Tumor Models in Cancer Research, Second Edition (Cancer Drug Discovery and Development)

Cancer Drug Discovery and Development Series editor Beverly A. Teicher Genzyme Corporation, Framington, MA, USA For o...

Author: Beverly A. Teicher

34 downloads 973 Views 16MB 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

Cancer Drug Discovery and Development

Series editor Beverly A. Teicher Genzyme Corporation, Framington, MA, USA

For other titles published in this series, go to www.springer.com/series/7625

Beverly A. Teicher Editor

Tumor Models in Cancer Research Second Edition

Editor Beverly A. Teicher Genzyme Corporation Boston, MA USA [email protected]

ISBN 978-1-60761-967-3 e-ISBN 978-1-60761-968-0 DOI 10.1007/978-1-60761-968-0 Springer New York Dordrecht Heidelberg London © Springer Science+Business Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Printed on acid-free paper Humana Press is part of Springer Science+Business Media (www.springer.com)

Preface

Progress in a given field is often dependent upon the development of appropriate, accurate models. In modern times, cancer research has been engaged in a focused search for such models for more than 50 years. The foremost problem in developing such models is that cancer is many, many diseases arising from nearly every tissue and metastasizing to many. A major breakthrough for model in cancer research was the development of transplantable rodent tumors. Many of the early tumor lines were carcinogen-induced, but other arose naturally in elderly animals from inbred strains of mice. These syngeneic tumors grown in the inbred host of origin allowed reproducible tumor growth and reproducible response to anticancer agents to be achieved. These tumor lines also frequently allowed the analysis of tumor metastasis in the host. The mutual needs for as large an array as possible of tumor types and expansion of true inbred strains of mice to carry these tumors lead to the identification of mutant mice with characteristics of deficient immunity suitable for the growth of human tumors as xenografts. The most frequently used of these mutant mouse strains are nude mice and SCID mice. Human tumor xenograft models were established from the many human tumor cell lines developed in the 1970s and 1980s and from fresh tumor explants. Since techniques for genetic manipulation have become more routine, animals expressing “oncogenes” or missing “tumor suppressor” genes have been developed, allowing a new level of understanding of the process of malignancy and new models for testing anticancer agent efficacy. Through the use of these techniques for some diseases and targets, it has been possible to establish specific animal models. Therapeutic index continues to be a critical variable for anticancer agents directed toward any cellular target related to proliferation. Animal models developed to determine potential normal tissue toxicities of new agents as well as the potential of normal tissue protectors have focused on proliferating normal tissues such as mucosa, gut, skin, and bone marrow although cardiac, renal, and lung toxi city can also be modeled. Still, it is the determination of meaningful experimental endpoints that defines the usefulness of models to a field. Increase-in-lifespan (survival) was an endpoint used by Dr. Howard Skipper and colleagues in their groundbreaking murine leukemia studies. Many current models, especially solid tumor models, are not amenable to a survival endpoint; therefore, other measures of tumor v

vi

Preface

response, usually involving tumor volume measurements are applied. Endpoints such as tumor growth delay and tumor growth inhibition closely mimic clinical endpoints, such as response time and time to recurrence. Other endpoints, such as ratio of treated group to control group, log kill, percent apoptosis, and tumor cell survival, depend upon the availability of an untreated or vehicle-treated control group in the experiment. The past 6 years since the first edition of this book have seen great progress in the development of genetically engineered mouse (GEM) models of cancer. These models are finding an important role in furthering our understanding of the biology of malignant disease. A comfortable position for GEM models in the routine conduct of screening for potential new therapeutics is slowly but surely coming. Increasing numbers of genetically engineered mice are available, some with conditional activation of oncogenes, some with multiple genetic changes providing mouse models that are moving closer to the human disease. While we wait for the perfection of the GEMs, the transplantable tumor remains the main resource for drug discovery and efficacy modeling. Though often maligned as models of human disease, antitumor activity in syngeneic mouse tumors and human tumor xenografts is a requirement for most therapeutics prior to entry into development. The criticism directed at these models is frequently a result of the differences between mice and humans. Drug pharmacokinetics in the mouse can be markedly different from pharmacokinetics for the same molecule in other species. The mouse is a remarkably resilient host often able to tolerate much higher doses of experimental therapeutics than human patients, thus allowing blood levels to be reached in mice that cannot be attained in humans frequently leading to disappointing clinical findings. These limitations of the host cannot readily be solved but are limitations which are recognized and are increasingly taken into account in decision making in selecting development candidates. An ideal tumor model would imitate in scale and mirror in response to the human disease. Though no such ideal models exist for the diseases that are cancer, the models described herein represent the efforts of many investigators for many years and approach with closer and closer precision examples that can serve as guides for the selection of agents and combinations for the treatment of human malignancy.

Beverly A. Teicher

Contents

Part I Introduction 1 Perspectives on the History and Evolution of Tumor Models........................................................................................ Shannon Decker and Edward Sausville

3

Part II Transplantable Syngeneic Rodent Tumors 2 Murine L1210 and P388 Leukemias........................................................ William R. Waud 3 Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice................................................................................. Lisa Polin, Thomas H. Corbett, Bill J. Roberts, Alfred J. Lawson, Wilbur R. Leopold III, Kathryn White, Juiwanna Kushner, Stuart Hazeldine, Richard Moore, James Rake, and Jerome P. Horwitz 4 B16 Murine Melanoma: Historical Perspective on the Development of a Solid Tumor Model.......................................... Enrique Alvarez

23

43

79

Part III Human Tumor Xenografts 5 Human Tumor Xenograft Efficacy Models............................................. Ming Liu and Daniel Hicklin

99

6 Imaging the steps of metastasis at the macro and cellular level with fluorescent proteins in real time......................... 125 Robert M. Hoffman 7 Patient-Derived Tumor Models and Explants......................................... 167 Heinz-Herbert Fiebig and Angelika M. Burger

vii

viii

Contents

8 The Pediatric Preclinical Testing Program............................................ 195 Christopher L. Morton and Peter J. Houghton 9 Imaging Efficacy in Tumor Models........................................................ 215 Vinod Kaimal, Wilbur R. Leopold, and Patrick McConville Part IV Carcinogen-Induced Tumors 10 Mammary Cancer in Rats....................................................................... 245 Henry J. Thompson Part V Disease and Target-Specific Models 11 Animal Models of Melanoma.................................................................. 259 Ene T. Fairchild and William E. Carson, III 12 Experimental Animal Models for Investigating Renal Cell Carcinoma Pathogenesis and Preclinical Therapeutic Approaches......................................................................... 287 Gilda G. Hillman 13 Animal Models of Mesothelioma............................................................ 307 Harvey I. Pass, Joseph B. Pincus, Michele Carbone, and Magdalena Plasilova 14 The Use of Mouse Models to Study Leukemia/Lymphoma and Assess Therapeutic Approaches...................................................... 325 William Siders 15 Spontaneous Companion Animal (Pet) Cancers................................... 353 David M. Vail and Douglas H. Thamm Part VI Genetically Engineered Mouse Models of Cancer 16 Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma...................................................................................... 377 Aram F. Hezel and Nabeel Bardeesy 17 Transgenic Adenocarcinoma of the Mouse Prostate: A Validated Model for the Identification and Characterization of Molecular Targets and The Evaluation of Therapeutic Agents...... 397 Sharon D. Morgenbesser

Contents

ix

18 The Utility of Transgenic Mouse Models for Cancer Prevention Research................................................................................ 423 Stephen D. Hursting, Laura M. Lashinger, Powel H. Brown, and Susan N. Perkins Part VII Metastasis Models 19 Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease............................................................................ 447 Suzanne A. Eccles Part VIII Normal Tissue Response Models 20 Animal Models of Toxicities Caused by Anti-Neoplastic Therapy...... 499 Stephen T. Sonis, Gregory Lyng, and Kimberly Pouliot 21 Bone Marrow as a Critical Normal Tissue that Limits Drug Dose/Exposure in Preclinical Models and the Clinic.................. 521 Ralph E. Parchment 22 Anesthetic Considerations for the Study of Murine Tumor Models........................................................................ 553 Thies Schroeder, Siqing Shan, and Mark W. Dewhirst Part IX Experimental Methods and Endpoints 23 Preclinical Tumor Response End Points................................................ 571 Beverly A. Teicher 24 Tumor Cell Survival................................................................................. 607 Sara Rockwell 25 Apoptosis In Vivo..................................................................................... 625 L.C. Stephens, L. Milas, K.K. Ang, K.A. Mason, and R.E. Meyn 26 Transparent Window Models and Intravital Microscopy: Imaging Gene Expression, Physiological Function and Therapeutic Effects in Tumors............................................................... 641 Rakesh K. Jain, Lance L. Munn, and Dai Fukumura Index.................................................................................................................. 681

Contributors

Enrique Alvarez Biomodels, 313 Pleasant street, Watertown, MA 02472, USA K.K. Ang The University of Texas M.D. Anderson Cancer Center, Houston, TX, USA Nabeel Bardeesy Massachusetts General Hospital Cancer Center, Boston, MA, USA Powel H. Brown University of Texas at Austin, Austin, TX, USA Angelika M. Burger School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA Michele Carbone New York University Medical Center, New York, NY, USA William E. Carson, III The Ohio State University, Columbus, OH 43210, USA Thomas H. Corbett School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA Shannon Decker University of Maryland Marlene & Stewart Greenebaum Cancer Center, Baltimore, MD, USA Mark W. Dewhirst Duke University Medical Center, Durham, NC, USA Suzanne A. Eccles McElwain Laboratories, CRC Center for Cancer, Institute of Cancer, Surrey, UK Ene T. Fairchild The Ohio State University, Columbus, OH, USA

xi

xii

Contributors

Heinz-Herbert Fiebig Oncotest GmbH Institute for Experimental Oncology 12. D-79108, Freiburg, Germany Dai Fukumura Department of Radiation Oncology, Massachusetts General Hospital Cancer Center, Boston, MA, USA Stuart Hazeldine School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA Aram F. Hezel Massachusetts General Hospital Cancer Center, Boston, MA, USA Daniel Hicklin Schering-Plough Research Institute, Kenilworth, NJ, USA Gilda G. Hillman Department of Radiation Oncology, School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI 48201, USA Robert M. Hoffman AntiCancer, Inc., and Department of Surgery, University of California, San Diego, CA, USA Jerome P. Horwitz School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA Peter J. Houghton Center for Childhood Cancer, Nationwide Children’s Hospital, Columbus, OH 43205, USA Stephen D. Hursting University of Texas at Austin, Austin, TX, USA Rakesh K. Jain Department of Radiation Oncology, Massachusetts General Hospital Cancer Center, Boston, MA, USA Vinod Kaimal Charles River Laboratories, Ann Arbor, MI, USA Juiwanna Kushner School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA Laura M. Lashinger University of Texas at Austin, Austin, TX, USA Alfred J. Lawson School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI 48108, USA

Contributors

Wilbur R. Leopold III Charles River Laboratories, Ann Arbor, MI, USA Ming Liu Schering-Plough Research Institute, Kenilworth, NJ 07033, USA Gregory Lyng Biomodels, Watertown, MA, USA Katherine A. Mason The University of Texas M.D. Anderson Cancer Center, Houston, TX, USA Patrick McConville Charles River Laboratories, Ann Arbor, MI, USA Raymond E. Meyn The University of Texas M.D. Anderson Cancer Center, Houston, TX, USA Luka Milas The University of Texas M.D. Anderson Cancer Center, Houston, TX, USA Richard Moore School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA Sharon D. Morgenbesser Genzyme Corporation, Framington, MA, USA Christopher L. Morton St. Jude Children’s research Hospital, Memphis, TN, USA Lance L. Munn Department of Radiation Oncology, Massachusetts General Hospital Cancer Center, Boston, MA, USA Ralph E. Parchment Laboratory of Human Toxicology and Pharmacology, SAIC-Frederick Inc., NCI-Frederick, Frederick, MD 21702, USA Harvey I. Pass New York University Medical Center, New York, NY, USA Susan N. Perkins University of Texas at Austin, Austin, TX, USA Joseph B. Pincus New York University Medical Center, New York, NY, USA Magdalena Plasilova New York University Medical Center, New York, NY, USA Lisa Polin School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI 48201, USA

xiii

xiv

Contributors

Kimberly Pouliot Biomodels, Watertown, MA, USA James Rake School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA Bill J. Roberts School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA Sara Rockwell Yale Cancer Center, New Haven, CT, USA Edward Sausville University of Maryland Marlene & Stewart Greenebaum Cancer Center, Baltimore, MD, USA Thies Schroeder Duke University Medical Center, Durham, NC, USA Siqing Shan Duke University Medical Center, Durham, NC, USA William Siders Genzyme Corporation, Framington, MA, USA Stephen T. Sonis Harvard-Farber Cancer Center, Boston, MA, USA; Biomodels, Watertown, MA, USA L.C. Stephens The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA Beverly A. Teicher Genzyme Corporation, Framington, MA, USA Douglas H. Thamm The Animal Cancer Center, Colorado State University, Fort Collins, CO, USA Henry J. Thompson Cancer Prevention Laboratory, Colorado State University, Fort Collins, CO, USA David M. Vail University of Wisconsin, School of Veterinary Medicine, Madison, WI, USA William R. Waud Southern Research Institute, Birmingham, AL 35205, USA Kathryn White School of Medicine, Karmanos Cancer Institute, Wayne State University, Detroit, MI, USA

Part I

Introduction

Chapter 1

Perspectives on the History and Evolution of Tumor Models Shannon Decker and Edward Sausville

Abstract Modern cancer therapeutic research is at crossroads in evolving our approaches to discovering, developing, and entering novel therapeutics into earlystage clinical trials. This chapter endeavors to summarize the customary use and interpretation of animal models used for prioritization of cancer treatments for entry into clinical trials through the end of the last century. We then consider the novel screening paradigms currently in use which exemplify the diverse types of challenging lead compounds for in vivo evaluation. Finally, we offer a strategic overview of steps to maximize utility of the animal model information in selecting agents for clinical study in the twenty-first century. Keywords Targeted in vivo models • Cancer drug development

1.1 Introduction and Statement of the Problem Modern cancer therapeutic research is at crossroads in evolving our approaches to discovering, developing, and entering novel therapeutics into early-stage clinical trials. The sequencing of the human genome [1] and the increasing awareness of the detailed sequence of numerous cancer cell genomes raises the possibility that the empiricism so characteristic of past cancer drug development will give rise to an approach more analogous to current AIDS or cardiovascular disease-related paradigms, where a precise knowledge of the structure of a putative target guides all aspects of a drug’s conceptualization, development, and clinical testing. Yet we have not arrived there yet, as it is currently not feasible in most diseases to employ clinically applicable testing to predict the value of novel agents, outside of fairly specific

E. Sausville (*) University of Maryland Marlene & Stewart Greenebaum Cancer Center, 22 S. Greene St, Baltimore, MD 21201, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_1, © Springer Science+Business Media, LLC 2011

3

4

S. Decker and E. Sausville

examples relevant to antibody-based therapeutics. Indeed, as an example, while the recently observed lack of value of anti-epidermal growth factor antibody therapy in patients with colon cancer with mutated K-ras alleles is understandable post hoc, an appreciation of that reality was apparent only in retrospective analysis [2]. Had it been possible to model reliably that circumstance, it is conceivable that a more efficient and focused development strategy could have been designed. Thus, the challenges facing the use and interpretation of animal models for human cancer drug development center on the predictability of the models in forecasting effects on tumor cells as well as in predicting tolerability of the agent by the host. The model use should occur within a product lifecycle that usually offers no more than 2 or 3 years in most industrial development paradigms after a lead has been identified, and ideally has a direct relevance to how the agent will be studied in early clinical trials. A related issue that will always inject an element of empiricism into the use and interpretation of animal models used in prioritizing human therapeutics for clinical study is the intrinsic unpredictability of animal (usually rodent) vs. human pharmacology and metabolism. While algorithms exist to predict susceptibility to, for example, cytochrome p450 metabolism features [3] or bioavailability [4], since molecules for cancer treatment, at least the classical cytotoxics, are usually employed at close to their maximum tolerated dose (MTD), even minor differences from the human in rodent compound handling parameters (absorption, plasma protein binding, clearance mechanisms, intrinsic susceptibility of host tissues) can translate into decreased relevance of murine dosing and efficacy information as predicting clinical value. Table 1.1 lists points of model departure from rodent vs. human behavior. In contrast, it is interesting to consider that certain classes of agents, particularly monoclonal antibodies with intrinsic anti-signaling of tumor cell tropism properties have for the most part rather reliably defined useful effects that were eventually borne out in humans [5] perhaps in part because of the bland interaction of human antibodies with both mouse and ultimately human physiology, Table 1.1 Potential points of divergence between rodent and human drug features Property Example Plasma protein Camptothecins [61]; result in stabilization of lactone in mice and binding therefore increased perception of activity 7-OH staurosporine [62]; much more avid protein binding in humans prolong half-life and diminish potential for activity Half-life MS-275 [63]; human clearance much slower than mouse; correlates with mice tolerating more frequent dosing schemes while humans do not Intrinsic drug target Neriifolin and cardiac glycosides [64]; murine Na/K ATPase intrinsically susceptibility less susceptible to agents therefore mouse model over-predict capacity for anti-tumor activity H-ras farnesylation intrinsically more sensitive to certain farnesyltransferase inhibitors and therefore not appropriate model for human Ki-ras associated tumors [65] Differing end-organ Bizelisin [66] murine marrow cells intrinsically less susceptible to antisusceptibility proliferative effects than humans therefore under-predict human toxicity

1 Perspectives on the History and Evolution of Tumor Models

5

and the capacity of certain antibodies such as trastuzumab to down-regulate the action of a target to whose action the relevant cell type is “addicted.” This chapter endeavors to summarize the customary use and interpretation of animal models used for prioritization of cancer treatments for entry into clinical trials through the end of the last century. We then consider the novel screening paradigms currently in use which exemplify the diverse types of challenging lead compounds for in vivo evaluation. Finally we will offer a strategic overview of steps to maximize utility of the animal model information in selecting agents for clinical study in the twenty-first century.

1.2 Tumor Models for Cancer Drug Development: Where We Were 1.2.1 Historical Basis Drug treatments for cancer arose from three distinct philosophical points of view. Classically, Ehrlich’s concept of “magic bullets” [6] that would cause toxicity to tumor cells while sparing normal cells arose from the observation that different dyestuffs had obvious affinity for different parts of the cell or different cell types. By this logic, therefore, screening for chemicals that might have a differential effect on tumor cells in comparison to normal cells might be a basis for deriving useful therapeutics related to cancer. A second potential direction was suggested by the observation of profound leucopenia as part of the symptom complex imparted by exposure to mustard gas during World War I. This suggested to some that in lower doses such chemicals might be useful in controlling tumors of (in that case) the hematopoietic system while not ablating all normal marrow elements [7]. Finally the observation that hormonal manipulation could cause useful regression of tumors derived from endocrine responsive organs [8] suggested that an understanding of the biological bases of tumor growth could impart strategies for treatment. This latter point of view, when coupled to the then emerging knowledge of the biochemistry of nucleic acids and the increase particularly in RNA content of tumor cells, led naturally to the efforts to develop what we now call anti-metabolites such as folate antagonists by Farber et al. [9] and purine and pyrimidine analogs by Elion, Hitchings, Heidelberger and a large number of colleagues [10]. Ironically, although such agents are now “lumped” into the category of “cytotoxics,” anti-metabolites were the rationally “targeted” therapeutics of the middle of the last century. The plethora of new chemicals potentially available for cancer treatment, along with relative indifference for cancer as a focus of opportunity by corporate pharmaceutical entities of the time created the perceived need to develop common platforms for evaluation of new molecules available for cancer treatment. This resulted in the evolution of tumor models that were geared for high throughput and mostly employed serially propagated tumor cells in syngeneic hosts. As recounted

6


e lsewhere by Zubrod et al. [11], such screening efforts in academia exemplified by Memorial Sloan Kettering were helpful but to keep up with demand for compound evaluation, Congress in 1955 directed the U.S. National Cancer Institute (NCI) to develop a publicly funded and publicly accessible resource that would promote both clinical testing and pre-clinical evaluation of novel anti-cancer agents. The former initiative was the precursor of the current national Cooperative Group approach to clinical trials. The latter initiative resulted in the formation of the Cancer Chemotherapy National Service Center (CC-NSC), whose “NSC” accession catalog of compounds continues to the present day at NCI’s Developmental Therapeutics Program as the successor to the CC-NSC. Compounds studied by NCI for the most part were synthesized by contractors or solicited from academic or commercial parties by an active compound acquisition program [12]. Encouragement to industry as well as academic participants was provided by confidentiality agreements that assured protection of the submitting party’s intellectual property. Results generated by the NCI screening effort could then be the basis for development of the compound to clinical trials sponsored either by the NCI through its Cooperative Groups or privately funded ventures.

1.2.2 Early Screening Models The models employed in efforts at NCI and at academic screening centers included and were exemplified by the L1210 and P388 mouse leukemias serially transplanted by the peritoneal route and treated by intraperitoneal injection of drug. The endpoint of the screening assay was survival of the treated vs. untreated or vehicle-treated groups of animals. A compound was considered to show preliminary evidence of activity if the mean or median lifespan of the treated animals was increased by 125%, with the control group survival set at 100%, and with the important caveat that “positive” compounds had to have acceptable therapeutic index with evidence of maintained or increasing body mass in treated animals and no untoward short-term toxic phenomena. Among the advantages of this model as a screening tool were its relative speed, with experiment evaluation generally complete by 2–3 weeks; capacity for high throughput allowing many compounds to be evaluated; and reproducibility of the model owing to high take rate and uniform growth rate. Using these and related models, important clinically relevant principles of cancer chemotherapeutic development were elucidated and formed the basis for construction of human chemotherapy regimens and practices. These principles include the demonstration that active agents produced with each dose increment reduction in the tumor cell population a reduction in tumor cell mass by logarithmically increasing increments. This led to the concept that valuable agents had to be applied in successive “cycles” to cause tumor-free animals to emerge. The inverse relationship of tumor cell inoculum to curability at a constant dose led likewise to the theoretical underpinnings of “adjuvant” treatment programs [13, 14].


7

Potential pitfalls of such models were numerous. Most obviously was the “same site” nature of the treated space, without a physiological barrier between the administration site and the locus of drug action. A second concern was the potential lack of relevance of such models to solid tumors. Both of these concerns were partially addressed by the use of syngeneic metastasizing murine models such as Lewis lung carcinoma or B16 melanoma. These models could be run either as solid allografts, with treatment intravenously or orally in a way that mimicked human treatment, or following a period of residence in a body part, generally an extremity, removal of the “primary” tumor could allow for observation of compound activity against the establishment or formation of metastases. Although such models were valuable adjuncts to evaluating positive compounds in the murine leukemia studies, an emerging concern throughout the later 1970s was that the paucity of agents emerging through such murine leukemia-based screens that ultimately had robust activity in human solid tumors. The limited value of agents detected in murine leukemia screening models when applied to human solid tumors resulted in enormous interest in the use of immunocompromised animals to study xenografts of human tumors through technology that was first applied on a large scale commencing ~1980 using athymic “nude” mice [15]. One initial hope was that agents thereby revealed to be active would be intrinsically more suitable for use in human solid tumors. An immediate problem in the use of these models, however, is their intrinsically less efficient throughput owing to a variety of factors including the mechanics of implanting and sizing tumors in a subcutaneous site; the fact that different human tumor cell lines had intrinsically different “take rates” and variable growth rates. This encouraged the development of prioritization criteria often after in vitro screening to assure that compounds entering into in vivo study already had evidence of cytotoxic potential. The “NCI 60” cell line panel is representative of one such large-scale effort of this type whose historical basis and output has been described elsewhere [16, 17]. Moreover, criteria for value of an agent in athymic mouse xenografts are problematic in that tumor growth delay is more frequently encountered than actual responses of established tumors, and the meaning of this to the clinical setting remains undefined in a precise way to this day. Looking at the performance of predominantly classical cytotoxic agents studied at the NCI in a variety of murine syngeneic and prototypic human xenograft systems, one can conclude that agents irrespective of their level of in vitro activity which have activity in less than 33% of the models tested had no “positive” phase 2 clinical trials. In contrasts, agents with activity in at least 33% of such models had an approximately 50% likelihood of positivity in phase 2 clinical trials [18]. Noteworthily, there was little histology-specific correlation of activity in models with activity in the clinic. As described above, the reason for this disconnect between animal and human experiences when ultimately understood has in the examples cited in Table 1.1 largely related to differences in animal and human pharmacological features or target susceptibility or importance to the host organism. This and related experiences [19] has reinforced that from a purely stochastic viewpoint there is value in prioritizing compounds for entry into the clinic by their behavior in some number of animal models.

8


The most profoundly dissatisfying aspect of this set of outcomes is that there is no tie on the part of the in vivo models used to evaluate the majority of screening experiences to the biology of the tumors studied. While one might argue that this is reasonable in light of the fact that most of the chemical entities selected from random screening experiences were not really designed around any key mechanism as relevant to the biology of a particular tumor, the present age has for the most part moved past the point where high enthusiasm for a compound arises solely by virtue of its behavior in a screening system. Rather novel approaches to cancer drug screening are generating lead compounds that will require distinctive approaches to further elicitation of activity in vivo, and will ideally be coupled to novel strategies to apply in early clinical trials.

1.3 Novel Screens Beget Novel In Vivo Model Challenges Traditionally as discussed above, lead compounds were selected for study in vivo based on evidence in vitro or expectation of cytotoxicity. The molecular target era has allowed the creation of a flood of new screening models. Importantly, some screens are aimed at identifying targets or pathways of interest as an initial step in then defining the effect of a compound on the target(s) or pathway(s) of interest, but not necessarily tied to initial evidence of cytotoxicity. Whether action of a lead against the target in one of these in vitro or non-traditional assay systems is enough to justify proceeding with the lead to in vivo models discussed throughout this volume is a key strategic issue to consider. These assay systems run the gamut from non-mammalian in vivo models in an array of organisms, to informationintensive screens capitalizing on the explosion of new “data mining” technologies, to cell-based in vitro assays looking for non-classical endpoints such as angiogenesis or invasion.

1.3.1 Non-mammalian Models In the last 10 years, efforts utilizing non-mammalian models to actually identify targets and drugs have proliferated. In some cases, such as for yeast and Drosophila, the organisms have been used for many decades as biological models, but have not traditionally served as a source of anti-cancer leads. Other organisms such as zebrafish have arisen relatively recently as models. 1.3.1.1 Unicellular Yeast screens have been widely used in cell biology and genetics studies. It was the first organism to have its genome sequenced [20]. Yeast strains are easily grown


9

and manipulated, allowing for facile studies of DNA damage repair, cell cycle progression and checkpoint control, among other well recognized utilities. Yeast screens have provided a facile way to identify sets of genes that contribute to sensitivity or resistance to particular drugs. For instance, for the synthetic tripeptide arsenical GSAO that inhibits angiogenesis and targets actively dividing but not quiescent endothelial cells, Hogg et al. identified 88 GSAO-sensitive Saccharomyces cerevisiae deletion strains by screening a genome-wide set of 4,546 such strains, thus identifying potential molecular targets of GSAO and allowing for confirmatory studies in mammalian cells [21]. Classical pathways well explored in yeast actually from a biological point of view have defined strains with alterations in cell cycle and cell cycle checkpoint control, particularly in response to DNA damage [22]. This observation has been capitalized on by numerous groups to screen for compounds that interfere with cell cycle control, thus potentially enhancing sensitivity to classical DNA-damaging chemotherapy or radiation therapy. The National Cancer Institute (NCI) Yeast Anticancer Drug Screen has screened tens of thousands of compounds in selected yeast strains mutated for cell cycle control or DNA damage repair [23]. One limitation of such screens though is the possibility that larger organisms do not rely on a single mechanism for repairing DNA damage. For example, mammals in some cases appear to have checkpoint-independent mechanisms for surviving radiation [24], so a compound identified in yeast as interfering with a checkpoint may be ineffective as a radiation sensitizer in humans. Thus, yeast serve to illustrate the caveat that a mammalian relevant in vivo model may need to be carefully constructed to provide evidence that the yeast-related screen output is an accurate reflection of a human circumstance. 1.3.1.2 Multicellular In an attempt to overcome some of the shortcomings of unicellular organism screens as predictors of in vivo activity, various groups have developed in vivo models in non-mammalian organisms ranging from zebrafish to nematodes to flies. The potential advantages of such screens generally are that they are cheaper and proceed more quickly than mammalian in vivo models, but still have the capacity to provide information about the ability of a drug lead to act in a live host. Drosophila strains have been used for over a century for genetic studies, and have a relatively small genome, making it an attractive model for studying various biological processes. A number of different cancer-related screening campaigns have now been run in Drosophila models, including transgenic models. For example, extending from the observation discussed above that whole organisms have checkpoint-independent mechanisms for surviving DNA damage from chemotherapy and radiation, Tin Su et al. ran a pilot screen for radiation sensitizers using wild-type and checkpoint mutants [24]. Drug candidates were mixed into food and placed in wells with Drosophila larvae, and survival was determined by counting the empty pupae cases. In another screening context, by looking at phenotypic

10


changes from Drosophila developing leg imaginal discs, Phanstiel et al. screened for drug–polyamine conjugates with polyamine transporter (PAT)-selective targeting ability, deriving from the observation that PAT is elevated in many tumor types and hypothesizing that drug–polyamine conjugates may be able to selectively attack tumor cells [25]. While the limited genetic redundancy of Drosophila lends itself to phenotypic endpoints and is part of the basis of its attractiveness as a model, it is also potentially a limitation of the model as the hits identified in such models may fail in more complex mammalian systems where redundancy is more frequent. The Caenorhabditis elegans nematode has been used as a model system for several decades. The worm goes from egg to fertile adult in 3 days, and each adult can produce 300 progeny making it a quick and inexpensive model system. Numerous knockout mutants exist and strains can be frozen for decades [26]. In one recent cancer application, Salgia et al. described a C. elegans nematode model in which transgenic worms were generated harboring either wild-type c-Met or mutations of c-Met commonly seen in lung cancer [27]. The worms expressing the mutant c-Mets consistently displayed the phenotypic outputs of abnormal vulval development and low fecundity. While this model can be used to investigate the role of gene mutations in a whole organism, invertebrates may not be appropriate models for certain cancer-related processes such as apoptosis due to their lesser complexity [28]. Avian embryo models have also been used in developmental biology for many years, but only more recently in cancer research with any frequency, most likely due in part to the recent sequencing of the chick genome. Advantages include the speed of the model in reproducing human tumor growth and angiogenesis. Researchers have validated that human glioblastoma grafted onto the chorioallantoic membrane (vascularized extra embryonic tissue; CAM) displays similar patterns of gene expression changes as the human disease [29]. Although they still have efficiency advantages over mouse models, CAM models also have disadvantages over other vertebrate non-mammalian systems such as a relatively lengthy assay (~10 days), higher cost than other models and difficulties in quantitation of the output [30]. Proponents of Xenopus tadpole models point to rapid extra uterine development, the transparency of developing tadpoles, permeability of the skin, and similarities to mammals in certain organ development, anatomy and physiology as advantages [31]. To identify molecules affecting angiogenesis and lymphangiogenesis, Brandli et al. screened 1,280 compounds in a Xenopus model looking first for edema as a phenotype and then used whole-mount in situ hybridization of Xenopus embryos to visualize blood and lymphatic vessel development for the 66 positive hits from the initial stage of the screen, with confirmatory endothelial cell proliferation and tube formation assays then conducted on the second level hits. The original Xenopus model, Xenopus laevis has a pseudotetraploid genome and a relatively long generation time, making the development of stable transgenic lines lengthy relative to other non-mammalian models, however work has also been done to use the diploid Xenopus tropicalis as a model for experimental genetics [32].


11

Zebrafish have been cited as having numerous advantages for screening, many of them shared with the advantages of Xenopus models above. The assays, while in vivo, are still relatively quick at approximately 3 days, relatively cheap, and have reasonably high throughput as they can be done on plates [9]. As demonstrated by Willett et al. using known angiogenesis inhibitors TNP470 and SU5416 [30], zebrafish, being transparent, lend themselves very well to angiogenesis-related assays as blood vessel formation can be assessed by visual inspection. Zebrafish have been used for models of drug sensitization and resistance. Transgenic models have been generated as well. Much less is known, however, about cancer-relevant issues such as DNA repair enzymes and the orthologs of human oncogenes and tumor suppressor genes in zebrafish than other model systems [33].

1.3.2 Technology-Intensive Screening Advances in fields such as computing technology, imaging, robotics, and miniaturization among others have helped spawn a range of new screening possibilities. All of these technology-intensive methodologies produce a wealth of information much more quickly than many classical screening techniques, but the challenge is in sifting through and capitalizing on the information. In many cases in vivo models applicable to the output of such screens will need to be constructed as a dedicated effort in parallel with the design and output of the ex vivo screen. 1.3.2.1 High-Throughput Screening High-throughput screening (HTS) methods became increasingly necessary as the number of potential molecular targets for cancer drugs grew virtually exponentially. In one possible format for an HTS assay, the activity of an enzyme is linked to an easily readable output, such as fluorescence or bioluminescence from luciferase. Cell-based HTS is also possible, many times with cell lines that have been transfected with a receptor or promoter of interest. Methodologies for HTS campaigns have been discussed extensively [34, 35] and the literature abounds with results from campaigns directed against particular enzymatic targets. In the fortunate circumstance where the role of the enzyme in a biologicalpathway relevant to human disease is well understood, where structural biology can show the development candidate interacting with the binding site of the enzyme, where the candidate has favorable drug-like characteristics, and where the action of the drug on the target can be tracked in cell culture and in vivo models, the path for development can be relatively straightforward. In the case where the output of a screening campaign using a cell-based assay where pathway activation or inhibition is the ultimate readout, caution must be urged in exploring activity in in vivo models prior to deconvolution of the lead compounds mechanism of action.

12


1.3.2.2 Chemogenomics Unlike a single HTS assay that has the ability to screen many compounds against a single target, chemogenomics represents the integration of a study of the effects of compounds on biological targets with modern genomics technologies, attempting to comprehensively discover and describe all possible drugs to all possible drug targets [36]. For instance, for a chemical genetics application, the function of proteins is probed by small molecules by adding a library of small molecules to cells, selecting those that produce the phenotype of interest and identifying the protein bound by the molecules [37]. Many of the newer applications of yeast screens fall into the chemogenomics category, helping to identify genes that can help explain the activity of known compounds [2]. 1.3.2.3 Proteome and Kinome Screens With the success of genome-wide screens, efforts next logically extended down to the proteome and a particular target class such as protein kinases (thus a “kinome” directed virtual screen) in the search for drug targets. In one such effort, Schreiber et al. combined a chemical genetics screen that identified small molecule modifiers of rapamycin activity with a probe of a yeast proteome chip to identify proteins that bound the small molecules [38]. One potential advantage of probing of the proteome over traditional affinity chromatography is the bias of chromatography toward high-abundance proteins. Some approaches have elected to limit the probe to the kinases rather than the whole proteome. Dagorn et al. screened the human kinome for all kinases involved in pancreatic cancer cell survival and gemcitabine resistance, identifying a set of potential targets for drug discovery campaigns [39]. Comprehensive screening of the whole yeast proteome has been undertaken to systematically identify protein–protein interactions, in an effort that might eventually assist in the development of small molecules that can disrupt key interactions [40]. Analysis of such protein–protein interaction data sets however requires significant bioinformatics resources, and the complexity will only increase when multicellular organism proteomes are screened. 1.3.2.4 Nanotechnology Considerable effort has been expended in recent years on integrating nanotechnology with more traditional biologically based methodologies. In one series of approaches nanoparticles such as quantum dots or magnetofluorescent particles are conjugated to peptides, antibodies, or small molecules to allow the targeting of the nanoparticle to specific cells, such as tumor cells. Some groups have had success in using such bioconjugates for imaging [41] and have demonstrated differential cellular uptake [42]. Others are using nanoparticles to produce formulations of compounds, ones with excellent in vitro activity but no systemic bioavailability, in an effort to make


13

such compounds clinically viable [43]. As a therapeutic approach however, these bioconjugates remain unproven clinically and numerous scientific, cost, and regulatory hurdles exist. 1.3.2.5 RNA Interference Since its discovery approximately a decade ago, RNA interference (RNAi) has found application in many aspects of cancer drug discovery including target identification and validation, identification of drug resistance and sensitization mechanisms, and synthetic lethal screening. Genome-wide RNAi screens have been used successfully in C. elegans and Drosophila to understand biological processes and work toward a comprehensive characterization of gene function [44]. For example, Woo et al. identified “driver genes” in hepatocellular carcinoma, each of which can now be considered for screening to define hepatocellular carcinoma-related drugs [45]. Iorns et al. suggest the utility of conducting chemical genetics and RNAi screens in parallel to simultaneously identify small molecule inhibitors and targets, giving as an example their use of an RNAi screen to identify the PDK1 pathway as a determinant of sensitivity to tamoxifen coupled with a screen to locate chemical inhibitors of the pathway [46]. Synthetic lethal screening is another potential application of RNAi. Two genes are “synthetic lethal” when cell death results from mutation of both genes even though the cell remains viable with mutation of either alone. One recent demonstration of the relevance of siRNA-related synthetic lethal screens arose from the observation that cells deficient in BRCA-1 were highly sensitive to concomitant PARP inhibition [47], based on the inability to repair DNA lesions utilizing homologous recombination.

1.3.3 In Vitro Models Cell-based in vitro models with cytotoxic endpoints that had the goal of identifying compounds for subsequent in vivo testing were used for several decades as primary screens (e.g., the NCI60 described above). More recently, cell-based models are being employed to either further filter hits from the high throughput and mammalian models discussed above or to look for other endpoints such as angiogenesis. As discussed above, yeast screens have been employed to identify compounds that act against yeast strains with specific genetic mutations that are believed to be relevant to cancer. The number of hits obtained from such assays though still requires further filtering before an in vivo mammalian model can be contemplated. In vitro cell-based models, particularly those where activity in a knockout cell line can be compared to the wild-type, can act as a further filter. For instance, Lamb et al. used a three-stage screen to first identify compounds inhibiting the growth

14


of double-strand break repair-deficient yeast cells, producing 28 hits, which were winnowed by looking for toxicity proportional to levels of topoisomerase I or II expression [48]. They then screened the remaining eight hits in two lines of chicken pre-B-cell line DT40, one wild-type and the other defective in doublestrand break repair. Other in vitro assays assess look at endpoints other than cytotoxicity, such as endothelial cell migration or cord formation, looking for compounds that affect processes such as angiogenesis or metastasis. By themselves such assays are not necessarily sufficient to warrant pursuit of in vivo models with identified compounds. In combination with other results, however, endothelial cell assays can be a source of lead compounds. For instance, Sekhar et al. combined observations from a chemistry-driven drug discovery screen for inhibitors of endothelial cell tubule formation with biochemical pathway screening and shRNA suppression to identify compounds to pursue as drug leads, and also validate ENOX1 as a target for enhancing radiation response of tumors [49]. Other endothelial cell strategies have looked to capitalize on differences between tumor and normal endothelial cells. Camussi et al. identified cyclic peptides that showed specific binding only to tumor but not normal endothelial cells to use as a mechanism for delivering antiangiogenic agents only to the tumor [50]. The integrin inhibitors can serve as an example, however, of how action on a molecular target coupled with endothelial cell assays for angiogenesis endpoints may not be enough to guarantee a drug candidate worthy of development. Screening for inhibitors of integrins, adhesion molecules considered important in angiogenesis, has been conducted in conjunction with numerous other angiogenesis assays [51]. In this case, while data existed to support the search for integrin inhibitors, certain of the integrins are promiscuous and the biology considerably more complicated than suggested in primary screening assays, such that development of an integrin inhibitor has been thus far unsuccessful [52].

1.4 Tumor Models for Cancer Drug Development: Where We Need to Be The above examples emerging from modern biology-driven potential cancer relevant screens illustrate the wide diversity of premises that need to be embodied in the in vivo models that might ultimately be used to further evaluate the value of such lead molecules in vivo. Given the fact that many of these leads may not be intrinsically cytotoxic but directed to particular targets, either directly in a molecular sense or as part of a pathway a readout which formed the basis of the screening efforts, how would the clinical development process be informed and fortified by knowledge gleaned from in vivo models exploring the activity of these agents? Following is a suggested series of steps that might be considered in the practice of in vivo models using such leads with the goal of pre-clinical evaluation of such a “targeted” compound. As illustrated in Table 1.2, it differs from the path that


15

Table 1.2 Distinction between cytotoxic and “targeted” in vivo model usage Classical Targeted Maximum targeted dose driven Biologic dose bracketing an optimized concentration Pharmacology frequently deferrable Early PK and PD crucial and build into correlates of clinical value Number of models active key to prioritization Limited number of models, but target enriched Need to define host cell susceptibility Need to define effect on host target in relation to toxicity observed

might have been applied to the traditional “cytotoxic” drug candidate of the last century in that the latter agents were generally developed according to an MTD model where pharmacological information could reasonably be obtained after initial confirmation of in vivo activity on a particular schedule. In contrast, most efficient and useful development steps for targeted agents would have a more early integration of pharmacological information, both kinetic and dynamic, into the early development strategy, and may actually not embark on evaluation of uncharacterized models with respect to target expression or pathway activation status. With this reasoning, the following steps might be usefully be allied to the process of in vivo model use with screening leads in the age of biologically tailored cancer drug screens.

1.4.1 “In Vitro” Area Under the Concentration × Time Curve for Target Effect In a range of cell culture models expressing the target, definition of the time until target modulation as a function of compound addition and removal, and the relationship of this to secondary endpoints such as cytotoxicity is critical, and helps to define initial dosing strategies. Ideally controls with respect to secondary endpoints would include cell lines not expressing the relevant target or pathway. This would also provide valuable information about “off target” effects.

1.4.2 Qualification of Compound for In Vivo Study A series of related molecules active in vitro can be further qualified for in vivo study by application of algorithms suitable for selection of oral bioavailability [53], if continuous exposure is the intended strategy. Alternatively, “cassette” type dosing schemes [54] allow preliminary assessment of pharmacological properties of a series, thereby narrowing choices of molecules for in vivo evaluation.

16


1.4.3 Initial Rodent Pharmacology and Model Selection Using a realistic dosing scheme, attempts to recreate at least the area under the concentration × time curve (AUC) defined in vitro by the results of studies described above using non-tumored animals should then occur. Transition to the use of tumored animals would initially use a tumor model with cells known to be dependent on the function of the target for growth, viability, or some easily assessed biologic readout. These may express the target endogenously or heterologously; in the latter event appropriate vector alone controls are necessary. In the event the target is expressed endogenously, consideration of a cell type related to the first where the target is absent or not functional would be an additional useful control.

1.4.4 Sample Size and Randomization of Animals Several considerations go into selecting the number of animals chosen for control and experimental groups, and consultation of a biostatistical expert in designing the experiments is useful. In part the sample size relates to the magnitude of the effect desired and the nature of the endpoint [55]. In the event that tumor is to be assessable at the initiation of the experiment, randomization of animals with different tumor sizes so that treatment groups are matched with respect to initial tumor size may be necessary.

1.4.5 Correlative Studies Ideally evaluation of efficacy in “hitting” the putative target should accompany in vivo evaluation of the compound, as well as in a most ideal case assessment of the pharmacologic properties of the agent achieving that effect (dose–response of effect on target in association with usual parameters such as plasma maximal concentration (Cmax), half life (t1/2), AUC, etc.). Determination of tumor drug levels corresponding to these phenomena would be a plus. Examples of successful integration of such information obtained in early in vivo studies with value when applied to the clinic would include bortezomib anti-tumor effect correlated with effect on proteosome inhibition [56] or more recently effect of dasatinib on bcr-abl kinase substrate phosphorylation in relation to plasma concentrations in mice [57], a set of observations which assisted initial clinical development.

1.4.6 Additional Desirable Studies While one intensively evaluated model (with respect to pharmacodynamics and pharmacology) may be useful in setting the “boundary conditions” and expectations for benchmarking initial compound use and performance in humans, particularly


17

if the tumor system studied is “artificial” with respect to the anticipated state of the target or pathway of interest in the clinic (e.g., heterologously expressed or otherwise manipulated cells), enthusiasm for the compound is increased generally if a range of non-manipulated cell types are exposed to the agent at the appropriate concentration dose and range with confirmation that in that circumstance there are expected effects on target function and consequences for cellular physiology. It may not be necessary to develop stable in vivo models from each cell type; such techniques as the “hollow fiber assay” [58] can be a way of usefully assessing in vivo effect without the time and expense of deriving independent models [59].

1.5 Conclusion The ultimate goal of in vivo model studies in the pre-clinical development of anticancer agents is to serve a variety of interests. First, from a strictly pragmatic standpoint, demonstration of unbiased, well understood in vivo activity serves to increase confidence in investing the considerably more time-consuming and expensive effort in developing the safety database to allow human early phase clinical testing. Valuable activity in an in vivo model should reflect pharmacological “action at a distance” across physiological and anatomical barriers in a way that has an acceptable therapeutic index on the clinical proposed dose range and schedule. Second, the in vivo model experience from a scientific standpoint becomes that which the early clinical trials would ideally seek to emulate precisely as a “mirror image” accurate reflection. Third, from an ethical standpoint, clear demonstration of in vivo activity on the part of a candidate anti-cancer agent is a basis for potentially justifying in a prospective patient’s mind their participation in such a study. Although recent studies have documented that modern phase I anti-cancer drug clinical trials are extremely safe and for many of the newer molecular entities have the prospect of benefit in perhaps as much as 30% of participants [60], the initial in vivo experiences in animals can serve as a talking point in assuring potential participants that there is the possibility of benefit at doses and schedules that have a modicum of expected safety and tolerability. The ideal for in vivo model use in therapeutics development is the assembly of a package of information that will guide in the design and ultimate interpretation of the initial human clinical trial. Conversely, it is also conceivable that once an initial appreciation of achieved human pharmacology emerges from the results of the initial early phase clinical trials in humans, a focused return to in vivo animal models with the intention of conscientiously modeling the achieved human pharmacology in the animals may allow a more realistic strategy to emerge before committing to an extensive human phase 2 program. In this way in vivo animal models can contribute not only to the initial qualification of a compound for human use but also to a more refined way of advancing it to having its best chance for positive later-stage clinical trial efforts.

18


References 1. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, et al. The sequence of the human genome. Science. 2001;291:1304–51. 2. Liévre A, Bachet JB, Le Corre D, Boige V, Landi B, Emile JF, et al. KRAS mutation status is predictive of response to cetuximab therapy in colorectal cancer. Cancer Res. 2006;66:3992–5. 3. Iwatsubo T, Hirota N, Ooie T, Suzuki H, Shimada N, Chiba K, et al. Prediction of in vivo drug metabolism in the human liver from in vitro metabolism data. Pharmacol Ther. 1997;73:147–71. 4. Bachmann KA, Ghosh R. The use of in vitro methods to predict in vivo pharmacokinetics and drug interactions. Curr Drug Metab. 2001;2:299–314. 5. Pegram MD, Lipton A, Hayes DF, Weber BL, Baselga JM, Tripathy D, et al. Phase II study of receptor-enhanced chemosensitivity using recombinant humanized anti-p185HER2/neu monoclonal antibody plus cisplatin in patients with HER2/neu-overexpressing metastatic breast cancer refractory to chemotherapy treatment. J Clin Oncol. 1998;16:2659–71. 6. Himmelweit F. ed. The collected papers of Paul Ehrlich, Vol. I. London: Pergamon; 1956. Originally published in 1900. 7. DeVita VT Jr, Carbone PP. Treatment of Hodgkin’s disease. Med Ann Dist Columbia. 1967;36:232. 8. Beatson GT. On the treatment of inoperable cases of carcinoma of the mamma: suggestions for a new method of treatment with illustrative cases. Lancet. 1896;2:104–7. 9. Farber S, Diamond LK. Temporary remissions in acute leukemia in children produced by folic acid antagonist, 4-aminopteroyl-glutamic acid. N Engl J Med. 1948;238:787–93. 10. Hitchings GH, Elion GB, Falco EA, Russell PB, Vanderwerff H. Studies on analogs of purines and pyrimidines. Ann N Y Acad Sci. 1950;52:1318–35 11. Zubrod CG, Schepartz SA, Carter SK. Historical background of the National Cancer Institute’s drug development thrust. Natl Cancer Inst Monogr. 1977;45:7–11. 12. Thambi P, Sausville EA. Working with the National Cancer Institute. In: Teicher BA, Andrews PA, editors. Anticancer drug development guide: preclinical screening, clinical trials, and approval. 2nd ed. 2004. Humana press, Totowa, NJ. 13. Skipper HE, Schabel FM Jr, Wilcox WS. Experimental evaluation of potential anticancer agents. XIII. On the criteria and kinetics associated with “curability” of experimental leukemia. Cancer Chemother Rep. 1964;35:1–111. 14. Skipper HE, Schabel FM Jr, Wilcox WS. Experimental evaluation of potential anticancer agents. XIV. Further study of certain basic concepts underlying chemotherapy of leukemia. Cancer Chemother Rep. 1965;45:5–28. 15. Plowman J, Dykes DJ, Hollingshead MG, Simpson-Herren L, Alley MC. Human tumor xenograft models in NCI drug development. In Teicher BA, editor. Anticancer drug development guide: preclinical screening, clinical trials, and approval; 1997. Totowa, NJ: Humana. p. 101–25. 16. Alley MC, Scudiero DA, Monks A, Hursey ML, Czerwinski MJ, Fine DL, et al. Feasibility of drug screening with panels of human tumor cell lines using a microculture tetrazolium assay. Cancer Res. 1988;48:589–601. 17. Monks A, Scudiero D, Skehan P, Shoemaker R, Paull K, Vistica D, et al. Feasibility of a highflux anticancer drug screen using a diverse panel of cultured human tumor cell lines. J Natl Cancer Inst.1991;83:757–66 18. Johnson JI, Decker S, Zaharevitz D, Rubinstein LV, Venditti JM, Schepartz S, et al. Relationships between drug activity in NCI preclinical in vitro and in vivo models and early clinical trials. Br J Cancer. 2001;84:424–1431 19. Voskoglou-Nomikos T, Pater JL, Seymour L. Clinical predictive value of the in vitro cell line, human xenograft, and mouse allograft preclinical cancer models. Clin Cancer Res. 2003;9:4227. 20. Williams N. Genome projects: yeast genome sequence ferments new research. Science. 1996;272:481.


19

21. Dilda PJ, Don AS, Tanabe KM, Higgins VJ, Allen JD, Dawes IW, et al. Mechanism of selectivity of an angiogenesis inhibitor from screening a genome-wide set of Saccharomyces cerevisiae deletion strains. J Natl Cancer Inst. 2005;97:1539–47. 22. Weinert TA, Hartwell LH. The RAD9 gene controls the cell cycle response to DNA damage in Saccharomyces cerevisiae. Science. 1988;241:317–22. 23. http://dtp.cancer.gov/yacds/default.html accessed 10/7/10. 24. Jaklevic B, Uyetake L, Lemstra W. Chang J, Leary W, Edwards A, et al.Contribution of Growth and Cell Cycle Checkpoints to Radiation Survival in Drosophila. Genetics 2006;174:1963–72. 25. Tsen C, Iltis M, Kaur N, Bayer C, Delcros J-G, Laurence von Kalm L, et al. A Drosophila model to identify polyamine−Drug conjugates that target the polyamine Transporter in an Intact Epithelium. J Med Chem. 2008;51:324–30. 26. Youds JL, Barber LJ, Boulton SJ. C. elegans: a model of Fanconi anemia and ICL repair. Mutat Res. 2009;668:103–16. 27. Siddiqui SS, Loganathan S, Krishnaw S, Faoro L, Jagadeeswaran R, Salgia R. C. elegans as a model organism for in vivo screening in cancer: effects of human c-Met in lung cancer affect C. elegans vulva phenotypes. Cancer Biol Ther. 2008;7:856–63. 28. Krumschnabel G, Podrabsky JE. Fish as model systems for the study of vertebrate apoptosis. Apoptosis. 2009;14:1–21. 29. Hagedorn M, Jave4rzat S, Gilges D, de Lafarge B, Eichmann A, Bikfalvi A. Accessing key steps of human tumor progression in vivo by using an avian embryo model. Proc Natl Acad Sci USA, 2005;102:1643–8. 30. Serbedzija GN, Flynn E, Willett CE. Zebrafish angiogenesis: a new model for drug screening. Angiogenesis. 1999;3(4):353–9. 31. Kalin RE, Banziger-Tobler NE, Detmar M, Brandli AW. An in vivo chemical library screen in Xenopus tadpoles reveals novel pathways involved in angiogenesis and lymphangiogenesis. Blood. 2009;114:1110–22. 32. Beck CW, Slack JMW. An amphibian with ambition: a new role for Xenopus in the 21st century. Genome Biol. 2001;2:reviews1029.1–reviews1029.5. 33. Spitsbergen JM, Kent ML. The state of the art of the zebrafish model for toxicology and toxicologic pathology research – advantages and current limitations. Toxicol Pathol. 2003; 31(Suppl):62–87. 34. Inglese J. Quantitative high-throughput screening: A titration-based approach that efficiently identifies biological activities in large chemical libraries. Proc Natl Acad Sci USA, 2006;103:11473–8. 35. Shoemaker RH, Scudiero DA, Melillo G, Currens MJ, Monks AP, Rabow AA, et al. Application of high-throughput, molecular-targeted screening to anticancer drug discovery. 2002;2:229–46. 36. Caron PR, Mullican MD, Mashal RD, Wilson KP, Su MS, Murcko MA. Chemogenomic approaches to drug discovery. Curr Opin Chem Biol. 2001;5:464–70. 37. Stockwell BR. Chemical genetics: ligand-based discovery of gene function. Nat Rev Genet. 2000;1:116–25. 38. Huang J, Zhu H, Haggarty SJ, Spring DR, Hwang H, Jin F, Snyder M, Schreiber SL. Finding new components of the target of rapamycin (TOR) signaling network through chemical genetics and proteome chips. Proc Natl Acad sci USA. 2004;101:16594–9. 39. Giroux V, Iovanna J, Dagorn JC. Probing the human kinome for kinases involved in pancreatic cancer cell survival and gemcitabine resistance. FASEB J. 2006;20:1982–91. 40. Schwikowski B, Uetz P, Fields S. A network of protein–protein interactions in yeast. Nat Biotechnol. 2000;18:1257–61. 41. Gao X, Nie S. Molecular profiling of single cells and tissue specimens with quantum dots. Trends Biotech. 2003;21:371–3. 42. Weissleder R, Kelly K, Sun EY, Shtatland T, Josephson L. Cell-specific targeting of nanoparticles by multivalent attachment of small molecules. Nat Biotechnol. 2005;23:1418–23. 43. Mukerjee A, Vishwanatha JK. Formulation, characterization and evaluation of curcuminloaded PLGA nanospheres for cancer therapy. Anticancer Res. 2009;29:3867–75.

20


44. Kuttenkeuler D, Boutros M. Genome-wide RNAi as a route to gene function in Drosophila. Brief Funct Genomics Proteomics. 2004;3:168–76. 45. Woo HG, Park ES, Lee JS, Lee YH, Ishikawa T, Kim YJ, Thorgeirsson SS. Identification of potential driver genes in human liver carcinoma by genomewide screening. Cancer Res. 2009;69:4059–66. 46. Iorns E, Lord CJ, Ashworth A. Parallel RNAi and compound screens identify the PDK1 pathway as a target for tamoxifen sensitization. Biochem J. 2009;417:361–70. 47. Farmer H, McCabe, N, Lord CJ, Tutt AN, Johnson DA, Richardson TB, et al. Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature. 2005;434:917–21. 48. Dunstan HM, Ludlow C, Goehle S, Cronk M, Szankasi P, Evans DR, et al. Cell-based assays for identification of novel double-strand break-inducing agents. J Natl Cancer Inst. 2002;94:88–94. 49. Geng L, Rachakonda G, Morré DJ, Morré DM, Crooks PA, Sonar VN, et al. Indolylquinuclidinols inhibit ENOX activity and endothelial cell morphogenesis while enhancing radiation-mediated control of tumor vasculature. FASEB J. 2009;23:2986–95. 50. Bussolati B, Grange C, Tei L, Deregibus MC, Ercolani M, Aime S, et al. Targeting of human renal tumor-derived endothelial cells with peptides obtained by phage display. J Mol Mef. 2007;85:8970906. 51. Belvisi L, Riccioni T, Marcellini M, Vesci L, Chiarucci I, Efrati D, et al. Biological and molecular properties of a new alpha(v)beta3/alpha(v)beta5 integrin antagonist. 2005;4:1670–80. 52. Hodivala-Dilke K. alphavbeta3 integrin and angiogenesis: a moody integrin in a changing environment. Curr Opin Cell Biol. 2008;20:514–9. 53. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. 1997; 23:3–25. 54. Raynaud FI, Fischer PM, Nutley BP, Goddard PM, Lane DP, Workman P. Cassette dosing pharmacokinetics of a library of 2,6,9-trisubstituted purine cyclin-dependent kinase 2 inhibitors prepared by parallel synthesis. Mol Cancer Ther. 2004;3:353–62. 55. Hollingshead MG. Antitumor efficacy testing in rodents. J Natl Cancer Inst. 2008;100:1500–10. 56. Adams J, Palombella VJ, Sausville EA, Johnson J, Destree A, Lazarus DD, et al. Proteasome inhibitors: a novel class of potent and effective antitumor agents. Cancer Res. 1999;59:2615–22. 57. Luo FR, Yang Z, Camuso A, Smykla R, McGlinchey K, Fager K, et al. Dasatinib (BMS354825) pharmacokinetics and pharmacodynamic biomarkers in animal models predict optimal clinical exposure. Clin Cancer Res. 2006;12:7180–6. 58. Decker S, Hollingshead M, Bonomi CA, Carter JP, Sausville EA. The hollow fibre model in cancer drug screening: the NCI experience. Eur J Cancer. 2004;40:821–6. 59. Zabludoff SD, Deng C, Grondine MR, Sheehy AM, Ashwell S, Caleb BL, et al. AZD7762, a novel checkpoint kinase inhibitor, drives checkpoint abrogation and potentiates DNAtargeted therapies. Mol Cancer Ther. 2008;7:2955–66. 60. Horstmann E, McCabe MS, Grochow L, Yamamoto S, Rubinstein L, Budd T, et al.. Risks and benefits of phase 1 oncology trials, 1991 through 2002. N Engl J Med. 2005;352:895–904. 61. Supko JG, Malspeis L. Pharmacokinetics of the 9-amino and 10,11-methylenedioxy derivatives of camptothecin in mice. Cancer Res. 1993;53:3062–9. 62. Sparreboom A, Chen H, Acharya MR, Senderowicz AM, Messmann RA, Kuwabara T, et al. Effects of alpha1-acid glycoprotein on the clinical pharmacokinetics of 7-hydroxystaurosporine. Clin Cancer Res. 2004;10:6840–6. 63. Acharya MR, Sparreboom A, Sausville EA, Conley BA, Doroshow JH, Venitz J, et al. Interspecies differences in plasma protein binding of MS-275, a novel histone deacetylase inhibitor. Cancer Chemother Pharmacol. 2006;57:275–81. 64. Lin Y, Dubinsky WP, Ho DH, Felix E, Newman RA. Determinants of human and mouse melanoma cell sensitivities to oleandrin. J Exp Ther Oncol. 2008;7:195–205. 65. Stayrook KR, McKinzie JH, Barbhaiya LH, Crowell PL. Effects of the antitumor agent perillyl alcohol on H-Ras vs. K-Ras farnesylation and signal transduction in pancreatic cells. Anticancer Res. 1998;18:823–8. 66. Volpe DA, Tomaszewski JE, Parchment RE, Garg A, Flora KP, Murphy MJ, et al. Myelotoxic effects of the bifunctional alkylating agent bizelesin on human, canine and murine myeloid progenitor cells. Cancer Chemother Pharmacol. 1996;39(1–2):143–9.

Part II

Transplantable Syngeneic Rodent Tumors

Chapter 2

Murine L1210 and P388 Leukemias William R. Waud

Abstract L1210 and P388 leukemia models have been extensively used over the last 50 years. The models are rapid, reproducible, and relatively inexpensive (in comparison to human tumor xenograft and transgenic models). However, as with any experimental animal tumor model, there are limitations. Neither leukemia is a satisfactory model for either human cancer in general or human leukemia in particular. Despite the limitations of murine leukemia models, these models have been useful in making progress in anticancer drug development, in the development of a number of therapeutic principles, and in understanding the biologic behavior of tumor and host. These models are still useful today in conducting detailed evaluations of new candidate anticancer drugs (e.g., schedule dependency, route of administration dependency, formulation comparison, analog comparison, and combination chemotherapy). The greatest utility of murine leukemias today is derived from evaluations of drug-resistant sublines for crossresistance and collateral sensitivity. Crossresistance data, coupled with knowledge of resistance mechanisms operative in drug-resistant leukemias, may yield insights into mechanisms of action of agents. Similarly, crossresistance data, coupled with mechanisms of action of various agents, may yield insights into resistance mechanisms operative in drugresistant leukemias. Furthermore, crossresistance data may identify potentially useful guides for patient selection for clinical trials of new antitumor drugs. Keywords Murine leukemia • L1210 • P388 • Drug resistance

W.R. Waud (*) Southern Research Institute, 2000 9th Avenue South, Birmingham, AL 35205, USA e-mail: [email protected]

B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_2, © Springer Science+Business Media, LLC 2011

23

24

W.R. Waud

2.1 Introduction Mouse leukemia models were a central component of the initial drug discovery programs employed by the Division of Cancer Treatment and Diagnosis (DCTD) of the National Cancer Institute (NCI) during the early 1960s and 1970s. The L1210 and P388 leukemias, developed in 1948 [1] and 1955 [2], respectively, played a major role in both screening and detailed evaluations of candidate anticancer agents. Today, 50 years later, these models are still used to evaluate anticancer activity, although at a greatly reduced level, and to study mechanisms of drug resistance. This chapter reviews their past contributions and updates their present role in the evaluation of anticancer drugs. Data for the drug sensitivity of these two leukemias and various drug-resistant P388 sublines to clinically useful drugs are summarized.

2.2 Role in Drug Screening Spontaneous tumors in animals were first used as models for screening potential anticancer agents. In fact, these types of studies occurred even prior to the beginning of the twentieth century [3] and provided the basis for modern drug screening programs. However, large-scale screening and the ability to conduct detailed drug evaluation studies with anticancer agents increased greatly in the 1920s by the development of inbred strains of mice that allowed investigators to propagate tumor lines by serial transplantation in vivo [4]. The United States Congress became interested in cancer research when it was recognized in the 1940s that systemic cancer could be influenced by drug treatment. This was demonstrated at Memorial Sloan-Kettering, which was one of the first of several institutions in the United States and Europe that began drug screening programs. In that program, the mouse sarcoma SA-180 was used as its screening model. However, as drugs exhibited anticancer activity and the supply of new candidate agents exceeded the screening capacity of that program, the need for additional drug development capability became apparent. With this impetus, Congress directed NCI to implement a national drug development program, which went into effect in 1955 as the Cancer Chemotherapy National Service Center (CCNSC). Initially, the CCNSC primary screening program consisted of L1210 leukemia, SA-180, and mammary adenocarcinoma 755 [5]. Over the years, the composition of the primary screen changed several times, i.e., from the original three tumors to L1210 and two arbitrarily selected tumors; to L1210 and Walker 256 carcinosarcoma; to L1210 and P388 leukemia; and finally to L1210, P388, and either B16 melanoma or Lewis lung carcinoma [6]. Several additional models were also used during this period for special detailed drug evaluation. The primary screening program underwent a major change in 1976 when DCTD incorporated the use of three human tumor xenograft models. The new screen now

2 Murine L1210 and P388 Leukemias

25

consisted of a panel of colon, breast, and lung tumors, both murine and human. However, all drugs going through this screen were still evaluated initially for activity against the sensitive P388 leukemia model [7]. During this period, the low number of drugs discovered with marked antitumor activity against human solid tumors led to a radical change in the screening program that had used murine leukemia models as the primary screen. In the mid-1980s, NCI developed a new primary screen based on the use of established human tumor cell lines in vitro [8]. The new and old screen programs were to be conducted in parallel so as to permit a comparison; however, in early 1987, budget cuts at NCI forced an end to largescale P388 screening [9].

2.3 Characteristics Both L1210 and P388 leukemias were chemically induced in a DBA/2 mouse by painting the skin with methylcholanthrene [1, 2]. Propagation of the leukemia lines is in the host of origin by intraperitoneal (i.p.) implantation of diluted ascitic fluid containing either 105 (L1210) or 106 (P388) cells per animal. Testing is generally conducted in a hybrid of DBA/2 (e.g., CD2F1 or B6D2F1), because the hybrids are somewhat heartier. However, DBA/2 mice may be used for special studies and should be used for serial in vivo propagation of the leukemias. Frequently used implant sites are i.p., subcutaneous (s.c.), intravenous (i.v.), or intracerebral (i.c.). For L1210 leukemia with an implant of 105 cells, the median days of death and the tumor doubling times for these implant sites are 8.8, 9.9, 6.4, and 7.0 days and 0.34, 0.46, 0.45, and 0.37 days, respectively. For P388 leukemia with an implant of 106 cells, the median days of death and the tumor doubling times for these implant sites are 10.3, 13.0, 8.0, and 8.0 days and 0.44, 0.52, 0.68, and 0.63 days, respectively. Skipper and coworkers at Southern Research Institute determined the rate of distribution and proliferation of L1210 leukemia cells using bioassays of untreated mice after i.p., i.v., and i.c. inoculation [10]. Following i.p. inoculation, most of the L1210 cells were found in the ascites fluid of the peritoneal cavity. Using the median day of death as the evaluation time point, the most infiltrated tissues were the bone marrow, liver, and spleen. Following i.v. inoculation, the majority of L1210 cells appeared in the bone marrow. On the median day of death from the i.v. implant, the most infiltrated tissues were also the bone marrow, liver, and spleen. After i.c. inoculation, most of the L1210 cells remained in the brain (for 3–5 days). On the median day of death from the i.c. implant, the spleen was heavily infiltrated (the extent of the leukemia in other tissues was not reported). Southern Research was one of the first institutions to become involved in the CCNSC screening program and was heavily involved in designing protocols for the program. One aspect essential to the operation of a screening program is the development of appropriate parameters for measuring antitumor activity. At Southern Research antitumor activity for leukemia studies is assessed on the basis of percent

26

W.R. Waud

median increase in life span (% ILS), net log10 cell kill, and long-term survivors. Calculations of net log10 cell kill are made from the tumor cell population doubling time that is determined from an internal tumor titration consisting of implants from serial tenfold dilutions [11]. Long-term survivors are excluded from calculations of % ILS and tumor cell kill. To assess tumor cell kill at the end of treatment, the survival time difference between treated and control groups is adjusted to account for regrowth of tumor cell populations that may occur between individual treatments [12]. The net log10 cell kill is calculated as follows:

Netlog10 cellkill =

(T − C) − (durationof treatmentindays) 3.32 × Td

where (T – C) is the difference in the median day of death between the treated (T) and the control (C) groups, 3.32 is the number of doublings required for a population to increase 1-log10 unit, and Td is the mean tumor doubling time (in days) calculated from a log-linear least-squares fit of the implant sizes and the median days of death of the titration groups.

2.4 Sensitivity to Clinical Agents Many of the clinically useful compounds in current use were first detected in the murine leukemia models. The sensitivities of L1210 and P388 leukemias (implanted i.p.) to most of these agents (administered i.p.) are shown in Figs. 2.1 and 2.2 and Figs. 2.3 and 2.4, respectively. Overall, P388 leukemia is somewhat more sensitive than L1210 leukemia. For alkylating agents, the sensitivities are similar. Notable exceptions are chlorambucil, mitomycin C, and carboplatin, for which P388 is markedly more sensitive. For antimetabolites, the sensitivities are also similar. Exceptions are floxuridine (P388 being markedly more sensitive) and hydroxyurea (L1210 being markedly more sensitive). For DNA-binding agents, P388 leukemia is clearly more sensitive (e.g., actinomycin D, mithramycin, daunorubicin, teniposide, doxorubicin, and amsacrine). For tubulin-binding agents, P388 leukemia is again clearly more sensitive. The vinca alkaloids are active against P388 leukemia but ineffective against L1210 leukemia. Although most of the sensitivity data are for i.p. implanted leukemia and i.p. administered drug, valuable information can be obtained from separating the implant site and the route of administration. Table 2.1 shows the activity of melphalan, administered i.p., against both L1210 and P388 leukemias implanted i.p., i.v., and i.c. Melphalan given i.p. is very effective against both i.p. implanted leukemias. The activity is reduced to less than one-half when the implant site is changed to i.v. The activity is further reduced when the implant site is changed to i.c.; however, melphalan is able to cross the blood–brain barrier to some extent. This principle is shown more extensively with the data in Figs. 2.5 (L1210) and 2.6 (P388) for the leukemias implanted i.c. and various clinically useful agents administered i.p. Thiotepa, CCNU,


27

NSC No. Rx Drug Alkylating Agents 750

D

Nitrogen Mustard 762

A

Chlorambucil

3088

D

Thiotepa

6396

A

Melphalan

8806

A

Hexamethylmelamine Cyclophosphamide Mitomycin C

13875

A

26271

A

26980

A

Dacarbazine

45388

A

Procarbazine

77213

D

CCNU

79037

A

Streptozotocin

85998

A

Busulfan

109724

A

Cisplatin

119875

A

Chlorozotocin

178248

A

Carboplatin

241240

D

BCNU

409962

A

Ifosfamide

Tubulin Binders Vinblastine

49842

G

Vincristine

67574

B

Miscellaneous Agents Gallium nitrate

15200

D

Bleomycin

125066

D

-5

-4

-3

-2

-1

0

1

2

3

4

5

6

7

8

Net Cell Kill (log10 units)

Fig. 2.1 Sensitivity of i.p. implanted L1210 leukemia to clinically useful alkylating agents, tubulin binders, and other miscellaneous agents. L1210 leukemia (105 cells except for hexamethylmelamine, which used 106 cells) was implanted i.p. on day 0. Beginning on day 1, the agents were administered i.p. using the indicated schedules. Treatment schedule (Rx): A = day 1; B = day 1, 5, 9; C = day 1–5; D = day 1–9; E = day 1, 4, 7, 10; F = q3h × 8, day 1, 5, 9; G = day 1–15

BCNU, and ara-C/palmO-ara-C, administered i.p., exhibit comparable activity against either i.p. or i.c. implanted leukemias. In addition to melphalan, cisplatin, cyclophosphamide, ifosfamide, and 6-mercaptopurine (L1210) have reduced activity when the implant site is changed to i.c. Several agents become inactive when the implant site is changed to i.c. (e.g., methotrexate (P388), 5-fluorouracil, floxuridine, actinomycin D, vincristine, doxorubicin, and etoposide). Some comparisons using different treatment schedules can be misleading. Even though all values have been expressed as net cell kill (i.e., corrected for the treatment schedule), one schedule can

28

W.R. Waud

NSC No. Drug Antimetabolites

Rx

Methotrexate

740

C

6-Thioguanine

752

A D

6-Mercaptopurine 755

A D

5-Fluorouracil

19893

C

Floxuridine

27640

D

Hydroxyurea

32065

F

Azacytidine

102816

D

2-Chlorodeoxy- 105014 adenosine 135962 PalmO-ara-C Deoxycoformycin 218321

A

Trimetrexate

249008

C

Fludarabine

312887

D

F F

DNA Binders Actinomycin D Mithramycin

3053

A

24559

D

82151

D

Teniposide

122819

B

Doxorubicin

123127

A

Etoposide

141540

B

Amsacrine

249992

D

Idarubicin

256439

D

Mitoxantrone

301739

D

Daunorubicin

-4

-3

-2

-1

0 1 2 3 4 Net Cell Kill (log10 units)

5

6

7

Fig. 2.2 Sensitivity of i.p. implanted L1210 leukemia to clinically useful antimetabolites and DNA binders. L1210 leukemia (105 cells, except for hydroxyurea, which used 104 cells and 6-thioguanine (day 1 only treatment) and daunorubicin, which used 106 cells) was implanted i.p. on day 0. Beginning on day 1 (day 2 for daunorubicin), the agents were administered i.p. using the indicated schedules. Treatment schedule (Rx): see legend for Fig. 2.1

be optimal, whereas another schedule is suboptimal. For nitrogen mustard, no conclusion can be drawn from the data about its ability to cross the blood–brain barrier. The agent is active against the i.p. implanted leukemia using a single i.p. injection (optimal) and is inactive against the i.c. implanted leukemia using 15 daily i.p. injections (suboptimal). This is further illustrated by chlorambucil which is active against i.c. implanted L1210 (using a single i.p. injection) and inactive against i.p. implanted L1210 (using nine daily i.p. injections). From work with these screening models, it became apparent that drug sensitivity was, in some cases, heavily dependent on drug concentration and exposure

2 Murine L1210 and P388 Leukemias Drug

29

NSC No. Rx

Alkylating Agents Busulfan

750

C

Nitrogen Mustard

762

C

Chlorambucil

3088

D

Thiotepa

6396

A

Melphalan Hexamethylmelamine Cyclophosphamide Mitomycin C

8806

A

13875

D

26271

A

26980

A

Dacarbazine

45388

A

Procarbazine

77213

D

CCNU

79037

A

85998

B

Ifosfamide

109724

A

Cisplatin

119875

B

Chlorozotocin

178248

B

Carboplatin

241240

B

BCNU

409962

A

Streptozotocin

Tubulin Binders Vinblastine

49842

B

Vincristine

67574

B

Paclitaxel

125973

C

Vinorelbine

608210

B

Gallium nitrate

15200

D

Mitotane

38721

D

Bleomycin

125066

D

Levamisole

177213

D

Miscellaneous

-3

-2

-1

0

1

2

3

4

5

6

7

8

Net Cell Kill (log10units)

Fig. 2.3 Sensitivity of i.p. implanted P388 leukemia to clinically useful alkylating agents, tubulin binders, and other miscellaneous agents. P388 leukemia (106 cells except for CCNU, which used 107 cells) was implanted i.p. on day 0. Beginning on day 1 (day 2 for CCNU, streptozotocin, and chlorozotocin), the agents were administered i.p. using the indicated schedules. Treatment schedule (Rx): see legend for Fig. 2.1

time which, in turn, was impacted by the in vivo treatment schedule. As an example, studies conducted with 1-β-d-arabinofuranosylcytosine (ara-C) pointed out the need for concentration and time of exposure studies. Using L1210 leukemia in mice, it was shown that the optimal dosage and schedule for ara-C was 15–20 mg/kg/dose, given every 3 h for eight doses, then repeated three times at 4-day intervals [13]. This regimen was “curative.” The single-dose LD10 for mice was between 2,500 and 3,000 mg/kg, and using a single dose within that range

30

W.R. Waud

Drug

Antimetabolites

NSC No.

Rx

Methotrexate

740

D

6-Thioguanine

752

D

6-Mercaptopurine

755

D

5-Fluorouracil

19893

D

Floxuridine

27640

D

Hydroxyurea

32065

F

Azacytidine

102816

D

PalmO-ara-C

135962

A

Deoxycoformycin 218321

D

Trimetrexate

249008

D

Fludarabine

312887

D

Gemcitabine

613327

E

DNA Binders 3053

A

Mithramycin

24559

D

Daunorubicin

82151

A

Teniposide

122819

B

Doxorubicin

123127

A

Etoposide

141540

B

Amsacrine

249992

B

Idarubicin

256439

B

Mitoxantrone

301739

B

Actinomycin D

-3

-2

-1

0

1

2

3

4

5

6

7

8


Fig. 2.4 Sensitivity of i.p. implanted P388 leukemia to clinically useful antimetabolites and DNA binders. P388 leukemia (106 cells) was implanted i.p. on day 0. Beginning on day 1, the agents were administered i.p. using the indicated schedules. Treatment schedule (Rx): see legend for Fig. 2.1

Table 2.1 Activity of melphalan administered as a single i.p. injection against L1210 and P388 leukemias implanted i.p., i.v., and i.c. Net cell kill (log10 units) Site Inoculum size L1210 P388 i.p. 106 4.7 >6.5 i.v. 106 2.0 2.9 1.2 2.4 i.c. 104

2 Murine L1210 and P388 Leukemias Drug

NSC No.

31

Rx

Alkylating Agents Busulfan

750 A

Nitrogen Mustard

762 G

Chlorambucil

3088 A

Thiotepa

6396 A

Melphalan

8806 A

CCNU Streptozotocin

79037 A 85998 D

Cisplatin

119875 A

BCNU

409962 A

Antimetabolites Methotrexate 6-Mercaptopurine

740 A 755 A

5-Fluorouracil

19893 D

Hydroxyurea

32065 F

Ara-C

63878 D

A

PalmO-ara-C

135962 A

DNA or Tubulin Binders Actinomycin D Vinblastine Daunorubicin Etoposide

3053 A 49842

G

82151

A

141540

B

-5

-4

-3

-2

-1

0

1

2

3

4

5

6


Fig. 2.5 Sensitivity of i.c. implanted L1210 leukemia to clinically useful agents. L1210 leukemia (104 cells except for CCNU, which used 105 cells) was implanted i.c. on day 0. Beginning on day 1 (day 2 for busulfan, chlorambucil, thiotepa, melphalan, hydroxyurea (single injection), cisplatin, BCNU, and daunorubicin), the agents were administered i.p. using the indicated schedules. Treatment schedule (Rx): see legend for Fig. 2.1

would effect a 3-log10-unit reduction in L1210 cells but was not “curative.” Although these in vivo results might give the appearance of a concentrationdependent effect, in vitro studies had clearly shown that cell kill of L1210 in culture was time dependent at the higher concentration levels employed. The apparent concentration dependence observed in vivo over a range of single doses had resulted from the extended time of exposure that resulted from those extremely high dosage levels.

32

W.R. Waud Drug

NSC No.

Rx

Alkylating Agents Thiotepa

6396

A

8806

A

Cyclophosphamide

26271

A

Procarbazine

77213

D

CCNU

79037

A

Ifosfamide

109724

A

Cisplatin

119875

A

BCNU

409962

A

Methotrexate

740

D

6-Thioguanine

752

D

Melphalan

Antimetabolites

755

D

19893

D

276401

D

63878

D

135962

A

6-Mercaptopurine 5-Fluorouracil Floxuridine Ara-C PalmO-ara-C

DNA or Tubulin Binders Actinomycin D Vincristine Doxorubicin

3053

A

67574

B

123127

B

-4

-3

-2

-1

0

1

2

3

4

5

6

7


Fig. 2.6 Sensitivity of i.c. implanted P388 leukemia to clinically useful agents. P388 leukemia (104 cells except for ifosfamide, methotrexate, 6-thioguanine, 6-mercaptopurine, 5-fluorouracil, and floxuridine, which used 103 cells and CCNU and ara-C, which used 105 cells) was implanted i.c. on day 0. Beginning on day 1 (day 2 for ifosfamide), the agents were administered i.p. using the indicated schedules. Treatment schedule (Rx): see legend for Fig. 2.1

2.5 Predictive Value Many investigators have questioned the use of experimental leukemias as primary screening models over the years. Some have argued that since L1210 or P388 leukemia was used for many years as the initial screening model, continued evaluation of compounds emerging from this screening configuration, even using solid tumor models for secondary evaluation, would only produce antileukemic drugs [14]. If compounds active against solid tumors were being missed by the primary screen


33

composed of leukemias, it would appear reasonable that in order to obtain agents that are active against specific tumor types or solid tumors in general, then the primary screen should consist of specific tumor types or solid tumors. Even though this would appear to be a reasonable approach, it will depend on whether or not there are existing agents or that agents can be developed that will selectively kill specific cancer histotypes. The correlation between drugs active against L1210 or P388 leukemia and solid experimental tumor models has not been good. For example, only 1.7% of 1,493 agents that were active against P388 leukemia were also active against murine Lewis lung carcinoma. Further, only 2% of 1,507 agents active against P388 leukemia were also active against murine colon 38 adenocarcinoma. Finally, only 2% of 1,133 agents that were active against P388 leukemia were also active against human CX-1 (HT29) colon tumor. However, when comparing leukemias, a less than expected correlation was obtained – only 15% of 1,564 active agents against P388 leukemia were also active against L1210 leukemia [15]. One observation often referred to is that there are drugs active against experimental solid tumors that are inactive against P388 leukemia. For example, 15% of 84 agents that were inactive against P388 leukemia were active against at least one of eight solid tumors tested [15]. Flavone acetic acid has been cited as an example [14]. This compound was inactive in the initial P388 screen even though it was later shown to exhibit activity against the leukemia when the appropriate treatment schedule was used [16]. This example points out a problem with large-scale screening programs in that it is not logistically feasible to conduct preliminary schedule dependency trials. One other observation is that there are experimental solid tumors (e.g., murine pancreatic 02 ductal adenocarcinoma) that are not responsive in vivo to any clinically used agents, including many P388-active agents [14]. It may be noted, however, that this tumor is sensitive to numerous clinical agents in vitro after a 24-h exposure [17], suggesting that the in vivo insensitivity of this tumor may not be due to cellular characteristics but rather physiological or architectural constraints of the animal. Southern Research has evaluated a spectrum of compounds in the i.p. implanted P388 model in order to evaluate this model as a predictor for the response of human tumor xenografts to new candidate antitumor agents (unpublished data). The P388 data collected were compared to the data for various s.c. implanted human tumor xenografts, which were selected on the basis of the results of the NCI in vitro screen. In general, compounds that were active against P388 leukemia were active to a lesser degree in one or more of the xenografts in the in vivo tumor panel. However, there were isolated examples of a P388-active agent being inactive in the human tumor xenograft models tested and vice versa. There was no indication that the P388 model could predict compound efficacy for specific tumor xenografts. Whether or not the murine leukemias are poor predictors of activity in solid tumors is still somewhat questionable and will only be determined when drugs without antileukemic model activity but of proven value in the treatment of human solid tumors become available.

34

W.R. Waud

2.6 Drug-Resistant Leukemias Panels of in vivo drug-resistant murine L1210 and P388 leukemia models have been developed at Southern Research for use in the evaluation of crossresistance and collateral sensitivity. These models have been used for the evaluation of new drugs of potential clinical interest. An extensive summary of in vivo drug resistance and crossresistance data has been published by Schabel and coworkers [18]. Their initial manuscript included results of in vivo crossresistance studies on 79 antitumor drugs in seven drug-resistant L1210 leukemias and 74 antitumor drugs in 12 drug-resistant P388 leukemias. Previously we expanded this crossresistance database for the drugresistant P388 leukemias to include two new drug-resistant lines and more clinically useful drugs. Also, we updated the database to include new candidate antitumor agents entering clinical trials [19]. Three additional drug-resistant P388 leukemias were added to this database [20]. In this section, we report this crossresistance database for 16 drug-resistant P388 leukemias and many of the clinically useful agents.

2.6.1 Resistance to Alkylating Agents The crossresistance profile of cyclophosphamide-resistant P388 leukemia (P388/ CPA) to 14 different clinical agents is shown in Table 2.2. The P388/CPA line was crossresistant1 to one of the five alkylating agents, no antimetabolites, no DNAbinding agents, and no tubulin-binding agents. Crossresistance of P388/CPA has also been observed for two other alkylating agents (chlorambucil and ifosfamide) [20]. Interestingly, there are differences among these three agents. Chlorambucil and ifosfamide, like cyclophosphamide, each have two chloroethylating moieties, whereas mitomycin C is from a different chemical classes. Whereas ifosfamide, cyclophosphamide, and mitomycin C require metabolic activation, chlorambucil does not. Although P388/CPA is crossresistant to two chloroethylating agents, the line is not crossresistant to other chloroethylating agents (melphalan and BCNU). Therefore, P388/CPA appears to be crossresistant only to a select group of alkylating agents with differing characteristics. P388/CPA appeared to be collaterally sensitive to fludarabine. The effect of 15 different clinical agents on melphalan-resistant P388 leukemia (P388/L-PAM) is shown in Table 2.2. The P388/L-PAM line was crossresistant to approximately one-half of the agents – two of four alkylating agents, one of four antimetabolites, three of five DNA-binding agents, and one of two tubulin-binding

Crossresistance is defined as decreased sensitivity (by >2-log10 units of cell kill) of a drug-resistant P388 leukemia to a drug compared to that observed concurrently in P388/0 leukemia. Similarly, marginal crossresistance is defined as a decrease in sensitivity of approximately 2-log10 units. Collateral sensitivity is defined as increased sensitivity (by >2-log10 units of cell kill) of a drugresistant P388 leukemia to a drug over that observed concurrently in P388/0 leukemia. 1


35

Table 2.2 Crossresistance of P388 sublines resistant to various alkylating agents and antimetabolites to clinically useful agentsa Drug NSC no. Rxb CPA L-PAM DDPt BCNU MMCc MTX 5-FU ARA-C Alkylating agents Melphalan 8806 A – % – – ± Cyclophosphamide 26271 A % – – – – % Mitomycin C 26980 A ± % – – % – % Procarbazine 77213 D – ±d Cisplatin 119875 B – % % – – % BCNU 409962 A – – – % – – Antimetabolites Methotrexate 740 D – –e % % % 6-Thioguanine 752 A –e – 6-Mercaptopurine 755 D – 5-Fluorouracil 19893 D – – –e % ∋ PalmO-ara-C 135962 A – – – –e – – % Trimetrexate 249008 D ± – – – Fludarabine 312887 D ∋ ∋ ∋ – ∋ % Gemcitabine 613327 E – – – – % DNA binders Actinomycin D 3053 A – ± – ± Doxorubicin 123127 A – – – % – – Etoposide 141540 B – – – %d – – Amsacrine 249992 B – % ∋ –d – – Mitoxantrone 301739 B % ∋ – – Tubulin binders Vinblastine 49842 A % Vincristine 67574 B – % – %d – % Paclitaxel 125973 C – – – – ARA-C, 1-β-D-arabinofuranosylcytosine; BCNU, N,N¢-bis(2-chloroethyl)-N-nitrosourea; CPA, cyclophosphamide; DDPt, cisplatin; 5-FU, 5-fluorouracil; L-PAM, melphalan; MMC, mitomycin C; MTX, methotrexate; NSC, National Service Center a CD2F1 mice were implanted i.p. with 106 P388/0 or drug-resistant P388 cells on day 0. Data presented are for i.p. drug treatment at an optimal (≤ LD10) dosage. Symbols: resistance/crossresistance, %; marginal crossresistance, ±; no crossresistance, –; and collateral sensitivity, ∋ b Treatment schedule (Rx): A = day 1; B = day 1, 5, 9; C = day 1–5; D = day 1–9; E = day 1, 4, 7, 10 c Data from Ref. [23] d Treatment schedule was day 1 e Treatment schedule was day 1 and 5

agents. The alkylating agents involved in crossresistance represent different chemical classes. Similarly, the DNA-interacting agents involved in crossresistance include agents with different mechanisms of action – inhibitors of DNA topoisomerase II (amsacrine and mitoxantrone) and a DNA-binding agent (actinomycin D). However, the melphalan-resistant line did not exhibit crossresistance to other inhibitors of DNA topoisomerase II (e.g., doxorubicin and etoposide) or another DNA-binding agent (e.g., doxorubicin).

36

W.R. Waud

The sensitivity of cisplatin-resistant P388 leukemia (P388/DDPt) to 17 different clinical agents is shown in Table 2.2. The P388/DDPt line was not crossresistant to any of these agents. Interestingly, the cisplatin-resistant line was collaterally sensitive to three agents (fludarabine, amsacrine, and mitoxantrone). Of these three agents, the latter two have been reported to interact with DNA topoisomerase II [21, 22]. The crossresistance data for N,N¢-bis(2-chloroethyl)-N-nitrosourea-resistant P388 leukemia (P388/BCNU) have been limited to the evaluation of alkylating agents. The crossresistance profile of P388/BCNU to four different clinical agents is shown in Table 2.2. The BCNU-resistant line was not crossresistant to melphalan, cyclophosphamide, mitomycin C, or cisplatin. The crossresistance profile of mitomycin C-resistant P388 leukemia (P388/ MMC) to 13 different clinical agents is shown in Table 2.2 [23]. The P388/MMC line was crossresistant to approximately one-half of the agents – one of three alkylating agents, zero of four antimetabolites, three of four DNA-binding agents, and two of two tubulin-binding agents. The pattern was similar to that observed for P388/L-PAM.

2.6.2 Resistance to Antimetabolites The effect of 14 different clinical agents on methotrexate-resistant P388 leukemia (P388/MTX) is shown in Table 2.2. The P388/MTX line was not crossresistant to any of these agents. The crossresistance data for 5-fluorouracil-resistant P388 leukemia (P388/5-FU) have been limited to antimetabolites. The sensitivity of the P388/5-FU to three different agents is shown in Table 2.2. The P388/5-FU line was not crossresistant to palmO-ara-C (a slow-releasing form of ara-C) or fludarabine (possible collateral sensitivity). Crossresistance was observed for methotrexate. The crossresistance profile of 1-β-d-arabinofuranosylcytosine-resistant P388 leukemia (P388/ARA-C) to 16 different clinical agents is shown in Table 2.2. The P388/ARA-C line was crossresistant to members of several functionally different classes of antitumor agents – four of five alkylating agents, three of five antimetabolites, zero of four DNA-binding agents, and one of two tubulin-binding agents. Interestingly, the line was collaterally sensitive to 5-fluorouracil.

2.6.3 Resistance to DNA- and Tubulin-Binding Agents The effect of 17 different clinical agents on actinomycin D-resistant P388 leukemia (P388/ACT-D) is shown in Table 2.3. P388/ACT-D was not crossresistant to any alkylating agents or antimetabolites. It was, however, crossresistant to all of the drugs tested that are involved in multidrug resistance except for amsacrine.


37

Table 2.3 Crossresistance of P388 sublines resistant to various DNA and tubulin binders to clinically useful agentsa Drug NSC no. Rxb ACT-D ADR AMSA DIOHA VP-16 CPTc VCR PTX Alkylating agents Melphalan 8806 A – – – – – Cyclophosphamide 26271 A – – – – – – Mitomycin C 26980 A ± – –f % Procarbazine 77213 D – – – Cisplatin 119875 C – – –d –d –f ± BCNU 409962 A – – – – Antimetabolites Methotrexate 6-Thioguanine 6-Mercaptopurine 5-Fluorouracil PalmO-ara-C Trimetrexate Fludarabine Gemcitabine

740 752 755 19893 135962 249008 312887 613327

D D D D A D D E

– – – – –

– – – –e – – ∋ –

± –e –

– – –

– – – – – – – –

DNA binders Actinomycin D Doxorubicin Etoposide Amsacrine Mitoxantrone

3053 123127 141540 249992 301739

A A B B B

% ± % – %

% % % % %

% % % % %

– – – % %

% % % % %

–f –f –f –f

– – – – –

% %

Tubulin binders Vinblastine 49842 B % % % % % Vincristine 67574 B % % % % % % % Paclitaxel 125973 C ± ± % – –f – % ACT-D, actinomycin D; ADR, doxorubicin; AMSA, amsacrine; CPT, camptothecin; DIOHA, mitoxantrone; NSC, National Service Center, PTX, paclitaxel; VCR, vincristine; VP-16, etoposide a CD2F1 mice were implanted i.p. with 106 P388/0 or drug-resistant P388 cells on day 0. Data presented are for i.p. drug treatment at an optimal (≤ LD10) dosage. Symbols: resistance/crossresistance, %; marginal crossresistance, ±; no crossresistance, –; and collateral sensitivity, ∋ b Treatment schedule (Rx): A = day 1; B = day 1, 5, 9; C = day 1–5; D = day 1–9; E = day 1, 4, 7, 10 c Data from Ref. [24] d Treatment schedule was day 1, 5, 9 e Treatment schedule was day 1–5 f Treatment schedule was day 1 and 5

The crossresistance profile of doxorubicin-resistant P388 leukemia (P388/ADR) to 21 different clinical agents is shown in Table 2.3. The P388/ADR line was not crossresistant to any of the antimetabolites and was marginally crossresistant to only one alkylating agent (mitomycin C). Resistance was observed for all of the drugs tested that are reported to be involved in multidrug resistance (actinomycin D, doxorubicin, etoposide, amsacrine, mitoxantrone, vinblastine, vincristine, and paclitaxel). P388/ADR was collaterally sensitive to fludarabine.

38

W.R. Waud

The sensitivity of amsacrine-resistant P388 leukemia (P388/AMSA) to 14 different clinical agents is shown in Table 2.3. P388/AMSA was not crossresistant to any of the alkylating agents and was marginally crossresistant to only one antimetabolite. Crossresistance was observed for all of the drugs tested that are involved in multidrug resistance. The crossresistance data for mitoxantrone-resistant P388 leukemia (P388/ DIOHA) have been limited mainly to agents involved in multidrug resistance. The sensitivity of P388/DIOHA to seven different clinical agents is shown in Table 2.3. The P388/DIOHA line exhibited mixed multidrug resistance – crossresistant to amsacrine and vincristine but was not crossresistant to actinomycin D, doxorubicin, etoposide, or paclitaxel. The crossresistance profile of etoposide-resistant P388 leukemia (P388/VP-16) to 13 different clinical agents is shown in Table 2.3. The P388/VP-16 line was not crossresistant to any of the alkylating agents or antimetabolites; however, it was crossresistant to all of the drugs tested that are reported to be involved in multidrug resistance. The sensitivity of camptothecin-resistant P388 leukemia (P388/CPT) to seven different clinical agents is shown in Table 2.3 [24]. P388/CPT was not crossresistant to any of these agents. The effect of 21 different clinical agents on vincristine-resistant P388 leukemia (P388/VCR) is shown in Table 2.3. The P388/VCR line was crossresistant to three of the agents – mitomycin C, cisplatin (marginal), and vinblastine. Unexpectedly, P388/VCR was not crossresistant to many of the drugs tested that are involved in multidrug resistance (e.g., actinomycin D, doxorubicin, etoposide, amsacrine, mitoxantrone, and paclitaxel). The crossresistance data for paclitaxel-resistant P388 leukemia (P388/PTX) have been limited to agents involved in multidrug resistance. The sensitivity of P388/PTX to three different clinical agents is shown in Table 2.3. The P388/PTX line was crossresistant to drugs that are involved in multidrug resistance (doxorubicin, etoposide, and vincristine).

2.7 Conclusions Currently, biotechnology continues to advance in an almost exponential fashion. Today, advanced techniques and tools allow us to conduct research not even imagined 50 years ago when the L1210 and P388 leukemia models were first used extensively (e.g., sequencing the human genome). Utilizing molecular biology techniques, the emphasis is now on the development of compounds designed for a specific target. Current NCI strategy suggests that models for evaluating these compounds contain the specific target either naturally or by gene transfection. Successful treatment of such models will theoretically provide the necessary proofof-concept required for continued development. This is a radical departure from the empirical approach to mass screening of compounds against murine leukemias.


39

The L1210 and P388 leukemia models do have some advantages – they are rapid, reproducible, and relatively inexpensive (in comparison to human tumor xenograft or transgenic models). However, as with any experimental animal tumor model, there are limitations. Neither leukemia is a satisfactory drug discovery model for either human cancer in general or human leukemia in particular. Of course this could be said of any animal tumor model. Of the two leukemias, P388 is the more sensitive but overpredicts drug activity for both preclinical human tumor xenograft models and the clinic. However, the question of whether P388 leukemia (or L1210) is a poor predictor for solid tumor-active drugs has yet to be sufficiently answered. Although the murine leukemia models have severe limitations, these models have been very useful in anticancer drug development, in the development of a number of therapeutic principles, and in understanding the biologic behavior of tumor and host. These models are still useful today in conducting detailed evaluations of new candidate anticancer drugs (e.g., schedule dependency, route of administration dependency, formulation comparison, analog comparison, and combination chemotherapy). Perhaps the greatest utility of the murine leukemias today is derived from the evaluations of the drug-resistant sublines for crossresistance and collateral sensitivity. Analysis of the crossresistance data generated at Southern Research for clinical agents has revealed possible noncrossresistant drug combinations. The P388 leukemia lines selected for resistance to alkylating agents (e.g., P388/CPA, P388/L-PAM, P388/DDPt, P388/BCNU, and P388/MMC) differed in crossresistant profiles, both with respect to alkylating agents and other functional classes. Similarly, P388 leukemia lines selected for resistance to antimetabolites (e.g., P388/MTX, P388/5-FU, and P388/ARA-C) differed in crossresistance profiles, both with respect to antimetabolites and other functional classes. Clearly, the spectrum of crossresistance of an alkylating agent or an antimetabolite will depend on the individual agent. P388 leukemia lines selected for resistance to large polycyclic anticancer drugs (e.g., P388/ACT-D, P388/ADR, P388/AMSA, P388/DIOHA, P388/VP-16, P388/CPT, P388/VCR, and P388/PTX) were not generally crossresistant to alkylating agents or antimetabolites. However, the crossresistance profiles to DNA- and tubulin-binding agents were variable. Five of the 16 drug-resistant leukemias exhibited collateral sensitivity to one or more drugs. These observations of collateral sensitivity suggest that a combination of one of the five drugs plus one of the corresponding agents for which collateral sensitivity was observed might exhibit therapeutic synergism. Crossresistance data, coupled with knowledge of the mechanisms of resistance operative in the drug-resistant leukemias, may yield insights into the mechanisms of action of the agents being tested. Similarly, crossresistance data, coupled with the mechanisms of action of various agents, may yield insights into the mechanisms of resistance operative in the drug-resistant leukemias [19]. Furthermore, crossresistance data may identify potentially useful guides for patient selection for clinical trials of new antitumor drugs [19]. In conclusion, the role of L1210 and P388 leukemias in the evaluation of anticancer agents has diminished considerably. Nevertheless, many of the clinical agents

40

W.R. Waud

now in use were first detected by the murine leukemias. These models are clearly still appropriate for answering certain questions, and the drug-resistant sublines can provide valuable information concerning crossresistance and collateral sensitivity. Acknowledgments The majority of this work was supported by Contracts NO1-CM-07315 and NO1-CM-47000 and predecessor contracts with the Developmental Therapeutics Program, DCTD, NCI. The studies of gemcitabine and P388/VP-16 leukemia were supported by Eli Lilly and Company and by Burroughs Wellcome Company, respectively. The authors gratefully acknowledge the technical assistance of the staff of the Cancer Therapeutics and Immunology Department. J. Tubbs assisted with data management, and K. Cornelius prepared the manuscript.

References 1. Law LW, Dunn DB, Boyle PJ, Miller JH. Observations on the effect of a folic-acid antagonist on transplantable lymphoid leukemias in mice. J Natl Cancer Inst. 1949;10:179–92. 2. Dawe CJ, Potter M. Morphologic and biologic progression of a lymphoid neoplasm of the mouse in vivo and in vitro. Am J Pathol. 1957;33(3):603. 3. Griswold DP Jr, Harrison SD Jr. Tumor models in drug development. Cancer Metastasis Rev. 1991;10:255–61. 4. Zubrod CG. Historic milestone in curative chemotherapy. Semin Oncol. 1979;6(4):490–505. 5. Goldin A, Serpick AA, Mantel N. A commentary, experimental screening procedures and clinical predictability value. Cancer Chemother Rep. 1966;50(4):173–218. 6. Carter S. Anticancer drug development progress: a comparison of approaches in the United States, the Soviet Union, Japan, and Western Europe. Natl Cancer Inst Monogr. 1974;40:31–42. 7. Goldin A, Venditti JM, Muggia FM, Rozencweig M, DeVita VT. New animal models in cancer chemotherapy. In: Fox BW, editor. Advances in medical oncology, research and education. Vol. 5. Basis for cancer therapy 1. New York: Pergamon; 1979. p. 113–22. 8. Alley MC, Scudiero DA, Monks A, Hursey ML, Czerwinski MJ, Fine DL, Abbott BJ, Mayo JG, Shoemaker RH, Boyd MR. Feasibility of drug screening with panels of human tumor cell lines using a microculture tetrazolium assay. Cancer Res. 1988;48:589–601. 9. Cancer Lett. 1987;13(2):507. 10. Skipper HE, Schabel FM Jr, Wilcox WS, Laster WR Jr, Trader MW, Thompson SA. Experimental evaluation of potential anticancer agents. XVII. Effects of therapy on viability and rate of proliferation of leukemic cells in various anatomic sites. Cancer Chemother Rep. 1965;47:41–64. 11. Schabel FM Jr, Griswold DP Jr, Laster WR Jr, Corbett TH, Lloyd HH. Quantitative evaluation of anticancer agent activity in experimental animals. Pharmacol Ther. 1977;1:411–35. 12. Lloyd HH. Application of tumor models toward the design of treatment schedules for cancer chemotherapy. In: Drewinko B, Humphrey RM, editors. Growth kinetics and biochemical regulation of normal and malignant cells. Baltimore: Williams & Wilkins; 1977. p. 455–69. 13. Skipper HE, Schabel FM Jr, Wilcox WS. Experimental evaluation of potential anticancer agents. XXI. Scheduling of arabinosylcytosine to take advantage of its S-phase specificity against leukemia cells. Cancer Chemother Rep. 1967;51:1625–55. 14. Corbett TH, Valeriote FA, Baker LH. Is the P388 murine tumor no longer adequate as a drug discovery model? Invest New Drugs. 1987;5:3–20. 15. Staquet MJ, Byar DP, Green SB, Rozencweig M. Clinical predictivity of transplantable tumor systems in the selection of new drugs for solid tumors: rationale for a three-stage strategy. Cancer Treat Rep. 1983;67:753–65.


41

16. Trader MW, Harrison SD Jr, Laster WR Jr, Griswold DP Jr. Cross-resistance and collateral sensitivity of drug-resistant P388 and L1210 leukemias to flavone acetic acid (FAA, NSC 347512) in vivo (Abstr). Proc AACR. 1987;28:312. 17. Wilkoff LJ, Dulmadge EA. Sensitivity of proliferating cultured murine pancreatic tumor cells to selected antitumor agents. J Natl Cancer Inst. 1986;77(5):1163–9. 18. Schabel FM Jr, Skipper HE, Trader MW, Laster WR Jr, Griswold DP Jr, Corbett TH. Establishment of cross-resistance profiles for new agents. Cancer Treat Rep. 1983;67:905–22 (see correction, Cancer Treat Rep. 1984;68:453–9). 19. Waud WR, Griswold DP Jr. Therapeutic resistance in leukemia. In: Teicher BA, editor. Drug resistance in oncology. New York: Marcel Dekker, Inc.; 1993. p. 227–50. 20. Waud WR. Murine L1210 and P388 leukemias. In: Teicher BA, Andrews PA, editors. Anticancer drug development guide: preclinical screening, clinical trials, and approval. 2nd ed. Totowa, NJ: Humana; 2004. p. 79–97. 21. Ho AD, Seither E, Ma DDF, Prentice G. Mitoxantrone-induced toxicity and DNA strand breaks in leukemic cells. Br J Haematol. 1987;65:51–5. 22. Nelson EM, Tewey KM, Liu LF. Mechanism of antitumor drug action: poisoning of mammalian DNA topoisomerase II on DNA by 4¢-(9-acridinylamino)methane-sulfon-m-anisidide. Proc Natl Acad Sci U S A. 1984;81:1361–5. 23. Rose WC, Huftalen JB, Bradner WT, Schurig JE. In vivo characterization of P388 leukemia resistant to mitomycin C. In Vivo. 1987;1:47–52. 24. Eng WK, McCabe FL, Tan KB, Mattern MR, Hofmann GA, Woessner RD, Hertzberg RP, Johnson RK. Development of a stable camptothecin-resistant subline of P388 leukemia with reduced topoisomerase I content. Mol Pharmacol. 1990;38:471–80.

Chapter 3

Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice Lisa Polin, Thomas H. Corbett, Bill J. Roberts, Alfred J. Lawson, Wilbur R. Leopold III, Kathryn White, Juiwanna Kushner, Stuart Hazeldine, Richard Moore, James Rake, and Jerome P. Horwitz

Abstract As preclinical chemotherapists, we are often asked to identify experimental tumor models that can accurately predict for the drug response characteristics of all tumors of a given cellular subtype or molecular target. Unfortunately, it is impossible to give satisfactory answers to these inquiries. Because of the unique character of each independently arising tumor (whether spontaneous or induced), it does not take very long to realize that each tumor is a unique biologic entity with its own tumor growth behavior, histological appearance, drug response and molecular expression profiles. This is true whether the tumor is an experimental animal model or one originally derived from a patient. Further, many factors can influence the tumor growth and therapy response of experimental tumor models. Still, in vivo models are needed to adequately assess pharmacodynamics, toxicity and efficacy of any potential novel therapy. Presented herein is what we hope will be useful information regarding the transplant characteristics of tumor models, with some of the “pitfalls” to look out for when using any given tumor model for chemotherapy evaluations. Although most of the examples given use syngeneic models, the methodologies for assessing the predictive worth and maintaining model usefulness can be applied to almost any given transplantable tumor system (whether syngeneic or xenograft). Keywords Preclinical model • Mice • Chemotherapy

L. Polin (*) Solid Tumor Drug Discovery Lab, Department of Oncology, Barbara Ann Karmanos Cancer Institute, Wayne State University, Prentis Building, Room 2232, 110 E. Warren Avenue, Detroit, MI 48201, USA e-mail: [email protected]


43

44

L. Polin et al.

3.1 Introduction For many decades, the results from transplantable tumor models have been viewed with considerable skepticism. The perception has long been that these models are excessively sensitive, and not predictive of the human disease. Although we do not intend to debate the many issues involved, it is our view that a better understanding of the transplant properties (e.g., take-rate) of the models – as well as a better understanding of the potential shortcomings in data presentation – will greatly aid the reader in the interpretation of these data. This chapter is an effort to summarize some of the basic operating characteristics of a wide range of solid-tumor models. Since this is a chemotherapy group, we can best explain some of the behavioral characteristics of these models within therapeutic experiments. Most of the data are drawn from the use of transplantable, syngeneic mouse tumors, but a few human tumors have been used for contrast.

3.2 Compatibility! Compatibility! Compatibility! Why consider using mouse-tumor models when there are so many human tumor models available? The reason is obvious: compatibility with the host animal. This one feature, above all else, allows the researcher a measure of dependability that can never be attained with the human tumor models in immune-deficient animals. Even if the researcher wishes to use the human tumor-xenograft models for a major portion of their studies, the ability to confirm a result in one or two syngeneic mouse-tumor models and healthy immune-competent mice will usually eliminate the many pitfalls awaiting the unwary [1–3].

3.3 Compatible But Not Perfect: Inbred Mice and Genetic Drift An inbred mouse has >99% homozygosity, and is defined as a product of 20 or more generations of brother–sister mating (each generation reducing heterozygosity by 19%). However, many of the inbred strains were developed between 1905 and 1915, and have undergone genetic drift in various breeding facilities. As an example, the first and third authors well recall the variations in C3H/He mice from various breeders that came into the laboratory in the early 1970s. These mice had different shades of coat-color and snouts of different shape from various suppliers. Since obvious physical attributes were so varied, it was clear that the quality control in breeding was lacking. During that period, the National Cancer Institute (NCI) undertook a program to standardize the common inbred strains (e.g., C3H/He, C57B1/6, DBA/2, Balb/c), and all are now a product of over 150 brother–sister matings. This standardization was accomplished, but some of our long-used

3 Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

45

C3H mammary tumors failed to grow and metastasize adequately in the newly standardized C3H/He strain. Consequently, new tumor models were isolated and characterized – e.g., Mammary Adenocarcinoma-16/C and 13/C [4]. During the mid 1970s to early 1980s, we also developed various colon adenocarcinomas and pancreatic ductal adenocarcinomas for use in chemotherapy studies [5–7]. These tumor models have remained trouble-free in our laboratories and many other laboratories for over 30 years (using only the standardized mice purchased though NCI). This is not necessarily true for some of these same tumors used in other countries. In the late 1970s, we supplied Colon Adenocarcinoma-51 to a highly competent investigator in Europe, who found the tumor to be exceptionally drug-sensitive to a large number of different agents, and quite curable. In our hands, it was (and continues to be) among the least drug-sensitive tumor available (Table 3.1) [6], and only curable with two agents (e.g., PCNU, Piposulfan). Likewise, Colon Carcinoma-26 has remained a drug-insensitive tumor in our laboratories (Table 3.1) [6], but appears to be easily curable with a large number of agents in studies carried out in Japan. One is never sure if the problem is tumor– host incompatibility or the development of deviant lines of tumors. The implications are obvious: syngeneic tumor-model systems can be produced and maintained, but only with adequate quality controls and access to the mice of origin (discussed vide infra).

3.4 Evidence of Tumor–Host Incompatibility and Consequences This topic has been previously reviewed [8, 9], but must be expanded, because so many manipulations are used to overcome the many problems encountered with incompatible models (especially in human tumor-xenograft studies). One suspects that incompatible systems are most frequently used because the treatment results are much more impressive, since the host immune system contributes substantially to the tumor-cell-kill [8]. However, if such a tumor had a unique property required for study (and an appropriate syngeneic system could not be found), the investigator could use the mouse tumor in a (severe combined immunodeficiency [SCID]) mouse host and probably eliminate much of the compatibility issue.

3.4.1 No-Takes Failure to have 100% takes of 30–60-mg trocar tumor fragments may be ascribed to several factors: 1 . Infection 2. Use of an extremely slow-growing, low malignancy, early-passage tumor

Table 3.1 In-vivo activity of standard and investigational agents against transplanted tumors of mice Squam Mam Mam IV Lung IV AML Leuk Mam 16/C/ 16/C/ In-vivo agent Mam 44 Colon 38 Mam 16/C Adr Taxol Colon 51 Panc 02 Panc 03 Mam 17 17/Adr Colon 26 Mel B16 LC12 Leuk L1210 Adriamycin – ++ ++++ ± +++ ± – +++ ++++ – ± + ++++ ++++ + Taxol – ++++ ++++ – – + – ++++ ++++ – ± ± – – – Camp/CPT – + +++ +++ NA + – ++++ NA – NA + – NA NA VP-16 ++ ++ ++++ – + ± – ++ ++ – – – ++++ NA +++++ Vinbl/Vinc – +++ ++ – – – – – – – – – – –a + 5-FU – +++ +++ +++ NA – – + + – ++ ++ – +a ++++ Ara-C – +++ ++ NA NA – – – ++ +++ – – – ++++ ++++ Gemzar + +++ ++++ NA NA NA – + ++ NA +++ NA – ++++ +++++ Cytoxan + ± +++ ++++ NA ++ – ++ +++ +++ ++ ++ ++++ ++++ +++++ CisDDPt + ± + + NA ++ – ++ +++ ++++ ++ ++ ++++ NA ++ BCNU +++ – – ± NA ++ – – – – ++++ ++++ + NA +++++ Cryptophycin8 ++++ ++++ +++ +++ NA +++ +++ ++++ NA ++++ +++ ++ + +++ ++ SR271425 ++ ++++ ++++ + +++ ++ ++++ ++++ ++++ + ++++ ++++ ++++ ++++ +++++ XK469 +++ ++++ ++++ +++ +++ +++ ++++ ++++ ++++ ++++ ++ +++ ++++ +++ +++++ NA not available a SRI/NCI data for Vinc and 5-FU against AML1498. Cryptophycin-8 testing: L1210 data were with Crypto52, and Mamm-44 data were with a close analog being considered for second-generation clinical trials. The L1210 activity rating was expanded because of this tumor’s higher sensitivity to a variety of agents (see Sect. 3.11 for conversion of log kill to activity ratings).


47

3 . Technical incompetence 4. Substantial tumor–host incompatibility The investigator can certainly identify and eliminate the problem of infection [9], as well as technical incompetence. Furthermore, it is unlikely that an investigator would be using an extremely slow-growing, low-malignancy, early-passage tumor for most efficacy studies. Thus, a substantial no-take frequency (5% or more) is usually a signal of tumor–host incompatibility. Interestingly, to get around the problem, one may see implanted tumors that are allowed to reach palpable size before being entered on study (e.g., 6–15 days post-implant, depending on the growth rate of the tumor). In this way, the no-takes are simply culled from the pool of trocared animals, and never entered onto the experiment (or mentioned).

3.4.2 Spontaneous Regressions or Tumors that Fail to Progress to Over 1,000-mg Size In a mouse-tumor model, the occurrence of either spontaneous regressions or a failure of the tumors to progress to over 1,000 mg in size (even at a low frequency of 1/100 mice) would suggest marked tumor–host incompatibility, and thus an invalid model. In human tumor-xenograft systems, a spontaneous regression and no growth thereafter may be the result of a mouse that has regained immune capacity (leaky). This can be confirmed by re-implantation of the tumor in the presumed leaky mouse. If it fails to grow, the leaky nature of the mouse can be verified, and the regression explained. Human tumors that fail to progress in immune-deficient mice are more common in athymic nude mice than in SCID mice (i.e., tumors that reach 250–500 mg and stay that size for many weeks, or even regress to zero) (Table 3.2). These tumors are usually caused by extensive diffuse necrosis within the tumor mass, which can be easily confirmed by histology (Table 3.2). Based on one of the definitions of malignancy (a malignant tumor must be able to grow and kill the host), such a non-progressing tumor model would be judged un-usable. Failure to see plots of all tumors in all treatment cages to >1,000 mg in size (or failure to provide data that include time to 1,000 mg, with range) may suggest that such an invalid human tumor model was used for data collection.

3.4.3 Excessive Curability Many tumor models of past decades were derived in random-bred mice or the original inbred strain subline is no longer available (e.g., Sarcoma 180, Ehrlich Ascites Carcinoma, Gardner Lymphosarcoma). Clearly, these models are excessively curable, with a variety of antitumor agents, because the immune system of the host provides additional cell-kill. The immune system kills by zero-order kinetics (meaning that it can kill a finite number of cells; e.g., 3 × 108 tumor cells). If the

Colon H116 80 (60–90) Colon H8 75 (70–80) Lung H125 35 (30–40) Prostate PC-3 20 (20) Breast MX-1 50 (40–60) Source: AACR 36:303, 1995

35 (10–50) 20 (15–20) 5 (0–5) 0 (0) 0 (0–20)

5.0 (3.0–5.0) 5 trials 5.5 (4.0–6.0) 4 trials 5.0 (3.0–8.8) 10 trials 6.0 (5.0–7.0) 2 trials 4.0 (2.2–6.0) 16 trials

4.0 (3.5–4.0) 5 trials 3.0 (3.0–4.0) 5 trials 3.0 (2.3–5.0) 14 trials 3.5 (3.0–5.0) 4 trials 2.7 (1.8–4.5) 13 trials

3/23 1/21 4/50 2/7 4/91

0/23 4/21 1/50 4/7 3/91

0/18 0/13 0/59 0/11 0/62

0/18 0/13 0/59 0/11 0/62

Table 3.2 Comparison of the behavior of transplanted human tumors in athymic nude mice and SCID mice Diffuse necrosis (determined by histology; H&E stained Exponential tumor-volume doubling time in sides) days SC tumors Progressive tumor growth SC and spontaneous regression Athymic nude mice SCID mice Athymic nude mice SCID mice Athymic nude mice SCID mice No. of tumors No. of No. of tumors reaching 250– No. of these reaching 250– tumors 500 mg but not these 500 mg but not tumors % Median regressing getting any regressing getting any necrosis % Median to zero larger to zero larger (range) necrosis (range) Median Td (range) Median Td (range) Human tumor

48 L. Polin et al.


49

drug treatment can reduce the population only slightly, the immune system can often handle the rest. One of the indicators of such an immunogenic system is that it is more curable by chemotherapy if the tumor is allowed to grow for about 7 days before treatment than if treatment is started 1–3 days after implant (the total opposite of a syngeneic model). These 7 days of growth allow the mouse to see and process the cell-surface antigens and begin to develop an immune response. Some tumor models have given rise to deviant sublines that are highly immunogenic. The most famous case was the deviant line of Lewis Lung Carcinoma (LLC) that was unfortunately widely distributed by NCI. This subline was highly curable by agents that historically had no activity against the tumor [8]. At least one agent N-phosphoacylase-l-aspartate disodium was advanced to clinical trials on the strength that it was able to cure LLC at four dose levels, which should have provided skepticism by itself. Other lines of LLC behave as expected [8], and we still use this highly metastatic, drug-insensitive tumor for selected studies. However, the widespread use of the deviant line of LLC has given it such a bad reputation that few researchers believe any of the chemotherapy data derived from this famous old tumor model. This is unfortunate, since LLC is one of the most metastatic mouse tumors ever isolated. Other tumor systems may or may not be immunogenic, and cures could simply be the result of treatment with a very effective agent. This is easily sorted out. The cures are simply re-challenged with 30–60 mg1 subcutaneous (SC) trocar fragments of the original tumor. If they take and grow to 1,000 mg (footnote 1) with the expected exponential volume doubling time, it is clear that immune factors were not involved in the original cure. No-takes of tumors implanted in the challenged mice are proof that immune factors were involved in the original cure. It is interesting to note how often tumor-free mice are declared cures without re-challenge. The time to re-challenge is sometimes an issue. The time for one tumor cell to populate to 1 g = [3.32 Td × 9] can be added to the last day of treatment. If the mouse is still tumor-free, a cure can be reasonably certain (see footnote 1). This allows at least three doublings beyond detection (easy detection is 100 mg = 0.1 g = 108 cells). However, some investigators (e.g., H. Skipper) point to the possibility of the survival of a slow Td cell or greater cell loss that can occur with low cell numbers, and the fact that some tumors can repopulate from one cell. For these reasons, they suggest that the time for one cell to grow to 1 g of cells be increased by 50% = 1.5 (3.32 × Td × 9) (which is then added to the last day of Rx) before cure is declared and the mouse can be re-challenged. Regardless, re-challenge of tumor-free survivors should be a prerequisite before an animal can be declared cured. This would clarify the nature of the model being used.

There is a relationship between tumor size and cell number. With syngeneic, transplanted mouse tumors, the tumor mass is usually >85% tumor cells while in exponential growth. A 1 g mass (1,000 mg) = 1 × 109 cells; a 0.1 g mass (100 mg) = 1 × 108 cells; a 0.01 g mass (10 mg) = 1 × 107 cells; a 1 mg mass = 1 × 106 cells, and so on. Thus, a 30 mg mass = 3 × 107 cells (30 million cells). Human tumor cells and mouse tumor cells are approximately the same size, but only a few xenografted human tumors contain >80% tumor cells. Td = exponential tumor-volume doubling time. 3.32 = number of doublings per log. Cure is usually obtained when the population is reduced to approximately 10 or 100 cells for most solid tumors.

1

50

L. Polin et al.

3.4.4 Lack of Invasion and Metastasis The failure to invade and metastasize is often a sign of strong host–tumor incompatibility. In addition, these tumors will often grow to an unusually large size (>6,000 mg) without affecting the health of the mouse. Most metastatic tumors will kill the mouse before they reach 6,000 mg, and some tumors will kill before the tumors reach 2,500 mg (e.g., Colon-26, Panc-02). A simple check for the metastatic capability of a tumor is as follows. The tumor (unilateral or bilateral) is allowed to reach approximately 1,500 mg in size (total burden). The mouse is euthanized, and the lungs are removed though the back (with care to avoid the primary tumor). The harvested lungs from each mouse are then implanted into a naive mouse of the same inbred strain. Growth of the lung tissue to a 1,000 mg mass will provide confirmation that metastatic cells were present and a histologic check of the mass will verify the tumor of origin. Highly metastatic tumors can be propagated in this manner with only a 30–60 mg-size trocar fragment of the lung tissue. The invasive nature of the tumor is also verified by the metastasis check, since non-invasive (often immunogenic) tumor models have little or no capacity to metastasize. Finally, the number of tumor cells required to establish tumor growth closely correlates with the metastatic and invasive nature of the tumor. In virtually all cases, 3 × 105 tumor cells are sufficient to establish SC growth with a highly metastatic and highly invasive mouse tumor, and often titers as low as 104 are adequate [10, 11]. Poorly metastatic tumors often require more than 106 tumor cells to establish SC growth in 100% of the mice [11]. All the mouse tumors we use are invasive and metastatic, although some are obviously more invasive and more highly metastatic (e.g., Pancreatic Ductal Adenocarcinoma-02, Mammary Adenocarcinoma-16/C, Colon Carcinoma-26, B16 melanoma, LLC, and Mammary Carcinoma-44). These six tumors can be used for surgery–chemotherapy adjuvant studies because lung metastases are present in nearly 100% of the mice by the time the tumor reaches 1,500 mg [4, 6, 7, 10–14]. The highly metastatic behavior can be encouraged and maintained by implanting lung fragments subcutaneously every passage or every few passages (instead of the primary SC tumor).

3.5 Compatibility Problems Unique to Human Tumors in Immune-deficient Mice Without a doubt, the most evident problem with human tumors implanted in athymic mice is a diffuse necrosis that can often make up 20–80% of the tumor mass, as determined by histology (Table 3.2). This diffuse necrosis (commonly referred to as shotgun necrosis) is evident even to the edges of the tumor mass, and even in tumors of small sizes (100–150 mg sizes). This type of diffuse necrosis is never seen in syngeneic mouse-tumor models, or in human tumors in humans. The diffuse


51

necrosis is less evident in SCID mice than in athymic nude mice, but still occurs (Table 3.2). Obviously, the extent of necrosis will explain the growth behavior of many human tumors in immune-deficient mice. If the cell loss from the diffuse necrosis matches or exceeds the cell gain from replication, one can have a nonprogressing or even a regressing mass. Examples of the marked differences in tumor-volume doubling time of tumors implanted in SCID mice and athymic nude mice have been published [3], with other examples shown in Table 3.2. This lack of compatibility exerts considerable selective pressure on the tumor, and probably accounts for the marked biologic and drug–response variations in sublines of the same human tumor. For example, we have investigated four different sublines of human-prostate LNCaP from different sources, all totally dissimilar in terms of growth behavior and take-rate.

3.6 Passage of Tumors in Cell Culture: Maintaining Genotype, Histology, Biologic Behavior, and Drug–Response Characteristics The passage of tumors in cell culture may or may not alter the behavior of the tumors when re-implanted in mice. For example, one of the authors (Roberts) re-implanted Colon-26 in mice after 291 population doublings in culture (culture passage carried out by E. Dulmadge and L. Wilkoff of Southern Research Institute). This culture-passaged subline behaved similarly to the preculture tumor with respect to histology, growth behavior, and drug response to three agents (sensitive to MeCCNU and CisDDPt, and insensitive to PalmoAraC). On the other hand, Roberts found that Colon Adenocarcinoma-38 changed markedly after culture passage (culture passage carried out by E. Dulmadge and L. Wilkoff at the same time). On re-implantation in mice, the culture-passaged subline of Colon-38 was markedly less differentiated histologically, grew faster, and was unresponsive to 5-fluorouracil (5-FU) and Anguidine (both drugs were highly active against the preculture tumor). In other studies, we have seen substantial changes in modal chromosome number, as well as a marked increase in the distribution of chromosome numbers within the cells of culture passage of tumors, unlike the mouse-passaged tumors that had no genotypic changes. Finally, malignancy properties can change with longer-term culture passage. We have found that culture passage reduced the take-rate more than 10,000-fold for L1210 and P388 leukemias, as well as slowing the growth rate and markedly decreasing invasive and metastatic behavior on re-implantation into the host of origin (DBA/2 mice). Obviously, there is little reason for tumors to maintain high-malignancy properties in a culture setting; they only need to be able to survive and replicate in the artificial cell-culture environment. The substantial increase in genetic instability in culture has obvious implications for the fidelity of the tumor model after

52

L. Polin et al.

r e-implantation in mice. With some effort, one can clone sublines of tumors from culture with widely differing properties. Some investigators (including the NCI) have used long-term-culture-passaged human tumor lines for new drug discovery since the early 1980s [15].

3.7 Cancer: a Cellular Disease, Take-Rate, Feeder Effect, Implications for Cure Many of the behavioral characteristics of transplantable tumor models are controlled by the take-rate behavior of the tumor. To illustrate this fact, it is necessary to discuss the establishment and behavior of tumors from counted cell implants. Counted cellimplant experiments (by the routes of administration planned for using the model) are one of the first steps that an investigator can use to characterize a tumor. An example of the take-rate of a highly metastatic mammary tumor implanted by various routes is shown in Table 3.3. It should be noted that the take-rate varies, depending on the location. Historically, intracerebral implants have the best takerates because the brain tissue acts as a feeder-layer. However, for this particular tumor, the intraperitoneal implants were almost as good (Table 3.3). An example of the take-rates of a low-viability, early-passage tumor is shown in Table 3.4. In this case, the (IC) take-rates were 3 logs better than the other three routes. Take-rates vary enormously depending on the tumor used. Examples are shown in Table 3.5. In general, the take-rate correlates closely with the invasive and metastatic capability of the tumor. In these examples, Mammary Carcinoma-44 is the most metastatic and the most invasive (>90% metastasis to the lungs from 1,000-mg-size SC tumors [11]. Colon 26 is virtually the same [6, 11]. Mammary 16/C is reasonably close, with >80% metastasis to the lungs from 800- to 1,000-mgsize SC tumors, and usually near 100% for tumors >1,200 mg [4, 10, 11]. Colon-51

Table 3.3 Comparison of take-rates of mammary adenocarcinoma-16/c by various implant routes Number of tumor-takes/number of mice implanted Number of cells implanted Subcutaneous Intravenous Intracranial Intraperitoneal 107 19/19 106 19/20 a 10/10 10/10 20/20 5 10 17/20 10/10 10/10 10/10 104 5/20 1/10 9/10 8/10 103 0/19 0/10 3/10 2/10 102 0/10 4 × 104 3 × 103 5 × 103 Take-rate level 4 × 104 One take-rate unit (by dilution) will establish a tumor in 63% of the mice [100 × (1 – 1/e)] This mammary adenocarcinoma is a highly invasive, highly metastatic, rapidly growing tumor a Leakage of the titered cell implant brei from the implant site probably accounts for the one notake. Three logs of incomplete takes should not occur with technically perfect implants. Data of B.J. Roberts


53

is metastatic to both the lungs and regional lymph nodes from a 1,500-mg SC tumor in 80% of the mice, but has been used only occasionally for surgical studies [11]. Colon Adenocarcinomas-9, 10, and 36 were only marginally metastatic in the early passages studied (2.8 2.0–2.8 1.3–1.9 0.7–1.2 125%) of initial value. All patients enrolled in these studies had measurable lesions, evaluation was usually performed after two treatment cycles or after maximal tumor regression. The evaluation of tumor response in nude mice and in the patients was performed by different physicians. For testing new compounds, the in vivo evaluation was performed using tumor volume obtained by the formula 0.5 × length × (width)2 according to Geran et al. [23]. Relative tumor volumes were calculated for individual tumors by dividing the tumor volume on dayt by the tumor volume on day0 multiplied with 100 for all time points. The minimum test/control in percent is considered as the optimal value (optimal T/C%).

7.2.3 Molecular Target Characterization of the Freiburg Patient-Derived Tumor Xenograft Panel A panel of 200 human tumor xenografts has been extensively studied for the expression of novel validated targets for cancer drug development by tissue microarray analysis and was profiled for mRNA expression of the whole annotated human genome [24–28]. Genomic Profiling. For mRNA preparation, tumors were grown in untreated mice. Following humane euthanasia, tumors were excised without delay and tumor pieces free of necrosis were flash frozen in liquid nitrogen. After mechanical tissue disruption, total tumor RNA was extracted using the RNeasy Mini kit (QIAGEN, Hilden, Germany). Prior to array analysis, one round of T7 promoter-based RNA amplification was performed. Affymetrix® HG-U133 Plus 2.0 mRNA expression arrays were used to determine the expression of 47,400 transcripts, corresponding to human 38,500 genes. The HG-U133 Plus 2.0 mRNA expression arrays have proven high reproducibility for mRNA expression analysis. CEL result files were preprocessed using the gc-RMA algorithm independently for training and validation sets. Chip normalization to the 50th percentile was performed afterwards. Predictive transcripts were identified by an iterative leave-one-out/intersection process utilizing Genespring BioScript SG2c-2 (Agilent, Santa Clara, USA). Support vector machines were used as the class prediction algorithm [31]. Tissue Microarrays. Microarrays were assembled from up to 150 paraffin embedded, formalin fixed human tumor xenografts by using a tissue microarrayer (Beecher Instruments, Sun Prairie, WI, USA). Fresh xenograft tissue was collected when tumors reached approximately 1.5 cm in size and immediately fixed in 10%

174

H.-H. Fiebig and A.M. Burger

PBS buffered formalin for 24 h followed by routine processing and embedding into paraffin [27, 28]. Whole tumor sections (4 mm) were cut and stained with H&E. Hematoxylin–Eosin sections of the xenografts were studied by light microscopy and representative areas marked on the slides. Xenograft biopsies, 0.6 mm in diameter, were taken from the corresponding area in the paraffin block and arrayed in duplicates into a new recipient block. Four micrometer sections of the microarray block were cut and transferred onto glass slides using the paraffin sectioning aid system (Instrumedics, Hackensack, NJ, USA). After rehydration, the endogenous peroxidase was blocked in 3% H2O2 solution. Antigen retrieval was accomplished through microwave pretreatment (20 min at 100°C) in citrate pH 6.0 or Tris/HCl pH 10 buffers, depending on the primary antibody. The detection system employed was based on streptavidin-peroxidase-diaminobenzidine and subsequent signal amplification with CuSO4 (Zymed, San Francisco, CA, USA) or strepavidin-peroxidase-Histogreen (Linaris, Wertheim, Germany). The arrays were stained with primary antibodies against targets of interest. Staining was analyzed by light microscopy (Zeiss Axiovert 100 Microscope, Darmstadt, Germany), according to the proportion of positive cells and intensity. A scoring system ranging from 0 to 3+ was used and the staining intensity evaluated by two independent observers [26–28].

7.2.4 Clonogenic Assay A modification of the double-layer soft-agar or clonogenic assay as described by Hamburger and Salmon was used [20, 30]. The cell population of this assay represents tumor stem and progenitor cells which are responsible for self-renewal and thus the unlimited growth of a tumor. An excellent correlation of drug response in the patient and in the clonogenic assay has been published by us and various other groups [17, 20, 21, 29]. 7.2.4.1 Preparation of Single-Cell Suspensions Solid human tumor xenografts were mechanically disaggregated and subsequently incubated with an enzyme cocktail consisting of collagenase 0.05%, DNAse 0.07%, and hyaluronidase 0.1% in RPMI 1640 at 37°C for 30 min. The cells were washed twice and passed through sieves of 200 and 50 mm mesh size. The percentage of viable cells was determined in a hemocytometer using trypan blue exclusion. 7.2.4.2 Culture Methods The tumor cell suspension was plated into 24-multiwell plates over a bottom layer consisting of 0.2 mL Iscove’s Modified Dulbecco’s Medium (IMDM) with 20%

7 Patient-Derived Tumor Models and Explants

175

fetal calf serum and 0.7% agar. 20,000–100,000 cells were added to 0.2 mL of the same culture medium, in 0.4% agar, and were plated onto the base layer. Drugs were added after 24 h (drug overlay) in 0.2 mL medium. In each assay, six control plates received the vehicle only, drug-treated groups were plated in triplicate in three or six concentrations. Cultures were incubated at 37°C and 7% CO2 in a humidified atmosphere for 6–18 days and monitored closely for colony growth using an inverted microscope. During this period, in vitro tumor growth led to the formation of colonies with a diameter of >50 mm. At the time of maximum colony formation, vital colonies were stained with a tetrazoliumchloride dye 24 h before counting them with an automated image analysis system (Bausch & Lomb OMNICON FAS IV). Drug effects were expressed as percentage of survival, obtained by comparison of the mean number of colonies in the treated plates with the mean colony count of the vehicle treated controls (colony count T/C × 100). A compound was considered active if it reduced colony formation to less than 30% of the control group value (T/C £ 30%). Furthermore, inhibitory concentrations IC50, IC70, and IC90 were calculated corresponding to T/C values of 50, 30, and 10%. Using these evaluation parameters the majority of clinically established anticancer agents was active at a concentration of 20%) and in resistant tumor types (PR rate 95% stable for a period of 8 h at room temperature.

Injecting rats with MNU

• MNU is injected intraperitoneally using a 1 ml disposable plastic syringe equipped with a 26 gauge, 3/8″ needle. • The amount injected is dictated by carcinogen dose and weight of the animal. There is no need for any type of animal restraining device during the injection procedure. • It is helpful to prepare an injection schedule with body weight listed in 5 g increments and the appropriate volume of carcinogen to be injected listed next to the body weight. Animals can be weighed at the time they are injected.

Cleanup following injection

• MNU in excess of that used during carcinogen administration should be chemically inactivated in an institutionally designated chemical fume hood. Typically, a saturated solution of sodium carbonate is used for this purpose. • A dilute solution of alkali can be used to decontaminate work surfaces. • In general other materials used during carcinogen administration can be disposed of via incineration in compliance with an institution’s Biosafety guidelines.

10.2.2 Carcinogen Dose and Age of Administration Currently there are two additional factors that must be considered in using MNU to induce mammary carcinomas: the dose of carcinogen to be administered and the age at which carcinogen is injected. The specific research question being addressed will, in part, dictate the choices made. The following information should be of value in making these decisions. A dose-dependent induction of mammary carcinomas is observed in response to MNU. The typical range of doses that has been used is from 12.5 to 50 mg MNU/kg body weight. At 50 mg/kg dose, which is the dose most typically injected, a high incidence and multiplicity of carcinomas is observed, and the latent period is short;

10 Mammary Cancer in Rats

249

mean time to tumor depends on the age at which MNU is injected. References [21, 24, 28] provide quantitative information relative to the changes in incidence and multiplicity of palpable mammary carcinomas over time following carcinogen administration at various doses of carcinogen injected at different ages. In general, if the research question simply requires the rapid induction of palpable mammary carcinomas to test a therapeutic agent, or if the research question is best tested when a robust carcinogenic process is induced with a short latent period and the occurrence of carcinomas in essentially all injected animals, then 50 mg MNU/kg is the dose of choice. At the other extreme is the carcinogenic response observed when a dose of 12.5 mg MNU/kg body weight is injected. In general 60% of the animals had mammary carcinomas that were detectable by palpation. This is in contrast to the longer latent periods (noted above) when MNU was injected at 50 days of age. Moreover, animals can be sacrificed at time points ranging from 14 to 35 days post-carcinogen, and a high incidence of pre-malignant lesions identified and evaluated in mammary gland wholemount preparations [28, 29]. The procedure for preparing wholemounts is described in detail in [30]. Based on our experience, if a high incidence of palpable mammary carcinomas with a short latent period is desired, then the model of choice is 50 mg MNU/kg body weight injected at 21 days of age. The added advantage of this model is that it also permits the investigation of the genesis and prevention of pre-malignant lesions. Whether MNU is injected at 21 or 50 days of age, there are many practical advantages to this method of mammary carcinoma induction. They are summarized in Table 10.2.

10.2.3 Typical Animal Protocols To facilitate the adaptation of the MNU model for those less experienced in experimental carcinogenesis, we describe two typical experimental protocols. 10.2.3.1 Chemoprevention Protocol In a typical chemoprevention experiment, the effects of a potential chemopreventive agent on the carcinogenic response are contrasted to the carcinogenic

250

H.J. Thompson

Table 10.2 Advantages of inducing mammary carcinomas using a single i.p. injection of 50 mg MNU/ kg body weight Category Attributes Induction • Minimal supplies are needed; readily available methodology • Convenient i.p. injection protocol requiring minimal manipulation of the animal • MNU has a short half life in the animal (3 years [36]. In another study, MacEwan et al. compared surgical resection of malignant melanoma tumors to surgery followed by therapy with Corynebacterium parvum. The heatkilled bacteria acted as a nonspecific immune stimulant, and it was hypothesized that this treatment might elicit an anti-tumor response. It was determined that the addition of this treatment led to a significant increase in overall survival (7.5 vs. 12 months) in animals with advanced-stage tumors (involvement of lymph nodes or distant organs) [37]. Other trials have targeted the monocyte/macrophage compartment with nonspecific stimulants, such as liposome-encapsulated muramyl tripeptide phosphatidylethanilamine (L-MTP-PE). L-MTP-PE stimulates the tumoricidal activity of canine macrophages and induces the secretion of proinflammatory factors. In a recently completed study it was found that the use of L-MTP-PE as an adjunct to surgical resection of Stage I oral melanomas resulted in a statistically significant increase in overall survival as compared to surgery alone, with 80% of the dogs treated with L-MTP-PE still alive at >2 years [38]. Canine oral melanomas also lend themselves to studies of gene therapy. Treatments that have been studied in this model include the intratumoral administration of a cDNA encoding recombinant human GM-CSF and an autologous tumor vaccine composed of irradiated tumor cells transfected with the GM-CSF gene ex vivo [39, 40]. Canines that had been injected with a combination of lipoplexes containing a plasmid coding the suicide gene herpes simplex thymidine kinase and irradiated xenogeneic cells secreting human GM-CSF and human IL-2 and simultaneously treated with ganciclovir exhibited prolonged survival compared to canines receiving control therapy [41]. Another study examined the efficacy of intratumoral injections of plasmid DNA-encoding staphylococcus enterotoxin B and either canine GM-CSF or IL-2 [42]. The overall response rate was 46%, although responses were more frequently seen in dogs with smaller tumors. Importantly, the survival times for animals bearing stage III tumors were significantly prolonged by the intratumoral administration of these genes as compared to animals receiving surgical treatment alone. The results obtained in these trials serve to highlight the importance of this model and its utility in the investigation of novel immunotherapeutic modalities. Canine melanomas are locally aggressive and are resistant to systemic treatments, just like their human counterparts. Therefore, it is quite likely that this model will continue to be used in the analysis of innovative biologic treatments.

11.5 Sinclair Swine The Sinclair, Hormel, Munich troll and melanoblastoma-bearing Libechov (MeLiM) breeds of miniature pig exhibit a strong predisposition toward the development of cutaneous melanoma [43]. This is a genetic trait that can be enhanced through

11 Animal Models of Melanoma

265

selective breeding, so that over 50% of swine will exhibit melanoma at birth and 85% will develop lesions by 1 year of age [44, 45]. Melanocytic nevi are also observed in these species. Two separate loci appear to be involved in the development of tumors. One is the B haplotype located within the swine leukocyte antigen (SLA) complex. Animals homozygous for this gene exhibit a strong tendency toward the development of melanoma tumors. A second gene family, unlinked to the SLA complex, is fully expressed only in animals exhibiting the B haplotype [46]. The second locus may actually be related to the retinoblastoma (Rb) locus in humans. Susceptible animals have inherited either an inactive form of this allele or are missing it altogether. In this case, tumor development occurs following somatic mutation of the second allele [47]. The inheritability of melanoma susceptibility has made these animals useful for genomic studies. Genome-wide quantitative trait loci (QTL) mapping in the MeLiM model detected several QTLs associated with ulceration, presence of melanoma at birth, lesion type, and number of aggressive melanomas. Comparative mapping has helped to reveal new regions of the human genome that may harbor melanoma susceptibility genes [48]. Interestingly, animals that have undergone oophrectomy (but not orchiectomy) exhibit reduced tumor growth, which can be reversed via the administration of estradiol. Despite these observations, it is clear that melanoma inheritance is not sex-linked [49]. Melanomas may develop in utero and animals are frequently born with congenital flat or raised black skin lesions, a proportion of which may represent malignant tumors. Malignant melanomas also arise throughout the life of the animal. These may develop from pre-existing pigment lesion and may exhibit areas of ulceration. Malignant lesions are frequently large and exophytic. Metastasis to lymph nodes and vital organs occurs in up to 25% of animals, but is rarely a significant cause of morbidity because most primary and metastatic lesions eventually undergo significant regression characterized by shrinkage of tumors and loss of pigmentation [49–51]. Tumor development has been carefully characterized by several investigators, and occurs in five distinct stages [2, 50, 51]. Stage I lesions are flat black macules that contain heavily pigmented melanocytes. The melanocytes are located singly or in nests just superficial to the epidermal basement membrane. Stage II lesions are raised pigmented nodules composed of melanocytes that have begun to invade the superficial dermal structures. Stage III lesions are exophytic and frequently ulcerated. These lesions are invasive and highly proliferative in nature. Melanocytes exhibit significant cellular atypia and may exhibit an epithelioid or spindle-shaped morphology. Mitotic figures are commonly seen. These tumors are composed of pigmented melanocytes that deeply infiltrate the dermis and epidermis and give rise to metastatic lesions. Careful histologic analysis of these tumors reveals that approximately 70% have features consistent with acral lentiginous melanomas of humans, with the remainder resembling superficial spreading melanomas (10%), or nodular mealnomas (20%) [52]. The ultrastructure of tumor cells found in Sinclair swine melanomas is very similar to that of human tumors. Thus, it is not surprising that these cells stain positively for S100 [52].

266

E.T. Fairchild and W.E. Carson, III

Stage IV represents the first stage of regression. Tumors are smooth and bluish in color, and may have a depigmented halo. T-laden macrophages invade the tumor mass. In the final phases of regression (Stage V), there is extensive infiltration of the tumor by lymphocytes and pigment-laden macrophages. Tumor cells disappear from the lesion, and depigmentation and dermal fibrosis become prominent histologic features. These findings suggest that tumor regression is largely mediated by immune mechanisms and may be related to an increased activity of host immune effectors [50]. Indeed, those animals with malignant melanomas exhibit enhanced leukocyte reactivity to tumor-cell lysates [53]. Other studies suggest that the cytotoxic response to the melanoma is mediated in part by gd T cells [43]. Experimental strategies such as analysis of gene expression microarrays and suppressive subtractive hybridization comparing regressing tumors and actively proliferating tumors from these miniature pigs have been employed to identify changes in the tumor that may contribute to regression. These studies have revealed a decreased expression of genes involved in cell cycle progression, DNA replication, DNA recombination, and DNA repair as well as an increase in the retinoic acid receptor-responsive gene RARRES1 [54, 55]. RARRES1 appears to be expressed in differentiated melanocytes and expression is lost as the tumor de-differentiates. Advantages of the Sinclair swine melanoma model include the observed genetic tendency, the high level of spontaneous transformation, and pathologic parallels to human disease. This model also provides an interesting experimental system for the study of spontaneous tumor regression.

11.6 Murine Models There are a number of murine models of malignant melanoma, which may vary with respect to several parameters, such as the origin of the tumor line employed, the site of inoculation, and the genetic background of the host. Each model has distinct strengths and weaknesses, and these must be taken into consideration prior to the selection of a murine model for use in a specific experimental system. In the following section, we will review models which employ chemical and physical carcinogens, transgenic mice, naturally occurring murine melanoma cell lines, and immunodeficient mice.

11.6.1 Induction with Physical Agents Melanoma tumors rarely develop spontaneously in rodents (e.g. mice, rats, hamsters, gerbils, or guinea pigs). Two general approaches are available to the researcher who wishes to induce the formation of melanomas or pigmented preneoplastic tumors in experimental animals. The first involves the repeated application of a carcinogen with or without subsequent applications of a tumor-promoter.


267

Alternatively, animals may be exposed to UVR followed by applications of a reagent with the ability to promote tumor development [2]. Carcinogens that have been employed in this fashion include 7,12-dimethylbenz(a)anthracene (DMBA), trimethylanthracene, nitrogen mustard, and nitrosurea compounds [2]. Initially there is transient hyperpigmentation of the epidermis. This is followed by the development of lesions derived from the dermis that have little or no metastatic potential. Application of tumor promoters such as croton oil or TPA can markedly enhance the development of carcinogen-induced melanomas. In fact, a single application of DMBA can induce the formation of melanomas if animals are treated repeatedly with one of these tumor-promoters [2]. A basic protocol involved the application of 100 mL of 0.4% DMBA in acetone to shaved skin followed by twice-weekly applications of 25 mL of 2.5% croton oil in acetone of dimethyl sulfoxide until the appearance of raised black skin lesions occurs [56]. The resulting tumors are locally aggressive, metastasize to multiple organs, and may be transplanted to new hosts or established in culture. The ability to obtain nevi and malignant melanomas with the application of a simple chemical carcinogen implies that UVR-induced DNA damage is not an essential step in the development of this cancer [2, 56]. This approach to tumor induction has been applied to multiple species, including transgenic murine strains, and has led to the development of several useful murine melanoma cell lines. The JB/MS and JB/RH melanomas were induced in C57BL/6 mice via a single application of DMBA to the scapular region of 4-day-old mice, followed by twice-weekly application of croton oil. These tumors developed at 16 and 23 weeks, respectively. The tumors continued to display a melanotic phenotype following transplantation to normal C57BL/6 mice, and metastasized spontaneously in these new hosts [56]. The K1735 cell line was induced in a C3H/HeN mouse via the application of UVR (ten 1-h exposures to a FS40 sunlamp over a 2-week period) followed by 92 weekly treatments with croton oil [57]. The K1735 cell line expresses MHC class I and class II molecules, and is capable of inducing a specific yet ineffective immune response in syngeneic mice [58]. This cell line may be grown in culture and implanted via the sc injection of a single-cell suspension, or may be passaged via the sc implantation of tumor fragments derived from a solid tumor raised in a mouse. This line has been used to great effect in studies examining the role cell-surface molecules such as B7 and Fas ligand that are able to influence the host-immune response to tumor [58, 59]. Other investigators have applied these protocols to transgenic mice in the hopes of producing a murine melanoma cell line with a specific genetic alteration. For example, utilizing a DMBA protocol, melanomas were successfully induced in mice transgenic for an activated human H-ras gene. Subsequent chromosomal analysis revealed translocations of chromosome 4 that led to reduced expression of the p16 gene product, a situation similar to that reported for human melanomas [60]. UVR may be employed in a variety of ways to induce melanoma formation. Interestingly, UVR can act as a promoter for carcinogen-induced melanoma in mice. Benign blue nevi induced with DMBA may be converted into malignant lesions via exposure of the pigmented areas to chronic low-dose UVR [61]. These results have been confirmed in other murine strains and UVA, UVB, and UVA plus

268


UVB appear to be equally effective promoters [62]. UVR also increases the rate of melanoma induction following applications of DMBA plus croton oil [63]. In general, the effect of UVR appears to be a local one, since application of UVR to nontreated skin is ineffective in the induction of tumors [63]. Despite its reputation as a complete carcinogen in humans, UVR alone cannot induce malignant melanoma in normal murine strains [64]. However, UVR treatment of normal skin can induce the formation of malignant melanomas if the treatment is followed by the application of a standard tumor-promoter [57]. These observations tend to confirm the carcinogenic potential of UVR seen in sun-exposed human populations, and validate the use of these models to study malignant melanoma. Chemically induced and UVR-induced melanomas of mice exhibit histologic features similar to those described for melanomas that arise spontaneously [62]. Hyperplasia of dermal melanocytes with dendritic characteristics is believed to be the first step in tumor development. Examination of benign tumors reveals the presence of large, polygonal cells that are heavily pigmented. Malignant lesions contain cells of similar morphology that in addition may exhibit nuclear atypia and variable pigmentation. Other rodent species (i.e. rat Chinese hamster, gerbil, and guinea pig) only rarely develop spontaneous melanoma. Rats and guinea pigs are relatively resistant to chemical carcinogens, whereas melanoma tumors can be reliably induced in Chinese hamsters and gerbils using standard protocols [2]. In contrast, malignant melanomas develop spontaneously with some frequency in the Syrian hamster (Mesocricetus auratus) [65]. This species (especially the golden and white variety) is also quite susceptible to the induction of tumors by chemical carcinogens, although they appear to be resistant to the effects of UVR [66]. Although the mouse is relatively resistant to the development of carcinogeninduced melanomas, there are several distinct advantages to the use of this laboratory animal in these protocols. There is a wealth of information relating to the murine immune system, immunologic and molecular reagents are widely available, and tumors may be induced in a variety of inbred and transgenic strains. The disadvantages of this approach are less obvious, but must be taken into consideration when contemplating this technique. Foremost, the tumors and cell lines obtained via DMBA treatments (e.g. JB/MS and JB/RH) are frequently nonpigmented, and this may represent a distinct phenotypical difference between these lesions and the majority of tumors that arise in humans. In addition, the incidence of melanoma formation following treatment with carcinogenic compounds appears to be strainand species-dependent. This suggests that the mechanism responsible for carcinogenesis may not be fully generalizable to humans [2].

11.6.2 Tumors Arising in Transgenic Mice The limitations associated with the induction of melanoma tumors in mice via the use of carcinogens have led investigators to develop transgenic mice with the potential to develop these tumors. One such model employs mice transgenic for


269

expression of the simian virus 40 (SV40) early region under the control of the melanocyte-specific tyrosinase gene. These mice express the small and large transforming (T) antigens and develop ocular melanomas, skin melanomas, and hyperplasia of neural-crest-derived pigmented cells [67]. The T antigen is expressed to a different degree in various inbred lines of mice derived from individual founders. Mice with light coats develop eye tumors that are rapidly growing and fatal at an early age, whereas mice with darker coats develop slow-growing eye tumors later in life. Thus, the cutaneous melanomas that do arise in susceptible transgenics are rare and usually benign at the time the animal dies because of the ocular lesions. To circumvent this problem, skin can be taken from the susceptible strains and grafted onto low-susceptibility transgenics with a longer lifespan. Interestingly, donor pigment cells in these grafts exhibited selective proliferation in those areas nearest the healing wound edges. This finding suggested that growth and malignant conversion of melanocytes expressing the T-antigen [67]. The T antigen functions to inactivate important tumor-suppressor genes (e.g. p53 and Rb), and the results obtained with this model suggest that these wellstudied pathways may be important in the development of melanoma in humans. The T antigen has been shown to potentiate the development of invasive and metastatic melanoma in other models including transgenic mice that express the ret oncogene in retinal epithelium. In the absence of T antigen expression, these mice only exhibit micropthalmia and benign tumors [68]. Thus, the T antigens have powerful effects on a number of pathways that might influence cell proliferation. This explains why mice transgenic for the T antigens may develop melanoma, but this also makes it difficult to pinpoint key steps in the oncogenic process. Interestingly, p53 is infrequently inactivated in human melanomas [69]. In contrast, Rb function may be inhibited in melanoma via loss of specific regulatory pathways [70]. Cell-cycle entry and progression are regulated by the cyclin-dependent kinases (cdks). Cdk4 and cdk6 mediate the phosphorylation of Rb, which inactivates it and permits transcription factors to translocate to the nucleus, where they induce the transcription and expression of specific growth-related genes. The (INK)4a and INK4b gene products (p16INK4a and p16INK4b) are specific inhibitors of cdk4 and cdk6. In the absence of functional INK4 proteins, cdk4 and cdk6 are driven towards a more activated state. Rb becomes hyperphosphorylated, with subsequent dysregulation of the cell cycle. In addition, other cdk inhibitors (e.g. the Cip/Kip proteins encoding p21 and p27, respectively) become sequestered from the cdk2/cyclin E in an indirect manner following the loss of INK4a function [71]. Interestingly, the INK4a gene can code for an unrelated protein through the use of an alternate open reading frame. The resulting gene product (p19ARF) functions as a potent growth-suppressor via its ability to stabilize p53 [70]. Serrano et al. have developed a mouse model with a deletion of exons 2 and 3 of the INK4a gene that eliminates expression of both p16INK4a and p19ARF. These mice are essentially dysfunctional for both the Rb and p53 pathways, and develop spontaneous cancers at an early age. Fibrosarcomas and B-cell lymphomas are the most common tumors that develop in these mice, yet despite the link between INK4a and melanoma, these transgenic mice fail to develop this particular tumor [72].

270


One gene that can accentuate the oncogenic potential of cells expressing an inactivated INK4a gene is oncogenic ras. Mice transgenic for the activated H-ras gene (H-rasV12G) under the control of the tyrosinase promoter develop cutaneous melanomas at a very low rate and only after a long period of observation. Interestingly, both INK4a alleles become deleted in these rare tumors [70]. In contrast, INK4aD2/3 mice expressing activated H-ras develop large numbers of cutaneous melanomas spontaneously after only a few weeks. A similar result was obtained with mice expressing activated N-ras [73]. These tumors are amelanotic, highly vascular, and similar in many respects to nodular melanomas [74]. It was subsequently shown that melanoma genesis and maintenance in this model are strictly dependent upon expression of H-rasV12G. These experiments were performed in INK4a-null mice that were transgenic for doxycycline-inducible H-rasV12G. Withdrawal of doxycycline and downregulation of inducible H-ras led to the clinical and histologic regression of primary and explanted tumors in this murine model. The initial stages of regression were marked by extensive apoptosis of tumor cells and host-derived endothelial cells; however, enhanced expression of vascular endothelial growth factor (VEGF) could not substitute for the loss of activated H-ras [75]. Additional models exploring the consequences of constitutively active oncogenes include transgenic mice that express the ret oncogene under the control of the ubiquitous metallothionein-1 (MT) promoter. Surprisingly, these MT-RET mice exhibited melanocytic tumors in the choroid of the eyes, the dermis around the nose and neck, and in the leg muscles, mediastinum, and retroperitoneal space. Also, mice with the MT-RET transgene expressed on a BALB/c background had a decreased incidence in these tumors as compared to mice with the transgene expressed on the C57Bl/6 background [76]. Interestingly, transgenic mice expressing RET under the melanocyte-specific promoter Tyrp1 did not develop melanoma tumors [77]. The clinical relevance of melanoma models over-expressing ret however is questionable as mutations in ret have not been identified in human melanomas. Transgenic mice have been created that over-express hepatocyte growth factor/ scatter factor (HGF/SF), a pleiotropic cytokine that stimulates melanocytes as well as other cell types. These mice have a unique distribution of melanocytes, which can be found in the dermis, the epidermal/dermal junction, and the basal layer of the epidermis as well as the base of hair follicles like wild-type mice [78]. Spontaneous melanomas arise in this model with a mean onset of 21 months and metastasize in 15% of animals. These melanomas demonstrate a dermal morphology that does not reflect that of human melanomas. However, when these mice are exposed to neonatal UV irradiation, the resulting melanomas exhibit a different histology with extension of tumor cells into the epidermis. These lesions also exhibit both junctional and dermal components [79]. These patterns are reflective of those in human melanomas and are likely due to the more superficial pattern of melanocyte distribution which leads to greater exposure to UV radiation and hair follicles. Further studies of UV radiation and HGF/SF mice have shown that UVB and not UVA is responsible for initiating melanoma formation in these mice [80].


271

11.6.3 Spontaneously Arising Murine Melanomas Malignant melanomas are a rare occurrence in mice. There are three melanomas that arose in mice and could be propagated either as a cell line or as a transplantable tumor. The Harding–Passey cell line is derived from a dermal melanoma that developed on the ear of an ICR mouse and S91 (or Cloudman) melanoma cell line developed in a DBA mouse. The transplantable B16 melanoma arose spontaneously in a C57BL/6J mouse in the 1950s and several different subclones are now available [81–83]. These lines can be maintained either in vitro under standard culture conditions or can be passaged in vivo as sc tumors. Many innovative murine models of melanoma have been developed utilizing these three cell lines. Some of the most important ones are described here. The Harding–Passey cell line has been utilized in numerous preclinical studies of novel anticancer therapies, including boron neutron capture therapy, hyperthermia, strategies employing attenuated herpes simplex virus I (HSV1), and radioiodinated antibodies [84–87]. This cell line can be implanted subcutaneously or intramuscularly [84, 88]. It has been used to generate a mouse brain-tumor model in which cells are injected intracranially into C57BL/6 [86]. Brain tumors develop in 100% of animals, and may be imaged in as few as 5 days. Melanotic and amelanotic variants of the Harding–Passey cell line also exist, thus it has been used in studies that correlate melanin content with biologic behavior [89]. The Cloudman line can be used to generate sc tumors [90] for use in the evaluation of novel anticancer strategies [90, 91]. In addition, these tumors can be harvested, mechanically disrupted (e.g. by forcing tissue fragments through a wire screen), cleared of cellular proteins and debris, and utilized as a single-cell suspension for injection [92]. An intracutaneous model of malignant melanoma was also developed using this cell line [93].

11.6.3.1 Models Employing the B16 Cell Line The B16 murine melanoma cell line arose spontaneously in the C57BL/6 mouse in 1954, and like most tumors, was probably monoclonal in origin. However, significant heterogeneity would likely have been generated within the primary tumor in a short period of time as the result of genetic instability and unique selective pressures (e.g. in vitro culture conditions and/or subsequent site of implantation). Subclones of the B16 line have been generated that exhibit an enhanced rate of proliferation, superior ability to colonize specific organs such as the lung following iv injection, and increased metastatic potential following sc implantation in a syngeneic host [94]. The ability to generate this variety of sublines suggested to some researchers that the B16 cell line may have accumulated a greater degree of heterogeneity than more recently developed melanoma cell lines. However, newer murine melanoma lines appear to exhibit just as much phenotypic diversity as ones developed in the early portion of the twentieth century. Studies by Fidler et al. revealed

272


that the K-1735 melanoma cell line (first isolated in 1979) is quite heterogeneous and contains subpopulations of cells exhibiting diverse biologic behavior. In fact, of the 22 subclones derived from the parental line, only two were indistinguishable from the original cell line [95]. Also, Stackpole et al. evaluated the phenotypic diversity of the B16 melanoma line using tumor fragments that had been kept in the frozen state for over 20 years. They were able to isolate a large number of clones that varied widely in their potential for dissemination (e.g. growth rate, metastatic potential, and ability to colonize organs). Moreover, cells readily converted their phenotype during growth in vitro and in vivo [96]. The subclones of the B16 cell line most commonly employed in murine experimental systems at the present time are B16F1 (low metastatic potential) and B16F10 (high metastatic potential). The differences in metastatic potential appear to relate to variations in cell-surface properties [97]. These clones were developed in experiments performed in the laboratory of J. Fidler in the 1970s [98]. The B16F10 tumor was obtained from Jackson Laboratories in 1970 and passaged in syngeneic mice as a sc tumor prior to being established in cell culture. After four to five in vitro passages, the tumor was frozen and stored in liquid nitrogen. Years later, the line was thawed, grown subcutaneously, and once again established in vitro. Two aliquots of cells were then prepared. One was further divided and used to inject a group of mice directly. The other was used to produce several clones. Aliquots of the unmanipulated cells and the distinct clone were then injected intravenously into syngeneic mice, and pulmonary metastases were counted at 18 days. The unmanipulated cells exhibited a metastatic potential similar to that of the parental line (median number of metastases = 40, range 8–131), whereas the clones differed markedly in their ability to colonize the lung (3–500). These experiments suggested that primary tumor-cell populations are enormously diverse with respect to their proclivity to disseminate to visceral organs. Another B16 variant with high metastatic potential is the BL6 subline [99]. The B16-BL6 cell line was generated by the injection of B16F10 cells into the urinary bladder of male C57BL/6 mice through the vas deferens. The bladder was then excised and cultured in vitro on semi-solid agar (37°C, 5% CO2). Tumor cells that invaded through the bladder wall into the agar were recovered and repassaged. This entire process was repeated six more times, and the resulting variant was designated BL6. As might be expected, the BL6 variant is highly invasive and highly tumorigenic, yet poorly immunogenic. In fact, this clone can invade through the bladder wall in just 24 h. Other variants that metastasize preferentially to the ovary (B16-O10), brain (B16-B15b), and liver have also been developed [100, 101]. The B16 variants – B16F1, B16F10, and B16-BL6 – may be grown subcutaneously via the injection of 106 cells in a volume of 20–100 mL followed by therapeutic manipulation in 7–10 days [102]. Alternatively, tumors may be treated after they achieved a given size (e.g. 5–10 mm in diameter). The B16F1 cell line will not metastasize to visceral organs, whereas visceral spread (primarily to the lungs) can be expected uniformly in mice receiving sc implants of the B16F10 or B16 BL6 variants. Tumor volume may then be measured using calipers and standard formulas, or survival may be used as an end point. Antimetastatic therapies may be


273

evaluated in mice that have been iv-injected (via the lateral tail vein) with 2 × 105 B16F10 cells [102, 103]. Treatment may then be started 1 d later. Following a period of treatment (e.g. 2 weeks), mice are euthanized, their lungs are harvested, and pulmonary surface colonies are enumerated with the aid of a dissecting microscope. Further metastasis to other visceral sites probably arise from the lung via hematogenous spread. In fact, most terminal-stage lung metastases themselves develop from lung lesions measuring only 1–2 mm in diameter [104]. The generation of visceral metastases via the iv injection of tumor cells or sc implantation of unique subclones that metastasize directly to the lungs does not accurately reflect the normal sequence of events in the clinical setting, in which nodal metastases play a prominent role. To address this problem, several unique models have been developed that employ the B16 melanoma cell line. Nathanson et al. injected B16 cells subcutaneously into the left foot pad of 6- to 8-week-old C57BL/6 mice [105]. They used the F1, F10, and BL6 variants, injecting 5 × 104 viable cells in a volume of 0.05 mL. Animals were inspected daily for the development of tumors, which generally became visible approximately 2 weeks following inoculation. At 18 days, the affected limb was amputated and the popliteal lymph nodes were isolated. Visceral metastases were enumerated at day 18 or at the time of death from systemic disease. Analysis of tumor-bearing mice demonstrated a direct correlation between the development of lymph node metastases and the size of the primary tumor. Also, pulmonary metastases correlated with tumor size with the BL6 variant and they exhibit a greater tendency to spread to nodal basins and then to the lung than either the F1 or F10 strains. The incidence of pulmonary metastases in mice whose regional lymph nodes did not contain tumor also correlated with increased size of the primary tumor, an apparent indicator of hematogenous spread. A similar model was developed by Markovic et al. utilizing 1 × 106 B16F10L cells implanted as single-cell suspension into the right food pad of mice [106]. The tumorbearing limb was amputated when the tumors routinely measured 6–8 mm in diameter (approximately 18 days). Animals with palpable inguinal lymphadenopathy were removed from the experimental group. In the absence of further therapy, mice characteristically died of metastatic pulmonary disease at about 35 days post-surgery [106]. An alternative site of injection is the web space of the hind foot (106 cells/0.1 mL), which gives rise to tumors of similar behavior [107]. Wanebo et al. have developed a variation of these models in which B16F10 tumor cells are injected subcutaneously in the mid-tail of syngeneic mice [108]. 5 × 105 cells are injected in a total volume of 0.025 mL. Fully 100% of mice exhibit local tumor growth within 2–3 weeks of inoculation, and the majority develop inguinal adenopathy caused by lymphatic spread by 5–7 weeks. Mean survival time for tumor-bearing mice was found to be approximately 54 days. At autopsy, inguinal lymph nodes were noted to be markedly enlarged, and multiple metastatic lesions were visible in the lungs. If desired, the tumor can be resected at 2 weeks post-inoculation via amputation of the tail 5–10 mm distal to the base. If tumors are allowed to grow, they will reach a diameter of 15–20 mm by 6 weeks, displaying areas of necrosis and ulceration. This model is particularly useful for examining the effects of specific treatments on residual nodal disease following resection of the

274


primary lesion. All of these local models of melanoma are valuable because they recreate the major clinical stages of malignant melanoma; local tumor growth, (Stage I), involvement of regional lymph nodes (Stage II), and metastases to distant organs such as the lungs (Stage III). Several other models of B16 melanoma with unique characteristics have been described. B16 melanoma cells may also be injected subcutaneously into the dorsal surface of the ear [109]. This results in the formation of black tumor nodules that are visible within 3 days. Tumor growth proceeds rapidly in this site, and all mice are dead by day 22. An alternative site of tumor inoculation is the peritoneal cavity. Intraperitoneal (ip) injection of B16F1 melanoma cells (106/0.1 mL) into 6- to 8-week-old C57BL/6 mice will give rise to tumor nodules that grow rapidly and develop into solid intra-abdominal tumors [110]. In the absence of treatment, mice succumb within 3 weeks. Treatment with interferon-alpha consistently prolongs survival of tumor-bearing mice by 7–10 d. The B16F1 or the B16 F10 clones may be employed, depending on the need to induce the formation of distant metastases. This model has been used with success by Fleischmann et al. in the analysis of the effects of interferon-alpha on the growth of the B16 line. Cytotoxic treatments may given intraperitoneally beginning 1–5 days post-inoculation and one obvious advantage of this model is the ability to deliver cytokines and chemotherapeutic agents directly to the site of tumor growth [111]. Another ip model involves the implantation of gelatin sponges containing B16F10 melanoma cells and recombinant human basic fibroblast growth factor (bFGF) onto the serosal surface of the left lateral hepatic lobe of syngeneic C57BL/6 mice. Initially, tumor growth is localized within the gelatin sponge. However, peritoneal implants eventually develop, giving rise to peritoneal carcinomatosis. This model permits the evaluation of the cellular infiltrate induced by cytokine combinations, as well as characterization of the pattern of vascularity before and after treatment [112]. Induction of tumors in mice through the application of physical agents is a powerful approach to the study of melanoma but is hampered by the long period of time required for the establishment of transplantable tumors. The use of transgenic approaches is gaining greater popularity, especially as we gain a better understanding of the molecular basis of malignant melanoma. The obvious disadvantage of this approach is the need to identify suitable molecular targets for manipulation and the requirement for transgenic capabilities. Spontaneously arising tumors represent an important resource for the animal researcher, and numerous models are available. Unfortunately, no one model precisely recapitulates the metastatic cascade observed in human tumors. Other concerns relating to models that employ murine cell lines involve the potential for alterations to occur following prolonged in vitro culture or in vivo passage.

11.6.4 Tumor Models that Employ Immunodeficient Mice It has long been a goal of cancer researchers to propagate human tumors in other species in order to facilitate the study of tumor biology and evaluation of novel


275

therapeutic modalities. Early attempts focused on the implantation of tumors into privileged sites, such as the anterior chamber of the eye, the hamster cheek pouch, and the parenchyma of the brain. Tumors have also been implanted into fetuses and newborn animals which are naturally immunocompromised, as well as adult animals that have been rendered immune-deficient via thymectomy and total body irradiation (TBI) [113]. An alternative approach became available in the mid-1960s with the description of the congenitally athymic nude mouse, which arose on Balb/c background as the result of a mutation in a winged helix protein gene on chromosome 11 (Hfh11) [114]. The thymus is almost totally absent in nude mice because of failure of development of the thymic anlage, which arises from the ectoderm of the third pharyngeal pouch [115]. Functional T cells cannot develop to maturity in the absence of the thymic microenvironment, and therefore the nude mouse cannot efficiently reject human xenografts consisting of normal or neoplastic tissue. Another mouse that is equally useful is the severe combined immunodeficiency (SCID) mouse which arose as the result of a mutation in the C.B-17 strain. This syndrome is characterized by a complete lack of functional T cells and B cells in the adult mouse. It was later determined that the SCID phenotype resulted from an inactivating (nonsense) mutation in the gene encoding DNA-activated protein kinase (Prkdc) located on mouse chromosome 16. DNA-activated protein kinase functions in double-stranded DNA break repair and in recombination among the variable, diversity, and joining segments of immunoglobulin and T-cell receptor genes. Loss of this gene results in arrested development of T and B cells. Immunodeficient mice may be injected intracranially, intradermally, subcutaneously, intraperitoneally, or intravenously with cultured melanoma cell lines or cell suspensions derived from primary melanoma tumors or experimental tumors. Fragments of human tumors may also be implanted in various anatomic sites, with the expectation that engraftment will occur in a significant percentage of cases. The histological, molecular, and biochemical characteristics for human tumor xenografts are generally maintained in the nude mouse. Melanomas and colon carcinomas grow well in the nude mouse, whereas prostate carcinomas and leukemic tumors often fail to engraft [113]. Previous work by Fogh et al. has demonstrated that xenografts derived from recurrent tumors or metastatic lesions are more likely to engraft (50% and 39%, respectively) than are primary tumors (approximately 20%) [116]. A large number of cells or a large volume of primary tumor are generally required in order to ensure successful tumor take in the sc position, however, sc xenografts rarely metastasize unless special techniques are employed [113]. Most human tumors grown in nude mice exhibit a proliferative pattern similar to that seen in original tumor, especially if the cells are grown orthotopically (i.e. within the same organ from which they originated). The SCID mouse has been useful for studies of human tumor specimens and tumor-cell lines that grow with difficulty in other strains. This system also provides a much-needed mechanism for developing improved models of tumor metastasis. In addition, SCID mice may be reconstituted with immune cells from syngeneic mice or with human hematologic tissues (adult peripheral blood lymphocytes) to generate the so-called hu-PBL-SCID model [117]. Human immune cells can survive and remain functional for a considerable period of time, as

276


easured by the production of human immunoglobulin by implanted B cells. PBL m and tumor-infiltrating lymphocytes derived from cancer patients may be engrafted along with the primary tumor to create a host-tumor model [118]. This model is complicated by the potential for reactivation of latent Epstein–Barr virus (EBV) infection if EBV+ donors are employed, and EBV-induced human B-cell lymphomas routinely develop unless specific measures are employed [119]. 11.6.4.1 Nude Mouse Models Numerous nude mouse models of malignant melanoma have been described. These vary primarily according to the nature of the tumor inoculums and the details of the implantation procedure (e.g. sc or iv injection, or orthotopic implantation). Recent reviews of this topic offer a comprehensive survey of this topic [113, 120]. Of particular interest are nude mouse models, which closely approximate specific clinical entities, and models which accurately reflect the metastatic cascade that progresses from primary lesion to sc sites and nodal basins and then on to visceral organs. Most melanoma cell lines, such as the MeWo human melanoma cell line, will grow subcutaneously only if large numbers of cells are injected. However, tumortake and lethality are markedly enhanced when mice received sc implants of lung cubes that had been impregnated with small numbers of MeWo cells as a result of prior in vitro coincubation. Numerous, large lung nodules were found in one mouse receiving such implants and sc transfer of the metastatic spread to the lungs. Cell lines from such metastases or from primary tumors that arose from implanted long cubes were remarkable for their ability to colonize the lung following iv injection [121]. Also, iv injection of MeWo sublines derived from metastatic lesions consistently gave rise to extrapulmonary metastases. Other investigators have isolated sublines of parental tumors that exhibit a distinct tendency toward pulmonary metastasis following sc implantation or iv inoculation. Examples of these lines include the HSR+ MeWo variant, MM-RU, 451Lu, and CRML [122–125]. Another approach to the development of metastatic models in nude mice involves the use of cells derived from particularly aggressive human tumors. The (BRO) human melanoma cell line was derived from a tumor that exhibited very rapid growth, a distinct tendency for local recurrence, and rapidly fatal metastasis to visceral organs. Nude mice inoculated intraperitoneally with 106 BRO cells survived only 14 d. BRO cells metastasized to the lung and diaphragm following ip or sc injection, and exhibited a doubling time of just 2.3 days [126]. Cell lines with increased metastatic potential may also be established via the iv-injected human tumor cells and selection of tumor deposits that have colonized visceral sites. The LOX-L amelanotic human melanoma cell line was established in this fashion from the LOX parental line normally, which does not spread to visceral sites after sc inoculation [127]. LOX-L cells metastasize rapidly to the lungs following ip or sc injection and have been used to evaluate novel chemotherapeutic regimens in vivo. Fodstad et al. utilized this same general strategy to develop the FEMX-1 human melanoma cell line that preferentially metastasizes to sc sites [128].


277

Other investigators have determined that orthotopic (i.e. intradermal) injection of human melanoma cell lines increases the likelihood that the xenografted tumors will behave in a manner similar to the parent tumor. Rofstad established cell lines from sc deposits obtained from four different patients [129]. These cells were injected intradermally and evaluated for several parameters. It was found that the growth rate, histopathologic character, and angiogenic potential of the parent tumors were maintained in the orthotopically located xenografts. In addition, the organ-specific metastatic pattern of the xenografts closely resembled that seen in the donor patients. Lines with organ-specific metastatic patterns may also be generated via in vitro manipulations. Human melanoma cells with the propensity to metastatsize to the brain following iv inoculation were developed by culturing the parental melanoma line in increasing concentrations of wheat germ agglutinin (a toxic lectin compound). The resulting subline (70-W) consistently gave rise to brain and sc metastases as well as to lesions in the bone marrow, ovaries, muscle, and intra-abdominal organs [130]. Other models of interest include cell lines that induce severe cancer cachexia in nude mice when grown subcutaneously [131], and a cell line that induces a syndrome of diffuse hyperpigmentation (melanosis) caused by the release of pigment granules from the xenografted tumor and uptake of the structures by macrophages throughout the body [132]. 11.6.4.2 SCID Mouse Models Human melanoma tumors engraft readily into SCID mice, and exhibit a tendency to metastasize more readily and grow more rapidly than when implanted into nude mice. As with the nude mouse, various cell lines have been adapted to this model, and a variety of ingenious methods for the inoculation of tumor have been devised [120]. SCID mouse models have been used extensively in the analysis of basic tumor biology as well as the evaluation of various therapeutic modalities [133, 134]. It was discovered early on that cell lines that metastasize spontaneously from sc tumors in nude mice will do so more rapidly and with greater frequency when implanted into SCID mice [120]. Other investigators have evaluated the behavior of freshly isolated tumor specimens from metastatic lesions [135]. 100% tumor take was observed when cell suspensions were injected subcutaneously into SCID mice, and two-thirds of these tumors could be transplanted successfully into new hosts. As few as 5 × 105 cells were required to yield a 100% incidence of tumor formation. Interestingly, seven of nine tumors metastasized to distant sites on the first or second passage. The lungs were the primary site of metastasis, but spread to the abdominal viscera and thoracic lymph nodes was also noted. The expression of specific surface antigens was found to be maintained over the course of passaging in SCID mice. Tumor-associated lymphocytes were identified in the original tumor inoculums, but the presence of these immune cells did not appear to influence the outgrowth or metastatic potential of sc tumors. A more extensive study conducted by Taylor et al. investigated the engraftment and dissemination of human melanoma cells obtained from various sources [136].

278


They examined two cell lines, four early-passage cell lines, and fresh or cryopreserved cells from nine patient biopsies. SCID mice were inoculated via the ip, sc, and iv routes. The take-rate was highest for established cell lines (77%), and early-passage cell lines and fresh tumor cells engrafted in 65% and 53% of mice, respectively. Administration of tumor cells via the ip route resulted in tumor growth in 77% of mice, as compared to 41% for sc injection and 48% for iv injections. A distinct correlation was noted to exist between the number of cells injected and the percent of mice developing tumor: Only 26% engraftment was obtained when 1 × 106 cells were injected, whereas 69% take could be achieved with the injection of very large numbers of cells (50 × 106). Each tumor engrafted in at least one animal. Dissemination of tumor cells to distant organs via hematogenous or lymphatic spread was common and reproducible, with the number of metastases per animal averaging 16.3. The lung, liver, spleen, abdomen (viscera and peritoneum), and kidneys were the most common sites of spread. Moreover, the histologic character of the patient biopsy specimens was maintained after extensive passage in SCID hosts. A very sophisticated orthotopic model was reported by Juhasz et al. in 1993 [137]. SCID mice and nude mice were given full-thickness human skin grafts measuring approximately 1.5 cm in diameter. After the grafts were completely healed, 2 × 106 melanoma cells were injected intradermally in a volume of 50 mL into the skin transplants. It was theorized that the human skin grafts would provide the melanoma cells with the unique dermal environment necessary for optimal engraftment. The skin grafts themselves were successful in over 90% of mice. Seven different cell lines were employed in this study, and all seven were engrafted without difficulty. Interestingly, several of the melanoma lines (WM164, WM9, and 451Lu) grew as multiple nodules that infiltrated the grafts without effecting major changes in the overall architecture of the dermis. Other cell lines (WM582, WM793, and 1205Lu) appeared to infiltrate the human dermis along collagen fibrils, and seemed to have induced the formation of endothelial vessels. In each case, the overall pattern of invasion was found to be quite similar to that of the original patient biopsy. Moreover, cell lines that produced metastases when implanted subcutaneously in SCID mice also formed metastases in this model. For example, the metastatic cell lines 1205Lu and 405Lu possessed the ability to invade human dermis and disseminated to the lungs in the majority of animals. Chudnovsky et al. have developed a variant of this model wherein freshly isolated human keratinocytes were mixed with melanocytes that had been genetically manipulated to express mutations found in human melanomas, overlaid on human dermis, and then implanted into immunocompromised mice [138]. The contribution of individual mutations could then be assessed in a three-dimensional tissue environment. Immunodeficient mice provide an important model for the study of melanoma in which human tumors may be studied directly. Tumors with different clinical behaviors may be employed, and the effect of novel treatments may be assessed prior to the initiation of clinical trials. The use of human skin grafts and intradermal injections of tumor cells is a novel approach that may help us to understand the earliest stages of tumor invasion. It is important to remember that these mice are not


279

entirely immune-deficient, and the effects of the remaining effector cells (e.g. macrophages and NK cells in SCID mice) cannot be completely ignored. The cost of acquiring and housing immunodeficient mice is another limitation that prevents these models from being more widely employed.

References 1. Schartl A, Malitschek B, Kazianis S, Borowsky R, Schartl M. Spontaneous melanoma formation in non-hybrid Xiphophorous. Cancer Res. 1995;55:159–65. 2. Kusewit DF, Ley RD. Animal models of melanoma. Cancer Surv. 1996;26:35–70. 3. Schartl A, Pagany M, Engler M, Schartl M. Analysis of genetic factors and molecular mechanisms in the development of hereditary and carcinogen-induced tumors of Xiphophorus. Recent Results Cancer Res. 1997;143:225–35. 4. Anders F, Schartl M, Barnekow A, Anders A. Xiphophorus as an in vivo model for studies on normal and defective control of oncogenes. Adv Cancer Res. 1984;42:191–275. 5. Wittbrodt J, Adam D, Malitschek B, Maueler W, Raulf F, Telling A, et al. Novel putative receptor tyrosine kinase encoded by the melanoma-inducting Tu locus in Xiphophorus. Nature. 1989;341:415–21. 6. WIttbrodt J, Lammers R, Malitschek B, Ullrich A, Schartl M. The Xmrk receptor tyrosine kinase is activated in Xiphophorus malignant melanoma. EMBO J. 1992;11:4239–46. 7. Malitschek B, Wittbrodt J, Fischer P, Lammers R, Ulrich A, Schartl M. Autocrine stimulation of the Xmrk receptor tyrosine kinase in Xiphophorus melanoma cells and identification of a source for the physiological ligand. J Biol Chem. 1994;269:10423–30. 8. Adam D, Dimitrijevic N, Schartl M. Tumor suppression in Xiphophorus by an accidentally acquired promoter. Science. 1993;259:816–9. 9. Maueler W, Schartl A, Schartl M. Different expression patterns of oncogenes and protooncogenes in hereditary and carcinogen-induced tumors of Xiphophorus. Int J Cancer. 1993;55:288–96. 10. Winkler C, Wittbrodt J, Lammers R, Ullrich A, Schartl M. Ligand-dependent tumor induction in medakafish embryos by a Xmrk receptor tyrosine kinase transgene. Oncogene. 1994;9:1517–25. 11. Baudler M, Schartl M, Altschmied J. Specific activation of a STAT family member in Xiphophorus melanoma cells. Exp Cell Res. 1909;249:212–20. 12. Morcinek JC, Weisser C, Geissinger E, Schartl M, Wellbrock C. Activation of STAT5 triggers proliferation and contributes to anti-apoptotic signaling mediated by the oncogenic Xmrk kinase. Oncogene. 2002;21:1668–78. 13. Wellbrock C, Schartl M. Activation of phosphatidylinositol 3-kinase by a complex of p59fyn and the receptor tyrosine kinase Xmrk is involved in malignant transformation of pigment cells. Eur J Biochem. 2000;267:3513–22. 14. Delfgaauw J, Duschl J, Wellbrock C, Froschauer C, Schartl M, Altschmied J. MITF-M plays and essential role in transcriptional activation and signal transduction in Xiphophorus melanoma. Gene. 2003;320:117–26 15. Ozato K, Wakamatsu Y. Multi-step genetic regulation of oncogene expression in fish hereditary melanoma. Differentiation. 1983;24:181–90. 16. Ley RD. Photoreactivation in mammalian skin: mouse, marsupial and man. Photodermatol. 1987;4:173–75. 17. Ley RD, Applegate LA, Fry RJ, Sanchez AB. Photoreactivation of ultraviolet radiationinduced skin and eye tumors of Monodelphis domestica. Cancer Res. 1991;51:6539–42. 18. Ley RD, Applegate LA, Padilla RS, Stuart TD. Ultraviolet radiation-induced malignant melanoma in Monodelphis domestica. Photochem Photobiol. 1989;50:1–5.

280


19. Kusewitt DF, Applegate LA, Ley RD. Ultraviolet radiation-induced skin tumors in a South American opossum (Monodelphis domestica). Vet Pathol. 1991;28:55–65. 20. Wang Z, Atencio J, Robinson ES, McCarrey JR. Ultraviolet B-induced melanoma in Mondodelphis domestica occurs in the absence of alterations in the structure or expression of the p53 gene. Melanoma Res. 2001;11:239–45. 21. Chan J, Robinson ES, Yeh IT, McCarrey JR. Absence of ras gene mutations in UV-induced malignant melanomas correlates with a dermal origin of melanocytes in Monodelphis domestica. Cancer Lett. 2002;184:73–80. 22. Kusewitt DF, Miska KB, Ley RD. S-100 immunoreactivity in melanomas of the South American opossum Mondelphis domestica. Vet Pathol. 1997;34:346–50. 23. Robinson ES, VandeBerg JL, Hubbard GB, Dooley TP. Malignant melanoma in ultraviolet irradiated laboratory opposums: initiation in suckling young, metastasis in adults, and xenograft behavior in nude mice. Cancer Res. 1994;54:5986–91. 24. Kusewitt DF, Applegate LA, Bucana CD, Ley RD. Naturally occurring malignant melanoma in the South American opossum (Monodelphis domestica). Vet Pathol. 1990;27:66–8. 25. Wang Z, Hubbard GB, Clubb FJ, Vandeberg JL. The laboratory opossum (Monodelphis domestica) as a natural mammalian model for human cancer research. Int J Clin Exp Pathol. 2009;2:286–99. 26. Wang Z, Hubbard GB, Pathak S, VandeBerg JL. In vivo opossum xenografts model for cancer research. Cancer Res. 2003;63:6121–4. 27. Modiano JF, Ritt MG, Wojcieszyn J. The molecular basis of canine melanoma: pathogenesis and trends in diagnosis and therapy. J Vet Intern Med. 1976;13:163–74. 28. MacEwan EG. Spontaneous tumors in dogs and cats: models for the study of cancer biology and treatment. Cancer Metastasis Rev. 1990;9:125–36. 29. Marino DJ, Matthiesen DT, Stefanacci JD, Moroff SD. Evaluation of dogs with digit masses: 117 cases (1981–1991). J Am Vet Med Assoc. 1995;207:726–8. 30. Carpenter JL, Arohnson MG. Distal extremity melanocytic nevi and malignant melanoma in dogs. J Am Anim Hosp Assoc. 1990;26:605–12. 31. Todoroff RJ, Brodey RS. Oral and pharyngeal neoplasia in the dog: a retrospective survey of 361 cases. J Am Vet Med Assoc. 1979;175:567–71. 32. Kurzman ID, MacEwan EG. Effect of extent of surgery on survival in canine oral melanomas. Proc Annu Mtg Vet Cancer Soc. 1997;17:85 (abstract). 33. Theon AP, Rodriguez C, Griffey S, Madewell BR. Analysis of prognostic factors and patterns of failure in dogs with periodontal tumors treated with megavoltage irradiation. J Am Vet Med Assoc. 1997;210:785–8. 34. Blackwood L, Dobson JM. Radiotherapy of oral malignant melanomas in dogs. J Am Vet Med Assoc. 1996;209:98–102. 35. Bateman KE, Catton PA, Pennock PW, Kruth SA. 0–7–21 radiation therapy for the treatment of canine oral melanoma. J Vet Intern Med. 1994;8:267–72. 36. Moore AS, Theilen GH, Newell AD, Madewell BR, Rudolf AR. Preclinical study of sequential tumor necrosis factor and interleukin 2 in the treatment of spontaneous canine neoplasms. Cancer Res. 1991;51:233–8. 37. MacEwen EG, Patnaik AK, Harvey HJ, Hayes AA, Matus R. Canine oral melanoma: comparison of surgery versus surgery plus Corynebacterium parvum. Cancer Invest. 1986;4:397–402. 38. MacEwen EG, Kurzman ID, Vail DM, Dubielzig RR, Everlith K, Madewell BR, et al. Adjuvant therapy for melanoma in dogs: results of randomized clinical trials using surgery, liposome-encapsulated muramyl tripeptide, and granulocyte-macrophage colony-stimulating factor. Clin Cancer Res. 1999;5:4249–58. 39. Hogge GS, Burkholder JK, Culp J, Albertini MR, Dubielzig RR, Yang NS, et al. Preclinical development of human granulocyte-macrophage colony-stimulating factor-transfected melanoma cell vaccine using established canine cell lines and normal dogs. Cancer Gene Ther. 1979;6:26–36.


281

40. Hogge GS, Burkholder JK, Culp J, Albertini MR, Dubielzig RR, Keller ET, et al. Development of human granulocyte-macrophage colony-stimulating factor-transfected tumor cell vaccines for the treatment of spontaneous canine cancer. Hum Gene Ther. 1998;9:1851–61. 41. Finocchiaro LM, Glikin GC. Cytokine-enhanced vaccine and suicide gene therapy as surgery adjuvant treatments for spontaneous canine melanoma. Gene Ther. 2008;15:267–76. 42. Dow SW, Elmslie RE, Willson AP, Roche L, Gorman C, Potter TA. In vivo transfection with superantigen plus cytokine genes induces tumor regression and prolongs survival in dogs with malignant melanoma. J Clin Investig. 1998;101:2406–14. 43. Misfeldt ML, Grimm DR. Sinclair miniature swine: an animal model of human melanoma. Vet Immunol Immunopathol. 1994;43:167–75. 44. Hook RR Jr, Aultman MD, Adelstein EH, Oxenhandler RW, Millikan LE, Middleton CC. Influence of selective breeding on the incidence of melanomas in Sinclair miniature swine. Int J Cancer. 1979;24:668–72. 45. Horak V, Fortyn K, Hruban V, Klaudy J. Hereditary melanoblastoma in miniature pigs and its successful therapy by devitalization technique. Cell Mol Biol. 1999;45:1119–29. 46. Tissot RG, Beattie CW, Amoss MS Jr, Williams JD, Schumacher J. Common swine leucocyte antigen (SLA) haplotypes in NIH and Sinclair miniature swine have similar effects on the expression of an inherited melanoma. Anim Genet. 1993;24:191–93. 47. Hook RR Jr, Berkelhammer J, Oxenhandler RW. Melanoma: Sinclair swine melanoma. Am J Pathol. 1982;108:130–33. 48. Du ZQ, Vincent-Naulleau S, Gilbert H, Vignoles F, Crechet F, Shimogiri T, Yasue H, Leplat JJ, Bouet S, Gruand J, Horak V, Milan D, Le Roy P, Geffrotin C. Detection of novel quantitative trait loci for cutaneous melanoma by genome-wide scan in the MeLiM swine model. Int J Cancer. 2007;120:303–20. 49. Amoss MS Jr, Ronan SG, Beattie CW. Growth of Sinclair swine melanoma as a function of age, histopathological staging, and gonadal status. Cancer Res. 1988;48:1708–11. 50. Greene JF Jr, Townsend JS 4th, Amoss MS Jr. Histopathology of regression in Sinclair swine model of melanoma. Lab Investig. 1994;71:17–24. 51. Flatt RE, Nelson LR, Middleton CC. Melanotic lesions in the internal organs of miniature swine. Arch Pathol. 1972;93:71–5. 52. Das Gupta TK, Ronan SG, Beattie CW, Shilkaitis A, Amoss MS Jr. Comparative histopathology of porcine and human cutaneous melanoma. Pediatr Dermatol. 1989;6:289–99. 53. Aultman MD, Hook RR Jr. In vitro lymphocyte reactivity to soluble tumor extracts in Sinclair melanoma swine. Int J Cancer. 1979;24:673–8. 54. Rambow F, Piton G, Bouet S, Leplat JJ, Baulande S, Marrau A, Stam M, Horak V, VincentNaulleau S. Gene expression signature for spontaneous cancer regression in melanoma pigs. Neoplasia. 2008;10:714–26. 55. Rambow F, Malek O, Geffrotin C, Leplat JJ, Bouet S, Piton G, Hugot K, Bevilacqua C, Horak V, Vincent-Naulleau S. Identification of differentially expressed genes in spontaneously regressing melanoma using the MeLiM swine model. Pigment Cell Melanoma Res. 2008;21:147–61. 56. Berkelhammer J, Oxenhandler RW, Hook RR, Hennessy JM. Development of a new melanoma model in C57BL/6 mice. Cancer Res. 1982;42:3157–63. 57. Kripke ML. Speculations on the role of ultraviolet radiation in the development of malignant melanoma. J Natl Cancer Inst. 1979;63:541–8. 58. Townsend SE, Allison JP. Tumor rejection after direct constimulation of CD8+ T cells by B7-transfected melanoma cells. Science. 1993;259:368–70. 59. Owen-Schaub LB, van Golen KL, Hill LL, Price JE. Fas and Fas ligand interactions suppress melanoma lung metastasis. J Exp Med. 1998;188:1717–23. 60. Gause PR, Lluria-Prevatt M, Keith WN, Balmain A, Linardopolous S, Warneke J, et al. Chromosomal and genetic alterations of 7,12-dimethylbenz[a]anthracene-induced melanoma from TP-ras transgenic mice. Mol Carcinogenesis. 1997;20:78–87. 61. Epstein JH, Epstein JL, Nakai T. Production of melanomas from DMBA-induced “blue nevi” in hairless mice with ultraviolet light. J Natl Cancer Inst. 1967;38:19–30.

282


62. Husain Z, Viola MV. Role of ultraviolet radiation in the induction of melanocytic tumors in hairless mice following 7,12-dimethyl(a)benzanthracene application and ultraviolet irradiation. Cancer Res. 1991;51:4964–70. 63. Donawho CK, Kripke ML. Photoimmunology of experimental melanoma. J Immunotherapy. 1991;10:177–88. 64. Van Weelden H, van der Putte SCJ, Toonstra J, van der Leun JC. Ultraviolet B-induced tumors in pigmented hairless mice, with an unsuccessful attempt to induce cutaneous melanoma. Photodermatol Photoimmunol Phtotomed. 1990;7:68–72. 65. Pour P, Althoff J, Salmasi SZ, Stephan K. Spontaneous tumors and common diseases in three types of hamsters. J Natl Cancer Inst. 1979;63:797–811. 66. Pawlowski A, Lea PJ. Nevi and melanoma induced by chemical carcinogenesis in laboratory animals: similarities and differences with human lesions. J Cutaneous Path. 1983;10:81–110. 67. Mintz B, Silvers WK. Transgenic mouse model of malignant skin melanoma. Proc Natl Acad Sci USA. 1993;90:8817–21. 68. Beermann F, Hunziker A, Folletti A. Transgenic mouse models for tumors of melanocytes and retinal pigment epithelium. Pigment Cell Res. 1999;12:71–80. 69. Saenz-Santamaria MC, McNutt NS, Bogdany JK, Shea CR. P53 expression is rare in cutaneous melanomas. Am J Dermatopathol. 17:344–9. 70. Chin L. Modeling malignant melanoma in mice: pathogenesis and maintenance. Oncogene. 199;18:5304–10. 71. Sherr CJ, Roberts JM. CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev. 1999;13:1501–12. 72. Serrano M, Lee H, Chin L, Cordon-Cardo C, Beach D, DePinho R. Role of the INK4a locus in tumor suppression and cell mortality. Cell. 1996;85:27–37. 73. Ackermann J, Frutschi M, Kaloulis K, McKee T, Trumpp A, Beerman F. Metastasizing melanoma formation caused by expression of activated N-RasQ61K on an INK4a-deficient background. Cancer Res. 2005;65:4005–11. 74. Chin L, Pomerantz J, Polsky D, Jacobson M, Cohen C, Cordon-Cardo C, et al. Cooperative effects of INK4a and ras in melanoma susceptibility in vivo. Genes Dev. 1997;11:2822–34. 75. Chin L, Tam A, Pomerantz J, Wong M, Holash J, Bardeesy N, et al. Essential role for oncogenic Ras in tumor maintenance. Nature. 1999;400:468–72. 76. Iwamoto T, Takahashi M, Ito M, Hamatani K, Ohbayashi M, Wajjwalku W, Isobe K, Nakashima I. Aberrant melanogenesis and melanocytic tumour development in transgenic mice that carry a metallothionein/ret fusion gene. EMBO J. 1991;10:3167–75. 77. Schmidt A, Tief K, Yavuzer U, Beermann F. Ectopic expression of RET results in microphthalmia and tumors in the retinal pigment epithelium. Int J Cancer. 199;80:600–5. 78. Noonan FP, Otsuka T, Bang S, Anver MR, Merlino G. Accelerated ultraviolet radiationinduced carcinogenesis in hepatocyte growth factor/scatter factor transgenic mice. Cancer Res. 2000;60:3738–43. 79. Noonan FP, Recio JA, Takayama H, Duray P, Anver MR, Rush WL, DeFabo EC, Merlino G. Nature. 2001;413:271–2. 80. De Fabo EC, Noonan FP, Fears T, Merlino G. Ultraviolet B but not ultraviolet A radiation initiates melanoma. Cancer Res. 2004;64:6372–6. 81. Harding HE, Passey RD. A transplantable melanoma of the mouse. J Pathol Bacteriol. 1930;33:417–27. 82. Cloudman AM. The effect of an extra-chromosomal influence upon transplanted spontaneous tumors in mice. Science. 1941;93:380–1. 83. Stackpole CW. Generation of phenotypic diversity in the B16 mouse melanoma relative to spontaneous metastasis. Cancer Res. 1983;43:3057–65. 84. Setiawan Y, Moore DE, Allen BJ. Selective uptake of boronated low-density lipoprotein in melanoma xenografts achieved by diet supplementation. Br J Cancer. 1996;74:1705–8. 85. Lopez Ballester JA, Penafiel R, Cremades A, Valcarcel MM, Solano F, Lozano JA. Effects of treatments with alpha-difluoromethylornithine and hyperthermia on the growth and polyamine metabolism of Harding-Passey murine melanoma. Anticancer Res. 1991;11:691–6.


283

86. Randazzo BP, Kesari S, Gesser RM, Alsop D, Ford JC, Brown SM, et al. Treatment of experimental intracranial murine melanoma with a neuroattenuated herpes simplex virus 1 mutant. Virology. 1995;211:94–101. 87. Fand I, Sharkey RM, Goldenberg DM. Use of whole-body autoradiography in cancer targeting with radiolabeled antibodies. Cancer Res. 1990;50:885s–891s. 88. Yong JH, Barth RF, Wyzlic IM, Soloway AH, Rotaru JH. In vitro and in vivo evaluation of o-carboranylalanine as a potential boron delivery agent for neutron capture therapy. Anticancer Res. 1995;15:2033–8. 89. Watts KP, Fairchild RG, Slatkin DN, Greenberg D, Packer S, Atkins HL, et al. Melanin content of hamster tissues, human tissues, and various melanomas. Cancer Res. 1981;41:467–72. 90. Knight GD, Laubscher KH, Fore ML, Clark DA, Scallen TJ. Vitalethine modulates erythropoiesis and neoplasia. Cancer Res. 1994;54:5623–35. 91. Horton HM, Anderson D, Hernandez P, Barnhart KM, Norman JA, Parker SE. A gene therapy for cancer using intramuscular injection of plasmid DNA encoding interferon alpha. Proc Natl Acad Sci USA. 1999;96:1552–8. 92. Nordlund JJ, Ackles A, Gershon RK. Quantitative factors which determine the effect of the immune response on the growth rate of the Cloudman Melanoma in the DBA/2 mice. Cell Immunol. 1980;56:258–72. 93. Gehlsen KR, Hadley ME, Levine N, Ray CG, Hendrix MJ. Effects of a melanotropic peptide on melanoma cell growth, metastasis, and invasion. Pigment Cell Res. 1992;5:219–23. 94. Alterman AL, Fornabaio DM, Stackpole CW. Metastatic dissemination of B16 melanoma: pattern and sequence of metastasis. J Natl Cancer Inst. 185;75:691–702. 95. Fidler IJ, Gruys E, Cifone MA, Barnes Z, Bucana C. Demonstration of multiple phenotypic diversity in a murine melanoma of recent origin. J Natl Cancer Inst. 1981;67:947–56. 96. Stackpole CW. Generation of phenotypic diversity in the B16 mouse melanoma relative to spontaneous metastasis. Cancer Res. 1983;43:3057–65. 97. Poste G, Nicolson GL. Arrest and metastasis of blood-borne tumor cells are modified by fusion of plasma membrane vesicles from highly metastatic cells. Proc Natl Acad Sci USA. 1980;77:399–403. 98. Fidler IJ, Kripke ML. Metastasis results from preexisting variant cells within a malignant tumor. Science. 1977;197:893–5. 99. Hart IR. The selection and characterization of an invasive variant of the B16 melanoma. Am J Pathol. 1979;97:587–600. 100. Nicolson GL, Dulski K, Basson C, Welch DR. Preferential organ attachment and invasion in vitro by B16 melanoma cells selected for differing metastatic colonization and invasive properties. Invasion Metastasis. 1985;5:144–58. 101. Nathanson SD, Nelson LT, Lee M. A spontaneous subcutaneous tumor in C57BL/6 mice that metastasizes to the liver. Clin Exp Metastasis. 1993;11:45–54. 102. Brunda MJ, Luistro L, Warrier RR, Wright RB, Hubbard BR, Murphy M, et al. Antitumor and antimetastatic activity of interleukin 12 against murine tumors. J Exp Med. 1993;178:1223–30. 103. Kobayashi M, Kobayashi H, Pollard RB, Suzuki F. A pathogenic role of Th2 cells and their cytokine products on the pulmonary metastasis of murine B16 melanoma. J Immunol. 1998;160:5869–73. 104. Stackpole CW. Intrapulmonary spread of established B16 melanoma lung metastases and lung colonies. Invasion Metastasis. 1990;10:267–80. 105. Nathanson SD, Haas GP, Mead MJ, Lee M. Spontaneous regional lymph node metastases of three variants of the B16 melanoma: relationship to primary tumor size and pulmonary metastases. J Surg Oncol. 1986;33:41–5. 106. Markovic SN, Murasko DM. Role of natural killer and T cells in interferon induced inhibition of spontaneous metastases of the B16F10L murine melanoma. Cancer Res. 1991;51:1124–8. 107. Dyce M, Sharif SF, Whalen GF. Search for anti-metastatic therapy: effects of phenytoin on B16 melanoma metastasis. J Surg Oncol. 1992;49:107–12.

284


108. Shrayer DP, Bogaars H, Wolf SF, Hearing VJ, Wanebo HJ. A new mouse model of experimental melanoma for vaccine and lymphokine therapy. Int J Oncol. 1998;13:361–74. 109. Shiohara T, Moellman G, Jacobson K, Kuklinska E, Ruddle NH, Lerner AB. Anti-tumor activity of class II MHC antigen-restricted cloned autoreactive T cells. II. Novel immunotherapy of B15 melanomas by local and systemic adoptive transfer. J Immunol. 1987;138:1979–86. 110. Fleischmann CM, Stanton GJ, Fleischmann WR Jr. Enhanced in vivo sensitivity to interferon with in vitro resistant B16 tumor cells in mice. Cancer Immunol Immunother. 1994;39:148–54. 111. Fleischmann WR Jr, Msoor J, Wu TY, Fleischmann CM. Orally administered IFN-alpha acts alone and in synergistic combination with intraperitoneally administered IFN-gamma to exert an antitumor effect against B16 melanoma in mice. J Interferon Cytokine Res. 1998;18:17–20. 112. Watanabe M, McCormick KL, Volker K, Ortaldo JR, Wigginton JM, Brunda MJ, et al. Regulation of local host-mediated anti-tumor mechanisms by cytokines: direct and indirect effects on leukocyte recruitment and angiogenesis. Am J Pathol. 1997;150:1869–80. 113. Giovanella BC, Fogh J. The nude mouse in cancer research. Adv Cancer Res. 1985;44:69–120. 114. Nehls M, Pfeifer D, Schorpp M, Hedrich H, Boehm T. New member of the winged-helix protein family disrupted in mouse and rat nude mutations. Nature. 1994;372:103–7. 115. Pantelouris EM. Absence of thymus in a mouse mutant. Nature. 1968;217:370–1. 116. Fogh J, Orfeo T, Tiso J, Sharkey FE, Fogh JM, Daniels WP. Twenty-three new human tumor lines established in nude mice. Exp Cell Biol. 1980;48(3):229–39. 117. Mosier DE, Gulizia RJ, Baird SM, Wilson DB. Transfer of a functional human immune system to mice with severe combined immunodeficiency. Nature. 1998;335:256–9. 118. Bankert RB, Umemoto T, Suguyama Y, Chen FA, Repasky E, Yokota S. Human lung tumors, patients’ peripheral blood lymphocytes and tumor infiltrating lymphocytes propagated in scid mice. Curr Topics Microbiol Immunol. 1989;152:201–10. 119. Baiocchi RA, Caligiuri MA. Low-dose interleukin 2 prevents the development of EpsteinBarr virus (EBV)-associated lymphoproliferative disease in scid/scid mice reconstituted i.p. with EBV-seropositve human peripheral blood lymphocytes. Proc Natl Acad Sci USA. 1994;91:5577–81. 120. Rofstad EK, Lyng H. Xenograft model systems for human melanoma. Mol Med Today. 1996;2:394–403. 121. Kerbel RS, Man MS, Dexter D. A model of human cancer metastasis: extensive spontaneous and artificial metastasis of a human pigmented melanoma and derived variant sublines in nude mice. J Natl Cancer Inst. 1984;72:93–108. 122. Gitelman I, Dexter DF, Roder JC. DNA amplification and metastasis of the human melanoma cell line MeWo. Cancer Res. 1987;47:3851–5. 123. Gattoni-Celli S, Byers RH, Calorini L, Ferrone S. Organ-specific metastases in melanoma: experimental animal models. Pigment Cell Res. 1993;6:381–4. 124. Heryln D, Adachi K, Koprowski H, Herlyn M. In vitro properties of human melanoma cells metastatic in nude mice. Cancer Res. 1990;50:2296–302. 125. Herlyn D, Adachi K, Koprowski H, Herlyn M, Nathanson L (ed.). Experimental model of human melanoma metastases. In: Melanoma research: genetics, growth factors, metastases, and antigens. Kluwer Academic Publishers, Boston, MA, 1991. p. 105–118. 126. Lockshin A, Giovanella BC, De Ipolyi PD, Williams LJ Jr, Mendoza JT, Yim SO, et al. Exceptional lethality for nude mice of cells derived from a primary human melanoma. Cancer Res. 1985;45:345–50. 127. Shoemaker RH, Dykes DJ, Plowman J, Harrison SD Jr, Griswold DP Jr, Abbott BJ, et al. Practical spontaneous metastasis model for in vivo therapeutic studies using a human melanoma. Cancer Res. 1991;51:2837–41. 128. Fodstad O, Kjonniksen I, Aamdal S, Nesland JM, Boyd MR, Pihl A. Extrapulmonary, tissuespecific metastasis formation in nude mice injected with FEMX-I human melanoma cells. Cancer Res. 1988;48:4382–8.


285

129. Rofstad EK. Orthotopic human melanoma xenografts model systems for studies of tumour angiogenesis, pathophysiology, treatment sensitivity and metastatic pattern. Br J Cancer. 194; 70:804–12. 130. Ishikawa M, Fernandez B, Kerbel RS. Highly pigmented human melanoma variant which metastasizes widely in nude mice, including to skin and brain. Cancer Res. 1988;48:4897–903. 131. Mori M, Yamaguchi K, Honda S, Nagasaki K, Ueda M, Abe O, et al. Cancer cachexia syndrome developed in nude mice bearing melanoma cells producing leukemia-inhibitory factor. Cancer Res. 1991;51:6656–9. 132. Spremulli EN, Bogaars HA, Dexter DL, Matook GM, Jolly GA, Kuhn RE, et al. Nude mouse model of the melanosis syndrome. J Natl Cancer Inst. 1983;71:933–9. 133. Jansen B, Schlagbauer-Wadl H, Brown BD, Bryan RN, van Elsas A, Muller M, et al. bcl-2 antisense therapy chemosensitizes human melanoma in SCID mice. Nat Med. 1998;4:232–4. 134. Naramura M, Gillies SD, Mendelsohn J, Reisfeld RA, Mueller BM. Therapeutic potential of chimeric and murine anti-(epidermal growth factor receptor) antibodies in a metastasis model for human melanoma. Cancer Immunol Immunother. 1993;37:343–9. 135. Hill LL, Korngold R, Jaworsky C, Murphy G, McCue P, Berd D. Growth and metastasis of fresh human melanoma tissue in mice with severe combined immunodeficiency. Cancer Res. 1991;51:4937–41. 136. Taylor CW, Grogan TM, Lopez MH, Leong SP, Odeleye A, Feo-Zuppardi FJ, et al. Growth and dissemination of human malignant melanoma cells in mice with severe combined immune deficiency. Lab Invest. 1992;67:130–7. 137. Juhasz I, Albelda SM, Elder DE, Murphy GF, Adachi K, Herlyn D, et al. Growth and invasion of human melanomas in human skin grafted to immunodeficient mice. Am J Pathol. 1993;143:528–37. 138. Chudnovsky Y, Adams AE, Robbins PB, Lin Q, Khavari PA. Use of human tissue to assess the oncogenic activity of melanoma-associated mutations. Nat Genet. 2005;37:745–9. 139. Fogh J, Dracopoli N, Loveless JD, Fogh H. Cultured human tumor cells for cancer research: assessment of variation and stability of cultural characteristics. Prog Clin Biol Res. 1982;89:191–223.

Chapter 12

Experimental Animal Models for Investigating Renal Cell Carcinoma Pathogenesis and Preclinical Therapeutic Approaches Gilda G. Hillman

Abstract Advanced metastatic renal cell carcinoma (RCC) is poorly responsive to conventional treatment including most chemotherapeutic drugs, hormones and radiation therapy. Although recent developments in anti-angiogenic therapy have improved targeting these highly vascularized tumors, the treatment of metastatic disease has been and remains a difficult clinical challenge. Pre-clinical studies in animal tumor models of RCC are essential to address new therapeutic approaches for metastatic disease. Previous studies have shown that animal models are also useful to improve our understanding of the progression and molecular genetics of the disease in order to tailor the proper therapy to each type and stage of RCC. The goals of this chapter are to review various experimental RCC models used in numerous preclinical studies and provide an update on recent developments in this field. Keywords Renal cell carcinoma • Animal models • Pathogenesis • Therapy

12.1 Introduction The incidence of renal cell carcinoma (RCC) is continuing to increase with approximately 57,760 new cases each year in the United States of America [1]. This increased RCC incidence may be linked to certain risk factors including smoking, obesity, high protein diets and hypertension [2, 3]. The disease is responsible for an estimated 12,980 deaths each year as updated by ACS for 2009 [1]. Nearly half of the patients present only with localized disease that can be treated by surgical removal [2–5]. However, one-third of the patients have metastatic disease at first presentation, and 30–50% of the patients treated for localized RCC subsequently

G.G. Hillman (*) Department of Radiation Oncology, Karmanos Cancer Institute, Wayne State University School of Medicine, Detroit, MI 48201, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_12, © Springer Science+Business Media, LLC 2011

287

288

G.G. Hillman

develop metastatic disease [4, 5]. Patients with metastatic RCC frequently present with pulmonary metastases [4, 5]. Historically, the median survival of patients with metastatic RCC ranged from 8 to 11 months [4–6]. Debulking nephrectomy has a positive survival impact in patients with metastatic RCC, as demonstrated in two randomized trials [6, 7]. Within the patients benefiting from debulking nephrectomy, the subgroup of patients with metastasis only to the lung appeared to have a survival benefit with a median survival of 14 months vs. 10 months in patients not undergoing nephrectomy. Metastatic RCC is poorly responsive to conventional treatment including most chemotherapeutic drugs, hormones and radiation therapy [2, 4, 5]. Although recent developments in anti-angiogenic therapy have improved targeting these highly vascularized tumors, the treatment of metastatic disease has been and remains a difficult clinical challenge. To develop new and alternative therapeutic modalities for metastatic disease, various animal models were developed and used in numerous preclinical studies. These models were also useful to investigate the metastatic progression of RCC and the molecular genetics of the disease. The properties of an ideal tumor model for RCC are spontaneous origin, histologically proven adenocarcinoma, predictable growth rate and ability to metastasize similarly to human RCC in a reasonable time frame [8, 9]. This chapter reviews several experimental models, including the murine Renca renal adenocarcinoma in Balb/c mice, the rat kidney carcinoma in Wistar–Lewis rat, the Eker rat model of hereditary RCC and human RCC tumor xenograft models in athymic nude mice. These models have been well characterized and extensively used to study the pathogenesis of RCC disease and to assess the efficacy and safety of novel treatment modalities. Novel recent animal models will be discussed in this updated chapter.

12.2 Murine Syngeneic Renal Adenocarcinoma: The Renca Model In the early 1970s, the Renca murine renal adenocarcinoma model was isolated and characterized by Hrushesky and Murphy [10]. The Renca tumor arose spontaneously in the kidney of a Balb/c mouse, and was found to induce metastatic kidney tumors when injected under the renal capsule of Balb/c mice [10]. Renca was histologically characterized as a poorly differentiated renal cortical adenocarcinoma of the granular type, pleomorphic with large nuclei [8–10]. Renca can be maintained by either in vitro culture or in vivo passage by intraperitoneal (i.p.) injection or by subcapsular renal injection in syngeneic Balb/c mice. The progression of Renca tumor following subcapsular implantation mimics that of human RCC because of the formation of a primary tumor mass on the kidney followed by the development of spontaneous metastases [10–12]. Metastases develop in the regional lymph nodes, lung, liver and peritoneum; thus Renca can be staged similarly to human RCC [10–12]. The renal Renca model allows evaluation of the therapy, on the primary tumor as well as on metastatic deposits. A nephrectomy of the tumor-bearing

12 Animal Models for Kidney Cancer

289

kidney can be performed [12], therefore, the model is applicable for the development of therapeutic protocols for advanced metastatic disease, similar to the clinical situation of postnephrectomy metastatic RCC patients. The mean survival time of Renca-bearing mice is approximately 46 days when 105 cells are implanted intrarenally, thus allowing for a therapeutic evaluation in a reasonable time frame. On the basis of these properties, Renca does represent an ideal RCC tumor model and indeed has been used extensively as a preclinical model to investigate various therapeutic approaches for metastatic RCC. In the 1970s, early studies in this model have included hormonal therapy, and chemotherapy with single or combined drugs that were found of limited efficacy [8, 9]. The immunogenicity of this tumor is relatively low, although some protection to rechallenge with viable Renca cells following immunization with crude membrane preparations from Renca cells has been reported [13]. This model is a valuable tool in testing immunotherapy approaches. Immunotherapy is a novel therapeutic approach developed in the 1980s for the treatment of disseminated cancers refractory to conventional treatments. This approach utilizes biological response modifiers (BRMs)/cytokines and immune cells capable of enhancing immune mechanisms directed against the tumor that may be present although ineffective in cancer-bearing hosts [14]. The most commonly used BRMs are interferons (IFN) and the lymphokine interleukin 2 (IL-2). In the 1980s, Wiltrout and co-workers have developed chemoimmunotherapy combining the administration of chemotherapeutic agents (doxorubicin hydrochloride or flavone acetic acid) with adoptive immunotherapy (IL-2 and lymphokine activated killer cells) for the treatment of Renca that resulted in significant antitumor responses [11, 12, 15]. These preclinical studies were translated into clinical trials for RCC patients [16]. Other cytokines such as IL-7, IL-1 or combination of cytokines including IFNa/IL-2 or IFNa/IFNg were tested [8, 9, 15]. Gregorian and Battisto have demonstrated the existence of immunosuppressive effects induced by the tumor cells in Renca-bearing mice [17, 18]. Generation of specific cytotoxic T lymphocytes to Renca cells is particularly difficult when using irradiated Renca cells in in vitro assays due to their immunosuppressive activity (personal communications). In the 1990s, we had further developed the Renca model to address the efficacy and mechanism of action of cytokine therapy alone or combined with radiation therapy [19–29]. We have defined the kinetics of the tumor model following Renca cell implantation in various sites. A concentration of 105 cells can be administered in 0.1-ml Hank’s balanced salt solution (HBSS) for i.p. or flank subcutaneous (s.c.) injections or in 0.5-ml HBSS for i.v. injection via a tail vein. For kidney implantation, the right kidney of anesthetized mice is exposed through a right flank incision and injected subcapsularly with 105 Renca cells in 50-µl HBSS using a 27-gauge needle. Intraperitoneal injection of Renca induced metastases in the mesenteric lymph nodes starting by day 16 and progressing to carcinomatosis [21]. Liver metastases were observed in 38% of the mice, and lung metastases were detected in 5% [21]. Following s.c. injections of Renca cells in the right flank, small tumors were detectable by day 14 and grew progressively reaching a size of 1–1.5 cm3 [21]. Large tumors showed a tendency to become ulcerative and necrotic. Tumors remained localized at the site of injection and metastases were not detectable in other organs.

290

G.G. Hillman

Following kidney implantation, a macroscopic primary tumor was detectable by days 7–10, which then grew progressively to 1–2 cm3 by day 20–21 [19]. Large tumors of 7–8 cm3 were present by days 25–35. Pulmonary metastases were first noted by days 15–20 following renal implantation. However, large variations in the number of lung metastases were observed from mouse to mouse. Metastases to the liver, hemorrhagic ascites and/or carcinomatosis were also observed in most animals after 21 days, confirming previous studies. Similar to previous reports mice began to die on day 21, with a 50% survival rate by day 37 and less than 10% mice survived more than 45 days [19]. To increase the incidence and number of pulmonary metastases, we have i.v. injected Renca cells and observed visible tumor nodules of 10 µm induced mesothelioma in 59.4% and crocidolite with 9 million fibers >10 µm in 56.3% of animals at the same dose. Wagner similarly showed that the length of the crocidolite fiber is directly proportional to its tumorigenic ability [21]. Samples that were milled for 4 and 8 h, and as a result contained far fewer crocidolite fibers >6.5 µm than 1- and 2-h milled fibers, induced mesothelioma in 34% of the animals compared to 80% for the 1- and 2-h group. A rat study by Miller found that injecting a dose containing 109 amosite particles >5 µm led to an incidence of mesothelioma in 88% of the animal group, confirming that a dose containing a large number of long fibers will be successful in promoting tumor growth [25]. In the rat species, the appearance of effusions and solid-tumor nodules after inoculation is often associated with the development of mesothelioma. When these are present, the tumors have been characterized as predominantly biphasic and

13 Animal Models of Mesothelioma

311

spindle-shaped (sarcomatous), but few reports indicate that all the patterns observed in humans – including tubular, papillary, solid, and spindle cell – are possible [23, 25]. In addition, several different patterns have been seen in the same tumor, and metastases are common. Intraperitoneally injected rats had a mean survival time between 509 and 1,002 days. In general, there is a direct relationship between administered dose, the ability of the fiber to induce mesothelioma, and average survival time (i.e. high doses and very toxic fiber administration resulted in a decrease of the life span of the animal). Amosite administration resulted in an average survival time between 462 and 889 days, crocidolite in 416–1,002 days, and chrysotile in 476–903 days – all depending on the administered dose, type of chrysotile, and milling time [23, 27].

13.2.5 Intraperitoneal Asbestos Injection: Mice Various rates of tumor induction have been reported in the murine mesothelioma model, ranging from 25% to 45% [20, 28]. Davis reported tumor growth in 25% of Balb/c mice and 45% of CBA mice treated with Wittenoon George crocidolite [28]. A study on the carcinogenic potential of amosite, chrysotile, and calindra chrysotile by Suzuki showed amosite to be the most mesotheliomatogenic fiber, inducing tumors in 40.5% of Balb/c mice, followed by Calidria chrysotile (25%) and standard chrysotile (0%) at the administered dose [20]. The relationship between fiber toxicity and number of long fibers, which has been observed in the rat mesothelioma model, holds true for the murine model as well. In the Suzuki study discussed here, amosite was the most mesotheliomatogenic fiber – causing tumors in 40% of the treated group – presumably because it had the highest percentage of long fibers (6% >7.5 µm). Calindra chrysotile, with the second highest percentage of long fibers (4.6% >5 µm), induced mesothelioma in just 25% of animals, and chrysotile with only 2% of fibers >3 µm was found to be nontumorigenic at the given dose [20]. In the murine animal model, the latency period has been established at 7 months [20, 28]. It has been speculated that this period is shorter in the mouse than in many other species, because of the animal’s relative short life span. Ultrastructural analysis of mesothelioma in the mouse species provides a blurred picture on the predominant cell type. Davis reports that while all three histological forms of human malignant mesothelioma were present, as with the human disease, the majority of malignant cells identified in the ascites were epithelial tumors exhibiting typical mesothelial differentiation, with long, thin microvilli, intermediate filaments, numerous microscopic vesicles, and much glycogen [28]. No evidence of metastases was noted. Suzuki similarly notes the presence of ascites in most mesothelioma cases, but the vast majority of fibrous tumors and the small minority of biphasic tumors reported are in stark contrast with both the history of human disease and Davis’ findings [20]. Tumors grew preferentially in the omentum, mesentery, and serosae of the gastrointestinal and

312

H.I. Pass et al.

genital organs, the diaphragm, the capsule of liver and spleen, and the abdominal wall peritoneum. Mesothelioma cell lines from rats and mice have been established in vitro, with the majority established in mice as described by Davis [28]. Mesothelial tumors of the Balb/c strain were more likely to be established than tumors of the CBA strain of mouse. All cell lines achieved greater than 32 passages, were in culture for at least 7 months, and exhibited a wide range of morphologies ranging from stellate-shaped cells to fibroblast-like cells. No correlation was found between morphology and doubling time, which ranged from 16 to 30 h. The tumorigenicity of all cell lines was tested by inoculation in syngeneic mice. The kinetics of tumor development varied substantially among cell lines, with the most tumorigenic lines (AB1 and AC29) producing ascites in 27 and 24 days, respectively, and solid tumors 34 and 25 days after subcutaneous (s.c.) inoculation. All cell lines produced solid-tumor growth, at times without concurrent ascite formation. In vivo aggressiveness did not correlate with in vitro morphology or growth rate.

13.2.6 Intrapleural Asbestos Injection Because most human mesothelioma is manifested in the pleural cavity, the intrapleural mode of induction is cited as being more relevant to the human disease than studies done in the peritoneum. However, its applicability to the human disease fibers introduced in this manner still bypasses the body’s natural defenses, and subjects mesothelial cells to doses of fiber that would not be encountered under normal circumstances with the human cases.

13.2.7 Intrapleural Asbestos Injection: Rat Tumor induction rates appear to be somewhat lower in the pleura than in the peritoneum, with chrysotile viewed as a more potent tumor initiator than crocidolite, inducing mesothelioma in 65% of injected animals compared to 45% induction for crocidolite [23]. Whitaker reports successful induction rates in 56% of the animals inoculated with Western Australian crocidolite [26]. Although few analyses have been performed to link the carcinogenicity of a fiber to its length, it is logical to assume that the same relationship holds true as with peritoneal studies. The time it takes to develop tumor in the pleura of rats is relatively long. Whitaker reports a latency of 56 weeks in animals treated with Western Australia crocidolite asbestos. The average survival time is longer in animals afflicted with pleural mesothelioma than in their peritoneal counterparts, with rats injected in the pleura surviving between 105 and 111 weeks [23]. Crocidolite administration resulted in an average survival time of 105 weeks, and animals injected with Canadian chrysotile lived an average of 111 weeks.


313

Malignant cell lines have been developed from rat pleural ascites. Whitaker reports establishing cell cultures in two of five pleural effusions for 2 and 8 months [26]. The presence of collagen fibers using Gieson’s stain. Examination of the cell junctions, profuse microvillous borders, and intermediate filaments further confirmed the mesothelial nature of the culture.

13.2.8 Inhalation of Asbestos in Animal Models Inhalation studies offer the advantage of introducing the carcinogen through the only significant pathway for humans because almost all exposures to asbestos in our species occur through breathing. However, these studies are often expensive, pose a hazard to the researchers conducting them, and have a very low incidence of mesothelioma induction, which makes them impractical for discriminating the fibers’ potential for mesothelioma production.

13.2.8.1 Inhalation Studies in Rats It has been known for some time that inhalation studies are not as efficient in producing mesothelioma as intracavital injections. Miller reported mesothelioma induction in 2 of 42 (5%) of animals, yet a study by Botham in which rats were exposed to high concentrations of Northwest Cape crocidolite failed to induce any mesothelioma in the animal group [25, 29]. In the Miller study, amosite asbestos with a significant proportion of fibers >25 µm was used to induce mesothelioma in 5% of the treated group.

13.2.8.2 Inhalation Studies: Guinea Pig A study by Botham et al. shows that West Cape crocidolite and Transvaal crocidolite are capable of producing mesothelioma in albino Guinea pigs [29].

13.2.8.3 Intratracheal Asbestos Administration Intratracheal instillations are considered to be nonphysiological and unsuitable for risk characterization because of the frequent, uneven distribution of fibers within the different lobes. One common result of this mode of introduction is the formation of areas of greater deposition leading to high local doses and acute inflammatory responses. However, like intracavital injections, this method can be considered for comparative risk assessment among different fiber types.

314

H.I. Pass et al.

13.2.8.4 Intratracheal Asbestos Administration: Syrian Golden Hamster There is some evidence that long fibers (i.e. fibers >8 µm) are not required for the induction of mesothelioma intratracheally. Mohr et al. reported an incidence of mesothelioma of 5.6% (8 of 142) in the Syrian golden hamsters after intratracheal instillations of crocidolite fibers, with 50% 8 µm seldom produce mesothelioma, and that about 600,000 fibers of this length are needed to produce substantial levels of mesothelioma. Ultrastructural and histological characteristics of erionite-induced mesothelioma were similar to those of asbestos fibers. Man-made vitreous fiber 21 (MMVF 21), or stonewool, has similarly been shown to be very mesotheliomatogenic. Miller reports a 95% incidence of mesothelioma in rats treated with MMVF 21 (versus 88% for amosite), and a mean survival time of 284 days (versus 509 days for amosite) [25]. The results correlate well with the hypothesis that long fibers are more toxic than shorter ones – MMVF 21 contains more than seven fold more fibers >10 µm than amosite.

316

H.I. Pass et al.

13.5 Novel Viral-Induced and Transgenic Knockout Models of Mesothelioma An association between mesothelioma and viruses has been previously reported. Malignant mesotheliomas, immunohistochemically and architecturally identical to those seen in humans, have been induced in chickens when a DNA fragment of the oncogene of the sarcoma virus was introduced intraperitoneally [33]. Our group was the first to report that 60% (25 of 48) of human mesotheliomas contain DNA for the amino terminus region of T-antigen, a protein associated with various DNA viruses, including SV40, which is capable of causing malignant transformation in some human cancers, including malignant mesothelioma [34]. Humans were exposed to SV40 through the administration of SV40-contaminated polio vaccines between 1955 and 1963. Today, SV40 continues to be transmitted horizontally and vertically, despite eradication of the contaminated vaccines. SV40 is a DNA-tumor virus that infects monkeys and causes malignant transformation of hamster and murine cells. SV40 produces two transforming proteins: the large T antigen (TAg), which is responsible for binding and inhibiting tumor suppressor genes, e.g. NF2, Rb, p53, CDKN2A/ARF, p107, p130 and transcription factors such as p300 and CBP leading to uncontrolled DNA replication and cellular proliferation. The second transforming protein called small T antigen (tAg) binds and inhibits cellular phosphatase which further contributes to malignant cell transformation (1A). In nonpermissive hosts, SV40 has been shown to be oncogenic [35]. No productive infection and virions result in nonpermissive hosts. SV40 is capable of transforming a number of different mammalian cells in vitro. Murine cells transformed by SV40 infection in vitro are capable of producing lethal tumors in vivo when transplanted back into the syngeneic host. Thus, SV40 murine transformed cells can be oncogenic in syngeneic hosts, and the tumors induced in vivo express SV40 tumor-specific transplantation antigens. These transplantation antigens include SV40 TAg and tAg, both of which are derived from a single early gene-product transcript [36].

13.5.1 SV40 Viral Hamster Models Newborn hamsters are also extremely susceptible to SV40-induced tumors. In 1962, Gerber and Kirschstein reported that SV40 induced ependymomas in hamsters. Since that discovery, wild-type SV40 has been known to be highly oncogenic in hamsters [37]. Newborn animals are particularly susceptible, and will develop fibrosarcomas at the injection site following subcutaneous inoculation of a low dose of SV40 [38]. When newborn hamsters are inoculated with intracerebral SV40, they develop ependymomas [37]. Weanling and adult animals may develop fibrosarcomas if injected subcutaneously with a high dose of virus [>109 plaque-


317

forming units (pfu)], but only with a low tumor incidence and after prolonged incubation periods [39]. When SV40 (>108.5 pfu) is injected intravenously into weanling hamsters, subjecting many cell types to high concentrations of virus, lymphocytic leukemia, lymphoma, and osteosarcoma will develop at sites distant from the injection [40]. Carcinomas – the most common tumors in humans – never develop following SV40 injection, suggesting that epithelial cells may be resistant to SV40 transformation. These data indicate that only specific types of cells are susceptible to SV40 transformation, mesothelial cells, osteoblasts/osteocytes, macrophages, and B-lymphocytes, and that the specific routes of SV40 inoculation used apparently play a key role in the induction of specific hamster tumors. It is actually fortuitous that a relationship between SV40 and mesothelioma was discovered in the hamster. Lipotich reported the use of an SV40-induced hamster mesothelioma-cell line (800TU) [41]. Before this study, mesotheliomas were not observed following s.c., intracerebral, and intravenous (i.v.) inoculation of SV40 in the hamster. Indeed, the development of the 800 TU line was an inoculation accident, for newborn hamsters in these experiments were injected between the scapulae with SV40, and all of the other animals in the experiment developed in situ fibrosarcomas. The development of the mesothelioma was caused by accidental pleural injection of the SV40 (R. C. Moyer, personal communication). Stimulated by these isolated yet intriguing pieces of data, Carbone investigated the oncogenicity of wild-type SV40 and SV40 small t-deletion mutants when injected into the hearts, pleura, or peritoneum of 21-day-old hamsters [42]. Mesotheliomas that could be characterized as macroscopically, microscopically, ultramicroscopically, and histochemically identical to those seen in humans, developed within a 3-month period in 30 of the 43 hamsters injected with wild-type SV40. All (n = 34) the hamsters injected with the small t-mutant SV40 developed true histiocytic or B-cell lymphomas, yet only 1 developed a mesothelioma. The decreased oncogenicity in the deletion mutant group was puzzling, because small T-antigen binds and inhibits the activity of the cellular phosphatase 2A that will subsequently prevent dephosphorylation of large T-antigen and the p53 protein product [43, 44]. It is theorized, therefore, that in addition to physical binding between large T-antigen and the products of p53 and retinoblastoma (Rb), alteration of the Rb and p53 phosphorylation state by small T-antigen may be required to completely inactivate their function, and thus allow the cell to progress to S phase during which transformation could occur. The SV40-induced hamster mesotheliomas spread along the pleural, pericardial, and peritoneal surfaces obliterating the cavities and infiltrating the diaphragm and the chest wall in the absence of distant metastases. Histologically, epithelial, spindlecell, and more often mixed-type mesotheliomas are seen. Ultramicro-scopically, the tumors and derived cell lines showed long, branching microvilli without core filaments, basal lamina, intracellular lumens, perinuclear tonofilaments, intercellular junctions, the absence of secretory granules, and limited cytoplasmic organelles, especially rough endoplasmic reticulum. These mesotheliomas are associated with

318

H.I. Pass et al.

hyaluronic acid, contain cytokeratins, and are immunohistochemically similar to human mesotheliomas. Southern blot hybridization of the DNAs extracted from the tumors reveals SV40 DNA sequences, and the cell lines derived from these tumors contain and express the early region of SV40 DNA. Immunohistochemical staining of the cell lines and tumors reveals the presence of intranuclear T-antigen.

13.5.2 SV40 Transgenic Mouse Models SV40 TAg transgenic mouse models have been developed to study the SV40 viral oncogenesis and co-carcinogenesis with asbestos exposure in the context of the development of malignant mesothelioma. Robinson et al. developed four lines of MexTAg mice that express SV40 TAg selectively in the mesothelial cells of the pleural, pericardial, and peritoneal cavities [45]. The four lines carry different numbers of the viral genome: high (100 copies), intermediate (32, 15 copies) and single copy. All high transgene copy mice showed some spontaneous tumor development (5%). Mesothelial cells from the mutant mice, cultivated in vitro, demonstrated immortality, with logarithmic growth for >100 doublings, as opposed to logarithmic growth in just six to eight doublings for normal mesothelial cells. These cells, isolated from high copy MexTAg mice, were able to form anchorage-independent colonies and grow in low-serum conditions, typical characteristics of a transformed cell line. After intraperitoneal injection of asbestos fibers, high-copy mice developed mesotheliomas that extremely aggressive, with a median survival rate of 35 weeks. This compares to 63 weeks for the wild-type and 55 weeks for the single copy line. The study demonstrated a direct relationship between the number of SV40 copies in the genome and the survival after exposure to asbestos. Additionally, increasing the dose of asbestos also accelerated the development of mesothelioma in high copy mice. The low level of spontaneous tumors suggests a two stage model, in which asbestos causes irritation, leading to mesothelial replication. The mesothelial cells that express TAg are more likely to become transformed in the proliferation process. These SV40 Tag transgenic mice, demonstrating a model of mesothelioma akin to the human disease, are good for studying early molecular changes during malignant transformation and can be used to test novel therapies.

13.5.3 Transgenic Murine Models, Utilizing Nf2, Ink4a/ARF, and P53 Knockout Mice Somatic gene changes involving inactivation of genes such as neurofibromatosis type 2 (Nf2), deletions of CDKN2A/Arf gene loci, and mutations of p53 gene are well described in human mesothelioma. Experimentation with the inactivation of


319

Nf2, along with various other tumor suppressor genes, was investigated in the hope of finding a murine parallel to the human disease. Altomare et al. reported that Nf2 (+/–) knockout mice exposed to asbestos develop malignant mesotheliomas at a much higher rate than wild-type mice [46]. Similar to human malignant mesothelioma, tumors from Nf2 (+/–) mice demonstrated homozygous deletions of tumor suppressor genes p16 (Ink4A), p14 (Arf)/p19 (Arf), p15 (Ink4B) and inactivation of p53/Arf. This suggests that these mutations permit malignant mesothelial cells to skip crucial cell-cycle checkpoints, proliferating without control. Additionally, the Nf2 (+/–) mice also displayed lower survival times because of the more aggressive nature of their malignant mesotheliomas, as compared to the wild-type mice. Recent development of conditional Nf2 combo knockout (CKO) mice further advanced transgenic mouse models of malignant mesothelioma. Jongsma et al. developed conditional mouse models by introducing mesothelial-specific loss of Nf2, Ink4A/Arf, and p53 transgenes by an intrathoracic injection of a viral AdenoCre recombinase specific to those genes [47]. This group also created compound homozygous and heterozygous Nf2/Ink4a/Arf; Nf2/p53 and Nf2/Ink4a/P53 knockout littermates. Without the introduction of any other carcinogenic elements, such as asbestos, thoracic tumors developed in 80–100% of homozygous combo knockout mice. The murine model does not demonstrate the same tendency to develop epitheloid tumors as the human model. Malignant mesothelioma was most common in Nf2;p53 or Nf2;InK4a/Arf CKO mice. The Nf2; p53 and Nf2;Ink4a/Arf had some epitheloid but none were seen in the Nf2;p53,Ink4a mice. Sarcomatoid and mixed were seen in all three groups. Nf2;Ink4a/Arf CKO tumors demonstrated to be the most aggressive genotype. These tumors are contrasted to the homozygous Nf2;p53 CKO tumors, which were not extremely aggressive. The evidence supports the idea that the loss of Ink4a is the factor that significantly reduces the latency period and contributes to more invasive behavior.

13.6 Orthotopic Transplants and Xenograft Xenografts offer the advantage of producing mesothelial tumors in the animal model at a faster rate and with higher success than asbestos animal models. Furthermore, the resulting tumors are of a similar histological type as their human counterparts and retain their functional and morphological features during several generations, and can therefore provide accurate information on the chemosensitivity of the human tumor. Mice have been successfully used to replicate human MPM. Chahinian reports the original transplant of human mesothelial tumor from two patients into nude mice of the Balb/c strain [48]. Intraperitoneal (i.p.) transplants did not grow, but s.c. xenografts/implants were able to produce tumor in an average of 46 days in the animal. The tumor transplants of the first generation grew in 6 of 20 mice (30%), with a take-rate of implants of 53%. Overall, tumors grew in 52 of 80 mice (65%)

320

H.I. Pass et al.

in a total of 169 of 266 implants (multiple implants were made on the same animal). Tumor examination of the first- and second-generation xenografts confirmed their histological similarity to the original epithelial tumors. Colt et al. commented on the implantation of intact human mesothelial tissue in the pleural space of four athymic nude mice and subcutaneously in another mouse [49]. The s.c. implantation resulted in a progressive tumor growth, with the tumor reaching a size of 10 × 12 mm. Of the four i.p. mice, one died early, another at 162 days, and the remaining two mice were sacrificed at 180 days after implantation. Both of the sacrificed animals demonstrated tumor growth at the implantation site as well as at the visceral, diaphragmatic, and mediastinal pleural surfaces. No metastases to distant organs were noted a feature similar to the natural history of human malignant mesothelioma. In the human tumor, sections of pleural biopsies showed malignant neoplasms composed of epitheloid cells, which on immunohistochemical study demonstrated diffuse and strong positive stainings for cytokeratin, vimentin, and epithelial-membrane antigen, but tested negative for carcinoembryonic antigen and Leu M1. As a result, the neoplasm was classified as monophasic epithelial-type mesothelioma. In the mouse, the immunohistochemical profile characterized by positive stainings for vimentin and cytokeratin, and negative CEA and Leu M1 stainings, strongly pointed to the similarity between the human and animal tumor. Rats have also been a species of choice for mesothelioma-cell transplants. Linden successfully inoculated athymic Rowett rats s.c. with a coarse cell suspension of mesothelioma cells from a human patient [50]. The take-rate was 93% (13 of 14) in the initial passages (P) and 100% (192/192) in P3-P9. There was a decrease in the tumor-volume doubling (TD) time during the serial passage of rats from 6 days in P2 to 3 days in P9, with no further reduction noted in later passages. The average latency, measured by the time needed to reach a specified tumor volume, was found to decrease sharply from P2 to P12, but not thereafter. It took 36 days in P2 to reach a volume of 2 cm3, and only 11–12 days in P10-P12. Morphological examination of the tumors revealed mesothelioma of an epithelial type. The histological pattern of the original tumor was retained in all xenograft generations of rats, with no differentiation noted. Prewitt et al. describe the orthotopic implantation of the tumor-cell line H-Meso 1 in pneumonectomized Fischer nu/nu rats [51]. Tumor reproducibly filled the chest cavity 6 weeks after implantation with 106 tumor cells, and were identical in histologic pattern to epithelioid mesothelioma. There have been reports of transplants of mesothelioma cells from rat to rat [52]. After the successful induction of mesothelioma in F344 rats with crocidolite fibers, the cell lines were cultured in vitro in RPMI-1640 and inoculated intrapleurally in F344 rats. The mesothelial origin of the cells was confirmed by the co-expression of keratin and vimentin, using an alkaline and anti-alkaline phosphatase. Polyclonal rabbit antibodies directed against human 56-kDa cytokeratin and monoclonal mouse anti-swine vimentin 57-kDa were also used. There was a clear-cut dose–response relationship when several concentrations of cells were inoculated, with the largest dose of mesothelioma cells administered


321

(5 × 106) able to induce the most tumors, as determined by chest radiographs performed on 15 and 30 days postinoculation. No rats showed abnormalities at 15 days postinoculation but by 30 days, most animals showed features suggesting massive tumor volumes. Most tumors predominantly invaded the mediastinum and pericardium, with diaphragm involvement, but no metastasis was noted. Transplants of mesothelioma cells from hamster to hamster have been performed [53]. In a study designed to test the usefulness of various chemotherapeutic agents against mesothelioma, Smith et al. report successfully transplanting mesothelioma induced by intrapleural injection of tremolite asbestos in one golden Lak:LVG Syrian hamster to others [53]. Ascites were noted 76 days after transplant in the first generation of three hamsters that had received intraperitoneal injections of peritoneal effusions from the original animal, with two of the animals sacrificed on day 76 and the third dying on day 90. The tumor was carried through 39 serial transplant generations by i.p. injections. The average survival time was found to decrease with continuous passage, leading to the death of new hosts within 21–38 days, and an average survival time of 28–32 days depending on the generation examined. The transplantable tumor line was defined as mesothelioma 10–24. Tumors of animals bearing transplants continued to resemble the epithelial nature of the mesothelioma in the original animal.

13.7 Conclusions A number of animal models for the investigation of mesothelioma are now available. These models have been used in a number of preclinical models for the treatment of mesothelioma, including gene therapy with suicide genes [54–57], antisense gene therapy [58], re-expression gene therapy [59], photodynamic therapy [60, 61], immunotherapy [62–67], and in vivo chemosensitivity [68–71]. Many cell lines are available, and the models are reproducible.

References 1. Antman KH, Pass HI, Schiff PB. Benign and malignant mesothelioma. In: De Vita VT Jr, Hellman S, Rosenberg SA, editors. Cancer: principles and practice of oncology. 5th ed. Philadelphia, PA: Lippincott-Raven; 1997. p. 1853–78. 2. Kaiser LR. New therapies in the treatment of malignant pleural mesothelioma. Semin Thorac Cardiovasc Surg. 1997;9:383–90. 3. Mark EJ, Yokoi T. Absence of evidence for a significant background incidence of diffuse malignant mesothelioma apart from asbestos exposure. In: Landrigen J, Kazemi H, editors. The third wave of asbestos disease: exposure to asbestos in place. NY: NY Academy of Sciences Press; 1960. p. 196–204. 4. Wagner JC, Sleggs CA, Marchand P. Diffuse pleural mesothelioma and asbestos exposure in the North Western Cape Province. Br J Ind Med. 1960;17:260–71.

322

H.I. Pass et al.

5. Selikoff IJ, Hammond EC, Seidman H. Latency of asbestos disease among insulation workers in the United States and Canada. Cancer 1980;46:2736–40. 6. Weitzman SA, Graceffa P. Asbestos catalyzes hydroxyl and superoxide radical generation from hydrogen peroxide. Arch Biochem Biophys. 1984;228:373–6. 7. Jaurand MC, Fleury J, Monchaux G, Nebut M, Bignon J. Pleural carcinogenic potency of mineral fibers (asbestos attapulgite) and their toxicity on cultured cells. J Natl Cancer Inst. 1987;79:797–804. 8. Weissman LB, Antman KH. Incidence, presentation and promising new treatments for malignant mesothelioma. Oncology 1989;3:67–72. 9. Mossman BT, Bignon J, Corn M, Seaton A, Gee JBL. Asbestos: scientific developments and implications for public policy. Science 1990;247:294–301. 10. Roggli VL, Santilippo F, Shelburne J. Mesothelioma. In: Roggli VL, Greenberg SD, Pratt PC, editors. Pathology of asbestos and associated diseases. Boston, MA: Little Brown Co.; 1995. p. 383–91. 11. Mossman B, Light W, Wei E. Asbestos: mechanisms of toxicity and carcinogenicity in the respiratory tract. Annu Rev Pharmacol Toxicol. 1983;23:595–615. 12. Stanton MF, Layard M, Tegeris A, Miller E, May M, Morgan E, et al. Relation of particle dimension to carcinogenicity in amphibole asbestoses and other fibrous minerals. J Natl Cancer Inst. 1981;67:965–75. 13. Leveston SA, McKeel DW Jr, Buckley PG, et al. Acromegaly and Cushing’s syndrome associated with a foregut carcinoid. J Clin Endocrinol Met. 1981;53:682–9. 14. Huncharek MD. Changing risk groups for malignant mesothelioma. Cancer 1992;69:2704–11. 15. Leigh J, Ferguson DA, Ackad M, Thompson R. Lung asbestos fiber content and mesothelioma cell type, site, and survival. Cancer 1991;68:135–41. 16. Rogers AJ, Leigh J, Berry G, Ferguson DA, Mulder HB, Ackad M. Relationship between lung asbestos fiber type and concentration and relative risk of mesothelioma. Cancer 1991;67:1912–20. 17. Murai Y, Kitagawa M. Asbestos fiber analysis in 27 malignant mesothelioma cases. Am J Ind Med. 1992;22:193–207. 18. Mowé G, Gylseth B, Hartveit F, Skaug V. Fiber concentration in lung tissue of patients with malignant mesothelioma. Cancer 1985;56:1089–93. 19. Mohr U, Potts F, Vonnahme F-J. Morphological aspects of mesotheliomas after intratracheal instillations of fibrous dusts in Syrian golden hamsters. Exp Pathol. 1984;26:179–83. 20. Suzuki Y, Kohyama M. Malignant mesothelioma induced by asbestos and zeolite in the mouse peritoneal cavity. Environ Res. 1984;35:277–92. 21. Wagner JC, Griffiths DM, Hill RJ. The effect of fibre size on the in vivo activity of UICC crocidolite. Br J Cancer. 1984;49:453–8. 22. Hill RJ, Edwards RE, Carthew P. Early changes in the pleural mesothelium following intrapleural inoculation of the mineral fibre erionite and the subsequent development of mesotheliomas. J Exp Pathol. 1990;71:105–18. 23. Minardi F, Maltoni C. Results of recent experimental research on the carcinogenicity of natural and modified asbestos. Ann Acad Sci. 1988;534:754–61. 24. Davis J, Bolton R, Miller B, Niven K. Mesothelioma dose response following intraperitoneal injection of mineral fibres. Int J Exp Pathol. 1991;72:263–74. 25. Miller G, Searl A, Davis J, Kenneth D. Influence of fibre length, dissolution and biopersistence on the production of mesothelioma in the rat peritoneal cavity. Ann Occup Hyg. 1999;43:155–66. 26. Whitaker D, Shilkin KB, Walters MN. Cytologic and tissue culture characteristics of asbestosinduced mesothelioma in rats. Acta Cytol. 1984;28:185–9. 27. Wagner JC, Griffiths DM, Hill RJ. The effect of fibre size on the in vivo activity of UICC crocidolite. Br J Cancer. 1984;48:453–8. 28. Davis MR, Manning LS, Whitaker D, Garlepp MJ, Robinson BW. Establishment of a murine model of malignant mesothelioma. Int J Cancer. 1992;52:881–6.


323

29. Botham S, Holt P. The effects of inhaled crocidolites from Transvaal and North-west Cape mines on the lungs of rats and guinea-pigs. Br J Exp Pathol. 1972;53:612–20. 30. Tanigawa H, Onodera H, Maekawa A. Spontaneous mesotheliomas in Fischer rats – a histological and electron microscopic study. Toxicol Pathol. 1987;15:157–63. 31. Rice JM, Kovatch RM, Anderson LM. Intraperitoneal mesotheliomas induced in mice by a polycyclic aromatic hydrocarbon. J Toxicol Environ Health. 1989;27:153–60. 32. Berman J, Rice J. Mesotheliomas and proliferative lesions of the testicular mesothelium produced in Fischer, Sprague-Dawley and Buffalo rats by methyl(acetoxymethyl)nitrosamine (DMN-OAc). Vet Pathol. 1979;16:574–82. 33. England JM, Panella MJ, Ewert DL, Halpern MS. Induction of a diffuse mesothelioma in chickens by intraperitoneal inoculation of v-src DNA. Virology 1991;182:423–9. 34. Carbone M, Pass HI, Rizzo P, Marinetti M, Di Muzio M, Mew DJY, et al. Simian virus 40-like DNA sequences in human pleural mesothelioma. Oncogene 1994;9:1781–90. 35. Eddy BE. Simian virus 40: an oncogenic virus. Prog Exp Tumor Res. 1964;4:1–26. 36. Tevethia SS. Immunology of simian virus 40. In: Klein G, editor. Viral oncology. New York, NY: Raven Press; 1980. p. 581–601. 37. Gerber P, Kirschstein RL. SV40-induced ependymomas in newborn hamsters. Virology 1962;18:582–8. 38. Lewis AM Jr, Martin RG. Oncogenicity of simian virus 40 deletion mutants that induce altered 17kilodalton t-proteins. Proc Natl Acad Sci USA. 1979;76:4299–302. 39. Allison AC, Chesterman FC, Baron S. Induction of tumors in adult hamsters with simian virus 40. J Natl Cancer Inst. 1967;38:567–77. 40. Diamandopoulous GT. Leukemia, lymphoma and osteosarcoma induced in the Syrian Golden hamster by simian virus 40. Science 1972;176:73–5. 41. Lipotich G, Moyer MP, Moyer RC. Rescue of SV40 following transfection of TC7 cells with cellular DNAs containing complete and partial SV40 genomes. Mol Gen Genet. 1982;186:78–81. 42. Cicala C, Pompetti F, Carbone M. SV40 induces mesotheliomas in hamsters. Am J Pathol. 1993;142:1524–33. 43. Scheidtmann KH, Mumby MC, Rundell K, Walter G. Dephosphorylation of simian virus 40 large-T antigen and p53 protein by protein phosphatase 2A: inhibition by small-t antigen. Mol Cell Biol. 1991;11:1996–2003. 44. Carbone M, Hauser J, Carty MC, Rundell K, Dixon K, Levine AS. Simian virus 40 small t antigen inhibits SV40 DNA replication in vitro. J Virol. 1992;66:1804–08. 45, Robinson C, van Bruggen I, Segal A, et al. A novel SV40 tag transgenic model of asbestosinduced mesothelioma: malignant transformation is dose dependent. Cancer Res. 2006;66:10786–94. 46. Altomare DA, Vaslet CA, Skele KL, et al. A mouse model recapitulating molecular features of human mesothelioma. Cancer Res. 2005;65:8090–5. 47. Jongsma J, van Montfort E, Vooijs M, et al. A conditional mouse model for malignant mesothelioma. Cancer Cell. 2008;13:261–71. 48. Chahinian AP, Beranek JT, Suzuki Y, Bekesi JG, Wisniewski L, Selikoff IJ, et al. Transplantation of human malignant mesothelioma into nude mice. Cancer Res. 1980;40:181–5. 49. Colt H, Astoui P, Wang Xiaoen, Yi E, Boutin C. Clinical course of human epithelial-type malignant pleural mesothelioma replicated in an orthotopic-transplant nude mouse model. Anticancer Res. 1996;16:633–9. 50. Linden C.-J. Progressive growth of a human pleural mesothelioma xenografted to athymic rats and mice. Br J Cancer. 1988;58:614–8. 51. Prewitt TW, Lubensky IA, Pogrebniak HW, Pass HI. Orthotopic implantation of mesothelioma in the pneumonectomized immune-deficient rat: a model for innovative therapies [letter]. Int J Cancer. 1993;55:877–80. 52. Pimpec-Barthes L, Bernard I, Alasmad ABD. Pleuro-pulmonary tumours detected by clinical and chest X-ray analyses in rats transplanted with mesothelioma cells. Br J Cancer. 1999;81:1344–50.

324

H.I. Pass et al.

53. Smith W, Huhert DD, Holiat SM, Sobel HJ, Davis S. An experimental model for treatment of mesothelioma. Cancer 1981;47:658–63. 54. Smythe WR, Kaiser LR, Hwang HC, Amin KM, Pilewski JM, Eck B, et al. Successful adenovirus-mediated gene transfer in an in vivo model of human malignant mesothelioma. Ann Thorac Surg. 1994;57:1395–401. 55. Smythe WR, Hwang HC, Elshami AA, Amin KM, Eck SL, Davidson BL, et al. Treatment of experimental human mesothelioma using adenovirus transfer of the herpes simplex thymidine kinase gene. Ann Surg. 1995;222:78–86. 56. Hwang HC, Smythe WR, Elshami AA, Kucharczuk JC, Amin KM, Williams JP, et al. Gene therapy using adenovirus carrying the herpes simplex-thymidine kinase gene to treat in vivo models of human malignant mesothelioma and lung cancer. Am J Respir Cell Mol Biol. 1995;13:7–16. 57. Kucharczuk JC, Elshami AA, Zhang HB, Smythe WR, Hwang HC, Tomlinson JS, et al. Pleural-based mesothelioma in immune competent rats: a model to study adenoviral gene transfer. Ann Thorac Surg. 1995;60:593–7. 58. Pass HI, Mew DJ, Carbone M, Matthews WA, Donington JS, Baserga R, et al. Inhibition of hamster mesothelioma tumorigenesis by an antisense expression plasmid to the insulin-like growth factor-1 receptor. Cancer Res. 1996;56:4044–8. 59. Frizelle SP, Grim J, Zhou J, Gupta P, Curiel DT, Geradts J, et al. Re-expression of p16INK4a in mesothelioma cells results in cell cycle arrest, cell death, tumor suppression and tumor regression. Oncogene 1998;16:3087–95. 60. Feins RH, Hilf R, Ross H, Gibson SL. Photodynamic therapy for human malignant mesothelioma in the nude mouse. J Surg Res. 1990;49:311–4. 61. Foster TH, Gibson SL, Raubertas RF. Response of Photofrin-sensitised mesothelioma xenografts to photodynamic therapy with 514 nm light. Br J Cancer. 1996;73:933–6. 62. Christmas TI, Manning LS, Garlepp MJ, Musk AW, Robinson BW. Effect of interferon-alpha 2a on malignant mesothelioma. J Interferon Res. 1993;13:9–12. 63. Leong CC, Marley JV, Loh S, Milech N, Robinson BW, Garlepp MJ. Transfection of the gene for B71 but not B72 can induce immunity to murine malignant mesothelioma. Int J Cancer. 1997;71:476–82. 64. Caminschi I, Venetsanakos E, Leong CC, Garlepp MJ, Scott B, Robinson BW. Interleukin-12 induces an effective antitumor response in malignant mesothelioma. Am J Respir Cell Mol Biol. 1998;19:738–46. 65. Astoul P, Boutin C. An experimental model of pleural cancer. Value of orthotopic implantation of human tumor tissue in nude mice. Rev Mal Respir. 1997;14:355–62. 66. Bielefeldt-Ohmann H, Fitzpatrick DR, Marzo AL, Jarnicki AG, Musk AW, Robinson BW. Potential for interferon-alpha-based therapy in mesothelioma: assessment in a murine model. J Interferon Cytokine Res. 1995;15:213–23. 67. Ohnuma T, Szrajer L, Holland JF, Kurimoto M, Minowada J. Effects of natural interferon alpha, natural tumor necrosis factor alpha and their combination on human mesothelioma xenografts in nude mice. Cancer Immunol Immunother. 1993;36:31–6. 68. Chahinian AP, Norton L, Holland JF, Szrajer L, Hart RD. Experimental and clinical activity of mitomycin C and cis-diamminedichloroplatinum in malignant mesothelioma. Cancer Res. 1984;44:1688–92. 69. Sklarin NT, Chahinian AP, Feuer EJ, Lahman LA, Szrajer L, Holland JF. Augmentation of activity of cis-diamine-dichloroplatinum (II) and mitomycin-C by interferon in human malignant mesothelioma xenografts in nude mice. Cancer Res. 1988;48:64–7. 70. Roboz J, Chahinian AP, Holland JF, Silides D, Azrajer L. Early diagnosis and monitoring of transplanted human malignant mesothelioma by serum hyaluronic acid. J Natl Cancer Inst. 1989;81:924–8. 71. Chahinian AP, Mandeli JP, Gluck H, Naim H, Teirstein AS, Holland JF. Effectiveness of cisplatin, paclitaxel, and suramin against human malignant mesothelioma xenografts in athymic nude mice. J Surg Oncol. 1998;67:104–11.

Chapter 14

The Use of Mouse Models to Study Leukemia/Lymphoma and Assess Therapeutic Approaches William Siders

Abstract The fields of leukemia and lymphoma research have significantly advanced over the recent years as a result of the establishment of cell lines that reflect several aspects of the human disease. These cell lines have been employed in all phases of preclinical research from the immunization of mice to generate new therapeutic antibodies to proof of concept and target validation experiments. In addition, several transgenic mouse or genetically engineered mouse models have been developed that recapitulate many aspects of both leukemia and lymphoma. These models are particularly well suited to the exploration of interactions between tumor and stromal cells and the progression of cancer as it relates to its microenvironment. Therapeutic antibodies including ofatumumab and epratuzumab are currently undergoing clinical trial evaluation based on their activity in models such as these. Xenograft tumor models have been especially instrumental in studies addressing mechanism of action and in evaluating combination therapies. This chapter will primarily explore the use of human cells in xenograft tumor systems as models for evaluating therapeutic approaches. Keywords Leukemia • Lymphoma • Mouse • Xenograft • Therapeutic

14.1 Introduction The ability to perform in vivo studies using mouse models has greatly enhanced our understanding of the development and progression of both leukemia and lymphoma disorders. Currently the mouse remains the best system for exploring several aspects of lymphomagenesis and therapeutic intervention. Several transgenic W. Siders (*) Cancer and Immunotherapy Research Group, Genzyme Corporation, 49 New York Avenue, Framington, MA 01701, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_14, © Springer Science+Business Media, LLC 2011

325

326

W. Siders

(Tg) mouse or genetically engineered mouse models (GEMMs) have been developed that recapitulate many aspects of both leukemia and lymphoma (for review see Refs. [1–5]). These systems often have the advantage of expressing oncogenes or dominant-negative tumor suppressor genes under the control of cell-specific promoters that restrict expression to the tissue of interest. These models are particularly well suited for exploring interactions between tumor and stromal cells and the progression of cancer as it relates to its microenvironment. As genetic manipulation of the mouse genome has become more common, sophisticated techniques have been developed that have allowed the field of cancer biology to advance as a result of being able to control gene expression not only in a cell-type-dependent manner but in a time-dependent manner as well. For instance, using the inducible tetracycline system [6], Huettner et al. [7] have generated a mouse model that conditionally expresses the BCR-ABL fusion protein resulting in the development of leukemia in 100% of the mice. Earlier attempts at creating BCR–ABL Tg mice resulted in embryonic lethality. In this system, expression of BCR–ABL can be induced at any stage of development in the mouse by simply withholding tetracycline administration. However, it is important to understand that these mouse systems still have limitations and that differences do exist in the development of cancer in mice relative to the human disease [8]. In addition to GEMMs, the study of hematologic malignancies and the development of targeted therapies have also been enhanced by the use of human tumor cells in immunodeficient mice as well as the ability to generate cell lines from patient samples. The creation of mouse models that are deficient in T cells (nude), T cells and B cells (SCID), NK cells (beige) or a combination of all three (SCID-beige) has significantly aided these efforts. Where it was once challenging to inject tumor cells from patients into mice and achieve successful engraftment, it is now common practice to develop novel models that in some instances reflect several aspects of the human disease. The extent of engraftment, however, may be dependent not only on the strain of mouse chosen, but also on whether additional immunosuppression such as radiation is required [9]. Cell lines from tumor samples isolated from patients that have been adapted to grow in vitro have been used extensively for several applications including (1) as immunogens in immunocompetent animals to create therapeutic antibodies; (2) to establish proof of concept for therapeutics targeting specific pathways or antigens in a particular tumor indication prior to the initiation of clinical trials; and (3) to understand the mechanisms underlying tumor cell migration and metastasis formation. This chapter will primarily explore the use of human cells in xenograft tumor systems as models for evaluating therapeutic approaches. Other approaches such as Tg mice that have been used for therapeutic evaluation or understanding the mechanism of action for a particular therapy have also been included. A table of several successfully xenografted human cell lines accompanies each of the primary hematologic indications discussed and serves as a reference point for investigators wishing to develop mouse models for therapeutic analysis.

14 The Use of Mouse Models to Study Leukemia/Lymphoma

327

14.1.1 Models of Myeloid Leukemia Myeloid leukemia is characterized by the proliferation of myeloid cells such as neutrophils, basophils or monocytes and may present either as an acute or chronic condition. Recent data from the American Cancer Society estimate that approximately 13,000 new cases of acute and 5,000 new yearly cases of chronic myeloid leukemia will be reported in the US. Acute myeloid leukemia (AML) is associated initially with the growth of myeloid progenitors in the bone marrow leading to the appearance of large numbers of immature white blood cells or blasts within the blood. A subtype of AML has also been identified, acute promyelocytic leukemia (APL), that arises from immature cells known as promyelocytes. In contrast, chronic myeloid leukemia (CML) is characterized by the development of progressively larger numbers of normal appearing white blood cells over time in both the bone marrow and the blood. Although genetic abnormalities have been described that result in the development of myeloid leukemia, most cases occur as a result of chromosomal translocations [10]. For example, the initiation event for APL is the t(15,17) translocation that fuses part of the promyelocytic leukemia (PML) gene in frame with part of the retinoic acid receptor a gene resulting in the PML–RARa fusion protein [11]. Likewise, CML development in a large number of patients is driven by the t(9,22) chromosomal translocation forming the “Philadelphia chromosome” and resulting in the generation of the BCR–ABL fusion gene. This fusion gene encodes the BCR–ABL protein with constitutive tyrosine kinase activity resulting in the activation of multiple signaling cascades leading to cell proliferation, resistance to apoptosis and leukemiagenesis [12]. Therapies that target this activity such as the tyrosine kinase inhibitor (TKI) imatinib have resulted in a significant response rate in patients [13]. Several attempts have been made to generate mouse models that recapitulate the onset and progression of myeloid leukemia to aid in studying the process of leukemiagenesis as well as identifying new cancer-specific targets. Initial efforts at engrafting primary human tumor samples from AML patients into SCID mice resulted in variable levels of engraftment and in most instances did not result in the development of reproducible disseminated disease even when the mice were treated with irradiation and immunosuppressive drugs [14, 15]. Although circulating AML cells could be detected in mice for short periods of time following injection, they quickly became undetectable [15]. The use of cytokines such as GM-CSF was shown to enhance engraftment [16–18] but several investigators were eventually successful in the absence of such help including Chelstrom et al. who demonstrated engraftment of primary tumor isolates from pediatric AML patients in sublethally irradiated SCID mice [19]. Given the mixed results with primary tumor samples, the development of myeloid leukemia cell lines that maintained a disease phenotype when injected into mice became paramount to advance the identification of new therapeutic strategies. Beran et al. [20] describe the development of one such line (KBM-5) from a CML patient that contains multiple copies of the Philadelphia chromosome. These cells

328

W. Siders

Table 14.1 Myeloid leukemia xenograft cell lines Cell line Dose Route Mouse strain Referencea KBM-3 2e7 Iv SCID [43] KBM-5 1e7 sc, iv, ip SCID [20] HL-60 ³5e6 Sc NOD/ SCID [44] K-562b 2e7 Sc SCID [29] MEG-01 5e7 Sc Nude [45] MV-4–11c 5e6 Sc Nude [42] Ba/F3 1e6 Iv SCID [28] a Reference cited is not always provided as the originator of the cell line and is included in most instances to provide an example of the tumor growth curve b Originally considered myeloid leukemia line but now recognized as an erythroid leukemia cell line c Biphenotypic cell line

were shown to be resistant to NK cell mediated lysis and to engraft tissues in the SCID mouse in a pattern similar to that observed in the human disease. Several cell lines currently exist that have been used initially in vitro and eventually in vivo to screen for the activity of therapeutic candidates including the murine Ba/F3 cell line and the human HL-60 and K-562 cell lines (Table 14.1). For CML treatment, the most successful therapeutic is the tyrosine kinase inhibitor imatinib which targets the BCR–ABL protein (for review see Ref. [21]). Imatinib (also known as Gleevec or STI571) was initially identified from a library of compounds screened as inhibitors of both PDGF and v-Abl activity [22]. Ultimately, potent activity against the BCR–ABL protein was also demonstrated in several cell lines both in vitro and in vivo and supported exploring imatinib efficacy in clinical trials of CML [23, 24]. Following accelerated approval in 2001, imatinib has proven to be successful in the treatment of CML [13]. However, imatinib- resistant BCR–ABL mutations have been identified [25] suggesting the need for secondgeneration therapies. To this end, dasatinib (BMS-354825) was developed and has been shown to have not only a significantly enhanced affinity for BCR–ABL over imatinib [26, 27] but also activity against most imatinib-resistant BCR–ABL mutants [28]. Using the murine pro B cell line Ba/F3, Shah and co-workers generated stable cell lines expressing all of the known imatinib-resistant BCR–ABL mutations and demonstrated that all but one of these mutants (T315I) were susceptible to treatment with dasatinib. To explore the in vivo efficacy of dasatinib, SCID mice were injected with the Ba/F3 cells expressing either wild-type BCR–ABL or the imatinib resistant forms along with the firefly luciferase gene. In vivo imaging revealed that treatment with dasatinib resulted in a decrease in bioluminescence that translated into a significant increase in survival indicating inhibition of tumor growth. Similar to the in vitro data, the T315I mutant was not responsive to dasatinib treatment demonstrating good correlation between in vitro assays and in vivo SCID mouse models. Pharmacokinetic studies in K-562 bearing SCID mice were also conducted and accurately predicted the clinical exposure required to inhibit BCR–ABL activity in CML patients [29]. Clinical studies with dasatinib have been conducted in both


329

BCR–ABL wild-type and imatinib-resistant CML patients leading to its approval for the treatment of CML patients who are resistant to imatinib therapy as well as Philadelphia chromosome positive acute lymphocytic leukemia (ALL). Although xenografts have proven successful as a predictive tool for some therapeutic approaches, more complex systems have also been examined including humanized-SCID mouse models such as the one developed by Kyoizumi et al. [30]. In this system, human fetal bone fragments are implanted into SCID mice to recapitulate the human hematopoietic microenvironment followed by injection of primary myeloid leukemia cells [31]. It was hypothesized that the human bone marrow may provide growth factors and cellular interactions required to support the engraftment and growth of the leukemic cells. Indeed, successful engraftment was observed in approximately 90% of the samples with growth occurring primarily in the human bone marrow compartment and not the mouse bone marrow. Taken together, these results suggest that understanding the interactions between grafted leukemic cells and their environment is crucial not only for the development of therapeutic models but also for understanding the development and progression of leukemia. Recently, studies in xenograft models have begun to explore the interaction between leukemic cells, their environmental niche and the factors that support tumor growth and have implicated the SDF-1/CXCR4 axis as playing an important role in these processes [32]. CXCR4 is expressed on hematopoietic stem cells (HSCs) and its interaction with SDF-1 produced by stromal cells supports HSCs in their bone marrow microenvironment. In addition to HSCs, CXCR4 is also present on some myeloid leukemia cell lines and to varying degrees on AML blasts [33]. However, conflicting data exist as to the role played by CXCR4 in the engraftment process of primary leukemia cells in NOD/SCID mice. Tavor et al. [34] demonstrated that treatment of mice with an antibody to CXCR4 following immediate injection of leukemic cells significantly interfered with the engraftment of primary AML cells into NOD/SCID mice. In contrast, Monaco et al. [35] observed significant engraftment in NOD/SCID mice in the absence of CXCR4 expression and pretreatment of CXCR4+ blasts with an anti-CXCR4 antibody did not affect the ability of the AML blasts to engraft. This suggests that a strict correlation may not exist between the surface expression of CXCR4 and engraftment and that other characteristics such as the patient’s disease state may also contribute to the engraftment process in NOD/SCID mice [36]. Although the involvement of CXCR4 in AML engraftment remains controversial, more recent studies have suggested that the SDF-1/ CXCR4 axis may play a role in maintaining leukemic cells in the bone marrow creating a protective environment and promoting AML growth and survival. Initial experiments demonstrated that disruption of this interaction by CXCR4 targeting agents such as AMD3100 (Mozobil) could have significant therapeutic implications. Treatment of primary AML samples or myeloid cell lines with AMD3100 in vitro significantly inhibited AML migration and growth in both a dose- and time-dependent manner [37, 38]. Using a mouse model of APL, Nervi et al. [39] explored the utility of AMD3100 as a mobilizing agent to disrupt the SDF-1/CXCR4 interaction of leukemia cells in the bone marrow and spleen. The APL mouse model was generated by knocking in

330

W. Siders

the human PML-RARa cDNA leading to the expression of the PML-RARa protein and the development of a fatal myeloid leukemia [40]. Leukemic blasts from these mice can be transferred to a recipient mouse where they migrate to the bone marrow and spleen leading to the rapid development of a fatal leukemia within 3 weeks. Although treatment with a single dose of AMD3100 can mobilize HSCs into the periphery, multiple doses of AMD3100 were required before mobilization of the APL cells was observed. Once in the periphery, the circulating APL cells became more sensitive to the cytotoxic effects of AraC treatment. Combination treatment with AMD3100 and AraC significantly increased median survival of the mice compared to either treatment alone suggesting that targeting the CXCR4 axis with Mozobil results in enhanced chemosensitization. Data from these studies have resulted in the initiation of a Phase I clinical trial in AML exploring Mozobil treatment in combination with chemotherapy. Successful therapies that target CXCR4 may become critically important given the recent finding that treatment with imatinib results in the upregulation of CXCR4 on CML cells resulting in their migration to the bone marrow and potentially increased chemoresistance [41].

14.1.2 Models of Acute Lymphocytic Leukemia Acute lymphocytic leukemia (ALL) results from the uncontrolled rapid growth of lymphocytes and is the most common form of leukemia in children under the age of 15. The malignant cells of ALL are thought to originate from early stages of B or T cells. Similar to chronic lymphocytic leukemia (CLL), a majority of cases of ALL express B cell lineage markers and are of B cell origin. Although current treatments have resulted in a 80–90% complete remission of newly diagnosed ALL in children, only 30–40% long-term disease-free survival has been observed in adult ALL patients. The etiology of ALL ranges from (1) aberrant expression of oncogenes, (2) chromosomal translocations such as the BCR/ABL t(9,22) or the TEL/ AML t(12, 21) or (3) hyperdiploidy (for review see Ref. [46]). Given the diversity of the alterations that contribute to the development of ALL, the use of primary leukemic cells isolated form children was viewed as the best approach to develop mouse models of ALL. Uckun et al. [47] initiated a comprehensive study to begin to identify potential correlations between the level of engraftment and treatment outcome of the ALL patients from which the cells were isolated [47]. SCID mice were injected intravenously with leukemic cells from 681 pediatric leukemia patients and monitored to determine the extent of engraftment. Injection of cells from only 104 of the 681 patients was able to engraft and proliferate in SCID mice with primary engraftment occurring within the bone marrow, liver, spleen, thymus and kidney. Although no overall correlation was observed between the extent of engraftment and the survival outcome of the ALL patients, a trend was observed in a subgroup of patients with poor outcome and the extent to which their leukemic cells engrafted in the SCID mice reflecting a more aggressive disease. In similar


331

experiments, the level of engraftment of T-ALL in SCID mice was shown to correlate with patient survival [48]. Twelve of 19 T-ALL samples resulted in successful engraftment leading to disseminated disease involving the bone marrow spleen and thymus. Interestingly, all of the seven samples that failed attempts at engraftment in SCID mice successfully engrafted into NOD/SCID mice suggesting that this mouse may be more receptive to engraftment of primary ALL cells potentially due to the increased immunodeficiencies present in this mouse. Others have followed up on this observation and demonstrated similar findings in some instances resulting in 100% engraftment of cells from multiple patients with few changes, if any, in the immunophenotype following injection of the cells [49–51]. Primary tumor models such as these are currently being used to evaluate anti-cancer agents such as mTOR inhibitors and compounds that target Bruton’s tyrosine kinase [52, 53]. In view of the limited availability of clinical samples for these types of studies, several investigators continue to pursue the use of ALL cell lines in xenograft models, especially for the validation of new therapeutic approaches (Table 14.2). For example, ABL chromosomal translocations have also been observed in T-ALL patients which involve fusion to the NUP214 gene resulting in increased kinase activity in these patients. Using NUP214–ABL positive models of ALL, QuintasCardama et al. [54] have demonstrated that agents targeting the BCR-ABL protein are similarly effective in this setting. Significant anti-tumor activity was observed with dasatinib in NUP214–ABL positive cell lines in xenograft studies but had little

Table 14.2 Lymphocytic leukemia xenograft cell lines Cell line Dose Route Mouse strain Referencea 380 Iv NOD/SCID [61] ³1e7 697 ³1e7 Iv NOD/SCID [61] BA31 5e6 Ip SCID [88] CCRF-CEM 5e6 Iv SCID [59] 4e7 Sc SCID [100] ED (HTLV)b HPB-ALL 1e7 Sc NOD/SCID [54] JOK-1 5e6 Iv SCID [59] Jurkat 1e7 ip, iv SCID [89] JVM-3 1e7 Iv nude [90] MET-1 (HTLV)b 1.5e7 Ip NOD/SCID [97–99] Molt-3 5e6 Sc NOD/SCID [91] Molt-4 5e6 Sc nude [92] Nalm-6 1e6 Iv SCID [93] RS4;11 5e7 ip, iv SCID [94] SIL-ALL 1e7 Sc NOD/SCID [54] TA27 5e6 Ip SCID [88] WSU-CLL 1e7 Sc SCID [95] a Reference cited is not always provided as the originator of the cell line and is included in most instances to provide an example of the tumor growth curve b Model of HTLV induced T cell leukemia

332

W. Siders

to no effect on NUP214–ABL negative tumor cells suggesting the clinical utility in T-ALL patients who express the NUP214–ABL fusion protein. Targeting leukemic cells in xenograft models through the use of TRAIL specific approaches has also been examined in attempts to overcome defects in the apoptotic pathway noted in lymphocytic leukemia (for review see Ref. [55]). Although TRAIL mediates its effects through ligation of death receptors on cells and triggering of apoptosis, a significant number of leukemic cells from pediatric ALL patients that express the TRAIL receptors have been shown to be refractory to TRAIL mediated apoptosis [56]. To overcome this issue, investigators have examined whether the inclusion of small molecules that target other proteins in the apoptotic pathway may enhance TRAIL mediated killing. Inhibitors of the antiapoptotic protein XIAP were shown to cooperate with TRAIL in vitro and demonstrated significant anti-tumor activity as single agents in vivo [57]. Treatment resulted in a decrease in tumor burden based on the number of leukemic blasts present in the blood and reduction in spleen weight. Other approaches aimed at enhancing TRAIL mediated killing include the use of a CD19–TRAIL fusion protein to target CD19 positive leukemic cells [58]. This targeted approach has resulted in significant anti-tumor activity and long-term survival in the Nalm-6 B-ALL xenograft model. Preclinical xenograft models of ALL are also being used to explore other therapeutic approaches including targeting cell surface antigens such as CD47 [59], histone deacetylase inhibitors such as vorinostat [60] and the use of CpG oligonucleotides to stimulate the anti-leukemic activity of NK cells [61].

14.1.3 Chronic Lymphocytic Leukemia A majority of the cases of chronic lymphocytic leukemia (CLL) are of the B cell phenotype (~95%) and result in the expansion and accumulation of malignant CD5+ /CD19+/IgM+ B cells in the blood, lymph nodes, spleen and bone marrow [62]. In contrast to other B cell malignancies that arise from chromosomal translocations, B-CLL often arises as a result of deletions with the most common being the 13q14 deletion on chromosome 13 accounting for approximately half of all B-CLL cases [63]. Cell lines from B-CLL patients were initially transformed for propagation in vitro using EBV induced transformation [64]. Similar to other lymphoid models, the study of B-CLL in mouse models was limited by the inefficient engraftment of cells into SCID mice (for review see Ref. [65]) but several groups were eventually successful at inducing engraftment into either SCID or NOD/SCID mice [66–68]. Several chronic T cell leukemias and lymphomas have been shown to be associated with chromosomal translocations at the Tcl1 (T cell leukemia/lymphoma 1) locus whose oncogenic potential is supported by data generated using a Tcl1 Tg mouse model of leukemia [69]. Recently a model of B-CLL was developed expressing the Tcl1 protein under the control of a B cell specific promoter [70]. This model results in the expansion of CD5+/IgM+ B cells and the development of


333

CLL like disease in older mice that resembles the human disease. The model was further validated by Johnson et al. [71] who demonstrated that the B-CLL like disorder developed by these mice was sensitive to fludarabine treatment. Recently, Tcl1 was found to function as a transcriptional regulator and its overexpression was shown to drive the progression of B-CLL through enhancement of NF-kappa B activity [72]. Planelles et al. [73] have described a similar model of B-CLL lymphomagenesis for APRIL Tg mice. APRIL (a proliferation inducing ligand) together with BAFF (B cell activating factor of the TNF family) play a role in B cell survival and differentiation and are expressed by B-CLL cells [74, 75]. Initially, APRIL Tg mice display normal B cell development in the spleen and lymph nodes [76]. However, as mice age, a progressive expansion of the B-1 B cell compartment can be observed, most notably in the lymph nodes and Peyers patches, reminiscent of the development of human B-CLL [73]. Therefore, both the Tcl1 and APRIL transgenic mice appear to be suitable models for understanding the lymphomagenesis of B-CLL as well as evaluating novel therapeutic approaches. CLL cells have been shown to be resistant to apoptotic-inducing therapies as a result of the overexpression of anti-apoptotic genes such Bcl-2 and XIAP [77]. Therapeutics such as monoclonal antibodies which target cell surface antigens and mediate their effects through additional activities such as antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC) have proven successful in the treatment of CLL (for review see Ref. [78]). Following binding of the antibody to its target antigen, ADCC is triggered through engagement of Fc receptors on effector cells such as NK cells, macrophages and neutrophils. CDC activity is also induced after ligand binding and involves cleavage of C1q and activation of the complement cascade. Immunodeficient mice, in particular the SCID mouse, have proven to be an invaluable tool for evaluating antibody therapeutics and understanding their mechanism of action since both ADCC and CDC pathways are present and functional. Alemtuzumab (Campath) is a monoclonal antibody approved for the treatment of B-CLL [79]. Alemtuzumab binds to the CD52 antigen present on T and B cells and can induce their depletion through both ADCC and CDC mechanisms in vitro. Although we and others have demonstrated the efficacy of antibody-mediated targeting of CD52 in xenograft models [80–82], the exact mechanism through which alemtuzumab mediates its activity in vivo remained undefined. To address this issue, we created a human CD52 (huCD52) Tg mouse that expresses huCD52 in a pattern similar to humans. Treatment of these mice with alemtuzumab resulted in a time- and dose-dependent depletion of CD52+ cells. Interestingly, removal of complement had little or no impact on the lymphocyte-depleting activity of alemtuzumab while removal of NK cells or neutrophils essentially ablated its activity indicating a prominent role for ADCC in lymphocyte depletion [83]. A role for neutrophils in the ADCC activity of alemtuzumab has not been previously described but has been reported as an effector mechanism in the activity of rituximab [84, 85]. Other cell surface antigens currently being explored for the antibody-mediated therapy of CLL include both CD20 targeted by ofatumumab (discussed in detail below) and CD23 targeted by lumiliximab. CD23 functions as a low-affinity

334

W. Siders

r eceptor for IgE and has been shown to be upregulated in B-CLL [86]. Treatment of primary CLL samples with lumiliximab results in the induction of apoptosis that is accompanied by the down regulation of anti-apoptotic proteins such as Bcl-2 and XIAP [87]. The anti-tumor activity of lumiliximab as a single agent was confirmed in the CD23+ SKW6.4 lymphoma xenograft model and was enhanced by the addition of fludarabine or rituximab.

14.1.4 HTLV-Related T Cell Leukemia/Lymphoma Adult T cell leukemia/lymphoma (ATLL) develops in a small number of people infected with the human T cell lymphotrophic virus (HTLV-1) and is characterized by the increased number of malignant activated T cells expressing several conventional T cell markers including CD2, CD3, CD25 and CD52. Conventional therapy has proven to be largely ineffective at treating ATLL patients suggesting the need for new therapies and models for therapeutic analysis. Using peripheral blood lymphocytes from normal donors, Feuer et al. [96] demonstrated that normal T cells infected with the HTLV-1 virus can grow in SCID mice and represent a potential model for assessing therapeutic intervention. Antibody approaches targeting the cell surface antigens expressed by ATLL cells have also been evaluated in xenograft models. These therapies have largely been evaluated using the MET-1 model developed by the Waldmann lab. The MET-1 model was developed by injection of NOD/ SCID mice with cells from an ATLL patient. Targeting markers present on activated T cells such as CD2 [97], CD25 [98] and CD52 [99] has been shown to provide a therapeutic benefit and to be mediated through killing mechanisms such as ADCC. Other xenograft ATLL models such as the ED model have demonstrated the utility of inhibiting proteasome activity through treatment with bortezomib [100].

14.2 Models of Hodgkin’s Lymphoma It is estimated by the American Cancer Society that approximately 8,500 new cases of Hodgkin’s lymphoma (HL) will be diagnosed in 2009. Classical HL which accounts for ~95% of all HL cases can be divided into four subtypes and is characterized by the presence of rare tumorigenic cells called Hodgkin’s and Reed/Sternberg (HRS) cells. HRS cells typically account for only 1% of the cells in the tumor tissue and in most cases represent transformed B cells that have lost typical B cell phenotypical and functional features (for reviews see Refs. [101, 102]). Although current radiation and polychemotherapeutic treatments for HL have resulted in cure rates of up to 90% [103], patients who do not respond to therapy or who eventually relapse have a poor prognosis with limited therapeutic options. As a result of the rare nature of the HRS cells, the development of HL cell lines has proven to be a difficult and rarely successful effort. Nevertheless, some HL cell lines have been generated that


335

Table 14.3 Hodgkin’s lymphoma xenograft cell lines Cell line Dose Route Mouse strain Referencea HDLM-2 Sc SCID [115] ³2e7 Hs445 5e6 Sc SCID [116] KM-H2 ³2e7 Sc SCID [115] L-428 ³2e7 Sc SCID [115] L-540 1e7 sc, iv SCID [107] L540cy 5e6 sc, iv SCID [106] L-1236 2e7 Sc SCID [117] RPMI6666 5e6 Sc SCID [116] a Reference cited is not always provided as the originator of the cell line and is included in most instances to provide an example of the tumor growth curve

can grow in immunodeficient mice (Table 14.3) and have proven useful for studies of tumor biology and the evaluation of therapeutic opportunities. Culturing of HRS cells resulted in the identification of CD30 as a viable target antigen due to its restricted expression pattern. CD30 is a member to the TNF receptor superfamily and is expressed on activated and virally transformed lymphocytes and is highly expressed by HRS cells [104]. Several anti-CD30 antibody targeting approaches have demonstrated benefit in HL xenograft models initially using antibodies conjugated to immunotoxins such as pseudomonas exotoxin A [105] and with naked antibodies mediating conventional killing activities such as ADCC [106, 107]. Using the unconjugated antibody approach, Wahl et al. [106] demonstrated that the anti-CD30 antibody SGN-30 could inhibit proliferation of HL cells in vitro and has potent antitumor activity in both subcutaneous and disseminated HL xenograft models. Within the same time frame, Borchmann et al. [107] described the generation of a fully human anti-CD30 antibody following immunization of the HuMAb mouse (transgenic for the human immunoglobulin genes) with the HL L540 cell line. The 5F11 clone (known as MDX-060) demonstrated strong ADCC activity against HL cell lines in vitro and in vivo. Both antiCD30 antibodies have demonstrated additive or synergistic activity with conventional chemotherapeutic agents in vitro (MDX-060 with gemcitabine and etoposide and SGN-30 with bleomycin and etoposide) with the activity of SGN-30 additionally being confirmed in vivo in the L540cy HL xenograft model [108, 109]. Using agents to deplete either NK cells or macrophages, the anti-tumor activity of SGN30 in the L540cy model was shown to be largely dependent on the presence of macrophages [110]. Based on the activity in HL xenograft models, both SGN-30 and MDX-060 have been evaluated in HL clinical trials. Following a multi-dose escalation phase I clinical trial [111], SGN-30 was evaluated in a phase II study of both HL and anaplastic large cell lymphoma (ALCL). Although no objective responses were observed in the HL group, 29% of the patients did exhibit stable disease while two complete responses and five partial responses were achieved in the ALCL group [112]. Similar activity with MDX-060 was observed in a phase I/ II clinical trial in HL and ALCL with two complete responses in both the ALCL and HL groups [113]. An antibody drug conjugate approach is also being pursued

336

W. Siders

for SGN-30 by coupling it with the tubulin inhibitor auristatin resulting in the SGN-35 antibody construct. Potent anti-tumor activity has been observed in HL xenograft models in combination with several chemotherapeutics including gemcitabine [114].

14.3 Models of Non-Hodgkin’s Lymphoma Non-Hodgkin’s lymphoma (NHL) is the most common type of hematologic malignancy and comprises approximately 35 different subtypes with the most prevalent types being diffuse large B cell lymphoma (DLBL), follicular lymphoma (FL) and mantle cell lymphoma (ML) [118, 119]. Although NHL can arise from both T and B cells, most cases of NHL occur as a result of disregulated B cell growth. Treatments for NHL include both chemotherapy such as CHOP (cyclophosphamide, adriamycin, vincristine and prednisone) and Rituxan which targets CD20 expressed on B cells or a combination of the two known as RCHOP. The success of Rituxan as a stand-alone agent as well in combination with chemotherapeutics has spurred the generation of antibodies that target other B cell specific antigens including CD19 and CD22 (epratuzumab) [120, 121]. In addition, second generation antibodies targeting CD20 are also currently being evaluated including veltuzumab and ofatumumab. Although more than 30 different subtypes of NHL exist, xenograft tumor models have not been generated that represent each subtype (Table 14.4). Instead, investigators have primarily relied on the use of several Burkitt’s lymphoma-derived lines such as the Raji, Daudi and Ramos models to evaluate therapeutic opportunities. Intravenous injection of these cells results in the development of a disseminated model of lymphoma. Tumors cells seed distal locations, grow and eventually cause the development of hind limb paralysis. To monitor the growth of these cells and understand the variability of tumor seeding, we have generated a Raji cell line that Table 14.4 Non-Hodgkin’s lymphoma xenograft cell lines Cell line Dose Route Mouse strain Referencea BJAB 2e7 Sc SCID [151] Daudi 5e6, 1.5e7 sc, iv SCID [129, 149] DoHH-2 1e7 Sc SCID [152] Granta-19 2e7 Sc SCID [151] Namalwa 5e6 Iv SCID [153] Raji 5e6, 1e6 sc, iv SCID [127, 132] Ramos ³5e6 sc, iv SCID [127, 140] SR 5e6 Sc SCID [154] SU-DHL-4 1e6 Iv SCID [155] WSU-FSCCL 2.5e6 Iv SCID [141] a Reference cited is not always provided as the originator of the cell line and is included in most instances to provide an example of the tumor growth curve


337

stably expresses luciferase under the control of the CMV promoter (Raji-luc). Similar to the parental Raji cell line, iv injection of the Raji-luc cells results in the development of hind limb paralysis. These cells can be visualized following an intraperitoneal injection of luciferin and imaging with the IVIS 200 imaging system (Caliper Life Sciences, Hopkinton, MA). Following iv injection, Raji-luc cells were found to disseminate to the lymph nodes, spleen and bone marrow of mice with a significant concentration of cells seeding the spinal cord, which is ultimately responsible for the development of hind limb paralysis (Fig. 14.1). Examination of various organs during necropsy revealed that Raji-luc cells also disseminated to the brain, liver and lung but not consistently in every mouse. Using this type of system, it is possible to begin to examine the effects of therapeutic intervention on tumors growing in multiple locations and determine which factors may influence susceptibility to therapy. In addition to the visual images, a total bioluminescence signal can be measured for each mouse at each time point representing the total tumor burden and establishing the overall efficacy of a therapeutic. The ability to target CD19 in NHL has been explored by several groups in multiple xenograft tumor models and included approaches such as unconjugated antibodies, antibody drug conjugates and bi-specific antibodies. Early proof of concept studies targeting CD19 using an idarubicin-conjugated antibody revealed significant activity in the Nalm-6 tumor model in nude mice [122]. To further validate CD19 as a potential target for antibody mediated therapy and understand its role in B cell development, a human CD19 Tg mouse was created that overexpresses CD19 in a lineage-specific manner [123]. These mice have been crossed with Em-cMyc

Fig. 14.1 Assessment of growth of Raji-luc cells following intravenous injection into SCID mice. SCID mice were injected intravenously with 2e6 Raji-luc cells. On days 14 and 21 post injection, mice were injected intraperitoneally with luciferin at 150 mg/kg, anesthetized and imaged using the IVIS 200 imaging system to visualize the tumor burden in each mouse. Intensity correlates with a higher tumor burden

338

W. Siders

transgenic mice [124, 125], which express the c-Myc oncogene under the control of the Ig heavy chain promoter, resulting in the development of B cell derived lymphomas that express human CD19. Yazawa et al. [126] have demonstrated that tumors from these mice can be isolated, characterized in vitro and adoptively transferred into immunodeficient mice resulting in the presence of circulating CD19+ lymphoblasts that are responsive to treatment with an anti-CD19 antibody. Although in this instance, using an unmodified antibody resulted in anti-tumor responses, efforts involving Fc engineering to enhance the cytolytic function of anti-CD19 antibodies have also been pursued. By encoding mutations within the Fc region to increase the affinity for the Fc receptor, a significant increase in ADCC activity was observed in vitro on several NHL cell lines including Raji, Namalwa and SU-DHL-6 [127]. These mutations also translated into enhanced anti-tumor activity in vivo in both the Raji and Ramos xenograft models. Bi-specific antibodies targeting additional cell surface molecules such as CD16 or CD22 in combination with CD19 have also been explored and have shown efficacy in mice bearing Daudi tumors [128, 129]. Experiments have also been conducted to explore targeting of the CD20 antigen expressed on a large majority of NHL tumors [130, 131]. The activity and mechanism of action of Rituxan alone or in combination with other agents has been explored in several xenograft tumor models including Raji [132], Daudi [133], Ramos [134], BJAB and DoHH2 [135]. Although complement appears to play a significant role in Rituxan-mediated cell lysis [136], several studies also suggest that in mouse xenograft models, neutrophils contribute significantly to this activity through ADCC [84, 85]. SCID mice bearing Raji tumors were treated with the GR-1 antibody to selectively depleted neutrophils and investigate their role in Rituxan-mediated anti-tumor protection. In the absence of these cells, Rituxan activity was severely attenuated indicating that in this model, neutrophils are crucial in mediating anti-tumor activity. In further support of this hypothesis, increasing the number of circulating neutrophils was also found to enhance Rituxan activity. Based on these findings and the clinical success of Rituxan, several efforts are underway to generate and evaluate second-generation anti-CD20 antibodies with enhanced or altered therapeutic profiles. Similar to the Raji-luc model described above, Bleeker et al. [137] have used the Daudi-luc model to examine the efficacy of targeting CD20 using ofatumumab. This antibody demonstrates significant CDC mediated activity on several NHL cell lines in vitro including SU-DHL-4 and Raji cells and binds to a different epitope of CD20 compared to Rituxan [138]. In the Daudi-luc model, treatment with a single dose of ofatumumab at 0.5 mg/kg, 5 days post tumor cell injection resulted in a significant inhibition of tumor growth which could be observed in individual mice by in vivo imaging as well as by a decrease in total bioluminescence. Anti-tumor protection was also observed when treatment was delayed until day 14. In addition to ofatumumab, veltuzumab is also being pursued as an alternative anti-CD20 antibody. Single agent activity has been observed in the Raji, Ramos, Daudi and WSU-FSCCL models either as an unconjugated antibody or conjugated with a radionuclide [139–142]. Based on their in vivo efficacy, ofatumumab and veltuzumab are currently being evaluated in


339

clinical trials of NHL. Currently, ofatumumab is undergoing FDA review for approval in the treatment of CLL. Unlike CD20, ligand binding to CD22 results in rapid internalization [143, 144], which makes this cell surface marker a good target for antibody–drug conjugate therapy. This property has been exploited in experiments examining the efficacy of the radio-labeled anti-CD22 antibody epratuzumab (for review see Refs. [145, 146]) that was originally created by immunizing mice with the Raji cell line [147]. Treatment of nude mice bearing Ramos subcutaneous tumors with 175 mCi of 90 Y-epratuzumab resulted in a significant degree of tumor regression [148]. However, tumors eventually re-grew suggesting that although treatment with 90Y-epratuzumab may induce tumor regression, it does not result in a durable response. Interestingly, the addition of veltuzumab enhanced the anti-tumor activity of 90Y-epratuzumab and resulted in long-term survival of tumor-free mice suggesting that targeting both CD20 and CD22 may have clinical utility. The approach is being explored further using a bispecific antibody engineered from epratuzumab and veltuzumab which has demonstrated efficacy in the Daudi tumor model [149, 150].

14.4 Models of Multiple Myeloma After non-Hodgkins lymphoma, multiple myeloma (MM) is the second most prevalent hematologic cancer representing ~1% of all cancers. According to the American Cancer Society, >20,000 new cases of MM will be diagnosed and >10,000 deaths will occur within the United State s in 2009. Multiple myeloma is a B cell malignancy characterized by accumulation of cancerous plasma cells in the bone marrow and often leads to the formation of bone lesions as a result of an increase in osteoclast activity [156, 157]. Current therapeutic options include treatment with chemotherapeutics such as thalidomide and lenalidomide, the proteasome inhibitor bortezomib and autologous stem cell transplant. Bisphosphonates are also often included in the therapy to delay the progression of bone lesions [158]. Attempts at modeling MM in mice have primarily involved xenograft experiments and the use of MM cell lines. Cell lines established more than 30 years ago including MC/CAR, IM-9, RPMI-8226 and ARH-77 are still proving useful in understanding the progression of MM as well as for the testing of new therapeutic strategies (Table 14.5). Early models of MM involved intraperitoneal injection of MM cells that resulted in circulating levels of human IgG and engraftment of tumor cells in the peritoneal cavity but did not result in the development of disseminated disease or migration to the bone marrow [159, 160]. In contrast, intravenous injection of the ARH-77 cell line into irradiated SCID mice did result in the development of a disseminated disease similar to MM in humans and was characterized by growth in the bone marrow, brain, kidney and liver [161]. In addition, growth of ARH-77 cells was associated with the development of osteolytic lesions further validating this as an adequate model for MM. Alsina et al. [162] further characterized this model with respect to the development of bone lesions and observed bone

340

W. Siders Table 14.5 Multiple myeloma xenograft cell lines Cell line Dose Route Mouse strain Referencea ARH-77 1e7 sc, iv SCID [162] CAG 1e7 Iv NOD/SCID [179] IM9 ³5e6 sc, iv SCID [180] KMS11 1e7 Iv SCID [179] KMS12-PE 1e7 Iv SCID [179] MC/CAR ³1e7 sc, iv SCID [177] MM.1S 2e6 Iv NOD/SCID [181] RPMI-8226 3e7, 5e6 sc, iv SCID [166, 170] S6B45 1e7 Sc Nu/xid/bg [182] a Reference cited is not always provided as the originator of the cell line and is included in most instances to provide an example of the tumor growth curve

destruction with increased osteolytic bone resorption in areas adjacent to the myeloma cells. Antisense blockade of MIP-1a in ARH-77 cells resulted in a decrease in bone destruction as well as a decrease in tumor burden within the bone marrow suggesting an involvement of this pathway in the progression of MM potentially by acting as a growth factor or by promoting interaction with stromal cells [163]. Inhibition of bone destruction was also observed in this model using a RANK-Fc fusion protein that blocks the activities of RANK-L which has been shown to play an important role in osteoclastgenesis [164]. Recently it has been suggested that in addition to inhibiting tumor growth, bortezomib therapy can stimulate bone formation [165]. Bortezomib is a proteasome inhibitor currently approved for the treatment of MM. In preclinical models of MM, treatment of mice bearing RPMI-8226 tumors with bortezomib resulted in an increase in apoptosis and a decrease in angiogenesis [166]. Initial observations by Giuliani et al. and others in MM patients indicate that bortezomib treatment could have both direct and indirect effects on bone formation [167]. This has been further explored in a rabbit-SCID mouse model in which primary human MM cells are injected into nonfetal rabbit bone that has been implanted into SCID mice [168]. In this model, treatment with bortezomib inhibited tumor growth and was associated with an increase in bone mineral density and the number of osteoblasts. Similar to myeloid leukemia, the interaction between the tumor cells and their microenvironment plays a key role in the development and progression of MM. Using more sophisticated mouse models, studies have been conducted to begin examining the mechanisms through which myeloma cells home to the bone marrow and interact with stromal cells. Urashima et al. [169] have developed a bilateral fetal bone hu-SCID mouse model to monitor homing of MM cells and explore which molecules may play a role in this process. Injection of MM cells into the bone marrow of one implant resulted in the migration of cells to the secondary implant but not to mouse bone marrow. Several MM cell lines such as ARH-77 and RPMI-8226 were shown to be able to migrate in this fashion. In vivo imaging techniques are also currently being incorporated into experiments to better visualize the migration and homing patterns of MM cells. Following injection, RPMI-8226 cells


341

expressing the green fluorescent protein (GFP) were also shown to migrate to the spine, skull and pelvis of mice leading to the development of osteolytic lesions similar to the pattern observed in MM patients [170]. The development of models such as these will allow for the testing of therapeutic strategies aimed at inhibiting migration or interrupting the tumor cell stromal cell interaction. To this end, Smallshaw et al. [171] have demonstrated that an antibody targeting ICAM has therapeutic efficacy in the ARH-77 model and results in a significant increase in survival. Other factors such as the SDF-1/CXCR4 axis described earlier are also thought to play an important role in the migration and homing patterns of MM cells. Inhibiting this interaction using AMD3100 reduces the homing of MM.1S cells in NOD/SCID mice [172]. In addition, the SDF-1/CXCR4 pathway may also contribute to the development of osteolytic lesions. Diamond et al. [173] have demonstrated that intra-tibial injection of RPMI-8226 cells resulted in the formation of osteolytic lesions that could be inhibited by systemic administration of agents that affect the SDF-1/CXCR4 interaction [173]. Therefore, disrupting the SDF-1/ CXCR4 interaction may have significant implication not only on the homing of MM cells to the bone marrow but also on the development of bone lesions. The therapeutic potential of targeting cell surface antigens such as CD74, which is expressed on almost 90% of MM tumors [174], has also been evaluated in several MM xenograft models. CD74, also known as the invariant chain, is a type II membrane protein that associates with the MHC class II molecule and is involved in trafficking and antigen loading (for review see Ref. [175]). An antibody targeting CD74 was generated by immunization of mice with the Raji Burkitt lymphoma cell line and was shown to bind to ARH-77 cells [147]. A humanized version of this antibody (hLL1) displays anti-proliferative effects on CD74+ MM cell lines such as ARH-77, MC/CAR and KMS12-PE cells and demonstrates significant anti-tumor activity in the MC/CAR subcutaneous xenograft model [176]. The observation that hLL1 can be rapidly internalized by MM cells following binding to CD74 makes it ideal for an antibody–drug conjugate platform. This potential has been explored in MM xenograft models by coupling hLL1 to doxorubicin [177] or ranpirnase, a novel ribonuclease with potent cytostatic activity [178]. Profound activity was observed in the MC/CAR model using the hLL1–doxorubicin conjugate and has resulted in the evaluation of hLL1, now known as milatuzumab, alone or conjugated to doxorubicin in MM clinical trials. Recently, Stein et al. [179] have demonstrated that the therapeutic activity of milatuzumab in several disseminated MM xenograft models such as CAG or KMS11 can be enhanced by the addition of bortezomib suggesting another potential clinical application.

14.5 Conclusions Advancement of the leukemia and lymphoma research fields has been driven in large part by the utilization of mouse models to understand disease progression and identify new therapeutic opportunities. The development of xenograft tumor models in immunodeficient mice has allowed for the testing of new antibodies and drug

342

W. Siders

combination strategies that are currently undergoing clinical evaluation. These models are crucial for establishing proof of concept data and gauging the efficacy of a therapy on human tumor cells growing in vivo especially in cases where therapies do not cross-react with the mouse homolog. Using these models, the interaction of tumor cells with their microenvironment as it relates to the growth of the primary tumor as well as the development of metastases are key areas of research for many groups (for review see Refs. [183, 184]). What was once thought of as only a supporting structure is now known to play a critical role in tumor growth and involve a mixture of cytokines and cell types. Therapies that affect this dynamic interaction such as AMD3100 which targets the CXCR4/SDF-1 axis are being tested clinically in combination with both biological and small molecule therapies. Although we continue to advance our ability to explore the targeting of new antigens on leukemia and lymphoma cells through mouse models, it is important to realize that each type of model comes with its own set of limitations. For example, in some instances the evaluation of therapeutics in xenograft models takes place in a setting where the target is only expressed on the xenografted cell line. Therefore, the impact of the therapy being analyzed on other cell types that express the antigen cannot be determined. In addition, it is important to understand the origin of the cell line being utilized as well as verify that it expresses the antigens common to the tumor from which it was developed. Recently Drexler et al. [185] have used DNA profiling to assess the validity of over 500 leukemia and lymphoma cell lines. Their results suggest that in several instances cell lines can be misidentified with respect to the actual tumor type from which it originated. In addition, cross-contamination from other rapidly growing cultures such as K562 cells can also result in the generation of false-positive or mixed cell lines. Therefore, it is important that we understand as much as possible the origin of the reagents being used as well as how well they mimic the human disease. As we continue to experiment and learn more about the progression of leukemia and lymphoma, it is clear that mouse models will play a pivotal role in all aspects of this research. Acknowledgements I would like to thank Yanping Hu, Robert Fogel and Gary Jacques for experiments involving in vivo imaging of the Raji-luc cell line. I would also like to thank Johanne Kaplan for critically reading the manuscript.

References 1. Seldin DC. New models of lymphoma in transgenic mice. Curr Opin Immunol. 1995;7:665–73. 2. Tarantul VZ. Transgenic mice as an in vivo model of lymphomagenesis. Int Rev Cytol. 2004;233:123–80. 3. Frese KK, Tuveson DA. Maximizing mouse cancer models. Nat Rev Cancer. 2007;7:645–58. 4. Bernardi R, Grisendi S, Pandolfi PP. Modeling hematopoietic malignancies in the mouse and therapeutical implications. Oncogene. 2002;21:3445–58. 5. Macor P, Secco E, Zorzet S., et al. An update on the xenograft and mouse models suitable for investigating new therapeutic compounds for the treatment of B-cell malignancies. Curr Pharmaceutl Des. 2008;14:2023–39.


343

6. Furth PA, St Onge L, Böger H, et al. Temporal control of gene expression in transgenic mice by a tetracycline-responsive promoter. Proc Natl Acad Sci. 1994;91:9302–6. 7. Huettner CS, Zhang P, Van Etten RA, Tenen DG. Reversibility of acute B-cell leukemia induced by BCR-ABL1. Nat Genet. 2000;24:57–60. 8. Rangarajan A, Weinberg RA. Comparative biology of mouse versus human cells: modeling human cancer in mice. Nat Rev Cancer. 2003;3:952–9. 9. Hudson WA, Li Q, Le C, Kersey JH. Xenotransplantation of human lymphoid malignancies is optimized in mice with multiple immunologic defects. Leukemia. 1998;12:2029–33. 10. Melo JV, Barnes DJ. Chronic myeloid leukemia as a model of disease evolution in human cancer. Nat Rev Cancer. 2007;7:441–53. 11. Melnick A, Licht JD. Deconstructing a disease: RARa, its fusion partners, and their roles in the pathogenesis of acute promyelocytic leukemia. Blood. 1999;93:3167–215. 12. Aichberger KJ, Mayerhofer M, Krauth MT, et al. Low-level expression of proapoptotic Bcl2-interacting mediator in leukemic cells in patients with chronic myeloid leukemia: role of BCR/ABL, characterization of underlying signaling pathways, and reexpression by novel pharmacologic compounds. Cancer Res. 2005;65:9436–44. 13. Druker BJ, Guillot F, O’Brien SG, et al. Five year follow-up of patients receiving Imatinib for chronic myeloid leukemia. N Engl J Med. 2006;355:2408–17. 14. Sawyers CL, Gishizky ML, Quan S, et al. Propagation of human blastic myeloid leukemias in the SCID mouse. Blood. 1992;79:2089–98. 15. De Lord C, Clutterbuck R, Powles R, et al. Growth of primary human acute lymphoblastic and myeloblastic leukemia in SCID mice. Leukemia Lymphoma. 1994;16:157–65. 16. Lapidot T, Pflumio F, Doedens M, et al. Cytokine stimulation of multilineage hematopoeisis from immature human cells engrafted in SCID mice. Science. 1992;255:1137–41. 17. Lapidot T, Sirard C, Vormoor J, et al. AML blast cells obtained from newly diagnosed patients engraft and proliferate in SCID mice in response to cytokines. Blood. 1992;80:32a. 18. Lange B, Valteiri M, Santoli D, et al. Growth factor requirements of childhood acute leukemia: establishment of GM-CSF-dependent cell lines. Blood. 1987;70:192–9. 19. Chelstrom LM, Gunther R, Simon J, et al. Childhood acute myeloid leukemia in mice with severe combined immunodeficiency. Blood. 1994;84:20–6. 20. Beran M, Pisa P, O’Brien S, et al. Biologic properties and growth in SCID mice of a new myelogenous leukemia cell line (KBM-5) derived from chronic myelogenous leukemia cells in the blastic phase. Cancer Res. 1993;53:3603–10. 21. Hunter T. Treatment for chronic myelogenous leukemia: the long road to imatinib. J Clin Investig. 2007;117:2036–43. 22. Buchdunger E, Zimmermann J, Mett H, et al. Inhibition of Abl protein-tyrosine kinase in vitro and in vivo by a 2-Phenylaminooyrimidine derivative. Cancer Res. 1996;56:100–4. 23. Druker BJ, Tamura S, Buchdunger E, et al. Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr-abl positive cells. Nat Med. 1996;2:561–6. 24. Le Coutre P, Mologni L, Cleris L, et al. In vivo eradication of human BCR/ABL-positive leukemia cells with an ABL kinase inhibitor. J Natl Cancer Inst. 1999;91:163–8. 25. Cowan-Jacob SW, Guex V, Fendrich G, et al. Imatinib (STI571) resistance in chronic myelogenous leukemia: molecular basis of the underlying mechanisms and potential strategies for treatment. Mini Rev Med Chem. 2004;4:85–99. 26. O’Hare T, Walters DK, Stoffegen EP, et al. In vitro activity of Bcr-Abl inhibitors AMN107 and BMS-354825 against clinically relevant imatinib-resistant Abl kinase domain mutants. Cancer Res. 2005;65:4500–5. 27. Verstovsek S. Preclinical and clinical experience with dasatinib in Philadelphia chromosome-negative leukemias and myeloid disorders. Leukemia Res. 2009;33:617–23. 28. Shah NP, Tran C, Lee FY, et al. Overriding imatinib resistance with a novel ABL kinase inhibitor. Science. 2004;305:399–401. 29. Luo FR, Yang Z, Camuso A, et al. Dasatinib (BMS-354825) pharmacokinetics and pharmacodynamic biomarkers in animal models predict optimal clinical exposure. Clin Cancer Res. 2006;12:7180–6.

344

W. Siders

30. Kyoizumi S, Baum CM, Kaneshima H, et al. Implantation and maintenance of functional human bone marrow in SCID-hu mice. Blood. 1992;79:1704–11. 31. Namikawa R, Ueda R, Kyoizumi S. Growth of human myeloid leukemias in the human marrow environment of SCID-hu mice. Blood. 1993;82:2526–36. 32. Burger JA, Kipps TJ. CXCR4: a key receptor in the crosstalk between tumor cells and their microenvironment. Blood. 2006;107:1761–7. 33. Mohle R, Bautz F, Rafii S, et al. The chemokine receptor CXCR4 is expressed on CD34+ hematopoietic progenitors and leukemic cells and mediates transendothelial migration induced by stromal cell-derived factor-1. Blood. 1998;91:4523–30. 34. Tavor S, Petit I, Porozov S, et al. CXCR4 regulates migration and development of human acute myelogenous leukemia stem cells in transplanted NOD/SCID mice. Cancer Res. 2004;64:2817–24. 35. Monaco G, Konopleva M, Munsell M, et al. Engraftment of acute myeloid leukemia in NOD/SCID mice is independent of CXCR4 and predicts poor patient survival. Stem Cells. 2004;22:188–201. 36. Pearce DJ, Taussig D, Zibara K, et al. AML engraftment in the NOD/SCID assay reflects the outcome of AML: implications for our understanding of the heterogeneity of AML. Blood. 2006;107:1166–73. 37. Liesveld JL, Bechelli J, Rosell K, et al. Effects of AMD3100 on transmigration and survival of acute myelogenous leukemia cells. Leukemia Res. 2007;31:1553–63. 38. Tavor S, Eisenbach M, Jacob-Hirsch J, et al. The CXCR4 antagonist AMD3100 impairs survival of human AML cells and induces their differentiation. Leukemia. 2008;22:2151–8. 39. Nervi B, Ramirez P, Rettig MP, et al. Chemosensitization of AML following mobilization by the CXCR4 antagonist AMD3100. Blood. 2009;113:6206–14. 40. Westervelt P, Lane AA, Pollock JL, et al. High-penetrance mouse models of acute promyelocytic leukemia with very low levels of PML-RARalpha expression. Blood. 2003;102:1857–65. 41. Jin L, Tave Y, Konoplev S, et al. CXCR4 up-regulation by imatinib induces chronic myelogenous leukemia (CML) cell migration to bone marrow stroma and promotes survival of quiescent CML cells. Mol Cancer Therapeut. 2008;7:48–58. 42. Auclair, D, Miller D, Yatsula V, et al. Antitumor activity of sorafinib in Flt3-driven leukemic cells. Leukemia. 2007;21:439–45. 43. Vey N, Giles FJ, Kantarjian H, et al. The topoisomerase inhibitor DX-8951f is active in a severe combined immunodeficient mouse model of human acute myelogenous leukemia. Clinical Cancer Res. 2000;6:731–6. 44. Nishioka C, Ikezoe T, Koeffler HP, Yokoyama A. Blockade of mTOR signaling potentiates the ability of histone deacetylase inhibitor to induce growth arrest and differentiation of acute myelogenous leukemia cells. Leukemia. 2008;22:2159–68. 45. Takeo T, Ogura M, Okumura M, et al. New xenografts of human megakaryoblastic cell line (MEG-01) for evaluating anti-tumor agents. Leukemia. 1993;7:1268–74. 46. Oui CH, Relling MV, Downing JR. Acute lymphoblastic leukemia. N Engl J Med. 2004;350:1535–48. 47. Uckun FM, Sather HN, Waurzyniak BJ, et al. Prognostic significance of B-lineage leukemic cell growth in SCID mice: a children’s cancer group study. Leukemia Lymphoma. 1998;30:503–14. 48. Steele JPC, Clutterbuck RD, Powles RL et al. Growth of human T cell lineage acute leukemia in severe combined immunodeficiency (SCID) mice and non-obese diabetic SCID mice. Blood. 1997;90:2015–19. 49. Lock RB, Liem N, Farnsworth ML, et al. The nonobese diabetic/severe combined immunodeficient (NOD/SCID) mouse model of childhood acute lymphoblastic leukemia reveals intrinsic differences in biologic characteristics at diagnosis and relapse. Blood. 2002;99:4100–8. 50. Borgmann A, Baldy C, Stackelberg, A et al. Childhood ALL blasts retain phenotypic and genotypic characteristics upon long-term serial passage in NOD/SCID mice. Pediatr Hematol Oncol. 2000;17:635–50.


345

51. Liem N, Papa RA, Milross CG, et al. Characterization of childhood acute lymphoblastic leukemia xenograft models for the preclinical evaluation of new therapies. Blood. 2004;103:3905–14. 52. Teachey DT, Sheen C, Hall J, et al. mTOR inhibitors are synergistic with methotrexate: an effective combination to treat acute lymphoblastic leukemia. Blood. 2008;112:2020–3. 53. Uckun FM. Clinical potential of targeting Bruton’s tyrosine kinase. Int Rev Immunol. 2008;27:43–69. 54. Quintas-Cardama A, Tong W, Manshouri T, et al. Activity of tyrosine kinase inhibitors against human NUP214-ABL1-positive T cell malignancies. Leukemia. 2008;22:1117–24. 55. Reed JC, Pellecchia M. Apoptosis-based therapies for hematologic malignancies. Blood. 2005;106:408–18. 56. Ehrhardt H, Fulda S, Schmid I, et al. TRAIL induced survival and proliferation in cancer cells resistant towards TRAIL-induced apoptosis mediated by NF-kB. Oncogene. 2003;22:3842–52. 57. Fakler M, Loeder S, Vogler M, et al. Small molecule XIAP inhibitors cooperate with TRAIL to induce apoptosis in childhood acute leukemia cells and overcome Bcl-2-mediated resistance. Blood. 2009;113:1710–22. 58. Steiglmaier J, Bremer E, Kellner C, et al. Selective induction of apoptosis in leukemic B-lymphoid cells by a CD19-specific TRAIL fusion protein. Cancer Immunol Immunother. 2008;57:233–46. 59. Uno S, Kinoshita Y, Azuma Y, et al. Antitumor activity of a monoclonal antibody against CD47 in xenograft models of human leukemia. Oncol Rep. 2007;17:1189–94. 60. Einsiedel HG, Kawan L, Eckert C, et al. Histone deacetylase inhibitors have antitumor activity in two NOD/SCID mouse models of B-cell precursor childhood acute lymphoblastic leukemia. Leukemia. 2006;20:1435–64. 61. Fujii H, Trudeau JD, Teachey DT, et al. In vivo control of acute lymphoblastic leukemia by immunostimulatory CpG oligonucleotides. Blood. 2007;109:2008–13. 62. Boelens J, Lust S, Vanhoecke B, Offner F. Chronic lymphocytic leukemia. Anticancer Res. 2009;29:605–16. 63. Dohner H, Stilgenbauer S, Benner A, et al. Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med. 2000;343:1910–6. 64. Lee CLT, Uniyal S, Fernandez LA, et al. Growth and spread in nude mice of Epstein-Barr virus transformed B-cells from a chronic lymphocytic leukemia patient. Cancer Res. 1986;46:2497–501. 65. Kobayashi R, Picchio G, Kirven M, et al. Transfer of human chronic lymphocytic leukemia to mice with severe combined immune deficiency. Leukemia Res. 1992;16:1013–23. 66. Kawata A, Han T, Dadey B, et al. Establishment and characterization of the tumors of chronic lymphocytic leukemia cell line in nude and SCID mice. Leukemia Res. 1993;17:883–94. 67. Shimoni A, Marcus H, Canaan A, et al. A model for human B-chronic lymphocytic leukemia in human/mouse radiation chimera: Evidence for tumor-mediated suppression of antibody production in low-stage disease. Blood. 1997;89:2210–8. 68. Mosier DE, Gulizia RJ, Baird SM, et al. Transfer of a functional human immune system to mice with severe combined immunodeficiency. Nature. 1989;335:256–9. 69. Pekarsky Y, Hallas C, Croce CM. The role of Tcl1 in human T-cell leukemia. Oncogene. 2001;20:5638–43. 70. Bichi R, Shinton SA, Martin ES, et al. Human chronic lymphocytic leukemia modeled in mouse by targeted TCL1 expression. Proc Natl Acad Sci. 2002;99:6955–60. 71. Johnson AJ, Lucas DM, Muthusamy N, et al. Characterization of the TCL1 transgenic mouse as a preclinical drug development tool for human chronic lymphocytic leukemia. Blood. 2006;108:1334–8. 72. Pekarsky Y, Palamarchuk A, Maximov V, et al. Tcl1 functions as a transcriptional regulator and is directly involved in the pathogenesis of CLL. Proc Natl Acad Sci. 2008;105:19643–8.

346

W. Siders

73. Planelles L, Carvalho-Pinto CE, Hardenberg G, et al. APRIL promotes B-1 cell associated neoplasm. Cancer Cell. 2004;6:399–408. 74. Kern C, Cornuel JF, Billard C, et al. Involvement of BAFF and APRIL in the resistance to apoptosis of B-CLL through an autocrine pathway. Blood. 2004;103:679–88. 75. Haiat S, Billard C, Quiney C, et al. Role of BAFF and APRIL in human B-cell chronic lymphocytic leukemia. Immunology. 2006;118:281–92. 76. Stein JV, Lopez-Fraga M, Elustondo FA, et al. APRIL modulates B and T cell immunity. J Clin Investig. 2002;109:1587–98. 77. Packham G, Stevenson FK. Bodyguards and assassins: Bcl-2 family proteins and apoptosis control in chronic lymphocytic leukemia. Immunology. 2005:114 441–9. 78. Christian BA, Lin TS. Antibody therapy for chronic lymphocytic leukemia. Semin Hematol. 2008;45:95–103. 79. Boyd K, Dearden CE. Alemtuzumab in the treatment of chronic lymphocytic leukemia. Expert Rev Anticancer Ther. 2008;8:525–33. 80. Siders WM, Shields J, Garron C, et al. Involvement of neutrophils and natural killer cells in the anti-tumor activity of alemtuzumab in xenograft tumor models. Submitted for publication. 81. Zhang Z, Zhang M, Goldman CK, et al. Effective therapy for a murine model of adult T-cell leukemia with the humanized anti-CD52 monoclonal antibody, Campath-1H. Cancer Res. 2003;63:6453–7. 82. Lapalombella R, Zhao X, Triantafillou G, et al. A novel Raji-Burkitt’s lymphoma model for preclinical and mechanistic evaluation of Cd52-targeted immunotherapeutic agents. Clin. Cancer Res. 2008;14:569–78. 83. Hi Y, Turner MJ, Shields J, et al. Investigation of the mechanism of action of alemtuzumab in a human CD52 transgenic mouse model. Immunology (in press). 84. Hernandez-Ilizaliturri F, Jupudy V, Ostberg J, et al. Neutrophils contribute to the biologic antitumor activity of rituximab in a non-Hodgkin’s lymphoma severe combined immunodeficiency mouse model. Clinical Cancer Res. 2003;9:5866–73. 85. Hernandez-Ilizaliturri F, Jupudy V, Reising S, et al. Concurrent administration of granulocyte colony-stimulating factor or granulocyte-monocyte colony-stimulating factor enhances the biological activity of rituximab in a severe combined immunodeficiency mouse lymphoma model. Leukemia Lymphoma. 2005;46:1775–84. 86. Lopez-Matas M, Rodriques-Justo M, Morilla R, et al. Quantitative expression of CD23 and its ligand CD21 in chronic lymphocytic leukemia. Hematologica. 2000;85:1140–5. 87. Pathan NI, Chu P, Hariharan K, et al. Mediation of apoptosis by and antitumor activity of lumiliximab in chronic lymphocytic leukemia cells and CD23+ lymphoma cell lines. Blood. 2008;111:1594–1602. 88. Yan Y, Steinherz P, Klingemann HG, et al. Antileukemia activity of a natural killer cell line against human leukemias. Clinical Cancer Res. 1998;4:2859–68. 89. Zengerle U, Hoffman-Fezer G, Schuh R, Thierfelder S. Suppression of normal and neoplastic human T cells with unconjugated monoclonal antibodies in SCID mouse chimeras. Bone Marrow Transpl. 1995;16:337–41. 90. Loisel S, Le Ster K, Quitin-Roue I, et al. Establishment of a novel B-CLL-like xenograft model in nude mouse. Leukemia Res. 2005;29:1347–52. 91. Indraccolo S, Steivano L, Minuzzo S, et al. Interruption of tumor dormancy by a transient angiogenic burst within the tumor microenvironment. Proc Natl Acad Sci. 2006;103:4216–21. 92. Peng L, Liu R, Andrei M, et al. In vivo optical imaging of human lymphoma xenograft using a library-derived peptidomimetic against a4b1 integrin. Mol Cancer Therapeut. 2008;7:432–7. 93. Brentjens RJ, Santos E, Nikhamin Y, et al. Genetically targeted T cells eradicate systemic acute lymphoblastic leukemia xenografts. Clinical Cancer Res. 2007;13:5426–35. 94. Jansen B, Uckun FM, Jaszcz WB, Kersey JH. Establishment of a human t(4;11) leukemia in severe combined immunodeficient mice and successful treatment using anti-CD19 (B43)pokeweed antiviral protein immunotoxin. Cancer Res. 1992;52:406–12.


347

95. Mohammad RM, Mohamed AN, Hamdan MY et al. Establishment of a human B-CLL xenograft model: utility as a preclinical therapeutic model. Leukemia. 1996;10:130–7. 96. Feuer G, Zack JA, Harrington WJ, et al. Establishment of human T-cell leukemia virus Type I T-cells lymphomas in severe combined immunodeficient mice. Blood. 1993;82:722–31 97. Zhang Z, Zhang M, Ravetch JV et al. Effective therapy for a murine model of adult T-cell leukemia with the humanized anti-CD2 monoclonal antibody MEDI-507. Blood. 2003;103:284–8. 98. Zhang M, Zhang Z, Garmestani K, et al. Activating Fc receptors are required for antitumor efficacy of the antibodies directed toward CD25 in a murine model of adult T-cell leukemia. Cancer Res. 2004;64:5825–9. 99. Zhang Z, Zhang M, Goldman CK, et al. Effective therapy for a murine model of adult T-cell leukemia with the humanized anti-CD52 monoclonal antibody, Campath-1H. Cancer Res. 2003;63:6453–7. 100. Satou Y, Nosaka K, Koya Y, et al. Proteasome inhibitor, bortezomib, potently inhibits the growth of adult T-cell leukemia cells both in vivo and in vitro. Leukemia. 2004;18:1357–63. 101. Kuppers R. The biology of Hodgkin’s lymphoma. Nat Rev Cancer. 2009;9:15–27. 102. Brauninger A, Schmitz R, Bechtel D, et al. Molecular biology of Hodgkin’s and Reed/ Sternberg cells in Hodgkin’s lymphoma. Int J Cancer. 2006;116:1853–61. 103. Diehl V, Thomas RK, Re D. Hodgkin’s lymphoma-diagnosis and treatment. Lancet Oncol. 2004;5:19–26. 104. Gruss HJ, Herrmann F. CD30 ligand, a member of the TNF ligand superfamily, with growth and activation control for CD30+ lymphoid and lymphoma cells. Leukemia Lymphoma. 1996;20:397–409. 105. Barth S, Huhn M, Matthey B, et al. Ki-4(scFv)-ETA’, a new recombinant anti-CD30 immunotoxin with highly specific activity against disseminated Hodgkin tumors in SCID mice. Blood. 2000;95:3909–14. 106. Wahl AF, Klussman K, Thompson JD, et al. The anti-CD30 monoclonal antibody SGN-30 promotes growth arrest and DNA fragmentation in vitro and affects activity in models of Hodgkin’s disease. Cancer Res. 2002;62:3736–42. 107. Borchmann P, Treml JF, Hansen H, et al. The human anti-CD30 antibody 5F11 shows in vitro and in vivo activity against malignant lymphoma. Blood. 2003;102:3737–42. 108. Hueck F, Ellermann J, Borchmann P, et al. Combinations of the anti-CD30 antibody 5F11 with cytostatic drugs enhances its antitumor activity against Hodgkin and anaplastic large cell lymphoma lines. J Immunother. 2004;27:347–53. 109. Cerveny CC, Law CL, McCormick RS, et al. Signaling via the anti-CD30 mAb SGN-30 sensitizes Hodgkin’s disease cells to conventional chemotherapeutics. Leukemia. 2005;19:1648–55. 110. Oflazoglu E, Stone IJ, Gordon KA, et al. Macrophages contribute to the antitumor activity of the anti-CD30 antibody SGN-30. Blood. 2007;110:4370–2. 111. Bartlett NL, Younes A, Carabasi MH, et al. A phase I multidose study of SGN-30 immunotherapy in patients with refractory or recurrent CD30+ hematologic malignancies. Blood. 2008;111:1848–54. 112. Forero-Torres A, Leonard JP, Ypounes A, et al. A phase II study of SGN-30 (anti-CD30 mAb) in Hodgkin lymphoma or systemic anaplastic large cell lymphoma. Br J Haematol. 2009; Epub. 113. Ansell SM, Horwitz SM, Engert A, et al. Phase I/II study of an anti-CD30 monoclonal antibody (MDX-060) in Hodgkin’s lymphoma and anaplastic large-cell lymphoma. J Clin Oncol. 2007;25:2764–9. 114. Oflazoglu E, Kissler KM, Sievers EL, et al. Combination of the anti-CD30-auristatin-E antibody-drug conjugate (SGN-35) with chemotherapy improves antitumor activity in Hodgkin lymphoma. Br J Haematol. 2008;142:69–73. 115. Von Kalle, C, Wolf J, Becker A, et al. Growth of Hodgkin cell lines in severely combined immunodeficient mice. Int J Cancer. 1992;52:887–91.

348

W. Siders

116. Modak S, Koehne G, Vickers A, et al. Rituxamab therapy of lymphoma is enhanced by orally administered (1 3),(1 4)-d-beta-glucan. Leukemia Res. 2005;29:679–83. 117. Wolf J, Kapp U, Bohlen H, et al. Peripheral blood mononuclear cells of a patient with advanced Hodgkin’s lymphoma give rise to permanently growing Hodgkin-Reed Sternberg cells. Blood. 1996;87:3418–28. 118. Armatiage JO, Weisenburger DD. New approach to classifying non-Hodgkin’s lymphomas: clinical features of the major histology subtypes. J Clin Oncol. 1998;16:2780–95. 119. Jaffe ES, Harris NL, Stein H, Isaacson PG. Classification of lymphoid neoplasms: the microscope as a toll for disease discovery. Blood. 2008;112:4384–99. 120. Castillo J, Winer E, Quesenberry P. Newer monoclonal antibodies for hematological malignancies. Exp Hematol. 2008;36:755–68. 121. Cheson BD, Leonard JP. Monoclonal antibody therapy for B-cell non-Hodgkin’s lymphoma. N Engl J Med. 2008;359:613–26. 122. Rowland AJ, Pietersz GA, McKenzie IF. Preclinical investigation of the antitumor effects of anti-CD19-idarubicin immunoconjugates. Cancer Immunol Immunother. 1993;37:195–202. 123. Zhou LJ, Smith HM, Waldschmidt TJ, et al. Tissue-specific expression of the human CD19 gene in transgenic mice inhibits antigen-independent B-lymphocyte development. Mol Cell Biol. 1994;14:3884–94. 124. Adams JM, Harris AW, Pinkert CA, et al. The c-myc oncogene driven by immunoglobulin enhancers induces lymphoid malignancy in transgenic mice. Nature. 1985;318:533–8. 125. Harris AW, Pinkert CA, Crawford M, et al. The Em-myc transgenic mouse: a model for highincidence spontaneous lymphoma and leukemia of early B-cells. J Exp Med. 1988;167:353–71. 126. Yazawa N, Hamaguchi Y, Poe JC, Tedder TF. Immunotherapy using unconjugated CD19 monoclonal antibodies in animal models for B lymphocyte malignancies and autoimmune disease. Proc Natl Acad Sci. 2005;102:15178–83. 127. Horton HM, Bernett MJ, Pong E, et al. Potent in vitro and in vivo activity of an Fc-engineered anti-CD19 monoclonal antibody against lymphoma and leukemia. Cancer Res. 2008;68:8049–57. 128. Bruenke J, Barbin K, Kunert S, et al. Effective lysis of lymphoma cells with a stabilized bispecific single-chain Fv antibody against Cd19 and FcgRIII (CD16). Br J Hematol. 2005;130:2218–28. 129. Vallera DA, Todhunter DA, Kuroki DW, et al. A bispecific recombinant immunotoxin, DT2219, targeting human CD19 and CD22 receptors in a mouse xenograft model of B-cell leukemia/lymphoma. Clin Cancer Res. 2005;11:3879–88. 130. Nadler LM, Ritx J, Hardy R, et al. A unique cell surface antigen identifying lymphoid malignancies of B cell origin. J Clin Investig. 1981;67:134–40. 131. Anderson KC, Bates MP, Slaughenhoupt BL, et al. Expression oh human B-cell associated antigens on leukemias and lymphomas: a model of human B cell differentiation. Blood. 1984;63:1424–33. 132. Maddipatla S, Hernandez-Ilialiturri FJ, Knight J, Czuczman MS. Augmented antitumor activity against B-cell lymphoma by a combination of monoclonal antibodies targeting TRAIL-R1 and CD20. Clin Cancer Res. 2007;13:4556–64. 133. Lopez de Menezes DE, Denis-Mize K, Tang Y, et al. Recombinant interleukin-2 significantly augments activity of rituximab in human tumor xenograft models of B-cell non-Hodgkin lymphoma. J Immunother. 2007;30:64–74. 134. Pagel JM, Laugen C, Bonham L, et al. Induction of apoptosis using inhibitors of lysophosphatidic acid acyltransferase-b and anti-CD20 monoclonal antibodies for treatment of human non-Hodgkin’s lymphomas. Clin Cancer Res. 2005;11:4857–66. 135. Daniel D, Yang B, Lawrence DA, et al. Cooperation of the proapoptotic receptor agonist rhApo2L/TRAIL with the Cd20 antibody rituximab against non-Hodgkin lymphoma xenografts. Blood. 2007;110:4037–46. 136. Zhou X, Hu W, Qin X. The role of complement in the mechanism of action of rituximab for B-cell lymphoma: implications for therapy. Oncologist. 2008;13:954–66.


349

137. Bleeker WK, Munk ME, Mackus WJM, et al. Estimation of dose requirements for sustained in vivo activity of a therapeutic human anti-CD20 antibody. Br J Hematol. 2007;140:303–12. 138. Teeling JL, Mackus WJM, Wiegman LJJM, et al. The biological activity of human CD20 monoclonal antibodies is linked to unique epitopes on CD20. J Immunol. 2006;177:362–71. 139. Stein R, Qu Z, Chen S, et al. Characterization of a new humanized anti-CD20 monoclonal antibody, IMMU-106 and its use in combination with the humanized anti-CD22 antibody, epratuzumab, for the therapy of non-Hodgkin’s lymphoma. Clin Cancer Res. 2004;10:2868–78. 140. Sharkey RM, Karacay H, Chang C-H, et al. Improved therapy of non-Hodgkin’s lymphoma xenografts using radionuclides pretargeted with a new anti-CD20 bispecific antibody. Leukemia. 2005;19:1064–9. 141. Goldenberg DM, Rossi EA, Stein R, et al. Properties and structure–function relationships of veltuzumab (hA20) a humanized anti-CD20 monoclonal antibody. Blood. 2009;113:1062–70. 142. Sharkey RM, Karacay H, Johnson CR, et al. Pretargeted versus directly targeted radioimmunotherapy combined with anti-CD20 antibody consolidation therapy of non-Hodgkin lymphoma. J Nucl Med. 2009;50:444–53. 143. Shih LB, Lu HH, Xuan H, Goldenberg DM. Internalization and intracellular processing of an anti-B cell monoclonal antibody, LL2. 1994;56:538–45. 144. Shan D, Press OW. Constitutive endocytosis and degradation of CD22 by human B cells. J Immunol. 1995;154:4466–75. 145. Coleman M, Goldenberg DM, Siegel AB, et al. Epratuzumab: targeting B-cell malignancies through CD22. Clin Cancer Res. 2003; 9:3991s–4s. 146. Leonard JP, Goldenberg DM. Preclinical and clinical evaluation of epratuzumab (anti-CD22 IgG) in B-cell malignancies. Oncogene. 2007;26:3704–13. 147. Pawlak-Byczowska EJ, Hansen HJ, Dion AS, Goldenberg DM. Two new monoclonal antibodies, EPB-1 and EPB-2, reactive with human lymphoma. Cancer Res. 1989;49:4568–77. 148. Mattes MJ, Sharkey RM, Karacay H, et al. Therapy of advanced B-lymphoma xenografts with a combination of 90Y-anti-CD22 IgG (epratuzumab) and unlabeled anti-CD20 IgG (veltuzumab). Clinical Cancer Res. 2008;14:6154–60. 149. Qu Z, Goldenberg DM, Cardillo TM, et al. Bispecific anti CD20/22 antibodies inhibit B-cell lymphoma proliferation by a unique mechanism. Blood. 2008;111:2211–9. 150. Rossi EA, Goldenberg DM, Cardillo TM, et al. Hexavalent bispecific antibodies represent a new class of anticancer therapeutics: 1. Properties of anti-CD20/22 antibodies in lymphoma. Blood. 2009;113:6161–71. 151. Polson AG, Yu S-F, Elkins K, et al. Antibody-drug conjugates targeted to CD79 for the treatment of non-Hodgkin lymphoma. Blood. 2007;110:616–23. 152. Schliemann C, Palumbo A, Zuberbuhler K, et al. Complete eradication of human B-cell lymphoma xenografts using rituximab in combination with immunocytokine L19-IL2. Blood. 2009;113:2275–83. 153. Sapra P, Moase EH, Ma J, Allen TM. Improved therapeutic responses in a xenograft model of human B lymphoma (Namalwa) for liposomal vincristine versus liposomal doxorubicin targeted via anti-CD19 IgG2a or Fab’ fragments. Clin Cancer Res. 2004;10:1100–11. 154. Cooper AC, Karp RM, Clark EJ, et al. A novel methionine aminopeptidase-2 inhibitor, PPI2458, inhibits non-Hodgkin’s lymphoma cell proliferation in vitro and in vivo. Clin Cancer Res. 2006;12:2583–90. 155. Schmidt-Wolf IGH, Negrin RS, Kiem H-P, et al. Use of a SCID mouse/human lymphoma model to evaluate cytokine-induced killer cells with potent antitumor cell activity. J Exp Med. 1991;174:139–49. 156. Roodman GD. Mechanisms of bone lesions in multiple myeloma and lymphoma. Cancer. 1997;80:1557–63. 157. Gertz MA, Ghobrial I, Luc-Harousseau J. Multiple myeloma: biology, standard therapy, and transplant therapy. Biol Blood Marrow Transpl. 2009;15:46–52. 158. Palumbo A, Rajkumar SV. Treatment of newly diagnosed myeloma. Leukemia. 2009;23:449–56.

350

W. Siders

159. Valeriote F, Grates H. MOPC-315 murine plasmacytoma as a model anticancer screen for human multiple myeloma. J Natl Cancer Inst. 1986;76:61–5. 160. Feo-Zuppardi FJ, Taylor CW, Iwato K, et al. Long-term engraftment of fresh human myeloma cells in SCID mice. Blood. 1992;80:2843–50. 161. Huang Y-W, Richardson JA, Tong AW, et al. Disseminated growth of a human multiple myeloma cell line in mice with severe combined immunodeficiency disease. Cancer Res. 1993;53:1392–6. 162. Alsina M, Boyce B, Devlin RD, et al. Development of an in vivo model of human multiple myeloma bone disease. Blood. 1996;87:1495–501. 163. Choi SJ, Oba Y, Gazitt Y, et al. Antisense inhibition of macrophage inflammatory protein 1-a blocks bone destruction on a model of myeloma bone disease. J Clin Investig. 2001;108:1833–41. 164. Sordillo EM, Pearse RN. RANK-Fc: a therapeutic antagonist for RANK-L in myeloma. Cancer. 2003;97:802–12. 165. Garrett IR, Chen D, Gutierrez G, et al. Selective inhibitors of the osteoblast proteasome stimulate bone formation in vivo and in vitro. J Clin Investig. 2003;111:1771–82. 166. LeBlanc R, Catley LP, Hideshima T, et al. Proteasome inhibitor PS-341 inhibits human myeloma cell growth in vivo and prolongs survival in a murine model. Cancer Res. 2002;62:4996–5000. 167. Giuliani N, Morandi F, Tagliaferri S, et al. The proteasome inhibitor bortezomib affects osteoblast differentiation in vitro and in vivo in multiple myeloma patients. Blood. 2007;110:334–8. 168. Pennisi A, Li X, Ling W, et al. The proteasome inhibitor bortezomib suppresses primary myeloma and stimulates bone formation in myelomatous and nonmyelomatous bones in vivo. Am J Hematol. 2009;84:6–14. 169. Urashima M, Chen BP, Chen S, et al. The development of a model for the homing of multiple myeloma cells to human bone marrow. Blood. 1997;90:754–65. 170. Mitsiades CS, Mitsiades NS, Bronson RT, et al. Fluorescence imaging of multiple myeloma cells in a clinically relevant SCID/NOD in vivo model: biologic and clinical implications. Cancer Res. 2003;63:6689–96. 171. Smallshaw JE, Coleman E, Spiridon C, Vitetta ES. The generation of anti-myeloma activity of a chimeric anti-CD54 antibody, cUV3. J Immunother. 2004;27:419–24. 172. Alsayed Y, Ngo H, Runnels J, et al. Mechanisms of regulation of CXCR4/SDF-1 (CXCL12)dependent migration and homing in multiple myeloma. Blood. 2007;109:2708–17. 173. Diamond P, Labrinidis A, Martin SK, et al. Targeted disruption of the CXCL12/CXCR4 axis inhibits osteolysis in a murine model of myeloma-associated bone loss. J Bone Miner Res. 2009; Epub ahead of print. 174. Burton JD, Ely S, Reddy PK, et al. CD74 is expressed by multiple myeloma cells and is a promising target for therapy. Clin Cancer Res. 2004;10:6606–11. 175. Stein R, Mattes MJ, Cardillo TM, et al. CD74: a new candidate target for the immunotherapy of B-cell neoplasms. Clin Cancer Res. 2007;13:5556s–63s. 176. Stein R, Qu Z, Cardillo TM, et al. Antiproliferative activity of a humanized anti-CD74 monoclonal antibody, hLL1, on B-cell malignancies. Blood. 2004;104:3705–11. 177. Sapra P, Stein R, Pickett J, et al. Anti-CD74 antibody-doxorubicin conjugate, IMMU-110, in a human multiple myeloma xenograft and in monkeys. Clin Cancer Res. 2005;11:5257–64. 178. Chang C-H, Sapra P, Vanama SS, et al. Effective therapy of human lymphoma xenografts with a novel recombinant ribonuclease/anti-CD74 humanized IgG4 antibody immunotoxin. Blood. 2005;106:4308–14. 179. Stein R Smith MR, Chen S, et al. Combining milatuzumab with bortezomib, doxorubicin, or dexamethasone improves responses in multiple myeloma cell lines. Clin Cancer Res. 2009;15:2808–17. 180. Cattan AR, Douglas E. The C.B.17 SCID mouse strain as a model for human disseminated leukemia and myeloma in vivo. Leukemia Res. 1994;18:513–22.


351

181. Mitsiades CS, Mitsiades NS, McMullen CJ, et al. Inhibition of the insulin-like growth factor receptor-1 tyrosine kinase activity as a therapeutic strategy for multiple myeloma, other hematologic malignancies, and solid tumors. Cancer Cell. 2004;5:221–30. 182. Mitsiades CS, Treon SP, Mitsiades NS, et al. TRAIL/Apo2L ligand selectively induces apoptosis and overcomes drug resistance in multiple myeloma: therapeutic applications. Blood. 2001;98:795–904. 183. Herreros B, Sanchez-Aguilera A, Piris MA. Lymphoma microenvironment: culprit or innocent. Leukemia 2008;22:49–58. 184. Psaila B, Lyden D. The metastatic niche: adapting the foreign soil. Nat Rev Cancer. 2009;9:285–92. 185. Drexler HG, Dirks WG, Matsuo y, and MacLeod RAF. False leukemia-lymphoma cell lines: an update on over 500 cell lines. Leukemia. 2003;17:416–26.

Chapter 15

Spontaneous Companion Animal (Pet) Cancers David M. Vail and Douglas H. Thamm

Abstract The inclusion of companion species with naturally occurring tumors provides significant opportunities for optimizing drug development pathways that other model systems cannot provide. Over the past decade, tremendous growth in the field of comparative oncology has occurred, including significant increases in organized consortium infrastructure, availability of investigational reagents and regulatory standardization. These advances are currently being applied to the development of novel cytotoxic, immunologic and biology-based anticancer therapies, innovative drug delivery systems, identification and validation of biological endpoints, noninvasive imaging techniques and surrogate markers critical to the design of Phase I and Phase II human clinical trials. The biotechnology and pharmaceutical industries recognize the utility of the model’s inclusion and several examples exist where they have initiated studies in companion species to assist in drug development. We are clearly at a period in time where the microscope is turned on this model and while currently a theory, the next 5 or 10 years should determine the degree to which information generated through the inclusion of companion animals with cancer is applicable to human cancer drug development. Keywords Companion animals • Dog • Cat • Canine • Feline • Cancer • Comparative oncology

15.1 Introduction At present, large animal translational models of cancer etiology, biology and therapy that serve to bridge more artificial cell and rodent-based models with the human cancer experience are either lacking or underutilized. Spontaneously D.M. Vail (*) Center for Clinical Trials and Research, School of Veterinary Medicine, University of Wisconsin-Madison, 2015 Linden Drive, Madison, WI 53706, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_15, © Springer Science+Business Media, LLC 2011

353

354

D.M. Vail and D.H. Thamm

o ccurring cancers in companion animal (pet) dogs and cats have the potential to serve as important bridging models to enhance both our understanding of cancer and the development of novel therapeutic modalities [1–5]. While rodent models provide important opportunities for investigating specific molecular and genetic pathways, they tend to lack the tumor–host heterogeneity that occurs in people and dogs with spontaneously occurring tumors. Taking a cafeteria approach to tumor modeling, further expanding knowledge gained in rodent models by supplementary investigations in species with spontaneous tumors that better recapitulate the heterogeneity of tumor development and progression makes intuitive sense. Due to the recent elucidation of the canine genome and the commitment to cooperative ventures between several academic, federal and private sector development groups, we stand poised at a phenomenal point of opportunity [6, 7]. The inclusion of companion animals into the drug development pathway with sufficient resources and organization could provide the mechanism to seize these opportunities. There is now a sufficient body of experience documenting companion animal owners’ willingness to allow their pets’ participation in tumor collections for biospecimen repositories and enrollment in trials to evaluate novel diagnostic and therapeutic modalities that may benefit both the animal patient and, eventually, humans with cancer. The organization of veterinarians, physicians and basic cancer researchers around the goal of including companion animal cancers within the mainstream of cancer research has begun. Programs in Comparative Oncology have increasingly been included within designated Comprehensive Cancer Centers, the pharmaceutical industry and regulatory bodies, including the FDA. These organizations have recognized the value and savings of both patient and financial resources offered through the integration of companion animal clinical trials into drug development paths. Safety and activity data generated can be invaluable to inform physicianbased human clinical trials that may follow. In addition, the National Cancer Institute has initiated the Comparative Oncology Program within the Center for Cancer Research designed to promote and include companion (pet) animals with spontaneous cancer in preclinical oncology investigations. This chapter serves to provide the reader with an overview of companion animal cancer, the potential opportunities and disadvantages of the model, the research communities involved, unique trial design issues including regulatory oversight and examples where companion species have been included in the drug development path.

15.2 Overview of Cancer in Companion Animal Species 15.2.1 Cancer Incidence and Availability of Veterinary Medical Care Cancer is the number one cause of morbidity and mortality in the aging pet population, as significant atherosclerotic cardiovascular diseases do not exist in the

15 Spontaneous Companion Animal (Pet) Cancers

355

Table 15.1 Comparative annual incidence rates (per 100,000) for common sites or types of cancer in dogs, cats, and humans [10–13] Site/type Dog Cat Human Oral 20.4 – 10.3 Skin 90.4 34.7 22.6 Connective tissue 35.8 17.0 3.3 Testes 33.9 – 5.5 Melanoma 25 – 21.1 Mammary/breast 198.8 25.4 123.0 Bone 7.9 4.9 0.5 Non-Hodgkin’s lymphoma 25 125 19.5 Leukemia – 35.6 11.9

companion species. Nearly half of all households in the United States include a companion animal. This places approximately 75 million dogs and 90 million cats at risk for developing cancers in the United States [8]. It is estimated that over 1 million new cases of cancer are diagnosed in pet dogs every year in North America. In a necropsy (autopsy) series of 2,000 dogs, 23% of all dogs, regardless of age, and 45% of dogs 10 years of age or older died of cancer [9]. Estimates of age-adjusted overall cancer incidence rates per 100,000 individuals/year at risk are 381 for dogs and 264 for cats; comparable to approximately 476 for humans (National Cancer Institute SEER Program) [10]. Cancer rates for dogs, cats and people based on site are presented in Table 15.1 [10–13]. For several histotypes, the incidence rates are significantly higher than that for humans and their relative abundance increases their model potential. In particular, the high incidence of canine osteosarcoma, soft-tissue sarcoma, non-Hodgkin’s lymphoma and malignant melanoma provide for a significant comparative population. There exists a growing body of residency-trained veterinary oncology specialists (approximately 250 in the United States) populating a similarly expanding number of active academic and private-sector veterinary oncology practices to meet the challenges of cancer management in veterinary species. Despite the increasing application and availability of treatment modalities used for years in humans, including surgical oncology, chemotherapy, immunotherapy and radiation therapy, highly effective “standards of care” do not exist for many tumor types in companion animals. This, in part, has led to a highly motivated, well-informed pet owning public that actively seeks high-quality medical care, including well-designed clinical trials for the animals under their charge.

15.2.2 Genetic and Molecular Basis of Cancer in Pet Dogs It is well documented that the genetic, epigenetic and molecular basis of cancer in people is as varied as it is multifactoral. While much has been learned over the past

356


decades through the use of existing cancer models and investigations in humans with cancer, significant gaps in our knowledge exist. Germane to this chapter, the more we explore the genetic and molecular pathways implicated in cancers of companion species, the more it becomes clear that they share a great deal with human cancers in terms of etiopathogenesis and biology. The publication of the canine genome in 2005, as well as the rapidly accelerating availability of analysis tools (e.g. canine gene microarrays, FISH probes, etc.), has revealed some remarkable similarities between the genetic basis of cancer in dogs and humans [6, 7]. In the canine genome, greater than 1,000 SNPs are identified and microsatellite markers identify 85 different breeds. Canine-specific CGH and gene expression mircoarrays now exist [14, 15]. Comparative expression pathway analysis has documented similarities between human and canine osteosarcoma [1, 16, 17], intracranial tumors [18], hematological tumors [19] and mammary tumors [20, 21]. Several investigations by Matthew Breen’s group and others have found evolutionarily conserved cytogenetic changes in many malignancies of dogs and humans [16–21]. The similarities of many gene families associated with cancer are much more closely related in dogs and humans than are rodents to either species. In fact, cluster analysis of orthologous gene signatures did not segregate human and canine osteosarcoma on the basis of species, illustrating highly similar gene expression patterns [1]. Mutations in many oncogenes and tumor suppressor genes commonly mutated in human cancer, such as p53 [22–31], rB [25], ras [32–34], myc [35, 36], met [16] and bcl-2 [37, 38] have been detected in a variety of canine and feline tumors. Likewise, overexpression of telomerase and matrix metalloproteases have been detected in several canine tumors [39–46]. Additionally, many signal transduction pathways known to be important in human cancers have been documented to play a role in tumor development and progression in companion species with spontaneous tumors. A variety of tyrosine kinase (TK) growth factors and their receptors have been detected in canine and feline tumors, and thus they may serve as pertinent targets for the preclinical development of small molecule inhibitors with relevance to human cancer. The growth factors HGF and IGF-1 and their respective receptors have been detected in a majority of canine osteosarcomas (OSA) [16, 47–49]. HGF was expressed in 17 of 19 clinical samples and c-Met was expressed in all samples evaluated (Fig. 15.1) [47]. Also, the addition of HGF to several human and canine OSA cell lines results in phosphorylation of c-Met, stimulates proliferation, anchorage-independent growth and invasion. In companion animal cats, it has been demonstrated that PDGF induces phosphorylation of PDGFR-b, enhances the proliferation, survival and chemotherapy resistance of several feline vaccine-associated soft tissue sarcoma (VAS) cell lines and that this stimulation can be reversed by the small-molecule TK inhibitor imatinib (Fig. 15.2) [50]. A large number of canine malignant mast cell tumors (MCT) display aberrant expression of c-kit, the receptor for stem cell factor and many canine MCT have mutations in c-kit that confer constitutive activation of the receptor in the absence of ligand binding [51–53]. Indeed, in vitro ligand activation and tumor cell growth can be inhibited in canine MCT cell lines using novel split-tyrosine kinase inhibitors and trials investigating these novel small


357

Fig. 15.1 Expression of HGF and c-Met mRNA in canine OSA tumor samples. All samples expressed c-Met and 17/19 expressed HGF

Fig. 15.2 Immunoprecipitates and immunoblots demonstrating that in several feline soft-tissue sarcoma cell lines, Imatinib blocked autophosphorylation induced by PDGF-BB in a dose-dependent manner, with near complete inhibition at a concentration of 2.5 mM

molecule inhibitors in pet dogs with naturally occurring MCT were instrumental in informing similar c-kit inhibitor trials in human clinical trials for treating c-kit tumors (e.g. GIST) [54–56]. Several studies have documented the presence of EGFR in canine normal and tumor tissues [57–61]. In particular, EGFR has been documented in canine mammary tumors, astrocytoma and lung tumors. We have documented EGFR expression in the majority of spontaneously arising canine bladder transitional cell carcinoma tissues we have evaluated (Fig. 15.3a) and in a transitional cell carcinoma cell line established from a dog (Fig. 15.3b). The documentation of drugable molecular targets in tumors from companion species that are nearly identical to those found in human tumors enhances the model potential. Analogs of most major angiogenic growth factors and their receptors also exist in dogs and cats, and elevations in serum or urine VEGF and/or bFGF have been detected in dogs with hemangiosarcoma (HSA) [62, 63], transitional cell carcinoma of the urinary bladder [64, 65] and osteosarcoma [66]. We have also demonstrated the presence and functional activity of several angiogenic growth factor TKs in canine HSA, a vascular endothelial-derived tumor [67]. HSA is a highly malignant and rapidly fatal tumor and may represent one of the most extreme examples of dysregulated angiogenesis. For example, the majority of canine HSA express receptors for bFGF, VEGF and the angiopoietins and inhibition of the VEGF signaling pathway results in a significant decrease in HSA cell growth in vitro [67].

358


Fig. 15.3 (a) Immunohistochemistry documenting EGFR expression in canine transitional cell carcinoma (Abcam EGFR antibody #ab2430). (b) Immunocytochemistry demonstrating EGFR expression in a cytospin preparation of a transitional cell carcinoma cell line derived from a dog

Similarly, a small molecule VEGF-receptor inhibitor has shown clinical utility in human angiosarcoma [68].

15.3 Potential Opportunities/Advantages of Including Companion Animals with Cancer as Models Several aspects of companion animal disease make pets attractive comparative models and are summarized in Table 15.2. Many have been touched upon in the preceding section on genetic and molecular similarities between the companion species and humans. Since tumors from companion species often possess the same molecular targets, regardless of histology, they can be readily utilized for proof-ofconcept, proof-of-target analysis. Companion species represent a more natural outbred population than laboratory animals and their malignancies develop spontaneously without experimental exposure to known carcinogens, transplantation or artificially induced immunologic or genetic modifications. Thus, they more closely recapitulate the intratumoral heterogeneity and cell-stromal interactions known to be of great importance in human tumor progression. Incidence rates for certain malignancies in companion species (e.g. canine osteosarcoma, non-Hodgkin’s lymphoma) are higher than those observed in people and provide a ready population for inclusion. The relative cost of veterinary-based trials, while often of the highest caliber, is substantially less than physician-based trials. Companion animal trials can also be performed in the pre-IND setting, which can have several advantages (see Sect. 15.6). Veterinary cancer patients tend to be less heavily pretreated, with better performance status at study entry than are most humans entered into phase I trials. This may allow a clearer understanding of the adverse event profiles of the treatment modalities under investigation as well as representing a more naive population with respect to acquired tumor resistance mechanisms and therefore provide a better measure of antitumor activity than would be observed in a more heavily pretreated


359

Table 15.2 Potential advantages of companion species cancer models Advantage Comments Similar genetic etiopathogenesis Greater synteny between dog and human vs. rodent and human Similar gene expression alterations in cancer Similar signal transduction pathways Similar drugable targets identified for TK pathways Similar angiogenic pathways Similar apoptotic pathways Spontaneous tumors in outbreed populations Recaputulates tumor heterogeneity and cell– stromal interactions observed in human tumors High incidence of cancer in companion species Similar to human SEER rates Some histologies more prevalent (e.g. OSA, NHL) Lower relative cost of therapeutic trials GMP quality not always necessary Pre-IND trials possible Less heavily pretreated patient population Higher patient performance scores More accurate AE assessment Less acquired resistance mechanisms Tumor cell kinetics more comparable than Rapid accrual of response assessment rodent systems Larger body size of companion species Similar imaging and treatment (e.g. radiation therapy units) modalities can be applied Greater opportunity for repeated tissue and fluid sampling over time More abundant tissue available for analysis Companion species share our environment Sentinel species Chemoprevention study potential Owner compliance Highly committed clients in the context of lack of financially available standard of care 80–90% necropsy compliance Intact immune system More relevant immunotherapy trial population Similar innate and adaptive immunity cues Similar tumor-associated antigens Minimal disease models (OSA, HSA) Allow for investigation of the metastatic phenotype Antimetastatic therapy models Hepatic enzyme homology more similar than More accurate PK/PD assessment most rodent models More accurate AE assessment Access to both laboratory normal and tumorSpecies-in-kind approach bearing populations More rapid toxicity to activity assessments Organized research efforts Biospecimen consortiums in place Clinical trial consortiums in place TK tyrosine kinase, OSA osteosarcoma, NHL non-Hodgkin’s lymphoma, HSA hemangiosarcoma, AE adverse events

patient. Companion animal cancers are also more comparable to human cancers than are rodent models in terms of size and cell kinetics and they generally progress at a more rapid rate than their human counterparts; therefore, the time course of trials with progression endpoints are of adequate length to allow comparison of

360


response times while short enough to ensure rapid accrual of data. For example, 5-year survival or progression-free temporal measures in human osteosarcoma and non-Hodgkin’s lymphoma populations are roughly equivalent to 1-year temporal measures in companion species [1]. The larger body size of companion species also is advantageous and allows sample collection (e.g. serum, urine, bronchoalveolar lavage, biopsies), surgical intervention, radiation therapy and imaging modalities to be more readily applied than in rodent models. Development of advanced radiation therapy delivery systems, such as Tomotherapy®, have taken advantage of this and used companion species with spontaneous tumors early in their clinical development [69]. Similarly, canine trials of inhalational chemotherapy and immunocytokines benefited from the similarities in body size and helped inform subsequent human trials [70, 71]. Additionally, because companion species require anesthesia for most imaging modalities, this allows for more aggressive and repeated serial collections of tumor and normal tissues during the course of treatment than most human clinical trials allow. By way of example, the antiproliferative effect of a novel cytotoxic chemotherapy for non-Hodgkin’s lymphoma was recently validated by correlating 3¢deoxy-3¢-[18F]fluorothymidine (FLT) PET/CT imaging with Ki-67 immunohistochemistry within the context of a canine clinical trial (Fig. 15.4) [72, 73]. Larger body size (and therefore larger tumor size) also allows for more abundant tumor tissue available for bench-derived molecular analysis, and often tissues can be submitted to several investigators concurrently. Companion species share common environmental exposures with people and therefore can act as sentinels or be included in chemoprevention studies related to exposure. Most companion animal owners are highly motivated and actively seek innovative and promising new therapies to treat their companions’ cancer. Compliance with treatment and recheck visits is exceptional, and necropsy (autopsy) compliance approaches 90%, significantly better than most human clinical trials [2]. Companion species also provide an opportunity to add exogenous substances (e.g. Matrigel plugs) to allow mechanistic assessment. Unlike many laboratory animal models, companion species with naturally occurring tumors have intact immune systems and the key players in both the innate and adaptive immune response are similar between companion animal species and humans. This provides inherent advantages for the inclusion of companion species in investigations of novel immunotherapuetics. Collectively, the species immune systems recognize the same pathogen-associated molecular patterns through similar Toll-like and peptidoglycan recognition receptors [74–76]. Modeling of the innate immune stimulant L-MTP-PE in pet dogs with naturally occurring osteosarcoma, hemangiosarcoma and malignant melanoma exemplifies this approach and helped inform the design of a subsequent clinical trial in children with osteosarcoma [77–80]. Canine and feline tumors also share important tumor-associated antigens with human tumors. For example, the expression of canine analogs of gp100, MART-1 and tyrosinase occurs in a majority of canine melanomas, and immunotherapeutic strategies designed to target canine tumor antigens have shown promise in clinical trials in dogs [81–84]. Canine osteosarcoma and malignant


361

Fig. 15.4 Representative 3¢-deoxy-3¢-18F-fluorothymidine (FLT) positron emission tomography/ computed tomography (PET/CT) of a dog with non-Hodgkin’s lymphoma before (a) and 5 days after (b) cytotoxic chemotherapy. Ki-67 immunohistochemistry of lymphoma tissue from the prescapular lymph node of a dog before (c) and 4 days after (d) cytotoxic chemotherapy in the same dog (600×. Note the significant decrease in FLT uptake in PET/CT scans following therapy correlates with Ki-67 immunoreactivity; both indicating an antiproliferative effect

melanoma, both extremely common tumors in pet dogs, also provide a unique opportunity to investigate the minimal residual disease setting and the metastatic phenotype. Both represent populations where the primary tumor is readily manageable (i.e., amputation, surgical excision) leaving micrometastatic disease in the majority of cases; in both histologies, while metastasis is not clinically evident at diagnosis, most dogs progress to develop gross metastatic disease within 4–6 months [85]. These minimal residual disease populations have been included in several mechanistic investigations of the metastatic phenotype as well as proof of concept trials investigating novel antimetastatic therapies.

362


Importantly, canine hepatic enzyme homology and organ-specific blood flow are more similar to people than are rodents, and this may allow more accurate assessment of pharmacokinetic and pharmacodynamic properties of novel therapeutics under development. For example, pharmacokinetic, safety and activity assessments for a novel chemotherapeutic prodrug requiring intracellular enzymatic activation was not possible in rodents as they, unlike humans or dogs, have high plasma levels of carboxyesterase which rapidly metabolized the drug in the extracellular compartment, effectively precluding rodent preclinical models [72, 86]. In particular, important information regarding the characterization and assessment of adverse event profiles of new agents in the development pathway may be more realistically assessed in companion species with similar metabolic pathways. Dose-limiting toxicities in tumor-bearing pet dogs treated off-label with chemotherapies commonly used in people are uniformly identical in both species [87]. Similarly, specific antitumor activity of chemotherapeutics commonly used in people have shown remarkable similarities in companion species tumors and appear more predictive than mouse xenograft models of activity [88]. For example, the most active agents for non-Hodgkin’s lymphoma (cyclophosphamide, doxorubicin, vincristine, prednisone) and osteosarcoma (platinum analogues) in people are the same as those showing the greatest activity in dogs [85, 89]. Likewise, drugs known to be relatively inactive for NHL in people (e.g. gemcitabine and cisplatin) are similarly inactive in dogs with NHL [87, 89, 90]. Finally, the fact that the dog is the only species with significant access to both a laboratory-normal population and a highly varied spontaneous tumor-bearing population is, in itself, a major advantage as it allows the evaluation of both safety and activity of novel therapies in the same species. Taking a “species in kind” approach, significant PK/PD, dosing, biomarker validation and adverse event issues can be significantly explored in laboratory dogs prior to moving promising agents into the “veterinary clinic” for assessment of activity and further characterization in the more relevant naturally occurring companion animal population. Data generated in laboratory dogs allow superior design of clinical trials in companion species without having to “best-guess” starting doses, better anticipation and measurement of adverse events and more rational design of PK/PD and biomarker assessment. Conversely, if an interesting finding requiring further characterization occurs during a study in tumor-bearing pet dogs, such as an unexpected adverse event or a potential biomarker is identified, more focused study can then be performed in purpose-bred laboratory dogs.

15.4 Caveats to Inclusion of Pet Dogs with Cancer as Models No one model of cancer is ideal, nor should it be expected to answer all the myriad of questions remaining to be answered in cancer biology and therapeutic development. The utility of including companion species in the global effort to understand and manage cancer in people is still theoretical and several caveats exist.


363

The greatest potential strength for inclusion of companion species with spontaneous tumors in cancer drug development, the heterogeneity that occurs in the population and the tumors under study, is also one of its weaknesses. Rodent modeling has been the workhorse of preclinical cancer research owing in part to the ability to regulate or restrict the genetic and molecular diversity of the study system and specifically control variables in a laboratory setting. This allows characterization of very specific genetic and epigenetic alterations to be performed, a process that is important and should ensure continued utility of rodent systems. Tumor-bearing companion species are clearly a different preclinical population compared to research animals; however, study entry characteristics (i.e. breed, sex, age, histology and presence of drugable target) can be restricted and ultimately treatment strategies that are further along in the development pathway could benefit from investigations in a model system that more closely recapitulates the tumor heterogeneity and cell–stromal interactions known to be of great importance in human tumor progression. While the incidence of cancer is high in companion species, the prevalence of common histologies is not similar to humans. The most common histologies in companion species include sarcomas and lymphoid malignancies while more common histologies affecting humans (e.g. breast, prostate, gastrointestinal and lung) are less common and require a multicenter approach and more time for accrual [1, 10–13]. That being said, often target trumps histology with respect to comparative importance as evidenced by the c-kit tyrosine kinase pathway which is often mutated in the common canine mast cell tumor (a histology that is exceedingly rare in people) and in human GIST tumors [51–56]. Despite the histological dissimilarity of these two tumors, significant characterization of the pathway, including development of currently available small molecule c-kit inhibitors in physician-based oncology were initially modeled through inclusion of pet dogs with spontaneous mast cell tumors. The availability and development of commercially available biological reagents and technology platforms specific for companion species have lagged behind those of rodent and human tissues. This weakness is diminishing now and the canine genome is known and commercially available products such as Affymetrix® canine gene chips, canine-based ELISA kits and canine tumor-specific antibodies are becoming available. The recent presence of several cooperative consortia (see Sect. 15.5) committed to the development and dissemination of canine- and feline-based reagents, cell lines, tumor tissue arrays and genetic and molecular probes has also served to lessen this caveat. In comparison to rodent modeling, companion species trials tend to have longer timelines and higher costs, albeit less so than similar human trials. While drug investigations in larger species such as dogs require larger quantities and ramping up production of the agent under investigation, this is tempered to a degree as GMP quality drug is not always necessary for companion species trials as long as the composition of matter is sterile, endotoxin free and of high quality and purity. Finally, there has been an historical aversion by biotechnology and pharmaceutical companies to “rock” the traditional development boat [1]. In other words,

364


guidelines for the control and reporting of data accrued in companion species trials including adverse event reporting have not been uniformly stated in the past and regulatory oversight and reporting guidelines have not been established for trials run in the pre- and post-IND setting. This has led to some confusion as well as a fear of having human trials stopped or retarded if an unforeseen adverse event is observed in parallel companion species trials. Recently, the interested parties, including the FDA have met to resolve many of these issues (see Sect. 15.6) [91].

15.5 The Vested Communities The concept of inclusion of companion species in drug development is not new and researchers have been involved in such investigations for several decades, dating back to the 1960s with work performed by Rainer Storb’s group on bone marrow transplantation techniques which included pet dogs with non-Hodgkin’s lymphoma [92]. However, such investigations have historically occurred in an isolated environment by a few individuals. In the last 10 years, however, the comparative oncology community has greatly expanded and become organized through several consortia, many through the efforts of the Comparative Oncology Program (COP, http://ccr. cancer.gov/resources/cop/default.asp) within the Center for Cancer Research at the National Cancer Institute (NCI-CCR). The COP, under the leadership of Chand Khanna, has initiated several programs aimed at better integration of current efforts as well as the development of bioresource and technology platforms aimed at comparative research. These include a biospecimen repository, an antibody validation project, gene microarray, tissue microarray and proteomics programs. An offshoot of the COP is the Canine Comparative Oncology and Genomics Consortium (CCOGC, http://www.ccogc.net/), a 501(c)3 not for profit whose primary objective is to facilitate strategic partnerships and collaborations across a diversity of disciplines, focused on the problem of cancer. Priorities of the CCOGC include advocacy for the field of Comparative Oncology, the development of mechanisms to share reagents and resources in the community, and to utilize the release of the Canine Genomics Project to characterize cancers in companion dogs using modern descriptors. The CCOGC leadership has determined that the most essential resource needed to make progress in the companion species comparative oncology field is the development of a well-described repository of tissues from tumor bearing dogs. This recognition is similar to that made in the development of a National Biospecimen Network that collects and distributes tissues from humans with cancer. A repository of well-described tissues provides opportunities not currently available to individual and institutional investigators interested in developing effective new treatments and prevention strategies for cancer. Through funding from Pfizer Inc., the American Kennel Club–Canine Health Foundation and the Morris Animal Foundation, the Pfizer– CCOGC Biospecimen Repostitory has become a reality and is part of the NCI Frederick Central Repository Services. The bioinformatics platform for the repository includes a relational database that connects clinical information on samples entered with a front and back end retrieval system. Biological data derived from samples in


365

the Biospecimen Repository are uploaded into the bioinformatics database and become part of the progressive value of the Repository. Another program to come out of the COP efforts at NCI-CCR is the Comparative Oncology Trials Consortium (COTC, http://ccr.cancer.gov/resources/cop/COTC. asp). The COTC is an active network of 18 academic comparative oncology centers (Fig. 15.5), centrally managed by the COP, that functions to design and conduct clinical trials in pet dogs with cancer to assess novel therapies. One key element of the COTC is a mechanism by which pharmaceutical companies can execute a single contract or memorandum of understanding with several academic trial centers for the conduct of investigative trials. Additionally, COTC trials utilize the same webbased bioinformatics backbone that exists for human trials and comply with NCI technology initiatives, such as caBIG. The overall goal of this effort is to answer biological questions geared to inform the development path of these agents for future use in human cancer patients. Trials conducted by the COTC are pharmacokinetically and pharmacodynamically intensive with the product of this work directly integrated into the design of current human Phase I and II clinical trials (Fig. 15.6). The inaugural pre-clinical COTC trial was recently concluded. This

Fig. 15.5 Comparative Oncology Trials Consortium (COTC) contributing institutions (Auburn University, Colorado State University, Cornell University, Michigan State University, North Carolina State University, Purdue University, Texas A&M University, The Ohio State University, Tufts University, University of California, Davis, University of Florida, University of Georgia, University of Illinois, University of Minnesota, University Of Missouri, University Of Pennsylvania, University Of Tennessee, University of Wisconsin). The COTC is an active network of 18 academic comparative oncology centers centrally managed by the Comparative Oncology Program within the NCI-Center for Cancer Research that functions to design and conduct clinical trials in pet dogs with cancer to assess novel therapies

366


Fig. 15.6 An integrated approach to the inclusion of companion animal species in the drug development pathway. Such studies can be performed prior to first-in-man trials (e.g. phase I trials) as well as in parallel to ongoing human clinical trials in order to further provide data necessary for more advanced trials to follow (e.g. phase II–III). Reproduced by permission from Macmillan Publishers Ltd: Nature Rev Cancer [1], copyright 2008

study demonstrated the utility of the COTC infrastructure to inform the development of new cancer drugs within large animal naturally occurring cancer models; specifically, the evaluation of a targeted AAV-phage vector delivering tumor necrosis factor (RGD-A-TNF) to aV integrins on tumor endothelium provided valuable and necessary data to complete the design of first-in-man studies [93].

15.6 Study Design Issues Specific to Companion Species Trials Companion species trials are best suited for proof-of-concept or proof-of-target investigations with an eye to informing future human clinical trials. Such studies can be performed prior to first-in-man trials (e.g. phase I trials) as well as in parallel to ongoing human clinical trials in order to further provide data necessary for more advanced trials to follow (e.g. phase II–III) as illustrated in Fig. 15.6 [1]. Ultimately such trials could help prioritize which agents continue on in the drug development pathway. Suggested guidelines for the performance of translational trials involving


367

veterinary patients have come, in large part, out of a meeting entitled “Translation of new cancer treatments from canine to human cancer patients”, sponsored by the National Cancer Institute in Bethesda, MD (June 2008) which was convened to discuss the potential value, opportunity, risks and rewards of an integrated and comparative drug and device development path for new cancer therapeutics that includes naturally occurring cancers in pet animals [91]. A summary of this meeting and subsequent discussion was recently published to provide clarity on the conduct of these studies and will contribute to the evaluation of this novel drug development opportunity [91]. The reader is referred to this review for a more indepth presentation of the suggested guidelines for companion species trials. Ethical considerations: Any trial to be undertaken in companion species must have the humane care of the companion patient as a priority. The appropriate use of an accredited institutional animal care and use committee (IACUC) is required and a well thought out informed consent and consenting process should be in place [94]. The scientific and translational motivation of the study must be balanced against the over-riding mandate for animal care. A data safety management function should be included similar to data safety management boards used in human clinical trials to provide protocol design oversight, evaluate progress and protect the data generated. Trial conduct: Clinical trials involving companion species should be designed to answer specific questions of translational interest with a high likelihood of informing future human trials. A priority should be placed on questions that cannot reasonably be fully answered in the context of conventional preclinical models or early human clinical trials. Specific endpoints should be identified and reviewed by persons with considerable background in both veterinary- and physician-based disciplines and trial design such that important preclinical questions can be effectively posed and answered. While GMP quality investigational agents may not be necessary, minimum standards of composition should be determined with respect to purity, sterility and quality. Similarly, trials should follow the spirit of Good Clinical Practice (GCP), which has been described for veterinary species through the VICH GCP and are summarized in Table 15.3 [95]. Table 15.3 Summary of attributes of the VICH GCP procedures and regulations relevant to the conduct of companion species trials Adequately developed study protocol with consent forms and consenting process Systems in place to manage protocol changes and modifications Adequate training of qualified participating investigators on study conduct including relevant standard operating procedures (SOPs) Facilities and institutional inspections necessary for study conduct Contemporaneous entry of data using paper based and/or web-based mechanisms Safety management system that includes monitoring and reporting of AEs and serious AEs to the IACUC, study coordinators, and if applicable, regulatory agencies System to verify the conduct and reporting of data within the study

368


Adverse event reporting: The reporting of adverse events (AE) encountered during trials of companion species has been of considerable concern among the biomedical and pharmaceutical community with regard to their impact on future or concurrent human clinical trials. A clear understanding of the standards for reporting AE data to regulatory authorities is needed and currently represents a concern within the field. While the majority of preclinical AE information comes from purpose-bred research animals, their assessment in tumor-bearing companion dogs should provide additionally valuable drug safety information. On the other hand, based on experience with over 30 human cytotoxic chemotherapeutic agents commonly used off-label to treat pet dogs, no adverse events have been identified in tumorbearing dogs that were not seen in purpose-bred research dogs (http://ccr.cancer. gov/resources/cop/). In the rare and unprecedented circumstance in which unexpected adverse events are defined in the conduct of a tumor-bearing dog study, it is reasonable that additional studies focused on any unexpected AEs should be conducted in either purpose-bred or tumor-bearing dogs before IND filing. Studies in companion species performed prior to Investigational New Drug (IND) applications should not require contemporaneous reporting of AEs to the regulatory authorities, rather they should be maintained as part of the legacy of the agent under development and included in a final study report as part of any IND application package should the agent progress through development. In companion species studies performed with novel human cancer agents in the post-IND setting, regulations already exist regarding AE reporting and are provided by IND Section 312.32 IND Safety Report [96]. Briefly, all serious and unexpected AEs must be reported within 15 days of development, whereas AEs that are either not serious or are expected, based on the protocol and informed consent, do not require expedited reporting. Of course, post hoc reporting of all AEs with attribution is provided with IND updates at the completion of the study. Because of the contemporaneous reporting requirement for post-IND agents, there exists some risk as to the impact AEs in companion species will have on ongoing human trials with the same agent. As discussed earlier, the risk of observing a unique and unexpected AE in tumor-bearing dogs that was not observed in purpose-bred laboratory dog toxicity trials is small [91]. Our current experience suggests that AE observations in companion species have not stopped ongoing human trials; rather, they have led to modifications in human trial design such as changing eligibility and exclusion criteria, additions to monitoring strategies or changes to informed consent.

15.7 Conclusions In summary, the inclusion of companion species with naturally occurring tumors provides significant opportunities for optimizing drug development pathways that other model systems cannot provide. Over the past decade, tremendous growth in the field of comparative oncology has occurred, including significant increases in


369

organized consortium infrastructure, availability of investigational reagents and regulatory standardization. These advances are currently being applied to the development of novel cytotoxic, immunologic and biology-based anticancer therapies, innovative drug delivery systems, identification and validation of biological endpoints, noninvasive imaging techniques and surrogate markers critical to the design of Phase I and Phase II human clinical trials. The biotechnology and pharmaceutical industries recognize the utility of the model’s inclusion and several examples exist where they have initiated studies in companion species to assist in drug development. We are clearly at a period in time where the microscope is turned on this model and while currently a theory, the next 5 or 10 years should determine the degree to which information generated through the inclusion of companion animals with cancer is applicable to human cancer drug development.

References 1. Khanna C, Paoloni M. Translation of new cancer treatments from pet dogs to humans. Nat Rev Cancer. 2008;8:7–16. 2. Vail DM, Thamm DH. Spontaneously occurring tumors in companion animals as models for drug development. In: Teisher BA, ed. Anticancer drug development guide. 2nd ed. Totowa: Humana Press, 2004;259–86. 3. Argyle DJ. The benefits of comparative medicine – a hundred years to come to our senses. Vet J. 2005;170: 147–8. 4. Porrello A, Cardelli P, Spugnini EP. Oncology of companion animals as a model for humans. an overview of tumor histotypes. J Exp Clin Cancer Res. 2006;25: 97–105. 5. Waters DJ, Wildasin K. Cancer clues from pet dogs. Sci Am. 2006;295:94–101. 6. Lindblad-Toh K, Wade CM, Mikkelsen TS, et al. Genome sequence, comparative analysis and haplotype structure of the domestic dog. Nature. 2005;438:803–19. 7. Ostrander EA, Wayne RK. The canine genome. Genome Res 2005;15:1706–16. 8. American Veterinary Medical Association. Results of the AVMA survey on companion animal ownership in US pet-owning households. J Am Vet Med Assoc. 2001;221:1572–5. 9. Bronson RT. Variation in age at death of dogs of different sexes and breeds. Am J Vet Res. 1982; 43:2057–9. 10. National Cancer Institute Surveillance Epidemiology and End Results, 2002–2006. (Accessed May 26, 2009 at http://seer.cancer.gov/statfacts/html/all.html). 11. Dorn CR, Taylor DON, Schneider R, Hibbard HH, Klauber MR. Survey of animal neoplasms in Alameda and Contra Costa Counties. Cancer morbidity in dogs and cats from Alameda County. J Natl Cancer Inst. 1968; 40:307–18. 12. Dorn CR. Epidemiology of canine and feline tumors. Comp Cont Ed Pract Vet. 1976; 12:307–12. 13. Priester WA, McKay FW. The occurrence of tumors in domestic animals. Natl Cancer Inst Monogr No. 54. Washington DC: US Government Printing Office, 1980:152. 14. Thomas R, Duke SE, Bloom SK, et al. A cytogenetically characterized, genome anchored 10-Mb BAC set and CGH array for the domestic dog. J Hered. 2007;98:474–84. 15. Holzwarth JA, Middleton RP, Roberts M, et al. The development of a high-density canine microarray. J Hered. 2005;96:817–20. 16. De Maria R, Miretti S, Lussich S, et al. MET oncogene activation qualifies spontaneous canine osteosarcoma as a suitable pre-clinical model of human osteosarcoma. J Pathol. 2009;218:399–408.

370


17. Tomas R, Wang HJ, Tsai P-C, et al. Influence of genetic background on tumor karyotypes: Evidence for breed-associated cytogenetic aberrations in canine appendicular osteosarcoma. Chromosome Res. 2009;17:365–77. 18. Thomas R, Duke SE, Wang HJ, et al. “Putting our heads together”: insights into genomic conservation between human and canine intracranial tumors. J Neurooncol. 2009;94:333–49. 19. Breen M, Modiano JF. Evolutionarily conserved cytogenetic changes in hematological malignancies of dogs and humans – man and his best friend share more than companionship. Chromosome Res. 2008;16:145–54. 20. Rao NAS, van Wolferen ME, van den Ham R, et al. cDNA microarray profiles of canine mammary tumour cell lines revel deregulated pathways pertaining to their phenotype. Anim Genet. 2008;39:333–45. 21. Uva P, Aurisicchio L, Watters J, et al. Comparative expression pathway analysis of human and canine mammary tumors. BMC Genom. 2009;10:135. 22. Roels S, Tilmant K, Ducatelle R. p53 expression and apoptosis in melanomas of dogs and cats. Res Vet Sci. 2001;70:19–25. 23. Korkolopoulou P, Oates J, Kittos C, Crocker J. p53, c-myc, p62 and proliferating cell nuclear antigen (PCNA) expression in non-Hodgkin’s lymphomas. J Clin Pathol. 1994; 47:9–14. 24. Levine RA, Fleischli MA. Inactivation of p53 and retinoblastoma family pathways in canine osteosarcoma cell lines. Vet Pathol. 2000; 37:54–61. 25. Mendoza S, Konishi T, Dernell WS, Withrow SJ, Miller CW. Status of the p53, Rb and MDM2 genes in canine osteosarcoma. Anticancer Res. 1998; 18:4449–53. 26. Nambiar PR, Jackson ML, Ellis JA, et al. Immunohistochemical detection of tumor suppressor gene p53 protein in feline injection site-associated sarcomas. Vet Pathol. 2001; 38:236–8. 27. Nasir L, Rutteman GR, Reid SW, Schulze C, Argyle DJ. Analysis of p53 mutational events and MDM2 amplification in canine soft-tissue sarcomas. Cancer Lett. 2001; 174:83–9. 28. Rungsipipat A, Tateyama S, Yamaguchi R, et al. Immunohistochemical analysis of c-yes and c-erbB-2 oncogene products and p53 tumor suppressor protein in canine mammary tumors. J Vet Med Sci. 1999; 61:27–32. 29. Veldhoen N, Stewart J, Brown R, Milner J. Mutations of the p53 gene in canine lymphoma and evidence for germ line p53 mutations in the dog. Oncogene. 1998; 16:249–55. 30. Veldhoen N, Watterson J, Brash M, Milner J. Identification of tumour-associated and germ line p53 mutations in canine mammary cancer. Br J Cancer. 1999; 81:409–15. 31. Wakui S, Muto T, Yokoo K, et al. Prognostic status of p53 gene mutation in canine mammary carcinoma. 2001;21:611–6. 32. Griffey SM, Kraegel SA, Madewell BR. Rapid detection of K-ras gene mutations in canine lung cancer using single-strand conformational polymorphism analysis. Carcinogenesis. 1998;19:959–63. 33. Merryman JI, Buckles EL, Bowers G, Neilsen NR. Overexpression of c-Ras in hyperplasia and adenomas of the feline thyroid gland: an immunohistochemical analysis of 34 cases. Vet Pathol. 1999;36:117–24. 34. Watzinger F, Mayr B, Gamerith R, Vetter C, Lion T. Comparative analysis of ras protooncogene mutations in selected mammalian tumors. Mol Carcinog 2001;30:190–8. 35. Ahern TE, Bird RC, Bird AE, Wolfe LG. Overexpression of c-erbB-2 and c-myc but not c-ras, in canine melanoma cell lines, is associated with metastatic potential in nude mice. Anticancer Res. 1993;13:1365–71. 36. Korkolopoulou P, Oates J, Kittos C, Crocker J. p53, c-myc, p62 and proliferating cell nuclear antigen (PCNA) expression in non-Hodgkin’s lymphomas. J Clin Pathol. 1994;47:9–14. 37. Madewell BR, Gandour-Edwards R, Edwards BF, Walls JE, Griffey SM. Topographic distribution of bcl-2 protein in feline tissues in health and neoplasia. Vet Pathol. 1999;36:565–73. 38. Madewell BR, Gandour-Edwards R, Edwards BF, Matthews KR, Griffey SM. Bax/bcl-2: cellular modulator of apoptosis in feline skin and basal cell tumours. J Comp Pathol. 2001;124:115–21.


371

39. Fukanoshi Y, Nakayama H, Uetsuka K, et al. Cellular proliferative and telomerase activity in canine mammary gland tumors. Vet Pathol. 2000;37:177–83. 40. Nasir L, Devlin P, Mckevitt T, Rutteman G, Argyle DJ. Telomere lengths and telomerase activity in dog tissues: a potential model system to study human telomere and telomerase biology. Neoplasia. 2001;3:351–9. 41. Yazawa M, Okuda M, Setoguchi A, et al. Measurement of telomerase activity in dog tumors. J Vet Med Sci. 1999;61:1125–9. 42. Yazawa M, Okuda M, Setoguchi A, et al. Telomere length and telomerase activity in canine mammary gland tumors. Am J Vet Res. 2001;62:1539–43. 43. Cakir Y, Hahn KA. Direct action by doxycycline against canine osteosarcoma cell proliferation and collagenase (MMP-1) activity in vitro. In Vivo. 1999;13:327–31. 44. Lana SE, Ogilvie GK, Hansen RA, et al. Identification of matrix metalloproteinases in canine neoplastic tissue. Am J Vet Res. 2000;61:111–4. 45. Leibman NF, Lana SE, Hansen RA, et al. Identification of matrix metalloproteinases in canine cutaneous mast cell tumors. J Vet Intern Med. 2000;14:583–6. 46. Yokota H, Kumata T, Taketaba S, et al. High expression of 92 kDa type IV collagenase (matrix metalloproteinase-9) in canine mammary adenocarcinoma. Biochim Biophys Acta. 2001;1568:7–12. 47. MacEwen EG, Kutzke J, Carew J, et al. C-Met tyrosine kinase receptor expression and function in human and canine osteosarcoma cells. Clin Exp Metastasis. 2003;20:421–30. 48. MacEwen EG, Pastor J, Kutzke J, et al. IGF-1 receptor expression and function contribute to the malignant phenotype in human and canine osteosarcoma. J Cell Biochem. 2004;92:77–91. 49. Ferracini R, Angelini P, Cagliero E, et al. MET oncogene aberrant expression in canine osteosarcoma. J Orthop Res. 2000;18:253–6. 50. Katayama R, Huelsmeyer MK, Marr AK, et al. Imatinib mesylate inhibits platelet-derived growth factor activity and increases chemosensitivity in feline vaccine-associated sarcoma. Cancer Chem Pharm. 2004;54:25–33. 51. London CA, Kisseberth WC, Galli SJ, Geissler EN, Helfand SC. Expression of stem cell factor receptor (c-kit) by the malignant mast cells from spontaneous canine mast cell tumors. J Comp Pathol. 1996;115:399–414. 52. Ma Y, Longley BJ, Wang X, Blount JL, Langley K, Caughey GH. Clustering of activating mutations in c-KIT’s juxtamembrane coding region in canine mast cell neoplasms. J Invest Dermatol. 1999;112:165–170. 53. Reguera MJ, Rabanal RM, Puigdemont A, Ferrer L. Canine mast cell tumors express stem cell factor receptor. Am J Dermatopathol. 2000; 22:49–54. 54. Pryer NK, Lee LB, Zadovaskaya R, et al. Proof of target for SU11654: inhibition of KIT phophorylation in canine mast cell tumors. Clin Cancer Res. 2003;15:5729–34. 55. London CA, Hannah AL, Zadovoskaya R, et al. Phase I dose – escalating study of SU11654, a small molecule receptor tyrosine kinase inhibitor, in dogs with spontaneous malignancies. Clin Cancer Res. 2003;9:2755–68. 56. London CA, Malpas PG, Wood-Follis SL, et al. Multicenter, placebo-controlled, doubleblind, randomized study of oral palladia (SU11654), a receptor tyrosine kinase inhibitor, for the treatment of dogs with recurrent mast cell tumors. Clin Cancer Res. 2009:15:3856–65. 57. Nerurkar VR, Seshadri R, Mulherkar R, et al. Receptors for epidermal growth factor and estradiol in canine mammary tumors. Int J Cancer. 1987;40:230–2. 58. Rutteman GR, Foekens JA, Protengen H, et al. Expression of epidermal growth factor receptor (EGFR) in non-affected and tumorous mammary tissue of female dogs. Breast Cancer Res Treat. 1994;30:139–46. 59. Gillett NA, Stegelmeier BL, Kelly G, Haley PJ, Hahn FF. Expression of epidermal growth factor receptor in plutonium-239-induced lung neoplasms in dogs. Vet Pathol. 1992;29:46–52. 60. Foster DJ, Yan X, Bellotto DJ, et al. Expression of epidermal growth factor and surfactant proteins during postnatal and compensatory lung growth. Am J Physiol Lung Cell Mol Physiol. 2002;283:L981–990.

372


61. Berens ME, Bjotvedt G, Levesque DC, et al. Tumorigenic, invasive, karyotypic, and immunocytochemical characteristics of clonal cell lines derived from a spontaneous canine anaplastic astrocytoma. In Vitro Cell Dev Biol. 1993;29A:310–8. 62. Clifford CA, Mackin AJ, Henry CJ. Treatment of canine hemangiosarcoma: 2000 and beyond. J Vet Intern Med. 2000;14:479–85. 63. Clifford CA, Hughes D, Beal MW, et al. Plasma vascular endothelial growth factor concentrations in healthy dogs and dogs with hemangiosarcoma. J Vet Intern Med. 2001;15:131–5. 64. Mohammed SI, Bennett PF, Craig BA, et al. Effects of the cyclooxygenase inhibitor, piroxicam, on tumor response, apoptosis, and angiogenesis in a canine model of human invasive urinary bladder cancer. Cancer Res. 2002;62:356–8. 65. Allen DK, Waters DJ, Knapp DJ, Kuczek T. High urine concentrations of basic fibroblast growth factor in dogs with bladder cancer. J Vet Intern Med. 1996;10:231–4. 66. Thamm DH, O’Brien MG, Vail DM. Serum vascular endothelial growth factor concentrations and postsurgical outcome in dogs with osteosarcoma. Vet Comp Oncol. 2008;6:126–32. 67. Thamm DH, Dickerson EB, Akhtar N, et al. Biological and molecular characterization of a canine hemangiosarcoma-derived cell line. Res Vet Sci. 2006;81:76–86. 68. Maki RG, D’Adamo DR, Keohan ML, et al. Phase II study of sorafenib in patients with metastatic or recurrent sarcomas. J Clin Oncol. 2009 May 18 (Epub ahead of print). 69. Forrest LJ, Rockwell T, Ruchala K, Turek M, Kapatoes J, Jaradat H, Hui S, Balog J, Vail DM, Mehta MP. The utility of megavoltage computed tomography images from a helical tomotherapy system for setup verification purposes. Int J Radiat Oncol Biol Phys. 2004;60:1639–44. 70. Khanna C, Vail DM. Targeting the lung: preclinical and comparative evaluation of anticancer aerosols in dogs with naturally occurring cancers. Current Cancer Drug Targets. 2003;3:265–73. 71. Hershey AE, Kurzman ID, Bohling C, et al. Inhalation chemotherapy for macroscopic primary or metastatic lung tumors: Proof of principle in a companion animal model. Clin Canc Res. 1999;5:2653–9. 72. Vail DM, Thamm DH, Tumas DB, et al. Assessment of GS-9219 in a pet dog model of nonHodgkin’s lymphoma. Clin Cancer Res. 2009;15:3503–10. 73. Lawrence J, Vanderhoek M, Barbee D, et al. Utility of 3¢-deoxy-3¢-[18F]Fluorothymidinepositron emission tomography/ computed tomography for evaluating response to cytotoxic chemotherapy in dogs with non-Hodgkin’s lymphoma. Vet Rad Ultrasound. 2009;50(6):660–8. 74. Burgener IA, Jungi TW. Antibodies specific for human or murine Toll-like receptors detect canine leukocytes by flow cytometry. Vet Immunol Immunopathol. 2008;124:184–91. 75. Turin L, Riva F. Toll-like receptor family in domestic species. Crit Rev Immunol. 2008;28:513–38. 76. Swerdlow MP, Kennedy DR, Kennedy JS, et al. Expression and function of TLR2, TLR4, and Nod2 in primary canine colonic epithelial cells. Vet Immunol Immunopathol. 2006;114:313–9. 77. Kurzman ID, MacEwen EG, Vail DM, et al. Adjuvant therapy for osteosarcoma in dogs: results of randomized clinical trials using combined liposome encapsulated muramyl tripeptide and cisplatin. Clin Canc Res. 1995;1:1595–601. 78. MacEwen EG, Kurzman ID, Vail DM, et al. Adjuvant therapy for melanoma in dogs: Results of randomized clinical trials using surgery, liposome-encapsulated muramyl tripeptide and granulocyte macrophage colony-stimulating factor. Clin Cancer Res. 1999; 5:4249–58. 79. Vail DM, MacEwen EG, Kurzman ID, et al. Liposome-encapsulated muramyl tripeptide phosphatidylethanolamine (L-MTP-PE) immunotherapy for splenic hemangiosarcoma in the dog: A randomized multi-institutional clinical trial. Clin Cancer Res. 1995;1:1165–70. 80. Myers PA, Schwarz CL, Krailo MD, et al. Osteosarcoma: the addition of muramyl tripeptide to chemotherapy improves overall survival – a report from the children’s oncology group. J Clin Oncol 2008;26:633–8. 81. Alexander AN, Huelsmeyer MK, Mitzey M, et al. Development of an allogeneic whole-cell tumor vaccine expressing xenogeneic gp100 and its implementation in a Phase II clinical trial in canine patients with malignant melanoma. Cancer Immunol Immunother. 2006;55:433–42.


373

82. Koenig A, Wojcieszyn J, Weeks BR, Modiano JF. Expression of S100a, vimentin, NSE, and melan A/MART-1 in seven canine melanoma cell lines and twenty-nine retrospective cases of canine melanoma. Vet Pathol. 2001;38:427–35. 83. Liao JC, Gregor P, Wolchok JD, et al. Vaccination with human tyrosinase DNA induces antibody responses in dogs with advanced melanoma. Cancer Immunol. 2006;21:8. 84. Bergman PJ, Camps-Palau MA, McKnight JA, et al. Development of a xenogeneic DNA vaccine program for canine malignant melanoma at the animal medical center. Vaccine. 2006;24:4582–5. 85. Dernell WS, Ehrhart N, Straw RC, Vail DM. Tumors of the skeletal system. In: Withrow SW, Vail DM, editors. Small animal clinical oncology, 4th ed.. St. Louis: Saunders/Elsevier. 2007. p 540–82. 86. Minagawa T, Kohno Y, Suwa T, Tsuji A. Species differences in hydrolysis of isocarbacyclin methyl ester (TEI-9090) by blood esterases. Biochem Pharmacol. 1995;49:1361–5. 87. Chun R, Garrett LD, Vail DM. Cancer chemotherapy. In: Withrow SW, Vail DM, editors. Small animal clinical oncology, 4th ed.. St. Louis: Saunders/Elsevier. 2007. p 163–92. 88. Johnson JI, Decker S, Zaharevitz D, et al. Relationships between drug activity in NCI preclinical in vitro and in vivo models and early clinical trials. Br J Cancer. 2001;84:1424–31. 89. Vail DM, Young KM. Canine lymphoma and lymphoid leukemias. In: Withrow SW, Vail DM, editors. Small animal clinical oncology, 4th ed.. St. Louis: Saunders/Elsevier. 2007. p 699–732. 90. Turner AI, Hahn KA, Rusk A, et al. Single agent gemcitabine chemotherapy in dogs with spontaneously occurring lymphoma. J Vet Intern Med. 2006;20:1384–8. 91. Khanna C, London CA, Vail DM, et al. Guiding the optimal translation of new cancer treatments from canine to human cancer patients. Clin Cancer Res. 2009;15:5671–7. 92. Epstein RB, Graham TC, Storb R, Thomas ED. Studies of marrow transplantation, chemotherapy and cross-circulation in canine lymphosarcoma. Blood. 1971;37:349–59. 93. Paoloni MC, Tandle A, Mazcko C, et al. Launching a novel preclinical infrastructure: comparative oncology trials consortium directed therapeutic targeting of TNF-alpha to cancer vasculature. PLoS One. 2009;4:e4972. doi:10.1371/journal.pone.0004972. 94. Welfare OoLA. Public health service policy on humane care and use of laboratory animals. 2002. 95. Committee VS. Good clinical practice VICH GL9 final guidance. 2001. 96. Services FaDADoHaH. Section 312.32 IND safety reports. 2008;5.

Part VI

Genetically Engineered Mouse Models of Cancer

Chapter 16

Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma Aram F. Hezel and Nabeel Bardeesy

Abstract Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer death in the United States. While the incidence of PDAC is low compared to that of the more common malignancies such as lung, breast, prostate, and colon cancer, it is the disease’s lethality – nearly all patients who develop the disease die from it – which makes it a significant health menace. PDAC is characterized by spread to other organs early in the course of disease and a general resistance to chemotherapy. Insight into the molecular pathogenesis of PDAC has come from analysis of the pathologic precursor lesions found adjacent to cancers in resected cases. The identification of KRAS, INK4A, P53, and SMAD4 mutations in these precursor lesions and advanced PDAC has provided the genetic framework for recent efforts to model the disease. The current models offer valuable tools for the study of PDAC and have been used to investigate critical signaling networks, stromal epithelial interactions, and potential cells of origin and early stages of disease [Hezel et al. Genes Dev. 20:1218–49, 2006]. In this chapter we review the development of genetically engineered mouse models (GEMMs) of PDAC and discuss how such models have given insight into disease biology and provided a foundation for preclinical testing. We also discuss emerging improvements in the PDAC models and how these will impact both basic and therapeutic research in the future. Keywords Pancreatic cancer • Genetics • Pancreas • Mouse models

N. Bardeesy (*) Massachusetts General Hospital Cancer Center and Harvard Medical School, Boston, MA 02114, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_16, © Springer Science+Business Media, LLC 2011

377

378

A.F. Hezel and N. Bardeesy

16.1 Pancreas Anatomy, Physiology, and Development The pancreas regulates protein and carbohydrate digestion and glucose homeostasis through both exocrine and endocrine compartments respectively. Acinar cells of the exocrine pancreas secrete digestive enzymes into a ductal network for transport to the GI tract. The endocrine pancreas regulates metabolism through the secretion of insulin and other hormones into the bloodstream and is organized into numerous highly vascularized islets embedded within the exocrine component (normal pancreatic histology is depicted in Fig. 16.1, left panel). In addition to PDAC, the most common and lethal type of tumor of the pancreas, a series of other types of pancreatic cancers occur in humans, including acinar carcinoma and various types of islet cell carcinoma (insulinoma, glucagonoma). Each of these tumor types has a characteristic histopathological progression and profile of oncogenic mutations and it is thought that these tumor types arise from the transformation of distinct cell populations within the pancreas. PDAC has generally been thought to arise from the pancreatic ductal cells based on their comparable histological features and on the development of PDAC precursor lesions within normal ducts – the issue of PDAC cell-of-origin will be discussed in greater detail further below.

Fig. 16.1 Histological progression of PDAC in genetically engineered mouse models. The histological progression of PDAC in the Pdx1-Cre (or P48-Cre) LSL-KrasG12D mouse model resembles that seen in the human disease. Left panel: normal wild type pancreas. The inset shows high power image of acinar issue (A), a duct (D), and an islet (I). Center panel: PanIN lesion. Activation of Kras promotes the formation of focal PanINs. Right panel: Invasive PDAC. The inactivation of p53 or Ink4a/Arf promotes the progression of PanIN to PDAC. Note that the tumor has invaded the duodenum (Duod.). Lower panel: IPMN. The combined activation of Kras and inactivation of Smad4 promotes cystic tumors resembling IPMN (or in some mouse strain backgrounds, resembling MCN). The cystic tumors progress to PDAC through inactivation of either p53 or Ink4/Arf

16 Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma

379

All the exocrine and endocrine compartments as well as the ducts arise from a common endodermal progenitor population (reviewed in [1]). These multipotent pancreatic progenitor cells express the pancreatic homeobox transcription factors, Pdx1, and the helix-loop-helix transcription factor, p48/Ptf-1, both of which are required for pancreatic development. The differentiated pancreatic cell types arise from the Pdx1+ and p48+ progenitors through coordinated expression of lineage specific transcriptional factors. In later stages of development and in the adult, Pdx1 becomes restricted to the differentiated beta-islet cells but can be re-activated in ductal and acinar cells following pancreatic injury [2]. p48 is confined to the acinar cells in the adult. Both the p48 and Pdx1 promoters have provided useful tools to drive Cre recombinase for the coordinated mutation of floxed engineered alleles in PDAC modeling efforts discussed in detail below.

16.2 Histological and Molecular Characteristics of Human PDAC The ideal mouse PDAC models should incorporate the gene mutations associated with human PDAC and should recapitulate the histopathologic progression of the human disease. Three types of PDAC precursor lesions have been identified; pancreatic intraepithelial neoplasm, mucinous cystic neoplasm (MCN), and intraductal papillary mucinous neoplasm (IPMN) [3, 4]. Each of these precursors appears to arise in association with the normal pancreatic ducts and can undergo progressive stages of dysplasia leading to PDAC. PanIN (graded from stages I–III), the most common and well-characterized type of precursor lesion, are microscopic and cannot be detected by noninvasive methods [4, 5]. Analysis of autopsy specimens have shown that early stage PanIN are present in ~30% of elderly individuals [6]. Molecular profiling studies have identified a series of common genetic mutations that appear with increasing frequency in progressively higher grade PanINs [7–15]. Activating K-RAS mutations are observed in the earliest PanIN lesions and are present in nearly all PDAC [10, 16, 17]. The tumor suppressors INK4A, ARF, p53, and SMAD4 are all commonly inactivated in PDAC. INK4A, encoding the cyclin-dependent kinase 4/6 inhibitor, p16Ink4a, is inactivated by intragenic point mutation, deletion, or promoter hypermethylation in most advanced PanINs and in ~ 80 to 95% of PDAC [17, 18]. Germline INK4A mutations are also associated with a significantly increased risk of developing PDAC. The ARF tumor suppressor (encoding P14ARF (Alternative Reading Frame) which promotes p53 stabilization by antagonizing MDM2-mediated p53 proteolysis) shares common exons with INK4A (but translated in a different reading frame) and is lost in the ~40% of PDAC that sustain deletion of the INK4 locus. p53 is also inactivated in ~50% of PDAC due to missense mutations of the DNA-binding domain. Finally, SMAD4, a transcription factor that is critical for transforming growth factor beta (TGFb) signaling, is deleted or sustains intragenic point mutations in ~50% of PDAC [19]. p53 and SMAD4 inactivation have also been documented in late stage PanINs [11, 12, 14]. These data collectively support a model whereby activating KRAS mutations promote PanIN initiation and that progression of these lesions is facilitated by loss of the

380


INK4A, P53, and SMAD4 tumor suppressors. These predictions have been validated in genetically engineered mouse models as discussed in detail below. The other types of PDAC precursors, MCN and IPMN, are large cystic tumors that can be detected radiographically. These tumors can either have a benign course or progress to PDAC; the probability of malignant progression is not certain, although it is estimated that ~30% of IPMN can progress to PDAC [3]. The mutational profiles of MCN and IPMN, and of the PDAC that arise from these cystic tumors, have not been fully defined, however, the existing evidence suggests that mutations in KRAS, INK4/ARF, p53, and SMAD4 also occur in these tumor types. Recurrent, but less prevalent mutations in a number of other tumor suppressors are observed in PDAC. Germline mutations in the familial breast and ovarian cancer genes, BRCA2, are associated with increased PDAC risk, although somatic BRCA2 mutations have not been reported in sporadic tumors [20]. Germline inactivating mutations in the BRCA2 interacting protein, PALB2, are also observed in a subset of familial PDAC cases [21]. The LKB1 tumor suppressor, encoding a serine-threonine kinase implicated in control of both energy metabolism and cell polarity, is mutationally inactivated in a subset of sporadic PDAC and familial PDAC cases [22] and in associated IPMN [23]. There are also a number of mutational events that disable the TGF-beta signaling pathway in PDAC. While SMAD4 mutation as described above the most frequent, recurrent inactivating mutations have been reported in the TGFbeta receptor type II receptor gene and in SMAD2 and SMAD3 (which encode transcriptional regulators that complex with SMAD4) [24]. In addition to the mutations in the well-characterized cancer-related genes discussed above, numerous other genetic alterations have been identified in PDAC. PDAC is characterized by centrosome abnormalities and high level of chromosomal aberrations and copy number alterations that may point to additional oncogenes/ tumor suppressors [25–35]. Genomic instability appears to occur relatively early in tumor progression and may reflect the capacity of shortened to promote chromosomal aberrations since shortened telomeres and anaphase bridging have been detected in low-grade PanINs [36]. Recent cancer genome sequencing efforts have identified recurrent mutations and copy number alterations in a series of genes whose functional roles in cancer are still under investigation [24]. Notably, while the individual genes are only altered in a small percentage of tumors, grouping them into oncogenic pathways suggests that a set of more than ten signaling pathways may be consistently targeted in PDAC. Genetically engineered mouse models should help in the functional analysis of novel candidate PDAC genes and in the study of telomere dynamics and genomic instability in tumor progression.

16.3 Modeling PDAC The development of mouse PDAC models has progressed from early efforts that employed transgenic expression of a series of viral or endogenous oncoproteins, to more recent systems that employed Cre-Lox technology to generate compound

16 Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma Table 16.1 Transgenic pancreatic cancer models Gene/promoter Phenotype of mouse Transgenics with predominantly acinar phenotypes T-Ag/elastase Acinar cell carcinoma Hras/elastase Acinar cell carcinoma TGF-a/ elastase Acinar cell carcinoma Develop mixed acinar-ductal tumors on a p53+/− background TGF-a/Metallothionein Tubular metaplasia. Develop lesions resembling serous cystadenomas on Ink4a/Arf or p53 null background c-myc/elastase Mixed Acinar-ductal tumors KrasG12D/Mist1 Acinar cell carcinoma Transgenics using the RCAS TVA system c-myc/elastase Islet cell tumors in Ink4a/Arf null mice PyMT/elastase Mixed Acinar-ductal tumors in Ink4a/Arf null mice and p53 null mice Adapted from Hezel et al. [48]

381

References [37, 38] [39] [40–42]

[43, 44]

[41] [45] [46] [47]

mutant mice with pancreas-specific activation or inactivation of genes specifically implicated in human PDAC. These different approaches targeted a spectrum of oncogenic mutations to selective cell lineages within the pancreas and resulted in numerous different phenotypic outputs. Collectively, these efforts provide a series of tools for studying pancreatic cancer and reveal information about the sensitivity of pancreatic cell types to malignant transformation and into the histopathologic phenotypes of the ensuing tumors (Table 16.1).

16.4 Transgenic Expression of Oncogenes in the Pancreas A number of transgenic models of exocrine pancreatic cancer have been developed. Acinar expression of SV40 large T antigen (T-Ag), activated H-RAS, or c-Myc under the control of the Elastase (Ela) promoter leads mainly to acinar cell carcinomas [37–40]. Mixed acinar–ductal tumors and cystic acinar tumor are induced in mice with p53 deficiency combined with acinar expression of the Epidermal Growth Factor Receptor ligand, TGF-a [42]. Acinar-targeted metallothioneinTGF-a (MT-TGF-a) transgenics with compound inactivating mutations in either p53 or Ink4a/Arf develop benign pancreatic ductal lesions resembling serous cystadenomas but do not develop carcinomas [43]. Finally, acinar expression of high levels of activated Kras (Elastase-KrasG12D mice) results in the development of preinvasive lesions with acinar and ductal features [49]. It is important to note that there may be some drawbacks in the use of elastase to drive transgene expression in these models since elastase is normally confined to differentiated acinar cells, and is not expressed in human PDAC. The fact that the tumor phenotype in these mouse models does not closely recapitulate the PanIN-to-PDAC sequence

382


characteristic of the human disease may reflect a silencing of the elastase promoter and consequent loss of oncogene expression as acinar differentiation is lost; the neoplastic lesions in these models may be directed to maintain some acinar features to maintain oncogene expression. KrasG12D has also been targeted to the acinar cells by knocking in this mutant allele into the open reading frame of the acinar transcription factor, Mist1. These mice also develop mixed acinar/ductal tumors again, perhaps reflecting the silencing of acinar genes in ductal tumors [45]. Finally, it is notable that the expression of KrasG12D in the differentiated pancreatic ductal cells (cytokeratin-19-KrasG12V) did not lead to any neoplastic change in the pancreas [50]. Since cytokeratin 19 is expressed in all stages of PDAC progression, these results may indicate that differentiated ducts are not readily transformed by KrasG12V although it is possible that the absence of a PDAC prone phenotype reflected aspects of the design of these transgenic mice.

16.5 Viral Delivery of Oncogenes The avian retroviral transduction system, RCAS-TVA, has been used for the somatic activation of oncogene expression in the pancreas [51, 52]. In this system, transgenic expression of TVA, the receptor for the avian leukosis sarcoma virus subgroup A (ALSV-A), under the control of Elastase (Elastase-tva mice) allows somatic delivery of oncogene-expressing avian retroviruses to the acinar cells [46]. Infection of neonatal elastase-tva; Ink4a/Arf null mice with RCAS vectors expressing c-Myc or polyoma virus middle T antigen (PyMT) led to islet cell tumors or to tumors of mixed acinar and ductal features, respectively. Furthermore, a subset of the PyMT-transduced mice developed PanIN-like lesions. The diverse representation of carcinomas in these models may have reflected targeting of a progenitor cell population since elastase-tva is expressed more broadly in the neonatal pancreas compared with the acinar-specific expression in the adult. p53 deficiency accelerated the development of PyMT-induced ductal and acinar tumors and promoted metastasis [47]. Although RCAS-TVA system has not gained wide usage to date, and the existing models using this system have not produced tumors that closely resemble human PDAC, the system could provide a rapid context for screening novel candidate pancreatic cancer-relevant genes.

16.6 Compound Inducible Mutants As noted above, the Pdx1 and Ptf1-p48 promoters are expressed in the common progenitors of all pancreatic cell lineages [53, 54]. Given that these promoters have relatively restricted expression outside of the pancreas (Pdx1 is also expressed in the developing duodenum and stomach, and Ptf1-p48 is expressed in the cerebellum) the creation of Pdx1-Cre and Ptf1-p48-Cre deleter strains has allowed for targeted somatic mutations of engineered (floxed) alleles in the pancreas.

16 Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma

383

An important tool for studying activated Kras in cancer models has been the generation of inducible knock-inactivated Kras alleles, designated LSL-KrasG12D and LSL-KrasG12ViresLacZ [55, 56]. These alleles are engineered to express mutant Kras (KrasG12D or KrasG12V) from the endogenous promoter and therefore at approximately physiological levels. The excision of a LoxP-flanked Stopper element via transgenic expression of Cre recombinase enables activation of the Kras knock-in allele in specific tissues (see Fig. 16.2). The activation of Kras from the endogenous locus recapitulates the acquisition of intragenic point mutations in evolving human cancers. Perhaps surprisingly, activation of LSL-KrasG12D in all pancreatic cells, from the earliest stages of pancreatic development using either Pdx1-Cre or p48-Cre, does not lead to any anomalies in pancreatic development. However, these mice progressively develop PanIN lesions starting in the first ~3 to 6 weeks of life [57, 58]. KrasG12D-mutants demonstrate a gradual progression of lesions resembling human PanINs-I–III (Fig. 16.1). These lesions develop against the backdrop of histologically normal pancreatic tissue despite the fact that all pancreatic cells harbor the activated KrasG12D allele. PDAC eventually develops in these mice although the latency is relatively long (>1 year). Together these observations indicate that Kras activation promotes the formation of PanIN lesions that can proceed to invasive PDAC, however, genetic/epigenetic alterations are required both to incite neoplasia and facilitate malignant progression. Preliminary analyses indicate that spontaneous deletion of Ink4a/Arf contributes to PDAC progression in this model. Pancreatic phenotypes have also been studied in a related knock-in KrasG12V-IRES-lacZ strain via crosses to a CMV-Cre transgene (which is broadly expressed in different tissues but has limited pancreatic expression). These mutant mice develop PanIN but only when crossed to mice that also express a Cdk4 mutant (R24C) that is refractory to Ink4a inhibition [55]. The mild pancreatic phenotype of these mice in comparison to the Pdx1-Cre KrasG12D mice may be due to the activation of Kras in fewer pancreatic cells or differences in the engineered KrasG12D knock-in based strategy PDX or p48 promoter

Cre

CRE

STOP

STOP

Ink4a/Arf

KRASG12D

p53

KRASG12D-IRES-LacZ

Smad4 Tgfbr2 STOP

Fig. 16.2 Cre/Lox based models relying on activated Kras alleles

p53R273H

384


Kras alleles (e.g., There may be differences in signaling activity between the KrasG12D and KrasG12V alleles or the replacement of the endogenous Kras 3-UTR with iresLacZ could affect expression levels). Irrespective of the different penetrance of neoplasia in these models, the data collectively show that expression of endogenous levels of activated Kras throughout the pancreas leads specifically to PanINs capable of progressing to PDAC. Importantly, there is no evidence of islet cell or acinar cell cancers in these mice. These findings contrast with the development of acinar or mixed acinar/ductal tumors in mice that over-express Kras under acinar-specific promoters.

16.6.1 Cooperation Between Kras and Tumor Suppressors in PDAC The p53 and Ink4a/Arf tumor suppressor genes have been evaluated for their roles in pancreatic development and homeostasis (Figs. 16.1 and 16.2). Inactivation of any of these tumor suppressors alone does not cause any defects in pancreatic development or promote tumorigenesis. However, when these genes are inactivated in combination with Kras activation there is a significant acceleration of PDAC formation, with each model showing differences in the precursor lesions, tumor latency, and resulting histologic phenotype. Pdx1-Cre LSL-Kras mice, with homozygous or heterozygous conditional Ink4a/Arf allele are born normally, but develop rapidly progressive PanIN and succumb to invasive PDAC with greatly decreased latency in comparison to animals with wild type Ink4a/Arf [57, 59]. Inactivation of Ink4a alone also accelerates Kras-driven PDAC but the effects are modest compared to dual Ink4a/Arf inactivation. Rapid PDAC progression also occurs when expression of the KrasG12D allele is combined with inactivation of p53, either through the expression of a mutant p53 knock-in allele (p53R273H) or deletion of a conditional p53 null allele [59, 60]. Together, these data show that p53 and Ink4a/Arf restrain the malignant progression of PanIN to PDAC while not playing a significant role in the initiation of PanIN lesions (Table 16.2). Most importantly, these models appear to accurately recapitulate many histopathologic and clinical features of the human disease. In addition to their evolution from the characteristic PanIN-PDAC progression sequence, the tumors have a similar immunophenotypic profile to human PDAC, including expression of specific pancreatic lineage markers (absence of islet cell markers, e.g., insulin, and acinar markers, e.g., amylase, and expression of ductal markers, e.g., cytokeratin-19), expression of distinct mucins, and activation of a series of relevant developmental/ oncogenic signaling pathways (e.g., Notch, Hedgehog, and EGFR). In terms of other pathologic and clinical features, the tumors show locally invasive growth, as well as regional and distal metastasis, and provoke the formation of a dense stroma (desmoplasia). Hence, these models appear to be well suited for studying many aspects of PDAC biology. The rest of this article will focus on the refinement, study, and translational application of these Kras-driven models.

16 Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma Table 16.2 Activated Kras knock-in engineered PDAC models Alleles Phenotype of mouse KrasG12DPdx1-Cre Spectrum of PanINs and some mice develop PDAC with long latency Average latency ~12 weeks with micrometastatic KrasG12DPdx1-Cre disease Ink4a/Arf−/− KrasG12DPdx1-Cre Average latency ~4 months with gross metastatic disease. LOH of WT p53 allele in p53+/− model, p53R273Hor p53+/− sporadic loss of Ink4a by methylation or deletion KrasG12DPdx1-Cre Average latency = ~5 to 6 months Ink4a/Arf+/− Gross metastatic disease and LOH of WT Ink4a/Arf allele KrasG12DPdx1-Cre LOH of WT p53 allele and loss of Ink4a expression p53+/−Ink4a+/− KrasG12DPdx1-Cre PDAC arising in setting of IPMN or MCN Smad4−/− KrasG12DPdx1-Cre PDAC TGFbr2−/− KrasG12DPdx1-Cre Latency of ~9 to 12 weeks, significantly accelerated compared with KrasG12DPdx1-CreSmad4−/− model Smad4−/−Ink4a+/− G12V PDAC only observed when Kras activation occurs in Kras -IRESlacZ setting of inflammation Ela-CreER

385

References [57, 58] [57] [59, 60]

[59]

[59] [61–63] [64] [61] [65]

The different models have some distinct features that will influence their utility for various applications. Mice with homozygous mutations of either Ink4a/Arf or p53 develop multifocal PDAC with very short latency ( studies that can screen thousands of transcripts simultaneously have great utility for further validating GEMMs, as well as for the identification of novel diagnostic markers, and therapeutic targets and pathways. Recently, three comprehensive and complementary studies utilizing a variety of samples, profiling techniques, and bioinformatic analyses have identified hundreds of transcripts and dozens of pathways that are significantly modulated during disease progression in the TRAMP model. First, Morgenbesser et al. [45] utilized several approaches to study [C57BL/6 TRAMP × FVB]F1 disease, beginning with Serial Analysis of Gene Expression (SAGE) [46] to compare expression profiles between AD and CR primary TRAMP tumors and normal prostate tissue from normal littermates; for each sample type, several specimens were pooled. SAGE is very quantitative and is an open system in that it provides an unbiased accounting of expression as it can detect all expressed transcripts and is not limited to those that have been cloned and presented on an array. This analysis identified several hundred transcripts that were significantly deregulated in all three pair-wise comparisons. The authors validated many of these mRNAs as

402

S.D. Morgenbesser

differentially expressed and expanded the study by utilizing microarray hybridization to analyze the aforementioned specimens, as well as PIN lesions and LN metastases, on a commercial array capable of detecting 10,357 cloned mouse transcripts. Hierarchical clustering analysis of the microarray profiles demonstrated that CR primary tumors were highly similar to both AD and CR LN metastases, and 180 transcripts were identified that distinguished these late-stage tumors from the earlystage samples. The expression of eight transcripts was further validated by real-time PCR (rtPCR). To identify genetic pathways that were altered during tumor progression, Gene Set Enrichment Analysis [47] was employed. The early- and late-clusters were characterized by differences in genes involved with oxidative phosphorylation (e.g., ATP/H+ transporting subunits and cytochrome c oxidase subunits), and the cell cycle (e.g., cyclins and cyclin-dependent kinases). Comparisons between AD and CR tumors revealed alterations in inflammatory response genes. The normal prostate was distinguished from the primary AD and CR tumors by differences in oxidative phosphorylation as well as anabolic enzymes, including those that regulate arachidonic acid and prostaglandin metabolism, which promotes tumor cell proliferation and metastasis; this is particularly interesting as subsequent data supported the possibility that phospholipase A2 group IIA (PLA2G2A) is involved in the highly aggressive nature of late-stage prostate tumors, and with the more severe phenotype on the [C57BL/6 × FVB]F1 background. PLA2G2A promotes arachidonic acid release and metabolism [48, 49], and is over-expressed in late-stage, clinical PC [50, 51]. SAGE revealed and rtPCR confirmed that PLA2G2A was only expressed in the late-stage TRAMP tumors. Moreover, C57BL/6 mice have a naturally occurring, inactivating mutation in PLA2G2A52, and the authors showed that FVB mice have wild-type PLA2G2A alleles and that [C57BL/6 TRAMP × FVB]F1 tumors primarily expressed the normal FVB-derived transcript. A role for PLA2G2A in TRAMP tumorigenesis was further supported by the observation that annexins A3 and A4, which negatively regulate PLA2 enzymes [53], were downregulated in the advanced tumors. Subsequently, Haram et al. [54] performed microarray hybridization on arrays presenting 34,000 murine genes, with RNA derived from eight tumors from C57BL/6 TRAMP mice or nine normal prostate tissues; the samples were studied individually. Several thousand genes were differentially expressed, and hierarchical clustering of the data segregated the samples into two groups consistent with their origin. To determine which biological processes were altered, the authors used the Database for Annotation, Visualization, and Integrated Discovery (DAVID) and identified 185 categories, most notably cell cycle; DNA replication, recombination, and repair; kinase regulation; and nucleotide metabolism. Ingenuity Pathway Analysis was used to place differentially expressed genes into functional categories, which identified many similar pathways as DAVID, but also other metabolic pathways as well as the aryl hydrocarbon receptor signaling pathway and chromatin assembly, segregation, and structural maintenance pathways. The expression of 25 genes was confirmed by rtPCR. Finally, the authors overlaid their expression profiles with similar data derived from human samples to identify common alterations, and they identified two mRNAs that were significant upregulated in PC in both species: Tubb2a, a b-tubulin subtype, and Sox4, the sex-determining region Y box 4 transcription factor.

17 Transgenic Adenocarcinoma of the Mouse Prostate

403

Most recently, Kela et al. [55] used laser capture microdissection to isolate normal and malignant epithelial cells from 28 [C57BL/6 TRAMP × FVB]F1 prostate samples, including low- and high-grade PIN, WD and PD tumors, and metastases. The mRNAs were hybridized to an array containing 45,035 murine probe sets. Transcripts with the highest variance were used to group the samples, all of which clustered according to their original identities except for the primary and metastatic tumors which grouped together. Additional analyses, including DAVID, were applied to identify functional clusters that characterized the histological subtypes. Low grade PIN were primarily defined by enhanced expression of cell-cycle and immune-related genes, high-grade PIN by an increase in apoptotic mRNAs, and the tumors by up-regulation of cell adhesion and cell cycle, and down-regulation of immune system and pro-apoptotic transcripts. To identify tumor aggressivenessassociated genes shared by mouse and man, seven human prostate array datasets were compared to their TRAMP expression profiles, and the analysis identified a common set of 933 genes that were differential between high-grade PIN and tumors. Further refinement of this gene list was achieved by identifying those transcripts that intersected with a list of 194 genes considered to represent the core expression of human PCs, which resulted in 64 genes that may be important in the acquisition of the aggressive phenotypes in both species. Of these, 27 were upregulated, with the most differential being Bub1, microtubule-associated protein tau, topoisomeraseII-a, Aurora kinase A, EZH2, and cell division cycle 6. Each of these studies identified many genes that that were previously observed to be differentially regulated in TRAMP, such as E-cadherin, cyclins, GSTs, and the Aurora kinases, and there were also many genes and pathways that were identified by more than one study, both of which indicate the validity of each analysis. A list of selected transcripts that were differential in at least two of these reports are presented in Table 17.1. There were also many transcripts that were only identified by one study, which may reflect differing methods for detecting transcripts (only one study used an open system) [45], algorithms for normalization, fold and statistical significance cutoffs, and pathway identification, as well as the tendency of arrays to underestimate differences in expression [56]. In addition, the samples possessed unique qualities related to their disease stage (e.g., only one study incorporated CR tumors) [45], celltype complexity (one study only used epithelial cells) [55], and genetic background (two studies utilized [C57BL/6 × FVB]F1 mice whereas one used C57BL/6). Therefore, transcripts identified by only one study are worthy of further exploration.

17.4 Epigenetic Regulation of Gene Expression In human and TRAMP PC, many mechanisms, including androgen-regulation, control gene expression. Alterations in DNA methylation also contribute to the etiology of PC, and simultaneously involve hypermethylation in the regulatory regions of some genes that affects their transcription, as well as widespread hypomethylation that can cause genomic instability. Hypermethylation usually occurs in

404

S.D. Morgenbesser

Table 17.1 Selected differentially expressed transcripts in TRAMP identified by more than one gene expression profiling study Transcript name Transcript symbol References Transcripts that are downregulated during TRAMP tumorigenesis Beta 2 microglobulin B2M [45, 55] Clusterin CLU [45, 55] Defensin beta 1 DEFB1 [45, 55] E-cadherin ECAD; CDH1 [45, 55] Glutathione-S-transferase mu 1 GSTMU1 [45, 55] Glutathione-S-transferase mu 2 GSTMU2 [45, 55] Microseminoprotein MSMB; PSP94 [45, 55] Myosin, light polypeptide kinase MYLK [45, 55] pdz domain containing 1 PDZK1 [45, 54] Probasin PBSN [45, 54] Prostate secretory glycoprotein p12 SPINK3 [45, 55] Serine protease inhibitor kazal-type 5 SPINK5 [45, 55] Smooth muscle myosin heavy chain 11 MYH11 [45, 55] Spermine binding protein SBP [45, 55] Tnf receptor-associated factor 1 TRAF1 [45, 55] Transcripts that are upregulated during TRAMP tumorigenesis Antigen identified by monoclonal antibody Ki-67 Aurora kinase a Aurora kinase b Brain-abundant, membrane attached signal protein 1 Budding uninhibited by benzimidazoles 1 homolog Budding uninhibited by benzimidazoles 1 homolog, beta Cdc2-associated protein CKS2 Cell division cycle 25 homolog C Chromogranin A Cyclin A2 Cyclin B1 Cyclin E1 Dopa decarboxylase Hematological and neurological expressed sequence 1 Ligase I, DNA, ATP-dependent Microtubule-associated protein homolog, transcript variant 3 Minichromosome maintenance homolog 2 Minichromosome maintenance homolog 4 Minichromosome maintenance homolog 6 Mitotic arrest-deficient 2, homolog 1 mutS homolog 2 (E. coli) Nucleoporin 62 Polymerase, DNA, epsilon Proliferating cell nuclear antigen Rac GTPase-activating protein 1 Secretogranin III Sex determining region Y box 4 Stathmin protein (oncoprotein p18)

MKI67 AURKA AURKB BASP1; NAP22 BUB BUB1B CKS2 CDC25C CHGA CCNA2 CCNB1 CCNE1 DDC HN1 LIG1 TPX2 MCM2, mMCM2 MCM4 MCM6 MAD2L1 MSH2 NUP62 POLE PCNA RACGAP1 SCG3 SOX4 STMN1

[45, 55] [54, 55] [54, 55] [45, 55] [54, 55] [45, 54] [54, 55] [45, 54, 55] [45, 55] [45, 54, 55] [54, 55] [54, 55] [54, 55] [45, 54, 55] [45, 54] [54, 55] [45, 54, 55] [54, 55] [54, 55] [54, 55] [45, 55] [54, 55] [54, 55] [54, 55] [45, 54, 55] [45, 55] [54, 55] [45, 55] (continued)

17 Transgenic Adenocarcinoma of the Mouse Prostate Table 17.1 (continued) Transcript name

405

Transcript symbol

References

Survivin (apoptosis inhibitor 4) AIP4 [45, 55] Timeless TIM; TIM1 [54, 55] Topoisomerase (DNA) II alpha TOP2, TOP2A [45, 54, 55] Trophinin (melanoma antigen, family D, 3) TRO, MAGED3 [45, 55] Please see the individual references for additional details, including degree of differential expression

the promoter regions to silence or significantly down-regulate the transcription of some negative tumor regulators including E-cadherin [57, 58], but it can also reside at other locations and correlate with increased expression of some tumor suppressor transcripts, such as p16 [59]. DNA methylation is controlled by three DNA methyltransferase (DNMT) enzymes, most notably DNMT1 whose expression and activity is enhanced in human prostate tumors. The DNMT1 inhibitor 5-aza-2¢deoxycytidine (5-aza) can reverse promoter hypermethylation and result in decreased tumor formation in cell lines and animals. In TRAMP, the E-cadherin protein is expressed in PIN and WD tumors and is reduced or absent in PD tumors [20]; it is also downregulated at the transcriptional level (Table 17.1) [45, 55] suggesting that DNA hypermethylation of this site occurs in TRAMP tumors. Recently, the status and relevance of DNA methylation have been studied in TRAMP by two groups. Day and colleagues initially studied DNMT1 protein expression in the C57BL/6 model, and determined that it was expressed in PIN lesions, and primary and metastatic tumors but not in normal prostate tissues [60]. Treatment of TRAMP C57BL/6 mice with 5-aza from age 6 weeks until 24 weeks resulted in the cessation of tumor progression at the PIN stage and a significant survival benefit. The mechanism of action was confirmed to be reduced DNA hypermethylation. These data suggest that DNA hypermethylation of key tumor suppressor genes is relevant in this model, and can be reversed with a positive preclinical outcome. Karpf and co-workers studied expression of DNMTs and DNA methylation in [C57BL/6 × FVB]F1 mice, primarily in intact (non-castrated) mice [61, 62]. At the PIN stage, they observed an increase in expression of all three DNMT proteins and low level of hypermethylation at specific sites, while PD tumors were characterized by specific hypermethylation events at a high frequency, and hypermethylated loci were remarkably heterogenous in the metastatic samples [61, 62]; the latter was also observed in CR tumors from castrated mice [63]. DNMT activity was increased in primary and metastatic tumors [60]. Nineteen loci were identified as being hypermethylated, and several of these were outside of promoter regions, including one downstream of the p16 transcriptional start site [61] that coincides with p16 mRNA overexpression in TRAMP tumors relative to normal controls [45, 62]. These studies indicate that the TRAMP model is valid for studying mechanisms regulating DNA methylation, discovering hypermethylated loci that might serve as biomarkers [64], and evaluating therapies that act by reversing DNA hypermethylation in PC.

406

S.D. Morgenbesser

17.5 Validating and Elucidating Gene Function Through Additional Genetic Engineering To determine the role of various genes of interest in prostate tumorigenesis, TRAMP mice have been mated to other genetically engineered mice that possess specific alterations, and these functional analyses have been very informative. For example, two groups have utilized TRAMP to address two very different, but key questions – one on genetic variation, and the other on spatial expression – regarding the role of the AR in the development of PC. Variations within the human AR gene, including length of the glutamine Q-tract in the N-terminus, have been associated with disease risk, though a precise correlation has not been established [65]. To determine whether the length of the Q-tract influences tumor progression, three novel lines of mice carrying humanized AR genes (h/mAR) with 12, 21, or 48 residue Q-tracts on a mixed B6:129/Sv background were mated with TRAMP C57BL/6 mice [66]. There were notable differences in the latency and survival, confirming that Q-tract length influences PC progression and suggesting the utility of the TRAMP model for further dissecting the influence of AR polymorphisms. As indicated earlier, the AR is expressed in epithelial luminal cells and stromal cells [8]. To study the influence of the AR during prostate tumorigenesis, TRAMP FVB mice were crossed with C57BL/6 mice harboring AR knockout (ARKO) alleles under the control of an inducible system (ind-ARKO) to abrogate AR either in both compartments or only in prostatic epithelial cells (pes-ARKO) [67, 68]; in particular, the ind-ARKO mice may provide an environment similar to that achieved during ADT. These studies suggested that the AR plays a proliferative role in the stroma and a tumor suppressive role in the epithelium, and illuminate the mechanisms of recurrent disease following ADT in man. In contrast, this approach has also been used to elucidate the relevance of a protein, PSCA, whose function is unknown [69]. PSCA is a cell surface, epithelialrestricted protein that is highly upregulated in most, but not all, human primary and metastatic tumors [70–73]. Mouse PSCA is expressed at high levels in some, but not all cells, in TRAMP C57BL/6 PIN lesions and WD tumors, and in most metastatic tumors; in some mice, PSCA expression could not be detected, and in others, surface expression was detected, but on a lower proportion of cells than on primary tumors [39–41]. When TRAMP C57BL/6 mice were mated to PSCA +/+, +/–, and –/– mice on a mixed 129/Sv background, the formation of primary tumors was unaffected by PSCA status, while metastatic tumors appeared much more frequently in the TRAMP/PSCA KO mice [69]. Examination of primary tumors identified a correlation between PSCA absence and increased cytoplasmic localization of aurora kinase B and survivin, which could confer elevated mitotic activity and promote their progression to later stages. These data demonstrate that reduced PSCA correlated with more advanced disease in TRAMP. While the functional role of PSCA in human and mouse prostate is undetermined, the differences in its expression in metastases between human samples and the TRAMP studies may reflect the influence of genetic background.


407

TRAMP mice have been mated to other mice with targeted gene disruptions to determine the importance of several genes whose expression in TRAMP and/or function in other settings, suggests roles in regulating prostate tumorigenesis. These include mice that are deficient for the PTEN tumor suppressor gene [74], FGF2 [75], the IGF-1 regulator GHRH-R [76], hepatic IGF-1 [77], the AR-regulated SRC-1 and SRC-3 proteins [78, 79], the polyamine biosynthesis regulator arginase II [80], caveolin-1 (which regulates vesicle transport) [81], the aryl hydrocarbon receptor Ahr [82] (which is part of a pathway implicated in TRAMP tumorigenesis by array analysis) [54], and cox-2 (which regulates the conversion of arachidonic acid to prostaglandins [83], an important pathway also implicated by differential gene expression profiling) [45].

17.6 Testing of Known and Putative Therapeutic Agents TRAMP mice have been used extensively to evaluate the efficacy of dozens of androgen regulators, cytotoxic agents, targeted therapies, immunotherapies, and dietary supplements, either as single agents or in combination with other therapies. Some examples of each will be discussed, and these and many others are also summarized in Table 17.2. The pathologic progression of tumorigenesis in TRAMP facilitates prevention trials, beginning when mice are between 4 and 6 weeks old; early or late intervention trials, initiating when mice are approximately 12 weeks or roughly 18 and 24 weeks of age; and regression trials that can also begin when mice are between 18 and 24 weeks of age [20, 84]. Efficacy in the TRAMP model has typically been measured by overall survival (in the case of long-term studies), as well as by assessing the urogenital (UG) weight, the overall burden and pathology of primary and metastatic prostatic lesions, and the number of proliferating and apoptotic cells within the prostate at a few specific times following treatment; the latter three endpoints require sacrificing the mice which limits the analysis of any given animal to a single point in time. Magnetic resonance imaging (MRI) has been used in some studies to measure tumor volume noninvasively, which facilitates the ability to frequently monitor disease progression in the same mice longitudinally and is more economical as fewer animals are needed [22, 85–87]. More recently, improvements in MRI have allowed for tumor grading, which greatly enhances the use of this model for preclinical studies [88], and ultrasound has been used very recently to identify and quantitate tumor burden [89]. In addition, molecular analyses have also been used to determine whether these drugs modulate their targets, or, for agents whose targets are unknown, other markers of PC to determine if their expression patterns were restored to those observed in the normal prostate or earlier-PC stages; in the absence of PSA, such surrogate biomarkers have included IGF-1 and IGFBP3 [86, 89], E-cadherin [10, 42, 90], and the metastasis-associated protein S100A4 [42]. Despite the limited success of ADT in the clinic, as indicated earlier, there are continued efforts to define the mechanisms that allow for CR tumors to emerge,

Immunotherapy

?

? B6 F1

p

r p ei

B6

B6

p

li

F1

ei

Allogeneic cell vaccination p + rIL-2f

CTLA-4 blockade + tumor cell line vaccines (+/- GM-CSF)d mTERT genetic vaccination GM-CSF- and HA -tumor cell vaccines + cyclophosphamided,e HA vaccine + radiotherapyb,e SV40-Tag peptidepulsed DCs IL-2 expressing oncolytic virusf Inhibition of prim and met tumor progression Inhibition of prim and met tumor progression; increased survival Inhibited prostate carcinogenesis

Reduced disease score

Effect on tumorigenesis was not evaluated

Reduced area effected by PIN lesions and tumors Reduced UG weights

Reduced tumor incidence and lowered tumor grade

Table 17.2 Selected therapeutic agents tested preclinically in TRAMP Agent (target/mechanism Type Class/subclass of action) of triala Strainb Effect on prostate tumorigenesisc Cytotoxic drugs Doxirubicin li/r F1 Reduced tumor volume Doxirubicin + NGR-TNF ei B6 Reduced UG weight and disease score li/r B6 Reduced UG weight and disease score; prevented appearance of NE foci and mets; no increase in survival B6 reduced UG and LN weights and R-fluriprofen (proliferative ei incidence of prim and mets arrest and induces apoptosis)

Increased CD4 and CD8+ T-cells

Virus detected

Evidence for primed anti-tumor response SV40 Tag-specific cytolytic activity Virus detected

antigen-specific T-cells infiltrated Treg cell depletion; activated DCs presentf

[116]

[115]

[114]

[113]

[112]

[43]

[111]

[95]

not determined

Inflammatory cells accumulated

References [88] [94]

Effect on target/ surrogate markersc Not determined Not determined Not determined

408 S.D. Morgenbesser

DNA methylation NSAIDsg

HDAC inhibitors

Angiogenesis

Targeted therapies

p

p li p

MS-275

OSU-HDAC42

5aza (DMNT1)

Celecoxib (COX-2)

r

2-Methoxyestradiol

F1

li/r

B6

F1

F1

?

?

B6 B6

F1

p li

F1 F1

r Angiostatin and endostatin p (ECs) ei

2-Methoxyestradiol

F1 F1

ei li

SU5416 (VEGFRs)

Decreased severity of PIN and prevented progression to PD Slowed progression (delayed emergence of CR disease); increased survival Reduced # of PIN lesions; increased apoptosis and decreased proliferation

No effect on onset but slowed progression; decreased proliferation

Reduced UG weight tumors regressed; significantly lowered disease score; decreased proliferation; apoptosis unchanged

No effect on tumor progression Slowed progression to WD but not later progression; PD tumors had increased apoptosis and decreased MVD No effect on tumor progression or survival rate slowed appearance of PIN and WD; decreased proliferation and MVD; increased survival Slowed appearance of PIN and WD; decreased proliferation and MVD; increased survival No effect on tumor progression and no improved survival Reduced UG weight; no indication of neoplasia Dramatic reduction in prostate volume

Lowered COX-2 and downstream targets

No indication of whether target is present Increased histone H3 acetylation Not determined

Not determined Decreased serum testosterone Decreased Sp1 and FLIP

Not determined

VEGFR2 reduced

Not determined VEGFR2 reduced

Not determined Not determined

(continued)

[106]

[101]

[105]

[104]

[100]

[87]

[99]

[84]

17 Transgenic Adenocarcinoma of the Mouse Prostate 409

Dietary supplements

Other

Genistein Green tea polyphenols

R-enantiomer of etodolac (retinoid X-receptor) B6 F1

B6 B6 B6 B6

p p ei li

B6

p p

ei

F1

p

Indomethacin (nonselective COXinhibitor)

F1

p

B6

B6

p/ei

p

Strainb

Type of triala

Exisulind (COX-2)

Table 17.2 (continued) Agent (target/mechanism Class/subclass of action)

Reduced progression to PD Prevent or delay in tumor development; reduced tumor growth, no mets; increased apoptosis and reduced prolif; increased survival Not evaluated; purpose was to evaluate S100A4 and E-cadherin as biomarkers Increased survival; reduced tumor burden; mostly normal with sparse PIN lesions Increased survival; reduced tumor burden; mostly HGPIN with some WD Increased survival; reduced tumor burden; WD and MD only

Reduced average tumor mass and freq of metas; increased apoptosis in prostate

Reduced incidence of palpable tumors and stabilized disease at WD, no mets; reduced prolif and increased apoptosis; increased survival No reduction in incidence of prim and met tumors; no change in time of tumor development or survival Reduced # of PIN lesions; increased apoptosis and decreased proliferation No reduction in incidence of prim and met tumors; no change in time of tumor development or survival

Effect on prostate tumorigenesisc

Moderate IGF-1 reduction

Reduced S100A4; restored E-cadherin Significant IGF-1 reduction Strong IGF-1 reduction

Not determined Lowered IGF-1; IGFB3 restored

[89]

[42]

[123] [86]

[107]

[83]

Not determined

Retinoid X receptor greatly reduced

[106]

[83]

Not determined (see legend) Decreased cox activity

[85]

References

Reduced cox-2 and activity of downstream marker

Effect on target/ surrogate markersc

410 S.D. Morgenbesser

r p

B6 B6

No effect on IGF-1 No effect on survival; no diff in tumor burden Decreased incidence of PIN and WD; reduced ND [122] proliferation and increased apoptosis Diallyl trisulfide p F1 Inhibits progression to PD; reduced pulmonary Decreased NE marker; [90] no effect on mets; decreased proliferation; no effect on E-cadherin apoptosis or angiogenesis [10] Sulforaphane p F1 Reduced incidence of WD, PD and pulmonary No effect on E-cadherin; mets; decreased proliferation; modest increased T-cell increase in apoptosis; no effect on infiltration angiogenesis Briefs details from some pre-clinical trials performed with TRAMP mice; please refer to the individual references for full experimental methods, results, and conclusions, and to the text for any abbreviations and acronyms not defined here a To indicate the type of trial, the following abbreviations are used: p, prevention; ei, early intervention; li, late intervention; r, regression b B6, C57BL/6; F1, [C57BL/6 × FVB]F1;?, not clear c Results presented are relative to age- and strain-matched, untreated or vehicle-control treated TRAMP mice. Abbreviations: prim, primary tumors; mets, metastases; UG, urogenital d CTLA-4, cytotoxic T-lymphocyte antigen 4; GM-CSF, granulocyte macrophage-colony stimulating factor e To introduce influenza hemagglutinin (HA) as a tumor antigen in these studies, a variant of TRAMP mice was utilized derived by mating TRAMP mice with Pro-HA mice which express HA under the control of the prostate-specific probasin promoter f IL-2, interleukin-2. Oncolytic viruses are selective for tumor cells g NSAIDS, non-steroidal anti-inflammatory drugs. The contrasting results with the COX-2 inhibitors correlate with differences in COX-2 expression and activity in the different mouse strains. In the studies in the TRAMP C57BL/6 model, Cox-2 expression and activity were shown to be elevated in the tumors relative to normal prostate tissue at multiple stages [85, 106], but in the TRAMP [C57BL/6 × FVB]F1 model, expression was actually found to be decreased, and knocking out COX-2 by gene deletion did not effect prostate tumorigenesis [106]. These differences are similar to that observed in human, suggesting that genetic heterogeneity plays a role, and may effect the clinical utility of these agents [126]

Inositol hexaphosphatate

17 Transgenic Adenocarcinoma of the Mouse Prostate 411

412

S.D. Morgenbesser

and to develop therapies to counteract aberrant activation of the androgen pathway [2, 3]. The observation that castration in TRAMP leads to the quick regression and eventual return of tumors [21, 22] suggests that TRAMP is suitable for studying alternative hormonal interventions to identify those that may achieve a positive clinical outcome. Indeed, two of the first therapeutics evaluated in TRAMP mice were flutamide (an anti-androgen) and toremifene (an antiestrogen), and both were shown to prevent tumor progression in TRAMP C57BL/6 mice [91, 92]. Cytotoxic agents are also utilized alone or in combination with other therapies to treat men with PC, including doxorubicin which has limited success in the clinic in part because it has a low tumor penetrance and is rapidly cleared from the bloodstream [93]. The efficacy of doxorubicin was evaluated in [C57BL/6 TRAMP × FVB]F1 mice bearing WD and PD tumors, and a reduction in volume of both tumor types was observed [88]. More recently, to improve the delivery of doxorubicin to the TRAMP tumors, the drug was combined with a peptide (NGF-TNF) that is capable of targeting drugs to the tumor vasculature [94]. The NGF peptide binds CD13, an EC surface protein, and TNF (tumor necrosis factor-a) alters vessel permeability. When 12- or 17-week-old TRAMP C57BL/6 mice, which were shown to express CD13 on the vasculature, were treated, the combination therapy was more effective than doxorubicin alone. More details on these studies as well as those with another cytotoxic agent that has been evaluated in autochthonous TRAMP mice can be found in Table 17.2 [95]. A variety of targeted therapies have been evaluated in TRAMP mice that act upon different pathways or cellular processes. As described earlier, active angiogenesis in TRAMP is associated with the expression of master-regulatory proteins that are being targeted clinically [29, 31], making it a suitable model for testing anti-angiogenic agents. SU5416 is a potent small molecule inhibitor of VEGRs that hinders proliferation and tube formation of human [96] and mouse [97] cultured ECs. In one study [C57BL/6 TRAMP × FVB]F1 mice were given SU5416 beginning when they were 10 or 16 weeks of age for a total of 6 weeks at which time they were sacrificed, in early or late intervention trials, respectively, or commencing when they had palpable abdominal masses until they were in distress in a regression trial [84]. In the late intervention trial, which was carried out concomitant with VEGF and VEGFR2 expression, SU5416 slowed the progression of PIN to WD tumors, but did not affect the development of PD and metastatic cancers. Within the tumors, there was increased apoptosis and decreased MVD, suggesting that SU5416 interferes with VEGFR2 signaling in PC. Differences in tumor progression were not detected in the early intervention trial consistent with observations that although VEGFR1 was expressed, VEGF levels were quite low, suggesting that targeted therapy to the VEGF axis would not be relevant at this stage of tumorigenesis. SU5416 was not effective in the regression trial, as similar incidence of metastatic disease and survival rate were observed. Remarkably, this study is consistent with the lack of efficacy observed with SU5416 in clinical trials with PC patients [98]. In addition, TRAMP has been utilized to study other antiangiogenic agents [87, 99, 100] (Table 17.2).


413

As discussed earlier, the DNMT1 inhibitor 5-aza (known clinically as decitabine) can reverse DNA methylation and stabilize tumor progression when given to TRAMP mice before PIN lesions appear. In addition, when 5-aza is given to [C57BL/6 TRAMP × FVB]F1 mice simultaneous with castration at 15 weeks, and then for an additional 10 weeks, there was a marked delay in the onset of CR disease with an overall improvement in survival, suggesting potential utility in treating men with this type of advanced PC [101]. Indeed, 5-aza has been evaluated in dozens of clinical trials to treat a wide variety of tumor types, including CR metastatic PC [102]. In this trial, there was a modest clinical benefit, which is promising considering the very advanced state of the disease, and the highly unstable nature of 5-aza in vivo [103]. Continued evaluation of 5-aza may yield improved outcomes, but this data also suggests the potential for other targeted drugs with a similar mechanism of action. Preclinical trialsin TRAMP with targeted therapies directed at other types of targets are summarized in Table 17.2 [83, 85, 104–107]. Immunotherapy, aimed at enhancing the body’s anti-tumor immune response, is also a potential strategy for treating human PC that has received a great deal of attention [108]. One approach involves vaccination against immunogenic PC-specific markers. In this regard, PSCA represents an attractive target based upon its selective over-expression on the surface of most prostate tumors. Indeed, when young TRAMP C57BL/6 mice with PIN lesions were primed with mouse PSCA cDNA and boosted with Venezuelan equine encephalitis virus replicons containing mouse PSCA, there was a protective immune response associated with a significant increase in survival [109]. The TRAMP tumors that were present were WD, had areas with many apoptotic cells, and were infiltrated by T cells, macrophages, and dendritic cells (DCs). The tumor cells had increased MHC class I expression and cytokine production, and decreased PSCA expression. The increase in survival and disease stabilization at the WD stage is interesting given that TRAMP tumors with reduced PSCA as a result of genetic KO were associated with progression to metastatic disease; it is not clear why metastases were not observed in this therapeutic study, but that observation is encouraging and suggests the potential utility for this treatment in man if applied early. Indeed, in a phase I/II clinical trial in which patients with advanced PC were vaccinated with DCs preloaded with PSCA peptides, there was evidence of infiltrating T-cells but with limited clinical benefit [110]. Some other immunologic-based approaches that have been evaluated in TRAMP are listed in Table 17.2 [43, 111–116]. Finally, TRAMP mice have been used extensively to evaluate the effects of various dietary supplements derived from vegetables and plants, mirroring the use of alternative medicines by nearly one-third of all men with PC [117]. The use of such agents, which have been associated with multiple anti-tumor effects including cellular growth arrest and apoptosis, is controversial, in part because these materials are frequently impure, though the approach is considered worthy of continued study given the epidemiologic evidence correlating diet and PC incidence and growth, particularly when the agents are of high purity. For example, silibin, which is isolated from milk thistle seeds and inhibits the growth of PC and other transformed cell lines in culture or propagated as xenografts, was recently tested in

414

S.D. Morgenbesser

TRAMP mice by dietary supplementation with highly purified material [118–120]. In a preventative trial, TRAMP C57BL/6 and [C57BL/6 TRAMP × FVB]F1 mice were fed silibin beginning at 4 weeks of age and continuing for 20 weeks [118]; in most mice, PIN lesions were observed with considerable decrease in tumor incidence, and prostate tissues were characterized by decreased proliferative and increased apoptotic rates. In two intervention trials, TRAMP C57BL/6 mice were fed diets supplemented with silibin or silybin-phytosome, which has superior bioavailability, at older ages, when PIN lesions or tumors were already present [119, 120]; tumor growth, progression, and metastasis were perturbed, and anti-proliferative and anti-angiogenic effects were observed. It was well tolerated in a phase I trial of patients with advanced PC [121] and a phase II trial was recently completed with the results pending. Examples of other alternative approaches that have been evaluated for their efficacy in TRAMP are described in Table 17.2 [10, 86, 89, 90, 122, 123]. Some of the preclinical results generated with TRAMP mice have led to further development; for instance the work with CTLA4-blockade and tumor cell vaccination [111] has led to clinical trials [108] and additional experimentation with the TRAMP model has led to the identification of PC T-cell antigens with relevance to the human disease [124]. In addition, the model was used to evaluate drug enhancements (such as to doxorubicin) or to follow negative responses by molecular studies to determine the reasons for failure, such as by determining the relevance of the target in prostate tumorigenesis by gene deletion [83] or by analyzing the expression of other key modulators [125]. In the case of COX-2 and its inhibitors, protein function and therapeutic efficacy were strain-dependent (Table 17.2 and legend). Therefore, the TRAMP model can be utilized to improve agents at the preclinical stage or for patient selection in the clinic. The TRAMP mouse is commercially available, with C57BL/6 mice and a congenic FVB strain that can be obtained from Jackson Laboratories and the National Cancer Institute Mouse Models of Human Cancers Consortium Repository, respectively [35].

17.7 Summary, Conclusions, and Future Directions The TRAMP mouse is a highly suitable model for the analysis of genes and pathways, testing of therapeutic agents, and identification of biomarkers, that have relevance for human PC. Clearly the role and regulation of the AR and its pathway, and of DNA methylation, can be further evaluated and targeted in these mice. In addition, the three recent, comprehensive gene expression profiling studies have provided the research community with hundreds of candidate diagnostic markers and therapeutic target proteins and pathways for further inquiry, including some that appear to play important roles in the progression to the advanced, aggressive stages of the disease, such as PLA2G2A, and arachodonic acid and prostaglandin metabolism. These lists contain proteins expressed on epithelial, stem, NE, EC, and stromal cells, and given their importance in disease progression in this model, there is an opportunity therein


415

to study inhibitors directed against any of them. Breeding of TRAMP mice to those mice that have genetically engineered alterations in proteins of interest is a powerful approach for elucidating their relevance in PC. For some proteins, exemplified by PSA, PSCA, PLA2G2A, and COX-2, expression and/or function may exhibit species- or strain-related differences which need to be considered in determining their utility as targets, or in evaluating preclinical efficacy data. Acknowledgments The author is grateful to Norman Greenberg for helpful discussions and critical review of this chapter.

References 1. http://www.cancer.gov/cancertopics/types/prostate. 2. Attar RM, Takimoto CH, Gottardis MM. Castration-resistant prostate cancer: locking up the molecular escape routesq. Clin Cancer Res. 2009;15:3251–5. 3. Van Poppel H, Joniau S, Van Cleynenbreugel B, Mottaghy FM, Oyen R. Diagnostic evaluation of PSA recurrence and review of hormonal management after radical prostatectomy. Prostate Cancer Prostatic Dis. 2009;12:116–23. 4. Andriole GL, Crawford ED, Grubb III RL, et al. Mortality results from a randomized prostate-cancer screening trial. N Eng J Med. 2009;360:1310–9. 5. Schroder FH, Hugosson J, Roobol MJ, et al. Screening and prostate-cancer mortality in a randomized european study. N Eng J Med. 2009;390:1320–8. 6. Pienta KJ, Abate-Shen C, Agus DB, et al. The current state of preclinical prostate cancer animal models. Prostate. 2008;68:629–39. 7. Hsieh C-L, Xie Z, Yu J, et al. Non-invasive bioluminescent detection of prostate cancer growth and metastasis in bigenictransgenic mouse model. Prostate. 2007;67:685–91. 8. Gingrich JR, Barrios RJ, Foster BA, Greenberg NM. Pathologic progression of autochthonous prostate cancer in the TRAMP model. Prostate Cancer Prostatic Dis. 1999;2:70–5. 9. Foster BA, Gingrich JR, Kwon ED, Madias C, Greenberg NM. Characterization of prostatic epithelial cell lines derived from transgenic adenocarcinoma of the mouse prostate (TRAMP) model. Cancer Res. 1997;57:3325–30. 10. Singh SV, Warin R, Xiao D, et al. Sulforaphane inhibits prostate carcinogenesis and pulmonary metastasis in TRAMP mice in association with increased cytotoxicity of natural killer cells. Cancer Res. 2009;69:2117–25. 11. Ozerdem U. Targeting of pericytes diminishes neovascularization and lymphangiogenesis in prostate cancer. Prostate. 2006;66:294–304. 12. Lee ECY, Frolov A, Li R, Ayala G, Greenberg NM. Targeting aurora kinases for the treatment of prostate cancer. Cancer Res. 2006;66:4996–5002. 13. Greenberg NM, DeMayo FJ, Sheppard PC, et al. The rat probasin gene promoter directs hormonally and developmentally regulated expression of a heterologous gene specifically to the prostate in transgenic mice. Mol Endocrinol. 1994;8:230–9. 14. Greenberg NM, DeMayo F, Finegold MJ, et al. Prostate cancer in a transgenic mouse. Proc Natl Acad Sci USA. 1995;92:3439–43. 15. Ludlow JW. Interactions between SV40 large-tumor antigen and the growth suppressor proteins pRb and p53. FASEB J. 1993;7:866–71. 16. Bookstein R, Rio P, Madreperla SA, et al. Promoter deletion and loss of retinoblastoma gene expression in human prostate carcinoma. Proc Natl Acad Sci USA. 1990;87:7762–6. 17. Navone NM, Troncoso P, Pisters LL, et al. p52 protein accumulation and gene mutation in the progression of human prostate carcinoma. J Natl Cancer Inst. 1993;85:1657–69.

416

S.D. Morgenbesser

18. Cotsiki M, Lock RL, Cheng Y, et al. Simian virus 40 large T antigen targets the spindle assembly checkpoint protein Bub1. Proc Natl Acad Sci USA. 2004;101:947–52. 19. Chen W, Possemato R, Cambell KT, Plattner CA, Pallas DC, Hahn WC. Identification of specific PP2A complexes involved in human cell transformation. Cancer Cell. 2004;5:127–36. 20. Kaplan-Lefko PJ, Chen T-M, Ittmann MM, et al. Pathobiology of autochthonous prostate cancer in a pre-clinical transgenic mouse model. Prostate. 2003;55:219–37. 21. Gingrich JR, Barrios RJ, Kattan MW, Nahm HS, Finegold MJ, Greenberg NM. Androgenindependent prostate cancer progression in the TRAMP model. Cancer Res. 1997;57:4687–91. 22. Eng MH, Charles LG, Ross BD, et al. Early castration reduces prostatic carcinogenesis in transgenic mice. Urology. 1999;54:1112–9. 23. Han G, Foster BA, Mistry S, et al. Hormone status selects for spontaneous somatic androgen receptor variants that demonstrate specific ligand and cofactor dependent activities in autochthonous prostate cancer. J Biol Chem. 2001;276:11204–13. 24. Han G, Buchanan G, Ittmann M, et al. Mutation of the androgen receptor causes oncogenic transformation of the prostate. Proc Natl Acad Sci USA. 2005;102:1151–6. 25. Niu Y-N, Xia S-J. Stroma-epithelium crosstalk in prostate cancer. Asian J Androl. 2009;11:28–35. 26. Dudley AC, Shih S-C, Cliffe AR, Hida K, Klagsbrun M. Attenuated p53 activation in tumorassociated stromal cells accompanies decreased sensitivity to etoposide and vincristine. Br J Cancer. 2008;99:118–25. 27. Gingrich JR, Barrios RJ, Morton RA, et al. Metastatic prostate cancer in a transgenic mouse. Cancer Res. 1996;56:4096–102. 28. Lissbrant IF, Lissbrant E, Damber JE, Bergh A. Blood vessels are regulators of growth, diagnostic markers, and therapeutic targets in prostate cancer. Scand J Urol Nephrol. 2001;35:437–52. 29. Huss WJ, Hanrahan CF, Barrios RJ, Simons JW, Greenberg NM. Angiogenesis and prostate cancer: identification of a molecular progression switch. Cancer Res. 2001;61:2736–43. 30. Ozawa MG, Yao VJ, Chanthery YH, et al. Angiogenesis with pericyte abnormalities in a transgenic model of prostate carcinoma. Cancer. 2005;104:2104–15. 31. Huss WJ, Barrios RJ, Foster BA, Greenberg NM. Differential expression of specific FGF ligand and receptor isoforms during angiogenesis associated with prostate cancer progression. Prostate. 2003;54:8–16. 32. Cindolo L, Cantile M, Vacherot F, Terry S, de la Taille A. Neuroendocrine differentiation in prostate cancer: from lab to bedside. Urologia Internationalis. 2007;79:287–96. 33. Wright ME, Tsai M-J, Aebersold R. Androgen receptor represses the neuroendocrine transdifferentiation process in prostate cancer cells. Mol Endocrinol. 2003;17:1726–37. 34. Huss WJ, Gray DR, Tavakoli K, et al. Origin of androgen-insensitive poorly differentiated tumors in the transgenic adenocarcinoma of mouse prostate model. Neoplasia. 2007;9:938–50. 35. Chiaverotti T, Couto SS, Donjacour AA, et al. Dissociation of epithelial and neuroendocrine carcinoma lineages in the transgenic adenocarcinoma of mouse prostate model of prostate cancer. Am J Pathol. 2008;172:236–46. 36. Foster BA, Kaplan PJ, Greenberg NM. Characterization of the FGF axis and identification of a novel FGFR1iiic isoform during prostate cancer progression in the TRAMP model. Prostate Cancer Prostatic Dis. 1999;2:76–82. 37. Kaplan PJ, Mohan S, Cohen P, Foster BA, Greenberg NM. The insulin-like growth factor axis and prostate cancer: lessons from the transgenic adenocarcinoma of mouse prostate (TRAMP) model. Cancer Res. 1999;59:2203–9. 38. Maddison LA, Huss WJ, Barrios RJ, Greenberg NM. Differential expression of cell cycle regulatory molecules and evidence for a “cyclin switch” during progression of prostate cancer. Prostate. 2003;58:335–44. 39. Ross S, Spencer SD, Lasky LA, Koeppen H. Selective expression of murine prostate stem cell antigen in fetal and adult tissues and the transgenic adenocarcinoma of the mouse prostate model of prostate carcinogenesis. Am J Pathol. 2001;158:809–16.


417

40. Yang D, Holt GE, Velders MP, Kwon ED, Kast WM. Murine six-transmembrane epithelial antigen of the prostate, prostate stem cell antigen, and prostate-specific membrane antigen: prostate-specific cell surface antigens highly expressed in prostate cancer of transgenic adenocarcinoma mouse prostate mice. Cancer Res. 2001;61:5857–60. 41. Dubey P, Wu H, Reiter RE, Witte ON. Alternative pathways to prostate carcinoma activate prostate stem cell antigen expression. Cancer Res. 2001;61:3256–61. 42. Saleem M, Adhami VM, Ahmad N, Gupta S, Mukhtar H. Prognostic significance of metastasis-associated protein S100A4 (Mts1) in prostate cancer progression and chemoprevention regimens in an authochthonous mouse model. Clinical Cancer Res. 2005;11:147–53. 43. Mennuni C, Ugel S, Mori F, et al. Preventative vaccination with telomerase controls tumor growth in genetically engineered and carcinogen-induced mouse models of cancer. Cancer Res. 2008;68:9865–74. 44. Frohlich DA, McCabe MT, Arnold RS, Day ML. The role of nrf2 in increased reactive oxygen species and DNA damage in prostate tumorigenesis. Oncogene. 2008;27:4353–62. 45. Morgenbesser SD, McLaren RP, Richards B, et al. Identification of genes potentially involved in the acquisition of androgen-independent and metastatic tumor growth in an autochthonous genetically engineered mouse prostate cancer model. Prostate. 2007;67:83–106. 46. Velculescu VE, Zhang L, Vogelstein B, Kinzler KW. Serial analysis of gene expression. Science. 1995;270:484–7. 47. Mootha V, Lindgren C, Eriksson K, et al. PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003;34:267–73. 48. Ablin RJ, Shaw MW. Prostaglandin modulation of prostate tumor growth and metastases. Anticancer Res. 1986;6:327–8. 49. Rose D, Connolly J. Effects of fatty acids and eicosanoid synthesis inhibitors on the growth of two human prostate cancer cell lines. Prostate. 1991;18:243–54. 50. Jiang J, Neubauder BL, Graff JR, et al. Expression of group IIa secretory phospholipase A2 is elevated in prostatic intraepithelial neoplasia and adenocarcinoma. Am J Pathol. 2002;160:667–71. 51. Graff JR, Konicek BW, Deddens JA, et al. Expression of group IIa secretory phospholipase A2 increases with prostate tumor grade. Clin Cancer Res. 2001;7:3857–61. 52. Kennedy BP, Payette P, Mudgett J, et al. A natural disruption of the secretory group II phospholipase A2 gene in inbred mouse strains. J Biol Chem. 1995;270:22378–85. 53. Gerke V, Moss SE. Annexins: from structure to function. Physiologic Rev 2001;82:331–71. 54. Haram KM, Peltier HJ, Lu B, et al. Gene expression profile of mouse prostate tumors reveal dysregulations in major biological processes and identifies potential murine targets for preclinical development of human prostate cancer therapy. Prostate. 2008;68:1517–30. 55. Kela I, Harmelin A, Waks T, Orr-Urteger A, Domany E, Eshhar Z. Interspecies comparison of prostate cancer gene-expression profiles reveals genes associated with aggressive tumors. Prostate. 2009;69(10):1034–44. 56. Yuen T, Wurmbach E, Pfeffer RL, Ebersole BJ, Sealfon SC. Accuracy and calibration of commercial oligonucleotide and custom cDN microarrays. Nucleic Acids Res. 2002; 30:e48. 57. Tomika K, van Bokhoven A, van Leenders GJLH, et al. Cadherin switching in human prostate cancer progression. Cancer Res. 2000;60:3650–4. 58. DeMarzo AM, Knudsen B, Chan-Tack K, Epstein JI. E-cadherin as a marker of tumor aggressiveness in routinely processed radical prostatectomy specimens. Urology. 1999;53:707–13. 59. Nguyen TT, Nguyen CT, Gonzales FA, Nichols PW, Yu MC, Jones PA. Analysis of cyclindependent kinase inhibitor expression and methylation patterns in human prostate cancers. Prostate. 2000;43:233–42. 60. McCabe MT, Low JA, Daignault S, Imperiale MJ, Wojno KJ, Day ML. Inhibition of DNA methyltransferase activity prevents tumorigenesis in a mouse model of prostate cancer. Cancer Res. 2006;66:385–92.

418

S.D. Morgenbesser

61. Morey SR, Smiraglia DJ, James SR, et al. DNA methylation pathway alterations in an autochthonous murine model of prostate cancer. Cancer Res. 2006;66:11659–67. 62. Morey Kinney SR, Smiraglia DJ, James SR, Moser MT, Foster BA, Karpf AR. Stagespecific alterations of DNA methyltransferase expression, DNA hypermethylation, and DNA hypomethylation during prostate cancer progressin in the transgenic adenocarcinoma of mouse prostate model. Mol Cancer Res. 2008;6:1365–74. 63. Camoriano M, Morey Kinney SR, Moser MT, et al. Phenotype-specific CpG island methylation events in a murine model of prostate cancer. Cancer Res. 2008;68:4173–82. 64. http://www.epigenomics.com/en/Press_Releases/2009/datednews/090428_Predictive_ Biosciences.html. 65. Ntais C, Polycarpou A, Tsatsoulis A. Molecular epidemiology of prostate cancer: androgens and polymorphisms in androgen-related genes. Eur J Endocrinol. 2003;149:469–77. 66. Albertelli MA, O’Mahony OA, Brogley M, et al. Glutamine tract length of human androgen receptors affects hormone-dependent and -independent prostate cancer in mice. Human Mol Genet. 2008;17:98–110. 67. Niu Y, Altuwaijri S, Lai K-P, et al. Androgen receptor is a tumor suppressor and proliferator in prostate cancer. Proc Natl Acad Sci USA. 2008;105:12182–7. 68. Niu Y, Altuwaijri S, Yeh S, et al. Targeting the stromal androgen receptor in primary prostate tumors at earlier stages. Proc Natl Acad Sci USA. 2008;105:12188–93. 69. Moore ML, Teitell MA, Kim Y, et al. Deletion of PSCA increases metastasis of TRAMPinduced prostate tumors without altering primary tumor formation. Prostate. 2008;68:139–51. 70. Reiter RE, Gu Z, Watabe T, et al. Prostate stem cell antigen: a cell surface marker overexpressed in prostate cancer. Proc Natl Acad Sci USA. 1998;95:1735. 71. Gu Z, Thomas G, Yamashiro J, et al. Prostate stem cell antigen (PSCA) expression increases with high gleason score, advanced stage and bone metastasis in prostate cancer. Oncogene. 2000;19:1288–96. 72. Lam JS, Yamashiro J, Shintaku IP, et al. Prostate stem cell antigen is overexpressed in prostate cancer metastases. Clin Cancer Res. 2005;11:2591–6. 73. Ross S, Spencer SD, Holcomb I, et al. Prostate stem cell antigen as therapy target: tissue expression and in vivo efficacy of an immunoconjugate. Cancer Res. 2002;62:2546–53. 74. Kwabi-Addo B, Giri D, Schmidt K, et al. Haploinsufficiency of the Pten tumor suppressor gene promotes prostate cancer progression. Proc Natl Acad Sci USA. 2001;98:11563–8. 75. Polnaszek N, Kwabi-Addo B, Peterson LE, et al. Fibroblast growth factor 2 promotes tumor progression in an autochthonous mouse model of prostate cancer. Cancer Res. 2003;63:5754–60. 76. Majeed N, Blouin M-J, Kaplan-Lefko PJ, et al. A germ line mutation that delays prostate cancer progression and prolongs survival in a murine prostate cancer model. Oncogene. 2005;24:4736–40. 77. Anzo M, Cobb LJ, Hwang DL, et al. Targeted deletion of hepatic Igf1 in TRAMP mice leads to dramatic alterations in the circulating insulin-like growth factor axis but does not reduce tumor progression. Cancer Res. 2008;68:3342–9. 78. Chung AC-K, Zhou S, Liao L, Tien JC-Y, Greenberg NM, Xu J. Genetic ablation of the amplified-in-breast cancer 1 inhibits spontaneous prostate cancer progression in mice. Cancer Res. 2007;67:5965–75. 79. Tien JC-Y, Zhou S, Xu J. The role of SRC-1 in murine prostate carcinogenesis is nonessential due to a possible compensation of SRC-3/AIB1 overexpression. Int J Biol Sci. 2009;5:256–64. 80. Mumenthaler SM, Rozengurt N, Livesay JC, Sabaghian A, Cederbaum SD, Grody WW. Disruption of arginase II alters prostate tumor formation in TRAMP mice. Prostate. 2008;68:1561–9. 81. Williams TM, Hassan GS, Li J, et al. Caveolin-1 promotes tumor progression in an autochthonous mouse model of prostate cancer. J Biol Chem. 2005;280:25134–45. 82. Fritz WA, Lin T-M, Cardiff RD, Peterson RE. The aryl hydrocarbon receptor inhibits prostate carcinogenesis in TRAMP mice. Carcinogenesis. 2007;28:497–505.


419

83. Wang X, Colby JKL, Yang P, Fischer SM, Newman RA, Klein RD. The resistance to the tumor suppressive effects of COX inhibitors and COX-2 gene disruption in TRAMP mice is associated with the loss of COX expression in prostate tissue. Carcinogenesis. 2008;29:120–8. 84. Huss WJ, Barrios RJ, Greenberg NM. SU5416 selectively impairs angiogenesis to induce prostate cancer-specific apoptosis. Mol Cancer Therapeut. 2003;2:611–6. 85. Gupta S, Adhami VM, Subbarayan M, et al. Suppression of prostate carcinogenesis by dietary supplementation of celecoxib in transgenic adenocarcinoma of the mouse prostate model. Cancer Res. 2004;64:3334–43. 86. Gupta S, Hastak K, Ahmad N, Lewin JS, Mukhtar H. Inhibition of prostate carcinogenesis in TRAMP mice by oral infusion of green tea polyphenols. Proc Natl Acad Sci USA. 2001;98:10350–5. 87. Garcia GE, Wisniewski H-G, Lucia MS, et al. 2-methoxyestradiol inhibits prostate tumor development in transgenic adenocarcinoma of mouse prostate: role of tumor necrosis factora-stimulated gene 6. Clin. Cancer Res. 2006;12:980–8. 88. Degrassi A, Russo M, Scanziani E, et al. Magnetic resonance imaging and histopathological characterization of prostate tumors in TRAMP mice as model for pre-clinical trials. Prostate. 2007;67:396–404. 89. Adhami VM, Siddiqui IA, Sarfaraz S, et al. Effective prostate cancer chemopreventive intervention with green tea polyphenols in the TRAMP models depends on the stage of the disease. Clin Cancer Res. 2009;15:1947–53. 90. Singh SV, Powolny AA, Stan SD, et al. Garlic constituent diallyl trisulfide prevents development of poorly differentiated prostate cancer and pulmonary metastasis multiplicity in TRAMP mice. Cancer Res. 2008;68:9503–11. 91. Raghow S, Kuliyev E, Steakley M, Greenberg NM, Steiner MS. Efficacious chemoprevention of primary prostate cancer by flutamide in an authochthonous transgenic model. Cancer Res. 2000;60:4093–7. 92. Raghow S, Hooshdaran MZ, Katiyar S, Steiner MS. Toremifene prevents prostate cancer in the transgenic adenocarcinoma of mouse prostate model. Cancer Res. 2002;62:1370–6. 93. Primeau AJ, Rendon A, Hedley D, Lilge L, Tannock IF. The distribution of the anticancer drug doxorubicin in relation to blood vesels in solid tumors. Clin Cancer Res. 2005;11:8782–8. 94. Bertilaccio MTS, Grioni M, Sutherland BW, et al. Vasculature-targeted tumor necrosis factor-alpha increases the therapeutic index of doxorubicin against prostate cancer. Prostate. 2008;68:1105–15. 95. Wechter WJ, Leipold DD, Murray Jr. ED, et al. E-7869 (R-flurbiprofen) inhibits progression of prostate cancer in the TRAMP mouse. Cancer Res. 2000;60:2203–8. 96. Fong TAT, Shawver LK, Sun L, et al. SU5416 is a potent and selective inhibitor of the vascular endothelial growth factor receptor (Flk-1/KDR) that inhibits tyrosine kinase catalysis, tumor vascularization, and growth of multiple tumor types. Cancer Res. 1999;59:99–106. 97. Walter-Yohrling J, Morgenbesser S, Rouleau C, et al. Murine endothelial cell lines as models of tumor endothelial cells. Clin Cancer Res. 2004;10(6):2179–89. 98. Stadler WM, Cao D, Vogelzang NJ, et al. A randomized phase II trial of the antiangiogenic agent SU5416 in hormone-refractory prostate cancer. Clin Cancer Res. 2004;10:3365–70. 99. Isayeva T, Chanda D, Kallman L, Eltoum I-EA, Ponnazhagen S. Effects of sustained antiangiogenic therapy in multistage prostate cancer in TRAMP model. Cancer Res. 2007;67:5789–97. 100. Ganapathy M, Ghosh R, Jianping X, et al. Involvement of FLIP in 2-methoxyestradiolinduced tumor regression in transgenic adenocarcinoma of mouse prostate model. Clin Cancer Res. 2009;15:1601–11. 101. Zorn CS, Wojno KJ, McCabe MT, Kuefer R, Gschwend JE, Day ML. 5-aza-2¢-deoxycytidine delays androgen-independent disease and improves survival in the transgenic adenocarcinoma of the mouse prostate mouse model of prostate cancer. Clin Cancer Res. 2007;13:2136–42.

420

S.D. Morgenbesser

102. Thibault A, Figg WD, Bergan RC, et al. A phase II study of 5-aza-2¢deoxycytidine (decitabine) in hormone independent metastatic (D2) prostate cancer. Tumorigenesis. 1998;84:87–9. 103. Rivard GE, Momparler RL, Demers J, et al. Phase I study on 5-aza-2’¢-deoxycytidine in children with acute leukemia. Leukemia Res. 1981;5:453–62. 104. Qian DZ, Wei Y-F, Wang X, Kato Y, Cheng L, Pili R. Antitumor activity of the histone deacetylase inhibitor MS-275 in prostate cancer models. Prostate. 2007;67:1182–93. 105. Sargeant AM, Rengel RC, Kulp SK, et al. OSU-HDAC42, a histone deacetylase inhibitor, blocks prostate tumor progression in the transgenic adenocarcinoma of the mouse prostate model. Cancer Res. 2008;68:3999–4009. 106. Narayanan BA, Narayanan NK, Pittman B, Reddy BS. Regression of mouse prostatic intraepithelial neoplasia by nonsteroidal anti-inflammatory drugs in the transgenic adenocarcinoma mouse prostate model. Clin Cancer Res. 2004;10:7727–37. 107. Kolluri SK, Corr M, James SY, et al. The R-enantiomer of the nonsteroidal antiinflammatory drug etodolac binds retinoid X receptor and induces tumor-selective apoptosis. Proc Natl Acad Sci USA. 2005;102:2525–30. 108. Harzstark AL, Small EJ. Immunotherapeutics in development for prostate cancer. The Oncologist. 2009;14:391–8. 109. Garcia-Hernandez MDLL, Gray A, Hubby B, Klinger OJ, Kast WM. Prostate stem cell antigen vaccination induces a long-term protective immune response against prostate cancer in the absence of autoimmunity. Cancer Res. 2008;68:861–9. 110. Thomas-Kaskel AK, Zeiser R, Jochim R, et al. Vaccination of advanced prostate cancer patients with PSCA and PSA peptide-loaded dendritic cells induces DTH responses that correlate with superior overall survival. Int J Cancer. 2006;119:2428–34. 111. Hurwitz AA, Foster BA, Kwon ED, et al. Combination immunotherapy of primary prostate cancer in a transgenic mouse model using CTLA-4 blockade. Cancer Res. 2000;60:2444–8. 112. Wada S, Yoshimura K, Hipkiss EL, et al. Cyclophosphamide augments antitumor immunity: studies in an authochthonous prostate cancer model. Cancer Res. 2009;69:4309–18. 113. Harris TJ, Hipkiss EL, Borzillary S, et al. Radiotherapy augments the immune response to prostate cancer in a time-dependent manner. Prostate. 2008;68:1319–29. 114. Degl’Innocenti E, Grioni M, Boni A, et al. Peripheral t cell toleratnce occurs early during spontaneous prostate cancer development and can be rescued by dendritic cell immunization. Eur J Immunol. 2005;35:66–75. 115. Varghese S, Rabkin SD, Nielsen GP, MacGarvey U, Liu R, Martuza RL. Systemic therapy of spontaneous prostate cancer in transgenic mice with oncolytic herpes simplex virus. Cancer Res. 2007;67:9371–9. 116. De Giovanni C, Croci S, Nicolett G, et al. Inhibition of prostate carcinogenesis by controlled active immunoprophylaxis. Int J Cancer. 2007;121:88–94. 117. Bhatt RS, Bubley GJ. The challenge of herbal therapies for prostate cancer. Clin Cancer Res. 2008;14:7581–2. 118. Raina K, Blouin M-J, Singh RP, et al. Dietary feeding of silibin inhibits prostate tumor growth and progression in transgenic adenocarcinoma of the mouse prostate model. Cancer Res. 2007;67:11083–91. 119. Raina K, Rajamanickam S, Singh RP, Deep G, Chittezhath M, Agarwal R. Stage-specific inhibitory effects and associated mechanisms of silibin on tumor progression and metastasis in transgenic adenocarcinoma of the mouse prostate model. Cancer Res. 2008;68:6822–30. 120. Singh RP, Raina K, Sharma G, Agarwal R. Silibin inhibits established prostate tumor growth, progression, invasion, and metastasis and suppresses tumor angiogenesis and epithelialmesenchymal transition in transgenic adenocarcinoma of the mouse prostate model mice. Clin Cancer Res. 2008;14:7773–80. 121. Flaig TW, Gustafson DL, Su L-J, et al. A phase I and pharmacokinetic study of silybinphytosome in prostate cancer patients. Invest New Drugs. 2007;25:139–46. 122. Raina K, Rajamanickam S, Singh RP, Agarwal R. Chemopreventive efficacy of inositol hexaphosphate against prostate tumor growth and progression in TRAMP mice. Clin Cancer Res. 2008;14:3177–84.


421

123. Mentor-Marcel R, Lamartiniere CA, Eltoum I-EA, Greenberg NM, Elgavish A. Genistein in the diet reduces the incidence of poorly differentiated prostatic adenocarcinoma in transgenic mice (TRAMP). Cancer Res. 2001;61:6777–82. 124. Fasso M, Waitz R, Hou Y, et al. SPAS-1 (stimulator of prostatic adenocarcinoma-specific T cells)/SH3GLB2: a prostate tumor antigen identifiedy by CTLA-4 blockade. Proc Natl Acad Sci USA. 2008;105:3509–14. 125. Isayeva T, Moore LD, Chanda D, Chen D, Ponnazhagen S. Tumoristatic effects of endostatin in prostate cancer is dependent on androgen receptor status. Prostate. 2009;69(10):1055–66. 126. Zha S, Yegnasubramanian V, Nelson WG, Isaacs WB, De Marzo AM. Cyclooxygenases in cancer: progress and perspective. Cancer Lett. 2004;215:1–20.

Chapter 18

The Utility of Transgenic Mouse Models for Cancer Prevention Research Stephen D. Hursting, Laura M. Lashinger, Powel H. Brown, and Susan N. Perkins

Abstract The development of effective cancer preventive interventions is being enhanced by the use of relevant animal models to confirm, refine and extend potential leads from clinical and epidemiologic studies. In particular, genetically altered mice, with specific cancer-related genes modulated, are providing powerful tools for studying carcinogenesis, as well as important conduits for translating basic research findings from the laboratory bench to the bedside. This review explores the utility of genetically altered mice for developing cancer preventive strategies that can offset increased cancer susceptibility resulting from specific genetic lesions. Examples will focus on preventing prostate, mammary, intestinal and pancreatic cancers by dietary interventions, particularly obesity prevention/ energy balance modulation, as well as chemoprevention, in mice with alterations in genes such as the p53 or Apc tumor suppressors, components of the Wnt, ErbB/Ras oncogenic pathways, and other pathways frequently altered in human cancer.

18.1 Introduction The development of genetically altered mouse models for cancer research over the past three decades has greatly facilitated efforts in understanding tumor biology and identifying the role of specific genes in carcinogenesis as well as in normal development [1]. These models, with specific cancer-related genes altered, also provide an intact and highly relevant biological system for evaluating the efficacy and underlying mechanisms of cancer preventive interventions [2]. There are numerous examples of how the development of effective cancer preventive interventions has been augmented by the use of relevant genetically altered mouse models to confirm, refine, and extend S.D. Hursting (*) Department of Nutritional Sciences, University of Texas at Austin, Austin, TX, USA and Department of Carcinogenesis, University of Texas MD Anderson Cancer Center, Smithville, TX, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_18, © Springer Science + Business Media, LLC 2011

423

424

S.D. Hursting et al.

potential leads from epidemiologic and clinical research. In particular, genetically altered mice are providing powerful tools for the preclinical evaluation of preventive dietary regimens or pharmacological agents (used singly or in combination). In addition, mutant mice engineered to mirror the genetic alterations characteristic of human tumors are facilitating the identification and validation of biomarkers for early detection of cancer. This chapter will discuss commonly used models of mutant mice for developing cancer preventive strategies. Examples will be provided from studies centered on preventing cancer by dietary (particularly energy balance modulation) and chemopreventive interventions, in a variety of strains of mutant mice. These include transgenic mice overexpressing an oncogene or other cancer-related gene; mice with targeted germ-line deletions of tumor suppressors or other key growth regulatory or metabolic genes; and mice with conditional alterations in targeted cancer-related genes.

18.2 Cancer Prevention in Transgenic Mice: Lessons Learned from Commonly Used Models 18.2.1 Mutant Mouse Models for Prostate Cancer Prevention Prostate cancer prevention research has been hampered by the lack of models that develop lesions analogous to those observed in human prostate cancer progression. Fortunately, numerous transgenic and knockout models have now been generated that develop lesions ranging from murine prostatic intraepithelial neoplasia (mPIN) to metastatic disease. In general, multiple signaling pathways must be altered in order to produce invasive lesions (adenocarcinoma) in the mouse, and mutant mouse models of prostate cancer can be broadly divided into two categories: (1) models with genetic modification of pathways implicated in human prostate cancer development, including PTEN-deficient and c-myc transgenic mice; and (2) models in which viral oncogene(s) are expressed in prostate tissues, particularly the Simian Virus (SV) 40 large T antigen. 18.2.1.1 PTEN Mutant Mouse Models PTEN (phosphatase and tensin homolog deleted on chromosome 10) is a tumor suppressor gene that is frequently mutated or lost in human cancers. The PTEN gene encodes for a lipid phosphatase, the major function of which is the dephosphorylation of phosphatidylinositol-3-phosphate, leading to downregulation of Akt/PKB [3] and other signaling molecules [4]. Deletion or mutation of PTEN is one of the most frequently observed genetic alterations in human prostate cancer, affecting up to 63% of metastatic tumors [5]. Attempts to generate knockout mouse models with conventional homozygous mutation of PTEN

18 The Utility of Transgenic Mouse Models for Cancer Prevention Research

425

resulted in an early embryonic lethal phenotype, and therefore only heterozygous PTEN-deficient (PTEN+/-) progeny could be assessed for prostate tumor development. PTEN+/- mice exhibit mPIN by 8–10 months of age, but display no progression beyond mPIN. However, PTEN has been shown to interact with other genetic alterations to induce prostatic lesions. Most notable among these are the PTEN+/- x Cdkn1b-/- and PTEN+/- x Nkx3.1-/- mutant mice. The combined heterozygous knockout of PTEN and homozygous knockout of Cdkn1b (p27) results in invasive adenocarcinoma in the prostate, as well as tumors in other tissues [6]. Knocking out Nkx3.1 expression in PTEN heterozygotes similarly causes invasive prostatic adenocarcinoma, primarily in the anterior prostate lobe, and metastasis to regional lymph nodes [7]. PTEN-floxed mice crossed with probasin-CRE transgenic mice results in a mouse line with homozygous deletion of the PTEN gene in prostate cells expressing the Cre transgene [8]. These mice have 100% incidence of mPIN by 6 weeks of age and invasive adenocarcinoma by 12 weeks of age in all prostatic lobes, with an incidence of metastasis to the lymph nodes and lung by 30 weeks of age. Prostate adenocarcinomas in this model are responsive to androgen ablation, suggesting that this model will be particularly useful for prostate cancer prevention studies targeting androgen signaling. 18.2.1.2 c-Myc Transgenic Mice The nuclear protein c-myc is a transcription factor that regulates proliferation and apoptosis. Myc expression is elevated in up to 30% of human prostate tumors. Ellwood-Yen et al. generated two founder lines of transgenic mice overexpressing c-myc. High-expressing (Hi-Myc) and low-expressing (Lo-Myc) mice, generated using the composite probasin promoter to drive expression of c-myc cDNA in the prostate, have similar phenotypes, with the Hi-Myc having a much shorter latency for the onset of preneoplastic and neoplastic changes in the prostate [9]. From 2 to 4 weeks of age, mPIN lesions appear in Hi-Myc mice, and these lesions progress to invasive adenocarcinomas by 3–6 months of age. Tumors occur primarily in the ventral and dorsolateral prostate lobes in these mice. 18.2.1.3 Viral Oncogene Models Prostate tumorigenesis in TRAMP (transgenic adenocarcinoma mouse prostate) mice is driven by prostate-specific expression of SV40 early genes (large and small T antigens). TRAMP mice develop histological mPIN in 100% of males by 8 weeks of age that progresses to poorly differentiated carcinoma with distant site metastasis by 16–32 weeks of age [10, 11]. Metastatic sites include the lymph nodes, liver, and lung and occasionally the kidney, adrenal glands, and spinal column [10]. These tumors are initially sensitive to androgen ablation, but develop androgen independence following castration [12].

426


The precisely defined course of disease progression and the overall high incidence of tumors in the TRAMP model allow prevention studies to be conducted in a reasonable length of time with fewer animals. Although the TRAMP model has been criticized as being too aggressive, and therefore perhaps not sufficiently sensitive to the actions of chemopreventive agents, there is little evidence to support this hypothesis. The TRAMP model has been utilized by many investigators for preclinical testing of chemoprevention strategies, including several positive studies with agents such as epigallocatechin gallate (a green tea polyphenolic compound), genistein (a soy isoflavone), toremifene (an estrogen modulator), difluoromethylornithine (DFMO, an irreversible inhibitor of ornithine decarboxylase), and the antiinflammatory drugs R-flurbiprofen and celecoxib [13–18]. Background strain can have a substantial effect on the prostatic phenotype resulting from expression of a transgene. For example, TRAMP mice are typically maintained on a pure C57BL/6 background. When C57BL/6 TRAMP mice are crossed with wild-type FVB mice, the resulting TRAMP[B6xFVB] F1 progeny exhibit an enhanced phenotype compared to the parental C57BL/6 strain [19]. Chemoprevention studies have been carried out using TRAMP mice of either strain background, but the strain used most often is the pure C57BL/6 TRAMP, even though TRAMP[B6xFVB] F1 mice have been more thoroughly characterized and have a number of advantages. Tumors in TRAMP[B6xFVB] F1 mice tend to form well-defined tumor nodules that can be externally palpated and easily dissected from adjacent tissues for weighing and analyses. Tumors in the C57BL/6 TRAMP mice are more diffuse and invade the seminal vesicles at a higher frequency, often resulting in occlusion. In addition, the incidence of phyllodes-like lesions is much higher in C57BL/6 TRAMP mice; these lesions do not appear to be relevant to human prostatic adenocarcinoma. There have been no direct comparisons of chemoprevention agent effects on the two different strain types. Therefore it is not known if there could be substantial differences between the strain types in terms of response to relevant chemoprevention agents. The incidence of early carcinoma in TRAMP is 100% and occurs within a very confined and predictable time period (6–8 weeks of age). Progression to poorly differentiated carcinoma is more stochastic, and occurs over a much broader time period (16–32 weeks of age). Length of treatment varies considerably, with some researchers conducting survival studies and other researchers collecting prostate tissues prior to the development of poorly differentiated tumors [20]. Treatments intended to prevent the progression of mPIN should be started between 4 and 6 weeks of age, while treatments designed to prevent progression of locally invasive carcinoma to poorly differentiated and metastatic disease should be started before 16 weeks of age. Typical studies designed to prevent progression to metastatic disease should include approximately 30 mice per treatment group in order to provide sufficient power to detect differences between treatment groups. The pathogenesis of prostate cancer in the Lady mouse model, or line 12T-10 of a group of long probasin promoter driven large T-antigen transgenic mice, is similar to that observed in the TRAMP model [21]. Lesions progress from mPIN to invasive adenocarcinoma, followed by conversion to metastatic carcinoma. The progression


427

from PIN to metastasis occurs at a somewhat lower rate than in the TRAMP model, so the Lady-12T-10 model may be considered slightly less aggressive than the TRAMP model. Venkateswaran and coworkers evaluated the effects of antioxidants and dietary fat levels on prostate cancer progression in Lady mice [21]. Interestingly, a high-fat diet increased the incidence of poorly differentiated tumors at 32 weeks of age from 74 to 100%. Addition of lycopene, selenium, and vitamin E to the diets reduced tumor incidence in response to both the low- and high-fat diets.

18.2.2 Mutant Mouse Models for Mammary Cancer Prevention Numerous transgenes have been used to generate mouse models to mimic human breast cancer. Most transgenic mouse models used for prevention studies were generated through gain of function or knockout of critical components in oncogenic pathways. The most commonly used models for breast cancer target growth factors or their receptors, cell-cycle regulators, signal transduction pathways, cellular differentiation regulators, oncogenes, and tumor suppressor genes. Most of these use tissue-specific promoters, including mouse mammary tumor virus-long terminal repeat (MMTV-LTR), whey acidic protein (WAP), the C(3)1 component of the rat prostate steroid binding protein, or bovine b-lactoglobulin (BLG) to drive mammary gland expression of the transgene. Each promoter has strengths and limitations, and the choice of promoter is dependent on the research question being addressed. 18.2.2.1 TGFa Models The transcription factor TGFa plays an important role in mammary development and is overexpressed in 30–70% of breast cancer cases, as reviewed by Rudland [22]. TGFa expression has been driven under MMTV-LTR, WAP, and metallothionein promoters [23–26]. The WAP-TGFa model has been shown to have diffuse mammary epithelial hyperplasia in pregnancy, multifocal hyperplastic alveolar nodules at latency of 2–6 months, and mammary tumors at 6–12 months [25–27]. Yet, the TGFa-induced mammary tumors are focal and relatively fewer in number [25], indicating additional tumorigenic mechanisms are needed to promote tumor development. 18.2.2.2 ErbB-2/HER2/neu Models ErbB2 (HER2, neu) is one of the most intensively studied genes in breast cancer biology. The gene is amplified in 15–20% of human breast cancer and is overexpressed in approximately 30% of breast cancers [28, 29]. ErbB-2 is an indicator for clinical prognosis, metastasis, and tamoxifen resistance [30–32]. ErbB-2 has been engineered to express under MMTV and WAP promoters. Wild-type and mutated ErbB-2 transgenic mice develop mammary tumors in several strains around

428


7 months of latency [33–38]. Multiple transgenic models have confirmed that the early ErbB-2 model carries a valine-to-glutamic acid substitution in the transmembrane domain that confers constitutive activation of the receptor in the absence of ligand [39]. More relevant to human breast pathology, a late wild-type ErbB-2 model develops mammary tumors that carry sporadic mutations in the transgene in the tumor, but not in the adjacent normal mammary tissue [38]. The mammary tumors in ErbB-2 transgenic mice are estrogen receptor (ER) negative, and their pathologic appearance resembles lobular and alveolar phenotypes, found in about 5% of human breast cancers [38]. 18.2.2.3 SV40 T-antigen Transgenic Models SV40 large and small T-antigens (SV40 Tag) induce mammary tumors by inactivating the p53 and Rb tumor suppressor genes. When SV40 Tag is expressed using the promoter C3(1) from the prostate steroid binding protein, the transgene induces mammary carcinomas in 100% of female mice and prostate tumors in all male mice [40–42]. The C3(1) model has several unique characteristics for breast cancer prevention studies. The model mimics a well-defined time course for progressive mammary lesions and tumorigenesis: atypical ductal epithelia at 8 weeks, mammary intraepithelial neoplasia (similar to human ductal carcinoma in situ (DCIS)) at 12 weeks, and invasive carcinoma by 16 weeks of age [40]. Most interestingly, virgin C3(1) mice all develop tumors without the need of additional hormonal stimulation from pregnancy, a superior attribute over several other transgenic models. Another valuable feature of this model is that the C3(1) promoter itself is not stimulated by estrogen or pregnancy, and the tumors are ER negative and estrogen independent. Therefore, the C3(1) model is especially useful for studying ER-negative mammary tumorigenesis. The SV40 Tag has also been expressed using the WAP promoter [43–45]. In this model, all female mice develop mammary tumors by 8–9 months of age. Histologic appearance of the tumor varies from wellto poorly differentiated phenotypes. Pregnancy enhances the tumor development due to the WAP promoter, and the first tumors appear at 6 months of age after one pregnancy. Similar to the C3(1) model, tumor development in the WAP T-antigen model is characterized by three distinct stages: initial proliferation, hyperplasia, and adenocarcinoma [46]. An interesting attribute of this model is the high level of proliferation, apoptosis, and fibrosis in the tumor. This model is potentially useful to explore the early events during mammary tumorigenesis, particularly with respect to cellular proliferation and cell death. 18.2.2.4 p53-mutant Mouse Models Alterations of the p53 tumor suppressor gene are frequently detected in human breast cancer, with up to 40–50% of all human breast cancers having p53 mutations [47]. Mutations or epigenetic alterations of the p53 tumor suppressor gene are commonly


429

observed in human cancer [48]. Donehower and colleagues [49] first reported in 1992 that homozygous p53-knockout (p53-/-) mice are viable, but susceptible to spontaneous tumorigenesis (particularly lymphomas) at an early age. Heterozygous p53-deficient (p53+/-) mice, with only one p53 allele inactivated, have some analogy to humans susceptible to heritable forms of cancer due to decreased p53 gene dosage, such as individuals with Li-Fraumeni Syndrome [50]. We and others have used these mice for evaluating the influence of diet or chemopreventive agents on genetic susceptibility to tumor development due to p53-deficiency [51–56]. However, the germline p53-mutant mice only rarely develop mammary tumors. Several animal models have been developed that either overexpress a mutant p53 gene in mammary tissue or have the endogenous p53 gene deleted or disrupted in mammary gland cells [57–59]. Mammary tumors are infrequently observed in homozygous p53-deficient (p53-/-) mice because the mice first develop lymphoma and die of these tumors before development of mammary gland tumors. Medina et al. developed a transplantable BALB/c-p53-/- mammary epithelium and demonstrated that lack of p53 function is sufficient to cause mouse mammary tumorigenesis, though hormone stimulation is an effective enhancer for the p53-/--induced tumorigenesis [60]. A WAP-p53172R-H transgenic mouse model was developed in which p53172R-H functions as a dominant-negative mutant [57, 59]. The WAP-p53172R-H mice develop tumors with shorter latency after DMBA treatment. Mice generated from crossing MMTV-ErbB2 with p53172R-H mice show significantly reduced latency [59]. An important characteristic of this model is that it develops mammary tumors similar to human high-grade breast adenocarcinoma in the presence of carcinogens and oncogenes. Thus, WAP-p53172R-H accelerates carcinogen- and oncogenemediated tumorigenesis, and is useful for cancer preventive intervention. 18.2.2.5 MMTV-Wnt-1 Transgenic Mouse Model Wnt-1 was originally found to be activated after MMTV infection, and the resulting mice had a high incidence of mammary tumors [61]. Wnt-1 is a glycoprotein that signals through the b-catenin pathway. Its expression is seen throughout mammary gland development, and deregulation of the downstream effectors in the Wnt-1 signaling pathway is involved in the tumorigenesis of several tumor types, including breast cancer [62]. MMTV-Wnt-1 expression causes ductal hyperplasia in late gestation and in prepubertal mice [63]. The Wnt-1 mice develop adenocarcinoma at 6–12 months of age [63, 64]. These tumors demonstrate a moderately differentiated ER-negative phenotype and are heterogeneous in ER-positive and/or ER-negative status. The MMTV-Wnt-1 mice have been crossed with MMTV-Fgf3, Sky-/-, p53-/-, ERa-/-, and TGFb mice [61]. There is a synergistic effect between Wnt-1 and Fgf3 [64], as these bigenic animals show shortened latency to developing mammary tumors. The female offspring of p53-/- mice bred with Wnt-1 mice develop mammary tumors significantly faster than their p53+/- counterparts [65]. Metastasis in Wnt-1 mice occurs to lymph node and lung, even after the primary tumors are removed. Therefore, the MMTV-Wnt-1 model is highly relevant to human breast cancer in two aspects: stromal signaling, characteristic of the

430


MMTV-Wnt-1 model, is important in breast tumorigenesis, since the human mammary gland has a significant proportion of stromal structure; and the metastatic route is similar to that of human breast cancer. In addition, MMTV-Wnt-1 mammary adenocarcinomas are not only reflective of human basal-like breast cancer gene signatures and pathologic characteristics, but also are highly enriched in cells expressing the CD44+/CD24low cell surface markers that correspond to mammary tumor-initiating cells in both human tumors and mouse models (Smith, Hursting, et al., personal communication [66, 67]). 18.2.2.6 Ras Mutant Models Ras mutation is infrequent in breast cancer [68]. However, wild-type ras is significantly activated in breast cancers overexpressing epidermal growth factor receptor (EGFR) and/or ErbB-2 [69]. Ras driven by WAP and MMTV is sufficient to induce hyperplasia, adenocarcinoma, accelerated tumorigenesis, and metastatic mammary tumors [70–73]. MMTV-Ha-ras transgenic mice develop mammary tumors from 5 weeks to 6 months of age [74]. 18.2.2.7 c-myc Transgenic Mice c-myc is a transcription factor that dimerizes with Max and regulates target gene promoters. A defined role for c-myc has been shown in cell-cycle regulation and apoptosis. c-myc regulates normal mammary development and hormone-related proliferation, and also controls involution and remodeling [75]. Further, c-myc is deregulated in many human breast cancers. The c-myc gene is amplified in approximately 15–20% of all human breast cancers and is overexpressed in up to 70% of breast cancers [76]. Several c-myc transgenic models have been developed in which the c-myc gene is expressed using MMTV or WAP promoters. The mice for each of these models develop mammary tumors at a high rate [25, 77, 78]. MMTV-c-myc mice develop spontaneous mammary adenocarcinomas within 4–8 months [78]. WAP-c-myc mice develop adenocarcinomas or solid carcinomas in 80% of female transgenic mice after multiparity, with a latency of 5–10 months [25, 77]. These c-myc-induced mammary tumors are ER-negative tumors. It is important to note that, as suggested by the long latency, the c-myc overexpression does not transform all mammary epithelial cells. This suggests that additional events are required for c-myc-induced transformation of mammary cells. In this regard, the c-myc model reflects the attributes of human breast carcinogenesis and hence is a potentially ideal mouse model for cancer preventive intervention. 18.2.2.8 Cyclin D1 Transgenic Mice Cyclin D1 is amplified in about 20% of human breast cancers [79], while the cyclin D1 protein is overexpressed in more than 50% of human breast cancers


431

[80–82]. In addition, loss of cyclin D1 interferes with mammary tumorigenesis. Sicinski and coworkers crossed cyclin D1-/- mice to four different mammary oncomice and found that cyclin D1 mediated mammary tumors induced by MMTV-c-neu and MMTV-v-Ha-ras, but not by MMTV-c-myc and MMTV-Wnt-1, suggesting that cyclin D1 is essential for the neu-ras pathway and the tumors dependent on cyclin D1 [83]. Cyclin D1-overexpressing breast cancers have been modeled by Wang et al. who developed an MMTV-cyclin D1 transgenic model. These mice have enhanced proliferation of mammary epithelial cells and develop mammary carcinomas at a mean age of 18 months [84]. Therefore, the cyclin D1 transgenic mouse models a significant proportion of human breast cancers and thus may be useful to study mammary carcinogenesis. 18.2.2.9 Inducible Models Inducible expression of transgenes has been used for mammary gland studies for more than a decade, although Chodosh’s more recent development of a reverse tetracycline-dependent transcriptional activator (rtTA) system with MMTV promoter achieves mammary-specific, tightly regulated homogeneous transgene expression in the presence of tetracycline or its derivative doxycycline. Using this system, the c-myc transgene was specifically induced in mammary epithelial cells [85, 86]. This system, although cumbersome because of the requirement of at least two transgenes, is highly mammary gland specific and inducible and has great potential for future cancer prevention studies. The development of the Cre-loxP strategy as an inducible and regulatable mammary gland-specific expression system is also providing a powerful approach for prevention studies. In this system, the Cre recombinase gene is under the control of MMTV or WAP promoters. Expression of the Cre gene causes conditional deletion of specific target genes. For example, deletion of the Brca1 gene by this system results in abnormal ductal development and activated apoptosis specifically in the mammary gland [87]. 18.2.2.10 Examples of Dietary or Chemopreventive Studies Using Transgenic Mouse Models of Mammary Cancer Selective Estrogen Receptor Modulators While classic selective estrogen receptor modulators (SERMs) such as tamoxifen and raloxifene are now being examined in the NASBP Study of Tamoxifen and Raloxifene (STAR) trial to compare their efficacy and safety profiles [88], other hormoneregulating agents are also being tested in animal models. The human clinical trials show that SERMs are able to prevent only ER-positive tumor formation. However, in preclinical studies using MMTV-ErbB2 mice, mammary tumor incidence was reduced significantly in mice given tamoxifen at an earlier age (8–18 weeks of age)

432


[89, 90]. In addition, a combination of tamoxifen and angiostatin achieved greater suppression of tumor growth than tamoxifen or angiostatin alone [91]. A further combination of tamoxifen, angiostatin, and TIMP-2 achieved 90% reduction of tumor incidence in the MMTV-ErbB-2 model [92]. These results suggest that, in some cases, anti-estrogenic SERMs can suppress the development of ER-negative cancers. The underlying mechanism is unknown at this time. Aromatase Inhibitors Aromatase is a key enzyme in the synthesis of endogenous estrogen in peripheral tissue. The transgenic model overexpressing aromatase demonstrates increased tissue estrogenic activity and induction of hyperplastic and dysplastic lesions in mammary glands with or without circulating estrogen [93, 94]. These preneoplastic changes appeared to be further stimulated by the carcinogen dimethylbenz[a] anthracene, leading to an increased incidence of mammary tumors in mice. Lowdose letrozole, an aromatase inhibitor, inhibits expression of ER, progesterone receptor, cell-cycle regulators, and reduces mammary cell hyperplasia and the index of the proliferation marker PCNA [93, 94]. These studies have provided a vivid example of how to use a transgenic mouse model to elucidate important underlying mechanisms of mammary tumorigenesis. Retinoids Retinoids are Vitamin A analogs that mediate transcriptional regulation with their receptors RAR and RXR. Studies in our laboratory have demonstrated that RXRselective retinoids are much less toxic than RAR-selective retinoids. LGD1069 (Bexarotene, Targretin), an RXR-selective retinoid, prevents ER-negative mammary tumors in C3(1) SV40T and MMTV-ErbB-2 transgenic mouse models [95, 96]. Another, newer RXR-selective retinoid, LG100268, has been reported recently by Suh and colleagues to reduce tumor incidence in the NMU-induced mammary cancer rat model by promoting TGFb-dependent apoptosis [97, 98]. The most striking finding in these studies is that when LG 100268 was used in combination with a third-generation SERM, arzoxifene, only very low dosages of both arzoxifene and LG100268 were needed to cause significant reduction of tumor burden [97]. Similar results were obtained using the MMTV-ErbB2 model. Tyrosine Kinase Inhibitors EGFR (HER1, ErbB-1) or other members of its receptor family (HER2, 3, and 4) are overexpressed in a portion of human breast cancers and are highly expressed in ER-negative tumors [99]. Tyrosine kinase inhibitors (TKIs) can effectively block the tumorigenic potentials that arise from the EGF signaling pathway. ZD1839


433

(IRESSA) is the prototype of this class of drugs [100]. Recent work in our laboratory has demonstrated that this drug prevents ER-negative tumor formation in MMTV-ErbB2 mice. The median time to tumor formation was approximately 230 days in vehicle-treated mice and more than 310 days in mice treated with ZD1839 at 100 mg/kg (P < 0.001). This effect was achieved by reducing proliferation and increasing expression of the cell-cycle regulator p27 [101]. Nonsteroidal Anti-Inflammatory Drugs/Cyclooxygenase-2 Inhibitors One most promising new class of chemopreventive agents is the cyclooxygenase (COX)-2 inhibitors. COX-2 is one of the rate-limiting enzymes in converting free arachidonic acid to prostaglandin (PG)G2. The two downstream products PGE1 and PGE2 enhance mitogenesis in mammary cells stimulated with EGF [102]. COX-2 is overexpressed in 56% of breast cancers, including DCIS as well as infiltrating ductal and lobular carcinomas [102, 103]. Mammary glands from transgenic, the MMTV-COX-2 mouse model, demonstrate hyperplasia, dysplasia, and development of metastatic tumors [104]. The specific COX-2 inhibitor, celecoxib, is currently being tested for its ability to prevent cancer in humans. When given at 500 ppm, celecoxib significantly suppresses tumor incidence and PGE2 levels in the MMTV-ErbB-2 model [105]. This drug is also under evaluation in our laboratory using other transgenic models. Energy Balance Interventions Obesity is an established risk factor for postmenopausal breast cancer and is associated with poor prognosis for both pre- and postmenopausal breast cancers. A number of transgenic mouse models have been used to assess the impact of obesity or energy balance interventions, including calorie restriction or exercise, on mammary tumor development or progression, as recently reviewed [106].

18.2.3 Apc-Deficient Models for Intestinal Cancer Prevention Studies The ApcMin mouse model [107] is an excellent example of the utility of animal models in understanding the connection between energy balance and genetic susceptibility to cancer, as well as for testing chemopreventive agents with potential anticancer effects against intestinal cancer. For example, intestinal polyp burden was significantly reduced in ApcMin mice following both a 4-week and a 10-week 40% calorie restriction regimen [108]. Additionally, insulin-like growth factor (IGF-1) levels, as well as inflammatory markers, were reduced in calorie restricted mice [108]. Further, an olive oil based diet high in fruits and vegetables reduced overall polyp numbers

434


in the 10-week intervention, suggesting a role for dietary based anti-inflammatory compounds in cancer prevention [108]. We also showed that running wheel exercise via a treadmill demonstrated no overall effect on polyp burden in ApcMin mice [109]. However, when male and female mice were analyzed separately, running wheel exercise led to fewer intestinal and total polyps in male mice [109]. A prime example of the efficacy of combination therapies is the administration of both celecoxib and immunotherapy, specifically a carcinoembryonic antigen (CEA)-based vaccine, in CEA.Tg/ApcMin mice [110]. A large majority of the CEA. Tg/ApcMin mice receiving the combination intervention remarkably remained polyp free for more than 2 years, resulting in better health status and increased survival [110]. This combination regimen is of particular relevance in high-risk individuals, such as those with the genetic syndromes of familial adenomatous polyposis, a hereditary syndrome in which individuals develop hundreds of adenomatous polyps, and hereditary nonpolyposis colorectal cancer. Treatment with celecoxib reduced polyp numbers in approximately 50% of patients with familial adenomatous polyposis studied over a 6-month period [111], but there were many nonresponders. The potential to combine celecoxib treatment with a cancer vaccine or other intervention in this high-risk population based on the results from preclinical animal studies is very encouraging. This model appears to be particularly sensitive to the preventive activity of anti-inflammatory agents such as nonsteroidal anti-inflammatory drugs (NSAIDs) and selective COX-2 inhibitors. For example, studies with the selective COX-2 inhibitors nimesulide [112] and celecoxib [113] in ApcMin mice demonstrated efficacy against intestinal tumorigenesis. Findings from these studies helped support human trials of celecoxib in individuals with familial adenomatous polyposis [111], and the combined mouse and human data encouraged the decision by the US Food and Drug Administration to approve this class of agent in persons with familial adenomatous polyposis. The findings that at least some COX-2 inhibitors at high doses increase risk of myocardial infarction has limited the use of these drugs as long-term cancer prevention agents [114]. However, their anti-cancer effects are unquestioned, providing a striking example of the development and translation of preclinical findings in relevant genetically altered mouse models to clinical application.

18.2.4 Emerging Models of Pancreatic Cancer Historically, pancreatic cancer research relied heavily on both orthotopic and chemically induced models of pancreatic cancer. Orthotopic models, unfortunately, constrain the field to investigations examining tumor progression, which are relevant to translational studies, but have limited utility for cancer prevention studies. Chemically induced models rarely develop pancreatic ductal adenocarcinomas (PDAC), the most common histological type of human pancreatic cancer, and thus they can provide little insight into the genetic alterations characteristic of


435

PDAC [115]. Early attempts at transgenic modeling of exocrine-specific pancreatic cancer employed promoters that directed acinar-specific transgene expression, i.e., elastase [116, 117] and Mist1 [118]. Although most of these transgenic models produce acinar cell neoplasms, some do develop an acinar-ductal metaplasia that results in mixed acinar ductal neoplasias with some clinical relevance. The subsequent generation of transgenic pancreatic cancer mouse models owes its advent to cumulative breakthroughs, including discovery of the genetic basis of PDAC, identification of pancreatic intraepithelial neoplasia (PanIN) as the neoplastic precursor to PDAC, and more recently the development of promoters specific to particular pancreatic cell lineages [119]. Consequently, clinically relevant mouse models have been developed that exploit the tumorigenic nature of activating K-ras mutations, the defining genetic lesion of PDAC. Hingorani et al. established the KrasG12D Pdx-1Cre model, which results in an array of mouse PanIN (mPanIN) lesions with a protracted latency period of PDAC, nicely mimicking the accumulation of K-ras anomalies in human disease [120]. In the pancreas, coupling K-ras activation with an Ink4a/Arf deletion [121] and/or a p53 point mutation [122] dramatically potentiates the aggressiveness and invasiveness of the extensive mPanIN lesions. In fact, invasive PDAC, which readily metastasizes, develops in KrasG12D Pdx-1Cre Ink4a/Arf-/- mice within 7–11 weeks. Similarly, KrasG12D Pdx-1Cre p53R273H [122] or p53+/- [123] mice form strikingly aggressive lesions with a short latency period. These mouse models have corroborated the sequential nature of genetic anomalies outlined by the step-wise carcinogenesis scheme, making them all extraordinarily relevant models. However, some may be too aggressive to effectively study pancreatic cancer prevention. The longer latency period and age-related progression of the KrasG12D Pdx-1Cre mouse model are characteristics that would offer an opportunity for modulation. In fact, investigators showed that a COX-2 inhibitor, nimesulide, hinders progression of mPanIN precursor lesions in these mice [124]. The link between enhanced COX-2 expression and pancreatic cancer is well established and has recently been exploited in the creation of another pertinent mouse model. Transgenic expression of the inflammation-induced enzyme COX-2, driven by the bovine keratin 5 promoter, in the BK5.COX-2 mouse model results in an extensive matrix of fibrosis, inflammatory cell infiltrate, and acinar-ductal metaplasia with 100% penetrance that leads to significant lesion development and the presence of adenocarcinoma [125]. Histological progression strongly recapitulates the evolution of chronic pancreatitis to human pancreatic cancer. Although moribundity occurs in these mice between 6 and 8 months of age, the authors demonstrated that treatment with the selective COX-2 inhibitor celecoxib abrogates the influence of COX-2 overexpression and significantly extends survival [125]. This model has also proven to be exquisitely sensitive to the anti-cancer effects of calorie restriction (Lashinger, personal communication). In conclusion, the development of genetically engineered mouse models applicable to comparisons with the progressive character of human pancreatic cancer provides exciting opportunities for pancreatic cancer prevention studies.

436


18.3 Summary and Conclusions Experimental models of cancer have been crucial to advancing our understanding of the tumorigenesis process, and recent progress in the field of molecular carcinogenesis has revealed multiple targets (Fig. 18.1) for the nutritional modulation and chemoprevention of cancer. We must now fully utilize this knowledge base, as well as capitalize on the availability of tools such as transgenic mice, to identify critical, modulatable targets and make important progress towards one of the major goals in contemporary cancer prevention research: the development of effective mechanismbased strategies for preventing human cancer. Successful attainment of this goal will require the integration of the very best science from multiple levels of investigation, including animal studies, clinical and epidemiologic research, and basic molecular and cellular biologic research. All three levels of investigation are essential in this effort, although in our view animal model studies play a critical central role. Thus, the availability of highly relevant animal models will greatly facilitate future progress in cancer prevention research. In this review, examples of cancer prevention studies that have utilized genetically altered mouse models were discussed. Numerous mouse models with cancer-related genes overexpressed or inactivated have been developed in recent years and are cataloged in online databases

Fig. 18.1 Molecular targets in mouse models for cancer prevention. Depicted in the cell are s everal key genes and their interacting pathways, which when altered in mice represent important models for studying cancer prevention strategies


437

such as the Induced Mutation Registry Database (http://www.jax.org/imr /index.html) and the Mouse Genome Informatics site (http:// www.informatics.jax.org ) maintained by the Jackson Laboratory. In addition, a systematic cataloging of potential mouse models with pertinent histopathology and other aspects of characterization has been undertaken under the auspices of the NCI-sponsored Mouse Models of Human Cancer Consortia, which can also be viewed on the Internet (http://emice.nci.nih.gov). Many of these models have been used effectively in studies focusing on toxicology and carcinogenesis, and some of the models are also being used for cancer prevention studies. In conclusion, mice with specific (and human-like) genetic susceptibilities for cancer provide powerful tools for developing and testing interventions which may inhibit the process of carcinogenesis in humans.

References 1. Van Dyke T, Jacks T. Cancer modeling in the modern era: progress and challenges. Cell. 2002;108(2):135–44. 2. Hursting SD, Lubet RA. The utility of transgenic mice in cancer prevention studies. Tumor Models Cancer Res 2001:263–74. 3. Cantley LC, Neel BG. New insights into tumor suppression: PTEN suppresses tumor formation by restraining the phosphoinositide 3-kinase/AKT pathway. Proc Natl Acad Sci U S A. 1999;96(8):4240–5. 4. Weng L, Brown J, Eng C. PTEN induces apoptosis and cell cycle arrest through phosphoinositol-3-kinase/Akt-dependent and -independent pathways. Hum Mol Genet. 2001;10(3):237–42. 5. Suzuki H, Freije D, Nusskern DR, Okami K, Cairns P, Sidransky D, Isaacs WB, Bova GS. Interfocal heterogeneity of PTEN/MMAC1 gene alterations in multiple metastatic prostate cancer tissues. Cancer Res. 1998;58(2):204–9. 6. Di Cristofano A, De Acetis M, Koff A, Cordon-Cardo C, Pandolfi PP. Pten and p27KIP1 cooperate in prostate cancer tumor suppression in the mouse. Nat Genet. 2001;27(2):222–4. 7. Abate-Shen C, Banach-Petrosky WA, Sun X, Economides KD, Desai N, Gregg JP, Borowsky AD, Cardiff RD, Shen MM. Nkx3.1; Pten mutant mice develop invasive prostate adenocarcinoma and lymph node metastases. Cancer Res. 2003;63(14):3886–90. 8. Wang S, Gao J, Lei Q, Rozengurt N, Pritchard C, Jiao J, Thomas GV, Li G, Roy-Burman P, Nelson PS, Liu X, Wu H. Prostate-specific deletion of the murine Pten tumor suppressor gene leads to metastatic prostate cancer. Cancer Cell. 2003;4(3):209–21. 9. Ellwood-Yen K, Graeber TG, Wongvipat J, Iruela-Arispe ML, Zhang J, Matusik R, Thomas GV, Sawyers CL. Myc-driven murine prostate cancer shares molecular features with human prostate tumors. Cancer Cell. 2003;4(3):223–38. 10. Gingrich JR, Barrios RJ, Morton RA, Boyce BF, DeMayo FJ, Finegold MJ, Angelopoulou R, Rosen JM, Greenberg NM. Metastatic prostate cancer in a transgenic mouse. Cancer Res. 1996;56(18):4096–102. 11. Kaplan-Lefko PJ, Chen TM, Ittmann MM, Barrios RJ, Ayala GE, Huss WJ, Maddison LA, Foster BA, Greenberg NM. Pathobiology of autochthonous prostate cancer in a pre-clinical transgenic mouse model. Prostate. 2003;55(3):219–37. 12. Gingrich JR, Barrios RJ, Kattan MW, Nahm HS, Finegold MJ, Greenberg NM. Androgenindependent prostate cancer progression in the TRAMP model. Cancer Res. 1997; 57(21):4687–91.

438


13. Mentor-Marcel R, Lamartiniere CA, Eltoum IE, Greenberg NM, Elgavish A. Genistein in the diet reduces the incidence of poorly differentiated prostatic adenocarcinoma in transgenic mice (TRAMP). Cancer Res. 2001;61(18):6777–82. 14. Raghow S, Hooshdaran MZ, Katiyar S, Steiner MS. Toremifene prevents prostate cancer in the transgenic adenocarcinoma of mouse prostate model. Cancer Res. 2002;62(5):1370–6. 15. Wechter WJ, Leipold DD, Murray ED, Jr., Quiggle D, McCracken JD, Barrios RS, Greenberg NM. E-7869 (R-flurbiprofen) inhibits progression of prostate cancer in the TRAMP mouse. Cancer Res. 2000;60(8):2203–8. 16. Gupta S, Ahmad N, Marengo SR, MacLennan GT, Greenberg NM, Mukhtar H. Chemoprevention of prostate carcinogenesis by alpha-difluoromethylornithine in TRAMP mice. Cancer Res. 2000;60(18):5125–33. 17. Gupta S, Hastak K, Ahmad N, Lewin JS, Mukhtar H. Inhibition of prostate carcinogenesis in TRAMP mice by oral infusion of green tea polyphenols. Proc Natl Acad Sci U S A. 2001;98(18):10350–5. 18. Gupta S, Adhami VM, Subbarayan M, MacLennan GT, Lewin JS, Hafeli UO, Fu P, Mukhtar H. Suppression of prostate carcinogenesis by dietary supplementation of celecoxib in transgenic adenocarcinoma of the mouse prostate model. Cancer Res. 2004;64(9):3334–43. 19. Gingrich JR, Barrios RJ, Foster BA, Greenberg NM. Pathologic progression of autochthonous prostate cancer in the TRAMP model. Prostate Cancer Prostatic Dis. 1999;2(2):70–5. 20. Narayanan BA, Narayanan NK, Pittman B, Reddy BS. Regression of mouse prostatic intraepithelial neoplasia by nonsteroidal anti-inflammatory drugs in the transgenic adenocarcinoma mouse prostate model. Clin Cancer Res. 2004;10(22):7727–37. 21. Venkateswaran V, Fleshner NE, Sugar LM, Klotz LH. Antioxidants block prostate cancer in lady transgenic mice. Cancer Res. 2004;64(16):5891–6. 22. Rudland PS, Fernig DG, Smith JA. Growth factors and their receptors in neoplastic mammary glands. Biomed Pharmacother. 1995;49(9):389–99. 23. Matsui Y, Halter SA, Holt JT, Hogan BL, Coffey RJ. Development of mammary hyperplasia and neoplasia in MMTV-TGF alpha transgenic mice. Cell. 1990;61(6):1147–55. 24. Jhappan C, Stahle C, Harkins RN, Fausto N, Smith GH, Merlino GT. TGF alpha overexpression in transgenic mice induces liver neoplasia and abnormal development of the mammary gland and pancreas. Cell. 1990;61(6):1137–46. 25. Sandgren EP, Schroeder JA, Qui TH, Palmiter RD, Brinster RL, Lee DC. Inhibition of mammary gland involution is associated with transforming growth factor alpha but not c-mycinduced tumorigenesis in transgenic mice. Cancer Res. 1995;55(17):3915–27. 26. Sandgren EP, Luetteke NC, Palmiter RD, Brinster RL, Lee DC. Overexpression of TGF alpha in transgenic mice: induction of epithelial hyperplasia, pancreatic metaplasia, and carcinoma of the breast. Cell. 1990;61(6):1121–35. 27. Rose-Hellekant TA, Sandgren EP. Transforming growth factor alpha- and c-myc-induced mammary carcinogenesis in transgenic mice. Oncogene. 2000;19(8):1092–6. 28. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science. 1987;235(4785):177–82. 29. Slamon DJ, Godolphin W, Jones LA, Holt JA, Wong SG, Keith DE, Levin WJ, Stuart SG, Udove J, Ullrich A, et al. Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science. 1989;244(4905):707–12. 30. Ravdin PM, Chamness GC. The c-erbB-2 proto-oncogene as a prognostic and predictive marker in breast cancer: a paradigm for the development of other macromolecular markers – a review. Gene. 1995;159(1):19–27. 31. Andrulis IL, Bull SB, Blackstein ME, Sutherland D, Mak C, Sidlofsky S, Pritzker KP, Hartwick RW, Hanna W, Lickley L, Wilkinson R, Qizilbash A, Ambus U, Lipa M, Weizel H, Katz A, Baida M, Mariz S, Stoik G, Dacamara P, Strongitharm D, Geddie W, McCready D. neu/erbB-2 amplification identifies a poor-prognosis group of women with node-negative breast cancer. Toronto Breast Cancer Study Group. J Clin Oncol. 1998;16(4):1340–9.


439

32. Pegram MD, Pauletti G, Slamon DJ. HER-2/neu as a predictive marker of response to breast cancer therapy. Breast Cancer Res Treat. 1998;52(1–3):65–77. 33. Muller WJ, Sinn E, Pattengale PK, Wallace R, Leder P. Single-step induction of mammary adenocarcinoma in transgenic mice bearing the activated c-neu oncogene. Cell. 1988;54(1):105–15. 34. Bouchard L, Lamarre L, Tremblay PJ, Jolicoeur P. Stochastic appearance of mammary tumors in transgenic mice carrying the MMTV/c-neu oncogene. Cell. 1989;57(6):931–6. 35. Stocklin E, Botteri F, Groner B. An activated allele of the c-erbB-2 oncogene impairs kidney and lung function and causes early death of transgenic mice. J Cell Biol. 1993;122(1):199–208. 36. Guy CT, Cardiff RD, Muller WJ. Activated neu induces rapid tumor progression. J Biol Chem. 1996;271(13):7673–8. 37. Guy CT, Webster MA, Schaller M, Parsons TJ, Cardiff RD, Muller WJ. Expression of the neu protooncogene in the mammary epithelium of transgenic mice induces metastatic disease. Proc Natl Acad Sci U S A. 1992;89(22):10578–82. 38. Siegel PM, Dankort DL, Hardy WR, Muller WJ. Novel activating mutations in the neu protooncogene involved in induction of mammary tumors. Mol Cell Biol. 1994;14(11):7068–77. 39. Hutchinson JN, Muller WJ. Transgenic mouse models of human breast cancer. Oncogene. 2000;19(53):6130–7. 40. Green JE, Shibata MA, Yoshidome K, Liu ML, Jorcyk C, Anver MR, Wigginton J, Wiltrout R, Shibata E, Kaczmarczyk S, Wang W, Liu ZY, Calvo A, Couldrey C. The C3(1)/SV40 T-antigen transgenic mouse model of mammary cancer: ductal epithelial cell targeting with multistage progression to carcinoma. Oncogene. 2000;19(8):1020–7. 41. Maroulakou IG, Anver M, Garrett L, Green JE. Prostate and mammary adenocarcinoma in transgenic mice carrying a rat C3(1) simian virus 40 large tumor antigen fusion gene. Proc Natl Acad Sci U S A. 1994;91(23):11236–40. 42. Shibata MA, Ward JM, Devor DE, Liu ML, Green JE. Progression of prostatic intraepithelial neoplasia to invasive carcinoma in C3(1)/SV40 large T antigen transgenic mice: histopathological and molecular biological alterations. Cancer Res. 1996;56(21):4894–903. 43. Li M, Hu J, Heermeier K, Hennighausen L, Furth PA. Apoptosis and remodeling of mammary gland tissue during involution proceeds through p53-independent pathways. Cell Growth Differ. 1996;7(1):13–20. 44. Li M, Hu J, Heermeier K, Hennighausen L, Furth PA. Expression of a viral oncoprotein during mammary gland development alters cell fate and function: induction of p53-independent apoptosis is followed by impaired milk protein production in surviving cells. Cell Growth Differ. 1996;7(1):3–11. 45. Santarelli R, Tzeng YJ, Zimmermann C, Guhl E, Graessmann A. SV40 T-antigen induces breast cancer formation with a high efficiency in lactating and virgin WAP-SV-T transgenic animals but with a low efficiency in ovariectomized animals. Oncogene. 1996;12(3):495–505. 46. Li M, Lewis B, Capuco AV, Laucirica R, Furth PA. WAP-TAg transgenic mice and the study of dysregulated cell survival, proliferation, and mutation during breast carcinogenesis. Oncogene. 2000;19(8):1010–9. 47. Osborne RJ, Merlo GR, Mitsudomi T, Venesio T, Liscia DS, Cappa AP, Chiba I, Takahashi T, Nau MM, Callahan R, et al. Mutations in the p53 gene in primary human breast cancers. Cancer Res. 1991;51(22):6194–8. 48. Petitjean A, Achatz MI, Borresen-Dale AL, Hainaut P, Olivier M. TP53 mutations in human cancers: functional selection and impact on cancer prognosis and outcomes. Oncogene. 2007;26(15):2157–65. 49. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA, Jr., Butel JS, Bradley A. Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature. 1992;356(6366):215–21. 50. Tennant RW, French JE, Spalding JW. Identifying chemical carcinogens and assessing potential risk in short-term bioassays using transgenic mouse models. Environ Health Perspect. 1995;103(10):942–50.

440


51. Hursting SD, Perkins SN, Brown CC, Haines DC, Phang JM. Calorie restriction induces a p53-independent delay of spontaneous carcinogenesis in p53-deficient and wild-type mice. Cancer Res. 1997;57(14):2843–6. 52. Hursting SD, Perkins SN, Haines DC, Ward JM, Phang JM. Chemoprevention of spontaneous tumorigenesis in p53-knockout mice. Cancer Res. 1995;55(18):3949–53. 53. Hursting SD, Perkins SN, Phang JM. Calorie restriction delays spontaneous tumorigenesis in p53-knockout transgenic mice. Proc Natl Acad Sci U S A. 1994;91(15):7036–40. 54. Perkins SN, Hursting SD, Haines DC, James SJ, Miller BJ, Phang JM. Chemoprevention of spontaneous tumorigenesis in nullizygous p53-deficient mice by dehydroepiandrosterone and its analog 16alpha-fluoro-5-androsten-17-one. Carcinogenesis. 1997;18(5):989–94. 55. Poetschke HL, Klug DB, Perkins SN, Wang TT, Richie ER, Hursting SD. Effects of calorie restriction on thymocyte growth, death and maturation. Carcinogenesis. 2000;21(11):1959–64. 56. Wang TT, Hursting SD, Perkins SN, Phang JM. Effects of dehydroepiandrosterone and calorie restriction on the Bcl-2/Bax-mediated apoptotic pathway in p53-deficient mice. Cancer Lett. 1997;116(1):61–9. 57. Li B, Rosen JM, McMenamin-Balano J, Muller WJ, Perkins AS. neu/ERBB2 cooperates with p53–172H during mammary tumorigenesis in transgenic mice. Mol Cell Biol. 1997;17(6):3155–63. 58. Jacks T, Remington L, Williams BO, Schmitt EM, Halachmi S, Bronson RT, Weinberg RA. Tumor spectrum analysis in p53-mutant mice. Curr Biol. 1994;4(1):1–7. 59. Li B, Murphy KL, Laucirica R, Kittrell F, Medina D, Rosen JM. A transgenic mouse model for mammary carcinogenesis. Oncogene. 1998;16(8):997–1007. 60. Jerry DJ, Kittrell FS, Kuperwasser C, Laucirica R, Dickinson ES, Bonilla PJ, Butel JS, Medina D. A mammary-specific model demonstrates the role of the p53 tumor suppressor gene in tumor development. Oncogene. 2000;19(8):1052–8. 61. Li Y, Hively WP, Varmus HE. Use of MMTV-Wnt-1 transgenic mice for studying the genetic basis of breast cancer. Oncogene. 2000;19(8):1002–9. 62. Nusse R. The Wnt gene family in tumorigenesis and in normal development. J Steroid Biochem Mol Biol. 1992;43(1–3):9–12. 63. Tsukamoto AS, Grosschedl R, Guzman RC, Parslow T, Varmus HE. Expression of the int-1 gene in transgenic mice is associated with mammary gland hyperplasia and adenocarcinomas in male and female mice. Cell. 1988;55(4):619–25. 64. Kwan H, Pecenka V, Tsukamoto A, Parslow TG, Guzman R, Lin TP, Muller WJ, Lee FS, Leder P, Varmus HE. Transgenes expressing the Wnt-1 and int-2 proto-oncogenes cooperate during mammary carcinogenesis in doubly transgenic mice. Mol Cell Biol. 1992;12(1):147–54. 65. Donehower LA, Godley LA, Aldaz CM, Pyle R, Shi YP, Pinkel D, Gray J, Bradley A, Medina D, Varmus HE. Deficiency of p53 accelerates mammary tumorigenesis in Wnt-1 transgenic mice and promotes chromosomal instability. Genes Dev. 1995;9(7):882–95. 66. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A. 2003;100(7):3983–8. 67. Wright MH, Calcagno AM, Salcido CD, Carlson MD, Ambudkar SV, Varticovski L. Brca1 breast tumors contain distinct CD44+/CD24- and CD133+ cells with cancer stem cell characteristics. Breast Cancer Res. 2008;10(1):R10. 68. Khleif SN, Abrams SI, Hamilton JM, Bergmann-Leitner E, Chen A, Bastian A, Bernstein S, Chung Y, Allegra CJ, Schlom J. A phase I vaccine trial with peptides reflecting ras oncogene mutations of solid tumors. J Immunother. 1999;22(2):155–65. 69. von Lintig FC, Dreilinger AD, Varki NM, Wallace AM, Casteel DE, Boss GR. Ras activation in human breast cancer. Breast Cancer Res Treat. 2000;62(1):51–62. 70. Hulit J, Di Vizio D, Pestell RG. Inducible transgenics. New lessons on events governing the induction and commitment in mammary tumorigenesis. Breast Cancer Res. 2001;3(4):209–12. 71. Nielsen LL, Discafani CM, Gurnani M, Tyler RD. Histopathology of salivary and mammary gland tumors in transgenic mice expressing a human Ha-ras oncogene. Cancer Res. 1991;51(14):3762–7.


441

72. Nielsen LL, Gurnani M, Tyler RD. Evaluation of the wap-ras transgenic mouse as a model system for testing anticancer drugs. Cancer Res. 1992;52(13):3733–8. 73. Nielsen LL, Gurnani M, Catino JJ, Tyler RD. In wap-ras transgenic mice, tumor phenotype but not cyclophosphamide-sensitivity is affected by genetic background. Anticancer Res. 1995;15(2):385–92. 74. Sinn E, Muller W, Pattengale P, Tepler I, Wallace R, Leder P. Coexpression of MMTV/vHa-ras and MMTV/c-myc genes in transgenic mice: synergistic action of oncogenes in vivo. Cell. 1987;49(4):465–75. 75. Strange R, Li F, Saurer S, Burkhardt A, Friis RR. Apoptotic cell death and tissue remodelling during mouse mammary gland involution. Development. 1992;115(1):49–58. 76. Nass SJ, Dickson RB. Defining a role for c-Myc in breast tumorigenesis. Breast Cancer Res Treat. 1997;44(1):1–22. 77. Schoenenberger CA, Andres AC, Groner B, van der Valk M, LeMeur M, Gerlinger P. Targeted c-myc gene expression in mammary glands of transgenic mice induces mammary tumours with constitutive milk protein gene transcription. EMBO J. 1988;7(1):169–75. 78. Stewart TA, Pattengale PK, Leder P. Spontaneous mammary adenocarcinomas in transgenic mice that carry and express MTV/myc fusion genes. Cell. 1984;38(3):627–37. 79. Dickson C, Fantl V, Gillett C, Brookes S, Bartek J, Smith R, Fisher C, Barnes D, Peters G. Amplification of chromosome band 11q13 and a role for cyclin D1 in human breast cancer. Cancer Lett. 1995;90(1):43–50. 80. Bartkova J, Lukas J, Muller H, Lutzhoft D, Strauss M, Bartek J. Cyclin D1 protein expression and function in human breast cancer. Int J Cancer. 1994;57(3):353–61. 81. Gillett C, Fantl V, Smith R, Fisher C, Bartek J, Dickson C, Barnes D, Peters G. Amplification and overexpression of cyclin D1 in breast cancer detected by immunohistochemical staining. Cancer Res. 1994;54(7):1812–7. 82. McIntosh GG, Anderson JJ, Milton I, Steward M, Parr AH, Thomas MD, Henry JA, Angus B, Lennard TW, Horne CH. Determination of the prognostic value of cyclin D1 overexpression in breast cancer. Oncogene. 1995;11(5):885–91. 83. Yu Q, Geng Y, Sicinski P. Specific protection against breast cancers by cyclin D1 ablation. Nature. 2001;411(6841):1017–21. 84. Wang TC, Cardiff RD, Zukerberg L, Lees E, Arnold A, Schmidt EV. Mammary hyperplasia and carcinoma in MMTV-cyclin D1 transgenic mice. Nature. 1994;369(6482):669–71. 85. D’Cruz CM, Gunther EJ, Boxer RB, Hartman JL, Sintasath L, Moody SE, Cox JD, Ha SI, Belka GK, Golant A, Cardiff RD, Chodosh LA. c-MYC induces mammary tumorigenesis by means of a preferred pathway involving spontaneous Kras2 mutations. Nat Med. 2001;7(2):235–9. 86. Gunther EJ, Belka GK, Wertheim GB, Wang J, Hartman JL, Boxer RB, Chodosh LA. A novel doxycycline-inducible system for the transgenic analysis of mammary gland biology. FASEB J. 2002;16(3):283–92. 87. Xu X, Weaver Z, Linke SP, Li C, Gotay J, Wang XW, Harris CC, Ried T, Deng CX. Centrosome amplification and a defective G2-M cell cycle checkpoint induce genetic instability in BRCA1 exon 11 isoform-deficient cells. Mol Cell. 1999;3(3):389–95. 88. Vogel VG. The NSABP Study of Tamoxifen and Raloxifene (STAR) trial. Expert Rev Anticancer Ther. 2009;9(1):51–60. 89. Menard S, Aiello P, Tagliabue E, Rumio C, Lollini PL, Colnaghi MI, Balsari A. Tamoxifen chemoprevention of a hormone-independent tumor in the proto-neu transgenic mice model. Cancer Res. 2000;60(2):273–5. 90. Yang X, Edgerton SM, Kosanke SD, Mason TL, Alvarez KM, Liu N, Chatterton RT, Liu B, Wang Q, Kim A, Murthy S, Thor AD. Hormonal and dietary modulation of mammary carcinogenesis in mouse mammary tumor virus-c-erbB-2 transgenic mice. Cancer Res. 2003;63(10):2425–33. 91. Sacco MG, Soldati S, Mira Cato E, Cattaneo L, Pratesi G, Scanziani E, Vezzoni P. Combined effects on tumor growth and metastasis by anti-estrogenic and antiangiogenic therapies in MMTV-neu mice. Gene Ther. 2002;9(19):1338–41.

442


92. Sacco MG, Soldati S, Indraccolo S, Cato EM, Cattaneo L, Scanziani E, Vezzoni P. Combined antiestrogen, antiangiogenic and anti-invasion therapy inhibits primary and metastatic tumor growth in the MMTVneu model of breast cancer. Gene Ther. 2003;10(22):1903–9. 93. Tekmal RR, Ramachandra N, Gubba S, Durgam VR, Mantione J, Toda K, Shizuta Y, Dillehay DL. Overexpression of int-5/aromatase in mammary glands of transgenic mice results in the induction of hyperplasia and nuclear abnormalities. Cancer Res. 1996;56(14):3180–5. 94. Luthra R, Kirma N, Jones J, Tekmal RR. Use of letrozole as a chemopreventive agent in aromatase overexpressing transgenic mice. J Steroid Biochem Mol Biol. 2003;86(3–5):461–7. 95. Wu K, Zhang Y, Xu XC, Hill J, Celestino J, Kim HT, Mohsin SK, Hilsenbeck SG, Lamph WW, Bissonette R, Brown PH. The retinoid X receptor-selective retinoid, LGD1069, prevents the development of estrogen receptor-negative mammary tumors in transgenic mice. Cancer Res. 2002;62(22):6376–80. 96. Wu K, Kim HT, Rodriquez JL, Hilsenbeck SG, Mohsin SK, Xu XC, Lamph WW, Kuhn JG, Green JE, Brown PH. Suppression of mammary tumorigenesis in transgenic mice by the RXR-selective retinoid, LGD1069. Cancer Epidemiol Biomarkers Prev. 2002;11(5):467–74. 97. Suh N, Lamph WW, Glasebrook AL, Grese TA, Palkowitz AD, Williams CR, Risingsong R, Farris MR, Heyman RA, Sporn MB. Prevention and treatment of experimental breast cancer with the combination of a new selective estrogen receptor modulator, arzoxifene, and a new rexinoid, LG 100268. Clin Cancer Res. 2002;8(10):3270–5. 98. Rendi MH, Suh N, Lamph WW, Krajewski S, Reed JC, Heyman RA, Berchuck A, Liby K, Risingsong R, Royce DB, Williams CR, Sporn MB. The selective estrogen receptor modulator arzoxifene and the rexinoid LG100268 cooperate to promote transforming growth factor beta-dependent apoptosis in breast cancer. Cancer Res. 2004;64(10):3566–71. 99. Mendelsohn J, Baselga J. The EGF receptor family as targets for cancer therapy. Oncogene. 2000;19(56):6550–65. 100. Moasser MM, Basso A, Averbuch SD, Rosen N. The tyrosine kinase inhibitor ZD1839 (“Iressa”) inhibits HER2-driven signaling and suppresses the growth of HER2-overexpressing tumor cells. Cancer Res. 2001;61(19):7184–8. 101. Shen Q, Brown PH. Novel agents for the prevention of breast cancer: targeting transcription factors and signal transduction pathways. J Mammary Gland Biol Neoplasia. 2003;8(1):45–73. 102. Howe LR, Subbaramaiah K, Brown AM, Dannenberg AJ. Cyclooxygenase-2: a target for the prevention and treatment of breast cancer. Endocr Relat Cancer. 2001;8(2):97–114. 103. Soslow RA, Dannenberg AJ, Rush D, Woerner BM, Khan KN, Masferrer J, Koki AT. COX-2 is expressed in human pulmonary, colonic, and mammary tumors. Cancer. 2000;89(12):2637–45. 104. Liu CH, Chang SH, Narko K, Trifan OC, Wu MT, Smith E, Haudenschild C, Lane TF, Hla T. Overexpression of cyclooxygenase-2 is sufficient to induce tumorigenesis in transgenic mice. J Biol Chem. 2001;276(21):18563–9. 105. Howe LR, Subbaramaiah K, Patel J, Masferrer JL, Deora A, Hudis C, Thaler HT, Muller WJ, Du B, Brown AM, Dannenberg AJ. Celecoxib, a selective cyclooxygenase 2 inhibitor, protects against human epidermal growth factor receptor 2 (HER-2)/neu-induced breast cancer. Cancer Res. 2002;62(19):5405–7. 106. Hursting SD, Lashinger LM, Wheatley KW, Rogers CJ, Colbert LH, Nunez NP, Perkins SN. Reducing the weight of cancer: mechanistic targets for breaking the obesity-carcinogenesis link. Best Pract Res Clin Endocrinol Metab. 2008;22(4):659–69. 107. Dove WF, Gould KA, Luongo C, Moser AR, Shoemaker AR. Emergent issues in the genetics of intestinal neoplasia. Cancer Surv. 1995;25:335–55. 108. Mai V, Colbert LH, Berrigan D, Perkins SN, Pfeiffer R, Lavigne JA, Lanza E, Haines DC, Schatzkin A, Hursting SD. Calorie restriction and diet composition modulate spontaneous intestinal tumorigenesis in Apc(Min) mice through different mechanisms. Cancer Res. 2003;63(8):1752–5. 109. Colbert LH, Mai V, Perkins SN, Berrigan D, Lavigne JA, Wimbrow HH, Alvord WG, Haines DC, Srinivas P, Hursting SD. Exercise and intestinal polyp development in APCMin mice. Med Sci Sports Exerc. 2003;35(10):1662–9.


443

110. Zeytin HE, Patel AC, Rogers CJ, Canter D, Hursting SD, Schlom J, Greiner JW. Combination of a poxvirus-based vaccine with a cyclooxygenase-2 inhibitor (celecoxib) elicits antitumor immunity and long-term survival in CEA.Tg/MIN mice. Cancer Res. 2004;64(10):3668–78. 111. Steinbach G, Lynch PM, Phillips RK, Wallace MH, Hawk E, Gordon GB, Wakabayashi N, Saunders B, Shen Y, Fujimura T, Su LK, Levin B. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med. 2000;342(26):1946–52. 112. Nakatsugi S, Fukutake M, Takahashi M, Fukuda K, Isoi T, Taniguchi Y, Sugimura T, Wakabayashi K. Suppression of intestinal polyp development by nimesulide, a selective cyclooxygenase-2 inhibitor, in Min mice. Jpn J Cancer Res. 1997;88(12):1117–20. 113. Jacoby RF, Seibert K, Cole CE, Kelloff G, Lubet RA. The cyclooxygenase-2 inhibitor celecoxib is a potent preventive and therapeutic agent in the min mouse model of adenomatous polyposis. Cancer Res. 2000;60(18):5040–4. 114. Graham DJ, Campen D, Hui R, Spence M, Cheetham C, Levy G, Shoor S, Ray WA. Risk of acute myocardial infarction and sudden cardiac death in patients treated with cyclo-oxygenase 2 selective and non-selective non-steroidal anti-inflammatory drugs: nested case-control study. Lancet. 2005;365(9458):475–81. 115. Roebuck BD, Baumgartner KJ, Longnecker DS. Growth of pancreatic foci and development of pancreatic cancer with a single dose of azaserine in the rat. Carcinogenesis. 1987;8(12):1831–5. 116. Grippo PJ, Nowlin PS, Demeure MJ, Longnecker DS, Sandgren EP. Preinvasive pancreatic neoplasia of ductal phenotype induced by acinar cell targeting of mutant Kras in transgenic mice. Cancer Res. 2003;63(9):2016–9. 117. Wagner M, Greten FR, Weber CK, Koschnick S, Mattfeldt T, Deppert W, Kern H, Adler G, Schmid RM. A murine tumor progression model for pancreatic cancer recapitulating the genetic alterations of the human disease. Genes Dev. 2001;15(3):286–93. 118. Tuveson DA, Zhu L, Gopinathan A, Willis NA, Kachatrian L, Grochow R, Pin CL, Mitin NY, Taparowsky EJ, Gimotty PA, Hruban RH, Jacks T, Konieczny SF. Mist1-KrasG12D knock-in mice develop mixed differentiation metastatic exocrine pancreatic carcinoma and hepatocellular carcinoma. Cancer Res. 2006;66(1):242–7. 119. Leach SD. Mouse models of pancreatic cancer: the fur is finally flying! Cancer Cell. 2004;5(1):7–11. 120. Hingorani SR, Petricoin EF, Maitra A, Rajapakse V, King C, Jacobetz MA, Ross S, Conrads TP, Veenstra TD, Hitt BA, Kawaguchi Y, Johann D, Liotta LA, Crawford HC, Putt ME, Jacks T, Wright CV, Hruban RH, Lowy AM, Tuveson DA. Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse. Cancer Cell. 2003;4(6):437–50. 121. Aguirre AJ, Bardeesy N, Sinha M, Lopez L, Tuveson DA, Horner J, Redston MS, DePinho RA. Activated Kras and Ink4a/Arf deficiency cooperate to produce metastatic pancreatic ductal adenocarcinoma. Genes Dev. 2003;17(24):3112–26. 122. Hingorani SR, Wang L, Multani AS, Combs C, Deramaudt TB, Hruban RH, Rustgi AK, Chang S, Tuveson DA. Trp53R172H and KrasG12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice. Cancer Cell. 2005;7(5):469–83. 123. Bardeesy N, Aguirre AJ, Chu GC, Cheng KH, Lopez LV, Hezel AF, Feng B, Brennan C, Weissleder R, Mahmood U, Hanahan D, Redston MS, Chin L, Depinho RA. Both p16(Ink4a) and the p19(Arf)-p53 pathway constrain progression of pancreatic adenocarcinoma in the mouse. Proc Natl Acad Sci U S A. 2006;103(15):5947–52. 124. Funahashi H, Satake M, Dawson D, Huynh NA, Reber HA, Hines OJ, Eibl G. Delayed progression of pancreatic intraepithelial neoplasia in a conditional Kras(G12D) mouse model by a selective cyclooxygenase-2 inhibitor. Cancer Res. 2007;67(15):7068–71. 125. Colby JK, Klein RD, McArthur MJ, Conti CJ, Kiguchi K, Kawamoto T, Riggs PK, Pavone AI, Sawicki J, Fischer SM. Progressive metaplastic and dysplastic changes in mouse pancreas induced by cyclooxygenase-2 overexpression. Neoplasia. 2008;10(8):782–96.

Part VII

Metastasis Models

Chapter 19

Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease Suzanne A. Eccles

Abstract This chapter will indicate molecular targets that may be appropriate for metastasis therapy and how they may be evaluated in appropriate preclinical models of malignant disease. The focus will be mainly on the evaluation of agents directed against oncogenic signaling pathways, which have provided the majority of new targeted therapies in the last decade. These approaches include small molecules and antibodies with a brief mention of gene therapeutic and immunological approaches. There is also a detailed description of a variety of tumor models (including syngeneic, xenogeneic, transgenic, and orthotopic) and their use in different therapeutic applications, with a brief discussion of various methods for measuring efficacy. Keywords Metastasis • Invasion • Syngeneic • Xenogeneic • Orthotopic • Transgenic • Receptors • Inhibitors • Cancer therapy • Signaling • Stem cells • Angiogenesis

19.1 Introduction Metastasis is the most frequent cause of cancer death and novel systemic therapies are required to improve patient outcome [1–3]. Unfortunately, preclinical drug development does not routinely include animal tumor models that mimic metastatic human cancer, and truly “adjuvant” preclinical studies are rare. This deficiency has recently been highlighted by the finding that some antiangiogenic therapies may enhance metastasis, emphasizing the need for more rigorous preclinical evaluation of new agents in appropriate models of disseminated disease [4]. Animal models have been criticized for failing to predict responses in clinical trials [5]. Notably,

S.A. Eccles (*) Tumour Biology and Metastasis, Cancer Research UK Centre for Cancer Therapeutics, The Institute of Cancer Research, Cotswold Rd, Belmont, Sutton, Surrey, SM2 5NG, UK e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_19, © Springer Science+Business Media, LLC 2011

447

448

S.A. Eccles

angiostatin was found to be effective in several syngeneic models, but failed in the clinic, possibly because the immunogenicity of the tumor contributed to its responsiveness [6]. A second “failure” was of matrix metalloprotease (MMP) inhibitors, now appreciated as being partially due to the inadvertent inhibition of tumor suppressor proteases and the greater dependency of micrometastases on MMPs for neoangiogenesis compared with established tumors [7]. Also, with the growing realization of the importance of cellular context and microenvironment [8], it is clear that the use of inappropriate models or a failure to take into account differences in oncogenic drivers can give misleading results [9]. Cytotoxic therapies, in particular, are likely to be more active in fast growing rodent tumors and the fact that the attrition rates are less for the more recent targeted therapies [10] provides encouragement that animal models are a valuable intermediary between target validation and clinical trials of novel agents. Preclinical models of cancer can contribute enormously to drug development in many ways: firstly by target validation (e.g. by gene knockdown) or in seeking novel targets by interrogating genetically defined cell lines with lentiviral RNAi libraries in a “synthetic lethality” approach. However, complete ablation of a gene is not the same as inhibition of a function (e.g. enzyme activity, ligand binding) with a drug or antibody. Secondly, animal models are invaluable for identifying biomarkers of response (which are now de rigueur in hypothesis-testing clinical trials) and for identifying off-target or potential harmful side effects [11]. The original mainstays of cancer research were often highly antigenic, transplanted rodent tumors. A few of these (most notably B16F10 and Lewis lung carcinoma) were used as metastasis models, generally by injecting the cells i.v. to give lung colonies. Later, human tumors were grown s.c. in immunodeprived mice but rarely metastasized. So the advantage of testing drugs against human molecular targets was counterbalanced by the lack of the definitive feature of malignant human cancers, i.e. distant metastasis. Fortunately, when human tumors are grown orthotopically (i.e. in the correct anatomical site) there is a higher probability of metastasis [12]. Athymic and SCID mice, however, have abnormal immune (and endocrine) systems, and therefore limitations for some therapeutic applications. Transgenic technology is finally enabling the development of more “patientlike” models of cancer; where a known genetic aberration is introduced into the germ line and is seen as “self” by the host. However, it remains to be proven in most cases that the strong over-expression of a single oncogene (or “knock out” of a suppressor gene) results in cancers that accurately mimic the human malignancy (especially in relation to metastasis) where multiple genetic, epigenetic, and environmental factors contribute to progression. To overcome this, double and triple transgenic animals are being developed whose tumors show increased malignant potential [13]. Also, well-established and/or simpler models still have their place for the high throughput that is required for modern drug development. The art of preclinical evaluation of targeted therapies is choosing the most appropriate model for the particular question being addressed. The use of genetically tagged cells now enables earlier detection and more accurate quantitation of micrometastases. However, there can be issues in

19 Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease

449

d etermining whether the tumor cells detected in blood, marrow, or other sites are clonogenic – and this applies to clinical studies too. In addition, increasingly sophisticated imaging techniques are enabling high resolution detection and functional readouts of internal tumors. Many are directly applicable to the clinic and others such as bioluminescence and fluorescence require the use of engineered cells or animals expressing the relevant transgene. In addition, methods for reporting on the delivery and activity of agents are being developed, as well as revealing, for example, intratumoral metabolic, angiogenic, and proteolytic functions.

19.2 Therapeutic Strategies for Targeting Metastases It is generally accepted that further genetic or epigenetic changes (beyond those involved in cellular transformation) are required for metastatic competence, although the recognition of metastasis-associated “signatures” in primary tumors has suggested that this may be a relatively early event [14–16]. Experimentally, it has been easier to identify metastasis suppressor genes than those whose activity can potentiate dissemination because the latter may require the activity of complementary genetic changes to manifest their potential. Interestingly, several of the best characterized metastasis suppressors prevent outgrowth of cells at secondary sites rather than inhibiting earlier phases [17]. This may be a reflection of the essential need of multicellular organisms to prevent ectopic growth of any cells that may escape from their normal tissue boundaries.

19.2.1 Target Identification and Validation Drug development is expensive and lengthy, so potential targets must be thoroughly validated [18]. Initial evidence for a “good” target often comes from studies of genetic aberrations in human cancers (translocations such as the Bcr-Abl oncogene and APC or PTEN loss leading to activated Wnt/b-catenin and PI3K pathway signaling, respectively) or mutations (e.g. BRAF in melanoma). Several were first identified in rodent tumors (e.g. c-ErbB2/neu in rats treated with a chemical carcinogen and Wnt from the insertion sites of oncogenic viruses). The process has become much more efficient with the advent of transgenic mice where genes can readily be mutated/hyperactivated and linked to tumor development. Also RNAi technology has provided a means to check the potential effects of pharmacological inhibition of a target (although with caveats) and increasingly as an approach to determine the “Achilles’ heel” of tumors with a specific genetic defect or to identify synergistic combinations. Since human cancers contain multiple genetic mutations, it is critical to discern the “drivers” (directly and causally linked to malignancy) from “passengers” resulting

450

S.A. Eccles

from genetic instability and whose presence is incidental. It is also important to determine the role(s) of these gene products before they are considered as targets for therapy. Some molecules are associated with early events (e.g. immortalization, initiation) but may not be essential for maintenance of the malignant phenotype and become redundant once a tumor (or metastasis) is established. Recently, several studies in experimental animal models and human cancers have identified site-specific metastasis signatures, e.g. for lung, bone, and brain [19–21]. This may reflect adaptation of tumor cells to a particular environment or preselection of cells with a survival advantage in specific tissues. Metastases (even those derived from the same primary but developing in different sites) may depend upon different genotypes/phenotypes for their establishment, continued expansion, or further spread. Therefore, although metastatic models may be more technically challenging than simple subcutaneous models, they are likely to give more accurate, sensitive, and predictive readouts of novel therapeutic approaches.

19.2.2 Molecular Targets The following section provides examples of established and emerging molecular targets with key roles in metastatic disease. Examples of targets in cancer cells, related to different steps in the metastatic cascade, and associated with specific organ tropisms are shown in Fig. 19.1a–c. 19.2.2.1 Oncogenic Receptor Tyrosine Kinase Signaling Pathways Monoclonal antibodies were initially the mainstays of targeted therapies against oncogenic cell surface receptors, but more recently have been supplemented by small molecule inhibitors, which have the advantage of oral bioavailability. Such drugs have been developed against proteins, which are mutated, overexpressed, or hyperactivated in cancers. Many are targeted against kinases which, because of the structure of their ATP binding sites, lend themselves to inhibition by small molecules. Selective drug (and some antibody-based) therapies targeting oncogenic signaling pathways have emerged as important new classes of anticancer agents and are showing activity in the clinic [22]. However, only subsets of patients respond and resistance remains an issue [23]. Key targets include receptor tyrosine kinases (RTK) such as EGFR, ErbB2/ HER2, PDGFR, c-KIT, c-MET, RON, RET, FGFR, IGF1R, etc. on tumor cells and VEGFR2 on vascular endothelial cells. In addition, elements of their downstream signaling pathways including PI3K, MTOR, AKT, BRAF, SRC, RAF, PLCg, etc. are receiving increasing attention [24–27]. Many of these pathways have been implicated in invasion, metastasis, and angiogenesis as well as primary tumor growth (Fig. 19.1a) [28–36]. Unexpected complexities may be uncovered in experimental models, for example, AKT1 co-expressed with PyVMT or ErbB2 in

a

RECEPTORS

ADHESION MOLECULES

CD44 Cadherins

INTEGRINS

PROTEASES

c-Kit, TGFβ, GPCR, Fzd Eph receptors Plexins Ion channels

EGFR, HER2 c-Met, RON PDGFR IGF1R FGFR

α6β4 α5β1/2 α9β1

MMPs, ADAMs uPA Cathepsins Heparanase

ECM P

CELL

Rho ROCK

Rac P P

P

P P P

TRANSCRIPTION

HIF1 HDAC AR, ER MYC NFKB, BMI1 Snail, Twist, ZEB1,2

b

gelsonin

HSP90 chaperone

P

SIGNALLING PATHWAYS

Ras, RAF FAK, PAK, ILK PI3-kinase- AKT-mTOR PLCγ, Memo c-Src, β-catenin Rho, Rac, ROCK, ezrin

profilin cofilin

P P

NUCLEUS

PRIMARY TUMOUR

METASTASES

angiogenesis VEGF/VEGFR αVβ3, αVβ5 PDGFR, Tie2 Robo/Slit

EMT

extravasation

c-MET, Wnt Vimentin, N,-cad Snail, Twist

Integrins Proteases COX-2

motility

hypoxia HIF pathway SIAH LOX PI3K

cancer ‘stem’ cells Hh, telomerase Notch, BMI1, CXCR4 Wnt, Kit, CD44 ABC transporters

C-MET, EGFR, HER2, PLCγ, ROCK, Rac, Rho, Src, PAK

intravasation ErbB2, MMP1,2 Epiregulin Cox-2

anoikis resistance TrkB PI3K/AKT

premetastatic niche VEGFR1 Fibronectin CXCR4 MMP9,LOX osteopontin

ectopic growth RTK-ligands GPCR-cytokines Chaperones JNK/p38

Fig. 19.1 (a) Molecular targets for therapy of metastasis, (b) The role of different molecules in specific elements of the metastatic cascade,

452

S.A. Eccles

c

SITES OF METASTASIS Bone

RANK/RANKL PTHrP ET1, BMP c-Src, GM-CSF Fzd-Wnts EphA2 FGFR/bFGF IGF1R/IGF PDGFRα/PDGFβ CXCR4/CXCL12 TGFR/TGFβ Osteopontin MMP2, MMP9

Brain

HER2-NRG IGF1R/IGF EphA2/ephrinA3 IL6R/IL6 EGFR(vIII)/EGF TGFR/TGF β uPA/uPAR/PAI-1

Lung

EGFR/EGF/EREG CXCR4/CXCL12 CCR5/CCL3 CXCR2/CXCL1 CXCR3/CXCL10 MMP1MMP2 , ANGPTL4 COX2 VCAM1 SPARC CD44 HSP27

Liver

EGFR-TGFα C-MET-HGF EphA2/ephrinA1 CXCR5/CXCL13 CXCR4/CXCL12 IGF1R/IGF1/2 PDGFRβ/PDGF CD44-HA MMP7 MMP9 HSP60

Lymph node

VEGFR3/VEGFC CCR7/CCL19/21 CXCR4/CXCL12 EphB4/ephrinB2 PDGFRα/PDGFBB IGF1R/IGF1 IGFBP7 PAX5 α1β5 integrin HSP110 GRP94 GRP78

Fig 19.1 (continued) (c) Key signaling pathways implicated in site-selective metastasis. Factors released by tumor cells activate osteoclasts and/or osteoblasts. A vicious cycle between these three cell types and growth factors released from the stroma potentiate tumor cell invasion and bone destruction/remodeling. Cytokines and their receptors may also contribute. In the brain, growth factors are produced, notably by astrocytes, which can stimulate the proliferation and invasion of tumor cells expressing the cognate receptors. Angiogenic factors released by tumor cells are also implicated. An important determinant of lung metastasis is the specific ability of tumor cells to effect transmigration of the lung endothelium. Factors implicated include those released by tumor cells which act upon endothelial cells (e.g. EREG and ANGPTL4), and paracrine stimulation of tumor cells mediated by bone marrow stem cell-derived CCL5. Within the lung, growth may be stimulated by CXCR4-CCL12 interactions. Receptors on tumor cells such as EGFR and MET may respond to high levels of their ligands in the liver. Paracrine interactions between tumor cells, host stromal cells, and endothelial cells involving ephrins and chemokines are also evident. The major signaling pathways implicated in lymphatic metastasis are the VEGF-C-VEGFR3 and CCL21-CCR7 systems. PDGF-BB acting through RTK receptors may also play a role

transgenic mice enhanced primary breast cancer growth, whereas AKT2 increased pulmonary metastases [37]. Animal models, primarily human tumor xenografts and some transgenic models have been used to evaluate novel RTK inhibitors (examples are shown in Tables 19.1–19.3) [38]. Tumor models should be selected with knowledge of their molecular pathology, e.g. testing PI3K pathway inhibitors in tumors where activation is induced by loss of PTEN, overexpressed or mutated RTK, or mutant P110a [39, 40]. It is also important, whenever possible, to evaluate novel agents in metastatic models as signaling pathway activation, drug access and responses can vary in different sites [3, 34, 41]. Recently, Lapatinib (EGFR/ErbB2 inhibitor) was shown to be effective in

M24met (mouse) B16 sublines (mouse)

Lewis lung 3LL (mouse)

R4OP (mouse)

C26 (mouse) MC38 (mouse) CC531 (rat)

K7M2 (mouse)

Melanoma

Lung

Pancreatic

Colon

Osteosarcoma

Walker 256 (rat) HOSP1 (rat)

i.v.

i.v. Spleen, portal vein Spleen Intrahepatic Lung

Lung Liver Liver Liver

Lymph nodes

Lung Lymph nodes

s.c. Lung Pancreas

Lung, liver

Lung Lung Liver Lung

Bone Bone Lung, lymph nodes

Metastases Lung Bone marrow Bone, lung Lung Local invasion, lung, LN

i.v.

Id i.v. Spleen i.v.

Intracardiac Intratibial Mfp

Table 19.1 Syngeneic metastatic tumor models Tumour type Name (species) Injection site Breast 4T1 (mouse) Mfp 4T1/E/M3 i.v. 4T1 66cl4 Mfp BN472 (rat) Mfp MTLn3 (rat) Mfp

CXCR4 peptide

CXCR3 CEA, TLR4, immunotherapy COX-2 inhibitor celecoxib

aVb3 targeted nanoparticles + doxorubicin

EP4 receptor antagonist ONO-AE3-208 MMPi MM1270 Celecoxib, MM1270

MM1270 MMPi HMGB1 vaccination Mab DC101 (VEGFR2) + mab TA99 (TYRP-1)

TGFb antibody 1D11, tranilast ALK5 kinase inhibitor SM16 aVb3 uPA inhibitor Multiphoton microscopy to visualize motility and intravasation, EGFR IKB inhibitors celastrol, parthenolide MMPs batimastat

Therapeutic target

[270]

[62] [268] [269]

[220]

[267] [221]

[266]

[262] [263] [264] [265]

[261]

(continued)

[203, 209, 260]

Reference(s) [255] [256] [257] [86, 258] [105] [187, 259]

19 Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease 453

RENCA

Dunning R3327 MATLyLu (rat) MAT-LyLu (rat) Dunning R-3327 AT6.3 (rat)

Kidney

Prostate

Prostate i.v. s.c.

Kidney

Injection site Bladder wall Lung, lymph nodes Bone Lung

Lung

Metastases Invasive IL-2 uPa inhibitor B-428 CEP-701 Trk inhibitor

VEGF Trap

Therapeutic target Cathepsin B, MMP9 uPA

Glioma

Hras/SV40T transformed astrocytes (mouse)

Brain

VEGFR2/PDGFR Invasion along blood vessels and leptomeninges C6 (rat) Brain Invasion Anti-HIF (YC-1) Mfp mammary fat pad, i.v. intravenous, Ipro intraprostate, i.c. intracardiac, TYRP-1 tyrosinase-related protein 1

Name (species) MB49-I (mouse)

Tumour type Bladder

Table 19.1 (continued)

[170]

[82, 95, 276]

[273] [274] [275]

[272]

Reference(s) [271]

454 S.A. Eccles

Breast Bone Lung Lung Lung, LN, liver Lung Lungs, bone, heart, liver, kidney Brain Lung

Intracardiac, i

i.v. Mfp Mfp Mfp Intracardiac

BR (brain) MDA MB 435#

Intracardiac s.c.

Bone, brain, lung

s.c. s.c./orthotopic Bone

LNCaP-H1 LNCaP C4-2b MDA Pca 2b

MDA MB 231 sublines, notably BO2 and BoM-1833 (bone)

Bone Lung and rib LN, bone Bone

s.c. or intraprostate Intracardiac

Metastases Bone LN LN LN LN, lung, kidney, liver, pancreas, adrenal, bone LN, lung Brain

LAPC-9 CWR22Rv-1 DU145

Table 19.2 Xenogeneic metastatic tumor models Tumour type Cell line Injection site Prostate PC-3M Intra-tibial+ PC-3LN3 Prostate PC-3MM2 Prostate PC-3AR-A1 Prostate ARCaP prostate Prostate

C-MET + Met decoy, NK4

LOX inhibitor BAPN MMP1i and anti-PAR1peptide RNA aptamer against osteopontin Sunitinib Axl knockdown and antibody CTCE-9908 (CXCR4 inhibitor) Lapatinib

VEGFR2 or VEGFR3 inhibitors Ras/RalGef transfected. High EGFR, antiVEGF therapy and MR imaging Selected by growth in hypoxia Detected by QRT-PCR Gene therapy, anti-integrins Implanted human bone fragments, high IGF1R KM1468 mab

Molecular features/therapeutic targets uPA+ 17-AAG, dasatinib NVP-AUY922 Dasatinib AR transfected Osteoblastic and osteolytic

[284] (continued)

[279] [21, 99] [19, 83, 213, 280] [47, 59, 281, 282] [42, 211, 283]

[204]

[207, 208, 278]

[277]

[223]

Reference(s) [219] [56] [52] [218] [208]


Colon

Portal vein

Spleen i.v. Cecal wall Spleen Caecum

Spleen

LoVo

HT29 TK-4 KM20 KM12 C

COLO 320DM HCT116

CN34-BoM2

Bowel Liver Spleen/portal vein

Mfp Mfp, i.v. mfp Intracardiac s.c./mfp Intracardiac

MDA MB 453 LvBr MCF10A+HRAS+BMI1 MDA MB 468LN BT474 GI101 CN34 BrM2

HCT116

Subrenal Intracardiac Internal carotid

Injection site

MDA MB 435 Br4

Table 19.2 (continued) Tumour type Cell line

Liver

Liver Lung Liver LN, liver

Lung Liver Liver Liver

Bone

Lung Lung, spleen liver, brain LN Bone Lung, LN Brain

Lung Bone, adrenal, brain, ovary Brain

Metastases

PIK3CA mutant RON+, RON shRNA 17-DMAG BLI Oncolytic adenovirus targeted via A33 antigen RON+, CXCR3+ CXCR2+, SU6668 (RTKi) effects on premet niche CMET+, activated PI3K pathway High TGFa increased VEGF, MMPs, lymphatic density NOG mice, FTI inhibitor CH4512600

Avastin, PD-0332991, BLI GFP tagged Activated notch, respond to g secretase inhibitor Erlotinib (EGFR) High activated MAPK High OPN Trastuzumab ST6GALNAC5 potentiated brain extravasation Src dependent survival in bone marrow

Molecular features/therapeutic targets

[195]

[62] [113] [288] [196]

[159]

[193] [53, 287]

[212] [191] [285] [43] [286] [19, 211]

[283] [205] [198]

Reference(s)

456 S.A. Eccles

DBM2 U251 U87

Glioma

Intracardiac Intracranial (nu rat)

PC14/B CRL-5904

IMR32 CHLA-255 SK-N-BE

Intracardiac

PC9BrM3

Neuro-blastoma

Intracardiac

H69VP H2030BrM3

i.v./intracranial

Intratibial Intrafemoral i.v.

s.c.

i.v.

SBC-3/DOX

OH1, SW2

i.v. i.v.

A549 H226 RERF-LC-AI

SCLC

Lung

Spleen i.v.

LS174T GW-39

Lung/brain invasion

Bone, liver Bone Liver

Lung

Brain Brain

Brain, bone

Brain, bone

Liver, kidneys, nodes

Lung Multiorgan

Liver Lung

[202] [161, 293]

[188]

[292] [201]

[73]

[106] [291]

Cells selected for lung metastasis showed [200] increased brain invasion. High IL6, IL8, MCP-1, GM-CSF. Treated with 17AAG (continued)

High IGF1R osteolytic NPG targeted osteoprotegerin Sunitinib + rapamycin

More metastases in pfp/rag2 mice than SCID

ONO-4817 MMPi MMP inhibitors > liver mets via stromal effects Widespread metastases in NK-depleted SCID mice VEGFR2 inhibitor ZD6474 Wnt/TCF-LEF1-HOXB9 associated with brain mets Wnt/TCF-LEF1-HOXB9 associated with brain mets High S100B, VEGF and MMP9 Kca channels influencing BBB

CEA+, mab A5B7 [289] Pretargeted SPECT (anti-CEAmab TF2/anti- [171] HSGbispecific mab + radiolabelled HSG peptide) and PET imaging Labetuzumab anti-CEA mab + GMCSF [290]


686LN

GDC185 OSC19

SN12-VC

SN12-SVR WM239A 113/6-4L

SCCHN

Renal ca

Melanoma

131/4-5B WM266.6 MeWo UISO-Mel6 FEMX-1 SKMEL-28 A375 A375Br

HT1080 SW620

Sarcoma colon ca

NBT-II

Orthotopic (subdermal) i.v. i.v. s.c. i.v. Intracardiac Internal carotid

Orthotopic (kidney)

Orthotopic--> (sublingual) Floor of mouth Tongue

Chick embryo CAM

Brain Lung Lung Lung, brain, kidney, liver Lung Brain Bone Brain

Liver lungs, LN

Lung, liver, adrenal, pancreas, spleen

LN LN

Lung

Lung, liver

Intracardiac, Bone, soft tissues, lung intratibial, bladder s.c. LN

(T24) TSU-Pr1 series

Lung, bone

Lung Lung

Metastases

Bladder

i.v. Intratibial

Injection site

i.v.

SAOS-LM7 TE85-143B

Ewing’s sarcoma TC-71

Osteo-sarcoma

Table 19.2 (continued) Tumour type Cell line

[60]

[186]

[24]

[296]

[295]

[61] [294]

Reference(s)

[52] [67, 83] [300, 301] [107] [302, 303]

[220] [299]

aVb3 Metronomic chemotherapy BRAFmut, HSP90i Hh VEGFR2 CEACAM1+, MRI Mab L235 to melanotransferrin Laminin antagonist peptide Increasted STAT3, bFGF and VEGF

[298]

CXCR4+, CXCL12, PDGF-D/PDGFR, Gleevec

[297] DC101 (VEGFR mab), cetuximab (EGFR) [78] + BLI

CXCR4+

High levels of angiogenic factors and MMPs

FGFR+ EMT reversion (MET). High MT1-MMP, MT2-MMP, MMP9 Src

IGF1R+

CXCR3+ Src

Molecular features/therapeutic targets

458 S.A. Eccles

SiHa

Cervix

Cervix

Pancreas Pancreas Pancreas Pancreas LN, lung, viscera

LN, liver LN, peritoneum

Invasion, ascites, liver

LN, spleen, liver, kidneys

17-DMAG, FRNK (FAK antagonist)

Hh+, IPI-269609 (Hhi)+/ - gemcitabine NK4 and anti-HGF antibody POP33 prodrug activated in hypoxic cells ZD6474+ gemcitabine Somatostatin receptor+ imaged by microPET and MR [306]

[69] [214, 215] [304] [305]

Leukaemia SHI-1 i.v. Bone marrow, brain Detected by RT-PCR and BLI [192, 307] SCCHN squamous cell carcinoma of the head and neck, HGF hepatocyte growth factor, # may be melanoma, FTI farnesyltransferase inhibitor, NOG NOD/ Shi-scid IL2Rgnul, VEGFR2 vascular endothelial growth factor receptor 2 (Kdr/Flk), BLI bioluminescent imaging, PET positron emission tomography, NPG neural progenitor cells

E3LZ10.7 Capan-1 SUIT-2 L3.6pl AR42J

Pancreatic


NRASQ61K, Ink4a− /− HGF+ survivin (+UV) HGF

Genetic manipulation PTEN loss +BRAFv600E

Liver and lung Lung Lung Mainly lung and some LN

Lung

Constitutively active RON MMTV-PyVmT MMTV Neu

MMTV or WAP also used to target Wnt-1, c-Met, Ras, Notch, HGF, IGF2, TGFa FGFs, Cox-2, BRCA1+/-, etc. Also composite transgenics eg PyMT/CD44-/-Neu/TGFb PyVMT/AKT Neu/S100A4 PyVMT/uPA-/Wnt1/p53+/BRCA1/p53+/ErbB2/PTEN-/MMTV-PyVMT + dox inducible TGFb

Pancreatic acinar

Breast

Liver

Elastase-tv/a/p53−/−

LN and liver Lymph node Liver

Gut, mesentery

Metastases LN and lung. BRAF alone – benign tumours LN and lung Liver

RIP1-Tag2 RIP1-Tag2 + IGF1R KPC Kras and p53 mutant

Pancreatic neurendocrine cancer RIP1-Tag2

Tumour type Melanoma

Table 19.3 Transgenic metastatic tumor models

Dominant negative PLCg Galardin MMPi Mab mu4D5 (trastuzumab precursor) EphA2 mab TGFb antisense

HSP90 and glycolysis inhibitors Anti-VEGFR mab (DC101) IPI-926 (Hh) + gemcitabine

Hh inhibitors

Notes/experimental studies Rapamycin (mTorc1), PD325901 (MEK)

[230] [282] [281] [236] [238] [37] [229] [250] [240] [231] [314]

[312] [232] [101] [313]

[311]

[310] [70]

[54, 82]

[67] [308] [309]

Reference(s) [31]

460 S.A. Eccles

MyrAkt, p53-/-

Mutant b1 integrin+ chem. carcinogenesis LN

Apc

SCCHN

SCC (skin)

Colon

TRb(PV/PV)/Pten(+/-)

Apc mut + Smad2+/-, Smad4+/-, Kras mut, EphB2+/Lung

Invasion

LN

N/a

[246]

[245]

[68]

[44]

Activated mTOR/AKT

[241]

Mostly adenomas; require [237] additional mutations for invasion, no metastases. COX-2 inhibitors (celecoxib)

NA

NA. PET imaging

Cyclopamine Hh Antag

sunitinib

Retinal-derived tumour TRP-1/SV40 Tag PAI-/Brain Upregulated FGF1 [103] MMTV mouse mammary tumour virus, WAP whey acidic protein, PyVmT polyoma virus middle T, TRb thyroid hormone receptor beta, TRP-1 tyrosinerelated protein 1, MSC mesenchymal stem cells, LN lymph node

Thyroid

Ptc+/-, p53-/

Medullo-blastoma

Min/-

Targeted deletion of PTEN and p53 in LN, spleen, liver, diaphragm Dergulated mTOR pathway, [319] bladder epithelium with adeno-Cre virus rapamycin

Bladder

LN

[316] [317] [318] [158] [160]

Kras/LkbL/L

Dominant negative PLCg Silibinin Oncolytic HSV adenovirus MSC targeted TGFb

Lung

Lung, Liver, kidney Lung Lung

[315] [169]

TRAMP/PTEN+/ TRAMP-C2

Sca1 stem cells PET imaging CD44 knockout

Prostate

Lung Lung Lung, liver

P53/RB mutants MYCN P53tm1+/CD44+/+

Osteo-sarcoma


462

S.A. Eccles

MDA MB 231 breast cancer brain metastases [42]. This is important because HER2 has been linked to an increased frequency of brain metastases and trastuzumab is unable to cross the blood–brain barrier, leading to a high incidence of relapse at this site. Trastuzumab was shown to be effective in a breast cancer xenograft bone metastasis model, but effects were limited if administered when lesions were already established [43]. Sunitinib, a multitargeted angiogenesis inhibitor, improved survival of transgenic mice bearing metastatic lung tumors induced by mutant Kras plus knock out of the suppressor Lkb1. However, the incidence of local and distant metastases was not reduced, suggesting that the benefits were primarily due to effects on the primary tumor [44]. Clearly, these data have implications for clinical studies. Increasingly, genetic analyses of metastases from patients and experimental models are yielding potential new targets for therapy [45]. Examples include TRKB, identified in a screen for inhibitors of anoikis [46] and Axl, implicated in metastasis and angiogenesis [47]. These data, combined with elegant validation studies will provide an armamentarium of selective inhibitors whose judicious use should help to overcome target redundancy or escape mechanisms and allow combination molecular therapies to rival and perhaps ultimately replace complex and toxic chemotherapy regimes. 19.2.2.2 HSP90 Chaperone The HSP90 chaperone, responsible for the correct location and folding of cellular proteins, has emerged as a key novel therapeutic target, and several drugs are now in clinical trial [48, 49]. Because its client proteins belong to multiple signaling pathways, a single inhibitor can provide the equivalent of “multitargeted” or combinatorial therapy, and resistance appears to be a relatively rare event [50]. Many client proteins are key oncogenes (such as mutant BRAF, ErbB2, AKT) and others are important in invasion and angiogenesis (e.g. HIF-1a, FAK, c-MET, VEGFR). Interestingly, extracellular HSP90 has been linked specifically with invasion and metastasis [51]. Several inhibitors have shown efficacy in preclinical metastasis models, for example, NVP-AUY922 in orthotopic PC3 prostate carcinoma (lymph nodes), BRAF mutant WM266.4 melanoma (lung) [52], and 17-DMAG in HCT116 colon carcinoma (liver) [53]. Geldanamycin also inhibited RIP-Tag2 liver metastasis as detected by MRI [54]. However, surprisingly, in MDA MD 231 [55] and PC3-M xenografts [56], 17-AAG enhanced bone metastases. In the latter case, this was linked to activation of c-Src in osteoclasts. It will be important to determine if nongeldanamycin drug classes have this undesirable profile and whether bone metastasis would be a contraindication in patients. 19.2.2.3 Chemokine Receptors Several chemokine receptors have been implicated in site-specific metastasis [57, 58]. CXCR4 is linked to breast cancer metastasis in nodes – liver, lung, and bone – sites expressing high levels of its ligand SDF-1a/CXCL12. Treatment with the CXCR4 inhibitor CTCE-9908 peptide prior to intracardiac or i.v. injection of MDA MB 231


463

breast carcinoma cells surprisingly failed to reduce the number of metastases, but decreased their size in all organs studied (e.g. lungs, bone, viscera) [59]. A different peptide (TN14003) inhibited the growth of orthotopic SCCHN xenografts via suppression of angiogenesis and thence reduced lung metastasis [60]. In further reports, AMG487, a small molecule inhibitor of CXCR3 reduced lung metastasis from human osteosarcoma SAOS2 [61], syngeneic C26, and xenogeneic HT29 colon carcinoma but not liver metastases [62]. In the latter case, efficacy was seen when animals were dosed several days after i.v. tumor cell inoculation, suggesting effects on survival/proliferation of extravasated cells, rather than on prevention of anoikis during dissemination [63] or initial homing/seeding. Other examples of chemokine receptors implicated in cancer dissemination include CCR7 (lymphatic metastasis in melanoma, squamous cell carcinoma, GI cancers and others; CNS infiltration of T-cell leukemia) and CCR10 (melanoma skin metastases). However, relatively few selective small molecule inhibitors exist and it is important to determine that any such agents inhibit outgrowth of metastases at multiple sites (rather than selectively or prophylactically) to give true survival advantages. Nevertheless, agents such as CXCR4 inhibitors CTCE-9908 and RCP-168 and the CXCR1/2 inhibitor meraxin are currently in clinical trial [64, 65]. 19.2.2.4 BCr-Abl: A Paradigm for Tumor-Specific Therapy The Bcr-Abl translocation is a paradigm for small molecule targeted therapy [66]. The fusion protein expressed from the Philadelphia chromosome initiates and, more importantly, maintains the malignant phenotype in some chronic myeloid leukemias. Imatinib, and later drugs designed to inhibit resistant cells with additional mutations, have been the inspiration for further molecularly targeted agents. However, in solid cancers, single, primary driver mutations are rare, and sometimes only the consequences of such deregulations (i.e. hyperactivated signaling pathways) are evident. Also, in many cases it is loss of a tumor (or metastasis) suppressor such as BRCA1, P53, PTEN, nm23, BRMS1, etc. that contributes to malignant progression. 19.2.2.5 Hedgehog (Hh) The Hh pathway has been shown to be important in Ras-driven melanomas [67]. Small molecule inhibitors were found to be effective in medulloblastomas in Ptc+/-, p53-/- transgenic mice [68] and inhibited systemic metastasis in orthotopic pancreatic cancer xenografts [69]. Interestingly, although primary tumor growth was not inhibited, the proportion of putative cancer “stem-like” cells was reduced, reinforcing the notion that Hh represents a key stem cell signaling pathway. Paracrine Hh signaling from tumor cells to stroma has been linked to desmoplasia. In an interesting approach, IPI-926, a drug that depletes tumor stroma by inhibition of the Hh signaling pathway, enhanced delivery of gemcitabine to transgenic KPC mice with pancreatic carcinomas. Improved responses to the cytotoxic agents were obtained with fewer liver metastases. This was linked to increased vascular density and perfusion [70].

464

S.A. Eccles

19.2.2.6 Wnt Pathway Wnt signaling plays major roles in stem cell maintenance and has been implicated in metastasis via canonical (b-catenin stabilization) and several noncanonical pathways [71]. In an orthotopic model of basal breast cancer, Wnt signaling was linked to epithelial–mesenchymal transition (EMT) and lung metastasis. Inhibiting signaling through LRP6 co-receptors reduced the self-renewal and lung colonization potential and induced differentiation markers [72]. In a lung xenograft tumor model, EMT was not observed, but the development of brain and bone metastases following intracardiac inoculation of cells was linked to WNT/TCF-LEF1-HOXB9 signaling [73]. Wnt signaling has also been suggested to contribute to prostate cancer bone metastasis [74] and to the acquisition of an invasive cancer stem cell phenotype [75]. So far, developing pharmacological inhibitors of the Wnt pathway(s) has proved challenging, but, providing that the function of normal stem cells (e.g. in the gut and bone marrow) can be spared, it could be an interesting therapeutic target. 19.2.2.7 Combination Therapies As with chemotherapy, where multiple drug regimens are the norm, several studies are exploring combination targeted therapy. For example, an angiogenic (aV) integrin antagonist plus an antibody–cytokine fusion protein gave synergistic activity against spontaneous liver metastases of a mouse neuroblastoma [76]. Similarly, an antiangiogenic urokinase-derived peptide combined with tamoxifen inhibited the growth and lymphatic metastasis of a rat mammary carcinoma better than single agents [77]. Due to the requirement for a reproducibly high incidence of metastases, most such complex studies have been performed in syngeneic systems. However recently, combinations of EGFR and VEGFR2 inhibitors were shown to be effective in an orthotopic model of oral cancer lymph node metastases [78].

19.2.3 Processes Linked to Metastasis 19.2.3.1 Angiogenesis and Hypoxia Neoangiogenesis is generally considered a prerequisite for sustained tumor growth and spread. Many different antagonists including natural inhibitors, antibodies, and soluble decoys targeting VEGF or small molecule inhibitors of VEGFR1 and 2 have been investigated and several have reached clinical trial (reviewed in [79–81]). Of concern is the fact that few antiangiogenic strategies were tested preclinically in metastatic models. It now seems that under certain circumstances, some agents can promote tumor progression. Downregulation of VEGF may lead to compensatory upregulation of alternative angiogenic factors, or successful inhibition (resulting in hypoxia) can unleash transcriptional programs enhancing cell survival, motility, and invasion.


465

Knockout or inhibition of VEGF signaling reduced growth (but enhanced invasion and/or metastasis) of orthotopic gliomas and RIP1-Tag2 transgenic pancreatic carcinomas via “adaptive-evasive” responses [82]. Similarly, short-term treatment with sunitinib or DC101 (anti-VEGFR2 antibody) resulted in accelerated multiorgan metastasis from MDA MB 231 cells in spite of potent inhibitory effects on localized tumors [83]. It was hypothesized that upregulation of pro-angiogenic cytokines facilitated an enhanced “premetastatic niche” involving circulating endothelial cell precursors, myeloid progenitors, CXCR4+ and VEGFR1+ bone marrow-derived cells (BMDC). Clinically, antiangiogenic agents have shown little benefit as single agents, but have improved survival when combined with chemotherapy. However, some patients with glioma develop resistance to anti-VEGF therapy and can relapse with extensive tumor spread. Taken together, these preclinical and clinical observations suggest that antiangiogenic agents should be more carefully evaluated in a range of metastatic models before clinical deployment, with a greater emphasis on developing sensitive biomarkers of response (or compensatory/pro-angiogenic rebound) [84]. What is more, it is likely that combinations of inhibitors will be most effective in controlling tumor progression. Antibodies and other inhibitors directed against integrins such as aVb3 or aVb5 expressed on activated endothelial cells have been developed. Again, caution is required since low concentrations of the RGD-mimetic peptide cilengitide targeting aVb3/aVb5 stimulated angiogenesis and growth of B16F10 melanoma or Lewis lung carcinoma [85]. Interestingly, aVb3 exogenously expressed on breast carcinoma cells can promote metastasis to bone [86]. Other integrins have been implicated in lymphangiogenesis and metastasis (notably in the premetastatic niche) and several inhibitors are in clinical development [87]. Also, VEGFR3 is expressed on lymphatic endothelial cells and may promote lymphatic metastasis [88, 89]. Interestingly, a multikinase inhibitor (E7080) targeting VEGFR2 and 3 was able to inhibit both lung and lymph node metastases from orthotopic MDA MB 231, whereas bevacizumab (anti-VEGF) significantly inhibited only lung metastases [90]. Hypoxia has been linked to invasion, metastasis, and resistance to therapy [91] and the HIF-1a pathway has been considered a viable target [92, 93]. Hypoxia increased HT1080 sarcoma lung metastasis by enhancing postextravasation survival [94]. HIF-1a can also recruit BMDC to the tumor site, including MMP9+ cells which, in glioblastoma, initiate the angiogenic switch. However, abrogation of HIF1a or VEGF in a mouse model of glioma enhanced deep invasion into the brain parenchyma [95]. Hypoxic cells can be targeted either by prodrugs specifically activated under low oxygen conditions, or by hypoxic radiosensitizers such as TX-1877, which showed some efficacy in metastatic orthotopic colon cancer and rectal cancer xenografts (although no prolongation of survival [96]). Recently, lysyl oxidases have been identified as key mediators of hypoxiainduced invasion and metastasis, and in particular are critical to premetastatic niche formation by crosslinking collagen IV and recruiting CD11b+ myeloid cells [97]. Antibodies against LOXL2 inhibited gastric carcinoma xenograft metastasis [98] and a pharmacological inhibitor (BAPN) reduced the development of bone

466

S.A. Eccles

etastases from MDA MB 231 cells injected into the heart. However, there was no m effect on established metastases, which together with Erler’s data suggests that LOX is mainly required for early organ colonization [99]. 19.2.3.2 Proteolysis in Invasion and Angiogenesis Proteolytic enzymes are considered important in tumor cell invasion, extravasation, and angiogenesis, and particular models (including knockout and transgenic animals) have been useful in determining their modes of action. First and second generation MMP inhibitors proved disappointing in the clinic. More extensive studies of the full panoply of tumor (and host) associated proteases have revealed hitherto unsuspected complexity and an appreciation that some proteases are “antitargets” whose inhibition could promote tumor progression. It is now clearer which MMPs (or ADAMs) could make good therapeutic targets, and efforts are underway to develop more selective inhibitors [7] and to measure protease activity noninvasively by imaging [100]. Galardin/GM6001, a broad-spectrum MMP inhibitor, significantly reduced tumor growth by twofold and spontaneous lung metastases by 100-fold [101]. The uPA/uPAR/PAI1 system has been implicated in tumor growth, metastasis, and angiogenesis. Interestingly, uPAR is expressed on disseminated cells in bone marrow and may regulate the shift between dormancy and proliferation via a fibronectin/ integrin-mediated process [102]. Components of the signaling complex are considered promising targets for therapy, but progress has been hampered, as in many protease systems, by the complexity of regulatory networks and apparently contradictory inhibitory or tumor promoting actions of several key players. For example, the uPA inhibitor PAI-1 was shown to contribute to brain metastasis from transgenic retinal tumors [103]. Nevertheless, new insights and structure–function relationships are emerging that will aid drug discovery [104]. WX-UK1, a derivative of 3-aminophenylaniline, has been shown to inhibit lung and lymph node metastases of syngeneic BN472 rat mammary carcinomas [105] and ONO-4817 inhibited experimental lung metastases of a variety of MMP-expressing human tumors [106]. 19.2.3.3 Intravasation and Extravasation Organ tropism is to some extent linked to adhesive interactions between tumor cells and vascular endothelia, but the ability to cross this barrier is essential if overt metastases are to develop. In an orthotopic rat breast cancer model, EGFR promoted tumor cell motility and invasion but not intravasation, whereas ERBB2 was required for the latter [35]. The ability of SK-Mel28 human melanoma cells to cross the blood–brain barrier is facilitated by melanotransferrin and prevented by targeted antibodies [107]. In contrast, in a breast cancer xenograft (CN34-BrM2) capable of brain metastasis from either intra-arterial injection or from orthotopic sites, this process was linked to expression of the brain-specific sialyltransferase ST6GALNAC5 [19]. These studies have been extended using an elegant series of organotropic models, mainly based on MDA MB 231 by Massague’s group. Some common factors were found to be utilized


467

by cancer cells for early steps in metastasis, e.g. molecules enhancing vascular permeability and mimicking lymphocyte extravasation, others were site selective. Molecules such as epiregulin, COX-2, MMP1, and MMP2 were identified from a “lung metastasis signature” and collectively mediate vascular remodeling and promote angiogenesis and tumor cell intravasation and extravasation [108]. Also the cytokine angiopoietin-like 4 (ANGPTL4) was specifically linked to lung metastasis (reviewed in [109]). However, although mechanistically interesting, the clinical possibilities of therapeutic interventions at such early stages are limited and factors that are required for survival and/or expansion of disseminated cells should be the main focus of attention for effective, potentially curative therapy. 19.2.3.4 The Premetastatic Niche It has recently been recognized that establishment of a successful metastasis may depend upon preconditioning of target organs by factors released by the primary tumor and involving recruitment of BMDC to a premetastatic niche [110, 111], elements of which may be targetable to prevent metastases. For example, osteopontin has been implicated in BMDC recruitment and activation of dormant metastases [112] and SU6668, by virtue of its activity against VEGFR1, was shown to decrease CXCL1 levels and neutrophil infiltration in premetastatic liver and to inhibit metastasis of orthotopic TK-4 colon carcinomas [113]. Metastasis of cancer stem-like cells, particularly to bone marrow, has also been linked to the SDF-1-CXCR4 axis [114]. However, again, it is not clear how long disseminated cells are dependent on support from these specialized microenvironments before becoming autonomous.

19.2.4 Resistance to Therapy Intrinsic or acquired resistance to conventional cytotoxic therapy is common and targeted therapies are not exempt. Both tumor and host factors can contribute via mechanisms including those mediated by cell–stroma and cell–cell interactions. This may be overcome by targeting the stromal cells, their secreted paracrine survival factors, or the proteasome (reviewed in [115]). Resistance in a mouse model of Bcr-Abl ALL to imatinib has been shown to be due to cytokines such as IL-7 released from the hematopoietic microenvironment [116]. Although perhaps rarer than resistance caused by secondary mutations or “kinase switching,” such mechanisms are likely to arise as multikinase inhibitors are increasingly deployed. 19.2.4.1 Cancer Stem-Like Cells Recently, much attention has been given to the possibility that tumor relapse, metastasis, and treatment failure may be due to the presence of cancer stem-like progenitor cells [117, 118], which in some cases has been linked to EMT [119]. Thus, methods to assay these cells in human tumor xenografts and other models have been

468

S.A. Eccles

d eveloped [120]. Invasive prostate cancer cell lines were found to express a stem cell-like phenotype (CD44+ CD24-) [121] and breast cancer cell lines contained a subpopulation of aldefluor+ cells with metastatic potential [122]. MDA MB 468 human mammary carcinoma cells selected for lymphatic metastasis also expressed a “stem cell”-like phenotype with an enhanced ability to survive in ectopic environments [123]. Cancer stem-like cells may have a low proliferative index, enhanced motility, express drug efflux pumps, reside in specialized niches, and are driven by different signaling pathways than more differentiated progeny [124, 299]. Recently, temozolomide was found to enhance the proportion of putative stem cells in a transgenic PTEN null glioma model and these residual cells had enhanced tumorigenic potential [125]. We must consider how such dangerous, residual cells, spared – or possibly even enriched – by conventional cytotoxic therapies, could be eliminated. Interestingly, HER2 overexpression increases the proportion of stem-like cells via Notch activation [126, 127] and thus trastuzumab or lapatinib efficacy may be linked to their ability to target these cells. Telomerase may also provide a suitable tumor stem cell therapeutic target [128], and an antagonist (GRN163L) has shown activity in prevention of A549 lung metastases [129]. Other possibilities include CD44, Hedgehog and Wnt pathways, TGFb, CXCR4, or Notch [117, 130]. 19.2.4.2 Dormant Metastases The presence of single tumor cells (or small clusters) in sites such as nodes or bone marrow suggests that early dissemination may seed micrometastases, which remain dormant until reactivated [124, 131–133]. Such cells may present a risk of recurrence, since if noncycling they are likely to be resistant to most forms of therapy; mechanisms responsible for their awakening therefore need to be identified and targeted. Dormancy may be due to immune restraint, lack of angiogenesis, or the tissue microenvironment failing to provide a congenial “soil” for the metastatic “seed” [134–137]. Genomic profiling of breast carcinoma, glioblastoma, and sarcoma xenografts emerging from dormancy recognized a transcriptional switch primarily linked to angiogenesis, but with upregulation of unexpected genes such as EGFR, IGF-1R, TIMP-3, and others [138]. Metastasis suppressor genes have also been implicated [139]. Alternatively, transition from quiescence to proliferation may be partly regulated by microenvironmental cues (e.g. fibronectin) leading to changes in cytoskeletal architecture. In murine mammary carcinomas and osteosarcomas, reactivated proliferation was inhibited by targeting b1 integrin or myosin light chain kinase [140]. In mouse mammary carcinoma models in which liver metastases arise after a long latency, dormant cells were resistant to doxorubicin, although actively growing macrometastases were inhibited [141, 142]. Dormant cells can be labeled with GFP or luciferase to follow their fate in vivo, and this approach (and patient studies) has demonstrated that viable, nonproliferating cells can persist for long periods [143]. Further investigation is required to evaluate the commonest mechanisms of dormancy in order to maintain this state or to target vulnerabilities that may differ from those in actively proliferating cells.


469

19.2.5 Immunological Approaches 19.2.5.1 Antibody-Based Therapies Immune effectors have exquisite specificity and should (in the absence of untoward cross-reactivity with normal tissues) give minimal toxicity. They are most effective against low tumor burden, i.e. minimal residual disease or micrometastases, as a useful adjunct to conventional debulking cytotoxic therapies. Many tumor antigens including CEA, PSMA, EGFR, c-erbB-2, gangliosides, etc. have been used as targets. Antibodies may be used to block growth factor receptors (EGFR, c-erbB-2, VEGFR) and/or to recruit complement or host effectors or to deliver toxins or cytotoxic drugs [144]. Radioimmunotherapy using anti-CEA antibodies has been explored in colorectal tumor lung and liver metastasis models [145, 146] and also in a transgenic breast cancer model [147]. Therapeutic effects were observed, although with these direct conjugates, normal tissue toxicity can be an issue. Antibody-directed enzyme prodrug therapy is one example of “pretargeting” (reviewed in [148]) where an antibody conjugate localizes to tumor deposits, then is allowed to clear from the circulation before a second moiety is administered. More complex therapeutic strategies have been devised which require correspondingly elegant models. Chimeric antibodies designed to target a human antigen and to recruit host effector cells have been assayed in a SCID/hu mouse model of neuroblastoma metastatic to liver [149], where the mouse bone marrow is reconstituted with human stem cells capable of maturing into effector cells. Other strategies have combined the targeting ability of antibodies with cytokines in pulmonary and hepatic metastasis models of syngeneic mouse melanomas, neuroblastomas, and colorectal carcinomas [150].

19.2.5.2 Vaccines, Cytokines, and Cell-Mediated Immunotherapy B700 antigen on B16F10 mouse melanoma cells was used in a vaccine that inhibited spontaneous lymph node and lung metastases and combination with cytokines such as IL-2, Il-12, or GM-CSF potentiated the effects. This model, while mimicking the human disease in terms of primary growth, bone marrow invasion, regional node involvement, and distant dissemination, nevertheless could not distinguish antimetastatic effects from indirect effects on the primary tumor [151]. Vaccination against HER2/neu has been tested in rat Mat-Ly-Lu Dunning prostate carcinoma [152] and orthotopic transgenic or xenograft metastatic murine breast tumors [153–155]. Adoptive immunotherapy with IL-2 and human effector cells injected into the tumor site (OSC-19 cells in the floor of the mouth) in athymic mice resulted in fewer lymph node metastases [156]. The main considerations for selecting models for trials of targeted immunotherapeutic agents are an understanding of the antigen(s) involved and their relative

470

S.A. Eccles

immunogenicity in rodents and humans. Also, it is important that the interventions are done (if possible) in animals bearing primary tumors (or those which have had them surgically removed) because the immunocompetence of naive animals may not equate to that of animals “conditioned” by the presence of a tumor. 19.2.5.3 Targeting Using Vectors or Peptides with Tumor Selectivity Delivery of prodrug activating enzyme genes such as Herpes simplex thymidine kinase, followed by therapy with ganciclovir can inhibit tumor growth and metastasis. However, because of the current difficulties of generating viral or nonviral vectors that are sufficiently stable to survive in the circulation and/ or their intrinsic immunogenicity, most gene therapy studies depend upon local or regional delivery. Oncolytic viral therapy has been achieved in a rat model of hepatic metastasis [157] and in TRAMP transgenic prostate lymphatic metastases [158]. Also, the A33 antigen was used to target virus to LoVo hepatic metastases [159]. Recently, mesenchymal stem cells have been used to target IFNb to lung metastases in the TRAMP model [160], and tumor-tropic neural progenitor cells expressing osteoprotegerin limited development of neuroblastoma bone metastases [161]. Vectors may be targeted to organs using tumor-specific promoters such as AFP (hepatoma), c-erbB-2 (breast, ovarian, and gastric cancer), and CEA (colorectal cancer). Alternatively, ligands or antibody fragments can redirect the vector to receptor-overexpressing tumor cells. Antisense approaches have mainly been directed against molecules associated with invasion and metastasis, e.g. CD44 v6 and matrilysin (MMP-7) in liver metastasis models of colorectal carcinoma. Increasingly, given the utility of siRNA to silence genes in vitro, attempts are being made to achieve in vivo delivery of stable hairpin RNAs or oligonucleotides, with some success in preclinical models [162, 163]. Novel targets have been identified on tumor cells, metastases, and tumor vasculature by selective binding of peptides or antibody fragments. A FITC-conjugated 5-amino acid peptide (TMTP1) homed to and detected PC3 lymph node and MKN gastric cancer liver micrometastases and could serve as a vector for therapeutic isotopes or drugs [164]. Such small peptides readily diffuse and penetrate tissues, are nonimmunogenic and easily synthesized.

19.3 Detection and Quantitation of Metastases and Determination of Therapeutic Benefit The most significant recent developments have been the use of tumor cells genetically tagged with fluorescent or luminescent markers, enabling routine localization and sequential measurements of metastases by optical imaging.


471

What is more, tumor-host interactions can be discerned by color coding different cell populations, including the generation of fluorescent transgenic mice [165]. Individual fluorescent tumor cells can be detected at high resolution by in vivo videomicroscopy [166]. In vivo measurements of gene expression, angiogenesis, and metabolism are possible using multiphoton laser scanning microscopy, and imaging in the skin or skull have extended the range of such intravital technologies [167]. In addition, such markers can be used in reporter assays of target function, for example, to indicate the activation status of a signaling pathway [168] or protease activity by using a quenched substrate [100]. Examples of different techniques used experimentally are given in Table 19.4. Other imaging modalities can be used to provide functional information: 18 FDG positron emission tomography (PET) and 18F-fluoride PET identified MCYN transgenic osteosarcoma bone and lung metastases. When the transgene was inactivated, glucose uptake was reduced and fluoride uptake increased, indicating bone remodeling. Such systems would be invaluable in experimental therapy studies to follow both proliferation and differentiation [169]. FDG-PET detected responses of orthotopic rat C6 gliomas to therapy with temozolomide and hypoxia inhibitors [170] and localized colon carcinoma lung metastases of ~0.3 mm diameter [171]. PET was used to track virus delivery to lymph node micrometastases in gene therapy of B16 melanoma [172] and to show downregulation of ErbB2/HER2 oncogene expression in response to an HSP90 inhibitor [173]. A strain of transgenic mouse with a fluorescent reporter has also been developed for studying PI (3,4,5)P(3) metabolism [168]. Magnetic resonance (MR) imaging and Doppler ultrasound are useful for measuring vascular density, permeability, and blood flow, and MRI was used to assay responses to Recentin (VEGFR inhibitor) in a DU145 brain metastasis model [174]. The fate of single MDA MB 231BR cells delivered to the brain via the left ventricle has also been demonstrated [175]. Recently, dual-modality uPAR targeted nanoparticles have been developed for molecular imaging (by optical and MR methods) of pancreatic xenografts and metastases [176]. 3D high frequency ultrasound is also used to measure tumor volume in liver [177] and also to ablate liver metastases in rat models [178]. With the increasing use of patient-like tumor models and targeted therapies, molecular imaging will become an intrinsic part of drug development, providing not only noninvasive measurements of tumor growth and spread but also enabling interrogation of tumor metabolism, proliferation, and vascularization [179–181]. Also, key functions associated with metastasis such as adhesion, matrix interactions, and even intracellular signaling are becoming accessible [182]. Ex vivo analyses of endothelial precursor cells or tumor cells in circulation, bone marrow, or nodes also add to our understanding of the dissemination and lodgment phases of metastasis and are applicable both to animal models and clinical studies [183, 184].

472

S.A. Eccles

19.4 Animal Models for Evaluating Targeted Therapy of Metastasis A multitude of preclinical tumor models are available to facilitate analysis of the molecular mechanisms of metastasis and for evaluating novel therapeutic approaches. Given the huge number (>200) of different human cancers, their intrinsic heterogeneity plus the confounding factors of host genetic backgrounds, no particular model can serve as an appropriate tool for all applications. The following sections illustrate the main types of animal tumor systems available (with specific examples), their strengths and weaknesses, and utilization in different experimental settings.

19.4.1 Syngeneic Rodent Tumor Models The first experimental tumors were chemically induced or arose in cancer prone strains, usually due to oncogenic viruses such as MMTV. Metastases were rare, although sometimes manifest if the primary tumor was surgically removed to allow longer survival of the host. Sublines of B16 melanoma, Lewis lung (3LL) carcinoma, and the 4T1 breast carcinoma with enhanced and/or organ-selective metastasis were derived. The importance of using strictly syngeneic, inbred animals for transplantation studies has been recognized, and indeed has allowed the identification of significant “metastasis modifier” genes in different mouse strains [185]. Most chemically induced tumors are highly immunogenic, unlike those arising spontaneously (in mouse or man). Several molecules first identified in rodent tumors are important in the malignant process in human cancers, and in some cases elicit cell-mediated and/or humoral responses, e.g. c-erbB-2/HER2, MUC-1, NG2, MAGE antigens. There are now a wider variety of syngeneic tumor models available (Table 19.1), although there is an increasing trend toward the use of human tumor xenografts and/or transgenic models.

19.4.2 Human Tumor Xenograft Models Table 19.2 illustrates a selection of reliable xenograft tumors, focusing on metastatic models. Although nu/nu or SCID mice hosts are the norm, nu/nu rats, and even more recently chick embryos are also used [186]. Metastatic models in zebra fish embryos are likely to emerge because transformed melanocytes with migratory ability have been generated [71]. SCID mice, lacking both B and T cell immunity, and bg mice with lower NK activity are often more susceptible to metastasis than athymic mice. Orthotopic 4T1 (mouse) and MTLn3 (rat) breast carcinoma metastasis was significantly increased in NK deficient Rag2(-/-)gC(-/-) mice [187]. Similar results were seen using human SCLC cells injected s.c. into pfp/rag2 mice which showed extensive lung metastases [188].

Cervical ca xenograft Transgenic Knockout mice lacking specific enzymes HIF – luc A431

HIF activity/hypoxia Angiogenesis

Application Liver metastasis growth + doxorubicin Liver metastasis, GFP-tagged cells, dormancy Brain metastasis Brain metastasis + anti-VEGF inhibitor AZD2171 Tumour motility and invasion Inhibition of liver and kidney mets by silibinin BM, LN and tumour cells in blood after mfp, i.v. or i.c. injection Lung metastases LN micrometastases Downregulation of HER2 detected by (64)Cu-DOTA-trastuzumab Lung metastases – osteopontin aptamers Subrenal – lung metastases – PD-0332991 (CDK4/6), avastin (VEGF) Bone metastases – dasatanib, 17-AAG Liver metastases Lung and CNS metastasis, also detects differentially labelled adoptively transferred NK-T cells Visceral metastases – HSP90 inhibitors Protease activity/inhibition PI3 kinase activity bioprobe [214] [322]

[306] [100, 168]

[287, 321]

[169] [172] [173] [213] [56, 211]

Reference(s) [177] [320] [175] [174] [259] [181, 317, 318] [183]

Mfp mammary fat pad, i.v. intravenous, i.c. intracardiac, PET positron emission tomography, NMR nuclear magnetic resonance, MRI magnetic resonance imaging, HF-VPDU high frequency volumetric power Doppler ultrasound

Luminescence-based Reporter assays HF-VPDU

Fluorescence Functional fluorescent optical imaging

MDA MB 231 breast ca MDA MB 435

Bioluminescence (BLI)

PC-3M HCT116 BCL1 and A20

MYCN transgenic osteosarcoma B16 + oncolytic virus targeting SKOV3 + 17AAG

PET imaging

Table 19.4 Evaluation of invasion, metastases, or response to therapy Method Model High frequency ultrasound B16F1, HT-29, MDA MB 435 Intravital videomicroscopy B16, colon26 MRI MDA MB231BR DU145 Multiphoton confocal microscopy MTLn3 NMR Metabolomics TRAMP FACS/LCS MDA MB 435HAL (GFP)


474

S.A. Eccles

Tumors grown s.c, although convenient for readily measuring responses to therapy, are less likely to metastasize than those grown orthotopically (see Sect. 19.4.4.1). Preserving the tissue architecture by implanting tumor fragments, either direct from patients or following a single s.c. passage can also maintain malignant potential better than using cultured cells and several key molecular drivers such as EGFRvIII or Hh may be rapidly lost in vitro. Malignant potential can be enhanced by selection of tumorigenic/metastatic variants and/or co-injection with matrigel, fibroblasts [189], or mesenchymal stem cells [190]. These procedures presumably mimic to some extent the orthotopic microenvironment or at least overcome some of the defects of the xenogeneic/ectopic environment. Metastatic xenograft tumor models have been used for target identification, validation, and evaluation of therapeutic agents. Some cell lines have been “custom built” to explore synergies between oncogenes and to generate more highly malignant tumors. For example, MCF10A mammary epithelial cells were transfected with H-RAS (which transforms cells) and with BMI-1 (which inhibits the INK4/ARF tumor suppressor locus and is implicated in stem cell maintenance). The doubly transfected cells generated both lung and brain metastases, and BMI1 knockdown reversed this trait [191].

19.4.3 Organ Colonization and Site-Selective Metastases The simplest “metastasis” assays, aiming to mimic late stages of metastasis (dissemination, extravasation, and colonization), are achieved by inoculation of a bolus of tumor cells directly into the peripheral circulation to give lung colonies. Tumors derived from cells that are naturally migratory (e.g. leukemias, lymphomas, plasmacytomas) more readily colonize multiple downstream sites including marrow, spleen, and liver. Xenograft models of CNS and visceral metastases of human monocytic leukemia are now available [192]. Tumor cells can also be directly injected into other vessels (or target organs) to generate additional models of disseminated disease (see below). These models are more widely used now that noninvasive optical imaging of internal tumors is readily achievable. 19.4.3.1 Lung Metastases Although i.v. injection of tumor cells yields “metastases” there is no direct correlation between lung colonizing ability and spontaneous metastatic potential or their responses to therapy. RON RTK was validated as a potential therapeutic target in HCT116 colon carcinoma cells as its knockdown inhibited metastasis from cecum to lung [193]. Minn et al. developed MDA MB 231 breast carcinoma sublines with enhanced pulmonary metastasis and identified a few key genes that either gave growth advantages at both primary and secondary sites (e.g. epiregulin, CXCL1, MMP-1, COX2) or were selective for enhancing growth in the lung microenvironment (e.g. SPARC, MMP2) [194].


475

Interestingly, inhibitors of any single target failed to control metastasis, whereas combinations, particularly of cetuximab and celecoxib, with or without the MMP inhibitor GM6001 were highly effective [108]. 19.4.3.2 Liver Metastasis Liver metastases (appropriate mainly for colon or pancreatic carcinomas) can be generated by inoculation of cells into a mesenteric vein. The spleen provides a simpler alternative since cells pass almost immediately into the portal circulation. However, growth and intraperitoneal spread from the primary tumor can be a confounding factor without splenectomy. Such colon liver metastases were used to show efficacy of a farnesyl transferase inhibitor [195]. When limited numbers of colonies are required (e.g. for focussed ultrasound therapy [178] or photodynamic therapy), tumor cells can be injected directly into the liver. Since the hepatic microenvironment reportedly upregulates EGFR and/or its ligands [196], EGFR inhibitors such as cetuximab or erlotinib could be beneficial and clinical trials seem promising [197]. 19.4.3.3 Brain Metastasis Brain metastasis models have recently been developed from the injection of tumor cells either into a carotid artery [198] or the left ventricle of the heart [19] with subsequent harvesting and recycling by the same means [199]. Interestingly, by selecting glioma cells first for lung colonization, sublines were generated which had a greater invasive potential when re-implanted into the brain. These DMB2 tumors were shown to respond to the HSP90 inhibitor 17-AAG, assessed in part by real-time ultrasound imaging [200]. MDA MB 435 LvBr2 and Br4 sublines injected into the carotid artery express high levels of angiogenic factors including angiopoietins and VEGF, and also elevated Notch signaling. Their invasive potential was reduced by g-secretase inhibitors [198]. Direct injection of human SCLC cells into nude rat brains can be used to mimic metastases and to investigate means of selectively permeabilizing the blood–brain barrier to enhance selective delivery of therapeutic agents [201].

19.4.3.4 Bone Metastasis Breast, prostate, neuroblastoma, and myeloma models of bone metastasis have been generated from direct implantation into the tibial marrow space [202, 203] into human bone fragments inserted s.c. [204] or by selection from the bone after intracardiac inoculation of cells [205]. Sublines of the syngeneic mouse 4T1 tumor or rat MatLyLu tumors spread to the skeleton after mfp [86] or i.v. injection [206].

476

S.A. Eccles

Interestingly, variants of LNCaP, which can metastasize to bone from s.c. orthotopic or intracardiac routes were obtained following various experimental selection procedures including exposure to hypoxia [207] or coculture with osteoblasts [208]. The resultant tumors lost androgen dependence and gained osteomimetic properties. Such tumors have been very instructive in developing novel gene therapies and methods to interfere with tumor–stromal interactions, e.g. with anti-integrin antibodies [208]. Inhibitors of IkB kinase inhibited bone metastases from rat W256 carcinosarcoma cells injected via the left ventricle [209] and in a similar xenograft model (PC3-ML), an anti-PDGFRa antibody inhibited early phases of bone metastasis establishment, prior to osteoclast recruitment [210]. In the case of neuroblastoma bone metastases, IGF-1R was implicated as a major mediator of osteolysis [202]. Latent metastases in bone marrow can be a major problem in breast and other cancers. Recently, using bone-tropic MDA MB 231 and CN34 human tumor xenografts (BoM-1833 and BoM2 sublines), Src was identified as a key factor enabling cells to respond to CXCL12 and resist TRAIL-mediated apoptosis in this microenvironment. Importantly, in clinical samples, a Src gene expression signature correlated with late bone relapse, whereas a TGFb signature was associated with lung metastases. Hence Src inhibition could be a valid therapeutic strategy to inhibit these latent cells [211]. The many models representing different sites of metastasis in both syngeneic and xenogeneic tumor systems have yielded important insights into the molecular mechanisms of organotropism. Such models are now frequently employed in preclinical evaluation of targeted therapies. However, results obtained in “experimental” metastasis assays should be treated with caution and confirmed in spontaneous metastasis assays. Most importantly, therapies for clinical use must be able to inhibit the growth of established metastases, often simultaneously developing in multiple organs. Many studies commence therapy on the same day as (or before) tumor cell inoculation or even pretreat tumor cells with inhibitory compounds or by genetic manipulation. While this is a reasonable strategy for early validation of the role of a specific molecule in particular stages of the metastatic cascade, such approaches cannot be used as a substitute for well-designed studies of therapeutic agents in established malignant disease.

19.4.4 Spontaneous Metastasis Models “Spontaneous” metastasis refers to the seeding of cells from a primary site to generate detectable lesions at distant sites. Several points are worthy of note: firstly, where possible, the primary tumor should be surgically excised to allow time for metastases to develop. Alternatively, micrometastases may be identified using “tagged” cells, although it should be confirmed that these are extravascular and clonogenic rather than simply in transit. Providing that the fluorescent or luminescent signal is stable (and this cannot be assumed), the presence and to some extent the size of metastatic lesions can be determined.


477

All suspected metastatic deposits should be confirmed histologically. It is unethical and unscientific to allow primary tumors, particularly those in vital organs, to reach such gross dimensions that “metastasis” into other sites cannot be distinguished from direct extension or transcelomic seeding. Clinically, most systemic therapy for metastases (detectable or suspected) is in an adjuvant setting, postsurgical removal or downstaging of the primary tumor. Ideally, animal experiments should be similarly designed. If therapy is commenced while the primary is in situ, and results in a significant inhibition of tumor growth, it cannot be concluded that an “antimetastatic” effect was not simply a consequence of a lower tumor burden and thus less cell shedding. 19.4.4.1 Orthotopic Implantation Models The orthotopic implantation of breast tumor cells into the mammary fat pad usually encourages the development of lung and/or lymphatic metastases, although spread to other sites common in human disease (e.g. bone, liver, and brain) is rare without further genetic manipulation or selection. Xenografted tumors also give a higher frequency of metastases when implanted orthotopically, and models of human colorectal cancer metastatic to liver, melanoma metastatic to nodes, prostate cancer metastatic to nodes and bone, pancreatic cancer metastatic to liver, and many other models are available (see Tables 19.1 and 19.2). These are increasingly being used in the preclinical workup of novel therapies [32–34]. In an example of adjuvant therapy, the EGFR inhibitor erlotinib was administered to mice following surgical removal of orthotopic MDA-MB-435 LvBr tumors and found to reduce subsequent lung metastasis [212]. In a novel approach, RNA aptamer blockade of osteopontin was shown to inhibit spontaneous lung metastases from orthotopic MDA-MB-231 breast carcinomas [213]. Liver metastasis from orthotopic SUIT-2 pancreatic carcinoma xenografts was inhibited using a prodrug specifically activated in hypoxic cells by caspase 3 activation and induction of apoptosis [214] and also by an anti-MET antibody or ligand antagonist [215]. 19.4.4.2 Lymph Node Metastases Lymphatic metastases are important for clinical staging, can induce significant morbidity, and potentially act as a bridgehead for wider dissemination [216, 217]. Several rat (e.g. HOSP1, MTLn3) and human (e.g. MDA MB 468, MDA MB 435) breast carcinomas give rise to spontaneous lymphatic metastases when grown in mammary fat pads. PTEN null PC3 orthotopic prostate carcinoma xenografts reliably metastasize to regional and distant lymph nodes and have been used to demonstrate efficacy of several novel targeted agents including inhibitors of Src (dasatinib) [218], HSP90 (NVP-AUY922) [52], and PI3 kinase (GDC-0941) [40]. Also, radioimmunoconjugates of alpha-emitting radioisotopes targeted to cell

478

S.A. Eccles

s urface uPAR inhibited primary tumor growth and lymph node metastases [219]. Other models include orthotopic human kidney and mouse pancreatic cancers in which nanoparticle-mediated delivery of doxorubicin targeted to aVb3 on tumor vasculature inhibited lymph node and/or visceral metastases [220]. With the discovery of important cytokine signaling pathways linked to lymphangiogenesis and lymphatic metastasis (e.g. VEGF-C:VEGFR3; CCL19:CCR7) and their consideration as targets for therapy, there is a need to enhance the models available. Lewis lung carcinoma cells express multiple lymphangiogenic cytokines and MMPs. Metastasis from lung to lymph nodes was inhibited by the MMP inhibitor MM1270 [221]. Transfection of VEGF-C into weakly metastatic LAPC-9 human prostate carcinoma cells (or cells with naturally high levels) showed extensive lymphangiogenesis and lymphatic metastasis and responded to antibodies targeted to VEGFR3 or a soluble ligand trap [222, 223]. In the same models, VEGFsiRNA or anti-VEGFR2 antibody reduced systemic metastasis but not nodal metastasis. This illustrates the importance of testing novel therapeutic agents against both hematogenous and lymphatic metastasis as they may not be effective in both. Indeed in a study of the antiangiogenic agent AGM-1470, hematogenous metastasis was reduced but lymphatic metastasis increased [224].

19.4.5 Transgenic Models This topic will be covered in Chap. 30, but models suitable for evaluation of targeted therapies against metastasis will be discussed briefly. Examples are shown in Table 19.3. Note that different genetic backgrounds (e.g. FVB vs. C57Bl/6J strains) can significantly influence tumor phenotypes and metastatic potential [185, 225]. Several cancer-prone transgenic mouse strains have been produced but development of clinically relevant tumor types and/or metastasis is not assured. Animals expressing a human oncogene such as c-erbB-2, where the gene not only initiates oncogenesis, but is also an ideal target for therapy are of particular value. Early work used the mutant rat neu gene to induce tumors at high frequency, which were metastatic. However, HER-2/neu was generally thought to be oncogenic simply due to gene amplification in humans, although alternatively spliced forms resembling the spontaneously mutated/activated forms have been identified in some human cancers [226]. MMTV-c-erbB-2/HER2 (wild type) transgenic mice develop mammary cancers with long latency that metastasize to the lung, although some carry additional somatic mutations in the transgene. Polyoma virus middle T (PyMT) transgenic mice rapidly develop metastatic mammary carcinomas [227, 228]. Both models have been quite extensively used to explore vaccination strategies, monoclonal antibodies, and targeted drugs [229]. However, in humans, c-erbB-2 is also expressed at low levels in certain normal cells hence tissue damage (e.g. cardiac toxicity with Herceptin) will not be detected. These models are also useful for target validation. For example, EphA2 receptor was identified as a promoter of


479

metastasis in MMTV-neu tumors by amplifying ErbB2 signaling, but not in MMTV-PyVMT tumors. Such information would help to define the patient population likely to respond to EphA2 inhibitors [230]. Similarly, conditional expression of TGFb in PyVMT models enhanced lung metastasis in the absence of effects on primary tumor growth. TGFb antisense [231] and dominant negative PLCg [232] inhibited lung metastases in PyVMT and TRAMP mice. In both models, AKT2 (but not AKT1) was found to enhance the incidence of metastases in the absence of effects on primary tumor latency [37]. On the other hand, genetic ablation of MMP-3 in PyVMT mammary tumors did not affect tumor growth or metastasis [233]. Interestingly, in an elegant recent study Husemann et al. determined that the initiation of metastasis in both MMTV-HER2 and PyVMT models was much earlier than previously thought, with disseminated cells detectable in blood, marrow, and lungs even before invasion was manifest [14]. Clinical studies also support the possibility that metastasis may be a relatively early event in cancer progression. These findings have implications both for the need to eradicate incipient metastases and show how the intelligent use of animal models may address these issues Table 19.4. The TRAMP model targets simian virus 40 T antigen to the prostate using the rat probasin gene. Mice develop metastases primarily in lung and lymph nodes, with up to 100% incidence by 28 weeks [234]. In a different approach, transgenic mice have been generated with a PSA tissue expression pattern very similar to that in humans [235] rather than confined to the prostate. In these immunocompetent mice, PSA is a normal “self” antigen, and if metastatic, this would represent an ideal model system for testing the feasibility of PSA targeted therapies. The main drawbacks of the transgenic systems for testing targeted therapies are their relatively high variability, long latency, incomplete penetrance, and development of multiple tumors. The frequency (and sites) of metastases may also be limited and/or unpredictable [236]. Indeed, none of the commonly used Apc (Min) mouse strains consistently develop metastases [237]. Alternative strategies have aimed to combine the benefits of controlled transgene oncogenesis with higher throughput and convenience. One approach is to transplant the autochthonous primary tumors or cell lines derived from them into recipient hosts [238]. TRAMP-C cell lines have been used in such models to evaluate therapies in an adjuvant setting [239]. Finally, producing double or multiply transgenic mice has been used to enhance the malignancy of developing tumors and/or to generate tumors with desired therapeutic targets. These approaches have indicated that, in prostate cancers at least, two “hits” are required for progression from benign to invasive tumors, and up to five for metastasis. PTEN deficiency also accelerated metastasis from transgenic MMTV-NIC ErbB2 breast cancers [240] and TRb thyroid cancers [241]. Such manipulations provide both mechanistic insights into tumor progression and also more reliable and realistic models for preclinical drug evaluation. Increasing use of “knock-in” systems is allowing the development of models exemplified by those such as the PB-Cre4 x PTEN(loxP/loxP) mice which provide a continuum from tumor initiation to metastasis [242].

480

S.A. Eccles

A further application has been the development of mice in which transgene expression can be controlled by doxycycline. This conditional expression of oncogenes (or knock out of suppressor genes) also overcomes the significant issue of embryonic lethality and enables the transforming effects to be manifest in adult tissues. By this means it was demonstrated that Her2/neu was essential for the maintenance of mammary tumors and lung metastases, not just their initiation, providing confidence that the model was appropriate for evaluation of targeted therapies in established disseminated disease. Nevertheless, Neu-independent tumors eventually emerged, which could provide a useful model of breaking dormancy [243]. New models of pancreatic cancer have been generated with conditional mutations in both p53 and Kras, which show appropriate metastatic patterns. This model was “credentialed” by showing similar gene and protein expression levels and responses to drugs used in the clinic (e.g. gemcitabine) [244]. Recently, additional metastatic transgenic models of different tumor types have been developed. These include squamous cell cancers with lymphatic metastases such as oral cancers driven by MyrAKT/Trp53-/-, which interestingly showed high levels of putative stem cells [245] and skin cancers in which a mutant b1 integrin collaborated with Ras mutations induced by a chemical carcinogen [246]. Several excellent reviews [38, 244, 247–252] discuss the relative merits of various animal models for drug development and are recommended for additional reading.

19.5 Summary and Conclusions Animal models have provided great insights into the process of metastasis, generated ideas for molecular targets, and have subsequently been used for preclinical evaluation of novel therapies. So why are we still faced, for several of the major cancers, with death rates which have changes little, in spite of advances in early detection and therapeutic options? Are the animal models to blame? Clearly, they represent imperfect systems and cannot represent the complexity and heterogeneity of human malignancies. However, there is much still to be learned about their optimization and rational deployment. Factors controlling tumor cell invasion, dissemination, preconditioning of metastatic “niches”, extravasation, and lodgement at secondary sites are of enormous scientific interest, but unless growth of established micrometastases is controlled, cures of disseminated cancer will remain limited. Cytotoxic therapy is designed to attack systemic disease, but fails perhaps because of metastatic heterogeneity [3], innate or acquired resistance, dormancy or the failure to evaluate such agents in appropriate metastatic tumor models. Too often, the claimed inhibition of metastasis in preclinical studies is secondary to reductions in primary tumor growth. In other cases, therapy is commenced before or at the same time as systemic injection of tumor cells. Although studies of metastasis prevention are of interest mechanistically, more attention must be given to the factors controlling ectopic tumor survival


481

and proliferation at secondary sites. For example, in spite of the key role of CXCR4 in promoting breast cancer xenograft metastasis, the potent inhibitor AMD3100 failed to prolong survival of mice bearing established lung metastases, showing that its role (perhaps as predicted) is primarily in early phases of metastasis [253]. Also, if stem-like tumor cells are responsible for treatment failure, we need to understand their molecular drivers and inhibit these to achieve complete control [34]. Indeed, we probably need to target the bulk population using cytoreductive therapies and the putative stem-like cells, which may be responsible for later relapse. Clearly, there are challenges ahead to discover key pivotal (or complementary) points for intervention and to identify molecular targets that subsume site selectivity, but with our rapidly increasing knowledge of basic molecular mechanisms, this may ultimately be achieved [32, 33, 254].

References 1. Iiizumi M, Liu W, Pai SK, Furuta E, Watabe K. Drug development against metastasis-related genes and their pathways: a rationale for cancer therapy. Biochim Biophys Acta. 2008;1786(2):87–104. 2. Perret GY, Crepin M. New pharmacological strategies against metastatic spread. Fundam Clin Pharmacol. 2008;22(5):465–92. 3. Steeg PS. Heterogeneity of drug target expression among metastatic lesions: lessons from a breast cancer autopsy program. Clin Cancer Res. 2008;14(12):3643–5. 4. Steeg PS, Anderson RL, Bar-Eli M, Chambers AF, Eccles SA, Hunter K., Itoh K., Matrisian LM, Sleeman JP, Theodorescu D., Thompson EW, Welch DR. Preclinical drug development must consider the impact on metastasis. Clin Cancer Res. 2009;15(14):4529–30. 5. Gura T. Systems for identifying new drugs are often faulty. Science. 1997;278(5340):1041–2. 6. Li M, Huang X, Zhu Z, Zhao Q, Wong M, Gorelik E. The therapeutic efficacy of angiostatin against weakly- and highly-immunogenic 3LL tumors. In Vivo. 2002;16(6):577–82. 7. Overall CM, Kleifeld O. Towards third generation matrix metalloproteinase inhibitors for cancer therapy. Br J Cancer. 2006;94(7):941–6. 8. Molloy T, van’t Veer LJ. Recent advances in metastasis research. Curr Opin Genet Dev. 2008;18(1):35–41. 9. Mazzocca A, Carloni V. The metastatic process: methodological advances and pharmacological challenges. Curr Med Chem. 2009;16(14):1704–17. 10. Walker I, Newell H. Do molecularly targeted agents in oncology have reduced attrition rates? Nat Rev Drug Discov. 2009;8(1):15–6. 11. Phan JH, Moffitt RA, Stokes TH, et al. Convergence of biomarkers, bioinformatics and nanotechnology for individualized cancer treatment. Trends Biotechnol. 2009;27(6):350–8. 12 Hoffman RM. Orthotopic metastatic mouse models for anticancer drug discovery and evaluation: a bridge to the clinic. Invest New Drugs. 1999;17(4):343–59. 13. Husemann Y, Klein CA. The analysis of metastasis in transgenic mouse models. Transgenic Res. 2009;18(1):1–5. 14. Husemann Y, Geigl JB, Schubert F, et al. Systemic spread is an early step in breast cancer. Cancer Cell. 2008;13(1):58–68. 15. Pantel K, Brakenhoff RH, Brandt B. Detection, clinical relevance and specific biological properties of disseminating tumour cells. Nat Rev Cancer. 2008;8(5):329–40. 16. Klein CA. Parallel progression of primary tumours and metastases. Nat Rev Cancer. 2009;9(4):302–12.

482

S.A. Eccles

17. Smith SC, Theodorescu D. Learning therapeutic lessons from metastasis suppressor proteins. Nat Rev Cancer. 2009;9(4):253–64. 18. Benson JD, Chen YN, Cornell-Kennon SA, et al. Validating cancer drug targets. Nature. 2006;441(7092):451–6. 19. Bos PD, Zhang XH, Nadal C, et al. Genes that mediate breast cancer metastasis to the brain. Nature. 2009;459(7249):1005–9. 20. Klein A, Olendrowitz C, Schmutzler R, et al. Identification of brain- and bone-specific breast cancer metastasis genes. Cancer Lett. 2009;276(2):212–20. 21. Minn AJ, Kang Y, Serganova I, et al. Distinct organ-specific metastatic potential of individual breast cancer cells and primary tumors. J Clin Invest. 2005;115(1):44–55. 22. Collins I, Workman P. New approaches to molecular cancer therapeutics. Nat Chem Biol. 2006;2(12):689–700. 23. Janne PA, Gray N, Settleman J. Factors underlying sensitivity of cancers to small-molecule kinase inhibitors. Nat Rev Drug Discov. 2009;8(9):709–23. 24. Green TP, Fennell M, Whittaker R, et al. Preclinical anticancer activity of the potent, oral Src inhibitor AZD0530. Mol Oncol. 2009;3:248–61. 25. Hoeflich KP, Herter S, Tien J, et al. Antitumor efficacy of the novel RAF inhibitor GDC-0879 is predicted by BRAFV600E mutational status and sustained extracellular signal-regulated kinase/ mitogen-activated protein kinase pathway suppression. Cancer Res. 2009;69(7):3042–51. 26. Hoeflich KP, O’Brien C, Boyd Z, et al. In vivo antitumor activity of MEK and phosphatidylinositol 3-kinase inhibitors in basal-like breast cancer models. Clin Cancer Res. 2009;15(14):4649–64. 27. Sarkar D, Boukerche H, Su ZZ, Fisher PB. mda-9/syntenin: recent insights into a novel cell signaling and metastasis-associated gene. Pharmacol Ther. 2004;104(2):101–15. 28 Cao Y, Cao R, Hedlund EM. Regulation of tumor angiogenesis and metastasis by FGF and PDGF signaling pathways. J Mol Med. 2008;86(7):785–9. 29. Chalhoub N, Baker SJ. PTEN and the PI3-kinase pathway in cancer. Annu Rev Pathol. 2009;4:127–50. 30. Comoglio PM, Giordano S, Trusolino L. Drug development of MET inhibitors: targeting oncogene addiction and expedience. Nat Rev Drug Discov. 2008;7(6):504–16. 31. Dankort D, Curley DP, Cartlidge RA, et al. Braf(V600E) cooperates with Pten loss to induce metastatic melanoma. Nat Genet. 2009;41(5):544–52. 32. Eccles SA. Parallels in invasion and angiogenesis provide pivotal points for therapeutic intervention. Int J Dev Biol. 2004;48(5–6):583–98. 33. Eccles SA. Targeting key steps in metastatic tumour progression. Curr Opin Genet Dev. 2005;15(1):77–86. 34. Eccles SA, Welch DR. Metastasis: recent discoveries and novel treatment strategies. Lancet. 2007;369(9574):1742–57. 35. Kedrin D, Wyckoff J, Boimel PJ, et al. ERBB1 and ERBB2 have distinct functions in tumor cell invasion and intravasation. Clin Cancer Res. 2009;15(11):3733–9. 36. Araujo J, Logothetis C. Targeting Src signaling in metastatic bone disease. Int J Cancer. 2009;124(1):1–6. 37. Dillon RL, Marcotte R, Hennessy BT, Woodgett JR, Mills GB, Muller WJ. Akt1 and akt2 play distinct roles in the initiation and metastatic phases of mammary tumor progression. Cancer Res. 2009;69(12):5057–64. 38. Troiani T, Schettino C, Martinelli E, Morgillo F, Tortora G, Ciardiello F. The use of xenograft models for the selection of cancer treatments with the EGFR as an example. Crit Rev Oncol Hematol. 2008;65(3):200–11. 39. Beloueche-Babari M, Jackson LE, Al-Saffar NM, et al. Identification of magnetic resonance detectable metabolic changes associated with inhibition of phosphoinositide 3-kinase signaling in human breast cancer cells. Mol Cancer Ther. 2006;5(1):187–96. 40. Raynaud FI, Eccles SA, Patel S, et al. Biological properties of potent inhibitors of class I phosphatidylinositide 3-kinases: from PI-103 through PI-540, PI-620 to the oral agent GDC0941. Mol Cancer Ther. 2009;8(7):1725–38.


483

41. Molinari F, Martin V, Saletti P, et al. Differing deregulation of EGFR and downstream proteins in primary colorectal cancer and related metastatic sites may be clinically relevant. Br J Cancer. 2009;100(7):1087–94. 42. Gril B, Palmieri D, Bronder JL, et al. Effect of lapatinib on the outgrowth of metastatic breast cancer cells to the brain. J Natl Cancer Inst. 2008;100(15):1092–103. 43. Khalili P, Arakelian A, Chen G, Singh G, Rabbani SA. Effect of Herceptin on the development and progression of skeletal metastases in a xenograft model of human breast cancer. Oncogene. 2005;24(44):6657–66. 44. Gandhi L, McNamara KL, Li D, et al. Sunitinib prolongs survival in genetically engineered mouse models of multistep lung carcinogenesis. Cancer Prev Res (Phila Pa). 2009;2(4):330–7. 45. Kluger HM, Chelouche Lev D, Kluger Y, et al. Using a xenograft model of human breast cancer metastasis to find genes associated with clinically aggressive disease. Cancer Res. 2005;65(13):5578–87. 46. Geiger TR, Peeper DS. Critical role for TrkB kinase function in anoikis suppression, tumorigenesis, and metastasis. Cancer Res. 2007;67(13):6221–9. 47. Li Y, Ye X, Tan C, et al. Axl as a potential therapeutic target in cancer: role of Axl in tumor growth, metastasis and angiogenesis. Oncogene. 2009;28:3442–55. 48. Koga F, Kihara K, Neckers L. Inhibition of cancer invasion and metastasis by targeting the molecular chaperone heat-shock protein 90. Anticancer Res. 2009;29(3):797–807. 49. Powers MV, Workman P. Inhibitors of the heat shock response: biology and pharmacology. FEBS Lett. 2007;581(19):3758–69. 50. Gaspar N, Sharp SY, Pacey S, et al. Acquired resistance to 17-allylamino-17-demethoxygeldanamycin (17-AAG, tanespimycin) in glioblastoma cells. Cancer Res. 2009;69(5):1966–75. 51. Tsutsumi S, Neckers L. Extracellular heat shock protein 90: a role for a molecular chaperone in cell motility and cancer metastasis. Cancer Sci. 2007;98(10):1536–9. 52. Eccles SA, Massey A, Raynaud FI, et al. NVP-AUY922: a novel heat shock protein 90 inhibitor active against xenograft tumor growth, angiogenesis, and metastasis. Cancer Res. 2008;68(8):2850–60. 53. Moser C, Lang SA, Kainz S, et al. Blocking heat shock protein-90 inhibits the invasive properties and hepatic growth of human colon cancer cells and improves the efficacy of oxaliplatin in p53-deficient colon cancer tumors in vivo. Mol Cancer Ther. 2007;6(11):2868–78. 54. Cao X, Jia G, Zhang T, et al. Non-invasive MRI tumor imaging and synergistic anticancer effect of HSP90 inhibitor and glycolysis inhibitor in RIP1-Tag2 transgenic pancreatic tumor model. Cancer Chemother Pharmacol. 2008;62(6):985–94. 55. Price JT, Quinn JM, Sims NA, et al. The heat shock protein 90 inhibitor, 17-allylamino-17demethoxygeldanamycin, enhances osteoclast formation and potentiates bone metastasis of a human breast cancer cell line. Cancer Res. 2005;65(11):4929–38. 56. Yano A, Tsutsumi S, Soga S, et al. Inhibition of Hsp90 activates osteoclast c-Src signaling and promotes growth of prostate carcinoma cells in bone. Proc Natl Acad Sci USA. 2008;105(40):15541–6. 57. Ben-Baruch A. Organ selectivity in metastasis: regulation by chemokines and their receptors. Clin Exp Metastasis. 2008;25(4):345–56. 58. Dithmar S, Rusciano D, Grossniklaus HE. A new technique for implantation of tissue culture melanoma cells in a murine model of metastatic ocular melanoma. Melanoma Res. 2000;10(1):2–8. 59. Richert MM, Vaidya KS, Mills CN, et al. Inhibition of CXCR4 by CTCE-9908 inhibits breast cancer metastasis to lung and bone. Oncol Rep. 2009;21(3):761–7. 60. Yoon Y, Liang Z, Zhang X, et al. CXC chemokine receptor-4 antagonist blocks both growth of primary tumor and metastasis of head and neck cancer in xenograft mouse models. Cancer Res. 2007;67(15):7518–24. 61. Pradelli E, Karimdjee-Soilihi B, Michiels JF, et al. Antagonism of chemokine receptor CXCR3 inhibits osteosarcoma metastasis to lungs. Int J Cancer. 2009;125:2586–94.

484

S.A. Eccles

62. Cambien B, Karimdjee BF, Richard-Fiardo P, et al. Organ-specific inhibition of metastatic colon carcinoma by CXCR3 antagonism. Br J Cancer. 2009;100(11):1755–64. 63. Kochetkova M, Kumar S, McColl SR. Chemokine receptors CXCR4 and CCR7 promote metastasis by preventing anoikis in cancer cells. Cell Death Differ. 2009;16(5):664–73. 64. Balakin KV, Ivanenkov YA, Tkachenko SE, Kiselyov AS, Ivachtchenko AV. Regulators of chemokine receptor activity as promising anticancer therapeutics. Curr Cancer Drug Targets. 2008;8(4):299–340. 65. Wu X, Lee VC, Chevalier E, Hwang ST. Chemokine receptors as targets for cancer therapy. Curr Pharm Des. 2009;15(7):742–57. 66. Muller BA. Imatinib and its successors – how modern chemistry has changed drug development. Curr Pharm Des. 2009;15(2):120–33. 67. Stecca B, Mas C, Clement V, et al. Melanomas require HEDGEHOG-GLI signaling regulated by interactions between GLI1 and the RAS-MEK/AKT pathways. Proc Natl Acad Sci USA. 2007;104(14):5895–900. 68. Romer J, Curran T. Targeting medulloblastoma: small-molecule inhibitors of the Sonic Hedgehog pathway as potential cancer therapeutics. Cancer Res. 2005;65(12):4975–8. 69. Feldmann G, Fendrich V, McGovern K, et al. An orally bioavailable small-molecule inhibitor of Hedgehog signaling inhibits tumor initiation and metastasis in pancreatic cancer. Mol Cancer Ther. 2008;7(9):2725–35. 70. Olive KP, Jacobetz MA, Davidson CJ, et al. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science. 2009;324(5933):1457–61. 71. Jessen JR. Noncanonical Wnt signaling in tumor progression and metastasis. Zebrafish. 2009;6(1):21–8. 72. DiMeo TA, Anderson K, Phadke P, et al. A novel lung metastasis signature links Wnt signaling with cancer cell self-renewal and epithelial-mesenchymal transition in basal-like breast cancer. Cancer Res. 2009;69(13):5364–73. 73. Nguyen DX, Chiang AC, Zhang XH, et al. WNT/TCF signaling through LEF1 and HOXB9 mediates lung adenocarcinoma metastasis. Cell. 2009;138(1):51–62. 74. Hall CL, Kang S, MacDougald OA, Keller ET. Role of Wnts in prostate cancer bone metastases. J Cell Biochem. 2006;97(4):661–72. 75. Neth P, Ries C, Karow M, Egea V, Ilmer M, Jochum M. The Wnt signal transduction pathway in stem cells and cancer cells: influence on cellular invasion. Stem Cell Rev. 2007;3(1):18–29. 76. Lode HN, Moehler T, Xiang R, et al. Synergy between an antiangiogenic integrin alphav antagonist and an antibody-cytokine fusion protein eradicates spontaneous tumor metastases. Proc Natl Acad Sci USA. 1999;96(4):1591–6. 77. Guo Y, Mazar AP, Lebrun JJ, Rabbani SA. An antiangiogenic urokinase-derived peptide combined with tamoxifen decreases tumor growth and metastasis in a syngeneic model of breast cancer. Cancer Res. 2002;62(16):4678–84. 78. Sano D, Choi S, Milas ZL, et al. The effect of combination anti-endothelial growth factor receptor and anti-vascular endothelial growth factor receptor 2 targeted therapy on lymph node metastasis: a study in an orthotopic nude mouse model of squamous cell carcinoma of the oral tongue. Arch Otolaryngol Head Neck Surg. 2009;135(4):411–20. 79. Cao Y. Tumor angiogenesis and molecular targets for therapy. Front Biosci. 2009;14:3962–73. 80. De Paepe B. Anti-angiogenic agents and cancer: current insights and future perspectives. Recent Pat Anticancer Drug Discov. 2009;4(2):180–5. 81. Grimm D, Bauer J, Schoenberger J. Blockade of neoangiogenesis, a new and promising technique to control the growth of malignant tumors and their metastases. Curr Vasc Pharmacol. 2009;7(3):347–57. 82. Paez-Ribes M, Allen E, Hudock J, et al. Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis. Cancer Cell. 2009;15(3):220–31. 83. Ebos JM, Lee CR, Cruz-Munoz W, Bjarnason GA, Christensen JG, Kerbel RS. Accelerated metastasis after short-term treatment with a potent inhibitor of tumor angiogenesis. Cancer Cell. 2009;15(3):232–9.


485

84. Jain RK, Duda DG, Willett CG, et al. Biomarkers of response and resistance to antiangiogenic therapy. Nat Rev Clin Oncol. 2009;6(6):327–38. 85. Reynolds AR, Hart IR, Watson AR, et al. Stimulation of tumor growth and angiogenesis by low concentrations of RGD-mimetic integrin inhibitors. Nat Med. 2009;15(4):392–400. 86. Sloan EK, Pouliot N, Stanley KL, et al. Tumor-specific expression of alphavbeta3 integrin promotes spontaneous metastasis of breast cancer to bone. Breast Cancer Res. 2006;8(2):R20. 87. Avraamides CJ, Garmy-Susini B, Varner JA. Integrins in angiogenesis and lymphangiogenesis. Nat Rev Cancer. 2008;8(8):604–17. 88. Achen MG, McColl BK, Stacker SA. Focus on lymphangiogenesis in tumor metastasis. Cancer Cell. 2005;7(2):121–7. 89. Makinen T, Alitalo K. Lymphangiogenesis in development and disease. Novartis Found Symp. 2007;283:87–98; discussion 105, 238–41. 90. Matsui J, Funahashi Y, Uenaka T, Watanabe T, Tsuruoka A, Asada M. Multi-kinase inhibitor E7080 suppresses lymph node and lung metastases of human mammary breast tumor MDA-MB-231 via inhibition of vascular endothelial growth factor-receptor (VEGF-R) 2 and VEGF-R3 kinase. Clin Cancer Res. 2008;14(17):5459–65. 91. Rapisarda A, Melillo G. Role of the hypoxic tumor microenvironment in the resistance to anti-angiogenic therapies. Drug Resist Updat. 2009;12(3):74–80. 92. Onnis B, Rapisarda A, Melillo G. Development of HIF-1 inhibitors for cancer therapy. J Cell Mol Med. 2009;13:2780–6. 93. Poon E, Harris AL, Ashcroft M. Targeting the hypoxia-inducible factor (HIF) pathway in cancer. Expert Rev Mol Med. 2009;11:e26. 94. Zhang L, Hill RP. Hypoxia enhances metastatic efficiency in HT1080 fibrosarcoma cells by increasing cell survival in lungs, not cell adhesion and invasion. Cancer Res. 2007;67(16):7789–97. 95. Du R, Lu KV, Petritsch C, et al. HIF1alpha induces the recruitment of bone marrow-derived vascular modulatory cells to regulate tumor angiogenesis and invasion. Cancer Cell. 2008;13(3):206–20. 96. Miyake K, Shimada M, Nishioka M, et al. Downregulation of matrix metalloprotease-9 and urokinase plasminogen activator by TX-1877 results in decreased tumor growth and metastasis on xenograft model of rectal cancer. Cancer Chemother Pharmacol. 2009;64(5):885–92. 97. Erler JT, Bennewith KL, Cox TR, et al. Hypoxia-induced lysyl oxidase is a critical mediator of bone marrow cell recruitment to form the premetastatic niche. Cancer Cell. 2009;15(1):35–44. 98. Ran Y, Peng L, Hu H, et al. Secreted LOXL2 is a novel therapeutic target that promotes gastric cancer metastasis via the Src/FAK pathway. Carcinogenesis. 2009;30:1660–9. 99. Bondareva A, Downey CM, Ayres F, et al. The lysyl oxidase inhibitor, beta-aminopropionitrile, diminishes the metastatic colonization potential of circulating breast cancer cells. PLoS One. 2009;4(5):e5620. 100. McIntyre JO, Matrisian LM. Optical proteolytic beacons for in vivo detection of matrix metalloproteinase activity. Methods Mol Biol. 2009;539:155–74. 101. Almholt K, Juncker-Jensen A, Laerum OD, et al. Metastasis is strongly reduced by the matrix metalloproteinase inhibitor Galardin in the MMTV-PymT transgenic breast cancer model. Mol Cancer Ther. 2008;7(9):2758–67. 102. Allgayer H, Aguirre-Ghiso JA. The urokinase receptor (u-PAR) – a link between tumor cell dormancy and minimal residual disease in bone marrow? APMIS. 2008;116(7–8):602–14. 103. Maillard CM, Bouquet C, Petitjean MM, et al. Reduction of brain metastases in plasminogen activator inhibitor-1-deficient mice with transgenic ocular tumors. Carcinogenesis. 2008;29(11):2236–42. 104. Vincenza Carriero M, Franco P, Vocca I, et al. Structure, function and antagonists of urokinase-type plasminogen activator. Front Biosci. 2009;14:3782–94. 105. Setyono-Han B, Sturzebecher J, Schmalix WA, et al. Suppression of rat breast cancer metastasis and reduction of primary tumour growth by the small synthetic urokinase inhibitor WX-UK1. Thromb Haemost. 2005;93(4):779–86.

486

S.A. Eccles

106. Shiraga M, Yano S, Yamamoto A, et al. Organ heterogeneity of host-derived matrix metalloproteinase expression and its involvement in multiple-organ metastasis by lung cancer cell lines. Cancer Res. 2002;62(20):5967–73. 107. Rolland Y, Demeule M, Fenart L, Beliveau R. Inhibition of melanoma brain metastasis by targeting melanotransferrin at the cell surface. Pigment Cell Melanoma Res. 2009;22(1):86–98. 108. Gupta GP, Nguyen DX, Chiang AC, et al. Mediators of vascular remodelling co-opted for sequential steps in lung metastasis. Nature. 2007;446(7137):765–70. 109. Nguyen DX, Bos PD, Massague J. Metastasis: from dissemination to organ-specific colonization. Nat Rev Cancer. 2009;9(4):274–84. 110. Kaplan RN, Psaila B, Lyden D. Bone marrow cells in the ‘pre-metastatic niche’: within bone and beyond. Cancer Metastasis Rev. 2006;25(4):521–9. 111. Psaila B, Lyden D. The metastatic niche: adapting the foreign soil. Nat Rev Cancer. 2009;9(4):285–93. 112. McAllister SS, Gifford AM, Greiner AL, et al. Systemic endocrine instigation of indolent tumor growth requires osteopontin. Cell. 2008;133(6):994–1005. 113. Yamamoto M, Kikuchi H, Ohta M, et al. TSU68 prevents liver metastasis of colon cancer xenografts by modulating the premetastatic niche. Cancer Res. 2008;68(23):9754–62. 114. Kucia M, Reca R, Miekus K, et al. Trafficking of normal stem cells and metastasis of cancer stem cells involve similar mechanisms: pivotal role of the SDF-1-CXCR4 axis. Stem Cells. 2005;23(7):879–94. 115. Meads MB, Gatenby RA, Dalton WS. Environment-mediated drug resistance: a major contributor to minimal residual disease. Nat Rev Cancer. 2009;9:665–74. 116. Williams RT, den Besten W, Sherr CJ. Cytokine-dependent imatinib resistance in mouse BCR-ABL+, Arf-null lymphoblastic leukemia. Genes Dev. 2007;21(18):2283–7. 117. Croker AK, Allan AL. Cancer stem cells: implications for the progression and treatment of metastatic disease. J Cell Mol Med. 2008;12(2):374–90. 118. Mimeault M, Batra SK. Functions of tumorigenic and migrating cancer progenitor cells in cancer progression and metastasis and their therapeutic implications. Cancer Metastasis Rev. 2007;26(1):203–14. 119. Polyak K, Weinberg RA. Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nat Rev Cancer. 2009;9(4):265–73. 120. Kang Y. Analysis of cancer stem cell metastasis in xenograft animal models. Methods Mol Biol. 2009;568:7–19. 121. Klarmann GJ, Hurt EM, Mathews LA, et al. Invasive prostate cancer cells are tumor initiating cells that have a stem cell-like genomic signature. Clin Exp Metastasis. 2009;26(5):433–46. 122. Charafe-Jauffret E, Ginestier C, Birnbaum D. Breast cancer stem cells: tools and models to rely on. BMC Cancer. 2009;9:202. 123. Pandit TS, Kennette W, Mackenzie L, et al. Lymphatic metastasis of breast cancer cells is associated with differential gene expression profiles that predict cancer stem cell-like properties and the ability to survive, establish and grow in a foreign environment. Int J Oncol. 2009;35(2):297–308. 124. Allan AL, Vantyghem SA, Tuck AB, Chambers AF. Tumor dormancy and cancer stem cells: implications for the biology and treatment of breast cancer metastasis. Breast Dis. 2006;26:87–98. 125. Bleau AM, Hambardzumyan D, Ozawa T, et al. PTEN/PI3K/Akt pathway regulates the side population phenotype and ABCG2 activity in glioma tumor stem-like cells. Cell Stem Cell. 2009;4(3):226–35. 126. Korkaya H, Paulson A, Iovino F, Wicha MS. HER2 regulates the mammary stem/progenitor cell population driving tumorigenesis and invasion. Oncogene. 2008;27(47):6120–30. 127. Magnifico A, Albano L, Campaner S, et al. Tumor-initiating cells of HER2-positive carcinoma cell lines express the highest oncoprotein levels and are sensitive to trastuzumab. Clin Cancer Res. 2009;15(6):2010–21. 128. Dikmen ZG, Ozgurtas T, Gryaznov SM, Herbert BS. Targeting critical steps of cancer metastasis and recurrence using telomerase template antagonists. Biochim Biophys Acta. 2009;1792(4):240–7.


487

129. Dikmen ZG, Gellert GC, Jackson S, et al. In vivo inhibition of lung cancer by GRN163L: a novel human telomerase inhibitor. Cancer Res. 2005;65(17):7866–73. 130. Ischenko I, Seeliger H, Schaffer M, Jauch KW, Bruns CJ. Cancer stem cells: how can we target them? Curr Med Chem. 2008;15(30):3171–84. 131. Akslen LA, Naumov GN. Tumor dormancy – from basic mechanisms to clinical practice. APMIS. 2008;116(7–8):545–7. 132. Goss P, Allan AL, Rodenhiser DI, Foster PJ, Chambers AF. New clinical and experimental approaches for studying tumor dormancy: does tumor dormancy offer a therapeutic target? APMIS. 2008;116(7–8):552–68. 133. Hedley BD, Chambers AF. Tumor dormancy and metastasis. Adv Cancer Res. 2009;102:67–101. 134. Joyce JA, Pollard JW. Microenvironmental regulation of metastasis. Nat Rev Cancer. 2009;9(4):239–52. 135. Lunt SJ, Chaudary N, Hill RP. The tumor microenvironment and metastatic disease. Clin Exp Metastasis. 2009;26(1):19–34. 136. Fidler IJ, Poste G. The “seed and soil” hypothesis revisited. Lancet Oncol. 2008;9(8):808. 137. Mendoza M, Khanna C. Revisiting the seed and soil in cancer metastasis. Int J Biochem Cell Biol. 2009;41(7):1452–62. 138. Almog N, Ma L, Raychowdhury R, et al. Transcriptional switch of dormant tumors to fastgrowing angiogenic phenotype. Cancer Res. 2009;69(3):836–44. 139. Horak CE, Lee JH, Marshall JC, Shreeve SM, Steeg PS. The role of metastasis suppressor genes in metastatic dormancy. APMIS. 2008;116(7–8):586–601. 140. Barkan D, Kleinman H, Simmons JL, et al. Inhibition of metastatic outgrowth from single dormant tumor cells by targeting the cytoskeleton. Cancer Res. 2008;68(15):6241–50. 141. Naumov GN, Bender E, Zurakowski D, et al. A model of human tumor dormancy: an angiogenic switch from the nonangiogenic phenotype. J Natl Cancer Inst. 2006;98(5):316–25. 142. Naumov GN, Folkman J, Straume O. Tumor dormancy due to failure of angiogenesis: role of the microenvironment. Clin Exp Metastasis. 2009;26(1):51–60. 143. Wikman H, Vessella R, Pantel K. Cancer micrometastasis and tumour dormancy. APMIS. 2008;116(7–8):754–70. 144. Chari RV. Targeted cancer therapy: conferring specificity to cytotoxic drugs. Acc Chem Res. 2008;41(1):98–107. 145. Behr TM, Salib AL, Liersch T, et al. Radioimmunotherapy of small volume disease of colorectal cancer metastatic to the liver: preclinical evaluation in comparison to standard chemotherapy and initial results of a phase I clinical study. Clin Cancer Res. 1999;5(10 Suppl):3232s–42s. 146. Saga T, Sakahara H, Nakamoto Y, et al. Radioimmunotherapy for liver micrometastases in mice: pharmacokinetics, dose estimation, and long-term effect. Jpn J Cancer Res. 1999;90(3):342–8. 147. Song H, Shahverdi K, Huso DL, et al. 213Bi (alpha-emitter)-antibody targeting of breast cancer metastases in the neu-N transgenic mouse model. Cancer Res. 2008;68(10):3873–80. 148. Bagshawe KD, Sharma SK, Begent RH. Antibody-directed enzyme prodrug therapy (ADEPT) for cancer. Expert Opin Biol Ther. 2004;4(11):1777–89. 149. Sabzevari H, Gillies SD, Mueller BM, Pancook JD, Reisfeld RA. A recombinant antibody-interleukin 2 fusion protein suppresses growth of hepatic human neuroblastoma metastases in severe combined immunodeficiency mice. Proc Natl Acad Sci USA. 1994;91(20):9626–30. 150. Lode HN, Xiang R, Varki NM, Dolman CS, Gillies SD, Reisfeld RA. Targeted interleukin-2 therapy for spontaneous neuroblastoma metastases to bone marrow. J Natl Cancer Inst. 1997;89(21):1586–94. 151. Shrayer DP, Cole B, Hearing VJ, Wolf SF, Wanebo HJ. Immunotherapy of mice with an irradiated melanoma vaccine coupled with interleukin-12. Clin Exp Metastasis. 1999;17(1):63–70. 152. Bhattachary R, Bukkapatnam R, Prawoko I, Soto J, Morgan M, Salup RR. Efficacy of vaccination with plasmid DNA encoding for HER2/neu or HER2/neu-eGFP fusion protein against prostate cancer in rats. Int Immunopharmacol. 2002;2(6):783–96.

488

S.A. Eccles

153. Foy TM, Fanger GR, Hand S, Gerard C, Bruck C, Cheever MA. Designing HER2 vaccines. Semin Oncol. 2002;29(3 Suppl 11):53–61. 154. Wang X, Wang JP, Rao XM, Price JE, Zhou HS, Lachman LB. Prime-boost vaccination with plasmid and adenovirus gene vaccines control HER2/neu+ metastatic breast cancer in mice. Breast Cancer Res. 2005;7(5):R580–8. 155. Smorlesi A, Papalini F, Amici A, et al. Evaluation of different plasmid DNA delivery systems for immunization against HER2/neu in a transgenic murine model of mammary carcinoma. Vaccine. 2006;24(11):1766–75. 156. Chikamatsu K, Reichert TE, Kashii Y, et al. Immunotherapy with effector cells and IL-2 of lymph node metastases of human squamous-cell carcinoma of the head and neck established in nude mice. Int J Cancer. 1999;82(4):532–7. 157. Kooby DA, Carew JF, Halterman MW, et al. Oncolytic viral therapy for human colorectal cancer and liver metastases using a multi-mutated herpes simplex virus type-1 (G207). FASEB J. 1999;13(11):1325–34. 158. Varghese S, Rabkin SD, Nielsen GP, MacGarvey U, Liu R, Martuza RL. Systemic therapy of spontaneous prostate cancer in transgenic mice with oncolytic herpes simplex viruses. Cancer Res. 2007;67(19):9371–9. 159. Cafferata EG, Maccio DR, Lopez MV, et al. A novel A33 promoter-based conditionally replicative adenovirus suppresses tumor growth and eradicates hepatic metastases in human colon cancer models. Clin Cancer Res. 2009;15(9):3037–49. 160. Ren C, Kumar S, Chanda D, et al. Cancer gene therapy using mesenchymal stem cells expressing interferon-beta in a mouse prostate cancer lung metastasis model. Gene Ther. 2008;15(21):1446–53. 161. Sims TL, Jr., Hamner JB, Bush RA, et al. Neural progenitor cell-mediated delivery of osteoprotegerin limits disease progression in a preclinical model of neuroblastoma bone metastasis. J Pediatr Surg. 2009;44(1):204–10; discussion 10–1. 162. Lee JH, Cho ES, Kim MY, et al. Suppression of progression and metastasis of established colon tumors in mice by intravenous delivery of short interfering RNA targeting KITENIN, a metastasis-enhancing protein. Cancer Res. 2005;65(19):8993–9003. 163. Bulk E, Hascher A, Liersch R, et al. Adjuvant therapy with small hairpin RNA interference prevents non-small cell lung cancer metastasis development in mice. Cancer Res. 2008;68(6):1896–904. 164. Yang W, Luo D, Wang S, et al. TMTP1, a novel tumor-homing peptide specifically targeting metastasis. Clin Cancer Res. 2008;14(17):5494–502. 165. Tran Cao HS, Reynoso J, Yang M, et al. Development of the transgenic cyan fluorescent protein (CFP)-expressing nude mouse for “Technicolor” cancer imaging. J Cell Biochem. 2009;107(2):328–34. 166. Naumov GN, Wilson SM, MacDonald IC, et al. Cellular expression of green fluorescent protein, coupled with high-resolution in vivo videomicroscopy, to monitor steps in tumor metastasis. J Cell Sci. 1999;112 (Pt 12):1835–42. 167. Kedrin D, Gligorijevic B, Wyckoff J, et al. Intravital imaging of metastatic behavior through a mammary imaging window. Nat Methods. 2008;5(12):1019–21. 168. Sasaki T, Sasaki J, Watanabe K, Suzuki A. Non-invasive visualization of the lipid product of class I PI3K in transgenic mouse models. Biochem Soc Trans. 2007;35(Pt 2):215–8. 169. Arvanitis C, Bendapudi PK, Tseng JR, Gambhir SS, Felsher DW. (18)F and (18)FDG PET imaging of osteosarcoma to non-invasively monitor in situ changes in cellular proliferation and bone differentiation upon MYC inactivation. Cancer Biol Ther. 2008;7(12):1947–51. 170. Assadian S, Aliaga A, Del Maestro RF, Evans AC, Bedell BJ. FDG-PET imaging for the evaluation of antiglioma agents in a rat model. Neuro Oncol. 2008;10(3):292–9. 171. Sharkey RM, Karacay H, Vallabhajosula S, et al. Metastatic human colonic carcinoma: molecular imaging with pretargeted SPECT and PET in a mouse model. Radiology. 2008;246(2):497–507. 172. Brader P, Kelly K, Gang S, et al. Imaging of lymph node micrometastases using an oncolytic herpes virus and [F]FEAU PET. PLoS One. 2009;4(3):e4789.


489

173. Niu G, Li Z, Cao Q, Chen X. Monitoring therapeutic response of human ovarian cancer to 17-DMAG by noninvasive PET imaging with (64)Cu-DOTA-trastuzumab. Eur J Nucl Med Mol Imaging. 2009;36(9):1510–9. 174. JuanYin J, Tracy K, Zhang L, et al. Noninvasive imaging of the functional effects of antiVEGF therapy on tumor cell extravasation and regional blood volume in an experimental brain metastasis model. Clin Exp Metastasis. 2009;26(5):403–14. 175. Heyn C, Ronald JA, Ramadan SS, et al. In vivo MRI of cancer cell fate at the single-cell level in a mouse model of breast cancer metastasis to the brain. Magn Reson Med. 2006;56(5):1001–10. 176. Yang L, Mao H, Cao Z, et al. Molecular imaging of pancreatic cancer in an animal model using targeted multifunctional nanoparticles. Gastroenterology. 2009;136(5):1514–25 e2. 177. Graham KC, Wirtzfeld LA, MacKenzie LT, et al. Three-dimensional high-frequency ultrasound imaging for longitudinal evaluation of liver metastases in preclinical models. Cancer Res. 2005;65(12):5231–7. 178. Chen L, ter Haar G, Hill CR, Eccles SA, Box G. Treatment of implanted liver tumors with focused ultrasound. Ultrasound Med Biol. 1998;24(9):1475–88. 179. Mather S. Molecular imaging with bioconjugates in mouse models of cancer. Bioconjug Chem. 2009;20(4):631–43. 180. Willmann JK, van Bruggen N, Dinkelborg LM, Gambhir SS. Molecular imaging in drug development. Nat Rev Drug Discov. 2008;7(7):591–607. 181. Raina K, Serkova NJ, Agarwal R. Silibinin feeding alters the metabolic profile in TRAMP prostatic tumors: 1H-NMRS-based metabolomics study. Cancer Res. 2009;69(9):3731–5. 182. Carragher NO. Profiling distinct mechanisms of tumour invasion for drug discovery: imaging adhesion, signalling and matrix turnover. Clin Exp Metastasis. 2009;26(4):381–97. 183. Goodale D, Phay C, Postenka CO, Keeney M, Allan AL. Characterization of tumor cell dissemination patterns in preclinical models of cancer metastasis using flow cytometry and laser scanning cytometry. Cytometry A. 2009;75(4):344–55. 184. Pantel K, Alix-Panabieres C, Riethdorf S. Cancer micrometastases. Nat Rev Clin Oncol. 2009;6(6):339–51. 185. Hunter K. Host genetics influence tumour metastasis. Nat Rev Cancer. 2006;6(2):141–6. 186. Deryugina EI, Quigley JP. Chick embryo chorioallantoic membrane model systems to study and visualize human tumor cell metastasis. Histochem Cell Biol. 2008;130(6):1119–30. 187. Le Devedec SE, van Roosmalen W, Maria N, et al. An improved model to study tumor cell autonomous metastasis programs using MTLn3 cells and the Rag2(–/–) gammac (–/–) mouse. Clin Exp Metastasis. 2009;26:673–84. 188. Sodeur S, Ullrich S, Gustke H, Zangemeister-Wittke U, Schumacher U. Increased numbers of spontaneous SCLC metastasis in absence of NK cells after subcutaneous inoculation of different SCLC cell lines into pfp/rag2 double knock out mice. Cancer Lett. 2009;282(2):146–51. 189. Noel A, De Pauw-Gillet MC, Purnell G, Nusgens B, Lapiere CM, Foidart JM. Enhancement of tumorigenicity of human breast adenocarcinoma cells in nude mice by matrigel and fibroblasts. Br J Cancer. 1993;68(5):909–15. 190. Spaeth EL, Dembinski JL, Sasser AK, et al. Mesenchymal stem cell transition to tumor- associated fibroblasts contributes to fibrovascular network expansion and tumor progression. PLoS One. 2009;4(4):e4992. 191. Hoenerhoff MJ, Chu I, Barkan D, et al. BMI1 cooperates with H-RAS to induce an aggressive breast cancer phenotype with brain metastases. Oncogene. 2009;28:3022–32. 192. Li Z, Chen Z, Lu J, et al. Establishment of a nude mice model of human monocytic leukemia with CNS and multiorgan extramedullary infiltration. Eur J Haematol. 2006;77(2):128–33. 193. Wang J, Rajput A, Kan JL, et al. Knockdown of Ron kinase inhibits mutant phosphatidylinositol 3-kinase and reduces metastasis in human colon carcinoma. J Biol Chem. 2009;284(16):10912–22. 194. Minn AJ, Gupta GP, Siegel PM, et al. Genes that mediate breast cancer metastasis to lung. Nature. 2005;436(7050):518–24.

490

S.A. Eccles

195. Hamada K, Monnai M, Kawai K, et al. Liver metastasis models of colon cancer for evaluation of drug efficacy using NOD/Shi-scid IL2Rgammanull (NOG) mice. Int J Oncol. 2008;32(1):153–9. 196. Sasaki T, Nakamura T, Rebhun RB, et al. Modification of the primary tumor microenvironment by transforming growth factor alpha-epidermal growth factor receptor signaling promotes metastasis in an orthotopic colon cancer model. Am J Pathol. 2008;173(1):205–16. 197. Ponz-Sarvise M, Rodriguez J, Viudez A, et al. Epidermal growth factor receptor inhibitors in colorectal cancer treatment: what’s new? World J Gastroenterol. 2007;13(44):5877–87. 198. Nam DH, Jeon HM, Kim S, et al. Activation of notch signaling in a xenograft model of brain metastasis. Clin Cancer Res. 2008;14(13):4059–66. 199. Barth RF, Kaur B. Rat brain tumor models in experimental neuro-oncology: the C6, 9L, T9, RG2, F98, BT4C, RT-2 and CNS-1 gliomas. J Neurooncol. 2009;94:299–312. 200. Xie Q, Thompson R, Hardy K, et al. A highly invasive human glioblastoma pre-clinical model for testing therapeutics. J Transl Med. 2008;6:77. 201. Hu J, Yuan X, Ko MK, et al. Calcium-activated potassium channels mediated blood-brain tumor barrier opening in a rat metastatic brain tumor model. Mol Cancer. 2007;6:22. 202. van Golen CM, Schwab TS, Kim B, et al. Insulin-like growth factor-I receptor expression regulates neuroblastoma metastasis to bone. Cancer Res. 2006;66(13):6570–8. 203. Mao-Ying QL, Zhao J, Dong ZQ, et al. A rat model of bone cancer pain induced by intra-tibia inoculation of Walker 256 mammary gland carcinoma cells. Biochem Biophys Res Commun. 2006;345(4):1292–8. 204. Goya M, Miyamoto S, Nagai K, et al. Growth inhibition of human prostate cancer cells in human adult bone implanted into nonobese diabetic/severe combined immunodeficient mice by a ligand-specific antibody to human insulin-like growth factors. Cancer Res. 2004;64(17):6252–8. 205. Harms JF, Welch DR. MDA-MB-435 human breast carcinoma metastasis to bone. Clin Exp Metastasis. 2003;20(4):327–34. 206. Singh AS, Figg WD. In vivo models of prostate cancer metastasis to bone. J Urol. 2005;174(3):820–6. 207. Butterworth KT, McCarthy HO, Devlin A, et al. Hypoxia selects for androgen independent LNCaP cells with a more malignant geno- and phenotype. Int J Cancer. 2008;123(4):760–8. 208. Zhau HE, Li CL, Chung LW. Establishment of human prostate carcinoma skeletal metastasis models. Cancer. 2000;88(12 Suppl):2995–3001. 209. Idris AI, Libouban H, Nyangoga H, Landao-Bassonga E, Chappard D, Ralston SH. Pharmacologic inhibitors of I{kappa}B kinase suppress growth and migration of mammary carcinosarcoma cells in vitro and prevent osteolytic bone metastasis in vivo. Mol Cancer Ther. 2009;8:2339–47. 210. Russell MR, Jamieson WL, Dolloff NG, Fatatis A. The alpha-receptor for platelet-derived growth factor as a target for antibody-mediated inhibition of skeletal metastases from prostate cancer cells. Oncogene. 2009;28(3):412–21. 211. Zhang XH, Wang Q, Gerald W, et al. Latent bone metastasis in breast cancer tied to Src-dependent survival signals. Cancer Cell. 2009;16(1):67–78. 212. Choi YJ, Nam SJ, Son MJ, et al. Erlotinib prevents pulmonary metastasis in curatively resected breast carcinoma using a mouse model. Oncol Rep. 2006;16(1):119–22. 213. Mi Z, Guo H, Russell MB, Liu Y, Sullenger BA, Kuo PC. RNA aptamer blockade of osteopontin inhibits growth and metastasis of MDA-MB231 breast cancer cells. Mol Ther. 2009;17(1):153–61. 214. Kizaka-Kondoh S, Itasaka S, Zeng L, et al. Selective killing of hypoxia-inducible factor-1active cells improves survival in a mouse model of invasive and metastatic pancreatic cancer. Clin Cancer Res. 2009;15(10):3433–41. 215. Tomioka D, Maehara N, Kuba K, et al. Inhibition of growth, invasion, and metastasis of human pancreatic carcinoma cells by NK4 in an orthotopic mouse model. Cancer Res. 2001;61(20):7518–24. 216. Eccles S, Paon L, Sleeman J. Lymphatic metastasis in breast cancer: importance and new insights into cellular and molecular mechanisms. Clin Exp Metastasis. 2007;24(8):619–36.


491

217. Sleeman JP. The lymph node as a bridgehead in the metastatic dissemination of tumors. Recent Results Cancer Res. 2000;157:55–81. 218. Park SI, Zhang J, Phillips KA, et al. Targeting SRC family kinases inhibits growth and lymph node metastases of prostate cancer in an orthotopic nude mouse model. Cancer Res. 2008;68(9):3323–33. 219. Li Y, Song E, Abbas Rizvi SM, et al. Inhibition of micrometastatic prostate cancer cell spread in animal models by 213Bilabeled multiple targeted alpha radioimmunoconjugates. Clin Cancer Res. 2009;15(3):865–75. 220. Murphy EA, Majeti BK, Barnes LA, et al. Nanoparticle-mediated drug delivery to tumor vasculature suppresses metastasis. Proc Natl Acad Sci USA. 2008;105(27):9343–8. 221. Nakamura ES, Koizumi K, Kobayashi M, Saiki I. Inhibition of lymphangiogenesis-related properties of murine lymphatic endothelial cells and lymph node metastasis of lung cancer by the matrix metalloproteinase inhibitor MMI270. Cancer Sci. 2004;95(1):25–31. 222. Brakenhielm E, Burton JB, Johnson M, et al. Modulating metastasis by a lymphangiogenic switch in prostate cancer. Int J Cancer. 2007;121(10):2153–61. 223. Burton JB, Priceman SJ, Sung JL, et al. Suppression of prostate cancer nodal and systemic metastasis by blockade of the lymphangiogenic axis. Cancer Res. 2008;68(19):7828–37. 224. Hori K, Li HC, Saito S, Sato Y. Increased growth and incidence of lymph node metastases due to the angiogenesis inhibitor AGM-1470. Br J Cancer. 1997;75(12):1730–4. 225. Davie SA, Maglione JE, Manner CK, et al. Effects of FVB/NJ and C57Bl/6J strain backgrounds on mammary tumor phenotype in inducible nitric oxide synthase deficient mice. Transgenic Res. 2007;16(2):193–201. 226. Siegel PM, Ryan ED, Cardiff RD, Muller WJ. Elevated expression of activated forms of Neu/ ErbB-2 and ErbB-3 are involved in the induction of mammary tumors in transgenic mice: implications for human breast cancer. EMBO J. 1999;18(8):2149–64. 227. Dimri G, Band H, Band V. Mammary epithelial cell transformation: insights from cell culture and mouse models. Breast Cancer Res. 2005;7(4):171–9. 228. Fluck MM, Haslam SZ. Mammary tumors induced by polyomavirus. Breast Cancer Res Treat. 1996;39(1):45–56. 229. Ursini-Siegel J, Schade B, Cardiff RD, Muller WJ. Insights from transgenic mouse models of ERBB2-induced breast cancer. Nat Rev Cancer. 2007;7(5):389–97. 230. Brantley-Sieders DM, Zhuang G, Hicks D, et al. The receptor tyrosine kinase EphA2 promotes mammary adenocarcinoma tumorigenesis and metastatic progression in mice by amplifying ErbB2 signaling. J Clin Invest. 2008;118(1):64–78. 231. Muraoka-Cook RS, Kurokawa H, Koh Y, et al. Conditional overexpression of active transforming growth factor beta1 in vivo accelerates metastases of transgenic mammary tumors. Cancer Res. 2004;64(24):9002–11. 232. Shepard CR, Kassis J, Whaley DL, Kim HG, Wells A. PLC gamma contributes to metastasis of in situ-occurring mammary and prostate tumors. Oncogene. 2007;26(21):3020–6. 233. Juncker-Jensen A, Romer J, Pennington CJ, Lund LR, Almholt K. Spontaneous metastasis in matrix metalloproteinase 3-deficient mice. Mol Carcinog. 2009;48(7):618–25. 234. Gingrich JR, Barrios RJ, Morton RA, et al. Metastatic prostate cancer in a transgenic mouse. Cancer Res. 1996;56(18):4096–102. 235. Willis RA, Wei C, Turner MJ, et al. A transgenic strategy for analyzing the regulatory regions of the human prostate-specific antigen gene: potential applications for the treatment of prostate cancer (Review). Int J Mol Med. 1998;1(2):379–86. 236. Khanna C, Hunter K. Modeling metastasis in vivo. Carcinogenesis. 2005;26(3):513–23. 237. McCart AE, Vickaryous NK, Silver A. Apc mice: models, modifiers and mutants. Pathol Res Pract. 2008;204(7):479–90. 238. Varticovski L, Hollingshead MG, Robles AI, et al. Accelerated preclinical testing using transplanted tumors from genetically engineered mouse breast cancer models. Clin Cancer Res. 2007;13(7):2168–77. 239. Hurwitz AA, Foster BA, Allison JP, Greenberg NM, Kwon ED. The TRAMP mouse as a model for prostate cancer. Curr Protoc Immunol. 2001;Chapter 20:Unit 20 5.

492

S.A. Eccles

240. Schade B, Rao T, Dourdin N, et al. PTEN deficiency in a luminal ErbB-2 mouse model results in dramatic acceleration of mammary tumorigenesis and metastasis. J Biol Chem. 2009;284(28):19018–26. 241. Guigon CJ, Zhao L, Willingham MC, Cheng SY. PTEN deficiency accelerates tumour progression in a mouse model of thyroid cancer. Oncogene. 2009;28(4):509–17. 242. Ahmad I, Sansom OJ, Leung HY. Advances in mouse models of prostate cancer. Expert Rev Mol Med. 2008;10:e16. 243. Moody SE, Sarkisian CJ, Hahn KT, et al. Conditional activation of Neu in the mammary epithelium of transgenic mice results in reversible pulmonary metastasis. Cancer Cell. 2002;2(6):451–61. 244. Olive KP, Tuveson DA. The use of targeted mouse models for preclinical testing of novel cancer therapeutics. Clin Cancer Res. 2006;12(18):5277–87. 245. Moral M, Segrelles C, Lara MF, et al. Akt activation synergizes with Trp53 loss in oral epithelium to produce a novel mouse model for head and neck squamous cell carcinoma. Cancer Res. 2009;69(3):1099–108. 246. Ferreira M, Fujiwara H, Morita K, Watt FM. An activating beta1 integrin mutation increases the conversion of benign to malignant skin tumors. Cancer Res. 2009;69(4):1334–42. 247. Becher OJ, Holland EC, Sausville EA, Burger AM. Genetically engineered models have advantages over xenografts for preclinical studies 10.1158/0008-5472.CAN-05-3827. Cancer Res. 2006;66(7):3355–9. 248. Fomchenko EI, Holland EC. Mouse models of brain tumors and their applications in preclinical trials. Clin Cancer Res. 2006;12(18):5288–97. 249. Sausville EA, Burger AM, Becher OJ, Holland EC. Contributions of human tumor xenografts to anticancer drug development 10.1158/0008-5472.CAN-05-3627. Cancer Res. 2006;66(7):3351–4. 250. Sharpless NE, DePinho RA. The mighty mouse: genetically engineered mouse models in cancer drug development. Nat Rev Drug Discov. 2006;5(9):741–54. 251. Talmadge JE, Singh RK, Fidler IJ, Raz A. Murine models to evaluate novel and conventional therapeutic strategies for cancer. Am J Pathol. 2007;170(3):793–804. 252. van Weerden WM, Bangma C, de Wit R. Human xenograft models as useful tools to assess the potential of novel therapeutics in prostate cancer. Br J Cancer. 2009;100(1):13–8. 253. Smith MC, Luker KE, Garbow JR, et al. CXCR4 regulates growth of both primary and metastatic breast cancer. Cancer Res. 2004;64(23):8604–12. 254. Steeg PS, Theodorescu D. Metastasis: a therapeutic target for cancer. Nat Clin Pract Oncol. 2008;5(4):206–19. 255. Nam JS, Terabe M, Mamura M, et al. An anti-transforming growth factor beta antibody suppresses metastasis via cooperative effects on multiple cell compartments. Cancer Res. 2008;68(10):3835–43. 256. Chakrabarti R, Subramaniam V, Abdalla S, Jothy S, Prud’homme GJ. Tranilast inhibits the growth and metastasis of mammary carcinoma. Anticancer Drugs. 2009;20(5):334–45. 257. Rausch MP, Hahn T, Ramanathapuram L, et al. An orally active small molecule TGF-beta receptor I antagonist inhibits the growth of metastatic murine breast cancer. Anticancer Res. 2009;29(6):2099–109. 258. Takahashi M, Furihata M, Akimitsu N, et al. A highly bone marrow metastatic murine breast cancer model established through in vivo selection exhibits enhanced anchorage- independent growth and cell migration mediated by ICAM-1. Clin Exp Metastasis. 2008;25(5):517–29. 259. Sidani M, Wyckoff J, Xue C, Segall JE, Condeelis J. Probing the microenvironment of mammary tumors using multiphoton microscopy. J Mammary Gland Biol Neoplasia. 2006;11(2):151–63. 260. Blouin S, Basle MF, Chappard D. Rat models of bone metastases. Clin Exp Metastasis. 2005;22(8):605–14. 261. Eccles SA, Box GM, Court WJ, Bone EA, Thomas W, Brown PD. Control of lymphatic and hematogenous metastasis of a rat mammary carcinoma by the matrix metalloproteinase inhibitor batimastat (BB-94). Cancer Res. 1996;56(12):2815–22.


493

262. Becker JC, Pancook JD, Gillies SD, Mendelsohn J, Reisfeld RA. Eradication of human hepatic and pulmonary melanoma metastases in SCID mice by antibody-interleukin 2 fusion proteins. Proc Natl Acad Sci USA. 1996;93(7):2702–7. 263. Kasaoka T, Nishiyama H, Okada M, Nakajima M. Matrix metalloproteinase inhibitor, MMI270 (CGS27023A) inhibited hematogenic metastasis of B16 melanoma cells in both experimental and spontaneous metastasis models. Clin Exp Metastasis. 2008;25(7):827–34. 264. Faham A, Bennett D, Altin JG. Liposomal Ag engrafted with peptides of sequence derived from HMGB1 induce potent Ag-specific and anti-tumour immunity. Vaccine. 2009;27:5846–54. 265. Patel D, Bassi R, Hooper AT, et al. Enhanced suppression of melanoma tumor growth and metastasis by combined therapy with anti-VEGF receptor and anti-TYRP-1/gp75 monoclonal antibodies. Anticancer Res. 2008;28(5A):2679–86. 266. Yang L, Huang Y, Porta R, et al. Host and direct antitumor effects and profound reduction in tumor metastasis with selective EP4 receptor antagonism. Cancer Res. 2006;66(19):9665–72. 267. Park W, Oh YT, Han JH, Pyo H. Antitumor enhancement of celecoxib, a selective Cyclooxygenase-2 inhibitor, in a Lewis lung carcinoma expressing Cyclooxygenase-2. J Exp Clin Cancer Res. 2008;27:66. 268. VanSaun MN, Lee IK, Washington MK, Matrisian L, Gorden DL. High fat diet induced hepatic steatosis establishes a permissive microenvironment for colorectal metastases and promotes primary dysplasia in a murine model. Am J Pathol. 2009;175(1):355–64. 269. de Heer P, Sandel MH, Guertens G, et al. Celecoxib inhibits growth of tumors in a syngeneic rat liver metastases model for colorectal cancer. Cancer Chemother Pharmacol. 2008;62(5):811–9. 270. Kim SY, Lee CH, Midura BV, et al. Inhibition of the CXCR4/CXCL12 chemokine pathway reduces the development of murine pulmonary metastases. Clin Exp Metastasis. 2008;25(3):201–11. 271. Lodillinsky C, Rodriguez V, Vauthay L, Sandes E, Casabe A, Eijan AM. Novel invasive orthotopic bladder cancer model with high cathepsin B activity resembling human bladder cancer. J Urol. 2009;182(2):749–55. 272. Verheul HM, Hammers H, van Erp K, et al. Vascular endothelial growth factor trap blocks tumor growth, metastasis formation, and vascular leakage in an orthotopic murine renal cell cancer model. Clin Cancer Res. 2007;13(14):4201–8. 273. Hautmann S, Huland E, Wullbrand A, Friedrich M, Huland H. Treatment of metastatic hormone-refractory prostate adenocarcinoma (MatLyLu) in Copenhagen rats with microosmotic interleukin-2 pumps. Anticancer Res. 2000;20(6B):4495–8. 274. Winter SF, Cooper AB, Greenberg NM. Models of metastatic prostate cancer: a transgenic perspective. Prostate Cancer Prostatic Dis. 2003;6(3):204–11. 275. Weeraratna AT, Dalrymple SL, Lamb JC, et al. Pan-trk inhibition decreases metastasis and enhances host survival in experimental models as a result of its selective induction of apoptosis of prostate cancer cells. Clin Cancer Res. 2001;7(8):2237–45. 276. Blouw B, Song H, Tihan T, et al. The hypoxic response of tumors is dependent on their microenvironment. Cancer Cell. 2003;4(2):133–46. 277. Yin J, Pollock C, Tracy K, et al. Activation of the RalGEF/Ral pathway promotes prostate cancer metastasis to bone. Mol Cell Biol. 2007;27(21):7538–50. 278. Havens AM, Pedersen EA, Shiozawa Y, et al. An in vivo mouse model for human prostate cancer metastasis. Neoplasia. 2008;10(4):371–80. 279. Rucci N, Recchia I, Angelucci A, et al. Inhibition of protein kinase c-Src reduces the incidence of breast cancer metastases and increases survival in mice: implications for therapy. J Pharmacol Exp Ther. 2006;318(1):161–72. 280. Yang E, Boire A, Agarwal A, et al. Blockade of PAR1 signaling with cell-penetrating pepducins inhibits Akt survival pathways in breast cancer cells and suppresses tumor survival and metastasis. Cancer Res. 2009;69(15):6223–31. 281. Ottewell PD, Coleman RE, Holen I. From genetic abnormality to metastases: murine models of breast cancer and their use in the development of anticancer therapies. Breast Cancer Res Treat. 2006;96(2):101–13.

494

S.A. Eccles

282. Fantozzi A, Christofori G. Mouse models of breast cancer metastasis. Breast Cancer Res. 2006;8(4):212. 283. Zhang C, Yan Z, Arango ME, Painter CL, Anderes K. Advancing bioluminescence imaging technology for the evaluation of anticancer agents in the MDA-MB-435-HAL-Luc mammary fat pad and subrenal capsule tumor models. Clin Cancer Res. 2009;15(1):238–46. 284. Michieli P, Mazzone M, Basilico C, et al. Targeting the tumor and its microenvironment by a dual-function decoy Met receptor. Cancer Cell. 2004;6(1):61–73. 285. Vantyghem SA, Allan AL, Postenka CO, et al. A new model for lymphatic metastasis: development of a variant of the MDA-MB-468 human breast cancer cell line that aggressively metastasizes to lymph nodes. Clin Exp Metastasis. 2005;22(4):351–61. 286. Daoud SS, Leathers CW, Hurst JM. p53 expression, growth, and spontaneous metastasis of the human GI 101 breast carcinoma in athymic nude mice. J Exp Ther Oncol. 2002;2(2):121–7. 287. Bouvet M, Tsuji K, Yang M, Jiang P, Moossa AR, Hoffman RM. In vivo color-coded imaging of the interaction of colon cancer cells and splenocytes in the formation of liver metastases. Cancer Res. 2006;66(23):11293–7. 288. Herynk MH, Stoeltzing O, Reinmuth N, et al. Down-regulation of c-Met inhibits growth in the liver of human colorectal carcinoma cells. Cancer Res. 2003;63(11):2990–6. 289. Fidarova EF, El-Emir E, Boxer GM, et al. Microdistribution of targeted, fluorescently labeled anti-carcinoembryonic antigen antibody in metastatic colorectal cancer: implications for radioimmunotherapy. Clin Cancer Res. 2008;14(9):2639–46. 290. Blumenthal RD, Osorio L, Hayes MK, Horak ID, Hansen HJ, Goldenberg DM. Carcinoembryonic antigen antibody inhibits lung metastasis and augments chemotherapy in a human colonic carcinoma xenograft. Cancer Immunol Immunother. 2005;54(4):315–27. 291. Yano S, Muguruma H, Matsumori Y, et al. Antitumor vascular strategy for controlling experimental metastatic spread of human small-cell lung cancer cells with ZD6474 in natural killer cell-depleted severe combined immunodeficient mice. Clin Cancer Res. 2005;11(24 Pt 1):8789–98. 292. Hu L, Zhang J, Zhu H, Min J, Feng Y, Zhang H. Biological characteristics of a specific brain metastatic cell line derived from human lung adenocarcinoma. Med Oncol. 2010;27:708–14. 293. Zhang L, Smith KM, Chong AL, et al. In vivo antitumor and antimetastatic activity of sunitinib in preclinical neuroblastoma mouse model. Neoplasia. 2009;11(5):426–35. 294. Hingorani P, Zhang W, Gorlick R, Kolb EA. Inhibition of Src phosphorylation alters metastatic potential of osteosarcoma in vitro but not in vivo. Clin Cancer Res. 2009;15(10):3416–22. 295. Chitnis MM, Yuen JS, Protheroe AS, Pollak M, Macaulay VM. The type 1 insulin-like growth factor receptor pathway. Clin Cancer Res. 2008;14(20):6364–70. 296. Chaffer CL, Brennan JP, Slavin JL, Blick T, Thompson EW, Williams ED. Mesenchymal-toepithelial transition facilitates bladder cancer metastasis: role of fibroblast growth factor receptor-2. Cancer Res. 2006;66(23):11271–8. 297. Sun R, Zhang JG, Guo CB. Establishment of cervical lymph node metastasis model of squamous cell carcinoma in the oral cavity in mice. Chin Med J (Engl) 2008;121(19):1891–5. 298. Pan J, Mestas J, Burdick MD, et al. Stromal derived factor-1 (SDF-1/CXCL12) and CXCR4 in renal cell carcinoma metastasis. Mol Cancer. 2006;5:56. 299. Cruz-Munoz W, Man S, Xu P, Kerbel RS. Development of a preclinical model of spontaneous human melanoma central nervous system metastasis. Cancer Res. 2008;68(12):4500–5. 300. Thies A, Mauer S, Fodstad O, Schumacher U. Clinically proven markers of metastasis predict metastatic spread of human melanoma cells engrafted in scid mice. Br J Cancer. 2007;96(4):609–16. 301. Thies A, Peldschus K, Ittrich H, et al. Magnetic resonance imaging of melanoma metastases in a clinical relevant human melanoma xenograft scid mouse model. Cancer Lett. 2009;274(2):194–200. 302. Nakai M, Mundy GR, Williams PJ, Boyce B, Yoneda T. A synthetic antagonist to laminin inhibits the formation of osteolytic metastases by human melanoma cells in nude mice. Cancer Res. 1992;52(19):5395–9.


495

303. Huang J, Chen K, Gong W, Dunlop NM, Wang JM. G-protein coupled chemoattractant receptors and cancer. Front Biosci. 2008;13:3352–63. 304. Conrad C, Ischenko I, Kohl G, et al. Antiangiogenic and antitumor activity of a novel vascular endothelial growth factor receptor-2 tyrosine kinase inhibitor ZD6474 in a metastatic human pancreatic tumor model. Anticancer Drugs. 2007;18(5):569–79. 305. Stelter L, Amthauer H, Rexin A, et al. An orthotopic model of pancreatic somatostatin receptor (SSTR)-positive tumors allows bimodal imaging studies using 3T MRI and animal PET-based molecular imaging of SSTR expression. Neuroendocrinology. 2008;87(4):233–42. 306. Schwock J, Dhani N, Cao MP, et al. Targeting focal adhesion kinase with dominant-negative FRNK or Hsp90 inhibitor 17-DMAG suppresses tumor growth and metastasis of SiHa cervical xenografts. Cancer Res. 2009;69(11):4750–9. 307. Lee EM, Bachmann PS, Lock RB. Xenograft models for the preclinical evaluation of new therapies in acute leukemia. Leuk Lymphoma. 2007;48(4):659–68. 308. Thomas J, Liu T, Cotter MA, et al. Melanocyte expression of survivin promotes development and metastasis of UV-induced melanoma in HGF-transgenic mice. Cancer Res. 2007;67(11):5172–8. 309. Otsuka T, Takayama H, Sharp R, et al. c-Met autocrine activation induces development of malignant melanoma and acquisition of the metastatic phenotype. Cancer Res. 1998;58(22):5157–67. 310. Lopez T, Hanahan D. Elevated levels of IGF-1 receptor convey invasive and metastatic capability in a mouse model of pancreatic islet tumorigenesis. Cancer Cell. 2002;1(4):339–53. 311. Morton JP, Klimstra DS, Mongeau ME, Lewis BC. Trp53 deletion stimulates the formation of metastatic pancreatic tumors. Am J Pathol. 2008;172(4):1081–7. 312. Zinser GM, Leonis MA, Toney K, et al. Mammary-specific Ron receptor overexpression induces highly metastatic mammary tumors associated with beta-catenin activation. Cancer Res. 2006;66(24):11967–74. 313. Finkle D, Quan ZR, Asghari V, et al. HER2-targeted therapy reduces incidence and progression of midlife mammary tumors in female murine mammary tumor virus huHER2-transgenic mice. Clin Cancer Res. 2004;10(7):2499–511. 314. You L, Kim J, He B, Xu Z, McCormick F, Jablons DM. Wnt-1 signal as a potential cancer therapeutic target. Drug News Perspect. 2006;19(1):27–31. 315. Berman SD, Calo E, Landman AS, et al. Metastatic osteosarcoma induced by inactivation of Rb and p53 in the osteoblast lineage. Proc Natl Acad Sci USA. 2008;105(33):11851–6. 316. Weber K, Doucet M, Kominsky S. Renal cell carcinoma bone metastasis – elucidating the molecular targets. Cancer Metastasis Rev. 2007;26(3–4):691–704. 317. Singh RP, Raina K, Sharma G, Agarwal R. Silibinin inhibits established prostate tumor growth, progression, invasion, and metastasis and suppresses tumor angiogenesis and epithelial-mesenchymal transition in transgenic adenocarcinoma of the mouse prostate model mice. Clin Cancer Res. 2008;14(23):7773–80. 318. Raina K, Rajamanickam S, Singh RP, Deep G, Chittezhath M, Agarwal R. Stage-specific inhibitory effects and associated mechanisms of silibinin on tumor progression and metastasis in transgenic adenocarcinoma of the mouse prostate model. Cancer Res. 2008;68(16):6822–30. 319. Puzio-Kuter AM, Castillo-Martin M, Kinkade CW, et al. Inactivation of p53 and Pten promotes invasive bladder cancer. Genes Dev. 2009;23(6):675–80. 320. MacDonald IC, Chambers AF. Breast cancer metastasis progression as revealed by intravital videomicroscopy. Expert Rev Anticancer Ther. 2006;6(9):1271–9. 321. Edinger M, Cao YA, Verneris MR, Bachmann MH, Contag CH, Negrin RS. Revealing lymphoma growth and the efficacy of immune cell therapies using in vivo bioluminescence imaging. Blood. 2003;101(2):640–8. 322. Palmowski M, Huppert J, Hauff P, et al. Vessel fractions in tumor xenografts depicted by flow- or contrast-sensitive three-dimensional high-frequency Doppler ultrasound respond differently to antiangiogenic treatment. Cancer Res. 2008;68(17):7042–9.

Part VIII

Normal Tissue Response Models

Chapter 20

Animal Models of Toxicities Caused by Anti-Neoplastic Therapy Stephen T. Sonis, Gregory Lyng, and Kimberly Pouliot

Abstract Radiation and chemotherapy induce a wide range of acute and chronic toxicities. Not only are these associated with poor health outcomes but they also limit patients’ ability to tolerate treatment and incur significant increases in resource use and cost. Universally, they impair patients’ quality of life (QoL). In addition to hematological complications such as anemia, thrombocytopenia, and neutropenia, cancer patients are also at risk for a wide range of non-hematological taxicities. These may occure during or soon after cancer treatment (acute toxicities), or they may not develop until well after the completion of treatment (# 100 days, late toxicities) and become chronic and linger for years after the patient’s disease is controlled. The overall incidence of some form of treatment toxicity is almost 100%. Toxicities include those that are tissuespecific such as mucosal injury of some or all of the parts of the gastrointestinal tract (mucositis), cutaneous damage (dermatitis), salivary gland dysfunction, and venous thrombosis. Alteratively, patients may develop more systemic forms of toxicity that result in conditions such as fatigue, depression, cognitive impairment, and cachexia. Keywords Toxicities • Mucositis • Dermatitis • Ostconecrosis • Fatigue • Animal models • Radiction • Chemotherapy

20.1 Introduction Radiation and chemotherapy induce a wide range of acute and chronic toxicities. Not only are these associated with poor health outcomes but they also limit patients’ ability to tolerate treatment and incur significant increases in resource use and cost. Universally, they impair patients’ quality of life (QoL). S.T. Sonis (*) Harvard-Farber Cancer Center, Boston, MA, USA and Biomodels, Watertown, MA, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_20, © Springer Science+Business Media, LLC 2011

499

500

S.T. Sonis et al.

In addition to hematological complications such as anemia, thrombocytopenia, and neutropenia, cancer patients are also at risk for a wide range of non-hematological toxicities. These may occur during or soon after cancer treatment (acute toxicities), or they may not develop until well after the completion of treatment (³100 days, late toxicities) and become chronic and linger for years after the patient’s disease is controlled. The overall incidence of some form of treatment toxicity is almost 100%. Toxicities include those that are tissue-specific such as mucosal injury of some or all of the parts of the gastrointestinal tract (mucositis), cutaneous damage (dermatitis), salivary gland dysfunction, and venous thrombosis. Alternatively, patients may develop more systemic forms of toxicity that result in conditions such as fatigue, depression, cognitive impairment, and cachexia. The clinical ramifications of regimen-related toxicities are diverse and significant. Some, such as oral mucositis, are associated with pain of such intensity as to require opioid analgesics and inability to eat [1]. Others, such as enteritis, cause diarrhea and put patients at risk for bacteremia or sepsis. Almost all negatively impact patients’ QoL and ability to tolerate treatment. Indeed, toxicities often necessitate less than optimal dosing regimens or early termination of treatment. Toxicities are also associated with higher mortality risks. Almost all toxicities adversely affect patients’ QoL and have a major impact on increased healthcare costs. Often patients with acute toxicities require unplanned office and emergency room visits, or hospital admissions for fluid support, pain control, or infection management. The economic impact of this extra use of resources is substantial. For example, the incremental cost of oral mucositis among patients being treated for cancers of the lung or head and neck is in excess of $17,000 [2]. Toxicities that linger in patients who have completed treatment such as fatigue and depression impair the ability to work and otherwise function normally. As a result of their impact on patients’ symptoms and QoL and especially because they indirectly limit an individuals’ treatment tolerance, toxicities are the subject of intense study. The discovery that many toxicities seem to cluster suggests shared pathoetiologies. Investigations defining the biology of toxicities have stimulated the quest for appropriate and effective pharmacologic and biologic interventions. The use of animal models has proliferated and a number of them now exist which mimic toxicities seen in humans and serve as predictive platforms for drug development. This chapter will focus on models of toxicities of epithelial injury and bisphosphonate osteonecrosis.

20.2 Models of Oral Mucositis Induced by Anti-Neoplastic Drugs and Radiation 20.2.1 Overview of the Condition Oral mucositis is one of the best studied acute toxicities of non-surgical cancer therapy. Since it affects about 40% of all patients being treated for non-cutaneous cancers, the need for a successful intervention remains a high priority [1]. At present only a single agent, palifermin, has been approved for this indication in the US and

20 Animal Models of Toxicities Caused by Anti-Neoplastic Therapy

501

palifermin’s applicability is limited to the small cohort of patients receiving stomatotoxic conditioning regimens in preparation for stem cell transplants to treat hematological malignancies (4% of patients at risk for the condition) [3]. Clinically, mucositis occurs with great frequency among patients being treated with radiation therapy, with or without concomitant chemotherapy, for cancers of the head and neck. Virtually 100% of patients with cancers of the mouth or oropharynx will develop mucositis. The incidence is slightly less among individuals being treated for hypopharyngeal or laryngeal tumors. Many of the conditioning regimens for stem cell transplant are stomatotoxic, especially those in which total body irradiation is a component. Lastly, mucositis impacts patients being treated with cycled therapy for the most common solid tumors (breast, colon, rectum, lung). In this group, the overall risk of mucositis in the first cycle of treatment is relatively low (about 15–20%), but if no effort is made to reduce chemotherapy dosing for subsequent cycles, the risk of mucositis increases dramatically, in many cases to more than 60%. The impact of mucositis is profound. Patients suffer marked pain, often requiring opioids, have to modify their diets, lose weight, have increased risk of local and systemic infection, require fluid support, and use consultation and emergency services more than patients who do not develop the condition [4]. Clinically mucositis develops in predictable stages. Initially, the mucosa is thinned and hyperemic. Although the tissue is intact, patients note some discomfort, often described as being analogous to a bad food burn. Symptoms can be reasonably controlled at this stage with a combination of topical analgesics and systemic agents such as acetaminophen of NSAIDs. The development of ulceration occurs next. This is the phase that is most symptomatic. Pain increases dramatically, often requiring morphine or fentynal. Eating a normal diet becomes impossible. Patients are limited to very soft or liquid diets and some may not be able to eat anything. Consequently, it is not unusual for nutrition to have to be provided by feeding tubes (gastrostomy tubes) or total parenteral tuition. In the majority of cases ulceration spontaneously resolves.

20.2.2 The Biology of Mucositis Historically, mucositis was considered to be solely the consequence of nonspecific direct clonogenic damage to basal cells of the oral mucosa. The paradigm held that as these “mother” cells were injured, normal renewal did not take place, the tissue became atrophic and ultimately ulcerated. Results of studies performed over the past decade paint a more complex picture of the biology of regimen-related mucosal injury [5]. A summary of our current understanding of the pathobiology indicates five phases [6]: 1. Initiation: In this phase, radiation or chemotherapy may directly injure DNA causing clonogenic cell death and, more significantly, cause the generation of reactive oxygen species. 2. Primary tissue response: Radiation, chemotherapy, and ROS trigger the activation of a number of transcription factors such as NF-kB, Wnt, and p53. At least 14 canonical pathways play a role in initiating mucosal injury.

502

S.T. Sonis et al.

3. Signal amplification: Many of the molecules induced during the primary response phase have the ability to positively or negatively feedback on activated pathways causing amplification of injury. 4. Ulceration: Ulceration is the most relevant clinical endpoint of the process and is responsible for virtually all of the outcomes with significant clinical meaningfulness. Bacterial colonization occurs during this phase with bacterial products percolating into the submucosa to stimulate macrophages to produce pro-inflammatory cytokines. 5. Healing: Signaling from the extracellular matrix guides epithelial healing, including differentiation, migration, and proliferation. This process is probably initiated by the generation of reactive oxygen species by exposure to stomatotoxic stimuli. Activation and expression of proinflammatory cytokines, particularly tumor necrosis alpha (TNF-a) and interleukin b (IL-b) and endothelial damage characterize the inflammatory/vascular phase.

20.2.3 Objectives of Animal Models of Mucositis There are four objectives for an effective animal model of mucositis to provide clinical meaningfulness: 1. The manifestations of mucositis should mimic the condition as it occurs in humans in its course, appearance, resolution, and dose response to stomatotoxic therapy. Its presentation should be robust enough as to not require microscopic or surrogate endpoints. 2. The pathogenesis of mucositis in the model should replicate, at the molecular, cellular, and tissue levels, the events that occur in humans. 3. Concurrent toxicities, especially those in which myelosuppression is an element, should occur in a measurable way. 4. The oral environment, especially the microscopic flora, should resemble that of humans and should respond to stomatotoxic therapy in a way that is the same as humans.

20.2.4 Current Models Three species have been and/or are used for studies of oral mucositis: mice, rats, and hamsters. Murine models have been used to study both radiation- and chemotherapyinduced mucositis. In general, the endpoints used to assess mucositis have relied heavily on histological outcomes since clinical changes tend to be subtle and focus on erythema, rather than ulceration as a primary endpoint. Rats have also been used to assess radiation and chemotherapy-induced mucositis, and both 5-FU and methatrexate have been used to induce mucosal injury, often accompanied by superficial irritation [7, 8]. Lesions in these models tend to be localized.


503

A number of studies focusing on the epithelial biology of oral radiation have been performed using murine lip, snout, or tongue models. Xu et al. described the effects of single and fractionated radiation schedules on the lip mucosa of mice. They found that acute reactions of the lip mucosa, i.e. focal desquamation, could be reliably scored [9]. Alternatively, Kilic et al. [10] have used a model in which the ventral surface of the tongues of mice are radiated by guiding the tongues of anesthetized animals through a 3-mm hole in an aluminum block. The dorsal tongue was then fixed with tape and an aluminum plate with a 3 × 3 mm2 window was placed over the target area on the ventral tongue. Importantly, strain-dependent variability in murine vulnerability to radiation injury has been reported. C3H/Neu mice have been used successfully. These models have been useful to define responses to various radiation regimens, including cell repopulation studies, yet the limited anatomic area available for evaluation, challenges associated with the use of topical formulations, and the subtlety of clinical changes have limited their applicability in interventional studies. While the clinical signal noted in murine models may be subtle, the ready availability of syngeneic animals, knock-outs, immune reagents, and gene chips makes the mouse a good choice for answering specific questions associated with the pathogenesis of mucosal injury. Rats have been the species of choice for studies of gastrointestinal mucositis, especially those induced by chemotherapy. Until recently, histological endpoints were mandated. However, we have recently applied endoscopy to assess mucosal injury of the lower GI tract (see section below). The rat has also been effective in studying radiation-induced proctitis. Given the limitations of the murine and rat models for oral mucositis, the hamster was evaluated as a potential species. The selection of the hamster was based on five major factors: 1. The hamster cheek pouch consists of a renewing squamous epithelium, which is similar to humans in many ways and has been studied for a number of other conditions, especially chemically induced carcinogenesis. 2. The cheek pouch mucosa provides a large mucosal surface for study that is easily accessible for examination and an anatomical site to which potential topical therapeutics can be easily applied. The pouch can be easily isolated with a lead shield from the rest of the animal’s head thereby permitting targeted radiation with no systemic toxicity. 3 . The hamster’s oral bacterial flora is similar to that of the humans in that Gram-positive bacteria are the dominant species. The cheek pouch has also been described as useful for the study of fungal infections, something that commonly occurs in humans undergoing cancer therapy. Shifts in hamster oral flora following cancer therapy mirror those described in humans – both quantitatively and qualitatively [11]. 4. The hamster is sensitive to chemotherapeutic agents that elicit toxicity in humans. The hamsters’ marrow response to these agents is similar to humans. Neutropenia develops in essentially the same time course as in humans. Thus the model is informative in studying the relationship between mucositis and myelosuppression.

504

S.T. Sonis et al.

5. The pathobiology of hamster mucositis is similar to humans. Immunohistochemistry of developing mucositis is easily done. 6. The size of hamster peripheral blood cells is similar to humans, thereby permitting use of automated, instrument run analyses or peripheral blood. 7. Mucositis development in the hamster is robust and follows a predictable course. Erythema, superficial necrosis, and frank ulceration are followed by spontaneous healing.

20.2.4.1 Screening Models for the Enablement of Pharmaceuticals and Biologicals Background Largely because of their predictive value, hamster models have become a workhorse for screening and efficacy testing for pharmaceuticals and biologicals as potential interventions for the prevention and treatment of mucositis. Four variations of the hamster model have been developed, tested, and validated. History of Radiation Model Development in Hamsters As noted above, the first animal models for radiation injury were based on the human paradigm in which an external radiation source induces intraoral injury. Earlier, we attempted to replicate this approach by radiating the faces and cheeks of mice, rats, and hamsters. The animals were placed in a plastic stint (a modified 50 cc polyethylene centrifuge tube with the conical end removed) and shielded with lead so that only the facial region was exposed. Doses of 25 or 30 Gy resulted in erythema, inflammation, and swelling of the mucosa, but within 1 week following radiation, marked peri-oral inflammatory changes were seen, and animals became moribund and demonstrated severe (>20%) weight loss. We concluded that this approach was unsatisfactory. Consequently, we developed a technique which allowed us to isolate the cheek pouch mucosa. A lead shield was fabricated with a slit at its base in the approximate length of the cheek pouch base. Following anesthesia, the shield was positioned so that only the cheek pouch mucosa protruded, and therefore was the only tissue exposed to radiation. The technique proved to be effective in producing consistent pseudomembranous ulcerative mucositis [12]. 20.2.4.2 The Hamster Model for Acute Radiation-Induced Mucositis Golden Syrian hamsters, aged 5–6 weeks and weighing between 80 and 100 g are obtained from Harlan Sprague Dawley or Charles River Laboratories, housed in small groups, fed standard hamster chow, and watered ad libitum. Mucositis is


505

induced with a single dose of radiation (40 Gy/dose) is administered on day 0 with a 160 kV potential (15-ma) Kimtron Polaris source at a focal distance of 40 cm, hardened with a 0.35 mm Al filtration system. Irradiation targets the left buccal pouch mucosa at a rate of 2.5 Gy/min. Prior to irradiation, animals are anesthetized with an intraperitoneal injection of ketamine (160 mg/ml) and xylazine (8 mg/ml). The left buccal pouch is everted, fixed, and isolated using a lead shield. Initial studies focused on establishing an optimal radiation dose that would consistently provide ulcerative mucositis that was clinically obvious, but not so severe that we would not be able to determine if a test drug exacerbated the response or was a radiosensitizer. Dose ranging studies were done in which we compared the time-course and clinical and clinical response from doses of 25–40 Gy. We noted that, even with doses of 25 Gy, mild mucositis, characterized by erythematous changes, were seen by 10 days following radiation. After day 10, dosedependent changes were noted. Earlier studies had demonstrated that a dose threshold of 20 Gy was required to produce consistent injury. A 6-point scale was developed to grade mucositis using outcomes that are analogous to clinical scoring (e.g. the NCI-CTC v3 scale) (Table 20.1 and Fig. 20.1–20.3). Little systemic impact was seen as animals were only radiated on one side, so food intake and

Table 20.1 Grading Scale to Describe Mucositis Severity in Hamster Models 0. 1. 2. 3. 4. 5.

Pouch completely healthy. No erythema or ulceration Erythema and vasodilation, but mucosa intact Severe erythema with superficial mucosal erosion Formation of mucosal ulceration with a cumulative size of about 25% of the pouch’s surface area Ulceration with a cumulative size of about 50% of the pouch’s surface area Contiguous ulceration involving almost the entire surface area of the pouch mucosa

Fig. 20.1 Normal hamster mucosa. Score of 0. The pouch is completely healthy with no erythema or ulceration

506

S.T. Sonis et al.

Fig. 20.2 Mild to moderate oral mucositis with a score of 2. Mucosal sloughing is present and erythema is noted. Nonetheless, the epithelial integrity is intact

Fig. 20.3 Severe oral mucositis. Ulceration with necrosis and pseudomembrane formation is clearly seen. The surface area of the ulcer exceeds 25% of the pouch service and pliability of the pouch is reduced. This lesion is scored as a 4

weight were not adversely impacted. This fact was important relative to the models’ acceptability by animal use committees. We have determined that an acute dose of 40 Gy produces the most replicable results and that the ability of test agents to attenuate mucositis at this dose is translatable to clinical efficacy in humans. Clinically relevant mucositis characterized by erythema and superficial sloughing is seen by day 12. Typically, ulceration (the most significant clinically relevant endpoint) is noted around day 15 and lasts for about 5 days. Mucositis spontaneously resolves by day 35. For interventional studies,


507

Acute Mucositis - Saline Controls: Combined data from 78 animals/10 studies 5

Mean Mucositis Scores (+/-SEM)

Saline Controls

4

3

2

1

0

4

6

8

10

12

14

16 18 Day

20

22

24

26

28

30

Fig. 20.4 Mucositis curve representing meta-analysis of 78 hamsters in 10 experiments following an acute dose of 400% ulcerative mucositis (grade ≥ 3) was seen in 42% of evaluable days

we have found that the most predictive endpoint is not peak mucositis score, but instead the ability of the test compound to favorably impact the number of days that animals demonstrate ulcerative mucositis (scores ³3) [13, 14]. A meta-analysis in which data from 10 experiments of 78 animals treated with saline was evaluated demonstrated that animals demonstrated ulcerative mucositis on 42% of study days (evaluated days 6–28) (Fig. 20.4). No significant impact was seen on survival. The acute model has proven to be most valuable as a screening tool for test compounds. A positive impact (25% or more reduction in ulcerative mucositis days) has been a consistent predictor of clinical effect. The model has also been valuable in optimizing dose and study drug formulation, particularly of topically applied agents. However, there are no clinical protocols in which patients receive such high focal doses of radiation. Consequently, other models were developed to optimize pre-clinical drug development. 20.2.4.3 Non-Clinical Endpoints The hamster cheek pouch has provided significant information relative to the descriptive and quantitative histologies, immunohistochemistry, and gene expression studies [12, 15]. The cheek pouch can be easily excised. For histological and immunohistochemical studies, it is best prepared by surgically opening the pouch to create a single layer of tissue and then using a cassette for fixation.

508

S.T. Sonis et al.

20.2.4.4 The Hamster Model for Fractionated Radiation-Induced Mucositis Radiation treatment for human head and neck cancer is a consistent cause of ulcerative mucositis. Radiation in these cases is typically given in small, daily fractionated doses of 2 Gy, 5 days per week, for cumulative doses of 60–70 Gy. This regimen results in a predictable pattern of mucositis with erythematous changes occurring at cumulative doses of 10–20 Gy and ulceration developing at doses of 30 Gy. Ulceration usually persists until 2–4 weeks after the completion of radiation. Although the acute radiation model described above provides an excellent method to assess acute toxicity, it has limited applicability to determine optimal scheduling of investigative agents destined to be applied to the head and neck cancer population. Consequently, we developed a model in which animals were irradiated for four consecutive days with daily doses ranging from 5–15 Gy and cumulative doses of 20–60 Gy per animal [16]. Like patients, mucositis induced by fractionated radiation dosing in hamsters is dose dependent. A reproducible regimen which produces significant ulcerative mucositis can be achieved using a cumulative radiation dose of 60 Gy. To replicate the human dosing schedule, the dose is divided into 7.5 Gy fractions delivered for four consecutive days, followed by a 2-day rest period, after which the 4-day regimen is repeated. The use of two cycles permits the study of interventional agents in schedules that are similar to humans. Since mucositis severity is dependent on the incremental and cumulative dose of radiation, daily fractions >7.5 Gy increase mucosal injury and extend the duration. Alternatively, increasing the number of cycles produces a similar result, but is logistically more taxing.

20.2.4.5 Hamster Models for Chemotherapy-Induced Mucositis With or Without Concomitant Radiation In many instances, including treatment of cancers of the head and neck, radiation is rarely given without concomitant chemotherapy. Since this approach involves agents that are synergistic in their ability to produce mucosal injury, a model combining the two is an important component in the development of any anti-mucositis agent [17].

20.2.4.6 Chemotherapy-Induced Oral Mucositis Oral mucosal toxicitiy is affected by the choice of chemotherapeutic agent, its dose and frequency, interval between dosing, and animal age. A number of preliminary studies have been performed to identify the optimal dose and schedule for 5-FU administration. A single large bolus dose of 5-FU caused mortality without mucositis. Multiple dosing at intervals of 5 days with moderate dosages was stomatotoxic with little mortality. Three intraperitoneal (IP) doses of 60 mg/kg of 5-FU administered


509

on days 0, 5, and 10 were initially used. Since the cheek pouch mucosa is anatomically protected from functional trauma as occurs in humans, superficial irritation of the mucosa was performed using an 18-gauge needle on days 1 and 2. Importantly, any evidence of irritation resolved in 24 h in animals not receiving 5-FU. The combination of three doses of 5-FU and superficial irritation consistently produced ulcerative mucositis without significant mortality. Slight, but insignificant, weight loss was noted following the second injection of 5-FU on day 5. Evaluation of peripheral white blood cells (WBC) showed marked myelosuppression on day 10 and second dip in WBC on day 14. No changes in animal activity levels were seen [18]. Early mucositis was noted on day 7, after the initial injection of 5-FU. Marked progressive mucositis characterized by large areas of epithelial disruption and surface necrosis was present in 71% of animals by day 9, and 100% of animals had robust lesions by day 14. Additional studies using this model have demonstrated that the maximum depression of femoral bone marrow cellularity was preceded by mucosal breakdown and that mucosal healing occurred by bone marrow recovery. 20.2.4.7 Concomitant Chemotherapy and Radiation Concomitant chemoradiation is significantly stomatotoxic. Mucositis is induced using 5-FU delivered as single ip doses (60 mg/kg) on days -4 and -2. A single dose of radiation (30 Gy/dose) is administered on day 0 [17].

20.3 Chemotherapy-Induced Mucositis of the GI Tract Chemotherapy-induced mucositis may affect any portion of the gastrointestinal tract. Unlike radiation therapy, which is targeted to the tumor site, chemotherapy is most often administered by intravenous infusion. Consequently, its effects are widespread, but often impact the small and large intestine resulting in tissue injury and symptoms of enteritis (i.e. diarrhea). The time course of small intestinal injury is more acute that oral mucositis and is largely governed by the lack of stratified epithelium.

20.3.1 Models of Intestinal Mucositis Chemotherapy-induced intestinal injury has been readily induced in mice and rats using a variety of drugs including methotrexate, cytarabine, 5-fluorouracil, CPT-11, and doxorubicin [19–22]. Keefe et al. have successfully used irinotecan (CPT-11) in dark agouti rats to study the pathobiology of intestinal mucositis and evaluate potential interventions [23]. In their model, female dark agouti rats (weights between 150 and 170 g) are injected intraperitoneally with 200 mg/kg of irinotecan in a

510

S.T. Sonis et al.

sorbitol/lactic acid buffer. Animals also receive 0.01 mg/kg of subcutaneous atropine to reduce the cholinergic reaction [24]. Endpoints with this model include diarrhea severity, measured on a 0 (no diarrhea) to 3 (staining over legs and higher abdomen, often with continual anal leakage), descriptive histology, quantitative histology, and immunohistochemistry. Widespread damage is easily noted. For the induction of intestinal mucositis in mice, the dose and schedule of CPT-11 have been modified. 20.3.1.1 Use of Endoscopy to Assess Chemotherapy-Induced Mucosal Injury In an effort to overcome the necessity of relying on histological endpoints to assess mucosal damage and permit in situ assessment of mucositis progression, we have applied video endoscopy to models of gastrointestinal mucositis. Mucositis is induced in mice via a single 60 mg/kg intraperitoneal injection of the chemotherapeutic, methotrexate, and the colon of each mouse is examined using a small animal video endoscope at 24, 48, and 72 h following treatment. Video endoscopy allows for daily visual assessment of the extent and severity of disease using clinically relevant endpoints as well as the tracking of mucosal healing following any potential therapeutic interventions. To conduct endoscopy the mice are anesthetized with isoflurane and the endoscope is slowly inserted into the rectum while the colon is insufflated with air. This method allows for clear imaging of the colon mucosal surface and both video and still images are recorded to help assess disease severity. Severity of disease is scored on 0–4 scale with 0 = normal, 1 = loss of vascularity, 2 = loss of vascularity and friability, 3 = friability and erosions, and 4 = ulcerations and bleeding. Changes in the colon observed endoscopically include modification of the vascular pattern, often accompanied by friability, erosion, and active bleeding (Fig. 20.5).

Methotrexate-Induced GI Mucositis: Video Endoscopy

Normal Colon

24 hours

48 hours

72 hours

Fig. 20.5 The sequence of development of methotrexate-induced mucosal injury of the colon as observed in the same animal using the video endoscopic technique described in the text. The use of the endoscope to evaluate injury precludes the use of sacrifice and histological sampling to determine the extent of injury. Tissue sampling is possible using the endoscope for access


511

20.4 Radiation-Induced Proctitis 20.4.1 Overview of the Condition Radiation-induced proctitis is a common complication associated with radiation directed at the lower abdomen or pelvis. Radiotherapy is usually a major portion of the treatment paradigm for rectal, prostate, or cervical malignancies. There are both acute and delayed forms of radiation proctitis that may present following radiation therapy. The symptoms of acute proctitis have been reported as a complication in as many as 75% of patients undergoing pelvic radiation [25] with symptoms beginning almost immediately following therapy and persisting for up to 3 months. Delayed or late radiation proctitis is a chronic disease that can occur well after the completion of the radiation therapy and result in symptoms that may persist for decades. Disease symptomology usually begins in the acute phase during the first or second week following radiation, with presentation of rectal bleeding that can be associated with diarrhea and/ or discharge of mucus. Late proctitis symptoms appear anywhere from a few months to several years following the radiation therapy and as with the acute disease, chronic rectal bleeding is the most common late symptom. Other late symptoms may include a discharge of mucus from the rectum and more instances of tenesmus, or a feeling or inability to empty the bowel upon defecation. Tenesmus is thought to be a result of extensive rectal tissue fibrosis or possibly the formation of rectal strictures. Diagnosis and treatment of radiation proctitis usually follow a colonoscopy to exclude the possibility of any other underlying pathology. Current treatment options for radiation proctitis vary greatly in method as well as rate of success [26]. While there are few, if any, compelling clinical studies in the treatment of radiation proctitis the majority of therapies are based on the results of small unblinded studies with mixed results. The most common first-line therapies have been adapted from the treatments used in inflammatory bowel disease (IBD), including 5-ASA, steroids, sucralfate, and metronidazole. Increasingly, endoscopic therapies are being employed to control bleeding associated with radiation proctitis which includes heat probes, lasers, and most commonly argon plasma coagulation (APC). APC involves the flow ionized argon gas and can target small or larger areas of bleeding without making physical contact to the tissue. While APC has been shown to be quite effective in a number of small studies [26], there are no large clinical studies to help support these results.

20.4.2 The Biology of Radiation-Induced Proctitis It is likely that the overall pathogenesis of mucositis described for the mouth and GI tract is similar to that of proctitis, although the mechanisms of radiation-induced proctitis have not been aggressively studied. Clearly, vascular changes and fibrosis are seen as end points of the condition. In the days to months following irradiation, the onset of acute proctitis begins with early apoptosis, disruptions of mitosis, and fibroblastic proliferation that leads to swelling and sloughing [27, 28], while the later

512

S.T. Sonis et al.

changes are primarily vascular and fibrotic in nature [29]. The neovascularization and telangiectasia that occur in the late phases of disease lead to persistent bleeding while increased fibrosis may lead to ischemia and eventually necrosis of the bowel tissue [29, 30]. While the exact mechanisms for the late changes in vascularity as well as fibrosis have yet to be elucidated, there is evidence that several growth factors including: platelet-derived growth factor, vascular endothelial growth factor, and fibroblast growth factor all have a roll in the pathology [28]. Other studies have shown that radiation-induced increased expression avb3 integrin has potent angiogenic effects and may contribute to the vascular changes observed in radiation proctitis [31]. Another recent study demonstrated that lesions in radiation proctitis are associated with increased angiogenesis and associated increased vascular expression of CD39 may be involved in the long-term sequel of radiation-induced proctitis [32]. Additional studies are needed to further understand the mechanisms of radiation proctitis and provide avenues for effective interventions.

20.4.3 Animal Model of Radiation-Induced Proctitis Recent modifications of rat models of radiation proctitis [33] have proven to be effective in creating an effective translational model of the condition. A single 17.5 Gy dose of radiation is directed at the rectum of a rat to induce proctitis. Critical to the success of the model has been the observation that angle of radiation exposure impact toxicity. Radiation is performed following the administration of ketamine and xylazine anesthesia. Animals are placed on a silicon rubber sheet and lead shielding is used to isolate the lower abdomen. The rectum is then irradiated with a single, acute, 17.5 Gy dose of radiation using a 160 kVp (15-ma) source at a focal distance of 30 cm, hardened with a 0.35 mm Cu filtration system at a rate of 2.5 Gy/min. Video endoscopy is used to assess the levels of disease at multiple time points following radiation (Fig. 20.3). Endosopic correlation with pathologic findings has been confirmed using models for inflammatory bowel disease [34]. While endoscopic data add an important visual and clinical component to the animal models of proctitis, the use of histology can still add valuable confirmatory information several weeks following the study completion (Fig. 20.2). Endoscopy has been valuable in evaluating the efficacy of potential therapeutics in radiation proctitis. 20.4.3.1 Radiation-Induced Dermatitis Overview of the Condition Dermatitis is a side effect of radiation therapy in patients in whom the skin is exposed to the radiation source. Acute dermatitis is most commonly reported in patients with cutaneous neoplasms, and head and neck, and breast cancer and usually begins to appear during the second to third week of fractionated radiation. Five to 10% of patients treated with radiation therapy for breast cancer develop moderate to severe radiation-induced dermatitis [35, 36]. The condition is unpleasant,


513

painful, and may contribute to poor QoL. In some cases, radiation dermatitis may become so severe as to necessitate interruption or cessation of radiation therapy. Radiation dermatitis usually begins with reddening of the skin, but may also include the epilation, dry desquamation, wet desquamation, decreased sweating, edema, ulceration, and bleeding. Areas of telangiectasia, hyper- and hypo-pigmentation are also common. Symptom progression depends on both treatment-related factors, such as the total radiation dose, fractionation, total duration of treatment, volume of tissue irradiated, and type of radiation delivered, as well as patient-related factors, such as age menopausal state and pre-disposing genetic factors [37]. Current treatment options for radiation dermatitis are primarily palliative and include a wide variety of products. In general, emollients or hydrating lotions are used during the early stages of the condition. If symptoms progress, however, topical corticosteroids, dressings, and/or radiation dose reduction may be necessary [38, 39]. There are no mechanistically based interventions.

20.4.4 The Biology of Radiation-Induced Dermatitis Although the mechanism has not been well studied, it is likely that the pathoetiology of radiation-induced dermatitis is similar to other epithelial toxicities. DNA damage caused by ionizing radiation generates free radicals, which in turn activate IkB kinase (IkK) inducing NF-kB translocation to the nucleus, and subsequently, transcription of pro-inflammatory mediators. This leads to a complex pattern of tissue injury and recruitment of inflammatory cells [38]. The p53 pathway may be activated, and cytokines such as interleukin 1 (IL-1), interleukin 6 (IL-6), and transforming growth factor b (TGF-b) are secreted by keratinocytes. Histological analysis of the tissue reveals a loss of keratinocyte polarization, a loss of sebaceous and sweat glands, and destruction of hair follicles.

20.4.5 Animal Model of Radiation-Induced Dermatitis We have modified the model originally described by Abe et al. [40]. A single 30 Gy dose of radiation directed at the dorsal skin of the mouse. Two days prior to irradiation, the skin on the backs of animals is prepared by removing the hair covering the entire back of each animal using an electric shaver and depilatory cream. Irradiation is performed under ketamine and xylazine anesthesia. Animals are placed on a silicon rubber sheet and the loose dorsal skin is gently stretched and secured with two 25 gauge needles. A lead shield is then placed over the animal so that an area of skin about 2 cm × 4 cm in size was exposed. The tissue is then irradiated with a single, acute, 30 Gy dose of radiation using a 160 kVp (15-ma) source at a focal distance of 30 cm, hardened with a 0.35 mm Cu filtration system at a rate of 2.5 Gy/min. In this model, dermatitis is scored every other day beginning the day after radiation using a 6-point scale (0–5) in which (Fig. 20.6) a score of 0 is defined by

514

S.T. Sonis et al.

Radiation-Induced Dermatitis: Scoring Scale Score = 0 Normal Tissue

Score = 1 Mild Erythema Minimal Dry Desquamation

Score = 2 Mild to Moderate Erythema Slight Desquamation

Score = 3 Moderate Desquamation ≤ 50%

Score = 4 Moderate to Severe Desquamation ≥ 50%

Score = 5 Frank Ulceration of Tissue

Fig. 20.6 BevDermatitis Scoring Scale


515

5

Mean Dermatitis Scores (+/-SEM)

Dermatitis30 Gy Radiation 4

3

2

1

0 0

2

4

6

8

10 12

14

16

18 20 22 24 26 28

30 32 34 36

Day

Fig. 20.7 Progression of radiation-induced dermatitis in a murine model of the condition. Animals receive a single dose of radiation (30 Gy) on day 0

normal skin, a score of 1 is defined by mild erythema, a score of 2 is defined by moderate to severe erythema with or without slight desquamation, a score of 3 is defined by desquamation of 25–50% of the irradiated area, a score of 4 is defined by desquamation of greater than 50% of the irradiated area, and a score of 5 is defined by frank ulceration of the skin. Onset of dermatitis in this model usually begins around 1 week following radiation and reaches peak severity another 7 days later (Fig. 20.7). One of the major benefits of this model is that dermatitis scores are obtained every other day and the disease persists for several weeks, making it an excellent model system to test potential therapeutics.

20.4.6 Bisphosphonate-Related Osteonecrosis of the Jaws 20.4.6.1 Introduction Bisphosphonates have been shown to be of marked benefit in reducing the bony complications of metastatic disease associated with multiple myeloma and breast cancer by inhibiting osteoclast activity, development, migration, and viability [41– 43]. A recognized side effect of bisphosphonate use is osteonecrosis of the jaws [44–46]. The risk of the condition has been placed somewhere between 1 and 10%. Aside from intravenous bisphosphonate infusion, dental manipulation has been iden-

516

S.T. Sonis et al.

tified as a significant risk factor for the condition. Studies of the natural history, pathogenesis, and risk factors for BRONJ had been limited by the need for human material and the lack of clinical predictability. Consequently, an animal model which replicated the condition has been an objective of a number of investigators. 20.4.6.2 Rat Model for Bisphosphonate-Associated Osteonecrosis of the Jaws In 2009, a description of a rat model was published in which BRONJ was elicited following a course of zoledronic acid and dexamethosone prior to tooth extraction [47]. Female Sprague–Dawley rats aged 10–12 weeks and weighing between 200 and 280 g underwent extractions of their left maxillary mandibular and maxillary molars following varying courses of zolendronic acid (7.5 mg/kg subcutaneously) and dexamethasone (1 mg/kg, subcutaneously). Clinical, radiographic, and histological endpoints were assessed. Extractions were performed 8, 15, or 22 days following the initiation of up to three courses of zoledronic acid and dexamethasone (Z/D) and animals were then evaluated 2 or 4 weeks later. The jaws were removed, split longitudinally, and photographed and radiographed using a standard dental X-ray source and digital imaging system. After excess tissue was trimmed and the bone decalcified, it was embedded in paraffin and stained for histological examination. Exposed necrotic bone and ulceration was noted 28 days following extractions in animals treated with the combination of zoledronic acid and dexamethasone, but not in animals treated with zoledronic acid alone (Fig. 20.8). Earlier studies had

Fig. 20.8 Gross observation of BRONJ in the rat model. These representative photographs d emonstrate the gross clinical appearance of the maxillary ridges of animals with intact epithelium (panel a – day 28 following extraction), or ulcerated mucosa overlying necrotic bone at days 14 (panel b), and 28 (panel c ) following extraction. Ulcerative areas are characterized by rolled mucosa, lack of drainage, and, by day 28 post-extraction, central areas of yellow/gray necrosis. From Sonis et al. Oral Oncol 2009;45:164–72


517

Fig. 20.9 Histology of specimen from not with evidence of osteonecrosis demonstrating fragments of necrotic bone

shown that the administration of dexamethasone did not confer a risk of osteonecrosis in the absence of the bisphosphonate. Histologic evidence of necrosis was also noted in the Z/D animals (Fig. 20.9). Refinement and modification of the model will likely evolve, but it provides a basis for studies relating to the mechanism of BRONJ and a way to evaluate potential interventions.

References 1. Sonis ST, Elting LS, Keefe DM, et al. Perspectives on cancer therapy-induced mucosal injury: pathogenesis, measurement, epidemiology and consequences for patients. Cancer 2004;100 (Suppl 9):1995–2025. 2. Nonzee NJ, Randade NA, Patel U, et al. Evaluating the supportive care costs of severe radiochemotherapy-induced mucositis and pharyngitis: results of a Northwestern University cost of Cancer Program pilot study with head and neck and nonsmall cell lung cancer patients who received care at a county hospital, a Veteran’s Administration hospital, or a comprehensive cancer center. Cancer 2008;113:1446–52. 3. Blijlevens N, Sonis S. Palifermin (recombinant keratinocyte growth factor-1): a pleiotropic growth factor with multiple biological activities in preventing chemotherapy- and radiotherapyinduced mucositis. Ann Oncol 2007;18:817–26. 4. Elting LS, Keefe DM, Sonis ST, et al. Patient-reported measurements of oral mucositis in head and neck cancer patients treated with radiotherapy with or without chemotherapy: demonstration of increased frequency, severity, resistance to palliation, and impact on quality of life. Cancer 2008;113:2704–13. 5. Sonis ST. The pathobiology of mucositis. Nat Rev Cancer 2004;4:277–84. 6. Sonis ST. Pathobiology of mucositis: novel insights and opportunities. J Support Oncol 2007;5:3–11.

518

S.T. Sonis et al.

7. Lee SW, Jung KI, Kim YW, et al. Effect of epidermal growth factor against radiotherapyinduced oral mucositis in rats. Int J Radiat Oncol Biol Phys 2007;67:1172–8. 8. Vivela-Goulat MG, Teixeira RT, Rangel DC, et al. Homogenious amniotic membrane as a biological dressing for oral mucositis in rats: histomorphic analysis. Arch Oral Biol 2008;53:1163–71. 9. Xu FX, van der Schueren E, Ang KK. Acute reactions of the lip of mice to fractionated irradiations. Radiother Oncol 1984;1:369–74. 10. Kilic Y, Rajewski K, Dorr W. Effect of post-exposure administration of keratinocyte growth factor (palifermin) on radiation effects in oral mucosa of mice. Radiat Environ Biophys 2007;46:13–19. 11. McMillan CD, Lowell VM. Experimental candidiasis in the hamster cheek pouch. Arch Oral Biol 1985;30:248–55. 12. Sonis ST, Peterson RL, Edwards LJ, et al. Defining mechanisms of action of Interleukin-11 on the progression of radiation-induced oral mucositis in hamsters. Oral Oncol 2000;36:372–81. 13. Murphy CK, Fey EG, Watkins BA, et al. Efficacy of superoxide dismutase mimetic M40403 in attenuating radiation-induced oral mucositis in hamsters. Clin Cancer Res 2008;14:4292–7. 14. Huang D, Popat R, Bragdon C, et al. Effects of ceramide inhibition on experimental radiationinduced oral mucositis. Oral Surg, Oral Med, Oral Pathol, Oral Radiol Endod 2005;100:321–9. 15. Sonis ST, Scherer J, Phelan E, et al. The gene expression sequence of radiated mucosa in an animal mucositis model. Cell Profif 2002;35 Suppl 1:93–102. 16. Ara G, Watkins BA, Zhong H, et al. Valifermin (rhFGF-20) reduces the severity and duration of hamster cheek pouch mucositis induced by fractionated radiation. Int J Radiat Biol 2008;84:401–12. 17. Alvarez E, Fey EG, Valax P, et al. Preclinical characterization of CG53135 (FGF-20) in radiation and concomitant chemotherapy/radiation-induced oral mucositis. Clin Cancer Res 2003;9:3453–61. 18. Sonis ST, Tracey C, Shklar G, et al. An animal model for mucositis induced by cancer chemotherapy. Oral Surg, Oral Med, Oral Pathol 1990;69:437–43. 19. Tooley KL, Howarth GS, Lymn KA, et al. Optimization of the non-invasive (13)C-sucrose breath test in a rat model of methotrexate-induced mucositis. Cancer Chemother Pharmacol 2010; 65:913–21. 20. Cheah KY, Howarth GS, Yazbeck R, et al. Grape seed extract protects IEC-6 cells from chemotherapy-induced cytotoxicity and improves parameters of small intestinal mucositis in rats with experimentally-induced mucositis. Cancer Biol Ther 2009;8:382–90. 21. Stringer Am, Gibson RJ, Logan RM, et al. Gastrointestinal microflora and mucins may play a critical role in the development of 5-fluorouracil-induced gastrointestinal mucositis. Exp Biol Med 2009;234:430–41. 22. Logan RM, Stringer AM, Bowen JM, et al. Is the pathobiology of chemotherapy-induced alimentary tract mucositis influenced by the type of mucotoxic drug administered? Cancer Chemother Pharmacol 2009;63:239–51. 23. Gibson RJ, Bowen JM, Alvarez E, et al. Establishment of a single-dose irinotecan model of gastrointestinal mucositis. Chemotherapy 2007;53:360–9. 24. Stringer AM, Gibson RJ, Logan RM, et al. Irinotecan-induced mucositis is associated with changes in intestinal mucins. Cancer Chemother Pharmacol 2009;64:123–32. 25. Counter SF, Froese DP, Hart MJ. Prospective evaluation of formalin therapy for radiation proctitis. Am J Surg 1999;177:396–8. 26. Leiper K, Morris AI. Treatment of radiation proctitis. Clin Oncol 2007;19:724–9. 27. Haboubi NY, Schofield PF, Rowland PL. The light and electron microscopic features of early and late phase radiation-induced proctitis. Am J Gastroenterol 1988;83:1140–44. 28. Brunn T, Fletcher CD. Postradiation vascular proliferations: an increasing problem. Histophathology 2006;48:106–14. 29. Fajardo LF. The pathology of ionizing radiation as defined by morphologic patterns. Acta Oncol 2005;44:13–22.


519

30. Haselton PS, Carr N, Schofield PF. Vascular changes in radiation bowel disease. Histopathology 1985;9:517–34. 31. Abdollani A, Griggs DW, Zieher H, et al. Inhibition of alpha (v) beta 3 integrin survival signaling enhances antiangiogenic and antitumor effects of radiotherapy. Clin Cancer Res 2005;11:6270–9. 32. Sheth S, Bleibel W, Thukral C, et al. Heightened NTPDase-1/CD39 expression and angiogenesis in radiation proctitis. Purinergic Signalling 2009;5:321–6. 33. Kang S, Chun M, Jin YM, et al. A rat model for radiation-induced proctitis. J Korean Med Sci 2000;15:682–89. 34. Lyng CD, Stevens AC, Gordon GJ, et al. Rodent video endoscopy and biopsy: a new method for the development of novel inflammatory bowel disease therapies. Gastroenterology 2008;134:A262. 35. Rosen EM, Fan S, Rockwell S, et al. The molecular and cellular basis of radiosensitivity: implications for understanding how normal tissues and tumors respond to therapeutic radiation. Cancer Invest 1999;17:56–72. 36. Fujishiro S, Mitumori M, Kokubo M, et al. Cosmetic results and complications after breast conserving therapy for early breast cancer. Breast Cancer 2000;7:57–63. 37. Isomura M, Oya N, Tachiiri S, et al. IL12RB2 and ABCA1 genes are associated with susceptibility to radiation dermatitis. Clin Cancer Res 2008;20:6683–89. 38. Hymes SR, Strom EA, Fife C. Radiation dermatitis: clinical presentation, pathophysiology and treatment. J Am Acad Dermatol 2006;54:28–46. 39. McQuestion M. Evidence-based skin care management in radiation therapy. Semin Oncol Nurs 2006;22:162–73. 40. Abe Y, Urano M. Fraction size-dependent acute skin reaction of mice after multiple twice-aday doses. Int J Radiat Oncol Biol Phys 1990;18:359–64. 41. Dunstan CR, Felsenberg D, Seibel MJ. Therapy insight: the risks and benefits of bisphosphonates for the treatment of tumor-induced bone disease. Nat Clin Pract Oncol 2007;4:42–55. 42. Lipton A. Efficacy and safety of intravenous bisphosphonates in patients with bone metastases caused by metastatic breast cancer. Clin Breast Cancer 2007;7(Suppl 1):S14–25. 43. Luftner D, Henshke P, Possinger K. Clinical value of bisphosphonates in cancer therapy. Anticancer Res 2007;27:1759–68. 44. Marx RE. Pamidronate (Aredia) and zoledronate (Zometa) induced avascular necrosis of the jaws: a growing epidemic. J Oral Maxillofac Surg 2003;61:1115–7. 45. Hellstein JW, Marek CL. Bisphosphonate osteochemonecrosis (Bis-Phossy Jaw): is this Phossy Jaw of the 21st century? J Oral Maxillofac Surg 2005;63:682–9. 46. Ruggiero SL, Merota B, Rosenberg TJ, et al. Osteonecrosis of the jaws associated with the use of bisphosphonates: a review of 63 cases. J Oral Maxillofac Surg 2004;62:527–34. 47. Sonis ST, Watkins BW, Lyng GD, et al. Bony changes in jaws of rats treated with zoledronic acid and dexamethasone before dental extractions mimic bisphosphonate-related osteonecrosis in cancer patients. Oral Oncol 2009;45:164–72.

Chapter 21

Bone Marrow as a Critical Normal Tissue that Limits Drug Dose/Exposure in Preclinical Models and the Clinic1 Ralph E. Parchment

Abstract Mouse models of cancer have played an important role in the discovery and development of cytotoxic and targeted anticancer agents. Initially, discovery models used transplantable syngeneic tumors that were treated with doses maximally tolerated by the normal tissues of the mouse. Thus, experimental compounds were selected for development based on their selectivity for murine malignant tissue over murine normal tissue, and the discovery method assumed that a murine therapeutic index closely approximates a human therapeutic index for most compounds. When mouse modeling migrated to the use of xenografted human malignancies in order that drug efficacy assessment would be more relevant for clinical disease, there was not a corresponding transition to human normal tissue to determine the maximum tolerated dose (MTD) to use in treating the mouse. Consequently, drug discovery in these models has been based on selectivity for human malignant tissue over murine normal tissue. This “xeno-therapeutic index” has an unknown relationship either to a mouse or human therapeutic index – these latter ones being determined by efficacy against cancer at the maximum dose tolerated by normal tissues from the same species of origin as the cancer. The biological process of producing new blood cells is termed hematopoiesis, and this process is a frequent target of toxicity of anticancer drugs, including toxicity that limits dose. Two recently established methods in experimental hematology and hematotoxicology are useful for determining the MTD of human hematopoiesis in the mouse. The methods are suitable for implementation in the drug discovery setting, and therefore could be used to discover new anticancer compounds that exhibit selectivity for malignant human 1 This project has been funded in whole or in part with federal funds from the National Cancer Institute, National Institutes of Health, under Contract No. HHSN261200800001E. This research was supported [in part] by the Developmental Therapeutics Program in the Division of Cancer Treatment and Diagnosis of the National Cancer Institute. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

R.E. Parchment (*) Laboratory of Human Toxicology and Pharmacology, SAIC-Frederick, Inc., NCI-Frederick, Frederick, MD 21702, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_21, © Springer Science+Business Media, LLC 2011

521

522

R.E. Parchment

tissue over normal, human hematopoiesis in mouse models. This chapter provides background information on hematopoiesis, explains why it frequently limits the dose of anticancer drugs, describes the two methods, and then proposes ways in which the methods might contribute to mouse modeling of cancer therapy to improve predictive accuracy for clinical outcome. Keywords Bone marrow • Hematopoiesis • Progenitor • Stem cell • Hematotoxicology • Neutropenia • Myelosuppression • Dose-limiting toxicity • Adverse drug effect • Therapeutic index • Maximum tolerated dose • Prediction models • CFU-GM • Engraftment

21.1 Introduction Many clinically approved chemotherapeutic agents were discovered with preclinical mouse models using transplantable syngeneic cancers of spontaneous or induced origin. In the discovery arena, the dose of an experimental agent was usually pushed to the level at which toxicity to a normal mouse tissue became lifethreatening, or just below that level. Since both the dose-limiting tissue and the malignancy were murine, experimental agents were prioritized and selected for development based on an approximate murine therapeutic index, i.e. efficacy at maximum tolerated toxicity or the toxic dose divided by the efficacious dose. The assumption behind the use of these murine models was a correlation between murine therapeutic index and human therapeutic index. Once both qualitative and quantitative biochemical differences between murine and human malignancies began to be discovered, there was naturally an impetus to develop murine models harboring human disease targets, either transplantable human tumor lines, primary transplants of surgical specimens, or more recently, genetically engineered mouse models harboring human molecular targets in the context of the mouse genetic background. These models, and their rationale as more relevant and hopefully more accurate models of clinical cancer, are given expert, detailed coverage throughout this volume. Of relevance to this chapter is the importance of noting that this move to “humanize” the mouse models stopped short of being complete. Once mouse models containing human disease targets were developed, there was little effort to humanize any other aspects of mouse modeling. Essentially the murine disease target was replaced with a human one, but the normal tissues that limit dose and make an equal contribution to therapeutic index remained murine. As a result, the murine therapeutic index has not yet been replaced by a human therapeutic index as the basis for selecting and prioritizing experimental compounds for development. Instead, compounds are being evaluated based on a “xeno-therapeutic” index that relates human efficacy to mouse dose tolerance, in essence modeling a situation in which the clinical treatment of human patients is somehow limited by the end organ dose tolerance of therapy in mice being treated in parallel. The current use of a chimeric mouse model – human disease and mouse normal tissue that

21 Bone Marrow as a Critical Normal Tissue that Limits Drug Dose/Exposure

523

limits dose – makes little sense, because the effectiveness of a drug is determined by its therapeutic index in man and therefore its selectivity for tumor over normal tissues within a species, not between two species. This chapter focuses on methods to get to an approximate human therapeutic index in the preclinical discovery setting using mouse models. Hematotoxicology is a wellestablished field, and both in vitro and in vivo methods are now available for quantifying differences between mouse and man in bone marrow tolerance of anticancer agents. The accepted philosophy of humanizing mouse models is thereby extended to doselimiting normal tissue by methods that replace mouse bone marrow with human bone marrow. Although there is not enough evidence yet to conclude that human therapeutic index-based selection and prioritization of experimental compounds from mouse modeling of human cancer provides a higher success rate in clinical development than murine or xeno-therapeutic index-based decisions, it seems reasonable to explore such modeling with key case studies – especially re-visiting very potent, “highly active” experimental compounds that failed clinical development due to severe clinical toxicity at doses/exposures far below that required for activity in mouse models. In addition, it seems unreasonable to select compounds for development that require doses above what human normal tissue like bone marrow will tolerate, unless they are planned for use only in the stem cell transplant setting. After introducing the field of hematotoxicology as related to cancer therapy, this chapter concludes with two specific methods for quantifying differences between murine and human bone marrow drug tolerance and integrating this information into mouse modeling for discovery and prioritization of new agents for development.

21.2 Blood Cytopenia as a Quantifiable Dose-Limiting Toxicity in the Oncology Clinic The blood of adult mammals contains a number of cell types that perform critical functions in health. The blood cells are classified into four major lineages based on morphology and ontogeny: granulocytes, lymphocytes, red blood cells and platelets. Homeostatic maintenance of blood cell concentrations, and in some cases specific ranges, is critical for achieving adequate oxygen delivery to tissue, immunity and hemostasis. Many antineoplastic agents, a number of other pharmaceutical agents, and some food contaminants can adversely affect the concentration of one or more blood cell types as determined by clinical laboratory measurements. The resulting severity of any clinical side effect will range from unremarkable to life-threatening, depending on the extent and duration of the change. It is important to recognize that anticancer therapy is often administered intentionally at maximum tolerated dose (MTD). One important objective of clinical Phase I dose-escalation trials is the identification of the dose that causes dose-limiting toxicity, defined as severe or lifethreatening toxicity caused by the investigational agent. The human MTD is defined as the dose just below that which causes dose-limiting toxicity, as long as patients recover from the moderate toxicity associated with that lower dose in time to maintain treatment schedules (“reversible toxicity”). Formal systems have been developed to

524

R.E. Parchment

standardize the classification of adverse toxic effects to a number of organ systems as well as the grading of their severity. The reader is referred to the most current version of the National Cancer Institute’s Common Toxicity Criteria and the Common Terminology Criteria for Adverse Events for classifying and grading severity of adverse events caused by investigational therapy during clinical trials [1]. For many anticancer drugs, the dose-limiting toxicity is a severe or life-threatening adverse effect on the hematologic system, as defined by the CTC. A decrease in a blood cell concentration, known as a cytopenia, is the most commonly encountered adverse effect of oncolytic drug therapy on the hematologic system. There are more specific terms used to describe various cytopenic conditions specifically affecting one blood cell lineage. The term “neutropenia” refers to a decrement in the blood concentration of a prevalent type of granulocyte called a neutrophil. Neutropenia is a subset of the condition known as leukopenia, which is literally a decrease in the blood concentration of all types of white blood cells counted together. Of all the leukocyte subsets that could be monitored during cancer therapy of human patients, neutropenia draws special attention, because decreased neutrophil granulocyte counts of even a few days’ duration are associated with increased risk of infection and septicemia. In the adult human, the normal concentration of neutrophils is 1,800 –7,700 ml–1 [2], and a life-threatening adverse effect of drug therapy (Grade 4) is defined as a drug-related drop in concentration below 500 for a few days or more [1]. The condition of severely decreased neutrophil counts approaching or reaching zero is termed agranulocytosis. Blood concentrations of monocytes may also decrease when neutropenia occurs, but isolated monocytopenia rarely occurs in man [3]. Except for leukemic disease, an overabundance of neutrophils or monocytes in the blood does not usually produce clinical side effects, although overabundant eosinophils are associated with clinical toxicity [3]. Of relevance to veterinary models in cancer drug development, it is important to note that there are substantial inter-species differences in the values of hematology parameters that indicate clinical toxicity. A similar 70–90% decline in the blood concentration of neutrophils, which causes significant clinical risk of infection in humans and dogs, may result in few if any clinical consequences in rodents, which have a much higher lymphocyte:neutrophil ratio than man (or dog) and are often housed in isolator/barrier facilities designed to minimize risk of infection. The white blood cell concentration in the mouse of ~6,000 ml–1 is similar to that in man, but the distribution of the various leukocyte cell types is different: ~75% of the white blood cells in the mouse are lymphocytes, i.e. a lymphocyte concentration of 4,500 ml–1 that is well above the typical human concentration [4]. A second, commonly encountered hematologic toxicity of cancer therapy is “thrombocytopenia,” which refers to a decrease in the blood concentration of thrombocytes (“platelets”). In man, the normal blood concentration of thrombocytes is 150,000–440,000 ml–1 [2], and a life-threatening (Grade 4) adverse effect of drug therapy is defined as a drop in the blood concentration of platelets below 25,000 ml–1 [1]. However, it is recognized that there is not a strong relationship between platelet count and bleeding manifestations, because clinical consequences of platelet counts are tempered by platelet function [5]. In patients with normally functioning platelets, bleeding time correlates with platelet count, whereas patients


525

with platelet abnormalities may show prolonged bleeding times despite normal platelet counts [5]. Of note, mouse exhibits normal platelet concentrations of ~1,200,000 ml–1, much higher than in healthy humans, but mouse platelets are smaller in size than human platelets [4]. Cytopenia in the erythroid or lymphoid lineage is a less frequent clinical concern of anticancer therapy than neutropenia and thrombocytopenia. However, chemotherapeutic agents destroy lymphocytes and result in a decreased count in the blood termed a lymphocytopenia. The normal blood lymphocyte concentration is 1,500– 4,000 ml–1, although counts as low as 1,000 ml–1 may be considered normal in adult humans [6, 7]. A Grade 4 adverse effect on the lymphocyte concentration is defined as a decrease to less than 200 ml–1 [1]. In addition, substantial clinical knowledge about immunodeficiency conditions and susceptibility to opportunistic infections has resulted in a specific definition of life-threatening drug toxicity as a decrease in blood concentration of CD4+ lymphocytes to 20,000 ml–1 [1]. Erythrocyte counts are usually unaffected by acute dosing schedules of cytotoxic chemotherapy, unless the agents cause hemolysis. In contrast to the other lineages mentioned above, toxicity is not judged solely on the basis of a decreased blood concentration of cells, i.e. “erythropenia.” Any adverse drug effect on the red cell lineage (anemia) is usually described by the magnitude of the resulting drop in hemoglobin concentration or in the hematocrit [1], which is the portion of blood volume occupied by erythrocytes, considering both red cell concentration and cell volume [2]. Hemoglobin concentration and packed red cell volume are relatively uniform across many mammalian species, even though erythrocyte concentrations vary [4].

21.3 Cancer Therapeutics as Toxicants to Highly Proliferative Hematopoietic Cells In the absence of infections, hemorrhagic events, hematologic disease or autoimmune reactions, the blood concentrations of the four hematologic lineages described above remain remarkably stable over many years of life. However, these stable concentrations belie the fact that mature cells are continually lost from the circulation, and new cells must therefore be produced to replace these losses. In fact, loss rates and therefore life-spans of various blood cell types vary widely (Table 21.1), ranging from the very short disappearance half-life of 7 h for neutrophils [8], to a 9–11 day life-span of platelets [9, 10], to a life-span of ~4 months for erythrocytes [11]. The rates of loss of mature blood cells from the circulation can be used to calculate rates of production of new, replacement cells required to maintain blood cell concentrations (Table 21.1). Because both cell concentrations and their specific life-spans in the blood are lineage-specific, the replacement rates also differ between the cell lineages [8, 9, 12]. As is apparent from the magnitude of these rates, there must be highly proliferative tissue(s) that is the source of production of these replacement cells.

526

R.E. Parchment

Table 21.1 Hematology values and calculated hematopoietic rates for the three blood cell lineages that are affected adversely by many anticancer agents Loss rate from Human cell type Blood concentration circulation Replacement rate Neutrophil [2, 8] 7-h half-life 0.9–1.6 × 109 kg–1 day–1 1.8-7.7 × 103 ml–1 Erythrocyte [2, 11, 12] 5,000 × 103 ml–1 120-d life span 3 × 109 kg–1 day–1 3 –1 Platelet [2, 9, 10] 10-d life span 3.3 × 109/kg/day 150-440 × 10 ml (5 × 104 µl–1 blood/day)

Fig. 21.1 Hematopoiesis in red marrow and its disruption by potential mechanisms of toxicity. (left) Photomicrograph of a histological section of hematopoietic (red) bone marrow in the mouse (courtesy of Dr. Miriam Anver, Pathology/Histotechnology Laboratory, the National Cancer Institute at Frederick). (right) Progenitor model of bone marrow hematopoiesis. The process begins with a multi-potential progenitor (purple cell) that is quiescent most of the time and therefore more resistance to drug toxicity. Via the influence of the cytokine milieu and interactions with stromal cells, this cell generates progeny committed to a particular blood cell lineage (blue, granulocyte/macrophage; orange, erythrocytes). Committed CFU-GM and CFU-G progenitors (blue circles-low stipples) encountering trophic factors (e.g. colony stimulating factors, blue boxes) survive and divide to produce differentiated, morphologically identifiable daughter cells (myeloblast, promyelocyte, myelocyte; blue circles-high stipples). Proliferation is tightly linked to differentiation to a post-mitotic state via paracrine/autocrine differentiation factors (yellow circles) induced by the trophic factors. Terminal differentiation in the post-mitotic, maturation compartment produces the metamyelocyte, band cell and the polymorphonuclear neutrophil (blue squares). A similar developmental process beginning with the multi-potential progenitor (purple) but under control of different trophic factors (e.g. erythropoietin) progresses through erythropoiesis (orange pathway)

Hematopoiesis is the term given to the renewable process of producing new blood cells from smaller numbers of undifferentiated precursor cells. In adult rodents, dogs and humans, both mature and maturing blood cells are found in the red marrow (Fig. 21.1a) primarily of flat bones (sternum, ribs, vertebrae, skull, pelvis) and in the proximal ends of long bones, whereas the yellow marrow of the


527

long bones and elsewhere is composed of adipose and stromal tissues with only rare blood cells [13]. Under certain circumstances, hematopoiesis can also occur in the spleen of rodents [14] and in the lungs of humans [9]. The red marrow contains morphologically identifiable, yet immature precursor cells of each lineage at various stages of phenotypic maturation, as well as a number of undifferentiated cells without any morphological features that could be used to assign them to particular lineages [13]. The renewable nature of blood cell production, in which large numbers of new cells are produced over the life-span of the organism without depleting red marrow cellularity, is consistent with a stem cell population that replenishes itself while producing progeny cells that terminally differentiate into the needed blood cells [15, 16]. In mouse models, some of these undifferentiated cells are capable of reconstituting blood cell production for longer than two life-spans during serial transplantation into marrow-ablated recipients [17]. Thus, the marrow appears to contain a complete system for producing new blood cells as well as for replenishing this capacity. There are cell specializations, and hierarchical relationships between precursors and progeny cells, within the “undifferentiated” population of cells, and methods other than histology are required to distinguish these different cell types. A number of subpopulations, differing in degree of lineage commitment and capacity to proliferate, can be detected, defined functionally, and quantified in vitro, when specific cytokines, cytokine combinations (or conditioned media), or cytokine sequences are used to promote cell survival in liquid cultures [18] or stimulate clonal colony formation in semi-solid media [19–22]. Because they respond to these specific cytokines by producing immature blood cells, these subpopulations are called hematopoietic progenitors, and they are named according to the blood cell lineage they produce. If they form clonal colonies containing recognizable blood cell types in semi-solid media, they are named colony-forming units (CFU) and designated by a suffix to indicate what lineage of blood cells are found in the colonies. For example, CFU-E produces hemoglobinized erythroid daughter cells, whereas CFU-G and CFU-M produce granulocytic and monocytic daughter cells. These CFUs exhibit developmental potential restricted to just one lineage. However, there are other, more immature CFUs that exhibit multi-lineage potential. For example, CFU-GM forms colonies of granulocytes and macrophages, and CFU-GEMM forms mixed colonies of granulocytes, erythrocytes, monocytes and megakaryocytes. The control of hematopoiesis lies with a complex network of cytokines [15, 23] that can act on progenitors individually, as well as synergistically and antagonistically (Fig. 21.1b). Under homeostatic conditions, normal concentrations of the blood cells are maintained by underutilized marrow capacity. Pathophysiological conditions as well as clinical trials of supra-physiological levels of recombinant cytokines have shown that there is marrow reserve that can increase production of new blood cells and thereby increase blood cell concentrations above normal levels, leading to the clinical condition known as “cytosis.” Some cytokines possess the ability to stimulate colony formation by particular CFUs, under culture conditions where there is not any spontaneous colony formation. These soluble proteins are usually known as “colony-stimulating factors” (CSF), except erythropoietin (Epo)

528

R.E. Parchment

and Interleukins-3 and -5 that were named based on reasons other than in vitro colony formation. In general, these CSFs appear to act as trophic factors for the CFUs by binding to cell surface receptors and thereby preventing a default program of apoptosis [24–27]. The surviving cells then pursue their committed developmental programs. Other factors act to modulate response to CSFs, whereas still others act synergistically on early progenitors with multi-lineage potential. Greater complexity is created by some cytokines that can act not only as trophic factors in the progenitor compartments, but also as modulators of the function of mature blood cells. Several cytokines that regulate myelopoiesis are produced by cell types found in the bone marrow and elsewhere in the body and may function both as paracrine factors from the bone marrow stromal and as endocrine factors from remote tissue locations when there is high demand for new blood cells. Epo is released from the kidney in response to hypoxia and is therefore an endocrine factor [28, 29]. Progenitor survival is tightly coupled to differentiation by CSF induction of paracrine/autocrine factors that direct the terminal differentiation program [15]. Cancer therapeutics are intended to interfere with aberrant, poorly controlled proliferation during one or more phases of the cell cycle, and therefore a priori they carry a mechanistic risk of toxicity in highly proliferative tissues, such as bone marrow, gastrointestinal epithelium and oral mucosa. Furthermore, cellular assays and preclinical models with high proliferation rates are often employed in drug discovery for the practical reasons that they are most likely to demonstrate a drug effect and that they provide study data in a much shorter time frame than models resembling clinical disease, where therapeutic outcomes may require months – years to be detectable. Using these models for lead selection further increases the bias toward experimental agents that will affect rapidly proliferating normal tissues. When coupled with the ease and frequency of monitoring blood cell counts during investigational cancer therapy, it is not surprising that adverse hematologic effects, which are usually manifestations of hematopoietic disruption in the bone marrow, are commonly reported toxicities of many classes of anticancer drugs, especially cytotoxic agents. Molecularly targeted agents that interfere with trophic factor signaling pathways (either blocking response or reducing cytokine availability), activate pro-apoptosis pathways, or cause chemical cytotoxicity via off-target effects could inhibit signal transduction and cause adverse hematological effects. The rapidly proliferating myelopoietic system (Table 21.2), including its lineage-committed progenitors, is a frequent target of cancer therapy, because even acute dosage regimens provide a sufficiently long duration of drug exposure to affect a large proportion of precursors in active cell cycle. More immature, multi-potential progenitors, as well as the hematopoietic stem cell, usually tolerate acute dosage regimens, because most of these cells are quiescent and are therefore not susceptible to the cell cycle phase-specific agents. However, the introduction of molecularly targeted agents that require protracted daily dosing brings into play potentially more serious toxicity to these multi-potential progenitors and stem cells. If the initial days’ dosing of the drug causes a loss of committed progenitors in the marrow and a cytopenic event in the blood, the multi-potential progenitors and stem cells may enter the cell cycle to replace the progenitor pool and compensate for the toxicity,

Duration of S-phase and cell cycle (h) Mitotic index

Stage of development Transit time (h) % S-Phase

14, 86 [32]

0.025

–

Myeloblast 23 85

CFU – Human: 26–43 [32–35] Mouse: 25–55 [36–38] – 0.015

–

26 [32]

Promyelocyte 26–78 65

0.011

–

Myelocyte 17–126 33

Metamyelocyte 8–108 Post-Mitotic

Band 12–96

PMN → neutrophil 0–120

Table 21.2 Proliferative properties of bone marrow progenitors, precursors and mature cells of the human granulocyte lineage Stage of development Proliferating (Amplification) compartment Maturation compartment Yield = 32x (~5 divisions) Transit time to blood = 5–7 days

Blood PMN neutrophil T½ = 7 h

21 Bone Marrow as a Critical Normal Tissue that Limits Drug Dose/Exposure 529

530

R.E. Parchment

at which point they could also become susceptible to the antiproliferative effects of the therapy. Studies of protracted daily dosing of different chemotherapeutic agents have shown that these early, multi-potential cells become sensitive to some, but not all, cytotoxic drugs when responding to acute neutropenia caused by previous doses [30, 31]. The result of this study also indicated that dosing of some anticancer drugs can be continued through the cytopenic episode without worsening the condition, but the author is not aware of any test or assay that can substitute for the empirical in vivo study for identifying drugs with this property. The process of generating granulocytes (mostly neutrophils, but also basophils and eosinophils) from committed progenitors in hematopoietic tissue is termed granulopoiesis, and it is frequently a target of toxicity of cancer therapy, both with chemotherapeutics as well as some molecularly targeted agents. Not surprisingly, the cellular hierarchy responsible for producing replacement granulocytes (Table 21.2) is highly proliferative in terms of both a rapid cell cycle time and a high growth fraction [32–38]. Coupled with the rapid rate of loss of granulocytes from the blood stream, it is apparent why disruption of granulopoiesis for even short periods of time leads to neutropenia.

21.4 Bone Marrow as a Critical Normal Tissue that Sets Maximum Human Dose/Exposure and Therefore Should Restrict Dose/Exposure Levels Used in Mouse Modeling Although severe neutropenia is considered a clinically manageable side effect of therapy, it nonetheless increases the risk of life-threatening sepsis and is therefore considered a dose-limiting toxicity [1]. The clinical dose that causes severe neutropenia is the highest possible dose that patients can tolerate without some type of hematopoietic stem cell support (transplantation or growth factor treatment), and this dose is known as the MTD. If pharmacokinetic data are available, a maximally tolerated systemic exposure (area under the plasma C × t curve) associated with the MTD will also be identified. In the absence of intentional support during therapy, the efficacy of an anticancer agent in the dose range above the MTD (exposure) will never be known, or need to be known. Efficacy at these unreachable dose levels is irrelevant to the clinical success or failure of an experimental drug. If the success or failure of an experimental drug in clinical trials is determined by its efficacy at doses/exposures that do not exceed maximum tolerated levels, why would one base preclinical development decisions on, or even want to know about, the efficacy of an experimental compound in mouse models at doses/exposures that lie above what human patients will tolerate? The answer to this question is obviously that one would not, so why hasn’t this philosophy gained traction in mouse modeling? There are significant modeling and technical challenges to identifying what the maximum tolerated human level of an experimental compound will be at the time of preclinical modeling – prior to any human clinical trial experience. Because the actual human maximum tolerated level would not be known until


531

early clinical trials are complete, the best one could do at the preclinical discovery/ development stage is to estimate or predict it. Then, this estimate could be incorporated into mouse modeling by treating only at doses (exposures) that do not exceed the predicted human maximally tolerated level. The technical challenge is establishing methodology that accurately predicts human MTD (exposure) early in the discovery setting, so this value is available by the time of mouse modeling. The switch from mice harboring mouse tumors to mice harboring human tumors was a substantial advance in modeling, because it provided an in vivo evaluation of human molecular target response with small quantities of compounds. However, the effort to “humanize” mouse models should not have stopped there, but should have continued into other aspects of animal modeling: replicating medical procedures for assessing tumor response, using anticipated clinical dosage regimens and routes, and limiting dose levels (exposures) to those predicted to be tolerated by normal human tissues. Potency against a human target has been a key factor in selection of experimental agents, and in some programs, potency against the human molecular target is further increased by iterative lead optimization studies. This focus on high human potency tends to create small molecules that are well tolerated in the mouse, either because of species-specific differences in drug binding to target, the functional role of the target, or its expression pattern in normal tissues, including bone marrow. Compounds highly selected for human potency may appear exceptionally active in mouse models, not because they are selective for tumor, but simply because they have very low potency against the mouse counterpart of the human target, or the mouse target is not critical for normal tissue function. Lead optimization of potency against a human molecular target usually does not include assessment of potency on the counterpart target of the mouse that will determine target-dependent organ toxicity, so it is not possible to distinguish tumor-selective from primate-selective agents. In this situation, it is not surprising that doses (exposures) required for efficacy in the mouse model are often not achievable in clinical trials. Since therapeutic index is a major determinant of the clinical success of an investigational agent, it is important that therapeutic index in the mouse models approximates its clinical counterpart. However, the therapeutic index in currently used xenograft mouse models is a highly artificial number: TImouse model = mouse maximum tolerated dose ÷ human efficacious dose The fundamental question is how mouse modeling can be modified to yield more accurate estimates of the actual human therapeutic index that will ultimately determine clinical success or failure: TImouse model = human maximum tolerated dose in mouse ÷ human efficacious dose in mouse Methods for estimating maximum tolerated human dose (exposure), and then limiting the evaluation of experimental compounds in mice to dose levels that do not exceed this maximum, hold promise for more accurate assessment of clinical therapeutic potential in the preclinical setting. Based on mouse modeling studies using xenografted pediatric solid tumors, the Houghton group has found that preclinical efficacy of topoisomerase I inhibitors, ixabepilone and irofulven closely resembles their clinical effectiveness only when the mouse models are treated

532

R.E. Parchment

at human tolerated dose levels, or doses that result in human tolerated systemic exposures [39–44]. However, these human dose/exposure levels were not estimate using preclinical methods, but were obtained from results of previous Phase 1 doseescalation clinical trials. This study was retrospective in nature, taking human clinical tolerability data back into the mouse models. Nonetheless, it proves the point that mouse modeling can give a more accurate picture of clinical outcome, if efficacy evaluations are restricted to doses/exposures that do not exceed those reachable in the clinic. It points to the potential payoff of incorporating human tolerated dose/ exposure into mouse modeling of experimental cancer therapeutics, but leaves open the question of how to predict/estimate human dose/exposure well before any human patients are treated. If one posits the reasonable assumption that granulopoiesis will be the clinical dose-limiting toxicity, then one can focus on the use of two emerging methods for predicting the tolerated dose (exposure) of this human tissue renewal process. These methods can be used sufficiently early in preclinical discovery/development to incorporate predicted human tolerated dose/exposure into mouse modeling of drug therapy.

21.5 Using Hematotoxicology to Limit Treatment of Mouse Models to Tolerated Human Doses/Exposures How can the anticipated human MTD/exposure for granulopoiesis be determined in the preclinical arena, prior to human clinical trials, so this information can be used to set the upper limit on dose used to treat mouse models of human cancer? If this is possible, then evaluation of experimental agents in mouse models can emphasize the efficacy of dose levels that do not exceed the predicted maximum tolerated human dose/exposure, without regard for how high a dose the mouse host will tolerate. The following sections describe two methods for determining the dose/exposure in mice that will resemble MTD/exposure in man. The reader will note that much of the discussion is focused on “dose” rather than “exposure” (area under the systemic Cp × T curve), and this is because in many situations, the promise of new agents must be judged by first-in-mouse studies prior to developing the bioanalytical methods to support pharmacokinetic studies. However, the reader should appreciate that maximum tolerated exposure of human hematopoietic tissue could also be used to limit the amount of drug administered to a tumored mouse. When using either of the methods described below, it is important to remember that the drug tolerance of human hematopoietic tissue will set an upper limit on dose/ exposure in the mouse model. Drugs requiring higher doses/exposures than the level tolerated by human marrow to be effective in mouse models would not be expected to be clinically effective. However, drugs that are active in mouse models at or below maximum tolerated human dose/exposure will not necessarily be active in the clinic, because organ systems other than the marrow could be more susceptible to toxicity and therefore limit dose/exposure to even lower levels than the


533

marrow can tolerate. Although the accuracy and value of these methods for setting relevant testing conditions in mouse models for a wide range of therapeutic and chemical classes of molecules are not yet established, they are presented as methods most likely to be useful in the immediate future.

21.5.1 Method 1: Treating Mouse Models at Maximum Tolerated Human Dose, Predicted Using CFU-GM Assays As discussed above, hematopoietic tissue contains a continuous spectrum of precursor cells within the granulocyte lineage (Table 21.2). Some precursor cells are morphologically recognizable based on the immature appearance of features found in mature granulocytes. The morphologically recognizable cells include a proliferative compartment containing myeloblasts, promyelocytes and myelocytes as well as a post-mitotic “maturation” compartment of terminal differentiating cells (metamyelocytes, band cells and mature PMN leukocytes). There are also more primitive, undifferentiated marrow precursor cells called “progenitors.” Although the progenitors do not morphologically resemble granulocytes, some nonetheless possess a commitment to produce a supply of myeloblasts, whereas others possess a bi-lineage potential to supply precursors into both granulocyte and monocyte lineages (Fig. 21.1). Not only are these lineage-specific progenitors impossible to recognize morphologically, but they are quite rate in the marrow mononuclear fraction – on the order of 0.01–0.5% – and until recently were not directly quantifiable. Insightful ex vivo studies used clonogenic assays in semi-solid media (agar, methylcellulose, etc) to identify these and other relatively rare progenitor cell populations based on their ability to form colonies containing morphologically recognizable precursor cells and sometimes even mature blood cells [19–22, 45, 46]. A cell whose original presence was proven by the formation of a colony was named a CFU, and pathways of development from more immature, multi-potential CFUs to unipotential CFUs, committed to producing daughter cells of a particular hematologic lineage, were discovered. Using different CFU assays, the cytokines that act alone to stimulate colonies of specific lineages (“colony stimulating factors,” including erythropoietin) were identified and eventually cloned to be available as recombinant proteins for use in the next generation assays. Additional cytokines were discovered and eventually cloned that modulate the CSFs in a network of positive and negative regulators of CFU proliferation, survival and apoptosis, or that are induced by CSFs to promote differentiation in a tightly linked feedback cycle [15, 16, 23]. Using enrichment approaches and complement-mediated destruction of CFUs, cluster of differentiation (CD) cell surface antigens that mark the various CFUs have been identified, and recent advances in cell sorting have made it possible to isolate highly enriched preparations of CFUs using CD markers. Not long after assays for CFUs of the granulocyte/monocyte lineage (CFU-GM, CFU-G, CFU-M) were reported, cancer researchers naturally began investigating the in vivo effects of cytotoxic chemotherapeutic agents on these cells.

534

R.E. Parchment

Chemotherapeutic agents that caused leukopenia and/or neutropenia in the mouse were shown to substantially decrease the number of myeloid CFUs in vivo prior to the onset of neutropenia [47–53]. Additional in vitro studies established that these chemotherapeutic agents were directly toxic to the CFUs outside of the host [54–61]. These data established the CFUs of the myeloid lineage, along with proliferating granulocyte precursors, as important targets of cancer chemotherapy, the inhibition of which leads to marrow hypocellularity and neutropenia. In vitro, drug exposure reduced the number of granulocyte/monocyte colonies in a concentrationdependent manner. Therefore, drug toxicity could be quantified based on the concentration that reduced colony formation by a particular amount, such as IC50 (the concentration required to reduce CFU number by 50% compared to vehicle treated cultures) (Fig. 21.2). In the early 1980s, a new nucleoside analog named fludarabine was advanced into clinical development [62, 63]. Rodent and non-rodent toxicology studies of the clinical dosage regimen indicated that myelosuppression would be dose limiting in man, and that a safe (non-toxic) first-in-human dose for Phase I clinical trials would be 1/10 the mouse LD10, equal to ~112 mg m–2 day–1 over 5 consecutive days. Despite being non-toxic in a non-rodent animal model, this dose level caused clinically significant neutropenia in the first patients treated in the Phase I trials. The failure of standard toxicology to provide a safe starting dose led to a quest to find additional toxicology tests that could flag such compounds that were singularly

+ drug:

CFUs (% Vehicle Ctl)

+ vehicle:

Drug Concentration

Fig. 21.2 Quantifying drug hematotoxicity using in vitro CFU assays. A small proportion of mononuclear cells cultivated in semi-solid medium will form clonal colonies over 1–2 weeks when stimulated with cytokines, and the particular blood cell lineage of the colony-forming unit (CFU) is identified by morphologically recognizable daughter cells found in the colony. CFUs from a particular lineage, e.g. CFU-GM of the granulocyte/monocyte lineage, can be specifically stimulated by properly selected cytokine(s). Using this in vitro method, drug toxicity against myeloid progenitors can be quantified by the concentration that reduces colony formation by a certain level, e.g. IC50 being the drug concentration that reduces the number of CFU-derived colonies by 50% (dashed line). Ideally, treatment with vehicle alone does not affect the number of clonal colonies


535

toxic in the human. It was reasonable to compare the in vitro hematotoxicity of fludarabine to human and rodent CFU-GM progenitors for a number of reasons: granulopoiesis was the target of drug toxicity, the enzyme that activates fludarabine is present at 10-times higher level in human than rodent marrow [64], and as mentioned above, CFU assays were available for all relevant species. As anticipated, human CFU-GM exhibits dramatically greater susceptible to drug toxicity than its murine counterpart (Fig. 21.3). In addition, the quantitative difference between human and mouse IC values closely approximated the difference between human and mouse tolerated dose for this myelosuppressive agent. Because the in vitro concentration–response curves for human and mouse CFU-GM to fludarabine were similarly shaped and parallel, a comparison of the IC50 values provides the most accurate quantitation of the inter-species difference in susceptibility to toxicity. But what about cases in which concentration–response curves are not parallel (differently shaped) and even cross? A clinical study correlating exposure-dependent CFU-GM toxicity in vitro with exposure-dependent clinical neutropenia in vivo revealed that the assay readout that correlates most closely with severe neutropenia is the IC90 rather than the IC50 [65]. Because severe neutropenia is a dose-limiting toxicity that defines MTD, one would anticipate

Fludarabine for Injection

CFU-GM (% vehicle)

125%

100%

75%

50%

25%

Hu 1.1 Hu 1.2 Hu 3.1 Hu 3.2 Mu 4.1 Mu 4.2 Mu 5.1 Mu 5.2 Mu 6.1 Mu 6.2

0% 0.00001 0.0001

0.001

0.01

0.1

Conc (mcg/mL)

1

10

100

Fig. 21.3 An in vitro comparison of the toxicity of the anticancer agent, fludarabine, to human and mouse CFU-GM. The drug was added directly into agarose cultures of human or mouse bone marrow mononuclear cells, which were stimulated to form colonies by species-specific rGM-CSF. After 7 (mouse, blue lines) or 14 (human, red lines) days, the mature colonies were counted, and the data expressed as the number of colonies relative to vehicle controls that were formed as a function of drug concentration. Each individual line indicates the results using an individual bone marrow preparation. Not only did the CFU-GM assay detect the much greater susceptibility of human than rodent granulopoiesis observed clinically but also the quantitative difference between the human and mouse IC50 values closely approximated the difference between human and rodent maximum tolerated dose

536

R.E. Parchment

that differences in IC90 values from CFU-GM assays across species should be similar to differences in their MTDs for myelosuppressive compounds. The correlation of the IC90 ratio from the CFU-GM assay with the MTD ratio between two species has been confirmed by a study of camptothecin-class topoisomerase I inhibitors [66] and a natural product cytotoxic compound [67]. The following equation that leads to a prediction model for human MTD in the mouse [68, 69] generally applies to IC90 values obtained from validated CFU assays of mouse and human CFU-GM developed by the US National Cancer Institute [22, 70] and the European Centre for the Validation of Alternative Methods (ECVAM) [71–74]:

IC90human MTD human = IC90mouse MTD mouse

(21.1)

Because all of the values in (21.1) except the human MTD can be obtained by running simple tests for CFU-GM toxicity and mouse MTD, and bioanalytical support is not required, it is readily apparent that the human MTD should be predictable early enough in the discovery setting to incorporate the information into mouse modeling, so to limit, or at least emphasize, efficacy testing at dose levels in the mouse that are predicted to be achievable in man. By rearranging (21.1), these simple preclinical values can be used to obtain a predicted human MTD that can set relevant doses in mouse models [68, 69]:

MTD human = MTD mouse × (IC90human ÷ IC90mouse )

(21.2)

Defining an innovative compound’s value based on efficacy at predicted human marrow MTD instead of efficacy at mouse MTD is a continuation of the philosophy of humanizing the mouse model that began with human tumor xenografting. However, it may be difficult to convince an organization to accept this new testing paradigm, especially when there are highly active compounds that exhibit their high activity only at doses above the predicted human MTD. In the new paradigm, the development of such compounds should be discontinued because they are inactive, rather than proceeding as “high priority.” Naturally, considerable justification needs to be provided for switching to this new paradigm of prioritizing discovery compounds. The reliability of CFU-GM assays to yield IC90 values was recently demonstrated by an international validation study of the analytical performance of SOP-defined mouse and human assays [73, 74]. The study demonstrated robust assay performance during SOP-driven transfers between laboratories. In addition, blinded testing by multiple laboratories yielded consistent human and mouse IC90 values for 9 of 10 chemical toxicants that inhibited CFU-GM colony formation by at least 90%, and 7 of 10 toxicants that were not toxic enough to reach 90% inhibition and needed IC90 values to be estimated. This validation study achieved one more milestone that is relevant for mouse modeling: it demonstrated the accuracy of the simple prediction model for human MTD represented by (21.2) above. “Successful prediction” was defined as a predicted human MTD within ±fourfold of the actual MTD established by Phase I clinical trials. The rationale for the fourfold


537

range was that pharmacokinetic differences between the species, which are not accommodated in the CFU-GM assay, could account for this degree of inaccuracy. When combined with the testing results from the prevalidation study, the validated CFU-GM assay correctly predicted the human MTD of 20 of 23 chemicals [74]. The development of a validated CFU-GM assay using public SOPs [75], and a companion method to predict human MTD from the data, creates the opportunity to apply this method to mouse modeling of cancer therapy. Needing only in vivo data from mouse dose-finding studies and in vitro data from human and mouse CFU-GM assays, it is possible to set a dose for mouse modeling studies that likely (87%) lies within ±fourfold of the eventual MTD that will be encountered in early clinical trials should that compound reach clinical development. Since it is possible to know the doses in the mouse that will also be achievable in the clinic, it seems reasonable to emphasize efficacy at those doses. In the drug discovery setting, it seems reasonable to err on the side of maximizing chances to find efficacious new structures, so the ±fourfold inaccuracy in the prediction model might cause the upper limit on mouse dose levels to be set at fourfold higher than the human MTD predicted by the model. Later in development, when candidate molecules are being prioritized, the requirements might be made more stringent – requiring efficacy at the predicted human MTD, or at that dose plus some fraction of it to find compounds with the widest therapeutic margin. Comparative CFU-GM testing results have just begun to be incorporated into mouse modeling of experimental cancer therapy and preclinical decision-making [76], so more time will be needed to know if this method will eventually improve the success rate of clinical development.

21.5.2 Method 2: Treating Mouse Models at Maximum Tolerated Human Dose/Exposure Determined Empirically Using NOD/SCID Mice Engrafted with Human Granulopoietic Tissue If a normal human dose-limiting tissue could be engrafted into mice, it would then be possible to determine a human MTD in that mouse strain. Using this empirical result to limit dose in efficacy studies would be a major step toward using mouse models to evaluate human therapeutic index, not just efficacy, and there has been considerable interest in creating such models (Table 21.3). Early studies took advantage of in vitro methodology for cultivating bone marrow CFUs in semi-solid media, and implanted diffusion chambers containing human hematopoietic cells embedded in agar into lethally irradiated mice to prevent immune rejection [46]. Although the mice survived only long enough to evaluate acute toxicity of chemotherapeutic agents to human CFUs in vivo [50], these studies proved that human CFUs can survive in vivo in the mouse, be susceptible to adverse drug effects, and provide a quantitative readout of drug toxicity. This method was extended to a comparison of the susceptibility of human, canine and mouse hematopoietic cells to chemotherapeutic agents [51]. Theoretically, this model could be used with syngeneic mouse tumors

SCID (C.B-17PrkdcSCID)

–(→+)

–(→+) Hi

BM, PB, CB → M,B

Fetal thymus+liver → T

Table 21.3 Mouse models with potential for evaluating human therapeutic index in the discovery setting Murine immune Co-engrafting system HSC and solid Proven human Human solid T B NK Ig hematopoietic engraftment tumor xenografts tumor Strain Not engrafted – MNCs loaded C57BL, CD-1 into diffusion chamber (lethally with semisolid agar irradiated)

Gain lymphocyte function with age Increased activity of alternative Complement pathway Modeling human therapy Low HSC engraftment rate

Mice

Very short-lived (9 day study) Modeling human therapy In vivo CFU assay Compare human and mouse Identified inter-species difference in drug tolerance

Notes Mice

[77–79]

References [46, 50, 51]

538 R.E. Parchment

NOD/LtSz-scid

–(→+)

–(→+) lo

lo

Fetal liver CD34+ w/fetal liver/thymus → M,B,T

CB CD34+ following in vitro ChemoRx

BM, CB → M,E,B

Human tumor lines (glioma, Ewings sarcoma)

Patient specimens of multiple myeloma into fetal bone

Patient specimens Autologous of breast Ca T-cell therapy of breast Ca

Pre-selected mouse marrow Serially transplantable MM Used in chemoRx studies (BCNU/ O6BG, busulfantreosulfan)

Radiosensitive Gain lymphocyte function with age C5 deficient Modeling human therapy

Short-lived Early lymphomas

Mice

Needs exogenous cytokines for human hematopoiesis

(continued)

[77–89]


NOD/LtSz-scid IL2R null

Strain

–

–

Murine immune system T B

Table 21.3 (continued)

–

NK

Ig

M,E,P,B,T Expanded CB CD34+ → M,E,B (TNFα-induced T-cell function)

Purified CB CD34+ →

G-CSF mob PBMC CD34+ → B,NK,M,T (IL7-induced T-cell phenotypes)

Proven human hematopoietic engraftment

Co-engrafting HSC and solid tumor

Single cells from patient specimens of melanoma

Canine transmissible venereal tumor Patient specimens of carcinoma and sarcoma

Human solid tumor xenografts

Poor human T-cell development Low neutrophil level Single cells from patient tumors can be cloned Engrafted tumors from patients are serially transplantable

Long-lived Radioresistant Low lymphoma rate Modeling human therapy

Mice

Notes

[90–96]

References

540 R.E. Parchment

–

NOD/LtSzRag1null

–

–

NOD/LtSz-scid- – 2m-/-

Lo

–

–

lo

CB CD34+ → M,E,B

PBMC → T, B

PBMC → B,T

CB CD34+ → N (inflammation required for human neutrophil detection)

Long-lived Later appearing follicular cell lymphomas Radioresistant

Mice

Shorter-lived than NOD/LtSz-scid High lymphoma rate Modeling human therapy High level of human engraftment Absence of MHC Class I expression Much higher levels of T cell engraftment Normalized CD4+/ CD8+ ratio

Mice

Human leukocytes/ stroma in tumor specimens survive long-term

(continued)

[77, 99]

[97, 98]


–

–

NK

–

Ig

PBMC → T

Proven human hematopoietic engraftment

Human solid tumor xenografts

Co-engrafting HSC and solid tumor Notes

Mice

Modeling human therapy Mostly CD8+ T-cells using PBMC [100]

References

CB MNC → T,B

Designed to eliminate activity of NK Cells Similar life-span to NOD/LtSz-Rag1null mice Modeling human therapy Mostly CD4+ T-cells using PBMC PB peripheral blood; CB cord blood; MNC mononuclear cell; CB cord blood; N neutrophil; T T-lymphocyte; B B-lymphocyte; M myeloid leukocyte; E erythroid; P platelet; HSC hematopoietic stem cell; NK natural killer cell

NOD/LtSz– Rag1nullPfpnull

Table 21.3 (continued) Murine immune system T B Strain

542 R.E. Parchment


543

implanted weeks before, but evaluating efficacy of human tissue-tolerated doses against murine transplantable tumors would have little therapeutic value. The advent of immunocompromised/immunodeficient mice (nu/nu and C.B-17PrkdcSCID) made it possible to xenograft human tumor tissue into subcutaneous and orthotopic sites, and study in vivo drug efficacy against human malignancies in the preclinical setting. Studies of hematopoietic engraftment in C.B-17-PrkdcSCID immunodeficient mice [77–79] found that cord blood, peripheral blood or bone marrow hematopoietic cells could reconstitute the B-lymphoid and myeloid lineages, but the presence of NK cells result in low engraftment efficiency and a lack of T-cell reconstitution, which in turn makes it necessary to provide exogenous cytokine support of human hematopoiesis. C.B-17-PrkdcSCID also exhibit elevated activity in the alternate pathway of complement. The NOD/LtSz-scid strain [80] has very low NK cell activity and is deficient in complement C5 [77–80], and therefore yields more successful and complete hematopoietic engraftment [81–83]. The mouse strain has been used to model T-cell therapy of breast cancer by transplanting surgical tumor specimens, and then subsequently treating the xenografted tumors with autologous lymphocytes from the same patient that had been processed, stored and re-stimulated prior to injection [84]. In addition to human breast cancer, clinical specimens of multiple myeloma [85], Ewing’s sarcoma and glioma human tumor lines [86, 87] and canine primary tumors [88] have all been successfully engrafted in this model. The NOD/scid strain has also been used as an in vivo assay system to quantify the loss of stem cell function following cytotoxic chemotherapy [89]. These data indicate the potential of a single designed mouse strain to host both human hematopoietic and tumor tissue, and in the case of autologous marrow and tumor engraftment, perhaps even the possibility of a reconstituted system that models an individual patient’s drug tolerance and response simultaneously, i.e. a human therapeutic index. However, the model has limited utility because the mice have a “leaky” immune system in which B- and T-cell function gradually appear with age [77–79]. Also, radiation is required to condition the mice for hematopoietic stem cell transplant, and NOD/LtSz-scid mice are radiosensitive, and prone to spontaneous lymphomas [78–80]. Limitations of the C.B-17-PrkdcSCID and nod/SCID strains mentioned above motivated the development of strains with additional defects in the immune system to eliminate residual NK cell activity, increase tolerance of radiation conditioning regimens to prepare the mice for hematopoietic stem cell transplant, and reduce or eliminate spontaneous lymphoma formation. The NOD/LtSz-scid IL2Rγ null strain, which carries a defective gamma chain of the IL-2 receptor, exhibits these properties [90, 91]. This strain affords high rates of multi-lineage hematopoietic engraftment following transplantation of cord blood- or G-CSF mobilized peripheral blood derived-CD34+ stem cells [90–92]. The rate and extent of development of a mature repertoire of human T-cells can be enhanced by treating the mice with interleukin-7 [90] or tumor necrosis factor-α [92]. Although blood cell counts of neutrophils are low, they can be induced to accumulate in areas of induced inflammation [93]. As expected due to negligible levels of NK cell activity, the NOD/LtSz-scid IL2Rγ null strain offers much improved engraftment rates of surgical specimens of human

544

R.E. Parchment

carcinoma and sarcoma, resulting in tumor xenografts that are serially transplantable [94]. In fact, the engraftment environment of these mice has supported cloning of patient melanoma specimens from unselected single cells [95]. Importantly, the tissue architecture of non-disrupted surgical specimens of primary human nonsmall cell lung cancer, including accompanying stroma, is preserved for prolonged periods of time after xenografting into these mice [96]. Tumor-derived fibroblasts and leukocytes (both CD3+ T-cells and CD138+ plasma cells) show long-term survival after implantation, and the leukocytes, including T-cells, are able to populate host organs [96]. Additional strains of mice have been produced by introducing defects into other molecular targets that play critical roles in immune function and graft rejection. Using a strategy similar to that resulting in elimination of interleukin-2 receptor γ-chain function in the NOD/LtSz-scid genetic background, genetic crosses between NOD/LtSz mice in which the gene encoding β2-microglobulin has been knocked out and NOD/LtSz-scid mice have yielded mice that are doubly homozymgous for the scid mutation and for the absence of the β2-microglobulin gene [97]. Because the resulting NOD/LtSz-scid-β2m-/- mice lack NK activity as well as impaired lymphocyte functions, they support much higher levels of human lymphomyeloid engraftment from peripheral blood mononuclear cells, including both CD4 and CD8 T-cells in appropriate ratio [98]. Another mouse model has been developed by backcrossing the null allele of the recombination-activating gene Rag1 into the NOD/LtSz background [99]. The resulting NOD/LtSz-Rag1null mice do not gain low level B- and T-cell functionality with age, because the loss of this gene function eliminates any ability to initiate V(D)J recombination of immunoglobulin or T-cell receptor genes. However, the NOD/ LtSz-Rag1null strain retains some residual NK cell activity, and in an effort to eliminate this entirely, Shultz et al. [100] have backcrossed this stain with mice harboring a mutation in the perforin structural gene, which is a key mediator of NK cell cytotoxicity. The resulting NOD/LtSz-Rag1nullPfpnull strain did not exhibit NK cell cytotoxicity and supported high level engraftment of human hematopoietic stem cells. Although NOD/LtSz-scid-β2m–/–, NOD/LtSz-Rag1null, and NOD/LtSz-Rag1null mice strains show promise as long-term models of xenografted human hematopoiesis, the author is not aware of any evaluations of these strains as hosts for human tumor xenografts. But, because these strains reliably support human multi-lineage hematopoietic engraftment, they provide a model for determining the mouse dose which is maximally tolerated by human bone marrow. For example, one can envision conducting a dose-ranging study in NOD/LtSz-scid IL2Rγnull, NOD/LtSzRag1nullPfpnull or NOD/LtSz-Rag1null mice stably engrafted with human hematopoietic tissue, and identifying the dose that causes a severe (90%) drop in human CFU-GM in the marrow, or a dose that causes a 90% drop in circulating human granulocytes if blood cell counts were sufficiently high and stable. Once identified, then additional mice harboring human tumors can be treated at that maximum tolerated human dose in the mouse to evaluate efficacy. It is exciting to think that such strains could be used to propagate both tumor and its autologous hematopoietic tissue from the same patient, so as needed for preclinical drug evaluations, these tissues could


545

be transplanted into mice for evaluation of human therapeutic index – the ratio of the toxic dose to the efficacious dose. Because hematopoietic reconstitution involves production of functional B- and T-lymphocytes, it may not be possible to study any hematopoietic tissue with any tumor line. The immune system produced by the engrafted hematopoietic tissue may recognize and reject human tumor tissue derived from different individuals than provided the hematopoietic stem cells.

21.6 Concluding Thoughts on Improving the Predictive Accuracy of Mouse Efficacy Models Using Human Hematotoxicology Data The switch from a murine to a pure human therapeutic index-based evaluation of experimental agents in mouse models of human cancer is certainly a new frontier, but is a logical next step to take in the continued quest to humanize these models and thereby improve their predictive accuracy for clinical outcome. We may need to admit that it may not be possible to design ethical prospective studies to prove that this approach is better at predicting clinical outcome than the xeno-therapeutic index approach currently in use, because it would be difficult to justify moving an experimental agent into clinical trials if it was efficacious only at the highest doses tolerated by the mouse host, but not at predicted maximum tolerated human dose/ exposure. So, we find ourselves in the difficult conundrum of philosophically needing to continue down the path of humanizing the mouse models, without ever being able to prove that this change represents an improvement in modeling and eliminates many clinical failures at an early stage in development. However, human xenograft models were adopted for selecting and prioritizing compounds without proof of superiority over models of syngeneic mouse tumors, despite the fact that the syngeneic models were responsible for the identification of many effective chemotherapeutic agents in clinical use today. The common claim that animal models do not predict clinical outcomes may relate much more to treatment regimens that only hit the clinically-relevant human dose/exposure by chance, than to the response of those human tumor models per se. Incorporating assessments of efficacy against human tumor models and toxicity against human normal tissue models into overall preclinical modeling will achieve an evaluation of human therapeutic index in the mouse and move us toward complete humanization of mouse modeling. What is clear from the fludarabine case study is that a large inter-species difference in dose tolerance does not indicate clinical success or failure, but only requires the drug to possess a favorable therapeutic index, i.e. to be effective against human malignancy despite the much lower dose tolerance of the human patient compared to the mouse models. Naturally, a favorable therapeutic index should be closely tied to clinical effectiveness. Such case studies can provide important justification and scientific foundation for incorporating human dose tolerance into mouse modeling of human malignancy. Houghton and colleagues have already proven that the efficacy of chemotherapeutic agents against human pediatric tumor xenografts in mouse models

546

R.E. Parchment

most closely correlates with their clinical effectiveness, when the dosage regimens and exposures in the mouse models replicate those tolerated by pediatric patients affected with the same malignancies [39–44]. Although these studies prove that limiting the doses/exposures used for efficacy assessments in mouse models to those tolerated by human patients improves the correlation between preclinical and clinical outcomes, it is important to note that the Houghton studies obtained the human dose/ exposure data from the results of Phase I clinical trials in patients – data that would not be available at the time of preclinical development studies. The two methods described in this chapter provide ways to estimate what this maximum tolerated human dose/exposure will be prior to the availability of any clinical results in man. Instead of using efficacy data for lead identification and optimization that comes from mouse models treated with doses/exposures that cannot be reached clinically, experimental agents could be compared based on efficacy in mouse models at their predicted human maximum doses/exposure levels. There are two frequent objections to the strategy of incorporating comparative hematoxicology data into preclinical modeling. The first is that hematopoiesis is not the only dose-limiting target of anticancer drugs; hepatic, renal and cardiac functions are all frequent targets for adverse drug effects. Of course, it would be a great advance in mouse modeling to be able to incorporate information about dose tolerance of all of these human organs into the selection of dose/exposure for evaluating efficacy. Only the most susceptible organ system is relevant to dose selection, because it will not be possible to escalate dose above the MTD of the most susceptible organ system. Other organ systems may exhibit even greater susceptibility to drug-induced toxicity than the bone marrow; therefore the maximum dose/exposure tolerated by hematopoiesis may not be reachable because of severe toxicities to these other organ systems that will occur at lower doses. The important point to recognize is that, regardless of which organ system turns out to be dose-limiting, it will not be possible to reach doses/exposures in the human that are higher than those tolerated by the granulopoietic cells because of dose-limiting neutropenia. Because it will not be possible to reach higher doses/exposures than those tolerated by human neutrophil progenitors, there is no need to exceed these levels for evaluating efficacy in mouse models. Although some compounds may be advanced into development based on efficacy at doses/exposures tolerated by human myelopoiesis that later show other organ toxicity at even lower levels, one is certainly avoiding compounds that have no chance at clinical effectiveness, i.e. those active at doses/exposures above marrow tolerated levels. The second common objection is the point that hematotoxicology is not relevant in the new age of targeted therapeutics, which are considered safer than cytotoxic chemotherapeutic agents. In response to this objection, it has been noted that myelosuppression has been identified as the clinical doselimiting toxicity of a number of targeted agents, perhaps not surprising given the fact that molecular pathways related to the biology of proliferating cell types are often targeted, and these pathways might be expected to play a role in the proliferation of normal cell types like hematopoietic progenitors and precursors. On a final note, because there are many examples of drugs where human progenitors show substantially greater susceptibility to toxicity than their mouse counterparts,


547

it would be expected that some drugs may show the opposite profile – being more toxic to mouse progenitors than their human counterpart or progenitors from other preclinical species. Since this property is not currently assessed in the discovery setting, many such compounds would be excluded from development because they would be classified as “toxic, inactive.” Only those compounds that possessed a very good therapeutic index would exhibit efficacy despite the dose reductions necessitated by the poor dose tolerance of the mouse. The rewards of developing these compounds could be great, because they will likely exhibit greater than expected clinical effectiveness as a result of being able to administer doses above those evaluated in the mouse models. The CFU-GM assay provides a way to identify those compounds for which the mouse may be the most susceptible species to bone marrow toxicity, and for which some other species may be more appropriate for efficacy studies. Acknowledgments The author would like to thank Drs. Charles K. Grieshaber, Joseph E. Tomaszewski and Adaline C. Smith for their seminal contributions to the field of in vitro hematotoxicology. The author would also like to acknowledge the critical importance of support of in vitro hematotoxicology and its application to drug and chemical safety evaluations by the Toxicology & Pharmacology Branch of the US National Cancer Institute and the European Centre for the Validation of Alternative Methods (ECVAM).

References 1. U.S. National Cancer Institute’s Common Toxicity Criteria and the Common Terminology Criteria for Adverse Events for classifying and grading severity of adverse events caused by investigational therapy during clinical trials, available at http://evs.nci.nih.gov/ftp1/CTCAE/ About.html. 2. Williams WJ. Examination of the blood. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 9–24. 3. Lichtman MA. Clinical manifestations of granulocyte and monocyte disorders. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 63–6. 4. Smith JE. Comparative hematology. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 117–25. 5. Williams WJ. Clinical manifestations of disorders of hemostasis. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 69–72. 6. Douglas SD. Morphology of lymphocytes. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 883–95. 7. Cassileth PA. Lymphocytopenia. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 955–6. 8. Golde DW. Production, distribution, and fate of neutrophils. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 759–65. 9. Paulus J-M, Aster RH. Production, distribution, life-span, and fate of platelets. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 1185–96. 10. Paulus J-M, Aster RH. Clinical evaluation of thrombokinetics. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 1196–201.

548

R.E. Parchment

11. Berlin NI. Erythrokinetics. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 395–405. 12. Erslev AJ. Production of erythrocytes. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 365–76. 13. Erslev AJ, Weiss L. Structure and function of the marrow. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 75–83. 14. Crosby WH. Structure and function of the spleen. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 89–97. 15. Sachs L. The proteins that control haemopoiesis and leukaemia. In: Bock G, Marsh J, editors. Molecular Control of Haemopoiesis (Ciba Foundation Symposium 148). Chichester, England: Wiley; 1990. p. 5–19. 16. Quesenberry PJ. Hemopoietic stem cells. In: Williams WJ, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 3rd ed. New York, NY: McGraw-Hill; 1983. p. 129–43. 17. Harrison DE. Normal production of erythrocytes by mouse marrow continuous for 73 months. Proc Natl Acad Sci USA 1973;70:3184–8. 18. Ginsburg H, Sachs L. Formation of pure suspensions of mast cells in tissue culture by differentiation of lymphoid cells from the mouse thymus. J Natl Cancer Inst. 1963;31:1–39. 19. Pluznik DH, Sachs L. The cloning of normal “mast” cells in tissue culture. J Cell Physiol 1965;66:319–24. 20. Bradley TR, Metcalf D. The growth of mouse bone marrow cells in vitro. Aust J Exp Biol Med Sci 1966;44:287–99. 21. Ichikawa Y, Pluznik DH, Sachs L. In vitro control of the development of macrophage and granulocyte colonies. Proc Natl Acad Sci USA 1966;56:488–95. 22. Volpe DA, Du DL, Pohl KP, Campbell JP, Murphy Jr MJ. Utility of human bone marrow obtained incidental to orthopedic surgery for hematopoietic clonal assays. Pathobiology 1991;59:53–6. 23. Moore MAS, Muench MO, Warren DJ, Laver J. Cytokine networks involved in the regulation of haemopoietic stem cell proliferation and differentiation. In: Bock G, Marsh J, editors. Molecular control of haemopoiesis (Ciba foundation symposium 148). Chichester, England: Wiley; 1990. p. 43–56. 24. Koury MJ, Bondurant MC. Erythropoietin retards DNA breakdown and prevents programmed death in erythroid progenitor cells. Science 1990;248:378–81. 25. Kelley LL, Koury MJ, Bondurant MC, Koury ST, Sawyer ST, Wickrema A. Survival or death of individual proerythroblasts results from differing erythropoietin sensitivities: a mechanism for controlled rates of erythrocyte production. Blood 1993;82:2340–52. 26. Askew DS, Ihle JN, Cleveland JL. Activation of apoptosis associated with enforced myc expression in myeloid progenitor cells is dominant to the suppression of apoptosis by interleukin-3 or erythropoietin. Blood 1993;82:2079–87. 27. Jacobsen FW, Stokke T, Jacobsen SE. Transforming growth factor-beta potently inhibits the viability-promoting activity of stem cell factor and other cytokines and induces apoptosis of primitive murine hematopoietic progenitor cells. Blood 1995;86:2957–66. 28. Jacobson LO, Goldwasser E, Fried W, Plzak L. Role of the kidney in erythropoiesis. Nature 1957;179:633–4. 29. Reissmann KR, Nomura T, Gunn RW, Brosius F. Erythropoietic response to anemia or erythropoietin injection in uremic rats with or without functioning renal tissue. Blood 1960;16:1411–23. 30. Keyes KA, Albella B, LoRusso PM, Bueren JA, Parchment RE. Cytotoxic chemotherapy regimens that increase dose per cycle (dose intensity) by extending daily dosing from 5 consecutive days to 28 consecutive days and beyond. Clin Cancer Res. 2000;6:2474–81. 31. Keyes KA, Segovia JC, Bueren JA, Parchment RE, Albella B. Latent hematopoietic stem cell toxicity associated with protracted drug administration. Exp Hematol. 2001;29:286–94. 32. Aglietta M, Piacibello W, Sanavio F, Stacchini A, Aprá F, Schena M, Mossetti C, Carnino F, Caligaris-Cappio F, Gavosto F. Kinetics of human hemopoietic cells after in vivo administration of granulocyte-macrophage colony-stimulating factor. J Clin Invest. 1989;83:551–7.


549

33. Aglietta M, Bussolino F, Piacibello W, Aprá F, Sanavio F, Stacchini A, Monzeglio C, Carnino F, Gavosto F. Human GM-CSF in vivo: identification of the target cells and of their kinetics of response. Int J Cell Cloning 1990;8 Suppl 1:283–90. 34. Nothdurft W, Baltschukat K, Selig C. Differences in radiation response between cells in S-phase and non-S-phase cells of the granulocyte/macrophage progenitor (GM-CFC) compartment. Int J Radiat Oncol Biol Phys. 1992;23:345–51. 35. Sköld S, Zeberg L, Gullberg U, Olofsson T. Functional dissociation between proforms and mature forms of proteinase 3, azurocidin, and granzyme B in regulation of granulopoiesis. Exp Hematol. 2002;30:689–96. 36. Tejero C, Testa NG, Hendry JH. Decline in cycling of granulocyte-macrophage colonyforming cells with increasing age in mice. Exp Hematol. 1989;17:66–7. 37. Blair A, Thomas DB. The proliferative status of haematopoietic progenitor cells in the developing murine liver and adult bone marrow. J Anat. 1998;193(Pt 3):443–7. 38. Fukuda S, Foster RG, Porter SB, Pelus LM. The antiapoptosis protein survivin is associated with cell cycle entry of normal cord blood CD34(+) cells and modulates cell cycle and proliferation of mouse hematopoietic progenitor cells. Blood 2002;100:2463–71. 39. Zamboni WC, Stewart CF, Thompson J, Santana VM, Cheshire PJ, Richmond LB, Luo X, Poquette C, Houghton JA, Houghton PJ. Relationship between topotecan systemic exposure and tumor response in human neuroblastoma xenografts. J Natl Cancer Inst. 1998;90:505–11. 40. Thompson J, Stewart CF, Houghton PJ. Animal models for studying the action of topoisomerase I targeted drugs. Biochim Biophys Acta. 1998;1400:301–19. 41. Furman WL, Stewart CF, Poquette CA, Pratt CB, Santana VM, Zamboni WC, Bowman LC, Ma MK, Hoffer FA, Meyer WH, Pappo AS, Walter AW, Houghton PJ. Direct translation of a protracted irinotecan schedule from a xenograft model to a phase I trial in children. J Clin Oncol. 1999;17:1815–24. 42. Kirstein MN, Houghton PJ, Cheshire PJ, Richmond LB, Smith AK, Hanna SK, Stewart CF. Relation between 9-aminocamptothecin systemic exposure and tumor response in human solid tumor xenografts. Clin Cancer Res. 2001;7:358–66. 43. Leggas M, Stewart CF, Woo MH, Fouladi M, Cheshire PJ, Peterson JK, Friedman HS, Billups C, Houghton PJ. Relation between Irofulven (MGI-114) systemic exposure and tumor response in human solid tumor xenografts. Clin Cancer Res. 2002;8:3000–7. 44. Peterson JK, Tucker C, Favours E, Cheshire PJ, Creech J, Billups CA, Smykla R, Lee FY, Houghton PJ. In vivo evaluation of ixabepilone (BMS247550), a novel epothilone B derivative, against pediatric cancer models. Clin Cancer Res. 2005;11:6950–8. 45. Pike BL, Robinson WA. Human bone marrow colony growth in agar-gel. J Cell Physiol. 1970;76:77–84. 46. Gordon MY, Blackett NM, Douglas ID. Colony formation by human haemopoietic precursor cells cultured in semi-solid agar in diffusion chambers. Br J Haematol. 1975;31:103–10. 47. Valeriote FA, Bruce WR, Meeker BE. Comparison of the sensitivity of normal hematopoietic and transplanted lymphoma colony-forming cells of mice to vinblastine administered in vivo. J Natl Cancer Inst. 1966;36:21–7. 48. Bruce WR, Meeker BE, Valeriote FA. Comparison of the sensitivity of normal hematopoietic and transplanted lymphoma colony-forming cells to chemotherapeutic agents administered in vivo. J Natl Cancer Inst. 1966;37:233–45. 49. Bruce WR, Meeker BE, Powers WE, Valeriote FA. Comparison of the dose- and time survival curves for normal hematopoietic and lymphoma colony-forming cells exposed to vinblastine, vincristine, arabinosylcytosine, and amethopterin. J Natl Cancer Inst. 1969;42:1015–25. 50. Gordon MV, Blackett NM. The sensitivities of human and murine hemopoietic cells exposed to cytotoxic drugs in an in vivo culture system. Cancer Res. 1976;36:2822–6. 51. Marsh JC. Comparison of the sensitivities of human, canine, and murine hematopoietic precursor cells to adriamycin and N-trifluoroacetyladriamycin-14-valerate. Cancer Res. 1979;39:360–4.

550

R.E. Parchment

52. Duménil D, Sainteny F, Frindel E. Some effects of chemotherapeutic drugs on bone marrow stem cells. II. Effect on non-Hodgkin lymphoma chemotherapy on various hemopoietic compartments of the mouse. Cancer Chemother Pharmacol. 1979;2:203–7. 53. Marsh JC. Comparative effects of methotrexate, two nonclassic folic acid antagonists, and cytarabine on hematopoietic precursor cells. Cancer Treat Rep. 1982;66:499–504. 54. Ogawa M, Bergsagel DE, McCulloch EA. Sensitivity of human and murine hemopoietic precursor cells to chemotherapeutic agents assessed in cell culture. Blood 1973;42: 851–6. 55. Ogawa M, Gale GR, Keirn SS. Effects of cis-diamminedichloroplatinum (NSC 119875) on murine and human hemopoietic precursor cells. Cancer Res. 1975;35:1398–401. 56. Spiro TE, Socquet M, Delforge A, Stryckmans P. Chemotherapeutic sensitivity of normal and leukemic hematopoietic progenitor cells to N-[4-(9-acridinylamino)-3-methoxyphenyl]methanesulfonamide, a new anticancer agent. J Natl Cancer Inst. 1981;66:615–8. 57. Koeffler HP, Yen J, Lowe L. An in vitro model for acute myelogenous leukemia chemotherapy. Cancer 1981;48:1958–63. 58. Smith SD, Trueworthy RC, Kisker SE, Noller LG, Lowman JT. In vitro sensitivity of normal granulocytic and lymphoma colonies to vinca alkaloids. Cancer 1983;51:417–22. 59. Neumann HA, Fiebig HH, Engelhardt R, Löhr GW. Cytostatic drug effects on human clonogenic tumor cells and human bone marrow progenitor cells (CFU-C) in vitro. Res Exp Med (Berl) 1985;185:51–6. 60. Du DL, Volpe DA, Grieshaber CK, Murphy Jr MJ. Comparative toxicity of fostriecin, hepsulfam and pyrazine diazohydroxide to human and murine hematopoietic progenitor cells in vitro. Invest New Drugs 1991;9:149–57. 61. Edelstein M, Corbett T, Valeriote F. In vitro screening model for the detection of agents active against myelogenous leukemia. Cancer Invest. 1998;16:303–8. 62. Casper ES, Mittelman A, Kelson D, Young CW. Phase I clinical trial of fludarabine phosphate (F-ara-AMP). Cancer Chemother Pharmacol. 1985;15:233–5. 63. Hutton JJ, Von Hoff DD, Kuhn J, Phillips J, Hersh M, Clark G. Phase I clinical investigation of 9-beta-D-arabinofuranosyl-2-fluoroadenine 5¢-monophosphate (NSC 312887), a new purine antimetabolite. Cancer Res. 1984;44:4183–6. 64. Collins JM, Grieshaber CK, Chabner BA. Pharmacologically guided phase I clinical trials based upon preclinical drug development. J Natl Cancer Inst 1990;(82):1321–6. 65. Parchment RE, Volpe DA, LoRusso PM, Erickson-Miller CL, Huang M, Murphy MJ Jr, Grieshaber CK. An in vivo-in vitro correlation of the myelotoxicity of 9-methoxypyrazoloacridine (pza, nsc-366140, pd115934) to myeloid and erythroid hematopoietic progenitors from human, murine, and canine marrow. J NatI Cancer Inst. 1994;86:273–80. 66. Erickson-Miller CL, May R, Tomaszewski J, Osborn B, Murphy Jr MJ, Page JG, Parchment RE. Ddifferential toxicity of camptothecin, topotecan and 9-aminocamptothecin to human, canine, and murine myeloid progenitors (cfu-gm) in vitro. Cancer Chemother Pharmacol. 1997;39:467–72. 67. Parent-Massin D, Parchment RE. Haematotoxicity of mycotoxins. Revue Med Veterinaire 1998;149:591–8. 68. Parchment RE. Alternative testing systems for evaluating non-carcinogenic, hematologic toxicity. Environmental Health Perspectives 1998;106 Suppl: 541–57. 69. Parchment RE, Gordon M, Grieshaber CK, Sessa C, Volpe D, Ghielmini M. Predicting hematological toxicity (myelosuppression) of cytotoxic drug therapy from in vitro tests. Annals Oncology 1998;9:357–64. 70. Du DL, Volpe DA, Murphy MJ Jr. Microcapillary clonogenic assays for human marrow hematopoietic progenitor cells. Int J Cell Cloning 1989;7:303–13 (Erratum in: Int J Cell Cloning 1989;7:395). 71. Bueren J, Deldar A, Gribaldo L, Hokland P, Meredith C, Moneta D, Mosesso P, Parchment R, Parent-Massin D, Pessina A, San Roman J, Schoeters G. Evaluation of hematotoxic effects using in vitro systems: the report and recommendations of ecvam workshop 17. Alternatives to Laboratory Animals (ATLA) 1996;24:211–31.


551

72. Gribaldo L, Casati S, Castoldi AF, Pessina A. Comparison of in vitro drug-sensitivity of human granulocyte-macrophage progenitors from two different origins: umbilical cord blood and bone marrow. Exp Hematol. 1999;27:1593–8. 73. Pessina A, Albella B, Bueren J, Brantom P, Casati S, Gribaldo L, Croera C, Gagliardi G, Foti P, Parchment R, Parent-Massin D, Sibiril Y, Schoeters G, Van Den Heuvel R. Prevalidation of a model for predicting acute neutropenia by colony forming unit-granulocyte/macrophage (cfu-gm) assay. Toxicol In Vitro 2001;15:729–40. 74. Pessina A, Albella B, Bayo M, Bueren J, Brantom P, Casati S, Croera C, Gagliardi G, Foti P, Parchment R, Parent-Massin D, Schoeters G, Sibiril Y, Van Den Heuvel R, Gribaldo L. Application of the CFU-GM assay to predict acute drug-induced neutropenia: an international blind trial to validate a prediction model for the maximum tolerated dose (MTD) of myelosuppressive xenobiotics. Toxicol Sci 2003;75:355–67. 75. Links to the INVITTOX Protocol and the ESAC Statement dated 21-March-2006 regarding The Colony Forming Unit-Granulocyte/Macrophage (CFU-GM) Assay for predicting acute neutropenia in humans can be found at http://ecvam.jrc.it/. 76. Kurtzberg LS, Battle T, Rouleau C, Bagley RG, Agata N, Yao M, Schmid S, Roth S, Crawford J, Krumbholz R, Ewesuedo R, Yu XJ, Wang F, Lavoie EJ, Teicher BA. Bone marrow and tumor cell colony-forming units and human tumor xenograft efficacy of noncamptothecin and camptothecin topoisomerase I inhibitors. Mol Cancer Ther. 2008;7:3212–22. 77. Gr einer DL, Hesselton RA, Shultz LD. SCID mouse models of human stem cell engraftment. Stem Cells 1998;16:166–77. 78. Macchiarini F, Manz MG, Palucka AK, Shultz LD. Humanized mice: are we there yet? J Exp Med. 2005;202:1307–11. 79. Manz MG, Di Santo JP. Renaissance for mouse models of human hematopoiesis and immunobiology. Nat Immunol. 2009;10:1039–42. 80. Shultz LD, Schweitzer PA, Christianson SW, Gott B, Schweitzer IB, Tennent B, McKenna S, Mobraaten L, Rajan TV, Greiner DL, et al. Multiple defects in innate and adaptive immunologic function in NOD/LtSz-scid mice. J Immunol. 1995;154:180–91. 81. Larochelle A, Vormoor J, Hanenberg H, Wang JC, Bhatia M, Lapidot T, Moritz T, Murdoch B, Xiao XL, Kato I, Williams DA, Dick JE. Identification of primitive human hematopoietic cells capable of repopulating NOD/SCID mouse bone marrow: implications for gene therapy. Nat Med. 1996;2:1329–37. 82. Melkus MW, Estes JD, Padgett-Thomas A, Gatlin J, Denton PW, Othieno FA, Wege AK, Haase AT, Garcia JV. Humanized mice mount specific adaptive and innate immune responses to EBV and TSST-1. Nat Med. 2006;12:1316–22. 83. Lan P, Tonomura N, Shimizu A, Wang S, Yang YG. Reconstitution of a functional human immune system in immunodeficient mice through combined human fetal thymus/liver and CD34+ cell transplantation. Blood 2006;108:487–92. 84. Feuerer M, Beckhove P, Bai L, Solomayer EF, Bastert G, Diel IJ, Pedain C, Oberniedermayr M, Schirrmacher V, Umansky V. Therapy of human tumors in NOD/SCID mice with patientderived reactivated memory T cells from bone marrow. Nat Med. 2001;7:452–8. 85. Huang SY, Tien HF, Su FH, Hsu SM. Nonirradiated NOD/SCID-human chimeric animal model for primary human multiple myeloma: a potential in vivo culture system. Am J Pathol. 2004;164:747–56. 86. Werner S, Mendoza A, Hilger RA, Erlacher M, Reichardt W, Lissat A, Konanz C, Uhl M, Niemeyer CM, Khanna C, Kontny U. Preclinical studies of treosulfan demonstrate potent activity in Ewing’s sarcoma. Cancer Chemother Pharmacol. 2008;62:19–31. 87. Kreklau EL, Pollok KE, Bailey BJ, Liu N, Hartwell JR, Williams DA, Erickson LC. Hematopoietic expression of O(6)-methylguanine DNA methyltransferase-P140K allows intensive treatment of human glioma xenografts with combination O(6)-benzylguanine and 1,3-bis-(2-chloroethyl)-1-nitrosourea. Mol Cancer Ther. 2003;2:1321–9. 88. Harmelin A, Pinthus JH, Katzir N, Kapon A, Volcani Y, Amariglio EN, Rehavi G. Use of a murine xenograft model for canine transmissible venereal tumor. Am J Vet Res. 2001;62: 907–11.

552

R.E. Parchment

89. Libura J, Ward M, Solecka J, Richardson C. Etoposide-initiated MLL rearrangements detected at high frequency in human primitive hematopoietic stem cells with in vitro and in vivo long-term repopulating potential. Eur J Haematol. 2008;81:185–95. 90. Shultz LD, Lyons BL, Burzenski LM, Gott B, Chen X, Chaleff S, Kotb M, Gillies SD, King M, Mangada J, Greiner DL, Handgretinger R. Human lymphoid and myeloid cell development in NOD/LtSz-scid IL2Rγnull mice engrafted with mobilized human hemopoietic stem cells. J Immunol. 2005;174:6477–89. 91. Ishikawa F, Yasukawa M, Lyons B, Yoshida S, Miyamoto T, Yoshimoto G, Watanabe T, Akashi K, Shultz LD, Harada M. Development of functional human blood and immune systems in NOD/SCID/IL2 receptor γ chainnull mice. Blood 2005;106:1565–73. 92. Giassi LJ, Pearson T, Shultz LD, Laning J, Biber K, Kraus M, Woda BA, Schmidt MR, Woodland RT, Rossini AA, Greiner DL. Expanded CD34+ human umbilical cord blood cells generate multiple lymphohematopoietic lineages in NOD-scid IL2rγnull mice. Exp Biol Med. 2008;233:997–1012. 93. Doshi M, Koyanagi M, Nakahara M, Saeki K, Saeki K, Yuo A. Identification of human neutrophils during experimentally induced inflammation in mice with transplanted CD34+ cells from human umbilical cord blood. Int J Hematol. 2006;84:231–7. 94. Fujii E, Suzuki M, Matsubara K, Watanabe M, Chen YJ, Adachi K, Ohnishi Y, Tanigawa M, Tsuchiya M, Tamaoki N. Establishment and characterization of in vivo human tumor models in the NOD/SCID/γcnull mouse. Pathol Int. 2008;58:559–67. 95. Quintana E, Shackleton M, Sabel MS, Fullen DR, Johnson TM, Morrison SJ. Efficient tumour formation by single human melanoma cells. Nature 2008;456:593–8. 96. Simpson-Abelson MR, Sonnenberg GF, Takita H, Yokota SJ, Conway Jr TF, Kelleher Jr RJ, Shultz LD, Barcos M, Bankert RB. Long-term engraftment and expansion of tumor-derived memory T cells following the implantation of non-disrupted pieces of human lung tumor into NOD-scid IL2Rγnull mice. J Immunol. 2008;180:7009–18. 97. Christianson SW, Greiner DL, Hesselton RA, Leif JH, Wagar EJ, Schweitzer IB, Rajan TV, Gott B, Roopenian DC, Shultz LD. Enhanced human CD4+ T cell engraftment in β2microglobulin-deficient NOD-scid mice. J Immunol. 1997;158:3578–86. 98. Glimm H, Eisterer W, Lee K, Cashman J, Holyoake TL, Nicolini F, Shultz LD, von Kalle C, Eaves CJ. Previously undetected human hematopoietic cell populations with short-term repopulating activity selectively engraft NOD/SCID-β2 microglobulin-null mice. J Clin Invest. 2001;107:199–206. 99. Shultz LD, Lang PA, Christianson SW, Gott B, Lyons B, Umeda S, Leiter E, Hesselton R, Wagar EJ, Leif JH, Kollet O, Lapidot T, Greiner DL. NOD/LtSz-Rag1null mice: an immunodeficient and radioresistant model for engraftment of human hematolymphoid cells, HIV infection, and adoptive transfer of NOD mouse diabetogenic T cells. J Immunol. 2000;164:2496–507. 100. Shultz LD, Banuelos S, Lyons B, Samuels R, Burzenski L, Gott B, Lang P, Leif J, Appel M, Rossini A, Greiner DL. NOD/LtSz-Rag1nullPfpnull mice: a new model system with increased levels of human peripheral leukocyte and hematopoietic stem-cell engraftment. Transplantation 2003;76:1036–42.

Chapter 22

Anesthetic Considerations for the Study of Murine Tumor Models Thies Schroeder, Siqing Shan, and Mark W. Dewhirst

22.1 Overview This chapter is to provide researchers with an overview over the requirements, challenges, and current solutions of rodent anesthesia in preclinical cancer research. Since the overwhelming majority of research is currently done in mouse models, rather than rats or other rodents, the review will focus predominantly on mouse strains. We will provide a range of hands-on protocols and suggestions on the application of the most commonly used rodent anesthesia procedures. Our target groups are both scientists that are new to the field of animal research in cancer and need help to establish SOPs, and PIs with previous expertise who wish to update or extend their knowledge about rodent anesthesia in cancer research. Our protocols are compliant with the current Duke University institutional guidelines. The chapter will cover the following sections: 1 . Rationale and requirements for animal anesthesia in cancer research 2. Guidelines of assessing depth and quality of anesthesia 3. Animal support 4. Currently used anesthesia solutions in cancer research

22.2 Rationale and Requirements for Animal Anesthesia in Cancer Research Anesthesia in research animals is carried out for two reasons: M.W. Dewhirst (*) Duke University Medical Center, Department of Radiation Oncology, Durham, NC 27710, USA e-mail: [email protected]


553

554

T. Schroeder et al.

22.2.1 Humane Reasons Anesthesia is required to eliminate pain and suffering due to invasive procedures, experimentally caused irritation, physical restraint, and certain types of euthanasia.

22.2.2 To Control Motion For most types of surgery, intravenous injections, and for imaging, it is very important to minimize motion during the procedures. Avoiding motion is also important in order to homogenize the response to treatment.

22.3 Special Requirements for Anesthesia in Cancer Research 22.3.1 Non-Survival Surgery One of the most common endpoints in cancer research studies in mice is the rate of delivery of an anticancer agent to the tumor and to organs during pharmacokinetic studies. Another typical endpoint is the effect of an anticancer agent on the representative growth rate of the tumor in a treatment group. These types of studies, typically conducted in mice, require anesthesia as a means of pain relief and physical restraint, enabling the experimentalist to surgically remove the tumor and organs in a humane manner. The requirements for anesthesia are relatively few: the onset of expressional changes of proteins (usually more than 30 min) exceeds the time needed to initiate anesthesia. In addition, considerations about post-surgical pain and distress, or carcinogenicity of the drug do not apply. Typical anesthetics used in this setting are the injectible drugs ketamine/xylazine and pentobarbital.

22.3.2 Survival Surgery Several cancer research applications require temporary anesthesia of the animal, with subsequent recovery: • orthotopic injection of tumor cells (e.g. injection of cells into the mammary fat pad), • permanent insertion of osmotic pumps (“Alzet pumps”) for studies that require steady rates of drug release, and

22 Anesthetic Considerations for the Study of Murine Tumor Models

555

• permanent insertion of devices, such as probes and sensors, or constructions for intravital microscopy. Anesthetics used for this purpose must primarily guarantee that the animal experiences no pain during the surgery and that it experiences minimal stress upon recovery. This excludes anesthetics that can cause permanent complications, such as urethane or chloralose. Typical anesthetics for these operations are pentobarbital, ketamine/xylazine, and isoflurane.

22.3.3 Functional Studies Particularly in an academic setting, anesthesia often has to meet additional demands: in order to acquire functional, mechanistic information from a treatment or diagnostic regimen, anesthesia, on one hand, has to provide efficient restraint of spontaneous movement and also must not change the physiology of the animal during data acquisition. Common applications involve functional imaging, using MicroPET, MicroMRI, MicroCT, ultrasound, or Laser Doppler measurements. These studies seek to obtain information about parameters such as tumor perfusion and blood flow, tumor cell proliferation, uptake of glucose and other metabolites, uptake of radiolabeled drugs, tumor oxygenation and hypoxia, extent of tumor necrosis, and expression of cellular markers. Unfortunately, every anesthetic strategy has inherent potential to change the physiologic condition of the tumor. At the same time, choosing anesthetic solutions based on minimal effects on animal physiology can be influenced by the accessibility of the animal during imaging (e.g. MRI). It can also be affected by the price of the setup, as in the case of the inhalable anesthetic isoflurane. Very commonly, injectables are used to maintain anesthesia during functional studies. A problem with this approach is that literally all injectibles exert some effect on the animal’s physiology, as will be discussed below. In addition, injectibles inevitably introduce variation into the study, because of the heterogeneity between individuals of even the same strain, in the reaction to the drug. This pertains in particular to differences in the depth and duration of anesthesia.

22.3.4 Controlled Delivery of Anticancer Treatment Certain anticancer treatments, such as continuously infused drugs, and radiotherapy, also require immobilization. This is because changes in tumor blood flow and oxygenation can influence the efficiency of the therapy.

556

T. Schroeder et al.

22.3.5 Assessment of Anesthetic Depth in Rodents Full anesthesia is defined as a pharmacologically induced, reversible state of amnesia (loss of memory), analgesia (freedom of pain), loss of consciousness, ataxia or chemical restraint (inability to move purposely), and loss of skeletal muscle reflexes. Among Guedel’s stages of anesthesia [1] (I: from induction of anesthesia to consciousness, II: from consciousness to onset of automatic breathing, III: from onset of automatic breathing to respiratory paralysis, VI: respiratory arrest until death), stage III, also called “surgical level of anesthesia,” is required for most applications in cancer research. Anesthetic depth in rodents must be continually monitored, for humane reasons and to allow for intervention if premature recovery from anesthesia threatens to introduce heterogeneity into the study. The following parameters can be used: • Toe pinch: A gentle pinch that does not injure the animal is sufficient. Any observed reaction to the pinch (withdrawing the paw) indicates that the animal is not sufficiently anesthetized. • Jaw retraction: This can serve as an indicator of muscle relaxation. The lower jaw is gently opened to its maximum extent. Any observed closing of the mouth is an indicator that the animal is not yet sufficiently anesthetized to do surgery. • Respiratory rate: This is a good indicator of anesthetic depth. Rapid, shallow breathing usually indicates the animal is above stage III anesthesia. Ventilation frequency can be quantified by observation (can be difficult, because fast), or using an advanced pulse oximeter that recalculates the breathing rate. • Heart rate: An increase in heart rate and/or blood pressure usually indicates a decrease in anesthetic depth. The heart rate and blood oxygenation status can be monitored using a pulse oxymeter, such as MouseOx. Using heart rate as an indicator of anesthetic depth is appropriate, but requires experience with the respective species and strain. However, even without such knowledge, a relative, consistent increase in heart rate during anesthesia can still be an indicator of anesthesia wearing off. • Eye blinking reflex: The medial edge of the cornea is very gently touched with a gauge sponge or cotton tip. Movement of the eyelids is an indication that the depth of anesthesia is not sufficient to do surgery.

22.4 Animal Support 22.4.1 Body Temperature Among all physiological consequences of prolonged anesthesia, hypothermia is the most common and significant side effect. Hypothermia during anesthesia has different reasons: on one hand, most types of anesthesia result in a drop of blood


557

pressure, and therefore for less efficient temperature homeostasis. As measurements in awake, restrained mice show, this can directly lead to a loss in tumor temperature [2]. On the other hand, loss of thermoregulation can occur due to infrared radiation to cooler surroundings, conduction to colder objects, through convection to surrounding air, and by surface evaporation of liquids [3]. Accordingly, one of the prime goals of measures of animal support is to maintain animal body temperature. This is best done using a temperature-controlled warm water-circulated rubber pad. The most sophisticated way is a heating device (electronic, water based), that is directly feedback controlled via a rectal thermometer or thermocouple. The use of heat lamps has the inherent danger of overheating the animals, and should therefore not be used, unless the light switch is controlled by a rectal thermal probe. In any case, continuous control of the animal’s body temperature using rectal monitoring is a good idea.

22.4.2 Respiration One of the most advanced developments in animal anesthesia is the control of respiratory movements and concomitant delivery of inhalable anesthesia, using an anesthesia machine and a ventilator. This setup solves numerous problems during functional studies: • The breathing rate and volume is passive, and can therefore be controlled. This homogenizes the conditions of data acquisition. This aspect is very important: less heterogeneity leads to decreased numbers of animals needed to obtain conclusive data. This saves time, effort, and money and avoids unnecessary use of animals. • Anesthetics can be delivered in an even more controlled manner and are therefore safer for the animal. • Animal physiology affecting the tumor is less influenced by the kinetics of anesthetics. • Exact control of respiratory movement is essential for cardiopulmonary imaging, e.g. during metastasis studies [4]. This solution is doubtlessly the best for functional studies. However, it is also expensive and requires skills and experience to set up the necessary facilities. The rat or mouse needs to be intubated with a matching endotracheal catheter that needs to be cut to the proper length. For rats, a 14–18 G catheter is appropriate. For mice, 20–24 G catheters work best. The insertion should be done by the help of an endoscope. Particularly for new users of this technique, it is very important to test whether the trachea, and not the esophagus, was catheterized. This is done by holding the teased end of a cotton swab in front of the opening of the catheter, and verifying its movement during in- and exhalation. It is recommended that after insertion, the catheter is sutured to the lips of the animal, to prevent it from sliding out.

558

T. Schroeder et al.

When purchasing a small animal ventilator, it has to be taken into account that MRI requires a specific setting, where long tubing has to be used, in order to keep the anesthesia machine away from the animal, or MR compatible equipment needs to be used.

22.4.3 Hydration During long periods and after fluid loss, the animals need to be rehydrated. This can be done by injecting milliliter amounts of pre-warmed sterile saline into the peritoneal cavity.

22.4.4 Analgesia It is important to alleviate pain in animals that arises due to an experiment, first and foremost for humane reasons and also to avoid stress-related bias of the experimental outcome. Pain can be most commonly defined as a stimulus that causes withdrawal and evasive action. Several pharmacological analgetic solutions are currently in use to alleviate pain in rodents, e.g. opioids such as Buprenorphine, amides such as Lidocaine and Bupivacaine, and other substances such as Carprofen and Ketoprofen. Buprenorphine is, for most applications in rats and mice, an acceptable analgesic, the dosing of which depends on the severity of the surgical procedure. In most cases, a subcutaneous dose of 0.05–0.1 mg/kg in mice and 0.01–0.05 mg/kg in rats, applied twice after 8–12 and 16–24 h, respectively, is acceptable. For further information on rodent analgesia (and anesthesia), the “Guide for the Care and Use of Laboratory Animals” by the National Research Council is an important resource to have at hand [5].

22.4.5 Inhalable Anesthetics Inhalation anesthetics are commonly used in rodent anesthesia, and are the most ideal choice in most applications in preclinical cancer research. The most widespread agents are currently halothane and isoflurane. Although most inhalable drugs are volatile at room temperature, controlled delivery of these substances to the animal is done using a calibrated vaporizer that allows to control the percentage (vol%) of the drug in the carrier gas, which is usually oxygen. The greatest advantage inhalable anesthetics offer is that their pharmacokinetics allow for the highest degree of user control, in terms of predictability and speed of adjustment, compared with other anesthetic solutions; recovery time is minimal, even after prolonged anesthesia. In addition, anesthesia can be safely delivered over several hours, if the


559

Table 22.1 Inhalable anesthetics and their metabolism

Agent

Chemical class

MAC (in % carrier gas) [6]

Halothane

Halogenated alkane

0.96 ± 0.07%

Enflurane

Halogenated methyl ethyl ether Halogenated methyl ethyl ether

1.95 ± 0.16%

Biometabolism (% recovered as metabolites) 15–20% trifluoroacetyl chloride 2.4% fluoride ion

1.34 ± 0.1%

0.17% fluoride ion

Isoflurane

animal is properly monitored and supported. However, inhalant anesthetics can also be delivered in a closed compartment, using the vapor/fluid equilibrium at room temperature. This very short-termed type of anesthesia is typically used for short operations, such as to deliver an injection to a rat. The key to the high user control of many inhalants is the low rate of cellular uptake and metabolism: this has the consequence that the tissue concentration of the drug is largely a function of the level of exposure, or in other words, the concentration of the inhalant in the alveolar space. Thus, an important measure for the anesthetic potency of an inhaled drug is the MAC (minimal alveolar concentration), which is the alveolar partial pressure at which 50% of animals, or humans, will not respond purposefully to a noxious stimulus, such as a surgical incision. An anesthetic dose can be expressed as a multiple of the MAC. Table 22.1 lists several inhalable drugs, together with their MAC and biometabolism. We will, in this chapter, limit ourselves to describing the use of isoflurane, the most important inhalable anesthetic drug in the field.

22.5 Isoflurane 22.5.1 Background About the Drug Isoflurane, a halogenated methyl ethyl ether, is probably the most commonly used inhalable anesthetic in research involving rodents. Its lipophilicity, high potency (MAC50 of 1.2%), low rate of metabolism (0.2%), and relatively low price make it a prime choice for almost any application. Isoflurane anesthesia allows for a high degree of investigator control: initiation of anesthesia normally does not take longer than 2–3 min, and upon withdrawal of anesthesia, animals recover from anesthesia in usually less than a minute. To apply isoflurane in mice, a calibrated vaporizer and flow regulator is necessary, with oxygen as a carrier gas, and a scavenger with filter, for waste gas. Isoflurane has a vasodilatory effect, therefore decreases blood pressure and increases the heart rate.

560

T. Schroeder et al.

The main advantage of using isoflurane in experimental animals is that it undergoes minimal biotransformation and is almost completely eliminated in exhaled air. This suggests that there will be little effect on liver microsomal enzymes and, hence, little interference in drug metabolism or toxicology studies. This characteristic, together with the rapid induction and recovery from anesthesia, has led to the widespread adoption of isoflurane in many research establishments.

22.5.2 Anesthetic Properties Isoflurane leads to analgesia and muscle relaxation [7]. Due to the above-mentioned low rate of metabolism, mice typically wake up within 1–2 min after cessation of isoflurane. It is important to consider the vasodilatory effect of isoflurane, which leads to a decrease in blood pressure, and a compensatory increase of the heart rate [8]. This leads to alterations of drug delivery, as well as loss of body temperature. It is easy to overdose and kill animals with isoflurane anesthesia; therefore, it is essential to observe the stability of respiration and heart rate. This can be done using oxymetric measurements and also by toe pinch and measurements of body temperature with a rectal thermistor.

22.5.3 Typical Applications Because of the high level of user control, isoflurane can be used for virtually any application, including most types of non-survival surgery. However, the proper control of anesthesia depth may require expensive machinery, such as an animal pulse oxymeter (e.g. MouseOx, STARR Life Sciences, Oakmont, PA). In addition, a calibrated anesthesia machine is necessary. Isoflurane is particularly useful in applications where long recovery times would adversely affect the feasibility of the study, such as when working with large group sizes and/or the need for multiple periods of anesthesia in short succession. Isoflurane has proven valuable in applications that require tight control of respiratory movements, such as imaging: MicroPET, MRI, and SPECT. In this case, it is applied over a ventilation machine, which controls respiratory movements of the animal. Isoflurane can be used to initiate anesthesia prior to injecting other drugs into rats, if the injection procedure has proven to involve significant stress for the animal. This will be discussed below. It is important, for reasons of occupational safety, to perform all operations where isofurane is not scavenged under a chemical fume hood.


561

22.5.4 Required Equipment • Ventilated box with grid, for initiation of anesthesia • Anesthesia machine: vaporizer with oxygen tank, calibrated flow controller, hose and nose cone, scavenger • Heating pad • Eye ointment

22.5.5 Protocol for Isoflurane in Mice and Rats 1 . Pre-warm the heating device to 37°C. 2. Initiate anesthesia by exposing mice to 2% isoflurane (rats 4%) in an anesthesia box with grid. Immediately after arrival at full anesthesia, transfer mice to 1.5% (rats 1.5–2%) isoflurane, delivered over a nose cone. The animals are placed on the heating device for the rest of the procedures. 3. Lubricate rectal thermistor with vaseline, then insert into the rectum of the animal. 4. Attach oximetry clamp to one of the hind legs. 5. Lubricate eyes. 6. After procedures, return animals to their cages. They should recover within 1–3 min.

22.5.6 Induction of Anesthesia in Rats with Isoflurane, Followed by Injectable Anesthesia The use of injectable anesthesia in rats is challenging because the initial injection, done intraperitoneally below through the hind legs, requires experience and confidence. Isoflurane can be used to induce ketamine and pentobarbital anesthesia even in the absence of an anesthesia machine. For this purpose, a sealable, transparent containment with a gridded bottom, such as a large desiccator, should be used. 1 . Prepare the respective injectable, based on the animal’s body weight. 2. Place the anesthesia container under a chemical fume hood. 3. Place 3–5 cotton sponges under the grid in the container. Soak sponges with isoflurane. Return grid on top of the sponges. 4. Transfer rat to the container and close the lid. Closely observe the animal until it is prone and loses the eye blink reflex. Immediately remove the animal from the containment, verify respiratory movement, and inject the anesthetic drug i.p. It is critical not to wait too long before removing the rat from the isoflurane,

562

T. Schroeder et al.

because there is only a very short time window from onset of isoflurane anesthesia to overdosing. After injection, continue to monitor the animal for anesthetic depth.

22.6 Injectible Anesthetics 22.6.1 Ketamine HCl with Xylazine 22.6.1.1 Background and Biochemistry Ketamine (Ketaset, Fort Dodge Animal Health, Iowa, USA) is a cyclohexamine that enjoys widespread popularity in its use in rodent anesthesia since its development 30 years ago. Parts of its popularity stem from its outstanding safety, efficiency, inexpensiveness, and vast pre-existing documentation. Ketamine causes a dosedependent CNS depression that is termed “dissociative,” meaning deep analgesia and amnesia, but not necessarily loss of consciousness. Ketamine appears to target the thalamoneocortical projection system in the brain, leading to a selective depression of neuronal function of the neocorticothalamic axis and the central nucleus of the thalamus, while it stimulates parts of the limbic system. With ketamine, ocular and pharyngeal reflexes are retained to a higher degree than with other drugs, which can lead to misinterpretation of anesthetic depth through observation of physical signs. Commercial ketamine is a racemic mixture of two optical enantiomers, R(–) and S(+), which differ in anesthetic potency and effect. Ketamine produces a range of pharmacological effects, spanning from interactions with N-methyl d-aspartate (NMDA) and non-NMDA glutamate/nitric oxide/cGMP receptors, as well as nicotinic and muscarinic cholinergic receptors and opioid receptors [9]. It also appears to interact with voltage-dependent Na+ and L-type Ca2+ channels. Ketamine is metabolized extensively by the hepatic cyt p450 system. Norketamine, the primary metabolite, is one-third to one-fifth as potent as ketamine and is excreted by the kidney. Therefore, reduced renal output can result in prolonged ketamine action. While ketamine produces dose-dependent analgesia and sedation/chemical restraint, it causes only little muscle relaxation. Therefore it is usually supplemented with an analgesic and/or sedative, such as the a-2 adrenergic agonist xylazine or a benzodiazepine, such as diazepam. The combination with xylazine provides satisfactory muscle relaxation and visceral analgesia.

22.6.1.2 Anesthetic Properties in Rodents The ratio of ketamine to xylazine is usually 1:20 to 1:10. Dosing in mice provides good results in i.p. injections at ketamine/xylazine at 100/5–10 mg/kg. Anesthesia


563

takes effect quickly, usually within 2–8 min after i.p. injection in mice and rats. Anesthesia can be refreshed at one-third of the initial dose, if necessary. Redosing should be done with ketamine only, not xylazine. Animal support must include body temperature management using, e.g. water circulated pads. The depth of anesthesia should be continually monitored by toe pinch and other methods outlined earlier. Recovery from anesthesia is also comparably quick, which is caused by rapid redistribution of ketamine from the CNS to all body tissues (primarily body fat, lung, liver, and kidney).

22.6.1.3 Impact of Ketamine on Rodent Physiology Ketamine alone induces short vasodepression, followed by a long-lasting pressor response [10]. Ketamine–xylazine mixtures generally decrease arterial blood pressure (MAP) and heart rate in both rats and mice [11, 12]. As a result, blood flow in an experimental tumor can decrease [13]. The influence of ketamine in combination with diazepam on heart rate and blood flow is less pronounced [11]. There is evidence that ketamine/xylazine has the potential of decreasing arterial pH slightly in some rat strains and in mice [11, 14]. Combined with diazepam, ketamine appears to decrease blood pH in rats as well [11]. Xylazine, alone or combined with ketamine, also has the potential to increase blood glucose levels [15–18]. This can lead to increased glucose levels and pH acidification in tumors [19]. There is also potential that the increased diuresis, caused by a-2 agonists such as ketamine, leads to an increase in hematocrit.

22.6.1.4 Anesthesia Protocol of Mice Using Ketamine/Xylazine Anesthesia Preparation • Prepare working solution: From 100 mg/ml solutions ketamine (ketaset) and xylazine, transfer 100 ml ketamine and 10 ml xylazine to a 15-ml falcon tube. Add 890 ml of saline. • Pre-heat heating pad to 37°C. • Provide several 1 ml syringes with 25-G needles.

Protocol 1. Measure out the body weight of the mouse. To apply 10/1 mg/kg ketamine/xylazine, multiply by 10 to get the amount of ketamine/xylazine working solution in microliters. For example, a 23-g mouse would require an injection of 230 ml of working solution.

564

T. Schroeder et al.

2 . Inject calculated amount i.p. Return the animal to its cage and wait until asleep. 3. Transfer animal to heating pad. Ensure surgical level anesthesia by toe pinch and corneal reflex. Continually test depth of anesthesia throughout procedures. 4. Redose, if needed, at one-third of the original volume, i.e. 80 ml.

22.7 Pentobarbital 22.7.1 Background and Biochemistry Pentobarbital, as all barbiturates, is a central nervous system depressant that is extensively used in rodent anesthesia. Although it is not a cheap drug any longer, its ongoing popularity roots in the amount of historical data and pre-existing experience with this substance at most research sites, its rapid onset of anesthesia (5–10 min in mice), and in the ease of its use through i.p. injection [20]. Pentobarbital produces deep hypnosis, but relatively poor analgesics. Barbiturates prolong the GABA induced opening of chloride channels in neurons. At anesthetic doses, this leads to the suppression of high-frequency neuronal firing via the inhibition of voltage-dependent Na+ channels [21–23]. Pentobarbital is metabolized by hepatic microsomal enzymes and hydroxylation of the 3-carbon methylbutyl side chain [24].

22.7.2 Anesthetic Properties in Rodents Pentobarbital often causes a mild phase of excitement, before and after it takes its full impact on the animal. This usually appears as an increase in breathing frequency. Another sign of onset or recovery from pentobarbital anesthesia in mice is the (unpurposeful) “stretching” of the animal, which is not an alarming sign. Depth of anesthesia can be readily assessed by observing the breathing frequency and by reaction to toe pinch. Successful pentobarbital anesthesia is also dependent on the quality of animal support: It is very important to provide adequate body temperature maintenance since pentobarbital reduces body temperature. Body temperature failure during pentobarbital anesthesia will lead to greatly extended recovery times (up to several hours), or the death of the animal. Pentobarbital can also be easily overdosed, as it has a narrow safety margin: 50 mg/kg has provided insufficient analgesia in mice, however, 60 mg/kg already shows some rate of mortality [14]. In our experience, 75–80 mg/kg provides safe, surgical level anesthesia between 25 and 45 min in female athymic nude, C57 black, and Balb/C mice, with the greatest risk of mortality due if redosing is necessary. However, the length of anesthesia can vary greatly, depending on mouse strains or genders [25, 26]. Redosing should be done in units of 1 mg/100 ml, and should not happen in intervals shorter than 20 min. Redosing should always be preceeded by adequate testing of anesthesia


565

depth, such as by toe pinch or observation of breathing rate. Overdosing usually manifests itself in shallow breathing, long breathing intervals, followed by gasping. The animal loses body temperature very quickly. In this state, which precedes death, it is often possible to rescussitate the animal, by placing it on its back and gently, but abruptly pushing the rib cage between index finger and thump, in series of 5–6. Repeat if necessary. In rats, nembutal at a dose of 40–60 mg/kg i.p. will produce surgical level anesthesia 20–50 min after injection. Redosing should be done only after careful testing for anesthetic depth, at a dose one-third of the original volume. Extended anesthesia for survival procedures has to be accompanied by application of eye ointment.

22.7.3 Physiological Impact of Pentobarbital Pentobarbital exerts strong cardiovascular effects that are less pronounced if administered i.p. than intravenously: blood flow in all organs except the kidney decreases under anesthesia in both rats and mice [12, 27, 28]. Pentobarbital also decreases the respiratory and heart rate in mice [14]. These mechanisms, together with hypothermia, may be responsible for reports of increases in tumor hypoxia [13, 29, 30]. Pentobarbital has consequently demonstrated radioprotective activity in several tumor types [31].

22.7.4 Applications The best application of pentobarbital is in both survival and nonsurvival surgery, where reliable, long anesthesia is required and free handling must not be hampered by attached nose cones and other machinery. Although pentobarbital is also a common choice where accessibility of the animal is hampered, such as during irradiation, or MRI and PET imaging, this application is less than optimal because of the known effects of pentobarbital on the animal physiology.

22.7.5 Protocol for Pentobarbital Anesthesia in Mice 22.7.5.1 Preparation • • • •

Prepare a heating pad, warm to 37°C, cover with fresh absorbent paper. Prepare 1-ml syringes w/25-G needles. Prepare 10 mg/ml nembutal, from 50 mg/ml stock, with saline. Butterfly needles 25 G, with 12¢ polyethylene catheters, and attached 1-ml syringe. Flush with 10 mg/ml nembutal.

566

T. Schroeder et al.

We recommend initiating pentobarbital anesthesia by injecting the appropriate amount of a 10 mg/ml solution intraperitoneally: 80 mg/kg equals the body weight in grams, times eight, in microliters. A 20-g mouse would therefore receive 160 ml of 10 mg/ml pentobarbital.

22.7.5.2 Protocol 1 . Initiate anesthesia i.p. as described. 2. Once the animal does not react to toe pinch, apply the butterfly needle catheter i.p. 3. Place the mouse on the heating pad. The animal should always be lying on its ventral side. 4. Frequently check retraction reflexes to toe pinch. Redose at 100 ml units via the catheter, if necessary. 5. After the procedures, allow the animal to recover on the heating pad, until sternal recumbency. Then transfer the animal back into its cage.

22.7.6 Other Injectibles Alpha-chloralose and urethane are other widespread used rodent anesthetics, often in combination with each other. Alpha-chloralose, a derivative of glucose, is a central nervous system depressant that is also used as a rodenticide and avicide. Its use is limited to non-survival procedures, because its late effect that can lead to significant animal suffering, e.g. due to its immunological effects [32]. Urethane, chemically ethyl carbamate, is an ammonia ethyl ester that produces deep anesthesia and analgesia in rodents, however, its use is limited to non-survival procedures due to its carcinogenicity and anti-proliferative effect [33]. Chloralose and urethane have been used in the following schedules in rats: anesthesia was induced with 4% isoflurane and maintained with alpha-chloralose at 30 mg/kg h [34]. In mice, induction of anesthesia was done with 3% isoflurane and then urethane (1,000 mg/kg) and chloralose (50 mg/kg) [35] were injected intravenously.

22.8 Summary In this article we presented different experimental situations that are typical for cancer research in rodents that require anesthesia. We have explained that anesthetic requirements are very different, depending on whether an organ extraction, a device implantation, or a functional study is conducted. We have listed different anesthetic solutions that are currently used, including benchtop grade protocols and guidelines.


567

As animal work in cancer research becomes more established and commercialized, we are currently experiencing an increased use of inhalable anesthetics, particularly isoflurane. The outstanding degree of user control that this drug offers makes it likely that it will take over many areas that are still covered by injectable anesthetics.

References 1. Guedel AE. Inhalation anesthesia. New York: Macmillan, 1951. 2. Pallavicini MG, Hill RP. Effect of tumor blood flow manipulations on radiation response. Int J Radiat Oncol Biol Phys. 1983;9:1321–5. 3. Suit HD, Sedlacek RS, Silver G, Dosoretz D. Pentobarbital anesthesia and the response of tumor and normal tissue in the C3Hf/sed mouse to radiation. Radiat Res. 1985:104;47–65. 4. Hedlund LW, Johnson GA. Mechanical ventilation for imaging the small animal lung. ILAR J. 2002:43;159–74. 5. The National Research Council. Guide for the care and use of laboratory animals. Washington, DC, 1996. 6. Mazze RI, Rice SA, Baden JM. Halothane, isoflurane, and enflurane MAC in pregnant and nonpregnant female and male mice and rats. Anesthesiology. 1985:62: 339–41. 7. Stokes EL, Flecknell PA,, Richardson CA. Reported analgesic and anaesthetic administration to rodents undergoing experimental surgical procedures. Lab Anim. 2009:43;149–54. 8. Clark SC, MacCannell KL. Vascular responses to anaesthetic agents. Can Anaesth Soc J. 1975:22;20–33. 9. Kohrs R, Durieux ME. Ketamine: teaching an old drug new tricks. Anesth Analg. 1998:87;1186–93. 10. Altura BM, Altura BT, Carella A. Effects of ketamine on vascular smooth muscle function. Br J Pharmacol. 1980:70;257–67. 11. Wixson SK, White WJ, Hughes HC, Jr, Lang CM, Marshall WK. The effects of pentobarbital, fentanyl–droperidol, ketamine–xylazine and ketamine–diazepam on arterial blood pH, blood gases, mean arterial blood pressure and heart rate in adult male rats. Lab Anim Sci. 1987:37;736–42. 12. Yang XP, et al. Echocardiographic assessment of cardiac function in conscious and anesthetized mice. Am J Physiol. 1999:277;H1967–74. 13. Menke, H, Vaupel, P. Effect of injectable or inhalational anesthetics and of neuroleptic, neuroleptanalgesic, and sedative agents on tumor blood flow. Radiat Res. 1988:114;64–76. 14. Erhardt, W, Hebestedt, A, Aschenbrenner, G, Pichotka, B, Blumel, G. A comparative study with various anesthetics in mice (pentobarbitone, ketamine–xylazine, carfentanyl–etomidate). Res Exp Med (Berl). 1984:184;159–69. 15. Aynsley-Green, A, Biebuyck JF, Alberti KG. Anaesthesia and insulin secretion: the effects of diethyl ether, halothane, pentobarbitone sodium and ketamine hydrochloride on intravenous glucose tolerance and insulin secretion in the rat. Diabetologia. 1973:9;274–81. 16. Hsu WH, Hembrough FB. Intravenous glucose tolerance test in cats: influenced by acetylpromazine, ketamine, morphine, thiopental, and xylazine. Am J Vet Res. 1982:43;2060–2061. 17. Reyes Toso CF, Linares LM, Rodriguez RR. Blood sugar concentrations during ketamine or pentobarbitone anesthesia in rats with or without alpha and beta adrenergic blockade. Medicina (B Aires). 1995:55;311–6. 18. Kawai, N, Keep RF, Betz AL. Hyperglycemia and the vascular effects of cerebral ischemia. Stroke. 1997:28;149–54. 19. Pavlovic, M, Wroblewski, K, Manevich, Y, Kim, S, Biaglow JE. The importance of choice of anaesthetics in studying radiation effects in the 9L rat glioma. Br J Cancer Suppl. 1996:27;S222–5. 20. Wixson SK. Anesthesia and analgesia in rodents. In: Kohn DF WS WJ, White GJ, Benson, editors. Anesthesia and analgesia in laboratory animals, pp 165–203 San Diego: Academic Press, 1997.

568

T. Schroeder et al.

21. Macdonald RL, McLean MJ. Anticonvulsant drugs: mechanisms of action. Adv Neurol. 1986:44;713–6. 22. Olsen RW. GABA–drug interactions. Prog Drug Res. 1987:31;223–41. 23. Saunders PA, Ho IK. Barbiturates and the GABAA receptor complex. Prog Drug Res. 1990:34;261–86. 24. Freudenthal RI, Carroll FI. Metabolism of certain commonly used barbiturates. Drug Metab Rev. 1973:2;265–78. 25. Lovell DP. Variation in pentobarbitone sleeping time in mice. 2. Variables affecting test results. Lab Anim. 1986:20;91–6. 26. Lovell DP. Variation in pentobarbitone sleeping time in mice. 1. Strain and sex differences. Lab Anim. 1986:20;85–90. 27. Kawaue Y, Iriuchijima J. Changes in cardiac output and peripheral flows on pentobarbital anesthesia in the rat. Jpn J Physiol. 1984:34;283–94. 28. Tuma RF, Irion GL, Vasthare US, Heinel LA. Age-related changes in regional blood flow in the rat. Am J Physiol. 1985:249;H485–91. 29. Rockwell S, Moulder JE, Martin DF. Tumor-to-tumor variability in the hypoxic fractions of experimental rodent tumors. Radiother Oncol. 1984:2;57–64. 30. Moulder JE, Rockwell S. Hypoxic fractions of solid tumors: experimental techniques, methods of analysis, and a survey of existing data. Int J Radiat Oncol Biol Phys. 1984:10;695–712. 31. Denekamp, J, Terry NH, Sheldon PW, Chu AM. The effect of pentobarbital anaesthesia on the radiosensitivity of four mouse tumours. Int J Radiat Biol Relat Stud Phys Chem Med. 1979:35;277–280. 32. Silverman, J, Muir WW, 3rd. A review of laboratory animal anesthesia with chloral hydrate and chloralose. Lab Anim Sci. 1993:43;210–216. 33. Field KJ, Lang CM. Hazards of urethane (ethyl carbamate): a review of the literature. Lab Anim. 1988:22;255–262. 34. Luckl, J, Keating, J, Greenberg JH. Alpha-chloralose is a suitable anesthetic for chronic focal cerebral ischemia studies in the rat: a comparative study. Brain Res. 2008:1191;157–167. 35. Kazerani HR, Furman BL. Comparison of urethane/chloralose and pentobarbitone anaesthesia for examining effects of bacterial lipopolysaccharide in mice. Fundam Clin Pharmacol. 2006:20;379–384.

Part IX

Experimental Methods and Endpoints

Chapter 23

Preclinical Tumor Response End Points Beverly A. Teicher

Abstract The first in vivo tumor models were developed in the mid-1960s. These models were mouse leukemia models grown as ascites. The growth pattern was like that of bacteria in vivo and therefore it was possible to apply similar mathematics of growth and response to these tumors as had been worked out for bacteria. Since the development of the murine leukemia models, investigators have devoted a large effort to modeling solid tumors in mice. There are now a variety of models including syngeneic mouse tumors and human tumor xenografts grown as subcutaneous nodules, syngeneic mouse tumors and human tumor xenografts grown in orthotopic sites, models of disseminated disease, ‘‘labeled’’ tumor models that can be visualized using varied technologies, and transgenic tumor models. The value of these models depends upon the application of rigorous experimental design and data analysis. The endpoints used can be in situ or excision. Each of these has advantages and disadvantages to the ‘‘drug hunter’’ searching for improved treatments. Keywords Tumor growth delay • Xenografts • Isobolograms • Drug distribution • Enzastaurin • Intratumoral vessels

23.1 Introduction The field of cancer research only recently came to the forefront of human scientific endeavor and took advantage of experience gained in studying other disease processes. Over many years prior to the formal investigation of malignant disease, researchers had worked out scientific methodology and recognized the importance of laboratory models for infectious diseases, allowing rapid progress in antibacterial

B.A. Teicher (*) Genzyme Corporation, 49 New York Avenue, Framington, MA 01701-9322, USA e-mail: [email protected] B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_23, © Springer Science+Business Media, LLC 2011

571

572

B.A. Teicher

drug development. Cancer research also benefited from the early research of the 1950s and 1960s, which took a very orderly and rigorously scientific approach to the development of in vivo models and to the development of the most informative endpoints available from experiments suing in vivo tumor models.

23.2 Ascites Tumors The science of preclinical modeling of anticancer therapies began in the 1950s, but the establishment of guidelines for experimental quality and end point rigor can be attributed in large part to the group headed by Howard Skipper at the KetteringMeyer Laboratory affiliated with Sloan-Kettering Institute, Southern Research Institute in Birmingham, Alabama. In the mid-1960s, this group published a series of reports on the criteria of “curability” and on the kinetic behavior of leukemia cells in animals and the effects of anticancer chemotherapy. The principles put forth in these reports were derived directly from the behavior of bacterial-cell populations exposed to antibacterial agents, and were based upon experimental findings in mice bearing intraperitoneally implanted L1210 or P388 leukemia [1–15]. The initial assumptions were: (1) one living leukemia cell can be lethal to the host. Therefore, to cure experimental leukemia, it is necessary to kill every leukemia cell in the animal, regardless of the number, anatomic distribution or metabolic heterogeneity, with treatment that spares the host. (2) The percentage, not the absolute number, of in vivo leukemic cell populations of various sizes killed by a given dose of a given anti-leukemic drug is reasonably constant. This phenomenon of a constant fractional (or percentage) drug-kill of a cell population, regardless of the population size, has been repeatedly observed and may be a general phenomenon. (3) The percentage of experimental leukemic cell populations of any size killed by single dose treatment of drug to the host is directly proportional to the dose level of the drug (the higher the dose, the higher than the percentage of cells killed). Thus, it is obviously necessary to kill leukemic cells faster than they are replaced by proliferation of the cells surviving the therapy if a “cure” is to be attempted [10–12]. The exponential killing of cells by drugs with time (mathematically equivalent to “a constant percentage kill of leukemic cells regardless of number”) was observed in bacterial cell populations around 1900, and has been investigated with many antibacterial agents [16–20]. Studies with bacterial cells exposed to anticancer agents confirmed that the first-order kinetics of cell kill by anticancer agents was similar to that of antibacterial agents [12]. The hypothesis that “the percentage, not the absolute number, of cells in populations of widely varying sizes killed by a given dose of a given anticancer drug is reasonably constant” was studied intensively and found, for the most part, to be valid [12]. Skipper and the group at the Kettering-Meyer Laboratory developed the mouse L1210 leukemia [21], as well as the mouse P388 leukemia [22], into highly sensitive and reasonably quantitative in vivo bioassay systems to study anatomic distribution, rate of leukemic cell proliferation, and effects of chemotherapy in tumor-bearing

23 Preclinical Tumor Response End Points

573

mice [14]. These studies were based on the notion that the principal mechanism of drug-induced increase in host life-span was leukemic cell kill, and not “inhibition of growth” of the leukemic cell population [23–26]. Leukemic cells that gained access to the brain and other areas of the central nervous system (CNS) were not markedly affected by certain peripherally administered anti-leukemic drugs. Therefore, if leukemic cells were present in the CNS at the time of treatment initiation, a drug that passes the blood–brain barrier would be required to achieve “cure” [27, 28]. The observation that there was a close relationship between the number of L1210 leukemic cells inoculated into BDF1 mice and the life-span of the mice was critical (Fig. 23.1) [23]. Thus, it was possible to estimate the in vivo doubling (or generation) time and the approximate lethal number of L1210 leukemic cells when L1210 leukemia cells were inoculated into the mice by various routes. By the intraperitoneal route (ip), the average doubling time for the leukemic cells was 0.55 days, and the lethal number of leukemic cells was approximately 1.5 billion [23]. When the L1210 leukemic cells were implanted intravenously (iv) or intracerebrally (ic), the doubling time and lethal number of cells was lower. This knowledge was used to develop an in vivo bioassay by ip implantation of unknown numbers of L1210

IP inoculation: generation time = 0.55 days lethal cell number = 1.5 x 109 IV inoculation: generation time= 0.43 days

IC inoculation generation time = 0.46 days

NUMBER OF LL1210 LEUKE EMIA CELLS IMPLANTED

Mean Survival Time of Mice Inoculated with L1210 Leukemia 1e+6

1e+5

1e+4

1e+3

1e+2

IP IMPLANT IV IMPLANT IC IMPLANT

1e+1

1e+0

0

2

4

6

8

10

12

14

16

18

20

MEAN SURVIVAL TIME, Days

Fig. 23.1 Mean survival time of mice inoculated with various numbers of murine L1210 leukemia cells injected intraperitoneally (ip), intravenously (iv) or intracranially (ic). These data form the basis for the in vivo bioassay method for determining the number of L1210 cells surviving after treatment of L1210 tumor-bearing mice with therapy. From these survival curves, it was determined that from: (1) ip inoculation the L1210 cell-generation time is 0.55 days; the lethal number of L1210 cells is 1.5 × 109; (2) iv inoculation the L1210 cell-generation time is 0.43 days, and (3) ic inoculation the L1210 cell-generation time is 0.46 days (adapted from Refs. [14, 23–31])

574

B.A. Teicher

leukemic cells isolated from various tissues of chemotherapy-treated L1210-bearing animals into fresh hosts and using the survival time of those mice to estimate the tumor cell killing by the chemotherapy. The estimated experimental error in this bioassay procedure was ±1 log of tumor cells. The method gave an order-of-magnitude estimate of the number of leukemic cells in various tissues and was sensitive to small, absolute numbers of viable L1210 leukemia cells [14]. Antitumor activity endpoints used in these mouse ascitic leukemia models were the percent mean or median increase in life-span (% ILS), net log10 cell kill, and number of long-term survivors [29, 30]. The percent mean or median increase in life-span (% ILS) is the ratio of the survival time of the treated mice (days) compared with the survival time of the untreated control mice (days). Calculation of net log10 cell kill is made from the tumor doubling time determined from an internal tumor titration consisting of implants from serial 10-fold dilutions (Fig. 23.1) [31]. Long-term survivors are excluded from the calculations of %ILS and net log10 tumor cell kill. To assess net log10 tumor cell kill at the end of treatment, the survival time (days) difference between treated and control groups is adjusted to account for re-growth of tumor cell populations that may occur between individual doses [32]. The net log10 cell kill is calculated as:

Net log10 cell kill = [ (T - C ) - (duration of treatment in days )] / 3.32 ´ Td , where (T – C) is the difference in the median day of death between the treated (T) and the control (C) groups, 3.32 is the number of doublings required for a population to increase 1 log10 unit, and Td is the mean tumor doubling time (days) calculated from a log-linear least squares fit of the implant sizes and the median days of death of the titration groups and accounts for any repopulation of the tumor during or after treatment (Fig. 23.1).

23.3 Solid Tumors As solid tumor models were developed, the response endpoints devised were tumor growth delay or tumor control of the implanted tumor. These assays require that drugs be administered at doses producing tolerable normal tissue toxicity, so that the response of the tumor to therapy can be monitored for a relatively long time. Treatment with test compounds can be initiated either prior to tumor development or after a tumor nodule has developed. If treatment begins the day after or on the day of tumor cell implant, the experiment is a tumor growth inhibition study. If treatment begins when an established tumor nodule (50–200 mm3) is present, the experiment is a tumor growth delay study. Activity in a tumor growth delay study is more persuasive than activity in a tumor growth inhibition study, and is a better model of clinical disease. Historically, in primary screening experiments in mouse solid tumor models, the tumor volumes in the treated and control groups were measured with calipers only


575

once, usually when the control tumors reached approximately 1 cm3 in volume (1 g by weight), at which time all the mice were sacrificed. Alternatively, the mice in all the groups were sacrificed when the tumors of the untreated or vehicle-treated control group reached approximately 1 cm3 in volume, and the tumors were excised and weighed. This traditional protocol design provided no kinetic data regarding tumor growth and response [33]. A more informative experimental design includes tumor volume measurements and body weight measurements of individual mice twice per week for the duration of the experiment. This experimental design elucidates the growth pattern of the tumor in control mice as well as elucidating the effect of the drug on the tumor growth pattern [33–35]. Tumor volumes are usually estimated from measurements of two diameters of the nodule:

Tumor volume (mm3 ) = (longer diameter ´ shorter diameter 2 ) ´ 0.5 where the diameters are the tumor length and width in mm, usually measured with calipers, respectively. Experiments in which tumor volume measurements are made over a relatively long period of time, until the tumors reach a volume of 1.5–2 cm3, allow the calculation of tumor growth delay and percent T/C at multiple time points and the tumor volume doubling time (Fig. 23.2). Tumor growth delay is the difference in days for treated versus control tumors to reach a specified, usually between 500 mm3 and 2 cm3. Therefore, tumor growth delay is simply T – C in days. T is the mean or median time (in days) required for the treatment group tumors to reach a predetermined size and C is the mean or median time (in days) for the control group tumors to reach the same size. Animals that are tumor-free at the time of the determination of the tumor growth delay are excluded from the calculation. Tumor growth delay may be the most important estimate of antitumor effectiveness, because it mimics most closely clinical endpoints and requires observation of the mice through the time of disease progression. TUMOR GROWTH DELAY 10,000

CONTROL #1

#2

1,000 500 TUMOR VOLUME mm 3

100

10 TIME,DAYS

Fig. 23.2 Schematic representation of the calculation of tumor growth delay in days is shown. Tumor growth delay is usually calculated at a specified tumor volume such as when the treated tumors reach 500, 1,000 or 1,500 mm3 in volume

576

B.A. Teicher

The importance of the duration of observation of tumors response is illustrated in Fig. 23.3. Mice bearing the mouse CT-26 colon carcinoma were treated with 10 compounds in a developing structure–activity relationship (SAR) in an effort to provide guidance regarding the relative antitumor activity of the compounds. When the experiment was terminated at 27 days the resolution amongst the compounds is very limited; however, when the tumor response observation time was extended to 48 days differentiation amongst the compounds became clear. If tumor growth is log-linear through the treatment and response phase of the experiment, the data can be used to calculate the log cell kill. The group at the Kettering-Meyer Laboratories used techniques developed in the mouse leukemia models for obtaining order-of-magnitude estimates of the absolute number, percent of

Tumor Volume (mm3)

2000 Untreated Vehicle cmpd 1 cmpd 2 cmpd 3 cmpd 4 cmpd 5 cmpd 6 cmpd 7 cmpd 8 cmpd 9 cmpd 10

1500

1000

500

0

0

3

6

9

12

15

18

21

24

27

30

4000

Untreated Vehicle cmpd 1 cmpd 2 cmpd 3 cmpd 4 cmpd 5 cmpd 6 cmpd 7 cmpd 8 cmpd 9 cmpd 10

Tumor Volume (mm3)

3500 3000 2500 2000 1500 1000 500 0

0

3

6

9

12

15

18

21

24

27

30

33

36

39

42

45

48

Day

Fig. 23.3 Growth curves for the murine CT-26 colon carcinoma treated with 10 compounds from a developing structure–activity relationship (SAR). In the upper panel, the observation of tumor response was terminated at 27 days which provides very little guidance regarding the efficacy of the compounds. In the lower panel, the observation of tumor response was extended to 48 days allowing clear delineation of the efficacy of the compounds to be discerned and maximizing the information obtained from the experiment


577

viable cancer cells remaining and log10 cell kill methodology for selected experimental solid tumors after a single dose of a drug [13]. The assumptions were that: (1) the mass of the tumor is in direct proportion to the number of malignant cells in the mass. (2) The cells killed by the cytotoxic agent immediately become nonviable. (3) The cells which remain viable despite treatment begin to grow again after a relatively short lag and proliferate at the same rate as tumor cells in untreated control animals. These assumptions appeared to be valid for two of the three tumors studied in the initial report, which included the hamster Plasmacytoma 1 tumor, the mouse Sarcoma 180, and the mouse Adenocarcinoma 755 tumor. Because both Plasmacytoma 1 and Sarcoma 180 control tumors grow logarithmically during the treatment period, a method for estimating cell killing could be applied to these tumors. However, many solid tumors do not maintain log-linear growth and this method cannot be applied to those tumors without modification. Furthermore, many cytotoxic therapies such as radiation therapy do not kill cells destined to die promptly but kill cells over several generations of proliferation and this methodology will not be accurate. Slopes were derived from tumor growth curves for untreated and treated groups and first order rate constants for tumor growth were derived to allow determination of the fraction of tumor cells killed or the fraction of viable cells remaining after the treatment. For subcutaneously growing tumors, the log10 cell kill is calculated as:

Log10 cell kill total (gross) = [T - C value in days / 3.32 ´ Td ] where T – C is the tumor growth delay and Td is the tumor volume doubling time (in days) of the control tumors in exponential growth over a volume range from approximately 100 mm3 to 1 cm3. The conversion of the T – C values to log10 cell kill is possible if the tumor maintains a log-linear growth pattern and if the Td of the tumors re-growing post-treatment approximates the Td values of the tumors in control group. The net log10 cell kill is derived by subtraction of the duration of the treatment period from the T – C value and then dividing by 3.32 × Td [36, 37]. Similarly, Norton and Simon proposed that cytotoxic agents were only active toward the growing tumor cell fraction. With the Norton–Simon model, it is possible to account the dose-dependence of a compound on tumor volume as a function of time [38, 39]. The Norton–Simon hypothesis predicts that the rate of tumor regression increases proportionally with increasing level of the drug. For most, solid tumors volume behavior is not a reliable endpoint with respect to tumor cell kill [4, 40–44]. A drug-induced regression in tumor mass of no more than 50% may represent a 99.99% reduction in clonogenic cells in a solid tumor mass. Many widely used human tumor xenograft models do not conform to the requirement of log-linear growth and many anticancer therapies do not kill promptly with little lag prior to the resumption of log-linear growth in the treated groups (Fig. 23.4). Despite this limitation, the NCI was able to correlate response in human tumor xenograft models with the activity of compounds in phase II clinical trial if the compound was an active anticancer agent in at least 33% of the models tested [45].

B.A. Teicher

MEAN TUMOR VOLUME, mm3


578

1000

100

CONTROL TUMORS DRUG DOSE #1 DRUGDOSE#2 DRUG DOSE #3

10

1000

100 CONTROL TUMORS DRUG DOSE #1 DRUGDOSE#2 DRUG DOSE #3

10 0

5

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70

10 15 20 25 30 35 40 45 50

DAYS POST TUMOR CELL IMPLANTATION

MEAN TUMORVOLUME, mm3



1000

1000

100

CONTROL TUMORS DRUG DOSE #1 DRUGDOSE#2 DRUG DOSE #3

100

CONTROL TUMORS DRUGDOSE#1 DRUG DOSE #2 DRUG DOSE #3

10

10 0

5

10

15

20

25

30

35

40


0

5

10 15 20 25 30 35 40 45 50


Fig. 23.4 Top: Growth curves for the human SW-2 small cell lung carcinoma and for the human Calu-6 non-small cell lung carcinoma grown as subcutaneous xenograft tumors in nude mice. Although small regions of the tumor growth curves may approach log-linear growth, marked deviations from log-linearity are clear. Bottom: Growth curves from two different studies for the human LNCaP prostate carcinoma grown as a subcutaneous xenograft tumor in nude mice. The control tumors approximate log-linear growth. The drug in the left-hand panel alters tumor growth variously depending upon dose. After the drug treatment in the right-hand panel, the tumors regain log-linear growth parallel to the control tumors after recovery from the regression phase

23.4 Combination Treatments That combination therapy regimens would be required to effectively treat malignant disease was realized in the very early scientific history of cancer research and rationale means of selecting drugs for combination were described. Sequential inhibition is the action of two or more inhibitors on different enzymes of a multi-enzyme pathway; concurrent blockade is the simultaneous inhibition by two or more agents


579

of alternative pathways to the same critical end-product and complementary inhibition is the combination of two or more agents that inhibit different loci involved in a critical metabolic process (Fig. 23.5) [46–48]. As it has become evident that “normal” cells involved in the malignant disease process are valid targets for therapeutic attack, horizontal combinations inhibiting different pathways in two or more cell types involved in malignant disease and vertical combinations inhibiting the same or related pathways in two or more cell types involved in malignant disease have been described (Fig. 23.5) [49, 50]. In the study of multimodality therapy or combination chemotherapy, it is important to determine whether the combined effects of two agents are additive or whether the combination is substantially different than the sum of the parts [51, 52]. Several methods have been developed to examine Enz1 A X B I1

Sequntial Blckade (Potter 1951)

I2

I1 Enz1 Enz2

A Conurrent Blockade (Elion 1954)

X

Synthesis

C

X

I2

B

Compimentary Blockad (Sartorelli 1964)

Enz2 X C

I1

I2 DNA RNA Protein

Horizontal Combinations (Kaelin) 2003; Sosman 2007)

VEGF

PDG

TGFα

VEGF Endothelial cells

PDGFR Pericyes

EGFR Tumor Cells

T

T

T

Drug B

Drug A

DrugC

VEGF

Verical Combinations

T

DrugA or

Growth Factor Receptor

Drug B

T

Growth Factor

T

Cirtical Molecule

Drug C

Fig. 23.5 Several rationale methods for developing anticancer drug combinations are shown. Sequential, concurrent and complimentary combinations were developed in the 1950s and 1960s [46–48]. Horizontal and vertical combinations were developed as cells other than the malignant cell were envisioned to be therapeutic targets [49, 50]

580

B.A. Teicher

combination therapy regimens for additive, sub-additive or greater than additive effects [53–61]. Among the most widely applicable analysis methods for data from combination regimens is the Combination Index method which requires that the two agents be combined in a constant ratio [55, 59, 61]. The uniform measures and response surface models allow more parameters to be taken into account [53, 56, 57]. Here examples will use isobologram analysis applied to in vivo tumor studies [54, 58]. Conceptual foundations for this analysis were based on the construction of an envelope of additivity in an isoeffect plot (isobologram). This approach provides a rigorous basis for defining regions of additivity, supra-additivity, sub-additivity, and protection [62]. This method of analysis is based on a clear conceptual formulation of the ways that two agents can show additivity [62]. For a selected level of effect (survival) on a log scale, the dose of Agent A to produce this effect is determined from a survival curve. A lower dose of Agent A is then selected, the difference in effect from the isoeffect level is determined and the dose of Agent B needed to make up this difference is determined from a survival curve for Agent B (Fig. 23.6a). For example, 3 mg of Agent A may be needed to produce 0.1% survival (3 logs of kill), the selected isoeffect. A dose of 2.5 mg Agent B produces 1.0% survival (2 logs of kill). The Mode I isoeffect point for Agent B would be the level of Agent B needed to produce 1 log of kill, to result in the same overall effect of 3 logs of kill. In this instance, 4 mg of Agent B are needed to produce 1 log of kill. Mode II additivity is conceptually more complex, and corresponds to the concepts of additivity discussed by Berenbaum [63]. For any given level of effect, the dose of Agent A needed to produce this effect is determined from the survival relationship. The isoeffect dose of Agent B is calculated as the amount of Agent B needed to produce the given effect, determined from the survival relationship. The isoeffect dose of Agent B is calculated as the amount of Agent B needed to produce the given effect, starting at the level of effect produced by Agent A (Fig. 23.6b). For example, 3 mg of Agent A may be needed to produce 0.1% survival (3 logs of kill). A dose of 2.5 mg of Agent A produces 1.0% survival (2 logs of kill). A dose of 6 mg of Agent B is needed to produce 3 logs of kill and 2 logs of kill are obtained with Agent B at 5 mg. Thus, the Mode II isoeffect point with Agent A at 2.5 mg is equal to the amount of Agent B needed to take Agent B from 2 logs of kill to 3 logs of kill (6 mg – 5 mg = 1 mg). This can be conceptualized by noting that Agent A should produce 2 logs of kill and is equal to 5 mg of Agent B. If Agent A + Agent B are identical in their mode of action, then 1 mg more of Agent B should be equivalent in effect to 6 mg of Agent B. Graphically, on a linear dose scale, Mode II additivity is defined as the straight line connecting the effective dose of Agent A alone and the effective dose of Agent B alone. This relationship is also described by the equation: Dose of A/ Ae + Dose of B/Be = 1 where Ae and Be are the doses of Agent A and Agent B, respectively, needed to produce the selected effect.


581

a ED50 Agent B

1

Greater-than-additive/ Synergistic

0.9 0.8

Agent B

0.7 0.6 Envelope of Additivity

0.5 0.4 0.3 0.2

Sub-Additivity

0.1 0

0

0.1 0.2

0.3 0.4

0.5 0.6

0.7 0.8

Agent A

b Surviving Fraction

1

Mode I A -> B

1

Mode II A -> B

Agent B

Agent A

Agent B

Agent A 0.01

0.01

0.001

Dose

D Dose

A

B

Mode II B -> A

0.1

Agent A

0.01

1 ED50 Agent A

Agent B 0.1

0.1

1

0.9

0.001

0.001

Dose

Dose

A

Dose

Dose

B

Dose

A

B

Fig. 23.6 Panel a: An isobologram for agent A and agent B at the ED50 for each agent is shown along with the envelope of additivity generated from the mode I and mode II forms of additivity. Panel b: The Mode I additivity curve is generated for each selected level of effect (survival) by selecting a dose of agent A as a starting point. A lower dose of agent A is then selected, the difference in effect from the isoeffect level is determined and the dose of agent B needed to make up the difference is derived from the survival curve for agent B. Mode II additivity is where Ae and Be are the doses of agent A and agent B, respectively, needed to produce the selected effect

Overall combinations that produce the desired effect that are within the boundaries of Mode I and Mode II are considered additive. Those displace to the left are supraadditive and those displace to the right are sub-additive (Fig. 23.6a). Combinations that produce effects outside of the rectangle defined by the intersections of Ae and Be

582

B.A. Teicher

are protective. This type of classical isobologram methodology is difficult to use experimentally because each combination must be carefully titrated to produce a constant level of effect. Dewey et al. described a form of analysis for the special case in which the dose of one agent was held constant [64]. Using full survival curves of each agent alone, this method produces envelopes of additive effect for various levels of the variable agent. It is conceptually identical to generating a series of isoeffect curves and then plotting the survivals from a series of these at constant dose of Agent A on a log effect by dose of Agent B coordinate system [65]. This approach can be applied to the experimental situation in a direct and efficient manner and isobolograms can be derived describing the expected effect (Mode I and Mode II) for any level of the variable agent and constant agent combinations. The schedule and sequence of drugs in combination can affect therapeutic outcome. The definition of additivity and therapeutic synergism has evolved with increasing stringency. In the early work of Schabel, Corbett, and Griswold, therapeutic synergism between two drugs was defined to mean that “the effect of the two drugs in combination was significantly greater than that which could be obtained when either drug was used alone under identical conditions of treatment” [66–75]. Using this definition, the combination of cyclophsophamide and melphalan administered simultaneously by intraperitoneal injection every 2 weeks was reported to be therapeutically synergistic in the Ridgeway osteosarcoma growth delay assay [66–70]. Similarly, the combination of cyclophosphamide and melphalan was reported to be therapeutically synergistic in L1210 and P388 leukemias [71]. Cyclophosphamide plus a nitrosourea (BCNU, CCNU or MeCCNU) were also reported to be therapeutically synergistic in increase in lifespan and growth delay assays using this definition [71]. However, in the EMT6 mouse mammary carcinoma, the maximum tolerated combination therapy of thiotepa (5 mg/kg × 6) and cyclophospahmide (100 mg/kg × 3) produced 25 days of tumor growth delay, which was not significantly different than expected for additivity of the individuals drugs by isobologram analysis [44, 51, 52]. The surivival of EMT6 tumor cells after the treatment of the animals with various single doses of thiotepa and cyclophosphamide was assayed. Tumor cell killing by thiotepa produced a very steep and linear survival curve through 5 logs. The tumor cell survival curve for cyclophosphamide up to 500 mg/kg gave linear tumor cell kill through almost 4 logs. In all cases, the combination treatment tumor cell survivals fell well within the envelope of additivity (Fig. 23.7). Both these drugs are somewhat less toxic toward bone marrow cells by the granulocyte-macrophage colony forming unit (CFU-GM) assay method than to tumor cells. The combination treatments were sub-additive or additive in bone marrow CFU-GM killing. When bone marrow is the dosing limiting tissue, there is a therapeutic advantage to the use of this drug combination [51]. The Lewis Lung Carcinoma (LLC) arose spontaneously as a carcinoma of the lung of a C57BL mouse in 1951 in the laboratory of Dr Margaret R. Lewis at the Wistar Institute. The Lewis lung carcinoma was among the earliest transplantable tumors used to identify new anticancer agents. Sugiura and Stock found that the Lewis lung carcinoma produced tumors 100% of the time yielding a highly malignant carcinoma. These investigators used the Lewis lung carcinoma along with


583

1

1 MURINE EMT6 MAMMARY CARCINOMA THIOTEPAANDCYCLOPHOSPHAMIDE

THIOTEPA(10mg/kg)+ CYCLOPHOSPHAMIDE mode 1 mode 2 CYCLOPHOSPHAMIDE

0.1 SURVIVING FRACTION

SURVIVING FRACTION

0.1

Combination Additive

0.01

0.001

Combination Additive

0.01

0.001 THIOTEPA (10mg/kg)+ CYCLOPHOSPHAMIDE mode 1 mode 2 CYCLOPHOSPHAMIDE

0.0001

MURINE EMT6 MAMMARY CARCINOMA THIOTEPA AND CYCLOPHOSPHAMIDE

0.0001 0

100 CYCLOPHOSPHAMIDE DOSE, mg/kg

200

0

100

200

CYCLOPHOSPHAMIDE DOSE, mg/kg

Fig. 23.7 Isobolograms for the combination treatment of the EMT6 mouse mammary carcinoma treated with 10 or 15 mg/kg of thiotepa and various doses of cyclophosphamide. The survival curve for the EMT6 tumors exposed to cycophosphamide only is shown with open triangles. The dotted lines define the envelope of additivity for the combination treatment. Tumor cell survivals for the combination treatment are shown in solid circles (adapted from Ref. [51])

several other transplantable mouse tumors to determine the antitumor activity of a series of phosphoramides from which the antitumor alkylating agent thiotepa emerged [76–78]. Twenty years later, DeWys standardized techniques for following primary tumor growth by tumor volume measurements and for assessing the response of lung metastases to therapeutic intervention [79]. DeWys observed the Gompertsian pattern of primary tumor growth, effects of tumor burden on therapeutic efficacy, and the effects of the presence of the primary tumor on the growth rate of lung metastases [80]. G. Gordon Steel and co-investigators continued work with the Lewis lung carcinoma and developed culture colony formation techniques, lung colony formation techniques and limiting dilution techniques to assess tumor response to new anticancer drugs and radiation therapy [81, 82]. The syngeneic Lewis lung carcinoma mimics the human disease because the primary tumor metastasizes to lungs, bone, and liver. It is nonimmunogenic and is grown in a host with a fully functional immune system. The rate of tumor growth is relatively rapid, with a tumor volume doubling time of 2.5 days and is lethal in 21–25 days when the tumor cell implant is 106 cells. Although the tumor growth rate is rapid, it is in line with the life-span of the host, which is about 2 years. Gemcitabine (LY18801; 2¢,2¢-difluorodeoxycytidine) is an analog of the natural pyrimidine. The mechanism of action and metabolism of gemcitabine have been well-characterized [83–87]. Gemcitabine is active against many solid tumor models, including the CX-1 human colon cancer xenograft and the LX-1 human lung carcinoma xenograft in nude mice [85–88]. In Phase I clinical trials, gemcitabine was evaluated in a variety of schedules. The greatest efficacy with the least toxicity was obtained with a weekly schedule [83]. In Phase II clinical trials, gemcitabine

584

B.A. Teicher

had activity against small cell lung, nonsmall cell lung, breast, ovarian, pancreatic, myeloma, prostatic, renal, and bladder cancer [89, 90]. Gemcitabine has demonstrated a 22% objective tumor response rate in a database of 331 patients diagnosed with nonsmall cell lung cancer (NSCLC) who received the drug on a weekly schedule in a dose range of 800–1,250 mg/m2. Vinorelbine (navelbine) is a semi-synthetic vinca alkaloid with antitumor activity related to microtubule depolymerization which dissolves mitotic spindles [91–98]. In a variety of human tumor cell lines, vinorelbine was cytostatic at nanomolar concentrations that are significantly below achievable plasma levels in patients [92, 96]. In a number of in vivo studies exploring activity in rodent tumor models and human tumor xenografts in athymic mice, vinorelbine demonstrated efficacy against P388, L1210, B16, and M5076 in vivo mouse models and in animals with human tumor xenografts. Phase I clinical trials showed the maximum tolerated dose of vinorelbine was 30 mg/m2 with weekly intravenous administration. Phase II clinical trials, employing weekly schedules of vinorelbine, demonstrated activity against SCLC, NSCLC and ovarian and breast cancer [96]. Single agent vinorelbine was studied in nonrandomized Phase II human trials as first-line therapy in NSCLC using a weekly schedule and showed good activity with 23 responders out of 70 evaluable patients producing a response rate of 32.8%. The median duration of response was 34 weeks [96–99]. Gemcitabine was an active anticancer agent in animals bearing the Lewis lung carcinoma. Gemcitabine was well-tolerated by the animals over the dosage range from 40 mg/kg × 3 to 80 mg/kg × 3 (Fig. 23.8a) [100]. Navelbine was administered in three different well-tolerated regimens with total doses of 10, 15, and 22.5 mg/kg. Both gemcitabine and navelbine produced increasing tumor growth delay with increasing drug dose. To assess the efficacy of the drug combination, the intermediate dosage regimen of navelbine was combined with each gemcitabine dose. These combination regimens were tolerated and the tumor growth delay increased with increasing gemcitabine dose. Isobologram methodology [51] was used to determine whether the combinations of gemcitabine and navelbine achieved additive antitumor activity (Fig. 23.8a). At gemcitabine doses of 40 and 60 mg/kg, the combination regimens achieved additivity, with the experimental tumor growth delay falling within the calculated envelope of additivity. At the highest gemcitabine dose, the combination regimen produced less than additive tumor growth delay [52, 100]. The untreated control animals in this study had a mean number of 35 lung metastases on day 20. Gemcitabine was highly effective against disease metastatic to the lungs, with a mean number of lung metastases on day 20 decreased to 1.0–1.5 or 3–4% of the number in the untreated controls. Each of the navelbine regimens decreased the number of lung metastases on day 20 to 10 or 11, or about 30% of the number in the untreated controls. The combination regimens were highly effective against Lewis lung carcinoma metastatic to the lungs, with a mean number of 80 Gy, and that an ovarian adenocarcinoma, OCa-1, was moderately sensitive, having a TCD50 of about 53 Gy. Tumors growing in the hind legs of mice were treated with a series of high doses of radiation and followed for relatively long times after irradiation as we had no preconception about the dose response and kinetics for radiation-induced apoptosis in vivo. The results showed that apoptosis occurred in the OCa-1 tumor but not in the HCa-1 tumor. In the OCa-1 tumor, the maximum percentage of apoptosis occurred at 6 h and fell with longer times. The dose response was already on a plateau with the lowest dose used, 25 Gy. The findings prompted a more detailed examination of apoptosis in the OCa-1 model [62]. There we found that the dose response for radiation-induced apoptosis plateaus at about 30–35% apoptotic cells following doses of 7.5 Gy or more. In addition, the apoptosis peaks very soon after irradiation, about 4 h, and then falls off dramatically. These first two series of experiments allowed the following conclusions to be drawn: (a) some tumors are susceptible to apoptosis but others are not; (b) apoptotic index peaks quickly after irradiation and then falls as the apoptotic bodies are phagocytosed; (c) low doses of radiation preferentially induce apoptosis; (d) there is a relatively large proportion of cells that are apparently resistant to apoptosis even within tumors that display apoptosis after irradiation. These conclusions were borne out in an expanded examination of 15 different murine tumors where we wanted to get a clearer picture about the heterogeneity in response [63]. In that study, we confirmed that some types of tumors, namely adenocarcinomas of the mammary gland and ovaries and lymphomas, display an apoptotic response to radiation in vivo whereas other types of tumors, namely squamous cell carcinomas, hepatocarcinomas, and fibrosarcomas, do not. Fortunately, other laboratory data related to the radiation response of these tumors were available allowing us to determine correlations of in vivo apoptotic response to tumor response. This analysis indicated that when radiation-induced apoptosis for all of the tumors was plotted against the respective tumor’s TCD50 value and specific growth delay, those tumors

634

L.C. Stephens et al.

that responded by apoptosis tended to have lower TCD50 values (0.1

Tumor Models in Cancer Research, Second Edition (Cancer Drug Discovery and Development)

Tumor Models in Cancer Research (Cancer Drug Discovery and Development)

Antiangiogenic Agents in Cancer Therapy (Cancer Drug Discovery and Development) (Cancer Drug Discovery and Development)

Cancer Drug Resistance (Cancer Drug Discovery and Development)

Genomics in Cancer Drug Discovery and Development

The Tumor Microenvironment (Cancer Drug Discovery and Development)

Farnesyltransferase Inhibitors in Cancer Therapy (Cancer Drug Discovery and Development)

Farnesyltransferase Inhibitors in Cancer Therapy (Cancer Drug Discovery and Development)

DNA Repair in Cancer Therapy (Cancer Drug Discovery and Development)

Checkpoint Responses in Cancer Therapy (Cancer Drug Discovery and Development)

Chemokine Receptors in Cancer (Cancer Drug Discovery and Development)

Fluoropyrimidines in Cancer Therapy (Cancer Drug Discovery and Development)

Camptothecins in Cancer Therapy (Cancer Drug Discovery and Development)

Deoxynucleoside Analogs in Cancer Therapy (Cancer Drug Discovery and Development)

Antiangiogenic Agents in Cancer Therapy (Cancer Drug Discovery and Development)

Chemoradiation in Cancer Therapy (Cancer Drug Discovery and Development)

Death Receptors in Cancer Therapy (Cancer Drug Discovery and Development)

DNA Topoisomerases and Cancer (Cancer Drug Discovery and Development)

Apoptosis, Senescence and Cancer (Cancer Drug Discovery and Development)

Stem Cells and Cancer (Cancer Drug Discovery and Development)

Tumor Suppressor Genes in Human Cancer (Cancer Drug Discovery and Development)

Tumor Suppressor Genes in Human Cancer (Cancer Drug Discovery and Development)

Targets for Cancer Chemotherapy (Cancer Drug Discovery and Development)

Regional Cancer Therapy (Cancer Drug Discovery and Development)

Gene Therapy for Cancer (Cancer Drug Discovery and Development)

Immunotherapy of Cancer (Cancer Drug Discovery and Development)

Antiangiogenic Agents in Cancer Therapy, Second Edition (Cancer Drug Discovery and Development)

Anticancer Drug Development Guide (Cancer Drug Discovery and Development)

Tumor Models in Cancer Research, Second Edition (Cancer Drug Discovery and Development)

Tumor Models in Cancer Research (Cancer Drug Discovery and Development)

Antiangiogenic Agents in Cancer Therapy (Cancer Drug Discovery and Development) (Cancer Drug Discovery and Development)

Cancer Drug Resistance (Cancer Drug Discovery and Development)

Genomics in Cancer Drug Discovery and Development

The Tumor Microenvironment (Cancer Drug Discovery and Development)

Farnesyltransferase Inhibitors in Cancer Therapy (Cancer Drug Discovery and Development)

Farnesyltransferase Inhibitors in Cancer Therapy (Cancer Drug Discovery and Development)

DNA Repair in Cancer Therapy (Cancer Drug Discovery and Development)

Checkpoint Responses in Cancer Therapy (Cancer Drug Discovery and Development)

Chemokine Receptors in Cancer (Cancer Drug Discovery and Development)

Fluoropyrimidines in Cancer Therapy (Cancer Drug Discovery and Development)

Camptothecins in Cancer Therapy (Cancer Drug Discovery and Development)

Deoxynucleoside Analogs in Cancer Therapy (Cancer Drug Discovery and Development)

Antiangiogenic Agents in Cancer Therapy (Cancer Drug Discovery and Development)

Chemoradiation in Cancer Therapy (Cancer Drug Discovery and Development)

Death Receptors in Cancer Therapy (Cancer Drug Discovery and Development)

DNA Topoisomerases and Cancer (Cancer Drug Discovery and Development)

Apoptosis, Senescence and Cancer (Cancer Drug Discovery and Development)

Stem Cells and Cancer (Cancer Drug Discovery and Development)

Tumor Suppressor Genes in Human Cancer (Cancer Drug Discovery and Development)

Tumor Suppressor Genes in Human Cancer (Cancer Drug Discovery and Development)

Targets for Cancer Chemotherapy (Cancer Drug Discovery and Development)

Regional Cancer Therapy (Cancer Drug Discovery and Development)

Gene Therapy for Cancer (Cancer Drug Discovery and Development)

Immunotherapy of Cancer (Cancer Drug Discovery and Development)

Antiangiogenic Agents in Cancer Therapy, Second Edition (Cancer Drug Discovery and Development)

Anticancer Drug Development Guide (Cancer Drug Discovery and Development)

Recommend Documents