Advances in DNA Sequence Specific Agents, Volume 3

ADVANCES IN DNA SEQUENCE-SPECIFIC AGENTS Volume3 9 1998 This Page Intentionally Left Blank ADVANCES IN DNA SEQUEN...

Author: M. Palumbo

49 downloads 696 Views 14MB 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

ADVANCES IN DNA SEQUENCE-SPECIFIC AGENTS

Volume3

9 1998

This Page Intentionally Left Blank

ADVANCES IN DNA SEQUENCE-SPECIFIC AGENTS Series Editor:

GRAHAM B. JON ES

Department of Chemistry Clemson University Clemson, South Carolina

Volume Editor: MANLIO PALUMBO Department of Pharmaceutical Sciences University of Padova Padova, Italy VOLUME3

9 1998

( ~ ) JAI PRESSINC. Greenwich, Connecticut

London, England

Copyright 0 1998 by JAI PRESSINC. 55 Old Post Road, No. 2 Greenwich, Connecticut 06836 ],41 PRESSLTD. 38 Tavistock Street Covent Garden London WC2E 7PB England

All rights reserved. No part of this publication may be reproduced, stored on a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, filming, recording, or otherwise without prior permission in writing from the publisher. ISBN: 0-7623-0203-8 ISSN: 1067-568)( Manufactured in the United States of America

CONTENTS

LIST OF CONTRIBUTORS PREFACE

Graham B. Jones and Manlio Palumbo

vii ix

INTRODUCTION TO DNA SEQUENCE-SPECIFIC AGENTS

Manlio Palumbo

SEQUENCE-SPECIFIC POISONS OF TYPE II DNA TOPOISOMERASES

Giovanni Capranico, Monica Binaschi, Maria E. Borgnetto, Mariagrazia Cornarotti, Emanuela Zagni, Manlio Palumbo, and Franco Zunino

TOPOISOMERASE I-TARGETING DRUGS: NEW DEVELOPMENTS IN CANCER PHARMACOLOGY

Barbara Gatto and Leroy Fong Liu

DNA SEQUENCE RECOGNITION ALTERED BIS-BENZlMIDAZOLE MINOR-GROOVE BINDERS

J. William Lown

39

67

SEQUENCE-SPECIFIC RECOGNITION AND MODIFICATION OF DOUBLE-HELICAL DNA BY MINOR-GROOVE BINDING CONJUGATES STRUCTURALLY RELATED TO NETROPSIN AND DISTAMYCIN

Christian Bailly

97

NEW DEVELOPMENTS IN THE USE OF NITROGEN MUSTARD ALKYLATING AGENTS AS ANTICANCER DRUGS

William A. Denny

157

vi

CONTENTS

DNA BINDING OF NONCLASSICAL PLATINUM ANTITUMOR COMPLEXES

Nicholas Farreil

179

KEDARCIDIN AND MADUROPEPTIN: TWO NOVEL ANTITUMOR CHROMOPROTEINS WITH SELECTIVE PROTEASE ACTIVITY AND DNA CLEAVING PROPERTIES

Nada Zein and Daniel R. Schroeder

201

ANTISENSE- AND ANTIGENE-BASED DRUG DESIGN STRATEGIES IN ONCOLOGY

Karl-Heinz Altmann, Doriano Fabbro, and Thomas Geiger

SEQUENCE-SPECIFIC RECOGNITION OF DOUBLE-STRANDED DNA BY PEPTIDE NUCLEIC ACIDS

Peter E. Nielsen

INDEX

227

267 279

LIST OF CONTRIBUTORS

Karl-Heinz Altmann

Novartis Pharma Inc. Basel, Switzerland

Christian Bailly

Institut de Recherches sur le Cancer de Lille INSERM Lille, France

Monica Binaschi

Division of Experimental Oncology Istituto Nazionale per Io Studio e la Cura dei Tumori Milan, Italy

Maria E. Borgnetto


Giovanni Capranico


Mariagrazia Cornarotti

DAKO SpA Milan, Italy

William A. Denny

School of Medicine University of Auckland Auckland, New Zealand

Doriano Fabbro

Oncology Department Ciba-Geigy Ltd. Basel, Switzerland

Nicholas Farrell

Department of Chemistry Virginia Commonwealth University Richmond, Virginia vii

viii

LIST OF CONTRIBUTORS

Barbara Gatto

Department of Pharmaceutical Sciences University of Padova Padova, Italy

Thomas Geiger

Oncology Department Ciba-Geigy Ltd. Basel, Switzerland

Leroy Fong Liu

Department of Pharmacology UMDNJ-R.W. Johnson Medical School Piscataway, New Jersey

J. William Lown

Department of Chemistry University of Alberta Edmonton, Alberta, Canada

Peter E. Nielsen

Department of Medical Biochemistry and Genetics The Panum Institute Copenhagen, Denmark

Manlio Palumbo

Department of Pharmaceutical Sciences University of Padova Padova, Italy

Daniel R. Schroeder

Pharmaceutical Research Institute Bristol-Myers Squibb Wallingford, Connecticut

Emanuela Zagni

Sandoz Prodotti Farmaceutici SpA Milan, Italy

Nada Zein

Oncology Drug Discovery Pharmaceutical Research Institute Bristol-Myers Squibb Princeton, New Jersey

Franco Zunino

Division of Experimental Oncoiogy Istituto Nazionale per Io Studio e la Cura dei Tumori Milan, Italy

PREFACE DNA sequence specificity plays a critical role in a number of biological processes, and influences a diverse range of molecular recognition phenomena, including protein-DNA, oligomer-DNA, and ligand-DNA interactions. This volume is intended to give the reader an up-to-date view of both the current status of, and developments to be expected in the near future, in research involving DNA interactive antitumor agents. In line with the intent of the series, special emphasis is thus placed on issues connected with sequence specificity and molecular recognition. Whereas volumes 1 and 2 were divided into subsections covering both analytical methods and applications, in this volume the entire focus is devoted to the macromolecule target specificity of DNA interactive developmental therapeutic agents of current interest. A brief introduction to DNA interactive anticancer agents is included for readers who may benefit from an overview surrounding the developments that have contributed to our general understanding of this field. The following nine chapters have been carefully chosen so that they describe topics which are at the forefront of development in DNA-targeted cancer chemotherapy. Issues that have been addressed include the mechanisms of selective DNA topoisomerase I and II poisoning by antitumor agents (Chapters 1 and 2), sequence-specific recognition of DNA by groove-binding drugs and drug-conjugates (Chapters 3 and 4), recent developments in nitrogen mustard alkylating agents and their potential use for antibody-directed enzyme-prodrug therapy (Chapter 5), nonclassical platinum

x

PREFACE

anticancer complexes, including dinuclear and trans-platinum derivatives (Chapter 6), DNA cleaving antitumor chromoproteins containing reactive enediyne moieties, which exhibit interesting free-radical chemistry along with selective targeting (Chapter 7), the potential of new sequence-specific antisense and antigene therapy in oncology (Chapter 8), and finally the conceivable chemotherapeutic use of mimetics of the DNA structure, obtained by substitution of the sugar-phosphate natural chain with a peptide backbone, the so-called peptide nucleic acids (Chapter 9). Important approaches being currently investigated for selective cancer treatment, such as gene therapy and immunochemotherapy, are not discussed in this volume since they fall beyond its scope. Graham B. Jones Series Editor Manlio Palumbo Volume Editor

INTRODUCTION TO DNA SEQUENCE-SPECIFIC AGENTS

Manlio Palumbo

I. Need for Specific Anticancer Drugs . . . . . . . . . . . . . . . . . . . . . . . . II. Nucleic Acid Binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III. The Search for Specificity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Short-Range Specificity . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Long-Range Specificity . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Newer Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 4 5 5 6

I. NEED FOR SPECIFIC ANTICANCER DRUGS Cancer is a genetic disease characterized by impaired regulation of cell proliferation and differentiation mechanisms. Normal cells contain specific genes, the protooncogenes, which encode proteins involved in biological processes, e.g. signal transduction and control of gene expression. These growth-promoting genes are counterbalanced by growth (tumor)-suppressing genes: the delicate equilibrium between activation and suppression allows a correct progression of the cell through its vital cycle. Mutations which potentiate proto-oncogene functions, turning them into oncogenes, induce uncontrolled growth in tumor cells. Similarly, DNA damage

Advances in DNA Sequence-Specific Agents Volume 3, pages 1-6 Copyright 9 1998 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0203-8

2

MANLIO PALUMBO

at the level of tumor suppressor genes lets cell proliferation proceed unconstrained, thus allowing development of malignancy. The hyperproliferative effects due to proto-oncogenes are generated in a number of ways in cancer cells. Proto-oncogenes may undergo point mutations in their sequence or they can be overexpressed as a result of amplification or translocation mechanisms. In the former case they produce "wrong" signal proteins, which impair the signal-transduction pathway; in the latter case the product of the oncogene differs only in the amount, not in the nature, from the product of the proto-oncogene found in normal cells. These observations clearly indicate how difficult it is to find a selective target for cancer chemotherapy. Many of the drugs used today for treating cancer patients are in fact practically nonselective and exhibit severe toxicity to normal tissues. Another limiting aspect of the current pharmacological therapy of cancer is the onset of resistance phenomena, which in many cases reflects an acquired capacity of developing defense mechanisms against drug treatment (multidrug resistance). Finally, chemotherapy with the presently available drugs can be valid for certain types of tumors but inadequate for common solid tumors. Hence, the requirement for novel effective and specific antineoplastic agents is stringent and pressing, as witnessed by intense efforts by many industrial and academic research teams.

II.

NUCLEIC A C I D B I N D I N G

Since DNA is the depository of genetic information and alterations in its expression are responsible for disease, the search for anticancer drugs directed against DNA, DNA processing enzymes, or their complexes has represented a logical and fruitful approach. Indeed, of the compounds widely employed in clinical practice since the late 1940s or presently in advanced clinical trials, a vast majority interact with DNA or with enzyme-DNA conjugates (Table 1). Even though antimetabolites are listed as "other drugs" in Table 1, they are also related to nucleic acids as they interfere with enzymatic reactions involved in DNA or RNA synthesis. The results of a very large number of investigations show that reversible binding to DNA, which may occur by intercalative or groove-binding mechanisms, is the initial important step of a cascade of chemical and biochemical events causing cell cytotoxicity. Subsequent to these reversible interactions, the DNA-damaging events generally occur. There are a number of ways DNA-directed anticancer drugs injure the nucleic acid. They can be reactive per se and generate damaging species such as free radicals or carbonium ions, which cleave or alkylate DNA, or they can mediate the transfer of a free radical to other species able to degrade the nucleic acid including oxygen or hydroxyl groups. A more elaborate mechanism involves the DNA cutting/rejoining enzymes of the topoisomerase family. In this case drugs interfere with an enzyme-DNA cleavable complex, thereby inducing DNA strand scission. The modes of DNA damage for clinically useful DNA binding anticancer agents are summarized in Table 2.

Introduction

3 Table 1.

ChronologicalDevelopmentof Anticancer Drugs

Drugs Interacting with DNA (or DNA-Enzyme Complex)

Other Drugs

1940s Nitrogen Mustard 1950s Methanesulfonates Busulfan Chlorambucil Cyclophosphamide 1960s Melphalan Actinomycin D

Methotrexate Mercaptopurine

Fluorouracil Vinca Alkaloids Thioguanine Cytosine Arabinoside

1970s Bleomycin Mitomycin Doxorubicin Dacarbazine Nitrosoureas cis-Platin 1980s Epipodophyllotoxins Mitoxantrone Carboplatin Ifosfamide 1990s Camptothecins" Enediynes a

Interferon

Taxola

Note: aDrugsundergoingclinicaltrials.

During the past few years important advances have been made in understanding the basic requirements to promote effective interactions in the drug-DNA complex, and in characterizing its structural features at a molecular level. In addition, biochemical information has become available on ternary interactions involving drug, nucleic acid, and nucleic acid processing enzymes. These include DNA and RNA polymerase, and, most relevant to anticancer activity, DNA topoisomerases. Although the initial events of drug-mediated DNA damage are now relatively well understood, a wealth of knowledge is still missing to connect them to the subsequent

4

MANLIO PALUMBO Table 2. Type of DNA Damage Generated by Clinically Useful

Anticancer Drugs Interacting with DNA

Drug

Type of DNA-Damage

Busulfan Chlorambucil Cyclophosphamide Dacarbazine Ifosfamide Melphalan Methanesulfonates Mitomycin Nitrogen Mustard Nitrosoureas

Alkylation of base nitrogens

Bleomycin Enediynes

Free radical strand-cleavage

Carboplatin

Platinum coordinaiton to purine nitrogens

cis-Platin Actinomycin D

Topoisomerase II-mediated strand-cleavage Interference with DNA processing enzymes

Amsacrine Doxorubicin Epipodofillotoxins Mitoxantrone

Topoisomerase II-mediated strand-cleavage

Camptothecins

Topoisomerase l-mediated strand-cleavage

cytotoxic events, which finally lead to cell death. In addition, as more is learned about the cell cycle and proliferation, other nucleic acid-related targets for chemotherapy are emerging that will require the development of new therapeutic strategies. Among the most promising appears to be telomerase, as it prevents the shortening of telomers and is activated in cancer tissues, but not in somatic tissues.

III. THE SEARCH FOR SPECIFICITY The development of antineoplastic agents possessing a selective and targeted action at the DNA level would represent a major advance in the pharmacological treatment of cancer. Sequence specificity should be considered at two different levels: short-range specificity, which is characteristic for low molecular weight ligands and allows preferential binding to 2-5 consecutive DNA base pairs; and long-range specificity, which may provide effective recognition of one DNA sequence in the

Introduction

5

entire genome. The latter requires matching of the ligand to a minimum of 15-20 base pair regions, i.e. to about 2 helical turns. Both types of specificity are relevant to DNA binding anticancer drugs.

A. Short-Range Specificity It is known that many intercalators and groove binders do not interact with DNA at random, but usually exhibit well-defined preferences for particular base combinations. The location of drug-induced DNA damage is obviously related to the location of the drug along the DNA chain. Different specificities would therefore generate chemical modifications (nitrogen base/strand-cleavage) at different positions, each lesion exhibiting its own biochemical and pharmacological properties. A similar difference in specificity is observed for topoisomerase poisons. Indeed, each family of drugs, characterized by distinctive structural and electronic features, exhibits individual preferences in stimulating/repressing enzyme-mediated DNA cleavage. Again, these differences could explain at least in part the wide range of cytotoxic responses observed for drugs having the same mechanism of action. The field of short-range specificity is therefore, still of general interest, and important advances are likely to be uncovered in the near future. In particular, modulation of sequence specificity of groove binders and development of hybrid molecules, combining different pharmacophoric groups or nonspecific reactive groups with DNA-specific binders, will be surely relevant for producing new, more effective, and less toxic cancer chemotherapeutic agents.

B. Long-Range Specificity Although a full understanding and mastering of short-range specificity will be valuable for directing covalent DNA modification, free radical damage, and topoisomerase-mediated poisoning, the achievement of in vivo recognition of DNA sequences in the 15-20 base pair range might represent a more significant advance to the therapeutic potential of DNA-damaging agents. The present approach to this level of specificity involves antisense and antigene strategies, according to which single- or double-stranded regions of DNA (or RNA) can be recognized by a complementary sequence forming either Watson and Crick or Hoogsteen hydrogen bonding with the target sequence. The knowledge base in this area is growing very quickly, and we may expect pharmacological applications in man in the not too distant future. However, the problems to be solved to fully reach this goal (delivery, stability, specificity, toxicity) are still formidable. Another way to approach the issue of long-range specificity could take advantage of the fact that gene expression is regulated by selective binding of appropriate peptide sequences in regulatory proteins to target DNA. Thus, DNA-specific drugs could involve peptides, rather than oligonucleotides as the recognition elements. This kind of approach is still in its infancy and requires a more thorough knowledge of the interplay between recognition elements from the nucleic acid and from the

6

MANLIO PALUMBO

protein and of the conformational changes occurring when the macromolecules interact. Peptidomimetics will surely represent a valid class of future drugs in this field of cancer selective agents. Not only can oligonucleotides and oligopeptides be used as such, they can also be potentiated by linking to them pharmacophoric groups able to generate permanent damage to DNA, as already indicated by the development of numerous conjugates between alkylators or strand breakers and antisense oligonucleotides.

IV. NEWER STRATEGIES Complimenting the traditional design methods for specific and effective drugs, newer strategies are also being successful in the field of DNA- and protein-specific agents. In particular the development of methodologies to generate peptide, oligonucleotide, or, in general, drug libraries and to screen them against a given target macromolecule (the so-called irrational drug design or epitope targeting) is providing us with new important tools to yield selective candidates for chemotherapeutic applications. Since viable targets have been identified, the design and evaluation of DNA sequence-specific agents continues at a fast pace. The following chapters highlight major advances in this field from a clinical perspective, and will also elude to important questions which remain to be answered in this rapidly developing area.

SEQUENCE-SPECIFIC POISONS OF TYPE II DNA TOPOISOMERASES

Giovanni Capranico, Monica Binaschi, Maria E. Borgnetto, Mariagrazia Cornarotti, Emanuela Zagni, Manlio Palumbo, and Franco Zunino I. II. III.

IV.

V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Structure and Function of DNA Topoisomerase II . . . . . . . . . . . . . . . . A. Amino Acid Sequence Homology . . . . . . . . . . . . . . . . . . . . . . B. Enzyme Phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . . . C. E n z y m e Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Drug-Resistant Topoisomerase II Mutants . . . . . . . . . . . . . . . . . . . . A. Prokaryotic Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Eukaryotic Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D N A Topoisomerase II Poisons . . . . . . . . . . . . . . . . . . . . . . . . . . A. Structural Determinants of Classical Poisons . . . . . . . . . . . . . . . . B. Non-cleaving Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Development of New Topoisomerase II-Directed Drugs . . . . . . . . . .


8 8 12 13 15 16 17 17 18 19 19 26 26

8

GIOVANNI CAPRANICO ET AL.

VI. Mechanismsof Drug Cell Killing . . . . . . . . . . . . . . . . . . . . . . . . A. Cellularand Molecular Mechanisms . . . . . . . . . . . . . . . . . . . . B. GeneticMechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I.

29 29 30 32 33

INTRODUCTION

DNA topoisomerases are nuclear enzymes that regulate DNA topology during transcription, replication, and recombination processes, and are essential for the integrity of the genetic material. Mammalian DNA topoisomerase II is the cellular target of important antitumor drugs, such as intercalating agents, minor-groove binders, and others. These agents interfere with topoisomerases by forming ternary DNA-drug--enzyme complexes in which DNA strands are broken and proteinlinked. Two isoforms (170 and 180 kDa) of type II DNA topoisomerase are present in human and murine cells. The expression of these isozymes, encoded by two distinct genes, are differentially modulated during cell proliferation and development. In recent years, X-ray diffraction studies as well as molecular and genetic analyses have allowed progress in the understanding of the structure and function of DNA topoisomerases. Drug structural determinants have been partially elucidated by studying the DNA sequence specificity of drug-mediated stimulation of topoisomerase II DNA cleavage and by novel structure-activity relationship studies. This will eventually allow design of new compounds targeted to selected genomic regions. Moreover, recent investigations on topoisomerase II inhibitors that do not stimulate DNA cleavage demonstrate that these agents may act against topoisomerases with a novel mechanism, suggesting new screening strategies to identify potential antitumor drugs. The initial molecular event, i.e. formation of the drug-stabilized cleavable complex, is a reversible process, thus how it can ultimately lead to cell death remains a very active area of research. Irreversible double-stranded DNA breaks, generated by interaction of cleavable complexes with ongoing DNA-dependent processes, may trigger a cell death program. This chapter surveys recent advances in the molecular pharmacology of sequence-specific antitumor topoisomerase II poisons.

II.

BACKGROUND

From unicellular organisms to human cells, DNA topology is governed by a large family of highly conserved enzymes known as DNA topoisomerases. 1-8 Regulation of the topology of the DNA molecule is an essential cellular task, thus these enzymes are necessary for the viability of all organisms. Since the discovery of the first topoisomerase, E. coli DNA topoisomerase I (or 0 protein), 2'9 several others have been isolated and characterized from both prokaryotic and eukaryotic cells. 5'1~ Recently, the crystal structure of a central fragment of the yeast type II

Poisons of Type II DNA Topoisomerases

9

DNA topoisomerase was resolved at a resolution of 2.7 ~, providing important insights for the catalytic mechanism. 12 Topoisomerases alter DNA conformations by making reversible DNA breaks in which a tyrosyl residue of the protein is covalently linked to a DNA phosphoryl group. 1-3 These enzymes may be divided into two major groups" type I enzymes introduce a single-stranded DNA cut at a time; type II enzymes on the other hand break both of the two strands of a duplex DNA fragment. Following DNA breakage, the enzyme catalyzes the passage of a single-stranded (type I) or double-stranded (type II) DNA fragment through the break, and finally reseals at the cleavage point. The known DNA topoisomerase I enzymes appear to belong to two distinct subfamilies based on sequence homology and biochemical properties, whereas all known type II enzymes show extensive homology of the amino acid sequences, suggesting that they are evolutionarily and structurally related to each other. 2'13 Type II DNA topoisomerases and eukaryotic topoisomerase I are poisoned by several compounds, some of which are widely used in the therapy of human cancers and infectious diseases (Figures 1 and 2). Antitumor drugs, such as anthracyclines, demethylepipodophyllotoxins, amsacrines, mitoxantrone, actinomycin D, streptonigrin, terpenoids, flavonoids, and camptothecins are poisons of eukaryotic topoisomerases II or I (Table 1), while quinolones, antibacterial agents, are poisons of DNA gyrase and topoisomerase IV, the bacterial type II DNA topoisomerases. These drugs have been shown to stabilize a covalent intermediate of the topoisomerase catalytic cycle, in which DNA strands are broken and the enzyme is covalently linked to the DNA. Drug action thus results in DNA cleavage stimulation that is believed to be the molecular basis of their cytotoxic activity. 14-e2 In recent years, the list of compounds that can interfere with DNA topoisomerase function has indeed become considerably longer. Often the word "inhibitor" is used for any of these compounds, however this may generate some confusion since there is more than one mechanism of drug action. Many agents stabilize the covalent complex, therefore stimulating DNA cleavage; others inhibit the catalytic activity without cleavage enhancement by impeding access of topoisomerases to DNA sequences bound by the drug, and still others interfere with protein conformational transitions by binding to the enzyme and freezing one of the possible protein conformations. Therefore, we define as poisons only the cleaving compounds (Table 1), and as inhibitors only those that inhibit the enzyme catalytic activity but do not stabilize the covalent DNA-enzyme complex. A peculiarity of known topoisomerase poisons is that their action is invariably DNA sequence-specific. DNA topoisomerases commonly produce DNA breakage at selected sites in a given DNA fragment, and a poison acts by stabilizing DNA-enzyme covalent complexes at a subset of these sites. Interestingly, chemically unrelated poisons generally stimulate DNA cleavage at distinct site subsets, resulting in drug-specific DNA cleavage intensity patterns. Detailed structureactivity relationship studies performed in our laboratories and by other groups have established structural determinants of the drug sequence specificity. Moreover,

H-N

d

0

I

H ‘

ww BISANTRENE

NHfi

OH

0

N

MITOXANTRONE

H

DOXORUBICIN

I

O

D

TENIPOSIDE (VM-26) H

‘

N s.Q

/I\

0

H

OH

H N ‘

0

I

d /\/OH NH

AMSACRINE (mAMSA)

ELLIPTICINE

Figure 1. Chemical structures of classical DNA topoisomerase I I poisons.

WCHS 8 0

OH

CLEROCIDIN

0

PLUM BAG IN

INTOPLICIN

ICRF-193

SAINTOPIN

AZATOX I N

STREPTONIGRIN

Figure 2. Chemical structures of other DNA topoisomerase I I poisons and inhibitors.

12


Table 1. Poisons of Eukaryotic DNA Topoisomerases a TOPO I

TOPO I & II

TOPO II

Intercalating drugs

Indolocarbozoles (ED- 1 1 0 ) Fagaronine

Actinomycin D Saintopin Intoplicin

Anthracyclines Amsacrines Anthracene derivatives Ellipticines Quinolones Amonafide Flavones Naphthoquinones

Non-intercalating drugs

Camptothecin Bulgarein Hoechst 33258

Indoloquinolinediones Demethylepipodophyllotoxins (AzalQD) Terpenoides Streptonigrin RO- 15-026 Isoflavones Catechins Azatoxin

Note: aMore information on these compounds can be found in 12o,148,149and references therein. Original data on

fagaronine and cathecins have been reported in Refs. 171,226, 227.

these investigations and cross-linking data with photoactivatable poisons have conclusively demonstrated that the drug receptors are located at the protein/DNA interface of the DNA cleavage site. Finally, knowledge of the high heterogeneity of topoisomerase II-mediated DNA cleavage in the chromatin of living cells has suggested that the sequence specific action of the poisons may have profound consequences for drug activity (vide infra). In past years, progress has been made in the elucidation of the mechanism of action of these agents that may be of help in the development of more effective poisons and inhibitors of type II DNA topoisomerase. The structural information on topoisomerase poisons may also be of great interest in the general search for sequence-specific DNA binders able to repress gene transcription at selected genomic loci. In this chapter we will focus on the biological activity and molecular action of the sequence-specific poisons of topoisomerase II, and on the structure and functions of the enzyme relevant to the action of antitumor poisons.

III.

STRUCTURE AND FUNCTION OF DNA TOPOISOMERASE II

Eukaryotic topoisomerase II is composed of two identical subunits, and it has been proposed to act as an ATP-dependent molecular clamp. 12'23-25 The enzyme can utilize an ATP-induced conformational change to capture a second DNA segment


13

to be transported through the DNA break, and this is thought to be achieved by forming a circular clamp around the double helix. 23 The catalytic cycle might thus be divided into the following steps: (1) the non-covalent binding of the enzyme to the DNA, in which the enzyme is in an open form of the clamp; (2) the enzyme then produces a double-strand break, in which each subunit remains covalently attached to a 5' phosphoryl terminus (base + 1), with each of the two 3' termini (bases-1) recessed by four bases; (3) ATP binding causes an allosteric conformational change that closes the protein clamp after the capture of a second DNA segment; (4) this intact DNA segment is then transported through the broken DNA and allowed to exit from the protein; and (5) ATP hydrolysis restores the open form of the protein clamp. 12'23-25 Experimental evidence and structural data supported the idea that topoisomerase II transports one DNA segment through another by a two-gate mechanism. The DNA segment being transported would exit from the inside of the protein molecule through a gate at the opposite side of the entrance gate. ~2,24,25 Two distinct isozymes of topoisomerase II are present in mammalian cells: topoisomerases Iltx and 1113that have different molecular weights, 170 and 180 kDa, respectively. 26 These isozymes are encoded by two genes mapped to human chromosomes 17 and 3, respectively. 27'28 The a form may be associated with cell proliferation, while the 13form is expressed also in cultured quiescent cells and in nonproliferating normal mouse tissues. 27-31 Although the cellular functions of topoismerase II isoforms may be different, their role in the antitumor activity of the poisons have been debated in recent years. Recently, human topoisomerases Ilct and III3 were purified from yeast cells overexpressing the corresponding plasmid-borne cDNA, and their drug sensitivities were measured with a DNA cleavage assay using 32p-labeled SV40 DNA fragments. 32 Both isozymes were sensitive to VM-26, mAMSA and dh-EPI, that stimulated similar cleavage intensity patterns in agarose and sequencing gels. Sequence specificities of the studied poisons were identical for both isozymes, and corresponded to those established previously for the native murine enzyme. 32 The data suggested that molecular drug interactions in the ternary complex are likely similar between the two isozymes, and that both topoisomerase II~ and 1II3 may be the cellular target of antitumor poisons. 3e

A. Amino Acid Sequence Homology A comparison of the amino acid sequences of several eukaryotic and prokaryotic type II topoisomerases provided evidence for a high degree of homology even between evolutionarily distant proteins. 13'33 The observation that common sequence motifs have been conserved during evolutions at corresponding positions in the various peptides strongly suggests that all type II topoisomerases are evolutionarily and structurally related. Investigations on prokaryotic enzymes may therefore provide information relevant for the eukaryotic enzymes as well. The

14


functional conservation of these enzymes is noteworthy: DNA topoisomerases II ofS. pombe, Drosophila, mouse, and the human topoisomerase IIa can complement a top2 mutant in S. cerevisiae. 34-37Moreover, a conditional overexpression of the wild-type Drosophila enzyme can restore the drug sensitivity of hamster cells resistant to drugs due to an endogenous mutated topoisomerase 11.38 The homology analysis of amino acid sequences provided evidence that the enzyme is likely constituted by three major domains: an N-terminal domain with the ATP binding site, a central portion with DNA breakage-rejoining activity, and a C-terminal domain that appears to be important for the regulation of enzyme activity. 13 Susceptibility to protease digestion also defines at least three domains. 39'4~A particularly interesting proteolysis site in the yeast topoisomerase II corresponds approximately to the junction of the Gyr A and Gyr B subunits of DNA gyrase. The proteolysis site is dependent on the presence of a nonhydrolyzable ATP analogue, 39 and may reflect a conformational change, such as the closing of the protein clamp (see above). 24 Wigley and colleagues reported the crystal structure of an N-terminal fragment of Gyr B, a protein subunit of a bacterial topoisomerase II, DNA gyrase. 41 The N-terminal fragment corresponds to the ATPase domain of DNA gyrase, and the authors demonstrated a central hole in the protein, the diameter of which is enough to accommodate a DNA duplex. 41 The amino acid sequence from 103 to 126 of Gyr B participates in the ATP binding site and includes a sequence motif conserved in all type II enzymes. In the human sequence, amino acids 149 to 172 correspond to amino acids 103 to 119 of Gyr B. 13'41 Three amino acid motifs of the central domain, EGDSA, PL(R/K)GK(I/L/M)LN, and IM(T/A)D(Q/A)D, are conserved in all known topoisomerase II sequences. 13 Interestingly, the EGDSA and PLRGK amino acid sequences may correspond to loops found in the structure of the T~i-resolvase, a protein that transfers a DNA strand to an acceptor DNA end. 13 Insertions of short linkers near these motifs reduce the enzyme activity, 42 and mutations within or close to the PLRGK sequence may alter the enzymes sensitivity to drugs 43-45 (vide infra). Taken together, these results strongly suggest that the conserved regions of the topoisomerase II sequence are involved in DNA binding and/or in the DNA breakage-reunion reaction, and are implicated in the drug binding to the enzyme/DNA complex. 13 The topoisomerase II region most sensitive to protease digestion defines the beginning of the C-terminal domain. 39'4~Regulatory functions of enzyme activity have been ascribed to the C-terminal one third of topoisomerase II. 46,47 Interestingly, this enzyme portion is the least homologous among known topoisomerases II. The relatively low degree of homology does not allow identification of large structural motifs in the C-terminus. In the case of the two human isozymes, p 170 and p 180, sequence homology is lower in the C-terminal quarter (46%) than in the first three-quarters (86%), 3~ thus suggesting that the C-terminal part of the protein may mediate different functions of the two isozymes in living cells.


15

B. EnzymePhosphorylation Nuclear localization signals and phosphorylation sites have been mapped to the C-terminal part of the protein. 46--49 Casein kinase II and protein kinase C were shown to be responsible for the phosphorylation of serine and threonine residues in the last 200 amino acids, therefore determining specific activity, and modulating both the mitotic functions and drug sensitivity of DNA topoisomerase 1I.50-52 Drug stimulation of DNA cleavage may be attenuated by enzyme phosphorylation, possibly due to faster enzyme turnover and ATP hydrolysis. 53-55 Indeed, it has been found that phosphorylation of topoisomerase II was increased 15-fold in a human KB cell line selected for resistance to VP-16. 56 Hyperphosphorylation could thus be a simple way for cultured cancer cells to adapt to growth in the presence of topoisomerase II poisons. Since the C-terminal portion of topoisomerase II is not essential for enzyme activity in vitro, and to a lesser extent also in vivo, 46'47 this seems to contrast with the apparent increase of the enzyme activity following phosphorylation in this portion of the protein. Therefore, it has been proposed that the C-terminal domain may act as a negative regulatory domain of enzyme activity, neutralized by phosphorylation. 57'58 At the same time, phosphorylation has been shown to affect DNA-protein as well as protein-protein interactions. 59'6~ The specificity of topoisomerase II action during the cell cycle suggests that the enzyme activity might be temporally regulated. This regulation could involve posttranslational modification of the protein. In fact, topoisomerase II is known to exist as a phosphoprotein in cells from both lower and higher eukaryotic species. Topoisomerase II from both Drosophila and budding yeast have been shown in vitro to be substrates for casein kinase II. 5~ Wells et al. 61 showed that human topoisomerase IIo~is phosphorylated at multiple sites in vivo and two of these sites are substrates for phosphorylation by casein kinase II in vitro. Topoisomerase IIo~ has been shown to be phosphorylated in vivo in a cell cycle-dependent manner with maximal phosphorylation occurring in the G2/M or specifically M phase. 51'62-64The distribution of the phosphate among multiple sites varies between G1 and M phase in yeast. 51 Chinese hamster ovary (CHO) topoisomerase IIc~ is primarily phosphorylated at a serine residue (with a minor phosphothreonine occurring in mitosis). 64 Determination of the sites of phosphorylation present only in interphase or only in mitosis are interesting in order to determine their effects on the function of the enzyme. In fact a change in the phosphorylation of topoisomerase IIc~ may be involved in the resistance of human leukemia cells to VP-16 and mAMSA. 65'66 Topoisomerase 1113also exists in vivo as a phosphoprotein 67 and in mitotic HeLa cells has a reduced electrophoretic mobility relative to cells at interphase. 68 Burden and Sullivan 69 showed that topoisomerase 1I[3 is posttranslationally modified in vivo in a cell cycle-dependent manner and it has been noted that the 13-isozyme is qualitatively different in mitosis when the nucleolus is disassembled. Phosphorylation of topoisomerase 1113in mitosis might be required for the stabilization of the enzyme during this process.

16


C. Enzyme Functions Topoisomerase II has been proposed to be an important factor that governs chromatin structure in living cells. The chromatin of eukaryotic cells is organized into structural loops that are attached to the nuclear matrix and the chromosome scaffold via specialized DNA elements known as SARs. DNA loops are dynamic structural units of chromatin, whose size and properties can vary among distinct genomic regions and during the cell cyle. 7~ Topoisomerase II might regulate chromatin structure by specific interactions with SARs, 70'72'73 which are enriched for sequences related to the in vitro topoisomerase II cleavage consensus 7~ and are preferentially cleaved by the enzyme in vitro. 74-77 Moreover, topoisomerase II preferentially aggregates SAR-containing DNA fragments through cooperative interactions. 74 Type II DNA topoisomerase has a vital role in the decatenation of daughter DNA helices at the end of replication and in the segregation of intertwined chromosomes. 78'79 Studies in a cell-free SV40 replication system suggest a role for DNA topoisomerase II in the elongation stages of DNA replication, 8~ nevertheless topoisomerase II does not seem to be required for the early stages, but it is necessary for the resolution of late replication intermediates. In support of this, S. cerevisiae and S. pombe top2 mutants accumulate multiply-intertwined and catenated dimers derived from newly replicated circular plasmids. 36'78'81 Indeed, the topoisomerase II gene has been shown to be essential in yeast. In the absence of an active enzyme, small chromosomes can apparently resolve intertwinings through simple unravelling of their ends, while larger chromosomes are more inclined to fragment. 82 Also, in Drosophila embryos microinjection of either antibodies against topoisomerase II or the drug VM-26 can inhibit the separation of chromosome daughter sets at anaphase. 83 Catenated dimers and late Cairns-type replication intermediates accumulate during the replication of SV40 DNA in vitro when topoisomerase II is inhibited or depleted from extracts 8~ or inhibited in vivo. 84'86 Interestingly, the requirement for topoisomerase II in chromosome condensation has been demonstrated in vitro by specific immunodepletion. 87'88 Using interphase nuclei of somatic cells and Xenopus egg extracts, chromosome condensation has been shown to be closely correlated with the level of the enzyme. Moreover, studies with ICFR- 193, a noncleavable complex forming inhibitor, showed that this drug mimics the phenotype of yeast top2 mutants by causing cell death during the G2/M phase. 85'89 Downes et al. 9~ showed that in mammalian cells ICRF-193 induced cell-cycle arrest is modulated through a system different from the damage-induced G2 arrest, and sensitive to the decatenation state of DNA. It is uncertain whether a similar system exists in yeast. Temperature-sensitive top2 mutants of S. pombe enter an abortive mitosis with very elongated chromosomes. They conclude that mammalian cells possess a G2 checkpoint system sensitive to the catenation state of DNA or to topoisomerase II activity either directly or through detection of consequent


17

changes in chromatin. In meiosis, top2 mutants enter arrest prior to meiosis I with no evidence of nuclear elongation. 91 This result suggests that meiotic cells differ from mitotic ones in having a checkpoint mechanism for ensuring that sufficient active topoisomerase II is present for efficient chromosome segregation.

IV. DRUG-RESISTANT TOPOISOMERASE II MUTANTS In recent years, several laboratories have investigated mutated DNA topoisomerases II that are resistant to their poisons. These studies were invaluable, allowing understanding of the structural enzyme domains directly responsible for the topological transformation of DNA molecules, and the amino acid residues that are critical for the formation of the ternary complex: DNA-drug-enzyme. The information of the specific amino acids important for effective drug interactions may certainly be useful for the rational design of more active compounds.

A. Prokaryotic Enzymes The very first studies were carried out on a eubacterial topoisomerase II, DNA gyrase, by mutagenesis of wild-type bacterial strains and then selection for resistance to quinolones, poisons of DNA gyrase. A large number of E. coli gyrase mutants have been identified with this strategy, 92 and they mainly belong to three groups: (1) mutations within the EGDSA motif of GyrB subunit (the D residue changed to N); (2) mutations in the PLKGKILN motif of the same subunit (the first K to E); and (3) mutations at Ser-83 of GyrA subunit--this amino acid residue is close to the catalytic Tyr- 122. Thus, amino acid mutations responsible for quinolone resistance were clustered in few regions of either GyrA or GyrB subunits. On the basis of computer predictions, these motifs correspond to hydrophilic regions of the polypeptide chains, likely situated on the protein surface, that may be important for drug and DNA interactions. Using gyrases reconstituted from both of the two mutated subunits, it has been demonstrated that the double-mutant enzyme was more resistant to oxolinic acid than either the single-mutant protein. 92 Though studies on cultured bacterial strains have provided valuable insights, the mechanism of quinolone resistance has also been investigated using clinical isolates. 93'94Using PCR techniques, eight uropathogenic E. coli strains were analyzed and all of them exhibited high levels of resistance, and had a mutation of Ser-83 to Leu or Trp in the GyrA subunit. 93'94 S. aureus clinical isolates from drug-resistant patients were shown to carry a mutation of Ser-84 to Leu of GryA, this is the equivalent residue of Ser-83 of E. coli GyrA. Interestingly, E. coli DNA gyrase proteins resistant to microcin B 17 have also been reported. 95 Microcin B 17 is a bactericidal peptide whose mode of action resembles that of quinolones. Two independent mutations conferring resistance to microcin B 17 have been isolated, and mapped to the GyrB subunit. The two mutations consisted of the same transition AT to GC at position 2251 of the GyrB

18


gene, thus causing the change of Trp-751 to Arg. 95 The mutation lies in the C-terminal half of GyrB that is thought to be responsible for the interaction with GyrA. The fact that mutations conferring resistance to quinolones also lie in the region of subunit A binding (even if in the center of Gyr B) may suggest that drugs and microcin interact with the same region of gyrase.

B. EukaryoticEnzymes The results obtained in bacterial systems may be helpful in understanding the structure and function of mammalian topoisomerases II since these enzymes share several structural domains and show high degrees of amino acid sequence homology (vide infra). In a mammalian system, analogous studies are more difficult to carry out, therefore several laboratories have isolated and characterized drugresistant mutated yeast or human topoisomerases II by using permeability mutants of S. cerevisiae. 96 This system is very useful since the human DNA topoisomerase II can be expressed in a functional form in yeast, and mutational analyses of the enzyme can thus be performed. Indeed, studies in yeast have demonstrated that drug stabilization of the cleavable topoisomerase II-DNA complex is the cellular event that triggers cell killing by anti-topoisomerase II agents, 96 thus demonstrating that topoisomerase II is the cellular target of antitumor drugs including etoposide, amsacrines, and others. Jannetipour et al. 97 have characterized a yeast temperature-sensitive topoisomerase II mutation that confers resistance to amsacrine and etoposide. They have overexpressed the protein to confirm that the enzyme is altered in its interaction with drugs performing cleavage in vitro and, sequencing the gene, three mutations (Arg-884 to Pro, Arg-886 to Ile and Met-887 to Ile) have been discovered. Even if none of these amino acids are absolutely conserved during evolution, this mutant may define a region of topoisomerase II that plays a critical role for drug interaction. Another set of mutations which cause resistance to drugs has been obtained with site-directed mutagenesis of regions within a plasmid-borne yeast top2 gene, and hydroxylamine mutagenesis of the whole plasmid. 98 Three classes of mutants have been identified: 1. Multiple changes in the PLRGK (MLN) sequence (position 474-481 of yeast topoisomerase II). Although the PLRGK motif is highly conserved, replacements within the motif are therefore possible without enzyme inactivation. In contrast, mutagenesis in the EGDSA motif failed to give any resistant mutant since substitutions in this region likely result in inactive enzymes. 98 2. Mutation at A642, which is in a region not present in the T4 topoisomerase II that is sensitive to amsacrine. Therefore, this mutation might cause a change in the protein conformation that prevents interaction with drug. 98 3. Truncation of parts of the carboxy-terminal domain of the enzyme. Recent studies of 3'-end truncated Drosophila topoisomerase II have demonstrated


19

that deletions up to 240 amino acids had no effect on the catalytic activity of the enzyme and VM-26 cleavage stimulation. Thus, it seems unlikely that this domain is important for the stabilization of cleavable complexes. Since nuclear localization signals are present in the C-terminal part of the protein, it is possible that reduced nuclear concentrations of the enzyme is the cause of drug resistance. Immunofluorescence studies indeed showed that none of the truncation mutants localize to the nucleus as efficiently as the wild-type enzyme. 42 Similar results have been obtained by truncation of S. cerevisiae topoisomerase 1I.48 Topoisomerase II mutations have also been investigated in mammalian cell lines selected for resistance to topoisomerase II poisons. An important observation from these studies is that a large part of the mutations fall in a very conserved protein region (amino acids 449-494) that corresponds to E. coli GyrB mutations conferring resistance to quinolones. 43,99-101 These results further support the idea that the region may be involved in the interaction between topoisomerase II and its poisons. Thus, a variety of drugs (quinolones, amsacrines, and epipodophyllotoxins) are likely to be located in a similar cleft in their interaction with the DNA-protein complex. Another mutation has also been reported in an etoposide resistant human leukemic cell line. 1~ This mutation lies close to the catalytic Tyr-804 residue of the protein (Lys-797 to Asn) and may interfere with drug induced trapping of the cleavable complex. Additionally, for quinolone resistance, mutations which lie in a conserved region of gyrase A protein adjacent to catalytic Tyr-122 have been mapped. Thus, all the mutations in mammalian topoisomerase II co-map with mutations in gyrase.

V.

D N A T O P O I S O M E R A S E II P O I S O N S

A. Structural Determinants of Classical Poisons Topoisomerase poisons often are natural compounds derived either from plants or microorganisms. Agents interacting with topoisomerase II are indeed widespread in nature, suggesting an evolutionary-conserved phenomenon of interference (modulation ?) of topoisomerase activity. The biological significance of this remains to be fully appreciated. In the past, antitumor topoisomerase poisons were classified as intercalating or as non-intercalating agents (see Table 1). It is noteworthy that both DNA intercalating agents and non-intercalative compounds are known to be poisons of eukaryotic DNA topoisomerase II as well as topoisomerase I. At the same time, streptonigrin, a topoisomerase II poison, may interact in the minor groove of the DNA, 1~ and both bulgarein and various Hoechst derivatives, poisons of topoisomerase I, are also DNA minor-groove binders. 1~176 Thus, hypotheses on the mechanism of drug action that are based simply on different modes of drug interaction

20


with the DNA seem questionable. With some exceptions (such as the terpenoid clerocidin), drugs have an intercalating moiety or a planar aromatic system that might fit into DNA base pairs in an intercalation-like manner. Interestingly, some agents are "pure intercalators" (ellipticines); some are not intercalators (VM-26, camptothecins), including minor-groove binders (Hoechst 33548, streptonigrin); and some have features of both intercalators and external binders (anthracyclines, actinomycin D, indolocarbazoles). In this latter class of compounds, historically considered intercalating agents, the external binder moiety of the drug molecule is crucial for drug activity. 1~ Present knowledge, indeed, suggests a common mode of action (interference with the cleavage-reunion reaction) for both "intercalating" and "non-intercalating" poisons, and receptor sites for the known topoisomerase II poisons are likely overlapping (vide infra). It is generally believed that the antitumor poisons stabilize the covalent intermediate by forming a ternary complex: DNA--drug-enzyme. Teniposide and other drugs were shown to inhibit the DNA religation step of topoisomerase II reaction both before and after strand passage. 114,115This results in a decrease of the resealing rate of DNA cleavage in in vitro systems. 114'I~5 The data thus suggests that topoisomerase II poisons may act primarily by hindering the religation step of the catalytic reaction. Nevertheless, some topoisomerase II poisons have been suggested to accelerate the forward cleavage reaction instead, since they do not decrease the rate of cleavage resealing. 116'117 Antitumor drugs of different chemical classes typically stimulate topoisomerase II DNA cleavage in a sequence selective manner, resulting in drug-specific cleavage intensity patterns in sequencing gels. 17'118-12~Although sites of DNA cleavage in a given DNA fragment may be uncommon, relative cleavage intensities greatly vary among different compounds. Each drug is thus able to stimulate DNA cleavage preferentially at certain sites, but not at all sites, recognized by the enzyme, thus suggesting that effective drug interactions in the ternary complex depend on the local base sequence. Several investigations have indeed revealed that specific nucleotides flanking the strand cut are required for drug stimulation of the cleavage (Table 2). 103'120-126Analogous observations have been made with camptothecin that prefers a guanine at position + 1, which for topoisomerase I represents the nucleotide noncovalently linked to the protein. ~27'128Based on these findings a stacking model of drug action has been proposed. The model predicts that drugs in the ternary complexes are placed at the interface between the DNA cleavage site and the enzyme active site, possibly with a planar ring system stacking with DNA base pairs. 12~ Consistent with this model, drug localization in the ternary complex has recently been shown using a photoactivatable mAMSA analogue. 129 Upon activation, the compound was covalently linked to DNA bases at the +1 and -1 positions, immediately adjacent to the cleaved phosphodiester bond, when T4 topoisomerase II was present in the reaction mixture. This suggests that the amsacrine receptor site is localized at the protein/DNA interface of the enzyme active site in the topoisomerase II-DNA complex. 129


21

Table 2. Sequence Specificity of in Vitro DNA Cleavage Stimulated by DNA

Topoisomerase II Poisonsa

Complete Preferred Sequence Position Spec!ficity -1

Drug Doxorubicin, Daunorubicin 3"-epi-daunorubicin V P - 1 6, VM-26 Mitoxantrone

1(/+ 1) +1

-

+2

Piroxantrone Ellipticinium Amonafidea Clerocidin Genistein Amsacrine Bisantrene Saintopin NSC 665517 Streptonigrin

Primary Requirement

-3

-2

-1

+1

+2

References

A

(A)

T

A

(A)

m

121

A(G)

(A)

G

A(G)

(A)

~

C(T)

--

--

C(Y)

C/Th C/Th Tc Ce G T A A G G T

113 122,131)

--

--

~ ~ ~ A/T

~ ~ ~ A/G

C/T C/T T C

~ ~ ~ A

G G

130 125 123,125

~

A m --

A m --

T T T

T A A

134 133 125 122,125 125

m

139 228 .

.

.

.

T

I03

Notes: aln parenthesis, secondarypreferences hWith the exclusionof adenines "With the exclusionof cytosines dThe completepreferred sequenceis required at both strands eWith the exclusionof thymines

This m o d e l has been a useful f r a m e w o r k for novel structure-activity relationship studies a i m e d at identification of the structural (drug) d e t e r m i n a n t s of t o p o i s o m erase p o i s o n i n g and (drug) s e q u e n c e specificity. Interesting data were obtained by s t u d y i n g the s e q u e n c e specificity of m i t o x a n t r o n e and VM-26.13~ T h e two drugs s h o w e d similar cleavage intensity patterns in s e q u e n c i n g gels and the same local base preferences; moreover, in spite of structural similarity and c o m p a r a b l e D N A b i n d i n g affinity, s e q u e n c e selectivities of m i t o x a n t r o n e and d o x o r u b i c i n were c o m p l e t e l y different. 13~ T h e s e results strongly indicate that intercalating and nonintercalating agents have a c o m m o n m o d e of action against the e n z y m e . M o r e o v e r , c o m p e t i t i o n studies a m o n g several t o p o i s o m e r a s e II-interacting agents have provided similar information on the mechanistic effects of the drug. t31 C o m p e t i t i o n a m o n g D N A cleaving poisons, such as V M - 2 6 , amsacrine, genistein, CP- 115,953, and n o n c l e a v i n g inhibitors, such as novobiocin, suggested that the receptor site of V M - 2 6 m a y overlap with those of the poisons, whereas it is distinct from those of the inhibitors. 131 T h r e e main structural d e t e r m i n a n t s of the sequence-specific action of the poisons have been characterized by investigating the s e q u e n c e specificity of drug c l e a v a g e

22


stimulation: (1) the spatial conformation of drug molecule, that determines the sequence specificity; (2) particular drug substituents, that can determine preferred bases; and (3) cooperative phenomena, that may markedly influence the specificity of drug effects.

Drug Pharmacophores The observation that mitoxantrone and VM-26, two structurally unrelated drugs, have similar sequence specificity 13~strongly suggests that intercalating and nonintercalating agents likely bind to the same receptor site in the ternary complex. Bisantrene and mAMSA, although chemically unrelated, also stimulate identical cleavage intensity patterns and show the same specific base requirements for cleavage stimulation (adenine at position + 1).125 Since these drugs are structurally unrelated but have the same sequence specificity, these findings allowed us to make an attempt to identify structural determinants of drug action. Comparison of energy-minimized drug structures by computer modeling simulation suggests that bisantrene and mAMSA share a specific pharmacophore that may be an important determinant of the "sequence position specificity". 125 If topoisomerase II poisons have different specific pharmacophores, precise drug receptor sites might thus be only partly overlapping. The case of streptonigrin is particularly interesting in this regard. This compound has a unique sequence specificity since it requires a thymine (or an adenine) at position + 2 (or + 3). 1~ Streptonigrin does not intercalate into DNA, and independent experimental results are consistent with a DNA binding mode similar to that of minor-groove binders. 1~ Therefore, streptonigrin might be placed, in the ternary complex, externally to the DNA, interacting with the T in a groove, and interfering with the DNA religation step of the enzyme reaction. 1~ The streptonigrin receptor site might be only in part overlapping with those of other poisons, and its pharmacophore may then be quite distinct from those of mAMSA and other drugs. 1~

Chemical Substituents Investigations on both anthracycline analogues and clerocidin suggest that specific drug substituents may have instead a role in determining preferred base pairs. Closely related drug congeners generally stimulate identical DNA cleavage intensity patterns, 1~176 and, indeed, in the case of anthracyclines, several modifications of the planar ring and/or the daunosamine did not alter doxorubicinspecific intensity patterns of cleavage. ~1~ However, a daunorubicin derivative with the 3'-NH 2 of daunosamine epimerized, stimulated topoisomerase II DNA cleavage with different cleavage intensity patterns in agarose and sequencing gels (with respect to the parent drug). 113 Several cleavage sites were detected with 3'-epi-daunorubicin, but were not observed with daunorubicin or doxorubicin. A statistical analysis of cleavage sites showed that the major difference between the derivative and parent drugs was that a guanine, instead of a thymine, was preferred at position -2 by the former. These findings, for the first time, demonstrated that


23

specific (drug) functional groups may be important structural determinants of preferred base pairs. 113It is thus possible that specific analogues of other topoisomerase poisons may also exhibit different sequence selectivity as compared with the corresponding parent drug. Interestingly, an anthracycline analogue with no substituent at the 3' position of the sugar, was able to stimulate DNA cleavage at sites stimulated by parent drugs as well as at those stimulated by 3'-epi-daunorubicin. 113 Thus, the presence of an amino group at the 3' position of the daunosamine restricts the drug activity to either the "parent drug" sites or the "analogue" sites depending on its orientation. These findings were rationalized in terms of interference of the anthracycline amino group with the amino group of guanines in the minor groove, as suggested by theoretical investigations of drug-DNA interactions. 132 This would be predicted to result in the exclusion of guanines at position -2 of the cleavage site by anthracyclines with the natural configuration, ll3 These data also indicate that the sugar moiety of anthracycline, placed in the DNA minor groove, interacts with base pairs up to position-2 and, possibly, with enzyme amino acid residues. More recently, we have also investigated the sequence specificity of the terpenoid clerocidin, a peculiar topoisomerase II poison since it lacks even a small planar drug moiety (see Figure 2), and stimulates salt-resistant DNA cleavage. 133 The results showed that clerocidin-dependent irreversible DNA cleavage was site selective, and required a guanine at position-1. Interestingly, some irreversible sites showed an abnormal electrophoretic mobility in sequencing gels (with respect to cleaved bands generated by mAMSA), suggesting a chemical alteration of the DNA strand. Since in ethanol solution clerocidin underwent chemical modifications of the C12-C15 side chain, losing in parallel the ability to stimulate irreversible DNA breaks, we thus suggested that the drug dicarbonyl function (Figure 2) may react with the guanine at the-1 position in the ternary complex resulting in cleavage irreversibility and altered DNA mobility in sequencing gels. 133

Enzyme Subunit Cooperativity Eukaryotic topoisomerase II subunits appear to act cooperatively during the catalytic reaction while undergoing conformational changes. 23'24 We have shown that cooperativity also occurs during the DNA cleavage step, and that one drug molecule interacting with one enzyme subunit may stimulate a strand cut mediated by the second subunit, due to cooperative effects. 126 This demonstrates that drug action at one subunit may be affected by the action of a second drug molecule acting at the other protein subunit. Consequently, site selectivity of a poison may be influenced, at least partially, by other factors in addition to drug structure itself, as is the case of amonafide. TM This drug has a strikingly high site selectivity since it stimulates about 60% of DNA cleavage at only one site in pBR322 DNA, and at two sites in SV40 DNA. 134 A statistical analysis of 94 sites, including the three prominent sites of pBR322 and SV40 DNAs, showed that amonafide strongly prefers a cytosine at position-1, and to a lesser extent an adenine at position + 1.134

24


However, we observed an inverted repeat sequence at the three sites exceptionally stimulated by the drug, with cytosines and adenines at positions-1 and +1, respectively. A base mutation analysis could rule out the involvement of a stem-loop structure of DNA in the drug action or in protein-DNA interactions. The data thus documented that exceptionally high levels of cleavage stimulation occur when amonafide molecules can optimally interact with both of the two enzyme subunits, which may then enhance drug stimulative effects through cooperative action. TM

Drug Receptor Sites in the Ternary Complex The novel information accumulated during recent years suggests a more defined model of a ternary complex, DNA-drug-topoisomerase II (Figure 3). We propose that the drug is placed at the interface protein/DNA of the enzyme active site making contacts with DNA base pairs very close to the cleaved phosphodiester bond (from -2 to +2 positions). To this broadly defined receptor, at least three pharmacophores may bind: (1) one represented by poisons with -1 position specificity; (2) a second represented by poisons with +1 position specificity; and (3) the third represented by poisons with +2 position specificity (Table 2)~ The three pharmacophores may indeed interact with partially different protein and DNA units, thus one may d.istinguish at least three partially overlapping receptor sites in the binary DNAenzyme complex. In Figure 3, this has been indicated by the shape and the position,

A

'! II

1

+2 +3

+4

'

$.

B

+5

$"

t +5

+4

+3

.2

+1 -1,~~'+2

.+3

l[

3'

+2

+1

C -2

-1

-2

-1 ~

+5

+1 +2 + 3

+4

+5

+1

-1

+4

+3

+2

'

+4

+5

1

-1

+1

-2

S'

+5

+4

+3

+2

-2

Figure3. Model of DNA-drug-topoisomerase II ternary complexes with overlapping drug receptor sites. DNA strands are represented by thin lines; drug molecules are shown as black objects.


25

relative to DNA strands, of the poison [shown as a black object]. Drugs might interact with the DNA in an intercalation-like manner (A and B models) or externally in the minor groove (C model); however, a moiety of the drug molecule would always be close to the 5' and/or 3' termini of the break, interfering with the breakage-rejoining reaction. In this schematic representation, no attempt was made to hypothesize drug interactions with the protein, since the structure of the enzyme active site is not at the present time available. Nevertheless, drugs most likely also interact with topoisomerases, and therefore the precise contacts among the three components-drug, DNA, and enzyme--remain to be fully elucidated. We suggested that in the case of anthracyclines, the sugar, and in particular the 3' amino group, makes contacts with the DNA in the minor groove. The side chain and the OH group at the C-9 position of the saturated ring might then likely interfere with the catalysis of DNA religation, ll3 Morjani et al. 135 reported an investigation of the interaction of intoplicin in the ternary complex by surface-enhanced Raman scattering spectroscopy. Intoplicin is a dual poison, since it can stimulate DNA cleavage by either topoisomerase I and II. 136'137Based on the Raman technique, the authors provided evidence that the drug interacts with topoisomerase II, and possibly that at least part of the drug molecule is buried in an internal part of the protein. Upon DNA binding, the drug is released from the enzyme and forms hydrogen bonds. 135These results indicate the presence of a hydrophobic pocket in the protein that may be easily accessible, and may constitute a high-affinity site for the drug. Such an enzyme pocket resembles the model proposed by Filipski, 138who hypothesized that base binding sites were present in the catalytic site of topoisomerase II, and that a drug molecule could be a competitive poison of DNA bases binding to these enzyme pockets. It is therefore possible that stacking between a drug molecule and DNA bases may occur in a hydrophobic environment constituted by an enzyme pocket. The sequence specificity of DNA cleavage stimulation by saintopin (Figure 2) was investigated in the presence of topoisomerase II and topoisomerase 1.139 The authors found that the drug highly prefers a guanine at position + 1 in the case of both enzymes, suggesting that drug-DNA interactions may be the same in the two cases. It is intriguing that some compounds are able to form ternary complexes with either topoisomerase II and I. Moreover, the observation that drug sequence specificity remains the same with both enzymes in the case of saintopin 139 raises the possibility of at least partial structural resemblance of the drug receptor sites in the two ternary complexes. This may be also supported by the finding that partial homology has been found between certain regions of the amino acid sequences of all DNA topoisomerases I and II. 13A similar cleft among different protein domains may be a common structural basis of the DNA breakage-rejoining catalysis of all DNA topoisomerases. 13

26


B. Non-cleaving Inhibitors Several agents are known to inhibit DNA topoisomerases without stabilization of the cleavable complex. 14~ This category of agents includes dioxopiperazines, merbarone, fostriecin, chebulagic acid, 13-1apachone, and others. 140'141'143-145These drugs inhibit the catalytic activity of purified enzymes and prevent DNA cleavage stimulated by teniposide or amsacrine. 14~ The mechanisms by which these drugs affect DNA topoisomerases are not known; however, the action of ICRF-193, a bisdioxopiperazine derivative (Figure 2), against topoisomerase II has recently been elucidated. 146In the absence of ATP, ICRF-193 has little effect on the binding of the enzyme to various forms of DNA, whereas in the presence of ATP, the drug converts the enzyme to a form which is capable of binding linear, but not circular, DNA molecules. 146The results have been interpreted in terms of the protein-clamp model: 23 ICRF- 193 may bind to the closed-clamp form of the enzyme and prevents its conversion to the open-clamp form. A DNA-bound enzyme in the closed-clamp form would be prevented from transporting DNA segments, and a free enzyme trapped in the closed-clamp form would be prevented from binding to DNA. Therefore, the drug action is expected to affect chromosomal condensation and decondensation, and the segregation of intertwined chromosomes during mitosis. Studies with mammalian cell lines show that cell exposures to ICRF-193 inhibit cell cycle progression at G2-M, TM and prevents chromosome segregation during anaphase. 89 Interestingly, ICRF-193 derivatives have been shown to be non-crossresistant with classical topoisomerase II antitumor poisons in a human VM-26resistant leukemia CEM cell line. 147 Thus, it may be of great interest in the development of topoisomerase II poisons with a different mechanism of action, which may be non-cross-resistant with conventional agents used in the therapy of human cancers.

C. Development of New Topoisomerase II-Directed Drugs In the last decade, several laboratories made efforts to discover and develop novel topoisomerase-interacting agents as more selective antitumor drugs. 20'120'148'149 New agents have been discovered both through random screening programs, 15~and from analogue development, based on arbitrary chemical modifications of a lead compound. 151'152 An attempt was made to prospectively design new compounds using molecular modeling, and based on a composite pharmacophore derived from known poisons. 153 The researchers thus synthesized azatoxin, a hybrid molecule of VP-16 and ellipticine. The approach appears successful, since azatoxin (that did not bind to DNA) was shown to be a poison of topoisomerase II and stimulated extensive DNA cleavage. 154 The makaluvamines, a series of pyrroloiminoquinones, were discovered and isolated from a sponge of the genus Zyzzya by following pharmacological activity against the human colon carcinoma line HCT-116.155 These compounds showed


27

enhanced toxicity toward a cell line sensitive to topoisomerase II poisons, and they were also shown to inhibit enzyme decatenation activity of kinetoplast DNA. ~55 Among the cleavable complex stabilizing agents, a new generation of antitumor intercalators related to mAMSA were constituted" acridine-4-carboxamide, 2-(4pyridyl) quinoline-8-carboxamide, and the 9-anilinoacridine derivatives. 156'15vThe first two tricyclic carboxamides induce DNA-protein cross-links and DNA double-strand breaks in mouse fibrosarcoma cells (line 935.1). 156 9-Anilinoacridines were synthesized and evaluated for their ability to inhibit the growth of Jurkat leukemia cells and to trap human DNA topoisomerase II. Derivatives bearing 1' substituents containing SO 2 moieties proved most potent in stimulating topoisomerase II-mediated DNA cleavage. A good correlation was found between the ability of these compounds to induce DNA breaks in vitro or in whole cells, and their cytotoxicity in human lymphoma Jurkat cells. 157 Another compound related to amsacrine is the anilinoacridine derivative bearing an N-methylpyrrolecarboxamide unit at position 1'. 158 Linear dichroism spectroscopy revealed that this compound intercalates its acridine chromophore between DNA base pairs, with a preference for GC-rich sequences, whereas both the analogue lacking the N-methylpyrrole unit and amsacrine itself intercalate into DNA without any strong sequence preference. This drug stabilizes the topoisomerase II-DNA covalent complex and stimulates DNA cutting at a subset of pre-existing topoisomerase II cleavage sites (which were different from those of anasacrine). 158 The nature of the substituent at position 1' of the anilinoacridine chromophore may therefore determine the site of DNA cleavage stimulation. 158 BE-22179, a cyclic depsipeptide having two 3-hydroxyquinoline moieties, has been shown to be more effective than VM-26 in the inhibition of the DNA-relaxing activity of L 1210 topoisomerase II. 159Nevertheless, it was ineffective in stimulating topoisomerase II DNA cleavage, but able to intercalate into DNA. However, since the compound inhibited topoisomerase II at far lower concentrations than those showing DNA-unwinding activity, this indicated that the anti-topoisomerase II activity of BE-22179 may likely be independent from DNA binding. The authors suggested that BE-22179 may directly interact with the enzyme, thus preventing its binding to DNA. 159 It remains to be established whether the potent cytotoxic activity of BE-22179 is related to topoisomerase II inhibition. Methyl-substituted indolo[2,3-b]quinolines 16~are c~-carboline derivatives, structurally similar to ellipticine. 16 1 Although structurally similar, the difference in cytotoxic activity between these compounds and ellipticines is striking. 16~ The indolo-quinolines showed a cytotoxic activity about 10 times as high as that of ellipticine; moreover these compounds were more active than ellipticines in stimulating in vitro topoisomerase II DNA breaks. These results demonstrated that minor variations of the structure of indolo[2,3-b]quinolines may yield remarkable changes in pharmacological activity. DNA minor-groove binders have been extensively studied for their capacity to bind to specific base sequences. One of the best known of these agents is the

28


antiviral antibiotic distamycin, which selectively binds to AT-rich sequences. 162 Distamycin is not able to interfere with the DNA breakage-reunion reaction of topoisomerase II by forming ternary complexes (c.f. anthracyclines, etoposide, and other antitumor drugs); nevertheless the groove binder may affect cleavage levels through modulation of the DNA binding activity of the enzyme. 163Several attempts have been made to rationally design a sequence-specific drug using distamycin or other minor-groove binders. The affinity of distamycin-like ligands for specific sequences can be enhanced by choosing pairs of ligand molecules with complementary DNA recognition properties. 164 A hybrid molecule, such as a distamycin intercalator, may have several potential advantages: the enhancement of the DNA binding strength and selectivity; the potential interference with topoisomerase functions; the improvement of the cellular drug transport; and an increased selectivity against cancer cells. 165 A minor-groove binder that is able to interfere with topoisomerase II activities is berenil, an antitrypanosomal drug. 166Topoisomerase II is unable to interact with DNA sites occupied by berenil or other minor-groove binders. Berenil is a relatively weak poison of type II topoisomerase as compared with other groove ligands. The difference might be related to the lower affinity of berenil for AT-rich DNA sequences with respect to distamycin or netropsin. 162'167 Studies of the molecular interaction of dual topoisomerase I and II poisons may shed light on the mechanism of action of topoisomerase-targeting drugs. Intoplicine (Figure 2) is a synthetic anticancer agent belonging to the series of 7-Hbenzo[e]pyridol[4,3-b]indoles, and is able to poison both enzymes. 137'168 Intoplicine was active in vitro against a variety of human tumor lines, including a subgroup of them that were insensitive to other agents. 169 Direct interaction between the drug and topoisomerase II has been shown, and it can play a critical role in the drug action. 135'17~ Many alkaloids of the benzo[c]phenanthridine family, such as fagaronine and nitidine, are also dual poisons. 171'172Fagaronine is a potent inducer of differentiation in various hematopoietic cell lines, and it has been shown to interact with DNA as well as with double-stranded regions of tRNAs. 173This drug can intercalate into DNA, and lower concentrations are required to inhibit topoisomerase I than topoisomerase 11.172 A particular class of antitumor compounds is the family of protein tyrosine kinase (PTK) poisons that are also DNA topoisomerase II poisons. Drugs which bind to the ATP site of PTK, such as genistein, are common poisons of both types of enzymes. 174Erbstatin and tyrphostine derivatives, which inhibit epidermal growth factor receptor PTK activity by competing with both the peptide substrate and ATE have the capacity to inhibit DNA topoisomerases I and II. 174In contrast to genistein, none of these molecules induced the stabilization of topoisomerase II DNA cleavable complexes.


VI.

29

M E C H A N I S M S OF D R U G CELL KILLING A. Cellular and Molecular Mechanisms

The stabilization of topoisomerase DNA cleavage by antitumor poisons is a reversible event, and the mechanism by which a transient DNA damage eventually results in cell killing is not completely clear. Thus, although the primary cellular targets of these anticancer drugs have been identified, less is known about how the process leading to selective cell death is activated by the primary lesions. Cell death induced by topoisomerase poisons, and also by other DNA-damaging agents, is likely due to their interference with DNA metabolism. Recent studies support an active cell participation in the process of cell death, since activation of new patterns of gene expression follows exposures to these agents. 175-18~ Indirect evidence suggest that this program is activated when drug-induced lesions are persistent and/or optimal repair is impaired. Inhibition of cell division is a common mechanism of response to DNA damage in prokaryotic and eukaryotic cells. To explain the cytotoxicity of drug-induced cleavage complexes, it has been argued that a cellular process, such as replication or transcription, converts the reversible cleavage complexes into permanent cytotoxic DNA lesions. 16 The role of replication and transcription in the cellular response to topoisomerase poisons has previously been investigated by measuring drug cytotoxicity under conditions where DNA or RNA syntheses were inhibited. 181'182A characteristic G2 arrest is commonly observed when cells are exposed to topoisomerase II poisons, which is likely achieved by downregulation of protein kinase. 183'184 Cell cycle arrest is believed to be an evolutionary conserved cell response to allow repair of DNA lesions before mitosis; 185'186nevertheless, heavily damaged cells might never reach mitosis and die in G2. Cell death might also be caused by incorrect repair of the damage. It has been demonstrated that etoposide and amsacrine induced deletions of a selectable genetic marker, possibly through nonhomologous DNA recombination, in cultured mammalian cells. 187'188The authors suggested that DNA deletions involving essential genes may be lethal to the cell 187'188 and that enzyme subunit exchanges may be the mechanism of large DNA deletions. Moreover, Bodley et al. 189 studied the effects of teniposide treatments on SV40 minichromosomes. Teniposi0e exposures of SV40-infected cells resulted in the rapid conversion of SV40 DNA into a high molecular weight form. The authors showed that this was due to the integration of viral DNA into the cellular genome, and found that integration sites on the viral DNA occurred at major sites of teniposide cleavage stimulation in an in vitro reaction with topoisomerase II. Similarly, Howard and Griffith 19~ showed that several strong sites of topoisomerase II DNA cleavage in HIV DNA were found less than 1000 bp from the integration site in the genome of host cells. These results suggested a direct role of topoisomerase II, and possibly of the ternary cleavable complex, in the integration process, and in chromosomal rearrangement pathways.

30


Helicase-driven separation of the double helix occurs during DNA replication and also during other DNA metabolic activities such as repair and recombination. 19) In recent work, a two-hybrid system has been used to demonstrate that the yeast Sgsl protein interacts with the C-terminal domain of yeast topoisomerase II. 192 S gsl protein has sequence homology with a bacterial helicase, and the authors proposed a model for the interaction of topoisomerase II with a helicase to account for the faithful segregation of newly replicated chromosomes. 192 Howard et al. 193 provided biochemical evidence that DNA helicases can convert drug-induced type II topoisomerase cleavage complexes to irreversible DNA lesions by displacing broken ssDNA fragments from the complex. DNA helicases may also play a role in various recombination and/or mutation events induced by DNA topoisomerases. Topoisomerase poisons that stimulate the cleavable complex have been shown to stimulate homologous recombination, sister chromatid exchange, gross genetic rearrangements, and frameshift mutations. 16'17'194For each of these genetic events, the helicase-mediated generation of discrete DNA breaks from a cleavage complex would be a reasonable first step. DNA helicases have recently been found to be sensitive to certain topoisomerase poisons and each helicase appears to have its own unique spectrum of sensitivity. 195'196 If a DNA helicase converts covalent topoisomerase-DNA complexes into cytotoxic lesions in vivo, then topoisomerase poisons that inhibit the helicase should be less cytotoxic than those that do not. The observation that each drug stimulates specific DNA cleavage patterns in in vitro systems raises the question of whether the drug sequence specificity has a role in drug cell killing activity. While drug cytotoxic potency correlates well with the extent of drug-stimulated double-strand breaks when considering closely related derivatives, this relation is lost when compounds of different classes are compared. 15'17'19A high degree of heterogeneity in the localization of topoisomerase II DNA cleavage stimulated by unrelated drugs may be observed in the chromatin of Drosophila Kc cells. 77 Interestingly, the reversibility of topoisomerase II DNA cleavage was slower in a transcriptionally active gene than in a silent genomic locus. Moreover, VM-26 and a doxorubicin analogue stimulated similar levels of doublestrand breaks in the whole cellular genome; nevertheless the anthracycline was 5to 10-fold more cytotoxic than VM-26. When specific loci were examined, the two drugs differed for the extent, location, and reversibility of DNA breakage. These findings suggest that the lethality of drug-stabilized topoisomerase II-DNA complexes might be different, depending on the specific DNA sequence targeted. 77

B. Genetic Mechanisms In recent years, several laboratories have focused on the process of apoptosis, a gene-directed cell death program operationally defined by biochemical and morphological events. 197 Different types of DNA-damaging agents, including T-rays and anticancer compounds, may trigger this genetic program, particularly in hematopoietic cells. 178'198Following drug-stimulated topoisomerase II DNA cleav-


31

age, the transient DNA breaks are presumably transformed into irreversible lesions that constitute the stimulus for activation of the cell death program. 199 The p5 3 gene product seems to play a crucial role in the drug-activated cell death pathway, z~176 The wild-type p53 protein is required either for the efficient activation of apoptosis or the cell cycle arrest at the G1/S border following exposures to DNA-damaging agents. 176'2~ The p53 product is a transcription factor that, in response to DNA damage, may regulate expression of other genes involved in the G 1 checkpoint. 2~176 The G 1 arrest has been ascribed to the ability of p53 to induce expression of the wafl/cipl gene that encodes a 21 kDa poison of G1 cyclindependent kinases. 2~ The degree of G 1 arrest observed with etoposide and X-rays correlated with the rate of p53 and p21 wafl/cipl protein accumulation, e~ However, whereas cells with wild-type p53 exhibited G 1 and G2 arrests after exposures to ionizing radiation, cells with mutated or deleted p53 retained only G2 arrest. 2~ Moreover, DNA-damaging agents tended to decrease survival to a greater extent in wild-type p53 cell lines as compared with mutant p53 cell lines. 2~ Wild-type p53 protein levels are raised dramatically in response to physical or chemical agents, and this is due to increased protein stability, e~ Studies of p53-deficient transgenic mice (p53 "knockouts") have also demonstrated that p53 is required for the induction of apoptosis by ~,radiation (and by some anticancer drugs, including doxorubicin) in thymocytes that are mainly in the G0/G 1 phase. 176'2~ The role of p5 3 in chemotherapeutic responses has also been evaluated in animal models. Using a transplantable fibrosarcoma, the authors showed that p5 3-deficient tumors treated with 7 radiation or adriamycin continued to enlarge and contained few apoptotic cells, in comparison to tumors with p53+/+ that regressed after drug treatments, e~ GADD genes (growth arrest and DNA damage-inducible genes) may play a role in growth arrest following DNA damage. 2~176 The GADD 45 and wafl/cipl gene expressions are specifically dependent on the wild-type (but not mutated) p53, which may bind to a conserved sequence motif found in an intron of the GADD 45 gene. 2~ Bcl-2 and Bax are homologous proteins that have opposite effects on cell life and death, with Bcl-2 serving to prolong cell survival and Bax acting as an accelerator of apoptosis. 211 Deregulated expression of Bcl-2 may thus protect lymphoma cells from chemotherapeutic agents. 2~2Consistently, Bcl-2 gene rearrangement seems to be correlated with a shorter disease-free survival in lymphoma patients. 212 Kamesaki et al. e~3 showed that overexpression of Bcl-2 protein inhibited etoposideinduced apoptosis and cytotoxicity without affecting the number of single- and double-strand breaks in cultured cells. The Bcl-2 product is likely a member of protein complexes with Bcl-2-related proteins, such as Bax. el4 Miyashita et al. el5 obtained evidence that restoration of p53 in a murine leukemia cell, M1, was associated with increases in Bax mRNA and protein. Furthermore, the increases in Bax gene expression were accompanied by simultaneous decreases in the steadystate levels of Bcl-2 mRNA and protein, el6 The effects of p53 on Bcl-2 gene expression may be mediated at least in part by a cis-acting p53 negative-response

32


element located in the 5' untranslated region (5'UTR) of the bcl-2 gene. 217 p53 can downregulate Bcl-2 gene expression in cultured cells and in some tissues in vivo, 215 and Bcl-2 can antagonize apoptosis induced by wild-type p53. 218 Functional analysis of the Bax gene promoter indicates that this apoptosis-inducing gene is a direct target of p53. 219 The results suggest the existence of multiple independent intracellular mechanisms of apoptosis, some of which can be prevented by Bcl-2, while others are unaffected by this gene product. Alternatively, some pathways may involve proteins that differentially regulate Bcl-2 function. Interestingly, the protein tyrosine kinase c-abl has been recently invoked in the pathway of cell death. 22~C-abl codifies for a membrane protein that, when activated (e.g. following a gene translocation), can be associated with leukemogenesis. Sawyers et al. 221 found that overexpression of c-abl leads to growth arrest. Overexpression of a dominant negative mutant leads to growth deregulation. 221 These findings demonstrate that the biological activity of c-abl contrasts with that of the abl-fused oncogene. 221 The c-abl gene may thus resemble a tumor suppressor gene. 222 Cooperativity between c-myc and bcl-2 genes was also studied, and the data indicated that their combined action may be important for transformation, as well as for the development of drug resistance. 223 The expression of Bcl-2 specifically abrogates c-myc-induced apoptosis, without affecting the c-myc mitogenic effect. 223 Moreover, Ryungsa et al. 224 have found that treatment of human leukemic lymphoblasts with teniposide (VM-26), under conditions that stabilize DNA-topoisomerase II complexes, caused the formation of internucleosomal DNA ladders. Under continuous exposure to VM-26, the internucleosomal DNA ladders were associated with the transient induction of c-jun mRNA in a dose-dependent fashion. The induction of c-jun mRNA by VM-26 apparently preceded DNA ladder formation. 224 In recent years, it has been reported that certain topoisomerase II poisons induce apoptotic cell death in human myeloid leukemic cells, characterized by internucleosomal DNA ladders and associated with the induction of transcription factors of c-jun and c-fos genes. 225 In the next few years, more information will certainly be accumulated on the genetic prog~'am of cell death, and how chemical and physical agents may trigger apoptosis in human tumor cells. The knowledge is expected to be fundamental for the complete understanding of the mechanism of the cell killing activity and selectivity of topoisomerase II poisons. Moreover, clinical research will need to establish the role of these factors in the unresponsiveness of certain human cancers to combination chemotherapy.

ACKNOWLEDGMENTS We gratefully acknowledge grant supports from Associazione Italiana per la Ricerca sul Cancro, Milan; Consiglio Nazionale delle Ricerche, Progetto Finalizzato ACRO, Rome; and Ministero della Sanita', Rome, Italy.


33

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. I I. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

Wang, J. C. Ann. Rev. Biochem. 1985, 54, 665. Wang, J. C. Ann. Rev. Biochem. 1996, 65, 635. Osheroff, N. Pharmacol. Ther. 1989, 41,223. Cozzarelli, N. R.; Wang, J. C. DNA Topology and its Biological Effects; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, 1990. Wang, J. C. J. Biol. Chem. 1991, 266, 6659. Watt, E M.; Hickson, I. D. Biochem. J. 1994, 303, 681. Andoh, T.; Ikeda, H.; Oguro, M. Molecular Biology of DNA Topoisomerases and its Application to Chemotherapy, CRC Press: Boca Raton, 1993. Osheroff, N.; Zechiedrich, E. L.; Gale, K. C. Bioessays 1991, 13, 269. Wang, J. C.; Becherer, K. Nuclueic Acids Res. 1983, 11, 1773. Bergerat, A.; Gadelle, D.; Forterre, E J. Biol. Chem. 1994, 269, 27663. Slesarev, A. I.; Lake, J. A.; Stetter, K. O.; Gellert, M.; Kozyavkin, S. A. J. Biol. Chem. 1994, 269, 3295. Berger, J. M.; Gamblin, S. J.; Harrison, S. C.; Wang, J. C. Nature 1996, 379, 225. Caron E R.; Wang, J. C.; Andoh, T.; Ikeda, H.; Oguro, M., Eds. Molecular Biology of DNA Topoisomerases and its Application to Chemotherapy; CRC Press: Boca Raton, 1993. Lock, R. B.; Ross, W. E. Anti-Cancer Drug Des. 1987, 2, 151. Pommier, Y.; Kohn, K. W. In Developments in Cancer Chemotherapy; Glazer, R. I. Ed.; CRC: Boca Raton, 1989. D'Arpa, E; Liu, L. E Biochim. Biophys. Acta 1989, 989, 163. Liu, L. E Ann. Rev. Biochem. 1989, 58, 35 I. Zhang, H.; D'Arpa, E; Liu, L. E Cancer Cells 1990, 2, 23, 1990. Zunino, E; Capranico, G. Anti-Cancer Drug Des. 1990, 5, 307. Pommier, Y. Cancer Cherru)ther. PharmacoI. 1993, 32, 103. Binaschi, M.; Zunino, E; Capranico, G. Stem Cells 1995, 13, 369. Corbett, A. H.; Osheroff, N. Chem. Res. ToxicoL 1993, 6, 585. Roca, J.; Wang, J. C. Cell 1992, 71,833. Roca, J.; Wang, J. C. Cell 1994, 77, 609. Orphanides, G.; Maxwell, A. Curr. Biol. 1994, 4, 1006. Drake, E H.; Hofmann, G. A.; Bartus, H. E; Mattern, M. R.; Crooke, S. T.; Mirabelli, C. K. Biochemistry 1989, 28, 8154. Jenkins, J. R.; Ayton, E; Jones, T.; Davies, S. L.; Simmons, D. L.; Harris, A. L.; Sheer, D.; Hickson, I. D. Nucleic Acids Res. 1992, 20, 5587. Tsai-Pflugfelder, M.; Liu, L. E; Liu, A. A.; Tewey, K. M.; Whang-Peng, J.; Knutsen, T.; Huebner, K.; Croce, C. M.; Wang, J. C. Proc. Natl. Acad. Sci. USA 1988, 85, 7177. Chung, T. D.; Drake, E H.; Tan, K. B.; Per, S. R.; Crooke, S. T.; Mirabelli, C. K. Proc. Natl. Acad. Sci. USA 1989, 86, 9431. Austin, C. A.; Sng, J.-H.; Patel, S.; Fisher, L. M. Biochim. Biophys. Acta Gene Struct. Expression 1993, 1172, 283. Capranico, G.; Tinelli, S.; Austin, C. A.; Fisher, M. L.; Zunino, E Biochim. Biophys. Acta 1992, 1132, 43. Cornarotti, M.; Tinelli, S.; Willmore, E.; Zunino E; Fisher, L. M.; Austin, C. A; Capranico, G. MoL Pharmacol. 1996, 50, 1463. Caron, E R.; Wang, J. C. In Advances in Pharmacology; Academic Press: New York, 1994. Wyckoff, E.; Hsieh, T. S. Proc. Natl. Acad. Sci. USA 1988, 85, 6272. Adachi, N.; Miyaike, M.; Ikeda, H.; Kikuchi, A. Nucleic Acids Res. 1992, 20, 5297. Uemura, T.; Ohkura, H.; Adachi, Y.; Morino, K.; Shiozaki, K.; Yanagida, M. Cell 1987, 50, 917. Wasserman, R. A.; Austin, C. A.; Fisher, U M.; Wang, J. C. Cancer Res. 1993, 53, 3591.

34


38. Eder, J. P., Jr.; Chan, V.T.-W.; Niemierko, E.; Teicher, B. A.; Schnipper, L. E. J. Biol. Chem. 1993, 268, 13844. 39. Lindsley, J. E.; Wang, J. C. Proc. Natl. Acad. Sci. USA 1991, 88, 10485. 40. Shiozaki, K.; Yanagida, M. Mol. Cell. Biol. 1991, 11, 6093. 41. Wigley, D. B.; Davies, G. J.; Dodson, E. J.; Maxwell, A.; Dodson, G. Nature 1991, 351,624. 42. Lee, M. P.; Hsieh, T. S. J. Mol. Biol. 1994, 235, 436. 43. Chan, V. T. W.; Ng, S.; Eder, J. P.; Schnipper, L. E. J. Biol. Chem. 1993, 268, 2160. 44. Danks, M. K.; Warmoth, M. R.; Friche, E.; Granzen, B.; Bugg, B. Y.; Harker, W. G.; Zwelling, L. A.; Futscher, B. W.; Suttle, D. P.; Beck, W. T. Cancer Res. 1993, 53, 1373. 45. Lee, M. S.; Wang, J. C.; Beran, M. J. Mol. Biol. 1992, 223, 837. 46. Crenshaw, D. G.; Hsieh, T. J. Biol. Chem. 1993, 268, 21328. 47. Crenshaw, D. G.; Hsieh, T. J. Biol. Chem. 1993, 268, 21335. 48. Caron, P. R.; Watt, P.; Wang, J.C. Mol. Cell. Biol. 1994, 14, 3197. 49. Shiozaki, K.; Yanagida, M. J. Cell Biol. 1992, 119, 1023. 50. Ackerman, P.; Glover, C. V.; Osheroff, N. Proc. Natl. Acad. Sci. USA 1985, 82, 3164. 51. Cardenas, M. E.; Dang, Q.; Glover, C. V.; Gasser, S. M. EMBO J. 1992, 11, 1785. 52. Bojanowski, K.; Filhol, O.; Cochet, C.; Chambaz, E. M.; Larsen, A. K. J. Biol. Chem. 1993, 268, 22920. 53. Corbett, A. H.; Femald, A. W.; Osheroff, N. Biochemistry 1993, 32, 2090. 54. Corbett, A. H.; DeVore, R. F.; Osheroff, N. J. Biol. Chem. 1992, 267, 20513. 55. DeVore, R. F.; Corbett, A. H.; Osheroff, N. Cancer Res. 1992, 52, 2156. 56. Takano, H.; Kohno, K.; Ono, M.; Uchida, Y.; Kuwano, M. Cancer Res. 1991, 51,3951. 57. Cardenas, M. E.; Gasser, S. M. J. Cell Sci. 1993, 104, 219. 58. Cardenas, M. E.; Walter, R.; Hanna, D.; Gasser, S. M. J. Cell Sci. 1993, 104, 533. 59. Dang, Q.; Alghisi, G.-C.; Gasser, S. M. J. Mol. Biol. 1994, 243,10. 60. Vassetzky, Y. S.; Dang, Q.; Benedetti, P.; Gasser, S. M. Mol. CelL Biol. 1994, 14, 6962. 61. Wells, N. J.; Addison, C. M.; Fry, A. M.; Ganapathi, R.; Hickson, I. D. J. Biol. Chem. 1994, 269, 29746. 62. Heck, M. M.; Hittelman, W. N.; Eamshaw, W.C.J. Biol. Chem. 1989, 264, 15161. 63. Saijo, M.; Ui, M.; Enomoto, T. Biochemistry 1992, 31,359. 64. Burden, D. A.; Goldsmith, L. J.; Sullivan, D. M. Biochem. J. 1993, 293, 297. 65. Ganapathi, R.; Zwelling, L.; Constantinou, A.; Ford, J.; Grabowski, D. Biochem. Biophys. Res. Commun. 1993, 192, 1274. 66. Ritke, M. K.; Allan, W. P.; Fattman, C.; Gunduz, N. N.; Yalowich, J. C. Mol. Pharmacol. 1994, 46, 58. 67. Taagepera, S.; Rao, P. N.; Drake, E H.; Gorbsky, G. J. Proc. Natl. Acad. Sci. USA 1993, 90, 8407. 68. Kimura, K.; Nozaki, N.; Saijo, M.; Kikuchi, A.; Ui, M.; Enomoto, T. J. Biol. Chem. 1994, 269, 24523. 69. Burden, D. A.; Sullivan, D. M. Biochemistry 1994, 33, 14651. 70. Laemmli, U. K.; K~, E.; Poljak, L.; Adachi, Y. Curr. Opin. Gen. Dev. 1992, 2, 275. 71. Roberge, M.; Gasser, S. M. Mol. Microbiol. 1992, 6, 419. 72. Gasser, S. M.; Laroche, T.; Falquet, J.; Boy de la Tour, E.; Laemmli, U. K. J. Mol. Biol. 1986, 188, 613. 73. Earnshaw, W. C.; Heck, M. M. J. Cell Biol. 1985, 100, 1716. 74. Adachi, Y.; K/is, E.; Laemmli, U. K. EMBO J. 1985, 8, 3997. 75. Pommier, Y.; Cockerill, P. N.; Kohn, K. W.; Garrard, W. T. J. Virol. 1990, 64, 419. 76. K~, E.; Laemmli, U. K. EMBO J. 1992, 11,705. 77. Borgnetto, M. E.; Zunino, E; Tinelli, S.; K/is, E.; Capranico, G. Cancer Res. 1996, 56, 1855. 78. DiNardo, S.; Voelkel, K.; Stemglanz, R. Proc. Natl. Acad. Sci. USA 1984, 81, 2616. 79. Holm, C.; Goto, T.; Wang, J. C.; Botstein, D. Cell 1985, 41,553. 80. Yang, L.; Wold, M. S.; Li, J. J.; Kelly, T. J.; Liu, L. E Proc. Natl. Acad. Sci. USA 1987, 84, 950.

Poisons of Type II DNA Topoisomerases 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121.

35

Holm, C.; Steams, T.; Botstein, D. Mol. Cell. Biol. 1989, 9, 159. Spell, R. M.; Holm, C. Mol. Cell. Biol. 1994, 14, 1465. Buchenau, P.; Saumweber, H.; Arndt-Jovin, D. J. J. Cell Sci. 1993, 104, 1175. Richter, A.; Strausfeld, U. Nucleic Acids Res. 1988, 16, 10119. Ishimi, Y.; Sugasawa, K.; Hanaoka, F.; Eki, T.; Hurwitz, J. J. Biol. Chem. 1992, 267, 462. Snapka, R. M. Mol. Cell. Biol. 1986, 6, 4221. Wood, E. R..; Earnshaw, W. C. J. Cell Biol. 1990, 111, 2839. Adachi, Y.; Luke, M.; Laemmli, U. K. Cell 1991, 64, 137. Clarke, D. J.; Johnson, R. T.; Downes, C. S. J. Cell Sci. 1993, 105, 563. Downes, C. S.; Clarke, D. J.; Mullinger, A. M.; Gimenez-Abian, J. E; Creighton, A. M.; Johnson, R. T. Nature 1994, 372, 467. Rose, D.; Holm, C. Mol. CelL Biol. 1993, 13, 3445. Nakamura, S.; Nakamura, M.; Kojima, T.; Yoshida, H. Agents Chemother. 1989, 33, 254. Cullen, M. E.; Wyke, A. W.; Kuroda, R.; Fisher, L. M. Agents Chemother. 1989, 33, 886. Oram, M.; Fisher, L. M. Antimicrob. Agents Chemother. 1991, 35, 387. Vizan, J. L.; Hemandez-Chico, C.; del Castillo, I.; Moreno, F. EMBO J. 1991, 10, 467. Nitiss, J.; Wang, J. C. Proc. Natl. Acad. Sci. USA 1988, 85, 7501. Jannatipour, M.; Liu, Y.; Nitiss, J. L. J. Biol. Chem. 1993, 268, 18586. Wasserman, R. A.; Wang, J. C. Cancer Res. 1994, 54, 1795. Hinds, M.; Deisseroth, K.; Mayes, J.; Altschuler, E.; Jansen, R.; Ledley, ED.; Zwelling, L.A. Cancer Res. 1991, 51, 4729. Bugg, B. Y.; Danks, M. K.; Beck, W. T.; Suttle, D. P. Proc. Natl. Acad. Sc'i. USA 1991, 88, 7654. Campain, J. A.; Gottesman, M. M.; Pastan, I. Biochemistry 1994, 33, 11327. Patel, S.; Fisher, L. M. J. Cancer 1993, 67, 456. Capranico, G.; Palumbo, M.; Tinelli, S.; Zunino, F. J. Biol. Chem. 1994, 269, 25004. Fujii, N.; Yamashita, Y.; Saitoh, Y.; Nakano, H. J. Biol. Chem. 1993, 268, 13160. Chen, A. Y.; Yu, C.; Gatto, B.; Liu, L. E Proc. Natl. Acad. Sci. USA 1993, 90, 8131. Nelson, E. M.; Tewey, K. M.; Liu, L. E Proc. Natl. Acad. Sci. USA 1984, 81, 1361. Rowe, T. C.; Chen, G. L.; Hsiang, Y. H.; Liu, L. E Cancer Res. 1986, 46, 2021. Zwelling, L. A.; Mitchell, M. J.; Satitpunwaycha, P.; Mayes, J.; Altschuler, E.; Hinds, M.; Baguley, B. C. Cancer Res. 1992, 52, 209. De Isabella, P.; Capranico, G.; Palumbo, M.; Sissi, C.; Krapcho, A. P.; Zunino, F. Mol. Pharmacol. 1993, 43, 715. Capranico, G.; Zunino, F.; Kohn, K. W.; Pommier, Y. Biochemistry 1990, 29, 562. Bodley, A.; Liu, L. E; Israel, M.; Seshadri, R.; Koseki, Y.; Giuliani, E C.; Kirschenbaum, S.; Silber, R.; Potmesil, M. Cancer Res. 1989, 49, 5969. Capranico, G.; Supino, R.; Binaschi, M.; Capolongo, L.; Grandi, M.; Suarato, A.; Zunino, F. Mol. Pharmacol. 1994, 45, 908. Capranico, G.; Butelli, E.; Zunino, E Cancer Res. 1995, 55, 312. Robinson, M. J.; Osheroff, N. Biochemistry 1990, 29, 2511. Robinson, M. J.; Osheroff, N. Biochemistry 1991, 30, 1807. Robinson, M. J.; Martin, B. A.; Gootz, T. D.; McGuirk, P. R.; Moynihan, M.; Sutcliffe, J. A.; Osheroff, N. J. Biol. Chem. 1991, 266, 14585. Sorensen, B. S.; Sinding, J.; Andersen, A. H.; Alsner, J.; Jensen, P. B.; Westergaard, O. J. Mol. Biol. 1992, 228, 778. Tewey, K. M.; Rowe, T. C.; Yang, L.; HaUigan, B. D.; Liu, L. E Science 1984, 226, 466. Chen, G. L.; Yang, L.; Rowe, T. C.; Halligan, B. D.; Tewey, K. M.; Liu, L. F. J. Biol. Chem. 1984, 259, 13560. Capranico, G.; Zunino, E Eur. J. Cancer 1992, 28A, 2055. Capranico, G.; Kohn, K. W.; Pommier, Y. Nucleic Acids Res. 1990, 18, 6611.

36


122. Pommier, Y.; Capranico, G.; Orr, A.; Kohn, K. W. (published erratum in Nucleic Acids Res. 1991, 19, 7003) Nucleic Acids Res. 1991,19, 5973. 123. Fosse, P.; Rene, B.; Le Bret, M.; Paoletti, C.; Saucier, J. M. Nucleic Acids Res. 1991, 19, 2861. 124. Freudenreich, C. H.; Kreuzer, K. N. EMBO J. 1993, 12, 2085. 125. Capranico, G.; Palumbo, M.; Tinelli, S.; Mabilia, M.; Pozzan, A.; Zunino, F. J. Mol. Biol. 1994, 235, 1218. 126. Bigioni, M.; Zunino, F.; Capranico, G. Nucleic Acids Res. 1994, 22, 2274. 127. Kjeldsen, E.; Mollerup, S.; Thomsen, B.; Bonven, B. J.; Bolund, L.; Westergaard, O. J. MoL Biol. 1988, 202, 333. 128. Jaxel, C.; Capranico, G.; Kerrigan, D.; Kohn, K. W.; Pommier, Y. J. Biol. Chem. 1991, 266, 20418. 129. Freudenreich, C. H.; Kreuzer, K. N. Proc. Natl. Acad. Sci. USA 1994, 91, 11007. 130. Capranico, G.; De Isabella, P.; Tinelli, S.; Bigioni, M.; Zunino, E Biochemistry 1993, 32, 3038. 131. Corbett, A. H.; Hong, D.; Osheroff, N. J. Biol. Chem. 1993, 268, 14394. 132. Pullman, B. Anti-Cancer Drug Design 1991, 6, 95. 133. Binaschi, M.; Zagotto, G.; Palumbo, M.; Zunino, F.; Farinosi, R.; Capranico, G. Cancer Res. 1997, 57, 1710. 134. De Isabella, P.; Zunino, F.; Capranico, G. Nucleic Acids Res. 1995, 23, 223. 135. Morjani, H.; Riou, J.-E; Nabiev, I.; Lavelle, E; Manfait, M. Cancer Res. 1993, 53, 4784. 136. Riou, J.-F.; Foss6, P.; Nguyen, C. H.; Larsen, A. K.; Bissery, M.-C.; Grondard, L.; Saucier, J.-M., Bisagni, E.; Lavelle, E Cancer Res. 1993, 53, 5987. 137. Poddevin, B.; Riou, J.-E; Lavelle, E; Pommier, Y. Mol. Pharmacol. 1993, 44, 767. 138. Filipski, J. FEBS Lett. 1983, 159, 6. 139. Leteurtre, E; Fujimori, A.; Tanizawa, A.; Chhabra, A.; Mazumder, A.; Kohlhagen, G.; Nakano, H.; Pommier, Y. J. Biol. Chem. 1994, 269, 28702. 140. Drake, E H.; Hofmann, G. A.; Mong, S. M.; Bartus, J. O.; Hertzberg, R. P.; Johnson, R. K.; Mattern, M. R.; Mirabelli, C. K. Cancer Res. 1989, 49, 2578. 141. Ishida, R.; Miki, T.; Narita, T.; Yui, R.; Sato, M.; Utsumi, K. R.; Tanabe, K.; Andoh, T. Cancer Res. 1991, 51, 4909. 142. Jensen, P. B.; Sorensen, B. S.; Demant, E. J.; Sehested, M.; Jensen, P. S.; Vindelov, L.; Hansen, H. H. Cancer Res. 1990, 50, 3311. 143. Bojanowski, K.; Lelievre, S.; Markovits, J.; Couprie, J.; Jacquemin-Sablon, A.; Larsen, A.K. Proc. Natl. Acad. Sci. USA 1992, 89, 3025. 144. Boothman, D. A.; Trask, D. K.; Pardee, A. B. Cancer Res. 1989, 49, 605. 145. Boritzki, T. J.; Wolfard, T. S.; Besserer, J. A.; Jackson, R. C.; Fry, D. W. Biochem. Pharmacol. 1988, 37, 4063. 146. Roca, J.; Ishida, R.; Berger, J. M.; Andoh, T.; Wang, J. C. Proc. Natl. Acad. Sci. USA 1994, 91, 1781. 147. Chen, M.; Beck, W. T. Cancer Res. 1993, 53, 5946. 148. D'Incalci, M.; Capranico, G.; Giaccone, G.; Zunino, E; Garattini, S. In Cancer Chemotherapy and Biological Response Modifiers Annual 14; Pinedo, H. M.; Longo, D. L.; Chabner, B. A. Eds.; Elsevier Science: Amsterdam, 1993, Chapter 6. 149. Capranico, G.; Giaccone, G.; Zunino, E; Garattini, S.; D'Incalci, M. In Cancer Chenu~therapy and Biological Response Modifiers Annual 15; Pinedo, H. M.; Longo, D. L.; Chabner, B. A. Eds.; Elsevier Science B.V: Amsterdam, 1994, Chapter 6. 150. Hecht, S. M.; Berry, D. E.; MacKenzie, L. J.; Busby, R. W.; Nasuti, C. A. J. Nat. Prod. 1992, 55, 401. 151. Larue, L.; Rivalle, C.; Muzard, G.; Paoletti, C.; Bisagni, E.; Paoletti, J. J. Med. Chem. 1988, 31, 1951. 152. Zhou, X. M.; Wang, Z. Q.; Chang, J. Y.; Chen, H. X.; Cheng, Y. C.; Lee, K. H. J. Med. Chem. 1991, 34, 3346. 153. Macdonald, T. L.; Lehnert, E. K.; Loper, J. T.; Chow, K.-C.; Ross, W. E. In DNA Topoisomerases in Cancer; Potmesil, M.; Kohn, K. W. Eds.; Oxford University: New York, 1991, Chapter 16.


37

154. Leteurtre, E; Madalengoitia, J.; Orr, A.; Guzi, T. J.; Lehnert, E.; Macdonald, T.; Pommier, Y.; Cuzi, T. J. G. Cancer Res. 1992, 52, 6136, ibid., 4478. 155. Barrows, L. R.; Radisky, D. C.; Copp, B. R.; Swaffar, D. S.; Kramer, R. A.; Waters, R. L.; Ireland, C. M. Makaluvamines. Anti-Cancer Drug Design 1993, 8, 333. 156. Woynarowski, J. M.; McCarthy, K.; Reynolds, B.; Beerman, T. A.; Denny, W. A. Anti-Cancer Drug Design 1994, 9, 9. 157. Figgitt, D. P.; Denny, W. A.; Gamage, S. A.; Ralph, R. K. Anti-Cancer Drug Design 1994, 9, 199. 158. Fosse, P.; Rene, B.; Saucier, J. M.; Henichart, J. P.; Waring, M. J.; Colson, P.; Houssier, C.; Bailly, C. Biochemistry 1994, 33, 9865. 159. Yoshinari, T.; Okada, H.; Yamada, A.; Uemura, D.; Oka, H.; Suda, H.; Okura, A. Japanese J. Cancer Res. 1994, 85, 550. 160. Peczynska-Czoch, W.; Pognan, F.; Kaczmarek, L.; Boratynski, J. J. Med. Chem. 1994, 37, 3503. 161. Auclair, C. Arch. Biochem. Biophys. 1987, 259, 1. 162. Zimmer, C.; Wahnert, U. Prog. Biophys. Mol. Biol. 1986, 47, 31. 163. Fesen, M.; Pommier, Y. J. Biol. Chem. 1989, 264, 11354. 164. Wade, W. S.; Mrksich, M.; Dervan, P. B. Biochemistry 1993, 32, 11385. 165. Baitly, C.; Michaux, C.; Colson, P.; Houssier, C.; Sun, J.-S.; Garestier, T.; H61~ne, C.; H6nichart, J.-P.; Rivalle, C.; Bisagni, E. et aI. Biochemistry 1994, 33, 15348. 166. Portugal, J. FEBS Lett. 1994, 344, 136. 167. Woynarowski, J. M.; Sigmund, R. D.; Beerman, T. A. Biochemistry 1989, 28, 3850. 168. Nguyen, C. H.; Lhoste, J.-M.; Lavelle, F.; Bissery, M.-C.; Bisagni, E. J. Med. Chem. 1990, 33, 1519. 169. Eckardt, J. R.; Burris, H. A., 3rd; Kuhn, J. G.; Bissery, M. C.; Klink-Alakl, M.; Clark, G. M.; Von Hoff, D. D. J. Natl. Cancer Inst. 1994, 86, 30. 170. Nabiev, I.; Chourpa, I.; Riou, J. E; Nguyen, C. H.; Lavelle, E; Manfait, M. Biochemistry 1994, 33, 9013. 171. Larsen, A. K.; Grondard, L.; Couprie, J.; Desoize, B.; Comoe, L.; Jardillier, J.-C.; Riou, J.-E Biochem. Pharmacol. 1993, 46, 1403. 172. Janin, Y. L.; Croisy, A.; Riou, J.-F.; Bisagni, E. J. Med. Chem. 1993, 36, 3686. 173. Pezzuto, J. M.; Antosiak, S. K.; Messmer, W. M.; Slaytor, B.; Honig, G. R. Chem. Biol. Interact. 1983, 43, 323. 174. Markovits, J.; Larsen, A. K.; Segal-Bendirdjian, E.; Fosse, P.; Saucier, J. M.; Gazit, A.; Levitzki, A.; Umezawa, K.; Jacquemin-Sablon, A. Biochem. Pharmacol. 1994, 48, 549. 175. Hickman, J. A. Cancer Metastasis Rev. 1992, 11, 121. 176. Clarke, A. R.; Purdie, C. A.; Harrrison, D. J.; Morris, R. G.; Bird, C. C.; Hooper, M. U; Wyllie, A. H. Nature 1993, 362, 849. 177. Sen, S.; D'Incalci, M. FEBS Lett. 1992, 307, 122. 178. Williams, T. G.; Smith, C. A. Cell 1993, 74, 777. 179. Sinkovics, J.; Horvath, J. Leukemia 1994, 8, Suppl 1, $98. 180. Fernandes, R. S.; Cotter, T. G. Biochem. PharmacoL 1994, 48, 675. 181. Holm, C.; Covey, J. M.; Kerrigan, D.; Pommier, Y. Cancer Res. 1989, 49, 6365. 182. D'Arpa, P.; Beardmore, C.; Liu, L. E Cancer Res. 1990, 50, 6919. 183. Lock, R. B.; Ross, W. E. Cancer Res. 1990, 50, 3767. 184. Lock, R. B.; Ross, W. E. Cancer Res. 1993, 53, 3761. 185. Weinert, T. A.; Hartwell, L. H. Science 1988, 241,317. 186. D'Ari, R.; Huisman, O. J. Bacteriol. 1983, 156, 243. 187. Berger, N. A.; Chatterjee, S.; Schmotzer, J. A.; Helms, S. R. Proc. Natl. Acad. Sci. USA 1991, 88, 8740. 188. Shibuya, M. U; Ueno, A. M.; Vannais, D. B.; Craven, P. A.; Waldren, C. A. Cancer Res. 1994, 54, 1092. 189. Bodley, A. L.; Huang, H.-C.; Yu, C.; Liu, L. F. MoL Cell Biol. 1993, 13, 6190.

38 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228.

GIOVANNI CAPRANICO ET AL. Howard, M. T.; Griffith, J. D. J. Mol. Biol. 1993, 232, 1060. Matson, S. W.; Kaiser-Rogers, K. A. Ann. Rev. Biochem. 1990, 59, 289. Watt, P. M.; Louis, E. J.; Borts, R. H.; Hickson, I. D. Cell 1995, 81,253. Howard, M. T.; Neece, S. H.; Matson, S. W.; Kreuzer, K. N. Proc. Natl. Acad. Sci. USA 1994, 91, 12031. Ripley, L. S.; Dubins, J. S.; deBoer, J. G.; DeMarini, D. M.; Bogerd, A. M.; Kreuzer, K. N. J. Mol. Biol. 1988, 200, 665. George, J. W.; Ghate, S.; Matson, S. W.; Besterman, J. M. J. Biol. Chem. 1992, 267, 10683. Naegeli, H.; Modrich, P.; Friedberg, E. C. J. Biol. Chem. 1993, 268, 10386. Wyllie, A. H.; Kerr, J. E R.; Currie, A. R. Cell 198t), 68, 251. Walker, P. R.; Smith, C.; Youdale, T.; Leblanc, J.; Whitfield, J. F.; Sikorska, M. Cancer Res. 1991, 51, 1078. Ryan, A. J.; Squires, S.; Strutt, H. L.; Evans, A.; Johnson, R. T. Carcinogenesis 1994, 15, 823. Hartwell, L. Cell 1992, 71,543. Lowe, S. W.; Ruley, H. E.; Jacks, T.; Housman, D. E. Cell 1993, 74, 957. Kastan, M. B.; Zhan, Q.; El-deiry, W. S.; Carrier, E; Jacks, T.; Walsh, W. V.; Plunkett, B. S.; Volgestain, B.; Fomace, A. J. Cell 1992, 71,587. Unger, T.; Nau, M. M.; Segal, S.; Minna, J. D. EMBO J. 1992, 11, 1383. El-deiry, W. S.; Tokino, T.; Velculescu, V. E.; Levy, D. B.; Parsons, R.; Trent, J. M.; Lin, D.; Mercer, W. E.; Kinzler, K. W.; Vogelstein, B. Cell 1993, 75, 817. Fan, S.; El-deiry, W. S.; Bae, I.; Freeman, J.; Jondle, D.; Bhatia, K.; Fomace, A. J.; Magrath, J. I.; Kohn, K. W.; O'Connor, P. M. Cancer Res. 1994, 54, 5824. Kuerbitz, S. J.; Plunkett, B. S.; Walsh, W. V.; Kastan, M. B. Proc. Natl. Acad. Sci. USA 1992, 89, 7491. Fritsche, M.; Haessler, C.; Brandner, G. Oncogene 1993, 8, 307. Lowe, S. W.; Schmitt, E. M.; Smith, S. W.; Osborne, B. A.; Jacks, T. Nature 1993, 362, 847. Lowe, S. W.; Bodis, S.; McClatchey, A.; Remington, L.; Ruley, H. E.; Fisher, D. E.; Housman, D. E.; Jacks, T. Science 1994, 266, 807. Luethy, J. D.; Holbrook, N. J. Cancer Res. 1992, 52, 5. Reed, J. C. J. Cell. Biol. 1994, 124, I. Yunis, J. J.; Mayer, M. G.; Arnesen, M. A.; Aeppli, D. E; Oken, M. M.; Frizzera, G. N. Engl. J. Med. 1989, 320, 1047. Kamesaki, S.; Kamesaki, H.; Jorgensen, T. J.; Tanizawa, A.; Pommier, Y.; Cossman, J. Cancer Res. 1993, 53, 425 I. Oltvai, Z. N.; Milliman, C. L.; Korsmeyer, S. J. Cell 1993, 74, 609. Miyashita, T.; Krajewsld, S.; Krajewska, M.; Wang, H. G.; Lin, H. K.; Liebermann, D. A.; Hoffman, B.; Reed, J. C. Oncogene 1994, 9, 1799. Selvakumaran, M.; Lin, H. K.; Miyashita, T.; Wang, H. G.; Krajewsld, S.; Reed, J. C.; Hoffman, B.; Liebermann, D. Oncogene 1994, 9, 1791. Miyashita, T.; Harigai, M.; Hanada, M.; Reed, J. C. Cancer Res. 1994, 54, 3131. Wang, Y.; Szekely, L.; Okan, I.; Klein, G.; Wiman, K. G. Oncogene 1993, 8, 3427. Miyashita, T.; Reed, J. C. Cell 1995, 80, 293. Chapman, R. S.;Whetton, A. D.; Dive, C. Cancer Res. 1994, 54, 5131. Sawyers, C. L.; McLaughlin, J.; Goga, A.; Havlik, M.; Witte, O. Cell 1994, 77, 121. Levine, A. J. Ann. Rev. Biochem. 1993, 62, 623. Lotem, J.; Sachs, L. Cell Growth Differ. 1993, 4, 41. Ryungsa, K.; Beck, W. T. Cancer Res. 1994, 54, 4958. Rubin, E.; Kharbanda, S.; Gunji, H.; Kufe, D. Mol. Pharmacol. 1991, 39, 697. Wang, L.-K.; Johnson, R. K.; Hecht, S. M. Chem. Res. Toxicol. 1993, 6, 813. Austin, C. A.; Patel, S.; Ono, K.; Nakane, H.; Fisher, L. M. Biochem. J. 1992, 282, 883. Gupta, M.; Abdel-Megeed, M.; Hold, Y.; Kohlhagen, G.; Paull, K.; Pommier, Y. Mol. Pharmacol. 1995, 48, 658.

TOPOISOMERASE I-TARGETING DRUGS" NEW DEVELOPMENTS IN CANCER PHARMACOLOGY

Barbara Gatto and Leroy Fong Liu

I.

II.

III.

IV.

V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Topoisomerases are Essential Enzymes . . . . . . . . . . . . . . . . . . . B. TopoisomeraseoTargeting Drugs: Two Classes of Compounds . . . . . . . Eukaryotic Topoisomerase I . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Physiological Roles and Cellular Regulation . . . . . . . . . . . . . . . . B. Mechanism of Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Camptothecins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Camptothecin is an Antitumor Agent . . . . . . . . . . . . . . . . . . . . . B. Camptothecin is a Topoisomerase I Poison . . . . . . . . . . . . . . . . . C. Camptothecin Binding Site is Defined by Topoisomerase I Mutants . . . . D. Camptothecin-Cleavable Complex Interactions: Design of New Derivatives.. New Topoisomerase I-Targeting Drugs . . . . . . . . . . . . . . . . . . . . . . A. Specific Topoisomerase I Poisons . . . . . . . . . . . . . . . . . . . . . . B. Dual Topoisomerase I and II Poisons . . . . . . . . . . . . . . . . . . . . Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Advances in DNA Sequence-Specific Agents Volume 3, pages 39-65 Copyright 9 1998 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0203-8 39

40 40 40 41 41 42 43 43 45 46 48 53 54 57 61 61

40

BARBARA GATTO and LEROY FONG LIU I.

INTRODUCTION

A. Topoisomerasesare Essential Enzymes Topoisomerase-targeting drugs are a vast group of compounds exerting various pharmacological activities, characterized by their ability to target the nuclear proteins topoisomerases or their prokaryotic analogues, n-4 Topoisomerases represent a class of nuclear enzymes deputed to solve the topological problems arising during normal cellular processing of nucleic acids. 1 Since the isolation of the first enzyme able to catalyze topological reactions, the protein co in E. coli, 5 proteins with topoisomerase activity have been found in almost every organism studied. Their ubiquity, as well as the significant homologies in their primary sequences, is linked to the "mechanical" problems posed by the complexity of the nucleic acids to any helix-processing activity. Detailed analysis of the reactions catalyzed by topoisomerases and their roles in fundamental processes like replication, transcription, and recombination have been the subject of many reviews 4'6-9 and therefore will not be covered here. All topoisomerases share a common activity toward their nucleic acid substrate, namely the nicking-closing activity, l~ Based on their mechanism of action, topoisomerases can be classified into two classes, type I and type II enzymes, x More recently, based on sequence homologies and on biochemical properties, all known topoisomerases have been divided into three subgroups. ~xThe group whose targeting agents are reviewed in this chapter is represented by eukaryotic topoisomerase I and by vaccinia virus topo I. A recent paper, however, based on differences in the domain organization of the two enzymes, suggests that they belong to separate subfamilies, ~2 a hypothesis that was supported by the partial crystal structure of both enzymes. 13,14

B. Topoisomerase-Targeting Drugs: Two Classes of Compounds Any drug capable of inhibiting the catalytic activity of the enzyme could theoretically result in inhibition of the cell growth. The first drugs with recognized ability to target topoisomerases were however not classical enzyme inhibitors, but rather compounds capable of transforming the enzyme into cellular poisons. 15-17 For example, doxorubicin, a well known antitumor agent, showed the ability to trap an essential intermediate of the reactions catalyzed by topoisomerase II, the "cleavable complex." In this drug-DNA-protein complex the nucleic acid is broken and covalently linked to the enzyme, and the presence of these drug-stabilized breaks in the nucleic acid triggers the events leading to cell death, l Many compounds behaving like doxorubicin have been discovered, with pharmacological actions ranging from antiproliferative to antibacterial, antifungal, and antitrypanosomal. ~'18-21 More recently, a different class of topoisomerase II inhibitors such as the ICRF compounds were identified. 22 We will follow the classification of topoisomerase I inhibitors proposed previously. II Hence, class I drugs are all

Topoisomerase I-Targeting Drugs

41

compounds which are able to stabilize the cleavable complex, known collectively as "topoisomerase poisons," while to class II belong all drugs that inhibit topoisomerases catalytic functions in other ways. While a conspicuous number of compounds, belonging to either class I or II drugs, are known to target the type II eukaryotic enzymes, camptothecin and its derivatives are until recently the only drugs with significant pharmacological activity linked to the poisoning of eukaryotic type I topoisomerase. However, research on eukaryotic topoisomerase I-targeting drugs has recently entered an explosive stage, leading to the isolation of new compounds belonging to diverse chemical classes. Before reviewing them, we will briefly examine the topoisomerase I mechanism of action and its significance as a target for antitumor chemotherapy.

II.

EUKARYOTIC TOPOISOMERASE

I

A. Physiological Roles and Cellular Regulation All type I eukaryotic enzymes are monomeric nuclear proteins capable of relaxing positive and negative DNA supercoils without energy cofactors, ll The first eukaryotic type I topoisomerase was identified by DNA superhelical untwisting activity in mouse cell extracts. 23 Its physiological importance is linked mainly to the swiveling activity necessary to unwind the DNA helix during replication, as well as to the relaxation of DNA supercoils generated by RNA polymerase during transcription. 24'25 Quite surprisingly however, topoisomerase I is not essential for cell viability in yeast, 26'27 although Drosophila topoisomerase I plays an essential role during fly development. 28 All eukaryotic topoisomerases I share significant homologies and catalyze the same type of in vitro reactions. 29 One of these suggests a role for topoisomerase I in illegitimate recombination: the enzyme can ligate oligonucleotides with free 5'-hydroxyls to the bound nicked-DNA strand, provided that sequence complementarities exist between the oligonucleotides and the noncleaved DNA strand. The involvement of eukaryotic topoisomerase I in recombination is further supported by an increase in illegitimate recombination events in yeast strains overexpressing topoisomerase I. Furthermore, the "hot spot" sequences for these integrations correspond to the preferred in vitro topoisomerase I induced cleavage sites. 3~ The activity of the enzyme is regulated in the cell by posttranslational modifications. The best characterized of these modifications is phosphorylation, 31'32 and differently phosphorylated forms of topoisomerase I appear during the cell cycle. 33'34 Certain types of phosphorylation can apparently enhance the relaxation activity of topoisomerase 1 35 and could be responsible for the lack of correlation between mRNA and protein levels during events like cell proliferation and cell differentiation. ADP ribosylation by poly(ADP-ribose)polymerase on the contrary abolishes topoisomerase I relaxing activity in vitro. 36'37 Since the addition of

42

BARBARA GATTO and LEROY FONG LIU

poly-ribose is stimulated by DNA strand breaks, 38 this modification may play a role in the cellular responses to topoisomerase I drugs. A recent paper describes the phosphorylation by mammalian topoisomerase I on proteins belonging to the SR family of splicing factors. 39 Topoisomerase I lacks a canonical ATP binding domain, but can indeed bind ATP and phosphorylate selected serine residues at the arginine-serine rich region of SR proteins; since this kinase activity is blocked by the topoisomerase I poison camptothecin, and phosphorylation regulates the activity of SR proteins, a possible role of topoisomerase I in RNA splicing was suggested. 39

B. Mechanism of Action The catalytic domain of known eukaryotic topoisomerase I is highly conserved, with a sequence homology higher than 50%. 12'4~It is located close to the C-terminal of the protein primary sequence, and contains a consensus sequence of six amino acids surrounding the tyrosine responsible for the DNA transesterification reactions. The mechanism of action of topoisomerase I has been classically subdivided into discrete steps: DNA binding, DNA cleavage, swiveling (or DNA passage), DNA religation, and, finally, enzyme turnover. 4j The best characterized steps, and the most relevant to our discussion on topoisomerase I-targeting drugs, are the DNA binding, cleavage, and religation reactions. 1~ DNA

Binding Specificity

Information on topoisomerase I binding preferences are usually inferred from DNA cleavage patterns, and might be biased by this experimental limitation. The enzyme prefers to bind double-stranded DNA, or single-stranded DNA in regions of secondary structures. 43 The footprint of topoisomerase I bound to a specific recognition site of an intergenic Tetrahymena DNA sequence has revealed a protection region of 16-20 base pairs roughly symmetrical with respect to the cleavage site, 44 with the main noncovalent contacts taking place in the upstream region. Specific contacts have also been suggested in the downstream region. 45 Another consensus sequence for enzyme binding has been deduced from the analysis of other strong in vitro cleavage sites; 43 the common element in the different proposed consensus sequences seems to be a strong preference for a T in position -1.46 Downstream of the cleavage site, the conformation of the DNA, rather than its primary sequence, appears important; if the bent A-tract downstream of the preferred topoisomerase I cleavage site in Tetrahymena DNA is mutated, the frequency of the enzyme-induced cleavage is reduced or abolished. 47 The preference of topoisomerase I for bent to nonbent sequences, as well as for supercoiled to linearized DNA, 48 is consistent with its role as a relaxing enzyme. Topoisomerase I, thus, is sensitive to the presence of writhe in DNA. 29

Topoisomerase I- Targeting Drugs

43

The Cleavable Complex Upon noncovalent binding to its DNA substrate, a tyrosine residue at the enzyme active site reacts with the phosphodiester backbone. The result of this nucleophilic attack is a complex in which the tyrosine is covalently linked to the 3'-phosphate of the broken DNA end. This reaction intermediate, easily isolated by the use of protein denaturants like alkali or SDS, has been called the "cleavable complex. ''l The cleavage-religation reaction is completely reversible, as demonstrated by conditions that favor enzyme dissociation from the nucleic acid substrate, like high salt, heat, DNA addition, or simple dilution. In other words, the cleavable complex is in equilibrium with the "noncleavable" complex. The religation step is driven by the energy accumulated in the phosphotyrosine linkage: the free 5'-OH of the broken DNA strand can attack the enzyme-linked phosphate to restore the phosphodiester bond. During a cleavage-religation cycle the enzyme can perform its physiological role in relaxation and strand passage events. Two hypotheses have been proposed to explain the strand passage mechanism. The simplest explanation calls for a free rotation of the intact DNA strand around the nicked one, while in the second model the intact strand passes through the gate created by the enzyme in the broken DNA. 49 A way to study the cleavable complex in vitro and in vivo, and hence to obtain more information about the enzyme mechanism, is through the use of compounds able to enhance formation of the reaction intermediate. These compounds, whose prototype is camptothecin, can shift the cleavage-religation equilibrium toward the cleaved state by inhibiting the religation step. 5~ The block is completely reversible, unless transformed into an irreversible event by enzyme denaturation with SDS or alkali. In this way it is possible to isolate DNA fragments covalently linked to the 3'-end of the enzyme through a phosphotyrosine link. 5~ Much of what is known about topoisomerase I cleavable complex comes from studies using camptothecin, although the discovery of new drugs able to target the enzyme in a different mode will also be likely to accelerate our understanding of the enzyme mechanism.

III.

CAMPTOTHECINS

The interest in topoisomerase I as a target for antitumor chemotherapy has its basis in the successful clinical development of the alkaloid camptothecin. Only recently new drugs structurally different from camptothecin have been identified. In order to examine in more detail the most recent findings on the basic and applied pharmacology of eukaryotic topoisomerase I-targeted drugs, we will outline the principal features of their activity.

A. Camptothecin is an Antitumor Agent Camptothecin, whose structure is shown in Figure 1, is a natural alkaloid isolated in 196652 whose broad spectrum of antitumor activity in animal models was related

44

BARBARA GATTO and LEROY FONG LIU NH 2

~

"

N

0

0 OH 0

OH 0

9-amino-20-(S)-Ca mptothecin

20-(S)-Camptothecin

O o

OH u

CPT-I1

OH u

Topotecan

Figure 1. Camptothecin derivatives currently in clinical trials. to the inhibition of DNA and RNA synthesis. After these encouraging observations, the drug entered phase I and II clinical studies; however, due to the alkaloids' low solubility, the chosen formulation was its carboxylate sodium salt, obtained by hydrolysis of the lactone form. The early clinical trials proved the drug to be too toxic and ineffective, and further investigations were abandoned. 53 Renewed interest came more than a decade later with the discovery that camptothecin can induce in vitro formation of a cleavable complex in the presence of mammalian topoisomerase 1;50 the requirement of the intact lactone form for activity 5~ explained the failure of the early clinical trials and strongly suggested the involvement of topoisomerase I poisoning in mediating the antitumor effect of campthotecin. This hypothesis was proved correct using different approaches. First, classic structureactivity relationship studies on camptothecin derivatives established a correlation between the ability to trap topoisomerase I in vitro and the observed cytotoxic effect. 5~ In vivo, a genetic approach using yeast as a model showed that those cells whose topoisomerase I gene is deleted are insensitive to camptothecin, but drug sensitivity is restored by transfection with the human topoisomerase I


45

gene. 26'56A third and conclusive piece of evidence pointing out topoisomerase I as being the only intracellular target for camptothecin came from the observation that the drug-resistant phenotype in human cell lines is mediated either by lower protein levels or by structural alterations resulting from point mutations in the topoisomerase I gene. 57-61 It was later established that the enzyme levels, although relatively constant throughout the cell cycle, 62'63 are elevated in several types of leukemia 64 and in some cancers like primary colon adenocarcinoma. 65 When these and other type of solid tumors, implanted as xenografts in nude mice, were treated with 9-aminocamptothecin, a synthetic derivative of camptothecin, tumor growth was arrested, and the treated animals showed prolonged remission. These impressive results opened the way to clinical studies: at the present moment 9-amino-CPT, 20(S)camptothecin, the prodrug CPT-11, and the water soluble topotecan, all shown in Figure 1, are at different stages of clinical trials. 66-73 New studies on tumor biopsies confirm that topoisomerase I levels and its catalytic activity are elevated in colorectal tumors and prostate cancer, but not in other types of cancer like kidney tumors, lung, and breast. 74'75 Topoisomerase I drugs are therefore good candidates for the chemotherapy of selected types of tumors overexpressing the enzyme either as a result of altered transcription/translation, or by increased mRNA stability.

B. Camptothecin is a Topoisomerase I Poison Camptothecin is a topoisomerase "poison," since it interferes with the cleavagereligation equilibrium of the enzyme on its nucleic acid substrate, shifting this equilibrium toward the cleaved state. The final effect is a persistent and lethal DNA damage induced by the enzyme, exclusively in the presence of drug. It is assumed that camptothecin inhibits the enzyme at the religation step, although the issue remains controversial. 76'77 Since the shifted cleavable complex is fully reversible, camptothecin needs the intervention of other processes to show its effect. The main cellular event responsible for the cell killing is ongoing DNA synthesis, as proved by the finding that replication inhibitors like aphidicolin abolish camptothecin toxicity. 78'79It is likely that the collision between the replication fork and the cleavable complex transforms reversible single-strand DNA breaks into irreversible double-strand breaks, sensed by the cell as DNA damage. 79 Yeast cells with mutations in the DNA repair gene RAD52 are in fact extremely sensitive to the antiproliferative effects of camptothecin. 26 The actual cellular processes of the damage induced by camptothecin are however still unknown, and much effort is directed toward this area. 8~ Studies in yeast cells, mutagenized and selected for mutations able to restore the cell viability in the presence of drug, led to the isolation of dominant SCT (suppressor of camptothecin toxicity) mutants. Differently from "classical" mutants, their drug resistance is

46


neither due to altered enzyme catalytic activity nor to a significative decrease in drug uptake. 8~The growth rate of such mutants is unaffected by drug treatment but they are still sensitive to DNA damage induced by UV light or by mutagens. Although a mechanism of increased double-strand DNA-breaks repair could be involved, further characterization of the mutants are needed to clarify the biochemical pathways involved in the lethal effect of camptothecin.

C. Camptothecin Binding Site is Defined by Topoisomerase I Mutants Molecular interactions of camptothecin with its partners are complex: it does not bind isolated DNA 81 or topoisomerase I unless in the context of the ternary complex. 54 The drug presumably contacts various protein domains 82 in the cleft that represents the enzyme catalytic site. The nature and complexity of this binding site can be indirectly deduced by the analysis of topoisomerase I mutants resistant to camptothecin. As shown in Figure 2, many mutations were detected either in cell lines exposed to increasing concentrations of camptothecin, or one of its analogues, or by site directed/random mutagenesis of human topoisomerase I gene, transformed and selected in bacteria or yeast. All these mutations cluster in defined regions in the enzyme primary sequence, and are likely to represent interdomainal points of contacts, delimiting the catalytic site and the camptothecin binding domain. 82 All the reported mutations map in regions, or "homology domains, ''4~ with the highest identity in the primary sequence of topoisomerase I from different organisms, namely Homo sapiens, Drosophila melanogaster, Arabidopsis thaliana, S. cerevisiae, and S. pombei. 4~ It is evident from Figure 2 that some of these sequences are conserved even in the vaccinia virus enzyme, and are likely to define domains critical for enzyme catalysis. Of particular interest is the study of vaccinia virus topoisomerase I, a small enzyme of 314 amino acids with significant homologies to its human and yeast counterparts, but naturally resistant to camptothecin. 83 When the two conserved amino acids (721 and 722) preceding the human (or yeast) catalytic tyrosine are substituted with the corresponding amino acids found in vaccinia virus, the "chimeric" enzymes become camptothecin-resistant. 84 The corresponding exchange of the yeast amino acids into the vaccinia virus enzyme however does not yield a camptothecin-sensitive vaccinia virus topoisomerase I, 83 suggesting that additional domains not present in the simple viral enzyme are needed to form the active camptothecin pocket. Topoisomerase I from CEM/C2 cells, derived by stepwise selection by camptothecin of human leukemia cells CCRF/CEM, bear a mutation at Asn-722 too, that results in a twofold reduction of the enzyme catalytic activity. 85 Other mutants isolated in different cell lines map in this region, immediately surrounding the catalytic site of topoisomerase I. Two mutations at 717 and 729 were isolated from the ovarian cancer line CPT-2000, derived from A2780, 86 and the same mutation at 729 was found in PC-7/CPT, a lung cancer cell line selected for resistance to CPT-11. 87,88 While in the latter case the mutant enzyme has a reduced relaxation


47 T729A D533G D221V N722A !i I

il

G363C I~12(. ,'' i

"

'

i

J 35~nPP. LrRGRG.HPK

: '

t~nr~a~k.]~e I

139n D T V G l l tL

.......

,,

I

I1 U. U

ICOOH

|

! ,,r~rsK.i~.bPa. t#] i 267pstSKr'aYmn--tt|

~

5s3GL.AKVFRTYNAS 216GirIKd_IR~..gvn

[ 52,FDFiGKDsI 162ikFvGKDkv

!

[

Figure 2. Mutations in the human topoisomerase I sequence responsible for camptothecin resistance. The mutation site evidenced as a black bar is responsible for camptothecin sensitivity of vaccinia virus topoisomerase I. The amino acidic residues surrounding the underlined mutation sites are boxed. The upper capitalized sequence shows the consensus found for 6 eukaryotic topoisomerases i (from Ref. 31), with numberings corresponding to the human enzyme. The lower sequence is vaccinia virus topoisomerase I. The amino acid residues conserved both in the eukaryotic consensus and in vaccinia virus sequences are capitalized. "Dots" correspond to nonconserved residues, while "hyphens" indicate gaps due to sequence alignment. * Indicates that the mutation has been isolated in Chinese hamster cells. # Indicates a nonconserved residue in the human topoisomerase I sequence. activity, other mutants have a fully functional enzyme in the absence of drug. The best characterized of these type of mutants is found in CPT-K5 cells, a human lymphoblastic leukemia cell line that bears two point mutations in the topoisomerase I gene, at 533 and 583. Only the mutation changing Asp-533 to Gly is responsible for resistance, 89'9~ which could be the result of the recession of the surrounding domain from the enzyme surface. 89 The hypothesis of a structural alteration of the drug binding site is supported by the finding that CPT-K5 topoisomerase I binds and cleaves a subset of cleavage sites with higher efficiency than the wild-type enzyme. 91 Other mutations map in the regions spanning residues 500-600. The mutation at position 503 isolated from a Chinese hamster cell line maps in a very conserved region, and corresponds to position 505 of the human enzyme. 58 The camptothecin sensitivity of the mutant enzyme could be the result of an altered binding of the drug to the DNA-enzyme complex. Similarly, the

48


replacement of Asp-221 of vaccinia virus topoisomerase I with Val, highly conserved in the corresponding position of much of the topoisomerase I gene (588 in human sequence), results in a camptothecin-sensitive vaccinia virus topoisomerase I. 92 A third domain involved in the formation of the camptothecin pocket is around position 363. 93 The mutation Gly363Cys isolated from a pool of mutagenized clones expressed in a drug permeable yeast strain does not result in alterations of mRNA or protein levels, or in decreased relaxation activity in the absence of drug. Similarly, a mutation Phe to Ser at residue 361 isolated in U-937/CR, a human myeloid leukemia cell line selected for resistance to 9-nitro-20(S)-camptothecin, 94 yields an enzyme whose cellular levels and activity are comparable to the wild type, but is 10-fold less active in the presence of drug. A comparison with the recently published crystal structure of yeast topo I shows that these two mutations in the human enzyme map in a I]-hairpin like loop 14 that is part of a highly conserved protein region contacting DNA. Biochemical and structural studies show how this domain is part of the surface of camptothecin cleft, involved both in the formation of a ternary complex, and in the DNA cleavage-religation reaction. A different set of yeast mutants show the ability to cause extensive DNA cleavage, resulting in cell growth inhibition and death in the absence of drug. 95 Despite the identification of the mutated residues, the molecular basis of action of these "camptothecin-mimicking" enzymes is still unclear, but resembles the nicking activity of mammalian topo I at mismatched sequences. 96 New camptothecin-resistant mutants are continuously been engineered and isolated for mechanistic studies, but clinically relevant ones are likely to become a major problem once the drugs enter the market. The elucidation of the biochemical properties of these mutants in the light of the crystal structure of the enzyme is therefore urgent: the definition of the drug binding domain in the drug-DNA-enzyme ternary complex will lead to the design and synthesis of new topoisomerase 1-targeting compounds.

D. Camptothecin-Cleavable Complex Interactions: Design of New Derivatives The extensive structure activity relationship conducted on camptothecin was aimed at: (a) establishing the structural requirements for activity, and (b) obtaining new potent semisynthetic derivatives with a better pharmacokinetic profile than the parent compound. Both aims have been fulfilled: potent water-soluble derivatives have been developed and are entering or completing clinical trials. In addition, the whole body of research has led to the definition of a "minimal" model for the drug interactions within the cleavable complex, whose main features are shown in Figure 3. Camptothecins' potency lies in what actually represents a pharmacokinetical disadvantage, the intact lactone at ring E: its pH-dependent hydrolysis renders the


49

planar ring system docking in DNA-enzymecleft ....

I

I

NU :

5

~i ~ ( C~ND._4,0 ~-" 12

1

interaction with V enzyme/DNA surface

~ 1 ~

"~

OH

!

inte raction with enzyme/DNAsurface : Nu

, I

stereospecific interaction with enzyme

Figure 3. "Minimal" model for the interaction of camptothecin with topoisomerase I and the nucleic acid in the cleavable complex (see text for details). drug inactive. 97 The lactone-carboxylate equilibrium shown in Figure 4 is shifted at physiological pH toward the open form. In whole blood the active closed form is only 5.3% of the total, 98 and represents a compromise between the preferential binding of the open form to human serum albumin 99 and the interaction of the lactone with lipid bilayer of the erythrocyte membrane. 98 The carbonyl group of the lactone form that reaches its nuclear target undergoes a reversible nucleophilic attack by the enzyme. The lactam analogue of camptothecin, whose carbonyl group is less prone to hydrolysis but also less reactive to nucleophilic additions, is in fact inactive. 55 Ring E interactions within the enzymeDNA complex is supported by its stereospecificity: only the S (hydroxyl group in C-20) shows activity, 55'1~176 but the ethyl moiety at C-20 can be substituted by bulkier groups without significant loss of activity. 1~ The requirement for this a-hydroxyl group has been recently challenged by the finding that 18-noranhydrocamptothecin, a camptothecin derivative with an exo-methylidene group at C-20, shown in Figure 5, retains the enzyme inhibitory activity. 1~ This result, if confirmed by

11

9

7

9

1

5

9 '~

~

7

11

0

12

14

OH O

1

5N

L OH

OH OO- Na+

2

Figure 4. Camptothecin lactone form (1) readily hydrolyzes to the inactive carboxylate form (2) at physiological pH.

50


I

N

O

CH 2 0

Figure 5. 18-noranhydrocamptothecin, a new derivative with major modifications at the stereospecific position 20 of camptothecin.

in vivo data, will open a whole new field on alkylating camptothecin analogues, useful both as new chemotherapeutics and as tools for mechanistic studies. The camptothecins currently under development retain the t~-hydroxylactone moiety at the E ring, but have extensive modifications at other positions of the pentacyclic structure. Each ring of the camptothecin molecule has been functionalized in the attempt of obtaining either water-soluble drugs or simplified molecules with considerable activity. The structural requirements of camptothecin are, however, strict. The ABCD ring system has been simplified to various tetra-, tri-, biand even monocyclic rings, 1~ with total or partial loss of activity; 1~ overall planarity of the "left" part of the molecule is undoubtedly an essential feature. 1~ In the most active pentacyclic structure, substitutions at C-12 and C-14 are absolutely detrimental to activity, suggesting steric limitations imposed by the enzyme or by the DNA-enzyme surface. 1~176 Positions C-9 and C-10, on the contrary, can be functionalized by a wide variety of substituents, 1~ as exemplified by the appreciable activity of 9-amino-camptothecin and topotecan. 65'11oRing A, the "left" part of the molecule, is a possible point of contact between drug and enzyme, since camptothecin derivatives with an alkylating group at position C-9 bind covalently to nucleophilic residues on the enzyme. Ill Position C-7 in ring B defines another possible surface of contact ll2 and is a good site for functionalization as demonstrated by CPT-11,113-114 a water-soluble prodrug whose active form SN-38 has shown good activity in clinical trials, ll5 More recently, a C-7 alkylated derivative of camptothecin showed the ability to react with DNA in the presence of topoisomerase 1.116 The drug can alkylate the purine immediately +1 to the topoisomerase I-induced cleavage site, demonstrating close proximity of this "side" of camptothecin and DNA in the enzyme-DNA interface. Since the reaction results in an irreversible trapping of the cleavable complex, this compound may represent a lead in the design of new potent camptothecin derivatives. In addition, the substitutions at C-7, C-9, and C-10 lower the association with human serum albumins and prolong the drug's half-life. 99 A new line of research couples the accessibility of position 7 with previous results evidencing how an additional methylenedioxy ring fused at positions C-10 and


51

C-11 can significantly enhance camptothecin activity. 117 The high toxicity of 10,11-methylene-dioxycamptothecin (10,11-MDC) precluded its therapeutic use, but is lowered by functionalization at position C-9.117 The new water-soluble compounds GI149893 and GI147211, shown in Figure 6, bear hydrophilic piperazino groups at position C-7 of 10,11-(methylendioxy) and (ethylenedioxy)-(20S)camptothecin, respectively, 118 and are 4-5 times more soluble than topotecan. ~9 Their in vitro and in vivo potency is comparable or superior to that shown by the reference compound topotecan, both in cultured tumor cell lines and in colon tumors implanted as xenografts in nude mice. 119By analogy to camptothecin, both GI compounds exhibit comparable cytotoxicity to sensitive and multi-drug resistant ovarian cell lines, making them good candidates for the treatment ofMDR overexpressing tumors. Clinical evaluation of the drug G1147211 is currently underway. 12~ A similar series of compounds bearing various aromatic quaternary ammonium salts at position C-7 are undergoing testing also, and have so far showed good in vitro a,nd in vivo activities. 121

ON O

10,11-methylendioxycamptothecin

/

O

/CH3

CH 3

O

c2

N---.

o

--+-4

ON O

ON O GI-149893

GI-147211

Figure 6. The new water soluble derivatives GI149893 and GI147211 are formerly derived from 10,11 -methylenedioxy-camptothecin.

52


The search for new camptothecin analogues with improved water solubility led to the development of substituted l l-aza-derivatives, 122 already known to have good activity in vitro in the cleavable complex assay. 5~ Though most of these water-soluble derivatives have good in vitro activity, this does not correlate with the in vivo data, probably due to pharmacokinetical limitations. An exception is represented by the compound shown in Figure 7A, which showed excellent potency in delaying tumor growth in nude mice xenografts. 122The possible formation of a penta- or hexacyclic ring by an intramolecular H-bond between the substituent at C-10 and the A ring nitrogen makes the overall structure of this compound similar to that of 10,11-methylenedioxy-camptothecin, and adds further evidence to the potential clinical development of rigid camptothecin analogues. Amore dramatic approach has been followed by Lackey et al. 1~ who synthesized a series of rigid analogues formally derived from the condensation of the CDE drug rings with an oxodihydroindolylidene moiety (Figure 7B). Various derivatives, tested in the cleavable complex assay, showed activity in the same concentration range as camptothecin. A noteworthy observation is that the structure of 9-amino10,11-methylene-dioxy-camptothecin and these oxodihydroindolyl-derivatives can be superimposed, and show striking similarities in steric terms. 1~ The structure-activity relationship established for this series of drugs in fact confirm the

.OH

H-NNNN~ HN

J1.0050.4

38.2

OCH,

HCI

15

GN Z ~ N('ff)ZN~. '~)--' ~ s~/'N~~ --N ' ~_N~,'~ ~ ell;

OCn, HCi

16

~;N,,~

~r

NT 7.02

17.8

NT 5.16

2.68

HO

17

,f~ N cll; N.....J ~

ocH,

HCI

Note:

a50%Inhibitorydose.

inhibitory against the human cell lines. Exceptions are seen in those cases where only weak binding to DNA is observed. 37 B. In Vivo Anticancer

Activity of Bis-benzimidazoles

The cytotoxic activity seen with certain of the bis-benzimidazoles in vitro is sustained in vivo. A lead compound 7 has an IC50 of 0.02 ktg/mL against L 1210 cells and of 0.289 kt/mL against KB cells in vitro. Because of the superior'potency against leukemic cells it was then tested in mice against P388 (i.p. injection) at a single dose of 100 mg/kg. This resulted in a 45.1% increased life span compared with controls. 41

Bis-Benzimidazole Minor-Groove Binders

89

C. Cellular Effects Inhibition of Nucleic and Biosynthesis In Whole Cells Having observed cytotoxic effects in vitro against a range of cancer cells and confirmed the activity in vivo we now wished to explore the cellular effects of the bis-benzimidazoles as a first step in elucidating their mode of action. The more potent examples were found to inhibit both DNA and RNA synthesis in whole cells with IC5o values in the range of 5-8 l.tM.41 This suggested that interaction with the cellular nucleic acid contributed to the cytotoxic potency.

Effects of Analogues of Hoechst 33258 On Topoisomerase I and II-Mediated Activities The importance of the topoisomerase enzymes as chemotherapeutic targets has become evident in recent years. 42--44Agents which inhibit or enhance the activity of these enzymes may play a pivotal role in cell survival. Hoechst 33258 is a DNA minor-groove agent which, with the exception of its use as an antihelmintic, 45 has not been developed as a therapeutic agent. 46-49 Nevertheless, significant effects on topoisomerase mediated activities have been observed. The evidence for the effects of a series of Hoechst 33258 analogues on these key enzymes provided a new avenue for defining which structural modifications of this compound were important in drug design. The Hoechst agents used in this study were extensively characterized in an earlier report. 24Although these analogues all bound to the minor groove of B-DNA, certain sequence selectivity differences were noted. For example, while strong footprinting with agents 5 and 7 was observed over the same sequences as with Hoechst compound 1, some alterations in base specificity within the binding sites were observed. Both 5 and 7 showed increased binding to G.C pairs compared with compound 1. By contrast, the parent compound can only bind GC at the terminus of its binding site. 2 A generally reduced footprinting intensity was observed with agents 2-7 compared With 1, and the order of binding strength for the Hoechst compounds using either calf thymus or poly(dA-dT) DNA was 1 > 7 > 5 > 2 > 6 > 3>4. How do base specificity and binding strength relate to the observed effects of Hoechst compounds on topoisomerase-mediated activities? Analysis of the intact cell and nuclear data indicated that Hoechst compounds 1, 5, and 7 were good inhibitors of VM-26-induced cross-links and all showed similar footprinting intensities. 24 These same agents also showed strong binding affinities for calf thymus and poly(dA-dT) DNA. 24 Similarly, agents 2 and 3, with reduced footprinting and overall binding intensity, were also less effective at inhibiting cross-links. The increased preference for GC binding sites did not reduce the effectiveness of either 5 or 7 compared with 1 in the present study.

90

J. WILLIAM LOWN

While the cell and nuclear data generally reflected the drug binding characteristics defined with isolated DNA fragments, a good correlation between the binding affinities (Ka) and cross-link inhibition was not observed. Thus, as is to be expected, modulating factors other than strict affinity of the drugs for DNA must be considered when evaluating topoisomerase-mediated nuclear or cellular lesions. In intact cells drug uptake mechanisms may well be a factor. For example, the presence of a methoxy group at C6 was shown earlier to enhance cellular uptake of Hoechst 33258. 24 This substitution probably contributed to the enhanced activity of 7 in intact cells. Additionally, interpretation of the effects of Hoechst on cross-link formation in nuclei or intact cells requires consideration of the sites of VM-26, m-AMSA, or camptothecin action. These agents associate with the topoisomerase-DNA complex via binding to enzyme or DNA or both. 5~ Thus, inhibition of nuclear cross-link formation may result from interference with either the enzyme-DNA recognition site, the VM-26 (or m-AMSA or camptothecin) DNA binding site (if any), or the cleavage site of the DNA-enzyme complex. The association of nuclear and cellular DNA with histone and non-histone chromatin proteins may also contribute to an alteration in enzyme reactivity. Recent evidence suggested that minor-groove binders alter the association of DNA with its nucleosome core proteins. 57 Such an alteration may in itself be sufficient for bisbenzimidazole-induced reduction in topoisomerase-mediated cross-links. Alternatively, topoisomerase activity may be localized to specific chromatin domains (e.g. replicating or transcribing regions. 58-6~ Unless Hoechst binding is highly specific for the same chromatin regions as are bound by these enzymes, one would not expect a direct correlation between Ka and inhibition of topoisomerasemediated activities in nuclei. In contrast to the above effects on cross-link formation, all four Hoechst compounds were potent inhibitors of topoisomerase I and II catalytic activity (Table 4). Agent 2, which was less effective in the nuclear system than would be predicted from its K a, strongly inhibited both enzymes. Agent 3, differing from agent 1 in that the nitrogen in the N2 position has been replaced by oxygen, was the least effective. According to the earlier footprinting study, this orientation alters hydrogen bonding by the Hoechst analogue such that N1 of agent 3 becomes a potential hydrogen bond-accepting nitrogen rather than a hydrogen bond donor as in 1. 24 Thus, both footprinting intensity and topoisomerase inhibition are enhanced by a nitrogen in the N2 position. As in the nuclear and cellular studies, agent 7 was a very effective inhibitor of topoisomerase activity. Despite a lower DNA binding affinity than 1 (Ka = 1.5 compared with 3.1), this agent showed increased binding at GC-rich sequences as a result of the substitution of N at the C16 position. 24 However, the relevance of such sites to topoisomerase-mediated activity remains unclear until more evidence derived from additional (GC) n directed agents are examined. Tight binding to the minor groove might be expected to make this region of DNA less accessible to enzyme interactions. Indeed, the good correlation between


91

Table 4. Effects of Hoechst Analogues on Catalytic Activity of Topoisomerase II and Ia

Drug Required for 25% Reduction in Relaxation of pBR322 Analogue (pM) b Compound

1

2

3

7

Topoisomerase II

10.8

11.2

28

8.2

Topoisomerase I

1.6

2.3

5

1.6

Notes:

~q'opoisomerase reactions were performed as described in Materials and Methods with Hoechst 33258 analogue concentrations ranging from 1.0 to 40 ~tM. Maximum relaxation was that obtained in the absence of Hoechst and equaled 45 (topoisomerase II) and 49% (topoisomerase I) conversion of supercoiled to relaxed form I. Each point represents the average of 4 to 6 independent enzyme assays. bCompound numbers refer to structures in Figure 3.

Hoechst analogue Ka'S and the amount of drug necessary to inhibit topoisomerase activity by 25% was indicative of this relationship. 4~ Are the above effects on topoisomerase-mediated activities related, at least in part, to the ionic charge of the Hoechst molecule? Both stimulation and inhibition of topoisomerases have been observed with differing concentrations of divalent cations (e.g. Ca 2+ and Mg2+)62'63 polycations (e.g. histones, polyamines) 63-65 and the polyanionic heparin molecule. 66 However, the enzyme inhibitory concentrations of Hoechst compounds (1-40 ktM) reported in the study are relatively low compared with those required for modulation by charged moieties (--0.1 mM). Harshman and Dervan represent Hoechst 33258 as monocationic under physiological conditions. 67 Additionally, charge variations among the Hoechst compounds tested were minimal 24 and did not correlate with drug potency. In view of the correlation with K a, changes in topoisomerase catalytic activity were more likely to be the result of Hoechst DNA binding affinity than molecular charge effects. Agent 7 was more effective than either 1, 2, or 3 at inhibiting VM-26 cross-link formation in intact cells. In view of this observation, future studies will examine a series of structural analogues of the lead agent 7 for their ability to modify topoisomerase activities. Such studies with compounds differing only by one or two molecular substitutions should refine the definition of structures capable of selective modulation of topoisomerase activity in intact cell and subcellular preparations.

Correlation between Interference of Bis-benzimidazoles with Topoisomerase II-Mediated Reactions in Whole Cells and Anticancer Potency Anticancer agents usually express their cytotoxicity by multiple cellular lesions and pathways. Thus it is seldom, if ever, possible to point to an individual cellular effect as being exclusively or even predominantly responsible for cytotoxicity. Nevertheless, as a first step in trying to understand the mechanism of action of an

92

J. WILLIAM LOWN Table 5. Correlation between Interference with Topoisomerase II-Mediated Reactions in Whole Cells and Anticancer Cytotoxicity of Bis-benzimidazoles DNA-Protein Cross-links Inhibition (%)a

Compound Hoechst 33258

35 (100)

5

50(100)

6

> 90 (100)

Cywtoxic Activity 16'5o (lab/) 13

1 0.28

N

Cil)Z V ~1

OCH~

HC!

CII~I

v

CH~O

OC]['l$

HCI

Note: aAt 100 ~M MGBD.

anticancer agent, it is often instructive to look for such correlations. In the case of the novel bis-benzimidazoles they all exhibit, to a greater or lesser extent, inhibitory action on topoisomerases both in vitro and in whole cells. 40 Acorrelation is observed between the ability of the drugs to interfere with topoisomerase II-mediated DNA-protein cross-link formation by drugs such as etoposide or camptothecin, and their cytotoxic potency over almost 2 orders of magnitude (Table 5). 41

Relation between DNA Sequence Recognition of Bis.benzimidazolesand Cellular Effects We have seen above (Sections II, III.A and B) that it is possible by rational structure alteration of the bis-benzimidazole molecule to alter, to some extent, the characteristic sequence preference. It has also been possible to increase the cytotoxic potency (compared with Hoechst 33258) by ca. 100-fold. The question then arises whether the cellular and/or pharmacological effects are related to the sequence preference. It is clear that other pharmacological factors concerned with solubility, cellular uptake, and intracellular stability will play significant roles. Assuming, for the moment, that such factors will be comparable for this series of drugs, one may then examine for factors concerned with DNA binding capacity and sequence preference. The contribution of the DNA binding affinity to, for example, t.opoisomerase inhibition has been discussed above (Section V.C). Factors other than drug binding affinity may also influence enzyme activity. For example, an examination of the pBR322 cleavage recognition sequence reported for Drosophila


93

topoisomerase I168 revealed the presence of two binding sites for the more potent topoisomerase inhibitors 1 and 7. This sequence, with the Hoechst binding sites underlined, is: ACAATG $ CGCTCATC. (The arrow denotes the enzyme cleavage site.) Neither 2 nor 3 showed significant binding at either of these sites. By contrast, no such identity between pBR322 topoisomerase I recognition sequences and Hoechst analogue binding sites was apparent. Thus it is possible that sequence recognition effects may also contribute to overall potency.

VI. CONCLUSIONS AND PROSPECTS It has been possible by rational structural modification to modify the strict (AT), recognition of Hoechst 33258 so that appropriate new bis-benzimidazoles also accept GC-rich sequences. Detailed analysis by high field NMR of drug-oligonucleotide complexes provided structural details that both supported the concept of base site acceptance and assisted subsequent drug design. The new bis-benzimidazoles inhibit DNA and RNA synthesis and proved to be uniquely active in inhibiting topoisomerase I and II activity in whole cells. They are also quite cytotoxic and have IC50 values against KB cells of-0.4 ktg/mL and against L1210 cells of ca. 1 x 10-7 M. The antiproliferative activity in vitro expresses itself as anticancer activity in vivo with an ILS of 145% in mice following i.p. injection of P388 cells. A tentative correlation can be established between the ability of the drugs to inhibit topoisomerase II-mediated processes and their cytotoxicities over a range of - 100. Finally there is evidence that, in addition to DNA binding affinity, sequence selectivity of the bis-benzimidazoles also appears to play a role in the cytotoxic potency. The stage is now set for the further development of therapeutically useful agents based on the bis-benzimidazole structure.

ACKNOWLEDGMENT We thank the National Cancer Institute of Canada for financial support.

REFERENCES 1. Hurley, L. H. In Annual Reports in Medical Chemistry; Vol. 22; Bailey, D., Ed.; Academic Press, New York, 1987, Vol. 259, Chapter 26. 2. Rao, E. K.; Lown, J. W. Chem. Res. Toxicol. 1990, 3, 262. 3. Hurley, L. H.; Needham-van de Vanter, D. R. Acc. Chem. Res. 1986, 19, 230. 4. Lee, M. D.; Dunne, T. S.; Siegel, M. M.; Chang, C. C.; Morton, G. O.; Borders, B. D. J. Am. Chem. Soc. 1987, 109, 2464. 5. Stubbe, J.; Kozarich, J. W. Chem. Rev. 1987, 87, 1107. 6. Lown, J. W. In Molecular Aspects of Anti-Cancer Drug Action; Neidle, S., Waring, M.J., Eds.; MacMillan Press: London, 1983, Vol. 3, p. 283. 7. Lown, J. W. Anticancer Drug Design 1988, 3, 25. 8. Lown, J. W. Org. Prep. and Proceed. Intl. 1989, 21, 1.

94

J. WILLIAM LOWN

9. Lown, J. W. In Molecular Basis of Specificity in Nucleic Acid-Drug Interactions; Pullman, B., Jortner, J., Eds.; Kluwer Academic Publishers: Dordrecht, Netherlands, 1990, p. 103. 10. Lown, J. W. Antiviral Res. 1992, 17, 179. 11. Lown, J. W. Chemtracts - Org. Chem. 1993, 6, 205. 12. Kissinger, I.; Krowicki, K.; Dabrowiak, J. C.; Lown, J. W. Biochemistry 1987, 26, 5590. 13. Lee, M.; Krowicki, K.; Hartley, J. A.; Pon, R. T.; Lown, J. W. J. Am. Chem. Soc. 1988, 110, 3641. 14. Bailly, C.; Catteau, J-P.; Henichart, J-P.; Rezska, K.; Shea, R. G.; Krowicki, K.; Lown, J. W. Biochem. Pharmacol. 1989, 38, 1625. 15. Harshman, K. D.; Dervan, P. B. Nucleic Acids Res. 1985, 13, 4825. 16. Commings, D. E. Chromosoma 1975, 52, 229. 17. Holmquist, G. Chromosoma 1975, 49, 333. 18. Pjura, P. E.; Grzeskowiak, I.; Dickrson, R. E. J. Mol. Biol. 1987, 197, 257. 19. Teng, M-K.; Usman, N.; Frederick, C. A.; Wang, A. H-J. Nucleic Acids Res. 1988, 16, 2671. 20. de C.T Carrondo, M. A. A. DE; Coil, M.; Aymami, J.; Wang, A. H-J.; van der Marel, G. A.; van Boom, J. A.; Rich, A. Biochemistry 1989, 28, 7849. 21. Parkinson, J. A.; Barber, J.; Douglas, K. T.; Rosamund, J.; Sharpies, D., Biochemistry 1990, 29, 10181. 22. Searle, M. S.; Embrey, K. J. Nucleic Acids Res. 1990; 18, 3753. 23. Fede, A.; Labhardt, A.; Bannwarth, W.; Leupin, W. Biochemistry 1991, 30, 11377. 24. Bathini, Y.; Rao, K. E.; Shea, R. G.; Lown, J. W. Chem. Res. Toxicol. 1990, 3, 268. 25. Singh, M. P.; Joseph, T.; Kumar, S.; Bathini, Y.; Lown, J. W. Chem. Res. Toxicol. 1992, 5, 597. 26. Wang, W.; Lown, J. W. J. Med. Chem. 1992, 35, 2890. 27. Bathini, YU.; Lown, J. W. Syn. Commun. 1990, 20, 955. 28. Bathini, Y.; Lown, J.W. Syn. Commun. 1991, 21, 215. 29. Stephens, E F.; Bower, J. D. J. Chem. Soc. 1949, 2971. 30. Dubey, P. D.; Ratnam, C. F. Indian J. Chem. 1979, 18B, 428. 31. Kuznetsov, V. A.; Galabadzhiu; Ginsburg, G. E J. Org. Chem. USSR 1987, 23, 576. 32. Morgan, A. R.; Lee, J. S.; Pulleyblank, D. E.; Murray, N. L.; Evans, D. H. Nucleic Acids Res. 1979, 7, 547. 33. Baguley, B. C. Mol. Cell. Biochem. 1982, 43, 167. 34. Kumar, S.; Joseph, T.; Singh, M. P.; Bathini, Y.; Lown, J. W. J. Biomol. Struct. Dyn. 1992, 9, 853. 35. Feigon, J.; Denny, W. A.; Leupin, W.; Keams, D. R. J. Med. Chem. 1984, 27, 450. 36. Pullman, A.; Pullman, B. Quart. Rev. Biophys. 1981, 14, 289. 37. Kumar, S.; Bathini, Y.; Zimmermann, J.; Pon, R. T.; Lown, J. W. J. Biomol. Struct. Dyn. 1990, 8, 331. 38. Adnet, F.; Liquier, J.; Taillandier, E.; Singh, M. P.; Rao, K. E.; Lown, J. W. J. Biomol. Struct. Dyn. 1992, 10, 565. 39. Bailly, C.; Colson, P.; Houssier, C.; Wangh, H.; Bathini, Y.; Lown, J.W.J. Biomol. Struct. Dyn. 1994, 12, 173. 40. Beerman, T. A.; McHugh, M. M.; Sigmund, R.; Lown, J. W.; Rao, K. E.; Bathini, Y. Biochim. Biophys. Acta 1992, 1131, 53. 41. Lown, J. W., unpublished results. 42. Hsiang, Y-H.; Wu, H-Y.; Liu, L. E Biochem. Pharmacol. 1988, 37, 1801. 43. Lock, R. B.; Ross, W. E. Anti-Cancer Drug Design 1987, 2, 151. 44. Wang, J. C. Biochim. Biophys. Acta 1987, 909, 1. 45. Friedheim, E. A. H. Bull. WHO 1974, 50, 572. 46. Woynarowski, J. M.; McHugh, M.; Sigmund, R. D.; Beerman, T. A. Mol. Pharmacol. 1989, 35, 177. 47. McHugh, M. M.; Woynarowski, J. M.; Sigmund, R. D.; Beerman, T. A. Biochem. Pharmacol. 1989, 38, 2323. 48. Woynarowski, J. M. Sigmund; R. D.; Beerman, T. A. Biochemistry 1989, 28, 3850.

Bis-Benzimidazole Minor-Groove Binders 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69.

95

McHugh, M. M.; Sigmund, R. D.; Beerman, T. A. Biochem. Pharmacol. 1990, 39, 7077. Hsiang, Y-H.; Hertzberg, R.; Hecht, S.; Liu, L. E J. Biol. Chem. 1985, 260, 14873. Nieson, E. M.; Tewey, K. M.; Liu, L. E Proc. Natl. Acad. Sci. USA 1984, 81, 1361. Chen G. L.; Yang, L.; Rowe, T.; Halligan, B. D.; Tewey, K. M.; Liu, L. J. Biol. Chem. 1984, 259, 13560. Tewey, K. M.; Chen, G. L.; Nelson, E. M.; Liu, L. J. Biol. Chem. 1984, 259, 9182. Rowe, T.; Kupfer, G.; Ross, W. Biochem. Pharmacol. 1985, 34, 2483. Pommier, Y.; Covey, J.; Kerrigan, D.; Mttes, W.; Markovits, J.; Kohn, K. Biochem. Pharmacol. 1987, 36, 3477. Hertzberg, R.; Caranfo, J. J.; Hecht, S. M. Biochemistry 1989, 28, 4629. Portugal, J.; Waring, M. J. Nucleic Acids Res. 1986, 14, 8735. Ness, P. J.; Koller, T.; Thoma, E J. Mol. Biol. 1988, 200, 127. Champoux, J. J. J. Virol. 1988, 62, 3675. Culotta, V.; Sollner-Webb, B. Cell 1988, 52, 585. Rowe, T. C. Wang; J. C.; Liu, L. E Mol. Cell. Biol. 1986, 6, 985. Osheroff, O. Biochemistry 1987, 26, 6402. Javaherian, K.; Liu, L. Nucleic Acids Res. 1983, 11, 461. Srivenugopal, K. S.; Morris, D. R. Biochemistry 1985, 24, 4766. Pommier, Y.; Kerrigan, D.; Kohn, K. Biochemistry 1989, 28, 995. Ishii, K.; Futaki, S.; Uchiyama, H.; Nagasawa, K.; Andoh, T. Biochem. J. 1987, 241, 111. Harshman, K. D.; Dervan, P. B. Nucleic Acids Res. 1985, 782, 4825. Andersen, A. H.; Christiansen, K.; Zechiedrich, E. L.; Jensen, P. S.; Osheroff, N.; Westergaard, O. Biochemistry 1989, 28, 6237. Gellert, M.; Mizuuchi, K.; O'Dea, M. H.; Nash, H. A. Proc. Natl. Acad. Sci. USA 1976, 73, 3872.


SEQUENCE-SPECIFIC RECOGNITION AN D MODIFICATION OF DOUBLE-HELICAL DNA BY MINOR-GROOVE BINDING CONJUGATES STRUCTURALLY RELATED TO N ETROPSI N AN D DISTAMYCI N

Christian Bailly I~ II. III.

IV.

Cancer, Genetics, and Chemotherapy . . . . . . . . . . . . . . . . . . . . . . . Binding to DNA and Antitumor Activity . . . . . . . . . . . . . . . . . . . . Sequence-Specific DNA Recognition . . . . . . . . . . . . . . . . . . . . . . A. Proteins and Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Minor-Groove Binding Drugs . . . . . . . . . . . . . . . . . . . . . . . Design of Sequence-Specific Minor-Groove Binding Drugs . . . . . . . . . . A. Recognition of Long AT Sequences . . . . . . . . . . . . . . . . . . . .

Advances in DNA Sequence-Specific Agents Volume 3, pages 97-156 Copyright 9 1998 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 0-7623-0203-8 97

98 100 102 102 105 107 108

98

CHRISTIAN BAILLY

B. The Lexitropsin Approach: From Monomers to Dimers . . . . . . . . . . V. Netropsin and Distamycin Derivatives Equipped with a Binding/Modifying Element . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Conjugation with Polyamines: The Microgonotropen Approach . . . . . . B. Conjugation with a DNA Alkylating Group . . . . . . . . . . . . . . . . C. Conjugation with a Photosensitive Group . . . . . . . . . . . . . . . . . D. Conjugation with a Metal-Complexing Group . . . . . . . . . . . . . . . E. Conjugation with an Enediyne Structure . . . . . . . . . . . . . . . . . . E Conjugation with a DNA Intercalating Drug . . . . . . . . . . . . . . . . VI. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

110 118 118

119 123 127 131 132

144 146 146

I. CANCER, GENETICS, AND CHEMOTHERAPY Cancer can be adequately described as a breakdown of one or several cellular growth controlling systems. Malignant tumors arise from a long sequence of intertwined molecular events. The first step in the sequence, as well as several other latter stages, is believed to result from DNA damage which affect either the expression or the biochemical function of particular genes. Cancer is a malady of genes. 1 In most cases, the causes of point mutations, rearrangement in DNA, gene amplification, or chromosomal translocation, and the exact mechanisms by which they occur, remain enigmatic. However, if these molecular events cannot yet be predicted they can often be detected. The continuously growing body of information on the molecular biology of cancer has brought a new vision with which to develop novel approaches to fight against cancer. By understanding the mechanisms responsible for the activation of oncogenes and the ensuing malignant transformation, we have acquired new devices for the prevention and diagnosis of cancer. With the advent of gene therapy we may acquire efficient devices for the treatment of cancer. Local therapies using surgery, radiotherapy, and certain forms of chemotherapy often permit tumor eradication when the malignant cells are confined to the treated area. This is so in about one-third of cancer patients but for the majority some form of systemic therapy is required. Patients with widespread cancer have to face treatment combining high-dose regimens of chemotherapy using cytotoxic drugs and sometimes radiotherapy. While there are numerous successes for Hodgkin's disease, non-Hodgkin's lymphoma, childhood leukemia, and certain forms of testicular cancers, only a small proportion of the advanced cancers such as those of the lung, breast, and colon can be cured. Even the most sophisticated, the most destructive, and the gloomiest treatments cannot get rid of certain forms of common cancers. It is sad to have to admit that despite the enormous efforts in new drug development, drug combination with cytokines, and the improved high-dose regi-

Recognition and Modification of Double-Helical DNA

99

mens, there still remain a significant proportion of cancers that do not respond to any of the available therapies. In most cases, if chemotherapy fails, death results. In recent years new hopes have been founded on the use of oligonucleotides to inhibit specifically the expression of certain genes and/or the production of cellular oncoproteins. The utilization of short DNA fragments to inhibit, or to facilitate in certain circumstances, the expression of oncogenes or tumor-suppressing genes may provide efficient ways to selectively control the development of tumor cells. Oligonucleotides have proved very useful as tools to help decipher the pathways controlling cell growth and differentiation. The use of natural and chemically modified oligonucleotides has largely contributed to our growing understanding of oncogene action and the numerous molecular events leading to malignancy. But so far, antisense and antigene oligonucleotides rather constitute tools for genetic analysis than new anticancer drugs. The antisense (RNA-targeted) and antigene (DNA-targeted) approaches offer rational means to interfere specifically with the function of oncogenes. 2-4 One may conceive and hope that soon oligonucleotides will take their place in the armament against cancer. However, there is still much to learn about the potential use of oligonucleotides before any therapeutic application can be contemplated. 5 Gene therapy is another emerging approach for cancer treatment. 6'7 Promising results have been obtained in a restricted number of trials dealing with advanced cancers. Genes coding for the tumor necrosis factor, interleukin-2, thymidine kinase, and the proteins K-ras and p53 are among those few genes which have been tested so far for gene transfer. In just a few years the use of molecular genetics to specifically manipulate tumor cells so as to provoke their destruction has become a palpable clinical reality. Gene therapy offers promising perspectives for enhancing tumor immunogenicity, vectoring cytokines to tumors, inserting drug activating genes, suppressing oncogene expression, replacing defective tumor suppressor genes, and marking tumor cells. 8 In the last 2 years we have witnessed the first successes of this approach but it is still too early to foresee the real potentiality of gene therapy in the daily practice of medical oncology. Our knowledge of the mechanisms of gene transfer are still too sparse for a true consensus picture to emerge. Alongside the .safety, ethical, and social concerns, there are numerous scientific problems still to solve before we can transfer genes efficiently and stably into tumor cells in vivo. This is not to deny that in the future it is likely that gene therapy will bring decisive clinical benefits to the treatment of chemo-resistant cancers and other genetic diseases. However, whatever the degree of hope, DNA as a pharmaceutical agentmbe it an antisense, an antigene oligonucleotide, or a transfer genemyet remains a cherished goal and not a reality. For more than 50 years and for the foreseeable future it has been and it will remain necessary to administer to patients a "small molecule" and in most cases a combination of drugs to fight the tumor and the disseminated malignant cells. 9 Over the last two decades impressive advances have occurred in the development of new molecules as well as refined applications of existing drugs. Nowadays, cancer

100

CHRISTIAN BAILLY

pharmacologists and clinicians dispense a somewhat confusing array of compounds with widely different mechanisms of action. There are about 60 potential therapies for the treatment of cancer among a total of more than 100 products under development worldwide. Each of the about 45 drugs which have been accepted as useful for the treatment of human tumors is the result of intense screening in vitro and in vivo, and has undergone years of intense clinical testing in many countries. Any new compound, as promising as it may appear after the primary screening, must be tested against a panel of tumor cells of different histological types to investigate selectivity of effect. The preclinical development is mandatory to identify acute and long-term toxic side effects, to determine a maximum tolerated dose, and to help define optimum schedules and therapeutic index. Only then can the phase I clinical evaluation begin, possibly followed by phase II and phase III clinical trials.

II. BINDING TO DNA AND ANTITUMOR ACTIVITY The different series of cytotoxic and noncytotoxic compounds used in chemotherapy are indicated in Table 1. Compounds which interact with nucleic acids represent by far the largest class of antitumor drugs. 1~ According to the mechanisms of interaction with DNA, the drugs are usually (and somewhat artificially) categorized into DNA alkylating, intercalating, groove binding, and cleaving agents. The major DNA-interacting anticancer drugs either approved for clinical use or still under clinical development are listed in Table 2. For all these drugs, DNA constitutes the

Table 1. Principal Series of Antitumor Drugs

Cytotoxic compounds antimetabolites antimitotic alkaloids DNA-alkylatingagents non-covalentDNA-bindingantibiotics topoisomeraseinhibitors tubulin interactingagents Noncytotoxic compounds hormones growth factors cytokines monoclonalantibodies modulatorsof resistance differentiation modifiers chemopreventionagents angiogenesiscompounds chemo- and radio-protectors.


101

Table 2. Principal Series of Antitumor Drugs Acting on DNA Used in the Clinic or Under Clinical Trials Alkylating agents nitrogen mustards mechlorethamine chlorambucil melphalan cyclophosphamide ifosfamide nitrosoureas carmustine lomustine aziridines thiotepa triethylenemelamine trenimon alkylsulfonates busulfan epoxides dianhydrogalactitol dibromodulcitol platinum analogues cisplatin carboplatin others procarbazine dacarbazine mitomycins

Intercalating agents anthracyclines daunomycin doxorubicin epirubicin idarubicin anthraquinones and derivatives mitoxantrone piroxantrone merbarone oxanthrazole acridines amsacrine nitracrine DACA others actinomycin D amonafide, mitonafide elliptinium DNA cleaving agents bleomycin antibiotics bleomycin A2 phleomycin tallysomycin Minor groove binders mitramycin adozelosin Others cristanol dynemicin misonidazole camptothecin derivatives

major but not unique intracellular target. Within each class of compounds one can find homologous series having similar, if identical, DNA binding capacity but displaying variable antitumor properties. For example, the literature is replete with DNA intercalating agents but only a few exhibit useful therapeutic properties. Moreover, the biological properties vary considerably: some intercalators are potent anticancer drugs (e.g. daunomycin, actinomycin, amsacrine), whereas other rather exhibit antiparasitic (e.g. lucanthone, pyronaridine), antiviral (e.g. tilorone),

102

CHRISTIAN BAILLY

or antimicrobial (e.g. ethidium) activities. II Other potent intercalators are carcinogenic (e.g. certain polycyclic aromatic hydrocarbons). Moreover, some intercalators display a GC selectivity, whereas others prefer AT sequences.12-16 Therefore, if in the 1960--1970s DNA was considered the principal, if not the unique target for antitumor drugs, nowadays DNA is widely considered as being a possible but neither a unique nor a mandatory intracellular target for antitumor action. This dogma was best illustrated recently with the discovery of potent antitumor drugs such as taxol which do not interact at all with DNA. Therefore, on the one hand, it seems possible to disregard DNA in the rational scheme of antitumor drug discovery but, on the other hand, the demonstration that the malignant transformation of a cellmbe it a cell of the blood, lung, liver, or brain---due to an alteration of particular genes, strongly call for the development of DNA-interacting substances. 17'18Moreover, DNA is probably the best structurally characterized target. Its large size and its seemingly repetitive and ordered structure make it an attractive receptor for drug design. A nonspecialist tracking the literature relevant to DNA structure might conclude that all aspects of DNA structures are now elucidated by the advances of X-ray crystallography, NMR, and many other sophisticated spectroscopic and biochemical techniques. It is true that we possess an overwhelming body of information on DNA conformation. Numerous three-dimensional structures of oligonucleotides have been solved to atomic resolution; the notion of flexibility, bendability, and supercoiling become more and more palpable. But despite all these important discoveries, DNA remains an elusive target for selective drug action. 19 The search for sequence-specific DNA binding ligands has attracted considerable interest over the last 15 years but at the present day there neither exists a drug capable of recognizing a unique sequence in DNA nor one can certify that chemotherapeutic selectivity will be improved by virtue of an increase in sequence selectivity.

III.

SEQUENCE-SPECIFIC D N A RECOGNITION

There are two channels of information in DNA: the minor-groove and major groove which present different characteristics in terms of hydrogen bonding and electrostatic potential. The conventional wisdom argues that proteins access the genetic information via the major groove, whereas drugs use the minor-groove to read DNA. Although this idea remains on the whole correct, this simplistic view is gradually changing with the discovery of proteins which interact within the minor groove and small molecules which can reside in the major groove upon binding to DNA. The design of sequence-selective small molecules has largely profited from the elucidation of the mechanism of protein-DNA recognition.

A. Proteins and Peptides DNA binding proteins such as activators, repressors, and certain endonucleases have the capacity to recognize DNA with a high level of sequence specificity. The


103

molecular recognition process between a given transcription factor and its specific target sequence permits precise regulation of gene expression and other genetic events. The sequence-specific protein-DNA recognition is mediated by structural elements such as the well characterized helix-turn-helix, leucine-zipper, and the Cys-Cys or Cys-His zinc-finger motifs. 2~ The three-dimensional shape of the motif, which is to a large extent determined by the amino acid sequence of the protein, ideally adapts to that of the target nucleotide sequence (and vice versa) but there is no simple recognition code whereby a given base pair on the DNA is recognized by a given type or combination of amino acids on the protein. DNAprotein recognition involves a complex network of interdependent hydrogen bonding, van der Waals contacts, and ionic interactions. The high level of complexity often observed in protein-DNA interactions makes it difficult but, however, possible to envisage the design of protein-like effectors capable of recognizing specified DNA sequences. 22 Moreover, during the past decade it has become obvious that in most cases, in addition to the direct recognition (digital readout), it is necessary to consider an indirect recognition (analogue readout) as a major determinant of sequence selectivity. The local conformation of the DNA molecule can have a profound, if not a dominant, influence on protein-DNA interactions. 23 For these reasons, the structural domains encountered in proteins do not yet offer many possibilities for designing sequence-specific DNA-reading small molecules. Proteins or peptide motifs have been coupled with DNA binding/cleaving functionalities but even if sequence-specific binding has been achieved these molecular tools are more valuable for genetic analysis than for therapeutic purposes. Alongside the regulatory proteins which make sequence-specific contacts with the edge of base pairs exposed in the major groove of DNA, there exist a category of proteins which recognize DNA sequences primarily via interactions within the minor groove. 24 For example, the bacterial integration host factor (IHF) and the non-histone chromosomal protein HMG-I bind to the minor groove of DNA. 25'26 A number of recent studies have revealed that these proteins often possess particular proline-rich amino acid sequences that may be considered as a kind of architectural domain capable of recognizing minor groove features. The list of proteins that interact with specific sequences via the minor groove of DNA is rapidly growing. The list includes several chromosomal proteins (e.g. histone HI) and related proteins involved in DNA condensation (e.g. the bacterial HU protein) as well as certain gene-regulatory proteins such as the RNApolymerase II transcription factor TFIID which, on binding to the TATA box element, results in DNA bending. 27-29 Most of these minor groove binding proteins bind preferentially and sometimes specifically to AT-rich sequences. 24'3~ It is likely that the proteins recognize a localized sequence-dependent DNA structure, i.e. a narrow minor groove or intrinsic bend than rather a primary AT sequence (analogue readout as opposed to digital readout). The best studied model for protein-DNA minor groove recognition is perhaps the SPXX motif (S for Ser; P for Pro, and X for a basic amino acid residue such as

104

CHRISTIAN BAILLY

K for Lys or R for Arg). 34-37 The SPKK peptide is believed to adopt an H-bond stabilized I]G-turn structure that confers to the peptide a crescent shape conformation, complementary to that of the minor groove and would thus permit the adjacent amide group to face the bases so as to hydrogen bond with A.T base pairs. 38'39 Although the precise structures of such SPKK motifs have yet to be experimentally established, it is clear that they actively participate in sequence-selective DNA recognition. Such basic peptide motifs have been shown to participate in the mechanisms of binding to DNA of TFIID and HMG nuclear proteins, but they are also encountered with major-groove binding proteins such as GCN4 and the 434 repressor. They are supposed to contribute actively to the interaction with DNA, perhaps also to the sequence recognition process. The degree of sequence selectivity of minor-groove binding proteins is generally several orders of magnitude lower than that of major-groove binding proteins. However, a moderate sequence selectivity such as that of HMG- 1 protein may be relevant to the in vivo function of the protein, providing that the target sequences are sufficiently abundant. 4~ As such, minor-groove binding proteins and SPKK peptide motifs provide interesting opportunities for designing drugs capable of recognizing AT sequences selectively.41-43 Minor groove recognition of DNA by proteins can also occur by oc-helix motifs. For example, the DNA binding domain of the purine repressor PurR contains both a helix-turn-helix motif and o~helices that make base specific contacts in the major and minor grooves, respectively. 44 A structural analogy between minor-groove binding drugs and SPKK-containing peptides was proposed to explain the selective interaction with AT sequences. 34'35 The antibiotics and drugs netropsin (1), distamycin (2), pentamidine (3), Hoechst 33258 (4), DAPI (5), and berenil (6) all adopt a crescent shape conformation roughly complementary to the spiral of the DNA double helix. They share the common properties of binding to DNA by displacing the spine of hydration within AT-rich sequences of the minor groove. 45-49 The bis-benzimidazole derivative Hoechst 33258 (4) effectively competes with a (SPKK)6 peptide for binding to

H2N'~NH

H2N"

NH

HN

(1) netropsin

H2N'~NH

HN,~.NH2

HN O

HN O

HN

HN H3

~r--N,.

HN

o

t

CH3 (2)

distamycin

H3

0 ~.,,,.~0 T

~

NH2

(3)

pentamidine


105

/CH3

N"

HN

HO~"-~ . . /

L)

,, NH2

(4)

(5)

Hoechst 33258

~

H2N

DAPI

(6) berenil

DNA 34'35and both ligands preferentially recognizes the same AT-rich sequences. 5~ These experimental results together with the predictive I]-turn structure of the SPKK peptide have reinforced the idea that the SPXX motif represents an "AThook. ''51 However, recent NMR studies argued against such a proposal but showed that related basic peptides containing PRGRP (but not PKGKP) motif bind specifically to AT sites. The authors postulated that the arginine guanidinium groups are major determinants of the AT specificity. Unlike the SPKKSPKK octapeptide, the SPRKSPRK peptide apparently does not fold into a I]-turn structure but adopts an extended conformation suitable to fit into the helicoidal cleft of the DNA, as observed with the well-characterized minor-groove binding drug netropsin. 52

B. Minor-Groove Binding Drugs Many low molecular weight molecules such as 1-6 fit within the minor groove of DNA but only a few exhibit antitumor properties. Such is the case for the antibiotics CC-1065 (7) and mithramycin (8) isolated from fermentation broths of Streptomyces zelensis and S. plicatus, respectively. Mithramycin (Plicamycin| belongs to the aureolic acid group of antibiotics and consist of a heterocyclic chromophore substituted with a hydrophilic side chain, and two carbohydrate moieties. Mithramycin has antitumor activity in disseminated embryonal cell carcinoma of the testis, but its primary uses are in the treatment of (1) hypercalcemia of malignancy unresponsive to other treatments, (2) Paget disease of bone, and (3) to promote maturation of myeloblasts in patients with the blastic phase of chronic myelogeneous leukemia. 53 The antitumor activity of mithramycin and its chemically related analogues, chromomycin and olivomycin, is believed to result from inhibition of DNA and RNA synthesis as a result of its complexation with DNA. It has been postulated that the antitumor activity of mithramycin results from selective blocking of formation of transcription initiation complexes of genes with GC-rich promoter regions, including the c-myc oncogene. Sequence-selective binding of mithramycin to GC-rich sequences blocks the binding of the transcription factor

106

CHRISTIAN BAILLY

OCH3

"~:"

~X, NH

o

/~~,,~O ~

/.,~ /.,&

H OCHa OH

"o

.,,~- O

/~,,.~

.~c" ",f ",f "n" o.o. o

(7)

(s)

cc- 1065

mith ramycin

OH

~'o-

H,c ?m

- ~ 6.u

o

H

Spl to a GC-rich element upstream to the TATA box of the promoter P1 of c-myc. 54'55 Mithramycin also inhibits the formation of an intermolecular triplex in the human c-Ki-ras promoter. 56 The sequence-specific recognition of DNA by mithramycin prevents the binding of important regulatory proteins (e.g. RNA polymerase II), thus inhibiting formation of transcription initiation complexes. Although early studies suggested that mithramycin intercalates into DNA or binds to the major groove, it is now clearly established that the antibiotic binds as a divalent cation coordinated head-to-tail dimer to the minor groove of DNA. 57'58 The antiparallel alignment of two Mg2§ mithramycin molecules generates a right-handed continuous hexasaccharide domain which spans six base pairs in the minor groove of DNA. The aureolic acid antibiotics are unique among minor-groove binders because they bind selectively to GC-rich sequences 59'6~ whereas all other groove binders display an AT selectivity. As for netropsin, the guanine 2-amino group is a key element for sequence-specific recognition of DNA by mithramycin. It would be most interesting to use mithramycin or chromomycin as a model for designing simpler molecules capable of reading GC sequences. However, given the absolute requirement for divalent cations such as Mg 2§ or Zn 2§ for binding to DNA, and the high complexity of the drug-DNA recognition process, at present it has not been possible to elaborate mithramycin-like molecules mimicking the carbohydrate side chains which actively participate in the GC specificity. Although both drugs interact within the minor groove of DNA, CC-1065 (7) is at the antipode of mithramycin in terms of DNA recognition. (+)-CC-1065 reacts with double-stranded DNA through N-3 of reactive adenine forming a covalent adduct that overlaps a 5 base pair region in the minor groove. CC-1065 reacts specifically with 5'-AAAAA* and 5'-PuNTs sequences and has a number of unique characteristics. It is one of the most potent cytotoxic drugs and shows efficacy against a wide variety of tumors in vitro. At the same time, it represents the "ideal" drug to bind to the minor groove of DNA. Among 1000 molecules tested for their ability to fit into the minor groove, CC- 1065 ranked first and thus appeared to be much more suited than netropsin to complement to the minor groove of an


107

AT-tract. 61 The discovery of unusual delayed lethality in mice precluded CC-1065 from clinical evaluation. However, its optimal adaptation to the minor groove coupled with its unique sequence recognition properties make CC-1065 an attractive model for designing less toxic analogues. A large number of derivatives have been synthesized and some of them such as the DNA-DNA interstrand cross-linker dimeric derivative U-7779, the prodrug U-80244, and adozelesin exhibit improved antitumor efficacy over CC-1065, and do not show delayed lethality. These drugs have now entered Phase I clinical trials. Comprehensive review on the DNA binding properties of CC-1065 (and the duocarmycins) and the development of antitumor analogues have been reported by Hurley 62'63 and Boger.64 The desire to synthesize molecules capable of recognizing specifically defined sequences in DNA has been a powerful stimulus during the last 10 years. Most of the approaches developed to create sequence-specific small molecules are based on the capacity of the antiviral drugs, netropsin (1) and distamycin (2), to bind selectively to AT-rich sequences. 65 Twenty years after the initial work of Wartell et al. 66 the study of the mechanism of binding of netropsin to DNA continues to arouse interest. The three-dimensional structures of netropsin and distamycin complexed to a variety of oligonucleotides containing an (A-T)4_6 tract have been solved to atomic resolution. 67-7~ In addition to crystallography, a multiplicity of spectroscopic techniques (visible, UV, IR, Raman, CD, LD, and NMR) have been used to further characterize netropsin-DNA complexes. Many thermodynamic and chemical studies have also been reported. Molecular modeling investigations have aided to define accurately the contribution of the electronic component of the DNA binding reaction and to design analogues. Taken together, these studies have led to a detailed understanding of the mechanisms whereby netropsin and distamycin engage in contact with the edge of A.T base pairs in the minor groove of B-DNA. Three main factors contribute favorably to the stability between AT sequences and netropsin or distamycin: (1) hydrogen bonding between the amide NH of the drug and thymine 0-2 and adenine N-3 atoms; (2) an overall shape complementarity as a result of close ligand-DNA van der Waals contacts; and (3) electrostatic interaction between the polyanionic DNA and the cationic drugs. It is assumed that the van der Waals interactions are pivotal in the sequence recognition. 72-74The knowledge of the mechanisms of interaction between AT sequences and the antibiotics netropsin and distamycin has greatly aided in the design of sequence-specific DNA binding molecules.

IV.

DESIGN OF SEQUENCE-SPECIFIC MINOR-GROOVE BINDING DRUGS

Several hundred derivatives of netropsin/distamycin have been synthesized since the first crystallographic studies. 67 This review is concerned with the different categories of netropsin and distamycin derivatives synthesized during the last 10 years and the many conjugates which are currently being developed.

108

CHRISTIAN BAILLY A. Recognition of Long AT Sequences

In 1985, oligo(N-methylpyrrolecarboxamide) molecules were synthesized with the aim of targeting large contiguous AT-rich sequences. 75 riffs-, tetra-, penta-, and hexa-N-methylpyrrolecarboxamide derivatives were shown to bind to sequences containing 5, 6, 7, and 8 contiguous A.T base pairs respectively. These studies led Dervan to deduce the following rule: the binding site size for an analogue containing n-1 pyrrole rings or n amide bonds is n + 1 base pair long. 76 This rule is valid for analogues containing up to six N-methylpyrrolecarboxamide residues. Longer molecules exhibit a lower affinity for DNA. A hepta(N-methylpyrrolecarboxamide)

NH

HN

N...jI

H

H3C'

H

-,,,o ~L.I~'CH3

O

flexible bis-netropsins (9) n= 1-10

HNH' CC~N:~NH NH

H3C

NHNH

~ i

~~N~c)HN

H

H

:

i ~

,

121

i

O (10)

o

i

rigid bis-netropsins

HN

~CH 3

109

Recognition and Modification of Double-Helical DNA HN'~NHz HN G/y G/y O H2N, 0 ~, ~ . 0 ~.-" /I~.~N" ..H\N'~,,"NH NH--..-/ ~ H / H Cys% Gly L ~N.~~ H3C~ N.'~ -n- ~ /.~NH H3c_NN/~ O--.'--,NH

l

HN

S/

O ",, JL Gty

I

HN,./~f'--NH

\

Gly ~1~ ":

0

Cys

('~N-CH3 ,,~ HN O

~....N

"CH3

H i /'~N,H.

"~0 "'_'~/N "~0 u

\ S

O

Gly

"~

Gly

2

Gly-Cys-(Gly)3 - - netropsin

netropsin - - (Gly)3-Cys-Gly NH2

(11)

,

derivative does not fit very well with the natural twist of the DNA presumably because the molecule gets out of phase with the base pairs along the minor groove floor of the double helix. Indeed, the pyrrolecarboxamide unit is approximately 20% longer than required to match perfectly the base pair rise in the minor groove. 77 Two alternative strategies to circumvent the lack of dimensional correspondence between the DNA and N-methylpyrrolecarboxamide-containing ligands have been envisaged. The first consists in joining two netropsin or distamycin molecules by a linker of suitable length to permit bidentate binding to DNA. A considerable number of bis-netropsins containing different types of linkers have been elaborated. 78-8~ Bis-netropsins coupled by a polymethylene tether (9) can engage in bidentate binding providing that the connector contains at least three methylene groups. However, owing to the flexibility of the aliphatic connector, monodentate binding of such bis-netropsins is also possible. In contrast, bis-netropsins possessing conformationally rigid linkers (10)can readily bind to sequences containing 8-10 consecutive AT base pairs without parasite binding to shorter sequences. 81'82 For chiral bis-netropsins, the stereochemistry of the linker is critical, and may be used for controlling the directionality in binding to DNA. 83The linker can also play an active role in the DNA recognition process. For example, the bis-netropsin analogue 11, in which the two netropsin residues are connected by a disulfide bond between two Gly-Cys-Gly peptides, binds better to sequences containing a central GC step such as (AT)3(GC)2(AT)3 than to a strict homologous (AT)n sequences. 84'85 The second strategy consists of replacing the carboxamide bond in netropsin and distamycin with shorter keto or amino linkages. Molecular modeling predicts that isohelical molecules such as 12 and 13 (called isolexins) would bind more tightly

110

CHRISTIAN BAILLY

H2N

H2NONH

H,~-H

~

~__N ~H '

HN~NH2

(12)

(13)

H2N'J~H

,~ 7"N N~

NH "~

0

.~0

HN==~ NH2 isolexins

H2N'~H

sCH3

ff

H H3C~NHf (14)

--

L.~N~N_CH3 HsC

vynilexins

"CH3 (15)

in the minor groove than their corresponding carboxamide analogues. 77'86 As for lexitropsins (Section IV.B), the use of imidazole or furan rings would favor recognition of GC sequences. In the same way, the computational studies predict that the replacement of the carboxamide bond of netropsin with an ethylene bond (compound 14) would permit an optimum fit to the minor-groove surface. 87 According to the computational measurements, molecules called vynilexins containing b o t h - C - - C - linkages and imidazole and/or furan heterocycles (e.g. compound 15) would display a considerable preference for GC-rich sequences. 88 Although very promising, so far neither the isolexin nor the vynilexin strategy has been experimentally tested.

B. The Lexitropsin Approach: From Monomers to Dimers Computational studies have suggested that the 2-amino group of guanines are primarily responsible for the lack of binding of netropsin and distamycin to GC sequences. 77 Indeed, the guanine 2-amino group protrudes into the minor groove and thus obstructs the access of these drugs to the floor of the groove. The fact that the 2-amino group constitutes a critical recognition element for binding of small molecules in the minor groove of DNA was recently clearly demonstrated. The expedient of preparing homologous DNA molecules containing inosine and/or 2,6-diaminopurine residues in place of guanine and/or adenine residues, respectively (Figure 1), has been used to investigate the exact role of the purine 2-amino group in determining the preferred binding sites on DNA for different groups of antibiotics including netropsin. 89'9~ Footprinting experiments using PCR-made DNA fragments containing A.T, G.C, DAP.T, or I.C base pairs chosen at will, revealed unambiguously that the binding of netropsin in the minor groove of DNA depends critically upon the presence or the absence of the 2-amino group of guanines. In normal DNA containing the four canonical bases, netropsin binds selectively to sequences lacking the 2-amino group, i.e. to AT sequences, as

Recognition and Modification of Double-Helical DNA N

H

.0~. _/CH3

~, A 'I~I........

"~..-N

N~.(

0

H

r-~N

...H-N ...... ....

/

"

N

111 H

.Q.

N~.~,

CH3

..... 0

H "N'H .......

H r~N

/

H

.... H-N .....

or

,

H.N,,H .... --

Figure 1. Structures of purine-pyrimidine Watson-Crick base pairs involving natural and unnatural nucleotides.

expected. DNA molecules in which all purine residues bear a 2-amino group (DAP-substituted DNA) do not bind netropsin at all, even at drug concentrations as high as 100 l.tM. On the opposite, the use of DNA molecules containing I.C and A.T base pairs (inosine DNA) results in a nonspecific binding of netropsin to DNA since in this case all sequences lack the purine 2-amino group. DNA in which the 2-amino group has been shifted from guanines to adenines (I+DAP doubly substituted DNA) show a complete redistribution of netropsin binding sites; netropsin binds exclusively to IC sequences. 89For all minor-groove binders tested (netropsin, distamycin, berenil, Hoechst 33258, DAPI, and pentamidine) as well as for intercalating and DNA-cleaving drugs, the repositionment of the guanine 2-amino group is sufficient to alter the drug-DNA recognition process. 89-94 Given both the strategic position of the guanine 2-amino group in the minor groove and its hydrogen-bonding capacity (it is the only H-bond donor exposed in the minor groove), it was proposed that the introduction of a H-bond acceptor heteroatom in the pyrrole rings of netropsin might permit the drug to bind to GC sequences. 77 Lown and coworkers have extensively exploited this concept and synthesized an impressive number of molecules christened lexitropsins or information-reading oligopeptides structurally related to netropsin and distamycin. 95'96 By substituting imidazole, thiazole, triazole, pyrazole, or oxazole heterocycles for the N-methylpyrrole rings of netropsin, it is possible to design drugs capable of binding to sequences containing one or two G.C pairs embedded in an AT sequence (Figure 2). Among the numerous lexitropsins synthesized so far, imidazole-lexitropsins such as compounds 16-18 display the most pronounced capacity for binding to GC-containing sequences. 97-1~ Thiazole-lexitropsins either accept or avoid a G.C base pair in their binding sites depending on the position of the sulfur atom. 1~ For example, the lexitropsin 19 with the sulfur atom directed into the minor groove does not bind to GC-containing sites, whereas the lexitropsin Thia-Nt 20 containing two thiazole rings with the sulfur atoms directed

112

CHRISTIAN BAILLY CH3

CH3

O fl

NI

~--/,~C\N,,~--ZN/

netropsin .... I'1'"-- C"

o .N_CH;2~ -~N ' ~H, ._ ' T : H i

NH

i

H

H

!

II

~

?

N-S

a inward '.

_

s ~

9 Z. I ~

,

~N

N~_.. N__

/

i~.

"~c

/

O/'%N

H,C/~'~j

.....

~

,-1

II

\\

O~ Z~

Ir

/a-

"

-

NH

CH3

pyrazole o

.~ ...........

..... ~'"

........,- ~ . _ o . ,

~c,~

nz"'C ~ t ' ~ - ~ ~ " ~ /'/'~_ lai

/I

.,

o ..."

c~(''" N-- N/ ,.~N_...// t

NHi I

'

/

iH

furan ,----.a

"

I%~X"

",, H '"-

""O:"";,'Nc H oxazole

.~; ~ _ i

%'~"~ " u , ,

St'~

II i

i

:' N~ C":"N~ o i \\ / i I

0 '" Y

thiophene

O

N-3

~.T" ~ c ,

.

CH3

'hi . . . .

H ~U"

~..~ / C 2~ 0-2 / C 2 "

thiazole s ~ward

lexitropsin

. /N~r,u_CH2 H 7"2 "C--NH2

"

.c~.. " A ' T

c+o

2P

NH v 2

~~C;

I

1 I

HiN-cz t

/\

N \ G

~

G I/C2.NH 2 ~

/

~ ,; N--~

imidazole

~=o

H.... N/"........ ~N

-..

I

H"..... ~..~o "N ........... N.,.,~

~

triazole "

pyrldlne

Figure 2. (Top)Representation ofthe binding to DNA of netropsin. (Bottom) Proposed representation of a model lexitropsin molecule with guanine residues in DNA. Heavy arrows are hydrogen bonds, from donor to acceptor. Dashed lines mark close van der Waals nonbonded contacts between DNA and drug.

outwards from the minor groove binds best to alternating purine-pyrimidine sequences such as 5'-TATGAC and 5'-TGCATGC. ll3 However, this compound remains capable of interacting with AT sequences. The same types of results were obtained with the furan-containing lexitropsin 21. ]14 To make a long story short, the lexitropsin approach based on 1:1 complex has led to minor groove binders with increased tolerance for G.C base pairs in the binding site, but did not lead to the design of a purely GC-specific molecule. 115 According to computational studies, the observation that most lexitropsins can accommodate both AT- and GC-containing sequences comes in part from the fact that the electrostatic potential in the minor


HzNLNH

H2N.~NH

HN ..0

N~

..Y

H3C'~N"CH3 L~

HN -O

N'CH3

(16)

113

HN .~

NZ N'CH3

N

.)-=/

.)=/

(17) imidazole-lexitropsins

"CH3

~"' (18)

CH~ H2N

H2N ~N

~CH~

HN

H

~

HN

~CH3

o

,So

,.

(19)

/,o

c H2N

thiazole-lexitropsins

(20)

(21) furan-lexitropsin

(22) pyridine-lexitropsin

groove of AT-rich regions is very negative. 116The electrostatic interactions between AT sites and the positively charged end groups in the lexitropsins presumably provide the initial attraction. The binding of mono- and dicationic minor-groove binders to AT-rich regions has a significant electrostatic component. 71'117'118However, that neutral lexitropsins show the same interaction with AT sequences as the monocationic distamycin tripeptide, argues against the role of electrostatics in sequence selectivity. 119 An alternative approach to design minor-groove binders capable of binding G.C pairs is to connect a netropsin-like molecule with GC-recognizing elements. For example, substitution of the terminal formamido-methylpyrrole group of distamycin for a pyridine group (compound 22) permits the interaction with a G.C base pair. However, as for the imidazole-lexitropsin, the binding to pure AT sequences is not abolished. ~2~ The bithiazole moiety of the antitumor drug bleomycin has offered opportunities for binding to GC sites. Indeed, we showed that a netropsinbithiazole hybrid 23 ligand binds specific sequences in DNA. The netropsin moiety

114

CHRISTIAN BAILLY CHa I _..1~

O

/O

H

c~O

14

NH2

H

~C2~

C-

/ ~,~

o-2 / ~ 2

T ~

o-2

~C2

N-3

A N o.z

/N

/

, H

HN

bithiazole-netropsin (23)

Figure 3. Proposed binding model of the netropsin-bithiazole 23 to the sequence 3'-CGTAT. Heavy arrows are hydrogen bonds, from donor to acceptor. Dashed lines mark close van der Waals nonbonded contacts between DNA and drug.

drives the conjugates to AT sites and allows the appended bithiazole unit to contact a pyrimidine-G-pyrimidine motif. Footprinting studies supported the model shown in Figure 3 for the interaction of the netropsin-bithiazole conjugate to the sequence 5'-TATGC. 121 This model supported previous modeling studies showing that the bithiazole-carboxamide moiety of metallobleomycin interacts with the guanosine residue of the 5'-TGC motif.122 The lexitropsin approach seemed in a somewhat wrecked state until it was reported that distamycin can form 2:1 complexes with short oligonucleotides. 123-127 The minor groove of the l l-mer d(CGCAAATTGCC) and the 12-mer d(CGCAAATTFGGC) can expand slightly to simultaneously accommodate two distamycin molecules associated side-by-side in an antiparallel head-to-head orientation (Figure 4). Subsequently, the 2:1 drugs:DNA motif was evidenced with different distamycin analogues including a carbamoyl tetrapyrrole derivative. 128 This landmark discovery immediately inspired the design of homo- and heterodimeric ligands. Imidazole- and pyridine-containing lexitropsins (compounds 24 and 25) were combined side-by-side so as to permit direct interaction with G-C pairs via hydrogen bonding with the 2-amino group of guanines (Figure 4). 129-132 The sequence selectivity, affinity, and geometry of the drug-DNA 2:1 complex can be optimized by adequately choosing pairs of ligand molecules with complementary recognition properties and by covalently linking the two DNA-reading elements (compounds 26, 27). 133-138 In recent years, important discoveries in this emerging field have been reported. A cross-linked dimeric lexitropsin possessing two distamycin-like residues connected by a heptakis(methylene) linkage exhibits a binding affinity for AT sequences approximately 1000 times higher than that of the monomer. 139'14~Hairpin pyrrole-imidazole polyamides, such as 28 and 29 which

Recognition and Modification of Double-Helical DNA CH=

o

k~

o

,.NH~

"

~

H,C / C - - N

~j~ N-"'~

9

F--~(* /

~,,,

I

..~

. ..............A . T /

I

..........

o I

"c'N--H ...........

L N~

I

/ /

o% ~ A ~ . T

, I ~

L-t

~CH=

~

H~i

~. /---'~ / \N--,~, c.=

"y

3 ' /

.... C . G .......-c.,

~ /

H

'~'N/---'

rlk

_

/""N"'C ~ 2)

"

i:

(24)

c,.

CHa

o

,~:-c,,

C',,N...,.J

1 / o ~'~ '. ...............A'T' ..c, , c - . , /, ~-~ " ............... / / A o

y

~C"

...... C . G ....~ j

._.

"'" ...............

. 2 G o C ............

. 5'~ T ~.A~ ............ % /

"*

H / H,C-.?N/----' +~

0.3

.~. N,.

L,',

\

"-% j~

11

~I~N / 0 ~H

(25)

=

/

0

' c ~'

~

oIL~-~- N.,..,..

Hac

Y

,.--o

/ .o.. --ff T . A ......... - , ~o

.L

3 ' /

/

9

~

N.......

........ / /

/

/

....... -:~.(2... ./'"-.-, v ".....H

,, I A ; ' T - - A 3' / H fN CH~

C

o"

4/

".......,,,, ,--c. a Y

.........."-,.'C-'o

G'C

. ........................ / C ' G/, , . ,

.=c. ,,

,

'- . "-. . . .

"

/~-~ H=IC~,N,~ ,~-

A .T/

oC--N

. ........../ . ............. / / _ % ........ O.. * G .

H"

hO

~'~c.. ,.-4"-C'o

distamycin

.

1

.........."-,

oH3

c,:

/

I1~

........ / / ..............Te A/.. ...... H--N /

[

CH~

/

. ........ A - T /

.~G.C ,,~T.A-~---

o

O

../-'~ 1 /

...............

MaC.N,/~.N."

.........

2,,,

__.,-c.J

H:IC% ? - - N N H

H

.............- A * T .... r / / " ".~lSC.o

/ . 5 , 1 # . lA l o / "w--...........

(•HI§

o

% P~

/

T.A,......,.o

H .........

115

".,

/

N--C~ " ~

.....~ o ~ /

C"=l

.._~"-R,/-%, \~

"N"..,

0

(30)

Figure 4. Homodimeric binding models for the complexes between a target DNA sequence and distamycin and the lexitropsin molecules 24, 25, and 30. Hydrogen bonds (dashed lines) can be formed (1) between the amide NH groups of the minor groove binders and the N-3 of purines and 0-2 of pyrimidines, and (2) between the nitrogen atoms of the imidazole or pyridine heterocycles of the drugs and the 2-amino group of guanine residues. bind to 5'-(A,T)GG(A,T) 2141,142 and 5'-AAAAAGACAAAAA, ]38 respectively, offer promising possibilities for high-affinity specific recognition of a broad sequence repertoire in the minor-groove of DNA. 141']42 An imidazole containing homodimer 30 binds specifically to GCGC sequences (Figure 4). This was the first instance where a minor groove binding lexitropsin was directed uniquely to GC sites. 143 No doubt, with the precise engineering of molecules forming 2:1 peptideDNA motifs, a decisive step has been realized towards the targeting of any

116

CHRISTIAN BAILLY

H3C

"N~

H3C

O

~/

HNf~

H3C'-NAON~'H ~-NN/~-NH O=~~lt~/

CH3

HN.~'

0//''NH

~I.

CH3 (26)

H3C /I"- ~ N , 0

CH3

imidazole-netropsindimers H3C,,N~.~N_ ~ O CH3 H3C ~'~~

H3C ~ O ~ N ~'~N

O N,"

HN ~=~

,,CH3 ~H3

NH (27)

pyridine-netropsin dimers

n- 1-4 designated DNA sequence with small molecules. By combining the end-to-end with the side-by-side dimeric motifs one will hopefully succeed in targeting 15-17 base pairs of unique sequence in a biological system. Moreover, the repertoire of sequences targeted by artificial ligands can be further extended by combining an oligonucleotide with a netropsin-like ~44or a hairpin polyamide derivative ~45 (or a Hoechst 33258 (4) analogue146'147). For example, the affinity of the conjugate 31 for the 18 base pair target site 5'-TGACATTAAAAAAGAAA-3' is 150-fold higher than that of the unlinked subunits. 148 A related ligand consisting of a pyrimidine oligonucleotide 11 bases in length covalently tethered to an imidazole-containing polyamide cooperatively binds as a dimer to 27 noncontiguous base pairs of double-helical DNA via the formation of a 2:1 ligand:DNA complex. 145 In both cases, the recognition process involves simultaneous recognition of the minor and major grooves of the double helix, thus mimicking the function of certain sequencespecific DNA binding proteins. Despite all these elegant structural studies of lexitropsin-DNA complexes, the biological activity of minor-groove binders has not been much improved. So far the lexitropsin approach has not led to clinically useful drugs although certain monomeric and dimeric lexitropsins exhibit interesting antiviral or anticancer


N--=.

H3C~ Hs~~ NH

0

117

0

/~

(28) NH'I~O N~ AcPyPyPy-7-1mlmPy-~-Dp~~/N'CH3 H3C,, H3C" NH ~N~c:) O~NH'~NH "~ O / H3 ~NH NH H3c'O 3 H3~-==~ - _NH ",--N NH (29) NH,,~O ~N,,CH3 i Nr~ CH3 imPypy.13.pypypy.D p ~N"CH3 H3C" NH H3C~ Ot2"~NH~,~ N~N~c s) NH

o~N~

N-...=.

O

H3c_N/~NH

02" NH

(31)

O

/

NH,~O

~N.CH3

NI~ ImPyPy-y-ImlPyPy-I~--H3C-'"'..~" NH O'~NH NH.~N~O N

o

r i I I I ' I I MeCMeCTI'T.3'

'ICH3

118

CHRISTIANBAILLu

activities in vitro and sometimes in vivo. The new generation of GC-specific lexitropsins may provide more efficient drugs since GC-rich sequences are frequently found in genes associated with proliferation including a number of oncogenes such as c-Ha-ras. 149 The human genome contains regions with more than 80% GC content which account for --1.3% of the genome. 15~By analogy with the GC-selective antitumor drug mithramycin discussed above, GC-specific lexitropsins may exhibit useful biological properties via recognition of key sequences in DNA.

V. NETROPSIN AND DISTAMYCIN DERIVATIVES EQUIPPED WITH A BINDING/MODIFYING ELEMENT A considerable number of antitumor agents, including some of clinical value, induce DNA damage either directly by alkylation (electrophilic modification of bases, e.g. nitrogen mustard, cis-platinum) or cleavage of DNA (oxidative degradation of the sugar backbone or hydrolysis of the phosphodiester bond, e.g. bleomycin, calicheamicin) or indirectly via inhibition of topoisomerase activities (e.g. daunomycin, amsacrine). The DNA lesions induced by these different categories of drugs are presumably responsible for their cellular toxicity. Irreversible damage of the genetic material in cells has been considered as the basis of the antitumor effect but it is more likely responsible, to some extent, for the toxic side effects towards normal cells (genotoxicity). Consequently, there is currently intense interest in the development of sequence-directed DNA damaging agents with a view to improving the therapeutic value of the drugs by virtue of a highly selective gene-specific recognition. The following sections describe some examples of netropsin/distamycin conjugates rationally designed to bind tightly to DNA and/or to produce irreparable damage at precisely defined genomic locations.

A. Conjugationwith Polyamines:The Microgonotropen Approach Minor-groove binding ligands such as distamycin can efficiently compete with the binding of protein in the major groove of DNA. 151-154 Inhibition of protein binding to the AT sequences is believed to contribute to the biological activity of minor-groove binders. 155'156However, under certain circumstances DNA can accommodate both a protein and a lexitropsin in the major and minor grooves, respectively. 157The capacity of a minor-groove binding agent to compete with the binding of a protein may be enhanced in different ways: (1) by increasing the affinity and sequence selectivity of the drug for DNA, (2) by equipping the drug with a major-groove binding element so as to interfere directly with the protein, and/or (3) by using drugs capable of inducing adequate structural changes in DNA. The bending and flexibility of DNA are actively exploited by sequence-specific DNA binding proteins. 21'23 Bruice and coworkers have conceived an elegant approach to increase the affinity of distamycin analogues for DNA while maintain-

Recognition and Modification of Double-Helical DNA CH3 H3C"N'~ HN

fH3 H3C'N~I, 0

HN

~3.a

119

HN

"~N~cH 3

n=3-5

(32) dien-microgonotropen

~CH3

l-IN

~3Ha

0

0

n=3-5

(33) tren-microgonotropen

\ NHz

ing the AT selectivity, and allowing potential recognition within the major groove of DNA. They designed a series of molecules called microgonotropens in which the methyl substituent on the central pyrrole ring of distamycin is replaced with a branched polyamine. 158-164The polyamine substituent on the dien- (32) and trenmicrogonotropens (33) projects outwards from the minor groove and acts as a hook to increase the affinity for DNA via interaction with the phosphodiester backbone or with the major groove of DNA. The tren [-NHCH2CH2N(CH2CH2NH2)] microgonotropen (33, n = 4) is much more effective than distamycin in promoting DNA bending and straightening. 165 The microgonotropen strategy represents a new avenue to the design of minor-groove-reading ligands having a very high affinity for DNA. In addition, the metal-chelating properties 166 of the dien and tren substituents may be exploited for the conception of catalysts for sequence-selective hydrolysis of DNA, i.e. for the design of artificial nucleases. 167'168

B. Conjugation with a DNA Alkylating Group The large majority of alkylating drugs react with DNA via an electrophilic attack of purine residues, in particular with guanines. Classical alkylating agents such as chlorambucil and cis-platinum react within the major groove of DNA at the N-7 position of guanines. 169Nucleophilic attack from N-7 of guanine is also implicated for the DNA alkylation by the pluramycin antibiotics (altromycin B, hedamycin). 17~ A few alkylating drugs of natural origin bind in the minor groove of DNA. Such is the case for (1) mitomycin C 171 and anthramycin 172'173 which both form adducts with the exocyclic amino group of guanine (mitomycin can also react with N-7 of guanine under acidic conditions but the biological significance of the N-7 adduct remains uncertain), and (2) the adenine-specific alkylators duocarmycin A and CC- 1065 discussed above. The accessibility of the major groove and the high intrinsic nucleophilicity of the N-7 heteroatom position contribute to the G-alkylation by nitrogen mustards.

120

CHRISTIAN BAILLY CH~

H2N~ H

"3c'N't.1 HN 0 ~N

~N...CHs

HN ..0

CI HN

"-CH3

HN

0

"CHa

HN

0 N...CH3

HN

~N~"--~O

ar j

H3

.N

(34)

(3s)

(~)

N-chloroacetyl-netropsin

N-bromoacetyl-distamycin

bis(chloroethyl)-netropsin

~H3

CH3

H3c"N'~,L

"l HN

"l

O

HN

N, , , ~ ~ ~N"

CH3

"-'~,/" " CH3

CI

O

N...CH3

~N...CH3 el HN ~==0

Ci CI~

H3c"N",,L

HN ~

H ~--N

..

/--~ HN"~N'cH3

o

0

cH3

(37)

(3a)

bis(chloroethyl)-distamycins

However, the sequence-selective alkylation of N-7 guanine by nitrogen mustards and nitrosoureas can be modified by linking the reactive group to a DNA reading element. The first generation of alkylating lexitropsins consisted of netropsin and distamycin analogues equipped with N-chloroacetyl (34) and N-bromoacetyl (35) substituents or with bis(chloroethyl)amino substituents (e.g. 36--38). 174'175 N-Bromoacetyl--distamycin 35 reacts with a single adenine in the sequence 5'GTTTA-5'-TA*AAC within a 167 base pair restriction fragment. 176'177The second generation of alkylating lexitropsins contained benzoyl- (39,40) or benzyl-mustards (41,42). 178-181 Following the discovery of the aforementioned antiparallel side-by-side motif, bismustard cross-linked lexitropsins such as 43 were synthesized. 182 The efficiency of DNA alkylation by chlorambucil and L-phenylalanine mustard can be considerably enhanced by linkage with DNA carriers such as spermidine, anilino-quinoline, and acridine derivatives. 183-187 Certain acridinelinked mustards exhibit a preference for adenine alkylation compared to guanine

121


H2N'~H

H3C,,NJCH3 HN ~._O

HN 0

c,

IN "OH3

~

c'-k_N~

HN H

HN

N j.

N

H~

H3 L)

I:H3

OH3

"O

(39)

(40)

(N-benzoyl-mustard)-distamycin FCE 24517

(N-benzoyl-mustard)-Iexitropsin

H3C"N'CH3

H3C"N"CH3

HN .,_O

HN ..O

N~N ' "OH3Cl yL ' ~N ~l

C~-~N-~'~ H N~O Cl/..~/ ~ " ~ O N.~N~o,."~ N-CH3

N 'cH3Y ~N ' N--~O

~ H "--~ N,~N.~N"~V N"CH3

I::H,

~:H,

(41)

(42)

chlorambucil-lexitropsin

(benzyl-mustard)-Iexitropsin

H3C,

H3/'--~~HN .0

"~N"CH3 /N~ O

S O~~

CI

'/ ~/~Cl (44)

40

~

H O

~/.~ (43) bismustard-lexitropsin

CH3

H~C'~~

CH3

J

HN~o

N0

})CH3

122

CHRISTIAN BAILLY

alkylation. A much higher level of sequence-selective alkylation/crosslinking has been achieved by attaching nitrogen mustards or nitrosoureas to oligonucleotides ]88'189and lexitropsin moieties. Distamycin analogues equipped with DNA alkylating functionalities such as compounds 39--42 show a remarkable sequence selectivity with, in some cases, an almost exclusive alkylation of adenines in the minor groove with no detectable guanine N-7 reaction. The lead compound in the series is the bis(2-chloroethyl)aminobenzoyl derivative of distamycin FCE24517 (39), also known as tallimustine, which demonstrates significant anticancer activity in animal models and is currently undergoing clinical trials. 190-196It remains to be demonstrated whether or not the potent antitumor activity of FCE24517 is attributable to its DNA binding selectivity. Perhaps, alkylation at specific A-containing sites (e.g. 5'-TI'I'IGA*) produce adducts which are poorly repaired and evolve into strand breaks lethal for the cell. Sequence-specific inhibition of the binding to DNA of transcription factors has also been postulated. 156']94Fifty years after they were introduced into medical practice in the treatment of neoplastic diseases, 197 nitrogen mustards are still among the most clinically useful anticancer drugs. With the rational design of tumor-active drugs like FCE24517, there is good reason to believe that for a considerable time nitrogen mustards will remain of major clinical importance. Other series of distamycin conjugates equipped with alkylating functionalities have been synthesized. Lown and collaborators ha.ve synthesized a series of lexitropsin-cyclopropylpyrroloindole (CPI) hybrids. 198 The CPI-N-methylpyrrolecarboxamide derivative (44) exhibits an exceptional cytotoxic potency against KB tumor cells in vitro (IC50 = 0.76 fg/L) and forms stable covalent adducts in the minor groove of DNA. ]99 Church et al. 2~176 have synthesized chloroethylnitrosoureas-lexitropsin conjugates such as 45 which alkylates adenine residues in the minor groove. Zhang et al. ]19 have prepared a series of noncationic Nmethylpyrrole dipeptides incorporating sulfonate ester terminal groups (46,47). Here also, efficient alkylation at adenine N-3 in the minor groove was observed. An elegant way to provoke DNA-DNA cross-links was recently reported using a CH3 "1 ~~=jN

c,

o

HN

0 .,...CH3

HN

__)=o

CI-ENU-netropsin

HN ---CH3

.-CH3

~''~N'~ /--~N O. HN 0~N' H N""~.~-"~0 H/N"'"~/"l'CH3 SO~"'~O

(45)

0

HN

H3

H3C"SO~')~ 0 HN

(46)

(47) methylsulfonate-netropsins

H3


123

...CH3

OH

~'-N"

~Ha

0

(48) [bis-(hydroxymethyl)-pyrrole]-distamycin

H2N

(49) anthramycin-netropsin

distamycin analogue coupled to a 2,3-bis(hydroxymethyl)pyrrole function which in part mimics the functionality present in reductively activated mitomycins or oxidatively activated pyrrolizidine alkaloids. Compound 48 efficiently produces interstrand cross-links by bridging the 2-amino groups of two-paired CpG steps. 2~176 More recently, Walker et al. 2~ have designed netropsin-anthramycin conjugates such as the chimera 49 which is expected to recognize the sequence RGAAAA from the HIV- 1 polyypurine tract.

C. Conjugation with a Photosensitive Group Sequence-selective minor-groove binding as a route to targeting photosensitive compounds to DNA has been considered in recent years. Among photoactive compounds of biological interest, psoralens (furocoumarins) are probably the most important group since some derivatives such as 8-methoxypsoralen are used in phototherapy for the treatment of various human skin diseases, chronic leukemia, and some infections connected with AIDS. The biological effect presumably derives from the action at the nucleic acids level. Psoralens can intercalate into DNA. When exposed to UV light, intercalated molecules react covalently with DNA to form cyclobutane linkage to pyrimidine bases, predominantly at 5'-TpA steps. 2~ Psoralen-derivatized oligonucleotides have been used to induce lightdependent sequence-specific reaction with RNA (antisense) and DNA (antigenes). 2~176 In the same vein, minor-groove binding oligopeptides were coupled with psoralens (50-53) and coumarins (54,55) to direct photoreaction at specific sites. 2~176 Pyrene-lexitropsin conjugates such as 56 were also synthesized. 2~ The cytotoxic activity towards human leukemia cells of this sequence-selective photosensitizer is significantly enhanced upon irradiation. The photo-induced DNA lesions apparently result from the production of singlet oxygen. 2~ Using a similar approach, Herfeld et al. 21~ have synthesized netropsin-isoalloxazine conjugates such as compounds 57 and 58. Upon photoactivation in the presence of molecular

124

CHRISTIANBAILLY HzN'~NH HN 0 ~ N 'CHa

(50) O~

H3C.,.N,CH3

~

HN

0

~m~N~CH3

(51)

HN 0 H3

HN .0

0

N~N"CH3

H2N'~ NH

(52)

HN 0 ~ N ""CH3 C~3 /OvCH3 HNI 0 0--/~'~ _H ,,~0 H,G.N,OH ' ~ . HN,~N.oH3 L~ CH:l o~"~'~0 HN : (53)

'0 ~

.N N-o

o

~

HN~N"cH3

0

.OH, (54)

HNY

HN 0

H3C,N.OH, I~ H:'~:....CH' (55) y N~ 0

0

oxygen, the flavin chromophore oxidizes and generates oxy radicals capable of causing DNA breaks. The linkage of netropsin to a flavin chromophore lead to AT-selective strand cleavage reactions. 211'212 Quinone-netropsin hybrids (e.g. 59,6t)) have been also designed. These conjugates are capable of inducing UV-mediated strand breaks. 213 Cationic porphyrins can also act as DNA photosensitizers. 214'215 Netropsinporphyrin conjugates (e.g. 61) have been synthesized. Molecular modeling predicts that the porphyrin moiety intercalates into DNA. 216'217UV-sensitive p-nitrobenzoyl groups attached to netropsin-acridine hybrids also act as photocleavers (Section V.F). Matsumoto et al. 218 have synthesized a series of oligo(N-methylpyrrolecarboxamide) derivatives linked to halogenated heteroaromatic groups. Efficient photoinduced DNA cleavage was observed with bis-pyrrole derivatives such as the bromofuran-netropsin conjugate 62. DNA cleavage is not mediated via active oxygen species (OH.) but may be due to the reaction of an aryl radical produced by photohomolysis of the carbon-halogen bond. 218A similar mechanism with drug radical production has been proposed for X-ray sensitive nitroaromatic compounds.


125

HzN,,,~..INH HN 0 HN N~'I

(56)

,

H3a ~

CH3

N~ C H 3

N?

N.. ~ N~)

HN

~

~/

O

HN~'~N_../~

H-y~

/CH3

HN ._O ~N"CH3

(60)

n3~"~~

o

~

ItO

~N.C:)

HN

N ~

~ "~ ' K ~ N ~ .

porphyrin-netropsin

/

O o

o

(c.,)&o

o

H N ' ~ sCH3

(61)

H3

H2N',~NH

(69)

HN..~N.cH3

HNNH=~\

" " Y "~"."~(CH,>,"-(,o quinone-netropsin

HN/--"

H3C~/'N.,~,N. ~ " ' ~ ~ k O OH3 O

HN L

(,,,

O

,

HNTO

CH3 I+~ N.

HN -0

~N....CH3

HN'

~~,

O NH

~'N~'CH3 (r T"- T"- ~ ~ -+N/~ '~N HN.~

o~

HN O H3C-k ~ r

'

"

%/,l

OH3

Metronidazole 63 and misonidazole 64 are typical examples of radiosensitizing agents used in the treatment of anaerobic infections and are under continuing investigation regarding their use in cancer therapy, acting as markers for hypoxic regions in tumors. 219 These sensitizers can react with DNA presumably via an electron-seeking radical (the metronidazole nitro radical anion RNO 2. has been identified in dimethylforfamide) but they exhibit very little binding to DNA. 22~ Targeting of a 2-nitroimidazole moiety to DNA via an acridine intercalator or a minor-groove binder has been investigated as a means of increasing sensitizer concentrations at defined sites on DNA. 222'223The affinity of 2-nitroimidazolenetropsin conjugate 65 for DNA is about 200-fold superior to that of misonidazole 64 but in spite of the improved interaction with DNA and improved cellular uptake

126

CHRISTIAN BAILLY

HzN,~NH

/'~NOz NyN~/~,OH

HN ~

HN 0~0 Br'

CH3

"CH3 9

(63) metronidazole

) H.N~ N ~ c H3

f'~ OH N,~,,,,N.v~OCH3 NO2

(62) bromofuran-netropsin

(64) misonidazole

H,N .

H3c.CHaN.t

HNTO ~ N ' cH3 HN NO2

'~0 (es)

misonidazole-netropsin

HN 0 ~N~CH3 HN CI (6s)

p-chlorobenzylsulfonamido-netropsin

capacity, the radiosenzitization efficiency of the netropsin-nitroarene conjugates remains relatively poor and not better than that of misonidazole. 223 Photoinduced DNA cleavage has been reported with oligo-N-methylpyrrole-carboxamide derivatives substituted with a benzyl-sulfonamido group such as the p-chlorobenzenesulfonamide-netropsin hybrid 66. The efficiency of single-strand cleavage under UV-A irradiation depends on the length of the pyrrolecarboxamide chain. Tetrapyrrole-sulfonamide conjugates are more efficient DNA cleavers than the corresponding analogues containing three or two pyrrole units. Conjugates with only one pyrrole are practically inactive. 224 The same conclusion was drawn for simple oligopeptides such as 67 and 68 which do not possess special side chains sensitive to UV light but which, nevertheless, can induce DNA cleavage under UV-A irradiation. 225 Conversely, for nitrated oligopyrrolecarboxamide derivatives such as 69 and 70, the DNA photocleavage efficiency is higher for the monopyrrole compounds than for the bis- and tris-peptides. 226 Another type of netropsin conjugate capable of inducing DNA cleavage upon X-ray ionization has been designed by Grokhovski and Zubarev. 227 Footprinting


HO'~,,NH

H2N~~,,N~NH z

HN~~ HN HN H~O

""CH3

HN H"~O

H3 (68)

CH3 H3c'N~,""I ..CH3

HN H~

(67)

CH3 H3c'N~,"~

HN~~ "CH3

127

HN

HN

~~N~I 3

~~N~c )

N 0 NO, (69)

.CH3

02N (70)

studies showed that the netropsin-platinum-netropsin conjugate 71 binds selectively to AT-rich sequences in DNA. X-ray radiation of drug-DNA complexes yield discrete cutting sites near the center of the platinum-bis(netropsin) binding sites. In this case the cleavage would result from the rupture of the deoxyribose residues upon attack by Auger electrons generated by the irradiated platinum atom. Such conjugate compounds capable of triggering sequence-specific DNA degradation might be of interest for X-ray therapy of tumors.

D. Conjugation with a MetaI-Complexing Group Ligand binding sites on DNA can be identified by a variety of techniques based on DNA sequencing methodologies. The most widely used method is the footprinting technique. 228The ligandmbe it a protein or a drugmis equilibrated with a singly end-labeled DNA restriction fragment and the complex is subjected to partial cleavage by an enzymic or chemical nuclease (Figure 5). DNAase I is the most frequently used endonuclease but transition metal complexes such as Fe.EDTA and Cu.phenanthroline can help to define more accurately the exact location and size of the binding sites. 229'23~An alternative strategy for mapping ligand binding sites on DNA consists of equipping the ligand with its own DNA cleaving functionality, e.g. to attach an Fe.EDTA complex to the test drug. This method, termed DNA affinity cleaving (Figure 5), proved very successful in analyzing binding sites for distamycin via the use of EDTA-distamycin conjugates such as compound 72. 23]'233 The technique has been extensively exploited by Dervan's group to study oligo(N-methylpyrrole-carboxamide) derivatives containing up to 12 pyrrole units as well as a variety of monomer and dimer lexitropsins. 234-239 Upon chemical activation with a reducing agent (e.g. dithiothreitol) the iron-EDTA portion generates hydroxyl radicals which react with deoxyriboses so as to provoke DNA strand cleavage. In contrast to Fe.EDTA footprinting, the active oxygen species are

128

CHRISTIAN BAILLY footprinting

nuclease

nuclease

m

m

DNA affinity cleaving

m

activation

m

Figure 5. Illustration of two methods used to locate drug binding sites on DNA. In the footprinting experiments, a DNA restriction fragment radioactively labeled on one strand is subjected to cleavage by a chemical or enzymic nuclease. In presence of a sequence-specific ligand, DNA cleavage is inhibited at the binding sites identified by the footprints on the autoradiogram. In the DNA affinity cleaving experiments, a DNA restriction fragment radioactively labeled on one strand is subjected to cleavage by the ligand itself which is equipped with a DNA cleaving functionality, e.g. a transition metal complex such as the EDTA-Fe(II) system activated by a reducing agent. Binding sites are identified by the cleavage products visible on the autoradiogram. In both cases, the resulting DNA cleavage products are analyzed using a standard DNA sequencing gel. The sequence of the DNA (and binding sites) can be identified by coelectrophoresis of oligonucleotide markers produced by cleavage of the DNA using standard Maxam-Gilbert sequencing methods.

Recognition and Modification of Double-Helical DNA CHa

o

O.;N.

H'C'N~NH O

(71)

HN

O

Q ~.. r ....__~Iv,. ~

NH ..... NH--..,/

.: H

~~'cu/~ ~

~ ~,~.'..~._/,,,

(72)

HN

~,,

H

0

~_ N'cH~ HN

HN~NHa O

~JNN-CH3

0

; N.H

H

L

, , ,~~ /

"~.--N

HN

Ha

CHa

distamycin-EDTA

L .~.~o

~,, ~'... o

....

~N f

cH3

HN

HN G/y

o

0

netropsin-platinum-netropsin

O-",,

'o...~..~ Y"'!'N

%

o

0 /-~...

O

.or7

~N.-CH,

~

Gly

129

~,,

c.,

(73)

Gly-Gly-His-(Gly)3-netropsin produced directly in the vicinity of the ligand binding sites on DNA and so are less susceptible to diffuse. Metal-complexing peptides can serve as a source of oxygen radicals. The growth-modulating tripeptide glycyl-histidyl-lysyl (GHK) and the related tripeptide glycyl-glycyl-L-histidine (GGH) both form complexes with copper, and upon activation can engender oxygen active species. 24~ Linkage of the GHK peptide to intercalating 245-247 and minor groove binding drugs has been considered as a means of inducing sequence-selective cleavage. The peptide moiety not only cleaves DNA but also contributes positively to the DNA binding reaction. 248 Two molecules of the GGH-netropsin conjugate 73 (one copper-free and one coppercomplexed) bind cooperatively to the sequence 5'-TITI'NCAA*AA. The two netropsin moieties occupy the minor groove of the two (A.T)4 tracts, whereas the GGH group is superimposed over the central G.C pair. In the proposed model, the 2-amino group of the central guanine residue hydrogen bonds with the histidine carbonyl group. 249 It is worth mentioning that GGH can also complex Ni(II) to produce activated oxygen species. 25~The oxidizing potential of Ni-chelated GGH has been exploited to study sequence-specific binding of the netropsin-GGH conjugates described above, as well as for proteins such as Hin recombinase 251 and peptide nucleic acids. 252

130

CHRISTIAN

BAILLY

The best characterized natural model for sequence-specific DNA cleavage is bleomycin. The bleomycin antibiotics are among the most useful antitumor drugs. Their therapeutic effects are believed to arise from their ability to cause DNA degradation in the presence of redox-active ions and a source of oxygen. The bleomycin.Fe(II) complex combines with 0 2 to produce a reactive oxygenated metallobleomycin species which is capable of abstracting a hydrogen atom from the deoxyriboses in DNA. Bleomycin generates mainly single-strand breaks at pyrimidine residues 3' to a guanine residue (i.e. at 5'-GpC and 5'-GpT sequences). 253-257 The long-established clinical utility of bleomycin (one of the few anticancer agents which has little bone marrow toxicity) sparked tremendous interest. A considerable number of bleomycin analogues and related structures have been designed including some equipped with DNA reading elements based on netropsin and distamycin. The structure of the antibiotic has been simplified to yield analogues such as PYML, PMAH, 258 and AMPHIS 259-261 which mimic efficiently

~NH [[~"7/>

.,,~iF~!? . ......

. . I " ~ "-~NH 0 d" (74)

CH3 HN ~

Amphis-netropsin

HN&.. o ~

~ O H3C.....j HN

-C o

A ~..N ~ "NHIF __NH'~'"~NJ ~ :',f/' : Fe : "

O

(7')

.,,.,.~1~"J~o O'))''N" V

g'~tr

N/CH3 HN

~ ~I'cH. HN&0

~O v

-HN.~o/ 'OH,

,.o

,,-~

HN"%~

H3C',N_.~ 0 # HsC,- ~

CONH2

(7'6)

CH3

OHN~N'CHa

,r

,,, o

0

.

,CH,

..

.o

.....

CH,""' (~H, H

~,,u.n=

PYML(6)-(AaM)-distamycin

(77)

,,,,NH'r~..N--~' J i-Fe i ~'-'O-IBu ~N'/-----~I~'(

..co o

O

~O a /~. /N~N'cH,

CH= PYML(6)-(APA)-distamycin


131

the metal-chelating/oxygen activation domain of bleomycin. Attachment of PYML 262'263 and AMPHIS 264-267 to lexitropsin carders led to the synthesis of bleomycin-like conjugates such as compounds 74-77 endowed with interesting sequence-specific DNA recognition properties. For bleomycin, the whole molecule spanning from the pyrimidine region to the bithiazole terminus appears to be responsible for specific recognition of DNA. 268 In contrast, the lexitropsin moiety is only responsible for specific base recognition of man-designed bleomycin conjugates. The fact that the synthetic PYML-distamycin hybrid 76 is more toxic than bleomycin itself towards L 1210 leukemia cells in vitro encourages the design of other related bleomycin-like molecules tailored with DNA reading elements. Bleomycin models have also been used to design sequence-specific DNA cleaving proteins 269 and oligonucleotides. 27~

E. Conjugation with an Enediyne Structure The discoveries in the late 1980s of tumor-active enediyne antibiotics have rapidly elicited extensive research activities in chemistry. A impressive number of synthetic analogues of dynemicin (78), neocarzinostatin (79), calichaemicin, and esperamicin have been described. 271-273 Upon activation, enediynes trigger a Bergman reaction which leads to highly reactive benzenoid diradicals and cause severe DNA damage. 256 Enediyne antibiotics are among the most cytotoxic compounds known to date, and their activity is likely attributable to their effect on DNA (at least in part). The excessive reactivity of the enediyne moiety prompted the development of analogues containing DNA delivery systems. To this end, netropsin was coupled with the neocarzinostatin chromophore (80) 274 and distamycin was attached to a dynemicin model (81). 271 More recently, the design of a simple cyclic enediyne attached to netropsin was reported. 275 The linkage of the two functionalities via acetate (82) and crotonate (83) tethers results in a hybrid series in which the cleavage efficiency is strongly enhanced (up to a 1000-fold compared to the enediyne alone). These encouraging results pave the way for the design of simpler related structures capable of triggering effective DNA cleavage via a radical mechanism. In this context, Bregant et al. 276 synthesized the netropsin analogue g4 equipped with a trimethylenemethane group (TMM) which can undergo cycloaddition to electron-deficient alkenes. Upon photolysis of the diazene moiety, the TMM-netropsin conjugate can transform to a diyl radical and cleave DNA, predominantly at AT-rich regions. 277 Propargylic sulfones are small synthetic molecules that mimic the chemical action of enediynes. They can cleave DNA in a pH-dependent fashion. Here again, linkage of lexitropsin carriers to propargylic sulfones (85-87) may permit direction of the cleavage of specific sequences in DNA. 278 Parenthetically, it is worth mentioning here the netropsin-quinocarcin conjugate 88 which can efficiently cleave DNA at AT-rich sequences via the production of a nondiffusible oxidant. 279

132

OH o

CHRISTIAN BAILLY

(79) neocarzinostatin ~--o OCH3 chromophore oN/I

~ 3

~~ ~ o

HN'_~:'~"~, ~~

y--C-y (78) OH 0 OH dynemicinA

CH3 H3c'N'~L

HO'~d

HN 0

HNTO

dynemicin-distamycin ~N.~CH'

~0

O~.N~~.~

o~S_O

~~) "o ~o~.~

HNy~ Hi

HNy~ HI

Oo

H3C"N"CH3 HN =O ~__N "CHs HN

HN~N O ...CH~ HN

HN O ~N .-CH3

(82)

"l

181)

NCH3

neocarzinostatin-netropsin~~N. c

CH3 Hac'N'~

Hd) CHa-

"~T~..

(80)

O

(83)

_

(84)

TMM-netropsin

F. Conjugationwith a DNA IntercalatingDrug By analogy with the lexitropsins, hybrid molecules combining a minor-groove binding element structurally related to netropsin or distamycin, and an intercalating chromophore are called combilexins. 28~

Combilexins and DNA Recognition In general, intercalating drugs exhibit a high affinity for double-stranded DNA but display little sequence selectivity. Simple intercalating agents such as ethidium and proflavine interact similarly with both AT- and GC-rich regions in DNA and


133

HaCO 0 HO~

(S5)

~N'~CH3_ HN

~,0

o

H N ~

N H3

HN

~Hs H3CO,.,~O /,OH ....

~

,'~N"CH~

"%0

N ~.~ H N ~ C H 3

CH3 fOH 0 /~ :sJ

H3CO~~...CHa (87)

HN~ ' / ~-o

ICH3 slightly prefer alternating purine-pyrimidine alternating sequences as compared to homopolymeric sequences. 16A large number of intercalating drugs, including some of major clinical value, possess an heterocyclic chromophore substituted with side chains of different nature that participate in DNA recognition. These chains represent a kind of hook for sequence-selective interaction within the minor groove of DNA. For example, the antitumor drug actinomycin (89) bears two sterically demanding cyclic peptides attached to a phenoxazinone chromophore. Upon bind-

2"' "NH

NH CH3

0

CH3

134

CHRISTIAN BAILLY

ing to DNA, the chromophore intercalates between base pairs leaving the two pentapeptide lactones lying in the minor groove of the double helix. 282-284The sharp selectivity of the drug for intercalating at 5'-GpC steps arises from prior recognition of a suitable groove geometry (e.g. an adequate minor groove width) followed by a reading of particular structural elements (e.g. the guanine 2-amino group) in the minor groove of the DNA helix by the antibiotic. 285 The symmetrically disposed peptides participate actively in the GC-specific recognition process through hydrogen bonding interactions between the N-3 atoms and 2-amino groups of the guanines, and the amide groups of the threonine residues. The amino acid residues of the drug play a considerable role in both the DNA binding properties and the biological activities. 286 Similar observations can be made for a variety of tumoractive intercalating drugs. Anthracyclines such as daunomycin and nogalamycin contain carbohydrate residues that serve as DNA recognition elements. 287 The peptide ring of the quinoxaline antibiotics echinomycin and triostin A are prime determinants for sequence-specific recognition by the drug. 288 The enediyne antibiotics dynemicin A (78) 289 and neocarzinostatin (79) 290 bind to DNA by intercalation of their chromophore (naphtoate for neocarzinostatin and anthraquinone for dynemicin), placing their reactive ene-diyne-containing bicyclic core moiety in the

o

0

=o~ o

CH2 H21C" ~C.,_ CH H3C ~C...CH.N ~u I 3 "N ""C~. ,,CH C'H~ CH "CH3

o=c,

?.

H3C.NcH C=O ,..d, C"O--cH'CH'NH _ C..t.,n o Hs ~., O CH,, ~C",0

CH= HaC/ '~C., 0 HsC, Q ~,, CH-c" CH .~CH ~-C' ' ~ "N" a I'l'aC" ~'CH ~ru

c=o

NH

0=~ N~ HN'CH-cH-O..--c'C~H O*C/ ~ CH- O CH , "CH3

II 1 1 "~o~o [

CH3 (89)CH3

actinomycin

~

H3C~

.~ HN

~

o

,_,, . ~ o 0

"3 - ~uN H

CH

.N.,,,.~O r / ~N'CH3 ,~='~"

....

N ~N_.~ 0 ~ " N~... ~

o~....~"l~^

~/~ N''CH3

O"-1F'~N.20

HN~',.0 "-~.._..~ N,CH.

distactin (gO)

0

"f

ell3

[bis-(EDTA-distamycin)]-phenoxazone


135

minor groove in a suitable position for selective DNA cleavage. The list can go on. Such considerations indicate that in many cases drugs usually referred to as "intercalators" in fact exhibit mixed modes of binding to DNA and therefore should be considered as intercalator-minor-groove-reading hybrid molecules or naturally occurring combilexins. Using actinomycin D as a model compound, Krivtsora et al. 291 designed a series of hybrid molecules called distactins in which the phenoxazone chromophore is substituted at positions 1 and 9 with one, two, or three N-methylpyrrole carboxamide units. Bidentate reaction with DNA involving intercalation of the chromophore and minor-groove binding of the DNA reading element was observed with distactins bearing one or two pyrrole rings but not with the analogues having three units. The model was reconsidered two years latter by Dervan 76 who elaborated a distactin molecule in which two distamycin tripeptides are connected to the phenoxazone by glycine tethers. DNA affinity cleaving studies of the bis-(EDTAdistamycin)-phenoxazone conjugate molecule 90 revealed a major cleavage site flanking the 10 base pair sequence 5'-TATAGGTI'AA, thus suggesting that intercalation of the tricyclic nucleus at the central GG step is accompanied with minor-groove binding of the distamycin moieties at the flanking (A.T)4 sites. Additional single cleavage loci were also observed. Depending on the recognized sequence, only one or both recognition elements engage in contact with DNA. These preliminary studies paved the way for the design of different categories of intercalator-minor-groove binder hybrid molecules endowed with sequencespecific recognition properties. Several reports have clearly established that a groove binder like netropsin and an intercalator (e.g. actinomycin, acridines) can bind to DNA simultaneously and in close proximity. 292-294 The affinity of distamycin for DNA can be considerably enhanced when the drug is attached to an intercalating agent. The chromophore not only provides extra strength of binding but can also notably influence the DNA recognition by the minor-groove element. Different behaviors have been observed depending on the chemical structure of the appended intercalating drug and the length and the flexibility of the linker between the two DNA binding functionalities. Bis-pyrrolecarboxamide moieties were linked to a 9-aminoacridine chromophore by alkyl linkers of variable length, z95 Optimum fit to DNA was obtained with the combilexin 91 with a butyroyl tether. This biscationic hybrid exhibits a strict AT preference as for distamycin. A structurally related acridine-distamycin ligand equipped with a photoactivatable p-nitrobenzoyl group 92 has been synthesized. 296 This conjugate can cut DNA upon activation with UV light (310 nm) with a preference for AT sequences. 297 The selectivity for AT sites was found to be significantly decreased when a truncated netropsin moiety was linked to a glycylanilinoacridine chromophore structurally related to the antileukemic drug amsacrine 93. In this case, the rigid connector between the bispyrrole unit and the acridine nucleus of the combilexin molecule 94 (NetGA) does not permit optimal intercalation of the acridine and apparently slightly restricts the netropsin moiety fitting

136

CHRISTIAN BAILLY

HN..~

netropsin-acridine

HN.~

HN

HN

HN~o CH= 9H3

H ~

HN~o CH3 0

H I~

.

o

02N" V

o~

deeply into the minor groove. Molecular modeling and experimental binding studies were consistent with a model in which the acridine ring is significantly tilted with respect to the helical axis of DNA. 298'299By contrast, the monopyrrolecarboxamide-anilinoacridine ligand MePyGA 95 (i.e. the mono-pyrrole counterpart of NetGA 94) behaves as a classical DNA intercalating drug as judged from electric linear dichroism measurements. 3~176 Moreover, this simpler analogue exhibits a marked preference for intercalating at GC sequences and the sequence selectivity correlates with its topoisomerase II inhibition capacities. The exact mechanism of binding of this pyrrole-anilinoacridine conjugate is not yet fully elucidated but we suspect that the intercalated acridine projects its anilino-glycyl-N-methylpyrrolecarboxamide substituent towards the exterior of the helix so as to place the carboxamide group in close contact with a guanine 2-amino group protruding in the minor groove. Covalent linkage of distamycin to the GC-selective ellipticine derivative 963~ afforded a monocationic hybrid molecule Distel(l+) (97) capable of bidentate binding to DNA. Complementary biochemical and spectroscopic studies showed that, in this case, the reaction with DNA was primarily driven by the charged ellipticine moiety of the hybrid. 3~ The geometry of the Distel(1 +)-DNA complexes can vary according to the target sequence. 3~ As a result, the hybrid exhibits little, if any, sequence preference. 3~ The ellipticine chromophore has markedly reinforced the affinity of the ligand for DNA, but the effect is at the expense of DNA sequence selectivity. Molecular modeling aided the design of a second-generation distamycin-ellipticine hybrid endowed with superior DNA recognition properties. Indeed, computational studies suggested that the addition of a positively charged group on the distamycin terminal group would favor binding to AT sequences. 3~ A recent experimental demonstration to this proposal clearly established that indeed substituting the terminal formamido group of Distel(l+) for an aminopropionamido group charged at neutral pH was the correct way to proceed in order to convert a nonspecific conjugate into a highly specific DNA reader. 3~ Both DNAase I and MPE.Fe(II) footprinting studies indicated that the combilexin


~NH-~-CHz I I DNA-binding domain

13 7

a m s a c r i n e (93)

Topoisomerase ll-targeted domain

=

, N~~ NH.~~ ~-'#

! o H N.~1.,,.,.~N~ I:1 O

NH2 C;H3

MePyGA(95)

NH2

H

~ DNA-intercalating moiety

t~1

O

skeletalcore

(~H3

NetGA(94)

DNA minor groove binding and topoisomerase-targeted moiety

molecule Distel(2+) (98) binds specifically to AT sites. A molecular model of the Distel(2+)-DNA complex is shown in Figure 6. In contrast to Distel(l+), the interaction of Distel(2+) with DNA seems to be driven as much by the distamycin moiety as by the ellipticine residue. A different situation has been reported with the biscationic netropsin-oxazolopyridocarbazole combilexin Net-OPC (99). Oxazolopyridocarbazoles are intercalating drugs derived from ellipticines and exhibit cytotoxic properties against malignant cells in vitro. 306'307 A detailed characterization of the complexes between o

,CH3

O

CH3

CH3 0

,o

H CH~ I

~"N"

CH3

(97)

HN

ellipticine derivative

Distel (1 +)

~'-r~ ~ HN.~./N ...CH3

H

138

CHRISTIAN BAILLY H

(~")

Distel (2+)

HNH~ OcH:= NHz

Not-oec HN

NH2

,,o~N

HN--~ "CH=

N

H

Net-OPC and DNA showed that the hybrid ligand adopts different configurations according to the sequence to which it binds. The hybrid binds much tighter to AT compared to GC sequences, as expected. At GC sites, the OPC chromophore intercalates into DNA but the netropsin remains unbound. At AT sites, the most energetically favored complex has both the netropsin and the OPC moieties inserted in the minor groove of a 7 base pair long sequence. In contrast, the second favored complex at (A.T) 4 sites involved intercalation of the OPC ring and minor groove recognition by the bispyrrole moiety. 3~ A somewhat similar coexistence of an intercalative and a nonintercalative binding mode was recently reported for the netropsin-porphyrin conjugate 61. 216 Theoretical studies reinforced the view that the connecting group between the two DNA binding elements play a critical role in the DNA recognition process. 217These typical examples illustrate the difficulties encountered in designing combilexin molecules. It is a challenging exercise which demands consideration of the notions of geometrical compatibility, hydrogen bonding capability, and the overall electronic properties of the interacting species.

Combilexins and Topoisomerase Inhibition As noted previously, drugs that intercalate into DNA are among the most useful cancer chemotherapeutic agents. However, the DNA intercalating properties of antitumor drugs such as acridines, anthracyclines, actinomycins, and anthracenediones are very similar to those of intercalators which have no therapeutic value such as ethidium and proflavine. There is mounting evidence suggesting that the antitumor activities of intercalating drugs is not due to interaction with DNA per se but is, at least in part, the result of the inhibition of enzymes that regulate DNA topology: the topoisomerases. 312'313 In most cases, topoisomerase poisons inhibit the reaction of topoisomerases with DNA by stabilizing an abortive reaction intermediate, termed the cleavable complex, whereby the DNA is cleaved but cannot be readily resealed. 314 The ATP-modulated protein-clamp model for topoisomerase II is an alternative model by which topoisomerase II inhibitors which do


139

Figure 6. View from the major groove of the best energy-minimized model of the Distel (2+)-[d(GCATATGC)2] complex. Complex formation induces a pronounced DNA bending towards the minor groove in the vicinity of the ellipticine binding site.302,303 not stabilize the cleavable complex may act. 315'316The cytotoxicity of drugs such as amsacrine, ellipticine, and daunomycin which trap the covalent enzyme-DNA complex is closely related to breakage of double-stranded DNA; for example as a result of the collision between replication forks and enzyme-DNA-drug ternary

140

CHRISTIAN BAILLY

complexes. Topoisomerase inhibitors can induce various types of lethal lesions in cells such as chromosomal aberrations, sister chromatid exchanges, nonhomologous recombinations, and genomic mutations. 317 Intercalation into DNA and the ensuing poisoning of topoisomerases affect the genetic stability and integrity of chromosome structure, and therefore contribute to the cytotoxic action of the drugs. Topoisomerases appear as one of the most promising targets for the design of more active antitumor agents. 318-32~ These observations prompted us to investigate the effect of combilexin molecules on the catalytic activities of mammalian DNA topoisomerases. We reasoned that the linkage of an intercalating chromophore, which stabilizes the topoisomerase II-DNA cleavable complex with a minor-groove binder related to distamycin which also affects topoisomerase functions, 321-325 may lead to novel categories of topoisomerase effectors and hopefully to the discovery of tumor-active compounds. In contrast to intercalating agents used in combilexin molecules (e.g. acridine, ellipticine), distamycin does not stabilize topoisomerase II-DNA cleavable complexes, but probably acts by impeding the access of topoisomerases to sequences in DNA selectively recognized by the drug. Therefore, a hybrid ligand may behave differently from its constituents. Of the two distamycin-ellipticine hybrids 97 and 98, only the monocationic hybrid Distel(l+) 97 proved to be a topoisomerase inhibitor. Its biscationic analogue Distel(2+) 98 showed practically no effect on both topoisomerase I and topoisomerase II despite its superior DNA binding properties. 326Distel(1 +) exhibits a moderate effect on topoisomerase II but exerts a significant effect on topoisomerase I. The effect is different from that observed with its constituents, namely distamycin and the ellipticine derivative. The poisoning of topoisomerase I by Distel(l+) detected using purified calf thymus enzyme is also observed in cells. Indeed, Distel(l+) is markedly cytotoxic towards P388 murine leukemia cells, whereas distamycin and Distel(2+) are totally inactive. Distel(l+) is even more cytotoxic than the ellipticine derivative. The toxicity is attributable to the effect of the hybrid ligand on topoisomerase I since P388CPT5 cells resistant to camptothecin (a powerful topoisomerase I inhibitor) display a notable cross-resistance to Distel(1 +).326 These results suggest that (t) tight and sequence-selective binding to DNA does not necessarily correlate with topoisomerase inhibition, and (2) of the two effects, binding to DNA and interference with topoisomerase functions, it is the latter which is important for the biological activity. In addition the results indicate that, as expected, combilexins may lead to new series of topoisomerase inhibitors. Amsacrine 93 (m-AMSA; 4'-(9-acridinylamino)-methanesulfon-m-anisidide) is an antitumor agent used in the clinic for the treatment of acute myelogenous leukemia. It is commonly accepted that the anticancer activity of amsacrine is connected with its ability to bind to DNA 327 and to interfere with topoisomerases. 312'328On the basis of structure-activity relationships, Baguley et al. 329 have proposed a model that divides amsacrine into two functional domains: the acridine


141

chromophore constitutes the DNA binding domain and the anilino group would represent the topoisomerase II binding domain. The methoxy group is not absolutely required for activity, but it can serve as a switch to enhance (meta position) or to block (ortho position) the activity of the drug. In fact, from recent studies it has been suggested that the l'-methanesulfonamide side chain attached to the anilino group is the putative site of interaction with topoisomerase 11.330 These observations prompted us to examine the issue of DNA binding and topoisomerase II poisoning by the mono- and bis-(N-methylpyrrolecarboxamide)-glycylanilinoacridine conjugates NetGA (94) and MePyGA (95). The studies showed that these two intercalating agents have the capacity to stabilize the topoisomerase II-DNA complex and stimulate the cutting of DNA at a subset of preexisting topoisomerase II cleavage sites. 300'331 The effect was attributed to the presence of the pyrrolecarboxamide moiety of the hybrids since the removal of the Nmethylpyrrole units (compound GA) abolishes the topoisomerase II-mediated DNA cleavage, but does not inhibit the intercalation of the drug into DNA. The spectroscopic and biochemical data lead to the conclusion that two functional domains can be identified in the combilexins 94 and 95: the anilino group can be regarded as a skeletal core to which are connected (1) the tricyclic acridine moiety which represents the DNA binding domain, and (2) the N-methylpyrrolecarboxamide moiety which constitutes the topoisomerase II-targeted domain. Therefore, the study supports the original model 329'330 in showing that the presence of a substituent at position 1' of the anilinoacridine chromophore is required to permit the drug to interfere with the catalytic activities of topoisomerase II. The existence of two functional domains in amsacrine and these combilexin molecules also accords with the structure-activity analyses on analogues of anthracyclines, 332 epipodophyllotoxin derivatives, 333 and several other structurally different topoisomerase II poisons. 334'335 Thus it may turn out that such a model points to a general principle valid for the majority of topoisomerase II inhibitors. The pyrrolecarboxamide-anilinoacridine derivatives 94 and 95 exhibit marked cytotoxic properties in vitro which correlate with their effects on topoisomerase 11.336 The cytotoxicity of compound 94 is remarkably dependent on topoisomerase II since the drug is 170 times more cytotoxic towards the Chinese hamster lung cells DC3F compared to DC3F/9-OHE cells which are resistant to the topoisomerase II inhibitor, 9-hydroxy-ellipticine. 337-339 DC3F/9-OHE cells contain about fivefold less topoisomerase IIt~ than the sensitive DC3F cells, and topoisomerase 1113is practically undetectable. 34~It is interesting to note that compound GA which has no effect on topoisomerase II is weakly cytotoxic. Moreover, unlike compound GA, the combilexin molecules 94 and 95 display noticeable antitumor properties in vivo (T/C--140). Although potential correlation does not mean causality, 341 it seems likely that topoisomerase II constitutes a primary target for such hybrid molecules in vivo. A moderate antitumor activity in vivo was observed with the bifunctional molecule ThiaNetGA 100 which combines features of both the lexitropsin and com-

142

CHRISTIAN BAILLY

bilexin approaches. 342 This hybrid ligand, in which are conjugated the thiazolelexitropsin 20 and the intercalating anilino-acridine chromophore GA, binds to DNA via a bimodal process involving minor-groove binding of the lexitropsin moiety and intercalation of the acridine moiety. Studies are in progress to determine whether the antitumor activity is connected with toposiomerase inhibition. Very recently, we have developed a second-generation combilexin which consists of a minor groove-binding netropsin-like moiety attached to an intercalating chromophore structurally related to amsacrine. This netropsin-amsacrine hybrid, called NetAmsa (101), differs from the netropsin-anilinoacridine derivative NetGA (94), bears a positively charged terminal side chain which contributes significantly to the AT-selectivity of such ligands, 3~ and retains the m-methoxy and methanesulphonamide substituents on the anilino ring which constitute key elements for the interference with topoisomerases, the maintenance of redox properties, and the biological properties of amsacrine. The netropsin moiety is connected to the acridine ring via a 4-carboxamide side chain, whereas it was previously attached directly to the anilino group. Linkage of a carboxamide side chain to position 4 of the acridine ring of amsacrine has earlier been shown to convert the drug from a classical intercalator to a threading intercalator. 343'344Structural and kinetic studies have revealed that the conjugate threads through the DNA double helix so as to intercalate its acridine chromophore, leaving the netropsin moiety and the methanesulfonanilino group positioned within the minor and major grooves of the double helix, respectively, as depicted in Figure 7. 345 In addition, the hybrid retains the susceptibility to copper-dependent oxidation, characteristic of the parent amsacrine moiety, as well as its ability to generate oxygen radicals which can mediate DNA damage, mainly at cytidine and guanosine nucleotides. 346It also retains the property of stimulating the formation of cleavable complexes with DNA in the presence of topoisomerase II, but its netropsin-like moiety confers little or no influence on the reaction with topoisomerase I. 346 Thus far, the combilexin strategy concerned netropsin- or lexitropsin-like minor groove binders attached to different intercalating chromophores. In 1995, the strategy was extended to another category of minor-groove binders, namely 1,3diaryltriazenes derived from the antiviral and antiprotozoal drug berenil (6). The acridine-triazene combilexin (1t)2) exhibits a distinct preference for AT base tracts rather than GC-rich sequences and is about 40-fold more cytotoxic than the triazene or acridine subunits towards L1210 mouse leukemia and A2780 human colon cancer cell lines. 347The results on combilexin molecules obtained so far encourage us tobelieve that this approach to DNA-targeted pharmacology has the potential to yield important developments in the search for new classes of topoisomerase inhibitors, and perhaps for new and better anticancer drugs. The epipodophyllotoxins etoposide (VP-16-213) and teniposide (VM-26) are very potent topoisomerase II inhibitors and are largely used in chemotherapy for the treatment of small cell lung cancer, testicular cancer, lymphoma, and leukemia. These compounds stabilize DNA-topoisomerase II complexes but interact only


143

Figure 7. View from the minor groove of an energy-minimized model of the complex between NetAmsa (101) and d(GCGCAATTGCGC)2. The molecular model illustrates the threading of the hybrid through the double helix so as to occupy both the minor and major grooves. The netropsin moiety of the hybrid lies in the minor groove of the AATT central core, the acridine ring is intercalated between the last A.T base pair and the adjacent C.G base pair and the methansulfonanilino group resides in the major groove. 345

weakly, if at all, with DNA in the absence of the enzyme. Lown and coworkers have attempted to direct epipodophyllotoxins to specific sequences via the conjugation with lexitropsins. A series of 4'-demethylepipodophyllotoxin-lexitropsin conjugates such as compound 103 were synthesized but these compounds were found to be much less cytotoxic than the parent compounds. 348

144

CHRISTIAN BAILLY o

~._j NH.~NH (100) ThiaNet-GA

N

I-~C-~-.NH O

~O (101) Net~m~

~--s Ns~O

-.CHa HN O

.N--t,

..~N,c F

H=C.,N,.C~H ,,,,J

H3C H.c'N"~

(102)

(103) N.,NH O/%O N H3CO ~

~ : ~ ~ C H s OO ~

.o~o H3co

o,,(,o )

. H ~,)'"~O ~N

"OCHa

.y~'-N*

VI. CONCLUSION Netropsin and distamycin have largely contributed to our understanding of the molecular basis of drug-DNA recognition. The understanding of how these two antibiotics recognize and bind to AT sequences in DNA gave perceptions of how to design sequence-specific DNA binding molecules. The knowledge arising from these two antibiotics has inspired the rational design of DNA reading molecules; hopefully, they will contribute to tackling the therapy of chemoresistant tumors. As shown in this review, the chemical structure of netropsin-like DNA reading elements may be varied in many synthetically convenient fashions, to produce different repertoires of sequence-selective DNA binding molecules with a range of functionalities. Sterically demanding groups can be attached to netropsin and distamycin without abolishing their sequence recognition properties. The fact that certain netropsin analogues can readily recognize a specific sequence in DNA (e.g. lexitropsins forming 2:1 complexes) shows DNA targeting with minor-groove binding drugs is no longer a possibility but a practical reality. The fact that certain netropsin conjugates capable of inducing DNA damages at specific sequences exhibit potent antitumor activities (e.g. netropsin-nitrogen mustard conjugates)


145

show that this DNA-targeted strategy has the potential to yield new and efficient antitumor agents. There is good reason to believe that netropsin and distamycin will continue to inspire the development of new anticancer agents. 349-352 In the many examples presented in this review, netropsin and distamycin are almost always used to deliver cytotoxic drugs to specific sequences in DNA, but they can also be used to deliver DNA to cells. Gene transfer with netropsin-lipid conjugates has been attempted. 353 Moreover, netropsin and distamycin analogues can also be used independently of their capacity to bind to AT-rich DNA sequences. For example, sulfonated and phosphonated ureido dimers of netropsin such as compound 104 containing six sulfonic acid units and the phosphonic acid-containing ligand 105 represent a new class of surface-acting antiviral agents. Their mechanism of action apparently does not require binding to DNA but may involve the disruption of the virus attachment to CD4+-susceptible cells via an effect on CD4-gp 120 interactions. 354 Although there are a few netropsin conjugates under clinical or preclinical evaluation, medicinal chemists still have much to do to design new clinically useful drugs endowed with the aptitude for reading any given DNA sequence, and inducing the required DNA modification to obtain the desired biological response. Computer-based methods will increasingly serve to delineate more precisely the molecular rules that govern drug-DNA recognition. The increasing evidence that DNA binding proteins such as topoisomerases and transcription factors mediate the effects of drugs suggest that in addition to elucidating the structures of drug-DNA complexes, it will be necessary to examine in detail the effects of drugs on protein-DNA complexes. Close collaboration between medicinal chemists and molecular biologists will be necessary if we are to gain an improved understanding PO3H2

PO3H2

(105)

NH H3

NH

S.O3H H O 3 S ~ T

v

H

H3C

HN CH3

Ha

~

"NH

S.O3H

"SO3H

HN

SO3HCc~N~ H3

H

HN

(104)

NH

H

H

H~c'

~~N~c) HN

'cN~ ureido dimers

S03H He

146

CHRISTIAN BAILLY

of the mechanisms whereby antitumor drugs affect the function of particular oncogenes. Over the last 15 years, research on small molecules acting on nucleic acids has not only led to therapeutically useful drugs but has also provided a inexhaustible source of structural and biological information on nucleic acids. The recent discovery of new tumor-active compounds, such as those based on netropsin and distamycin reported here, has restored small molecules to the forefront of cancer therapy. It is still premature to state that netropsin and distamycin derivatives will provide new generations of sequence-specific antitumor drugs. However, the results obtained thus far are suggestive and promising. They augur exciting developments for years to come.

ACKNOWLEDGMENTS I w~sh to thank my colleagues, past and present, who contributed to many of the studies cited in this chapter. Research performed in the author's laboratory are supported by the INSERM, the Association pour la Recherche contre le Cancer (ARC) and the Ligue Nationale Frangaise Contre le Cancer (Comit6 du Nord).

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

Bishop, J. M. Science 1987, 235, 305. Stein, C. A.; Cohen, J. S. Cancer Res. 1988, 48, 2659. H61~ne,C. Eur. J. Cancer 1991, 27, 1466. Thuong, N. T.; H61~ne,C. Angew. Chem. Int. Ed. Engl. 1993, 32, 666. Saison-Behmoaras, T.; H61~ne,C. In The Searchfor New Anticancer Drugs; Waring, M. J.; Ponder, B. A. J., Eds.; Kluwer Academic: London, 1992, p. 259. Anderson, W. E Science 1992, 256, 808. Thompson, L. Science 1992, 258, 744. Sikora, K. Gene Therapy 1994, 1, 149. Stevens, M. E G. In The Search for New Anticancer Drugs; Waring, M. J.; Ponder, B. A. J., Eds.; Kluwer Academic: London, 1992, p. 1. Waring, M. J. Annu. Rev. Biochem. 1981, 50, 159. Wakelin, L. E G.; Waring, M. J. In Comprehensive Medicinal Chemistry; Sammes, E G., Ed.; Pergamon: Oxford, 1990, Vol. 2, p. 703. Feigon, J.; Denny, W. A.; Leupin, W.; Kearns, D R. J. Med. Chem. 1984, 27, 450. Bailly, C.; Waxing, M. J. Biochemistry 1993, 32, 5985. Bailly, C.; Waxing, M. J. Antiviral Chem. Chemother. 1993, 4, 113. Bailly,C.; Denny, W.A.; Mellor, L.; Wakelin, L. E G.; Waxing,M. J. Biochemistry 1992, 31, 3514. Bailly, C.; H6nichart, J. E; Colson, E; Houssier, C. J. Mol. Recognit. 1992, 5, 155. Baguley, B. C. Anti-Cancer Drug Des. 1991, 6, 1. Denny, W. A. In The Search for New Anticancer Drugs; Waring, M. J.; Ponder, B. A. J., Eds.; Kluwer Academic: London, 1992, p. 19. Hurley, L. H. J. Med. Chem. 1989, 32, 2027. Steitz, T. A. Structural Studies of Protein-Nucleic Acid Interaction; University Press: Cambridge, 1993. Travers, A. A. DNA-Protein Interactions; Chapman and Hall: London, 1993.

Recognition and Modification of Double-Helical DNA 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53.

54. 55. 56. 57. 58. 59. 60. 61.

147

Pellegrini, M.; Ebright, R. H. J. Am. Chem. Soc. 1996, 118, 5832. Travers, A. A. Annu. Rev. Biochem. 1989, 58, 427. Churchill, M. E. A.; Travers, A. A. Trends Biochem. Sci. 1991, 16, 92. Bonnefoy, E.; Rouvi~re-Yaniv, J. EMBO J. 1991, 10, 687. Wang, S.; Cosstick, R.; Gardner, J. F.; Gumport, R. I. Biochemistry 1995, 34, 13082. Lee, D. K.; Horikoshi, M.; Roeder, R. G. Cell 1991, 67, 1241. Start, D. B.; Hawley, D. K. Cell 1991, 67, 1231. Juo, Z. S.; Chiu, T. K.; Leiberman, P. M.; Baikalov, I.; Berk, A. J.; Dickerson, R. E. J. Mol. Biol. 1996, 261,239. Reardon, B. J.; Winters, R. S.; Gordon, D.; Winter, E. Proc. Natl. Acad. Sci. USA 1993, 11327. King, C.-Y.; Weiss, M. A. Proc. Natl. Acad. Sci. USA 1993, 90, 11990. Wong, J. M.; Bateman, E. Nucleic Acids Res. 1994, 22, 1890. Ades, S. E.; Sauer, R. T. Biochemistry 1995, 34, 14601. Suzuki, M. EMBO J. 1989, 8, 797. Suzuki, M. J. Mol. Biol. 1989, 207, 61. Suzuki, M. In Nucleic Acids. and Molecular Biology; Eckstein, F.; Lilley, D. M. J., Eds.; Springer-Verlag: Berlin, 1991, Vol. 5, p. 126. Lindsey, G. G.; Thompson, P. J. Biol. Chem. 1992, 267, 14622. Suzuki, M.; Yagi, N. Proc. R. Soc. Lond. B 1991, 246, 231. Suzuki, M.; Gerstein, M.; Johnson, T. Protein Engr. 1993, 6, 565.. Strauss, F.; Varshavsky, A. Cell 1984, 37, 889. Bailly, E; Bailly, C.; Waring, M. J.; H6nichart, J. P. Biochem. Biophys. Res. Commun. 1992, 184, 930. Flock, S.; Bailly, E; Bailly, C.; Waring, M. J.; H6nichart, J. P.; Colson, P.; Houssier, C. J. Biomol. Struct. Dyn. 1994, 11,881. Bourdouxhe-Housiaux, C.; Colson, P.; Houssier, C.; Bailly, C. Anti-Cancer Drug Design, 1996, in press. Schumacher, M. A.; Choi, K. Y.; Zalkin, H.; Brennan, R. G. Science 1994, 266, 763-770. Van Dyke, M. W.; Hertzberg, R. P.; Dervan, P. B. Proc. Natl. Acad. Sci. USA 1983, 79, 5470. Harshman, K. D.; Dervan, P. B. Nucleic Acids Res. 1985, 13, 4825. Portugal, J.; Waring, M. J. Eur. J. Biochem. 1987, 167, 281. Bailly, C.; Donkor, I. O.; Gentle, D.; Thomalley, M.; Waring, M. J. Mol. Pharmacol. 1994, 46, 313. Abu-Daya, A.; Brown, P. M.; Fox, K. R. Nucleic Acids Res. 1995, 23, 3385. Churchill, M. E. A.; Suzuki, M. EMBO J. 1989, 8, 4189. Reeves, R.; Nissen, M. S. J. Biol. Chem. 1990, 265, 8573. Geierstanger, B. H.; Volkman, B. F.; Kremer, W.; Wemmer, D. E. Biochemistry 1994, 33, 5347. Verweij, J.; den Hartigh, J.; Pinedo, H. M. Antitunu~r Antibiotics in Cancer Chemotherapy--Principles and Practice; Chabner, B. A.; Collins, J. M., Eds.; J. B. Lippincott: Philadelphia, 1990, p. 382. Ray, R.; Snyder, R. C.; Thomas, S.; Koller, C. A.; Miller, D. M. J. Clin. Invest. 1989, 83, 2003. Snyder, R. C.; Ray, R.; Blume, S.; Miller, D. M. Biochemistry 1991, 30, 4290. Vigneswaran, N.; Mayfield, C. A.; Rodu, B.; James, R.; Kim, H.-G.; Miller, D. M. Biochemistry 1996, 35, 1106. Keniry, M. A.; Banville, D. L.; Simmonds, P. M.; Shafer, R. J. Mol. Biol. 1993, 231,753. Sastry, M.; Patel, D. J. Biochemistry 1993, 32, 6588. Van Dyke, M. W.; Dervan, P. B. Biochemistry 1983, 22, 2373. Liu, C.; Chen, F.-M. Biochemistry 1994, 33, 1419. Grootenhuis, P. D. J.; Kollman, P. A.; Seibel, G. L.; DesJarlais, R. L.; Kuntz, I. D. Anti-Cancer Drug Des. 1990, 5, 237.

148

CHRISTIAN BAILLY

62. Hurley, L. H.; Draves, E H. In Molecular Aspects ofAnticancer Drug-DNA Interactions; Neidle, S.; Waring, M. J., Eds.; Macmillan: London, 1993, Vol. l, pp. 89-133. 63. Hurley, L. H.; Sun, D. J. Mol. Recognit. 1994, 7, 123. 64. Boger, D. L.; Johnson, D. S. Angew. Chem. Int. Ed. Engl. 1996, 35, 1438. 65. Zimmer, C.; W~ihnert, U. Prog. Biophys. Mol. Biol. 1986, 47, 3 I. 66. Wartell, R. M.; Larson, J. E.; Wells, R. D.; Netropsin. J. Biol. Chem. 1974, 249, 6719. 67. Kopka, M. L.; Yoon, C.; Goodsell, D.; Pjura, E; Dickerson, R. E. J. Mol. Biol. 1985, 183, 553. 68. Coil, M.; Frederick, C. A.; Wang, A. H. J.; Rich, A. Proc. Natl. Acad. Sci. USA 1987, 84, 8385. 69. Coll, M.; Aymami, J.; van der Marel, G. A.; van Boom, J. H.; Rich, A.; Wang, A. H.-J. Biochemistry 1989, 28, 3 lO. 70. Tabemero, L.; Verdaguer, N.; Coil, M.; Fita, I.; van der Marel, G. A.; van Boom, J. H.; Rich, A.; Aymami, J. Biochemistry 1993, 32, 8403. 7 I. Goodsell, D. S.; Leung, Ng H.; Kopka, M. L.; Lown, J. W.; Dickerson, R. E. Biochemistry 1995, 34, 16654. 72. Kopka, M. L.; Yoon, C.; Goodsell, D.; Pjura, E; Dickerson, R. E. Proc. Natl. Acad. Sci. USA 1985, 82, 1376. 73. Chen, X.; Ramakrishnan, B.; Sundaralingam, M. Nature Struct. Biol. 1995, 2, 733. 74. Goodsell, D. S.; Kopka, M. L.; Dickerson, R. E. Biochemistry 1995, 34, 4983. 75. Youngquist, R. S.; Dervan, E B. Proc. Natl. Acad. Sci. USA 1985, 82, 2565. 76. Dervan, E B. Science 1986, 232, 464. 77. Goodsell, D.; Dickerson, R. E. J. Med. Chem. 1986, 29, 727. 78. Khodin, A. A.; Krylov, A. S.; Grokhovsky, S. L.; Zhuze, A. L.; Zasedatelev, A. S.; Gursky, G. V.; Gottikh, B. E FEBS Lett. 1980, 118, 311. 79. Lown, J. W.; Krowicki, K.; Balzarini, J.; Newman, R. A.; De Clercq, E. J. Med. Chem. 1989, 32, 2368. 80. Wang, W.; Lown, J. W. J. Med. Chem. 1992, 35, 2890. 81. Kissinger, K. L.; Dabrowiak, J. C.; Lown, J. W. Chem. Res. Toxicol. 1990, 3, 162. 82. Rao, K. E.; Zimmermann, J.; Lown, J. W. J. Org. Chem. 1991, 56, 786. 83. Singh, M. E; Plouvier, B.; Hill, G. C.; Gueck, J.; Pon, R. T.; Lown, J. W. J. Am. Chem. Soc. 1994, 116, 7006. 84. Leinsoo, T. A.; Nikolaev, V. A.; Grokhovskii, S. L.; Surovaya, A. M.; Sidorova, N. Y.; Streltsov, S. A.; Zasedatelev, A. S.; Zhuze, A. L. Mol. Biol. 1989, 23, 1616. 85. Nikolaev, V. A.; Grokhovsky, S. L.; Surovaya, A. N.; Leinsoo, T. A.; Sidorova, N. Y.; Zasedatelev, A. S.; Zhuze, A. L.; Strahan, G. A.; Shafer, R. H.; Gursky, G. V. J. Biomol. Struct. Dyn. 1996, 14, 31. 86. Zakrzewska, K.; Pullman, B. J. Biomol. Struct. Dyn. 1988, 5, 1043. 87. Pullman, B. In Perspectives in Quantum Chemistry; Jortner, J.; Pullman, B., Eds.; Kluwer Academic: Dordrecht, 1989. 88. Zakrzewska, K.; Randrianarivelino, M.; Pullman, B. J. Biomol. Struct. Dyn. 1988, 6, 331. 89. Waring, M. J.; Bailly, C. Gene 1994, 149, 69. 90. Bailly, C.; Waring, M. J. Nucleic Acids Res. 1995, 23, 885. 91. Marchand, C.; Bailly, C.; McLean, M. J.; Moroney, S. E.; Waring, M. J. Nucleic Acids Res. 1992, 20, 5601. 92. Bailly, C.; Marchand, C.; Waring, M. J. J. Am. Chem. Soc. 1993, 115, 3784. 93. Bailly, C.; Payet, D.; Travers, A. A.; Waring, M. J. Proc. Natl. Acad. Sci. USA 1996, 93, 13623. 94. Waring, M. J.; Bailly, C. J. Mol. Recognit., in press. 95. Lown, J. W. Anti-Cancer Drug Des. 1988, 3, 25. 96. Lown, J. W. J. Mol. Recognit. 1994, 7, 79. 97. Lown, J. W.; Krowicki, K.; Balzarini, J.; De Clercq, E. J. Med. Chem. 1986, 29, 1210. 98. Lown, J. W.; Krowicki, K.; Bhat, U. G.; Skorobogaty, A.; Ward, B.; Dabrowiak, J. C. Biochemistry 1986, 25, 7408.

Recognition and Modification of Double-Helical DNA 99. 100. 101. 102.

149

Krowicki, K.; Lown, J. W. J. Org. Chem. 1987, 52, 3493. Kissinger, K. L.; Krowicki, K.; Dabrowiak, J. C.; Lown, J. W. Biochemistry 1987, 26, 5590. Lee, M.; Pon, R. T.; Krowicki, K.; Lown, J. W. J. Biomol. Struct. Dyn. 1988, 5, 939. Lee, M.; Coulter, D. M.; Pon, R. T.; Krowicki, K.; Lown, J. W. J. Biorru)l. Struct. Dyn. 1988, 5, 1059. 103. Lee, M.; Hartley, J. A.; Pon, R. T.; Krowicki, K.; Lown, J. W. Nucleic Acids Res. 1988, 16, 665. 104. Lee, M.; Chang, D. K.; Hartley, J. A.; Pon, R. T.; Krowicki, K.; Lown, J. W. Biochemistry 1988, 27, 445. 105. Lee, M.; Krowicki, K.; Hartley, J. A.; Pon, R. T.; Lown, J. W. J. Am. Chem. Soc. 1988, 110, 3641. 106. Burckhardt, G.; Luck, G.; Zimmer, C.; Sttid, J.; Krowicki, K.; Lown, J. W. Biochim. Biophys. Acta 1989, 1009, I 1. 107. Lee, M.; Rhodes, A. L.; Wyatt, M. D.; Forrow, S.; Hartley, J. A. Biochemistry 1993, 32, 4237. 108. Wyatt, M. D.; Garbiras, B. J.; Lee, M.; Forrow, S. M.; Hartley, J. A. Bioorg. Med. Chem. Lett. 1994, 4, 801. 109. Rao, K. E.; Bathini, Y.; Lown, J. W. J. Org. Chem. 1990, 55, 728. 110. Rao, K. E.; Shea, R. G.; Yadagiri, B.; Lown, J. W. Anti-Cancer Drug Des. 1990, 5, 3. 111. Kumar, S.; Jaseja, M.; Zimmermann, J.; Yadagiri, B.; Pon, R. T.; Sapse, A. M. J. Biorm)l. Struct. Dyn. 1990, 8, 99. 112. Kumar, S.; Bathini, Y.; Joseph, T.; Pon, R. T.; Lown, J. W. J. Biomx)l. Struct. Dyn. 1991, 9, 1. 113. Plou~,ier, B.; Bailly, C.; Houssin, R.; Rao, K. E.; Lown, J. W.; H6nichart, J.-P.; Waring, M. J. Nucleic Acids Res. 1991, 19, 5821. 114. Lee, M.; Krowicki, K.; Shea, R. G.; Lown, J. W.; Pon, R. T. J. Mol. Recognit. 1989, 2, 84. 115. Kopka, M. L.; Larsen, T. A. In Nucleic Acid Targeted Drug Design; Propst, C. L.; Perun, T. J., Eds.; Marcel Dekker: New York, 1992, pp. 303-374. 116. Pullman, A.; Pullman, B. Q. Rev. Biophysics 1981, 14, 289. 117. Marky, L. A.; Breslauer, K. J. Proc. Natl. Acad. Sci. USA 1987, 84, 4359. 118. Singh, S. B.; Wemmer, D. E.; Kollman, P. A. Proc. Natl. Acad. Sci. USA 1994, 91, 7673. 119. Zhang, Y.; Chen, F.-X.; Mehta, P.; Gold, B. Biochemistry 1993, 32, 7954. 120. Wade, W. S.; Dervan, P. B. J. Am. Chem. Soc. 1987, 109, 1574. 121. Bailly, C.; Colson, P.; Houssier, C.; Houssin, R.; Mrani, D.; Gosselin, G.; Imbach, J. L.; Waring, M. J.; Lown, J. W.; H6nichart, J. P. Biochemistry 1992, 31, 8349. 122. Kuwahara, J.; Sugiura, Y. Proc. Natl. Acad. Sci. USA 1988, 85, 2459. 123. Pelton, J. G.; Wemmer, D. E. Proc. Natl. Acad. Sci. USA 1989, 86, 5723. 124. Pelton, J. G.; Wemmer, D. E. J. Am. Chem. Soc. 1990, 112, 1393. 125. Fagan, P.; Wemmer, D. E. J. Am. Chem. Soc. 1992, 114, 1080. 126. Chen, X.; Ramakrishnan, B.; Rao, S. T.; Sundaraligam, M. Struct. Biol. 1994, 1, 169. 127. Rentzeperis, D.; Marky, L. A. Biochemistry 1995, 34, 2937. 128. Animati, E; Arcamone, E M.; Conte, M. R.; Felicetti, P.; Galeone, A.; Lombardi, P.; Mayol, L.; Paloma, L. G.; Rossi, C. J. Med. Chem. 1995, 38, 1140. I29. Wade, W. S.; Mrksich, M.; Dervan, P. B. Biochemistry 1993, 32, 11385. 130. Geierstanger, B. H.; Dwyer, T. J.; Bathini, Y.; Lown, J. W.; Wemmer, D. E. J. Am. Chem. Soc. 1993, 115, 4474. 131. Geierstanger, B. H.; Jacobsen, J. P.; Mrksich, M.; Dervan, P. B.; Wemmer, D. E. Biochemistry 1994, 33, 3055. 132. Mrksich, M.; Wade, W. S.; Dwyer, T. J.; Geierstanger, B. H.; Wemmer, D. E.; Dervan, P. B. Proc. Natl. Acad. Sci. USA 1992, 89, 7586. 133. Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1993, 115, 9892. 134. Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1994, 116, 3663. 135. Dwyer, T. J.; Geierstanger, B. H.; Mrksich, M.; Dervan, P. B.; Wemmer, D. E. J. Am. Chem. Soc. 1993, 115, 9900. 136. Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1994, 116, 7983.

150

CHRISTIAN BAILLY

137. 138. 139. 140. 141. 142. 143. 144.

Cho, J.; Parks, M. E.; Dervan, E B. Proc. Natl. Acad. Sci. USA 1995, 92, 10389. Trauger, J. W.; Baird, E. E.; Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6160. A1-Said, N. H.; Lown, J. W. Tetrahedron Lett. 1994, 35, 7577. Chen, Y.-H.; Lown, J. W. J. Am. Chem. Soc. 1994, 116, 6995. Parks, M. E.; Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6147. Parks, M. E.; Baird, E. E. J. Am. Chem. Soc. 1996, 118, 6153. Geierstanger, B. H.; Mrksich, M.; Dervan, P. B.; Wemmer, D. E. Science 1994, 266, 646. Sinyakov, A. N.; Lokhov, S. G.; Kutyavin, I. V.; Gamper, H. B.; Meyer, R. B. Jr. J. Am. Chem. Soc. 1995, II 7, 4995. Szewczyk, J. W.; Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6778. Robles, J.; Rajur, S. B.; McLaughlin, L. W. J. Am. Chem. Soc. 1996, 118, 5820. Wiederholt, K.; Rajur, S. B.; Giuliano, J. Jr.; O'Donnell, M. J.; McLaughlin, L. W. J. Am. Chem. Soc. 1996, 118, 7055. Szewczyk, J. W.; Baird, E. E.; Dervan, P. B. Angew. Chem. Int. Ed. Engl. 1996, 35, 1487. Mattes, W. A.; Hartley, J. A.; Kohn, K. W.; Matheson, D. W. Carcinogenesis 1988, 9, 2065. Zerial, M.; Salinas, J.; Filipski, J.; Bemardini, G. Eur. J. Biochem` 1986, 160, 479. Broggini, M.; Ponti, M.; Ottolenghi, S.; D'Incalci, M.; Mongelli, N.; Mantovani, R. Nucleic Acids Res. 1989, 17, 1051. Gambari, R.; Barbieri, R.; Nastruzzi, C.; Chiorboli, V.; Feriotto, G.; Natali, G. P.; Giacomini, P.; Arcamone, E Biochem. Pharmacol. 1991, 41,497. Dorn, A.; Affolter, M.; Muller, M.; Gehring, W. J.; Leupin, W. EMBO J. 1992, 11,279. Chai, S.; Alonso, J. C. Nucleic Acids Res. 1996, 24, 282. Gambari, R.; Nastruzzi, C. Biochem. Pharmacol. 1994, 47, 599. Broggini, M.; D'Incalci, M. Anti-Cancer Drug Des. 1994, 9, 373. Oakley, M. G.; Mrksich, M.; Dervan, P. B. Biochemistry 1992, 31, 10969. Browne, K. A.; He, G.-X.; Bruice, T. C. J. Am. Chem. Soc. 1993, 115, 7072. Blask6, A.; Bruice, T. C. Proc. Natl. Acad. Sci. USA 1993, 90, 10018. Blask6, A.; Browne, K. A.; Bruice, T. C. J. Am. Chem. Soc. 1994, 116, 3726. He, G.-X.; Browne, K. A.; Blaskb, A.; Bruice, T. C. J. Am. Chem. Soc. 1994, 116, 3716. He, G.-X." Browne, K. A.; Groppe, J. C." Blask6, A.; Mei, HI-Y.; Bruice, T. C. J. Am. Chem. Soc. 1993, 115, 7061. Blaskb, A.; Browne, K. A.; Bruice, T. C. BioOrg. Med. Chem. 1995, 3, 631. Xue, T.; Browne, K. A.; Bruice, T. C. Bioconjugate Chem. 1995, 6, 82. Hansma, H. G.; Browne, K. A.; Groppe, J. C.; Bezanilla, M.; Bruice, T. C. Biochemistry 1994, 33, 8436. Browne, K. A.; Bruice, T. C. J. Am. Chem. Soc. 1992, 114, 4951. Bruice, T. C.; Mei, H.-H.; He, G.-X.; Lopez, V. Proc. Natl. Acad. Sci. USA 1992, 89, 1700. Sigman, D. S.; Mazumder, A.; Perrin, D. M. Chem. Rev. 1993, 93, 2295. Hartley, J. A. In Molecular Aspects ofAnticancer Drug-DNA Interactions; Neidle, S.; Waring, M. J., Eds.; Macmillan: London, 1993, Vol. 1, p. 1. Hansen, M.; Hurley, L. Acc. Chem. Res. 1996, 29, 249. Tomasz, M. In Molecular Aspects ofAnticancer Drug-DNA Interactions; Neidle, S.; Waring, M. J., Eds.; Macmillan: London, 1994, Vol. 2, p. 312. Boyd, E L.; Cheatham, S. E; Remers, W.; Hill, G. C.; Hurley, L. H. J. Am. Chem. Soc. 1990, 112, 3279. Kopka, M. L." Goodsell, D.; Baikalov, I." Grzeskowiak, K." Cascio, D.; Dickerson, R. E. Biochemistry 1994, 33, 13593. Krowicki, K.; Balzarini, J.; De Clercq, E.; Newman, R. A.; Lown, J. W. J. Med. Chem. 1988, 31, 341.

145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174.


151

175. Arcamone, E A.; Animati, E; Barbieri, B.; Configliacchi, E.; D'Alesio, R.; Geroni, C.; Giulani, F. C.; Lazzari, E.; Menozzi, M.; Mongelli, N.; Penco, S.; Verini, M. A. J. Med. Chem. 1989, 32, 774. 176. Baker, B. E; Dervan, P. B. J. Am. Chem. Soc. 1985, 107, 8266. 177. Baker, B. E; Dervan, P. B. J. Am. Chem. Soc. 1989, 111, 2700. 178. Lee, M.; Rhodes, A. L.; Wyatt, M. D.; Forrow, S.; Hartley, J. A. Anti-Cancer Drug Des. 1993, 8, 173. 179. Xie, G.; Gupta, R.; Lown, J. W. Anti-Cancer Drug Des. 1995, 10, 389. 180. Wyatt, M. D.; Lee, M.; Garbirasd, B. J.; Souhami, R. L.; Hartley, J. A. Biochemistry 1995, 34, 13034. 181. Ciucci, A.; Manzini, S.; Lombardi, P.; Arcamone, E Nucleic Acids Res. 1996, 24, 311. 182. Chen, H. Y.; Liu, J. X.; Lown, J. W. BioOrg. Med. Chem. Lett. 1995, 5, 2223. 183. Gourdie, T. A.; Valu, K. K.; Gravatt, G. L.; Boritzki, T. J.; Baguley, B. C.; Wakelin, L. P. G.; Wilson, W. R.; Woodgate, P. D.; Denny, W. A. J. Med. Chem. 1990, 33, 1177. 184. Palmer, B. D.; Lee, H. H.; Johnson, P.; Baguley, B. C.; Wicklam, G.; Wakelin, L. P. G.; McFadyen, W. D.; Denny, W. A. J. Med. Chem. 1990, 33, 3008. 185. Prakash, A. S.; Denny, W. A.; Gourdie, T. A.; Valu, K. K.; Woodgate, P. D.; Wakelin, L. P. G. Biochemistry 1990, 29, 9799. 186. Valu, K. K.; Gourdie, T. A.; Boritzki, T. J.; Gravatt, G. L.; Baguley, B. C.; Wilson, W. R.; Wakelin, L. P. G.; Woodgate, P. D.; Denny, W. A. J. Med. Chem. 1990, 33, 3014. 187. Wickham, G.; Prakash, A. S.; Wakelin, L. P. G.; McFadyen, W. D. Biochim. Biophys. Acta 1991, 1073, 528. 188. Vlassov, V. V.; Gaidamakov, S. A.; Zarytova, V. F.; Knorre, D. G.; Levina, A. S.; Nikonova, A. A.; Podust, L. M.; Fedorova, O. S. Gene 1988, 72, 313. 189. Kutyavin, I. V.; Gamper, H. B.; Gall, A. A.; Meyer, R. B. Jr. J. Am. Chem. Soc. 1993, 115, 9303. 190. Broggini, M.; Erba, E.; Ponti, M.; Ballinari, D.; Geroni, C.; Spreafico, E; D'Incalci, M. Cancer Res. 1991, 51,199. 191. Montecucco, A.; Fontana, M.; Focher, E; Lestingi, M.; Spadari, S.; Ciarrochi, G. Nucleic Acids Res. 1991, 19, 1067. 192. Fontana, M.; Lestingi, M.; Mondello, C.; Braghetti, A.; Montecucco, A.; Ciarrochi, G.Anti-Cancer Drug Des. 1992, 7, 131. 193. Broggini, M.; Coley, H. M.; Mongelli, N.; Pesenti, E.; Wyatt, M. D.; Hartley, J. A.; D'Incalci, M. Nucleic Acids Res. 1995, 23, 81. 194. Broggini, M.; Moncollin, V.; D' Incalci, M.; Mongelli, N.; Mantovani, R. Nucleic Acids Res. 1995, 23, 1657. 195. Pezzoni, G.; Grandi, M.; Biasoli, G.; Capolongo, L.; Ballinari, D.; Giuliani, E C.; Barbieri, B.; Pastori, A.; Pesenti, E.; Mongelli, N. Br. J. Cancer 1991, 64, 1047. 196. Coley, H. M.; Mongelli, N.; D'Incalci, M. Biochem. Pharmacol. 1993, 45, 619. 197. Rhoads, C. J. JAMA 1946, 131,656. 198. Wang, Y.; Gupta, R.; Huang, L.; Luo, W.; Lown, J. W. Anti-Cancer Drug Des. 1996, 11, 15. 199. Fregeau, N. L.; Wang, Y.; Pon, R. T.; Wylie, W. A.; Lown, J. W. J. Am. Chem. Soc. 1995, 117, 8917. 200. Church, K. M.; Wurdeman, R. L.; Zhang, Y.; Gold, B. Biochemistry 1990, 29, 6827. 201. Sigurdsson, S. T.; Rink, S. M.; Hopkins, P. B. J. Am. Chem. Soc. 1993, 115, 12633. 202. Fagan, P. A.; Spielmann, H. P.; Sigurdsson, S. Th.; Rink, S. M.; Hopkins, P. B.; Wemmer, D. E. Nucleic Acids Res. 1996, 24, 1566. 203. Walker, W. L.; Kopka, M. L.; Filipowsky, M. E.; Dickerson, R. E.; Goodsell, D. S. Biopolymers 1995, 35, 543. 204. Nielsen, P. E. J. Mol. Recognit. 1990, 3, 1. 205. Takasugi, M.; Guendouz, A.; Chassignol, M.; Decout, J. L.; Lhomme, J.; Thuong, N. T.; H61~ne, C. Proc. Natl. Acad. Sci. USA 1991, 88, 5602.

152

CHRISTIAN BAILLY

206. Godard, G." Francois, J. C.; Duroux, I.; Asseline, U.; Chassignol, M.; Thuong, N.; H616ne, C." Saison-Behmoaras, T. Nucleic Acids Res. 1994, 22, 4789. 207. Hartley, J. A.; McAdam, S. R.; Das, S.; Roldan, M. C.; Haskell, M. K.; Lee, M. Anti-Cancer Drug Des. 1994, 9, 181. 208. Rao, K. E.; Gosselin, G." Mrani, D.; P6rigaud, C.; lmbach, J. L.; Bailly, C.; H6nichart, J. P." Colson, P.; Houssier, C.; Lown, J. W. Anti-Cancer Drug Des. 1994, 9, 221. 209. Hartley, J. A.; Webber, J.; Wyatt, M. D.; Bordenick, N.; Lee, M. BioOrg. Med. Chem. Lett. 1995, 3, 623. 210. Herfeld, P.; Helissey, P.; Giorgi-Renault, S. Bioconjugate Chem. 1994, 5, 67. 211. Bouziane, M.; Ketterl6, C.; Helissey, P.; Herfeld, P.; Le Bret, M.; Giorgi-Renault, S.; Auclair, C. Biochemistry 1995, 34, 14051. 212. Ketterl6, C.; Gabarro-Arpa, J.; Ouali, M.; Bouziane, M.; Auclair, C.; Helissey, P.; Giorgi-Renault, S.; Le Bret, M. J. Biomol. Struct. Dyn. 1996, 13, 963. 213. Helissey, P.; Bailly, C.; Vishwakarma, J. N.; Auclair, C.; Waring, M. J.; Giorgi-Renault, S. Anti-Cancer Drug Des. 1996, 11,527-551. 214. Meunier, B. Chem. Rev. 1992, 92, 141. 215. Munson, B. R.; Fiel, R. J. Nucleic Acid Res. 1992, 20, 1315. 216. Anneheim-Herbelin, G.; Perr6e-Fauvet, M.; Gaudemer, A.; Helissey, P.; Giorgi-Renault, S. Tetrahedron Lett. 1993, 34, 7263. 217. Perr6e-Fauvet, M.; Gresh, N. J. Biomol. Struct. Dyn. 1994, 11, 1203. 218. Matsumoto, T.; Utsumi, Y.; Sakai, Y.; Toyooka, K.; Shibuya, M. Heterocycles 1992, 34, 1697. 219. Wilson, W. R. In The Search for New Anticancer Drugs; Waring, M. J.; Ponder, B. A. J., Eds.; Kluwer Academic: London, 1992, p. 87. 220. Tocher, J. H.; Edwards, D. I. Biochem. Pharmacol. 1994, 48, 1089. 221. Kappen, L. S.; Lee, T. R.; Yang, C.-C.; Goldberg, I. H. Biochemistry 1989, 28, 4540. 222. Nishiwaki, E.; Lee, H.; Matsumoto, T.; Toyooka, K.; Sakurai, H.; Shibuya, M. Tetrahedron Lett. 1990, 31, 1299. 223. Parrick, J.; Porssa, M.; Davies, L. K.; Dennis, M. E; Patel, K. B.; Stratford, M. R. L.; Wardman, P. BioOrg. Med. Chem. Lett. 1993, 3, 1697. 224. Matsumoto, T.; Sakai, Y.; Toyooka, K.; Shibuya, M. Heterocycles 1992, 33, 135. 225. Nishiwaki, E.; Nakagawa, H.; Takasaki, M.; Matsumoto, T.; Sakurai, H.; Shibuya, M. Heterocycles 1990, 31, 1763. 226. Nishiwaki, E.; Lee, H.; Matsumoto, T.; Toyooka, K.; Sakurai, H.; Shibuya, M. Tetrahedron Lett. 1990, 31, 1299. 227. Grokhovsky, S. L.; Zubarev, V. E. Nucleic Acids Res. 1991, 19, 257. 228. Galas, D. J.; Schmitz, A. Nucleic Acids Res. 1978, 5, 3157. 229. Dabrowiak, J. C.; Stankus, A. A.; Goodisman, J. In Nucleic Acid Targeted Drug Design; Propst, C. L.; Perun, T. J., Eds.; Marcel Dekker: New York, 1992, p. 93. 230. Bailly, C.; Waring, M. J. J. Biomol. Struct. Dyn. 1995, 12, 869. 231. Schultz, P. G.; Dervan, P. B. J. Biomol. Struct. Dyn. 1983, 1, 1133. 232. Taylor, J. S.; Schultz, P. G.; Dervan, P. B. Tetrahedron 1984, 40, 457. 233. Schultz, P. G.; Dervan, P. B. Proc. Natl. Acad. Sci. USA 1983, 80, 6834. 234. Schultz, P. G.; Taylor, J. S.; Dervan P. B. J. Am. Chem. Soc. 1982, 104, 6861. 235. Schultz, P. G.; Dervan, P. B. J. Am. Chem. Soc. 1983, 105, 7748. 236. Youngquist, R. S.; Dervan, P. B. J. Am. Chem. Soc. 1985, 107, 5528. 237. Youngquist, R. S.; Dervan, P. B. J. Am. Chem. Soc. 1987, 109, 7564. 238. Wade, W. S.; Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1992, 114, 8783. 239. Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1993, 115, 2572. 240. Camerman, N.; Camerman, A.; Sarkar, B. Can. J. Chem. 1976, 54, 1309. 241. Pickart, L.; Freedman, J. H.; Loker, W. J.; Peisach, J.; Perkins, C. M.; Stenkamp, R. E.; Weinstein, B. Nature 1980, 288, 715.

Recognition and Modification of Double-Helical DNA 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278. 279. 280.

15 3

Freedman, J. H.; Pickart, L.; Weinstein, B.; Mims, W. B.; Peisach, J. Biochemistry 1982, 21, 4540. Chikira, M.; Sato, T.; Antholine, W. E.; Petering, D. H. J. Biol. Chem. 1991, 266, 2859. Sigman, D. S.; Mazumder, A.; Perrin, D. M. Chem. Rev. 1993, 93, 2295. Bailly, C.; Bernier, J. L.; Houssin, R.; Helbecque, N.; H6nichart, J. P. Anti-Cancer Drug Des. 1987, 1,303. Morier-Teissier, E.; Bailly, C.; Bemier, J. L.; Houssin, R.; Helbecque, N.; Catteau, J. P.; Colson, P.; Houssier, C.; H6nichart, J. P. Anti-Cancer Drug Des. 1989, 4, 37. Morier-Teissier, E.; Bernier, J. L.; Lohez, M.; Catteau, J. P.; H6nichart, J. P. Anti-Cancer Drug Des. 1990, 5, 291. Bailly, C.; Sun, J. S.; Colson, P.; Houssier, C.; H61~ne, C.; Waring, M. J.; HEnichart, J. P. Bioconjugate Chem. 1992, 3, 100. Grokhovskii, S. L.; Nikolaev, V. A.; Zubarev, V. E.; Surovaya, A. N.; Zhuze, A. L.; Chemov, B. K.; Sidorova, N. Y.; Zasedatelev, A. S.; Gurskii, G. V. Molecular Biol. 1993, 6, 839. Chen, X. C.; Rokita, S. E.; Burrows, C. J. J. Am. Chem. Soc. 1991, 113, 5884. Mack, D. P.; Dervan, P. B. J. Am. Chem. Soc. 1990, 112, 4604. Footer, M.; Egholm, M.; Kron, S.; Coull, J. M.; Matsudaira, P. Biochemistry 1996, 35, 10673. Hecht, S. M. Acc. Chem. Res. 1986, 19, 383. Stubbe, J.; Kozarich, J. W. Chem. Rev. 1987, 87, 1107. Natrajan, A.; Hecht, S. M. In Molecular Aspects ofAnticancer Drug-DNA Interactions; Neidle, S.; Waring, M. J., Eds.; Macmillan: London, 1994, Vol. 2, p. 197. Pratviel, G.; Bernadou, J.; Meunier, B. Angew. Chem. Int. Ed. Engl. 1995, 34, 746. Stubbe, J.; Kozarich, J. W.; Wu, W.; Vanderwall, D. E. Acc. Chem. Res. 1996, 29, 322. Farinas, E.; Tan, J. D.; Baidya, N.; Mascharak, P. K. J. Am. Chem. Soc. 1993, 115, 2996. H6nichart, J. P.; Houssin, R.; Bemier, J. L.; Catteau, J. P. J. Chem. Soc., Chem. Commun. 1982, 1295. H6nichart, J. P.; Bemier, J. L.; Houssin, R.; Lohez, M.; K6nani, A.; Catteau, J. P. Biochem. Biophys. Res. Commun. 1985, 126, 1036. Kohda, J.; Shinozuka, K.; Sawai, H. Tetrahedron Lett. 1995, 36, 5575. Otsuka, M.; Masuda, T.; Haupt, A.; Ohno, M.; Shiraki, T.; Sugiura, Y.; Maeda, K. J. Am. Chem. Soc. 1990, 112, 838. Owa, T.; Haupt, A.; Otsuka, M.; Kobayashi, S.; Tomioka, N.; Itai, A.; Ohno, M.; Shiraki, T.; Uesugi, M.; Sugiura, Y.; Maeda, K. Tetrahedron 1992, 48, 1193. Huang, L.; Morgan, R.; Lown, J. W. BioOrg. Med. Chem. Lett. 1993, 3, 1751. Huang, L.; Quada, J. C.; Lown, J. W. Tetrahedron Lett. 1994, 35, 5323. Huang, L.; Quada, J. C. Jr.; Lown, J. W. Bioconjugate Chem. 1995, 6, 21. Huang, L.; Quada, J. C. Jr.; Lown, J. W. BioOrg. Med. Chem. 1995, 3, 647. Kane, S. A.; Natrajan, A.; Hecht, S. M. J. Biol. Chem. 1994, 269, 10899. Oakley, M. G.; Turnbull, K. D.; Dervan, P. B. Bioconjugate Chem. 1994, 5, 242. Zarytova, V. F.; Sergeyev, D. S.; Godovikova, T. S. Bioconjugate Chem. 1993, 4, 189. Nicolaou, K. C.; Dai, W.-M. Angew. Chem. Int. Ed. Engl. 1991, 30, 1387. Nicolaou, K. C.; Dai, W.-M.; Tsay, S.-C.; Estevez, V. A.; Wrasidlo, W. Science 1992, 256, 1172. Nicolaou, K. C.; Smith A. L.; Yue, E. W. Proc. Natl. Acad. Sci. USA 1993, 90, 5881. Tokuda, M.; Fujiwara, K.; Gomibuchi, T.; Hirama, M.; Uesugi, M.; Sugiura, Y. Tetrahedron Lett. 1993, 34, 669. Semmelhack, M. F.; Gallagher, J. J. J. Org. Chem. 1993, 59, 4357. Bregant, T. M.; Groppe, J.; Little, R. D. 1994, 116, 3635. Spielmann, H. P.; Fagan, P. A.; Bregant, T. M.; Little, R. D.; Wemmer, D. E. Nucleic Acids Res. 1995, 23, 1576. Xie, G.; Morgan, A. R.; Lown, J. W. BioOrg. Med. Chem. Lett. 1993, 3, 1565. Flanagan, M. E.; Rollins, S. B.; Williams, R. M. Chem. Biol. 1995, 2, 145. Bailly, C.; H6nichart, J. P. Bioconjugate Chem. 1991, 2, 379.

154

CHRISTIAN BAILLY

281. Bailly, C.; H6nichart, J. P. In Molecular Aspects of Anticancer Drug-DNA Interactions; Neidle, S.; Waring, M. J., Eds.; Macmillan: London, 1994, Vol. 2, p. 162. 282. Kamitori, S.; Takusagawa, E J. Mol. Biol. 1992, 225, 445. 283. Kamitori, S.; Takusagawa, E J. Am. Chem. Soc. 1994, 116, 4154. 284. Chen, H.; Liu, X.; Patel, D. J. J. Mol. Biol. 1996, 258, 457. 285. Bailly, C.; Ridge, G.; Graves, D. E.; Waring, M. J. Biochemistry 1994, 33, 8736. 286. Chu, W.; Shinomiya, M.; Kamitori, K. Y.; Kamitori, S.; Carlson, R. G.; Weaver, R. E; Takusagawa, E J. Am. Chem. Soc. 1994, 116, 797 I. 287. Wang, A. H. J. Curr. Opin. Struct. Biol. 1992, 2, 36 I. 288. Waring, M. J. In Molecular Aspects ofAnticancer Drug-DNA Interactions; Neidle, S.; Waring, M. J., Eds.; Macmillan: London, 1993, Vol. l, p. 213. 289. Sugiura, Y.; Shiraki, T.; Konishi, M.; Oki, T. Proc. Natl. Acad. Sci. USA 1993, 87, 383 I. 290. Lee, S. H.; Goldberg, I. H. Biochemistry 1989, 28, 1019. 291. Krivtsora, M. A.; Moroshkina, E. B.; Glibin, E. Mol. Biol. 1984, 18, 950. 292. Wartell, R. M.; Larson, J. E.; Wells, R. D. J. Biol. Chem. 1975, 250, 2698. 293. Patel, D. J.; Kozlowski, S. A.; Rice, J. A.; Broka, C.; Itakura, K. Proc. Natl. Acad. Sci. USA 1981, 78, 7281. 294. Bourdouxhe, C.; Colson, P.; Houssier, C.; H6nichart, J. P.; Waring, M. J.; Denny, W. A.; Bailly, C. Anti-Cancer Drug Des. 1994, 10, 131. 295. Eliadis, A.; Phillips, D. R.; Reiss, J. A.; Skorobogaty, A. J. Chem. Soc., Chem. Commun. 1988, 1049. 296. Shinomiya, M.; Kuroda, R. Tetrahedron Lett. 1992, 33, 2697. 297. Kuroda, R.; Satoh, H.; Shinomiya, M.; Watanabe, T.; Otsuka, O. Nucleic Acids Res. 1995, 23, 1524. 298. Bailly, C.; Pommery, N.; Houssin, R.; H6nichart, J. E J. Pharm. Sci. 1989, 78, 910. 299. Bailly, C.; Helbecque, N.; H6nichart, J. P.; Colson, P.; Houssier, C.; Rao, K. E.; Shea, R. G.; Lown, J. W. J. Mol. Recognit. 1990, 3, 26. 300. Foss6, P.; Ren6, B.; Saucier, J. M.; H6nichart, J. P.; Waring, M. J.; Colson, P.; Houssier, C.; Bailly, C. Biochemistry 1994, 33, 9865. 301. Bailly, C.; OhUigin, C.; Rivalle, C.; Bisagni, E.; H6nichart, J. P.; Waring, M. J. Nucleic Acids Res. 1990, 18, 6283. 302. Bourdouxhe, C.; Colson, P.; Houssier, C.; Sun, J.-S." Montenay-Garestier, T.; H61~ne, C.; Rivalle, C.; Bisagni, E.; Waring, M. J.; H6nichart, J.-P.; Bailly, C. Biochemistry 1992, 31, 12385. 303. Bailly, C." Lecl~re, V." Pommery, N.; Colson, P.; Houssier, C." Rivalle, C.; Bisagni, E.; H6nichart, J. P. Anti-Cancer Drug Des. 1993, 8, 145. 304. Bailly, C.; OhUigin, C.; Houssin, R.; Colson, P.; Houssier, C.; Rivalle, C.; Bisagni, E.; H6nichart, J. P.; Waring, M. J. Mol. Pharmacol. 1992, 41,845. 305. Bailly, C.; Michaux, C.; Colson, P.; Houssier, C.; Sun, J. S.; Garestier, T.; H61~ne, C.; H6nichart, J. P.; Rivalle, C.; Bisagni, E.; Waring, M. J. Biochemistry 1994, 33, 15348. 306. Auclair, C. Arch. Biochem. Biophys. 1987, 258, 1. 307. Auclair, C.; Schwaller, M. A.; Ren6, B.; Banoun, H.; Saucier, J. M.; Larsen, A. K. Anti-Cancer Drug Des. 1988, 3, 133. 308. Mrani, D.; Gosselin, G.; Auclair, C.; Balzarini, J.; De Clercq, E.; Paoletti, C.; Imbach, J. L. Eur. J. Med. Chem. 1991, 26, 481. 309. Subra, F.; Carteau, S.; Pager, J.; Paoletti, J.; Paoletti, C.; Auclair, C.; Mrani, D.; Gosselin, G.; Imbach, J. L. Biochemistry 1991, 30, 1642. 310. Subra, E; Mouscadet, J. E; Lavignon, M.; Roy, C.; Auclair, C. Biochem. Pharmacol. 1993, 45, 93. 311. Goulaouic, H.; Carteau, S.; Subra, E; Mouscadet, J. E; Auclair, C.; Sun, J.-S. Biochemistry 1994, 33, 1412.


155

312. Ralph, R. K.; Judd, W.; Pommier, Y.; Kohn, K. W. In Molecular Aspects ofAnticancer Drug-DNA Interactions; Neidle, S.; Waring, M. J., Eds.; Macmillan: London, 1994, Vol. 2, p. 1. 313. Wang, J. C. Annu. Rev. Biochem. 1996, 65, 635. 314. Berger, J. M.; Gamblin, S. J.; Harrison, S. C.; Wang, J. C. Nature 1996, 379, 225. 315. Roca, J.; Wang, J. C. Cell 1992, 71,833. 316. Roca, J.; Wang, J. C. Cell 1994, 77, 609. 317. Pommier, Y.; Bertrand, R. In The Causes and Consequences of ChromosomaI Aberrations; Kirsch, I. R., Ed.; CRC: London, 1993, p. 277. 318. Liu, L. F. Annu. Rev. Biochem. 1989, 58, 351. 319. Zunino, E; Capranico, G. Anti-Cancer Drug Des. 1990, 5, 307. 320. Capranico, G.; Zunino, E Curr. Pharm. Des. 1995, 1, 1. 321. Fesen, M.; Pommier, Y. J. Biol. Chem. 1989, 264, 11354. 322. McHugh, M. M.; Woynarowski, J. M.; Sigmund, R. D.; Beerman, T. A. Biochem. Pharmacol. 1989, 38, 2323. 323. Woynarowski, J. M.; McHugh, M.; Sigmund, R. D.; Beerman, T. A. MoI. Pharmacol. 1989, 35, 177. 324. Woynarowski, J. M.; Sigmund, R. D.; Beerman, T. A. Biochemistry 1989, 28, 3850. 325. McHugh, M.; Sigmund, R. D.; Beerman, T. A. Biochem. Pharmacol. 1990, 39, 707. 326. Riou, J. E; Grondard, L.; Naudin, A.; Bailly, C. Biochem. Pharmacol. 1995, 50, 424. 327. Wilson, W. R.; Baguley, B. C.; Wakelin, L. P. G.; Waring, M. J. MoL Pharmacol. 1981, 20, 404. 328. Nelson, E. M.; Tewey, K. M.; Liu, L. E Proc. Natl. Acad. Sci. USA 1984, 81, 1361. 329. Baguley, B. C.; Holdaway, K. M.; Fray, L. M. J. Natl. Cancer Inst. 1990, 82, 398. 330. Zwelling, L. A.; Mitchell, M. J.; Satitpunwaycha, P.; Mayes, J.; Altschuler, E.; Hinds, M.; Baguley, B. Cancer Res. 1992, 52, 209. 331. Bailly, C.; Collyn-d'Hooghe, M.; Lantoine, D.; Fournier, C.; Hecquet, B.; Foss6, P.; Saucier, J. M.; Colson, P.; Houssier, C.; H6nichart, J. P. Biochem. Pharmacol. 1992, 43, 457. 332. Jensen, P. B.; SCrensen, S. B.; Sehested, M.; Demant, E. J. E; Kjeldsen, E.; Friche, E.; Hansen, H. H. Biochem. Pharmacol. 1993, 45, 2025. 333. Chow, K. C.; MacDonald, T. L.; Ross, W. E. Mol. Pharmacol. 1989, 34, 467. 334. Capranico, G.; Palumbo, M.; Tinelli, S.; Mabilia, M.; Pozzan, A.; Zunino, F. J. Mol. Biol. 1994, 235, 1218. 335. Palumbo, M.; Mabilia, M.; Pozzan, A.; Capranico, G.; Tinelli, S.; Zunino, E J. Mol. Recognit. 1994, 7, 227. 336. Ren6, B.; Foss6, P.; Kh61ifa, T.; Jacquemin-Sablon, A.; Bailly, C. Mol. Pharmacol. 1996, 49, 343. 337. Charcosset, J. Y.; Bendirdjian, J. P.; Lantieri, M. E; Jacquemin-Sablon, A. Cancer Res. 1985, 45, 4229. 338. Pommier, Y.; Schwartz, R. E.; Zwelling, L. A.; Kerrigan, D.; Mattern, M. R.; Charcosset, J. Y.; Jacquemin-Sablon,.A.; Kohn, K. W. Cancer Res. 1986, 46, 611. 339. Charcosset, J. Y.; Saucier, J. M.; Jacquemin-Sablon, A. Biochem. Pharmacol. 1988, 37, 2145. 340. Jacquemin-Sablon, A.; Bojanowski, K.; Casabianca-Pign~de, M. R.; Cr6mier, S.; Delaporte, C.; Khelifa, T.; Markovits, J.; Ren6, B.; Saucier, J. M.; Larsen, A. K. Bull. Cancer 1994, 81,381. 341. Gewirtz, D. A. Biochem. Pharmacol. 1991, 42, 2253. 342. Plouvier, B.; Houssin, R.; Hecquet, B.; Colson, P.; Houssier, C.; Waring, M. J.; H6nichart, J. P.; Bailly, C. Bioconjugate Chem. 5, 475. 343. Denny, W. A.; Wakelin, L. P. G. Cancer Res. 1986, 46, 1717. 344. Bailly, C.; Denny, W. A.; Mellor, L.; Wakelin, L. P. G.; Waring, M. J. Biochemistry 1992, 31, 3514. 345. Bourdouxhe-Housiaux, C.; Colson, P.; Houssier, C.; Waring, M. J.; Bailly, C. Biochemistry 1996, 35, 4251. 346. H6nichart, J. P.; Waring, M. J.; Riou, J. E; Denny, W. A.; Bailly, C. Mol. Pharmacol. 1997, 51, 448. 347. McCaunaughie, A. W.; Jenkins, T. C. J. Med. Chem. 1995, 38, 3488.

156

CHRISTIAN BAILLY

348. 349. 350. 351. 352. 353. 354.

Gupta, R.; AI-Said, N. H.; Oreski, B.; Lown, J. W. Anti-Cancer Drug Des. 1996, 11,325. Lown, J. W. Drug Development Res. 1995, 34, 145. Geierstanger, B. H.; Wemmer, D. E. Annu. Rev. Biophyx. Biomol. Struct. 1995, 24, 463. Zunino, E; Animati, E; Capranico, G. Curr. Pharm. Des. 1995, 1, 83. Kahne, D. Chem. Biol. 1995, 2, 7. Remy, J.-S.; Sirlin, C.; Vierling, E; Behr, J.-E Bioconjugate Chem. 1994, 5, 647. Clanton, D. J.; Buckeit, R. W. Jr.; Terpening, S. J.; Kiser, R.; Mongelli, N.; Borgia, N. L.; Schultz, R.; Narayanan, V.; Bader, J. E; Rice, W. G. Antiviral Res. 1995, 27, 335.

NEW DEVELOPMENTS IN THE USE OF NITROGEN MUSTARD ALKYLATING AGENTS AS ANTICANCER DRUGS

William A. Denny

II.

III.

IV.

"Classical" Nitrogen Mustards . . . . . . . . . . . . . . . . . . . . . . . . . 158 A. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 B. Sequence Selectivity of D N A Alkylation . . . . . . . . . . . . . . . . . 158 C. Mechanism of D N A Alkylation . . . . . . . . . . . . . . . . . . . . . . 159 D. Limitations of "Classical" Nitrogen Mustards as Anticancer Drugs . . . . 160 DNA-Targeted Mustards . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 A. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 B. D N A Intercalating Mustards . . . . . . . . . . . . . . . . . . . . . . . . 162 C. D N A Minor-Groove Binding Mustards . . . . . . . . . . . . . . . . . . 164 Mustards as Effectors in Hypoxia-Activated Prodrugs . . . . . . . . . . . . . 167 A. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 B. Advantages of Mustards as Effectors for Hypoxia-Activated Prodrugs . . 167 C. Mustards as Effectors in Hypoxia-Activated Prodrugs . . . . . . . . . . 168 Mustards as Effectors in Prodrugs for Antibody-Directed E n z y m e - P r o d r u g Therapy (ADEPT) . . . . . . . . . . . . . . . . . . . . . . 171 A. Mustard Prodrugs for Carboxypeptidase . . . . . . . . . . . . . . . . . . 172


157

158

WILLIAM A. DENNY

B. Mustard Prodrugs for 13-Lactamase . . . . . . . . . . . . . . . . . . . . . C. Mustard Prodrugs for Penicillin G Amidase . . . . . . . . . . . . . . . . D. Mustard Prodrugs for 13-Glucuronidase . . . . . . . . . . . . . . . . . . . E. Mustard Prodrugs for NR2 Nitroreductase . . . . . . . . . . . . . . . . . V. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I.

173 173 173 173 174 175

"CLASSICAL" NITROGEN MUSTARDS

A. Introduction Compounds that alkylate DNA have long been of interest for their biological properties. They constitute major classes of both anticancer drugs and carcinogens. While a number of different types of chemical are able to alkylate DNA, historically the most important "simple" alkylating agents functioning as anticancer drugs are the nitrosoureas, platinum complexes, and particularly the nitrogen mustards [N,N-bis(2-chloroethyl)amines]. Nitrogen mustards (mustards) were among the earliest classes of agents used as systemic anticancer drugs. The simplest such compound, mechlorethamine (1), was one of the earliest clinical drugs. Derivatives still used today are the aniline mustards chlorambucil (2) and melphalan (3), and particularly the phosphoramide mustard cyclophosphamide (4), which is a component of many combination chemotherapy regimens. The biologically important initial lesion formed by mustards in cells is interstrand cross-links between different DNA bases, 1 although there is evidence that they also cause termination of transcription. 2 Many excellent reviews exist on the classical mustards, 3'4 and it is not the purpose of the present chapter to cover this in detail. Instead, only points required as background for the later sections are discussed.

B. Sequence Selectivity of DNA Alkylation The major site of DNA alkylation is at the N7 position of guanine, particularly at guanines in contiguous runs of guanines 5 that have the lowest molecular electrostatic potentials. 6 However, the level of selectivity of the initial attack by mustards (to form monoadducts) is quite low, with evidence 7 that most guanines are attacked. Recent studies with compounds 1-3 have also indicated significant levels of alkylation at the N3 position of adenine. 8-1~The sequence selectivity of cross-link formation by mustards is necessarily higher, because of the requirement to have two suitable sites juxtaposed. Early work ~1'12on the interaction of mechlorethamine with DNA isolated the bis-guanine adduct (5), and it was widely assumed that the cross-links formed between adjacent guanines (i.e. at 5'-GC or 5'-CG sites). However, recent work 13 has shown that the preferred cross-links are between nonadjacent guanines (at 5'-GNC sites).

Nitrogen Mustard Alkylating Agents

159

CI

MeN

,___/

R

1 CI Me

o HN~

,,--/

,Ju.., N

H2NJ'eN" -14>

5

CI

N

cO

N

CI 2 : R = (CH2)3COOH 3 : R = CH2CH(NH2)COOH

CI

,---"

N

4

o

bl~N,.1,t

N monofunctional > d(AG) > d(GG). 49 The diamine ligand slowed conversion of monoadduct to diadduct by a factor of 1.8 in comparison to the [Pt(NH3) 2] group.

DWA 2114R The relationship between the total number of platinum atoms binding to DNA and cytotoxicity indicated that DWA2114R was 2-3 times more effective cytotoxin than cis-DDP in HeLa S-3 cells. 5~ In agreement with the presence of the slowly displaced dicarboxylate group, 10 times the dose of DWA2114R is needed to produce the same cytotoxic effect. An in vitro assay showed that DWA 2114R adducts terminated DNA polymerase template activity in a manner similar to that of cis-DDR rendering it likely that the two compounds form a very similar array of adducts. 51 In agreement with this conclusion, CT DNA modified by DWA2114R shows similar conformational changes with a kinetic preference for dG nucleoside binding. 52

An Orally Active Platinum Complex The complex JM216 is currently in clinical trials. 53 A predicted metabolite is the Pt(II) complex, cis-[PtC12(NH3)(cyclohexylamine)]. The major product (54%) of the proposed metabolite with CT DNA is the d(GpG) intrastrand cross-link, with much smaller contributions from interstrand and long-range intrastrand crosslinks. 54 As this complex has two inequivalent amines (NH 3 and C6HllNH2),the adducts were distinguished by the novel finding of orientational isomers with either the 5' or 3' end of the dinucleotide oriented toward the NH 3 group. Replication of site specifically platinated M13 genomes with T7 DNA polymerase showed that the orientational isomers were less efficient than cis-DDP in inhibition of replication. 49

IV. DINUCLEAR PLATINUM COMPLEXES The weight of evidence indicates that all cisplatin analogues eventually produce an array of DNA adducts similar to those formed from cis-DDP. Differences in antitumor activity and toxicity may be dictated by more favorable pharmacokinetics of the analogues, rather than enhanced inhibition of target (DNA) function. Sequence specificity, conformational changes, and interactions with DNA processing

186

NICHOLAS FARRELL

proteins of the analogues appear to be qualitatively similar to the parent drug. Thus, the need arises to study "nonclassical" platinum complexes to examine the effectiveness of adducts structurally different to those of cis-DDP and its congeners. This section summarizes the DNA binding properties so far elucidated for two new classes of platinum drugs---dinuclear complexes containing two Pt-amine units linked by a flexible diamine chain and transplatinum complexes. The results clearly demonstrate that antitumor activity can be obtained with chemical structures distinct from cis-DDP. 55 Likewise, the array of DNA adducts produced by these two distinct series differ significantly from those formed by the cis-[PtX:(amine)2 ] family. Studies on the binding modes of these new clinical candidates allow us to examine how different modes of DNA binding may contribute to a different spectrum of antitumor activity. The following discussions build on recently published reviews. 56-58 A. Biological Effects of Dinuclear Platinum Complexes Dinuclear platinum complexes represent a large class of general formula

[{PtClm(NH3)3_m(H2N-R-NH2)] 2(2-m)+ (m = 0-3 and R is a linear or substituted

Pt

el'

/\

/\

I~

NH2(CH2)~I2N

el

2,2/c,c

o \ /~", Pt

H3N

/\

.,.\ /

-'12+

1~

/\

NH2(CH~)aH2 N

NH;

H3N

/\

/\

Pt

Pt

NH2(CH2)nH2N

NH3

l,l/e~

l,l/t,t

o\/.., Cl

/\

Pl

/\

NH2(CH2)~H2N

1,2/t,c

Pl NH3

CI

/\

Pl

/\

NH2(CH2)nH2N

1,2/c,c

Figure 3. Dinuclear platinum complexes.

Pl NH3

Platinum Antitumor Complexes

187

aliphatic linker). Within this broad scope, bifunctional, trifunctional, and tetrafunctional complexes are all possible (Figure 3). Examples of all types have been synthesized. The high antitumoral activity of dinuclear complexes in both cis-DDPsensitive and -resistant cells 59'6~as well as in human tumor cell lines inherently insensitive to cis-DDP 61 gives this series significant clinical potential. The first complex chosen from this series will enter clinical trials in late 1997 and is based on the 1,1/t,t geometry. The dinuclear platinum complexes demonstrate the first alternative mechanism of action for platinum complexes, with unique modes of DNA binding inaccessible to the mononuclear complexes. 57'58'62'63 The structures of the distinct bis(platinum)-DNA adducts will be described below. The UVrABC repair complex recognizes and excises damage induced by dinuclear platinum complexes. 64'65The efficiency of equivalent numbers of lesions in inhibition of Taq DNA polymerase and inhibition of transcription were measured in human cells using a host-cell r e a c t i v a t i o n assay.66 Inhibition of transcription was essentially equivalent for all dinuclear complexes and similar to that of cis-DDP but much more effective than lesions caused by either trans-DDP or the monofunctional [PtCI(NH3)3]+ (Figure 4). In contrast, inhibition of replication between the 1,1/t,t and 1,1/c,c geometric isomers differed by almost an order of magnitude. The relative repair efficiencies of all dinuclear platinum complexes in normal (NF) and repair-deficient. (XPA) human fibroblasts were less than that of cis-DDP.

II1 cO

r-----I Transcription Replication

7 e-

"~ .9

6 s

l= -9-

~ ~ ,,.

4

i._

.ag.

2

z

1

E

cisDDP

transDDP

triamine (Pt)

bis(Pt) 1,11t, t

bis(Pt) 2,21c,c

L_

bis(Pt) 1,11c,c

Figure 4. A comparison of the number of platinum lesions required to block RNA or DNA synthesis for a series of platinum complexes. The dinuclear complexes are abbreviated as in Figure 3. The extent of platinum-lesion "by-pass" by native RNA polymerase or Taq DNA polymerase was determined for each of the Pt complexes by comparing results from host-cell reactivation and QPCR assays. See Ref. 58 for details.

188

NICHOLAS FARRELL

Cellular studies using alkaline elution techniques in both murine L1210 and human tumor A2780 cells showed that dinuclear platinum complexes form significantly higher numbers of interstrand cross-links in comparison to cis-DDP, and that these persist longer than those of the mononuclear compound. 67'68 An interesting effect of geometric isomerism is also observed with the 1,1/c,c complex, producing significantly more interstrand cross-links than the 1,1/t,t isomer and being the predominant lesion in this case. 69 The interstrand cross-links also persist in cisDDP-resistant cell lines. This point is particularly relevant in view of the observations of gene-specific removal of cis-DDP interstrand cross-links in CHO cells, and the implications that acquired cellular resistance to cis-DDP may be associated with increased DNA repair of the interstrand cross-link. 7~ In summary, the lesions caused by the family of dinuclear compounds are in general as effective as those of cis-DDP, but relative efficiencies vary with the structure of the compound and thus the exact nature of the lesions formed. The lesions, or some thereof, appear also to be responsible for overcoming acquired

(Pt, Pt)lnterstrand Cross-link

(Pt, Pt)Intrastrand Cross-link

2+

2+ CI

H3N

X/ Pt /N

NH3

HaN

X/ Pt /\

NH2(CH2)nH2N

1,1/ t,t

C!

H3N

NH3

H3N

\/ PI /\

CI

CI

\/ Pt /\

NH2(CH2)nH2N

1,1/

NHz

NH3

c,c

Figure 5. Schematic representation of the limiting modes of bifunctional DNA binding of the cis- and trans-isomers of the dinuclear platinum complexes [{PtCI(NH3)2}21~-H2N(CH2)4NH2]2+. In the case of cis-DDP-like coordination spheres (See Figure 3), further reaction of these minimal structures will lead to tri- and tetrafunctional DNA-DNA and DNA-protein cross-links.


189

resistance to the parent cis-DDP. What then is the nature of these lesions, their global conformational changes, and how do they compare with those of cis-DDP?

B. DNA Interactions of Dinuclear Platinum Complexes Binding of dinuclear platinum complexes to DNA leads to: (1) DNA (Pt,Pt) interstrand cross-links by binding of one Pt atom to each strand of DNA; (2) (Pt,Pt) intrastrand cross-links by binding of the two Pt atoms to the same strand, as well as (3) DNA-protein cross-links, as shown in Figure 5. In the case of trifunctional and tetrafunctional complexes, further binding sites are possible and indeed a novel ternary complex of a DNA-DNA cross-link with elements of the UVrABC repair system produces unique DNA-protein adducts. 57

C. Conformational Changes Upon Formation of (Pt,Pt) Interstrand Cross-links Interstrand cross-link formation by dinuclear platinum complexes is significantly greater than for cis-DDp.56'72 The cross-linking is also dependent on the geometry around the Pt atom 61 and the length of the diamine chain. 73 cis-DDP, and indeed most alkylating agents, give rise to only 1,2-interstrand cross-links. In contrast, both the length and the flexibility of the diamine chain in dinuclear compounds allows the targeting of much larger DNA sequences for cross-link formation. 74'75 For binding between guanines on opposite strands, in addition to 1,2 cross-links, 1,3 and 1,4 interstrand cross-links are possible. In 1,3 and 1,4 cross-links the guanines are separated by one and two base pairs, respectively, whereas a 1,2 cross-link is formed from guanines on neighboring base pairs. In an interesting study of a 1,2 cross-link produced specifically in the 8-mer 5'-CATGCATG-3' by trans-[ {PtCI(NH3) 2}2kt-NH2(CH2)4NH2] 2+, a novel dumbbell structure is observed. 76 This structure arises from first folding the singlestranded DNA octamer into a mini-hairpin with a three-base pairs stem and a two-base loop. Now, since there are two such mini-hairpins arising from platination on both strands, placing the two hairpins end-over-end produces the dumbbell structure. A possible consequence for cruciform formation in longer oligonucleotides is shown in Figure 6. The formation of such structures may be facilitated by the syn conformation induced in the G sugar upon platination. The syn conformation of a platinated guanine has also been observed in the hairpin structure induced by cis-DDP binding in the palindromic sequence d(ATGGGTACCAT). 77 The induction of the syn conformation in both strands of DNA upon cross-linking by dinuclear platinum complexes may also contribute to the remarkable ease with which these compounds induce the B --> Z transition in poly(dG-dC).poly(dGdC). 65 As few as one 1,1/t,t bis(Pt) lesion per 25 bases is needed for complete conversion to the Z form. The syn conformation is an essential feature of the left-handed helix. 78 The effect of different adduct structures on conformational changes within a similar sequence may be appreciated by the fact that cis-DDP

190

NICHOLAS FARRELL

C'4 GC

1,l/t,t

Figure 6. Formation of a four-way DNA junction through a 1,2-intrastrand (Pt,Pt)cross-link.

stabilizes B-form poly(dG-dC).poly(dG-dC). 79'8~ The m o n o f u n c t i o n a l [PtCl(dien)] § also facilitates the B --> Z DNA transition at low r b but does not, by itself, fully induce the conformational change. 71 A further interesting feature of this B --> Z transition is that it is irreversible upon interstrand cross-linking. 81 The ability to lock the Z-DNA conformation could have important consequences in regulation of transcription, where the relevance of Z-DNA regions has been implicated. 82'83 It is noteworthy that d(CGTA'CG) 2 (A' =2-NHzA) modified by [PtCI(NH3) 3+ maintains the hexanucleotide in a Z-form. 84

D. Conformational Changes of the (Pt, Pt) Intrastrand Cross-link In a dinuclear complex, initial monofunctional binding of one center leads to possible (Pt,Pt) interstrand cross-linking by binding of the second center to the opposite strand or, alternatively, (Pt,Pt) intrastrand cross-linking by binding of the second Pt atom to the same strand, as shown in Figure 5. These preferences may be controlled by the geometry of the bis(platinum) complex as well as by the nature of the ligands coordinated to the platinum. 61 Model studies with the dinucleotide d(GpG) confirmed the formation of (Pt,Pt) 1,2-intrastrand cross-links with the dinuclear [ {trans-PtCl(NH3) 2}2HzN(CHz)6NH212§ 85 The formation of intrastrand cross-links was also observed in double-stranded DNA. 66 Unlike the structure of the cis-DDP adduct, 11'13 the (Pt,Pt) adduct features a syn conformation of the 3' base with the sugar pucker being more N type, although not 100% (c.f. the anti, anti conformation and the 100% N-type pucker of the cis-DDP structure). 86 Importantly, the (Pt,Pt) intrastrand cross-link appears to produce a less rigid adduct than cis-DDP. This is a result not only of the fact that one monofunctional Pt center binds to each of the two guanines, but also the flexibility of the diamine chain. A"stepped" head-to-head structure is proposed. 86

E. Effects of Conformational Changes of Bis(platinum)-DNA Interactions The conformational differences between the (Pt,Pt) intrastrand adduct and the intrastrand d(GpG) cross-link of cis-DDP could have important biological conse-


191

quences. The latter is not removed as efficiently as interstrand cross-links in gene-specific repair, 63 and other workers have confirmed that site-defined d(GpG)cis-DDP adducts are refractory to repair in cell extracts. 86 Differential repair of the (Pt,Pt) intrastrand adduct relative to that of cis-DDP could affect, for example, cytotoxicity in cis-DDP-resistant cells, where enhanced DNA repair is implicated in the mechanism of resistance. 87 The global conformational changes of importance to protein recognition of DNA are unwinding and bending. For the formally bifunctional platinum complexes [{ PtCI(NH3)2}zla-(HzN(CHz)4NH2)] 2+ the unwinding caused in supercoiled DNA is similar to that of cis-DDE We would not expect unwinding, therefore, to be the cause of any differential biological effects between dinuclear complexes and cis-DDE The bending of site-specifically modified oligonucleotides is not as

I cis-DDP 1 +

F]

I

~

......

~.':.T : :

/ ......

Figure 7. Proposal for differential pathways for protein recognition of (Pt,Pt)-DNA adducts in comparison to those of cis-DDP. The protein as depicted would recognize a bend as induced by cis-DDP intrastrand or interstrand cross-links. The flexibility and structural differences of the (Pt,Pt) adducts do not give the conformational changes required for protein recognition and are thus bypassed.

192

NICHOLAS FARRELL

significant due in part to the flexibility of the (Pt,Pt) intrastrand adduct, ss This structural difference between dinuclear platinum complexes and cis-DDP is further manifested by the significantly reduced recognition of (Pt,Pt)-DNA adducts by HMG proteins. 61 Bearing in mind that bending by cis-DDP is a significant structural feature contributing to recognition by structure-specific proteins, it is reasonable to conclude that the "downstream" effects of (Pt,Pt)-DNA adducts will be distinct from those discussed above. Figure 7 shows this feature in a schematic mannernwe are now beginning to see that different structures have different biological consequences with respect to interaction with DNA processing enzymes. These features, if correlated with a different spectrum of antitumor activity, point the way toward rational molecular mechanisms to design new drugs complementary to those currently employed in the anticancer field.

V. NOVEL ANTITUMOR ACTIVE TRANS-PLATINUM

COMPLEXES

Very early studies showed that the trans-DDP isomer was devoid of antitumor activity. 89 No major differences in uptake and bioavailability have been found between the cis- and trans-DDP isomers. 9~These observations led to emphasis on the differences in DNA binding to explain the dramatic difference in antitumor activity between the two compounds. Both kinetic and structural effects contribute to the inability of the trans-DDP-DNA adducts to effectively inhibit replication and transcription.

A. DNA Binding of trans.DDP The trans-[PtCl2(NH3) 2] is kinetically more reactive than its cis-isomer. The monofunctional adduct of trans-DDP may react readily with glutathione preventing closure to the toxic bifunctional adduct. 91'92 This chemical repair could be assisted by the persistence of long-lived monofunctional trans-DDP adducts. B ifunctional cis-DDP-DNA adducts may inhibit DNA replication to a greater extent than those formed from trans-DDP 93 or the array of DNA adducts formed by trans-DDP may be repaired more rapidly. 94 These explanations assume that an equal number of bifunctional lesions are formed by the two isomers. The trans-isomer is sterically inhibited from formation of 1,2-intrastrand cross-links but instead 1,3-d(GNG) cross-links are formed. These adducts also result in bending of the DNA, but the deformation is distinct from that caused by the cis-DDP structure and is best interpreted in terms of flexible hinge joints. 15'95 The 1,3-GAG binding site within the CCTCGAGTCTCC sequence also produces localized disruption of base pairing and destacking of the platinated bases, as indicated from NMR and molecular dynamic modeling studies. 96


193

The structure of the trans-DDP interstrand cross-link is between the GN7 and CN3 of the same base pair, rather than the two opposite GN7 atoms of a GC-C'G' duplex, as for cis-DDP. 97'98 Thus, the structure of the interstrand cross-links formed by cis- and trans-DDP are not the same. The conformational distortions of the interstrand cross-link have been measured as an unwinding of approximately 12 ~ and a bending of 260. 90

B. Novel trans.Platinum Complexes Use of a more sterically hindering nonleaving group than the simple NH3 may retard the kinetic reactivity of the trans geometry. Results from a number of laboratories now show that use of sterically hindered ligands in the trans geometry produces antitumor activity and preclinical activity distinct from c i s - D D E 99 T h e two most studied structural groups are trans-[PtC12(L)(L')], where L - L ' - p y r i d i n e or thiazole, or L = planar ligand such as pyridine, quinoline, or thiazole and L' NH 3 (Figure 8). The features of cytotoxicity of trans planar amine (TPA) complexes in both murine and human tumor cell lines include cytotoxicity at least equivalent to their corresponding cis-isomers and, indeed, equivalent to cis-DDP itself. Retention of activity in cell lines rendered resistant to cis-DDP and a different pattern of

Cl~ /N ~ /

Pt

t rans- [PtCl2(pyddine)2]

CI~/NH3 N/ ~CI H2 t,t,t.-[PtCI2(OH)2(NHz)(cha)]

CI~ /NI'I3 Pt

c N / \CI trans-[PtCl2(M4z)(quinoline)] Me ~C/OMe // CI~ / N~

H~ / P t ~ c I H

MeO,..--/C ~/N Me trans-[PtCI2(iminoether)2]

Figure 8. Structures of trans-platinum complexes displaying antitumor activity.

194

NICHOLAS FARRELL

cytotoxicity over a range of cell lines in comparison to that of cis-DDP is observed. 5~176176The activity of the trans-platinum geometry has been confirmed by studies on t,t,t-[PtC12(OH)2(NH3)(cyclohexylamine)] l~ and on [transPtC12{ HN=CRR')2}2] (R = Me, R ' = OMe) 1~ (Figure 8). These results are of particular interest because the trans-platinum complexes are completely or partially non-cross-resistant to cis-DDP in all cell lines we have tested. Factors contributing to cisplatin resistance include drug accumulation, selective repair of Pt-DNA adducts, and modulation of the Pt-DNA adduct level by intracellular thiols such as glutathione. The cellular pharmacology of the new TPA complexes is significantly different to that of cis-DDP with respect to all important parameters of cytotoxicity and cellular resistance.

C. DNA Strand Breakage is a Consequence of trans-Platinum Damage in Cells An unparalleled finding in platinum antitumor chemistry has been that the predominant effect of TPA complexes upon cellular DNA is the induction of protein-associated breaks. 1~ Interstrand cross-links are also formed and the relative balance of strand breakage and interstrand cross-linking is dependent on structure. Strand breakage is not observed with either cis- or trans-DDP. In contrast, elution of DNA from trans-[PtCla(py)2]-treated cells indicate the presence of DNA strand breaks beyond those introduced by radiation. Further studies have shown that the DNA strand breaks induced by trans-[PtCl2(py)2] are protein-associated, occur in the absence of irradiation, and are observable at early time points. These effects are similar to those seen with filter elution studies of topoisomerase inhibitors such as etoposide, Adriamycin, m-AMSA, and camptothecin. The phenomenon of proteinassociated DNA strand breaks is general--the order is trans-[PtCl2(NH3)(quin)] < trans-[PtCl2(NH3)(thiazole)] < trans-[PtCl2(py)2]. The corresponding cis-isomers produce a significantly smaller amount of strand breaks.

D. DNA Binding of trans-Platinum Complexes with Planar Amines The relative proportion of strand breakage and interstrand cross-linking is dependent on the exact structure of the trans-plafinum complex. The mechanism of induction of the protein-associated strand breaks indicates, at this point, a pattern similar to that of topoisomerase inhibitors. Other candidate proteins would be helicases and those involved in DNA repair. The production of transient strand breaks by DNA processing enzymes suggests a plausible mode of attack for drugs which, in the case of topoisomerase poisons, may stabilize the drug-protein complex. If a monofunctional adduct on DNA is sufficiently long-lived, then a competition is set up between DNA interstrand cross-linking and formation of ternary DNA-protein cross-links, as illustrated in Figure 9. Both kinetic and structural features of platinum complexes can be systematically modified to dictate one pathway over the other. The observation of protein-associated strand breakage


195

I! / _T P A

Figure 9. Competition between formation of an interstrand cross-link and a DNAprotein cross-link upon initial formation of a monofunctional adduct of a trans-platinum complex. TPA stands for trans-planaramine.

from trans-isomers, and to a lesser extent from cis-isomers containing planar amines, is consistent with a steric effect retarding closure to a bifunctional DNADNA intrastrand or interstrand cross-link. In studies using DNA from various sources (CT, plasmid, or defined oligonucleotide sequence) interstrand cross-linking is considerably enhanced by trans[PtC12(py)2] over both its cis-isomers. 56' 101 At 10 ILtM (rb = 0.005) interstrand cross-links represent 0.23 or 23% of the total lesions (c.f. cis-DDP with < 5% interstrand cross-links). Modeling indicates that the GG "cis-DDP-like" interstrand cross-link in d(GC) sequences is favored due to the steric constraints of the planar ligands. Thus, the steric effects favor the GG interstrand cross-link over the GC cross-link favored by trans-platinum, as shown in Figure 10. The presence of planar rings has a considerable effect on DNA binding properties in comparison to the [PtCI2(NH3)2] isomers. Local unwinding of the DNA helix, as measured by topological unwinding of supercoiled plasmids, is significantly more effective for trans-[PtCl2(py) 2] than trans-DDP, with unwinding angles of

196

NICHOLAS FARRELL

H2N "NH'#O 0

1-12N Nil O NH3NH2

Figure 10. Schematic diagrams of coordination spheres of interstrand cross-links formed from trans-[PtCi2(py)2] (a GC/CG cross-link) and trans-[PtCi2(NH3)2] (a crosslink between G and C of the same base pair). 17 ~ and 9 ~ respectively. 99 The sequence specificity for trans-platinum complexes with planar ligands is for alternating (GC) sites. The adducts from trans[PtC12(NH3)(quinoline)] and trans-[PtCl2(py) 2] are not recognized by HMG proteins, which may indicate little or no bending of the DNA. 56 In summary, the structural and kinetic effects of substitution of a planar ligand for NH 3 in the trans-platinum geometry leads to an array of structurally distinct Pt-DNA adducts, which may be responsible for different pathways of protein recognition (such as proteins involved in strand breakage) in comparison to that delineated from cis-DDP.

VI. SUMMARY B ifunctional binding of platinum complexes to DNA is considered essential for manifestation of antitumor activity. Studies on direct analogues of cis-DDP confirm that all cis-[PtX2(amine) 2] complexes will produce a similar array of adducts with similar consequences. This fact may well be a reason for the similarity in clinical profile of those analogues which have entered clinical trials. On the other hand, work on dinuclear platinum and trans-platinum complexes indicate that complexes structurally different to cis-DDP and acting on DNA in a manner distinct to the clinically used agent may display an altered spectrum of antitumor activity. A rich array of new structures has been discovered worthy of further exploration to develop new platinum complexes genuinely complementary to the currently used drugs. 109

ACKNOWLEDGMENTS I thank my many co-workers for the results presented. This work is supported by The American Cancer Society (Grant DHP-2D) and Boehringer Mannheim Italia.


197

REFERENCES 1. Sherman, S. E.; Lippard, S. J. Chem. Rev. 1987, 87, 1153. 2. Reedijk, J.; Fichtinger-Schepman, A. M. J.; van Oosterom, A. T.; van de Putte, P. Structure and Bonding 1987, 67, 53. 3. Lepre, C. A.; Lippard, S. J. In Nucleic Acids and Molecular Biology; Eckstein, E; Lilley, D. M. J., Eds.; Springer-Verlag: Berlin, 1990, Chapter 2. 4. Comess, K. M.; Lippard, S. J. In Molecular Aspects of Anticancer Drug-DNA Interactions; Neidles, S.; Waring, M., Eds.; CRC Press: Boca Raton, 1993, Chapter 5. 5. Reed, E.; Kohn, K. W. In Cancer Chemotherapy- Principles and Practice; Chabner, B. A.; Collins, J., Eds.; J.B. Lippincott: Philadelphia, 1990, pp. 465--490. 6. Christian, M. C. Sem. Oncol. 1992, 19, 720. 7. Eisenhauer, E.; Swerton, K.; Sturgeon, J.; Fine, S.; O'Reilly, S.; Canetta, R. In Carboplatin: Current Perspectives and Future Directions; Bunn, P.; Canetta, R.; Ozols, R.; Rozencweig, M., Eds.; W.B. Saunders: Philadelphia, pp. 133-140. 8. Dorr, R. T.; Noel, K. Prin. Prac. Gynecol. OncoL Updates 1993, 1, 1. 9. Bumouf, D.; Daune, M.; Fuchs, R. P. P. Proc. Natl. Acad. Sci. USA 1987, 84, 3758. 10. Naser, L. J.; Pinto, A. L.; Lippard, S. J.; Essigman, J. M. Biochemistry 1998, 27, 4357. 11. 5;herman, S. E.; Gibson, D.; Wang, A. H.-J.; Lippard, S. J. Science 1998, 230, 412. 12. Coil, M.; Sherman, S. E.; Gibson, D.; Wang, A. H.-J." Lippard, S. J. J. Biomol. Struc. Dyn. 1990, 8,315. 13. den Hartog J.H.J.; Altona, C.; Chottard, J. C.; Girault, J. P.; Chottard, G.; Lallemand, J. Y.; Mansuy, D.; de Leeuw, E A. A. M.; Marcelis, A. T. M.; Reedijk, J. Nucleic Acids Res. 1982, 10, 4715-4730. 14. Bloemink, M. J.; Reedijk, J. Metal Ions Biol. Sys. 1996, 32, 641. 15. Bellon, S. E; Coleman, J. H.; Lippard, S. J. Biochemistry 1991, 30, 8026. 16. Takahara, P. M.; Rosenzweig, A. C.; Frederick, C. A.; Lippard, S. J. Nature 1995, 377, 649. 17. Yang, D.; van Boom, S. S. G. E.; Reedijk, J.; van Boom, J. H.; Wang, A. H.-J. Biochemistry 1995, 34, 12912. 18. Schwartz, A.; Sip, M.; Leng, M. J. Am. Chem. Soc. 1990, 112, 3673. 19. Lemaire, M.-A.; Schwartz, A.; Rahmouni, A. R.; Leng M. Proc. Natl. Acad. Sci. USA 1991, 88, 1982. 20. Rahmouni, A.; Leng, M. Biochemistry 1987, 26, 7229. 21. Hopkins, P. B.; Milliard, J. T.; Woo, J.; Weidner, M. E; Kirchner, J. J.; Sigurdsson, S. Th.; Raucher, S. Tetrahedron 1991, 47, 2475. 22. Sip, M.; Schwartz, A.; Vovelle, F.; Ptak, M.; Leng, M. Biochemistry 1992, 31, 2508. 23. Huang, H.; Zhu, L.; Reid, B. R.; Drobny, G. P.; Hopkins, P. B. Science 1995, 270, 1842. 24. van Garderen, C. J.; van Houte, L. P. A.; van den Elst, H.; van Boom, J. H.; Reedijk, J. J. Amer. Chem. Soc. 1989, 111, 4123. 25. van Garderen, C. J.; Altona, C.; Reedijk, J. Inorg. Chem. 1990, 29, 1481. 26. Brabec, V.; Reedijk, J.; Leng, M. Biochemistry 1992, 31, 12397. 27. ViUani, G.; Hubscher, U.; Butour, J.-L. Nucleic Acids Res. 1988, 16, 4407. 28. Pinto, A. I.; Lippard, S. J. Proc. Natl. Acad. Sci. USA 1985, 82, 4616. 29. Gralla, J. D.; Sasse-Dwight, S.; Poljak, L. G. Cancer Res. 1987, 47, 5092. 30. Bissett, D.; McLaughlin, K.; Kelland, L. R.; Brown, R. Br. J. Cancer 1993, 67, 742. 31. Chu, G.; Chang, E. Proc. Natl. Acad. Sci. USA 1990, 87, 3324. 32. Billings, P. C.; Davis, R. J.; Engelsberg, B. N.; Skov, K. A.; Hughes, E. N. Biochem. Biophy. Res. Comm. 1992, 188, 1286. 33. Brown, S. J.; Kellett, P. J.; Lippard, S. J. Science 1993, 261,603. 34. Pil, P. M.; Lippard, S. J. Science 1992, 256, 234. 35. Whitehead, J. P.; Lippard, S. J. Metal Ions Biolog. Syst. 1996, 32, 687. 36. Lilley, D. M. J. Nature 1992, 357, 282.

198

NICHOLAS FARRELL

37. Chu, G. J. Biol. Chem. 1994, 269, 787. 38. Kasparkova, J.; Brabec, V. 7th International Symposium on Platinum and Other Metal Compounds in Cancer Chemotherapy; Amsterdam, 1995, Abstract S098. 39. Mello, J. A.; Acharya, S.; Fishel, R.; Essigmann, J. Chem. Biol. 1996, 3, 579. 40. Knox, R. J.; Friedlos, F.; Lydall, D. A.; Roberts, J. J. Cancer Res. 1986, 46, 1972. 41. Micetich, K. C.; Barnes, D.; Erickson, L. C. Cancer Res. 1985, 45, 4043. 42. Frey, U.; Ranford, J. D.; Sadler, P. J. Inorg. Chem. 1993, 32,1333. 43. Canovese, L.; Cattalini, L.; Chessa, G.; Tobe, M. L. J. Chem. Soc., Dalton Trans. 1988, 2135. 44. Burchenal, J. H.; Kalaher, K.; Dew, K.; Lokys, L.; Gale, G. Biochimie 1978, 60, 961. 45. Kidani, Y.; Inagaki, K. J. Med. Chem. 1978, 21, 1315. 46. Boudny, V.; Vrana, O.; Gaucheron, F.; Kleinwachter, V.; Leng, M.; Brabec, V. Nucleic Acids Res. 1992, 20, 267. 47. Jennerwein, M.; Eastman, A.; Khokhar, A. Chem.-Biol. Inter. 1989, 70, 39. 48. Inagaki, K.; Nakahra, H.; Alink, M.; Kidani, Y. lnorg. Chem. 1990, 29, 4496. 49. Page, J. D.; Husain, I.; Sancar, A.; Chaney, S. G. Biochemistry 1990, 29, 1016. 50. Akaboshi, M.; Kawai, K.; Ujeno, Y.; Takada, S.; Miyahara, T. Jpn. J. Cancer Res. 1994, 85, 106. 51. Iwata, M.; Izuta, S.; Suzuki, M.; Kojima, K.; Furuhashi, Y.; Tomoda, Y.; Yoshida, S. Jpn. J. Cancer Res. 1994, 82, 433. 52. Morikawa, K.; Matsumotot, T.; Matsuoka, Y.; Koizumi, K. Chem. Pharm. Bull. 1992, 40, 21. 53. Kelland, L. R.; Mistry, P.; Abel, G.; Loh, S. Y.; O'Neill, C. E; Murrer, B. A.; Harrap, K. R. Cancer Res. 1992, 52, 3857. 54. Hartwig, J. E; Lippard, S. J. J. Am. Chem. Soc. 1992, 114, 5646. 55. Farrell, N. Cancer Invest. 1993, 11,578. 56. Farrell, N. Metal hms Biolog. Syst. 1996, 32, 603. 57. Farrell, N. Advances in DNA Sequence Spec~[ic Agents. Hurley, L. H.; Chaires, J. B., Eds.; JAI Press: Greenwich, CT, 1996, Vol. 2, p. 187. 58. Farrell, N. Comments Inorg. Chem. 1995, 16, 373. 59. Hoeschele, J. D.; Kraker, A. J.; Qu, Y.; Van Houten, B.; Farrell, N. In Molecular Ba~visof Specificity in Nucleic Acid-Drug Interactions; Pullman, B.; Jortner, J., Eds.; Kluwer Academic Press: Dordrecht, 1990, Vol. 23, p. 301 ff. 60. Farrell, N.; Qu, Y.; Hacker, M. E J. Med. Chem. 1990, 33, 2179. 61. Manzotti, C.; Pezzoni, G.; Giuliani, E; Valsecchi, M.; Farrell, N.; Tognella, S. Proc. A.A.C.R. 1990, 35, 2628. 62. Mellish, K. J.; Qu, Y., Scarsdale, N.; Farrell, N. Nucleic Acids Res. 1997, 25, 1265. 63. Kharatishvili, M.; Mathieson, M.; Farrell, N. Inorg. Chim. Acta 1997, 255, I. 64. Roberts, J. D.; van Houten, B.; Qu, Y.; Farrell, N. E Nucleic Acids Res. 17, 9719. 65. Van Houten, B.; IUenye, S.; Qu, Y.; Farrell, N. Biochemistry 1993, 32, 11794. 66. Johnson, A. L.; Illenye, S.; Farrell, N.; Van Houten, B. Proc. A.A.C.R. 1994, 35, 2634. 67. Lehnert, S.; Farrell, N. Proc. A.A.C.R. 1995, 36, 2376. 68. Farrell, N.; Roberts, J. D.; Qu, Y.; Zou, Y.; Marples, B.; Skov, K. A.; Tognella, S. Proc. A.A.C.R. 1994, 35, 2637. 69. Farrell, N.; Appleton, T. G.; Qu, Y.; Roberts, J. D.; Soares Fontes, A. E; Skov, K. A.; Wu, E; Zou, Y. Biochemistry 1995, 34, 15480. 70. Zhen, W.; Link Jr., C. J.; O'Connor, E M.; Reed, E.; Parker, R.; Howell, S. B.; Bohr, V. A. Mol. Cell Biol. 1992, 12, 3689. 71. Jones, J. C.; Zhen, W.; Reed, E.; Parker, R. J.; Sancar, A.; Bohr, V. A. J. Biol. Chem. 1991, 266, 7101. 72. Farrell, N.; Qu, Y.; Feng, L.; Van Houten, B. Biochemistry 1990, 29, 9522. 73. Johnson, A.; Qu, Y.; Van Houten, B.; Farrell, N. Nucleic Acids Res. 1992, 20, 1697. 74. Zou, Y.; Van Houten, B.; Farrell, N. Biochemistry 1994, 33, 5404. 75. Gruff, E. S.; Orgel, L.E. Nucleic Acids Res. 1991, 19, 6849.


199

76. Yang, D.; van Boom, S.; Reedijk, J.; van Boom, J.; Farrell, N.; Wang, A. H.-J. Nature Struc. Biol. 1995, 2, 577. 77. Iwamoto, M.; Mukunda, S. J.; Marzilli, L. G. J. Am. Chem. Soc. 1994, 116, 6238. 78. Rich, A.; Nordheim, A.; Wang, A. H.-J. Ann. Rev. Biochem. 1984, 53, 791. 79. Malfoy, B.; Hartmann, B.; Leng, M. Nucleic Acids Res. 1981, 9, 5659. 80. Ushay, M.; Santella, R. M.; Grunberger, D.; Lippard, S. J. Nucleic Acids Res. 1982, 10, 3573. 81. Kharatishvili, M.; Qu, Y.; Farrell, N. Biochemistry, submitted. 82. Liu, L. E; Wang, J. C. Proc. Natl. Acad. Sci. USA 1987, 84, 7024. 83. Rich, A. In Proceedings of The Robert A. Welch Foundation; Houston, 1993, Vol. 37, pp. 13-34. 84. Parkinson, G. N.; Arvanitis, G. M.; Lessinger, L.; Ginell, S. I.; Jones, R.; Gaffney, B.; Berman, H. Biochemistry 1995, 34, 2960. 85. Bloemink, M. J.; Reedijk, J.; Farrell, N.; Qu, Y.; Stetsenko, A. I. J. Chem. Soc., Chem. Commun. 1992, 1002. 86. Qu, Y.; Bloemink, M. J.; Reedijk, J.; Hambley, T. W.; Farrell, N. J. Am. Chem. Soc. 1992, 118, 9307. 87. Syzmkowski, D. E.; Yarema, K.; Essigmann, J. M.; Lippard, S. J.; Wood, R. D. Proc. Natl. Acad. Sci. USA 1992, 89, 10772. 88. Richon, W. M.; Schulte, N. A.; Eastman, A. Cancer Res. 1987, 47, 2056. 89. Kasparkova, J.; Mellish, K. J.; Qu, Y.; Brabec, V.; Farrell, N. Biochemistry 1996, 35, 16705. 90. Rosenberg, B. In Nucleic Acid - Metal hm Interactions; Spiro, T.G. Ed.; Wiley: New York, 1980, Vol. 1, pp. 1-29. 91. Hoeschele, J. D.; Butler, T. A.; Roberts, J. A. Inorganic Chemistry in Biology and Medicine; ACS Symposium Series; Martell, A. E., Ed.; 1980, Vol. 140, p. 181ff. 92. Eastman, A.; Barry, M. A. Biochemistry 1987, 26, 3303. 93. Bancroft, D. P.; Lepre, C. A.; Lippard, S. J. J. Am. Chem. Soc. 1990, 112, 6860. 94. Roberts, J. J.; Friedlos, F. Cancer Res. 1987, 47, 31. 95. Ciccarelli, R. B.; Solomon, M. J.; Varshavsky, A.; Lippard, S. J. Biochemistry 1985, 24, 7533. 96. Bellon, S. E; Lippard, S. J. Biophys. Chem. 1990, 35, 179. 97. Lepre, C. A.; Chassot, L.; Costello, C. E.; Lippard, S. J. Biochemistry 1990, 29, 811. 98. Brabec, V.; Leng, M. Proc. Natl. Acad. Sci. USA 1993, 90, 5345. 99. Brabec, V.; Sip, M.; Leng, M. Biochemistry 1993, 32, 11676. I00. Van Beusichem, M.; Farrell, N. lnorg. Chem. 1992, 31,634. 101. Farrell, N.; Kelland, L. R.; Roberts, J. D.; Van Beusichem, M. Cancer Res. 1992, 52, 5065. 102. Kelland, L. R.; Bamard, C. F. J.; Mellish, K. J.; Jones, M.; Goddard, P. M.; Valenti, M.; Bryant, A.; Murrer, B. A.; Harrap, K. R. Cancer Res. 1994, 54, 5618. 103. Coluccia, M.; Nassi, A.; Loseto, F.; Boccarelli, A.; Mariggio, M. A.; Giordano, D.; Intini, E P.; Caputo, P.; Natile, G. J. Med. Chem. 1993, 36, 510.


KEDARCI DI N AN D M A D U ROPEPTI N" TWO NOVEL ANTITU MOR CH ROMOPROTEI NS WITH SELECTIVE PROTEASE ACTIVITY AND DNA CLEAVING PROPERTIES

Nada Zein and Daniel R. Schroeder

I~ II.

III.

IV.

Introduction

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

202

Kedarcidin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

202

A.

Kedarcidin: The Holoantibiotic

. . . . . . . . . . . . . . . . . . . . . .

202

B.

Kedarcidin: The C h r o m o p h o r e . . . . . . . . . . . . . . . . . . . . . . .

203

C.

Kedarcidin: The Apoprotein

Maduropeptin

. . . . . . . . . . . . . . . . . . . . . . . .

209

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

214

A.

Maduropeptin: The Holoantibiotic . . . . . . . . . . . . . . . . . . . . .

214

B. C.

The C h r o m o p h o r e and Solvent Artifacts . . . . . . . . . . . . . . . . . . Maduropeptin: The Apoprotein . . . . . . . . . . . . . . . . . . . . . . .

215 220

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

223

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

223


201

202

NADA ZEIN and DANIEL R. SCHROEDER I.

INTRODUCTION

Enediyne antitumor antibiotics represent a novel chemical class of compounds with an outstanding array of biological and chemical properties. In vitro studies on various cell lines and in vivo testing on murine solid tumors and human xenograft models demonstrate that the enediynes are the most potent antitumor agents ever isolated. As an example, C1027 is 107 times more potent than adriamycin, one of the most clinically effective antitumor drugs used to date. The enediynes contain two acetylenic groups conjugated to a double bond or incipient double bond within a 9- or 10-membered ring. This diyne--ene unit is often referred to as the warhead. Upon proper activation in the presence of DNA, the warhead undergoes rearrangement yielding sp 2 carbon-centered diradicals that cleave DNA. This radical-mediated DNA damaging capability is believed to account for the potent antitumor activity of the enediynes. 1 To date, the 9-membered enediyne antibiotics have been isolated as a complex with an acidic apoprotein; they are referred to as the chromoproteins and include auromomycin, actinoxanthin, neocarzinostatin, kedarcidin, maduropeptin, and C 1027. 2-18 This is in contrast to esperamicin, calicheamicin, and dynemicin, agents that contain 10-membered enediyne rings and, to date, have been isolated without associated apoproteins. 19-22 The chemistry, biology, and the mechanism of action of the 10-membered enediyne-containing antitumor agents have been extensively discussed in several recent review articles. 23 This chapter will focus on the latest studies conducted on kedarcidin and maduropeptin, the two 9-membered enediyne containing chromoproteins recently isolated and characterized at Bristol-Myers Squibb. 9-14 For a thorough review on neocarzinostatin, refer to the work of Goldberg and coworkers who paved the way for many of the mechanistic studies performed on the enediyne chromophores. 24

II.

KEDARCIDIN

Kedarcidin chromoprotein was isolated from the culture supernatant of an actinomycete strain obtained from soils collected in the Maharastra State in India. This compound exhibits potent in vivo antitumor activity against P388 leukemia and B 16 melanoma in murine models at an optimal dose of 3.3 and 2~tg/kg, respectively. The potency of kedarcidin against murine tumors is similar to that of calicheamicin and esperamicin. Kedarcidin also shows potent activity against Gram-positive bacteria but no activity against Gram-negative bacteria. Exposure of H C T l l 6 human colon cancer cell lines to kedarcidin results in an IC50 value of 10-9 MmlCso is the drug concentration that causes 50% cell kill. 9'1~

A. Kedarcidin: The Holoantibiotic Kedarcidin is an acidic complex (pI 3.65) with an apparent molecular mass of 12,400 Da as determined by SDS-PAGE. It consists of an apoprotein and a highly

Kedarcidin and Maduropeptin

203

ASAAVSVSPA

TGLADGATVT

VSASGFATST

SATALQCAIL

ADGRGACNVA

EFHDFSLSGG

EGTTSVVVRR

SFTGYVMPDG

PEVGAVDCDT

APGGCQIVVG

GNTGEYGNAA

I SFG

i 14

Sequence 1. Proposed amino acid sequence of the kedarcidin apoprotein. labile, non-peptidic chromophore. The major apoprotein is a single-chain polypeptide consisting of 114 amino acid residues (Sequence 1).l~ The chromophore is a solvent-extractable molecule having a molecular weight of 1029. ll'12 Depending on fermentation conditions, the kedarcidin complex varies with respect to amino acid composition in the apoprotein. The molecular formulae of the chromophores and the ratio of apoprotein to chI:omophore are as follows: Amino acid sequencing studies have revealed the presence of three variants of the kedarcidin apoprotein. The major variant polypeptide consists of 114 amino acid residues and is the one used in our studies (Sequence 1). Two minor variants were also identified; one lacks the first alanine of the major variant and the second lacks the first two amino acids of the major variant. HPLC-UV and LC-MS showed the presence of three different chromophores of varying ratios and molecular weights. The major chromophore 1 (see Scheme 1) of molecular weight 1029 has been fully characterized and is the subject of our discussion in this part of the chapter. The ratio of apoprotein to chromophore in the complex varied from 1" 1 to 18:1 as shown by a quantitative UV method of analysis. The complex, chromophore and apoprotein used in our studies, came from a lot that had a 16:1 ratio of apoprotein to chromophore. Like neocarzinostatin, kedarcidin chromophore is believed to reside in a hydrophobic cleft in the apoprotein where it is shielded from degradation. This situation keeps the otherwise insoluble chromophore in aqueous solutions and may aid in exporting it out of the producing organism. The three-dimensional structure of the kedarcidin holoantibiotic has not yet been determined; however recent NMR studies on the apokedarcidin revealed that apokedarcidin adopts an overall structure similar to that elucidated for neocarzinostatin, macromomycin, and apoactinoxanthin. 25-3~ This structure consists of a seven-stranded antiparallel [3-barrel domain linked to a subdomain composed of two 13-hairpin ribbons. Interestingly, this motif resembles the variable domain of immunoglobulins. 31

B. Kedarcidin:The Chromophore Chemical Structure and Activation Initial attempts to characterize the intact kedarcidin chromophore 1 were unsuccessful due to its low stability in most organic solvents. Considerable substructural information was obtained by acidic methanolysis experiments from which the

204

NADA ZEIN and DANIEL R. SCHROEDER

structures of the two carbohydrate residues and the bridging azatyrosine residue could be deduced. Reduction of the crude chromophore fraction followed by isolation of the two major products gave invaluable information on the structure of the enediyne core and the attachment of the carbohydrate and bridging amino acid residues as well as important clues regarding its activation. 11'12 Close examination of the structure shows that the functionalities on the aglycone are all strategically situated with respect to each other to participate in the formation of a transient 1,4 diyl by a chemically induced Bergman-type bond reorganization. 32 The cycloaromatization is initiated by attack of a reducing thiol on the less-hindered side of C- 12 followed by bond migration and opening of the epoxide ring, as shown in Scheme 1. Subsequent cyclization of the enediyne gives the 1,4-benzenoid diradical, 3, the species believed to cause DNA damage. The opening of the epoxide reduces the strain energy developed in the transition state leading to cycloaromatization. This mechanism of action contrasts with that of the neocarzinostatin chromophore, in which the epoxide ring opening generates a cumulene intermediate that aromatizes to a C-2/C-6 indacene diradical. 8

Cleavage of Plasmid DNA Agarose gel electrophoresis studies demonstrate that overnight incubation of the kedarcidin chromophore at 37 ~ with supercoiled pM2 DNA (Form I) in the presence of ~3-mercaptoethanol results in the conversion of Form I DNA to mostly Form II (i.e. open circular DNA). A plot of the consumption of Form I and the formation of Forms II and III (i.e. linear DNA), as a function of drug concentration, showed that at 37 ~ single-strand breaks accumulate on average four times more rapidly than double-strand breaks, consistent with the behavior of a single-strand cutter. Treatment of the cleavage products with hydrazine and putrescine had no effect on the results suggesting absence of alkali-labile lesions. A time-dependent study at one concentration of chromophore showed that the chromophore cleaved the DNA within 5 minutes of incubation time. Omitting 13-mercaptoethanol from the reaction mixture greatly decreased the DNA cleavage, consistent with the proposed mechanism for the formation of the diradical active intermediate 3 (Scheme 1). A temperature-dependent study at 4 and 37 ~ showed the rate of accumulation of both forms to be about 5 times more rapid at 37 than at 4 ~ A comparative study between the cleavage of various forms of OX174 indicated that the kedarcidin chromophore does not recognize the single-strand form and that it favors supercoiled DNA over the open circular and the linearized forms. 33

DNA Cleavage Site, Sequence Specifidty, and Cleavage Chemistry 5"-End-Labeled Studies. Cleavage specificity was examined by the reaction of the kedarcidin chromophore with five different 5'-end-labeled pBR322 and pUC 18 restriction fragments in presence of 13-mercaptoethanol. Comparison of the electrophoretic mobility of the drug-induced cleavage products with the Maxam-

205

Kedarcidin and Maduropeptin CH3y C H ~ O ~ O H C H , O ~ OCH,

I

CI~ ..,N.. ~

. IH

0

ArCO

CH, .... 'N'CH'

o~

I II "I r''H

Cl ~,~N ~ , / , , ~ . NH

OH

o"~~

~ pSH 0

sugar.O~"

OH

--

HO._U-7",,.~ CH3

ArCO

I

sugar-O w -

~

~"~" OH |

g

---

ArCO

I II "I r""H O.,"';',,~ O,,'~M.O,,,',,,~,,\O"suga, ~::",:, H \ "~~ H

Cl ~,~N ~ , , , , ~ . . NH

"

o~ o ' ' ' x '~~

"H

DNA

.~r'~"

sugar-O

.,-:,o.

,...,

H4' "OH

3

....

Scheme 1.

Gilbert chemically produced markers 34 shows that the DNA cleavage caused by the chromophore is highly sequence-specific. Also, as observed from the agarose gel studies, the kedarcidin chromophore cleaves DNA in a single-stranded manner. A preferred site is the 3' nucleotide (N = T, C, G, A) adjacent to the TCCT tetramer, as seen in Figure 1. Another favored site is TCGTN, where N is a C or a G. The 5-mer TCATN is less preferred than TCCTN (

)

Q

DNA I

,

./"

-...~.,~ m-RNA

Protein

.-H~N o9

~

IT N

"

N~C. 1'

. . . . N.~" , , , , ~~_~l N

H'" O

H~NtH.....

~C-1'

I C N"O'" "H'~N~H Figure 1. (A) General principle of antisense inhibition of protein expression: Binding of an antisense oligonucleotide to a complementary region on a target mRNA suppresses translation and thus prevents synthesis of the corresponding target protein. (Figure adapted from Moser, H. In Perspectives in Medicinal Chemistry; Testa, B.; Kyburz, E.; Fuhrer, W.; Giger, R.; Eds.; Verlag Helvetica Chimica Acta: 1993, p. 279.) (B) Recognition of nucleic acids by Watson-Crick hydrogen-bonding: A(denosine) pairs with T(hymidine) and G(uanosine) pairs with C(ytidine).

Antisense- and Antigene-BasedDrug Design

231

among the strongest known in biological chemistry) (Figure 1B), 12 in theory the use of AS-ODNs should allow for the elimination of any given protein from the cell in a very specific manner as long as the mRNA sequence for the target protein is known and the target base sequence on this RNA is unique for the entire RNA population of the cell. On a purely statistical basis, any particular RNA sequence longer than 18 bases can be expected to occur only once throughout the entire human genome, and oligonucleotides containing at least 18 nucleotide units should therefore suffice to achieve specificity. 1 In practical terms, very high specificity might even be achievable with shorter oligonucleotides (as long as they display high enough RNA binding affinity), since large parts of a mRNA molecule are rendered inaccessible for binding of an AS-ODN due to extensive secondary structure formation. 1-8 The potential for exquisite specificity in the inhibition of protein expression (and thus protein function) represents the most attractive feature of antisense-based drug development strategies and constitutes the most important single advantage of the antisense approach over traditional medicinal chemistry approaches directed towards the inhibition of protein function. Tight complexes between nucleic acids may not only be formed by interactions between single-stranded DNA or RNA molecules via Watson-Crick base pairing, but under certain circumstances oligonucleotides can also bind to double-stranded DNA, i.e. Watson-Crick double helices, in a sequence-specific fashion, leading to the formation of triple helices or triplexes (Figure 2A). 4'13-15For the most important pyr-pur-pyr structural motif binding of a third strand requires one strand of the DNA/DNA duplex to incorporate a large majority of purine nucleotides, and recognition occurs via Hoogsten base pairing between T's and (protonated) C's in the third strand to A's and G's, respectively, in the polypurine strand of the double helical component (Figure 2B). The binding of the polypyrimidine third strand occurs parallel to the polypurine strand of the double helix (Figure 2A). The practical applicability of the triplex approach for drug design is currently still limited by the requirement for polypurine tracts within genomic DNA regions, but examples do exist in the literature where triplex-forming oligonucleotides have been used to specifically block transcription in vitro;4"16-23furthermore, attempts have been made to extend the triplex alphabet by the design of non-natural nucleic acid bases that may eventually allow the design of oligonucleotides capable of binding to DNA/DNA double helices of any given base sequence. 24-28

B. Mechanism of Action of Antisense Oligonucleotides There are two different basic mechanisms by which the binding of an AS-ODN to a target mRNA can lead to inhibition of protein expression, namely physical blockage of a variety of processes involved in the sequence of events leading from DNA to protein, or RNAse H-mediated RNA degradation (Figure 3). 1-7 In the former case, binding of the AS-ODN, e.g., in the vicinity of the AUG start codon,

232

KARL-HEINZALTMANN, DORIANO FABBRO, and THOMAS GEIGER

WATSON-CRICK HYDROGENBOND DOUBLE-S~F.D DNA

MAJORGROOVE

t. .::::,::-" 5" 3'

,,C-1' 9 9

H

H-~I~

H

O

H

H

.

"

"0

O""

I ~ NI TT ~ O~H .....- ',,,~N (~-1'

~C-I'

" ~

H~ ~H"'"" ~ N ~ c ' I ' N

~CZ ~,

,.I~G I

N"'"

C-1'

0-

I

...."H~N~H

Figure 2. (A) Schematic representation of triple helix formation. The polypyrimidine third strand is bound in the major groove of double-stranded DNA via Hoogsteen base pairing. The orientation of the third strand is parallel to the poly-purine strand of the DNA/DNA duplex. (B) Recognition of AT and GC base pairs in double-stranded DNA by T and (protonated) C, respectively, via Hoogsteen hydrogen bonds.

blookedtranm~lptlon 5'Me

'ill I 3' ~ 5 ' blocked spiking

onzyrr~tlc 3'

enzymatic attack

I

attack by rlbozyme8 ~ ' ~

.oo

5'

os-

~~ffexprt attackby ribozymes

bindinglit CA~Por Inltlatlonsltea 5'Me ,

~I~NI..~ ~ N I - ~ 5,Me~ ~ ~ ~ r ~

&AAA

block ribosomeassemblyor initiation

dbosomemigrationblocked

2

NH2

I~roteinfunktion blodmd by binding of nuoleioacid

Figure 3. Possible points of interference of antisense and antigene oligonucleotides with protein expression. Note that the binding of antisense oligonucleotides to single-stranded mRNA may lead to suppression of protein synthesis either by physical blockage or by RNAse H-mediated degradation of bound RNA. (From Murray, J. A. H.; Crockett, N. In Antisense DNA and RNA; Murray, J. A. H., Ed.; Wiley-Lyss: 1992, p. 21.) 233

234

KARL-HEINZ ALTMANN, DORIANO FABBRO, and THOMAS GEIGER

may prevent binding of initiation factors or the assembly of ribosomal subunits by a steric effect. The latter mechanism, on the other hand, involves degradation of the RNA strand of the DNA/RNA duplex formed upon binding of the AS-ODN to its RNA target sequence by the endogenous RNAse H enzyme(s). It would appear that due to its irreversible and catalytic nature, RNA cleavage should be the most efficient mechanism of AS-ODN-mediated inhibition of protein expression and it has in fact been demonstrated in a variety of cases that potent inhibition of translation by AS-ODNs depends on RNAse H-mediated destruction of the target RNA. 29-36 It should be noted, however, that specific RNA degradation conceivably could also be induced by AS-ODNs equipped with an appropriate (synthetic) chemical cleaver moiety, and successful attempts to design such cleaver moieties have been recently reported in the literature (although no cell culture or in vivo data are yet available for such cleaver-oligonucleotide conjugates). 37'38 The activation of RNAse L (a single-stranded ribonuclease) by 2'-5'-oligoadenylates attached to an antisense oligonucleotide via an appropriate spacer group has been demonstrated to lead to RNA cleavage only in the vicinity of the bound oligonucleotide. 39'4~An antisense oligonucleotide of this type was also shown to be a potent inhibitor of the expression of the double-stranded RNA-dependent protein kinase PKR in HeLa cells, 4~ resulting in the unresponsiveness of these cells to activation of nuclear factor-~cB (NF-nB) by poly(I):Poly(C). Obviously, any inhibitory effect of triplex-forming oligonucleotides on gene transcription must depend on the physical blockage of transcription factor or RNA polymerase binding, or, alternatively, on the presence of an artificial (ds) DNA cleaving agent delivered with the antigene oligonucleotide. Although such modified oligonucleotides have been used in model systems to cleave double-stranded DNA, 14'41'42 experiments involving animals have not been reported for triplexforming oligonucleotides in conjunction with either type of mechanism.

C. Chemical Modifications of Antisense Oligonucleotides For the basic concepts outlined in the previous discussion to be transformed into a viable drug design and drug development strategy, it is clear that an antisense or antigene oligonucleotide has to fulfill a number of requirements; failure to meet any single one most likely will abolish, or at least severely limit, its therapeutic utility. 1. The oligonucleotide has to exhibit a reasonable metabolic stability in a physiological environment. 2. It has to bind to its target RNA (or double-stranded DNA) with high affinity and specificity. 3. It must be taken up by the target cells and tissues at a reasonable rate and to a reasonable extent. 4. It has to show appropriate pharmacokinetic and pharmacodynamic behavior.


235

Natural oligonucleotides (i.e. short pieces of single-stranded DNA based on a phosphodiester backbone) are very rapidly degraded under physiological conditions by the action of single-stranded nucleases (primarily 3'-exonucleases) and therefore are not suitable for applications in vivo. ~-7 It is for this reason that extensive efforts have been made over the past few years to design and synthesize chemically modified oligonucleotides or oligonucleotide analogues which would exhibit improved nuclease resistance, but retain the ability to bind to complementary RNA with high affinity and specificity (Figure 4). 43--46 From these studies a number of sugar, base, and backbone modifications have emerged, some of which dramatically increase the stability of oligonucleotides against nucleolytic degradation and/or make them tighter binders to RNA than natural DNA itself. It should be noted that the affinity of oligonucleotides for complementary RNA and especially for double-stranded DNA may be improved by simple conjugation with intercalating groups, an approach that has been extensively followed by H61~ne and coworkers in the design of antigene oligonucleotides. 4'47 On the other hand, several affinity-enhancing modifications of the sugar or base moiety do not lead to improved nuclease resistance, and their favorable binding properties can only be exploited in combination with stabilizing backbone modifications (e.g. a phosphorothioate backbone). 33'48 One of the major shortcomings of almost all the chemical modifications of oligonucleotides reported to date is their inability to activate RNAse H, i.e. the corresponding DNA/RNA duplexes do not serve as RNAse H substrates. 1-7'29'31-33 In fact, so far phosphorothioates are the only class of modified oligonucleotides with very favorable nuclease resistance, sufficiently tight binding to RNA, and the ability to elicit RNAse H cleavage of bound RNA. However, strategies have been devised to overcome this problem of lack of RNAse H activation; these involve the design of chimeric oligonucleotides containing non-RNAse H activating (but possibly RNA binding affinity enhancing) modifications only in the terminal parts of the sequence ("wings"), while the central part ("gap") is based on a (2'-deoxyribose) phosphorothioate backbone. 32'49'5~The part of the bound RNA located across the 2'-deoxy phosphorothioate section of the AS-ODN in the oligonucleotide/RNA duplex is then still susceptible to RNAse H cleavage. The feasibility of this strategy has been impressively demonstrated by Monia et al. in in vitro studies on the inhibition of the H-ras oncogene in T 24 cells. 32'36 It is not only for this reason, however, but also for reasons of synthetic accessibility of phosphorothioates that this class of compounds has been most extensively studied for antisense effects in vitro and, more recently, also in vivo. There are now many examples in the literature where phosphorothioates have proven to act as very potent and specific inhibitors of gene expression, 1-7'29-33'51 but it also has to be noted that this first generation of potential antisense drugs still suffers from a variety of problems associated with it. Due to chirality on phosphorus and the generally stereo-random mode of their synthesis, phosphorothioates are mixtures of 2 n diastereoisomers (where n denotes the number of phosphorothioate

""6

"~0

~0

~,0

o/~

in

co

__./

o,,

o-~'~

,,__.~.~o,,,

w

"r

E::~. 9 m

.~

"a

~ o

e-

:= Q" Q~

oe -

z:

0

.o=

Q)

~ E

~

~8

ul

LU m

o\-~

0

o

N

m

/

2-~

< m.--/~

0

h -~

-', ',0 ~.a,,

~

0

0,,

\J-~--O

d

I

~ "

/

236

C

2-minoadenine

5-propyny luracil

5'-thiazolyluracil

RNA-Affinity:

H73

VH74

VH75

Stability:

L

L3'

NR

Figure 4. Selected types of chemical modifications of oligonucleotides. (A) Modifications of the ribose moiety. (B) Backbone modifications. (C)Base modifications. "RNA-Affinity" refers to the binding affinity of the corresponding modified oligonucleotides for complementary single-stranded RNA. "Stability" refers to nuclease resistance. H = high; VH = very high; M = medium; L = low; V = variable, i. e. there is a strong sequence dependence; NR = not reported.

238


groups; e.g. for a fully modified 20-mer phosphorothioate there are 524,288 possible isomers). It should be mentioned, however, that regulatory institutions have accepted this situation and in most cases have raised no objections to the use of these diastereomeric mixtures in clinical trials. Furthermore, phosphorothioates certainly display suboptimal affinity for complementary RNA, with the duplexes between phosphorothioate oligonucleotides and RNA being less stable than the corresponding duplexes incorporating an unmodified DNA strand. 43'52 However, the major drawback in the use of phosphorothioates for therapeutic purposes may consist in a certain potential for toxic side effects due to unspecific protein binding, which is a general and characteristic phenomenon for this class of oligonucleotide analogues. 53-58 Phosphorothioates bind to serum albumin with an apparent affinity that is in the same range as that of clinically used drugs such as aspirin and penicillin. As indicated above, new generations of modified oligonucleotides are now available which have higher affinity for target RNA than phosphorothioates or even natural DNA, and which also do not inevitably depend on the presence of phosphorothioate groups for nuclease resistance. Even if complete elimination of sulfur may presently be difficult to sustain because of the RNAse H-dependent mechanism of action of many active AS-ODNs (see above), it should nevertheless be possible in the immediate future to evaluate nuclease resistant high-affinity oligonucleotides with reduced sulfur content for biological potency in vitro and in vivo.

D. Cellular Uptake of Antisense Oligonucleotides Given the state of research on chemically modified oligonucleotides with improved RNA binding affinity and nuclease resistance, our understanding of the cellular uptake of oligonucleotides and especially our ability to favorably influence the uptake of antisense oligonucleotides into cells, as well as to control their intracellular distribution, may still be considered to rest in its infancy. The solution to this problem (although there may not be a single and general solution) may well represent the ultimate challenge for a successful general implementation of the antisense strategy in a wide variety of therapeutic areas. Indications exist in the literature that certain cell types possess specific cellular receptor proteins for oligonucleotides which mediate oligonucleotide internalization by an endocytotic mechanism. 76'77 On the other hand, there are numerous examples where oligonucleotides (e.g. phosphorothioates) are not at all taken up by cells from the medium efficiently, as judged by the complete lack of observable biological activity (i.e. downregulation of target RNA or protein). Alternatively, oligonucleotides after cellular uptake may be trapped in endosomal vesicles and thus not be available for intracellular binding to their target mRNA. 78 Cellular uptake of antisense oligonucleotides and the associated biological activities can be improved to a dramatic extent by the use of cationic lipids (e.g.


239

DOTAP, DOTMA) as uptake mediators. 79'8~Cellular uptake of oligonucleotides is highly cell-type specific and optimization of uptake properties for certain cell lines does not necessarily lead to improved uptake in others. Improvement of oligonucleotide internalization has been reported for cholesterol 81'82 and polylysine 83 conjugates of oligonucleotides, for oligonucleotide-transferrin conjugates 84 (involving binding to the transferrin receptor), and for oligonucleotides encapsulated in liposomes, 85-87 but it is unclear how generalizable these results are, and especially whether they can be extrapolated to in vivo experiments. As indicated above, in vitro cell culture experiments are often conducted in the presence of cationic lipids which (generally) greatly enhance oligonucleotide uptake and thus allow for the potency of various oligonucleotides to be assessed outside of the limitations imposed by insufficient cellular uptake. In this way, active sequences can be identified in a rather straightforward fashion, but the question remains whether the same efficacy can be achieved in vivo by simple i.v. or i.p. administration of the same oligonucleotide, even if nuclease resistance is not an issue. The question of cellular uptake of oligonucleotides is further complicated by the fact that the effects of cationic lipids themselves are highly cell-type specific, thus leading to variable results between different cell lines.

III. ANTISENSE OLIGONUCLEOTIDES AS ANTICANCER AGENTS A. Potential Targets in Antisense-Based Cancer Therapy In the past decade, major progress has been made in understanding the molecular basis of the pathophysiology of cancer. The unrestricted proliferation of cells is now perceived to be the result of the activation of oncogenes or the loss of tumor suppressors, both of which confer a selective growth advantage to cancer cells. Oncogene activation itself in most cases is the consequence of structural alterations in the chromosomal organization of the affected cells, such as amplification of genes, translocations, or point mutations. 88-91The consequences of these alterations are either the deregulated expression of structurally normal oncogene products, or the expression of an abnormal oncogene (point mutations and translocations that alter the affected oncogene). Many of the genes that are mutated or lost in cancer cells, including both the oncogenes and tumor suppressors, encode proteins that are crucial regulators in various signal transduction pathways, like growth factors and their receptors, GTP-binding proteins, transcription factors, adhesion molecules, angiogenic factors, or tumor suppressor genes such as p53 and Rb. Almost every extracellular signal transduced either by receptor-activated tyrosine phosphorylations or by receptor coupling to trimeric GTP-binding proteins is amplified and diversified inside cells by protein kinase cascades. The constituent protein kinases building up these highly conserved networks are important players in propagating, integrating, and delivering incoming mitogenic stimuli to the

240


cell-cycle regulators. Mutant alleles of these protein kinase genes or of oncogenic ras as well as the anti-oncogene p53 or Rb affect these protein kinase cascades. This perturbs the entire signaling network, leading to the deregulation of both transient processes such as changes in cell shape or motility, or less reversible processes such as cell differentiation, division, and programmed cell death (apoptosis). Thus, the multistage carcinogenic process, which ultimately leads to a clinically identifiable tumor mass, can be conceptualized as a multistep and progressive disordering of the signal transduction processes inside the cell, with the set of genes responsible for the deregulation of mitogenic signaling most likely being involved in the rate-limiting steps in cellular tumor growth. Based on these concepts it is clear that the selective inhibition of deregulated oncogene products (e.g. a variety of protein kinases involved in signal transduction) or the restoration of function of mutated tumor suppressors should result in antitumor activity. In contrast to classical antitumor chemotherapy, such a molecular approach to the growth inhibition of tumor cells may be expected to yield highly efficacious growth inhibitors with superior tolerability as compared to standard antineoplastic agents. Although the specific inhibition of oncogene product function may in principle also be accomplishable by appropriate small molecules (e.g. highly specific protein kinase inhibitors), it is obvious from the above discussions that antisense technology is ideally suited to neutralize mutated or deregulated oncogenes in neoplastic cells, since it offers the unique opportunity to knock out target genes by a highly specific mechanism. It is therefore reasonable to expect that antisense-based anticancer agents should be devoid of the undesired side effects associated with classical anticancer chemotherapy; however, it has to be kept in mind that other types of toxicities may surface for antisense or antigene oligonucleotides, which may also depend on the exact chemical nature of the (modified) oligonucleotide in question, and which are not predictable at this stage. A great variety of nuclear factors and oncogenes have been targeted by antisense oligonucleotides or antisense vector constructs in cell culture studies, including the transcription factors c-myc, 92-97 c-myb, 98-103 c-fos, 104-106 c-jun, ~~ and NF-KB, 107-109and additional nuclear factors such as p 120 nucleolar antigen ll~ or components of the cell cycle machinery like PCNA lll'll2 and cdc2.113'114 In addition, various growth factors and their receptors 115-124as well as growth-related kinases involved in cellular signal transduction have been chosen as targets for antisense approaches, including protein kinase C, 125-127 protein kinase A subunits, 122'128'129tyrosine kinases, 13~ and c-raf. 51'138-141. Also oncogenes like Ha-ras, 31'32 K-ras, 142'143 and N-ras, 144 as well as tumor suppressors like p53, have been targeted. In many cases, inhibition of protein synthesis and/or downregulation of the targeted mRNA have been observed even at submicromolar oligonucleotide concentration (especially with nuclease-resistant phosphorothioates), and several of these studies have additionally demonstrated pronounced anti-proliferative effects of AS-ODNs in cell culture experiments. These studies have recently been excellently reviewed by Pierga and Magdalenat 145and their arguments will not be

Antisense- and Antigene-Based Drug Design

241

reiterated here in any detail. Rather, the following sections will focus on some examples of relevant studies in animal models, involving either ex vivo treatment of affected cells by A S - O D N s and subsequent reimplantation, or systemic applications of A S - O D N s (Table 1). Potential therapeutic applications of A S - O D N s in oncology have only recently been conceived, but the increasing interest in antisense technology in the treatment of cancer and neoplastic diseases is reflected by a growing list of review articles in the international literature that cover this exploding field. 145-157

Targets Related to Hematological Malignancies The most suitable target genes (respectively the corresponding m R N A s ) for A S - O D N s are those that are exclusively or at least preferentially expressed in malignant cells, but are not (or only at strongly reduced levels) in normal nontransformed cells. M a n y malignant diseases of the hematologic and lymphatic systems

Table 1. Inhibition of Oncology Targets in'Animals by Antisense Oligonucleotides Target Protein

Indication TargetRegion

bcr-abl cdc2 PCNA c-myb c-myb

CML restenosis restenosis AML/CML cancer

cyclin B 1 bcl-2

restenosis B-cell lymphoma cancer cancer

Ha-ras Ha-ras TGF-t~ EGF receptor NF-~B, p65 NF-~cB, p65 ICAM- 1 p 120 PKCtx, murine PKC-ct, human c-raf basic FGF PKA, RIt~

cancer cancer cancer cancer melanoma metastasis cancer cancer cancer cancer Kaposi's sarcoma cancer

OligoModification

Concept Reference Validation Number

breakpoint AUG AUG AUG 5'-UTR, coding AUG AUG

26-mer,P"-S 18-mer, P=O 18-mer, P--O 21-mer, P=S 20-mer, P--S

+ + + +

169 225 225 170 230

18-mer, P--O 20-mer, P--O P=S cap

-

114 177

5'-UTR coding (codon 12) AUG AUG AUG AUG 3'-UTR

15-mer, P--O 13-mer, P--O, absorbed to nanoparticles 39-mer, P=O P--S cap 39-mer, P=O P---S cap 24-mer, P=S 20-mer, P--S 20-mer, P--'S

3'-UTR AUG 3'-UTR 3'-UTR AUG

20-mer, P'-S 20-mer, P--S 20-mer, P=S 20-mer, P-'S 24-mer, P--S

110 30 165 62 229

AUG

18-mer, P=S

168

179 180 + -

227 227 107 228 181

242


fulfill these requirements and can therefore be considered as ideal targets for antisense approaches. The bcr-abl fusion protein is typically found in patients with a Philadelphia chromosome-positive leukemic disease. At the molecular level the abl proto-oncogene is translocated from the long arm of chromosome 9 onto the bcr gene on chromosome 22 to form a hybrid bcr-abl fusion gene that is expressed as the fusion protein p210 bcr-abl. This chromosomal translocation is the hallmark of chronic myelogenous leukemia (CML), 158but was also detected later in a cohort of Philadelphia-positive acute lymphocytic leukemia patients (ALL). 159The hybrid p210 bcr-abl protein has deregulated tyrosine kinase activity and is therefore important for the pathogenesis and progression of CML. The characteristic of CML is a clonal disorder of the hematopoietic stem cell system, which evolves in two different phases, a chronic phase and the acute blast crisis. The chronic phase of CML is characterized by an increase in immature and mature hematopoietic stem cells in the peripheral blood and in the bone marrow. The subsequent CML blast crisis involves a marked degree of differentiation arrest of leukemic cells in the peripheral blood and bone marrow, and due to the fact that the blasts are refractory to any of the established therapeutic regimens, ultimately leads to death. Philadelphia chromosome-positive CML blast cells have been targeted both in vitro and in vivo with antisense ODNs against the bcr-abl fusion region. 16~ Inhibition of cell proliferation has also been achieved by treatment of leukemia cells with a bcl-abl-specific ribozyme, which resulted in specific cleavage of bcr-abl mRNA at the targeted site. 166'167The tendency to form blast colonies in vitro was inhibited 80-90% by unmodified AS-ODNs (15-30 ~tM) compared to cells treated with control oligonucleotides. TM This study used a mismatched control oligonucleotide and it was demonstrated that the introduction of mismatches into the antisense sequence inhibited the specific effect of the antisense sequence on blast cell colony formation. Furthermore, it was also shown that if normal bone marrow and blast cells were mixed, exposed to the oligonucleotide, and assayed for residual colony formation, the majority of residual cells consisted largely of differentiated cells suggesting the selective elimination of leukemic blast cells. 161'168 In another study, the CML blast crisis cell line BV173 was injected into SCID mice where a disease process developed that closely resembled that observed in leukemic patients with blast crisis. 169 One to three weeks after injection of 106 BV 173 cells into mice, bcr-abl transcripts were detected in the bone marrow, blood, spleen, liver, and lungs. Systemic treatment of the leukemic mice with a 26-mer bcr-abl AS-ODN for 9 days (from day 7 or day 21 after tumor inoculation) led to a marked decrease in bcr-abl transcripts and reduced the number of leukemic cells. AS-ODNs, but not sense or mismatched controls, prolonged the survival of the BV173-injected mice from 8-13 weeks to 18-23 weeks. This study provided compelling evidence for the in vivo effectiveness of anticancer therapy based on antisense ODNs targeting a tumor-specific gene. In addition, bcr-abl antisense oligodeoxynucleotides in combination with a low dose of mafosfamide drastically reduced clonogenicity and growth of BV173 cells in nude mice, in a 1:1 mixture

Antisense- and Antigene-Based Drug Design

243

with normal bone marrow. 168Such combination treatment was by far more effective in the elimination of tumor cells than a high dose mafosfamide treatment, and spared much higher numbers of normal progenitors. This experiment demonstrated that AS-ODNs could potentially be used as nontoxic and highly effective agents for ex vivo bone marrow purging in autologous bone marrow transplantation to eliminate leukemia cells specifically. The c-myb proto-oncogene encodes proteins that have been shown to be important for hematopoietic cell proliferation. In vivo treatment of human leukemia SCID mouse chimeras that had been inoculated with the human K562 leukemia cell line with c-myb antisense phosphorothioate ODNs by osmotic minipumps at a dose of 5 mg/kg body weight significantly prolonged the survival of the mice. 17~ Once inoculated leukemic blast cells were detected, the survival of untreated mice or mice treated with control ODNs was only 6 + 3 days, whereas the mice that were treated with AS-ODNs to c-myb survived 3 to 8 times longer. In addition, the AS-ODN treated mice had significantly less leukemic cell infiltration in the central nervous system and the ovary, the two sites that show manifestation of disease most frequently. Another important gene fusion event occurs after the translocation t(15;17) in acute promyelocytic leukemia. This translocation involves the gene for the retinoic acid receptor-c~ and a gene designated PML. Burkitt's lymphoma is a B-cell malignancy characterized by the chromosomal translocation t(8:14). This translocation is found in 90% of all cases and juxtaposes c-myc and immunoglobulin heavy-chain genes. These fusion proteins are only found in malignant cells, to which they confer a growth advantage, and although no antisense work has been reported on those targets they should be ideally suited for inhibition with AS-ODNs. Problematic side effects of such a therapy due to the elimination of these proteins in normal cells is unlikely, as is the development of resistance if the presence of the fusion protein is necessary for the persistence of tumor growth. Another potential therapeutic target that has been examined for the treatment of acute myeloid leukemia is p53, the most frequently mutated gene in cancer cells. 171' 172 p53 is known to be a tumor suppressor gene that is mutated in leukemic cells and consequently leads to a transformed phenotype by endowing the leukemic cells with a selective growth advantage. The p53 AS-ODN treatment resulted in a drastic decrease in viable leukemia blast cells. The aberrant expression of protooncogenes in tumor cells can also be caused by another recognizable mechanism of gene activation--gene amplification. A 20--40fold amplification of the c-myc gene has been observed in HL-60 promyelocytic leukemia cells. 173 Exposure of HL-60 cells to a phosphodiester oligonucleotide designed to hybridize to the translational start codon of c-myc, in a dose-dependent manner, inhibited c-myc expression and proliferation of HL-60 cells by inhibition of S-phase entry, 174 whereas various control oligonucleotides with unrelated sequences had no effects. These results demonstrate that the growth advantage that c-myc overexpression confers to HL-60 cells can be eliminated by inhibiting

244


overexpression of the oncogene. An 18-mer AS-ODN complementary to codons 2 to 7 of c-myb mRNA inhibited c-myb RNA and protein expression in myelogenous leukemia cell lines and primary leukemia cells obtained from patients. 98 The downregulation of c-myb expression was accompanied by inhibition of proliferation of the leukemic cell lines and by inhibition of clonogenic growth in freshly obtained leukemia cells. However, in contrast to c-myc, inhibition of c-myb expression in HL-60 cells was not accompanied by the induction of terminal differentiation, suggesting that c-myc and c-myb regulate leukemia cell proliferation and differentiation at different levels. The Wilms tumor gene (WT1) has been recently targeted with antisense oligonucleotides in leukemia cells and inhibition of proliferation and induction of programmed cell death was observed. 175'176These data indicate that the Wilms tumor gene plays an important role in leukemogenesis and therefore should represent an attractive target for the treatment of leukemia with antisense oligonucleotides.

Targets Related to Solid Tumors The activation of oncogenes that confer a growth advantage to the tumor cell is also an important mechanism in solid tumors. In vitro treatment of a B-cell lymphoma cell line that had a bcl-2 translocation with AS-ODNs directed to the translational start codon of the bcl-2 mRNA before injection into SCID mice resulted in a reduction in bcl-2 expression and prevented the establishment of lymphoma disease in the mice, whereas sense and nonsense control oligonucleotides were without effect. 177 AS-ODNs have been targeted to the p65 subunit of NF-~B, which is implicated in the induction of adhesion molecule expression. In this very elegant study, AS-ODNs were shown to inhibit the tumorigenicity of a fibrosarcoma cell line in vivo and led to tumor regression in nude mice. 1~ In this particular study, the mechanism of action of the AS-ODNs was explicitly verified (i.e. the antisense concept was validated) by means of control oligonucleotides that were without effect, and by the use of tumor cells transfected with a dexamethasone-inducible antisense construct. The effect of the inducible antisense construct in inhibiting in vivo tumor growth was comparable to the effect of systemically administered oligonucleotides. In another study, NF-r,B Rel A expression could be successfully inhibited by phosphorothioate antisense oligonucleotides resulting in the suppression of urokinase-type plasminogen activator (uPA) but not of its inhibitor PAI- 1. These findings indicate that Rel A may play an important role in tumor spread and metastasis. 178 In follicular lymphomas, the bcl-2 gene is overexpressed due to its juxtaposition to immunoglobulin heavy-chain sequences as a result of a t(14; 18) translocation. Since bcl-2 inhibits programmed cell death, the elimination of its aberrant expression in these lymphomas could potentially lead to tumor regression. The design of a tumor-specific oligonucleotide targeting the region of bcl-2 that is juxtaposed to immunoglobulin sequences should be achievable as in the case of oligonucleotides that target the bcr-abl fusion gene in CML.


245

Oligonucleotides against the nucleolar antigen p120 were shown to inhibit the growth and tumorigenesis of human melanoma cells (LOX) injected i.p. into nude mice. 11~ In this study, the oligonucleotides were injected i.p. into tumor-beating mice at days 1, 3, and 5 after inoculation in the presence of cationic lipids. This study represents the first report showing enhancement of antisense effects by cationic lipids in vivo in an animal model. The ras oncogene is one of the most frequently mutated oncogenes in several different types of tumors. In vitro treatment of Ha-ras transformed NIH3T3 cells with AS-ODNs to the 5'-flanking region of the ras mRNA for 3 consecutive days (50 ktM) resulted in more than 90% reduction in Ha-ras levels. 179When these cells were implanted into nude mice, tumor growth of the AS-ODN treated cells was dramatically reduced for up to 14 days, compared to cells treated with nonspecific control ODNs. The treatment of transformed cells with Ha-ras AS-ODNs reversed the transformed phenotype and this reversal of transformation was a long-lasting effect after termination of the AS-ODN treatment. Oligonucleotides directed to the codon 12 point mutation of Ha-ras selectively inhibited the proliferation of cells expressing the mutated Ha-ras. 18~This study used phosphodiester ODNs which are rapidly degraded by nucleases. To increase their nuclease resistance, the ODNs were absorbed onto polyalkylcyanoacrylate nanoparticles, which resulted in a 100-fold increase in potency compared to free ODNs. When cells expressing mutated Ha-ras were implanted into nude mice, they formed ras-dependent tumors and AS-ODNs absorbed to nanoparticles were also capable of inhibiting tumor growth in vivo. In this study 18~it was demonstrated that inhibition ofras oncogene can block tumor development even though ras oncogenic activation is probably an early event in tumor progression. In addition, the stability and cellular activity of phosphodiester oligonucleotides was shown to be dramatically improved by absorption of the oligonucleotide onto polyalkylcyanoacrylate nanoparticles. Cytokine treatment (TNF-o~ and IFN-T) of human metastatic melanoma cells (C8161) increased the formation of lung metastasis in nude mice up to fourfold. 181 Treatment of C8161 cells in vitro with AS-ODNs to the cell adhesion protein ICAM-1 inhibited ICAM-1 expression > 90% and inhibited metastasis formation in vivo, indicating that ICAM- 1 is involved in melanoma cell metastasis formation. Osteopontin is a secreted, calcium-binding phosphoprotein that frequently has been associated with the transformed phenotype. Antisense inhibition of osteopontin expression in murine PAP2 cells, a metastatic ras-transformed NIH3T3 cell line, reduced the tumorigenic and metastatic properties of tumors in vivo, thus demonstrating that osteopontin is involved in tumorigenesis. ~82 The regulatory subunits of protein kinase A, RI, and RII have been implicated in tumor growth. RI is a growth-stimulatory protein, whereas RII is a growth inhibitory and differentiation-inducing protein. 154Rltx is presumably responsible for the neoplastic cell growth of human colon, breast, gastric cancer, and neuroblastoma cells. 154'183Exposure of the cell lines to a 21-mer phosphorothioate oligonucleotide

246


to human Rloc exhibited growth inhibition with an IC50 of 6 lxM without signs of toxicity. Oligonucleotide treatment was correlated with a decrease in RIo~ mRNA and protein suppression and a concomitant increase in RIII3 expression. 183 The erbB2 oncogene (also called neu and Her-2) codes for a 185-kDa transmembrane growth factor receptor with an intracellular tyrosine kinase catalytic domain. 184'185The erbB2 gene copy number is amplified in many adenocarcinomas; erbB-2 is also amplified and overexpressed in approximately 25% of human breast carcinoma samples. Amplification and overexpression oferbB2 in malignant breast tumors has been correlated with nodal metastases, early relapse, and shortened survival. 186 erbB2 antisense phosphodiester oligonucleotides were used to inhibit the growth and DNA synthesis of breast cancer cell lines 184with an amplified erbB2 gene, and this growth inhibition was accompanied by a dose-dependent inhibition of erbB2 expression, as measured by immunohistochemistry. Control sense oligonucleotides did not inhibit cellular proliferation at the same concentration, but showed inhibitory effects at much higher concentration. There was no effect of erbB2 oligonucleotides on breast cancer cell lines that had no amplification of erbB2, indicating that the erbB2 overexpression is important for the proliferation of breast cancer cells that have been selected for erbB2 amplification. Malignant melanomas, unlike normal melanocytes, can proliferate in the absence of exogenous basic fibroblast growth factor. Exposure of primary melanomas in the vertical growth phase and metastatic melanomas to unmodified AS-ODNs, targeted against 3 different sites of human bFGF mRNA, inhibited cell proliferation and colony formation in soft agar. 187 In addition, the human malignant melanoma cell line HTZ19 was demonstrated to be dependent on PDGF-AA as an autocrine growth factor. 188 Transfection of tumor cells with antisense constructs has been used to investigate the role of target proteins and oncogenes in tumorigenesis. Expression of antisense RNA to TGF-13 with a metallothionein promoter construct 189 reduced the anchorage-independent growth of murine mesothelioma cells in vitro and their tumorigenesis in vivo, but did not influence anchorage-dependent growth. Maximal reductions in TGF-I3 mRNA of 60 to 80% were achieved in this study. This finding suggests that very strong promoter systems have to be used for vector-based antisense inhibition of target genes in order to achieve the estimated required ratio of 1:20 between target mRNA and antisense RNA for efficient inhibition of gene expression. In another study, an antisense construct to the IGF-1 receptor ~9~was introduced into FO-1 human melanoma cells under the control of the HSP70 promoter. The expression of the antisense construct resulted in a marked reduction in the number of IGF-1 cell surface receptors, and the growth of the transfected cells implanted into nude mice was strongly inhibited or even abrogated. When tumors arose after long delay in nude mice, the tumor cells had often lost the expression plasmid and the IGF-1 receptor levels had returned to wild-type levels. The growth-inhibitory effect of vector-transfected cells in the mice in vivo was remarkable, because the


247

growth of FO-1 cells in monolayer culture was not affected by the expression of antisense RNA. Inhibition of tumorigenesis in this study was also observed when FO-1 cells were treated in vitro with synthetic AS-ODNs prior to the injection into nude mice. These results strongly indicate that this effect was due to antisense inhibition of IGF- 1 receptor expression. Urokinase plasminogen activator (uPA) and its surface receptor (PAR) have been shown to correlate strongly with an invasive tumor cell phenotype. A human malignant epidermoid carcinoma cell line (Hep3) was transfected with an antisense construct to PAR in a constitutive promoter construct. 191'192 The antisense clones showed 50-75% reduction in PAR mRNA and protein expression. The clones with the lowest PAR expression showed a significantly lower level of invasion (chorioallantoic membrane) and a drastically reduced tumorigenicity and local invasion in vivo. It was concluded therefore that diminished expression of surface PAR in tumor cells leads to a reduction in the invasiveness and an increase in tumor latency, which makes PAR an attractive target for inhibition by AS-ODNs in different tumor cell lines. Another target for AS-ODNs in solid tumors is. c-raf kinase, a kinase that is located downstream of Ras and upstream of Map kinase in the Map kinase signaling pathway. It has previously been demonstrated that transfection of the radiationresistant human squamous carcinoma cell line SW-20B with a c-raf-1 antisense construct significantly reduced the malignant potential, in comparison to cells containing a construct with a sense orientation. 193The raf antisense clones appeared to have an increased radiation sensitivity, indicating that reduced expression of raf kinase is sufficient to modulate both the tumorigenicity and the radiation phenotype of squamous carcinoma cells. In a recent study Monia et al. have reported the identification of a potent 20-mer phosphorothioate antisense oligonucleotide targeted to the 3'-UTR of human c-raf mRNA. In the presence of cationic lipids this oligonucleotide, CGP 69846ABSIS 5132, inhibits c-rafkinase expression in cancer cells in vitro with an IC50 below 100 nM, and also shows potent antitumor activity against a variety of human tumors subcutaneously transplanted into nude mice in the dose range of 0.006 to 6 mg/kg (Figure 5). 51 Employing analogues of CGP69846A/ISIS5132 incorporating an increasing number of mismatched bases, it was also demonstrated that the antitumor activity of this oligonucleotide is a highly sequence-specific phenomenon (Figure 6). 66b In combination with other data, especially the observed downregulation of c-raf mRNA levels in the tumors of CGP69846A/ISIS5132-treated animals, 51 this latter f i n d i n g s t r o n g l y s u g g e s t s that the p o t e n t a n t i t u m o r a c t i v i t y of CGP69846A/ISIS5132 is mediated by a true antisense mechanism of action involving RNAse H-mediated degradation of bound target mRNA. C G P 6 9 8 4 6 A / I S I S 5 1 3 2 has been modified by incorporation of 2'-0methoxyethyl modified building blocks (cf. Figure 4A), which has resulted in an analogue (CGP69845A) which in comparison to the 2'-deoxyphosphorothioate displays higher biological activity in vitro and is at least an equipotent inhibitor of

248


tumor growth in vivo. 66 CGP69845A may be considered as the prototype of a second-generation antisense oligonucleotide which exhibits a significantly reduced phosphorothioate content, and therefore may be less prone to side effects that are caused by the presence of phosphorothioate linkages. With regard to other targets involved in signal transduction, Dean et al. have recently described a 20-mer phosphorothioate oligonucleotide targeted to the 3'-UTR of human PKC-o~ mRNA (CGP64128A/ISIS3521). 127 As in the case of CGP69846A/ISIS5132, CGP64128A/ISIS3521 is a potent inhibitor of PKC-o~ expression in vitro and displays high antitumor activity against various human tumors transplanted into nude mice. 127 It should be noted that CGP64128A/ISIS3521 is a true isozyme-specific inhibitor of PKC-o~ expression and does not inhibit the expression of other PKC isozymes, i.e. PKC-~5,-E, -11, and _~.126 Downregulation of PKC-a expression in A549 tumors by CGP 64128A has also been demonstrated (J. Phillipps, unpublished observation, this laboratory).

0.8

CGP 69846A 0.7

~

--

~ _.~

0.6-

,

Control 0.006 mg/kg 0.06 mg/kg 0.6 mg/kg 6.0 mg/kg

0.50.4O

~'

0.3-

~

0.2-

0.1

--

0.0 8

I

I

I

I

I

I

10

12

14

16

18

20

22

days after tumor transplantation

Figure 5. Antitumor activity of CGP 69846A (A) and a scrambled control ODN (B) against A549 human lung carcinomas transplanted s.c. into nude mice. 51 A549 human lung carcinoma tumor fragments (25 mg) were implanted subcutaneously into nude mice at day_ 0. The tumors were allowed to grow until they reached a mean volume of 100 mm 3. CGP 69846A and a control ODN were injected i.v. once daily from day 9 to day 21 at doses of 6, 0.6, 0.06, and 0.006 mg/kg. Tumor growth was monitored twice weekly and tumor volumes were calculated.


249

The antitumor activity of CGP69846A and CGP64128A has also been studied in combination with standard chemotherapeutic agents (cisplatin, mitomycin-C, vinblastine, tamoxifen, 5-fluorouracil, adriamycin, and ifosfamide) against a variety of human tumors transplanted into nude mice (Figure 7). For most of the combinations studied, antitumor effects were essentially additive, but pronounced synergism was observed for the combination of CGP69846A/ISIS5132 as well as CGP64128A/ISIS3521 with mitomycin and cisplatin (Geiger et al., unpublished observations, this laboratory). For PC-3 human prostate carcinomas, treatment with CGP69846A/ISIS5132 in combination with cisplatin resulted in tumor cures. Both CGP69846A and CGP64128A are in development as anticancer agents and are presently in phase I clinical trials. The expression of the metalloproteinase matrilysin in human colon carcinoma cells (SW480, SW420) correlates with the ability to invade an artificial basement membrane in vitro and to metastasize to the liver following injection into the cecum of nude mice in vivo. 194 Reduction of matrilysin expression by AS-ODNs decreased tumorigenicity in vivo and inhibited subsequent metastasis to the liver, indicating

B

0.9-

Control ODN 0.8

-

0.7

-

Control --II-- 0.006 mg/kg 0.06 mg/kg 0.6 mg/kg 6.0 mg/kg

CO

E 0.6o

O

E 0.5O O

E 0.40.3

--

0.2

--

0.1

--

8

i

I

i

I

I

I

10

12

14

16

18

20

days after tumor transplantation Figure 5. (Continued)

22

250


that matrilysin is involved in tumorigenicity and metastasis of colon carcinoma cell lines. Inhibition of matrilysin expression by antisense oligonucleotides was also observed in BM314 colon cancer cells, resulting in inhibition of invasiveness in vitro. 195

Multidrug Resistance Emergence of drug-resistant tumors after chemotherapy is a persistent problem in the treatment of cancer. The most common and serious manifestation of this problem is the phenomenon of both acquired and intrinsic pleiotropic drug resistance or multidrug resistance (MDR). MDR is a complex system of biochemical

A

100 @ 5 mismatche,, ~' 4 mismatches

A 2

tO O O

80

60

v

Advances in DNA Sequence Specific Agents, Volume 3

Sequence-specific DNA Binding Agents

Advances in DNA Sequence Specific Agents, Volume 2

Bioinformatics for DNA sequence analysis (Volume 537)

Bioinformatics for DNA Sequence Analysis

Advances in Agronomy, Volume 3

Advances in Computers, Volume 3

Advances in Pharmacology Volume 3

Advances in Pharmacology Volume 3

Advances in Immunology Volume 3

Advances in Genetics Volume 3

Advances in Computers, Volume 3

Advances in Atomic Spectroscopy, Volume 3 (Advances in Atomic Spectroscopy)

Probability models for DNA sequence evolution

Advances in Experimental Social Psychology, Volume 3

Advances in Applied Mechanics, Volume 3

Advances in Developmental Biochemistry, Volume 3

Advances in Applied Microbiology, Volume 3

Advances in Asymmetric Synthesis, Volume 3, Volume 3 (Advances in Asymmetric Synthesis)

Advances in Porous Media. Volume 3

Advances in Heterocyclic Chemistry, Volume 3

Advances in Ecological Research, Volume 3

Advances in Cancer Research, Volume 3

Advances in Theoretically Interesting Molecules: Volume 3

Advances in Dendritic Macromolecules. Volume 3

Advances in Low-Temperature Biology, Volume 3

Advances in Chemical Engineering, Volume 3

Advances in Applied Mechanics, Volume 3

Probability models for DNA sequence evolution

DNA REPLICATION - CURRENT ADVANCES

Advances in DNA Sequence Specific Agents, Volume 3

Sequence-specific DNA Binding Agents

Advances in DNA Sequence Specific Agents, Volume 2

Bioinformatics for DNA sequence analysis (Volume 537)

Bioinformatics for DNA Sequence Analysis

Advances in Agronomy, Volume 3

Advances in Computers, Volume 3

Advances in Pharmacology Volume 3

Advances in Pharmacology Volume 3

Advances in Immunology Volume 3

Advances in Genetics Volume 3

Advances in Computers, Volume 3

Advances in Atomic Spectroscopy, Volume 3 (Advances in Atomic Spectroscopy)

Probability models for DNA sequence evolution

Advances in Experimental Social Psychology, Volume 3

Advances in Applied Mechanics, Volume 3

Advances in Developmental Biochemistry, Volume 3

Advances in Applied Microbiology, Volume 3

Advances in Asymmetric Synthesis, Volume 3, Volume 3 (Advances in Asymmetric Synthesis)

Advances in Porous Media. Volume 3

Advances in Heterocyclic Chemistry, Volume 3

Advances in Ecological Research, Volume 3

Advances in Cancer Research, Volume 3

Advances in Theoretically Interesting Molecules: Volume 3

Advances in Dendritic Macromolecules. Volume 3

Advances in Low-Temperature Biology, Volume 3

Advances in Chemical Engineering, Volume 3

Advances in Applied Mechanics, Volume 3

Probability models for DNA sequence evolution

DNA REPLICATION - CURRENT ADVANCES

Recommend Documents