No title

Introduction

Volume 1 of Advances in Cancer Research appeared in 1953, with Jesse Greenstein and Alexander Haddow as editors. Greenstein was a prominent biochemist, head of the Biochemistry Department at the NCI and particularly known for his pivotal book “Biochemistry of Cancer.” Alexander Haddow was the Head of the Chester Beatty Research Institute in London and a pioneer in the field of carcinogenesis. The articles of the first volume reflect the spirit of the times. They were written and assembled prior to the publication of the double helix structure of DNA. There was increasing evidence suggesting that genetic information may be encoded in the nucleic acids, rather than in the proteins as previously believed, but this was not yet generally accepted. The main paradigmatic shift was still a couple of years off. Characteristically, DNA is still called “thymonucleic acid” in one of the reviews. Its significance is unsuspected. The same review, concerned with epidermal cancerogenesis, notes that ribonucleic acid is “increased in some cancers” but does not see the connection to either cell growth or protein synthesis. In 1953, the virus‐cancer field has recently emerged from its deep sleep of several decades. It is thoroughly discussed in two reviews, one on the mammary tumor virus, referred to as the milk agent and the other on “the agent of Rous no. 1 sarcoma.” They are not yet heralding the panvirological era that will come one or two decennia later. At that time, most experimental cancers will be attributed to tumor viruses. This will dwindle in another decennium, opening the way for cell biology, a discipline that is conspicuous by its absence in Volume 1. So is the multistep nature carcinogenesis and the Darwinian microevolution of tumor cell populations. The classical discipline of carcinogenesis gets no less than five reviews. They are all excellent reflections of the time and provide basic information that survives to our days. The role of hormonal imbalance in causing tumors in endocrine glands, and their targets and the transition from hormone dependence to independence, is strongly brought out by one review. Some special aspects of the relationships between the host and the tumor cell are considered in two reviews, one on nutrition and the other on plasma proteins.

v

vi

Introduction

Would the authors of the reviews in Volume 1 understand the articles in Volume 100? Could they move freely around in the new cancer biology, like Icelandic Vikings of the year 1000 who could communicate without the slightest difficulty in modern Iceland or ancient Israelites in modern Israel, or would they be more like the ancient Greeks who would understand little if anything in modern Greece? I fear the latter or even worse, because the old Greeks would at least recognize the alphabet. If scientists from pre–Watson and Crick times would look at all those sequences or if they would read about genes being knocked in and knocked out, constructs being switched on and off, they would probably come to the conclusion that they are dealing with some advanced form of science fiction. But if they would encounter a clinical article with some patient pictures and X‐rays, they would no doubt realize that the game is still more or less the same, although at a much improved level of therapeutic success in the clinic, and an incredibly more sophisticated science in the laboratory. George Klein

From Volume 1, published in 1953, to Volume 100, published in 2008, Advances in Cancer Research (ACR) has covered a remarkable period of discovery that encompasses the beginning of the revolution in biology. The first ACR volume came out in the year that Watson and Crick reported on the central dogma of biology, the DNA double helix. In the first 100 volumes are found many contributions by some of those who helped shape the revolution and who made many of the remarkable discoveries in cancer research that have developed from it. We have tried to capture the importance of the early contributions in our “Foundations in Cancer Research” part of the series. However, during a revolution, it is hard, from moment to moment, to identify the leading edge. In 1953, we lacked any understanding of the fundamental basis of cancer. The knowledge we did have was descriptive and phenomenological, but it was all we had as we began—slowly, but at an ever accelerating rate—to characterize biological systems in previously unattainable detail. What is largely unappreciated outside the science world is that, to understand the causes of cancer, a knowledge of the complex mechanisms of the processes within “normal” cells, tissues, and organisms is required, and not in descriptive terms, but in molecular terms. The ability to approach a detailed understanding of causality required us to ask new questions in new ways, leading to new fields of study and new technologies. The boundaries of traditional biology disciplines blurred, fields such as molecular biology and molecular genetics, genomics, and proteomics arose, and technologies such as blotting technologies, polymerase chain reaction, and recombinant DNA were developed.

Introduction

vii

The many disciplines involved in the study of cancer causes—developmental and cell biology, genetics, biochemistry, virology, microbiology, and immunology—merged their techniques and knowledge in pursuit of cellular/ molecular mechanisms and led to the first phase of success, discovery. Previously, these disciplines were “stovepipes” of knowledge, unable to cross‐fertilize, in large part due to terminology barriers and biases of perspective. Seemingly overnight, however, differences vanished and disciplines merged through the discovery of “oncogenes,” made possible by new genetic technologies. And over a longer span, we gradually learned that cancer and its genetic instability is the antithesis of normal biological function. The excitement and stimulus of tying together results from collaborations between two or more disciplines, and the ever‐widening vistas they uncovered, can hardly be understated. Such events were defining moments of the discovery phase. The second phase continues today, with application of the molecular and mechanistic knowledge we have acquired to the development of increasingly sophisticated technologies and analytic methods, and to the pursuit of effective ways of modulating the pathways and mechanisms we now recognize. This phase has been powered by information technology and database management using the principles of physics, and engineering, and has given rise to genomics, proteomics, and computational biology. The cancer‐related aspects of this phase include targeted drug therapy, clinical trials of novel classes of drugs, and personalized medicine. The Apollo 11 moon landing in 1969 was a tremendous accomplishment that had as its foundation several hundred years of physics, astronomy, and engineering knowledge. Yet the crux of that feat—solving a multibody problem in celestial mechanics and a set of multiple‐redundancy problems in engineering—was relatively simple when compared with the interdependent complexities of the causes of, and paths to cures for, cancer in living beings. The revolution in biology has led to the unraveling of some long‐standing mysteries: how life and living organisms work; of the pathogens that infect them; of the genetic variations that predispose them to resistance or disease; and of how the environment imposes its pressure for natural selection. We have established an initial foundation of molecular biological knowledge over the past fifty‐some years. We continue to expand that foundation, while we simultaneously apply the knowledge in novel ways. The third phase, which we hope has begun, will see this accumulation of knowledge and effort culminate in widespread and successful preventive measures and cures for cancer and other diseases. George Vande Woude

Contributors

Numbers in parentheses indicate the pages on which the authors’ contributions begin.

Nissim Benvenisty, Stem Cells Unit, Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem 91904, Israel (133) Barak Blum, Stem Cells Unit, Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem 91904, Israel (133) Yihai Cao, Laboratory of Angiogenesis Research, Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, 171 77 Stockholm, Sweden (113) Shannon L. Duffy, Leukaemia Foundation Research Unit, Queensland Institute of Medical Research, Herston, Qld 4029, Australia; UMR144 Curie/CNRS, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France (85) H. Shelton Earp, Department of Medicine and Pharmacology, UNC Lineberger Comprehensive Cancer Center, University of North Carolina School of Medicine, Chapel Hill, NC (35) Silvia Fre, UMR144 Curie/CNRS, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France (85) Douglas K. Graham, Department of Pediatrics, University of Colorado at Denver and Health Sciences Center, Aurora, CO (35) Klaus-Peter Janssen, Department of Surgery, Klinikum rechts der Isar, TUM, 81675 Munich, Germany (85) Amy K. Keating, Department of Pediatrics, University of Colorado at Denver and Health Sciences Center, Aurora, CO (35) Rachel M. A. Linger, Department of Pediatrics, University of Colorado at Denver and Health Sciences Center, Aurora, CO (35) Daniel Louvard, UMR144 Curie/CNRS, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France (85) Belinda E. Peace, Department of Surgery, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0558 (1) Sylvie Robine, UMR144 Curie/CNRS, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France (85)

xiii

xiv

Contributors

Harry Rubin, Department of Molecular and Cell Biology, Life Sciences Addition, University of California, Berkeley, CA 94720-3200 (159) Marie Schoumacher, UMR144 Curie/CNRS, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France (85) Danijela Vignjevic, UMR144 Curie/CNRS, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France (85) Purnima K. Wagh, Graduate Program in Cell and Cancer Biology, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0558; Department of Surgery, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0558 (1) Susan E. Waltz, Department of Research, Shriner’s Hospital for Children, Cincinnati, OH; Graduate Program in Cell and Cancer Biology, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0558; and Department of Surgery, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0558 (1) Isaac P. Witz, Department of Cell Research and Immunology, The George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv, 69978, Israel (203)

Met‐Related Receptor Tyrosine Kinase Ron in Tumor Growth and Metastasis Purnima K. Wagh,*,{,} Belinda E. Peace,*,} and Susan E. Waltz*,{,z *Department of Surgery, University of Cincinnati College of Medicine, Cincinnati, OH 45267‐0558;

{Graduate Program in Cell and Cancer Biology, University of Cincinnati College z

of Medicine, Cincinnati, OH 45267‐0558; Department of Research, Shriner’s Hospital for Children, Cincinnati, OH; } Contributed equally to this work

I. II. III. IV. V. VI. VII. VIII. IX. X. XI. XII. XIII. XIV. XV.

Ron Structure and Function Ron Ligand Structure and Function Ron Chromosomal Location and Cancer Ron in Macrophages: Inflammation and Cancer Developmental Roles of Ron and Tumor Properties Epithelial to Mesenchymal Transition Oncogenic Potential of the Ron Receptor Loss of Function Mouse Models for Ron Loss of Ron Function and Tumorigenesis Gain of Function Mouse Models for Ron Overexpression in Tumors Mechanisms of Ron‐Induced Tumorigenesis: Signaling Through the Ron Receptor Receptor Cross‐Talk and Ron Activity in Tumorigenesis Angiogenesis Genomic Instability and Cell Cycle Disruption Ron Expression in Human Tumors and Tumor‐Derived Cell Lines A. Breast Cancer B. Prostate Cancer C. Pancreatic Cancer D. Renal Tumors E. Bladder Cancer F. Ovarian Cancer G. Lung Cancer H. Gastrointestinal Tumors I. Liver Cancers J. Short Form Ron XVI. Ron as a Target of Cancer Therapy XVII. Conclusions References The Ron receptor is a member of the Met family of cell surface receptor tyrosine kinases and is primarily expressed on epithelial cells and macrophages. The biological response of Ron is mediated by binding of its ligand, hepatocyte growth factor‐like Advances in CANCER RESEARCH Copyright 2008, Elsevier Inc. All rights reserved.

1

0065-230X/08 $35.00 DOI: 10.1016/S0065-230X(08)00001-8

Purnima K. Wagh et al.

2

protein/macrophage stimulating‐protein (HGFL). HGFL is primarily synthesized and secreted from hepatocytes as an inactive precursor and is activated at the cell surface. Binding of HGFL to Ron activates Ron and leads to the induction of a variety of intracellular signaling cascades that leads to cellular growth, motility and invasion. Recent studies have documented Ron overexpression in a variety of human cancers including breast, colon, liver, pancreas, and bladder. Moreover, clinical studies have also shown that Ron overexpression is associated with both worse patient outcomes as well as metastasis. Forced overexpression of Ron in transgenic mice leads to tumorigenesis in both the lung and the mammary gland and is associated with metastatic dissemination. While Ron overexpression appears to be a hallmark of many human cancers, the mechanisms by which Ron induces tumorigenesis and metastasis are still unclear. Several strategies are currently being undertaken to inhibit Ron as a potential therapeutic target; current strategies include the use of Ron blocking proteins, small interfering RNA (siRNA), monoclonal antibodies, and small molecule inhibitors. In total, these data suggest that Ron is a critical factor in tumorigenesis and that inhibition of this protein, alone or in combination with current therapies, may prove beneficial in the treatment of cancer patients. # 2008 Elsevier Inc.

I. RON STRUCTURE AND FUNCTION Cell surface growth factor receptors play a vital role in translating signals from the extracellular environment into an intracellular biologic response. One such receptor is the Ron receptor tyrosine kinase. Ron, also referred to as macrophage stimulating 1‐receptor (MST1R), is a receptor tyrosine kinase (RTK) of the hepatocyte growth factor (HGF)/Met receptor family. Ron was first identified as a novel protein tyrosine kinase by screening a library prepared from a mixture of human tumors. The full‐length Ron cDNA was then identified using a human foreskin keratinocyte library (Ronsin et al., 1993). The Ron ortholog in the mouse was first cloned from hemapoietic stems cells and is also referred to as stem cell derived tyrosine kinase (STK) (Iwama et al., 1994). Met and Ron are the only two members of this RTK family, in contrast to other receptor tyrosine kinase families with multiple members (Manning et al., 2002). Ron was classified based upon its homology to Met and also by its homology to the Sea receptor found in chicken. The c‐Sea receptor is the cellular homolog of the avian oncoprotein v‐sea, and is structurally similar to Ron and Met (Huff et al., 1993; Huff et al., 1996). Sea is activated by chicken macrophage stimulating‐protein (MSP) (Wahl et al., 1999). To date, homologs of Ron and its ligand have been identified by sequence analysis in many mammalian species including Rattus norvegicus (rat), Canis lupus (dog), Bos taurus (cow), Equus caballus (horse), and Macaca mulatta (rhesus monkey) (BLAST sequence analysis, 2007). Homologs of Ron have also been found in nonmammalian species, including Fugu rubripes (puffer fish) and Strongylocentrotus purpuratus (sea urchin) (Cottage et al., 1999; Lapraz et al., 2006). The Ron and Met receptors are structurally very similar. Both Ron and Met receptors contain an extracellular ligand binding domain, a single pass

Ron Receptor in Cancer Biology

3

hydrophobic membrane spanning domain, and an intracellular region containing a tyrosine kinase domain. Ron is synthesized as a 185 kDa precursor glycosylated protein and is further processed by furin‐like proteases before being delivered as a mature receptor to the cell surface (Gaudino et al., 1994). On the cell surface, Ron exists as a heterodimeric receptor, consisting of a 35 kDa alpha chain and 150 kDa beta chain. The alpha chain is entirely extracellular whereas the beta chain contains the extracellular, transmembrane, and intracellular regions of the receptor (Gaudino et al., 1994). The 50‐amino acid tyrosine kinase domain of Ron shares 80% identity to the Met tyrosine kinase domain and overall the receptors exhibit 34% identity (BLAST sequence analysis, 2007) (Fig. 1). Human and murine Ron cDNAs share about 74% identity overall, with about 88% identity in the intracellular domains (Iwama et al., 1994). The human Ron transcript consists of 20 exons while murine Ron codes for 19 exons. Altered splicing of the murine Ron gene creates a deletion of a small juxtamembrane region that is present in the human Ron gene (Wei et al., 2005). An analysis of the mouse Ron gene promoter region showed the presence of a number of putative transcription factor binding sites important in tumor progression, including binding sites for NF‐k , Ets‐1, and the estrogen receptor (Waltz et al., 1998).

II. RON LIGAND STRUCTURE AND FUNCTION The ligand for Ron is hepatocyte growth factor‐like (HGFL) protein and is also known as MSP. HGFL was originally cloned from a human genomic library by screening for the characteristic kringle domains present in prothrombin and several other proteins in the blood coagulation system (Han et al., 1991). The protein sequence of the isolated gene was predicted to contain four kringle domains followed by a serine protease‐like domain. On the basis of domain structure, this protein was predicted to be similar to HGF, the ligand for the Met receptor. By sequence comparison, however, HGF and HGFL are only about 45% identical (BLAST sequence analysis, 2007) (Fig. 1). This newly identified protein was localized to human chromosome 3p21, a region that often displays loss of heterozygosity in cancerous tissue. The mouse gene and cDNA for HGFL were then isolated from mouse liver (Degen et al., 1991). The mouse homolog of HGFL was predicted to display the same domain structure as human HGFL and to be about 80% identical. The expression pattern of HGFL was determined by Northern analysis of tissues in the pregnant rat. The liver represented the primary site of expression for HGFL, with low levels detected in lung, adrenal gland, and placenta. Another group similarly cloned a cDNA for MSP from a library prepared from HepG2 cells, a human hepatocarcinoma

4

Purnima K. Wagh et al.

Fig. 1 The Ron and Met receptor tyrosine kinases exhibit important similarities and differences between receptors. Structurally, Ron and Met are similar in that both receptors are single‐ pass, disulfide‐linked / heterodimers. However, the amino acid identity between Ron and Met is not high (34% overall) but the intracellular region involved in signal transduction is conserved (63%). The ligands for Ron and Met, HGFL and HGF respectively, also share a similar structure and have an overall amino acid identity of 45%. In contrast to their structural similarity, HGFL and HGF are secreted ligands, which originate from different cell types, with HGFL produced as an endocrine molecule secreted primarily from hepatocytes and HGF produced from meschemycal cells operating in a paracrine fashion. Binding of HGFL or HGF to their corresponding receptor induces receptor dimerization and trans‐autophosphorylation of tyrosine residues

Ron Receptor in Cancer Biology

5

cell line (Yoshimura et al., 1993). The probe for this clone was derived from the peptide sequence of MSP that had previously been isolated from human serum. The predicted amino acid sequence of MSP also included four kringle domains and was subsequently found, like HGFL, to be most similar to HGF. HGFL and MSP were soon determined to be identical (Shimamoto et al., 1993). Two independent groups later determined HGFL to be the ligand for the Ron receptor. Further, in spite of sequence similarities, no cross activation is seen between HGFL and Met, or HGF and Ron (Gaudino et al., 1994; Wang et al., 1994b). Despite the structural similarity of HGF and HGFL, their production and mechanism of action differ. HGF is generally produced by mesenchymal cells and primarily activates the Met receptor in a paracrine fashion. HGFL is primarily produced by hepatocytes and is secreted from the liver into the blood at a concentration of about 400 ng/ml, and works in an endocrine fashion at distant sites to activate Ron (Fig. 1). These differences in ligand activation may reflect the localization of Met and Ron in normal tissue. In an analysis of normal bronchiolar ciliated epithelium of the lung, the Met receptor was localized to the basolateral cell membrane, while the Ron receptor was localized on the apical cell membrane (Sakamoto et al., 1997). Dysregulation of the spatial localization of Ron and HGFL, as well as dysregulation in the quantity of the receptor and ligand, may be important in tumor tissue growth. Since the identification of HGFL, further work has elucidated details about the promoter sequence associated with its gene. Promoter analyses have suggested that the transcription factor hepatocyte nuclear factor‐4 is important for the liver‐specific expression of HGFL (Waltz et al., 1996). Specific elements in the first intron of HGFL have also been found to regulate liver‐ and kidney‐specific expression (Wetzel et al., 2003). In addition, experiments performed in one cell type derived from large‐cell lung carcinoma have demonstrated the ability of mutant p53 to associate with the HGFL promoter and repress its transcriptional activity, leading to a decrease in HGFL mRNA and secreted protein and increased cell survival after exposure to a chemotherapeutic agent (Zalcenstein et al., 2006). Further experiments will elucidate whether this effect is conserved in other cell lines. Like HGF, HGFL is secreted as a single chain inactive precursor molecule of 80 kDa. The pro‐HGFL molecule exhibits no biological activity, nor does (1238/1239 Ron and 1234/1235 Met) in the tyrosine kinase domain, leading to the tyrosine phosphorylation of key C‐terminal residues (1353/1360 Ron and 1349/1356 Met). Activation of either receptor results in recruitment of several downstream adaptor molecules and initiation of robust signaling responses. Signaling pathways that are impacted by these receptors include the PI3‐K, Akt, ‐catenin, Ras, MAPK, and JAK/STAT pathways which induce pleiotropic biologic events such as proliferation, migration, invasion, cell scattering and branching morphogenesis.

6

Purnima K. Wagh et al.

it bind the receptor. Proteolytic cleavage results in the formation of a disulfide‐linked heterodimer of HGFL composed of an alpha (50 kDa) and a beta (35 kDa) chain. The alpha chain of HGFL contains four kringle domains while the beta chain contains a serine protease like domain. The two protein chains have distinct functions. The alpha chain is important for regulating the functional activities of Ron whereas the beta chain is important for binding of HGFL to its receptor (Danilkovitch et al., 1999a; Waltz et al., 1997). Proteases of the coagulation cascade, such as kallikrein, factor XIIa, and factor XIa, are capable of cleaving pro‐HGFL into HGFL (Wang et al., 1994c). Membrane bound proteases produced by macrophages were shown to have specific and nonspecific pro‐HGFL proteolytic activity, such that both activation and degradation of pro‐HGFL occurred at the cell surface (Wang et al., 1996c). The inhibitor of the HGFL degrading enzyme was identified as alpha 1‐antichymotrypsin (Skeel and Leonard, 2001). Interestingly, increased expression of alpha 1‐antichymotrypsin in human breast tumors, which might allow for the increased activation of HGFL, was associated with significantly poorer prognosis of patients with grade 2 and 3 breast adenocarcinomas (Hurlimann and van Melle, 1991). Estradiol treatment of breast cancer cells has also been shown to increase the production of alpha 1‐antichymotrypsin (Massot et al., 1985). Recently, the specific membrane‐bound protease that is responsible for the activation of pro‐HGFL at the cell surface has been identified by transcriptional profiling of normal tissues, cancer cell lines, and multiple types of cancer tissues, and validated by biochemical and functional testing. This enzyme is known as membrane type serine protease 1 (MT‐SP1) or matriptase (Bhatt et al., 2007). Matriptase is highly expressed in many breast, ovarian, prostate, and colon cancer cell lines (Bhatt et al., 2003). An examination of 330 node‐negative breast cancer specimens showed an association between expression of matriptase and poor patient outcome (Kang et al., 2003). An analysis of microarray gene expression data from 162 primary tumors was also analyzed for expression of Ron, HGFL, and matriptase. Overexpression of all the three was associated with significantly shorter relapse‐free survival when compared with other patients. The overexpression of all three genes also significantly improved the accuracy of prediction of a 70‐gene signature predicting poor outcome (Welm et al., 2007). Overexpression of HGFL has recently been shown to promote breast tumor growth and promote metastasis to multiple sites in a model of oncogene‐induced mouse mammary tumors (Welm et al., 2007). In this model system, orthotopically transplanted cells expressed the polyoma middle T antigen under the control of the mouse mammary tumor virus (MMTV) promoter, with or without the addition of HGFL overexpression. The additional HGFL expression significantly increased the initial growth rate of the

Ron Receptor in Cancer Biology

7

mammary tumors, but the most striking effect of ligand overexpression was the increased range of metastasis. Cells overexpressing HGFL metastasized not only to the lung, but also to lymph nodes, spleen, liver, and bone.

III. RON CHROMOSOMAL LOCATION AND CANCER Interestingly, the genes for each of the two receptor–ligand pairs, that is Met and its ligand HGF and Ron and its ligand HGFL, are located close together on the same chromosomes. Met is located on 7q31.2, and HGF is located on 7q21.11; Ron and HGFL are both located on 3p21.31 (Human Protein Atlas Version: 3.0, 2007). Both the murine Ron gene and the HGFL murine counterpart are also located on chromosome 9qF2 (UCSC Genome Browser, 2007). The human chromosome 3p21 region has been frequently observed to undergo loss of heterozygosity in cancer specimens and cell lines, suggesting that this region may harbor tumor suppressor genes. Using the sensitive detection method of quantitative real‐time PCR to examine cervical carcinoma, it was recently shown that aberrations in the 3p21 region are complex and may involve gene amplification as well as deletion (Senchenko et al., 2003). Aberrations in the 3p21 chromosome region have also been examined in lung cancer cell lines, and renal cell and breast carcinoma biopsy material. Amplification of 3p is a common event in these cancers, occurring in 15–42.5% of the samples examined (Senchenko et al., 2004). This amplification of the chromosome region containing Ron and HGFL is consistent with the overexpression of Ron seen in many human tumor types, although direct evidence for the amplification of Ron in these human tumors has not yet been produced.

IV. RON IN MACROPHAGES: INFLAMMATION AND CANCER The determination of the expression of Ron in normal tissues and cells has helped to define its normal roles and the signaling pathways that are activated during transformation from normal cell to tumor cell. The initial characterization of the effect of HGFL was on mouse resident peritoneal macrophages. Stimulation by this ligand caused shape changes, altered response to chemoattractants, and stimulated phagocytosis in macrophages (Skeel et al., 1991). Through absorption studies, it was determined that HGFL was binding to a receptor and activating mature resident macrophages (Skeel and Leonard, 1994). Further studies demonstrated Ron to be expressed on

8

Purnima K. Wagh et al.

human alveolar, peritoneal macrophages, and monocyte‐derived macrophages, but not on circulating human monocytes (Brunelleschi et al., 2001). Ron, through HGFL stimulation, was shown to play an inhibitory role in regulating nitric oxide production by macrophages (Wang et al., 1994a). Further, mice with a defect in Ron signaling have altered inflammatory responses in vivo (Waltz et al., 2001). The link between Ron, inflammation, and cancer has had little attention. However, it is becoming increasingly evident that chronic inflammatory processes contribute to the development of cancer (Federico et al., 2007; Perwez Hussain and Harris, 2007). Many papers have described the ability of nitric oxide, a known mediator of inflammation, to alter neoplastic effects (Hussain et al., 2004). Ron has been shown to be a negative regulator of nitric oxide in epithelial cells as well as macrophages (Hess et al., 2003b). Moreover, in macrophages, the various effects of Ron, including superoxide anion production, apoptotic resistance, and phagocytosis, are induced through interactions with diverse signal molecules, including Src, Erk, p38, and PI3‐K/Akt, which have been implicated in tumorigenesis (Brunelleschi et al., 2001; Chen et al., 1998; Lutz and Correll, 2003).

V. DEVELOPMENTAL ROLES OF RON AND TUMOR PROPERTIES The expression of Ron in normal development also may indicate some future role in tumorigenesis. The expression of Ron mRNA was determined in normal mouse tissues at different stages of development (Gaudino et al., 1995; Quantin et al., 1995). Expression of Ron was found in the liver as early as day 12.5, but expression in other tissues appeared at later stages of development, from day 13.5 to 16.5, and was present in the adult. There have been some contradictory reports concerning the expression of Ron in different tissues and in cell lines. One reason for this discrepancy may be the very low level of Ron that is present in normal tissue. An estimation of the number of Ron receptors per cell was calculated by determining the saturation kinetics of binding of HGFL to BK‐1 cells, a normal keratinocyte cell line. Keratinocytes had been shown to express Ron and were responsive to HGFL stimulation in functional assays. The estimated receptor number per cell using this method was about 600–1200 (Wang et al., 1996a). Several other keratinocyte cell lines showed equivalent binding levels. In contrast, a receptor binding study was used to estimate the number of epidermal growth factor receptors (EGFRs) in NIH3T3 fibroblasts to be about 70,000 per cell (Roque et al., 1992).

Ron Receptor in Cancer Biology

9

Ron expression has been seen in the glandular epithelium of the gastrointestinal tract, including the stomach and colon, adrenal glands, testis; kidney, the central and peripheral nervous system, and ossification centers of developing bone. Ron is expressed in ovaries and in mammary tissue (Chodosh et al., 2000; Hess et al., 2003b). Ron protein is also expressed in tumor cells from the breast, colon, pancreas, liver, gastric system, kidney and lung, and haematopoietic cells (Gaudino et al., 1994). Ron appears to be expressed in nearly every tissue tested, at low levels, and good agreement from several studies finds that Ron is expressed in most epithelial tissues. Although the role that Ron plays in tumor formation and growth are still under investigation, some of the functions of Ron in normal development suggest mechanisms by which Ron may influence cancer progression. Ron is expressed in reproductive, hormone‐dependent mouse tissues, including uterus, placenta, testis, and epididymis, and HGFL transcripts are present in the cervix, placenta, epididymis, and testis. Ron is expressed during the process of mouse embryo implantation and placentation. In vivo, Ron is expressed in the invading ectoplacental cone and trophoblast giant cell regions surrounding the implanting embryo. Using several murine trophoblast cell lines, HGFL stimulation has been shown to increase invasion through a basement membrane component material (Matrigel) and to enhance cell survival (Hess et al., 2003a). In liver progenitor cells, the Ron receptor induces additional cell responses in response to ligand stimulation, including cell scattering (motility), DNA synthesis, and extracellular matrix invasion (Medico et al., 1996). These normal cellular responses are also mechanisms by which tumor cells propagate, invade, and metastasize.

VI. EPITHELIAL TO MESENCHYMAL TRANSITION Another hallmark of the progression from normal epithelium to tumor development is termed the epithelial to mesenchymal transition (EMT). EMT is a process that is characterized by loss of epithelial differentiated morphology and reversion to mesenchymal phenotype. Cells undergoing EMT demonstrate a transition from cuboidal to spindle‐shaped morphology, a reorganized actin cytoskeleton, and the expression of mesenchymal cellular marker proteins. Ron activation by HGFL has been shown to induce a motile‐invasive phenotype marked by dissociation or cell scattering and matrix invasion, characteristics resembling EMT. The characteristics that mark EMT were also evaluated in MDCK cells expressing Ron. Constitutive expression of Ron was shown to induce EMT, marked by phenotypic changes and alterations in cell motility (Wang et al., 2004). A collaborative effect of HGFL and TGF‐ 1 in EMT was also demonstrated. These results

10

Purnima K. Wagh et al.

demonstrate that Ron overexpression alone or in combination with HGFL stimulation can induce traits that promote tumorigenic properties such as EMT, cell migration, and matrix invasion.

VII. ONCOGENIC POTENTIAL OF THE RON RECEPTOR The oncogenic potential of Ron and its role in cellular transformation has been investigated with in vitro and in vivo experimental systems. Stable expression of wild‐type and constitutively active murine Ron mutants in NIH3T3 mouse fibroblast cells were investigated for transforming potential. The point mutations in the Ron gene were analogous to those found in the Met receptor tyrosine kinase in hereditary papillary renal carcinoma (HPRC), and had also been found in somatic mutations in renal carcinoma. Two of the point mutations were also analogous to activating mutations in the Ret and Kit oncogenes. Both overexpression of wild‐type murine Ron and the activating mutations induced receptor phosphorylation and transformation of the fibroblasts, as determined by phenotypic changes and foci formation. These transformed cells also demonstrated increased proliferation rates and increased motility. The NIH3T3 cells overexpressing wild‐type or mutant Ron formed tumors when injected into nude mice. Cells expressing a point mutation in the kinase domain (M1231T) and those expressing wild‐type Ron showed equivalent tumor latency and 100% tumor formation in the nude mice. To determine whether these transformed cells exhibited metastatic potential in vivo, NIH3T3 cells injected into nude mice were tested for both spontaneous and experimental metastasis. Mutation M1231T was the most aggressive form, and showed spontaneous and experimental metastasis to lungs (Peace et al., 2001). The oncogenic potential of similar point mutations in the human Ron gene has also been investigated (Williams et al., 1999). The point mutations D1232V and M1254T in the tyrosine kinase domain of the Kit and Ret receptor respectively are found in human malignancies mastocytosis and multiple endocrine neoplasia type 2B. Mouse NIH3T3 fibroblasts transfected with these Ron mutants produced transformed cells that formed foci. Constitutive phosphorylation of Ron and kinase activity of the receptor was shown for both the mutants and for the wild‐type overexpressed receptor, although the mutant forms were more active. These same mutant forms were also examined for tumor formation when injected into nude mice. Both mutant forms produced tumors in nude mice and were highly metastatic. Overexpression of both these mutant receptors in fibroblasts induced constitutive Ron receptor phosphorylation. Phosphorylation and constitutive activation of Ron also led to activation of its downstream target, the mitogen‐activated protein kinase (MAPK) (Santoro et al., 1998).

Ron Receptor in Cancer Biology

11

A constitutively active form of Ron has also been produced as a Tpr‐Ron chimera that mimics the oncogenic form Tpr‐Met (Santoro et al., 1996). The properties of this constitutively active Ron were also examined after transfection into NIH3T3 fibroblasts. The constitutive activation of Ron produced by this chimera produced a phenotype that is highly relevant to tumor progression and metastasis, marked by cell scattering, cellular motility, and invasion of an extracellular matrix.

VIII. LOSS OF FUNCTION MOUSE MODELS FOR RON To dissect the function of Ron in vivo, several different mouse models with defects in Ron were produced. A mouse model with total loss of Ron protein was produced by a global deletion of exon 1–14 of the mouse Ron gene. This strategy knocks‐out completely a large genomic region of Ron containing Ron 50 ‐flanking sequences, the extracellular domain, the transmembrane domain, and a portion of the intracellular domain of the Ron gene. Strikingly, mice with this large deletion of Ron are lethal at an early stage of embryo development (e7.5) (Muraoka et al., 1999). Mice that were hemizygous for this deletion of Ron were viable and fertile, but displayed an enhanced response to inflammation. The hemizygous mice were more susceptible to endotoxic shock and displayed an impaired ability to regulate nitric oxide, demonstrating the role of Ron in regulating these functions. Nevertheless, the lethality of this mutant line made it impossible to further dissect the role of Ron in different tissues in vivo. Therefore, a mouse model in which the signaling function of Ron could be ablated was designed and produced. A mouse model was produced in which the extracellular and transmembrane domains of Ron are preserved, along with eight amino acids of the intracellular domain, while the ablation of the remainder of the cytoplasmic domain of Ron results in complete loss of Ron intracellular signaling (Waltz et al., 2001). Homozygous mice with this germline deletion, referred to at the TK/ mice, are viable, fertile, and display no gross phenotypic abnormalities. However, the Ron receptor plays an important role in macrophage‐ mediated inflammatory response by limiting nitric oxide production and thereby attenuating its harmful effects. In the absence of Ron signaling, the Ron TK/ mice show an enhanced response to both acute and cell‐ mediated inflammatory stimuli. This model has also been used to examine the role of Ron signaling in oncogene‐mediated tumorigenesis. A similar enhanced response to inflammation was observed in another mutant Ron mouse model (Correll et al., 1997). In this case, the gene targeting strategy inserted a ‐galactosidase gene into exon 1 of the mouse Ron gene, so that transcription of the reporter would arise from the endogenous reporter and would be translated from its own start site. With this

12

Purnima K. Wagh et al.

strategy, homozygous mutant mice were produced that were viable and phenotypicaly normal. In this model, the insertion in exon 1 probably produced a functionally hypomorphic allele. Although this insertion ablated the activity of Ron arising from ligand binding, it is probable that some functions of Ron were still preserved by the production of known alternate splicing forms that did not require exon 1. Nevertheless, the preponderance of evidence in all three mutant mouse models demonstrates that the Ron gene plays a significant role in the negative regulation of inflammatory responses. The expression of the ligand for Ron, HGFL, has also been deleted in a mouse model (HGFL/) (Bezerra et al., 1998). The global deletion of HGFL in mice leads to no gross phenotypic abnormalities, and the mice were fertile. Histological examination of mouse tissues revealed the presence of lipid‐filled vacuoles in hepatocytes in the HGFL/ mice, but the significance of these vacuoles has not been determined at this time. The impact of ligand‐mediated signaling in Ron‐overexpressing tumors has not been determined at this time.

IX. LOSS OF RON FUNCTION AND TUMORIGENESIS To examine the significance of Ron in mammary tumorigenesis and metastasis, mice with a global deletion of the Ron tyrosine kinase intracellular signaling domain (Ron TK/) were crossed with mice predisposed to mammary cancer through expression of polyoma virus middle T antigen (pMT) under the control of the MMTV promoter (MMTV‐pMT) (Peace et al., 2005). The MMTV‐pMT mouse is a well‐characterized model in which 100% of the mice develop mammary tumors by three months of age. The mammary tumors in MMTV‐pMT mice metastasize to the lung. In this model, loss of Ron signaling (MMTV‐pMT/Ron TK/) markedly impacted mammary tumor latency, tumor growth, and metastasis compared to mice with intact Ron signal function (MMTV‐pMT/Ron TKþ/þ). Loss of Ron signaling significantly delayed tumor initiation and growth, and reduced metastasis. Loss of Ron signaling reduced tumor angiogenesis, decreased cell proliferation, and increased tumor apoptosis. In this model, the experiments demonstrated that Ron impacted tumorigenesis through the MAPK and Akt signaling pathways. Loss of Ron signaling was also examined in the context of skin carcinogenesis using a model of chemically‐induced Ras‐mediated skin cancer (Chan et al., 2005). Mice expressing a mutated Ras transgene (v‐Ha‐Ras; Tg.AC) were crossed to mice deficient in the Ron tyrosine kinase domain (TK/). Mice expressing the mutated Ras transgene and deficient in Ron signal function (Tg.ACþ//Ron TK/) and mice expressing the mutated Ras

Ron Receptor in Cancer Biology

13

transgene with wild‐type Ron signal function (Tg.ACþ//Ron TKþ/þ) were treated with 12‐O‐tetradecanoylphorbol‐13‐acetate (also known as TPA or PMA). This chemical treatment of the Ha‐Ras‐transgenic mice has been shown to induce the formation of papillomas, some of which undergo malignant conversion. Loss of Ron signaling resulted in an increased number of papillomas, but these papillomas showed significantly reduced growth. Most notably, loss of Ron signaling significantly reduced the number of papillomas that underwent malignant conversion, as well as reducing the number of other malignant tumor types found in these mice. The expression of Ron protein was found to be upregulated during TPA treatment. As had been found previously in the mammary carcinogenesis model, loss of Ron signaling impacted tumorigenesis through the MAPK and Akt signal pathways.

X. GAIN OF FUNCTION MOUSE MODELS FOR RON OVEREXPRESSION IN TUMORS Two mouse models that overexpress Ron in different organ systems have been developed, and the effect of the overexpression of Ron on tumor development in those organs has been analyzed. One model overexpressed the human Ron gene in the lung by driving expression of Ron with the lung‐ specific surfactant C promoter (SPC) (Chen et al., 2002). Multiple adenomas developed at an early age in these mice. However, these adenomas did not progress to a malignant state. The adenomas were analyzed for point mutations in p53 and K‐Ras, since mutations in these genes are frequently associated with lung tumors in mouse models, but no mutations were found in these genes in the time period under study. However, some indication of limited genomic instability was seen in individual tumors. In addition, the expression level of Ras, an important oncogene, was elevated in these adenomas. These data suggest that while Ron overexpression in the lung has oncogenic potential, progression to a malignant lesion may require additional genetic alterations in the lung. A mouse model overexpressing murine Ron, driven by the MMTV promoter, was developed in order to analyze the role of Ron overexpression in mammary tumorigenesis (Zinser et al., 2006). These mice developed hyperplastic mammary glands by 12 weeks of age. Ron overexpression was sufficient for the development of mammary tumors in 100% of the female animals. The tumors overexpressing Ron were also found to be highly metastatic to liver and lung, and nearly 90% of the animals developed metastases. Ron overexpression was associated with receptor phosphorylation and kinase activity. The tumors were also found to overexpress cyclin

14

Purnima K. Wagh et al.

D1 and c‐myc, which have been associated with poor prognosis in human breast tumors. In addition, overexpressed Ron was associated with tyrosine phosphorylated ‐catenin. The association of Ron and activated ‐catenin, and the consequent upregulation of the ‐catenin target genes cyclin D1 and c‐myc, produces one plausible mechanism for the tumorigenic activity of Ron in breast cancer.

XI. MECHANISMS OF RON‐INDUCED TUMORIGENESIS: SIGNALING THROUGH THE RON RECEPTOR The pathways by which the Ron receptor conducts signals from the extracellular environment to the intracellular environment have been studied. However, the relationships of these different pathways to the specific biologic responses that are relevant to tumor formation are still poorly defined. Certain pathways appear to be commonly activated in many tumor types, whereas the responses of other signals may be cell‐type specific. Ron activation by ligand binding and signaling via downstream adapter molecules has been shown to promote pleotrophic effects dependent on cell type (Iwama et al., 1996). The most prominent oncogenic pathways implicated in Ron signaling to date are activation of PI3‐K/Akt, MAPK, Ras, Src, and ‐catenin. A preponderance of evidence in a number of tumor types indicates that a major mode of action of Ron in cancer is to promote cell survival via resistance to apoptosis. Both the MAPK and the PI3‐K signal pathways have been implicated in this antiapoptotic action, with both pathways contributing to the effect generated by ligand stimulation of Ron (Danilkovitch et al., 2000). The activation of PI3‐K leads to activation of Akt, which has been shown to enhance cell survival, but is not required for metastasis (Hutchinson et al., 2001). Ligand binding of HGFL to the Ron receptor leads to phosphorylation of tyrosine residues on the C terminus of the chain of Ron. As with other receptor tyrosine kinases, activation of the kinase domain of Ron is thought to depend on receptor dimerization and trans‐autophosphorylation of tyrosine residues. The phosphorylation of two tyrosine residues within the carboxyl‐terminus (Y1353 and Y1360) is required for the biological activities of Ron (Fig. 1). These tyrosine residues serve as docking sites for signaling molecules having Src homology‐2 (SH2) and the phosphotyrosine binding domains (PTB). Grb2 via its SH2 domain binds directly to the activated Ron receptor and allows recruitment of Son of sevenless (SOS) to the SH3 domain of Grb2. SOS activates Ras, which recruits Raf to the membrane. Raf in turn activates MEK, leading to Erk activation and the

Ron Receptor in Cancer Biology

15

transcription of pro‐proliferative genes (Li et al., 1995). SHC via its PTB domain also binds directly to the phosphorylated tyrosine in the C‐terminal region of Ron and SHC‐Grb2‐SOS together can also activate the Ras pathway in response to HGFL. Grb2 may also act as an adapter to indirectly recruit multiple proteins to Ron (as is the case for Met), including the docking protein Grb2‐associated binding protein‐1 (Gab1) and Cbl ubiquitin ligases. Gab1 can also bind to membrane phosphotidyl‐inositol 3,4,5‐ triphosphate (PIP3) via pleckstrin homology domain. When Gab1 and Ron are expressed in COS cells, Gab1 directly associates with tyrosine phosphorylated Ron through the Met binding domain (MBD) of Gab1. Gab1 can also directly associate with variety of signal transducers including PI3K, phospholipase‐C (PLC‐ ), and SHP2 phosphatase (van den Akker et al., 2004). Gab1‐mediated signaling is important for inducing the branching morphogenesis (Maroun et al., 2000). PI3K can also interact with Ron receptor either directly or through adaptor molecules (Danilkovitch et al., 2000). Activation of the Ras pathway is important for the program leading to invasive growth and PI3‐K‐dependent activation of Akt activation is important for cell migration and survival (Danilkovitch et al., 1999b; Wang et al., 1996b). Ron is a strong inducer of both PI3‐K and MAPK signaling pathways in vivo and in vitro. Tumor cell lines with a knockdown of Ron exhibit a diminution of basal phosphorylated MAPK and Akt (Wagh and Waltz, unpublished results). Moreover, in mammary tumors from mice expressing pMT under MMTV promoter, loss of Ron receptor signaling leads to a significant decrease in pMAPK and pAkt in tumor lysates compared with that in mice with wild‐type Ron (Peace et al., 2005). These studies demonstrate the reliance of MAPK and PI3K/Akt signaling on Ron receptor expression. Many human cancers have high cellular levels of ‐catenin, and ‐catenin plays a dual role in cell adhesion as well as acting as a transcription factor. Overexpression of activating Ron mutants M1254T and D1232V in NIH3T3 cells caused increase in cellular accumulation of ‐catenin, which thereby upregulated ‐catenin responsive oncogenes c‐myc and cyclin D1. Mutant Ron kinase caused tyrosine phosphorylation of ‐catenin thereby increasing its stability and preventing degradation by the axin/GSK‐3 complex (Danilkovitch‐Miagkova et al., 2001). The interaction of Ron with the extracellular matrix is important for the characteristic biological activity of Ron in promoting cell migration, and may also be important for its activity in promoting cell survival. Both of these biological activities, migration and enhanced survival, may contribute to the role that Ron plays in metastasis. Ron has been shown to directly interact with integrins (Danilkovitch‐Miagkova et al., 2000). Cellular adhesion to extracellular matrix induced phosphorylation of Ron, and this

16

Purnima K. Wagh et al.

activity was dependent on the kinase activity of Ron and of Src. In keratinocytes, HGFL stimulation of Ron was shown to lead to phosphorylation of the Ron receptor and also phosphorylation of 6 4 integrin (Santoro et al., 2003). This interaction leads to the generation of 14‐3‐3 binding sites on Ron and the integrin, and the linkage of these molecules through the dimeric 14‐3‐3. This interaction is important for cell spreading and migration.

XII. RECEPTOR CROSS‐TALK AND RON ACTIVITY IN TUMORIGENESIS Another means of activating Ron signaling may be through the interaction of Ron with other receptors. This interaction between receptors of different types has been termed receptor cross‐talk. Interaction between dissimilar receptors may play a role in stimulating receptor activity independent of ligand activity. However, receptor cross‐talk may also retain responsiveness to ligand‐induced activation. Both direct and indirect evidence exists that Ron interacts with other receptor types. This receptor cross‐talk may be especially important for tumor progression, since other interacting receptors have also been shown to be upregulated in tumors. Ron is of course most closely related to the Met receptor, which is a known protooncogene. Accordingly, the regulation, expression, and interaction of Ron and Met have been studied in several normal tissues and tumor types. The regulation of expression of Met and Ron was examined in normal liver, hepatocellular carcinoma (HCC) tissues, and cell lines derived from HCC. Both Ron and Met were expressed in normal liver tissue. Both receptors were also overexpressed in a subset of HCC tumor tissues. The expression of Met and Ron was induced by the treatment of HCC cell lines with HGF, interleukin‐1 and ‐6, and tumor necrosis factor alpha. Met and Ron expression appeared to be modulated in liver tumors by a similar cytokine network. The interaction between Met and Ron was investigated by expressing full‐ length and kinase‐inactive combinations of the two receptors in COS cells (Follenzi et al., 2000). When wild‐type Met and Ron receptors were transiently expressed in COS cells, trans‐autophosphorylation of tyrosine residues occurred in ligand‐independent manner. However, treatment with either HGF or HGFL ligand increased the trans‐autophosphorylation of the two receptors. By expressing a wild‐type Ron receptor with a Met receptor in which the docking site tyrosines were deleted, or vice versa, it was demonstrated that transphosphorylation of Ron and Met occurred directly, rather than through a secondary signal transduction molecule. Through cross‐ linking of the proteins, Met–Ron complexes were detected on the cell surface,

Ron Receptor in Cancer Biology

17

prior to ligand‐induced dimerization. Kinase‐dead Ron‐inhibited (mutant) Met induced transforming ability of NIH3T3 cells, suggesting that Ron increases transforming ability of mutant Met (Follenzi et al., 2000). The cross‐talk between Ron and Met is also relevant to ovarian cancer (Maggiora et al., 2003). When a panel of human ovarian carcinoma tissues was evaluated, Ron and Met were significantly coexpressed in 42%. The mechanism by which cross‐talk of Met and Ron could impact ovarian cancer was examined in vitro. The motility and invasiveness of ovarian cancer cells was stimulated by the addition of ligand for either receptor, but was synergistically enhanced by the coadministration of both ligands. The cross‐talk between Ron and Met in ovarian cancers that overexpress both receptors may promote tumor progression. A similar situation exists for cross‐talk between Ron and Met in breast cancer (Lee et al., 2005). When Ron and Met expression was determined by immunohistochemistry on a panel of human invasive ductal breast carcinoma tissue samples, it was found that Ron and Met expression were independent predictors of distant metastasis. This clinical property correlates well with the observation that Ron influences cell scattering, motility, and invasiveness. Overall the synergism between Ron and Met can confer an aggressive phenotype to breast cancer. A multivariate retrospective analysis of clinical outcome was performed to determine the risk of the overexpression of Ron and Met in breast cancer. This analysis controlled for tumor size; tumor grade; and estrogen receptor, bcl‐2, HER2/neu, and p53 status. In patients with overexpression of both Ron and Met, the likelihood of 10‐year disease‐free survival was only 11.8%, compared to 79.3% in patients with tumors that were negative for both receptors. Decreased survival was also significantly associated with coexpression of Ron and Met in 19.1% of a cohort of 183 patients with transitional‐cell bladder cancer (Cheng et al., 2005). Overexpression of Ron in bladder cancer cell lines increased cell proliferation, motility, and survival. There is mounting evidence that cross‐talk between Ron and Met may be a significant factor in subsets of various types of epithelial tumors. Another tyrosine kinase receptor that is frequently overexpressed in many different tumor types, and has been a target for cancer therapeutic drug development for this reason, is the EGFR. To determine the role of EGFR in Ron‐induced cellular transformation, a dominant‐negative form of human EGFR was overexpressed in cells stably expressing mouse Ron (Peace et al., 2003). This dominant‐negative EGFR markedly reduced the scattering of these cells that is the normal response to treatment with HGFL ligand. Cell scattering was also reduced when EGFR was chemically inhibited. Cotransfection of dominant‐negative Ron with wild‐type EGFR, or cotransfection of dominant‐negative EGFR with wild‐type Ron both produced significantly fewer transformed foci compared to transfection of wild‐type

18

Purnima K. Wagh et al.

Ron or wild‐type EGFR receptor alone. Transphosphorylation of both receptors was induced when cells overexpressing murine Ron and expressing endogenous EGFR were stimulated with either HGFL or the EGFR ligand epidermal growth factor (EGF). Coimmunoprecipitation and activation of phosphatidyl inositol 3‐kinase (PI3‐K), a downstream signal molecule that has been shown to play a role in cell motility, was observed after stimulation with either ligand. The coexpression of Ron and EGFR also has clinical significance in primary transitional‐cell carcinoma of the bladder (Hsu et al., 2006). In a cohort of bladder cancer patients, Ron and EGFR expression was found in 33.3% of the tumor samples analyzed. Receptor coexpression was significantly associated with tumor invasion, risk of local recurrence, and decreased survival. The interaction between Ron and EGFR was also examined in a bladder cancer cell line that expresses high levels of both Ron and EGFR. The interaction between Ron and EGFR was found to be ligand‐ independent. The knockdown of either Ron or EGFR expression via the transfection of small interfering RNA (siRNA) reduced ligand‐independent phosphorylation of both receptors, although interestingly, the reduction in phosphorylation of EGFR by knockdown of Ron was greater than the reverse. The inhibition of EGFR activity by either siRNA or by treatment with small molecule inhibitors of EGFR also impacted biological effects mediated by Ron, with a reduction in proliferation, migration, survival, and foci formation. In total, these results indicate that cross‐talk between Ron and EGFR may be an important mode of activation and stimulation of biological activities mediated by Ron in both a ligand‐dependent and ligand‐ independent manner. Ron cross‐talk has also been shown to occur with two other classes of receptors that are less well‐characterized for relevance to cancer. Ron has been shown to interact with the interleukin‐3 (IL‐3) receptor common chain (Mera et al., 1999). Cross‐talk between these receptors after HGFL‐ ligand stimulation was shown to modulate downstream signal pathways through activation of the JAK2 signal transduction molecule, and to tip the balance of cellular activity toward shape change that is relevant to cell motility rather than to cell proliferation. Another class of receptors that may cross‐talk with Ron are the plexins. Plexins are transmembrane receptors for semaphorins, a class of secreted molecules that were first characterized for axonal growth cone guidance. However, plexins are also overexpressed in variety of human cancers, including pancreas, colon, and liver. Ron shares structural and functional similarities with plexins. Sema 4D, a ligand for B1 plexin, caused an increase in the invasiveness of NIH‐3T3 cells expressing Ron. Saturating concentrations of HGFL and 100 nM of Sema 4D synergistically increased NIH3T3 cell invasion as compared to controls (Conrotto et al., 2004).

Ron Receptor in Cancer Biology

19

XIII. ANGIOGENESIS It has been well established that progressive tumor growth requires de novo blood vessel production, and that tumors produce angiogenic chemokines to fulfill the recruitment and growth of these blood vessels. The development of antiangiogenic tyrosine kinase inhibitors, such as those that target vascular endothelial growth factor receptors (VEGFR), are an area of intensive research, and have moved rapidly into patient treatment (Kesisis et al., 2007). The role of the Ron receptor tyrosine kinase in mediating angiogenic signals is an intriguing area that has had little attention to date. The first report that Ron may play a role in tumor angiogenesis was produced in an examination of Ron signal function in mammary carcinogenesis (Peace et al., 2005). Tumors induced by polyoma middle T expression, with or without Ron signaling, were examined in blood vessels by immunohistological staining. It was demonstrated that the ablation of Ron signaling was associated with a significant reduction in microvessel density. Further studies are required to define the significance of Ron in tumor angiogenesis.

XIV. GENOMIC INSTABILITY AND CELL CYCLE DISRUPTION In recent work, the effect of Ron overexpression on genomic instability in the mouse model of mammary tumorigenesis has been examined (Zinser et al., 2006). Primary cells derived from tumors were shown to display aberrant cell cycle kinetics and mitotic defects. These tumor‐derived cells showed a high level of inherent DNA damage, as evidenced by the phosphorylation of substrates of ATM, and an accumulation of the cell cycle checkpoint protein Cdc25A. The accumulation of Cdc25A prompted the examination of Chk2, a cell cycle modulator of Cdc25A stability. Chk2 was also of interest, since point mutations in this gene have been shown to be a risk factor for human breast cancer. An interaction between Ron and Chk2 that converges on the Cdc25A protein was determined. This work explores a previously unexamined role for Ron in genomic stability in cancer.

XV. RON EXPRESSION IN HUMAN TUMORS AND TUMOR‐DERIVED CELL LINES The growing awareness of the potential role for Ron in human cancer has lead to a recent examination of Ron expression in a range of human tumor types and tumor‐derived cell lines (O’Toole et al., 2006). Panels of human

20

Purnima K. Wagh et al.

tumor tissue were analyzed for the extent and intensity of Ron staining, and covered tumors of the breast, lung, prostate, gastric tissue, pancreas, and colon. The number of tumor tissues in these arrays ranged from 38 to 55. The percent of tissues that were positive for Ron expression ranged from 65% in colon cancer to 100% in breast cancer, with high staining intensities found in epithelial cells. A large number of cancer‐derived cell lines were also analyzed for Ron expression, and positive cell lines were found that were derived from breast, lung, prostate, pancreas, and colon, ovary, stomach, and liver. The involvement of overexpressed Ron in tumors of epithelial origin reflects its wide distribution in epithelial cells. The following sections will briefly describe the current information that is available about Ron expression in different tumor types.

A. Breast Cancer The most compelling information about the overexpression of Ron in tumor tissue is demonstrated in breast cancer. The first report examining the expression of Ron in human breast tumor tissue showed that Ron is overexpressed in about 50% of breast tumors. Its expression is very low in normal mammary gland and in benign lesions but is significantly higher in primary breast carcinomas (Maggiora et al., 1998). Ron receptor is highly expressed in epithelial breast cancer cells including T47D and ZR 75‐1 cells (Gaudino et al., 1994). HGFL is able to induce Ron activation in T47D cells, and stimulation of Ron receptor in ZR 75.1 cells causes increased cell proliferation, invasion, and about a 12‐fold increase in migration (Gaudino et al., 1994; Maggiora et al., 1998). Interestingly, a feline form of Ron was found to be overexpressed in about 33% of archival feline mammary carcinoma samples tested (De Maria et al., 2002). Mouse models have also demonstrated an important role for Ron in mammary tumorigenesis (Peace et al., 2005; Zinser et al., 2006).

B. Prostate Cancer The recent work showing that Ron is expressed in breast cancer suggests that Ron may be important in prostate cancer as well. A recent survey of Ron expression in human cancer showed that Ron is expressed in 92% of the prostate tumor tissues examined, and that Ron was highly expressed in several prostate cancer cell lines including PC‐3, DU145, and LnCAP (O’Toole et al., 2006). The Waltz laboratory has examined a panel of human prostate tissue in order to determine the relationship of Ron expression to tumor stage. Preliminary data indicates that Ron expression is limited in the normal prostate epithelium, and expression increases progressively with stage of disease in benign prostate hyperplasia, compared to prostate

Ron Receptor in Cancer Biology

21

adenocarcinomas and prostate metastases. Preliminary results also indicate that Ron may play an important role in prostate cancer in vivo as well.

C. Pancreatic Cancer Since Ron appears to be associated with ductal epithelium, such as breast tissue, the expression and function of Ron in pancreatic cancer was examined. Ron receptor is highly expressed in several human pancreatic cell lines, including BxPC‐3, CFPAC‐1, FG, and L3.6 pl and murine pancreatic cancer cell lines 4964PDA, 4964LM, 5143PDA, and 5143LM (Camp et al., 2007; Thomas et al., 2007). Ron activation in the human pancreatic cell line L3.6 pl leads to activation of the Erk and Akt pathways that are downstream of Ron and showed characteristics of EMT, including HGFL‐induced L3.6 pl cell shape changes, migration, and invasion. Migration and invasion in these cell lines in response to ligand stimulation was blocked by a neutralizing monoclonal antibody against Ron. The L3.6 pl cells also showed loss of E‐cadherin and increased nuclear translocation of ‐catenin in response to HGFL stimulation. Monoclonal antibody blockage of Ron signaling successfully decreased subcutaneous and orthotropic tumors growth formed by injecting human L3.6 pl cells into nude mice (Camp et al., 2007). The response of migration and invasion after ligand stimulation was also shown in several other pancreatic cell lines (Thomas et al., 2007). Ron expression was also examined in human pancreatic tumor tissue samples by immunohistochemistry. Ron expression was very low in normal ductal epithelia, and significantly increased in invasive and metastatic cancers. In one report, 93% of the human pancreatic cancer tissues showed overexpression of Ron relative to normal ductal epithelium (Camp et al., 2007). In another report, 79% of the primary pancreatic cancers and 83% of the metastatic lesions overexpressed Ron. In addition, 100% of eight invasive carcinoma tissue specimen tested for phophorylated Ron receptor had positive staining (Thomas et al., 2007). A mouse model of pancreatic cancer (PdxCre/LSL‐KRASG12D) was also examined for overexpression of Ron by immunohistochemistry at 6 months of age. In this model, the increase in Ron expression with the progression of disease was similar to that seen in human tissue samples, with normal pancreatic ducts showing very low‐level Ron expression that increased with tumor grade (Thomas et al., 2007).

D. Renal Tumors Both Ron and HGFL have been shown to be expressed in normal human renal tissue (Rampino et al., 2002). Although the liver is the primary site of HGFL ligand production, it has also been shown that HGFL is produced by

22

Purnima K. Wagh et al.

cultured tubular cells of the kidney in vitro, and that cultured human mesangial cells express Ron and are activated by HGFL from tubular cell supernatant. This ligand stimulation was found to induce proliferation, migration, and invasion of the mesangial cells, properties which are important in tumorigenesis. The expression of Ron was then studied in a number of different renal tumor types (Rampino et al., 2003). Ron was strongly expressed in a phosphorylated form in oncocytomas, a benign tumor, and was not found in renal carcinomas. The mechanism by which Ron promoted this tumor growth appeared to be predominately by opposing apoptosis rather than inducing proliferation.

E. Bladder Cancer Immunohistochemical analysis of a panel of bladder cancer specimens showed that Ron was overexpressed in 32.8% of the primary tumors, and 23.3% of these positive tumors showed high levels of expression. Overexpression of Ron in vitro in a bladder cancer cell line increased cell proliferation, motility, and resistance to apoptosis (Cheng et al., 2005). Ron cross‐ talk with the Met receptor and with the EGFR receptor was shown in bladder cancer cell lines (Cheng et al., 2005; Hsu et al., 2006).

F. Ovarian Cancer Ron expression was detected in 55% of the human ovarian cancer tissues specimens that were examined. HGFL stimulation was examined in this tissue type in vitro, and caused increased motility and invasion of SK‐OV3 ovarian carcinoma cells that have high‐level Ron expression (Maggiora et al., 2003).

G. Lung Cancer Ron is expressed in normal lung, and is localized to the apical surface of ciliated epithelium. Stimulation of Ron with its ligand HGFL increased ciliary beat frequency, and therefore Ron may play a role in mucociliary lung clearance. Ron has also been identified in small cell carcinoma of the lung (SCLC), in a pulmonary carcinoid cell line, and in a SCLC cell line (Willett et al., 1997). Ron was also examined in nonsmall cell lung cancer (NSCLC) (Willett et al., 1998). Ron was expressed in both primary human tumor tissue and in NSCLC cell lines. In vitro tests of Ron activity in these cell lines showed that ligand stimulation induced Ron phosphorylation,

Ron Receptor in Cancer Biology

23

showing that Ron was active, and that ligand stimulation increased cell motility, an important component of metastasis.

H. Gastrointestinal Tumors The distribution and expression level of Ron in normal human gastrointestinal organs was examined by immunohistochemistry, and a comparison was made of expression between adult and fetal tissue (Okino et al., 2001). High‐level expression was seen in the esophagus, small intestine, and colon, but in gall bladder was negative. Immunoreactivity for Ron was strong in fetal stomach and pancreas, but was faint in these organs in the normal adult tissues. This result suggests that Ron may be associated with differentiation in these organs, and that it may function as an oncofetal protein. Several splicing variants of Ron were initially isolated from gastrointestinal origin cell lines and tissues. The first was isolated from a gastric cancer cell line, and was termed
Ron (Collesi et al., 1996). This splicing variant has a molecular weight of 165 kDa. It dimerizes in the intracellular compartment and is constitutively active. The same form was later identified in normal and malignant human colonic tissues (Okino et al., 1999). The expression of Ron was in general related to the degree of differentiation of the tissue. Other splice variants, Ron
160 and Ron
155, were also originally isolated from colorectal cancer cells, and then identified in human primary adenocarcinomas (Wang et al., 2000; Zhou et al., 2003) These variants caused cellular transformation as tested by focus‐formation assay when expressed in NIH3T3 cells. Ron
165 and Ron
155 stable transfectants also gave multiple colonies when grown in soft agar, showing their transforming potential. These activated forms of Ron also showed transforming potential in vivo. Ron
160 and Ron
155 expressed in NIH3T3 cells formed tumors when xenografted on the flank of nude mice.

I. Liver Cancers Ron mRNA and protein is expressed in normal human liver, and its expression has been localized to hepatocytes and Kupffer cells, the resident macrophage population of the liver. Indeed, ablation of Ron receptor activity negates the detrimental effect of bacterial lipopolysaccaride (LPS) in a murine model of acute liver failure, a process induced by the action of LPS on Kupffer cells (Leonis et al., 2002). Ron has also been shown to be overexpressed in 2 of 7 HCC tissue samples (Chen et al., 1997), and approximately one‐half of 45 hepatoblastoma tumor specimens analyzed (Leonis and Waltz, unpublished observations).

24

Purnima K. Wagh et al.

Various cytokines, such as IL‐1, IL‐6, and TNF‐, and growth factors like HGF, increase Ron expression in the HCC cell line HepG2. These cytokines are known to be upregulated in liver disease, including the LPS‐induced murine model of acute liver failure described above, and thus alterations in the production of these cytokines may play an important role in inducing liver tumors, in part by modulating Ron receptor expression (Chen et al., 1997; Leonis et al., 2002).

J. Short Form Ron Another truncated form of the Ron gene was first identified in mice from the locus that confers susceptibility to Friend virus‐induced erythroleukemia (Persons et al., 1999). This short form of the receptor is deleted at the C terminus, but retains the transmembrane and intracellular domains of the protein. In mice, the short form Ron was found to interact with the envelope glycoprotein of Friend virus (Nishigaki et al., 2001). An equivalent truncated short form has also been identified for the human Ron gene (Bardella et al., 2004). This form has been identified in both normal and cancer cells, including ovarian, pancreatic, gastrointestinal, and leukemic cells. Expression of this short form Ron induces characteristics of EMT, including shape change, motility, and anchorage‐independent growth. This short form may be responsible for cell motility (Ghigna et al., 2005). A specific involvement of a virus equivalent to Friend virus in humans has not been identified.

XVI. RON AS A TARGET OF CANCER THERAPY In the last 10 years, progress has been made in developing new drug therapies for cancer by targeting specific overexpressed growth factor receptors that characteristically appear in solid tumors. Most of these growth factor receptors, like the Ron receptor, are activated by and transmit signal cascades by tyrosine phosphorylation. The drug therapies include both monoclonal antibodies and small molecule inhibitors. Some of the recently approved or experimental drug targets include the EGFR (Cohen et al., 2004), human epidermal receptor 2 (HER2/neu) (Barlesi et al., 2005; Rabindran et al., 2004), or drugs that target more than one receptor (Plosker and Keam, 2006; Wong et al., 2006); and both platelet‐derived growth factor (PDGFR) and vascular endothelial growth factor (VEGFR) receptors (Caponigro et al., 2005; Izzedine et al., 2007). One of the first receptor tyrosine kinase that was targeted is the EGFR. Both small molecule inhibitors and anti‐EGFR antibodies have been

Ron Receptor in Cancer Biology

25

approved for clinical use and have been used together, and in combination with chemotherapy or radiation (Huang et al., 2004). The development of EGFR inhibitors is very important to future development of drug therapies against the Ron receptor, since Ron and EGFR have been shown to be closely connected and to interact. Combinatorial therapies have also been shown to be highly effective when targeting receptor tyrosine kinases. Combination therapy may involve small molecules that inhibit several receptor tyrosine kinases (Izzedine et al., 2007). The addition of receptor‐targeted drugs to chemotherapeutic agents may also be an effective strategy. For instance, a combination therapy of tamoxifen, angiostatin, and TIMP‐2 (tissue inhibitor of metalloproteinase‐2) administered to mice with breast tumors, the MMTV‐neu mice, significantly reduced primary tumor growth (90% inhibition, P ¼ 0.01) and metastasis free survival of up to 6 months in the experimental group as compared to 33% in control group, suggesting an overall survival advantage with this combinatorial therapy (Sacco et al., 2003). Biologic drugs that target the Ron receptor are in early stages of development. A humanized monoclonal antibody that blocks the interaction of Ron with HGFL has been developed (O’Toole et al., 2006). This antibody not only inhibits the binding of HGFL to Ron, but also diminished Ron phosphorylation and its downstream signaling. In addition, this antibody also significantly decreased tumor growth of murine xenografts from subcutaneously injected lung, colon, and pancreatic cancer cell lines in nude mice. The mechanism by which Ron promotes tumor growth, and a potential combination therapy, was examined in vitro using a commercially‐available mouse monoclonal blocking antibody (R&D systems). Treatment of BxPC‐3 pancreatic cancer cells with this monoclonal antibody against Ron, followed by 0.1 mol/l of gemcitabine, resulted in 32% increase in apoptosis as compared to gemcitabine alone (Thomas et al., 2007). This interesting result suggests that the function of Ron in tumors may be to increase cell survival, and that blockage of Ron signaling might be used to increase apoptosis induced by classical chemotherapeutic drugs. Additional antibodies have been used to block Ron signaling. Monoclonal antibodies named as ID‐1 and ID‐2 inhibited binding of HGFL to Ron and also diminished HGFL‐induced HT‐29‐D4 human intestinal cell migration, suggesting that these antibodies are efficient in blocking Ron‐mediated oncogenic signaling (Montero‐Julian et al., 1998). The extracellular region of the Ron chain contains a Sema domain, a plexin, semaphorins, and integrins (PSI) domain, and also four IPTs (immunoglobulins like fold shared by plexins and transcription factors) domains (Danilkovitch‐Miagkova, 2003). Both Ron–Sema and Ron–PSI were able to inhibit binding of HGFL to Ron. In addition, they also blocked HGFL‐induced Ron tyrosine phosphorylation and inhibited growth of HCT116 colon cancer cells (Angeloni et al., 2004).

26

Purnima K. Wagh et al.

Chemotherapeutic agents that impair Hsp (heat shock protein) functions are geldanamycins. Hsps are important chaperone proteins that facilitate correct protein folding and assembly. Several receptor tyrosine kinases including Ron are sensitive to these drugs (Germano et al., 2006). These drugs may be useful for combination therapy in concert with Ron‐receptor‐ targeted drugs. Another approach to reducing Ron activity has used gene silencing. Use of a siRNA against Ron expressed in human colorectal carcinoma significantly reduced cancer cell proliferation, motility, and increased apoptotic susceptibility of the cells (Xu et al., 2004). Other types of combination therapy may also be beneficial. Since Ron driven tumors are highly metastatic, a combination of a Ron inhibitor along with angiostatin (drug that prevents tumor angiogenesis) may be efficient in reducing tumor growth and subsequent metastasis, because tumor cells can invade the primary site through newly formed blood vessels. Inhibitors of PI3‐K and the NF‐B pathway in combination with Ron inhibitors can be a useful combinatorial therapy, since these pathways are upregulated in different cancers in response to Ron activation.

XVII. CONCLUSIONS In conclusion, accumulating evidence shows that Ron plays an important role in human cancers. Data summarized here elucidate critical signaling pathways that are downstream of Ron and are important mediators of Ron‐ induced tumorigenesis. In the future, more precise anticancer drugs that block Ron activity may be important additions to cancer therapy.

ACKNOWLEDGEMENTS The authors would like to acknowledge the contribution of Claudia Hinzman for the artwork provided in this manuscript.

REFERENCES Angeloni, D., Danilkovitch‐Miagkova, A., Miagkov, A., Leonard, E. J., and Lerman, M. I. (2004). The soluble sema domain of the RON receptor inhibits macrophage‐stimulating protein‐induced receptor activation. J. Biol. Chem. 279, 3726–3732. Bardella, C., Costa, B., Maggiora, P., Patane, S., Olivero, M., Ranzani, G. N., et al. (2004). Truncated RON tyrosine kinase drives tumor cell progression and abrogates cell‐cell adhesion through E‐cadherin transcriptional repression. Cancer Res. 64, 5154–5161.

Ron Receptor in Cancer Biology

27

Barlesi, F., Tchouhadjian, C., Doddoli, C., Villani, P., Greillier, L., Kleisbauer, J. P., et al. (2005). Gefitinib (ZD1839, Iressa) in non‐small‐cell lung cancer: A review of clinical trials from a daily practice perspective. Fundam. Clin. Pharmacol. 19, 385–393. Bezerra, J. A., Carrick, T. L., Degen, J. L., Witte, D., and Degen, S. J. (1998). Biological effects of targeted inactivation of hepatocyte growth factor‐like protein in mice. J. Clin. Invest. 101, 1175–1183. Bhatt, A. S., Takeuchi, T., Ylstra, B., Ginzinger, D., Albertson, D., Shuman, M. A., et al. (2003). Quantitation of membrane type serine protease 1 (MT‐SP1) in transformed and normal cells. Biol. Chem. 384, 257–266. Bhatt, A. S., Welm, A., Farady, C. J., Vasquez, M., Wilson, K., and Craik, C. S. (2007). Coordinate expression and functional profiling identify an extracellular proteolytic signaling pathway. Proc. Natl. Acad. Sci. USA 104, 5771–5776. Brunelleschi, S., Penengo, L., Lavagno, L., Santoro, C., Colangelo, D., Viano, I., et al. (2001). Macrophage stimulating protein (MSP) evokes superoxide anion production by human macrophages of different origin. Br. J. Pharmacol. 134, 1285–1295. BLAST: Basic local sequence alignment tool at ncbi.nlm.nih.gov [homepage on the Internet]. Bethesda, MD: [updated 2007 Oct 17; cited 2007 Dec 31]. Available from: http://www.ncbi. nlm.nih.gov/ Camp, E. R., Yang, A., Gray, M. J., Fan, F., Hamilton, S. R., Evans, D. B., et al. (2007). Tyrosine kinase receptor RON in human pancreatic cancer: Expression, function, and validation as a target. Cancer 109, 1030–1039. Caponigro, F., Formato, R., Caraglia, M., Normanno, N., and Iaffaioli, R. V. (2005). Monoclonal antibodies targeting epidermal growth factor receptor and vascular endothelial growth factor with a focus on head and neck tumors. Curr. Opin. Oncol. 17, 212–217. Chan, E. L., Peace, B. E., Collins, M. H., Toney‐Earley, K., and Waltz, S. E. (2005). Ron tyrosine kinase receptor regulates papilloma growth and malignant conversion in a murine model of skin carcinogenesis. Oncogene 24, 479–488. Chen, Q., Seol, D. W., Carr, B., and Zarnegar, R. (1997). Co‐expression and regulation of Met and Ron proto‐oncogenes in human hepatocellular carcinoma tissues and cell lines. Hepatology 26, 59–66. Chen, Y. Q., Fisher, J. H., and Wang, M. H. (1998). Activation of the RON receptor tyrosine kinase inhibits inducible nitric oxide synthase (iNOS) expression by murine peritoneal exudate macrophages: Phosphatidylinositol‐3 kinase is required for RON‐mediated inhibition of iNOS expression. J. Immunol. 161, 4950–4959. Chen, Y. Q., Zhou, Y. Q., Fu, L. H., Wang, D., and Wang, M. H. (2002). Multiple pulmonary adenomas in the lung of transgenic mice overexpressing the RON receptor tyrosine kinase. Recepteur d’origine nantais. Carcinogenesis 23, 1811–1819. Cheng, H. L., Liu, H. S., Lin, Y. J., Chen, H. H., Hsu, P. Y., Chang, T. Y., et al. (2005). Co‐expression of RON and MET is a prognostic indicator for patients with transitional‐ cell carcinoma of the bladder. Br. J. Cancer 92, 1906–1914. Chodosh, L. A., Gardner, H. P., Rajan, J. V., Stairs, D. B., Marquis, S. T., and Leder, P. A. (2000). Protein kinase expression during murine mammary development. Dev. Biol. 219, 259–276. Cohen, M. H., Williams, G. A., Sridhara, R., Chen, G., McGuinn, W. D., Jr., Morse, D., et al. (2004). United States Food and Drug Administration Drug Approval summary: Gefitinib (ZD1839; Iressa) tablets. Clin. Cancer Res. 10, 1212–1218. Collesi, C., Santoro, M. M., Gaudino, G., and Comoglio, P. M. (1996). A splicing variant of the RON transcript induces constitutive tyrosine kinase activity and an invasive phenotype. Mol. Cell Biol. 16, 5518–5526. Conrotto, P., Corso, S., Gamberini, S., Comoglio, P. M., and Giordano, S. (2004). Interplay between scatter factor receptors and B plexins controls invasive growth. Oncogene 23, 5131–5137.

28

Purnima K. Wagh et al.

Correll, P. H., Iwama, A., Tondat, S., Mayrhofer, G., Suda, T., and Bernstein, A. (1997). Deregulated inflammatory response in mice lacking the STK/RON receptor tyrosine kinase. Genes Funct. 1, 69–83. Cottage, A., Clark, M., Hawker, K., Umrania, Y., Wheller, D., Bishop, M., et al. (1999). Three receptor genes for plasminogen related growth factors in the genome of the puffer fish Fugu rubripes. FEBS Lett. 443, 370–374. Danilkovitch‐Miagkova, A. (2003). Oncogenic signaling pathways activated by RON receptor tyrosine kinase. Curr. Cancer Drug Targets 3, 31–40. Danilkovitch‐Miagkova, A., Angeloni, D., Skeel, A., Donley, S., Lerman, M., and Leonard, E. J. (2000). Integrin‐mediated RON growth factor receptor phosphorylation requires tyrosine kinase activity of both the receptor and c‐Src. J. Biol. Chem. 275, 14783–14786. Danilkovitch‐Miagkova, A., Miagkov, A., Skeel, A., Nakaigawa, N., Zbar, B., and Leonard, E. J. (2001). Oncogenic mutants of RON and MET receptor tyrosine kinases cause activation of the beta‐catenin pathway. Mol. Cell. Biol. 21, 5857–5868. Danilkovitch, A., Donley, S., Skeel, A., and Leonard, E. J. (2000). Two independent signaling pathways mediate the antiapoptotic action of macrophage‐stimulating protein on epithelial cells. Mol. Cell. Biol. 20, 2218–2227. Danilkovitch, A., Miller, M., and Leonard, E. J. (1999a). Interaction of macrophage‐stimulating protein with its receptor. Residues critical for beta chain binding and evidence for independent alpha chain binding. J. Biol. Chem. 274, 29937–29943. Danilkovitch, A., Skeel, A., and Leonard, E. J. (1999b). Macrophage stimulating protein‐ induced epithelial cell adhesion is mediated by a PI3‐K‐dependent, but FAK‐independent mechanism. Exp. Cell Res. 248, 575–582. De Maria, R., Maggiora, P., Biolatti, B., Prat, M., Comoglio, P. M., Castagnaro, M., et al. (2002). Feline STK gene expression in mammary carcinomas. Oncogene 21, 1785–1790. Degen, S. J., Stuart, L. A., Han, S., and Jamison, C. S. (1991). Characterization of the mouse cDNA and gene coding for a hepatocyte growth factor‐like protein: Expression during development. Biochemistry 30, 9781–9791. Federico, A., Morgillo, F., Tuccillo, C., Ciardiello, F., and Loguercio, C. (2007). Chronic inflammation and oxidative stress in human carcinogenesis. Int. J. Cancer 121, 2381–2386. Follenzi, A., Bakovic, S., Gual, P., Stella, M. C., Longati, P., and Comoglio, P. M. (2000). Cross‐ talk between the proto‐oncogenes Met and Ron. Oncogene 19, 3041–3049. Gaudino, G., Avantaggiato, V., Follenzi, A., Acampora, D., Simeone, A., and Comoglio, P. M. (1995). The proto‐oncogene RON is involved in development of epithelial, bone and neuro‐ endocrine tissues. Oncogene 11, 2627–2637. Gaudino, G., Follenzi, A., Naldini, L., Collesi, C., Santoro, M., Gallo, K. A., et al. (1994). RON is a heterodimeric tyrosine kinase receptor activated by the HGF homologue MSP. Embo J. 13, 3524–3532. Germano, S., Barberis, D., Santoro, M. M., Penengo, L., Citri, A., Yarden, Y., et al. (2006). Geldanamycins trigger a novel Ron degradative pathway, hampering oncogenic signaling. J. Biol. Chem. 281, 21710–21719. Ghigna, C., Giordano, S., Shen, H., Benvenuto, F., Castiglioni, F., Comoglio, P. M., et al. (2005). Cell motility is controlled by SF2/ASF through alternative splicing of the Ron protooncogene. Mol. Cell 20, 881–890. Han, S., Stuart, L. A., and Degen, S. J. (1991). Characterization of the DNF15S2 locus on human chromosome 3: Identification of a gene coding for four kringle domains with homology to hepatocyte growth factor. Biochemistry 30, 9768–9780. Hess, K. A., Waltz, S. E., Chan, E. L., and Degen, S. J. (2003a). Receptor tyrosine kinase Ron is expressed in mouse reproductive tissues during embryo implantation and is important in trophoblast cell function. Biol. Reprod. 68, 1267–1275.

Ron Receptor in Cancer Biology

29

Hess, K. A., Waltz, S. E., Toney‐Earley, K., and Degen, S. J. (2003b). The receptor tyrosine kinase Ron is expressed in the mouse ovary and regulates inducible nitric oxide synthase levels and ovulation. Fertil. Steril. 80(Suppl 2), 747–754. Hsu, P. Y., Liu, H. S., Cheng, H. L., Tzai, T. S., Guo, H. R., Ho, C. L., et al. (2006). Collaboration of RON and epidermal growth factor receptor in human bladder carcinogenesis. J. Urol. 176, 2262–2267. Huang, S., Armstrong, E. A., Benavente, S., Chinnaiyan, P., and Harari, P. M. (2004). Dual‐agent molecular targeting of the epidermal growth factor receptor (EGFR): Combining anti‐EGFR antibody with tyrosine kinase inhibitor. Cancer Res. 64, 5355–5362. Huff, J. L., Jelinek, M. A., Borgman, C. A., Lansing, T. J., and Parsons, J. T. (1993). The protooncogene c‐sea encodes a transmembrane protein‐tyrosine kinase related to the Met/ hepatocyte growth factor/scatter factor receptor. Proc. Natl. Acad. Sci. USA 90, 6140–6144. Huff, J. L., Jelinek, M. A., Jamieson, T. A., and Parsons, J. T. (1996). Expression and maturation of the cellular sea receptor, a member of the hepatocyte growth factor (HGF) receptor family of protein tyrosine kinases. Oncogene 12, 299–307. Hurlimann, J., and van Melle, G. (1991). Prognostic value of serum proteins synthesized by breast carcinoma cells. Am. J. Clin. Pathol. 95, 835–843. Hussain, S. P., Trivers, G. E., Hofseth, L. J., He, P., Shaikh, I., Mechanic, L. E., et al. (2004). Nitric oxide, a mediator of inflammation, suppresses tumorigenesis. Cancer Res. 64, 6849–6853. Hutchinson, J., Jin, J., Cardiff, R. D., Woodgett, J. R., and Muller, W. J. (2001). Activation of Akt (protein kinase B) in mammary epithelium provides a critical cell survival signal required for tumor progression. Mol. Cell. Biol. 21, 2203–2212. Human Protein Atlas v. 3.0 [homepage on the Internet]. HUPO Human Antibody Initiative (HAI), Stockholm, Sweden: [updated 2007 Oct 09; cited 2007 Dec 22]. Available from: http://www.proteinatlas.org/index.php. Iwama, A., Okano, K., Sudo, T., Matsuda, Y., and Suda, T. (1994). Molecular cloning of a novel receptor tyrosine kinase gene, STK, derived from enriched hematopoietic stem cells. Blood 83, 3160–3169. Iwama, A., Yamaguchi, N., and Suda, T. (1996). STK/RON receptor tyrosine kinase mediates both apoptotic and growth signals via the multifunctional docking site conserved among the HGF receptor family. Embo J. 15, 5866–5875. Izzedine, H., Buhaescu, I., Rixe, O., and Deray, G. (2007). Sunitinib malate. Cancer Chemother. Pharmacol. 60, 357–364. Kang, J. Y., Dolled‐Filhart, M., Ocal, I. T., Singh, B., Lin, C. Y., Dickson, R. B., et al. (2003). Tissue microarray analysis of hepatocyte growth factor/Met pathway components reveals a role for Met, matriptase, and hepatocyte growth factor activator inhibitor 1 in the progression of node‐negative breast cancer. Cancer Res. 63, 1101–1105. Kesisis, G., Broxterman, H., and Giaccone, G. (2007). Angiogenesis inhibitors. Drug selectivity and target specificity. Curr. Pharm. Des. 13, 2795–2809. Lapraz, F., Rottinger, E., Duboc, V., Range, R., Duloquin, L., Walton, K., et al. (2006). RTK and TGF‐beta signaling pathways genes in the sea urchin genome. Dev. Biol. 300, 132–152. Lee, W. Y., Chen, H. H., Chow, N. H., Su, W. C., Lin, P. W., and Guo, H. R. (2005). Prognostic significance of co‐expression of RON and MET receptors in node‐negative breast cancer patients. Clin. Cancer Res. 11, 2222–2228. Leonis, M. A., Toney‐Earley, K., Degen, S. J., and Waltz, S. E. (2002). Deletion of the Ron receptor tyrosine kinase domain in mice provides protection from endotoxin‐induced acute liver failure. Hepatology 36, 1053–1060. Li, B. Q., Wang, M. H., Kung, H. F., Ronsin, C., Breathnach, R., Leonard, E. J., et al. (1995). Macrophage‐stimulating protein activates Ras by both activation and translocation of SOS nucleotide exchange factor. Biochem. Biophys. Res. Commun. 216, 110–118.

30

Purnima K. Wagh et al.

Lutz, M. A., and Correll, P. H. (2003). Activation of CR3‐mediated phagocytosis by MSP requires the RON receptor, tyrosine kinase activity, phosphatidylinositol 3‐kinase, and protein kinase C zeta. J Leukoc. Biol. 73, 802–814. Maggiora, P., Lorenzato, A., Fracchioli, S., Costa, B., Castagnaro, M., Arisio, R., et al. (2003). The RON and MET oncogenes are co‐expressed in human ovarian carcinomas and cooperate in activating invasiveness. Exp. Cell Res. 288, 382–389. Maggiora, P., Marchio, S., Stella, M. C., Giai, M., Belfiore, A., De Bortoli, M., et al. (1998). Overexpression of the RON gene in human breast carcinoma. Oncogene 16, 2927–2933. Manning, G., Whyte, D. B., Martinez, R., Hunter, T., and Sudarsanam, S. (2002). The protein kinase complement of the human genome. Science 298, 1912–1934. Maroun, C. R., Naujokas, M. A., Holgado‐Madruga, M., Wong, A. J., and Park, M. (2000). The tyrosine phosphatase SHP‐2 is required for sustained activation of extracellular signal‐ regulated kinase and epithelial morphogenesis downstream from the met receptor tyrosine kinase. Mol. Cell. Biol. 20, 8513–8525. Massot, O., Baskevitch, P. P., Capony, F., Garcia, M., and Rochefort, H. (1985). Estradiol increases the production of alpha 1‐antichymotrypsin in MCF7 and T47D human breast cancer cell lines. Mol. Cell Endocrinol. 42, 207–214. Medico, E., Mongiovi, A. M., Huff, J., Jelinek, M. A., Follenzi, A., Gaudino, G., et al. (1996). The tyrosine kinase receptors Ron and Sea control “scattering” and morphogenesis of liver progenitor cells in vitro. Mol. Biol. Cell 7, 495–504. Mera, A., Suga, M., Ando, M., Suda, T., and Yamaguchi, N. (1999). Induction of cell shape changes through activation of the interleukin‐3 common beta chain receptor by the RON receptor‐type tyrosine kinase. J. Biol. Chem. 274, 15766–15774. Montero‐Julian, F. A., Dauny, I., Flavetta, S., Ronsin, C., Andre, F., Xerri, L., et al. (1998). Characterization of two monoclonal antibodies against the RON tyrosine kinase receptor. Hybridoma 17, 541–551. Muraoka, R. S., Sun, W. Y., Colbert, M. C., Waltz, S. E., Witte, D. P., Degen, J. L., et al. (1999). The Ron/STK receptor tyrosine kinase is essential for peri‐implantation development in the mouse. J. Clin. Invest. 103, 1277–1285. Nishigaki, K., Thompson, D., Hanson, C., Yugawa, T., and Ruscetti, S. (2001). The envelope glycoprotein of friend spleen focus‐forming virus covalently interacts with and constitutively activates a truncated form of the receptor tyrosine kinase Stk. J. Virol. 75, 7893–7903. O’Toole, J. M., Rabenau, K. E., Burns, K., Lu, D., Mangalampalli, V., Balderes, P., et al. (2006). Therapeutic implications of a human neutralizing antibody to the macrophage‐stimulating protein receptor tyrosine kinase (RON), a c‐MET family member. Cancer Res. 66, 9162–9170. Okino, T., Egami, H., Ohmachi, H., Takai, E., Tamori, Y., Nakagawa, K., et al. (1999). Presence of RON receptor tyrosine kinase and its splicing variant in malignant and non‐malignant human colonic mucosa. Int. J. Oncol. 15, 709–714. Okino, T., Onda, M., Matsukura, N., Inada, K. I., Tatematsu, M., Suzuki, S., et al. (2001). Sequential histopathological changes in vivo after suicide gene therapy of gastric cancer induced by N‐methyl‐N’‐nitro‐N‐nitrosoguanidine in rats. Jpn. J. Cancer Res. 92, 673–679. Peace, B. E., Hill, K. J., Degen, S. J., and Waltz, S. E. (2003). Cross‐talk between the receptor tyrosine kinases Ron and epidermal growth factor receptor. Exp. Cell. Res. 289, 317–325. Peace, B. E., Hughes, M. J., Degen, S. J., and Waltz, S. E. (2001). Point mutations and overexpression of Ron induce transformation, tumor formation, and metastasis. Oncogene 20, 6142–6151. Peace, B. E., Toney‐Earley, K., Collins, M. H., and Waltz, S. E. (2005). Ron receptor signaling augments mammary tumor formation and metastasis in a murine model of breast cancer. Cancer Res. 65, 1285–1293.

Ron Receptor in Cancer Biology

31

Persons, D. A., Paulson, R. F., Loyd, M. R., Herley, M. T., Bodner, S. M., Bernstein, A., et al. (1999). Fv2 encodes a truncated form of the Stk receptor tyrosine kinase. Nat. Genet. 23, 159–165. Perwez Hussain, S., and Harris, C. C. (2007). Inflammation and cancer: An ancient link with novel potentials. Int. J. Cancer 121, 2373–2380. Plosker, G. L., and Keam, S. J. (2006). Trastuzumab: A review of its use in the management of HER2‐positive metastatic and early‐stage breast cancer. Drugs 66, 449–475. Quantin, B., Schuhbaur, B., Gesnel, M. C., Doll’e, P., and Breathnach, R. (1995). Restricted expression of the ron gene encoding the macrophage stimulating protein receptor during mouse development. Dev. Dyn. 204, 383–390. Rabindran, S. K., Discafani, C. M., Rosfjord, E. C., Baxter, M., Floyd, M. B., Golas, J., et al. (2004). Antitumor activity of HKI‐272, an orally active, irreversible inhibitor of the HER‐2 tyrosine kinase. Cancer Res. 64, 3958–3965. Rampino, T., Collesi, C., Gregorini, M., Maggio, M., Soccio, G., Guallini, P., et al. (2002). Macrophage‐stimulating protein is produced by tubular cells and activates mesangial cells. J. Am. Soc. Nephrol. 13, 649–657. Rampino, T., Gregorini, M., Soccio, G., Maggio, M., Rosso, R., Malvezzi, P., et al. (2003). The Ron proto‐oncogene product is a phenotypic marker of renal oncocytoma. Am. J. Surg. Pathol. 27, 779–785. Ronsin, C., Muscatelli, F., Mattei, M. G., and Breathnach, R. (1993). A novel putative receptor protein tyrosine kinase of the met family. Oncogene 8, 1195–1202. Roque, R. S., Caldwell, R. B., and Behzadian, M. A. (1992). Cultured Muller cells have high levels of epidermal growth factor receptors. Invest. Ophthalmol. Vis. Sci. 33, 2587–2595. Sacco, M. G., Soldati, S., Indraccolo, S., Cato, E. M., Cattaneo, L., Scanziani, E., et al. (2003). Combined antiestrogen, antiangiogenic and anti‐invasion therapy inhibits primary and metastatic tumor growth in the MMTVneu model of breast cancer. Gene Ther. 10, 1903–1909. Sakamoto, O., Iwama, A., Amitani, R., Takehara, T., Yamaguchi, N., Yamamoto, T., et al. (1997). Role of macrophage‐stimulating protein and its receptor, RON tyrosine kinase, in ciliary motility. J. Clin. Invest. 99, 701–709. Santoro, M. M., Collesi, C., Grisendi, S., Gaudino, G., and Comoglio, P. M. (1996). Constitutive activation of the RON gene promotes invasive growth but not transformation. Mol. Cell. Biol. 16, 7072–7083. Santoro, M. M., Gaudino, G., and Marchisio, P. C. (2003). The MSP receptor regulates alpha6beta4 and alpha3beta1 integrins via 14‐3‐3 proteins in keratinocyte migration. Dev. Cell 5, 257–271. Santoro, M. M., Penengo, L., Minetto, M., Orecchia, S., Cilli, M., and Gaudino, G. (1998). Point mutations in the tyrosine kinase domain release the oncogenic and metastatic potential of the Ron receptor. Oncogene 17, 741–749. Senchenko, V., Liu, J., Braga, E., Mazurenko, N., Loginov, W., Seryogin, Y., et al. (2003). Deletion mapping using quantitative real‐time PCR identifies two distinct 3p21.3 regions affected in most cervical carcinomas. Oncogene 22, 2984–2992. Senchenko, V. N., Liu, J., Loginov, W., Bazov, I., Angeloni, D., Seryogin, Y., et al. (2004). Discovery of frequent homozygous deletions in chromosome 3p21.3 LUCA and AP20 regions in renal, lung and breast carcinomas. Oncogene 23, 5719–5728. Shimamoto, A., Kimura, T., Matsumoto, K., and Nakamura, T. (1993). Hepatocyte growth factor‐like protein is identical to macrophage stimulating protein. FEBS Lett. 333, 61–66. Skeel, A., and Leonard, E. J. (1994). Action and target cell specificity of human macrophage‐ stimulating protein (MSP). J. Immunol. 152, 4618–4623. Skeel, A., and Leonard, E. J. (2001). Alpha 1‐antichymotrypsin is the human plasma inhibitor of macrophage ectoenzymes that cleave pro‐macrophage stimulating protein. J. Biol. Chem. 276, 21932–21937.

32

Purnima K. Wagh et al.

Skeel, A., Yoshimura, T., Showalter, S. D., Tanaka, S., Appella, E., and Leonard, E. J. (1991). Macrophage stimulating protein: Purification, partial amino acid sequence, and cellular activity. J. Exp. Med. 173, 1227–1234. Thomas, R. M., Toney, K., Fenoglio‐Preiser, C., Revelo‐Penafiel, M. P., Hingorani, S. R., Tuveson, D. A., et al. (2007). The RON receptor tyrosine kinase mediates oncogenic phenotypes in pancreatic cancer cells and is increasingly expressed during pancreatic cancer progression. Cancer Res. 67, 6075–6082. UCSC Genome Browser [homepage on the Internet]. Santa Cruz, CA: [updated 2007 Jul 01; cited 2007 Dec 31]. Available from: http://genome.ucsc.edu. van den Akker, E., van Dijk, T., Parren‐van Amelsvoort, M., Grossmann, K. S., Schaeper, U., Toney‐Earley, K., et al. (2004). Tyrosine kinase receptor RON functions downstream of the erythropoietin receptor to induce expansion of erythroid progenitors. Blood 103, 4457–4465. Wahl, R. C., Hsu, R. Y., Huff, J. L., Jelinek, M. A., Chen, K., Courchesne, P., et al. (1999). Chicken macrophage stimulating protein is a ligand of the receptor protein‐tyrosine kinase Sea. J. Biol. Chem. 274, 26361–26368. Waltz, S. E., Eaton, L., Toney‐Earley, K., Hess, K. A., Peace, B. E., Ihlendorf, J. R., et al. (2001). Ron‐mediated cytoplasmic signaling is dispensable for viability but is required to limit inflammatory responses. J. Clin. Invest. 108, 567–576. Waltz, S. E., Gould, F. K., Air, E. L., McDowell, S. A., and Degen, S. J. (1996). Hepatocyte nuclear factor‐4 is responsible for the liver‐specific expression of the gene coding for hepatocyte growth factor‐like protein. J. Biol. Chem. 271, 9024–9032. Waltz, S. E., McDowell, S. A., Muraoka, R. S., Air, E. L., Flick, L. M., Chen, Y. Q., et al. (1997). Functional characterization of domains contained in hepatocyte growth factor‐like protein. J. Biol. Chem. 272, 30526–30537. Waltz, S. E., Toms, C. L., McDowell, S. A., Clay, L. A., Muraoka, R. S., Air, E. L., et al. (1998). Characterization of the mouse Ron/Stk receptor tyrosine kinase gene. Oncogene 16, 27–42. Wang, D., Shen, Q., Chen, Y. Q., and Wang, M. H. (2004). Collaborative activities of macrophage‐stimulating protein and transforming growth factor‐beta1 in induction of epithelial to mesenchymal transition: Roles of the RON receptor tyrosine kinase. Oncogene 23, 1668–1680. Wang, M. H., Cox, G. W., Yoshimura, T., Sheffler, L. A., Skeel, A., and Leonard, E. J. (1994a). Macrophage‐stimulating protein inhibits induction of nitric oxide production by endotoxin‐ or cytokine‐stimulated mouse macrophages. J. Biol. Chem. 269, 14027–14031. Wang, M. H., Dlugosz, A. A., Sun, Y., Suda, T., Skeel, A., and Leonard, E. J. (1996a). Macrophage‐stimulating protein induces proliferation and migration of murine keratinocytes. Exp. Cell. Res. 226, 39–46. Wang, M. H., Kurtz, A. L., and Chen, Y. (2000). Identification of a novel splicing product of the RON receptor tyrosine kinase in human colorectal carcinoma cells. Carcinogenesis 21, 1507–1512. Wang, M. H., Montero‐Julian, F. A., Dauny, I., and Leonard, E. J. (1996b). Requirement of phosphatidylinositol‐3 kinase for epithelial cell migration activated by human macrophage stimulating protein. Oncogene 13, 2167–2175. Wang, M. H., Ronsin, C., Gesnel, M. C., Coupey, L., Skeel, A., Leonard, E. J., et al. (1994b). Identification of the ron gene product as the receptor for the human macrophage stimulating protein. Science 266, 117–119. Wang, M. H., Skeel, A., and Leonard, E. J. (1996c). Proteolytic cleavage and activation of pro‐ macrophage‐stimulating protein by resident peritoneal macrophage membrane proteases. J. Clin. Invest. 97, 720–727.

Ron Receptor in Cancer Biology

33

Wang, M. H., Yoshimura, T., Skeel, A., and Leonard, E. J. (1994c). Proteolytic conversion of single chain precursor macrophage‐stimulating protein to a biologically active heterodimer by contact enzymes of the coagulation cascade. J. Biol. Chem. 269, 3436–3440. Wei, X., Hao, L., Ni, S., Liu, Q., Xu, J., and Correll, P. H. (2005). Altered exon usage in the juxtamembrane domain of mouse and human RON regulates receptor activity and signaling specificity. J. Biol. Chem. 280, 40241–40251. Welm, A. L., Sneddon, J. B., Taylor, C., Nuyten, D. S., van de Vijver, M. J., Hasegawa, B. H., et al. (2007). The macrophage‐stimulating protein pathway promotes metastasis in a mouse model for breast cancer and predicts poor prognosis in humans. Proc. Natl. Acad. Sci. USA 104, 7570–7575. Wetzel, C. C., Degen, S. J., and Waltz, S. E. (2003). Cis‐acting elements in the hepatocyte growth factor‐like protein gene regulate kidney and liver‐specific expression in mice. DNA Cell Biol. 22, 293–301. Willett, C. G., Smith, D. I., Shridhar, V., Wang, M. H., Emanuel, R. L., Patidar, K., et al. (1997). Differential screening of a human chromosome 3 library identifies hepatocyte growth factor‐ like/macrophage‐stimulating protein and its receptor in injured lung. Possible implications for neuroendocrine cell survival. J. Clin. Invest. 99, 2979–2991. Willett, C. G., Wang, M. H., Emanuel, R. L., Graham, S. A., Smith, D. I., Shridhar, V., et al. (1998). Macrophage‐stimulating protein and its receptor in non‐small‐cell lung tumors: Induction of receptor tyrosine phosphorylation and cell migration. Am. J. Respir. Cell Mol. Biol. 18, 489–496. Williams, T. A., Longati, P., Pugliese, L., Gual, P., Bardelli, A., and Michieli, P. (1999). MET (PRC) mutations in the Ron receptor result in upregulation of tyrosine kinase activity and acquisition of oncogenic potential. J. Cell. Physiol. 181, 507–514. Wong, T. W., Lee, F. Y., Yu, C., Luo, F. R., Oppenheimer, S., Zhang, H., et al. (2006). Preclinical antitumor activity of BMS‐599626, a pan‐HER kinase inhibitor that inhibits HER1/HER2 homodimer and heterodimer signaling. Clin. Cancer Res. 12, 6186–6193. Xu, X. M., Wang, D., Shen, Q., Chen, Y. Q., and Wang, M. H. (2004). RNA‐mediated gene silencing of the RON receptor tyrosine kinase alters oncogenic phenotypes of human colorectal carcinoma cells. Oncogene 23, 8464–8474. Yoshimura, T., Yuhki, N., Wang, M. H., Skeel, A., and Leonard, E. J. (1993). Cloning, sequencing, and expression of human macrophage stimulating protein (MSP, MST1) confirms MSP as a member of the family of kringle proteins and locates the MSP gene on chromosome 3. J. Biol. Chem. 268, 15461–15468. Zalcenstein, A., Weisz, L., Stambolsky, P., Bar, J., Rotter, V., and Oren, M. (2006). Repression of the MSP/MST‐1 gene contributes to the antiapoptotic gain of function of mutant p53. Oncogene 25, 359–369. Zhou, Y. Q., He, C., Chen, Y. Q., Wang, D., and Wang, M. H. (2003). Altered expression of the RON receptor tyrosine kinase in primary human colorectal adenocarcinomas: Generation of different splicing RON variants and their oncogenic potential. Oncogene 22, 186–197. Zinser, G. M., Leonis, M. A., Toney, K., Pathrose, P., Thobe, M., Kader, S. A., et al. (2006). Mammary‐specific Ron receptor overexpression induces highly metastatic mammary tumors associated with beta‐catenin activation. Cancer Res. 66, 11967–11974.

TAM Receptor Tyrosine Kinases: Biologic Functions, Signaling, and Potential Therapeutic Targeting in Human Cancer Rachel M. A. Linger,* Amy K. Keating,* H. Shelton Earp,{ and Douglas K. Graham* *Department of Pediatrics, University of Colorado at Denver and Health Sciences Center, Aurora, CO Department of Medicine and Pharmacology, UNC Lineberger Comprehensive Cancer Center, University of North Carolina School of Medicine, Chapel Hill, NC {

I. Introduction II. Molecular Biology of TAM Receptors A. Cloning/Nomenclature B. Expression Patterns C. Ligands and Crystal Structures D. Regulation of Receptor Kinase Activity E. Cellular Functions F. TAM Receptor Signaling Pathways III. Involvement of TAM Receptors in Cancer A. Migration and Invasion B. Angiogenesis C. Cell Survival and Tumor Growth D. TAM Receptors as Prognostic Factors IV. Potential Therapeutic Applications A. Small Molecule Inhibitors B. Soluble Receptors C. Antibodies D. Liabilities of TAM Receptor Antagonism V. Conclusions References Tyro‐3, Axl, and Mer constitute the TAM family of receptor tyrosine kinases (RTKs) characterized by a conserved sequence within the kinase domain and adhesion molecule‐ like extracellular domains. This small family of RTKs regulates an intriguing mix of processes, including cell proliferation/survival, cell adhesion and migration, blood clot stabilization, and regulation of inflammatory cytokine release. Genetic or experimental alteration of TAM receptor function can contribute to a number of disease states, including coagulopathy, autoimmune disease, retinitis pigmentosa, and cancer. In this chapter, we first provide a comprehensive review of the structure, regulation, biologic functions, and downstream signaling pathways of these receptors. In addition, we discuss recent evidence which

Advances in CANCER RESEARCH Copyright 2008, Elsevier Inc. All rights reserved.

35

0065-230X/08 $35.00 DOI: 10.1016/S0065-230X(08)00002-X

Rachel M. A. Linger et al.

36

suggests a role for TAM receptors in oncogenic mechanisms as family members are overexpressed in a spectrum of human cancers and have prognostic significance in some. Possible strategies for targeted inhibition of the TAM family in the treatment of human cancer are described. Further research will be necessary to evaluate the full clinical implications of TAM family expression and activation in cancer. # 2008 Elsevier Inc.

I. INTRODUCTION Receptor tyrosine kinases (RTKs) are transmembrane proteins which transduce signals from the extracellular environment to the cytoplasm and nucleus. In this manner, RTKs regulate normal cellular processes, including survival, growth, differentiation, adhesion, and motility. Abnormal expression or activity of RTKs can render them transforming in cellular and animal models. Furthermore, increased RTK expression or activation has been directly implicated in the pathogenesis of myriad human cancers leading to intense interest in the development and testing of tyrosine kinase inhibitors as cancer therapeutics. The 58 RTKs in the human genome are classified into 20 families by amino acid sequence identity within the kinase domain and structural similarities within their extracellular regions (Robinson et al., 2000). The focus of this review is the TAM family which includes Tyro‐3, Axl, and Mer, three receptors which share the vitamin K‐dependent ligands Gas6 and Protein S. Signaling pathways employed by the TAM family have been recently elucidated and shown to mediate diverse cellular functions, including macrophage clearance of apoptotic cells, platelet aggregation, and natural killer (NK) cell differentiation. This review will highlight the role of these RTKs in normal cellular function as well as the mechanisms employed by the TAM family to promote oncogenesis. In addition, we will discuss possible means of targeted inhibition of the TAM family in the treatment of human cancer.

II. MOLECULAR BIOLOGY OF TAM RECEPTORS Like all RTKs, Tyro‐3, Axl, and Mer contain an extracellular domain, a transmembrane domain, and a conserved intracellular kinase domain. The TAM family is distinguished from other RTKs by a conserved sequence, KW (I/L)A(I/L)ES, within the kinase domain and adhesion molecule‐like domains in the extracellular region (Fig. 1A). More specifically, two immunoglobulin‐ like (Ig) domains and two fibronectin type III (FNIII) domains comprise nearly the entire ectodomain of each family member. These motifs are thought to be important in cell–cell contacts and mimic the structure of

37

TAM Receptor Tyrosine Kinases

A

Tyro-3/Axl/Mer

B

Gas6/protein S Loop region

NH2

Gla domain EGF repeat LG1 domain LG2 domain

NH2

C

COOH

i

Ig-like domain FNIII domain Kinase domain Gas6-LG2 Gas6-LG1 Axl-Ig1 Axl-Ig2

ii KW(I/L)A(I/L)ES

COOH

D

iii ii

iv v

i

Fig. 1 Structure, binding, and activation of TAM receptors are their ligands. (A) Domain organization of Tyro‐3, Axl, and Mer. The conserved sequence within the kinase domain is indicated. (B) Domain structure of the TAM receptor ligands, Gas6 and Protein S. Protein S contains thrombin cleavage sites in the loop region and has not been shown to activate Axl. (C) Axl binds to Gas6 with 2:2 stoichiometry as shown from the side (i) and from the top (ii). No ligand/ligand or receptor/receptor contacts were observed in crystals of the minimal complex containing the two LG domains of Gas6 and the two Ig domains of Axl. (D) Possible means of TAM receptor activation include: (i) ligand‐independent dimerization, (ii) ligand‐dependent dimerization, (iii) heteromeric dimerization of two different TAM receptors, (iv) heterotypic dimerization with a non‐TAM receptor, and (v) trans‐cellular binding of extracellular domains.

38

Rachel M. A. Linger et al.

neural cell adhesion molecule (NCAM), which contains five Ig domains and two FNIII domains (Yamagata et al., 2003). Among the RTKs, Tie (Tie1) and Tek (Tie2) are the only other receptors that contain both Ig and FNIII extracellular domains. The FGF, VEGF, and PDGF growth factor receptor families contain Ig domains while the Ephrin and Insulin families contain FNIII domains. Although the TAM receptors share extracellular motifs with the above RTKs, the MET RTK family (composed of Met and Ron) is most closely related to the TAM family on the basis of amino acid sequence of the kinase domain (Robinson et al., 2000). The MET and TAM receptors activate common signaling molecules resulting in similar functions of the two RTK families (Birchmeier et al., 2003; Hafizi and Dahlback, 2006a). Thus, both the extracellular domain and the intracellular kinase domain are important determinants of the cellular processes regulated by specific RTKs. The TAM receptor genes share similar genomic structure encoding transcripts which range in size from 3 to 5 kb (Graham et al., 1994, 1995; Mark et al., 1994; O’Bryan et al., 1991). Within the TAM family, Tyro‐3 and Axl appear to have the most similar genomic structure sharing the same number, 20, and size of exons (Lewis et al., 1996b; Lu et al., 1999; Schulz et al., 1993). Mer is encoded by 19 exons (Gal et al., 2000). Axl and Tyro‐3 contain alternative splice sites, although the location and outcome of splicing are different. A splice variant of Mer has been suggested but not fully characterized (Graham et al., 1995). Alternative splicing of Tyro‐3 near the 50 end results in three different splice variants containing either exon 2A, exon 2B, or exon 2C (Biesecker et al., 1995; Lewis et al., 1996b; Lu et al., 1999). These exons encode a signal peptide, suggesting that the presence of these splice variants may impact posttranslational processing, localization, and/or function of Tyro‐3. Two Axl variants have been observed resulting from alternative splicing of exon 10 (Neubauer et al., 1994; O’Bryan et al., 1991; Schulz et al., 1993). This exon encodes part of the second FNIII domain just upstream from the transmembrane region (Lu et al., 1999). It remains unknown whether the Tyro‐3 and Axl variants are produced from a single transcript or from multiple promoters. However, analysis of Axl and Mer sequences upstream of their respective translation initiation sites revealed a GC‐rich promoter region lacking traditional TATA or CAAT boxes (Schulz et al., 1993; Wong and Lee, 2002). Further analysis of the Mer promoter suggests that several transcription factors, including Sp1, Sp2, and E2F, may regulate promoter activity (Wong and Lee, 2002). In contrast to the striking similarity of genomic structure between Tyro‐3 and Axl, Axl and Mer have the most similar tyrosine kinase domain amino acid sequence (Graham et al., 1995; Robinson et al., 2000). Overall, the protein sequences of the human TAM receptors share 31–36% identical (52–57% similar) amino acids within the extracellular region. The intracellular

39

TAM Receptor Tyrosine Kinases

domains share 54–59% sequence identity (72–75% similarity) with higher homology in the tyrosine kinase domain (Graham et al., 1995). The full‐ length Tyro‐3, Axl, and Mer proteins contain 890, 894, and 999 amino acids, respectively. Although the predicted protein sizes are 97, 98, and 110 kDa for Tyro‐3, Axl, and Mer, respectively, the actual molecular weights range from 100 to 140 kDa for Axl and Tyro‐3 and 165–205 kDa for Mer due to posttranslational modifications, including glycosylation, phosphorylation, and ubiquitination (Lu et al., 1999; O’Bryan et al., 1991; Sather et al., 2007; Valverde, 2005). Such modifications are possible mediators of tissue‐ and cell type‐specific variations in TAM receptor function (Heiring et al., 2004; Ling et al., 1996) (see Section II.D).

A. Cloning/Nomenclature In addition to sequence and structural similarities, the TAM receptor kinases are unusual in that the entire family was discovered within a span of 3 years. In the early 1990s, each TAM receptor gene was cloned from multiple species by independent groups resulting in confusing nomenclature (Table I). Axl was first detected in 1988 as an unidentified transforming gene in two patients with chronic myelogenous leukemia (CML) (Liu et al., 1988). Three years later, two independent groups reported cloning of the human gene from patients with CML (O’Bryan et al., 1991) and chronic myeloproliferative disorder (Janssen et al., 1991). One group named the gene Axl from the Greek word for uncontrolled, anexelekto (O’Bryan et al., 1991), and the other called the gene UFO indicating the unknown function of its protein product (Janssen et al., 1991). Around the same time, a third group cloned the murine gene and named it Ark (adhesion‐related kinase) Table I Kinase

TAM Receptor Nomenclature Synonyms

Tyro‐3

Brt (m), Dtk (m), Rse, Sky, Tif, Etk‐2 (m), Rek (ch)

Axl

Ark (m), Ufo, Tyro‐7 (r)

Mer

Eyk (ch), MerTK, Nyk, Tyro‐12 (r)

ch, chicken; m, mouse; r, rat.

References Biesecker et al. (1993), Biscardi et al. (1996), Crosier et al. (1994), Dai et al. (1994), Fujimoto and Yamamoto (1994), Lai and Lemke (1991), Lai et al. (1994), Mark et al. (1994), Ohashi et al. (1994), Polvi et al. (1993) Janssen et al. (1991), Lai and Lemke (1991), Liu et al. (1988), O’Bryan et al. (1991), Rescigno et al. (1991) Graham et al. (1994), Graham et al. (1995), Jia et al. (1992), Jia and Hanafusa (1994), Lai and Lemke (1991), Ling and Kung (1995)

40

Rachel M. A. Linger et al.

(Rescigno et al., 1991). In the same year, 13 novel PCR fragments comprising 50–60 amino acids of the conserved tyrosine kinase catalytic domain were isolated from rat brain and named Tyro‐1 to ‐13 (Lai and Lemke, 1991). Interestingly, the authors grouped Tyro‐3, Tyro‐7, and Tyro‐12 into a novel subfamily based on the unique amino acid sequence found in their kinase domains. It would later be discovered that Tyro‐7 is the same gene as Axl/UFO, Tyro‐12 is the same gene as Mer, and Tyro‐3 constituted the third member of the TAM family. In 1992, a second member of the TAM family, v‐ryk, was isolated from the chicken retrovirus RLP30 (Jia et al., 1992). The cellular protooncogene, c‐ryk, was later cloned from embryonic chicken brain and renamed c‐eyk in order to avoid confusion with an unrelated tyrosine kinase also called ryk (Jia and Hanafusa, 1994). Later that same year, our lab cloned the human gene from a B‐lymphoblastoid
gt11 expression library and named it c‐mer because it was found in monocytes as well as in epithelial and reproductive tissues (Graham et al., 1994). We cloned murine c‐mer the following year (Graham et al., 1995). The human gene was cloned by a separate group and called Nyk for NCAM‐related tyrosine kinase (Ling and Kung, 1995). Mer was also named MerTK for Mer tyrosine kinase in a paper which mapped the human gene to chromosome 2q14.1 (Weier et al., 1999). In addition to the earlier mentioned PCR fragment isolated from rat (Lai and Lemke, 1991), fragments of murine Tyro‐3, called Etk‐2 (Biesecker et al., 1993), and human Tyro‐3 (Polvi et al., 1993) were cloned from mouse embryonic stem cells and human teratocarcinoma cell, bone marrow, and melanocyte cDNA libraries, respectively. In 1994, the murine and human genes were cloned by multiple labs. The murine gene was named Dtk (Crosier et al., 1994), Brt (Fujimoto and Yamamoto, 1994), Rse (Mark et al., 1994), and Tyro‐3 (Lai et al., 1994) while the human gene was called Sky (Ohashi et al., 1994), Tif (Dai et al., 1994), or Rse (Mark et al., 1994). Subsequent sequence analysis revealed that Dtk and Brt were alternative splice variants (Lewis et al., 1996b). The chicken ortholog was cloned in 1996 but was given the name Rek because of limited amino acid sequence identity with the mouse and human genes (66% and 68%, respectively) (Biscardi et al., 1996). While many of these names were used initially in the literature, Tyro‐3, Axl, and Mer (or MerTK) have become the most commonly published and will be used exclusively throughout the remainder of this review.

B. Expression Patterns Although expression of TAM receptor mRNA has been observed in embryonic tissues (Crosier et al., 1996; Faust et al., 1992; Graham et al., 1995; Lai and Lemke, 1991), single, double, and even triple knockouts are

TAM Receptor Tyrosine Kinases

41

viable without obvious signs of developmental defect at birth (Lemke and Lu, 2003; Lu and Lemke, 2001; Lu et al., 1999). These data suggest that the TAM RTKs are largely nonessential for embryogenesis. Conversely, TAM adult knockout mice develop diverse phenotypes in a wide range of tissues revealing some of the most prominent cellular functions of TAM receptors (discussed in Section II.E). In adult tissues, Tyro‐3, Axl, and Mer exhibit widespread distribution with overlapping but unique expression profiles. Tyro‐3 is most abundantly expressed in the nervous system, and is also found in ovary, testis, breast, lung, kidney, osteoclasts, and retina as well as a number of hematopoietic cell lines including monocytes/macrophages and platelets (Angelillo‐ Scherrer et al., 2001; Katagiri et al., 2001; Lai et al., 1994; Lu and Lemke, 2001; Mark et al., 1994; Prasad et al., 2006). Axl is expressed ubiquitously (O’Bryan et al., 1991), with notable levels found in the hippocampus and cerebellum (Bellosta et al., 1995) as well as monocytes/macrophages, platelets, endothelial cells, heart, skeletal muscle, liver, kidney, and testis (Angelillo‐Scherrer et al., 2001; Graham et al., 1995; Neubauer et al., 1994). Within the hematopoietic lineages, Mer is expressed in monocytes/ macrophages, dendritic cells, NK cells, NKT cells, megakaryocytes, and platelets (Angelillo‐Scherrer et al., 2001; Behrens et al., 2003; Graham et al., 1994). High levels of Mer expression are also detected in ovary, prostate, testis, lung, retina, and kidney. Lower levels of Mer are found in heart, brain, and skeletal muscle (Graham et al., 1994, 1995; Prasad et al., 2006). Tyro‐3, Axl, and Mer also display ectopic or overexpression in numerous cancers, including myeloid and lymphoblastic leukemias, melanoma, breast, lung, colon, liver, gastric, kidney, ovarian, uterine, and brain cancers (Table II). However, the pattern differs for each family member, e.g. Mer is found in lymphoid leukemia while Axl is not (Graham et al., 1994, 2006; Neubauer et al., 1994).

C. Ligands and Crystal Structures The vitamin K‐dependent protein Gas6 was first identified as a ligand for Axl in 1995 (Stitt et al., 1995; Varnum et al., 1995). The related vitamin K‐dependent anticoagulation factor, Protein S, was described as a ligand for Tyro‐3 (Stitt et al., 1995). Although numerous subsequent studies confirmed that Gas6 binds to and activates all three members of the TAM receptor family, the validity of Protein S as a ligand for any of the TAM receptors became subject to extensive debate (Chen et al., 1997; Godowski et al., 1995; Mark et al., 1996; Nagata et al., 1996; Ohashi et al., 1995). At the heart of the dispute was the issue of physiological relevance as the initial study used human Protein S to activate murine Tyro‐3. Further studies were

Rachel M. A. Linger et al.

42 Table II

TAM Receptor Expression in Human Cancers Cancer

Myeloid leukemias (AML, CML)

Axl

Mer

þ

Lymphoid leukemias (ALL)

Tyro‐3 þ

Ect þ þ

Erythroid leukemia Megakaryocytic leukemia Mantle cell lymphoma Multiple Myeloma Uterine endometrial cancer Gastric cancer Colon cancer Prostate cancer

þ þ þ þ

Thyroid cancer

þ

Lung cancer

þ

Breast cancer

þ

Ovarian cancer

þ

Liver cancer Renal cell carcinoma Astrocytoma/Glioblastoma Pituitary adenoma Melanoma

þ þ þ

Osteosarcoma Rhabdomyosarcoma

þ

þ

þ þ þ

þ

þ þ

þ

þ

References Challier et al. (1996), Crosier et al. (1995), Liu et al. (1988), Neubauer et al. (1994), Rochlitz a et al. (1999) Graham et al. (1994), Graham et al. (2006), Yeoh et al., (2002) Challier et al. (1996) Challier et al. (1996) Ek et al. (2002) De Vos et al. (2001) Sun et al. (2003) b Lin et al. (1999), Wu et al. (2002) Craven et al. (1995) Jacob et al. (1999), Mahajan et al. (2005), Sainaghi et al. (2005), Wu et al. (2004) Ito et al. (1999, 2002), Tanaka et al. (1998) c Shieh et al. (2005), Wimmel et al. (2001) Berclaz et al. (2001), Meric et al. (2002), Zantek et al. (2001), Tavazoie et al., (2008) Macleod et al. (2005), Sun et al. (2004) Tsou et al. (1998) Chung et al. (2003) Vajkoczy et al. (2006) Evans et al. (2001) Gyorffy and Lage (2007), Quong et al. (1994), van Ginkel et al. (2004) Nakano et al. (2003) Khan et al. (1999)

aOverexpression of Axl correlated with poor prognosis. bCoexpression of Axl and Mer correlated inversely with patient prognosis. cOverexpression of Axl correlated with metastatic cancer and poor prognosis.

Over‐ (þ) or ectopic expression (Ect) of TAM receptors has been reported in numerous human cancers.

unable to demonstrate that Protein S could activate a TAM receptor of the same species, possibly due to the need for additional cofactor(s) or modification of the Protein S ligand. However, it was recently determined that purified recombinant murine Protein S does bind to and activate both endogenous murine Mer and heterologously expressed murine Tyro‐3 (Prasad et al., 2006). There is currently no evidence that Protein S activates

TAM Receptor Tyrosine Kinases

43

Axl. A large number of additional studies have investigated the interspecies affinities of Gas6 and Protein S for TAM receptors (reviewed in Hafizi and Dahlback, 2006b). Studies which evaluated the Kd values for human Gas6 binding to each of the three human TAM receptors in vitro suggest that Axl and Tyro‐3 bind Gas6 with roughly equal affinity while Mer affinity for Gas6 is 3–10‐fold lower (Chen et al., 1997; Fisher et al., 2005). Gas6 and Protein S share 43% amino acid sequence identity and have the same domain structure with the exception of thrombin cleavage sites which are present in Protein S but not Gas6 (Dahlback and Villoutreix, 2005; Stenflo et al., 1987) (Fig. 1B). The N‐terminal domain contains glutamic acid residues which must be carboxylated in a vitamin K‐dependent reaction before Gas6 and Protein S are biologically active (Stenhoff et al., 2004). The ‐carboxyglutamic acid (Gla) domain is followed by four EGF‐like repeats and two C‐terminal globular laminin G‐like (LG) domains. The Gla domain mediates Ca2þ‐dependent binding to negatively charged membrane phospholipids exposed on the surface of apoptotic cells. The LG domains form a V‐shaped structure stabilized by a calcium‐binding site and mediate ligand–receptor interactions (Mark et al., 1996; Sasaki et al., 2002). Solution for the crystal structure of a Gas6 fragment containing the two LG domains revealed an unusual ‐helix within LG2 located at the edge of the ‐sandwich fold typical of all LG domains. In addition, five amino acids within LG2 constitute a patch of surface‐exposed hydrophobic residues located near the crook of the “V” created by LG1 and LG2. These residues are also in close proximity to the stabilizing calcium‐binding site. It has not been determined whether the calcium‐binding site contributes to RTK binding. Mutagenesis studies and receptor activation assays suggested that the hydrophobic residues within LG2 comprise at least part of the Axl binding site (Sasaki et al., 2002). However, LG2 alone does not bind to or activate Axl, and a later study by the same group determined that only LG1 of Gas6 binds Axl (Sasaki et al., 2006). The authors suggest that the hydrophobic residues may still affect ligand/receptor binding indirectly. Direct binding between Axl and the LG1 domain of Gas6 was first demonstrated by Fisher et al. (2005). An anti‐Gas6 monoclonal antibody diminished Gas6 binding to Axl and the antibody binding epitope was mapped to residues 403–414 within the J–K loop of LG1. Notably, this region is located near the edge of the LG1 ‐sandwich fold, distant from the hydrophobic patch within LG2. The crystal structure of a Gas6/Axl complex finally revealed that the LG1 domain of Gas6 makes two separate contacts with the IG1 and IG2 domains of Axl (Sasaki et al., 2006). Each contact is characterized by antiparallel alignment of edge ‐strands such that continuous ‐sheets span the molecular junction. Interestingly, no ligand/ligand or receptor/receptor contacts were reported in this minimal complex containing the LG domains of Gas6 and the Ig domains of Axl (Fig. 1C). Additional experiments suggest

44

Rachel M. A. Linger et al.

that ligand‐mediated TAM receptor dimerization occurs via a two‐step mechanism whereby one molecule of Gas6 binds one receptor molecule with high affinity at the LG1/IG1 “major” contact. Lateral diffusion of these 1:1 ligand/receptor complexes results in dimerization of two 1:1 complexes via the LG1/IG2 “minor” contact. Thus, a 2:2 ligand/receptor complex is formed. Further evidence to support two Gas6/Axl binding sites was provided by receptor binding studies, which demonstrated that Gas6 can simultaneously bind Axl–Fc and a neutralizing Gas6 antibody (Fisher et al., 2005). Receptor binding studies of an N‐terminal fragment of Tyro‐3 demonstrated that one site of Gas6/Tyro‐3 receptor interaction is localized to the two Ig domains. Although the crystal structure of the Tyro‐3 fragment and sequence alignment of the three TAM receptors predict the existence of a Gas6‐binding site near the interface of the two Ig domains, no empirical evidence regarding the actual ligand binding site(s) was provided (Heiring et al., 2004). Thus, additional studies are required to determine whether Tyro‐3 and Mer bind Gas6 in the same manner as does Axl. Given that there is no current information describing Protein S as a ligand for Axl, it will be particularly interesting to see how Protein S interacts with Mer and Tyro‐3. Until recently, no structural information was available for the kinase domains of TAM receptors. The crystal structure of the catalytic domain of human Mer has been solved and may provide new insight into numerous aspects of TAM receptor biology, including mechanisms of receptor activation and interaction with downstream signaling molecules (Walker et al., 2007).

D. Regulation of Receptor Kinase Activity 1. CONVENTIONAL ACTIVATION Typical activation of RTKs involves ligand binding to the extracellular domain (Schlessinger, 2000). Ligand binding induces receptor dimerization and subsequent trans‐autophosphorylation of tyrosine residues within the cytoplasmic domain (Fig. 1D). The result of autophosphorylation is twofold: (1) increased catalytic efficiency leads to phosphorylation of other substrates and (2) tyrosine‐phosphorylated RTKs and other proteins constitute docking sites that recruit signaling molecules containing SH2, PTB, or other phosphotyrosine‐binding domains. This allows RTKs and other proteins to form macromolecular signaling complexes. For Mer, three tyrosine residues (Y‐749, Y‐753, and Y‐754 in the human sequence) within the activation loop of the kinase domain have been identified as the primary sites of autophosphorylation (Ling et al., 1996). Interestingly, in vitro kinase

TAM Receptor Tyrosine Kinases

45

assays utilizing peptides with two of the three tyrosines mutated to phenylalanine residues as substrates for WT Mer demonstrated that tyrosine 749 is the preferred site of autophosphorylation. Additional in vitro kinase assays evaluated WT Mer versus mutant Mer phosphorylation of a synthetic peptide containing tyrosines 749, 753, and 754. Single mutations of tyrosines 749, 753, and 754 to phenylalanine reduced Mer kinase activity to 39%, 10%, and <6% of WT Mer, respectively, suggesting that all three residues are required for complete functional activity of the kinase (Ling et al., 1996). These three tyrosines are conserved among the TAM receptors and correspond to residues 681, 685, and 686 in the human sequence on Tyro‐3 and residues 698, 702, and 703 in the human sequence of Axl. Autophosphorylation of Tyro‐3 and Axl have not been reported at these residues. Three alternative tyrosine residues (Y‐779, Y‐821, and Y‐866) within the C‐terminal domain of Axl have been proposed as potential autophosphorylation sites (Braunger et al., 1997). These three sites, and in particular Y‐821, mediate interaction of Axl with a number of signaling molecules including phospholipase C (PLC), phosphatidyl inositol 3 kinase (PI3K), and Grb2 (Braunger et al., 1997; Fridell et al., 1996). All of the interactions identified were dependent on Axl tyrosine kinase activity; however, the studies do not provide clear evidence that tyrosine residues 779, 821, and 866 are indeed sites of autophosphorylation. The residue equivalent to Axl Y821 in Mer (Y‐867/872 in the murine/human sequences) is also a probable site of interaction with multiple signaling molecules. Mutation of tyrosine 867/872 to phenylalanine did not reduce tyrosine phosphorylation of Mer, suggesting that this site does not regulate kinase activity efficiency (Georgescu et al., 1999). Furthermore, Axl mutants lacking tyrosine 821 display normal ligand‐induced tyrosine phosphorylation (Fridell et al., 1996). Alternative to these tyrosines being sites of autophosphorylation, they may be phosphorylated by another kinase recruited by autophosphorylation at different residues. Src‐family non‐RTKs (SFKs) are potential candidates for this activity as they have been shown to interact with both Axl and Tyro‐3 (Braunger et al., 1997; Toshima et al., 1995). The combination of site‐directed mutagenesis and in vitro kinase activity assays allows more definitive assignment of tyrosines 749, 753, and 754 as Mer autophosphorylation sites (Ling et al., 1996). However, it remains possible that these and additional tyrosine or serine/threonine residues are phosphorylated by other kinases. It is also possible that a unique complement of residues becomes phosphorylated in response to specific stimuli within the cellular microenvironment. Expression of TAM receptors in certain cell types may also lead to distinct phosphorylation patterns. Future generation of phospho‐ site‐specific antibodies will greatly aid our ability to address these types of questions.

46

Rachel M. A. Linger et al.

2. ATYPICAL ACTIVATION In some cases, ligand‐independent receptor dimerization and activation can occur (Fig. 1D). For example, overexpression of Axl leads to cell aggregation via homophilic binding of the extracellular domains on neighboring cells (Bellosta et al., 1995). Although cell aggregation correlated with increased tyrosine phosphorylation of Axl, activation of the kinase domain was not required for homophilic binding (Bellosta et al., 1995). Because the specific residue(s) responsible for the observed increase in tyrosine phosphorylation remain unknown, it is possible that phosphorylation occurred at a site unrelated to receptor activation. Studies of Axl and Tyro‐3 overexpression suggest that these receptors are also capable of ligand‐independent dimerization and autophosphorylation (Burchert et al., 1998; Taylor et al., 1995a). Further evidence to support ligand‐independent dimerization was provided by crystal structures of a Tyro‐3 fragment containing the two N‐terminal Ig domains (Heiring et al., 2004). Importantly, a distinction must be made between dimerization of two receptors on the surface of one cell and homophilic binding of receptors on neighboring cells (i versus v in Fig. 1D) as exogenous expression of Tyro‐3 in Sf9 cells (Toshima et al., 1995) and basal expression of Axl in NIH3T3 cells (Bellosta et al., 1995) are not sufficient to induce homophilic binding. Thus, it remains unknown whether this phenomenon occurs with any endogenous TAM receptor. An increasingly common theme in cell signaling literature is cross‐talk between receptor systems. Ligand‐independent heterotypic receptor dimerization of Axl with interleukin‐15 receptor alpha (IL‐15R) has been reported in immortalized and primary fibroblasts (Budagian et al., 2005b) (Fig. 1D). Binding of IL‐15 to IL‐15R, not Axl, leads to Axl‐mediated phosphorylation of IL‐15R as well as Axl phosphorylation, although it is not known whether this is a direct action of the Axl kinase domain. Thus, IL‐15 transactivates the Axl receptor and downstream signaling molecules, including PI3K, Akt, and ERK. Heterotypic dimerization of Axl with cytokine receptors seems to be specific to IL‐15R as Axl does not coprecipitate IL‐2, IL‐4, IL‐7, IL‐9, or IL‐21 receptor subunits, even in the presence of ligand (Budagian et al., 2005b). To date, similar heterotypic receptor interactions have not been reported for Mer or Tyro‐3. Another unexplored possibility is an unusual heteromeric interaction among the three TAM receptors (Fig. 1D). Homo‐ and heterodimerization have been reported for other RTK families such as EGFR family members. Recent studies suggest that Gas6‐mediated phosphorylation/activation of one TAM receptor may require the presence of one or both of the other TAM receptors in some circumstances (Angelillo‐Scherrer et al., 2005; Seitz et al., 2007). Interestingly, Western blotting studies suggest that relatively equal amounts of Axl total protein can be detected in whole cell lysates of

TAM Receptor Tyrosine Kinases

47

platelets from WT and Tyro‐3/ mice. However, flow cytometry experiments demonstrated that surface expression of Axl is significantly reduced in Tyro‐3/ and Mer/ mice (Angelillo‐Scherrer et al., 2005). Taken together, these data suggest that Axl may require the presence of Mer or Tyro‐3 or both for functional surface delivery and stabilization within the plasma membrane.

3. MECHANISMS OF DEACTIVATION Cellular control of RTK signal attenuation is important as aberrant or continued receptor signaling can lead to numerous pathological states, including cancer. Cells have developed numerous methods for inactivation of RTKs, including antagonistic ligands, hetero-oligomerization with kinase inactive mutants, phosphorylation of inhibitory residues by other kinases, dephosphorylation of activating residues by phosphatases, and receptor endocytosis accompanied by ligand dissociation, receptor degradation, or both (Schlessinger, 2000). Only a few of these pathways have been explored as possible mechanisms of TAM receptor regulation. Many tyrosine kinases are negatively regulated by phosphorylation of an inhibitory residue. For example, phosphorylation of tyrosine 527 near the C‐terminus of Src prevents activation of the kinase by promoting intramolecular binding to the SH2 domain, thus rendering the active site inaccessible. Interestingly, it has been postulated that tyrosine 866 on Axl, one of the same residues proposed as a site of autophosphorylation, might constitute an inhibitory phosphorylation site akin to C‐terminal tyrosines found in SFKs and the EGFR (Burchert et al., 1998). However, the same study concluded that the absence or mutation of this residue did not impact the ability of Axl‐retroviruses to transform NIH3T3 cells. A second phosphorylation‐mediated mechanism of receptor downregulation is receptor dephosphorylation by protein tyrosine phosphatases. The putative tyrosine phosphatase C1‐TEN has been shown to bind Axl and overexpression of C1‐TEN correlates with reduced cell survival, proliferation, and migration of 293 cells (Hafizi et al., 2002, 2005b). Although neither enzymatic activity of C1‐TEN nor direct dephosphorylation of Axl have been demonstrated, these results are consistent with C1‐TEN‐mediated Axl inactivation. Soluble forms of Axl and Mer, produced by proteolytic cleavage and release of the ectodomain, can be detected in murine and human plasma (Budagian et al., 2005a; Costa et al., 1996; O’Bryan et al., 1995; Sather et al., 2007). Although a truncated form of Tyro‐3 was found in the cytoplasm when expressed in 293 cells (Taylor et al., 1995a), extracellular soluble Tyro‐3 was not detected in human plasma (Sather et al., 2007). Soluble Mer can also be produced by alternative splicing of the Mer transcript (our unpublished data). Although alternative splicing of Axl (O’Bryan et al.,

48

Rachel M. A. Linger et al.

1991; Schulz et al., 1993) and Tyro‐3 (Biesecker et al., 1995; Lewis et al., 1996b) have been reported, the transcripts generated encode transmembrane proteins. Soluble TAM receptors bind to Gas6 and can act as a ligand sink and inhibit normal cellular functions of the full‐length RTK (Budagian et al., 2005a; Sather et al., 2007). In the same regard, soluble TAM receptors may have therapeutic potential in pathological conditions, such as cancer, where TAM receptor activity is upregulated. This topic will be further explored in Section IV. Evidence supporting endocytosis as a mechanism of TAM receptor downregulation was provided by a report which demonstrated that Gas6 stimulates interaction of Axl with the ubiquitin ligase c‐Cbl and ubiquitination of Axl (Valverde, 2005), a process that has been demonstrated with other RTKs such as the EGFR. Clearly the study of mechanisms which regulate TAM receptor function and turnover is an area that needs further investigation.

E. Cellular Functions Stimulation of TAM receptors can produce diverse cellular functions depending on the ligand–receptor combination as well as the cell type and microenvironment. Initial studies of individual TAM receptors suggested that each kinase performs unique functions in specific cell types. However, as the number of publications investigating two or three TAM receptors in the same system increases, it is becoming evident that the TAM receptors can serve overlapping and possibly cooperative roles. While it is beyond the scope of this review to discuss every cell type which expresses TAM receptors, several cellular functions of TAM receptors are discussed here according to specific cell types.

1. MACROPHAGES/DENDRITIC CELLS TAM‐receptor knockouts develop autoimmune diseases, including rheumatoid arthritis and lupus (Cohen et al., 2002; Lemke and Lu, 2003). Loss of Mer alone confers susceptibility to autoimmunity (Scott et al., 2001). However, the phenotype is more pronounced in double knockouts and most severe in triple knockouts (Lemke and Lu, 2003). These phenotypes likely result from accumulation of apoptotic cells and subsequent tissue necrosis combined with constitutive activation of the immune system. Studies of single, double, and triple mutants suggest that these defects are a result of TAM receptor loss from macrophages/dendritic cells (Lu and Lemke, 2001).

a. Clearance of Apoptotic Cells Cell death via apoptosis is a necessary process for maintenance of normal cell number and health. Clearance of apoptotic cells plays an important role in many biological processes, including tissue development and homeostasis,

TAM Receptor Tyrosine Kinases

49

lymphocyte maturation, and pathological responses such as inflammation. Progressive accumulation of apoptotic cells leads to tissue necrosis and release of intracellular contents into the local environment. Because it is more difficult for immune cells to locate and clear this cellular debris, necrosis leads to inflammation and, in some cases, activation of autoantibody production. Although a number of different types of professional phagocytes can ingest infectious microorganisms and particles, clearance of apoptotic cells is primarily mediated by macrophages and, to a lesser degree, dendritic cells. Because the surface of apoptotic cells and the phagocytes which digest them are both negatively charged, proteins must mediate the processes of cell recognition and engulfment. Specifically, apoptotic cells express phosphatidylserine (PS) on their surface, which has been shown to bind directly to phagocytes via PS receptors or indirectly via binding to one of several soluble proteins, including the TAM receptor ligands Gas6 and Protein S (Anderson et al., 2003; Nakano et al., 1997). Macrophages express all three TAM receptors (Graham et al., 1994; Lu and Lemke, 2001; Neubauer et al., 1994), suggesting a mechanism whereby TAM receptors and their ligands might mediate macrophage recognition of apoptotic cells. Protein S binds to and stimulates phagocytosis of apoptotic cells (Anderson et al., 2003). However, there is currently no empirical evidence which directly correlates Protein S‐mediated phagocytosis with activation of a TAM receptor. Conversely, in vitro studies demonstrated that Gas6 stimulates macrophage uptake of PS liposomes and uptake is blocked by the extracellular domain of Axl (Ishimoto et al., 2000). Similarly, soluble Mer bound to the Fc domain of human immunoglobulin G (Mer–Fc) inhibits macrophage phagocytosis of apoptotic cells presumably by sequestering Mer ligand (Sather et al., 2007). Several lines of evidence suggest that Mer is not required for binding to apoptotic cells but is essential for cell shape changes associated with engulfment of the apoptotic cell (Cohen et al., 2002; Guttridge et al., 2002; Hu et al., 2004; Scott et al., 2001; Todt et al., 2004). The TAM ligands are proposed to mediate phagocytosis of apoptotic cells by bridging an interaction between PS‐expressing cells and TAM receptor‐ expressing macrophages. Thus, the tyrosine kinase domains of TAM receptors, in particular Mer, likely activate downstream signaling events, including integrins such as v5, which leads to cytoskeletal changes necessary for engulfment of apoptotic cells (Wu et al., 2005). It is likely that unique mechanisms mediate clearance of apoptotic cells depending on the type of phagocyte involved and the tissue microenvironment. Accordingly, a recent study by Seitz et al. (2007) suggests that TAM receptor involvement in clearance of apoptotic cells varies according to cell and organ type. They found that Mer, and to a lesser degree Axl and Tyro‐3, mediates macrophage clearance while dendritic cell clearance of apoptotic

50

Rachel M. A. Linger et al.

cells is largely mediated by Axl and Tyro‐3. These findings are consistent with an earlier study which showed that dendritic cells from mice lacking Mer protein exhibit normal phagocytosis of apoptotic cells (Behrens et al., 2003). One of the most intensely studied examples of TAM receptor‐mediated macrophage clearance of apoptotic cells is phagocytosis of photoreceptor outer segment membranes by retinal pigment epithelium (RPE) cells. The role of Mer in RPE phagocytosis was initially elucidated through the study of the Royal College of Surgeons (RCS) rat, a widely studied model of recessively inherited retinal degeneration and animal model for the human disease retinitis pigmentosa. Two groups independently discovered that the genetic basis for RPE dysfunction in the RCS rat was due to a deletion of the second exon of Mer leading to aberrant transcription with a frameshift and translation termination signal 20 codons after the AUG (D’Cruz et al., 2000; Nandrot et al., 2000). In a similar manner, transgenic mice (MerKD) containing a truncated form of the Mer gene lacking the kinase domain exhibit total loss of Mer protein expression and a retinal phenotype similar to that of the RCS rat (Duncan et al., 2003). Subsequent work demonstrated that loss of function mutations in human Mer are present in a small subset of patients with severe and progressive retinitis pigmentosa (Gal et al., 2000; McHenry et al., 2004; Thompson et al., 2002). It would be interesting to determine whether these patients exhibit other similarities to Mer knockout mice, such as predisposition to autoimmune disease. Recent studies have demonstrated that viral gene transfer of Mer into the RCS rat retina results in correction of the RPE phagocytosis defect and preservation of photoreceptors, suggesting the exciting possibility of gene therapy for retinitis pigmentosa patients with Mer mutations (Tschernutter et al., 2005; Vollrath et al., 2001).

b. Cytokine Secretion Cytokines are soluble proteins which mediate communication between cells of the immune system. Cytokines are released in response to extracellular stimuli, including microorganisms and antigens. A number of different cell types, including macrophages, secrete cytokines, and these soluble signaling molecules usually act over short distances. Cytokine levels indicate the status of the immune system and are subject to stringent regulation in order to avoid inappropriate immune responses. When cytokine levels are not held in check, constitutive activation of the immune system can occur resulting in development of autoimmunity. As mentioned previously, TAM receptor knockout mice develop autoimmune diseases likely due, at least in part, to abnormal regulation of cytokine release. MerKD mice are more susceptible to lethal septic shock following lipopolysaccharide (LPS) challenge. LPS binds to surface receptors and activates nuclear factor (NF)‐B, which then initiates production of proinflammatory

TAM Receptor Tyrosine Kinases

51

cytokines, including TNF. Pretreatment with anti‐TNF antibody protects against LPS‐induced death, suggesting that TNF is a key upstream regulator of lethal septic shock. Following LPS treatment, MerKD mice have elevated NFB and TNF levels relative to wild‐type controls (Camenisch et al., 1999). In addition, a recent study demonstrated that Mer activation stimulates the PI3K/Akt pathway which negatively regulates NFB activation, thus decreasing TNF production in dendritic cells (Sen et al., 2007). These data suggest that one of the normal functions of Mer in macrophages and dendritic cells is attenuation of proinflammatory cytokine responses following exposure to bacterial endotoxin. TAM receptors may also mediate other antiinflammatory macrophage responses. For example, interferon (IFN)  has been shown to upregulate expression of Axl and Gas6 in human macrophages resulting in reduced TNF production (Sharif et al., 2006). A role for TAM receptors in a broad spectrum of antiinflammatory responses is further supported by the observation of hyperactive macrophages in TAM receptor triple knockouts which produce higher levels of the proinflammatory cytokine IL‐12 than do wild‐type counterparts (Lu and Lemke, 2001). TAM receptor regulation of the inflammatory response may be disrupted in various pathologies as microarray analysis of Mer kinase activation (via stimulation of FMS–Mer receptor chimera containing the extracellular domain of the M‐CSF receptor and the transmembrane and cytoplasmic domains of Mer) in human prostate cancer cells indicated upregulation of proinflammatory cytokine genes, including IL‐8, IL‐11, and IL‐24 (Wu et al., 2004).

2. NATURAL KILLER CELLS NK cells are lymphocytes which do not express any of the antigen receptors characteristic of T‐ or B‐cells. NKT cells exhibit characteristics similar to both NK and T cells. Expression of Mer in both NK and NKT cells was first reported by Behrens et al. (2003), also demonstrating that the Mer tyrosine kinase domain is critical for normal cytokine release from NKT cells. A later study showed that NK cells also express Axl and Tyro‐3 and all three TAM receptors are required for normal differentiation and functional maturation of NK cells (Caraux et al., 2006).

3. PLATELETS The first evidence to suggest a role for TAM receptors in platelet function came from studies of Gas6 knockout mice. Gas6/ mice were protected against thrombosis and exhibited defective platelet aggregation (Angelillo‐ Scherrer et al., 2001). In the same study, RT‐PCR analysis demonstrated that platelets express Tyro‐3, Axl, and Mer. A follow‐up study used single

52

Rachel M. A. Linger et al.

knockouts of Tyro‐3, Axl, and Mer to demonstrate that all three receptors are required for normal platelet aggregation (Angelillo‐Scherrer et al., 2005). Loss of any one of the TAM receptors or application of soluble Axl protects against fatal thrombosis. These findings are supported by a study from our lab, which demonstrated that soluble Mer (Mer–Fc) reduces platelet aggregation in vitro and protects against collagen/epinephrine‐ induced thrombosis in vivo (Sather et al., 2007). Furthermore, a recent study demonstrated that double and triple TAM receptor knockouts exhibit more severe impairment of platelet function than single knockouts (Wang et al., 2007).

4. VASCULAR SMOOTH MUSCLE CELLS Some of the first studies which evaluated cellular function of TAM receptors were conducted in vascular smooth muscle cells (VSMCs). In these early studies, expression of Axl and Gas6 was increased following vascular injury (Melaragno et al., 1998). In additional experiments, Gas6 stimulation induced migration of Axl‐overexpressing VSMCs (Fridell et al., 1998). Furthermore, Gas6 protects VSMCs from apoptosis induced by serum starvation in an Axl kinase‐dependent manner (Melaragno et al., 2004). These results suggest that TAM receptors may play a role in vascular diseases, such as atherosclerosis, which are characterized by accumulation of VSMCs. Indeed, Gas6 has been shown to stimulate scavenger receptor expression in normal VSMCs (Murao et al., 1999). Scavenger receptors facilitate uptake of low‐density lipoprotein (LDL) which may lead to transformation of the VSMCs into foam cells and development of atherosclerosis. In advanced atherosclerotic lesions, however, TAM receptors may help slow the progression of disease by mediating ingestion of apoptotic macrophages and attenuating the proinflammatory response (Li et al., 2006).

5. OTHER Given their broad expression patterns, it is likely that TAM receptors perform important functions in numerous other cells types. For example, Tyro‐3, Axl, Mer, and their mutual ligand Gas6 are all expressed in the central nervous system but their normal biological activity has not been widely studied in the brain (Lai and Lemke, 1991; Mark et al., 1994; Prieto et al., 1999, 2000). One exception is an established line of evidence demonstrating a role for Axl in survival and migration of gonadotropin‐ releasing hormone (GnRH) neurons (Allen et al., 1999, 2002; Nielsen‐Preiss et al., 2007). Similarly, Gas6 has been shown to reduce cell death of Tyro‐ 3‐expressing hippocampal neurons following serum starvation (Funakoshi et al., 2002). Taken together, these studies suggest that TAM receptors may

TAM Receptor Tyrosine Kinases

53

activate neurotrophic signaling pathways in specific regions of the central nervous system. It also appears that the three TAM receptors act in concert to regulate spermatogenesis, as triple knockouts are infertile because of progressive degeneration of germ cells beginning one week prior to sexual maturity (Lu et al., 1999). The mechanism of germ cell death remains unknown except that it likely involves reduced communication between the TAM receptor‐expressing Sertoli cells which line the seminiferous tubules and the interstitial Leydig cells which express Gas6 and Protein S. TAM receptor regulation of GnRH neurons may also contribute to the infertility of these knockouts as impaired migration of GnRH neurons inhibits sexual maturation.

F. TAM Receptor Signaling Pathways The first hint towards understanding TAM receptor signaling came from studies of FMS–Mer receptor chimera by Ling and Kung in 1995. Around the same time, studies of EGF–Axl receptor chimera were published by an independent group (Fridell et al., 1996). When the studies began, the ligand for TAM receptors was unknown, necessitating the use of receptor chimera composed of, in the latter report, the EGFR receptor ectodomain and transmembrane domain fused to the intracellular kinase domain of Axl. During the course of the studies, Gas6 was discovered as a ligand for Axl and Tyro‐3 and additional work was conducted with the native Axl receptor. Two important findings came out of this seminal work. First, signaling pathway(s) downstream from the Mer and Axl kinase domains were determined to include PI3K, Ras, and ERK. Second, studies of the Axl receptor chimera compared to the native Axl RTK demonstrated that variation in the extracellular domain has a significant impact on downstream signaling. In the 12 years since, an abundance of research has been conducted with the goal of outlining signaling pathways downstream of TAM receptors. Most of these experiments utilize Gas6 to stimulate TAM receptor function but discuss relevance to only one TAM receptor, usually Axl. It should be noted that Gas6 will also activate other TAM receptors endogenously expressed by the cells under investigation. For example, all three TAM receptors are expressed in platelets and are required for normal function of these cells (see Section II.E.3). The downstream signaling pathway whereby TAM receptors mediate platelet aggregation likely involves cross‐talk with the integrin family of receptors as platelets from TAM receptor knockouts exhibit impaired spreading after adhesion to fibrinogen. Indeed, Gas6 stimulates phosphorylation of 3 integrin, PI3K, and Akt in resting platelets from WT, but not TAM receptor knockout mice (Angelillo‐Scherrer et al.,

54

Rachel M. A. Linger et al.

2005). Importantly, the specific contributions of each TAM receptor to this signaling pathway have yet to be clarified. To avoid uncertainty regarding which TAM receptor is responsible for the observed effects, some studies have continued to use the receptor chimera approach, fusing a TAM receptor intracellular kinase domain to an extracellular receptor kinase domain not normally expressed in the cells being studied. Although the use of chimeric receptors allows for determination of signaling pathways downstream from a single TAM receptor kinase, data from such experiments must be interpreted conservatively, given evidence provided by Fridell et al. (1996), suggesting that the extracellular domain impacts downstream signaling. This issue along with inducible expression of TAM receptors in various cell types and unknown variables such as heterodimerization has made characterization of TAM receptor signaling pathways a complex task.

1. MER SIGNALING Much of the evidence delineating Mer signaling pathways is provided by studies of chimeric receptors. This approach originated out of necessity as the ligand for Mer was unknown when many of the studies began. Three well‐known signaling pathways, those downstream from PI3K/Akt, PLC, and MAPK/ERK (Fig. 2), were linked to Mer tyrosine kinase activation by early studies of chimeric Mer receptors expressed in NIH3T3 fibroblasts (Ling and Kung, 1995). In this context, ligand stimulation of Mer kinase led to cellular transformation exemplified by increased proliferation and DNA synthesis. Additional experiments indicated that activation of the MAPK/ERK pathway correlated with activation of Raf and p90RSK kinases as well as phosphorylation of Shc and association of Grb2 with Mer (Ling and Kung, 1995). Later studies identified Gas6 as a ligand for Mer and confirmed that ligand‐dependent activation of endogenous Mer stimulates phosphorylation of ERK1/2 (Chen et al., 1997). Phosphorylation and activation of PLC may occur through direct binding of one of its SH2 domains to endogenous phospho‐Mer (Todt et al., 2004). Similarly, there is evidence to suggest that PI3K may interact with Mer via an SH2 domain (Sen et al., 2007). However, the coimmunopreciptiation experiments of the previous studies do not demonstrate direct binding and it is possible that association of PI3K and PLC with Mer is mediated by interaction of Mer tyrosine 872 with additional adapter proteins such as Grb2 (Georgescu et al., 1999). The ultimate downstream targets of the PI3K/Akt, PLC, and MAPK/ERK pathways may differ according to several factors, including cell type and the tissue microenvironment. In some cells, the PI3K/Akt and MAPK/ERK pathways may act in parallel. In leukemia cells, for example, ligand‐dependent activation of an EGFR–Mer chimeric receptor stimulated phosphorylation

55

TAM Receptor Tyrosine Kinases

b3 Integrin

Mer

p38 MAPK pY749, pY753, pY754 pY872

Platelet aggregation

? IKK

X IkB NFkB

Akt

PI3K

Sequestered in cytoplasm

S6K Shc

PLC

Survival

Grb2 Ras

Src

? Raf1

Hsp90b

? PKC bII

MEK-1

ERK 1/2

Ack1 ? Vav1

Wwox

Wwox

Degradation

Ub

FAK

p90RSK

X

Survival Proinflammatory cytokine production

Rac1/cdc42 p130CAS

c-Fos/ c-Jun

IL-8

Actin reorganization/ cell migration

CrkII

X NFkB

TNFa

Dock180 Rac1

Nucleus

Fig. 2 Mer signaling pathways lead to platelet aggregation, cell survival, regulation of proinflammatory cytokine production, and regulation of the actin cytoskeleton. Molecules in blue have been shown to associate with Mer through either a direct or indirect interaction. Tyrosines 749, 753, and 754 (yellow circles) within the Mer kinase domain are most likely sites of autophosphorylation. Vav1 binds to the region of Mer containing these phosphorylation sites (AA 697–754). It remains undetermined whether the interaction with Mer is direct or mediated by additional adapter proteins. Coimmunoprecipitation experiments suggest that several signaling molecules associate with phosphorylated tyrosine 872 of Mer via their SH2 domains. The kinase(s) which phosphorylate Mer at tyrosine 872 remain unknown. See text for full details. Amino acid designations are from the human sequences. Ub ¼ ubiquitin.

of Akt, ERK1/2, and p38 MAPK resulting in reduced apoptosis without a change in proliferation (Guttridge et al., 2002). The presence of multiple Mer signaling pathways which converge on the same prosurvival outcome gives these cells a strong advantage over noncancerous lymphocytes. In other instances, the PI3K/Akt and MEK/Erk pathways may act in opposition. Similar to the study of leukemia cells discussed earlier, the PI3K/Akt and MAPK/ERK pathways were activated by ligand stimulation of an FMS–Mer chimeric receptor in prostate cancer cells. Additional experiments demonstrated that the Raf and p90RSK kinases act upstream and downstream, respectively, of MAPK/ERK, leading to transcriptional activation of IL‐8 via c‐Fos/c‐Jun binding to the AP‐1 promoter region

56

Rachel M. A. Linger et al.

(Wu et al., 2004). Preincubation with a MEK inhibitor produced the expected result of decreased IL‐8 production. However, preincubation with a PI3K inhibitor increased IL‐8 production. The authors therefore speculated that the PI3K/Akt pathway may attenuate the effects of the MAPK/ERK pathway by phosphorylating and inhibiting Raf. In this case, activation of Mer may both stimulate and reduce proinflammatory cytokine production. It should be noted that other studies have suggested that Mer reduces production of proinflammatory cytokines in noncancerous cells (Camenisch et al., 1999; Sen et al., 2007). Ectopic expression of Mer in prostate cancer cells may therefore result in activation of altered downstream signaling pathways. The tonic strength of normal versus aberrant signaling may therefore determine the oncogenic potential of Mer activation and the ultimate phenotypic fate of the tissue. Yet another possibility exists whereby activation of Mer stimulates a unique complement of signaling events under specific conditions, thus altering the downstream effect(s) of each individual pathway. For example, some studies of Mer signaling suggest that the PI3K/Akt pathway activates NFB while others suggest that NFB is inhibited by the PI3K/Akt pathway. Expression of a constitutively active CD8–Mer chimera in pro‐B cells resulted in transcriptional activation of NFB via PI3K/Akt (Georgescu et al., 1999). Additional signaling pathways activated by CD8–Mer included p38/MAPK and MEK1. These cells were protected from apoptosis and became IL‐3‐independent. Conversely, pretreatment of dendritic cells with apoptotic cells prior to LPS exposure induces Mer‐mediated stimulation of PI3K/Akt. Under these experimental conditions, the p38/MAPK, MEK1, and JNK signaling pathways were active but unaffected by Mer stimulation. The phenotypic result in this case was reduced production of the proinflammatory cytokine, TNF, following exposure to LPS (Sen et al., 2007). Additional experiments in the same study demonstrated that PI3K/Akt negatively regulates NFB by inhibiting IKK activity and thus preventing degradation of IB. As is observed with Axl‐mediated survival (explained later), PI3K/Akt is classically thought to phosphorylate and activate IB kinase (IKK), leading to phosphorylation and degradation of inhibitor of B (IB) releasing NFB from the inhibitory complex. However, different isoforms of IKK have been discovered that are differentially phosphorylated by Akt (Gustin et al., 2004). Thus, there are many factors that define the downstream effects of TAM signaling pathways, including the isoforms of numerous kinases involved and the concomitant activity of additional signaling pathways. Clearly, further investigation is needed to elucidate the myriad signaling pathways activated by Mer kinase. In addition to the well‐known pathways mediated by PI3K/Akt, PLC, and MAPK/ERK, some atypical signaling pathways have been proposed as a link between Mer and the actin cytoskeleton. Yeast two‐hybrid experiments

TAM Receptor Tyrosine Kinases

57

revealed Mer interactions with Grb2, SHC, and Vav1, the latter is a guanine nucleotide‐exchange factor regulating Rac and cdc42 GDP to GTP exchange. Surprisingly, the Mer interaction with Vav1 involved the Vav1 SH2 domain but was constitutive and phosphotyrosine‐independent (Mahajan and Earp, 2003). Subsequent Mer activation induced both Vav1 tyrosine phosphorylation and release of Vav1 from Mer. GDP/GTP exchange on Rac1 and cdc42 followed. These small G proteins are commonly recognized as regulators of the actin cytoskeleton. The initial experiments cited earlier were conducted using an EGFR–Mer chimera expressed in 32D cells. Further study, however, demonstrated that Gas6 stimulation of endogenous Mer in human macrophages also results in Vav1 release and subsequent Rac1 and cdc42 GTP loading (Mahajan and Earp, 2003). These data suggest a potential mechanism whereby activation of Mer may induce spatially focused regulation of the actin cytoskeleton, thus providing a model whereby Mer may mediate changes in cellular morphology necessary for phagocytosis of apoptotic cells bound at specific sites on the macrophage surface. Interestingly, the site of Vav1 interaction was mapped to amino acids 697–754 of Mer. This region contains the three putative Mer autophosphorylation sites (see Section II.D.1). As tyrosine phosphorylation of Vav1 was not sufficient for release from Mer, it is enticing to speculate that another SH2 domain‐ containing protein, perhaps with higher affinity for phosphorylated Mer, is required to release Vav1 and initiate cytoskeletal rearrangement. However, to our knowledge no other proteins have been suggested to interact with Mer in this region. Another study suggests that Mer regulates the actin cytoskeleton via PLC2 and Src. Upon exposure of macrophages to apoptotic cells, PLC2 associates with Mer and becomes phosphorylated (Todt et al., 2004). PLC can activate classical protein kinase Cs (PKCs) such as PKC II, which is required for PS receptor‐dependent phagocytosis in macrophages (Todt et al., 2002). In addition, the Gas6–Mer system may also cooperate with the soluble bridging molecule milk fat globule‐EGF factor 8 protein (MFG‐E8) and its receptor v5 integrin to stimulate the lamellipodia formation necessary for phagocytic engulfment of apoptotic cells. Studies utilizing constitutively active Mer chimera and kinase dead mutant Mer demonstrated that Mer stimulates Src‐mediated phosphorylation of FAK and p130CAS/CrkII/Dock180 complex activation of Rac1 in an v5 integrin‐dependent manner (Wu et al., 2005). This pathway may also involve PLC2 as FAK association with v5 integrin is dependent on PKC (Lewis et al., 1996a). Mer activation has also been linked to cell survival via atypical signaling pathways. Gas6 stimulation of a human prostate adenocarcinoma cell line resulted in phosphorylation of a 120‐kDa protein that was identified as Cdc42‐associated kinase (Ack1) by mass spectrometry (Mahajan et al.,

58

Rachel M. A. Linger et al.

2005). Constitutive association of Mer and Ack1 could be detected by coimmunoprecipitation of the endogenous proteins. Experiments with constitutively active and kinase dead mutant constructs of Ack1 demonstrated that Ack1 is not a direct Mer substrate, but that Ack1 autophosphorylation (and presumably activation) is facilitated by ligand activation of cell surface Mer. Continued Ack1 kinase activity required the chaperone activity of heat shock protein 90 (Hsp90). Additional mass spectrometry sequencing of constitutively active Ack immunoprecipitates identified the tumor suppressor Wwox as an Ack1‐interacting protein. Further investigation suggests that Ack1 induces phosphorylation, ubiquitination, and degradation of Wwox. Downregulation of this proapoptotic tumor suppressor may be one mechanism by which Ack1 and perhaps Mer relay survival signals in cancer cells. Since the physiologic function of the high levels of Mer expressed in normal prostate is not known, it is difficult to assess the normal role of the Mer–Ack axis.

2. AXL SIGNALING Gas6/Axl signaling promotes the growth and survival of numerous cell types, including cardiac fibroblasts (Stenhoff et al., 2004). These effects are likely mediated by Gas6/Axl‐induced activation of the MAPK/ERK and PI3K signaling pathways (Fig. 3). Early studies utilized a chimeric EGFR/ Axl receptor expressed in a leukemic cell line. These experiments demonstrated that ligand stimulation of the chimeric receptor leads to cell proliferation via activation of Grb2, Ras, Raf1, MEK‐1, and ERK1/2 (Fridell et al., 1996). Interestingly, Grb2 can be activated either by direct binding to tyrosine 821 on Axl or by association with Shc, which is phosphorylated upon ligand stimulation but does not associate with Axl. Later studies confirmed that the Ras/ERK pathway is essential for Gas6‐induced mitogenesis of NIH3T3 cells (Goruppi et al., 1999). Importantly, NIH3T3 cells also express Tyro‐3 and therefore this mitogenic pathway may be activated by multiple TAM receptors. Although more than one study has suggested that weak or partial activation of the Ras/ERK pathway contributes to Axl‐ mediated survival (Bellosta et al., 1997; Fridell et al., 1996), more recent data indicate that Ras is dispensable for survival resulting from Gas6 stimulation of native TAM receptors in NIH3T3 cells (Goruppi et al., 1999). However, the MAPK/ERK pathway may be important for Gas6/ TAM receptor‐mediated survival in certain cell types, including GnRH neurons (Allen et al., 1999). While the MAPK/ERK pathway typically results in Axl‐mediated proliferation, Axl binding to and activation of PI3K has been linked to multiple downstream pathways converging on increased cell survival. One pathway involves classical PI3K stimulation of Akt and S6K (Goruppi et al., 1997).

59

TAM Receptor Tyrosine Kinases

b3 Integrin

Platelet aggregation

Axl C1-TEN

IKK NFkB pY779 pY821 pY866

Bad

Ras

Rac/Rho

Grb2

Caspase-3

Rac Pak

Ca2+ signaling

IkB NFkB

Degradation

S6K

Shc

PLC

Akt

PI3K

MAPKAP kinase 2 Raf1

?

JNK

p38 MAPK

Ras

Survival

HSP25

Actin reorganization/ cell migration

RanBPM MEK-1 Src/Lck ?

Nck2

SOCS-1 ? Axl expression induced by IFNa

ERK 1/2

?

Proliferation ?

NFkB

Proinflammatory cytokine production Twist expression

Bcl-xL or Bcl-2

X

NFkB Twist

TNFa

Nucleus

Fig. 3 Axl signaling pathways lead to platelet aggregation, cell survival, proliferation, regulation of proinflammatory cytokine production, and regulation of the actin cytoskeleton. Molecules in blue have been shown to associate with Axl through either a direct or indirect interaction. Tyrosines 779, 821, and 866 of Axl are phosphorylated (yellow circles) and mediate interactions with a number of signaling molecules. It remains unknown whether these residues are sites of autophosphorylation or whether they are substrates for another protein tyrosine kinase. See text for full details. Amino acid designations are from the human sequences.

Gas6 also stimulates phosphorylation of Bad, a target of Akt commonly associated with prosurvival signaling (Goruppi et al., 1999; Lee et al., 2002). Other survival pathways downstream of Gas6–Axl signaling via PI3K/Akt include phosphorylation of NFB, increased expression of antiapoptotic proteins such as Bcl‐2 and Bcl‐xL, and inhibition of proapoptotic proteins such as caspase 3 (Demarchi et al., 2001; Hasanbasic et al., 2004). Transcriptional activation of Bcl‐xL occurs via the cannonical NFB activation pathway whereby Akt phosphorylates and activates IKK, leading to phosphorylation and degradation of IB releasing NFB from the inhibitory complex (Demarchi et al., 2001). NFB then enters the nucleus where it binds to the promoter region of Bcl‐xL. Interestingly, this mechanism of NFB regulation by Axl/PI3K/Akt differs from Mer activation of PI3K/ Akt, which has been shown to inhibit IKK resulting in downregulation of NFB‐dependent transcription of TNF (explained later). Another

60

Rachel M. A. Linger et al.

Gas6/Axl‐induced survival pathway may involve PI3K activation of the small GTPases Rac and Rho as well as the downstream kinases Pak and JNK (Goruppi et al., 1999). Many of these experiments were conducted in NIH3T3 cells which express both Axl and Tyro‐3. However, Gas6 stimulation of fibroblasts from Axl/ mice did not result in increased cell survival relative to Axl WT cells (Bellosta et al., 1997). These results suggest that Axl is required for Gas6‐mediated survival in some cell types. Additional studies suggest that Gas6/Axl receptor signaling activates PI3K‐dependent survival pathways in numerous other cells types, including lens epithelial cells, vascular smooth muscle cells, GnRH neurons, and oligodendrocytes (Allen et al., 1999; Melaragno et al., 2004; Shankar et al., 2003; Valverde et al., 2004). Further study in oligodendrocytes from WT, Axl/, and Tyro‐3/ mice suggest that Axl is required for Gas6–PI3K–Akt‐mediated survival (Shankar et al., 2006). In addition to the prototypic growth and survival pathways described earlier, Gas6/Axl signaling has also been linked to additional cellular functions such as neuronal cell migration and cytokine production. Studies of GnRH neurons suggest that Axl directs migration of these cells from the olfactory placode to the forebrain via a signaling pathway involving PI3K, Ras, Rac, p38 MAPK, MAPKAP kinase 2, and HSP25, which results in actin reorganization (Allen et al., 2002; Nielsen‐Preiss et al., 2007). Interestingly, Axl is not expressed in postmigratory GnRH neurons (Allen et al., 1999). With respect to cytokine production, IFN‐induced upregulation of Axl and Gas6 expression in human macrophages leads to increased Twist expression and reduced TNF production (Sharif et al., 2006). Twist is a basic helix loop helix protein that likely inhibits NFB‐mediated transcription of TNF by binding to the E box region within the TNF promoter. Given that macrophages also express Tyro‐3 and Mer, these receptors may also regulate Twist expression. Consistent with this idea, Protein S (which has not been shown to activate Axl) stimulated Twist expression in the presence of IFN. A number of studies have suggested a physical association between Axl and various signaling molecules. For example, coimmunoprecipitation experiments demonstrated association of EGFR/Axl chimera and several coexpressed GST fusion proteins in 293 cells. In the same study, Far–Western analysis of mutant EGFR/Axl receptors as well as competition assays with phosphorylated Axl peptides revealed that tyrosine 821 of Axl mediates binding to PLC, p85 and p85 subunits of PI3K, Grb2, Src, and Lck (Braunger et al., 1997). Axl tyrosine 866 also contributes to PLC binding while tyrosine 779 may constitute a nonessential, low affinity site of interaction with p85 and p85. The interaction of Src and Lck likely involves additional contacts in vivo as the Axl mutant receptor Y821F effectively coimmunoprecipitated both SFKs from 293 cells. Yeast two‐hybrid experiments confirmed the interaction of Axl with PI3K and Grb2 while identifying four novel proteins which

TAM Receptor Tyrosine Kinases

61

potentially interact with Axl: suppressor of cytokine signaling (SOCS)‐1, Nck2, Ran‐binding protein in microtubule organizing center (RanBPM), and C1‐TEN (Hafizi et al., 2002). In many cases, such as the Grb2 and PI3K pathways described earlier, the signaling events downstream of these interactions have been subject to intense investigation. Conversely, Src‐family kinase activity has been associated with Gas6‐mediated mechanisms of proliferation and survival as well as neuronal migration, but the upstream and downstream components of these signaling pathway(s) have not been determined (Goruppi et al., 1997; Nielsen‐Preiss et al., 2007). Many of the other Axl‐interacting proteins have not been studied beyond their association with activated receptor. Nonetheless, there are reasonable hypotheses as to how some of these proteins may be involved in TAM receptor signaling. C1‐TEN, for example, contains a tyrosine phosphatase motif. Thus Axl and other TAM receptors may be found in complex with both tyrosine kinases (SFKs) and phosphatases. Overexpression of C1‐TEN in 293 cells has been shown to inhibit Akt signaling resulting in reduced cell survival, migration, and proliferation (Hafizi et al., 2005b). These data are consistent with Axl inactivation mediated by the putative phosphatase C1‐TEN. Furthermore, Axl signaling has been associated with attenuation of cytokine production (see Section II.E.1.b), including attenuation of proinflammatory cytokine production following exposure to LPS, a potential role for Axl SOCS‐1 signaling as SOCS‐1 is implicated in negative regulation of LPS‐induced signaling (Kinjyo et al., 2002; Nakagawa et al., 2002).

3. TYRO‐3 SIGNALING The Tyro‐3 receptor is the least studied of the TAM receptors and the signaling pathways downstream of Tyro‐3 activation are poorly understood. Nonetheless, a handful of studies provided clues as to the molecules which mediate Tyro‐3 signaling (Fig. 4). Coimmunoprecipitation of Tyro‐3 transiently expressed in COS cells revealed a potential interaction with a phosphorylated SFK (Toshima et al., 1995). Because of cross‐reactivity of the antibody used, it remains unknown which SFK(s) (Src, Yes, and/or Fyn) interact with Tyro‐3. Importantly, all three of these SFKs are highly expressed in tissues of the central nervous system where they are likely to be found colocalized with Tyro‐3. Yeast two‐hybrid studies identified a number of proteins that potentially interact with Tyro‐3, including RanBPM, protein phosphatase 1 (PP1), and the p85 ‐subunit of PI3K (Hafizi et al., 2005a; Lan et al., 2000). Sequencing of the DNAs encoding the interacting proteins demonstrated that PI3K binds Tyro‐3 via one of its SH2 domains and the interaction was confirmed in vitro and in vivo by GST pull‐down assay and coimmunoprecipitation, respectively. Furthermore, ligand stimulation of an

Rachel M. A. Linger et al.

62 b3 Integrin

Tyro-3

PI3K

?

ERK 1/2 ?

RanBPM

Akt

Platelet aggregation

Cell transformation

Osteoclastic bone resorption

SFK PP1 ? ? ?

Fig. 4 Tyro‐3 signaling pathways mediate platelet aggregation, cell transformation, and osteoclastic bone resorption. Molecules in blue have been shown to associate with Tyro‐3 through either a direct or indirect interaction. Phosphorylation of Tyro‐3 at specific residues remains uncharacterized.

EGFR/Tyro‐3 chimera induces phosphorylation of Tyro‐3, PI3K, and Akt resulting in a transformed phenotype. A MAPK signaling pathway has also been linked to Tyro‐3 activation as phosphorylation of ERK1/2 was increased by Gas6 stimulation of NIH3T3 cells which express endogenous Tyro‐3 (Chen et al., 1997). Phosphorylation of ERK1/2 was also upregulated by Gas6 stimulation of endogenous Tyro‐3 in mouse osteoclasts, resulting in bone resorption (Katagiri et al., 2001). Importantly, phosphorylation of Tyro‐3 at specific residues has not been described. Clearly, further investigation is necessary to elucidate the signaling pathways downstream of Tyro‐3 activation.

III. INVOLVEMENT OF TAM RECEPTORS IN CANCER There are many ways that protooncogenes such as TAM receptors can be activated, including gene amplification and mutations, proteolytic cleavage, and altered protein expression. These modifications have all been described for TAM receptors and each may result in generation of a constitutively active enzyme and/or over‐ or ectopically expressed proteins that are not subject to normal cellular regulation. Most of the TAM receptor gene mutations reported involve Mer and retinal degeneration (D’Cruz et al., 2000; Gal et al., 2000; McHenry et al., 2004; Tada et al., 2006; Tschernutter et al., 2006). To date, no activating TAM receptor human mutations have been associated with development of cancer. Although random

TAM Receptor Tyrosine Kinases

63

retrovirus‐induced mutations of Axl correlated with increased transformation of NIH3T3 cells, gene sequencing revealed that the mutations were silent and overexpression of Axl was determined to be a major contributor to cellular transformation (Burchert et al., 1998). This idea is consistent with evidence discussed later, which suggests that the oncogenic potential of TAM receptors is related to aberrant regulation of the same signaling pathways and cellular processes in which these receptors normally play a role. The oncogenic potential of the TAM receptor kinases was immediately evident as each family member was originally cloned from cancer cells and early studies demonstrated that these RTKs exhibit the ability to transform NIH3T3 fibroblasts and BaF3 lymphocytes in vitro (Georgescu et al., 1999; Lan et al., 2000; Ling and Kung, 1995). Some of the most convincing early evidence, however, comes from studies of the avian ortholog of Mer, Eyk (Jia and Hanafusa, 1994). A truncated version of Eyk containing only the tyrosine kinase domain mediates the transforming ability of the virus RLP30, which causes fibrosarcomas, endotheliomas, and visceral lymphomatosis in chickens (Jia et al., 1992). Numerous studies have since used a variety of techniques, including immunohistochemistry, Western blotting, microarrays, RT‐PCR, and flow cytometry to demonstrate that TAM receptors are ectopically or overexpressed in a wide array of human cancers. Tyro‐3 expression has been associated with acute myeloid leukemia (AML) and multiple myeloma. Altered Axl expression has been reported in lung cancer, uterine cancer, breast cancer, ovarian cancer, gastric cancer, colon cancer, prostate cancer, thyroid cancer, liver cancer, renal cell carcinoma, AML, CML, erythroid leukemia, megakaryocytic leukemia, melanoma, osteosarcoma, and glioblastoma. Aberrant expression of Mer has been linked to B‐ and T‐cell acute lymphoblastic leukemias, mantle cell lymphoma, melanoma, rhabdomyosarcoma, pituitary adenoma, gastric cancer, and prostate cancer (Table II). Hanahan and Weinberg (2000) have proposed six primary cellular functions as “Hallmarks of Cancer” which normal cells acquire during oncogenesis: self‐sufficiency in growth signals, insensitivity to antigrowth signals, limitless replicative potential, tissue invasion and metastasis, sustained angiogenesis, and evasion of apoptosis. In this section we will discuss evidence which suggests that TAM receptors contribute to at least three of these six fundamental mechanisms of malignancy.

A. Migration and Invasion As discussed in Section II, TAM receptor signaling pathways have been linked to regulation of the actin cytoskeleton. The resultant changes in cellular morphology are likely to contribute to TAM receptor regulation of normal

64

Rachel M. A. Linger et al.

cellular processes such as platelet spreading and phagocytosis (Angelillo‐ Scherrer et al., 2005; Mahajan and Earp, 2003). In glioblastoma cells which express elevated levels of Axl, transfection of a dominant negative Axl (Axl‐DN) lacking the kinase domain results in reduced motility, altered morphology characterized by loss of filopodia, and loss of cell‐to‐cell interactions (Vajkoczy et al., 2006). Conversely, stimulation of an ectopically expressed EGF/Mer chimera in a murine leukemic cell line induces rapid (8–24 h) changes in cell morphology, including cell flattening, extension of dendrite‐ like processes, and adherence (Guttridge et al., 2002). Thus, ectopic expression or overexpression of TAM receptors and resultant downstream changes in cellular morphology may contribute to mechanisms of oncogenesis. In addition to kinase‐mediated links to the actin cytoskeleton, the extracellular domains of TAM receptors contain adhesion molecule‐like motifs suggesting that they may be involved in cell–cell contacts. Overexpression of murine Axl in insect cells results in cell aggregation (Bellosta et al., 1995). In vitro experiments with fluorescently labeled Axl‐expressing cells and unlabeled Axl‐negative cells did not result in mixed aggregates suggesting that the observed cell aggregation is mediated by homophilic binding of Axl receptors on neighboring cells. Additional experiments demonstrated that Axl‐expressing cells bind to immobilized Axl extracellular domain (ECD) providing further evidence that Axl ECDs are capable of homophilic binding. In these experiments, Axl–Axl interaction was independent of calcium and Axl kinase activity. Interestingly, overexpression of human Axl in a mammalian leukemia cell line (32D) is not sufficient to induce cell aggregation (McCloskey et al., 1997). Rather, addition of Gas6 is required to induce cell aggregation and this effect is blocked by excess Axl ECD peptide. Experiments with truncated Gas6 demonstrated that either the Gla domain and/or the EGF motifs of Gas6 bind to 32D cells leaving the LG domains of Gas6 available for interaction with Axl. These results suggest that cell aggregation is the result of a Gas6‐mediated interaction between Axl and neighboring cells. It is unknown whether phospholipids or another integral membrane protein mediates the interaction of Gas6 with 32D cells. These studies suggest that TAM receptors can mediate cell–cell contacts via receptor– receptor or receptor–ligand interaction. The determinants of adhesion likely involve cell type and tissue microenvironment. Adhesion molecules are important not only for cell–cell contacts, but also for interaction of cells with their extracellular environment. Axl expression correlates with adherence of lung cancer cell lines (Wimmel et al., 2001). However, forced expression of Axl did not induce an adherent phenotype in small cell lung cancer (SCLC) cells which normally grow in suspension. One interpretation of these data is that Axl expression occurs as a consequence of cellular adhesion rather than playing a causative role in the adherent phenotype. Alternatively, additional factors may be required in order for Axl to

TAM Receptor Tyrosine Kinases

65

mediate cell adhesion. Axl ligands are the most obvious candidates, although the presence of Gas6 did not induce adherence in the previous experiments and Protein S has not been shown to bind Axl. Another candidate is soluble Axl (sAxl), which is present at high levels in vivo. Incubation of Axl‐expressing cells with immobilized Axl–Fc induces phosphorylation of Axl and PI3K suggesting that Axl–Fc/Axl interaction activates the kinase domain of the full‐length receptor (Budagian et al., 2005a). Thus, sAxl may mediate interactions between full‐length Axl and the extracellular matrix. Given that regulation of cell adhesion and morphology are precursors to more complex cellular processes such as cellular migration, the aforementioned studies suggest that both the extracellular and kinase domains of TAM receptors may contribute to oncogenic mechanisms such as cellular migration and tissue invasion. Indeed, Axl has been shown to be involved in normal migration of GnRH neurons from the olfactory placode to the hypothalamus (Allen et al., 2002). Interestingly, Axl expression in GnRH neurons is present during migration but then disappears once the cells reach their destination. Aberrant TAM receptor expression could therefore lead to new migratory function and increase invasiveness of cancer cells. In a dorsal skinfold xenograft model, human glioblastoma cells transfected with WT Axl showed significantly greater tumor growth and tissue invasion than cells transfected with truncated Axl lacking the kinase domain (Vajkoczy et al., 2006). These results suggest that Axl kinase contributes to tumorigenesis in vivo. In vitro studies confirmed that reduced migration and invasion resulted from loss of Axl kinase domain and was not an artifact of reduced tumor cell load (Vajkoczy et al., 2006). Soluble Axl bound to the extracellular matrix may constitute a chemoattractant for Axl‐mediated migration as scratch tests revealed that immobilized Axl–Fc promotes migration of primary fibroblasts prepared from Axl WT mice (Budagian et al., 2005a). Primary fibroblasts prepared from Axl/ mice exhibited reduced migration. Further study is necessary to determine whether increased production of sAxl correlates with tumor invasiveness.

B. Angiogenesis Formation of new blood vessels is a normal process important during development as well as wound healing. In addition, angiogenesis promotes tumor growth and malignant transformation. Proliferation and migration of vascular smooth muscle cells (VSMCs) are key events required during normal angiogenesis. VSMCs express Gas6 and exogenous application of purified or recombinant Gas6 promotes proliferation and migration of VSMCs (Fridell et al., 1998; Nakano et al., 1995). Gas6‐induced migration of VSMCs was blocked by inclusion of recombinant Axl–ECD. Furthermore,

66

Rachel M. A. Linger et al.

overexpression of Axl increased migration 2–5‐fold whereas expression of a kinase dead mutant reduced migration 50% relative to parental VSMCs (Fridell et al., 1998). These results demonstrate that migration of VSMCs correlates with the level of Axl kinase activity. It has also been suggested that Axl plays a role in flow‐induced vascular remodeling by regulating VSMC apoptosis (Korshunov et al., 2006). The role of TAM receptors in angiogenesis is not restricted to VSMCs as a genetic screen identified Axl as a regulator of human umbilical vein endothelial cell (HUVECs) migration (Holland et al., 2005). Genetic silencing of Axl or Gas6 significantly reduced migration of HUVECs, whereas overexpression of Axl WT protein increased HUVEC growth and tube formation. Overexpression of a kinase dead Axl mutant, however, reduced HUVEC growth but had no effect on tube formation (Holland et al., 2005). Consistent with these results, glioblastoma xenografts containing a Axl‐DN construct exhibited blood vessel density and diameter similar to WT Axl xenografts (Vajkoczy et al., 2006). These results suggest that Axl kinase activity is important for regulation of endothelial cell growth, whereas tube formation is likely regulated by Axl in a kinase‐independent manner. Finally, stable shRNA knockdown of Axl reduces blood vessel formation and functional circulation in a mouse model of angiogenesis supporting a role for Axl in angiogenic processes in vivo (Holland et al., 2005). Although Tyro‐3 and Mer are expressed in endothelial cells (Chan et al., 2000; Sather et al., 2007), their involvement in angiogenesis has not been investigated.

C. Cell Survival and Tumor Growth Several lines of evidence (see Section II.F) suggest that TAM receptors activate prosurvival signaling pathways in both normal and cancerous cells. In some cases, TAM receptor signaling pathways prevent apoptosis without stimulating proliferation (Guttridge et al., 2002). On the other hand, TAM receptors have also been shown to increase proliferation without inhibiting apoptosis (Sainaghi et al., 2005). A third situation exists, whereby TAM receptors promote both survival and proliferation (van Ginkel et al., 2004). Each mechanism provides a means by which TAM receptors may contribute to tumor growth. As discussed in Section II.F.1, one mechanism of Mer‐mediated cell survival involves activation of Ack1 and subsequent downregulation of the tumor suppressor Wwox (Mahajan et al., 2005). In the same study, expression of constitutively active Ack1 in human prostate adenocarcinoma cells induced anchorage‐independent growth in vitro and dramatically increased tumor growth in an ectopic xenograft model. Furthermore, patient samples of androgen‐independent prostate cancer (AICaP), an advanced stage of prostate

TAM Receptor Tyrosine Kinases

67

cancer with no curative treatment, exhibited 4–5‐fold more phosphorylated (i.e., active) Ack1 and approximately 6‐fold less Wwox protein than does normal prostate. These data suggest that Mer stimulation not only activates prosurvival pathways, but also downregulates proapoptotic pathways in this cell type. Mer is not expressed in normal mouse or human lymphocytes but is ectopically expressed in T-cell leukemias and E2A-PBX1 positive B-cell leukemias (Graham et al., 2006; Yeoh et al., 2002). To further investigate the oncogenic potential and causal role of Mer in malignancy, our lab developed a Mer transgenic (MerTg) mouse model (Keating et al., 2006). Using a C57Bl/6 background mouse, the expression of full‐length Mer cDNA was introduced under the control of a Vav1 promoter in mouse lymphocytes and thymocytes. MerTg lymphocytes exhibited a functional survival advantage in vitro compared with wild‐type lymphocytes when treated with glucocorticoids, a standard leukemia therapy. The MerTg lymphocytes also exhibited phosphorylation of Mer and robust activation of antiapoptotic pathways, including Akt and Erk1/2. Additionally, ectopic expression of Mer in lymphocytes, as is found in T‐cell lymphoblastic leukemia patient samples, promoted the development of T‐cell predominant leukemia and lymphoma in 55–58% of mice compared to a WT rate of only 12%. Interestingly, Mer expression was significantly correlated with the immature thymocyte stage (CD3, CD4, and CD8 negative), which is a subset of leukemia that portends a poor prognosis with difficulty to reach and retain remission. Resistance to conventional therapies, such as glucocorticoids, in Mer positive leukemias may also indicate a poor prognosis. The Mer transgenic mouse model of T‐cell acute lymphoblastic leukemia (T‐ALL) therefore suggests that TAM receptors may provide a novel target for future therapy development. Further investigation is necessary to substantiate a correlation between TAM receptor induced cell survival and carcinogenesis and tumor growth in vivo.

D. TAM Receptors as Prognostic Factors Elevated Axl expression correlated with adherence, motility, and invasiveness of osteosarcoma cell lines selected for their high metastatic ability in an in vivo model of lung metastasis (Nakano et al., 2003). In addition, lung metastasis has been correlated with reduced overall survival of osteosarcoma patients (Tsuchiya et al., 2002). The previous results therefore suggest that Axl expression may correlate with poor prognosis in osteosarcoma. Similarly, analysis of 58 adenocarcinoma patient samples revealed that Axl expression significantly correlated with metastatic cancer of advanced clinical stage (Shieh et al., 2005). Axl expression also correlated with invasiveness of lung cancer cell lines in vitro. In 54 patient samples of AML, Axl

Rachel M. A. Linger et al.

68

expression correlated with worse progression‐free and overall survival (Rochlitz et al., 1999). Interestingly, coexpression of both Mer and Axl correlates with poor prognosis in gastric cancer (Wu et al., 2002), suggesting that cooperativity of multiple TAM receptors may play a role in progression and metastasis of some cancers. These data suggest that TAM receptor signaling may play a role in the progression of multiple cancers, including the development of metastasis.

IV. POTENTIAL THERAPEUTIC APPLICATIONS Several studies have validated the therapeutic potential of targeting the TAM family in cancer therapy (Table III). Axl RTK and the ligand Gas6 are overexpressed in human glioma cell lines and malignant glioma patient samples when evaluated by microarray, Northern blot, Western blot, and immunohistochemistry analysis (Vajkoczy et al., 2006), leading to baseline constitutive activation of Axl RTK. In order to further investigate the role of Axl in glioblastoma tumorigenesis, these researchers transfected SF126 cells, a human glioma cell line exhibiting high levels of Axl RTK expression, with either wild‐type Axl (Axl‐WT) or Axl‐DN which lacks the tyrosine kinase domain. Cells expressing Axl‐DN demonstrated inhibition of Axl RTK activity, decreased proliferation, and reduced invasive potential relative to Axl‐WT. Orthotopic implantation of Axl‐DN cells resulted in markedly reduced tumor growth in vivo compared to Axl‐WT. Furthermore, mice which received Axl‐WT died within 30 days of implantation whereas 50% of the mice which received Axl‐DN were still alive 70 days after

Table III Axl as a Therapeutic Target Inhibitor Axl‐DN construct

shRNA Axl construct Axl small molecule inhibitor (MP470)

Outcome Reduced glioblastoma growth and invasiveness in vitro and in vivo; Increased overall survival after orthotopic implantation of glioblastoma cells containing Axl‐DN Decreased growth of breast carcinoma tumors in an ectopic xenograft model May inhibit in vitro Axl kinase activity with limited selectivity; cytotoxic to gastrointestinal stromal tumor cells in vitro

Reference Vajkoczy et al. (2006)

Holland et al. (2005) Mahadevan et al. (2007)

TAM Receptor Tyrosine Kinases

69

implantation. This study suggests that inhibition of Axl kinase activity reduces glioblastoma tumor growth and invasiveness and improves overall survival. Axl RTK is expressed in lung adenocarcinoma cell lines, and the level of Axl expression correlates with the invasive ability of these cell lines in vitro (Lay et al., 2007; Shieh et al., 2005). Ectopic overexpression of Axl in adenocarcinoma cell lines leads to increased formation of filipodia, migration, and drug resistance. Conversely, shRNA knockdown of Axl protein levels results in decreased migration (Lay et al., 2007). These studies suggest that blockade of Axl signaling would offer a new therapeutic strategy for this tumor type. Overexpression of TAM receptors has also been reported in breast cancer (Berclaz et al., 2001; Taylor et al., 1995b). Stable shRNA knockdown of Axl significantly reduced tumor growth in a xenograft model of breast carcinoma (Holland et al., 2005). In the same study, inhibition of Axl with small interfering RNA in human umbilical vein endothelial cells (HUVECs) blocked endothelial tube formation in vitro suggesting that inhibition of Axl may restrict mechanisms of angiogenesis required for breast cancer tumor cell growth. The aforementioned studies suggest that downregulation of Axl and its family members with currently existing or new biologically targeted tyrosine kinase inhibitors may prove to be a viable treatment option for historically difficult to treat cancers such as glioblastoma and drug‐resistant lung adenocarcinoma. Because the TAM family of tyrosine kinases has been implicated in the pathophysiology of several malignancies, they offer unique targets for new therapeutics. Several nonspecific tyrosine kinase inhibitors (e.g., Gleevec, Erlotinib, Dasatinib, and others) are already in use for a variety of malignancies, and have proven to be both efficacious and less toxic than standard chemotherapies. These tyrosine kinase inhibitors may prevent activation of the TAM family kinases in addition to other RTKs, leading to downregulation of cell survival pathways, thereby slowing growth and metastasis of malignancy. Additional mechanisms of TAM receptor inhibition could include soluble receptors that soak up available ligand, or direct binding to the receptor by monoclonal antibodies. The latter might block activation, desensitize, or downregulate the surface receptor, or call in an immune response. Many of these types of activities have been ascribed to the anti‐HER2 monoclonal antibody, Herceptin.

A. Small Molecule Inhibitors To date, only one small molecule inhibitor designed to inhibit TAM receptor function has been reported in the literature. MP470 is a potential Axl inhibitor but also blocks other tyrosine kinases within the same

70

Rachel M. A. Linger et al.

concentration range (Mahadevan et al., 2007). Sensitivity of Mer and Tyro‐3 to MP470 has not been tested. Thus, it is not clear how selectively this molecule inhibits Axl. Nevertheless, MP470 reduces the metabolic activity of an Axl‐expressing, drug‐resistant, gastrointestinal stromal tumor (GIST) cell line, suggesting that this novel drug may provide new treatment strategies for drug‐resistant cancers. In addition, these studies further validate Axl as a therapeutic target for treatment of cancer and provide promising evidence for future selective small molecule inhibitors of TAM receptors.

B. Soluble Receptors As discussed previously in Section II.D.3, the TAM family members undergo alternative splicing or shedding or both of their extracellular portion, leading to production of soluble receptors. These soluble receptors lack the tyrosine kinase domain, and act as a ligand sink to sequester ligand, thereby limiting signaling through the full‐length RTK (Costa et al., 1996). For example, treatment of NIH3T3 cells with soluble Axl ectodomains led to inhibition of Axl signaling and a decrease in DNA synthesis (Costa et al., 1996). In investigating soluble ectodomains as a potential treatment mechanism in malignancy, Sainaghi et al. (2005) evaluated proliferation of Axl‐expressing prostate carcinoma cell lines. They found that treatment with secreted Axl ectodomains abrogated Gas6‐induced stimulation and cell proliferation.

C. Antibodies Antibody therapy for cancer treatment has been theoretically promising but only successful in limited areas, such as Rituximab (anti‐CD20) used in treating lymphoma (Foran et al., 2000a,b) or anti‐GD2 used in neuroblastoma (Handgretinger et al., 1992, 1995). The discovery of novel important targets in the oncogenic process that can be effectively inhibited by antibody presence has been the limiting step. The ectopic surface expression of TAM family members, such as occurs with Mer in lymphoblastic leukemia and lymphoma, makes inhibition of TAM receptors with antibodies attractive. Angelillo‐Scherrer et al. (2001) have provided in vivo proof of concept of inhibition of TAM receptors with an anti‐Gas6 antibody to prevent fatal thrombosis in mouse models. Therefore, in the case of cancer therapeutics this antibody could potentially be used to neutralize Gas6 activity and thus reduce signaling through all of the TAM family members. Further investigations into antibody use and application are warranted.

TAM Receptor Tyrosine Kinases

71

D. Liabilities of TAM Receptor Antagonism Because the TAM family of RTKs performs several normal cellular functions (see Section II.E), there is potential for concern regarding the effects of inhibiting these RTKs. For example, mutations in the Mer gene lead to defective phagocytosis of photoreceptor outer segments by the RPE resulting in retinal degeneration. However, rodent studies suggest that retinal degeneration only occurs after prolonged Mer inhibition. These data are supported by reports of human patients with deactivating mutations of Mer (Gal et al., 2000). In three patients described with either heterozygous or homozygous Mer mutations, poor vision and night blindness were first noted in childhood. Thus retinal degeneration caused by Mer inhibition is likely to develop gradually over the course of several years and there may be a “therapeutic window” during which short‐term therapy with agents biologically targeted to inhibit Mer may be a feasible strategy for treatment of Mer positive cancers such as lymphoblastic leukemia. Furthermore, if vision changes were observed, studies in rodents suggest that cessation of therapy would restore normal vision (Vollrath et al., 2001). While the above example is specific to Mer antagonism, inhibition of Axl and Tyro‐3 also elicit potential concerns. As discussed in Section II.E.1, TAM knockout mice develop autoimmune diseases including lupus‐like syndrome (Cohen et al., 2002; Lemke and Lu, 2003). In a similar manner, chronic antagonism of TAM receptors may lead to autoimmune disease in humans. However, such effects are unlikely with short‐term therapeutic inhibition of TAM receptors. Furthermore, autoimmunity phenotypes were most pronounced in double and triple TAM receptor knockouts suggesting that the development of biological therapeutics which selectively target individual TAM family members would reduce the likelihood of adverse effects. Conventional chemotherapies cause a multitude of serious toxicities, most notably bone marrow suppression, kidney and liver dysfunction, and neuropathies. There is no evidence to suggest that TAM RTK inhibition would have overlapping toxicity profiles with current conventional therapies. In fact, one might expect that selective inhibitors of TAM receptors would exhibit minimal systemic toxicity. Nevertheless, this is an area that requires further preclinical investigation. Certainly the safety of many nonspecific tyrosine kinase inhibitors is already well proven. It remains to be seen whether these inhibitors affect the TAM RTK family, thereby inhibiting cell survival pathways as is hypothesized. More direct and specific inhibition of Mer, Axl, and Tyro‐3 using selective small molecule tyrosine kinase inhibitors, soluble receptors, and/or antibodies has not been accomplished in animal models of cancer as yet, and therefore more research is needed.

72

Rachel M. A. Linger et al.

V. CONCLUSIONS In the last decade, research has established the link between abnormal TAM receptor expression and oncogenesis. All three receptors are over‐ or ectopically expressed in a wide spectrum of human cancers, and overexpression of TAM receptors is sufficient to transform cells. TAM receptor inhibitions in animal xenograft tumor models of glioblastoma and breast cancer have provided preliminary validation of this receptor family as a cancer therapy target. More than half of the known RTKs have been directly implicated in human cancer. Although cancer is a multistep process, the success of targeted RTK inhibition in clinical cancer trials has demonstrated that blocking activity of a single dominant activated RTK can affect tumor growth, leading to widespread development of this class of drugs. In fact, in the last 15 years, novel targeted therapies led to the FDA approval of more cancer drugs than in the preceding 40 years combined. Some of the targeted therapies against tyrosine kinases, such as inhibitors of abl, EGFR, and VEGFR, have clearly improved patient survival with minimal additional toxicity. Many posit that this phase of tyrosine kinase inhibitory drug development is winding down. But as the ability to molecularly type human tumors and uncover or implicate additional tyrosine kinases as targets in tumor subsets improves, we believe this group of targets and their inhibitors will continue to supplement cytotoxic therapy.

ACKNOWLEDGEMENTS Research in the authors’ laboratories is supported by grants from the G&P Foundation for Cancer Research (#030, DKG), the Brent Eley Foundation (AKK) and the NCI (HSE). Douglas K. Graham is the Damon Runyon-Novartis Clinical Investigator supported in part by the Damon Runyon Cancer Research Foundation (CI‐39–07). The authors thank Drs. James DeGregori, Deborah DeRyckere, and Peter Henson for critical review of the manuscript.

REFERENCES Allen, M. P., Zeng, C., Schneider, K., Xiong, X., Meintzer, M. K., Bellosta, P., Basilico, C., Varnum, B., Heidenreich, K. A., and Wierman, M. E. (1999). Growth arrest‐specific gene 6 (Gas6)/adhesion related kinase (Ark) signaling promotes gonadotropin‐releasing hormone neuronal survival via extracellular signal‐regulated kinase (ERK) and Akt. Mol. Endocrinol. 13(2), 191–201.

TAM Receptor Tyrosine Kinases

73

Allen, M. P., Linseman, D. A., Udo, H., Xu, M., Schaack, J. B., Varnum, B., Kandel, E. R., Heidenreich, K. A., and Wierman, M. E. (2002). Novel mechanism for gonadotropin‐ releasing hormone neuronal migration involving Gas6/Ark signaling to p38 mitogen‐ activated protein kinase. Mol. Cell Biol. 22(2), 599–613. Anderson, H. A., Maylock, C. A., Williams, J. A., Paweletz, C. P., Shu, H., and Shacter, E. (2003). Serum‐derived protein S binds to phosphatidylserine and stimulates the phagocytosis of apoptotic cells. Nat. Immunol. 4(1), 87–91. Angelillo‐Scherrer, A., de Frutos, P., Aparicio, C., Melis, E., Savi, P., Lupu, F., Arnout, J., Dewerchin, M., Hoylaerts, M., Herbert, J., Collen, D., Dahlback, B., et al. (2001). Deficiency or inhibition of Gas6 causes platelet dysfunction and protects mice against thrombosis. Nat. Med. 7(2), 215–221. Angelillo‐Scherrer, A., Burnier, L., Flores, N., Savi, P., DeMol, M., Schaeffer, P., Herbert, J. M., Lemke, G., Goff, S. P., Matsushima, G. K., Earp, H. S., Vesin, C., et al. (2005). Role of Gas6 receptors in platelet signaling during thrombus stabilization and implications for antithrombotic therapy. J. Clin. Invest. 115(2), 237–246. Behrens, E. M., Gadue, P., Gong, S. Y., Garrett, S., Stein, P. L., and Cohen, P. L. (2003). The mer receptor tyrosine kinase: Expression and function suggest a role in innate immunity. Eur. J. Immunol. 33(8), 2160–2167. Bellosta, P., Costa, M., Lin, D. A., and Basilico, C. (1995). The receptor tyrosine kinase ARK mediates cell aggregation by homophilic binding. Mol. Cell Biol. 15(2), 614–625. Bellosta, P., Zhang, Q., Goff, S. P., and Basilico, C. (1997). Signaling through the ARK tyrosine kinase receptor protects from apoptosis in the absence of growth stimulation. Oncogene 15(20), 2387–2397. Berclaz, G., Altermatt, H. J., Rohrbach, V., Kieffer, I., Dreher, E., and Andres, A. C. (2001). Estrogen dependent expression of the receptor tyrosine kinase axl in normal and malignant human breast. Ann. Oncol. 12(6), 819–824. Biesecker, L. G., Gottschalk, L. R., and Emerson, S. G. (1993). Identification of four murine cDNAs encoding putative protein kinases from primitive embryonic stem cells differentiated in vitro. Proc. Natl. Acad. Sci. USA 90(15), 7044–7048. Biesecker, L. G., Giannola, D. M., and Emerson, S. G. (1995). Identification of alternative exons, including a novel exon, in the tyrosine kinase receptor gene Etk2/tyro3 that explain differences in 50 cDNA sequences. Oncogene 10(11), 2239–2242. Birchmeier, C., Birchmeier, W., Gherardi, E., and Vande Woude, G. F. (2003). Met, metastasis, motility and more. Nat. Rev. Mol. Cell Biol. 4(12), 915–925. Biscardi, J. S., Denhez, F., Buehler, G. F., Chesnutt, D. A., Baragona, S. C., O’Bryan, J. P., Der, C. J., Fiordalisi, J. J., Fults, D. W., and Maness, P. F. (1996). Rek, a gene expressed in retina and brain, encodes a receptor tyrosine kinase of the Axl/Tyro3 family. J. Biol. Chem. 271(46), 29049–29059. Braunger, J., Schleithoff, L., Schulz, A. S., Kessler, H., Lammers, R., Ullrich, A., Bartram, C. R., and Janssen, J. W. (1997). Intracellular signaling of the Ufo/Axl receptor tyrosine kinase is mediated mainly by a multi‐substrate docking‐site. Oncogene 14(22), 2619–2631. Budagian, V., Bulanova, E., Orinska, Z., Duitman, E., Brandt, K., Ludwig, A., Hartmann, D., Lemke, G., Saftig, P., and Bulfone‐Paus, S. (2005a). Soluble Axl is generated by ADAM10‐ dependent cleavage and associates with Gas6 in mouse serum. Mol. Cell Biol. 25(21), 9324–9339. Budagian, V., Bulanova, E., Orinska, Z., Thon, L., Mamat, U., Bellosta, P., Basilico, C., Adam, D., Paus, R., and Bulfone‐Paus, S. (2005b). A promiscuous liaison between IL‐15 receptor and Axl receptor tyrosine kinase in cell death control. EMBO J. 24(24), 4260–4270. Burchert, A., Attar, E. C., McCloskey, P., Fridell, Y. W., and Liu, E. T. (1998). Determinants for transformation induced by the Axl receptor tyrosine kinase. Oncogene 16(24), 3177–3187.

74

Rachel M. A. Linger et al.

Camenisch, T. D., Koller, B. H., Earp, H. S., and Matsushima, G. K. (1999). A novel receptor tyrosine kinase, Mer, inhibits TNF‐alpha production and lipopolysaccharide‐induced endotoxic shock. J. Immunol. 162(6), 3498–3503. Caraux, A., Lu, Q., Fernandez, N., Riou, S., Di Santo, J. P., Raulet, D. H., Lemke, G., and Roth, C. (2006). Natural killer cell differentiation driven by Tyro3 receptor tyrosine kinases. Nat. Immunol. 7(7), 747–754. Challier, C., Uphoff, C. C., Janssen, J. W., and Drexler, H. G. (1996). Differential expression of the ufo/axl oncogene in human leukemia‐lymphoma cell lines. Leukemia 10(5), 781–787. Chan, M. C., Mather, J. P., McCray, G., and Lee, W. M. (2000). Identification and regulation of receptor tyrosine kinases Rse and Mer and their ligand Gas6 in testicular somatic cells. J. Androl. 21(2), 291–302. Chen, J., Carey, K., and Godowski, P. J. (1997). Identification of Gas6 as a ligand for Mer, a neural cell adhesion molecule related receptor tyrosine kinase implicated in cellular transformation. Oncogene 14(17), 2033–2039. Chung, B. I., Malkowicz, S. B., Nguyen, T. B., Libertino, J. A., and McGarvey, T. W. (2003). Expression of the proto‐oncogene Axl in renal cell carcinoma. DNA Cell Biol. 22(8), 533–540. Cohen, P. L., Caricchio, R., Abraham, V., Camenisch, T. D., Jennette, J. C., Roubey, R. A., Earp, H. S., Matsushima, G., and Reap, E. A. (2002). Delayed apoptotic cell clearance and lupus‐like autoimmunity in mice lacking the c‐mer membrane tyrosine kinase. J. Exp. Med. 196(1), 135–140. Costa, M., Bellosta, P., and Basilico, C. (1996). Cleavage and release of a soluble form of the receptor tyrosine kinase ARK in vitro and in vivo. J. Cell. Physiol. 168(3), 737–744. Craven, R. J., Xu, L. H., Weiner, T. M., Fridell, Y. W., Dent, G. A., Srivastava, S., Varnum, B., Liu, E. T., and Cance, W. G. (1995). Receptor tyrosine kinases expressed in metastatic colon cancer. Int. J. Cancer 60(6), 791–797. Crosier, P. S., Lewis, P. M., Hall, L. R., Vitas, M. R., Morris, C. M., Beier, D. R., Wood, C. R., and Crosier, K. E. (1994). Isolation of a receptor tyrosine kinase (DTK) from embryonic stem cells: Structure, genetic mapping and analysis of expression. Growth Factors 11(2), 125–136. Crosier, P. S., Hall, L. R., Vitas, M. R., Lewis, P. M., and Crosier, K. E. (1995). Identification of a novel receptor tyrosine kinase expressed in acute myeloid leukemic blasts. Leuk. Lymphoma 18(5–6), 443–449. Crosier, P. S., Freeman, S. A., Orlic, D., Bodine, D. M., and Crosier, K. E. (1996). The Dtk receptor tyrosine kinase, which binds protein S, is expressed during hematopoiesis. Exp. Hematol. 24(2), 318–323. D’Cruz, P. M., Yasumura, D., Weir, J., Matthes, M. T., Abderrahim, H., LaVail, M. M., and Vollrath, D. (2000). Mutation of the receptor tyrosine kinase gene Mertk in the retinal dystrophic RCS rat. Hum. Mol. Genet. 9(4), 645–651. Dahlback, B., and Villoutreix, B. O. (2005). Regulation of blood coagulation by the protein C anticoagulant pathway: Novel insights into structure‐function relationships and molecular recognition. Arterioscler. Thromb. Vasc. Biol. 25(7), 1311–1320. Dai, W., Pan, H., Hassanain, H., Gupta, S. L., and Murphy, M. J., Jr. (1994). Molecular cloning of a novel receptor tyrosine kinase, tif, highly expressed in human ovary and testis. Oncogene 9(3), 975–979. De Vos, J., Couderc, G., Tarte, K., Jourdan, M., Requirand, G., Delteil, M. C., Rossi, J. F., Mechti, N., and Klein, B. (2001). Identifying intercellular signaling genes expressed in malignant plasma cells by using complementary DNA arrays. Blood 98(3), 771–780. Demarchi, F., Verardo, R., Varnum, B., Brancolini, C., and Schneider, C. (2001). Gas6 anti‐ apoptotic signaling requires NF‐kappa B activation. J. Biol. Chem. 276(34), 31738–31744.

TAM Receptor Tyrosine Kinases

75

Duncan, J. L., LaVail, M. M., Yasumura, D., Matthes, M. T., Yang, H., Trautmann, N., Chappelow, A. V., Feng, W., Earp, H. S., Matsushima, G. K., and Vollrath, D. (2003). An RCS‐like retinal dystrophy phenotype in mer knockout mice. Invest. Ophthalmol. Vis. Sci. 44(2), 826–838. Ek, S., Hogerkorp, C. M., Dictor, M., Ehinger, M., and Borrebaeck, C. A. (2002). Mantle cell lymphomas express a distinct genetic signature affecting lymphocyte trafficking and growth regulation as compared with subpopulations of normal human B cells. Cancer Res. 62(15), 4398–4405. Evans, C. O., Young, A. N., Brown, M. R., Brat, D. J., Parks, J. S., Neish, A. S., and Oyesiku, N. M. (2001). Novel patterns of gene expression in pituitary adenomas identified by complementary deoxyribonucleic acid microarrays and quantitative reverse transcription‐ polymerase chain reaction. J. Clin. Endocrinol. Metab. 86(7), 3097–3107. Faust, M., Ebensperger, C., Schulz, A. S., Schleithoff, L., Hameister, H., Bartram, C. R., and Janssen, J. W. (1992). The murine ufo receptor: Molecular cloning, chromosomal localization and in situ expression analysis. Oncogene 7(7), 1287–1293. Fisher, P. W., Brigham‐Burke, M., Wu, S. J., Luo, J., Carton, J., Staquet, K., Gao, W., Jackson, S., Bethea, D., Chen, C., Hu, B., Giles‐Komar, J., et al. (2005). A novel site contributing to growth‐arrest‐specific gene 6 binding to its receptors as revealed by a human monoclonal antibody. Biochem. J. 387(Pt 3), 727–735. Foran, J. M., Cunningham, D., Coiffier, B., Solal‐Celigny, P., Reyes, F., Ghielmini, M., Johnson, P. W., Gisselbrecht, C., Bradburn, M., Matthews, J., and Lister, T. A. (2000a). Treatment of mantle‐cell lymphoma with Rituximab (chimeric monoclonal anti‐CD20 antibody): Analysis of factors associated with response. Ann. Oncol. 11(Suppl 1), 117–121. Foran, J. M., Rohatiner, A. Z., Cunningham, D., Popescu, R. A., Solal‐Celigny, P., Ghielmini, M., Coiffier, B., Johnson, P. W., Gisselbrecht, C., Reyes, F., Radford, J. A., Bessell, E. M., et al. (2000b). European phase II study of rituximab (chimeric anti‐CD20 monoclonal antibody) for patients with newly diagnosed mantle‐cell lymphoma and previously treated mantle‐cell lymphoma, immunocytoma, and small B‐cell lymphocytic lymphoma. J. Clin. Oncol. 18(2), 317–324. Fridell, Y. W., Jin, Y., Quilliam, L. A., Burchert, A., McCloskey, P., Spizz, G., Varnum, B., Der, C., and Liu, E. T. (1996). Differential activation of the Ras/extracellular‐signal‐regulated protein kinase pathway is responsible for the biological consequences induced by the Axl receptor tyrosine kinase. Mol. Cell. Biol. 16(1), 135–145. Fridell, Y. W., Villa, J., Jr., Attar, E. C., and Liu, E. T. (1998). GAS6 induces Axl‐mediated chemotaxis of vascular smooth muscle cells. J. Biol. Chem. 273(12), 7123–7126. Fujimoto, J., and Yamamoto, T. (1994). brt, a mouse gene encoding a novel receptor‐ type protein‐tyrosine kinase, is preferentially expressed in the brain. Oncogene 9(3), 693–698. Funakoshi, H., Yonemasu, T., Nakano, T., Matumoto, K., and Nakamura, T. (2002). Identification of Gas6, a putative ligand for Sky and Axl receptor tyrosine kinases, as a novel neurotrophic factor for hippocampal neurons. J. Neurosci. Res. 68(2), 150–160. Gal, A., Li, Y., Thompson, D. A., Weir, J., Orth, U., Jacobson, S. G., Apfelstedt‐Sylla, E., and Vollrath, D. (2000). Mutations in MERTK, the human orthologue of the RCS rat retinal dystrophy gene, cause retinitis pigmentosa. Nat. Genet. 26(3), 270–271. Georgescu, M. M., Kirsch, K. H., Shishido, T., Zong, C., and Hanafusa, H. (1999). Biological effects of c‐Mer receptor tyrosine kinase in hematopoietic cells depend on the Grb2 binding site in the receptor and activation of NF‐kappaB. Mol. Cell Biol. 19(2), 1171–1181. Godowski, P. J., Mark, M. R., Chen, J., Sadick, M. D., Raab, H., and Hammonds, R. G. (1995). Reevaluation of the roles of protein S and Gas6 as ligands for the receptor tyrosine kinase Rse/Tyro 3. Cell 82(3), 355–358.

76

Rachel M. A. Linger et al.

Goruppi, S., Ruaro, E., Varnum, B., and Schneider, C. (1997). Requirement of phosphatidylinositol 3‐kinase‐dependent pathway and Src for Gas6‐Axl mitogenic and survival activities in NIH 3T3 fibroblasts. Mol. Cell Biol. 17(8), 4442–4453. Goruppi, S., Ruaro, E., Varnum, B., and Schneider, C. (1999). Gas6‐mediated survival in NIH3T3 cells activates stress signalling cascade and is independent of Ras. Oncogene 18(29), 4224–4236. Graham, D. K., Dawson, T. L., Mullaney, D. L., Snodgrass, H. R., and Earp, H. S. (1994). Cloning and mRNA expression analysis of a novel human protooncogene, c‐mer. Cell Growth Differ. 5(6), 647–657. Graham, D. K., Bowman, G. W., Dawson, T. L., Stanford, W. L., Earp, H. S., and Snodgrass, H. R. (1995). Cloning and developmental expression analysis of the murine c‐mer tyrosine kinase. Oncogene 10(12), 2349–2359. Graham, D. K., Salzberg, D. B., Kurtzberg, J., Sather, S., Matsushima, G. K., Keating, A. K., Liang, X., Lovell, M. A., Williams, S. A., Dawson, T. L., Schell, M. J., Anwar, A. A., et al. (2006). Ectopic expression of the proto‐oncogene Mer in pediatric T‐cell acute lymphoblastic leukemia. Clin. Cancer Res. 12(9), 2662–2669. Gustin, J. A., Ozes, O. N., Akca, H., Pincheira, R., Mayo, L. D., Li, Q., Guzman, J. R., Korgaonkar, C. K., and Donner, D. B. (2004). Cell type‐specific expression of the IkappaB kinases determines the significance of phosphatidylinositol 3‐kinase/Akt signaling to NF‐kappa B activation. J. Biol. Chem. 279(3), 1615–1620. Guttridge, K. L., Luft, J. C., Dawson, T. L., Kozlowska, E., Mahajan, N. P., Varnum, B., and Earp, H. S. (2002). Mer receptor tyrosine kinase signaling: Prevention of apoptosis and alteration of cytoskeletal architecture without stimulation or proliferation. J. Biol. Chem. 277(27), 24057–24066. Gyorffy, B., and Lage, H. (2007). A Web‐based data warehouse on gene expression in human malignant melanoma. J. Invest. Dermatol. 127(2), 394–399. Hafizi, S., Alindri, F., Karlsson, R., and Dahlback, B. (2002). Interaction of Axl receptor tyrosine kinase with C1‐TEN, a novel C1 domain‐containing protein with homology to tensin. Biochem. Biophys. Res. Commun. 299(5), 793–800. Hafizi, S., Gustafsson, A., Stenhoff, J., and Dahlback, B. (2005a). The Ran binding protein RanBPM interacts with Axl and Sky receptor tyrosine kinases. Int. J. Biochem. Cell Biol. 37(11), 2344–2356. Hafizi, S., Ibraimi, F., and Dahlback, B. (2005b). C1‐TEN is a negative regulator of the Akt/PKB signal transduction pathway and inhibits cell survival, proliferation, and migration. FASEB J. 19(8), 971–973. Hafizi, S., and Dahlback, B. (2006a). Signalling and functional diversity within the Axl subfamily of receptor tyrosine kinases. Cytokine Growth Factor Rev. 17(4), 295–304. Hafizi, S., and Dahlback, B. (2006b). Gas6 and protein S. Vitamin K‐dependent ligands for the Axl receptor tyrosine kinase subfamily. FEBS J. 273(23), 5231–5244. Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100(1), 57–70. Handgretinger, R., Baader, P., Dopfer, R., Klingebiel, T., Reuland, P., Treuner, J., Reisfeld, R. A., and Niethammer, D. (1992). A phase I study of neuroblastoma with the anti‐ganglioside GD2 antibody 14.G2a. Cancer Immunol. Immunother. 35(3), 199–204. Handgretinger, R., Anderson, K., Lang, P., Dopfer, R., Klingebiel, T., Schrappe, M., Reuland, P., Gillies, S. D., Reisfeld, R. A., and Neithammer, D. (1995). A phase I study of human/mouse chimeric antiganglioside GD2 antibody ch14.18 in patients with neuroblastoma. Eur. J. Cancer 31(2), 261–267. Hasanbasic, I., Cuerquis, J., Varnum, B., and Blostein, M. D. (2004). Intracellular signaling pathways involved in Gas6‐Axl‐mediated survival of endothelial cells. Am. J. Physiol. Heart Circ. Physiol. 287(3), H1207–1213.

TAM Receptor Tyrosine Kinases

77

Heiring, C., Dahlback, B., and Muller, Y. A. (2004). Ligand recognition and homophilic interactions in Tyro3: Structural insights into the Axl/Tyro3 receptor tyrosine kinase family. J. Biol. Chem. 279(8), 6952–6958. Holland, S. J., Powell, M. J., Franci, C., Chan, E. W., Friera, A. M., Atchison, R. E., McLaughlin, J., Swift, S. E., Pali, E. S., Yam, G., Wong, S., Lasaga, J., et al. (2005). Multiple roles for the receptor tyrosine kinase axl in tumor formation. Cancer Res. 65(20), 9294–9303. Hu, B., Jennings, J. H., Sonstein, J., Floros, J., Todt, J. C., Polak, T., and Curtis, J. L. (2004). Resident murine alveolar and peritoneal macrophages differ in adhesion of apoptotic thymocytes. Am. J. Respir. Cell Mol. Biol. 30(5), 687–693. Ishimoto, Y., Ohashi, K., Mizuno, K., and Nakano, T. (2000). Promotion of the uptake of PS liposomes and apoptotic cells by a product of growth arrest‐specific gene, gas6. J. Biochem. (Tokyo) 127(3), 411–417. Ito, M., Nakashima, M., Nakayama, T., Ohtsuru, A., Nagayama, Y., Takamura, N., Demedchik, E. P., Sekine, I., and Yamashita, S. (2002). Expression of receptor‐type tyrosine kinase, Axl, and its ligand, Gas6, in pediatric thyroid carcinomas around chernobyl. Thyroid 12(11), 971–975. Ito, T., Ito, M., Naito, S., Ohtsuru, A., Nagayama, Y., Kanematsu, T., Yamashita, S., and Sekine, I. (1999). Expression of the Axl receptor tyrosine kinase in human thyroid carcinoma. Thyroid 9(6), 563–567. Jacob, A. N., Kalapurakal, J., Davidson, W. R., Kandpal, G., Dunson, N., Prashar, Y., and Kandpal, R. P. (1999). A receptor tyrosine kinase, UFO/Axl, and other genes isolated by a modified differential display PCR are overexpressed in metastatic prostatic carcinoma cell line DU145. Cancer Detect. Prev. 23(4), 325–332. Janssen, J. W., Schulz, A. S., Steenvoorden, A. C., Schmidberger, M., Strehl, S., Ambros, P. F., and Bartram, C. R. (1991). A novel putative tyrosine kinase receptor with oncogenic potential. Oncogene 6(11), 2113–2120. Jia, R., and Hanafusa, H. (1994). The proto‐oncogene of v‐eyk (v‐ryk) is a novel receptor‐type protein tyrosine kinase with extracellular Ig/GN‐III domains. J. Biol. Chem. 269(3), 1839–1844. Jia, R., Mayer, B. J., Hanafusa, T., and Hanafusa, H. (1992). A novel oncogene, v‐ryk, encoding a truncated receptor tyrosine kinase is transduced into the RPL30 virus without loss of viral sequences. J. Virol. 66(10), 5975–5987. Katagiri, M., Hakeda, Y., Chikazu, D., Ogasawara, T., Takato, T., Kumegawa, M., Nakamura, K., and Kawaguchi, H. (2001). Mechanism of stimulation of osteoclastic bone resorption through Gas6/Tyro 3, a receptor tyrosine kinase signaling, in mouse osteoclasts. J. Biol. Chem. 276(10), 7376–7382. Keating, A. K., Salzberg, D. B., Sather, S., Liang, X., Nickoloff, S., Anwar, A., Deryckere, D., Hill, K., Joung, D., Sawczyn, K. K., Park, J., Curran‐Everett, D., et al. (2006). Lymphoblastic leukemia/lymphoma in mice overexpressing the Mer (MerTK) receptor tyrosine kinase. Oncogene Epub ahead of print, doi: 10.1038/sj.onc.1209633. Khan, J., Bittner, M. L., Saal, L. H., Teichmann, U., Azorsa, D. O., Gooden, G. C., Pavan, W. J., Trent, J. M., and Meltzer, P. S. (1999). cDNA microarrays detect activation of a myogenic transcription program by the PAX3‐FKHR fusion oncogene. Proc. Natl. Acad. Sci. USA 96(23), 13264–13269. Kinjyo, I., Hanada, T., Inagaki‐Ohara, K., Mori, H., Aki, D., Ohishi, M., Yoshida, H., Kubo, M., and Yoshimura, A. (2002). SOCS1/JAB is a negative regulator of LPS‐induced macrophage activation. Immunity 17(5), 583–591. Korshunov, V. A., Mohan, A. M., Georger, M. A., and Berk, B. C. (2006). Axl, a receptor tyrosine kinase, mediates flow‐induced vascular remodeling. Circ. Res. 98(11), 1446–1452.

78

Rachel M. A. Linger et al.

Lai, C., and Lemke, G. (1991). An extended family of protein‐tyrosine kinase genes differentially expressed in the vertebrate nervous system. Neuron 6(5), 691–704. Lai, C., Gore, M., and Lemke, G. (1994). Structure, expression, and activity of Tyro 3, a neural adhesion‐related receptor tyrosine kinase. Oncogene 9(9), 2567–2578. Lan, Z., Wu, H., Li, W., Wu, S., Lu, L., Xu, M., and Dai, W. (2000). Transforming activity of receptor tyrosine kinase tyro3 is mediated, at least in part, by the PI3 kinase‐signaling pathway. Blood 95(2), 633–638. Lay, J. D., Hong, C. C., Huang, J. S., Yang, Y. Y., Pao, C. Y., Liu, C. H., Lai, Y. P., Lai, G. M., Cheng, A. L., Su, I. J., and Chuang, S. E. (2007). Sulfasalazine suppresses drug resistance and invasiveness of lung adenocarcinoma cells expressing AXL. Cancer Res. 67(8), 3878–3887. Lee, W. P., Wen, Y., Varnum, B., and Hung, M. C. (2002). Akt is required for Axl‐Gas6 signaling to protect cells from E1A‐mediated apoptosis. Oncogene 21(3), 329–336. Lemke, G., and Lu, Q. (2003). Macrophage regulation by Tyro 3 family receptors. Curr. Opin. Immunol. 15(1), 31–36. Lewis, J. M., Cheresh, D. A., and Schwartz, M. A. (1996a). Protein kinase C regulates alpha v beta 5‐dependent cytoskeletal associations and focal adhesion kinase phosphorylation. J. Cell. Biol. 134(5), 1323–1332. Lewis, P. M., Crosier, K. E., Wood, C. R., and Crosier, P. S. (1996b). Analysis of the murine Dtk gene identifies conservation of genomic structure within a new receptor tyrosine kinase subfamily. Genomics 31(1), 13–19. Li, Y., Gerbod‐Giannone, M. C., Seitz, H., Cui, D., Thorp, E., Tall, A. R., Matsushima, G. K., and Tabas, I. (2006). Cholesterol‐induced apoptotic macrophages elicit an inflammatory response in phagocytes, which is partially attenuated by the Mer receptor. J. Biol. Chem. 281(10), 6707–6717. Lin, W. C., Li, A. F., Chi, C. W., Chung, W. W., Huang, C. L., Lui, W. Y., Kung, H. J., and Wu, C. W. (1999). tie‐1 protein tyrosine kinase: A novel independent prognostic marker for gastric cancer. Clin. Cancer Res. 5(7), 1745–1751. Ling, L., and Kung, H. J. (1995). Mitogenic signals and transforming potential of Nyk, a newly identified neural cell adhesion molecule‐related receptor tyrosine kinase. Mol. Cell Biol. 15(12), 6582–6592. Ling, L., Templeton, D., and Kung, H. J. (1996). Identification of the major autophosphorylation sites of Nyk/Mer, an NCAM‐related receptor tyrosine kinase. J. Biol. Chem. 271(31), 18355–18362. Liu, E., Hjelle, B., and Bishop, J. M. (1988). Transforming genes in chronic myelogenous leukemia. Proc. Natl. Acad. Sci. USA 85(6), 1952–1956. Lu, Q., Gore, M., Zhang, Q., Camenisch, T., Boast, S., Casagranda, F., Lai, C., Skinner, M. K., Klein, R., Matsushima, G. K., Earp, H. S., Goff, S. P., et al. (1999). Tyro‐3 family receptors are essential regulators of mammalian spermatogenesis. Nature 398(6729), 723–728. Lu, Q., and Lemke, G. (2001). Homeostatic regulation of the immune system by receptor tyrosine kinases of the Tyro 3 family. Science 293(5528), 306–311. Macleod, K., Mullen, P., Sewell, J., Rabiasz, G., Lawrie, S., Miller, E., Smyth, J. F., and Langdon, S. P. (2005). Altered ErbB receptor signaling and gene expression in cisplatin‐ resistant ovarian cancer. Cancer Res. 65(15), 6789–6800. Mahadevan, D., Cooke, L., Riley, C., Swart, R., Simons, B., Della Croce, K., Wisner, L., Iorio, M., Shakalya, K., Garewal, H., Nagle, R., and Bearss, D. (2007). A novel tyrosine kinase switch is a mechanism of imatinib resistance in gastrointestinal stromal tumors. Oncogene 26(27), 3909–3919. Mahajan, N. P., and Earp, H. S. (2003). An SH2 domain‐dependent, phosphotyrosine‐ independent interaction between Vav1 and the Mer receptor tyrosine kinase: A mechanism for localizing guanine nucleotide‐exchange factor action. J. Biol. Chem. 278(43), 42596–42603.

TAM Receptor Tyrosine Kinases

79

Mahajan, N. P., Whang, Y. E., Mohler, J. L., and Earp, H. S. (2005). Activated tyrosine kinase Ack1 promotes prostate tumorigenesis: Role of Ack1 in polyubiquitination of tumor suppressor Wwox. Cancer Res. 65(22), 10514–10523. Mark, M. R., Scadden, D. T., Wang, Z., Gu, Q., Goddard, A., and Godowski, P. J. (1994). rse, a novel receptor‐type tyrosine kinase with homology to Axl/Ufo, is expressed at high levels in the brain. J. Biol. Chem. 269(14), 10720–10728. Mark, M. R., Chen, J., Hammonds, R. G., Sadick, M., and Godowsk, P. J. (1996). Characterization of Gas6, a member of the superfamily of G domain‐containing proteins, as a ligand for Rse and Axl. J. Biol. Chem. 271(16), 9785–9789. McCloskey, P., Fridell, Y. W., Attar, E., Villa, J., Jin, Y., Varnum, B., and Liu, E. T. (1997). GAS6 mediates adhesion of cells expressing the receptor tyrosine kinase Axl. J. Biol. Chem. 272(37), 23285–23291. McHenry, C. L., Liu, Y., Feng, W., Nair, A. R., Feathers, K. L., Ding, X., Gal, A., Vollrath, D., Sieving, P. A., and Thompson, D. A. (2004). MERTK arginine‐844‐cysteine in a patient with severe rod‐cone dystrophy: Loss of mutant protein function in transfected cells. Invest. Ophthalmol. Vis. Sci. 45(5), 1456–1463. Melaragno, M. G., Wuthrich, D. A., Poppa, V., Gill, D., Lindner, V., Berk, B. C., and Corson, M. A. (1998). Increased expression of Axl tyrosine kinase after vascular injury and regulation by G protein‐coupled receptor agonists in rats. Circ. Res. 83(7), 697–704. Melaragno, M. G., Cavet, M. E., Yan, C., Tai, L. K., Jin, Z. G., Haendeler, J., and Berk, B. C. (2004). Gas6 inhibits apoptosis in vascular smooth muscle: Role of Axl kinase and Akt. J. Mol. Cell Cardiol. 37(4), 881–887. Meric, F., Lee, W. P., Sahin, A., Zhang, H., Kung, H. J., and Hung, M. C. (2002). Expression profile of tyrosine kinases in breast cancer. Clin. Cancer Res. 8(2), 361–367. Murao, K., Imachi, H., Sayo, Y., Hosokawa, H., Sato, M., Ishida, T., Nakano, T., Kodama, T., Sasaguri, Y., and Takahara, J. (1999). A product of growth arrest‐specific gene 6 modulates scavenger receptor expression in human vascular smooth muscle cells. FEBS Lett. 459(3), 363–366. Nagata, K., Ohashi, K., Nakano, T., Arita, H., Zong, C., Hanafusa, H., and Mizuno, K. (1996). Identification of the product of growth arrest‐specific gene 6 as a common ligand for Axl, Sky, and Mer receptor tyrosine kinases. J. Biol. Chem. 271(47), 30022–30027. Nakagawa, R., Naka, T., Tsutsui, H., Fujimoto, M., Kimura, A., Abe, T., Seki, E., Sato, S., Takeuchi, O., Takeda, K., Akira, S., Yamanishi, K., et al. (2002). SOCS‐1 participates in negative regulation of LPS responses. Immunity 17(5), 677–687. Nakano, T., Higashino, K., Kikuchi, N., Kishino, J., Nomura, K., Fujita, H., Ohara, O., and Arita, H. (1995). Vascular smooth muscle cell‐derived, Gla‐containing growth‐potentiating factor for Ca(2þ)‐mobilizing growth factors. J. Biol. Chem. 270(11), 5702–5705. Nakano, T., Ishimoto, Y., Kishino, J., Umeda, M., Inoue, K., Nagata, K., Ohashi, K., Mizuno, K., and Arita, H. (1997). Cell adhesion to phosphatidylserine mediated by a product of growth arrest‐specific gene 6. J. Biol. Chem. 272(47), 29411–29414. Nakano, T., Tani, M., Ishibashi, Y., Kimura, K., Park, Y. B., Imaizumi, N., Tsuda, H., Aoyagi, K., Sasaki, H., Ohwada, S., and Yokota, J. (2003). Biological properties and gene expression associated with metastatic potential of human osteosarcoma. Clin. Exp. Metastasis 20(7), 665–674. Nandrot, E., Dufour, E. M., Provost, A. C., Pequignot, M. O., Bonnel, S., Gogat, K., Marchant, D., Rouillac, C., Sepulchre de Conde, B., Bihoreau, M. T., Shaver, C., Dufier, J. L., et al. (2000). Homozygous deletion in the coding sequence of the c‐mer gene in RCS rats unravels general mechanisms of physiological cell adhesion and apoptosis. Neurobiol. Dis. 7(6 Pt B), 586–599.

80

Rachel M. A. Linger et al.

Neubauer, A., Fiebeler, A., Graham, D. K., O’Bryan, J. P., Schmidt, C. A., Barckow, P., Serke, S., Siegert, W., Snodgrass, H. R., Huhn, D., et al. (1994). Expression of axl, a transforming receptor tyrosine kinase, in normal and malignant hematopoiesis. Blood 84(6), 1931–1941. Nielsen‐Preiss, S. M., Allen, M. P., Xu, M., Linseman, D. A., Pawlowski, J. E., Bouchard, R. J., Varnum, B. C., Heidenreich, K. A., and Wierman, M. E. (2007). Adhesion‐related kinase induction of migration requires phosphatidylinositol‐3‐kinase and ras stimulation of rac activity in immortalized gonadotropin‐releasing hormone neuronal cells. Endocrinology 148(6), 2806–2814. O’Bryan, J. P., Frye, R. A., Cogswell, P. C., Neubauer, A., Kitch, B., Prokop, C., Espinosa, R., 3rd, Le Beau, M. M., Earp, H. S., and Liu, E. T. (1991). axl, a transforming gene isolated from primary human myeloid leukemia cells, encodes a novel receptor tyrosine kinase. Mol. Cell Biol. 11(10), 5016–5031. O’Bryan, J. P., Fridell, Y. W., Koski, R., Varnum, B., and Liu, E. T. (1995). The transforming receptor tyrosine kinase, Axl, is post‐translationally regulated by proteolytic cleavage. J. Biol. Chem. 270(2), 551–557. Ohashi, K., Mizuno, K., Kuma, K., Miyata, T., and Nakamura, T. (1994). Cloning of the cDNA for a novel receptor tyrosine kinase, Sky, predominantly expressed in brain. Oncogene 9(3), 699–705. Ohashi, K., Nagata, K., Toshima, J., Nakano, T., Arita, H., Tsuda, H., Suzuki, K., and Mizuno, K. (1995). Stimulation of sky receptor tyrosine kinase by the product of growth arrest‐specific gene 6. J. Biol. Chem. 270(39), 22681–22684. Polvi, A., Armstrong, E., Lai, C., Lemke, G., Huebner, K., Spritz, R. A., Guida, L. C., Nicholls, R. D., and Alitalo, K. (1993). The human TYRO3 gene and pseudogene are located in chromosome 15q14‐q25. Gene 134(2), 289–293. Prasad, D., Rothlin, C. V., Burrola, P., Burstyn‐Cohen, T., Lu, Q., Garcia de Frutos, P., and Lemke, G. (2006). TAM receptor function in the retinal pigment epithelium. Mol. Cell Neurosci. 33(1), 96–108. Prieto, A. L., Weber, J. L., Tracy, S., Heeb, M. J., and Lai, C. (1999). Gas6, a ligand for the receptor protein‐tyrosine kinase Tyro‐3, is widely expressed in the central nervous system. Brain Res. 816(2), 646–661. Prieto, A. L., Weber, J. L., and Lai, C. (2000). Expression of the receptor protein‐tyrosine kinases Tyro‐3, Axl, and mer in the developing rat central nervous system. J. Comp. Neurol. 425(2), 295–314. Quong, R. Y., Bickford, S. T., Ing, Y. L., Terman, B., Herlyn, M., and Lassam, N. J. (1994). Protein kinases in normal and transformed melanocytes. Melanoma Res. 4(5), 313–319. Rescigno, J., Mansukhani, A., and Basilico, C. (1991). A putative receptor tyrosine kinase with unique structural topology. Oncogene 6(10), 1909–1913. Robinson, D. R., Wu, Y. M., and Lin, S. F. (2000). The protein tyrosine kinase family of the human genome. Oncogene 19(49), 5548–5557. Rochlitz, C., Lohri, A., Bacchi, M., Schmidt, M., Nagel, S., Fopp, M., Fey, M. F., Herrmann, R., and Neubauer, A. (1999). Axl expression is associated with adverse prognosis and with expression of Bcl‐2 and CD34 in de novo acute myeloid leukemia (AML): Results from a multicenter trial of the Swiss Group for Clinical Cancer Research (SAKK). Leukemia 13(9), 1352–1358. Sainaghi, P. P., Castello, L., Bergamasco, L., Galletti, M., Bellosta, P., and Avanzi, G. C. (2005). Gas6 induces proliferation in prostate carcinoma cell lines expressing the Axl receptor. J. Cell Physiol. 204(1), 36–44. Sasaki, T., Knyazev, P. G., Cheburkin, Y., Gohring, W., Tisi, D., Ullrich, A., Timpl, R., and Hohenester, E. (2002). Crystal structure of a C‐terminal fragment of growth arrest‐specific protein Gas6. Receptor tyrosine kinase activation by laminin G‐like domains. J. Biol. Chem. 277(46), 44164–44170.

TAM Receptor Tyrosine Kinases

81

Sasaki, T., Knyazev, P. G., Clout, N. J., Cheburkin, Y., Gohring, W., Ullrich, A., Timpl, R., and Hohenester, E. (2006). Structural basis for Gas6‐Axl signalling. EMBO J. 25(1), 80–87. Sather, S., Kenyon, K. D., Lefkowitz, J. B., Liang, X., Varnum, B. C., Henson, P. M., and Graham, D. K. (2007). A soluble form of the Mer receptor tyrosine kinase inhibits macrophage clearance of apoptotic cells and platelet aggregation. Blood 109(3), 1026–1033. Schlessinger, J. (2000). Cell signaling by receptor tyrosine kinases. Cell 103(2), 211–225. Schulz, A. S., Schleithoff, L., Faust, M., Bartram, C. R., and Janssen, J. W. (1993). The genomic structure of the human UFO receptor. Oncogene 8(2), 509–513. Scott, R. S., McMahon, E. J., Pop, S. M., Reap, E. A., Caricchio, R., Cohen, P. L., Earp, H. S., and Matsushima, G. K. (2001). Phagocytosis and clearance of apoptotic cells is mediated by MER. Nature 411(6834), 207–211. Seitz, H. M., Camenisch, T. D., Lemke, G., Earp, H. S., and Matsushima, G. K. (2007). Macrophages and dendritic cells use different Axl/Mertk/Tyro3 receptors in clearance of apoptotic cells. J. Immunol. 178(9), 5635–5642. Sen, P., Wallet, M. A., Yi, Z., Huang, Y., Henderson, M., Mathews, C. E., Earp, H. S., Matsushima, G., Baldwin, A. S., Jr., and Tisch, R. M. (2007). Apoptotic cells induce Mer tyrosine kinase‐dependent blockade of NF‐kappaB activation in dendritic cells. Blood 109(2), 653–660. Shankar, S. L., O’Guin, K., Cammer, M., McMorris, F. A., Stitt, T. N., Basch, R. S., Varnum, B., and Shafit‐Zagardo, B. (2003). The growth arrest‐specific gene product Gas6 promotes the survival of human oligodendrocytes via a phosphatidylinositol 3‐kinase‐dependent pathway. J. Neurosci. 23(10), 4208–4218. Shankar, S. L., O’Guin, K., Kim, M., Varnum, B., Lemke, G., Brosnan, C. F., and Shafit‐ Zagardo, B. (2006). Gas6/Axl signaling activates the phosphatidylinositol 3‐kinase/Akt1 survival pathway to protect oligodendrocytes from tumor necrosis factor alpha‐induced apoptosis. J. Neurosci. 26(21), 5638–5648. Sharif, M. N., Sosic, D., Rothlin, C. V., Kelly, E., Lemke, G., Olson, E. N., and Ivashkiv, L. B. (2006). Twist mediates suppression of inflammation by type I IFNs and Axl. J. Exp. Med. 203(8), 1891–1901. Shieh, Y. S., Lai, C. Y., Kao, Y. R., Shiah, S. G., Chu, Y. W., Lee, H. S., and Wu, C. W. (2005). Expression of axl in lung adenocarcinoma and correlation with tumor progression. Neoplasia 7(12), 1058–1064. Stenflo, J., Lundwall, A., and Dahlback, B. (1987). Beta‐Hydroxyasparagine in domains homologous to the epidermal growth factor precursor in vitamin K‐dependent protein S. Proc. Natl. Acad. Sci. USA 84(2), 368–372. Stenhoff, J., Dahlback, B., and Hafizi, S. (2004). Vitamin K‐dependent Gas6 activates ERK kinase and stimulates growth of cardiac fibroblasts. Biochem. Biophys. Res. Commun. 319(3), 871–878. Stitt, T. N., Conn, G., Gore, M., Lai, C., Bruno, J., Radziejewski, C., Mattsson, K., Fisher, J., Gies, D. R., Jones, P. F., et al. (1995). The anticoagulation factor protein S and its relative, Gas6, are ligands for the Tyro 3/Axl family of receptor tyrosine kinases. Cell 80(4), 661–670. Sun, W., Fujimoto, J., and Tamaya, T. (2004). Coexpression of Gas6/Axl in human ovarian cancers. Oncology 66(6), 450–457. Sun, W. S., Fujimoto, J., and Tamaya, T. (2003). Coexpression of growth arrest‐specific gene 6 and receptor tyrosine kinases Axl and Sky in human uterine endometrial cancers. Ann. Oncol. 14(6), 898–906. Tada, A., Wada, Y., Sato, H., Itabashi, T., Kawamura, M., Tamai, M., and Nishida, K. (2006). Screening of the MERTK gene for mutations in Japanese patients with autosomal recessive retinitis pigmentosa. Mol. Vis. 12, 441–444.

82

Rachel M. A. Linger et al.

Tanaka, K., Nagayama, Y., Nakano, T., Takamura, N., Namba, H., Fukada, S., Kuma, K., Yamashita, S., and Niwa, M. (1998). Expression profile of receptor‐type protein tyrosine kinase genes in the human thyroid. Endocrinology 139(3), 852–858. Tavazoie, S. F., Alarcon, C., Oskarsson, T., Padua, D., Wang, Q., Bos, P. D., Gerald, W. L., and Massague, J. (2008). Endogenous human microRNAs that suppress breast cancer metastasis. Nature 451, 147–152. Taylor, I. C., Roy, S., and Varmus, H. E. (1995a). Overexpression of the Sky receptor tyrosine kinase at the cell surface or in the cytoplasm results in ligand‐independent activation. Oncogene 11(12), 2619–2626. Taylor, I. C., Roy, S., Yaswen, P., Stampfer, M. R., and Varmus, H. E. (1995b). Mouse mammary tumors express elevated levels of RNA encoding the murine homology of SKY, a putative receptor tyrosine kinase. J. Biol. Chem. 270(12), 6872–6880. Thompson, D. A., McHenry, C. L., Li, Y., Richards, J. E., Othman, M. I., Schwinger, E., Vollrath, D., Jacobson, S. G., and Gal, A. (2002). Retinal dystrophy due to paternal isodisomy for chromosome 1 or chromosome 2, with homoallelism for mutations in RPE65 or MERTK, respectively. Am. J. Hum. Genet. 70(1), 224–229. Todt, J. C., Hu, B., Punturieri, A., Sonstein, J., Polak, T., and Curtis, J. L. (2002). Activation of protein kinase C beta II by the stereo‐specific phosphatidylserine receptor is required for phagocytosis of apoptotic thymocytes by resident murine tissue macrophages. J. Biol. Chem. 277(39), 35906–35914. Todt, J. C., Hu, B., and Curtis, J. L. (2004). The receptor tyrosine kinase MerTK activates phospholipase C gamma2 during recognition of apoptotic thymocytes by murine macrophages. J. Leukoc. Biol. 75(4), 705–713. Toshima, J., Ohashi, K., Iwashita, S., and Mizuno, K. (1995). Autophosphorylation activity and association with Src family kinase of Sky receptor tyrosine kinase. Biochem. Biophys. Res. Commun. 209(2), 656–663. Tschernutter, M., Schlichtenbrede, F. C., Howe, S., Balaggan, K. S., Munro, P. M., Bainbridge, J. W., Thrasher, A. J., Smith, A. J., and Ali, R. R. (2005). Long‐term preservation of retinal function in the RCS rat model of retinitis pigmentosa following lentivirus‐mediated gene therapy. Gene Ther. 12(8), 694–701. Tschernutter, M., Jenkins, S. A., Waseem, N. H., Saihan, Z., Holder, G. E., Bird, A. C., Bhattacharya, S. S., Ali, R. R., and Webster, A. R. (2006). Clinical characterisation of a family with retinal dystrophy caused by mutation in the Mertk gene. Br. J. Ophthalmol. 90(6), 718–723. Tsou, A. P., Wu, K. M., Tsen, T. Y., Chi, C. W., Chiu, J. H., Lui, W. Y., Hu, C. P., Chang, C., Chou, C. K., and Tsai, S. F. (1998). Parallel hybridization analysis of multiple protein kinase genes: Identification of gene expression patterns characteristic of human hepatocellular carcinoma. Genomics 50(3), 331–340. Tsuchiya, H., Kanazawa, Y., Abdel‐Wanis, M. E., Asada, N., Abe, S., Isu, K., Sugita, T., and Tomita, K. (2002). Effect of timing of pulmonary metastases identification on prognosis of patients with osteosarcoma: The Japanese Musculoskeletal Oncology Group study. J. Clin. Oncol. 20(16), 3470–3477. Vajkoczy, P., Knyazev, P., Kunkel, A., Capelle, H. H., Behrndt, S., von Tengg‐Kobligk, H., Kiessling, F., Eichelsbacher, U., Essig, M., Read, T. A., Erber, R., and Ullrich, A. (2006). Dominant‐negative inhibition of the Axl receptor tyrosine kinase suppresses brain tumor cell growth and invasion and prolongs survival. Proc. Natl. Acad. Sci. USA 103(15), 5799–5804. Valverde, P., Obin, M. S., and Taylor, A. (2004). Role of Gas6/Axl signaling in lens epithelial cell proliferation and survival. Exp. Eye Res. 78(1), 27–37. Valverde, P. (2005). Effects of Gas6 and hydrogen peroxide in Axl ubiquitination and downregulation. Biochem. Biophys. Res. Commun. 333(1), 180–185.

TAM Receptor Tyrosine Kinases

83

van Ginkel, P. R., Gee, R. L., Shearer, R. L., Subramanian, L., Walker, T. M., Albert, D. M., Meisner, L. F., Varnum, B. C., and Polans, A. S. (2004). Expression of the receptor tyrosine kinase Axl promotes ocular melanoma cell survival. Cancer Res. 64(1), 128–134. Varnum, B. C., Young, C., Elliott, G., Garcia, A., Bartley, T. D., Fridell, Y. W., Hunt, R. W., Trail, G., Clogston, C., Toso, R. J., et al. (1995). Axl receptor tyrosine kinase stimulated by the vitamin K‐dependent protein encoded by growth‐arrest‐specific gene 6. Nature 373(6515), 623–626. Vollrath, D., Feng, W., Duncan, J. L., Yasumura, D., D’Cruz, P. M., Chappelow, A., Matthes, M. T., Kay, M. A., and LaVail, M. M. (2001). Correction of the retinal dystrophy phenotype of the RCS rat by viral gene transfer of Mertk. Proc. Natl. Acad. Sci. USA 98(22), 12584–12589. Walker, J. R., Huang, X., Finerty, P. J., Jr., Weigelt, J., Sundstrom, M., Arrowsmith, C. H., Edwards, A. M., Bochkarev, A., and Dhe‐Paganon, S. (2007). Catalytic Domain of the Proto‐ oncogene Tyrosine‐protein Kinase MER Retrieved May 30, 2007, fromhttp://www.rcsb.org/ pdb/explore.do?structureId=2P0C. Wang, H., Chen, S., Chen, Y., Wang, H., Wu, H., Tang, H., Xiong, W., Ma, J., Ge, Y., Lu, Q., and Han, D. (2007). The role of Tyro 3 subfamily receptors in the regulation of hemostasis and megakaryocytopoiesis. Haematologica 92(5), 643–650. Weier, H. U., Fung, J., and Lersch, R. A. (1999). Assignment of protooncogene MERTK (a.k.a. c‐mer) to human chromosome 2q14.1 by in situ hybridization. Cytogenet. Cell Genet. 84 (1–2), 91–92. Wimmel, A., Glitz, D., Kraus, A., Roeder, J., and Schuermann, M. (2001). Axl receptor tyrosine kinase expression in human lung cancer cell lines correlates with cellular adhesion. Eur. J. Cancer 37(17), 2264–2274. Wong, C. C., and Lee, W. M. (2002). The proximal cis‐acting elements Sp1, Sp3 and E2F regulate mouse mer gene transcription in Sertoli cells. Eur. J. Biochem. 269(15), 3789–3800. Wu, C. W., Li, A. F., Chi, C. W., Lai, C. H., Huang, C. L., Lo, S. S., Lui, W. Y., and Lin, W. C. (2002). Clinical significance of AXL kinase family in gastric cancer. Anticancer Res. 22(2B), 1071–1078. Wu, Y., Singh, S., Georgescu, M. M., and Birge, R. B. (2005). A role for Mer tyrosine kinase in alphavbeta5 integrin‐mediated phagocytosis of apoptotic cells. J. Cell Sci. 118(Pt 3), 539–553. Wu, Y. M., Robinson, D. R., and Kung, H. J. (2004). Signal pathways in up‐regulation of chemokines by tyrosine kinase MER/NYK in prostate cancer cells. Cancer Res. 64(20), 7311–7320. Yamagata, M., Sanes, J. R., and Weiner, J. A. (2003). Synaptic adhesion molecules. Curr. Opin. Cell Biol. 15(5), 621–632. Yeoh, E. J., Ross, M. E., Shurtleff, S. A., Williams, W. K., Patel, D., Mahfouz, R., Behm, F. G., Raimondi, S. C., Relling, M. V., Patel, A., Cheng, C., Campana, D., et al. (2002). Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression profiling. Cancer Cell 1, 133–143. Zantek, N. D., Walker‐Daniels, J., Stewart, J., Hansen, R. K., Robinson, D., Miao, H., Wang, B., Kung, H. J., Bissell, M. J., and Kinch, M. S. (2001). MCF‐10A‐NeoST: A new cell system for studying cell–ECM and cell–cell interactions in breast cancer. Clin. Cancer Res. 7(11), 3640–3648.

Epithelial Morphogenesis and Intestinal Cancer: New Insights in Signaling Mechanisms Silvia Fre,* Danijela Vignjevic,* Marie Schoumacher,* Shannon L. Duffy,*,{ Klaus‐Peter Janssen,z Sylvie Robine,* and Daniel Louvard* {

I. II. III. IV. V. VI. VII.

VIII. IX. X. XI. XII. XIII. XIV.

*UMR144 Curie/CNRS, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France; Leukaemia Foundation Research Unit, Queensland Institute of Medical Research, Herston, Qld 4029, Australia; z Department of Surgery, Klinikum rechts der Isar, TUM, 81675 Munich, Germany

The Intestinal Epithelium Wnt Signaling in the Intestine Notch Signaling in the Intestine Hedgehog Signaling in the Intestine BMP (Bone Morphogenetic Protein) Signaling in the Intestine PTEN (Phosphatase and Tensin Homologue) in the Intestine Receptor Tyrosine Kinases A. The Eph/Ephrin Signaling Pathway in the Intestine B. K‐Ras Signaling in the Intestine An Example of Integration Between Signaling Pathways: K‐Ras Promotes Wnt Signaling Metastasis Initial Step of Metastasis—Invasion Through the EMT or Collective Cell Migration? Two‐Phase Model for ‐Catenin Target Gene Activation ‐Catenin Target Genes at the Invasive Front EMT is a Reversible Process Concluding Remarks and Perspectives References

In this review, the major signal transduction pathways that have been shown to play an important role in intestinal homeostasis are highlighted. Each of them, the Wnt, Notch, Hedgehog, and Bone Morphogenetic Protein, as well as growth‐factor regulated Receptor Tyrosine Kinases are depicted with a special emphasis through their involvement in stem cell maintenance and their role in intestinal tumorigenesis. Finally, we discuss recent data on the final steps of tumor progression, notably the formation of distant metastases. This multistep process is highly complex and still far from being understood while being of major importance for the survival of patients with digestive cancer. # 2008 Elsevier Inc.

Advances in CANCER RESEARCH Copyright 2008, Elsevier Inc. All rights reserved.

85

0065-230X/08 $35.00 DOI: 10.1016/S0065-230X(08)00003-1

Silvia Fre et al.

86

I. THE INTESTINAL EPITHELIUM The intestinal epithelium is a paradigm of rapid tissue renewal in which undifferentiated cells are spatially separated from the differentiated compartment. The lining of the mammalian gastrointestinal tract is composed of a highly specialized monostratified epithelium with finger‐like projections, called villi, present in the lumen of the small intestine that penetrate the stroma to form deep invaginations known as crypts of Lieberku¨hn (henceforth referred to as crypts). Undifferentiated cells predominate in the crypts and receive proliferative signals from the adjacent mesenchyme. The majority of crypt cells belong to the transit amplifying compartment, which is characterized by uncommitted, actively proliferating cells derived from a small number of stem cells that are believed to be localized towards the bottom of the crypt, just above the Paneth cells. Paneth cells are found exclusively in the small intestine and are the only differentiated cell type to reside in the crypt compartment. The other three cell types present in the gut: enterocytes, enteroendocrine, and Goblet cells, differentiate upon reaching the crypt–villus border (Fig. 1). A tightly controlled network of interactions exists between the epithelial and the surrounding mesenchymal cells, and the lack of contact with diffusible determinants present in the crypt niche allows epithelial differentiating cells to enter cell cycle arrest and to commit toward a specific cell fate, as they undertake apical migration along the villus. Villi are composed mostly of epithelial columnar enterocytes, dedicated to absorptive functions, whereas the remaining three cell types of the intestine are all secretory. Signaling pathways that participate in the control and modeling of the intestine are tightly regulated in order to ensure precise patterns of cell distribution as well as the rapid turnover observed in the intestinal epithelium. A complex interplay between diverse cellular responses such as proliferation, differentiation, migration, apoptosis, and adhesion exists and is coordinated by a relatively small number of evolutionary conserved signal transduction cascades. Interestingly, the same signaling pathways that are important in broader terms for the development of the organism as a whole are also used to maintain intestinal homeostasis. It is conceivable that these essential signal transduction cascades are also involved in the maintenance of the stem cell population necessary to ensure tissue renewal as well as regeneration upon injury. Fine tuning of these signals is crucial for the maintenance of normal tissue homeostasis, and deregulation of the mechanisms that maintain the balance between proliferation, apoptosis, and differentiation leads to neoplastic pathologies. The intestinal epithelium represents one of the best systems to examine one of the central problems in biology: how cells integrate multiple inputs from different signaling

A

B

Lactase

Alcian blue

Enteroendocrine cell Differentiation

Goblet cell

ki67

Enterocytes

Enterocyte

Goblet cells

Proliferative cells

Proliferation Enteroendocrine cells

Chromogranin Paneth cell Mesenchymal cells Basal membrane

Paneth cells

Lysozyme

88

Silvia Fre et al.

pathways to produce a coordinated response from the genome dictating specific developmental programs. In this chapter, we concisely describe the major signaling pathways that have been shown to play an important role in intestinal homeostasis and address their involvement in stem cell maintenance and in intestinal tumorigenesis.

II. WNT SIGNALING IN THE INTESTINE The current evidence indicates that the Wnt cascade is essential in establishing cell fate along the crypt–villus axis of the intestinal epithelium (Reya and Clevers, 2005). ‐Catenin is the major cytoplasmic signal transducer of the canonical Wnt pathway (Nusse and Varmus, 1992). Accumulation of nuclear ‐catenin, the hallmark of active Wnt signaling, is evident in the crypt cells of the normal intestine, whereas differentiated villus cells present ‐catenin at their basolateral membrane, where it is important to ensure cell adhesion (van de Wetering et al., 2002). Deletion of the transcription factor T cf4, the most prominent effector of Wnt signaling in the gastrointestinal tract, produces a severe intestinal phenotype accompanied by neonatal lethality (Korinek et al., 1998). While the villous epithelial compartment is practically unaffected in these mice, intestinal crypts are completely absent and no proliferating cells are observed, indicating that Wnt signals are required for the maintenance of the crypt proliferative compartment. In agreement with these findings, inhibition of the Wnt receptor by expression of the potent secreted Wnt antagonist Dickkopf‐1 in adult mice leads to a complete loss of crypts (Kuhnert et al., 2004; Pinto et al., 2003). Thus, in physiological conditions, the ‐catenin / T cf4 complex drives the transcription of a set of target genes that determine the characteristics of the intestinal crypt cells. Wnt signals are turned off in differentiated cells present in the villi, hence establishing the crypt–villus boundary. The only terminally Fig. 1 Distribution of epithelial cell types in the mammalian small intestine. (A) Schematic view of the crypt–villus unit, depicting the position of the different intestinal cell types. The dotted line marks the border that separates the proliferative region (crypt) from the differentiated part (villus) of the intestinal epithelium. This diagram was adapted from Rizvi, A. Z. et al., (2005, Fig. 2A, p. 153). (B) Immunohistological stainings showing: Proliferative cells located exclusively in the crypts and identified by the cell cycle marker Ki67 in pink; enterocytes identified by the apical concentration of lactase in white; Goblet cells highlighted by an Alcian Blue staining in dark blue; enteroendocrine cells scattered throughout the villus and stained by an anti‐chromogranin antibody in red; and Paneth cells located at the bottom of the crypts and marked by an anti‐lysozyme antibody in red. All nuclei are stained in blue with the DAPI DNA marker in the fluorescent images (Ki67, enteroendocrine and Paneth cells) and are counterstained with hematoxylin in the Alcian blue picture.

Epithelial Morphogenesis and Intestinal Cancer

89

differentiated cells in the intestine that present nuclear ‐catenin and active Wnt signaling are the Paneth cells, and this specific feature might account for their downward migration towards the bottom of the crypts, contrary to all the other differentiated cell types of the gut. The same genetic program that is physiologically active in crypt progenitor cells appears to be driving colon carcinogenesis. Indeed, loss of the tumor suppressor Apc (adenomatous polyposis coli), a key negative regulator of Wnt signals that functions by controlling ‐catenin degradation, is the signature of the great majority of human intestinal tumors, both in hereditary syndromes and in sporadic colorectal cancers (Kinzler and Vogelstein, 1998). In the rare cases where Apc is not inactivated, human intestinal tumors have been found to arise from activating mutations in ‐catenin itself (Morin et al., 1997), or from loss of function mutations in Axin 2, a protein that participates with Apc in ‐catenin degradation (Liu et al., 2000). These data underscore the central role of Wnt signaling activation in the transformation of intestinal epithelial cells. Wnt signaling exerts its effects both on progenitor/stem cells of the intestinal crypts in physiological conditions and on cell transformation in the context of intestinal tumors. This leads us to suggest that the canonical Wnt pathway can influence self‐renewal of gut stem cells as well as tumor initiating cells, also referred to as cancer stem cells. Indeed, the Wnt cascade has been implicated in controlling the fate of stem cells in other tissues, such as the skin and the hematopoietic system (Clevers, 2006), and if, on one side, stem cells ensure tissue renewal and repair, on the other side, deregulation of their proliferative potential can lead to dramatic phenotypes. In this regard it is relevant to mention that it has been suggested that the initial events leading to tumorigenesis would occur only in cells that persist long enough to accumulate multiple oncogenic events, and therefore adult stem cells, in any given tissue, would be the best candidates for representing the tumor initiating cells. The concept of cancer stem cells is particularly intuitive in a rapidly regenerating tissue like the intestine, where the majority of the cells do not survive more than a week. The presumptive tumor initiating cells should have some of the functional properties of a stem cell: they should retain an undifferentiated phenotype, be capable of asymmetric division and self‐renewal, and they should represent only a minor fraction of the bulk of more differentiated cells in the tumor.

III. NOTCH SIGNALING IN THE INTESTINE Notch signaling controls pattern formation and cell fate decisions in all eukaryotes (Artavanis‐Tsakonas et al., 1999). Multiple Notch pathway components are expressed in the mouse intestine (Sander and Powell, 2004;

90

Silvia Fre et al.

Schroder and Gossler, 2002). In general, it appears that the Notch ligands are mostly expressed in the mesenchymal cells surrounding the intestinal tube, whereas Notch receptors, as well as some downstream target genes, are almost exclusively found in the gut epithelial layer. The Notch pathway was initially implicated in the regulation of cell fate decisions in the intestine through the analysis of mice deficient for the basic helix–loop– helix proteins Hes‐1, Math‐1, and neurogenin‐3, all of which are transcriptional targets of Notch signaling (Jenny et al., 2002; Jensen et al., 2000; Yang et al., 2001). Hes‐1 expression in the mouse intestine is restricted to the crypt compartment, and analysis of the developing fetal intestine of Hes‐1/ mutant mice revealed a relative increase in mucosecreting and enteroendocrine cells at the expense of absorptive enterocytes (Jensen et al., 2000). Notch signaling appears to be inactive in precursors of the secretory cell type, allowing for the expression of the proendocrine bHLH proteins Math‐1 and neurogenin 3 (ngn3). Math‐1, a target gene of Hes‐1‐mediated repression, seems to be the first factor involved in secretory cell lineage differentiation, committing cells to become mucosecreting, Paneth, and enteroendocrine. In fact, the intestines of Math‐1‐deficient mice show a relatively normal crypt–villus architecture that is populated entirely by enterocytes, indicating that Math‐1 is essential for the acquisition of the secretory cell fate (Yang et al., 2001). This phenotype is somewhat reciprocal to the phenotype observed in Hes‐1/ mice, and these results indicate that Hes‐1 and Math‐1 are able to skew the fate of differentiating cells exiting from the transit amplifying compartment toward either an enterocyte or a secretory phenotype. Taken together, these results suggested that Notch signals regulate a binary decision between adsorptive and secretory cell fates. Other bHLH proteins appear to further refine these fate decisions. For example, mice deficient in the bHLH factor neurogenin‐3 specifically lack enteroendocrine precursors (Jenny et al., 2002). The demonstration of a direct role for Notch signals in controlling the segregation of each mature lineage from undifferentiated progenitor cells, as well as in the maintenance of the proliferating intestinal cell pool, came from the analysis of mice carrying either a loss of function (van Es et al., 2005) or a gain of function (Fre et al., 2005) Notch pathway mutation. When Notch signaling is inhibited by deletion of the major downstream effector RBPJ (Su(H) in Drosophila), all crypt cells cease to proliferate and differentiate into Goblet cells. Reciprocally, expression of a constitutively active form of the Notch receptor in the developing intestinal epithelium markedly impairs cell differentiation and increases the proportion of dividing cells, which extend outside of the crypt proliferative compartment, and are found all along the vertical axis of the villi. Interestingly, the involvement of Notch signals in cell fate determination in the intestine seems to be conserved in all eukazyotes: in zebrafish, inhibition

Epithelial Morphogenesis and Intestinal Cancer

91

of Notch signals achieved by two different genetic lesions (the DeltaD and mindbomb mutants) leads to an overproduction of secretory cells at the cost of absorptive cells, accompanied by an increase in apoptotic cells, demonstrating a strong evolutionary conservation of the role of Notch signaling in the intestine (Crosnier et al., 2005). In addition, an important function for the Notch pathway in cell fate specification and control of proliferation in the invertebrate gut of Drosophila melanogaster has been recently shown (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006), further underscoring the essential role of Notch signaling in this organ. Although the oncogenic potential of Notch signals has been documented in a variety of experimental systems and in several tissues (Bellavia et al., 2000; Capobianco et al., 1997; Gallahan and Callahan, 1997; Gallahan et al., 1996; Kiaris et al., 2004; Pear et al., 1996; Rohn et al., 1996; Soriano et al., 2000), a direct role for Notch in intestinal tumorigenesis has not been clearly established. A connection between Notch and intestinal tumors arises from studies showing that Hes‐1, a signature of Notch signal activation, is overexpressed in mouse adenomas (van Es et al., 2005). In addition, ‐secretase inhibitors able to downregulate the Notch cascade by blocking cleavage and activation of the receptor can reduce proliferation and induce Goblet cells differentiation (van Es et al., 2005) in adenomas that occur spontaneously in Multiple intestinal neoplasia (Min) mice (Moser et al., 1990; Su et al., 1992; Yamada et al., 2002), suggesting that the maintenance of the proliferative potential of adenoma cells depends on Notch activity. The Notch signaling pathway has been shown to positively influence the maintenance and self‐renewal of adult stem cells in several different tissues (Luo et al., 2005; Molofsky et al., 2004; Suzuki and Chiba, 2005; Yoon and Gaiano, 2005). As already discussed for the Wnt cascade, the notion that Notch signals are required to maintain proliferation in the intestinal crypts raises the intriguing possibility that Notch might control the developmental outcome of gut stem cells, known to reside in this region.

IV. HEDGEHOG SIGNALING IN THE INTESTINE Hedgehog signaling plays an important role in the morphogenesis of the intestine, where the ligands Sonic Hedgehog (Shh) and Indian Hedgehog (Ihh) are localized to the gut epithelial cells and communicate in a paracrine fashion with mesenchymal cells expressing the Patched‐1 and ‐2 receptors, as well as the effectors of Hedgehog signaling, Gli1, Gli2, and Gli3 (Ramalho‐Santos et al., 2000) (Fig. 2). Blocking Hedgehog signals through the pan inhibitor Hhip (Hedgehog interacting protein) markedly compromises the crypt–villus

Silvia Fre et al.

92

Ihh

Ptch Gli Ihh BMP4 BMPR1A N N

P-PTEN

BMPR1A BMP4

Wnt Wnt

Wnt

BMP4

Fig. 2 Signaling pathways in the small intestine. Diagram representing the different signaling cascades implicated in intestinal homeostasis. The expression pattern of some of the pathway components is schematically shown. Wnt ligands, expressed in crypt mesenchymal cells, send proliferative signals to the crypt cells. The Notch pathway is activated in undifferentiated crypt cells and is turned off at the crypt–villus border. Ihh, expressed in epithelial cells of the villus, communicates with the underlying stromal cells expressing Patched (Ptch) and the Hh effectors Gli; Ihh also contributes to BMP activation in the mesenchyme (dashed red arrow). BMP4 is expressed in mesenchymal cells both in the villus and around the crypt and it interacts with BMPR1A receptors expressed in epithelial cells. Crypt progenitor/stem cells show high levels of the inactive, phosphorylated form of PTEN (p‐PTEN). This drawing was modified from Rizvi, A. Z. et al., (2005, Fig. 2A, p. 153).

architecture and leads to a complete absence of villi in an often pseudostratified and hyperproliferating epithelium (Madison et al., 2005). A milder phenotype is observed when Hh signaling is suppressed by the injection of an anti‐Hh antibody, and is characterized by the presence of ectopic proliferation and abnormally branched villi (van den Brink et al., 2004). The general conclusion is that Hedgehog signals from the epithelium to the mesenchyme are required for epithelial remodeling as well as for restraining proliferation to the intervillus regions, where the proteins Shh and Ihh become concentrated as development proceeds. This process seems to involve a two‐way communication from the mesenchymal cells that signal back to the epithelium through some other signaling pathways (van den Brink, 2004).

Epithelial Morphogenesis and Intestinal Cancer

93

The involvement of Hedgehog in pathological conditions of the gastrointestinal tract, most notably colorectal carcinogenesis, is indirect and appears to be mediated by Wnt signals. Indeed, in mice expressing the Hhip inhibitor, Wnt target genes that are normally restricted to the crypt compartment, such as Cdx1, Cd44, and EphB2, demonstrate ectopic expression throughout the villus epithelium and may be responsible for the reported hyperproliferation (Madison et al., 2005). These results indicate that Hh signals negatively regulate Wnt signaling under normal conditions. Indeed, in human adenomas of familial adenomatous polyposis (FAP) patients, expression of Ihh is lost in the neoplastic tissue, where Wnt signaling is constitutively active due to a mutation of Apc (van den Brink, 2004). In conclusion, the role of Hh in restricting Wnt signaling to the crypts seems to be lost by a feedback loop mechanism in neoplastic conditions that leads to ectopic Wnt activation. The finding that the constitutive knock‐out of Ihh leads to perinatal lethality with depletion of the proliferative compartment suggests that Ihh might be important for the maintenance of intestinal stem cells (Ramalho‐Santos et al., 2000). Notwithstanding the clear function of Hh in controlling stem cell number in the Drosophila ovary (Zhang and Kalderon, 2001), suggestive of a parallel role in mammalian tissues, a direct role for the Hh pathway in affecting the stem cell compartment of the intestine is yet to be established.

V. BMP (BONE MORPHOGENETIC PROTEIN) SIGNALING IN THE INTESTINE Bone morphogenetic proteins (BMPs) are members of the transforming growth factor‐ (TGF‐ ) superfamily, which use BMP receptor and SMAD transcription factors to transduce signals that dictate cell differentiation, proliferation, and apoptosis (Itoh et al., 2000; Shi and Massague, 2003). The BMP pathway has been shown to have a key role during intestinal development as well as in homeostasis of the adult gastrointestinal tract (Haramis et al., 2004). BMP4 is expressed primarily in the intestinal mesenchyme, both in the villus compartment, where it appears to signal in a paracrine fashion to the adjacent differentiated epithelial cells of the villus, and in the intercrypt mesenchymal cells, including those surrounding the intestinal stem cells (He et al., 2004) (Fig. 2). The findings that transgenic mice expressing the BMP inhibitor Noggin form ectopic crypt–villus units perpendicular to the original axis have provided evidence that BMP signals are involved in shaping the intestine and the intestinal stem cell niche (Haramis et al., 2004). BMPs and their receptors have been linked to the pathogenesis of certain solid tumors, including those of the intestine, and BMP2 is able to inhibit the growth of normal colonic epithelial cells by promoting apoptosis and

94

Silvia Fre et al.

differentiation (Hardwick et al., 2004). In addition, mutations in several BMP pathway genes have been found in the hereditary juvenile polyposis (JP) syndrome, a condition that markedly increases the risk for developing colon cancer in affected individuals (Howe et al., 1998). This disorder is characterized by the presence of multiple polyps and the formation of ectopic proliferating crypts along the villi. Blocking BMP signals in mice either by inactivation of the Bmpr1a receptor or by overexpression of the BMP antagonist Noggin result in phenotypes mimicking human JP syndrome (Haramis et al., 2004). BMP signaling also appears to play an important role in controlling stem cell numbers in the intestinal and hematopoietic stem cell compartments (He et al., 2005). In the intestine, expression of BMP4 is found in mesenchymal cells surrounding the crypts (the presumed stem cell niche), whereas the BMP receptor 1A (BMPR1A) is expressed in the presumptive intestinal stem cells themselves, defined as BrdU label retaining crypt cells, which also show expression of phosphorylated SMADs, a hallmark of active BMP signaling (He et al., 2004).

VI. PTEN (PHOSPHATASE AND TENSIN HOMOLOGUE) IN THE INTESTINE The tumor suppressor gene phosphatase and tensin homologue (PTEN) is a negative regulator of the serine–threonine kinase Akt and inhibits signals downstream of Akt that promote cell proliferation (Stiles et al., 2004). PTEN has been shown to be implicated in intestinal homeostasis in a more indirect way than in the signaling pathways discussed earlier. Indeed, PTEN appears to be able to regulate Wnt signaling by specifically inhibiting ‐catenin activity through downregulation of Akt (Persad et al., 2001). PTEN is a paradigm for the importance of understanding the interplay between different pathways in the control of normal tissue homeostasis as well as in affecting their impact on pathological conditions. In fact, BMP signals have been shown to enhance PTEN activity (Waite and Eng, 2003). In the intestine of mice carrying a deletion of the BMP receptor Bmpr1, the levels of phospho‐PTEN and, consequently, of phosho‐Akt are higher than that in control mice, suggesting that in the absence of BMP signaling, PTEN is inhibited and this relieves its negative regulation on Akt, thereby resulting in the accumulation of nuclear ‐catenin. In conclusion, it appears that BMP negatively affects Wnt signaling by enhancing the inhibitory activity of PTEN (He et al., 2004). The conditional inactivation of PTEN in mouse intestinal epithelium leads to the formation of numerous polyps throughout the intestinal tract

Epithelial Morphogenesis and Intestinal Cancer

95

(He et al., 2007). A similar phenotype is observed in human patients carrying PTEN mutations responsible for Cowden disease, a pathological polyposis condition also characterized by an excess number of Goblet and Paneth cells (Liaw et al., 1997). Importantly, expression of the phosphorylated, inactive form of PTEN has been observed specifically and exclusively in intestinal stem cells, identified as long‐term BrdU‐retaining crypt cells (Fig. 2). These cells show high levels of active Akt, as well as nuclear accumulation of ‐catenin, consistent with the notion that PTEN normally inhibits Akt and that this results in a negative regulation of Wnt signals. In PTEN knock‐out mice, a consistent increase in the number of proliferating cells in the intestinal crypts has been observed. These data demonstrate that PTEN is required as a negative cell cycle regulator in the transit amplifying compartment of intestinal crypts (He et al., 2007).

VII. RECEPTOR TYROSINE KINASES A. The Eph/Ephrin Signaling Pathway in the Intestine Signals transmitted by RTKs have multiple, varied effects on assorted cellular processes, including proliferation, differentiation, and cell movement. Pathological expression of RTKs in the intestine, therefore, has the potential to confer significant survival advantages to cells through effects on a large number of signaling mechanisms controlling cell behavior and growth; this complexity has contributed to the difficulty in defining the precise role of RTKs in intestinal tumorigenesis. This review describes two examples of RTK signaling in the intestine, in order to illustrate the broader principles that are characteristic of and unique to the RTK signaling cascades/family: the Eph/ephrin and K‐Ras signaling pathways. The Eph receptors comprise the largest subfamily of RTKs, and interaction with their membrane‐bound ephrin ligands initiates signaling pathways that regulate cell adhesion and migration during normal development and in cancer (Coulthard et al., 2002). Eph and ephrin proteins are uniquely distinguished from other RTKs by their ability to transduce intracellular signals bidirectionally in interacting cells and signaling pathways affecting the cellular cytoskeleton are a major target of activated Eph receptors. The molecular basis for these responses lies in the regulation of the Rho and Ras families of small GTPases. Extensive expression of Ephs and ephrins, including expression of EphA1, EphA2, EphA3, EphA7, EphB2, EphB3, EphB4, ephrin‐A1, and ephrin‐B1, is found during normal physiological and pathological processes of the

96

Silvia Fre et al.

intestine (Hirai et al., 1987; Kataoka et al., 2004; Kiyokawa et al., 1994; Stephenson et al., 2001). The EphB2/B3 receptors and their ligands ephrin‐ B1 and ‐B2 are expressed in mutually exclusive regions of the intestinal mucosa, with the receptors found at the bottom of the crypts, in a complementary pattern to the ligands, which are distributed in a gradient peaking at the crypt–villus junction (Batlle et al., 2002). Eph/ephrin signals are required for the normal migration and maturation of enterocytes in embryonic and neonatal mice and for sorting of differentiated cells in the adult intestine (Batlle et al., 2002). In an example of crosstalk between the Wnt and the Eph signaling pathways, expression of EphB2/B3 and ephrin‐B1 is inversely controlled by ‐catenin and TCF, and functions to restrict intermingling of the proliferative and differentiative cells migrating along the crypt–villus axis (Batlle et al., 2002; Kiyokawa et al., 1994; Mariadason et al., 2001). Eph‐ephrin signals, in concert with ‐catenin/TCF, may supply the molecular mechanism that links the migratory behavior of cells to their proliferation and differentiation programs. While expression of EphB2 and EphB3 is required for normal epithelial organization of the intestine, absence of these signals does not lead to aberrations in adult intestinal architecture (Batlle et al., 2002), suggesting that other gastrointestinally‐expressed Eph family members also play important roles in intestinal homeostasis. In addition to their functions in normal development, Ephs and ephrins are implicated in tumorigenesis, where they may play roles in tumor invasion and metastasis. Eph/ephrin signals are known to target cell– extracellular matrix interactions, and this is in part effected by the modulation of integrin function (Zou et al., 1999). Adhesion of cellular extensions to components of the extracellular matrix, for example, involves interactions between the Eph, integrin, and ERK/MAPK signaling pathways. Activated H‐Ras initiates a signaling cascade resulting in phosphorylation and activation of MAPK, including Erk1 and 2, and serine/threonine kinases such as Raf1 and Mek1 (Chang and Karin, 2001; Johnson and Lapadat, 2002). Both the EphA and EphB receptor subclasses are capable of downregulating this pathway in most cell types, including embryonic fibroblasts and transformed epithelial cell lines (Miao et al., 2001; Nagashima et al., 2002). In addition to directly suppressing signaling through the Ras–MAPK pathway, Ephs are also able to downregulate Ras–MAPK signaling downstream of other receptors such as integrins, vascular endothelial growth factor (VEGF), and epidermal growth factor (EGF), implying further cross‐communication between activated Eph receptors and other RTKs (Grunwald et al., 2001; Kim et al., 2002; Miao et al., 2001). A role for Ephs and ephrins in blocking tumor progression has recently been suggested in intestinal neoplasia, where it was found that expression of the EphB2 and EphB3 receptors is strongly reduced in most colorectal cancer cell lines as well as in human colon carcinomas (Batlle et al., 2005). This finding is

Epithelial Morphogenesis and Intestinal Cancer

97

in apparent contradiction with the notion that EphB2 is a Wnt target gene (van de Wetering et al., 2002), since adenocarcinomas showing prominent nuclear ‐catenin accumulation exhibit decreased levels of EphB2, and suggests the existence of silencing mechanisms able to downregulate EphB expression during cancer progression (Clevers and Batlle, 2006). Moreover, tumor progression in Apcmin mice was found to be strongly accelerated when these mice were crossed with transgenic animals expressing a dominant negative form of EphB2 or with EphB3 null mice (Batlle et al., 2005). These findings provide evidence for a causal role of Eph silencing in colorectal cancer progression. Evidence of a role for Ephs and ephrins in stem cell biology comes from studies on asymmetric cell division and cell fate specification in the ascidian embryo (Picco et al., 2007). Convergence of the Eph/ephrin and the Ras/ ERK signaling pathways is also apparent in this system, in which Eph/ ephrin‐mediated differential regulation of Ras/ERK signaling is proposed to control the fate specification of daughter cells during asymmetric cell division. In the mouse intestine, a relationship between the Eph/ephrin system and stem cells can be inferred from studies on the expression of EphB2 and EphB3 receptors and on the phenotype displayed by deletion of these genes in knock‐out mice (Batlle et al., 2002). In the small intestine, EphB3 is found in cells at the bottom of the crypts, which include the Paneth as well as progenitor cells, and EphB2 is expressed by all proliferative crypt cells. In the colon, expression of both receptors is restricted to progenitor/ stem cells present at the crypt base. EphB3/ mice feature anomalous localization of Paneth cells that are found scattered along the crypt and villus, and loose their specific localization at the crypt base, where they are presumed to play a role in protecting the intestinal stem cell population (Muller et al., 2005). In double deficient EphB2/EphB3 mice, progenitor cells seem to randomly intermingle with differentiated cells (Batlle et al., 2002). These results are indicative of a role for Eph/ephrin signaling in maintaining the stem cell niche or in correctly positioning and sorting of crypt cells or both, but a clear demonstration for how essential this function might be in intestinal stem cell maintenance is still missing.

B. K‐Ras Signaling in the Intestine The three mammalian ras genes, K‐, N‐, and H‐ras, encode proteins that are members of the guanine nucleotide binding protein superfamily (Malumbres and Barbacid, 2003). Ras proteins are situated at the inner face of the plasma membrane, where they act as molecular switches during normal cellular signaling. They are activated by diverse extracellular signals,

98

Silvia Fre et al.

for example, by growth factors such as EGF, and platelet‐derived growth factor (PDGF). Oncogenic Ras proteins are locked in the active, GTP‐bound state and are capable of constitutively activating their downstream effectors (Trahey and McCormick, 1987). Among the best studied Ras effectors are the serine–threonine kinases of the Raf family and their downstream target, the mitogen‐activated protein kinase (MAPK); other important effectors are the PI3‐kinases (Khosravi‐Far et al., 1998; McCormick, 1999). The evolutionarily conserved MAPK cascade consists of the serine/threonine kinases of the Raf family, which phosphorylate and hence activate MEK1 and MEK2, which in turn phosphorylate ERK1 and ERK2. Activated ERK1/2 translocate to the nucleus, where they play a major role in the mitogenic action of oncogenic Ras. In general, activation of the MAPK cascade is thought to result in phosphorylation and activation of transcription factors such as c‐Jun, c‐Myc and c‐Fos, causing enhanced transcription of genes that are associated with cell proliferation and survival (Hancock, 2003; Khosravi‐Far et al., 1998; Malumbres and Barbacid, 2003). The important role or Ras signaling in intestinal tumorigenesis is well documented. Activating mutations of K‐Ras are found in one‐third of all human cancers (Bos, 1989). The frequency of K‐Ras mutations in human colorectal tumors varies between 38% and 50% (Forrester et al., 1987; Vogelstein et al., 1988; Walker et al., 1999). Moreover, tumor cells depend on oncogenic K‐Ras even after transformation has taken place. When the expression of oncogenic K‐Ras is inhibited by small interfering RNAs, human tumor cells loose their oncogenic potential (Brummelkamp et al., 2002). Taken together, several lines of evidence indicate that, even though there may be considerable functional overlap between the Ras isoforms, K‐ras has a unique, distinct and essential role in tumor formation. Activating mutations in K‐Ras are more frequently detected in human large adenomas and carcinomas (Rajagopalan et al., 2002; Vogelstein et al., 1988), implying a role for K‐Ras in tumor progression rather than in initiation. However, mouse models designed to address the role of oncogenic K‐Ras in intestinal tumor formation have demonstrated that expression of a constitutively activated form of K‐Ras is sufficient for tumorigenesis in this species (Janssen, 2003; Janssen et al., 2002). Notwithstanding the general notion that Ras/MAPK signals promote cell proliferation and survival in a variety of tissues and experimental organisms, very little data have been gathered about the physiological role of this pathway in cell renewal of the intestinal epithelium. In accordance with a possible role in promoting intestinal cell proliferation, localization of the active, phosphorylated form of MAPK is restricted to the nuclei of proliferative undifferentiated precursors in the intestinal crypts (Aliaga et al., 1999). In addition, the EGF receptor (EGFR), an upstream activator of the Ras/MAPK cascade, plays a crucial role in wound repair processes in the intestinal epithelium. It has

Epithelial Morphogenesis and Intestinal Cancer

99

recently been shown that EGFR signaling induce the expression of p21waf1/cip1 via an activation of the MAPK cascade, leading to increased epithelial cell proliferation (Sheng et al., 2006). The role of EGFR signaling in the broader context of intercellular communication and cell fate decisions has been reviewed elsewhere (Shilo, 2005). It is noteworthy, however, that none of the transgenic mice expressing activated forms of K‐Ras in the intestinal epithelium display major abnormalities in cell renewal or differentiation.

VIII. AN EXAMPLE OF INTEGRATION BETWEEN SIGNALING PATHWAYS: K‐RAS PROMOTES WNT SIGNALING Understanding how different signaling cascades crosstalk and integrate in order to ensure the correct positioning of cells along the crypt–villus axis and the constant flow of cell renewal in the intestine has become the real challenge in deepening our knowledge of the molecular mechanisms controlling gut homeostasis. In addition, it has long been established that intestinal tumorigenesis progresses through the sequential acquisition of several genetic alterations, which include the loss of tumor suppressor genes and the activation of oncogenes (Hanahan and Weinberg, 2000; McCormick, 1999). These mutations lead a given epithelial cell to loose cell cycle control and hyperproliferate, then to form a dysplastic crypt, which may become a polyp and can eventually progress to the more aggressive stages of adenocarcinoma in situ and invasive carcinoma (Fearon and Vogelstein, 1990). Along the adenoma– carcinoma sequence, the loss of Apc function is usually followed by the oncogenic activation of K‐Ras. Moreover, the combination of K‐Ras activation and nuclear ‐catenin accumulation at the invasive front of a tumor is predictive of poor prognosis among patients with colorectal cancer (Zhang et al., 2003). The analysis of the dependence of intestinal tumors on combinatorial signaling has already proven useful in elucidating some of the synergies acting on cancer development and progression. One example of such studies is represented by the analysis of compound transgenic mice mimicking the combination of mutations in the Apc tumor suppressor and in the oncogene K‐Ras (Janssen et al., 2006). Transgenic mice expressing the constitutively active human K‐Ras oncogene (K‐RasV12G) under the control of the villin promoter (Janssen et al., 2002) were crossed with the Apcþ/1638N mouse model, which carries a targeted mutant allele at the endogenous Apc locus (Fodde et al., 1994).

100

Silvia Fre et al.

Double transgenic animals show a 10‐fold increase in tumor multiplicity and accelerated tumor progression, accompanied by a strongly enhanced mortality when compared with single transgenic littermates. Tumors from these mice proliferate faster and show lower levels of apoptosis than do tumors derived from mice harboring either an Apc or a K‐Ras mutation alone. Several lines of evidence indicate that the observed increase in tumor multiplicity and malignant transformation is caused by the synergistic activation of Wnt/ ‐catenin signaling in cells harboring gain of function K‐Ras and loss‐of‐function Apc mutations. Accordingly, intestinal tumors from the compound mutant mice show a significant increase in cells presenting nuclear accumulation of ‐catenin, accompanied by a reduction in its membranous staining when compared to Apcþ/1638N animals. In addition, transcriptional reporter assays and immunofluorescence analysis indicated that cells with both oncogenic K‐Ras and loss‐of‐function Apc mutations are characterized by enhanced Wnt signaling. This phenomenon is likely to underlie the observed striking increase in tumor onset and malignant transformation. Activated K‐Ras is known to induce the phosphorylation of ‐catenin on tyrosine residues, leading to its release from E‐cadherin at the adherens junction and to increased Wnt signaling (Kinch et al., 1995). Thus, it is likely that mutations in other genes of the Ras/ MAPK pathway, such as B‐Raf (Davies et al., 2002; Rajagopalan et al., 2002), which are frequently found in sporadic colorectal cancers, also enhance Wnt signaling through ‐catenin tyrosine phosphorylation. However, the cross‐talk between the Ras/MAPK and the Wnt signaling pathways may not be limited to increased nuclear availability of ‐catenin. Of note, the Ras/MAPK pathway has been shown to attenuate the function of the transcriptional corepressor Groucho (reviewed in Hasson and Paroush, 2006). Groucho represents a nodal point of integration among several signaling networks and has been shown to associate with repressors downstream of both the Wnt and the Notch pathways (Brantjes et al., 2001; Hasson et al., 2005). Attenuation of Groucho by the EGFR/Ras/MAPK cascade is thought to lead to the derepression of downstream target genes, leading to cell fate changes in a physiological setting or to the activation of genes relevant for tumorigenesis. In conclusion, the K‐RasV12G/Apcþ/1638N mouse model recapitulates some of the steps of tumor progression and is useful for the characterization of the molecular events underlying the synergy between these two signaling pathways (Fodde and Brabletz, 2007; Janssen et al., 2006). However, even though disseminated tumor cells were found in the liver of compound mutant mice, macrometastases were never observed. To date, no genetically modified mouse models exist that can recapitulate metastasis formation from digestive cancer; new approaches are therefore urgently needed to identify the genes that are critically involved in late stages of tumor progression.

Epithelial Morphogenesis and Intestinal Cancer

101

IX. METASTASIS A great number of reviews deal with the initial steps of intestinal tumor formation and progression; however, we have chosen to concentrate here on the final steps of tumor progression, notably the formation of distant metastases, which are the decisive factor for the survival of patients with digestive cancer. The multistep process of metastasis formation is highly complex and still far from being understood. Here we discuss the various steps of metastasis formation and the crucial involvement of the canonical Wnt signaling cascade in these processes.

X. INITIAL STEP OF METASTASIS—INVASION THROUGH THE EMT OR COLLECTIVE CELL MIGRATION? Tumor cell invasion into the surrounding tissue either as individual cells or as sheets occurs at the interface between the tumor mass and its microenvironment; the so called invasive front. Most solid tumors, including CRC, are not homogeneous and often exhibit an invasive front that differs in cell morphology and molecular composition from a central more differentiated region (Brabletz et al., 2001). In the center of a CRC, cells maintain their epithelial morphology and retain membranous and cytoplasmic ‐catenin localization, whereas cells at the invasive front demonstrate comparatively a greater accumulation of nuclear ‐catenin. The formation of secondary tumors, or metastases, in distant organs is a critical step in cancer progression and is the major cause of tumor‐related mortality. Metastases are formed from cancer cells that detach from the primary tumor bulk, invade the surrounding tissues, disseminate through the body, then arrest in a target organ where they proliferate and build another tumor. The dissemination of tumor cells requires the capacity for cell migration and invasion as a prerequisite for metastasis. Such changes in cell phenotype can be achieved either by loss of epithelial characteristics and migration of individual cells (Brabletz et al., 2005a, 2005b), or by the migration of sheets of attached cells in a process known as collective cell migration (Friedl, 2004). Single cell migration is mainly dependent on signaling pathways within the migrating cells themselves and is accompanied by the disruption of cell–cell contacts suggested to be achieved by a process reminiscent of the epithelial to mesenchymal transition (EMT) (Thiery, 2002). This involves downregulation of epithelial‐specific proteins, including E‐cadherin and cytokeratins, and expression of mesenchymal‐specific

102

Silvia Fre et al.

molecules, such as vimentin and fibronectin (Lee et al., 2006). In contrast, collective cell migration requires the maintenance of cell–cell adhesion and a certain level of multicellular organization within the migrating tumor sheet. Full EMT and invasion of single cancer cells was observed in human colorectal cancer (CRC) samples by Brabletz and colleagues (Spaderna et al., 2006). However, a significant proportion of tumors (Wicki et al., 2006) that are characterized as invasive and malignant by pathological criteria fail to display the molecular signatures of EMT. An example is the analysis of human CRC biopsies that revealed a phenotype of large, invading tumor fronts with rare single cells, corresponding best to the phenotype of collective cell migration (Vignjevic et al., 2007; Wicki et al., 2006). Nevertheless, these data do not necessarily contradict each other, since some of the invasive cells might undergo an EMT later in the invasion process. Cancer cells are indeed able to change their migration mode, from collective to individual, to enter and escape vessels (Friedl, 2004). In addition, cancer cells in the EMT acquire a fibroblast‐like morphology and express mesenchymal markers and since they are mixed with the stromal compartment, it becomes difficult to discriminate them from the stromal cells. Finally, a newly emerging idea is that invasive cells undergo a partial EMT, gaining some mesenchymal properties such as secretion of matrix metalloproteases (MMPs), which allow them to form specific membrane protrusions and invade through the extracellular matrix (ECM), while still retaining some of their epithelial characteristics. Taken together, these observations strengthen our perception of the extraordinary plasticity of invasive tumor cells and exemplify the complexity of understanding the regulation of these composite transitions.

XI. TWO‐PHASE MODEL FOR b‐CATENIN TARGET GENE ACTIVATION The transcriptional program driven by Wnt signals has been shown to have a critical role not only during tumor initiation, but also in the invasive process. Indeed, Brabletz and colleagues suggested a two‐phase model for ‐catenin target gene activation (Brabletz et al., 2005b), where the expression of Wnt target genes is controlled both temporally and spatially. In phase I, early in tumorigenesis, low levels of nuclear ‐catenin may be sufficient for the persistent activation of proliferation‐associated genes, which are expressed throughout tumor progression (Fig. 3). During progression from adenoma to carcinoma, a further increase in nuclear ‐catenin, reaching maximal levels in cells at the invasive front of the carcinoma (Brabletz et al., 2000) is believed to lead to the transient induction of metastasis‐associated genes and is referred to as the phase II of ‐catenin activation (Fig. 3).

Normal epithelium

Adenoma

Carcinoma in situ

Invasive carcinoma

Liver metastasis

H&E

b-catenin

Nuclear b-catenin level

Phase I c-myc, cyclinD1

Phase II L1, Nr-CAM, fascin MMP-7, MT1-MMP, uPAR

Fig. 3 Different steps of colorectal tumorigenesis. Hematoxylin and Eosin (H&E) staining of a normal mouse small intestine (normal epithelium), an adenoma and the progression to the more aggressive stages of carcinoma in situ and invasive carcinoma; a distant metastasis in the liver is also shown (liver metastasis). The same sections are stained in the lower panel with an anti‐ ‐catenin antibody in brown, illustrating the changes in ‐catenin accumulation and subcellular localization. The inset in the normal epithelium represents a higher magnification of a crypt showing nuclear localization of ‐catenin. ‐catenin is found in the nucleus of tumor cells already in the adenoma and its levels are strongly increased at the invasive front

104

Silvia Fre et al.

The mechanisms responsible for the accumulation of ‐catenin in the nucleus of cancer cells specifically at the invasive front are not fully understood. One hypothesis proposes that destabilization of the E‐cadherin junctions during EMT leads to the release of membranous ‐catenin into the cytoplasm, thus increasing the amount of protein available for nuclear translocation (Brabletz et al., 2005b). Alternatively, environmental signals from the stroma may be responsible for the increased nuclear translocation of ‐catenin and the subsequent recruitment of coactivators, which would drive the expression of new genes leading to the dissemination of tumor cells. In addition, Wnt target gene expression might reinforce the inhibition of E‐cadherin function and therefore the dissociation of adherens junctions, promoting EMT through a positive feedback loop (Shtutman et al., 2006).

XII. b‐CATENIN TARGET GENES AT THE INVASIVE FRONT Recent data from several laboratories have demonstrated a direct link between high nuclear ‐catenin levels, specifically in cells at the invasive front of a tumor and colorectal cancer progression. The enhanced transcriptional activity of ‐catenin at the invasive front appears to lead to the local and transient expression of genes directly or indirectly involved in the invasion and migration with or without EMT. For example, ‐catenin activates the expression of Slug, a transcriptional repressor of E‐cadherin, resulting in the disruption of cell–cell contacts (Conacci‐Sorrell et al., 2003). ‐Catenin also activates several genes involved in cell migration and adhesion to the extracellular matrix, like Nr‐CAM, osteopontin, and L1 (Conacci‐Sorrell et al., 2002, 2003; Gavert et al., 2005; Rimkus et al., 2006). L1 and Nr‐CAM are transmembrane proteins normally expressed in the nervous system, where they induce neuronal outgrowth, sensory guidance, and path‐finding (Kamiguchi et al., 1998; Sakurai et al., 2001). Both L1 and Nr‐CAM are absent in the normal colonic epithelium but become locally expressed at the invasive front of CRC. Expression of L1 or Nr‐CAM in cultured cells promotes cell migration and invasion through mechanisms believed to be closely related to those operating in normal neural cells. Cleavage of L1 by ADAM10, a metalloprotease secreted by the stromal

of carcinomas (inset in the invasive carcinoma image). Note that while the normal epithelium and the adenoma are mouse sections, the more aggressive stages of carcinoma and the metastasis represent here human specimens. The two phases of ‐catenin activation and the respective target genes are illustrated as a diagram at the bottom.

Epithelial Morphogenesis and Intestinal Cancer

105

cells, generates an extracellular subunit that binds to integrins and enhances the migration of carcinoma cells (Mechtersheimer et al., 2001). One hypothesis proposes that cleaved L1 plays a role in the guidance of invasive cells to the vessels by recognizing the surrounding nerve cells which then act as axon tracks for tumor cell dissemination (Gavert et al., 2005). The intracellular part of L1 may also play a functional role by activating signaling pathways, such as the MAPK cascade (Faivre‐Sarrailh et al., 1999; Schaefer et al., 1999), and as a consequence induce other mechanisms involved in cell migration and invasion. Fascin is another example of a neuronal protein whose expression is induced in human colorectal tumors in response to Wnt signals (Vignjevic et al., 2007). Fascin is an actin‐binding protein necessary for the formation of long membrane protrusions termed filopodia (Vignjevic et al., 2006), which are thought to act as guidance organelles for directional cell migration. Recently, it has been found that the expression of fascin promotes the invasive capacity of colon cancer cells. Cultured CRC cells expressing fascin degrade the basement membrane matrix matrigel and penetrate through filters more efficiently than controls. Also, fascin expressing cells have an enhanced ability to disseminate to the lungs and to form distant metastases upon injection into the tail vein of mice; in addition, fascin is expressed at high levels in human colon carcinomas, most significantly in cancer cells at the invasive front of tumors (Vignjevic et al., 2007). Furthermore, in order to invade adjacent cell layers, malignant cells need to remodel the local tissue environment by excavating passageways through the ECM and removing cells that obstruct their path. The degradation of the basal membrane and the ECM is achieved by the action of matrix metalloproteinases (MMPs). In carcinomas, the majority of these proteases are secreted by recruited stromal cells, notably macrophages and fibroblasts, rather than by the cancer epithelial cells. However, through changes in transcriptional activity, cancer cells also begin secreting MMPs themselves. The Wnt signaling pathway, for example, induces expression of MMP‐7 (Brabletz et al., 1999; Crawford et al., 1999), MT1‐MMP (Takahashi et al., 2002) and uPAR (Mann et al., 1999) and, as a consequence, ECM degradation is enhanced and the efficiency of cell invasion is increased.

XIII. EMT IS A REVERSIBLE PROCESS At the invasive front of the tumor, the architecture of the intestinal epithelium is disrupted and the basal membrane is degraded to enable the dissemination of tumor cells (Spaderna et al., 2006). However, epithelial organization is reestablished in the metastatic site where invasive cells form

106

Silvia Fre et al.

the secondary tumor (Spaderna et al., 2006). Thus, the EMT process is a transient and reversible event (Kirchner and Brabletz, 2000; Thiery, 2002) that must be at least in part controlled by external noncell autonomous signals rather than acquired mutations alone. Once cells reach their destination in the target organ, migration and invasion are no longer required, and they resume proliferation and redifferentiate (Brabletz et al., 2005b). This process is characterized by the relocalization of ‐catenin to adherens junctions, together with E‐cadherin, concomitant with a reduction of nuclear ‐catenin levels and Wnt signals (Brabletz et al., 2005b; Vignjevic et al., 2007). As a consequence, expression of the metastatic genes that are regulated by ‐catenin, such as L1, Nr‐CAM and fascin, is downregulated. This suggests that during colorectal tumorigenesis the expression of metastatic genes is tightly regulated in a spatiotemporal manner. Their transient upregulation promotes the acquisition of migratory and invasive features leading to metastasis, but their expression is downregulated once the tumor cells reach their metastatic destination, where migration ceases and proliferation is enhanced.

XIV. CONCLUDING REMARKS AND PERSPECTIVES Since the seminal work of Vogelstein and colleagues demonstrating the multiple genetic steps necessary for the progression of colorectal cancer, a great deal of information has been unraveled on the genetic changes associated with loss or gain of function of key genes involved in this disease. Not surprisingly, regulators of all major signaling pathways have been identified as key players both for the maintenance of gut homeostasis and for tumor progression and the mechanisms involved in these processes are remarkably conserved throughout evolution. Indeed, with the help of innovative genetic tools applied to mouse models, in vivo studies have been undertaken, thus allowing the direct experimental assessment of the contribution of genes suspected to play a role in the development of human cancers. The molecular mechanisms contributing to the dissemination of tumor cells, leading to aggressive metastasis, remain poorly understood. The challenge for future studies is to identify the key players in the final stages of tumor progression. Similarly, the properties of the stem cells in the gut that ensure its rapid renewal throughout life are yet to be discovered. This research is indeed crucial if one hopes to establish the origin of the tumor initiating cells (cancer stem cells); such a breakthrough will pave the way to new strategies for early diagnosis and innovative targeted therapies in colorectal cancer.

Epithelial Morphogenesis and Intestinal Cancer

107

ACKNOWLEDGEMENTS We are grateful to Melissa Wong for giving permission to modify the diagram of the intestine depicted in Figs. 1 and 2. S.F. is supported by a Marie Curie Intra‐European Fellowship; D.V. receives a Human Frontiers Science Program Organization Fellowship; M.S. is funded by an “Allocation couple´e” from the Ecole Normale Supe´rieure, Paris; S.L.D. currently receives a “Bourse du Conseil Scientifique” from the Ville de Paris; K.P.J. is supported by grants from the Deutsche Forschungsgemeinschaft (DFG/SFB456), Kommission fu¨r Klinische Forschung (KKF/MRI) and the Wilhelm‐Sander Stiftung. Our work is supported by grants from Institut National du Cancer (PL043) and Association pour la Recherche sur le Cancer (2976) to S.R. We apologize to those whose papers are not cited here because of limited space.

REFERENCES Aliaga, J. C., Deschenes, C., Beaulieu, J. F., Calvo, E. L., and Rivard, N. (1999). Am. J. Physiol. 277, G631–G641. Artavanis‐Tsakonas, S., Rand, M. D., and Lake, R. J. (1999). Science 284, 770–776. Batlle, E., Bacani, J., Begthel, H., Jonkheer, S., Gregorieff, A., van de Born, M., Malats, N., Sancho, E., Boon, E., Pawson, T., Gallinger, S., Pals, S., et al. (2005). Nature 435, 1126–1130. Batlle, E., Henderson, J. T., Beghtel, H., van den Born, M. M., Sancho, E., Huls, G., Meeldijk, J., Robertson, J., van de Wetering, M., Pawson, T., and Clevers, H. (2002). Cell 111, 251–263. Bellavia, D., Campese, A. F., Alesse, E., Vacca, A., Felli, M. P., Balestri, A., Stoppacciaro, A., Tiveron, C., Tatangelo, L., Giovarelli, M., Gaetano, C., Ruco, L., et al. (2000). EMBO J. 19, 3337–3348. Bos, J. L. (1989). Cancer Res. 49, 4682–4689. Brabletz, T., Herrmann, K., Jung, A., Faller, G., and Kirchner, T. (2000). Am. J. Pathol. 156, 865–870. Brabletz, T., Hlubek, F., Spaderna, S., Schmalhofer, O., Hiendlmeyer, E., Jung, A., and Kirchner, T. (2005a). Cells Tissues Organs 179, 56–65. Brabletz, T., Jung, A., Dag, S., Hlubek, F., and Kirchner, T. (1999). Am. J. Pathol. 155, 1033–1038. Brabletz, T., Jung, A., Reu, S., Porzner, M., Hlubek, F., Kunz‐Schughart, L. A., Knuechel, R., and Kirchner, T. (2001). Proc. Natl. Acad. Sci. USA 98, 10356–10361. Brabletz, T., Jung, A., Spaderna, S., Hlubek, F., and Kirchner, T. (2005b). Nat. Rev. Cancer 5, 744–749. Brantjes, H., Roose, J., van De Wetering, M., and Clevers, H. (2001). Nucleic Acids Res. 29, 1410–1419. Brummelkamp, T. R., Bernards, R., and Agami, R. (2002). Cancer Cell 2, 243–247. Capobianco, A. J., Zagouras, P., Blaumueller, C. M., Artavanis‐Tsakonas, S., and Bishop, J. M. (1997). Mol. Cell Biol. 17, 6265–6273. Chang, L., and Karin, M. (2001). Nature 410, 37–40. Clevers, H. (2006). Cell 127, 469–480. Clevers, H., and Batlle, E. (2006). Cancer Res. 66, 2–5. Conacci‐Sorrell, M., Simcha, I., Ben‐Yedidia, T., Blechman, J., Savagner, P., and Ben‐Ze’ev, A. (2003). J. Cell Biol. 163, 847–857. Conacci‐Sorrell, M. E., Ben‐Yedidia, T., Shtutman, M., Feinstein, E., Einat, P., and Ben‐Ze’ev, A. (2002). Genes Dev. 16, 2058–2072.

108

Silvia Fre et al.

Coulthard, M. G., Duffy, S., Down, M., Evans, B., Power, M., Smith, F., Stylianou, C., Kleikamp, S., Oates, A., Lackmann, M., Burns, G. F., and Boyd, A. W. (2002). Int. J. Dev. Biol. 46, 375–384. Crawford, H. C., Fingleton, B. M., Rudolph‐Owen, L. A., Goss, K. J., Rubinfeld, B., Polakis, P., and Matrisian, L. M. (1999). Oncogene 18, 2883–2891. Crosnier, C., Vargesson, N., Gschmeissner, S., Ariza‐McNaughton, L., Morrison, A., and Lewis, J. (2005). Development 132, 1093–1104. Davies, H., Bignell, G. R., Cox, C., Stephens, P., Edkins, S., Clegg, S., Teague, J., Woffendin, H., Garnett, M. J., Bottomley, W., Davis, N., Dicks, E., et al. (2002). Nature 417, 949–954. Faivre‐Sarrailh, C., Falk, J., Pollerberg, E., Schachner, M., and Rougon, G. (1999). J. Cell. Sci. 112(Pt 18), 3015–3027. Fearon, E. R., and Vogelstein, B. (1990). Cell 61, 759–767. Fodde, R., and Brabletz, T. (2007). Curr. Opin. Cell Biol. 19, 150–158. Fodde, R., Edelmann, W., Yang, K., van Leeuwen, C., Carlson, C., Renault, B., Breukel, C., Alt, E., Lipkin, M., Khan, P. M., and Kucherlapati, R. (1994). Proc. Natl. Acad. Sci. USA 91, 8969–8973. Forrester, K., Almoguera, C., Han, K., Grizzle, W. E., and Perucho, M. (1987). Nature 327, 298–303. Fre, S., Huyghe, M., Mourikis, P., Robine, S., Louvard, D., and Artavanis‐Tsakonas, S. (2005). Nature 435, 964–968. Friedl, P. (2004). Curr. Opin. Cell Biol. 16, 14–23. Gallahan, D., and Callahan, R. (1997). Oncogene 14, 1883–1890. Gallahan, D., Jhappan, C., Robinson, G., Hennighausen, L., Sharp, R., Kordon, E., Callahan, R., Merlino, G., and Smith, G. H. (1996). Cancer Res. 56, 1775–1785. Gavert, N., Conacci‐Sorrell, M., Gast, D., Schneider, A., Altevogt, P., Brabletz, T., and Ben‐Ze’ev, A. (2005). J. Cell Biol. 168, 633–642. Grunwald, I. C., Korte, M., Wolfer, D., Wilkinson, G. A., Unsicker, K., Lipp, H. P., Bonhoeffer, T., and Klein, R. (2001). Neuron 32, 1027–1040. Hanahan, D., and Weinberg, R. A. (2000). Cell 100, 57–70. Hancock, J. F. (2003). Nat. Rev. Mol. Cell Biol. 4, 373–384. Haramis, A. P., Begthel, H., van den Born, M., van Es, J., Jonkheer, S., Offerhaus, G. J., and Clevers, H. (2004). Science 303, 1684–1686. Hardwick, J. C., Van Den Brink, G. R., Bleuming, S. A., Ballester, I., Van Den Brande, J. M., Keller, J. J., Offerhaus, G. J., Van Deventer, S. J., and Peppelenbosch, M. P. (2004). Gastroenterology 126, 111–121. Hasson, P., Egoz, N., Winkler, C., Volohonsky, G., Jia, S., Dinur, T., Volk, T., Courey, A. J., and Paroush, Z. (2005). Nat. Genet. 37, 101–105. Hasson, P., and Paroush, Z. (2006). Br. J. Cancer 94, 771–775. He, X. C., Yin, T., Grindley, J. C., Tian, Q., Sato, T., Tao, W. A., Dirisina, R., Porter‐ Westpfahl, K. S., Hembree, M., Johnson, T., Wiedemann, L. M., Barrett, T. A., et al. (2007). Nat. Genet. 39, 189–198. He, X. C., Zhang, J., and Li, L. (2005). Ann. N Y Acad. Sci. 1049, 28–38. He, X. C., Zhang, J., Tong, W. G., Tawfik, O., Ross, J., Scoville, D. H., Tian, Q., Zeng, X., He, X., Wiedemann, L. M., Mishina, Y., and Li, L. (2004). Nat. Genet. 36, 1117–1121. Hirai, H., Maru, Y., Hagiwara, K., Nishida, J., and Takaku, F. (1987). Science 238, 1717–1720. Howe, J. R., Roth, S., Ringold, J. C., Summers, R. W., Jarvinen, H. J., Sistonen, P., Tomlinson, I. P., Houlston, R. S., Bevan, S., Mitros, F. A., Stone, E. M., and Aaltonen, L. A. (1998). Science 280, 1086–1088. Itoh, S., Itoh, F., Goumans, M. J., and Ten Dijke, P. (2000). Eur. J. Biochem. 267, 6954–6967. Janssen, K. P. (2003). Cell. Mol. Life Sci. 60, 495–506.

Epithelial Morphogenesis and Intestinal Cancer

109

Janssen, K. P., Alberici, P., Fsihi, H., Gaspar, C., Breukel, C., Franken, P., Rosty, C., Abal, M., El Marjou, F., Smits, R., Louvard, D., Fodde, R., and Robine, S. (2006). Gastroenterology 131, 1096–1109. Janssen, K. P., el‐Marjou, F., Pinto, D., Sastre, X., Rouillard, D., Fouquet, C., Soussi, T., Louvard, D., and Robine, S. (2002). Gastroenterology 123, 492–504. Jenny, M., Uhl, C., Roche, C., Duluc, I., Guillermin, V., Guillemot, F., Jensen, J., Kedinger, M., and Gradwohl, G. (2002). EMBO J. 21, 6338–6347. Jensen, J., Pedersen, E. E., Galante, P., Hald, J., Heller, R. S., Ishibashi, M., Kageyama, R., Guillemot, F., Serup, P., and Madsen, O. D. (2000). Nat. Genet. 24, 36–44. Johnson, G. L., and Lapadat, R. (2002). Science 298, 1911–1912. Kamiguchi, H., Hlavin, M. L., and Lemmon, V. (1998). Mol. Cell Neurosci. 12, 48–55. Kataoka, H., Igarashi, H., Kanamori, M., Ihara, M., Wang, J. D., Wang, Y. J., Li, Z. Y., Shimamura, T., Kobayashi, T., Maruyama, K., Nakamura, T., Arai, H., et al. (2004). Cancer Sci. 95, 136–141. Khosravi‐Far, R., Campbell, S., Rossman, K. L., and Der, C. J. (1998). Adv. Cancer Res. 72, 57–107. Kiaris, H., Politi, K., Grimm, L. M., Szabolcs, M., Fisher, P., Efstratiadis, A., and Artavanis‐ Tsakonas, S. (2004). Am. J. Pathol. 165, 695–705. Kim, I., Ryu, Y. S., Kwak, H. J., Ahn, S. Y., Oh, J. L., Yancopoulos, G. D., Gale, N. W., and Koh, G. Y. (2002). Faseb J. 16, 1126–1128. Epub 2002 May 21. Kinch, M. S., Clark, G. J., Der, C. J., and Burridge, K. (1995). J. Cell Biol. 130, 461–471. Kinzler, K. W., and Vogelstein, B. (1998). Science 280, 1036–1037. Kirchner, T., and Brabletz, T. (2000). Am. J. Pathol. 157, 1113–1121. Kiyokawa, E., Takai, S., Tanaka, M., Iwase, T., Suzuki, M., Xiang, Y. Y., Naito, Y., Yamada, K., Sugimura, H., and Kino, I. (1994). Cancer Res. 54, 3645–3650. Korinek, V., Barker, N., Moerer, P., van Donselaar, E., Huls, G., Peters, P. J., and Clevers, H. (1998). Nat. Genet. 19, 379–383. Kuhnert, F., Davis, C. R., Wang, H. T., Chu, P., Lee, M., Yuan, J., Nusse, R., and Kuo, C. J. (2004). Proc. Natl. Acad. Sci. USA 101, 266–271. Lee, J. M., Dedhar, S., Kalluri, R., and Thompson, E. W. (2006). J. Cell Biol. 172, 973–981. Liaw, D., Marsh, D. J., Li, J., Dahia, P. L., Wang, S. I., Zheng, Z., Bose, S., Call, K. M., Tsou, H. C., Peacocke, M., Eng, C., and Parsons, R. (1997). Nat. Genet. 16, 64–67. Liu, W., Dong, X., Mai, M., Seelan, R. S., Taniguchi, K., Krishnadath, K. K., Halling, K. C., Cunningham, J. M., Boardman, L. A., Qian, C., Christensen, E., Schmidt, S. S., et al. (2000). Nat. Genet. 26, 146–147. Luo, D., Renault, V. M., and Rando, T. A. (2005). Semin. Cell Dev. Biol. 16, 612–622. Madison, B. B., Braunstein, K., Kuizon, E., Portman, K., Qiao, X. T., and Gumucio, D. L. (2005). Development 132, 279–289. Malumbres, M., and Barbacid, M. (2003). Nat. Rev. Cancer 3, 459–465. Mann, B., Gelos, M., Siedow, A., Hanski, M. L., Gratchev, A., Ilyas, M., Bodmer, W. F., Moyer, M. P., Riecken, E. O., Buhr, H. J., and Hanski, C. (1999). Proc. Natl. Acad. Sci. USA 96, 1603–1608. Mariadason, J. M., Bordonaro, M., Aslam, F., Shi, L., Kuraguchi, M., Velcich, A., and Augenlicht, L. H. (2001). Cancer Res. 61, 3465–3471. McCormick, F. (1999). Trends Cell Biol. 9, M53–56. Mechtersheimer, S., Gutwein, P., Agmon‐Levin, N., Stoeck, A., Oleszewski, M., Riedle, S., Postina, R., Fahrenholz, F., Fogel, M., Lemmon, V., and Altevogt, P. (2001). J. Cell Biol. 155, 661–673. Miao, H., Wei, B. R., Peehl, D. M., Li, Q., Alexandrou, T., Schelling, J. R., Rhim, J. S., Sedor, J. R., Burnett, E., and Wang, B. (2001). Nat. Cell Biol. 3, 527–530. Micchelli, C. A., and Perrimon, N. (2006). Nature 439, 475–479.

110

Silvia Fre et al.

Molofsky, A. V., Pardal, R., and Morrison, S. J. (2004). Curr. Opin. Cell Biol. 16, 700–707. Morin, P. J., Sparks, A. B., Korinek, V., Barker, N., Clevers, H., Vogelstein, B., and Kinzler, K. W. (1997). Science 275, 1787–1790. Moser, A. R., Pitot, H. C., and Dove, W. F. (1990). Science 247, 322–324. Muller, C. A., Autenrieth, I. B., and Peschel, A. (2005). Cell Mol. Life Sci. 62, 1297–1307. Nagashima, K., Endo, A., Ogita, H., Kawana, A., Yamagishi, A., Kitabatake, A., Matsuda, M., and Mochizuki, N. (2002). Mol. Biol. Cell 13, 4231–4242. Nusse, R., and Varmus, H. E. (1992). Cell 69, 1073–1087. Ohlstein, B., and Spradling, A. (2006). Nature 439, 470–474. Pear, W. S., Aster, J. C., Scott, M. L., Hasserjian, R. P., Soffer, B., Sklar, J., and Baltimore, D. (1996). J. Exp. Med. 183, 2283–2291. Persad, S., Troussard, A. A., McPhee, T. R., Mulholland, D. J., and Dedhar, S. (2001). J. Cell Biol. 153, 1161–1174. Picco, V., Hudson, C., and Yasuo, H. (2007). Development 134, 1491–1497. Pinto, D., Gregorieff, A., Begthel, H., and Clevers, H. (2003). Genes Dev. 17, 1709–1713. Rajagopalan, H., Bardelli, A., Lengauer, C., Kinzler, K. W., Vogelstein, B., and Velculescu, V. E. (2002). Nature 418, 934. Ramalho‐Santos, M., Melton, D. A., and McMahon, A. P. (2000). Development 127, 2763–2772. Reya, T., and Clevers, H. (2005). Nature 434, 843–850. Rizvi, A. Z., and Wong, M. H. (2005). Stem Cells 23, 150–165. Rohde, F., Rimkus, C., Friederichs, J., Rosenberg, R., Marthen, C., Doll, D., Holzmann, B., Siewert, J. R., and Janssen, K. P. (2007). Int. J. Cancer 121, 1717–1723. Rohn, J. L., Lauring, A. S., Linenberger, M. L., and Overbaugh, J. (1996). J. Virol. 70, 8071–8080. Sakurai, T., Lustig, M., Babiarz, J., Furley, A. J., Tait, S., Brophy, P. J., Brown, S. A., Brown, L. Y., Mason, C. A., and Grumet, M. (2001). J. Cell Biol. 154, 1259–1273. Sander, G. R., and Powell, B. C. (2004). J. Histochem. Cytochem. 52, 509–516. Schaefer, A. W., Kamiguchi, H., Wong, E. V., Beach, C. M., Landreth, G., and Lemmon, V. (1999). J. Biol. Chem. 274, 37965–37973. Schroder, N., and Gossler, A. (2002). Gene Expr. Patterns 2, 247–250. Sheng, G., Bernabe, K. Q., Guo, J., and Warner, B. W. (2006). Gastroenterology 131, 153–164. Shi, Y., and Massague, J. (2003). Cell 113, 685–700. Shilo, B. Z. (2005). Development 132, 4017–4027. Shtutman, M., Levina, E., Ohouo, P., Baig, M., and Roninson, I. B. (2006). Cancer Res. 66, 11370–11380. Soriano, J. V., Uyttendaele, H., Kitajewski, J., and Montesano, R. (2000). Int. J. Cancer 86, 652–659. Spaderna, S., Schmalhofer, O., Hlubek, F., Berx, G., Eger, A., Merkel, S., Jung, A., Kirchner, T., and Brabletz, T. (2006). Gastroenterology 131, 830–840. Stephenson, S. A., Slomka, S., Douglas, E. L., Hewett, P. J., and Hardingham, J. E. (2001). BMC Mol. Biol. 2, 15. Stiles, B., Groszer, M., Wang, S., Jiao, J., and Wu, H. (2004). Dev. Biol. 273, 175–184. Su, L. K., Kinzler, K. W., Vogelstein, B., Preisinger, A. C., Moser, A. R., Luongo, C., Gould, K. A., and Dove, W. F. (1992). Science 256, 668–670. Suzuki, T., and Chiba, S. (2005). Int. J. Hematol. 82, 285–294. Takahashi, M., Tsunoda, T., Seiki, M., Nakamura, Y., and Furukawa, Y. (2002). Oncogene 21, 5861–5867. Thiery, J. P. (2002). Nat. Rev. Cancer 2, 442–454. Trahey, M., and McCormick, F. (1987). Science 238, 542–545.

Epithelial Morphogenesis and Intestinal Cancer

111

van de Wetering, M., Sancho, E., Verweij, C., de Lau, W., Oving, I., Hurlstone, A., van der Horn, K., Batlle, E., Coudreuse, D., Haramis, A. P., Tjon‐Pon‐Fong, M., Moerer, P., et al. (2002). Cell 111, 241–250. van den Brink, G. R. (2004). Nat. Genet. 36, 1038–1039. van den Brink, G. R., Bleuming, S. A., Hardwick, J. C., Schepman, B. L., Offerhaus, G. J., Keller, J. J., Nielsen, C., Gaffield, W., van Deventer, S. J., Roberts, D. J., and Peppelenbosch, M. P. (2004). Nat. Genet. 36, 277–282. van Es, J. H., van Gijn, M. E., Riccio, O., van den Born, M., Vooijs, M., Begthel, H., Cozijnsen, M., Robine, S., Winton, D. J., Radtke, F., and Clevers, H. (2005). Nature 435, 959–963. Vignjevic, D., Kojima, S., Aratyn, Y., Danciu, O., Svitkina, T., and Borisy, G. G. (2006). J. Cell Biol. 174, 863–875. Vignjevic, D., Schoumacher, M., Gavert, N., Janssen, K. P., Lae, M., Louvard, D., Ben‐Ze’ev, A., and Robine, S. (2007). Cancer Res. 67, 6844–6853. Vogelstein, B., Fearon, E. R., Hamilton, S. R., Kern, S. E., Preisinger, A. C., Leppert, M., Nakamura, Y., White, R., Smits, A. M., and Bos, J. L. (1988). N. Engl. J. Med. 319, 525–532. Waite, K. A., and Eng, C. (2003). Hum. Mol. Genet. 12, 679–684. Walker, L., Lynch, M., Silverman, S., Fraser, J., Boulter, J., Weinmaster, G., and Gasson, J. C. (1999). Stem Cells 17, 162–171. Wicki, A., Lehembre, F., Wick, N., Hantusch, B., Kerjaschki, D., and Christofori, G. (2006). Cancer Cell 9, 261–272. Yamada, Y., Hata, K., Hirose, Y., Hara, A., Sugie, S., Kuno, T., Yoshimi, N., Tanaka, T., and Mori, H. (2002). Cancer Res. 62, 6367–6370. Yang, Q., Bermingham, N. A., Finegold, M. J., and Zoghbi, H. Y. (2001). Science 294, 2155–2158. Yoon, K., and Gaiano, N. (2005). Nat. Neurosci. 8, 709–715. Zhang, B., Ougolkov, A., Yamashita, K., Takahashi, Y., Mai, M., and Minamoto, T. (2003). Clin. Cancer Res. 9, 3073–3079. Zhang, Y., and Kalderon, D. (2001). Nature 410, 599–604. Zou, J. X., Wang, B., Kalo, M. S., Zisch, A. H., Pasquale, E. B., and Ruoslahti, E. (1999). Proc. Natl. Acad. Sci. USA 96, 13813–13818.

Molecular Mechanisms and Therapeutic Development of Angiogenesis Inhibitors Yihai Cao Laboratory of Angiogenesis Research, Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, 171 77 Stockholm, Sweden

I. II. III. IV. V. VI. VII. VIII. IX. X.

Introduction Tumor Angiogenic Factors and Blood Vessels Therapeutic Targets of Angiogenic Factors Drug Resistance Issues Side Effects Mechanisms of Broad‐Spectrum Angiogenesis Inhibitors Antiangiogenic Therapy vs. Chemotherapy Patient Selection and Timescale of Treatment Biomarkers Future Perspectives References

Bevacizumab (Avastin), a vascular endothelial growth factor antagonist, is the first approved antiangiogenic drug for the treatment of human cancers. Endostatin, a broad‐ spectrum endogenous angiogenesis inhibitor, has recently been approved in China for cancer therapy. Today, hundreds of antiangiogenic molecules targeting different signaling pathways are being tested for their anticancer efficacies at preclinical and clinical stages. The underlying mechanisms by which these antiangiogenic cancer drugs used in combination with chemotherapy confer survival advantages for cancer patients are not fully understood. Thus, deeper understanding the mechanisms of tumor angiogenesis and actions of these therapeutic molecules is crucial for designing more potent anticancer drugs. # 2008 Elsevier Inc.

I. INTRODUCTION The tumor vasculature has become an increasingly attractive target for development of new anticancer drugs. Emerging evidence shows that the development of the tumor vasculature is controlled by a number of cell growth regulatory machineries, including various growth factors, transcriptional factors, adhesion molecules, the notch signaling system, and hypoxia (Blouw et al., 2003; Li and Harris, 2005). Pharmaceutical intervention of the functions of tumor‐produced angiogenic factors has successfully brought the first anti‐VEGF (vascular endothelial growth factor) drug, Avastin, to the market for the treatment of human cancers after showing a survival benefit Advances in CANCER RESEARCH Copyright 2008, Elsevier Inc. All rights reserved.

113

0065-230X/08 $35.00 DOI: 10.1016/S0065-230X(08)00004-3

114

Yihai Cao

in a phase III metastatic colorectal study (Hurwitz et al., 2004). This recent success has validated the early concept that tumor growth is dependent on angiogenesis, and that antiangiogenesis is a valid approach for cancer therapy (Folkman, 1971). It is estimated that more than US$4 billion have been invested so far in angiogenesis research and drug development, and probably 500 million people globally will benefit from future pro‐ or antiangiogenic therapy (Carmeliet, 2005). Tumors produce multiple growth factors that individually or synergistically stimulate vessel growth (Cao, 2005). Thus, combinations of various angiogenesis inhibitors targeting different signaling pathways would therapeutically be more effective and evoke less resistance. Endostatin, as a paradigm of a broad‐spectrum angiogenesis inhibitor, in combination with chemotherapy has recently been approved by the FDA, China, for the treatment of cancer (Sun et al., 2005). In addition, several previously approved drugs for the treatment of cancer and nonmalignant diseases are antiangiogenic. Examples include celecoxcib (a COX‐2 inhibitor), Tarceva (a growth factor receptor inhibitor), Velcade (a proteosome inhibitor), and interferon‐ (Folkman, 2006). However, recent preclinical and clinical studies with antiangiogenesis agents have raised several crucially unresolved issues, which include: (1) Mechanisms of action; why is antiangiogenic monotherapy less effective than predicted in human cancer patients? What are the underlying mechanisms of the survival benefit gained by using antiangiogenic agents in combination with chemotherapy? (2) Selection of patients; what should be the criteria for selection of cancer patients who are likely to respond to antiangiogenic therapy? (3) Timescale of treatment; how long period should the patients be treated with antiangiogenic agents? (4) Drug resistance; do patients develop drug resistance toward antiangiogenic therapy? (5) Side effects; what are the short‐term and long‐term side effects of antiangiogenic therapy? (6) Monitoring therapeutic efficacy; what kind of biomarkers should be used to monitor therapeutic effects? In this review, I will highlight our recent understanding around these crucial issues and suggest future avenues.

II. TUMOR ANGIOGENIC FACTORS AND BLOOD VESSELS The malignant tissue is composed of several cell types that include tumor cells, blood vessel endothelial cells, pericytes/smooth muscle cells, inflammatory cells, stromal cells, and probably lymphatic endothelial cells. The diversity of cell types determines the complex patterns of gene expression within the tumor tissue. Indeed, various angiogenic factors, including members of the

Molecular Mechanisms and Therapeutic Development

115

VEGF, fibroblast growth factor (FGF), platelet‐derived growth factor (PDGF), insulin‐like growth factor (IGF), hepatocyte growth factor (HGF), transforming growth factor‐beta (TGF‐ ), angiopoietin (Ang), and tumor necrosis factor‐alpha (TNF‐) families, are frequently expressed at high levels in tumors (Ferrara and Kerbel, 2005) (Fig. 1). The genomic instability and heterogeneity of malignant cells could further increase the diversity of gene expression profiles of multiple angiogenic factors during progression of tumors (Cao, 2005). Thus, advanced stages of tumors often simultaneously express several angiogenic factors at high levels (Cao, 2005; Folkman, 2002). In addition, the hypoxic environment found in nearly all tumors could further elevate expression levels of angiogenic factors. For example, the level of VEGF‐ A is remarkably increased by hypoxia via the hypoxia inducible factor‐1alpha (HIF‐1)‐induced pathway (Makino et al., 2001). High levels of VEGF‐A not only promote tumor angiogenesis but also increase vascular leakage, which leads to the development of high interstitial fluid pressure (IFP) in tumors (Dvorak, 2006). The high IFP might act as a positive feedback factor to further increase the hypoxic pressure by restricting blood flow in tumor vessels (Jain, 2005). VEGF‐A also acts as a chemoattractant for monocytes/macrophages and neutrophils that express VEGFR‐1, thus promoting the inflammatory process in tumors. Infiltrated and activated monocytes/macrophages as well as stromal cells also significantly contribute to tumor angiogenesis via production of angiogenic stimulators (Cursiefen et al., 2004; Dong et al., 2004; Lee et al., 2004). Further, VEGF‐A has also been reported to synergistically induce angiogenesis together with other factors such as FGF‐2 and PDGF‐BB (Richardson et al., 2001). Other non‐VEGF factors such as FGF‐2 and PDGF‐ BB can also synergistically induce blood vessel growth (Cao et al., 2003). The complex reciprocal interactions between different factors should therefore be taken into consideration for antiangiogenic drug development. VEGF‐A can also remarkably increase the population of bone marrow‐derived circulating endothelial progenitor cells (BM‐CEPCs) that contribute to tumor neovascularization (Lyden et al., 2001; Rafii and Lyden, 2003). Moreover, tumor‐ produced VEGF‐A might also target several neuronal and nonneuronal healthy tissues through its receptors, VEGFR‐1, VEGFR‐2, and the neuropilins (Fig. 1). The imbalanced production of VEGF‐A and other angiogenic factors leads to malformation of tumor blood vessels. Although growing tumors retain the ability to switch on the angiogenic phenotype by mechanisms of angiogenesis and vasculogenesis, there are some fundamental differences of the structure and function of blood vessels found in tumors versus those in healthy tissues. The tumor vasculature usually consists of disorganized, leaky, and tortuous blood vessels that tend to be hemorrhagic. Owing to the leaky feature of tumor vessels, the tumor tissue resembles a classical inflamed tissue characterized by edema, high IFP, and infiltration of high

Tumor

TCs

BECs SCs VSMCs

ICs

LECs

PDGF

IGF

HGF

Blood vessels

Growth factor/ receptor antagonists Ligand antagonists

VEGF

FGF

Broad-spectrum angiogenesis inhibitors Endostatin Angiostatin

Anti-receptor antibodies

TSP-1 Tumstatin

Features of tumor vasculature - Disorganized, tortuous, leaky, hemorrhagic and primitive vascular plexuses - Lack clear separation between arteriol and venous vessels - Chaotic blood flow - Mosaic cell type

Tyrosine kinase inhibitors

NRP

VEGFR Downstream inhibitors

PDGFR FGFR Signaling

Transcription

IGFR HGFR

Molecular Mechanisms and Therapeutic Development

117

numbers of inflammatory cells. Further, in solid tumors, the microvessels usually exhibit interconnected sinusoidal structures, lacking clear separation between the arterial and venous systems, and form primitive vascular plexuses (Fig. 1). The tumor blood flow travels chaotically at a low speed within these sinusoidal vascular structures. Furthermore, tumor vessels are usually poorly coated with mural cells (pericytes/vascular smooth muscle cells) and may consist of mosaic cell types, including tumor cells (vascular mimicry) (Maniotis et al., 1999). Similar to the tumor cells, the endothelial cells infiltrated in tumor tissues might contain an aberrant genome (Hida et al., 2004). However, the mechanism and significance of an abnormal endothelial genome in contribution to tumor growth is currently not known.

III. THERAPEUTIC TARGETS OF ANGIOGENIC FACTORS Designing protein molecules or small chemical compounds that antagonize growth factor‐induced angiogenesis is an obvious approach of drug development. Owing to its central role in promoting tumor growth, VEGF‐A has become a key therapeutic target and its functions can be blocked at different levels of the signaling pathways (Fig. 1). Because similar approaches can also be applied to the development of other angiogenic factor antagonists, I here use VEGF‐A antagonists as examples for further discussion. In tumor cells, blockage of VEGF‐A biosynthesis using small interference RNA (RNAi) or antisense RNA‐based approaches seems to be valid mainly in in vitro settings, although their in vivo efficacy remains to be investigated (Ryo et al., 2005). Second, gene therapy approaches aimed to block VEGF‐A secretion in tumor cells also present an attractive approach for suppression of tumor growth (Bjorndahl et al., 2004). Third, anti‐ VEGF‐A neutralizing antibodies or aptamers (Pegaptinib, an aptamer that binds and inhibits VEGF165, is approved for the treatment of the wet type of age‐related macular degeneration) block the interaction of VEGF‐A with its receptors expressed on endothelial cells (Gragoudas et al., 2004). Indeed, the first anti‐VEGF‐A cancer drug, Avastin, is a humanized anti‐VEGF Fig. 1 Tumor blood vessels and therapeutic targets of tumor‐derived angiogenic factors. The diversity of cell types in a malignant tumor contributes to the production of multiple angiogenic factors that activate several signaling pathways in vascular endothelial and/or mural cells by binding to their respective receptors. Therapeutic approaches targeting ligands, receptors, and downstream signaling components are under development for antiangiogenic treatment of cancer. Broad‐spectrum endogenous angiogenesis inhibitors might antagonize multiple angiogenic factor‐induced angiogenic responses. TCs, tumor cells; BECs, blood vessel endothelial cells; SCs, Stromal cells; SMCs, Vascular smooth muscle cells; ICs, inflammatory cells; LECs, lymphatic endothelial cells.

118

Yihai Cao

neutralizing antibody with a relatively long half‐life in the blood stream (Hurwitz et al., 2004). Fourth, soluble VEGFRs consisting of the extracellular ligand binding domains acting as traps for tumor‐produced VEGF‐A molecules are under clinical evaluation (Holash et al., 2002). Similarly, anti‐VEGFR‐2 or VEGFR‐1 neutralizing antibodies have also entered into clinical trials (Jain et al., 2006). However, antibodies or soluble VEGFR‐ antibody conjugates might have poor penetration capacity within tumor tissues. One of the most important groups of anti‐VEGF‐A molecules is the small chemical compounds that block the receptor tyrosine kinase (RTK) activities of VEGFRs (Garber, 2006; Schlessinger, 2005). A phase III trial study shows that a RTK inhibitor, SU11248, produced unexpectedly positive results for prolonging survivals in renal carcinoma patients (Schlessinger, 2005). The US FDA approved an anti‐VEGF small molecule, Sunitinib (Sutent, Pfizer) as a new antiangiogenic drug for the treatment of gastrointestinal stromal tumors and advanced renal‐cell carcinoma. This inhibitor targets not only VEGFR‐2, but also PDGFR‐ , KIT, and FLT‐3 kinases. In addition to its tyrosine kinase receptors, VEGF‐A also binds to neuropilins (NRPs), NRP‐1 and NRP‐2, expressed on endothelial and nonendothelial cells (Soker et al., 1998). Although the NRPs do not seem to transduce active angiogenic signals for VEGF‐A, they may modulate the angiogenic activity mediated by the VEGFR‐2 receptor (Soker et al., 2002). Thus, NRPs could also be important therapeutic targets for the development of VEGF antagonists. Fig. 1 summarizes current therapeutic approaches by blocking functions of tumor‐produced angiogenic factors.

IV. DRUG RESISTANCE ISSUES The genomic instability and heterogeneity of tumor cell populations might lead to a shift of the expression of angiogenic factors during tumor progression. Indeed, various tumors at an advanced stage express multiple angiogenic factors (Cao, 2005; Folkman, 2002; Relf et al., 1997). At this stage, anti‐VEGF‐A or other antiangiogenic factor monotherapy would be predicted to be ineffective or even encounter a drug resistance problem. In fact, anti‐VEGF‐A monotherapy might even help tumors to switch on VEGF‐A‐ unrelated angiogenic pathways for growing blood vessels (Ferrara and Kerbel, 2005). Consistent with this notion, tumors have been reported to escape from anti‐VEGF‐A therapy after a relative long‐term treatment in a transgenic mouse tumor model (Casanovas et al., 2005). Another possibility is that anti‐VEGF‐A agents might “normalize” tumor vessels, which then acquire drug resistance (Jain, 2005). The hypoxic tumor microenvironment could also select for tumor cells populations that are less dependent on

Molecular Mechanisms and Therapeutic Development

119

angiogenesis and resistant to antiangiogenic therapy. The mosaic cell types of tumor vessels (vascular mimicry) might also potentially compromise the efficacy of antiangiogenic drugs (Maniotis et al., 1999). A recent study demonstrates that endothelial cells isolated from tumor vessels might also contain unstable genome, aneuploidy, and aberrant gene expression, which might also provide another mechanism of drug escape (Hida et al., 2004). Finally, recruitment of bone marrow‐derived circulating endothelial cells may also account for escape from antiangiogenic therapy if antiangiogenic agents do not target vasculogenesis. Taken together, although endothelial cells within tumor tissues are considered as normal cells, they might nevertheless escape from antiangiogenic therapy via various mechanisms that are different from those of chemotherapy or radiotherapy. However, it should be emphasized that VEGF‐A expression is most likely preserved at sustained high levels during tumor progression. Thus, it is desirable to design therapeutic strategies by combining anti‐VEGF therapy with other antiangiogenic therapies.

V. SIDE EFFECTS Virtually all angiogenic factors have broad cellular targets in the body. Although VEGF‐A is the most specific angiogenic factor that selectively targets endothelial cells, it also acts on nonendothelial cells, including nerve cells, hematopoietic cells, and hepatocytes (Carmeliet and Tessier‐ Lavigne, 2005; LeCouter et al., 2003; Lyden et al., 2001; Rafii and Lyden, 2003) (Fig. 2). The broad‐spectrum targets and ubiquitous distribution of angiogenic factors suggest that they may play important roles in physiological conditions. Indeed, gene deletion of most angiogenic factors or their tyrosine kinase receptors leads to phenotypes with severe vascular defects under physiological and pathological conditions (Betsholtz, 2004; Carmeliet et al., 1996; Carmeliet et al., 2001; Ferrara et al., 1996; Hoch and Soriano, 2003). Thus, it is not unexpected that systemic delivery of inhibitory agents antagonizing functions of these growth factors could result in unwanted effects. For example, Avastin can increase hypertension and cardiovascular thrombosis, leading to high risks of cardiac infarction and stroke (Ferrara and Kerbel, 2005; Garber, 2002; Jain et al., 2006). Similarly, anticancer clinical trials of other anti‐VEGF agents, including PTK inhibitors, have also faced similar cardiovascular side effects (Garber, 2002). Although anti‐ VEGF agents affect the coagulation system, the exact and detailed molecular basis of clinical thrombosis is not known. It is also unclear why such a side effect has not been seen in mice during preclinical studies. One possibility could be that anti‐VEGF agents destabilize the preexisting atherosclerotic plaques that subsequently cause clinically related severe disorders in humans.

120

Yihai Cao

Tumor cells

Collagen XVIII

Inflammantory cells Stromal cells

Proteases

NP-1

VEGF-A

Endostatin

Vascular Vasculogenesis Neurotrophic Hepatic Angiogenesis permeability protection effect

Non-VEGF targets Leaky vessels

Bone marrow

Neuron

Liver

Blood vessels

Chemotheraputic agents

Fig. 2 Multiple in vivo targets of VEGF‐A and chemotherapy. In addition to its angiogenic activity, VEGF‐A also induces vascular permeability, hematopoiesis, neurotrophic activity, and protection of hepatocytes. Chemotherapeutic agents have destructive effects on VEGF‐A‐ dependent and non‐VEGF‐A‐dependent targets.

Perhaps, the most commonly discussed side effects of antiangiogenic drugs include problems in wound healing and pregnancy—the two physiological processes that depends on angiogenesis (Hanahan and Folkman, 1996).

Molecular Mechanisms and Therapeutic Development

121

The straight answer to this concern is that suppression of angiogenesis would certainly affect these systems, unless antiangiogenic agents are able to distinguish the tumor vasculature from physiological angiogenesis elsewhere in the body. Most notably, VEGF‐A plays a crucial role in the development of the ovarian corpus luteum and during ovulation (Ferrara et al., 1998). In preeclampsia (a common pregnancy disorder (affecting about 3.5%) with characteristics of a local hypoxic placenta and systemic responses in the mother, including hypertension, proteinuria, increased blood clotting, and other organ dysfunctions that threaten both the mother and the baby), the circulating levels of soluble VEGFR‐1 (sVEGFR1) are increased and VEGF‐A or placenta growth factor (PlGF) is correspondingly reduced (Levine et al., 2004). Consistent with this notion, the delivery of sVEGFR‐1, an antagonist for VEGF‐A and PlGF, in pregnant rats imitates the clinical symptoms of preeclampsia (Maynard et al., 2003). In addition, VEGF‐A is also involved in maintenance, differentiation, and reciprocal communications of hematopoietic stem cells and endothelial cells in the bone marrow (Coultas et al., 2005). Thus, it is possible that anti‐VEGF‐A agents might cause dysfunction of the hematopoietic system. Systemic delivery of anti‐VEGF‐A drugs might also affect the function of other tissues and organs including the liver and the nervous system because VEGF‐A acts as a trophic factor for nerve cells and hepatocytes (Carmeliet and Tessier‐Lavigne, 2005; LeCouter et al., 2003). A particular concern for antiangiogenic therapy is the treatment of pediatric cancers. In developing children, the growth, remodeling, and maturation of various tissues and organs, and the maintenance of their normal functions are all dependent on appropriate development of the vasculature. Thus, antiangiogenic therapy in principle could adversely affect the development of various organs in children. At present, we know almost nothing about these and similar long‐term side effects of antiangiogenic drugs.

VI. MECHANISMS OF BROAD‐SPECTRUM ANGIOGENESIS INHIBITORS Broad‐spectrum angiogenesis inhibitors suppress angiogenesis by blocking general mechanisms that govern endothelial cell growth. In addition to intact protein molecules such as thrombospondin‐1 (TSP‐1), platelet factor‐4, and pigment epithelial cell‐derived factor (PEDF), most endogenous angiogenesis inhibitors are cryptic proteolytic fragments derived from large precursor molecules, which usually lack antiangiogenic activity (Cao, 1998; Folkman, 2006; Nyberg et al., 2005). For example, angiostatin, endostatin, arrestin, canstatin, kringle 5, and tumstatin are all proteolytic fragments. Interestingly, the generation of these proteolytic angiogenesis inhibitors are associated with tumors

122

Yihai Cao

(O’Reilly et al., 1994, 1997). However, it is unclear why tumors produce angiogenesis inhibitors and how the protease activity is regulated. These findings demonstrate that the proteolytic process is crucial for inhibition of angiogenesis. Unlike proangiogenic factors, the molecular mechanisms of endogenous angiogenesis inhibitors remain generally unknown although various endothelial cell surface molecules have been claimed to interact with these inhibitors (Folkman, 2006; Nyberg et al., 2005). Endostatin is probably the most well‐studied endogenous angiogenesis inhibitor that potently inhibits the growth of various tumors in mice. Both mouse and human studies demonstrate that endostatin produces virtually no toxicity after long‐term delivery. Many published reports show that endostatin mainly suppresses pathological angiogenesis and does not affect wound healing or reproduction (Folkman, 2006). However, its antitumor activity remains a controversial issue, which is probably due to differences in expression systems, protein folding and solubility, heparin‐binding affinities, zinc binding, dosages, bolus versus sustained delivery, and gene therapy versus protein therapy. Gene expression profile analysis shows that endostatin downregulates a number of angiogenic factors, including VEGF‐A and FGF‐2, and upregulates other known endogenous angiogenesis inhibitors such as maspin and TSP‐1 (Abdollahi et al., 2004). Among all proposed mechanisms, it is worth to mention that the integrins v 5 v 3 and 5 1 integrins have been linked to the antiangiogenic activity of endostatin (Nyberg et al., 2005). Because some of these integrins are selectively expressed in growing endothelial cells, it might explain the specific inhibitory activity of endostatin on angiogenic endothelial cells. The broad molecular targets of endostatin demonstrate that multiple signaling systems are involved in mediation of its antiangiogenic action (Fig. 3). Another interesting and related issue is that many known anticancer drugs, including celecoxib (COX‐2 inhibitor), prednisolone, cyclophosphamide, doxycyclin, and rosiglitazone, could increase the levels of endogenous angiogenesis inhibitors such as endostatin and TSP‐1.

VII. ANTIANGIOGENIC THERAPY

VS.

CHEMOTHERAPY

Chemotherapeutic drugs are broad‐spectrum antiproliferative agents that nonspecifically target almost all dividing cells (Fig. 2). Proliferating endothelial cells in the tumor environment are no exception, and are affected by chemotherapeutic drugs (Browder et al., 2000; Man et al., 2002). In fact, dividing endothelial cells seem more sensitive than tumor cells in response to chemotherapeutic agents (Kerbel and Kamen, 2004). Consistent with this notion, persistent delivery of low‐dose chemotherapeutic drugs (metronomic

123

Molecular Mechanisms and Therapeutic Development

Non-proteolytic inhibitors PF-4 Interleukin Interferons TIMPs PEDF TSP-1

Malignant tissue Tumor cells Stromal cells Inflammantory cells

Elastase MMPs Plasmin Serineprotease Collagenase Cathepsins

Non-malignant tissue Healthy and pathological

Proteolytic inhibitors Collagen Plasminogen Fibronectin Thrombin MMPs

Cylophosphamide Celecoxib

Doxycycline

Endostatin

Prednisolone

Rosiglitazone

Angiostatin Tumostatin Canstatin Arresten Kringle 5

Broad-spectrum angiogenises inhibitors

Angiogenic vessel FAK

V

EG

s

FGFR

a5b1

a Vb

3 a V

2 FR

b5

TN

FR

Ras Raf MAPKs

EC

Proliferation Migration Survival

Fig. 3 Possible mechanisms of endostatin and other broad‐spectrum angiogenesis inhibitors. Proteolytic and nonproteolytic endogenous angiogenesis inhibitors might interact with their endothelial cell surface receptors/binding molecules to inhibit new blood vessel growth.

chemotherapy) could shift the target from tumor cells to the endothelial cell compartment and convert a drug‐resistant tumor to become sensitive again in response to the same drug (Kerbel and Kamen, 2004). If a chemotherapeutic agent is sufficiently potent to suppress angiogenesis, why would angiogenesis inhibitors such as Avastin produce synergy with chemotherapy in prolonging

124

Yihai Cao

patient lives? Although the exact mechanism of synergism between antiangiogenesis and chemotherapy remains unknown, there are several speculative possibilities, which include: (1) Antiangiogenesis agents suppress the growth of tumor vessels whereas chemotherapy shows evidence of destruction of these vessels. Thus, the underlying mechanisms of these therapies are different although they both target the same compartment; (2) Antiangiogenic agents such as Avastin could potentially “normalize” tumor blood vessels, which usually consist of disorganized and leaky vessels. Normalization of the tumor vasculature could reduce IFP and increase blood perfusion, leading to increased chemotherapeutic drug delivery (Jain et al., 2006). In tumor models, normalization of tumor vessels could be achieved by Avastin. (3) Normalization of tumor vessels by antiangiogenic agents could inhibit metastasis by preventing dissemination of tumor cells into the circulation; (4) Antiangiogenic agents inhibit the growth of metastatic tumors or induce tumor dormancy, and micrometastases are more sensitive to chemotherapy (Naumov et al., 2006); (5) Antiangiogenic agents might potentially protect tissues or organs against the toxic side effects of chemotherapy.

VIII. PATIENT SELECTION AND TIMESCALE OF TREATMENT Almost all clinical trials for evaluation of therapeutic efficacy of antiangiogenic agents are conducted in patients with advanced stage cancers and metastases. For example, combinations of Avastin and chemotherapy were designed for the treatment of advanced colorectal, breast, and renal cancers (Ferrara and Kerbel, 2005). Advanced tumors usually produce multiple angiogenic factors while small tumor nodules at the early stage of malignancy might only express one or two angiogenic factors. For example, a breast cancer at the early stage might only predominantly express VEGF‐A, but switches on the expression of multiple other angiogenic factors during malignant progression (Relf et al., 1997). Thus, treatment of tumors with Avastin at early clinical stages would in principle be more effective while the treatment of advanced stage tumors might increase the risks of encountering drug resistance or escape. For patient selection, it would be desirable to analyze expression profiles and levels of angiogenic factors. For instance, Avastin would have particular therapeutic advantages for patients whose tumors selectively express high levels of VEGF‐A. However, the therapeutic efficacy of growth factor antagonists does not always correlate with the expression levels of their corresponding growth factors or signaling molecules. For example, the survival advantage of lung cancer patients subjected to antiepidermal growth factor (EGF) receptor therapy has negatively been correlated with EGFR expression (Sridhar et al., 2003). Another important issue is the order of drug

Molecular Mechanisms and Therapeutic Development

125

delivery. In combination therapy, it is unclear whether antiangiogenic agents should be used prior to or simultaneously with chemotherapy. If normalization of tumor vessels contributes to the underlying mechanism of antiangiogenic therapy, antiangiogenic drugs should be administrated before initiation of chemotherapy. As antiangiogenic drugs in principle only suppress but not destroy tumor vessels, relentless delivery of these drugs for the rest of patient lives would have to be considered. However, in combination with chemotherapy that might destroy the tumor vasculature, it is unclear what the optimal time window of antiangiogenic therapy would be.

IX. BIOMARKERS In general, reliable surrogate markers for patient selection, optimizing dosages, and monitoring the therapeutic efficacy of antiangiogenic drugs have been lacking. The levels of various tumor‐produced angiogenic factors, including VEGF‐A and FGF‐2, in the circulation and body fluids of cancer patients have been used as indicators of the progression of malignant diseases (Nguyen et al., 1994). However, VEGF‐A, FGF‐2, and most likely several other angiogenic factors have unusually short half‐lives in the range of minutes in the circulation, which makes it very difficult to estimate the actual expression of these growth factors in tumors. Because of their high affinity to heparin, the biological actions of these growth factors are mainly restricted to their sites of production through interactions with heparan sulfate proteoglycans in the extracellular matrix. Thus, probably the only available relevant surrogate marker for anti‐VEGF‐A therapy is alteration of the number of circulating endothelial progenitor cells (CEPCs) in peripheral blood (Shaked et al., 2005). CEPCs derived from bone marrow plays pivotal roles in contribution to tumor neovascularization, growth, and metastasis. VEGF‐A can remarkably increase the number of CEPCs in the blood and anti‐VEGF‐A agents could effectively block this vasculogenic effect, which correlates with progression of malignant progression. Interestingly, endostatin treatment of animals could also reduce the number of CEPCs, validating the reliability of this marker (Schuch et al., 2003).

X. FUTURE PERSPECTIVES Cancer is a complex disease that might not be cured by a single therapeutic agent. Combinations of various anticancer drugs that target different tumor compartments would in principle be more effective than single agents. Conventional anticancer therapies such as chemotherapy and radiotherapy

126

Yihai Cao

are generally toxic for multiple organs and tissues. In fact, it has been estimated that about 25% of cancer patients die of therapy‐related side effects but not of malignant disease. Thus, there is an imperative need to develop novel nontoxic therapeutic strategies that might replace conventional chemotherapy or radiotherapy. Alternatively, these novel agents could serve as adjuvant therapies that can potentiate therapeutic efficacies, increase tolerance, and/or reduce toxicities of chemotherapy and radiotherapy. Unlike conventional therapeutic drugs, angiogenesis inhibitors target relatively “normal” endothelial cells within the tumor tissue and they are nondestructive to the vasculature. Current guidelines of antiangiogenic cancer therapy recommend combinatorial therapy with chemotherapeutic agents and the efficacy of combination therapy is dictated as improvement of survival over monotherapy. Although the tumor vasculature seems to be affected by antiangiogenic drugs such as Avastin, there is no direct correlation between the degrees of tumor suppression and prolongation of patient survival. Despite several candidate mechanisms, including suppression of tumor vessel growth, normalization of the tumor vasculature, and reduction of the number of CEPCs, the exact mechanisms by which Avastin and other antiangiogenic agents such as endostatin improve patient survival are not fully understood. As these new drugs become available for the treatment of cancers and other human disorders, it is crucial to understand exactly how they work. Thus, future studies should be focused on mechanistic understanding of the molecular action of these antiangiogenic drugs in cancer patients. Selection of optimal drug combinations, cancer types, dosages, and patients are essential for designing successful clinical trials for antiangiogenic therapy. Our current knowledge on these issues is largely based on preclinical studies, which in most cases enroll tumor transplant models in mice. In general, mouse tumor models are artificially fast growing systems that do not adequately reflect human spontaneous malignancy. The tumor growth rate might be a critical factor that determines the architecture of the tumor vascular network. Although the mice bearing a rapidly growing highly angiogenic tumor might be more responsive to antiangiogenic agents, the spontaneous human tumors with a relatively low growth rate might be more resistant to antiangiogenic therapy. Thus, it seems that although an antiangiogenic agent may be effective for the treatment of mouse tumors, it may not necessarily be effective for the treatment of human cancers. For this reason, spontaneous mouse tumor models are more valuable and relevant for preclinical evaluation of the therapeutic efficacy of antiangiogenic drugs. There is an urgent need to identify biomarkers that allow us to select potentially responsive patents, to design the therapeutic protocols and to monitor the therapeutic efficacies. These potential markers could include alterations of gene expression, cell types, and protein levels in the body fluids

Molecular Mechanisms and Therapeutic Development

127

as well as the development of high‐resolution imaging techniques such as magnetic resonance or positron emission tomography imaging for noninvasive studies of the tumor vasculature. For designing optimal treatment regimens, it is expected that oncologists in the near future will in have multiple choices of antiangiogenic drugs to use in combination with various chemotherapeutic agents. In addition, combinations of antiangiogenic drugs targeting different signaling pathways could be therapeutically synergistic. The important issue is how to design these smart combinations for the future. Our deeper understanding of the antiangiogenic actions of these new drugs will be crucial for future therapeutic designs. Although most patients tolerate antiangiogenic therapy well, rare but severe toxicities, including hypertension, cardiovascular thrombosis, gastrointestinal perforation, delayed wound healing and hemorrhages, have been observed in Avastin‐treated patients. It is speculated that broad‐spectrum tyrosine kinase inhibitors also might cause broader side effects. It would be interesting to see if endogenous angiogenesis inhibitors such as endostatin cause any similar side effects in patients although preclinical studies suggest this type of broad‐spectrum endogenous inhibitors do not cause side effects in mice. In conclusion, the antiangiogenic drugs have become an important therapeutic element for the treatment of various types of human cancers. Combinations of different antiangiogenic drugs alone or with conventional chemotherapy may significantly extend the survival of cancer patients and approach a possible curative treatment.

ACKNOWLEDGEMENTS I thank Dr. Yuan Xue for making the artistic work, Drs. Ebba Bra˚kenhielm and Lasse Dahl Ejby Jensen for critical reading of the manuscript. Yihai Cao’s laboratory is supported by the Swedish Research Council, the Swedish Heart and Lung Foundation, the Swedish Cancer Foundation, the Karolinska Institute Fund, the EU integrated projects of Angiotargeting (Contract No. 504743), VascuPlug (Contract No. STRP 013811), and the So¨derberg Foundation. Y. Cao is supported by The Swedish Research Council.

REFERENCES Abdollahi, A., Hahnfeldt, P., Maercker, C., Grone, H. J., Debus, J., Ansorge, W., Folkman, J., Hlatky, L., and Huber, P. E. (2004). Endostatin’s antiangiogenic signaling network. Mol. Cell 13, 649–663. Betsholtz, C. (2004). Insight into the physiological functions of PDGF through genetic studies in mice. Cytokine Growth Factor Rev. 15, 215–228.

128

Yihai Cao

Bjorndahl, M., Cao, R., Eriksson, A., and Cao, Y. (2004). Blockage of VEGF‐induced angiogenesis by preventing VEGF secretion. Circ. Res. 94, 1443–1450. Blouw, B., Song, H., Tihan, T., Bosze, J., Ferrara, N., Gerber, H. P., Johnson, R. S., and Bergers, G. (2003). The hypoxic response of tumors is dependent on their microenvironment. Cancer Cell. 4, 133–146. Browder, T., Butterfield, C. E., Kraling, B. M., Shi, B., Marshall, B., O’Reilly, M. S., and Folkman, J. (2000). Antiangiogenic scheduling of chemotherapy improves efficacy against experimental drug‐resistant cancer. Cancer Res. 60, 1878–1886. Cao, R., Brakenhielm, E., Pawliuk, R., Wariaro, D., Post, M. J., Wahlberg, E., Leboulch, P., and Cao, Y. (2003). Angiogenic synergism, vascular stability and improvement of hind‐limb ischemia by a combination of PDGF‐BB and FGF‐2. Nat. Med. 9, 604–613. Cao, Y. (1998). Endogenous angiogenesis inhibitors: angiostatin, endostatin, and other proteolytic fragments. Prog. Mol. Subcell. Biol. 20, 161–176. Cao, Y. (2005). Opinion: emerging mechanisms of tumour lymphangiogenesis and lymphatic metastasis. Nat. Rev. Cancer 5, 735–743. Carmeliet, P. (2005). Angiogenesis in life, disease and medicine. Nature 438, 932–936. Carmeliet, P., Ferreira, V., Breier, G., Pollefeyt, S., Kieckens, L., Gertsenstein, M., Fahrig, M., Vandenhoeck, A., Harpal, K., Eberhardt, C., et al. (1996). Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 380, 435–439. Carmeliet, P., Moons, L., Luttun, A., Vincenti, V., Compernolle, V., De Mol, M., Wu, Y., Bono, F., Devy, L., Beck, H., et al. (2001). Synergism between vascular endothelial growth factor and placental growth factor contributes to angiogenesis and plasma extravasation in pathological conditions. Nat. Med. 7, 575–583. Carmeliet, P., and Tessier‐Lavigne, M. (2005). Common mechanisms of nerve and blood vessel wiring. Nature 436, 193–200. Casanovas, O., Hicklin, D. J., Bergers, G., and Hanahan, D. (2005). Drug resistance by evasion of antiangiogenic targeting of VEGF signaling in late‐stage pancreatic islet tumors. Cancer Cell 8, 299–309. Coultas, L., Chawengsaksophak, K., and Rossant, J. (2005). Endothelial cells and VEGF in vascular development. Nature 438, 937–945. Cursiefen, C., Chen, L., Borges, L. P., Jackson, D., Cao, J., Radziejewski, C., D’Amore, P. A., Dana, M. R., Wiegand, S. J., and Streilein, J. W. (2004). VEGF‐A stimulates lymphangiogenesis and hemangiogenesis in inflammatory neovascularization via macrophage recruitment. J. Clin. Invest. 113, 1040–1050. Dong, J., Grunstein, J., Tejada, M., Peale, F., Frantz, G., Liang, W. C., Bai, W., Yu, L., Kowalski, J., Liang, X., et al. (2004). VEGF‐null cells require PDGFR alpha signaling‐ mediated stromal fibroblast recruitment for tumorigenesis. Embo J. 23, 2800–2810. Dvorak, H. F. (2006). Discovery of vascular permeability factor (VPF). Exp. Cell Res. 312, 522–526. Ferrara, N., Carver‐Moore, K., Chen, H., Dowd, M., Lu, L., O’Shea, K. S., Powell‐Braxton, L., Hillan, K. J., and Moore, M. W. (1996). Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene. Nature 380, 439–442. Ferrara, N., Chen, H., Davis‐Smyth, T., Gerber, H. P., Nguyen, T. N., Peers, D., Chisholm, V., Hillan, K. J., and Schwall, R. H. (1998). Vascular endothelial growth factor is essential for corpus luteum angiogenesis. Nat. Med. 4, 336–340. Ferrara, N., and Kerbel, R. S. (2005). Angiogenesis as a therapeutic target. Nature 438, 967–974. Folkman, J. (1971). Tumor angiogenesis: therapeutic implications. N. Engl. J. Med. 285, 1182–1186. Folkman, J. (2002). Looking for a good endothelial address. Cancer Cell 1, 113–115.

Molecular Mechanisms and Therapeutic Development

129

Folkman, J. (2006). Antiangiogenesis in cancer therapy‐endostatin and its mechanisms of action. Exp. Cell Res. 312, 594–607. Garber, K. (2002). Angiogenesis inhibitors suffer new setback. Nat. Biotechnol. 20, 1067–1068. Garber, K. (2006). The second wave in kinase cancer drugs. Nat. Biotechnol. 24, 127–130. Gragoudas, E. S., Adamis, A. P., Cunningham, E. T., Jr., Feinsod, M., and Guyer, D. R. (2004). Pegaptanib for neovascular age‐related macular degeneration. N. Engl. J. Med. 351, 2805–2816. Hanahan, D., and Folkman, J. (1996). Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86, 353–364. Hida, K., Hida, Y., Amin, D. N., Flint, A. F., Panigrahy, D., Morton, C. C., and Klagsbrun, M. (2004). Tumor‐associated endothelial cells with cytogenetic abnormalities. Cancer Res. 64, 8249–8255. Hoch, R. V., and Soriano, P. (2003). Roles of PDGF in animal development. Development 130, 4769–4784. Holash, J., Davis, S., Papadopoulos, N., Croll, S. D., Ho, L., Russell, M., Boland, P., Leidich, R., Hylton, D., Burova, E., et al. (2002). VEGF‐Trap: A VEGF blocker with potent antitumor effects. Proc. Natl. Acad. Sci. USA 99, 11393–11398. Hurwitz, H., Fehrenbacher, L., Novotny, W., Cartwright, T., Hainsworth, J., Heim, W., Berlin, J., Baron, A., Griffing, S., Holmgren, E., et al. (2004). Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N. Engl. J. Med. 350, 2335–2342. Jain, R. K. (2005). Normalization of tumor vasculature: An emerging concept in antiangiogenic therapy. Science 307, 58–62. Jain, R. K., Duda, D. G., Clark, J. W., and Loeffler, J. S. (2006). Lessons from phase III clinical trials on anti‐VEGF therapy for cancer. Nat. Clin. Pract. Oncol. 3, 24–40. Kerbel, R. S., and Kamen, B. A. (2004). The anti‐angiogenic basis of metronomic chemotherapy. Nat. Rev. Cancer 4, 423–436. LeCouter, J., Moritz, D. R., Li, B., Phillips, G. L., Liang, X. H., Gerber, H. P., Hillan, K. J., and Ferrara, N. (2003). Angiogenesis‐independent endothelial protection of liver: Role of VEGFR‐1. Science 299, 890–893. Lee, C. G., Link, H., Baluk, P., Homer, R. J., Chapoval, S., Bhandari, V., Kang, M. J., Cohn, L., Kim, Y. K., McDonald, D. M., and Elias, J. A. (2004). Vascular endothelial growth factor (VEGF) induces remodeling and enhances TH2‐mediated sensitization and inflammation in the lung. Nat. Med. 10, 1095–1103. Levine, R. J., Maynard, S. E., Qian, C., Lim, K. H., England, L. J., Yu, K. F., Schisterman, E. F., Thadhani, R., Sachs, B. P., Epstein, F. H., et al. (2004). Circulating angiogenic factors and the risk of preeclampsia. N. Engl. J. Med. 350, 672–683. Li, J. L., and Harris, A. L. (2005). Notch signaling from tumor cells: a new mechanism of angiogenesis. Cancer Cell 8, 1–3. Lyden, D., Hattori, K., Dias, S., Costa, C., Blaikie, P., Butros, L., Chadburn, A., Heissig, B., Marks, W., Witte, L., et al. (2001). Impaired recruitment of bone‐marrow‐derived endothelial and hematopoietic precursor cells blocks tumor angiogenesis and growth. Nat. Med. 7, 1194–1201. Makino, Y., Cao, R., Svensson, K., Bertilsson, G., Asman, M., Tanaka, H., Cao, Y., Berkenstam, A., and Poellinger, L. (2001). Inhibitory PAS domain protein is a negative regulator of hypoxia‐inducible gene expression. Nature 414, 550–554. Man, S., Bocci, G., Francia, G., Green, S. K., Jothy, S., Hanahan, D., Bohlen, P., Hicklin, D. J., Bergers, G., and Kerbel, R. S. (2002). Antitumor effects in mice of low‐dose (metronomic) cyclophosphamide administered continuously through the drinking water. Cancer Res. 62, 2731–2735.

130

Yihai Cao

Maniotis, A. J., Folberg, R., Hess, A., Seftor, E. A., Gardner, L. M., Pe’er, J.,Trent, J. M., Meltzer, P. S., and Hendrix, M. J. (1999). Vascular channel formation by human melanoma cells in vivo and in vitro: Vasculogenic mimicry. Am. J. Pathol. 155, 739–752. Maynard, S. E., Min, J. Y., Merchan, J., Lim, K. H., Li, J., Mondal, S., Libermann, T. A., Morgan, J. P., Sellke, F. W., Stillman, I. E., et al. (2003). Excess placental soluble fms‐like tyrosine kinase 1 (sFlt1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia. J. Clin. Invest. 111, 649–658. Naumov, G. N., Bender, E., Zurakowski, D., Kang, S. Y., Sampson, D., Flynn, E., Watnick, R. S., Straume, O., Akslen, L. A., Folkman, J., and Almog, N. (2006). A model of human tumor dormancy: An angiogenic switch from the nonangiogenic phenotype. J. Natl. Cancer Inst. 98, 316–325. Nguyen, M., Watanabe, H., Budson, A. E., Richie, J. P., Hayes, D. F., and Folkman, J. (1994). Elevated levels of an angiogenic peptide, basic fibroblast growth factor, in the urine of patients with a wide spectrum of cancers. J. Natl. Cancer Inst. 86, 356–361. Nyberg, P., Xie, L., and Kalluri, R. (2005). Endogenous inhibitors of angiogenesis. Cancer Res. 65, 3967–3979. O’Reilly, M. S., Boehm, T., Shing, Y., Fukai, N., Vasios, G., Lane, W. S., Flynn, E., Birkhead, J. R., Olsen, B. R., and Folkman, J. (1997). Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 88, 277–285. O’Reilly, M. S., Holmgren, L., Shing, Y., Chen, C., Rosenthal, R. A., Moses, M., Lane, W. S., Cao, Y., Sage, E. H., and Folkman, J. (1994). Angiostatin: A novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 79, 315–328. Rafii, S., and Lyden, D. (2003). Therapeutic stem and progenitor cell transplantation for organ vascularization and regeneration. Nat. Med. 9, 702–712. Relf, M., LeJeune, S., Scott, P. A., Fox, S., Smith, K., Leek, R., Moghaddam, A., Whitehouse, R., Bicknell, R., and Harris, A. L. (1997). Expression of the angiogenic factors vascular endothelial cell growth factor, acidic and basic fibroblast growth factor, tumor growth factor beta‐1, platelet‐derived endothelial cell growth factor, placenta growth factor, and pleiotrophin in human primary breast cancer and its relation to angiogenesis. Cancer Res. 57, 963–969. Richardson, T. P., Peters, M. C., Ennett, A. B., and Mooney, D. J. (2001). Polymeric system for dual growth factor delivery. Nat. Biotechnol. 19, 1029–1034. Ryo, A., Uemura, H., Ishiguro, H., Saitoh, T., Yamaguchi, A., Perrem, K., Kubota, Y., Lu, K. P., and Aoki, I. (2005). Stable suppression of tumorigenicity by Pin1‐targeted RNA interference in prostate cancer. Clin. Cancer Res. 11, 7523–7531. Schlessinger, J. (2005). SU11248: Genesis of a new cancer drug. The Scientist Abstract‐GEOBASE 17. Schuch, G., Heymach, J. V., Nomi, M., Machluf, M., Force, J., Atala, A., Eder, J. P., Jr., Folkman, J., and Soker, S. (2003). Endostatin inhibits the vascular endothelial growth factor‐induced mobilization of endothelial progenitor cells. Cancer Res. 63, 8345–8350. Shaked, Y., Bertolini, F., Man, S., Rogers, M. S., Cervi, D., Foutz, T., Rawn, K., Voskas, D., Dumont, D. J., Ben‐David, Y., et al. (2005). Genetic heterogeneity of the vasculogenic phenotype parallels angiogenesis; Implications for cellular surrogate marker analysis of antiangiogenesis. Cancer Cell 7, 101–111. Soker, S., Miao, H. Q., Nomi, M., Takashima, S., and Klagsbrun, M. (2002). VEGF165 mediates formation of complexes containing VEGFR‐2 and neuropilin‐1 that enhance VEGF165‐receptor binding. J. Cell Biochem. 85, 357–368. Soker, S., Takashima, S., Miao, H. Q., Neufeld, G., and Klagsbrun, M. (1998). Neuropilin‐1 is expressed by endothelial and tumor cells as an isoform‐specific receptor for vascular endothelial growth factor. Cell 92, 735–745.

Molecular Mechanisms and Therapeutic Development

131

Sridhar, S. S., Seymour, L., and Shepherd, F. A. (2003). Inhibitors of epidermal‐growth‐factor receptors: a review of clinical research with a focus on non‐small‐cell lung cancer. Lancet Oncol. 4, 397–406. Sun, Y., Wang, J., Liu, Y., Song, X., Zhang, Y., Li, K., Zhu, Y., Zhou, Q., You, L., and Yao, C. (2005). Results of phase III trial of rh‐endostatin (YH‐16) in advanced non‐small cell lung cancer (NSCLC) patients. J Clin Oncol Abstract‐GEOBASE.

The Tumorigenicity of Human Embryonic Stem Cells Barak Blum and Nissim Benvenisty Stem Cells Unit, Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem 91904, Israel

I. Introduction II. Spontaneous and Experimental Teratomas and Teratocarcinomas A. Spontaneous Teratomas and Teratocarcinomas B. Experimental Teratomas, Teratocarcinomas and ES Cells C. Definition of Experimental Teratomas and Teratocarcinomas III. Cellular and Molecular Aspects of HESC Tumorigenicity A. In Vivo Differentiation of Embryonic Carcinoma Cells B. Culture Adaptation of HESCs In Vitro C. Molecular Biology of Culture Adaptation in HESCs D. Tumorigenicity of Nonadapted HESCs IV. HESC‐Induced Teratomas as a Model for Early Human Development A. Modeling Normal Embryogenesis B. Modeling Genetic Diseases C. Utilizing HESC‐Induced Teratomas as a Surrogate Human Environment for Cancer Research V. HESC‐Induced Teratomas as a Clinical Hurdle A. General Ablation of Teratoma Cells B. Differentiation to Eliminate Tumorigenic Cells C. Sorting for Nontumorigenic Populations or against Pluripotent Cells VI. Concluding Remarks References Human embryonic stem cells (HESCs) are the in vitro descendants of the pluripotent inner cell mass (ICM) of human blastocyst stage embryos. HESCs can be kept undifferentiated in culture or be differentiated to tissues representing all three germ layers, both in vivo and in vitro. These properties make HESC‐based therapy remarkably appealing for the treatment of various disorders. Upon transplantation in vivo, undifferentiated HESCs rapidly generate the formation of large tumors called teratomas. These are benign masses of haphazardly differentiated tissues. Teratomas also appear spontaneously in humans and in mice. When they also encompass a core of malignant undifferentiated cells, these tumors are defined as teratocarcinomas. These malignant undifferentiated cells are termed embryonic carcinoma (EC), and are the malignant counterparts of embryonic stem cells. Here we review the history of experimental teratomas and teratocarcinomas, from spontaneous teratocarcinomas in mice to induced teratomas by HESC transplantation. We then discuss cellular and molecular aspects of the tumorigenicity of HESCs. We also describe the utilization of HESC‐induced teratomas for the modeling of

Advances in CANCER RESEARCH Copyright 2008, Elsevier Inc. All rights reserved.

133

0065-230X/08 $35.00 DOI: 10.1016/S0065-230X(08)00005-5

134

Barak Blum and Nissim Benvenisty

early human embryogenesis and for modeling developmental diseases. The problem of HESC‐induced teratomas may also impede or prevent future HESC‐based therapies. We thus conclude with a survey of approaches to evade HESC‐induced tumor formation. # 2008 Elsevier Inc.

I. INTRODUCTION Human embryonic stem cells (HESCs) are pluripotent cells derived from the inner cell mass (ICM) of a human blastocyst stage embryo (Thomson et al., 1998). They are characterized by their ability to self‐renew by cellular divisions and their ability to differentiate to all somatic tissue of the embryo (pluripotency). HESCs grow in tightly packed colonies and, if supplied with specific culture requirements such as a supportive feeder layer (usually mitotically arrested mouse embryonic fibroblasts), can remain undifferentiated indefinitely. HESCs are also defined by the expression of a battery of typical genes, the most renown among them are Oct4, Nanog, Sox2, high telomerase activity, and typical cell surface markers such as SSEA3, SSEA4, TRA‐1–60, TRA‐1–81, and tissue‐specific alkaline phosphatase (Adewumi et al., 2007). HESCs currently open some of the most promising avenues in the field of regenerative medicine, and efforts are being made to establish HESC‐based therapy for various diseases such as Parkinson’s disease, heart failures, and diabetes (Blum and Benvenisty, 2005). Accordingly, reports on the successful differentiation of HESCs to CNS neurons, cardiomyocytes, insulin‐secreting cells, and many other cell types are rapidly accumulating (Blum and Benvenisty, 2005). HESCs can spontaneously differentiate in vitro in the form of embryoid bodies (EBs) (Itskovitz‐Eldor et al., 2000) (Fig. 1). However, HESCs ability to spontaneously differentiate is best manifested when these cells are transplanted in vivo into immunosuppressed mice, where they form typical gross looking tumors termed teratomas, in which the cells differentiate disorderedly to various tissue types of the embryo (Przyborski, 2005) (Fig. 1). This tumorigenic nature of HESCs is considered a major hurdle for their clinical utilization, but it can be valuable for other purposes, such as studying early human development. This tumorigenic nature of HESCs is also important for the assessment of the differentiation potential of newly derived pluripotent cells, since blastocyst injection of HESCs is obviously impractical (Lensch et al., 2007). Almost two decades before HESCs were first successfully derived from human embryos (Thomson et al., 1998), the first mouse embryonic stem (ES) cells were derived (Evans and Kaufman, 1981; Martin, 1981). These successful establishments of blastocyst‐derived ES cells are based on work previously performed on pluripotent cells that were isolated from teratocarcinomas

135

Tumorigenicity of HESCs

In vivo differentiation

1 cm Teratoma

In vitro differentiation

50 mM Embryoid bodies

Fig. 1 HESCs spontaneously differentiate in vitro in the form of EBs (right), and in vivo in the form of a teratoma (left).

(Andrews, 2002; Damjanov, 2005; Solter, 2006). These tumor cells are very close counterparts of HESCs because of their ability to self‐renew and to differentiate in culture. HESCs are unique in that they are tumorigenic yet perfectly normal in every other aspect. They thus can make an excellent tool for the understanding of tumorigenicity. Accordingly, HESC and many tumor cells hold some similarities, such as the aforementioned self‐renewal and undifferentiated phenotype, along with the expression of telomerase, the ability for in vivo angiogenesis, shortened cell cycle and, of course, the ability to generate tumors upon transplantation in vivo (Dreesen and Brivanlou, 2007). In this review we will discuss the relations between HESCs and their tumor counterparts, outlining the differences and similarities between them. We will also discuss cellular and molecular aspects of HESC tumorigenicity, the use of HESC‐induced tumors for studying human embryonic development, and conclude the discussion by reviewing some of the approaches for evading the appearance of these tumors in clinical utilization.

136

Barak Blum and Nissim Benvenisty

II. SPONTANEOUS AND EXPERIMENTAL TERATOMAS AND TERATOCARCINOMAS A. Spontaneous Teratomas and Teratocarcinomas Spontaneously occurring teratomas and teratocarcinomas represent a unique set of tumors. They are categorized among the group of germ cell tumors (GCTs), and are characterized by the presence of haphazardly arranged differentiated tissues representing the three embryonic germ layers. This points to their origin from a pluripotent precursor (Ulbright, 2005). GCTs appear both in gonadal and extragonadal sites along the body midline, and are classified into five pathological groups (Looijenga et al., 2007; Oosterhuis et al., 2007). Thus, type I GCTs are teratomas and yolk sac tumors of infants, which mostly occur on extragonadal sites; type II are seminomas and nonseminomas (amidst which are both teratomas and teratocarcinomas), which occurs mainly in the testes of young adult males, and also occasionally on extragonadal sites along the midline of the body (along the trail of germ cell migration during embryogenesis). Type III GCTs are spermatocytic seminomas, which occur only in the testes of adult males. Type IV are dermoid cysts, and type V are hydatiform moles, both occurring only in females. Hence, teratomas can be classified as type I or type II GCTs. On the basis of their site of appearance, the age and gender of the patient, the repertoire of the cells that comprise them, and the cytogenetic (i.e., karyotype) and epigenetic (i.e., genomic imprinting, gene expression) characteristics of the tumor, the supposed cell of origin of each of these neoplasms can be traced (Looijenga et al., 2007; Oosterhuis et al., 2007). According to this classification, the authors propose a different cell of origin for different tumor categories. For example, while type I teratomas are benign and possess normal diploid karyotype, teratomas and teratocarcinomas categorized within type II nonseminomas are aneuploid (Codesal et al., 1991; Mayer et al., 2003; Looijenga et al., 2007). It is thus hypothesized that teratomas classified as type I GCTs originate from a cell akin to ES cell whereas the teratomas or teratocarcinomas classified as type II GCTs originate from a cell closer to a primordial germ cell (PGC) (Looijenga et al., 2007).

B. Experimental Teratomas, Teratocarcinomas and ES Cells In 1954, Stevens and Little reported that a certain mouse strain, named 129, is specifically prone to develop testicular GCTs (Stevens and Little, 1954) (Fig. 2). Later, they and others were able to produce experimental

137

Tumorigenicity of HESCs

ICM/ES cells

Egg cylinder stage embryo

Genital ridge

1 cm

Spontaneous in human and mice

Fig. 2 Sources of teratoma formation in mouse and human. Teratomas can be experimentally induced from ES cells, whole embryos at the egg cylinder stage or genital ridges of embryos between E11 ando E13.5. Teratomas also occurred spontaneously in both mouse and human. Note the many differentiated structures in the histological section.

teratocarcinomas by transplanting the undifferentiated core of the tumor back into the mouse, demonstrating the pluripotent tumor initiating cell, termed embryonal carcinoma (EC) (Stevens, 1958; Kleinsmith and Pierce, 1964). EC cells were observed to be localized in small foci (“nests”) of morphologically undifferentiated embryonic like cells. Most EC cells were

138

Barak Blum and Nissim Benvenisty

demonstrated to have an abnormal karyotype, usually in the form of near diploid aneuploidity (McBurney, 1976; McBurney and Rogers, 1982). It was also observed that tumors that did not contain these EC nests were never transplantable, grew slower, and eventually ceased to grow (Stevens and Hummel, 1957; Stevens, 1959; Martin, 1975). On the other hand, tumor fractions containing EC cells were always transplantable. When single EC cells were transplanted, the resulting tumor was not composed of EC cells only, but was again a teratocarcinoma containing many differentiated tissues (Kleinsmith and Pierce, 1964), proving the pluripotency of the EC cells. Accordingly, some EC lines grew in culture as two distinct types of colonies. While one type had a fibroblastic phenotype and did not form tumors, the other type grew as tight colonies of undifferentiated cells that could give rise to both types upon clonality assays, and could form complete teratocarcinomas in vivo (Martin and Evans, 1974). Teratomas and teratocarcinomas were also experimentally formed by ectopic transplantations of normal mouse embryos or embryonic genital ridges to adult host, usually beneath the kidney capsule (Solter and Damjanov, 1979; Solter et al., 1970; Stevens, 1964, 1967, 1968; Stevens and Hummel, 1957) (Fig. 2), demonstrating the existence of tumor initiating cells also in normal embryos. These experimental teratomas were dependent upon the age of the transplanted embryos. Thus, teratomas could be obtained only from mouse embryos younger than 7 days (Damjanov et al., 1987) and from genital ridges of embryos between embryonic day 11 and 13.5 (Andrews, 2002). In common to these time points is the fact that after this time, no pluripotent cells are present in the transplant (presumably primitive ectoderm in the early embryos or PGCs in the genital ridges). EC cells were also isolated from spontaneous human GCTs and were characterized (Andrews et al., 1984). Notably, these cells differ in several aspects, mainly cell surface antigens, from mouse EC cells (Andrews, 2002; Andrews et al., 1987). In 1981, pluripotent ES cells were isolated from mouse embryos (Evans and Kaufman, 1981; Martin, 1981). These ES cells appeared similar to most mouse EC cells in their cell surface markers and in their growth requirements. However, differently from EC cells, the new ES cells were karyotypically normal, and displayed broader capacity to differentiate (Andrews, 2002; Bradley et al., 1984). Since 1995, ES cells were established from nonhuman primates (Thomson et al., 1995, 1996), and finally, in 1998, from human embryos (Thomson et al., 1998). These human ES cells were slightly different from mouse ES cells in their growth requirements and the expression of some, but not all, cell surface molecules, but resembled very closely the human EC cells. This was evident by the expression of specific pluripotency markers like SSEA3, SSEA4, TRA‐1–60, TRA‐1–81, telomerase, and alkaline phosphatase, and by the ability to form teratoma in immunodeficient mice (Thomson et al., 1998). In later years, human ES

Tumorigenicity of HESCs

139

and EC cells were also compared for gene expression using microarray analysis (Liu et al., 2006; Sperger et al., 2003), the results showing many similarities, but also noticeable differences. Thus, human ES cells are the nonmalignant equalities of human EC cells, whereas mouse ES cells are the nonmalignant counterparts of mouse EC cells. This difference in some patterns between mouse and human ES and EC cells, along with the notion that GCT pathogenesis probably differs between mouse and human (Clark, 2007; Oosterhuis and Looijenga, 2005; Walt et al., 1993), has led to the suggestion that HESCs may come of an utterly different origin than the mouse ES cells, and are not, after all, truly equivalent interspecies counterparts (Zwaka and Thomson, 2005). With some species, like the rat, for example, it has been consistently very hard to obtain ES cells from (Skreb and Svajger, 1975). Intriguingly, these species had also much less‐reported incidents of spontaneous GCTs (Damjanov, 1993; Damjanov et al., 1987).

C. Definition of Experimental Teratomas and Teratocarcinomas There is much confusion regarding the terminology of teratoma/teratocarcinoma in the experimental setting, partially owing to inconsistencies in the use of medical terminology (Damjanov and Andrews, 2007; Lensch and Ince, 2007). From a histopathological point of view, benign GCTs with differentiation to all embryonic germ layers are termed “teratomas.” These can be mature teratomas (which contain only mature, well‐differentiated tissues) or immature teratomas (which contain tissues of more embryonic, less‐differentiated nature). If the tumors also contain clusters of totally undifferentiated, highly malignant embryonic carcinoma (EC) cells, than they are defined as “teratocarcinomas” (Gonzalez‐Crussi, 1982; Pierce et al., 1960). The presence of EC cells is currently best detected histologically by their immunopositivity for the expression of Oct4 (Jones et al., 2004).

III. CELLULAR AND MOLECULAR ASPECTS OF HESC TUMORIGENICITY A. In Vivo Differentiation of Embryonic Carcinoma Cells As is known from every knockout mouse thus far made, normal euploid mouse ES cells lose their tumorigenicity by incorporation into age‐ complemented embryonic environment. This is conclusively manifested by their incorporation into the blastocyst to form completely normal, germ line

140

Barak Blum and Nissim Benvenisty

transmitting chimera (Bradley et al., 1984). Aneuploid mouse teratocarcinoma‐ derived EC cells, however, could not be completely reversed by injection to the blastocyst. It appeared that some of the chimera progeny develop tumors (mostly embryonic carcinoma) and are not germ line competent (Hochedlinger and Jaenisch, 2006; Papaioannou et al., 1975; Rossant and McBurney, 1982). In a straightforward approach, Blelloch et al. (2004) determined the developmental capacity of several mouse teratocarcinoma cell lines to contribute to a normal chimera. They have performed cloning by nuclear transfer of three mouse teratocarcinoma cell lines, and produced ES cells from the resulted blastocysts. Strikingly, the derived ES cells were identical in their differentiation potential to the parental EC cells, demonstrating that the reduced differentiation potential of an aneuploid EC cell cannot be overruled. In this experiment, the only EC line that could still contribute to all embryonic tissues was found to be karyotypicaly normal, and was thus an exception to the rule. Similarly, human EC lines vary in their differentiation capacity between partial pluripotency to complete nullypotecy (Andrews et al., 1987). Hence, aneupolidity in EC cells is deleterious to their ability to differentiate, and the more genomic alternation a cell has, the less differentiation capacity is displayed by this cell. More importantly, and as mentioned earlier, the aneuploidity of EC cells also reflects the malignant nature of teratocarcinomas.

B. Culture Adaptation of HESCs In Vitro When first isolated, HESCs were thought to maintain normal karyotype for many passages in vitro (Amit et al., 2000; Thomson et al., 1998). However, it has been subsequently found that karyotypic changes do occur, the rate of being attributed by many to be dependent on passage number. Gain of chromosomes 12, 17, and X is frequently observed, but other karyotypic changes have also been reported (Baker et al., 2007; Draper et al., 2004; Imreh et al., 2006). In some HESC lines that were grown extensively to high passage, small genomic aberrations that were not apparent on regular G banding karyotyping were discovered using genomic array methods (Maitra et al., 2005). These were also accompanied with aberrations in mitochondrial DNA and impaired imprinting, all previously associated with cancers. Enver et al. (2005) have compared gene expression in a single HESC line before and after culture adaptation. Their results indicate that insufficient X chromosome inactivation is another way by which culture‐adapted HESCs can virtually obtain an additional copy of this chromosome even without apparent genetic changes. Interestingly, the same karyotypic abnormalities that accompany culture‐ adapted HESCs are also frequent in human GCTs and human teratocarcinoma

Tumorigenicity of HESCs

141

cell lines (Almstrup et al., 2006; Andrews et al., 2005; Baker et al., 2007). Particularly observed in these tumors is the gain of an isochromosome 12p (Dal Cin et al., 1989; de Bruin et al., 1994; Speleman et al., 1990), but also the addition of the X and 17 chromosomes has been reported (Kraggerud et al., 2002; Looijenga et al., 1997). An interesting finding, further stressing the similarity between culture‐adapted HESCs and EC cells, is the discovery of Herszfeld et al. (2006) that a specific EC surface antigen, CD30, is exclusively expressed on culture‐adapted, karyotypically abnormal HESCs, but not euploid low passaged cells. Mouse ES cells were also reported to acquire chromosomal changes during high passage adaptation in culture. These changes were directly correlated to their differentiation potential, as cells harboring them were impaired in contributing to germ line transmitting chimera, compared to euploid counterparts (Liu et al., 1997; Longo et al., 1997). Blastocyst injection of normal or culture‐adapted HESCs to asses their differentiation potential could obviously not be performed. However, their capacity to differentiate and their degree of malignancy can be studied through the hisopathological examination of the tumors that they make in immunodeficient mice. As expected from aneuploid mouse EC and ES cells studies, it was shown that culture‐adapted HESCs that were injected into immunodeficient mice develop tumors of a less‐differentiated nature. This was evident by the less mature tissues in them and the detection of undifferentiated nests of cells resembling EC (Andrews et al., 2005) and was most prominent in the works of Herszfeld et al. (2006) and Plaia et al. (2006). They have compared the tumorigenicity of normal unadapted HESCs to the tumorigenicity of CD30‐positive HESCs harboring trisomy in various chromosomes (Herszfeld et al., 2006; Plaia et al., 2006) and found that the karyotypically abnormal cells generated tumors with much primitive, undifferentiated tissues. Plaia et al. (2006) also reports on clusters of Oct4 expressing cells within the tumor. It should be mentioned, however, that there is a single work which specifically reported that HESCs‐bearing trisomy of chromosome 12 did not exhibit more aggressive or less mature teratomas than wild type cells (Gertow et al., 2007), but the differentiation was somewhat biased toward mesodermal lineages.

C. Molecular Biology of Culture Adaptation in HESCs The karyotypic instability of HESCs in culture could be related to an uncoupling between the G2/M checkpoint and the apoptosis machinery. It was recently discovered by Mantel et al. (2007) that in both mouse and human ES cells the mitotic spindle checkpoint is functional, but fails to induce apoptosis upon cell division arrest. As a result, the cells obtain a

142

Barak Blum and Nissim Benvenisty

tetraploid karyotype, which is reduced to the typical near diploid aneuploidity in subsequent divisions. This may facilitate the selection of cells with higher proliferation rate or better propensity for self‐renewal, which may lessen their capacity to differentiate. ES cells lose this ability to survive mitotic arrest upon differentiation, and thus differentiated ES cells are destined to apoptosis upon mitotic checkpoint activation (Mantel et al., 2007). Numerous candidate genes, most of them located on the specific regions of chromosomes 12, 17, and X that are typically associated with culture adaptation were suggested to be involved in the tumorigenicity of the in vitro adapted HESCs. Recent reviews by Harrison et al. (2007) and Baker et al. (2007) inclusively lists important genes on these chromosomal loci, which are suspected to contribute to the in vivo tumorigenicity of culture‐adapted HESCs. One of the most interesting among these genes, located on chromosome 12, is Nanog. It resides on 12p13 (Clark et al., 2004) and is highly expressed in human GCTs and embryonic carcinoma (Hart et al., 2005). Overexpression of Nanog was found to prohibit the differentiation of mouse ES cells under feeder free conditions (Chambers et al., 2003; Mitsui et al., 2003). Similarly, overxpressing Nanog in HESCs allowed them to sustain an undifferentiated phenotype even without feeder cells (Darr et al., 2006). Oct4 is one of the most accurate diagnostic markers in the assessment of the malignancy of human GCTs, as it is specifically expressed in the undifferentiated EC core of teratocarcinomas (Jones et al., 2004). Oct4 was found to work together also with Nanog for the maintenance of pluripotency in ES cells (Wang et al., 2006). Previously, it was suggested that Oct4 itself is a driving force oncogene in mouse GCTs, and that overexpression of it can promote malignancy in mouse ES cells derived tumors (Gidekel et al., 2003). It is currently not known if Nanog and Oct4 are consequently detected in EC cells of teratocarcinomas due to the pluripotency of these cells or whether they are a driving force for tumorigenicity, but it is noteworthy that Nanog and Oct4 are among the few genes found to be required for reprogramming somatic human cells into pluripotent (and tumorigenic) fate (Yu et al., 2007). Another interesting gene is the apoptosis inhibitor, cell cycle regulator survivin. Located on 17q25, survivin is highly expressed in early embryonic tissues, but is virtually absent from adult, terminally differentiated tissues (Adida et al., 1998; Ambrosini et al., 1997; Islam et al., 2000; Li et al., 1998). However, it is also expressed in human testicular GCTs (Weikert et al., 2005) and many other cancers, and is thus seen as a tumorigenic marker. Also on the long arm chromosome 17, the gene BCAS3, recently reported to be expressed in HESCs and in several cancers (Siva et al., 2007), might be implicated in angiogenesis processes of the tumors. Another interesting gene that might be related to the tumorigenicity of HESCs is the human homologue of the mouse ERas. Located on the

Tumorigenicity of HESCs

143

X chromosome, this gene was thought to be a nonfunctional pseudogene called HRasp. However, Takahashi et al. (2003) have found that it is indeed expressed and functional in both mouse and human ES cells. When ERas was added to NIH3T3 fibroblasts, their tumorigenc transformation was promoted. Reciprocally, when mouse ES cells were deprived of ERas, their tumorigenicity was markedly reduced, although they could still form small teratomas. Interestingly, the protein was found to be constitutively active in normal diploid ES cells, but absent from differentiated cells.

D. Tumorigenicity of Nonadapted HESCs In contrast to less mature teratomas generated by culture‐adapted HESCs, nonadapted HESCs were thus far reported to exclusively make mature teratomas. This was determined histologically (Reubinoff et al., 2000), and in some cases by the absence of immunostaining for pluripotency markers like Oct4 (Blum and Benvenisty, 2007; Gertow et al., 2004) and TRA‐1–60, TRA‐1–81, and SSEA4 (Gertow et al., 2004). We have recently reported that BrdU incorporation in established tumors derived from unadapted HESCs demonstrate that within the teratoma all cell types are proliferative, and that no exclusive nests or foci of highly proliferating cells were observed (Blum and Benvenisty, 2007). In the same work we further demonstrated that the tumor as a whole is not clonally derived. Specific differentiated structures within the HESC‐induced teratoma were derived from different cells of origin, thus arguing against clonal selection of a more aggressive HESC clone that takes over the tumor. In contrast, clonality is sometime reported to occur in spontaneous human GCTs (Gillis et al., 1994; Rothe et al., 1999). In a wide ranging survey that was performed by the International Stem Cell Initiative (ISCI) in order to standardize HESC research (Adewumi et al., 2007) many HESC‐induced teratomas were analyzed. It was reported that three of the teratomas actually did include cells that somewhat look like the EC core of teratocarcinoma, but the nature of these EC‐like cells could not be determined, and they were suggested to represent either residual, undifferentiated HESCs, or true embryonic carcinoma cells originated from transformed HESCs. Transplantations of normal, euploid mouse ES cells into adult mice do result sometimes in the formation of tumors that were described as malignant tumors instead of teratomas (Gidekel et al., 2003). Also transplantation of normal mouse embryos gives rise to teratocarcinomas (Solter et al., 1970). The differences observed with human ES cells transplanted to the mouse might be attributed to differences between the transformation of mouse and human cells. Specifically, this may apply that in contrast to mouse ES cells, HESCs are not easily transformed in vivo, perhaps because human

144

Barak Blum and Nissim Benvenisty

cells in general are more resilient to transformation than are mouse cells (Rangarajan and Weinberg, 2003; Rangarajan et al., 2004). On the other hand, this difference seen between the tumorigenic capabilities of mouse and human ES cells may be attributed to the difference of species between the transplanted HESCs and the murine environment. It is reasonable to assume that HESC‐derived tumors in the mouse are facing suboptimal conditions regarding growth factors mismatch, difficulties in anastomosing with the host blood system and other discrepancies. Indeed, interspecies host difference in the tumorigenicity of mouse ES cells was demonstrated (Erdo et al., 2003). Erdo co‐workers have transplanted mouse ES cells into the brain of rat and mice hosts. Although mouse ES cells injected to rats hardly resulted in tumor formation, transplanting the same ES cells into mouse brains resulted in the formation of large, very aggressive teratocarcinomas. HESC teratoma experiments in human are obviously not feasible. However, one group has reported on experiments in nonhuman primates’ (i.e., cynomologus macaque) ES cells transplantation back into hosts of the same species (Asano et al., 2003; Shibata et al., 2006). To evade immune rejection in the host, the ES cells were injected in utero, using ultrasound guidance, into the liver or abdominal cavity of monkey fetuses. In these experiments almost all the fetuses developed tumors. The reported tumors were massive, and sometimes killed the host. However, it is not clear whether intraspecies transplantation of primate ES cells results only in benign teratoma or in teratocarcinoma like some of the mouse autologous experiments. Another approach to the question of tumorigenicity of HESCs within human hosts was the injection of HESCs directly into three different human fetal tissues (thymus, lung, and pancreas), which were transplanted earlier in SCID mice (Shih et al., 2007). Surprisingly, it was found that the HESCs transplanted into the human fetal tissues formed aggressive undifferentiated tumors, in which cells still expressed Oct4. In contrast, transplantation of the same cells into the hind leg or under the kidney capsule of the SCID mouse itself resulted in the formation of only mature teratomas. It is noteworthy that the two HESC lines that were used in this experiment displayed normal karyotype before transplantation and also after tumor formation, and did not express CD30. Nonetheless, it still may be that the fetal human tissues used in this experiment were for some unknown reason supportive to HESC self‐renewal (or suppressive for differentiation), similar to other embryonic tissues such as mouse embryonic fibroblasts and human embryonic fibroblasts. It is also interesting that not only species difference but also difference in anatomical location within the same host can be relevant to the tumorigenic outcome; HESCs that were injected into the liver formed tumors that looked more aggressive and less differentiated than those that formed when the same cells were transplanted subcutaneously (Cooke et al., 2006).

Tumorigenicity of HESCs

145

More provocatively, when HESCs were transplanted into the brain ventricle of neonatal mice, they differentiated to functional neurons and integrated into the host brain without teratoma formation (Muotri et al., 2005). This may be explained by the presence of an environment rich in developmental signals, thus directing the HESCs to take part in the surrounding tissue. The tumorigenic feature of HESCs may perhaps be explained by the fact that ES cells already normally possess some of the features considered necessary for tumor formation, such as immortality, telomerase expression, resistance to contact inhibition, and abrogated cell cycle. However, there is still much to discover regarding this unique phenomenon.

IV. HESC‐INDUCED TERATOMAS AS A MODEL FOR EARLY HUMAN DEVELOPMENT In line with the notion that normal HESCs can differentiate to form teratoma‐like tumors without being transformed, HESC‐derived tumors are seen by some not as a tumor at all, but rather as failed progress of normal embryonic development, due to the incorrect localization of the developing cells (Lensch and Ince, 2007; Lensch et al., 2007). In that context, this tumorigenic activity of HESCs can function as a very promising tool by which one can study the very early stages of human embryonic development. These are actually inaccessible to research, due to the unavailability of normal human embryos at theses stages, and the obvious ethical restriction regarding human experimentation. Moreover, as the early developmental stages of mouse and human embryos differ in many ways (Dvash and Benvenisty, 2004), the use of HESCs, as a tool for developmental studies, both of normal development and as a model for developmental diseases, is sometimes the only possibility. This is further stressed by the fact that in vitro differentiation of HESCs fails to reach the level of tissue complexity seen upon in vivo differentiation (Blum and Benvenisty, 2005).

A. Modeling Normal Embryogenesis Gertow et al. (2004) were the first to try and describe in detail the development of various tissues within HESC‐induced teratoma using systematic histological examinations and a large set of antibodies recognizing differentiated tissues. They report that within the teratoma, tissue types known to be inductive of each other were in close proximity. Furthermore, they observed that the degree of differentiation within the teratoma is higher in tumors that were grown for a longer time in the mouse. Intertissue induction was also

146

Barak Blum and Nissim Benvenisty

suggested from the observation by Lavon et al. (2004), who noticed that hepatic‐like cells were localized adjacent to cardiomyocytes in HESC‐induced teratomas. These observations suggested that the teratoma tissue is developing in a way that resembles normal events in the developing embryo. In order to assess this issue experimentally, we have performed experiments in which distinct differentiated structures were excised (using laser capture microdissection) from teratomas that were generated from a combination of three different HESC lines, and proved that they are formed by the assembly of cells from different origins, meaning that the teratoma environment is indeed inductive (Blum and Benvenisty, 2007). In contrast, Gerecht‐Nir et al. (2004), studying vasculogenesis in HESC‐ induced teratomas, have found that most of the blood vessels within the tumor originate from the murine host. They have stated that vascular development within HESC‐induced teratomas does not completely mimic normal human development and that, at least in this case, in vitro models might prove more useful.

B. Modeling Genetic Diseases Eiges et al. (2007) have recently isolated HESCs from embryos carrying the fragile X syndrome. These cells were used to study the silencing of FMR1. This model is very relevant, since mouse model fail to recapitulate the inactivation of FMR1 that occurs in the fragile X patients. Here, the use of teratomas induced from the mutated HESCs was instrumental in the study of the developmental silencing of FMR1 in neuronal cells. Recently, teratomas from mouse ES cells carrying an additional copy of human chromosome 21 were reported (Mensah et al., 2007). The study of such teratomas was aimed at understanding the means by which this specific trisomy affects neuronal differentiation in Down syndrome patients. Histological and immunohistochemical examination of these teratomas show less neuronal differentiation in the affected teratomas than in the parental cell line. However, a more convincing proof of this phenomenon in human Down syndrome will be to compare neuronal differentiation in teratomas from human embryos with Down syndrome that will be identified through PGD.

C. Utilizing HESC‐Induced Teratomas as a Surrogate Human Environment for Cancer Research Another creative use of HESC‐induced teratomas has been their use as a tool to study the relationship between normal human tissues and cancer development. These relationships are very difficult to study using

Tumorigenicity of HESCs

147

conventional mouse models. Thus, Tzukerman et al. (2003, 2006) established human teratomas in immunodeficient mice hosts, and then injected genetically labeled human cancer cells directly into the teratoma. By tracing the behavior of the caner cells within the teratomas, they extrapolated conclusions on the alleged behavior of these malignant cells in human environment.

V. HESC‐INDUCED TERATOMAS AS A CLINICAL HURDLE There are still several obstacles on the way to the implementation of HESCs in cellular therapy, the most prominent being immune rejection and the tumorigenicity of HESCs‐based tissues (Parson, 2006; Vogel, 2005). Recently, the issue of immune rejection of HESC‐based grafts has made enormous progress towards resolution with the report of two different approaches to generating patient‐specific pluripotent stem cells. First, Byrne et al. (2007) have succeeded in the cloning of a nonhuman primate, a task that has been considered technically impossible (Simerly et al., 2003). Second, Takahashi et al. (2007) and Yu et al. (2007) have independently reported the successful reprogramming of somatic human cells into pluripotent cells perfectly resembling HESCs. These scientific advancements have brought the concern regarding the tumorigenicity of the implanted cells into the prime spotlight. The formation of a teratoma as a clinical outcome of HESC transplantation in human patients is completely unacceptable, regardless of its perception as a less‐differentiated tumor from culture‐adapted cells or as a disorganized bulk of normal embryonic tissues. Accordingly, several strategies have been implemented in order to tackle the dangerous tumorigenic potential of HESC‐ induced transplants (Fig. 3). These have been thoroughly reviewed by Hentze et al. (2007), who comprehensively summarize and categorize the clinical hurdles facing utilization of HESC‐based grafts in the clinic. We will give general examples of approaches designed to evade teratoma formation in grafted HESCs.

A. General Ablation of Teratoma Cells To be able to control and ablate the transplanted cells, if a teratoma forms after grafting, Schuldiner et al. (2003) have genetically engineered a HESC line to carry as a transgene the viral thynidine kinase (HSV‐tk) gene. Upon treatment with ganciclovir, an antiviral drug designed to induce apoptosis in the presence of herpes simplex thynidine kinase, the cells carrying this so‐called “suicide gene” are eliminated. Using this system, Schuldiner et al.

148

Barak Blum and Nissim Benvenisty

Terminal differentiation/eliminating residual pluripotent cells

Pre-transplantation

Interfering with tumor progression genes

Tumor formation

Suicide genes

Tumor detection

Fig. 3 Strategies for preventing teratoma formation from HESC‐based grafts. (A) The cell culture can be terminally differentiated prior to transplantation. Residual tumorigenic cells can be excluded from the transplant using various sorting techniques. (B) Different mechanisms that are involved in teratoma formation can be targeted to produce HESCs which are intrinsically unable to generate tumors after transplantation. (C) Cells carrying a “suicide genes” can be eliminated after teratoma formation using specific drugs.

were able to significantly eradicate in vivo established tumors in mice burdening HESC‐induced teratomas. This pioneering approach with human ES cells was recently repeated with slight modifications using mouse ES cells. Thus, Cao et al. (2006) infected mouse ES cells with a lentivirus carrying a triple fusion transgene incorporating firefly luciferase, monomeric red fluorescence protein, and a truncated thymidine kinase. This enabled them to follow the kinetics of teratoma formation after injection of the cells into the myocardium of immunosuppressed rats. With this system, no tumors were created when the rats were treated with ganciclovir. In yet another report, Jung et al. infected mouse ES cells with a lentivirus carrying TK and green fluorescence protein. Again, they were able to completely eliminate teratomas arising from transplantation of the modified ES cells to the flank and into the CNS of SCID mice after treatment with ganciclovir (Jung et al., 2007). One drawback of this method is that by applying ganciclovir, all proliferating cells carrying the suicide gene are eliminated, thus while killing the tumor the normal differentiating cells will also be affected. Another drawback is that genetically modified cells are used, increasing their chance for transformation.

Tumorigenicity of HESCs

149

B. Differentiation to Eliminate Tumorigenic Cells As discussed earlier, neither euploid nor culture‐adapted HESCs generate teratomas after being terminally differentiated prior to transplantation. Hence, a large number of efforts are being made to completely differentiate HESCs before grafting them, and to eliminate residual pluripotent cells that may hide in the culture. This is especially important giving the fact that a work on mouse ES cells aimed at finding the minimal amount of ES cells sufficient to form a teratoma revealed that even single pluripotent ES cells are able to generate this tumor (Lawrenz et al., 2004). Most of the approaches thus far tested to prevent the formation of teratomas from both mouse and human ES cells were aimed at terminally differentiating them in order to get rid of the pluripotent tumorigenic population, but not all were reported successful. For example, Leor et al. (2007) have transplanted small pieces of HESCs derived beating cardiomyocytes into the heart of athymic nude rats. They report that the injected cells did not integrate fully into the infracted myocardium, but differentiated to various types of fibrotic and myofibrotoc tissues and a teratoma occurred in one of the animals. Similarly, Fujikawa et al. (2005) report on the development of a teratoma after transplantation of allegedly insulin‐secreting cells. Here, mouse ES cells were differentiated in vitro and were shown to express insulin mRNA as well as C peptide, and were able to complement hyperglycemia in the host for 3 weeks. Again, incomplete differentiation has led to the formation of a teratoma and to the end of the experiment. On the other hand, other methods of differentiation have proven more successful. Interestingly, these were usually designated to produce neurons, suggesting the possibility that the brain is either less permissive on teratoma formation (maybe owing to a more inductive environment), or that neurons are easier to differentiate purely. Two examples are works of Reubinoff et al. (2001) and Zhang et al. (2001), who succeeded, by different methods composed of a complex succession of steps, to generate neural precursors that did intercalate into mice brains, but did not create teratomas.

C. Sorting for Nontumorigenic Populations or against Pluripotent Cells Another approach is to specifically sort out residual pluripotent tumorigenic HESCs after differentiation, or sort for a desired progenitor population. Several years ago, a fluorescent marker under the regulation of a pluripotent marker (Rex‐1) was introduced into HESCs (Eiges et al., 2001). These new cell lines were fluorescent as undifferentiated cells and

150

Barak Blum and Nissim Benvenisty

lost their fluorescence upon differentiation. It was shown that the undifferentiated tumorigenic cells can be sorted out from a mixed population of HESCs and their differentiated progenies. Another effort to specifically ablate pluripotent cells after transplantation without genetic manipulation has been the use of ceramide analogues that specifically induce apoptosis in pluripotent cells (Bieberich et al., 2004). It was discovered that the expression of the PAR‐4 gene, whose encoded protein is involved in the apoptotic response of ES cells to ceramides during neural differentiation, is colocalized with the expression of the OCT4 gene in differentiating ES cells. By incubation of the differentiating culture with a specific ceramide analogue, residual pluripotent cells were indeed eliminated, resulting in no teratoma formation after transplantation of these cultures into mice. Another example is the work by Shibata et al. (2006). Here, cynomolgus monkey ES cells differentiated into hematopoietic cells generated teratoma upon in vivo transplantation into embryos of the same species. Sorting this population for SSEA4‐negative cells by means of flow cytometry resulted in a tumorigenic‐free population. Alternatively, sorting the differentiated population according to a tissue‐ specific marker can provide highly purified populations of nontumorigenic cells. Fukuda et al. (2006) and Chung et al. (2006) used FACS to purify neural progenitors for sox1GFP‐positive cells. These sox1‐positive cells engrafted well and did not form tumors, in contrast to the sox1GFP‐negative fraction, which was tumorigenic. Similarly, Barberi et al. (2007) succeeded in generating a pure population of nontomorigenic myoblasts by sorting for CD73þ/NCAMþ cells.

VI. CONCLUDING REMARKS Blastocyst‐derived ES cells are the in vitro counterparts of the malignant EC cells found in spontaneous teratocarcinomas. The study of HESCs has emerged from pivotal experiments and observations on teratocarcinoma and teratoma GCTs, and is now coming of age. Being the in vitro successors of the pluripotent ICM cells, it is hypothesized that the differentiation of nonadapted HESCs in vivo resembles normal embryonic processes, albeit in a disorganized manner. It is thus attractive to use HESC‐induced teratomas to study development in humans, as early stages of human embryo, especially around the time of implantation, are inaccessible. This is also true in the case of developmental disorders, where an animal model is sometimes insufficient. HESCs can acquire karyotypic changes in culture upon stressful prolonged growth. These culture‐adapted HESCs then resemble, karyotypically, malignant EC cells. Specific genes like Nanog and ERas, some of which are

Tumorigenicity of HESCs

151

reported to relate to cancer, are located on the chromosomes that are involved in this culture adaptation, and the adapted cells are suggested to generate less mature tumors. Taken together, the reports regarding the outcome of the transplantation of naı¨ve or culture‐adapted HESCs that were surveyed here indicate that there may actually be different levels of tumorigenicity displayed by HESCs. Thus, as naı¨ve, unadapted cells generate only mature teratomas, the tumorigenicity of culture‐adapted HESCs is less mature, and is suggested to be more aggressive. The ability of HESCs to differentiate has led to high expectations regarding the use of them in medicine. These are somewhat clouded by HESCs notorious habit to generate teratomas. Any of the techniques thus far described is not sufficient to completely preclude the formation HESC‐induced teratoma in human patients. However, methods that will aim against tumorigenic genes that are normally expressed in unadapted HESCs may prevent this process. Clinical translation of HESCs cannot tolerate any type of tumorigenicity of HESC‐based grafts. In view of the enormous progress in the field of regenerative medicine following the soon expected ability to generate patient‐specific pluripotent stem cells, there is an urgent need for more experiments that will solve the problem of HESC‐induced tumorigenicity.

ACNOWLEDGMENTS We thank Yoav Mayshar and Rina Klinov for critically reading the manuscript and Tamar Golan‐Lev for assistance with preparation of the figures. This work was partially supported by funds from Bereshit Consortium, the Israeli Ministry of Trade and Industry (Grant number 37675), and by the European Community (ESTOOLS, Grant number 018739).

REFERENCES Adewumi, O., Aflatoonian, B., Ahrlund‐Richter, L., Amit, M., Andrews, P. W., Beighton, G., Bello, P. A., Benvenisty, N., Berry, L. S., Bevan, S., et al. (2007). Characterization of human embryonic stem cell lines by the International Stem Cell Initiative. Nat. Biotechnol. 25, 803–816. Adida, C., Crotty, P. L., McGrath, J., Berrebi, D., Diebold, J., and Altieri, D. C. (1998). Developmentally regulated expression of the novel cancer anti‐apoptosis gene survivin in human and mouse differentiation. Am. J. Pathol. 152, 43–49. Almstrup, K., Sonne, S. B., Hoei‐Hansen, C. E., Ottesen, A. M., Nielsen, J. E., Skakkebaek, N. E., Leffers, H., and Rajpert‐De Meyts, E. (2006). From embryonic stem cells to testicular germ cell cancer—should we be concerned? Int. J. Androl. 29, 211–218. Ambrosini, G., Adida, C., and Altieri, D. C. (1997). A novel anti‐apoptosis gene, survivin, expressed in cancer and lymphoma. Nat. Med. 3, 917–921.

152

Barak Blum and Nissim Benvenisty

Amit, M., Carpenter, M. K., Inokuma, M. S., Chiu, C. P., Harris, C. P., Waknitz, M. A., Itskovitz‐Eldor, J., and Thomson, J. A. (2000). Clonally derived human embryonic stem cell lines maintain pluripotency and proliferative potential for prolonged periods of culture. Dev. Biol. 227, 271–278. Andrews, P. W. (2002). From teratocarcinomas to embryonic stem cells. Philos. Trans. R Soc. Lond. B Biol. Sci. 357, 405–417. Andrews, P. W., Damjanov, I., Simon, D., Banting, G. S., Carlin, C., Dracopoli, N. C., and Fogh, J. (1984). Pluripotent embryonal carcinoma clones derived from the human teratocarcinoma cell line Tera‐2. Differentiation in vivo and in vitro. Lab. Invest. 50, 147–162. Andrews, P. W., Matin, M. M., Bahrami, A. R., Damjanov, I., Gokhale, P., and Draper, J. S. (2005). Embryonic stem (ES) cells and embryonal carcinoma (EC) cells: Opposite sides of the same coin. Biochem. Soc. Trans. 33, 1526–1530. Andrews, P. W., Oosterhuis, J., and Damjanov, A. (1987). Cell lines from human germ cell tumors. In “Teratocarcinomas and Embryonic Stem Cells: A Practical Approach” (E. J. Robertson, Ed.), pp. 207–248. IRL Press, Oxford. Asano, T., Ageyama, N., Takeuchi, K., Momoeda, M., Kitano, Y., Sasaki, K., Ueda, Y., Suzuki, Y., Kondo, Y., Torii, R., Hasegawa, M., Ookawara, S., et al. (2003). Engraftment and tumor formation after allogeneic in utero transplantation of primate embryonic stem cells. Transplantation 76, 1061–1067. Baker, D. E., Harrison, N. J., Maltby, E., Smith, K., Moore, H. D., Shaw, P. J., Heath, P. R., Holden, H., and Andrews, P. W. (2007). Adaptation to culture of human embryonic stem cells and oncogenesis in vivo. Nat. Biotechnol. 25, 207–215. Barberi, T., Bradbury, M., Dincer, Z., Panagiotakos, G., Socci, N. D., and Studer, L. (2007). Derivation of engraftable skeletal myoblasts from human embryonic stem cells. Nat. Med. 13, 642–648. Bieberich, E., Silva, J., Wang, G., Krishnamurthy, K., and Condie, B. G. (2004). Selective apoptosis of pluripotent mouse and human stem cells by novel ceramide analogues prevents teratoma formation and enriches for neural precursors in ES cell‐derived neural transplants. J. Cell. Biol. 167, 723–734. Blelloch, R. H., Hochedlinger, K., Yamada, Y., Brennan, C., Kim, M., Mintz, B., Chin, L., and Jaenisch, R. (2004). Nuclear cloning of embryonal carcinoma cells. Proc. Natl. Acad. Sci. USA 101, 13985–13990. Blum, B., and Benvenisty, N. (2005). Differentiation in vivo and in vitro of human embryonic stem cells. In “Stem Cells: From Bench to Bedside” (A. Bongso and E. H. Lee, Eds.), pp. 123–143. World Scientific, Singapore. Blum, B., and Benvenisty, N. (2007). Clonal analysis of human embryonic stem cell differentiation into teratomas. Stem Cells 25, 1924–1930. Bradley, A., Evans, M., Kaufman, M. H., and Robertson, E. (1984). Formation of germ‐line chimaeras from embryo‐derived teratocarcinoma cell lines. Nature 309, 255–256. Byrne, J., Pedersen, D., Clepper, L., Nelson, M., Sanger, W., Gokhale, S., Wolf, D., and Mitalipov, S. (2007). Producing primate embryonic stem cells by somatic cell nuclear transfer. Nature. 450, 497–502. Cao, F., Lin, S., Xie, X., Ray, P., Patel, M., Zhang, X., Drukker, M., Dylla, S. J., Connolly, A. J., Chen, X., Weissman, I. L., Gambhir, S. S., et al. (2006). In vivo visualization of embryonic stem cell survival, proliferation, and migration after cardiac delivery. Circulation 113, 1005–1014. Chambers, I., Colby, D., Robertson, M., Nichols, J., Lee, S., Tweedie, S., and Smith, A. (2003). Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. Cell 113, 643–655.

Tumorigenicity of HESCs

153

Chung, S., Shin, B. S., Hedlund, E., Pruszak, J., Ferree, A., Kang, U. J., Isacson, O., and Kim, K. S. (2006). Genetic selection of sox1GFP‐expressing neural precursors removes residual tumorigenic pluripotent stem cells and attenuates tumor formation after transplantation. J. Neurochem. 97, 1467–1480. Clark, A. T. (2007). The stem cell identity of testicular cancer. Stem Cell Rev. 3, 49–59. Clark, A. T., Rodriguez, R. T., Bodnar, M. S., Abeyta, M. J., Cedars, M. I., Turek, P. J., Firpo, M. T., and Reijo Pera, R. A. (2004). Human STELLAR, NANOG, and GDF3 genes are expressed in pluripotent cells and map to chromosome 12p13, a hotspot for teratocarcinoma. Stem Cells 22, 169–179. Codesal, J., Paniagua, R., Regadera, J., Fachal, C., and Nistal, M. (1991). Significance of DNA quantification in testicular germ cell tumors. Andrologia 23, 381–385. Cooke, M. J., Stojkovic, M., and Przyborski, S. A. (2006). Growth of teratomas derived from human pluripotent stem cells is influenced by the graft site. Stem Cells Dev. 15, 254–259. Dal Cin, P., Drochmans, A., Moerman, P., and Van den Berghe, H. (1989). Isochromosome 12p in mediastinal germ cell tumor. Cancer Genet. Cytogenet. 42, 243–251. Damjanov, I. (1993). Teratocarcinoma: Neoplastic lessons about normal embryogenesis. Int. J. Dev. Biol. 37, 39–46. Damjanov, I. (2005). The road from teratocarcinoma to human embryonic stem cells. Stem Cell Rev. 1, 273–276. Damjanov, I., and Andrews, P. W. (2007). The terminology of teratocarcinomas and teratomas. Nat. Biotechnol. 25, 1212. Damjanov, I., Damjanov, A., and Solter, D. (1987). Production of teratocarcinomas from embryos transplanted to extra‐uterine sites. In “Teratocarcinomas and Embryonic Stem Cells: A Practical Approach” (E. J. Robertson, Ed.), pp. 1–18. IRL Press, Oxford. Darr, H., Mayshar, Y., and Benvenisty, N. (2006). Overexpression of NANOG in human ES cells enables feeder‐free growth while inducing primitive ectoderm features. Development 133, 1193–1201. de Bruin, T. W., Slater, R. M., Defferrari, R., Geurts van Kessel, A., Suijkerbuijk, R. F., Jansen, G., de Jong, B., and Oosterhuis, J. W. (1994). Isochromosome 12p‐positive pineal germ cell tumor. Cancer Res. 54, 1542–1544. Draper, J. S., Smith, K., Gokhale, P., Moore, H. D., Maltby, E., Johnson, J., Meisner, L., Zwaka, T. P., Thomson, J. A., and Andrews, P. W. (2004). Recurrent gain of chromosomes 17 q and 12 in cultured human embryonic stem cells. Nat. Biotechnol. 22, 53–54. Dreesen, O., and Brivanlou, A. H. (2007). Signaling pathways in cancer and embryonic stem cells. Stem Cell Rev. 3, 7–17. Dvash, T., and Benvenisty, N. (2004). Human embryonic stem cells as a model for early human development. Best Pract. Res. Clin. Obstet. Gynaecol. 18, 929–940. Eiges, R., Schuldiner, M., Drukker, M., Yanuka, O., Itskovitz‐Eldor, J., and Benvenisty, N. (2001). Establishment of human embryonic stem cell‐transfected clones carrying a marker for undifferentiated cells. Curr. Biol. 11, 514–518. Eiges, R., Urbach, A., Malcov, M., Frumkin, T., Schwartz, T., Amit, A., Yaron, Y., Eden, A., Yanuka, O., Benvenisty, N., and Ben‐Yosef, D. (2007). Developmental study of fragile X syndrome using human embryonic stem cells derived from preimplantation genetically diagnosed embryos. Cell Stem Cell 1, 568–577. Enver, T., Soneji, S., Joshi, C., Brown, J., Iborra, F., Orntoft, T., Thykjaer, T., Maltby, E., Smith, K., Dawud, R. A., Jones, M., Matin, M., et al. (2005). Cellular differentiation hierarchies in normal and culture‐adapted human embryonic stem cells. Hum. Mol. Genet. 14, 3129–3140.

154

Barak Blum and Nissim Benvenisty

Erdo, F., Buhrle, C., Blunk, J., Hoehn, M., Xia, Y., Fleischmann, B., Focking, M., Kustermann, E., Kolossov, E., Hescheler, J., Hossmann, K. A., and Trapp, T. (2003). Host‐ dependent tumorigenesis of embryonic stem cell transplantation in experimental stroke. J. Cereb. Blood Flow Metab. 23, 780–785. Evans, M. J., and Kaufman, M. H. (1981). Establishment in culture of pluripotential cells from mouse embryos. Nature 292, 154–156. Fujikawa, T., Oh, S. H., Pi, L., Hatch, H. M., Shupe, T., and Petersen, B. E. (2005). Teratoma formation leads to failure of treatment for type I diabetes using embryonic stem cell‐derived insulin‐producing cells. Am. J. Pathol. 166, 1781–1791. Fukuda, H., Takahashi, J., Watanabe, K., Hayashi, H., Morizane, A., Koyanagi, M., Sasai, Y., and Hashimoto, N. (2006). Fluorescence‐activated cell sorting‐based purification of embryonic stem cell‐derived neural precursors averts tumor formation after transplantation. Stem Cells 24, 763–771. Gerecht‐Nir, S., Osenberg, S., Nevo, O., Ziskind, A., Coleman, R., and Itskovitz‐Eldor, J. (2004). Vascular development in early human embryos and in teratomas derived from human embryonic stem cells. Biol. Reprod. 71, 2029–2036. Gertow, K., Cedervall, J., Unger, C., Szoke, K., Blennow, E., Imreh, M. P., and Ahrlund‐ Richter, L. (2007). Trisomy 12 in HESC leads to no selective in vivo growth advantage in teratomas, but induces an increased abundance of renal development. J. Cell Biochem. 100, 1518–1525. Gertow, K., Wolbank, S., Rozell, B., Sugars, R., Andang, M., Parish, C. L., Imreh, M. P., Wendel, M., and Ahrlund‐Richter, L. (2004). Organized development from human embryonic stem cells after injection into immunodeficient mice. Stem Cells Dev. 13, 421–435. Gidekel, S., Pizov, G., Bergman, Y., and Pikarsky, E. (2003). Oct‐3/4 is a dose‐dependent oncogenic fate determinant. Cancer Cell 4, 361–370. Gillis, A. J., Looijenga, L. H., de Jong, B., and Oosterhuis, J. W. (1994). Clonality of combined testicular germ cell tumors of adults. Lab. Invest. 71, 874–878. Gonzalez‐Crussi, F. (1982). “Exrtragonadal Teratomas.” Armed Forces Institute of Pathology, Washington, DC. Harrison, N. J., Baker, D., and Andrews, P. W. (2007). Culture adaptation of embryonic stem cells echoes germ cell malignancy. Int. J. Androl. 30, 275–281; discussion 281. Hart, A. H., Hartley, L., Parker, K., Ibrahim, M., Looijenga, L. H., Pauchnik, M., Chow, C. W., and Robb, L. (2005). The pluripotency homeobox gene NANOG is expressed in human germ cell tumors. Cancer 104, 2092–2098. Hentze, H., Graichen, R., and Colman, A. (2007). Cell therapy and the safety of embryonic stem cell‐derived grafts. Trends Biotechnol. 25, 24–32. Herszfeld, D., Wolvetang, E., Langton‐Bunker, E., Chung, T. L., Filipczyk, A. A., Houssami, S., Jamshidi, P., Koh, K., Laslett, A. L., Michalska, A., Nguyen, L., Reubinoff, B. E., et al. (2006). CD30 is a survival factor and a biomarker for transformed human pluripotent stem cells. Nat. Biotechnol. 24, 351–357. Hochedlinger, K., and Jaenisch, R. (2006). Nuclear reprogramming and pluripotency. Nature 441, 1061–1067. Imreh, M. P., Gertow, K., Cedervall, J., Unger, C., Holmberg, K., Szoke, K., Csoregh, L., Fried, G., Dilber, S., Blennow, E., and Ahrlund‐Richter, L. (2006). In vitro culture conditions favoring selection of chromosomal abnormalities in human ES cells. J. Cell Biochem. 99, 508–516. Islam, A., Kageyama, H., Takada, N., Kawamoto, T., Takayasu, H., Isogai, E., Ohira, M., Hashizume, K., Kobayashi, H., Kaneko, Y., and Nakagawara, A. (2000). High expression of Survivin, mapped to 17q25, is significantly associated with poor prognostic factors and promotes cell survival in human neuroblastoma. Oncogene 19, 617–623.

Tumorigenicity of HESCs

155

Itskovitz‐Eldor, J., Schuldiner, M., Karsenti, D., Eden, A., Yanuka, O., Amit, M., Soreq, H., and Benvenisty, N. (2000). Differentiation of human embryonic stem cells into embryoid bodies compromising the three embryonic germ layers. Mol. Med. 6, 88–95. Jones, T. D., Ulbright, T. M., Eble, J. N., Baldridge, L. A., and Cheng, L. (2004). OCT4 staining in testicular tumors: A sensitive and specific marker for seminoma and embryonal carcinoma. Am. J. Surg. Pathol. 28, 935–940. Jung, J., Hackett, N. R., Pergolizzi, R. G., Pierre‐Destine, L., Krause, A., and Crystal, R. G. (2007). Ablation of tumor‐derived stem cells transplanted to the central nervous system by genetic modification of embryonic stem cells with a suicide gene. Hum. Gene Ther. 18, 1182–1192. Kleinsmith, L. J., and Pierce, G. B., Jr. (1964). Multipotentiality of single embryonal carcinoma cells. Cancer Res. 24, 1544–1551. Kraggerud, S. M., Skotheim, R. I., Szymanska, J., Eknaes, M., Fossa, S. D., Stenwig, A. E., Peltomaki, P., and Lothe, R. A. (2002). Genome profiles of familial/bilateral and sporadic testicular germ cell tumors. Genes Chromosomes Cancer 34, 168–174. Lavon, N., Yanuka, O., and Benvenisty, N. (2004). Differentiation and isolation of hepatic‐like cells from human embryonic stem cells. Differentiation 72, 230–238. Lawrenz, B., Schiller, H., Willbold, E., Ruediger, M., Muhs, A., and Esser, S. (2004). Highly sensitive biosafety model for stem‐cell‐derived grafts. Cytotherapy 6, 212–222. Lensch, M. W., and Ince, T. A. (2007). The terminology of teratocarcinomas and teratomas. Nat. Biotechnol. 25, 1211. Lensch, M. W., Schlaeger, T. M., Zon, L. I., and Daley, G. Q. (2007). Teratoma formation assays with human embryonic stem cells: A rationale for one type of human–animal chimera. Cell Stem Cell 1, 253–258. Leor, J., Gerecht, S., Cohen, S., Miller, L., Holbova, R., Ziskind, A., Shachar, M., Feinberg, M. S., Guetta, E., and Itskovitz‐Eldor, J. (2007). Human embryonic stem cell transplantation to repair the infarcted myocardium. Heart 93, 1278–1284. Li, F., Ambrosini, G., Chu, E. Y., Plescia, J., Tognin, S., Marchisio, P. C., and Altieri, D. C. (1998). Control of apoptosis and mitotic spindle checkpoint by survivin. Nature 396, 580–584. Liu, X., Wu, H., Loring, J., Hormuzdi, S., Disteche, C. M., Bornstein, P., and Jaenisch, R. (1997). Trisomy eight in ES cells is a common potential problem in gene targeting and interferes with germ line transmission. Dev. Dyn. 209, 85–91. Liu, Y., Shin, S., Zeng, X., Zhan, M., Gonzalez, R., Mueller, F. J., Schwartz, C. M., Xue, H., Li, H., Baker, S. C., Chudin, E., Barker, D. L., et al. (2006). Genome wide profiling of human embryonic stem cells (hESCs), their derivatives and embryonal carcinoma cells to develop base profiles of U.S. Federal government approved HESC lines. BMC Dev. Biol. 6, 20. Longo, L., Bygrave, A., Grosveld, F. G., and Pandolfi, P. P. (1997). The chromosome make‐up of mouse embryonic stem cells is predictive of somatic and germ cell chimaerism. Transgenic Res. 6, 321–328. Looijenga, L. H., Gillis, A., Stoop, H., Hersmus, R., and Oosterhuis, W. (2007). Chromosomes and expression in human testicular germ cell tumors: Insight into their cell of origin and pathogenesis. Ann. N Y Acad. Sci. 1120, 187–214. Looijenga, L. H., Gillis, A. J., van Gurp, R. J., Verkerk, A. J., and Oosterhuis, J. W. (1997). X inactivation in human testicular tumors. XISTexpression and androgen receptor methylation status. Am. J. Pathol. 151, 581–590. Maitra, A., Arking, D. E., Shivapurkar, N., Ikeda, M., Stastny, V., Kassauei, K., Sui, G., Cutler, D. J., Liu, Y., Brimble, S. N., Noaksson, K., Hyllner, J., et al. (2005). Genomic alterations in cultured human embryonic stem cells. Nat. Genet. 37, 1099–1103. Mantel, C., Guo, Y., Lee, M. R., Kim, M. K., Han, M. K., Shibayama, H., Fukuda, S., Yoder, M. C., Pelus, L. M., Kim, K. S., and Broxmeyer, H. E. (2007). Checkpoint‐apoptosis uncoupling in human and mouse embryonic stem cells: A source of karyotpic instability. Blood 109, 4518–4527.

156

Barak Blum and Nissim Benvenisty

Martin, G. R. (1975). Teratocarcinomas as a model system for the study of embryogenesis and neoplasia. Cell 5, 229–243. Martin, G. R. (1981). Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc. Natl. Acad. Sci. USA 78, 7634–7638. Martin, G. R., and Evans, M. J. (1974). The morphology and growth of a pluripotent teratocarcinoma cell line and its derivatives in tissue culture. Cell 2, 163–172. Mayer, F., Stoop, H., Sen, S., Bokemeyer, C., Oosterhuis, J. W., and Looijenga, L. H. (2003). Aneuploidy of human testicular germ cell tumors is associated with amplification of centrosomes. Oncogene 22, 3859–3866. McBurney, M. W. (1976). Clonal lines of teratocarcinoma cells in vitro: Differentiation and cytogenetic characteristics. J. Cell Physiol. 89, 441–455. McBurney, M. W., and Rogers, B. J. (1982). Isolation of male embryonal carcinoma cells and their chromosome replication patterns. Dev. Biol. 89, 503–508. Mensah, A., Mulligan, C., Linehan, J., Ruf, S., O’Doherty, A., Grygalewicz, B., Shipley, J., Groet, J., Tybulewicz, V., Fisher, E., Brandner, S., and Nizetic, D. (2007). An additional human chromosome 21 causes suppression of neural fate of pluripotent mouse embryonic stem cells in a teratoma model. BMC Dev. Biol. 7, 131. Mitsui, K., Tokuzawa, Y., Itoh, H., Segawa, K., Murakami, M., Takahashi, K., Maruyama, M., Maeda, M., and Yamanaka, S. (2003). The homeoprotein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells. Cell 113, 631–642. Muotri, A. R., Nakashima, K., Toni, N., Sandler, V. M., and Gage, F. H. (2005). Development of functional human embryonic stem cell‐derived neurons in mouse brain. Proc. Natl. Acad. Sci. USA 102, 18644–18648. Oosterhuis, J. W., and Looijenga, L. H. (2005). Testicular germ‐cell tumours in a broader perspective. Nat. Rev. Cancer 5, 210–222. Oosterhuis, J. W., Stoop, H., Honecker, F., and Looijenga, L. H. (2007). Why human extragonadal germ cell tumours occur in the midline of the body: Old concepts, new perspectives. Int. J. Androl. 30, 256–263; discussion 263–264. Papaioannou, V. E., McBurney, M. W., Gardner, R. L., and Evans, M. J. (1975). Fate of teratocarcinoma cells injected into early mouse embryos. Nature 258, 70–73. Parson, A. (2006). The long journey from stem cells to medical product. Cell 125, 9–11. Pierce, G. B., Jr., Dixon, F. J., Jr., and Verney, E. L. (1960). Teratocarcinogenic and tissue‐ forming potentials of the cell types comprising neoplastic embryoid bodies. Lab. Invest. 9, 583–602. Plaia, T. W., Josephson, R., Liu, Y., Zeng, X., Ording, C., Toumadje, A., Brimble, S. N., Sherrer, E. S., Uhl, E. W., Freed, W. J., Schulz, T. C., Maitra, A., et al. (2006). Characterization of a new NIH‐registered variant human embryonic stem cell line, BG01V: A tool for human embryonic stem cell research. Stem Cells 24, 531–546. Przyborski, S. A. (2005). Differentiation of human embryonic stem cells after transplantation in immune‐deficient mice. Stem Cells 23, 1242–1250. Rangarajan, A., Hong, S. J., Gifford, A., and Weinberg, R. A. (2004). Species‐ and cell type‐ specific requirements for cellular transformation. Cancer Cell 6, 171–183. Rangarajan, A., and Weinberg, R. A. (2003). Opinion: Comparative biology of mouse versus human cells: Modelling human cancer in mice. Nat. Rev. Cancer 3, 952–959. Reubinoff, B. E., Itsykson, P., Turetsky, T., Pera, M. F., Reinhartz, E., Itzik, A., and Ben‐Hur, T. (2001). Neural progenitors from human embryonic stem cells. Nat. Biotechnol. 19, 1134–1140. Reubinoff, B. E., Pera, M. F., Fong, C. Y., Trounson, A., and Bongso, A. (2000). Embryonic stem cell lines from human blastocysts: Somatic differentiation in vitro. Nat. Biotechnol. 18, 399–404.

Tumorigenicity of HESCs

157

Rossant, J., and McBurney, M. W. (1982). The developmental potential of a euploid male teratocarcinoma cell line after blastocyst injection. J. Embryol. Exp. Morphol. 70, 99–112. Rothe, M., Albers, P., and Wernert, N. (1999). Loss of heterozygosity, differentiation, and clonality in microdissected male germ cell tumours. J. Pathol. 188, 389–394. Schuldiner, M., Itskovitz‐Eldor, J., and Benvenisty, N. (2003). Selective ablation of human embryonic stem cells expressing a “suicide” gene. Stem Cells 21, 257–265. Shibata, H., Ageyama, N., Tanaka, Y., Kishi, Y., Sasaki, K., Nakamura, S., Muramatsu, S., Hayashi, S., Kitano, Y., Terao, K., and Hanazono, Y. (2006). Improved safety of hematopoietic transplantation with monkey embryonic stem cells in the allogeneic setting. Stem Cells 24, 1450–1457. Shih, C. C., Forman, S. J., Chu, P., and Slovak, M. (2007). Human embryonic stem cells are prone to generate primitive, undifferentiated tumors in engrafted human fetal tissues in severe combined immunodeficient mice. Stem Cells Dev. 16, 893–902. Simerly, C., Dominko, T., Navara, C., Payne, C., Capuano, S., Gosman, G., Chong, K. Y., Takahashi, D., Chace, C., Compton, D., Hewitson, L., and Schatten, G. (2003). Molecular correlates of primate nuclear transfer failures. Science 300, 297. Siva, K., Venu, P., Mahadevan, A., Shankar, K. S., and Inamdar, M. S. (2007). Human BCAS3 expression in embryonic stem cells and vascular precursors suggests a role in human embryogenesis and tumor angiogenesis. PLoS ONE 2, e1202. Skreb, N., and Svajger, A. (1975). Experimental teratomas in rats. In “Teratomas and Differentiation” (M. I. Sherman and D. Solter, Eds.), pp. 83–97. Academic Press, New York. Solter, D. (2006). From teratocarcinomas to embryonic stem cells and beyond: A history of embryonic stem cell research. Nat. Rev. Genet. 7, 319–327. Solter, D., and Damjanov, I. (1979). Teratocarcinomas rarely develop from embryos transplanted into athymic mice. Nature 278, 554–555. Solter, D., Skreb, N., and Damjanov, I. (1970). Extrauterine growth of mouse egg‐cylinders results in malignant teratoma. Nature 227, 503–504. Speleman, F., De Potter, C., Dal Cin, P., Mangelschots, K., Ingelaere, H., Laureys, G., Benoit, Y., Leroy, J., and Van Den Berghe, H. (1990). i(12 p) in a malignant ovarian tumor. Cancer Genet. Cytogenet. 45, 49–53. Sperger, J. M., Chen, X., Draper, J. S., Antosiewicz, J. E., Chon, C. H., Jones, S. B., Brooks, J. D., Andrews, P. W., Brown, P. O., and Thomson, J. A. (2003). Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors. Proc. Natl. Acad. Sci. USA 100, 13350–13355. Stevens, L. C. (1958). Studies on transplantable testicular teratomas of strain 129 mice. J. Natl. Cancer Inst. 20, 1257–1275. Stevens, L. C. (1959). Embryology of testicular teratomas in strain 129 mice. J. Natl. Cancer Inst. 23, 1249–1295. Stevens, L. C. (1964). Experimental production of testicular teratomas in mice. Proc. Natl. Acad. Sci. USA 52, 654–661. Stevens, L. C. (1967). Origin of testicular teratomas from primordial germ cells in mice. J. Natl. Cancer Inst. 38, 549–552. Stevens, L. C. (1968). The development of teratomas from intratesticular grafts of tubal mouse eggs. J. Embryol. Exp. Morphol. 20, 329–341. Stevens, L. C., and Hummel, K. P. (1957). A description of spontaneous congenital testicular teratomas in strain 129 mice. J. Natl. Cancer Inst. 18, 719–747. Stevens, L. C., and Little, C. C. (1954). Spontaneous testicular teratomas in an inbred strain of mice. Proc. Natl. Acad. Sci. USA 40, 1080–1087. Takahashi, K., Mitsui, K., and Yamanaka, S. (2003). Role of ERas in promoting tumour‐like properties in mouse embryonic stem cells. Nature 423, 541–545.

158

Barak Blum and Nissim Benvenisty

Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., and Yamanaka, S. (2007). Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell. 131, 861–872. Thomson, J. A., Itskovitz‐Eldor, J., Shapiro, S. S., Waknitz, M. A., Swiergiel, J. J., Marshall, V. S., and Jones, J. M. (1998). Embryonic stem cell lines derived from human blastocysts. Science 282, 1145–1147. Thomson, J. A., Kalishman, J., Golos, T. G., Durning, M., Harris, C. P., Becker, R. A., and Hearn, J. P. (1995). Isolation of a primate embryonic stem cell line. Proc. Natl. Acad. Sci. USA 92, 7844–7848. Thomson, J. A., Kalishman, J., Golos, T. G., Durning, M., Harris, C. P., and Hearn, J. P. (1996). Pluripotent cell lines derived from common marmoset (Callithrix jacchus) blastocysts. Biol. Reprod. 55, 254–259. Tzukerman, M., Rosenberg, T., Ravel, Y., Reiter, I., Coleman, R., and Skorecki, K. (2003). An experimental platform for studying growth and invasiveness of tumor cells within teratomas derived from human embryonic stem cells. Proc. Natl. Acad. Sci. USA 100, 13507–13512. Tzukerman, M., Rosenberg, T., Reiter, I., Ben‐Eliezer, S., Denkberg, G., Coleman, R., Reiter, Y., and Skorecki, K. (2006). The influence of a human embryonic stem cell‐derived microenvironment on targeting of human solid tumor xenografts. Cancer Res. 66, 3792–3801. Ulbright, T. M. (2005). Germ cell tumors of the gonads: A selective review emphasizing problems in differential diagnosis, newly appreciated, and controversial issues. Mod. Pathol. 18(Suppl 2), S61–S79. Vogel, G. (2005). Cell biology. Ready or not? Human ES cells head toward the clinic. Science 308, 1534–1538. Walt, H., Oosterhuis, J. W., and Stevens, L. C. (1993). Experimental testicular germ cell tumorigenesis in mouse strains with and without spontaneous tumours differs from development of germ cell tumours of the adult human testis. Int. J. Androl. 16, 267–271. Wang, J., Rao, S., Chu, J., Shen, X., Levasseur, D. N., Theunissen, T. W., and Orkin, S. H. (2006). A protein interaction network for pluripotency of embryonic stem cells. Nature 444, 364–368. Weikert, S., Schrader, M., Krause, H., Schulze, W., Muller, M., and Miller, K. (2005). The inhibitor of apoptosis (IAP) survivin is expressed in human testicular germ cell tumors and normal testes. Cancer Lett. 223, 331–337. Yu, J., Vodyanik, M. A., Smuga‐Otto, K., Antosiewicz‐Bourget, J., Frane, J. L., Tian, S., Nie, J., Jonsdottir, G. A., Ruotti, V., Stewart, R., Slukvin, II, and Thomson, J. A. (2007). Induced pluripotent stem cell lines derived from human somatic cells. Science. 318, 1917–1920. Zhang, S. C., Wernig, M., Duncan, I. D., Brustle, O., and Thomson, J. A. (2001). In vitro differentiation of transplantable neural precursors from human embryonic stem cells. Nat. Biotechnol. 19, 1129–1133. Zwaka, T. P., and Thomson, J. A. (2005). A germ cell origin of embryonic stem cells? Development 132, 227–233.

Contact Interactions Between Cells That Suppress Neoplastic Development: Can They Also Explain Metastatic Dormancy? Harry Rubin Department of Molecular and Cell Biology, Life Sciences Addition, University of California, Berkeley, CA 94720‐3200

I. II. III. IV. V. VI. VII. VIII. IX. X. XI.

XII. XIII. XIV.

Introduction Suppression of Transformation Among Fibroblasts Suppression of Transformation Among Epithelial Cells Is GJC Required in Cell–Cell Suppression of Tumor Development? The Role of Plasma Membrane Activity in Regulation of Cell Growth Suppressive Effects of Mesenchymal Tissue on Normal and Neoplastic Epithelial Proliferation The Prototype of Progression to Metastasis as seen in Human Malignant Melanoma Characteristics of Cultured Human Melanocytes Isolated from Different Stages of Melanoma Progression Is there a Relationship Between the Cell Contact Interactions that Suppress Neoplastic Development and the Phenomenon of Metastatic Dormancy? Characteristics of Metastatic Dormancy Tumor Cell Adhesion to Cells in Distant Organs A. Endothelial Cells B. Parenchymal Cells Possible Alternative Explanations of Metastatic Dormancy Molecular Basis of Cell–Cell Adhesion Conclusions References

A comprehensive listing with accompanying discussions is given for established cases of interactions between normal and neoplastic cells of the same histotype that suppress neoplastic development. General principles that apply to the process are: (a) the requirement for a large excess of normal cells in direct contact with the neoplastic cells; (b) the effectiveness of suppression decreases with the degree of malignant progression of the neoplastic cells; and (c) the transformability of normal cells decreases under long‐term negative selection, which also increases their contact suppression of neoplastic cells. Although suppression requires adhesive contact, it does not require gap junction communication, and it represents the first line of defense against tumor development. In contrast, potentially metastatic cells released from primary carcinomas into the circulation are activated to multiply when they form heterotypic adhesions with

Advances in CANCER RESEARCH Copyright 2008, Elsevier Inc. All rights reserved.

159

0065-230X/08 $35.00 DOI: 10.1016/S0065-230X(08)00006-7

160

Harry Rubin

endothelial and parenchymal cells of distant organs. The great majority of the disseminated cancer cells (DCCs) fail to develop heterotypic cells adhesions in distant organs, and remain metastatically dormant as single cells. The lack of growth factors for the DCC in foreign territory also contributes to metastatic dormancy. The major insights about suppression of neoplastic development by homotypic contact arose from strictly operational experiments with living cells in culture. Molecular characterization of the cell–cell adhesions that underlie neoplastic suppression and metastasis activation has had only limited success, probably because of the complex variety of molecules involved. Hence, a program is outlined for further operational experiments on cell–cell interactions in tumor suppression to deepen our understanding of the neoplastic process, and provide possible avenues for its control. # 2008 Elsevier Inc.

I. INTRODUCTION The concepts of initiation and promotion were adopted to characterize the observations that the single application of a chemical carcinogen to the skin of mice produced no neoplastic lesion, but the repeated application to the initiated area of a promoting agent any time thereafter resulted in the appearance of multiple papillomas of epidermal origin, some of which progressed to carcinomas (reviewed in Rubin, 2001, 2003). Related observations have been made in a wide variety of organs in experimental animals and in the development of human neoplasms (Pitot, 2002). The generally accepted implication of these observations is that the initiating treatment induces tumor‐related mutations in epidermal cells that retain their normal phenotype until their microenvironment was perturbed by promoters, which are themselves nonmutagenic. With the advent of molecular genetics it was shown that tumor‐related mutations occur even in normal untreated tissues (Cha et al., 1994; Kasami et al., 1997; Tsai et al., 1996). The first hint of the microenvironmental relations that might underlie the suppression of initiated cells came from studies of fibroblasts in culture which maintained their normal morphology and growth behavior despite infection with tumor‐inducing viruses (Rubin, 1960a,b; Stoker, 1964; Stoker et al., 1966). Those studies showed that contact of solitary infected cells with confluent, contact‐inhibited cultures of normal fibroblasts prevents, and even reverses, the transformation of the infected cells. There followed many other examples of suppression of transformation in fibroblasts, but the full significance of these observations was not appreciated until the same kind of contact suppression was reported with initiated epidermal cells surrounded by normal keratinocytes (Hennings et al., 1990; Strickland et al., 1992), which provided an explanation for the ubiquitous occurrence of initiated cells in diverse epithelial tissues (Rubin, 2006, 2007). Recently, additional, important examples of contact‐related suppression of neoplasia by normal homotypic cells have come to light, which are included in the comprehensive review of the field that forms the first part of the present article.

Cell–Cell Contact in Suppression of Tumor Development

161

The foregoing results raised the question whether contact suppression of the neoplastic state might play a role in the phenomenon of metastatic dormancy. In metastatic dormancy, cells released into the circulation from primary tumors lodge in distant organs, but fail to form metastatic tumors there even though a very small fraction of cells from the same clonal population might manage to do so. There are, of course, differences from the contact suppression of initiated cells. Cells released into the circulation from primary tumors have not only lost their capacity to be regulated by homotypic normal cells, but are no longer retained even within their homotypic primary tumor. Any suppression provided by contact with normal cells in distant tumors would have to be not only heterotypic, but diversely heterotypic to be effective in every tissue the disseminated tumor cells encounter. Despite these reservations, I made a literature search of the conditions that underlie metastatic dormancy, which are discussed in the second part of this article. Not surprisingly, the metastatic literature is mainly concerned with the local conditions that enhance rather than suppress metastatic growth, but absence of, or failure to respond to those conditions might help to understand the origins of metastatic dormancy. The findings indicate that adhesive interactions of circulating tumor cells with heterotypic cells of distant organs promote metastasis formation, which is the opposite of the homotypic cell interactions that suppress primary tumor development. In addition to the adhesive interactions, the absence of appropriate soluble growth factors in distant organs could contribute to metastatic dormancy. The basic biological principle of order in the large over heterogeneity in the small which governs primary tumor development (Rubin, 2006, 2007) appears to be inoperative in the development of metastases. Further exploration of the principle of ordered heterogeneity in both situations is suggested to clarify their operational relationships.

II. SUPPRESSION OF TRANSFORMATION AMONG FIBROBLASTS The first indication of the nature of tumor suppression came from studies on the interactions between normal and neoplastic cells of fibroblastic origin (Table I). Chick embryo fibroblasts (CEFs), infected with a low dose of the Bryan strain of Rous sarcoma virus (RSV) within 24 h after seeding of the cells, would undergo neoplastic transformation resulting in their continued proliferation to form discrete, multilayered foci after the surrounding noninfected cells had stopped proliferating. The number of transformed cells in a culture could be assayed by their capacity to initiate foci when mixed in suspension with a large excess of uninfected cells (Rubin, 1960a).

162 Table I

Harry Rubin Suppression of Transformation Among Fibroblasts

Source of neoplastic cells RSV‐infected 1
CEF

Polyoma virus‐infected line of hamster fibroblasts Chemically transformed line of 10T1/2 mouse fibroblasts 5 sublines of an REF line, each transformed by a different oncogene Spontaneously transformed NIH 3T3 line of mouse fibroblasts Mouse glial cells transformed by transfected viral oncogene Spontaneously transformed Balb/c 3T3 mouse fibroblasts Spontaneously transformed Balb/c 3T3 mouse fibroblasts

Normalizing conditions

References

Normal CEF Normal CEF in high [CS] or moderate [FBS] Normal hamster or mouse fibroblasts Non‐transformed 10T1/2 line of 10T1/2 mouse fibroblasts

Rubin (1960a,b)

Early passage REF and an established REF line

Martin et al. (1991)

Sublines of NIH 3T3 mouse fibroblasts

Rubin (1994)

Non‐transformed glial cells

Alexander et al. (2004)

Self‐normalization of the transformed cells at very high density Lowered Mg2þ concentration in medium

Rubin and Chu (1982)

Stoker (1964, 1967) Mehta et al. (1986)

Rubin (1982)

CEF, chick embryo fibroblasts; CS, calf serum; FBS, fetal bovine serum; REF, rat embryo fibroblasts.

If, however, a culture of noninfected CEF was allowed to grow to confluence and undergo contact inhibition before adding cells already transformed at 3 days after RSV infection, no transformed foci appeared. The results suggested that the transformed cells were normalized and their growth suppressed by contact with the growth‐inhibited, confluent monolayer of normal cells. This conclusion was supplemented by the observation that the development of transformed foci was suppressed when RSV was added at the time of seeding the normal cells in the standard transformation assay for RSV, if the medium contained unusually high concentrations of calf serum (15–20%, v/v) or conventional concentrations of fetal bovine serum (5–10%, v/v) (Rubin, 1960b). Suppression of transformation occurred only when the concentrations of RSV were low enough to form discrete, countable foci surrounded by normal cells. If high enough concentrations of RSV were added to infect most of the cells, the entire culture underwent transformation regardless of how high the serum concentration was. Growth curves of

Cell–Cell Contact in Suppression of Tumor Development

163

the number of infected cells assayed from serum‐suppressed cultures revealed that they assumed the regulatory behavior of normal uninfected cells since they stopped proliferating within a day after the culture as a whole did. In contrast, the infected cells in permissively low concentrations of serum showed no such constraint at confluence. The regulated growth of infected cells in the first case was attributed to direct interaction between them and the surrounding normal cells, which was enhanced by high concentrations of serum. The serum effect was later interpreted as inhibition of proteases released at the cell surface (Rubin, 1970). This interpretation was supported by the digestion of fibrin by Rous sarcomas in plasma clots (Fischer, 1946), the identification of a surface protease of RSV‐ transformed cells in monolayer culture (Unkeless et al., 1973), and the capacity of proteases to stimulate the mitotic cycle of CEFs (Sefton and Rubin, 1970). These findings were extended to mammalian cells in experiments using a clone of established hamster fibroblasts transformed by infection with polyoma virus (Stoker, 1964). Colonies of the clone exhibited random orientation and unrestricted growth when in contact with one another on a bare surface, or when growing amid a low density of normal hamster or mouse fibroblasts. They made no visible colonies or foci, however, when seeded on a contact‐inhibited layer of normal fibroblasts (Stoker et al., 1966). There was transfer of hypoxanthine from a dense sheet of normal mouse embryo cells to polyoma‐transformed hamster cells, raising the possibility that there was also transfer of growth‐regulatory molecules (Stoker, 1967). That possibility seemed to gain support from experiments mainly with the 10T1/2 line of mouse fibroblasts transformed by methylcholanthrene, which were suppressed after seeding on top of a confluent layer of the original nontransformed 101/2 line (Mehta et al., 1986). The suppression was correlated with gap junction communication (GJC) as probed by dye transfer between the normal and transformed cells. A variety of treatments were applied to enhance or inhibit GJC among the cell combinations. It was concluded that heterologous GJC among the cells is required to suppress growth of the transformed cells. The generality of this conclusion was questioned in studies with five sublines of an immortal line of rat fibroblasts, each transformed by a different oncogene and cocultured with either the original nontransformed line, or early passages of normal rat embryo fibroblasts (REFs) (Martin et al., 1991). There was no correlation of transformed cell suppression with GJC. Furthermore, total inhibition of fluorescent dye transfer between normal and transformed cells failed to relieve the growth suppression of any of the transformed cell populations in combination with either of the two nontransformed populations. The results are obviously inconsistent with those of the previous paragraph (Mehta et al., 1986).

164

Harry Rubin

Spontaneous transformation occurs in the NIH 3T3 line of mouse fibroblasts by selection for overgrowth in serial rounds of prolonged confluence (Rubin, 1994). In contrast, negative selection apparently occurred in about 330 short‐ term, serial passages of the original NIH 3T3 cells at very low density and gave rise to a subline that resisted transformation. Moreover, confluent cultures of the subline suppressed the proliferation, and normalized the morphology of transformed cells derived from another subline. In contrast, susceptible sublines that had undergone only 30–60 low‐density passages permitted continuous expansion of transformed foci. The suppression by the higher passage cells did not assert itself until several days after they became confluent, indicating that the establishment of strong contact inhibition of those cells with each other was required before they could inhibit the transformed cells. When that happened, the small colonies of transformed cells that had already formed blended in with the nontransformed background, indicating that the cells had taken on a normal fibroblastic appearance, and were themselves inhibited from further proliferation. However, they resumed their transformed phenotype when assayed on a permissive background. Fibroblasts, presumably of glial origin, were obtained from the brains of mice that had a homozygous null connexin 43 knockout, which substantially reduced GJC (Alexander et al., 2004). A group of these cells was transformed by transfection with the src kinase gene. Growth and transformed morphology of the transformed cells were suppressed by contact with confluent monolayers of the nontransformed fibroblasts even when the already low GJC was further reduced by an inhibitor. Not only was there no dye transfer between cells, but electrical conductance between most of the heterologous pairs was also eliminated. The results indicated that significant GJC was not required for contact normalization, and that other intercellular junctions mediated this growth‐regulatory process. Pure cultures of transformed mouse fibroblasts rapidly deplete their medium when grown at high density with the conventional ratio of cells to medium, even when the medium is changed daily. However, if the transformed cells were restricted to the area of a small coverslip in a large culture dish with more than ample medium, they multiplied to a 10 times higher saturation density than do nontransformed cells, and they became contact‐ inhibited in the absence of nontransformed cells (Rubin and Chu, 1982). In doing so, the cells took on the appearance of a multilayered culture of normal cells, and displayed a markedly reduced rate of DNA synthesis. They had a greatly reduced capacity to produce colonies when subcultured in agar, and retained the flattened appearance of normal cells for about 1 day when reseeded at low density on plastic before resuming their transformed appearance. This homologous contact normalization indicates that there is no need for the transfer of regulatory molecules from normal cells for the transformed cells to assume a normal phenotype.

165

Cell–Cell Contact in Suppression of Tumor Development

Among other changes in the normalized cells was a reduction in the intracellular content of Mg2þ. Indeed, normalization could be brought about in less crowded cultures of the transformed cells by drastically lowering the Mg2þ concentration of the medium with a consequent reduction of the Mg2þ content of the cells (Rubin, 1982). The combined results were consistent with a central role of Mg2þ in the coordinate control of cell proliferation (Rubin, 1975, 2005).

III. SUPPRESSION OF TRANSFORMATION AMONG EPITHELIAL CELLS Cancers of epithelial origin or carcinomas constitute over 90% of human neoplasia (Cairns, 1978). The first indication of tumor suppression among epithelial cells came from studies of the comparative growth of hyperplastic alveolar nodules (HANs), which are precursors of mammary cancer in mice, after injection into intact mammary gland fat pads versus injection into fat pads that had been cleared of mammary epithelium (Table II; Faulkin and

Table II

Suppression of Transformation Among Epithelial Cells

Source of neoplastic cells

Normalizing cells

Hyperplastic mouse mammary alveolar epithelium

Normal mammary epithelium

Preneoplastic and neoplastic rat tracheal epithelium Neoplastic mouse keratinocytes Neoplastic human keratinocytes

Normal tracheal epithelium Normal keratinocytes

a

Human melanoma cells

Normal human keratinocytes

Rat hepatocarcinoma cells

Normal human keratinocytes Rat liver in vivo

Neoplastic imaginal disc epithelium of Drosophila

Normal imaginal disc epithelium

References DeOme et al. (1978), Faulkin and DeOme (1960), Medina et al. (1978) Gillett et al. (1989), Terzaghi‐Howe (1987) Hennings et al. (1990, 1992), Strickland et al. (1992) Alt‐Holland et al. (2005), Javaherian et al. (1998), Mudgil et al. (2003) Hsu et al. (2000b) Coleman et al. (1993), McCullough et al. (1998) Bilder (2004), Brumby and Richardson (2003)

aTransfected with and overexpressing E‐cadherin. Although melanocytes originate in the neural ridge and migrate to the epidermis during development, they and their neoplastic derivatives are considered here along with epithelial cells because they are mainly situated in the basal layer of the epidermis in humans, and are regulated by contact with keratinocytes.

166

Harry Rubin

DeOme, 1960). The transplanted HAN grew much more rapidly in the gland‐free fat pads than in the intact fat pads, indicating that normal mammary epithelium suppressed growth of the hyperplastic tissue. Unlike the suppression of neoplastic growth among fibroblasts (Table I) which required direct contact between cells, the suppression of the HANs occurred at a distance, suggesting the release of inhibitory substances from the normal mammary epithelium. There was similar evidence for the inhibition of mammary carcinoma growth by diffusible substances from normal epithelium but it was less effective than the inhibition of HAN growth. Such diffusible inhibitors are also responsible for maintaining the distance between normal mammary tubules (Faulkin and DeOme, 1960) and between branches of tubules (Nelson et al., 2006). The cells of HANs are also subject to growth inhibition by contact with normal mammary epithelium. The mixture of enzymatically dissociated HAN cells with an excess of dissociated normal mammary epithelial cells from virgin, pregnant, or lactating mice significantly decreased the tumorigenicity of the HAN cells upon injection into the gland‐free fat pad (Medina et al., 1978). Further evidence of contact inhibition of hyperactive growth came from experiments in which mammary epithelial cells from mammary cancer‐prone mice were removed, dissociated, and injected into the gland‐ free fat pads of young mice. HANs were produced within 2–3 months in the injected mice whereas 8–9 months were required for their appearance in the undisturbed host (DeOme et al., 1978). If the mammary epithelium was transplanted as intact pieces, each containing many cells, the onset of HANs was delayed over that produced by the dissociated cells (Medina et al., 1978), supporting a suppressive role of direct cell–cell interaction. Studies were later done on the suppressive effect of normal tracheal epithelium on carcinogen‐altered rat tracheal epithelial cells (Terzaghi‐ Howe, 1987). Ten thousand or more normal tracheal epithelial cells were inoculated with a 2‐ to 100‐fold excess of a carcinogen‐exposed aneuploid neoplastic cell line into tracheas in which the resident epithelia had been destroyed, and the tracheas were then transplanted subcutaneously into syngeneic hosts. The regenerated epithelium comprised normal diploid epithelial cells (almost entirely), and no tumors developed. Substitution of esophageal epithelial cells for normal tracheal cells also suppressed growth of the neoplastic tracheal cells. In contrast, tumors developed quickly upon adding neoplastic cells alone to deepithelialized tracheas, or to a small scratched area of otherwise intact tracheal epithelium, so the neoplastic cells in both cases occupied a contiguous surface of submucosa. The suppressive effect of the normal epithelial cells therefore appeared to require contact with the neoplastic line. The results differed from those of the aforementioned mixing experiments in which suppression was obtained with a minority of normal cells rather than an excess, which suggested that

Cell–Cell Contact in Suppression of Tumor Development

167

the normal cells attached more firmly to the denuded tracheal submucosa and displaced adjacent neoplastic cells. Preferential attachment might also explain the heterotypic suppression of tracheal neoplasia by a minority of esophageal cells. Mixing experiments were done with the same deepithelialized tracheal technique, but using different lines of preneoplastic and neoplastic tracheal cells (Gillett et al., 1989). The results were complex and irregular, but significant suppression occurred only with 100‐fold or greater excesses of normal to neoplastic cells, and suppression was uniformly successful only with the preneoplastic cells. The variable extent of suppression among the cell lines suggested that they represented different stages of neoplasia. It was surmised that progression of the neoplastic state decreases the susceptibility of the neoplastic cells to suppression by normal cells. Studies of the mixtures in cell cultures rather than the deepithelialized trachea identified neoplastic cells that persisted even where no lesions could be seen in the trachea. This suggested that, unlike the earlier experiments in suppression of tracheal neoplasia (Terzaghi‐Howe, 1987), the neoplastic cells were not eliminated by the normal cells, but were kept in a dormant state. Some of the most clearly defined and informative experiments on suppression of epithelial tumor development by normal epithelial cells were done on epidermal neoplasia, in part because they involved questions generated by the exhaustively studied in vivo system of initiation and promotion. Mouse keratinocytes derived from untreated skin terminally differentiate in cell culture without feeder layers in medium containing more than 0.1 mM calcium, but evade differentiation and multiply rapidly in lower concentrations (Hennings et al., 1980). Keratinocytes cultured from skin exposed to carcinogens in vivo yield occasional foci of proliferating cells that are resistant to calcium‐induced terminal differentiation (Kulesz‐Martin et al., 1980). These calcium‐resistant foci were not tumorigenic when first formed, but some cell lines derived from them progressed to tumor formation (Kulesz‐Martin et al., 1983). Colony formation in calcium by such cells was markedly delayed when they were cocultured with a large excess of normal keratinocytes, but the delay was shortened in the continuous presence of a promoting agent (Hennings et al., 1990). In contrast, there was no inhibition of colony formation from the initiated cells by adding a large excess of fibroblasts, or growing them in medium that had been conditioned on keratinocyte cultures. Similarly, only normal keratinocytes inhibited tumor formation on mouse skin by grafts of cells from papillomas that had been induced in the initiation– promotion procedure (Strickland et al., 1992). However, the keratinocytes failed to inhibit the growth of malignant carcinoma cells when the two cell types were mixed in skin grafts. A line of initiated keratinocytes that multiplied in high calcium, but formed normal epidermis when grafted in vivo, had lost

168

Harry Rubin

the capacity to inhibit tumor development by papilloma cells. This loss of inhibitory capacity for papilloma development is potentially of great significance because it suggests that the apparently normal epidermal cells that constitute the cellular microenvironment surrounding erstwhile tumorigenic cells in carcinogen‐initiated skin may be more permissive for tumor growth than are keratinocytes from untreated skin. It is noteworthy that there was no GJC between normal keratinocytes and the papilloma cells whose growth they inhibited (Hennings et al., 1992), indicating that contact between the normal and benign tumor cells is sufficient to suppress growth of the latter. An organotypic reconstruction of human skin in culture has been used to study the effect of normal keratinocytes on the growth of low‐grade malignant epidermal cells (Javaherian et al., 1998). Ratios of normal to malignant cells of 4:1 or higher eliminated the malignant cells from the organotypic cultures by displacing them from the basal layer and leading them to terminal differentiation. In a 1:1 ratio, the malignant cells persisted in expanding foci from which they eventually invaded the collagen substratum. The elimination of the malignant cells in the excess of normal keratinocytes, however, is not a model for the indefinite persistence of initiated cells in the skin, nor does it correspond to the failure of keratinocytes to suppress carcinoma cell proliferation in vivo (Strickland et al., 1992). Indeed, when the collagen substratum for the epidermis in organotypic cultures was covered with human basement membrane proteins, the low‐grade malignant cells were not eliminated, but invaded the substratum (Alt‐Holland et al., 2005). The results indicated that the normal keratinocytes outcompeted the neoplastic cell for attachment to collagen but not to basement membrane. The situation resembles that of the tracheal carcinoma cells that were suppressed by even a small minority of normal tracheal epithelium as a result of competition for attachment to the mesenchyme of deepithelialized tracheas (Terzaghi‐Howe, 1987). Epidermal tumors mainly induced by ultraviolet light are the most common form of neoplasia in humans (Ziegler et al., 1993). Clones of keratinocytes carrying mutations in the p53 tumor suppressor gene (TSG) occur at high frequency in the skin and can involve as much as 4% of the sun‐exposed epidermis of humans (Jonason et al., 1996). Sunlight acts both as a tumorigenic mutagen and a tumor promoter which favors clonal expansion of p53‐ mutated cells. Stem cell components act as physical barriers to clonal expansion of the p53 mutant keratinocyte (Zhang et al., 2001). Sustained ultraviolet irradiation enables the p53 mutant keratinocyte to expand to adjacent stem cell compartments without incurring an additional mutation. Studies in mice indicate that apoptosis of normal keratinocytes in the stem cell compartment will drive clonal expansion of the p53‐mutated clones (Zhang et al., 2005). Mixtures of p53‐mutated intraepithelial human tumor cells and normal keratinocytes were used to fabricate organotypic

Cell–Cell Contact in Suppression of Tumor Development

169

skin cultures, and then exposed to ultraviolet light (Mudgil et al., 2003). The ultraviolet treatment induced apoptosis in the normal keratinocytes but not in the tumor cells, which allowed expansion of the tumor cell colonies, thereby confirming the role of apoptosis in the expansion of p53 tumor clones found in mice. Melanomas are neoplasms that arise from melanocytes, which are interspersed as single cells in human skin among basal keratinocytes of the epidermis. Undifferentiated, but not differentiated, keratinocytes control growth, morphology, and antigen expression through direct cell–cell contact (Valyi‐Nagy et al., 1993). One of the most significant features accompanying melanoma development and progression is the expression of cell surface antigens not expressed in melanocytes. The expression of melanoma antigens is also induced in melanocytes after their isolation from skin and subsequent rapid division in culture (Herlyn et al., 1987), but expression is downregulated when the melanocytes are cocultured with keratinocytes (Shih et al., 1994). Neither fibroblasts nor medium from keratinocytes exert these effects on cultured melanocytes. Unlike the melanocytes, melanoma cells from primary and metastatic lesions continue to express those antigens constitutively in the presence of keratinocytes. The loss of regulatory dominance by keratinocytes occurs with downregulation of E‐cadherin expression and replacement by N‐cadherin in melanoma cells, resulting in loss of their adhesion to keratinocytes (Hsu et al., 2000b). Transduction of E‐cadherin expression in the melanoma cells leads to their adhesion to keratinocytes, thereby rendering them susceptible to keratinocyte‐mediated control. E‐cadherin expression in the presence of keratinocytes also inhibits invasion of the melanoma cells into the dermis by downregulating invasion‐related adhesion receptors, and inducing apoptosis. Normal melanocytes, but not melanoma cells, establish GJC with keratinocytes (Hsu et al., 2000a). However, the melanoma cells establish GJC among themselves and with fibroblasts. Although the loss of E‐cadherin is associated with malignancy, the increase of another cell adhesion protein, melanoma cell adhesion molecule (MCAM), correlates with malignancy (Li et al., 2004). Ectopic expression of E‐cadherin or the E‐cadherin alpha catenin fusion protein in melanoma cells restores their adhesion to keratinocytes. Deletion of the E‐cadherin cytoplasmic domain blocks restoration of that adhesion. Although GJC is associated with the normalization by E‐cadherin of melanoma cells when they come in contact with undifferentiated keratinocytes, it is noteworthy that the expression of E‐cadherin induces adhesion to keratinocytes (Hsu et al., 2000a; Kanno et al., 1984). Therefore, it is consistent with other findings that involve regulatory interactions between neoplastic and normal cells (Alexander et al., 2004; Hennings et al., 1992; Martin et al., 1991) that adhesion per se between the plasma membranes of E‐cadherin‐expressing melanoma cells and

170

Harry Rubin

keratinocytes is sufficient to normalize the phenotype of the melanoma cells. (Note: In mice, melanocytes are in the dermis, and a whole different set of regulatory relations must hold (Fukunaga‐Kalabis et al., 2006).) Another informative series of experiments about the modulating effect of normal tissues on the expression of neoplastic cells arose from transplanting rat hepatocarcinoma cells derived from stem‐like liver cell cultures into the liver or the subcutaneous tissue of normal rats. Some of the transplanted hepatocarcinoma cells of either a moderately aggressive or a highly aggressive line migrated into the hepatic plates (Coleman et al., 1993). Solitary cells in the hepatic plates from the moderately aggressive hepatocarcinoma line took on the appearance and regulated growth behavior of normal mature hepatocytes in young rats. However, cells from the more aggressive line multiplied in the hepatic plates, albeit slowly, and largely maintained their neoplastic morphology. Both cell lines multiplied rapidly into tumors after transplantation into subcutaneous tissue, where they assumed a bipolar mesenchymal cell appearance. Solitary cells from the less aggressive line multiplied into small foci when transplanted in the liver of old (18–24 months) rats (McCullough et al., 1998). The initially normalized hepatocarcinoma cells in the young rat retained that phenotype until they began to multiply at 14 months after their transplantation, forming small foci in the liver. The overall results of these experiments showed that the intact liver of young rats has the capacity to normalize hepatocarcinoma cells originating from stem‐like normal liver cells but that subcutaneous connective tissue has no such capacity. The liver loses the normalizing capacity as the rat ages, which suggests that the large increase of incidence in cancer with age results, at least in part, from the decrease in the normalizing capacity of the immediate epithelial cell microenvironment with age. The capacity of the young intact liver to normalize solitary hepatocarcinoma cells contrasts with the failure of normal keratinocytes to suppress the growth of epidermal carcinoma cells. The liver is also capable of maintaining metastasizing mammary cancer cells in a dormant state for extended periods of time (Naumov et al., 2002), although it is not known whether this capacity declines with age. It would appear that mixing keratinocytes with epidermal carcinoma cells in skin grafts is a less effective procedure for normalizing cancer cells than is the intact liver with its large masses of essentially pure hepatocytes. Drosophilae are ideal organisms for the study of tumor genetics because their genomes are so fully defined, and readily manipulated. Recessive‐lethal mutations that cause tumors in Drosophila larvae were first reviewed in 1978 (Gateff, 1978) and their study along molecular lines later elaborated (Bilder, 2004). Zygotic mutations in any of about 15 genes cause tumors in all proliferating tissues of larvae, but have been most extensively studied in epithelial organs called imaginal discs that serve as primordia for most adult structures. Unlike vertebrate tumors, which require combinations of

Cell–Cell Contact in Suppression of Tumor Development

171

genetic changes for their full development, a single zygotic mutation of the normal allele in flies carrying a heterozygous TSG is sufficient for neoplastic growth of the cells throughout all the imaginal discs. The biallelic TSG mutations have a very high penetrance, resulting in manifold increases in size of the imaginal discs and death of the larvae. Since all the cells carry the mutation, there are no wild‐type epithelial cells to modulate their neoplastic proliferation. The three most carefully studied TSGs encode cytoplasmic proteins found at the cell membrane, leading to a loss of ability to organize an epithelial monolayer. More recently, it has been found that mutations of the components of the endocytic machinery, which is required to internalize cell surface proteins, also produce larval tumors (Bilder, personal communication). Mutated TSG clones can be made in an otherwise wild‐type imaginal disc through mitotic recombination (Bilder, 2004). These clones grow more slowly than their wild‐type neighbors and are eliminated by the process of cell competition (Bilder, 2004). The elimination accounts for the extreme rarity of spontaneous tumors in adult Drosophila, and introduces the question of how the cells carrying the biallelic mutations are eliminated. The TSG scribble mutant clones do not overgrow because of cell death mediated by the Jun N‐terminal signaling pathway and the surrounding wild‐type tissue (Brumby and Richardson, 2003). In contrast, when oncogenic Ras or Notch is expressed within the scribble clones, cell death is prevented, and tumors develop. The mechanism of the wild‐type tissue in causing apoptosis of scribble clones is unknown, except to note that it differs from most cases of the suppression of tumor development in vertebrate organs in which the potential tumor cells persist indefinitely. It is also noteworthy that the imaginal discs are pure epithelium so the suppressive effects of the wild‐type tissue on TSG mutant clones are not complicated by possible mesenchymal effects, e.g., integrin mutations produce no tumors (Bilder, 2004).

IV. IS GJC REQUIRED IN CELL–CELL SUPPRESSION OF TUMOR DEVELOPMENT? It was apparent from the earliest studies among fibroblasts and their transformed counterparts that (a) contact between the two cell types is required for suppression of neoplastic development and (b) the normal cells had to be contact‐inhibited among themselves for the neoplastic suppression of transformed cells to be effective (Rubin, 1960a; Stoker, 1964). The discovery that there was exchange of molecules in the normal and neoplastic cells raised the question whether such GJC is necessary for suppression of neoplastic development (Stoker, 1967). Experiments directed

172

Harry Rubin

toward this question reported a correlation between GJC and suppression of fibroblast transformation, and suggested that cyclic AMP is the regulatory molecule (Mehta et al., 1986). However, subsequent experiments with rat fibroblasts transformed by five different oncogenes found no correlation of suppression with GJC, and that total inhibition of dye transfer did not prevent the suppression (Martin et al., 1991). Shortly thereafter it was found that there was no GJC between mouse papilloma cells and the normal keratinocytes that suppressed their proliferation (Hennings et al., 1992). In a related matter consistent with the lack of a role for GJC in suppression, TPA relieved the suppression of papilloma growth by keratinocytes but had no effect on homologous GJC among either the papilloma cells themselves or among the keratinocytes (Hennings et al., 1992). More recently, the modern techniques of molecular genetics have been combined with the earlier physiological methods to distinguish between the roles of contact per se and of GJC (Alexander et al., 2004). Brain fibroblasts, presumably glial cells which lacked GJC, were taken by Caesarean section from connexin‐knockout mice, and transformed by transfection with the v‐Src kinase. Direct contact between transformed and nontransformed cells was required for growth suppression of the former without the need for dye transfer between the cells. There was also evidence against the requirement of electrical communication by small ion transfer between the cells. In addition, there was an increase in expression of serum deprivation response protein, consistent with contact inhibition of membrane activity as the suppressive mechanism of transformation. Further evidence against the requirement for GJC in suppression of transformation came from the observation that suppression can be induced among the transformed cells themselves, without intervention by nontransformed cells. As noted earlier, spontaneously transformed mouse fibroblasts undergo contact inhibition when grown to very high density on a small coverslip in a large volume of medium to avoid depletion of the medium (Rubin and Chu, 1982). The cells take on the appearance and other growth characteristics of nontransformed cells, and these are maintained up to 1 day after subculture at low density. This result shows there is no need for transfer of regulatory molecules from nontransformed molecules from nontransformed cells, or stimulatory molecules from the transformed cells to effect their regulation. With the exception of one paper reporting correlation between GJC and suppression (Mehta et al., 1986), the overall results are consistent with contact inhibition of plasma membrane activity as the underlying mechanism of suppressing the transformed phenotype of fibroblasts and epithelial cells. It should be noted that GJC itself requires contact between cells, and the reported correlation of GJC with suppression (Mehta et al., 1986) could be an indicator of the extent of adhesion between the normal and the transformed cells.

Cell–Cell Contact in Suppression of Tumor Development

173

V. THE ROLE OF PLASMA MEMBRANE ACTIVITY IN REGULATION OF CELL GROWTH It is well established that protein growth factors such as insulin initiate their stimulatory action on cells by attaching to specific receptors on the plasma membrane (Cuatrecasas, 1969). It has also been shown that trypsin immobilized on polystyrene beads initiates division of chicken embryo fibroblasts without entering the cell or being released to the medium (Carney and Cunningham, 1977). Calcium pyrophosphate stimulates the division of an established line of mouse fibroblasts but only when insoluble floccules are formed by the agent (Bowen‐Pope and Rubin, 1983). These floccules attach to the cell surface but can be completely recovered by briefly lowering the pH, thereby blocking further action on the cells. Further evidence of the role of plasma membrane activity in regulating the proliferation of cells comes from the relation between population density of fibroblasts and their growth rate. The fibroblasts proliferate at a maximal rate at relatively low population densities at which they move over the surface of the culture dish while exhibiting an actively ruffling membrane (Abercrombie and Heaysman, 1954). The ruffling stops when the membrane forms an adhesion with another fibroblast (Abercrombie and Ambrose, 1958). The cell may break its adhesive contact with the other cell and move away, thereby maintaining a high growth rate. When the population density becomes high enough to form a confluent sheet, all ruffling and independent movement of the cells are inhibited, and the rate of proliferation of the population decreases markedly in the phenomenon of contact inhibition of proliferation (Todaro et al., 1965). Epithelial cells also exhibit contact inhibition (Castor, 1968; Eagle and Levine, 1967) but they form tighter adhesions with one another than do fibroblasts, and do not break contacts once they are established (Middleton, 1973). Proliferation is therefore maximized at the free edge of a coherent sheet or of a colony of epithelial cells. Epithelial cells neither inhibit fibroblasts, nor are they inhibited by them (Eagle and Levine, 1967). Less is known about whether epithelial cells from different tissues inhibit each other except that esophageal epithelia can replace tracheal epithelia in competing with a line of tracheal carcinoma cells for attachment to tracheal mesenchyme (Terzaghi‐Howe, 1987). When fibroblasts are transformed into sarcoma cells they assume a rounded morphology in culture and do not form ruffled membranes (Abercrombie and Ambrose, 1958). The sarcoma cells do have a very actively moving surface in the form of fine, rather spiky processes extending and retracting around the periphery. There is little or no mutual adhesion or reduction in

174

Harry Rubin

surface activity or locomotion when one sarcoma cell meets another or meets a fibroblast, nor do these changes occur in a fibroblast when it encounters a sarcoma cell. The proliferation of sarcoma cells continues after they contact one another or fibroblasts, and form multilayered colonies of transformed foci on a confluent sheet of fibroblasts, which are confined to a monolayer. As noted earlier, if the fibroblasts had themselves formed a fully contact‐inhibited sheet before adding the dissociated sarcoma cells, the latter in some combinations assume the morphology of fibroblasts and stop proliferating. As also noted, sarcoma cells can be forced to undergo contact inhibition at densities an order of magnitude higher than fibroblasts do, and normalize their appearance (Rubin and Chu, 1982). The overall results of these cell interactions indicate that normal homophilic cells inhibit their proliferation by forming adhesions between their surface membranes which immobilize the membranes, thereby accounting for the inhibition. Transformation to malignancy markedly reduces the homophilic adhesions and permits continuing proliferation. The role of the plasma membrane in regulating cell proliferation was directly demonstrated by adding cell‐free membrane preparations to multiplying cells. Membrane preparations from 3T3 fibroblasts inhibited DNA synthesis of intact 3T3 fibroblasts, but not that of transformed cells (Wittenberger and Glaser, 1977). Inhibitory effects on sparsely seeded human lung fibroblasts were obtained by adding membranes isolated from confluent sheets of the same cells (Wieser et al., 1985). Removing the oligosaccharides of glycoproteins from the lung fibroblast membranes inactivated their growth‐inhibiting effect, indicating those components mediate the contact inhibition of the cells. Purified rat liver plasma membranes inhibited growth of rat hepatocytes at low density (Nakamura et al., 1983). These results strongly support the sufficiency of direct cell–cell membrane interactions in regulating cell proliferation among normal cells, and between normal and neoplastic cells to regulate the latter. It should be borne in mind, however, that the susceptibility of neoplastic cells to regulation by normal cells decreases with increase in their malignancy. This was most clearly evident in the epidermal cell system in which keratinocytes suppress the proliferation of papilloma cells, but not of carcinoma cells (Hennings et al., 1990; Strickland et al., 1992). This finding is correlated with the observation that papilloma cells have almost as firm adhesion to one another as normal cells do, but squamous carcinoma cells have much lower mutual adhesiveness (Coman, 1944). It was also seen in the transplantation of rat hepatocarcinoma stem cells into the normal rat liver where cells of the moderately malignant line assumed the phenotype of a normal hepatocyte, but those of the highly malignant line formed slow‐growing tumors (Coleman et al., 1993). Similar but less striking results were obtained with preneoplastic and neoplastic tracheal epithelium transplanted to the trachea

Cell–Cell Contact in Suppression of Tumor Development

175

(Gillett et al., 1989). While those correlations are consistent when the phenotypes of adhesiveness or malignancy of living cells are considered, they are less consistent when correlations are sought between identified adhesive molecules, and the extent of cell–cell adhesion or other neoplastic characters (Foty and Steinberg, 2004; Rubin, 2007).

VI. SUPPRESSIVE EFFECTS OF MESENCHYMAL TISSUE ON NORMAL AND NEOPLASTIC EPITHELIAL PROLIFERATION It is commonly thought that the basement membrane of epithelial tissue prevents invasion of the underlying connective tissue by normal and benign epithelium (Weinberg, 2007). However, the basement membrane is a thin, fragile structure that, in the case of the epidermis, is breached every time there is a bleeding wound, yet there is no invasive growth of the epidermis into the connective tissue. More than half the injections by a needle through the skin carry fragments of epidermis into the subcutaneous tissue, where most fail to survive, although some produce epidermoid cysts (Gibson and Norris, 1958). Benign papillomatous growth, by definition, is noninvasive; growth of the papillomas is entirely in an upward direction (Cramer and Stowell, 1942) even when papilloma cells are grafted in suspension along with an excess of fibroblasts in the absence of basement membrane (Strickland et al., 1992). It would therefore appear that a combination of strong homotypic adhesion among epidermal cells, and the alien environment of the underlying connective tissue maintain the noninvasive architecture of the normal epidermis and of papilloma cells. About half of the fragments of mouse mammary epithelium that are transplanted into subcutaneous tissue retain their viability for many months, most of the others regress (DeOme et al., 1978). In some instances, the transplants develop into HANs, but do not progress to invasive tumors. Most HANs maintain their viability after transplantation into the subcutis, but a significant minority regress, and very few progress to invasive tumors. In contrast, about 10% of normal mammary epithelia injected into mammary fat pads that had been cleared from a resident epithelium developed into invasive tumors, while almost 50% of transplanted HANs did so. The results indicate that connective tissue is a more hostile place for growth of mammary epithelium and its development into invasive tumors than is the adipose tissue of the mammary fat pads that forms the normal surrounding of the mammary ducts. The development of a malignant epithelial tumor involves the capacity to flourish in a previously restricted environment, presumably by genetic or

176

Harry Rubin

genomic changes in the tumor cells. It does not require any permissive change in the connective tissue itself, since the tumor can now be transplanted into the subcutaneous tissue of normal isogenic animals. Invasiveness is facilitated not only by the capacity to proliferate in connective tissue but also by a decrease in the adhesiveness of the tumor cells to one another (Coman, 1944; McCutcheon et al., 1948) and to the nontumor epithelial cells surrounding them (Coman, 1953; Mareel et al., 1990). There is a further reduction in adhesiveness associated with metastasis, which allows separation of cells from the invasive tumor mass and entry into the circulation, as well as a capacity to proliferate in a different organ. Taken to an extreme, ascites tumor cells which multiply while suspended as individuals in peritoneal fluid, can be developed by selection from solid tumors (Klein, 1951). There is a consistent difference in the rate of conversion to the ascites form between different tumors, indicating an underlying genetic variation. In the selective process, the tumor cells exhibit increasing negative charge as they increase their tendency to multiply in the ascites form (Purdom and Ambrose, 1958). At the same time there is an increase in the capacity of the cells for metastasis (Ringertz et al., 1957), indicating a negative correlation between adhesiveness and metastasis.

VII. THE PROTOTYPE OF PROGRESSION TO METASTASIS AS SEEN IN HUMAN MALIGNANT MELANOMA Perhaps the most thoroughly studied human progression models, extending from normal cells through benign and malignant neoplasia to metastasis, is the melanoma model. The visibility of these tumors in the skin and their pigmentation make them an ideal subject for continuous observation. The Pigmented Lesion Clinic started at the Massachusetts General Hospital by Wallace Clark, continued in Philadelphia at Temple University, and succeeded by the Pigmented Lesion Group at the University of Pennsylvania, recorded hundreds of clinical and pathology attributes in over 1300 patients (Clark, 1994, and personal communication). One of the earlier reports from this group involved 261 patients with primary melanoma and followed for a minimum of 5½ years (Clark et al., 1984). Later on, there were over 1000 new patients with normal and abnormal nevi or melanomas who were observed for shorter times (Clark et al., 1986). These studies led to classifying melanoma progression into five stages. The first of these is the common acquired melanocytic nevus or mole, in which there is focal proliferation of benign cells. The second stage is a melanocytic nevus with aberrant differentiation and an aberrant form of melanocytic hyperplasia. The third stage

Cell–Cell Contact in Suppression of Tumor Development

177

exhibits nuclear atypia, i.e., dysplasia. The fourth stage is known as the radial growth phase (RGP) and is the first sign of primary melanoma. It is characterized by net enlargement of the tumor at its periphery, but is not associated with metastasis. The fifth stage is the vertical growth phase (VGP) which appears as spheroidal nodules that extend into the dermis. The cells of this phase are those that give rise to the sixth phase of metastasis or the ability to maintain autonomous growth that is discontinuous from the primary site. To elaborate on these stages, the first evident focal proliferation of melanocytes in humans is the common acquired melanocytic nevus (Clark et al., 1984). It is also the initial proliferative lesion seen in response to the carcinogen DMBA in certain guinea pig strains (Pawlowski et al., 1976). It therefore represents the first stage of melanocytic tumor progression. After a few years, many of the nevi undergo a pathway of differentiation into Schwann cells. At the same time the nevi may extend into the dermis, extinguishing the basement membrane boundary between epidermis and dermis. However, the lesion is considered a compound melanocytic nevus, and rather than being invasive, Schwannian differentiation may occur in the dermis, followed by disappearance of the lesion. Another pathway followed by the nevi is hyperplasia of cells in the periphery of the nevus (Clark et al., 1984). The proliferation occurs in the basilar intraepidermal melanocytes, and there is no differentiation. Atypical melanocytes may appear in two forms in the area of melanocytic growth at the shoulder of the nevus, which constitute melanocytic dysplasia. It identifies patients at heightened risk for the development of malignant melanoma. The chief marker of dysplasia is nuclear atypia. It is often followed by the RGP of malignant melanoma, which shows partial growth autonomy and is considered the first phase of a primary neoplasm. If untreated such lesions are quite likely to progress further to the next step of the VGP. Prior to the autonomous RGP, progression of individual lesions from one class of lesion to another is a very rare event. However, from the RGP of autonomous growth, forward progression is the rule not the exception (Clark et al., 1984). Progression from melanocytic dysplasia to the RGP of superficial spreading melanoma is focal, qualitative, and has the characteristics of a mutational event. Generally there is an associated broad dense plaque of lymphocytes that underlies invasive melanoma cells. This pattern of invasiveness, however, does not lead directly to metastasis. The VGP starts as the focal appearance within the RGP of a new population of cells that tend to grow as an expanding spheroidal nodule similar to the growth of a metastasis. The cellular aggregates of the VGP are larger than the clusters of cells that form the RGP, and grow at a faster rate. They extend more deeply than those of the radial phase, but the local lymphocytic immune response disappears (Clark et al., 1986).

178

Harry Rubin

It was assumed by many in the 1960s that malignant melanoma is the most malignant of cancers, but the actual fatality rate is 50–60% (Clark, 1994). It does have the capacity to metastasize to a wide variety of organs throughout the body (Rusciano and Burger, 1992). Unlike most other cutaneous cancers, tumor progression in melanocyte neoplasia commonly leads to metastasis (Clark et al., 1984). The pigmentation and surface location allows identification of early lesions when only 0.1–0.2 mm diameter and the progression from such small benign lesions to metastatic malignancy allows systematic study of the full range of lesional entities. Such studies have taken melanocytic neoplasia from one of the less understood forms of malignant progression to probably the most completely described form. One particularly significant distinction made in these studies is that metastases only occur when cells acquire the competence to proliferate in the dermis as seen in the VGP, as opposed to their simple extension into the dermis, which can occur even in the nonneoplastic stages.

VIII. CHARACTERISTICS OF CULTURED HUMAN MELANOCYTES ISOLATED FROM DIFFERENT STAGES OF MELANOMA PROGRESSION Melanocytic cells from normal skin, common acquired congenital nevi, radial (RGP) and vertical (VGP) growth phases of primary melanoma, and from metastatic melanoma were reported to have properties in cell culture that reflect their original state of tumor progression in vivo (Herlyn et al., 1985b). The relative phenotypic stability of cultured melanocytic cells can be explained by their karyotypic stability because they maintain the same chromosomal abnormalities in freshly isolated melanomas and both short‐ and long‐term cultured cell lines derived from the same lesions. The growth characteristics of cells from various stages of progression are shown in Table III. It can be seen in Table III that the normal melanocytes from newborn foreskins and from the nonmalignant nevi assume a spindle shape in cell culture, but so do cells from the single RGP primary tumor, and 4 of the 7 VGP tumors, the other three being epithelioid. Cells from the two metastatic tumors were epithelioid. The normal melanocytes and those from the nevi had a limited lifespan in culture, whereas the cells from the primary tumors and from metastases grew indefinitely. It should be noted, however, that indefinite growth may arise from a minority of cells in the original tumors, especially in the case of the RGP, which has a very low capacity for colony formation in agar and produce no tumors in nude mice, nor do they lead directly to metastases (Clark et al., 1984, 1986). In contrast, there is significant colony formation over a wide range in agar by cells from the VGP that

179

Cell–Cell Contact in Suppression of Tumor Development

Table III Growth of Cultured Melanocytic Cells Isolated from Newborn Skin, Nevus, and Primary and Metastatic Melanoma

Origin

Morphology

Normal melanocytes from Newborn Spindle foreskin (2) Common Spindle acquired nevus (2) Congenital Spindle nevus (2)

Indefinite growth in vitro

% colony formers in agar

Tumor formation in nude mice

No

0.012–0.024

No

No

0.014–1.03

No

No

0.47–0.84

No

RGP Early primary melanoma (1)

Spindle

Yes

<0.001

No

VGP Late primary melanoma (7) Metastatic melanoma (2)

3 spindle and 4 epithelioid Epithelioid

Yes

3.7–19.1

Yes

Yes

22.7–31.8

Yes

Values in parentheses indicate number of original samples. Abstracted from Herlyn 1985b.

is exceeded by the cells from the two metastases, and all the tumors in these two categories produced tumors in nude mice. Chromosome studies were done on direct preparations, early and late passage cell cultures from the various stages of melanoma progression (Balaban et al., 1984). Nevus cells had a normal karyotype whereas the early passage tumor of the RGP was pseudodiploid with an extra copy of chromosome 6p translocated onto chromosome 22, as well as other abnormalities. The VGP‐advanced primary and metastatic melanomas were aneuploid with multiple aberrations and nonrandom abnormalities of chromosomes 1, 6, and 7. The overall similarity of karyotype between VGP with its growth in the dermis and metastasis is consistent with the observation that metastases arise only from cells of the VGP, although metastatic spread from this phase is not the rule (Clark, 1994). As noted earlier, it is only these two stages that have a high efficiency of colony formation in soft agar and are tumorigenic in nude mice (Herlyn et al., 1985b). Their cells had similar morphology and plating efficiency in cell culture (Herlyn et al., 1985a). However, the metastatic cells generally had a shorter population‐doubling time, growth to a higher cell density, higher tyrosinase activity, and more pigmentation than do cells of the VGP.

180

Harry Rubin

The colony‐forming efficiency of the metastatic cells in agar was also somewhat higher than that of the VGP cells, and they could be found growing in peritoneal exudates, indicative of a loose relationship to one another. On the one hand, monoclonal antibodies to VGP and metastatic cells bound poorly to nevus cells. On the other hand, antinevus antibodies bound to melanocytes, nevus cells and RGP primary melanoma cells but not to VGP or metastatic cells (Herlyn et al., 1985b).

IX. IS THERE A RELATIONSHIP BETWEEN THE CELL CONTACT INTERACTIONS THAT SUPPRESS NEOPLASTIC DEVELOPMENT AND THE PHENOMENON OF METASTATIC DORMANCY? The major characteristics of the suppression of the neoplastic phenotype can be summarized as follows. Studies with fibroblasts in culture revealed that cells transformed by viruses (Rubin, 1960a; Stoker, 1964), carcinogenic chemicals (Mehta et al., 1986), or spontaneously (Rubin, 1994) assume the morphology and growth‐regulatory behavior of normal cells when surrounded by, and in full contact with an excess of the normal fibroblasts. The latter had to be confluent, and under contact inhibition themselves before the suppression could take hold (Rubin, 1960a, 1994). The suppression is brought about by contact between the plasma membranes of the normal and transformed cells in the same manner that the normal cells induce contact inhibition among themselves, by restricting the membrane activity of the cells. Although no specific attempt was made to determine whether epithelial cells could suppress transformation in fibroblasts, it seems unlikely since there is no contact inhibition between normal cells of the two cell types (Eagle and Levine, 1967). Basically similar results are found when neoplastic epithelial cells are brought in contact with homotypic normal cells. Papilloma cells are suppressed in contact with an excess of normal keratinocytes both in vitro and in vivo, but not by fibroblasts (Hennings et al., 1990; Strickland et al., 1992). However, epidermal carcinoma cells are not suppressed by the keratinocytes (Strickland et al., 1992). Rat hepatocarcinoma cells of stem cell origin take on the phenotype of normal hepatocytes when transplanted into the normal liver as solitary cells which come into contact with hepatic plates (McCullough et al., 1998). Human melanoma cells that are metastasis‐ competent in severely immunodeficient mice have obviously escaped contact with and regulation by keratinocytes, but they resume that contact and are normalized when transfected to overexpress the E‐cadherin gene (Hsu et al., 2000b). Such observations raise the question whether metastatic

181

Cell–Cell Contact in Suppression of Tumor Development

dormancy results from contact of the neoplastic cells with the heterotypic cells of the organ in which they are lodged.

X. CHARACTERISTICS OF METASTATIC DORMANCY Evidence for metastatic dormancy arose when rats were injected intraportally with as few as 50 carcinosarcoma cells but failed to develop liver tumors by 20 weeks unless they were subjected to weekly laparotomy and hepatic manipulation, beginning at 12–13 weeks (Fisher and Fisher, 1959). The 100% incidence of liver tumors in the treated animals indicated the neoplastic cells remained dormant, yet retained all of their former capacity to proliferate. Dormancy was recognized clinically in human cancer with the occurrence of metastases years after removal of the primary tumor (Demicheli et al., 1996). However, the status of the cancer between the time of the primary treatment and metastatic recurrence was unknown. Several factors were suggested as possible contributors to metastatic dormancy, one of which was the survival in tissue of solitary cancer cells that are neither proliferating nor undergoing apoptosis. Concrete evidence for such solitary cells came from experiments in which melanoma cells labeled with fluorescent nanospheres were injected into the mesenteric vein of mice to target the liver (Luzzi et al., 1998). Eighty percent of the injected cells survived in the liver microcirculation and extravasated by day 3, but that number decreased to 36% at 13 days (Fig. 1). Only 1 in 40 of the extravasated cells formed micrometastases (4–16 cells) by day 3, and only 1 in 100 of these

Time (p.i.)

Solitary cells

Micrometastases Tumors (4–16 cells)

Total loss (%)

0

100%

Inject 3⫻105 B16F1 cells

0

90 min

87.4

3d

81.4

2.04

13 d

36.1

0.07

12.6

16.6

0.02

63.8

Fig. 1 Flow chart summarizing survival data in the liver of mouse melanoma cells injected intraportally to target mouse liver. See text for details. From Luzzi et al. (1998).

182

Harry Rubin

progressed to macroscopic tumors by day 13. In contrast to the solitary cells and the micrometastases, more than 90% of the cells in the macroscopic tumors were proliferating. Thus, in this model, metastatic inefficiency was principally determined by failure of solitary cells to initiate growth, and failure of early micrometastases to continue growth into microscopic tumors. The results agree with the earlier finding that virtually all melanoma cells injected in the chick embryo chorioallantoic membrane survive in the microcirculation and successfully extravasate by 24 h (Koop et al., 1995). In the chick embryo model, extravasation is independent of metastatic ability, and occurs even when normal mouse embryo fibroblasts are injected (Koop et al., 1996). The implication of these studies is that the primary determinants of metastatic inefficiency are the postextravasation survival and growth of cells. The results indicate that early cell destruction in the microcirculation, an inability of cells to extravasate and failure of angiogenesis are not major contributors to metastatic inefficiency. Two lines of mouse mammary cancer cells, one that was highly metastatic and the other poorly metastatic for the liver, produced tumors after injection into the mammary fat pad (Naumov et al., 2002). Solitary cells from both cell lines were recognizable in the liver after appearance of the primary tumors, indicating that the cells had carried out the first steps of metastasis. To study the kinetics of solitary cell survival, the cells were labeled with nanospheres and injected intraportally. By 3 days, all observed cancer cells of both types had extravasated (Fig. 2). There was a small loss of the poorly metastatic cells shortly after injection, but there was no further loss of cells up to 3 weeks, and about 80% of them remained as undivided cells.

% Solitary cell survival

100 80 60 40 20 0 90 min

3

10

14

18

21

77

Time post-injection (days)

Fig. 2 Long‐term survival in the liver of poorly metastatic solitary mammary carcinoma cells injected intraportally to target the mouse liver. See text for details. From Naumov et al. (2002).

Cell–Cell Contact in Suppression of Tumor Development

183

Even after 11 weeks, about 50% of the poorly metastatic cells that were injected remained as solitary cells, with low levels of both apoptosis and proliferation. In the case of the highly metastatic line, all cells had extravasated by 3 days, and most of them (64%) remained as solitary, undivided cells at 10 days. A small population (0.6%) had begun to form metastases by 10 days, but their number decreased 100‐fold by 14 days. By contrast, the number of solitary cells fell by only a factor of 3, indicating that the overall rate of loss of early metastases was almost 2 orders of magnitude greater than that for solitary cells. Although the tumor burden increased to 70% of the liver in 21 days, surprisingly large numbers of the dormant solitary cells could still be seen, and they were similar in appearance to those of the poorly metastatic line. The proportion of remaining solitary cells of the highly metastatic line could not be estimated beyond 21 days because of the large tumor burden in the liver, but those of the poorly metastatic line remained fully viable in the liver parenchyma through the last point taken at 11 weeks. They could therefore account for the occasional metastases in the liver in some mice after long latency periods. The line of melanoma cells used to study the fate of disseminated cells in the liver was used for the same purpose after targeting to the lung (Cameron et al., 2000). As in the liver, a large proportion of melanoma cells extravasated and solitary cells showed an initially slow decline to 74% by day 3, a rapid decline to 25% by day 4, followed by a steady decline to 3.5% by days 12–14. This contrasts with 36% solitary cells at 13 days in the liver (Luzzi et al., 1998). However, the percentage of multicellular foci at 13–14 days in the lung was much higher than those in the liver at the same time, and the overall metastatic efficiency was about 10% in the lung as compared with 0.002% in the liver. The results indicate a sharp difference in growth response of the melanoma cells in the microenvironment of the two organs. The above experiments were initiated by intravenous injection of mouse cancer cells targeted to the liver or lung, but a different set of experiments began with human mammary cancer cells that were inoculated into the mammary glands of nude mice to form primary tumors there (Goodison et al., 2003). After tumors appeared in the mammary gland from isogenic, nonmetastatic, and metastatic lines, cells that escaped spontaneously from both cell lines were seen in the lungs. No metastases were seen in the lungs by 6 months with the nonmetastatic line despite the continuing presence of scattered cancer cells. These self‐disseminated human tumor cells were retrievable from the tissues long after resection of the primary tumor, as manifested by indefinite proliferation in vitro and local tumorigenicity in the mammary gland. Scattered tumor cells from the lung were still immortal but were rendered indefinitely quiescent by the microenvironment conditions in the lung tissue. In contrast, many cells from the metastatic line grew into metastases in the lungs, but others remained solitary and quiescent.

184

Harry Rubin

Therefore, even in a clonally‐derived cell population with metastatic properties, many cells do not mobilize the organ‐specific growth properties needed to generate metastases. The studies with the human mammary cancer cells were extended to include a variety of organs in which no spontaneous metastases were formed from primary tumors, in addition to the lungs and lymph nodes where the metastatic line usually produced growth (Suzuki et al., 2006). Dormant cells were recovered from metastasis‐free organs 3 months after injection into mammary glands, including cells in those lungs and lymph nodes that had no metastases. Cells that grew indefinitely in culture and could induce primary mammary tumors were retrieved from blood, lungs, lymph nodes, spleen, bone marrow, liver, and kidneys from mice in which the primary tumor was not resected, but not from the blood when the tumors had been resected. Cells recovered from the high metastatic line also produced metastases in the lung and lymph nodes. Although the resolution of light microscopy was not always sufficient to establish unequivocally whether the spontaneously metastasized labeled cells were intra‐ or extravascular, there was convincing evidence in some instances that the tumor cell had departed from the blood and entered the tissue of the host organ. Further evidence for extravasation and entry into the tissue was the failure to detect the cancer cells in blood cultures of the mice 1–4 weeks after resection of the primary tumors. The results showed that the growth of disseminated cancer cells in all the organs tested can be suspended for 6 months or longer by the microenvironmental conditions, although growth was still active in other organs of the same host.

XI. TUMOR CELL ADHESION TO CELLS IN DISTANT ORGANS A. Endothelial Cells The question to be addressed at this point is whether metastatic dormancy has any relation to the suppression of tumor development by cell–cell interactions between preneoplastic and neoplastic cells with an excess of their normal homotypic counterparts. Because the best established cases of homotypic suppression of tumor development result from direct contact between the normal and neoplastic cells, evidence was first sought for heterotypic contact interactions between disseminated cancer cells (DCC) and the cells of their host organs that might account for metastatic dormancy. As might be expected, the studies of adhesion in metastasis have focused upon its role in the promotion rather than suppression of metastatic lesions. However, the major features of the metastatic cell adhesion literature will be examined to

Cell–Cell Contact in Suppression of Tumor Development

185

learn whether it reveals hidden evidence of suppression. The first tissue encountered between DCC and an organ in which they may lodge is, of course, the endothelial cell surface of the organ’s vasculature. It has long been considered that adhesion of the DCC to specific organ microvessel endothelial cells is an important event in determining organ‐specific metastatic growth (Nicolson, 1988). B16a melanoma cells injected into the tail vein of mice appeared in the lung and were arrested there by contact with the endothelial plasma membrane (Crissman et al., 1985). The endothelial cells were gradually displaced by tumor cells which achieved contact with the vascular basement membrane at 4 h. Mitotic figures were evident by 24 h, and the tumor appeared to proliferate intravascularly along the basement membrane. Extravasation occurred through tumor cell proliferation and destruction of the vascular basement membrane. In the case of B16F10 melanoma cells, there was a partial retraction of the endothelial cells following attachment of the tumor cells, which then attached to the basement membrane and the basolateral endothelial cells (Lapis et al., 1988). The endothelial cells extended to cover the tumor cells, which proliferated to fill the lumen. The endothelial layer became mechanically disrupted and the tumor cells extravasated. In both cases contact of the tumor cells with endothelial cells led to proliferation of tumor cell growth. The specificity of adhesion between murine tumor cells and capillary epithelium was addressed with a panel of eight different histological types of tumor cells interacting with endothelial cell monolayers from four different organs (Auerbach et al., 1987). The tumor cells differed in adhesive propensity for different endothelial cells. Some, but not all, of the adhesive preferences correlated with the in vivo metastatic behavior of the tumors, indicating that endothelial cell surface‐associated specificities may play a significant role in determining the pattern of metastases. Two sublines of a large cell lymphoma line were selected for enhanced liver and lung colonization and tested for adhesion to mouse liver sinusoidal endothelium, lung microvessel endothelium and bovine aortic epithelium (Belloni et al., 1986). Only the selected lung‐colonizing melanoma cells adhered preferentially to lung endothelium, but both the lung and liver‐ colonizing lines adhered to hepatic sinus endothelium significantly more than the parental line did. These and other results suggested that organ specificity of metastasis may be determined in part by specific tumor cell‐endothelial cell interactions (Nicolson, 1988). Human fibrosarcoma cells injected intravenously into rodents attached to the endothelia of pulmonary precapillary arterioles and capillaries (Al‐Mehdi et al., 2000). Proliferation of the tumor cells in this case originated intravascularly, and was restricted to those attached to the endothelium. The intravascular origin of the proliferation was made especially clear by the technique of labeling the vascular surface of the endothelial cells with

186

Harry Rubin

acetylated low‐density lipoprotein, and labeling the tumor cells with green fluorescent protein. The accumulated results show that attachment of DCC to endothelial cells results in their proliferation rather than dormancy.

B. Parenchymal Cells An alternative to the origin of metastatic dormancy of DCC by adhesion to endothelial cells is their adhesion to the parenchymal cells of the host organs after extravasation. Experiments suggest that endothelial cells express the same or nearly the same organ‐specific molecules as do parenchymal cells from the same organ (Nicholson, 1988). Cell adhesion studies have been conducted with malignant cells and parenchymal cells from their target and nontarget organs. Sublines of B16 mouse melanoma cells were developed that had undergone increasing numbers of selective subcutaneous or intravenous passages for capacity to produce increasing metastatic growths in the lungs of mice (Nicolson and Winkelhake, 1975). Cells from the lung metastases were suspended and mixed with cells from different organs prepared to exclude membrane debris, erythrocytes, platelets, and cell aggregates. The degree of aggregation between the melanoma cells and lung cells increased with increasing rounds of selection of the melanoma cells, which also produced increasing lung metastases. In contrast, there was no correlation of aggregation with metastasis in mixtures of the lung‐ colonizing melanoma cell lines with cells from liver, spleen, and kidney. Furthermore, the maximum binding of the melanoma cells to any of these organs was much lower than it was to lung cells. Many other studies indicate that highly metastatic cells adhere at greater rates or more extensively to parenchymal cells of target rather than nontarget organs. For example, liver‐ metastasizing lymphoma cells bound to isolated liver parenchymal cells in proportion to their capacity to colonize the liver in vivo (McGuire et al., 1984; Schirrmacher et al., 1980). Sublines of large cell lymphoma cells were selected by in vivo colonization of the liver and produced extensive metastasis to the liver. They also bound much more selectively to aggregates (McGuire et al., 1984) or frozen sections (Kieran and Longenecker, 1983) of liver cells than to nontarget organ cells or tissues. Spleen‐metastasizing leukemia cells bound to isolated spleen cells but not to lung cells (Phondke et al., 1981). That such preferential adhesion to target organs is mediated by specific cell receptors was shown by blocking adhesion to liver cells in vitro and liver metastasis in vivo with antibodies to cell surface determinants on lymphoma cells (McGuire et al., 1984). In conclusion, tumor cell adhesion mechanisms for endothelial and parenchymal cells are mediated by different multiple adhesion molecules operating in parallel, and they commonly use normal cell adhesion processes

Cell–Cell Contact in Suppression of Tumor Development

187

(Belloni and Nicolson, 1988; Nicolson, 1988). Since the adhesion of DCC to heterotypic cells of distant organs enhances the proliferation of the tumor cells in metastasis formation, it stands in sharp contrast to the suppression of the tumor development by adhesion between early stage neoplastic cells and an excess of homotypic normal cells, as most clearly illustrated in the interactions between initiated and normal epidermal cells (Hennings et al., 1990; Strickland et al., 1992).

XII. POSSIBLE ALTERNATIVE EXPLANATIONS OF METASTATIC DORMANCY A hint of a possible corollary to metastatic dormancy may be found in the failure of epidermal cells to invade the underlying dermis despite frequent breaches in the basement membrane (Tarin, 1972). It is also seen in the failure of epidermal cells to grow in the underlying mesenchyme when commonly deposited there by ordinary needle injections through skin (Gibson and Norris, 1958). A closer example to metastatic dormancy comes from experiments in which enzymatically dissociated normal thyroid cells were intravenously injected into mice, and lodged in the lung (Taptiklis, 1968). Continuous ingestion of methylthiourea, which inhibited the production of thyroxin by the host thyroid gland in the experimental mice, prevented the feedback inhibition of thyroid‐stimulating hormone. The resulting overproduction of thyroid‐stimulating hormone led to proliferation and follicle formation of the thyroid cells that had lain dormant in the lung for as long as one year. The evidence indicated that the normal thyroid cells penetrated the endothelial barrier and remained dormant in the interstitial tissue of the lung. Subsequent experiments with intravenous injection of dissociated normal, hyperplastic, and neoplastic thyroid cells into methylthiourea‐ingesting mice revealed a common ability to penetrate the endothelium by intravascular proliferation, and migration of the thyroid cells to an extravascular position where they proliferated (Taptiklis, 1969). The vascular penetration appears to occur by unstimulated cells of all three types, and was later also found to occur in unstimulated chick embryos inoculated intravenously with normal and transformed mouse fibroblasts (Koop et al., 1996). Even the neoplastic thyroid cells remained dormant in the absence of thyroid‐stimulating hormone, indicating their hormone dependence. In the presence of the hormone, the neoplastic thyroid cells lost polarity, grew much more rapidly than either the normal or the hyperplastic cells, invaded the bronchi, and completely obliterated the acinar spaces. The overall results are consistent with the idea that metastatic dormancy arises from the paucity of growth‐stimulatory

188

Harry Rubin

factors in host organs, rather than any suppressive effect of heterotypic cell–cell interactions. Further evidence for a role of growth factor deficiency in maintaining metastatic dormancy comes from experiments on the invasiveness of human epidermoid carcinoma, which is enhanced by the binding of urokinase plasminogen activator to its receptor on the cancer cells (Yu et al., 1997). If the expression of the receptor is blocked, the carcinoma cells enter a state of dormancy, which resembles that observed with human cancer metastasis. Although the dormancy was protracted, the cells eventually emerged to initiate progressive growth and form metastases, indicating that other factors can compensate for the lack of a full complement of surface receptors. High expression of the receptors and binding of the plasminogen activator entrains signaling pathways which increase the pro‐ proliferative balance of a high ratio of extracellular signal‐regulated kinase ERK to p38 (Aguirre‐Ghiso et al., 2003). The organ specificity of metastasis development has been associated with receptors for soluble growth factors in the vasculature of target organs (Pasqualini and Ruoslahti, 1996). Endothelial cells appear to express phosphorylated receptors for platelet‐derived growth factor when they are exposed to tumor cells that produce the growth factor (Uehara et al., 2003). Hence, such tumor cells that are carried to an organ that does not have the receptors may remain dormant. A discontinuous 70 cM (centimorgan) region of human chromosome 17 contributes to the dormancy of a rat line of prostatic cancer cells (Chekmareva et al., 1998). The growth inhibition seems to result from an effect of one or more genes at the metastatic site, and not from a circulating angiogenesis inhibitor. However, angiogenesis plays a limiting role in the growth of selected model systems at the primary site of cell inoculation (Naumov et al., 2006), and might play a similar role in limiting the size of metastases. It is unlikely to be the limiting factor in the dormancy of solitary disseminated cells which feature so prominently in models of metastatic dormancy (Luzzi et al., 1998; Naumov et al., 2002), since they have the same nutritional support as the surrounding normal cells. In any case, the role of angiogenesis in limiting tumor size in the decades‐ long development of some human cancers is less apparent than it is in experimental model systems since most of the cell growth in these tumors occurs at their periphery, where they are fed by diffusion from preexisting normal capillaries (Caspersson and Santesson, 1942); see also Sardari‐Nia et al. (2007) and Pezzella et al. (1997). As a result, the general biosynthetic activity and viability of cells in primary human tumors decreases with distance from the tumor periphery whereas cell size increases followed by necrosis near the center. However, small proliferating cells with high biosynthetic activity can be found more deeply within the human tumors in close proximity to the occasional capillaries that penetrate the tumors.

Cell–Cell Contact in Suppression of Tumor Development

189

There is evidence that soluble factors released within organs may play a role in determining the organ specificity of metastatic growth. Medium conditioned by the growth of each of a variety of mouse organ cultures was used to affect the survival and attachment to the culture dish of cells from 52 spontaneous mammary carcinomas (Horak et al., 1986). Medium conditioned by lung fragments was most successful in supporting the overall survival and attachment of the mammary carcinoma cells, followed by ovary and kidney conditioned media. Liver and thyroid conditioned media had negative effects on almost all the mammary carcinomas. Since mouse mammary tumors develop metastases mainly in the lungs and occasionally in the ovaries or kidneys (if inoculated via the aorta), the effects of these organs on the tumor cells were considered in good agreement with the in vivo observations. Using B16 melanoma cells, it was found that cells with high lung‐ colonizing capacity were growth stimulated by lung‐conditioned medium significantly more than by conditioned medium from other tissues (Nicolson and Dulski, 1986). Similar results were obtained with high ovary‐colonizing capacity B16 cells stimulated by ovary‐conditioned medium. In contrast, however, the growth of brain‐colonizing B16 cells was not stimulated by factors released from brain tissue. Hence, it appears that metastasis of B16 melanoma cells to specific organ sites is dependent not only on preferential target–organ adhesion but may be supplemented by soluble organ‐derived factors.

XIII. MOLECULAR BASIS OF CELL–CELL ADHESION Cell–cell adhesion plays a central role in developmental biology, normal cell function, and neoplasia, and efforts to identify the underlying molecular components of this adhesion have been made in all these fields. Conspicuous efforts were made to identify the membrane proteins involved in the specific sorting out of cells from different tissues during embryonic development, and were recently reviewed (Rubin, 2007). Emphasis was placed on the role of cadherins (Gumbiner and Yamada, 1993; Takeichi, 1991), but the number of molecules of this type is large, and cross adherence among them is so common that distinctive association with particular cell–cell associations could not be achieved (Foty and Steinberg, 2004). As a result, it was recommended that the understanding of adhesive relationships would be better served by focusing on functional measurements of differential cohesive and adhesive relationships of living cells in cell sorting and tumor development (Foty and Steinberg, 2004). This conclusion is reinforced by the large number of other classes of adhesion molecules, which along with

190

Harry Rubin

cadherins, are decreased in some tumors (Mareel et al., 1992) and increased in others (Auersperg et al., 1999; Bindels et al., 2000). Somewhat surprisingly in view of the loss of adherence between cells in human colon cancer (McCutcheon et al., 1948), there is a typical increase in the adhesion molecule carcinoembryonic antigen (CEA) in these cases (Shuster et al., 1980). This increase has been attributed to the shifting of the small amounts of CEA in the luminal surface of normal columnar colorectal epithelium to large amount along the borders between the cells in cancers where they are thought to displace other adhesion molecules (Benchimol et al., 1989). As noted earlier, there is a downregulation of E‐cadherin during the transformation of melanocytes into malignant melanoma cells and its replacement by N‐cadherin (Hsu et al., 2000b). The decrease in E‐cadherin is accompanied by a loss of adhesion to keratinocytes which regulate the growth of melanocytes, but adhesion and regulation can be restored by transducing and overexpressing E‐cadherin in the melanoma cells. It should be noted, however, that the loss of E‐cadherin is considered but one of the mechanisms for the malignant behavior of melanoma cells (Hsu et al., 2000b), since many other surface antigens are lost during transformation (Herlyn et al., 1987). In the case of metastatic cell adhesion to endothelial receptors, at least seven major luminal glycoproteins common to microvessels derived from six different organs have to be taken into account (Fig. 3; Belloni and Nicolson, 1988). Five of the glycoproteins appeared to be expressed differentially in particular organs. Nicolson concluded (Nicolson, 1988) that “tumor cell adhesion mechanisms (for endothelial cells, basement membranes, parenchymal cells, platelets, etc.) are mediated by different multiple adhesion molecules, operating in parallel, and many if not all tumor cells probably use normal cell adhesion mechanisms. In addition, different adhesion systems are involved in tumor cell adhesion to different structures (endothelium versus basement membrane, etc.). Thus organ specificity is determined at one or more levels of tumor interaction such as at the level of microvessel endothelial cell, basement membrane, parenchymal cell and so on.” He also concluded that differential adhesion properties alone are probably insufficient for determining the organ specificity of metastasis, and other mechanisms such as tumor autocrine and paracrine growth factors are likely to be involved. Nicolson’s observations on the number of plasma membrane components likely to be involved in metastatic formation are similar to those made about the multiplicity of cadherins in the homophilic adhesion in the specific sorting out of cells in embryonic development (Foty and Steinberg, 2004). Integrins are also involved in metastatic adhesion, as in the case of 31 integrin interaction with laminin 5 of the vascular basement membrane (Wang et al., 2004). There again the integrin–laminin binding is only part

191

Cell–Cell Contact in Suppression of Tumor Development

700

600

Kidney

500

Intensity

Spleen 400 Liver 300 Lung 200 Heart 100 Brain 0 200

116 97.5

68

43

Apparent Mr

Fig. 3 Densitometric scan of iodinated microvascular endothelial cell surface proteins from various mouse organs. Arrows indicate differentially‐expressed surface proteins. From Belloni and Nicolson (1988) and Nicolson (1988).

of the story since anti‐integrin antibody only partially reduced pulmonary arrest of metastatic cells. The presence of multiple adhesion molecules on the surface of parenchymal cells would buffer the cells of intact tissues against mutations in any one of them, and help to explain the maintenance of their normal phenotype in the face of myriad genetic changes found in every cell (Rubin, 2006).

XIV. CONCLUSIONS 1. Virtually all the examples of suppression of the neoplastic phenotype in Tables I and II involve contact between an excess of normal regulated cells in contact with solitary homotypic cells of neoplastic potential. The basic principles of cell contact suppression were first adduced from studies of

192

2.

3.

4.

5.

Harry Rubin

fibroblast cocultures, but since there were no different degrees of transformation in the fibroblasts, the examples shed no light on the relative susceptibility to suppression at progressive stages of transformation. The most informative observations on the latter point came from studies on normal keratinocytes mixed in excess with initiated papilloma cells or with carcinoma cells. There was strong suppression of the papilloma cells by normal keratinocytes but no suppression of the epidermal carcinoma cells (Strickland et al., 1992). Cell contact was required for suppression of the papilloma cells but GJC was not (Hennings et al., 1992). Hepatocarcinoma cells of moderate malignancy were fully normalized when transplanted as solitary cells in the liver plates, but a more malignant line of hepatocarcinoma cells was only partly modulated in the same conditions (Coleman et al., 1993; McCullough et al., 1998). The implication is that a large intact organ predominantly composed of a single type of parenchymal cells is a more effective modulator of neoplastic behavior than mixtures made in 2‐dimensional cell cultures like those of keratinocytes and neoplastic epidermal cells. There is broad tissue specificity in the suppression of papilloma cells, since fibroblasts do not suppress them (Hennings et al., 1990; Strickland et al., 1992). Similarly, the hepatocarcinoma cells proliferated at a maximal rate when transplanted subcutaneously or leaked from the liver into the peritoneal cavity (Coleman et al., 1993; McCullough et al., 1998). However, in neither case was any attempt made to ascertain the effect of epithelial cells other than those of the same tissue origin as the neoplastic cells. In one atypical case, esophageal cells inhibited tracheal carcinoma cells, but only a small minority of the normal cells was required and most likely involved competition for attachment to the tracheal mesenchyme (Terzaghi‐Howe, 1987). The normalizing effect of liver on the hepatocarcinoma cells declines as the organism ages. This suggests that the ordering capacity of normal cells decreases with age, which may contribute to the widely observed increased incidence of solid epithelial cancers with age. Epidermal cells in an early stage of initiation allows them, unlike normal keratinocytes, to multiply in vitro in adequate calcium, but to make normal epidermis rather than papillomas in grafts. Despite their normal appearance and function, they have lost the capacity to suppress papilloma growth. This also suggests that the immediate cellular microenvironment in chemical carcinogenesis plays a dynamic role in determining the development of tumors. It has, for example, been shown that sustained exposure of skin to ultraviolet light results in the expansion of intraepithelial neoplastic clones, mainly by damaging the suppressive normal epidermis rather than mutagenizing the clones (Mudgil et al., 2003; Zhang et al., 2001).

Cell–Cell Contact in Suppression of Tumor Development

193

6. Little work has been done to identify the adhesive molecules in contact suppression of neoplasia by normal cells. Analysis of adhesion‐driven sorting out of embryonic cells from different tissues indicates that the large number of cadherin subtypes and their overlapping function are too dynamic and complex to evaluate the contribution of each type to the process (Foty and Steinberg, 2004). The authors propose that quantitative, functional measurements of cell–cell adhesion provide a more reliable predictor of tissue‐specific sorting out than does quantitation of various adhesion molecules. The same conclusion may be applicable in tumor suppression in light of the inconsistency of relating particular adhesion molecules to malignant transformation (Rubin, 2007). 7. Almost all the major insights on cell–cell interaction in tumor suppression in Tables I and II came from operational experiments without regard to molecular reduction. This type of operational experiment seems to have largely disappeared in recent years, but its usefulness has hardly been exhausted. There are a number of operational experiments that can be done to further evaluate the role of the tumor suppressive activity of cell contact interactions in neoplastic development. Some such experiments are listed as follows in terms of epidermal carcinogenesis in which the in vivo dynamics of initiation and promotion are best understood, but the basic design can be adapted to other neoplastic cell systems. They include: (a) the effect of age on suppression of papilloma cells by normal keratinocytes; (b) the capacity of other epithelial cells such as those from the esophagus to suppress epidermal papilloma cells; (c) comparison of the suppressive capacity of keratinocytes from sensitive and resistant mouse strains; and (d) the effect of carcinogen treatment on the suppressive capacity of keratinocytes. Given the insights already obtained from the normalization of hepatocarcinoma cells transplanted into the liver, the development of a cell culture model of this system would permit a range of quantitative, functional experiments that were not possible to do in vivo. 8. The tumor suppressive effects of a large number of contacting homotypic normal cells may be subsumed under the basic biological principle of “order in the large over heterogeneity in the small” (Elsasser, 1998; Rubin, 2006, 2007). The great heterogeneity of cancer cells has been demonstrated for many behavioral characteristics (Fidler, 1978; Heppner, 1984), and more recently for mRNA sequences (Brulliard et al., 2007). Cellular (Fry et al., 1966) and molecular genetic (Bahar et al., 2006) heterogeneity are also characteristic of aging. Such age‐ related increase in the heterogeneity of cells could account for the reduced capacity of aging organs to normalize the behavior of neoplastic cells (Coleman et al., 1993; McCullough et al., 1998). 9. There is no credible evidence that heterotypic cell–cell adhesion induces metastatic dormancy. Rather, the adhesion of DCC to endothelial or

194

Harry Rubin

parenchymal cells activates metastatic growth, but fortunately involves only a small fraction of the disseminated cells (Luzzi et al., 1998; Naumov et al., 2002; Suzuki et al., 2006). Significant factors in metastatic dormancy are the lack of capacity of most cancer cells to adhere to heterotypic receptors of host organ cells, and the lack of specific soluble growth factors at the site. The requirement for adhesion in metastasis may be related to the requirement for cells to attach to and spread on a solid substratum in culture in order to proliferate. 10. Efforts have been made to characterize the cell surface molecules of endothelial cells that might bind DCC to initiate metastases (McGuire et al., 1984; Nicolson and Winkelhake, 1975). A large number of such molecular species are common to many different organs but vary in quantity; there are however few that are specific to individual organs (Fig. 3; Belloni and Nicolson, 1988; Nicolson, 1988). Antibody to a particular endothelial surface integrin only partially reduced the frequency of metastases to the lung of fibrosarcoma cells (Wang et al., 2004). As was the case with cadherins in cell sorting, the number of molecules that bind metastatic cells is likely to be large and cross‐ reactive, so a full accounting of their role in promoting adhesion will be complex. Some idea of the complexity of the factors involved in metastasis is seen in Fig. 4 (Nicolson, 2002), which was later considered

Tumor and host microenvironment Platelets Growth factors Coagulation factors Enzymes

T cells Cytokines Cytolytic factors Degradative enzymes Macrophages Cytolytic factors Cytostatic factors Mutagenic factors Growth factors

Matrix

Tumor cells Cytolic factors Chemotactic factors Degradative enzymes Procoagulants, MIFs Secreted antigens Parenchymal cells Growth factors Growth inhibitors Nutritional factors Hormones, proteins

Endothelial cells Growth factors Growth inhibitors Fibrinolytic factors Enzymes, nutrients

NK & LAK Cells Cytokines Cytolytic factors Degradative enzymes

Mast cells Growth factors Glycosaminoglycans Extracellular matrix Cytokines, histamine Degradative enzymes

Fibroblasts Growth factors Differentiation factors Inductive factors Other cytokines Extracellular matrix Degradative enzymes Enzyme inhibitors

G. L. Nicolson 8/92

Fig. 4 Tumor and host microenvironmental factors involved in metastasis. From Nicolson (2002).

Cell–Cell Contact in Suppression of Tumor Development

195

to represent an underestimate of the number of molecules involved (Mareel, 2004). Physical measurements of the strength of adhesions among living cells, e.g., (Coman, 1944; Foty et al., 1996; McCutcheon et al., 1948) could be a more productive approach to their meaning for metastasis. Gene expression profiles characteristic of the metastatic process indicate a formidable degree of complexity. About 75 host‐regulated genes were found to differ in expression between highly and poorly metastatic clones from the same human breast carcinoma (Montel et al., 2006). An expression pattern of 128 genes is required to distinguish primary and metastatic carcinomas (Ramaswamy et al., 2003). No doubt others will be found in other metastatic situations. The problem falls into Elsasser’s rule that “causal chains [in organisms] cannot be traced beyond a terminal point because they are lost in unfathomable complexity. . .” (Elsasser, 1998). Therefore, the pervasive preoccupation with searching for causal chains in neoplasia at the molecular level should be complemented by operational studies on cell–cell interactions to understand the primary microenvironmental aspects of tumor development. The importance of an overarching biological theory (Elsasser, 1998) for structuring and unifying the diversity of observations that are sure to follow cannot be overemphasized.

ACKNOWLEDGMENTS I am grateful for the manuscript preparation and editing by Dorothy M. Rubin. Correspondence with Ann Chambers, Ruth Muschel, Garth Nicolson and David Tarin provided useful insights on modern views of tumor metastasis. Discussion with David Bilder clarified points about Drosophila tumors. George Klein made helpful suggestions about the manuscript.

REFERENCES Abercrombie, M., and Ambrose, E. J. (1958). Interference microscope studies of cell contacts in tissue culture. Exp. Cell Res. 15, 332–345. Abercrombie, M., and Heaysman, J. E. M. (1954). Observations on the social behavior of cells in tissue culture. II. “Monolayering” of fibroblasts. Exp. Cell Res. 6, 293–306. Aguirre‐Ghiso, J. A., Estrada, Y., Hiu, D., and Ossowski, L. (2003). ERKMAPK activity as a determinant of tumor growth and dormancy: regulation of p38SAPK1. Cancer Res. 63, 1684–1695. Al‐Mehdi, A. B., Tozawa, K., Fisher, A. B., Shientag, L., Lee, A., and Muschel, R. J. (2000). Intravascular origin of metastasis from the proliferation of endothelium‐attached tumor cells: A new model for metastasis. Nature Med. 6, 100–102. Alexander, D. B., Ichikawa, H., Bechberger, J. F., Valiunas, V., Ohki, M., Naus, C. C. G., Kunimoto, T., Tsuda, H., Miller, W. T., and Goldberg, G. S. (2004). Normal cells control the growth of neighboring transformed cells independent of gap junctional communication and Src activity. Cancer Res. 64, 1347–1358.

196

Harry Rubin

Alt‐Holland, A., Zhang, W., Margulis, A., and Garlick, J. (2005). Microenvironmental control of premalignant disease: The role of intercellular adhesion in the progression of squamous cell carcinoma. Sem. Cancer Biol. 15, 84–96. Auerbach, R., Lu, W. C., Pardon, E., Gumkowski, F., Kaminska, G., and Kaminski, M. (1987). Specificity of adhesion between murine tumor cells and capillary endothelium: An in vitro correlate of preferential metastasis in vivo. Cancer Res. 47, 1492–1496. Auersperg, N., Pan, J., Grove, B. D., Peterson, T., Fisher, J., Maines‐Bandiera, S., Somasiri, A., and Rosekelley, C. D. (1999). E‐cadherin induces mesenchymal‐to‐epithelial transition in human ovarian epithelium. Proc. Natl. Acad. Sci. USA 96, 6249–6254. Bahar, R., Hartmann, C., Rodriguez, K. A., Denny, A. D., Busuttil, R. A., Dolle´, M. E. T., Calder, R. B., Chisholm, G. B., Pollock, B. H., Klein, C. A., and Vijg, J. (2006). Increased cell‐ to‐cell variation in gene expression in ageing mouse heart. Nature 441, 1011–1014. Balaban, G. B., Herlyn, M., Guerry, D., Bartolo, R., Koprowski, H., Clark, W. H., Jr., and Nowell, P. C. (1984). Cytogenetics of human malignant melanoma and premalignant lesions. Cancer Genet. Cytogenet. 11, 429–439. Belloni, P. N., and Nicolson, G. L. (1988). Differential expression of cell surface glycoproteins on various organ‐derived microvascular endothelia and endothelial cell cultures. J. Cell Physiol. 136, 398–410. Belloni, P. N., Tressler, R. J., and Nicolson, G. L. (1986). Identification of endothelial cell surface glycoproteins associated with preferential adhesion of metastatic murine RAW117 large cell lymphoma to organ‐derived endothelial cells. J. Cell Biol. 103, 461a. Benchimol, S., Fuks, A., Jothy, S., Beauchemin, N., Shirota, K., and Stanners, C. P. (1989). Carcinoembryonic antigen, a human tumor marker, functions as an intercellular adhesion molecule. Cell 57, 327–334. Bilder, D. (2004). Epithelial polarity and proliferation control: Links from Drosophila neoplastic tumor suppressors. Genes Dev. 18, 1909–1925. Bindels, E. M. J., Vermey, M., van den Beemd, R., Dinjens, W. N. M., and Van Der Kwast, T. H. (2000). E‐cadherin promotes intraepithelial expansion of bladder carcinoma cells in an in vitro model of carcinoma in situ. Cancer Res. 60, 177–183. Bowen‐Pope, D. F., and Rubin, H. (1983). Growth stimulatory precipitates of Ca2þ and pyrophosphate. J. Cell. Physiol. 117, 51–61. Brulliard, M., Lorphelin, D., Collignon, O., Lorphelin, W., Thouvenot, B., Gothie, E., Jacquenet, S., Ogier, V., Roitel, O., Monnez, J.‐M., Vallois, P., Yen, F. T., et al. (2007). Nonrandom variations in human cancer ESTs indicate that mRNA heterogeneity increases during carcinogenesis. Proc. Natl. Acad. Sci. USA 104, 7522–7527. Brumby, A. M., and Richardson, H. S. (2003). scribble mutants cooperate with oncogenic Ras or Notch to cause neoplastic overgrowth in Drosophila. EMBO J. 22, 5769–5779. Cairns, J. (1978). “Cancer: Science and Society.” W.H. Freeman, San Francisco. Cameron, M. D., Schmidt, E. E., Kerkvliet, N., Nadkarni, K. V., Morris, V. L., Groom, A. C., Chambers, A. F., and MacDonald, I. C. (2000). Temporal progression of metastasis in lung: Cell survival, dormancy and location dependence of metastatic inefficiency. Cancer Res. 60, 2541–2546. Carney, D. H., and Cunningham, D. D. (1977). Initiation of chick cell division by trypsin action at the cell surface. Nature 268, 602–606. Caspersson, T., and Santesson, L. (1942). Studies on protein metabolism in the cells of epithelial tumors. Acta Radiol. Suppl. 46, 5–105. Castor, L. N. (1968). Contact regulation of cell division in an epithelial‐like cell line. J. Cell Physiol. 72, 161–172. Cha, R. S., Thilly, W. G., and Zarbl, H. (1994). N‐Nitroso‐N‐methylurea‐induced rat mammary tumors arise from cells with preexisting oncogenic Hras 1 gene mutations. Proc. Natl. Acad. Sci. USA 91, 3749–3753.

Cell–Cell Contact in Suppression of Tumor Development

197

Chekmareva, M. A., Kadkhodaian, M. M., Hollowell, C. M., Kim, H., Yoshida, B. A., Stadler, W. M., and Rinker‐Schaeffer, C. W. (1998). Chromosome 17‐mediated dormancy of AT6.1 prostate cancer micrometastases. Cancer Res. 58, 4963–4969. Clark, W. H. (1994). From the melanocyte to melanoma to tumor biology. Adv. Cancer Res. 65, 113–140. Clark, W. H., Elder, D. E., and Van Horn, M. (1986). The biologic forms of malignant melanoma. Human Pathol. 17, 443–450. Clark, W. H., Jr., Elder, D. E., Guerry, D., IV, Epstein, M. N., Greene, M. H., and Van Horne, M. (1984). A study of tumor progression: The precursor lesions of superficial spreading and nodular melanoma. Human Pathol. 15, 1147–1165. Coleman, W. B., Wennerberg, A. E., Smith, G. J., and Grisham, J. W. (1993). Regulation of the differentiation of diploid and some aneuploid rat liver epithelial (stemlike) cells by the hepatic microenvironment. Am. J. Pathol. 142, 1373–1382. Coman, D. R. (1944). Decreased mutual adhesiveness, a property of cells from squamous cell carcinomas. Cancer Res. 4, 625–629. Coman, D. R. (1953). Mechanisms responsible for the origin and distribution of blood‐borne tumor metastases: A review. Cancer Res. 13, 397–404. Cramer, W., and Stowell, R. E. (1942). The early stages of carcinogenesis by 20‐methylcholanthrene in the skin of the mouse. I. Experimental technique and macroscopic changes. J. Natl. Cancer Inst. 2, 369–378. Crissman, J. D., Hatfield, J., Schaldenbrand, M., Sloane, B. F., and Honn, K. V. (1985). Arrest and extravasation of B16 amelanotic melanoma in murine lungs. A light and electron study. Lab. Invest. 53, 470–478. Cuatrecasas, P. (1969). Interaction of insulin with the cell membrane: The primary action of insulin. Proc. Natl. Acad. Sci. USA 63, 450–457. Demicheli, R., Abattista, A., Miceli, R., Valagussa, P., and Bonadonna, G. (1996). Time distribution of the recurrence risk for breast cancer patients undergoing mastectomy: Further support for the concept of tumor dormancy. Breast Cancer Res. Treat. 41, 177–185. DeOme, K., Miyamoto, M., Osborn, R., Guzman, R., and Lum, K. (1978). Detection of inapparent nodule‐transformed cells in the mammary gland tissue of virgin female BALB/cfC3H mice. Cancer Res. 38, 2103–2111. Eagle, H., and Levine, E. M. (1967). Growth regulatory effects of cellular interaction. Nature 213, 1102–1106. Elsasser, W. M. (1998). “Reflections on a Theory of Organisms.” Johns Hopkins University Press, Baltimore, MD. Faulkin, L. J., Jr., and DeOme, K. B. (1960). Regulation of growth and spacing of gland elements in the mammary fat pad of the C3H mouse. J. Natl. Cancer Inst. 24, 953–968. Fidler, I. J. (1978). Tumor heterogeneity and the biology of cancer invasion and metastasis. Cancer Res. 38, 2651–2660. Fischer, A. (1946). “Biology of Tissue Cells.” Cambridge University Press, Cambridge, England. Fisher, B., and Fisher, E. R. (1959). Experimental evidence in support of the dormant tumor cell. Science 130, 918–919. Foty, R. A., Pfleger, C. M., Forgacs, G., and Steinberg, M. S. (1996). Surface tensions of embryonic tissues predict their mutual envelopment behavior. Development 122, 1611–1620. Foty, R. A., and Steinberg, M. S. (2004). Cadherin‐mediated cell–cell adhesion and tissue segregation in relation to malignancy. Int. J. Dev. Biol. 48, 397–409. Fry, R. J. M., Tyler, S. A., and Lesher, S. (1966). In “Radiation and Ageing” (P. J. Lindop and G. A. Sacher, Eds.), pp. 43–55. Taylor & Francis, London. Fukunaga‐Kalabis, M., Martinez, G., Liu, Z.‐J., Kalabis, J., Mrass, P., Weninger, W., Firth, S. M., Planque, N., Perbal, B., and Herlyn, M. (2006). CCN3 controls 3D spatial localization of melanocytes in the human skin through DDR1. J. Cell Biol. 175, 563–569.

198

Harry Rubin

Gateff, E. (1978). Malignant neoplasms of genetic origin in Drosophila melanogaster. Science 200, 1448–1459. Gibson, T., and Norris, W. (1958). Skin fragments removed by injection needles. Lancet 4, 983–985. Gillett, N. A., Thomassen, D. G., Buchholz, A., and Kelly, G. (1989). Cell‐line‐dependent inhibition of phenotypic expression of preneoplastic and neoplastic tracheal epithelial cells by normal tracheal cells in denuded tracheal transplants. Carcinogenesis 10, 1147–1154. Goodison, S., Kawai, K., Hibara, J., Jiang, P., Yang, M., Urquidi, V., Hoffman, R. M., and Tarin, D. (2003). Prolonged dormancy and site‐specific growth potential of cancer cells spontaneously disseminated from nonmetastatic breast tumors as revealed by labeling with green fluorescent protein. Clin. Cancer Res. 9, 3808–3814. Gumbiner, B. M., and Yamada, K. M. (1993). Cell‐to‐cell contact and extracellular matrix. Curr. Opin. Cell Biol. 5, 769–771. Hennings, H., Lowry, D. T., Robinson, V. A., Morgan, D. L., Fujiki, H., and Yuspa, S. H. (1992). Activity of diverse tumor promoters in a keratinocyte co‐culture model of initiated epidermis. Carcinogenesis 13, 2145–2151. Hennings, H., Michael, D., Cheng, C., Steinert, P., Holbrook, K., and Yuspa, S. H. (1980). Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell 19, 245–254. Hennings, H., Robinson, V. A., Michael, D. M., Pettit, G. R., Jung, R., and Yuspa, S. H. (1990). Development of an in vitro analogue of initiated mouse epidermis to study tumor promoters and antipromoters. Cancer Res. 50, 4794–4800. Heppner, G. H. (1984). Tumor heterogeneity. Cancer Res. 44, 2259–2265. Herlyn, M., Rodeck, U., Mancianti, M.‐L., Cardillo, F. M., Lang, A., Ross, A. H., Jambrosic, J., and Koprowski, H. (1987). Expression of melanoma‐associated antigens in rapidly dividing human melanocytes in culture. Cancer Res. 47, 3057–3061. Herlyn, M., Balaban, G., Bennicelli, J., Guerry, D., Halaban, R., Herlyn, D., Elder, D. E., Maul, G. G., Steplewski, Z., Nowell, P. C., Clark, W. H., and Koprowski, H. (1985a). Primary melanoma cells of the vertical growth phase: Similarities to metastatic cells. JNCI 74, 283–289. Herlyn, M., Thurin, J., Balaban, G., Bennicelli, J., Herlyn, D., Elder, D. E., Bondi, E., Guerry, D., Nowell, P. C., Clark, W. H., and Koprowski, H. (1985b). Characteristics of cultured human melanocytes isolated from different stages of tumor progression. Cancer Res. 45, 5670–5676. Horak, E., Darling, D. L., and Tarin, D. (1986). Analysis of organ‐specific effects on metastatic tumor formation by studies in vitro. J. Natl. Cancer Inst. 76, 913–922. Hsu, M., Andl, T., Li, G., Meinkoth, J. L., and Herlyn, M. (2000a). Cadherin repertoire determines partner‐specific gap junctional communication during melanoma progression. J. Cell Sci. 113, 1535–1542. Hsu, M.‐Y., Meier, F. E., Nesbit, M., Hsu, J.‐Y., Van Belle, P., Elder, D. E., and Herlyn, M. (2000b). E‐cadherin expression in melanoma cells restores keratinocyte‐mediated growth control and down‐regulates expression of invasion‐related adhesion receptors. Am. J. Pathol. 156, 1515–1525. Javaherian, A., Vaccariello, M., Fusenig, N. E., and Garlick, J. A. (1998). Normal keratinocytes suppress early stages of neoplastic progression in stratified epithelium. Cancer Res. 58, 2200–2208. Jonason, A. S., Kunala, S., Price, G. J., Restifo, R. J., Spinell, H. M., Persing, J. A., Leffell, D. J., Tarone, R. E., and Brash, D. E. (1996). Frequent clones of p53‐mutated keratinocytes in normal human skin. Proc. Natl. Acad. USA 93, 14025–14029. Kanno, Y., Sasaki, Y., Shiba, Y., Yoshida‐Noro, C., and Takeichi, M. (1984). Monoclonal antibody ECCD‐1 inhibits intercellular communication in teratocarcinoma PCC3 cells. Exp. Cell Res. 152, 270–274.

Cell–Cell Contact in Suppression of Tumor Development

199

Kasami, M., Vnencak‐Jones, C. L., Manning, S., Dupont, W. D., and Page, D. L. (1997). Loss of heterozygosity and microsatellite instability in breast neoplasia. No obligate correlation of these genetic alterations with subsequent malignancy. Am. J. Pathol. 150, 1925–1932. Kieran, M. W., and Longenecker, B. M. (1983). Organ specific metastasis with special reference to avian systems. Cancer Metastasis Rev. 2, 165–182. Klein, G. (1951). Comparative studies of mouse tumors with respect to their capacity for growth as “ascites tumors” and their average nucleic aid content per cell. Exp. Cell Res. 2, 518–573. Koop, S., MacDonald, I. C., Luzzi, K., Schmidt, E. E., Morris, V. I., Grattan, M., Khokha, R., Chambers, A. F., and Groom, A. C. (1995). Fate of melanoma cells entering the microcirculation: over 80% survive and extravasate. Cancer Res. 55, 2520–2523. Koop, S., Schmidt, E. E., MacDonald, I. C., Morris, V. I., Khokha, R., Grattan, M., Leone, J., Chambers, A. F., and Groom, A. C. (1996). Independence of metastatic ability and extravasation: Metastatic ras‐transformed and control fibroblasts extravasate equally well. Proc. Natl. Acad. Sci. USA 93, 11080–11084. Kulesz‐Martin, M. F., Kilkenny, A. E., Holbrook, K. A., Digernes, V., and Yuspa, S. H. (1983). Properties of carcinogen altered mouse epidermal cells resistant to calcium‐induced terminal differentiation. Carcinogenesis 4, 1367–1377. Kulesz‐Martin, M. F., Koehler, B., Hennings, H., and Yuspa, S. H. (1980). Quantitative assay for carcinogen altered differentiation in mouse epidermal cells. Carcinogenesis 1, 995–1006. Lapis, K., Paku, S., and Liotta, L. A. (1988). Endothelialization of embolized tumor cells during metastasis formation. Clin. Exp. Metastasis 6, 73–89. Li, G., Fukunaga, M., and Herlyn, M. (2004). Reversal of melanocytic malignancy by keratinocytes is an E‐cadherin‐mediated process overriding ‐catenin signaling. Exp. Cell Res. 297, 142–151. Luzzi, K. J., MacDonald, I. C., Schmidt, E. E., Kerkvliet, N., Morris, V. L., Chambers, A. F., and Groom, A. C. (1998). Dormancy of solitary cells after successful extravasation and limited survival of early micrometastases. Am. J. Path. 153, 865–873. Mareel, M. (2004). Molecular cell biology and cancer metastasis. An interview with Garth Nicolson. Int. J. Dev. Biol. 48, 355–363. Mareel, M., Van Roy, F., and De Baetselier, P. (1990). The invasive phenotypes. Cancer Metastasis Rev. 9, 45–62. Mareel, M., Vermeulen, S., Bracke, M., and Van Roy, F. (1992). E‐cadherin expression: A counterbalance for cancer cell invasion. Bull. Cancer 79, 347–355. Martin, W., Zempel, G., Hu¨lser, D., and Willecke, K. (1991). Growth inhibition of oncogene‐ transformed rat fibroblasts by cultured normal cells: Relevance of metabolic cooperation mediated by gap junctions. Cancer Res. 51, 5348–5354. McCullough, K. D., Coleman, W. B., Ricketts, S. L., Wilson, J. W., Smith, G. J., and Grisham, J. W. (1998). Plasticity of the neoplastic phenotype in vivo is regulated by epigenetic factors. Proc. Natl. Acad. Sci. USA 95, 15333–15338. McCutcheon, M., Coman, D. R., and Moore, F. B. (1948). Studies on invasiveness of cancer. Adhesiveness of malignant cells in various human adenocarcinomas. Cancer 1, 460–467. McGuire, E. J., Mascali, J. J., Grady, S. R., and Nicolson, G. L. (1984). Involvement of cell–cell adhesion molecules in liver colonization by metastatic murine lymphoma/lymphosarcoma variants. Clin. Exp. Metastasis 2, 213–222. Medina, D., Shepherd, F., and Gropp, T. R. (1978). Enhancement of the tumorigenicity of preneoplastic mammary nodule lines by enzymatic dissociation. J. Natl. Cancer Inst. 60, 1121–1126. Mehta, P. P., Bertram, J. S., and Loewenstein, W. R. (1986). Growth inhibition of transformed cells correlates with their junctional communication with normal cells. Cell 44, 187–196. Middleton, C. A. (1973). In “Locomotion of Tissue Cells” (R. Porter and D. W. Fitzsimons, Eds.), pp. 251–270. Elsevier, Amsterdam.

200

Harry Rubin

Montel, V., Mose, E. S., and Tarin, D. (2006). Tumor‐stromal interactions reciprocally modulate gene expression patterns during carcinogenesis and metastasis. Int. J. Cancer 119, 251–263. Mudgil, A. V., Segal, N., Andriani, F., Wang, Y., Fusenig, N. E., and Garlick, J. A. (2003). Ultraviolet B irradiation induces expansion of intraepithelial tumor cells in a tissue model of early cancer progression. J. Invest. Dermatol. 121, 191–197. Nakamura, T., Yoshimoto, K., Nakayama, Y., Tomita, Y., and Ichihara, A. (1983). Reciprocal modulation of growth and differentiated functions of mature rat hepatocytes in primary culture by cell–cell contact and cell membranes. Proc. Natl. Acad. Sci. USA 80, 7229–7233. Naumov, G. N., Bender, E., Zurakowski, D., Kang, S.‐Y., Sampson, D., Flynn, E., Watnick, R. S., Straume, O., Akslen, L. A., Folkman, J., and Almog, N. (2006). A model of human tumor dormancy: An angioswitch from the nonangiogenic phenotype. JNCI 98, 316–325. Naumov, G. N., MacDonald, I. C., Weinmeister, P. M., Kerkvliet, N., Nadkarni, K. V., Wilson, S. M., Morris, V. L., Groom, A. C., and Chambers, A. F. (2002). Persistence of solitary mammary carcinoma cells in a secondary site. Cancer Res. 62, 2162–2168. Nelson, C. M., VanDuijn, M. M., Inman, J. L., Fletcher, D. A., and Bissell, M. J. (2006). Tissue geometry determines sites of mammary branching morphogenesis in organotypic cultures. Science 314, 298–300. Nicolson, G. L. (1988). Cancer metastasis: tumor cell and host organ properties important in metastasis to specific secondary sites. Biochem. Biophys. Acta 948, 175–224. Nicolson, G. L. (2002). Autocrine and paracrine growth mechanisms in cancer progression and metastasis. In “Encyclopedia of Cancer,” Vol. 1, 2nd edn, pp. 165–177. Elsevier, Amsterdam. Nicolson, G. L., and Dulski, K. M. (1986). Organ specificity of metastatic tumor colonization is related to organ‐selective growth properties of malignant cells. Int. J. Cancer 38, 289–294. Nicolson, G. L., and Winkelhake, J. L. (1975). Organ specificity of blood‐borne tumour metastasis determined by cell adhesion? Nature 255, 230–232. Pasqualini, R., and Ruoslahti, E. (1996). Organ targeting in vivo using phage display peptide libraries. Nature 380, 364–365. Pawlowski, A., Haberman, H. F., and Menon, I. A. (1976). Junctional and compound pigmented nevi induced by 9,10‐dimethyl‐1,2‐benzanthracene in skin of albino guinea pigs. Cancer Res. 36, 2813–2821. Pezzella, F., Pastorino, U., Tagliabue, E., Andreila, S., Sozzi, G., Gasparini, G., Menard, S., Gatter, K. C., Harris, A. L., KFox, S., Buyse, M., Pilotti, S., et al. (1997). Non‐small‐cell lung carcinoma tumor growth without morphological evidence of neo‐angiogenesis. Am. J. Path. 151, 1417–1423. Phondke, G. P., Madyastha, K. R., Madyastha, P. R., and Barth, R. F. (1981). Relationship between concanavalin A‐induced agglutinability of murine leukemia cells and their propensity to form heterotypic aggregates with syngeneic lymphoid cells. JNCI 66, 643–647. Pitot, H. C. (2002). “Fundamentals of Oncology.” Marcel Dekker, New York, NY. Purdom, L., and Ambrose, E. J. (1958). A correlation between electrical surface charge and some biological characteristics during the stepwise progression of a mouse sarcoma. Nature 181, 1586–1587. Ramaswamy, S., Ross, K. N., and Golub, T. R. (2003). A molecular signature of metastasis in primary solid tumors. Nature Genet. 33, 49–54. Ringertz, N., Klein, E., and Klein, G. (1957). Histopathologic studies of peritoneal implantation and lung metastasis at different stages of the gradual transformation of the MC1M mouse sarcoma into ascites form. J. Natl. Cancer Inst. 18, 173–196. Rubin, H. (1960a). An analysis of the assay of Rous sarcoma cells in vitro by the infective center technique. Virology 10, 29–49.

Cell–Cell Contact in Suppression of Tumor Development

201

Rubin, H. (1960b). The suppression of morphological alterations in cells infected with Rous sarcoma virus. Virology 12, 14–31. Rubin, H. (1970). Overgrowth stimulating factor released from Rous sarcoma cells. Science 167, 1271–1272. Rubin, H. (1975). Central role for magnesium in coordinate control of metabolism and growth in animal cells. Proc. Natl. Acad. Sci. USA 72, 3551–3555. Rubin, H. (1982). Effect of magnesium content on density‐dependent regulation of the onset of DNA synthesis in transformed 3T3 cells. Cancer Res. 42, 1761–1768. Rubin, H. (1994). Cellular epigenetics: Control of the size, shape and spatial distribution of transformed foci by interactions between the transformed and non‐transformed cells. Proc. Natl. Acad. Sci. USA 91, 1039–1043. Rubin, H. (2001). Synergistic mechanisms in carcinogenesis by polycyclic aromatic hydrocarbons and tobacco smoke: A biohistorical perspective with updates. Carcinogenesis 22, 1903–1930. Rubin, H. (2003). Microenvironmental regulation of the initiated cell. Adv. Cancer Res. 90, 1–62. Rubin, H. (2005). Central roles of Mg2þ and of MgATP2 in the regulation of protein synthesis and cell proliferation: Significance for neoplastic transformation. Adv. Cancer Res. 93, 1–58. Rubin, H. (2006). What keeps cells in tissues behaving normally in the face of myriad mutations? BioEssays 28, 515–524. Rubin, H. (2007). Ordered heterogeneity and its decline in cancer and aging. Adv. Cancer Res. 98, 117–147. Rubin, H., and Chu, B. (1982). Self‐normalization of highly transformed 3T3 cells through maximized contact interaction. Proc. Natl. Acad. Sci. USA 79, 1903–1907. Rusciano, D., and Burger, M. M. (1992). Why do cancer cells metastasize into particular organs. BioEssays 14, 185–194. Sardari‐Nia, P., Van Marck, E., and Van Schuil, P. (2007). Re: A model of human tumor dormancy: An angiogenic escape from the nonangiogenic phenotype. J. Natl. Cancer Inst. 99, 331. Schirrmacher, V., Cheingsong‐Popov, R., and Arnheiter, H. (1980). Hepatocyte‐tumor cell interaction in vitro. I. Conditions for rosette formation and inhibition by anti‐H‐2 antibody. J. Exp. Med. 151, 984–989. Sefton, B. M., and Rubin, H. (1970). Release from density dependent growth inhibition by proteolytic enzymes. Nature 227, 843–845. Shih, I.‐M., Elder, D. E., Hsu, M.‐Y, and Herlyn, M. (1994). Regulation of Mel‐CAM/MUC18 expression on melanocytes of different stages of tumor progression by normal keratinocytes. Am. J. Pathol. 145, 837–845. Shuster, J., Thomson, D. M. P., Fuks, A., and Gold, P. (1980). Immunologic approaches to the diagnosis of malignancy. Prog. Exp. Tumor Res. 25, 89–139. Stoker, M. (1964). Regulation of growth and orientation in hamster cells transformed by polyomavirus. Virology 24, 165–174. Stoker, M. (1967). Transfer of growth inhibition between normal and virus transformed cells: Autoradiographic studies using marked cells. J. Cell Sci. 2, 293–304. Stoker, M., Shearer, M., and O’Neill, C. (1966). Growth inhibition of polyoma‐transformed cells by contact with static normal fibroblasts. J. Cell. Sci. 1, 297–310. Strickland, J. E., Ueda, M., Hennings, H., and Yuspa, S. H. (1992). A model for initiated mouse skin: Suppression of cells in grafts on athymic nude mice. Cancer Res. 52, 1439–1444. Suzuki, M., Mose, E. S., Montel, V., and Tarin, D. (2006). Dormant cancer cells retrieved from metastasis‐free organs regain tumorigenic and metastatic potency. Am. J. Path. 169, 673–681.

202

Harry Rubin

Takeichi, M. (1991). Cadherin: Cell adhesion receptors as a morphogenetic regulator. Science 251, 1451–1455. Taptiklis, N. (1968). Dormancy by dissociated thyroid cells in the lings of mice. Eur. J. Cancer 4, 59–66. Taptiklis, N. (1969). Penetration of the vascular endothelial barrier by non‐neoplastic thyroid cells in circulation. Eur. J. Cancer 5, 445–457. Tarin, D. (1972). In “Tissue Interactions in Carcinogenesis” (D. Tarin, Ed.), pp. 227–289. Academic Press, London and New York. Terzaghi‐Howe, M. (1987). Inhibition of carcinogen‐altered rat tracheal epithelial cell proliferation by normal epithelial cells in vivo. Carcinogenesis 8, 145–150. Todaro, G. J., Lazar, G. K., and Green, H. (1965). The initiation of cell division in a contact‐ inhibited mammalian cell line. J. Cell. Comp. Physiol. 66, 325–334. Tsai, Y. C., Lu, Y., Nichols, P. W., Zlotnikov, G., Jones, P. A., and Smith, H. S. (1996). Contiguous patches of normal human mammary epithelium derived from a single stem cell. Cancer Res. 56, 402–404. Uehara, H., Kim, S. J., Karashima, T., Shepherd, D. L., Fan, D., Tsan, R., Killion, J. J., Logothetis, C., Mathew, P., and Fidler, I. J. (2003). Effects of blocking platelet‐derived growth factor‐receptor signaling in a mouse model of experimental prostate cancer bone metastases. JNCI 95, 458–470. Unkeless, J. C., Tobia, A., Ossowski, L., Quigley, J. P., Rifkin, D. B., and Reich, E. (1973). An enzymatic function associated with transformation of fibroblasts by oncogenic viruses I. Chick embryo fibroblast cultures transformed by avian RNA tumor viruses. J. Exp. Med. 137, 85–111. Valyi‐Nagy, I. T., Hirka, G., Jensen, P. J., Shih, I.‐M., Juhasz, I., and Herlyn, M. (1993). Undifferentiated keratinocytes control growth, morphology, and antigen expression of normal melanocytes through cell–cell contact. Lab. Investigation 69, 152–159. Wang, H., Fu, W., Im, J. H., Zhou, Z., Santoro, S. A., Iyer, V., DiPersio, C. M., Yu, Q.‐C., Quaranta, V., Al‐Mehdi, A. B., and Muschel, R. J. (2004). Tumor cell 31 integrin and vascular laminin‐5 mediate pulmonary arrest and metastasis. J. Cell Biol. 164, 935–941. Weinberg, R. A. (2007). “The Biology of Cancer.” Garland Science, New York, NY. Wieser, R. J., Heck, R., and Oesch, F. (1985). Involvement of plasma membranes glycoproteins in the contact‐dependent inhibition of growth of human fibroblasts. Exp. Cell Res. 158, 493–499. Wittenberger, B., and Glaser, L. (1977). Inhibition of DNA synthesis in cultures of 3T3 cells by isolated surface membranes. Proc. Natl. Acad. Sci. USA 74, 2251–2255. Yu, W., Kim, J., and Ossowski, L. (1997). Reduction in surface urokinase receptor forces malignant cells into a protracted state of dormancy. J. Cell Biol. 137, 767–777. Zhang, W., Hanks, A. N., Boucher, K., Florell, S. R., Allen, S. M., Alexander, A., Brash, D. E., and Grossman, D. (2005). UVB‐induced apoptosis drives clonal expansion during skin tumor development. Carcinogenesis 26, 249–257. Zhang, W., Remenyik, E., Zelterman, D., Brash, D. E., and Wikonkai, N. M. (2001). Escaping the stem cell compartment: Sustained UVB exposure allows mutant keratinocytes to colonize adjacent epidermal proliferating units without incurring additional mutations. Proc. Natl. Acad. Sci. USA 98, 13948–13953. Ziegler, A., Leffell, D. J., Kunala, S., Sharma, H. W., Gailani, M., Simon, J. A., Halperin, A. J., Baden, H. P., Shapiro, P. E., Bale, A. E., and Brash, D. E. (1993). Mutation hotspots due to sunlight in the p53 gene of nonmelanoma skin cancers. Proc. Natl. Acad. USA 90, 4216–4220.

Tumor–Microenvironment Interactions: Dangerous Liaisons Isaac P. Witz Department of Cell Research and Immunology, The George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv, 69978, Israel

I. II. III. IV. V. VI. VII. VIII.

The Tumor Microenvironment Stress Responses Interactions With Fibroblasts Tumor–Endothelium Interactions Tumor–Macrophage Interactions The Premetastatic Niche Tumor–Immunoglobulin Interactions Examining the Big Picture References

The interaction between microenvironmental components and tumor cells is bidirectional. Tumor cells and their products are capable of regulating and altering gene expression in nontumor cells residing in or infiltrating into the microenvironment and exert selective pressures on such cells, thereby shaping their phenotype. Conversely, microenvironmental components regulate gene expression in tumor cells thereby directing the tumor into one or several possible molecular evolution pathways, some of which may lead to metastasis. This review summarizes six instances in which the tumor liaises with different components of its microenvironment. These liaisons result, in most cases, in enhanced tumor progression. In these cases (responses of tumor and nontumor cells to microenvironmental stress, the interaction of the tumor with fibroblasts, endothelial cells and macrophages, the formation of the metastatic niche, and the interaction of the tumor with immunoglobulins) the tumor, directly or indirectly, alters the phenotype of its interaction partners thereby enlisting them to promote its progression. Does the tumor need all these pathways to form metastasis? Is there a hierarchy of interactions with respect to impact on tumor progression? These questions remain open. They may be answered by approaches employed in the analysis of hypercomplex systems. # 2008 Elsevier Inc.

I. THE TUMOR MICROENVIRONMENT The tumor tissue is composed of two compartments intimately associated with each other. The first compartment constitutes the malignant cells. The second is the tumor microenvironment. This compartment is composed of resident fibroblasts, endothelial cells, and other nonmalignant cells; of infiltrating fibroblasts, lymphocytes, or macrophages; and of numerous Advances in CANCER RESEARCH Copyright 2008, Elsevier Inc. All rights reserved.

203

0065-230X/08 $35.00 DOI: 10.1016/S0065-230X(08)00007-9

204

Isaac P. Witz

molecules such as those of the extracellular matrix, growth factors, cytokines, chemokines, antibodies, proteases, other types of enzymes, and various metabolites. All these molecules may be released from the tumor cells or the nonmalignant cells. Products released from necrotic cells (mainly tumor cells) are also present. Artificially administered molecules such as anticancer drugs may also localize in the microenvironment. The tumor microenvironment is perceived to function as an active “educational/inductive/selection” venue in which the tumor is directed into one or several possible molecular evolution pathways by microenvironmental factors. Some of these pathways may lead to metastasis. The interaction between microenvironmental components and tumor cells is, however, bidirectional: Tumor cells and their products are capable of regulating gene expression in nontumor cells residing in or infiltrating into the microenvironment thereby shaping their phenotype. It is now widely accepted that interactions of cancer cells with components of their microenvironment are crucial determinants in the decision whether cancer cells will progress towards metastasis or whether they will stay dormant or disappear altogether. There are scores of tumor–microenvironment interactions that play anti‐ or promalignancy roles. These include interactions that lead to or drive cell proliferation or death: angiogenesis, motility, chemotaxis, invasion, protective immunity, inflammation, and metastasis, to name a few. Although additional interactions probably await discovery and the significance of other interactions has still to be elucidated, it is safe to tentatively conclude that many, if not most, of these interactions constitute dangerous liaisons: They enhance tumor progression and promote metastasis. This review focuses on the late stages of cancer i.e. on progression towards metastasis. In these stages which abnormalities in cancer genes do not play major roles (Sager, 1997; Vogelstein and Kinzler, 2004) and in which some (or most) of the surveillance mechanisms that resist tumor formation (Klein et al., 2007) are not functioning anymore. We will briefly review several cases in which the tumor alters, conditions, or educates cells and molecules in its microenvironment making them its accomplices in progression towards metastasis. Tumor–microenvironment interactions involving adhesion molecules, cytokines, chemokines and other growth factors, proteases and other degradative enzymes, components of the innate and adaptive immunity as well as other interactions situated in the forefront of tumor biology research such as angiogenesis, invasion, inflammation, epithelial to mesenchymal transition, etc. will not be dealt with because they are extensively reviewed. However, several of the cells and molecules involved in these interactions will, none the less, be mentioned in the specific context of the cases reviewed below.

Tumor–Microenvironment Interactions

205

II. STRESS RESPONSES A study on genomic and phenotypic diversity under environmental stress reported that the south‐facing slopes of canyons north of the equator receive higher solar radiation than the do north‐facing slopes. The former slopes are therefore subjected to more stressful insults such as high solar radiation, heat, and drought than are the latter ones (Nevo, 2001). Since the components of the terrain were similar in both slopes and since the distance between the bottoms of the slopes was not more than 100 m, microclimate remained the major interslope divergent factor. A significant biodiversity was observed when the two slopes were compared. In general, the south‐facing slope manifested a considerably higher degree of heterogeneity in numerous genetic and other parameters of most organisms studied than do the north‐facing slope. For example, mutation rates in the fungus Sordaria fimicola were 3‐fold higher in the south‐facing slope than in the north‐facing slope. These and other results showing that genetic and epigenetic diversity is promoted by stress support the well‐accepted paradigm that stress functions as an evolutionary driving force. In conformity with these results, the studies summarized in this section indicate that stress promotes tumor heterogeneity and as such functions as an important determinant of tumor progression. It has been long recognized that tumors are heterogeneous with respect to essentially every measured parameter, including cell morphology and size, gene expression patterns, antigenicity, drug resistance, and metastatic potential (Carter, 1978; Fidler, 1978; Fidler and Hart, 1982; Langley and Fidler, 2007; Macaluso et al., 2003; Miller, 1982; Pathak, 1990; Poste and Greig, 1982; Woodruff, 1983). Highly heterogeneous tumors contain, most probably, a higher number of potentially metastatic cell variants than less heterogeneous tumors. In view of the above, one may hypothesize that tumor heterogeneity drives tumor progression and that heterogeneity amplifying mechanisms will promote metastasis. The molecular pathways underlying the heterogeneity of tumors are very complex. Genetic, epigenetic, and microenvironmental components are involved. Acquired genetic instability of tumor cells (Nowell, 1986, 1989) was, and still is, the accepted and dominating dogma. Indeed tumor cells harbor quite many genomic changes when compared with the corresponding normal cells (Jackson and Loeb, 1998, 2001). Alterations in epigenetic signaling, including DNA methylation and histone modifications, are also a frequent feature of tumor cells (Baylin and Ohm, 2006).

206

Isaac P. Witz

The magnitude of genetic instability increases in parallel with tumor progression. For example, the spontaneous mutation rate in metastatic cell populations is higher than in nonmetastatic ones (Bailly et al., 1993; Cifone and Fidler, 1981; Cillo et al., 1987; Hill, 1990). Several findings indicate that stress induces tumor heterogeneity and as such would be a major factor that promotes tumor progression toward metastasis (Rofstad, 2000; Subarsky and Hill, 2003; Weber and Ashkar, 2000; Xie and Huang, 2003). The microenvironment of solid tumors in cancer patients and in experimental animals is universally characterized by a state of hypoxia, low extracellular pH, and high glycolysis. These factors being functionally connected via hypoxia (Acker and Plate, 2002; Brahimi‐Horn and Pouyssegur, 2006) have an enormous potential to create a stressful environment which promotes tumor progression (Acker and Plate, 2002; Brahimi‐ Horn and Pouyssegur, 2006; De Milito and Fais, 2005; Gatenby and Gillies, 2004; Harris, 2002; Pouyssegur et al., 2006; Rofstad, 2000; Semenza, 2000, 2003; Subarsky and Hill, 2003). Indeed these studies and numerous others indicate that hypoxia plays a critical and fundamental role in metastasis formation. Tumors exhibiting extensive hypoxia tend to be more metastatic than corresponding oxygenized tumors (Chan and Giaccia, 2007; Sullivan and Graham, 2007). Hypoxia drives tumor progression by influencing both the tumor cells as well as the tumor microenvironment. It increases genomic instability and heterogeneity in tumor cells, selects resistant tumor variants, and induces dedifferentiation (Axelson et al., 2005; Helczynska et al., 2003; Weber and Ashkar, 2000; Xie and Huang, 2003). One of the major effects of hypoxia on the microenvironment is the stimulation of angiogenesis (Lewis et al., 2007; Liao and Johnson, 2007). Hypoxia alters the expression of several genes, mainly that of hypoxia inducible factor (HIF). HIF is a heterodimeric transcription factor, composed of constitutively expressed HIF‐1 and HIF‐1 or HIF‐2 subunits. Control of HIF‐1 function occurs primarily through posttranslational oxygen‐ dependent enzymatic hydroxylation of the  subunit. (Chan and Giaccia, 2007; Sullivan and Graham, 2007). Under hypoxia, the  subunit is stabilized, translocates to the nucleus, and dimerizes with HIF‐1 . The stable dimer stimulates the transcription of a large variety of key target genes controlling cell survival, anaerobic metabolism, migration, and angiogenesis, to name just a few (Brahimi‐Horn and Pouyssegur, 2006; Pouyssegur et al., 2006; Semenza, 2000, 2003). Under physiological circumstances the coordinated function of these genes enables cells exposed to hypoxia to cope with this stress (Avivi et al., 2007; Liu and Simon, 2004). In cancer, these genes would confer upon tumor cells various metastasis‐promoting characteristics (Denko et al., 2003).

Tumor–Microenvironment Interactions

207

By surviving and propagating under hypoxia, resistant tumor cells further aggravate the state of tumor hypoxia. This in turn increases genomic instability and further stabilizes and activates HIF. This vicious cycle drives tumor progression. While hypoxia and HIF proteins mediate progression‐promoting effects on tumor cells and on the microenvironment, they may also mediate progression‐ restraining functions by the induction of genes that cause antitumor effects, for example, cell death (Acker et al., 2005; Greijer and van der Wall, 2004). These results necessitate a thorough and careful evaluation of whether HIF proteins can be used as targets for antitumor drugs. Microenvironmental stress also induces or upregulates the expression in cancer cells of heat shock proteins (HSP) (Calderwood et al., 2006). The physiological function of these molecular chaperons is to help damaged proteins to refold to their native conformation. In cancer cells, HSP proteins (e.g., HSP90) maintain the stability and function of several signaling oncoproteins such as mutated p53 or ErbB2 that promote the growth or survival of these cells (Tsutsumi and Neckers, 2007). In addition to their chaperon activity, some HSP such as HSP27 and HSP70 also confer upon the tumor cells heightened resistance against apoptosis mediated by anticancer drugs and immune effectors. This selective pressure promotes progression towards metastasis (Soti and Csermely, 1998). HSP also participate in the proteasome‐mediated degradation of proteins under stress conditions (Garrido et al., 2006). These and other functions of HSP may contribute to cancer progression. There is a linkage between HIF and HSP especially HSP90. HIF‐1 is a client protein of HSP90, it depends on HSP90 for its stability, and the secretion of HSP90 is stimulated by HIF‐1 (Isaacs et al., 2002). Furthermore, the HSP90 inhibitor geldanamycin inhibited HIF‐1a accumulation in human tumor cells (Koga et al., 2007). Taken together the studies on hypoxia and other stress‐mediated effects on tumor progression support the paradigm that stress drives this process. It is therefore logical to hypothesize that targeting molecules, such as HIF or HSP, that confer upon cancer cells the ability to adapt successfully to microenvironmental stress may be a promising cancer therapy modality (Tsutsumi and Neckers, 2007).

III. INTERACTIONS WITH FIBROBLASTS Fibroblasts produce most of the structural proteins of the extracellular matrix, which serves as a source and reservoir of bioactive molecules such as growth factors, proteases, and their inhibitors. Fibroblasts respond to

208

Isaac P. Witz

different physiological and pathological stimuli that may alter their structural and functional phenotype (Sappino et al., 1990). The concept that fibroblasts play an important role in tumor progression was initially introduced by pathologists who reported that desmoplasia, the proliferation of fibroblasts accompanied by an accumulation of extracellular matrix, is a typical feature in many solid tumors (Micke and Ostman, 2004). A rather large body of experimental results confirm that fibroblasts play pivotal promalignancy roles in tumorigenesis, tumor progression, and metastasis (Bhowmick et al., 2004; Cunha et al., 2003; De Wever and Mareel, 2003; Elenbaas and Weinberg, 2001; Kalluri and Zeisberg, 2006; Mareel and Madani, 2006; Micke and Ostman, 2004; Orimo and Weinberg, 2006; Ruiter et al., 2002; Tlsty and Hein, 2001; van den Hooff, 1988). Fibroblasts may generate signals that induce oncogenomic changes involved in cancer initiation (Kuperwasser et al., 2004; Schor et al., 1986) and in driving proliferation, angiogenesis, and motility of epithelial cells, and in suppressing epithelial cell death (Tlsty, 2001; Tlsty and Hein, 2001). Cancer‐associated fibroblasts (CAFs) can be defined as myofibroblasts as they express ‐smooth‐muscle actin, vimentin, and smooth muscle myosin (van den Hooff, 1988). CAFs are activated fibroblasts (Cukierman, 2004; De Wever and Mareel, 2003), biologically different from fibroblasts present in benign microenvironments in several important aspects (Orimo et al., 2005; Tlsty and Hein, 2001; van den Hooff, 1988). CAFs are not passive bystanders in tumorigenesis and metastasis, but contribute actively to these processes. This was shown by experiments demonstrating that the growth of prostate, breast, and colorectal tumor cells in immune‐deficient mice could be significantly accelerated by mixing the inoculated tumor cells with fibroblasts (Chung, 1991; Dimanche‐Boitrel et al., 1994; Noel et al., 1993; Olumi et al., 1999; Picard et al., 1986). Soluble factors from such fibroblasts were active in promoting the growth of prostate cancer cells in nude mice (Chung, 1991). The ability of the CAFs to promote the growth of admixed cultured breast carcinoma cells in nude mice was significantly higher than that of normal mammary fibroblasts derived from the same patients (Orimo et al., 2005). CAF‐derived CXCL12 (SDF‐1) mediated tumor enhancement through the CXCR4 receptor expressed by the carcinoma cells. The CXCL12–CXCR4 axis also supported angiogenesis by recruiting endothelial progenitor cells into the carcinomas. Interestingly, the myofibroblastic phenotype and the ability to enhance tumor growth in vivo were stably maintained in the CAFs even in the absence of contact between them and tumor cells (Orimo et al., 2005). Several additional mechanisms operate in CAF‐mediated enhancement of tumor progression. These include synthesis of support matrices for the growing tumor, production of promalignancy growth factors, and promotion of angiogenesis (Smalley et al., 2005); stimulation of inflammatory

Tumor–Microenvironment Interactions

209

pathways and of secretion of extracellular matrix‐remodeling proteases by the CXCL‐12–CXCR‐4 axis and by an interaction between MMP‐1 and PAR1 (Radisky and Radisky, 2007); secretion of hepatocyte growth factor which stimulates the invasion of cancer cells (Matsumoto and Nakamura, 2006); immune suppression by CAF‐derived cytokines and chemokines (Silzle et al., 2004); promotion of matrix, perineural, and muscular invasion and transendothelial migration (De Wever and Mareel, 2003). Concomitant with tumor progression and with the molecular evolution of the tumor, the fibroblasts in the tumor microenvironment undergo an evolutionary pathway of their own (Beacham and Cukierman, 2005; Cukierman, 2004; Lee and Herlyn, 2007; Orimo and Weinberg, 2006). As the tumor progresses, the infiltration of fibroblasts into the tumor microenvironment increases, and the fibroblasts (which should be now labeled as CAFs) acquire a phenotype of activated cells and exert several promalignancy functions as detailed earlier (Beacham and Cukierman, 2005). Three alternative models were proposed for the evolution of the fibroblasts present within the microenvironment of carcinomas (Orimo and Weinberg, 2006). 1. A small population of fibroblasts that have undergone genetic alterations are selected to become CAFs. 2. Recruitment of normal stroma fibroblasts into the tumor followed by evolution to CAFs without acquiring any genetic alterations. 3. Recruitment of specialized circulating progenitor fibroblasts and evolution of these to CAFs. Which of these evolutionary scenarios, if any, does indeed operate remains unknown. From the studies summarized earlier and from several others it is clear that the tumor is heavily involved in the recruitment of myofibroblasts into the microenvironment, in the activation of these cells, and in the induction of promalignancy functions in them. The signals employed by the tumor in order to mediate these activities are similar to those involved in wound healing (Micke and Ostman, 2004) and are probably different from one type of cancer to another. Whether or not tumor‐specific signaling cascades participate in the recruitment and activation of fibroblasts remains to be seen. Transforming growth factor beta (TGF‐ ) is considered to be most important in the recruitment and generation of tumor‐enhancing CAFs (De Wever et al., 2004; Forsberg et al., 1993; Micke and Ostman, 2004; Shao et al., 2000). TGF‐ has also been postulated to regulate fibroblast‐to‐myofibroblast trans differentiation (Gabbiani, 2003; Untergasser et al., 2005). Platelet‐derived growth factor (PDGF) also plays an important role in the generation and activation of promalignancy CAFs (Forsberg et al., 1993;

210

Isaac P. Witz

Micke and Ostman, 2004; Shao et al., 2000). This cytokine, secreted by many types of cancer cells, induces the proliferation of fibroblasts. Fibroblast growth factor 2 (also known as basic fibroblast growth factor) is involved in tissue fibrosis and probably also in fibroblast proliferation (Strutz et al., 2000). As such it could play a role in the formation of desmoplasia in tumors. Vascular endothelial growth factor (VEGF) released by tumor cells may increase vascular permeability thereby causing the leaking of plasma proteins such as fibrin. This protein, in turn, attracts an influx of fibroblasts (and other cells) into the tumor microenvironment (Brown et al., 1999; Dvorak et al., 1991). Mouse colon tumor cells were found to secrete fibronectin, which induces migration and activation of host fibroblasts (Morimoto and Irimura, 2001). A recent study indicated that human colon epithelial cells release galectin‐3. This molecule activated NF‐kB in colonic lamina propria fibroblasts and induced the secretion of IL8 from these cells (Lippert et al., 2007). Pleiotrophin is a cytokine secreted from many breast cancers. Using a mouse mammary carcinoma model it was shown that this protein activates fibroblasts and increases the secretion of ECM proteins such protocollagens and elastin from these cells. Pleiotrophin promotes the progression of these tumors and induces a phenotype similar to that of scirrhous carcinoma, one of the most aggressive breast cancers in humans (Chang et al., 2007). The studies summarized earlier, as well as additional ones, clearly indicate that cancer‐associated fibroblasts play significant roles in tumorigenesis and especially in tumor progression. The information at our disposal regarding the tumor–CAF crosstalk is however much more extensive with regard to the influences of CAFs on the tumor cells than the information on tumor‐ mediated influences on the fibroblasts. Since signals delivered by the tumor to fibroblasts enlist them to function as supporters of its progression, it is imperative to obtain comprehensive information on the signaling cascades involved in this function.

IV. TUMOR–ENDOTHELIUM INTERACTIONS The interactions between tumor cells and endothelium are major and pivotal driving forces of tumor progression and metastasis. Angiogenesis is the most conspicuous tumor–endothelium interaction and as such is extensively studied and reviewed (Adams and Alitalo, 2007; Alitalo and Carmeliet, 2002; Augustin, 2003; Ferrara, 2002; Folkman, 1974, 2002, 2007; Hlatky et al., 2002; Jain, 2002; Jain et al., 2007; Jung et al., 2002; Kerbel and Folkman, 2002; Liao and Johnson, 2007; Rafii et al., 2002;

Tumor–Microenvironment Interactions

211

Shaked and Kerbel, 2007; Shchors and Evan, 2007; Vlodavsky et al., 2002). Other critical tumor–endothelial interactions, possibly not less important, are invasion into the blood or lymph node vessels (intravasation) (Bockhorn et al., 2007; Condeelis and Pollard, 2006; Dua et al., 2005; Weidner, 2002; Wong and Hynes, 2006) and exit of the cancer cells from the circulation otherwise known as transendothelial migration or extravasation (Brandt et al., 2005, 2007; Deryugina and Quigley, 2006; Dvorak and Feng, 2001; Groom et al., 1999; Gupta and Massague, 2004; Heyder et al., 2006; Houle and Huot, 2006; Kannagi, 1997; Kannagi, 2001; Laferriere et al., 2002; Thurin and Kieber‐Emmons, 2002; Witz, 2008a). It has been long recognized that endothelial cells in tumors express a different phenotype from that of endothelial cells in the corresponding normal tissue (Arap et al., 1998; Sasaki et al., 1991; St Croix et al., 2000). Using a modified serial analysis of gene expression assay, St. Croix and colleagues compared endothelial cells isolated from colorectal cancer and from normal colorectal mucosa. This analysis led to the identification of tumor endothelial markers (TEMs) which were upregulated in the endothelium of colorectal cancer and of other types of tumors (St Croix et al., 2000). Several TEMs were abundantly expressed in blood vessels of developing embryos of mice and humans (Carson‐Walter et al., 2001; Seaman et al., 2007). Certain TEMs were identified as cell surface proteins and others are involved in the formation of and interactions with extracellular matrix, and in its remodeling (Arap et al., 1998; Carson‐Walter et al., 2001; Nanda et al., 2004; Seaman et al., 2007; Tomkowicz et al., 2007; Vallon and Essler, 2006). Epigenetic modifications mediated by DNA methyltransferase (Hellebrekers et al., 2006) and histone deacetylase (Kim et al., 2001) are involved in the regulation of endothelial gene expression and in tumor angiogenesis. A recent study suggested that clusterin, fibrillin 1, and quiescin Q6, functioning as angiogenesis‐suppressing genes, become silenced in endothelial cells exposed in vitro to conditions mimicking tumor‐induced angiogenesis (Hellebrekers et al., 2007). The results of this study led to the hypothesis that tumor endothelium may not be subject to physiologic constrains that normally limit angiogenesis. The findings that endothelial cells in tumors are different in several aspects from endothelial cells in the corresponding normal tissues were confirmed by a large group of investigators (Baluk et al., 2005; Bielenberg et al., 2006; Buckanovich et al., 2007; Castronovo et al., 2006; Hida et al., 2004; Hida and Klagsbrun, 2005; Kumagai et al., 2002; Lu et al., 2007; Parsons‐ Wingerter et al., 2005; Peters et al., 2005; Rybak et al., 2007; Teicher, 2007; van Beijnum et al., 2006; Zhang et al., 2005). The obvious conclusion of the studies cited above is that the tumor, directly or indirectly, embosses its foot imprints on the endothelial cells thereby shaping their phenotype.

212

Isaac P. Witz

What are the signaling pathways that execute this effect? Molecules known to regulate normal physiological angiogenesis such as VEGF, bFGF, angiopoietins, and their receptors are certainly involved. Specific tumor microenvironment‐derived molecules may be involved in driving the phenotypic alterations of tumor endothelial cells (Carmeliet et al., 2001; Hellebrekers et al., 2007; Michiels et al., 2000; Rehman and Wang, 2006; Seaman et al., 2007). It is not unlikely that additional tumor‐specific endothelium regulators will be identified. Do the tumor‐mediated alterations of endothelial cells serve to promote tumor progression? This question remains open for the time being. The fact that endothelial cells in tumors and those in the corresponding normal tissue are phenotypically distinct, provides new challenges for effective cancer‐associated antivascular therapies (Arap et al., 1998; Bazan‐Peregrino et al., 2007; Duan et al., 2007; Jung et al., 2002; Nanda and St Croix, 2004).

V. TUMOR–MACROPHAGE INTERACTIONS Tumor‐associated macrophages (TAMs) are a major constituent of the leukocyte infiltrate in solid tumors (Allavena et al., 2007; Mantovani et al., 2004; Pollard, 2004). Tumor‐derived CCL2 (MCP‐1), frequently expressed in and secreted from various types of cancer, is probably the main chemoattractant responsible for monocyte infiltration into the tumor microenvironment (Allavena et al., 2007; Ben‐Baruch, 2006; Bottazzi et al., 1983; Gazzaniga et al., 2007; Mantovani et al., 2004; Neumark et al., 2003; Ueno et al., 2000). Tumor‐ derived VEGF and macrophage‐colony stimulating factor (M‐CSF) also contribute to macrophage recruitment into the tumor microenvironment (Barleon et al., 1996; Nowicki et al., 1996). The infiltration of the monocytes into the tumor microenvironment is followed by their differentiation to mature macrophages (Allavena et al., 2007). Responding to outside‐in signals, macrophages exhibit versatility and plasticity. Certain stimuli induce a polarized differentiation of macrophages to either M1 or M2 phenotypes (Gordon, 2003; Mantovani et al., 2002; Talmadge et al., 2007). M1 macrophages produce TNF; IL‐1 , IL‐12, IL‐23; and the IFN‐ ‐ inducible chemokines CXCL9, CXCL10, and CXCL11. These cytokines are produced in response to microbial infections, LPS, IFN‐ , or microbial‐ induced cytokines. M1 macrophages are therefore effectors mediating immune responses in conjunction with T1 lymphocytes. M2 macrophages typically produce IL‐10, IL‐1 receptor antagonist (IL‐1ra), and TGF‐ . M2 cells are capable of downregulating inflammatory responses

Tumor–Microenvironment Interactions

213

and adaptive Th1 immunity, they scavenge debris, promote angiogenesis, and induce tissue remodeling (Mantovani et al., 2002; Talmadge et al., 2007). Many, if not most components of the tumor microenvironment may, under certain circumstances exert anti malignancy activities whereas under different circumstances, they exert pro malignancy effects (Witz, 2008b). This general finding holds true also for TAM (Brigati et al., 2002; Dirkx et al., 2006; Lamagna et al., 2006; Mantovani et al., 1992; Talmadge et al., 2007). However, according to the contemporary literature, the promalignancy effects of TAM overshadow the antimalignancy functions of these cells (Allavena et al., 2007; Condeelis and Pollard, 2006; Gazzaniga et al., 2007; Lin and Pollard, 2004; Mantovani et al., 2006; Pak and Fidler, 1991; Pawelek et al., 2006; Pollard, 2004; Whitworth et al., 1990). TAMs express a M2 macrophage phenotype. This is in line with the promalignancy activities of these cells, as summarized earlier. TAMs are both a source as well as a target for cytokines and chemokines present in the tumor microenvironment. These TAM‐derived molecules are among those which regulate tumor proliferation, progression, and invasion and the interaction with immune components. The numerous promalignancy activities of TAM include: production of growth and survival factors for tumor cells such as EGF, IL‐6, and CXCL8 (Mantovani et al., 2002; Pollard, 2004; Sica et al., 2006); production of various angiogenic factors e.g. VEGF, PDGF, and IL‐8 (Bingle et al., 2006; Dirkx et al., 2006; Knowles and Harris, 2007; Lamagna et al., 2006; Mantovani et al., 2002; Talmadge et al., 2007); degradation of extracellular matrix and tissue remodeling by the release of several proteases (Coussens et al., 2000; Egeblad and Werb, 2002); production of immunosuppressive mediators and recruitment into the tumor microenvironment of regulatory T cells (Kim et al., 2006; Knowles and Harris, 2007; Kryczek et al., 2006; Talmadge et al., 2007). The studies summarized above document another instance in which the tumor recruits, conditions, and educates microenvironmental components, macrophages in this case, to facilitate its progression.

VI. THE PREMETASTATIC NICHE The molecular and cellular mechanisms underlying selective organ tropism of different cancer types and site‐specific metastasis are relatively unexplored. A groundbreaking study (Kaplan et al., 2005), employing transplantable syngeneic mouse tumors that form organ‐specific metastasis, focused on the earliest steps in site‐specific metastasis. It was demonstrated that in tumor bearers a VEGF‐mediated crosstalk between tumor cells and VEGFR1‐expressing bone‐marrow‐derived hematopoietic progenitor cells

214

Isaac P. Witz

takes place. This crosstalk leads to the recruitment of the bone marrow cells to organs that serve as specific destination targets for metastasizing cells originating from these tumors. The recruited bone marrow cells apparently prepare a permissive nest for the incoming metastatic cells which appear in the nest some time later. The factors that activate the premetastatic site are present in conditioned media of tumor cells and in the plasma of tumor‐ bearing mice (Kaplan et al., 2005). The tumor seems to be the primary, possibly the only, mediator that directs the influx of the bone marrow cells to the location of the future metastatic site. By using a similar tumor system to that used by Kaplan et al., it was shown that tumor‐derived VEGF, TGF , or TNF induced the expression of the inflammatory chemoattractants S100A8 and S100A9 in the future metastatic site (Hiratsuka et al., 2006). These proteins seemingly attracted bone‐marrow‐derived hematopoietic progenitor cells to this site. How do tumor‐derived signals select a certain organ to serve as a future organ‐specific metastatic site? This is still an open question. It was suggested that cancer cells released into the blood from nonmetastatic primary tumors arrive at the future metastatic site but are unable to undergo metastasis. Such cells may, however, modify the microenvironment of the future metastatic site and make it hospitable for colonization by subsequent waves of released tumor cells (Bidard et al., 2007). The tumor‐derived factors that activate the premetastatic site induce also the secretion of fibronectin from resident fibroblasts. This protein facilitates the anchoring, at the site, of incoming cells such as bone‐marrow‐derived hematopoietic progenitor cells or tumor cells, via a fibronectin–integrin interaction (Kaplan et al., 2005). Many more tumor‐specific molecular and cellular interactions are probably involved in the establishment of the premetastatic niche. The targeting of these could slow or halt metastasis.

VII. TUMOR–IMMUNOGLOBULIN INTERACTIONS Analyzing antitumor immune responses occurring at the tumor microenvironment, we noticed that several tumor types propagating in mice, rats, and humans are coated with immunoglobulins (Ig). The predominant Ig isotype in the coat was IgG. The IgG molecules coating various nonlymphoid mouse tumor cells including polyoma virus induced tumors could be classified into three types based on their functional activities. The first type were IgG molecules in immune complexes bound to Fc receptors expressed by the tumor cells or by infiltrating inflammatory host cells (Braslawsky et al., 1976a,b,c; Gergely and Sarmay, 1994). The second type of IgG molecules were antibodies mediating complement‐dependent cytotoxicity

Tumor–Microenvironment Interactions

215

of tumor cells (Braslawsky et al., 1976a; Moav and Witz, 1978; Ran et al., 1976). These antibodies were found to be directed against serologically detectable polyoma virus induced tumor antigens (Klein et al., 1979; Witz et al., 1976). The third type of tumor‐associated IgG molecules were cross‐ reactive antibodies that mediated complement‐dependent cytotoxicity of mouse lymphocytes (Ran et al., 1978, 1980). IgG eluted from carcinogen‐induced tumors enhanced the in vivo growth of corresponding tumors (Ran and Witz, 1972). Possible mechanisms by which the eluted IgG enhanced tumor growth include masking of tumor epitopes by the coating IgG molecules (Dorval et al., 1976a,b), immunosuppression by cytotoxic antilymphocyte antibodies present in this coat (Ran et al., 1978, 1980) or the delivery of tumor‐enhancing signals to tumor cells through Fc receptors expressed by these cells (Eshel et al., 2002; Ran and Witz, 1985; Ran et al., 1991, 1992, 1988; Witz and Ran, 1992a,b; Zusman et al., 1996a,b). Adaptive B cell responses were recently assigned a pivotal role in the initiation of promalignancy chronic inflammation (de Visser et al., 2005, 2006). Transgenic mice expressing the HPV16 early‐region genes, including the E6/E7 oncogenes, under the control of the human keratin‐14 promoter/ enhancer (HPV16 mice) develop premalignant hyperplasia followed by dysplasia and squamous cell carcinoma (Coussens et al., 1996). This progression was accompanied by chronic inflammation and angiogenic vasculature. The premalignant skin in HPV16 mice was chronically infiltrated by inflammatory cells, predominantly mast cells and granulocytes, and the premalignant cervix was infiltrated by macrophages. T‐ and B‐cell deficiency reduced the inflammatory infiltrate into the premalignant lesions of HPV16 mice, the blood vasculature remained quiescent, there was no hyperproliferation of keratinocytes and the overall incidence of invasive carcinomas was decreased. The adoptive transfer of B cells from untreated HPV16 mice into T‐ and B cell‐deficient HPV16 mice restored the infiltration of the inflammatory cells. Observations that in vivo propagating cancer cells may be coated with Igs (Witz, 1977) were confirmed by showing that IgG was deposited on premalignant skin lesions of HPV16 mice (de Visser et al., 2005, 2006). Passive transfer of serum from untreated HPV16 mice into T‐ and B‐cell‐deficient HPV16 mice had a similar effect to that of adoptively transferred B lymphocytes in restoring the infiltration of the inflammatory cells and in recapitulating inflammation‐associated keratinocyte carcinogenesis (de Visser et al., 2005, 2006). It would be interesting to find out if Ig molecules eluted from the premalignant skin lesions of HPV16 mice are also capable of recapitulating the inflammation‐associated keratinocyte carcinogenesis in these mice. A positive finding would support the notion that antibodies capable of interacting in vivo with tumor cells or with premalignant cells may enhance tumor growth (Ran and Witz, 1972).

216

Isaac P. Witz

Taken together, these findings offer an additional and novel mechanism by which B lymphocytes and especially their Ig products promote tumorigenesis: The Fc portion of IgG molecules bound to premalignant cells or to tumor cells may activate inflammatory cells by engaging their Fc receptors. This activation may trigger a full blown promalignancy inflammatory response as described earlier. In this connection, it is interesting to note that the engagement of Fc RII on monocytes and the costimulation of these cells with IL‐1 resulted in differentiation to a subtype of M2 macrophages (Sironi et al., 2006). M2 macrophages constitute most of the tumor‐associated macrophage population and are among the major promalignancy inflammatory leukocytes (Sica et al., 2006) (see also Tumor–Macrophage Interaction section). Sautes‐Fridman and coworkers reported that melanoma cells from humans may express Fc receptors. A metastatic melanoma cell line expressed the low affinity Fc RIIa, which, under normal physiological conditions, is expressed by immunocytes (Cassard et al., 2000). Tyrosine phosphorylation of several proteins was induced in the melanoma cells by cross‐linking of Fc RIIa by anti‐Fc RIIa antibodies. Among the phosphorylated proteins was the receptor itself. The cross‐linking also caused internalization of the receptor and its migration to class II‐containing compartments. Does the Fc RIIa play a role in the progression of melanoma? It is not unlikely that immune complexes composed of melanoma antigens bound to the corresponding antibodies (Herlyn and Koprowski, 1988) can be bound by the melanoma‐expressed Fc RIIa. In view of the fact that part of the internalized Fc RIIa localized in class II‐ containing compartments of the melanoma cells, the authors raised the possibility that the Fc RIIa may be involved in the presentation of exogenous antigenic peptides by melanoma cells. In the absence of costimulatory molecules on the tumor cells, this presentation could lead to immune anergy. On the other hand and in the presence of costimulatory molecules, it could also lead to a protective antimelanoma immune reactivity (Cassard et al., 2000). In a subsequent study, Cassard et al. (2002) reported that melanin‐ and S100‐expressing metastatic human melanoma cells in fresh tumor biopsies as well as cultured melanoma cells expressed inhibitory Fc RIIB1. The expression of this receptor was associated with an inhibition of melanoma development in nude mice. The authors suggested that therapeutic antimelanoma antibodies may inhibit tumor growth by two nonmutually exclusive mechanisms. The first is by induction of complement‐dependent cytotoxicity or by the activation of ADCC‐mediating killer lymphocytes. The second mechanism postulates that antimelanoma antibodies might directly suppress tumor development by engaging inhibitory Fc receptors on the melanoma cells.

Tumor–Microenvironment Interactions

217

Tumor cells may interact with Igs either by expressing antigenic epitopes that bind antitumor antibodies via their antigen‐binding site or by expressing Fc receptors that bind such antibodies via their Fc portion. In these interactions the tumor cells play a “passive role.” However, cells in the tumor microenvironment may also actively participate in these interactions by interfering with antitumor functions of cyototoxic antibodies and cyototoxic lymphocytes. Proteolytic enzymes derived from such cells were shown to degrade antitumor antibodies and to generate fragments devoid of cytotoxic activity, but being capable of binding to membrane antigens. These antibody fragments protected the tumor cells from intact cytotoxic antibodies and from cytotoxic lymphocytes (Dauphinee et al., 1974; Keisari and Witz, 1973, 1975, 1978). These experiments support the notion that the interaction of tumor cells with microenvironmental components having tumor‐restraining capacities may transform such components and induce them to perform promalignancy functions.

VIII. EXAMINING THE BIG PICTURE There are No Facts, Only Interpretations—Friedrich Nietzsche Cancer research has and still does undergo an interesting transition: from an exclusive emphasis on the tumor cell as the villain responsible for the A and the Z of tumorigenesis and tumor progression, to an increasing focus on the tumor microenvironment as playing the pivotal role especially with regard to progression towards metastasis. In some contemporary studies there is a strong tendency to “put the blame” for tumor progression primarily on nontumor components of the microenvironment. However, these components can be blamed only for serving as accessories in tumor progression. The blame should be shared between them and the tumor being the original perpetuator. The molecular composition of the tumor microenvironment is established jointly by tumor cells as well as by resident and infiltrating nontumor cells. Moreover, tumor cells and nontumor cells in the same microenvironment regulate and shape each other’s phenotype (Joyce, 2005; Witz, 2008b; Witz and Levy‐Nissenbaum, 2006). Tumor‐induced stress, the tumor‐mediated recruitment of fibroblasts and macrophages into the microenvironment and their subsequent activation, the tumor‐mediated enlistment of hematopoietic progenitor cells to create a premetastatic niche, the phenotypic alterations of endothelial cells by the tumor, and the consequences of its interactions with B cells and their immunoglobulin products represent six cases of tumor–microenvironment interactions. They illustrate that the tumor indeed takes part in shaping the

218

Isaac P. Witz

phenotype of nontumor cells in the microenvironment and that it can manipulate these cells and harness them to support its progression. Is any one of these interactions sufficient for metastasis formation or do tumor cells need all (or a subgroup) of them in order to progress? Are we able to identify which of these six interactions (and of many others not reviewed here) plays the most important role in tumor progression and should be thus, therapeutically targeted? Do these interactions integrate through intertwined signaling cascades, and through shared molecules such as NF‐kB, VEGF, TGF , TNF to a single interaction network? The answers to these questions are obviously of enormous importance in the design of future cancer therapy drugs. However, the immense multitude of candidate microenvironmental factors, the extreme complexity of the signaling cascades operating in the microenvironment, the interactive crosstalk between these cascades (Hornberg et al., 2006) and the inter‐ and intratumor heterogeneity pose a formidable challenge for those attempting to provide answers to these questions. Attempting to comprehend the big picture of tumor progression, we should realize that a single molecule or a single signaling pathway is just one component, out of many comprising an immense network. This realization should lead to the abandonment of reductionism and to the employment of approaches used in the analysis of hyper complex systems (Aderem, 2005; Ao et al., 2007; Xiong et al., 2005).

ACKNOWLEDGEMENTS I am indebted to the former and present members of my team for their devotion, talent, creativity, and diligence. The following foundations and individuals are thanked for generous grant support: The Dr. Miriam and Sheldon G. Adelson Medical Research Foundation (Needham, MA, USA), The Jacqueline Seroussi Memorial Foundation for Cancer Research (Basel, Switzerland), The Ela Kodesz Institute for Research on Cancer Development and Prevention, Tel Aviv University; The Fainbarg Family Fund (Orange County, CA, USA); Bonnie and Steven Stern (New York, NY, USA), The Fred August and Adele Wolpers Charitable Fund (Clifton, NJ, USA), Natan Blutinger (West Orange, NJ, USA), Arnold and Ruth Feuerstein (Orange County, CA, USA), The Pikovsky Fund (Jerusalem, Israel); and James J. Leibman and Rita S. Leibman Endowment Fund for Cancer Research (New York, NY, USA).

REFERENCES Acker, T., Diez‐Juan, A., Aragones, J., Tjwa, M., Brusselmans, K., Moons, L., Fukumura, D., Moreno‐Murciano, M. P., Herbert, J. M., Burger, A., Riedel, J., Elvert, G., et al. (2005). Genetic evidence for a tumor suppressor role of HIF‐2alpha. Cancer Cell 8, 131–141.

Tumor–Microenvironment Interactions

219

Acker, T., and Plate, K. H. (2002). A role for hypoxia and hypoxia‐inducible transcription factors in tumor physiology. J. Mol. Med. 80, 562–575. Adams, R. H., and Alitalo, K. (2007). Molecular regulation of angiogenesis and lymphangiogenesis. Nat. Rev. Mol. Cell Biol. 8, 464–478. Aderem, A. (2005). Systems biology: Its practice and challenges. Cell 121, 511–513. Alitalo, K., and Carmeliet, P. (2002). Molecular mechanisms of lymphangiogenesis in health and disease. Cancer Cell 1, 219–227. Allavena, P., Sica, A., Solinas, G., Porta, C., and Mantovani, A. (2007). The inflammatory micro‐environment in tumor progression: The role of tumor‐associated macrophages. Crit. Rev. Oncol. Hematol. Epub ahead of print PMID 17913510. Ao, P., Galas, D., Hood, L., and Zhu, X. (2007). Cancer as robust intrinsic state of endogenous molecular‐cellular network shaped by evolution. Med Hypotheses. 70, 678–684. Arap, W., Pasqualini, R., and Ruoslahti, E. (1998). Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model. Science 279, 377–380. Augustin, H. G. (2003). Translating angiogenesis research into the clinic: The challenges ahead. Br. J. Radiol. 76(Spec No 1), S3–S10. Avivi, A., Ashur‐Fabian, O., Joel, A., Trakhtenbrot, L., Adamsky, K., Goldstein, I., Amariglio, N., Rechavi, G., and Nevo, E. (2007). P53 in blind subterranean mole rats–loss‐ of‐function versus gain‐of‐function activities on newly cloned Spalax target genes. Oncogene 26, 2507–2512. Axelson, H., Fredlund, E., Ovenberger, M., Landberg, G., and Pahlman, S. (2005). Hypoxia‐ induced dedifferentiation of tumor cells—A mechanism behind heterogeneity and aggressiveness of solid tumors. Semin. Cell Dev. Biol. 16, 554–563. Bailly, M., Bertrand, S., and Dore, J. F. (1993). Increased spontaneous mutation rates and prevalence of karyotype abnormalities in highly metastatic human melanoma cell lines. Melanoma Res. 3, 51–61. Baluk, P., Hashizume, H., and McDonald, D. M. (2005). Cellular abnormalities of blood vessels as targets in cancer. Curr. Opin. Genet. Dev. 15, 102–111. Barleon, B., Sozzani, S., Zhou, D., Weich, H. A., Mantovani, A., and Marme, D. (1996). Migration of human monocytes in response to vascular endothelial growth factor (VEGF) is mediated via the VEGF receptor flt‐1. Blood 87, 3336–3343. Barthel, S. R., Gavino, J. D., Descheny, L., and Dimitroff, C. J. (2007). Targeting selectins and selectin ligands in inflammation and cancer. Expert Opin. Ther. Targets 11, 1473–1491. Baylin, S. B., and Ohm, J. E. (2006). Epigenetic gene silencing in cancer—A mechanism for early oncogenic pathway addiction? Nat. Rev. Cancer 6, 107–116. Bazan‐Peregrino, M., Seymour, L. W., and Harris, A. L. (2007). Gene therapy targeting to tumor endothelium. Cancer Gene Ther. 14, 117–127. Beacham, D. A., and Cukierman, E. (2005). Stromagenesis: The changing face of fibroblastic microenvironments during tumor progression. Semin. Cancer Biol. 15, 329–341. Ben‐Baruch, A. (2006). The multifaceted roles of chemokines in malignancy. Cancer Metastasis Rev. 25, 357–371. Bhowmick, N. A., Neilson, E. G., and Moses, H. L. (2004). Stromal fibroblasts in cancer initiation and progression. Nature 432, 332–337. Bidard, F. C., Pierga, J. Y., Vincent‐Salomon, A., and Poupon, M. F. (2007). A “class action” against the microenvironment: Do cancer cells cooperate in metastasis? Cancer Metastasis Rev. 27, 5–10. Bielenberg, D. R., Pettaway, C. A., Takashima, S., and Klagsbrun, M. (2006). Neuropilins in neoplasms: Expression, regulation, and function. Exp. Cell Res. 312, 584–593. Bingle, L., Lewis, C. E., Corke, K. P., Reed, M. W., and Brown, N. J. (2006). Macrophages promote angiogenesis in human breast tumour spheroids in vivo. Br. J. Cancer 94, 101–107.

220

Isaac P. Witz

Bockhorn, M., Jain, R. K., and Munn, L. L. (2007). Active versus passive mechanisms in metastasis: Do cancer cells crawl into vessels, or are they pushed? Lancet Oncol. 8, 444–448. Bottazzi, B., Polentarutti, N., Acero, R., Balsari, A., Boraschi, D., Ghezzi, P., Salmona, M., and Mantovani, A. (1983). Regulation of the macrophage content of neoplasms by chemoattractants. Science 220, 210–212. Brahimi‐Horn, C., and Pouyssegur, J. (2006). The role of the hypoxia‐inducible factor in tumor metabolism growth and invasion. Bull. Cancer 93, E73–E80. Brandt, B., Heyder, C., Gloria‐Maercker, E., Hatzmann, W., Rotger, A., Kemming, D., Zanker, K. S., Entschladen, F., and Dittmar, T. (2005). 3D‐extravasation model—Selection of highly motile and metastatic cancer cells. Semin. Cancer Biol. 15, 387–395. Braslawsky, G., Ran, M., and Witz, I. P. (1976a). Tumor bound immunoglobulins: The relationship between the in vivo coating of tumor cells by potentially cytotoxic anti‐tumor antibodies, and the expression of immune complex receptors. Int. J. Cancer 18, 116–121. Braslawsky, G. R., Serban, D., and Witz, I. P. (1976b). Receptors for immune complexes on cells within a polyoma virus‐induced murine sarcoma. Eur. J. Immunol. 6, 579–583. Braslawsky, G. R., Yaackubowicz, M., Frensdorff, A., and Witz, I. P. (1976c). Receptors for immune complexes on cells within a non‐lymphoid murine tumor. J. Immunol. 116, 1571–1578. Brigati, C., Noonan, D. M., Albini, A., and Benelli, R. (2002). Tumors and inflammatory infiltrates: Friends or foes?. Clin. Exp. Metastasis 19, 247–258. Brown, L. F., Guidi, A. J., Schnitt, S. J., Van De Water, L., Iruela‐Arispe, M. L., Yeo, T. K., Tognazzi, K., and Dvorak, H. F. (1999). Vascular stroma formation in carcinoma in situ, invasive carcinoma, and metastatic carcinoma of the breast. Clin. Cancer Res. 5, 1041–1056. Buckanovich, R. J., Sasaroli, D., O’Brien‐Jenkins, A., Botbyl, J., Hammond, R., Katsaros, D., Sandaltzopoulos, R., Liotta, L. A., Gimotty, P. A., and Coukos, G. (2007). Tumor vascular proteins as biomarkers in ovarian cancer. J. Clin. Oncol. 25, 852–861. Calderwood, S. K., Khaleque, M. A., Sawyer, D. B., and Ciocca, D. R. (2006). Heat shock proteins in cancer: Chaperones of tumorigenesis. Trends Biochem. Sci. 31, 164–172. Carmeliet, P., Moons, L., Luttun, A., Vincenti, V., Compernolle, V., De Mol, M., Wu, Y., Bono, F., Devy, L., Beck, H., Scholz, D., Acker, T., et al. (2001). Synergism between vascular endothelial growth factor and placental growth factor contributes to angiogenesis and plasma extravasation in pathological conditions. Nat. Med. 7, 575–583. Carson‐Walter, E. B., Watkins, D. N., Nanda, A., Vogelstein, B., Kinzler, K. W., and St Croix, B. (2001). Cell surface tumor endothelial markers are conserved in mice and humans. Cancer Res. 61, 6649–6655. Carter, R. L. (1978). Metastatic potential of malignant tumours. Invest. Cell Pathol. 1, 275–286. Cassard, L., Cohen‐Solal, J. F., Galinha, A., Sastre‐Garau, X., Mathiot, C., Galon, J., Dorval, T., Bernheim, A., Fridman, W. H., and Sautes‐Fridman, C. (2002). Modulation of tumor growth by inhibitory Fc(gamma) receptor expressed by human melanoma cells. J. Clin. Invest. 110, 1549–1557. Cassard, L., Dragon‐Durey, M. A., Ralli, A., Tartour, E., Salamero, J., Fridman, W. H., and Sautes‐Fridman, C. (2000). Expression of low‐affinity Fc gamma receptor by a human metastatic melanoma line. Immunol. Lett. 75, 1–8. Castronovo, V., Waltregny, D., Kischel, P., Roesli, C., Elia, G., Rybak, J. N., and Neri, D. (2006). A chemical proteomics approach for the identification of accessible antigens expressed in human kidney cancer. Mol. Cell Proteomics 5, 2083–2091. Chan, D. A., and Giaccia, A. J. (2007). Hypoxia, gene expression, and metastasis. Cancer Metastasis Rev. 26, 333–339. Chang, Y., Zuka, M., Perez-Pinera, P., Astudillo, A., Mortimer, J., Berenson, J. R., and Deuel, T. F. (2007). Secretion of pleiotrophin stimulates breast cancer progression through remodeling of the tumor microenvironment. Proc. Natl. Acad. Sci. USA 104, 10888–10893.

Tumor–Microenvironment Interactions

221

Chung, L. W. (1991). Fibroblasts are critical determinants in prostatic cancer growth and dissemination. Cancer Metastasis Rev. 10, 263–274. Cifone, M. A., and Fidler, I. J. (1981). Increasing metastatic potential is associated with increasing genetic instability of clones isolated from murine neoplasms. Proc. Natl. Acad. Sci. USA 78, 6949–6952. Cillo, C., Dick, J. E., Ling, V., and Hill, R. P. (1987). Generation of drug‐resistant variants in metastatic B16 mouse melanoma cell lines. Cancer Res. 47, 2604–2608. Condeelis, J., and Pollard, J. W. (2006). Macrophages: Obligate partners for tumor cell migration, invasion, and metastasis. Cell 124, 263–266. Coussens, L. M., Hanahan, D., and Arbeit, J. M. (1996). Genetic predisposition and parameters of malignant progression in K14‐HPV16 transgenic mice. Am. J. Pathol. 149, 1899–1917. Coussens, L. M., Tinkle, C. L., Hanahan, D., and Werb, Z. (2000). MMP‐9 supplied by bone marrow‐derived cells contributes to skin carcinogenesis. Cell 103, 481–490. Cukierman, E. (2004). A visual‐quantitative analysis of fibroblastic stromagenesis in breast cancer progression. J. Mammary Gland Biol. Neoplasia 9, 311–324. Cunha, G. R., Hayward, S. W., Wang, Y. Z., and Ricke, W. A. (2003). Role of the stromal microenvironment in carcinogenesis of the prostate. Int. J. Cancer 107, 1–10. Dauphinee, M. J., Talal, N., and Witz, I. P. (1974). Generation of non‐complement‐fixing, blocking factors by lysosomal extract treatment of cytotoxic anti‐tumor antibodies. J. Immunol. 113, 948–953. De Milito, A., and Fais, S. (2005). Tumor acidity, chemoresistance and proton pump inhibitors. Future Oncol. 1, 779–786. de Visser, K. E., Eichten, A., and Coussens, L. M. (2006). Paradoxical roles of the immune system during cancer development. Nat. Rev. Cancer 6, 24–37. de Visser, K. E., Korets, L. V., and Coussens, L. M. (2005). De novo carcinogenesis promoted by chronic inflammation is B lymphocyte dependent. Cancer Cell 7, 411–423. De Wever, O., and Mareel, M. (2003). Role of tissue stroma in cancer cell invasion. J. Pathol. 200, 429–447. De Wever, O., Westbroek, W., Verloes, A., Bloemen, N., Bracke, M., Gespach, C., Bruyneel, E., and Mareel, M. (2004). Critical role of N‐cadherin in myofibroblast invasion and migration in vitro stimulated by colon‐cancer‐cell‐derived TGF‐beta or wounding. J. Cell Sci. 117, 4691–4703. Denko, N. C., Fontana, L. A., Hudson, K. M., Sutphin, P. D., Raychaudhuri, S., Altman, R., and Giaccia, A. J. (2003). Investigating hypoxic tumor physiology through gene expression patterns. Oncogene 22, 5907–5914. Deryugina, E. I., and Quigley, J. P. (2006). Matrix metalloproteinases and tumor metastasis. Cancer Metastasis Rev. 25, 9–34. Dimanche‐Boitrel, M. T., Vakaet, L., Jr., Pujuguet, P., Chauffert, B., Martin, M. S., Hammann, A., Van Roy, F., Mareel, M., and Martin, F. (1994). In vivo and in vitro invasiveness of a rat colon‐cancer cell line maintaining E‐cadherin expression: An enhancing role of tumor‐associated myofibroblasts. Int. J. Cancer 56, 512–521. Dirkx, A. E., Oude Egbrink, M. G., Wagstaff, J., and Griffioen, A. W. (2006). Monocyte/ macrophage infiltration in tumors: Modulators of angiogenesis. J. Leukoc. Biol. 80, 1183–1196. Dorval, G., Wilz, I. P., Klein, E., and Wigzell, H. (1976a). Tumor‐bound immunoglobulins: An in vivo phenomenon of masked specificity. J. Natl. Cancer Inst. 56, 523–527. Dorval, G., Witz, I. P., Klein, E., and Wigzell, H. (1976b). Tumor‐bound immunoglobulins. I. Further analysis of the characteristics of binding of immunoglobulins to in vivo‐grown tumor cells. Int. J. Cancer 17, 109–119.

222

Isaac P. Witz

Dua, R. S., Gui, G. P., and Isacke, C. M. (2005). Endothelial adhesion molecules in breast cancer invasion into the vascular and lymphatic systems. Eur. J. Surg. Oncol. 31, 824–832. Duan, H. F., Hu, X. W., Chen, J. L., Gao, L. H., Xi, Y. Y., Lu, Y., Li, J. F., Zhao, S. R., Xu, J. J., Chen, H. P., Chen, W., and Wu, C. T. (2007). Antitumor activities of TEM8‐Fc: An engineered antibody‐like molecule targeting tumor endothelial marker 8. J. Natl. Cancer Inst. 99, 1551–1555. Dvorak, A. M., and Feng, D. (2001). The vesiculo‐vacuolar organelle (VVO). A new endothelial cell permeability organelle. J. Histochem. Cytochem. 49, 419–432. Dvorak, H. F., Sioussat, T. M., Brown, L. F., Berse, B., Nagy, J. A., Sotrel, A., Manseau, E. J., Van de Water, L., and Senger, D. R. (1991). Distribution of vascular permeability factor (vascular endothelial growth factor) in tumors: Concentration in tumor blood vessels. J. Exp. Med. 174, 1275–1278. Egeblad, M., and Werb, Z. (2002). New functions for the matrix metalloproteinases in cancer progression. Nat. Rev. Cancer 2, 161–174. Elenbaas, B., and Weinberg, R. A. (2001). Heterotypic signaling between epithelial tumor cells and fibroblasts in carcinoma formation. Exp. Cell Res. 264, 169–184. Eshel, R., Neumark, E., Sagi‐Assif, O., and Witz, I. P. (2002). Receptors involved in microenvironment‐driven molecular evolution of cancer cells. Semin. Cancer Biol. 12, 139–147. Ferrara, N. (2002). VEGF and the quest for tumour angiogenesis factors. Nat. Rev. Cancer 2, 795–803. Fidler, I. J. (1978). Tumor heterogeneity and the biology of cancer invasion and metastasis. Cancer Res. 38, 2651–2660. Fidler, I. J., and Hart, I. R. (1982). Biological diversity in metastatic neoplasms: Origins and implications. Science 217, 998–1003. Folkman, J. (1974). Tumor angiogenesis. Adv. Cancer Res. 19, 331–358. Folkman, J. (2002). Role of angiogenesis in tumor growth and metastasis. Semin. Oncol. 29, 15–18. Folkman, J. (2007). Angiogenesis: An organizing principle for drug discovery? Nat. Rev. Drug Discov. 6, 273–286. Forsberg, K., Valyi‐Nagy, I., Heldin, C. H., Herlyn, M., and Westermark, B. (1993). Platelet‐ derived growth factor (PDGF) in oncogenesis: Development of a vascular connective tissue stroma in xenotransplanted human melanoma producing PDGF‐BB. Proc. Natl. Acad. Sci. USA 90, 393–397. Gabbiani, G. (2003). The myofibroblast in wound healing and fibrocontractive diseases. J. Pathol. 200, 500–503. Garrido, C., Brunet, M., Didelot, C., Zermati, Y., Schmitt, E., and Kroemer, G. (2006). Heat shock proteins 27 and 70: Anti‐apoptotic proteins with tumorigenic properties. Cell Cycle 5, 2592–2601. Gatenby, R. A., and Gillies, R. J. (2004). Why do cancers have high aerobic glycolysis? Nat. Rev. Cancer 4, 891–899. Gazzaniga, S., Bravo, A. I., Guglielmotti, A., van Rooijen, N., Maschi, F., Vecchi, A., Mantovani, A., Mordoh, J., and Wainstok, R. (2007). Targeting tumor‐associated macrophages and inhibition of MCP‐1 reduce angiogenesis and tumor growth in a human melanoma xenograft. J. Invest. Dermatol. 127, 2031–2041. Gergely, J., and Sarmay, G. (1994). Fc gamma receptors in malignancies: Friends or enemies? Adv. Cancer Res. 64, 211–245. Gordon, S. (2003). Alternative activation of macrophages. Nat. Rev. Immunol. 3, 23–35. Greijer, A. E., and van der Wall, E. (2004). The role of hypoxia inducible factor 1 (HIF‐1) in hypoxia induced apoptosis. J. Clin. Pathol. 57, 1009–1014.

Tumor–Microenvironment Interactions

223

Groom, A. C., MacDonald, I. C., Schmidt, E. E., Morris, V. L., and Chambers, A. F. (1999). Tumour metastasis to the liver, and the roles of proteinases and adhesion molecules: New concepts from in vivo videomicroscopy. Can. J. Gastroenterol. 13, 733–743. Gupta, G. P., and Massague, J. (2004). Platelets and metastasis revisited: A novel fatty link. J. Clin. Invest. 114, 1691–1693. Harris, A. L. (2002). Hypoxia—A key regulatory factor in tumour growth. Nat. Rev. Cancer. 2, 38–47. Helczynska, K., Kronblad, A., Jogi, A., Nilsson, E., Beckman, S., Landberg, G., and Pahlman, S. (2003). Hypoxia promotes a dedifferentiated phenotype in ductal breast carcinoma in situ. Cancer Res. 63, 1441–1444. Hellebrekers, D. M., Jair, K. W., Vire, E., Eguchi, S., Hoebers, N. T., Fraga, M. F., Esteller, M., Fuks, F., Baylin, S. B., van Engeland, M., and Griffioen, A. W. (2006). Angiostatic activity of DNA methyltransferase inhibitors. Mol. Cancer Ther. 5, 467–475. Hellebrekers, D. M., Melotte, V., Vire, E., Langenkamp, E., Molema, G., Fuks, F., Herman, J. G., Van Criekinge, W., Griffioen, A. W., and van Engeland, M. (2007). Identification of epigenetically silenced genes in tumor endothelial cells. Cancer Res. 67, 4138–4148. Herlyn, M., and Koprowski, H. (1988). Melanoma antigens: Immunological and biological characterization and clinical significance. Annu. Rev. Immunol. 6, 283–308. Heyder, C., Gloria‐Maercker, E., Hatzmann, W., Zaenker, K. S., and Dittmar, T. (2006). Visualization of tumor cell extravasation. Contrib. Microbiol. 13, 200–208. Hida, K., Hida, Y., Amin, D. N., Flint, A. F., Panigrahy, D., Morton, C. C., and Klagsbrun, M. (2004). Tumor‐associated endothelial cells with cytogenetic abnormalities. Cancer Res. 64, 8249–8255. Hida, K., and Klagsbrun, M. (2005). A new perspective on tumor endothelial cells: Unexpected chromosome and centrosome abnormalities. Cancer Res. 65, 2507–2510. Hill, R. P. (1990). Tumor progression: Potential role of unstable genomic changes. Cancer Metastasis Rev. 9, 137–147. Hiratsuka, S., Watanabe, A., Aburatani, H., and Maru, Y. (2006). Tumour‐mediated upregulation of chemoattractants and recruitment of myeloid cells predetermines lung metastasis. Nat. Cell Biol. 8, 1369–1375. Hlatky, L., Hahnfeldt, P., and Folkman, J. (2002). Clinical application of antiangiogenic therapy: Microvessel density, what it does and doesn’t tell us. J. Natl. Cancer Inst. 94, 883–893. Hornberg, J. J., Bruggeman, F. J., Westerhoff, H. V., and Lankelma, J. (2006). Cancer: A systems biology disease. Biosystems 83, 81–90. Houle, F., and Huot, J. (2006). Dysregulation of the endothelial cellular response to oxidative stress in cancer. Mol. Carcinog. 45, 362–367. Isaacs, J. S., Jung, Y. J., Mimnaugh, E. G., Martinez, A., Cuttitta, F., and Neckers, L. M. (2002). Hsp90 regulates a von Hippel Lindau‐independent hypoxia‐inducible factor‐1 alpha‐ degradative pathway. J. Biol. Chem. 277, 29936–29944. Jackson, A. L., and Loeb, L. A. (1998). The mutation rate and cancer. Genetics 148, 1483–1490. Jackson, A. L., and Loeb, L. A. (2001). The contribution of endogenous sources of DNA damage to the multiple mutations in cancer. Mutat. Res. 477, 7–21. Jain, R. K. (2002). Tumor angiogenesis and accessibility: Role of vascular endothelial growth factor. Semin. Oncol. 29, 3–9. Jain, R. K., di Tomaso, E., Duda, D. G., Loeffler, J. S., Sorensen, A. G., and Batchelor, T. T. (2007). Angiogenesis in brain tumours. Nat. Rev. Neurosci. 8, 610–622. Joyce, J. A. (2005). Therapeutic targeting of the tumor microenvironment. Cancer Cell 7, 513–520.

224

Isaac P. Witz

Jung, Y. D., Ahmad, S. A., Liu, W., Reinmuth, N., Parikh, A., Stoeltzing, O., Fan, F., and Ellis, L. M. (2002). The role of the microenvironment and intercellular cross‐talk in tumor angiogenesis. Semin. Cancer Biol. 12, 105–112. Kalluri, R., and Zeisberg, M. (2006). Fibroblasts in cancer. Nat. Rev. Cancer 6, 392–401. Kannagi, R. (1997). Carbohydrate‐mediated cell adhesion involved in hematogenous metastasis of cancer. Glycoconj. J. 14, 577–584. Kannagi, R. (2001). Transcriptional regulation of expression of carbohydrate ligands for cell adhesion molecules in the selectin family. Adv. Exp. Med. Biol. 491, 267–278. Kaplan, R. N., Riba, R. D., Zacharoulis, S., Bramley, A. H., Vincent, L., Costa, C., MacDonald, D. D., Jin, D. K., Shido, K., Kerns, S. A., Zhu, Z., Hicklin, D., et al. (2005). VEGFR1‐positive haematopoietic bone marrow progenitors initiate the pre‐metastatic niche. Nature 438, 820–827. Keisari, Y., and Witz, I. P. (1973). Degradation of immunoglobulins by lysosomal enzymes of tumors. I. Demonstration of the phenomenon using mouse tumors. Immunochemistry 10, 565–570. Keisari, Y., and Witz, I. P. (1975). The specific blocking of humoral immune cytolysis mediated by anti‐tumor antibodies degraded by lysosomal enzymes of tumor origin. Eur. J. Immunol. 5, 790–795. Keisari, Y., and Witz, I. P. (1978). Degradation of cell‐bound antibodies by tumor cells from C57BL/6 mice. J. Natl. Cancer Inst. 61, 1135–1141. Kerbel, R., and Folkman, J. (2002). Clinical translation of angiogenesis inhibitors. Nat. Rev. Cancer 2, 727–739. Kim, M. S., Kwon, H. J., Lee, Y. M., Baek, J. H., Jang, J. E., Lee, S. W., Moon, E. J., Kim, H. S., Lee, S. K., Chung, H. Y., Kim, C. W., and Kim, K. W. (2001). Histone deacetylases induce angiogenesis by negative regulation of tumor suppressor genes. Nat. Med. 7, 437–443. Kim, R., Emi, M., Tanabe, K., and Arihiro, K. (2006). Tumor‐driven evolution of immunosuppressive networks during malignant progression. Cancer Res. 66, 5527–5536. Klein, G., Ehlin, B., and Witz, I. (1979). Serological detection of a polyoma‐tumor‐associated membrane antigen. Int. J. Cancer 23, 683–690. Klein, G., Imreh, S., and Zabarovsky, E. R. (2007). Why do we not all die of cancer at an early age? Adv. Cancer Res. 98, 1–16. Knowles, H. J., and Harris, A. L. (2007). Macrophages and the hypoxic tumour microenvironment. Front Biosci. 12, 4298–4314. Koga, F., Tsutsumi, S., and Neckers, L. M. (2007). Low dose geldanamycin inhibits hepatocyte growth factor and hypoxia‐stimulated invasion of cancer cells. Cell Cycle 6, 1393–1402. Kryczek, I., Zou, L., Rodriguez, P., Zhu, G., Wei, S., Mottram, P., Brumlik, M., Cheng, P., Curiel, T., Myers, L., Lackner, A., Alvarez, X., et al. (2006). B7‐H4 expression identifies a novel suppressive macrophage population in human ovarian carcinoma. J. Exp. Med. 203, 871–881. Kumagai, Y., Toi, M., and Inoue, H. (2002). Dynamism of tumour vasculature in the early phase of cancer progression: Outcomes from oesophageal cancer research. Lancet Oncol. 3, 604–610. Kuperwasser, C., Chavarria, T., Wu, M., Magrane, G., Gray, J. W., Carey, L., Richardson, A., and Weinberg, R. A. (2004). Reconstruction of functionally normal and malignant human breast tissues in mice. Proc. Natl. Acad. Sci. USA 101, 4966–4971. Laferriere, J., Houle, F., and Huot, J. (2002). Regulation of the metastatic process by E‐selectin and stress‐activated protein kinase‐2/p38. Ann. N Y Acad. Sci. 973, 562–572. Lamagna, C., Aurrand‐Lions, M., and Imhof, B. A. (2006). Dual role of macrophages in tumor growth and angiogenesis. J. Leukoc. Biol. 80, 705–713. Langley, R. R., and Fidler, I. J. (2007). Tumor cell‐organ microenvironment interactions in the pathogenesis of cancer metastasis. Endocr. Rev. 28, 297–321.

Tumor–Microenvironment Interactions

225

Lee, J. T., and Herlyn, M. (2007). Microenvironmental influences in melanoma progression. J. Cell Biochem. 101, 862–872. Lewis, C. E., De Palma, M., and Naldini, L. (2007). Tie2‐expressing monocytes and tumor angiogenesis: Regulation by hypoxia and angiopoietin‐2. Cancer Res. 67, 8429–8432. Liao, D., and Johnson, R. S. (2007). Hypoxia: A key regulator of angiogenesis in cancer. Cancer Metastasis Rev. 26, 281–290. Lin, E. Y., and Pollard, J. W. (2004). Role of infiltrated leucocytes in tumour growth and spread. Br. J. Cancer 90, 2053–2058. Lippert, E., Falk, W., Bataille, F., Kaehne, T., Naumann, M., Goeke, M., Herfarth, H., Schoelmerich, J., and Rogler, G. (2007). Soluble galectin‐3 is a strong, colonic epithelial‐ cell‐derived, lamina propria fibroblast‐stimulating factor. Gut 56, 43–51. Liu, L., and Simon, M. C. (2004). Regulation of transcription and translation by hypoxia. Cancer Biol. Ther. 3, 492–497. Lu, C., Bonome, T., Li, Y., Kamat, A. A., Han, L. Y., Schmandt, R., Coleman, R. L., Gershenson, D. M., Jaffe, R. B., Birrer, M. J., and Sood, A. K. (2007). Gene alterations identified by expression profiling in tumor‐associated endothelial cells from invasive ovarian carcinoma. Cancer Res. 67, 1757–1768. Macaluso, M., Paggi, M. G., and Giordano, A. (2003). Genetic and epigenetic alterations as hallmarks of the intricate road to cancer. Oncogene 22, 6472–6478. Mantovani, A., Allavena, P., Sozzani, S., Vecchi, A., Locati, M., and Sica, A. (2004). Chemokines in the recruitment and shaping of the leukocyte infiltrate of tumors. Semin. Cancer Biol. 14, 155–160. Mantovani, A., Bottazzi, B., Colotta, F., Sozzani, S., and Ruco, L. (1992). The origin and function of tumor‐associated macrophages. Immunol. Today 13, 265–270. Mantovani, A., Schioppa, T., Porta, C., Allavena, P., and Sica, A. (2006). Role of tumor‐ associated macrophages in tumor progression and invasion. Cancer Metastasis Rev. 25, 315–322. Mantovani, A., Sozzani, S., Locati, M., Allavena, P., and Sica, A. (2002). Macrophage polarization: Tumor‐associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol. 23, 549–555. Mareel, M., and Madani, I. (2006). Tumour‐associated host cells participating at invasion and metastasis: Targets for therapy? Acta Chir. Belg. 106, 635–640. Matsumoto, K., and Nakamura, T. (2006). Hepatocyte growth factor and the Met system as a mediator of tumor‐stromal interactions. Int. J. Cancer 119, 477–483. Michiels, C., Arnould, T., and Remacle, J. (2000). Endothelial cell responses to hypoxia: Initiation of a cascade of cellular interactions. Biochim. Biophys. Acta 1497, 1–10. Micke, P., and Ostman, A. (2004). Tumour‐stroma interaction: Cancer‐associated fibroblasts as novel targets in anti‐cancer therapy? Lung Cancer 45(Suppl 2), S163–S175. Miller, F. R. (1982). Intratumor immunologic heterogeneity. Cancer Metastasis Rev. 1, 319–334. Moav, N., and Witz, I. P. (1978). Characterization of immunoglobulins eluted from murine tumor cells: Binding patterns of cytotoxic anti‐tumor IgG. J. Immunol. Methods 22, 51–62. Morimoto, M., and Irimura, T. (2001). Fibroblast migratory factor derived from mouse colon carcinoma cells: Potential roles of fibronectin in tumor stroma formation. J. Cell Biochem. 80, 635–646. Nanda, A., Buckhaults, P., Seaman, S., Agrawal, N., Boutin, P., Shankara, S., Nacht, M., Teicher, B., Stampfl, J., Singh, S., Vogelstein, B., Kinzler, K. W., et al. (2004). Identification of a binding partner for the endothelial cell surface proteins TEM7 and TEM7R. Cancer Res. 64, 8507–8511. Nanda, A., and St Croix, B. (2004). Tumor endothelial markers: New targets for cancer therapy. Curr. Opin. Oncol. 16, 44–49.

226

Isaac P. Witz

Neumark, E., Sagi‐Assif, O., Shalmon, B., Ben‐Baruch, A., and Witz, I. P. (2003). Progression of mouse mammary tumors: MCP‐1‐TNFalpha cross‐regulatory pathway and clonal expression of promalignancy and antimalignancy factors. Int. J. Cancer 106, 879–886. Nevo, E. (2001). Evolution of genome‐phenome diversity under environmental stress. Proc. Natl. Acad. Sci. USA 98, 6233–6240. Noel, A., De Pauw‐Gillet, M. C., Purnell, G., Nusgens, B., Lapiere, C. M., and Foidart, J. M. (1993). Enhancement of tumorigenicity of human breast adenocarcinoma cells in nude mice by matrigel and fibroblasts. Br. J. Cancer 68, 909–915. Nowell, P. C. (1986). Mechanisms of tumor progression. Cancer Res. 46, 2203–2207. Nowell, P. C. (1989). Chromosomal and molecular clues to tumor progression. Semin. Oncol. 16, 116–127. Nowicki, A., Szenajch, J., Ostrowska, G., Wojtowicz, A., Wojtowicz, K., Kruszewski, A. A., Maruszynski, M., Aukerman, S. L., and Wiktor‐Jedrzejczak, W. (1996). Impaired tumor growth in colony‐stimulating factor 1 (CSF‐1)‐deficient, macrophage‐deficient op/op mouse: Evidence for a role of CSF‐1‐dependent macrophages in formation of tumor stroma. Int. J. Cancer 65, 112–119. Olumi, A. F., Grossfeld, G. D., Hayward, S. W., Carroll, P. R., Tlsty, T. D., and Cunha, G. R. (1999). Carcinoma‐associated fibroblasts direct tumor progression of initiated human prostatic epithelium. Cancer Res. 59, 5002–5011. Orimo, A., Gupta, P. B., Sgroi, D. C., Arenzana‐Seisdedos, F., Delaunay, T., Naeem, R., Carey, V. J., Richardson, A. L., and Weinberg, R. A. (2005). Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF‐1/CXCL12 secretion. Cell 121, 335–348. Orimo, A., and Weinberg, R. A. (2006). Stromal fibroblasts in cancer: A novel tumor‐promoting cell type. Cell Cycle 5, 1597–1601. Pak, C. C., and Fidler, I. J. (1991). Molecular mechanisms for activated macrophage recognition of tumor cells. Semin. Cancer Biol. 2, 189–195. Parsons‐Wingerter, P., Kasman, I. M., Norberg, S., Magnussen, A., Zanivan, S., Rissone, A., Baluk, P., Favre, C. J., Jeffry, U., Murray, R., and McDonald, D. M. (2005). Uniform overexpression and rapid accessibility of alpha5beta1 integrin on blood vessels in tumors. Am. J. Pathol. 167, 193–211. Pathak, S. (1990). Cytogenetic abnormalities in cancer: With special emphasis on tumor heterogeneity. Cancer Metastasis Rev. 8, 299–318. Pawelek, J., Chakraborty, A., Lazova, R., Yilmaz, Y., Cooper, D., Brash, D., and Handerson, T. (2006). Co‐opting macrophage traits in cancer progression: A consequence of tumor cell fusion? Contrib. Microbiol. 13, 138–155. Peters, D. G., Kudla, D. M., Deloia, J. A., Chu, T. J., Fairfull, L., Edwards, R. P., and Ferrell, R. E. (2005). Comparative gene expression analysis of ovarian carcinoma and normal ovarian epithelium by serial analysis of gene expression. Cancer Epidemiol. Biomarkers Prev. 14, 1717–1723. Picard, O., Rolland, Y., and Poupon, M. F. (1986). Fibroblast‐dependent tumorigenicity of cells in nude mice: Implication for implantation of metastases. Cancer Res. 46, 3290–3294. Pollard, J. W. (2004). Tumour‐educated macrophages promote tumour progression and metastasis. Nat. Rev. Cancer 4, 71–78. Poste, G., and Greig, R. (1982). On the genesis and regulation of cellular heterogeneity in malignant tumors. Invasion Metastasis 2, 137–176. Pouyssegur, J., Dayan, F., and Mazure, N. M. (2006). Hypoxia signalling in cancer and approaches to enforce tumour regression. Nature 441, 437–443. Radisky, E. S., and Radisky, D. C. (2007). Stromal induction of breast cancer: Inflammation and invasion. Rev. Endocr. Metab. Disord. 8, 279–287.

Tumor–Microenvironment Interactions

227

Rafii, S., Lyden, D., Benezra, R., Hattori, K., and Heissig, B. (2002). Vascular and haematopoietic stem cells: Novel targets for anti‐angiogenesis therapy? Nat. Rev. Cancer 2, 826–835. Ran, M., Katz, B., Kimchi, N., Halachmi, E., Teillaud, J. L., Even, J., Berko‐Flint, Y., Atlas, E., Fridman, W. H., and Witz, I. P. (1991). In vivo acquisition of Fc gamma RII expression on polyoma virus‐transformed cells derived from tumors of long latency. Cancer Res. 51, 612–618. Ran, M., Klein, G., and Witz, I. P. (1976). Tumor‐bound immunoglobulins. Evidence for the in vivo coating of tumor cells by potentially cytotoxic anti‐tumour antibodies. Int. J. Cancer 17, 90–97. Ran, M., Langer, A. B., Eliassi, I., Gohar, O., Gonen, B., Gradsztajn, S., Fridman, W. H., Teillaud, J. L., and Witz, I. P. (1992). Possibilities of interference with the immune system of tumor bearers by non‐lymphoid Fc gamma RII expressing tumor cells. Immunobiology 185, 415–425. Ran, M., Teillaud, J. L., Fridman, W. H., Frenkel, H., Halachmi, E., Katz, B., Gips, M., Shlomo, Y., Barzilay, J., and Witz, I. P. (1988). Increased expression of Fc gamma receptor in cancer patients and tumor bearing mice. Mol. Immunol. 25, 1159–1167. Ran, M., and Witz, I. P. (1972). Tumor‐associated immunoglobulins. Enhancement of syngeneic tumors by IgG2‐containing tumor eluates. Int. J. Cancer 9, 242–247. Ran, M., and Witz, I. P. (1985). FcR derived from without the immune system—A potential escape mechanism for cells propagating in a hostile immunological environment. Contrib. Gynecol. Obstet. 14, 83–89. Ran, M., Yaakubowicz, M., Amitai, O., and Witz, I. P. (1980). Tumor‐localizing lymphocytotoxic antibodies. Contemp. Top. Immunobiol. 10, 191–211. Ran, M., Yaakubowicz, M., and Witz, I. P. (1978). Lymphocytotoxic autoantibodies eluted from in vivo propagating sarcoma cells of mice. J. Natl. Cancer. Inst. 60, 1509–1513. Rehman, A. O., and Wang, C. Y. (2006). Notch signaling in the regulation of tumor angiogenesis. Trends Cell Biol. 16, 293–300. Rofstad, E. K. (2000). Microenvironment‐induced cancer metastasis. Int. J. Radiat. Biol. 76, 589–605. Ruiter, D., Bogenrieder, T., Elder, D., and Herlyn, M. (2002). Melanoma‐stroma interactions: Structural and functional aspects. Lancet Oncol. 3, 35–43. Rybak, J. N., Roesli, C., Kaspar, M., Villa, A., and Neri, D. (2007). The extra‐domain A of fibronectin is a vascular marker of solid tumors and metastases. Cancer Res. 67, 10948–10957. Sager, R. (1997). Expression genetics in cancer: shifting the focus from DNA to RNA. Proc. Natl. Acad. Sci. USA 94, 952–955. Sappino, A. P., Schurch, W., and Gabbiani, G. (1990). Differentiation repertoire of fibroblastic cells: Expression of cytoskeletal proteins as marker of phenotypic modulations. Lab. Invest. 63, 144–161. Sasaki, K., Kiuchi, Y., Sato, Y., and Yamamori, S. (1991). Morphological analysis of neovascularization at early stages of rat splenic autografts in comparison with tumor angiogenesis. Cell Tissue Res. 265, 503–510. Schor, S. L., Haggie, J. A., Durning, P., Howell, A., Smith, L., Sellwood, R. A., and Crowther, D. (1986). Occurrence of a fetal fibroblast phenotype in familial breast cancer. Int. J. Cancer 37, 831–836. Seaman, S., Stevens, J., Yang, M. Y., Logsdon, D., Graff‐Cherry, C., and St Croix, B. (2007). Genes that distinguish physiological and pathological angiogenesis. Cancer Cell 11, 539–554. Semenza, G. L. (2000). Hypoxia, clonal selection, and the role of HIF‐1 in tumor progression. Crit. Rev. Biochem. Mol. Biol. 35, 71–103. Semenza, G. L. (2003). Targeting HIF‐1 for cancer therapy. Nat. Rev. Cancer 3, 721–732. Shaked, Y., and Kerbel, R. S. (2007). Antiangiogenic strategies on defense: On the possibility of blocking rebounds by the tumor vasculature after chemotherapy. Cancer Res. 67, 7055–7058. Shao, Z. M., Nguyen, M., and Barsky, S. H. (2000). Human breast carcinoma desmoplasia is PDGF initiated. Oncogene 19, 4337–4345.

228

Isaac P. Witz

Shchors, K., and Evan, G. (2007). Tumor angiogenesis: Cause or consequence of cancer? Cancer Res. 67, 7059–7061. Sica, A., Schioppa, T., Mantovani, A., and Allavena, P. (2006). Tumour‐associated macrophages are a distinct M2 polarised population promoting tumour progression: Potential targets of anti‐cancer therapy. Eur. J. Cancer 42, 717–727. Silzle, T., Randolph, G. J., Kreutz, M., and Kunz‐Schughart, L. A. (2004). The fibroblast: Sentinel cell and local immune modulator in tumor tissue. Int. J. Cancer 108, 173–180. Sironi, M., Martinez, F. O., D’Ambrosio, D., Gattorno, M., Polentarutti, N., Locati, M., Gregorio, A., Iellem, A., Cassatella, M. A., Van Damme, J., Sozzani, S., Martini, A., Sinigaglia, F., Vecchi, A., et al. (2006). Differential regulation of chemokine production by Fcgamma receptor engagement in human monocytes: Association of CCL1 with a distinct form of M2 monocyte activation (M2b, Type 2). J. Leukoc. Biol. 80, 342–349. Smalley, K. S., Brafford, P. A., and Herlyn, M. (2005). Selective evolutionary pressure from the tissue microenvironment drives tumor progression. Semin. Cancer Biol. 15, 451–459. Soti, C., and Csermely, P. (1998). Molecular chaperones in the etiology and therapy of cancer. Pathol. Oncol. Res. 4, 316–321. St Croix, B., Rago, C., Velculescu, V., Traverso, G., Romans, K. E., Montgomery, E., Lal, A., Riggins, G. J., Lengauer, C., Vogelstein, B., and Kinzler, K. W. (2000). Genes expressed in human tumor endothelium. Science 289, 1197–1202. Strutz, F., Zeisberg, M., Hemmerlein, B., Sattler, B., Hummel, K., Becker, V., and Muller, G. A. (2000). Basic fibroblast growth factor expression is increased in human renal fibrogenesis and may mediate autocrine fibroblast proliferation. Kidney Int. 57, 1521–1538. Subarsky, P., and Hill, R. P. (2003). The hypoxic tumour microenvironment and metastatic progression. Clin. Exp. Metastasis 20, 237–250. Sullivan, R., and Graham, C. H. (2007). Hypoxia‐driven selection of the metastatic phenotype. Cancer Metastasis Rev. 26, 319–331. Talmadge, J. E., Donkor, M., and Scholar, E. (2007). Inflammatory cell infiltration of tumors: Jekyll or Hyde. Cancer Metastasis Rev. 26, 373–400. Teicher, B. A. (2007). Newer vascular targets: Endosialin (review). Int. J. Oncol. 30, 305–312. Thurin, M., and Kieber‐Emmons, T. (2002). SA‐Lea and tumor metastasis: The old prediction and recent findings. Hybrid Hybridomics 21, 111–116. Tlsty, T. D. (2001). Stromal cells can contribute oncogenic signals. Semin. Cancer Biol. 11, 97–104. Tlsty, T. D., and Hein, P. W. (2001). Know thy neighbor: Stromal cells can contribute oncogenic signals. Curr. Opin. Genet. Dev. 11, 54–59. Tomkowicz, B., Rybinski, K., Foley, B., Ebel, W., Kline, B., Routhier, E., Sass, P., Nicolaides, N. C., Grasso, L., and Zhou, Y. (2007). Interaction of endosialin/TEM1 with extracellular matrix proteins mediates cell adhesion and migration. Proc. Natl. Acad. Sci. USA 104, 17965–17970. Tsutsumi, S., and Neckers, L. (2007). Extracellular heat shock protein 90: A role for a molecular chaperone in cell motility and cancer metastasis. Cancer Sci. 98, 1536–1539. Ueno, T., Toi, M., Saji, H., Muta, M., Bando, H., Kuroi, K., Koike, M., Inadera, H., and Matsushima, K. (2000). Significance of macrophage chemoattractant protein‐1 in macrophage recruitment, angiogenesis, and survival in human breast cancer. Clin. Cancer Res. 6, 3282–3289. Untergasser, G., Gander, R., Lilg, C., Lepperdinger, G., Plas, E., and Berger, P. (2005). Profiling molecular targets of TGF‐beta1 in prostate fibroblast‐to‐myofibroblast transdifferentiation. Mech. Ageing Dev. 126, 59–69. Vallon, M., and Essler, M. (2006). Proteolytically processed soluble tumor endothelial marker (TEM) 5 mediates endothelial cell survival during angiogenesis by linking integrin alpha(v) beta3 to glycosaminoglycans. J. Biol. Chem. 281, 34179–34188.

Tumor–Microenvironment Interactions

229

van Beijnum, J. R., Dings, R. P., van der Linden, E., Zwaans, B. M., Ramaekers, F. C., Mayo, K. H., and Griffioen, A. W. (2006). Gene expression of tumor angiogenesis dissected: Specific targeting of colon cancer angiogenic vasculature. Blood 108, 2339–2348. van den Hooff, A. (1988). Stromal involvement in malignant growth. Adv. Cancer Res. 50, 159–196. Vogelstein, B., and Kinzler, K. W. (2004). Cancer genes and the pathways they control. Nat. Med. 10, 789–799. Vlodavsky, I., Goldshmidt, O., Zcharia, E., Atzmon, R., Rangini‐Guatta, Z., Elkin, M., Peretz, T., and Friedmann, Y. (2002). Mammalian heparanase: Involvement in cancer metastasis, angiogenesis and normal development. Semin. Cancer Biol. 12, 121–129. Weber, G. F., and Ashkar, S. (2000). Stress response genes: The genes that make cancer metastasize. J. Mol. Med. 78, 404–408. Weidner, N. (2002). New paradigm for vessel intravasation by tumor cells. Am. J. Pathol. 160, 1937–1939. Whitworth, P. W., Pak, C. C., Esgro, J., Kleinerman, E. S., and Fidler, I. J. (1990). Macrophages and cancer. Cancer Metastasis Rev. 8, 319–351. Witz, I. P. (1977). Tumor‐bound immunoglobulins: In situ expressions of humoral immunity. Adv. Cancer Res. 25, 95–148. Witz, I. P. (2008a). The selectin‐selectin ligand axis in tumor progression. Cancer Metastasis Rev. 27, 19–30. Witz, I. P. (2008b). Yin‐yang activities and vicious cycles in the tumor microenvironment. Cancer Res. 68, 9–13. Witz, I. P., Lee, N., and Klein, G. (1976). Serologically detectable specific and cross‐reactive antigens on the membrane of a polyoma virus‐induced murine tumor. Int. J. Cancer 18, 243–249. Witz, I. P., and Levy‐Nissenbaum, O. (2006). The tumor microenvironment in the post‐PAGET era. Cancer Lett. 242, 1–10. Witz, I. P., and Ran, M. (1992a). FcR may function as a progression factor of nonlymphoid tumors. Immunol Res. 11, 283–295. Witz, I. P., and Ran, M. (1992b). The involvement of Fc receptors in tumor progression. AIDS Res. Hum. Retroviruses 8, 806–811. Wong, S. Y., and Hynes, R. O. (2006). Lymphatic or hematogenous dissemination: How does a metastatic tumor cell decide? Cell Cycle 5, 812–817. Woodruff, M. F. (1983). Cellular heterogeneity in tumours. Br. J. Cancer 47, 589–594. Xie, K., and Huang, S. (2003). Regulation of cancer metastasis by stress pathways. Clin. Exp. Metastasis 20, 31–43. Xiong, M., Feghali‐Bostwick, C. A., Arnett, F. C., and Zhou, X. (2005). A systems biology approach to genetic studies of complex diseases. FEBS Lett. 579, 5325–5332. Zhang, T., Sun, H. C., Xu, Y., Zhang, K. Z., Wang, L., Qin, L. X., Wu, W. Z., Liu, Y. K., Ye, S. L., and Tang, Z. Y. (2005). Overexpression of platelet‐derived growth factor receptor alpha in endothelial cells of hepatocellular carcinoma associated with high metastatic potential. Clin. Cancer Res. 11, 8557–8563. Zusman, T., Gohar, O., Eliassi, H., Avivi, Y., Lisansky, E., Sautes, C., Even, J., Bonnerot, C., Fridman, W. H., Witz, I. P., and Ran, M. (1996a). The murine Fc‐gamma (Fc gamma) receptor type II B1 is a tumorigenicity‐enhancing factor in polyoma‐virus‐transformed 3T3 cells. Int. J. Cancer 65, 221–229. Zusman, T., Lisansky, E., Arons, E., Anavi, R., Bonnerot, C., Sautes, C., Fridman, W. H., Witz, I. P., and Ran, M. (1996b). Contribution of the intracellular domain of murine Fc‐gamma receptor type IIB1 to its tumor‐enhancing potential. Int. J. Cancer 68, 219–227.

Index

A

Adenomatous polyposis coli (Apc), 89 Alpha 1-antichymotrypsin enzyme, 6 Androgen-independent prostate cancer (AICaP), 67 Angiogenesis, 210 inhibitors, mechanisms of, 121–122 VEGF-A role of, 115 Antiangiogenic therapy biomarkers, 125 cancer therapy guidelines, 126 vs. chemotherapy, 122–124 drugs, and side effects, 119–121 future perspectives, 125–127 patient selection and timescale of treatment, 124–125 Anti-integrin antibody, 191 Anti-VEGF therapy and CEPCs, 125 tumors, 118–119 Apoptotic cells phagocytosis, 51 Avastin, 113, 124 side effects of, 119 Axin/GSK-3 complex, 15 Axin 2 protein, 89 Axl, TAM family, 36–39 signaling pathways, 59–61

B

BCAS3 gene, 142 Benign papillomatous growth, 175 Bladder carcinoma, coexpression Ron and EGFR, 18 BMPs. See Bone morphogenetic proteins Bone marrow-derived circulating endothelial progenitor cells (BM-CEPCs), 115 Bone morphogenetic proteins receptor 1A, 94 signaling, in intestine, 93–94

Bos taurus, 2 Breast cancer, Ron expression in, 20

C

CAFs. See Cancer-associated fibroblasts Calcium pyrophosphate, for stimulate division, 173 Cancer and TAM receptors, 62–63 angiogenesis, 65–66 cell survival and tumor growth, 66–67 migration and invasion, 64–65 Cancer-associated fibroblasts, 208 Cancer cells, interactions of, 204 Canis lupus, 2 Carcinoembryonic antigen (CEA), 190 -Catenin target gene activation, model for, 102–104 in tumor and colorectal cancer progression, 104–105 in Wnt pathway, 88 Cbl ubiquitin ligases, 15 Cdc25A protein, in tumor-derived cells, 19 CD30-positive HESCs, 141 Cell-cell adhesion, molecular basis of, 189–191 anti-integrin antibody, role of, 191 3 1 integrin interaction, 190 E-cadherin and CEA level, 190 interactions, 180–181 Cell surface growth factor receptors, 2 Chemotherapeutic drugs, 122 Chick embryo fibroblasts (CEFs), 161 Chronic myelogenous leukemia (CML), 39 Circulating endothelial progenitor cells (CEPCs), 125 Clusterin protein, 211 Colorectal cancer, 102 Colorectal tumorigenesis, steps for, 103 Cowden disease, 95

231

232 CRC. See Colorectal cancer Crypts of Lieberku¨hn, 86 C1-TEN proteins, 61 Cultured human melanocytes, characteristics of, 178–180 chromosome studies, 179 colony-forming efficiency, 180 CXCL-12–CXCR-4 axis, 208–209

D

Disseminated cancer cell (DCC), 160, 184–187, 193–194 DNA methyltransferase, 211 Down syndrome, 146 Drosophila melanogaster, 91

E

E-Cadherin, 190 associated with malignancy, 169 gene, 180 in melanoma cells and keratinocytes, 169–170 EGF. See Epidermal growth factor Embryoid bodies (EBs), 134 Embryonal carcinoma (EC), 137 Embryonic stem (ES) cells, 134 EMT. See Epithelial to mesenchymal transition Endostatin, 114 antiangiogenic activity, 122–123 Endothelium and tumor progression, 210–212 Eph/ephrin protein, 95 signaling pathway, in intestine, 96–97 Epidermal growth factor receptors (EGFRs), 96, 98 in NIH3T3 fibroblasts, 8 and Ron-induced cellular transformation, 17 Epidermoid cysts, 175 Epithelial cells, suppression of transformation in, 165–171 calcium-resistant foci, 167 cell-cell interaction, 166 deepithelialized tracheal technique, 167 Drosophila, for tumor genetics, 170–171 TSG mutant clones, for integrin, 171 E-cadherin and N-cadherin, role in, 169 HANs cell, growth of, 165–166

Index liver vs. subcutaneous connective tissue, 170 malignant epidermal cells growth, normal keratinocytes effect, 168 melanoma antigens, expression of, 169 papilloma development, inhibitory capacity for, 168 p53 mutant keratinocyte, 168–169 in skin grafts, growth inhibition of, 167–168 Epithelial to mesenchymal transition, 9, 101, 105–106 Equus caballus, 2 E-Ras gene and tumorigenicity of HESCs, 142–143 Extravasation. See Transendothelial migration

F

Familial adenomatous polyposis (FAP), 93 Fascin protein, 105 Fc RIIa antibody, 216 Fibrillin 1 protein, 211 Fibroblast growth factor (FGF), 2, 115, 210 Fibroblasts, suppression of transformation in depletion of medium, 164 fluorescent dye transfer inhibition and hypoxanthine, 163 intracellular content of Mg2þ, 165 in null connexin 43 knockout cell and spontaneous transformation, 164 RSV infection in CEF, 161 concentrations of RSV, 162 (GJC) probing of, 163 growth curves and inhibition of proteases, 163 transformed foci suppression, 162 and tumor progression, 207–210 Fibronectin, 210 type III (FNIII), 36–37 Fugu rubripes, 2

G

Galectin-3, 210 Ganciclovir, 147–148 Gap junction communication (GJC), 163 Gas6/Axl complex, crystal structure, 43 Gastrointestinal tumor and Ron expression, 23

233

Index Germ cell tumors (GCTs), 136 GJC in suppression of transformation, 171 cyclic AMP and depletion of medium, 172 glial cells and plasma membrane activity, 172 Glioblastoma tumorigenesis, Axl role of, 68 Glycoproteins, 190 Grb2-associated binding protein-1 (Gab1), 15 Gut, cell types in, 86

H

Ha-Ras-transgenic mice and papillomas formation, 13 Heat shock protein functions, geldanamycins, 26 HSP90 protein, 207 Hedgehog interacting protein, 91 signaling, in intestine, 91–93 Hepatocellular carcinoma (HCC), 16 Hepatocyte growth factor (HGF), 2 Hepatocyte growth factor-like protein (HGFL), 1–3 disulfide-linked heterodimer formation of, 6 liver and kidney specific expression of, 5 Herceptin anti-HER2 monoclonal antibody, 69 Hereditary papillary renal carcinoma (HPRC), 10 Hes-1 protein, 90 HIF. See Hypoxia inducible factor Histone deacetylase, 211 HSP. See Heat shock proteins Human colorectal carcinoma and Ron expression, 26 foreskin keratinocyte library, 2 malignancies mastocytosis, 10 ovarian cancer and Ron expression, 22 protein S and murine Tyro-3, 41–42 umbilical vein endothelial cell (HUVECs), 66 Human embryonic stem cells (HESCs) definition and nature of, 134 induced teratomas in cancer research, 146–147 as clinical hurdle, 147 differentiation of, 149 genetic diseases, modeling for, 146

normal embryogenesis, modeling for, 145–146 teratoma cells ablation, 147–148 nontransformed, tumorigenicity of, 143–145 tumorigenicity and culture adaptation of, 140–141 embryonic carcinoma cells, in vivo differentiation of, 139–140 genes associated with, 142–143 human embryonic development study, 145 Hyperplastic alveolar nodules (HANs), 165–166 Hypoxia inducible factor, 206–207 HIF-1-induced pathway, 115

I

ID-1 and ID-2, monoclonal antibodies and Ron signaling, 25 IL-1 receptor antagonist (IL-1 ra), 212 Immunoglobulin and tumor progression, 214–217 Indian Hedgehog (Ihh), 91 Integrin-laminin binding, 190 Interference RNA (RNAi), 117 International stem cell initiative (ISCI), 142–143 Interstitial fluid pressure (IFP), in tumors, 115 Intestine BMPs signaling in, 93–94 eph/ephrin signaling pathway in, 95–97 epithelial cell types in, 86–87 Hedgehog signaling in, 91–93 K-ras signaling in, 97–99 notch signaling in, 89–91 PTEN in, 94–95 Wnt pathway in, 88–89

J

Jun N-terminal signaling pathway, 171 Juvenile polyposis syndrome (JP), 94

K

Kallikrein, factor XIIa and factor XIa, 6. See also Hepatocyte growth factor-like protein (HGFL); Macrophage stimulating-protein

234 Keratinocytes, 160 K-ras signaling, in intestine, 97–99

L

Laminin 5, in cell-cell adhesion, 190 Liver cancers, Ron expression in, 23–24 progenitor cells and Ron receptor, 9 L1 protein, 104–105 Lung, Ron overexpression in, 13

M

Macaca mulatta, 2 Macrophage and tumor progression, 212–213 Macrophage-colony stimulating factor, 212 Macrophages and dendritic cells, TAM receptor loss from apoptotic cells and, 50–52 and cytokine secretion, 52–53 natural killer cells and platelets, 53 vascular smooth muscle cells (VSMCs), 53–54 Macrophage stimulating-protein, 1–3 Macrophage stimulating 1-receptor, 2 Malignant cells, 203 Mammalian gastrointestinal tract composition of, 85–86 epithelial cell types in, 87 Mammary tumorigenesis and Ron overexpression, 13 MAPK. See Mitogen-activated protein kinase MAPK signal pathway, 14 Math-1 protein, 90 Matriptase, 6. See also Hepatocyte growth factor-like protein (HGFL); Macrophage stimulating-protein Matrix metalloproteases, 102 MCP-1, 212 M-CSF. See Macrophage-colony stimulating factor Melanoma cell adhesion molecule (MCAM), 169 Melanomas, 169 Membrane type serine protease 1 (MT-SP1), 6 Mer signaling pathways, 55–59 Mer, TAM family, 36–39 MerTK for Mer tyrosine kinase, 40

Index Mesenchymal tissue, suppressive effects adhesiveness reduction and genomic changes, 176 transplants, develop into HANs, 175 Metastases, 161 prototype of progression, in malignant melanoma, 176–178 melanocytic dysplasia to RGP and nevi hyperplasia, 177 pigmentation and surface location, 178 types of stages, 176–177 tumor and host microenvironmental factors in, 194 Metastatic dormancy, 161 alternative explanations of B16 melanoma cells, metastasis of, 189 role of growth factor deficiency, 188 soluble growth factors, 188–189 thyroid cells, role in, 187 characteristics of disseminated cells, fate of, 183 human mammary cancer cells, 184 solitary cell survival, kinetics of, 182–183 survival in tissue, of solitary cancer cells, 181–182 Metastatis, formation of, 101–102 Met binding domain (MBD) of Gab1, 15 Met receptor tyrosine kinases, 2–3 hereditary papillary renal carcinoma (HPRC) in, 10 Ron receptor tyrosine kinases and, 4 Mice expressing mutated Ras transgene (v-Ha-Ras; Tg.AC), 12 Milk fat globule-EGF factor 8 protein (MFG-E8), 58 Mitogen-activated protein kinase, 10, 98 MMPs. See Matrix metalloproteases MMTVpromoter (MMTV-pMT), 12 Mouse mammary tumor virus (MMTV), 6–7 Mouse model and Ron protein, 11–12 pancreatic cancer (PdxCre/ LSL-KRASG12D), 21 NIH3T3 fibroblasts and Ron mutants, 10 resident peritoneal macrophages and HGFL effect, 7 MP470 Axl inhibitor, 70 MSP. See Macrophage stimulating-protein MST1R. See Macrophage stimulating 1-receptor

235

Index Multiple endocrine neoplasia type 2B, 10 Murine Ron transcript, 3 Mutated Ras transgene with wild-type Ron signal function (Tg.ACþ//Ron TKþ/þ), 12–13

N

Nanog gene, 142 N-Cadherin, in transformation of melanocytes, 169, 190 Nck2 proteins, 61 Necrotic cells, 204 Neoplasia, contact-related suppression of, 160 Neurogenin-3 protein, 90 Neuropilins (NRPs) and VEGF-A, 118 NIH3T3 mouse fibroblast cells and murine Ron mutants, 10 Nitric oxide, production by macrophages and Ron, 8 Notch signaling, in intestine, 89–91 Nr-CAM protein, 104 Nyk for NCAM-related tyrosine kinase, 40

O

Oct4 gene and human GCTs, 142 Ovary, Mer expression in, 41

P

Pancrea cancer and BxPC-3 monoclonal antibody treatment, 25 Ron expression in, 21 Paneth cells, 86 Papilloma cells, 180 PDGF. See Platelet-derived growth factor Pediatric cancers and antiangiogenic therapy, 121 Pegaptinib, in age-related macular degeneration, 117 Phosphatase and tensin homologue, 94 Phospholipase-C (PLC- ), signal transducers, 15 Phosphotidyl-inositol 3, 4, 5-triphosphate (PIP3), 15 Phosphotyrosine binding domains (PTB), biological activities of Ron, 14–15 Pigment epithelial cell-derived factor (PEDF), 121

PI3-kinases, 98 PI3-K signal pathway, 14–15 Placenta growth factor (PlGF), 121 Plasma membrane activity, role in cell growth, 173–175 adhesiveness and malignancy of living cells, phenotypes of, 175 contact inhibition and tighter adhesions, 173 direct cell-cell membrane interactions and homophilic adhesions, 174 fibroblasts, transformed into sarcoma cells, 173–174 papilloma cells vs. squamous carcinoma cells, 174 population density and growth rate, 173 Platelet-derived growth factor, 98, 115, 209–210 Plexins, transmembrane receptors for semaphorins, 18 Polyoma virus middle T antigen (pMT), 12 Preeclampsia and VEGF-A, 121 Primordial germ cell (PGC), 136 Prostate cancer, Ron expression in, 20 Protein S, 51 PTEN. See Phosphatase and tensin homologue PTEN, in intestine, 94–95

Q

Quiescin Q6 protein, 211

R

Radial growth phase (RGP), 177–178 Ran-binding protein in microtubule organizing center (RanBPM), 61 Ras/ERK signaling pathway, 97 Ras/MAPK pathway, in intestine, 98, 100 Ras proteins, 97 Ras transgene and deficient in Ron signal function (Tg.ACþ//Ron TK/), 12–13 Rat embryo fibroblasts (REFs), 163 Rattus norvegicus, 2 Receptor kinase activity, regulation, 44 Axl signaling pathways, 46 Mer signaling pathways, 45 Receptor tyrosine kinases (RTKs), 2, 36, 118 in intestine, 95–99 Renal tumors and Ron expression, 21–22 Ret receptor, mutations, 10

236 Rex-1, pluripotent marker, 149 Rituximab (anti-CD20), 70 Ron receptor, Met family of cell surface receptor tyrosine kinases, 1 angiogenesis and, 19 chromosomal location and cancer, 7 developmental roles and tumor properties of, 8–9 expression, in bladder cancer, 22 function loss and tumorigenesis, 12–13 genomic instability and cell cycle disruption, 19 human tumors and tumor-derived cell lines expression, 19–24 induced tumorigenesis, 14–16 ligand, structure and function of, 3–7 in macrophages, inflammation and cancer, 7–8 and Met, cross-talk in cancer, 17 Met receptor tyrosine kinases and, 4 mouse models function loss for, 11–12 overexpression in tumors gain of, 13–14 oncogenic potential of, 10–11 overexpression in organ systems, mouse models, 13 receptor cross-talk and activity in tumorigenesis, 16–18 structure and function of, 2–3 as target for cancer therapy, 24–26 transition of, 9–10 Ron tyrosine kinase intracellular signaling domain (Ron TK/), 12 Rous sarcoma virus (RSV), 161 c-ryk, cellular protooncogene, 40

S

Sea receptor, 2. See also Ron receptor, Met family of cell surface receptor tyrosine kinases

-Secretase, 91 Sf9 cells Tyro-3 expression, 47 Shh. See Sonic Hedgehog Small cell carcinoma of lung (SCLC) and Ron expression, 22 Soluble VEGFR-1 (sVEGFR1) and preeclampsia, 121 Sonic Hedgehog, 91 Son of sevenless (SOS), biological activities of Ron, 14–15

Index Sordaria fimicola, 205 Spleen-metastasizing leukemia cells, 186 Src homology-2 (SH2), biological activities of Ron, 14–15 Stem cell derived tyrosine kinase (STK), 2 Strongylocentrotus purpuratus, 2 Suicide gene, 147–148 Sunitinib, antiangiogenic drug, 118 Suppressor of cytokine signaling (SOCS)–1 proteins, 61 SU11248, RTK inhibitor, 118 Survivin, as tumorigenic marker, 142

T

TAM Receptors and cancer, 62–68 cellular functions and, 49–54 cloning/nomenclature, 39–40 expression patterns of, 40–41 ligands and crystal structures, 41–44 regulation of, 44–49 signaling pathways, 54–62 structure and activation of, 36–38 therapeutic potential of, 68–71 Tcf4, in Wnt pathway, 88 T47D and ZR 75–1 cells and Ron receptor, 20 Teratocarcinomas, 138–139 Teratomas, 134 TGF- . See Transforming growth factor- Thrombospondin-1 (TSP-1), 121–122 Transendothelial migration, 210 Transforming growth factor- , 93, 209 Tumor angiogenic factors, 114–115 antiangiogenic drugs, side effects of, 119–121 and blood vessels, 115–117 drug resistance, 118–119 therapeutic targets, 117–118 angiogenesis and Ron receptor tyrosine kinase, 19 cyclin overexpression D1 and c-myc, 13–14 endothelium interaction, 210–212 fibroblasts interaction, 207–210 immunoglobulin interaction, 214–217 inducing viruses, 160 macrophage interaction, 212–213 microenviroment, 203–204

237

Index premetastatic niche in, 213–214 related mutations, 160 tissue, composition of, 203–204 types and tumor-derived cell lines, Ron expression, 19–20 stress responses in, 205–207 suppression, 161 vasculature, 115–117 Tumor-associated macrophages (TAMs), 212 Tumor cell adhesion, 184 endothelial cells, 184–186 adhesion, murine tumor cells and capillary epithelium, 185 attachment of DCC to, 186 parenchymal cells, 186–187 Tumor endothelial markers (TEMs), 211 Tumor formation and progression, in intestine, 89 metastasis formation in, 101 notch signaling, 91 Ras signaling in, 98 researches and future studies on, 106 RTKs in, 95 Tumor suppressor gene (TSG), 168 Type I and II GCTs, 136 Tyro-3 receptor signaling, 61–62 and TAM family, 36–39

Tyrosine kinase, D1232V and M1254T domain, 10

V

Vascular endothelial growth factor (VEGF), 96, 113, 210 VEGF-A antagonists for, 117 anti-VEGF therapy, 118–119 role in tumor cells, 115 Vertical growth phases (VGP), 178 Villi, in gastrointestinal tract, 86 Vitamin K-dependent protein Gas6, 41–44

W

Wild-type Axl (Axl-WT), 69 Wild-type EGFR, cotransfection of dominant-negative Ron with, 17–18 Wild-type Met and Ron receptors expression in COS cells, 16 Wnt signaling pathway, in intestine, 88–89 Hh signals and, 92–93 K-Ras in, 99–100 and PTEN, 94 Wwox tumor suppressor, 66–67