Abstract
AIMS: Glucagon-like peptide 1 receptor agonists (GLP1RA), used to treat type 2 diabetes and obesity, have been associated with off-target behavioural effects. We systematically assessed genetic variation in the GLP1R locus for impact on mental ill-health (MIH) and cardiometabolic phenotypes across diverse populations within UK Biobank.</p>
MATERIALS AND METHODS: All genetic variants with minor allele frequency >1% in the GLP1R locus were investigated for associations with MIH phenotypes and cardiometabolic phenotypes. Linear or Logistic regression analyses (adjusted for age, sex, population structure and genotyping chip) were conducted separately in unrelated individuals of self-reported white British (N = 408 774), white European (N = 50 314), South Asian (N = 7667), multiple-ancestry groups (N = 10 437) or African-Caribbean (N = 7641) subsets. All ancestries were subsequently combined in an inverse variance-weighted fixed effects meta-analysis. Bonferroni correction for multiple testing was applied (for number of independent genetic variants).</p>
RESULTS: Associations were identified between GLP1R variants and body mass index (BMI), blood pressure and type 2 diabetes in all ancestries. All ancestries except South Asian had significant MIH associations (mood instability: rs111265626-G, odds ratio [OR] 0.851 [confidence interval, CI 0.79-0.92], risk-taking behaviour: rs75408972-T, OR 1.05 [CI 1.03-1.08] or chronic pain: rs9296280-C, OR 0.645 [CI 0.54-0.78]). The trans-ancestry meta-analysis showed mainly consistent effect sizes and directions for metabolic traits, but discordant directions MIH associations. Only signals for chronic pain, stroke and BMI influenced expression of GLP1R.</p>
CONCLUSIONS: GLP1R variants have consistent cardiometabolic effects across ancestries, but effects on MIH phenotypes are more varied. Any observed behavioural changes with GLP1RA are likely not acting directly through GLP1R.</p>