SOX4, PRNP, and FOXO4 are key biomarkers linking PCa, T2DM, and ageing, and the diagnostic model shows robust performance and clinical utility, providing insights into microenvironment-driven mechanisms and personalized treatment strategies for PCa patients with T2DM.
Machine learning analysis of PCa and T2DM transcriptomic data identified three key ageing-related genes: SOX4, PRNP, and FOXO4.
A diagnostic nomogram model based on SOX4, PRNP, and FOXO4 showed high diagnostic accuracy across cohorts.
The three genes showed distinct functional roles in tumor biology and immune context.
Spatial transcriptomics and single-cell RNA sequencing revealed cell-type-specific expression of the key genes and fibroblast-epithelial communication.
Drug sensitivity profiling suggested different therapeutics for high- and low-risk patients.
This study used computer-based analysis of publicly available gene-activity data to look for genes that link prostate cancer, type 2 diabetes, and ageing. Using several machine learning methods, the researchers narrowed the candidates to three genes: SOX4, PRNP, and FOXO4. They built a diagnostic model from these genes that distinguished patients from controls with moderate to high accuracy (AUCs of 0.726 to 0.905) across multiple datasets. The analyses also suggested how each gene might work. SOX4 appeared to help tumors evade the immune system, FOXO4 appeared to act as a tumor suppressor, and PRNP's role seemed to depend on context. Single-cell and spatial analyses indicated that these genes are active in specific cell types and that fibroblasts and epithelial cells in the tumor environment communicate through certain signaling molecules (MIF, LRP1, FZD1). Drug sensitivity predictions pointed to abiraterone for patients classified as high-risk and docetaxel for those classified as low-risk. This research suggests that these three genes could help with early detection and risk stratification in people with prostate cancer and type 2 diabetes. The findings come from bioinformatic analysis of existing datasets, so the biomarkers and drug suggestions would need further laboratory and clinical testing before any use in patient care.
Zhang W, Zhang X, Zhang C, Shi X, Deng Y, Gao Y, et al.. (2026). Identification of ageing-related genes in prostate cancer with type 2 Diabetes using machine learning: a multi-omics study focusing on SOX4, PRNP, and FOXO4.. Artificial cells, nanomedicine, and biotechnology. https://doi.org/10.1080/21691401.2026.2737622