Genome-wide association analysis of coffee drinking suggests association with CYP1A1/CYP1A2 and NRCAM

Abstract

Coffee consumption is a model for addictive behavior. We performed a meta-analysis of genome-wide association studies (GWASs) on coffee intake from 8 Caucasian cohorts (N=18 176) and sought replication of our top findings in a further 7929 individuals. We also performed a gene expression analysis treating different cell lines with caffeine. Genome-wide significant association was observed for two single-nucleotide polymorphisms (SNPs) in the 15q24 region. The two SNPs rs2470893 and rs2472297 (P-values=1.6 × 10(-11) and 2.7 × 10(-11)), which were also in strong linkage disequilibrium (r(2)=0.7) with each other, lie in the 23-kb long commonly shared 5' flanking region between CYP1A1 and CYP1A2 genes. CYP1A1 was found to be downregulated in lymphoblastoid cell lines treated with caffeine. CYP1A1 is known to metabolize polycyclic aromatic hydrocarbons, which are important constituents of coffee, whereas CYP1A2 is involved in the primary metabolism of caffeine. Significant evidence of association was also detected at rs382140 (P-value=3.9 × 10(-09)) near NRCAM-a gene implicated in vulnerability to addiction, and at another independent hit rs6495122 (P-value=7.1 × 10(-09))-an SNP associated with blood pressure-in the 15q24 region near the gene ULK3, in the meta-analysis of discovery and replication cohorts. Our results from GWASs and expression analysis also strongly implicate CAB39L in coffee drinking. Pathway analysis of differentially expressed genes revealed significantly enriched ubiquitin proteasome (P-value=2.2 × 10(-05)) and Parkinson's disease pathways (P-value=3.6 × 10(-05)).

Figure 1

Regional association plot for 15q24. The vertical axis shows the negative logarithm of the association P-values, and the horizontal axis shows the position in mega bases. Each dot represents an SNP and the colors of the dots represent the extent of linkage disequilibrium with the top SNP, which is colored in purple. Genes in the region are shown below the horizontal axis.

Figure 2

Forest plot for the top hits from the GWAS. Vertically left, the populations included in the GWAS and the replication phase. The boxes represent precision and horizontal lines represent the confidence intervals. The diamond represents the pooled effect estimate from the meta-analysis of all cohorts. The horizontal axis shows the scale of the effects.