Recent studies have revealed glucagon's pivotal role in stimulating insulin secretion from intact beta cells²⁵. During the fed state, glucagon functions as an insulinotropic hormone, complementing rather than opposing insulin action, thereby maintaining euglycemia^25,43–45. To mimic the hyperglycemia and hyperglucagonemia characteristic of insulin resistance and pre-diabetes⁴⁶, we cultured human islets for 48 hours in high-glucose (10 mmol/L) with or without glucagon (10 pmol/L), using islets maintained at basal glucose (5.5 mmol/L) as controls. Notably, liraglutide enhanced GSIS in normoglycemic islets exposed to combinated of high glucose and exogenous glucagon, but not in islets treated with either basal or high glucose alone (Fig. 2A). In islets from donors with obesity and/or glucose intolerance, liraglutide similarly enhanced GSIS following chronic exposure to high glucose plus glucagon, comparable to the enhancement observed with either basal or high glucose alone (Fig. 2B,C). Consistent with previous studies^25,45, these findings suggest that glucagon stimulates insulin secretion in human islets under high-glucose conditions (10 mM) primarily through GLP-1R activation. Moreover, chronic exposure to high glucose and glucagon elevated both GLP1R and INS mRNA levels in islets from donors with glucose intolerance or T2D (Fig. 2D,E), potentiating liraglutide's insulinotropic effects through enhanced receptor availability.

GLP-1 7–36 and GLP-1 9–36, Enhances Glucose-Stimulated Insulin Secretion from Normoglycemic Human and Mouse islets.

While liraglutide did not enhance GSIS in islets from normoglycemic donors, previous studies have demonstrated native GLP-1 (GLP-1 7–36) to be a potent insulin secretagogue under similar conditions^37,47–49. To explore this discrepancy, we assessed GSIS using the same donor islet preparations from normoglycemic donors and chow-fed mice, comparing liraglutide with GLP-1 7–36 as a positive control. Of note, GLP-1 7–36 significantly potentiated GSIS at both 10 pmol/L and 25 nmol/L, in contrast to liraglutide (Supplementary Fig. 2A,B). This effect was replicated in mouse islets under high glucose conditions (Supplementary Fig. 2C,D), accompanied by elevated intracellular Ca²⁺ levels at both basal and high glucose concentrations compared to vehicle controls (Supplementary Fig. 2E). To evaluate GLP-1R-mediated intracellular signaling independently of glucose concentrations, we measured intracellular cAMP levels at 5-minute intervals following exposure to increasing doses of liraglutide or GLP-1 7–36. Both ligands elicited comparable dose-dependent increases in intracellular cAMP (Supplementary Fig. 2F,G), suggesting that the differential secretory responses to liraglutide and GLP-1 7–36 stem from their distinct effects on intracellular Ca²⁺ mobilization rather than variations in GLP-1R cAMP signaling. The differential effects of GLP-1 7–36, semaglutide, and liraglutide on GSIS in normoglycemic donor islets suggested distinct mechanisms of action, potentially involving dipeptidyl peptidase 4 (DPP4) degradation⁵⁰. DPP4, predominantly expressed in human and mouse alpha cells ^51,52, rapidly converts GLP-1 7–36 to GLP-1 9–36, a weak insulinotropic peptide that inhibits glucagon secretion and hepatic glucose production^53,54. Using islets from normoglycemic donors (n = 5), we compared the effects of liraglutide (10 pmol/L) and GLP-1 7–36 (10 pmol/L), with and without the DPP4 inhibitor sitagliptin, on insulin secretion under high glucose conditions (15 mmol/L). While GLP-1 7–36 enhanced GSIS, liraglutide showed no effect. Notably, sitagliptin co-treatment attenuated GLP-1 7-36-mediated GSIS enhancement (Supplementary Fig. 2H), suggesting a role for GLP-1 9–36 in this process. Indeed, direct GLP-1 9–36 treatment augmented GSIS at both basal (6 mmol/L) and high glucose (20 mmol/L) concentrations (Supplementary Fig. 2I). Using mouse islet perifusion, we compared insulin secretory dynamics in response to GLP-1 7–36, GLP-1 9–36 (both 25 nmol/L), and their combination with sitagliptin. Sitagliptin inhibition of GLP-1 7–36 degradation reduced GSIS compared to GLP-1 7–36 alone, aligning with the enhanced GSIS observed with GLP-1 9–36 treatment (Supplementary Fig. 2J). Notably, the differential GSIS response to GLP-1 7–36 versus liraglutide seen in normoglycemic donor islets was absent in glucose-intolerant donor islets, where both peptides enhanced GSIS (Supplementary Fig. 2K) with comparable increases in intracellular Ca²⁺ (Supplementary Fig. 2L) and dose-dependent cAMP elevation (Supplementary Fig. 2M,N).

Tanycytic Transport of Liraglutide Is Required for Glucose-Induced Insulin Release in Chow-Fed Mice, but not in glucose-intolerant mice fed a high-fat diet.

Previous studies, including our own, have demonstrated that liraglutide increases plasma insulin levels in healthy mice^2,55, despite not directly enhancing GSIS in isolated pancreatic islets under these metabolic conditions (Fig. 1A,B). To investigate liraglutide's hypothalamic-mediated effects on insulin release and glucose homeostasis, we employed two mouse models: GLP-1r^TanycyteKD mice with tanycyte-specific GLP-1R ablation, and iBot mice expressing botulinum toxin serotype B-light chain protein in tanycytes to suppress their transport function^15,56. Previous studies have established liraglutide at 0.3 mg/Kg effectively reduces glycemia in both chow- and high-fat diet mice^24,57–59. A single intraperitoneal dose of liraglutide reduced glycemia during oral glucose tolerance tests (OGTTs) in both control and GLP-1rTanycyteKD mice (Fig. 3A). However, while liraglutide increased plasma insulin levels at 15 minutes post-glucose challenge in control mice, this effect was absent in GLP-1r^TanycyteKD mice (Fig. 3B). iBot mice showed similar patterns of glycemic response and insulin secretion compared to controls (Supplementary Fig. 3A,B). These findings demonstrate that liraglutide’s enhancement of plasma insulin release in healthy mice depends on tanycytes-mediated central transcytosis, while its glucose-lowering effects persist even when transcytosis is blocked. Previous studies have identified the autonomic nervous system as a key mediator for specific GLP-1 effects^60,61. To investigate the vagus nerve’s role in liraglutide-mediated plasma insulin elevation during OGTTs⁶² – given its central function in brain-pancreas-gut axis regulation⁶³, we performed a celiac vagotomy (CVGX), severing central connections from the esophagus to stomach and pancreas, with sham-operated mice serving as controls. Intraperitoneal liraglutide reduced glycemia during OGTTs in both CVGX and sham-operated mice (Supplementary Fig. 3C). CVGX mice, unlike sham-operated controls, showed no increase in plasma insulin levels (Supplementary Fig. 3D), paralleling our observations in GLP-1r^TanycyteKD and iBot mice. We then evaluated liraglutide's dose-dependent effects on plasma insulin and glycemia by conducting OGTTs in chow-fed mice across a broad dose range (0.006–0.4 mg/kg body weight). All tested doses reduced glycemia both during fasting and after glucose challenge. The EC50 for liraglutide's glucose-lowering effect was 0.01 mg/kg, both during fasting and OGTT conditions (Supplementary Fig. 3E,F). At this EC50 dose, like at 0.3 mg/kg, liraglutide reduced OGTT glycemia in both control and GLP-1r^TanycyteKD mice (Supplementary Fig. 3G). However, the insulin-stimulating effect observed in control mice was absent in GLP-1r^TanycyteKD mice (Supplementary Fig. 3H). Under normoglycemic conditions, tanycytic GLP-1R-mediated transcytosis of liraglutide appears crucial for its central regulation of plasma insulin release^21,23, and sympathetic pancreatic outflow^23,64,65. However, in mice fed a high-fat diet for 12 weeks, liraglutide enhanced both glucose reduction and insulin secretion during OGTTs in both GLP-1r^TanycyteKD and control mice compared to saline-treated controls (Fig. 3C,D). Liraglutide's effect was mediated through direct pancreatic action, as islets isolated from both control and GLP-1r^TanycyteKD mice showed enhanced GSIS in response to the agonist (Fig. 3E). Extended high-fat feeding (27 weeks) led to significant elevations in fasting glucose (Supplementary 3I), fasting plasma insulin (Supplementary 3J), and insulin resistance - assessed by homeostatic model assessment of insulin resistance (HOMA-IR) (Supplementary 3K), compared to either 12-week high-fat-fed or chow-fed mice. Under these conditions, liraglutide treatment reduced glycemia and failed to enhance plasma insulin levels in both genotypes (Fig. 3F,G). Notably, isolated islets from both control and GLP-1r^TanycyteKD mice maintained liraglutide-responsive GSIS after 27 weeks of high-fat feeding (Fig. 3H), indicating preserved islet responsiveness despite impaired in vivo regulation.

Liraglutide Reduces Glycemia Through Decreased Hepatic Gluconeogenesis and Enhanced Peripheral Glucose Uptake Independent of Plasma Insulin Elevation

Liraglutide acutely lowered fasting glycemia without elevating plasma insulin in lean, normoglycemic mice (Fig. 4A, B), suggesting its metabolic effects might be mediated through altered gluconeogenesis or enhanced peripheral glucose uptake. To test gluconeogenic regulation, we performed pyruvate tolerance tests (PTTs) in overnight-fasted wild-type mice with or without liraglutide treatment. Pyruvate administration increased plasma glucose levels in fasting mice, confirming active endogenous glucose production (Fig. 4C). Liraglutide-treated mice showed reduced glycemia and attenuated pyruvate-induced glucose elevation compared to saline controls (Fig. 4C). While hepatic glucose production is regulated by hypothalamic POMC neurons⁶⁶, and central GLP-1 administration reduces hepatic glucose production⁶⁷, liraglutide suppressed pyruvate-induced gluconeogenesis equally in GLP-1rTanycyteKD and wild-type mice (Fig. 4C). We further analyzed hepatic gluconeogenic gene expression in wild-type and GLP-1rTanycyteKD mice following overnight fasting and treatment with either liraglutide or saline. Liraglutide significantly reduced Pgc and G6pase mRNA levels in both genotypes, while Pc and PcK1 expression remained unaffected (Fig. 4D). These findings indicate that liraglutide reduces fasting blood glucose in both genotypes by inhibiting hepatic gluconeogenesis⁶⁸, independent of tanycytic GLP-1R expression. To examine liraglutide's effects on tissue glucose uptake, we performed PET scanning to track [18F]2-fluoro-2-deoxy-d-glucose (18-FDG) utilization across tissues. Mice received intraperitoneal liraglutide (0.3 mg/kg) or vehicle, followed by an oral 18-FDG bolus (13.8 pg/kg) one hour later, and underwent PET imaging under isoflurane anesthesia (Fig. 4E), according to established protocols⁶⁹. PET analysis revealed increased global tissue 18-FDG utilization in liraglutide-treated mice (Fig. 4F), without affecting gastric emptying or renal excretion (Supplementary Fig. 4A, B). Specifically, liraglutide enhanced 18-FDG uptake in the gut, white and brown adipose tissues, and brain, but not in muscle (Fig. 4G). Given the influence of gut motility on peripheral glucose utilization⁷⁰, we examined liraglutide's effects on ex vivo isotonic contractions of gut segments following 30-minute pre-incubation (Fig. 4H). Liraglutide-induced glucose lowering (Fig. 4I) coincided with reduced ileal contraction amplitude (Fig. 4J), while duodenal, jejunal, and colonic contractions remained unchanged (Supplementary Fig. 4C-E). The ileum contains approximately 20% of mechanosensory gut neurons and vagal afferents from GLP-1-sensitive nodose ganglia neurons ^71,72. These findings suggest that liraglutide's insulin-independent glucose-lowering effect in the pre-prandial state may be partially mediated through ileal actions that promote tissue glucose utilization via gut-brain axis activation.

This study did not generate new unique reagents.

Human pancreatic islets were harvested from brain-dead adult human donors in the context of the traceability requirements for our clinical islet transplantation program (clinicaltrials.gov, NCT01123187, NCT00446264, NCT01148680). Islets isolated from lean, normoglycemic donors (BMI < 25, HbA1c < 5.7%), chronic smokers, obese normoglycemic donors (BMI > 30, HbA1c < 5.7%), glucose intolerant donors (HbA1c < 6.5%), newly diagnosed T2D (HbA1c < 7%) or overt T2D (HbA1c > 7%) were used in our study. The phenotypic characteristics of human islet donors are reported in Table 1. All human islet preparations underwent quality control assessment to demonstrate GSIS (Supplementary Fig. 1A). Islets were cultured in a glucose-free medium (Gibco, Life Technologies, Paris, France) supplemented with 0.625% HSA, 1% P/S, and 5.5 mM glucose. The dynamics of insulin secretion in response to different glucose and drug exposure were studied using a perifusion technique as previously described.⁹⁴ In brief, the KREB solution used during the experiment contained: 120 NaCl, 4.8 KCl, 2.5 CaCl2, 1.2 MgCl2, and 24 NaHCO3 mmol/L. It was continuously gassed with a mixture of O2/CO2 (94:6% ratio) and was supplemented with 1 mg/ml BSA and the temperature was maintained at 37°C. Approximately, 300 islet equivalents (IEQ) were transferred into two reaction chambers and preincubated with low glucose concentration (3 mmol/L) for 50 min. After equilibration time, islets were exposed to 1 mmol/L, 6 mmol/L, and 15 mmol/L glucose concentrations throughout the experiment with or without liraglutide (10 pmol/L or 25 nmol/L). Samples were collected every 2 min to measure insulin concentration in the effluent fractions. During all processes, outflow, pressure, temperature (37°C), and oxygen addition were maintained constant. At the end of the experiment, islets were recovered from the chambers and transferred into acid-ethanol for intracellular insulin extraction. Insulin concentrations were expressed as % of insulin content that was secreted per minute. For insulin secretory dose-response experiments to liraglutide, human islets were exposed for 1h to glucose-free RPMI 1640 media (Gibco, Netherlands) supplemented with 0.625%HA, 1% P/S, and 1 mmol/L, 6 mmol/L, and 15 mmol/L glucose concentrations in the presence or absence of several concentrations of liraglutide (0.01 nmol/L, 0.1 nmol/L, 0.5 nmol/L, 1 nmol/L, 10 nmol/L or 20 nmol/L) using static incubation techniques as described previously.⁹⁵ Each condition was performed in quadruple from each donor islet preparation and hormone secretion was normalized to the percentage of intracellular content. To study the effect of liraglutide on enhancing insulin secretion after chronic treatment with glucagon and glucose, human islets were cultured in CMRL media (Gibco, UK) supplemented with 5.5 mM glucose or 10 mM glucose with exogenous glucagon (10 pM). After 48 hours of culture, GSIS was performed by perifusion technique. Islets were stimulated with low (3 mmol/L), and high glucose (10 mmol/L) concentrations with or without liraglutide (25 nmol/L). Insulin secretion was normalized to intracellular insulin content and insulin secretion values were expressed as a percentage of content.

All mice were housed under specific pathogen-free conditions in a temperature-controlled room (21–22°C) with a 12h light/dark cycle and ad libitum access to food and water. All experiments were performed in 8–10 weeks-old male mice and were fed with either chow or high fat diet (D12492, Research Diet). All mice tissue and serum collection were kept at -80ºC

Male db/db mice (RRID:IMSR_TAC:DB), were obtained from Charles River. Tg(CAG-BoNT/B,EGFP)U75-56wp/J (BoNTB-EGFP^{loxP−STOP−loxP} ) mice have been engineered by Dr. Franck Pfrieger (University of Strasbourg, France; JAX Stock No. 018056; (RRID:IMSR_JAX:018056) as previously described.⁹⁶ Male C57Bl/6J (RRID:MGI:5657312) mice were purchased from Charles River. Animal studies were performed with the approval of the Institutional Ethics Committees for the Care and Use of Experimental Animals of the University of Lille and the French Ministry of National Education, Higher Education and Research (APAFIS#2617-2015110517317420 v5) and under the guidelines defined by the European Union Council Directive of September 22, 2010 (2010/63/EU).

Animal procedures at Imperial College London were approved by the British Home Office under the UK Animals (Scientific Procedures) Act 1986. Male C57BL/6J mice (Charles River, UK) were group housed and held in a condition-controlled room at 21–23°C with 12h:12h light to dark cycles (lights on at 07:00). All animals had ad libitum access to water and feed. Lean mice were fed standard chow (RM1(E); Special Diet Services, UK), and diet-induced-obese (DIO) mice, a 60% kcal high-fat diet (D12492; Research Diets, USA).

Tanycytic specific knockdown of GLP-1R (GLP-1r^TanycyteKD) was performed in isoflurane-anesthetized 8-week old WT C57Bl/6J male mice (Charles Rives), by stereotaxically injection of either AAV1/2-shRNA-GLP-1R (AAV1/2-EGFP-U6-mGLP-1R-shRNA (ACC GCG TCA ACT TTC TTA TCT TC ACT CGA GTG AAG ATA AGA AAG TTG ACG C TTTTT); serotype 1:2 chimeric; titer = 5.7 x 10¹⁰; Vector Biolabs) or AAV1/2-GFP (AAV1/2-EGFP-U6(SapI); serotype 1:2 chimeric; titer = 6.5 x 10¹⁰; Vector Biolabs) in the lateral ventricle (2 µl; at 0.2 µl/min; anteroposterior, -0.3 mm; midline, -1 mm; dorsoventral, -2.5 mm), 2 weeks before starting the experiments as previously described.¹⁵

A total of 20 mice, consisting of 10 control mice (WT) and 10 GLP-1r^TanycyteKD mice, were subjected to a high-fat diet (HFD) for either 12 or 27 weeks. These mice were divided into 4 groups, with each group consisting of 5 mice. Prior to drug administration, the mice underwent a fasting period. Subsequently, liraglutide (0.3 mg/kg) or saline was administered intraperitoneally (IP), and their blood glucose levels were monitored before and after the drug administration. After 1 hour, the mice were given a glucose bolus orally (2g/kg), and their glycemia was measured at regular intervals of 0, 15, 30, 60, and 90 minutes. Blood samples were collected before the glucose challenge and 15 minutes after the bolus to measure insulin levels. Blood was collected from the tail and stored in tubes using capillarity. Following centrifugation, plasma was separated and stored at -80°C until hormone measurement. Each experiment was repeated twice.

Islets were isolated from C57bl6J lean mice fed a chow or HFD, db/db mice, and GLP-1r^TanycyteKD mice. Each mouse was fully anesthetized and euthanized via cervical dislocation. The mouse was positioned with the abdomen facing upwards, and the skin was sterilized using 70% ethanol. An incision was made around the upper abdomen to expose the pancreas and common bile duct. Cold enzyme collagenase P was infused into the pancreas through the common bile duct. Following perfusion, the pancreas was extracted, placed in a 50 ml tube containing 2 ml of enzyme collagenase, and digested in a water bath at 37°C for 8–10 minutes. The enzymatic digestion was halted by adding cold Hanks' Balanced Salt solution with 1% albumin. To obtain a high-purity fraction of islets, a density gradient using polysucrose 1,132/1,108/1,096/1,069/1,000 (Mediatech) was performed. The isolated islets demonstrated a purity of over 90%, with minimal presence of exocrine tissue. They were then cultured in RPMI-1640 medium (Sigma Aldrich) supplemented with 2 g of glucose, 10% FBS, and 1% P/S for 18 hours before treatment. The perifusion technique, as described earlier for human islets, was utilized to study the insulin secretion dynamics in response to varying glucose and drug exposure.

At Imperial College London, pancreatic islets were isolated from lean or DIO C57BL/6J mice after 8 weeks on regular chow or high-fat diet, as above. Pancreata were inflated with RPMI-1640 medium (R8758, Sigma-Aldrich) containing 1 mg/mL collagenase from Clostridium histolyticum (S1745602, Nordmark Biochemicals), dissected and incubated in a water-bath at 37oC for 12 min. Islets were subsequently washed and purified using a Histopaque gradient (Histopaque-1119, 11191, Sigma-Aldrich, and Histopaque-1083, 10831, Sigma-Aldrich). Isolated islets were allowed to recover overnight in RPMI-1640 supplemented with 10% v/v FBS (F7524, Sigma-Aldrich) and 1% v/v Penicillin-Streptomycin solution (15070-063, Invitrogen).

The surgical procedure was performed in 6h fasted male C57Bl/6J mice under isoflurane anesthesia using standard aseptic procedures. After median laparotomy, the stomach was isolated outside the abdominal cavity, and all gastric connections to the spleen and liver were released. The left lobes of the liver were gently pulled aside, and the stomach was retracted caudally to expose the distal segment of the esophagus. At this level, the subdiaphragmatic left (anterior) vagal trunk was identified as descending parallel to the esophagus. The nerve was identified at the Hiss's angle. In mice receiving celiac vagotomy, all the connections around this area (between the esophagus, stomach [posterior wall], and the pancreas) were sectioned. In sham-operated animals, the left vagal branch was visualized but not sectioned. The stomach and the lobes of the liver were then repositioned back within the abdominal cavity. For all procedures, the laparotomy was closed using 4.0 polyglycolide sutures in two layers (peritoneum-aponeurosis and skin). A dose of xylocaine (1mg/kg) was applied along the sutures to reduce pain. Sham-operated and celiac vagotomised mice were fasted overnight (12 hours), after which they received an intraperitoneal injection of either liraglutide (Victoza, NovoFine®, Novo Nordisk) (0.3 mg/kg) or vehicle (0.9% NaCl solution)^24,57,58. One hour later an oral glucose tolerance test (OGTT) was performed (2g/kg glucose) and glycemia was monitored before and after drug and glucose administration. Blood samples were collected during fasting and 15 min after the glucose challenge for plasma insulin measurements. Samples were kept on ice during the experiment before being centrifuged (4°C, 600 rpm, 15 min) for sample collection and stored at -80°C until their use.

Human insulin concentrations were measured using the human DxI Access Immunoassay System (Beckman Coulter), and mouse insulin concentrations were measured using the mouse Insulin Elisa kit (Mercodia AB; Uppsala Sweden; RRID: AB_2636872), according to the manufacturer's instructions.

Imaging of whole-islet Ca²⁺ dynamics was performed 24 hours after isolation. Islets from individual C57BL/6J mice were pre-incubated for 1 hour in Krebs-Ringer Bicarbonate-HEPES (KRBH) buffer (140 mM NaCl, 3.6 mM KCl, 1.5 mM CaCl2, 0.5mM MgSO4, 0.5 mM NaH2PO4, 2 mM NaHCO3, 10 mM HEPES, saturated with 95% O2/5% CO2; pH 7.4) containing 0.1% w/v BSA (10775835001, Roche) and 6 mM glucose, and the Ca²⁺ responsive dye Calbryte590-AM (AAT Bioquest). Wide-field fluorescence imaging of islet Ca2 + responses was performed at 37C in black-walled microplates, using a modified Nikon Ti-2E with LED illumination, motorized stage and 10X air objective. After a baseline period at 6 mM glucose, liraglutide, glucose, or KCl were added to achieve the desired in-well concentrations of each stimulus. Typically, 30 islets were imaged per biological replicate. Fluorescence intensity traces from whole islet regions-of-interest (ROIs) after motion correction were extracted using Fiji (version 1.53t). Responses were double normalized, firstly as fractional change relative to baseline and then as a percentage of the maximum response to KCl.

Islets from C57BL/6J mice were dispersed by trituration in 0.05% trypsin-EDTA for 3 min at 37°C. The cell suspension was seeded onto black 96-well plates coated with 0.01% poly-D-lysine and 25 µg/ml mouse laminin (Thermo Fisher, UK) and transduced with the Green Up cADDis biosensor (10 µL per well; Montana Molecular, USA) according to the manufacturer’s instructions. 24 hours after viral transduction cells were washed and incubated in Kreb’s buffer with 6 mM glucose and 0.1% BSA for 30 min before the assay. Cells were imaged at 37°C using an automated epifluorescence microscope. Several wells were imaged in parallel with multiple FOVs per well acquired (comprising hundreds of cells) at set intervals before and after stepwise addition of agonist to achieve a 10-fold increase in concentration every 5 minutes. IBMX (500 µM) plus forskolin (FSK; 50 µM) were added to each well at the end of the incubation to maximally stimulate the sensor. A maximum intensity projection was derived and thresholded to define cell-containing regions. The fluorescence intensity was then measured at each timepoint and expressed as a fractional change from baseline. Individual cells that did not show a pre-defined response to IBMX/FSK were excluded. Intensity profiles were then normalised with the IBMX/FSK response set as 100%. The AUC for each concentration step was calculated and used to derive concentration-responses using 3-parameter fitting.

Human islets, obtained from donors with varying metabolic characteristics (as outlined in Table 1), were used for RNA extraction. The RNeasy Mini Kit (Qiagen, Courtaboeuf, France) was employed for total RNA extraction, and the concentration of each sample's RNA was measured using a Nano-drop. Subsequently, first-strand cDNA synthesis was carried out utilizing Superscript IV reverse transcriptase (Life Technologies, Lithuania) in accordance with the manufacturer's instructions. For quantitative RT-PCR, the Bio-Rad MyiQ Single-Color Real-Time PCR Detection System and Bio-Rad SYBR Green Supermix (Bio-Rad Laboratories, CA, USA) were employed. Primers for specific target genes (GLP-1R: Forward 5'-TTGTTCTGCAACCGGACCTT-3', Reverse 5'-CTCCTCGCACTCCGACAAGT-3'; INSULIN: Forward 5'-TGCATCAGAAGAGGCCATCA-3', Reverse 5'-CGTTCCCCGCACACTAGGTA-3'; RPL27: Forward 5'-TCTGGTGGCTGGAATTGACC-3', Reverse 5'-CCTTGTGGGCATTAGGTGATTG-3') were designed using Primer3 software and the Primer-design-tool from IDT Technologies. The final concentration of all primers used was 500 nM (250 nM forward + 250 nM reverse). The cDNA was diluted 1:10 with ultrapure water, and 2 µL of the diluted cDNA was added to each reaction, resulting in a final reaction volume of 10 µL. The Biorad CFX Connect Real-Time thermal cycler was utilized for amplification. Subsequently, a melt curve analysis was performed to assess primer and RT-PCR reaction quality. The gene expression levels were normalized to the housekeeping gene RPL27.

All PET (Positron Emission Tomography) scans were obtained using a microPET Inveon (Siemens Medical Solution, Knoxville, USA). After overnight fasting, one hour before the acquisition animals were intraperitoneally injected with either saline of liraglutide (0.3 mg/kg) and 5 minutes before the start of the PET scan, 0.72 ± 0.15MBq of 18F-FDG diluted in 0.2mL of water was orally administered in mice. For the PET experiments, the mice were anesthetized and maintained with a mixture of 2% isoflurane and air. PET data were acquired during one hour, all scans were performed in three-dimensional list mode, and acquisitions were reconstructed in 12 frames of 5 minutes. The CT (computerized tomography) scan was performed after PET imaging and provided anatomic and tissue attenuation coefficients. PET data were reconstructed with iterative algorithms of OSEM3D and corrected for attenuation and scatter. Imaging data analyses were performed on all frames by use of IRW software (IRW: Inveon Research Workplace version 4.2, Siemens Medical Solution).

In our study, peripheral tissue is defined as areas that do not include the stomach, intestine, bladder, and kidneys. The FDG measured in this peripheral tissue region was calculated as the difference between the FDG activity in the whole-body region and the sum of the FDG activities present in the stomach, kidneys, urinary bladder, and in intestine.