Using Gene Regulation In Human Brain To Move Genome-Wide Association Study Discoveries Forward For Nicotine Dependence

Abstract

Cigarette smoking is the leading cause of preventable mortality worldwide. The likelihood of quitting smoking is reduced by nicotine dependence (ND), a heritable trait with a complex genetic architecture. The genome-wide association study (GWAS) era has been successful for ND, bringing about several important discoveries of specific common variants in nicotine acetylcholine receptor genes (CHRNA5-A3-B4, CHRNB3-A6, and CHRNA4) and other loci. Follow-up of these GWAS findings has shown that the genetics underlying ND involves multiple genes and variants exerting effects on gene regulation across different brain tissues. GWAS-identified variants associated with ND in or near CHRNA5 were found to include a missense variant that alters function of the CHRNA5 protein (rs16969968) and other variants that tag quantitative trait loci (QTL) altering CHRNA5 methylation (rs11636753) and expression (rs680244 and rs880395) in prefrontal cortex. As another example of neurobiology uncovered through GWAS discovery, CHRNA4 was found to harbor a ND-associated splice site acceptor variant (rs2273500) that correlates with CHRNA4 expression in intralobular white matter. To identify additional loci, our latest and the largest ever GWAS of ND involved combining results for 19 million 1000 Genomes-imputed variants tested for association with mild/moderate/severe ND across 15 independent studies: total N=38,602 smokers (28,677 Europeans/European Americans and 9,925 African Americans). This GWAS yielded a novel association in the DNA methyltransferase (DNMT3B) gene. No DNMT3B variants have been reported before for any addiction phenotype. For our top variant rs910083, its C allele was found to associate with increased ND risk across both ancestries: P=3.7E-8, odds ratio (95% confidence interval) = 1.06 (1.04–1.07) for severe vs. mild ND. The association was independently confirmed in the UK Biobank (N=48,931) using heavy vs. never smoking as a proxy phenotype (P=3.6E-4). Moreover, rs910083 was identified as a methylation QTL variant associated with DNMT3B methylation in fetal brain (P=2.3E-26) and an expression QTL variant associated with DNMT3B RNA expression in adult cerebellum of two independent studies (P ranging from 2.1E-7 to 8.8E-5). This finding offers a new direction to better understand genetic regulatory variants that contribute to ND, as the cerebellum is an often-overlooked tissue for addiction. Current studies are focused on moving beyond sequential analyses of different ‘omics (GWAS → methylation → expression), by using concurrent integration of GWAS, methylation, and expression as an anchor for discovery. Increasingly larger sample sizes will likely enable more GWAS discoveries, but equally important is the utilization of gene regulatory information in human brain to gain insight into the consequences of common genetic perturbations that influence risk of developing ND.