The number of patients with diabetes worldwide has increased to 537 million, with a 3.6-fold increase occurring in the last 20 years. The prevalence of diabetes has reached 10.5%, affecting 1 out of 10 adults, and the global diabetes population is predicted to increase to 783 million by 2045 (IDF Diabetes Atlas 10th edition). Common reasons for the increased prevalence of type 2 diabetes include a lack of exercise and physical inactivity and greater numbers of overweight and obese individuals.
Diabetes is commonly associated with gastrointestinal symptoms, such as nausea, vomiting, diarrhea, abdominal pain, and constipation (Bytzer et al., 1996; Maleki et al. 2000), and affects most parts of the gastrointestinal tract. Gastric motility is impaired in the stomach of diabetics, with delayed gastric emptying and gastroparesis being reported (Camilleri, 2007; Intagliata & Koch, 2007). Furthermore, it was reported that diabetes mellitus is associated with structural changes in the connective tissue matrix and in the muscles in the wall of intestine and colon and further causes biomechanical remodeling (Zhao et al., 2017). Although the diabetes-induced dysfunctions in the small and large intestines are becoming clearer (Horváth et al., 2015), the effects of diabetes on the functions of the small intestine and colon have yet to be clarified in detail. The prevalence of constipation and diarrhea is high in patients with diabetes. A previous study suggested that the intestinal transit time was slower in animal models of diabetes, leading to bacterial overgrowth and, ultimately, diarrhea (Sellin & Hart, 1992), while other studies reported an accelerated intestinal transit time in some models (Rosa-e-Silva et al., 1996) and the induction of oxidative stress (Yarandi & Srinivasan, 2014). The colonic transit time is often increased, and constipation is a common symptom in patients with diabetes mellitus (Imaeda et al., 1998).
Due to the overwhelming number of patients with diabetes, the development of antidiabetic drugs is still actively conducted. Some traditional diabetes medicines include biguanides, a-glucosidase inhibitors, and thiazolidinediones, while others are relatively new, such as dipeptidyl peptidase-4 (DPP-4) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, and sodium–glucose cotransporter 2 (SGLT2) inhibitors, which were launched after 2000.
Although the main route of administration of GLP-1 receptor agonists is subcutaneous, orally administered GLP-1 receptor agonists (Ozempic, semaglutide, Novo Nordisk Pharma Ltd.) were approved in the United States in 2019 and in Japan in 2020. Since GLP-1 receptor agonists have been shown to not only improve blood glucose levels but also reduce body weight (Okamoto et al., 2021), they are used as diet medications in the United States and may be taken by non-diabetic patients. Therefore, the use of orally administered formulations of GLP-1 receptor agonists will continue to increase in the future.
GLP-1 receptor agonists exert inhibitory effects on oxidative stress-induced disorders, such as renal ischemia-reperfusion injury (Zheng et al., 2019; He et al., 2020; Tiba et al., 2023); therefore, they may also be used to treat digestive diseases caused by diabetes-induced oxidative stress. On the other hand, the administration of GLP-1 receptor agonists for weight loss has been shown to delay gastric emptying in non-diabetic patients (Klein & Hobai, 2023), and GLP-1 receptor agonists were found to be cytotoxic at high concentrations in a human neuroblastoma cell line (Salles et al., 2023). These findings indicate that GLP-1 receptor agonists exert negative effects on the gastrointestinal tract and, thus, may promote digestive disorders associated with diabetes. Therefore, the effects of GLP-1 receptor agonists on the gastrointestinal tract are controversial. Since these effects were observed from a vascular approach after subcutaneous administration, limited information is currently available on their effects from an apical approach of gastrointestinal epithelial cells (Takizawa et al., 2022). The bioavailability of orally administered formulations of GLP-1 receptor agonists is less than 1% (Overgaard et al., 2021), indicating that the majority of an ingested GLP-1 receptor agonist will remain in the gastrointestinal tract and continue to be exposed to gastrointestinal epithelial cells. Therefore, the effects of GLP-1 receptor agonists on gastrointestinal epithelial cells need to be clarified in order to confirm their safety and efficacy. The present study investigated the effects of GLP-1 receptor agonists on gastrointestinal epithelial cells, particularly barrier functions.
Ozempic® (semaglutide, M.W.: 4113.6, Novo Nordisk) and Trulicity® (dulaglutide, M.W.: 63,000, Eli Lilly) were used as GLP-1 receptor agonists. All other reagents were of an analytical grade or higher.
The Caco-2 cell line was obtained from the Riken Cell Bank (Ibaraki, Japan) and kept in a humidified incubator at 37°C with 5% CO2. Cells were maintained in DMEM-High glucose (Fujifilm Wako, 044-29765, Osaka, Japan) supplemented with 10% fetal bovine serum (SELBORNE, FBS-04, Tasmania, Australia), 0.1 mM non-essential amino acids (Fujifilm Wako, 139-15651, Osaka, Japan), 2 mM L-glutamine (Fujifilm Wako, 073-05391, Osaka, Japan), 100 U/mL of penicillin, 100 mg/ml of streptomycin, and 250 ng/mL of amphotericin B (Fujifilm Wako, 161-23181, Osaka, Japan) and were used between passages 13–18.
In the growth assay, cells were seeded on a 96-well culture plate (1.0 × 104 cells/well) and cultured for 24 h to allow for adherence. Thereafter, the culture medium was replaced with a fresh medium with/without each concentration of semaglutide and dulaglutide and incubated for 96 h.
To assess cell proliferation, relative cell numbers were measured using crystal violet staining for adherent cells. Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 10 min, stained with a 0.04% crystal violet aqueous solution for 30 min, and dissolved in a 1% SDS solution. Cell viability was estimated by measuring absorbance on a microplate reader (SpectraMax iD5, Molecular Devices, LLC., CA, USA) at a wavelength of 560 nm.
In the cell viability assay, cells were seeded on a 96-well culture plate (1.0 × 104 cells/well) and cultured to confluency (approximately 96 h). Cells were then treated with each concentration of semaglutide and dulaglutide and incubated for 24 h.
To assess cell viability, relative cell numbers were measured using crystal violet staining for adherent cells. Cells were fixed with 4% paraformaldehyde in PBS for 10 min, stained with a 0.04% crystal violet aqueous solution for 30 min, and dissolved in a 1% SDS solution. Cell viability was estimated by measuring absorbance on a microplate reader (SpectraMax iD5, Molecular Devices, LLC., CA, USA) at a wavelength of 560 nm.
Cells were seeded on 12-well plates (1.0 × 105 cells/well), incubated until confluent, treated with each concentration of semaglutide and dulaglutide, and then incubated for 48 and 96 h. Cells were lysed with Laemmli Sample Buffer (Sigma Aldrich Co., Ltd., Tokyo, Japan). Proteins were separated by SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked in 5% skim milk in Tris-buffered Saline with Tween 20 (TBST) at room temperature for 1 h and incubated with primary antibodies (GLP-1 receptor: GLP1R Rabbit pAb, A13990, ABclonal, P-gp: P-Glycoprotein antibody [C219], GTX23364, GeneTex, Claudin-4: Claudin 4 Polyclonal Antibody, bs-2790R, Bioss Antibodies, Claudin-7: Claudin 7 Polyclonal Antibody, bs-8482R, Bioss Antibodies, MUC1: MUC1 Rabbit pAb, A0333, ABclonal, MUC2: MUC2 Rabbit mAb, A4767, ABclonal, β-actin: Beta Actin Mouse Monoclonal antibody, 66009-1-Ig, ProteinTech) at 4°C overnight. After washing three times in TBST, PVDF membranes were incubated with a secondary antibody solution at room temperature for 1 h. PVDF membranes were reacted with a chemiluminescent reagent (Pierce ECL Western Blotting Substrate, 32106, Thermo Scientific), and their protein bands were visualized on ChemiDoc touch (Bio-Rad). Each protein expression level was normalized by β-actin.
Cells were seeded on 12-well plates (1.0 × 105 cells/well), incubated until confluent, treated with each concentration of semaglutide and dulaglutide, and then incubated for 48 and 96 h. Total RNA isolation from Caco-2 cells were performed using TRIzol reagent (Thermo Fisher Scientific K.K, 15596026, Tokyo, Japan) following the manufacturer’s instruction. To obtain cDNA, 1 μg of RNA was subjected to reverse transcription using the ReverTra Ace qPCR RT kit (TOYOBO Co., Ltd., Tokyo, Japan).
Real-time RT-PCR was performed using THUNDERBIRD SYBR qPCR Mix (TOYOBO Co., Ltd., Tokyo, Japan) on CFX Maestro (Bio-Rad Laboratories, Inc., CA, USA). The specificity of the primers was theoretically tested by the BLAST database. The sequences of all primers are shown in Table 1. Each experiment was performed in duplicate and repeated for each condition tested. The mRNA levels of all genes were normalized using the housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as an internal control. All quantification were performed three times independently.
Sequences of primers used for real-time RT-PCR
| Gene | Sequence (5′-3′) | Amplicon Size | GenBank ID | |
|---|---|---|---|---|
| ABCB1 | Forward | CAGAGGGGATGGTCAGTGTT | 87 bp | AF016535.1 |
| Reverse | CCTGACTCACCACACCAATG | |||
| Claudin-4 | Forward | CATCTCCTCTGTTCCGGGTA | 91 bp | NM_001305.5 |
| Reverse | AAGGCCTCAGCCATACTCCT | |||
| Claudin-7 | Forward | CCCTCCACCTTTTGTTTGCC | 116 bp | NM_001307.6 |
| Reverse | GCACAGGGAGTAGGATACGC | |||
| GAPDH | Forward | ACCAGGGCTGCTTTTAACTCTG | 104 bp | NM_001256799.3 |
| Reverse | TGGGTGGAATCATATTGGAACAT | |||
All results are expressed as the mean ± standard deviation (S.D.). The significance of differences between groups was analyzed using Tukey's test; P <0.05 was considered to be significant.
Caco-2 cells were exposed to 5, 50, and 500 ng/mL and 15, 150, and 1500 ng/mL of semaglutide and dulaglutide, respectively, and neither affected the proliferation rate or ratio of Caco-2 cells (Fig. 1a, 1b). Furthermore, confluent Caco-2 cells were exposed to semaglutide and dulaglutide in order to assess their cytotoxicities, and no effects were noted for either (Fig. 1c, 1d). On the other hand, the protein expression level of the GLP-1 receptor was not significantly changed by semaglutide, whereas dulaglutide exerted significant concentration- and time-dependent effects (Fig. 1e, 1f).
Effects of semaglutide and dulaglutide on Caco-2 cells. (a, b) Cell proliferation, (c, d) cell viability, and (e, f) GLP-1 receptor protein expression. Data represent means and S.D. (n = 3-6 for each condition). *P <0.05 vs. the control (CTRL) condition
The present study examined the effects of GLP-1 receptor agonists (exposure for 48 and 96 h) on the mRNA expression levels of epithelial cell barrier factors in Caco-2 cells. The mRNA expression level of ABCB1, which codes P-glycoprotein (P-gp), was not significantly changed by the 48- or 96-h exposure to semaglutide and dulaglutide (Fig. 2a, 2b). Although the mRNA expression levels of Claudin-4 and Claudin-7, components of tight junctions (TJ), were not also significantly affected by the 48- or 96-h exposure to semaglutide and dulaglutide (Fig. 2c, 2d, 2e, 2f), slight decreases were observed in a semaglutide concentration-dependent manner. In contrast, slight increases were noted in a dulaglutide concentration-dependent manner after the 96-h exposure.
Effects of semaglutide and dulaglutide on mRNA expression levels of epithelial barrier function-related factors in Caco-2 cells. (a, b): ABCB1, (c, d): Claudin-4, and (e, f): Claudin-7. Exposure to a GLP-1 receptor agonist for 48 h (a, c, e) and 96 h (b, d, f). Data represent means and S.D. (n = 3 for each condition).
Next, the effects of GLP-1 receptor agonists (exposure for 48 and 96 h) on the protein expression levels of epithelial cell barrier factors in Caco-2 cells were examined, and their typical bands are shown in Figs. 3a and 3b. The protein expression levels of MUC1 and MUC2, components of unstirred water layer (UWL), were decreased by the 48-h exposure to semaglutide (Fig. 3i), whereas those of other proteins were not. Furthermore, the 96-h exposure to semaglutide did not affect the expression levels of any of the proteins examined (Fig. 3). In contrast, in cells exposed to dulaglutide, significant concentration- and exposure time-dependent increases were noted in the protein expression levels of P-gp, Claudin-4, Claudin-7, MUC1, and MUC2 (Fig. 3).
Effects of semaglutide and dulaglutide on protein expression levels of epithelial barrier function-related factors in Caco-2 cells. (a, b): Typical bands for Western blotting, (c, d): P-gp, (e, f): Claudin-4, (g, h): Claudin-7, (i. j): MUC1, and (k, l): MUC2. Exposure to a GLP-1 receptor agonist for 48 h (a, c, e, g, i, k) and 96 h (b, d, f, h, j, l). Data represent means and S.D. (n = 3 for each condition). *P <0.05 vs. the control (CTRL) condition.
In this study, the Caco-2 cell line, an intestinal epithelial model of small intestinal epithelial cell-like functions, was used. Semaglutide and dulaglutide concentrations were set to 5, 50, and 500 ng/mL and 15, 150, and 1500 ng/mL, respectively, based on steady-state plasma concentrations during subcutaneous administration (Interview Form of Ozempic® and Trulicity®). Neither semaglutide nor dulaglutide affected the proliferation rate or ratio and cell viability of Caco-2 cells (Fig. 1). We previously reported that liraglutide, another GLP-1 receptor agonist, did not affect the proliferation of Caco-2 cells (Takizawa et al., 2022). Collectively, these findings and the present results indicate that the oral administration of GLP-1 receptor agonists is safe at in vitro cell proliferation and viability assay levels, even if unabsorbed medicine is exposed to gastrointestinal epithelial cells for a prolonged period of time.
On the other hand, the protein expression level of the GLP-1 receptor was not significantly changed by semaglutide, whereas dulaglutide exerted significant concentration- and time-dependent increase effects (Fig. 1E, 1F). There is currently no evidence to support increases in the expression of GLP-1 receptors in intestinal epithelial cells (Caco-2) following their exposure to GLP-1 receptor agonists. Interestingly, only dulaglutide exerted significant effects. Therefore, the mechanisms responsible for this change in expression may provide useful information for the effective use of GLP-1 receptor agonists.
Since the GLP-1 receptor is a 7-transmembrane G protein-coupled receptor (Göke & Conlon, 1988; Brubaker & Drucker, 2002), the present results also suggest that dulaglutide changes the expression of other membrane proteins, such as drug transporters.
Intestinal epithelial cells directly interact with the external environment; therefore, various barrier functions exist, which may be broadly classified into two types: biological and physical. Biological barriers include ABC transporters, such as P-gp, which excrete foreign substances taken into cells (Suzuki, 1999; Watanabe et al., 2013), and metabolic enzymes, such as CYP (Canaparo et al., 2007). On the other hand, physical barriers include TJ in intercellular spaces and UWL on the cell surface, which are formed by membrane proteins, such as members of the Claudin family and MUC family (Madara et al., 1993; Chiou, 1994).
The effects of GLP-1 receptor agonists on the mRNA expression levels of these factors in Caco-2 cells were examined (Fig. 2). No significant changes were detected in the mRNA expression levels of each epithelial barrier function regulator upon exposure to semaglutide or dulaglutide, suggesting that the luminal approach of GLP-1 receptor agonists did not affect the regulatory signals for the expression of various factors in these cells.
Although no significant changes were observed in the mRNA expression levels of each factor, protein expression levels were examined under the same conditions as those for mRNA. It was observed that fluctuations in protein expression of epithelial cell barrier factor by GLP-1 receptor agonist exposure were greater than fluctuations in their mRNA expression. In particular, dulaglutide significantly increased the protein expression levels of many epithelial cell barrier factors in concentration- and exposure time-dependent manner (Fig. 3).
We previously reported that liraglutide did not affect MUC1 or Claudin-4 protein expression levels but increased P-gp protein expression levels in concentration- and exposure time-dependent manners (Takizawa et al., 2022). Collectively, these findings and the present results demonstrated that the effects of GLP-1 receptor agonists on membrane proteins differed depending on the type of GLP-1 receptor agonist. On the other hand, the protein expression levels of AMP-activated protein kinase a and phosphatidylinositol-5-phosphate 4-pinase type 2 beta, which function as intracellular ATP and GTP sensors, were not significantly changed by semaglutide or dulaglutide (data not shown). These results confirmed that dulaglutide did not induce the expression of all proteins. On the other hand, there are reports that GLP-1 receptor agonists alter the expression of proteins such as transporters and enzymes (Zhang et al., 2021; Wang et al, 2023), but the detailed mechanism is not clear. Therefore, further investigation is required, including the mechanistic theory of the phenomenon observed in this study.
In the present study, dulaglutide was used at a higher concentration (ng/mL) than semaglutide, whereas that of semaglutide was higher when considered in terms of mol concentration (nmol/L). Since both concentrations were above the EC50 for GLP-1R (Assessment report of Ozempic® and Trulicity®), the effects of dulaglutide do not appear to be mediated by GLP-1 receptors. Moreover, semaglutide did not induce the expression of these proteins, suggesting that dulaglutide-mediated changes were not due to a common target with semaglutide. The mechanism by which dulaglutide increases the expression of these proteins is unknown. Although the expression of proteins on the cell surface that function as defense factors increased across the board, the presence of dulaglutide, the molecular weight of which is very high (M.W.: 63,000), on the cell surface caused the strong recognition of foreign substances by cells, resulting in a temporary increase in the expression of these epithelial barrier factors. Liraglutide (M.W.: 3751.26), a GLP-1 receptor agonist with a smaller molecular weight than that of semaglutide (M.W.: 4113.6), induced the protein expression of P-gp in Caco-2 cells at high concentrations and exerted protective effects against external oxidative stress (Takizawa et al., 2022). These findings suggest that the induction of membrane proteins related to epithelial cell barrier functions is due to a defense mechanism at the cell surface.
On the other hand, previous studies reported a decrease in P-gp expression in the liver and significant increases in P-gp mRNA and protein levels in the small intestine in a type 2 diabetic mellitus model rat (Yao et al., 2020; Gwak et al., 2020). If there is a mixture of the up-regulation of P-gp due to pathological conditions and that due to therapeutic agents, an additive effect may occur. Therefore, further studies are needed to elucidate the mechanisms underlying the up-regulation of P-gp by GLP-1 receptor agonists, such as dulaglutide, which will provide important information for the appropriate use of concomitant medicines that are P-gp substrates.
In addition, the protein expression levels of TJ components, such as Claudin-1, Occludin, and ZO-1, were found to be decreased in the colon of a type 2 diabetic mellitus model mouse (Rehman et al., 2022). Furthermore, Occludin mRNA levels in the small intestine were higher, whereas MUC2, MUC12, MUC20, and MUC21 mRNA levels were significantly lower in patients with type 1 diabetes than those in a normal group (Lo Conte et al., 2023). Moreover, since inflammation of the gastrointestinal tract during diabetes is an important factor in its pathogenesis, some studies have attempted to control the disease by amplifying MUC in the gastrointestinal tract (Eliuz Tipici et al., 2023). Therefore, the increases observed in MUC levels by dulaglutide may suppress disease progression due to diabetes; however, further studies are needed.
In this study, semaglutide has a relatively small effect on epithelial cells barrier, while dulaglutide was shown to have the potential to improve epithelial cell barrier function. Since various gastrointestinal symptoms develop during diabetes and epithelial cell barrier functions change, medicines that promote barrier functions, such as dulaglutide, may be effective treatments. However, their mechanisms of action remain unknown; therefore, further detailed studies are warranted.