Abstract
BACKGROUND AND AIMS: Hepatocellular carcinoma (HCC) is a prevalent malignancy with poor prognosis. This study uses integrated bioinformatic analyses to explore potential competing endogenous RNA (ceRNA) network chains in HCC. METHODS: HCC expression profile data were obtained from the Gene Expression Omnibus dataset, and differential expression analysis was conducted to identify differentially expressed mRNAs (DEmRNAs), microRNAs (DEmiRNAs), and long non-coding RNAs (DElncRNAs) between HCC and normal liver tissue samples. Univariate Cox regression analysis was performed to identify mRNAs associated with the prognosis of HCC patients. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses were used to classify the identified genes functionally. Cytoscape software was used to construct a protein-protein interaction network. Using the intersection method, a ceRNA network was established to align data from two databases (miRTarBase and miRcode). Pearson correlation analysis was conducted to evaluate the relationships between lncRNAs and mRNAs. RESULTS: A total of 106 prognosis-related DEmRNAs were identified between HCC and normal samples. A total of 132 DEmiRNAs and 42 DElncRNAs were dysregulated in HCC. A ceRNA network of three lncRNAs, six miRNAs, and eight mRNAs was constructed. High expression of MCM10, CDKN3, RRM2, KIF3A, and ALYREF correlated with a poor prognosis, while high expression of CPEB2, COLEC10, and PBLD was associated with a better prognosis for HCC patients. Expression analysis confirmed the differential expression of these genes in HCC samples. Correlation analysis revealed that a MAGI2-AS3/hsa-miR-450b-5p/COLEC10 axis might play a crucial role in the progression of HCC. CONCLUSION: The ceRNA network constructed could provide insight into HCC tumorigenesis and might lead to new molecular biomarkers for diagnosing and treating HCC.