EZH2 polymorphism and benefit from bevacizumab in colorectal cancer: another piece to the puzzle.

We have recently shown that one allelic variant (rs3757441 CC versus TC/TT) of an EZH2 single nucleotide polymorphism (SNP) is significantly associated with shorter progression-free survival (PFS) and overall survival (OS) in metastatic colorectal cancer (mCRC) patients treated with first-line FOLFIRI plus bevacizumab [1]. EZH2 silences several tumor-suppressor genes, thereby enhancing cancer progression and chemoresistance [2]. Recently, EZH2 has been shown to regulate tumor angiogenesis via a paracrine circuit promoted by the vascular endothelial growth factor (VEGF) [3]: VEGF increases endothelial EZH2 expression, thus promoting angiogenesis by methylating and silencing vasohibin1. In order to explore the value of the 626-394C>T SNP in predicting benefit of the anti-VEGF antibody bevacizumab,we genotyped a parallel group of 104mCRCpatients treated with first-line FOLFIRI without bevacizumab ([1] for details about sequencing technique and statistics). FOLFIRI-treated patients received the same schedule of chemotherapy described in published paper [1]. Compared with the bevacizumab cohort in the FOLFIRI group, median age was slightly higher (68 versus 63 years) and performance status (PS) poorer (Eastern Cooperative Oncology Group PS 1 or 2: 21% versus 8%), while there was no difference with regards to the number of metastatic sites (>1: 46% versus 49%). Among FOLFIRI-treated patients, 46 achieved an objective response (response rate, RR: 44%), 37 (36%) had disease stabilization as best response and 21 (20%) progressed during treatment. At a median follow-up of 47.8 months, all patients experienced disease progression and 92 died: median PFS and OS were 8.2 and 20.6 months, respectively. As regards the 626-394C>T SNP, 2 patients harbored the CC genotype, while 102 had at least one T allele (36 TC and 66 TT, respectively). As in the bevacizumab cohort, also in the chemotherapy-only group, no significant association (P = 0.502) between EZH2 genotype and RR was found (probably due to the limited number of CC patients: in fact, RR was 0% and 45% among CC and TC/TT patients, respectively). Again, outcome did not differ between TC and TT patients (data not shown), while both PFS and OS were significantly shorter among CC patients compared with other genotypes (Figure 1). We conducted a multivariate analysis including all the 214 patients genotyped: variables were represented by administered regimen (FOLFIRI with or without bevacizumab) and EZH2 genotype (CC or TC/TT) and the analysis included a total of 14 CC patients. CC genotype remained an independent predictor of both PFS [hazard ratio (HR) 5.752, 95% confidence interval (CI) 1.358–24.369; P = 0.018] and OS (HR 6.091, 95% CI 1.447–25.635; P = 0.014). As expected, addition of bevacizumab to FOLFIRI was associated with a significant survival advantage (data not shown). We therefore tested if EZH2 genotype predicted a benefit from adding bevacizumab to chemotherapy (genotype-treatment interaction): EZH2 genotype was not predictive of bevacizumab efficacy for either PFS (HR 0.476, 95% CI 0.092–2.478; P = 0.3782) or OS (HR 0.334, 95% CI 0.064–1.738; P = 0.193). Even though the small number of CC patients prevents any conclusive confirmation of our findings, these data suggest that the 626-394C>T SNP does not represent a potential predictive marker for bevacizumab efficacy but may play a role as either a prognostic variable or a predictive factor for first-line irinotecan-based chemotherapy in mCRC patients. The latter conclusion is supported also by the results of the Oncomine analysis reported elsewhere [1], which shows that EZH2 target Figure 1. Progression-free survival (PFS) (A) and overall survival (OS) (B) according to EZH2 626-394C>T genotype in FOLFIRI-treated patients. (A) Median PFS among CC patients: 4.2 months; median PFS among TC/ TT patients: 8.2 months [hazard ratio (HR) 0.512, 95% confidence interval (CI) 0.266–0.759; P = 0.035, log-rank test]. (B) Median OS among CC patients: 8.25 months; median OS among TC/TT patients: 21.2 months (HR 0.389, 95% CI 0.143–0.635; P = 0.035, log-rank test). le tt e rs to