Methylomics analysis identifies epigenetically silenced genes and implies an activation of β-catenin signaling in cervical cancer

Using DNA methylation biomarkers in cancer detection is a potential direction in clinical testing. Some methylated genes have been proposed for cervical cancer detection; however, more reliable methylation markers are needed. To identify new hypermethylated genes in the discovery phase, we compared the methylome between a pool of DNA from normal cervical epithelium (n = 19) and a pool of DNA from cervical cancer tissues (n = 38) using a methylation bead array. We integrated the differentially methylated genes with public gene expression databases, which resulted in 91 candidate genes. Based on gene expression after demethylation treatment in cell lines, we confirmed 61 genes for further validation. In the validation phase, quantitative MSP and bisulfite pyrosequencing were used to examine their methylation level in an independent set of clinical samples. Fourteen genes, including ADRA1D, AJAP1, COL6A2, EDN3, EPO, HS3ST2, MAGI2, POU4F3, PTGDR, SOX8, SOX17, ST6GAL2, SYT9, and ZNF614, were significantly hypermethylated in CIN3+ lesions. The sensitivity, specificity, and accuracy of POU4F3 for detecting CIN3+ lesions were 0.88, 0.82, and 0.85, respectively. A bioinformatics function analysis revealed that AJAP1, EDN3, EPO, MAGI2, and SOX17 were potentially implicated in β-catenin signaling, suggesting the epigenetic dysregulation of this signaling pathway during cervical cancer development. The concurrent methylation of multiple genes in cancers and in subsets of precancerous lesions suggests the presence of a driver of methylation phenotype in cervical carcinogenesis. Further validation of these new genes as biomarkers for cervical cancer screening in a larger population-based study is warranted. What's New? The identification of novel genes that are hypermethylated in cancer and precancerous lesions is needed in order to achieve a better sensitivity and specificity in cervical cancer screening. Using a genome-wide approach, here the authors identified 14 genes that were frequently hypermethylated in CIN3+ and might thus become useful biomarkers in future molecular cervical cancer screening. A bioinformatics function analysis revealed that five of these genes were potentially implicated in β-catenin signaling, suggesting the epigenetic dysregulation of Wnt signaling during cervical cancer development. The concurrent hypermethylation of multiple genes also suggests the involvement of a CpG island methylator phenotype.