Detection of critical genes associated with poor prognosis in breast cancer via integrated bioinformatics analyses

Purpose: To explore the potential prognostic differentially expressed genes (DEGs) in breast cancer (BC) via bioinformatic analysis and elucidate possible mechanisms underlying the effects on BC progression. Methods: Three datasets (GSE21422, GSE31192 and GSE42568) were extracted from Gene Expression Omnibus (GEO) information bank. The GEO2R tool and Venn diagram softwares were used for data filtration, GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) analysis method were used to functionally annotate the selected DEGs. Protein-protein interaction (PPI) network of the selected DEGs was visualized by Cytoscape. Lastly, Kaplan-Meier (KM) plotter and Profiling Interactive Analysis (GEPIA) were employed to validate the values of the DEGs. Results: A total of 46 up-regulated and 65 down-regulated DEGs were identified. Of these, up-regulated DEGs were enriched in pathways related to cancer, p53 signaling path-way, ECM-receptor interaction, PI3K-Akt signaling pathway, while down-regulated DEGs were enriched in pathways involved in PPAR signaling pathway, proteoglycans in cancer, focal adhesion. 24 genes were selected from the PPI network analysis by Molecular Complex Detection (MCODE), and 20 vital genes were found to be correlated to poorer overall survival (OS) rates in BC. The prognostic values of these genes were validated by both KM and GEPIA. Finally, the CCNE2, CCNB1 and RRM2 genes were found to be markedly enriched in the p53 signaling pathway through the DAVID analysis. Conclusion: This study revealed that the p53 signaling pathway could be an important pathway in BC progression. The three p53-related genes CCNE2, CCNB1 and RRM2 may represent candidate therapeutic gene targets for the treatment of BC.