[CANCER RESEARCH 60, 6253– 6258, November 15, 2000] Advances in Brief Complete Inhibition of Rhabdomyosarcoma Xenograft Growth and Neovascularization Requires Blockade of Both Tumor and Host Vascular Endothelial Growth Factor Hans-Peter Gerber, Joe Kowalski, Daniel Sherman, David A. Eberhard, and Napoleone Ferrara 1 Departments of Molecular Oncology [H-P. G., J. K., D. S., N. F.] and Pathology [D. A. E.], Genentech Incorporated, South San Francisco, California 94080 Abstract Growth of the human rhabdomyosarcoma A673 cell line in nude mice is substantially reduced but not completely suppressed after systemic administration of the antihuman vascular endothelial growth factor (VEGF) monoclonal antibody (Mab) A.4.6.1. Potentially, such escape might be attributable to incomplete local penetration of the antibody because of a diffusion barrier associated with tumor growth. Alterna- tively, it might reflect a compensatory up-regulation of murine VEGF, produced by the stroma of the host, or of other angiogenic factor genes. To test these potential mechanisms, systemic administration of Mab A.4.6.1. was performed in conjunction with intratumoral administration of an irrelevant antibody, an antihuman VEGF Fab or mFlt(1-3)-IgG that neutralizes both human and murine VEGF. Tumor growth in the systemic- plus-intratumoral anti-VEGF group was not different from that in the systemic anti-VEGF-plus-intratumoral-control antibody group, arguing against the possibility that bioavailability is the factor that limits the antitumor efficacy of Mab A.4.6.1. However, intratumoral mFlt(1-3)-IgG administration dramatically enhanced the activity of systemic anti-VEGF Mab and resulted in complete suppression of tumor growth, which indi- cated that host VEGF significantly contributes to tumor growth. Systemic administration of mFlt(1-3)-IgG alone replicated these findings. Histolog- ical analysis of residual tumor tissues revealed an almost complete absence of host-derived vasculature and massive tumor-cell necrosis in the mFlt(1- 3)-IgG groups. Such extensive necrotic areas were not present in the other groups. Real-time reverse transcription-PCR analysis of total RNA de- rived from tumor tissues indicated strong up-regulation of both human and murine VEGF as well as other genes regulated by hypoxia. Our findings emphasize the need to completely block VEGF for maximal inhibition of tumor growth. Introduction There is extensive evidence that the development of a neovascular supply is required for a variety of proliferative processes (1). VEGF 2 is a key regulator of physiological and pathological angiogenesis, and even partial inactivation of the VEGF gene results in early embryonic lethality (2, 3). Inhibition of VEGF activity by neutralizing antibodies or other inhibitors results in significant tumor suppression in a broad variety of tumor cell lines (4, 5). Also, administration of a chimeric murine soluble VEGF receptor protein, mFlt(1-3)-IgG, induces growth arrest and lethality in neonatal mice (6). However, safety evaluation studies in fully developed rodents or primates have failed to detect any significant toxicity after VEGF blockade, except inhi- bition of corpus luteum angiogenesis (6 – 8). This favorable safety profile, combined with considerable antitumor efficacy in preclinical models, has made VEGF inhibitors attractive candidates for the treat- ment of solid tumors. A humanized anti-VEGF Mab has completed Phase I and Phase II trials in cancer patients, and, currently, Phase III studies are under way. Phase II studies have shown initial evidence of clinical efficacy in non-small cell lung and colorectal carcinoma patients (9, 10). Small molecules inhibiting VEGF receptor signal transduction are undergoing clinical testing as well (for review, see Ref. 11). Systemic administration of the murine Mab A.4.6.1, directed against human VEGF, causes considerable inhibition of human rhab- domyosarcoma xenografts growth in immunodeficient mice (4). When the treatment is initiated at the same time or shortly after tumor-cell inoculation, the growth inhibition is dramatic, exceeding 90 –95%. However, if tumors are allowed to reach a significant size prior to the beginning of the treatment, the level of inhibition is less complete, and tumors escape from the inhibition. This phenomenon might reflect molecular and cellular alterations resulting from genomic instability and/or increased mutation rates in the tumor cells. This might provide the molecular framework for a compensatory up-regulation of angiogenic molecules other than VEGF, or alterna- tively, for the down-regulation of antiangiogenic genes under growth- selective conditions. Such escape could also be potentially mediated by increased production of VEGF by the stroma of the host, because Mab A.4.6.1. does not neutralize murine VEGF (4, 12). Another factor that may potentially account for tumor escape after systemic administration of Mab A.4.6.1 is the blood/tumor barrier, attributable in part to interstitial hypertension, which may prevent systemically delivered therapeutic agents from achieving optimal concentrations in the extravascular space (13). The availability of several inhibitors enabled us to study the relative contribution of all these potential mechanisms involved in angiostatic escape. Our findings indicate that neither insufficient bioavailability nor compensatory regulation of antiangiogenic or angiogenic factor, but, rather, the up-regulation of host-derived VEGF is primarily responsible for the angiostatic escape observed with the A673 rhabomyosarcoma. Materials and Methods In Vivo Experiments. Human A673 rhabdomyosarcoma cells (HTB 1598) were cultured as described previously (4). Five  10 6 cells in 0.1 ml of Matrigel were injected s.c. in the dorsal flank region of beige nude mice (Harlan Sprague Dawley). Five days after tumor cell inoculation, when the xenografts were clearly established and had reached a volume of 50 –100 mm 3 , i.p. administration of Mab A.4.6.1 (4) was initiated, at a dose of 10 mg/kg. The antibody was then given i.p. at the same dose twice weekly. In addition, animals received direct intratumoral injections of an affinity metered recom- binant humanized Fab fragment (14) originally derived from Mab A.4.6.1, murine Flt(1-3)-IgG (8) or a control murine Mab directed against HSV gly- coprotein D (Clone 3L8) of the same isotype as Mab A.4.6.1., each at the dose of 25 mg/kg. Injections were made directly into the tumor mass, from the side and underneath, using a 28-gauge needle and a 0.5-ml tuberculin syringe. Received 6/5/00; accepted 10/5/00. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 To whom requests for reprints should be addressed, at Genentech Incorporated, Department of Molecular Oncology, 1 DNA Way, South San Francisco, CA 94080. Phone: (650) 225-2968; Fax: (650) 225-6327; E-mail: nf@gene.com. 2 The abbreviations used are: VEGF, vascular endothelial growth factor; Mab, mono- clonal antibody; RT-PCR, reverse transcription-PCR; GAPDH, glyceraldehyde-3-phos- phate dehydrogenase. 6253 Research. on December 2, 2015. © 2000 American Association for Cancer cancerres.aacrjournals.org Downloaded from