Three-microRNA signature identified by bioinformatics analysis predicts prognosis of gastric cancer patients

AIM: To identify multiple microRNAs (miRNAs) for predicting the prognosis of gastric cancer (GC) patients by bioinformatics analysis. METHODS: The original microarray dataset GSE93415, which included 20 GC and 20 tumor adjacent normal gastric mucosal tissues, was downloaded from the Gene Expression Omnibus database and used for screening differentially expressed miRNAs (DEMs). The cut-off criteria were P < 0.05 and fold change > 2.0. In addition, we acquired the miRNA expression profiles and clinical information of 361 GC patients from The Cancer Genome Atlas database to assess the prognostic role of the DEMs. The target genes of miRNAs were predicted using TargetScan, miRDB, miRWalk, and DIANA, and then the common target genes were selected for functional enrichment analysis. RESULTS: A total of 110 DEMs including 19 up-regulated and 91 down-regulated miRNAs were identified between 20 pairs of GC and tumor adjacent normal tissues, and the Kaplan-Meier survival analysis found that a three-miRNA signature (miR-145-3p, miR-125b-5p, and miR-99a-5p) had an obvious correlation with the survival of GC patients. Furthermore, univariate and multivariate Cox regression analyses indicated that the three-miRNA signature could be a significant prognostic marker in GC patients. The common target genes of the three miRNAs are added up to 108 and used for Gene Functional Enrichment analysis. Biological Process and Molecular Function analyses showed that the target genes are involved in cell recognition, gene silencing and nucleic acid binding, transcription factor activity, and transmembrane receptor activity. Cellular Component analysis revealed that the genes are portion of nucleus, chromatin silencing complex, and TORC1/2 complex. Biological Pathway analysis indicated that the genes participate in several cancer-related pathways, such as the focal adhesion, PI3K, and mTOR signaling pathways. CONCLUSION: This study justified that a three-miRNA signature could play a role in predicting the survival of GC patients.