Potential drug targets for prostate cancer ： A mendelian randomization study and application for target-derived drug design

Background The advent of targeted cancer therapies has led to a decline in prostate cancer (PCa) incidence and mortality rates. Nevertheless, challenges persist due to the long-term single-agent therapeutic insensitivity and resistance encountered in PCa treatment. Therefore, there is an urgent need for novel drug targets to address these challenges in PCa therapy. Method We analyzed 731 plasma proteins and PCa summary GWAS data from Prostate Cancer Association Group to investigate cancer associated genomic alterations ( control: case = 61 106: 79 148 ). Cis-acting Mendelian randomization and Bayesian analysis was applied to reveal the causality between protein and PCa. Additionally, protein-protein interaction (PPI) was performed to discover the potential coactions between identified target proteins and established drug targets approved by FDA for the treatment of PCa. Furthermore, we utilized Alpha Fold 2 to predict the 3D complex structure between identified proteins and established drug targets. Finally, these findings were validated using data from UK Biobank and the European Bioinfomatics Institute, and six promising target proteins were categorized into three tiers. Results Six potential causal proteins including MSMB, IGF2R, KDELC2, TNFRSF10B, GSTP1, and SPINT2 were discovered through drug target Mendelian randomization analysis. Among them, MSMB (Odds ratio (OR) = 0.81; 95% confidence interval (CI) : 0.80–0.82; P  = 2.52×10 ^− 148 ), IGF2R (OR = 0.92; 95% CI: 0.90–0.94; P  = 4.57×10 ^− 10) , KDELC2 (OR = 0.89; 95% CI: 0.86–0.93; P  = 1.89×10 ^− 8 ), TNFRSF10B (OR = 0.74; 95% CI: 0.65–0.83; P  = 2.41×10 ^− 7 ), and GSTP1 (OR = 0.82; 95% CI: 0.75–0.90; P  = 4.22×10 ^− 5 ) were inversely associated with PCa risk, and upregulate level of SPINT2 (OR = 1.05; 95% CI: 1.03–1.05; P  = 1.49×10 ^− 6 ) increased PCa risk. None of six proteins had reverse causality. MSMB and KDELC2 shared the same variant with PCa by co-localization analysis (PPH4 > 0.8). During external validation, five proteins were replicated in at least one dataset except IGF2R. Conclusions Our study has highlighted that a constellation of plasma proteins including MSMB, KDELC2, GSTP1, and TNFRSF10B have been identified as potential drug targets for PCa, which might provide valuable insights for the rational design of novel drugs in PCa therapy.