Liraglutide Increases Gastric Fundus Tonus and Reduces Food Intake in Type 2 Diabetic Rats

Abstract

Background/Objectives: Incretin-based therapies have demonstrated benefits in glycemic control and the prevention of long-term complications of diabetes. In addition to glucose-dependent insulin secretion stimulation, glucagon-like peptide-1 (GLP-1) also inhibits gastric acid secretion, delays gastric emptying, inhibits gut motility and induces satiety. We aimed to understand the modulation of gastric fundus motility by GLP-1 receptor agonists (GLP-1RA). Methods: We have studied the relaxation to sodium nitroprusside (SNP) and noradrenaline (NA) of gastric fundus isolated from Wistar rats and Goto-Kakizaki (GK) rats, an animal model of spontaneous non-obese type 2 diabetes, after Liraglutide treatment (200 μg/kg s.c., b.i.d., 14 days). Results: Decreased relaxation induced by SNP and NA (0.01–889 μM) was observed in treated groups, with no significant changes in SNP maximum relaxation or in nNOS/p-nNOS levels between treated and non-treated rats of both animal models. Accordingly, in rat gastric fundus pre-contracted with 5 µM of carbachol, GLP-1RA (0.05–111.1 nM) induced contractile responses that were GLP-1R-dependent and -independent. Exenatide showed more intrinsic activity, while Liraglutide showed more potency than GLP-1 in Wistar rats. Moreover, GLP-1 showed more intrinsic activity in diabetic rats compared to control ones. Conclusions: Liraglutide-induced increased gastric muscle tone may contribute to the significant decrease in caloric intake and body weight in all treated rats, causing a reduction in gastric accommodation during food intake. Thus, the increased gastric fundus tone induced by GLP-1RA may constitute a peripheral mechanism by which they can reduce food intake and induce satiety.

1. Introduction

Type 2 Diabetes Mellitus (T2DM) is the most common type of diabetes, accounting for a higher risk of cardiovascular diseases in later life according to the last International Diabetes Federation (IDF) report [1]. Projections of the global prevalence of diabetes in the 20–79 age group will be 853 million people worldwide in 2050 [1]. T2DM is a long-term metabolic disorder with high rates of socioeconomic costs, morbidity and mortality. Person-centered pharmacological interventions, lifestyle/behavioral changes and moderate-intensity physical activity are used in preventing or delaying it and improving other cardiometabolic markers [2].

Incretin-based antidiabetic therapies, namely glucagon-like peptide (GLP)-1 receptor agonists (GLP-1RA) and dipeptidyl peptidase-4 (DPP-4) inhibitors, have demonstrated benefits in glycemic control by stimulating nutrient-induced insulin secretion while inhibiting glucagon secretion. They also prevent diabetes long-term complications [3], although the cardioprotective effects associated with the GLP-1 metabolite, GLP-1 (9–36) amide—which is inactive at the GLP-1 receptor (GLP-1R)—are absent in DPP-4-resistant GLP-1 analogs and DPP-4 inhibitors [4]. This alerted the scientific community to the possibility that GLP-1’s actions could be mediated by GLP-1R-independent and -dependent mechanisms [5]. GLP-1RA are a particularly attractive choice for T2DM treatment because these drugs preserve β-cell function and have beneficial effects on hypertension, hyperlipidemia and obesity, being recommended as a preferred treatment for patients with pre-existing atherosclerotic cardiovascular disease and microvascular diabetic complications. Moreover, they do not cause side effects observed for other antidiabetic drugs, such as hypoglycemia and weight gain [6]. On the contrary, all approved GLP-1RA have the potential to induce weight loss by decreasing appetite and increasing satiety, mainly through an interaction with GLP-1R in brain areas involved in energy homeostasis. Indeed, murine models show that liraglutide— an acylated analog of human GLP-1, sharing 97% amino acid sequence homology and exhibiting non-covalent binding to serum albumin in circulation, thereby conferring high resistance to DPP-4-mediated degradation. [3,6,7,8]—can access specific brain areas involved in appetite regulation. It binds to GLP-1R on pro-opiomelanocortin and cocaine- and amphetamine-regulated transcript (POMC/CART) expressing neurons in the arcuate nucleus, thereby increasing satiety. GLP-1RA are approved for both long-term weight management and T2DM [9] and are associated with reductions in appetite and hunger, lower preference for energy-dense foods, alteration in food reward pathways, decrease in food craving, and improvement in eating control [10].

Corroborating these findings, GLP-1RA treatment has also been shown to delay gastric emptying within the first postprandial hour, reducing food intake and post-meal rises in glycaemia. However, tachyphylaxis has been described for long-acting GLP-1RA, such as once-daily subcutaneous liraglutide, suggesting additional mechanisms of action in GLP-1RA-mediated weight loss [11,12]. In contrast, short-acting GLP-1RA, such as twice-daily subcutaneous exenatide—which shares 53% amino acid sequence identity with endogenous GLP-1 and is also resistant to DPP-4-mediated degradation—have been associated with gastrointestinal adverse effects, including nausea, vomiting, and diarrhea, which appear to be transient with long-acting ones [6,13].

Besides pancreatic islets and peripheral and central nervous systems, the GLP-1R is expressed in the kidney, lung, heart, gastrointestinal tract, adipose tissue and smooth muscle [14,15,16]. In vascular smooth muscle, GLP-1-induced activation of the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB/Akt) signaling pathway leads to phosphorylation and activation of the endothelial isoform of nitric oxide synthase (eNOS), resulting in nitric oxide (NO) release and subsequent vascular smooth muscle relaxation [17].

In the gut, GLP-1 inhibits gastric acid secretion, delays gastric emptying, as mentioned above, and inhibits gut motility through activation of GLP-1R on enteric neurons, possibly via NO release [18,19,20,21,22]. However, its effect on the gastric fundus motility was never addressed before. Given that this region is responsible for accommodating ingested food and that reducing its tone will allow an increase in gastric volume, without a substantial increase in intragastric pressure [23], we hypothesized that GLP-1RA may modulate the contractile response of the gastric fundus, reducing gastric compliance and thereby limiting the accommodation of food in the stomach. This contributes to satiety through a peripheral mechanism independent of delayed gastric emptying.

The present study aimed to understand the mechanisms involved in the modulation of gastric fundus motility by GLP-1RA, with a focus on Liraglutide, as well as the alterations induced by T2DM in such mechanisms. To this end, relaxation to sodium nitroprusside (SNP) and noradrenaline (NA) was studied in gastric fundus isolated from Wistar rats and Goto-Kakizaki (GK) rats, an animal model of spontaneous non-obese T2DM, after Liraglutide treatment. The contractile responses of gastric fundus strips, isolated from untreated animals and pre-contracted with carbachol (CCh), to native GLP-1, the long-acting GLP-1A Liraglutide, and the short-acting agonist exenatide were further evaluated.

2. Materials and Methods

2.1. Drugs and Chemicals

The following drugs were used in functional studies: GLP-1 (7–36) amide, Exenatide (Exendin-4), Exendin-3 (9–39) amide, from Tocris, Bristol, UK; Liraglutide (Victoza^®) from Novo Nordisk, Bagsværd, Denmark; NA and SNP from Sigma-Aldrich, St Louis, MO, USA. Solutions of these drugs were prepared fresh on the day of the experiment with an adequate solvent.

The antibodies used in molecular studies were Calnexin (Sicgen, Coimbra, Portugal), GLP-1R (bs-1559R, Bioss, Woburn, MA, USA), nNOS and nNOS-S1417 (ab5586, ab5583, Abcam, Waltham, MA, USA). The secondary antibodies were anti-rabbit and anti-goat (Bio-Rad, Hercules, CA, USA).

Other reagents used in this study were purchased from Panreac (Barcelona, Spain) and Sigma-Aldrich (St. Louis, MO, USA).

2.2. Animals

Male Wistar and GK rats were obtained from our local breeding colonies at the Faculty of Medicine, University of Coimbra, Portugal. All animals were maintained under controlled environmental conditions (day-night cycles of 12 h, temperature of 22–24 °C and humidity of 50–60%) with ad libitum access to food (standard diet A03, SAFE^®, Augy, France) and water. The experimental protocol was approved by the local Institutional Animal Care and Use Committee, and all the procedures were performed by licensed users of the Federation of Laboratory Animal Science Associations (FELASA).

2.3. In Vivo Studies

Wistar and GK rats with 14 weeks of age were blinded allocated into four groups: Wistar (n = 7) and GK (n = 4) rats treated with Liraglutide (200 μg/kg s.c.), twice daily [24], for fourteen days [25] (WL and GKL, respectively); and Wistar (n = 7) and GK (n = 4) rats treated with saline (NaCl 0.9% s.c.) during the same period (WC and GKC, respectively). All animals were maintained under controlled environmental conditions with ad libitum access to food and water during the treatment.

Fasting glycemia, triglycerides, total cholesterol and intraperitoneal insulin tolerance test (ITT) were evaluated on the first and last days of treatment in blood from the tail vein and after a 6 h fasting period. For ITT an intraperitoneal injection of insulin (250 mU/kg) was performed, and the evaluation of glycaemia was performed at 0, 15, 30 and 60 min using a glucose meter and test strips (Accu-Chek Aviva, Roche, Basel, Switzerland). The area under the curve (AUC) was calculated. Triglycerides and total cholesterol were evaluated through Accutrend Plus Meter (Roche, Basel, Switzerland). Body weight and caloric intake were evaluated daily and weekly, respectively.

2.4. Functional Studies

2.4.1. Relaxation Induced by SNP or NA of Gastric Fundus Strips Isolated from Treated and Non-Treated Rats with Liraglutide

At day 15, at the end of treatment with Liraglutide, animals were anesthetized with ketamine (75 mg/kg body weight; Pfizer Inc., New York, NY, USA)/chlorpromazine (3 mg/kg body weight; Vitória Laboratories, Amadora, Portugal) and then sacrificed by cervical displacement.

The stomach was isolated, and the fundus region was dissected and washed in ice-cold Krebs–Henseleit solution (mM: NaCl 118.67; KCl 5.36; MgSO₄·7H₂O 0.57; CaCl₂·2H₂O 1.90; KH₂PO₄ 0.90; NaHCO₃ 25; glucose 11.1), pH 7.4, aerated with 5% CO₂–95% O₂. Four longitudinal fundal strips (approximately 15 mm) were obtained, following the methodology described by Riazi-Farzad et al. [26] and based on that firstly described by Vane [27], and suspended on stainless steel hooks under 19.6 mN of tension in 10 mL organ baths (Panlab, Barcelona, Spain) filled with aerated Krebs–Henseleit solution maintained at 37 °C. Gastric fundus strips were allowed to equilibrate for 2 h and then were submitted to isometric relaxations induced by two successive cumulative concentration-response (CR) curves for SNP and NA (0.01–889 μM) after pre-contraction with CCh 5 μM [28], using Panlab isometric transducers (Barcelona, Spain) connected to a four-channel polygraph (Polygraph 4006, Letica Scientific Instruments, Barcelona, Spain). The two CR curves were performed with a time interval of one hour to avoid tachyphylaxis.

2.4.2. Responses to GLP1-RA of Pre-Contracted Gastric Fundus Strips Isolated from Non-Treated Rats

Following the above-described procedure, gastric fundus strips isolated from non-treated Wistar (n = 4) and GK (n = 4) rats were mounted in organ baths, and after the equilibration period of 2 h, with periodic washings, isometric contractions of cumulative CR curves for GLP-1 (0.05–111.1 nM) were recorded after pre-contraction with CCh 5 μM (Figure 1). The efficacy and potency of the two drugs, exenatide and Liraglutide, were also evaluated in comparison to the endogenous agonist using the same experimental protocol. For adequate pharmacological characterization, a second cumulative CR curve of the full agonist GLP-1 was performed in the presence or absence of Exendin-3 (300 nM), a selective GLP-1R antagonist added thirty minutes before.

2.4.3. Analysis of Results

Relaxation responses to SNP and NA were expressed as a percentage of inhibition of the pre-contraction obtained with CCh. Contractile responses to GLP-1RA were expressed as a percentage of the pre-contraction obtained with CCh and were also analyzed in terms of mN of tension. Contractile responses of the GLP-1 second CR curve (performed in the presence or absence of Exendin-3) were expressed as a percentage of the maximum contraction obtained in the first CR curve of the respective strips. Considering that in each assay, gastric fundus strips pre-incubated with the appropriate solvent of each drug were used as controls, results were then expressed as a percentage of the control strips maximum response.

In both experimental protocols, the maximum contractile/relaxation response (E_max) and the pEC₅₀ (negative logarithm of the molar concentration of agonist inducing half maximum response) were determined. E_max and pEC₅₀ values represent, respectively, the intrinsic activity and the potency of a drug in eliciting a response. Values of pEC₅₀ were obtained by interpolation in each CR curve in a half-logarithmic scale (percentage of maximum contraction/relaxation vs. logarithm of concentration in mol/L), using a computer program (CurveExpert 2.2 version for Windows, Hyams Development, London, UK).

Differences between NO donor or NA E_max and pEC50-values for the four animal groups (control and diabetic rats treated and non-treated with Liraglutide) and for different GLP-1RA were statistically evaluated.

Differences between dose-responses to CR curves for native GLP-1 performed in the absence or in the presence of Exendin-3 were also analyzed.

2.5. Western Blot Analysis

Gastric fundus samples isolated from the four groups of animals were cut into small pieces and were homogenized by mechanical disruption with lysis buffer (0.25 M Tris HCl, 125 mM NaCl, 1% Triton-X-100, 0.5% SDS, 1 mM EDTA, 1 mM EGTA, 20 mM NaF, 2 mM Na₃VO₄, 10 mM β-glycerophosphate, 2.5 mM of sodium pyrophosphate, 10 mM of PMSF and 40 µL of protease inhibitor) followed by centrifugation (14,000 rpm, 20 min, 4 °C). The supernatants were collected, and a second centrifugation was performed. Protein concentration of the lysates was determined using the BCA Protein Assay Kit (Bio-Rad, Hercules, CA, USA). Then the supernatants were mixed with Laemmli buffer (62.5 mM Tris-HCl, 10% glycerol, 2% SDS, 5% β-mercaptoethanol, 0.01% bromophenol blue). Polyacrylamide gels (SDS-PAGE) were loaded with the obtained samples, transferred to polyvinylidene fluoride (PVDF) membranes, which were washed and blocked with TBS-T 0.01% and 5% BSA, and incubated overnight with the primary antibodies followed by a 2 h-incubation with secondary antibodies. Membranes were revealed using ECL substrate in a Versadoc system (Bio-Rad, Hercules, CA, USA) and analyzed with Image Quant^® (Molecular Dynamics, San Jose, CA, USA). To confirm equal protein loading, membranes were reprobed for calnexin and results were normalized against calnexin and then expressed as a percentage of the WC group protein density.

2.6. Immunohistochemistry

Gastric fundus samples isolated from animals of all experimental groups were also fixed in 4% formalin buffered at pH 6.9 and embedded in paraffin. Histological sections (3 μm thick) were deparaffinized in a BOND™ Dewax solution (Leica Biosystems, Chicago, IL, USA), rehydrated in 100% alcohol and then washed with a BOND™ Wash solution (Leica Biosystems). Antigen retrieval was performed using sodium citrate buffer 10% (v/v), at pH 6, and endogenous peroxidase quenching was performed using 3–4% (v/v) hydrogen peroxide (Novocastra, Peroxidase Block solution, Leica Biosystems).

Sections were then incubated with the primary rabbit polyclonal anti-GLP-1R and secondary antibodies diluted in suitable solutions (Leica Biosystems) and revealed with the chromogen diaminobenzidine (Mixed DAB Refine BOND™, Leica Biosystems) for 10 min. Finally, slides were counterstained with hematoxylin (Hematoxylin BOND™, Leica Systems) for 5 min, followed by diaphanization, dehydration and mounting in a DPX synthetic medium. Immunostaining analysis was performed using an optic microscope Nikon Eclipse 80i (Nikon, Tokyo, Japan). Histological sections were also stained with standard hematoxylin/eosin for microanatomy analysis.

2.7. Statistical Analysis

Results are presented as mean ± standard error of the mean (S.E.M.) of the number (n) of experiments indicated. Given the relatively small sample size for biological and biochemical studies (n = 4–7 animals per group) and for functional studies [s (number of gastric fundus strips tested) = 5–16 of n = 3–4 animals per group], the non-parametric Kruskal–Wallis test (all pairwise multiple comparation) was applied to determine all statistical differences between more than two groups and statistical differences between two groups were assessed using the non-parametric Mann–Whitney test for unpaired data and the Wilcoxon signed-rank test for paired data using the IBM SPSS 29 software (IBM, Armonk, NY, USA) and GraphPad Prism 9 PC Software (Boston, MA, USA). The alpha level of significance for all experiments was 0.05 and p < 0.05 was considered as the criterion for significance.

3. Results

3.1. Liraglutide Improves Metabolic Profile in Diabetic GK Rats

Liraglutide treatment significantly reduced (p < 0.05) caloric intake in Wistar and GK rats (Figure 2B), resulting in body weight loss as demonstrated by a significant reduction in body weight gain during the treatment period (WC = 0.2% ± 0.5 vs. WL = −9.6% ± 0.7, p < 0.01 and GKC = 1.2% ± 0.6 vs. GKL = −5.4% ± 2.3; p < 0.05) (Figure 2A). Regarding the analyzed biochemical parameters, comparison of initial and final values within the same animals using the Wilcoxon signed-rank test for paired data (for which § p < 0.05 means significance) revealed that Liraglutide treatment in GK rats (GKL group) led to a significant reduction only in triglyceride levels (GKL Final = 111.3. ± 6.6 vs. GKL Initial = 161.8 ± 18.4, n = 4) (Figure 2E). Additionally, we observed potential off-target effects in Wistar control and Liraglutide-treated animals with respect to the AUC during ITT (Figure 2D).

When comparing the four groups at the same time point using the Kruskal–Wallis test, the Wistar control group was used as the main reference. In this context, at baseline (Initial), GK rats exhibited significantly elevated fasting blood glucose (p < 0.01) and triglyceride levels (p < 0.05) compared to Wistar control rats at the same time point (Figure 2C and 2E, respectively). Following treatment with Liraglutide (GKL group), both fasting glycemia and triglyceride concentrations were reduced, as demonstrated by the lower statistical significance of the differences in fasting glycemia (p < 0.05), and the already mentioned significant reduction in triglyceride levels by paired data analysis. Furthermore, in untreated diabetic rats (GKC group), we observed a trend toward a worsening lipid profile (triglycerides and cholesterol) over time (Figure 2E,F, respectively), which, in contrast, may further support the beneficial effects of Liraglutide observed in the GKL group. Additionally, at baseline, GK rats exhibited reduced insulin sensitivity, as indicated by a significantly higher AUC during ITT compared to Wistar control rats at the same time point (GKL = 191.3 ± 34.1 vs. WC = 84.21 ± 2.6, p < 0.05); this difference was no longer observed following Liraglutide treatment (Final) (GKL = 143.1 ± 7.9 vs. WC = 75.41 ± 1.9, p > 0.05) (Figure 2D). Although this may suggest an improvement in insulin sensitivity in the GKL group, no definitive conclusions can be drawn based solely on the paired data analysis.

Liraglutide also altered the levels of GLP-1 receptor in the gastric fundus. A significant decrease in GLP-1R protein levels was observed in gastric fundus samples isolated from Wistar (p < 0.05) and GK (p < 0.05) rats treated with Liraglutide compared to non-treated ones. No significant changes were observed between control and diabetic rats (Figure 3A). GLP-1 receptor-positive cells were found in ganglion and capillary endothelial cells, but not on smooth muscle cells of muscularis mucosa and endothelial cells of arterioles (Figure 3B and Figure S1).

3.2. Liraglutide Treatment Reduces NO- and NA-Induced Relaxation of Gastric Fundus from Wistar and Diabetic GK Rats

SNP and NA induced smooth muscle relaxation in rat gastric fundus strips of both Wistar and GK rats in a concentration-dependent manner. Unexpectedly, Liraglutide treatment reduced the relaxation response induced by both substances when compared to the non-treated groups (Figure 4A,B). In SNP-induced relaxation, no statistically significant changes were observed in the maximum response (E_max) elicited by this NO donor between the four experimental groups in a multiple comparison test; however, a significant reduction in its potency (pEC₅₀) was observed in Liraglutide-treated diabetic rats (GKL) compared to non-treated Wistar and GK rats (

Table 1. E_max and pEC₅₀ values for SNP and NA in gastric fundus isolated from Wistar and GK rats treated and non-treated with Liraglutide.

	Sodium Nitroprusside (SNP)			Noradrenaline (NA)
	Emax (%CCh)	pEC50	s/n	Emax (%CCh)	pEC50	s/n
WC	64.64 ± 7.06	6.26 ± 0.16	14/4	85.38 ± 5.21	6.12 ± 0.09	16/4
WL	43.60 ± 4.92	6.12 ± 0.12	15/4	61.63 ± 5.22 *	5.7 ± 0.15 *	16/4
GKC	55.75 ± 8.61	5.85 ± 0.10	15/4	60.44 ± 4.86 *	5.62 ± 0.06 **	14/4
GKL	47.42 ± 5.12	5.22 ± 0.09 **** ###	16/4	72.20 ± 6.10	5.27 ± 0.07 ***	16/4

E_max = maximum relaxation in percentage of inhibition of CCh induced contraction; pEC₅₀ = negative logarithm of the molar concentration of SNP/NA inducing half maximum relaxation; s/n = number of gastric fundus strips tested (s)/number of rats (n) from whom tissues were obtained; WC = Wistar rat injected with saline; WL = Wistar rat injected with Liraglutide (200 μg/kg s.c., twice daily for 14 days); GKC = Goto-Kakizaki rat injected with saline; GKL = Goto-Kakizaki rat injected with Liraglutide (200 μg/kg s.c., twice daily for 14 days). Results are presented as mean ± S.E.M of the indicated tissue strips isolated from n animals. Statistical differences were evaluated by the Kruskal–Wallis test. * p < 0.05, ** p < 0.01, *** p < 0.005, **** p < 0.001 vs. WC; ^### p < 0.005 vs. GKC.

). On the other hand, a significant decrease in maximum relaxation response induced by NA was observed in Wistar-treated rats compared to non-treated ones (p < 0.005), but not between GK-treated and non-treated rats (Table 1). Furthermore, NA showed less potency in inducing relaxation of gastric fundus strips from Wistar and GK rats treated with Liraglutide (Table 1). Regarding levels of nNOS and its phosphorylated form (p-nNOS), no significant changes were observed between the gastric fundus of treated and non-treated rats (in both animal models) (Figure 4C).

3.3. GLP-1 Receptor Agonists Induce Distinct Contractile Responses of Rat Gastric Fundus Through GLP-1R-Dependent and -Independent Pathways

Due to the results obtained with the relaxation of the gastric fundus induced by a NO donor and NA, we decided to assess whether GLP-1RA would induce contraction under the same experimental conditions. Also unexpectedly, GLP-1, Exenatide and Liraglutide, caused tonic contractions of gastric fundus strips from both animal models pre-contracted with CCh 5 μM in a concentration-dependent (0.05–111.1 nM) manner (

Table 2. E_max and pEC₅₀ values for GLP-1RA in the gastric fundus isolated from Wistar and GK rats.

	Emax (mN)		pEC50		s/n
	Wistar	GK	Wistar	GK	Wistar	GK
GLP-1	5.61 ± 0.69	9.09 ± 1.44 *	8.33 ± 0.16	8.58 ± 0.16	16/4	12/4
Exenatide	11.67 ± 1.08 ####	10.37 ± 1.99	8.78 ± 0.16	8.92 ± 0.21	11/3	12/3
Liraglutide	4.69 ± 0.81 §§§§	3.76 ± 1.01 # §§	9.62 ± 0.24 #### §	9.06 ± 0.51	13/4	5/3

E_max = maximum contraction in mN of tension; pEC₅₀ = negative logarithm of the molar concentration of agonist inducing half maximum contraction; s/n = number of gastric fundus strips tested (s)/number of rats (n) from whom tissues were obtained. Results are presented as mean ± S.E.M. Statistical differences were evaluated by the Kruskal–Wallis test and Mann–Whitney test. * p < 0.05 vs. Wistar; ^# p < 0.05, ^#### p < 0.001 vs. GLP-1; ^§ p < 0.05, ^§§ p < 0.01, ^§§§§ p < 0.001 vs. Exenatide.

). The Exenatide showed more intrinsic activity (E_max) and Liraglutide more potency (pEC₅₀) than GLP-1 in Wistar rats. Moreover, in GK rats, GLP-1 showed more intrinsic activity (E_max) than Liraglutide, and comparing both animal models, GLP-1 showed more intrinsic activity (E_max) in diabetic rats than in control ones (Table 2). Accordingly, the selective GLP-1 receptor antagonist Exendin-3 (300 nM) caused a significant reduction (p < 0.01) of the GLP-1 maximum effect only in GK rats [GLP-1 + Exendin-3 (W) = 107.79% ± 13.48, n = 8–7 strips/4 rats and GLP-1 + Exendin-3 (GK) = 55.08% ± 9.25, n = 5–6 strips/4 rats], without significant effect in potency (Figure 5). To rule out the involvement of acetylcholine, the primary neurotransmitter responsible for gastrointestinal contractility, we also conducted CR experiments for Liraglutide, in the presence and absence of 10 µM atropine, a non-selective muscarinic receptor antagonist, obviously using non-pre-contracted gastric fundus strips isolated from Wistar rats. Our findings suggest that the Liraglutide-induced tonic contractions are not of cholinergic origin (see Figure S2a,b).

4. Discussion

In the present study, we have focused on the pharmacological mechanism underlying the modulation of gastric fundus motility by GLP-1RA, since peripheral GLP-1 is known to delay gastric emptying and small intestinal motility and secretion through an interaction with brain centers or afferent neural pathways relaying to the vagal motor nuclei [29,30,31]. We have also evaluated the alterations of gastric fundus motility induced by this class of incretin mimetics in a rat model of T2DM, the GK rat.

In rats treated with Liraglutide (200 μg/kg, twice daily for fourteen days), we demonstrated that administration of this long-acting GLP-1A improved the lipidic profile in diabetic rats and caused a significant reduction in caloric intake and body weight in both diabetic and non-diabetic Liraglutide-treated animals. Our results are similar to those previously reported with obese and/or diabetic rats treated with Liraglutide [24,25,32]. Unexpectedly, Liraglutide reduced the relaxation induced by SNP and NA in gastric fundus strips isolated from treated animals, compared to non-treated groups, with effects differing depending on the presence or absence of diabetes. With respect to NA-induced relaxation, Liraglutide treatment significantly reduced both the intrinsic activity and potency of NA in gastric fundus strips from control and diabetic rats. Regarding NO-induced relaxation, no significant changes were observed in the maximum relaxation to SNP; only a significant reduction in its potency was observed, caused by the disease and treatment. These pharmacological findings should indicate a greater increase in NO concentration to achieve the same level of efficacy. However, no differences were observed in the levels of total and phosphorylated forms of nNOS in either animal model. This contrasts with our histological findings in the rat gastric fundus, which revealed GLP-1R-positive staining in ganglia, suggesting a neuronal role for GLP-1R, and with previous reports indicating the loss of nitrergic neurons—and consequently reduced NO synthesis—in the gastrointestinal tract of various diabetic models [33,34,35], as well as studies showing NO-dependent inhibition of gastrointestinal motility by GLP-1 [18,19,20,21,22]. Such differences should be further investigated.

In view of these results, which question the role of GLP-1RA as facilitators of NO and sympathetic-induced gastric emptying, we decided to verify whether GLP-1RA would induce contraction instead, under the same experimental conditions. Interestingly, the tested GLP-1RA caused tonic concentration-dependent contraction of the gastric fundus isolated from Wistar and GK rats after a CCh-induced pre-contraction, Liraglutide being the most potent of the agonists tested, but the least effective, especially compared to the short-acting GLP-1RA exenatide.

Additionally, native GLP-1 showed more intrinsic activity in diabetic rats than in control ones, which is consistent with the observed trend for higher GLP-1R protein levels in the gastric fundus of GK rats. This may reflect a compensatory mechanism resulting from lower endogenous GLP-1 production in this animal model, as already reported [36]. However, other reports point to a decrease in disease states. For instance, reduced GLP-1R expression in gastric glands of patients with T2DM was observed by Broide et al. [37] and significantly lower Glp-1r expression levels in endothelial and smooth muscle cells were observed in obese type 2 diabetic db/db mice compared to non-diabetic db/m mice [38]. Very little data has been reported regarding the effects of GLP-1RA treatment on receptor expression that can support our reduction in GLP-1R protein levels by Liraglutide treatment, regardless of the presence or absence of diabetes. In a report by Sanada et al. [39] with the long-acting incretin mimetic dulaglutide, Glp-1r expression varies depending on whether it is an early or late pharmacological intervention. Indeed, Glp-1r expression in the abdominal aorta of streptozotocin-induced diabetic Apoe knockout mice was reduced, becoming comparable to non-diabetic mice if an early intervention dulaglutide (twice weekly for 8 weeks) treatment was performed. However, after a late treatment (twice weekly for 16 weeks), Glp-1r expression was lower in controls, and dulaglutide failed to recover it. On the other hand, the reduced GLP-1R protein levels observed in Liraglutide-treated Wistar and GK rats may help explain the tachyphylaxis described for long-acting GLP-1RA on gastric emptying [11,12]. Although GLP-1RAs initially inhibit gastric motility, their efficacy in maintaining this effect diminishes over time, and the underlying mechanisms responsible for this attenuation remain unclear. All GLP-1RAs effectively slowed gastric emptying in the early stages of treatment, but within 1–2 weeks, their efficacy waned, and it has been suggested to be associated with a decline of Glp-1r mRNA expression in neurons of the nucleus of the solitary tract which plays a fundamental role in regulating metabolism and gastric motility [40].

Finally, the contractile effects of GLP-1RA on the rat gastric fundus are not mediated by acetylcholine—the primary excitatory neurotransmitter in the gastrointestinal tract—as they are unaffected by the non-selective muscarinic receptor antagonist atropine and persist in the presence of pre-contraction induced by CCh. On the other hand, they are unexpected, considering the GLP-1RA vasorelaxation via endothelium-derived NO, which has been described in the cardiovascular system. Despite the ongoing debate regarding the presence of GLP-1R in vascular smooth muscle cells (VSMCs) and endothelial cells (ECs), several studies have demonstrated that direct exposure of VSMCs and ECs to native GLP-1 or GLP-1RA increases eNOS phosphorylation and NO production via a 5′AMP-activated protein kinase (AMPK)-dependent mechanism [17,41]. However, and according to other studies exploring the location of GLP-1R in human, rat and monkey [23,37,42], our histological analysis of the rat gastric fundus did not reveal GLP-1R immunoreactivity in smooth muscle cells nor endothelial cells of arterioles, thus providing no anatomical evidence to support such a relaxation effect. In fact, it has been recognized that GLP-1 modulates antro-pyloro-duodenal motility by stimulating pyloric motility and inhibiting antral motility [43]. The precise mechanism by which GLP-1 does this is not entirely clear.

Moreover, our data indicates that the contractile response of the gastric fundus to GLP-1RA involves both GLP-1R-dependent and -independent mechanisms, as evidenced by the significant reduction in the maximum effect of GLP-1 following incubation with the selective GLP-1R antagonist Exendin-3 in GK rats, but not in control animals. This would not be unprecedented since some GLP-1 effects in the cardiovascular system were suggested to be independent of this receptor [5]. Supporting this, animal studies have shown no direct effect of GLP-1 on gastric smooth muscle cells [44]. This reinforces the notion that suppression of vagal input from neural afferents is involved in the inhibition of gastric motor functions by exogenous GLP-1 [29,30,31]. Additional studies are needed to elucidate the mechanisms mediating GLP-1-induced gastric motor effects in vivo.

This study provides evidence of a contractile response to GLP-1RA in rat gastric fundus that is partially mediated by GLP-1R. In diabetic rats, the intrinsic activity of GLP-1 was higher, which may be attributed to a compensatory mechanism induced by T2DM, as suggested by a trend toward increased levels of receptor proteins. The significant reduction in GLP-1R protein levels following Liraglutide treatment was associated with diminished NA- and SNP-mediated relaxations, contributing to an overall increase in gastric fundus tone. Thus, given the satiating effect of gastric distension [45], our results suggest that the increased gastric fundus tone induced by this long-acting GLP-1RA may reduce gastric accommodation during a meal. These mechanisms may contribute to the GLP-1RA-induced reduction in food intake, alongside the well-established central pathways, and support the therapeutic benefits of GLP-1RAs in the management of obesity and T2DM.