Prognostic implications of aberrantly expressed methylation‑driven genes in hepatocellular carcinoma: A study based on The Cancer Genome Atlas.

Abstract

RNA‑Sequencing and methylation data for hepatocellular carcinoma (HCC) were downloaded from The Cancer Genome Atlas (TCGA). The aberrantly expressed methylation‑driven genes in HCC and normal tissues were identified using the Limma package and the MethylMix algorithm. The Database for Annotation, Visualization and Integrated Discovery and ConsensusPathDB were used for Gene Ontology (GO) enrichment and pathway analysis. Univariate and multivariate Cox regression analyses were used to construct a prognostic risk model of HCC. Survival curve and receiver operating characteristic (ROC) curves were applied to evaluate the clinical utility of the risk model. A total of 238 methylation‑driven genes were successfully identified from cancer and normal tissues. GO enrichment analysis indicated that these genes functioned in the extracellular space, interfering with lipid metabolism in hepatocytes and regulating adaptive immune responses. In total, 14 relevant pathways were identified. The following prognostic risk model was generated: Risk score=CALML3 (degree of methylation) x (‑4.860) + CCNI2 x (2.071) + TNFRSF12A x (‑3.369) + IFITM1 x (1.203) + ENPP7P13 x (‑1.366) + DDT x (2.139) + RASAL2‑AS1 x (‑1.384) + ANKRD22 x (‑3.215). The median risk score (0.970) derived from this model was set as cutoff value for assigning patients to high‑ or low‑risk group. The 5‑year survival rate was 35.8% [95% confidence interval (CI)=27.1‑47.4%] in the high‑risk group and 61.7% (95% CI=51.4‑74.2%) in the low‑risk group (P<0.0001). The ROC curve showed an area under the curve of 0.742, indicating that this model is appropriate for predicting the survival rate of patients. Furthermore, the methylation and expression levels of two key genes, tumor necrosis factor superfamily member 12A and D‑dopachrome decarboxylase, were significantly associated with prognosis and were correlated with cg00510447, cg26808293, cg11060661 and cg16132339 methylation. In conclusion, a prognostic risk model for HCC is proposed based on the bioinformatic analysis of methylation‑driven genes. The findings of the present study may improve understanding of the pathogenesis and prognosis of HCC.