Combining evidence from human genetic and functional screens to identify pathways altering obesity and fat distribution

Overall adiposity and body fat distribution are heritable traits associated with altered risk of cardiometabolic disease and mortality. Performing rare-variant (minor allele frequency <1%) association testing using exome-sequencing data from 402,375 participants of European ancestry in the UK Biobank for nine overall and tissue-specific fat distribution traits, we identified 19 genes where putatively damaging rare variation associated with at least one trait (Bonferroni-adjusted p < 1.58 × 10^-7) and 50 additional genes at false discovery rate (FDR) ≤1% (p ≤ 4.37 × 10^-5). These 69 genes exhibited significantly higher (one-sided t test p = 3.58 × 10^-18) common-variant prioritization scores for association with body mass index (BMI), waist-to-hip ratio adjusted for BMI, and body fat percentage than genes not significantly enriched for rare putatively damaging variation, with evidence of monotonic allelic series (dose-response relationships) among ultra-rare variants (minor allele count ≤10) in 22 genes. Combining rare and common variation evidence, allelic series and longitudinal analysis, we selected 14 genes for CRISPR knockdown in human white adipose tissue cell lines. In two target genes, knockdown significantly (two-sided t test p < 0.05/14) decreased lipid accumulation: PPARG (fold change [FC] = 0.25, p = 5.52 × 10^-7) and SLTM (FC = 0.51, p = 1.91 × 10^-4); knockdown of COL5A3 (FC = 1.72, p = 0.0028) resulted in significantly increased lipid accumulation. Integrating across population-based genetic and in vitro functional evidence, we highlight therapeutic avenues for altering obesity and body fat distribution by modulating lipid accumulation.