Objective To investigate the predictive value of radiation-regulated cell cycle-related signature genes in prostate cancer (PCa) and to develop a corresponding prognostic model.

Methods Single-cell RNA sequencing and gene set enrichment analysis were performed on PCa cell lines LNCaP and PC3 with the 10×Genomics Chromium platform. Cells were exposed to single-dose irradiation with ¹³⁷Cs γ-rays at a dose rate of 0.883 Gy/min. The administered doses were 0, 0.5, 2, and 8 Gy (0 Gy served as the control group). In addition, molecular subtyping was conducted using a nonnegative matrix factorization (NMF) clustering algorithm based on The Cancer Genome Atlas-prostate adenocarcinoma (TCGA-PRAD). Subsequently, a least absolute shrinkage and selection operator (LASSO)-Cox regression model was used in the construction of a cell cycle-related prognostic model for radiation-sensitive PCa. Patients were assigned to low-risk (n=218) and high-risk (n=217) groups according to the median threshold of risk scores in TCGA-PRAD. The predictive efficacy of the model for biochemical recurrence-free survival in TCGA-PRAD was assessed using Kaplan-Meier survival and time-dependent receiver operating characteristic curves. The final model was validated using an independent external validation dataset from the Memorial Sloan-Kettering Cancer Center (MSKCC)-PCa cohort, and identical performance metrics were calculated to test the model's generalizability in clinical practice. Immune function and immune cell infiltration in tumor samples from TCGA-PRAD were systematically assessed through single-sample gene set enrichment analysis. Finally, the half-maximal inhibitory concentration of drugs in high-risk and low-risk tumors was calculated using oncoPredict for the identification of risk-specific sensitive drugs, and the molecular mechanisms and therapeutic targets of radiation regulation in PCa were further explored. Paired sample t-test and Wilcoxon rank-sum test were used for between-group comparisons.

Results Single-cell RNA sequencing revealed that the LNCaP cell lines exhibited multicluster distribution after 8 Gy ¹³⁷Cs γ-ray irradiation, and notable enrichment was observed in core cell cycle-related pathways, including cell cycle, regulation of cell cycle process, and S and G2/M phase transition (all P<0.01). By contrast, PC3 cell lines showed a weak cell cycle response but stronger radioresistance. After the integration of differentially expressed genes in tumor and normal tissues, radiation-regulated genes, and cell cycle-related genes from TCGA-PRAD, a total of 68 intersecting genes were obtained, of which 63 genes had a hazard ratio>1. Consensus clustering of TCGA-PRAD was performed using the 68 intersecting genes, and two tumor subtypes were identified by the NMF clustering algorithm. LASSO-Cox regression analysis established a PCa prognostic model consisting of seven core genes (extra spindlepole bodies like 1, cell division cycle associated 5, Polo-like kinase 1, aurora kinase B, forkhead box M1, kinesin family member (KIF) 18B, and KIF2C). The area under the curve (AUC) of the risk score for predicting 3 year biochemical recurrence-free survival was 0.691, which outperformed conventional indicators, including the Gleason score (AUC=0.661), M stage (AUC=0.511), and N stage (AUC=0.527). Validation in the MSKCC-PCa external dataset confirmed the favorable stability and generalizability of the model (AUC=0.723 for 3 year biochemical recurrence-free survival). Single-cell RNA sequencing further verified that ¹³⁷Cs γ-rays downregulated the expression of high-risk prognostic genes (t=2.20–30.88, all P<0.001). Immune cell infiltration analysis indicated limited benefit from immune checkpoint blockade therapy in the high-risk group. Meanwhile, oncoPredict prediction demonstrated that high-risk tumor samples were more sensitive to inhibitors targeting the protein kinase B/Wnt signaling pathways (all P<0.001). This prognostic model can serve as a basis for decision-making in integrating radiotherapy, immunotherapy, and targeted therapy.

Conclusion The PCa prognostic model based on the radiation-regulated cell cycle can predict the proliferative activity of PCa cells, providing a potential decision-making for individualized interventions combining radiotherapy, immunotherapy, and targeted therapy.