To identify the valid biomarkers of diagnosis, prognosis and treatment in breast cancer, we have analyzed four datasets via bioinformatical methods in the present work. A total of 53 common DEGs were found by comparing breast cancer samples with normal samples in the four datasets. Then, GO enrichment and KEGG pathway analysis were performed on the common DEGs to investigate the different biological information related to these genes. It was observed that up-regulated genes were mainly involved in nuclear division, organelle fission and mitotic nuclear division related biological processes, and down-regulated genes were mainly involved in regulation of cellular response to growth factor stimulus related biological process. Meanwhile, the result proved that the up-regulated DEGs were involved in the breast cancer-related pathways such as P53 signaling pathway , platinum drug resistance  and bladder cancer . Next, we selected 16 genes from the densely regions of PPI networks, and found that 15 of them were associated with poor prognosis of breast cancer, which the expression level in breast cancer was significantly higher than that in normal tissue samples. For further study, we re-constructed the PPI networks with the common DEGs which were screened from the individual comparisons of three grades breast cancer with normal samples. The number of common DEGs of grade 3 was more increasingly than other grades, and the poor prognosis genes in the above research were reappeared in densely regions of grade PPI networks. It was conferred that the genes in grade 3 lead the dominant role in breast cancer tumorigenesis and progression. Therefore, we chose these 15 poor prognosis genes for the subsequent study. After gene set enrichment analysis and KEGG analysis, 5 genes (AURKA, BIRC5, CCNB2, CDK1 and TOP2A) were identified to be the potential targets for diagnosis and prognosis of breast cancer. Results showed that the cancer-related signaling pathways regulated by these 5 genes were also involved in grade 3 breast cancer. Finally, we explored the interaction of genes and chemotherapeutics dugs and found these 5 genes were strongly correlated with anthracyclines and cyclophosphamide which has been proven to be the effective treatment drugs of breast cancer [22, 23].
Because of the high expression in many cancers, increasingly researches has been focused on AURKA as a target for cancer therapy . Clinical trials of AURKA inhibitors are currently undergoing and have shown the effective anti-tumor activity in multiple cancers , but the studies on breast cancer have progressed slowly. AURKA has been well proved involved in the cell cycle and cell division , it functions as a key role in mitotic entry, centrosome functions and bipolar spindle assembly especially regulates the G2/M transition . Evidences indicated that AURKA is associated with higher tumor stage and cancer cell migration . Additionally, the breast cancer-related studies have reported that AURKA induces different carcinogenic properties in malignant cells by regulating stem cell function . In a word, the research prospects of AURKA inhibitors are promising in the breast cancer therapy.
Early studies observed that BIRC5 is involved in mitosis by forming the CPC (chromosomal passenger complex) and regulating microtubule dynamics and nucleation [30, 31]. Several researches demonstrated that BIRC5 has been considered as a prognostic biomarker for many cancers in that its closely correlation with chemotherapy resistance, neoplasm metastasis and tumor recurrence [32–34]. To date, the role of BIRC5 as a poor prognostic factor for breast cancer has been confirmed in the past studies [35, 36]. Additionally, in the field of tumor targeted therapy, BIRC5 has also been received much attention . Research data suggested that the BIRC5 inhibitor could increase the antineoplastic activity of paclitaxel in breast cancer treatment . Taken together, BIRC5 has a bright perspective as biomarker for prognostic and targeted therapy of breast cancer in the future.
In the regulation of cell cycle phase, the complex formed by CCNB2 binds to CDK1 acts a key role in G2/M transition [39, 40].Besides, cell cycle disorder in the process of tumorigenesis is associated with dysfunction of CCNB2 and CDK1 . Previous reports demonstrated that CCNB2 not only has negative impact on the survival of breast cancer patients , but also could be correlated with the tumor recurrence . Also, CDK1 has been proved to be a predictor of clinical outcome for breast cancer and related with the unfavorable clinicopathologic features . For decades, numerous clinical studies have been performed on CDK1 inhibitors, but the results are disappointed . The reason could be the CDK1 targeted drugs cause the cell toxicity, thus restricting the therapeutic window . However, the role of CDK1 in the breast cancer treatment is still worth exploring. Although the current results of CCNB2 and CDK1 in breast cancer therapy research are dissatisfied, they could have well prospects as factors for diagnosis, prognostic and treatment in future work.
It was reported that TOP2A is typically highly expressed in rapidly proliferating cells and functioned in DNA replication and cell division . Recent study showed that TOP2A was closely related with tumor size differentiation and lymph node metastasis in triple-negative breast cancer patients . While in estrogen receptor positive breast cancer patients, TOP2A could be a evaluation factor of prognostic and recurrence . Previously, numerous researches have been conducted on the correlation between TOP2A and anthracycline-based chemotherapy treatment [50, 51], but the predication effectiveness is uncertain and follow-up studies are required . As the close location with the HER2 gene on chromosome , TOP2A should be well studied for prognosis an therapy in breast cancer.
Since the safety and efficacious of anthracyclines and cyclophosphamide has been shown in chemotherapy for breast cancer [54, 55], we performed the gene-drug interaction analysis for further study. Indeed, our work demonstrated that these 5 genes were interacted with the drugs through different ways. These data provide evidence for future research on the influence of these 5 genes on breast cancer treatment.
In the present study, detailed observation showed that the common genes of grade 3 are the most numerous and closely linked in breast cancer. As the heterogeneity of cancer, we conferred that genes in grade 3 were dominated the progression of breast cancer. Thus, the 5 core genes and the cancer-related pathways might be the key factors in tumorigenesis and development of breast cancer. In brief, we hope these data could be helpful for future study on the pathogenesis, prognosis and treatment of breast cancer.