Abstract
Functional DNA methylation abnormalities are a hallmark of human cancers and may be a promising biomarker for their early diagnosis. Moreover, the largest methylation differences can improve the sensitivity of noninvasive diagnoses of solid tumors. We combined whole-genome bisulfite sequencing (WGBS) and mRNA-seq data from 33 paired hepatocellular carcinoma (HCC) and adjacent tissues to identify methylation markers that could be used for noninvasive diagnosis in blood samples. Methylation markers were selected according to the following criteria: differentially methylated regions (DMR) located in the promoter region with large differences in methylation (Δβ > 0.3) and inverse correlation with matched gene expression (cor < -0.3). Cell-free DNA (cfDNA) from 48 patients with HCC and 24 normal participants was used to verify the performance of meTSPYL5 using qMSP. Integrated WGBS and transcriptomic data analysis identified eight target promoter hyper-DMRs. After confirming the WGBS profiles of genes in peripheral blood mononuclear cells, meTSPYL5 was selected to further verify the plasma cfDNA samples by qMSP. The results of plasma validation showed that the methylation detection of meTSPYL5 was sensitive for identifying HCC, with a sensitivity and specificity of 85.4% and 100%, respectively. Pan-cancer analysis found that the methylation level of TSPYL5 was elevated in multiple cancer types, indicating that it lacks cancer-type specificity; however, this result does not affect its application value in monitoring high-risk populations of HCC. By analyzing and integrating all available high-throughput epigenomic and transcriptomic data from human HCC tissues, we identified eight regions as potential diagnostic biomarkers for HCC. Integrative analyses of epigenomic and transcriptomic data provide an efficient method to identify diagnostic biomarkers for human cancers. Methylated TSPYL5 in plasma is a promising biomarker for the detection and screening of HCC.