MicroRNA/mRNA interactions underlying colorectal cancer molecular subtypes

Colorectal cancer (CRC) molecular subtypes have been recently identified by gene
expression profiling. To search for microRNAs potentially driving the subtypes, we
designed an analytical pipeline, microRNA Master Regulator Analysis (MMRA). As
input, MMRA requires a paired microRNA/mRNA expression dataset, with samples
subdivided in two or more subgroups, and gene expression signatures specific for
each subgroup. MMRA then identifies candidate regulator microRNAs by assessing
their subtype-specific expression, target gene enrichment in subtype signatures
and network analysis-based contribution to subtype gene expression. MMRA was
applied to a CRC dataset of 450 samples, assigned to various subtypes by three
different transcriptional classifiers. In total, 24 microRNA were associated to
subtypes, in most cases negatively contributing to the stem/serrated/mesenchymal
(SSM) poor prognosis subtype. Functional validation in CRC cell lines confirmed
downregulation of the SSM subtype by miR-194, miR-200b, miR-203 and miR-429,
and highlighted shared target genes and pathways mediating this effect.