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 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

S. Decker and E. Sausville

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].

1  Perspectives on the History and Evolution of Tumor Models

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

S. Decker and E. Sausville

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

1  Perspectives on the History and Evolution of Tumor Models

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

S. Decker and E. Sausville

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].

1  Perspectives on the History and Evolution of Tumor Models

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

S. Decker and E. Sausville

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

1  Perspectives on the History and Evolution of Tumor Models

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

S. Decker and E. Sausville

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

1  Perspectives on the History and Evolution of Tumor Models

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

S. Decker and E. Sausville

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

1  Perspectives on the History and Evolution of Tumor Models

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

S. Decker and E. Sausville

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.

1  Perspectives on the History and Evolution of Tumor Models

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

S. Decker and E. Sausville

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,

2  Murine L1210 and P388 Leukemias

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

Net Cell Kill (log10 units)

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

Net Cell Kill (log10 units)

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

Net Cell Kill (log10 units)

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

2  Murine L1210 and P388 Leukemias

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 

2  Murine L1210 and P388 Leukemias

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.

2  Murine L1210 and P388 Leukemias

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.

2  Murine L1210 and P388 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.

2  Murine L1210 and P388 Leukemias

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]

B.A. Teicher (ed.), Tumor Models in Cancer Research, Cancer Drug Discovery and Development, DOI 10.1007/978-1-60761-968-0_3, © Springer Science+Business Media, LLC 2011

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).

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

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.

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

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

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

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

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

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 (<10% from 1,000 mg tumors). It may be inferred from a casual reading of the literature that one tumor cell can be implanted, and will take and grow into a tumor mass that will kill the host. Although this is true for some tumors such as L1210 or P388 leukemias, these are exceptions. The lack of better tumor-takes at a particular location (other than the brain) is caused by 1–3 log-cell loss of tumor cells at the implant site, which occurs within 24 h of implant [16]. This cell loss can be prevented by admixing X-ray-killed tumor cells with the implants, or by admixing a brain-tissue brei with the tumor cells. Table 3.4  Comparison of take-rates of Colon Adenocarcinoma-10 (a low viability, poor t­ake-rate tumor) by various implant routes Number of tumor-takes/number of mice implanted Number of cells implanted Subcutaneous Intravenous Intracranial Intraperitoneal 8 10 8/9 10/10 107 2/10 6/9 6/10 106 1/10 1/10 10/10 0/10 105 0/10 1/10 9/10 0/10 104 6/9 0/10 103 3/10 0/9 102 Take-rate level 4 × 107 1 × 107 1 × 104 1 × 107 One take-rate unit (by dilution) will establish a tumor in 63% of the mice [100 × (1 – 1/e)] This colon adenocarcinoma was a slow growing, low viability tumor. It was in the 11th passage for this implant trial Note that the tumor-takes were 3 logs better by the IC route than by any other route of implant Table 3.5  Comparison of subcutaneous take-rates of different transplantable tumors Number of cells Number of tumor-takes/number of mice implanted implanted Mam-44 Colon-26 Mam-16/C Colon-51 Colon-09 Colon-10 Colon 36a 3/3 8/9 107 19/19 10/10 9/9 2/10 2/10 106 10/10 10/10 19/20b 10/10 2/10 1/10 0/10 105 10/10 10/10 17/20 3/10 0/10 0/10 0/10 104 10/10 9/10b 5/20 1/10 0/10 103 9/10 10/10 0/19 0/10 102 Not done 2/10 3 × 102 3 × 104 3 × 105 3 × 106 4 × 107 5 × 107 63% take-rate <103 level One take-rate unit (by dilution) will establish a tumor in 63% of the mice [100 × (1 – 1/e)] a In separate experiments, 2 × 107 cells SC produced 6/20 takes with IP implantation, confirming that this was a poor take-rate tumor b Leakage of the titered cell implant brei from the implant site probably accounts for the one no-take in each of these groups. Three logs of incomplete take should not occur with technically perfect implants (as occurred in Mam-16/C)

54

L. Polin et al.

An example is shown in Table 3.6. This phenomenon is known as the feeder effect or the Revesz effect, after the original investigator [17, 18]. The feeder cells (radiation-killed cells) can be different from the titered tumor cells being assayed, but work best if they are from the same inbred strain and are derived from a highly metastatic/fast-growing/high-take-rate tumor (Table 3.7). The only normal tissue Table 3.6  Comparison of B16 melanoma tumor-takes with and without added feeder layers Co60 killed B16 cells Brain Brei from No feeder layer C57Bl/6 mice added to titered tumor cells added to titered titered tumor cells only tumor cells B16 TumorB16 TumorB16 Tumortakes/ Median takes/ Median takes/ Median number number time to time to number Number of cells time to implanted a 1.0 g implanted a 1.0 g implanted a 1.0 g implanted 106 5/5 24 105 4/5 30 4/4 18 4/4 17 104 1/10 34 10/10 21 10/10 22 103 0/10 9/10 27 10/10 26 102 0/10 5/10 (27–35 6/10 (28–38 range) range) 101 0/10 1/10 Feeder layer only 0/10 One take-rate unit (by dilution) will establish a tumor in 63% of the mice [100 × (1 – 1/e)]. This is the take-rate level a B16 Tumors took and grew to 1.0 g in size. Feeder layers (approximately 5 × 107 killed cells/ implant) made with irradiated tumor cells or brain tissue and admixed with the titered tumor cells improved the take-rate by over 2 logs each Table 3.7  Comparison of Adenocarcinoma-755 tumor-takes in C57Bl/6 mice with and without added feeder layers: note, one feeder layer was made with Co60 killed B16 cells; the other was made from a brain tissue brei 10,000R Co60 killed B16 Brain Brei from C57Bl/6 mice added to titered No feeder layer titered cells added to titered tumor cells tumor cells tumor cells only CA755 CA755 CA755 Tumor-takes/ Median tumor-takes/ Median tumor-takes/ Median number time to number time to Number of cells number time to Implanteda implanted a implanted a 1.0 g 1.0 g implanted 1.0 g 106 5/5 18 5/5 11 5/5 16 105 5/5 24 5/5 14 5/5 17 104 0/10 10/10 16 5/5 22 103 0/10 10/10 20 5/5 33 102 7/8 26 3/5 38 101 0/10 0/5 Feeder layer only 0/10 Adenocarcinoma-755 tumors took and grew to 1.0 g in size. Feeder layers made with irradiated tumor cells (of a different tumor) or brain tissue and admixed with the titered tumor cells improve the take-rate by more than 2 logs each a

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

55

that provides the feeder effect to all tumors is brain tissue. The feeder effect provided by brain tissue probably accounts for some of the difficulty in curing intracranial malignancies, since an additional 1–3 logs of tumor-cell-kill would be necessary to reduce the population below the take-rate level at that site. The feeder effect can only be provided by intact, metabolizing cells, and must be admixed with the titered cells and not at a distant location. Furthermore, greater than 106 X-ray killed cells are needed to produce the effect; with 5 × 107 to 1 × 108 used by most investigators. In selected cases, tumor cells killed by an antitumor agent can provide feeder activity [19]. It is likely that some feeder effect is taking place post-treatment with either chemotherapy or radiation treatment of any intact tumor mass. Immune-suppression does not change the take-rate of a syngeneic tumor model (Table 3.8). Thus the feeder effect is unrelated to subverting immune function. Other examples of the lack of influence of immune-suppression on take-rates are available [20]. As one approaches within 1 log of the take-rate level (or cure level), an accelerated tumor-cell loss is clearly taking place (compared to implants that are 3–4 logs above the take-rate). This is evident in virtually any end point or tumor location [10]. There is also a consequent shift in the shapes of the curves (survival or time to a specific tumor size). At 3–4 logs above the take-rate, curves are normal (i.e., Gaussian) (Fig. 3.1) [10]. With a Gaussian curve, the top half of the curve can be inverted and can then be superimposed on the bottom half of the curve. At approximately 0.5–1.5 logs above the take-rate, the curves become log-normal, in which case the data can be replotted using the log of time, and the curve will become normal or Gaussianshaped [10]. Near the take-rate level, which yields 20–50% no-takes or survivors after therapy, the curve is first-order kinetic (Fig. 3.1) [10]. This knowledge can be used to predict the outcome of a mouse experiment, or even a partially completed human trial. If the survival curve plunges in a Gaussian fashion, it is clear that one is far above the take-rate level, and no cures will be obtained. On the other hand, if the curve begins to bend-out in a first-order kinetic fashion, one can project that some cures will be obtained. Detailed examples of normal, log-normal, and firstorder-kinetic plots, with explanations and mortality curve analyses have been published [10]. Other examples of first-order-kinetic behavior are common in physics and biology: e.g., light-bulbs burn out by first-order kinetics (a constant fraction of the bulbs burn out per unit of time); most antitumor agents kill by first-order kinetics (a constant fraction of the cells are killed per given dose of a drug) [21]; and disinfectants kill by first-order kinetics (a constant concentration of the disinfectant in a test tube kills a constant fraction of the bacterial cells per unit of time) [21]. Once an investigator understands the take-rate behavior of a particular tumor, understands the relationship between tumor size and cell number, and understands the feeder effect, the meaning of cure for that particular tumor will become clear, regardless of the stage of the disease and the nature of the therapy. This knowledge can also be used in experimental planning and projecting the outcome of various treatment regimens. With this perspective, we can now examine the results from treatment of high take-rate tumors and poor take-rate tumors. The first example is shown in Table 3.9. A total of 21 cures were produced in 77 mice with poor take-rate tumors by therapies

17/17 2/19 1/20 0/19

Tumortakes/total

106 105 104 103

Syngeneic Syngeneic Syngeneic Syngeneic

ALS ALS ALS ALS

0.075 0.075 0.075 0.075

day –3 to day +7 day –3 to day +7 day –3 to day +7 day –3 to day +7

36 55.5 46 None

29–42 45–66 46 None

0/20 18/20 18/19 20/20

20/20 2/20 1/19 0/20

B16, 60 mg Allogeneic fragments SC B16, 60 mg Allogeneic fragments SC B16, 60 mg Syngeneic fragments SC No Rx

No Rx

No Rx

No Rx

0.075

ALS

No Rx

No Rx

day –3 to day +7

0/8

10/10

1/10

8/8

0/10

9/10

Immune-suppression was sufficient to allow B16 melanoma to take and grow in a foreign (allogeneic) host mouse (C3H) in 9/10 mice (8/10 of these grew to over 1,000 mg)

BDF1

C3H

C3H

Controls to determine the degree of immune-suppression

One take-rate unit by dilution will establish a tumor in 63% of the mice: For the immune-suppressed mice (ALS treated) shown above, the take-rate level was calculated to be 6 × 105 B16 tumor cells; identical with that of the saline treated controls

BDF1 BDF1 BDF1 BDF1

Immune-suppressed; antilymphocyte serum (ALS) started 3 days before tumor implant and continued for 7 days after tumor implant

One take-rate unit by dilution will establish a tumor in 63% of the mice: For the controls shown above, the take-rate level was calculated to be 6 × 105 B16 tumor cells

Table 3.8  Lack of influence of immune-suppression on the take-rate of B16 melanoma cells implanted intravenously in a syngeneic mouse Dosage; No.  of B16 Tumor free Median day of volume in Immunecells tumor death Range of tumor survivors on ml/mouse/ Schedule of suppressive implanted IV Nature of day 205 (dying only) deaths day treatment agent transplant Host strain on day 0 Controls BDF1 106 Syngeneic None (saline) 0.075 day –3 to day +7 34 26–41 0/17 BDF1 105 Syngeneic None (saline) 0.075 day –3 to day +7 52.5 45–60 17/19 BDF1 104 Syngeneic None (saline) 0.075 day –3 to day +7 56 56 19/20 BDF1 103 Syngeneic None (saline) 0.075 day –3 to day +7 None None 19/19

56 L. Polin et al.

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

57

Fig. 3.1  Normal, log-normal, and first-order kinetic curves

that produced only one cure of 159 mice with high take-rate tumors; the log10 ­tumor-cell-kills were similar. A comparison of the curability, as a function of a threefold increase in tumor size of a poor take-rate tumor, is instructive (Table 3.10). Clearly, the tumor (Colon-36) was slightly less sensitive at a larger size (note growth delay and net cell-kill), but not markedly less. Curability, however, decreased dramatically, with only a threefold increase in tumor-cell number. The calculations can be used to approximate the cure level (take-rate level and the feeder effect). Starting with the 200-mg masses, the median is 2 × 108 cells to start (8.3 logs). Treatment killed 2.9 logs, and thus the median population is reduced to 5.4 logs = 2.5 × 105 (producing 10/15 cures). At the 600-mg size masses, the median is 6 × 108 cells to start (8.78 logs). Treatment killed 2.1 logs. Thus, the median population is reduced to 6.68 logs = 4.8 × 106, and only 1/8 cures. The data can be plotted on semi-log paper, and various results can be estimated. If the population is reduced to below 4 × 104, it should yield all cures. Above 1 × 107, there should be no cures. The 63% cure level is 3 × 105 cells. The feeder contribution for this dose-schedule of treatment and established tumors at the start of therapy was 2.22 logs (7.7 logs–5.48 logs). The cure of a poor take-rate tumor by chemotherapy is shown in Table 3.11. Again, the tumor-cell-kills were rather modest, considering the cures. Based on the last tumor to regrow in the treatment groups, it is possible to calculate that the cures produced by Procarbazine and Maytansine represent only slightly more than a 1.8 and 1.6 net log-kill, respectively. The cure of a high take-rate tumor by chemotherapy is clearly a challenge. We have never obtained cures of 30 mg SC trocar-implanted tumors such as Mam-44 or Panc-02 with chemotherapy, even when treatment begins the day after implant. A more responsive tumor, such as Mam-16/C, has had only an occasional chemotherapeutic cure in hundreds of trials. Examples of treatment of Mam-16/C with

Curesa 48 33 18

7/8 8/8 8/8 6/8

7/8 7/8 7/8 0/8

Curesa 0.8 0.55 0.3

Curesa 2.4 1.6 0.9

Curesa   2.2   1.45   0.65

5/8 2/8 2/8 0/8

Time for median control tumor Colon-10 to reach 750 mg post-trocar implant = 36 days. Exponential Td = 6.0 days63% take-rate level = 4 × 10 7 Curability was, in part, a function of the viability (take-rate) of a tumor. All tumor models were synegeneic a  Only two mice regrew tumors, growth delay not meaningful (five cured, one Rx death)

1/8 0/8 0/8 0/8

1,100 845 650 500

Colon Adenocarcinoma-10 at 100–400 mg

3,300 2,535 1,950 1,500

845 2,535 0/15 37 13/15 10/15 1.07 3.2   2.9 10/15 650 1,950 0/15 20 11/15 6/15 0.59 1.77   1.4 0/15 500 1,500 1/15 17 4/15 4/15 0.50 1.5   1.15 2/15 Time for median control tumor Colon-36 to reach 750 mg post-trocar implant = 22 days. Exponential Td = 3.4 days63% take-rate level = 5 × 107

Colon Adenocarcinoma-36 at 100–400 mg

Curable, poor take-rate tumors

845 2,535 0/8 17 3/8 1/8 0.77 2.3   1.8 0/8 650 1,950 0/8 12 3/8 0/8 0.55 1.64   1.2 0/8 500 1,500 0/8 10 0/8 0/8 0.46 1.37   1.0 0/8 Time for median control tumor Colon-51 to reach 750 mg post-trocar implant = 15 days. Exponential Td = 2.2 days (20 mice)63% takerate level = 3 × 105

Log10 kill per Log10 kill Log10 kill fraction Gross Net Cures

Colon Adenocarcinoma-51 at 100–400 mg

CRs excluding toxic deaths

1,000 3,000 3/26 13 12/26 8/26 0.9 2.8   1.7 0/26 750 2,250 1/42 12 20/42 16/42 0.86 2.6   1.5 0/42 500 1,500 0/49 7 14/49 6/49 0.5 1.5   0.4 1/49 200 600 0/18 3 0/18 0/18 0.23 0.7 −0.4 0/18 Time for median control tumor Mam-16/C to reach 750 mg post-trocar implant = 11–15 days. Exponential Td = 1.4 days63% take-rate level = 3 × 104

PRs excluding toxic deaths

Mammary Adenocarcinoma-16/C at 100–400 mg

Table 3.9  Treatment of transplantable tumors of mice with X-irradiation Treatment Median tumor RADS per growth delay Total dosage related Tumor (SC) size at first fraction in days deaths (Q2dx3) in RADS Rx Non-curable, high take-rate tumors

845

2,535

0/8

27.3

6/8

1/8

0.8

1.07

2.4

3.2

2.1

2.9

1/8

10/15

Log10 kill Log10 kill Log10 kill per fraction gross net Cures

function of take-rate and tumor size at first

Time for median control tumor Colon-36 to reach 750 mg post-trocar implant = 22 days. Exponential Td = 3.4 days Curability was, in part, a function of the viability (take-rate) of the tumor. Note that the net log tumor-cell-kill for Colon 36 did not change markedly with the change in tumor size at first Rx (and the log kill was not impressively large). The curability was simply related to the log kill necessary to reduce the population below the take-rate level with a feeder

300–1,000 mg (600 mg median = 6 × 10 8 cells

Table 3.10  Treatment of transplantable Colon Adenocarcinoma-36 in mice with X-irradiation curability as a treatment Treatment- Median tumor PRs RADS per Total CRs excluding growth delay excluding related dosage fraction toxic deaths toxic deaths in days Tumor (SC) size at first Rx (Q2dx3) in RADS deaths Colon-36 100–400 mg (200 mg 845 2,535 0/15 37 13/15 10/15 median = 2 × 108 cells) )

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice 59

Not tabulated

0/10

10/10 CRs

0%

Not tabulated

46

–

Day 3,10 (IV) 0.44

– Not tabulated

0/10

–

0/10

–

1,000 mg on – day 23 63 mg on day 6% 23

1,000 mg on – day-25 100 mg on 10% day 25

37(27–52)

23(20–35)

36 (28–51)

25 (21–32)

55 (52–58)

14

–

11

–

34

27

Tumor growth delay in days –

1.2

–

0.83

–

3.2

2.5

Median log10 kill gross –

0.6

–

0.45

–

2.5

2.2

Median log10 kill net –

2/10 on day 93

0/10

2/10 on day 67

0/10

4/10 day 208 8/10 day 208

Tumor– free “Cures” 0/20

Td = 3.6 days: Trocar implanted SC on day 0 This is an excellent example of the utility of converting tumor growth delay data to log10 tumor-cell-kill. The cures do not represent a huge log kill. Only Ara-C and it=s depot form Palmo-Ara-C had substantial activity. The other two agents were only modestly active against this poor take-rate tumor.

Colon-36 No Rx (1499Q1) Maytansine

Not tabulated

Not tabulated

–

Td = 4.0 days. Trocar implanted SC on day 0.

Colon-36 No Rx – (1499J1) Procarbazine qd3–8 (PO)

The day of first treatment, the tumors were 231 mg median on day 16 (range 144–320) = 2.3 × 10 8 Exponential tumor-volume doubling time (Td) = 3.2 days. The 63% take-rate level = 5 × 107

q3hx8 (IP) days 16,20 Palmo-AraC qd16–23 (IP) 13.5

Ara-C

Tumor (Exp) Treatment Colon-36 No Rx (1564G1)

Schedule (route) –

Median time in days for tumors to Median % Body-weight reach 1,000 mg Drug tumor mass mg/kg/ loss at nadir % T/C mass (range) injection (day of nadir) deaths on (days) – 21 (17.5– – Not tabulated – 1,000 mg 27.5) on day 21 11.8 Not tabulated 0/10 9/10 CRs 0% 48 (45–65)

Table 3.11  Chemotherapeutic treatment of a poor take-rate tumor

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

61

highly active agents are shown in Table 3.12. The log-kill values were substantial, but cures were not forthcoming, except in one case. This cure was re-challenged on day 182, and grew to over 2 g in 11 days, indicating that immune factors were not involved in the cure. The log-kill values – gross and net – were also consistent with a cure. It should be noted that other dose-schedules of this agent failed to produce cures or this impressive log-kill (thus, even this one cure is an outlier). In general, in order to obtain cures of such high take-rate tumors, surgery ­followed by chemotherapy has been successful, especially with highly active drug combinations [6, 7, 10, 11]. However, in order to produce cures of such models as Mam-16/C without surgery, we would need to resort to some tricks. For example, we could IP-implant the tumor as a counted cell brei, and start IP treatment the next day. This would improve the tumor-cell-kill by approximately 3 logs, and effect cures if the agent was highly active. This regimen has been successful for multiple agents, but does not reflect the usual clinical problem or treatment requirement. Likewise, we could implant the tumor only slightly above the take-rate, and begin treatment the next day by another route. Again, some cures could be expected if the agent could produce a 1 log net cell-kill, but to what aim? A casual reading of the literature may prompt the opinion that L1210 leukemia is a highly chemo-sensitive and curable tumor. It is indeed chemo-sensitive, but the cures have mainly been obtained with IP-implanted tumor and IP-drug ­administration the day after 105 cell implants. One can easily remove the possibility of a cure: the tumor can be IV-implanted with 106 cells, with treatment starting the next day. With this titer and implant route, cures are out of the question (Table 3.13). (In  many respects, cure is the worst and most misleading endpoint that an investigator­ can use. In most cases, the reader will have no concept of the actual log-kill required to produce the cure).

3.8 Results Tabulation of Chemotherapy Trials: Desired Information from a Tumor Model It is often said that a picture is worth a thousand words. However, the presentation of chemotherapeutic results in plots can be among the most misleading methods of presentation if all the information is not provided to the reader. Several examples follow. The first example (Fig. 3.2) would seem to indicate self-evident antitumor activity of Gemzar against a very slow-growing, early-passage pancreatic adenocarcinoma. However, the plot is truncated in time to pick the best appearance, and we have provided very little information (intentionally). The following data are obviously missing: (1) What was the body-weight loss from the treatment? (2) What happened to the treated mice later in time? If the reader is not told what happened to all of the treated tumors to a size of 1,000 mg, the presenter is remiss. Sacrifice for some assay or histology is not a valid explanation. Historically, if other findings are needed, extra mice are added to the trial, and the mice used for efficacy determination are not summarily killed until all the vital information is obtained. (3) What was the log-kill from

#1361 Passage 62

#1286 Passage 98

–

–

–

bid6,8,11,13 (IV) 124

d6,8,11,13 (IV)

No Rx

Taxol

VP-16

84

18

day 2,5,8 (IV)

Adriamycin

–

–

144

–

No Rx

Cryptophycin d1–8 (IV) #55

#2064 No Rx Passage 116

–

Loss 7%

Loss 8%

Gain 5%

Loss 4% day 9

Gain 6% day 8

0/7

0/7

–

0/5

–

Loss 7.5% (day0/5 8)

Unchanged day 8

1,790 (1,296– – 2,537) on day 9 0 (all zeros) 0% on day 9 0 (0–63) on 0% day 9

1,539 (446– – 2,440) on day 11 0 (all zeros) 0% on day 11

1,710 (723– – 2,232) on day 11 0 (all zero) on 0% day 11

13.5

21 (20–22)

21 (20.3–22.5) 13.5

–

16

–

13

–

7.5 (4.5–8.5)

24 (20–63)

8 (8–14.5)

23 (20.5–24)

10(8.5–13)

3.3

3.3

–

4.8

–

3.9

–

63% Take-rate level = 3 × 104; The exponential tumor-volume doubling time = 1.0 days; Tumors Implanted as 30–60 mg fragments on day 0

Table 3.12  Chemotherapeutic treatment Mammary Adenocarcinoma-16/C (a high take-rate tumor)

2.0

2.0

–

3.0

–

2.6

–

0/7

0/7

0/6

0/5

0/5

0/5

0/5

XK469

–

–

No Rx

0/5

Gain 4% day 7 –

Unchanged day 9

Gain 7% day 8 –

2,460 (309– – 3,560) on day 11 0 (0–95) 0%

1,263 (735– – 1,450) on day 9 0 (all zeros) 0% on day 9 8.5 (7.5–14)

22 (19.3–39)

8.6 (8.5–9.5)

–

13.4

–

–

4.0

–

d1,3,5,8,10, 350 Unchanged day0/5 41.7 (25.5–66.5)33.2 10 12,14 (IV) 14 With many dozens of trials, almost no cures could be found of this high take-rate tumor (indicating the difficulty in producing enough log kill)

#1620 Passage 134

420

–

qd 2–8 (PO)

–

SR271425

#2278 No Rx Passage 118

6

–

2.2

–

1/5

0/5

0/5

0/5

64

L. Polin et al.

Table 3.13  Non-curability of IV implanted L1210/0 A high take-rate tumora Activity L1210/0 rating leukemia MTD total Cures/no. against Gross dosage of mice L1210 log kill (mg/kg) Agent Exp no. implanted XK469 2473 300 6.4 +++++ 0/5 SR271425 2304 350 6.4 +++++ 0/5 VP-16 2473   70 8.0 +++++ 0/5 PALMO ARA-C 2443 140 5.7 +++++ 0/5 GEMZAR 2483   84 6.0 +++++ 0/5 BCNU 2483   28 6.7 +++++ 0/5 5-FU 2473 180 4.0 ++++ 0/5 CISDDPT 2473   12 2.4 ++ 0/5 ADRIAMYCIN 2304   17 1.6 + 0/6 VINBLASTINE 2473 6.0 1.6 + 0/5 TAXOL 2304   63 0.4 C 0/6 Mice were implanted IV on day 0 with 106 leukemic cells in the treatment groups (5 mice/group) and the corresponding Control group. Titered controls were also included (104 and 102) in order to verify the tumor doubling time and log10 kill. Treatment started on day 1 a  One cell will establish tumors with L1210

Fig. 3.2  Exp 2475: Treatment of a human pancreatic ductal adenocarcinoma

the treatment? An unwillingness to do the log-kill can indicate that the presenter is attempting to conceal how poor the activity really is. If the treatment affects the tumor growth after cessation of Rx (e.g., radiation treatment of the tumor and tumor bed can

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

65

slow regrowth), the exponential Td is taken from the treated tumors – and not the control tumors – for the purpose of the log-kill calculation. This is the same trial presented in Fig. 3.2, but supplemented with additional information (including the details to the end of trial). The first important piece of information is the weight loss (-8.9% body-weight loss, with the nadir on day 89). Many antitumor agents and toxic agents make a mouse sufficiently sick that it fails to adequately eat or drink fluids. This causes a nonspecific weight loss. It is well known that weight loss alone will cause substantial growth inhibition of a solid tumor [22]. A -11% body-weight loss produced by 9 days of caloric restriction will produce a tumor-mass inhibition (T/C) = 47% for Mammary Adenocarcinoma16/C (with no deaths). Caloric restriction will have even greater effects on other solid tumors [22]. The NCI cutoff for a minimum indication of antitumor activity is a T/C of 42%. This cutoff was adopted because of the various caloric restriction studies (Ref. [22] is the most cited). Importantly, a T/C = 20% was possible with caloric restriction with no deaths [22], although the weight loss was punishing. A weight loss of -20% or more is considered frankly toxic by NCI standards, but weight losses of -15% are usually attended by some lethality. Historically, NCI requires a T/C (tumor mass) less than 10% to indicate high antitumor activity for a solid tumor. As an aside, weight loss does not add to survival improvement for a leukemia such as L1210 or P388, and is not an issue for survival testing. As Fig. 3.3 is examined, it is apparent that the treatment must have been near the lethal level, since the weight loss was -8.9% and the nadir occurred nearly 40 days after

Fig. 3.3  Exp 2475: Treatment of a human pancreatic ductal adenocarcinoma

66

L. Polin et al.

t­reatment (indicating the treated mice were probably in relatively poor condition for an extended period). The T/C (mass) = 50% on day 129, which is inactive by NCI standards. Statistical significance values are often applied to solid-tumor treatment data. A T/C in the 42–50% range would usually meet statistical significance at the 5% level, but is hardly of biologic importance in light of weight-loss information. In some cases, tumor-mass data can be presented as a mean value (average value). The mean provides greater proportional emphasis on the largest tumors in the group. In some cases it may be justified, but alternatively the presentation of a mean value may be done to place the best spin on the results, not the most conservative-spin. Historically, NCI prescribes that median values be used for tumor-mass information and mean values for weight-loss information. The next item to note is the dispensation of all the treated mice up to a size of 1,000 mg (Fig. 3.3). In this example, the median did not reach 1,000 mg. One mouse died from tumor burden on day 136. Two mice had tumors over 1,200 mg on day 143, and were sacrificed. The last measurement of all mice showed a median tumor mass of 580 mg (Fig. 3.3). The tumor-growth delay was 49 days at this time-point. This is a substantial delay, until the weight loss and the slow Td of the tumor (Td = 22 days) are considered. The last item is the conversion of the growth delay into gross log-kill (see Sect. 3.11 for additional details): Log − kill (gross) =

tumor growth delay in days 3.32 × (exponential tumor − volume doubling time )

Log − kill (gross) =

49 = 0.62(inactive ) 3.32 × (22 )

Now that the results are quantified and explained, the self-evident antitumor activity vanishes. This is a very good, highly active antitumor agent, but it does not work against this particular tumor. Figure 3.4 is another example of an incomplete plot (treated tumors not plotted up to 1,000 mg). On superficial examination, the agent may be considered active. If the mice are sacrificed on day 11, there would be no information to refute the purported activity. In actual fact, this is simply a truncated presentation of an agent that produced delayed lethality. The following actually occurred: 225 mg/kg total: all five mice were drug deaths, dying on day 11, 12, 13, 13, 14 150 mg/kg total: all five mice were drug deaths, dying on day 13, 14, 16, 16, 16 100 mg/kg total: there were two drug deaths (both day 16). The other three mice were sacrificed because of tumor burden. The optimum T/C mass = 700/1,830 = 38% on day 9 (marginally active). The log-kill (gross) of this treatment group = 0.75 (­marginally active). Thus, in the final analysis of the agent used in Fig. 3.4, all treatment groups were toxic, and even the dose that produced 40% drug deaths was only marginally active.

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

67

In many cases, mice are sacrificed if they are dying from toxicity (and this is not explained to the reader). Delayed toxicity is an often overlooked topic in the use of tumor models [23]. In some cases, the delayed lethality can occur 30–120 days after treatment, even with clinically available agents [23]. It is for this reason that investigators will include non-tumor toxicity control mice in some trials; i.e., non-tumor normal mice injected with the maximum tolerated dosages (MTD based on short-term toxicity findings) of the drug and held for 150 days post-treatment to verify that no longdelayed toxicity exists [23]. Often, cured mice are held for the same purpose. One of the hallmarks of a delayed lethality drug is the failure of the mice to gain weight satisfactorily or to add skeletal mass. Normally, a mouse will continue to gain skeletal size throughout its life span, and at least reach 30 g within 3 months posttreatment with most antitumor agents. If the mice stay in the 19–21 g size for an extended period of time, one can suspect serious non-reversible organ toxicities (e.g., lung, kidney). The use of non-tumor toxicity-control mice to clearly separate tumor effects from treatment effects is standard practice [23]. It should be noted that the plot shown in Fig. 3.4, as well as the other plots, are done on Cartesian plots. This tends to give the appearance of more separation between the control and treated than there actually is (especially if the treatment groups are not plotted to 1,000 mg). Historically, it is better to use semi-log plots (with the median tumor masses plotted on the log axis). The exponential doubling-time portion of the growth curve will then be essentially linear, allowing the calculation of the exponential Td with more accuracy.

Fig. 3.4  Treatment of SC mammary adenocarcinoma 16/c

68

L. Polin et al.

Fig. 3.5  Treatment of MX-1 mammary tissue in female athymic nude mice (NCI data)

Figure 3.5 shows the NCI data. If it was not explained, it would appear that adriamycin was nicely active against the human breast adenocarcinoma MX-1 [T/C = 12%; 2/6 tumor-free mice cures; median of 1.7 log-kill among those not cured (++ activity rating)]. In fact, adriamycin is not active against this tumor in our testing [1, 3], and furthermore, it is markedly more toxic in immune-deficient mice than in conventional mice [3]. The dose used in the experiment shown in Fig. 3.5 is approximately three times the MTD in immune-deficient mice [3]. This is a superb educational example because the weights of all the mice were provided up to and including the day of sacrifice (day 39). The first item to note is the large body-weight loss (-12.4% of body weight) occurred with the nadir on day 39 (the day of sacrifice). At face value, the tumors could be considered to cause some of the weight loss, except that the mice were continuously losing weight post-treatment, including two mice that were listed as tumor-free. Furthermore, the tumors in this cage were too small to cause the mice to lose weight (based on the controls). However, the actual weight losses of the tumor-free mice essentially tell the whole story. One lost 5.0 g (-26% of its body weight), and the other mouse lost -4.9 g (-22% of its body weight). The mice were obviously sacrificed because they were moribund and dying of drug toxicity. In addition, it is unacceptable to kill tumor-free mice so soon after treatment. These mice would need to be held for at least 2 more months to be assured a cure. Lastly, the cures would have needed to be re-challenged with the tumor to prove that the cures were not the result of immune rejection in a leaky mouse. The next lower dose of adriamycin was also a toxic dose, and had no antitumor activity (T/C = 103%; no log-kill).

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

69

3.9 Characterization of a Tumor Model and Quality-Control Monitoring To a large extent, the types of characterization end points that the investigator needs depend on the usage of the model. Furthermore, the tumor model is not a static entity. The exponential tumor-volume doubling time usually shortens substantially during the first 10 transplant generations, and can shorten more with additional passage [10]. For example, Mam-16/C had a 4.5-day Td in the 4th passage and a 1.25-day Td in the 29th passage. The cause is a decrease in the cell-loss factor [24], which is consistent with an observed improvement in the take-rate of a tumor with serial passage. Interestingly, the growth fraction and the intermitotic times do not change markedly with serial passage of new tumors [24]. Overall, it would appear that a transplantable tumor becomes fast-growing and less differentiated for two reasons: (1) the normal components that make up a portion of the original tumor and impart the highly differentiated character are lost in the first transplant (normal cells do not transplant well); and (2) the cell-loss factor decreases (either through an improvement in the fidelity of replication or by adapting to the implant environment). Either way, it is hard to envision that any of these changes could ­substantially alter the predictive worth of the model, other than making it more challenging. The Td has a bottom limit of 1.0 day for solid tumors of mice, and 0.35 days for leukemias of mice. We propose the following suggestions for using solid-tumor models for chemotherapy purposes. In most cases, this information will provide adequate characterization and quality-control monitoring. 1. Use tumors that have been in passage for over 10 serial passages (there is too much tumor to tumor-growth variability with fewer passages). 2. Use tumors from in vivo passage and not culture passage. Tumors directly from culture tend to magnify responses, because they are less adapted to the in vivo environment and thus more fragile. 3. Determine the take-rate of the tumor with titered tumor-cell implants at the various locations intended for use in the model. One should then use a titer for chemotherapy that is at least 3 logs above the 63% take-rate level (more if high antitumor activity is expected). 4. Obtain histologic sections of the tumor at least every three or four passages (every passage if it is a human tumor because of the diffuse necrosis problem). 5. Determine the exponential tumor-volume doubling time, and monitor this every few passages (every passage if it is a human tumor, since human tumors can become contaminated with spontaneous lymphoma cells, resulting in a marked decrease in the Td; or undergo excessive diffuse necrosis and have a marked increase in the Td). 6. Determine the metastatic behavior of the tumor at various sizes. 7. Passage the tumors at a relatively small size (e.g., 600–1,000 mg), while the animal is still healthy and the tumor is still in exponential growth. Passage at larger sizes increases the chance of passaging an infected tumor.

70

L. Polin et al.

8. Passage mouse tumors in the inbred strain of origin only (F1 hybrids should not be used). 9. Passage mouse tumors in mice that have been in the laboratory for at least 2 months (so that the mice are large and healthy and have well-developed immune systems). These older, larger mice have been exposed to, and have conquered, various extraneous bacteria and viruses. On the other hand, human tumors should be passaged in young immune-deficient mice (6–7 weeks) to minimize development of leaky mice; and also to avoid the possibility of spontaneous lymphomas that will contaminate human tumors (athymic mice and Balb/c SCIDs have less problems with spontaneous lymphomas than ICR SCIDs). 10. The texture and other appearance characteristics of the tumors should be noted. Photos are helpful if the tumor is being used for the first time (this is critical for human tumors which can, and do change easily over time). 11. Be very selective with tumor fragments used to maintain human tumors as xenograft transplants. A chosen fragment should possess a pearly appearance with clearly visible vascularity (red colored areas that include the tumor’s blood vessels). 12. Tumor fragments should be maintained in frozen storage vials in two different locations (10% dimethyl sulfoxide in media with 10% serum is standard). 13. Fragments of the tumor being passaged should be checked for bacterial contamination (incubate at 37°C for 72 h in a growth media). 14. The chemotherapeutic responses to standard agents should be determined. 15. For mouse tumors, only syngeneic models should be used. Verification can be done by treatment with IL-2, which is a simple and inexpensive technique [25]. More classic methods are also available [26]. 16. Simple vital statistics obtained: passage generation; histologic appearance and grade (degree of differentiation); host of origin; tissue and cell of origin; how induced; when induced; time required to reach 500 and 1,000 mg from a trocar implant; unusual characteristics (e.g., mucin-producing; common metastatic sites that vary with implant location), appropriate references. 17. Molecular characterizations of the tumor models will undoubtedly become increasingly critically important to a better understanding of the biologic and response behavior of the models. Since each independently arising tumor is a separate and unique biologic entity, no one tumor model is a perfect predictor for any other model or tissue type. For this reason, most investigators are forced to use multiple models for study [1]. Nonetheless, it is still better for the new investigator to understand a few models very well than many models superficially.

3.10 Summary Clearly, it is difficult to even introduce the topic of solid-tumor models of mice. Many of the details can best be learned by an apprenticeship in a laboratory with an experienced staff, because only a small portion of the practical information and

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

71

reasons for the various methodologies is published. Lastly, many of the issues are not covered at all, or only superficially, because of previous publications. As such, the following references are suggested for additional details in the use of tumor models: [1, 3–11, 21, 23, 27–38]. Nonetheless, the information provided in this work concerning the transplant characteristics of the tumor models, as well as the few examples of data presentation pitfalls, should aid investigators in future readings. With these perspectives, the reader should not be misled into believing that transplantable mouse-tumor models are exceptionally sensitive to everything. In fact, there are very few core molecules with substantial antitumor activity. One should consider the numbers: Worldwide, over 2,000,000 agents have been tested in tumor-bearing animals, and countless additional agents have been tested in culture or other assays to reduce the number to two million. Only about 60 agents are currently available for clinical use, of which less than 15 are used with regularity. Even fewer have substantial activity against common solid tumors of humans. Likewise, in our experience, fewer than 15 agents (clinically available and investigational) from separate chemical cores have a high antitumor ++++ rating against more than one of the highly metastatic, highly invasive, high-take-rate syngeneic solid-tumor models of mice (Panc-02; Mam-16/C; Mam-44; Colon-51; Colon-26; B16 melanoma; LL Carcinoma); with the implant more than 3 logs above the takerate and the drug and tumor injected by different routes, or both (Table 3.1) [1, 4, 6, 7, 32, 34, 35] and unpublished results. It should be noted that a ++++ rating only requires >2.8 log-kill (gross); see Sect. 3.11 for rating scale.

3.11 Methods The methods of protocol design, drug treatment, toxicity evaluation, data analysis, quantitation of tumor-cell-kill, tumor-model systems, cross-resistance behavior, and the biological significance of the drug treatment results with transplantable tumors have been presented previously [1, 3, 9, 26–29, 31, 33–37]. A brief description of the methods as they apply to this work follows.

3.11.1 Tumor Maintenance Tumors were maintained in the mouse strain of origin and were transplanted into the appropriate F1 hybrid or inbred for the trials. Individual mouse body weights for each experiment were within 5 g, and all mice were more than 18 g at the start of therapy. The mice were supplied food and water ad libitum.

3.11.2 Origins of Mouse Tumors Used or Discussed Colon Adenocarcinomas-26,38,51 [5, 6], Mammary Adenocarcinoma-16/C [4], Mammary Adenocarcinoma-16/C/Adr [37, 39], Mammary Adenocarcinomas-17/0

72

L. Polin et al.

and 17/Adr [39, 40], Mammary Adenocarcinoma-16/C/Taxol (induced in vivo with repeated Taxol treatments over six passage generations, first usage published in Ref. [32]), Pancreatic Ductal Adenocarcinomas-02,03 [7], Squamous-Cell Lung LC-12 [41, 42], B16 melanoma (first chemotherapeutic treatment) [43], LLC (first chemotherapeutic treatment) [43], L1210 Leukemia [44], P388 Leukemia [45], AML-C1498 leukemia (arose spontaneously in a 10-month-old C57B1/6 female mouse at Jackson Memorial Labs in 1941 [46].

3.11.3 Chemotherapy of Leukemias: L1210/0 The animals were pooled, IV-implanted with counted numbers of leukemic cells (106, 104, 102) prepared from a spleen brei containing approximately 80% replacement of the spleen with leukemic cells (as judged by the eightfold enlarged size in the leukemic passage DBA/2 mice). Survival was the end point; with moribund mice sacrificed. Cause of death was verified by necropsy, and there were no drug deaths among the groups shown in Table 3.13. In all cases, treatment was started the day after implantation of the leukemic cells.

3.11.4 X-Irradiation of Solid Tumors X-irradiation was delivered through a Picker Vanguard X-ray unit operated at 280 kVp, 20 mA, 1.33 mm Cu HVL, and TSD of 47.5 cm. Anesthetized mice were irradiated (1.1 gy/min, 1 gy = 100 rads) through a treatment port of 1.2 × 4.1 cm that included the tumor site, axillary nodes, and 75% right-lung volume. Maximum scatter and leakage radiation to the shielded animal was less than 2.0%.

3.11.5 Chemotherapy of Solid Tumors The animals were pooled, SC-implanted bilaterally with 30–60 mg tumor fragments by a 12-gauge trocar, and again pooled before unselective distribution to the various treatment and control groups. For early-stage treatment, chemotherapy was started within 1–3 days after tumor implantation, when the number of cells was relatively small (107–108 cells). For more advanced staged disease, the tumors were allowed to reach 100–800 mg (depending on the trial) before the start of treatment. Tumors were measured with a caliper twice weekly. Mice were sacrificed when their tumors reached 1,500 mg. Tumor weights are estimated from two-dimensional measurements:

( )

Tumor weight (mg) = a b 2 ⁄ 2

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

73

where a and b are the tumor length and width in (mm), respectively. In those cases in which bilateral SC tumors were used, both tumors on each mouse were added together, and the total mass per mouse was used for calculations of end points. Note: Some investigators present data as measurements (partly because tumor masses are obtained by measurements). However, all NCI protocols of the past have mandated the use of tumor weights in mg. We have often found that readers do not relate to measurements, and in some cases the measurements obscure the fact that unusually small tumors are being shown and end points are obtained at sizes that are not acceptable in standard NCI protocols.

3.11.6 End Points Assessing Antitumor Activity Solid Tumors The following quantitative end points are used to assess antitumor activity.

3.11.6.1 Tumor-Growth Delay (T–C Value) T is the median time (in days) required for the treatment group tumors to reach a predetermined size (e.g., 1,000 mg), and C is the median time (in days) for the control group tumors to reach the same size. Tumor-free survivors are excluded from these calculations (cures are tabulated separately). In our judgment, this value is the single most important criterion of antitumor effectiveness because it allows the quantitation of tumor-cell-kill.

3.11.6.2 Percent Increase Life Span For leukemic mice only: Percent Increase Life Span (%ILS) = [(T - C)/C] × 100, where C is the median day of death of the control group and T is the median day of death of the treated group. There were no cures in any of the trials shown herein.

3.11.7 Calculation Tumor-Cell-Kill For subcutaneously growing tumors, and leukemic survival trials, the log10 cell-kill is calculated from the following formula: Thelog10 cell − kill total (gross) =

T − C value in days (3.32)(Td )

74

L. Polin et al.

Thelog10 cell - kill total (net ) =

(T − C ) − (duration of treatment ) (3.32)(Td )

The log10 cell-kill/dose = log-kill gross/number of injections (if each injection has the same dose). For solid tumors, the T–C is the tumor-growth delay described here, and Td is the tumor-volume doubling time (in days), estimated from the best-fit straight line from a log-linear growth plot of the control group tumors in exponential growth (100–800 mg range). The conversion of the T–C values to log10 cell-kill is possible because the Td of tumors regrowing post-treatment (Rx) approximates the Td values of the tumors in untreated control mice. If the regrowing tumors do not (e.g., irradiation of tumor and tumor bed), the Td value is determined from the regrowing tumors. This equation for log-kill is also used for leukemia testing, where T is the median day of death for the treated group and C is the median day of death for the control group. The Td is determined from differences in the median days of death of the titered control groups. Four log10 dilution controls were included in all trials. There were five mice each in all treatment and control groups.

3.11.8 Activity Rating Solid Tumors For comparison of activity with standard agents and comparisons of activity between tumors, the log10 kill values were converted to an arbitrary activity rating [35]. It should be noted that +++ or ++++ activity would be required to produce tumor regressions in most models. As such, ++ activity would not be scored by a clinician as more than stable disease. Duration of treatment 5–20 days Antitumor activity Gross log10 tumor-cell-kill Highly active ++++ +++ ++ + Inactive

>2.8 2.0–2.8 1.3–1.9 0.7–1.2 <0.7

3.11.8.1 Activity Rating for Leukemia L1210/0 Because the L1210 leukemia was markedly more sensitive to the antitumor agents than the solid tumors, an expanded activity rating was required (Tables 3.1 and 3.13).

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

75

Duration of treatment 1–9 days Antitumor activity Gross log10 tumor-cell-kill Highly active +++++ ++++ +++ ++ + Inactive

³ 5.0 4.0–4.9 3.0–3.9 2.0–2.9 1.0–1.9 <1.0

3.11.9 Non-quantitative Determination Antitumor Activity Tumor-Growth Inhibition (T/C Value) The treatment and control groups are measured when the control-group tumors reach approximately 700–1,200 mg in size (median of group). The median tumor weight of each group is determined (including zeros). The T/C value in percent is an indication of antitumor effectiveness. A T/C value equal to or less than 42% is considered significant antitumor activity by the Drug Evaluation Branch of the Division of Cancer Treatment (NCI). A T/C value <10% is considered to indicate highly significant­ antitumor activity, and is the level used by NCI to justify a clinical trial if toxicity­, formulation, and certain other requirements are met (termed DN-2 level activity). A body-weight loss nadir (mean of group) of greater than 20% or greater than 20% drug deaths is considered to indicate an excessively toxic dosage. Acknowledgments  This work was supported by CA12623, CA82341, CA53001, Sanofi-Aventis previously Sanofi Synthelabo Research, and Eli Lilly Corporation.

References 1. Corbett T, Valeriote F, LoRusso P, Polin L, Panchapor C, Pugh S, et al. Tumor models and the discovery and secondary evaluation of solid tumor active agents. Int J Pharmacogn. 1995;33(Suppl.):102–22. 2. LoRusso P, Demchik L, Dan M, Polin L, Gross JL, Corbett TH. Comparative efficacy of DMP840 against mouse and human solid tumor models. Invest New Drugs. 1995;13:195–203. 3. Polin L, Valeriote F, White K, Panchapor C, Pugh S, Knight J, et  al. Treatment of human prostate tumors PC-3 and TSU-PRI with standard and investigational agents in SCID mice. Invest New Drugs. 1997;15:99–108. 4. Corbett TH, Griswold DP, Jr, Roberts BJ, Peckham JC, Schabel FM, Jr. Biology and therapeutic response of a mouse mammary adenocarcinoma (16/C) and its potential as a model for surgical adjuvant chemotherapy. Cancer Treat Rep. 1978;62(10):1471–88. 5. Corbett TH, Griswold DP, Jr, Roberts BJ, Peckham JC, Schabel FM, Jr. Tumor induction relationships in development of transplantable cancers of the colon in mice for chemotherapy assays, with a note on carcinogen structure. Cancer Res. 1975;35:2434–9.

76

L. Polin et al.

6. Corbett TH, Griswold DP, Jr, Roberts BJ, Peckham JC, Schabel FM, Jr. Evaluation of single agents and combinations of chemotherapeutic agents in mouse colon carcinomas. Cancer. 1977;40(5):2660–80. 7. Corbett TH, Roberts BJ, Leopold WR, Peckham JC, Wilkoff LJ, Griswold DP, Jr, et  al. Induction and chemotherapeutic response of two transplantable ductal adenocarcinomas of the pancreas in C57BL/6 mice. Cancer Res. 1984;44:717–26. 8. Corbett TH, Valeriote FA. Rodent models in experimental chemotherapy. In: Kallman RF, editor. The use of rodent tumors in experimental cancer therapy: conclusions and recommendations. New York, NY: Pergamon; 1987. p. 233–47. 9. Corbett T, Valeriote F, LoRusso P, Polin L, Panchapor C, Pugh S, et al. In vivo methods for screening and preclinical testing; use of rodent solid tumors for drug discovery. In: Teicher B, editor. Anticancer drug development guide: preclinical screening, clinical trials, and approval. Totowa, NJ: Humana Press Inc.; 1997. p. 75–99. 10. Corbett TH, Griswold DP, Jr, Roberts BJ, Schabel FM, Jr. Cytotoxic adjuvant therapy and the experimental model. In: Stoll BA, editor. New aspects of breast cancer. Vol. 4, Systemic therapy in breast cancer. London: William Heinemann Medical Books; 1981. p. 204–43. 11. Corbett TH, Roberts BJ, Lawson AJ, Leopold WR, III. Curative chemotherapy of advanced and disseminated solid tumors of mice. In: Jacobs JR, Al-Sarraf M, Crissman J, Valeriote F, editors. Scientific and clinical perspectives in head and neck cancer management: strategies for cure. New York, NY: Elsevier Scientific; 1987. p. 175–92. 12. Griswold DP, Jr, Corbett TH. Use of experimental models in the study of approaches to treatment of colorectal cancer. In: Lipkin M, Good RA, editors. Gastrointestinal tract cancer. New York, NY: Medical Book; 1978. p. 399–418. 13. Karrer K, Humphreys SR. Continuous and limited courses of cyclophosphamide (NSC26271) in mice with pulmonary metastasis after surgery. Cancer Chemother Rep. 1967;51:439–49. 14. Schabel FM Jr. Concepts for treatment of micrometastases developed in murine systems. Am J Roentgenol Radium Ther Nucl Med. 1976;126:500–11. 15. Boyd MR. The NCI in vitro anticancer drug discover screen, concept, implementation, and operation, 1985 to 1995. In: Teicher B, editor. Anticancer drug development guide: preclinical screening, clinical trials, and approval. Totowa, NJ: Humana; 1997. p. 75–99. 16. Wallace AC. Effect of delayed addition of irradiated cells to small viable tumor inocula. Cancer Res. 1965;25:355–7. 17. Revesz L. Effect of tumor cells killed by X-rays upon growth of admixed viable cells. Nature. 1956;178:1391–2. 18. Revesz L. Effect of lethally damaged tumor cells upon the development of admixed viable cells. J Natl Cancer Inst. 1958;20:1157–86. 19. Dykes DJ, Griswold DP, Jr, Schabel FM, Jr. Growth support of small B16 melanoma implants with nitrosourea-sterilized fractions of the same tumor. Cancer Res. 1978;36:2031–4. 20. Annual Progress Report to Division of Cancer Treatment, National Cancer Institute on Primary Screening and Development and Application of Secondary Evaluation Procedures for Study of New Materials with Potential Anticancer Activity. Lack of Influence of immune suppression on the number of cells required to establish takes of solid tumors. Southern Research Institute, Contract NO1-CM-43756, March 15, 1982. p. 1–9. 21. Skipper HE. The effects of chemotherapy on the kinetics of leukemic cell behavior. Cancer Res. 1965;25:1544–50. 22. Laster WR, Jr, Schabel FM, Jr, Skipper HE, Wilcox WS, Thomson JR. Experimental evaluation of potential anticancer agents IV. Host weight loss as it related to false positives in drug evaluation. Cancer Res. 1961;21:895–906. 23. Annual Progress Report to Division of Cancer Treatment, National Cancer Institute on Primary Screening and Development and Application of Secondary Evaluation Procedures for Study of New Materials with Potential Anticancer Activity. Section 27. Unusual toxicity problems ­encountered in the evaluation of new antitumor agents: implications for data analysis and protocol­ design. Southern Research Institute, Contract NO1-CM-43756, March 15, 1982. p. 1–32.

3  Transplantable Syngeneic Rodent Tumors: Solid Tumors in Mice

77

24. Houghton JA, Taylor DM. Growth characteristics of human colorectal tumors during serial passage in immune-deprived mice. Br J Cancer. 1978;37:213–23. 25. LoRusso PM, Polin L, Aukerman SL, Redman BG, Valdivieso M, Biernat L, et al. Antitumor efficacy of Interleukin-2 alone and in combination with Adriamycin and Dacarbazine in murine solid tumor systems. Cancer Res. 1990;50:5876–82. 26. Motycka K, Bostik J, Danek PF, Chudomel V. Immunogenicity of the Gardner lymphosarcoma for the mice of the strain C3H (H-2k). I. The effect of the 60Co-irradiation of recipients and of the interval between the immunization and transplantation of the tumor on the antitumor resistance of immunized mice. Neoplasma 1986;33(2):167–75. 27. Simpson-Herren L, Corbett TH, Griswold DP, Jr. The cell population kinetics and response to an S-phase specific agent of three transplantable colon tumor lines. Cell Tissue Kinet. 1980;13:613–24. 28. Corbett TH, Leopold WR, Dykes DJ, Roberts BJ, Griswold DP, Jr, Schabel FM, Jr. Toxicity and anticancer activity of a new triazine antifolate (NSC-127755). Cancer Res. 1982;42:1707–15. 29. Corbett TH, Roberts BJ, Trader MW, Laster WR, Jr, Griswold DP, Jr, Schabel FM, Jr. Response of transplantable tumors of mice to anthracenedione derivatives alone and in combination with clinically useful agents. Cancer Treat Rep. 1982;66(5):1187–200. 30. Corbett T, Valeriote F, Baker L. Is the P388 murine tumor no longer adequate as a drug discovery model? Invest New Drugs. 1987;5:3–20. 31. Corbett TH, Bissery M-C, LoRusso P-M, Polin L. 5-Fluorouracil containing combinations in murine tumor systems. Invest New Drugs 1989;7:37–49. 32. Corbett TH, Valeriote FA, Demchik L, Polin L, Panchapor C, Pugh S, et  al. Preclinical ­anticancer activity of cryptophycin-8. J Exp Ther Oncol. 1996;1:95–108. 33. Corbett TH, Valeriote FA, Demchik L, Lowichik N, Polin L, Panchapor C, et al. Discovery of cryptophycin-1 and BCN-183577: examples of strategies and problems in the detection of antitumor activity in mice. Invest New Drugs. 1997;15:207–18. 34. Corbett TH, LoRusso P, Demchick L, Simpson C, Pugh S, White K, et al. Preclinical antitumor efficacy of analogs of XK-469: Sodium-(2-[4-(7-chloro-2-quinoxalinyloxy)phenoxy] propionate. Invest New Drugs. 1998;16:129–39. 35. Corbett TH, Panchapor C, Polin L, Lowichik N, Pugh S, White K, et al. Preclinical Efficacy of thioxanthone SR-271425 against transplanted solid tumors of mouse and human origin. Invest New Drugs. 1999;17:17–27. 36. Schabel FM, Jr, Trader MW, Jr, Laster WR, Jr, Corbett TH, Jr, Griswold DP, Jr. Cisdichlorodiammineplatinum (II): combination chemotherapy and cross-resistance studies with tumors of mice. Cancer Treat Rep. 1979;63:1459–73. 37. 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. 38. Skipper HE, Schabel FM, Jr, Wilcox WS. Experimental evaluation of potential anticancer agents XIII: on the criteria and kinetic associated with curability of experimental leukemias. Cancer Chemother Rep. 1964;35:3–111. 39. Kessel D, Corbett TH. Correlations between anthracycline resistance, drug accumulation and membrane glycoprotein patterns in solid tumors of mice. Cancer Lett. 1985;28:187–93. 40. Biernat L, Polin L, Corbett T. Adaptation of mammary tumors of mice to a soft agar assay for use in drug discovery. Seventh NCI-EORTC Symposium on New Drugs in Cancer Therapy. Abstract #185, Amsterdam, March 1992. p. 17–20. 41. Smith WE, Yazdi E, Miller L. Carcinogenesis in pulmonary epithelia in mice on different levels of Vitamin A. Environ Res. 1972;5:152–63. 42. Tapazoglou E, Polin L, Corbett TH, Al-Sarraf M. Chemotherapy of the squamous cell lung cancer LC-12 with 5-fluorouracil, cisplatin, carboplatin or iproplatin combinations. Invest New Drugs. 1988;6(4):259–64. 43. Sugiura K, Stock CC. Studies in a tumor spectrum. III. The effect of phosphoramides on the growth of a variety of mouse and rat tumors. Cancer Res. 1955;15:38–51.

78

L. Polin et al.

44. Law LW, Dunn TB, Boyle PJ, Miller JH. Observations on the effect of folic-acid antagonists on transplantable lymphoid leukemias in mice J Natl Cancer Inst. 1949;10:179–92. 45. Dawe CJ, Potter M. Morphologic and biologic progression of a lymphoid neoplasm of the mouse in vivo and in vitro (abstr). Am J Pathol. 1957;33:603. 46. Bradner WT, Pindell MH. Myeoid leukemia C1498 as a screen for cancer chemotherapeutic agents Cancer Res. 1966;April(pt 2):375–90.

Chapter 4

B16 Murine Melanoma: Historical Perspective on the Development of a Solid Tumor Model Enrique Alvarez

The determination of weight of a factor in producing metastases can not be judged from single experiences on man, as it is impossible to eliminate conflicting conditions. Only by the use of a homogeneous material which the size of the cells, their histological and biological qualities, and the vascularity of the surrounding tissue, etc., are practically constant can valid conclusions be drawn, and this elimination of variables is possible to obtain only by the use of animal tumors of a long transplanted strain, so that the morphological and biological characters are well known. Dr. Leila C. Knox, 1922 [1]

4.1 Introduction The development of reliable models of disease mechanisms largely depends on our understanding of the characteristic processes of the disease being modeled. Syngeneic tumors in mice offer an important model for oncology research. Although the use of murine models of neoplastic disease has been raised to the level of a fundamental paradigm in oncology research, it is to be considered with all the care and diligence afforded to us by any biological model. Syngeneic murine tumors offer potential advantages as well as limitations, which should always be present in the mind of the investigator. Every year, countless studies are published describing new models and/or new cell lines available to scientists. Many of these are highly specialized, and have a narrow application to the broad community. Through time, several models are developed, which present us with a newer tool to truly advance our understanding. In many instances, these models fill a specific unmet need. If fortunate enough, the model is also relatively easy to reproduce, thus providing for rapid dissemination among the community. The overall relative importance of an experimental model depends largely on two important aspects.

E. Alvarez (*) Biomodels, 313 Pleasant Street, Watertown, MA 02472, 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_4, © Springer Science+Business Media, LLC 2011

79

80

E. Alvarez

First, how does the model recapitulate the process it attempts to emulate? Second, how does the model itself offer the flexibility to expand our knowledge relating to the pathologic process being evaluated? This chapter pinpoints how the B16 murine melanoma line has proven itself a valuable model on both points. This tumor line provides researchers with the ability to model the process of solid tumor formation and the following metastatic process seen in animals and man. Upon establishment of this model of metastasis, a more detailed understanding of the steps involved in tumor dissemination was gained. Astute manipulation of the cell line and a rational use of experimental animal models were essential for this process. Overall, this effort has helped the modern description of fundamental processes in metastasis, invasion, and antitumor drug development. The body of knowledge derived from the B16 murine melanoma line is ample, and even at present continues to grow. The emphasis of this overview is limited to its origin and to the seminal work attributable to the early work done with this tumor line and some of the tumor models that have arisen from it. In the span of biological research, this model is relatively new. But within one lifetime, it has helped shape our understanding of oncology. In science, many posed questions are old and numerous, but the tools we need to thoroughly explore them and ask new ones are in continuous evolution. These new tools encourage even more questions. The B16 melanoma is one of those tools.

4.2 Historical Context The process of scientific discovery never occurs outside the context of contemporary knowledge. Contemporary knowledge is relative to the time of the work itself. In the early 1970s, at the time of the establishment of the B16 as a model for metastasis research, the state of metastasis research from a technical standpoint was immature when compared to today’s standards. Although many recent developments in the area of metastasis have depended on the use of relatively recent technological advances (e.g., molecular biology, protein chemistry, and bioinformatics), many questions regarding the nature of metastasis are very old. Up to the early 1970s, there had been numerous qualitative (i.e., autopsies) and some quantitative studies (i.e., rodent models) relating to the natural history of the metastatic process, but a readily accessible murine model was lacking – specifically, a model that would offer a predictable metastatic pattern. A periodic survey of the historical record serves to focus our attention to the correct context of the research. Within this context, the significance of the questions asked becomes clearer, as well as the importance of the techniques being used. The development and current use of the B16 melanoma line should be incorporated into a larger scope of oncology research by looking at a previous generation of researchers and their professional contributions to the field. An often-quoted work by Dr. Stephen Paget was seminal in establishing our appreciation for the complexity of the processes involved in metastasis [2]. The concept of Seed and Soil, as it applies to the tumor embolus and its potential site of

4  B16 Murine Melanoma

81

invasion, is firmly entrenched in the minds of researchers in the field of metastasis. Simple, allegorical, yet effective at setting the concepts of a complicated pathological process, the idea of seed and soil has been fundamental for over 110 years. The careful evaluation of human autopsies derived from breast-cancer patients and obvious nonrandom pattern of metastasis, noted by Paget, demonstrated that breast-cancer patients had a notable predisposition for bony metastasis. This bony metastasis was nonrandom in its distribution, since, as Dr. Paget commented: “Who has ever seen the bones of the hands or the feet attacked by secondary cancer?” This observation is critical in two important aspects. First, the bony tissue clearly presents the tumor with a preferred invasion site (soil to the seed). Second, bones are not just bones; there appears to be further discrimination by the tumor embolus as to which bone offers an optimal site for colonization. Effective tools, such as animal models that could be used to experimentally understand the human condition, were not yet available to Paget. It would still be many years until such specific disease models were widely available for use. Our contemporary description of the specific relationship between the cell and its host tissue uses the term “tumor microenvironment” to essentially name the same phenomena described in 1889. Successful tumor metastasis is a complex, nonrandom, multifactorial event that requires contributing components from both the cancer cell and its host. As with many areas in science, in oncology we also find a periodic re-evaluation of fundamental themes. The “Seed and Soil” concept has not been immune to this effect. In 1982, Hart specifically reflected on the mechanisms of metastasis as they apply to murine models [3]. This work reiterated the clear importance of circulation and physical-cell distribution on the outcome of a metastatic event. Hart concluded: Patterns of metastasis primarily appear to be a direct consequence of the delivery of an optimal dose of tumor cells to the first organ encountered along the lymphatic or venous pathway. Nonetheless, the very same tumors that use this mechanism as a predominant mode of spread will, on other occasions, exhibit true organ tropism.

This conclusion did not exclude the already noted metastatic preference of tumors to certain organs, but gave a stronger importance to circulatory parameters in the final disposition of tumor metastasis. This attempt at revisionism occurred almost 100 years after the publication by Paget. What had transpired in that interval to call for the reevaluation of the concept? In the first half of the 1900s, there was a general effort to describe the nature of circulating tumor cells in man and rodent models. In 1913, while studying a carcinoma model of the Japanese waltzing mouse, Tyzzer found a correlation between the tumor size, duration of tumor presence, peculiar conditions furnished by the host, and the number of metastases [4]. These important clinically applicable correlations were being laid down by scientists using syngeneic models. In 1915, Iwasaki presented a paper, which when carefully read can serve to introduce today’s researcher to many of the important areas of current interest [5]. By using microscopy to describe the tumor embolus–host interaction in necropsy cases, Iwasaki finely described the disposition of several tumor types in the vessels of patients. Most of the work presents the reader with a description of the relationship

82

E. Alvarez

between the tumor embolus, thrombus formation/organization, and the formation of tumor metastasis. A particular description of tumor–vessel interaction by Iwasaki is of particular interest today: The tumour cell mass is here covered by a thin layer of endothelium continuous with the intima. Though a small number of leucocytes may be found between the tumour cells, the latter appeared to be normal: no sign of degradation can be seen. Some connectivetissue fibres from the wall, as well as cells from the covering endothelium, are thrust into the mass: these serve as stroma. I consider them as the first stage of metastases formation; at the same time, I consider that the penetration of the wall by the tumor cells from inside can sometimes occur without such an endothelial cover.

Eighty-five years ago we find an accurate description of the process of extravasation by a metastatic cell, while at the same time describing a relationship between tumor cells and endothelium. In the same paper, Iwasaki also studied the fate of intravenously (iv)-injected tumor cells in rats and mice: In the cases of mouse carcinoma I again observed the appearances of which I have described in cases of human sarcoma, that is, the tumour cell group on a vessel wall, having a single endothelial layer directly covering it. The carcinoma cells of such fixed emboli appear to be in healthy condition in every respect, and in the cases where connective tissue bundles thrust into and divide them up, they have the common appearance of alveolar carcinoma and are not to be considered as degenerating or organizing at all.

This statement serves to offer initial validation of the use of murine models from a comparative and histological perspective. When comparing his own tumor implantation results with that of fellow researchers (who obtained a lower tumor take-rate than himself) Iwasaki concluded: From my results I hold that it is not difficult to cause tumours by inoculating foci directly into vessels. Nevertheless, all tumor cells introduced into blood vessels do not necessarily disintegrate, as several authors believe, but if the technique is suitable, the lung will be attacked in a very considerable number of cases.

Iwasaki’s work was important in describing the lung as the preferred site of tumortakes after iv injection. Much work followed to determine the reasons for this apparent affinity. Warren and Gates clearly demonstrated that the successful establishment of tumors in rodent lungs after iv injection is strongly correlated with the cell viability of the sample being used [6]. At that time, the authors were attempting to standardize the procedure of iv tumor-cell implantation. This was demonstrated using the Walker 256 carcinoma line in rats. Cloudman in 1947 studied the – as they were formerly called – organophilic tendencies of murine hepatic tumors [7]. At that time, the author specifically looked at the dissemination patterns of two tumor lines (C954 liver carcinoma and C198 reticuloendothelioma). The work employed the use of subcutaneous (SC) implantations using a trochar as well as the use of parabiosis in mice. Interestingly, Cloudman’s method of tumor passage may have contributed to the organophilic tendencies noted. The mode of tumor maintenance for the C198 line required that tumor-affected pieces of liver from mice be SC-implanted in naive hosts. This SC tumor would lead to disseminated disease in the mice. When the initial line was being established, simple SC implantation led

4  B16 Murine Melanoma

83

to poor tumor-take. But the hepatic metastatic spread of the tumor offered an optimal tissue for reproducibly metastatic tumor whenever this tissue was SC-implanted. In the case of the C954 tumor line, it was maintained via simple SC passage. The C954 had no metastatic properties. Retrospectively, it appears that Cloudman unintentionally selected for a tumor-line variant by only using the tumor stock from metastatic sites. The research compared the metastatic potential of the C198 line vs. the nonmetastatic C954 line. In addition to a direct comparison between both tumor lines, by using various mouse strains, there was a demonstration of the different metastatic patterns (organophilic tendencies) attributable to the mouse strain being used. This led to the conclusion: The appearance of tumor metastasis within a specific internal organ is probably dependent upon a host–tumor interrelationship rather than upon either the tumor type or the host type alone.

This is an important holistic approach to in vivo modeling. But strictly speaking, it was probably not a comparison of equal lines. The successful establishment of the tumor is attributed to both the tumor cell and the selected host. In 1949, Coman et al. presented work, which in the context of Paget’s hypothesis, discussed the apparent inability of the V2 rabbit carcinoma to metastasize to muscle [8]. By injecting tumor cells into the arterial circulation and demonstrating tumor growth in the muscle mass, the authors confirmed the ability of the tumor to recognize the tissue as favorable. In their conclusion, the reason for the apparent lack of tumors in certain organs was the result of a filtering effect by the lungs. In 1950 and 1952, Ziedman studied important parameters for the use of a murine tumor line in vivo at the University of Pennsylvania [9, 10]. Ziedman specifically studied the relationship between the number of viable tumor emboli iv-implanted and the total number of resulting metastatic nodules. In 1950, using the mouse Sarcoma 241 line, Ziedman concluded: That the number of metastases is directly proportional to the number of viable tumor cell emboli released into the circulation. The longer a primary tumor existed the greater the number of emboli released, as judged by the number of metastases appearing in the lungs [9].

In his studies, the author also compared two rabbit tumor lines (V2 squamous-cell carcinoma, Brown-Pearce carcinoma) and one rat tumor line (Walker 256 carcinoma). One matter that still had to be resolved regarding iv injections of tumor cells relates to the pulmonary disposition of tumor cells. Were the lungs acting as a mere sieve? To test this hypothesis, rabbits received an iv injection of tumor cells, while aortic outflow was captured. This collected blood was then iv-injected into a naive host; this second animal was then followed for tumor formation. Tumor formation in the second hosts demonstrated that tumor cells successfully passed through the pulmonary circulatory system. All three tumors were able to pass through the pulmonary vasculature and form tumors in the second host, thus negating the pulmonary sieve hypothesis. In 1961, Ziedman used microcinematic techniques to further evaluate the flow patterns of tumor cells in circulation [11]. In this work, he demonstrated the higher

84

E. Alvarez

degree of membrane flexibility/deformability of the Brown-Pearce tumor vs. the V2 tumors. This work was done by studying the vascular patterns in the mesenteric vessels of rabbits. It illustrated how the Brown-Pearce cells, upon reaching an arteriolocapillary junction, could deform and pass more freely than the V2 cells. This added to the body of evidence for the tumor’s enhanced ability for transpulmonary passage, thus highlighting an intrinsic cellular difference between both lines. Descriptive pathology and in vivo work, as many of the works presented here have been, continue to be essential for the development of our understanding of metastasis. Even at the present time, in the era of genomics, an attempt to elucidate the underlying biological forces involved in the dissemination of prostatic carcinoma required the careful evaluation of autopsies over 19,000 patients [12]. Approximately 1,500 cases of prostatic carcinoma were identified in this cohort. Of those cases, 35% showed evidence of metastatic spread. Interestingly, over 100 years after the seminal work of Dr. Paget, this paper offers an attempt to describe in detail the common metastatic patterns of dissemination found in prostatic carcinoma patients using autopsy records.

4.3 B16 Melanoma In the context of the Seed–Soil hypothesis, the use of the B16 murine melanoma line as a model for both solid-tumor formation and metastasis was an important development in oncology research. Although the introduction of this model for metastasis research can be traced to 1970, the cell line itself had been identified and characterized as a tumorigenic line years before. The B16 murine melanoma cell line originated in 1954. The tumor spontaneously arose in a C57BL/6J mouse at the Jackson Laboratories in Maine. There is no record of the sex of the originating mouse. The initial neoplastic lesion arose in the skin at the base of the ear. The following is a histological description of the tumor from the Handbook on Genetically Standardized Jax Mice by Dr. Earl Green: Gross: soft gray tissue, frequently hemorrhagic. Microscopic: tumor cells polyhedral or spindle-shaped, arranged in perivascular mantles and diffuse masses; some cells contain fine pigmented granules, a few are obscured by large, very dark globules of pigment; stoma delicate and vascular. Pigment greatly decreased in comparison with early-transplant generation [13].

This offers the first histological description of the tumor line in its host. The tumor line was maintained at the Jackson Laboratories by continuous passage in vivo. At the time of the report, the tumor is described as metastatic to lung, liver, and spleen. This metastatic pattern was present after SC implantation. Using the B16 melanoma line from the Jackson Laboratories, Dr. Isaiah Fidler carefully documented the final disposition of the melanoma cells after iv injection in mice in 1970 [14]. A number of protocols used before this time to aid in the understanding of the metastatic process proved to be either cumbersome, unreliable, or both. There was an unmet need for a mouse line, and this line offered an important opportunity for the field. In the 1970 study, B16 murine melanoma cells

4  B16 Murine Melanoma

85

were initially cultured in  vitro. The cells used for iv injection were labeled with 125 I-5-iodo-2-deoxyuridine. This radioactive label provided a clear and specific way to monitor the organs in which tumor cells arrested. This ability to monitor cells was specifically caused by the affinity of the radioactive label for viable cells. Cell death would lead to the excretion of the label from the animal, thus precluding the possibility of labeling the host’s tissues. The iv injection of killed labeled B16 cells served to demonstrate the inability of the host to reutilize the radioactive label, thus preventing a false-positive reading. In vitro, 200,000 labeled cells would produce an average of 40,000 counts/min. After inoculating the cells into naive hosts, the number of viable tumor cells in each organ was determined from radioactivity count, by the use of the ratio of cpms to cells in the original inoculum. Simple and efficient, the labeling of B16 murine melanoma cells could now open the door for the description of a complex tumor–host interaction. The B16 melanoma cells could now be iv-implanted and followed in the host animal, providing a faster, less cumbersome method than the previously used histopathological approaches. Tables 4.1 and 4.2 from the 1970 paper show the final organ of cellular arrest and temporal distribution of the viable/dead and labeled B16 murine melanoma cells after a single iv injection [14]. From the table, it can be easily established that in the earliest postinjection time-points, the majority of the cells find themselves in Table 4.1  Fate of 125IUDR-labeled tumor emboli organ distribution of 200,000 melanoma cells injected intravenously into C57BL/6J mice Number of cellsa Time of death Lung Liver Spleen Kidney Bloodb 1 min 2 min 5 min 7 min 10 min 15 min 30 min 45 min 1 h 2 h 4 h 8 h 12 h 1 day 2 days 3 days 7 days 14 days

136, 750 128, 500 106, 700 103, 500 130, 500 122, 800 117, 900 105, 000 100, 550 89, 590 46, 700 18, 000 5, 500 1, 700 610 450 450 400

2, 230 5, 500 7, 600 7, 570 9, 350 8, 390 4, 260 4, 140 3, 570 5, 830 5, 300 1, 340 700 600 260 200 230 0

200 230 260 280 270 310 250 390 330 680 870 320 580 140 160 90 40 0

Mean number of cells (20 mice per time interval) 1.0 cc of blood c Urinary bladder and contained urine a

b

300 270 270 290 310 310 370 460 370 530 770 300 230 130 130 20 0 0

3, 750 1, 590 2, 200 2, 300 2, 600 3, 340 3, 500 3, 900 3, 800 4, 300 4, 660 4, 500 1, 050 580 140 40 0 0

125

IUDR-labeled

Urinec – – – – 100 350 4,300 4,100 10,000 7, 000 33,900 23,500 10,500 1,700 1,400 20 0 0

86

E. Alvarez Table 4.2  Fate of 125IUDR-labeled tumor emboli organ distribution of 200,000 125IUDRlabeled melanoma cells injected intravenously into C57BL/6J mice Number of cellsa Time of death Lung Liver Spleen Kidney Bloodb Urinec 1 min 96, 700 4, 860 40 120 2, 314 – 2 min 97, 250 5, 420 100 570 845 – 5 min 99, 600 4, 810 90 120 1, 190 – 10 min 91, 700 5, 700 300 270 2, 070 345 15 min 64, 400 4, 000 220 200 2, 600 380 30 min 41, 500 9, 300 420 590 5, 600 3, 530 45 min 19, 500 12, 430 900 930 6, 790 20, 060 1 h 2, 100 8, 800 940 740 9, 340 25, 000 2 h 550 1, 700 450 435 13, 200 24, 500 4 h 200 2, 310 640 360 4, 010 22, 290 8 h 0 750 330 0 1, 950 7, 260 24 h 0 40 120 0 100 190 2 days 0 30 40 0 0 120 7 days 0 0 0 0 0 60 14 days 0 0 0 0 0 0 Mean number of cells (10 mice per time interval) 1 cc of blood c Urinary bladder and contained urine a

b

the pulmonary tissue, but some are also localized in other organs. After 14 days, only the lungs contain labeled cells, now seen as tumor nodules. Liver, spleen, kidneys, and blood all showed the early presence of the labeled cells, but none of these tissues show the establishment of tumors at 14 days postinjection. Intravenously-injected B16 melanoma cells in mice could lead to rapid accumulation of cells in the pulmonary tissue, lead to early high levels of tumor cells in circulation caused by transpulmonary passage, or ultimately produce a very low rate of tumor formation in the animals. In the case of the B16 line, this tumor formation was limited to the pulmonary tissue. As a control, the author used killed B16 cells, which had had also been labeled with 125I-5-iodo-2-deoxyuridine. Some of the obvious questions presented by this work include: Why are so few cells if the injected cells able to form tumors? Why such apparent inefficiency? How do the cells select the target organ for colonization? Again, in the case of the B16 tumor, the lungs were not acting as a passive sieve. This work demonstrated that the mere presence of neoplastic cells in the circulation of the mice is not a guarantee of successful tissue colonization by the tumor, confirming earlier work done with mouse Sarcoma 241 tumors in 1950 by Ziedman [9]. The work with the B16 line is a good example of how research is done in the context of an already established scientific framework. But strictly speaking, the process recapitulated in the B16 model is only representative of what happens to neoplastic cells, which have escaped a primary tumor into circulation. At this point in time, the iv model was not truly representative of the

4  B16 Murine Melanoma

87

entire set of steps now established as requirements for a cell to leave the primary tumor and successfully colonize a distal site. In 1973, Dr. Fidler published another paper, which described a model for metastatic neoplasia using B16 melanoma cells iv-injected in syngeneic mice [15]. Until this time, metastatic tumor models usually relied on the desegregation of cells from solid tumors (a heterogeneous mixture of neoplastic and normal cells). This study focused on the relationship between the number of implanted cells, tumor emboli size, and the resulting pulmonary metastatic nodules. Cell selection was aided by the use of in vitro culturing of the cells in order to select the most optimal cell population for implantation. Building upon prior work by Ziedman, the author studied the effects of cell viability and cell clumping in the formation of pulmonary nodules (Tables 4.3 and 4.4). The value of this effort was as important as it was simple. Using the B16 line, the results demonstrate a proportional increase in lung metastasis formation in animals injected with higher numbers of cells. Although this is a rather intuitive point (and one previously tested) it served to solidify a base of knowledge around the B16 as a murine model. In the context of the contemporary studies, this effort established a standard to be followed by contemporary scientists. The careful selection of viable single cells in this model was again shown to be essential for the establishment of a reproducible model of in vivo metastasis. These 1970 and 1973 papers presented what should be considered, an introduction of the B16 tumor line to the scientific community [14, 15]. Importantly, the works themselves clarified an optimal protocol for the establishment of metastatic lesion in the lungs of C57-black mice, using a relatively simple technical procedure. Also published in 1973, the process that gave rise to the B16 (F1, F10) sublines with different metastatic potentials. This was a short two-page paper with important ramifications [16]. The work describes the in vivo–in vitro selection process used Table 4.3  Relationship of the number of viable B16 melanoma cells injected iv in C57 mice with the number of resultant pulmonary metastases Number of viable B16 cells iv-injected Average number of pulmonary metastasesa ±SD 100 1, 000 10, 000 50, 000 100, 000

0.3 (0–1) 12.1 ± 3.4 (9–18) 71.6 ± 16.2 (52–96) 205.5 ± 34.3 (166–260) 394.8 ± 51.6 (338–502)

a Nine mice per each group. Pulmonary metastases were counted on day 14 post iv injection with the aid of a dissecting microscope

Table 4.4  The effect of B16 melanoma embolic size on resultant pulmonary metastases in C57 mice Total number of embolic Number of cells per Average number of resultant B16 cells iv-injected embolic clump pulmonary metastasesa ± SD 50, 000 10, 000–12, 000

1 4–5

11.5 ± 3.4 (5–14) 33.3 ± 8 (21–41)

a Eight mice per each group. Pulmonary metastases were counted on day 14 post iv injection with the aid of a dissecting microscope

88

E. Alvarez

to identify cell variants with a high degree of preference for metastatic growth in the lungs. The experimental protocol started by SC implantation of the tumor cells into mice. The cells would grow into a tumor and spontaneously metastasize. The author then selected a resulting metastatic nodule from the lung. This tumor was dissociated and cultured as a monolayer in vitro. Upon expansion of the selected cells, these were then iv-implanted into a new host. Again, the resulting pulmonary nodules were harvested and cultured in vitro. After five cycles through the selection process, a line was derived. At this point the B16-F10 melanoma line had been specifically selected to metastasize to the lungs after iv injection. Table 4.5 shows the illustration used to explain the selection process. In 1973, Fidler concluded: As the tumour cell viability, size and homogeneity, and the syngeneic recipient did not change from one preparation to another, the differences in metastatic incidence could only be attributed to properties intrinsic to the various tumour cell lines. The clonal selection of tumours from successive metastases apparently results in cells better capable of survival and formation of secondary growths. This indicates that survival of circulating tumour emboli is not a random phenomenon [16].

Fifty years had passed since Dr. Leila Knox had proposed the criteria for an experimental murine model of metastasis. The B16 model system was incorporated into the American Type Culture Collection in June 1978 (Mr. Andrew Redman, personal communication). In the 1983 edition of the Catalog of Transplantable Animal and Human Tumors published by the Division of Cancer Treatment, National Cancer Institute (NCI) (Maryland, USA) a letter by Dr. Fidler is included [17]. This letter approves the distribution of the deposited B16 melanoma lines by the NCI. The catalog shows on the summary sheets for each tumor line that on October 11, 1979 the B16-F1, -F10, -F10LR-6, and -BL-6 were incorporated into the tumor repository. At this point, this important model was available to the entire scientific community. After the demonstration of the lung-colonizing ability of the B16-F10 lines, many in vivo lines were derived, using a similar selection approach. This approach of iv tumor cell administration would be followed for target organ seeding. Those tumors arising at the target organ would be selected and again reimplanted into fresh hosts, this would force a selection process. Examples of this methodology can be found in Brunson et al. [18], Nicolson et al. [19], and Nicolson [20]. These particular examples show how the selected lines would preferentially metastasize to the brains of C57 mice. The braincolonizing lines were shown to invade the meninges or the forebrain (lines were designated B16-B10b or B10n). Line B16-B10n showed preference for cerebral vasculature. By this route, the cells would gain access to the cerebral cortex. In 1979, Brunson and Nicolson selected another variant of the B16 melanoma line. This time the selected lines would invade the ovarian tissue when implanted in mice [21]. An observation was made that the selection process produced a less melanotic cell line than the originating cells. This ovary-colonizing variant was designated the B17-010 line. In 1980, Raz and Hurt used the cytochalasing B and colchicines as selection agents in the B16 melanoma to promote the cytoskeletal changes, which the authors suspected believed to be important in the development of differential patterns of metastases in this particular line [22]. The selected lines (designated B16-F10-B1, B2, and B3) the selected lines in addition to

4  B16 Murine Melanoma

89

Table 4.5  Reproduction of serial lung implantation diagram Schematic representation of the tissue culture and animal transplantation system. The SC B16 melanoma tumor in the syngeneic mouse, C57BI/6J, was adapted to grow in tissue culture as described previously [2, 3]. Confluent monolayers of tumor cells were collected by 2-min treatment of 0.25% trypsin and vigorous shaking. Cell suspensions were diluted to give an inoculum dose of 50,000 viable cells (trypan blue excluding cells) in 0.25 ml Hank’s basic salt solution. Mice were injected iv and were killed with ether 3 weeks later, and submerged in consecutive washes of 7% iodine, 70% ethanol and sterile saline, and then placed in a laminar airflow hood. Their lungs, which contained melanoma nodules, were removed aseptically. Several pulmonary metastases were dissected free of the lungs, gently pressed through a number 70 stainless-steel mesh sieve and filtered through gauze. After centrifugation, the cells were resuspended in supplemented media2, plated in several Petri dishes, and incubated at 37°C, 5% CO2. Three to four days later, small colonies could be observed with the aid of an inverted microscope; one to five colonies with melanin granules were selected in each dish and their position marked. All other cells or colonies were removed by scraping and washed off with media. The selected attached colonies were incubated for an additional 3–5 days, then trypsinized, combined, and replated into 75C2 m Falcon flasks (Falcon Plastics). When the cultures became confluent, cells were collected, diluted to 25,000 cells (0.25 ml–1), and injected iv into new C57 mice. Three weeks later, these mice were killed, and their pulmonary metastases adapted to grow in culture as described above. This procedure was repeated five times. The original line was designated as line no. 26 (our melanoma clone 26), and its daughter lines and their progenies designated as lines 27, 28, 29, and 30. T.c., Grown in tissue culture; injected intravenously.

showing a higher capacity for brain colonization. The cells also showed an increased mean number of chromosomes when compared to the originating parental line. In addition to this selection process, which led to a line with higher propensity for brain colonization, the line also demonstrated a much higher rate of pulmonary growth when iv-injected. In 1988, another model of brain metastasis was presented using the B16 melanoma line [23]. This time, the B16 clones used produced metastatic spread to the brain parenchyma from the vessels found in the leptomeninges. In 1988, Arguello et al., using the B16 melanoma line, were able to modify its injection into mice to produce a model of bone-marrow metastasis [24]. Injection of a relatively small number of cells into the left ventricle produced a metastatic pattern strongly directed to the axial skeleton and bone marrow. In the description of the metastatic pattern of the line, the authors commented: “Metastases were also commonly found in the proximal large bones of the extremities. No metastases were ever seen in the most distal small bones such as carpals and tarsals.” This pattern of dissemination is reminiscent of the pattern described by Paget [2]. Now a model closely resembling the seminal work of Paget was available with the B16 melanoma line. In a publication by Dithmar et al., a novel selection of the B16 model was again demonstrated [25]. In this study, the B16-LS9 line selected by Rusciano et  al. is used to establish a model of uveal melanoma in mice [26]. The transcorneal implantation method used produces extra-ocular metastasis to lung and liver tissue. This pattern is similar to the human uveal melanoma condition. Again, the use of the B16 line helped in the development of a model for human pathology.

90

E. Alvarez

Another area of cancer research that has directly benefited from the development of the B16 tumor line as a model for metastatic disease is cancer immunology. In some of the earliest work with the B16 melanoma line, Fidler and Ziedman studied the effect of host irradiation on the line’s metastatic rate [27]. At that time, the increased metastatic rate of the cells in mice that had received whole-body irradiation was not attributed to immunodepression, but to endothelial-cell damage. This increased colonizing ability was referred to as “enhanced trapping effect.” Interestingly, in their own discussion, the authors cite how previous work had demonstrated that the use of cortisone could also increase the metastatic ability of cells. It is likely that both the use of whole-body radiation and cortisone treatment of the animals also led to a degree of immunosuppression, which enhanced lung colonization by tumor cells. In some of the earliest studies of B16 melanoma cells and immune cells, the relationship between lymphocytes, macrophages, and B16 melanoma cells [28–31]. The direct repercussion of this work can be seen in the development of various methods used to activate the patient’s own immune system against the circulating cancer cells. Specifically, it is seen in protocols that use muramyl dipeptide or muramyl tripeptide in a liposome-based vehicle for iv administration [32–34]. Through activation of tissue macrophages, an immunemediated antitumor effect was elicited in vivo. This has led to a better understanding of the complex host–immune response to tumor cells. The B16 melanoma line was also instrumental in expanding our knowledge of the ability of metastases to spread. With the use of parabiosis systems in mice, it was demonstrated that metastatic spread can originate from a metastatic nodule. In a clever experiment, a mouse was sc-implanted with B16 melanoma, and when this first animal underwent surgical removal of the primary tumor site (leg amputation) and was then surgically joined to a naive host, the naive host developed lung tumors. These lung tumors originated from metastatic nodules present in the first animal. This finding offers proof that once a solid tumor has successfully spread to other tissues, the metastatic cascade perpetuates itself in the host. This problem can serve to emphasize the basic problem with solid tumors: metastases present the biggest therapeutic challenge.

4.4 Conclusions A recurring theme of this chapter is that the work that we now see as a final product is the summation of many people working at various points in time. This is evident from Paget’s hypothesis in the late 1800s to the current development of newer models of metastasis [2]. As a final example, I would like to offer an image giving a visual description of the now commonly accepted tumor-cell endothelialization extravasation steps for the B16-F10 melanoma line (Figs. 4.1 and 4.2). Lapis et al. published this sophisticated microscopy work in 1988 [35]. This is a clear example of how a new technology was able to shed light on ongoing scientific research. But microscopy work, specifically, studying the relationship between the endothelium

4  B16 Murine Melanoma

91

and circulating tumor cells in a rodent model can be traced back to work presented in 1915 by Iwasaki (Figs. 4.3 and 4.4) [5]. While Iwasaki mostly studied human samples, a series of images were drawn of murine tumor cells interacting with the vasculature. Back in 1915, Iwasaki recognized that the importance of the tumor– endothelial interactions was essential for the establishment of the tumor embolus. In Lapi’s paper, newer technologies are applied to answer old questions. Over 70 years had passed since the Iwasaki’s description of extravasating tumor cells. At that time, two different reactions of the endothelium were noted, and contact of the tumor cells with the vessel wall would lead to either a separation of endothelial cells or an engulfment by the endothelial lining. The images here present us with a visual representation of works separated by many years, but of equal impact to our knowledge. In conclusion, the success of the B16 melanoma tumor model is clearly evident today. As an in vivo model for solid tumor formation and metastasis, the B16 has fulfilled two important criteria presented earlier. The model has been instrumental in the dissection of the many steps now associated with tumor establishment and the metastatic cascade. A wide range of disciplines have originated from the use of this tumor line. The eventual relevance of the many sublines that have been selected so far still remains unclear. The development of the solid-tumor model is a dynamic process, and one that attempts to produce a very close representation of the disease process itself. Currently, orthotopic tumor implantation appears to be gaining favor in the scientific community. Knowledge regarding embolization, tumor-cell viability, cell number, organ of arrest, and host strain are all important basic concepts required for the refinement of solid-tumor models. Another important topic in the ultimate development of metastatic models is how the site of implantation of tumor cells in the host animal will influence the metastatic pattern in murine models. Whereas many of the murine tumors that are SC-implanted spread to limited organs (mostly lungs), tumors implanted orthotopically generally behave in a manner more consistent with the natural metastatic progression of the human counterpart for in-depth coverage on this topic. An argument has recently been made regarding the need to use the advanced orthotopic tumor models for the screening and development of chemotherapeutic agents [36, 37]. While the orthotopic models might provide the researcher with a model that more accurately depicts the natural history of the tumor, the model itself is more technically demanding and costly than syngeneic metastatic systems. The knowledge gained from these efforts has directly led to the use of the B16 model to ask more questions regarding the nature of metastasis, with the added benefit of time to provide newer tools to answer such questions. It is also important to keep in mind that the overall selection and the acceptance of the model in the scientific community have been influenced by the already existing scientific framework of the time. The B16 melanoma tumor line has most definitely made a profound impact in the field of oncology research. It has served as a model for basic biological research, and has helped in the identification of distinct pathways now available as potential therapeutic targets.

92

E. Alvarez

Fig. 4.1  An endothelialized thrombus attaching to the vessel wall (arrow heads), 24 h after inoculation of tumor cells. Part of a tumor cell (*) and endothelium

Fig. 4.2  A group of tumor cells completely surrounded by an endothelial covering (E) in the lumen of an arteriole, 2 weeks after inoculation of tumor cells. Scale bar: 10 µm, ×1,060

4  B16 Murine Melanoma

93

Fig. 4.3  Experiment 37/189E: iv inoculation, March 3, 1915; killed 3 days later. Commencing growth of a young sarcoma embolus. Vacuolation of endothelial cells is seen at one point, and early distension of the artery is indicated by flattening of the folds of the elastic laminae. Cells resuming spindle shape (×525)

Fig. 4.4  Experiment 37/189B: iv inoculation, March 3, 1915; killed 5 days later. A later stage than Fig. 4.3. Increasing distension of vessel and entrance of capillaries into embolus on left upper surface (×525)

94

E. Alvarez

4.5 Dedication To Olivia.

References 1. Charlton Knox L. The relationship of massage to metastasis in malignant tumors. Ann Surg. 1922;75(2):129–42. 2. Paget S. Distribution of secondary growths in cancer of the breast. Lancet. 1889;571–3. 3. Hart I. ‘Seed and soil’ revisited: mechanisms of site-specific metastasis. Cancer Metastasis Rev. 1982;1:5–16. 4. Tyzzer E. Factors in the production and growth of tumor metastases. J Med Res. 1913;28:309–33. 5. Iwasaki T. Histological and experimental observations on the destruction of tumor cells in the blood vessels. J Pathol Bacteriol. 1915;20:85–105. 6. Warren S, Gates O. The fate of intravenously injected tumor cells. Am J Cancer. 1936;27:485–92. 7. Cloudman A. Organophilic tendencies of two transplantable tumors of the mouse. Cancer Res. 1947;7:585–91. 8. Coman D, Eisneberg RB, McCutcheon M. Factors affecting the distribution of tumor metastases experiments with V2 carcinoma of the rabbit. Cancer Res. 1949;9:649–54. 9. Ziedman I, McCutcheon M, Coman DL. Factors affecting the number of tumor metastases. Experiments with a transplantable mouse tumor. Cancer Res. 1950;10:357–9. 10. Ziedman I., Buss JM. Transpulmonary passage of tumor cell emboli. Cancer Res. 1952;12:731–3. 11. Ziedman I. The fate of circulating tumor cells I. Passage of cells through capillaries. Cancer Res. 1961;21:38–9. 12. Bubendorf L, Schopfer A, Wagner U, Sauter G, Moch M, Willi N, et al. Metastatic patterns of prostate cancer: an autopsy study of 1589 patients. Hum Pathol. 2000;31(5):578–83. 13. Green E. Handbook of genetically standardized JAX mice. Bar Harbor, ME: The Jackson Laboratory; 1968. p. 57, 58. 14. Fidler I. Metastasis: quantitative analysis of distribution and fate of tumor emboli labeled with 125I-5-iodo-2¢-deoxyuridine. J Natl Cancer Inst. 1970;45(4):773–82. 15. Fidler I. The relationship of embolic homogeneity, number, size and viability to the incidence of experimental metastasis. Eur J Cancer. 1973;9:223–7. 16. Fidler I. Selection of successive tumour lines for metastasis. Natl New Biol. 1973;242:148–149. 17. National Cancer Institute, Division of Cancer Treatment Tumor Repository. Catalogue of transplantable animal and human tumors. Frederick: National Cancer Institute, 1983. 18. Brunson K, Beattie G, Nicolson GL. Selection and altered properties of brain-colonising metastatic melanoma. Nature. 1978;272:543–6. 19. Nicolson G, Bronson KW, Fidler IJ. Specificity of arrest, survival and growth of selected metastatic variant lines. Cancer Res. 1978;38:4105–11. 20. Nicolson G. Experimental tumor metastasis: characteristics and organ specificity. BioScience. 1978;28:441–7. 21. Brunson K, Nicolson GL. Selection of malignant melanoma variant cell lines for ovary colonization. J Supramo Struct. 1979;11:517–28. 22. Raz A, Hurt IR. Murine melanoma: a model for intracranial metastasis. Br J Cancer. 1980;42:331–41.

4  B16 Murine Melanoma

95

23. Alterman A, Stackpole CH. B16 melanoma spontaneous brain metastasis: occurrence and development within leptomeninges blood vessels. Clin Exp Metastasis. 1989;7(1):15–23. 24. Arguello F, Baggs RB, Frantz CN. A murine model of experimental metastasis to bone and bone marrow. Cancer Res. 1988;48:6876–81. 25. 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;10:2–8. 26. Rusciano D, Logenzoni P, Burger MM. Murine models of liver metastasis. Invasion Metastasis. 1994;14:349–61. 27. Fidler I, Ziedman I. Enhancement of experimental metastasis by x-ray: a possible mechanism. J Med. 1972;3:172–7. 28. Fidler I. Inhibition of pulmonary metastasis by intravenous injection of specifically activated macrophages. Cancer Res. 1974;34:1074–8. 29. Fidler I. Immune stimulation – inhibition of experimental cancer metastasis. Cancer Res. 1974;34:491–8. 30. Fidler I, Darnell JH, Budmen MB. Tumoricidal properties of mouse macrophages activated with mediators from rat lymphocytes stimulated with concanavalin A. Cancer Res. 1976;36:3608–15. 31. Fidler IJ, Bucana C. Mechanism of tumor cell resistance to lysis by syngeneic lymphocytes. Cancer Res. 1977;37:3945–56. 32. Fidler IJ. Therapy of disseminated melanoma by liposome-activated macrophages. World J Surg. 1992;16:270–6. 33. Killion J, Fidler IJ. Systemic targeting of liposome-encapsulated immunomodulators to macrophages for treatment of cancer metastasis. Immunomethods. 1994;4:273–9. 34. Killion J, Fidler IJ. Therapy of cancer metastasis by tumoricidal activation of tissue macrophages using liposome-encapsulated immunomodulators. Pharmacol Ther. 1998;78(3):141–54. 35 Lapis K, Paku S, Liotta LA. Endothelialization of embolized tumor cells during metastasis formation. Clin Exp Metastasis. 1988;6(1):73–89. 36. Kerbel R. What is the optimal rodent model for anti-tumor drug testing? Cancer Metastasis Rev. 1999;17:301–4. 37. Killion J, Radinsky R, Fidler IJ. Orthotopic models are necessary to predict therapy of transplantable tumors in mice. Cancer Metastasis Rev. 1999;17:279–84.

Part III

Human Tumor Xenografts

Chapter 5

Human Tumor Xenograft Efficacy Models Ming Liu and Daniel Hicklin

Abstract  Mouse models of cancer have consistently been used to qualify new anticancer drugs for development of human clinical trials. The most used models are xenografts of human tumors grown in immunodeficient mice. Retrospective preclinical–clinical correlation studies indicate that xenograft models are very useful and the models continue to make contributions to critical clinical development choices. However, improvements must be made to increase their values. Here, we review the value and the limitations of xenograft models, and discuss how to enhance their role in developing new anticancer treatments. Keywords  Xenograft model • Efficacy • Cancer drug discover

5.1 Introduction Before the initiation of human trials of novel anticancer agents, it is necessary to perform preclinical trials to evaluate the agent’s pharmacologic and toxic effects. Preclinical tumor models play an essential role in the evaluation of efficacy and optimization of lead compounds in the discovery and development of anticancer drugs. A robust, dependable animal model of human disease is critical to the evaluation of potential anticancer candidates. A typical example of an in-vivo screening in the development of targeted cancer therapeutics is shown in Fig. 5.1. Although other animal species (e.g. rat, hamster, rabbit dog) have been used as animal models for cancer research, the mouse has been the most important contributor due to its accessibility, short generation time, ease of propagation, lower consumption of test compounds, and the advances in mouse genetics.

M. Liu (*) Merck Research Laboratories, Schering-Plough Research Institute, K-15-2-2700, 2015 Galloping Hill Road, Kenilworth, NJ 07033, 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_5, © Springer Science+Business Media, LLC 2011

99

100

M. Liu and D. Hicklin Drug Candidate (post biochemical and cellular potency tests) Mouse PK Screening MTD testing in Mouse

Initial Efficacy Evaluation in Appropriate Xenograft Models

PK/PD profiling ADME profiling

Optimization of Dose & Schedule

Additional Efficacy testing in Models: • GEM • Primary Tumor Isolates • Orthotopic

PK/PD profiling

Combination Studies

PK/PD profiling

Toxicology Studies

Fig. 5.1  In-vivo flow of the development of targeted cancer therapeutics

The discovery of cancer drugs through in vivo screening methods traditionally utilized syngeneic transplantable murine tumors. Since 1955, the US National Cancer Institute (NCI) has provided screening support to cancer researchers worldwide [1]. The earliest in  vivo screens were the fast-growing murine leukemias, L1210 and P388, implanted intraperitoneally. These tumors were derived from leukemias originally induced chemically in the DBA/2 mouse by painting the skin with methylcholanthrene [2, 3]. Using survival as the endpoint, these tumors provided a rapid and reproducible means for identifying potential anticancer drugs [4]. From 1975 till 1985, the in  vivo P388 mouse leukemia model was used almost exclusively as the initial or primary screen at the NCI. It became evident that there were marked similarities in the drugs emerging from the murine leukemia screen. The classes of agents found active in the mouse tumor models were limited, mainly comprising alkylating agents and DNA interacting drugs [5]. Subsequently, panels of syngeneic murine solid tumors and human tumor xenografts have largely replaced the murine leukemias in anticancer drug screens. Syngeneic rodent tumor models provide an experimental model for evaluating the anticancer effects of therapeutic agents in animals with an intact immune system. However, their relevance to human cancer may be overestimated [6] as subsequent analysis revealed that almost 30% of the compounds that were disregarded by the syngeneic models were active in at least one human tumor xenograft, e.g. paclitaxel, and some classes

5  Human Tumor Xenograft Efficacy Models

101

of compounds, e.g. brefeldins and some minor groove DNA binders, seem to have less activities in syngeneic models [7]. Successful xenografting of human tumors into nude mice was first reported in the late 1960s [10, 11]. Nude mouse models are now extensively used in the development of potential anticancer drugs and studies of tumor biology Fig. 5.1. Moreover, mice with severe combined immunodeficiencies (e.g. SCID, beige, xid) have enlarged the spectrum of possible models and enabled engraftments of human tumors that were previously difficult to explant, such as those of the hematopoietic system. Thus, xenograft/human explants have become the workhorse in cancer drug development, and their use is highly recommended by various regulatory agencies [12]. Cancer is a complex family of diseases, characterized by the deregulation or dysregulation of normal control pathways for cellular growth, differentiation, and apoptosis. In the late 1980s and early 1990s, the focus of new drug development shifted to molecularly targeted/disease-directed treatment strategies [13]. In 1985, the NCI initiated a new project assessing the feasibility of employing human tumor cell lines for large-scale drug screening [14]. Cell lines derived from seven cancer types (brain, colon, leukemia, lung, melanoma, ovarian, and renal) were acquired from a wide range of sources and subjected to battery of in vitro and in vivo evaluations. In 1993, the composition of the cell line screen, often referred to as “the NCI panel of 60 cell lines,” NCI-60, was modified to include various prostate and breast tumor lines [15]. As part of the evaluation, in vivo tumor models derived from this panel were used to assess the antitumor efficacy of new compounds [16]. Because more than 85% of the screened compounds did not show activity, a highly sensitive three-cell-line prescreen was adopted in 1999 [7, 8]. In addition to the progress in the human xenograft models, the shift from “compound oriented” to “disease oriented” drug discovery at the NCI in 1989 also prompted a realization that there was a need to identify more target-defined ­models. Specifically designed and bred transgenic/knockout mice have proven very useful to satisfy this need. During the past 20 years, an impressive range of tools has become available to the mouse geneticist, as well as tumor biologists. Groundbreaking experiments in several laboratories established the first transgenic mouse tumor models by expressing viral [17, 18] or cellular [19, 20] oncogenes in specific tissues. Germline inactivation of the prototype tumor suppressors gene Rb [21] and p53 [22] using gene targeting technology in mouse embryonic stem cells provided additional tools to model the scope of mutations in human tumors [23]. Both xenograft and transgenic/knockout models are being increasingly utilized in the discovery of anticancer agents. However, both animal models have been criticized for failing to predict the response of human patients to new agents [24–26]. It is generally believed that requirements for successful preclinical animal-tumor models should include the following characteristics: reproduction of the biology of human cancer; objective and quantitative evaluation of cellular and molecular events associated with cancer progression; reliability, availability, and affordability. In this chapter, we compare the utility and the limitations of the xenograft models, particularly from the viewpoint of the pharmaceutical drug development, hoping to capture the advantages and disadvantages of the models for cancer drug evaluation.

102

M. Liu and D. Hicklin

5.2 Xenograft Tumor Models for Efficacy Evaluation 5.2.1 Immunodeficient Mice The growth of human tumors in a different species (e.g. mouse) requires ­immunodeficiency in the host animal to prevent rejection of the transplanted ­foreign ­tissues. There are many strains of immunodeficient mice containing single mutations (e.g. nude, scid, beige, xid, rag-1 null, rag-2 null) or combined ­mutations, (e.g. bg/nu, bg/nu/xid, nude/scid, nod/scid) available for cancer research [6, 27]. These mice have mutations leading to different degrees of immunodeficiency in natural killer (NK) cells, lymphokine-activated killer (LAK) cells, macrophages, B cells, T cells, and blood immunoglobulin production. Table 5.1 includes an outline of some of the immunologic characteristics of the nude and severe combined immune deficiency (SCID) mice, with single mutations, in comparison to the NOD-SCID and the NIH III nu/beige/xid mice with combined mutations. Nude mice and SCID mice are the strains of immunodeficient mice most commonly used as the recipient for human tumor xenografts, with the nude mouse being more heavily utilized. Both nude mice and SCID mice are easily accessible Table 5.1  Immunodeficiencies in nude, scid, nod-scid and nude/beige/xid mice NIH III nude/ Nude Scid Nod-scid beige/xid Prkdc and nod on nu, beige on Gene defect Prkdc on Nu on chromosone chromosome chromosome chromosome 13, and xid on 18 16 11 X-chromosome Pre-B and B cells Pre-B and B cells Pre-B and B cells Defective absent absent B cells maturation/ absent precursor B cells normal Nonfunctional Nonfunctional Nonfunctional T cells Mature and functional T cells low or absent NK cells High Normal Nonfunctional Nonfunctional Macrophages Normal Normal Defective Defective LAK cells Normal Normal Absent IgG low IgG low IgG low IgG low Serum immuno­ globulin IgM normal IgM low IgM low IgM low IgA low IgA low IgA low IgA low Lymphoid Athymic Small lymphoid Small lymphoid Athymic organs organs organs Hair No Yes Yes No

5  Human Tumor Xenograft Efficacy Models

103

in large quantities from ­commercial sources (e.g. Charles River Laboratories, Jackson Laboratory, and Harlan Bio-products). The SCID mouse is in general more immunodeficient (reduced NK cell, nonfunctional T cell, and more defective B cell) than the nude mouse, therefore, higher take rates and more metastasis are achieved in SCID mice for many human tumor cell lines [28, 29]. Since the development of the human-SCID chimera mouse models [30], SCID mice have been broadly used in studies of anticancer immunotherapy by engrafting human tumor with various combinations of human peripheral blood leukocytes (HPBL), subsets of the HPBL, and/or intact human stromal tissue adjacent to the tumor tissue [31]. The nude mouse, due to its milder temper and hairless feature, is generally easier to handle and easier to observe/quantify the growth of transplanted tumors. The SCID mouse is more sensitive to the toxic effects of irradiation and some cytotoxic agents [32, 33]. The SCID mouse is also more expensive than the nude mouse, which may be a practical reason that it is used less frequently than the nude mouse for drug screening. Other immunodeficient mice, especially mice with combined mutations, are frequently used to explore the immunological mechanisms underlying tumor progression and compound efficacy [31, 34, 35]. Over the past ten years, NOD-SCID mice have been the “gold standard” for studies of human hematolymphoid engraftment in small animal models [36].

5.2.2 Cultured Tumor Cells Vs. Tumor Fragments Subcutaneous injection of cultured tumor cells into immunodeficient mice is widely practiced. In general, cultured tumor cells have a much higher take rate when ­inoculated as suspension into nude mice than human solid tumors of the same ­histological type that are transplanted directly from the patient [139]. A wide variety of human cancer cells can be procured from institutions such as American Type Culture Collection (ATCC, www.atcc.org), German Collection of Microorganisms and Cell Culture (DSMZ, www.dsmz.de), and European Collection of Cell Cultures (ECACC, www.ecacc.org.uk). Usually, the cryo-preserved cell lines are thawed and cultured in medium (e.g. DMEM or RPMI) supplemented with heat-inactivated fetal bovine serum, and expanded until the population is sufficient. Cells are harvested and implanted subcutaneously into various regions (e.g. axillary, flank, or back) of the immunodeficient mouse. Different cell lines require different cell number inoculums for optimal tumor growth. Between one and ten million cells per mouse is a suitable range for the majority of cell lines. To increase take rates, basement membrane matrix (e.g. BD Matrigel) is frequently incorporated in the inoculums. With or without the basement membrane matrix, proper in vivo growth kinetics induced by appropriate amount of inoculated cells needs to be titrated. Implanted animals are commonly monitored two to three times weekly for tumor growth using caliper measurements to determine the length (L), width (W), and height (H) of the tumor. The treatment of the implanted tumors can be started either immediately after inoculation (nonstaged model) or postponed until the tumor reaches a certain size (staged model).

104

M. Liu and D. Hicklin

Tumor fragments have also been subpassaged in animals and used as a tool to evaluate anticancer therapeutics by the NCI [16] and others [12]. The initial solid tumors established in mice are maintained in the mouse by serial passage of 30–40 mg tumor fragments implanted subcutaneously. For each line of tumor, both range and mean values of tumor doubling time are provided to demonstrate the inherent variability of growth. Serial passage is not allowed to exceed a defined range with replacement starting from the frozen stocks around the tenth generation [16]. Xenografts derived directly from patient biopsies appear to have better retention of the morphological and molecular markers of the source tumors [7]. Issues with the use of tumor fragments include reports indicating contamination with malignant mouse cells, changes in hormone sensitivity, changes in histology patterns, and changes in response to anticancer agents following serial passage in mice [37–39]. Despite these issues, research groups using human tumor xenografts established in serial passage believe that such methodology has a higher correlation with clinical drug response. In addition, serial passage of tumor fragments from clinical specimens also allows for preselection of responsive tumor types for follow-up studies [12].

5.2.3 Subcutaneous Vs. Orthotopic Transplantation The subcutaneous xenograft model is easy to monitor and quantify, however, it is ectopic, (i.e. out of the native place). The inhibition of the growth of a tumor implanted in the subcutaneous tissue space after administration of a cytotoxic compound may be a reliable assay for antitumor activity in  vivo, but the same assay may be inappropriate to identify agents against cellular target molecules that are only expressed when the tumor resides orthotopically in original organs [40, 41]. In order to obtain improved models over subcutaneously growing human tumors, there have been efforts [42, 43] to develop techniques of surgical orthotopic implantation (SOI) to transplant histologically intact fragments of human cancers, including tumors taken directly from the patient, to the corresponding organ of immunodeficient rodents. The SOI models include spontaneous bone metastasis models of prostate cancer, breast cancer, and lung cancer; spontaneous liver and lymph node metastatic models of colon cancer; and spontaneous brain metastasis of melanoma and prostate cancers. Comparisons of the SOI models with transgenic mouse models of cancer have indicated that the SOI models have more features of clinical metastatic cancers. Xenograft tumors seem to yield a much higher frequency of metastases when implanted orthotopically [44]. Introduction of renal cancers, as well as tumors of other origins, into the subrenal compartment of immunodeficient mice seems to increase engraftment take rates and metastasis rates. After serial transplantations under renal capsule, xenografted tumors maintain genetic stability and responses to cytotoxic chemotherapy [45]. It is hoped that above models will increasingly be used in preclinical evaluation of potential therapeutics.

5  Human Tumor Xenograft Efficacy Models

105

One factor in selecting the orthotopic location of the transplanted xenograft is the hormonal dependency of the tumor. Many breast cancer (e.g. MCF-7, ZR-75-1, Br-10) and prostate cancer models (e.g. LNCaP and CWR-22) are dependent on estrogen or testosterone and they require supplementation with sex hormones or intact sexual organs to grow in the mouse [46, 47]. Usually hormone supplements are given by implanting time-release pellets of hormone (commercially available, e.g. Innovative Research of America, Sarasota FL) via trocar needle into the subcutaneous tissue of mice. The hormones in these pellets are designed to be constantly released and can last a defined period of time, from days to weeks. It is believed that it is better to have human breast cancers grow in ovariectomized female mice supplemented with extra estrogen due to the influence of fertility cycles on the primary tumor growth rate and metastasis rate [48]. One should note, however, that the extra estrogen can cause toxicity or even death to the mouse, so caution and especially close monitoring of the treated mouse (e.g. by physical signs or radioimmunoassays) needs to be exerted. Combinations of castration/ovariectomy and supplementation of testosterone/estrogen in various sequences have been utilized to mimic hormonal blocking therapies, hormonal supplementation therapies, and different drug and hormone resistant conditions [49–51].

5.2.4 Tumor Metastasis It is widely acknowledged that tumors grown subcutaneously are less likely to metastasize than those grown in the anatomically relevant or orthotopic sites, as mentioned earlier. Various studies have shown that transfection with human angiogenic cytokines and subcutaneous implantation of xenogeneic tumor cells with Matrigel and/or helper host cells such as fibroblasts can increase tumor growth and/ or malignant potential [52–54]. The simple “experimental metastasis” assays are intended to mimic late stage of metastasis (dissemination, extravasation, and colonization). The most commonly used method is the injection of cells into the tail vein of mice. In these cases, the resultant tumor colonies are most commonly confined to the lung – which harbors the first capillary bed encountered – due to mechanical trapping. Tumor cells growing in culture frequently have a good plating efficiency and are generally easy to clone; cloning efficiency is thought to be related to the ability to metastasize [55]. The intravenous inoculation of cultured cell suspensions has been widely used in the hope of obtaining pulmonary metastasis. Curiously, though, this rarely happens, and the majority of cell lines do not give rise to pulmonary metastases even following intravenous injection of several million cells. Only a limited number of cell lines produce metastases, and most of these do so with low frequency. It should be noted that there is no direct correlation between lung colonization and spontaneous metastasis, and in some tumor models there are significant discrepancies between these two functions [54, 56]. It is perhaps less surprising that tumor cells derived from cells that are naturally migratory, such as leukemias, lymphomas, and ­plasmacytomas,

106

M. Liu and D. Hicklin

more readily metastasize and form colonies in multiple sites, including bone ­marrow, spleen, and liver. Other methods have been used to recapitulate the process of metastasis. Cells can be introduced into the portal circulation for liver colonies or the left ventricle of the heart for bone colonies. Inoculation of cultured cells into the left ventricle has proven to be a useful route for determining the specific organ tropism of the inoculated cells [57]. The spleen provides an alternative site for the injection of tumor cells than the portal or mesenteric vein. Injected cells pass almost immediately into the portal circulation. Inoculation of human tumor cell suspension into the mouse spleen can produce liver metastases using tumors (e.g. colorectal carcinoma) selected for liver implantation. A few metastases generated from the initial splenic implantation are isolated and cultured for further rounds of splenic inoculation/selection/cultivation. Cell lines generated this way will metastasize exclusively to liver and cause significant amount of micro-metastases [58]. Tumor cells can be injected directly into the liver parenchyma for circumstances where a small number of colonies are required [59]. Tumor cells have also been injected directly into the pleura, peritoneal cavity, bone marrow, and brain, but there is a risk of morbidity and mortality, and quantization of tumor burden is difficult [54, 60].

5.2.5 Monitoring Tumor Progression and Determining Efficacy The growth of the primary tumor is routinely quantified by in situ caliper ­measurements of the three perpendicular dimensions. Various formulas are used to determine the volume of the tumor, with L (length, longest diameter, in mm) × W (width, in mm) × H (height, in mm) × p/6 or (L)2 × W × p/6 the most commonly used formulas, both of which assume that tumors are ellipsoid, and both of which correlate well with the tumor weight measured upon necropsy [61]. Weight of the tumor can also be converted from measurements (mm) of two perpendicular dimensions, L and W, using the formula, L × (W)2 × 1/2, for a prolate ellipsoid that assumes a specific gravity of 1.0 g/cm3 [16, 62]. Parameters commonly used to quantify the effects of the test agents include the following: • • • • •

Dose levels Relative growth: % T (treated)/C (control) [1] Growth inhibition: 1 – % T/C [63] Growth delay: % (T – C)/C [63] Net log cell kill: [(T – C) – treatment duration] × 0.301/median doubling time [64] • Rate of partial or complete regression (PR or CR) [65] • Percent tumor free animals [65] • Increase in life span [65].

5  Human Tumor Xenograft Efficacy Models

107

Many researchers use computers in connection with electronic calipers, ­ eighing balances, and telemetry devices to facilitate and automate the operations. w There are various computer programs designed for in vivo tumor biology applications available commercially, e.g. LABCAT (www.LABCAT.com) and STUDYLOG (www.studylog.com). It is simple to evaluate efficacy by measuring subcutaneous tumors with calipers, thus accounting for their popularity. However, there are factors that, e.g. tumors of irregular shape, inconsistent measurement of tumor axes, and differences in tumor measurement between investigators, influence the validity of caliper measurements. Therefore, automated methodologies such as laser scanning are being developed [66] to reduce inter-operator variability and improve reproducibility. The therapeutic response in xenograft models is typically based on reduced tumor growth as opposed to tumor regression required in humans. Thus, the predictive value could be improved in mouse models by adopting tumor regressions as the efficacy end point. On the other hand, chronic slowness of tumor growth rate can be considered a relevant outcome as compared with tumor regression if our goal is to delay tumor progression [7]. To overcome the possibilities of leaving useful data due to setting arbitrary cutoff of T/C ratio, tumor biologists have been proposing the use of tools such as adjusted area under the curve ratio to quantify tumor inhibition [67]. Histology has traditionally been used to study tumor progression and both qualitative and quantitative evaluation of the changes of various tumor markers. Immunological techniques such as ELISA, immunoblotting, immunohistochemical staining, and nucleic acid based techniques such as Southern and Northern blottings, and quantitative PCR all have been utilized widely to study the changes of biological markers in the tumor samples. These techniques have been mostly used as end-point assays. It is uncommon to have biopsies done on tumors of preclinical studies. There have been efforts to overcome such limitations by transfecting tumor cell lines that are used to generate grafted tumors with a tumor marker gene (e.g. prostate specific antigen), and then monitoring the growth of the tumor in vivo by quantifying the tumor marker in the serum of the mouse host [31, 68]. However, despite such efforts, more sensitive technology such as molecular profiling of ­circulating tumor cells [9] for the study of tumor progression and metastasis would be desired. Recent work has established the feasibility of analyzing tumor growth and growth inhibition in live animals. Table 5.2 includes examples of in vivo techniques that have been utilized to monitor tumor progression, as well as efficacy by therapeutics, in xenograft models. In vivo imaging of cells tagged with light-emitting probes, such as green fluorescent proteins (GFP) [69] or firefly luciferase [70, 71] has been shown to be a powerful technology that enables imaging of single tumor cells and metastases and a wide range of other biological characterizations in tumor-bearing mice. GFP- or luciferase expressing tumors (e.g. colon, prostate, brain, liver, lung, bone, and others) were visualized externally through quantitative trans-cutaneous whole-body imaging. Since these technologies can be conducted in a real-time fashion on live animals, they provide particularly useful tools for the

108

M. Liu and D. Hicklin

Table 5.2  In-vivo techniques used to monitor tumor progression/efficacy in xenograft models Subcutaneous Orthotopic tumor growth tumor growth Metastasis Angiogenesis Caliper measurement + − − − + + + + Optical imaging: bioluminescence and fluorescence MRI, Micro CT/PET/ + + + − SPECT Ultrasound + ± ± − Subcutaneous window + − − + chamber

study of the process of tumor progression. For example, direct observation of the metastatic process can be performed using in  vivo video microscopy, and early steps in the metastatic process related to cell survival or extravasation have been successfully imaged with this technique [72, 73]. Various transparent window preparations offer useful methods to allow noninvasive, continuous measurement of tumor growth, angiogenesis, blood flow, and expression of tagged genes in living tissue [74]. In addition, tumor tissue oxygen content can be measured using a polarographic needle microelectrode (pO2-Histograph, Eppendorf, Inc.) to document responses to various anticancer agents (e.g. cyclophosphamide, cisplatin) in combination with agents that aim at reversing hypoxia in tumor tissue [75]. Initial efforts are also being made to study tumor responses, such as angiogenesis, extracellular volume, and microvascular permeability, to anticancer agents with magnetic resonance imaging (MRI) [76, 77], positron emission tomography (PET) [78], ultrasound (US) [79], computed tomography (CT) [80], and single photon emission computed tomography (SPECT) [81] techniques. The area of in vivo tumor imaging technology, which offers exciting potential for studies of cancer physiology and treatment regimens, continues to develop quite rapidly.

5.2.6 Examples of Single Agent and Combination Preclinical Trials Human tumor xenograft assays were successfully exploited at the NCI to facilitate the discovery of more than a dozen clinically useful cytotoxic anticancer drugs [16]. Current anticancer drugs that were evaluated in these systems include melphalan, cyclophosphamide, dacarbazine, BCNU, mitomycin C, cisplatin, actinomycin, doxorubicin, bleomycin, methotrexate, 5-fluorouracil, vinblastine, and paclitaxel. These drugs were evaluated in a panel of tumor lines that belong to different organ/disease types (colon, CNS, leukemia, nonsmall cell lung, small cell lung, melanoma, and ovary). Drugs were administered intraperitoneally, over different schedules such as:

5  Human Tumor Xenograft Efficacy Models

• • • •

109

qd × 4 (once daily for consecutive 4 days) qd × 5 (once daily for consecutive 5 days) q4d × 3 (once every 4 days for three times) q7d × 3 (once every 7 days for three times)

At each dose or schedule level, each drug was ranked by five efficacy levels: • 0 = inactive, % T/C (change in tumor weight, each treated [T] or control [C] group of mice) > 40 • 1 = tumor inhibition, % T/C range 1–40 • 2 = tumor stasis, % T/C range 0 to −49 • 3 = tumor regression, % T/C range −50 to −100 • 4 = % T/C range −50 to −100 with > 30% tumor-free mice To screen and prioritize compounds for testing in the xenograft models, so-called hollow fiber assays were used, in which tumor cells are cultured in sealed hollow fibers and implanted either subcutaneously or intraperitoneally in the nude mouse. After drug treatment for 6–8 days, cell survival is quantified by MTT dye conversion measurements [82]. The in vivo drug sensitivity profiles of these human tumor xenografts have served as worldwide benchmarks for the testing of new agents. More recently, xenograft tumor models have been used to evaluate molecularly targeted therapies. Efforts devoted to such “target-oriented” drug discovery have produced fruitful results [83–86, 87]. Table 5.3 lists examples of therapeutics and xenograft models representing such achievements. One successful example of this effort has focused on the epidermal growth factor receptor tyrosine kinase (EGFR-TK). A variety of human tumor xenografts, including prostate (TSU-PRI, PC-3. DU-145, CWR-22), ovarian (OVCAR-3), breast (MCF-7, ZR-75-1), colon (GEO), vulva (A431), and lung carcinoma (A549, SK-LC-16, LX-1), grown in nude mice, were used to evaluate the EGFR-TK inhibitor (EGFR-TKI) termed gefitinib, also known as Iressa [130–132]. The EGFR signaling pathway contributes to a number of processes important to tumor progression, including cell proliferation and apoptosis. EGFR is highly expressed in many tumors and is associated with poor disease prognosis. Oral administration of gefitinib produced dose-dependent reversible growth inhibition in a wide range of tumor xenograft models. Gefitinib has also shown growth inhibitory activity against xenografts initiated from ductal carcinoma in situ tissues, indicating that EGFR inhibition may have a role in the treatment of early stage breast cancer [133]. In addition to small molecular chemicals, biologics such as monoclonal antibodies have also been shown widely to be efficacious in xenograft models. Anti-EGFR antibody cetuximab was validated to be efficacious in NSCLC models expressing both wide type and mutated EGFR [134]. Human tumor xenograft models have also been used to evaluate the ­combi­natorial efficacy of molecular therapeutics in combination with cytotoxic drugs [135, 136] or gene therapeutic agents, such as a p53 recombinant adenovirus [140]. Enhanced in  vivo efficacy was observed when farnesyltransferase

110

M. Liu and D. Hicklin

Table 5.3  Examples of cancer therapeutics and xenograft responses Target Therapeutics mechanism Clinical indications Gefitinib EGFR Nonsmall cell (Iressa®) lung cancer Erlotinib EGFR Nonsmall cell lung (Tarceva®) cancer Pancreatic cancer Cetuximab EGFR Colorectal cancer (Erbitux®) Squamous cell carcinoma Sorafenib Renal cell Raf/Mek/Erk (Nexavar®) carcinoma pathway, PDGF, Hepatocellular VEGFRs carcinoma Sunitinib Advanced renal PDGFRs, (Sutent®) cell ca. VEGFRs, KIT, RET, Gastrointestinal flt3 stromal tumor Trastuzumab HER2/neu Metastatic breast (Herceptin®) cancer Lapatinib (Tykerb®)

EGFR, HER2/neu

Temozolomide (Temodar®)

Nucleotide methylation

Imatinib (Gleevec®)

Bcr-abl, c-Kit

Bevacizumab (Avastin®)

VEGF

Advanced or metastatic breast cancer Newly diagnosed glioblastoma multiforme Refractory anaplastic astrocytoma Philadelphia positive chronic myeloid leukemia, acute lymphoblastic leukemia Kit+ gastrointestinal stromal tumor Metastatic colorectal cancer Nonsquamous nonsmall cell lung cancer Metastatic breast cancer Glioblastoma Metastatic renal cell carcinoma

models that recapitulate clinical

Xenograft models A549, NCI-H1155, NCI-H23, NCI-H441 [89, 90] A549, NCI-H1299, NCI-H1975 [91, 92] BxPC-3 [93] GEO, HT-29, HCT-8, SW620, HCT-116 [94, 95] A431, UMSCC-1, SCC1, CAL27, FaDu [96–98] 786-O [99] PLC/PRF/5, LCI-D20 [100, 101] 786-O, KCI-18 [102] GIST-T1 [103] BT-474, MDA-MB-231-BR, JIMT-1, Her2(±)-MCF7 [104–107] MDA-MB-231-BR, MDA-MB-231-4173, MTLn3 [105, 108] U-87 MG, D-54 MG, D-245 MG, GBM6, GBM14 [109–111] U373 MG

KBM5, KBM7, K562 [112, 113]

GIST882 [114] GEO, LS174T, WiDr, HCT116, KM12SM [115–117] PC14PE6 [118]

KPL-4, MDA-MB-231 [119, 120] Hs683 [121] Caki-1 [122] (continued)

5  Human Tumor Xenograft Efficacy Models

111

Table 5.3  (continued) Therapeutics Docetaxel (Taxotere®)

Target mechanism Microtubules

Clinical indications

Xenograft models

Breast cancer

R-27, MX-1, SK-BR-3 [123, 124] A549, NCI-H460 [125]

Nonsmall cell lung cancer Prostate cancer Gastric adenocarcinoma Head and neck cancer

22Rv1, LNCaP, DU-145 [126] MKN-28, MKN-45, KKLS [127] UMSCC2, HNC-3, KB3-1 [128, 129]

inhibitor ­lonafarnib was combined with paclitaxel, cyclophosphamide, 5-FU, or vincristine in the human nonsmall cell lung cancer NCI-H460 xenograft model. Significantly greater combined efficacy for lonafarnib and the p53 adenovirus was also observed, compared to either agent alone, in both the intraperitoneal and subcutaneous DU-145 human prostate xenograft models. The EGFR-TKI gefitinib in combination with a range of cytotoxic agents has also shown promising activities in several human tumor xenografts. Gefitinib combination therapy was associated with a significant inhibition of tumor growth and a significant increase in survival of nude mice in the GEO colon model, especially when gefitinib was combined with paclitaxel [137]. Gefitinib in combination with cytotoxic agents (carboplatin, paclitaxel, or edatrexate) enhanced antitumor activity and in some cases produced tumor regression in nude mice bearing prostate tumor xenografts TSU-PrI and PC-3, [138]. It is also established that antibodies targeting the epidermal growth factor receptor, Cetuximab, and vascular endothelial growth factor receptor-2, DC101, are synergistic with regards to antitumor effects, in a BxPc3 subcuanteous xenograft model for pancreatic cancer [141]. Combination treatments with cytotoxics also increased the efficacy of cetuximab [134]. In addition, an enhancement of efficacy was shown in a head and neck cancer orthotopic model CAL33 with combined effects of anti-VEGF antibody bevacizumab, anti-EGFR inhibitor erlotinib, and irradiation [142]. These examples illustrate that, in addition to offering an efficacy screening tool, xenograft tumor models can also be used to analyze the action of molecular targeted therapeutics. Furthermore, xenograft models can be used not only to evaluate anticancer agents but also to facilitate target validation and proof of principle by exploring how manipulating the presumptive therapeutic target(s) in tumor cells alters their biological response to such agents. In conjunction with a variety of technologies (e.g. siRNA, functional blocking antibodies, dominant negative approaches), xenograft efficacy studies can help to identify responsive tumor subtypes, to stratify patient selection, and to guide clinical path [83]. The examination of human breast tumor lines in this way, contributed to the identification of trastuzumab, which is used clinically to treat breast cancers that overexpress the ErbB2 target protein [143].

112

M. Liu and D. Hicklin

5.3 Pros and Cons No model is ideal. The most useful models would be those that reflect the natural history and histopathology of human disease, provide more accurate understanding of cellular and molecular mechanisms, and allow for discovery and development of therapeutics that are clinically effective. In the cancer arena, while not all chemotherapeutic agents that test positively in mouse models are efficacious in humans, agents that are efficacious in humans are generally effective in mice [25]. Both xenograft and transgenic models have strengths and limitations in this regard. A comparison of the general advantages and disadvantages of each type of model is listed in Table 5.4. However, we have to always bear in mind that both are models in the mouse. Mice tend to be more resilient to many xenobiotics and supranormal concentrations of protein molecules than are patients [139]. It is critical for a new drug to define the selectivity profile against human tumors of different organs. In this respect, human tumor xenografts are considered by many as the most relevant models [12]. In addition, xenograft models have been characterized and calibrated by many tumor biologists, and they have been widely used in large-scale screening for many years. Due to their extensive use, both the mouse hosts and the tumor cell lines are generally in the public domain and can be easily

Table 5.4  A comparison of xenograft (subcutaneous and orthotopic) and genetically engineered models Subcutaneous xenograft Orthotopic xenograft Transgenic/knockout Easy to set up Need surgical expertise Need TG/KO expertise Relatively inexpensive More expensive More expensive Labor economic Labor intensive Labor intensive Time economic Longer time needed Longest, months to years needed Widely used and better Database being Database being accumulated understood accumulated Easy to monitor tumor Not as easy to monitor tumor Not as easy to monitor tumor burden and progression burden and progression burden and progression Cannot study immune Cannot study immune Immune responses can be responses responses studied Gene expression is not organ Organ-specific gene Organ-specific gene expression specific expression available Relevant tumor cell and host Relevant tumor cell and host Least relevant tumor cell and host organ interactions organ interactions organ interactions Lack of natural metastasis Metastasis can be studied Metastasis can be studied Not as good for target Not as good for target Good for target validation and validation validation proof of principle Not as good for study of early Not as good for study Early tumorigenesis events can events of early events be studied Good for prophylactic therapy Not as good for Not as good for and chemoprevention prophylactic therapy prophylactic therapy and chemoprevention and chemoprevention

5  Human Tumor Xenograft Efficacy Models

113

obtained. Especially for the subcutaneous models, the methods are relatively easy to set up, less labor-intensive, less costly, and highly reproducible. Therefore, it is relatively straightforward to screen through a large panel of xenograft models representing various cancers of different organ types to evaluate tissue sensitivity of a particular series of anticancer drugs. Another advantage of xenografts is reflected in the naturally occurring mutational changes that take place during the process of carcinogenesis. Each cell line likely carries multiple mutations, and each set of mutations differ among cell lines, mirroring the clinical situation. The efficacy results may be more predictive if the anticancer candidates are tested against a panel of tumors within a given histology. Xenograft systems are not suitable for testing agents that work through immunebased or species-specific mechanisms that involve host cell interactions. Although the implanted cancer cells are of human origin, their growth depends on the mouse host (e.g. blood and nutrient supply, infiltration of extra-cellular matrix, interactions with host hormones and growth factors). Lack of spontaneous metastasis is another major drawback of xenograft models, although orthotopic implants can address this issue to some extent. However, while orthotopic procedures result in an environment that is more natural for the tumor, increasing the likelihood of metastatic progression, orthotopic models are far more labor intensive and require more surgical expertise. This feature makes it impractical to perform orthotopic assays for large-scale screening. Incorporation of green fluorescence protein or luciferase into xenografts make these models more powerful, by offering simple methods to monitor even very small and systemically distributed tumors by bioluminescence imaging techniques. It should be noted that, after the establishment of the new tumor cell line with the indicator gene, a thorough characterization needs to be conducted to make sure the new cell line still has the same phenotype as the original line. Although orthotopic xenograft models may be more representative than subcutaneous models, both share an intrinsic problem. The tumors usually are derived from cell lines that have been selected for in vitro and in vivo growth. They grow rapidly, so they can be used in higher throughput testing in a timely fashion; however, this rapid growth certainly does not reflect the progression of clinical cancers which are typically relatively slow growing. The interactions between xenograft tumor cells and vasculature or stromal microenvironment also are not well defined, and may not reflect the interactions that occur during cancer development in human patients.

5.4 Pharmacology and Pharmacokinetic Correlations Cancer drug discovery has entered a new era. Recent developments in molecular biology and chemistry (e.g. genomics and bioinformatics, improvements in cloning/expression technologies, structural biology, high-throughput screening, and combinatorial chemistry) have led to a significant increase of new drugable targets.

114

M. Liu and D. Hicklin

In the year 2002, there were nearly 500 molecules undergoing clinical studies in cancer and this number could well reach 1,000 in 2003 [144]. Due to the rapid increase of candidates, new methods and criteria to prioritize the most promising candidates are needed before and during clinical trials. In addition to the traditional phase I clinical trial goals (maximal tolerated dose determination, phase II dose recommendation, safety and pharmacokinetic evaluation), current phase I studies also need to examine biomarkers and to stratify patients and clinical endpoints with these biomarkers [145]. Therefore, intense efforts are needed to correlate pharmacokinetic profiles and pharmacodynamic changes in the drug target with preclinical efficacy, such that adequate information can be gained with regard to the problems of dose determination and patient stratification. Current drug discovery programs commonly incorporate such integrated approaches at early phases of the program. Anticancer drugs traditionally have been given to patients by intravenous ­infusion; while during preclinical testing most experimental agents have been given by tail vein or intraperitoneal injections. However, there is a growing trend towards developing anticancer drugs that can be administered via additional routes of administration, in particular by oral routes. Therefore, pharmacokinetic studies of compounds delivered by various dosing routes and using various formulations (e.g. saline, 0.4% methylcellulose, 5–20% hydroxy-propyl-betacyclodextrin) should be tested before in vivo efficacy evaluations are started. Different pharmacokinetic parameters for experimental compounds, including half-life, AUC, Cmax, Cmin, and bioavailability can all be readily established in mouse models. For settings where constant delivery of a compound is desired, such as through the use of the Alzet Osmotic Pump, a pharmacokinetic profile utilizing appropriately sized pumps and compound formulations is also readily established. To determine the most suitable dose for reaching maximum in vivo efficacy, drug exposure should be compared with IC50 and IC90 concentrations required to achieve efficacy in cell-based assays. Mouse pharmacokinetic studies also provide important preliminary information which complements any subsequent studies in higher mammal species [86, 88, 146]. To correlate pharmacokinetic parameters with efficacy, and to help determine clinical dose levels/dose responses more accurately, changes in suitable molecular markers (pharmacodynamic markers) need to be monitored carefully. Taking an example, the EGFR-TKI gefitinib, a series of preclinical studies identified surrogate markers of EGFR activity, including EGFR phosphorylation and phosphorylation of the downstream molecules MAPK, AKT, and p27KIP1 [147]. In tumor xenograft models, including head and neck carcinoma, gastric adenocarcinoma, and breast adenocarcinoma, a relationship between gefitinib efficacy, EGFR level, and downstream markers (e.g. phosphorylated MAPK) was established. Preliminary analysis of serial skin biopsies from phase I patients has confirmed that gefitinib results in substantial changes in EGFR-dependent molecules such as phosphorylated MAPK and p27. However, the level of expression of EGFR in cells or tumors was not found to predict sensitivity to gefitinib [148]. Thus, additional biomarkers that can specifically indicate gefitinib sensitivity need to be further defined. As a matter of fact, one of the most important shortcomings in the quest of targeted

5  Human Tumor Xenograft Efficacy Models

115

therapy is the failure to identify and validate predictive markers. In order to identify predictive markers, it is necessary to understand the mechanisms of targeted agents as it is likely that the predictive markers may be specific for a disease type, biological process, and mechanism action of a specific targeted therapy [149]. A combination of multiple detection modalities to enhance the monitoring of biological marker changes have been exploring through the field [150]. The advancing developments in genomic, proteomic, molecular biology, chemistry, and imaging tools will help streamline marker identification and quantification, promoting a more rapid rate of drug testing. Collectively, murine tumor models are critical in drug development, but require a rational and hierarchical approach beginning with toxicology and pharmacology studies to identify therapeutic targets and to compare drugs using rigorous and clinically relevant outcome parameters, e.g. improved survival, better quality of life, and prevention of recurrence [151].

5.5 Future Perspectives It is estimated that the development of a new anticancer agent cost US$ 800 million to 1 billion and takes more than a decade between conception and approval [152]. However, 90% of novel anticancer drugs fail in the clinic despite evidence of antitumor efficacy in preclinical models [153]. In vivo tumor models have been frequently criticized for lack of predictive value. However, the failure of animal tumor systems to serve as predictive models for human cancer does not diminish their potential utility, given the absolute necessity of precisely defining the question that the model will be used to answer [154]. As every investigator is aware, the art of choosing an appropriate model follows upon the art of framing the most appropriate question. One should carefully select xenograft models that are relevant to the study objective, selecting models that represent the pathological or mechanistic setting which fits the goals of the study best. Well-established and/or simpler models still have their place. For example, while a bolus intravenous injection of enzymatically separated tumor cells may not be appropriate to study the process of metastasis, the tumor colonies of relatively uniform number, size, and organ location which form after injection may be invaluable to compare the access and activity of drugs or biological agents [54]. The NCI-60 is a set of human cancer cell lines derived from diverse tissues: brain, blood and bone marrow, breast, colon, kidney, lung, ovary, prostate, and skin. These cell lines have been subjected to a battery of xenograft experiments in numerous studies. Institute such as Wellcome Trust Sangers Institute has been sequencing NCI-60 for mutations in known human cancer genes and making the database available to public in their Web site (www.sanger.ac.uk). Such efforts will guide the use of xenograft models intelligently and efficiently. Table 5.5 lists what successful elements tumor models should possess. In the future, it is likely that a combination of both transgenic and xenograft models will allow investigators to reach a more complete understanding of tumor inhibition, thereby facilitating the drug discovery process. The transgenic model can be extremely

116

M. Liu and D. Hicklin Table 5.5  Key requirements of successful preclinical tumor models Recaptulate the biology of human tumorigenesis Can be evaluated objectively and quantitatively Predictive of clinical response Readily available, reliable, and affordable

valuable to ask mechanism-related questions, however, its accessibility as a tool for large-scale drug discovery remains an issue. In contrast, the xenograft model is less defined in nature, but remains more accessible than the transgenic models. With increasing molecular characterization of each tumor line and the use of orthotopic methodology, the xenograft model will continue to make great contributions to target/ disease oriented cancer research, as well as remaining useful in large-scale drug screening. The marriage of mouse tumor models with rapidly evolving methods to profile genetic and epigenetic alterations in tumors, and to finely map genetic modifier loci, will continue to provide insight into the key pathways leading to tumorigenesis. The utility of both types of models continues to offer great promise for identifying relevant drug targets to treat human cancers [22]. Current anticancer drug discovery is focused on target-oriented and tumor cellspecific approaches; however, there is still an ongoing need for a better understanding of tumor biology. It is widely believed that the successful application of molecular cancer therapeutics will require accurate genetic profiling of tumors as well as the identification of novel, tractable, more promising, and more tumor-typespecific therapeutic targets. Enhanced efforts to identify and utilize pharmacodynamic markers during tumor progression and treatment would be expected to enable faster, more efficient, and more accurate studies of drug efficacy. Further advancement and availability of in  vivo imaging technologies will also be very helpful to the design, analysis, and interpretation of results generated from both transgenic and xenograft studies, especially for the detection of early events in tumor progression and metastasis. One benefit of these technologies will also be a reduction in the number of animals needed for experimental paradigms. It has been proposed that it would be fruitful for the NCI to partner with the FDA and the pharmaceutical industry to play a centralized role in identifying and establishing a standardized set of transgenic and knockout models of proven reliability and predictability to be used in preclinical cancer efficacy tests [25]. Collaborations such as The Mouse Models of Human Cancers Consortium (MMHCC) headed by NCI and the collaboration of the NIEDHS with the FDA, ILSI (International Life Science Institute), and several major pharmaceutical companies will be crucial to validate and promote such goals. Progress in these areas is already being made, and new resources including useful Web site references (http://emice.nci.nih.gov), databases for cancer models (http://cancermodels.nci.nih.gov), and cancer images (http://cancerimages. nci.nih.gov) are available to be freely shared by the cancer research community. A broadening of such centralized and concerted efforts to include an extensive panel of well-characterized human tumor xenograft models representative of different ­tissue types would be a welcome addition to the collaborations being pursued.

5  Human Tumor Xenograft Efficacy Models

117

References 1. Grever MR, Schepartz SA, Chabner BA. The National Cancer Institute: Cancer drug ­discovery and development program. Semin Oncol. 1992;19:622–38. 2. Law LW, Dunn TB, 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. 3. Dawe CJ, Potter M. Morphologic and biological progression of a lymphoid neoplasm of the mouse in vivo and in vitro (abstr). Am J Pathol. 1957;33:603. 4. Teicher BA. In vivo tumor response end points. In: Teicher BA, editor. Tumor models in cancer research. Totowa, NJ: Humana Press. 2002. p. 593–616. 5. Fiebig HH, Dengler WA, Roth T. Human tumor xenografts: predictivity, characterization, and discovery of new-anticancer agents. In: Fiebig HH, Burger AM, editors. Relevance of tumor models for anti-cancer drug development. Contrib Oncol, Vol 54, Basel: Karger; 1999. p. 29–50. 6. Amadori A, Belardelli F, Cavallo F, Ferraris PC, De Rossi A, Doria G, Forni G, Forni M, Giavazzi R, Jemms C, Lollini P-L, Lusso P, Mezzanzanica D, Nanni P, Parmiani G, Roncells S, Scala G, Sensi ML. Mind the mouse! A consensus vie on the use of immunodeficient mice in immunology and oncology. J Immunol Res. 1992;4:1–5. 7. Rubio-Viqueira B, Hidalgo M. Direct in vivo xenograft tumor model for predicting chemotherapeutic drug response in cancer patients. Clin Pharmacol Ther. 2009;85:217–21. 8. 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;65:200–211. 9. Rosell R, de Las Penas R, Balana C, Santarpia M, Salazar F, de Aguirre I, Reguart N, Villa S, Wei J, Ramirez JL, Molina MA, Ramon y Cajal S, Jablons D, Taron M. Translational research in glioblastoma multiforme: molecular criteria for patient selection. Future Oncol. 2008;4:219–28. 10. Rygaard J, Povlsen CO. Heterotransplantation of a human malignant tumor to the mouse mutant “nude.” Acta Pathol Microbiol Scand. 1969;77:758–60. 11. Giovanella BC, Yim SO, Stgehlin JS, Williams LJ, Jr. Development of invasive tumors in the “nude” mouse after injection of cultured human melanoma cells. J Natl Cancer Inst. 1972;48:1531–33. 12. Fiebig HH, Burger AM. Human tumor xenografts and explants. In: Teicher BA, editor. Tumor models in cancer research. Totowa, NJ: Humana Press. 2002, p. 113–37. 13. Sausville EA, Feigal E. Evolving approaches to cancer drug discovery and development at the National Cancer Institute. Ann Oncol. 1999;10:1287–92. 14. Boyd MR. Status of the NCI preclinical antitumor drug discovery screen. In: DeVita VT Jr., Hellman S, Rosenberg SA, editors. Cancer: principles and practice of oncology, updates, vol. 3. Philadelphia: Lippincott. 1989, p. 1–12. 15. Shoemaker RH. The NCI60 human tumor cell line anticancer drug screen. Nat Rev Cancer 2006;6:813–23. 16. Plowman J, Dykes DJ, Melinda H, Simpson-Herren L, Alley MC. Human tumor xenograft models in NCI Drug Development. In: Teicher BA, editor. Anticancer drug development guide. Totowa, NJ: Humana Press. 1997, p. 101–26. 17. Brinster RL, Chen HY, Messing A, van Dyke T, Levine AJ, Palmiter RD. Transgenic mice harboring SV40 T-antigen genes develop characteristic brain tumors. Cell. 1984;37: 367–79. 18. Hanahan D. Heritable formation of pancreatic beta-cell tumours in transgenic mice ­expressing recombinant insulin/simian virus 40 oncogenes. Nature. 1985;315:115–22. 19. Adams JM, Harris AW, Pinkert CA, Corcoran LM, Alexander WS, Cory S, Palmiter RD, Brinster RL. The c-myc oncogene driven by immunoglobulin enhancers induces lymphoid malignancy in transgenic mice. Nature. 1985;318:533–38. 20. Stewart TA, Pattengale PK, Leder P. Spontaneous mammary adenocarcinomas in transgenic mice that carry and express MTV/myc fusion genes. Cell. 1984;38:627–37.

118

M. Liu and D. Hicklin

21. Jacks T, Fazeli A, Schmitt EM, Bronson RT, Goodell MA, Weinberg RA. Effects of an Rb mutation in the mouse. Nature. 1992;359:295–300. 22. Donehower LA, Harvey M, Slagele BL, McArthur MJ, Montgomery CA, Jr, Butel, JS, Bradley A. Mice deficient for p53 are developmentally normal but susceptible to ­spontaneous tumours. Nature. 1992;356:215–21. 23. Jackson-Grusby L. Modeling cancer in mice. Oncogene. 2002;21:5504–14. 24. Gura T. Systems for identifying new drugs are often faulty. Science. 1997;278:1041–42. 25. Rosenberg MP, Bortner, D. Why transgenic and knockout animal models should be used (for drug efficacy studies in cancer). Cancer Metastasis Rev. 1999;17:295–99. 26. Kerbel RS. What is the optimal rodent model for anti-tumor drug testing? Cancer Metastasis Rev. 1999;17:301–4. 27. Clarke R, Dickson RB. Animal models of tumor onset, growth, and metastasis. In: Bertino JR, et al. editors. Encyclopedia of cancer. New York: Academic Press. 1997, p. 10–21. 28. Clark EA, Shultz LD, Pollack SB. Mutations in mice that influence natural killer (NK) cell activity. Immunogenetics. 1981;12:601–13. 29. Williams SS, Alosco TR, Croy BA, Bankert RB. The study of human neoplastic disease in severe combined immunodeficient mice. Lab Anim Sci. 1993;43:139–46. 30. Mosier DE, Gulizia RJ, Baird SM, Wilson DB. Transfer of a functional human immune system to mice with severe combined immunodeficiency. Nature. 1988;335:256–59. 31. Bankert RB, Hess SD, Egilmez NK. SCID mouse models to study human cancer pathogenesis and approaches to therapy: potential, limitations, and future directions. Front Biosci. 2002;7:c44–62. 32. Schuler W, Bosma MJ. Nature of the scid defect: a defective VDJ recombinase system. Curr Top Microbiol Immunol. 1989;152:55–62. 33. Croy BA, Linder KE, Yager JA. Primer for non-immunologists on immune-deficient mice and their applications in research. Comp Med. 2001;51:300–13. 34. Mc Cormack E, Bruserud O, Gjertsen BT. Animal models of acute myelogenous leukemia – development, application, and future perspectives. Leukemia. 2005;19:687–706. 35. Yang S, Dong Q, Yao M, Shi M, Ye J, Zhao L, Su J, Gu W, Xie W, Wang K, Du Y, Li Y, Huang Y. Establishment of an experimental human lung adenocarcinoma cell line SPC-A1BM with high bone metastases potently by (99m)Tc-MDP bone scintigraphy. Nucl Med Biol. 2009;36:313–21. 36. Pearson T, Greiner DL, Shultz LD. Humanized SCID mouse models for biomedical research. Curr Top Microbiol Immunol. 2008;324:25–51. 37. Gao J, Tombal B, Isaacs JT. Rapid in situ hybridization technique for detecting malignant mouse cell contamination in human xenograft tissue from nude mice and in vitro cultures from such xenografts. Prostate. 1999;39:67–70. 38. Horvath G, Ferno M, Baldetorp B, Cameron R, Ranstam J. Progression of human endometrial adenocarcinoma heterotransplanted into nude mice from hormone-sensitive to hormone resistant growth. In Vivo. 1991;5:185–90. 39. Rydell R, Lybak S, Wennerberg J, Willen R. Increased response to cisplatin after long-term serial passage of a squamous cell carcinoma xenograft. In Vivo. 1991;5(1):23–27. 40. Killion JJ, Radinsky R, Fidler IJ. Orthotopic models are necessary to predict therapy of transplantable tumors in mice. Cancer Metastasis Rev. 1999;17:279–84. 41. Fidler IJ. Seed and soil revisited: contribution of the organ microenvironment to cancer metastasis. Surg Oncol Clin N Am. 2001;10:257–69. 42. Hoffman RM. Orthotopic metastatic mouse models for anticancer drug discovery and ­evaluation: a bridge to the clinic. Invest New Drugs. 1999;17:343–59. 43. 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;68:4500–5. 44. Man S, Munoz R, Kerbel RS. Development of models in mice of advanced visceral ­metastatic disease for anti-cancer drug testing. Cancer Metastasis Rev. 2007;26:737–47. 45. Press JZ, Kenyon JA, Xue H, Miller MA, De Luca A, Miller DM, Huntsman DG, Gilks CB, McAlpine JN, Wang YZ. Xenografts of primary human gynecological tumors grown under

5  Human Tumor Xenograft Efficacy Models

119

the renal capsule of NOD/SCID mice slow genetic stability during serial transplantation and respond to cytotoxic chemotherapy. Gynecol Oncol. 2008;110:256–64. 46. Brodie A, Jelovac D, Long BJ. Predictions from a preclinical model: studies of aromatase inhibitors and antiestrogens. Clin Cancer Res. 2003;9:455S–59S. 47. van Weerden WM, Romijn JC. Use of nude mouse xenograft models in prostate cancer research. Prostate. 2000;43:263–71. 48. Bove K, Lincoln DW, Wood PA, Hrushesky WJ. Fertility cycle influence on surgical breast cancer cure. Breast Cancer Res Treat. 2002;75:65–72. 49. Buhler KR, Santucci RA, Royai RA, Whitney SC, Vessella RL, Lange PH, Ellis WJ. Intermittent androgen suppression in the LuCaP 23.12 prostate cancer xenograft model. Prostate. 2000;43:63–70. 50. Gleave ME, Miyake H, Zellweger T, Chi K, July L, Nelson C, Rennie P. Use of antisense oligonucleotides targeting the antiapoptotic gene, clusterin/testosterone-repressed prostate message 2, to enhance androgen sensitivity and chemosensitivity in prostate cancer. Urology. 2001;58(2 Suppl 1):39–49. 51. Clarke R. Human breast cancer cell line xenografts as models of breast cancer. The immunobiologies of recipient mice and the characteristics of several tumorigenic cell lines. Breast Cancer Res Treat. 1996;39:69–86. 52. Kleinman HK, McGarvey ML, Hassell JR, Star VL, Cannon FB, Laurie GW, Martin GR. Basement membrane complexes with biological activity. Biochemistry. 1986;25:312–18. 53. Noel A, De Pauw-Gillet MC, Purnell G, Nusgens B, Lapiere C-M, Foidart J-M. Enhancement of tumorigenicity of human breast adenocarcinoma cells in nude mice by matrigel and fibroblasts. Br J Cancer. 1993;68:909–15. 54. Eccles SA. Models for evaluation of targeted therapies of metastatic disease. In: Teicher BA, editor. Tumor models in cancer research. Totowa, NJ: Humana Press. 2002, p. 293–319. 55. Giovanella BC. Xenotransplantation of human cell cultures in nude mice. In: Teicher BA, editor. Tumor models in cancer research. Totowa, NJ: Humana Press. 2002, p. 93–97. 56. Welch DR. Technical considerations for studying cancer metastasis in  vivo. Clin Exp Metastasis. 1997;15:272–306. 57. Verschraegen C, Giovanella BC, Mendoza JT, Kozielski Aj, Stehlin JS, Jr. Specific organ metatases of human melanoma cells injected into the arterial circulation of nude mice. Anticancer Res. 1991;11:529–36. 58. Potmesil M, Vardeman D, Kozielski AJ, Mendoza J, Stehlin JS, Giovanells BC. Growth inhibition of human cancer metastases by camptothecins in newly developed xenograft models. Cancer Res. 1995;55:5637–41. 59. Chen L, Ter Haar G, Hill CR, Eccles SA, Box G. Treatment of implanted liver tumors with focused ultrasound. Ultrasound Med Biol. 1998;24:1475–88. 60. Nam D-H, Jeon H-M, Kim S, Kim MH, Lee Y-J, Lee MS, Kim H, Joo KM, Lee D-S, Price JE, Bang SI, Park W-Y. Activation of notch signaling in a xenograft model of brain ­metastasis. Clin Cancer Res. 2008;14:4059–66. 61. Euhus DM, Hudd C, LaRegina MC, Johnson FE. Tumor measurement in the nude mouse. J Surg Oncol. 1986;31:229–34. 62. Geran RI, Greenberg NH, Macdonald MM, Schumacher AM, Abbott BJ. Protocols for screening chemical agents and natural products against animal tumors and other biological systems. Cancer Chemother Rep. 1972;3(Part 3):51. 63. Zubrod CG, Schepartz S, Leiter J, Endicott KM, Carrese LM, Baker CG. The chemotherapy program of the National Cancer Institute: history, analysis, and plans. Cancer Chemother Rep. 1966;50:349–540. 64. Goldin A, Schepartz SA, Venditti JM, DeVita VT Jr. Historical development and current ­strategy of the National Cancer Institute Drug Development Program. In: DeVita VT Jr, Busch H, editors. Methods in cancer research, vol. XVI. New York: Academic. 1979, p. 165–245. 65. DeVita VT Jr, Goldin A, Oliverio VT, Rubin D, Muggia FM, Henney J, Wiernik PW, Schepartz S, Ziegler J. The drug development and clinical trials programs of the Division of Cancer Treatment, National Cancer Institute. Cancer Clin Trials. 1979;2:195–216.

120

M. Liu and D. Hicklin

66. Girit IC, Jure-Kunkel M, McIntyre KW. A structured light-based system for scanning subcutaneous tumors in laboratory animals. Comp Med. 2008;58:264–70. 67. Wu J, Houghton PJ. Interval approach to assessing antitumor activity for tumor xenograft studies. Pharmaceut Statist. 2009; 755–762. 68. Conway TF Jr, Sabel MS, Sugano M, Frelinger JG, Egilmez NK, Chen F, Bankert RB. Growth of human tumor xenografts in SCID mice quantified using an immunoassay for tumor marker protein in serum. J Immunol Methods. 2000;233:57–65. 69. Hoffman R. Green fluorescent protein imaging of tumour growth, metastasis, and angiogenesis in mouse models. Lancet Oncol. 2002;3:546–56. 70. Contag CH, Jenkins K, Contag PR, Negrin RS. Use of reporter genes for optical measurements of neoplastic disease in vivo. Neoplasia. 2000;2:41–52. 71. Klerk CP, Overmeer RM, Niers TM, Versteeg HH, Richel DJ, Buckle T, Van Noorden CJ, van Tellingen O. Biotechniques. 2007;43:7–13. 72. Chambers AF, MacDonald IC, Schmidt EE, Morris VL, Groom AC. Preclinical assessment of anti-cancer therapeutic strategies using in vivo videomicroscopy. Cancer Metastasis Rev. 1998–99;17:263–69. 73. MacDonald IC, Groom AC, Chambers AF. Cancer spread and micrometastasis development: quantitative approaches for in vivo models. Bioessays. 2002;24:885–93. 74. Jain RK, Schlenger K, Hockel M, Yuan F. Quantitative angiogenesis assays: progress and problems. Nat Med. 1997;3:1203–8. 75. Teicher BA, Ara G, Herbst R, Takeuchi H, Keyes S, Northey D. PEG-hemoglobin: effects on tumor oxygenation and response to chemotherapy. In Vivo. 1997;11:301–11. 76. Furman-Haran E, Grobgeld D, Margalit R, Degani H. Response of MCF7 human breast cancer to tamoxifen: evaluation by the three-time-point, contrast-enhanced magnetic resonance imaging method. Clin Cancer Res. 1998;4:2299–2304. 77. Leach MO. Application of magnetic resonance imaging to angiogenesis in breast cancer. Breast Cancer Res. 2001;3:22–27. 78. Gupta N, Price PM, Aboagye EO. PET for in vivo pharmacokinetic and pharmacodynamic measurements. Eur J Cancer. 2002;38:2094–2107. 79. Malich A, Bohm T, Fritsch T, Facius M, Freesmeyer MG, Anderson R, Fleck M, Kaiser WA. Animal-based model to investigate the minimum tumor size detectable with an electrical impedance scanning technique. Acad Radiol. 2003;10:37–44. 80. Kennel SJ, Davis IA, Branning J, Pan H, Kabalka GW, Paulus MJ. High resolution computed tomography and MRI for monitoring lung tumor growth in mice undergoing radioimmunotherapy: correlation with histology. Med Phys. 2000;27:1101–7. 81. Waterhouse RN, Chapman J, Izard B, Donald A, Belbin K, O’Brien JC, Collier TL. Examination of four 123I-labeled piperidine-based sigma receptor ligands as potential melanoma imaging agents: initial studies in mouse tumor models. Nucl Med Biol. 1997;24:587–93. 82. Hollingshead MG, Alley MC, Camalier RF, Abbott BJ, Mayo JG, Malspeis L, Grever MR. In vivo cultivation of tumor cells in hollow fibers. Life Sci. 1995;57:131–41. 83. Curwen JO, Wedge SR. The use and refinement of rodent models in anti-cancer drug ­discovery: a review. Altern Lab Anim. 2009;37:173–80. 84. Sausville EA, Burger AM. Contributions of human xenografts to anticancer drug ­development. Cancer Res. 2006;66:3351–54. 85. Hao D, Rowinsky EK. Inhibiting signal transduction: recent advances in the development of receptor tyrosine kinase and Ras inhibitors. Cancer Invest. 2002;20:387–404. 86. Liu M, Bishop WR, Nielsen LL, Bryant MS, Kirschmeier P. Orally bioavailable farnesyl transferase inhibitors as anticancer agents in transgenic and xenograft models. Methods Enzymol. 2001;333:306–18. 87. Liu M, Bryant MS, Chen J, Lee S, Yaremko B, Lipari P, Malkowski M, Ferrari E, Nielsen L, Prioli N, Dell J, Sinha D, Syed J, Korfmacher WA, Nomeir AA, Lin CC, Wang L, Taveras AG, Doll RJ, Njoroge FG, Mallams AK, Remiszewski S, Catino JJ, Girijavallabhan VM, Bishop WR. Antitumor activity of SCH 66336, an orally bioavailable tricyclic inhibitor of

5  Human Tumor Xenograft Efficacy Models

121

farnesyl protein transferase, in human tumor xenograft models and wap-ras transgenic mice. Cancer Res. 1998;58:4947–56. 88. Liu M, Bryant MS, Chen J, Lee S, Yaremko B, Li Z, Dell J, Lipari P, Malkowski M, Prioli N, Rossman RR, Korfmacher WA, Nomeir AA, Lin CC, Mallams AK, Doll RJ, Catino JJ, Girijavallabhan VM, Kirschmeier P, Bishop WR. Effects of SCH 59228, an orally bioavailable farnesyl protein transferase inhibitor, on the growth of oncogene-transformed fibroblasts and a human colon carcinoma xenograft in nude mice. Cancer Chemother Pharmacol. 1999;43:50–58. 89. Wang YC, Kulp SK, Wang D, Yang CC, Sargeant AM, Hung JH, Kashida Y, Yamaguchi M, Chang GD, Chen CS. Targeting endoplasmic reticulum stress and Akt with OSU-03012 and gefitinib or erlotinib to overcome resistance to epidermal growth factor receptor inhibitors. Cancer Res. 2008;68:2820–30. 90. Wu W, O’Reilly MS, Langley RR, Tsan RZ, Baker CH, Bekele N, Tang XM, Onn A, Fidler IJ, Herbst RS. Expression of epidermal growth factor (EGF)/transforming growth factoralpha by human lung cancer cells determines their response to EGF receptor kinase inhibition in the lung of mice. Mol Cancer Ther. 2007;6:2652–63. 91. Das AK, Sato M, Story MD, Peyton M, Graves R, Redpath S, Girard L, Gazdar AF, Shay JW, Minna JD, NIrodi CS. Non-small-cell lung cancers with kinase domain mutations in the epidermal growth factor receptor are sensitive to ionizing radiation. Cancer Res. 2006;66:9601–8. 92. Tang Z, Du R, Jiang S, Wu C, Barkauskas DS, Richey J, Molter J, Lam M, Flask C, Gerson S, Dowlati A, Liu L, Lee Z, Halmos B, Wang Y, Kern JA, Ma PC. Dual MET-EGFR combinatorial inhibition against T790M-EGFR-mediated erlotinib-resistant lung cancer. Br J Cancer. 2008;99:911–22. 93. Morgan MA, Parsels LA, Kollar LE, Normolle DP, Maybaum J, Lawrence TS. The combination of epidermal growth factor receptor inhibitors with gemcitabine and radiation in pancreatic cancer. Clin Cancer Res. 2008;14:5142–49. 94. Rose WC, Wild R. Therapeutic synergy of oral taxane BMS-275183 and cetuximab versus human tumor xenografts. Clin Cancer Res. 2004;10:7413–17. 95. Balin-Gauthier D, Delord JP, Rochaix P, Mallard V, Thomas F, Hennebelle I, Bugat R, Canal P, Allal C. In vivo and in vitro antitumor activity of oxaliplatin in combination with cetuximab in human colorectal tumor cell lines expressing different level of EGFR. Cancer Chemother Pharmacol. 2006;57:709–18. 96. Milas L, Mason K, Hunter N, Petersen S, Yamakawa M, Ang K, Mendelsohn J, Fan Z. In vivo enhancement of tumor radioresponse by C225 antiepidermal growth factor receptor antibody. Clin Cancer Res. 2000;6:701–8. 97. Feng FY, Lopez CA, Normolle DP, Varambally S, Li X, Chun PY, Davis MA, Lawrence TS, Nyati MK. Effect of epidermal growth factor receptor inhibitor class in the treatment of head and neck cancer with concurrent radiochemotherapy in vivo. Clin Cancer Res. 2007;13:2512–18. 98. Rosenthal EL, Kulbersh BD, Duncan RD, Zhang W, Magnuson, JS, Carroll WR, Zinn K. In vivo detection of head and neck cancer orthotopic xenografts by immunofluorescence. Laryngoscope. 2006;116:1636–41. 99. Chang YS, Adnane J, Trail PA, Levy J, Henderson A, Xue D, Bortolon E, Ichetovkin M, Chen C, McNabola A, Wilkie D, Carter CA, Taylor IC, Lynch M, Wilhelm S. Sorafenib (BAY 43-9006) inhibits tumor growth and vascularization and induces tumor apoptosis and hypoxia in RCC xenograft models. Cancer Chemother Pharmacol. 2007;59:561–74. 100. Liu L, Cao Y, Chen C, Zhang X, McNabola A, Wilkie D, Wilhelm S, Lynch M, Carter C. Sorafenib blocks the RAF/MEK/ERK pathway, inhibits tumor angiogenesis, and induces tumor cell apoptosis in hepatocellular carcinoma model PLC/PRF/5. Cancer Res. 2006;66:11851–58. 101. Wang Z, Zhou J, Fan J, Qiu SJ, Yu Y, Huang XW, Tang ZY. Effect of rapamycin alone and in combination with sorafenib in and orthotopic model of human hepatocellular carcinoma. Clin Cancer Res. 2008;14:5124–30.

122

M. Liu and D. Hicklin

102. Hillman GG, Singh-Gupta V, Zhang H, Al-Bashir AK, Katkuri Y, Li M, Yunker CK, Patel AD, Abrams J, Haacke EM. Dynamic contrast-enhanced magnetic resonance imaging of vascular changes induced by sunitinib in papillary renal cell carcinoma xenograft tumors. Neoplasia. 2009;11:910–20. 103. Ikezoe T, Yang Y, Nishioka C, Bandobashi K, Nakatani H, Taguchi T, Koeffler HP, Taguchi H. Effect of SU11248 on gastrointestinal stromal tumor-T1 cells: enhancement of growth inhibition via inhibition of 3-kinase/Akt/mammalian target of rapamycin signaling. Cancer Sci. 2006;97:945–51. 104. 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;24:6657–66. 105. Gril B, Palmieri D, Bronder JL, Herring JM, Vega-Valle E, Feigenbaum L, Liewehr DJ, Steinberg SM, Merino MJ, Rubin SD, Steeg PS. Effect of lapatinib on the outgrowth of metastatic breast cancer cells to the brain. J Natl Cancer Inst. 2008;100:1092–1103. 106. Barok M, Balazs M, Nagy P, Rakosy Z, Treszl A, Toth E, Juhasz I, Park JW, Isola J, Vereb G, Szollosi J. Trastuzumab decreases the number of circulating and disseminated tumor cells despite trastuzumab resistance of the primary tumor. Cancer Lett. 2008;260:198–208. 107. Hardee ME, Eapen RJ, Rabbani ZN, Dreher MR, Marks J, Blackwell KL, Dewhirst MW. Her2/neu signaling blockade improves tumor oxygenation in a multifactorial fashion in Her2/neu+ tumors. Cancer Chemother Pharmacol. 2009;63:219–28. 108. Kedrin D, Wyckoff J, Boimel PJ, Coniglio SJ, Hynes NE, Arteaga CL, Segall JE. ERBB1 and ERBB2 have distinct functions in tumor cell invasion and intravasion. Clin Cancer Res. 2009;15:3733–39. 109. Hoffmann J, Fichtner I, Lemm M, Lienau P, Hess-Stumpp H, Rotgeri A, Hoffmann B, Klar U. Sagopilone crosses the blood-brain barrier in  vivo to inhibit brain tumor growth and metastases. Neuro Oncol. 2009;11:158–66. 110. Keir ST, Dolan ME, Pegg AE, Lawless A, Moschel RC, Bigner DD, Friedman HS. O6-benzylguanine-mediated enhancement of nitrosourea activity in Mer-central nervous system tumor xenografts – implications for clinical trials. Cancer Chemother Pharmacol. 2000;45:437–40. 111. Kitange GJ, Carlson BL, Mladek AC, Decker PA, Schroeder MA, Wu W, Grogan PT, Giannini C, Bellman KV, Buckner JC, James CD, Sarkaria JN. Evaluation of MGMT promoter methylation status and correlation with temozolomice response in orthotopic glioblastoma xenograft model. J Neurooncol. 2009;92:23–31. 112. Golemovic M, Verstovsek S, Giles F, Cortes J, Manshouri T, Manley PW, Mestan J, Dugan M, Alla d L, Griffin JD, Arlinghaus RB, Sun T, Kantarjian H, Beran M. AMN107, a novel aminopyrimidine inhibitor of Bcr-Abl, has in vitro activity against imatinib-resistant chronic myeloid leukemia. Clin Cancer Res. 2005;11:4941–47. 113. Luo FR, Yang Z, Camuso A, Smykla R, McGlinchey K, Fager K, Flefleh C, Castaneda S, Inigo I, Kan D, Wen ML, Kramer R, Blackwood-Chirchir A, Lee FY. Dasatinib (BMS354825) pharmacokinetics and pharmacodynamics biomarkers in animal models predict optimal clinical exposure. Clin Cancer Res. 2006;12:7180–86. 114. Prenen H, Deroose C, Vermaelen P, Sciot R, Debiec-Rychter M, Stroobants S, Mortelmans L, Schoffski P, Van Oosterom A. Establishment of a mouse gastrointestinal stromal tumor model and evaluation of response to imatinib by small animal positron emission tomography. Anticancer Res. 2006;26:1247–52. 115. Damiano V, Caputo R, Garofalo S, Bianco R, Rosa R, Merola G, Gelardi T, Racioppi L, Fontanini G, De Placido S, Kandimalla ER, Agrawal S, Ciardiello F, Tortora G. TLR9 agonist acts by different mechanisms synergizing with bevacizumab in sensitive and cetuximabresistant colon cancer xenografts. Proc Natl Acad Sci USA. 2007;104:12468–473. 116. Lee FY, Covello KL, Castaneda S, Hawken DR, Kan D, Lewin A, Wen ML, Ryseck RP, Fairchild CR, Fargnoli J, Kramer R. Synergistic antitumor activity of ixabepilone (BMS247550) plus bevacizumab in multiple in  vivo tumor models. Clin Cancer Res. 2008;14:8123–31.

5  Human Tumor Xenograft Efficacy Models

123

117. Mizobe T, Ogata Y, Murakami H, Akagi Y, Ishibashi N, Mori S, Sasatomi T, Shirouzu K. Efficacy of the combined use of bevacizumab and irinotecan as a postoperative adjuvant chemotherapy in colon carcinoma. Oncol Rep. 2008;20:517–23. 118. Zhong W, Hansen R, Li B, Cai Y, Salvador C, Moore GD, Yan J. Effect of yeast-derived beta-glucan in conjugation with bevacizumab for the treatment of human lung adenocarcinoma in subcutaneous and orthotopic xenograft models. J Immunother. 2009;32:703–12. 119. Higgins B, Kolinsky K, Linn M, Adames V, Zhang YE, Moisa C, Dugan U, Heimbrook D, Packman K. Antitumor activity of caecitabine and bevacizumab combination in a human estrogen receptor-negative breast adenocarcinoma xenograft model. Anticancer Res. 2007;27:2279–87. 120. 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;14:5459–65. 121. Mathieu V, De Neve N, Le Mercier M, Dewelle J, Gaussin J-F, Dehoux M, Kiss R, Lefranc F. Combining bevacizumab with temozolomide increases the antitumor efficacy of temozolomide in a human glioblastoma orthotopic xenograft model. Neoplasia. 2008;10:1383–92. 122. Siemann DW, Shi W. Dual targeting of tumor vasculature: combining avastin and vascular disrupting agents (CA4P or Oxi4503). Anticancer Res. 2008;28:2027–31. 123. Egawa T, Kubota T, Suto A, Otani Y, Furukawa T, Saikawa Y, Watanabe M, Kumai K, Kitajima M. Antitumor activity of doxorubicin in combination with docetaxel against human breast cancer xenografts. In Vivo. 2003;17:23–28. 124. Ciomi M, Croci V, Stellari F, Amboldi N, Giavarini R, Pesenti E. Antitumor activity of edtecarin in breast carcinoma models. Cancer Chemother Pharmacol. 2007;60:229–35. 125. Fulzele SV, chatterjee A, Shaik MS, Jackson T, Ichite N, Singh M. 15-Deoxy-delta 12, 14-prostaglandin J2 enhances docetaxel anti-tumor activity against A549 and H460 nonsmall-cell lung cancer cell lines and xenograft tumors. Anticancer Drugs. 2007;18:65–78. 126. Liu G, Taylor SA, Marrinan CH, Hsieh Y, Bishop WR, Kirschmeier P, Long BJ. Continuous and intermittent dosing of lonafarnib potentiates the therapeutic efficacy of docetaxel on preclinical human prostate cancer models. Int J Cancer. 2009;125:2711–20. 127. Tanaka M, Obata T, Sasaki T. Evaluation of antitumor effects of docetaxel (taxotere) on human gastric cancers in vitro and in vivo. Eur J Cancer. 1996;32A:226–30. 128. Bradshaw-Pierce EL, Steinhauer CA, Raben D, Gustafson DL. Pharmacokinetic-directed dosing of vandetanib and docetaxel in a mouse model of human squamous cell carcinoma. Mol Cancer Ther. 2008;7:3006–17. 129. Yamada H, Maki H, Takeda Y, Orita S. Evaluation of combined nedaplatin and docetaxel therapy for human head and neck cancer in vivo. Anticancer Res. 2006;26(2A):989–94. 130. Blackledge G, Averbuch S, Kay A, Barton J. Anti-EGF receptor therapy. Prostate Cancer Prostatic Dis. 2000;3:296–302. 131. Ciardiello F, Tortora G. A novel approach in the treatment of cancer: targeting the epidermal growth factor receptor. Clin Cancer Res. 2001;7:2958–70. 132. Sirotnak FM, She Y, Lee F, Chen J, Scher HI. Studies with CWR22 xenografts in nude mice suggest that ZD1839 may have a role in the treatment of both androgen-dependent and androgen-independent human prostate cancer. Clin Cancer Res. 2002;8:3870–76. 133. Morris C. The role of EGFR-directed therapy in the treatment of breast cancer. Breast Cancer Res Treat. 2002;75:S51–S55. 134. Steiner P, Joynes C, Bassi R, Wang W, Tonra JR, Hadari YR, Hicklin DJ. Tumor growth inhibition with cetuximab and chemotherapy in non-small cell lung cancer xenografts expressing wild-type and mutated epidermal growth factor receptor. Clin Cancer Res. 2007;13:1540–51. 135. Liu G, Marrinan CH, Taylor, SA, Black S, Basso BD, Kirschmeier P, Bishop WR, Liu M, Long BJ. Enhancement of the antitumor activity of tamoxifen and anastrozole by the farnesyltransferase inhibitor lonafarnib (SCH66336). Anticancer Drugs. 2007;18:923–31.

124

M. Liu and D. Hicklin

136. Teicher BA. In vivo/ex vivo and in situ assays used in cancer research: a brief review. Toxicol Pathol. 2009;37:114–22. 137. Ciardiello F, Caputo R, Bianco R, Damiano V, Pomatico G, De Placido S, Bianco AR, Tortora G. Antitumor effect and potentiation of cytotoxic drugs activity in human cancer cells by ZD-1839 (Iressa), an epidermal growth factor receptor-selective tyrosine kinase inhibitor. Clin Cancer Res. 2000;6:2053–63. 138. Sirotnak FM, Zakowski MF, Miller VA, Scher HI, Kris MG. Efficacy of cytotoxic agents against human tumor xenografts is markedly enhanced by coadministration of ZD1839 (Iressa), an inhibitor of EGFR tyrosine kinase. Clin Cancer Res. 2000;6:4885–92. 139. Giovanella BC, Vardeman DM, Williams LJ, Taylor DJ, et  al. Heterotransplantation of human breast carcinomas in nude mice: correlation between nude takes, poor prognosis and overexpression of the HER-2/neu oncogene. Int J Cancer. 1991;47:66–71. 140. Nielsen LL, Gurnani M, Porter G, Trexler S, Emerson D, Tyler RD. Development of a nude mouse model of ras-mediated neoplasia using WR21 cells from a transgenic mouse salivary tumor. In Vivo. 1994;8:295–302. 141. Tonra JR, Prewett M, Corcoran E, Hicklin DJ, Zhu Z. In vivo method for establishing ­synergy between antibodies to epidermal growth factor receptor and vascular endothelial growth factor receptor-2. Methods Mol Biol. 2009;525:545–57. 142. Bozec A, Sudaka A, Fischel J-L, Brunstein M-C, Etienne-Grimaldi M-C, Milano G. Combined effects of bevacizumab with erlotinib and irradiation: a preclinical study on a head and neck cancer orthotopic model. Br J Cancer. 2008;99:93–99. 143. Shepard HM, Jin P, Slamon DJ, Pirot Z, Maneval DC. Herceptin. Handb Exp Pharmacol, Ther Antibodies. 2008;181:183–219. 144. Sikora K. Surrogate endpoints in cancer drug development. Drug Discov Today. 2002;7:951–56. 145. Wagner JA. Overview of biomarkers and surrogate endpoints in drug development. Dis Markers. 2002;18:41–46. 146. Kim H, Likhari P, Lin CC, Nomeir AA. High-performance liquid chromatographic analysis of the anti-tumor agent SCH 66336 in cynomolgus monkey plasma and evaluation of its chiral inversion in animals. J Chromatogr B Biomed Sci Appl. 1999;728:133–41. 147. Albanell J, Rojo F, Baselga J. Pharmacodynamic studies with the epidermal growth factor receptor tyrosine kinase inhibitor ZD1839. Semin Oncol. 2001;28(5 Suppl 16):56–66. 148. Wakeling AE, Guy SP, Woodburn JR, Ashton SE, Curry BJ, Barker AJ, Gibson KH. ZD1839 (Iressa): an orally active inhibitor of epidermal growth factor signaling with potential for cancer therapy. Cancer Res. 2002;62:5749–54. 149. Ellis LM, Hicklin DJ. VEGF-targeted therapy: mechanisms of anti-tumor activity. Nat Rev Cancer. 2008;8:579–91. 150. Cheung, AM, Brown, AS, Cucevic V, Roy M, Needles A, Yang V, Hicklin DJ, Kerbel RS, Foster FS. Detecting vascular changes in tumor xenografts using micro-ultrasound and micro-CT following treatment with VEGFR-2 blocking antibodies. Ultrasound Med Biol. 2007;33:1259–68. 151. Talmadge JE, Singh RK, Fidler IJ, Raz A. Murine models to evaluate novel and conventional therapeutic strategies for cancer. Am J Pathol. 2007;170:793–804. 152. Dimasi JA, Hansen RW, Grabowski HG. The price of innovation: new estimates of drug development costs. J Health Econ. 2003;22:151–85. 153. Von Hoff DD. There are no bad anticancer agents, only bad clinical trial designs. Clin Cancer Res. 1998;4:1079–86. 154. Fidler IJ. The role of host factors and tumor heterogeneity in the testing of therapeutic agents. In: Fidler IJ, White RJ, editors. Design of models for testing cancer therapeutic agents. New York, NY: Van Nostrand Reinhold. 1982; p. 239–47.

Chapter 6

Imaging the Steps of Metastasis at the Macro and Cellular Level with Fluorescent Proteins in Real Time Robert M. Hoffman

6.1 Introduction In 2008, the Nobel Prize for chemistry was awarded for discovery and modification of green fluorescent protein (GFP) [1]. The Nobel announcement cited two uses of GFP (Nobel background), one of which was the use of GFP to track cancer cells in vivo, which was pioneered in our laboratory [2–6]. GFP was discovered in the bioluminescent jellyfish Aequorea victoria by Shimomura [7]. The GFP gene was cloned from A. victoria by Doug Prasher which enabled GFP to become the most powerful tool in cell biology [8–13]. The GFP cDNA encodes a 283-amino acid polypeptide with a molecular weight of 27 kDa [14, 15]. The monomeric GFP requires no other Aequorea proteins, substrates, or cofactors to fluoresce [16]. GFP gene gain-of-function mutants have been generated using various techniques [17–20]. For example, the GFPS65T clone has the serine-65 codon substituted with a threonine codon, which results in a single excitation peak at 490  nm [17]. Moreover, to develop higher expression in human and other mammalian cells, a humanized hGFP-S65T clone was isolated [21]. The much brighter fluorescence in the mutant clones allows for easy detection of GFP expression in transfected cells. Matz et  al. cloned six fluorescent proteins homologous to GFP. The proteins were isolated from the coral Discosoma. A red protein has been further developed for expression in mammalian cells and is now known as DsRed-2 [22]. Monomer variants of the Discosoma red protein have been developed with distinguishable colors from yellow-orange to red-orange with fruit names similar to their color, m-Cherry, m-Tomato, etc. [23]. Fluorescent proteins that can be converted from one color to another or become fluorescent upon light activation have also been developed [12, 24].

R.M. Hoffman (*) AntiCancer, Inc, Department of Surgery, University of California San Diego, 7917 Ostrow Street, San Diego, CA 92111 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_6, © Springer Science+Business Media, LLC 2011

125

126

R.M. Hoffman

6.2 The In Vivo Revolution Sparked by Fluorescent Proteins 6.2.1 Isolation of Stable High-Level Expression GFP and/or RFP Expressing Tumor Cell Lines We have isolated more than 100 GFP and red fluorescent protein (RFP) transformants of human and animal cancer cells that are stable in vitro and in vivo [2, 25–29] using the above GFP and RFP genes. The transformants are highly fluorescent in vivo in tumors formed from the cells. Using these fluorescent cancer cell lines, orthotopic transplant animal models [28–32] were developed for imaging the metastatic processes at the macro- and the single-cell level. The main techniques for producing bright stable fluorescent protein-expressing cells are described below. 6.2.1.1 Production of GFP Retrovirus The pLEIN retroviral vector (Clontech Laboratories, Inc., Palo Alto, CA) expressing GFP and the neomycin resistance gene on the same bicistronic message was used as a GFP expression vector. PT67, an NIH3T3-derived packaging cell line, expressing the 10 Al viral envelope, was purchased from Clontech Laboratories, Inc. PT67 cells were cultured in DMEM (Irvine Scientific, Santa Ana, CA) supplemented with 10% heat-inactivated fetal bovine serum (Gemini Bio-products, Calabasas, CA). For vector production, packaging cells (PT67), at 70% confluence, were incubated with a precipitated mixture of DOTAP reagent (Boehringer Mannheim, Indianapolis, IN) and saturating amounts of pLEIN plasmid for 18 h. Fresh medium was replenished at this time. The cells were examined by fluorescence microscopy 48 h after transfection. For selection, the cells were cultured in the presence of 500–2,000 mg/ml of G418 (Life Technologies, Inc., Grand Island, NY) for 7 days to select for a clone producing high amounts of a GFP retroviral vector (PT67-GFP) [33]. 6.2.1.2 Production of RFP Retroviral Vector For RFP retrovirus production, the HindIII/NotI fragment from pDsRed2 (Clontech Laboratories, Inc.), containing the full-length RFP (DsRed2) cDNA, was inserted into the HindIII/NotI site of pLNCX2 (Clontech Laboratories) that has the neomycin resistance gene to establish the pLNCX2-DsRed2 plasmid. PT67, an NIH3T3-derived packaging cell line (Clontech Laboratories) expressing the 10 Al viral envelope, was cultured in DMEM (Irvine Scientific) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gemini Bio-products). For vector production, PT67cells, at 70% confluence, were incubated with a precipitated mixture of LipofectAMINE reagent (Life Technologies, Inc., Grand Island, NY) and saturating amounts of pLNCX2DsRed2 plasmid for 18 h. Fresh medium was replenished at this time. The cells were examined by fluorescence microscopy 48 h after transfection. For selection of a clone

6  Imaging the steps of metastasis at the macro and cellular level

127

producing high amounts of a RFP retroviral vector (PT67-DsRed2), the cells were cultured in the presence of 200–1,000 mg/ml of G418 (Life Technologies) which was increased stepwise in order to select for brighter cells [34]. 6.2.1.3 Production of Histone H2B-GFP Vector The histone H2B gene has no stop codon, thereby enabling the ligation of the H2B gene to the 5¢-coding region of the A. victoria EGFP gene (Clontech Laboratories). The histone H2B-GFP fusion gene was then inserted at the HindIII/ClaI site of the pLHCX (Clontech Laboratories) that contains the hygromycin resistance gene. To establish a packaging cell clone producing high amounts of a histone H2B-GFP retroviral vector, the pLHCX histone H2B-GFP plasmid was transfected in PT67cells using the same methods described above for PT67-DsRed2. The transfected cells were cultured in the presence of 200–400 mg/ml of hygromycin (Life Technologies) for 15 days to establish stable PT67H2B-GFP packaging cells [33, 34]. 6.2.1.4 GFP or RFP Gene Transduction of Cancer Cells For GFP or RFP gene transduction, 70% confluent cultures of cancer lines are used. Cancer cells were incubated with a 1:1 precipitated mixture of retroviral supernatants of PT67-GFP or PT67-RFP cells and RPMI 1640 (Mediatech, Inc.) containing 10% FBS for 72 h. Fresh medium was replenished at this time. Cells were harvested with trypsin/EDTA 72 h after transduction and subcultured at a ratio of 1:15 in selective medium, which contained 200  mg/ml of G418. The level of G418 was increased stepwise up to 800 mg/ml. GFP- or RFP-expressing cancer cells were isolated with cloning cylinders (Bel-Art Products) using trypsin/EDTA and amplified by conventional culture methods [34]. 6.2.1.5 Establishment of Dual-Color Cancer Cells For establishing dual-color cells, RFP-expressing cancer cells were incubated with a 1:1 precipitated mixture of retroviral supernatants of PT67H2B-GFP cells and culture medium. To select the double transformants, the cells were incubated with hygromycin 72 h after transfection. The level of hygromycin was increased stepwise up to 400 mg/ml. Clones of dual-color cancer cells were isolated with cloning cylinders under fluorescence microscopy. These clones were amplified by conventional culture methods. These sublines stably expressed GFP in the nucleus and RFP in the cytoplasm [34]. Nuclear GFP expression enabled visualization of nuclear dynamics, whereas simultaneous cytoplasmic RFP expression enabled visualization of nuclear cytoplasmic ratios as well as simultaneous cell and nuclear shape changes. Thus, total cellular dynamics can be visualized in the living dualcolor cells in real time. The cell cycle position of individual living cells was readily

128

R.M. Hoffman

visualized by the nuclear–cytoplasmic ratio and nuclear morphology. Real-time induction of apoptosis was observed by nuclear size changes and progressive nuclear fragmentation. Thus, the dual-color cells are a useful tool for visualizing living-cell dynamics in vivo as well as in vitro. Drugs that could specifically perturb these processes can now be readily screened in real time in vivo [33, 35].

6.2.2 Imaging Sites of Metastasis 6.2.2.1 Patterns of Contralateral and Regional Lung Tumor Metastases Visualized by GFP Expression in Orthotopic Models After surgical orthotopic implantation (SOI) of GFP-expressing ANIP-973 human lung cancer in the left lung of nude mice, the primary tumor grew. GFP expression allowed visualization of the advancing margin of the tumor spreading throughout the ipsilateral lung. All animals explored had evidence of chest-wall invasion and local and regional spread. Metastatic contralateral tumors involved the mediastinum, contralateral pleural cavity, and the contralateral visceral pleura. While the ipsilateral tumor had a continuous and advancing margin, the contralateral tumor seems to have been formed by multiple seeding events. These observations were made possible by the stable, high-GFP fluorescence of the tumor cells [25, 26]. Contralateral hilar lymph nodes were also involved, as well as cervical lymph nodes visualized by GFP expression [25, 26]. When non-GFP-transfected ANIP was compared with GFP-transformed ANIP for metastatic capability, similar results were seen [25]. 6.2.2.2 GFP-Expressing Bone Metastases of Lung Cancer in Orthotopic Models SOI of H460-GFP human lung cancer in the left lung of nude mice was performed [28]. The implanted mice were sacrificed at 3–4 weeks at the time of significant decline in performance status. GFP fluorescence demonstrated metastases in the left lung, the contralateral lung, the chest wall, and the skeletal system. It was determined by GFP fluorescence that the vertebrae were the most involved skeletal site of metastasis. Metastasis could also be visualized in the tibia and femur marrow by GFP fluorescence [28]. 6.2.2.3 Prostate-Cancer Bone and Visceral Metastasis Visualized by GFP in Orthotopic Models A stable high GFP-expressing clone of human prostate carcinoma PC-3 was orthotopically implanted surgically in nude mice. Subsequent skeletal metastasis included the skull, rib, pelvis, femur, and tibia. All the tumors metastasized to the

6  Imaging the steps of metastasis at the macro and cellular level

129

lung, pleural membrane, and kidney. Four of five tumors metastasized to the liver, and two of five tumors metastasized to the adrenal gland. In two mice, cancer cells or small colonies were seen in the brain, and in one mouse, a few cells could be seen in the spinal cord by GFP fluorescence [29]. 6.2.2.4 GFP-Expressing Melanoma Bone and Organ Metastasis Models We have characterized metastatic properties of bright, highly stable GFP-expressing B16 mouse-melanoma and LOX human melanoma [36]. The highly fluorescent malignant-melanoma cell lines allowed the visualization of skeletal and multiorgan metastases after intravenous (iv) injection of B16 cells in C57BL/6 mice and intradermal (id) injection of LOX cells in nude mice. The melanoma cell lines were transduced with the pLEIN-expression retroviral vector containing the GFP and neomycin resistance genes. Extensive bone and bone-marrow metastases of B16F0 were visualized by GFP expression when the animals were sacrificed 3 weeks after cell implantation. This was the first observation of experimental skeletal metastases of melanoma, which was made possible by GFP expression. For both cell lines, metastases were visualized in many other organs, including the brain, lung, pleural membrane, liver, kidney, adrenal gland, lymph nodes, muscle, and skin, by GFP fluorescence. 6.2.2.5 GFP-Expressing Brain Metastasis in Orthotopic Models With the use of GFP imaging, we have observed spontaneous metastasis to the brain in three orthotopic nude mouse model systems of human cancer: the PC-3 human prostate cancer cell line [29], the LOX human melanoma cell line [36], and spinal cord glioma model using the U87 human glioma cell line [37, 38]. 6.2.2.6 GFP-Expressing Experimental Multi-organ Metastases in Nude Mice CHO-K1 GFP transformants, injected via the tail vein, were visualized by GFP expression in the peritoneal wall vessels down to the single-cell level [39]. These cells formed emboli in the capillaries of the lung, liver, kidney, spleen, ovary, adrenal gland, thyroid gland, and brain. ANIP GFP cells were injected into the tail vein of nude mice, which were sacrificed at 4 and 8 weeks. In both groups, numerous micrometastatic colonies were detected in the lung by GFP expression in fresh tissue [39]. Even 8 weeks after injection, in most of the mice, colonies were not obviously further developed compared to that in mice sacrificed at 4 weeks [39]. Numerous small colonies, which ranged in size to less than 10 cells, were detected at the lung surface in both groups. After 8 weeks, metastases in the brain, the submandibular gland, the lung, the pancreas, the bilateral adrenal glands, the peritoneum, and the pulmonary hilum lymph nodes were visualized by GFP

130

R.M. Hoffman

expression. Actively colonizing as well as dormant tumor cells were visualized in the lung [39]. Dormant micrometastasis is one of the most important steps in tumor progression [40]. In recent studies, the mechanism of dormancy was studied with regard to angiogenesis and other regulators of tumor colonization using GFP [41].

6.2.3 Whole-Body Fluorescence Optical Tumor Imaging of Tumor Growth and Metastasis We reported in 2000 for the first time, noninvasive imaging, in real time, of fluorescent tumors growing and metastasizing in live mice. The whole-body optical imaging system is external, simple, and noninvasive, and allows images to be captured from freely moving animals with internal GFP-expressing tumors. It affords unprecedented continuous visual monitoring of malignant growth and spread within intact animals. B16F0-GFP mouse melanoma cells were injected into the tail vein or portal vein of 6-week-old C57BL/6 and nude mice. Whole-body optical images showed metastatic lesions in the brain, liver, and bone of B16F0-GFP that were used for real-time, quantitative measurement of tumor growth in each of these organs. Whole-body optical images showed, in real time, growth of the GFPexpressing tumor on the colon and metastatic lesions in the liver and skeleton. Imaging was with either a trans-illuminated epifluorescence microscope or a fluorescence light box and thermoelectrically-cooled, color charge-coupled device (CCD) camera. The depth to which metastasis and micrometastasis could be imaged depended on their size and brightness. The simple, noninvasive, and highly selective imaging of growing tumors, made possible by strong GFP fluorescence, enables the detailed imaging of tumor growth and metastasis formation [3].

6.2.4 Whole-Body Imaging of RFP Pancreatic Cancer Progression A highly fluorescent, RFP-expressing pancreatic cancer model was orthotopically established in nude mice. The MIA-PaCa-2 human pancreatic cancer cell line was transduced with RFP and grown subcutaneously. Fluorescent tumor fragments were then surgically transplanted into the nude mouse pancreas. Groups treated with intraperitoneal gemcitabine or intravenous irinotecan were sequentially imaged to compare, in real time, the antimetastatic and antitumor effects of these agents compared with untreated controls. Rapid tumor growth and widespread metastases developed in untreated mice within 2 weeks, leading to a median survival of 21 days. Significant tumor growth suppression and consequent increase in survival (32.5 days, P = 0.009) were achieved with CPT-11. Gemcitabine highly improved survival (72 days, P = 0.004) by inducing transient tumor regression over the first 3 weeks. However, at this time, growth and dissemination occurred despite continued

6  Imaging the steps of metastasis at the macro and cellular level

131

treatment, suggesting the development of tumor resistance. The antimetastatic efficacy of each drug was followed noninvasively in real time by imaging the RFPexpressing tumor and metastases, and was confirmed by fluorescent open imaging of autopsy specimens. This highly metastatic model reliably simulates the aggressive course of human pancreatic cancer. Noninvasive, sequential imaging permits the quantification of tumor growth and dissemination and, thereby, real-time evaluation of therapeutic efficacy [42].

6.2.5 Whole-Body Imaging of RFP Prostate Cancer Progression A whole-body imageable spontaneous metastatic model of human prostate cancer was developed by SOI and visualized by RFP expression. Human prostate cancer PC-3 cells were transduced with the pLNCX2-DsRed-2-RFP retroviral vector containing the RFP and neomycin resistance genes. A stable RFP-expressing PC-3 clone was injected subcutaneously in nude mice. Stable high-level expression of RFP was maintained in the subcutaneously-growing tumors. To utilize RFP expression for metastasis studies, fragments of the subcutaneously-growing tumor, which were comprised of RFP-expressing cells, were implanted by SOI in the prostate of nude mice. Primary tumor growth, progression, and subsequent lymphatic metastases were visualized in live, intact animals in real time by wholebody RFP fluorescence imaging. In total, 100% of the experimental animals developed lymphatic metastasis, the growth of which was monitored in real time by whole-body imaging. The aggressive lymphatic metastasis in this model reflects one of the major metastatic routes of prostate cancer in human patients. Intravital RFP imaging visualized single cancer cells in the lung and bladder. Open RFP imaging at autopsy visualized extensive primary growth and highly disseminated lymph node metastases [43].

6.2.6 Whole-Body Imaging of GFP Colon Cancer Progression Human colon cancer HCT-116 cells were transfected with GFP and subcutaneously injected into BALB/c nude male mice. Once subcutaneous xenografts were established, they were excised and orthotopically implanted into 32 other male BALB/c nude mice using SOI. The animals were serially imaged and euthanized at 6–8 weeks post-implantation. Tissues were procured and processed for hematoxylin and eosin analysis. All 32 animals demonstrated primary tumor growth, invasion, and peritoneal spread. Liver metastases were identified in 15/32 (47%), and lung metastases were confirmed in 13/32 (41%). In total, 19/32 (59%) animals demonstrated distant metastatic colony formation. This orthotopic model of colon cancer demonstrates local invasion and distant colony formation in the process of metastases [44].

132

R.M. Hoffman

In summary, many different types of cancer models can be used for whole-body imaging, including orthotopic, colon [3], prostate [43], pancreas [42], bone [45], brain [46], and other cancer models [4]. The relative transparency of the footpad reduces the scatter of fluorescent light emitted from the tumor, and the presence of relatively few resident blood vessels makes it an excellent tumor transplantation site for whole-body tumor angiogenesis imaging [47]. Whole-body imaging of fluorescent cells is possible in essentially any organ. Examples include Fig. 6.1, which shows whole-body imaging of GFP-labeled cancer cells in the liver and skull [3]. Figure 6.2 is an example of “molecular imaging,” with whole-body images of a mouse lymphoma arising from a transgenic Em-Myc mouse. The virulence of the tumor is determined by genes controlling apoptosis, such as p53 and BCL2. Fluorescent tumor cells can be whole-body imaged in numerous lymphatic and non-lymphatic organs [48]. Figure 6.3 demonstrates the use of RFP (DsRed2). In this case, human glioma U87 is implanted in the brain of nude mice and whole-body imaging [46] demonstrates growth of the tumor over a 5-week period. Figure 6.4 demonstrates whole-body imaging of a highly metastatic pancreatic cancer-expressing RFP compared to open imaging of the same tumor. Comparison of the two-dimensional image with the measured volume of the tumor in the opened animal results in a straight line (r = 0.89). These results demonstrate that the easily acquired whole-body images correlate with tumor volume, thereby validating the imaging method for studying tumor growth and the efficacy of antitumor drugs [42, 49].

6.3 Imaging Bacterial Targeting of Tumors We have developed an effective bacterial cancer therapy strategy by targeting viable tumor tissue using Salmonella typhimurium auxotrophs that we have generated which grow in viable and necrotic areas of tumors. However, the auxotrophy severely restricts the growth of these bacteria in normal tissue. The S. typhimurium A1-R mutant, which is auxotrophic for leu-arg and has high antitumor virulence, was developed in our laboratory. In vitro, A1-R infects tumor cells and causes nuclear destruction. A1-R was initially used to treat metastatic human prostate and breast tumors that had been orthotopically implanted in nude mice. Of the treated mice, 40% were cured completely and survived as long as non-tumor-bearing mice. A1-R administered iv to nude mice with primary osteosarcoma and lung metastasis was highly effective, especially against metastasis. A1-R was also targeted to both axillary lymph and popliteal lymph node metastases of human pancreatic cancer and fibrosarcoma, respectively, as well as lung metastasis of the fibrosarcoma in nude mice. The bacteria were delivered via a lymphatic channel to target the lymph node metastases and systemically via the tail vein to target the lung metastasis. The metastases were cured without the need of chemotherapy or any other treatment. A1-R was administered intratumorally to nude mice with an orthotopically-transplanted human pancreatic tumor. The primary pancreatic cancer regressed without

Fig. 6.1  Comparison of external and internal images of bone metastasis. (a) External images of tumors in the skeletal system including the skull (small arrowheads), scapula (thick arrows), spine (fine arrows), and liver metastasis (hollow arrowheads) in a dorsal view of a live, intact nude mouse. (b–i) Series of external fluorescence images of metastatic lesions in the skull, rips, spine and tibia (b, d, f and h) compared with corresponding images of the exposed skeletal metastases (c, e, g and i) Bars 1,280 mm [49]

134

R.M. Hoffman

Fig. 6.2  Whole-body fluorescence imaging allows the visualization of lymphoma dissemination. p53-null and Bcl2-overexpressing lymphomas are highly disseminated, infiltrating liver, kidneys, lungs (marked), and brain, whereas the control (ctrl.) lymphoma is restricted to the lymphoid compartment [49]

additional chemotherapy or any other treatment. A1-R was also effective against pancreatic cancer liver metastasis when administered intrasplenically to nude mice. The approach described here, where bacterial monotherapy effectively treats primary and metastatic tumors, is a significant improvement over previous bacterial tumor therapy strategies that require combination with toxic chemotherapy. Three promoter clones engineered in S. enterica typhimurium were identified to have enhanced expression in bacteria growing in tumors relative to those growing

6  Imaging the steps of metastasis at the macro and cellular level

135

Fig. 6.3  Whole-body imaging of a brain tumor. Real-time whole-body imaging of a U87-GFP human glioma growing in the brain of a nude mouse at (a) 1 week, (b) 3 weeks, and (c) 5 weeks after surgical orthotopic implantation [49]

in the spleen. The expression of therapeutics in Salmonella under the regulation of one or more promoters that are activated preferentially in tumors has the potential to improve the efficacy of Salmonella tumor therapy. Exploitation of the tumorkilling capability of Salmonella has great promise for a new paradigm of cancer therapy [50].

6.4 Advantages of GFP Imaging The GFP approach has several important advantages over other optical approaches to imaging tumor growth in vivo. In comparison with the luciferase reporter, GFP has a much stronger signal, and therefore can be used to image unrestrained animals; irradiation with non-damaging blue light is the only step needed. Images can be captured with fairly simple apparatus and there is no need for total darkness. The fluorescence intensity of GFP is strong [13, 17–19] and the protein sequence of GFP has also been “humanized,” which enables it to be highly expressed in mammalian cells [21]. Importantly, unlike luciferase, fluorescent proteins come in a multitude of colors [23], allowing for multiple events to be imaged. In addition, GFP fluorescence is relatively unaffected by the external environment, as the chromaphore is protected by the three-dimensional structure of the protein [16]. A triple fusion reporter vector harboring a Renilla luciferase reporter gene, a reporter gene encoding a monomeric RFP, and a mutant herpes simplex virus type thymidine kinase was tested in vivo. A highly sensitive cooled CCD camera that is compatible with both luciferase and fluorescence imaging compared these two signals from the fused reporter gene expressed with a lentivirus vector in 293T cells implanted in nude mice. The signal from RFP was found to be approximately 1,000 times stronger than that from luciferase [51]. The weak signal from luciferase necessitates photon counting, with the construction of a pseudo-image in  vivo rather than true imaging, therefore greatly reducing resolution and precluding the

136

R.M. Hoffman

Fig. 6.4  Correlation of whole-body and open images. (a) High-level expression of red fluorescent protein (RFP) in MIA-PaCa-2 cells in vitro. (b) External and (c) open images of a single, representative control mouse at autopsy on day 17 after surgical orthotopic implantation. A strong correlation can be observed between the fluorescence visualized externally and that obtained after laparotomy. (d) The liver, (e) spleen, and other solid organs were removed at autopsy and examined for evidence of metastatic disease. (f) Hematoxylin and eosin staining. (g) Intra-abdominal ascites. (h) Colonies of MIA-PaCa-2-RFP clones were easily retrieved and cultured from aspirated ascites fluid [49]

6  Imaging the steps of metastasis at the macro and cellular level

137

Table 6.1  Comparison of optical imaging methods for tumor growth and metastasis [49] Parameter GFP/RFP Luciferase References Strength of signal 6.6 × 109 pixels/s/cm2/ 7.4 × 106 pixels/s/cm2/ [51] steradian steradian Minimum number of cells 1 300 [2, 83] imageable in vitro Minimum number of cells 1 3,000 [2, 53, 64,84] imageable in vivo Need for substrate No Yes [2, 53, 84] Need for anesthesia No Yes [53, 84] Method of visualization Direct imaging Photon counting (pseudo- [3, 84] image) Multi-color imaging Yes No [53] Stability of signal Yes No [4, 52] Need for excitation light Yes No [4, 84]

in vivo cellular imaging that is an important feature of GFP imaging. In addition, the rapid clearance of the injected luciferase results in an unstable signal that makes comparison of data difficult [52]. The stronger signals from fluorescent proteins allow much more cost-efficient instrumentation. To overcome limits on fluorescent protein imaging imposed by the skin, reversible skin-flap window models have been developed that allow single-cell imaging on most organs of the mouse [53]. The main advantage of luciferase-based imaging is that no excitation light is required. See Table 6.1 for a comparison of GFP and luciferase imaging [49].

6.5 Viral Labeling of Tumors with GFP in Live Animals 6.5.1 Selective In Vivo Tumor Delivery of the Retroviral Green Fluorescent Protein Gene to Report Future Occurrence of Metastasis The GFP gene was administered to intraperitoneally-growing human stomach cancer in nude mice in order to visualize future regional and distant metastases. GFP retroviral supernatants were ip-injected from day 4 to day 10 following ip implantation of the cancer cells. Tumor and metastasis fluorescence was visualized every other week using fluorescence optics via a laparotomy on the tumorbearing animals. No normal tissues were found to be transduced by the GFP retrovirus. Within 2 weeks after retroviral GFP delivery, GFP-expressing tumor cells were observed in gonadal fat, greater omentum, and intestine, indicating that these primary intraperitoneally-growing tumors were efficiently transduced by the GFP gene and could be visualized by its expression. At the second and third laparotomies at 4 and 6 weeks, respectively, GFP-expressing tumor cells were

138

R.M. Hoffman

found spreading to lymph nodes in the mesentery. At the fourth laparotomy at 8 weeks, widespread tumor growth, including liver metastasis, was observed. Thus, reporter-gene transduction of the primary tumor enabled detection of its subsequent metastasis [54].

6.5.2 Telomerase-Dependent Adenovirus to Label Tumors In Vivo for Surgical Navigation Cancer surgery requires the complete and precise identification of malignant tissue margins, including the smallest disseminated lesions. Internal GFP fluorescence can intensely illuminate even single cells but requires GFP sequence transcription within the cell. Introducing and selectively activating the GFP gene in malignant tissue in vivo are made possible by the development of OBP-401, a telomerase-dependent, replication-competent adenovirus expressing GFP. This potentially powerful adjunct to surgical navigation was demonstrated in two nude mouse models that represent difficult surgical challenges – the resection of widely-disseminated cancer. HCT-116, a model of intraperitoneal disseminated human colon cancer, was labeled by virus injection into the peritoneal cavity. A549, a model of pleural dissemination of human lung cancer, was labeled by virus administered into the pleural cavity. Only the malignant tissue fluoresced brightly in both models. In the intraperitoneal model of disseminated cancer, fluorescence-guided surgery enabled resection of all tumor nodules labeled with GFP by OBP-401. These results suggest that adenoviralGFP labeling of tumors in patients can enable fluorescence-guided surgical navigation (Fig. 6.5) [55].

Fig. 6.5  Fluorescence-guided surgical removal of peritoneal disseminated HCT-116 tumors after GFP labeling with OBP-401 in vivo. Noncolored HCT-116 human colon cancer cells were injected into the abdominal space of nude mice. Ten days later, 1 × 108 PFU OBP-401 were ip injected. (a) Disseminated nodules were efficiently labeled and noninvasively visualized by GFP expression 5 days after virus administration. (b) Under general anesthesia, laparotomy was performed to remove intra-abdominal disease under GFP-guided navigation. (c) After removal of Disseminated nodules visualized by GFP-guided navigation. (Scale bars: a–c, 10 mm) [55]

6  Imaging the steps of metastasis at the macro and cellular level

139

6.6 Color-Coded Imaging of the Tumor Microenvironment 6.6.1 Color-Coded Imaging of the Tumor–Host Interaction Using Colored Host Mice Fluorescent proteins have the properties of being very bright with high quantum yield and are available in many colors. Tumor–host models can be made with transgenic nude mice expressing GFP or RFP or cyan fluorescent protein (CFP) in almost all cells and tissues (Fig. 6.6). Colored transgenic nude mice are particularly useful, as they can accept human tumors cells. For example, when tumor cells expressing RFP are implanted in mice expressing GFP, various types of tumor–host interactions can be observed, including those involving host blood vessels, lymphocytes, tumor-associated fibroblasts, macrophages, dendritic cells, and others. The “color-coded” tumor–host models enable imaging and, therefore, a deeper understanding of the host cells involved and their function in tumor progression [56]. 6.6.1.1 Transgenic GFP Nude Mouse We have developed a transgenic GFP nude mouse with ubiquitous GFP expression. The GFP nude mouse was obtained by crossing nontransgenic nude mice with the transgenic C57/B6 mouse in which the beta-actin promoter drives GFP expression in essentially all tissues. In crosses between nu/nu GFP male mice and nu/+ GFP female mice, the embryos fluoresced green. Approximately 50% of the offspring of these mice were GFP nude mice. Newborn mice and adult mice fluoresced very bright green and could be detected with a simple blue-light-emitting diode

Fig. 6.6  Interactions of host stromal GFP-expressing fibroblast cell (short arrow) and Dunning RFPexpressing rodent prostate cancer cells (long arrows)in live tumor tissue. Scale bar, 20 mm [56]

140

R.M. Hoffman

flashlight with a central peak of 470 nm and a bypass emission filter. In the adult mice, all the organs, including the heart, lungs, spleen, pancreas, esophagus, stomach, and duodenum, brightly expressed GFP. The following systems were dissected out and shown to have brilliant GFP fluorescence: the entire digestive system from tongue to anus; the male and female reproductive systems; brain and spinal cord; and the circulatory system, including the heart and major arteries and veins. The skinned skeleton highly expressed GFP. Pancreatic islets showed GFP fluorescence. The spleen cells were also GFP positive. RFP-expressing human cancer cell lines, including PC-3-RFP prostate cancer, HCT-116-RFP colon cancer, MDA-MB-435-RFP breast cancer, and HT1080-RFP fibrosarcoma, were transplanted to the transgenic GFP nude mice. All of these human tumors grew extensively in the transgenic GFP nude mouse. Dual-color fluorescence imaging enabled visualization of human tumor–host interaction by whole-body imaging and at the cellular level in fresh and frozen tissues (Fig. 6.7) [57]. 6.6.1.2 Transgenic RFP Nude Mouse The RFP nude mouse was obtained by crossing nontransgenic nude mice with the transgenic C57/B6 mouse in which the beta-actin promoter drives RFP (DsRed2) expression in essentially all tissues. In crosses between nu/nu RFP male mice and nu/+ RFP female mice, the embryos fluoresced red. Approximately 50% of the offspring of these mice were RFP nude mice. In the RFP nude mouse, all the organs, including the heart, lungs, spleen, pancreas, esophagus, stomach, duodenum,

Fig. 6.7  GFP expression in the tissues and cells of the transgenic GFP nude mouse. (a) Mice fluoresce brilliant, bright green under blue light excitation. The fluorescence could be detected with a simple blue light-emitting diode flashlight with a central peak of 470 nm and a bypass emission filter. (b) The dissected circulatory system, including the heart and major arteries and veins, had brilliant green fluorescence on blue light excitation. (c) The entire skeleton was dissected and could be seen to fluoresce brilliant green on blue light excitation. (d) Spleen cells also could be seen to fluoresce brilliant green on blue light excitation [57]

6  Imaging the steps of metastasis at the macro and cellular level

141

the male and female reproductive systems, brain and spinal cord, and the circulatory system, including the heart, and major arteries and veins, brightly expressed RFP. The skinned skeleton highly expressed RFP. The bone marrow and spleen cells were also RFP positive. GFP-expressing human cancer cell lines, including HCT116-GFP colon cancer and MDA-MB-435-GFP breast cancer, were orthotopically transplanted to the transgenic RFP nude mice. These human tumors grew extensively in the transgenic RFP nude mouse. Dual-color fluorescence imaging enabled visualization of human tumor–host interaction (Fig. 6.8) [58]. 6.6.1.3 Transgenic CFP Nude Mouse The CFP nude mouse was developed by crossing nontransgenic nude mice with the transgenic CK/ECFP mouse in which the beta-actin promoter drives expression of CFP in almost all tissues. In crosses between nu/nu CFP male mice and nu/+ CFP female mice, approximately 50% of the embryos fluoresced blue. In the CFP nude mice, of all internal organs, the pancreas and reproductive organs displayed the strongest fluorescent signals, which varied in intensity. Orthotopic implantation of XPA-1 human pancreatic cancer cells expressing RFP, or GFP in the nucleus and RFP in the cytoplasm, was performed in female nude CFP mice. Color-coded fluorescence imaging of these human pancreatic cancer cells

Fig. 6.8  Transgenic RFP nude mouse. (a) Whole-body image of transgenic RFP nude mouse. (b) Digestive tract of RFP nude mice. (c) Whole skeleton of RFP nude mouse. (d) Bone marrow cells of the RFP nude mouse. (e) Splenocytes of the RFP nude mouse [58]

142

R.M. Hoffman

implanted into the bright blue fluorescent pancreas of the CFP nude mouse afforded novel insight into the interaction of the pancreatic tumor and the normal pancreas, in particular the strong desmoplastic reaction of the tumor. The naturally enhanced blue fluorescence of the pancreas in the CFP mouse serves as an ideal background for color-coded imaging of the interaction of implanted cancer cells and the host (Fig. 6.9) [59].

Fig. 6.9  Imaging the orthotopic human pancreatic cancer CFP nude mouse model with XPA-1RFP cells. Orthotopic injection of human pancreatic cancer cells was performed in CFP nude mice; these images were obtained at week 6 post-implantation. Exposure time for CFP signals was 1 s. (a) In this series, whole-body dual-color imaging permits identification of the tumor in the whole animal. This brightly red fluorescent tumor had grown extensively into the blue pancreas. This could be seen after isolation of the GI tract, when only small slivers of blue pancreas could be seen around the red tumor. (b) The dual-colored XPA-1-GFP-RFP tumor was grown from human pancreatic cancer cells engineered to express GFP in nucleus and RFP in the cytoplasm. Wholebody tri-color imaging permitted identification of the tumor within the mouse. After isolation of the GI tract, a dual-colored tumor on a blue-fluorescent pancreas can been seen [59]

6  Imaging the steps of metastasis at the macro and cellular level

143

6.6.2 Color-Coded Imaging in the Tumor Microenvironment We have developed a simple yet powerful technique for delineating the morphologic events of tumor-induced angiogenesis and other tumor-induced host processes with dual-color fluorescence. The method clearly images implanted tumors and adjacent stroma, distinguishing unambiguously the host and tumorspecific components of the malignancy. The dual-color fluorescence imaging is effected by using RFP-expressing tumors growing in GFP-expressing transgenic mice. This model shows with great clarity the details of the tumor–stroma interaction, especially tumor-induced angiogenesis and tumor-infiltrating lymphocytes. The GFP-expressing tumor vasculature, both nascent and mature, could be readily distinguished interacting with the RFP-expressing tumor cells. GFPexpressing dendritic cells were observed contacting RFP-expressing tumor cells with their dendrites. GFP-expressing macrophages were observed engulfing RFPexpressing cancer cells. GFP lymphocytes were seen surrounding cells of the RFP tumor, which eventually regressed. Dual-color fluorescence imaging visualizes the tumor–host interaction by whole-body imaging and at the cellular level in fresh tissues, dramatically expanding previous studies in fixed and stained preparations [60]. For example, when tumor cells expressing RFP are implanted into mice expressing GFP (as in Fig. 6.10), various types of tumor interactions can be observed [57, 60–62]. In fresh tissue specimens, tumor vessels expressing GFP can be visualized vascularizing tumors expressing RFP in primary and metastatic sites (Fig. 6.11). Dendritic cells expressing GFP can be seen in close contact with tumor cells with their dendritic processes (Fig. 6.12). Stromal fibroblasts expressing GFP can be seen in contact with multiple cancer cells through their pseudopodia. Lymphocytes, expressing GFP, can be observed in the process of rejecting tumor cells from growing in immunocompetent mice. Macrophages, expressing GFP, can be observed engulfing tumor cells expressing RFP (Fig. 6.13). The color-coded tumor–host models will enable imaging and, therefore, help elucidate to a much greater extent the function of the stromal host cells involved in tumor progression [56].

6.6.3 Noninvasive Color-Coded Imaging of the Tumor Microenvironment To noninvasively image cancer cell/stromal cell interaction in the tumor microenvironment and drug response at the cellular level in live animals in real time, we developed a new imageable three-color animal model. The model consists of GFP-expressing mice transplanted with dual-color cancer cells labeled with GFP in the nucleus and RFP in the cytoplasm. The Olympus IV100 Laser Scanning Microscope, with ultranarrow microscope objectives (“stick objectives”), was used for three-color wholebody imaging of the two-color cancer cells interacting with the GFP-expressing

144

R.M. Hoffman

Fig. 6.10  Whole-body image of orthotopically-growing HCT 116 RFP+ human colon cancer in a nude GFP mouse. Image was acquired in a fluorescence light box with a CCD camera 10 weeks after orthotopic implantation of HCT 116 RFP+ cells. Original magnification, ×1 [56]

Fig. 6.11  Visualization of angiogenesis in live tumor tissue 3 weeks after sc injection of B16F10 RFP+ melanoma cells in a transgenic GFP mouse. Well-developed, host-derived, GFP-expressing blood vessels in an RFP-expressing mouse melanoma. Scale bar, 50 mm [56]

stromal cells. In this model, drug response of both cancer and stromal cells in the intact live animal was also imaged in real time. Various in vivo phenomena of tumor– host interaction and cellular dynamics were imaged, including mitotic and apoptotic tumor cells, stromal cells interacting with the tumor cells, tumor vasculature, and tumor blood flow. This new model system enables the first cellular and subcellular images of unperturbed tumors in the live intact animal (Fig. 6.14) [34].

6  Imaging the steps of metastasis at the macro and cellular level

145

Fig. 6.12  Visualization of the interaction of host dendritic cells and tumor cells in fresh tumor tissue. Many host-derived GFP-expressing dendritic cells directly contacting B16F10 RFP-expressing melanoma cells with their dendrites (arrows) are visualized. Dendritic cell–lymphocyte clusters can be seen in certain regions of the image (arrowhead) 3 weeks after tumor implantation. Scale bar, 50 mm [56]

Fig. 6.13  Visualization of macrophage–tumor interactions. (a) Host GFP+ macrophage (arrowhead) contacting a PC-3 RFP+ human prostate cancer cell (arrow). (b) GFP+ macrophage (arrowhead) engulfing an RFP+ cancer cell (arrow). (c) RFP+ cancer cell (arrow) engulfed by a GFP+ macrophage (arrowhead). (d) RFP+ cancer cell (arrows) digested by a GFP+ macrophage (arrowhead). Scale bars, 20 mm [56]

146

R.M. Hoffman

Fig. 6.14  Whole-body, noninvasive, subcellular imaging of drug response of dual-color mouse mammary cancer cells and GFP stromal cells in the live GFP nude mouse with and without doxorubicin. Dual-color MMT cells were injected in the footpad of GFP transgenic nude mice. (a) Whole-body image of untreated dual-color MMT cells in the footpad of a live GFP mouse. Note the numerous spindle-shaped dual-color MMT cells interdispersed among the GFP host cells. (b) Whole-body image of MMT dual-color cancer cells in a live GFP nude mouse 12 h after treatment with doxorubicin (10 mg/kg). The cancer cells lost their spindle shape and the nuclei appear contracted. (c) Whole-body image of dual-color MMT tumor. Numerous dual-color spindle-shaped MMT cells interacted with GFP-expressing host cells. Well-developed tumor blood vessels and real-time blood flow were visualized by whole-body imaging (arrows). (d) In vivo drug response of dual-color MMT tumor 12 h after iv injection of 10 mg/kg of doxorubicin. All of the visible MMT cells lost their spindle shape. Many of the cancer cells fragmented (arrows). Tumor blood vessels were damaged (dashed black lines) and the number of cancer cells was dramatically reduced 12 h after chemotherapy. Bar, 20 mm [34]

6.7 Imaging the Cell Biology of Metastasis In Vivo 6.7.1 Dynamic Imaging of GFP or RFP-Expressing Cancer Cells in Blood Vessels and Lymphatics Several new strategies now exist for imaging cancer cell interactions with both blood vessels and lymphatics in living animals. Signal attenuation by overlying tissue can be markedly reduced by opening a reversible skin flap in the light path, increasing detection sensitivity [63].

6  Imaging the steps of metastasis at the macro and cellular level

147

6.7.2 Imaging Cancer Cell Trafficking in Blood Vessels The mechanism of cancer cell deformation and migration in narrow vessels is incompletely understood. In order to visualize the cytoplasmic and nuclear dynamics of cancer cells migrating in capillaries, RFP was expressed in the cytoplasm, and GFP, linked to histone H2B, was expressed in the nucleus of the cancer cells. Immediately after the cells were injected in the heart of nude mice, a skin flap on the abdomen was made. With a color CCD camera, we observed highly elongated cancer cells and nuclei in capillaries in the skin flap in living mice (Fig. 6.15). The migration velocities of the cancer cells in the capillaries were measured by capturing images of the dual-color fluorescent cells over time. The cells and nuclei in the capillaries elongated to fit the width of these vessels. The average length of the major axis of the cancer cells in the capillaries increased to approximately four times their normal length. The length of the nuclei in the capillaries increased 1.6 times. Cancer cells in capillaries over 8  mm in diameter could migrate up to 48.3 mm/h. The data suggest that the minimum diameter of capillaries where cancer cells are able to migrate is approximately 8 mm [64]. Dual-color cancer cells were also injected by a vascular route in an abdominal skin flap in nude mice. The mice were imaged with an Olympus OV100 small animal imaging system. We observed the nuclear and cytoplasmic behavior of cancer

Fig. 6.15  Classification of the deformation of HT-1080 dual-color cells in the vessels in the skin. (a) Nondeformed cells are within a microvessel. The cells in the microvessel are round and the nuclei oval. The cells occupy the full diameter of the vessel. (b) The cells and nuclei are elongated to fit a capillary. (c) The cells are arrested at the capillary bifurcation. Because of the difference of the deformability between cytoplasm and nucleus, only the cytoplasm is bifurcated. The nucleus is also deformed but remains intact. (d) Cytoplasmic fragmentation in very thin capillary; bar, 50 mm [64]

148

R.M. Hoffman

Fig. 6.16  Intravascular trafficking of dual-color HT-1080 human fibrosarcoma cells labeled in the nucleus with histone H2B-GFP and in the cytoplasm with RFP. (c) Dual-color HT-1080 cell, trafficking at low velocity, advances between other cells and the vessel wall. The moving cancer cell contacts the other cells (arrow). The cell deforms slightly and continues to move without adhesion. Bar, 100 mm. Right, schematics (c). Images were taken every 1.04 s. Bar, 100 mm. Images were acquired in real time with the Olympus OV100 [16]

cells in real time in blood vessels as they moved by various means or adhered to the vessel surface in the abdominal skin flap (Fig. 6.16). Real-time dual-color imaging showed that during extravasation, cytoplasmic processes of the cancer cells exited the vessels first, with nuclei following along the cytoplasmic projections. Both cytoplasm and nuclei underwent deformation during extravasation. Different cancer cell lines seemed to vary strongly in their ability to extravasate. With the dualcolor cancer cells and the highly sensitive imaging system, the subcellular dynamics of cancer metastasis could be observed in live mice in real time [5].

6.7.3 Imaging Cancer Cell Trafficking in Lymphatic Vessels We have developed real-time imaging of cancer cell trafficking in lymphatic vessels. Cancer cells labeled with both GFP in the nucleus and RFP in the cytoplasm, or with GFP only or RFP only were injected into the inguinal lymph node of nude mice. The labeled cancer cells trafficked through lymphatic vessels where they were imaged via a skin flap in real time at the cellular level until they entered the axillary lymph node. The bright fluorescence of the cancer cells and the real-time microscopic imaging capability of the Olympus OV100 small-animal imaging system enabled visualization of the trafficking cancer cells in the lymphatics. Using this imaging strategy, two different cancer cell lines, one expressing GFP and the other expressing RFP, were simultaneously injected in the inguinal lymph node. Fluorescence imaging readily distinguished the two color-coded cell lines and their different abilities to survive in the lymphatic system. Using this imaging technology, we also investigated the role of pressure on tumor-cell shedding into lymphatic vessels. Pressure was generated by placing 25- and 250-g weights for 10 s on the bottom surface of a tumor-bearing footpad. Tumor cell fragments, single cells, and emboli shed from the footpad tumor were easily distinguished using the labeled cells and OV100 imaging system. Increasing pressure on the tumor increased the numbers of shed cells, fragments, and emboli. Pressure also deformed the shed emboli, increasing their maximum major axis [6].

6  Imaging the steps of metastasis at the macro and cellular level

149

6.7.4 Determining Clonality of Metastasis Using Color-Coded Cancer Cells We used GFP-labeled and RFP-labeled HT-1080 human fibrosarcoma cells to determine clonality by simple fluorescence visualization of metastatic colonies after mixed implantation of the red and green fluorescent cells. Resulting pure red or pure green colonies were scored as clonal, whereas mixed yellow colonies were scored as nonclonal. In a spontaneous metastasis model originating from footpad injection in severe combined immunodeficient (SCID) mice, 95% of the resulting lung colonies were either pure green or pure red, indicating monoclonal origin, whereas 5% were of mixed color, indicating polyclonal origin. In an experimental lung metastasis model established by tail-vein injection in severe combined immunodeficient mice, clonality of lung metastasis was dependent on cell number. With a minimum cell number injected, almost all (96%) colonies were pure red or green and, therefore, monoclonal. When a large number of cells were injected, almost all (87%) colonies were of mixed color and, therefore, heteroclonal. We concluded that spontaneous metastasis may be clonal because they are rare events, thereby supporting the rare-cell clonal origin of metastasis hypothesis. The clonality of the experimental metastasis model depended on the number of input cells. The simple fluorescence method of determining clonality of metastases described here can allow large-scale clonal analysis in numerous types of metastatic models (Fig. 6.17) [65, 66]. Color-coded lung metastases were also visualized by external fluorescence imaging in live animals through skin-flap windows over the chest wall. Lung metastases were observed on the lung surface. SCID mice well tolerated multiple surgical procedures for direct-view imaging via skin-flap windows. Real-time metastatic growth of the color-coded clones in the same lung was externally imaged with resolution and quantification of green, red, or yellow colonies in live animals. The color coding enabled the determination of whether the colonies grew clonally or were seeded as a mixture with one cell type eventually dominating, or whether the colonies grew as a mixture. The simultaneous real-time dual-color imaging of metastatic colonies makes color-coded imaging of clones of cancer cells carrying various forms of genes of interest possible [66].

6.8 Imaging Lateral Gene Transfer Between Cancer Cells 6.8.1 Color-Coded Imaging of Circulating Cancer Cells Hormone-refractory human prostate carcinoma growing orthotopically efficiently delivers viable metastatic cells in the host circulation. This is in contrast to the ectopic tumors of the same lineage, which do not deliver live cells into the circulation. To investigate the malignant potential of viable circulating carcinoma cells,

150

R.M. Hoffman

Fig. 6.17  Dual-color fluorescent images of metastatic colonies on the extracted lung. Most appeared green or red and, therefore, clonal. Numerous small metastases could be detected on the lung under fluorescence microscopy. (a) Low magnification view. in b1. b2 is a non-clonal metastatic coclony. (b1, b2) High magnification views. White arrows indicate the border of green and red colonies. Bars, 200 mm [66]

6  Imaging the steps of metastasis at the macro and cellular level

151

we developed a novel dual-color orthotopic co-implantation model of PC-3 human prostate cancer metastasis in nude mice. This model is comprised of coinjection of an equivalent mixture of isolated and cultured GFP-expressing prostate cancer cells isolated from the circulation and parental RFP-expressing human prostate carcinoma cells. In the dual-color model, the selected GFP-labeled viable circulating cells have an increased metastatic propensity relative to the RFPlabeled parental cells [67].

6.8.2 Color-Coded Imaging of Gene Transfer Between Cancer Cells Interacting In Vivo Color-coded imaging visualized circulating yellow fluorescent prostate-cancer metastatic cells that were readily isolated from the circulation of tumor-bearing mice after mixtures of RFP- and GFP-expressing PC-3 human prostate carcinoma cells were implanted in the nude mouse prostate. The yellow fluorescent cells were purified from the circulation of nude mice to 99% homogeneity by FACS, expanded in culture, and re-implanted in the prostate of nude mice. The yellow fluorescent phenotype was heritable and stably maintained in vitro and in vivo by tumor cells for many generations. In the animals implanted with the yellow fluorescing cells, 100% developed aggressive metastatic cancer. Lung metastases were demonstrated in 100% of the animals as early as 4 weeks after injection of the yellow fluorescing cells in the mouse prostate. In contrast, when the GFP- and RFP-expressing parental cells were inoculated into the mouse prostate separately, none of the animals developed lung metastasis. All animals had almost exclusively yellow fluorescent cells in the circulation and bone marrow. These results are consistent with the idea that spontaneous genetic exchange between tumor cells in vivo contributes to genomic instability and creation of highly metastatic cells [68].

6.8.3 Color-Coded Imaging of Gene Transfer from High- to Low-Metastatic Osteosarcoma Cells In Vivo The 143B-GFP cell line with high metastatic potential and the MNNG/HOS-RFP cell line with low metastatic potential, both derived from the TE85 human osteosarcoma cell line, were either co-transplanted or transplanted alone in the tibia in nude mice. Upon mixed transplantation of the two differently-labeled sublines, resulting metastatic colonies were single colored, either red or green, thereby demonstrating their clonality and enabling facile color-coded quantification. When MNNG/HOS-RFP and 143B-GFP were cotransplanted in the tibia, the number of lung metastases of MNNG/HOS-RFP increased eightfold compared to that when MNNG/HOS-RFP was transplanted alone (P < 0.01). In contrast, no enhancement of MNNG/HOS-RFP metastases occurred when MNNG/HOS-RFP and 143B-GFP were transplanted separately in the right and left tibiae, respectively. This result

152

R.M. Hoffman

suggests that the presence of 143B-GFP increased the metastatic potential of MNNG/HOS-RFP within the mixed tumor. We observed transfer of the Ki-ras gene from 143B-GFP to MNNG/HOS-RFP after they were co-implanted, suggesting that the Ki-ras played a role in increasing the metastatic potential of MNNG/HOSRFP in the presence of 143B-GFP [69]. These experiments further suggest the possible role of in vivo gene transfer in enhancing the metastatic potential of cancer cells. The data also further demonstrated the power of color-coded imaging to visualize cancer cell/cancer cell interactions in vivo.

6.9 Methods 6.9.1 Imaging Apparatus 6.9.1.1 In Vivo Imaging with an LED Flashlight and Filters Low-cost instrumentation and standard GFP and RFP biomarkers can be used to visualize tumors completely noninvasively. Utilizing a blue LED flashlight with a 470-nm excitation filter, tumors in – and on – several organs (including liver, pancreas, colon, bone, and brain) could be clearly imaged. The clearest demonstration of the power of this technique are the data showing that when the image of a surgically exposed colon tumor was analyzed and compared to the same tumor from a whole-body (unopened) image, the intensity of the GFP signal from the unopened mouse was 70% that of the opened. Furthermore, despite the expected distortion caused by light scattering, the image sizes were comparable [70]. A blue LED flashlight (LDP LLC, Woodcliff Lake, NJ, USA; www.maxmax. com/OpticalProducts.htm) with an excitation filter (midpoint wavelength peak of 470  nm) and an emission D470/40 filter (Chroma Technology, Brattleboro, VT, USA) for viewing could be used for whole-body imaging of mice with GFP- and RFP-expressing tumors growing in or on internal organs [3, 70]. Figure 6.18 shows whole-body imaging of two tumors, one expressing GFP and the other expressing RFP implanted in the brain. The image shows that the GFP and RFP tumors were simultaneously excited with the blue LED flashlight and readily imaged [70]. 6.9.1.2 Simple Light-Box Imaging Whole-body imaging that visualized the entire animal at lower magnification was carried out in a light box illuminated by blue light fiberoptics (Lightools Research, Inc., Encinitas, CA) and imaged using a thermoelectrically-cooled color CCD camera [3].

6  Imaging the steps of metastasis at the macro and cellular level

153

Fig. 6.18  Whole-body imaging of green fluorescent protein (GFP) and red fluorescent protein (RFP) tumors in the brain in a nude mouse. GFP- and RFP-expressing tumors implanted in the brain in a single nude mouse. The excitation light was produced with a simple blue-LED flashlight equipped with an excitation filter with a central peak of 470 nm. The image was acquired with a Hamamatsu charge-coupled device (CCD) camera [70]

6.9.1.3 In Vivo Imaging with a Fluorescence Dissecting Microscope A Leica fluorescence stereo microscope, model LZ12, equipped with a 50-W mercury lamp, was used for high-magnification imaging of GFP-expressing tumors and metastasis in situ or for whole-body imaging of animals with GFP-expressing tumors. Selective excitation of GFP was produced through a D425/60 band-pass filter and 470 DCXR dichroic mirror. Emitted fluorescence was collected through a long-pass filter GG475 (Chroma Technology, Brattleboro, VT) on a Hamamatsu C5810 3-chip cooled color CCD camera (Hamamatsu Photonics Systems, Bridgewater, NJ). Images were processed for contrast and brightness and analyzed with the use of Image Pro Plus 3.1 software (Media Cybernetics, Silver Springs, MD). Images of 1,024  ×  724 pixels were captured directly on an IBM PC or continuously through video output on a high-resolution Sony VCR model SLVR1000 (Sony Corp., Tokyo, Japan). 6.9.1.4 In Vivo Cellular Imaging with a Variable Magnification Imaging Chamber An Olympus OV100 Small animal imaging system with a sensitive CCD camera and four objective lenses, parcentered and parfocal, enabling imaging from macrocellular to subcellular was developed (Fig. 6.19). Nuclear and cytoplasmic behavior of cancer

154

R.M. Hoffman

Fig. 6.19  (a) Olympus OV100 Small animal imaging system. (b) OV100 covers a wide range of magnifications, from x0.14 (63 × 47 mm imaging area) to x16 (0.6 × 0.45 mm). Four individually optimized objectives, parcentered and parfocal, provide a 106-fold magnification range for seamless imaging of the entire body down to the subcellular level without disturbing the animal. The optics and antireflective coating ensure optimal imaging of multiplexed fluorescent reporters in small animal models. Subcellular structure can be visualized clearly by fluorescence imaging [5]

cells expressing GFP in the nucleus and RFP in the cytoplasm was observed in real time in blood vessels as they moved by various means or adhered to the vessel surface in the abdominal skin flap. During extravasation, real-time dual-color imaging showed that cytoplasmic processes of the cancer cells exited the vessels first, with nuclei following along the cytoplasmic projections. Both cytoplasm and nuclei underwent deformation during extravasation. With the dual-color cancer cells and the highly sensitive imaging system, the subcellular dynamics of cancer metastasis can be observed in live mice in real time [5]. 6.9.1.5 Imaging Chambers Designed for Whole-Body Imaging The FluorVivo small animal imaging system (INDEC Systems, Inc.) can be used for whole-body imaging in live mice. FluorVivo uses extremely bright, solid state, LED illuminators and a full color CCD camera to provide high-speed, multi-color imaging of up to three animals with single exposures. The instrument’s high speed acquisition permits in vivo monitoring of both static and dynamic processes, as well as real-time recordings of fluorescence-guided surgeries. FluorVivo’s fully integrated software provides complete control of the instrument, ease of use, and powerful analytical tools for extracting quantitative data from acquired images [58].

6  Imaging the steps of metastasis at the macro and cellular level

155

The UVP iBox Small Animal Imaging System is capable of fluorescent protein imaging with a range of cameras that use front and back illuminated CCDs with sizes up to a 43-mm diagonal, greatly expanding the applications for high-resolution, large-field-of-view, and increased-throughput imaging. The iBox imaging system can be configured with both monochrome and color CCDs, with CCD resolution currently up to 8.3 megapixels and sensitive to a wide range of spectrum (CFP to near infrared). The range of fast lenses includes several interchangeable, fully automated optics: a 50-mm f1.2, 28-mm f1.8, and a 24–70-mm f2.8 zoom lens. These lenses give maximum imaging flexibility, with the field of view ranging from one to several animals. At f1.2, the typical exposures are less than 50 ms, minimizing the effect of animal movement. The camera, optics, sample platform position, and excitation and emission filters are under full software control, permitting reproducible and rapid imaging with software presets and macros [59].

6.10 Patient-Like Orthotopic Tumor Models 6.10.1 Surgical Orthotopic Implantation A most important development in mouse models of cancer was the availability of nude mice, which are athymic and T-cell deficient, enabling human tumors to grow in these mice [71, 72]. A second important development was the availability of orthotopic mouse models of cancer, which were first described by Miller et  al, 1981; Wang et  al, 1982; Ibrahiem et al, 1983; Naito et al, 1986; McLemore et al, 1987; Bresalier et al, 1987; and Fidler, 1990 [73–79]. These models yield significantly more information of novel compounds, especially on human tumors in orthotopic models than other models of cancer. Our laboratory developed orthotopic models of patient colon cancer in nude mice using SOI of histologically intact patient specimens, a major improvement since it allows higher metastatic rates [80–82]. 6.10.1.1 Ovarian Cancer Tumor fragments (1 mm3) derived from the nude mouse sc CHO-K1-GFP tumors were implanted by SOI on the ovarian serosa in six nude mice. The mice were anesthetized by isofluran inhalation. An incision was made through the left lower abdominal pararectal line and peritoneum. The left ovary was exposed, and part of the serosal membrane was scraped with a forceps. Four 1-mm3 tumor pieces were fixed on the scraped site of the serosal surface with an 8–0 nylon suture (Look, Norwell, MA). The ovary was then returned into the peritoneal cavity, and the abdominal wall and the skin were closed with 6–0 silk sutures. About 4 weeks later, the mice were killed, and the lungs and the other organs removed. All procedures of

156

R.M. Hoffman

the operation described above were performed with a ×7 magnification microscope (Olympus) [82]. 6.10.1.2 Lung Cancer Tumor fragments (1 mm3) derived from the ANIP-GFP or H460-GFP sc tumor growing in the nude mouse were implanted by SOI on the left lung in nude mice. The mice were anesthetized by isofluran inhalation. The animals were put in a position of right lateral decubitus, with the four limbs restrained. A 0.8-cm transverse incision of the skin was made in the left chest wall. Chest muscles were separated by sharp dissection, and costal and intercostal muscles were exposed. A 0.4–0.5-cm intercostal incision between the third and fourth rib on the chest wall was made, and the chest wall was opened. The left lung was taken up by a forceps and tumor fragments were promptly sewn into the upper lung with one 8–0 suture. The lung was then returned into the chest cavity. The incision in the chest wall was closed by a 6–0 surgical suture. The closed condition of the chest wall was examined immediately, and if a leak existed, it was closed by additional sutures. After closing the chest wall, an intrathoracic puncture was made by using a 3-ml syringe and 25 and II gauge needle to withdraw the remaining air in the chest cavity. After the withdrawal of air, a completely inflated lung could be seen through the thin chest wall of the mouse. Then the skin and chest muscle were closed with a 6–0 surgical suture in one layer. All procedures of the operation described here were performed with a ×7 magnification microscope (Olympus) [82]. 6.10.1.3 Prostate Cancer Two tumor fragments (1 mm3) from a high GFP-fluorescent sc tumor from a single animal were implanted by SOI in the dorsolateral lobe of the prostate in five nude mice. After proper exposure of the bladder and prostate following a lower midline abdominal incision, the capsule of the prostate was opened, and the two tumor fragments were inserted into the capsule. The capsule was then closed with an 8–0 surgical suture. The incision in the abdominal wall was closed with a 6–0 surgical suture in one layer. The animals were kept under isoflurane anesthesia during surgery. All procedures of the operation described here were performed with a × 7 magnification microscope (Olympus) [82]. 6.10.1.4 Colon Cancer Colonic Transplantation For transplantation, nude mice were anesthetized, and the abdomen was sterilized with iodine and alcohol swabs. A small midline incision was made and the colorectal part of the intestine was exteriorized. Serosa of the site where tumor pieces were to

6  Imaging the steps of metastasis at the macro and cellular level

157

be implanted was removed. Two to four pieces of 1-mm3 size tumor are implanted on the top of the animal intestine. An 8–0 surgical suture was used to penetrate these small tumor pieces and attach them on the wall of the intestine. The intestine was returned to the abdominal cavity, and the abdominal wall was closed with 7–0 surgical sutures. The animals were kept in a sterile environment. Tumors of all stages and grades can be utilized [82]. Intrahepatic Transplantation An incision was made through the left upper abdominal pararectal line and peritoneum. The left lobe of the liver was carefully exposed, and the liver was cut about 3 mm with scissors. Two to three tumor pieces of 1–2-mm3 size were put on the nude mouse liver and attached immediately with double sutures using 8–0 nylon with an atraumatic needle. After confirmation of no bleeding, the liver was then returned to the peritoneal cavity. The abdomen and skin were then closed with 6–0 black silk sutures [82].

6.11 Technical Details 6.11.1 RFP Retrovirus Production 1.  Insert the HindIII-NotI fragment from pDsRed2, containing full-length RFP cDNA, into the HINDIII-NotI site of pLNCX2, which contains a neomycin resistance gene, to establish the pLNCX2-DsRed2 plasmid. 2. Use PT67, an NIH3T3-derived packaging cell line expressing the 10 A1 viral envelope, to produce retrovirus. Culture 3 × 105 PT67 cells in a 25-mm2 flask with DMEM supplemented with 10% heat-inactivated FBS. It takes approximately 3 days for the cells to reach about 70% confluence. 3. For vector production, use PT67 packaging cells at 70% confluence. Plate PT67 cells on a 60-mm culture dish at 60–80% confluence 12 h before transfection. Use 10 mg of pLNCX2-DSRed2 DNA and the Lipofectamine PLUS transfection kit. Add 7 ml of pLNCX2-DsRed2 DNA to 87 ml serum-free medium in a tube and then add 6  ml Lipofectamine reagent, mix, and incubate for 15  min at 22–26°C (room temperature). 4. Dilute 4 ml of Lipofectamine reagent in 96 ml of serum-free medium in a second tube. Mix and incubate for 15 min at room temperature. 5. Combine the DNA prepared in Step 3 and diluted Lipofectamine reagent, then mix and incubate for 15 min at room temperature. 6. While the complexes are forming, replace the medium on the cells with 800 ml of serum-free DMEM. Add the DNA–Lipofectamine complex to the dish with cells containing fresh DMEM. Mix the complexes into the medium gently and incubate for 4 h at 37°C in 5% CO2.

158

R.M. Hoffman

7. After 4 h of incubation, increase the volume of the medium to 5 ml and incubate for 24 h at 37°C. 8. After 24 h of incubation, clone the packaging cells by limiting dilution in 96-well plates. 9. For selection of a PT67 packaging cell clone producing large amounts of RFP retroviral vector (PT67-DsRed2), culture the cells in the presence of 100– 1,000 mg/ml of G418. Culture the cells for 1–2 days in each concentration of G418. Clones of PT67-DsRed2 cells with high viral titer production are identified with 3T3 cells used for virus titering. Clones with a titer higher than 1  ×  106 plaque-forming units per ml are used for RFP vector production.

6.11.2 GFP Retrovirus Production 1. Use the pLEIN or equivalent retroviral vector expressing enhanced GFP or equivalent GFP and the neomycin resistance gene, on the same bicistronic message, as a GFP expression vector. 2. Use PT67, an NIH3T3-derived packaging cell line expressing the 10 A1 viral envelope, to produce retrovirus. Culture 3 × 105 PT67 cells in a 25-mm2 flask with DMEM supplemented with 10% heat-inactivated FBS. It takes approximately 3 days for the cells to reach ~70% confluence. 3. For vector production, use packaging cells (PT67) (Clontech) at 70% confluence. Plate PT67 cells on a 60-mm dish at 60–80% confluence 12 h before transfection. Use 10 mg pFB-GFP (Clontech) with the Lipofectamine Plus transfection kit. Add 7 ml precomplexed pFB-GFP DNA in 87 ml of serum-free medium and then add 6 ml Lipofectamine reagent in a tube; mix and incubate at room temperature (22–26°C) for 15 min. 4. Dilute 4 ml Lipofectamine in 96 ml serum-free medium in a second tube. Mix and incubate at RT for 15 min. 5. Combine precomplexed DNA and diluted Lipofectamine reagent; then mix and incubate at RT for 15 min. 6. While the complexes are forming, replace the medium on the cells with 800 ml serum-free DMEM. Add the DNA–Lipofectamine reagent complex to the dish with cells containing fresh DMEM. Mix the complexes into the medium gently; incubate in a humidified incubator at 37°C and 5% CO2 for 4 h. 7. After 4 h of incubation, increase the volume of the medium to 5 ml. Incubate in the same conditions for 24 h. 8. After 24 h incubation, clone the packaging cells by limiting dilution in 96-well plates. 9. Examine the cells by fluorescence microscopy 48 h post-transduction. 10. For selection, culture the cells in the presence of 500–2,000 mg/ml G418 to select for a clone producing high amounts of a GFP retroviral vector (PT67-GFP). Culture the cells for 1–2 days in each concentration of G418. High-viral titer production clones of GFP PT67 cells are identified with 3T3 cells used for virus titering. Clones with titer higher than 106 plaque-forming units per ml are used for GFP vector production.

6  Imaging the steps of metastasis at the macro and cellular level

159

6.11.3 RFP or GFP Gene Transduction of Tumor Cell Lines 1. For RFP or GFP gene transduction, use cancer cells that are 20% confluent. Plate cancer cells at a density of 1 × 105 to 2 × 105 cells per 60-mm plate 12–18 h before infection with RFP retrovirus. 2 . For retroviral infection, collect conditioned medium from packaging cells (PT67-DsRed2 or PT67-eGFP) and filter medium through a 0.45-mm polysulfonic filter. Add virus-containing filtered medium to the target cells. Add polybrene to a final concentration of 8 mg/ml. Incubate the cells for 24 h at 37°C. 3. Replace the medium with DMEM and 10% FBS after 24 h of incubation and check for RFP-expressing cells by fluorescence microscopy. 4. Collect tumor cells with trypsin-EDTA and subculture them at a ratio of 1:15 in selective medium, which contains 50 mg/ml G418. 5. To select brightly fluorescent cells, increase the concentration of G418 to 800 mg/ml in a stepwise way. Culture the cells for 1–2 days in each concentration of G418. 6. Isolate clones expressing RFP with cloning cylinders using trypsin-EDTA and amplify them in DMEM in the absence of the selective agent. Then select cells for brightness and stability.

6.11.4 Cell Injection to Establish an Experimental Metastasis Model 1. Collect fluorescent protein-expressing tumor cells by trypsinization for 3 min at 37°C with 0.25% trypsin. 2. Wash the cells three times with cold serum-free medium using a tabletop centrifuge at 500 × g. 3. Resuspend the cells in approximately 0.2 ml serum-free medium. 4. Within 30 min of collecting cells, inject 1 × 106 tumor cells in a total volume of 0.2 ml into 6-week-old C57BL/6 GFP mice or nude (nu/nu) GFP mice in the lateral tail vein, or subcutaneously using a 1-ml 27G2 latex-free syringe. 5. For liver colonization, inject fluorescent protein-expressing cells directly into the portal vein in anesthetized mice.

6.11.5 Surgical Orthotopic Implantation to Establish a Spontaneous Metastasis Model 1. Induce anesthesia with a “ketamine mixture” (10 ml ketamine HCl, 7.6 ml xylazine, and 2.4 ml acepromazine maleate, injected sc). 2. Use a microscope (Leica MZ6) with magnification of about x6 to about x40 for all procedures of the operation.

160

R.M. Hoffman

3. Isolate fluorescent protein-expressing tumor fragments (1  mm3) from subcutaneously growing tumors, formed by injection of RFP-expressing tumor cells by mincing tumor tissue into 1-mm3 fragments. After proper exposure of the target organ, implant three tumor fragments per transgenic GFP mouse. 4. With 8–0 surgical suture, penetrate the tumor fragments and suture the fragments onto the target organ. 5. Keep the mice in a barrier facility under high-efficiency particulate air filtration.

6.11.6 Imaging 6.11.6.1 Fluorescence Microscopy 1. Use an Olympus BH2-RFCA fluorescence microscope equipped with a mercury 100-W lamp power supply or its equivalent. 2. To visualize both GFP and RFP fluorescence at the same time, produce excitation light using a D425/60 band-pass filter and a 470 DCXR dichroic mirror. 3. Collect emitted fluorescence light through a GG475 long-pass filter. 4. Capture high-resolution images of 1,024 × 724 pixels with a Hamamatsu C5810 three-chip cooled color CCD camera or its equivalent, and store directly on an IBM PC or its equivalent. 5. Process images for contrast and brightness using Image-Pro Plus 4.0 software or its equivalent.

6.11.6.2 Fluorescence Stereomicroscopy 1. Use a Leica fluorescence stereomicroscope (model LZ12) equipped with a mercury 50-W lamp power supply or its equivalent. 2. Produce selective excitation of GFP and/or RFP via a D425/60 band-pass filter and 470 DCXR dichroic mirror. 3. Collect emitted fluorescence through a long-pass filter (GG475) on a Hamamatsu C5810 three-chip cooled color CCD camera or its equivalent. 4. Process images for contrast and brightness with using Image-Pro Plus 4.0 software or its equivalent. 5. Capture high-resolution images of 1,024 × 724 pixels directly on an IBM PC or continuously through video output on a high-resolution Sony VCR model SLVR1000 or its equivalent. 6. For C57BL/6 mice, remove hair with Nair or by shaving before images are obtained

6  Imaging the steps of metastasis at the macro and cellular level

161

6.11.7 Chamber Imaging Systems 6.11.7.1 Olympus OV100 1. Perform whole-body or intravital imaging with an Olympus OV100 imaging system using 470-nm excitation light originating from an MT-20 light source. 2. Collect emitted fluorescence through appropriate filters configured on a filter wheel using a DP70 CCD camera. Variable magnification imaging can be done with a series of four objective lenses for macro- or cellular and subcellular imaging in vivo. 3. Capture images on a PC (Fujitsu-Siemens) and process images for contrast and brightness with Paint Shop Pro 8 and cellR.

6.11.7.2 INDEC FluorVivo 1. Perform whole-body imaging with an INDEC FluorVivo imaging system using 470-nm excitation light. 2. Collect emitted fluorescence with the instrument’s full color CCD camera, using the appropriate emission filter. 3. Use the integrated FluorVivo software to adjust acquisition parameters. 4. Capture still and streaming images to the PC’s hard disc using the FluorVivo software. 5. Make required spatial and intensity measurements with the FluorVivo software.

6.11.7.3 UVP iBox 1. Turn on the power of the imaging system and (optional) turn on the warming pad. 2. Place the animal in the imaging system (optional) with their nose in the anesthesia cone. Illuminate the animal without excitation filter first to capture a wide spectrum (white-light) image. Set the F number of the camera lens to over 10 and the exposure time at 200 ms. Use the preview function of the imaging system while adjusting the height of the platform, supporting the animal to obtain a comfortable field of view. Adjust the focus of the lens until the image is clear. Reduce the intensity of illumination if horizontal strips are seen in the white-light image (blooming). Press the capture button when the preview image is satisfactory. 3. Change to the appropriate excitation and emission filter for fluorescence capturing. Adjust the focus with the aperture wide open (smallest number). The camera

162

R.M. Hoffman

exposure time is typically about 1 s. The exposure time can be lengthened to increase the brightness. 4. (Optional) Use the VisionWorks LS software to create an overlay of the whitelight and the fluorescent images.

6.11.8 Tumor Tissue Sampling 1. Obtain tumor tissue biopsies from 3 days to 4 weeks after inoculation of tumor cells. Biopsies of tumor tissue can be obtained from anesthetized mice by removal of a small piece of tumor tissue (1 mm3 or less) with a scalpel. Staunch bleeding by pressing the wound with sterile gauze. Alternatively, the mouse can be killed and the tissue can be collected and processed for analysis. 2. Cut fresh tissue into pieces of about 1  mm3 and gently press onto slides for fluorescence microscopy. This procedure is done manually on normal slides. 3. To analyze tumor angiogenesis, digest the tissues with trypsin-EDTA for 5 min at 37°C before examination. 4. After trypsinization, place the tissues on precleaned microscope slides and cover with another microscope slide.

6.11.9 Measurement of GFP-expressing Tumor Blood Vessel Length and Evaluation of Antiangiogenetic Agents 1. Give mice daily ip injections of doxorubicin (5 mg per gram body weight in a 2-mg/ml solution of 0.9% NaCl) or other drugs or 0.9% NaCl solution (vehicle controls) on days 0, 1, and 2 after implantation of tumor cells. 2. Anesthetize mice with the ketamine mixture and obtain biopsies on days 10, 14, 21, and 28 after implantation (Step 18 provides biopsy sample details). 3. Gently flatten the tumor tissue between the slide and coverslip. 4. Quantify angiogenesis in the tumor tissue by measuring the length of GFPexpressing blood vessels in all fields using fluorescence microscopy. 5. Obtain measurements in all fields at x40 or x100 magnification to calculate the total length of GFP-expressing blood vessels. 6. Calculate the vessel density by dividing the total length of GFP-expressing blood vessels (in mm) by the tumor volume (in mm3).

6.11.10 Immunohistochemical Staining 1. “Snap-freeze” fresh tissue with liquid nitrogen, then orient and embed the frozen tissue in optimum cutting temperature blocks and store at -80°C. Cut the frozen sections to a thickness of 5 mm with a Leica CM1850 cryostat.

6  Imaging the steps of metastasis at the macro and cellular level

163

2. Detect colocalization of GFP fluorescence, CD31, and nestin in the frozen skin sections of mice transgenic for nestin enhancer-driven GFP expression using the anti-rat immunoglobulin and anti-mouse immunoglobulin horseradish peroxidase detection kits following the manufacturer’s instructions. 3. Use monoclonal anti-CD31 (1:50 dilution) and monoclonal anti-nestin (1:80 dilution) as primary antibodies. To identify the GFP-expressing tumor-infiltrating natural killer cells, macrophages, and dendritic cells, detect localization of GFP together with cell surface markers using immunohistochemical staining with monoclonal antibodies to NK1.1, CD111b, and CD11c, respectively. 4. Use staining with substrate-chromogen 3,3¢-diaminobenzidine for antigen detection.

References 1. Tsien RY. Constructing and exploiting the fluorescent protein paintbox (Nobel Lecture). Angew Chem Int Ed Engl. 2009;48:5612–26. 2. Chishima T, Miyagi Y, Wang X, et al. Cancer invasion and micrometastasis visualized in live tissue by green fluorescent protein expression. Cancer Res. 1997;57:2042–7. 3. Yang M, Baranov E, Jiang P, et al. Whole-body optical imaging of green fluorescent proteinexpressing tumors and metastases. Proc Natl Acad Sci U S A. 2000;97:1206–11. 4. Hoffman RM. The multiple uses of fluorescent proteins to visualize cancer in vivo. Nat Rev Cancer. 2005;5:796–806. 5. Yamauchi K, Yang M, Jiang P, et al. Development of real-time subcellular dynamic multicolor imaging of cancer cell-trafficking in live mice with a variable-magnification whole-mouse imaging system. Cancer Res. 2006;66:4208–14. 6. Hayashi K, Jiang P, Yamauchi K, et al. Real-time imaging of tumor-cell shedding and trafficking in lymphatic channels. Cancer Res. 2007;67:8223–8. 7. Shimomura O. Discovery of green fluorescent protein (GFP) (Nobel Lecture). Angew Chem Int Ed Engl. 2009;48:5590–602. 8. Betzig E, Patterson GH, Sougrat R, et al. Imaging intracellular fluorescent proteins at nanometer resolution. Science. 2006;313:1642–5. 9. Chalfie M. GFP: Lighting up life (Nobel Lecture). Angew Chem Int Ed Engl. 2009;48:5603–11. 10. Cheng L, Fu J, Tsukamoto A, Hawley RG. Use of green fluorescent protein variants to monitor gene transfer and expression in mammalian cells. Nat Biotechnol. 1996;14:606–9. 11. Condeelis J, Segall JE. Intravital imaging of cell movement in tumours. Nat Rev Cancer. 2003;3:921–30. 12. Sakaue-Sawano A, Kurokawa H, Morimura T, et al. Visualizing spatiotemporal dynamics of multicellular cell-cycle progression. Cell. 2008;132:487–98. 13. Morin J, Hastings J. Energy transfer in a bioluminescent system. J Cell Physiol. 1971;77: 313–18. 14. Prasher DC, Eckenrode VK, Ward WW, Prendergast FG, Cormier MJ. Primary structure of the Aequorea victoria green-fluorescent protein. Gene. 1992;111:229–33. 15. Yang F, Moss LG, Phillips GN Jr. The molecular structure of green fluorescent protein. Nat Biotechnol. 1996;14:1246–51. 16. Cody CW, Prasher DC, Welstler VM, Prendergast FG, Ward WW. Chemical structure of the hexapeptide chromophore of the Aequorea green fluorescent protein. Biochemistry. 1993;32:1212–8. 17. Heim R, Cubitt AB, Tsien RY. Improved green fluorescence. Nature. 1995;373:663–4. 18. Delagrave S, Hawtin RE, Silva CM, Yang MM, Youvan DC. Red-shifted excitation mutants of the green fluorescent protein. Biotechnology. 1995;13:151–4.

164

R.M. Hoffman

19. Cormack B, Valdivia R, Falkow S. FACS-optimized mutants of the green fluorescent proten (GFP). Gene. 1996;173:33–8. 20. Crameri A, Whitehorn EA, Tate E, Stemmer WPC. Improved green fluorescent protein by molecular evolution using DNA shuffling. Nat Biotechnol. 1996;14:315–9. 21. Zolotukhin S, Potter M, Hauswirth WW, Guy J, Muzycka N. “Humanized” green fluorescent protein cDNA adapted for high-level expression in mammalian cells. J Virol. 1996;70:4646–54. 22. Matz MV, Fradkov AF, Labas YA, et  al. Fluorescent proteins from nonbioluminescent Anthozoa species. Nat Biotechnol. 1999;17:969–73. 23. Shaner NC, Campbell RE, Steinbach PA. et al. Improved monomeric red, orange and yellow fluorescent proteins derived from Discosoma sp. red fluorescent protein. Nat Biotechnol. 2004;22:1567–72. 24. Verkhusha VV, Lukyanov KA. The molecular properties and applications of Anthozoa fluorescent proteins and chromoproteins. Nat Biotechnol. 2004;22:289–96. 25. Chishima T, Miyagi Y, Wang X, et al. Metastatic patterns of lung cancer visualized live and in process by green fluorescent protein expression. Clin Exp Metastasis. 1997;15:547–52. 26. Chishima T, Miyagi Y, Wang X, et al. Visualization of the metastatic process by green fluorescent protein expression. Anticancer Res. 1997;17:2377–84. 27. Chishima T, Miyagi Y, Li L, et al. The use of histoculture and green fluorescent protein to visualize tumor cell host interaction. In Vitro Cell Dev Biol 1997;33:745–7. 28. Yang M, Hasegawa S, Jiang P, et al. Widespread skeletal metastatic potential of human lung cancer revealed by green fluorescent protein expression. Cancer Res. 1998;58:4217–21. 29. Yang M, Jiang P, Sun FX, et al. A fluorescent orthotopic bone metastasis model of human prostate cancer. Cancer Res. 1999;59:781–6. 30. Kaufman RJ, Davies MV, Wasley LC, Michnick D. Improved vectors for stable expression of foreign genes in mammalian cells by use of the untranslated leader sequence from EMC virus. Nucleic Acids Res. 1991;19:4485–90. 31. Astoul P, Colt HG, Wang X, Hoffman RM. A “patient-like” nude mouse model of parietal pleural human lung adenocarcinoma. Anticancer Res. 1994;14:85–92. 32. Wang X, Fu X, Hoffman RM. A new patient-like metastatic model of human lung cancer constructed orthotopically with intact tissue via thoracotomy in immunodeficient mice. Int J Cancer. 1992;51:992–5. 33. Yamamoto N, Jiang P, Yang M, et al. Cellular dynamics visualized in live cells in vitro and in vivo by differential dual-color nuclear-cytoplasmic fluorescent-protein expression. Cancer Res. 2004;64:4251–6. 34. Yang M, Jiang P, Hoffman RM. Whole-body subcellular multicolor imaging of tumor-host interaction and drug response in real time. Cancer Res. 2007;67:5195–200. 35. Jiang P, Yamauchi K, Yang M, et al. Tumor cells genetically labeled with GFP in the nucleus and RFP in the cytoplasm for imaging cellular dynamics. Cell Cycle. 2006;5:1198–1201. 36. Yang M, Jiang P, An Z, Baranov E, Li L, Hasegawa S, et al. Genetically fluorescent melanoma bone and organ metastasis models. Clin Cancer Res. 1999;5:3549–59. 37. Hayashi K, Yamauchi K, Yamamoto N, et al. A color-coded orthotopic nude-mouse treatment model of brain-metastatic paralyzing spinal cord cancer that induces angiogenesis and neurogenesis. Cell Prolif. 2009;42:75–82. 38. Hoffman RM. Comment re: Development of a preclinical model of spontaneous melanoma metastasis. Cancer Res. 2009;69:719. 39. Yang M, Chishima T, Wang X, et al. Multi-organ metastatic capability of Chinese hamster ovary cells revealed by green fluorescent protein (GFP) expression. Clin Exp Metastasis. 1999;17:417–22. 40. Holmgren L, O’Reilly MS, Folkman J. Dormancy of micrometastases: balanced proliferation and apoptosis in the presence of angiogenesis suppression. Nat Med. 1995;1:149–53. 41. Goodison S, Kawai K, Hihara J, et al. Prolonged dormancy and site-specific growth potential of cancer cells spontaneously disseminated from non-metastatic breast tumors as revealed by labeling with green fluorescent protein. Clin Cancer Res. 2003;9:3808–14. 42. Katz MH, Takimoto S, Spivack D, et al. A novel red fluorescent protein orthotopic pancreatic cancer model for the preclinical evaluation of chemotherapeutics. J Surg Res. 2003;113:151–60.

6  Imaging the steps of metastasis at the macro and cellular level

165

43. Yang M, Jiang P, Yamamoto N, et  al. Real-time whole-body imaging of an orthotopic metastatic prostate cancer model expressing red fluorescent protein. Prostate. 2005;62:374–9. 44. Rajput A, Dominguez San Martin I, Rose R, et al. Characterization of HCT116 human colon cancer cells in an orthotopic model. J Surg Res. 2008;147:276–81. 45. Peyruchaud O, Winding B, Pécheur I, Serre CM, Delmas P, Clézardin P. Early detection of bone metastases in a murine model using fluorescent human breast cancer cells: application to the use of the bisphosphonate zoledronic acid in the treatment of osteolytic lesions. J Bone Miner Res. 2001;16:2027–34. 46. Hoffman RM. Green fluorescent protein imaging of tumor cells in mice. Lab Anim. 2002;31:34–41. 47. Yang M, Baranov E, Li X-M, et al. Whole-body and intravital optical imaging of angiogenesis in orthotopically implanted tumors. Proc Natl Acad Sci U S A. 2001;98:2616–21. 48. Schmitt CA, Fridman JS, Yang M, et al. Dissecting p53 tumor suppressor functions in vivo. Cancer Cell. 2002;1:289–98. 49. Hoffman RM, Yang M. Whole-body imaging with fluorescent proteins. Nat Protoc 2006;1:1429–38. 50. Hoffman RM. Tumor-targeting amino acid auxotrophic Salmonella typhimurium. Amino Acids. 2009;37:509–21. 51. Ray P, De A, Min JJ, Tsien RY, Gambhir SS. Imaging tri-fusion multimodality reporter gene expression in living subjects. Cancer Res. 2004;64:1323–30. 52. Burgos JS, Rosol M, Moats RA, et al. Time course of bioluminescent signal in orthotopic and heterotopic brain tumors in nude mice. Biotechniques. 2003;34:1184–8. 53. Yang M, Baranov E, Wang J-W, et  al. Direct external imaging of nascent cancer, tumor progression, angiogenesis, and metastasis on internal organs in the fluorescent orthotopic model. Proc Natl Acad Sci U S A. 2002;99:3824–9. 54. Hasegawa S, Yang M, Chishima T, et al.In vivo tumor delivery of the green fluorescent protein gene to report future occurrence of metastasis. Cancer Gene Ther. 2000;7:1336–40. 55. Kishimoto H, Zhao M, Hayashi K, et al. In vivo internal tumor illumination by telomerasedependent adenoviral GFP for precise surgical navigation. Proc Natl Acad Sci U S A. 2009;106:14514–7. 56. Hoffman RM, Yang M. Color-coded fluorescence imaging of tumor-host interactions. Nat Protoc. 2006;1:928–35. 57. Yang M, Reynoso J, Jiang P, et al. Transgenic nude mouse with ubiquitous green fluorescent protein expression as a host for human tumors. Cancer Res. 2004;64:8651–6. 58. Yang M, Reynoso J, Bouvet M, Hoffman RM. A transgenic red fluorescent protein-expressing nude mouse for color-coded imaging of the tumor microenvironment. J Cell Biochem. 2009;106:279–84. 59. 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;107:328–34. 60. Yang M, Li L, Jiang P, et  al. Dual-color fluorescence imaging distinguishes tumor cells from induced host angiogenic vessels and stromal cells. Proc Natl Acad Sci U S A. 2003;100:14259–62. 61. Amoh Y, Li L, Yang M, et al. Hair follicle-derived blood vessels vascularize tumors in skin and are inhibited by doxorubicin. Cancer Res. 2005;65:2337–43. 62. Amoh Y, Yang M, Li L, et al. Nestin-linked green fluorescent protein transgenic nude mouse for imaging human tumor angiogenesis. Cancer Res. 2005;65:5352–7. 63. McElroy M, Bouvet M, Hoffman RM. Color-coded fluorescent mouse models of cancer cell interactions with blood vessels and lymphatics. Methods Enzymol. 2008;445:27–52. 64. Yamauchi K, Yang M, Jiang P, et al. Real-time in vivo dual-color imaging of intracapillary cancer cell and nucleus deformation and migration. Cancer Res 2005;65:4246–52. 65. Yamamoto N, Yang M, Jiang P, et al. Determination of clonality of metastasis by cell-specific color-coded fluorescent-protein imaging. Cancer Res. 2003;63:7785–90. 66. Yamamoto N, Yang M, Jiang P, et  al. Real-time imaging of individual fluorescent protein color-coded metastatic colonies in vivo. Clin Exp Metastasis. 2003;20:633–8.

166

R.M. Hoffman

67. Glinskii AB, Smith BA, Jiang P, et al. Viable circulating metastatic cells produced in orthotopic but not ectopic prostate cancer models. Cancer Res. 2003;63:4239–43. 68. Glinsky GV, Glinskii AB, Berezovskaya O, et al. Dual-color-coded imaging of viable circulating prostate carcinoma cells reveals genetic exchange between tumor cells in vivo, contributing to highly metastatic phenotypes. Cell Cycle. 2006;5:191–7. 69. Tome Y, Tsuchiya H, Hayashi K, et al. In vivo gene transfer between interacting human osteosarcoma cell lines is associated with acquisition of enhanced metastatic potential. J Cell Biochem. 2009;108:362–7. 70. Yang M, Luiken G, Baranov E, Hoffman RM. Facile whole-body imaging of internal fluorescent tumors in mice with an LED flashlight. Biotechniques. 2005;39:170–2. 71. Rygaard J, Povisen CO. Heterotransplantation of a human malignant tumour to “Nude” mice. Acta Pathol Microbiol Scand. 1969;77:758–60. 72. Rygaard J. Immunobiology of the mouse mutant “Nude.” Preliminary investigations. Acta Pathol Microbiol Scand. 1969;77:761–2. 73. Miller FR, Medina D, Heppner GH. Preferential growth of mammary tumours in intact mammary fatpads. Cancer Res. 1981;41:3863–7. 74. Wang WR, Sordat B, Piguet D, Sordat M. Human colon tumors in nude mice: implantation site and expression of the invasive phenotype. In: Sordat B, editor. Immune-deficient animals. Lausanne, Switzerland: Karger; 1982. p. 239–45. 75. Ibrahiem EH, Nigam VN, Brailovsky CA, et al. Orthotopic implantation of primary N-[4-(5-Nitro2-furyl)-2-thiazolyl] formamide-induced bladder cancer in bladder submucosa: an animal model for bladder cancer study. Cancer Res. 1983;43:617–22. 76. Naito S, von Eschenbach AC, Giavazzi R, Fidler IJ. Growth and metastasis of tumor cells isolated from a human renal cell carcinoma implanted into different organs of nude mice. Cancer Res. 1986;46:4109–15. 77. McLemore TL, Liu MC, Blacker PC, et al. Novel intrapulmonary model for orthotopic propagation of human lung cancers in athymic nude mice. Cancer Res. 1987;47:5132–40. 78. Bresalier RS, Raper SE, Hujanen ES, Kim YS. A new animal model for human colon cancer metastasis. Int J Cancer. 1987;39:625–30. 79. Fidler IJ. Critical factors in the biology of human cancer metastasis: twenty-eighth G.H.A. Clowes Memorial Award Lecture. Cancer Res. 1990;50:6130–8. 80. Fu X, Besterman JM, Monosov A, Hoffman RM. Models of human metastatic colon cancer in nude mice orthotopically constructed by using histologically intact patient specimens. Proc Natl Acad Sci U S A. 1991;88:9345–9. 81. Fu X, Guadagni F, Hoffman RM. A metastatic nude-mouse model of human pancreatic cancer constructed orthotopically from histologically intact patient specimens. Proc Natl Acad Sci U S A. 1992;89:5645–9. 82. Hoffman RM. Orthotopic metastatic mouse models for anticancer drug discovery and evaluation: a bridge to the clinic. Invest New Drugs. 1999;17:343–59. 83. Dusich JM, Oei YA, Purchio T, Jenkins DE. In vivo detection of lung colonization and metastasis using luciferase-expressing human A549 lung cells. Proc Am Assoc Cancer Res. 2002;43:1059. 84. Vooijs M, Jonkers J, Lyons S, Berns, A. Noninvasive imaging of spontaneous retinoblastoma pathway-dependent tumors in mice. Cancer Res. 2002;62:1862–7.

Chapter 7

Patient-Derived Tumor Models and Explants Heinz-Herbert Fiebig and Angelika M. Burger

Keywords  Human Xenografts • Patient derived tumor models • Molecular characterization • Drug sensitivity testing • Gene signature

7.1 Introduction 7.1.1 Historical Perspective Since the first report of the successful xenografting of a human tumor into nude mice in 1969, there have been numerous studies conducted throughout the world using the nude moue as a tool to answer a variety of questions regarding the cause, prevention, and therapy of cancer. Thus, the role of immunodeficient animals in oncology has continuously increased and the athymic nude mouse has proven to be an outstanding host for many human solid tumor xenografts [1, 2]. The latter are now extensively used in the development of potential anticancer drugs, new antineoplastic treatment modalities, and studies on tumor biology [3–7]. Moreover, mice with severe combined immunodeficiency (SCID) and non-obese diabetic (NOD)/SCID mice have enlarged the spectrum of possible applications in cancer research. SCID mice have enabled engraftments of human tumors which were previously difficult to explant such as those of the hematopoietic system [8]. Prior to development of immunodeficient mice, syngenic transplantable mouse tumor systems or autochthonous rat tumors were employed as the main – or the only – tools in the development of antitumor agents [5, 7, 10]. Most of the chemotherapeutic agents currently used in the clinic have been developed in these rodent tumor models (reviewed in Chaps. 1–4). The most frequently used murine tumors were the leukemias L1210 and P388, the melanoma B16, and the Lewis lung cancer H.-H. Fiebig (*) Oncotest GmbH Institute for Experimental Oncology, Am Flughafen 12, D-79108, Freiburg, Germany 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_7, © Springer Science+Business Media, LLC 2011

167

168

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

model. Yet the classes of agents found active in the mouse tumor models, were limited and mainly comprised alkylating agents, antimetabolites, and other DNA interacting drugs [5, 11, 12]. Nevertheless, today transplantable syngeneic murine-tumor models remain particularly valuable for studying biological response modifiers or certain agents that need to be evaluated in a syngeneic environment, such as those targeting distant organ metastasis or requiring NK- or T lymphocytes [9]. In the late 1980s and early 1990s, new drug development moved from applying general cytotoxic principles to molecular target-directed treatment strategies [13]. Consequently, there was a need to identify tumor types/individual patient tumors that express the target and could benefit from more selective therapies in phase II/ III clinical trials. Thus, the in vivo models used in preclinical development today are “disease oriented” and target characterized and are either patient- or cell linederived xenografts or spontaneous tumors in specifically bred transgenic mice [14]. However, because of their high running costs and limited availability, transgenic mice are not suitable for large-scale drug testings. Thus, xenografts/human explants have become the gold standard in cancer drug development. Their use is highly recommended by regulatory agencies such as the EMEA (European Agency for the Evaluation of Medicinal Products) in the “note for guidance on the preclinical evaluation of anticancer medicinal products” [15, 16].

7.1.2 The Strength of Human Models Derived from Patient Explants While the spectrum of transplantable and autochthonous murine tumor models is confined to certain entities such as melanoma, colon, breast, bladder, or lung ­carcinomas and genetically engineered mouse models cannot cover all human ­cancers, patient explants and stably growing xenografts derived thereof can be generated from the vast majority of solid tumor types. Table 7.1 shows the broad panel of human tumor types included in the Freiburg xenograft collection. Moreover, the use of fresh surgical patient material or human tumor engraftments growing subcutaneously in nude mice enables chemosensitivity screening procedures in vitro and in vivo, and thus can be used to predict clinical response [11, 16–19]. Human tumor models established in serial passage have demonstrated a particularly high correlation of drug response compared to that in the clinic (Table  7.2). If evaluated in the nude mouse in  vivo, the correct prediction for response (positive predictive value) was 90% [18, 19]. If tested in vitro using the clonogenic assay, the correct prediction was 62%. The prediction for resistance was 97% in  vivo and 92% in  vitro. This allows for preselection of responsive tumor types in follow-up studies [20, 21]. It will be crucial for a new drug to demonstrate a differential selectivity against human tumors compared to the most sensitive normal tissue. In this respect, humantumor xenografts are being considered as the most relevant models because the

7  Patient-Derived Tumor Models and Explants

169

Table 7.1  Patient-derived tumor models established in Freiburg Bladder CNS Leukemias Melanomas Prostate

Breast Colon Cervix uteri Gall bladder Gastric Head and neck Liver Lung Lymphoma Pleura mesothelioma Pancreas Ovarian Renal Sarcoma Testicle Uterus More than 1,700 tumors of cancer patients have been implanted s.c. into nude mice More than 300 human tumors of the following tumor types have been established 150–200 models have been molecularly characterized

Table 7.2  Comparison of drug response in human tumors growing subcutaneously in nude mice and in the patient

Nude mouse Patient Total Remission Remission 19 No remission Remission  2 No remission No remission 57 Remission No remission  2 Eighty comparisons were obtained in 55 tumors. Xenografts predicted correctly for response in 90% (19/21), resistance in 97% (57/59)

patient-derived tumors (a) grow as a solid tumor, (b) develop a stroma, vasculature­, as well as central necrosis, (c) show more or less differentiation, and (d) have clinically relevant response rates [19]. The tumor-xenograft architecture, the cell morphology, and molecular characteristics mirror the original patient cancers Fin most cases (Fig. 7.1). This is in marked contrast to xenografts derived from cell lines, which in general show an undifferentiated histology and as a result are very resistant to most of the standard agents [11] (Fig. 7.2). This is most likely a result of the high selection pressure in vitro during long-term culture resulting in ­aggressive subclones. In this chapter, we report on our experience with the growth of more than 1,700 patient tumors; the correlation of drug response against clinically used agents; the molecular characterization of parameters relevant for tumor development, growth, and dissemination; examples of target directed compounds discovered by our group; and the concept of discovering gene signatures predicting activity of a drug.

7.2 Materials and Methods 7.2.1 Establishment of Human Tumor Xenografts from Patient Explants All animal experiments were performed according to the project license number G-97/30 following German Animal License regulations (Tierschutzgesetz) which closely reflect the recently published UKCCCR guidelines for the welfare of ­animals in experimental neoplasia [22].

170

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

Fig. 7.1  Histological appearance of the colon carcinoma CXF 158 and the large cell lung carcinoma LXFL 529 as original patient tumors depicted in a and b, respectively, and xenografts established thereof as s.c. solid tumors in serial passage shown in c–d. c represents the CXF 158 human tumor xenograft in passage 10, d the LXFL 529 explant in passage 5. Although stromal parts and blood supply are derived from the murine host, architecture and morphology of the tumors resemble closely the original patient specimens (H&E sections ×400)

Fig. 7.2  Histological appearance of cell line-derived human tumor xenografts established by subcutaneous injection of (a) COLO 205 colon-cancer cells obtained from the NCI Central Repository or (b) DU145 prostate cancer cells obtained from the American Type Culture Collection (ATCC). Both human tumor xenografts are undifferentiated and only sparsely vascularized. They have lost morphological elements which would enable to identify them as either colonor prostate cancer histologies (H&E sections ×400)

7  Patient-Derived Tumor Models and Explants

171

7.2.1.1 Animals Outbred athymic nude mice of the NMRI genetic background from our own breeding­ facility or purchased from accredited breeders (Taconic, Charles River, Harlan, Janvier) were used for all studies described herin. The mice were kept in Macrolon™ cages on laminar-flow shelves and in recent years in individually ­ventilated cages from Tecniplast. In the first passage, tumors derived from male cancer patients were implanted into male nude mice, tumors from women into female mice. All therapeutic experiments with the exception of testicular and ­prostate cancers were carried out in female mice. 7.2.1.2 Tumors Over the past 20 years more than 1,700 resected human solid malignancies were implanted subcutaneously into nude mice. Tumor slices of approximately 5 × 5 × 0.5 mm to 1.0 mm in diameter were implanted in the flanks of the animals. Usually 16 fragments of patient material were grafted into four mice in the first passage. For therapeutic experiments, two tumor fragments of 1.5–2 mm in diameter were implanted subcutaneously between the fore and hind flank of each mouse near the arteria mammaria interna. 7.2.1.3 Tumor Growth Measurements Tumor growth was followed biweekly by serial caliper measurements determining two perpendicular diameters. The product of the two diameters was taken as a ­measure for tumor size. Drug treatment efficacy was evaluated as tumor volume inhibition. Tumor volume was calculated according to the formula: lw2/2, where l is the longer and w the perpendicular diameter. Relative tumor values were calculated for each individual tumor according to tumor size on dayX divided by tumor size on day0 at the time of ­randomization multiplied by 100. The median tumor size was taken for evaluation.

7.2.2 Experimental Design of Drug Testing 7.2.2.1 Study Design Drug testing was performed in serial passage when the tumor growth was regular. Drug response in the nude mouse versus that in the patient was compared when xenografts grew in passages between two and ten. In vivo testing of novel agents was performed with human tumor xenografts which had not been propagated beyond ten passages from the frozen master stock. Mice were selected randomly for vehicle control or test groups after 2–6 weeks when tumors were

172

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

palpable and had reached mean diameters greater than 4 mm. Tumors with a yellow color indicating a high amount of fibrous tissue were excluded. Using these criteria, we did not observe spontaneous regressions or stationary growth behavior in the controls. Each test group consisted of five to six animals comprising between six and ten evaluable tumors. Experiments were terminated when tumors reached a diameter of 1.5 cm. 7.2.2.2 Chemotherapy Treatment regimens used in nude mice were adapted to clinically employed schedules with the exception that intermittent treatment was repeated in mice after 2 weeks instead of 3 to 4 weeks in patients. Drugs, dose, schedule, and route of administration are shown in Table 7.3. A dose around the LD10 after 14 days and around the LD20 28 days after initiation of therapy was considered the maximum tolerated dose (MTD) in tumor-bearing nude mice. If available, agents were used as clinical formulation, experimental drugs were either dissolved in normal saline or prepared, e.g. in 10% dimethylsulfoxide/phosphate-buffered saline (DMSO/PBS) and 0.05% Tween 80™. Table 7.3  Growth behavior of patient-derived tumors in nude mice Serial passageb Rapid growtha (%) Tumor origin Total (n) n (%) n Esophagus 10 7 70 6 60 Cervix uteri 10 7 70 6 60 Colorectal 152 88 58 78 51 Corpus uteri 8 4 50 4 50 Melanomas 63 39 62 27 43 Lung, small cell 39 14 36 16 41 Lung, NSC 227 118 52 87 38 Sarcomas, soft tissue 79 36 46 29 37 Ovary 22 7 32 8 36 Head and neck 47 16 34 16 34 Pancreas 39 16 41 13 33 Brain 6 2 33 2 33 Liver 3 1 33 1 33 Neuroblastoma 3 1 33 1 33 Osteosarcoma 13 5 38 4 31 Testicle 48 14 29 12 25 Pleuramesothelioma 36 14 39 9 25 Stomach 68 17 25 17 25 Bladder 44 17 39 10 23 Renal 124 37 30 24 19 Mammary 74 13 18 13 18 Prostate 41 7 17 2 5 Miscellaneous 129 42 33 34 26 Total 1,260 513 41 411 33 a 2 Tumor size (a × b) >60 mm 90 days after s.c. implant b At least three passages

7  Patient-Derived Tumor Models and Explants

173

7.2.2.3 Evaluation Parameters for Tumor Response For comparing tumor responses in nude mice and in patients, the product of the two diameters was taken as a measure of tumor size. Tumors in nude mice were evaluated after maximum tumor regression or after 21–28 days in nonregressing tumors. The effect of treatment was classified in the xenograft system and in the patient as remission (the product of two diameters <50% of initial value), minimal regression (50– 75%), no change (>75–125%), and progression (>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 <1 mg/mL. 5-Fluorouracil (5-FU) was used as positive reference compound at ultra pharmacological doses of 100–1,000 mg/mL. The coefficient of variation in the control group was <50%. The average cloning efficiency increased markedly in recent years and is now 0.72% [20].

7.2.5 Gene Signatures Patient-derived tumor models were used to investigate the hypothesis that correlating drug response to gene expression would identify gene signatures that can predict the drug response of individual tumors to these agents. Between 74 and 28 xenograft tumors were characterized for their sensitivity towards 11 standard cytotoxic anticancer agents and the VEGF inhibitor Bevacizumab. The mean number of tumors treated with any of the various drugs was 54 (range 31–78). Tumors were considered as responsive if the drugs effected a tumor volume inhibition to less than 11–41% of control tumors. The median cut-off for all drugs was a T/C of 25%. Using these criteria, on average one-third of the test tumors were sensitive (responders) and two-thirds were resistant (nonresponders). The tumor xenografts’ gene expression profiles were determined using the Affymetrix HG-U133 plus 2.0 mRNA expression array representing ~38,500 human genes. Predictive gene signatures were found and subsequently validated using the leave-one-out cross-validation (LOOCV) technique and for three drugs in independent testing sets. To identify predictive transcripts, an iterative leave-one-out/intersection process was used as described by us before [31]. Briefly, in a first step all transcripts with

176

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

a standard deviation >0.75 for at least one of the two classification groups (R, NR) were filtered out because they were unlikely to discriminate between the two groups. Gene lists consisting of the 300 most specific transcripts were generated by iteratively removing one tumor at a time from a training set. The final list of transcripts consisted of the intersection of the gene lists. The purpose of this iterative process was to reduce the risk of underestimating the prediction error by reducing the number of false positive transcripts. Welch ANOVA and Fisher’s rank tests were used to describe the rank of every transcript according to the hypergeometric Fisher’s exact test. This calculation was performed by Genespring’s BioScript SG2c-2 (Agilent, Santa Clara, USA). Support vector machines (SVMs) are powerful pattern recognition techniques assigning objects unequivocally to one of two classes. They build robust prediction models even when the dimensionality is high and the number of training samples is small. A Gaussian SVM (radial basis) was used as class prediction algorithm. This type of SVM gave the best prediction results in the LOOCV and, consequently, was used to predict the independent dataset as well [31].

7.3 Results and Discussion 7.3.1 Take Rates and Growth Behavior of Patient Tumor Explants in Nude Mice Of 1,700 patient tumors, the growth behavior of 1,260 different human malignancies growing subcutaneously in nude mice has been characterized in detail and is presented in Table 7.3. Histological examination of the tumors showed that 79% contained viable tumor tissue. However, only 41% of the tumors showed a rapid growth with tumor diameters of more than 8 mm after 90 days (Table 7.3). Thus, only these rapidly growing tumors were suitable for establishing stably growing models and for further investigations. Nonetheless, it must be noted that the growth of human tumors of the previously described category is still much slower than seen in established murine models. Four hundred and eleven (33%) of the 513 “rapid” growing human tumor explants were successfully transferred in serial passage and most of them were cryopreserved in liquid nitrogen with a recovery rate of 90%. Tumors in serial passage can be divided into three categories. The rate was ­highest (40–60%) in tumors of the esophagus, cervix and corpus uteri, colon, small cell lung, and melanomas. Intermediate growth rates between 20 and 40% were observed for carcinoma of the lung (non-small cell), soft tissue sarcoma, and carcinoma of the ovaries, head and neck, pancreas, testicle, pleuramesothelioma, stomach, and bladder. The lowest rates (5–19%) were observed in renal cancers and in the hormone dependent mammary and prostate carcinomas. The growth of breast and prostate cancers was increased when intramuscular estrogen or dihydrotestosterone depots were given every 2 weeks.

7  Patient-Derived Tumor Models and Explants

177

Among the 400 models about 200 have been selected for further characterization­, e.g. the chemosensitivity profile and molecular characterization. These models showed a regular growth behavior and typical characteristics in terms of histology and differentiation. Rapid- and slow-growing tumors were included, as well as sensitive and resistant models against standard agents. All the tumors can and have been frozen in liquid nitrogen with a recovery rate of 90%. At the University of California, in a similar effort to establish human tumor models from surgical specimens a total of 323 patient explants were collected [4]. These yielded the development of only 27 (8%) of transplantable models. Although the California group had similar positive results with respect to retaining patient characteristics, their markedly lower success rate (8%) in obtaining stable, transplantable models versus that of our program (33%) might be a result of the processing of the specimens. The California group minced and then injected the tumor suspensions subcutaneously into nude mice without paying attention to the gender of the patient from which the tumors were derived or the age of the mice, but we established our models from tumor fragments and used mice of the same gender as the donor patient. Thus, we assured that the xeno-engrafted tumor specimens were provided with the most optimal environment such as their own extracellular matrix, appropriate hormonal conditions, and lowest probability of graft rejection. The Pediatric Preclinical Testing Program established by the NCI has recently reported that 95 pediatric xenografts were established that are currently used for preclinical drug testing [32]. They are suitable models representing very well the tumor type of origin. It was found that the characteristic expression patterns of the primary tumors were maintained in the corresponding xenografts for the majority of samples. The generated gene expression profiles as well as tissue microarrays for 75 models were used for the development of novel targeted therapies. However, it remains unclear how the xenografts were established and what the success rate was of tumors transplanted versus models established.

7.3.2 Comparison of Drug Response of a Tumor Growing in the Nude Mouse and in the Patient 7.3.2.1 Cytotoxic Agents In 55 tumors, 80 comparisons were performed (Table 7.2). The patients and the xenotransplanted tumors were treated with similar schedules at their respective MTD. In the nude mouse, two treatment cycles were given. In colorectal cancers, 24 comparisons were evaluated using single-agent chemotherapy with 5-FU or a nitrosourea and in three tumors combination chemotherapy. The other tumor types were treated in 40 cases with combination chemotherapy and in 13 cases with single-agent chemotherapy. Mammary cancers were treated initially with tamoxifen and after progression with the combination of cyclophosphamide, methotrexat, and 5-FU or adriamycin and cyclophosphamide. Stomach cancers were treated

178

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

with 5-FU + adriamycin + mitomycin-C and small cell lung cancers (SCLC) with the ­combination of adriamycin + vincristine + cyclophosphamide and after progression with cisplatin and etoposide. The overall results are given in Table 7.2. A total of 21 patients reached a remission. The same result was observed in 19 tumors growing in the nude mouse. A total of 59 patients did not respond to treatment and the same result was found in 57 cases in the nude mouse system. Overall, xenografts gave a correct prediction for resistance in 97% (57/59) and for tumor responsiveness in 90% (19/21) of cases. Combination chemotherapy was more successful than single-agent therapy. Of the 43 combinations studied, 16 (37%) effected a remission in the patient as compared to 5 of 35 (14%) of single-agent chemotherapies. Single-agent therapy was successful in four patients with colorectal cancer and in one with stomach cancer. 7.3.2.2 Targeted Agents With targeted drugs, comparisons of drug response in the same tumor growing in the nude mouse and in the patient are difficult to obtain because the targeted agents Bevacizumab and the EGFR inhibitor Cetuximab are registered and given clinically in combination with cytotoxic agents. We have one example of single-agent therapy in a colon cancer which we established in the nude mouse. The established model CXF 1784 was treated with Cetuximab and Bevacizumab as single agents. Cetuximab effected a partial remission with a T/C of 10% of the control. Bevacizumab was less active with an inhibition to 44% of the controls. The patient had already received the FOLFOX and FOLIRI combinations and as third-line single-agent BV under which he showed a slow progression of the tumor markers CEA and CA 19–9 (Fig. 7.3). In the nude mouse the tumor showed a similar slow progression under Bevacizumab, however, delaying the tumor growth since the tumor volume of untreated controls was twice as large. The clinicians were informed that Cetuximab was very effective in this tumor. The patient was treated with 6 weekly injections under which the tumor markers CEA and CA 19–9 decreased to 35 and 53% (Fig. 7.4). In the example the tumor CXF 1784 responded exactly in the same way in the nude mouse as the patient in the clinic.

7.3.3 Clinically Used Cytotoxic Agents Active in the Freiburg Xenograft Panel 7.3.3.1 Cytotoxic Drugs A second approach to validate the xenograft system as a model for drug development is retrospective chemosensitivity profiling. Thus, xenografts were ­a nalyzed in view of their ability to identify compounds active in the clinic. A detailed analysis of ten standard agents of the older generation of cytotoxic drugs studied in 508 tumors was reported elsewhere [11, 19]. Here, we report on the response ­pattern of taxol, gemcitabine, taxotere, vindesine, and topotecan in clinically­

7  Patient-Derived Tumor Models and Explants

179

Patient

CEA [ng\ml]

500 400 300 200 100 0 0

CXF 1784 SA (Xenograft)

Days

1000

CEA

Avastin therapy

3500

100

10

CA 19-9 [U/ml]

Rel. Tumor Volume [%]

10 20 30 40 50 60 70 80

Avastin T/C 44%

0

5

10

15

20

25

3000 2500 2000 1500 1000 500 0

0

10 20 30 40 50 60 70 80

Days

Days after Randomisation CA 19-9

Vehicle control

Avastin therapy

AVASTIN; 40/20 mg/kg/day iv; Days 1/8, 15

K514

Fig. 7.3  Effect of Bevacizumab (Avastin®) in the colon cancer CXF 1784 growing in the nude mouse and in the patient administered as third line therapy

responsive and resistant tumor types. Criteria for evaluation were the same as in the clinic. Each agent was evaluated in tumor types which are considered to be sensitive (CR + PR rate >20%) and in resistant tumor types (PR rate <10%). Overall, the five standard agents studied here (Table 7.4) induced remissions in 24% (45 of 187), and minor regressions or no change in 13% (25 of 187) of the cases examined. Yet 117 of 187 tumors (63%) progressed (Table 7.5). Thus, the overall response rate is similar to what is observed with monotherapy regimens in the patient. In the subgroup of sensitive tumors, the five compounds effected remissions in 37%. In contrast, only 4% remissions were seen in the resistant tumor entities. A progression was observed in 49% of the sensitive tumor types compared to 84% of the resistant tumor types (Table 7.5).

7.3.3.2 Clinically Used Targeted Agents As an example we investigated the angiogenesis inhibitor Bevacizumab [33], an antibody against human vascular endothelial growth factor (VEGF), in 99

180

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

Patient 500

CEA [ng/ml]

400 300 200 100 0 0

10

20

30

40

50

60

Days CEA

Erbitux therapy

3500 3000

100 Erbitux T/C 10%

10

CA 19-9 [U/ml]

R e l. T u mor V o lu m e [%]

CXF 1784 SA (Xenograft) 1000

2500 2000 1500 1000 500 0

0

5 10 15 20 25 Days after Randomisation

0

10

30

40

50

60

Days CA 19-9

Vehicle control

20

Erbitux therapy

ERBITUX; 30 mg/kg/day ip; Days 1,8,15

K514

Fig. 7.4  Effect of Cetuximab (Erbitux®) in the colon cancer CXF 1784 growing in the nude mouse and in the patient administered as fourth line therapy Table 7.4  Maximum tolerated dosesa of selected anticancer drugs in tumor-bearing nude mice Drug Dose (mg/kg/day) Schedule Route Gemcitabine 300 1,8,15 i.v. Taxol 20 1,8,15 i.v. Taxotere 20 1,8,15 i.v. Topotecan 2.5 1–3 i.v. Vindesine 1.5 1.5 i.v. a Until 7–10 days after last therapy

patient-derived tumor xenografts, growing subcutaneously in nude mice in order to compare the preclinical with the clinical activity pattern. Avastin was applied i.v. at a loading dose of 40 mg/kg/day on day 0, followed by 20 mg/kg/day on day 7 and 14, respectively. Antitumor activity was evaluated at the day of minimum T/C value or after 28 days. A marked inhibition (optimal T/C <20%) occurred in nine tumors (10%), 20 tumors (21%) showed T/C values between 20 and 34%, whereas 64 tumors (69%)

7  Patient-Derived Tumor Models and Explants

181

Table 7.5  Anticancer activity of five drugs in clinically responsivea and resistant tumor types Responsive tumors

Resistant tumors

Drug

Total

CR + PR

MR + NC

P

Total

CR + PR

MR + NC

P

Gemcitabine Taxol Taxotere Topotecan Vindesine

17 13 3 8 72

1 7 2 3 29

5 1 1 1 8

11 5 0 4 35

10 6 6 11 41

0 0 1 0 2

2 0 5 0 2

8 6 0 11 37

Total

113 (100%)

42 16 55 74 3 (4%) 9 (12%) 62 (37%) (14%) (49%) (100%) (84%) a For definition see text; CR complete remission, PR partial regression, MR minor regression, NC no change, P progression, for further details see methods

were resistant to single-agent treatment with Bevacizumab with T/C values >34%. Therefore, Bevacizumab as a single agent showed moderate activity at the stronger cut-off point of T/C <20% and a broader activity at the weaker cut-off of T/C <34%. Antitumor activity (T/C <34%) was found in 4/20 colon, 12/36 NSCLC, 3/11 breast, 2/10 renal, and 7/22 miscellaneous tumor entities mimicking the activity pattern seen in the clinic. Nevertheless, most of the tumors grew progressively and doubled the initial tumor volume within 2–3 weeks. A regression of more than 50% was found only in one non-small cell lung cancer. In the clinic the remission rates observed in different tumor types were always <5% when Bevacizumab was administered as single agent. In conclusion, the preclinical antitumor profil and the degree of activity of Bevacizumab corresponded well to clinical experience with efficacy in colon, NSCL, breast, and renal cancers and also additional tumor entities responded. In the clinic, the real benefit was evident in combination with cytotoxic drugs when the progressionfree survival or the overall survival was significantly prolonged in colon, breast, lung, and renal cancers leading to registration in the United States, Europe, and Japan. 7.3.3.3 Comparison of Freiburg Experience to Other Groups These in vivo chemosensitivity data and the high correlation of drugs response in the same tumor growing in the nude mouse and in the patient demonstrate the strength of the patient-derived tumor xenograft model as a predictor of clinical response. Thus, this patient-derived model system provides a valuable tool to screen novel experimental agents to identify sensitive tumor types prior to entering phase II clinical trials. The sensitivity and predictivity data of patient-derived tumor explants generated with the Freiburg xenograft panel in recent years, however, have not always been mirrored in other programs using xenograft models like that of the U.S. National Cancer Institute [5, 6]. Status reports of latter institution about human tumor xenograft testings have not been favorable and as a result large-scale in vivo screening has been dramatically reduced. In the 1980s, the NCI used three patient-derived xenograft tumor models in nude mice (MX1, LX1, and CX1) which were partly

182

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

able to identify compounds active in the clinic [12]: the mammary tumor MX-1 10 of 21, the lung cancer LX-1 7 of 21, and the colon cancer CX-1 0 of 22. Xenografts established from cell lines of the NCI 60 cell line panel were less successful. They responded well to alkylating agents, but the majority of drugs with other modes of action showed some growth delay only. Tumor remissions were an extremely rare event [11, 12]. An explanation for this phenomenon could be found in the undifferentiated histology of the vast majority of cell line-derived xenografts as shown in Fig. 7.2. Compared to patient-derived models, these tumors lost most differentiation and tissue architecture, vascularization is minimal, and it appears that only one homogeneous cell population persisted. The patient-derived xenografts maintained donor characteristics at the morphological and molecular level (Fig. 7.1) and the clinical response characteristics. Whenever xenografts were established from fresh patient material – such as in our labs or, for example, the Institute for Cancer Research in London where six different tumor types (teratoma, small cell lung, breast, melanoma, colon, and non-small cell lung) were studied for chemotherapeutic response – they were shown to be predictive and valuable for new drug development [7, 16, 34].

7.3.4 Molecular Characterization of Human Tumor Xenografts for Target-Oriented Drug Discovery Recent advances in molecular medicine have deepened our understanding of the pathological basis of oncogenesis. This development has had far reaching consequences for drug development in oncology. While traditional procedures for evaluation of drug efficacy were based on empirical screens measuring cytotoxicity, the current algorithms emphasize additional aspects of malignancy such as angiogenesis, apoptosis, cellular senescence, self-renewal, metastasis, and signal transduction [20, 21, 28, 34, 35]. The rapid characterization of new target proteins and their validation for new drug development would not have been possible without two key technologies that have eased the bottleneck of targeted drug development: (a) whole human genome gene expression profiling [25, 31] and (b) high-throughput tumor tissue microarrays [24–28]. 7.3.4.1 Significance of Gene Signatures for Anticancer Therapy Patient tumors established subcutaneously in serial passage in nude mice were characterized for their sensitivity towards one targeted and ten standard cytotoxic anticancer agents [25, 33]. The latter include the alkylating agents cyclophosphamide, ifosfamide, mitomycin-C, and cisplatin, the antimetabolites 5-FU, and gemcitabine; the topoisomerase II inhibitors adriamycin and etoposide as well as the tubulin binders paclitaxel and docetaxel. The mean number of tumors treated with any of the various drugs was 54 (range 31–78). The tumor xenografts’ gene

7  Patient-Derived Tumor Models and Explants

183

expression profiles were determined using the Affymetrix HG-U133 plus 2.0 mRNA expression array representing ~38,500 human genes. The hypothesis was that the correlation of drug response to gene expression would identify gene signatures that can predict the drug response of individual tumors to these agents. Predictive gene signatures were found and subsequently verified using the LOOCV technique and for three drugs in independent testing sets. Tumors were considered as responsive if the drugs effected a tumor volume inhibition to less than 11–41% of control tumors. The median cut-off for all drugs was a T/C of 25%. Using these criteria, on average one-third of the test tumors were sensitive (responders) and two-thirds were resistant (nonresponders). The bioinformatic analysis yielded predictive gene signatures consisting of 20–129 genes (mean for the ten drugs 87 genes). On average, the response rate for predicted responders (83%) was 2.45-fold higher than that for all test tumors (random testing, 34%). This increase of response rates, following signature-guided testing, was consistent for all agents. Conversely, 94% of the predicted nonresponders (range 84–100%) proved to be nonresponders in the nude mouse [25, 31]. The majority of genes (59%) making up the predictive gene signatures had an unknown function. Known genes were implicated in cell proliferation, apoptosis, DNA repair, cell cycle, metabolism, and transcription. The predictive gene signatures presented here for 12 cytotoxic agents have the potential to substantially increase tumor response rates compared to empirical drug treatment. However, they need to be further validated. In a similar way a predictive gene signature was developed for the VEGF inhibitor Bevacizumab. The antibody was tested in vivo in 72 tumor models including tumor types in which Bevacizumab is registered in the clinic (colon, non-small cell lung, breast, and renal cancers) [36]. Bevacizumab was applied in 3 weekly injections i.v. when the tumors reached ~100 mm³. Using a T/C of <34% as a cut-off for activity 29% (21/72) of all tumors were sensitive, 12/34 NSCLC, 4/18 colon, 3/10 breast, and 2/10 renal cancers. A gene signature consisting of 35 genes was identified from which 21 genes are associated with angiogenesis. The signature was validated using the LOOCV and in the independent testing set. In the LOOCV, 15/47 tumors were predicted to respond, from which in the real testings ten responded. Therefore, the response rate increased from 30% (14/47) to 67%. From the predicted 32 nonresponders 28 (88%) were resistant in real testings. In the independent testing set (25 tumors), the correct prediction was 71% for the predicted responders and 89% for the nonresponders. This gene signature is currently being evaluated in a clinical pilot study. 7.3.4.2 Significance of Tumor Tissue Microarrays Since their introduction in the late 1990s, tissue microarrays have become a wellestablished method for the parallel evaluation of gene and protein expression in hundreds of tissue biopsies [37]. Fluorescent in situ hybridization (FISH) and immunohistochemistry allow a classification of tissues according to gene expression­,

184

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

protein levels, and histology. Moreover, the relationship between gene expression, pathological variables, and clinical outcome data can be studied, which permits the assessment of the target’s relevance for therapy, diagnosis, and prognosis of cancer. Thus, tissue microarrays have proven to be a valuable tool for the study of the human oncoproteome. We have applied tissue microarray technology to our collection of human tumor xenografts [26–28]. For example, we have profiled the expression of five proteins involved in cell migration and/or angiogenesis: VEGF, matrix metalloproteinase 1 (MMP1), protease activated receptor (PAR1), cathepsin B, and b1 integrin and transferring receptor in a panel of over 150 tumors and compared their expression levels to available patient outcome data as reported elsewhere [28, 34]. The basic characterization of the angiogenic proteins included overall occurrence, associations with tumor histology, and prognostic significance. In the case of cathepsin B, a high expression was found in a wide range of tumors. Our tissue microarray results were compared to cathepsin B mRNA data (r = 0.48, p < 0.001, n = 94). The degree of correlation was moderate, which might reflect the inherent posttranslational processing between mRNA and intracellular protein. Cathepsin B overexpression was found in pleural mesotheliomas, urethral carcinomas, cancers of the prostate and lung adenocarcinomas. Overall, we have identified 22 xenografts in our tumor collection with highly elevated cathepsin B levels. These might be employed for preclinical testing of compounds targeting the enzyme or polymer prodrugs such as polymeric prodrug N-(2-hydroxypropyl) methacrylamide (HPMA) copolymer–Gly–Phe–Leu–Gly–doxorubicin conjugate PK1 or HPMA-conjugated geldanamycin that are aimed at utilizing tumor associated proteases for specific release of cytotoxic agents in the tumor cell [38, 39]. 7.3.4.3 Target Prevalence in the Freiburg Tumor Collection The development of anticancer agents is Focused on the discovery of novel targets or pathways which play a critical role in the pathogenesis of tumors. Therefore, we characterized our tumor collection on the genomic and proteomic level with different methods (Table 7.6). Gene expression was determined with the Affymetrix chip HG-U133 plus 2.0 mRNA expression array representing ~38,500 human genes. For protein analyses, immunohistochemistry is done on tissue microarrays or with bead suspension assays using protein lysates. The majority of the known signal transduction proteins (total and activated) were determined with these techniques. Mutation status of selected genes is ongoing. As an example gene expression for the AKT gene is shown in 220 tumor models (Fig. 7.5). At the protein level the data are shown for 150 tumor models (Fig. 7.6) and for the colon panel phospho-(p)AKT and total (t)AKT are presented (Fig. 7.7). The latter allows us to carefully select tumor models that are likely to respond to AKT pathway inhibitors and those that are not and thus can be used as “negative” controls. All of our molecular data are stored in a database, allowing the selection of models overexpressing targets of interest. Furthermore the expression of molecular

7  Patient-Derived Tumor Models and Explants

185

Table 7.6  Patient-derived tumor models available for genomic and proteomic studies DNAGrowing in vivo Protein Gene repository lysate December expression Tissue for mutation 2007–2008 2008 Affy-chip microarray repository analyses Tumor Lung nonsmall cell 24 43 37 25 41 Colon 20 27 21 24 30 Melanoma 16 19 19 14 18 Kidney 11 17 12 7 12 Breast 10 11 11 11 10 Pancreas 8 9 9 10 10 Gastric 6 8 6 4 7 Lung small cell 3 8 5 4 8 Head and neck 5 7 5 4 7 Sarcomas 4 6 6 5 9 Ovary 3 7 4 5 9 Pleuramesothelioma 4 6 4 5 5 Bladder 4 6 6 5 4 Prostate 2 2 2 2 2 Uterus cervix 2 3 2 1 2 Uteruscorpus 0 1 1 1 2 Liver 1 1 1 1 1 Glioblastoma 1 1 1 1 1 GIST tumor 1 0 0 0 1 Testicular 1 1 2 1 2 Anal 1 1 0 1 1 Miscellaneous 3 0 6 5 4 Total >200 130 184 160 136 186 In addition, the gene expression data of 44 cell lines growing in vivo as xenografts are available

targets can be correlated with in  vitro or in  vivo activity of new drugs. This ­bioinformatic approach is based on the COMPARE algorithm developed by the U.S. National Cancer Institute [40]. This allows to determine if there is any statistical interaction (Spearman’s rank coefficient) with the potential molecular target, thus providing information regarding a possible mechanism of action of a drug.

7.3.5 Assessment of In Vivo Efficacy of Anticancer Agents Drug Discovery Using the Freiburg Xenograft System In recent years we have developed a combined in vitro/in vivo testing ­procedure for anticancer drug development [18–20]. Broad in  vitro testings using fresh human tumor xenograft tissues in the clonogenic assay or slow growing ­p ermanent cell lines derived thereof have been employed in our laboratory as

186

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

10

AXF BXF CEXF CNXF CXF GXF HNXF LEXF LIXF LXFA LXFE LXFL LXFS LYXF MAXF MEXF MMXF OVXF PAXF PRXF PXF RXF SXF TXF UXF

mRNA expression

100

Fig. 7.5  Characterization of the Freiburg xenograft panel for AKT gene expression in 220 tumor models. AXF anal cancer, BXF bladder, CEXF cervix uteri, CNXF central nervous system, CXF colon, GXF gastric, HNXF head and neck, LIXF liver, LXF lung cancers adeno, epidermoid, large and small cell, MAXF mammary, MEXF melanomas, OVXF ovarian, PAXF pancreas, PRXF prostate, PXF pleuramesotheliomas, RXF renal, SXF soft tissue sarcomas, TXF testicule, UXF uterus cancers 100

normalized MFI

10

1

0.1

AXF BXF CEXF CNXF CXF GXF HNXF LIXF LXFA LXFE LXFL LXFS MAXF MEXF OVXF PAXF PRXF PXF RXF SKXF SXF TXF UXF

0.01

Fig. 7.6  Protein levels for Phospho-AKT in 150 human tumor xenografts growing s.c. in nude mice

a ­p rescreen preceding the in vivo testings in nude mice in order to identify the most sensitive tumor types and individual tumors, thus reducing the number of ­animal experiments. The clonogenic assay offers the advantage that tumor cells grow in three dimensions, and only tumor stem cell-like cells are capable of proliferating. Furthermore, the tumor source – the patient-derived in  vivo growing xenograft – resembles the tumor as it would grow in a patient. In ­addition,

7  Patient-Derived Tumor Models and Explants

187

10

p-Akt t-Akt

8 6 4 2 0 CXF 1096 CXF 1788 CXF 647 CXF 676 CXF 1133 CXF 742 CXF 1034 CXF 233 CXF 260 CXF 504 CXF 280 CXF 269 CXF 1297 CXF 94LX CXF 1103 CXF 1784 CXF 243 CXF 1729 CXF 609 CXF 1783 CXF 533 CXF 1044 CXF 280 CXF 1086 CXF 158 CXF 1753 CXF 264 CXF 975 CXF 883

normalized mean fluorescence intensity (nMFI)

12

10

p-Akt t-Akt

1

0,1

0,01 CXF 1096 CXF 1788 CXF 647 CXF 676 CXF 1133 CXF 742 CXF 1034 CXF 233 CXF 260 CXF 504 CXF 280 CXF 269 CXF 1297 CXF 94LX CXF 1103 CXF 1784 CXF 243 CXF 1729 CXF 609 CXF 1783 CXF 533 CXF 1044 CXF 280 CXF 1086 CXF 158 CXF 1753 CXF 264 CXF 975 CXF 883

normalized mean fluorescence intensity (nMFI)

100

Fig. 7.7  Akt/PKB detection in patient-derived colon carcinoma xenografts. Native tumor lysates were analyzed for levels of Akt/PKB [total (t) and S473 (p)] by BioPlex® analysis. Values are given as normalized mean fluorescence intensity values. Normalization control is an Oncotest tumor lysate pool, generated from more than 100 lysates

the clonogenic assay has been well validated in the past in terms of response rates of fresh tumor explants versus that of patients in the clinic with a highly accurate prediction found for drug resistance and an intermediate predictivity for drug sensitivity [17, 19]. The first stage of the in vitro/in vivo xenograft model system is the evaluation of new compounds in 6–12 and the second stage in 24–48 human tumor xenografts for their inhibitory effects on colony growth in soft agar. The 3–6 most sensitive xenografts in vitro are used for subsequent in vivo testings of compounds against subcutaneously growing tumors in nude mice. One of the limitations of the clonogenic assay, however, is its labor intensity. Nevertheless, it is very valuable as a secondary screen for testing several hundred compounds per year, and for significantly reducing the number of animal tests.

188

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

7.3.6 Selected Examples of Anticancer Agents in Clinical Trials Which Have Been Discovered in a Target-Oriented Approach For the past 20 years, our group has received compounds from academia, the European Organization for Research and Treatment of Cancer (EORTC), the NCI, and pharmaceutical industry [41, 42]. They belong to various structural classes and act by different mechanisms. An overview of compounds which showed activity in our models and where we contributed to the decision to develop the compounds clinically is shown in Table 7.7. Compounds that have entered clinical trials are methotrexate coupled to human serum albumin (MTX-HSA) as synthesized at the German Cancer Research Center or the telomere targeting agent RHPS4, and the minor groove binder DRH 417. The CDK-inhibitor flavopiridol, the DNA binder Quinocarmycin, the glycoprotein synthesis inhibitor Spicamycin, the HSP90 inhibitors 17-allylamino-geldanamycin (17AAG), and 17-dimethylaminoethylamino-17-demethoxy-geldanamycin (17DMAG) were obtained from the U.S. National Cancer Institute; the tubulin binder rhizoxin, and the bioreductive agent EO9 from the EORTC. From the pharmaceutical industry we obtained the platin derivative lobaplatin, the minor groove binder Trabectedin/ ET743, the recombinant human monoclonal antibody 2C4 (Omnitarg, pertuzumab), the telomerase inhibitor sodium metaarsenite/KML001, and various other compounds from companies. The minor groove binder Trabectedin was approved for the treatment of soft tissue sarcoma and lobaplatin for the treatment of SCLC.

Table 7.7  Anticancer drugs discovered as active in the Freiburg human tumor xenograft panel Mode of action References From academia

From the US-NCI

From the EORTC From companies

MTX–has RHPS4 DRH 417 Flavopiridol Quinocarmycin Spicamycin 17-AAG/tanespimycin 17-DMAG Rhizoxin EO9 Lobaplatin ET743/Trabectedin 2C4/pertuzumab KML001 Many discrete compounds

Antimetabolite, EPR effect Telomere targeting agent Minor groove binder Cyclin dependent kinase inhibition DNA binding Glycoprotein synthesis Heatshock protein 90 modulation Heatshock protein 90 modulation Tubulin binder Bioreductive alkylating DNA cross-linking and adducts DNA minor groove alkylation Anit-HER2 antibody Telomerase inhibitor/targets telomeres

[49–52] [53, 54] [55] [56] [57] [58] [43–46] [47, 48] [62] [63] [64] [59–61] [65] [66]

7  Patient-Derived Tumor Models and Explants

189

7.3.7 Possible Future Impact of Patient-Derived Tumor Xenografts and Explants Inasmuch as new drug development is already focused on target-oriented and tumor cell specific approaches based on currently available knowledge, there is still an ongoing demand for a better understanding of tumor biology. Thus, basic research is directed toward genetic tumor profiling and the finding of novel more promising, tumor type specific therapeutic agents, or even individualizing cancer therapy. In accomplishing this, DNA array technology now provides a large amount of possible new gene candidates for therapeutic intervention, but the validation of these genes in a wide range of human tumor specimens seems still to prove a bottleneck. In particular, the lack of sufficient tumor tissue material to perform and thoroughly repeat a set of confirmatory analyses is often problematic. Here, our xenograft collection not only provides a rich source of the human genome, but also provides material of persistent quality. Tissues can be collected freshly and immediately flash-frozen, which results in extremely high total RNA, mRNA, protein, and phosphor protein quality. Moreover, the optimal tissue processing procedure can be evaluated, and later on, novel treatments – either small molecules or antibodies – can be tested in the same tissue from which a target was isolated and validated. In addition, pre- and posttreatment specimens for a pharmaco­ genomic approach can easily be collected. New biomarkers or gene signatures predicting the activity of new agents are continuously being developed. In view of these possibilities, patient-derived tumor xenografts and explants will retain their importance in preclinical cancer research and will continue to increase.

7.4 Summary In Freiburg, we have implanted more than 1,700 human tumors during the past 20 years and experimental models haven been developed for all major solid human tumor types by engrafting patient tumors into immunodeficient mice. The percentage of tumors established in serial passage was highest (40–60%) for cancers of the esophagus, cervix and corpus uteri, colon, SCLCs, and melanomas. Take rates between 20 and 39% were found for cancers of the lung (NSCLC), ovary, head and neck, pancreas, testicle, stomach and bladder, soft tissue sarcomas, and pleuramesothelimas. Only 5–19% of cancers of the kidney, breast, and prostate, however, could be transferred into serial passage. Two hundred tumors were characterized in detail for the sensitivity against standard agents as well as expression of oncogenes, suppressor genes, growth factors and their receptors, parameters of angiogenesis, invasion, and metastasis as well as resistance associated proteins. The gene expression profiles of 200 models were determined using the Affymetric HG-U 133 plus 2.0 expression array representing 38,500 genes. The response of human tumor xenografts to standard agents in the nude mouse in comparison to patient tumors was very similar, identical results were found in 90% (19/21) of sensitive and 97% (57/59) of resistant tumors. This highly accurate predictivity

190

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

validates the xenograft system for drug development. Moreover, since xenografts retain most of the molecular and pathophysiological patient-tumor characteristics, they are suitable tools to study target-oriented therapeutic approaches. We have used these models for both biological and therapeutic investigations. Thus, initial testings are performed in vitro using the clonogenic assay and xenograft as donor material for fresh tissue specimens. Novel compounds are initially tested in 6–12 and up to a total of 24–48 human tumor xenografts. The in  vitro testings further allow the determination of the most sensitive tumors for subsequent in vivo studies, thus reducing the number of in vivo experiments. Compounds in clinical trial that have been found active by our patient-derived xenograft systems include a broad spectrum of agents with various modes of action and very specific targets such as methotrexate coupled to human serum albumin, RHPS4, DRH 417, flavopiridol, quinocarmycin, spicamycin, 17-AAG and 17-DMAG, rhizoxin, and EO9. The minor groove binder Trabectetin was registered in soft tissue sarcoma. Acknowledgments  We are grateful to our colleagues Julia Schüler, Thomas Metz, Veronika Jung, Elke Tetling, and Anke Behnke for their important contributions to this project. We thank Andre Korrat and Hildegard Willmann for software development and data evaluation. The work was supported by grants from the German Ministry for Research and Education BMBF and the National Cancer Institute, Biological Testing Branch.

References 1. Rygaard J, Povlsen CO. Heterotransplantation of a human malignant tumour to the mouse mutant “nude”. Acta Pathol Microbiol Scand. 1969;77:758–60. 2. Povlsen CO, Sordat B, Tamaoki N. Human tumors serially transplanted in nude mice. Report Copenhagen: The Nude Mouse Secretariat, 1977. 3. Houchens DP, Ovejera AA. Proceedings of the symposium on the use of athymic (nude) mice in cancer research. New York: Gustav Fischer Verlag; 1978. 4. Reid LM, Holland J, Jones C, et al. Some of the variables affecting the success of transplantation of human tumors into the athymic nude mouse. In: Houchens DP, Ovejera AA, editors. The use of athymic (nude) mice in cancer research. New York: Gustav Fischer Verlag; 1978. p. 107–22. 5. Ovejera AA. The use of human tumor xenografts in large scale drug screening. In: Kallman RF, editor. Rodent tumor models in experimental cancer therapy. Oxford: Pergamon Press; 1987, p. 218–20. 6. Venditti JM, Weseley RA, Plowman J. Current NCI preclinical antitumor screening in vivo: results of tumor panel screening, 1976–1982, and future directions. In: Garattini S, Goldin A, Hawking F, editors. Advances in pharmacology and chemotherapy. Vol 20. New York: Academic Press; 1984. p. 2–20. 7. Staquet MJ, Byar DP, Green SB, et al. Clinical predictivity of transplantable tumor systems in the selection of new drugs for solid tumors. Rational of a three-stage strategy. Cancer Treat Rep. 1983;67:753–65. 8. Fichtner I, Goan S, Becker M, et al. Transplantation of human haematopoietic of leukaemic cells into SCID and NOD/SCID mice. In: Fiebig HH, Burger AM, editors. Relevance of tumor models for anticancer drug development. Contrib Oncol, Vol 54. Basel: Karger; 1999. p. 207–17. 9. Bonnet D, Dick JE. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med. 1997;3:730–37.

7  Patient-Derived Tumor Models and Explants

191

10. Bibby MC. Transplantable tumours in mice – the way forward. In: Fiebig HH, Burger AM, editors. Relevance of tumor models for anticancer drug development. Contrib Oncol, Vol 54. Basel: Karger; 1999. p. 1–13. 11. Fiebig HH, Dengler WA, Roth T. Human tumor xenografts: predictivity, characterization and discovery of new anticancer agents. In: Fiebig HH, Burger AM, editors. Relevance of tumor models for anticancer drug development. Contrib Oncol, Vol 54. Basel: Karger; 1999. p. 29–50. 12. Plowman J, Dykes DJ, Hollingshead M, et al. Human tumor xenograft models in NCI drug development. In: Teicher BA, editor. Anticancer drug development guide. Totawa: Humana Press; 1997. p. 101–25. 13. Sausville EA, Feigal E. Evolving approaches to cancer drug discovery and development at the National Cancer Institute. Ann Oncol. 1999;10:1287–92. 14. Malakoff D, Vogel G, Marshall E. The rise of the mouse, biomedicine’s model mammal. Science. 2000;288:248–57. 15. Note for guidance on the pre-clinical evaluation of anticancer medicinal products. (http:// www.eudra.org/emea.html). 16. Sausville EA, Burger AM. Use of human tumor xenografts in anti-cancer drug development. Cancer Res. 2006;66:3351–54. 17. Scholz CC, Berger DP, Winterhalter BR, et al. Correlation of drug response in patients and in the clonogenic assay using solid human tumor xenografts. Eur J Cancer. 1990;26:901–5. 18. Fiebig HH, Schmid JR, Bieser W, et al. Colony assay with human tumor xenografts, murine tumors and human bone marrow. Potential for anticancer drug development. Eur J Cancer Clin Oncol. 1987;23:937–48. 19. Fiebig HH, Berger DP, Dengler WA, et  al. Combined in  vitro/in  vivo test procedure with human tumor xenografts. In: Fiebig HH, Berger DP, editors. Immunodeficient mice in oncology. Contrib Oncol Vol 42. Basel: Karger; 1992. p. 321–51. 20. Fiebig HH, Maier A, Burger AM. Clonogenic assay with established human tumor xenografts: Correlation of in vitro to in vivo activity as a basis for anticancer drug discovery. Eur J Cancer. 2004;40:802–20. 21. Chumsri S, Phatak P, Edelman MJ, et al. Cancer stem cells and individualized therapy. Cancer Genomics Proteomics. 2007;4:165–74. 22. Workman P, Twentyman P, Balkwill F, et al. United Kingdom Co-Ordinating Committee on Cancer Research (UKCCCR) guidelines for the welfare of animals in experimental neoplasy (ed 2). Br J Cancer. 1998;77:1–10. 23. Geran RI, Greenberg NH, MacDonald MM, et al. Protocols for screening chemical agents and natural products against tumor and other biological systems. Cancer Chemother Rep. 1972;3:1–103. 24. Baumgarten AJ, Fiebig HH, Burger AM. Molecular analysis of xenograft models of human cancer cachexia – possibilities for therapeutic intervention. Cancer Genomics Proteomics. 2007;4:223–31. 25. Fiebig HH, Schüler J, Bausch N, et al. Gene signatures developed from patient tumor explants grown in nude mice to predict tumor response to 11 cytotoxic drugs. Cancer Genomics Proteomics. 2007;4:197–209. 26. Kuesters S, Maurer M, Burger AM, et  al. Correlation of ErbB2 gene status, mRNA- and ­protein expression in a panel of human tumour xenografts. Onkologie. 2006;29:249–56. 27. Wirth GJ, Schandelmaier K, Smith V, et al. Microarrays of 41 human tumor cell lines for the characterization of new molecular targets: expression patterns of cathepsin B and the transferrin receptor. Oncology. 2006;71:86–94. 28. Smith V, Wirth GJ, Fiebig HH, et al. Tissue microarrays of human tumor xenografts: characterization of proteins involved in migration and angiogenesis for applications in the development of targeted anticancer agents. Cancer Genomics Proteomics. 2008;5:263–73. 29. Von Hoff DD. In vitro predictive testing: the sulfonamide era. Int J Cell Cloning. 1987;5:179–90. 30. Hamburger AW, Salmon SE. Primary bioassay of human tumor stem cells. Science. 1977;197:461–63. 31. Korrat A, Greiner T, Maurer M, et al. Gene signature-based prediction of tumor response to cyclophosphamide. Cancer Genomics Proteomics. 2007;4:187–95.

192

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

32. Whiteford CC, Bilke S, Greer BT, et al. Credentialing preclinical pediatric xenograft models using gene expression and tissue microarray analysis. Cancer Res. 2007;67:32–40. 33. Fiebig HH, Metz T, Greiner T, et al. Antitumor activity of Avastin® in a panel of 100 patient derived tumor models in  vivo in relation to proteomic biomarker profiles. San Francisco: AACR-NCI-EORTC Molecular Targets and Cancer Therapeutics; 2007, p. A28. 34. Steel G. How well do xenografts maintain the therapeutic response characteristics of the source tumor in the donor patient? In: Kallman RF, editor. Rodent tumor models in experimental cancer therapy. Oxford: Pergamon Press; 1987, p. 205–8. 35. Klostermeyer A, Schüler JB, Fiebig HH, et al. Expression patterns of metastasis associated proteins(MMPs, CD44) in a panel of human tumor xenografts. Ann Hematol. 1998;77:220. 36. Fiebig HH, Schüler JB, Greiner T, et al. Determination of a 35 gene signature predictive for the effectiveness of Bevacizumab. San Francisco: AACR-NCI-EORTC Molecular Targets and Cancer Therapeutics, Congress; 2007, p. B10. 37. Kononen J, Bubendorf L, Kallioniemi A, et  al. Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med. 1998;4:844–7. 38. Satchi R, Connors TA, Duncan R. PDEPT: polymer-directed enzyme prodrug therapy. I. HPMA copolymer-cathepsin B and PK1 as a model combination. Br J Cancer. 2001;85:1070–76. 39. Borgman MP, Ray A, Kolhatkar RB, et al. Targetable HPMA copolymer-aminohexylgeldanamycin conjugates for prostate cancer therapy. Pharm Res. 2009;26:1407–18. 40. Paull KD, Shoemaker RH, Hodes L, et  al. Display and analysis of patterns of differential activity of drugs against human tumour cell lines: development of mean graph and COMPARE algorithm. J Natl Cancer Inst. 1989;81:1088–92. 41. Fiebig HH, Berger DP, Winterhalter BR, et  al. In-vitro and in-vivo evaluation of US-NCI compounds in human tumor xenografts. Cancer Treat Rev. 1990;17:109–17. 42. Hendriks HR, Berger DP, Dengler WA, et  al. New anticancer drug development: interim results of the cooperative program between the Freiburg preclinical anticancer drug development group and the EORTC new drug development office. Contrib Oncol. 1996;51:108–14. 43. Stebbins CE, Russo AA, Schneider C, et  al. Crystal structure of an Hsp90-geldanamycin complex: targeting of a protein chaperone by an antitumor agent. Cell. 1997;89:239–50. 44. Schnur RC, Corman ML, Gallachun RJ, et al. erbB-2 oncogene inhibition by geldanamycin derivatives: synthesis, mechanism of action, and structure-activity relationships. J Med Chem. 1995;38:3813–20. 45. Burger AM, Fiebig HH, Newman DJ, et al. Antitumor activity of 17-allylaminogeldanamycin (NSC 330507) in melanoma xenografts is associated with decline in Hsp90 protein expression. Ann Oncol. 1998;9 Suppl 2:132. 46. Burger AM, Fiebig HH, Stinson SF, et al. 17-(allylamino)-17-demethoxy-geldanamycin activity in human melanoma models. Anticancer Drugs. 2004;15:377–87. 47. Smith V, Sausville EA, Camalier RF, et al. Comparison of 17-dimethylaminoethylamino-17demethoxy-geldanamycin (17DMAG) to 17-allylamino-demethoxygeldanamycin (17AAG) in  vitro: effects on Hsp90 and client proteins in melanoma models. Cancer Chemother Pharmacol. 2005;56:126–37. 48. Hollingshead M, Alley M, Burger AM, et  al. In vivo anti-tumor efficacy of 17-DMAG (17-dimethylaminoethylamino-17-demethoxygeldanamycin hydrochloride), a water-soluble geldanamycin derivative. Cancer Chemother Pharmacol. 2005;56:115–25. 49. Stehle G, Sinn H, Wunder A, et al. Plasma protein (albumin) catabolism by the tumor itself – implications for tumor metabolism and the genesis of cachexia. Crit Rev Oncol Hematol. 1997;26:77–100. 50. Matsumara Y, Maeda HA. A new concept for macromolecular therapeutics in cancer ­chemotherapy: mechanism of tumoritropic accumulation of proteins and the antitumor agent smancs. Cancer Res. 1986;46:6387–92. 51. Hartung G, Stehle G, Sinn H, et al. Phase I trial of methotrexate-albumin in a weekly intravenous bolus regimen in cancer patients. Clin Cancer Res. 1999;5:753–59. 52. Burger AM, Hartung G, Stehle G, et al. Preclinical evaluation of a methotrexate-albumineconjugate (MTX-HSA) in human tumor xenografts in vivo. Int J Cancer. 2001;92:718–24.

7  Patient-Derived Tumor Models and Explants

193

53. Cookson JC, Dai F, Smith V, et  al. Pharmacodynamics of the G-quadruplex-stabilizing telomerase inhibitor 3,11-difluoro-6,8,13-trimethyl-8H-quino[4,3,2-kl]acridinium methosulfate (RHPS4) in vitro: activity in human tumor cells correlates with telomere length and can be enhanced, or antagonized, with cytotoxic agents. Mol Pharmacol. 2005;68:15551–58. 54. Phatak P, Cookson JC, Dai F, et al. Telomere uncapping by the G-Qadruplex ligand RHPS4 inhibits clonogenic tumour cell growth in vitro and in vivo consistent with a cancer stem cell targeting mechanism. Br J Cancer. 2007;96:1223–33. 55. Burger AM, Loadman PM, Thurston DE, et al. Preclinical pharmacology of the pyrrolobenzodiazepine (PBD) monomer DRH-417 (NSC 709119). J Chemother. 2007;19:66–78. 56. Drees M, Dengler W, Roth T, et  al. Flavopiridol (L86-8275): selective antitumor activity in vitro and in vivo for prostate carcinoma cells. Clin Cancer Res. 1997;3:273. 57. Fiebig HH, Berger DP, Dengler WA, et al. Cyanocyclin A and the Quinocarmycin analog NSC 607 097 demonstrate selectivity against melanoma xenografts in-vitro and in-vivo. Proc Am Assoc Cancer Res. 1994;2794. 58. Burger AM, Kaur G, Hollingshead M, et al. Antiproliferative activity in vitro and in vivo of the spicamycin analog KRN5500 with altered glycoprotein expression in vitro. Clin Cancer Res. 1997;3:455–63. 59. Hendriks HR, Fiebig HH, Giavazzi R, et al. High antitumor activity of ET743 against human tumour xenografts from melanoma, non-small-cell lung and ovarian cancer. Ann Oncol. 1999;10:1233–40. 60. Grosso F, Sanfilippo R, Virdis E, et al. Trabectedin in myxoid liposarcomas (MLS): a longterm analysis of a single-institution series. Ann Oncol. 2009; 439–44. 61. Fiebig HH, Maier A, Bausch N, et al. Preclinical evaluation of Trabectedin to support the selection of tumor indications for further clinical development. Proc Am Assoc Cancer Res. 2009;2678. 62. Winterhalter BR, Berger DP, Dengler WA, et  al. High antitumors activity of rhizoxin in a combines in-vitro and in-vivo test procedure with human tumor xenografts. Proc Am Assoc Cancer Res. 1993;34:376. 63. Hendriks HR, Pizao PE, Berger DP, et al. E09, a novel bioreductive alkylating indoloquinone with preferential solid tumor activity and lack of bone marrow toxicity in preclinical models. Eur J Cancer. 1993;29:897–906. 64. Klenner T, Voegeli R, Fiebig HH, et al. Antitumor effect of the platinum complex D-19466 (Inn: Lobaplatin) against the transplantable osteosarcoma of the rat and other experiments. J Cancer Res Clin Oncol. 1992;118 Suppl:149. 65. Friess T, Bauer S, Burger AM, et al. In vivo activity of recombinant humanized monoclonal antibody 2C4 in xenografts is independent of tumor type and degree of HER2 overexpression. Eur J Cancer. 2002;38:S149. 66. Phatak P, Dai F, Nandakumar MP, et al. KML001 (sodium meta arsenite) cytotoxic activity is associated with its binding to telomeric sequences and telomere erosion in prostate cancer cells. Clin Cancer Res. 2008;14:4593–4602.

Chapter 8

The Pediatric Preclinical Testing Program Christopher L. Morton and Peter J. Houghton

Abstract  Cancer in children is rare, and tumors differ from more common ­epithelial carcinomas most frequently diagnosed in adults. Current multimodality treatments are curative in approximately 70% of children with cancer. Further, even for patients who ultimately succumb to their disease, initial responses to treatment may be good. Thus, the number of patients eligible for evaluating new agents is limited. Consequently, developing novel drugs, drug combinations, or strategies presents unique problems, as relatively few clinical trials can be conducted. The Pediatric Preclinical Testing Program (PPTP) is an experiment to determine whether panels of childhood cancers, either in vitro or in vivo growing as subcutaneous xenografts in mice, can accurately identify novel agents or combinations of agents that will have significant clinical activity. Here we describe the process of model selection, molecular characterization, in vivo drug evaluation, and summarize results to date. Keywords  Pediatric preclinical testing program • Human tumor xenografts • Brain tumors • Sarcomas • Neuroblastoma • Kidney tumors • Acute lymphoblastoid ­leukemia • Drug screening • Molecular characterization

8.1 Introduction Development of new therapies for children with cancer presents certain challenges unique to this population. For any given pediatric cancer type, the number of patients diagnosed annually is much smaller than for most adult cancers [1,

P.J. Houghton (*) Center for Childhood Cancer, Nationwide Children’s Hospital, 700 Children’s Drive, Columbus, OH 43205, 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_8, © Springer Science+Business Media, LLC 2011

195

196

C.L. Morton and P.J. Houghton

2]. Overall, there are approximately 12,500 cases of cancer diagnosed in patients under 21 years old annually in the United States. For example, rhabdomyosarcoma is the most common sarcoma in children, and fewer than 400 cases are diagnosed each year among children and adolescents less than 20 years of age [3]. Cure rates, overall, are about 70% with current multimodality therapy, and the overall 5-year survival rate for children with cancer is approaching 80% [2]. This combination of low incidence and increasingly effective primary therapy results in relatively few children eligible for evaluating experimental therapies, and those that are available have generally been extensively treated and have highly resistant disease. Consequently, the population of patients available for Phase 1 and -2 clinical testing may not be sufficient and/or ideal for identifying novel therapies that may have significant benefit if used at diagnosis [4]. The challenge is to identify which novel agents (of the 400 plus being developed as cancer therapeutics) should be prioritized for pediatric testing. The magnitude of the problem is greatly amplified when one considers potential drug combinations, let  alone issues of drug scheduling, etc. The pioneering work by Skipper and Schabel, alluded to in the preface of this book, demonstrated that preclinical models of leukemia could serve as a guide to developing effective curative treatment for children with acute lymphoblastic leukemia. Several groups within the pediatric cancer community have systematically tested the validity of preclinical models to identify novel agents that may have clinical activity in childhood cancers [4–15] as an approach to overcoming the limitations in childhood cancer drug development noted above. On the basis of these experiences, the National Cancer Institute (NCI) implemented the Pediatric Preclinical Testing Program (PPTP), comprising a consortium of investigators who integrate in vivo and in vitro testing with expression and genomic profiling. A 2001 meeting organized by the Children’s Oncology Group (COG) and the NCI identified the need for a systematic approach to pediatric preclinical testing to allow the identification of preclinical models that can be used to reliably inform clinical prioritization decisions. Here, we describe the process used for selecting cell lines and xenografts for inclusion into the PPTP, and the current procedures and results from the in vivo screening component.

8.2 Selection of Preclinical Models The objective of the PPTP is to identify agents with significant activity in the panels of pediatric preclinical models as a potential mechanism for prioritizing agents for advancement to clinical trials for children with specific cancers, thereby expediting the discovery of more effective therapies for children with cancer – but how to best accomplish this? Certain syngeneic models or genetically engineered models of pediatric cancer were available, but by far the most comprehensive panel of tumors was represented by patient derived cell lines

8  The Pediatric Preclinical Testing Program

197

and xenografts in mice. The first initiative implementing the recommendations of the COG/NCI meeting was the Pediatric Oncology Preclinical Protein-Tissue Array Project (POPP-TAP), a collaborative effort between the NCI and the Children’s Oncology Group [16]. Xenografts and cell lines of pediatric tumors were solicited for the POPP-TAP project and a total of 75 high-quality xenograft samples representing eight tumor types were collected. Objectives of POPP-TAP included determining the gene expression profiles (cDNA arrays) of these preclinical models and developing xenograft tissue microarrays for protein expression of a panel of pediatric xenografts. In the course of this study, a few xenografts were identified that were not good representations of their primary tumors (i.e. their mRNA profile did not capture the characteristic RNA signature typical for the primary [patient] tumors), and these lines were excluded from the final PPTP tumor panel. Hierarchical clustering showed that the majority of the xenografts grouped according to their specific tumor types. Secondly, to formally validate that the expression profiles of the xenografts reflect those of the corresponding primary tumors, Ewing’s tumor, rhabdomyosarcoma, and neuroblastoma were used to test if the characteristic patterns discriminating different tumor types in primary tumors were preserved in the xenograft models. The artificial neural network trained with profiles of primary tumors accurately diagnosed the xenograft tumors for the majority of xenograft models. To determine whether xenografts or cell lines most closely represented the expression profiles of patient tumors, the Euclidean and Pearson’s distance of expression profiles was used as the most immediate way to measure profile similarity. The average distance of model systems from primary tumors indicates how well the model represents the primary tumor. Comparison of the results for the two model systems of neuroblastoma showed that xenografts were significantly closer to primary tumors than the cell lines were (Fig. 8.1). This latter point is of importance and the majority of models (75%) incorporated into the PPTP screen have been derived by directly engrafting patient tumor into mice and avoiding in vitro culture. These models have more recently been termed “tumorgrafts” [17]. As a result of the POPP-TAP initiative the Rh1 line was reclassified from rhabdomyosarcoma to Ewing’s sarcoma (it was subsequently shown to express the type 1 EWS/FLI transcript and t(11;22) translocation). The second round of characterization used the commercially available Affymetrix HG U133 plus 2 platform and 100K SNP arrays to detect genomic alterations. Profiling led to exclusion of two tumor lines that were of mouse origin and five osteosarcoma lines that did not cluster with human or xenograft osteosarcoma samples. Expression profiles of the remaining 87 models with profiles from 112 clinical samples representing the same histologies showed that model tumors cluster with the appropriate clinical histotype, once “immunosurveillance” genes (contributed by infiltrating immune cells in clinical samples) were eliminated from the analysis [18]. Profiling also led to reclassification of SK-NEP1 (putative anaplastic Wilms tumor) to Ewing sarcoma of the kidney [19].

198

C.L. Morton and P.J. Houghton

Fig. 8.1  (a) Hierarchical clustering of xenografts. All quality clones (N = 38,789) of the preclinical pediatric xenograft models were subjected to hierarchical clustering with average linkage using Pearson’s correlation coefficient as the metric. Distinct colors were used for each tumor type

8  The Pediatric Preclinical Testing Program

199

8.3 Molecular Characterization of Tumor Models The “major” difference between the gene expression of these models and their human counterparts was the signature contributed by stromal cells, which would be absent from cell lines, and contributed by murine elements in xenografts (these signals would not be identified on the HG U133 array). However, of equal importance is whether the models accurately represented the known genetic complexity of the respective childhood cancers. Preliminary data had shown that for several Wilms tumor xenografts similar copy number alterations (CNA) were represented in both patient derived tissue and in the models after heterografting for several passages (serial transfer from mouse to mouse). Although the models were not compared directly with patient tissue from which they were derived, analysis of copy number alterations using the Affymetrix 100K single nucleotide polymorphism GeneChip showed that the models have similar copy number alterations to their “generic” clinical counterparts. Several consistent copy number changes not reported previously were also found (e.g. gain at 22q11.21 that was observed in 5 of 7 glioblastoma samples, loss at 16q22.3 that was observed in 5 of 9 Ewing’s sarcoma and 4 of 12 rhabdomyosarcoma models, and amplification of 21q22.3 that was observed in 5 of 7 osteosarcoma models). Osteosarcoma models demonstrated the highest incidence of CNA, whereas acute lymphoblastoid leukemias and rhabdoid tumor demonstrated the lowest rate of CNAs. Importantly, consistent CNAs were non-random, and distributed differentially across the genome dependent on tumor type (Fig. 8.2). Extensive sequencing of these models is planned; however, one consistent feature of all human studies sequencing cancer genomes to date has been the diversity of genes mutated within individual tumor types. These data challenge the proposed gene-centric models of tumorigenesis [20] and caution one’s interpretation of the translational value of current “simplistic” genetically engineered rodent models.

Fig. 8.1 (continued) to enhance readability. (*) Xenografts that did not cluster with the majority of the same cancer type. Solid vertical lines, two pairs of samples each derived from a common cell line (SK-N-AS: NB-X75 and NB-X107; SMS-KCNR: NB-X75 and NB-X108) but obtained from different laboratories and derived from the s.c. (X75 and X107) or orthotopic (intra-adrenal; X108 and X64) route. (b) Artificial neural network (ANN) average committee votes from a feedforward resilient back-propagation multilayer perceptron ANN with three layers: an input layer of the top ten principal components of the data; a hidden layer with five nodes; and an output layer generating a committee vote for each of the three input classes. (c) clone cutter: sensitivity and specificity with increasing number of the top-ranking clones removed from the training and testing data sets. Quality-filtered clones were ranked by determining the sensitivity of prediction of the training samples with respect to a change in the gene expression level of each clone. Then, after classification using the clone-cutter ANN, the sensitivities (true positives) and specificities (true negatives) of the ANN to predict the xenograft samples were calculated with each successive removal of the top-ranking ANN clones and plotted. From Whiteford et al. [29]

200

C.L. Morton and P.J. Houghton

Fig. 8.2  Correlation of DNA copy and expression across the PPTP samples. (a) Correlation of expression variation and DNA copy variation according to genomic location. SNP markers are ordered across the X axis, whereas expression probe sets are ordered on the Y axis according to their mapped positions. The heat map displays Pearson’s coefficient (R) of the association between copy change and relative expression. Results from the 493 unique transcripts with R 0.6. (b) Distribution of the 493 copy-disrupted transcripts (R 0.6) across the genome. Y axis, frequency of observations after adjustment for chromosome length and number of genes per chromosome. Chromosomes are labeled by number and colored by their apparent grouping. (c) Distribution of the 493 copy-disrupted transcripts according to genome location and frequency of observation in tumor histotypes. The number of transcripts is plotted (Y axis) against chromosome position (X axis). Colors within the bars indicate the tumor histotype with the most frequent number of changes at a given chromosome location. From Neale et al. [16]

8  The Pediatric Preclinical Testing Program

201

The samples, from SNP and expression analysis of the models, were then integrated to determine whether changes in copy number were reflected by coordinate changes in gene expression. This analysis identified 493 copy number–altered genes that are nonrandom and appear to identify histotype-specific programs of genetic alterations. These data, although from relatively small sample sets, imply a genetic program leading to the complex genotypes of many childhood cancers. Again, this is an important consideration when developing genetically engineered models to recapitulate the complex genetic alterations found in childhood cancers. Full details of the molecular characterization of the models used in the PPTP are at http://pptp.stjude.org/demographcs. Genomic and expression files are available at http://pptp.stjude.org/affyData.php.

8.3.1 Model Fidelity On the basis of the molecular characteristics, as well as in vivo growth characteristics of the models, the PPTP in vivo panel constitutes 75% of models derived into mice directly from patient tumor (i.e. no in vitro culture), and 42% of the models were derived from previously treated patients. The methods for initiating primary tumors from patient samples [21] and tumor propagation methods used in the PPTP for both solid tumors and leukemia lines have recently been published [22]. The major concerns regarding tumor fidelity are through genetic drift, contamination with spontaneous lymphoma that occurs in SCID mice, and cross-contamination of tumor lines during transplant. Genetic drift can be minimized if tumors are serially passaged (mouse to mouse) for a limited duration (say 20 serial transfers), before reinitiating the line from cryopreserved tissue. Genetic drift can be monitored by high-resolution SNP analysis. More problematic is the potential overgrowth by murine lymphoma (particularly in NOD/SCID mice) or the potential for cross contamination. Recently a DNA fingerprinting technique to determining the fidelity of tumor lines using PCR has been adopted by the PPTP. DNA fingerprinting provides a simple, precise, reproducible method to characterize cell lines and tumors. This method determines the number of tandem repeat sequences (“variable number of tandem repeats” [VNTR]) in specific loci of the genome. The MCT118 locus (also termed D1S80) was characterized first, since this is a highly polymorphic site. The YNZ22 (also called D17S5) and COL2A1 were also chosen as potentially useful VNTR loci because they are composed of 70-base pair and 31–34-base pair repeats. Analysis of the MCT118 and YNZ22 loci was sufficient to distinguish 19 of the 21 lines. However, two cell lines (JR1 and SKNSH) contained the same number of repeat sequences at both the MCT118 and YNZ22 loci. Analysis of a third locus, COL2A1, readily distinguished these two cell lines from each other. We conclude that analysis of 2–3 VNTR loci is sufficient to characterize tumor and cell lines. This method gives added confidence that inadvertent cross-contamination between cell lines can be prevented.

202

C.L. Morton and P.J. Houghton

8.3.2 Primary Screening Because of the limited capacity to test more than 12 agents per year, the PPTP has focused on agents that are in early stage adult clinical trials, or where an IND has been filed. An overall schema for drug testing is shown in Fig. 8.3. The objective of the initial evaluation (Stage 1) is to define whether an agent has significant antitumor activity. For most of the agents screened the company or academic center submitting the agent has dose and schedule optima defined, and these data are factored into how the drug is administered, although preliminary toxicity testing may be required if existing data were obtained in different mouse strains. In addition, NOD/SCID mice used to propagate the ALL models tend to be more sensitive to many agents, and toxicity testing is done for most drugs prior to full screening. Where there is available drug pharmacokinetic data for human and mouse, a dose level and schedule that give relevant clinical drug exposure in mice are selected. Otherwise, for cytotoxic agents the maximum tolerated dose is used, and for noncytotoxic agents a “biologically effective dose” (BED) is administered. All agents are coded and screened “blinded” to avoid investigator bias, but also to allow screening of agents known to be active against certain childhood cancers (model credentialing). A third reason for blinded screening is that agents can be retested periodically to assess the stability of responses in the models.

Fig. 8.3  Schema showing the two stage testing process used by the PPTP. From Houghton et al. [6]

8  The Pediatric Preclinical Testing Program

203

8.3.3 Response Criteria: Solid Tumors The primary in vivo screen for new agents is against 44 solid tumor models and 8 ALL models [23]. Hence very large data sets are generated. To uniformly analyze experiments, we opted for a simple experimental design (treatment for two cycles of therapy over 6 weeks) and on use of several metrics to define tumor “response” using “pseudo-clinical” criteria. For cytotoxic agents where clinical objective response (i.e. tumor regression) is used for assessing drug activity, regressions in preclinical models would be an important determinant of activity. In contrast, for a cytostatic or antiangiogenic agent, prolonged tumor growth inhibition or disease stabilization (SD) would merit further testing. Response criteria for the solid tumor panels are summarized in Table 8.1. For individual tumors, progressive disease (PD) was divided into PD1 (<50% growth inhibition defined by T/C) and PD2 (>50% growth inhibition, but >25% increase in tumor volume). Other criteria are listed in Table 8.1, but mimic stable disease (SD), partial response (PR), complete response (CR) and maintained CR (MCR), as pseudo-clinical endpoints. Each individual mouse is assigned a score from 0 to 10 based on their response: PD1 = 0, PD2 = 2, SD = 4, PR = 6, CR = 8, and MCR = 10 (Table 8.1). The median scores for control and treatment groups are determined and overall group responses according to this median score are assigned. Studies in which the control group is not categorized as SD or PD are considered inevaluable and are excluded from analysis. Studies in which toxicity is greater than 25% are considered inevaluable and are excluded from analysis. Treatment groups with PR, CR, or MCR are considered to have had an objective response. Agents inducing objective responses are considered highly active against the tested line, while agents inducing stable disease or PD2 are considered to have intermediate activity, and agents producing PD1 are considered to have a low level of activity against the tested line.

Table 8.1  Tumor response definitions for solid tumor studies Response Definition PD1 Progressive <50% regression at all measurements and >25% disease 1 increase in tumor volume at the end of the study period, TGD value of £1.5 PD2 Progressive <50% regression at all measurements and >25% disease 2 increase in tumor volume at the end of the study period, TGD value of >1.5 SD Stable disease <50% regression at all measurements and £25% increase at the end of the study PR Partial response ³50% regression but with tumor volume ³0.1 cm3 during study CR Complete response Tumor volume <0.1 cm3 for at least one study measurement MCR Maintained complete Tumor volume <0.1 cm3 at the end of study response TGD tumor growth delay value

Score 0

2

4 6 8 10

204

C.L. Morton and P.J. Houghton

8.3.4 Response Criteria: Acute Lymphoblastic Leukemia Xenograft Models For the ALL xenograft models, percentages of human CD45 (hCD45) cells in the peripheral blood are measured relative to total mouse CD45 plus hCD45 cells. An event is defined as a tumor reaching 25% or greater of hCD45 cells in the peripheral blood. Event-free survival is defined as the time interval from initiation of the study to the first event or to the end of the study period for animals that do not event. Times to event are determined using interpolation. Response criteria for the ALL panel are summarized in Table 8.2. Responses are coded as progressive disease (PD), stable disease (SD), partial response (PR), or complete response (CR). A tumor is considered as having PD if the percentage of hCD45 cells never dropped below 1% and if the tumor has an event before the end of the study period. A response is classified as SD if the percentage of hCD45 cells never dropped below 1% and no event occurs before the end of the study. PR is assigned if the percentage of cells drops below 1% for any one time point regardless of whether the percentage reached 25%. A CR is assigned if the percentage of hCD45 cells drops below 1% for two consecutive weeks of the study and regardless of whether the percentage reached 25% or not. A CR is considered maintained if the percentage of hCD45 is less than 1% for the last three measurements of the study. For treatment groups, PD is further classified into PD1 and PD2 according to the tumor growth delay (TGD) value. 8.3.4.1 Event-Free Survival

An event in the solid tumor xenograft models is defined as a quadrupling of tumor volume from the initial tumor volume. Event-free survival is defined as the time interval from initiation of study to the first event or to the end of the study period for tumors that do not quadruple in volume. The time to event is determined using interpolation.

Table 8.2  Tumor response definitions for acute lymphoblastic leukemias Response Definition PD1 Progressive Disease 1 CD45% never drops below 1%, events before end of study, TGD* value of £1.5 PD2 Progressive Disease 2 CD45% never drops below 1%, events before end of study, TGD value of >1.5 SD Stable Disease CD45% never drops below 1%, no events before end of study PR Partial Response CD45% drops below 1% for only 1 week CR Complete Response CD45% drops below 1% for 2 consecutive weeks MCR Maintained Complete CD45% drops below 1% for last 3 consecutive Response measurements of the study *TGD = tumor growth delay value

Score 0 2 4 6 8 10

8  The Pediatric Preclinical Testing Program

205

8.3.4.2 Tumor Growth Delay Value

For treatment groups only, a TGD value is used to further classify mice with PD as PD1 or PD2. TGD values are calculated based on the numbers of days to event. For each individual mouse that has PD and has an event in the treatment groups, a TGD value is calculated by dividing the time to event for that mouse by the median time to event in the respective control group. Median times to event are estimated based on the Kaplan– Meier event-free survival distribution. If a mouse had a TGD value £1.5, that mouse is considered PD1. If the TGD value was >1.5, the mouse is considered PD2. Mice that have PD but do not have an event at the end of the study are coded as PD2. 8.3.4.3 Tumor Volume T/C Value A treated/control (T/C) value for tumor volume is calculated for each solid tumor study. Relative tumor volumes (RTV) for control (C) and treatment (T) mice are calculated at day 21 or when all mice in the control and treated groups still have measurable tumor volumes (if less than 21 days). The mean RTV for control and treatment mice for each study are then calculated and the T/C value is the mean RTV for the treatment group divided by the mean RTV for the control group. For the tumor volume T/C response measure, agents producing a T/C of £15% are considered highly active, those with a mean tumor volume T/C of £45% but >15% are considered to have intermediate activity, and those with mean T/C values >45% are considered to have low levels of activity [24]. The median RTV (or median hCD45% for ALL cell lines) at the last day of the study is also calculated. 8.3.4.4 EFS T/C Value The EFS T/C value is defined by the ratio of the median time to event of the treatment group and the median time to event of the respective control group. If no median time to event exists for the control group, then EFS T/C is undefined. If the treatment group does not have a median time to event, then EFS T/C is defined as greater than the ratio of the last day of the study for the treatment group divided by the median time to event for the control group. For the EFS T/C measure, agents are considered highly active if they meet three criteria: (a) an EFS T/C >2; (b) a significant difference in EFS distributions (p£0.050); and (c) a net reduction in median tumor volume for animals in the treated group at the end of treatment as compared to at treatment initiation (i.e. median RTV <1 at the end of study). Agents meeting the first two criteria, but not having a net reduction in median tumor volume for treated animals at the end of the study are considered to have intermediate activity. Agents with an EFS T/C <2 are considered to have low levels of activity. Xenografts in which the median EFS for the control line is greater than one-half of the study period or in which the median EFS for the control line does not exist are considered not evaluable for the EFS T/C measure of activity.

206

C.L. Morton and P.J. Houghton

8.4 Data Presentation Because the screen generates large data sets, we have used two approaches to present these in a “manageable” form. The first is the “Heat Map” (Fig. 8.4) where responses (PD1 ® MCR) are color-coded. While this has limited utility for single agents, it allows visualization of the activity in data sets obtained from agents having diverse mechanisms of action, and characteristics from small molecules to lytic viruses. For example, one can readily compare the activity of agents with specific mechanisms of action (e.g. antiangiogenic drugs, topoisomerase poisons etc.), or identify agents that have broadspectrum activity (cyclophosphamide, a known active agent, and standard component of many solid tumor protocols) or restricted histotype activity (e.g. bortezemib, or ABT263). The data set presented in Fig. 8.4 comprises 28 drugs/biologicals tested against 52 in vivo models (44 solid tumor, 8 ALL). There are 343 missing data points as some tumor models were not available for use during the early testing period, or data were excluded due to excessive toxicity (>25% deaths). For the 1113 data sets there were 163 objective responses for an overall response rate (defined as ³PR) of 14.7%. Excluding results from PR104 (see “false positives” below), the response rate was 12.2% (136/1,113). Notably, the screen identified the known active agents (vincristine, cyclophosphamide, cisplatin, topotecan) that contributed 46 of the responses in this data set.

Fig. 8.4  Cumulative “heat map” data for published studies. In vivo tumor panels are shown at the top, and agents tested are shown in the left column. Median group response for each treatment is color-coded according to the key (bottom). Black boxes indicate missing data

8  The Pediatric Preclinical Testing Program

207

Hence the response rate for “known active agents” was 30.7% (46/150), whereas the overall response rate for “novel” agents was 9.4% (90/963). The second method of graphic representation is the COMPARE format, modeled after the format used to present NCI in vitro data from the 60 cell line panel (Fig. 8.5). This format is more useful for visualizing individual agent activity, or for comparing differential or similar activity between two agents. The example shown demonstrates the single agent activity for cyclophosphamide and the ­experimental agent MLN8237, an inhibitor of Aurora kinase A. The point illustrated is that while cyclophosphamide has little activity against neuroblastoma models derived from patients that had received cyclophosphamide treatment [25], MLN8237 has significant activity against several neuroblastoma lines. The PPTP has also undertaken limited evaluation of agents where histotype selectivity was anticipated (anti-CD19 toxin against ALL models or cloretazine against glioblastomas), or where a “graded” approach was considered to ­demonstrate differential activity of a novel agent having the same target as an agent previously evaluated (e.g. topoisomerase 1) before evaluating the agent against the full in vivo panel.

Fig. 8.5  COMPARE representation of tumor sensitivity based on the difference of individual tumor lines from the midpoint response (stable disease). Bars to the right of the median represent lines that are more sensitive, and to the left are tumor models that are less sensitive. Red bars indicate lines with a significant difference in EFS distribution between treatment and control groups, while blue bars indicate lines for which the EFS distributions were not significantly different. Left: activity spectrum for cyclophosphamide; Right: activity of MLN8237. From Houghton et al. [23] and unpublished data

208

C.L. Morton and P.J. Houghton

8.5 Secondary Screening As mentioned above, the primary function of the Stage 1 evaluation is to identify significant antitumor activity, and to avoid “false negative” results. False negatives may be a consequence of low mouse tolerance to drug relative to humans, whereas false positives may be a consequence of high tolerance to drug of the tumor host (mouse) relative to humans [5, 26]. Hence, in secondary screening it is imperative to address this fundamental issue by defining the dose–effect relationship, and through pharmacokinetic determination of drug exposure in mice compared to that achieved at the MTD in phase 1 clinical trials [27, 28]. As shown in Fig. 8.4, the pre-prodrug PR104 demonstrated very high activity against multiple tumors when administered at its MTD. However, it lost essentially all activity at 50% MTD dosing, suggesting a very narrow therapeutic window. Pharmacokinetic modeling for PR104 and its active metabolite indicated that the mouse exposure relative to that achieved at the phase 1 MTD was ~10-fold greater per 21-day cycle of treatment. Hence, it is unlikely that effective drug exposures will be achieved to induce significant antitumor activity in the solid tumor models (false positive), although this is less clear with respect to the ALL panel. These results can at least be a one factor in considering whether to advance this agent for clinical trials in children. The problem of “false negative” results is no less problematic. For example, topoisomerase II poisons (epipodophyllotoxins, anthracyclines) are highly effective clinical agents, but tend to demonstrate only moderate antitumor activity in human xenograft models. In part, this is a consequence of low tolerance to anthracyclines (especially in SCID mice), and both low tolerance and also rapid clearance of etoposide in mice relative to that in children. As agents tested in the PPTP screen are either close to clinical evaluation, or have advanced to clinical trial, the great majority will have demonstrated some level of preclinical activity against mouse models (usually carcinoma models derived from adult tumor cell lines). Reference to Fig. 8.4 reveals many such agents that demonstrate minimal activity against childhood tumor models. For example small molecule inhibitors alvespimycin, lapatinib, vorinostat, AZD6244 that target HSP90, ERBB1/2, HDACs, and MEK1/2, respectively, or the TRAIL agonist antibody HGS-ETR2 show minimal activity in Stage 1 testing. This raises the issue of potential false-negative results as a consequence of under-dosing and not achieving adequate systemic exposure. Alternatively, the drug target may not be essential or critical to the growth or survival of these tumor cells in vivo. In the latter case the drug acts as a chemical probe to elucidate the underlying biology of these tumors. One approach to distinguishing between under dosing and the relevance of the drug target is to conduct pharmacodynamic studies to ascertain the degree and duration of target inhibition. For alvespicamycin, biomarkers for HSP90 inhibition are loss of client proteins, and induction of HSP70, and these changes were observed in tumors irrespective of their response to alvespimycin, suggesting that target inhibition was achieved. Similarly, the MEK1/2 inhibitor AZD6244 completely abrogated phosphorylation of its downstream substrates (ERK1/2), despite tumor progression, suggesting that

8  The Pediatric Preclinical Testing Program

209

this pathway has little relevance to tumor growth (at least alone). In contrast to its lack of activity against the primary in vivo panel, AZD6244 induced regressions in a pilocytic astrocytoma xenograft model where mutant BRAF (V600E) is constitutively activated, but demonstrated no activity against a pilocytic astrocytoma xenograft model that has amplified wild-type BRAF.

8.6 Combination Drug Testing The PPTP now evaluates approximately four experimental drugs in combination with standard agents per year. We considered experimental designs that would allow formal testing of synergy or antagonism between agents [29, 30]. However, such designs require multiple dose groups for each agent, severely restricting the number of tumor models that could be evaluated. Thus, rather than testing a combination rigorously in one or two models, the PPTP approach has been to determine whether a combination has a more general effect that can be extrapolated to multiple models in vivo. A typical experimental design used for testing rapamycin (Table 8.3), although sub-optimal for assessing true synergy or antagonism, allows the combinations to be tested against a reasonable number of models. The rationale for combining experimental agents with standard cytotoxic drugs used in the treatment of childhood cancer is that any new agent introduced into the clinic will almost certainly be added to standard of care protocols that incorporate these agents. The experimental design presented allows the determination of whether adding a novel agent enhances the antitumor activity of the standard agent over that achieved at the MTD for the standard agent alone. Testing the combination with the lower dose of standard agent (0.5 × MTD) allows for determination of enhanced therapeutic activity over that achieved at the MTD, and where the experimental agent may exacerbate Table 8.3  Experimental design for combination drug testing (example rapamycin) RAP @ VCR @ VCR @ Cytoxan Cytoxan @ Cisplatin Cisplatin Group 5 mg/kg MTD 0.5MTD @MTD 0.5MTD @MTD @0.5MTD A – – – – – – – B – X – – – – – C – – X – – – – D – – – X – – – E – – – – X – – F – – – – – X – G – – – – – – X H X – – – – – – I X X – – – – – J X – X – – – – K X – – X – – – L X – – – X – – M X – – – – X – N X – – – – – X

210

C.L. Morton and P.J. Houghton

host toxicity. For example, rapamycin enhanced the toxicity of cisplatin given at its MTD, but the ­combination with cisplatin administered at a lower dose level with rapamycin was superior to single agent cisplatin administered at the MTD in all models evaluated. The experimental design also allows determination of whether the cytotoxic agent impacts on the single agent activity of the experimental drug.

8.7 Secondary Models The PPTP Stage 1 and Stage 2 evaluations are conducted using subcutaneous xenografts for solid tumors, whereas the ALL panel is for disseminated disease. Potentially, the site of tumor growth could impact drug sensitivity, either positively or negatively. For example, a brain tumor growing in the flank of a mouse may respond to an agent that has poor penetration to the brain due to the blood–brain barrier. Within the PPTP there is capability to undertake secondary screening against orthotopic and disseminated models of brain tumors (intracerbral injection), rhabdomyosarcoma, neuroblastoma [31], and osteosarcoma models. Several transgenic models (neuroblastoma, rhabdomyosarcoma, and medulloblastoma) are also incorporated into the PPTP; however, use of these models is severely restricted due to the Dupont patent restrictions and licensing fees imposed on the usage of these genetically engineered mice for drug development. The PPTP is in a novel position to test the validity of these engineered models in the context of pediatric drug ­development, but it is unlikely that significant testing will be undertaken due to the restrictions imposed by Dupont, as described above.

8.8 Integrating Molecular Data with Drug Sensitivity The PPTP has established two databases for expression profiles and SNP analysis of untreated tumors. Potentially, it will be possible to interrogate these data with respect to tumor sensitivity with the object of identifying signatures that predict drug sensitivity. An example of this approach for analysis of MLN8237 is shown in Fig. 8.6. In this example a signature based on 261 probe sets distinguishes between tumors that regress from those that progress. Taking an approach similar to that used in the POPP-TAP [16], it may be possible to train an artificial neural network with such profiles of drug sensitive or resistant tumors and to potentially identify other xenograft tumors that would be sensitive to this agent.

8.8.1 Submitting Agents to the PPTP The PPTP has the capacity to test approximately 12 agents or combinations of agents annually in its in  vivo preclinical models. Agents are selected for PPTP ­testing based on their potential relevance in the childhood cancer setting and based

8  The Pediatric Preclinical Testing Program

211

Fig. 8.6  Expression signatures that distinguish responsive tumors from non-responders treated with MLN8237. Affymetrix U133 plus 2 expression profiles (untreated tumors) from ten tumor lines with objective responses (>PR) or nine non-responders (PD1) were analyzed. Left panel: differentiation based on 261 probesets that distinguish sensitive from resistant tumors. Top panel, Principal component analysis (PCA) representation of differential showing tumor histotypes (color key at bottom). Lower panel: PCA showing differentiation of sensitive (green) and resistant (red) tumors independent of histotype

on their stage of clinical development. Details of the submission process, and Materials Transfer Agreement (MTA) between the NCI and organization submitting the agent can be http://pptp.nchresearch.org/documents.html. The Pediatric Drug Development Group (PedDDG) is the group that recommends whether agents should be prioritized for evaluation by the PPTP. The PedDDG is primarily composed of NCI staff with drug development and pediatric oncology expertise, the Principal Investigator for the PPTP and the Chair of the Children’s Oncology Group Phase 1 Consortium. The application should address the specificity of the agent, preclinical studies (in  vitro and in  vivo), drug pharmacokinetics (preclinical and clinical where available), commitment to clinical development, as well as intellectual property issues.

8.9 Closing Remarks One significant concern is that the PPTP in vivo panel does not accurately represent the diversity of molecular subtypes for each histotype. At present the tumor panels are relatively small (for example only four glioblastomas), and additional models have been

212

C.L. Morton and P.J. Houghton

developed since inception of the PPTP that could be incorporated should resources become available. Under-represented tumors include anaplastic Wilms tumors, nonrhabdomyosaroma, soft tissue sarcomas, and leukemias other than ALL. Ideally, each panel should reflect clinical subtypes, and include at least 12 models representing a disease type would allow more informative “preclinical” phase 2 trials. A second challenge will be to integrate the molecular databases (expression profiling, SNP analysis) and drug response data. There exists the potential to identify markers of tumor response that are either histotype specific or cross histology groups. Identification and prospective validation of expression signatures, or copy number alterations that correlate with response could potentially prove informative for stratification of patient therapy. It is too early to determine whether the PPTP approach is valid for prioritizing agents for pediatric clinical trials. Importantly, results generated through the PPTP do not serve as a restriction point for drug evaluation in pediatric populations. Further clinical evaluation of PPTP “actives” and “inactives” will be required to ultimately determine the validity of this approach. On the basis of PPTP data, MLN8237, the Aurora kinase A inhibitor, was “fast-tracked” and has completed phase 1 testing in children. The lytic virus NTX-001 may be tested relatively soon. The rapamycin analog, temsirolimus, is in phase 2 testing, and combination testing has started. Lapatinib as a single has been extensively tested in patients with brain tumors, hence comparison of its clinical activity relative to the PPTP brain tumor data should soon be available.

References 1. Reis LA, Percy CL, Bunin GR. Cancer incidence and survival among children and adolescents. United States SEER Program. 1975–1995. NIH (Pub. No. 99–4649). Bethesda, MD: National Cancer Institute; 1999. p. 1–15. 2. Ries LA, Eisner MP, Kosary CL, et al. SEER Cancer Statistics Review 1975–2002. Bethesda, MD: National Cancer Institiute; 2005. 3. Gurney JG, Young JL, Roffers SD, Smith MA, Bunin GR. Cancer incidence and survival among children and adolescents: United States SEER Program. 1975–1995 NIH (Pub. 99–4649). Bethesda, MD: National Cancer Institute; 1999. p. 111–24. 4. Horowitz ME, Etcubanas E, Christensen ML, et al. Phase II testing of melphalan in children with newly diagnosed rhabdomyosarcoma: a model for anticancer drug development. J Clin Oncol. 1988;6:308–14. 5. Peterson JK, Houghton PJ. Integrating pharmacology and in vivo cancer models in preclinical and clinical drug development. Eur J Cancer. 2004;40:837–44. 6. Houghton PJ, Adamson PC, Blaney S, et  al. Testing of new agents in childhood cancer preclinical models: meeting summary. Clin Cancer Res. 2002;8:3646–57. 7. Boland I, Vassal G, Morizet J, et al. Busulphan is active against neuroblastoma and medulloblastoma xenografts in athymic mice at clinically achievable plasma drug concentrations. Br J Cancer. 1999;79:787–92. 8. Friedman HS, Colvin OM, Ludeman SM, et  al. Experimental chemotherapy of human medulloblastoma with classical alkylators. Cancer Res. 1986;46:2827–33. 9. Friedman HS, Houghton PJ, Schold SC, Keir S, Bigner DD. Activity of 9-dimethylaminomethyl-10-hydroxycamptothecin against pediatric and adult central nervous system tumor xenografts. Cancer Chemother Pharmacol. 1994;34:171–4.

8  The Pediatric Preclinical Testing Program

213

10. 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. 11. Liem NL, 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. 12. Santos A, Calvet L, Terrier-Lacombe MJ, et  al. In vivo treatment with CPT-11 leads to differentiation of neuroblastoma xenografts and topoisomerase I alterations. Cancer Res. 2004; 64:3223–9. 13. Shusterman S, Grupp SA, Barr R, Carpentieri D, Zhao H, Maris JM. The angiogenesis inhibitor tnp-470 effectively inhibits human neuroblastoma xenograft growth, especially in the setting of subclinical disease. Clin Cancer Res. 2001;7:977–84. 14. Jaboin J, Wild J, Hamidi H, et al. MS-27–275, an inhibitor of histone deacetylase, has marked in  vitro and in  vivo antitumor activity against pediatric solid tumors. Cancer Res. 2002; 62:6108–15. 15. Peterson JK, Tucker C, Favours E, et al. In vivo evaluation of ixabepilone (BMS247550), a novel epothilone B derivative, against pediatric cancer models. Clin Cancer Res. 2005;11:6950–8. 16. Whiteford CC, Bilke S, Greer BT, et al. Credentialing preclinical pediatric xenograft models using gene expression and tissue microarray analysis. Cancer Res. 2007;67:32–40. 17. Garber K. From human to mouse and back: “tumorgraft” models surge in popularity. J Natl Cancer Inst. 2009;101:6–8. 18. Neale G, Su X, Morton CL, et al. Molecular characterization of the pediatric preclinical testing panel. Clin Cancer Res. 2008;14:4572–83. 19. Smith MA, Morton CL, Phelps D, Girtman K, Neale G, Houghton PJ. SK-NEP-1 and Rh1 are Ewing family tumor lines. Pediatr Blood Cancer. 2008;50:703–6. 20. Fox EJ, Salk JJ, Loeb LA. Cancer genome sequencing – an interim analysis. Cancer Res. 2009;69:4948–50. 21. Morton CL, Houghton PJ. Establishment of human tumor xenografts in immunodeficient mice. Nat Protoc. 2007;2:247–50. 22. Morton CL, Papa RA, Lock RB, Houghton PJ. Preclinical chemotherapeutic tumor models of common childhood cancers: solid tumors, acute lymphoblastic leukemia and disseminated neuroblastoma. Curr Protocol Pharmacol. 2007;39:14.8.1–.21. 23. Houghton PJ, Morton CL, Tucker C, et al. The pediatric preclinical testing program: description of models and early testing results. Pediatr Blood Cancer. 2007;49:928–40. 24. Plowman J CR, Alley M, Sausville E, Schepartz S. US-NCI testing procedures. In: Feibig HHBA, editor. Relevance of tumor models for anticancer drug development: Basel: Karger; 1999. p. 121–35. 25. Thompson J, Zamboni WC, Cheshire PJ, et al. Efficacy of systemic administration of irinotecan against neuroblastoma xenografts. Clin Cancer Res. 1997;3:423–31. 26. Kurmasheva RT, Houghton PJ. Pediatric oncology. Curr Opin Chem Biol. 2007;11:424–32. 27. Kirstein MN, Houghton PJ, Cheshire PJ, et  al. Relation between 9-aminocamptothecin systemic exposure and tumor response in human solid tumor xenografts. Clin Cancer Res. 2001;7:358–66. 28. Leggas M, Stewart CF, Woo MH, et al. Relation between Irofulven (MGI-114) systemic exposure and tumor response in human solid tumor xenografts. Clin Cancer Res. 2002;8:3000–7. 29. Tan M, Fang HB, Tian GL, Houghton PJ. Experimental design and sample size determination for testing synergism in drug combination studies based on uniform measures. Stat Med. 2003;22:2091–100. 30. Tan M, Fang HB, Tian GL, Houghton PJ. Repeated-measures models with constrained parameters for incomplete data in tumour xenograft experiments. Stat Med. 2005;24:109–19. 31. Thompson J, Guichard SM, Cheshire PJ, et al. Development, characterization and therapy of a disseminated model of childhood neuroblastoma in SCID mice. Cancer Chemother Pharmacol. 2001;47:211–21.

Chapter 9

Imaging Efficacy in Tumor Models Vinod Kaimal, Wilbur R. Leopold, and Patrick McConville

Abstract  Imaging modalities such as magnetic resonance imaging (MRI), positron emission tomography (PET), single-photon emission computed tomography (SPECT), X-ray computed tomography (CT), ultrasound imaging and optical imaging techniques are increasingly being used to assess disease progression and response to therapeutics in animal models of disease. In oncology, the applications of these modalities range from volumetric assessment of tumor burden to imaging of functional parameters such as tumor vascularity and permeability, cell proliferation and tissue hypoxia. Quantitative assessment of molecular events such as receptor occupancy or upregulation of enzymes (e.g.matrix metalloproteinases) is also possible. Further, the use of imaging has enabled the use of more realistic models of disease (e.g. orthotopic/metastatic models) and the generation of more predictive data in a noninvasive manner. This chapter aims to provide a brief overview of the various imaging modalities, followed by a discussion of the most relevant oncology endpoints that are accessible by imaging. Keywords  Imaging • MRI • PET • SPECT • CT • Bioluminescence • Optical • ultrasound

9.1 Introduction The past two decades have seen rapid development and adoption of imaging modalities for diagnosis and assessment of progression and therapeutic response in a multitude of human diseases. More recently, imaging has found widespread use in animal models, with oncology being the most prevalent area of focus. A broad

W.R. Leopold (*) Charles River Laboratories, Inc., 800 Technology Drive, Ann Arbor, MI 48108, 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_9, © Springer Science+Business Media, LLC 2011

215

216

V. Kaimal et al.

array of imaging technologies are routinely used for anatomical detection of tumors and assessment of tumor growth and response in the full spectrum of cancer models and are relied on particularly in orthotopic, metastatic, and transgenic tumor models. Additionally, today’s molecular targeted therapies have increased the use of imaging technologies for quantifying drug-induced changes in physiology (commonly termed functional imaging) and for detecting drug-mediated cellular and molecular level events (commonly termed molecular imaging). Imaging methods can enable quantitative assessment of target modulation, and simultaneous measurement of traditional growth and lifespan based endpoints. The increasing focus on imaging endpoints to increase efficiency and predictive power in clinical trials further motivates the use of preclinical imaging to validate and optimize clinical imaging strategies. This chapter will begin with a description of the unique challenges posed by in vivo preclinical imaging, followed by an overview of the most prevalent preclinical imaging modalities and the basis for signal detection and generation of spatially resolved signal. The chapter will then discuss the most relevant oncology endpoints that are accessible by imaging.

9.2 Challenges in Preclinical Imaging Increased reliance on imaging technologies in clinical use has been one of the major driving factors in adaption of clinical imaging hardware and software standards to preclinical use. However, preclinical translation of even the most mature clinical imaging technologies has presented significant challenges related to achieving the higher resolution required in small animals (most commonly mice and rats) [1, 2]. The last decades have seen substantive advances in preclinical imaging hardware and the introduction of commercially available MRI, CT, PET [3], SPECT [4], and ultrasound imaging systems dedicated to small animals [5–7]. An increasing number of imaging technologies currently restricted to preclinical use have also recently been developed, many based on optical imaging technologies. Even though in vivo preclinical imaging technologies have been commercially available for the better part of the last decade, their widespread use in testing drug candidates has been more recent. Throughput, sensitivity, and resolution limitations have contributed to the cautious adoption of imaging technologies for testing efficacy [8], particularly by the pharmaceutical and contract research industry. Generally, optimization of any one of sensitivity, resolution, or imaging time requires sacrifice in at least one of the other two (Fig. 9.1). Despite these challenges, in the last 5 years, almost all large pharmaceutical companies have established their own imaging centers and imaging programs, and, even more recently, the major in vivo contract research organizations (CROs) have established in-house imaging programs. This trend is driving more rapid development and optimization of imaging technology, protocols, biomarkers, and probes

9  Imaging Efficacy in Tumor Models

217

ACQUISITION TIME Throughput Efficiency Statistical Power

SENSITIVITY Limit of detection Dynamic range Signal to Noise ratio

RESOLUTION Spatial information Tissue delineation

Fig. 9.1  Imaging is generally a compromise between the three parameters depicted above. An improvement in any one of these parameters results in some sacrifice in one or both of the others. For example, averaging of the MRI signal is a common approach to improving the SNR but results in longer scans. A higher resolution image leads to lower SNR, which has to be ­compensated by longer imaging time (i.e. averaging) assuming all hardware remains the same

for testing efficacy. The following sections of this chapter will highlight some of the more prevalent and accepted preclinical imaging technologies and approaches for determination of drug efficacy in oncology.

9.3 Imaging Modalities: Technical Overview and Use in Oncology Models 9.3.1 Magnetic Resonance Imaging Magnetic resonance imaging [9] is a clinically relevant imaging modality that relies on signals received from nuclei (most commonly protons) placed in a magnetic field and perturbed using radio-frequency (rf) pulses. The use of magnetic field gradients enables spatial resolution of signal and reconstruction of images in three dimensions. Major work in magnetic resonance began with Rabi et  al. [10] and the principle of nuclear magnetic resonance (NMR) was discovered independently by Bloch [11] and Purcell [12] in 1946. In the 1970s, several accomplishments led to the development of MRI [13, 14], with Lauterbur and Mansfield sharing the Nobel prize for its discovery in 2003. Systems specifically optimized for preclinical in  vivo imaging were introduced commercially about 20 years ago. MRI enables relatively high spatial resolution (~10 mm) but suffers from relatively poor sensitivity and motion artifacts (e.g. due to respiration and the cardiac cycle). Although prevention of motion artifacts can be achieved by a

218

V. Kaimal et al.

breath hold in clinical patients, animal imaging poses greater challenges with respect to motion. However, MRI enables unparalleled flexibility with choice of tissue contrast, leading to a broad spectrum of MRI-determined endpoints that are used in preclinical and clinical efficacy testing. For example, the MRI signal can be manipulated to measure tissue water content [9], water diffusion [15], oxygenation and blood flow [16], perfusion [17], temperature [18], elastivity [19], and metabolite concentrations [20]. An extensive array of exogenous probes (e.g. 19F, FeOx, Gd, CEST, Mn, Cr) [21–23] can also be ­targeted to specific cells, tissues or processes to enable molecular imaging [24] and cell tracking [25, 26]. The broad flexibility associated with tissue contrast in MRI has led to its use as a clinical standard in oncology in most major tissues, as well as in imaging the full spectrum of preclinical tumor models. In preclinical models, MRI is particularly well suited to imaging of the brain, and lower abdominal tissues. Liver MRI is also particularly well enabled due to the high natural iron content of normal liver tissue. This provides high contrast between the liver and lesions with abnormal pathology such as liver metastases or primary liver tumors [27]. The use of MRI in tissues near the thorax generally requires the use of respiratory and/or cardiac gating, to synchronize the MRI signal acquisition to the respiratory or cardiac cycle, thereby preventing motion-induced artifacts [28–30].

9.3.2 Computed Tomography Computed Tomography (CT) is based on the attenuation of X-rays as they pass through the subject. A three-dimensional image can be reconstructed from a series of projections acquired as an X-ray source and detector rotate together around the subject. Following their discovery by Wilhelm Rontgen around 1895, X-rays have been used in imaging applications. Although several individuals were instrumental in developing tomographic imaging using X-rays, Allan Cormack and Godfrey Housenfield are credited with the invention of the modern CT [31]. Preclinical CT, commonly known as microCT, allows resolution to approximately 5 mm and is generally not limited by sensitivity. Modern technologies with continuously moving source and detector allow rapid scanning at full resolution. Soft tissue contrast in CT is generally poor due to poorly differentiated X-ray attenuating properties of normal soft tissues. Therefore, traditionally, CT has been used for imaging the skeletal system. However, the development of CT contrast agents with differential tissue uptake properties and dual energy X-ray technologies has led to recent advances in soft tissue CT. In preclinical cancer models, CT imaging of bone invasion or remodeling by bone metastases [32] and multiple myeloma [33] is prevalent. Naturally high CT contrast in the airways has also led to extensive use of microCT in lung tumor models [34], commonly with respiratory and/or cardiac gating [35]. The application of high-resolution X-ray CT to soft tissue tumors (including liver [36], ­pancreas [37], spleen [36], kidney [38], and mammary fat pads [39]) has increased

9  Imaging Efficacy in Tumor Models

219

with the use of CT contrast agents [24, 40]. The introduction of rapid scan enabled preclinical CT systems and blood pooling contrast agents has increased the use of CT to image the vasculature [41] and tissue perfusion [42, 43].

9.3.3 Positron Emission Tomography Positron Emission Tomography is based on the detection of gamma photons that are emitted after positron–electron annihilation. Unstable nuclei containing a high number of protons (e.g. 18F) stabilize by the emission of positrons, which are captured by the surrounding electron cloud with the emission of photons. Coincidence counting and calculation of origin of the annihilation events based on trajectories of the gamma photons enable reconstruction of a spatially accurate image. Commonly used PET isotopes include 18F, 13N, 14C, 64Cu, and 124I. PET imaging was first discovered in the early 1950s and is primarily credited to Gordon Brownwell [44]. The first systems truly optimized for small animal imaging were released in the early 1990s [45, 46]. PET has relatively low resolution (~1–1.5 mm at best) but relatively high sensitivity. PET imaging is inherently quantitative as the PET signal is a direct measure of the tracer concentration, although accuracy depends on a complex series of corrections for detector non-idealities, random coincidences, scatter, and attenuation. Sensitivity and resolution are also highly dependent on the reconstruction algorithm, with iterative numerical methods often preferred [47, 48]. PET is used to study tumor biology using a wide variety of tracers with specific properties that enable differentiation of cancerous tissue from normal tissue, by preferential uptake into, or longer clearance from tumor tissue compared with surrounding normal tissues. Commonly available PET tracers include those used to measure cellular glucose metabolism (18F-fluorodeoxyglucose, FDG), cellular proliferation ([18F]-fluorothymidine, FLT), DNA synthesis ([18F]-1-(2¢-deoxy-2¢fluoro-b-d-arabinofuranosyl)thymine, FMAU) [49], cell membrane synthesis ([11C]- methionine), protein synthesis ([18F]-tyrosine, 18F-FTyr), and tissue hypoxia ([18F]-fluoromisonidasol, 18F-MISO]) [50–56]. Radioimmunotracers targeting tumor-specific antigens can also be used to differentiate normal and malignant tissues [57, 58]. The applications of PET imaging in preclinical models [59] are limited by (i) resolution, as tumors of less than several millimeters in dimension may not be robustly detected or may lead to partial volume errors [60], (ii) tumor uptake of the tracer (sensitivity), and (iii) tumor to background signal ratio (tumor specificity) which affects the accuracy of image segmentation. Beyond these technical limits, PET is limited only by the development, availability, and validation of PET-enabled radiotracer molecules. The use of traditional subcutaneous tumor models has been particularly successful and widespread in preclinical PET as this helps to overcome the resolution limitation (as tumor sizes of at least 3–4 mm in diameter are typically used) and the background signal limitation, as the tumor protrudes from the body and can be segmented with high accuracy.

220

V. Kaimal et al.

9.3.4 Single Photon Emission Computed Tomography Single photon emission computed tomography (SPECT) is based on detection of gamma photons emitted from the subject after systemic administration of a radioactively labeled tracer (analogous to PET). A three-dimensional distribution of radioactivity can be computed after generation of multiple two-dimensional projections (analogous to CT). SPECT imaging is accomplished by gamma camera detection of lower energy photons (compared with PET), but allows a much broader spectrum of energies and the detection of multiple isotopes of different energies simultaneously. Sub-millimeter resolution is feasible using modern small animal SPECT systems, through the use of pinhole collimators. Sensitivity is also relatively high, and can be improved by the use of multiple detectors. 99mTc is the most commonly used isotope in SPECT, although other isotopes such as 67Ga, 111In, and 123I are also prevalent [24]. The detector technology used in SPECT is essentially the same as that used in traditional gamma camera imaging, and therefore SPECT can be applied to a broad spectrum of approved and commonly used radiopharmaceuticals which is a key advantage. SPECT isotopes are also generally cheaper, more broadly accessible and longer lived than PET isotopes. Similar to PET, the reliance on exogenous radiotracers means that many of the same limitations of PET are present for SPECT as well. The accuracy of the image depends on correction for scatter, attenuation, and detector non-ideality. As with PET, the use of SPECT in preclinical oncology models can be limited by resolution (tumor size), tumor uptake, and background tissue uptake (tracer specificity). These factors can limit preclinical SPECT imaging in metastatic and orthotopic models, where tumor sizes are typically smaller than those in subcutaneously implanted tumors.

9.3.5 In Vivo Optical Imaging In vivo optical imaging is now widely prevalent in preclinical oncology efficacy testing through bioluminescence imaging (BLI) and fluorescence imaging (FLI). BLI is a relatively recently developed optical imaging method that relies on detection of light from luciferase-expressing cells in an animal [40, 61]. This is commonly achieved through implantation of cells engineered to express luciferase constituitively (e.g. tumor cells) or by use of transgenic animals that express luciferase in one or more tissues of interest. Emission of light at 460–630 nm [62] from these cells or tissues occurs following access of the luciferase substrate, luciferin, which is generally introduced by a systemic injection. In the presence of the luciferase catalyst, ATP and oxygen, luciferin is rapidly converted to oxyluciferin, producing light. FLI relies on light emission from a fluorophore, after excitation by a light source at the appropriate wavelength for the fluorophore in question. Excitation and emission wavelengths are typically in the 500–900 nm range. FLI can be applied to cell lines or transgenics expressing fluorescent ­proteins

9  Imaging Efficacy in Tumor Models

221

(analogous to luciferase) or by the use of exogenous probes labeled with fluorophores. In vivo optical imaging technologies rely on a light tight imaging box with a CCD camera array cooled to approximately liquid nitrogen temperature to minimize noise and maximize sensitivity for low levels of light [62, 63]. Resolution is limited to a few millimeters, but sensitivity is commonly not limiting, for BLI or FLI. In BLI, systemic luciferin at saturation levels (15 mg/kg in mice) results in the emission of up to 10–104 photons/sec/cell in cells engineered to express luciferase and can enable in vivo detection of a single cell [64]. Imaging times for BLI are typically seconds to a few minutes. In FLI, the emitted light intensity is measured as a proportion of the excitation light intensity, with imaging times of milliseconds to seconds typically adequate to obtain the required sensitivity. The limitations of optical imaging are related to the poor depth of penetration of light (~mm) in the visible to near-infrared spectrum, and the efficient scattering properties of normal tissue for these photons. While BLI has the advantage of the absence of a systemic background in normal animals, FLI suffers from high background autofluorescence in the visible to nearinfrared region.Excitations at multiple wavelengths coupled with spectral deconvolution software are generally required to correct for background autofluorescence. In vivo optical imaging is most commonly a two-dimensional modality, and images from multiple animal positions can be necessary to detect signals throughout the body and to unambiguously identify the anatomical location of a given signal. However, newer in vivo optical imaging technologies [65] enable three-dimensional imaging using tomographical techniques (e.g. by the use of mirrors) [62] or by rastering [66] techniques. The use of subcutaneous xenograft models can overcome the poor depth of penetration and lack of three-dimensional information. In FLI, the use of newer near-IR fluorophores that enable narrow bandwidth and well-separated excitation and emission minimizes background autofluorescence and maximizes tissue penetration of the light [67]. Both BLI and FLI can offer the advantage of imaging several animals simultaneously, making in  vivo optical imaging a relatively high-throughout modality. Both modalities are widely used for monitoring tumor burden using engineered tumor cell lines [68] or exogenous tumor-specific probes [69].

9.3.6 Ultrasound High-frequency sound propagates through tissue as pressure waves and is reflected at tissue interfaces. The reflected signal can be used to reconstruct images in real time, forming the basis for ultrasound imaging. Contrast between tissue types is determined by differences in acoustic impedance, which causes differential backscatter of the sound. Contrast can be improved using agents based on microbubbles [70, 71], or echogenic liposome formulations [72]. Ultrasound is a clinically prevalent imaging technology which was first used clinically over 50 years ago. The first specialized small animal ultrasound systems were delivered commercially in the last 5–10 years and commercial systems

222

V. Kaimal et al.

now offer real-time resolution down to approximately 30 mm. Traditionally, ultrasound has been a two-dimensional modality, but clinically developed technologies have evolved to enable three-dimensional imaging. The limitations of ultrasound have centered around poor soft tissue contrast and variability from manual positioning of the transducer, and the variable positions of internal organs. Despite this, ultrasound has seen increased use in oncology models and, in addition to anatomical detection of tumors, can be used for detection and quantification of tissue blood flow, through color Doppler [73], power Doppler [74], and pulsed wave Doppler imaging [75]. Ultrasound is therefore increasingly prevalent in the assessment of vascular development and response to therapy [76]. Microbubble contrast agents have also enabled dynamic assessment of tissue perfusion [71].

9.4 Imaging End Points in Oncology Models 9.4.1 Anatomical Imaging 9.4.1.1 Tumor Detection and Staging The ability for imaging techniques to distinguish and spatially localize cancerous tissue from normal tissue has led to it being a clinical standard for tumor diagnosis and detection of primary tumors and metastases. Preclinically, imaging has only become prevalent as a detection technique with the increasing use of orthotopic, metastatic, and transgenic mouse models. In these models, imaging is often the only non-invasive way to detect the appearance of a tumor, and in transgenic or metastasis models, detection and incidence can be the primary endpoint. Any imaging modality that can distinguish cancerous from normal tissue can be used for detection, and for this reason all the major modalities previously discussed in this chapter have commonly been used for this purpose. Due to the inherent flexibility of MRI contrast choice, it is often the modality of choice for detection of tumors in preclinical models, particularly in brain [77, 78], liver [79, 80], and kidney models [81]. 18 F-FDG PET can offer high tumor to background tissue contrast by the generally higher metabolic rates in tumor tissues and can therefore be used to detect tumors. Similarly, PET tracers such as 18F-FLT and 18F-MISO are frequently used for detection by their tumor specificities. In vivo optical imaging offers an attractive method for tumor detection, staging, and group matching prior to treatment due to its high throughput, and the relative ease of image generation and analysis, compared with other modalities. Luciferase or fluorescent protein expressing tumor lines are commonly used for this purpose [82, 83], particularly in metastatic models [84–86], and exogenous, tumor-specific fluorescent probes can also be used in this manner.

9  Imaging Efficacy in Tumor Models

223

9.4.1.2 Tumor Burden The adoption of major clinical imaging modalities for tumor diagnosis has undoubtedly positioned imaging techniques strongly for taking the next step: quantifying tumor growth or tumor burden and tracking it over time. Traditionally, tumor burden has been measured in terms of volume or weight. Other than lifespan, there is no more accepted method for assessing efficacy of a therapy than by quantification of tumor volume over time. The higher-resolution imaging modalities, MRI (Fig. 9.2a) and CT, have seen very common use for measuring tumor size, and are accurate methods for this purpose in a wide variety of models. To a lesser extent, ultrasound is used for similar means, with increased use in parallel with development of ultrasound transducer technology and software for automated definition of tissue boundaries. Contrast agents, such as gadolinium or iron-oxide based agents [23] for MRI, iodine-based agents for CT, and microbubble-based agents [71] for ultrasound can be used to improve tumor delineation and increase the accuracy of tumor burden assessment. Furthermore, imaging methods enable tracking of multiple tumors in a single animal over time. This is particularly important in transgenic models and metastasis models, which typically involve more than one, and commonly dozens of tumors [87, 88]. Not only does this allow assessment of total tumor burden, it also enables tissue specific effects to be differentiated. This is important for the development of tissue targeted agents. This approach can infer not only tumor growth inhibition, but successful tissue specific targeting [89]. While the traditionally anatomical modalities have been the mainstay for imagebased tumor burden assessment, the inherently functional and molecular imaging modalities are seeing increasing use for this purpose. For example, tumor-specific PET or SPECT tracer uptake can be used as a surrogate for tumor burden in certain cases (Fig 9.2b). However, increasing tumor size commonly results in differential vascularization (tracer access) and poorer viability or necrosis in the tumor core,

Fig. 9.2  Estimation of tumor burden and anatomy by a variety of imaging modalities. (a) MRI of an adrenal tumor in a PTEN-/- mouse (b) FDG PET in a PC-3 peri-tibial implant model (c) PC-3-luc bone metastasis imaging using BLI

224

V. Kaimal et al.

which can uncouple the tumor tracer activity from tumor size in a PET or SPECT study, ­particularly in subcutaneous tumors which are traditionally staged and followed to relatively large sizes, compared with deep tissue tumors [90]. Similar to their use in tumor detection, optical imaging methods are attractive for tracking tumor burden due to high efficiency and high sensitivity for doing so in a variety of models [82, 90–92] (Fig. 9.2c). However, since this methodology relies on a strong correlation between tumor size and light produced, the readout can become uncoupled with development of normal heterogeneities as tumors become larger. This can be a problem for genetically engineered luciferase and fluorescent protein expressing tumor lines which rely on supply, oxygenation, and tumor ATP levels, all of which commonly vary spatially in a tumor and in response to therapy. Exogenous substrate (e.g. luciferin) or probe access can also be an issue, as tissue perfusion is significantly heterogeneous in relatively large tumors. Additionally, increasing tumor size creates greater attenuation for light emanating from the center of the tumor, leading to decreasing total light per volume as tumor size increases. For the same reason, the positioning of the mouse and variability in internal organ position or even body weight can affect the optical readout. Despite these limitations, optical imaging can greatly improve the efficiency and accuracy of tumor burden assessment in orthotopic and metastasis models. Recently introduced luciferase vectors have improved expression levels and luciferase activity in mammalian cells and can result in detection of a single cell in vivo. This facilitates optical imaging in deep tissue models, particularly cell seeding models commonly used in studying metastasis.

9.4.2 Functional Imaging The rapid adoption of imaging technologies for drug discovery and development has truly occurred only with the development of imaging techniques that probe tumor physiology, known commonly as “functional imaging.” While these methods provide an opportunity to assess drug mechanism and prove method of action, they can also measure efficacy before traditional tumor growth or lifespan based endpoints can be quantified. Functional imaging methods may also increase specificity (correlation with response as determined by a traditional or clinically relevant end point) or ­sensitivity (magnitude of the response modulation) of the efficacy endpoint, compared with traditional end points. The following sections will discuss some of the prominent tumor physiological properties that can be accessed by imaging.

9.4.2.1 Imaging Tumor Metabolism and Metabolite Levels PET imaging using [18F]-FDG provides one of the most common imaging appro­ aches to in  vivo assessment of tumor metabolism. 18F-FDG is an ­FDA-approved

9  Imaging Efficacy in Tumor Models

225

tracer that is the mainstay of PET imaging in oncology. Since tumors generally have elevated glucose metabolism, [18F]-FDG can be used to quantify changes in tumor metabolism in response to therapy [93] by measurement of tumor 18F-FDG uptake. However, these types of images must be interpreted with care because elevated glucose metabolism may be associated with other physiological processes such as macrophage activity and inflammation. The additional dependence of FDG uptake on properties such as cell kill and vascular supply (which could be construed as a lack of specificity) has led to it being broadly successful in detecting response to therapy [94]. Since modulation of metabolism commonly precedes tumor size changes, 18F-FDG PET can provide a time-saving advantage. The limitations of 18 F-FDG PET include the generally high normal tissue background (decreases sensitivity for tumor), the variability associated with competition between the tracer and circulating glucose and variability introduced by muscular uptake in awake animals. Commonly, therefore, animals are fasted prior to imaging to deplete circulating glucose levels, and the uptake protocol must be carefully designed to minimize variability of tumor uptake from animal to animal [95] (Fig. 9.3b). Fluorescent reporters have also been coupled to deoxyglucose in place of the radionuclide 18F, enabling imaging of tumor metabolism using optical techniques [96, 97]. This approach addresses some of the issues with PET imaging, such as expensive instrumentation, availability, and short half life of the isotope.

Tumor Luciferase Signal 100 Luciferase Signal + SE (x106 photons/sec)

Mean Tumor Burden (mg) + SE

Tumor Burden (MRI) 100

10

1

0.1 10

15

20

25

30

35

Days Post Tumor Implant

40

45

10 1 0.1 0.01

10

15

20

25

30

35

40

45

Days Post Tumor Implant

Fig. 9.3  Tracking a tumor over time using MRI and BLI in a U87-MG-luc intracranial implant tumor model: T2 weighted MRI is used to calculate tumor volume. The tumor luciferase signal can also be used to track tumor burden. MRI enables higher resolution and three-dimensional assessment of tumor burden over time, while BLI enables higher throughput and greater sensitivity in a similar assessment

226

V. Kaimal et al.

Magnetic resonance spectroscopy (MRS) (and related MRS imaging, MRSI) involves the combination of MR imaging techniques with conventional NMR spectroscopic methods to provide spatially localized spectra (Fig. 9.4). Since the MR signal is dependent on the local chemical environment of nuclei, MR spectroscopy can resolve nuclei from different metabolites by their chemical shift, enabling ­quantification of metabolite levels over time. Common MRS imaging protocols can provide information on metabolite levels from single volume elements as small as a few cubic millimeters. Pathologic tumor biology provides unique spectral signatures of various metabolites which distinguish tumors from normal tissue, enabling metabolite levels to be used as a surrogate for efficacy. For example, MRS methods can be used to quantify the levels of metabolites important to tumor function such as choline, lactate, and lipids using 1H MRS [98], ATP, and phosphocreatine using 31 P MRS [40, 99, 100]. The concentrations of these metabolites can be used as end points for tumor response to therapy.

Choline

3.5

3 2.5 Frequency (ppm)

Choline

3.5

NAA

3

2

NAA

2.5

2

Frequency (ppm)

Fig. 9.4  Bioluminescence image overlaid on a T2-weighted anatomical MRI in a mouse ­intracranial glioma model (U87-MG-luc). Proton MR spectroscopy performed on the indicated voxel reveals reduction in tumor choline concentration in the treated tumor (below) compared to the untreated control (above). This example demonstrates the use of bioluminescence for relatively high throughput staging, screening and tumor burden tracking, while the anatomical MRI facilitates voxel placement for proton MRS-based quantification of tumor choline concentration and response to therapy

9  Imaging Efficacy in Tumor Models

227

9.4.2.2 Cell Proliferation DNA synthesis in the tumor can be quantified by PET imaging of an 18F labeled form of the nucleoside thymidine, 18F-Fluoro-thymidine (18F-FLT) [51, 101] (Fig. 9.5). 18F-FLT is the subject of a multicenter investigational new drug (IND) application and the subject of many ongoing clinical trials in oncology. 18F-FLT is phosphorylated by thymidine kinase and trapped in proliferating cells. Although it is not incorporated into DNA in the relatively short timeframe for a PET study, it is a measure of thymidine uptake and phosphorylation. Tumor 18F-FLT uptake can therefore be used as a biomarker for tumor proliferation [102]. One of the key advantages of FLT over the more commonly used PET tracer, 18F-FDG is its tumor specificity [103–105]. Bioluminescence imaging has also been used to image tumor cell proliferation in the transgenic Ef-luc mouse model, where the gene encoding luciferase is controlled by the human E2F1 promoter and the BLI signal correlates not only with cell number but also the proliferative capacity of the tumor [106]. 9.4.2.3 Vascular Imaging, Permeability, and Blood Flow Vascular end points including blood flow, blood volume, and vascular permeability can be assessed in tumors using a variety of MRI-based techniques such as dynamic contrast-enhanced MRI (DCE MRI) [107, 108] (Fig. 9.6), arterial spin labeling [109, 110], and iron-oxide-based contrast MRI [111]. These methods have been widely used to detect response to increasingly prevalent anti-angiogenic [112–114] and vascular disruption [115] therapies in mouse tumor models [40, 116].

Fig. 9.5  (a) A T2-weighted MR image in an intracranial tumor model showing definition of the tumor boundary. (b) Corresponding FLT PET image in the same tumor coregistered with the same MRI image filtered to highlight the edges for anatomical localization. This example highlights the inherently high resolution of MRI for defining tumor boundaries, volume, and heterogeneity, while FLT PET provides an inherently quantitative functional assessment of the tumor proliferation. Future use of integrated PET/MR systems will facilitate greater utility in tumor imaging studies

228

V. Kaimal et al. 0.07

[Gd] in mmol/L

0.06 0.05 0.04 0.03 0.02 0.01 0.00 0

100

200

300

400

500

600

700

Time (seconds) Fig. 9.6  DCE MRI: A gadolinium concentration [Gd] time course, calculated from a T1-weighted image signal intensity time course after gadolinium bolus injection. Sample T1-weighted images are shown at four points along the [Gd] time course. The [Gd] time course is fitted to a kinetic model, enabling quantification of parameters sensitive to blood flow, vascular permeability, and vascular surface area, including Ktrans and initial area under the curve (IAUC)

Of these methods, DCE-MRI (Fig. 9.6) is the most clinically relevant and is a technique that employs rapid T1-weighted imaging coupled with the injection of a T1 modifying contrast agent such as Gd-DTPA in order to characterize blood flow, vascular surface area, and vessel permeability by the dynamics of tracer exchange between the vessels and the interstitial space [117, 118]. By manipulation of the tracer type and size, the sensitivity to flow vs. permeability can be manipulated [119, 120]. A variety of modeling approaches exist to generate meaningful end points that can quantify a vascular response [121]. With the introduction of microCT systems with rapid scanning capabilities, DCE imaging has also been used with CT contrast agents in preclinical models [122]. Advantages of CT-based DCE approaches include higher resolution than can be achieved through rapid MRI. In CT, tissue tracer concentration is more directly related to image signal than it is in MRI, where tissue relaxation times contribute. Small animal ultrasound [123] can also be used to assess tissue perfusion through dynamic uptake and clearance of microbubble contrast agents [73], and blood flow by Doppler methods [75]. Optical imaging techniques for dynamic contrast assessment of vasculature are also under development. Vessel size imaging [124, 125] can be used to quantify the distribution of microvasculature in tumors by using superparamagnetic iron-oxide contrast agents (SPIOs). SPIO-induced changes in tissue relaxivity have been shown to be proportional to the average vessel radius and can be used to calculate blood volume. Blood pooling contrast agents are also being used for generating vascular maps using MRI [126] and microCT [127], although commonly this is more easily accomplished in excised fixed tumor samples than in vivo where motion artifacts, contrast

9  Imaging Efficacy in Tumor Models

229

clearance, and sensitivity are limiting factors. Tumor blood flow measurements have also been performed using clinical PET imaging of [15O]-labeled water [128]. Arterial spin labeling (ASL) [129] is an MR imaging technique that can be used to quantify blood flow and tissue perfusion. ASL involves magnetically “tagging” water protons in the blood supply for a particular tissue and measuring their appearance downstream in the tissue of interest. ASL techniques have been used in ­preclinical models such as renal cell carcinoma [130] where high perfusion regions in the images were shown to correlate with viable tissue while regions of ­diminished perfusion was shown to be necrotic [129]. Targeted probes can also be used to assess the structure and integrity of the vasculature. One approach is to label a contrast agent with a vascular targeted antibody or peptide. For example, the integrin anb3 is expressed in angiogenic vessels and can be imaged by using probes that are coupled with anti-anb3 antibodies [131]. Such probes have been developed using both Gd and iron-oxide contrast agents [132]. Another target for imaging angiogenesis is vascular endothelial growth factor (VEGF), which is expressed by most solid tumors and plays a role in the formation of new vessels. PET has been used to image VEGF receptors by conjugating VEGF-121 with radioactive 64Cu [133–135]. Targeted optical imaging probes also exist for imaging vasculature [82]. These probes are distributed passively through blood vessels to image vascularity, blood pooling, and vascular leakage [136]. 9.4.2.4 Tumor Cellularity and Cell Kill Diffusion MRI (dMRI) is used in oncology applications for the early evaluation of tumor response to therapy. dMRI provides a measure of the apparent translational mobility of water in tissue, or apparent diffusion coefficient (ADC). The tissue water ADC is influenced by diffusion barriers such as cytoplasmic structures, organelles, cell membranes, and the extracellular matrix which break down in response to treatment. Several studies of different tumor types and models have shown that ADC is correlated with tissue cell density [137–140]. When a tumor responds to therapy, an early change in ADC can be observed, often before a measurable decrease in tumor volume [138, 139, 141]. Recent clinical trials suggest a positive prognostic value for early ADC change in brain tumors [142]. 9.4.2.5 Receptor Occupancy and Gene Expression Radio-labeled small molecule reporter systems can be used for direct imaging of molecular targets using PET or SPECT. For example, neuroendocrine tumors expressing somatostatin receptors (SSTRs) can be imaged using SST analogs like octreotide labeled with iodine (123I) using SPECT [143]. Radiolabeled antibodies and antibody fragments also enable receptor imaging. For example, the Her2/neu tyrosine kinase receptor is often expressed on the surface of breast cancer cells [144] and is a target for therapeutic agents such as Herceptin. Her2/neu expression

230

V. Kaimal et al.

and monitoring efficacy of therapy using PET has been accomplished using 68Ga, a positron emitter [145]. Herceptin has also been labeled with SPIO using a pretargeting approach and imaged using MRI [146] by exploiting the strong affinity of the biotin–streptavidin complex. In this work, the biotinylated antibody was first administered followed by streptavidin bound SPIO, leading to significant increase in the MRI relaxation times in Her2/neu expressing tumor tissue. 9.4.2.6 Tissue Hypoxia and pH Acidic and hypoxic microenvironments are typical of solid tumors, generally attributed to poor perfusion and elevated metabolic rates [147]. Hypoxia, in turn, drives processes such as angiogenesis-mediated by hypoxia-inducible factor-1 (HIF-1). MR techniques sensitive to blood or tissue oxygenation, e.g. BOLD (blood oxygen level dependant) contrast imaging [148] can be used to quantify hypoxia. However, BOLD images are also affected by tissue pH, the hematocrit and flow [149]. 19F MRI, following injection of perfluorocarbons (PFCs), has also been shown to be sensitive to pO2. This is due to the high solubility of molecular oxygen in PFCs, leading to T1 enhancement of the 19F nuclei [150]. Tissue hypoxia can be measured by the PET tracer 18F-Fluoromisonidazole 18 ( F-FMISO) [151] and a series of new analogs currently in development [152, 153]. This tracer is taken up by cells, where it is acted upon and trapped by nitroreductase enzymes expressed during hypoxia. Tissue PET signal is therefore a measure of hypoxia. Measurement of pHe (extracellular pH) and pHi (intracellular pH) is possible using various MR techniques including 31P MRS, and 1H MRS using imidazoles or aromatics [147, 154–158]. Optical techniques could also be used to image tissue oxygenation, potentially in tumors. For example, a transgenic mouse model, ROSA26 ODD-Luc/+ that ubiquitously expresses a bioluminescent reporter consisting of firefly luciferase fused to a region of HIF that is subject to oxygen-dependant degradation has been developed [159]. 9.4.2.7 Apoptosis Several in vivo approaches for imaging of apoptosis have been developed utilizing BLI [160], MRI [161–163], and PET/SPECT [164]. One technique using BLI makes use of tumor cells that have been transfected with a hybrid recombinant reporter consisting of luciferase linked to the estrogen receptor regulatory domain (ER) via a cleavage site for caspase-3 (DEVD) [160]. The presence of ER in the reporter renders the luciferase inactive. On activation of caspase-3 during apoptosis, the DEVD site is cleaved and the luciferase becomes active, signaling the onset of apoptosis in the presence of luciferin. Other luciferase-based approaches usually involving caspase 3 specificity have been reported [165]. Another approach to imaging apoptosis is to target phosphatidyl-serine (PS), which is normally present in the cytoplasmic leaflet of the cell membrane but is externalized to the outer

9  Imaging Efficacy in Tumor Models

231

l­eaflet in apoptotic cells [166]. Proteins like annexin-V bind with high affinity to PS. Therefore, labeling annexin-V with radioligands for PET/SPECT [160, 164, 167], fluorescent dyes [168], or MRI contrast agents such as SPIO [161–163] has been used to image the apoptotic cells.

9.4.2.8 Imaging of Oncogenic Pathways As drug discovery has evolved to be a target driven endeavor, strategies have been developed in parallel to enable imaging of target modulation, many based on approaches originally developed for in  vitro assays. Some of the imaging strategies already discussed demonstrate the use of molecular imaging reporters to enable quantitative imaging of cellular and molecular processes including receptor occupancy [143–146], apoptosis [160–166], and angiogenic targets [131–135]. However, there has been broader use of reporters that drive imaging signal conditionally based on modulation of oncogenic pathways, with many of the most prevalent oncology targets being the focus of image-based reporters [169]. For example, proximity assays based on fluorescence or bioluminescence resonance energy transfer (FRET/BRET) rely on protein interactions to bring the donor and acceptor fluorophores into close proximity [170]. Another approach is a two-hybrid system, in which interaction between proteins leads to reconstitution of an active transcription factor, leading to expression of a reporter gene [171–173]. In protein-­complementation assays (PCA), two fusion proteins are expressed, each carrying a fragment of the split reporter such as luciferase [174] or a fluorescent protein [175]. The interaction of these proteins leads to a functional reporter [176–180]. These approaches can enable imaging of modulation of targets such as Akt [181–183], EGFR [184], mTOR [185], RAS [186], Raf [187], and ERK/MEK [188].

9.4.2.9 Other Imaging Applications in Oncology Imaging proteolytic activity. A wide variety of diseases, including cancer and metastases formation, show an upregulation of proteases such as MMPs (matrix metalloproteinases) and excessive protein degradation [189]. Fluorescence-based imaging approaches have been used to quantify MMP activity [69]. Imaging of tumor stem cells. Tumor stem cells, also known as cancer initiating cells, are hypothesized to be a sub-population of solid tumor cells that are capable of repopulating an entire tumor, but with their progeny lacking this ability [190]. These cells can be identified based on an increasing number of cell surface markers using techniques such as fluorescence activated cell sorting (FACS). Emerging in  vivo fluorescence imaging techniques have been used to image these cells after genetically modifying the cells to express ZsGreen fluorescent protein [191].

232

V. Kaimal et al.

9.5 Imaging Efficacy in Oncology Models: Future Outlook Comprehensive imaging centers in pharmaceutical and contract research industry drug discovery programs have become prevalent only in the last 5 years. The increasingly rapid adoption levels and parallel optimization of image-based applications have moved preclinical imaging beyond the anatomical to routine use in quantifying drug function and tumor physiology. Oncology drug discovery now critically relies on image-based methodology for increasing the speed, sensitivity, and specificity in generation of endpoints determining drug efficacy. In parallel with this evolution, use of imaging is increasingly “modality independent.” With the introduction of extensive multimodality core imaging facilities in academic and teaching environments and more recently in industry, the imaging approach has evolved to be driven as much by the pharmacologists, biologists, clinicians, and chemists as by the imaging scientists themselves. This has provoked an endpoints based approach where imaging is used only if it can provide a tangible advantage over traditional methods. The modality or imaging protocol is chosen based on suitability and prior validation, rather than preferred use of a specific modality due to modality-centric expertise. Preclinical MRI, PET, SPECT, CT, optical, and ultrasound imaging technologies are now widely available, and underlying technologies have largely reached the point of stability and strong performance. Major advances are now focusing increasingly on new applications of these technologies and novel software to enable better automated and more quantitative endpoints. New software approaches for leveraging inherent image heterogeneities that have traditionally been viewed as confounding are under development and have been shown to increase sensitivity and specificity of image-based endpoints [192, 193]. New imaging probes and tracers across all modalities are increasing the spectrum of applications and the specificity of image-based approaches. Supporting hardware such as benchtop radiopharmaceutical production systems and multiple animal imaging approaches [194] are also headed toward wide use. PET/CT [195, 196] and SPECT/CT [197] integrated multimodality imaging systems that are now the preclinical as well as clinical standard for PET and SPECT have highlighted the advantage of integrated multimodality approaches. MRI/PET [198, 199] and MRI/SPECT [200] and integration of optical detectors with MRI, PET, SPECT, and CT comprise new approaches that will further potentiate imaging in oncology drug discovery and development.

References 1. Balaban RS, Hampshire VA. Challenges in small animal noninvasive imaging. ILAR J. 2001;42:248–62. 2. Riemann B, Schafers KP, Schober O, Schafers M. Small animal PET in preclinical studies: opportunities and challenges. Q J Nucl Med Mol Imaging. 2008;52:215–21. 3. Levin C, Zaidi H. Current trends in preclinical PET system design. PET Clin. 2007;2:125–60.

9  Imaging Efficacy in Tumor Models

233

4. Chatziioannou AF. Instrumentation for molecular imaging in preclinical research: microPET and Micro-SPECT. Proc Am Thor Soc. 2005;2:533–6, 10–11. 5. Renault G, Bonnin P, Marchiol-Fournigault C, et al. High-resolution ultrasound imaging of the mouse. J Radiol. 2006;87:1937–45. 6. Hu X, Norris DG. Advances in high-field magnetic resonance imaging. Annu Rev Biomed Eng. 2004;6:157–84. 7. Ritman EL. Micro-computed tomography-current status and developments. Annu Rev Biomed Eng. 2004;6:185–208. 8. McConville P, Moody JB, Moffat BA. High-throughput magnetic resonance imaging in mice for phenotyping and therapeutic evaluation. Curr Opin Chem Biol. 2005;9:413–20. 9. Haacke EM BR, Thompson MR, Venkatesan R. Magnetic resonance imaging. Physical principles and sequence design. New York: Wiley; 1999. 10. Rabi II, Zacharias JR, Millman S,Kusch P. A new method of measuring nuclear magnetic moments. Phys Rev. 1938:318. 11. Bloch FAH, Packard M. The nuclear induction experiment. Phys Rev. 1946;70:474. 12. Purcell EM, Torrey HC, Pound RV. Resonance absorption by nuclear magnetic moments in a solid. Phys Rev. 1946;69:37–38. 13. Damadian R. Tumor detection by nuclear magnetic resonance. Science. 1971;171: 1151–1153. 14. Lauterbur PC. Image formation by induced local interactions: examples employing NMR. Nature. 1973;242:190–191. 15. Stejskal EO, Tanner JE. Spin diffusion measurements: spin echoes in the presence of timedependent field gradient. J Chem Phys. 1965;42:288–92. 16. Thulborn KR, Waterton JC, Matthews PM, Radda GK. Oxygenation dependence of the transverse relaxation time of water protons in whole blood at high field. Biochim Biophys Acta. 1982;714:265–70. 17. Axel L. Blood flow effects in magnetic resonance imaging. Magn Reson Annu. 1986: 237–44. 18. Rieke V, Butts Pauly K. MR thermometry. J Magn Reson Imaging. 2008;27:376–90. 19. Muthupillai R, Ehman RL. Magnetic resonance elastography. Nat Med. 1996;2:601–3. 20. Howe FA, Maxwell RJ, Saunders DE, Brown MM, Griffiths JR. Proton spectroscopy in vivo. Magn Reson Q. 1993;9:31–59. 21. Mathur-De Vre R, Lemort M. Invited review: biophysical properties and clinical applications of magnetic resonance imaging contrast agents. Br J Radiol. 1995;68:225–47. 22. Sobol WT. Magnetic resonance contrast agents. Physical basis of relaxation. Neuroimaging Clin N Am. 1994;4:27–42. 23. Strijkers GJ, Mulder WJ, van Tilborg GA, Nicolay K. MRI contrast agents: current status and future perspectives. Anti-cancer Agents Med Chem. 2007;7:291–305. 24. Rudin M. Molecular imaging. 1st ed. Imprerial College Press: Covent Garden, London 2005. 25. Kim D, Hong KS, Song J. The present status of cell tracking methods in animal models using magnetic resonance imaging technology. Mol Cells. 2007;23:132–7. 26. Arbab AS, Liu W, Frank JA. Cellular magnetic resonance imaging: current status and future prospects. Expert Rev Med Devices. 2006;3:427–39. 27. Zech CJ, Reiser MF, Herrmann KA. Imaging of hepatocellular carcinoma by computed tomography and magnetic resonance imaging: state of the art. Dig Dis. 2009;27:114–24. 28. Spuentrup E, Botnar RM. Coronary magnetic resonance imaging: visualization of the vessel lumen and the vessel wall and molecular imaging of arteriothrombosis. Eur Radiol. 2006;16:1–14. 29. Epstein FH. MR in mouse models of cardiac disease. NMR Biomed. 2007;20:238–55. 30. Fishbein KW, McConville P, Spencer RG. The lever-coil: a simple, inexpensive sensor for respiratory and cardiac motion in MRI experiments. Magn Reson Imaging. 2001;19:881–9. 31. Filler A. The history, development and impact of computed imaging in neurological ­diagnosis and neurosurgery: CT, MRI, and DTI. Nature Precedings (2009). doi:10.1038/ npre.2009.3267.5.

234

V. Kaimal et al.

32. Scutellari PN, Antinolfi G, Galeotti R, Giganti M. Metastatic bone disease. Strategies for imaging. Miner Med. 2003;94:77–90. 33. Delorme S, Baur-Melnyk A. Imaging in multiple myeloma. Eur J Radiol. 2009;70:401–8. 34. Winkelmann CT, Figueroa SD, Rold TL, Volkert WA, Hoffman TJ. Microimaging characterization of a B16-F10 melanoma metastasis mouse model. Mol Imaging. 2006;5:105–14. 35. Badea C, Hedlund LW, Johnson GA. Micro-CT with respiratory and cardiac gating. Med. Phys. 2004;31:3324–9. 36. Willekens I, Lahoutte T, Buls N, et al. Time-course of contrast enhancement in spleen and liver with Exia 160, Fenestra LC, and VC. Mol Imaging Biol. 2009;11:128–35. 37. Ma X, Setty B, Uppot RN, Sahani DV. Multiple-detector computed tomographic angiography of pancreatic neoplasm for presurgical planning: comparison of low- and high-concentration nonionic contrast media. J Comput Assist Tomogr. 2008;32:511–7. 38. Dillman JR, Caoili EM, Cohan RH. Multi-detector CT urography: a one-stop renal and ­urinary tract imaging modality. Abdom Imaging. 2007;32:519–29. 39. Kim H, Chaudhuri TR, Buchsbaum DJ, Wang D, Zinn KR. High-resolution single-photon emission computed tomography and X-ray computed tomography imaging of Tc-99mlabeled anti-DR5 antibody in breast tumor xenografts. Mol Cancer Ther. 2007;6:866–75. 40. Contag CH. Bioluminescence imaging of mouse models of human cancer. In: Holland EC, editor. Mouse models of human cancer. Hoboken, New Jersey: Wiley; 2004. p. 363–73. 41. Won C, Chon D, Tajik J, et al. CT-based assessment of regional pulmonary microvascular blood flow parameters. J Appl Physiol. 2003;94:2483–93. 42. Chugh BP, Lerch JP, Yu LX, et al. Measurement of cerebral blood volume in mouse brain regions using micro-computed tomography. Neuroimage. 2009 Oct 1;47(4):1312–8. 43. Kamena A, Streitparth F, Grieser C, et al. Dynamic perfusion CT: optimizing the temporal resolution for the calculation of perfusion CT parameters in stroke patients. Eur J Radiol. 2007;64:111–8. 44. Sweet WH. The use of nuclear disintegration in the diagnosis and treatment of brain tumor. N Engl J Med. 1951;245:875–8. 45. Cutler PD, Cherry SR, Hoffman EJ, Digby WM, Phelps ME. Design features and performance of a PET system for animal research. J Nucl Med. 1992;33:595–604. 46. Chatziioannou A, Tai YC, Doshi N, Cherry SR. Detector development for microPET II: a 1 microl resolution PET scanner for small animal imaging. Phys Med Biol. 2001;46:2899–910. 47. Muehllehner G, Karp JS. Positron emission tomography. Phys Med Biol. 2006;51:R117–37. 48. Lartizien C, Kinahan PE, Swensson R, et al. Evaluating image reconstruction methods for tumor detection in 3-dimensional whole-body PET oncology imaging. J Nucl Med. 2003;44:276–90. 49. Sun H, Mangner TJ, Collins JM, Muzik O, Douglas K, Shields AF. Imaging DNA synthesis in vivo with 18F-FMAU and PET. J Nucl Med. 2005;46:292–6. 50. Spence AM, Mankoff DA, Muzi M. Positron emission tomography imaging of brain tumors. Neuroimaging Clin N Am. 2003;13:717–39. 51. Shields AF, Grierson JR, Dohmen BM, et al. Imaging proliferation in vivo with [F-18]FLT and positron emission tomography. Nat Med. 1998;4:1334–6. 52. Shields AF, Mankoff DA, Link JM, et al. Carbon-11-thymidine and FDG to measure therapy response. J Nucl Med. 1998;39:1757–62. 53. Jager PL, Que TH, Vaalburg W, Pruim J, Elsinga P, Plukker JT. Carbon-11 choline or FDGPET for staging of oesophageal cancer? Eur J Nucl Med. 2001;28:1845–9. 54. Jager PL, Vaalburg W, Pruim J, de Vries EG, Langen KJ, Piers DA. Radiolabeled amino acids: basic aspects and clinical applications in oncology. J Nucl Med. 2001;42:432–45. 55. Furumoto S, Takashima K, Kubota K, Ido T, Iwata R, Fukuda H. Tumor detection using 18 F-labeled matrix metalloproteinase-2 inhibitor. Nucl Med Biol. 2003;30:119–25. 56. Gambhir S, Herschman H, Cherry S, et al. Imaging transgene expression with radionuclide imaging technologies. Neoplasia. 2000;2:118–38. 57. Lee FT, Hall C, Rigopoulos A, et  al. Immuno-PET of human colon xenograft- bearing BALB/c nude mice using 124I-CDR-grafted humanized A33 monoclonal antibody. J Nucl Med. 2001;42:764–9.

9  Imaging Efficacy in Tumor Models

235

58. Sundaresan G, Yazaki PJ, Shively JE, et al. 124I-labeled engineered anti-CEA minibodies and diabodies allow high-contrast, antigen-specific small-animal PET imaging of xenografts in athymic mice. J Nucl Med. 2003;44:1962–9. 59. Aboagye EO. Positron emission tomography imaging of small animals in anticancer drug development. Mol Imaging Biol. 2005;7:53–8. 60. Rousset OG, Ma Y, Evans AC. Correction for partial volume effects in PET: principle and validation. J Nucl Med. 1998;39:904–11. 61. Edinger M, Cao YA, Hornig YS, et  al. Advancing animal models of neoplasia through in vivo bioluminescence imaging. Eur J Cancer. 2002;38:2128–36. 62. Contag CH, Bachmann MH. Advances in in vivo bioluminescence imaging of gene expression. Annu Rev Biomed Eng. 2002;4:235–60. 63. Dothager RS, Flentie K, Moss B, Pan MH, Kesarwala A, Piwnica-Worms D. Advances in bioluminescence imaging of live animal models. Curr Opin Biotechnol. 2009;20:45–53. 64. Lim E, Modi KD, Kim J. In vivo bioluminescent imaging of mammary tumors using IVIS  spectrum. J Vis Exp. 26. 2009 http://www.jove.com/index/details.stp?id=1210, doi: 10.3791/1210. 65. Haustein E, Schwille P. Trends in fluorescence imaging and related techniques to unravel biological information. HFSP J. 2007;1:169–80. 66. Hall D, Ma G, Lesage F, Wang Y. Simple time-domain optical method for estimating the depth and concentration of a fluorescent inclusion in a turbid medium. Opt Lett. 2004;29:2258–60. 67. Jaffer FA, Weissleder R. Molecular imaging in the clinical arena. JAMA. 2005;293: 855–62. 68. Zhang W, Feng JQ, Harris SE, Contag PR, Stevenson DK, Contag CH. Rapid in vivo functional analysis of transgenes in mice using whole body imaging of luciferase expression. Transgenic Res. 2001;10:423–34. 69. Weissleder R, Tung CH, Mahmood U, Bogdanov A, Jr. In vivo imaging of tumors with protease-activated near-infrared fluorescent probes. Nat Biotechnol. 1999;17:375–8. 70. Foster FS, Burns PN, Simpson DH, Wilson SR, Christopher DA, Goertz DE. Ultrasound for the visualization and quantification of tumor microcirculation. Cancer Metastasis Rev. 2000;19:131–8. 71. Schutt EG, Klein DH, Mattrey RM, Riess JG. Injectable microbubbles as contrast agents for diagnostic ultrasound imaging: the key role of perfluorochemicals. Angew Chem Int Ed. 2003;42:3218–35. 72. Alkan-Onyuksel H, Demos SM, Lanza GM, et  al. Development of inherently echogenic liposomes as an ultrasonic contrast agent. J Pharm Sci. 1996;85:486–90. 73. Needles A, Goertz DE, Karshafian R, et  al. High-frequency subharmonic pulsed-wave Doppler and color flow imaging of microbubble contrast agents. Ultrasound Med Biol. 2008;34:1139–51. 74. Hoskins PR. Measurement of arterial blood flow by Doppler ultrasound. Clin Phys Physiol Meas. 1990;11:1–26. 75. Christopher DA, Burns PN, Starkoski BG, Foster FS. A high-frequency pulsed-wave Doppler ultrasound system for the detection and imaging of blood flow in the microcirculation. Ultrasound Med Biol. 1997;23:997–1015. 76. Bruce M, Averkiou M, Tiemann K, Lohmaier S, Powers J, Beach K. Vascular flow and perfusion imaging with ultrasound contrast agents. Ultrasound Med Biol. 2004;30: 735–43. 77. McConville P, Hambardzumyan D, Moody JB, et al. Magnetic resonance imaging determination of tumor grade and early response to temozolomide in a genetically engineered mouse model of glioma. Clin Cancer Res. 2007;13:2897–904. 78. Ross BD, Chenevert TL, Rehemtulla A. Magnetic resonance imaging in cancer research. Eur J Cancer. 2002;38:2147–56. 79. Silva AC, Evans JM, McCullough AE, Jatoi MA, Vargas HE, Hara AK. MR imaging of hypervascular liver masses: a review of current techniques. Radiographics. 2009;29:385–402.

236

V. Kaimal et al.

80. Beuf O, Lartizien C, Milot L, et al. Multimodal imaging for the detection and characterization of liver lesions in a mouse model of neuroendocrine tumor. Gastroenterol Clin Biol. 2008;32:32–40. 81. Martiniova L, Kotys MS, Thomasson D, et al. Noninvasive monitoring of a murine model of metastatic pheochromocytoma: a comparison of contrast-enhanced microCT and nonenhanced MRI. J Magn Reson Imaging. 2009;29:685–91. 82. Montet X, Ntziachristos V, Grimm J, Weissleder R. Tomographic fluorescence mapping of tumor targets. Cancer Res. 2005;65:6330–6. 83. Brader P, Riedl CC, Woo Y, et al. Imaging of hypoxia-driven gene expression in an orthotopic liver tumor model. Mol Cancer Ther. 2007;6:2900–8. 84. Gee MS, Upadhyay R, Bergquist H, et al. Human breast cancer tumor models: molecular imaging of drug susceptibility and dosing during HER2/neu-targeted therapy. Radiology. 2008;248:925–35. 85. Miretti S, Roato I, Taulli R, et al. A mouse model of pulmonary metastasis from spontaneous osteosarcoma monitored in vivo by luciferase imaging. PloS One. 2008;3:e1828. 86. 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;15:238–46. 87. Quaglino E, Mastini C, Forni G, Cavallo F. ErbB2 transgenic mice: a tool for investigation of the immune prevention and treatment of mammary carcinomas. Curr Protoc Immunol. 2008 Aug;Chapter 20:Unit 20.9.1-20.9-10. 88. Jenkins DE, Oei Y, Hornig YS, et al. Bioluminescent imaging (BLI) to improve and refine traditional murine models of tumor growth and metastasis. Clin Exp Metastasis. 2003;20:733–44. 89. Park YJ, Nah SH, Lee JY, et al. Surface-modified poly(lactide-co-glycolide) nanospheres for targeted bone imaging with enhanced labeling and delivery of radioisotope. J Biomed Mater Res. 2003;67:751–60. 90. Ballou B, Ernst LA, Waggoner AS. Fluorescence imaging of tumors in  vivo. Curr Med Chem. 2005;12:795–805. 91. Choy G, Choyke P, Libutti SK. Current advances in molecular imaging: noninvasive in vivo bioluminescent and fluorescent optical imaging in cancer research. Mol Imaging. 2003;2: 303–12. 92. Choy G, O’Connor S, Diehn FE, et al. Comparison of noninvasive fluorescent and bioluminescent small animal optical imaging. Biotechniques. 2003;35:1022–6, 8–30. 93. Phelps ME. Inaugural article: positron emission tomography provides molecular imaging of biological processes. Proc Natl Acad Sci USA. 2000;97:9226–33. 94. Hicks RJ. Beyond FDG: novel PET tracers for cancer imaging. Cancer Imaging. 2004;4:22–4. 95. Fueger BJ, Czernin J, Hildebrandt I, et al. Impact of animal handling on the results of 18FFDG PET studies in mice. J Nucl Med. 2006;47:999–1006. 96. Kovar JL, Volcheck W, Sevick-Muraca E, Simpson MA, Olive DM. Characterization and performance of a near-infrared 2-deoxyglucose optical imaging agent for mouse cancer models. Anal Biochem. 2009;384:254–62. 97. Cheng Z, Levi J, Xiong Z, et al. Near-infrared fluorescent deoxyglucose analogue for tumor optical imaging in cell culture and living mice. Bioconjug Chem. 2006;17:662–9. 98. Lindskog M, Kogner P, Ponthan F, et  al. Noninvasive estimation of tumour viability in a xenograft model of human neuroblastoma with proton magnetic resonance spectroscopy (1H MRS). Br J Cancer. 2003;88:478–85. 99. Maxwell RJ, Nielsen FU, Breidahl T, Stodkilde-Jorgensen H, Horsman MR. Effects of ­combretastatin on murine tumours monitored by 31P MRS, 1H MRS and 1H MRI. Int J Radiat Oncol Biol Phys. 1998;42:891–4. 100. Kristensen CA, Askenasy N, Jain RK, Koretsky AP. Creatine and cyclocreatine treatment of human colon adenocarcinoma xenografts: 31P and 1H magnetic resonance spectroscopic studies. Br J Cancer. 1999;79:278–85.

9  Imaging Efficacy in Tumor Models

237

101. Been LB, Suurmeijer AJ, Cobben DC, Jager PL, Hoekstra HJ, Elsinga PH. [18F]FLT-PET in oncology: current status and opportunities. Eur J Nucl Med Mol Imaging. 2004;31:1659–72. 102. Salskov A, Tammisetti VS, Grierson J, Vesselle H. FLT: measuring tumor cell proliferation in  vivo with positron emission tomography and 3¢-deoxy-3¢-[18F]fluorothymidine. Semin Nucl Med. 2007;37:429–39. 103. Murayama C, Harada N, Kakiuchi T, et al. Evaluation of D-18F-FMT, 18F-FDG, L-11C-MET, and 18F-FLT for monitoring the response of tumors to radiotherapy in mice. J Nucl Med. 2009;50:290–5. 104. Pan MH, Huang SC, Liao YP, et  al. FLT-PET imaging of radiation responses in murine tumors. Mol Imaging Biol. 2008;10:325–34. 105. Yang YJ, Ryu JS, Kim SY, et al. Use of 3¢-deoxy-3¢-[18F]fluorothymidine PET to monitor early responses to radiation therapy in murine SCCVII tumors. Eur J Nucl Med Mol Imaging. 2006;33:412–9. 106. Uhrbom L, Nerio E, Holland EC. Dissecting tumor maintenance requirements using ­bioluminescence imaging of cell proliferation in a mouse glioma model. Nat Med. 2004;10:1257–60. 107. Taylor JS, Tofts PS, Port R, et al. MR imaging of tumor microcirculation: promise for the new millennium. J Magn Reson Imaging. 1999;10:903–7. 108. Leach MO, Brindle KM, Evelhoch JL, et al. Assessment of antiangiogenic and antivascular therapeutics using MRI: recommendations for appropriate methodology for clinical trials. Br J Radiol. 2003;76 Spec No 1:S87–91. 109. Moffat BA, Chenevert TL, Hall DE, Rehemtulla A, Ross BD. Continuous arterial spin labeling using a train of adiabatic inversion pulses. J Magn Reson Imaging. 2005;21:290–6. 110. Sun Y, Schmidt NO, Schmidt K, et al. Perfusion MRI of U87 brain tumors in a mouse model. Magn Reson Med. 2004;51:893–9. 111. Moffat BA, Reddy GR, McConville P, et al. A novel polyacrylamide magnetic nanoparticle contrast agent for molecular imaging using MRI. Mol Imaging. 2003;2:324–32. 112. Zwick S, Strecker R, Kiselev V, et al. Assessment of vascular remodeling under antiangiogenic therapy using DCE-MRI and vessel size imaging. J Magn Reson Imaging. 2009;29:1125–33. 113. Robinson SP, McIntyre DJ, Checkley D, et  al. Tumour dose response to the antivascular agent ZD6126 assessed by magnetic resonance imaging. Br J Cancer. 2003;88:1592–7. 114. Rosen MA, Schnall MD. Dynamic contrast-enhanced magnetic resonance imaging for assessing tumor vascularity and vascular effects of targeted therapies in renal cell carcinoma. Clin Cancer Res. 2007;13:770s-6s. 115. Nielsen T, Murata R, Maxwell RJ, Stodkilde-Jorgensen H, Ostergaard L, Horsman MR. Preclinical studies to predict efficacy of vascular changes induced by combretastatin a-4 disodium phosphate in patients. Int J Radiat Oncol Biol Phys. 2008;70:859–66. 116. Miller JC, Pien HH, Sahani D, Sorensen AG, Thrall JH. Imaging angiogenesis: applications and potential for drug development. J Natl Cancer Inst. 2005;97:172–87. 117. Tofts PS, Brix G, Buckley DL, et al. Estimating kinetic parameters from dynamic contrastenhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging. 1999;10:223–32. 118. Cheng HL. Dynamic contrast-enhanced MRI in oncology drug development. Curr Clin Pharmacol. 2007;2:111–22. 119. Feng Y, Jeong EK, Mohs AM, Emerson L, Lu ZR. Characterization of tumor angiogenesis with dynamic contrast-enhanced MRI and biodegradable macromolecular contrast agents in mice. Magn Reson Med. 2008;60:1347–52. 120. Wu X, Feng Y, Jeong EK, Emerson L, Lu ZR. Tumor characterization with dynamic contrast enhanced magnetic resonance imaging and biodegradable macromolecular contrast agents in mice. Pharm Res. 2009. 121. McGrath DM, Bradley DP, Tessier JL, Lacey T, Taylor CJ, Parker GJ. Comparison of modelbased arterial input functions for dynamic contrast-enhanced MRI in tumor bearing rats. Magn Reson Med. 2009;61:1173–84.

238

V. Kaimal et al.

122. Raatschen HJ, Fu Y, Brasch RC, Pietsch H, Shames DM, Yeh BM. In vivo monitoring of angiogenesis inhibitory treatment effects by dynamic contrast-enhanced computed tomography in a xenograft tumor model. Invest Radiol. 2009;44:265–70. 123. Liang HD, Blomley MJ. The role of ultrasound in molecular imaging. Br J Radiol. 2003;76 Spec No 2:S140–50. 124. Dennie J, Mandeville JB, Boxerman JL, Packard SD, Rosen BR, Weisskoff RM. NMR imaging of changes in vascular morphology due to tumor angiogenesis. Magn Reson Med. 1998;40:793–9. 125. Jensen JH, Chandra R. MR imaging of microvasculature. Magn Reson Med. 2000;44: 224–30. 126. Handelsman H. Magnetic resonance angiography: vascular and flow imaging. Health Technol Assess. 1994;(3):1–20. 127. Collins R, Cranny G, Burch J, et  al. A systematic review of duplex ultrasound, magnetic resonance angiography and computed tomography angiography for the diagnosis and assessment of symptomatic, lower limb peripheral arterial disease. Health Technol Assess. 2007;11(iii-iv, xi-xiii): 1–184. 128. Anderson H, Price P. Clinical measurement of blood flow in tumours using positron emission tomography: a review. Nucl Med Commun. 2002;23(2):131–138 2002. 129. Petersen ET, Zimine I, Ho YC, Golay X. Non-invasive measurement of perfusion: a critical review of arterial spin labelling techniques. Br J Radiol. 2006;79:688–701. 130. Pedrosa I, Alsop DC, Rofsky NM. Magnetic resonance imaging as a biomarker in renal cell carcinoma. Cancer. 2009;115:2334–45. 131. Montet X, Montet-Abou K, Reynolds F, Weissleder R, Josephson L. Nanoparticle imaging of integrins on tumor cells. Neoplasia. 2006;8:214–22. 132. Cai W, Rao J, Gambhir SS, Chen X. How molecular imaging is speeding up antiangiogenic drug development. Mol Cancer Ther. 2006;5:2624–33. 133. Cai W, Chen K, Mohamedali KA, et al. PET of vascular endothelial growth factor receptor expression. J Nucl Med. 2006;47:2048–56. 134. Chen K, Cai W, Li ZB, Wang H, Chen X. Quantitative PET imaging of VEGF receptor expression. Mol Imaging Biol. 2009;11:15–22. 135. Wang H, Cai W, Chen K, et al. A new PET tracer specific for vascular endothelial growth factor receptor 2. Eur J Nucl Med Mol Imaging. 2007;34:2001–10. 136. Choi Y, Weissleder R, Tung CH. Selective antitumor effect of novel protease-mediated ­photodynamic agent. Cancer Res. 2006;66:7225–9. 137. Chenevert TL, Stegman LD, Taylor JMG, et al. Diffusion magnetic resonance imaging: an early surrogate marker of therapeutic efficacy in brain tumors. J Natl Cancer Inst. 2000;92: 2029–36. 138. Chenevert TL, McKeever PE, Ross BD. Monitoring early response of experimental brain tumors to therapy using diffusion magnetic resonance imaging. Clin Cancer Res. 1997;3:1457–66. 139. Lyng H, Haraldseth O, Rofstad EK. Measurement of cell density and necrotic fraction in human melanoma xenografts by diffusion weighted magnetic resonance imaging. Magn Reson Med. 2000;43:828–36. 140. Sugahara T, Korogi Y, Kochi M, et  al. Usefulness of diffusion-weighted MRI with echoplanar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging. 1999;9:53–60. 141. Ross BD, Moffat BA, Lawrence TS, et  al. Evaluation of cancer therapy using diffusion magnetic resonance imaging. Mol Cancer Ther. 2003;2:581–7. 142. Moffat BA, Chenevert TL, Lawrence T, et al. Functional Diffusion Map: a noninvasive MRI biomarker for early stratification of clinical brain tumor response. Proc Natl Acad Sci USA. 2005;102(15):5524–9. Epub 2005 Apr 1. 143. Olias G, Viollet C, Kusserow H, Epelbaum J, Meyerhof W. Regulation and function of somatostatin receptors. J Neurochem. 2004;89:1057–91. 144. Slamon DJ, Godolphin W, Jones LA, et  al. Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science. 1989;244:707–12.

9  Imaging Efficacy in Tumor Models

239

145. Smith-Jones PM, Solit DB, Akhurst T, Afroze F, Rosen N, Larson SM. Imaging the pharmacodynamics of HER2 degradation in response to Hsp90 inhibitors. Nat Biotechnol. 2004; 22:701–6. 146. Artemov D, Mori N, Okollie B, Bhujwalla ZM. MR molecular imaging of the Her-2/neu receptor in breast cancer cells using targeted iron oxide nanoparticles. Magn Reson Med. 2003;49:403–8. 147. Gillies RJ, Raghunand N, Karczmar GS, Bhujwalla ZM. MRI of the tumor microenvironment. J Magn Reson Imaging. 2002;16:430–50. 148. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci USA. 1990;87:9868–72. 149. Howe FA, Robinson SP, Rodrigues LM, Griffiths JR. Flow and oxygenation dependent (FLOOD) contrast MR imaging to monitor the response of rat tumors to carbogen breathing. Magn Reson Imaging. 1999;17:1307–18. 150. Parhami P, BM F. Fluorine-19 relaxation study of perfluoro chemicals as oxygen carriers. J Physiol Chem. 1983;87:1928–31. 151. Bentzen L, Keiding S, Horsman MR, Falborg L, Hansen SB, Overgaard J. Feasibility of detecting hypoxia in experimental mouse tumours with 18F-fluorinated tracers and positron emission tomography-a study evaluating [18F]Fluoro-2-deoxy-D-glucose. Acta Oncol. 2000;39:629–37. 152. Dunphy MP, Lewis JS. Radiopharmaceuticals in preclinical and clinical development for monitoring of therapy with PET. J Nucl Med. 2009;50 Suppl 1:106S–21S. 153. Kumar R, Dhanpathi H, Basu S, Rubello D, Fanti S, Alavi A. Oncologic PET tracers beyond [(18)F]FDG and the novel quantitative approaches in PET imaging. Q J Nucl Med Mol Imaging. 2008;52:50–65. 154. Gillies RJ, Raghunand N, Garcia-Martin ML, Gatenby RA. pH imaging. A review of pH  measurement methods and applications in cancers. IEEE Eng Med Biol Mag. 2004;23:57–64. 155. Bhujwalla ZM, Artemov D, Ballesteros P, Cerdan S, Gillies RJ, Solaiyappan M. Combined vascular and extracellular pH imaging of solid tumors. NMR Biomed. 2002;15:114–9. 156. Evelhoch JL, Gillies RJ, Karczmar GS, et al. Applications of magnetic resonance in model systems: cancer therapeutics. Neoplasia. 2000;2:152–65. 157. Gillies RJ, Bhujwalla ZM, Evelhoch J, et al. Applications of magnetic resonance in model systems: tumor biology and physiology. Neoplasia. 2000;2:139–51. 158. van Sluis R, Bhujwalla ZM, Raghunand N, et al. In vivo imaging of extracellular pH using 1H MRSI. Magn Reson Med. 1999;41:743–50. 159. Safran M, Kim WY, O’Connell F, et al. Mouse model for noninvasive imaging of HIF prolyl hydroxylase activity: assessment of an oral agent that stimulates erythropoietin production. Proc Natl Acad Sci USA. 2006;103:105–10. 160. Laxman B, Hall DE, Bhojani MS, et al. Noninvasive real-time imaging of apoptosis. Proc Natl Acad Sci USA. 2002;99:16551–5. 161. Jung HI, Kettunen MI, Davletov B, Brindle KM. Detection of apoptosis using the C2A domain of synaptotagmin I. Bioconjug Chem. 2004;15:983–7. 162. Schellenberger EA, Bogdanov A, Jr., Hogemann D, Tait J, Weissleder R, Josephson L. Annexin V-CLIO: a nanoparticle for detecting apoptosis by MRI. Mol Imaging. 2002;1:102–7. 163. Zhao M, Beauregard DA, Loizou L, Davletov B, Brindle KM. Non-invasive detection of apoptosis using magnetic resonance imaging and a targeted contrast agent. Nat Med. 2001;7:1241–4. 164. Lahorte CM, Vanderheyden JL, Steinmetz N, Van de Wiele C, Dierckx RA, Slegers G. Apoptosis-detecting radioligands: current state of the art and future perspectives. Eur J Nucl Med Mol Imaging. 2004;31:887–919. 165. Ray P, De A, Patel M, Gambhir SS. Monitoring caspase-3 activation with a multimodality imaging sensor in living subjects. Clin Cancer Res. 2008;14:5801–9. 166. Martin SJ, Reutelingsperger CP, McGahon AJ, et  al. Early redistribution of plasma membrane phosphatidylserine is a general feature of apoptosis regardless of the initiating stimulus: inhibition by overexpression of Bcl-2 and Abl. J Exp Med. 1995;182:1545–56.

240

V. Kaimal et al.

167. Blankenberg FG, Vanderheyden JL, Strauss HW, Tait JF. Radiolabeling of HYNIC-annexin V with technetium-99m for in vivo imaging of apoptosis. Nat Protocols. 2006;1:108–10. 168. Petrovsky A, Schellenberger E, Josephson L, Weissleder R, Bogdanov A, Jr. Near-infrared fluorescent imaging of tumor apoptosis. Cancer Res. 2003;63:1936–42. 169. Villalobos V, Naik S, Piwnica-Worms D. Current state of imaging protein–protein interactions in vivo with genetically encoded reporters. Annu Rev Biomed Eng. 2007;9:321–49. 170. Wang Y, Wang G, O’Kane DJ, Szalay AA. A study of protein–protein interactions in living cells using luminescence resonance energy transfer (LRET) from Renilla luciferase to Aequorea GFP. Mol Gen Genet. 2001;264:578–87. 171. Luker GD, Luker KE. Optical imaging: current applications and future directions. J Nucl Med. 2008;49:1–4. 172. Tkaczyk ER, Zhong CF, Ye JY, et al. In vivo monitoring of multiple circulating cell populations using two-photon flow cytometry. Opt Commun. 2008;281:888–94. 173. Luker GD, Sharma V, Pica CM, Prior JL, Li W, Piwnica-Worms D. Molecular imaging of protein–protein interactions: controlled expression of p53 and large T-antigen fusion proteins in vivo. Cancer Res. 2003;63:1780–8. 174. Ozawa T, Kaihara A, Sato M, Tachihara K, Umezawa Y. Split luciferase as an optical probe for detecting protein–protein interactions in mammalian cells based on protein splicing. Anal Chem. 2001;73:2516–21. 175. Ozawa T, Nogami S, Sato M, Ohya Y, Umezawa Y. A fluorescent indicator for detecting protein–protein interactions in vivo based on protein splicing. Anal Chem. 2000;72:5151–7. 176. Michnick SW, Ear PH, Manderson EN, Remy I, Stefan E. Universal strategies in research and drug discovery based on protein-fragment complementation assays. Nat Rev. 2007;6:569–82. 177. Paulmurugan R, Gambhir SS. Monitoring protein–protein interactions using split synthetic renilla luciferase protein-fragment-assisted complementation. Anal. Chem. 2003;75:1584–9. 178. Paulmurugan R, Umezawa Y, Gambhir SS. Noninvasive imaging of protein–protein interactions in living subjects by using reporter protein complementation and reconstitution strategies. Proc Natl Acad Sci USA. 2002;99:15608–13. 179. Ray P, Pimenta H, Paulmurugan R, et al. Noninvasive quantitative imaging of protein–protein interactions in living subjects. Proc Natl Acad Sci USA. 2002;99:3105–10. 180. Westwick JK, Michnick SW. Protein-fragment complementation assays (PCA) in small GTPase research and drug discovery. Methods Enzymol. 2006;407:388–401. 181. Zhang L, Bhojani MS, Ross BD, Rehemtulla A. Molecular imaging of protein kinases. Cell Cycle. 2008;7:314–7. 182. Zhang L, Lee KC, Bhojani MS, et al. Molecular imaging of Akt kinase activity. Nat Med. 2007;13:1114–9. 183. Zhang L, Bhojani MS, Ross BD, Rehemtulla A. Enhancing Akt imaging through targeted reporter expression. Mol Imaging. 2008;7:168–74. 184. Li W, Li F, Huang Q, Frederick B, Bao S, Li CY. Noninvasive imaging and quantification of epidermal growth factor receptor kinase activation in vivo. Cancer Res. 2008;68:4990–7. 185. Paulmurugan R, Massoud TF, Huang J, Gambhir SS. Molecular imaging of drug-modulated protein–protein interactions in living subjects. Cancer Res. 2004;64:2113–9. 186. Tang Y, Kim M, Carrasco D, Kung AL, Chin L, Weissleder R. In vivo assessment of RASdependent maintenance of tumor angiogenesis by real-time magnetic resonance imaging. Cancer Res. 2005;65:8324–30. 187. Dangi-Garimella S, Yun J, Eves EM, et al. Raf kinase inhibitory protein suppresses a metastasis signalling cascade involving LIN28 and let-7. EMBO J. 2009;28:347–58. 188. Konopleva M, Shi Y, Steelman LS, et  al. Development of a conditional in  vivo model to evaluate the efficacy of small molecule inhibitors for the treatment of Raf-transformed hematopoietic cells. Cancer Res. 2005;65:9962–70. 189. Rawlings ND, O’Brien E, Barrett AJ. MEROPS: the protease database. Nucleic Acids Res. 2002;30:343–6.

9  Imaging Efficacy in Tumor Models

241

190. Reya T, Morrison SJ, Clarke MF, Weissman IL. Stem cells, cancer, and cancer stem cells. Nature. 2001;414:105–11. 191. Vlashi E, Kim K, Lagadec C, et al. In vivo imaging, tracking, and targeting of cancer stem cells. J Natl Cancer Inst. 2009;101:350–9. 192. Hamstra DA, Chenevert TL, Moffat BA, et al. Evaluation of the functional diffusion map as an early biomarker of time-to-progression and overall survival in high-grade glioma. Proc Natl Acad Sci USA. 2005;102:16759–64. 193. Moffat BA, Chenevert TL, Lawrence TS, et al. Functional diffusion map: a noninvasive MRI biomarker for early stratification of clinical brain tumor response. Proc Natl Acad Sci USA. 2005;102:5524–9. 194. Bock NA, Konyer NB, Henkelman RM. Multiple-mouse MRI. Magn Reson Med. 2003;49:158–67. 195. Pichler BJ, Judenhofer MS, Pfannenberg C. Multimodal imaging approaches: PET/CT and PET/MRI. Handb Exp Pharmacol. 2008;(185 Pt 1):109–32. 196. Mawlawi O, Townsend DW. Multimodality imaging: an update on PET/CT technology. Eur J Nucl Med Mol Imaging. 2009;36 Suppl 1:S15–29. 197. Townsend DW. Multimodality imaging of structure and function. Phys Med Biol. 2008;53:R1–39. 198. Schmidt GP, Kramer H, Reiser MF, Glaser C. Whole-body magnetic resonance imaging and positron emission tomography-computed tomography in oncology. Top Magn Reson Imaging. 2007;18:193–202. 199. Antoch G, Bockisch A. Combined PET/MRI: a new dimension in whole-body oncology imaging? Eur J Nucl Med Mol Imaging. 2009;36 Suppl 1:S113–20. 200. Wagenaar DJ, Kapusta M, Li J, Patt BE. Rationale for the combination of nuclear medicine with magnetic resonance for pre-clinical imaging. Technol Cancer Res Treat. 2006;5:343–50.

Part IV

Carcinogen-Induced Tumors

Chapter 10

Mammary Cancer in Rats Henry J. Thompson

Abstract  A need exists for a useful, practical, model for breast cancer not only to rapidly produce invasive, autochthonous tumors that can be used for screening of new compounds for treatment of advanced disease, but also for evaluation of candidate agents for the prevention of breast cancer. Such a model exists, is easy to set up in the laboratory, relatively inexpensive in that it requires only a single dose of a carcinogen, reproducible when adequately powered statistically, and relevant to human disease. This chapter will focus on this rat model for breast cancer induced by injection of the carcinogen, 1-methyl-1-nitrosourea, as the initiating agent. Keywords  Mammary carcinogenesis • 1-Methyl-1-nitrosourea • Model for breast cancer • Rat

10.1 Introduction Human breast cancer is a very heterogeneous disease and presents itself in many forms [1]. As recently reviewed, one important lesson derived from the use of animal models is that each model serves to illuminate some aspect of breast cancer and mimics a subset of the many subsets of the human disease [1, 2]. Thus, many different models will be needed to fully enable the understanding of the cellular and molecular basis for breast cancer and eventually provide the critical preventive and therapeutic approaches being developed to conquer this cancer [1, 3–8]. The 1-methyl-1-nitrosurea (MNU)-induced rat model for breast cancer is a useful, practical, model that has been deployed not only to produce invasive, autochthonous tumors that can be used for screening of new compounds for treatment of advanced disease, but also for evaluation of candidate agents for prevention of breast cancer.

H.J. Thompson (*) Cancer Prevention Laboratory, Colorado State University, Fort Collins, CO, 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_10, © Springer Science+Business Media, LLC 2011

245

246

H.J. Thompson

This latter approach to cancer control has seen major clinical advances, with the demonstration of the clinical efficacy of agents such as tamoxifen, raloxifene, and fenretinide, in chemoprevention of breast cancer in selected groups of women at high risk for development of the disease [9–13]. These experimental clinical trials in chemoprevention have been extremely costly to perform, since they have involved the administration of the above drugs to thousands of women over many years. It is therefore of the utmost importance to be certain that there is a strong pre-clinical rationale for the use of a new chemopreventive agent before it is introduced into clinical trials. In addition, negative outcome from the large clinical trials that attempted to use b-carotene to prevent cancer before there was strong evidence that this compound was truly an effective chemopreventive agent in experimental animals, underscores the importance of pre-clinical animal studies prior to clinical testing. The above considerations serve to highlight the importance of having a useful experimental model for breast cancer. It is indeed fortunate that such a model exists, and that it is easy to set up in the laboratory, relatively inexpensive as it requires only a single dose of a carcinogen, reproducible when adequately powered, and above all, relevant to human disease. This last criterion has been demonstrated by the experimental data which have shown that all three of the chemopreventive agents that have been found to be clinically effective in preventing breast cancer were first shown to be highly active agents in the animal model [9, 14, 15] which uses the carcinogen, MNU, as the initiating agent. This chapter will focus on the use of this model.

10.2 Historical Perspective Modern studies of mammary carcinogenesis in rats began with the introduction of the use of polynuclear hydrocarbons as carcinogens, most notably in the pioneering work of Huggins et al. [16]. In a classic paper published in 1961, data were reported on the effective use of 3-methylcholanthrene (3-MC) and 7,12-dimethylbenz(a) anthracene (DMBA). Both agents were shown to be potent mammary carcinogens when administered orally at appropriate single doses to Sprague–Dawley rats; tumor incidence of 100% was readily attained. However, although some of the mammary tumors induced by DMBA are indeed carcinomas, most of the tumors developing in rats treated with DMBA or 3-MC are benign, and should be classified as fibroadenomas [17]. Moreover, the mammary carcinomas induced by these polynuclear hydrocarbons in rats do not metastasize. These important limitations on the usefulness of the DMBA model therefore led to a search for a more effective mammary carcinogen that would induce tumors that are almost exclusively carcinomas. This search culminated with the introduction of MNU in 1975, in a classic paper by Gullino et al. [18]. Although the model, as presently used, has been modified to make it simpler and easier to implement, its basic elements remain unchanged, and it remains the system of choice at the present time. Unless one is specifically

10  Mammary Cancer in Rats

247

i­nterested in the effects of carcinogenic hydrocarbons on the mammary gland, i.e., modulation of the initiation of carcinogenesis­, there is essentially no good justification for the continued use of DMBA to induce mammary cancer in rats. The original system proposed by Gullino et al. used three intravenous injections of MNU, each at a dose of 50 mg/kg, given 4 weeks apart, beginning at 50 days of age. Inbred BUF/N rats were the strain of choice (mean latent period, 77 days), although the authors also showed high tumor yields in Sprague–Dawley and F344 rats (mean latent periods, 86 and 94 days, respectively). In sharp contrast to the tumors induced by DMBA, all of the 60 tumors from rats treated with MNU that were examined histologically were adenocarcinomas or papillary carcinomas. Rats bearing tumors induced by MNU were hypercalcemic, and bony metastasis was seen in some animals. Tumors were transplantable, and sensitive to estrogen. In several hundred animals with primary mammary carcinomas induced by MNU, the authors did not find tumors of other organs growing to a macroscopic size. The need for a total of three injections of MNU was not evaluated in this initial paper. As other investigators became interested in using this model for chemoprevention studies, it became important to shorten the time required for initiation of carcinogenesis, thus allowing a wider window for evaluation of the anti-promoting activity of new agents. A first modification was to eliminate the third intravenous dosing with MNU, and also to give the second dose 1 week (rather than 1 month) after the first [19]. Then it was shown that a single intravenous dose of MNU was just as effective as two doses [20]. Finally, to make the model even easier to use, it was shown that intraperitoneal injection of MNU was just as effective as intravenous administration of the carcinogen [21]. Thus it is now possible to give the complete initiating dose of carcinogen to a cohort of 100 rats in only a few hours, with a minimum of technical skill required. Furthermore, the marked instability of MNU in dilute alkaline solution makes this an ideal carcinogen to handle. Finally, the model has been standardized in the readily available Sprague–Dawley rat. The use of this outbred rat as a model remains relevant even though inbred strains have shown promise in studying genetic susceptibility to breast cancer [22, 23].

10.2.1 Induction of Mammary Carcinomas Using MNU Mammary carcinomas can be induced in female Sprague–Dawley rats by administering a single dose of MNU intraperitoneally [21, 24]. The use of this outbred rat strain is strongly recommended since it is both readily available and highly susceptible to chemically induced mammary carcinogenesis. The characteristics of the carcinogenic response are also best characterized in this strain. A discussion of the genetic susceptibility of this and other rat strains to chemically induced mammary carcinogenesis can be found in [25, 26]. An extensive description of the technical details by which MNU is prepared, injected, and inactivated has been reported recently [27]. A summary of the most critical aspects of this protocol is presented in Table 10.1.

248

H.J. Thompson

Table 10.1  Protocol for the induction of mammary carcinomas using MNU Storage and analysis of MNU

• MNU, obtained from a commercial source, is stored on desiccant below –10°C prior to use. • A solution of 14 mg/ml dissolved in water is prepared prior to use for the carcinogenesis protocol for spectral analysis to confirm purity of the compound (use a 1:1,000 dilution). • Concentration of MNU can be confirmed by measuring the absorbance at lmax = 231 nm; the extinction coefficient for MNU is log e3.77 (in water).

Preparation of MNU for injection

• MNU for animal injection is prepared in 0.9% NaCl acidified to pH 5.0 using acetic acid. • A satisfactory working concentration is 14 mg MNU/ml; more dilute solutions can be prepared, but this increases the injection volume required. • Care must be exercised to insure that MNU is completely dissolved in the saline solution; this is achieved by gentle heating with hot tap water accompanied by vigorous shaking. This process should be carried out in a sealed injection vial. • It is usually recommended that MNU be used immediately after preparation. However, spectral analysis indicates that the MNU solution is >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 <30% of injected animals are observed with palpable mammary carcinomas after a latency in excess of 6 months. The age at which MNU is injected has a significant impact on the rapidity of the carcinogenic response. From the time of the initial report of Huggins et al. [16], the accepted age at which to inject, at least the initial dose of carcinogen if multiple injections were used, was 50 days of age. Clearly, the animal is very responsive to carcinogenic insult at this age. However, reported data have shown that the injection of MNU at ages prior to 50 days results in robust tumor development [24]. We have reported on the carcinogenic response when rats were injected with 50 mg MNU/ kg at 21 days of age [28]. Greater than 90% incidence of invasive carcinomas was detected in wholemount preparations within 35 days post-carcinogen, and at this time >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 (<2 h); this reduces the management issues and costs associated with containment of carcinogen treated animals and disposal of animal-related carcinogen waste • Chemical inactivation of MNU is easily accomplished • Overall, the procedures are convenient and economical Carcinogenic • Short latency to tumor emergence response • High incidence and multiplicity of mammary carcinomas and low prevalence of benign mammary tumors • Low incidence of tumors at other organ sites • Mammary tumors can be detected by palpation throughout the time course of an experiment • Rate of tumor growth and/or regression can be monitored via caliper measurements Experimental • Ability to operationally distinguish between carcinogenic initiation and design promotion/progression • In general experimental groups comprised 25–30 animals provide adequate statistical power • Ability to complete a study in as little time as 5 weeks following MNU injection

response observed in a control group of animals that are administered a placebo treatment. The endpoints of the carcinogenic response that are usually compared are the incidence and multiplicity of carcinoma and the latency to their detection as assessed by palpation. For statistical reasons, we recommend studying the effects of an agent using 30 animals per experimental group. If rats are injected with 50 mg MNU/kg body weight at 21 days of age, the experiment can be usually completed within 7–8 weeks if the occurrence of palpable mammary tumors, histopathologically classified, is used as an endpoint. If the carcinogenic response is assessed in wholemount preparations, the study can be completed within 5–6 weeks of carcinogen treatment. However, this approach incurs more effort and resources and is recommended only if there is an interest in quantifying the occurrence of pre-malignant as well as malignant mammary gland pathologies. If MNU treatment is delayed until 50 days of age, the typical duration of an experiment is 6 months, although some investigators have reduced this interval by 1–2 months. Given that similar information is obtained from both protocols relative to screening a compound for potential chemopreventive activity, the savings in time, effort and resources can be considerable if the 21 days of age injection protocol is used. This experimental timeframe is especially important if custom diet formulations containing chemopreventive agents are of limited supply [31]. Additional information related to the design of experiments and their execution is provided in [27, 30].

10  Mammary Cancer in Rats

251

10.2.3.2 Therapeutic Protocol A primary objective of a tumor induction protocol to test a therapeutic agent is to rapidly generate animals bearing palpable mammary tumors of measurable dimensions. It is presumed that the shorter the time period between carcinogen administration and the detection of tumors with measurable dimensions, the more expediently the experiment can be completed. The injection of 21-day-old female Sprague–Dawley rats with 50 mg MNU/kg body weight should yield a majority of the animals with palpable tumors within 7–8 weeks following carcinogen administration. As reported in [24], it is possible to further shorten latency by injecting up to 75 mg MNU/kg body weight. What is generally more variable than latency is the time that it takes for palpable mammary tumors to reach dimensions that can be reproducibly measured using vernier calipers. Animals are randomized to treatment regimes as they develop tumors with measurable dimensions. Some investigators wanting a small number of tumor-bearing rats frequently inject only a small number of animals, and at times are disappointed by failure to rapidly obtain tumors that are of measurable size. We generally recommend injecting up to twice the number of animals required for a particular experimental protocol so that animals with tumors of measurable dimensions can be randomized to a treatment protocol over a narrow time frame. We also note that some investigators have not reported tumor yields equivalent to those mentioned above. Through many personal communications with investigators adapting this model system, we have learned that other than errors in carcinogen preparation or administration, that two other factors can impact tumor latency and yield: excessive noise in an animal facility and disruption of the animals’ normal routines within an animal room, particularly if such disruptions include frequent transport of animals to different rooms within the facility and/or the frequent administration of anesthesia. Care should also be taken to adjust the light cycle to which animals are exposed; a 12 h light/12 h dark cycle is generally recommended in the animal holding room.

10.2.4 Biological Characteristic of Mammary Carcinomas Induced by MNU While the biological characteristics of chemically induced mammary carcinomas in rats have been reviewed [32, 33], we briefly draw attention to several of these characteristics and summarize them in Table 10.3. Chemically induced mammary carcinomas in the rat originate from ductal mammary epithelial cells, most likely from cells in the terminal end bud. This is significant since the majority of human breast cancers have a ductal histogenesis [34]. Moreover, most evidence indicates that these carcinoma have a similar pathogenesis to the human disease, i.e., they progress from ductal hyperplasias without or with atypia to ductal carcinoma in situ to invasive carcinoma. As reported by Thompson et al. [28], pre-malignant as well as malignant

252

H.J. Thompson

Table 10.3  Biological characteristics similar to the human disease process Biological characteristics Features Histogenesis Ductal epithelial cells Pathogenesis Ductal hyperplasia, ductal carcinoma in situ, invasive carcinoma Histopathology Carcinoma with cribiform, papillary, and mixed morphology are most common Ovarian hormone dependence Greater than 70% of mammary tumors that reach a palpable size regress in response to ovariectomy Pregnancy Full term pregnancy prior to carcinogenic initiation protects against tumor formation Metastasis While infrequent, metastases to the lung have been noted Pathogenetic characteristics Altered expression of TGFa, ErbB2, cyclin D1, and gelsolin has been reported

stages of the disease can be studied in animals injected with MNU at 21 days of age. Histologically, the carcinomas induced and their pre-malignant precursors have many similarities to the human disease, although there are also obvious differences [34]. A  histological comparison of the lesions induced in this model relative to those observed in the human disease has been published recently [35]. Other characteristics of the human disease process reflected in the MNU model are the occurrence of both ovarian hormone-dependent and hormone-independent carcinomas, and the protection against tumor development conferred by a full term live birth prior to carcinogenic initiation. One characteristic that is of major concern in the human disease and that is observed infrequently in this model, is the occurrence of metastases. The lung appears to be a primary site of occurrence when metastases are observed. An understanding of the molecular biological characteristics of mammary carcinomas induced by MNU is only beginning to emerge. This topic has been reviewed in [1, 36]. To date it appears that MNU-induced mammary carcinomas, like their human counterparts, have altered expression of TGFa, ErbB2, cyclin D1, and gelsolin. Cytogenetic analyses clearly demonstrate that tumor cells are aneuploid, and structural alterations of chromatid, iso-chromatid breaks and translocations involving chromosome 1, 2, and 4 are detectable [37]. A G to A transition mutation in codon 12 of the Ha-ras gene is observed in MNU-induced mammary carcinomas; however, a common misconception is that most MNU induced mammary carcinomas harbor this mutation. At the dose of MNU most commonly used, they do not, an observation of importance since this mutation is uncommon in human breast cancer [38]. Interestingly MNU-induced mammary carcinomas do not appear to have misregulated p53 activity as is commonly seen in the human disease, although the signaling pathway of which phosphatidyl inositol-3 kinase (PI-3 kinase) is a component is activated in MNU-induced mammary carcinomas [39] and this pathway is misregulated in the majority of breast cancers in women [40, 41]. It is clear that additional work is required to clarify the pathogenetic basis of the disease process induced in the mammary gland by MNU. The continued study of the genetic susceptibility, especially regarding the study of quantitative trait loci (QTLs) remains an important area [22, 42].

10  Mammary Cancer in Rats

253

10.2.5 Value of this Model Relative to Genetically Engineered Models for Mammary Cancer The advantage of the MNU-induced model is that autochthonous mammary carcinomas can be induced rapidly, reproducibly, and relatively inexpensively. These carcinomas have many characteristics similar to the human disease, and this model has been shown to have value in identifying agents that are effective in the control and treatment of the human disease process. While the availability of transgenic and knockout models for the study of mammary carcinogenesis clearly provides an invaluable and complementary set of tools to investigate the genesis, prevention, and treatment of breast cancer, currently, none of these models can fully replace the advantages offered in the MNU-induced mammary carcinogenesis model in the rat. Acknowledgement  The contribution of Michael B. Sporn to the writing of the original version of this chapter is gratefully acknowledged.

References 1. Medina D. Chemical carcinogenesis of rat and mouse mammary glands. Breast Dis. 2007;28:63–8. 2. Vargo-Gogola T, Rosen JM. Modelling breast cancer: one size does not fit all. Nat Rev Cancer. 2007;7(9):659–72. 3. Li Y, Brown PH. Prevention of ER-negative breast cancer. Recent Results Cancer Res. 2009;181:121–34. 4. Uray IP, Brown PH. Prevention of breast cancer: current state of the science and future opportunities. Expert Opin Investig Drugs. 2006;15(12):1583–600. 5. William WN, Jr., Heymach JV, Kim ES, Lippman SM. Molecular targets for cancer chemoprevention. Nat Rev Drug Discov. 2009;8(3):213–25. 6. Howe LR, Lippman SM. Modulation of breast cancer risk by nonsteroidal anti-inflammatory drugs. J Natl Cancer Inst. 2008;100(20):1420–3. 7. Sporn MB. Dichotomies in cancer research: some suggestions for a new synthesis. Nat Clin Pract Oncol. 2006;3(7):364–73. 8. Sporn MB, Liby KT. Cancer chemoprevention: scientific promise, clinical uncertainty. Nat Clin Pract Oncol. 2005;2(10):518–25. 9. Gottardis MM, Jordan VC. Antitumor actions of keoxifene and tamoxifen in the N-nitrosomethylurea-induced rat mammary carcinoma model. Cancer Res. 1987;47(15):4020–4. 10. Chlebowski RT, Col N, Winer EP et  al. American Society of Clinical Oncology technology assessment of pharmacologic interventions for breast cancer risk reduction including tamoxifen, raloxifene, and aromatase inhibition. J Clin Oncol. 2002;20(15):3328–43. 11. Cummings SR, Eckert S, Krueger KA et al. The effect of raloxifene on risk of breast cancer in postmenopausal women: results from the MORE randomized trial. Multiple Outcomes of Raloxifene Evaluation. JAMA. 1999;281(23):2189–97. 12. Decensi A, Zanardi S, Argusti A, Bonanni B, Costa A, Veronesi U. Fenretinide and risk reduction of second breast cancer. Nat Clin Pract Oncol. 2007;4(2):64–5. 13. Zanardi S, Serrano D, Argusti A, Barile M, Puntoni M, Decensi A. Clinical trials with retinoids for breast cancer chemoprevention. Endocr Relat Cancer. 2006;13(1):51–68.

254

H.J. Thompson

14. Anzano MA, Peer CW, Smith JM et al. Chemoprevention of mammary carcinogenesis in the rat: combined use of raloxifene and 9-cis-retinoic acid. J Natl Cancer Inst. 1996;88(2):123–5. 15. Moon RC, Thompson HJ, Becci PJ et al. N-(4-Hydroxyphenyl)retinamide, a new retinoid for prevention of breast cancer in the rat. Cancer Res. 1979;39(4):1339–46. 16. Huggins H, Grand LC, Brillantes FP. Mammary cancer induced by a single feeding of polymucular hydrocarbons, and its suppression. Nature. 1961;189:204–7. 17. McCormick GM, Moon RC. Effect of pregnancy and lactation on growth of mammary tumors induced by 7,12-dimethylbenz(a)anthracene (DMBA). Br J Cancer. 1965;19:160–6. 18. Gullino PM, Pettigrew HM, Grantham FH. N-Nitrosomethylurea as mammary gland carcinogen in rats. J Natl Cancer Inst. 1975;54(2):401–14. 19. Moon RC, Grubbs CJ, Sporn MB, Goodman DG. Retinyl acetate inhibits mammary carcinogenesis induced by N-methyl-N-nitrosourea. Nature. 1977;267(5612):620–1. 20. McCormick DL, Adamowski CB, Fiks A, Moon RC. Lifetime dose–response relationships for mammary tumor induction by a single administration of N-methyl-N-nitrosourea. Cancer Res. 1981;41(5):1690–4. 21. Thompson HJ, Adlakha H. Dose-responsive induction of mammary gland carcinomas by the intraperitoneal injection of 1-methyl-1-nitrosourea. Cancer Res. 1991;51(13):3411–5. 22. Adamovic T, McAllister D, Rowe JJ, Wang T, Jacob HJ, Sugg SL. Genetic mapping of mammary tumor traits to rat chromosome 10 using a novel panel of consomic rats. Cancer Genet Cytogenet. 2008;186(1):41–8. 23. Lan H, Kendziorski CM, Haag JD, Shepel LA, Newton MA, Gould MN. Genetic loci controlling breast cancer susceptibility in the Wistar-Kyoto rat. Genetics. 2001;157(1):331–9. 24. Thompson HJ, Adlakha H, Singh M. Effect of carcinogen dose and age at administration on induction of mammary carcinogenesis by 1-methyl-1-nitrosourea. Carcinogenesis. 1992;13(9):1535–9. 25. Shepel LA, Gould MN. The genetic components of susceptibility to breast cancer in the rat. Prog Exp Tumor Res. 1999;35:158–69. 26. Shull JD. The rat oncogenome: comparative genetics and genomics of rat models of mammary carcinogenesis. Breast Dis. 2007;28:69–86. 27. Thompson HJ. The induction of mammary carcinogenesis in the rat using either 7,12 dimethylbenz[a]anthracene or 1-methyl-1-nitrosourea. In: Ip M, Asch B, editors. Methods in mammary gland biology and breast cancer research. New York: Kluwer/Plenum; 2000. p. 19–30. 28. Thompson HJ, McGinley JN, Rothhammer K, Singh M. Rapid induction of mammary intraductal proliferations, ductal carcinoma in situ and carcinomas by the injection of sexually immature female rats with 1-methyl-1-nitrosourea. Carcinogenesis. 1995;16(10):2407–11. 29. Thompson HJ, McGinley JN, Wolfe P, Singh M, Steele VE, Kelloff GJ. Temporal sequence of mammary intraductal proliferations, ductal carcinomas in situ and adenocarcinomas induced by 1-methyl-1-nitrosourea in rats. Carcinogenesis. 1998;19(12):2181–5. 30. Thompson HJ, Singh M, McGinley J. Classification of premalignant and malignant lesions developing in the rat mammary gland after injection of sexually immature rats with 1-methyl1-nitrosourea. J Mammary Gland Biol Neoplasia. 2000;5(2):201–10. 31. Thompson MD, Thompson HJ, Brick MA et al. Mechanisms associated with dose-dependent inhibition of rat mammary carcinogenesis by dry bean (Phaseolus vulgaris, L.). J Nutr. 2008;138(11):2091–7. 32. Welsch CW. Host factors affecting the growth of carcinogen-induced rat mammary carcinomas: a review and tribute to Charles Brenton Huggins. Cancer Res. 1985;45(8):3415–43. 33. Russo J, Russo IH. Experimentally induced mammary tumors in rats. Breast Cancer Res Treat. 1996;39(1):7–20. 34. Russo J, Gusterson BA, Rogers AE, Russo IH, Wellings SR, van Zwieten MJ. Comparative study of human and rat mammary tumorigenesis. Lab Invest. 1990;62(3):244–78. 35. Singh M, McGinley JN, Thompson HJ. A comparison of the histopathology of premalignant and malignant mammary gland lesions induced in sexually immature rats with those occurring in the human. Lab Invest. 2000;80(2):221–31.

10  Mammary Cancer in Rats

255

36. Medina D, Thompson HJ. A comparison of the salient features of mouse, rat and human mammary tumorigenesis. In: Ip M, Asch B, editors. Methods in mammary gland biology and breast cancer research. New York: Kluwer/Plenum; 2000. p. 31–6. 37. Goepfert TM, Moreno-Smith M, Edwards DG, Pathak S, Medina D, Brinkley WR. Loss of chromosomal integrity drives rat mammary tumorigenesis. Int J Cancer. 2007;120(5):985–94. 38. Zhang R, Haag JD, Gould MN. Reduction in the frequency of activated ras oncogenes in rat mammary carcinomas with increasing N-methyl-N-nitrosourea doses or increasing prolactin levels. Cancer Res. 1990;50(14):4286–90. 39. Jiang W, Zhu Z, Thompson HJ. Modulation of the activities of AMP-activated protein kinase, protein kinase B, and mammalian target of rapamycin by limiting energy availability with 2-deoxyglucose. Mol Carcinog. 2008;47(8):616–28. 40. Velculescu VE. Defining the blueprint of the cancer genome. Carcinogenesis. 2008;29(6):1087–91. 41. Gustin JP, Cosgrove DP, Park BH. The PIK3CA gene as a mutated target for cancer therapy. Curr Cancer Drug Targets. 2008;8(8):733–40. 42. Shepel LA, Lan H, Haag JD et al. Genetic identification of multiple loci that control breast cancer susceptibility in the rat. Genetics. 1998;149(1):289–99.

Part V

Disease and Target-Specific Models

Chapter 11

Animal Models of Melanoma Ene T. Fairchild and William E. Carson, III

Abstract  A wide variety of animal models of malignant melanoma are currently available to the researcher interested in this disease process. The Xiphophorus fish represents a genetic form of melanoma in which tumor formation is tied to the activity of a tyrosine kinase and specific tumor-suppressor genes. The South American opossum Monodelphis domestica possesses the photoreactivation pathway of pyrimidine dimer repair, and has been used extensively in photobiology studies of melanoma. Melanomas arise spontaneously in dogs, and this model has been used to evaluate novel cytokine combinations and gene-therapy protocols. The Sinclair swine also develops melanoma tumors, and the tumor regressions seen in this model provide a unique model for the study of interactions between tumor cells and immune effectors. Murine models are by far the most widely employed animal model of melanoma, and include the induction of tumors via application of physical agents or transgenic manipulations as well as the inoculation of mice with tumor-cell lines such as the spontaneously occurring Harding–Passey, Cloudman, and B16 cell lines. Finally, the more recently described models which employ immunodeficient mice provide a method for studying human tumors under ­controlled conditions. Keywords  Melanoma • Xiphophorus monodelphis • B16 • MT-RET • Sc10

11.1 Introduction The worldwide incidence of malignant melanoma is rising faster than any other cancer, and in the United States alone, almost 69,000 new cases of melanoma were projected for the year 2009. While thin primary melanomas are highly curable with surgery, the prognosis for patients with advanced disease is poor: Over 70% of patients with primary melanomas thicker than 4 mm or metastasis to the regional W.E. Carson, III (*) The Ohio State University, Doan Hall, N924, 410 W. 10th Avenue, Columbus, OH 43210, 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_11, © Springer Science+Business Media, LLC 2011

259

260

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

lymph nodes will die of disseminated disease within 5 years of diagnosis. The treatment of metastatic disease has undergone significant changes over the past three decades with the introduction of high-dose cytokine therapy (e.g. interferon-alpha and interleukin-2), biochemotherapeutic regimens, novel vaccination strategies, and sentinel lymph node biopsy techniques. Despite these advances and a better understanding of the molecular lesions underlying the development of the malignant phenotype, the therapeutic options for patients with metastatic disease remain limited. High-dose cytokine therapy infrequently produces durable responses and is often poorly tolerated by patients with advanced disease. Aggressive chemotherapeutic regimens have not consistently proved superior to therapy with single agents, and the improved response rates achievable with biochemotherapy come at the price of significant toxicity. Other approaches to the patient with advanced disease, such as peptide and anti-ganglioside vaccines, remain investigational. Thus, it is critical that novel regimens with favorable toxicity profiles enter into clinical development. In order to move new treatments rapidly into the clinic and to gain a better understanding of their mechanism of action, it is essential that experiments be conducted in animal models of malignant melanoma. This chapter reviews those models that are currently in use and explores their strengths and disadvantages.

11.2 Xiphophorus Species Central American freshwater fish of the genus Xiphophorus do not normally develop pigmented tumors [1]. However, hybrids between different species of this genus may spontaneously develop melanomas in skin structures called “macromelanophores” that are formed by terminally differentiated melanocytes. Tumor development is markedly enhanced following treatments with ultraviolet radiation (UVR), X-rays, carcinogens, or differentiating agents (reviewed in Ref. [2]). A series of studies in Xiphophorus hybrids has revealed that the development of melanomas is genetically determined [2, 3]. The responsible genes (termed “Tu” for tumor) map to the sex chromosomes and are necessary for the formation of normal black spots within the skin as well as melanoma lesions. The Tu genes exhibit oncogenic potential in the absence of any moderating factors and thus act in a dominant fashion. Tumor development is normally inhibited via the action of tumor-suppressor genes (“R” for regulating) which must be present in two copies in order to be fully functional. The R genes may or may not be linked to the Tu genes; however, R genes are not found in species that lack Tu genes. Melanoma results when the R genes of a Tu-bearing species are deleted from the genome by hybridization with a species lacking Tu genes [2, 4]. Wittebrodt et  al. have cloned and characterized an important Tu gene (Xmrk) that is homologous to a number of other Tu alleles in Xiphophorus [5–8]. Xmrk apparently arose as the result of non-homologous recombination involving a highly related proto-oncogene that is present as a single copy in all species of Xiphophorus [8]. The recombination event resulted in a second copy of the gene that was placed

11  Animal Models of Melanoma

261

under the control of a completely separate promoter region and thus was not subject to the same controls as the parent proto-oncogene. Overexpression of the Xmrk oncogene is necessary and sufficient for neoplastic transformation of pigment cells in Xiphophorus, and levels of Xmrk expression correlate with the malignancy of the melanoma tumor that eventually develops [9]. The protein encoded by the Xmrk gene is a novel receptor tyrosine kinase that is closely related to the receptor for human epidermal growth factor [5]. This kinase is constitutively autophosphorylated and can induce tumor formation in the embryos of medakafish when driven by a strong promoter sequence [10]. There is evidence to suggest that transformation occurs in response to the binding of an unknown ligand to the extracellular domain of the Xmrk gene product [10]. Furthermore, Xmrk kinase induction results in signaling through multiple pathways by activating STAT5, phophotidylinositol 3-kinase, and Ras-Raf-ERK. These pathways have also been implicated in the pathogenesis of human melanoma [11–13]. In addition to altering cell proliferation, survival, and motility, Xmrk suppresses differentiation by interfering with signaling via the micropthalmia-associated transcription factor (MITF) [14]. Certain hybrids and backcrosses are particularly sensitive to melanoma induction by carcinogens or tumor promoters, and others readily develop tumors following exposure to UVR wavelengths in the range of 290–304 nm. The mechanism involved appears to be inactivation of tumor-suppressor genes (such as R gene), which are present in a single copy in these fish [2]. Malignant melanomas arise from the macromelanophore: a flattened, multinucleated, terminally differentiated cell that is 100–400 mm in diameter, contains a number of mature melanosomes, and is derived from the melanocyte population of the skin. Malignant lesions ­contain macromelanophores and small, spindle-shaped melanocytes with few melanosomes but numerous mitotic figures. Microscopic analysis reveals that these melanocytes differentiate from skin melanoblasts [15]. These tumors are locally aggressive, and invasion of the underlying muscle is common. Malignant melanomas are rapidly fatal to affected fish. Xiphophorus are easily maintained, are possessed of a short generation time, and are genetically well-characterized. This system is therefore well-suited for the identification of oncogenes and tumor-suppressor genes involved in the development of malignant melanoma. However, the highly aggressive and invasive nature of these tumors and their genesis in pigmented structures unique to fish must be taken into account when interpreting any results obtained in this model.

11.3 South American Opossum The South American opossum Monodelphis domestica possesses the photoreactivation pathway of pyrimidine dimer repair and has been used extensively in photobiology studies. Pyrimidine dimers induced by UVR are recognized by the photolyase enzyme, which catalyzes the enzymatic monomerization of dimers using the energy of long-wavelength visible light. Thus, exposure of animals to photoreactivating

262

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

light (PRL) can remove 80–90% of pyrimidine dimers induced by suberythemal doses of UVR [16, 17]. In fact, exposure of UVR-treated Monodelphis to PRL reduces the incidence of melanoma from near 40% to well under 10% – clear evidence that UVR-induced DNA damage is important in the induction of melanoma [18]. Experimental comparisons can be made between UVR-exposed animals that have been treated with PRL and those that have received a control treatment of equal duration. In this manner, the effects of UVR-induced DNA damage can be specifically evaluated. UVR alone can induce melanoma tumors in Monodelphis. This is the only nontransgenic animal in which UVR alone acts as a complete carcinogen [19]. Genetic analysis of these UVR-induced melanoma tumors has revealed that, similar to lesions arising in humans, mutations in p53 rarely occur and that these lesions progress in a p53-independent manner [20]. However, unlike human melanomas, UVR-induced primary lesions in Monodelphis retain wild-type copies of both H-ras and N-ras. It has been hypothesized that this difference is due to the dermal origin of the Monodelphis tumors and the absence of the vertical growth phase found in human lesions [21]. Interestingly, application of carcinogenic compounds such as 9,10-dimethyl-1,2-benz(a)anthracene are also highly effective in the induction of malignant melanoma in this species [22]. Monodelphis is a small (approximately 100 g) pouchless opossum that is easily maintained and has been well-characterized via numerous studies relating to the damaging effects of UVR [23]. It is clear that the development of tumors is not genetically determined. Melanomas do not arise spontaneously in this species with any great regularity; however, application of UVR (e.g. 125–250 J/m2, three times per week, for 70–80 weeks) will induce tumors in 10–20% of animals, of which about one-third will metastasize to lymph nodes [18]. All tumors, regardless of their origin (e.g. naturally occurring, carcinogen-induced, and UVR-induced) possess the ability to metastasize to lymph node basins, whereas only spontaneous and carcinogen-induced tumors can spread to visceral sites [2]. Melanomas arising in Monodelphis are dermal in location and have their origins in lesions of melanocytic hyperplasia [19, 22, 24]. Dendritic melanocytes enlarge during the development of malignant lesions and assume a more polygonal shape. Eventually, they will give rise to clusters of infiltrative cells exhibiting deep pigmentation and epithelioid characteristics. The resulting lesions are characterized by cellular atypia, nuclear pleomorphism, and areas of depigmentation. Kusewitt et al. have also demonstrated the presence of S-100 immunoreactivity in the foci of melanocytic hyperplasia, UVR- and carcinogen-induced primary tumors, as well as in metastatic lesions [24]. Tumors can be established in tissue culture and even implanted into immunodeficient mice for further analysis. Additionally, murine and human melanoma cell lines can be injected subcutaneously (sc) into young (0–4 days) opossums. Due to the immunological immaturity of these young animals, the cell lines are able to engraft, form tumors, and in some cases metastasize. However, when the pups reach 5 weeks of age, the engrafted tumors begin to show signs of regression as their immune systems mature [25, 26]. Older animals appear to be more susceptible to the effects of UVR exposure and develop melanocytic lesions with much greater frequency than younger animals.

11  Animal Models of Melanoma

263

Interestingly, approximately 40% of newborn opossums exposed to UVR will later develop metastatic melanoma as adults. This suggests that early exposure to sunlight may act as an initiator in the progression to melanoma. The susceptibility to UVR-induced melanomas can be modulated by dietary factors. Animals receiving a diet high in saturated fat do not develop skin lesions following exposure to UVR, whereas no protection is observed in animals receiving high levels of unsaturated fat in their feed [2]. The susceptibility of Monodelphis to UVR, the ability to modulate DNA repair by photoreactivation, and the presence of large, easily characterized chromosomes makes this animal model attractive for studies of photobiology. Disadvantages include the dermal location of melanoma lesions, the infrequent rate of spontaneous regression, and the lack of information regarding the immune system of this species.

11.4 Canine Melanoma Canine melanomas arising in the oral cavity, mucocutaneous junction, and distal extremities are highly aggressive tumors that grow rapidly, invade locally, and metastasize readily to lymph nodes and distant organs; cutaneous canine melanomas are usually benign. Melanoma may arise in any breed, but it is especially prevalent in purebred dogs such as Golden Retrievers, Irish Setters, Scottish Terriers, Doberman Pinschers, and Schnauzers [27, 28]. Canine melanomas are difficult to control even with radical surgery, and chemotherapy and radiation therapy are not considered effective in the control of either primary tumors or metastatic disease, although significant palliation may occasionally be achieved. In the absence of distant metastases, the median survival following aggressive surgical resection of primaries located in the oral cavity or on the digits is approximately 3–12 months [29–32]. Radiation treatments may be useful as an adjuvant to surgery, but not in the face of significant residual disease [33–35]. Chemotherapeutic regimes employing dacarbazine (DTIC) with or without cyclophosphamide are employed most commonly in the setting of advanced disease but generally yield only partial response of short duration [28]. The lungs are the most common site of metastasis, and spread to this organ is the usual cause of death. Thus, it should come as no surprise that 1 year survival rate for canine oral melanoma is just 25%. In the absence of distant disease, the size of the primary lesion and the ability to obtain local control are the most important determinants of survival [31]. A high mitotic rate and presence of aneuploidy may also possess prognostic significance; however, other factors such as breed, age, sex, and anatomic location do not appear to affect prognosis [27]. Like their human counterparts, metastatic canine melanomas tend to be resistant to conventional therapies. Numerous in vitro studies have indicated that immunologic approaches to the treatment of this tumor may be effective [27]. Therefore, canines with malignant melanoma have been employed in several large studies of biological response modifiers and other immunotherapeutic modalities. For ­example, ­administration of inter-

264

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

leukin-2 (IL-2) in combination with recombinant tumor necrosis factor alpha has resulted in tumor regression in five of 13 dogs bearing malignant melanoma tumors. One of the five dogs exhibited a complete response that persisted for >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.

11  Animal Models of Melanoma

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

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

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

11  Animal Models of Melanoma

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

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

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].

11  Animal Models of Melanoma

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

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

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

11  Animal Models of Melanoma

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

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

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

11  Animal Models of Melanoma

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

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

­ 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].

11  Animal Models of Melanoma

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

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

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

11  Animal Models of Melanoma

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

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

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.

11  Animal Models of Melanoma

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

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

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.

11  Animal Models of Melanoma

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

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

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.

11  Animal Models of Melanoma

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 <1 mm diameter by day 15. By day 20, numerous lung metastases (100–200) were consistently enumerated in both lobes and increased in size [20, 21, 23]. Tumor was not detectable in other organs. Mice became moribund by day 30, and a median survival of 38 days was observed. This more predictable model of lung metastases was essential for quantitative evaluation of new therapeutic approaches for the treatment of lung metastases [20, 23, 24, 27, 29]; thus mimicking the clinical situation of postnephrectomy patients presenting with lung metastases. Using these various Renca models, we have studied the interaction between local tumor irradiation and systemic (i.p.) IL-2 immunotherapy. In the kidney tumor model, a greater therapeutic effect was demonstrated on the primary tumor and distant metastases by local irradiation of the tumor-bearing kidney followed by IL-2 therapy than with each modality by itself [19]. In the pulmonary metastases model, irradiation of the left tumor-bearing lung followed by systemic IL-2 therapy resulted in increased tumor reduction in both lungs suggesting that radiation enhances the systemic effect of IL-2 [20, 23]. We demonstrated the requirement for T cell and NK cell functions in the mechanism of action of the combined therapy [23]. In histological studies of lung tumor nodules, we found that radiation caused not only tumor cell apoptosis but also vascular damage allowing for an influx of macrophages and mononuclear cells in the tumor nodules and surrounding tissues [23, 24]. The combination of both therapies induced a greater vascular damage and massive infiltration of immune cells that may play a role in tumor destruction [23, 24]. Our data in the Renca models suggest that radiation therapy causes changes in the tumor cells and the tumor environment which increase the tumor susceptibility to destruction by the immune system activated by IL-2. The Renca system was an ideal animal model to investigate these complex issues, and these studies were translated into a clinical trial for metastatic RCC [25]. Independent studies recently demonstrated enhanced therapeutic effect of intratumoral administration of dendritic cells combined with localized radiotherapy in subcutaneous Renca tumors, confirming the role of radiation in enhancing immunotherapy approaches [30]. Amplified antitumor responses were associated with enhanced systemic cytokine production from splenocytes in response to tumor cells. A new original study combined i.p. IL-2 injections with oral administration of the histone deacetylase (HDAC) inhibitor MS 275 which have shown antitumor activity and immunomodulatory properties [31]. These studies were performed in mice bearing orthotopic Renca kidney tumors expressing luciferase to image mice

12  Animal Models for Kidney Cancer

291

at different time points after treatment and assess tumor response in the kidneys and inhibition of lung metastases [31]. This bioluminescence approach to follow in situ tumor growth in different organs is a novel and well accepted technique. These studies demonstrated a synergistic effect of IL-2 with HDAC inhibitor and decreased Foxp3+ regulatory T cells [31]. The Renca kidney model and lung metastases model were also helpful in the evaluation of other cytokines, including IL-4 and IFNg [27–29]. Following in vitro studies on human RCC cell lines demonstrating growth inhibition induced by IL-4 and IFNg [26], we have tested the therapeutic effect of these cytokines in different Renca models in vivo. We found that systemic treatment of either cytokine-induced regression of pulmonary metastases in a dose-dependent manner and increased mouse survival [27, 29]. We showed that IL-4 and IFNg have different antitumor mechanisms of action; consequently we found that sequential administration of IFNg followed by IL-4 resulted in a greater therapeutic effect [29]. During these studies, we developed a new treatment approach of intratumoral injection of IL-4 in Renca kidney tumors [28]. On day 16 following Renca cell renal implantation, the tumor-bearing kidney was re-exposed through a midline abdominal incision, and the kidney tumor (of 4–6 mm diameter) was injected with 50-µl IL-4. This treatment was safely performed and resulted in a marked inhibition of tumor growth in the kidney, with minimal effect on the progression of metastases in other sites [28]. Thus, the Renca kidney model can be adapted to therapeutic manipulations. Other studies used IL-4 for transfection of Renca tumor cells and showed that IL-4 transduced Renca cells injected s.c. caused systemic immunity [8, 9, 15, 32]. In the 1990s, the newly discovered cytokine IL-12 seemed to be the most promising, because it was found to be particularly effective at mediating a specific immune antitumor response in various tumor models including the Renca model [33, 34]. Subsequently, the Renca model was extensively used to study IL-12 mechanism of action, IL-12 gene therapy and IL-12 combination with other cytokines. IL-12 antitumor effects were shown to be increased by tumor irradiation [34], similar to our findings with IL-2 and tumor irradiation [19–24]. Using the Renca kidney model, Wiltrout et al. showed that IL-12 administered with pulse IL-2 was safe and induced a rapid and complete regression of primary and metastatic Renca tumors that was greater than with each cytokine alone [35]. This effect was found to be related to enhanced macrophage production of nitric oxide following studies in the s.c. Renca model [36]. Furthermore, in the same s.c. Renca model, the mechanism of IL-12 antitumor activity was found to depend on induced expression of IFNg by T and NK cells, leading to IFNg induced expression of chemokines IP-10 and Mig within tumor tissue. These chemokines act as chemoattractants for activated T lymphocytes [37, 38]. IL-12 was also found to potentiate the cytolytic effector function of recruited CD8+ T cells [37]. These studies in the Renca model led to monitoring IFNg, IP-10 and Mig mRNA in biopsies of RCC tumors and peripheral blood mononuclear cells from IL-12-treated patients demonstrating augmented levels of those molecules after therapy [39]. Thus, the Renca model is helpful for testing new treatments for translational purpose, and it also allows the investigation of mechanism of these treatments and determination of response parameters for patient monitoring.

292

G.G. Hillman

A novel Fas-mediated antitumor pathway involving also IFNg was discovered in studies performed with the Renca model [40, 41]. Renca cells express low levels of Fas that can be enhanced by IFNg and TNFa, making the cells susceptible to apoptosis induced by Fas ligand (FasL)-expressing hybridomas (dIIS), cross-linking of anti-Fas Abs or soluble Fas (FasL) [40]. The Fas/FasL pathway was found to be involved in cell-mediated killing of Renca cells by activated T cells, whereas a granule-mediated pathway predominated in killing of Renca cells by activated NK cells [40]. When Fas-overexpressing Renca cells (post-IFNg and TNFa treatment) were injected into mice, there was a consistent and significant delay in tumor progression, reduced metastasis and prolonged survival [41]. These findings show that therapeutic manipulation of Fas expression may represent a means of tumor treatment as well as a novel mechanism for IFNg-mediated antitumor effect. More recently, a new Renca model expressing a defined human cancer-associated antigen was generated [42]. The Renca/CA-IX model was developed by transduction of Renca cells with human carbonic anhydrase (CA-IX) gene coding for a cell surface tumor-associated antigen expressed by most clear RCC [42]. This model maintains antigen expression, forms metastases and has tumor growth kinetics equivalent to that of the original Renca cell line. This new model will be useful to study immunotherapeutic approaches and specific immune response to a human RCC antigen in immuno-competent Balb/c mice. The Renca model was also very useful for addressing various approaches of cytokine gene therapy to circumvent the toxicity caused by systemic injection of cytokines. For generation of cancer vaccines, Renca cells were transfected with cytokine genes. Renca cells transfected with IL-2, IFNg, granulocyte–macrophage colony-stimulating factor (GM-CSF) or IL-12 failed to produce s.c. tumors [43, 44]. We found that s.c. vaccination with cytokine transfected Renca cells combined with lung radiation significantly reduced the number of pulmonary metastases in the Renca lung metastases model [44]. Others showed that IL-12 transfected Renca cells used as a cancer vaccine inhibited the growth of parental Renca cells injected at a distant site and were synergistic with systemic IL-18 treatment [43]. We and others investigated the direct use of cytokine gene therapy for established Renca tumors using different vectors. An IL-12 adenovirus construct, administered i.v., inhibited Renca hepatic metastases in a Renca hepatic metastasis tumor model induced by intrasplenic Renca cell injection followed by splenectomy [45]. This effect seemed to be mediated by macrophages and neutrophils [45]. Recently, we demonstrated the efficacy of adenoviral vectors expressing cytokine cDNA genes encoding the human IL-2 cDNA (Ad-IL-2) and murine IFN-g gene (Ad-IFN-g) [46]. These constructs were particularly effective when three intratumoral injections were administered in Renca s.c. tumors compared to single dose gene therapy. We found that tumor irradiation enhanced the therapeutic efficacy of Ad-IL-2 and Ad-IFN-g intratumoral gene therapy. Tumor irradiation, administered one day prior to three doses of Ad-IL-2 treatment, was more effective than radiation or Ad-IL-2 alone, resulting in tumor growth arrest in all mice, increased survival and a consistent increase in complete tumor regression response rate [46]. Complete responders rejected Renca tumor challenge and had specific cytotoxic T cell activity;

12  Animal Models for Kidney Cancer

293

demonstrating induction of specific tumor immunity. The effect of radiation combined with three doses of Ad-IFN-g was less pronounced and did not lead to tumor immunity [46]. Histological observations showed that irradiation of the tumor prior to gene therapy increased tumor destruction and inflammatory infiltrates in the tumor nodules [46]. These findings demonstrate that tumor irradiation improves the efficacy of Ad-IL-2 gene therapy for induction of antitumor immune response [46]. These studies also showed the increased effectiveness of intratumoral IL-2 gene therapy preceded by tumor irradiation. This therapeutic approach did not cause toxicity and resulted in complete responders with induction of specific tumor immunity in about 50% of the mice while our previous studies on tumor irradiation and systemic injections of IL-2 did not lead to complete responses [19, 20, 23, 24]. Recently, we found that tumor irradiation enhanced the level and duration of IL-2 cytokine gene expression, both in tumor lysates and serum of mice treated with radiation and Ad-IL-2 (Hillman et  al., manuscript in preparation). Therefore, we suggest that tumor irradiation administered prior to gene-mediated immunotherapy is an effective strategy to debulk the tumor, enhance the probability of cell transfection and increase the level and duration of gene expression in the tumor to generate effective and potent therapeutic cancer vaccines. In another study, we used intratumoral injection of adenoviral vectors to modify the phenotype of Renca tumors to express MHC class II antigens with an IFN-g gene and suppress the invariant Ii chain with an antisense reverse Ii-RGC gene construct [47, 48]. This approach converted, in situ, Renca tumor cells into antigenpresenting cells (APC) in order to stimulate T helper cells by expression of class II antigens and by suppression of Ii protein that normally blocks the antigenic peptide-binding site of MHC class II molecules during synthesis in the endoplasmic reticulum [47, 48]. A single recombinant adenoviral vector containing IFN-g gene and an antisense Ii-RGC gene (rAV/IFN-g/Ii-RGC) efficiently induced the MHC Class II+/Ii- phenotype in Renca tumor nodules, and when combined with a suboptimal dose of Ad-IL-2 induced a potent antitumor immune response [47, 48]. This study shows that the phenotype of tumor cells can be modified by gene therapy in situ to induce an immune tumor response. A different approach involves the use of plasmids containing cytokine cDNA inserts. Intradermal injections of murine IL-12 plasmid DNA-induced systemic biological effects characteristic of IL-12, including enhanced NK activity and IL-12 inducible IFNg genes [49, 50]. Pretreatment of mice with IL-12 plasmid DNA induced tumor growth delay when mice were re-challenged with s.c. Renca [49]. Another approach used IL-2 plasmid DNA complexed with a cationic lipid for intratumoral injection of Renca tumors and showed antitumor activity and induction of specific immunity [51]. In conclusion, the Renca system has been of great value for testing multiple new therapeutic approaches for RCC-including novel gene therapies – and in the investigation of their mechanism of action, as well as in studying the role of several components of the immune system involved in the antitumor response. Other issues related to progression and angiogenesis of RCC were addressed in the Renca system including ways to inhibit these processes. Transfection of Renca

294

G.G. Hillman

cells with basic fibroblast growth factor (bFGF) gene increased their metastatic potential when injected i.v. or intrarenally, probably through the production of metalloproteinase 2 (MMP-2) [52]. The ratio of MMP-2 and tissue of inhibitor of metalloproteinase-2 (TIMP-2) was found to be a critical factor in the invasion and metastases of RCC based on Renca studies [53]. Absence of expression of TGFb type II receptor in Renca was also associated with an aggressive growth pattern [54]. Transfection of Renca cells with this receptor conferred tumor suppressive activity emphasizing the role of TGFb type II receptor as a mediator in Renca tumorigenicity [55]. Transfection of mouse endostatin into Renca cells decreased the rate of tumor growth when these cells were injected s.c. or i.v., suggesting that gene delivery of the angiogenesis inhibitor endostatin may be an effective strategy to prevent progression of RCC disease [56]. VEGF Trap is a newly developed VEGF blocking agent that is a soluble decoy receptor comprising parts of VEGFR-1 and VEGFR-2 based on a human IgG1 backbone. VEGF Trap binds VEGF and suppresses VEGF signaling and angiogenesis. This agent was shown to be a potent inhibitor of primary tumor growth and lung metastasis in the orthotopic Renca model and induced a decrease in vascular leaking into the tumor microenvironment [57].

12.3 Rat Renal Carcinoma Models 12.3.1 The Wistar–Lewis Rat Renal Adenocarcinoma The Wistar–Lewis rat renal adenocarcinoma, like murine Renca, also arose spontaneously in the kidney of a Wistar–Lewis rat and originated in the renal cortical tubules. Its growth and expansion were characterized by White and Olsson and found to be nonhormonal dependent [reviewed in [9]]. This tumor is maintained by transplantation in the flank of syngeneic rats. The s.c. tumor takes 3 weeks to develop and does not metastasize, and the mean survival is unpredictable [9]. To induce metastases, rats were first splenectomized. Then, 24 h later a tumor fragment was placed into the peritoneal cavity. Nine weeks later, all animals displayed ascites and metastatic disease into the diaphragm, bowel and muscle throughout the abdominal cavity [9]. Alternatively, the rat cancer cells could be injected into the renal capsule and resulted into 80% tumor occurrence. However, the time required for widespread metastases was unpredictable. Chemotherapeutic studies were performed on s.c. tumors as well as on the i.p. metastatic tumors and showed selective responses to drugs [8, 9]. Thus, although this rat RCC tumor model was less predictable and required a longer time frame study than Renca, it was useful to test chemotherapeutic agents. Renal implantation of the rat tumor showed a rapid growth rate of the primary kidney tumor, and within 90 days metastasized to the lungs even when the primary tumor was resected. Electron microscopic studies in this model revealed that the rat RCC cells shared many ultrastructural features with human RCC cells, such as the presence of large nuclei, abundant glycogen granules and numerous large vacuoles

12  Animal Models for Kidney Cancer

295

[58]. The primary tumor was used to study the tumor cholesterol metabolism, and also showed some similarity with human RCC tumors in its accumulation of an excess of esterified cholesterol due to an increased rate of cholesterol synthesis [58].

12.3.2 The Eker Rat Model In 1954 and 1961, Eker and Mossige described a dominantly inherited cancer syndrome in Wistar rats in which bilateral multicentric RCC develops at an early age [59]. The hereditary tumors that develop in the Eker rat model have many similarities to their human counterparts. They have similar histology, are bilateral, over express transforming growth factor (TGFa) and do not exhibit a high frequency of ras oncogene activation [60, 61]. Loss of sequences on rat chromosomes 4 (q11 through qter), 5 (monosomy) and 6 (q24) occur in these tumors and tumor-derived cell lines. Animals carrying the Eker mutation develop hemangiosarcomas in the spleen (males and females) and uterine leiomyosarcomas as second primary tumors later in life [60, 61]. Vascular neoplasms (hemangioblastomas) and second primary tumors are also associated with RCC in human von Hippel–Lindau (VHL) disease [61]. Genetic analysis of Eker rats showed that the familial tumors were due to an alteration in a single gene, which caused heterozygote carriers of the mutation to develop spontaneous RCCs between 4 and 12 months of age, whereas rats that are homozygous for the wild-type allele rarely develop spontaneous RCC (<0.5%) [60]. When homozygous, the mutation is lethal prenatally at 9–10 days of gestation [60]. Eker rats were found to have a 70-fold increase in the susceptibility to chemical carcinogenesis [61] and increased sensitivity to radiation [60, 62], resulting in a greater cancer incidence. A carcinogen that targeted both renal epithelial and mesenchymal cells caused an increase in tumors of epithelial origin in susceptible animals; the number of carcinogen-induced mesenchymal tumors was unaffected by the presence of the mutation at the susceptibility locus [61]. The cancer susceptibility gene was identified in the 1990s as the rat homolog of the tuberous sclerosis gene 2 (Tsc2) and the mutation involved a 6.3-kb insertion within an intron of the gene on chromosome 10q [63, 64]. This region was found to be homologous with human chromosome band 16p13.3, the site of human Tsc 2 gene [63]. Rat Tsc2 functions as a tumor suppressor gene as normal kidneys of heterozygote carriers express both normal and abnormal Tsc2 mRNA, whereas primary tumors and tumor-derived cell lines exhibit only the mutant transcript [60]. Tumor suppressor genes represent a class of cancer susceptibility genes in humans. Inheritance of a mutation in one allele of a tumor suppressor gene predisposes individuals to develop tumors after sustaining an additional spontaneous mutation in the remaining normal allele of that gene. In humans, several tumor suppressor genes including the VHL gene, the WTI gene and the Tsc2 gene have been implicated in the development of sporadic as well as hereditary tumors in the kidney [2]. In the Eker rats, a single gene mutation (Tsc2) predisposes to multiple bilateral RCCs with an autosomal-dominant pattern of inheritance. Therefore, animals carrying the Eker mutation serve as a model for hereditary RCC.

296

G.G. Hillman

In the 1990s, the Eker rat model was extensively studied to identify key genetic components in the pathogenesis of RCC and to dissect the molecular mechanisms underlying tubular epithelial carcinogenesis. Very early preneoplastic stages in tubular transformation in hereditary RCC can be identified in Eker rats. By laser microdissection procedure, loss of heterozygosity was detected in the region of Tsc2 locus on rat chromosome 10 in preneoplastic renal tubular lesions in Eker rats [65]. These studies suggested that, in heterozygotes, at least two events (one inherited and one somatic) are necessary to produce large adenomas and carcinomas [65]. Studies on patients with tuberous sclerosis complex (TSC) documented that it is an autosomal-dominant disorder characterized by seizures, mental retardation and hamartomas. These TSC patients have increased frequency for developing RCC at an early age, similar to Eker rats [66]. The role of Tsc2 gene was further elucidated using the Eker rat model. This gene encodes a large membrane-associated GTPaseactivated protein (GAP) designated tuberin, and its biological activity was determined by transfection of Eker rat-derived RCC cells with wild-type Tsc2 gene. These cells demonstrated a decreased ability to form colonies in vitro and tumors in vivo providing evidence for the tumor suppressor function of Tsc2 gene [67]. The Tsc2 gene may also contribute to regulation of the cell cycle and cell survival [68]. Recently, a Tsc2 knockout mouse was generated to further characterize the tuberin function in  vivo [69]. Tsc2 heterozygous mice developed renal carcinomas, and Tsc2 homozygous mice died at 10-day embryonic age, similar to the observations made in Eker rats [69]. These studies emphasize an essential function for tuberin in mouse and rat embryonic development as well as in RCC incidence. Other studies have investigated the role of the VHL gene in the rat RCC pathogenesis. The rat VHL gene was identified and found to be 90% homologous to the human VHL gene [70]. However, alterations were not detected in the rat VHL gene in Eker rat RCC cell lines, suggesting that RCC development in the rat is independent of VHL mutation or deletion in contrast to human sporadic RCC [70]. The influence of hormonal and dietary factors on RCC incidence was also investigated in the Eker rat model. Estrogen treatment was found to enhance the development of renal tumors in Eker rats [71] and a high fat diet increased renal preneoplasia [72]. In conclusion, the Eker rat model is an excellent example of a Mendeliandominant predisposition to RCC cancer in an experimental animal that allows the study of genes and molecular events involved in the development of hereditary RCC and secondary tumors.

12.4 Xenografts of Human RCC Tumor Cell Lines in Immunodeficient Mice Athymic nude mice are deficient in immune T cell functions and do not reject xenografts therefore they are used for implantation of human tissues. In the beginning of the 1990s, human RCC cells have been successfully heterotransplanted in nude mice and grew locally as a solid tumor after s.c. injection [73].

12  Animal Models for Kidney Cancer

297

The tumorigenicity of tumor cells in nude mice is used as a criterion for characterizing the malignancy of the cells isolated from a tumor specimen [73]. In addition, human RCC xenografts established by s.c. or i.p. injection in nude mice have been tested for their responsiveness to therapy. Human RCC tumors growing in nude mice were studied for localization and treatment by monoclonal antibodies (MAbs) including the G250 and DAL K29 MAbs directed specifically against human renal tumor lines. van Dijk et al. have demonstrated that G250 MAb or the bispecific MAb CD3/G250, injected i.v., preferentially localized in human renal tumor transplanted s.c. in nude mice [74]. In a subsequent study, G250 administered i.p. induced a significant growth inhibition of this tumor, and this effect was enhanced by intratumoral injection of IFNa and TNFa resulting in macrophage infiltrates into the tumor stroma. Singh et al., found that human RCC cells administered i.p. in nude mice developed into ascites tumor and could be targeted by specific MAb DAL K29 linked to liposomes containing methotrexate resulting in a higher drug uptake [75]. Other groups explored the subcapsular renal implantation of human RCC cells in the nude mouse kidneys. Based on the concept that neoplasms are heterogeneous and contain subpopulations of cells with different biological behavior patterns, including metastatic potential, Fidler has demonstrated the usefulness of orthotopic implantation human RCC in the kidney of nude mice [76]. Fidler has implanted human RCC cells, obtained from surgical specimens, into different organs of nude mice and showed that growth at different organs selected for different subpopulations of human RCC cells. However, the cells did not metastasize unless they were implanted orthotopically in the kidney [76]. The injection of human RCC cells into the kidney produces the highest incidence of tumorigenicity, suggesting that the kidney is a better environment than the skin, spleen or peritoneum. Some of these tumors spontaneously metastasize to the lung, pancreas, diaphragm and mesenteric lymph nodes [76]. Tumor cells were isolated from the kidney tumors to establish RCC cell lines in culture and study their biological activity in vitro and in vivo [76]. Other studies have confirmed the tumorigenicity and metastatic potential of human RCC cells obtained from cell lines or surgical specimens [77, 78]. Not every cell line or cells from different surgical specimens are tumorigenic. However, established human RCC tumors in the mouse kidney usually showed histological characteristics of the original primary tumor, including positivity to cytokeratins and vimentin [78, 79]. Orthotopic xenograft RCC models were used to test new anticancer drugs (BCH-4556) or gene therapy [80–82]. Human RCC cells transfected with IFNb did not induce localized tumors, kidney tumors and spontaneous lung metastases when injected in the kidney, or s.c., or i.v. [81]. These IFNb-transfected RCC tumor cells stimulated a high level of nitric oxide production by murine macrophages in vitro and in vivo that was cytotoxic to tumor cells [81]. This study led to testing retroviral vectors encoding murine macrophage inducible nitric oxide synthase (iNOS) gene for the treatment of RCC [82]. Infection of metastatic human RCC cells with iNOS retrovirus decreased their tumor growth and metastasis when injected i.v. or in the kidney of nude mice by producing nitric oxide that enhanced the rate of apoptosis of the tumor cells [82]. More recently, treatment of s.c. or i.p.

298

G.G. Hillman

human clear RCC xenografts with intratumoral injections of capsid-modified oncolytic adenoviruses reduced tumor growth and increased survival [83]. We have developed a new xenograft RCC experimental tumor model in nude mice. We have established a tumor cell line designated KCI-18 RCC from a primary renal tumor specimen obtained from a patient with papillary RCC (nuclear grade III/IV) [79]. Chromosome analysis of KCI-18 RCC cell line (passage 8 in vitro) revealed a hypertriploid karyotype with multiple clonal aberrations: 75–85, XX, -X, add (1) (p36) x2, +2, +3, +5, +i (5) (q10), +6, +del (7) (q11), +8, der (9;14) (q10;q10), -9, +10, +12, -14, add (15) (q26), +16, +17, +21, +mar, ace [cp6]. Cells were injected into the kidney (intra kidney, IK) of Balb/c nude mice. Renal tumors were resected, recultured in vitro, and repassaged in the kidney in vivo to produce KCI-18/IK cell lines [79]. Cells from the KCI-18/IK lines preserved the karyotype of the original KCI-18 RCC cell line. The KCI-18/IK lines were highly tumorigenic, and grew in the kidney with faster kinetics than the original KCI-18 RCC cell line. Tumor nodules of 0.2 cm were detectable on the kidney by day 11 postcell injection and grew to large tumors of 2–2.5 cm invading the kidney by day 44. Metastases were detectable by day 37 and observed mostly in the lungs and occasionally in the liver [79]. Mouse survival was about 44–50 days. The tumor presence in kidney and lungs was histologically confirmed and defined as a high-grade carcinoma with a sinusoidal vascular pattern. Tumor cells were characterized by large pleiomorphic nuclei, prominent nucleoli and abundant eosinophilic cytoplasm. Tumor cells co-expressed cytokeratin and vimentin. These features resembled those of the original human tumor specimen [79]. These studies demonstrate that orthotopic models of heterotransplanted human RCC cells are representative of human RCC disease progression, and preserve the karyotypic and histological characteristics of the human primary tumors. These models can be used to study human RCC progression and metastasis and in vivo responsiveness of human RCC tumors to treatment. The KCI-18 tumor model was used to test if the soy isoflavone genistein promotes metastasis but when combined with tumor irradiation could enhance radiotherapy. We had previously reported the potentiation of radiotherapy by pure genistein for prostate cancer in  vitro and in orthotopic prostate tumor models in vivo [84]. However, when genistein was used as single therapy in murine prostate tumor models, it promoted metastasis to regional para-aortic lymph nodes [84]. To clarify whether these intriguing adverse effects of genistein are intrinsic to the orthotopic prostate tumor model, we tested the effect of pure genistein alone or combined with kidney tumor irradiation in our KCI-18 RCC model [79]. Treatment of established kidney tumors with genistein demonstrated a tendency to stimulate the growth of the primary kidney tumor and increase the incidence of metastasis to the mesentery lining the bowel. In contrast, when given in conjunction with kidney tumor irradiation, genistein significantly inhibited the growth and progression of established kidney tumors. These findings confirm that, in an orthotopic RCC model, genistein treatment alone promotes metastasis, whereas it acts as a radiosensitizer when combined with tumor irradiation, similar to its effects in orthotopic prostate cancer models [79, 84, 85].

12  Animal Models for Kidney Cancer

299

In a novel and recently published study, we demonstrated that orthotopic RCC xenograft models are also useful to investigate the effect of anti-angiogenic drugs on tumor vasculature using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) [86]. Human RCC tumors are highly vascularized and recent developments in anti-angiogenic therapy have improved the targeting of RCC metastatic disease. The drug sunitinib is a small molecule receptor tyrosine kinases (RTK) inhibitor that has demonstrated antitumor and anti-angiogenic activities in mouse xenograft models [87–89]. Sunitinib has been particularly effective in recent clinical trials with an objective response rate of 47% and 11 months median progression-free survival [90–92]. However, long-term control of the disease is still not achieved and adverse effects of cardiotoxicity were documented in some of the patients, probably as a result of alterations to normal vasculature [93]. To improve the efficacy of sunitinib and decrease its impact on healthy vital organs, we investigated the effect of lower and less toxic doses of sunitinib on tumor vasculature in the KCI-18 RCC model [86]. Histological observations of KCI-18 kidney tumors showed the typical morphology of highly vascularized human RCC tumors, exhibiting abnormal vessels that are enlarged, disorganized and leaky as a result of defective basement membrane [86]. These structural defects of tumor vessels cause increased interstitial tissue pressure, impaired blood supply and decreased oxygen supply in tumors compromising the delivery and efficacy of cytotoxic drugs and radiotherapy [94, 95]. The challenge is to develop imaging technologies to monitor early vascular changes and induction of tumor vasculature “normalization,” namely regularization of vessels and blood flow, by anti-angiogenic drugs for scheduling cytotoxic therapy. Imaging of KCI-18 kidney tumor-bearing mice using DCE-MRI was successful at evaluation of vascular changes induced by various doses of sunitinib [86]. These vascular changes were observed in the right tumor-bearing kidney, in which KCI-18 cells were implanted, as well as in the left contralateral normal kidney by DCE-MRI. Sunitinib, at a suboptimal but effective dose of 20 mg/kg/day, caused increased tumor perfusion and decreased vascular permeability associated with thinning and regularization of tumor vessels while mildly affecting normal vessels in the normal contralateral kidney [86]. Sunitinib inhibited KCI-18 kidney tumor growth and caused modulation of its molecular targets VEGFR-2 and PDGFR-b, which are important receptors involved in signaling pathways of angiogenesis and found to be highly expressed in human RCC [86]. These preclinical studies emphasize the potential of human xenograft RCC tumors to assess the clinical use of DCE-MRI for selecting the dose and schedule of anti-angiogenic compounds to combine with cytotoxic therapies. Other anti-angiogenic compounds or approaches were also recently investigated in different xenograft preclinical RCC models. RCC is characterized by the loss of VHL tumor suppressor gene resulting in dysregulation of growth factor signaling including VEGF, PDGF and Raf pathways. Sorafenib, a multikinase inhibitor drug, also targets RTKs including VEGFR-1, -2, and -3 and PDGFR-b and serine/threonine kinases including Raf-1 and B-Raf that are molecules involved in angiogenesis and tumor progression. Sorafenib inhibited angiogenesis and tumor growth in both Renca (VHL+/+) and human 786-O (VHL-/-) RCC xenografts in athymic nude mice, confirming the key role of angiogenesis in supporting RCC tumor growth [96].

300

G.G. Hillman

These findings corroborate previous studies demonstrating that drugs which disrupt established tumor blood vessels (ZD6126) and drugs which interfere with neovascularization (ZD6474) are effective at causing tumor growth delays in the human clear RCC Caki-1 s.c. xenograft model [97]. Blockade of VEGFR-2 signaling pathway by TSU-68 (SU6668) inhibitor caused inhibition of kidney tumor growth but not of bone tumor xenografts generated in nude mice using the human RBM1-IT4 RCC cell line, obtained from an RCC bone lesion, that has mutated VHL [98]. The role of VHL gene in angiogenesis is further emphasized in a study showing that hypomethylating agents causing re-expression of VHL gene in human UOK 121 clear RCC s.c. xenografts reduced tumor growth [99]. A novel metastatic xenograft tumor model expressing a defined human tumor antigen was established from a primary tumor of a patient with sporadic type 2 chromophil (papillary) RCC who responded well to IL-2 immunotherapy [100]. Serial passages in SCID mice under the renal capsule generated a metastatic RCC model LABAZ1 expressing the membrane-bound tumor-associated antigen carbonic anhydrase type 9 (CA IX) [100]. This model is valuable to study the role of CA IX tumor antigen in RCC progression and metastasis [100]. In conclusion, animal models generated by xenograft implantation of human tumor RCC cell lines in immunodeficient mice have proven to be adequate for investigating novel chemotherapeutic or angiogenic drugs, gene therapy and imaging of human RCC tumors in vivo. The disadvantages of using heterotransplants in nude mice or SCID mice reside in their deficient immune system, which does not allow the study of T cell responses and induction of antitumor-specific immune responses. For demonstrating efficacy of cancer vaccines and immunotherapybased approaches, the Renca model in immunocompetent mice has been valuable in our studies and others. New generations of syngeneic animal models that express human tumor antigens could also be promising for studying these therapeutic approaches, dependent on an integral immune system. The Eker rat model has been helpful to investigate the genetic basis and molecular events involved in the development of hereditary RCC and secondary tumors. Acknowledgments  We thank Dr. A. Konski, Chairman of Radiation Oncology Dept., for supporting our studies. We thank Dr. F. Sarkar, Dr. M. Haacke, Dr. V. Singh-Gupta, Dr. J. Raffoul, Dr. H. Zhang, C. Yunker, A. Al-Bashir and A. Patel for excellent contribution to our new studies. Our studies were supported by grants from the Pardee Foundation, American Institute for Cancer Research, American Cancer Society, The Fund for Cancer Research, and Karmanos Cancer Institute; and grants from the companies Transgene SA (Strasbourg, France), Antigen-Express, Inc. (MA, USA), and Pfizer, Inc. (NY, USA).

References 1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2008. CA Cancer J Clin. 2008;58:71–96. 2. Mulders P, Figlin R, deKernion JB, et al. Renal cell carcinoma: recent progress and future directions. Cancer Res. 1997;57:5189–95.

12  Animal Models for Kidney Cancer

301

3. Chow WH, Gridley G, McLaughlin JK, et al. Protein intake and risk of renal cell cancer. J Natl Cancer Inst. 1994;86:1131–9. 4. Haas GP, Hillman GG. Update on the role of immunotherapy in the management of kidney cancer. Cancer Control. 1996;3:536–41. 5. Motzer RJ, Bander NH, Nanus DM. Renal-cell carcinoma. N Engl J Med. 1996;335:865–75. 6. Flanigan RC, Salmon SE, Blumenstein BA, et al. Nephrectomy followed by interferon alfa2b compared with interferon alfa-2b alone for metastatic renal-cell cancer. N Engl J Med. 2001;345:1655–9. 7. Mickisch GH, Garin A, van Poppel H, et al. Radical nephrectomy plus interferon-alfa-based immunotherapy compared with interferon alfa alone in metastatic renal-cell carcinoma: a randomised trial. Lancet. 2001;358:966–70. 8. Hillman GG, Droz JP, Haas GP. Experimental animal models for the study of therapeutic approaches in renal cell carcinoma. In Vivo. 1994;8:77–80. 9. Hillman GG. Experimental animal models for renal cell carcinoma. In Tumor Models in Cancer Research. Edited by: Teicher BA. Totowa, NJ, Humana Press; 2002:493–505. 10. Hrushesky WJ, Murphy GP. Investigation of a new renal tumor model. J Surg Res. 1973;15:327–36. 11. Salup RR, Wiltrout RH. Adjuvant immunotherapy of established murine renal cancer by interleukin 2-stimulated cytotoxic lymphocytes. Cancer Res. 1986;46:3358–63. 12. Salup RR, Back TC, Wiltrout RH. Successful treatment of advanced murine renal cell cancer by bicompartmental adoptive chemoimmunotherapy. J Immunol. 1987;138:641–7. 13. Huben RP, Connelly R, Goldrosen MH, et  al. Immunotherapy of a murine renal cancer. J Urol. 1983;129:1075–8. 14. Hillman GG, Haas GP, Wahl WH, Callewaert DM. Adoptive immunotherapy of cancer: biological response modifiers and cytotoxic cell therapy. Biotherapy. 1992;5:119–29. 15. Wiltrout RH, Gregorio TA, Fenton RG, et al. Cellular and molecular studies in the treatment of murine renal cancer. Semin Oncol. 1995;22:9–16. 16. Holmund JT, Kopp WC, Wiltrout RH, et al. A Phase I clinical trial of flavone-8-acetic acid in combination with interleukin 2. J Natl Cancer Inst. 1995;87:134–6. 17. Gregorian SK, Battisto JR. Immunosuppression in murine renal cell carcinoma. I. Characterization of extent, severity and sources. Cancer Immunol Immunother. 1990;31:325–34. 18. Gregorian SK, Battisto JR. Immunosuppression in murine renal cell carcinoma. II. Identification of responsible lymphoid cell phenotypes and examination of elimination of suppression. Cancer Immunol Immunother. 1990;31:335–41. 19. Dybal EJ, Haas GP, Maughan RL, et al. Synergy of radiation therapy and immunotherapy in murine renal cell carcinoma. J Urol. 1992;148:1331–7. 20. Chakrabarty A, Hillman GG, Maughan RL, et al. Radiation therapy enhances the therapeutic effect of immunotherapy on pulmonary metastases in a murine renal adenocarcinoma model. In Vivo. 1994;8:25–31. 21. Chakrabarty A, Hillman GG, Maughan RL, et al. Influence of tumor site on the therapy of murine kidney cancer. Anticancer Res. 1994;14:373–8. 22. Younes E, Haas GP, Dezso B, et al. Radiation-induced effects on murine kidney tumor cells: role in the interaction of local irradiation and immunotherapy. J Urol. 1995;153:2029–33. 23. Younes E, Haas GP, Dezso B, et al. Local tumor irradiation augments the response to IL-2 therapy in a murine renal adenocarcinoma. Cell Immunol. 1995;165:243–51. 24. Dezso B, Haas GP, Hamzavi F, et al. The mechanism of local tumor irradiation combined with interleukin 2 therapy in murine renal carcinoma: histological evaluation of pulmonary metastases. Clin Cancer Res. 1996;2:1543–52. 25. Redman BG, Hillman GG, Flaherty L, et al. Phase II trial of sequential radiation and interleukin 2 in the treatment of patients with metastatic renal cell carcinoma. Clin Cancer Res. 1998;4:283–6.

302

G.G. Hillman

26. Hillman GG, Puri RK, Kukuruga MA, et al. Growth and major histocompatibility antigen expression regulation by IL-4, interferon-gamma (IFN-gamma) and tumour necrosis factoralpha (TNF-alpha) on human renal cell carcinoma. Clin Exp Immunol. 1994;96:476–83. 27. Hillman GG, Younes E, Visscher D, et  al. Systemic treatment with interleukin-4 induces regression of pulmonary metastases in a murine renal cell carcinoma model. Cell Immunol. 1995;160:257–63. 28. Younes E, Haas GP, Visscher D, et al. Intralesional treatment of established murine primary renal tumor with interleukin-4: localized effect on primary tumor with no impact on metastases. J Urol. 1995;153:490–3. 29. Hillman GG, Younes E, Visscher D, et al. Inhibition of murine renal carcinoma pulmonary metastases by systemic administration of interferon gamma: mechanism of action and potential for combination with interleukin 4. Clin Cancer Res. 1997;3:1799–806. 30. Huang J, Wang Y, Guo J, et al. Radiation-induced apoptosis along with local and systemic cytokine elaboration is associated with DC plus radiotherapy-mediated renal cell tumor regression. Clin Immunol. 2007;123:298–310. 31. Kato Y, Yoshimura K, Shin T, et  al. Synergistic in  vivo antitumor effect of the histone deacetylase inhibitor MS-275 in combination with interleukin 2 in a murine model of renal cell carcinoma. Clin Cancer Res. 2007;13:4538–46. 32. Golumbek PT, Lazenby AJ, Levitsky HI, et  al. Treatment of established renal cancer by tumor cells engineered to secrete interleukin-4. Science. 1991;254:713–6. 33. Brunda MJ, Luistro L, Warrier RR, et al. Antitumor and antimetastatic activity of interleukin 12 against murine tumors. J Exp Med. 1993;178:1223–30. 34. Teicher BA, Ara G, Menon K, Schaub RG. In vivo studies with interleukin-12 alone and in combination with monocyte colony-stimulating factor and/or fractionated radiation treatment. Int J Cancer. 1996;65:80–4. 35. Wigginton JM, Komschlies KL, Back TC, et al. Administration of interleukin 12 with pulse interleukin 2 and the rapid and complete eradication of murine renal carcinoma. J Natl Cancer Inst. 1996;88:38–43. 36. Wigginton JM, Kuhns DB, Back TC, et  al. Interleukin 12 primes macrophages for nitric oxide production in  vivo and restores depressed nitric oxide production by macrophages from tumor-bearing mice: implications for the antitumor activity of interleukin 12 and/or interleukin 2. Cancer Res. 1996;56:1131–6. 37. Tannenbaum CS, Wicker N, Armstrong D, et  al. Cytokine and chemokine expression in tumors of mice receiving systemic therapy with IL-12. J Immunol. 1996;156:693–9. 38. Tannenbaum CS, Tubbs R, Armstrong D, et  al. The CXC chemokines IP-10 and Mig are necessary for IL-12-mediated regression of the mouse RENCA tumor. J Immunol. 1998;161:927–32. 39. Bukowski RM, Rayman P, Molto L, et al. Interferon-gamma and CXC chemokine induction by interleukin 12 in renal cell carcinoma. Clin Cancer Res. 1999;5:2780–9. 40. Sayers TJ, Brooks AD, Lee JK, et al. Molecular mechanisms of immune-mediated lysis of murine renal cancer: differential contributions of perforin-dependent versus Fas-mediated pathways in lysis by NK and T cells. J Immunol. 1998;161:3957–65. 41. Lee JK, Sayers TJ, Brooks AD, et al. IFN-gamma-dependent delay of in vivo tumor progression by Fas overexpression on murine renal cancer cells. J Immunol. 2000;164:231–9. 42. Shvarts O, Janzen N, Lam JS, et al. RENCA/carbonic anhydrase-IX: a murine model of a carbonic anhydrase-IX-expressing renal cell carcinoma. Urology. 2006;68:1132–8. 43. Hara I, Nagai H, Miyake H, et al. Effectiveness of cancer vaccine therapy using cells transduced with the interleukin-12 gene combined with systemic interleukin-18 administration. Cancer Gene Ther. 2000;7:83–90. 44. Nishisaka N, Maini A, Kinoshita Y, et  al. Immunotherapy for lung metastases of murine renal cell carcinoma: synergy between radiation and cytokine-producing tumor vaccines. J Immunother. 1999;22:308–14.

12  Animal Models for Kidney Cancer

303

45. Siders WM, Wright PW, Hixon JA, et  al. T cell- and NK cell-independent inhibition of hepatic metastases by systemic administration of an IL-12-expressing recombinant adenovirus. J Immunol. 1998;160:5465–74. 46. Hillman GG, Slos P, Wang Y, et al. Tumor irradiation followed by intratumoral cytokine gene therapy for murine renal adenocarcinoma. Cancer Gene Ther. 2004;11:61–72. 47. Hillman GG, Kallinteris NL, Li J, et  al. Generating MHC Class II+/Ii- phenotype after adenoviral delivery of both an expressible gene for MHC Class II inducer and an antisense Ii-RNA construct in tumor cells. Gene Ther. 2003;10:1512–8. 48. Hillman GG, Kallinteris NL, Lu X, et  al. Turning tumor cells in situ into T-helper cellstimulating, MHC class II tumor epitope-presenters: immuno-curing and immuno-consolidation. Cancer Treat Rev. 2004;30:281–90. 49. Tan J, Newton CA, Djeu JY, et al. Injection of complementary DNA encoding interleukin-12 inhibits tumor establishment at a distant site in a murine renal carcinoma model. Cancer Res. 1996;56:3399–403. 50. Watanabe M, Fenton RG, Wigginton JM, et al. Intradermal delivery of IL-12 naked DNA induces systemic NK cell activation and Th1 response in vivo that is independent of endogenous IL-12 production. J Immunol. 1999;163:1943–50. 51. Saffran DC, Horton HM, Yankauckas MA, et al. Immunotherapy of established tumors in mice by intratumoral injection of interleukin-2 plasmid DNA: induction of CD8+ T-cell immunity. Cancer Gene Ther. 1998;5:321–30. 52. Miyake H, Hara I, Yoshimura K, et al. Introduction of basic fibroblast growth factor gene into mouse renal cell carcinoma cell line enhances its metastatic potential. Cancer Res. 1996;56:2440–5. 53. Miyake H, Hara I, Gohji K, et al. Relative expression of matrix metalloproteinase-2 and tissue inhibitor of metalloproteinase-2 in mouse renal cell carcinoma cells regulates their metastatic potential. Clin Cancer Res. 1999;5:2824–9. 54. Kundu SD, Kim IY, Zelner D, et al. Absence of expression of transforming growth factorbeta type II receptor is associated with an aggressive growth pattern in a murine renal carcinoma cell line, Renca. J Urol. 1998;160:1883–8. 55. Engel JD, Kundu SD, Yang T, et al. Transforming growth factor-beta type II receptor confers tumor suppressor activity in murine renal carcinoma (Renca) cells. Urology. 1999;54:164–70. 56. Yoon SS, Eto H, Lin CM, et al. Mouse endostatin inhibits the formation of lung and liver metastases. Cancer Res. 1999;59:6251–6. 57. 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;13:4201–8. 58. Clayman RV, Bilhartz LE, Buja LM, et al. Renal cell carcinoma in the Wistar–Lewis rat: a model for studying the mechanisms of cholesterol acquisition by a tumor in vivo. Cancer Res. 1986;46:2958–63. 59. Eker R. Familial renal adenomas in Wistar rats: a preliminary report. Acta Pathol Microbiol Scand. 1954;34:554–62. 60. Everitt JI, Goldsworthy TL, Wolf DC, Walker CL. Hereditary renal cell carcinoma in the Eker rat: a unique animal model for the study of cancer susceptibility. Toxicol Lett. 1995;82–83:621–5. 61. Walker C, Goldsworthy TL, Wolf DC, Everitt J. Predisposition to renal cell carcinoma due to alteration of a cancer susceptibility gene. Science. 1992;255:1693–5. 62. Hino O, Klein-Szanto AJ, Freed JJ, et al. Spontaneous and radiation-induced renal tumors in the Eker rat model of dominantly inherited cancer. Proc Natl Acad Sci U S A. 1993;90:327–31. 63. Yeung RS, Xiao GH, Jin F, et al. Predisposition to renal carcinoma in the Eker rat is determined by germ-line mutation of the tuberous sclerosis 2 (TSC2) gene. Proc Natl Acad Sci U S A. 1994;91:11413–16.

304

G.G. Hillman

64. Kobayashi T, Hirayama Y, Kobayashi E, et al. A germline insertion in the tuberous sclerosis (Tsc2) gene gives rise to the Eker rat model of dominantly inherited cancer. Nat Genet. 1995;9:70–4. 65. Kubo Y, Klimek F, Kikuchi Y, et al. Early detection of Knudson’s two-hits in preneoplastic renal cells of the Eker rat model by the laser microdissection procedure. Cancer Res. 1995;55:989–90. 66. Bjornsson J, Short MP, Kwiatkowski DJ, Henske EP. Tuberous sclerosis-associated renal cell carcinoma. Clinical, pathological, and genetic features. Am J Pathol. 1996;149:1201–8. 67. Jin F, Wienecke R, Xiao GH, et al. Suppression of tumorigenicity by the wild-type tuberous sclerosis 2 (Tsc2) gene and its C-terminal region. Proc Natl Acad Sci U S A. 1996;93:9154–9. 68. Orimoto K, Tsuchiya H, Sakurai J, et al. Identification of cDNAs induced by the tumor suppressor Tsc2 gene using a conditional expression system in Tsc2 mutant (Eker) rat renal carcinoma cells. Biochem Biophys Res Commun. 1998;247:728–33. 69. Kobayashi T, Minowa O, Kuno J, et al. Renal carcinogenesis, hepatic hemangiomatosis, and embryonic lethality caused by a germ-line Tsc2 mutation in mice. Cancer Res. 1999;59:1206–11. 70. Walker C, Ahn YT, Everitt J, Yuan X. Renal cell carcinoma development in the rat independent of alterations at the VHL gene locus. Mol Carcinog. 1996;15:154–61. 71. Wolf DC, Goldsworthy TL, Donner EM, et al. Estrogen treatment enhances hereditary renal tumor development in Eker rats. Carcinogenesis. 1998;19:2043–7. 72. Miki S, Miki Y, Mitani H, et al. Hyperlipidemia enhancement of renal preneoplasia but not tumor formation is due to elevated apoptosis in the Eker rat. Cancer Res. 1997;57:4673–6. 73. Ebert T, Bander NH, Finstad CL, et al. Establishment and characterization of human renal cancer and normal kidney cell lines. Cancer Res. 1990;50:5531–6. 74. van Dijk J, Zegveld ST, Fleuren GJ, Warnaar SO. Localization of monoclonal antibody G250 and bispecific monoclonal antibody CD3/G250 in human renal-cell carcinoma xenografts: relative effects of size and affinity. Int J Cancer. 1991;48:738–43. 75. Singh M, Ghose T, Mezei M, Belitsky P. Inhibition of human renal cancer by monoclonal antibody targeted methotrexate-containing liposomes in an ascites tumor model. Cancer Lett. 1991;56:97–102. 76. Fidler IJ. Orthotopic implantation of human colon carcinomas into nude mice provides a valuable model for the biology and therapy of metastasis. Cancer Metastasis Rev. 1991;10:229–43. 77. Grossi FS, Zhao X, Romijn JC, et al. Metastatic potential of human renal cell carcinoma: experimental model using subrenal capsule implantation in athymic nude mice. Urol Res. 1992;20:303–6. 78. Beniers AJ, Peelen WP, Schaafsma HE, et al. Establishment and characterization of five new human renal tumor xenografts. Am J Pathol. 1992;140:483–95. 79. Hillman GG, Wang Y, Che M, et  al. Progression of renal cell carcinoma is inhibited by genistein and radiation in an orthotopic model. BMC Cancer. 2007;7:4. 80. Kadhim SA, Bowlin TL, Waud WR, et al. Potent antitumor activity of a novel nucleoside analogue, BCH-4556 (beta-l-dioxolane-cytidine), in human renal cell carcinoma xenograft tumor models. Cancer Res. 1997;57:4803–10. 81. Xie K, Bielenberg D, Huang S, et al. Abrogation of tumorigenicity and metastasis of murine and human tumor cells by transfection with the murine IFN-beta gene: possible role of nitric oxide. Clin Cancer Res. 1997;3:2283–94. 82. Juang SH, Xie K, Xu L, et al. Suppression of tumorigenicity and metastasis of human renal carcinoma cells by infection with retroviral vectors harboring the murine inducible nitric oxide synthase gene. Hum Gene Ther. 1998;9:845–54. 83. Guse K, Ranki T, Ala-Opas M, et  al. Treatment of metastatic renal cancer with capsidmodified oncolytic adenoviruses. Mol Cancer Ther. 2007;6:2728–36. 84. Hillman GG, Wang Y, Kucuk O, et al. Genistein potentiates inhibition of tumor growth by radiation in a prostate cancer orthotopic model. Mol Cancer Ther. 2004;3:1271–9.

12  Animal Models for Kidney Cancer

305

85. Raffoul JJ, Banerjee S, Singh-Gupta V, et  al. Down-regulation of apurinic/apyrimidinic endonuclease 1/redox factor-1 expression by soy isoflavones enhances prostate cancer radiotherapy in vitro and in vivo. Cancer Res. 2007;67:2141–9. 86. Hillman GG, Singh-Gupta V, Zhang H, et  al. Dynamic contrast-enhanced magnetic resonance imaging of vascular changes induced by sunitinib in papillary renal cell carcinoma xenograft tumors. Neoplasia. 2009;11:910–20. 87. Mendel DB, Laird AD, Xin X, et al. In vivo antitumor activity of SU11248, a novel tyrosine kinase inhibitor targeting vascular endothelial growth factor and platelet-derived growth factor receptors: determination of a pharmacokinetic/pharmacodynamic relationship. Clin Cancer Res. 2003;9:327–37. 88. Xu L, Tong R, Cochran DM, Jain RK. Blocking platelet-derived growth factor-D/plateletderived growth factor receptor beta signaling inhibits human renal cell carcinoma progression in an orthotopic mouse model. Cancer Res. 2005;65:5711–9. 89. Abrams TJ, Lee LB, Murray LJ, et  al. SU11248 inhibits KIT and platelet-derived growth factor receptor beta in preclinical models of human small cell lung cancer. Mol Cancer Ther. 2003;2:471–8. 90. Motzer RJ, Hutson TE, Tomczak P, et al. Sunitinib versus interferon alfa in metastatic renalcell carcinoma. N Engl J Med. 2007;356:115–24. 91. Motzer RJ, Michaelson MD, Redman BG, et al. Activity of SU11248, a multitargeted inhibitor of vascular endothelial growth factor receptor and platelet-derived growth factor receptor, in patients with metastatic renal cell carcinoma. J Clin Oncol. 2006;24:16–24. 92. Motzer RJ, Hutson TE, Tomczak P, et al. Overall survival and updated results for sunitinib compared with interferon alfa in patients with metastatic renal cell carcinoma. J Clin Oncol. 2009;27:3584–90. 93. Schmidinger M, Zielinski CC, Vogl UM, et al. Cardiac toxicity of sunitinib and sorafenib in patients with metastatic renal cell carcinoma. J Clin Oncol. 2008;26:5204–12. 94. Jain RK. Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science. 2005;307:58–62. 95. Jain RK. Normalizing tumor vasculature with anti-angiogenic therapy: a new paradigm for combination therapy. Nat Med. 2001;7:987–9. 96. Chang YS, Adnane J, Trail PA, et al. Sorafenib (BAY 43-9006) inhibits tumor growth and vascularization and induces tumor apoptosis and hypoxia in RCC xenograft models. Cancer Chemother Pharmacol. 2007;59:561–74. 97. Siemann DW, Shi W. Efficacy of combined antiangiogenic and vascular disrupting agents in treatment of solid tumors. Int J Radiat Oncol Biol Phys. 2004;60:1233–40. 98. Karashima T, Inoue K, Fukata S, et al. Blockade of the vascular endothelial growth factorreceptor 2 pathway inhibits the growth of human renal cell carcinoma, RBM1-IT4, in the kidney but not in the bone of nude mice. Int J Oncol. 2007;30:937–45. 99. Alleman WG, Tabios RL, Chandramouli GV, et  al. The in vitro and in vivo effects of reexpressing methylated von Hippel–Lindau tumor suppressor gene in clear cell renal carcinoma with 5-aza-2¢-deoxycytidine. Clin Cancer Res. 2004;10:7011–21. 100. Zisman A, Pantuck AJ, Bui MH, et al. LABAZ1: a metastatic tumor model for renal cell carcinoma expressing the carbonic anhydrase type 9 tumor antigen. Cancer Res. 2003;63:4952–9.

Chapter 13

Animal Models of Mesothelioma Harvey I. Pass, Joseph B. Pincus, Michele Carbone, and Magdalena Plasilova

Abstract  Animal models for mesothelioma mainly stem from exposure of different species to environmental fibers that are carcinogenic. The majority take months for a mesothelioma to develop in the abdomen or the chest. The models are invaluable, however, for studying molecular pathways. Other models in transgenic animals have used novel promoters or other molecules involved in mesothelioma generation to produce the disease. Keywords  Mesothelioma • Rodents • Asbestos • Fibers

13.1 Introduction Mesotheliomas are tumors that originate from the serosal lining of the pleural, pericardial, and peritoneal cavities [1]. Malignant pleural mesotheliomas (MPMs) are among the most aggressive human tumors, with a survival of less than 1 year after diagnosis. None of the current therapeutic approaches has been shown to alter the natural history of this disease [2]. The continued rise in the incidence of mesotheliomas in the United States and abroad (2,000–3,000 cases per year in the United States) has been related to the widespread commercial use of asbestos during the last 50 years and its latency in manifestation [3–5]. The exact biochemical mechanism that causes asbestos to induce mesothelioma is unclear [6–10]. In tissue culture, asbestos fibers can cause mutagenic events, including DNA-strand breaks and deletion mutations through the production of hydroxyl radicals and superoxide anions. They can also alter chromosome morphology and ploidy by mechanically interfering with their segregation during mitosis. Furthermore, macrophages produce DNA-damaging oxyradicals H.I. Pass (*) Department of Cardiothoracic Surgery, New York University Medical Center, Suite 9V, 530 First Avenue, New York, NY 10016, 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_13, © Springer Science+Business Media, LLC 2011

307

308

H.I. Pass et al.

following phagocytosis of asbestos fibers, and will elaborate lymphokines that may depress the host immune response. Although the association between asbestos and mesothelioma is indisputable, less than 10% of the people exposed to high doses of asbestos will develop ­mesotheliomas. Moreover, approximately one-half of the reported cases of ­mesothelioma have no documented exposure to asbestos [10]. In areas where asbestos mines exist, more than 70% of mesothelioma patients have a positive exposure history, but in areas with no substantial asbestos-using industries, as few as 10% of patients with mesothelioma have a positive exposure history. These findings and reports of mesothelioma developing in childhood – or even in utero – as well as the continued description of the disease in individuals too young to have the usual latency period (25–40 years) from asbestos exposure to MPM development suggest that there are probably other unknown factors involved in the pathogenesis of mesothelioma. Thus, researchers have long sought additional carcinogens that may be responsible for mesotheliomas in nonasbestos-exposed individuals, or that could render particular individuals more susceptible to the carcinogenic effect of asbestos. Nonasbestos fibers and ionizing radiation have been described to cause the malignant transformation of mesothelial cells. This chapter focuses on the development of animal models for mesothelioma. Since conventional wisdom dictates that the etiology of most mesotheliomas is asbestos, a large part of the discussion will focus on animal models of ­mesothelioma using asbestos fibers. However, since there is interest in viral-induced mechanisms for mesothelioma in humans, viral-induced mesotheliomas in hamster and murine models are also discussed. Finally, orthotopic transplantation of established ­mesothelioma cell lines in various positions in immunocompromised animals is described. The individual sections of the chapter discuss the individual means by which the tumor can be induced or propagated in the animal. Routes of delivery and species specificity are highlighted.

13.2 Asbestos-Induced Mesothelioma in Animal Models 13.2.1 Types of Asbestos Fibers There are two families of asbestos fibers: serpentine (chrysotile), which is curly and pliable, and the rod-like amphiboles, which include crocidolite, amosite, and anthrophylite, tremolite, and actinolyte [11]. The size and thickness of these fibers will vary considerably, and the carcinogenic effects are related to their physical characteristics, with greater tumor promotion seen in the long, thin fibers that are readily phagocytosed and are stable in tissue [12]. Inhaled fibers are cleared from the tracheobronchial tree by macrophages and ciliary action, and the remaining fibers accumulate in the lower one-third of the lungs, adjacent to the visceral pleura [13]. As opposed to crocidolite and other amphiboles, the serpentine fibers break

13  Animal Models of Mesothelioma

309

down into smaller subunits that deteriorate or dissolve in the lung, and are then eliminated by lymphatics. The relative carcinogenicity of the fibers remains controversial. It is safe to say that in humans there seems to be a “gradient” from crocidolite to chrysotile in potency in order to induce mesothelioma and – despite the ability of chrysotile to induce mesothelioma in laboratory animals – it is generally agreed that crocidolite is of greater carcinogenicity in humans [14]. Nevertheless, chrysotile constitutes the majority (90%) of the asbestos used worldwide, and there are marked differences in the disease incidence in chrysotile workers [15–17]. Brake-lining workers chiefly exposed to chrysotile fibers and other automobile maintenance workers are estimated to represent 20,000 deaths from asbestos-related cancer over the next 40 years. This finding may be a function of difference in fiber size or contamination with crocidolite fibers. The necessity to delineate the risk of chrysotile-induced mesothelioma is even clearer when one considers that most buildings use this fiber for cement products and insulation. Attempts at delineating the dose intensity and relation of fiber type to mesothelioma development have been performed using fiber analysis of tissues. This fiber content, measured by light and electron microscopic techniques, depends on the amount of fiber deposition (a function of the duration and dose intensity) as well as fiber clearance. Recent data suggest that the fiber concentration exceeding 1 million fibers per gram of dried tissue may be associated with increased mesothelioma risk. Moreover, epithelial mesothelioma may be associated with significantly lower fiber content than sarcomatoid mesothelioma [18].

13.2.2 Asbestos-Induced Animal Models of Mesothelioma: General Comments Animal models for mesothelioma have been developed using all types of asbestos fibers [19–22]. Just as in the human disease, the animal models have demonstrated that fiber length, diameter, shape, and durability are more important than the type of fiber. All types of asbestos can induce tumor formation in animals through intraperitoneal (i.p.), intrathoracic, or intratracheal inoculation. The observation that some fibers are more mesotheliomatogenic than others is not coincidental. There seems to be a relationship between the carcinogenicity of a specific fiber and the proportion of long, thin fibers it contains. The general consensus is that long, thin fibers over 8 µm in length and thinner than 1 µm have the strongest carcinogenic effects because they cannot be readily phagocytosed by alveolar macrophages for mechanical clearance via the mucociliary escalator. However, threshold toxicity has yet to be defined, because asbestos dust comes in variety of particle lengths and it is difficult to administer a fiber of some precise length without also injecting fibers of other lengths. As a result, most studies have attempted to characterize carcinogenic potential as a function of number of particles over or under a certain length.

310

H.I. Pass et al.

13.2.3 Intraperitoneal Asbestos Injection Intraperitoneal (i.p.) injections have the disadvantage of introducing the asbestos into the body in an unnatural manner. It has been argued that the conclusions drawn from carcinogenic studies are not applicable to humans because the physiological route of entry into the respiratory tract via the conducting airways is effectively bypassed. Thus, mesothelial cells are exposed to massive local doses of fiber that are not relevant for human exposure. Advocates of this method of induction point to it as being less labor-intensive than inhalation studies, and view it as an excellent opportunity to test the carcinogenic potential and potency of a fiber. Studies using the peritoneal mode of induction also offer the advantage of producing tumors in a large number of the injected animals.

13.2.4 Intraperitoneal Asbestos Injection: Rats Tumor induction rates vary between 56% and 97.5% in the rat mesothelioma model. A variety of naturally occurring asbestos fibers – such as crocidolite, standard chrysotile, Canadian chrysotile, Rhodesian chrysotile, and amosite – have been studied for their ability to induce mesothelioma [23, 24]. In these studies, where a fixed, notadjusted-for fiber-length dose in milligrams of a fiber has been administered, there is conflicting evidence as to whether standard chrysotile or standard crocidolite is more tumorigenic. Davis found chrysotile and amosite to be more carcinogenic than crocidolite [24]. However, a study by Minardi and Maltoni suggests that standard crocidolite is the most hazardous material used in the study, inducing mesothelioma in 97.5% of animals tested, followed by amosite (90%), Rhodesian chrysotile (82.5%), Canadian chrysotile (80%), and Californian chrysotile (72.5%) [23]. Davis demonstrated that the hazard of a particular fiber (e.g. chrysotile, amosite, crocidolite) is related to the number of long fibers it contains [24]. Thus, chrysotile with over 60 million fibers >10 µm produced mesothelioma in 68% of injected animals at the 2.5-mg dose. In contrast, amosite with only 10 million fibers >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.

13  Animal Models of Mesothelioma

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% <2.1 and <0.2 µm in diameter [19]. Light microscopy observation has revealed tumors consisting of large, polyhedral cells of the epitheloid type, with abundant amphophylic cytoplasm and sharply defined cell contours [19]. Areas of papillary formations have also been observed. Scanning electron microscopy analysis has shown individual, minute polypoid alterations, with surfaces composed of normal mesothelial cells, and polymorphic cells which varied in size and shape. The microvilli of these cells displayed clear pleomorphism.

13.3 Spontaneous Models of Mesothelioma Spontaneous mesotheliomas are generally rare in experimental animals. No spontaneous mesotheliomas have been reported in mice and hamsters. However, Fisher 344 rats have been reported to have a 3–4% natural occurrence of tumors by a number of investigators, and it is generally accepted that spontaneously occurring tumors are quite similar to those induced by asbestos. Tanigawa reported that spontaneously induced mesothelial tumors were specific to males, with an incidence of 4.3%, or 17 of 395 [30]. The majority of tumors (16 of 17) were confined to the genital serosa and peritoneum, with one tumor occurring in the pleural cavity. Expansion of the peritoneal cavity caused by ascites is typical in this disease, as is protrusion of the scrotum. Microscopically, the tumors are notable for complex papillary growth and sessile nodular growths resembling sarcomas. Immunohistochemical localization of keratin proteins and histological patterns of the tumors examined suggest that spontaneous mesothelioma essentially resembles the epithelial type of human mesothelioma.

13.4 Other Agents for Animal Production of Mesothelioma 13.4.1 Chemical Although most cases of mesothelioma can be linked to asbestos exposure, some cannot, which raises the question of whether other agents cause mesothelioma or augment the effects of asbestos [31]. Rice reports that the polycyclic aromatic hydrocarbon carcinogen 3-methylcholanthrene (MC) is capable of inducing mesothelioma when injected intragastrically in certain strains of mice. The C3H strain was the most susceptible of the six

13  Animal Models of Mesothelioma

315

strains tested, with a mesothelioma incidence of 39% (12 of 31 treated animals) and an average survival time of 8.3 months. The Balb/c strain had an incidence of 28% (9 of 32) and average survival of 10.4 months. Because both groups of animals sustained high rates of peritoneal injury, there is some uncertainty as to whether the mesothelial proliferation was caused primarily by the mechanical insult caused by the 12 injections received by each group. A few of the mesotheliomas were predominantly mesothelial. However, most were mixed and fibrous in nature with epithelial elements. The neoplasms grew within the peritoneal cavity over the surfaces of the viscera, and were frequently invasive of other organs, including the diaphragm. Methyl (acetoxymethyl) nitrosamine (DMN-OAc) has been established as an agent capable of inducing mesothelioma in male Sprague–Dawley and Buffalo rats. Berman and Rice report the induction of 25 testicular mesotheliomas in 78 treated rats (32%) [32]. It is notable that intraperitoneal (i.p.) administration of the chemical agent did not produce mesotheliomas in any other location except the testes, and there was no evidence of metastases or local invasion. Although some tumors contained massive amounts of stroma, there were no examples of the sarcomatoid variant of tumor found in humans among the examined mesotheliomas. All examined tumors stained positive by the colloidal iron method.

13.4.2 Man-Made Fibers Experimental studies suggest that certain man-made fibers have a greater toxicity than naturally occurring asbestos fibers. This greater mesotheliomatogenic effect is believed to be a result of the greater proportion of long fibers and their long durability. In intracavital injection studies in rats, erionite-induced mesothelioma at rates of 54.5–93% of treated animals percentages considerably higher than the incidence produced by either amosite, crocidolite, or chrysotile [20, 28]. Suzuki suggests that erionite’s toxicity is linked to its high content of long fibers (4% >9.5 µm) – more than any of the tested asbestos fibers. The median survival time for the animals treated with erionite, which can be indicative of the toxicity of a fiber, was also significantly shorter (513 days) for animals treated with erionite than for rats treated with crocidolite, chrysotile, and amosite [28]. Davis suggests that doses of less than 160,000 fibers >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-

13  Animal Models of Mesothelioma

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

13  Animal Models of Mesothelioma

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

13  Animal Models of Mesothelioma

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.

13  Animal Models of Mesothelioma

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

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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.

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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.

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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.

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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.

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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.

14  The Use of Mouse Models to Study Leukemia/Lymphoma

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

D.M. Vail and D.H. Thamm

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

15  Spontaneous Companion Animal (Pet) Cancers

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

D.M. Vail and D.H. Thamm

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

15  Spontaneous Companion Animal (Pet) Cancers

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

D.M. Vail and D.H. Thamm

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

15  Spontaneous Companion Animal (Pet) Cancers

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

D.M. Vail and D.H. Thamm

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.

15  Spontaneous Companion Animal (Pet) Cancers

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

D.M. Vail and D.H. Thamm

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

15  Spontaneous Companion Animal (Pet) Cancers

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

D.M. Vail and D.H. Thamm

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

15  Spontaneous Companion Animal (Pet) Cancers

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

D.M. Vail and D.H. Thamm

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

15  Spontaneous Companion Animal (Pet) Cancers

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

D.M. Vail and D.H. Thamm

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.

15  Spontaneous Companion Animal (Pet) Cancers

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

D.M. Vail and D.H. Thamm

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.

15  Spontaneous Companion Animal (Pet) Cancers

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

A.F. Hezel and N. Bardeesy

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

A.F. Hezel and N. Bardeesy

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-in- activated 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

A.F. Hezel and N. Bardeesy

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 (<10 weeks) and do not commonly show gross metastasis probably due to the rapid progression of the primary tumors. These features may make these models less suitable for preclinical therapeutic studies than the models with p53 heterozygosity (of the null or R273H allele) which have more gradual progression and exhibit single tumor foci. The R273H model may be particularly suitable for studying metastasis as recent studies have defined a pathway whereby this mutant p53 allele can promote invasive growth by interfering with p63 function [66]. The mouse models also exhibit some histologic variation. Human PDAC ­characteristically have a ductal histological phenotype (with a range between welland poorly differentiated features). The mouse models described above all recapitulate these histological features, although tumors with histological variants are also observed. Undifferentiated sarcomatoid histology, an uncommon histological ­variant of human pancreatic cancer, is observed in a subset of Kras-Ink4a/Arf mice, and less frequently in Kras-p53 mice. Anaplastic histology (pleomorphic tumor cells, with significant nuclear irregularities) is also a variant observed in a subset of Kras-p53 and Kras-Ink4a/Arf mice. The final highly recurrent mutation in human PDAC is inactivation of Smad4. In mouse models, homozygous inactivation of Smad4 in the pancreas does not lead to development defects or tumor predisposition, but synergistically promotes PDAC development in combination with Kras activation [61–63]. Notably, the course of

386

A.F. Hezel and N. Bardeesy

the tumorigenesis is distinct from the models described above. In particular the Kras-Smad4 mutants develop cystic tumors resembling either IPMN or MCN (Fig. 16.1). These tumors can progress to PDAC although with longer latency than the Kras-Ink4/Arf or Kras-p53 models. The PDAC that emerges appears to have sustained spontaneous Ink4a/Arf deletion, and breeding these mice onto a conditional heterozygous Ink4a/Arf background significantly accelerates progression to PDAC without influencing the IPMN phenotype. Mutation of the Type 2 TGF-beta receptor in cooperation with KrasG12D also leads to PDAC; in this model the tumors arise from progressive PanIN lesions rather than IPMN or MCN [64]. It is notable that the PDAC in the Kras-TGFRII model arise more rapidly than those in Kras-Smad4 mice (although the differences in mouse strain backgrounds in these models prevent definitive comparisons). The different phenotypes between the Smad4 and TGFRII mutant models were somewhat unanticipated since Smad4 is a central effecter of TGFbeta signaling. The differences in phenotype are likely due to the capacity of Smad4 to integrate signals from other pathways (activins, BMPs) and for TGFRII to signal independently of Smad4 [67]. In both model ­systems, the PDAC predominantly show well or moderately differentiated ductal histology, which may suggest that active TGFbeta signaling contributes to the loss of epithelial features (i.e., epithelial-to-mesenchymal transition) in PDAC.

16.6.2 Preclinical Studies In addition to their value for studying the role of individual gene mutations in PDAC pathophysiology, the models described above provide value systems for relevant preclinical studies. For example, these mouse models hold considerable promise for testing novel therapeutics, serving as potential intermediate or complimentary system between in vitro and xenograft studies and human clinical trials. Multiple approaches have been employed with varying relevance to human cancer therapeutics. Recent studies have tested the impact of a variety of agents including gamma-secretase inhibitors, COX2 inhibitors, and Hedgehog inhibitors on PanIN progression to PDAC [68–71]. In these studies treatment was initiated at defined time-points to cohorts of animals evaluated how drugs affect the onset and ­incidence of PDAC. Thus, they have provided useful information regarding the pathways that regulate PDAC initiation. It is more of a challenge to address whether a therapeutic approach has an impact on established PDAC – a situation more akin to what occurs clinically. A number of imaging methods have been shown to be effective in detecting PDAC in mouse models, including MRI, FDG-PET, and ultrasound, enabling monitoring responsiveness in the setting of advanced disease (see Fig. 16.3). In this context an intriguing recent finding is that PDAC may be refractory to treatment at least in part due to the pronounced desmoplasia associated with these tumors. In this study, it was shown that there is very poor uptake of gemcitabine in the engineered mouse models [70]. Correspondingly, there is generally a limited clinical response to ­gemcitabine.

16  Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma

387

Fig. 16.3  Detection of murine PDAC with MRI. Gadolinium enhanced MRI imaging may be used to detect focal tumors in the Pdx1-Cre LSL-KrasG12D model. Both axial and coronal images are shown here. S and L denote stomach and liver, respectively

Previous studies have provided evidence that stromal proliferation in PDAC is regulated by the Hedgehog signaling pathway. Notably, treatment with inhibitors of Smoothened, a key mediator of Hh signaling reduced PDAC stroma, resulting in increased gemcitabine uptake, transient tumor regression, and increased survival in the GEMM. Since PDAC xenografts were highly responsive to gemcitabine treatment alone and did not exhibit extensive stromal tissue, it appears that GEMM may be more a more relevant model system for predicting clinical responses. Another translational application of these models is in the development of novel approaches for early detection of PDAC. This is particularly important challenge in this disease as most patients have locally advanced or metastatic disease at diagnosis and therefore are not eligible for potentially curative resection. While the identification of serum biomarkers for early PDAC detection is currently an area of significant investigation, there is significant difficulty in obtaining suitable human specimens. The mouse models could facilitate serum biomarker discovery given their predictable course of tumor progression and the capacity to generate tumors with defined genetic alterations and on inbred backgrounds. One study used identified a high-throughput mass spectrometric method, surfaced enhanced laser desorption ionization time-of-flight (SELDI-TOF), to identify a consistent signature of altered peptides in serum from mice harboring PanIN compared to serum from wild type mice [58, 72]. Another study employed protein fractionation approach coupled to mass spectrometry analysis to identify specific proteins overexpressed in mice harboring advanced PanIN or focal PDAC [73]. This approach led to the identification of novel serum biomarkers that are also upregulated in PDAC patients, and therefore may help in the development of early detection tests. In addition to serum biomarkers, the development of imaging probes for ­incipient PDAC could also enable earlier detection. The mouse models have also provided a useful system for the identification of new targets for molecular imaging of PDAC.

388

A.F. Hezel and N. Bardeesy

In order to develop novel images probes and identify cell surface proteins that are specifically upregulated in locally advanced PDAC, early PDAC cell lines were generated from the Kras-p53 mouse model and used in subtractive phage display peptide screens using normal mouse pancreatic ductal cells in the subtraction step [74]. These screens resulted in the identification of plectin-1 as a marker specifically upregulated on the cell surface of PDAC cells. When the peptides that target plectin-1 were conjugated to magnetofluorescent nanoparticles, it was found that PDAC in GEM models and human xenograft specimens could be effectively detected in  vivo. This novel imaging probe and the discovery of plectin-1 as a novel biomarker, may have clinical utility in the diagnosis and management of PDAC in humans.

16.7 Ongoing and Future Modeling Efforts More recent modeling efforts have led to refinements in our understanding of PDAC biology and have provided a range of tractable new systems. One series of studies has begun to address (a) whether acute activation of Kras in the adult mouse pancreas can give rise to PDAC and (b) which specific cell types in the pancreas are susceptible to Kras-mediated transformation into PDAC. The pdx-1 and p48 Cre stains used initially in creating PDAC models target all cells within the pancreas. In order to answer these questions additional Cre strains utilizing specific promoters with activity in defined pancreatic subcellular compartments have been used to activate Kras. The Nestin-Cre transgene is active in acinar progenitors and, less efficiently, in duct cells, but not in islet cells or centroacinar cells. Nestin-Cre LSL-Kras animals develop a PanIN phenotype that resembles that seen in Pdx1-Cre LSL-Kras mice [75]. Adult nonductal cells have also been shown to be susceptible to Kras-mediated transformation to PanIN lesions. Specifically, activation of the Kras-G12D allele in adult acinar cells or acinar and centroacinar cells using the Mist1-CreERt and Elastase-CreERt strains, respectively, led to the rapid development of PanINs with comparable features to those seen in the Pdx1-Cre Kras model [76, 77]. Slightly different interpretations were drawn from experiments with the KrasG12V-IresLacZ allele. When expressed in the acinar cells and centroacinar cells during late embryonic development or at early postnatal time points, this allele results in rapid development of PanIN that progress to PDAC by 1 year of age [65]. On the other hand, activation of this allele in the adult pancreatic acinar and ­centroacinar cells failed to result in PanINs unless coupled with pancreatic injury induced by cerulein. While differences in genetic background, recombination ­frequency, specific Kras mutation (G12D vs. G12V), and method of Cre activation (spontaneous, doxycylcine-regulated, or Tamoxifen-induced), prevent direct ­comparisons of these models, some broad themes appear to emerge. In particular, the data indicate that the adult acinar cells, and possibly centroacinar cells, can give rise to PanIN that presumably are capable of progression to PDAC. In addition, pancreatic injury and inflammation can enhance PDAC progression. This idea is also supported by epidemiologic studies that link pancreatic injury and ­inflammation

16  Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma

389

with risk of cancer. Overall, it is clear that there is considerable developmental plasticity since acinar cells can be programed to take on an endocrine character in vivo and convert to a ductal phonotype in vitro, and acinar cells can de-differentiate to a duct-like state in vitro and in vivo [78]. Finally, there is evidence that progenitorlike cells can be induced in the pancreas following pancreatic injury [79]. Future studies using additional lineage-specific Cre strains will be needed to determine whether other pancreatic cell types (e.g., ductal cells, centroacinar cells, differentiated endocrine cells, and endocrine precursors) can give rise to PDAC. There are a number of areas where refinements of the mouse models would be advantageous for gaining further insights into PDAC pathogenesis and for having improved systems for preclinical studies. In terms of understanding the molecular pathogenesis of PDAC, refinements of the existing mouse models could enable an improved understanding of the mechanisms regulating multistage progression of PanIN to PDAC. The existing models either undergo gradual PanIN progression through acquiring spontaneous molecular alterations such as Ink4a/Arf deletion, or have simultaneous activation of Kras and tumor suppressor inactivation. A system whereby Kras activation and TSG inactivation could be temporally controlled would provide a valuable context to study the molecular and cellular alterations by which TSG loss facilitates progression. The combined use of the Cre-Lox and FlpFrt recombination systems could be suitable for this application. Inducible transgenic systems have given insights into the role-specific oncogenes in tumor maintenance in other cancers and could be useful in studies of PDAC. The capacity to turn off the expression of these oncogenes at any stage during tumor progression enables an assessment of the biological roles of the oncogene in evolving tumors and to validate whether tumorigenesis requires sustained expression of these genes [80]. In PDAC, it would be of particular importance to assess the requirement of activated Kras for tumor maintenance in vivo. Another area of refinement could be in the development of models in which tumorigenesis occurs focally in the backdrop of genetically wild type pancreas. In the current models, Kras activation and TSG loss is induced either in all cells of the pancreas, or in subsets of acinar cells. In either case, tumorigenesis appears to be multifocal and there is also a very large population of cells at varying decrease of tumor progression as well as in the histologically normal pancreas that harbor the mutant alleles. A model in which mutations are confined to a much smaller fraction of pancreatic cells could provide an improved context for therapeutic studies. In lung cancer, the use of adenoviruses that express Cre recombinase have proven to be effective in inducing tumorigenesis more focally and to be useful for testing therapeutic compounds [81–83]. Such an approach should also be feasible in the studies of PDAC. An additional area of interest for both biological studies and preclinical studies is in developing genomically unstable models of PDAC. As noted above, human PDAC is characterized by extensive genomic instability that may be facilitated by progressive telomere shortening early in tumor progression. Mouse PDAC shows some chromosomal aberrations; however, the degree of such alterations appears to be somewhat less than that observed in human PDAC [59, 60]. Telomeres in the

390

A.F. Hezel and N. Bardeesy

normal mouse are much longer than in humans and do not exhibit significant shortening. Crossing of mice with mutations in components of the telomerase enzyme with the PDAC models could provide systems to study genomic instability in the mouse and thereby give insights into how such instability contributes to tumor progression and response to therapeutic agents [84]. These systems may also provide relevant models for cancer gene discovery.

16.8 Conclusions The utility of GEMMs in therapeutic applications and preclinical testing of promising drugs and biologic agents is an emerging field. The shortcomings of studies relying on cell lines is evident by the poor clinical track records of many drugs which initially appeared efficacious. It is clear there are significant advantages to these models over xenografted human cell lines namely, that the tumors occur in the correct cellular context and microenvironment and these models offer control over genetic context [85, 86]. On the other hand, GEMMs also have limitations, which must be recognized. Given the creation of the many models with differing underlying genetics and associated varying degrees of histological cellular atypia and resemblance to human PDAC, a consensus on the pathologic interpretation of these models is emerging [87]. This is critical to the integration of these models into preclinical drug evaluation. Further refinement in these models allowing pinpointed evaluations of drugs in focal lesions at differing stages of the disease from early precursor lesions through the growth of metastatic deposits and across a range of genetic settings will be invaluable. Neither human cell line-based models nor GEMMS will likely be adequate in cancer drug evaluation by themselves, but will be needed to be used in concert to guide clinical development. Ultimately, the success of engineered mouse models in pancreatic cancer therapeutics will be determined by the coming results of clinical trials, now in the early stages, whose design and implementation were guided by such models. Acknowledgments  Aram Hezel is supported by an NCI K08 CDA and HHMI early career award. Nabeel Bardeesy is supported by grants from the NIH and by the AACR-PanCAN award.

References 1. Murtaugh LC, Melton DA. Genes, signals, and lineages in pancreas development. Annu Rev Cell Dev Biol. 2003;19:71–89. 2. Jensen JN, Cameron E, Garay MV, Starkey TW, Gianani R, Jensen J. Recapitulation of ­elements of embryonic development in adult mouse pancreatic regeneration. Gastroenterology. 2005;128:728–41. 3. Brugge WR, Lauwers GY, Sahani D, Fernandez-del Castillo C, Warshaw AL. Cystic ­neoplasms of the pancreas. N Engl J Med. 2004;351:1218–26.

16  Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma

391

4. Maitra A, Fukushima N, Takaori K, Hruban RH. Precursors to invasive pancreatic cancer. Adv Anat Pathol. 2005;12:81–91. 5. Hruban RH, Adsay NV, Albores-Saavedra J, Compton C, Garrett ES, Goodman SN, et  al. Pancreatic intraepithelial neoplasia: a new nomenclature and classification system for ­pancreatic duct lesions. Am J Surg Pathol. 2001;25:579–86. 6. Kozuka S, Sassa R, Taki T, Masamoto K, Nagasawa S, Saga S, et al. Relation of pancreatic duct hyperplasia to carcinoma. Cancer. 1979;43:1418–28. 7. Heinmoller E, Dietmaier W, Zirngibl H, Heinmoller P, Scaringe W, Jauch KW, et al. Molecular analysis of microdissected tumors and preneoplastic intraductal lesions in pancreatic ­carcinoma. Am J Pathol. 2000;157:83–92. 8. Hruban RH, Iacobuzio-Donahue C, Wilentz RE, Goggins M, Kern SE. Molecular pathology of pancreatic cancer. Cancer J. 2001;7:251–8. 9. Hruban RH, Wilentz RE, Kern SE. Genetic progression in the pancreatic ducts. Am J Pathol. 2000;156:1821–5. 10. Klimstra DS, Longnecker DS. K-ras mutations in pancreatic ductal proliferative lesions. Am J Pathol. 1994;145:1547–50. 11. Luttges J, Galehdari H, Brocker V, Schwarte-Waldhoff I, Henne-Bruns D, Kloppel G, et al. Allelic loss is often the first hit in the biallelic inactivation of the p53 and DPC4 genes during pancreatic carcinogenesis. Am J Pathol. 2001;158:1677–83. 12. Maitra A, Adsay NV, Argani P, Iacobuzio-Donahue C, De Marzo A, Cameron JL, et  al. Multicomponent analysis of the pancreatic adenocarcinoma progression model using a ­pancreatic intraepithelial neoplasia tissue microarray. Mod Pathol. 2003;16:902–12. 13. Moskaluk CA, Hruban RH, Kern SE. p16 and K-ras gene mutations in the intraductal ­precursors of human pancreatic adenocarcinoma. Cancer Res. 1997;57:2140–3. 14. Wilentz RE, Iacobuzio-Donahue CA, Argani P, McCarthy DM, Parsons JL, Yeo CJ, et al. Loss of expression of Dpc4 in pancreatic intraepithelial neoplasia: evidence that DPC4 inactivation occurs late in neoplastic progression. Cancer Res. 2000;60:2002–6. 15. Yamano M, Fujii H, Takagaki T, Kadowaki N, Watanabe H, Shirai T. Genetic progression and divergence in pancreatic carcinoma. Am J Pathol. 2000;156:2123–33. 16. Almoguera C, Shibata D, Forrester K, Martin J, Arnheim N, Perucho M. Most human ­carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell. 1988;53:549–54. 17. Rozenblum E, Schutte M, Goggins M, Hahn SA, Panzer S, Zahurak M, et al. ­Tumor-suppressive pathways in pancreatic carcinoma. Cancer Res. 1997;57:1731–4. 18. Hustinx SR, Leoni LM, Yeo CJ, Brown PN, Goggins M, Kern SE, et al. Concordant loss of MTAP and p16/CDKN2A expression in pancreatic intraepithelial neoplasia: evidence of homozygous deletion in a noninvasive precursor lesion. Mod Pathol. 2005;18:959–63. 19. Hahn SA, Schutte M, Hoque AT, Moskaluk CA, da Costa LT, Rozenblum E, et al. DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. [see comments]. Science. 1996;271:350–3. 20. Murphy KM, Brune KA, Griffin C, Sollenberger JE, Petersen GM, Bansal R, et al. Evaluation of candidate genes MAP2K4, MADH4, ACVR1B, and BRCA2 in familial pancreatic cancer: deleterious BRCA2 mutations in 17%. Cancer Res. 2002;62:3789–93. 21. Jones S, Hruban RH, Kamiyama M, Borges M, Zhang X, Parsons DW, et  al. Exomic sequencing identifies PALB2 as a pancreatic cancer susceptibility gene. Science. 2009;324:217. 22. Su GH, Hruban RH, Bansal RK, Bova GS, Tang DJ, Shekher MC, et al. Germline and somatic mutations of the STK11/LKB1 Peutz-Jeghers gene in pancreatic and biliary cancers. Am J Pathol. 1999;154:1835–40. 23. Sato N, Rosty C, Jansen M, Fukushima N, Ueki T, Yeo CJ, et al. STK11/LKB1 Peutz-Jeghers gene inactivation in intraductal papillary-mucinous neoplasms of the pancreas. Am J Pathol. 2001;159:2017–22. 24. Jones S, Zhang X, Parsons DW, Lin JC, Leary RJ, Angenendt P, et al. Core signaling ­pathways in human pancreatic cancers revealed by global genomic analyses. Science. 2008;321:1801–6.

392

A.F. Hezel and N. Bardeesy

25. Armengol G, Knuutila S, Lluis F, Capella G, Miro R, Caballin MR. DNA copy number changes and evaluation of MYC, IGF1R, and FES amplification in xenografts of pancreatic adenocarcinoma. Cancer Genet Cytogenet. 2000;116:133–41. 26. Gorunova L, Hoglund M, Andren-Sandberg A, Dawiskiba S, Jin Y, Mitelman F, et  al. Cytogenetic analysis of pancreatic carcinomas: intratumor heterogeneity and nonrandom pattern of chromosome aberrations. Genes Chromosomes Cancer. 1998;23:81–99. 27. Gysin S, Rickert P, Kastury K, McMahon M. Analysis of genomic DNA alterations and mRNA expression patterns in a panel of human pancreatic cancer cell lines. Genes Chromosomes Cancer. 2005;44:37–51. 28. Harada T, Baril P, Gangeswaran R, Kelly G, Chelala C, Bhakta V, et  al. Identification of genetic alterations in pancreatic cancer by the combined use of tissue microdissection and array-based comparative genomic hybridisation. Br J Cancer. 2007;96:373–82. 29. Harada T, Chelala C, Bhakta V, Chaplin T, Caulee K, Baril P, et al. Genome-wide DNA copy number analysis in pancreatic cancer using high-density single nucleotide polymorphism arrays. Oncogene. 2008;27:1951–60. 30. Mahlamaki EH, Hoglund M, Gorunova L, Karhu R, Dawiskiba S, Andren-Sandberg A, et al. Comparative genomic hybridization reveals frequent gains of 20q, 8q, 11q, 12p, and 17q, and losses of 18q, 9p, and 15q in pancreatic cancer. Genes Chromosomes Cancer. 1997;20:383–91. 31. Nowak NJ, Gaile D, Conroy JM, McQuaid D, Cowell J, Carter R, et al. Genome-wide aberrations in pancreatic adenocarcinoma. Cancer Genet Cytogenet. 2005;161:36–50. 32. Sato N, Mizumoto K, Nakamura M, Nakamura K, Kusumoto M, Niiyama H, et al. Centrosome abnormalities in pancreatic ductal carcinoma. Clin Cancer Res. 1999;5:963–70. 33. Sato N, Mizumoto K, Nakamura M, Maehara N, Minamishima YA, Nishio S, et al. Correlation between centrosome abnormalities and chromosomal instability in human pancreatic cancer cells. Cancer Genet Cytogenet. 2001;126:13–9. 34. Schleger C, Arens N, Zentgraf H, Bleyl U, Verbeke C. Identification of frequent chromosomal aberrations in ductal adenocarcinoma of the pancreas by comparative genomic hybridization (CGH). J Pathol. 2000;191:27–32. 35. Sirivatanauksorn V, Sirivatanauksorn Y, Gorman PA, Davidson JM, Sheer D, Moore PS, et al. Non-random chromosomal rearrangements in pancreatic cancer cell lines identified by spectral karyotyping. Int J Cancer. 2001;91:350–8. 36. van Heek NT, Meeker AK, Kern SE, Yeo CJ, Lillemoe KD, Cameron JL, et  al. Telomere shortening is nearly universal in pancreatic intraepithelial neoplasia. Am J Pathol. 2002;161:1541–7. 37. Glasner S, Memoli V, Longnecker DS. Characterization of the ELSV transgenic mouse model of pancreatic carcinoma. Histologic type of large and small tumors. Am J Pathol. 1992;140:1237–45. 38. Ornitz DM, Hammer RE, Messing A, Palmiter RD, Brinster RL. Pancreatic neoplasia induced by SV40 T-antigen expression in acinar cells of transgenic mice. Science. 1987;238:188–93. 39. Quaife CJ, Pinkert CA, Ornitz DM, Palmiter RD, Brinster RL. Pancreatic neoplasia induced by ras expression in acinar cells of transgenic mice. Cell. 1987;48:1023–34. 40. Sandgren EP, Luetteke NC, Qiu TH, Palmiter RD, Brinster RL, Lee DC. Transforming growth factor alpha dramatically enhances oncogene-induced carcinogenesis in transgenic mouse pancreas and liver. Mol Cell Biol. 1993;13:320–30. 41. Sandgren EP, Quaife CJ, Paulovich AG, Palmiter RD, Brinster RL. Pancreatic tumor pathogenesis reflects the causative genetic lesion. Proc Natl Acad Sci U S A. 1991;88:93–7. 42. Wagner M, Greten FR, Weber CK, Koschnick S, Mattfeldt T, Deppert W, et  al. A murine tumor progression model for pancreatic cancer recapitulating the genetic alterations of the human disease. Genes Dev. 2001;15:286–93. 43. Bardeesy N, Morgan J, Sinha M, Signoretti S, Srivastava S, Loda M, et al. Obligate roles for p16(Ink4a) and p19(Arf)-p53 in the suppression of murine pancreatic neoplasia. Mol Cell Biol. 2002;22:635–43.

16  Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma

393

44. 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;61:1121–35. 45. Tuveson DA, Zhu L, Gopinathan A, Willis NA, Kachatrian L, Grochow R, et  al. Mist1KrasG12D knock-in mice develop mixed differentiation metastatic exocrine pancreatic carcinoma and hepatocellular carcinoma. Cancer Res. 2006;66:242–7. 46. Lewis BC, Klimstra DS, Varmus HE. The c-myc and PyMT oncogenes induce different tumor types in a somatic mouse model for pancreatic cancer. Genes Dev. 2003;17:3127–38. 47. Morton JP, Klimstra DS, Mongeau ME, Lewis BC. Trp53 deletion stimulates the formation of metastatic pancreatic tumors. Am J Pathol. 2008;172:1081–7. 48. Hezel AF, Kimmelman AC, Stanger BZ, Bardeesy N, Depinho RA. Genetics and biology of pancreatic ductal adenocarcinoma. Genes Dev. 2006;20:1218–49. 49. 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;63:2016–9. 50. Brembeck FH, Schreiber FS, Deramaudt TB, Craig L, Rhoades B, Swain G, et al. The mutant K-ras oncogene causes pancreatic periductal lymphocytic infiltration and gastric mucous neck cell hyperplasia in transgenic mice. Cancer Res. 2003;63:2005–9. 51. Orsulic S, Li Y, Soslow RA, Vitale-Cross LA, Gutkind JS, Varmus HE. Induction of ovarian cancer by defined multiple genetic changes in a mouse model system. Cancer Cell. 2002;1:53–62. 52. Pao W, Klimstra DS, Fisher GH, Varmus HE. Use of avian retroviral vectors to introduce transcriptional regulators into mammalian cells for analyses of tumor maintenance. Proc Natl Acad Sci U S A. 2003;100:8764–9. 53. Gu G, Brown JR, Melton DA. Direct lineage tracing reveals the ontogeny of pancreatic cell fates during mouse embryogenesis. Mech Dev. 2003;120:35–43. 54. Kawaguchi Y, Cooper B, Gannon M, Ray M, MacDonald RJ, Wright CV. The role of the transcriptional regulator Ptf1a in converting intestinal to pancreatic progenitors. Nat Genet. 2002;32:128–34. 55. Guerra C, Mijimolle N, Dhawahir A, Dubus P, Barradas M, Serrano M, et al. Tumor induction by an endogenous K-ras oncogene is highly dependent on cellular context. Cancer Cell. 2003;4:111–20. 56. Johnson L, Mercer K, Greenbaum D, Bronson RT, Crowley D, Tuveson DA, et al. Somatic activation of the K-ras oncogene causes early onset lung cancer in mice. Nature. 2001;410:1111–6. 57. Aguirre AJ, Bardeesy N, Sinha M, Lopez L, Tuveson DA, Horner J, et al. Activated Kras and Ink4a/Arf deficiency cooperate to produce metastatic pancreatic ductal adenocarcinoma. Genes Dev. 2003;17:3112–26. 58. Hingorani SR, Petricoin EF, Maitra A, Rajapakse V, King C, Jacobetz MA, et al. Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse. Cancer Cell. 2003;4:437–50. 59. Bardeesy N, Aguirre AJ, Chu GC, Cheng KH, Lopez LV, Hezel AF, et al. From the Cover: Both p16Ink4a and the p19Arf-p53 pathway constrain progression of pancreatic adenocarcinoma in the mouse. Proc Natl Acad Sci U S A. 2006;103:5947–52. 60. Hingorani SR, Wang L, Multani AS, Combs C, Deramaudt TB, Hruban RH, et al. Trp53R172H and KrasG12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice. Cancer Cell. 2005;7:469–83. 61. Bardeesy N, Cheng KH, Berger JH, Chu GC, Pahler J, Olson P, et al. Smad4 is dispensable for normal pancreas development yet critical in progression and tumor biology of pancreas cancer. Genes Dev. 2006;22:3130–46. 62. Izeradjene K, Combs C, Best M, Gopinathan A, Wagner A, Grady WM, et al. Kras(G12D) and Smad4/Dpc4 haploinsufficiency cooperate to induce mucinous cystic neoplasms and invasive adenocarcinoma of the pancreas. Cancer Cell. 2007;11:229–43. 63. Kojima K, Vickers SM, Adsay NV, Jhala NC, Kim HG, Schoeb TR, et al. Inactivation of Smad4 accelerates Kras(G12D)-mediated pancreatic neoplasia. Cancer Res. 2007;67:8121–30.

394

A.F. Hezel and N. Bardeesy

64. Ijichi H, Chytil A, Gorska AE, Aakre ME, Fujitani Y, Fujitani S, et al. Aggressive pancreatic ductal adenocarcinoma in mice caused by pancreas-specific blockade of transforming growth factor-beta signaling in cooperation with active Kras expression. Genes Dev. 2006;20:3147–60. 65. Guerra C, Schuhmacher AJ, Canamero M, Grippo PJ, Verdaguer L, Perez-Gallego L, et al. Chronic pancreatitis is essential for induction of pancreatic ductal adenocarcinoma by K-Ras oncogenes in adult mice. Cancer Cell. 2007;11:291–302. 66. Adorno M, Cordenonsi M, Montagner M, Dupont S, Wong C, Hann B, et al. A Mutant-p53/ Smad complex opposes p63 to empower TGFbeta-induced metastasis. Cell. 2009;137:87–98. 67. Massague J, Blain SW, Lo RS. TGFbeta signaling in growth control, cancer, and heritable disorders. Cell. 2000;103:295–309. 68. Feldmann G, Habbe N, Dhara S, Bisht S, Alvarez H, Fendrich V, et al. Hedgehog inhibition prolongs survival in a genetically engineered mouse model of pancreatic cancer. Gut. 2008;57:1420–30. 69. Mukherjee P, Basu GD, Tinder TL, Subramani DB, Bradley JM, Arefayene M, et  al. Progression of pancreatic adenocarcinoma is significantly impeded with a combination of vaccine and COX-2 inhibition. J Immunol. 2009;182:216–24. 70. Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D, Honess D, et al. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science. 2009;324:1457–61. 71. Plentz R, Park JS, Rhim AD, Abravanel D, Hezel AF, Sharma SV, et al. Inhibition of gammasecretase activity inhibits tumor progression in a mouse model of pancreatic ductal adenocarcinoma. Gastroenterology. 2009;136:1741–9, e6. 72. Hingorani SR, Tuveson DA. In search of an early warning system for pancreatic cancer. Cancer Biol Ther. 2003;2:84–6. 73. Faca VM, Song KS, Wang H, Zhang Q, Krasnoselsky AL, Newcomb LF, et al. A mouse to human search for plasma proteome changes associated with pancreatic tumor development. PLoS Med. 2008;5:e123. 74. Kelly KA, Bardeesy N, Anbazhagan R, Gurumurthy S, Berger J, Alencar H, et al. Targeted nanoparticles for imaging incipient pancreatic ductal adenocarcinoma. PLoS Med. 2008;5:e85. 75. Carriere C, Seeley ES, Goetze T, Longnecker DS, Korc M. The Nestin progenitor lineage is the compartment of origin for pancreatic intraepithelial neoplasia. Proc Natl Acad Sci U S A. 2007;104:4437–42. 76. Habbe N, Shi G, Meguid RA, Fendrich V, Esni F, Chen H, et al. Spontaneous induction of murine pancreatic intraepithelial neoplasia (mPanIN) by acinar cell targeting of oncogenic Kras in adult mice. Proc Natl Acad Sci U S A. 2008;105:18913–8. 77. Shi G, Zhu L, Sun Y, Bettencourt R, Damsz B, Hruban RH, et al. Loss of the acinar-restricted transcription factor Mist1 accelerates Kras-induced pancreatic intraepithelial neoplasia. Gastroenterology. 2009;136:1368–78. 78. Baeyens L, Bouwens L. Cellular plasticity of the pancreas. Biol Chem. 2009;390:975–1001. 79. Xu X, D’Hoker J, Stange G, Bonne S, De Leu N, Xiao X, et al. Beta cells can be generated from endogenous progenitors in injured adult mouse pancreas. Cell. 2008;132:197–207. 80. Chin L, Tam A, Pomerantz J, Wong M, Holash J, Bardeesy N, et al. Essential role for oncogenic Ras in tumour maintenance. Nature. 1999;400:468–72. 81. Engelman JA, Chen L, Tan X, Crosby K, Guimaraes AR, Upadhyay R, et al. Effective use of PI3K and MEK inhibitors to treat mutant Kras G12D and PIK3CA H1047R murine lung cancers. Nat Med. 2008;14:1351–6. 82. Jackson EL, Willis N, Mercer K, Bronson RT, Crowley D, Montoya R, et al. Analysis of lung tumor initiation and progression using conditional expression of oncogenic K-ras. Genes Dev. 2001;15:3243–8. 83. Ji H, Ramsey MR, Hayes DN, Fan C, McNamara K, Kozlowski P, et al. LKB1 modulates lung cancer differentiation and metastasis. Nature. 2007;448:807–10.

16  Genetically Engineered Mouse Models of Pancreatic Ductal Adenocarcinoma

395

84. Artandi SE, Chang S, Lee SL, Alson S, Gottlieb GJ, Chin L, et  al. Telomere dysfunction promotes non-reciprocal translocations and epithelial cancers in mice. Nature. 2000;406:641–5. 85. Olive KP, Tuveson DA. The use of targeted mouse models for preclinical testing of novel cancer therapeutics. Clin Cancer Res. 2006;12:5277–87. 86. Sharpless NE, Depinho RA. The mighty mouse: genetically engineered mouse models in cancer drug development. Nat Rev Drug Discov. 2006;5:741–54. 87. Hruban RH, Adsay NV, Albores-Saavedra J, Anver MR, Biankin AV, Boivin GP, et  al. Pathology of genetically engineered mouse models of pancreatic exocrine cancer: consensus report and recommendations. Cancer Res. 2006;66:95–106.

Chapter 17

Transgenic Adenocarcinoma of the Mouse Prostate: A Validated Model for the Identification and Characterization of Molecular Targets and The Evaluation of Therapeutic Agents Sharon D. Morgenbesser Abstract  Prostate cancer remains a serious health issue worldwide. Genetically engineered mouse models have been utilized extensively to better understand human prostate cancer at the genetic and cellular levels, to identify new diagnostic markers, therapeutic targets, and biomarkers, and to evaluate the efficacy of potential therapeutic agents. Of these, the transgenic adenocarcinoma of the mouse prostate (TRAMP) model is extremely well characterized, highly regarded as a model for human prostate cancer, and the most frequently used. This chapter will review the histological and molecular characterization of pathologic progression in TRAMP, and focus on its use to identify genes that regulate multiple transition points in tumor development, to understand the contributions of the androgen-pathway and DNA methylation as regulators of gene expression, to explore the role of numerous proteins in prostate tumorigenesis, and to evaluate a wide variety of anti-tumor agents. The influence of murine genetic background on disease progression, and on gene expression and function, and the utilization of these differences to model the genetic heterogeneity that exists in the human disease are also discussed. Keywords  Prostate cancer • Androgen • Androgen receptor • Castration-resistant • Gene expression profiling • DNA methylation • Therapeutic • Preclinical trials

17.1 Introduction Prostate cancer (PC) continues to be a major problem affecting men world wide. In 2008 in the US, an estimated 186,320 men were diagnosed with the disease, and approximately 28,660 died as a result [1]. Detection and treatment are generally successful when tumors are confined to the prostate and their growth and survival S.D. Morgenbesser (*) 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_17, © Springer Science+Business Media, LLC 2011

397

398

S.D. Morgenbesser

are androgen-dependent (AD), as radical prostatectomy is generally very effective at early stages. If there is evidence of new tumor growth, or PC is initially detected at an advanced stage, androgen-deprivation therapy (ADT), which involves surgical or chemical castration, is employed and prostate tumors initially regress [2, 3]. However, within 2–3 years, castration-resistant (CR) tumors emerge in most patients, and CR and metastatic prostate tumors are associated with high mortality rates [2]. Therefore, there remains an urgent need for better diagnostic markers, as monitoring prostate-specific antigen (PSA) levels has limited benefit [4, 5], as well as additional therapeutic targets and more effective therapies. For many years, clinical material in the form of established cell lines, tissues propagated in immunodeficient mice (xenografts), and frozen or fixed tumor specimens had been the sources for interrogation. While they have great utility, they are limited by a high degree of genetic heterogeneity, artifacts commonly associated with cell lines that are serially passaged by growing in culture or implanted outside of the prostate microenvironment, and the difficulty of obtaining primary tumor tissues from patients before treatment, as well as in procuring early prostatic intraepithelial neoplasia (PIN) lesions, CR and metastatic tumors, and suitable normal control tissues. To overcome these limitations, nearly 30 genetically engineered mouse models (GEMMs) of PC have been generated by either enforced expression of oncogenes or inactivation of tumor suppressor genes in the mouse prostate [6]. Although these models display many features commonly found in clinical disease, there are some inherent limitations to their utility as the mouse and human prostates possess some differences. For example, the human organ comprises contiguous ­tissue that can be phenotypically resolved to the central, transition, and peripheral zones while the mouse prostate develops as pairs of ventral, dorsal lateral, and anterior lobes. In addition, mice do not carry Kal3, the gene that encodes PSA. Nevertheless, many of the models develop spontaneous and progressive ­autochthonous PC and some display metastatic and CR disease. While each model possesses strengths and weaknesses, the mice known as ­transgenic adenocarcinoma of the mouse prostate (TRAMP) are extremely wellcharacterized histologically and molecularly, and as such these mice are highly regarded as a model for human disease and the most widely used. In particular, TRAMP is one of the few GEMMS to give rise to lymphatic and hematologic ­metastatic disease with reproducible evidence of bone lesions [7, 8]. Moreover, disease progression in TRAMP occurs within a reasonable time frame for experimental research. Three epithelial cell lines derived from TRAMP have also been established [9] and validated as useful models in in vitro assays [10] and in vivo as grafts [11, 12]. This chapter will review the pathologic characterization of the model, and explore its use for the identification and validation of genes and ­pathways involved in prostate tumorigenesis, which may provide useful diagnostic markers and therapeutic targets, and for use as a platform for testing known and potential therapeutic agents and identifying clinically relevant biomarkers. Though studies that employed the TRAMP cell lines have been informative, only those utilizing the authochthonous mouse model are discussed.

17  Transgenic Adenocarcinoma of the Mouse Prostate

399

17.2 Generation of the TRAMP Model and Phenotypic Characterization The TRAMP model was generated by utilizing the minimal regulatory elements of the androgen-regulated rat probasin promoter [13] to enforce prostate-specific expression of the potent SV40 T and t oncoproteins in C57BL/6 inbred mice [14]. The large T can inactivate the tumor suppressor proteins pRb and p53 [15], which are deficient in many human PCs [16, 17]. In addition, the large T can also inhibit Bub1 [18], which is a spindle assembly checkpoint protein, and the little t negatively regulates protein phosphatase 2A, a tumor suppressor [19]. Expression of the transgene initiates between 4 and 6 weeks of age and tumors arise spontaneously in the prostatic epithelial compartment with a short latency. TRAMP C57BL/6 mice have been mated to FVB mice, and comparisons demonstrate that the TRAMP phenotype in C57BL/6 mice is more mild than in [C57BL/6 × FVB]F1 mice [8], suggesting the influence of genetic modifiers and indicating that these mice can model the diversity that exists within the human disease. In the [C57BL/6 × FVB] F1 mice, PIN lesions are evident by 8–12 weeks of age, well differentiated (WD) carcinomas appear by 18 weeks, and poorly differentiated (PD) tumors emerge by 18–24 weeks [20]. Metastatic disease can be observed as early as 12 weeks with lesions primarily in the lymph nodes (LN), but also in lung, kidney, and adrenal gland; bone metastases have only been observed in TRAMP mice on the mixed background [7, 8]. TRAMP [C57BL/6 × FVB]F1 and C57BL/6 mice succumb to the disease by 33 and 36–40 weeks, respectively [8, 20, 21]. TRAMP tumor development is initially AD as tumors in males castrated at early stages (e.g., when PIN lesions have appeared) initially undergo massive apoptosis and regress, but CR primary and metastatic tumors eventually emerge in most mice several weeks later [21, 22]. A significant effort has been devoted towards defining the molecular mechanisms leading to CR growth in human PC, particularly in the context of ADT, including analysis of the expression, mutational status, and activity of the androgen receptor (AR), as well as its various co-factors [2]. Upon binding androgens, the AR assumes an active conformation capable of binding to specific response elements in the promoter regions of androgen-responsive genes, which either enhances or represses their transcription. Mutations have been observed in the AR of clinical samples that are associated with enhanced activities, such as the ability to bind other ligands, which may confer tumor growth after ADT. In TRAMP, both transformed epithelial cells and nascent stromal cells express the AR [8], and naturally occurring, somatic AR mutations have been identified in prostate tumors from intact and castrated TRAMP [C57BL/6 × FVB]F1 mice [23]. Notably, all the mutations from the tumors of intact mice were in the ligand-binding domain (LBD), whereas the vast majority of those from castrated mice were derived from the transactivation domain, indicating that ADT can affect the particular type of alterations. Many of these mutations were similar to those observed in man, and functional analyses revealed their enhanced activities. In fact, enforced expression of one variant, AR-E231G, was subsequently shown to be capable of causing PC in

400

S.D. Morgenbesser

a GEMM [24]. This phenotype was likely not simply related to over expressing AR since enforced expression of wild-type AR or another mutation in the LBD commonly observed in human cancer cell lines were unable to cause tumors in GEMMs. Clearly, the emergence of CR prostate tumor growth and the acquisition of these AR mutations demonstrate that TRAMP is highly suitable for further investigation into this aspect of human PC. Human prostate tumor progression is enhanced by a reactive stroma that ­provides various extracellular matrix proteins and growth factors that activate key signaling pathways within the epithelial cells [25]. The stromal compartment is nominally AR-positive, and primarily includes fibroblasts, as well as myofibroblasts and smooth muscle cells. In the tumor microenvironment, fibroblasts often increase in number and are genetically unstable. In TRAMP, tumor-associated stromal cells similarly express AR [8], and about 30% are aneuploid and have ­multiple centrosomes [26]. A stromal response in TRAMP is further indicated by a thickening and remodeling of the stromal compartment in PIN lesions, followed by additional thickening and remodeling with hypercellularity in AD and CR tumors [8, 21, 27]. Human prostate tumorigenesis relies on angiogenesis for a sufficient blood ­supply to enable growth of tumors beyond 1 mm3, with signals provided by both epithelial and stromal cells for the formation of new blood vessels from endothelial cells (ECs) [28]. Tumors in TRAMP [C57BL/6 × FVB]F1 mice have a much higher microvessel density (MVD) relative to those in TRAMP C57BL/6 mice, an observation that correlates with their differences in disease progression [8]. The process of angiogenesis has been particularly well-characterized in the TRAMP [C57BL/6 × FVB]F1 mice [29, 30] and demonstrates high similarity to that observed in clinical PC. New vessels first appear in the stromal space in low grade PIN, are more numerous and reside consistently with the infolded mesenchyme adjacent to the epithelium in high grade PIN, and become increasingly more scattered, disorganized, and abundant in both regions as tumors progress through the WD, PD, and CR tumor stages [29]. While pericytes are tightly associated with vessels at the earlier stages, these cells becomes more numerous in WD tumors, and are loosely associated and abnormally shaped in PD lesions [30]. At the molecular level, expression of the pro-angiogenic factors VEGF (vascular endothelial growth factor) and FGF2 (fibroblast growth factor 2) and their cognate receptors arise during the vascularization process. FGF2 and VEGF receptors (VEGFR) are primarily expressed on ECs, suggesting involvement of both axes during transformation. VEGFR1 protein expression coincides with the initiation of the early angiogenic increase in high-grade PIN, whereas VEGF and VEGFR2 protein expression are detected once PD tumors emerge [29]. FGF2 is expressed in PIN lesions and more advanced tumors, FGFRiiib is expressed in the neovasculature in these tumors, and FGFR2iic is elevated in the epithelial cells within PIN [31]. In the prostate, neuroendocrine (NE) cells arise from a common epithelial stem cell, but differentiate into AR-negative cells to influence the growth and differentiation the prostate [32, 33]. The density of NE foci is much higher in PC than in normal prostate tissue, particularly in patients who have undergone ADT, and this NE density correlates with tumor grade, rapid tumor growth, increased angiogenesis,

17  Transgenic Adenocarcinoma of the Mouse Prostate

401

loss of androgen sensitivity, and more importantly, a poor prognosis. Within the tumor environment, and particularly in the context of ADT, the expansion of the NE compartment may be driven by differentiation of prostate tumor stem cells and/or by a transdifferentiation of neoplastic epithelial cells. Similarly, NE cells were observed in TRAMP prostate tumors when this model was initially being characterized [20]. Recent studies have examined the precise timing and mechanism of the emergence of these cells. Huss et al. show that in TRAMP [C57BL/6 × FVB]F1 mice with PD tumors, the NE cells derive from malignant progenitor cells as a rare, stochastic, late event [34]. In contrast, Chiaverotti et  al. studied the TRAMP ­phenotype on the C57BL/6 and FVB backgrounds and concluded that NE carcinomas arose from progenitor cells at an early stage of tumorigenesis to constitute the primary type of lethal malignancy that emerges in these mice [35]. These somewhat contradictory findings likely reflect the different mouse strains used and the subjectivity associated with tumor classification as well as other differences. However, both studies underscore that, like the analysis of NE features in human PC, ­additional investigation is warranted and that multiple molecular mechanisms may drive this phenotype.

17.3 Gene Expression Profiling Studies The expression of approximately 24 transcripts and proteins of particular interest in human PC had been studied and contributed to the validation of the model at the molecular level, as the expression patterns were generally similar to those observed in the human disease. The expressed genes studied include E-cadherin [20], insulin growth factor (IGF) and FGF ligands, their receptors, and binding partners [31, 36, 37], cyclins and cyclin-dependent kinases [38], prostate stem cell antigen (PSCA) [39–41], the metastasis-associated protein S100A4 [42], the Aurora kinases [12], mouse telomerase reverse transcriptase (mTERT) [43], and glutathione S-transferases (GSTs) [44]. Gene expression profiling--> 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.

17  Transgenic Adenocarcinoma of the Mouse Prostate

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 anti-­ estrogen), 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).

17  Transgenic Adenocarcinoma of the Mouse Prostate

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

17  Transgenic Adenocarcinoma of the Mouse Prostate

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.

17  Transgenic Adenocarcinoma of the Mouse Prostate

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.

17  Transgenic Adenocarcinoma of the Mouse Prostate

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.

17  Transgenic Adenocarcinoma of the Mouse Prostate

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

S.D. Hursting et al.

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

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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

S.D. Hursting et al.

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

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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

S.D. Hursting et al.

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

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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

S.D. Hursting et al.

[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

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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

S.D. Hursting et al.

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

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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

S.D. Hursting et al.

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

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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

S.D. Hursting et al.

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.

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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

S.D. Hursting et al.

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.

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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

S.D. Hursting et al.

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.

18  The Utility of Transgenic Mouse Models for Cancer Prevention Research

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]

19  Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease 455

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]

19  Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease 457

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

19  Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease 459

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

19  Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease 461

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

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

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.

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

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

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

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.

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

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.

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

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)

19  Models for Evaluation of Targeted Therapies of Invasive and Metastatic Disease 473

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].

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

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.

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

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

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

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

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

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.

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

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.

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

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.

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

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.

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

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.

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

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.

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

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.

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

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