Tumor and stromal pathways mediating refractoriness/resistance to anti-angiogenic therapies Yongping Crawford and Napoleone Ferrara Genentech, Incorporated, 1 DNA Way, South San Francisco, CA 94080, USA Identification and characterization of VEGF as an import- ant regulator of angiogenesis, and FDA approval of the first anti-angiogenic drugs, has enabled significant advances in the therapy of cancer and neovascular age-related macular degeneration. However, similar to other therapies, inherent/acquired resistance to anti- angiogenic drugs may occur in patients, leading to dis- ease recurrence. Recent studies in several experimental models suggest that tumor and non-tumor (stromal) cell types may be involved in the reduced responsiveness to the treatments. The present review examines the role of tumor- as well as stromal cell-derived pathways involved in tumor growth and in refractoriness to anti- VEGF therapies. Introduction It is generally accepted that angiogenesis (Box 1) is a rate- limiting process in tumor growth [1–3]. Over the last few years, several clinical trials have demonstrated the clinical benefits conferred by anti-angiogenic agents on cancer and age-related macular degeneration (reviewed in [2,4]). Vascular endothelial growth factor (VEGF) and its receptors represent the most commonly targeted signaling pathway in angiogenesis [5]. VEGF-A (VEGF hereafter) is the prototype member of a gene family that also includes VEGF-B, VEGF-C, VEGF-D and placental growth factor (PlGF). VEGF binds to two tyrosine kinase receptors: VEGFR1 and VEGFR2. Despite its weaker binding affinity to VEGF compared with that of VEGFR1, it is VEGFR2 that primarily mediates VEGF signaling within endo- thelial cells [6]. The current FDA-approved anti-angiogenic agents inhibit the VEGF pathways. The first anti-angio- genic agent to be approved is bevacizumab (Avastin, Gen- entech), a humanized anti-VEGF monoclonal antibody. Administration of bevacizumab, in combination with cyto- toxic chemotherapy, conferred benefits to patients with metastatic colorectal cancer, non-squamous, non-small- cell lung carcinoma and metastatic breast cancer [7–9]. Most recently, bevacizumab has been FDA-approved also for the therapy of renal cell carcinoma (in combination with interferon-alfa) and glioblastoma multiforme. Addition- ally, two small-molecule inhibitors targeting VEGFRs and other kinases, sorafenib (Bayer and Onyx pharmaceu- ticals) and sunitinib (Pfizer), have been approved based on their efficacy in treating renal or hepatocellular cancer [10–13]. However, not all cancer patients benefit from such anti-angiogenic therapies, and some who benefit initially might develop resistance during the treatment as well as showing some adverse effects [11,13,14]. Hence, there is an urgent need to understand the mechanisms (intrinsic or acquired) that mediate refractoriness/resistance to anti- angiogenic agents. This article will discuss current understanding of tumor- and stromal-derived molecular pathways mediat- ing VEGF-independent tumor angiogenesis. Significant overlaps occur in several cases. Vascular co-option and acquisition of an invasive phenotype as potential mechanisms of resistance to anti-angiogenic therapy Several mechanisms of inherent refractoriness or acquired resistance to antiangiogenic agents have been identified in preclinical models (reviewed in [11,15]). Tumor cells may become less sensitive to hypoxia or nutrient deprivation induced by anti-angiogenic agents via selection of resistant clones [16]. Tumor vessels can also be or become less sensitive to anti-angiogenic agents, and sustained tumor angiogenesis could occur through VEGF-independent mechanisms [11,13,15,17]. Also, in some cases, the tumor (at least initially) co-opts existing vessels and thus obviates the need to promote neovascularization [18,19]. VEGF-pathway inhibitors have been shown to result in suppression of primary and metastatic tumor growth in many studies [20–32], but the possibility that VEGF block- ade may, in some circumstances, result in an invasive phenotype has recently received considerable attention [33,34]. The earliest evidence suggesting such a possibility is from a study in an intracranial glioblastoma model in Review Glossary RIP-Tag: Transgenic mouse strain that develops insulinoma (tumor of the endocrine pancreas). Tumorigenesis is driven by the SV40 large T-antigen under control of the insulin promoter. FOLFOX: Chemotherapy regimen used for treatment of colorectal cancer. It comprises folinic acid (FOL), 5-fluorouracil (F) and oxaliplatin (OX). Depending on dosing and scheduling, several variants of the FOLFOX regimen are currently used. CT26: An N-nitroso-N-methylurethane-(NNMU)-induced, undifferentiated co- lon carcinoma cell line in Balb/c mice. EL4: T-lymphoma cell line established from a tumor induced by 9,10-dimethyl- 1,2-benzanthracene in C57/BL/6N mice. LLC: Lewis lung carcinoma. A tumor cell line established from the lung of a C57BL mouse bearing a tumor resulting from implantation of primary Lewis lung carcinoma. Corresponding author: Ferrara, N. (nf@gene.com). 624 0165-6147/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.tips.2009.09.004