Genome-Wide Association Study (GWAS) and Genome-Environment Wide Interaction Study (GEWIS) of Late-Life Depressive Symptoms in Women

Monday, 18 August 2014
Exhibit hall (Dena'ina Center)
Erin C Dunn , Massachusetts General Hospital, Boston, MA
Anna Wiste, MD , Massachusetts General Hospital, Boston, MA
Farid Radmanesh, MD , Massachusetts General Hospital, Boston, MA
Karestan C Koenen, PhD , Columbia University, New York, NY
Shaun Purcell, PhD , Massachusetts General Hospital, Boston, MA
Marilyn Cornelis, PhD , Harvard School of Public Health, Boston, MA
Jonathan Rosand, MD , Massachusetts General Hospital, Boston, MA
Peter Kraft, PhD , Harvard School of Public Health, Boston, MA
Sylvia Wassertheil-Smoller, PhD , Albert Einstein College of Medicine, Bronx, MA
Jordan Smoller, MD , Massachusetts General Hospital, Boston, MA
INTRODUCTION: Depression is a common mental health problem among women. Although the role of environmental exposures is well-established for depression, little progress has been made in identifying specific genes that increase risk. We undertook a genome-wide association study (GWAS) and genome-environment wide interaction study (GEWIS) of depressive symptoms using data from the Women’s Health Initiative, a large population-based, longitudinal study of post-menopausal women.  

METHODS: We examined the joint effect of common genetic variants (e.g., single nucleotide polymorphisms; SNPs) with two environmental exposures, stressful life events and social support.  We used data from African American (n=8,565) and Hispanic (n=3,709) women, making this the first large-scale GWAS of depression in these groups.  Women were ages 50-79 at baseline.  Participants were genotyped using the Affymetrix 6.0 chip; additional SNPs were imputed using the 1000 Genomes reference panel.  Depressive symptoms were assessed using six items from the Center for Epidemiological Studies of Depression Scale.

RESULTS: No SNPs achieved genome-wide significance (p<5x10-8) in either sample.  The strongest association signal in African Americans was for rs73531535 (p=5.75x10-8), an imputed SNP located 20kb from GPR139 (G protein-coupled receptor 139) and in Hispanics, for the intronic SNP rs4542757 (p=7.31x10-7) in DCC (deleted in colorectal cancer).  Stressful life events were positively associated with depressive symptoms (African American: β=1.42; p<0.001; r2=0.06; Hispanic:β=1.62; p<0.001; r2=0.06); social support was inversely associated (African American: β=1.46; p<0.001; r2=0.06; Hispanic: β=2.12; p<0.001; r2=0.09).  After Bonferroni correction for multiple testing of the top 1% of GWAS hits (p<2.47x10-6 African American; p<3.71x10-6 Hispanics), there were 8 statistically significant SNPs in the GEWIS of social support in African Americans and 5 significant SNPs in Hispanics.

CONCLUSIONS: Social support may be an important buffer of genetic factors associated with depression.  GEWIS appears to be a promising and unbiased approach to identify novel loci associated with depression.  The results await replication.