使用Infinium HumanMethylation450 BeadChip（HM450）技术对前列腺癌DNA甲基化进行全基因组分析

[Genome-wide analysis of DNA methylation in prostate cancer using the technology of Infinium HumanMethylation450 BeadChip (HM450)].

作者信息

Babalyan K A, Sultanov R, Generozov E V, Zakharzhevskaya N B, Sharova E I, Peshkov M N, Vasilev A O, Govorov A V, Pushkar D Yu, Prilepskaya E A, Danilenko S A, Babikova E A, Larin A K, Govorun V M

出版信息

Vopr Onkol. 2016;62(1):122-32.

PMID:30444590

Abstract

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.

摘要

利用DNA芯片技术Infinium HumanMethylation 450 BeadChip，对12对前列腺腺癌及形态改变组织样本进行了DNA甲基化定量分析。对基因组差异甲基化区域的分析显示，21610个低甲基化和3852个高甲基化的CpG位点与异常状态相关。癌症基因组中高甲基化位点占主导地位且主要定位于基因调控区域，这表明它们在前列腺癌（PCa）发病机制中基因抑制机制的实施可能发挥作用。对于所研究的14个基因，其启动子区域的阵列最大高甲基化值（>50%的CpG位点）与治疗组之间的高甲基化差异水平（>40%）相结合。其中一些基因高甲基化的作用：AOX1、KLF8、ZNF154、TMEM106A在前列腺癌发病机制中的作用先前已被证实。ACSS3、TAC1、TUBA4B、ZSCAN12基因的高甲基化在前列腺癌中先前未被报道，但与其他癌症相关。反过来，GPRASP1、NKX2-6、ARX、CYBA、EPSTI1、RHCG基因甲基化水平的差异在多项肿瘤基因组研究中已有记录，但尚未详细研究。为了评估前列腺癌表观遗传标记物的诊断潜力，对最能可靠地区分肿瘤样本与非肿瘤样本组的单个CpG位点进行了无偏选择。在基于逻辑回归的选定诊断模型中纳入了9个CpG位点。该模型在癌症基因组图谱（TCGA）前列腺癌项目的40对样本的独立甲基化数据集上进行了验证，该数据集在同一版本的DNA芯片上进行分析。验证后获得的模型诊断信息汇总率（特异性95%，敏感性97%，诊断试验曲线下面积（ROC）-0.96），使我们能够将这些CpG位点视为前列腺癌分子诊断的潜在标记物。