Identification and validation of an E2F-related gene signature for predicting recurrence-free survival in human prostate cancer

Background: It is well-established that biochemical recurrence is detrimental to prostate cancer (PCa). In the present study, we explored the mechanisms underlying PCa progression. Methods: Five cohorts from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus databases were used to perform gene set variation analysis (GSVA) between nonrecurrent and recurrent PCa patients. We obtained the intersection of pathway enrichment results and extracted the corresponding gene list. LASSO Cox regression analysis was used to identify recurrence-free survival (RFS)-related significant genes and establish an RFS prediction gene signature and nomogram. MTT and colony formation assays were conducted to validate our findings. Results: The E2F signaling pathway was activated in recurrent PCa patients compared to nonrecurrent patients. We established an E2F-related gene signature for RFS prediction based on the four identified E2F-related genes (CDKN2C, CDKN3, RACGAP1, and RRM2) using LASSO Cox regression in the Memorial Sloan Kettering Cancer Center (MSKCC) cohort. The risk score of each patient in MSKCC was calculated based on the expression levels of CDKN2C, CDKN3, RACGAP1, and RRM2. PCa patients with low-risk scores exhibited higher RFS than those with high-risk scores. Receiver operating characteristic (ROC) curve analysis validated the good performance and prognostic accuracy of the E2F-related gene signature, which was validated in the TCGA-prostate adenocarcinoma (TCGA-PRAD) cohort. Compared to patients with low Gleason scores and early T stages, PCa patients with high Gleason scores and advanced T stages had high-risk scores. Moreover, the E2F-related gene signature-based nomogram yielded good performance in RFS prediction. Functional experiments further confirmed these results. Conclusions: The E2F signaling pathway is associated with biochemical recurrence in PCa. Our established E2F-related gene signature and nomogram yielded good accuracy in predicting the biochemical recurrence in PCa.