Comprehensive analysis of competitive endogenous RNAs networks reveals potential prognostic biomarkers associated with epithelial ovarian cancer

Ovarian cancer (OC) is a major health threat to females, as it has high morbidity and mortality. Evidence has increasingly demonstrated that long non-coding RNAs (lncRNAs) regulate OC progression and they may have value as early diagnostic biomarkers, prognostic biomarkers and/or therapeutic targets. In the present study, the regulatory mechanisms and prognosis associated with cancer-specific lncRNAs and their related competing endogenous (ce)RNA network in OC were investigated. The differential expression profiles and prognostic significance of lncRNAs and mRNAs were systematically explored based on data from 359 OC cases from The Cancer Genome Atlas and 180 healthy individuals from the Genotype-Tissue Expression database. Functional enrichment analyses, RNA-RNA interactome prediction, ceRNA network analysis, correlation analysis and survival analysis were utilized to identify hub lncRNAs and biomarkers associated with OC diagnosis or prognosis. A total of 1,049 differentially expressed lncRNAs and 6,516 differentially expressed mRNAs between OC and healthy tissues were detected. An lncRNA-micro (mi)RNA-mRNA regulatory network in OC was further established, containing 91 lncRNAs, 23 miRNAs and 179 mRNAs. After survival analysis based on the expression of the RNAs in the ceRNA network, 8 lncRNAs, 4 miRNAs and 11 mRNAs that were significantly associated with OC patient survival (P<0.05) were obtained. Using least absolute shrinkage and selection operator-penalized Cox regression, an eight-lncRNA risk score model was generated, which was able to readily discriminate between OC and healthy individuals and predict the survival of patients with OC. In addition, the differential expression of several key lncRNAs and mRNAs was verified by reverse transcription-quantitative PCR and western blot analysis. The current study presents a novel lncRNA-miRNA-mRNA network, which provides insight into the potential pathogenesis of OC and allows the identification of prognostic biomarkers and treatment strategies for OC.