The HNSCC cell line, FaDu was stably transfected with control vector (FaDu) or with plasmid expressing small interfering RNA against EMMPRIN (FaDu/siE). Tumor cells were treated with bevacizumab (0, 25, 50, and 75 ng/ml) in vitro, and then cell counts were performed at 72 h. For in vivo analysis, tumor cells were xenografted onto the flank of SCID mice, and were treated with 100 microg bevacizumab twice weekly for three weeks. Xenograft samples from the control and treatment groups were analyzed for microvessel density. Escalating doses of bevacizumab had no effect on the growth of tumor cells in vitro (P.or=0.086). However, tumor xenografts expressing EMMPRIN responded to bevacizumab treatment (P=0.0013), whereas the EMMPRIN knockdown cell line did not (P=0.7942). Immunohistochemical analysis demonstrated that microvascular density was reduced in the treated FaDu tumors (P=0.005), but not in the FaDu/siE tumors (P=0.48). Currently there is limited information on biomarkers to predict response to bevacizumab. By demonstrating effectiveness of bevacizumab therapy in tumors that express EMMPRIN, but not in tumors with silenced EMMPRIN expression, this study suggests that EMMPRIN may serve as a biomarker for response to bevacizumab treatment.