Abstract
Background
Prostate cancer is one of the most common malignancies among men worldwide. Anoikis is a form of programmed cell death that is potentially negatively correlated with tumor progression; however, its relationship with prostate cancer remains inconclusive.
Methods
The transcriptomic and clinical data for this study were obtained from the TCGA and GEO databases. The prediction model was established using univariate Cox, multivariate Cox, and LASSO regression. Receiver operating characteristic (ROC) curves determined the predictive performance, and the GEO database was used for external validation. Patients were stratified into different risk groups, and their prognoses were compared using Kaplan-Meier analysis. We also analyzed immune cell infiltration and sensitivity to immunotherapeutic drugs in prostate cancer patients. The BUB1 gene was selected for in vitro experimental validation.
Results
We constructed a prognostic risk prediction model using four ARGs: BUB1, PTGS2, RAC3, and IRX1. Patients in the high-risk group had worse overall survival than those in the low-risk group, with significant differences in immune cell infiltration, immune checkpoint expression, and sensitivity to immunotherapeutic drugs. Using NMF, we categorized TCGA prostate cancer patients into two subgroups, with cluster2 having better prognoses. Gene expression and immune cell infiltration were compared between the subgroups. Knocking down the BUB1 gene in PC3 and C4-2 cell lines reduced prostate cancer cell proliferation and invasion and altered EMT-related protein expression.
Conclusion
After external validation, our study shows that the ARG-based predictive model accurately forecasts prostate cancer prognosis. In vitro experiments revealed that the BUB1 gene significantly affects prostate cancer cell proliferation, invasion, and the expression of specific EMT-related proteins. Thus, BUB1 is a potential therapeutic target.