The development of more effective drugs for treating pa- tients with cancer has been a major human endeavor over the past 50 years, and the 21st century now promises some dra- matic new directions. While improvements in surgery and ra- diotherapy have had a major impact on cancer treatment, the concept of systemic chemotherapy, specific for cancer cells and free of major side effects, remains a critical goal for the future. The issues underlying the achievement of this goal are complex, extending from an understanding of how cancer growth is controlled, through the technology of drug synthe- sis and testing, to the multifactorial requirements for clinical trial. For anyone working in any single area of anticancer drug development, it is important to have an overview of the whole process. This book aims to provide such an overview. The opening chapters discuss possible targets for drug design, including the cell division cycle, growth signal transduction, apoptosis induction, and the manifold interactions between tumor cells and host tissues. Succeeding chapters then consider tech- niques of identifying new potential drugs, including molecu- lar modeling, chemical synthesis, and screening. The con- cluding chapters detail the required steps that any new potential anticancer agent must go through before it can be considered for routine clinical treatment. In each of these ar- eas, a number of eminently qualified contributors have pro- vided commentaries. Inevitably there are areas of overlap, but these have been retained because they reflect the interdepen- dence of different areas of research. We hope that this book will provide a useful commentary, including both overviews and specific detail, on this vital but fascinating subject. We also hope that it will stimulate origi- nal thought and further encourage those from both scientific and medical backgrounds who are committed to improving the outlook of cancer patients worldwide. Acknowledgments This book represents the results of a considerable amount of work by many talented contributors. As editors we thank all of these contributors for their enthusiasm and patience. We thank our research colleagues for providing advice when needed and the staff of Academic Press for their support over many months. B.C.B particularly acknowledges the support of the Auckland Division of the Cancer Society of New Zealand, and through them the members of the public who have donated money for cancer research. Bruce C. Baguley David J. Kerr x i i i P R E F A C E 9863.FM 8/28/01 7:22 AM Page xiii Numbers in parentheses indicate page numbers on which authors’ contributions begin. Eric Ofori Aboagye (353) Cancer Research Campaign, Im- perial College School of Medicine, Hammersmith Hospi- tal, London W12 0NN, United Kingdom Wynne Aherne (249) Cancer Research Campaign Centre for Cancer Therapeutics, Institute of Cancer Research, Sutton, Surrey SM2 5NG, United Kingdom Bruce C. Baguley (1, 269) Auckland Cancer Society Re- search Centre, The University of Auckland, Auckland 1000, New Zealand Alex Bridges (31) Pfizer Global Research and Develop- ment, Ann Arbor, Michigan 48105 Angelika M. Burger (285) Tumor Biology Centre, Univer- sity of Freiburg, D-79110 Freiburg, Germany Jian Cao (91) School of Medicine, State University of New York at Stony Brook, Stony Brook, New York 11794 Joseph M. Covey (301) Division of Cancer Treatment and Diagnosis, National Cancer Institute, National Institutes of Health, Rockville, Maryland 20892 Thomas Davis (157) Oncology, Medarex, Annandale, New Jersey 08801 Stuart Decker (31) Pfizer Global Research and Develop- ment, Ann Arbor, Michigan 48105 Maja J. A. de Jonge (329) Department of Medical Oncol- ogy, Rotterdam Cancer Institute, 3075 EA Rotterdam, The Netherlands William A. Denny (187) Auckland Cancer Society Re- search Centre, Faculty of Medical and Health Sciences, University of Auckland, Auckland 1000, New Zealand Susan J. Donohue (301) Division of Cancer Treatment and Diagnosis, National Cancer Institute, National Institutes of Health, Rockville, Maryland 20892 Nathalie Droin (55) INSERM U517, Faculty of Medicine, 21000 Dijon, France Patrick Ducoroy (55) INSERM U517, Faculty of Medi- cine, 21000 Dijon, France David Ferry (371) Department of Oncology, New Cross Hospital, Wolverhampton, West Midlands WV10 0QP, United Kingdom Heinz-Herbert Fiebig (285) Tumor Biology Centre, Uni- versity of Freiburg, D-79110 Freiburg, Germany Rodolphe Filomenko (55) INSERM U517, Faculty of Medicine, 21000 Dijon, France David W. Fry (31) Pfizer Global Research and Develop- ment, Ann Arbor, Michigan 48105 Michelle Garrett (249) Cancer Research Campaign Centre for Cancer Therapeutics, Institute of Cancer Research, Sutton, Surrey SM2 5NG, United Kingdom Cherry L. Herald (203) Cancer Research Institute, Arizona State University, Tempe, Arizona 85287 Kevin O. Hicks (269) Auckland Cancer Society Research Centre, The University of Auckland, Auckland 1000, New Zealand Fiona Hogan (203) Cancer Research Institute, Arizona State University, Tempe, Arizona 85287 Robert C. Jackson (171) Cyclacel Limited, Dundee DD1 5JJ, United Kingdom David J. Kerr (371) Institute for Cancer Medicine, Univer- sity of Oxford, 0X2 6HE, United Kingdom C O N T R I B U T O R S x i 9863.FM 8/28/01 7:22 AM Page xi Kurt W. Kohn (13) Laboratory of Molecular Pharmacol- ogy, Center for Cancer Research, National Institutes of Health, Bethesda, Maryland 20892 Elianne Koop (123) Laboratory of Medical Oncology, Di- vision of Medical Oncology, University Medical Centre Utrecht, 3584 CX Utrecht, The Netherlands Alan Kraker (31) Pfizer Global Research and Develop- ment, Ann Arbor, Michigan 48105 W. R. Leopold (31) Pfizer Global Research and Develop- ment, Ann Arbor, Michigan 48105 Walter J. Loos (329) Department of Medical Oncology, Rotterdam Cancer Institute, 3075 EA Rotterdam, The Netherlands Ted McDonald (249) Cancer Research Campaign Centre for Cancer Therapeutics, Institute of Cancer Research, Sutton, Surrey SM2 5NG, United Kingdom Christopher J. Molloy (91) 3-Dimensional Pharmaceuti- cals, Inc., Exton, Pennsylvania 19341 Ion Niculescu-Duvac (137) Cancer Research Campaign Centre for Cancer Therapeutics, Institute of Cancer Re- search, Sutton, Surrey SM2 5NG, United Kingdom George R. Pettit (203) Cancer Research Institute, Arizona State University, Tempe, Arizona 85287 Stephanie Plenchette (55) INSERM U517, Faculty of Medicine, 21000 Dijon, France Yves Pommier (13) Laboratory of Molecular Pharmacol- ogy, Center for Cancer Research, National Institutes of Health, Bethesda, Maryland 20892 Patricia M. Price (353) Cancer Research Campaign, Impe- rial College School of Medicine, Hammersmith Hospital, London W12 0NN, United Kingdom Cédric Rebe (55) INSERM U517, Faculty of Medicine, 21000 Dijon, France Julie K. Rhie (301) Division of Cancer Treatment and Di- agnosis, National Cancer Institute, National Institutes of Health, Rockville, Maryland 20892 Azeem Saleem (353) Cancer Research Campaign, Imperial College School of Medicine, Hammersmith Hospital, London W12 0NN, United Kingdom Karen M. Schweikart (301) Division of Cancer Treatment and Diagnosis, National Cancer Institute, National Insti- tutes of Health, Rockville, Maryland 20892 Judith Sebolt-Leopold (31) Pfizer Global Research and Development, Ann Arbor, Michigan 48105 Sachdev Sidhu (237) Department of Protein Engineering, Genentech, Inc., South San Francisco, California 94080 Adaline C. Smith (301) Division of Cancer Treatment and Diagnosis, National Cancer Institute, National Institutes of Health, Rockville, Maryland 20892 Eric Solary (55) INSERM U517, Faculty of Medicine, 21000 Dijon, France; and Hematology Unit, CHU le Bocage, BP1542, 21034 Dijon, France Olivier Sordet (55) INSERM U517, Faculty of Medicine, 21000 Dijon, France Alex Sparreboom (329) Department of Medical Oncology, Rotterdam Cancer Institute, 3075 EA Rotterdam, The Netherlands Caroline J. Springer (137) Cancer Research Campaign Centre for Cancer Therapeutics, Institute of Cancer Re- search, Sutton, Surrey SM2 5NG, United Kingdom Mario Sznol (157) Clinical Affairs, Vion Pharmaceuticals, New Haven, Connecticut 06511 Joseph Tomaszewski (301) Division of Cancer Treatment and Diagnosis, National Cancer Institute, National Insti- tutes of Health, Rockville, Maryland 20892 Akihiro Tomida (77) Institute of Molecular and Cellular Biosciences, University of Tokyo, Tokyo 113-0032, Japan Takashi Tsuruo (77) Institute of Molecular and Cellular Biosciences, University of Tokyo, Tokyo 113-0032, Japan Jaap Verweij (329) Department of Medical Oncology, Rot- terdam Cancer Institute, 3075 EA Rotterdam, The Netherlands Emile E. Voest (123) Laboratory of Medical Oncology, Di- vision of Medical Oncology, University Medical Centre Utrecht, 3584 CX Utrecht, The Netherlands Gregory A. Weiss (237) Department of Protein Engineering, Genentech, Inc., South San Francisco, California 94080 William R. Wilson (269) Auckland Cancer Society Re- search Centre, The University of Auckland, Auckland 1000, New Zealand Paul Workman (249) Cancer Research Campaign Centre for Cancer Therapeutics, Institute of Cancer Research, Sutton, Surrey SM2 5NG, United Kingdom Anne Wotawa (55) INSERM U517, Faculty of Medicine, 21000 Dijon, France Qiang Yu (13) Laboratory of Molecular Pharmacology, Center for Cancer Research, National Institutes of Health, Bethesda, Maryland 20892 Stanley Zucker (91) Department of Veterans Affairs Med- ical Center, Northport, New York 11768 x i i C O N T R I B U T O R S 9863.FM 8/28/01 7:22 AM Page xii C H A P T E R 1 A B R I E F H I S T O R Y O F C A N C E R C H E M O T H E R A P Y Bruce C. Baguley Auckland Cancer Society Research Centre The University of Auckland Auckland, New Zealand Summary 4. Host–Tumor Interactions 1. Introduction 5. Conclusions 2. Genotoxic (Cytotoxic) Therapy References 3. Growth Control Pathways Copyright © 2002 by Academic Press. ANTICANCER DRUG DEVELOPMENT 1 All rights of reproduction in any form reserved. Summary Clinical cancer chemotherapy in the 20th century has been dominated by the development of genotoxic drugs, initiated by the discovery of the anticancer properties of nitrogen mus- tard and the folic acid analogue aminopterin in the 1940s. The development of inbred strains of mice in the early part of the 20th century led to the use of transplantable tumors for the screening of very large numbers of compounds, both nat- ural and synthetic, for experimental antitumor activity. Such screening led to the identification of clinically useful drugs at a rate of approximately one every 2 years. New targets for cy- totoxicity were identified in this program, including tubulin and DNA topoisomerases I and II. The huge expansion in our basic knowledge of cancer has facilitated the development of two new anticancer strategies: the inhibition of specific cel- lular growth pathways and the inhibition of growth of cancer as a tissue. One of the most important principles to emerge is that loss of growth control of cancer cells is mechanistically associated with an increased tendency to undergo pro- grammed cell death, or apoptosis. Thus, cancer growth is a balance between cell birth and cell death. The balance is maintained not only by the genetic status of the cancer cell but by interactions with host cells and extracellular matrix components in the tumor environment. The identification of estrogen as a factor for stimulating the growth led to anties- trogens as therapeutic agents and, more recently, to antago- nists of growth factor receptor-mediated pathways. The early use of bacterial toxins in cancer treatment has led to strate- gies based on host–tumor interactions, such as antiangiogenic and immune approaches. Current research has underlined the enormous complexity not only of growth and death control systems within the tumor cell but of interactions of tumor cells with vascular endothelial, immune, and other cells in cancer tissue. The challenge of future development of low- molecular-weight anticancer drugs is to apply knowledge gained in basic studies to develop new strategies. 1. Introduction It is difficult to assign a date to the beginning of the treat- ment of cancer with drugs because herbal and other prepara- tions have been used for cancer treatment since antiquity. However, the 1890s, a decade that represents an extraordi- narily creative period in painting, music, literature, and tech- nology, encompassed discoveries that were to set the scene for developments in cancer treatment in the 20th century. The discovery of penetrating radiation, or x-rays, by Roentgen in Germany in 1895 was complemented 3 years later by the discovery of radium by Marie and Pierre Curie. The discov- ery of ionizing radiation led not only to radiotherapy as form of cancer treatment but eventually to the development of anticancer drugs that mimicked the effect of radiation by 9863.ch01 8/28/01 7:22 AM Page 1 damaging DNA. The discovery by George Beatson, working in Scotland in 1896, that the growth of a breast cancer could be halted by removal of the ovaries indicated that the growth of cancer cells in the body could be influenced by external factors. This provided the basis for cancer treatment strate- gies that changed the regulation of cancer cell growth. The demonstration by William Coley in 1898 that the administra- tion to cancer patients of a bacterial extract, sterilized by pas- sage through a porcelain filter, caused regressions in lym- phoma and sarcoma indicated that activation of the body’s defense systems might provide a strategy for cancer treat- ment. Each of these three advances lent weight to the bold assumption, made by Paul Ehrlich and others in the early part of the 20th century, that low-molecular-weight drugs might be used in the management of cancer as well as infectious diseases. This chapter considers each of these three ap- proaches in turn. 2. Genotoxic (Cytotoxic) Therapy The first practical anticancer drugs were discovered acci- dentally. One such discovery was an outcome of war, stem- ming from the finding that sulfur mustard gas, used as a toxic vesicant in the First World War, caused myelosuppression. Although gas warfare was not employed in the Second World War, a considerable stock of mustard gas canisters was main- tained in the Mediterranean area. An accident in the Italian port of Bari, involving leakage of one of these canisters, rekindled interest in the myelosuppressive effect of nitrogen mustard, leading to clinical trials in lymphoma patients (Karnofsky et al., 1948; Kohn, 1996). The identification of vitamins as small low-molecular- weight enzyme cofactors was an important biochemical achievement in the early part of the 20th century. The struc- tural elucidation and crystallization of folic acid in 1946 led, as with other isolated vitamins, to studies on its effect on the course of a number of diseases. Unexpectedly, administration to leukemia patients of folic acid and its glutamylated deriv- atives resulted in an increase in tumor growth. While the use of low-folate diets in the management of leukemia was in- vestigated, the development of the folic acid analogue aminopterin provided a significant advance in the manage- ment of childhood acute leukemia (Farber et al., 1948; Bertino, 1979). The link between these two disparate types of drugs and their biological activity was found to be related to their dam- aging effect on DNA. Although Friedrich Miescher had char- acterized DNA as a substance in 1862, the informational complexity and significance to life of DNA was not appreci- ated until the 1940s. The elucidation in 1953 by James Wat- son and Francis Crick of the double-helical structure of DNA had a singular impact on strategies of anticancer drug devel- opment. The cancer chemotherapeutic agent nitrogen mus- tard was found to react chemically with DNA (Kohn et al., 1966). Studies on aminopterin indicated that it interrupted DNA biosynthesis and in so doing caused DNA damage. The next two decades brought a massive development of new drugs that affected the integrity of the cell’s genetic material, with approximately one new drug entering widespread clini- cal use every 2 years. Many of these drugs, which revolu- tionized the treatment of many types of cancer, are shown in Figure 1. A. Development of in Vivo Cancer Screening Systems Developments in chromatography and analytical chem- istry in the first half of the 20th century allowed compounds of defined structure to be isolated from a variety of plants, an- imals, and microorganisms (see Chapter 12). The evolution of synthetic organic chemistry over this time provided anti- cancer drugs in addition to antimicrobial and other medicinal drugs (see Chapter 11). It was quickly realized that it would be impossible to test such a large number of compounds in cancer patients and that some type of model tumor system was required. Transplantable animal cancers became accepted as the best basis for the screening of such drugs. This was made possible by the availability of inbred mouse strains, which had their beginnings in the early part of the 20th cen- tury. Three inbred strains of particular importance to anti- cancer drug screening—DBA, BALB/c, and C57BL—were introduced in 1909, 1916, and 1921, respectively. Sponta- neous or carcinogen-induced tumors in these strains could be transplanted from one inbred mouse to another, allowing re- peated testing of potential anticancer drugs (Stock, 1954). A detailed description of the role of animal testing in drug de- velopment is provided in Chapter 16. In the 1950s and 1960s, most testing programs used the transplantable L1210 and P388 murine leukemia models for primary screening and transplantable solid tumors for more advanced testing (Goldin et al., 1981). The discovery of the athymic “nude” mouse, which had lost its ability to mount a cell-mediated immune response, allowed the testing of new drugs against human tu- mor material growing as xenografts in such mice (Rygaard and Povlsen, 1969). B. Mitotic Poisons Many of the early drugs that were screened for anticancer activity were derived from natural products. The plant prod- uct colchicine, isolated from the autumn crocus, was one of the first of these to demonstrate activity against experimental murine tumor models. It was found to induce arrest of cul- tured cells in mitosis and demonstrated a new mode of in- duction of genomic damage: that of disturbing the correct distribution of genetic material into daughter cells at mitosis. Colchicine, although useful at lower doses for the treatment 2 B A G U L E Y 9863.ch01 8/28/01 7:22 AM Page 2 of gout, proved too toxic for use as an anticancer drug, and early attention focused on the Vinca alkaloids from the peri- winkle plant (Johnson et al., 1963). Two such alkaloids, vin- cristine and vinblastine, had a major impact on the early treatment of patients with malignancy (Rowinsky and Done- hower, 1991). The protein tubulin was identified as the target for colchicine and the Vinca alkaloids (reviewed by Uppuluri et al., 1993) and is the subject of considerable anticancer drug research. The broadening of the spectrum of tumor types susceptible to spindle poisons resulted from the discovery of the taxane class of compounds. Paclitaxel, discovered as a component of some Taxus species (Wani et al., 1971), es- caped detailed investigation until it was found to have a novel biochemical action distinct from that of the Vinca alkaloids, involving promotion rather than inhibition of microtubule as- sembly (Schiff et al., 1979). Paclitaxel (Rowinsky et al., 1990) and docetaxel (Bissery et al., 1991) have a prominent place in cancer therapy today. C. DNA-Reactive Drugs Nitrogen mustard was the basis for the synthesis of a large series of clinically useful derivatives, including melphalan and cyclophosphamide, all found to exert their antitumor ef- fects by alkylation of DNA. Natural products also yielded a number of clinically useful compounds that reacted chemi- cally with DNA, such as mitomycin C (Whittington and Close, 1970), and bleomycin, which required the presence of oxygen and ferric ions to react (Crooke and Bradner, 1976). A par- ticularly important development in DNA-reactive drugs came with the discovery of cisplatin, which had its origins in the chance observation that bacterial growth was inhibited around one of the platinum electrodes of an electrophoresis appara- tus containing ammonium chloride in the buffer (Rosenberg et al., 1965). Platination of DNA became a new mode of DNA damage induction and formed the basis for developing new analogues of cisplatin with reduced host toxicity. D. Inhibitors of DNA Replication A consequence of the elucidation of the structure of DNA was the rational design of analogues of the DNA bases, which were hypothesized to exert their anticancer activity by dis- ruption of DNA replication. These included the thymine ana- logue 5-fluorouracil (Heidelberger et al., 1957) and the purine analogues 6-mercaptopurine and 8-azaguanine (Hitchings and Elion, 1954). The cytotoxic effects of aminopterin and methotrexate were traced to their inhibition, through their A B R I E F H I S T O R Y O F C A N C E R C H E M O T H E R A P Y 3 FIGURE 1 Chronology for the development of some of the anticancer drugs currently in use today. The abbreviations are N mustard (nitrogen mustard), cyt. arabin. (cytosine arabinoside), and BCNU (bischloroethylnitrosourea). Year of introduction 1950 1960 1970 1980 1990 2000 Clinical cytotoxic drugs commonly available 0 5 10 15 20 25 N Mustard Methotrexate Chlorambucil Thioguanine Cyclophosph. Vinblastine 5-Fluorouracil Vincristine Melphalan Daunorubicin Cyt. arabin. Bleomycin Doxorubicin Mitomycin Dacarbazine BCNU Cisplatin Etoposide Teniposide Amsacrine Carboplatin Epirubicin Mitoxantrone Paclitaxel Gemcitabine Docetaxel Topotecan Irinotecan 9863.ch01 8/28/01 7:22 AM Page 3 effect on the enzyme dihydrofolate reductase, of the conver- sion of deoxyuridine monophosphate to thymidine monophos- phate. The phenomenon of “thymineless death,” whereby bacteria unable to synthesize the DNA base thymine died in its absence, was found to have a parallel in mammalian cells and shaped the rationale for the development of the an- timetabolite class of drugs. As the individual enzymes re- sponsible for DNA replication were identified it became clear that the successful operation of the DNA replicase complex relied on a constant supply of the triphosphate precursors of DNA and that interruption of this supply resulted in damage to newly synthesized DNA. Natural products also played a role in the development of anticancer drugs acting on DNA replication. Arabinose nucleosides from the sponge Crypto- thethya were found to have experimental antitumor activity, and one of them, the antimetabolite cytosine arabinoside, has found extensive use in the treatment of persons with leukemia (Ellison et al., 1968). The testing of chemical analogues of cytosine arabinoside led more recently to drugs such as gem- citabine, which has activity against carcinoma (Plunkett et al., 1996). E. DNA Topology as a Target for Drug Development In the 1960s, two logical problems concerning DNA struc- ture became evident. The first was that the unwinding of DNA associated with DNA replication appeared to be ther- modynamically impossible in the time frame involved, since the average chromosome would have tens of millions of he- lical twists. The second, familiar to anyone who has tried to untangle a fishing line, was that the separation of daughter DNA strands produced by DNA replication, prior to cell di- vision, was thermodynamically impossible. Closed circular duplex DNA in some bacterial and mammalian viruses pro- vided smaller molecular weight models for the study of these problems. Such DNA was found to exist in several distinct forms, each with the same sequence but with a different num- ber of helical twists, and these were called topoisomers, based on topology (the branch of mathematics dealing with such differences in shape). In 1976, a new ATP-requiring enzyme, DNA gyrase, with the property of being able to change the topology of closed circular duplex DNA, was discovered in bacterial cells (Gellert et al., 1976). DNA gyrase was found to have an essential role in DNA replication, and because it could pass one strand of DNA through another, it elegantly solved the problems of how DNA could be rapidly unwound during replication and how the daughter DNA strands could be separated after replication. Subsequent studies of mam- malian cells demonstrated two main classes of enzymes. The first, topoisomerase I, changed DNA topology by breaking and rejoining a single DNA strand (Been and Champoux, 1980). The second, topoisomerase II, with some of the char- acteristics of bacterial gyrase, broke both strands of one double-stranded DNA to allow passage of a second double- helical strand through the breakage point (Miller et al., 1981). The discovery of the DNA topoisomerases also solved a problem concerning the activity of a number of natural prod- ucts with anticancer activity that caused DNA damage but did not appear to react chemically. Actinomycin D, identified from Streptomyces cultures, found extensive early clinical use particularly in pediatric tumors (Farber et al., 1960) and was found to bind DNA by intercalating its polycyclic chro- mophore between the base pairs of the DNA double helix (Müller and Crothers, 1968). It was of great biochemical in- terest because of its potent inhibition of RNA synthesis, but this did not appear to explain its antitumor activity. Subse- quently, two anthracycline derivatives, daunorubicin and dox- orubicin, were also found to bind DNA by intercalation of their chromophores, but their effects on RNA synthesis were less than those of actinomycin D. The clinical activity of daunorubicin was generally confined to hematologic malig- nancies but that of doxorubicin was broader (Arcamone, 1985). The synthetic compound amsacrine, which had clini- cal activity against acute leukemia (Arlin, 1989), bound DNA by intercalation of its acridine chromophore (Wilson et al., 1981) but had little or no effect on RNA synthesis. Both am- sacrine and doxorubicin were found to induce covalent links between DNA and proteins (Zwelling et al., 1981), and sub- sequent work demonstrated that this protein was in fact the enzyme topoisomerase II (Nelson et al., 1984). The drugs acted as poisons of this enzyme, subverting its normal func- tion to one of inducing DNA damage. A parallel development in plant natural product research provided podophyllotoxin analogues derived from the man- drake root (Stahelin and Von Wartburg, 1991). Podophyllo- toxin itself, like colchicine, bound to tubulin, but some semi- synthetic glycosidic derivatives, termed epipodophyllotoxins, were found to have superior experimental antitumor activity to podophyllotoxin itself. Etoposide, first tested clinically in 1971, was found to be useful against a variety of malignan- cies (Issell and Crooke, 1979). Investigation of the action of etoposide and of the related drug teniposide revealed that they had reduced binding to tubulin but induced DNA dam- age and poisoned the enzyme topoisomerase II. The plant product camptothecin, which did not bind DNA and previ- ously had no known function, was found to be a specific poi- son for topoisomerase I (Hsiang et al., 1985). Water-soluble analogues of camptothecin, such as topotecan and irinotecan, have clinical potential, and topoisomerase I is now an estab- lished tumor target (Pommier, 1993). F. The Search for Selectivity While the selectivity of radiotherapy was progressively in- creased by localization of the radiation field to specific areas 4 B A G U L E Y 9863.ch01 8/28/01 7:22 AM Page 4 of tumor growth, the selectivity of cytotoxic therapy was de- pendent on particular properties of cancer tissue. The use of microbial models gave rise to the important concept that alkylating drugs killed cells in an exponential fashion, with a certain percentage of the cell population killed with each dose (Pittillo et al., 1965). For some drugs, cytotoxicity was found to be maximal at a particular phase of the cell cycle. Skipper and colleagues (Skipper, 1967) used animal models to select administration schedules with optimal cytotoxicity for the cell-cycle-selective agents, and drug combination schedules that allowed optimal intensity of treatment. The spacing between treatments and the rate of appearance of re- sistant populations could make the difference between suc- cess and failure of treatment (Carl, 1989). Such reasoning was applied with success to hematologic malignancies, which often had a high rate of cell division, but was less successful in the management of solid tumors. Another basis for selectivity was to exploit an enzyme or drug transport mechanism that was present to a different ex- tent in tumor and normal cells. Many antimetabolites were found to exert their selectivity by such mechanisms. Topo- isomerase enzymes provided a particularly good example of such selectivity since high cellular activity, which tended to occur in rapidly dividing cell populations, was associated with greater sensitivity to topoisomerase-directed anticancer drugs (Pommier, 1993). More recently, selectivity has been generated by the development of prodrugs, which have no cy- totoxicity until an enzyme or other agent activates them. Ini- tially, naturally occurring cellular enzymes, such as nitrore- ductases, were considered as candidates for activating prodrugs, but more recently the concept of introducing a dif- ferent activating enzyme by means of a localizing antibody or gene therapy has been investigated. These concepts are discussed in Chapters 8 and 11. 3. Growth Control Pathways While cytotoxic agents dominated the development of clinical cancer chemotherapy, the alternative approach of al- tering the signals that determine cancer growth was not for- gotten. The demonstration by Beatson in 1896 of the role of the ovary in the progression of some types of breast cancer raised the question of whether the growth of all cancer types might be controlled by circulating hormones. Subsequent surgical studies showed that, apart from the case of prostate cancer, removal of endocrine glands was generally ineffective in cancer treatment. It was another 30 years before estrogen, one of the main hormones accounting for Beatson’s result, was identified (Frank et al., 1925), but the biochemical path- ways linking steroid hormones to cell growth stimulation re- mained a mystery. During the 1960s and 1970s, studies of cultured cells, both normal and tumor, indicated that a diverse series of polypeptide growth factors were essential for cell growth, many specific for certain tissue types (reviewed by James and Bradshaw, 1984). An understanding of the action of such factors first required the elucidation of the molecular mechanism of regulation DNA replication and cell division. A. The Cell Cycle Clock One of the most fascinating questions posed by dividing normal and cancer cells was the nature of the molecular clock that instructed the cell as to when it would replicate its DNA and when it would divide. Early studies of cancer tissue iden- tified mitotic cells by their morphology and DNA-synthesizing (S-phase) cells by their uptake of tritium-labeled thymidine, and these phases were found to be separated by periods of cell enlargement, termed G1 and G2 phase (reviewed by Tan- nock, 1978). The first clues to the nature of the oscillator that ran the molecular clock were provided by the discovery in developing sea urchin eggs of a protein termed cyclin. The cellular concentration of this protein increased up to the time of cell division and then abruptly decreased (Evans et al., 1983). Studies of the division of fertilized frog eggs also in- dicated the presence of a cyclin, the synthesis of which was necessary to cell division (Cross et al., 1989). A second com- ponent of the clock was identified from two lines of research, one using frog embryos and one using yeast mutants (Cross et al., 1989). A specific enzyme, termed a cyclin-dependent kinase (cdk), was found both to associate physically with a cyclin and to be activated by it, providing a link between the oscillator and the actuator. Several distinct cdk’s and cyclins were found to be present in mammalian cells. A third com- ponent of the clock mechanism was a protease (in the form of a proteasome) that was responsible for the degradation of the cyclin and thus the resetting of the clock (Glotzer et al., 1991; King et al., 1996). In the general scheme of the cell cy- cle (Fig. 2), the multiple functions required for cell division and DNA replication are exquisitely coordinated by cdk’s 1 and 2, respectively, each activated at the appropriate time by specific associated cyclins. The cell cycle time for human tu- mors varies from around 2 days to several weeks (Wilson et al., 1988). B. Stopping and Starting the Cell Cycle Clock While the majority of the body’s cells are in a nondivid- ing state, certain cells, such as blood cell precursors in the bone marrow and epithelial cells in the gut, are capable of di- viding rapidly. Some mechanism must therefore regulate the passage of cells from a quiescent state to a dividing state. Most early studies utilized cultured fibroblasts to investigate this process. Time-lapse studies (Smith and Martin, 1973) in- dicated that the commitment to DNA replication and mitosis was determined by a stochastic (random) mechanism, and a A B R I E F H I S T O R Y O F C A N C E R C H E M O T H E R A P Y 5 9863.ch01 8/28/01 7:22 AM Page 5 “restriction point” in the G1 phase of the cell cycle was de- fined (Pardee, 1974), past which cells were irreversibly com- mitted. A clear requirement was established for the presence of external polypeptide growth factors to allow passage of cells past the restriction point. Such factors were found to in- teract with membrane-bound surface receptors on target cells, and by the end of the 1970s it was established that at least some of these receptors, including that for epidermal growth factor, became phosphorylated as a consequence of growth factor engagement (Carpenter et al., 1979). Internal cellular proteins were also phosphorylated in response to growth fac- tors, but the identification of the complex linkages between growth factor receptors and commitment to DNA replication and cell division required new findings. A major step forward in the identification of the pathway for initiation of cell growth had its origins in Peyton Rous’s study of cancer-causing viruses in birds, mice, and rats in the early part of the 20th century (Rous, 1983). In 1983, the extraordinary finding was made that a gene in a tumor- transforming virus of simian apes was similar or identical to that specifying a known growth factor for cultured cells (Doolittle et al., 1983; Waterfield et al., 1983). Subsequent work defined a variety of genes that could be transmitted by retroviruses to the tissue of a variety of birds and animals, causing a tumorigenic change. As the functions of these cor- responding gene products were elucidated, it was found that they mapped to biochemical pathways linking the binding of growth factors to the commitment of cells to DNA replica- tion and cell division. The above work led to identification not only of a network of regulatory proteins but also of new cyclins (D-cyclins) and cdk’s (4 and 6) that control the passage of the cell from a qui- escent phase into the cell cycle. The retinoblastoma protein and the transcription factors E2F and c-myc were also impli- cated in a complex control system that resulted in the up- regulation of the E and A cyclins and subsequent initiation of DNA replication. As the elements of this control system were identified, it also became clear that the function of one or more of these elements was defective in cancer. For instance, many human cancers were found to be associated with a mutated ras onco- gene such that the cells behaved as though they were being continuously stimulated by growth factors (Pronk and Bos, 1994). Furthermore, many cancer cells lacked the proper function of proteins, such as the retinoblastoma protein, that regulated entry into S phase (Herwig and Strauss, 1997). C. Programmed Cell Death In 1972, a pivotal hypothesis was advanced that cell death was, like progress through the cell cycle, a product of precise cellular programming (Kerr et al., 1972). This hypothesis was to have a profound effect not only in explaining the loss of cells during the development of the embryo but in ad- vancing our understanding of cancer growth. Apoptosis was found to be an energy-dependent process whereby a cell was converted to fragments that could be absorbed by surround- ing tissue without the initiation of an inflammatory response (Wyllie, 1993). The molecular mechanisms of apoptosis are described in Chapter 4. In a multicellular organism, loss of a single cell is gener- ally unimportant. On the other hand, loss of growth control in a single cell could lead, in the absence of any protective mechanism, to unrestricted and catastrophic growth. The body appears to have a protective mechanism to ensure that any such cells losing growth control are eliminated. The mechanistic links are not yet fully defined, but transcription factors such as E2F and c-myc, which are involved in driving cells into the cell cycle, are also involved in driving cells into apoptosis (Evan and Littlewood, 1998; King and Cidlowski, 1998). Thus, cancer growth is a balance between cell birth and cell death, each initiated by the same pathway (Fig. 3). Cancer cells (as well as normal cells) can be prevented from undergoing apoptosis by so-called survival factors (Evan and Littlewood, 1998), such as insulin-like growth factor 1 (Juin et al., 1999) and cell–matrix interactions (Meredith et al., 1993). D. The Cell Cycle Calendar The pioneering studies of Hayflick showed that when hu- man fibroblasts were cultured, they would die after a certain 6 B A G U L E Y G1 -phase S-phase G2-phase Mitosis DNA replication Cell division Cyclins E/A Cyclin B Cyclin D (pathway often defective in tumor cells) FIGURE 2 The cell cycle clock, with two key alternating processes: DNA replication and cell division. The timing of these processes is controlled pri- marily by oscillations (one per cell cycle) in the cellular levels of cyclin E/A and cyclin B, respectively. Activation of these processes is controlled by cyclin-dependent kinases 2 and 1, respectively. Normal cells have a further control on the decision to enter the cell cycle, timed by cyclin D and acti- vated by cyclin-dependent kinases 4/6. This control system is deficient in cancer cells. 9863.ch01 8/28/01 7:22 AM Page 6 number of generations by a process called senescence (Hayflick and Moorhead, 1961; Hayflick, 1965). The number of gener- ations of doublings, typically 60–80 for fetal cells, progres- sively reduced as the age of the donor of the fibroblasts in- creased, suggesting that some kind of calendar was running during the lifetime of an individual cell. More recently, evi- dence for a slightly different calendar mechanism has been ob- tained in breast epithelial cells (Romanov et al., 2001). Tumor cells in culture, in contrast to normal cells, ap- peared to grow indefinitely, suggesting that the lack of a functioning calendar may be a common characteristic (Gold- stein, 1990). An understanding of the mechanism behind the calendar was provided by research on the question of how a linear chromosome could be replicated. Characterization of the DNA replication machinery indicated that DNA poly- merase had to start from a priming strand, either DNA or RNA, bound to a DNA strand in a duplex conformation (Lingner et al., 1995). Thus, DNA polymerase could not replicate the ends of chromosomes, and with each successive round of DNA replication the chromosomes would become progressively shorter. The solution to this “end-replication” problem was provided by the discovery of the enzyme telom- erase, which added repetitive sequences to the ends or telom- eres of chromosomes (Morin, 1989). Normal tissues, unlike embryonic tissues, generally lack telomerase activity. Thus, loss of telomeric DNA sequences might provide a mechanism to ensure that cells that mutate to allow inappropriate cell di- vision, even if not killed directly by apoptosis, die as a result of telomeric DNA loss. It appears that the acquisition of an active telomerase is a very common feature of cancer tissue (Kim et al., 1994). E. Therapeutic Possibilities One of the earliest therapeutic successes in cell growth control was the discovery of triphenylethylene derivatives that antagonized the effects of estrogens (Ward, 1973). These so- called antiestrogens revolutionized the management of some types of breast cancer. The basis for the antitumor action of antiestrogens is complex and multifactorial. Estrogen was found to bind to an internal protein that acted as a transcrip- tion factor (McKnight et al., 1975), and further work indicated that the products of transcription had multiple effects on tar- get cells, including changes in the expression of growth fac- tor receptors (Curtis et al., 1996). Thus, one possible mecha- nism of action of the antiestrogens is to decrease the action of survival factors and thus increase the rate of cell death. More recently, research suggesting that cancer growth rep- resents a balance between cell birth and cell death suggested new approaches. Thus, either a drug-induced decrease in cell birth rate or an increase in cell death rate might result in tu- mor regression. Low-molecular-weight drugs have been de- veloped that either reduce the progress of cells through the cell cycle, thereby decreasing the cell birth rate, or increase the rate of apoptosis. One important consequence of this re- search was the finding that cytotoxic drugs, discussed in ear- lier sections of this chapter, could increase the rate of apop- tosis. Cisplatin was one of the first cytotoxic drugs demonstrated to induce cultured cells to undergo apoptosis (Barry et al., 1990). Damaged DNA appears to activate spe- cific enzymes, such as Ataxia Telangiectasia Mutated (ATM) kinase, which in turn initiate the series of steps leading to apoptosis (Meyn, 1995). This topic is further discussed in Chapters 2–4. The reinstatement of cancer cell mortality constitutes an- other possible therapeutic approach to cancer treatment. If the “calendar” could be reactivated, such as by inactivation of telomerase activity, cancer cells might once again have a finite lifetime. The design of therapies aimed at inhibiting telomerase activity is still at an early stage and has not been reviewed specifically in this volume. However, a number of excellent reviews can be found (Neidle and Kelland, 1999; Perry and Jenkins Thomas, 1999). 4. Host–Tumor Interactions Cancer cells do not exist as isolated units but as compo- nents of a tissue containing both host and tumor cells. Host cells not only provide, through the vascular endothelium, the precursors of cancer cell growth but also secrete factors that have profound effects on the life and death of tumor cells. The early finding that a mixture of bacterial exotoxins and endotoxins present in “Coley’s toxins” were effective in some types of cancer, particularly lymphomas and sarcomas (re- viewed by Wieman and Starnes, 1994), prepared the way for A B R I E F H I S T O R Y O F C A N C E R C H E M O T H E R A P Y 7 Cell birth Cell death Survival factors FIGURE 3 The balance of cell birth and death in a tumor cell population. In human tumors there is a high rate of turnover such that the overall dou- bling time of tumor tissue is generally considerably longer than the cell cy- cle time of the tumor cells. The rate of cell death is controlled to some ex- tent by external factors called survival factors. 9863.ch01 8/28/01 7:22 AM Page 7