The discovery (case-only, N=1,046) and replication (case-control; N=206 cases, 176 controls) study populations were non-Hispanic whites. The discovery samples were genotyped on an Illumina HumanExome BeadChip and imputed to 1.2 million SNPs. Cases were considered exposed to TS if they were a smoker ≥5yrs prior onset. Logistic regression determined SNPs marginally associated with TS (2-sided p<0.05). The discovery variants were formally tested for interactions with TS in the replication data, where TS exposure was defined as ever/never smoker. The replication samples were genotyped on the Illumina MEGAEx BeadChip and imputed to 7.5 million SNPs. Logistic regression with main effects for TS and each discovery SNP and an interaction term were conducted. A replication 1-sided alpha of 0.05 was considered significant. Pathway enrichment analyses were conducted for genic replicating variants.

57,619 SNPs were marginally associated with TS in the discovery analysis, and 97.7% of these variants were tested for GxE replication in the case-control analysis. 3,513 variants had significant replicating GxE interactions, including 1,977 genic variants across 317 genes. Variants within BRWD1, CAPN8, PSMG1, and WASL has the most significant GxE associations. Interestingly, BRWD1 is a master orchestrator of B-cell development, and was recently associated with tobacco use. Also noteworthy is WASL, which confers risk for Parkinson disease and is associated with smoking cessation. Fascinatingly, pathway analyses demonstrated a significant enrichment of genes associated with tobacco use disorder (p=2.7x10^-6), xenobiotic glucuronidation (p=2x10^-14), and drug metabolism (p=1x10^-9).

These results demonstrate the genetic variation in processes related to tobacco use (initiation and cessation) and TS metabolism confers increased risk for MS among smokers.