Semen sEV tRF-Based Models Increase Non-Invasive Prediction Accuracy of Clinically Significant Prostate Cancer among Patients with Moderately Altered PSA Levels

Abstract

PSA screening has led to an over-diagnosis of prostate cancer (PCa) and unnecessary biopsies of benign conditions due to its low cancer specificity. Consequently, more accurate, preferentially non-invasive, tests are needed. We aim to evaluate the potential of semen sEV (small extracellular vesicles) tsRNAs (tRNA-derived small RNAs) as PCa indicators. Initially, following a literature review in the OncotRF database and high-throughput small RNA-sequencing studies in PCa tissue together with the sncRNA profile in semen sEVs, we selected four candidate 5 ' tRF tsRNAs for validation as PCa biomarkers. RT-qPCR analysis in semen sEVs from men with moderately elevated serum PSA levels successfully shows that the differential expression of the four tRFs between PCa and healthy control groups can be detected in a non-invasive manner. The combined model incorporating PSA and specific tRFs (5 '-tRNA-Glu-TTC-9-1_L30 and 5 '-tRNA-Val-CAC-3-1_L30) achieved high predictive accuracy in identifying samples with a Gleason score >= 7 and staging disease beyond IIA, supporting that the 5 ' tRF fingerprint in semen sEV can improve the PSA predictive value to discriminate between malignant and indolent prostate conditions. The in silico study allowed us to map target genes for the four 5 ' tRFs possibly involved in PCa. Our findings highlight the synergistic use of multiple biomarkers as an efficient approach to improve PCa screening and prognosis.