Inference and validation of an integrated regulatory network of autism

Abstract

AbstractBackgroundAutism is a complex neurodevelopmental disorder. Functional roles and mechanisms of several non-coding transcripts including long noncoding RNAs (lncRNAs) have been shown to influence the pathobiology of autism. We hypothesize that there are more autism-associated lncRNAs to be discovered. In this work, we utilize a systems biology approach to identify novel lncRNAs that might play a role in the molecular pathogenesis of autism.MethodsBased on the data provided by the Simons Foundation Autism Research Initiative (SFARI), a three-component regulatory network comprising mRNAs, microRNAs (miRNAs) and lncRNAs was constructed. Functional enrichment analysis was performed to identify molecular pathways potentially mediated by components of the network. The potential association of four candidate lncRNAs with autism was investigated experimentally by developing and verifying a valproic acid (VPA)-exposed mouse model of autism.ResultsWe composed a network of 33 mRNA nodes, 25 miRNA nodes and 4 lncRNA nodes associated with neurologically relevant pathways and functions. We then verified the differential expression of four candidate lncRNAs: Gm10033, 1500011B03Rik, A930005H10Rik and Gas5 in the brain of VPA-exposed mice. We furthermore identified a novel splice variant of Gm10033, designated as Gm10033-ΔEx2, which was expressed in various mouse tissues.ConclusionsThe integrative approach combines the analysis of a three-component regulatory network with experimental validation of targets in an animal model of autism. As a result of the analysis, we prioritized a set of candidate autism-associated lncRNAs. These links add to the common understanding of the molecular and cellular mechanisms that are involved in disease etiology, specifically in the autism.