Abstract 2543: Concurrent profiling of canonical and modified miRNAomes from TCGA and TARGET cohorts leads to enhanced resolution in cancer

Abstract

Abstract MicroRNAs (miRs), a class of regulatory single-stranded small non-coding RNAs, are frequently deregulated in cancer and have been proposed as diagnostic and prognostic biomarkers in this fatal disease. Recent improvements in high-throughput sequencing (HTS) technologies have allowed the discovery of post-transcriptional modification phenomena involving miRs (e.g., isomiRs, A-to-I editing sites) able to alter the miR canonical sequence and thus their function. However, albeit earlier studies have proposed the exploitation of modified miRs as potential cancer biomarkers, their comprehensive and concurrent profiling, in cancer, still lacks. In light of this, we have applied the miRge2 pipeline (PubmedID:30153801) to more than 12K samples across 38 different cancer types from TCGA and TARGET, aiming at profiling the expression of all detectable canonical and modified miR isoforms in cancer. We identified over 29K unique miR isoforms (21 novel ones) expressed in at least one tumor type, whereby about 85% contains at least one modification (e.g., SNPs, RNA editing). According to our results, taking into account both canonical and modified miR isoforms throughout the analysis leads to an increase of about 20-fold in the identification of potential diagnostic and prognostic miRs with respect to conventional methods, which exclusively consider canonical miRs (miRBase v22). Then, leveraging on an independent cohort of 38 colorectal cancers and 21 adjacent healthy tissues (EGAS00001001127), we found that ~600 significantly deregulated miR isoforms were in common with the ones derived from the TCGA COAD dataset. Focusing on two miR isoforms, the canonical miR-151a-3p and its isoform (miR-151a-3p_-2|-2) with two left-shifted residues in both 5' and 3' end, which we found to be up-regulated in both cohorts, we performed a target prediction coupled with a miR/protein anti-correlation analysis. Interestingly, we found that the two miRs have a distinct set of targets. Finally, employing an Ingenuity(R) Pathway Analysis on their predicted targets, we observed that miR-151a-3p_-2|-2, in contrast to the canonical counterpart, is significantly involved in oncogenic pathways, such as PAK and CXCR4 signaling. This suggests that the final tumor phenotype could be substantially affected by modified miR isoforms. Here, we present one of the most in-depth and extensive miRNome profiling in cancer performed to date, revealing one magnitude of novel potential diagnostic and prognostic cancer biomarkers and enlightening the contribution of miR isoforms to the wider miRNome in cancer pathogenesis. Citation Format: Rosario Distefano, Luisa Tomasello, Gioacchino P. Marceca, Marina Bagnoli, Alessandro Lagana', Mario Acunzo, Chi Song, Pierluigi Gasparini, Giovanni Nigita, Carlo M. Croce. Concurrent profiling of canonical and modified miRNAomes from TCGA and TARGET cohorts leads to enhanced resolution in cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2543.