MicroRNA modulation networks in Glioblastoma Multiforme subtypes

Abstract

Because Glioblastoma Multiforme (GBM) usually results in death for most patients diagnosed, it is extremely important to delineate the molecular subtypes of GBM and determine molecular targets for therapy. Though there have been various previous studies that have determined subtypes for GBM, there are no studies that have determined molecular subtypes with a specific gene signature, as well as a specific miRNA signature. This study seeks to expand on the generally accepted and recently discovered subtype determinations, Neural, Proneural, Mesenchymal and Classical by elucidating a miRNA signature for each. In order to determine possible signature miRNAs, functional miRNA-target interactions were integrated with miRNA and mRNA/gene expression to infer mRNA-mediated miRNA-miRNA interactions, using conditional mutual information, for each subtype. The generated network represent miRNA modulation via common target mRNA. Each subtype had a group of most highly connected miRNAs. Because many of these highly connected have experimentally validated evidence linking their regulation or dysregulation in GBM, those without any experimentally validated evidence could also be linked to GBM initiation or progression. Thus, these miRNAs could become future targets for therapeutic discovery and innovation. Further analysis is required to definitively classify miRNAs as linked to GBM in a significant fashion.