Identification of a novel senescence-associated signature to predict biochemical recurrence and immune microenvironment for prostate cancer.

Background: Prostate cancer (PCa) is an age-associated malignancy with high morbidity and mortality rate, posing a severe threat to public health. Cellular senescence, a specialized cell cycle arrest form, results in the secretion of various inflammatory mediators. In recent studies, senescence has shown an essential role in tumorigenesis and tumor development, yet the extensive effects of senescence in PCa have not been systematically investigated. Here, we aimed to develop a feasible senescence-associated prognosis model for early identification and appropriate management in patients with PCa.

Method: The RNA sequence results and clinical information available from The Cancer Genome Atlas (TCGA) and a list of experimentally validated senescence-related genes (SRGs) from the CellAge database were first obtained. Then, a senescence-risk signature related with prognosis was constructed using univariate Cox and LASSO regression analysis. We calculated the risk score of each patient and divided them into high-risk and low-risk groups in terms of the median value. Furthermore, two datasets (GSE70770 and GSE46602) were used to assess the effects of the risk model. A nomogram was built by integrating the risk score and clinical characteristics, which was further verified using ROC curves and calibrations. Finally, we compared the differences in the tumor microenvironment (TME) landscape, drug susceptibility, and the functional enrichment among the different risk groups.

Results: We established a unique prognostic signature in PCa patients based on eight SRGs, including CENPA, ADCK5, FOXM1, TFAP4, MAPK, LGALS3, BAG3, and NOX4, and validated well prognosis-predictive power in independent datasets. The risk model was associated with age and TNM staging, and the calibration chart presented a high consistency in nomogram prediction. Additionally, the prognostic signature could serve as an independent prediction factor due to its high accuracy. Notably, we found that the risk score was positively associated with tumor mutation burden (TMB) and immune checkpoint, whereas negatively correlated with tumor immune dysfunction and exclusion (TIDE), suggesting that these patients with risk scores were more sensitive to immunotherapy. Drug susceptibility analysis revealed differences in the responses to general drugs (docetaxel, cyclophosphamide, 5-Fluorouracil, cisplatin, paclitaxel, and vincristine) were yielded between the two risk groups.

Conclusion: Identifying the SRG-score signature may become a promising method for predicting the prognosis of patients with PCa and tailoring appropriate treatment strategies.