Genetically proxied glucagon-like peptide-1 receptor perturbation and risk of mood disorders: a Mendelian randomization study

Abstract

Background: Glucagon-like peptide-1 receptor (GLP1R) agonists have gained attention for their role in diabetes treatment along with their diverse effects, such as appetite suppression, suggesting potential psychiatric benefits. This study aimed to assess the effect of GLP1R perturbation on mood disorders based on protein and biomarker levels using Mendelian randomization (MR) approach.

Methods: We conducted two-sample MR using summary statistics for GLP1R plasma levels (n = 3,301) from the INTERVAL study, glycated hemoglobin (HbA1c) levels (n = 128,610) from the Meta-Analyses of Glucose and Insulin-related traits Consortium, and bipolar disorder (BD: 371 cases/360,823 controls) and major depressive disorder (MDD: 776 cases/360,418 controls) incidences from the UK Biobank. Genetic variants associated with the plasma levels of GLP1R and HbA1c were used as proxies for the variation in GLP1R.

Results: GLP1R level was significantly associated with a reduced risk of MDD (odds ratio [OR] = 0·9988, 95% confidence interval [CI] = 0·9978-0·9999, P = 0·0291) and of BD (OR = 0·9990, 95% CI = 0·9982-0·9998, P = 0·0182). GLP1R's HbA1c level-lowering effect was significantly associated with a decreased risk of BD (OR = 0·9786, 95% CI = 0·9613-0·9962, P = 0·0175) but not with MDD.

Conclusions: GLP1R perturbation may have protective effects on MDD and BD through different mechanisms, although additional clinical trials are required to determine the therapeutic implications.

Trial registration: Clinical trial number not applicable.

Keywords: Bipolar disorder; GLP1R; Glycemic control; Major depressive disorder; Mendelian randomization.

Fig. 1

Graphs of drug target MR. a In the protein–drug target MR plot, exposure is the plasma protein level (P). Dashed line indicates that an assumption, Φ_G = 0, is needed. This model estimates the value of µθ, which is the effect of multiple pathways of the protein. b In biomarker-drug target MR, the effect of downstream biomarker (X) is evaluated. This model estimates the value of θ and assesses effects through downstream biomarker pathways. Assumptions Φ_G = 0 and Φ_P = 0 are needed. Abbreviations: G, genetic variant; P, protein levels; X, downstream biomarker levels; D, disease; U, common causes of both P, X, D; Φ_G, direct effect of a genetic variant; Φ_P, direct effect of a protein; MR, Mendelian randomization

Fig. 2

Overview of the drug target MR study between GLP1R perturbation and mood disorders

Fig. 3

Forest plot of the protein-drug target MR for associations between GLP1R levels and mood disorders. GLP1R level, BD, and MDD data were used from INTERVAL and UK Biobank. Estimates of robust IVW MR are given as ORs with 95% CI. Abbreviations: OR, odds ratio; CI, confidence interval; MR, Mendelian randomization

Fig. 4

Forest plot of the biomarker-MR for associations of GLP1R activity, glycemic control and mood disorders. Association between GLP1R activity, glycemic control, and (a) BD and (b) MDD. HbA1c data were used from MAGIC and linked to BD and MDD data from UK Biobank. Estimates of robust IVW MR are given as OR with 95% CI. OR indicates the odds ratio, diff refers to the difference of estimates, and ‘diff.p’ is the p value indicating the significance of the difference. Abbreviations: OR, odds ratio; CI, confidence interval; HbA1c, glycated hemoglobin; MAGIC, Meta-Analyses of Glucose and Insulin-related traits Consortium

Fig. 5

Forest plot of the one-sample MR for associations of GLP1R activity and glycemic control.Association between GLP1R activity and glycemic control with (a) BD and (b) MDD. HbA1c, BD, and MDD data were used from UK Biobank. Estimates are given as OR with 95% CI. OR indicates the odds ratio, diff refers to the difference of estimates. ‘diff.p’ is the p value that indicates the significance of the difference. Abbreviations: OR: odds ratio, CI: confidence interval, HbA1c: glycated haemoglobin