MicroRNA and transcription factor co-regulatory network analysis: a combined bioinformatics and molecular strategy for uncovering genetic hub regulatory elements in multiple sclerosis pathogenesis

Introduction: As the main regulators of gene expression, microRNAs (miRNAs) and transcription factors (TFs) play key roles in complex multifactorial diseases like Multiple Sclerosis (MS). Recently, miRNA/TF-based feed-forward loops (FFLs) have been identified as key elements of biological network motifs in MS. Objectives: Starting from preliminary results obtained by a High-Throughput Next-Generation Sequencing (HT-NGS) approach on an independent MS patients cohort, we aimed to search for miRNA-TF co-regulatory networks involved as hub genetic elements in MS pathogenesis. Methods: The study was performed on peripheral blood samples belonging to 58 MS patients and 20 healthy controls (HCs).To validate the differential expressed (DE) miRNAs resulted from HT-NGS analysis, total RNA extraction was followed by microfluidic qPCR. The subsequent analysis included: miRNA-target gene interaction evaluation using experimentally validated miRNA-target interaction databases (DIANA-Tarbase and miRTarBase) and the output of 5 algorithms (miRanda, DIANA-microT-CDS, RNA22, miRDB and TargetScan) for target genes prediction; miRNA-target experimentally validation using dual luciferase reporter assay; TF-miRNA co-regulatory network identification using miRNA-TF interaction tools (Harmonizome and TransmiR) and a gene-TF interaction database (TRRUST); pathway enrichment analysis using DAVID tool. Results: Microfluidic qPCR analysis revealed 6 significantly DE miRNAs (miR-320a, miR-125a-5p, miR-652-3p, miR-185-5p, miR-942-5p and miR-25-3p) in MS patients compared to HCs. The integrated analysis of miRNAs and computationally predicted mRNAs uncovered 638 miRNA-target pairs validated by reporter gene assays and/or predicted by at least 4 of the 5 interaction tools. Luciferase reporter assays indicated that miR-125a-5p was able to target and modulate DIP2A, ADD2 and E2F2 expression, thus confirming that the in silico predicted target were effectively subjected to miRNA-mediated post-transcriptional modulation. In addition, 409 TF-miRNA interactions were identified, including 198 miRNA-TF FFLs. Furthermore, the enrichment analysis categorized in functional pathways revealed significant representations of networks mostly related to neurotrophin signalling pathway, ErbB signalling pathway and axon guidance. Conclusions: These results uncovered several hub regulatory elements allowing to shed further lights in the gene regulation mechanisms that may underlie the pathogenesis of MS.