Purpose: We investigated the aberrant methylation profile of prostate cancers and correlated the data with clinical findings.
Experimental design: Gene promoter methylation was analyzed in 101 prostate cancer samples. In addition, we analyzed 32 nonmalignant prostate tissue samples, which included 25 with benign disease, benign prostatic hypertrophy, or prostatitis, and 7 normal tissues adjacent to cancer. The methylation status of 10 genes was determined. The methylation index (MI) was calculated as a reflection of the methylated fraction of the genes examined.
Results: Methylation percentages of the genes tested in prostate cancers were: RARbeta, 53%; RASSF1A, 53%; GSTP1, 36%; CDH13, 31%; APC, 27%; CDH1, 27%; FHIT, 15%; p16(INK4A), 3%; DAPK, 1%; and MGMT, 0%. Methylation percentages in nonmalignant tissues were much lower. For clinicopathological correlations, we divided the cancer cases into low (6 or less) or high (7 or more) Gleason score (GS) groups, and into low (8 ng/ml or less) or high (greater than 8 ng/ml) preoperative serum prostate-specific antigen (PSA) groups. Methylation of RASSF1A, GSTP1, RARbeta, and CDH13 genes was significantly more frequent in the high GS group than in the low GS group. Methylation of RASSF1A, CDH1, and GSTP1 genes was significantly more frequent in the high PSA group than in the low PSA group. The median MIs were significantly higher in the high GS and the high PSA groups. According to the Spearman rank-correlation test, there was significant correlation between MI and GS (coefficient = 0.43, P < 0.0001) and the preoperative serum PSA (coefficient = 0.37, P = 0.0003).
Conclusions: Our results indicate that the methylation profile of prostate cancers correlates with clinicopathological features of poor prognosis.