RESUMO
PURPOSE: The analysis of exosome/microvesicle (extracellular vesicles (EVs)) and the RNA packaged within them (exoRNA) has the potential to provide a non-invasive platform to detect and monitor disease related gene expression potentially in lieu of more invasive procedures such as biopsy. However, few studies have tested the diagnostic potential of EV analysis in humans. EXPERIMENTAL DESIGN: The ability of EV analysis to accurately reflect prostate tissue mRNA expression was examined by comparing urinary EV TMPRSS2:ERG exoRNA from pre-radical prostatectomy (RP) patients versus corresponding RP tissue in 21 patients. To examine the differential expression of TMPRSS2:ERG across patient groups a random urine sample was taken without prostate massage from a cohort of 207 men including prostate biopsy negative (Bx Neg, n = 39), prostate biopsy positive (Bx Pos, n = 47), post-radical prostatectomy (post-RP, n = 37), un-biopsied healthy age-matched men (No Bx, n = 44), and young male controls (Cont, n = 40). The use of EVs was also examined as a potential platform to non-invasively differentiate Bx Pos versus Bx Neg patients via the detection of known prostate cancer genes TMPRSS2:ERG, BIRC5, ERG, PCA3 and TMPRSS2. RESULTS: In this technical pilot study urinary EVs had a sensitivity: 81% (13/16), specificity: 80% (4/5) and an overall accuracy: 81% (17/21) for non-invasive detection of TMPRSS2:ERG versus RP tissue. The rate of TMPRSS2:ERG exoRNA detection was found to increase with age and the expression level correlated with Bx Pos status. Receiver operator characteristic analyses demonstrated that various cancer-related genes could differentiate Bx Pos from Bx Neg patients using exoRNA isolated from urinary EVs: BIRC5 (AUC 0.674 (CI:0.560-0.788), ERG (AUC 0.785 (CI:0.680-0.890), PCA3 (AUC 0.681 (CI:0.567-0.795), TMPRSS2:ERG (AUC 0.744 (CI:0.600-0.888), and TMPRSS2 (AUC 0.637 (CI:0.519-0.754). CONCLUSION: This pilot study suggests that urinary EVs have the potential to be used as a platform to non-invasively differentiate patients with prostate cancer with very good accuracy. Larger studies are needed to confirm the potential for clinical utility.