[Genome-wide analysis of DNA methylation in prostate cancer using the technology of Infinium HumanMethylation450...

Using the technology of DNA chips Infinium HumanMethylation 450 BeadChip it was analyzed quantitative DNA methylation status in 12 paired samples of prostate adenocarcinoma, and morphologically altered tissues. Analysis of differentially methylated regions of the genome showed an association with abnormal status for 21610 and 3852 hypomethylated hyper-methylated CpG sites. Dominance in the cancer genome hypermethylated sites and their predominant localization in the regulatory regions of genes indicate their possible role in the implementation of mechanisms of gene suppression in the pathogenesis of prostate cancer (PCa). For 14 genes studied were characterized array maximum values hypermethylation in promoter region (> 50% CpG sites) in combination with a high level of methylation differences between treatment groups (> 40%). Role of hypermethylation in some of them: AOX1, KLF8, ZNF154, TMEM106A in the pathogenesis of prostate cancer has been showed previously. Hypermethylation of genes ACSS3, TAC1, TUBA4B, ZSCAN12 not previously been shown for prostate cancer, but is characterized by the association with other cancers. In turn, the differences in the levels of methylation in genes GPRASP1, NKX2-6, ARX, CYBA, EPSTI1, RHCG been documented as a result of a number of genome-research oncology, but has not been studied in detail. To assess the diagnostic potential of epigenetic markers of prostate cancer there was carried out unbiased selection of individual CpG sites most reliably discriminate against tumor samples from a group of no tumor samples. In selected diagnostic model based on logistic regression included 9 CpG sites. Validation of the model was carried out on an independent dataset of methylation of 40 paired samples from the prostate cancer project Atlas of Cancer Genome (TCGA) analyzed on the same version of the DNA chip. Summarized rates of diagnostic informativeness of a model (specificity 95%, sensitivity of 97%, the area under the curve of the diagnostic test (ROC) - 0,96), obtained after validation, allow us to consider these CpG Sites as potential markers for molecular diagnosis of prostate cancer.