Effects of rare coding variants in severe early-onset obesity genes in the population-based UK Biobank study.

Objectives: Clinical case-based studies have identified rare pathogenic variants in several genes as causes of severe early-onset obesity, but their penetrance and interaction with polygenic susceptibility in the general population remain unclear. We analysed the UK Biobank (UKBB) whole exome sequence data to assess the effects of heterozygous variants in nine previously reported genes on adult BMI and recalled childhood adiposity.

Methods: Among 419,581 UKBB participants, we identified heterozygous carriers of coding variants that were i) experimentally-characterised as loss-of-function (LoF), or ii) bioinformatically-predicted as rare (minor allele frequency <0.1%) LoF. We assessed variant-level and gene-level population penetrance of obesity and associations with adult BMI and recalled childhood adiposity and tested the statistical interaction between rare variants carriage and a BMI polygenic score.

Results: Considering experimentally-characterised LoF variants (excluding MC4R), we identified 22 heterozygous and 2 homozygous variants in three autosomal recessive genes (POMC, PCSK1, LEPR), and three autosomal dominant genes (SH2B1, SIM1, KSR2) with at least 10 carriers in the UKBB. Obesity penetrance among carriers ranged from 8% to 29% (median 23%), and none was significantly different from non-carriers (24%, all P >0.05). For bioinformatically-predicted rare LoF variants, gene-based burden tests showed that carriage of heterozygous variants in MC4R, PCSK1, and POMC was associated with higher adult BMI (effect sizes ranged from 0.5 to 2.5 kg/m2, all P <0.003), with no significant interaction effects with common variant polygenic risk of BMI.

Conclusions: This study provides the population-specific report of variant penetrance of known obesity genes and confirmed the heterozygous rare variant effects in MC4R, POMC and PCSK1. We also underscore the utility of population-based studies in supporting variant classifications.