Regulatory Network Analysis Reveals Prognosis-Related CircRNAs in Pancreatic Cancer Cells

Abstract

Objective: To construct a key prognosis-related regulatory network and to identify the epigenomic alterations of RNAs that have crucial functions in cancer pathogenesis.Methods: RNA expression profiles of miRNAs, mRNAs and circRNAs were extracted from the NCBI GEO and EBI ArrayExpress databases. The identification of differentially expressed genes was performed by R language. Databases such as starBase and TCGA were used for annotation, visualization, and integrated discovery. GO and KEGG pathway enrichment analyses were performed by DAVID 6.8. The key prognosis-related regulatory network was constructed with Cytoscape software. The involved circRNAs were verified by quantitative real-time PCR (qRT-PCR) in five pancreatic cancer cell lines. Proliferation ability was assessed by the CCK-8 assay, apoptosis was detected by flow cytometry, and migratory and invasive abilities were assessed by Transwell assays.Results: In total, 798 differentially expressed genes, consisting of 85 circRNAs, 19 miRNAs and 694 mRNAs, were obtained. A miRNA interaction network was predicted and used to construct a prognosis-related regulatory network. GO annotation and KEGG pathway analysis revealed important biological processes and pathways in pancreatic cancers. The key regulatory factors in the network were miR-146b-5p and miR-152. Eight circRNAs included in the network were verified, and hsa_circ_0006502 was downregulated in the five tested pancreatic cancer cell lines and functioned as a tumor suppressor, as predicted.Conclusion: In conclusion, a key prognosis-related regulatory network in pancreatic cancers was constructed through bioinformatics analysis of public RNA databases, and miR-146b-5p and miR-152 were the key regulatory factors. Hsa_circ_0006502 was identified as a tumor suppressor that interacted closely with miR-146b-5p and might serve as a potential therapeutic agent.