Identification of novel biomarkers in breast cancer via integrated bioinformatics analysis and experimental validation

Breast cancer (BC), an extremely aggressive malignant tumor, causes a large number of deaths worldwide. In this study, we pooled profile datasets from three cohorts to illuminate the underlying key genes and pathways of BC. Expression profiles GSE42568, GSE45827, and GSE124646, including 244 BC tissues and 28 normal breast tissues, were integrated and analyzed. Differentially expressed genes (DEGs) were screened out based on these three datasets. Functional analysis including Gene Ontology (GO) and Kyoto Encyclopedia of Gene and Genome (KEGG) pathway were performed using The Database for Annotation, Visualization and Integrated Discovery (DAVID). Moreover, Cytoscape with Search Tool for the Retrieval of Interacting Genes (STRING) and Molecular Complex Detection (MCODE) plugin were utilized to visualize protein protein interaction (PPI) of these DEGs. The module with the highest connectivity of gene interactions was selected for further analysis. All of these hub genes had a significantly worse prognosis in BC by survival analysis. Additionally, four genes (CDK1, CDC20, AURKA, and MCM4) dramatically were enriched in oocyte meiosis and cell cycle pathways through re-analysis of DAVID. Moreover, the mRNA and protein levels of CDK1, CDC20, AURKA, and MCM4 were significantly increased in BC patients. In addition, knockdown of CDK1 and CDC20 by small interfering RNA remarkably suppressed cell migration and invasion in MCF-7 and MDA-MB-231 cells. In conclusion, our results suggested that CDK1, CDC20, AURKA, and MCM4 were reliable biomarkers of BC via bioinformatics analysis and experimental validation and may act as prospective targets for BC diagnosis and treatment.