Bioinformatics Identification of Prognostic Factors Associated with Breast Cancer

Abstract

Background: Breast cancer (BRCA) remains one of the most common forms of cancer and is the most prominent driver of cancer-related death among women. The mechanistic basis for BRCA, however, remains incompletely understood. In particular, the relationships between driver mutations and signaling pathways in BRCA are poorly characterized, making it difficult to identify reliable clinical biomarkers that can be employed in diagnostic, therapeutic, or prognostic contexts.Methods: First, we downloaded publically available BRCA datasets (GSE45827, GSE42568, and GSE61304) from the Gene Expression Omnibus (GEO) database. We then compared gene expression profiles between tumor and control tissues in these datasets using Venn diagrams and the GEO2R analytical tool. We further explore the functional relevance of BRCA-associated differentially expressed genes (DEGs) via functional and pathway enrichment analyses using the DAVID tool, and we then constructed a protein-protein interaction network incorporating DEGs of interest using the Search Tool for the Retrieval of Interacting Genes (STRING) database. Modules within this PPI network were then identified using Cytoscape, leading to the identification of key candidate genes. The prognostic relevance of these candidate genes was then established through Kaplan-Meier survival analyses and further Gene Expression Profiling Interactive Analysis (GEPIA) validation. Then, key gene-target miRNA regulatory network and transcription factor-key gene regulatory relationships were established using the online miRWalk2.0, TargetScan7.2, miRDB and TRRUST tools. Moreover, four representative key molecules (AURKA, RRM2, BIRC5, and E2F1) were optionally chosen for verification by using quantitative real-time polymerase chain reaction (RT-PCR) and western blot.Results: We identified 85 BRCA-related DEGs across these three datasets. The 31 upregulated DEGs were found to be enriched for pathways and functions including mitotic nuclear division, cell division, G2/M transition of mitotic cell cycle, collagen catabolic process, endodermal cell differentiation, oocyte meiosis, ECM-receptor interactions, and p53 signaling pathway. The 54 downregulated DEGs were, in contrast, enriched in pathways and functions such as lipid metabolic processes, lipid transport, regulation of cardiac muscle contraction by regulation of the release of sequestered calcium ions, positive regulation of cell proliferation, positive regulation of cell-matrix adhesion, tyrosine metabolism, cytochrome P450 drug metabolism, protein digestion and absorption, and PPAR signaling. We were further able to select 16 upregulated candidate genes of interest from our PPI network, and in subsequent Kaplan-Meier analyses we were able to determine that elevated expression of 14 of these genes was associated with a poorer BRCA patient prognosis. We then employed GEPIA to validate these 14 gene candidates, confirming them to all be expressed at elevated levels in BRCA relative to normal tissue controls. In addition, a regulatory network consisting of 9 genes, 10 miRNAs and 3 TFs was constructed, enabling the identification of potential biomarkers of BRCA, including AURKA, RRM2, BIRC5, and E2F1. RT-PCR results suggested that significantly elevated AURKA, RRM2 and BIRC5 mRNAs expressed in the breast cancer cells than in the normal cells. Western blot results shown that E2F1 protein was highly expressed in breast cancer cells compared to normal cells. In conclusion, these candidate molecules may offer insight regarding the underlying pathogenesis of BRCA and highlight a number of potential therapeutic avenues for the treatment of breast cancer patients.