A linkage analysis of cigarette and alcohol consumption in an unselected Mexican American population

A genome wide association study of fast beta EEG in families of European ancestry

BACKGROUND

Differences in fast beta (20-28Hz) electroencephalogram (EEG) oscillatory activity distinguish some individuals with psychiatric and substance use disorders, suggesting that it may be a useful endophenotype for studying the genetics of disorders characterized by neural hyper-excitability. Despite the high heritability estimates provided by twin and family studies, there have been relatively few genetic studies of beta EEG, and to date only one genetic association finding has replicated (i.e., GABRA2).

METHOD

In a sample of 1564 individuals from 117 families of European Ancestry (EA) drawn from the Collaborative Study on the Genetics of Alcoholism (COGA), we performed a Genome-Wide Association Study (GWAS) on resting-state fronto-central fast beta EEG power, adjusting regression models for family relatedness, age, sex, and ancestry. To further characterize genetic findings, we examined the functional and behavioral significance of GWAS findings.

RESULTS

Three intronic variants located within DSE (dermatan sulfate epimerase) on 6q22 were associated with fast beta EEG at a genome wide significant level (p<5×10-8). The most significant SNP was rs2252790 (p<2.6×10-8; MAF=0.36; β=0.135). rs2252790 is an eQTL for ROS1 expressed most robustly in the temporal cortex (p=1.2×10-6) and for DSE/TSPYL4 expressed most robustly in the hippocampus (p=7.3×10-4; β=0.29). Previous studies have indicated that DSE is involved in a network of genes integral to membrane organization; gene-based tests indicated that several variants within this network (i.e., DSE, ZEB2, RND3, MCTP1, and CTBP2) were also associated with beta EEG (empirical p<0.05), and of these genes, ZEB2 and CTBP2 were associated with DSM-V Alcohol Use Disorder (AUD; empirical p<0.05).'

DISCUSSION

In this sample of EA families enriched for AUDs, fast beta EEG is associated with variants within DSE on 6q22; the most significant SNP influences the mRNA expression of DSE and ROS1 in hippocampus and temporal cortex, brain regions important for beta EEG activity. Gene-based tests suggest evidence of association with related genes, ZEB2, RND3, MCTP1, CTBP2, and beta EEG. Converging data from GWAS, gene expression, and gene-networks presented in this study provide support for the role of genetic variants within DSE and related genes in neural hyperexcitability, and has highlighted two potential candidate genes for AUD and/or related neurological conditions: ZEB2 and CTBP2. However, results must be replicated in large, independent samples.

Confirmation and generalization of an alcohol-dependence locus on chromosome 10q

Several genome scans on alcohol dependence (AD) and AD-related traits have been published. In this article, we present the results of a genome-wide linkage scan on AD and several related traits in 322 European-American (EA) families, and results of additional analysis in 335 African-American (AA) families that were the subject of a previous report. All families were initially ascertained for cocaine and/or opioid dependence. Non-parametric linkage analysis in the EA sample revealed suggestive linkages on chromosomes 7 (LOD=2.1 at 82.8 cM, p=0.0009) and 10 (LOD=3.0 at 137.7 cM, p=0.0001). The chromosome 10 linkage peak is 20 cM distal from a genome-wide significant linkage peak we observed previously in the AA sample. Parametric linkage analysis on chromosome 10 (assuming a recessive model, 80% penetrance, disease allele frequency=0.3) resulted in LOD scores of 2.7 at 136.7 cM and 1.9 at 121.7 cM in the EA and AA samples, respectively, with a combined sample genome-wide significant LOD score of 4.1 at 131.7 cM. To reduce heterogeneity of the AD phenotype, we also assessed linkage of chromosome 10 markers with the presence of alcohol withdrawal symptoms, one of the seven components of the DSM-IV diagnosis of AD. Suggestive evidence for linkage was observed in both populations with only 5 cM separating the location of the peak LOD scores despite a loss of power due to a smaller number of families informative for this trait. Results of our study confirm a chromosome 10 risk locus for AD in two genetically distinct populations and suggest that this locus may correspond more precisely to a specific component of the disorder.