Genome-wide association analyses and beyond for major depressive disorder and antiepileptic drug-induced severe cutaneous adverse reactions

Abstract

While accumulating genome-wide association studies are increasingly achieving their goal of identifying susceptibility loci for complex traits, the translation of genome-wide association (GWA) results to meaningful biological mechanisms remains a challenge. Two complex traits, major depressive disorder (MDD) and antiepileptic drug-induced severe cutaneous adverse reactions (AED-induced SCARs) have been studied with novel methods and analyses to explore the potential interpretation of GWASs.

In a study of major depression, the genomic and phenotypic data from the UK’s large-scale biobank data of ~500,000 individuals (UKBiobank) were employed to elucidate the genetic and phenotypic risk factors for depression. A subset of 79,839 White-British individuals who completed a mental health questionnaire were valid participants. One significant single nucleotide polymorphism (SNP) was detected (rs4906348; p < 1.5 ) and evaluated in terms of tissue expression through expression quantitative trait loci analyses. High expression of a nearby gene, BAG5, was found in tissues residing in the anterior cingulate cortex and in the hypothalamic-pituitary-adrenal axis, regions critical to depressive symptoms. Subsequent analyses of Mendelian randomisation and mediation models pinpointed putative intermediate phenotypes including neuroticism (and significant sub-items: negativity, mood instability, loneliness and tenseness) and sleeplessness. It was found that the effect of a depression-risk variant, rs28986303, was mediated by negativity and mood instability, and a nearby gene, GABBR1, is crucial for synaptic plasticity of GABAergic transmission, which plays a regulatory role in the hypothalamic-pituitary-adrenocortical (HPA) axis.

The study of AED-induced SCARs aims to establish the genetic architecture of this complex trait with next-generation sequencing technologies to examine the whole genome at finer grain. The common variants were analysed with standard GWAS procedure based on whole-genome sequencing data of 77 cases and 338 controls in a Hong Kong population, identifying two genome-wide significant SNPs, which could be mapped into genes LINC01149 (rs147099772, p < ; within the HLA complex genomic region) and SUCLG2 (rs139867579, p < ). A complementary AED stratified GWAS using a subset of 50 carbamazepine-induced SCARs cases reinforced the variant rs147099772 (p < ) with a doubled odds ratio (OR = 7.3). For rare variants, annotation and prioritisation were conducted, and then gene-based association tests were performed. The top four genes showed significant discrepancies in the case-control genotype frequency table. Genes ZNF676 and SPATA31C1 were speculatively related to the immune system and cutaneous conditions.

Put together, these two studies were examples of extending and interpreting GWA results to elucidate phenotypic functions. Downstream analyses with putative intermediate phenotypes and whole-genome analyses might facilitate a deeper understanding of the genetics of complex traits.

(429 words)