Identification of a novel microRNA signature for prediction of clinical prognosis and drug sensitivity/resistance analysis in Hepatocellular Carcinoma

Abstract

Abstract Prognostic stratification of hepatocellular carcinoma (HCC) remains challenging. It is found that microRNAs play a crucial role in pathogenesis of HCC. Our aims to construct a useful prognostic model based on miRNA signature to determine patient at high risk of HCC progression. Detailed information of miRNA were collected from The Cancer Genome Atlas (TCGA). We categorized HCC patients into two groups (training group and testing group) and identified the prognostic miRNA signature by univariate Cox regression analysis. A novel model based on miRNA was developed for HCC prognostic prediction. Then we apply ROC curve, KM curve and multivariate Cox regression analysis to assess the prognostic performance of the novel model. The results showed that AUC values based on ROC analysis of one, three and five-year overall survival (OS) for patients were 0.742, 0.723, 0.841, and 0.706, 0.653, 0.581, and 0.721, 0.702, 0.716 for training group, testing group and whole group, respectively. It indicated that the predictive performance of prognostic model was good. The nine-miRNA signature (miR-139, miR-125b-2, miR-125b-1, miR-4746, miR-1180, miR-592, miR-9-2, miR-9-1 and miR-9-3) could independently predict OS by using cox regression analysis method. The survival rate of patients in the high-risk group was lower than that in the low-risk group (all p < 0.001). In order to predict the potential biological functions of nine-miRNA signature, a total of 356 overlapping miRNA-target genes were analyzed by pathway enrichment analysis. In addition, the results showed that the DEGs mainly enriched in cancer-related pathways, including microRNAs in cancer, Longevity regulating pathway, AMPK signaling pathway, FoxO signaling pathway, Yersinia infection and T cell receptor signaling pathway. Function analysis of miRNAs showed that the nine miRNAs are related to various critical biological function modules and interact closely based on their functional similarity. Besides, most of miRNAs can promote sensitivity or induce resistance to some anti-cancer drugs.Taken together, these findings suggested that the nine-miRNA signature model could be conductive to the prognostic diagnosis and treatment of HCC.