Identification of invasive key genes in breast cancer by bioinformatics analysis.

Purpose Of all breast cancers, triple-negative and HER-2 positive are the most aggressive breast cancer subtypes with a high risk of recurrence and worse prognosis. The study's purpose was to further assess the molecular mechanisms underlying aggression of breast cancer. Methods The microarray gene expression datasets of GSE29431 and GSE53752 were obtained from the GEO (Gene Expression Omnibus) database, which include HER-2 positive breast cancer, triple-negative breast cancer (TNBC) and normal breast tissue samples. Differentially expressed genes (DEGs) were determined using the LIMMA package of R software and subsequently functional enrichment analysis were performed by the ClusterProfiler package in the R platform. The STRING database was used to construct a protein-protein interaction (PPI) network. The most significant module and key genes were identified by Cytoscape software. Utilizing the Kaplan-Meier plotter and UALCAN database, we defined the key genes associated with prognotic values and molecular subtypes as invasive genes. Results In total, 428 common DEGs were identified, including 143 upregulated and 285 downregulated. GO and KEGG pathway enrichment analysis indicated that the upregulated genes were associated with mitotic nuclear division and cell cycle, whereas the downregulated genes were significantly associated with response to peptide and PPAR signaling pathway, respectively. A PPI network with 57 nodes and 335 edges was established, from which one most significant module was identified. Moreover, 12 key genes selected from the module with high degree centrality more than 21 were highly associated with high clinical aggressiveness and worse overall survival rate. Conclusions Our studies could enhance the understanding of the molecular mechanism of breast cancer aggressiveness, and the identification of invasive key genes promoted the individualized and comprehensive treatment.