Validation of the four-miRNA biomarker panel MiCaP for prediction of long-term prostate cancer outcome

Abstract

Improved prostate cancer prognostic biomarkers are urgently needed. We previously identified the four-miRNA prognostic biomarker panel MiCaP ((miR-23a-3p × miR-10b-5p)/(miR-133a-3p × miR-374b-5p)) for prediction of biochemical recurrence (BCR) after radical prostatectomy (RP). Here, we identified an optimal numerical cut-off for MiCaP dichotomisation using a training cohort of 475 RP patients and tested this in an independent cohort of 281 RP patients (PCA281). Kaplan-Meier, uni- and multivariate Cox regression analyses were conducted for multiple endpoints: BCR, metastatic-(mPC) and castration-resistant prostate cancer (CRPC), prostate cancer-specific (PCSS) and overall survival (OS). Functional effects of the four MiCaP miRNAs were assessed by overexpression and inhibition experiments in prostate cancer cell lines. We found the numerical value 5.709 optimal for MiCaP dichotomisation. This was independently validated in PCA281, where a high MiCaP score significantly [and independent of the Cancer of the Prostate Risk Assessment Postsurgical (CAPRA-S) score] predicted BCR, progression to mPC and CRPC, and PCSS, but not OS. Harrell's C-index increased upon addition of MiCaP to CAPRA-S for all endpoints. Inhibition of miR-23a-3p and miR-10b-5p, and overexpression of miR-133a-3p and miR-374b-5p significantly reduced cell survival. Our results may promote future implementation of a MiCaP-based test for improved prostate cancer risk stratification.

Figure 1

Kaplan–Meier analyses in PCA475. Kaplan–Meier analysis of patients stratified by MiCaP score (low vs. high) relative to three different end-points [(a) biochemical recurrence (BCR), (b) prostate cancer-specific survival (PCSS), and (c) overall survival (OS)]. p values from log-rank test.

Figure 2

Kaplan–Meier analyses in PCA281. Kaplan–Meier analysis of patients stratified by MiCaP score (low vs. high) relative to five different end-points. (a) Biochemical recurrence (BCR), (b) metastatic prostate cancer (mPC), (c) castration-resistant prostate cancer (CRPC), (d) overall survival (OS), and (e) prostate cancer-specific survival (PCSS). p values from log-rank test.

Figure 3

Overexpression and inhibition studies of miRNAs. Functional studies of miRNAs in prostate cancer cell lines. (a) Inhibitory effect on prostate cancer cell viability by single miRNA mimics and inhibitors in PC3 and DU145. Each mimic or inhibitor was compared to the corresponding control mimic or inhibitor in the same cell line. Results from alamarBlue viability assay (72 h post-transfection), plotted as mean ± SE of three independent experiments performed in triplicate. (b) Significant inhibitory effect on real-time proliferation by miR-374b-5p mimic transfections in PC3 and DU145 using the xCELLigence instrument. Results from one representative experiment performed in triplicate (three experiments in total) are plotted as mean ± SD for each time point. Student’s two-sided t-test, *p < 0.05, **p < 0.01, ***p < 0.001.