Contributors Numbers in parentheses indicate the pages on which the authors’ contributions begin. Stephan E. Baldus, Institute of Pathology, Medical Faculty of the University of Cologne, 50924 Cologne, Germany (201) Nabeel Bardeesy, Department of Adult Oncology, Dana-Farber Cancer In- stitute, 44 Binney Street, Boston, Massachussetts 02115 (123) Zsofia Berke, Division of Clinical Virology F68, Department of Immunolo- gy, Microbiology, Pathology and Infectious Diseases, Karolinska Insti- tutet, Huddinge University Hospital, 141 86 Huddinge, Sweden (249) Federico Caligaris-Cappio, Department of Biomedical Sciences and Human Oncology, University of Torino, Torino; Division of Clinical Immunology, Ospedale Mauriziano Umberto I, Torino; and Laboratory of Tumor Im- munology, IRCC, Candiolo, Italy (157) Ann F. Chambers, Department of Oncology, Department of Medical Bio- physics, Department of Microbiology & Immunology, University of West- ern Ontario; and London Regional Cancer Centre, London, Ontario N6A 4L6, Canada (91) Julie M. Cherrington, SUGEN, Inc., South San Francisco, California 94080 (1) Lynda Chin, Department of Adult Oncology, Dana-Farber Cancer Institute, and Department of Dermatology, Harvard Medical School, Boston, Mass- achussetts 02115 (123) Tina Dalianis, Division of Clinical Virology F68, Department of Immunol- ogy, Microbiology, Pathology and Infectious Diseases, Karolinska Insti- tutet, Huddinge University Hospital, 141 86 Huddinge, Sweden (249) Ronald A. DePinho, Department of Adult Oncology, Dana-Farber Cancer Institute, and Department of Medicine (Genetics), Harvard Medical School, Boston, Massachussetts 02115 (123) Patrick W. B. Derksen, Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands (39) Paolo Ghia, Department of Biomedical Sciences and Human Oncology, Uni- versity of Torino, Torino; Division of Clinical Immunology, Ospedale ix Mauriziano Umberto I, Torino; and Laboratory of Tumor Immunology, IRCC, Candiolo, Italy (157) Alan C. Groom, Department of Medical Biophysics, University of Western Ontario, London, Ontario N6A 5C1, Canada (91) Franz-Georg Hanisch, Institute of Biochemistry, Medical Faculty of the Uni- versity of Cologne, 50924 Cologne, Germany (201) Richard Longnecker, Northwestern University Medical School, Chicago, Illi- nois 60611 (175) Ian C. MacDonald, Department of Medical Biophysics, University of West- ern Ontario, London, Ontario N6A 5C1, Canada (91) Vincent L. Morris, Department of Oncology, Department of Medical Bio- physics, Department of Microbiology & Immunology, University of West- ern Ontario, Ontario N6A 5C1, Canada (91) Steven T. Pals, Department of Pathology, Academic Medical Center, Uni- versity of Amsterdam, Amsterdam, The Netherlands (39) Eric E. Schmidt, Department of Medical Biophysics, University of Western Ontario, London, Ontario N6A 5C1, Canada (91) Laura K. Shawver, SUGEN, Inc., South San Francisco, California 94080 (1) Marcel Spaargaren, Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands (39) Laurie M. Strawn, SUGEN, Inc., South San Francisco, California 94080 (1) Taher E. I. Taher, Department of Pathology, Academic Medical Center, Uni- versity of Amsterdam, Amsterdam, The Netherlands (39) Ronald van der Neut, Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands (39) Robbert van der Voort, Department of Pathology, Academic Medical Cen- ter, University of Amsterdam, Amsterdam, The Netherlands (39) Kwok-Kin Wong, Department of Adult Oncology, Dana-Farber Cancer In- stitute, 44 Binney Street, Boston, Massachussetts 02115 (123) x Contributors New Paradigms for the Treatment of Cancer: The Role of Anti-Angiogenesis Agents Julie M. Cherrington, Laurie M. Strawn, and Laura K. Shawver SUGEN, Inc., South San Francisco, CA 94080 I. Introduction II. Growth Factors and Receptor Tyrosine Kinases A. VEGF and Receptors B. PDGF and Receptors C. Angiopoietin and Ties D. FGF and Receptors E. HGF/SF and Met F. EGF and Receptors G. RTK Inhibitors in Clinical Development III. Platelet-Derived Endothelial Cell Growth Factor/Thymidine Phosphorylase A. PD-ECGF/TP in Angiogenesis B. PD-ECGF/TP Inhibitors IV. Matrix Metalloproteinases A. MMPs in Angiogenesis B. MMP Inhibitors V. Plasminogen Activator/Plasmin System VI. Integrins A. Integrins in Angiogenesis B. Inhibitors of Integrins VII. Other A. Cytokines B. Pleiotrophin VIII. Conclusions References Angiogenesis, the sprouting of new blood vessels, plays a role in diverse disease states including cancer, diabetic retinopathy, age-related macular degeneration, rheumatoid arthritis, psoriasis, atherosclerosis, and restenosis. With regard to cancer, the clinical association of tumor vascularity with tumor aggressiveness has been clearly demon- strated in numerous tumor types. The observation of increased microvessel density in tumors not only serves as an independent prognostic indicator, but also suggests that anti-angiogenic therapy may be an important component of treatment regimens for cancer patients. The complexity of the angiogenic process, which involves both posi- tive and negative regulators, provides a number of targets for therapy. Many positive Advances in CANCER RESEARCH Copyright © 2000 by Academic Press. 0065-230X/00 $35.00 All rights of reproduction in any form reserved. regulators, including growth factor receptors, matrix metalloproteinases, and integrins, have been correlated with increased vascularity of tumors and poor prognosis for pa- tient survival. Thus, these serve as ideal targets for anti-angiogenesis therapy. Many in- hibitors of these targets are currently undergoing clinical evaluation as potential anti- cancer agents. In this article, we discuss the role of positive regulators in angiogenesis and tumor growth and describe the anti-angiogenic agents under development. © 2000 Academic Press. I. INTRODUCTION In recent years, angiogenesis has become a rich area of research due to the role this process plays in a number of diseases as well as the potential for therapeutic intervention. Angiogenesis occurs during development and in normal adults during wound healing, pregnancy, and corpus luteum forma- tion. Although angiogenesis is limited in normal adults, it is induced in many disease states including cancer, diabetic retinopathy, rheumatoid arthritis, psoriasis, atherosclerosis, and restenosis (reviewed in Folkman, 1995). The requirement for angiogenesis in tumor growth was first hypothesized by Folkman in 1971. Ample evidence has since been collected to show that tumors require angiogenesis to grow beyond 1–2 mm3 (Folkman, 1990). A switch to an angiogenic phenotype occurs when the balance of positive and negative regulators shifts, causing tumors to progress from hyperplasia to neoplasia (Hanahan and Folkman, 1996). The increased blood flow to the tumor allows for continued growth as well as metastasis because successful metastasis requires the presence of blood vessels to allow for the tumor cells to enter the circulation (Fig. 1, see color plate). The close interplay between angiogenesis and metastasis contributes to the poor prognosis seen in pa- tients with highly angiogenic tumors. The main players in angiogenesis are the endothelial cells that line the blood vessels. Endothelial cells release pro- teases that degrade the basement membrane, which enables their migration through the membrane to form sprouts and their proliferation to extend the sprouts (Figs. 1A and B). The sprouts fuse to form loops, allowing blood to flow to the tissue (Fig. 1C). In metastasis, proteases released from the tumor cells also degrade the basement membrane as the tumor cells migrate from the tumor to a vessel (Fig. 1D) or from a vessel to a new site (Fig. 1E). Once the tumor cells are established in a new organ (Fig. 1F), the angiogenic process may begin again to allow growth of the metastases. Tumor angiogenesis has been investigated in numerous studies in which archival tumor material has been analyzed by immunohistochemical stain- ing of vasculature markers such as Factor VIII, CD31 or CD34. Increased tumor vessel density relative to normal tissue has been shown in many tu- mor types including brain (Wesseling et al., 1998), colon (Takahashi et al., 2 Julie M. Cherrington et al. 1995), and breast tumors (Weidner et al., 1991). Furthermore, vessel densi- ty frequently correlates with progression and severity of disease. For exam- ple, astrocytomas and anaplastic astrocytomas have limited microvascula- ture, but higher grade glioblastoma multiforme have significantly increased microvasculature (Wesseling et al., 1998). In non-small-cell lung carcinoma (NSCLC), dysplasias and in situ carci- nomas had higher vessel counts than less severe hyperplastic/metaplastic le- sions (Fontanini et al., 1996). Furthermore, 94 out of 253 NSCLC patients studied who developed metastases had higher vessel counts than the patients who did not develop metastases (Fontanini et al., 1995). In two additional studies of NSCLC patients, microvessel count was found to be a better prog- nostic indicator than tumor size or lymph node status (Fontanini et al., 1997; Matsuyama et al., 1998). Intratumoral vessel count was also found to be a prognostic indicator in gastric carcinoma (Tanigawa et al., 1997). Metastatic colon cancers have been shown to have approximately 70% more vessels than nonmetastatic tu- mors (Takahashi et al., 1995). In node-negative breast tumors, low micro- vessel density correlated with longer disease-free survival (Karaiossifidi et al., 1996; Obermair et al., 1997), whereas high vessel density associated with lymph node positivity and other metastasis (Karelia et al., 1997; Fox et al., 1997; Weidner et al., 1991). Clearly, neovascularization is required for solid tumor growth to supply the proliferating tumor cells with nutrients. Other tumor types such as Ka- posi’s sarcomas (KS) and hemangioblastomas originate from vessels and are characterized by abnormal, highly permeable vessels. KS lesions contain abundant and abnormally permeable vasculature (Masood et al., 1997). The spindle-shaped cells in the lesions are of endothelial and macrophage origin (Ensoli and Sirianni, 1998). A virus, Kaposi’s sarcoma-associated herpes virus (KSHV), is found in all KS lesions and carries an oncogene that encodes a G-protein-coupled receptor (Bais et al., 1998). This receptor is constitu- tively activated and it couples to signaling pathways that lead to the ex- pression of angiogenic factors. These factors stimulate proliferation of the endothelial cell-derived KS cells as well as mediating angiogenesis. Von Hip- pel–Lindau (VHL) disease is also characterized by vascular lesions. VHL is frequently manifested by hemangioblastomas of the retina and central ner- vous system (for review, see Maher and Kaelin, 1997; Wizigmann-Voos and Plate, 1996). These lesions consist of a capillary network made up of vascu- lar endothelial cells, stromal cells, and pericytes. Angiogenic factors have also been implicated in development of and vessel permeability in these le- sions. The inhibition of angiogenesis through one of its positive regulators may serve as a valuable, novel therapy for solid tumors, KS, and VHL disease. The low incidence of angiogenesis in healthy adults would be expected to Role of Anti-Angiogenesis Agents 3 minimize the potential for side effects by agents that specifically block the angiogenic process. Because angiogenesis is controlled by many positive and negative regulators, there are a number of potential targets for inhibi- tion. Well-characterized inducers include growth factors, cytokines, and proteinases. Endogenous inhibitors of angiogenesis have also been described, including thrombospondin, platelet factor-4, tissue inhibitors of matrix metalloproteinase (TIMPs), transforming growth factor-� (TGF-�), inter- feron �, placental proliferin-related proteins, interleukin 12, angiostatin, and endostatin (reviewed in O’Reilly et al., 1997; Folkman, 1995). Specific tar- gets for anti-angiogenesis agents that have been investigated in clinical trials include inhibitors of growth factor receptor tyrosine kinases, matrix metal- loproteinases, and integrins. These are reviewed in more detail here, and oth- er important targets such as platelet-derived endothelial cell growth factor/ thymidine phosphorylase (PD-ECGF/TP) and the plasminogen activator (PA) system are discussed. II. GROWTH FACTORS AND RECEPTOR TYROSINE KINASES Some of the most well-characterized regulators of angiogenesis are growth factors and receptor tyrosine kinases (RTKs) involved in the migration and proliferation of endothelial cells (Fig. 2, see color plate). More than 50 mem- bers of the RTK family are characterized by an extracellular ligand binding domain, a transmembrane domain and an intracellular catalytic domain that transfers phosphate from ATP to substrate proteins initiating a signaling cas- cade (reviewed in Shawver et al., 1997). Of primary interest for angiogene- sis are Flt-1 and Flk-1/KDR, the receptors for vascular endothelial growth factor (VEGF), as well as Tie1 and Tie2/Tek, the receptors for angiopoietins. These four receptors are expressed primarily on endothelial cells and play a direct role in angiogenesis. Additional RTKs with broader expression patterns implicated in angiogenesis are platelet-derived growth factor recep- tors (PDGFRs); fibroblast growth factor receptors (FGFRs); the hepatocyte growth factor/scatter factor (HGF/SF) receptor, Met; and epidermal growth factor receptors (EGFRs). A. VEGF and Receptors VEGF is a dimeric protein also known as vascular permeability factor be- cause it acts on endothelial cells to regulate permeability as well as prolifer- ation. These two activities are mediated through its tyrosine kinase recep- 4 Julie M. Cherrington et al. tors, VEGFR1/Flt-1 and VEGFR2/Flk-1/KDR (KDR is the human homolog of Flk-1). VEGF and its receptors are expressed in angiogenic tissues during development, wound healing, and other situations when angiogenesis oc- curs. The temporal and spatial patterns of expression of VEGF and its re- ceptors as well as the results of targeted mutagenesis support that they are required for angiogenesis during development. Similarly, the role of VEGF in tumor angiogenesis has been clearly demonstrated using tumor models in rodents (reviewed in Hanahan, 1997; Shawver et al., 1997). An extensive literature exists linking VEGF with human cancer. In human tumors, VEGF mRNA or protein has been identified by reverse transcriptase polymerase chain reaction, in situ hybridization, or immunohistochemistry in primary gliomas (Plate et al., 1994), colon cancer (Takahashi et al., 1995; Tokunaga et al., 1998; Landriscina et al., 1998), NSCLC (Fontanini et al., 1999; Takahama et al., 1998), pulmonary adenocarcinoma (Takanami et al., 1997), renal cell tumors (Takahashi et al., 1994), and KS (Cornali et al., 1996). In pulmonary adenocarcinoma (Takanami et al., 1997) and NSCLC (Fontanini et al., 1999; Takahama et al., 1998; Ohta et al., 1996), survival of patients with VEGF-positive tumors was significantly less than patients with VEGF-negative tumors. For example, in one study of NSCLC, patients with low VEGF levels had a median survival time of 151 months, whereas those with high VEGF expression had a mean survival time of only 8 months (Ohta et al., 1996). Likewise, poor prognosis, as well as increased micro- vessel counts were also found in patients with high VEGF expression at ei- ther the mRNA or protein level in many studies of both node-negative and node-positive breast cancer patients (Eppenberger et al., 1998; Scott et al., 1998; Anan et al., 1998; Relf et al., 1997; Brown et al., 1995). The presence of VEGF in serum (Landriscina et al., 1998; Fujisaki et al., 1998; Kumar et al., 1998) as well as in tumors (Landriscina et al., 1998; Fujisaki et al., 1998) was correlated with stage of disease in colon cancer patients, where liver metastasis (Tokunaga et al., 1998) and poor prognosis correlated to VEGF levels (Tokunaga et al., 1998; Ishigami et al., 1998; Hyodo et al., 1998). In KS cells, VEGF expression is induced by the G-protein-coupled recep- tor oncogene expressed by KSHV. VEGF mediates angiogenesis in these le- sions and also stimulates the growth of the spindle cells (Bais et al., 1998). VEGF has also been detected in hemangiomas in VHL disease patients (Wiz- igmann-Voos et al., 1995). This factor appears to be secreted from stromal cells and is likely to act through a paracrine mechanism to stimulate capil- lary formation and to increase vessel permeability in the lesions. KDR and Flt-1 mRNA have also been detected in tumors such as gliomas (Plate et al., 1994), neuroblastomas (Rossler et al., 1999), colon cancer (Takahashi et al., 1995), and adenocarcinomas (Takanami et al., 1997). KDR was expressed at higher levels in malignant breast tissue than sur- rounding tissue (Brown et al., 1995; Kranz et al., 1999). In 51 cases of in- Role of Anti-Angiogenesis Agents 5 testinal-type gastric cancer, vessel count correlated with high KDR expres- sion on endothelium, and vessel count in turn correlated with progression of disease and liver metastasis (Takahashi et al., 1996). High Flt-1 expression in pulmonary adenocarcinoma correlated with poor survival (Takanami et al., 1997). In these tumors, as with the intestinal-type gastric tumors, the re- ceptors were detected on the endothelial cells of the vessels and not the tu- mor cells. This strongly suggests a paracrine mechanism in which VEGF se- creted from tumor cells stimulates migration and proliferation of endothelial cells. In KS, KDR was found to be expressed on the tumor cells as well as on en- dothelial cells in stromal vessels (Brown et al., 1996; Masood et al., 1997). VEGF produced by tumor cells and squamous epithelium in these lesions may stimulate growth of the tumor itself as well as the vessels. This is sup- ported by experiments with VEGF antisense oligonucleotides (Masood et al., 1997) and antibodies (Nakamura et al., 1997), both of which inhibited the growth of KS cells in culture. Recently, it was discovered that the Tat pro- tein of human immunodeficiency virus can bind to and activate KDR (Gan- ju et al., 1998), thus contributing to angiogenesis and the proliferation of spindle cells in lesions of AIDS-associated KS patients. In hemangiomas from VHL disease patients, KDR and Flt-1 were shown to be coexpressed on en- dothelial cells (Wizigmann-Voos et al., 1995). Although VEGF and its re- ceptors act by a paracrine mechanism in these lesions, it is unclear how they are regulated (Wizigmann-Voos and Plate, 1996). VEGF and Flt-1 may function via an autocrine loop mechanism in hematopoietic neoplasms. Five of 12 human hematopoietic tumor cell lines were found to express mRNA for both VEGF and Flt-1, although there was no expression of KDR mRNA in these cell lines (Bellamy et al., 1999). Al- though Flt-1 and KDR were not detected on normal stem cells, VEGF stim- ulated proliferation of hematopoietic cells from approximately 15% of chronic and acute myeloid leukemia (AML) patients studied (Ratajczak et al., 1998). Leukemic blasts from AML patients were shown to express VEGF, Flt-1, and KDR transcripts in another study (Fiedler et al., 1997). The role of angiogenesis in leukemias is an important area of further investiga- tion. B. PDGF and Receptors PDGF is a pleiotropic factor that exists as a homo- or heterodimer of two polypeptides, the A and B chains, which interact with two receptor subtypes, the PDGF � receptor (PDGFR�) and PDGF � receptor (PDGFR�). PDGF re- ceptors have been found to be expressed in microvascular endothelium in vivo when endothelial cell activation and angiogenesis occur. However, the 6 Julie M. Cherrington et al. mechanism whereby PDGF stimulates angiogenesis has been controversial. PDGF has been found to induce tube formation by endothelial cells (Batte- gay et al., 1994) and an angiogenic response in the chick chorioallantoic membrane assay (Wilting et al., 1992). In some studies, PDGF induced the proliferation of endothelial cells (Marx et al., 1994; Koyama et al., 1994) al- though not all endothelial cells respond mitogenically to PDGF (Sato et al., 1993). Endothelial cell migration in response to the homodimer PDGF-BB has been shown using video time-lapse microscopy (Thommen et al., 1997). PDGF also has been reported to upregulate other angiogenic factors such as VEGF (Enholm et al., 1997) and thus it has been postulated that its activat- ing role in angiogenesis is indirect. Another indirect mechanism was discov- ered by studies in PDGF-BB-deficient mice (Lindahl et al., 1997). These mice were found to lack microvascular pericytes, small cells that form part of the capillary wall. It appears that neovessels were unable to attract PDGFR�-ex- pressing pericyte progenitor cells, leading to instability of the capillaries. PDGF and its receptors have been detected in cancers such as gliomas, (Plate et al., 1992; Hermanson et al., 1992), lung carcinomas (Antoniades et al., 1992), melanoma (Antoniades et al., 1992), prostate carcinomas (Story, 1991), and esophageal carcinomas (Yoshida et al., 1993). In breast cancer patients, circulating PDGF was found to be higher in patients with meta- stases (Ariad et al., 1991). In another study, PDGF-A mRNA was detected in a higher number of tumor samples than nontumor breast tissue and ves- sel counts correlated with PDGF-A expression (Anan et al., 1996). PDGFR has been shown to be expressed on vascular endothelial cells in breast tu- mors (Bhardwaj et al., 1996). During tumor progression, PDGFRs on tumor neovasculature were upregulated (Plate et al., 1992). In colorectal (Lindmark et al., 1993) and breast carcinomas (Bhardwaj et al., 1996), PDGF and PDGFR� were detected on stroma. Because PDGF and its receptors are ex- pressed in tumor cells, stroma, and pericytes, they may play both direct and indirect roles in angiogenesis. C. Angiopoietin and Ties Tie1 and Tie2 are expressed in the vascular endothelium during embry- onic development. Tie2 (also known as TEK) knockout and transgenic mice suggest that Tie2 is important for vasculogenesis, while Tie1 may be important for maintaining vascular integrity (reviewed in Hanahan, 1997; Shawver et al., 1997). Also, Tie2 and its ligand, angiopoietin-1 (ang-1), ap- pear to be required for recruitment of pericytes and smooth muscle cells. Angiopoietin-2 (ang-2) can also bind to Tie2; interestingly ang-1 induces autophosphorylation of Tie2 while ang-2 does not. Instead, ang-2 competi- tively inhibits ang-1-induced Tie2 phosphorylation, thereby negatively reg- Role of Anti-Angiogenesis Agents 7