Genome-wide association study of lipids in Greenlandic Inuit show large-effect size and independent associations and reveals a unique genetic architecture (PrgmNr 2036)

Objective: Perturbation of the lipid homeostasis is a key risk factor for cardiovascular disease (CVD), the leading cause of death worldwide. We aimed to investigate the role genetics plays for lipid levels, and via these CVD, in Greenlanders.
Methods: Using imputed data from 4473 Greenlanders, we performed a genome-wide association study (GWAS) of levels of six blood lipids and lipoproteins: triglycerides, LDL-cholesterol, HDL-cholesterol, total cholesterol, apolipoprotein A1, and apolipoprotein B. For relevant genome-wide significant variants, we then tested for associations with additional traits and outcomes, including risk of CVD and RNA expression. Finally, we investigated the combined effect of the independent genome-wide significant variants using genetic risk scores (GRSs).
Results: The GWAS identified 11 genome-wide significant loci, of which 9 were known. Two were potentially novel, however, they were more frequent in Europeans and not Bonferroni significant and therefore need independent replication to be validated. In the PCSK9 locus, we identified an association signal, which was independent of the well-described PCSK9 loss-of-function variant rs11591147. The derived allele of the lead SNP, rs4927191, in this locus was associated with lower levels of LDL-cholesterol (beta(SD)=-0.226, p=1.24e-12) and total cholesterol (beta(SD)=-0.169, p=2.28e-9). Interestingly, it was also associated with a lower risk of peripheral artery disease (hazard ratio=0.30 per allele, p=0.01), and was a top eQTL signal for PCSK9 across all tissues in GTEx. A variant (rs12117661) in LD with the lead SNP showed similar results, and there were furthermore indications of rs12117661 being located in a predicted regulatory element. When combining the GWAS hits into GRSs using 10-fold cross-validation, these scores alone explained up to 16.3% of the variance of the lipid traits. Furthermore, the GRSs for LDL-cholesterol, total cholesterol and apolipoprotein B were associated with increased risk of several types of CVD events.
Conclusion: We identified a possible causal variant that affected lipid levels and CVD risk by regulating the expression of PCSK9. Furthermore, we found that in the Greenlandic population much more of the variance in lipid levels can be explained with markedly fewer variants than in large European populations, supporting a marked difference in genetic architecture.