Construction of a long non-coding RNA-associated ceRNA network reveals potential prognostic lncRNA biomarkers in hepatocellular carcinoma

Objectives Hepatocellular carcinoma (HCC) is the sixth leading cause of cancer-related mortality in the world. Accumulating evidence has highlighted the regulatory roles of long non-coding RNAs (lncRNAs) acting as competing endogenous RNAs (ceRNAs) in HCC. Methods The lncRNA expression data and corresponding patient information were obtained from The Cancer Genome Atlas (TCGA) database. Competing lncRNA-mRNA interactions were identified using the hypergeometric test. Co-expression analysis was implemented using the Spearman correlation coefficient. Multivariate Cox regression survival analysis was utilized to extract prognostic lncRNAs in the network. Results Based on the “ceRNA hypothesis”, a global lncRNA-associated ceRNA network (LCeNET) in HCC was constructed. Nine lncRNAs were identified as hubs and found to be enriched in various cancer-related biological processes. In addition, ceRNA pairs associated with survival were screened to construct a lncRNA-miRNA-mRNA sub network. Finally, we developed a sixteen-lncRNA model that could classify patients into high- and low-risk subgroups with different survival outcomes, and MCM3AP-AS1 functioned as a hub in both LCeNET and prognostic model. Conclusions Our work will improve the understanding of lncRNA-mediated ceRNA regulatory mechanisms in HCC pathogenesis and facilitate the identification of candidate prognostic biomarkers for HCC.