Oral metformin transiently lowers post-prandial glucose response by reducing the apical expression of sodium-glucose co-transporter 1 in enterocytes

**Background:** Metformin is a first-line glucose-lowering agent for type 2 diabetes, but its mechanisms of action remain incompletely understood. While traditionally thought to act primarily via hepatic gluconeogenesis suppression and increased insulin sensitivity, emerging evidence points to the gut as a key site of action. The effect of metformin on intestinal glucose absorption, particularly via the sodium-glucose co-transporter 1 (SGLT1), has been controversial. This study aimed to determine whether a single dose of metformin acutely reduces post-prandial glucose response by modulating SGLT1-mediated glucose absorption in enterocytes.

**Methods:** The study employed a multi-model approach. In vitro, polarized Caco-2/TC7 human intestinal cells were incubated with 10 mM metformin for 1 hour, approximating the apical concentration after oral ingestion. SGLT1 abundance in the brush border membrane was assessed by immunofluorescence and quantified by fluoroscopy. Functional glucose uptake was measured using 14C-labeled α-methyl-D-glucopyranoside (AMG), a non-metabolizable SGLT1-specific substrate. Washout experiments evaluated reversibility at 3 and 9 hours. In vivo, Goto-Kakizaki (GK) diabetic rats received a single oral dose of metformin (200 mg/kg) 1 hour before an oral glucose tolerance test (OGTT; 1 g/kg). SGLT1 abundance and phosphorylated PKA substrate levels in jejunum were assessed by immunofluorescence. Additional OGTTs at 6 and 12 hours tested reversibility. Dose-response (50–200 mg/kg) and chronic dosing (7 days daily) experiments were performed. Intraperitoneal glucose tolerance tests (IPGTT) distinguished gut from systemic effects. Experiments were repeated in C57BL/6 mice and in Sglt1-knockout and Glut2-knockout mice. In Göttingen mini-pigs (n=10), metformin (25 mg/kg) or vehicle was administered via jejunostomy 30 minutes before a glucose + D-xylose bolus; plasma glucose, insulin, D-xylose, and GLP-1 were measured. PET-CT dynamic imaging with intraluminal 18F-FDG was performed in GK rats using surgically created closed compartments (pylorus, ileocecal valve, and urethra occluded) to directly visualize intestinal glucose retention.

**Key Results:** In Caco-2/TC7 cells, 1-hour metformin incubation significantly reduced apical SGLT1 abundance (p < 0.001) and AMG uptake (p < 0.05). This effect was reversible: SGLT1 returned to vehicle levels at 3 hours and fully normalized by 9 hours. No changes in SLC5A1 or other gluco-transporter mRNA expression were observed. In GK rats, a single metformin dose (200 mg/kg) reduced apical SGLT1 density in jejunum (p < 0.0001) and decreased phosphorylated PKA substrates (p < 0.0001). PGR was significantly lowered during OGTT at 1 hour post-dose (p < 0.05), attenuated at 6 hours, and fully normalized at 12 hours. The effect was dose-dependent, maximal at 200 mg/kg but effective at 50 mg/kg. Chronic daily dosing for 7 days did not alter PGR when glucose was given 12 hours after the last dose, but an additional acute dose 1 hour before OGTT again reduced PGR. IPGTT showed no significant difference between metformin and vehicle, confirming a gut-specific effect. In C57BL/6 mice, metformin reduced PGR (p < 0.05). In Glut2-KO mice, metformin still significantly decreased PGR (p < 0.05), but in Sglt1-KO mice, metformin had no effect on PGR, demonstrating SGLT1 dependence. In mini-pigs, metformin reduced PGR (p < 0.05), plasma insulin (p < 0.05), and plasma D-xylose appearance (p < 0.05), while increasing GLP-1 release (p < 0.05). PET-CT imaging in GK rats showed that metformin retained 18F-FDG in the intestinal lumen (p = 0.0001 for time-activity curves) and reduced its appearance in the bladder (p = 0.03).

**Clinical Implications:** This study provides direct evidence that a single oral dose of metformin transiently reduces intestinal glucose absorption by decreasing apical SGLT1 density in enterocytes, leading to a lower post-prandial glucose response. The effect is rapid, reversible, dose-dependent, and SGLT1-dependent. These findings align with clinical observations that metformin reduces PGR by approximately 21–23% when given 1 hour before glucose. The results challenge current recommendations to administer metformin at bedtime or after meals and suggest that pre-meal dosing may enhance gut-mediated glucose-lowering effects. The study also highlights SGLT1 inhibition in the gut as a clinically relevant mechanism, consistent with data showing that SGLT1 partial loss-of-function haplotypes reduce T2D incidence by ~2% per year. Limitations include the supraphysiological metformin concentration used in vitro (10 mM), though this approximates luminal concentrations after oral dosing, and the lack of direct AMPK activation measurements in intestinal samples.