Breast cancer (BRCA) is the most common malignant tumor that seriously threatens the health of women worldwide. Senescence has been suggested as a pivotal player in the onset and progression of tumors as well as the process of treatment resistance. However, the role of senescence in BRCA remains unelucidated.
The clinical and transcriptomic data of 2994 patients with BRCA were obtained from The Cancer Genome Atlas and the METABRIC databases. Consensus clustering revealed senescence-associated subtypes of BRCA patients. Functional enrichment analysis explored biological effect of senescence. We then applied weighted gene co-expression network analysis (WGCNA) and LASSO regression to construct a senescence scoring model, Sindex. Survival analysis validated the effectiveness of Sindex to predict the overall survival (OS) of patients with BRCA. A nomogram was constructed by multivariate Cox regression. We used Oncopredict algorithm and real-world data from clinical trials to explore the value of Sindex in predicting response to cancer therapy.
We identified two distinct senescence-associated subtypes, noted low senescence CC1 and high senescence CC2. Survival analysis revealed worse OS associated with high senescence, which was also validated with patient samples from the National Cancer Center in China. Further analysis revealed extensively cell division and suppression of extracellular matrix process, along with lower stromal and immune scores in the high senescence CC2. We then constructed a 37 signature gene scoring model, Sindex, with robust predictive capability in patients with BRCA, especially for long time OS beyond 10 years. We demonstrated that the Sene-high subtype was resistant to CDK inhibitors but sensitive to proteosome inhibitors, and there was no significant difference in paclitaxel chemotherapy and immunotherapy between patients with different senescence statuses.
We reported senescence as a previously uncharacterized hallmark of BRCA that impacts patient outcomes and therapeutic response. Our analysis demonstrated that the Sindex can be used to identify not only patients at different risk levels for the OS but also patients who would benefit from some cancer therapeutic drugs.