Polymorphisms in VEGF, eNOS, COX-2, and IL-8 as predictive markers of response to bevacizumab.

Abstract

e13502 Background: There are no validated predictive markers for selecting patients who would respond to bevacizumab. Previous studies analyzed the role of SNPs in VEGF, eNOS, COX-2, IL-8, but no definite prognostic and predictive association has been established. We evaluated the association of genetic variability with outcome in patients (pts) treated with bevacizumab. Methods: Fifty metastatic pts affected by colorectal (35/50), breast (11/50), and kidney (4/50) cancer were included in this study. VEGF936 C/T and VEGF2578 C/A, eNOS894 G/T, eNOS786 T/C, COX-2 8473 C/T, IL-8 251 T/A polymorphisms were selected. All pts were treated with bevacizumab plus chemotherapy (colorectal and breast cancer) or IFN (renal cancer). Response rate was related to polymorphic gene expression. The relationship of TTP/OS to SNPs is too early because most pts are still alive. The association between response and genotypes was evaluated using the Chi-square test. Results: 29/50 (58%) pts had a partial response, one patient had a complete response and 20 (40%) pts a stable disease. 39/50 (78%) pts were VEGF936 CC (wt) and 11 (22%) VEGF936 CT. 20/50 (40%) pts were VEGF2578 CC (wt), 19 (38%) VEGF2578 CA and 11 (22%) VEGF2578 AA. 16/50 (32%) pts were eNOS894 GG (wt), 27 (54%) eNOS894 GT and 7 (14%) eNOS894 TT. 17/50 (34%) pts were eNOS786 TT (wt), 24 (48%) eNOS786 CT, 9 (18%) eNOS786 CC. All patients were COX-2 8473 CC (wt). The SNP IL-8 251 T/A was tested only in patients with colorectal cancer (35 pts) and 20/35 resulted IL-8 251 TT (wt), 6 pts were IL-8 251 TA and 9 pts IL-8 251 AA. The eNOS786 TT genotype (wt), compared to genotype CC, was associated with a best response to treatment in breast (RR 36.4% vs 18.2%; p=0.036) and colorectal cancer (RR 28.6% vs. 8.6%; p=0.04). The IL-8 251 TT genotype (wt), was associated with a best response to bevacizumab in colorectal cancer, when compared with the genotype AA (RR 25.7% vs 14.3%; p=0.04). Conclusions: With increasing number of antiangiogenic agents, the need of biomarkers is becoming critical, either due to hight cost of these treatments and to response variability. Our preliminary results suggest that the study of SNPs of some genes involved in angiogenesis might be useful in improving the selection of potentially responsive pts. No significant financial relationships to disclose.