Abstract
OBJECTIVES: Metformin is the first-line treatment for type 2 diabetes (T2D), yet interindividual variability in glycemic response remains poorly understood. Preclinical studies demonstrate that low-dose metformin lowers glucose via Rap1 signaling in ventromedial hypothalamic neurons, a pathway not required for sulfonylurea or thiazolidinedione efficacy. We investigated whether a common intronic RAP1A variant (rs7525578) modifies glycemic response to metformin in humans.
METHODS: We analyzed men with T2D and HbA1c >6% in the UK Biobank who were prescribed metformin (n=7002), pioglitazone (n=587), or gliclazide (n=2654). Genotypes were obtained from imputed array data. The mean HbA1c by genotype (CC, CT, and TT) was compared within each drug cohort using 1-way ANOVA with Tukey post hoc tests. Effect sizes were quantified using eta-squared (η2).
RESULTS: Among men prescribed metformin, rs7525578 was significantly associated with HbA1c (F=5.644, P=0.004, η2=0.0016). Heterozygotes (CT) had a higher mean HbA1c (7.52%) than CC homozygotes (7.39%, P=0.002). No significant associations were observed among pioglitazone (P=0.444) or gliclazide (P=0.233) users, and no effects were detected in women. The metformin-specific association remained after adjustment for combination therapy.
CONCLUSIONS: The RAP1A variant rs7525578 is associated with modestly higher HbA1c among men treated with metformin, but not with pioglitazone or gliclazide. This drug-specific pharmacogenetic association mirrors mechanistic data in mice, supporting a role for Rap1 signaling in metformin action and highlighting pathway-based approaches to T2D pharmacogenetics.