) and regional loss of sarcomere and myoﬁbril structure (arrow in H). Unless speciﬁed, the scale bar 0.5 µm. (From Seachrist et al. [59], with permission.)

530

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

Figure 13 Ultrastructurally, swollen mitochondria and cristaelysis (arrowhead) can be seen in rabbit skeletal muscle after 4 weeks of simvaststin exposure (50 mg/kg via gavage). (mag. 8300x). (From Nakahara et al. [58], with permission.)

[65]. When NASH patients were treated for 10 months with rosiglitazone, which also slightly uncouples OXPHOS (albeit less potently that troglitazone and others [57]), the number of crystalline inclusions increased signiﬁcantly [61]. 3.7. Morphology of Mitochondria in Cancer Although the number of mitochondria varies among individual neoplasms, malignant cells generally have diminished oxidative phosphorylation (but see [66] for a counterpoint) and therefore, mitochondria with short stubby cristae. Proteins, particularly the large cytochromes, may disintegrate and repolymerize to form intramitochondrial lattices. Oncocytomas (or oxyphilic tumors) are characterized by the abnormal proliferation of mitochondria (Figure 16). These tumors affect a wide variety of human tissues, including the thyroid, salivary, and parotid glands, liver, kidneys, and lungs. The most frequent site is the thyroid, where oncocytomas are often referred to as Hurthle cell tumors. In such tumors about 75% of oncocytes are characterized by a ﬁne eosinophilic granular cytoplasm due to the presence of a large number of cytoplasmic mitochondria. It has been suggested that this type of mitochondrial proliferation represents a compensatory mechanism against mitochondrial complex 1 or ATP synthesis dysfunction. Consistent with this hypothesis is the fact that oncocytes have increased histochemical reactivity of the respiratory chain complexes II and IV while the overall respiration and phosphorylation capacity are diminished. The ﬁrst phases of drug development are typically conducted in transformed cell lines where mitochondrial structure and function are not necessarily normal. For example, many cell lines are highly reliant on glycolysis and not OXPHOS

531

MITOCHONDRIOPATHIES

A

B

C

D

Figure 14 Mitochondrial ultrastructure in human immortalized hepatocytes exposed to PPARγ ligands. Compared to control cells (A), troglitazone (50 µM for 5 days) (B, C) yielded swollen mitochondria with loss of cristae and electron-dense myelinlike bodies, presumably from degraded mitochondrial membrane. Under these same condtions, pioglitazone (D) induced much less mitochondrial swelling and loss of cristae, in accord with the clinical disposition of these two drugs. (From Caldwell et al. [61], with permission from Wiley.)

for ATP, and as such are resistant to known mitochondrial toxicants such as rotenone, antimycin, and cyanide [67]. 3.8. Mitochondrial DNA Mutations and Diseases Mitochondrial disorders caused by mtDNA mutations range from relatively mild late-onset conditions, such as progressive external ophtalmoplegia (discussed earlier) or sensorineural hearing loss, to devastating syndromes such as MELAS (mitochondrial encephalomyopathy with lactic acidosis and strokelike seizures) and MERRF (myoclonus epilepsy and ragged-red ﬁbers). Although mitochondrial mutations, or mutations in proteins encoded by nuclear DNA and then

532

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

Figure 15 Linear mitochondrial crystal inclusions are shown in nearly all mitochondria present (black arrows), along with fat droplet and foci of the dilated endoplasmic reticulum (asterisk; 20,000×). The inset shows a detail of the crystals that displaced the normal cristae (60,000×). From Caldwell et al. [61], with permission from Wiley.)

Figure 16 Renal oncocytoma ultrastructure showing numerous mitochondria ﬁlling the cytoplasm of the tumor cell. The nucleus is on the left edge. (From Vara Castrodeza et al. [68], with permission.)

MITOCHONDRIOPATHIES

533

imported into mitochondria, are distributed systemically, frank pathology arises when a threshold of mutated mtDNA and consequent mitochondrial dysfunction is exceeded and bioenergetics is undermined. Pathology typically arises ﬁrst in energy-demanding tissues such as central nervous system, myocardium, skeletal muscle, and the β-cells of the pancreas (causing diabetes) (see Chapters 4 and 11). Microscopy plays a particularly important role in the diagnosis of these mitochondrial pathologies. Brantov´a et al. [69] showed that the TEM of ﬁbroblasts can provide a differential diagnosis of mitochondrial disorders caused by mtDNA mutations, even in the absence of alterations in mitochondrial function or amounts of OXPHOS proteins. In these cases, TEM reveals a mixture of mainly abnormal, partially swollen mitochondria, heterogeneous in size and shape, containing few or no cristae. As discussed above, immunohistochemistry of porin and the OXPHOS complexes in ﬁbroblasts can also provide insight into mechanism(s) of mitochondrial impairment in patients with mitochondrial diseases [20]. Patients affected by the MERRF syndrome present a mixture of neurological and myopathic symptoms, including myoclonic seizures, ataxia, muscle weakness, and lactic acidosis due to a compensatory shift toward glycolysis. In skeletal muscle biopsies, ragged-red ﬁbers are prominent features on Gomori’s modiﬁed trichrome-stained sections (Figure 17) and reﬂect the massive subsarcolemmal mitochondrial proliferation secondary to mitochondrial dysfunction in muscle. Because comparable mitochondrial accumulation can also be found in type 1 ﬁbers of children and occasionally in unrelated mitochondrial diseases, ragged-red ﬁbers are pathognomonic of MERF when large and present in at least 25% of

Figure 17 Histological appearance of ragged-red ﬁber in patient with a mitochondrial myopathy. This image is from frozen muscle sections stained with Gomori trichrome stain. The red color of these ﬁbers is due to large numbers of subsarcolemmal mitochondria that have proliferated to compensate for repressed OXPHOS. The abnormal ﬁbers appear coarse and disorganized. [From Neuropathology Web site (www.neuropathologyweb.org), by D. P. Agamanolis, Akron Children’s Hospital, Northeastern Ohio Universities College of Medicine, with permission.] (See insert for color representation of ﬁgure.)

534

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

Figure 18 Ultrastructural appearance of mitochondria in a ragged-red ﬁber from a patient with a mitochondrial myopathy. Note the widespread loss or disorganization of cristae, as well as electron-dense precipitates in the matrix of many. Myoﬁlaments appear in transverse section. [From Neuropathology Website (www.neuropathologyweb.org), by D. P. Agamanolis, Akron Children’s Hospital, Northeastern Ohio Universities College of Medicine, with permission.]

the ﬁbers. When ragged red myoﬁbers are identiﬁed, it is recommended that the composition of these areas be conﬁrmed by TEM, where one can expect subsarcolemmal accumulation of misshapen mitochondria, concentrically organized cristae or no cristae at all, and paracrystalline transformation (Figure 18). Supporting evidence is that of concomitant increase in lipid or glycogen within the myoﬁbers. Histochemistry is also often used to assess myoﬁber C-IV, NADH, and SDH status. 4. CONCLUSIONS Mitochondrial function depends on organellar structural integrity, particularly of the inner membrane. Agents or conditions that disrupt this integrity undermine function, and if sufﬁciently severe, will induce cell death via necrosis or apoptosis. Accordingly, assessment of mitochondrial structure is a viable surrogate index of function, and such assessments can entail immunohistochemistry and a host of other microscopy techniques. Importantly, drug-induced mitochondrial dysfunction can also be detected, so that histopathology will continue to play an important role in making drugs of the future safer. REFERENCES 1. Perkins GA Electron tomography of large, multicomponent biological structures. J Struct Biol. 1997;120:219–227.

REFERENCES

535

2. Perkins GA. Electron tomography of mitochondria from brown adipocytes reveals cristae junctions. J Bioenerg Biomembr. 1998;30:431–442. 3. Vogel F, Bornh¨ovd C, Neupert W, Reichert AS. Dynamic subcompartmentalization of the mitochondrial inner membrane. J Cell Biol. 2006;175:237–247. 4. Glauert AM, ed. Practical Methods in Electron Microscopy. Amsterdam: Elsevier; 1984. 5. Luft JH. Permanganate; a new ﬁxative for electron microscopy. J Biophys Biochem Cytol. 1956;2:799–802. 6. Hayat MA. Fixation for Electron Microscopy, vol 1. New York: Van Nostrand Reinhold, 1970. 7. Normal F. Cheville NF. Ultrastructural Pathology: An Introduction to Interpretation. Ames, IA: Iowa State University Press; 1994. 8. Mannella CA. Topology of the mitochondrial inner membrane: dynamics and bioenergetic implications. IUBMB Life. 2001;52:93–100. 9. Riva A, Tandler B, Loffredo F, Vazquez E, Hoppel C. Structural differences in two biochemically deﬁned populations of cardiac mitochondria. Am J Physiol Heart Circ Physiol. 2005: 289:H868–H872. 10. Palmer JW, Tandler B, Hoppel CL. Biochemical properties of subsarcolemmal and interﬁbrillar mitochondria isolated from rat cardiac muscle. J Biol Chem. 1977;252:8731–8739. 11. Riva A, Tandler B, Lesnefsky EJ, et al. Structure of cristae in cardiac mitochondria of aged rat. Mech Ageing Dev. 2006;127:917–921. 12. Takekura H, Yoshioka T. Speciﬁc mitochondrial responses to running training are induced in each type of rat single muscle ﬁbers. Jpn J Physiol. 1989;39:497–509. 13. Muller W. Subsarcolemmal mitochondria and capilarization of soleus muscle ﬁbers in young rats subjected to an endurance training: a morphometric study of semithin sections. Cell Tissue Res. 1976;174:367–389. 14. Chan DC. Mitochondrial fusion and ﬁssion in mammals. Annu Rev Cell Dev Biol. 2006;22:79–99. 15. Chen H, Chan DC. Emerging functions of mammalian mitochondrial fusion and ﬁssion. Hum Mol Genet. 2005;14 (spec 2):R283–R289. 16. Pearse AGE. Histochemistry: Theoretical and Applied , vol 2. Baltimore: Williams & Wilkins; 1972. 17. Dubowitz V. Muscle Biopsy: A Practical Approach. London: Bailliere Tindall; 1985. 18. Nachlas MM, Tsou KC, DeSouza E., Cheng CS, Seligman AM. Cytochemical demonstration of succinic dehydrogenase by the use of a new p-nitrophenyl substituted ditetrazole. J Histochem Cytochem. 1957;5:420–436. 19. Farber R, Sternberg WH, Dunlop CE, Histochemical localization of speciﬁc oxidative enzymes: I. Tetrazolium stains for diphosphopyridine nucleotide diaphorase and triphosphopyridine nuceleotide diaphorase. J Histochem Cytochem. 1956;4:254–265. 20. Capaldi RA, Murray J, Byrne L, Janes MS, Marusich MF. Immunological approaches to the characterization and diagnosis of mitochondrial disease. Mitochondrion. 2004;4:417–426. 21. Parone PA, Martinou JC. Mitochondrial ﬁssion and apoptosis: an ongoing trial. Biochem Biophys Acta. 2006;1763:522–530.

536

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

22. Cereghetti GM, Scorrano L. The many shapes of mitochondrial death. Oncogene. 2006;25:4717–4724. 23. Santel A. Get the balance right: mitofusins roles in health and disease. Biochem Biophys Acta. 2006;1763:490–499. 24. Gosslau A, Chen M, Ho Ci-, Chen KY. A methoxy derivative of resveratrol analogue selectively induced activation of the mitochondrial apoptotic pathway in transformed ﬁbroblasts. Br J Cancer. 2005;92:512–521. 25. Muller FL, Liu Y, Van Remmen H. Complex III releases superoxide to both sides of thee inner mitochondrial membrane. J Biol Chem. 2004;279:49064–49073. 26. Dykens JA. Isolated cerebral and cerebellar mitochondria produce free radicals when exposed to elevated CA2+ and Na+ : implications for neurodegeneration. J Neurochem. 1994;63:584–591. 27. Anderson EJ, Neufer PD. Type II skeletal myoﬁbers possess unique properties that potentiate mitochondrial H2 O2 generation. Am J Physiol Cell Physiol. 2006;290:C844–C851. 28. Westwood F R, Bigley A, Randall K, Marsden A M, Scott R C. Statin-induced muscle necrosis in the rat: distribution, development, and ﬁbre selectivity. Toxicol Pathol. 2005;33:246–257. 29. Copin JC, Gasche Y, Chan PH. Overexpression of copper/zinc superoxide dismutase does not prevent neonatal lethality in mutant mice that lack manganese superoxide dismutase. Free Radic Biol Med. 2000;28:1571–1576. 30. Li Y, Wo JM, Su RR, Ray MB, Martin RC. Loss of manganese superoxide dismutase expression and activity in rat esophagus with external esophagal perfusion. Surgery. 2007;141:359–367. 31. Kinnula VL, Torkkeli T, Kristo P, et al. Ultrastructural and chromosomal studies on manganese superoxide dismutase in malignant mesothelioma. Am J Respir Cell Mol Biol. 2004;31:147–153. 32. Qi X, Lewin AS, Sun L, Hauswirth WW, Guy J. Suppression of mitochondrial oxidative stress provides long-term neuroprotection in experimental optic neuritis. Invest Ophthalmol Vis Sci. 2007;48:681–691. 33. Pearce LL, Bominaar EL, Hill BC, Peterson J. Reversal of cyanide inhibition of cytochrome c oxidase by the auxiliary substrate nitric oxide: an endogenous antidote to cyanide poisoning? J Biol Chem. 2003;278:52139–52145. 34. Garcia-Ruiz, I. Uric acid and anti-TNF antibody improve mitochondrial dysfunction in ob/ob mice. Hepatology. 2006;44:581–591. 35. Telek G, Regoly-Merei J, Kovacs GC, et al. The ﬁrst histological demonstration of pancreatic oxidative stress in human acute pancreatitis. Hepatogastroenterology. 2001;48:1252–1258. 36. Hybertson BM, Lee YM, Cho HG, Cho OJ, Repine JE. Alveolar type II cell abnormalities and peroxide formation in lungs of rats given IL-1 intratracheally. Inﬂammation. 2000;24:289–303. 37. Zhu JH, Horbinski C, Guo F, Watkins S, Uchiyama Y, Chu CT. Regulation of autophagy by extracellular signal-regulated protein kinases during 1-methyl-4-phenylpyridinium-induced cell death. Am J Pathol. 2007;170:75–86. 38. Rodriguez-Enriquez S. Tracker dyes to probe mitochondrial autophagy (mito phagy) in rat hepatocytes. Autophagy. 2006;2:39–46.

REFERENCES

537

39. Munaf´o DB, Colombo MI. A novel assay to study autophagy: regulation of autophagosome vacuole size by amino acid deprivation. J Cell Sci. 2001;114:3619–3629. 40. Herrera F, Martin V, Carrera P, et al. Tryptamine induces cell death with ultrastructural features of autophagy in neurons and glia: possible relevance for neurodegenerative disorders. Anat Rec A Discov Mol Cell Evol Biol. 2006;288:1026–1030. 41. Bampton ET, Goemans CG, Niranjan D, Mizushima N, Tolkovsky AM. The dynamics of autophagy visualized in live cells: from autophagosome formation to fusion with endo/lysosomes. Autophagy. 2005;1:23–36. 42. Lemasters JJ. Selective mitochondrial autophagy, or mitophagy, as a targeted defense against oxidative stress, mitochondrial dysfunction, and aging. Rejuvenation Res. 2005;8:3–5. 43. Dykens JA. Mitochondrial free radical production and the etiology of neurodegenerative disease. In Neurodegenerative Diseases: Mitochondria and Free Radicals in Pathogenesis MF Beal, I Bodis-Wollner, N Howell, eds. New York: Wiley; 1997:29–55. 44. Venugopal B, Wong KT, Goto YI, Bhattacharjee MB. Mitochondrial disorder, diabetes mellitus, and ﬁndings in three muscles, including the heart. Ultrastruct Pathol. 2006;30:135–141. 45. Bhattacharjee M, Venugopal B, Wong KT, Goto YI, Bhattacharjee MB. Mitochondrial disorder, diabetes mellitus, and ﬁndings in three muscles, including the heart. Ultrastruct Pathol. 2006;30:481–487. 46. Buja LM, Ferrans VJ, Levitsky S. Occurrence of intramitochondrial glycogen in canine myocardium after prolonged anoxic cardiac arrest. J Mol Cell Cardiol. 1972;4:237–254. 47. Alaraj M, Gajkowska B, Cholewinski M, Lazarewicz JW. Hyperglycaemia and intramitochondrial glycogen granules in the brain of mice: ultrastructural study. Folia Neuropathol. 2004;42:113–118. 48. Wakabayashi T. Megamitochondria formation: physiology and pathology. J Cell Mol Med. 2002;6:497–538. 49. Le TH, Caldwell SH, Redick JA, et al. The zonal distribution of megamitochondria with crystalline inclusions in nonalcoholic steatohepatitis. Hepatology. 2004;39:1423–1429. 50. Albring M, Radsak K, Thoenes W. Giant mitochondria. II. Induction of matrix enriched megamitochondria in mouse liver parenchymal cells by ethidium bromide. Virchows Arch B Cell Pathol. 1973;14:373–377. 51. Szewczyk A, Wojtczak L. Mitochondria as a pharmacological target. Pharmacol Rev. 2002;54:101–127. 52. Fromenty B, Pessayre D. Impaired mitochondrial function in microvesicular steatosis: effects of drugs, ethanol, hormones and cytokines. J Hepatol. 1997;26:43–53. 53. Pruimboom I, Brees D, Will Y. A case of hepatic microvesicular steatosis in the liver of dogs during in vivo toxicology studies. Presented at Classic Examples in Toxicologic Pathology XIV, Hannover, Germany, 2007. 54. Medina DJ, Tsai CH, Hsiung GD, Cheng YC. Comparison of mitochondrial morphology, mitochondrial DNA content, and cell viability in cultured cells treated with three anti-human immunodeﬁciency virus dideoxynucleosides. Antimicrob Agents Chemother. 1994;38:1824–1828.

538

ASSESSMENT OF MITOCHONDRIAL DYSFUNCTION BY MICROSCOPY

55. Sugiyama S. HMG CoA reductase inhibitor accelerates aging effect on diaphragm mitochondrial respiratory function in rats. Biochem Mol Biol Int. 1998;46:923–931. 56. East C, Alivizatos PA, Grundy SM, Jones PH, Farmer JA. Rhabdomyolysis in patients receiving lovastatin after cardiac transplantation. N Engl J Med. 1988;318:47–48. 57. Nadanaciva S, Dykens JA, Bernal A, Capaldi RA, Will Y. Mitochondrial impairment by PPAR agonists and statins identiﬁed via immunocaptured OXPHOS complex activities and respiration. Toxicol Appl Pharmacol. 2007;223:277–287. 58. Nakahara K, Kuriyama M, Sonoda Y, et al. Myopathy induced by HMG-CoA reductase inhibitors in rabbits: a pathological, electrophysiological, and biochemical study. Toxicol Appl Pharmacol. 1998;152:99–106. 59. Seachrist JL, Loi CM, Evans MG, Criswell KA, Rothwell CE. Roles of exercise and pharmacokinetics in cerivastatin-induced skeletal muscle toxicity. Toxicol Sci. 2005;88:551–561. 60. Shishido S, Koga H, Harada M, et al. Hydrogen peroxide overproduction in megamitochondria of troglitazone-treated human hepatocytes. Hepatology. 2003;37:136–147. 61. Caldwell SH, Patrie JT, Brunt EM, et al. The effects of 48 weeks of rosiglitazone on hepatocyte mitochondria in human nonalcoholic steatohepatitis. Hepatology. 2007;46:1101–1107. 62. Bhuyan KC, Bhuyan DK. Molecular mechanism of cataractogenesis: III. Toxic metabolites of oxygen as initiators of lipid peroxidation and cataract. Curr Eye Res. 1984;3:67–81. 63. Torii H, Yoshida K, Tsukamoto T, Tanayama S. Disposition in rats and dogs of ciglitazone, a new antidiabetic agent. Xenobiotica. 1984;14:259–268. 64. Aleo MD, Doshna CM, Navetta KA. Ciglitazone-induced lenticular opacities in rats: in vivo and whole lens explant culture evaluation. J Pharmacol Exp Ther. 2005;312:1027–1033. 65. Sternlieb I, Berger JE. Optical diffraction studies of crystalline structures in electron micrographs. J Cell Biol. 1969;43:448–455. 66. Zu XL, Guppy M. Cancer metabolism: facts, fantasy, and ﬁction. Biochem Biophys Res Commun. 2004;313:459–465. 67. Marroquin LD, Hynes J, Dykens JA, Jamieson JD, Will Y. Circumventing the Crabtree effect: replacing media glucose with galactose increases susceptibility of HepG2 cells to mitochondrial toxicants. Toxicol Sci. 2007;97:539–547. 68. Vara Castrodeza A, Ablanedo Ablanedo P, Garc´ıa Garc´ıa et al. Adrenal oncocytoma. B Arch Esp Urol. 1999;52:525–528. 69. Brantov´a O, Tesarov´a M, Hans´ıkov´a H, Elleder M, Zeman J, Sl´adkov´a J. Ultrastructural changes of mitochondria in the cultivated skin ﬁbroblasts of patients with point mutations in mitochondrial DNA. Ultrastruct. Pathol. 2006;30:239–245. 70. Van Huyen D, et al. Alteration of cytochrome oxidase subunit 1 labeling is associated with severe mitochondriopathy in NRTI-related hepatotoxicity in HIV patients. Modern Pathol. 2006;19:1277–1288. 71. Acevedo LM, Rivero JLL. New insights into skeletal muscle ﬁbre types in the dog with particular focus towards hybrid myosin phenotypes. Cell Tissue Res. 2006;323:283–303. 72. Martin TP, Vailas AC, Durivage JB, Edgerton VR, Castleman KR. Quantitative histochemical determination of muscle enzymes: biochemical veriﬁcation. J Histochem Cytochem. 1985;33:1053–1069.

23 DEVELOPMENT OF ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY Urs A. Boelsterli Department of Pharmaceutical Sciences, University of Connecticut School of Pharmacy, Storrs, Connecticut

Yie Hou Lee Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore

1. Introduction 2. Animal models of mitochondrial dysfunction 2.1. Animal models with acquired mitochondrial abnormalities 2.2. Animal models with mtDNA mutations 2.3. Animal models with nDNA mutations 3. Development of the Sod2+/− knockout mouse model for drug testing 3.1. Sod2+/− mice sensitized to hepatotoxic drugs that target mitochondria 4. Mitochondrial abnormalities: a possible link to idiosyncratic drug hepatotoxicity 5. Implications and signiﬁcance of the use of animal models of mitochondrial toxicity in drug safety assessment

539 540 542 543 543 546 548 550 551

1. INTRODUCTION Animal models of mitochondrial toxicity are increasingly important for mechanistic and investigative toxicology. This renewed appreciation is fostered by Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

539

540

ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY

the adverse effects of many important drugs that are associated with coincident “off-target” mitochondrial dysfunction. This drug-induced mitochondrial toxicity occurs mainly in the central nervous system, liver, heart, kidney, and skeletal muscle (reviewed by [1–4]). Another reason is the growing awareness that in humans, silent mitochondrial impairment may be more common than previously appreciated, and this preexisting condition could further modify the toxicity of drugs that affect mitochondrial function. To detect, understand, and predict drug-induced changes in mitochondrial function, a number of in vitro and in vivo models are currently available. Typically, mitochondria are isolated from animal tissues and exposed to drugs in vitro (see Chapters 12, 13, 15, and 17). It is considerably more difﬁcult to study mitochondrial injury in vivo, as there are only a limited number of endpoints with which one can readily attribute the toxicity to mitochondrial injury (see Chapters 21 and 22). Certain drugs clearly cause mitochondrial toxicity and organ damage in animal models. However, in other cases, normal, healthy rats or mice as used in standard safety studies are refractory to drugs that produce mitochondrial injury and organ toxicity in susceptible humans (at least within reasonable dose ranges) [5]. One reason why potential toxicity is missed is that mild mitochondrial damage is not readily detectable and does not necessarily pose a hazard, because individual cells are endowed with hundreds or thousands of mitochondria and respond to a toxic insult only if a certain critical threshold has been reached. Species-speciﬁc differences in the underlying mechanisms of drug toxicity may also result in poor predictability of mitochondrial liability. To circumvent some of these difﬁculties, new approaches have been used to develop animal models that feature an underlying preexisting mitochondrial abnormality to render these animals susceptible to drugs known to impair mitochondrial function in humans. By using recombinant gene deletion (knockout) or gene silencing (knockdown with small interfering RNA) techniques, it is possible to eliminate or silence key gene products selectively and thus simulate certain human situations where mitochondrial function is impaired by inherited and/or acquired mechanisms. We provide an overview of animal models of mitochondrial toxicity used currently, with an emphasis on the use of animals featuring underlying mitochondrial defects that are sensitized to drug-induced toxicity and are, correspondingly, more faithful predictors of the clinical disposition of drugs.

2. ANIMAL MODELS OF MITOCHONDRIAL DYSFUNCTION There are two general approaches to how animal models of mitochondrial dysfunction can be applied to toxicology and safety assessment. The ﬁrst utilizes direct xenobiotic modiﬁcations of mitochondrial function. This is commonly aimed at inducing organ-speciﬁc damage, and such animal models are generally used to study the pathogenesis or therapeutic treatment of a disease whose

ANIMAL MODELS OF MITOCHONDRIAL DYSFUNCTION

541

etiology includes mitochondrial dysfunction. For example, chemical induction of Parkinson-like syndrome can be induced in mice by the dopaminergic neuron-targeting model toxicant 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) [6,7]. MPTP is converted by mitochondrial MAO-B into 1-methyl4-phenylpyridinium (MPP+ ), which is taken up selectively by dopaminergic neurons via the dopamine transporter, particularly in the substantia nigra. When accumulated in these cells, MPP+ binds to mitochondrial respiratory complex I, thereby inducing oxidant stress and bioenergetic deﬁcits that ultimately kill the neurons. Another example is acute hepatic injury secondary to high doses of the analgesic acetaminophen [8]. Acetaminophen is converted to a reactive intermediate (N -acetyl-p-benzoquinoneimine), which arylates mitochondrial proteins and accelerates production of reactive oxygen species (ROS) such as superoxide plus reactive nitrogen species (RNS) such as peroxynitrite. In response, the jun-N -terminal kinase (JNK) pathway becomes activated, resulting in mitochondrial permeabilization and hepatocyte death [9–12]. A third example is cardiomyopathy induced by the antineoplastic agent doxorubicin. Doxorubicin, which is associated with cardiolipin in the inner mitochondrial membrane, has a quinone moiety and can divert electrons away from the electron transport chain (ETC), thus producing large amounts of ROS. The resulting oxidant stress cannot only oxidize critical targets in mitochondria, such as the adenine nucleotide translocator (ANT), implicated in formation of the mitochondrial permeability transition pore (mPT) [13], but also gradually damages the mitochondrial DNA (mtDNA), leading to abnormal ETC subunits and persistent oxidant stress through a vicious cycle mechanism [14] (see Chapter 6). Such animal models of chemically induced mitochondrially mediated pathology have been used primarily to characterize mitochondrial drug liabilities and to study mechanisms of mitochondrial injury. Ideally, such animal models should have a reporter system with which one can easily detect mitochondrial toxicity. Some of the currently used endpoints to assess mitochondrial liability in vivo or ex vivo are summarized in Table 1. The second approach involves transgenic or other techniques to induce mitochondrial dysfunction more akin to realistic circumstances in human patients. This can be either a systemic (i.e., affecting all tissues) or an organ-selective effect when organ-selective protein isoforms are manipulated. Overall, the advantage of such transgenic animal models for toxicological research is that they can be used to detect and study drug-induced mitochondrial toxicity because the animals are sensitized to superimposed mitochondria-targeting drug effects. This approach is aimed at modeling either an inherited or an acquired abnormality in mitochondrial function, and the normal organ wear and tear typical of most patients. Genetic models are normally utilized to study inherited mitochondrial disease, which emulates certain human mitochondrial conditions (mutations or deletions). Such abnormalities are inherited by the classical mechanisms if they involve genetic alterations encoded by nuclear DNA (nDNA). However, mutations and deletions can also involve mtDNA, which encodes for 13 peptide subunits of the ETC as well as for the mitochondria-speciﬁc 12S and 16S rRNAs and 22

542

ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY

TABLE 1 Endpoints Commonly Used to Assess Mitochondrial Liability in Animal Modelsa Clinical chemistry

Histopathology/electron microscopy

Tissue analysis

Functional assays with ex vivo isolated cells or mitochondria

Plasma AST activity (mitochondrial form) relative to ALT (cytosolic) Plasma lactate levels Microvesicular steatosis Alterations in mitochondrial structure, loss of cristae Megamitochondria Change in abundance ATP content Mitochondrial GSH/GSSG Thioredoxin-2 redox state Mitochondrial/cytosolic cytochrome c translocation Oxygen consumption De novo ATP biosynthesis Decrease in m Ca2+ -induced swelling (mPT) Aconitase activity Complex I, II, III, IV activity

a AST, aspartate aminotransferase; ALT, alanine aminotransferase; m , mitochondrial transmembrane potential.

tRNAs. Since these mtDNA mutations are inherited maternally, they have to be induced in germline mitochondria by transmitochondrial techniques (see below and Chapter 2). Alternatively, other animal models reﬂect noninherited (acquired) mitochondrial changes. For example, some animal models exist that carry somatic mitochondrial abnormalities in particular tissues; this can occur in certain diseases or in old age. 2.1. Animal Models with Acquired Mitochondrial Abnormalities Animal models that are based on acquired mitochondrial function are rare. One example is the carnitine deﬁciency rat model, in which rats were chronically exposed to the carnitine analog trimethylhydraziniumpropionate [15,16]. This leads to hepatic microvesicular steatosis as a result of massive inhibition of mitochondrial fatty acyl β-oxidation because mitochondria cannot import long-chain fatty acids into the matrix. Some animal models exhibit abnormalities in mitochondrial function secondary to certain diseases. Such animals normally serve as a model for these particular human diseases rather than as a model for mitochondrial dysfunction. For example, several animal models are in use for type 2 diabetes, including ob/ob mice, which are deﬁcient in the leptin gene; db/db mice, which are deﬁcient in the leptin receptor gene; KKAy mice, in which a mutation in the raly gene results in overexpression of the agouti gene; and the fa/fa rat, which is deﬁcient in the

ANIMAL MODELS OF MITOCHONDRIAL DYSFUNCTION

543

leptin receptor gene. In addition to the obesity, hyperglycemia, hyperlipidemia, and hyperinsulinemia characteristic of these models, mitochondrial dysfunction is found in all of them. Diabetes type 2 can also be induced by a high-fat diet combined with a low-dose streptozotozin in rats, leading to obesity and hyperglycemia [17]. Another model with secondary mitochondrial dysfunction is to use old animals, because during aging, oxidative damage to mtDNA gradually accumulates, leading to impaired mitochondrial function. However, such animals with gradually acquired mitochondrial damage have not been widely used as models for drug toxicity studies. Rather, the pathophysiological changes associated with degenerative diseases or aging are usually the focus. 2.2. Animal Models with mtDNA Mutations The development of transmitochondrial mice started relatively recently, during the late 1990s, and the technology to introduce foreign or altered mtDNA into somatic or germ cells remains challenging (reviewed in [18–20]). One approach was aimed at producing a heteroplasmic germline in mice ova (i.e., containing a mixture of wild-type and mutant mitochondria). However, it turned out that it was extremely difﬁcult to deliver foreign mtDNA genomes into intact mitochondria, in part because two membranes have to be crossed, and in part because an ovum contains approximately 105 wild-type mitochondria, and the fate of the mutated foreign mtDNA is uncertain. Despite these challenges, it was possible to create mito-mice as a new animal model of mtDNA-based diseases [21]. Speciﬁcally, mitochondria with a 4696-bp mtDNA deletion that included six tRNA genes and seven structural genes have been introduced successfully into mouse embryos. The deletion, which was transmitted maternally, was able to induce mitochondrial dysfunction in certain tissues. The mito-mice exhibited low body weights and lactic acidosis, and they died within several months, due to renal failure. Such mito-mice have not yet been used to study drug toxicity. 2.3. Animal Models with nDNA Mutations Animals with nDNA mutations of genes encoding for mitochondrial proteins have been generated by standard transgenic knock-in or knock-out techniques. A number of animal models currently available are described below. Among these, the phenotypes are similar and include impaired mitochondrial function and organ damage, although some mutations yield damage to selective organs, whereas others have more general effects. Tfam-Deﬁcient Mice These mice are deﬁcient in the mitochondrial transcription factor A. Using a loxP-ﬂanked Tfam allele in combination with a cre-recombinase transgene under the control of the muscle creatinine kinase promoter, Tfam was disrupted selectively in the heart and muscle. The homozygous negative mutation is lethal in utero, whereas heterozygous mice are viable. The latter exhibit a 50% reduction in Tfam transcript and protein levels, reduced mtDNA copy number,

544

ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY

and reductions in complex I and IV activities. Mutant animals developed a mosaic cardiac-speciﬁc progressive ETC deﬁciency, leading to dilated cardiomyopathy and death at 2 to 4 weeks post partum. This animal model mimics the hallmarks of mtDNA mutation disorders in humans [22,23]. ANT1 Knockout Mice The adenosine nucleotide translocator (ANT) is an abundant protein in the inner mitochondrial membrane that exchanges newly synthesized ATP in the matrix for cytosolic ADP. ANT may also play an important role in mediating cell death because it has been implicated as part of the multiprotein complex of the mitochondrial permeability transition (mPT) pore. Mice have two ANT genes (Ant1 and Ant2 ) [24]. Ant1 −/− mice are viable but exhibit overt myopathy and cardiomyopathy. Since both the ETC and the TCA cycle are inhibited, these mice have increased serum levels of lactate and other metabolic substrates. Mouse ANT1 is expressed in skeletal muscle and heart, while ANT2 is expressed in all tissues except skeletal muscle. Therefore, Ant1 -deﬁcient mice have normal ANT levels in the liver [25]. The consequence of this mutation is an inhibition of the F1 FO -ATP synthase due to substrate depletion, which leads to a stop in the proton ﬂux through the ATP synthase, and consequently, to an overall inhibition of the ETC. This in turn increases production of superoxide generated from the ETC, which will induce mtDNA damage. To compensate for the increased levels of ROS, there is an induction of MnSOD and glutathione peroxidase in muscle tissue. This animal model has not yet been used in drug toxicology. UCP-Knockout Mice Mammals have three genes for the uncoupling protein (UCP): Ucp1, 2, and 3 . The major function of UCP is to increase the inner membrane permeability to protons, resulting in the back-ﬂux of protons into the matrix, circumventing the F1 FO -ATPase and resulting in heat production, but also reducing production of ETC-derived ROS. Ucp2 - or Ucp3 -knockout mice exhibit increased m and hence increased oxidant stress [26]. This mouse model has occasionally been used to explore speciﬁc effects of drugs, such as methamphetamine-induced hyperthermia [27], but it has not been widely applied to drug toxicity studies. GGT-Knockout Mice γ-Glutamyltranspeptidase (GGT) is involved in glutathione (GSH) turnover and is expressed abundantly in kidney and in hepatobiliary epithelia. Ggt-gene knockout mice exhibit markedly reduced levels of hepatic GSH and increased transcript levels of the genes involved in GSH biosynthesis [28]. Because the cytosolic GSH pool is the source of mitochondrial GSH, GGT-deﬁcient mice have less than 50% of normal mitochondrial GSH levels. In addition, they have impaired ETC function and decreased mitochondrial ATP content [29]. Interestingly, this sustained mitochondrial oxidant stress, due to reduced GSH levels, does not have immediate adverse effects in young mice, but becomes apparent as the damage accumulates, eventually resulting in cell injury and death at less than 20 weeks

ANIMAL MODELS OF MITOCHONDRIAL DYSFUNCTION

545

of age. The sustained oxidative stress also leads to cumulative DNA damage [30]. This animal model has been utilized to reveal the mechanistic role of GGT or cysteine in selected cases of drug toxicity (e.g., for cisplatin) [31], but it has not been used to screen for mitochondrial toxicity with other drugs. GPx-1-Knockout Mice Glutathione peroxidase 1 (GPx-1) is highly expressed in liver, brain, and kidney, but is only weakly expressed in skeletal muscle and heart. The major function of GPx is to reduce H2 O2 to water by cooxidation of GSH, thereby avoiding potential hydroxyl production from Fenton chemistry. Gpx-1 −/− mice are viable but grow slowly. Liver mitochondria of such knockout animals have about fourfold higher H2 O2 levels than those of wild-type mitochondria [32]. Although used to demonstrate that organic hydroperoxides (substrates for GPx-1) were more toxic in Gpx-1 -null mice than in wild-type controls [33], this animal model has yet not found appreciable application in drug toxicology. SOD2-Knockout Mice Superoxide dismutase (SOD) catalyzes the dismutation of two superoxide anion radicals to hydrogen peroxide. The enzyme thus regulates both superoxide and hydrogen peroxide signaling and is a key determinant of the cellular redox status [34]. There are three isoforms; SOD1 ( = Cu, Zn-SOD) is dimeric and present in both the cytoplasm and mitochondrial intermembrane space, SOD2 ( = Mn-SOD) is abundant in the mitochondrial matrix, and SOD3 is an extracellular form akin to SOD1 but tetrameric. SOD2 is a homotetramer, and a knockout mouse model has been generated by deleting the exon that encodes for tetramer formation and manganese binding, thereby rendering it completely inactive [35,36]. Interestingly, recombinant deletion of the Sod1 or Sod3 gene has little effect on viability [37], although lack of SOD1 is associated with a number of abnormalities, including age-dependent skeletal muscle atrophy due to oxidant stress [38]. In contrast, Sod2 -null mice die postnatally due to dilated cardiomyopathy and/or neuronal degeneration, as these tissues are highly susceptible to mitochondrial oxidant stress [39,40]. Two different knockout models are available: the Sod2 tm1Cje mouse on the CD1 background [35] and the Sod2 tm1Leb mouse on the C57BL/6 background [41]. Because of the high and early mortality of homozygous knockout mice, most studies have used heterozygous mice (Sod2 +/− ). These mice have approximately 50% of the normal SOD2 mRNA and protein levels. The model has been well characterized, and some prominent biochemical or functional changes are summarized in Table 2. To determine alterations in mitochondrial protein expression in these mice on a systemic level, we have recently performed a proteomics analysis of hepatic mitochondria isolated from Sod2 +/− and wild-type mice [42]. This analysis demonstrates that a number of mitochondrial proteins are selectively up-regulated, whereas others, such as SOD1, exhibit lower expression levels in the Sod2 +/− mice (Figure 1). The data are complementary to a study with isolated cortical mitochondria from Sod2 -null mice [43]. However, in the latter study, the homozygous knockout mice had to be treated with low doses of mitochondria-targeting

546

TABLE 2

ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY

Hepatic Mitochondrial Abnormalities in Sod2 +/− Mice

Abnormal Function Increased superoxide net levels Decreased basal ATP production Decreased expression of Sod2 Decreased activity of Sod2 Decreased complex I activity Increased levels of 8-OH-dG Decreased mitochondrial total glutathione Decreased aconitase activity Increases in oxidatively modiﬁed proteins

Assay System Isolated mitochondria, DHE ﬂuorescence Isolated mitochondria, luciferin chemiluminescence Western blot Liver, biochemical assay Isolated mitochondria Isolated mitochondria, HPLC Isolated liver mitochondria, biochemical assay Isolated mitochondria, biochemical assay Isolated mitochondria, Western blot

Refs. 49 49 36,49 46 46 46 46 46 46

antioxidants and superoxide mimetics in order to survive long enough to complete the experiment (Sod2 -null mice are not viable). In contrast, heterozygous Sod2 +/− mice that we used for the proteomics analysis are viable and appear phenotypically normal; they exhibit normal growth curves, breed normally, and do not exhibit any clinical signs of abnormality [36]. The observed mitochondrial changes, all due to chronically increased levels of mitochondrial superoxide production, are subtle but accumulate over time. Many are similar to those occurring during normal aging, but they appear at a younger age, fostering use of Sod2 +/− mice for aging research [44]. The major consequences of this sustained mitochondrial oxidative stress include damage to Fe–S clusters in aconitase and complexes I, II, and III that impair the electron transport system. This lowers m and represses ATP production correspondingly. Mitochondrial macromolecules are damaged more extensively, and cells are more susceptible to various cytotoxic stressors (summarized in Figure 2). Because Sod2 +/− mice are sensitized to xenobiotics that impose an additional stress on mitochondria, they are ideally suited as a model for drug toxicity. 3. DEVELOPMENT OF THE Sod2+/− KNOCKOUT MOUSE MODEL FOR DRUG TESTING Recently, Sod2 +/− knockout mice have been shown to be more sensitive than wild-type mice to toxicants that impose oxidant stress and that impair mitochondrial function. For example, paraquat, which redox cycles to form superoxide anion, increased mortality in these mice [45]. Similarly, isolated mitochondria from Sod2 +/− mice undergo more extensive mPT after exposure to the prooxidant, t-butyl hydroperoxide [46].

DEVELOPMENT OF THE Sod2+/− KNOCKOUT MOUSE MODEL

547

Molecular weight (kDa) 150 75

Superoxide dismutase 2 Fold change: -2.17

50

Sod2 +/+

Sod2 +/-

25 20

Superoxide dismutase 1 Fold change: -1.77

15 Sod2 +/+

3

Isoelectric point (pI)

Sod2 +/-

11

Figure 1 Two-dimensional electrophoresis proﬁling of hepatic mitochondrial proteins from Sod2 +/+ (wild-type, shown left) and Sod2 +/− mice. Proteins were separated on the ﬁrst dimension using pH 3 to 11 nonlinear immobilized pH gradient (IPG) strips, followed by second-dimension separation by SDS-PAGE on 13% polyacrylamide gels. A −2.17-fold decrease in superoxide dismutase 2 (SOD2) protein relative to wild-type mice corresponds to a 50% reduction in SOD2 activity in the heterozygous Sod2 +/− mouse. Superoxide dismutase 1 (SOD1) protein was found to be less abundant relative to wild-type Sod2 +/+ mice. (See insert for color representation of ﬁgure.)

An important clue suggesting potential use of organ-selective toxicity has emerged from another model of SOD2 deﬁciency, where such loss is due to a deletion of the T-associated maternal effect (Tme) locus. This results in the removal of the coding region for SOD2 [47]. Treatment of Tme-deﬁcient mice with the neurotoxicant 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) at doses of 15 mg/kg intraperitoneally four times at intervals of two hours yielded massive brain damage after 3 weeks. A dose of 100 mg MPTP/kg was lethal to these SOD2-deﬁcient mice but not to wild-type animals. An interesting reason why the toxicity was induced selectively in the brain is the glial cell–mediated bioactivation of MPTP to MPP+ (see above), which is then taken up preferentially by neurons via the dopamine carrier and inhibits mitochondrial complex I [48]. These toxicokinetic features may explain the organotropic effects and led us to

548

ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY

O2 H+

I

SOD1

H+ H+ H+ F1F0ATPase

III e

Drug H+ DrugH

Drug H+

GSSG O2

DrugH

I

III IV

UCP-2

H2O Ask-1

GPx

Trx2red ProtS-S H2O2 Trx2 Red SOD2 (Fe) HO Trx2ox ProtSH H+ ONOO P-Ask-1 JNK pathway NO2 activation NO GSH

O2

Drug ‘Peroxidase’

H2O2

Drug

mtNOS ANT mPT VDAC pore Bcl-2

Cytc Smac AIF EndoG Bax

Figure 2 Role of SOD2 in mitochondria and consequences of abnormal expression of Sod2 . The major consequences include effects on the mitochondrial GSH pool and metabolism, the hydrogen peroxide–dependent effects on the Trx2/JNK system, and peroxynitrite-dependent reactions. For details, see the text.

speculate whether this mode of action could be adapted for other organs as well, including the liver. 3.1. Sod2+/− Mice Sensitized to Hepatotoxic Drugs that Target Mitochondria We have recently adopted the heterozygous Sod2 +/− mouse model for the evaluation of mitochondrial changes, and possible induction of hepatic-selective toxicity, after administration of certain drugs that have caused idiosyncratic liver injury in patients. One such drug is the COX-2 inhibitor nimesulide, which has been implicated in rare but serious cases of hepatotoxicity in susceptible patients. Repeated administration of low doses of nimesulide (10 mg/kg i.p. twice a day for 4 weeks) to young adult Sod2 +/− mice resulted in mitochondrial injury in the mutant but not in the wild-type mice [49]. These doses are comparable to clinical exposures after correction for interspecies differences using a dose scaling factor. Speciﬁcally, hepatic mitochondrial aconitase activity was decreased by 51% versus vehicle controls, and mitochondrial protein carbonyls were increased by 34%, indicating signiﬁcant levels of oxidant stress. In accord with this was a 28% net increase in mitochondrial glutathione (GSH) levels compared to vehicle controls, which reﬂects a compensatory up-regulation by prolonged mild oxidant

DEVELOPMENT OF THE Sod2+/− KNOCKOUT MOUSE MODEL

549

stress. Cytosolic levels of cytochrome c and caspase-3 activity were increased by 87 and 38%, respectively, indicating that mitochondrial permeabilization was occurring, thereby triggering caspase-mediated cell death. In accordance with this, the number of TUNEL-positive hepatocytes was greatly increased compared to vehicle controls. Another compound used to validate this Sod2 +/− model is troglitazone, a ﬁrst-generation antidiabetic thiazolidinedione that was withdrawn from the market in 2000 because of an unacceptable risk of hepatic injury. Troglitazone shows no hepatotoxicity in wild-type animals, but has dramatic effects in Sod2 +/− animals. When administered at doses, and reaching exposure levels, comparable to human dosage [50], troglitazone caused cumulative hepatic mitochondrial injury [51]. For example, 4-week treatment of Sod2 +/− mice with 30 mg troglitazone/kg daily i.p. caused a 45% decrease in aconitase activity and 46% decrease in complex I activity, paralleled by a 58% increase in protein carbonyls in isolated hepatic mitochondria. Moreover, this was accompanied by mild but signiﬁcant increases in serum ALT activity, a marker of liver injury, and focal hepatic necrosis (Figure 3). The data are compatible with treatment-related mitochondrial oxidative stress, which is revealed in animals having an underlying mitochondrial deﬁciency, but not in normal healthy animals with drug-naive livers. These ﬁndings were corroborated in isolated hepatocytes from Sod2 +/− mice where exposure to troglitazone increased mitochondrial superoxide production, as demonstrated

Figure 3 Liver injury (focal hepatic necrosis, arrows) following repeated treatment of Sod2 +/− mice with troglitazone (30 mg/kg daily i.p. for 4 weeks). Note that there were no apparent hepatic abnormalities in the liver of these mice treated for 2 weeks, nor were any changes seen in normal wild-type (Sod2 +/+ ) mice treated with troglitazone for 4 weeks. H&E staining, original objective magniﬁcation ×20.

550

ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY

by the mitochondria-selective ﬂuorescent probe MitoSox (see Chapter 18). Interestingly, Sod2 +/− knockout mice treated with troglitazone for only 2 weeks did not exhibit any clinical chemical or histopathological signs indicative of liver injury, suggesting that the compound-induced hepatotoxic response sets in abruptly after a lag time of several weeks, rather than early and developing gradually. Again, this is similar to the observed time course of troglitazone-associated liver injury in patients. This pattern also underscores the notion that cumulative silent mitochondrial injury leads to overt liver injury only after a critical threshold of mitochondrial dysfunction and failure has been reached.

4. MITOCHONDRIAL ABNORMALITIES: A POSSIBLE LINK TO IDIOSYNCRATIC DRUG HEPATOTOXICITY Drug-induced liver injury (DILI) has been the single major cause for withdrawal of drugs launched successfully or for discontinuation of development of a potential new drug. In most cases, the toxicity clearly has an idiosyncratic (i.e., speciﬁc host-dependent) basis, but the underlying mechanisms of injury are not known. As contended in this book, a novel concept implicates mitochondrial dysfunction in this toxicity, with preexisting mitochondrial impairment and organ history serving as key determinants of idiosyncratic susceptibility. In this model, drug-induced mitochondrial dysfunction pushes the cell beyond a bioenergetic and/or oxidative threshold, triggering overt hepatotoxicity [52]. Thus, mitochondrial impairment of genetic or xenobiotic etiology can be a key factor dictating the intrinsic toxicity of a mitochondria-targeting compound taken up and activated not only in the liver, but also in other tissues. Some of the most obvious mitochondrial abnormalities in humans are overt (see Chapter 11), but it seems likely that a relatively high number of clinically silent mutations in mtDNA and nuclear DNA encoding proteins destined for mitochondrial import, or acquired abnormalities in mitochondrial proteins, remain undetected (see Chapter 4). Moreover, heteroplasmy (coexistence of normal wild-type and mutated mtDNA in the same cell) will protect the cell until a threshold of abnormal mitochondria has been crossed, which will precipitate a sudden and abrupt onset of pathology, a pattern typical in idiosyncratic DILI. The most sensitive organs are highly aerobically poised, including the central nervous system, the cardiovascular system, and the kidneys, but also the liver, which is typically exposed to higher concentrations of oral drugs because of hepatic portal circulation. Previously, it has been difﬁcult to develop a reliable and reproducible animal model that mimics these silent mitochondrial deﬁciencies and that could therefore be used to study the mechanisms and pathways of DILI and predict clinical disposition. However, the Sod2 +/− model, with its phenotype of gradually accumulating oxidative damage to the mtDNA and ETC proteins, resembles the phenotype of a variety of human mitochondrial disorders and reveals drug toxicity in accord with clinical evidence. As such, this model may more faithfully

REFERENCES

551

predict idiosyncratic DILI where genetic or acquired mitochondrial impairment contributes to etiology. 5. IMPLICATIONS AND SIGNIFICANCE OF THE USE OF ANIMAL MODELS OF MITOCHONDRIAL TOXICITY IN DRUG SAFETY ASSESSMENT It is reasonable to assume that underlying mitochondrial impairment arising from aging, diabetes, or xenobiotic exposure can greatly inﬂuence superimposed drug effects, so that animal models of mitochondrial disease would offer a great opportunity to study etiology and treatment. Ideally, deleterious drug effects would be detected in such animal models but not in normal healthy animals, as is the case for troglitazone in the Sod2 +/− model. A remaining variable is the choice of the right model, which is dictated by the question at hand. For example, mito-mice, featuring large deletions, or speciﬁc gene knockout or knock-down models such as the Sod2 +/− mouse could be used to identify potential organ-selective drug-induced toxic responses. In any event, structural drug analogs not implicated in human adverse effects should be run concurrently as negative controls in all such animal studies. In addition, animal models of other pathology, such as diabetes or neurodegenerative disease, could also be evaluated to determine how well they reveal drug-induced mitochondrial toxicity. Despite the great potential of the animal models discussed here, one should add a note of caution. To avoid false positive and/or false negative data, it is mandatory to assess possible species-speciﬁc toxicokinetic and toxicodynamic differences, such as metabolism and disposition of drugs, as well as interaction of a drug with its target. Unfortunately, such information is not required by regulatory agencies, so no historical data are available with which to compare new results. The search for animal models of mitochondrial toxicity will undoubtedly continue, and it is hoped that novel models will continue to elucidate the molecular mechanisms of drug toxicity. Acknowledgements This work was supported in part by the Boehringer Ingelheim Endowed Chair in Mechanistic Toxicology at UCONN, as well as grants from the Biomedical Research Council Singapore (R-184-000-096-305 to U.A.B.) and NUS Ofﬁce of Life Sciences Toxicology Program (R-184-000-079-712 to U.A.B.). Y.H.L. is a NUS Graduate Research Scholarship recipient. REFERENCES 1. Szewczyk A, Wojtczak L. Mitochondria as a pharmacological target. Pharmacol Rev. 2002;54:101–127.

552

ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY

2. Amacher DE. Drug-associated mitochondrial toxicity and its detection. Curr Med Chem. 2005;12:1829–1839. 3. Kass GEN. Mitochondrial involvement in drug-induced hepatic injury. Chem Biol Interact. 2006;163:145–159. 4. Chan K, Truong D, Shangari N, O’Brien PJ. Drug-induced mitochondrial toxicity. Expert Opin Drug Metab Toxicol. 2005;1:655–669. 5. Boelsterli UA, Animal models of human disease in drug safety assessment. J Toxicol Sci. 2003;28:109–121. 6. Heikkila RE, Hess A, Duvoisin RC. Dopaminergic neurotoxicity of 1-methyl-4phenyl-1,2,5,6-tetrahydropyridine in mice. Science. 1984;224:1451–1453. 7. Hayley S, Crocker SJ, Smith PD, et al. Regulation of dopaminergic loss by Fas in a 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine model of Parkinson’s disease. J Neurosci. 2004;24:2045–2053. 8. Jaeschke H, Bajt ML. Intracellular signaling mechanisms of acetaminophen-induced liver cell death. Toxicol Sci. 2006;89:31–41. 9. Kon K, Kim JS, Jaeschke H, Lemasters JJ. Mitochondrial permeability transition in acetaminophen-induced necrosis and apoptosis of cultured mouse hepatocytes. Hepatology. 2004;40:1170–1179. 10. Masubuchi Y, Suda C, Horie T. Involvement of mitochondrial permeability transition in acetaminophen-induced liver injury in mice. J Hepatol. 2005;42:110–116. 11. Gunawan B, Liu ZX, Han D, Hanawa N, Gaarde WA, Kaplowitz N. c-Jun N-terminal kinase plays a major role in murine acetaminophen hepatotoxicity. Gastroenterology. 2006;131:165–178. 12. Latchoumycandane C, Goh CW, Ong MMK, Boelsterli UA. Mitochondrial protection by the JNK inhibitor leﬂunomide rescues mice from acetaminophen-induced liver injury. Hepatology. 2007;45:412–421. 13. Oliveira PJ, Wallace KB. Depletion of adenine nucleotide translocator protein in heart mitochondria from doxorubicin-treated rats: relevance for mitochondrial dysfunction. Toxicology. 2006;220:160–168. 14. Zhou S, Palmeria CM, Wallace KB. Doxorubicin-induced persistent oxidative stress to cardiac myocytes. Toxicol Lett. 2001;121:151–157. 15. Spaniol M, Brooks H, Auer L, et al. Development and characterization of an animal model of carnitine deﬁciency. Eur J Biochem. 2001;268:1876–1887. 16. Spaniol M, Kaufmann P, Beier K, et al. Mechanisms of liver steatosis in rats with systemic carnitine deﬁciency due to treatment with trimethylhydraziniumpropionate. J Lipid Res. 2003;44:144–153. 17. Sawant SP, Dnyanmote AV, Warbritton A, Latendresse JR, Mehendale HM. Type 2 diabetic rats are sensitive to thioacetamide hepatotoxicity. Toxicol Appl Pharmacol. 2006;211:221–232. 18. Inoue K, Ogura A, Hayashi JI. Production of mitochondrial DNA transgenic mice using zygotes. Methods. 2002;26:358–363. 19. Pinkert CA, Trounce IA. Production of transmitochondrial mice. Methods. 2002;26:348–357. 20. Pinkert CA, Trounce IA. Animal modeling: from transgenics to transmitochondrial models. In Mitochondria in Health and Disease (CD Berdanier, ed). Boca Raton, FL: Taylor & Francis; 2005:559–580.

REFERENCES

553

21. Nakada K, Inoue K, Hayashi JI. Mito-mice: animal models for mitochondrial DNA-based diseases. Semin Cell Dev Biol. 2001;12:459–465. 22. Larsson NG, Wang G, Wilhelmsson H, et al. Mitochondrial transcription factor A is necessary for mtDNA maintenance and embryogenesis in mice. Nat Genet. 1998;18:199–200. 23. Wang J, Wilhelmsson H, Graff C, et al. Dilated cardiomyopathy and atrioventricular conduction blocks induced by heart-speciﬁc inactivation of mitochondrial DNA gene expression. Nat Genet. 1999;21:133–137. 24. Levy SE, Chen YS, Graham BH, Wallace DC. Expression and sequence analysis of the mouse adenine nucleotide translocase 1 and 2 genes. Gene. 2000;254:57–66. 25. Graham B, Waymire K, Cottrell B, Trounce IA, MacGregor GR, Wallace DC. A mouse model for mitochondrial myopathy and cardiomyopathy resulting from a deﬁciency in the heart/skeletal muscle isoform of the adenine nucleotide translocator. Nat Genet. 1997;16:226–234. 26. Arsenijevic D, Onuma H, Pecqueur C, et al. Disruption of the uncoupling protein-2 gene in mice reveals a role in immunity and reactive oxygen species production. Nat Genet. 2000;26:435–439. 27. Sprague JE, Mallett NM, Rusyniak DE, Mills E. UCP3 and thyroid hormone involvement in methamphetamine-induced hyperthermia. Biochem Pharmacol. 2004;68:1339–1343. 28. Habib GM, Shi ZZ, Ou CN, Kala G, Kala SV, Lieberman MW. Altered gene expression in the liver of γ-glutamyl transpeptidase-deﬁcient mice. Hepatology. 2000;32:556–562. 29. Will Y, Fischer KA, Horton RA, et al. γ-Glutamyltranspeptidase-deﬁcient knockout mice as a model to study the relationship between glutathione status, mitochondrial function, and cellular function. Hepatology. 2000;32:740–749. 30. Rojas E, Valverde M, Kala SV, Kala G, Lieberman MW. Accumulation of DNA damage in the organs of mice deﬁcient in γ-glutamyltranspeptidase. Mutat Res. 2000;447:305–316. 31. Hanigan MH, Lykissa ED, Townsend DM, Ou CN, Barrios R, Lieberman MW. γ-Glutamyl transpeptidase–deﬁcient mice are resistant to the nephrotoxic effects of cisplatin. Am J Pathol. 2001;159:1889–1894. 32. Esposito LA, Kokoszka JE, Waymire K, Cottrell B, MacGregor GR, Wallace DC. Mitochondrial oxidative stress in mice lacking the glutathione peroxidase-1 gene. Free Radic Biol Med. 2000;28:754–766. 33. Liddell JR, Dringen R, Crack PJ, Robinson SR. Glutathione peroxidase 1 and a high cellular glutathione concentration are essential for effective organic hydroperoxide detoxiﬁcation in astrocytes. Glia. 2006;54:873–879. 34. Buettner GR, Ng CF, Wang M, Rodgers VGJ, Schafer FQ. A new paradigm: Manganese superoxide dismutase inﬂuences the production of H2 O2 in cells and thereby their biological state. Free Radic Biol Med. 2006;41:1338–1350. 35. Li Y, Huang TT, Carlson EJ, et al. Dilated cardiomyopathy and neonatal lethality in mutant mice lacking manganese superoxide dismutase. Nat Genet. 1995;11:376–381. 36. Van Remmen H, Salvador C, Yang H, Huang TT, Epstein CJ, Richardson A. Characterization of the antioxidant status of the heterozygous manganese superoxide dismutase knockout mouse. Arch Biochem Biophys. 1999;363:91–97.

554

ANIMAL MODELS OF DRUG-INDUCED MITOCHONDRIAL TOXICITY

37. Morten KJ, Ackrell BAC, Melov S. Mitochondrial reactive oxygen species in mice lacking superoxide dismutase 2. J Biol Chem. 2006;281:3354–3359. 38. Muller FL, Song W, Liu Y, et al. Absence of CuZn superoxide dismutase leads to elevated oxidative stress and acceleration of age-dependent skeletal muscle atrophy. Free Radic Biol Med. 2006;40:1993–2004. 39. Huang TT, Carlson EJ, Raineri I, Gillespie AM, Kozy H, Epstein CJ. The use of transgenic and mutant mice to study oxygen free radical metabolism. Ann N Y Acad Sci. 1999;893:95–112. 40. Melov S, Coskun P, Patel M, et al. Mitochondrial disease in superoxide dismutase 2 mutant mice. Proc Natl Acad Sci U S A. 1999;96:846–851. 41. Lebovitz RM, Zhang H, Vogel H, et al. Neurodegeneration, myocardial injury, and perinatal death in mitochondrial superoxide dismutase-deﬁcient mice. Proc Natl Acad Sci U S A. 1996;93:9782–9787. 42. Lee YH, Boelsterli UA, Lin Q, Chung MCM. Proteomics proﬁling of hepatic mitochondria in heterozygous Sod2 +/− mice, an animal model of discreet mitochondrial oxidant stress. Proteomics. 2008;8:555–568. 43. Hinerfeld D, Traini MD, Weinberger RP, et al. Endogenous mitochondrial oxidative stress: neurodegeneration, proteomics analysis, speciﬁc respiratory chain defects, and efﬁcacious antioxidant therapy in superoxide dismutase 2 null mice. J Neurochem. 2004;88:657–667. 44. Kokoszka JE, Coskun P, Esposito L, Wallace DC. Increased mitochondrial oxidative stress in the Sod2 +/− mouse results in the age-related decline of mitochondrial function culminating in increased apoptosis. Proc Natl Acad Sci U S A. 2001;98:2278–2283. 45. Van Remmen H, Qi W, Sabia M, et al. Multiple deﬁciencies in antioxidant enzymes in mice result in a compound increase in sensitivity to oxidative stress. Free Radic Biol Med. 2004;36:1625–1634. 46. Williams MD, Van Remmen H, Conrad CC, Huang TT, Epstein CJ, Richardson A. Increased oxidative damage is correlated to mitochondrial function in heterozygous manganese dismutase knockout mice. J Biol Chem. 1998;273:28510–28515. 47. Cortopassi G, Wang E. Modelling the effects of age-related mtDNA mutation accumulation: complex I deﬁciency, superoxide and cell death. Biochim Biophys Acta. 1995;1271:171–176. 48. Cleeter MW, Cooper JM, Schapira AH. Irreversible inhibition of mitochondrial complex I by 1-methyl-4-phenylpyridinium: evidence of free radical involvement. J Neurochem. 1992;58:786–789. 49. Ong MMK, Wang AS, Leow KY, Khoo YM, Boelsterli UA. Nimesulide-induced hepatic mitochondrial injury in heterozygous Sod2 +/− mice. Free Radic Biol Med. 2006;40:420–429. 50. New LS, Saha S, Ong MMK, Boelsterli UA, Chan ECY. Pharmacokinetic study of intraperitoneally administered troglitazone in mice using UPLC/MS/MS. Rapid Commun Mass Spectrom. 2007;21:982–988. 51. Ong MMK, Latchoumycandane C, Boelsterli UA. Troglitazone-induced hepatic necrosis in an animal model of silent mitochondrial abnormalities. Toxicol. Sci. 2007;97:205–213. 52. Boelsterli UA, Lim PLK. Underlying mitochondrial abnormalities: a link to idiosyncratic drug hepatotoxicity? Toxicol Appl Pharmacol. 2007;220:92–107.

24 NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION USING NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY Robert W. Wiseman Biomedical Imaging Research Center, Departments of Physiology and Radiology, Michigan State University, East Lansing, Michigan

J. A. L. Jeneson Biomedical NMR Laboratory, Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands

1. Introduction 2. Spin physics and the NMR experiment 3. Probing mitochondrial function using 31 P-NMR spectroscopy 3.1. Information from the intracellular pH measurement 3.2. Information from phosphocreatine kinetics 3.3. Reconstruction of the mitochondrial transfer function from multiple perturbations 3.4. Phosphocreatine recovery dynamics following a single metabolic perturbation 4. Conclusions

555 557 561 561 564 567 568 569

1. INTRODUCTION Techniques to assess and monitor mitochondrial function have been described in a number of the earlier chapters. These include monitoring O2 consumption via Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

555

556

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

several technologies (Chapters 12 to 14), and assessment of membrane potential and organellar swelling (Chapter 17). Determining the phosphate potential, essentially the amounts of ATP, ADP, AMP, phosphocreatine (PCr), and inorganic phosphate (Pi ), provides an alternative means of assessing mitochondrial function. Under most conditions this potential is heavily biased toward ATP, which is buffered by PCr and creatine kinase in mammals and by other analogs in invertebrates. One could freeze-clamp the tissue and perform biochemical assays for the various components, but this technology is laborious and provides a mere snapshot of what is a dynamic equilibrium between several reactants. Nuclear magnetic resonance (NMR) techniques take advantage of the magnetic asymmetry that is inherent in many naturally occurring atoms, notable the hydrogen nuclei in water, 13 C, and 31 P, among many others (Table 1). These are stable, nonradioactive atoms, and the in vivo abundance of some is sufﬁcient to permit direct detection without exogenous supplementation. The 31 P signals of ATP, PCr, and Pi are spectroscopically distinct, and the dynamic interplay between them provides insight into cellular and/or tissue bioenergetic equilibria. As described in this chapter, one can monitor the phosphate potential while imposing a bioenergetic demand, such as depolarization for neurons or contraction for muscle, and follow the depletion of PCr and eventually, ATP. Unless excessively severe, withdrawal of the stimulus yields a return to the normal steady state. These changes and the kinetics of recovery provide much insight into mitochondrial capacity and function. In this way, NMR provides a noninvasive, real-time view of bioenergetic resilience and kinetics that directly reﬂect mitochondrial function. Magnetic resonance imaging (MRI) is essentially a series of spectroscopic observations acquired in the presence of magnetic ﬁeld gradients to resolve anatomy in addition to detection of metabolite concentration, and both spectroscopic and imaging techniques are well suited for assessment of mitochondrial function in vivo. NMR also provides noninvasive determination of lactate levels in vivo, a hallmark of mitochondrial deﬁciency [1–3]. Based on similar technologies, following signal shifts with various metabolites containing 13 C allows the mapping of metabolic pathways and monitoring carbon ﬂux in real time. Mitochondrial pathways, such as Krebs cycle ﬂux, among others, are readily probed using this technique. The application of NMR spectroscopic and MRI techniques to biology and medicine has spanned decades and made considerable contributions to our understanding of the underlying mechanisms of disease and their detection. The breadth of these contributions far exceeds the scope of the present chapter, and the interested reader is directed to excellent reviews of work in spectroscopy [4,5] and imaging [6] for more details on theory and application to biological systems. In the present chapter we describe NMR spectroscopy techniques applied to the detection of mitochondrial function in excitable tissues. A modicum of spin physics is presented to appreciate both theoretical and practical considerations for the typical NMR experiment investigating mitochondrial function noninvasively.

557

SPIN PHYSICS AND THE NMR EXPERIMENT

2. SPIN PHYSICS AND THE NMR EXPERIMENT Magnetic resonance techniques rely on a characteristic of the atomic nucleus termed the angular momentum, a property that causes it to spin gyroscopically. Figure 1 shows the random distribution of magnetic dipoles present in a tube of water (Figure 1A). These particles possess a magnetic dipole moment depending on charge state, which results in these spins aligning with an externally applied magnetic ﬁeld (Figure 1B). The spin populations align along the main axis of the static magnetic ﬁeld either parallel (low energy) or antiparallel (higher energy) to the magnet axis, but because the lower-energy state is slightly favored, the bulk vector aligns with the static magnetic ﬁeld (B 0 ). The magnetic ﬁeld exerts torque on these molecules, resulting in precession, much like a child’s top. The

A

+ +

+

H2O

+

+ +

+

+

+

+ B

S B0 N H2O

S

+

z

+

+ +

y

+

x

+ N

Figure 1 The principles of magnetic moments and precession can be simply illustrated with a test tube of water. (A) Test tube and a sampling of water molecules. The charge state of the nucleus allows these molecules to act as dipoles that possess a magnetic moment (akin to atomic bar magnets). The arrows through the center of each water molecule indicate the direction of spin, which in the absence of an external magnetic ﬁeld, is random. (B) Effect on the same sample in the presence of a static magnetic ﬁeld (B 0 ). The spin systems align with B 0 , and more spins align parallel to the magnetic ﬁeld than antiparallel. Thus, the bulk magnetization M z is aligned with the z -axis in three-dimensional space. The lower-energy-state spins will result in the NMR signal that we observe.

558

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

rate of precession is characteristic for the nucleus and the ﬁeld strength that is applying the torque, and is termed the Larmor frequency: ω = γB0

(1)

where ω is the Larmor frequency, γ the gyromagnetic ratio, and B 0 the static magnetic ﬁeld strength. The gyromagnetic ratio (γ) is a proportionality between the angular momentum of the dipole and the applied magnetic ﬁeld. As a practical matter, one can easily deduce from inspection of equation (1) that precessional frequency increases with magnetic ﬁeld strength. Furthermore, there is an increase in the apparent signal-to-noise ratio (SNR) as well. Table 1 shows a list of nuclei of biological interest together with their spin quantum numbers, natural abundance, and gyromagnetic ratios. The spin population is sampled by application of a second applied magnetic ﬁeld (B 1 ) that is orthogonal or perpendicular to the static magnetic ﬁeld B 0 . This second ﬁeld is delivered by a transmitter coil adjacent to the sample and tuned to resonate at the frequency of biological interest (see Table 1 for examples). The spin system absorbs this energy and realigns along the XY plane to spin about its axis and in doing so, generate a small current that can be detected by the same coil acting as a receiver (Figure 2). Eventually, the resonant energy is slowly lost to the surrounding environment and the signal disappears, termed the free induction decay (FID; Figure 2). The raw NMR signal is transformed from the time domain to frequency using Fourier transform methods to generate a spectrum that contains two valuable pieces of information: peak amplitude (signal) and peak position (chemical shift). Peak amplitudes are directly proportional to the concentration within the sample but can be inﬂuenced by their relaxation properties and the acquisition parameters. There are two relaxation rates, which result in dissipation of the signal through losing energy to the system (spin–lattice relaxation) and loss of coherence of the spin population (spin–spin relaxation). If the spin systems are sampled at too high a rate relative to their relaxation(s), then quantitation, although not impossible, is much more complicated and can lead to misestimates of chemical exchange

TABLE 1

Common Nuclei Used in Biology and Their Magnetic Properties Spin

Natural Abundance (%)

γ/2π (MHz/T)

1 2

99.98

42.576

P

1 2

100.0

17.235

Sodium

23 Na

3 2

100.0

11.262

Carbon

13

C

1 2

1.1

10.705

Nitrogen

15

N

1 2

0.36

−4.3156

Name

Symbol

Hydrogen

1

Phosphorus

31

H

559

SPIN PHYSICS AND THE NMR EXPERIMENT

S

z

H2O

x

y The NMR signal! MY t

N

Figure 2 Generation of the NMR signal derives from perturbing M z with a second magnetic ﬁeld orthogonal or perpendicular to B 0 using a radio-frequency transceiver coil. This applied ﬁeld (B 1 ) transmits at the Larmor frequency for the nuclei to be detected and results in the bulk magnetization vector moving from the z -axis to either the x - or y-axis, depending on the pulse programmed to be delivered. As M xy rotates about the z -axis it generates current in the transceiver coil as a decaying sine wave. The free induction decay (FID) is converted from the time domain to the frequency domain using a Fourier transform to create the NMR spectrum. The decay of the NMR signal results from a return of the bulk magnetization to M z (T1 relaxation) and a dephasing of the bulk vector immediately after the applied B 1 pulse (T2 relaxation).

dynamics. Details of the practical aspects of sampling rates and signal-to-noise considerations can be found in an excellent work by DeGraaf [5]. Spin–lattice or T1 relaxation refers to the rate at which spins return to their equilibrium state after having absorbed energy from the sampling pulse delivered orthogonal to the static magnetic ﬁeld. For phosphorus NMR studies the range of values for spin–lattice relaxation times of the biologically interesting phosphates (ATP, phosphocreatine, and inorganic phosphate) are on the order of 2 to 5 seconds, depending on the species and the ﬁeld strength at which these rates are measured [7–10]. From a practical perspective, T1 values set the outside limit of a quantitative experiment for signal averaging because to avoid saturation artifacts, the delay between sampling pulses must be ﬁve times the longest T1 value within the spectrum. For dynamic spectroscopy where time resolution is the primary concern, shorter delays can be used, but shorter sampling pulses as well, to avoid signal saturation. Spin–spin or T2 relaxation results from a dephasing of the bulk magnetization following the sampling pulse in the XY plane and disperse based on their precessional frequencies. In practical terms T2 values often set the maximal sampling rate for the NMR experiment. In phosphorus experiments typical T2 values are tens of milliseconds and do not vary greatly with ﬁeld strength [11–14]. The second piece of information that is important in biological NMR spectroscopy is the chemical shift position of the resonant species. The

560

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

electron cloud surrounding each atom creates a unique chemical environment that resonates at a slightly higher (if less electron dense) or lower (if more electron dense) frequency than the carrier frequency. This creates a dispersion of frequencies so that each atom in a functional group resonates at a same frequency but distinctly different from other functional groups. Termed the chemical shift and expressed in dimensionless units of parts per million, this parameter affords the advantage that it is independent of magnetic ﬁeld strength. The chemical shift (δ) is calculated as follows: δ=

νs − νr × 106 νr

(2)

where νs is the resonance frequency of the sample and νr is the resonance frequency of the reference. For in vivo phosphorus spectroscopy in striated muscle, the spectrum presents information on the inorganic phosphate, phosphocreatine, and ATP content as well as intracellular pH from chemical shift information (see below). The area under each peak is directly proportionate to its concentration when acquired under the proper conditions. A typical phosphorus spectrum from rodent hindlimb muscle is presented in Figure 3. Note that the inorganic

PCr

ATP

Pi

10

5

0

–5

–10 PPM

–15

–20

–25

Figure 3 Rat hindlimb phosphorus metabolites detected by 31 P NMR spectroscopy. A 0.8-cm surface coil was placed adjacent to the superﬁcial gastrocnemius muscle. Data were acquired using a horizontal-bore 4.7-T GE spectrometer with a sweep width of 4000 Hz and 4096 data points (45◦ pulse width, 2.6-s delay, and 64 data transients). Data were apodized with a 15-Hz exponential ﬁlter and zero-ﬁlled once prior to the Fourier transform. Peak identiﬁcation from left to right: inorganic phosphate (Pi ), phosphocreatine (PCr), and γ, α, and β ATP. Chemical content of the sample is directly proportional to the area under each peak. Additional information that can be derived from this spectrum are the intracellular pH [equation (3)], free ADP [equation (7)] and the phosphate chemical potential or free energy of ATP hydrolysis (G ATP ) [equation (13)].

PROBING MITOCHONDRIAL FUNCTION USING

31 P-NMR

SPECTROSCOPY

561

phosphate (Pi ), phosphocreatine (PCr), and the γ, α, and β resonances of adenosine triphosphate (ATP) are all in different positions along the chemical shift axis. The ﬁnal practical consideration for the NMR experiment relates to the inherent insensitivity of the method itself. NMR detection derives from the difference in energy states between low- and high-energy nuclei and the number of nuclei in the lower-energy state able to be excited by the sampling pulse. This number is relatively small. As such, the ability of the system to absorb energy limits detection to a range of 0.1 mM. The signal acquired is directly proportional to the size of the free induction decay, reﬂecting the number of spins in the receiver range. Therefore, the signal-to-noise ratio in a given experiment depends on a number of parameters, most notably the size of the sample and the number of acquisitions. Biological questions requiring high time resolution (short delays between spectra) require larger samples and higher-ﬁeld-strength instruments. Current state-of-the-art clinical scanners operate at 3.0 Tesla magnetic ﬁeld strength and provide excellent SNR for most human studies. For smaller mammals (such as transgenic mice), even higher ﬁeld strengths (9.4 to 11 Tesla) are typically used. 3. PROBING MITOCHONDRIAL FUNCTION USING SPECTROSCOPY

31 P-NMR

Phosphorus metabolites important to mitochondrial function have been shown to be fully visible to detection by NMR spectroscopy. The depots of adenine nucleotides (ATP, ADP, and AMP), inorganic phosphate (Pi ), and phosphocreatine (PCr) can be measured in vivo in real time and have been shown to diffuse freely within the cytosol [15–19]. On this basis, two indices of mitochondrial function can be determined routinely using phosphorus NMR spectroscopy. The ﬁrst index is the cellular phosphate chemical potential, or Gibbs free energy of ATP hydrolysis (G ATP ) that can be maintained by the mitochondrial pool. By default, this is a steady-state concentration measurement and requires quantiﬁcation of the metabolites as well as calculation of the intracellular pH and free ADP. The second index is the kinetic capacity, which can be evaluated by measuring the dynamic change in the concentration of any given metabolite (but typically, PCr and Pi ) following a metabolic perturbation, and requires sufﬁcient signal and time resolution for these measurements. The quantiﬁcation of any mitochondrial property, however, ﬁrst requires estimates of cytosolic buffering capacities for both protons and adenylates before absolute changes in chemical content and cytosolic thermodynamic potentials can be determined. 3.1. Information from the Intracellular pH Measurement Intracellular pH (pHi ) determinations have been used extensively in NMR since publication of the landmark paper by Moon and Richards in 1973 [20], and the biological implications for pH regulation of cell function has been reviewed extensively [21]. The measurement relies on the protonation state of the Pi resonance. The pK a of phosphate at physiologic temperatures is 6.75, and therefore at

562

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

normal pHi values the intracellular inorganic phosphate resonance exists in two states, H2 PO4 − and HPO4 2− . However because the chemical exchange is faster than the time resolution of the NMR measurement Pi appears as one resonance and is simply the algebraic sum of the amounts of both species present in the sample. When intracellular pH decreases, due to glycolytic activity, for example, the relative amount of H2 PO4 − increases and HPO4 2− decreases. Because the resonance frequencies are distinct but the exchange of protons between these two protonation states is swift, the apparent chemical shift of Pi will move relative to the other peaks within the spectrum. The intracellular pH can thus be calculated as follows: δ0 − δa pHi = pKa + log (3) δb − δ0 where pK a is the second dissociation constant for inorganic phosphate, δa the chemical shift of the acid form (HPO4 2− ), δb the chemical shift of the basic form (H2 PO4 − ), and δ0 the Pi chemical shift observed. For mammalian systems the pK a and chemical shift endpoints have been determined empirically by titration [22,23]. The values for the chemical shift observed are referenced to a second phosphate-containing peak, ideally where there is no dissociable proton or the pK a is several orders of magnitude below the pHi being measured. These criteria are satisﬁed by phosphocreatine in mammalian systems [23] and phosphoarginine in invertebrates [24]. However, as pHi values approach pH 6, where the pH sensitivity of the reference should be also considered, one can solve for pHi using the equations presented in Wiseman et al. [23]. The proton-buffering capacity can be determined in tissues by monitoring the intracellular pH and then employing one of several methods: permeant acids and bases [24] or exercise-induced acidosis [25]. The acid or base pulsed methods rely on the dissociation state of the permeant species. Adapted from microelectrode measurements, these techniques use a prepulse of ammonium chloride (NH4 Cl) or a pulse of a permeant acid such as 5,5-dimethyloxyzolidine-2,4-dione to cause a calculable base or acid load, respectively [24]. The proton load calculated relies on knowledge of the pK a of the species being utilized, its concentration, and the temperature [21]. The buffering capacity is calculated, knowing the proton load and the measured pHi change, by 31 P-NMR. The advantage of these two methods is that the method can be carried out on resting tissue and is independent of other chemical changes. The second class of methods was developed by Adams et al. [25] using a contracting skeletal muscle preparation. This method relies on precise knowledge of the stoichiometric coefﬁcient for protons in the creatine kinase reaction (α) [25]: αH+ + PCr + ADP ↔ ATP + Cr

(4)

The initial hydrolysis of PCr consumes protons and causes a transient alkalinization due to the differing protonation states of the phosphate that is transferred

PROBING MITOCHONDRIAL FUNCTION USING

31 P-NMR

SPECTROSCOPY

563

from creatine to ADP. By monitoring the extent of PCr consumed as well as the change in pHi , one can calculate buffering capacity as follows: β=

PCr × α(pH) pH

(5)

where PCr is ([PCr])t = i − ([PCr])t = 0 and pHi is (pH)t = i − (pH)t = 0 with t = 0 corresponding to the end time point of the initial alkalinization, and α(pH) is the proton stoichiometric coefﬁcient for the creatine kinase reaction [equation (4)] [25]. Figure 4 shows a representative time series of pHi changes in contracting skeletal muscle from rodent hindlimb muscle. The initial transient alkalinization used to calculate buffer capacity is shown as a dotted line starting at time zero (no stimulation). The ﬁnal piece of information that can be obtained from intracellular pH measurements by 31 P-NMR spectroscopy is an estimate of the nonmitochondrial glycolytic ATP synthesis ﬂux. It can be calculated from the rate of pH change during acidiﬁcation of the cell during metabolic perturbation but after the initial alkalinization period (Figure 4, solid line). This rate is calculated as follows: JGLY = 1.5 × β ×

dpHi dt

(6)

β Intracellular pH

7.1 7.0 ~JATP Gly

6.9 6.8 6.7 0

100

200 Time (sec)

300

400

Figure 4 A plot of intracellular pH changes versus time provides information on proton buffering (β) (dotted line) and glycolytic proton production (solid line). To generate this plot a series of spectra were acquired from the rat hindlimb (as in Figure 3); however, the sciatic nerve was electrically paced to induce a series of contractions. Intracellular pH was calculated according to [equation (3)] and plotted over the time course of the experiment. The initial alkalinization is the result of a change in pK a for phosphate when bound to PCr versus ATP. Knowledge of this coefﬁcient and a measure of the amount of PCr consumed yields β [equation (5)]. Analysis of the glycolytic proton production (solid line) yields GLY estimates of ATP synthesis capacity by this pathway (JATP ) using equation (6).

564

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

where 1.5 is the stoichiometric coefﬁcient for ATP/H+ , β the buffering capacity, and d (pHi )/dt the ﬁrst derivative of the pHi function during contractile activity after the transient alkalinization period. As deﬁned, this calculation relies on ﬁxed stoichiometry between metabolic acid load due to lactate formation and the change measured in pHi . This will be quantitative only in a closed system (i.e., where any exchange of acid with the extracellular environment is negligible) [26,27]. In an open system where lactate and protons can exchange with the extracellular environment, this stoichiometry may no longer hold [28]. In any case, whereas lactate production is obviously the best index of glycolytic ATP production, the derived measurement presented here is proportionate with glycolytic activity and therefore is a useful index of mitochondrial capacity. It is also a necessary value to obtain if mitochondrial ATP production is the desired goal in the NMR experiment. Intracellular pH kinetics have been utilized to advance studies of mitochondrial disease in skeletal muscle, where dysfunction results in exacerbated pH changes [29–32]. 3.2. Information from Phosphocreatine Kinetics Through the equilibration of creatine kinase with its substrates and products in the cytosol, phosphocreatine is linked to the adenylate pool and thus provides information on ATP utilization and synthesis rates as well as the steady-state adenosine diphosphate (ADP) levels. Free ADP concentrations in the cytosol are below the detection limits (micromolar) and only NMR visible under extreme conditions [33,34]. However, free ADP can be calculated through the creatine kinase equilibrium from the phosphorus spectrum using additional information from either chemical measurements or simple assumptions about total creatine content [19,35–43]. Creatine kinase catalyzes the reversible phosphorylation of creatine to produce ATP from ADP. By rearrangement of equation (4), free ADP can be calculated as follows: free ADP =

[ATP][Cr] [PCr]Keq [H+ ]

(7)

The K eq value for creatine kinase is temperature, magnesium, and pH dependent and must be adjusted for these conditions before free ADP can be calculated accurately [44,45]. This assumes that the creatine kinase reaction freely equilibrates with the cytosolic adenylate pool and has the kinetic capacity to fully buffer ADP loads imposed on the cell. The evidence for this is borne out in kinetic studies using spin transfer methods [46–48] and synthetic creatine analogs [19,35] in striated muscle [19,35,42,49,50], heart [41,51–53], and liver [38]. However, it is necessary to test the equilibrium conditions in any system under investigation before free ADP values can be relied on for mechanistic studies. Initial PCr hydrolysis provides data on the nonphosphate proton buffering capacity when taken together with pHi changes, but because of the close coupling to the adenylate pool, PCr hydrolysis also measures ATP hydrolysis. The coupled

PROBING MITOCHONDRIAL FUNCTION USING

31 P-NMR

SPECTROSCOPY

565

reactions for ATPase activity and CK activity yield a summed reaction that shows a stoichiometric increase in free creatine (Cr) and inorganic phosphate during ATPase activity: ATP −→ ADP + Pi + βH+

ATPase activity: CK activity: Summed RXN:

+

αH + ADP + PCr ←→ Cr + ATP +

γH + PCr ←→ Cr + Pi

(8) (9) (10)

Representative data for a stimulation protocol are presented in Figure 5 and illustrate the initial time course of PCr hydrolysis and entering into a new steady state for a rodent hindlimb muscle. Time zero is the resting value, which decreases progressively until a new steady-state level is reached as the ATPase demands are matched by metabolic supply (glycolytic and mitochondrial ATP generation). The ATPase activity can be derived from these data by ﬁtting the time course to an exponential function and evaluating this function at time zero, shown graphically in Figure 5: dPCr JATP ase = BATP ase + (11) dt t=0 where (d PCr/dt)t = 0 is the value of the derivative of the [PCr] at time zero and BATP ase is the basal ATPase rate (rest) in units of mmol ATP/L per second. Basal ATPase rates can be estimated from oxygen consumption assuming an ATP/O2 value from the literature or measured directly from the preparation itself using an ischemic event [27]. For calculations of JATP ase in human skeletal muscle, a value of 0.008 mmol ATP/L per second has been measured [27]. The mitochondrial oxidative phosphorylation (MOP) ﬂux at any given cellular ATP turnover rate i can then be calculated by using the kinetic information already derived for ATPase and glycolytic ﬂuxes (Figures 4 and 5). At a particular steady-state i where the PCr content remains constant at a given workload, the metabolic load imposed by ATP use (JATP ase ) must necessarily be matched by ATP synthesis ﬂuxes (glycolysis and mitochondrial oxidative phosphorylation). Glycolytic ﬂux is subtracted from the overall ATP balance to derive the mitochondrial ATP synthesis ﬂux (JATP MOP ) at this particular steady state: (JATP ase )i = (JATP MOP )i − (JATP GLY )i

(12)

Having introduced the practical and conceptual aspects of cellular energetics using magnetic resonance, one can now derive from 31 P-NMR spectroscopic data sets the two indices of mitochondrial function introduced earlier in this chapter. The ﬁrst index, the Gibbs free energy of ATP hydrolysis maintained in the cytosol of the metabolically quiescent cell (e.g., skeletal muscle in resting state), is calculated from information in the 31 P-NMR spectrum according to [ATP] o GATP = GATP + RT ln (13) [ADP][Pi ]

566

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

Normalized PCr (content)

0.9

JATP mito = JATP ase - JATP Gly

0.8 0.7 0.6

STEADY-STATE

0.5 0.4 JATPase

0.3 0.2 0.1 0

100

200 300 Time (sec)

400

Figure 5 A plot of phosphocreatine kinetics over the time course of an electrically paced hindlimb muscle (same as in Figure 4) permits analysis of both ATPase and synthesis capacities. The ATPase load is derived by ﬁtting the entire progress curve for PCr from rest to the steady state and evaluating that function at time zero (instantaneous cost per unit work) [equation (11)]. The steady-state value of sustainable PCr content can now be used to derive mitochondrial ATP synthesis capacity. In steady state the ATPase load must be matched by the mitochondrial ATP synthesis capacity (JATP mito ), subtracting out any contribution of glycolytic ATP synthesis capacity (JATP Gly ), which is derived from data on pH (Figure 4).

where GoATP is typically on the order of −32 kJ/mol for mammalian tissues [54,55] and [ADP] is calculated from equation (7). This calculation requires knowledge of the total creatine and adenine nucleotide pool sizes in the particular cell type under investigation either assumed or by direct chemical measure [56]. Typically, 42 and 8 mM, respectively, have been used in studies of human skeletal muscle yielding resting G ATP estimates on the order of −65 kJ/mol [26,27]. In comparison, signiﬁcantly lower values in the range −62 to −58 kJ/mol have been found by 31 P-MRS in resting skeletal muscle of patients with inherited disorders of mitochondrial metabolism [57–61]. However, rodent muscles appear to display a remarkably wide range (−64 to −58 kJ/mol) of G ATP values in fast- and slow-twitch muscle phenotypes, respectively [19,35,62,63]. The second index of mitochondrial function, the maximal mitochondrial capacity for ATP synthesis rate or Q max [64], can be determined from dynamical 31 P NMR spectroscopic measurements during and/or following metabolic perturbation using any one of two experimental designs: single, low-metabolic load perturbations [65] or multiple perturbations covering the full range of sustainable steady states. The former yields a 31 P-NMR data set for estimation of mitochondrial capacity from PCr recovery rate [66], while the latter allows for reconstruction of the mitochondrial transfer function [67]. Both approaches are based mechanistically in feedback control of mitochondrial respiration involving a

PROBING MITOCHONDRIAL FUNCTION USING

31 P-NMR

SPECTROSCOPY

567

function of adenine nucleotide concentrations, the kinetic control model involving ADP stimulation of respiration and the nonequilibrium thermodynamic control model involving a linear dependence of respiration on G ATP [66].

3.3. Reconstruction of the Mitochondrial Transfer Function from Multiple Perturbations Chance ﬁrst described quantitative determination of the oxidative capacity of a tissue from the mitochondrial input–output or transfer function [67]. Experimentally, this particular method involves 31 P-NMR spectroscopic measurement of steady-state phosphate concentrations over the full range of sustainable steady states of oxidative ATP metabolism. From this data set, the mitochondrial transfer function can be reconstructed by correlation of the adenine nucleotide feedback signal strength calculated (i.e., [ADP] or G ATP ) with mitochondrial ATP synthesis ﬂux (or a proportional measure such as the power output of contraction in case of skeletal muscle) for each steady state studied. The nature of the transfer function is sigmoidal for both the kinetic (Figure 6) [68] and thermodynamic (Figure 7) [61,66,69] formulations. An estimate of Q max is typically obtained by nonlinear curve-ﬁtting analysis of the data points [70]. This method has been used successfully in the clinical investigation of mitochondrial function in human disease, such as mitochondrial myopathy (Figure 7) [57–61].

Figure 6 Representative plot of ADP versus JATP mito in a single healthy male subject acquired from the forearm muscles. Data were acquired according to the parameters presented in Wu et al. [61]. The solid line is a two-parameter ﬁt of Hill function to data using a second-order function [68]. The ﬁxed parameters were nH = 2, Min = −Max/10. The regression parameters: J p (mM/s) = −0.03 + 0.29([ADP]/0.028)2 /(1 + ([ADP]/0.028)2 ). ADP concentrations are reported in mmol/liter cell water and mitochondrial synthesis capacity in mmol/liter cell water per second.

568

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

Figure 7 Mitochondrial ﬂow-force relation derived from phosphorus NMR data and depicting mitochondrial function in human muscle disease. Data were acquired from eight healthy males and three complex I–deﬁcient subjects from the same family (two males, one female). Acquisition parameters are reported in Wu et al. [61]. The solid line is a three-parameter ﬁt of a Hill function to the data. Fixed parameters: Min = −Max/10. The regressions show: control; CON (circles): J p (mM/s) = −0.03 + 0.3 [1/(1 + (Gp/− 60.6)−28 )]; CID ([]): J p (mM/s) = −0.02 + 0.2[1/(1 + (G p /−56.0)−14 )]. G p in kJ/mol. The data illustrate an important biological point for mitochondrial function. Over the sustainable ATPase load imposed by muscle contraction while the ATP free energy varies over the same range, the sustainable mitochondrial ATP synthesis capacity in CID patients is almost 50% of the healthy control subjects.

3.4. Phosphocreatine Recovery Dynamics Following a Single Metabolic Perturbation In a pivotal paper, Meyer showed that the time course of phosphocreatine concentration during and following a single mild perturbation of the cellular metabolic steady state followed approximately monoexponential kinetics characterized by time constant τPCr [66]. Importantly, this parameter provided a proportional measure of the ATP synthesis capacity of the cellular mitochondrial pool Q max [66] according to τPCr = (total creatine pool) × Qmax (14) Because of this proportionality in combination with the simpler experimental design, one instead of multiple sets of serial 31 P-MRS measurements in the organ at rest during mild metabolic perturbation and subsequent recovery, this particular approach has been applied widely in clinical studies of mitochondrial function in human skeletal muscle disease, including mitochondrial myopathy [71–82], heart failure [83–85], and more recently, diabetes [86–91]. Of course, straightforward interpretation of PCr recovery kinetics in this type of clinical

REFERENCES

569

study requires that any difference in creatine pool between patients and control population is negligible. The equivalent but more robust method is analysis of the time course of the free ADP concentration [79]. However, implicit in this approach is an exact knowledge of the phenotypic creatine pool because cytosolic [ADP] is calculated through creatine kinase equilibration we have shown.

4. CONCLUSIONS Mitochondria sense and match the available ATP free energy that can be measured directly and noninvasively by 31 P-NMR spectroscopy. Together with a quantitative method to perturb the ATP free energy as described here in exercising muscle, mitochondrial (dys)function can be deduced with relative ease, providing information on mitochondrial ATP synthesis capacity (mitochondrial output function). In this chapter we provided a description of the critical parameters that need to be considered when undertaking such studies, using primarily skeletal muscle. However, the experimental approach can be applied to any organ system with a quantiﬁable physiologic perturbation of the energetic demands. Substrate metabolism, a mitochondrial input function, can be determined using this same experimental paradigm but with carbon-labeled substrates.

REFERENCES 1. Macri MA, et al. Regional changes in the metabolite proﬁle after long-term hypoxia–ischemia in brains of young and aged rats: a quantitative proton MRS study. Neurobiol Aging. 2006;27(1):98–104. 2. Bianchi MC, et al. Proton MR spectroscopy of mitochondrial diseases: analysis of brain metabolic abnormalities and their possible diagnostic relevance. Am J Neuroradiol. 2003;24(10):1958–1966. 3. Bianchi MC, et al. Brain magnetic resonance in the diagnostic evaluation of mitochondrial encephalopathies. Biosci Rep. 2007;27(1–3):69–85. 4. Gadian DG, Nuclear Magnetic Resonance and Its Application to Living Systems. New York: Oxford University Press; 1982. 5. de Graaf R. In Vivo Biological NMR Spectroscopy: Principles and Techniques. New York: Wiley; 1999: 530 pp. 6. Mansﬁeld P, Morris PG. NMR Imaging in Biomedicine. New York: Academic Press; 1982: 354 pp. 7. Bottomley PA, Ouwerkerk R. Optimum ﬂip-angles for exciting NMR with uncertain T1 values. Magn Reson Med. 1994;32(1):137–141. 8. Horska A, Spencer RG. Measurement of spin-lattice relaxation times and kinetic rate constants in rat muscle using progressive partial saturation and steady-state saturation transfer. Magn Reson Med. 1996;36(2):233–240. 9. Newcomer BR, Boska MD. T1 measurements of 31 P metabolites in resting and exercising human gastrocnemius/soleus muscle at 1.5 tesla. Magn Reson Med. 1999;41(3):486–494.

570

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

10. Ellington WR, Wiseman RW. Nuclear magnetic resonance spectroscopic techniques for study of cellular function. In Advances in Comparative and Environmental Physiology, R. Gilles, ed. New York: Springer-Verlag; 1989; pp. 76–111. 11. Brooks WM, et al. In vivo determination of 31 P spin relaxation times (T1, T2, T1 rho) in rat leg muscle: use of an off-axis solenoid coil. Magn Reson Imaging. 1986;4(3):245–250. 12. Jehenson P, et al. MR imaging as a potential diagnostic test for metabolic myopathies: importance of variations in the T2 of muscle with exercise. Am J Roentgenol. 1993;161(2):347–351. 13. Jung WI, et al. 31 P transverse relaxation times of the ATP NMR signals of human skeletal muscle in vivo. Magn Reson Med. 1992;28(2):305–310. 14. Marincek B, Friedburg H. Calculated T1 and T2 MR images and 31 P MR spectra in rabbit thigh V2 carcinoma to monitor the effect of steroid administration. Invest Radiol. 1985;20(7):701–707. 15. Hubley MJ, Locke BR, Moerland TS. The effects of temperature, pH, and magnesium on the diffusion coefﬁcient of ATP in solutions of physiological ionic strength. Biochim Biophys Acta. 1996;1291(2):115–121. 16. Hubley, MJ, Locke BR, Moerland TS. Reaction-diffusion analysis of the effects of temperature on high-energy phosphate dynamics in goldﬁsh skeletal muscle. J Exp Biol. 1997;200(Pt 6):975–988. 17. Hubley MJ, Rosanske RC, Moerland TS. Diffusion coefﬁcients of ATP and creatine phosphate in isolated muscle: pulsed gradient 31 P NMR of small biological samples. NMR Biomed. 1995;8(2):72–78. 18. Kinsey ST, Moerland TS. Metabolite diffusion in giant muscle ﬁbers of the spiny lobster Panulirus argus. J Exp Biol. 2002;205(Pt 21):3377–3386. 19. Wiseman RW, Kushmerick MJ. Creatine kinase equilibration follows solution thermodynamics in skeletal muscle: 31 P NMR studies using creatine analogs. J Biol Chem. 1995;270(21):12428–12438. 20. Moon RB, Richards JH. Determination of intracellular pH by 31 P magnetic resonance. J Biol Chem. 1973;248(20):7276–7278. 21. Roos A, Boron WF. Intracellular pH. Physiol Rev. 1981;61(2):296–434. 22. Kushmerick MJ, et al. 31 P NMR spectroscopy, chemical analysis, and free Mg2+ of rabbit bladder and uterine smooth muscle. J Biol Chem. 1986;261(31):14420–14429. 23. Wiseman RW, Beck TW, Chase PB. Effect of intracellular pH on force development depends on temperature in intact skeletal muscle from mouse. Am J Physiol. 1996;271(3 pt 1):C878–C886. 24. Wiseman RW, Ellington WR. Intracellular buffering capacity in molluscan muscle: superfused muscle versus homogenates. Physiol Zool. 1989;62(2):541–558. 25. Adams, GR, Foley JM, Meyer RA. Muscle buffer capacity estimated from pH changes during rest-to-work transitions. J Appl Physiol. 1990;69(3):968–72. 26. Blei ML, et al. Individual variation in contractile cost and recovery in a human skeletal muscle. Proc Natl Acad Sci U S A. 1993;90(15):7396–400. 27. Blei ML, Conley KE, Kushmerick MJ. Separate measures of ATP utilization and recovery in human skeletal muscle. J Physiol. 1993;465:203–22. 28. Wiseman RW, Ellington WR. Effects of extracellular pH and d-lactate efﬂux on regulation of intracellular pH during isotonic contractions in a molluscan muscle: a 31 P-nuclear magnetic resonance study. J Exp Zool. 1989;252:228–236.

REFERENCES

571

29. Chance B, et al. Phosphorus magnetic resonance spectroscopy studies of the role of mitochondria in the disease process. Ann N Y Acad Sci. 1986;488:140–153. 30. Sako EY, et al. ATP synthesis kinetics and mitochondrial function in the postischemic myocardium as studied by 31 P NMR. J Biol Chem. 1988;263(22):10600–10607. 31. Lodi R, et al. Inﬂuence of cytosolic pH on in vivo assessment of human muscle mitochondrial respiration by phosphorus magnetic resonance spectroscopy. Magma. 1997;5(2):165–171. 32. Chen JT, et al. Modeling in vivo recovery of intracellular pH in muscle to provide a novel index of proton handling: application to the diagnosis of mitochondrial myopathy. Magn Reson Med. 2001;46(5):870–878. 33. Hancock CR, et al, 31 P-NMR observation of free ADP during fatiguing, repetitive contractions of murine skeletal muscle lacking AK1. Am J Physiol Cell Physiol. 2005;288(6):C1298–C1304. 34. Fisher MJ, Dillon PF. Direct determination of ADP in hypoxic porcine carotid artery using 31 P NMR. NMR Biomed. 1988;1(3):121–126. 35. Wiseman RW, Kushmerick MJ. Phosphorus metabolite distribution in skeletal muscle: quantitative bioenergetics using creatine analogs. Mol Cell Biochem. 1997;174(1–2):23–28. 36. Dykens JA, Wiseman RW, Hardin CD. Preservation of phosphagen kinase function during transient hypoxia via enzyme abundance or resistance to oxidative inactivation. J Comp Physiol B. 1996;166(6):359–368. 37. Nioka S, et al. Age dependence of steady state mitochondrial oxidative metabolism in the in vivo hypoxic dog brain. Neurol Res. 1991;13(1):25–32. 38. Koretsky AP, et al. NMR detection of creatine kinase expressed in liver of transgenic mice: determination of free ADP levels. Proc Natl Acad Sci U S A. 1990;87(8):3112–3116. 39. Rees D, et al. In vivo functioning of creatine phosphokinase in human forearm muscle, studied by 31 P NMR saturation transfer. Magn Reson Med. 1989;9(1):39–52. 40. Chance B, et al. Multiple controls of oxidative metabolism in living tissues as studied by phosphorus magnetic resonance. Proc Natl Acad Sci U S A. 1986;83(24):9458–9462. 41. Gard JK, et al. Quantitative 31 P nuclear magnetic resonance analysis of metabolite concentrations in Langendorff-perfused rabbit hearts. Biophys J. 1985;48(5):803–813. 42. Meyer RA, Brown TR, Kushmerick MJ. Phosphorus nuclear magnetic resonance of fast- and slow-twitch muscle. Am J Physiol. 1985;248(3 Pt 1):C279–C287. 43. Keogh JM, et al. A phosphorus-31 nuclear magnetic resonance study of effects of altered thyroid state on cardiac bioenergetics. Adv Myocardiol. 1985;6:299–309. 44. Teague WE, Jr, Golding EM, Dobson GP. Adjustment of K for the creatine kinase, adenylate kinase and ATP hydrolysis equilibria to varying temperature and ionic strength. J Exp Biol. 1996;199(pt 2):509–512. 45. Goudemant JF, et al. pH and temperature effects on kinetics of creatine kinase in aqueous solution and in isovolumic perfused heart: a 31 P nuclear magnetization transfer study. NMR Biomed. 1994;7(3):101–110. 46. McFarland EW, Kushmerick MJ, Moerland TS. Activity of creatine kinase in a contracting mammalian muscle of uniform ﬁber type. Biophys J. 1994;67(5):1912–1924. 47. Martin JF, et al. Myocardial creatine kinase exchange rates and 31 P NMR relaxation rates in intact pigs. Magn Reson Med. 1989;11(1):64–72.

572

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

48. Brindle KM, Porteous R, Radda GK. A comparison of 31 P-NMR saturation transfer and isotope-exchange measurements of creatine kinase kinetics in vitro. Biochim Biophys Acta. 1984;786(1–2):18–24. 49. Kushmerick MJ, Meyer RA, Brown TR. Phosphorus NMR spectroscopy of cat biceps and soleus muscles. Adv Exp Med Biol. 1983;159:303–325. 50. Meyer RA, Sweeney HL, Kushmerick MJ. A simple analysis of the “phosphocreatine shuttle.” Am J Physiol. 1984;246(5 pt 1):C365–C377. 51. Kupriyanov VV, et al. Combination of 31 P-NMR magnetization transfer and radioisotope exchange methods for assessment of an enzyme reaction mechanism: ratedetermining steps of the creatine kinase reaction. Biochim Biophys Acta. 1990; 1020(3):290–304. 52. Katz LA, et al. Relation between phosphate metabolites and oxygen consumption of heart in vivo. Am J Physiol. 1989;256(1 pt 2):H265–H274. 53. Balaban RS, et al. Relation between work and phosphate metabolite in the in vivo paced mammalian heart. Science. 1986;232(4754):1121–1123. 54. Harkema SJ, Meyer RA. Effect of acidosis on control of respiration in skeletal muscle. Am J Physiol. 1997;272(2 pt 1):C491–C500. 55. Harkema SJ, Adams GR, Meyer RA. Acidosis has no effect on the ATP cost of contraction in cat fast- and slow-twitch skeletal muscles. Am J Physiol. 1997;272(2 pt 1):C485–C490. 56. Wiseman RW, et al. High-performance liquid chromatographic assays for free and phosphorylated derivatives of the creatine analogues beta-guanidopro pionic acid and 1-carboxy-methyl-2-iminoimidazolidine (cyclocreatine). Anal Biochem. 1992;204(2):383–389. 57. Argov Z, et al. Muscle energy metabolism in McArdle’s syndrome by in vivo phosphorus magnetic resonance spectroscopy. Neurology. 1987;37(11):1720–1724. 58. Argov Z, et al. Muscle energy metabolism in human phosphofructokinase deﬁciency as recorded by 31 P nuclear magnetic resonance spectroscopy. Ann Neurol. 1987;22(1):46–51. 59. Argov Z, et al. Bioenergetic heterogeneity of human mitochondrial myopathies: phosphorus magnetic resonance spectroscopy study. Neurology. 1987;37(2):257–262. 60. Chance B, et al. Magnetic resonance spectroscopy of normal and diseased muscles. Am J Med Genet. 1986;25(4):659–679. 61. Wu F, Jeneson JA, Beard DA. Oxidative ATP synthesis in skeletal muscle is controlled by substrate feedback. Am J Physiol Cell Physiol. 2007;292(1):C115–C124. 62. Phillips SK, et al. Neither changes in phosphorus metabolite levels nor myosin isoforms can explain the weakness in aged mouse muscle. J Physiol. 1993;463:157–167. 63. Kushmerick MJ, Moerland TS, Wiseman RW. Two classes of mammalian skeletal muscle ﬁbers distinguished by metabolite content. Adv Exp Med Biol. 1993;332:749–760. discussion 760–761. 64. Kemp GJ. Interactions of mitochondrial ATP synthesis and the creatine kinase equilibrium in skeletal muscle. J Theor Biol. 1994;170(3):239–246. 65. Slade JM, et al. A gated 31 P NMR method for the estimation of phosphocreatine recovery time and contractile ATP cost in human muscle. NMR Biomed. 2006;19(5):573–580.

REFERENCES

573

66. Meyer RA. A linear model of muscle respiration explains monoexponential phosphocreatine changes. Am J Physiol. 1988;254(4 Pt 1):C548–C553. 67. Chance B, et al. Control of oxidative metabolism and oxygen delivery in human skeletal muscle: a steady-state analysis of the work/energy cost transfer function. Proc Natl Acad Sci U S A. 1985;82(24):8384–8388. 68. Jeneson JA, et al. The signal transduction function for oxidative phosphorylation is at least second order in ADP. J Biol Chem. 1996;271(45):27995–27998. 69. Jeneson JA, et al. Quasi-linear relationship between Gibbs free energy of ATP hydrolysis and power output in human forearm muscle. Am J Physiol. 1995;268(6 pt 1):C1474–C1484. 70. Jeneson JA, Wiseman RW, Kushmerick MJ. Non-invasive quantitative 31 P MRS assay of mitochondrial function in skeletal muscle in situ. Mol Cell Biochem. 1997;174(1–2):17–22. 71. Jeppesen TD, et al. 31 P-MRS of skeletal muscle is not a sensitive diagnostic test for mitochondrial myopathy. J Neurol. 2007;254(1):29–37. 72. Trenell MI, et al. Aerobic exercise and muscle metabolism in patients with mitochondrial myopathy. Muscle Nerve. 2006;33(4):524–531. 73. Mattei JP, Bendahan D, Cozzone P. P-31 magnetic resonance spectroscopy: a tool for diagnostic purposes and pathophysiological insights in muscle diseases. Reumatismo. 2004;56(1):9–14. 74. Lodi R, et al. Reduced cytosolic acidiﬁcation during exercise suggests defective glycolytic activity in skeletal muscle of patients with Becker muscular dystrophy: an in vivo 31 P magnetic resonance spectroscopy study. Brain. 1999;122(pt 1):121–130. 75. Lodi R, et al. Normal in vivo skeletal muscle oxidative metabolism in sporadic inclusion body myositis assessed by 31 P-magnetic resonance spectroscopy. Brain. 1998;121(pt 11):2119–2126. 76. Taivassalo T, et al. Effects of aerobic training in patients with mitochondrial myopathies. Neurology. 1998;50(4):1055–1060. 77. Argov Z. Functional evaluation techniques in mitochondrial disorders. Eur Neurol. 1998;39(2):65–71. 78. Mattei JP, et al. P-31 magnetic resonance spectroscopy demonstrates unaltered muscle energy utilization in Polymyalgia rheumatica. Arthritis Rheum. 1997;40(10):1817–1822. 79. Argov Z, De Stefano N, Arnold DL. ADP recovery after a brief ischemic exercise in normal and diseased human muscle: a 31 P MRS study. NMR Biomed. 1996;9(4):165–172. 80. Gold R, et al. Phosphorus magnetic resonance spectroscopy in the evaluation of mitochondrial myopathies: results of a 6-month therapy study with coenzyme Q. Eur Neurol. 1996;36(4):191–196. 81. Taylor DJ, et al. A 31P magnetic resonance spectroscopy study of mitochondrial function in skeletal muscle of patients with Parkinson’s disease. J Neurol Sci. 1994;125(1):77–81. 82. Argov Z, Bank WJ. Phosphorus magnetic resonance spectroscopy (31 P MRS) in neuromuscular disorders. Ann Neurol. 1991;30(1):90–97. 83. Ye Y, et al. High-energy phosphate metabolism and creatine kinase in failing hearts: a new porcine model. Circulation. 2001;103(11):1570–1576.

574

NONINVASIVE ASSESSMENT OF MITOCHONDRIAL FUNCTION

84. Kemp GJ, et al. Quantitative analysis by 31 P magnetic resonance spectroscopy of abnormal mitochondrial oxidation in skeletal muscle during recovery from exercise. NMR Biomed. 1993;6(5):302–310. 85. van der Ent M, et al. A non-invasive selective assessment of type I ﬁbre mitochondrial function using 31 P NMR spectroscopy: evidence for impaired oxidative phosphorylation rate in skeletal muscle in patients with chronic heart failure. Eur Heart J. 1998;19(1):124–131. 86. Szendroedi J, et al. Muscle mitochondrial ATP synthesis and glucose transport/ phosphorylation in type 2 diabetes. PLoS Med. 2007;4(5):e154. 87. Befroy DE, et al. Impaired mitochondrial substrate oxidation in muscle of insulinresistant offspring of type 2 diabetic patients. Diabetes. 2007;56(5):1376–1381. 88. Rabol R, Boushel R, Dela F. Mitochondrial oxidative function and type 2 diabetes. Appl Physiol Nutr Metab. 2006;31(6):675–683. 89. Morino K, Petersen KF, Shulman GI. Molecular mechanisms of insulin resistance in humans and their potential links with mitochondrial dysfunction. Diabetes. 2006;55(suppl 2):S9–S15. 90. Petersen KF, et al. Mitochondrial dysfunction in the elderly: possible role in insulin resistance. Science. 2003;300(5622):1140–1142. 91. Kemp GJ, et al. Mitochondrial function and oxygen supply in normal and in chronically ischemic muscle: a combined 31 P magnetic resonance spectroscopy and near infrared spectroscopy study in vivo. J Vasc Surg. 2001;34(6):1103–1110.

25 Targeting Antioxidants to Mitochondria by Conjugation to Lipophilic Cations Michael P. Murphy MRC Dunn Human Nutrition Unit, Wellcome Trust, Cambridge, UK

1. 2. 3. 4. 5. 6. 7. 8.

Introduction ROS and drug design MitoQ and MitoE Potential toxicity Bioavailability Other approaches Pharmaceutical development of MitoQ10 Conclusions

575 576 578 579 581 582 583 583

1. INTRODUCTION “Off-target” adverse drug effects on mitochondrial function that in many cases reﬂect bioaccumulation of the xenobiotic have been described in several chapters in this book. For statins, this may be via selective plasma membrane transporters; and for biguanides, and no doubt for many other drugs, mitochondrial accumulation, via either mitochondrial membrane potential or transporters, can yield organellar concentrations that greatly exceed those in other compartments (see Chapter 17). However, the reciprocal also holds: selective mitochondrial bioaccumulation of xenobiotics offers a therapeutic opportunity. Mitochondrial oxidative damage contributes to a range of degenerative diseases. Consequently, the selective inhibition of mitochondrial oxidative Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

575

576

TARGETING ANTIOXIDANTS TO MITOCHONDRIA

damage is a promising therapeutic strategy [1]. One way to do this may be to use antioxidants that are selectively accumulated into mitochondria within patients. Such mitochondria-targeted antioxidants have been developed by conjugating the lipophilic triphenylphosphonium cation to an antioxidant moiety such as ubiquinol or α-tocopherol. These compounds pass easily through all biological membranes, including the blood–brain barrier, and into muscle cells, and thus reach those tissues most affected by mitochondrial oxidative damage. Furthermore, because of their positive charge they are accumulated several-hundredfold within mitochondria driven by the membrane potential, enhancing the protection of mitochondria from oxidative damage. These compounds protect mitochondria from damage following oral delivery and may therefore form the basis for mitochondria-protective therapies. Here we review the background and work to date on this class of mitochondria-targeted antioxidants.

2. ROS AND DRUG DESIGN In nearly all cases where mitochondrial dysfunction contributes to disease, a major cause of damage is reactive oxygen species (ROS) produced by mitochondria, either directly or as a secondary consequence of other malfunctions [2–4]. The proximal ROS is superoxide, produced by the respiratory chain, probably at complexes I and III [2–4]. Superoxide dismutates to hydrogen peroxide, which can react with ferrous iron to form the very reactive hydroxyl radical. Mitochondrial ROS cause damage to mitochondrial protein, lipid, and DNA, thereby disrupting mitochondrial function and also causing ROS to ﬂow to the cytosol [5–7]. There are a series of mitochondrial antioxidant defenses to intercept ROS and to minimize oxidative damage [8], but excessive production of ROS or disruption to the antioxidant defences leads to extensive oxidative damage to mitochondria [3]. As mitochondrial oxidative damage is either a primary cause or a signiﬁcant secondary factor leading to cell damage and death in degenerative diseases, a general therapy to decrease mitochondrial oxidative damage would be of use in a range of clinical situations [3,8,9]. However, most small molecule antioxidants will distribute around the body, with only a small fraction being taken up by the mitochondria. Pharmaceutically tractable and stable small molecule antioxidants are required which have the following properties: acceptable oral bioavailability, selective uptake by mitochondria within those organs most affected by mitochondrial oxidative damage; efﬁcient blocking of oxidative damage within mitochondria, the ability to be recycled to the active antioxidant form within mitochondria, and action as a clinically effective antioxidant at concentrations well below those that cause toxic side effects. One approach to addressing these challenges is to target antioxidants to mitochondria by conjugation to a lipophilic cation such as the triphenylphosphonium (TPP) cation [1,8,10–13]. This procedure leads to orally bioavailable molecules that accumulate into the cell driven by the plasma membrane potential

577

ROS AND DRUG DESIGN

and accumulate further into the mitochondria, where the antioxidant moiety can protect from oxidative damage and also be recycled back to its active form. Two features of lipophilic cations make them effective at delivering antioxidants to mitochondria: They can pass directly through phospholipid bilayers without requiring a speciﬁc uptake mechanism, and they accumulate substantially within mitochondria due to the large membrane potential [14–16]. Lipophilic cations can move easily through phospholipid bilayers because the activation energy for movement of lipophilic cations through the hydrophobic barrier of a biological membrane is far lower than for other cations [15–17]. During membrane transport the cations initially adsorb to the membrane in the potential energy well on the outer surface of the membrane. They then pass rapidly through the hydrophobic core of the membrane to the potential energy well on the other membrane surface, before then desorbing from the membrane. Lipophilic cations will be taken up from a positively charged compartment into a negatively charged compartment until a sufﬁciently large concentration gradient is built up to equalize the electrochemical potential of the molecules in the two compartments. At this point, compound uptake has equilibrated with the membrane potential () and the ratio of the concentrations of the free, unbound cations in the two compartments is described by the Nernst equation: =

2.303RT [cationin ] log10 F [cationout ]

As 2.303RT /F is 59.5 to 61.5 mV at 25 to 37◦ C, there will be about a 10-fold accumulation of the cation within mitochondria for about every 60 mV increase in . In addition, as the plasma membrane potential is about 30 to 60 mV (negative inside), so lipophilic cations will accumulate 5- 10-fold into the cytoplasm. As the mitochondrial membrane potential in cells is typically 140 to 180 mV [18,19], the cations within the cytosol will accumulate further several-hundredfold within mitochondria, localizing selectively within the mitochondria. The overall conclusion is that it should be possible for correctly designed antioxidants, linked to lipophilic cations, to be taken up into cells within the body and then be further accumulated into mitochondria, thus selectively protecting mitochondria against oxidative damage. Many different lipophilic cations have a sufﬁciently large hydrophobic surface area to permeate membranes and accumulate within mitochondria. In principle, any lipophilic cation could be used to direct an attached antioxidant to mitochondria. The uptake of TPP cations by energized mitochondria was introduced over three decades ago by Skulachev and co-workers to investigate the mitochondrial [20,21], and since then TPP compounds have been used routinely to measure the mitochondrial [18,19] and there are a number of synthetic chemical routes to making a range of variously substituted alkyltriphenylphosphonium compounds [12]. Therefore, we selected the TPP cation as the main delivery vector for targeting antioxidants to mitochondria.

578

TARGETING ANTIOXIDANTS TO MITOCHONDRIA

3. MITOQ AND MITOE In principle, a wide range of antioxidants could be targeted to mitochondria by conjugation to the TPP moiety. As lipid peroxidation is important in many forms of mitochondrial oxidative damage, and because the alkyl-TPP conjugates strongly associate with the mitochondrial inner membrane, the initial focus has been on antioxidants which are effective against lipid peroxidation: in particular, targeted versions of ubiquinol (MitoQ) and, to a lesser extent, α-tocopherol, the main active c25 of natural vitamin E (MitoE). The ﬁrst mitochondria-targeted antioxidant investigated was MitoE2 , which comprises the α-tocopherol moiety of vitamin E conjugated to a TPP cation by a two-carbon chain [22]. The α-tocopherol moiety is an effective chain-breaking antioxidant that prevents lipid peroxidation, with the resulting α-tocopheroxyl radical then being recycled by the endogenous mitochondrial the coenzyme Q pool [23,24]. MitoE2 was taken up rapidly by isolated mitochondria, where it was more effective at preventing lipid peroxidation than an untargeted α-tocopherol [22]. This approach was extended to targeting the ubiquinol moiety of coenzyme Q to mitochondria, as ubiquinol is known to be an effective chain-breaking antioxidant that can be recycled by the respiratory chain [24]. In this case the initial work was carried out on a molecule in which the ubiquinol moiety was linked to the TPP cation by a 10-carbon alkyl bridge to form MitoQ10 [25]. MitoQ10 is taken up rapidly by isolated mitochondria driven by the , and within mitochondria nearly all the accumulated MitoQ10 is adsorbed to the matrix surface of the inner membrane [25,26]. MitoQ10 is reduced to the active ubiquinol antioxidant by complex II in the respiratory chain, but it is not a good substrate for complex I [27] or electron transfer ﬂavoprotein–ubiqunone oxidoreductase [26]. MitoQ10 cannot restore respiration in mitochondria lacking coenzyme Q, because the reduced form of MitoQ10 is poorly oxidized by complex III; consequently, all the effects of MitoQ10 are likely to be due to the accumulation of the antioxidant ubiquinol form [27]. Furthermore, when the ubiquinol form of MitoQ10 acts as an antioxidant, it is oxidized to the ubiquinone form, which is then rapidly reduced by complex II, restoring antioxidant efﬁcacy [27]. This is important, as the recycling of an antioxidant back to its active form after it has neutralized a ROS is a critical factor in the efﬁcacy of many antioxidants [24,28–31]. As MitoQ10 is largely found adsorbed to the mitochondrial inner membrane, and its side chain enables it to penetrate deeply into the membrane core, it was anticipated to be an effective antioxidant against lipid peroxidation. This has been conﬁrmed for isolated mitochondria [25,27,31]. MitoQ10 has also been shown to detoxify peroxynitrite and it can react with superoxide, although, as with other ubiquinols, its reactivity with hydrogen peroxide is negligible [25,27]. The dependence of efﬁcacy on chain length was tested by creating and examining a series of MitoQ molecules with different numbers of carbons [32–35] in their chains. It was found that the shorter-chain MitoQ compounds were less effective antioxidants than MitoQ10 [27,31], due in part to the slower reduction to the ubiquinol form. This may be a consequence of poor access of the short-chain

POTENTIAL TOXICITY

579

analogs to the active sites of ubiquinone reductases in the respiratory chain [27]. In addition, as illustrated in Figure 3 for MitoQ10 , the antioxidant moiety can access most of the hydrophobic core of the membrane, whereas the shorter-chain compounds would have far more limited access. As yet, only short-chain MitoE analogs have been tested and have been found to be less effective than MitoQ10 [36], but the model above predicts that the longer-chain MitoE analogs currently under development should be far more effective than MitoE2 at preventing lipid peroxidation [22]. In summary, in isolated mitochondria, MitoQ10 seems to fulﬁll most of the requirements for a mitochondria-targeted antioxidant.

4. POTENTIAL TOXICITY Toxicity is the ﬁrst issue to examine in determining whether MitoQ10 or MitoE2 can act as antioxidants in cells. The extensive accumulation of lipophilic cations within isolated mitochondria at concentrations approaching millimolar levels can disrupt membrane integrity, respiration, and ATP synthesis [13,18,19,21]. These effects are thought to be largely a result of adsorption of the lipophilic cations to the matrix surface of the inner membrane, disrupting membrane permeability and affecting enzyme and transporter activity. Supporting this idea, the more hydrophobic TPP cations can disrupt mitochondrial function at lower concentrations, and the degree of disruption correlates with the amount of compound adsorbed to the inner membrane. Thus, MitoQ10 starts to increase the respiration rate of isolated mitochondria when about 2.5 to 5 µM of the compound is added to the incubation, while 5- to-10 fold higher concentrations of TPMP are required to show the same effects [25]. The nonspeciﬁc effects of MitoQ10 on mitochondria are assessed using the control compound decyl-TPP, which is similar in hydrophobicity to MitoQ10 (octanol–PBS partition coefﬁcients of 5000 [27] and 2760 [31], respectively) but lacks the antioxidant ubiquinol moiety [27]. We ﬁnd that the nonspeciﬁc mitochondrial disruption of MitoQ10 and decyl-TPP occurs at similar concentrations [37]. These nonspeciﬁc effects on mitochondrial function will always limit the amounts of TPP-derived targeted antioxidants that can be used, and it is therefore essential that the compounds be effective antioxidants at concentrations well below those that disrupt function. In moving to cells from isolated mitochondria the targeted antioxidants will be accumulated into the cytosol 5 to 10-fold relative to the extracellular environment by the plasma membrane potential [13,19,32]. Consequently, a given concentration of targeted antioxidant added to the incubation will disrupt mitochondria in cells more than for isolated mitochondria. In yeast, toxicity due to mitochondrial disruption can be assessed by observing the effects of cell growth on nonfermentable medium. Under these conditions, MitoQ10 and decyl-TPP show similar toxicity with no evident effects at 0.1 µM, but at 1 µM, toxicity is apparent and becomes severe at 10 µM. In comparison, TPMP is far less toxic (unpublished observations and [37]). For mammalian cells in culture we ﬁnd that concentrations in the range 0.1 to 1 µM generally avoid short-term toxicity (unpublished

580

TARGETING ANTIOXIDANTS TO MITOCHONDRIA

observations). However, this varies considerably with cell density, type, and incubation conditions and should be checked for all new experimental arrangements. In establishing these parameters for the non-speciﬁc toxic effects of MitoQ10 , we ﬁnd decyl-TPP to be a useful, routine control [37], although another group uses mitoundecanol [38]. Finally, in many cancer cells, expression of the plasma membrane P-glycoprotein, the product of the multidrug resistance gene (MDR-1) in the plasma membrane leads to resistance to anticancer drugs [39]. As this protein can export tetraphenylphosphonium and TPMP from cells [40,41], it might also decrease accumulation of alkyltriphenylphosphonium cations by cells. To date there is no evidence for MitoQ10 efﬂux from cells by this pathway; however, the possible role of glycoprotein-P in anomalous cation uptake should always be considered. It is well known that tetraphenylphosphonium and TPMP can be taken up into cells through the plasma membrane [18,19]. The uptake by MitoQ10 is faster than that of TPMP, presumably due to its greater hydrophobicity, lowering its activation energy for passage through the plasma membrane [25]. MitoQ10 uptake into cells is blocked largely by abolishing the mitochondrial membrane potential by the uncoupler carbonylcyanide-p-triﬂuoromethoxyphenylhydrazone (FCCP), consistent with uptake being primarily into the mitochondria and not to other cell compartments [25]. These data are consistent with rapid equilibration of MitoQ10 across the plasma membrane, followed by accumulation into mitochondria. Rapid cell subfractionation in 143B cells indicated that at least 90% of MitoE2 was in the mitochondria [22]. However, it is technically difﬁcult to conﬁrm that a lipophilic cation taken up by cells is actually located within mitochondria, as the mitochondria depolarize and release the compounds rapidly during cell subfractionation. An alternative way of visualizing TPP cations within cells is by using the 4-iodobutyltriphenylphosphonium (IBTP) cation, in which TPP is linked to an iodoalkyl system that reacts with protein thiols to form a stable thioether linkage [42]. This chemical bond prevents loss of the functionalized TPP cation on cell ﬁxation, and the location can be visualized by using TPP-speciﬁc antiserum. The results from these experiments indicate that there is almost total mitochondrial uptake within cells with very little remaining outside the mitochondria [42]. This suggests that in cells in culture, nearly all lipophilic cations accumulated are present within mitochondria, however, this approach may underestimate nonmitochondrial uptake [42]. One further indication of the membrane potential–dependent mitochondrial concentration of TPP-containing molecules within cells was the FCCP-sensitive protection afforded by MitoQ10 in a cell model of Friedreich’s ataxia, whereas FCCP did not affect the potency of decyl-Q or idebenone [36]. This is fully consistent with MitoQ10 protecting against the damage in this cell model due to its -dependent uptake into mitochondria. Therefore, there exists strong evidence that on incubation with cells in culture, MitoQ10 is accumulated predominantly within the mitochondria, but the amount of MitoQ10 present throughout the cell is currently difﬁcult to quantify.

BIOAVAILABILITY

581

5. BIOAVAILABILITY MitoQ10 and MitoE2 have been used in a large range of mitochondrial and cell models REF [1], where they show protection against oxidative damage. The interaction of MitoQ10 with mitochondrial ROS within rotenone-treated ﬁbroblasts has been studied in detail [43]. MitoQ10 did not decrease superoxide production as measured by dihydroethidium oxidation, but it did prevent lipid peroxidation as measured by the ﬂuorescent probe C11 -Bodipy [43]. This ﬁnding is consistent with the model for MitoQ10 action developed from studies with isolated mitochondria: namely, that the main antioxidant action of MitoQ10 is to prevent lipid peroxidation. It remains to be seen if this is the major mechanism by which MitoQ10 acts as a protective agent in all cell types and forms of oxidative stress. To function as therapies, mitochondria-targeted antioxidants must be delivered to mitochondria within cells in patients, preferably following oral administration. As TPP cations pass easily through phospholipid bilayers, they should be able to pass from the gut to the bloodstream and from there to most tissues. It has been shown that the TPMP cation is taken up into energized mitochondria within the perfused heart [44,45], liver [46], and skeletal muscle [47–49]. When simple alkyl-TPP compound MitoE2 or MitoQ10 is administered to mice by intravenous injection, they are cleared rapidly from the plasma and accumulate in the heart, brain, skeletal muscle, liver, and kidney [50–52]. These experiments clearly indicate that once in the bloodstream, alkyl-TPP compounds redistribute rapidly into organs. Importantly, TPP-derived compounds are orally bioavailable to mice, as was shown by feeding mice tritiated TPMP, MitoE2 , or MitoQ10 in their drinking water, which led to uptake into the plasma and from there into the heart, brain, liver, kidney, and muscle [50]. The TPMP was shown to be cleared from all organs at a similar rate by a ﬁrst-order process with a half-life of about 1.5 days [50]. Therefore, these studies are consistent with orally administered alkylTPP compounds distributing to all organs, due to their facile permeation through biological membranes. The nonspeciﬁc toxicity of alkyl-TPP cations found in mitochondria and cells will also occur in vivo, and this will probably be the major factor limiting the amounts of the compound that can be administered safely. In crude toxicity assessments [50], TPMP and MitoE2 showed no toxicity at 300 nmol intravenous (∼4 to 6 mg/kg) but did show toxicity at 500 nmol (∼6 to 10 mg/kg). MitoQ10 was marginally better tolerated, with no toxicity at 750 nmol (∼20 mg/kg) but toxicity evident at 1000 nmol (∼27 mg/kg). Administering 500 µM TPMP, MitoE2 , or MitoQ10 to mice in their drinking water could be maintained for several weeks: No toxic effects were noted for TPMP for at least 43 days, MitoE2 for at least 14 days, and MitoQ10 for at least 14 days [50]. The similarities in toxicities are consistent with the hypothesis that the toxic effects are due largely to nonspeciﬁc disruption to mitochondria caused by accumulation of large amounts

582

TARGETING ANTIOXIDANTS TO MITOCHONDRIA

of the lipophilic cation. Although the published investigations outlined above are relatively limited proof-of-principle studies, more formal toxicity determinations have also been undertaken in developing MitoQ10 for clinical trials, and these are outlined below under the pharmaceutical development of MitoQ10 . To summarize, it is possible to administer alkyltriphenylphosphonium compounds to animals orally and they are taken up into the plasma with reasonable bioavailability and then cleared rapidly from the plasma accumulating in mitochondria within tissues. Having shown that the long-term administration of mitochondria-targeted antioxidants is possible, the next step is to determine whether the amount of compound accumulated is sufﬁcient to act as an antioxidant in vivo. To date, only a few trials of this nature have been published. When 500 µM MitoQ10 was administered to rats in their drinking water for 2 weeks and the hearts then isolated and exposed to ischemia–reperfusion injury in a Langendorff perfusion system, there was protection against the loss of heart function, tissue damage, and mitochondrial function compared with TPMP or short-chain quinol controls [53]. The most probable reason for the protection observed in these experiments is that lipid peroxidation in the mitochondrial inner membrane was being prevented by MitoQ10 [54]. However, this has yet to be conﬁrmed by showing that MitoQ10 can block an increase in markers of oxidative damage in mitochondria. 6. OTHER APPROACHES Most work on mitochondria-targeted antioxidants has been on MitoE2 and MitoQ10 , which are designed to prevent lipid peroxidation. However, targeting to mitochondria by conjugation to the lipophilic TPP cation can clearly be applied to many other small, neutral antioxidants. A range of further mitochondria-targeted antioxidants has been developed. These are a mitochondria-targeted version of the SOD mimetic M40403 [55] MitoSOD, which can degrade superoxide; a mitochondria-targeted version of the peroxidase mimetic ebselen MitoPeroxidase [56]; and a mitochondria-targeted version of the nitroxide TEMPOL. The MitoTEMPOL molecule may act as a SOD mimetic and also promote the detoxiﬁcation of ferrous iron by oxidizing it to ferric iron, while simultaneously being reduced to a hydroxylamine, which can then be converted back to a nitroxide. We have also prepared a mitochondria-targeted version of the spin trap phenylbutylnitrone (MitoPBN), and this has been shown to interact with mitochondrial carbon-centered radicals [57]. Clearly, many other antioxidants can be targeted to mitochondria by this approach, which has been exploited by a number of other groups [38,57–60]. Work is continuing toward developing an understanding of the interaction of these molecules with mitochondria and determining which ones have useful properties. The ultimate aim is to create a suite of mitochondria-targeted antioxidants which can be used to intervene at several stages of the mitochondrial ROS cascade and perhaps be used in a complementary fashion.

CONCLUSIONS

583

7. PHARMACEUTICAL DEVELOPMENT OF MITOQ10 The development of MitoQ10 as a pharmaceutical is somewhat different from that of most other pharmaceuticals. Typically, in medicinal chemistry a large number of compounds are investigated that are based on a lead compound that interacts with a speciﬁc target, such as a receptor binding site. In assessing these compounds, the “rule of ﬁve” is often used as a preliminary screen to ensure that drug candidates are soluble, bioavailable, and can pass through phospholipid bilayers [61]. However, mitochondria-targeted antioxidants based on TPP lipophilic cations are less constrained by these traditional guidelines, as they have the unusual property of being both relatively water soluble and membrane permeant. Even though the molar mass of MitoQ10 is relatively large for a pharmaceutical, and it has a high octanol/PBS partition coefﬁcient [31], it is readily bioavailable and passes easily through biological membranes. A further unusual feature of the TPP-targeted compounds is that they are targeted to an organelle to interfere in a general rather than a speciﬁc process (i.e., oxidative damage). Therefore, if lipophilic cations such as MitoQ10 prove to be effective pharmaceuticals, it represents an unusual approach to medicinal chemistry and drug discovery. MitoQ10 is now being developed as a pharmaceutical [62]. For a commercially satisfactory stable formulation, it was found beneﬁcial to make the compound with the methanesulfonate counter anion and to inhibit decomposition further by complexation with β-cyclodextrin. This preparation was readily made into tablets and has passed through conventional animal toxicity with no observable adverse effect level at 10.6 mg/kg. The oral bioavailability was determined at about 10%, and major metabolites in urine are glucuronides and sulfates of the reduced hydroquinone form along with demethylated compounds. In phase 1 trials, MitoQ10 showed good pharmacokinetic behavior with oral dosing at 80 mg (1 mg/kg), resulting in a plasma C max = 33.15 ng/mL and T max ∼1 hour. This formulation has good pharmaceutical characteristics and is now in phase 2 trials for Parkinson’s disease and hepatitis C [62].

8. CONCLUSIONS The use of the TPP cation to increase the antioxidant defences of mitochondria has been demonstrated to be a viable strategy in vitro. It has also been shown that compounds such as MitoQ10 can be formed into pharmaceuticals that can successfully be delivered orally to humans. Animal experiments have indicated that MitoQ10 has antioxidant efﬁcacy in tissues, and therefore the scene is set for testing this and related compounds in human diseases. It will be important to ascertain deﬁnitively whether these chemicals are acting as effective antioxidants in vitro and whether by so doing they improve the outcome of the disease pathology. An intriguing aspect of the use of mitochondria-targeted antioxidants is that they could in principle be applied to a range of diseases and organs, because

584

TARGETING ANTIOXIDANTS TO MITOCHONDRIA

mitochondrial oxidative damage contributes to so many disorders. In addition, they could be applied to acute injuries such as ischemia–reperfusion injury during surgery, to semiacute situations such as liver damage from steatohepatitis, or to chronic degenerative disease such as Parkinson’s disease, Friedreich’s ataxia, or type II diabetes. It may even be possible to administer these compounds prophylactically. Hopefully, work over the next few years will indicate in which organs these compounds are effective, whether they can decrease mitochondrial oxidative damage in diseases, and whether this positively affects the outcome for the patient. Finally, there is considerable scope to ﬁne-tune the chemical biology of these compounds to target speciﬁc ROS and other mitochondrial targets. Acknowledgments Work in the author’s laboratory is supported by the Medical Research Council and by the European Community’s 6th Framework Programme for Research, Priority 1, Life Sciences, Genomics and Biotechnology for Health, contract LSHM-CT-2004-503116. REFERENCES 1. Murphy MP Smith RAJ. Targeting antioxidants to mitochondria by to mitochondria by conjugation to lipophilic cations. Annu Rev Pharmacol Toxicol. 2007;47:629–656. 2. Raha S, Robinson BH. Mitochondria, oxygen free radicals, disease and ageing. Trends Biochem Sci. 2000;25:502–508. 3. Finkel T. Radical medicine: treating ageing to cure disease. Nat Rev Mol Cell Biol. 2005;6:971–976. 4. Balaban RS, Nemoto S, Finkel T. Mitochondria, oxidants, and aging. Cell. 2005; 120:483–495. 5. Sawyer DT, Valentine JS. How super is superoxide? Acc Chem Res. 1981;14:393–400. 6. Vasquez-Vivar J, Kalyanaraman B, Kennedy MC. Mitochondrial aconitase is a source of hydroxyl radical: an electron spin resonance investigation. J Biol Chem. 2000;275:14064–14069. 7. Fridovich I. Superoxide anion radical (O2 −· ), superoxide dismutases, and related matters. J Biol Chem. 1997;272:18515–18517. 8. Murphy MP, Smith RAJ. Drug delivery to mitochondria: the key to mitochondrial medicine. Adv Drug Deliv Rev. 2000;41:235–250. 9. Skulachev VP. How to clean the dirtiest place in the cell: cationic antioxidants as intramitochondrial ROS scavengers. IUBMB Life. 2005;57:305–310. 10. Murphy MP. Development of lipophilic cations as therapies for disorders due to mitochondrial dysfunction. Exp Opin Biol Ther. 2001;1:753–764. 11. Smith RA, Kelso GF, Blaikie FH, et al. Using mitochondria-targeted molecules to study mitochondrial radical production and its consequences. Biochem Soc Trans. 2003;31:1295–1299. 12. Smith RA, Kelso GF, James AM, Murphy MP. Targeting coenzyme Q derivatives to mitochondria. Methods Enzymol. 2004;382:45–67.

REFERENCES

585

13. Murphy MP. Targeting bioactive compounds to mitochondria. Trends Biotechnol. 1997;15:326–330. 14. Ross MF, Kelso GF, Blaikie FH, et al. Lipophilic triphenylphosphonium cations as tools in mitochondrial bioenergetics and free radical biology. Biochemistry (Mosc). 2005;70:222–230. 15. Ketterer B, Neumcke B, Laeuger P. Transport mechanism of hydrophobic ions across through lipid bilayers. J Membr Biol. 1971;5:225–245. 16. Liberman EA, Topali VP, Tsoﬁna LM, Jasaitis AA, Skulachev VP. Mechanism of coupling of oxidative phosphorylation and the membrane potential of mitochondria. Nature. 1969;222:1076–1078. 17. Flewelling RF, Hubbell WL. Hydrophobic ion interactions with membranes. Biophys J. 1986;49:531–540. 18. Azzone GF, Pietrobon D, Zoratti M. Determination of the proton electrochemical gradient across biological membranes. Curr Top Bioenerg. 1984;13:1–77. 19. Brand MD. Measurement of mitochondrial protonmotive force. In Bioenergetics: A Practical Approach (GC Brown, CE Cooper, eds). Oxford: IRL Press; 1995: 39–62. 20. Liberman EA, Skulachev VP. Conversion of biomembrane-produced energy into electric form: IV. General discussion. Biochim Biophys Acta. 1970;216:30–42. 21. Bakeeva LE, Grinius LL, Jasaitis AA, et al. Conversion of biomembrane-produced energy into electric form: II. Intact mitochondria. Biochim Biophys Acta. 1970;216:13–21. 22. Smith RAJ, Porteous CM, Coulter CV, Murphy MP. Targeting an antioxidant to mitochondria. Eur J Biochem. 1999;263:709–716. 23. Maguire JJ, Wilson DS, Packer L. Mitochondrial electron transport-linked tocoperoxyl radical reduction. J Biol Chem. 1989;264:21462–21465. 24. James AM, Smith RA, Murphy MP. Antioxidant and prooxidant properties of mitochondrial coenzyme Q. Arch Biochem Biophys. 2004;423:47–56. 25. Kelso GF, Porteous CM, Coulter CV, et al. Selective targeting of a redox-active ubiquinone to mitochondria within cells. J Biol Chem. 2001;276:4588–4596. 26. James AM, Sharpley MS, Frerman FE, Hirst J, Smith RAJ, Murphy MP. Interaction of the mitochondria-targeted antioxidant MitoQ10 with phospholipid bialyers and ubiquinone reductases. J Biol Chem. 2007;282:14708–14718. 27. James AM, Cocheme HM, Smith RA, Murphy MP. Interactions of mitochondriatargeted and untargeted ubiquinones with the mitochondrial respiratory chain and reactive oxygen species: implications for the use of exogenous ubiquinones as therapies and experimental tools. J Biol Chem. 2005;280:21295–21312. 28. Frei B, Kim MC, Ames BN. Ubiquinol-10 is an effective lipid-soluble antioxidant at physiological concentrations. Proc Natl Acad Sci U S A. 1990;87:4879–4883. 29. Sies H. Strategies of antioxidant defense. Eur J Biochem. 1993;215:213–219. 30. Ames BN, Shigenaga MK, Hagen TM. Oxidants, antioxidants, and the degenerative diseases of aging. Proc Natl Acad Sci U S A. 1993;90:7915–7922. 31. Asin-Cayuela J, Manas AR, James AM, Smith RA, Murphy MP. Fine-tuning the hydrophobicity of a mitochondria-targeted antioxidant. FEBS Lett. 2004;571:9–16. 32. Nicholls DG, Ferguson SJ. Bioenergetics, Vol. 3. 2002. London: Academic Press; 2002. 33. Zong WX, Thompson CB. Necrotic death as a cell fate. Genes Dev. 2006;20:1–15.

586

TARGETING ANTIOXIDANTS TO MITOCHONDRIA

34. Taylor RW, Turnbull DM. Mitochondrial DNA mutations in human disease. Nat Rev Genet. 2005;6:389–402. 35. Beal MF. Mitochondria take center stage in aging and neurodegeneration. Ann. Neurol. 2005;58:495–505. 36. Jauslin ML, Meier T, Smith RAJ, Murphy MP. Mitochondria-targeted antioxidants protect Friedreich ataxia ﬁbroblasts from endogenous oxidative stress more effectively than untargeted antioxidants. FASEB J. 2003;17:1972–1974. 37. James AM, Cocheme HM, Murphy MP. Mitochondria-targeted redox probes as tools in the study of oxidative damage and ageing. Mech Ageing Dev. 2005;126:982–986. 38. Kalivendi SV, Konorev EA, Cunningham S, et al. Doxorubicin activates nuclear factor of activated T-lymphocytes and FAS ligand transcription: role of mitochondrial reactive oxygen species and calcium. Biochem J. 2005;389:527–539. 39. Gottesman MM, Pastan I. Biochemistry of the multidrug resistance mediated by the multidrug transporter. Annu Rev Biochem. 1993;62:385–427. 40. Gros P, Talbot F, Tang-Wai D, Bibi E, Kaback HR. Lipophilic cations: a group of model substrates for the multidrug-resistance transporter. Biochemistry. 1992; 31:1992–1998. 41. Lampidis TJ, Shi YF, Calderon CL, Kolonias D, Tapiero H, Savaraj N. Accumulation of simple organic cations correlates with differential cytotoxicity in multidrug-resistant and -sensitive human and rodent cells. Leukemia. 1997;11:1156–1159. 42. Lin TK, Hughes G, Muratovska A, et al. Speciﬁc modiﬁcation of mitochondrial protein thiols in response to oxidative stress: a proteomics approach. J Biol Chem. 2002;277:17048–17056. 43. Koopman WJ, Verkaart S, Visch HJ, et al. Inhibition of complex I of the electron transport chain causes O2 −· -mediated mitochondrial outgrowth. Am J Physiol. 2005;288:C1440–C1450. 44. Kauppinen R. Proton electrochemical potential of the inner mitochondrial membrane in isolated perfused rat hearts, as measured by exogenous probes. Biochim Biophys Acta. 1983;725:131–137. 45. Wan B, Doumen C, Duszynski J, Salama G, LaNoue KF. A method of determining electrical potential gradient across mitochondrial membrane in perfused rat hearts. Am J Physiol. 1993;265:H445–H452. 46. Steen H, Maring JG, Meijer DK. Differential effects of metabolic inhibitors on cellular and mitochondrial uptake of organic cations in rat liver. Biochem Pharmacol. 1993;45:809–818. 47. Rolfe DFS, Brand MD. Contribution of mitochondrial proton leak to skeletal muscle respiration and to standard metabolic rate. Am J Physiol. 1996;271:C1380–C1389. 48. Rolfe DFS, Brand MD. Proton leak and control of oxidative phosphorylation in perfused, resting rat skeletal muscle. Biochim Biophys Acta. 1996;1276:45–50. 49. Rolfe DFS, Newman JM, Buckingham JA, Clark MG, Brand MD. Contribution of mitochondrial proton leak to respiration rate in working skeletal muscle and liver and to SMR. Am J Physiol. 1999;276:C692–C699. 50. Smith RAJ, Porteous CM, Gane AM, Murphy MP. Delivery of bioactive molecules to mitochondria in vivo. Proc Natl Acad. Sci U S A. 2003;100:5407–5412. 51. Fukuda H, Syrota A, Charbonneau P, et al. Use of 11 C-triphenylmethylphosphonium for the evaluation of membrane potential in the heart by positron-emission tomography. Eur J Nucl Med. 1986;11:478–483.

REFERENCES

587

52. Srivastava PC, Hay HG, Knapp FF. Effects of alkyl and aryl substitution on the myocardial speciﬁcity of radioiodinated phosphonium, arsonium, and ammonium cations. J Med Chem. 1985;28:901–904. 53. Adlam VJ, Harrison JC, Porteous CM, et al. Targeting an antioxidant to mitochondria decreases cardiac ischemia-reperfusion injury. FASEB J. 2005;19:1088–1095. 54. Paradies G, Petrosillo G, Pistolese M, Di Venosa N, Federici A, Ruggiero FM. Decrease in mitochondrial complex I activity in ischemic/reperfused rat heart: Involvement of reactiv oxygen species and cardiolipin. Circ Res. 2004;94:53–59. 55. Salvemini D, Wang Z-Q, Zweier JL, et al. A nonpeptidyl mimic of superoxide dismutase with therapeutic activity in rats. Science. 1999;286:304–306. 56. Filipovska A, Kelso GF, Brown SE, Beer SM, Smith RA, Murphy MP. Synthesis and characterization of a triphenylphosphonium-conjugated peroxidase mimetic: insights into the interaction of ebselen with mitochondria. J Biol Chem. 2005;280:24113–24126. 57. Murphy MP, Echtay KS, Blaikie FH, et al. Superoxide activates uncoupling proteins by generating carbon-centered radicals and initiating lipid peroxidation: studies using a mitochondria-targeted spin trap derived from alpha-phenyl-n-tert-butylnitrone. J Biol Chem. 2003;278:48534–48545. 58. Dhanasekaran A, Kotamraju S, Kalivendi SV, et al. Supplementation of endothelial cells with mitochondria-targeted antioxidants inhibit peroxide-induced mitochondrial iron uptake, oxidative damage, and apoptosis. J Biol Chem. 2004;279:37575–37587. 59. Dessolin J, Schuler M, Quinart A, De Giorgi F, Ghosez L, Ichas F. Selective targeting of synthetic antioxidants to mitochondria: towards a mitochondrial medicine for neurodegenerative diseases? Eur J Pharmacol. 2002;447:155–161. 60. Dhanasekaran A, Kotamraju S, Karunakaran C, et al. Mitochondria superoxide dismutase mimetic inhibits peroxide-induced oxidative damage and apoptosis: role of mitochondrial superoxide. Free Radic Biol Med. 2005;39:567–583. 61. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev. 1997;23:3–25. 62. Antipodean Pharmaceuticals Inc. http://www.antipodeanpharma.com/.

INDEX

Abacavir (ABC), 182, 187, 244, 274–275 Absorption, 264 Acetaldehyde, 76 4-Acetamido-4’-maleimidylstilbene-2,2’-disulfonic acid (AMS), 442 Acetaminophen, 97, 99, 177–178, 299, 438, 440, 449, 541 Acetoxymethyl (AM), 425, 427 Acetyl-CoA, 176, 318 Acetyl-p-aminophenol agent, 99 Acid cycle inhibitors, 469 Actinomycins, 178 Active state, 16 Acute hepatitis, 145 Acute tubular necrosis, 292, 301, 304–305 Acyclovir, 302 Acyl chain, 20 Acyl groups, 89 Acyl-CoA: characteristics of, 73, 75, 168–169, 174 dehydrogenase, 437 Acylates, 100 Acylcarnitine, 166 Adefovir, 279, 281, 301, 303 Adenine nucleotides, 211 Adenine nucleotide translocator (ANT): animal models, 541, 544 cardiovascular toxicity, 208–209, 217 cell viability, 27 drug-associated toxicity, 103, 106 drug-induced liver injury (DILI), 174, 181–183 high-resolution respirometry, 343

mitochondrial permeability transition (MPT), 423 mitochondrial replication, 56 nephrotoxicity, 294, 296–297 synthase, 186 Adenosine diphosphate (ADP): availability, 294 characteristics of, 67, 174, 205, 345–346, 361 cystolic, 182 depletion of, 510 NMR spectroscopy, 556, 560, 563–564 phosphorylation, 10, 12, 16, 354, 376 respiration, 567 ROS production, 19 transformation into ATP, 152 Adenosine triphosphate (ATP): availability, 295 cardiac, 204 characterized, 6–7, 174, 398, 524, 556, 563 cytosolic, 14 decrease in, 511 depletion, 83, 96, 99, 100–101, 105, 107, 279 formation of, 8, 10, 82–83, 158–159 generation, 253, 258, 262, 265, 293–294, 373, 512–513 hydrolysis, 12, 15, 53–55, 295 levels, implications of, 103, 215, 242 measurements, 363 mitochondrial, 182 NMR spectroscopy, 560 production, 110, 114, 211, 218, 262, 376, 379, 398, 438–439, 478, 523, 564 reduction, 211–212

Drug-Induced Mitochondrial Dysfunction, Edited by James A. Dykens and Yvonne Will Copyright  2008 John Wiley & Sons, Inc.

589

590 Adenosine triphosphate (ATP): (Continued ) reserves, 211 synthase, see ATPase/ATP synthase synthesis, see ATP synthesis turnover, 294, 348 Adenylate: kinase, 6, 276 phosphorylation, 377 Adipocytes, 276 Adipose tissue, 16, 274, 278, 282, 480, 499 ADP/ATP: exchange, 498 translocase inhibitor, 469 Adriamycin, 30, 105, 524 ADT synthase, 343 Advanced glycation end products (ALEs), 76 Aerobic metabolism, 20 African trypanosomiasis, 101 Aged mitochondria, 29 Aging, mitochondrial theory of, 4 AIDs, 100–101, 474 Air saturation, 334–337 Alammethicin, 425 Alanine aminotransferase (ALT), 146–147 Alatroﬂoxacin, 179, 187 Albumin, 159 Alcohol intoxication, 176–177 Alcoholic liver disease (ALD), 440, 495 Alkaline phosphatase (AP), 148 Alkaloids, 236 Alkyl groups, 81 Alkylphenols, 103–104, 110 Alkyltriphenylphosphonium compounds, 577 Alleles, mitochondrial translation, 50 Allopurinol, 281 Alpers’ syndrome, 319 Alzheimer’s disease (AD), 409, 445 Amikacin, 259 Amine agents, 93 Amine groups, 111 Amineptine, 75–78, 170, 186 Amino acids, 57, 185, 404 Aminoaciduria, 301 Aminoglycoside antibiotics, 302, 304–305, 316–317 Aminoglycoside-induced deafness (AID), 316–317, 319 Aminoglycosides, 128–130, 259–260, 464–466 Aminophenol, 104–105 Aminopyrine, 494 Amiodarone, 75, 79, 80–82, 105, 111, 171–172, 186, 403 Amitriptyline, 82 Amodiaquine, 173

INDEX AMP, 6, 556, 560 Amphetamines, 520 Amphiphile drugs, 101 Amphotericin, 302 Amprenavir, 475 Amsacrine, 178 Amyotrophic lateral sclerosislike syndrome, 264 Amytal, 89–90 Anaerobic metabolism, 93, 347–348 Analgesic drugs, 83, 99, 148, 440, 541 Anemia: aplastic, 185 characteristics of, 282, 465–466 sideroblastic, 313 transient, 279 Anesthetics, 91, 93, 933, 101, 103–104, 111 Animal models: CNS injury, 262 drug safety testing, Sod2 +/− knockout mouse model, 546–550 mitochondrial abnormalities, idiosyncratic drug hepatotoxicity, 550of mitochondrial dysfunction, 540–546 mitochondrial liability endpoints, 542 signiﬁcance of, 539–540, 551 Animal toxicity, 583 Anorexia, 168 ANOVA, 403, 501 Anoxia, 18 ANT gene, 314 ANT1 knockout mice, 544 Anthracyclines, 105, 178 Anthralin, 390–391 Anthropological studies, 40 Anti-Alzheimer’s drugs. See Tacrine Antiandrogen/antiprostate cancer drug, 96 Antianginal drugs, 78–80 Antibacterial drugs, 178 Antibiotics, see speciﬁc antibiotic drugs characteristics of, 77, 96, 99–100, 107, 149, 165, 179, 185, 300, 302, 316–319, 404–407, 410, 464–466, 469–470 mitochondrial disease, 316–318 susceptibility studies, 470 toxicity, 468 Antibodies, 149, 157, 442 Anticancer drugs, 28, 30, 105–106, 178 Anticonvulsants, 93–94, 319 Antidepressant drugs, 77–78, 83, 103, 148 Antidiabetic drugs, 91, 94–95 Antiepileptic drugs, 83–84, 167 Antiestrogens, 103 Antifungal drugs, 105, 295 Antigen-presenting cells (APCs), 154–155

INDEX Anti-HIV drugs, 100, 273 Antihyperglycemic drugs, 95–96 Antihyperlipidemic drugs, 94 Antimalarial drugs, 173 Antimycin A, 18–19, 96–97, 339, 342–343, 385, 523, 531 Antineoplastic therapy, 212–219 Antioxidants: breath testing, 503 characterized, 20 cardiovascular toxicity, 219 efﬁcacy of, 578 drug-associated toxicity, 99–100, 106, 109, 113 drug-induced liver injury (DILI), 176 nervous system, 260 redox signaling and control, 436–437, 443–444, 447 targeting mitochondria by conjugation to lipophilic cations, 575–583 Antipsychotic drugs, 90–913, 148 Antipyretic drugs, 440 Antipyretics, 99 Antiretroviral drug therapy, 100, 244–245, 498–502. See also Highly active retroviral therapy (HAART) Antisense oligodeoxynucleotides, 297 Antituberculosis drugs, 106–107 Antitumor drugs, 105 Antiviral drugs: characteristics of, 105, 292, 302 current treatments, 474–475 DNA polymerase γ hypothesis, 478–480 highly active, see Highly active antiretroviral therapy (HAART) lactic acidosis, 476 lipodystrophies, 476–475 nucleotide, 304 Apaf-1, 30 Apolipoprotein B, 163 Apoproteins, 77 Apoptogen, 261 Apoptosis, triggers for: cardiovascular toxicity, 211 characteristics of, 4–5 20, 25–32 diagnosis, 478 drug-associated mitochondrial toxicity, 75, 81, 83, 88, 94, 103, 106, 108 drug-induced liver injury, 153, 163, 170, 184 lipoatrophy, 275 membrane potential and permeability transition, 424 microscopic studies, 516–518, 523 mitochondrial disease, 313

591 mitochondrial respiratory complex function, 390, 392 nephrotoxicity, 294, 296–300, 302, 304–305 redox signaling and control, 440–441, 450–452 skeletal muscle, 238, 261 Apoptosis-inducing factor (AIF), 28, 30, 300 Apoptosome, 26 Apoptotic bodies, 516 APTX gene, 314–315 Arachidonate, 74 Arachidonic acid, 150 Aromatics: amines, 101 hydrocarbons, 112 lipophilic rings, 78 Arsenate, 469 Arsenite, 524 2-Arylpropionate anti-inﬂammatory drugs, 169, 186 2-Arylpropionic acid agents, 77 Ascorbate, 341 Ascorbic acid, 20 Asp51, 99 Aspirin, 101, 167–169, 186–187 Atazanavir, 475 Atorvastatin, 243 ATP/ADP trafﬁcking, 509 ATPAF gene, 314 ATPase/ATP synthase: animal models, 544 calcium levels, 245 drug effects in patients, 312 drug-induced liver injury (DILI), 152, 158, 173, 181, 186 ﬂuorescent probe assessments, 362 high-resolution respirometry, 332, 341 historical perspectives, 4 energy coupling, 12–15 membrane potential and permeability transition, 417 microscopic studies, 512–513 mitochondrial replication, 55–56, 59 mitochondrial translation system, 48 nephrotoxicity, 296 in nervous system, 252 NMR spectrometric studies, 565–566 oxidative phosphorylation, 8 subunit structure, 13–14 ATP synthesis: antioxidants, 579 cardiovascular toxicity, 206, 214 cell viability, 16–18, 22–23 drug effects in patients, 315

592 ATP synthesis: (Continued ) drug-induced liver injury (DILI), 152–153 lipoatrophy, 274 microscopic studies, 530 nephrotoxicity, 303 NMR spectroscopy studies, 563, 565–566, 568–569 OXPHOS protein levels, 398, 409 protein synthesis, 464, 468 redox signaling and control, 445 Atractyloside, 331, 469 Attenuation, 443 Auditory nerve, 257 Autoantibodies, 157 Autoﬂuorescence, 359, 380 Autolysis, 26 Autooxidation, 73–74, 434 Autophagosomes, 31, 517, 520 Autophagy, 29, 31, 520 Autosomal dominant progressive external ophthalmoplegia (adPEO), 44 Autoxidation, 99, 332 Axonal neuropathy, 277 Azide, 99 3 -Azido-3 -dexoythymidine (AZT). See AZT AZT, 78, 100, 180–181, 220–221, 243, 278–279, 474–476, 481–482 Back-diffusion, high-resolution respirometry, 338 Background subtraction, image processing, 421 Bacteria/bacteria cells, 42–43, 48, 466, 316, 366 Bactericides, 104 Bacteriophages, mitochondrial genome, 44, 47 Bak protein, 28, 452, 517 Barbiturates, 89 Basal laminae, 303, 516 Base excision pathway, 46 Base pairing, 181, 466 bax, 107 Bax protein, 25, 28, 30, 75, 150–151, 155–156, 208, 297, 299, 305, 452, 517 Baycol, 114 bc1 complex, 48, 463 Bcl-2 protein, 25–30, 392, 517 Bcl-XL , 29, 304 BCS1 gene, 388 BCSIL gene, 314 Benoxaprofen, 102 Benzarone, 79, 173 Benzbromarone, 79, 81, 173 Benzofurans, 79–82, 105 Benzoic acids, 110, 494 β-carotene, 20 β-cyclodextrin, 583

INDEX β-oxidation, 74–78, 165–169, 170–173, 175, 186–187, 438, 513, 521, 523, 542 β-thujaplicin, 98, 105 Betulinic acid, 162, 524 Bezaﬁbrate, 84, 240, 243 Bicarbonaturia, 301 Bid protein 25, 28, 155–156, 517 Biguanide, 94–95, 319 Bik protein, 517 Bile acids, 72, 85–87 Bilirubin levels, 146 Bim protein, 75, 150–151, 517 Binding studies, 467 Bioaccumulation, 519 Bioactivation, 547 Bioavailability, 576, 581–583 Biochemical disturbances, 171 Bioenergetic studies, 37, 219–221, 293, 317, 319, 347, 377, 380, 533 Biogenesis, 42, 60–61, 74, 89, 183, 257, 468 Biopsies, 237, 244, 280, 302–303, 404, 498–499 Biothiol, 435 Biotransformation, 362–363 BLAST software, 56 Blebbing, 516–517 Blindness, 317 Blood dyscrasias, 184 BMD188, 30 Body mass index (BMI), 131 Bone marrow: depression, 465 suppression, 100, 282, 466 Bone mineral density, 280 Bongkrekic acid, 174 Bovine heart mitochondria, 399–401 Boyer, Paul D., 4 Brain: mitochondria, 252, 262 plasticity, 24 stem, 315 BrdU, 43 Breast cancer, 78, 103 Breath test, as diagnostic tool, 84, 282, 494–502 Brivudin, 281 Brown adipocytes, 510 Brown fat, 16 BSEP, 87 Budd-Chiari syndrome, 148 Buffalo hump, indications of, 477–478 Buformin, 91, 94, 114 Bulk analysis, 360 Bupivacaine, 91, 93, 98, 103, 111, 236 Buprenorphine, 79, 83, 173 Buspirone, 111

INDEX Butacaine, 101, 111 Buthionine sulfoxime (BSO), 440 14 C,

494–495 Ca2+ : accumulation, 513, 521 ATPases, 150–151 cationic amphiphile drugs (CADs), 112 cationic drug uncouplers, 101 cell death, 27, 30, 32 cystolic, 520–521 free, 150 mitochondrial permeability transition (MPT), 423–424, 450–452 mitochondrial replication, 38, 42 nephrotoxicity, 296–297 oxidative phosphorylation, 17 reactive oxygen species (ROS) production, 23–24 Calcein, 426–427 Calcineurin: characterized, 296 cytosolic, 517 -inhibitor toxicity, 296–297 Calcium: channel blockers, 237 cytosolic, 210 densities, 520–521 heart mitochondria, 205–206 homopantothenate, see Calcium hopantenate hopantenate, 170 levels, implications of, 60, 74, 150–151, 157, 280, 295, 298, 354, 451 phosphates, 516, 521 signaling, 23–25, 106, 364, 367 Calorespirometry, 348 Calpains, 24, 150 Calreticulin, 451 cAMP, 60 Cancer, 28, 184, 212, 318, 427, 530–531. See also speciﬁc types of cancers Capsaicin, 89–90 Carbamazepine, 167 Carbon, magnetic properties, 558 Carbon cyanide p-triﬂuoromethoxyphenylhydrazone (FCCP), 342–343, 366, 377–378 Carbon dioxide, 494–496 Carbon tetrachloride, 108 Carbonyl cyanide m-chlorophenylhydrazone (CCCP), 16, 19 Carbonyl cyanide p-triﬂuoromethoxyphenylhydrazone (FCCP), 16

593 Carbonyls, 76 Carboxyatractyloside, 210 Carboxydichloroﬂuorescein, 424–425 Carboxylic groups, 158 Cardevilol, 105 Cardiac ischemia and reperfusion (IR), 209–212 Cardiac muscle, 451, 498 Cardiac myocytes, 425, 427 Cardiolipin, 25–26, 29, 106, 211, 314, 439 Cardiomyocytes, 510 Cardiomyopathy, 89, 100, 217, 313, 474–475, 483, 498, 544 Cardiovascular toxicity: cardiac ischemia and reperfusion, 209–212 cardiac oxidative stress, 206–209 cardiovascular system, 203–205 cell calcium homeostasis, 205–206 xenobiotic induction, 212–221 Carnitine: characterized, 84, 170, 319, 498 palmitoyltransferase I, 72 Carticaine, 111 Carvedilol, 218 Caspases: activation, 30, 75, 88, 102, 108, 305 caspase-2, 297, 299 caspase-3, 74, 238–239, 297, 299, 304, 442–443 caspase-9, 26, 86, 155–156, 297 caspase-12, 297, 299 Caspase-activated deoxyribonuclease (CAD), 150–151 Caswell No. 392, 104 Catabolism, 187 Catalases (CATs), 21, 30, 106, 109, 444 Cataracts, 528 Cathepsin B, 75, 8, 163 Cathepsin D, 88 Cationic agents, 83 Cationic amphiphile drugs (CADs), 78–83, 111–113 Cationic drug uncouplers, 101 Cations, oxidative phosphorylation, 17 Cause-and-effect relationship, 263 CD4: lymphocytopenia, 278 T-cell depletion, 100 CD4+ lymphocytes, 499 CD8 lymphocytopenia, 278 CD8+ lymphocytes, 499 cDNA, 47, 52 Cefalotin, 100 Cefazolin, 100

594 Cell: cycle, 275 death, 20, 25–32. See also Apoptosis; Necrosis division, mitochondrial replication, 58–60 lines, 363 proliferation, 105 viability, 21, 263, 373–374, 376–377, 480 Cellular acidosis, 210 Cellular bioenergetics, see Bioenergetics Cellular integrity, 21 Central nervous system (CNS): antibiotics effects, 465 characteristics of, 252–253 injury to, 261–264 microscopic studies, 533 Cephaloridine, 97, 99–100 Cephalosporin antibiotics, 99–100 Ceriva, 238–239 Cerivastatin, 114, 237, 529 cGMP, 60 Chaperones, 53–55, 423, 520 Chemical permeabilization, 345 Chemical protonophores, 19 Chemiosmotic theory, 110 Chemopreventive antiestrogenic drugs, 78 Chemotherapeutic agents, 20, 103, 475 Chenodeoxycholic acid, 85, 87 Children: antiretroviral therapy, 278–279 cholestasis in, 86 mitochondrial diseases in, 315, 533 Reye’s syndrome, 167–168 Chloramphenicol (Cm), 108, 185–187, 316–317, 404, 406–407, 464–466, 469–470, 522, 527 Chlorcyclizine, 82 Chlorodiazepam, 163 Chloroform, 108 Chlorophenicol, 257–258 Chlorophenols, 104–105, 110 Chloroplasts, 48 Chloroquine, 79, 82–83, 173 Chlorpromazine, 79, 82, 86, 90, 92, 101, 111, 402–404 Chlortetracycline, 75 Cholangiolitis, 148 Cholestasis, drug-induced: characteristics of, 86–87, 148 prevention and therapy of, 87–88 Cholestatic hepatitis, 148, 185 Cholesterol levels, 6, 102 Cholesterol 7a-hydroxylase (CYP7A), 86 Chromatin margination, 516

INDEX Chromosomes, 40 Chronic hepatitis, 146 Chronic progressive external ophthalmoplegia (CPEO), 521 Cidofovir, 281, 301–302, 304–305 Cigliazone, 95, 438, 528 Cilastain, 305 Cinnarizine, 89, 93 Ciproﬁbrate, 92 Ciproﬂoxacin, 107, 179 Circular mtDNA, 44 Cirrhosis, 76, 166, 497 Cisplatin, 259, 295, 300, 304, 440 Citric acid cycle, 7, 10 Citrulline, 17 c-Jun N-terminal kinase, 177 Clark electrodes, 354, 374 Classical cloverleaf structure, 49 Claude, Albert, 4 Cleavage, 28, 46, 48, 178 Clinaﬂoxacin, 107 Clinical trials, see speciﬁc drugs Cloﬁbrate, 84, 92, 94 Clozapine, 90, 92 [13 C]methionine, exhalation: breath test drug-induced mitochondrial toxicity, 498–502 indications of, 282, 495–497 clinical assessment of mitochondrial function, 494–495 exhalation assays, development of, 494–495 future research directions, 502–503 Co-morbidity, 186–187 Cochlear, 129 Coding sequences, mitochondrial genome, 48–49 Codons, 49–51, 464 Coenzyme A (CoA): functions of, 72, 102, 166, 524 drug-induced sequestration, 83–84 reductase, 236, 256 Coenzyme Q (CoQ), 18, 73, 89, 281, 314, 392, 436 Coma, 187 Complex I (NADH-ubiquinone oxidoreductase), 10, 16, 18–19, 48, 57, 89–96, 175, 181, 262–263, 313–314, 317, 342, 379, 384–386, 393, 398–399, 401–404, 406, 438, 463, 467, 513–515, 520, 530 Complex II (succinate-ubiquinone oxidoreductase), 10, 49, 56–58, 93, 96–97, 313–315, 347, 385–386, 393, 398–399, 401–403, 407, 463, 513–515, 520, 530, 578

INDEX Complex III (ubiquinol-ferricytochrome c oxidoreductase), 16, 19, 48, 57–58, 73, 96–97, 99, 156, 175, 181, 252, 262, 313–315, 342, 347, 385–386, 388, 393, 403, 409, 436, 467, 513, 515, 521 Complex IV (cytochrome c oxidase), 16, 18–19, 22, 58, 73, 89, 97, 99–100, 111, 156, 258–259, 262, 313–315, 347, 386–387, 398, 400–403, 405–406, 408–409, 467, 513, 519, 521, 530 Complex V, 97, 250, 314–315, 398, 400–404 406, 528 Computer software progams: DatLab, 347–348 image analysis, 421 oxygen calibration, 336 quench-ﬂuorescent probes, 361 COMT inhibitor drugs, 104 Confocal laser scanning microscopy, 365 Confocal microscopy, 420–421, 428 Control region, 40 Convulsions, 184 Copper, 12, 21, 30, 58, 519 CoQ10 deﬁciency, 320 COQ2 gene, 314–315 Cori cycle, 277 Cotranslation, mitochondrial genome, 50–51 Councilman hyaline body, 516 Coupling factor, 12–13 COX genes, 314–315 Cox1p, 51 Cox2 inhibitor, 102 Cox2p, 51 Cox3p, 51 Creatine kinase, 205, 562–563 Creatine kinase, mitochondrial (mtCK), 208. See also Creatine kinase Creatine phosphokinase (CK), 236, 238 Creatinine levels, 280, 304 Cristae, 4–5, 57, 99. See also Cristae membrane Cristae membrane (CM), 509, 512, 521 Crystalluria, 292, 302 CSF protein, 317 C-terminal, mitochondrial replication, 55–56 Cuprizone, 108, 522 Curcumin, 173 Cyanide, 99, 259, 342–343, 531 Cyclines, 275 Cyclooxygenase, 73, 102 Cyclophilin, 296, 300 Cyclophilin D (Cycl. D), 27, 208–209, 217, 294, 423–424 Cyclosporin, 74, 96

595 Cyclosporin A (CsA), 86, 210, 216–217, 292, 295–297, 423–424, 426 CYP2C, 105 CYP2D6, 186 CYP2E1, 73–74 CYP3A, 92–93, 96 CYP450, 99 Cys-62, 448 Cysteines, 56, 439, 447, 451 Cysteinyl groups, 434 Cysteinylation, 452 Cystic ﬁbrosis, 257 Cystic ﬁbrosis transmembrane conductance regulator (CFTR), 440 Cystolic copper-zinc superoxide dismutase, 219 Cystolic signaling, 362 Cystosol, mitochondrial genome, 53 Cytochrome, 437 Cytochrome b, 463 Cytochrome bc 1 complex, 96 Cytochrome c: characterized, 10–12, 22, 25–30, 38, 54, 58, 74–75, 88, 93, 99, 102, 108, 153, 155–156, 159, 175, 211, 218–219, 258, 280, 293–295, 297, 299, 305, 385–386, 390–391, 440, 483, 521, 523 mitochondrial genome, 54 mitochondrial replication, 58 oxidase (COX), 89, 99, 111, 244, 277, 320, 405, 513 Cytochrome P450, 73, 162, 166–167 Cytokines, 72, 76, 113, 157, 172, 304, 391, 424 Cytolytic hepatitis, 147–148 Cytomegalovirus (CMV), 184, 301 Cytoplasmic blebbing, 516 Cytosol levels, 4, 6, 17, 25, 28, 30, 39, 57–59, 153, 156, 171, 174, 205, 305, 315, 318, 417, 422, 425, 427, 444 Cytotoxic drugs, 292 Cytotoxic prooxidant phenoxyl radicals, 110 Cytotoxicity, 72, 74, 79, 85–86, 96, 103, 105, 107, 109–111, 212, 359–360 Dantrolene, 92–94 DARS2 gene, 314–315 Databases, MmtDB, 40 DatLab, 347–348 db/db mice, 542 ddC, see Zalcitabine (ddC) “D drugs,” 98, 180, 182–183,2 74, 276, 278–279, 474–476, 498–499 Deafness, 128–130, 313, 470 Deafness-dystonia-optic atrophy syndrome, 56

596 Death, drug-induced liver injury (DILI), 187 Decyl-TPP, 579–580 Defective human X-linked gene (DDP1), 56 Degenerative diseases, 543, 576 Deglutathionylation, 439 Dehydrogenases, 73 Delta over base (DOB), 497 Delta p, 16–17 Demeclocycline, 75 Demerol, 89 Deoxycholic acid, 87 Deoxygenation, 356, 359 2-Deoxyglucose (2-DG), 376 Deoxyguanosine kinase, 276 Deoxyribonucleic acid (DNA), see DNA Dephosphorylation, 517 Depolarization, 209, 261–262, 418, 424 Depolymerization, 516 Depression, 184 Deprotonation, 83 Dequalinium chloride, 89 Dermal immunosuppression, 391 Detoxiﬁcation, 441 Dexamethasone, 75 Dexrazoxane, 106 ddI, see Didanosine (ddI) d4T, see Stavudine (d4T) DGUOK gene, 314 Diabetes, 75 characteristics of, 75, 212, 313, 319, 362, 522, 568 non-insulin-dependent, 479 Type 2 diabetes, 72, 109, 409, 474–475, 542–543, 584 Type 2 diabetes mellitus (T2DM), 72, 94, 130–135, 498 Diabetic glomerulopathy, 301 Diablo, 27 Dialysis, 295 Diamide, 451 Dibucaine, 101 Dicarboxylic acids, 187, 521 Dicarboxylic carrier (DIC), 441 Dichloroacetate (CDA), 318 Dichlorophenolindophenol, 385, 399 Diclofenac, 98, 102, 105, 110, 159, 162, 403–404 Dicyclohexyl-2-(-piperidyl) ethane agent perhexiline maleate, 82 Didanosine (ddI), 98, 100, 180, 182, 244, 274–276, 279–281, 474–476, 481, 498 Dideoxynucleosides, 187, 277, 527 Dienes, 84

INDEX Diet: high-fat, 84–85 ketogenic, 320–321 supplementation of, 100 Western sucrose/fat-rich, 113 withdrawn drugs, 104 Diethylaminoethoxyexestrol (DEAEH), 78–80, 171–172 Differential diagnosis, 277 Diffusion, 363 Diﬂunisal, 102, 110 Digitonin, 6, 345, 416 Dihydralazine, 149 Dihydroethidium oxidation, 581 Dihydroorotate dehydrogenase (DHODH), 274–275, 282 Dihydrorhodamine 123 (DH123), 445 Diltiazem, 101 Dimethoate, 245 Dimethylphenylacetamide agents, 103 Dinitrophenol (DNP), 16, 19, 101, 104, 331, 342, 376, 390, 469 Dinitrophenylhdyrazine (DNPH), 445 Dioxygen (O2 ), 18–19, 355 Dioxymethamphetamine, 236 Dipalmitoylphosphatidic acid (DPPA), 112 Diphenylamine: characterized, 102, 111 NSAIDs, 101 Diphenylamine-containing agents, 102–103 Diphenyleneiodonium, 99, 390 Dipyridamole, 303 Distal symmetric polyneuropathy, 281 Disulﬁde bonds, 56 Dithiol, 209 Dithiopyridines, 451 D-loop, 44, 49, 464 DNA: characterized, 106 diseases, 531–534 fragmentation, 30, 86, 89, 150, 177, 219 mutations, 128–130, 531–534 polymerase, 43, 45–46, 180 polymerase γ, 219–220, 404, 478–481 proviral, 300 replication, 178 synthesis, 178 transcription, 178 viral, 244 DNMIL gene, 315 2, 4-DNP, see Dinitrophenol (DNP) dNTP, 283 Domain V, 316 Dopamine receptor antagonists, 92

INDEX Double-blind studies, 282 Down-regulation, 78, 107 Doxorubicin (DOX), 20, 105–106, 212–219, 438, 449, 524 Doxycyclin, 57, 75 Drosophila, 59 Drp-1, 517 Drug-associated mitochondrial toxicity: fatty-liver and steatohepatitis, 72–85 hepatic cholestatic injury, 85–88 oxidative stress and tissue toxicity, 88–109 structure-activity relationships, 109–113 Drug-induced liver injury (DILI): adenine nucleotide translocator, inhibition of, 174 ATP synthase inhibition, 173 co-morbidity factors, 186 diagnosis, 148–149 features of, 145–150 frequency of, 145 idiosyncratic, 550–551 immunoallergic, 149 interference with mtDNA and/or mitochondrial transcripts, 174–185 mechanisms, 150 metabolic factors, 185–186 parent drug-mediated permeability transition, 157–163 primary impairment of mitochondrial β-oxidation: characteristics of, 163–170, 187 and respiration, 170–173 reactive metabolite-mediated mitochondrial disruption, 144–157 treatment, 149–150 Drug-induced oxidative stress: drugs independent of respiratory inhibition, 105–109 electron transport chain inhibition, 88–100 mitochondrial DNA synthesis inhibition, 100–105 prevention and therapy for, 109 Drugs: clinical trials, 146 development, 354 530 drug-drug interactions, 103 efﬁcacy, 503 metabolism, peroxidase-catalyzed, 88 metabolites, 110 safety testing, animal models, 546–551 screening, 380 testing, 345 toxicity, see Drug toxicity Drug toxicity, pharmacogenetics:

597 aminoglycosides, 129–130 deafness, 128–130 DNA mutations, 128–130 16189 mtDNA polymorphism, type 2 diabetes, 130–135 overview of, 127–128, 135–136 Dynamic spectroscopy, 559 Dynamin-related proteins (DRP/DLP1), 59–60, 315, 512 Dynein, 43, 59 Dystonia, 313 Ear: inner, 259, 465–466 mitochondrial DNA mutations, 129–130 ototoxicity, 465–466 Effector caspases, 26 Effector cells, 155 Elderly patients, research studies, 93, 145, 148–149 Electrochemical potential, 8–9 Electrogenic exchange, oxidative phosphorylation, 17 Electron, see Electron transfer; Electron transport chain acceptors, 8, 10 donors, 8, 10 Electron microscopy, 4, 12, 77, 82, 183, 277–278 Electron transfer: chain inhibitors, 252 ﬂavoprotein (ETF), 437 function (ETF), 73 system, 171, 377 Electron transport: chain (ETC), 56–57, 74, 88–89, 259, 262, 354, 378, 389, 438, 541, 544 reversed, 15–17, 19–20 system (ETS), 89, 342, 434, 468 Electroneutral exchange, 17 Electrophoresis, 101, 159, 171–172, 407–408 Electrophoretic transport, 78 Electroporation, 363 Ellipticines, 178 Elvucitabine, 474 Embryogenesis, 25 Emtricitabine (FTC), 244, 274–275, 474 End-stage renal disease (ESRD), 295 Endocytosis, 366 Endonuclease cleavage, 46 Endonuclease G, 28 Endoplasmic reticulum (ER): calcium storage, 74

598 Endoplasmic reticulum (ER): (Continued ) characterized, 23–24, 38, 53, 363, 435–436, 449–450, 520, 532 oxidative protein folding, 450–452 redox signaling, 450–452 Endoplasmic reticulum oxidase 1 (ERO1), 450–451 Endosymbiot hypothesis, 316 Endothelial cells, 295, 414 Energy coupling, 9 Energy-linked transhydrogenase (MDTH), 207 Energy production, 187 Enﬂurane, 93 Enfuviritide, 475 Entacapone, 98, 104 Enzymatic pathways, 494 Enzymatic studies, 303 Enzymes, mtDNA replication, 43–44 Eosinophils, hepatic, 162 Epidermal growth factor (EGF), 435, 442 Epilepsy, 313 Epithelial cells, 305 Erv1, 56 Erythromycins, 185, 187, 259, 316–317, 494, 522 Escherichia coli, 43, 51, 523 ESSS protein, 57 Estrogen receptor (ER), 103 ETFDH gene, 314–315 Ethambutol, 258 Ethanol, 75–76, 108, 176, 522 Ethionine, 75 Etidocaine, 103 Etoposide, 178, 524 Euglena gracilis, 52 Eukaryotes 7, 42–44, 316 European Medicines Agency (EMEA), 474 Excitotoxicity, 261–262, 521 Excretion, 264 Exercise, 277 Extracellular acidiﬁcation rate (ECAR), 374–379 Extracellular ﬂux (XF) technology, 374. See also XF24 analyzer Extrapyrimidal toxicity, 92 Eye: blindness, 317 cataract, 528 infection, 465 Leber’s hereditary optic neuropathy (LHON), 89, 252, 257, 313, 317 optic atrophy, 512 optic neuritis, 519 optic neuropathies, 259

INDEX progressive external ophtalmoplegia, 531 vision loss, 257, 317 FAD, 10 FADH, 274 Fanconi syndrome, 279–280, 301–302, 483 Farnesylpyrophosphate, 237 Fas: expression, 297 ligand, 88, 156 Fatty acid: β-oxidation, 89, 167 branched-chain, 166 characterized, 11, 187, 303 circulating levels of, 478 detoxiﬁcation, 114 as endogenous toxins in NASH, 72–75 free, 73, 114, 159, 168 hepatic, 177 long-chain, 16, 72, 75, 84, 167 medium-chain, 84 nonesteriﬁed, 17, 524 oxidation, 84, 164–166, 170, 404 polyunsaturated, 25, 28 saturated, 75 short-chain, 77, 83–84, 170 unsaturated, 20, 108 very-long-chain, 315 Fatty liver: as endogenous toxins in NASH, 72–75 ethanol-induced, 75 implications of ,103 steatosis/NASH, 75–85 Fat wasting, 277–278 Fe-S cluster, 436 Federal Food, Drug, and Cosmetic Act, 104 Feedback loop, 299 Fenoﬁbrate, 92, 94, 240 Fenoprofen, 102, 110 Ferritin, 96 Ferrous ions, 21 Fialuridine (FIAU), 183–184, 187, 243 Fibrates: characteristics of, 94, 237, 240–243 statins and, 242–243 Fibroblasts, 49, 163, 304, 344, 481, 515, 518, 533 Fibrosis, 76, 81, 113, 170, 295, 302 Filter paper disk, 468 Fission process, 9, 58–60, 257, 263, 315, 511–512, 517 FK506, 295–296, 300 Flavin mononucleotide (FMN), 10, 57, 436

INDEX Flavoprotein-ubiquinone oxidoreductase (ETF-QO), 445 Flavoproteins, 20, 347 Flow cytometry, 96, 428 Flufenamic acid, 102, 110 Flunarizine, 93 Fluorescence imaging, 59, 355, 360, 364, 421–422. 452 Fluorescence lifetime imaging microscopy (FLIM), 367 Fluorescence microscopy, 15, 418, 421 Fluorescence protocols, 348 Fluorescent dyes, 5–6, 416 Fluorometers, 359 Fluorophores, 417–420, 424–427 Fluoroquinolone agents, 107 Fluoxetine, 79, 82–83, 98, 103, 111 Fluphenazine, 90, 92 Flurbiprofen, 170 Flutamide, 91, 96, 108 Forensic studies, 40 Foscarnet, 302 Fourier transform, 558 Fractionation, 447 Frataxin, 405–406, 408 Free energy, 569 Free radicals, 11, 18, 23, 108, 207, 218, 220, 294, 434, 513, 519 Friedreich’s ataxia, 89, 584 FTase inhibitor, 238 Fumarate, 10, 19 Fungicides, 104 Fusion process, 39, 511–512 Fzo knockout, 59 Gabapentin, 319 GABA receptor, 103 Gallbladder stones, 148 Gamma (γ), see DNA polymerase γ characterized, 219–220 radiation, 28, 439 γ-Glutamyltranspeptidase (GGT), 544–545 Ganciclovir, 184, 187, 281 Ganglion cells, 258 Gas isotope ratio mass spectrometer, 496 Gatiﬂoxacin, 107 Gemﬁbrozil, 84, 240, 243, 526 Gene expression: proﬁling, 242 research studies, 127 Gene knockouts, 58 Genetic diseases, 128 Genistein, 173 Genome, nuclear, 463

599 Genomic analyses, 438 Gentamicin, 105, 259, 316–317 Gepirone, 111–112 GFM1 gene, 314 GGT-knockout mice, 544–545 GGTase inhibitors, 238 Gibbs free energy, 12, 560, 565 Gibbs potential, 8 Glibenclamide, 87 Glitazones, 438–439 Glucocorticoids, 170, 186, 236 Gluconeogenesis, 17, 165, 187 Glucose-deﬁcient cells, 448 Glucose levels, 77, 257, 305, 319, 479 Glucosuria, 301 Glucuronides, 186, 583 Glutamate, 95–96, 166, 262–263, 346–347, 439, 448, 521 Glutamyl cysteine synthetase, 87 Glutaraldehyde, 508 Glutaredoxin-1 (Grx-1), 439 Glutaredoxin-2 (Grx2), 439–440 Glutathione: characterized, 20–22, 83, 99, 150–151, 434–435, 447, 451, 544 oxidized, 207, 439–441 redox signaling, 439–441 reduced (GSH), 96, 100, 102, 105–107, 207 reductase (GRed), 160, 207 synthase, 439 Glutathione disulﬁde (GSSG), 21–22, 160, 207, 434, 439–441 Glutathione peroxidase (Gpx), 30, 176, 207, 219, 439 Glutathione peroxidase-1 (Gpx-1) knockout mice, 545 Glutathionylation, 452 GLUT4, 479 Glyburide, 87 Glycerol-3-phosphate dehydrogenase (G3PDH), 437, 513 Glycochemodeoxycholic acid, 85–87 Glycogen: levels, 521, 534 synthase, 444 Glycolic ﬂux, 373 Glycolysis, 15, 104, 204, 210, 240, 318, 376–377, 380, 513, 516, 530 Glycosylation, 49 Golgi apparatus, 77 Granulomas, hepatic, 162 Grapefruit juice, 237 Gray baby syndrome, 465 Green ﬂuorescent protein (GFP), 42, 59, 446

600 Grepaﬂoxacin, 107 Growth factors, 253 Growth hormone, 319 GSH/GSSG, 440, 447–450, 452 GSTπ pathway, 88 GTPase, 59, 512 Guanidinium group, 95 Guanosine, 182 GW-501516, 241 Hairpin loops, 46 Halobenzene, 104 Halogenated ether agents, 99 Halohydrocarbon agents, 108 Haloperidol, 89–90, 92 Halophenols, 110 Halothane, 91, 93, 149 Haptens, 93 HDL cholesterol, 94 Hearing impairment (HI), 259–260 Hearing loss, 185, 531 Heart disease, 475 Heart failure, 568 Heart mitochondria, 74, 414 Heat-shock proteins (HSPs), 28 Heavy-strand promoter (HSP), 45–46, 48 HeLa cells, 46, 364, 439, 446 Hematopoietic progenitor cells, 278 Hematopoietic toxicity, 282 Heme groups, 58, 99, 107 Hepatic cholestatic injury, drug-induced: cholestasis, characteristics of, 86–87 endogenous bile acid toxins, 85–86 prevention and therapy for, 87–88 Hepatic tissue, 499 Hepatitis: characterized, 82, 86, 147 clinical, 156–157 cytolytic, 173 fulminant, 93 immune, 156–157 metabolic alcohol, 522 viral, 499 Hepatitis B, 184, 279, 495 Hepatitis C, 184 Hepatocellular carcinoma (HCC), 72 Hepatocellular injury, 148 Hepatocyte: characteristics of, 29, 276 lysosomes, 78 mitochondria, 74 triglyceride secretion, 75 Hepatoma cells, 438 Hepatotoxic drugs, 548–550

INDEX Hepatotoxicity, 84, 93, 114, 474–475 HepG2 cells, 81–83, 95, 363, 377–379, 406, 408, 481 Herbicides, 20, 104 Herpes virostatic, 281 Heterodimers, mitochondrial genome, 47–48 Heteroplasmy, 131 Hexokinase (HK), 27, 208 High mobility group (HMG) proteins, 42, 47 High-performance liquid chromatography, 448 High-resolution fractionation, 38 High-resolution respirometry, 328–329, 332–340 High-resolution scanning electron microscopy, 510 High-throughput screening, 317 Highly active antiretroviral therapy (HAART): characteristics of, 78, 100, 192, 273–274, 276–278, 282, 300–305, 473–474 HIV-infected patients, 498–502 Histidine, 99 Histochemistry, 513, 530, 534 HIV: current treatment of, 474–475 infection, 78, 100, 180, 276–277, 281–282, 301–302, 362, 476, 483, 494, 498, 500–501 reverse transcriptase, 300, 404 transmission, 278 HIV-1, 473–474, 478, 480 HK-2, 304 [3 H]leucine, 404 HL-60 cells, 102 Homeostasis, 78, 203–206, 353, 443 HOMO-LUMO (highest occupied molecular orbital-lower unoccupied molecular orbital), 109, 111 Homogenization, 415 Homologs, 43 Homooligomerization, 26 Hot spots, 24 HSP-70, 28, 53, 300 H-strand synthesis (OH ), 44 Human blood progenitor cells, 282 Human growth factor (HGF), 304 Human immunodeﬁciency virus (HIV), see HIV Human lymphocyte mitochondria, 77 Human mitochondrial DNA database, 40 Human transcription factor, 42 Huntington’s disease, 252, 415 Hurthle cell tumors, 530 Hy’s rule, 146–147 Hydantoin agents, 93–94 Hydantoine, 92 Hyderazine, 106–108, 522

INDEX Hydrocarbon agents, 93 Hydrochloroquine, 163 Hydrogen: ion, respiratory chain, 8–9 magnetic properties, 558 peroxide (H2 O2 ), 21, 18, 20, 88, 105, 159–160, 437–438, 441–445, 447, 519, 522, 528, 576 sulﬁde, 99 Hydrogenosomes, 58 Hydrolysis: ATP, 560, 564–565 phosphocreatine, 564–565 Hydrophobicity, 111 8-Hydroxydeoxyguanosine, 183 Hydroxyl: groups, 180, 183–184, 274 radical, 20–21 Hydroxymethylglutaryl (HMG), 236, 526 3-Hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase: characterized, 320 inhibitors, 443 Hydroxyproline, 81 4-Hydroxytamoxifen, 103 Hydroxyurea, 281 Hyperglycemia, 319, 543 Hyperinsulinemia, 543 Hyperlactatemia, 277, 279, 475–476, 483, 498–501 Hyperlactemia, 467, 480 Hyperlipidemia, 72, 543 Hypertensive nephrosclerosis, 301 Hyperuricemia, 279 Hypnotics, 148 Hypogonadism, 319 Hypolipodemic drugs, 526 Hypoparathyroidism, 319 Hypothyroidism, 319 Hypoxia, 18, 61, 354, 367, 522 IAPs (inhibitors of apoptosis proteins), 26–27 Ibuprofen, 76–77, 169–170 Idebenone, 89, 90 IF-1/IF-2/IF-3, 51–52 IGF-1, 304 IMAC, 78 Imipenem, 300 Imipramine, 79, 101, 111 Imiquimod, 391–392 Immune system, 155 Immunization, 154–155, 185 Immunocytochemical analysis, 407–408, 419

601 Immunohistochemical analysis, 242, 515, 520, 533 Impaired glucose tolerance (IGT), 131–133 Impaired mitochondria, 29 Indinavir, 301, 303, 475 Indomethacin, 98, 102–103, 110 Inherited diseases, 257 Initiation: mtDNA replication, 45–46 mtDNA translation, 51 Initiator caspases, 26 Inner boundary membrane (IBM), 509, 512 Inner membrane (IM): characteristics of, 4, 6–10, 12–13 human mtDNA, 41–42 microscopy studies, 509, 513, 524 mitochondrial genome, 52 mitochondrial translation, 49, 51 mitochondrial replication, 52, 55–58 permeabilization, 426 potential, 101 Inorganic phosphate (Pi ), 560–561 iNOS/iNOS inhibitors, 75, 88, 300 Inositol 1,4,5-triphosphate receptor (InsP3 R), 450–452 Insecticides, organophosphates, 235, 245 Insertion, mitochondrial replication, 58 Insulin: production of, 130, 238 resistance, 78, 83–84, 186, 475 sensitivity, 282 Interferon-α (INF-α), 168–169, 184, 187 Interleukin (IL-8), 88 Intermembrane compartment, 4–6, 8, 11 Intermembrane space (IMS), 52 Internet, as information resource: mitomap.org, 40 123.genomics.com, 53 Intracellular signaling, 4, 23–25 Introns, 312 In vitro and in vivo studies: animal models, 539–551 antioxidants, 575–584 antiviral drugs, 473–483 assessment of mtDNA depletion, 480–481 exhalation, 493–503 ﬂuorescent probes, oxygen-sensitive, 353–367 using microscopy, 507–534 mitochondrial dysfunction assessment, 373–380 mitochondrial membrane potential, 413–428 mitochondrial respiratory complex function, 383–393 using NMR spectroscopy, 555–569

602 In vitro and in vivo studies: (Continued ) OXPHOS complex activity, 397–410 polarographic oxygen sensors, oxygraph, and high-resolution respirometry, 327–348 protein synthesis, 463–470 redox signaling, 433–453 xenobiotic-induced myopathies, 238–240 4-Iodobutyltriphenylphosphonium (IBTP) cation, 579 Ionizating radiation, 20, 28, 30 Ionization, 113 Ion transport, 10 Iproniazid, 149 Iron, 108 Iron/polyunsaturated lipid-mediated oxygen activation, 107 Iron-sulfur (Fe-S) clusters, 57 Ischemia, 25, 29, 209–212 Ischemia-reperfusion injury, 584 Isocitrate, 22 Isoﬂurane, 93, 97, 99 Isolated mitochondria: intact cells, 467–468 quality of, 468–469 respiratory control ratio, 468–469 tissue, choice of, 469–470 Isoleucine, 464 Isoniazid, 105–107 Isothermal ampliﬁcation technology, 482 Isotope-coded afﬁnity tag (ICAT), 444 J-aggregates, 418–419 Jaundice, 146 JC-1, 418–419 JNK activation, 88 JNK1, 75 Jun-N -terminal kinase (JNK), 444, 541 Jurkat cells, 365–366, 438 K+ , 17 Kanamycin, 259, 316 Karryolysis, 516 Karryorrhexis, 516 KCN, 385–386, 401, 487 Kearns-Sayre syndrome, 317, 319–321 Keratinocytes, 391 Ketoconazole, 86 α-Ketobutyrate decarboxylase, 495 α-Ketoglutarate dehydrogenase (KGDH), 437 α-Ketoisocaproic acid (KIKA), 494–495 Ketoprofen, 169 Kidney: characterized, 279 chronic disease, 292

INDEX malfunction, 483 renal disease, 301 renal dysfunction, 100 renal failure, 280, 292, 301, 304 renal replacement therapy, long-term, 304 renal tissue, 499 renal toxicity, 465 KIKA breath test, 494–495 Kinesin, 43, 59 Kjer disease, 257 KKAy mice, 54 Krebs cycle, 7, 10, 20, 42, 100, 107, 206, 398, 510–511, 521 Krebs, Hans, 4 Kruskall-Wallis test, 501 Kupffer cells, 76–77, 157, 176, 414 Lactate: dehydrogenase, 83 levels, 277, 279–280 production, 282 Lactic acidemia, 315 Lactic acidosis, 94–95, 100, 104, 107, 128, 175, 183, 185, 187, 244, 279, 281, 303, 319, 466, 474–476, 498–499, 543 Lamivudine (3TC), 180, 182–183, 244, 274–275, 474 Lamotrigine, 319 Langendorff perfusion system, 582 Langerhan cells, 391 Large-molecular-weight compounds, 6 l-arginine, 320 Laser scanning cytometry, 513 Laser scanning microscopy, 420 Lateral ﬂow dipsticks, 397–410 Lattice: intramitochondrial, 530 spin physics, 558 Lazeroid, 85 l-carnitine, 281, 502, 524 LDL cholesterol, 94 Leaky mitochondria, 468 Leber’s hereditary optic neuropathy (LHON), 89, 252, 257, 313, 317 Legislation, Federal Food, Drug, and Cosmetic Act, 104 Leigh’s syndrome (LS), 313, 513, 515 Leishmaniasis, 101 Lesions: liver, 148, 524 mitochondrial, 184, 186 Leucopoenia, 282 Leukoencephalography, 315 Levoﬂoxacin, 107

603

INDEX Lidocaine, 91, 93, 111 Lifetime-based imaging, 366–367 Ligand binding, 89 Light microscopy, 5, 39, 183, 516 Light-strand promoter (LSP), 45–46, 48 Lincosamides, 464, 466–467 Linezolid, 185, 317–318 404, 407–408, 466–467, 469–470 Lipid(s): bilayer, 55, 58–59, 103, 171 functions of, 20, 211, 529 metabolic dysfunction, 498 mitochondrial, 176–177 oxidation, 438 peroxidation, 74–77, 84, 94, 96, 100, 107, 165, 170, 172, 177, 187, 299–300, 578 Lipidosis, lysosomal, 78 Lipoapoptosis, 75 Lipoatrophy, 278, 281, 478–479, 483 Lipodystrophies, 273, 277–280, 303, 474–477, 499 Lipodystrophy syndrome, 78 Lipohypertrophy, 478 Lipohypotrophy, 483 Lipolysis, 73, 168 Lipomas, cervical, 317 Lipophilic cations, targeting mitochondria by conjugation to: bioavailability, 581–583 MitoE, 578–579 MitoE2 , 578–579 MitoQ, 578–583 overview of, 575–576 potential toxicity, 579–580 ROS and drug design, 576–577 Lipophilicity, 110 Liposomal transfer, 363, 365 Lipoxygenases, 73 Liquid scintillation counting, 468 Lithocolic acid, 87, 242 Liver: alcohol-related disease, 528 cystolic metabolism, 494 diseases, 80, 582 dysfunction, 318 failure, 187 fat removal from, 163 fatty, 187 injury, 109. See also Drug-induced liver injury (DILI) mitochondria, 4–5, 13, 17, 96 necrosis, 93 toxicity, 465

Long-chain-3-hdyroxyacyl-CoA dehydrogenase (LCHAD), 169 Lonidamide, 162, 524 Loop diuretics, 259 Lovastatin, 237 Low-molecular-weight compounds, 6, 20, 27 LRPPRC gene, 314 Lungs, 78 Lymphoblastoma cells, 344 Lymphocytes, functions of, 212, 280. See also speciﬁc types of lymphocytes Lymphomas, 212 Lymphopenia, 278–279 Lypodystrophy syndrome, 498 Lysosomal lipoapoptosis, 75 Lysosomes, 29, 171, 304–305, 425 Lysosomotropic toxins, 88 Macrolide antibiotics, 316 Macrolides, 464, 466–467 Macrophages, 78, 82, 300, 366, 478 Macrophagy, 520 Macrovascuolar steatosis, 78, 165, 497, 524 Magnetic beads, 15 Magnetic resonance imaging (MRI), 556 Major histocompatibility (MHC) class I/II molecules, 154–156 Malate, 17, 22, 95–96, 166, 346–347 MALDI-TOF analysis, 445 Malic enzyme, 22 Mallory bodies, 76, 80 Malonate, 97 Malondialdehyde, 84 Malonyl CoA, 78 Mammalian cells, 59, 179, 366 Mammalian mitochondria: gene products in, 314 mtDNA in, 39–42, 46 translation system, 49 Mammals, neonatal, 16 Manganese, 21 Manganese superoxide dismutase (MnSOD), 160, 519, 544 Maraviroc, 475 Mas37, 55 Mass spectrometry, 39, 244, 496 Mass spectroscopy, 38–39 Maternal inheritance, mtDNA mutations, 128–135 Matrix chaperones, 55 Matrix compartment, 4, 6–8, 11, 13, 20 Matrix proteins, 54, 152 MCT-1, 245 Mefenamic acid, 102, 110

604 Megamitochondria (Mgs): characterized, 185, 522 formation: drug radical-induced, 108 free radicals, 523 Melatonin, 176 Membrane blebbing, 517 Membrane permeabilization, see Membrane permeability transition (MPT) characterized, 88 ROS-induced mitochondrial, 440 Membrane potential, signiﬁcance of, 17, 19–20, 24, 27, 416–420, 428, 523, 528, 544. See also Membrane potential, measurement using ﬂuorescent reporters Membrane potential, measurement using ﬂuorescent reporters: ﬂuorescent probes, Nernstian distribution of, 416 isolated mitochondria: characteristics of, 414–416 monitoring, 417–420 single intact cells, imaging, 420–423 overview, 414, 577 plasma membrane, 428 Meperidine, 89 MERRF (myoclonus epilepsy with ragged-red ﬁbers), 128, 313, 317, 533 Messenger RNA (mRNAs): characterized, 40, 183 mitochondrial genome, 40–41, 45–47, 49–50, 53 mitochondrial replication, 58 Metabolic disease, 522–523 Metabolism, signiﬁcance of, 264, 313 Metabolite transport, 10 Metal ions, 56 Metalloporphyrins, phosphorescent, 355 Metalloprotease, 53 Metamorphosis, 25 Metazoans, 5, 44 Metformin, 91, 94–95, 109, 281, 319, 402 Methanol, 258–259 Methionine, 464 Methods of Enzymology, 54 Methylation, 47 5-Methylchloroﬂuorescein (CMF-DA), 446 3,4-Methylenedioxymethamphetamine (MDMA), 244–245 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine), see MPTP 1-Methyl-4-phenylpyridinium, 520 Methylgloxal bis(guanine hydrazone), 108, 179 4-Methylpyrazole, 176

INDEX Methyl thiazol tetrazolium (MTT) assay, 389 Mfn-2, 517 Mg2+ , 424 Mia40, 56 Michaelis, Leonor, 4 Microarray analysis, 127 Microcalorimetry, 348 Microdeletions, 312 Microscopy, see speciﬁc types of microscopy background subtraction, 421 calculation of , 422–423 image acquisition and processing, 420–421 mitochondrial dysfunction studies, see Microscopy studies, mitochondrial dysfunction assessment Microscopy studies, mitochondrial dysfunction assessment: mitochondriopathies: autophagy, 520 calcium densities, 520–521 cancer, 530–531 DNA mutations and diseases, 531–534 drug-induced, 523–530 glycogen and fat deposition, 521 metabolic disease, 522–523 mitophagy, 520 oxidative stress, 518–520 overview, 508, 534 Microsomal triglyceride transfer protein (MTTP), 163, 166 Microsomes, anti-liver, 149 Microvesicular steatosis, 75, 77–78, 80, 165, 169, 175–178, 186–187, 280, 497, 524–525. See also Fatty liver Migraine headache, 320 Mild toxicity, in hepatitis, 154 Mitchell, Peter, 4, 8 Mitochondria, generally: cell death and, 25–32, 513–518 compartments, 4, 6–7 complexity of, 136 ﬁssion and fusion, 511–512 functions of, 38 historical perspectives, 3–4, 38 normal morphology, 4–6, 508–511 OXPHOS and assays, 512–513 replication of, see Replication structure, 6–7 tightly coupled, 18 Mitochondria-drug interactions, 74 Mitochondrial diseases: drug toxicity: direct effects, 316–318 hypothetical effects, 319–320

INDEX indirect effects, 318 features of, 4, 48, 311–312 mtDNA mutations, 312–313 multisystem disorders, 465 nDNA mutations, 313–316 potential therapies, 320–321 Mitochondrial DNA (mtDNA): characteristics of, 7 defects in, 4 degradation: by acetaminophen, 177–178 by alcohol, 176–177 depletion, 182, 276–280, 282, 303–304, 478, 480–481, 498–499 interference with, 174–185 maintenance of, 314 measurement technologies, 481–483 mutations, 183, 245–246, 260, 312–316, 320, 410, 533 polymorphisms, 130–135 reactive oxygen species (ROS), see Reactive oxygen species (ROS) repair of, 46 replication of, 42–46, 180–184, 187, 264, 313, 468, 498, 526 resynthesis, 178 synthesis, 98, 100, 172, 527 transcription, 313 translation, 313–314 Mitochondrial dysfunction, quantitative assessment of extracellular ﬂux of oxygen and protons, 373–376 Mitochondrial encephalopathy with lactic acidosis and stroke-like episodes (MELAs), 89, 252, 260, 313, 318, 320, 531 Mitochondrial form of superoxide dismutase (MnSOD), 106, 437, 443–444, 545 Mitochondrial function, quantitative assessment: overview, 376–377 mitochondrial respiration: inhibitors of, 378–380 uncouplers of, 377–378 Mitochondrial function analysis: future directions for, 367 high-resolution respirometry: intact cells, 340–347 isolated mitochondria, 346 multisensor applications, 347–348 Oxygraph-2k, 333–340, 347 permeabilized cells and tissues, 345–346 phosphorylation control protocol (PC protocol), 340–346 polarographic oxygen sensor, 329–333 titration protocols, 346–347

605 traditional oxygraphy, 329–333 overview of, 328–329 oxygen-sensing probes, 360–367 quantitative, see Mitochondrial function, quantitative assessment quenched-ﬂuorescence oxygen sensing, 355–360 Mitochondrial genome: mitochondrial translation system, 48–52 mtDNA in mammal mitochondria, 40–42 oxidative damage in, 38 replication of mtDNA, 42–46 transcription, 46–48 Mitochondrial import protein (MIP), 53–54 Mitochondrial manganese superoxed dismutase (MnSOD), 207 Mitochondrial matrix, 58–59 Mitochondrial medicine, 4 Mitochondrial network, 5–6 Mitochondrial oxidative phosphorylation (MOP), 101–105, 565–566 Mitochondrial permeability transition (MPT): carboxydichloroﬂuorescein release, 424–425 cardiac oxidative stress, 206–209, 211 drug-associated toxicity, 86, 88, 99, 102, 114–115 drug-induced liver injury (DILI), 151–153, 156, 166–167, 174, 179, 186 models of, 423–424 nephrotoxicity, 294, 295–298 plasma membrane permeability, 428 skeletal muscle, 240 swelling assay, 424 visualizing in intact cells, 425–427 Mitochondrial processing peptidase (MPP), 53, 541, 547 Mitochondrial protein synthesis, assessment in drug toxicity screening: antibiotics, 464–466 effects of drugs, 447 isolated mitochondria: intact, 467–468 quality and respiratory control ratio, 468–469 tissue, choice of, 469–470 overview, 463–464 Mitochondrial protonmotive force (p), 8–10, 12, 19, 22 Mitochondrial respiration inhibition, cationic amphiphilic drug-induced, 78–83 Mitochondrial respiratory complex function assessment: anthrallin, 390–391 imiquimod, 391–392

606 Mitochondrial respiratory complex function assessment: (Continued ) MitoQ, 393 photodynamic therapy, 390 resveratrol, 392–393 spectrophotometric measurement: assay limitations, 388–389 complex I (NADH-ubiquinone oxidoreductase activity), 384–385 complex II (succinate-ubiquinone oxidoreductase activity), 385 complex III (ubiquinol-ferricytochrome c oxidoreductase), 385–386 complex IV (cytochrome c oxidase), 386–387 drug-mediated toxicity, 389–393 succinate-cytochrome c reductase, 388 succinate-hexacyanoferrate reductase, 387 ultraviolet radiation, 392 Mitochondrial toxicity: autooxidation and, 73–74 drug-associated, 86, 88, 99, 102, 114–115, 281 monitoring and prediction of, 280 by perioxisomal ROS, 75 therapy, 281–283 Mitochondrial transcription factor A (mtTFA), 42, 47–48, 60, 183–184, 264 Mitochondrial transcription factor B (mtTFB), 47–48, 50 Mitochondrial transcription termination factor (mTERF), 48 Mitochondrial transfer function, 567–568 Mitochondrial transmembrane potential, 299 Mitochondriology, 37 Mitochondriopathies, drug-induced: alcohol, 524 anticancer drugs, 524 HAART, 526 microvesicular steatosis, 524–525 overview of, 523 PPARγ agonists, 528, 530 statins, 526, 528 Mitocondrial myopathy, 567–568 MitoE, 578–579 MitoE2 , 578–579, 581 Mitofusin (Mfn), 512 Mitofusin-1, 107 Mitomycin C, 105–106 Mitophagy, 520 Mitoplasts, 6 MitoQ, 383, 578–579 MitoQ10 : characteristics of, 578–582 pharmaceutical development of, 583

INDEX Mitoribosomes, 315 MitoSOD, 582 MitoSOX, 445, 550 MitoTracker, 419–420 MIV310, 474 Mixed acute hepatitis, 148 Mixed hepatitis, 185 MKK4, 88 Molecular connectivity index, 111 Molecular genetic studies, 38 Molecular oxygen (O2 ), 8, 11, 450 Monocarboxylate transporter (MCT4), 245, 443 Monoclonal antibodies (mAbs), 405 Monodansylcadaverine, 520 Motor neuron disease, 313 Motor proteins, 43 Mouse studies: fatty liver, 77 liver mitochondria, 77 ROS production, 21 MPP+ , 91 MPTP, 90–92, 440, 442, 450, 452, 541, 547 MPV17 gene, 314–315 MRPS16 gene, 314 MSF (mitochondrial import stimulating factor), 54 Multidrug resistance gene (MDR-1), 480 Multidrug resistance proteins (MRPs), 293, 303 Multiphoton microscopy, 420, 428 Multiple myeloma, 438 Muscle mitochondria, 4 Mutations, see speciﬁc genes mitochondrial genome, 47–48 mitochondrial replication, 56 mitochondrial translation, 50 MWFE protein, 57 Myalgias, 184 Mycoplasma, 52 Myoblasts, 204 Myocardium, 205, 531 Myocyte cell death, 211 Myoﬁber: C-IV, 534 degeneration, 526 Myoﬁbrils, 414 Myopathy: characterized, 75, 183, 303, 410, 465, 475, 498 congenital, 245–246 metabolic, 279 mitochondrial, 277 skeletal muscle, 236–245

INDEX Myosin, 59 Mytoxicity, ﬁbrate-induced, 240 Myxothiazol, 99, 388, 390 N -acetyl-p-benzoquinonimine (NAPQI), 99, 438, 541 NAD+ , 175–176, 210 NADH: characterized, 10–11, 19–20, 76, 96, 160–161, 175–176, 205, 274, 354, 384–385, 514, 534 dehydrogenase, 89, 95, 215 levels, 93 oxidation, 257 NADH-cytochrome c reductase, 240 NADP, 160 NADPH, 22, 88, 107, 160, 162, 441 Nafazodone, 87 Nagarse, 414 Nalidixic acid, 179 Nanomotor, 14 Naproxen, 102, 110, 169 NASH, 84–85, 100, 103, 108, 113, 529–530 Native proteins, 55 NDMA receptor, 258 NDUFA gene, 314 NDUFS genes, 314 Necrosis: avoidance strategies, 153 hepatocyte, 76, 83 liver cells, 78, 177 triggers for, 20, 29, 31, 81, 74, 77, 88, 93, 99, 103, 107, 166, 170, 211, 424, 513–516, 521 Nefazodone, 402–403 Negative staining, 12–13 Neomycin, 259, 316 Nephritis, interstitial, 302 Nephrolithiasis, 292, 302 Nephrotoxicity: antibiotic therapies, 304 calcineurin inhibitor, 295–300 characteristics of, 292, 304 future research directions, 305 glucose and, 292, 305 HAART and, 300–304 respiratory chain, 293–295 tubular cells, 292–293 Nephrotoxins, 299 Nernst equation, 416, 577 Nerves, auditory, 128–130 Nervous system: CNS injury, 261–264 ototoxicity, 259–260

607 overview of, 251–253 parasympathetic, 478 peripheral neuropathy, 185, 253–257, 303, 318, 466, 474–475, 483, 498–499 retinal drug toxicity, 257–259 N -ethylmaleimides, 451 Neuroblastoma, 438 Neurodegenerative disease, 25, 252 Neurodegenerative disorders, 29 Neuroleptics, 90–93 Neuropathy, peripheral, see Peripheral neuropathy Neuropsychiatric disorders, 475 Neurospora crassa, 57 Neurotoxicity, 93 Neurotoxins, 253, 261, 265, 520 Neutropenia, 279 Neutrophils, 76, 80, 102 Nevirapine, 475 Niacin, 236–237 Nicotinamide adenine dinucleotide (NADH), see NADH Nicotinamide nucleotides, 10 Nilutamide, 108 Nimesulide, 98, 102, 108, 159 Nitric oxide (NO): implications of, 262–263, 294–295, 320, 348, 519–520 synthase (NOS): characterized, 206, 296 endothelial (eNOS), 60 mitochondrial isoform of nNOS (mtNOS), 294 Nitro Kleenup, 104 Nitroaromatics: agents, 96, 102 drug radicals, 108–109 Nitrocatechol agents, 104 Nitrofurantoin, 108 Nitrogen: clearance, 494 dioxide, 88 magnetic properties, 558 Nitroglycerin, 320 Nitrophenols, 104, 109, 294, 390 3-Nitropropionate, 415 Nitrosic stress, 294 Nonalcoholic steatohepatitis (NASH): breath tests, 495 deﬁned, 72 drug-induced, 75–85 fatty acids and, 72–75 Non-insulin-dependent diabetes mellitus (NIDDM), 319

608 Nonmelanoma skin cancer (NMSC), 391–392 Nonnucleoside reverse transcriptase inhibitor (NNRTIs), 474–475 Nonsteroidal anti-inﬂammatory drugs (NSAIDs): with 2-arylpropionate structure, 169–170 characteristics of, 77, 101, 110–111, 186, 147–148, 159, 259 drug-induced liver injury, 186 nephrotoxicity, 303 Norﬂuoxetine, 103 Normalization protocols, 376 Northern blotting, 481 Novobiocine, 178 NOXA, 150–151 Nrf2/Keap1, 451 Nrf2 pathway, 74 NRK-52E, 441 N-terminal, 53 Nuclear DNA (nDNA): characterized, 513, 515, 531 deﬁned, 312 mutations, 313–316, 541, 543–546 replication, 43 Nuclear factor of activated T cells (NFAT), 296 Nuclear localized d-amino acid oxidase (NLS-DAO), 448 Nuclear magnetic resonance spectroscopy: biological, 559 characteristics of, 494, 555–557 phosphorus, 561 probing mitochondrial function using 31 P-NMR, 561–569 spin physics, 557–561 Nuclear protein, 58 Nuclear regulatory factors, 60 Nuclear respiratory factors (NRF-1/NRF-2), 60 Nuclear targeted d-amino acid oxidase (NLS-DAAO), 446, 449 Nucleic acid sequence base ampliﬁcation (NASBA) protocol, 482 Nucleoid formation, 41 NucleomaxX, 282 Nucleoside reverse transcriptase inhibitors (NRTIs): antiviral therapies, 481 characteristics of, 100, 212, 219–220, 235, 243–245, 273, 474–475, 478, 498 clinically licensed, 274 development of, 474 hepatotoxicity, 526 mtDNA damage, 503 OXPHOS activity assays, 410 toxicity, pathogenesis of, 219–221, 274–276 triphosphates, 274

INDEX Nucleotide analog reverse transcriptase inhibitor, 279 Nucleotides, 44, 49 Nutrition, signiﬁcance of, 72. See also Diet ob/ob mice, 542 Obesity, 72, 186–187, 543 Obstructive renal disease, 292 O-dealkylation, 105 Oleate, 73 Oligomers, 405 Oligomycin, 97, 263, 342, 469, 523 Omp85 protein, 55 OM protein, 511–512 Oncocytomas, 530, 532 16189 mutation, 130–135 Opa1 protein, 517 Opthalmoplegia, sporadic progressive external, 321 Organ drug toxicity: cardiovascular toxicity, 203–221 drug-induced liver injury, 143–187 lipoatrophy, 273–283 mitochondrial disease, 311–321 nephrotoxicity, 291–305 skeletal muscle, 235–246 Organic acid transporter (OAT), 293, 303–304 OROBoPOS, 334, 337 Oscillations, calcium signaling, 25 Osmosis, 28–29 Osmotic shock, 363 Osteomalacia, 279 Ototoxicity, 465–466 OUMS-29 cells, 528 Outer membrane (OM): characteristics of, 4, 6–8, 509 mitochondrial genome, 52–53 mitochondrial replication, 54–55 permeabilization, 299 rupture of, 152, 156, 161, 208 Oxaloacetate, 97 Oxazolidinones, 316, 464, 466 Oxidation, formic acid, 259 Oxidative pathology, 22 Oxidative phosphorylation (OXPHOS): antiretroviral therapeutics, 274 cardiovascular toxicity, 204 complex activity assays, see OXPHOS activity assays conformational model, 15 control of, 60 dipstick immunoassays for protein level assessment, 404–408 drug-associated toxicity, 94

609

INDEX drug effects, 311–313 efﬁciency of, 23 ﬂuorescent probe studies, 365–366 general principles of, 7–8 high-resolution respirometry, 338 historical perspectives, 4 impairment, 85, 187 inhibiting, 3898 microscopy studies, 509, 521, 524, 526, 530, 533 mitochondrial ATPase/ATP synthase, energy coupling, 12–15 mitochondrial carriers, 17 mitochondrial function analysis, 343, 345 mitochondrial genome, 52 mitochondrial replication, 39 nephrotoxicity, 297, 303 nervous system and, 252, 262 respiratory chain, as proton pump, 8–12 reversed electron transport, coupling/uncoupling, 15–17 skeletal muscle, 237 uncoupling, 77, 80, 104, 389–390 Oxidative stress: cardiac, 206–209 drug-induced, see Drug-induced oxidative stress formation of, 72, 74 inﬂuential factors, 20–22, 42, 46, 60, 72, 84, 107, 209, 263, 294, 518–520, 523–524, 529 prevention and therapy for, 109 thiol-disulﬁde systems, 436 Oxidoreductases, 435, 450 2-Oxoglutarate: carrier (OGC), 441 dehydrogenase, 20 OXPHOS activity assays, immunocapture-based: characteristics of, 398–399 complex I, 398–399, 401–403 complex II, 398–399, 401–403 complex III, 403 complex IV, 398, 400–403 complex V, 398, 400–403 data analysis, 400–401 drug screening using, 398, 401–404 future research directions, 409–410 speciﬁcity of, classical inhibitor effects, 401 Oxygen: calibration, 334–337 consumption analysis: characteristics of, 354, 468 extracellular assessment, 357–358, 360, 362–363

intracellular assessment, 357–358, 360, 363–366 probes, quench-ﬂuorescence, 358–360 quench-ﬂuorescence oxygen sensing, 355–360 solid-state oxygen-sensitive coating consumption rate (OCR), 374–379 Oxygraph-2k, 332–340, 347 Oxygraphy, traditional, 329–333 31

P, 494 Palmitate, 73 PAM complex, 55 Panadiplon, 170 Pancreas, 5, 533 Pancreatitis, 100, 498, 520 Pancytopenia, 466 Panadiplon, 170 Papillary necrosis, 302 Paracetamol, 177–178, 187 Paracrystalline inclusions, 522–523 Paraquat, 20, 439 Parasympathetic nervous system, 478 Parenchymal cell death, 99 Paresthesias, 184 Parkinsonism, 93 Parkinson’s disease, 89, 104, 252, 264, 409, 440 Paroxetine, 402 PARP cleavage, 86 Parsing, 348 Partition coefﬁcient, 583 Passive proton leak, 10 Pathogenesis, 508 PDSS genes, 314–315 Pentacarboxylic acid, 425 Pentachlorophenol, 104 Pentamidine, 98, 101, 179, 302 PEO gene, 314 Peptides, 55–56, 154–155, 363, 365, 524 Peptidylprolyl cis-trans isomerases (PPIases), 296 Percoll density gradient centrifugation, 416 Perhexiline, 79, 82, 171–172, 186–187 Peripheral benzodiazepine receptor (PBR), 27, 163 Peripheral blood cells (PBCs), 183 Peripheral blood mononuclear cells (PBMCs), 280, 404, 479, 481, 502 Peripheral neuropathy: antiviral drugs, 474–475, 483 clinical assessments, 498–499 drug effects, 318 drug-induced liver injury (DILI), 185 inﬂuential factors, 100, 253–257 reverse transcriptase inhibitor, 253–255

610 Peripheral neuropathy: (Continued ) microtubule-modifying agents and mitochondria, 255–256 nephrotoxicity, 303 protein synthesis and, 466 statins and, 256–257 PERK, 451 Permeability: calcium signaling, 24 ﬂuorescent probe studies, 362–363 transition, parent drugs: anionic uncouplers, 158–162 examples of drug-induced mitochondrial dysfunction, 162–163 Permeability transition pore (PTP), 27–30, 32 Permeabilization, 101, 299, 304, 341, 343–346 Pernoline, 114 Peroxidases, 21, 95 Peroxides, 439 Peroxidredoxins (Prx), 434, 441, 452 Peroxiredoxin system, 441–443 Peroxisomal disorder, 315 Peroxisome proliferation γ (PARPγ), nuclear, 95 Peroxisome proliferator-activated receptors (PPARs), 240–241, 523 Peroxisomes, 21 Peroxynitrate, 88 Peroxynitrite, 99, 109, 177–178, 294, 296, 438, 520 Pesticides, organophosphates, 235 p50, 447–448 p53, 150, 275 P450, 83, 93, 102, 108 pH: implications of, 17, 104, 375, 441, 445, 560–564 intracellular measurement (pHi ), 561–562 Phage polymerase, 48 Phage T7, 48 Phagocytosis, 366 Pharmaceuticals, 20, 37. See also speciﬁc drugs Phenformin, 91, 94, 114 Phenolic groups, 89 Phenols, 101 Phenylalanine, 494 Phenylbutylnitrone, mitochondria-targeted version (MitoPBN), 582 1-Phenylpiperzine, 111 Phenylpropylamine derivatives, 112 Phenytoin, 92–94, 167, 319 Phosgene, formation of, 108 Phosphate:

INDEX dehydrogenase (PDH), 513, 515 groups, 95, 205 levels, 280 transporter, 343 Phosphatidic acid, 102 Phosphatidylserine, 113 Phosphaturia, 301 Phosphocreatine (PCr), NMR spectroscopy: characterized, 560–561 kinetics, 564–567 recovery dynamics following single metabolic perturbation, 568–569 Phosphodiesters, 178 Phosphoenolpyruvate, 17 Phospholipase A2, 74, 150 Phospholipid(s): characterized, 7, 28, 39, 78, 101, 112, 171 membrane bilayers, 6–7, 525 phosphate groups, 95 Phospholipidosis, 78–79, 82–83, 111–114, 117, 172–173 Phosphonates, acyclic nucleotide, 301 Phosphorus, magnetic properties, 558 Phosphorylation: applications of, 100, 221, 276, 318, 376, 524, 526, 531 control protocol (PC protocol): experimental example, 342–343 ﬂux control ratios, 343–344 with intact cells, 340–342 high-resolution respirometric analysis, 340 oxidative, see OXPHOS (oxidative phosphorylation) Photobleaching, 357 Photodynamic therapy (PDT), 390 Photomultiplier tubes, 421 Photosensitivity, 465 Photostability, 366 Phototoxicity, 359, 366 Phytoalexin, 84 Piericidin, 89–90, 390 Pioglitazone, 95, 438, 531 Piperazine, 93 Pirfenidone, 81 Pirprofen, 75–77, 169–170 Pivampicillin, 170 Pixels, microscopy systems, 420–422 pK a , 109–112 Plasma transaminase, 77 Platelet contamination, 280 Pneumocystis infection, 101 Pneumonitis, 106 31 P-NMR spectroscopy: intracellular pH measurement, 561–564

INDEX phosphocreatine kinetics, 564–567 phosphocreatine recovery dynamics following single metabolic perturbation, 568–569 reconstruction of mitochondrial transfer function, 567–568 Point mutations, 312 Poison topoisomerases, 178 Polarization, 423 Polarographic oxygen sensors, 329–338 POLG genes, 314, 319 PolRMT, 47–49 Poly(A) tail, 50, 464 Polycistrionic transcripts, 46–47 Polymerase chain reaction (PCR), 303, 481–482 Polymerase gamma (γ): characterized, 283, 303, 474, 526 inhibitors, 274–276, 280–281 PolγA, 43 PolγB, 43 Polymine analogs, 179 Polymorphisms, 40, 130–135, 318, 409–410, 466, 470 Polyneuropathy, 277, 281 Polypeptides, 10, 300, 312–313 Polypharmy, 72, 305 PolyU RNA, 467 Porﬁromycin, 106 Porin, 6, 27, 533 Post hoc analysis, 135 Posttransductional modiﬁcations, 294 Potentiometric systems, 354 Poxvirus, 301 Pramocaine, 111 Preconditioning, deﬁned, 99 Pregnancy, 77, 168–169, 186, 278–279 Prilocaine, 111 Primaquine, 173 Primers, 483 Priming, mtDNA replication, 45–46 Proaptotic protein, 27–28 Probenicid, 303, 305 Probes: ﬂuorescent, 445–447 quench-ﬂuorescence, 358–360 Procaine, 111, 364 Prodrugs, NRTIs, 276 Programmed cell death, 27, 38. See also Apoptosis Progressive external ophtalmoplegia, 531 Prokaryotes, 7, 38–39, 48–51, 57, 523 Proliferator receptor gamma coactivator 1 (PGC1), 183 Proliferator-activated receptor (PPAR) γ: agonists, 528–530

611 characteristics of, 479 Proliferators-activated receptor gamma coactivator-α (PGC-1α), 60 Promoters, mitochondrial genome, 45–47 Proof-of-concept studies, 304 Proof-of-principle studies, 444 Proofreading, 181–182 Prooxidant α-hydroxyethyl radicals, 76 Propanolol, 79–80, 101, 111 Propidium iodide, 427 Propionic acid derivatives, 110 Propofol, 98, 103–104 Propranolol, 111 Prostaglandins, 102, 169 Protease inhibitors (PIs), 278, 281, 300, 302, 474–475, 479 Protein disulﬁde isomerase (PDI), 450–451 Protein kinase C, 24, 100 Protein synthesis: bacterial, 259, 404, 466 components of, 47–48, 50, 53, 74, 77, 89, 166, 185, 187, 398, 527 mitochondrial, 185, 257 Proteins, see speciﬁc proteins ancillary, 315 assessment using lateral ﬂow dipstick immunoassays, 404–408 folding, oxidative, 450–452 green ﬂuorescent (GFP), 42, 59 import into mitochondria, 52 kinases, 205 high-mobility-group (HMG), 47 import into mitochondria, 52–58 synthesis, see Protein synthesis transporting, 8 Proteinurea, 301 Proteomics: analyses, 438 import of proteins into mitochondria mitochondrial targeting signal, 53–54 translation through/into inner membrane, 55–56 translation through/into outer membrane, 54–55 supercomplexes, 56–58 redox, 452 techniques, 444–445 Proton(s): characterized, 152 leak, 27 pump, 7–12, 14 translocation, 159 Protonation, 562 Protonophores, 16, 101, 341, 418

612 Protozoa, 5 p38 MAPK, 74 PUMA, 150–151 Purines, 244, 279 PUS1 gene, 314 Pyknosis, 516 Pyridines, 20, 209 Pyridoxamine, 109 Pyrimidines: characterized, 244 metabolism, 281–282 nucleosides, intracellular, 274–275 1-Pyrimidinylpipersine, 111 Pyruvate: dehydrogenase, 24 functions of, 176, 187, 211, 346–347 Pyruvate dehydrogenase complex (PDHC), 318 Pyruvate dehydrogenase kinase (PDK), 318 Q-cycle, 11, 347 Quantitative analysis, 362 Quantitative structure-activity relationship (QSAR) studies, 109–115 Quenching, 355–360, 417 Quercetin, 173 Quinidine, 79, 82–83, 101 Quinine, 82, 101, 173, 259 Quinisocaine, 111 4-Quinolones, 178 Quinolone antibiotics, 107, 179 Quinone, 96, 105, 110 Quinoneimine, 99, 102, 105 Quinuclidine agents, 82–83 Racker, Efraim, 12 Radioactive materials, 380 Raji cells, 438 Ranolazine, 89 Rat studies, see Animal models Rb protein, 275 Reactive metabolite-mediated mitochondrial disruption in DILI: direct toxicity, 150–153 immune reactions, 154–157 tolerance, 157 Reactive metabolites, 186 Reactive nitrogen species (RNS), 519–520, 541 Reactive oxygen species (ROS): antiviral therapy and, 479–480 calcium and, 261 components of, 18–19, 23–24, 80, 114, 252 cytotoxic, 76, 215–216 drug design, 576–577

INDEX formation, 86, 89–109, 170–171, 174, 176, 182–183, 293–295 ﬂuorescent probes, in detection of, 445–447 generation of, 252, 260, 262–263, 389 mitochondrial fatty acid oxidation, drug-induced inhibition, 73 mitochondrial function analysis, 341 mitochondrial uncoupling, 73 nephrotoxicity, 293–295, 299 perioxisomal, 75 signal transduction, 99 sources in mitochondria, 436–439 Reactive superoxide radicals, 294 Real-time PCR, 481 Redox cycling: characterized, 161–162, 218, 221 oxygen activation, 105–106 Redox-sensitive green ﬂuorescent protein (roGFP), 446–447 Redox signaling: and control, compartmentation in, 447–450 mechanisms, 434–435 Reduction-oxidation (redox) processes, 4, 11–12 Reductive stress, 96 Ref1 system, 448–449 Reiske iron-sulfur protein, 215 Renal, see Kidneys Reoxygenation, 18 Reperfusion, 61 Replication, molecular biology of: ﬁssion of mitochondria, 58–60 import of proteins, 52–58 long-term inhibition of, 114 mitochondrial biogenesis, control of, 60–61 mitochondrial genome, 40–52 overview, 37–40 segregation during cell division, 58–60 Replisomes, 43–44 Reporter proteins, 59 Residual bodies, 520 Respirasome, 57 Respiration: components of, 170 mitochondrial permeability transition (MPT), 159–161 uncouplers of, 377–378 Respiratory chain: dysfunction, 282 function, 7–12, 319 Respiratory control ratio (RCR), 342, 344, 362, 468–469 Respiratory dysfunction, pathogenic mtDNA mutations, 312 Respirometry, principle of, 329

INDEX Resveratrol (R-3), 173, 392–393, 518 Reticulocytes, 282 Reverse transcriptase (RT): characterized, 43, 180, 474 HIV, 282–283 inhibitor, 253–255 Reversed electron transfer, 19 Reye’s syndrome, 167–169, 186, 494 Rezulin, 114 Rhabdomyolysis, 236–245, 526 Rhodamine, 417 Ribavirin, 281 Riboﬂavin, 281 Ribonucleic acid, see RNA Ribosomal RNA (rRNA): characterized, 40–41, 46, 48, 128, 260, 312–313, 316, 463, 466, 470, 482–483 methyltransferases, 39, 47 Ribosomes, mitochondrial genome, 53 Rieske protein, 11 Rifampicin, 87 RIP-TRAF2 pathway, 88 Risperidone, 90, 92 Ritonavir, 100, 301–303, 475 R-loop, 45 RNA, see speciﬁc types of RNAs polymerase, 45, 47 synthesis, 178 viral, 300 RNase L, 184 Rolitetracycline, 77 Ropivacaine, 93, 103 ROS/RNS signaling, 452–453 Rosiglitazone, 91, 95, 241, 281–282, 402, 438, 530 Rotenone, 18–19, 89–90, 252, 339, 342–343, 345–346, 376, 378, 384, 387–388, 390, 401, 523, 531 Rottlerin, 100 Rule of ﬁve, 583 Ryanodin, 24 Ryanodine receptor (RyR), 450–451 Saccharomyces cerevisiae, 39 Safranine, 417 Salicylate, 186 Salicylic acids, 101, 110 Sam35, 55 Sam37, 55 Sam50, 55 Saponin, 345 Saquinavir, 183, 475 Sarcosine, 495–496 Scanning electron microscopy, 420

613 SCO genes, 314 SDHA gene, 314–315 S -(1,2-dichlorovinyl)-l-cysteine (DCVC), 441 Seizures/seizure disorders, 83, 166, 319 Self-digestion, 30 Semiubiquinone, 11 Sertraline, 82 Serum albumin, 72 Serum creatine kinase, 100, 277 Serum transaminase, 156–157 Sevoﬂurane, 93, 97, 99 Sex hormones, female, 169, 186 S -glutathionylated protein (Pr-SSG), 448–449 Shine-Dalgarno sequence, 51 Shy1 gene, 58 Signaling pathways, 74 Signal-to-noise ratio (SNR), 558, 561 Signal transduction pathways, 42, 367 Simvastatin, 237, 402, 409, 443 Sinusoidal endothelial cells (SECs), 154 SIRT1, 84–85 Skeletal muscle: characterized, 235–236, 451, 498, 511, 526, 531, 565–568 myopathy, xenobiotic-induced, 236–245 overview of, 235–236 Skeletal myopathy, 100 SLC25A3, 315 Slope, oxygraph, 329, 331 Smac (second mitochondrial activator of caspases), 27 Small interfering RNA (siRNA), 540 Smallpox, 301 [35 S]methionine, 404, 468 SOD2-knockout mice, drug safety testing, 545–551 Sodium, properties of, 23, 558 Sparks, calcium signaling, 25 Spectroﬂuorometry, 364 Spectrophotometric measurement: assay limitations, 388–389 complex I (NADH-ubiquinone oxidoreductase activity), 384–385 complex II (succinate-ubiquinone oxidoreductase activity), 385 complex III (ubiquinol-ferricytochrome c oxidoreductase), 385–386 complex IV (cytochrome c oxidase), 386–387 drug-mediated toxicity, 389–393 succinate-cytochrome c reductase, 388 succinate-hexacyanoferrate reductase, 387 Spermatogenesis, 59 S-phase, in cell cycle, 42 Spikes, calcium signaling, 25

614 Spin-lattice relaxation, 558–559 Spin physics, 557–561 Spin-spin relaxation, 559 Spinal cord, 315 Src-Gab1 pathway, 88 SREBP-1, 278 Staining: lateral ﬂow dipstick immunoassays, 407 in microscopy, 508, 520 mitochondrial function analysis, 364 techniques, 12–13, 58–59 State 3, 16 State U, 16 Statins: characteristics of, 236–240, 242–243, 245, 256–257, 320, 443 ﬁbrate synergy, 242–243 Statistics, paired-comparison, 376 Staurosporine, 445 Stavudine (d4T), 78, 98, 100, 180, 182–183, 244, 274–276, 278–279, 281, 390, 474–476, 478, 481, 498 Steatohepatitis, 75–85, 146, 171, 186, 282, 584 Steatosis: alcohol-induced, 177 characterized, 75–85, 148, 170, 172 drug-induced, 163–165 hepatic, 184, 186–187, 279, 303, 465, 474–476, 483, 498 Stern-Volmer calibration function, 359 Steroids, androgenic, 878 Stevens-Johnson syndrome, 475 S-thiylation, 434 Stigmatellin, 96–97 Stoichiometry, 10, 562, 564–565 Strand displacement model, 44 Streptomyces peucetius caesius, 212 Streptomycin, 259, 316, 466, 470 Streptozotocin, 212 Strokes, 264, 320 Structure-activity relationships: mitochondrial/lysosomal accumulation by cationic amphiphile drugs, 111–113 mitochondrial toxic drugs, 109–110 quantitative (QSAR) studies, 109 signiﬁcance of, 81, 410 Subarcolemmal mitochondria (SSM), 204 Succinate, 10–11, 19, 96–97, 100, 345–346, 498 Succinate-cytochrome c reductase (SCR), 388–389 Succinate dehydrogenase (SDH), 107, 416, 511, 534 Succinate-Q reductase, 89

INDEX SUGLA2 gene, 314 Sulfadiazine, 302 Sulfates, 583 Sulfhydryl groups, 99, 434 Sulfo Black B, 104 Sulfur, 12 Supercomplexes, mitochondrial replication, 56–58 Superoxide: anions, 20–21, 109, 170 production, 393, 581 radical, 96, 156, 519 Superoxide dismutase (SOD), 18, 21, 30, 84, 106, 109, 159, 443, 519, 545, 576 SURF gene, 314 SURF-1, 58 Systems biology, 444 Tacrine, 79, 83, 178–179, 187 Tacrolimus, 295, 297 Tafazzin (TAZ) protein, 315 Taft-Kutter-Hansch constant, 109–110 Tamoxifen, 76, 78, 97, 103, 172, 179, 186–187, 402–404 Tasmar, 114 T-associated maternal effect (TME), 547 Taurochenodeoxycholic, acid, 87 Taurocholate, 86 Taurohydrodeoxycholic acid, 88 TAZ gene, 314–315 t-butyl hydroperoxide, 426 T-cell receptor (TCR), 154, 156 3TC, see Lamivudine (3TC) TEMPOL, 582 Tenofovir, 244, 274, 279–281, 301–302, 474, 483, 499 Tetracaine, 101 Tetracyclines, 75–77, 165–166, 186, 464–465 Tetramethoxystilbene (MR-4) analog, 518 Tetramethylrhodamine methyl (TMRM), 417–418, 422, 426 Tfam-deﬁcient mice, 543–544 Thermodynamics, impact of, 347, 567 Thermogenesis, 510 Thiamine, 281 Thiamphenicol, 185, 465 Thiazolidinediones, 73, 75, 91, 95–96, 522, 528 Thiol-disulﬁde redox, 435, 447 Thiols/thiol groups, 91, 106, 150, 208–209, 237, 444–445, 449–450, 452 Thioredoxin reductase 1, 107 Thioredoxins (Trx), 434–435, 441–443, 446, 448–449, 452

INDEX Thrombocytopenia, 465–466 Thrombopenia, 279 Thymidine: kinases (TKs), 181, 275–276 monophosphate (TMP), 181–182 triphosphate (TTP), 181–182 Tiagabine, 319 Tianeptine, 76–78, 170, 186 Tiaprofenic acid, 170 Tim proteins, 54–56 TIP-2k, 342 Tissue toxicity, 25, 88 TK2 gene, 314 T lymphocytes, 155, 277–278 TMC125, 474–475 TMPD (N, N, N , N tetramethyl-p-phenylenediamine dihydrochloride), 331, 339, 341, 362 TNF-α, 75, 155–156, 162–163, 168, 442, 479, 498 TNF-α related apoptosis-induced ligand (TRAIL), 155–156 Tobramycin, 259 α-Tocopherol, 20, 162, 523 Tolcapone, 98, 104, 108, 114 Tolfenamic acid, 102, 110 TOM (translocation through inner membrane) complex, 54–55 T1 relaxation, 559 Topiramate, 319 Topoisomerase II, 178 Topoisomerase-interfering drugs, 178 Torsades de pointes, 81 Toxicity: metabolic factors, 185–186 mitochondrial, see Drug-associated mitochondrial toxicity; Mitochondrial toxicity Toxicology, 316, 354, 444–445 TP gene, 314 TPMP (tetraphenylphosphonium) cation, 579–582 Trafﬁcking of mitochondria, 252–253, 258, 263 Transcription: factors, 74 mitochondrial replication, 39, 46–48 Transfer RNA (tRNA), 40, 45–47, 49, 51–52, 130, 312–313, 464 Transforming growth factor (TGF-β), 81–82, 88 Transhydrogenation, 22 Transition metals, 28, 88 Translation:

615 mitochondrial replication, 39 mitochondrial system, 48–52 mitochondrial targeting signal, 53 Translocation, 8, 11, 17, 50 Translocation through inner membrane (TIM), 54–56 Transmembrane potential, 19, 73 Transmethylation, 493, 495 Transmission electron microscopy (TEM), 508, 517, 519, 533–534 Transphosphorylation, 6 Transplantation, liver, 145, 157 Transporters, types of, 6–7, 9, 283, 293, 302–303, 305 Trauma, 29 Tricarboxylic acid (TCA) cycle, 4, 177, 347, 378, 469, 495 Trichloroethylene, 441 Tricyclic agents, 77, 103 Triglycerides, 77, 80, 107, 524 Trimethoprim-sulfamethoxazole, 302 Triose phosphate isomerase (TPI), 445 Triphenylphosphonium (TPP) cation, 576–583 TRMU gene, 50, 129–130 tRNAL, 483 Troglitazone, 87, 91, 95, 114, 162, 187, 438, 530–531, 549–550 Tropolone, 98, 105 Trovaﬂoxacin, 105, 107, 179, 187 Truncated Bid (tBid), 155–156 Tryptamine, 520 Tryptophan, 464 TSFM gene, 314 TTFA (thenoyl triﬂuoroacetate), 392, 401–402 T2 relaxation, 559 Tubular cells: apoptosis, 304 peculiarities of, 292–293 proximal, schematic representation of, 293 renal, 295, 299 TUFM gene, 314 Tukey’s post hoc test, 403 Tumor(s): characteristics of, 318, 530 cells, 106 viruses, 43 Tumor necrosis factor (TNF), 76, 88 TUNEL, 238–239, 241, 243, 549 Twenty-ﬁfth chromosome, 40 Twin CX3C, 56 Twinkle proteins, 43–44 Two-dimensional agarose gel electrophoresis, 44 Tyrosine, 99, 520 Tyrosine kinase, 447

616 Ubiqinone, 57 Ubiquinol, 10–11, 578 Ubiquinol-cytochrome c reductase, 89 Ubiquinone, 10, 89, 96, 176, 237–239, 384–385 Ubisemiquinone, 99, 436–437 UCP-knockout mice 544 Ulceration, 101 U-loop, 49 Ultrasound, as diagnostic instrument, 83–84, 280, 497 Ultraviolet radiation (UV), 20, 28, 30, 391–392 Uncoupled state, 16 Uncouplers, 469 Uncoupling: characterized, 16, 516, 524, 530 control ratio (UCR), 344 mitochondrial, 73 proteins, see Uncoupling protein Uncoupling protein (UCP1/UCP2/UCP3/UCP4), 16–17, 19, 544 Unfolded protein: characterized, 56 response (UPR), 451–452 U.S. Food and Drug Administration (FDA), 212, 474 Up-regulation, 60, 87, 318, 450 UQCRB gene, 314 Urea, 5 Urea synthesis, cystolic, 17 Uridine, 275, 277, 282–283 Ursodeoxycholic acid, 87 3 -UTR, 50 5 -UTR, 50–51 Valinomycin, 365 Valproate, 75, 84, 187, 497 Valproic acid, 75–76, 83, 159, 166–167, 186–187, 319 Vancomycin, 295, 300 Venoocclusive disease, 148 Veratine, 244 Very low density lipoproteins (VLDLs), 75–76, 163, 166 Vigabatrin, 319 Viral infections, 186–187 Viral protein R (vpr), 498 Vision loss, 257, 317 Vitamins: supplementation, 20 vitamin E deﬁciencies, 84, 100, 176, 522 Voltage-dependent anion channel (VDAC), 6, 27, 55, 208–209, 294, 296–297, 423

INDEX Walker, John E., 4 Warburg effect, 318 Warburg, Otto, 4 Water (H2 O), 108 Watson-Crick base pairing, 46 Waves, calcium signaling, 25 Western blot analysis, 406, 442–443, 448, 452 White blood cells, 177 Wilson’s disease, 528 Wood preservatives, 104 World Health Organization (WHO), 476 WY-14643, 240, 243 Xanthine dehydrogenase, 76 X chromosome, 57 Xenobiotics: characterized, 4, 20, 28, 88–89, 113–114, 424, 523, 540 detoxiﬁcation, 435 effects of, 257 ﬁbrates, 237, 240–243 3,4-methylenedioxymethamphetamine (MDMA), 244–245 nucleoside reverse transcriptase inhibitors (NRTIs), 235, 243–245, 273–276 statins, 236–240, 242–243, 245 toxic, 108 XF24 analyzer, 374–376, 379–380 X-linked genes, 56–57 X-ray crystallography, 96 X-rays, 28 Yeast: autophagy-related signaling pathways, 29 characteristics of, 5 ﬁssion process, 59 genetic studies, 56 mitochondrial function analysis, 366 mitochondrial genome, 46–47, 50, 52 mitochondrial replication, 46–47, 58 mitochondrial translation, 50 Zalcitabine (ddC), 98, 100, 180, 182, 244, 274–276, 279, 474, 480–481, 498 ZDV, 498 Zidovudine (ZDV), 98, 100, 174, 180, 182–183, 243–244, 274, 276–278, 282, 303, 390, 474. See also AZT Zimmerman, Hyman, 146 Zinc, 21, 30, 258, 263, 519 zVAD, 297, 299

A

B

Chapter 1, Figure 2 Mitochondrial network. Human osteosarcoma cell stained with ﬂuorescent dyes MitoTracker CMXRos (red) for mitochondria, phalloidin-FITC (green) for actin ﬁlaments, and DAPI (blue) for nuclear DNA. (A), Whole cell; (B), higher magniﬁcation of a cell fragment with mitochondria. The bars correspond to 12 µm.

Chapter 1, Figure 6 Subunit structure of mitochondrial ATPase/ATP synthase. Subunits c of FO are assembled as a ring plunged into the inner membrane. They allow protons to return to the mitochondrial matrix. Transient and sequential protonations of each of the 10 c subunits causes a clockwise rotation (when viewed from the membrane side) of subunit γ, driving a cycle of conformational changes of the α3 β3 assembly of F1 . Full 360◦ rotation requires 10 protons to pass across the inner mitochondrial membrane. This allows for phosphorylation of three molecules of ADP. OSCP (oligomycin sensitivity-conferring protein), together with subunits a and b, comprises a stator that prevents the α3 β3 assembly to rotate together with subunit γ. Note that the OSCP subunit is distant from FO and is not the oligomycin-binding site. However, it makes a link between subunit b and the α3 β3 assembly and prevents the latter from undergoing conformational changes in the presence of oligomycin.

Intermembrane

Chapter 1, Figure 10 Model of the contact site between the outer and inner mitochondrial membranes that may function as the permeability transition pore. Indications: VDAC, voltage-dependent anion channel (mitochondrial porin); ANT, adenine nucleotide translocase; Cyp D, cyclophilin D; HK, hexokinase; PBR, peripheral benzodiazepine receptor; Bcl-2, antiapoptotic protein Bcl-2. Cytochrome c molecules associated partially with the outer face of the inner membrane and partially free in the intermembrane space are indicated by red circles.

(A) Electron transport chain H+

H+

H+

H+

H+

H+

A- H+

Cyt c

+++++ Q ETF

I Matrix

---

III

IV

Complex I inhibitors: Rotenone, Piericidin, Capsaicin Antihyperlipidemics Anesthetics Antidiabetics Anticonvulsants, Idebenone Complex II MPTP, Antipsychotics Malonate Flutamide Oxaloacetate Isoniazid

Complex III inhibitors: Antimycin A Stigamatellin Acetaminophen quinoneimine Isoflurane

II

Complex IV inhibitors: ADP

ATP

Cephaloridine

AH

H+

inhibitors:

Complex V/ ATP synthase Uncouplers: SCoA

R O

FAD

Inhibitor: CoA

R

H+

Free fatty acids, bile acids Pentamidine NSAIDS. Tamoxifen Tolcapone, Propofol

O FADH2

(C) Mt protein synthesis and biogenesis

V

oligomycin

S OH O

Cytochrome complexes

CoA R

Translation

O2

S

Inhibitors of fatty acid β– oxidation:

O

O

CoA Transcription

R

Antivirals: Zidovudine

MtDNA

O R

CoA S

S

+ R

Tetracycline, NSAIDs, antidepressants, tamoxifen

(D) Oxidative stress caused by:

citric acid cycle

O

Doxorubicin-semiquinone Gentamicin Trovafloxacin

CoA S

(B) Fatty acid β-oxidation

2CO2

Chapter 3, Figure 1 Mitochondria–drug interactions: (A) inhibition of mitochondrial electron transport chain complexes; (B) inhibition of fatty acid β-oxidation; (C) inhibition of protein synthesis and biogenesis; (D) formation of mitochondrial oxidative stress.

H++ H++ H++ cc II Q II III III SUST RATO SUBSTRATE

O2 CO22+H +H22O

ADP ADP CsA ++ Ca++

H22O H O ++ Ca Ca ++

CyD cifD

ATP CyD ci cifD

H H++ Ca++ UniPorter VDAC

H2O

ADP

ATP

MPT PTM

ANTc

ANTm

Chapter 10, Figure 2 The transient mitochondrial permeability pore (MPT) is formed by the action of calcium when type 1 amine nucleotide translocase (ANT) in its c conformation combines with the voltage-dependent anion channel (VDAC). Interaction of intramitochondrial cyclophilin D (CyD) with ANT facilitates such binding, although this is not essential. CsA binding to CyD prevents its binding to the ANT and MPT pore activation.

7.45

150

7.40

OCR, pmol/min

140

7.35

ECAR, mpH/min

130

7.30 7.25

120 100

pH

Oxygen (mm Hg)

160

1

2

3

4 5 Time (min)

6

7

7.20

Well 1

Chapter 14, Figure 1 are made.

Compound Sensors Analyte (O2) Cell

Well 2

XF24 analyzer: how real-time measurements of OCR and ECAR

Complex IV / Frataxin (%)

A

110 100 90 80 70 60 50 40 30 20 10 0 0 1 3 5 Cell population doublings after addition of 40 µM linezolid 1

2

3

4

B Complex V α subunit Porin

Complex II-30 kD subunit

Complex IV-subunit Complex I-20 kD subunit

C

1

2

3

4

5

6

Chapter 16, Figure 5 Effect of 40 µM linezolid on HepG2 cells. HepG2 cells were treated with the vehicle (0.1% DMSO) or 40 µM linezolid for ﬁve cell population doublings. (A) 2 µg protein from detergent-solubilized cells was analyzed with complex IV+ frataxin PQuant dipsticks. 100% represents the complex IV/frataxin ratio in vehicle-treated cells. Data are expressed as mean± SD (cell cultures were grown in duplicates and each culture was analyzed with duplicate dipsticks). (B) 20 µg protein from detergent-solubilized cells was loaded on a 10 to 20% acrylamide gel, subjected to SDS polyacrylamide gel electrophoresis, transferred to a polyvinylidine diﬂuoride membrane, and probed with a cocktail of mAbs (anti-complex V α subunit mAb, anti-porin mAb, anti-complex II 30-kDa subunit mAb, anti-complex IV subunit 2 mAb, anti-complex I 20-kDa subunit mAb). Lanes 1, 2, 3: cells grown in 40 µM linezolid for one, three, and ﬁve cell population doublings, respectively. Lane 4, cells grown in the vehicle (0.1% DMSO). (C) Immunocytochemistry analysis of HepG2 cells treated with either 40 µM linezolid for ﬁve cell population doublings (images 1 to 3) or the vehicle, 0.1% DMSO (images 4 to 6). Cells were ﬁxed with 4% paraformaldehyde, permeabilized with 0.1% Triton X100 in PBS, and stained with an anti-complex IV mAb and Alexa Fluor 594 goat anti-mouse IgG2a antibody (images 1 and 4), an anti-porin mAb and Alexa Fluor 488 goat anti-mouse IgG2b antibody (images 2 and 5), and the nucleic acid stain DAPI (4 ,6-diamidino-2-phenylindole). The vehicle-treated cells shown in image 6 (the merged image of images 4 and 5) appear yellow and show that complex IV (image 4, red) and porin (image 5, green) co-localize. In contrast, the linezolid-treated cells shown in image 3 (the merged image of 1 and 2) appear green and emphasize the reduced level of complex IV (image 1) and normal level of porin (image 2).

5 mm Chapter 17, Figure 1 Red and green mitochondrial ﬂuorescence after loading with JC-1. A cultured mouse hepatocyte was loaded with 100 nM JC-1 for 30 minutes in Krebs–Ringer–HEPES buffer (KRH) at 37◦ C, and green and red ﬂuorescence were imaged by multitrack confocal microscopy using 488- and 543-nm excitation light, respectively. Red inclusions within green ﬂuorescing mitochondria are JC-1 J-aggregates.

Chapter 17, Figure 2 Distribution of electrical potential in a cardiac myocyte. An adult feline cardiac myocyte was loaded with 200 nM TMRM for 20 minutes at 37◦ C, and TMRM ﬂuorescence was imaged by confocal microscopy using 543-nm excitation and a 565 to 615-nm emission ﬁlter. The distribution of is displayed in pseudocolor, as described in the text.

Chapter 17, Figure 4 Increased mitochondrial inner membrane permeability in a rat hepatocyte induced by t-butyl hydroperoxide. A cultured rat hepatocyte was loaded with TMRM (left panel) and calcein (right panel). Note that dark round voids in the green calcein ﬂuorescence coincide with red TMRM labeling of mitochondria. After 9 minutes exposure to 100 µM t-butyl hydroperoxide, dark mitochondrial voids ﬁlled with green calcein ﬂuorescence. Simultaneously, mitochondria release red TMRM ﬂuorescence. These events signiﬁed the onset of the MPT.

Chapter 17, Figure 5 Inner membrane permeabilization after ischemia/reperfusion in rat myocytes visualized by mitochondrial calcein release after cold ester loading/warm incubation. An adult rat cardiac myocyte was cold-loaded with calcein AM and subjected to 3 hours of simulated ischemia at 37◦ C at pH 6.2 followed by reperfusion at pH 7.4 for 10 and 20 minutes. Green calcein ﬂuorescence was retained by mitochondria at the end of ischemia. After reperfusion, mitochondria began to release calcein, signifying inner membrane permeabilization. For experimental details, see [45].

A Mitochondrial-targeted

Nuclear-targeted

Redox–sensitive GFP (roGFP)

B

− D-amino acid

+ D-amino acid

Wildtype

Trx2 overexpression

CMF-DA stained NLS-DAAOtransfected

C

MitoSox Stained

Chapter 18, Figure 3 Visualization of compartmental redox pools by redox-sensitive ﬂuorescent probes. (A) HeLa cells transfected with redox-sensitive GFP (roGFP) targeted to the mitochondria (green) and nucleus (white), providing mitochondria and nuclear-speciﬁc assessment of the redox state, as described by Hanson et al. [112] and Dooley et al. 113, respectively. (B) HeLa cells transfected with nuclear targeted D-amino acid oxidase (NLS-DAAO), treated with or without D-amino acids and stained with the thiol-speciﬁc ﬂuorescent dye 5-methylchloroﬂuorescein (CMF-DA), as described by Halvey et al. [115]. A selective decrease in nuclear thiol staining occurs during nuclear oxidative stress. (C) HeLa cells were transfected with an empty vector or a thioredoxin-2 (Trx2) overexpressing plasmid and stained with the mitochondrial ROS-speciﬁc indicator dye MitoSOX (Invitrogen Corporation) (unpublished data).

500 450 400 350 300 250 200 150 100 50 0

11 10 9 8 7 6 5 4 3 2 1 0

: Therapy with ddI/d4T/NFV

360

330

300

270

240

210

180

150

120

90

60

30

0

−30

−60

Time after start of therapy (days)

Chapter 20, Figure 1

Results of combination therapy.

Lactate (mmol/L)

Copies mtDNA/cell

Stop therapy

Mitochondrial DNA-pol-g

NRTIs

Glucose

free fatty acids

Energy production

Fat metabolism

ATP Glucose

Pyruvate

ATP Pyruvate

NADH NAD+

Krebs cycle

Acetyl-CoA

lactate

Mitochondrion

serum lactate

lipid droplets

Krebs cycle b-oxidation Acetyl-CoA of fatty acids

Mitochondrion

free fatty acids

Chapter 20, Figure 3 Relation between mitochondrial DNA and mitochondrial function: Due to NRTI treatment, DNA polymerase γ inhibition leads to depletion of mtDNA and mtDNA-encoded proteins, thereby impairing mitochondrial function. Ultimately, this leads to changes in energy production and fat metabolism with increased serum lactate levels and increased free fatty acids.

Chapter 22, Figure 1 Dendritic mitochondrion from chick cerebellum. Electron tomography reveals unprecedented details of mitochondrial structure, including ﬁne structure of junctions between cristae and the inner membrane. The coarse texture apparent on the cristae is caused by the respiratory complexes. The outer membrane is shown in purple, the inner boundary membrane in aqua, and the cristae in yellow, green-gray, and red to demonstrate variety of cristal structure.

A

B

C

D

E

F

Chapter 22, Figure 4 Histochemistry assessment of myoﬁber type in fast-twitch rat extensor digitorum longus (EDL; panels A, C, and E) and slow-twitch postural soleus muscle (panels B, D, and F), stained for complex 1 activity (NADH stain; top pair), complex II (succinate dehydrogenase; middle pair), and Complex IV (cytochrome c oxidase; bottom pair). Note the heterogeneous distribution of ﬁber type in EDL, with larger, anaerobically poised, fast-twitch ﬁbers appearing fainter than the aerobically poised, mitochondrially enriched slow ﬁbers. Note also the relatively more homogeneous ﬁber population in soleus, consisting of ﬁbers containing intermediate mitochondrial levels. Magniﬁcation in all is 200x.

Patient 1

Patient 2

Patient 3

CI

CII

CIII

CIV

CV

PDH

Chapter 22, Figure 5 Immunohistochemical analysis of mitochondrial dysfunction. Fibroblasts from three patients with Leigh’s syndrome, due to three different nuclear DNA mutations, are labeled with a porin mAb (red) as a mitochondrial marker, and a second mAb (green) against the OXPHOS complex I, II, III, or IV, ATP synthase, or PDH indicated. Nuclei are stained with DAPI (blue). The merged red, green, and blue images are shown. Cells with a reduced labeling of a particular mitochondrial complex appear red, while normal levels of a particular complex appear yellow.

A

B

Chapter 22, Figure 9 Oil red O stain for lipid in normal rat liver (A), and showing the lipid accumulation characteristic of microvescular steatosis (B). In this case the steatosis was induced by a drug in development that potently inhibited OXPHOS complex V at submicrometer concentrations.

Chapter 22, Figure 17 Histological appearance of ragged-red ﬁber in patient with a mitochondrial myopathy. This image is from frozen muscle sections stained with Gomori trichrome stain. The red color of these ﬁbers is due to large numbers of subsarcolemmal mitochondria that have proliferated to compensate for repressed OXPHOS. The abnormal ﬁbers appear coarse and disorganized. Molecular weight (kDa) 150 75

Superoxide dismutase 2 Fold change: -2.17

50

Sod 2+/+

Sod 2+/+

25 20

Superoxide dismutase 1 Fold change: -1.77

15 Sod 2+/+ 3

Isoelectric point (pI)

Sod 2+/+

11

Chapter 23, Figure 1 Two-dimensional electrophoresis proﬁling of hepatic mitochondrial proteins from Sod2+/+ (wild-type, shown left) and Sod2+/− mice. Proteins were separated on the ﬁrst dimension using pH 3 to 11 nonlinear immobilized pH gradient (IPG) strips, followed by second-dimension separation by SDS-PAGE on 13% polyacrylamide gels. A −2.17-fold decrease in superoxide dismutase 2 (SOD2) protein relative to wild-type mice corresponds to a 50% reduction in SOD2 activity in the heterozygous Sod2+/− mouse. Superoxide dismutase 1 (SOD1) protein was found to be less abundant relative to wild-type Sod2+/+ mice.