Abstract
Prostate cancer (PCa) diagnosis is hampered by the limited specificity of current methods, necessitating more reliable biomarkers. To identify causal protein biomarkers and therapeutic targets in humans, we conducted a proteome-wide Mendelian randomization (MR) study. We first performed a meta-analysis of two independent genome-wide association studies, including 94,397 individuals with PCa and 192,372 controls, which identified five possible susceptibility loci (JAZF1, PDILM5, WDPCP, EEFSEC, TNS3) for PCa. Subsequently, MR and colocalization analyses were performed using genetic instruments for 4907 plasma proteins from deCODE Genetics (N=35,559) and 2940 plasma proteins from UK Biobank Pharma Proteomics Project (UKB-PPP) (N=54,219). Among 3722 human proteins analyzed, 193 were associated with PCa risk, with 20 high-risk proteins (including KLK3) validated across both cohorts. Functional annotation implicated immune and inflammatory responses and cell-cell interaction pathways. Druggability analyses nominated several potential drug targets for PCa, such as HSPB1, RRM2B, and PSCA. Our findings reveal novel risk loci and candidate protein biomarkers, providing new etiological insights and potential avenues for PCa early detection and therapy.