Background: The heritability of Alcohol Dependence (AD) has been estimated to be roughly 50% by twin studies and 30% by whole-genome SNP-based studies (Prescott & Kendler, 1999; Mbarek et al., 2015). However, genome-wide... [ view full abstract ]
Background: The heritability of Alcohol Dependence (AD) has been estimated to be roughly 50% by twin studies and 30% by whole-genome SNP-based studies (Prescott & Kendler, 1999; Mbarek et al., 2015). However, genome-wide association studies (GWAS) have yet to account for a large proportion of this heritability, and they lack the ability to provide insight into functional pathways. However, gene co-expression networks can be used to leverage GWAS results and provide functional information. Controlled experiments have identified significantly ethanol-responsive gene networks in mouse brain. Direct integration of human GWAS and protein-protein interaction (PPI) data with mouse gene expression data has the potential to identify novel associated gene networks and the mechanistic frameworks through which they function.
Methods: The present analysis used Edge-Weighted dense module searching for Genome Wide Association Studies (EW-dmGWAS) to co-analyze GWAS data from the Irish Affected Sib-Pair Study of Alcohol Dependence (IASPSAD), human PPI data, and acute-ethanol-exposed mouse gene expression data, to identify and prioritize ethanol-regulated gene networks (i.e. modules) with respect to AD risk contribution. Expression data was obtained from the Ventral Tegmental Area (VTA), Nucleus Accumbens (NAc), and Prefrontal Cortex (PFC), due to their role in reward pathways, and previous findings of significantly ethanol-regulated networks in these regions. Largely overlapping modules from the EW-dmGWAS output were then merged to form Mega Modules (MMs). To validate our results, we tested modules for overrepresentation of genes with nominal GWAS p-values (p<0.001) from an alternative GWAS dataset (Avon Longitudinal Study of Parents and Children; ALSPAC). To determine functional relatedness of genes within significantly overrepresented modules, we analyzed their functional enrichment using ToppGene.
Results: After correction for multiple testing, there were 276, 171, and 314 significant MMs for the VTA, NAc, and PFC, respectively. One significant MM from each brain region was significantly overrepresented with ALSPAC-nominally significant genes. Top functional enrichment categories for these MMs included: ubiquitination, and ligase activity for the VTA MM; telomere maintenance and syndecan-mediated signaling for the NAc MM; and chromatin organization, immune response, and NFKB and Wnt signaling pathways for the PFC MM.
Discussion: The results indicate that integration of mouse gene expression data and human genetic data via EW-dmGWAS allows identification of novel alcohol-associated gene networks in the VTA, PFC, and NAc with overrepresentation of nominally associated genes from an independent GWAS dataset. This suggests that IASPSAD and ALSPAC have identified different AD-associated genes that are contained in the same ethanol-responsive networks in these three brain regions. Functional enrichment results suggest that these networks potentially play a role in several brain region-specific pathways involved in gene and protein regulation, cell signaling, and immune response. The exact mechanisms through which these pathways are related to alcohol dependence will be explored in future studies.
Animal models , Gene Finding Strategies , Statistical Methods , Substance use: Alcohol, Nicotine, Drugs
