Systematic analysis of prognostic miRNAs and isomiRs in prostate cancer

Abstract

There are no reliable prognostic indicators to distinguish between indolent and aggressive prostate cancer (PCa). Consequently, 42–66% of patients with indolent PCa are over-treated. Additionally, 15-45% of patients treated with radical prostatectomy (RP) experience biochemical recurrence (BCR) within 5-years, highlighting an urgent need for reliable prognostic biomarkers.

MiRNAs (miRs) and isomiRs (miR isoforms) are non-coding regulatory RNAs that hold ideal biomarker properties such as detection in circulation, tissue and tumour specific expression profiles, and correlation with PCa development and progression. I hypothesised that miR species (canonical miRs and isomiRs) can be utilised as biomarkers for reliable PCa prognostication.

A novel database of prognostic PCa miRs was built by performing a systematic review of relevant publications in the PubMed database. MiRs significantly associated with BCR were also identified following a meta-analysis of six datasets. MiR-148a-3p and miR-582-4p were identified as potential biomarker candidates as they were consistently prognostic in both the review and meta-analysis.

The ability of miR species to predict BCR post-RP was tested with elastic net regularisation models using The Cancer Genome Atlas PCa dataset (recurrent=61, non-recurrent=330). Models based on a combination of isomiRs and clinical markers achieved marginally greater predictive power (AUC=0.795) than the model solely based on clinical markers (AUC=0.748), demonstrating that isomiRs could contribute additional prognostic value to the clinical markers currently used.

The mechanism by which miR-27a-3p, a PCa-specific putative oncomiR, promotes tumour growth was investigated using RNA-seq data from LNCaP tumour xenograft models treated with a miR-27a-3p inhibitor (n=3) and control (n=3). 11 significantly dysregulated genes involved in apoptosis and oncogenic signalling were identified as likely mir-27a-3p targets.

This study has not only furthered our understanding of the importance of miRs in PCa, but also identified potential prognostic miR biomarkers and showed the inclusion of miR species increases the utility of current markers.