Genetically modeled GLP1R and GIPR agonism reduce binge drinking and alcohol-associated phenotypes: a multi-ancestry drug-target Mendelian randomization study

Abstract

Pharmacological modulation of glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP) through dual GIP/GLP-1 receptor agonists, commonly used for diabetes and obesity, shows promise in reducing alcohol consumption. We applied drug-target Mendelian randomization (MR) using genetic variation at these loci to assess their long-term effects on problematic alcohol use (PAU), binge drinking, alcohol misuse classifications, liver health, and other substance use behaviors. Genetic proxies for lowered BMI, modeling the appetite-suppressing and weight-reducing effects of variants in both the GIPR and GLP1R loci ("GIPR/GLP1R"), were linked with reduced binge drinking in the primary (β = -0.44, 95% CI [-0.72, -0.15], P = 2.42 × 10^-3) and replication data (β = -0.13, [-0.22, -0.04], P = 0.0058). HbA1c lowering via GIPR/GLP1R variants was associated with reduced risk of heavy drinking with psychiatric comorbidities versus low-risk drinking (odds ratio [OR] = 0.62, [0.45, 0.85], P = 0.0031), with replication in independent HbA1c data (OR = 0.71, [0.60, 0.84], P = 5.22 × 10^-5) and directional consistency with reduced PAU. Analysis of individual loci indicated that both GIPR and GLP1R were protective against heavy drinking, underscoring the importance of both targets. While estimates for other substance use disorders (tobacco, cannabis, opioid) were consistently null, food preference analyses revealed that BMI lowering via GIPR/GLP1R reduced fatty food liking (β = -1.58, [-2.01, -1.14], P = 1.62 × 10^-12) and increased vegetarian food liking (β = 2.08, [1.17, 2.99], P = 8.22 × 10^-6), implicating metabolic and appetite regulation pathways for the alcohol consumption findings. For liver health, HbA1c lowering via GIPR/GLP1R was associated with reduced NAFLD (β = -0.34, [-0.50, -0.18], P = 2.74 × 10^-5) and lower ALT levels (β = -0.26, [-0.38, -0.15], P = 8.39 × 10^-6), with replication supporting these findings. Consistency across multiple MR methods and colocalization analyses strengthened causal inference. Mediation analysis suggested reductions in hazardous alcohol consumption partially explain the cardioprotective effects of these agonists. Multi-ancestry analyses supported directionally aligned relationships in non-European cohorts. These findings support further clinical exploration of GLP1R, GIPR, and dual agonists in addiction medicine.

Fig. 1. Comparison of drug-target Mendelian randomization (DTMR) with randomized control trials (RCTs) (adapted from Chauquet et al. [90]).

The design for a a hypothetical RCT that investigates the impact of GIPR/GLP1R agonists on alcohol consumption behaviors, liver health, and food preferences is compared to b a drug-target MR analysis assessing the impact of genetically proxied associations of GIPR/GLP1R activity. In b, Single nucleotide polymorphisms (SNPs) located within or near the genomic loci and associated with the downstream physiological impact emulate GIPR/GLP1R therapeutics (reduced glycated hemoglobin [HbA1c] or lowered body mass index [BMI]).

Fig. 2. Study overview.

This study used summary-level GWAS data relating to glycated hemoglobin (HbA1c) and body mass index (BMI) to construct genetic instruments modeling GLP1R and GIPR agonism. We constructed three instrument types: one proxying GLP1R agonism; one proxying GIPR agonism; and one combined instrument proxying dual GLP1R and GIPR agonism. Each instrument type included multiple exposure sources mimicking the expected physiological responses to pharmacological modulation of the targets (lowered glycated hemoglobin [HbA1c], reduced body mass index [BMI], and GLP1R or GIPR gene expression in the cortex). Instrument sets for each BMI and HbA1c exposure were constructed in two independent GWAS summary statistics (UK Biobank [plus GIANT for BMI] and the Million Veterans Program [MVP]). After instrumentation and validation with the primary clinical indications for GLP1R and GIPR agonism (type 2 diabetes and obesity), and assessing their impact on liver health, we obtained a selection of outcomes related to alcohol use disorder (AUD) and alcohol consumption behavior to assess the impact of GLP1R and GIPR agonism. We contextualized the alcohol-related analyses by analyzing other substance use disorders and investigating outcomes related to self-reported food liking. Because of the availability of large sample sizes and the most relevant endpoints, we used data from European ancestry as the main analysis set, but we also performed analyses using East Asian and African ancestry data sources. Finally, for all drug-target MR estimates demonstrating evidence of a relationship (main drug-target MR method P < 0.05), we performed colocalization analyses to assess evidence of shared causal variants between the biomarker exposures and outcomes in the GLP1R and GIPR genomic loci. MR Mendelian Randomization, GLP1R Glucagon-like peptide-1 receptor, GIPR glucose-dependent insulinotropic polypeptide receptor, NAFLD Non-alcoholic fatty liver disease, ALD Alcohol-related liver disease, SNP Single nucleotide polymorphism, BMI Body mass index.

Fig. 3. Drug-target MR estimates of GIPR and GLP1R agonism on alcohol consumption outcomes.

This figure presents Mendelian Randomization (MR) estimates assessing the effects of GIPR, GLP1R, and dual GIPR/GLP1R agonism on Problematic Alcohol Use, self-reported binge drinking behavior, and drinks per week. Results are shown separately for genetically proxied reductions in BMI and HbA1c levels, two primary mechanisms through which these agonists exert their clinical effects. Beta values and 95% confidence intervals (CI) are displayed, with MR estimates derived from biomarker (BMI or HbA1c) data using the primary and replication data sources outlined in the Methods and Table 1. GIPR glucose-dependent insulinotropic polypeptide receptor, GLP1R Glucagon-like peptide-1 receptor, HbA1c glycated hemoglobin, BMI body mass index, CI confidence interval.

Fig. 4. Drug-target MR estimates of GIPR and GLP1R agonism on alcohol misuse classifications comparing high-risk drinking to other drinking patterns.

This figure presents MR estimates assessing the effects of GIPR, GLP1R, and combined GIPR/GLP1R agonism on high-risk drinking behaviors compared to other drinking patterns, including low risk drinking (4v1), Internalizing alcohol use (4v2), and heavy drinking (4v3). Results are displayed separately for genetically proxied reductions in BMI (left) and HbA1c levels (right), reflecting distinct physiological mechanisms of these agonists. Odds ratios (OR) and 95% confidence intervals (CI) are plotted for both primary and replication datasets. Independent replication data sources are described in the Methods and Table 1. GIPR glucose-dependent insulinotropic polypeptide receptor, GLP1R glucagon-like peptide-1 receptor, BMI body mass index, HbA1c glycated hemoglobin, CI confidence interval.

Fig. 5. Drug-target MR estimates of GIPR and GLP1R agonism on liver disease risk and serum enzyme levels.

This figure presents MR estimates assessing the effects of GIPR, GLP1R, and dual GIPR/GLP1R agonism on the multi-trait NAFLD GWAS (see Methods), serum alanine aminotransferase levels, aspartate aminotransferase, and serum gamma-glutamyl transpeptidase levels. Results are shown separately for genetically proxied reductions in BMI and HbA1c levels, two primary mechanisms through which these agonists exert their clinical effects. Beta values and 95% confidence intervals (CI) are displayed, with MR estimates derived from biomarker (BMI or HbA1c) data using the primary and independent replication data sources (described in the Methods and Table 1). GIPR glucose-dependent insulinotropic polypeptide receptor, GLP1R Glucagon-like peptide-1 receptor, HbA1c glycated hemoglobin; BMI body mass index, CI confidence interval.

Fig. 6. Summary of protective associations for GIPR/GLP1R, GIPR, and GLP1R targets across alcohol-related, dietary, and liver traits.

This figure presents traits with consistent protective or beneficial associations across both primary and replication datasets for each receptor target (GIPR/GLP1R, GIPR, GLP1R), using either BMI (blue) or HbA1c (yellow) as the exposure biomarker. Arrows indicate direction of effect. Asterisks (*) denote associations that surpassed the Bonferroni-corrected significance threshold (P < 0.0025). Results with dashed borders surpassed testing for multiple comparisons in either the primary or replication exposure data but had P-value > 0.05 in the other. Traits shown include alcohol use behaviors (e.g., binge drinking, high-risk drinking), dietary preferences (e.g., fatty, fried, vegetarian, and low-calorie foods), and liver-related outcomes (e.g., NAFLD, ALT, AST, GGT). GIPR glucose-dependent insulinotropic polypeptide receptor, GLP1R Glucagon-like peptide-1 receptor, HbA1c glycated hemoglobin, BMI body mass index.