Inhibition of GSK-3β restores delayed gastric emptying in obesity-induced diabetic female mice

Abstract

Diabetic gastroparesis (DG) is a clinical syndrome characterized by delayed gastric emptying (DGE). Loss of nuclear factor erythroid 2-related factor 2 (Nrf2) is associated with reduced neuronal nitric oxide synthase-α (nNOSα)-mediated gastric motility and DGE. Previous studies have shown that nuclear exclusion and inactivation of Nrf2 is partly regulated by glycogen synthase kinase 3β (GSK-3β). In the current study, the molecular signaling of GSK-3β-mediated Nrf2 activation and its mechanistic role on DG were investigated in high-fat diet (HFD)-induced obese/Type 2 diabetes (T2D) female mice. Adult female C57BL/6J mice were fed with HFD or normal diet (ND) with or without GSK-3β inhibitor (SB 216763, 10 mg/kg body wt ip) start from the 14th wk and continued feeding mice for an additional 3-wk time period. Our results show that treatment with GSK-3β inhibitor SB attenuated DGE in obese/T2D mice. Treatment with SB restored impaired gastric 1) Nrf2 and phase II antioxidant enzymes through PI3K/ERK/AKT-mediated pathway, 2) tetrahydrobiopterin (BH₄, cofactor of nNOS) biosynthesis enzyme dihydrofolate reductase, and 3) nNOSα dimerization in obese/T2 diabetic female mice. SB treatment normalized caspase 3 activity and downstream GSK-3β signaling in the gastric tissues of the obese/T2 diabetic female mice. In addition, GSK-3β inhibitor restored impaired nitrergic relaxation in hyperglycemic conditions. Finally, SB treatment reduced GSK3 marker, pTau in adult primary enteric neuronal cells. These findings emphasize the importance of GSK-3β on regulating gastric Nrf2 and nitrergic mediated gastric emptying in obese/diabetic rodents.

NEW & NOTEWORTHY Inhibition of glycogen synthase kinase 3β (GSK-3β) with SB 216763 attenuates delayed gastric emptying through gastric nuclear factor erythroid 2-related factor 2 (Nrf2)-phase II enzymes in high-fat diet-fed female mice. SB 216763 restored impaired gastric PI3K/AKT/ β-catenin/caspase 3 expression. Inhibition of GSK-3β normalized gastric dihydrofolate reductase, neuronal nitric oxide synthase-α expression, dimerization and nitrergic relaxation. SB 216763 normalized both serum estrogen and nitrate levels in female obese/Type 2 diabetes mice. SB 216763 reduced downstream signaling of GSK-3β in enteric neuronal cells in vitro.

INTRODUCTION

Gastroparesis diabeticorum or diabetic gastroparesis (DG) defined as abnormally delayed gastric emptying of solid food in the absence of mechanical obstruction, remains a diagnostic and management challenge (34). Gastroparesis occurs in both Type 1 and 2 diabetes and may not, necessarily, be indicative of a poor prognosis (39). Data from the United States indicate that emergency visits rose from 12.9 to 27.3% between 2006 and 2014, and hospitalizations rose 100-fold between 1994 and 2014 for gastroparesis, which may reflect a true increase in incidence and/or greater clinical awareness of the condition (21, 23, 60). Glycogen synthase kinase-3β (GSK-3β) has recently emerged as a promising target for development of drugs against chronic diseases because of causative associations with diabetes mellitus and neuropsychiatric disorders (8, 29, 41).

Two isoforms of GSK-3 (α and β) have been shown to play a role in various cellular mechanisms (73). Of these, increases in GSK-3β has been studied extensively in two distinct signal transduction pathways: 1) the phosphatidylinositol (PtdIns)-3–kinase-dependent pathway that is triggered by insulin and growth factors, and 2) the Wnt signaling pathway that is required for embryonic development and is linked to pathological diseases, including, but not limited to, cancer and metabolic and neurological disorders (7, 12). GSK-3β is a serine/threonine protein kinase that regulates fundamental cellular pathways, depending on the substrates it phosphorylates (15, 28). Particularly, the functional role of GSK-3β in insulin signaling and glucose metabolism makes it a particularly intriguing candidate target for treatment of Type 2 diabetes. Activated GSK-3β phosphorylates and, thereby, inactivates glycogen synthase, an enzyme involved in converting glucose to glycogen for storage. Insulin can relieve GSK-3-mediated inhibition on glycogen synthase by binding to its receptor and activating the phosphatidylinositol-3-kinase /protein kinase B (PI3K/Akt) signaling pathway. The activity and expression of GSK-3β have been reported to be elevated in adipose tissue of insulin-resistant obese rodent models and in skeletal muscle of obese Type 2 diabetic patients (17, 47). Overexpression of human GSK-3β in skeletal muscle of mice results in impaired glucose tolerance and suppressed glycogen synthase activity and glycogen synthesis, further supporting a role for GSK-3β in Type 2 diabetes (51). Furthermore, overexpression of GSK-3β has been shown to attenuate insulin signaling due to phosphorylation and downregulation of insulin receptor substrate-1 (8). Therefore, it has been suggested that drugs inhibiting GSK-3β could mimic the ability of insulin to promote the conversion of glucose to glycogen, overcoming the resistance to insulin. Under physiological conditions, GSK-3β phosphorylates and triggers degradation of several transcription factors and proto-oncoproteins such as β-catenin (37).

Emerging research suggests that GSK-3β participates in the cellular response to oxidative stress, a hallmark of several nervous system disorders. GSK-3β modulates this response through its interaction with nuclear factor erythroid 2-related factor 2 (Nrf2) (27). Nrf2 is strongly associated with immune regulation and oxidative stress, rendering it a reasonable therapeutic target (2). We have previously reported that loss of Nrf2 in female mice (Nrf2 KO, Nfe2/2^−/−) resulted in elevated gastric GSK-3β, decreased tetrahydrobiopterin (BH₄) levels, inhibition of neuronal nitric oxide (NO) (nNOS, nitrergic neuron)-mediated gastric relaxation, and reduction in nitrite levels, subsequently causing delayed gastric emptying (19, 20, 43, 57). Most studies have shown a higher prevalence of gastroparesis in women than in men (4, 34). Hence, in the current study, we have investigated the mechanistic role of a small-molecule GSK-3β inhibitor (SB 216763) in gastric nitrergic relaxation and solid gastric emptying in a high-fat diet-induced (HFD) obesity and Type 2 diabetic gastroparesis using adult female wild-type (WT, C57BL/6J) mice.

MATERIALS AND METHODS

Experimental animals.

All animal experiments in this study (no. 17-09-764) were approved by The Institutional Animal Care and Use Committee at Meharry Medical College. Adult (9 wk old) female C57BL/6J were purchased from Jackson Laboratories (Bar Harbor, ME). The mice were fed either with a high-fat diet (HFD, 70% energy as fat, 19% protein, and 11% carbohydrate, HFD; 5SSV, Test Diet, St. Louis, MO) or a normal diet (ND, 6.2% energy as fat, Teklad Global 2018, Teklad Diets, Madison, WI) and were allowed access to food and water ad libitum.

Experimental design.

Adult female mice were randomly assigned to one of three treatment groups (n = 6–8 per group): 1) WT-ND, 2) WT-HFD, and 3) WT-HFD + SB 216763 [SB 216763 (SB); 10 mg/kg ip; Tocris Biosciences, Minneapolis, MN]. SB was dissolved in 5% DMSO and administered intraperitoneally three times a week for 3 wk, beginning at the 14th week in HFD mice. The control group WT-ND and WT-HFD were supplemented with vehicle (5% DMSO), administered intraperitoneally three times a week for 3 wk, beginning at the 14th wk. Body weights were recorded every week. At the end of week 16, all mice were euthanized by CO₂ asphyxiation. Stomach tissues, livers, and epididymal fat pads were rapidly isolated, weighed, and were snap-frozen and stored at −80°C.

Fasting blood glucose determination and glucose tolerance test.

Absolute fasting blood glucose levels were monitored at the end of every alternate week for the duration of the study period using a standard protocol (75). Intraperitoneal glucose tolerance test (IPGTT) were performed at week 16, as described previously (75).

Analysis of insulin, nitrite, and 17β-estradiol in the serum.

Insulin levels in the serum were measured by ELISA (Crystal Chem, Elk Grove Village, IL), as described in the manufacturer’s protocol. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated using the following formula as per Bonora et al. (5). Serum nitrite and 17β-estradiol (E₂) were performed, as described previously (57). Nitrite levels in the serum were analyzed as total nitrite (metabolic by-product of NO), according to manufacturer’s protocol (BioVision, Milpitas, CA). Serum E₂ levels were assayed using an ELISA kit (Enzo Life Sciences, Farmingdale, NY) following the manufacturer’s instructions. Absorbance at 405 nm was obtained using a microplate reader (BioTek, Winooski, VT).

Solid gastric emptying studies.

Solid gastric emptying (GE) studies were performed, as described previously (43). Briefly, the mice were first deprived of solid food overnight (water consumption allowed). The next day, each mouse were placed separately in a single cage. A premeasured bolus of food (ND or HFD) was then provided with water ad libitum for 3 h. The mice were then placed in a separate clean cage and starved for 2 h without food and water. The remaining food was dried and weighed to establish the amount of total food intake (FI). Because the solid food is dried adequately, we believe that this procedure minimizes to avoid measuring gastric secretions that may differ between the groups. To measure the rate of gastric emptying, mice were euthanized by cervical dislocation, and stomachs were carefully dissected. To minimize the variations, the same experimenter performed all of these procedures. Full and empty stomach weights were recorded for each animal. The difference estimated the remaining gastric content (GC) after 2 h of fasting and the rate of GE was calculated as follows: GE (% in 2 h) = 1 – [gastric content (GC) /food intake (FI)] × 100. This protocol calculates only the solid food content but not any other gastric secretions that may incur upon HFD diet.

Organ bath studies.

Electric field stimulation (EFS)-induced nonadrenergic noncholinergic relaxation (NANC) was examined in circular gastric antrum neuromuscular strips in WT mice (n = 4/group), as previously described (57). To investigate the in vitro effect of SB (10 μM; 30 min) under normoglycemic or hyperglycemic conditions (50 mM, 30 min), muscle strips were preincubated and NANC-dependent nitrergic relaxation (nNOS function) was determined. The circular gastric antrum neuromuscular strips were mounted in 10 mL Krebs buffer at 37°C, and NANC-nNOS function was determined at 2 Hz (57) (DMT Technologies, Nottingham, UK). The NO dependence of nitrergic relaxation was confirmed with N^G-nitro-l-arginine methyl ester treatment (l-NAME, 100 µM, 30 min). Comparison between groups was performed by measuring the area under the curve (AUC/mg of tissue) of the EFS-induced relaxation (AUCR) curve at 1 min and the baseline (AUCB) curve at 1 min, as follows: (AUCR – AUCB)/weight of tissue (mg) = AUC/mg of tissue.

nNOSα dimerization in the gastric neuromuscular tissues.

Levels of nNOSα monomer and dimer were quantified by Western blot analysis via low temperature (LT)-PAGE of gastric antrum homogenates, as described previously (19). A polyclonal anti-nNOSα antibody (NH₂-terminus) (1:500, Thermo Fisher Scientific, Waltham, MA) and anti-rabbit IgG conjugated with horseradish peroxidase (1:10,000, Sigma Chemical, St. Louis, MO) were used as the primary and secondary antibodies, respectively.

Western blot analysis.

Proteins from gastric antrum neuromuscular tissue lysates were separated by SDS-PAGE. The membrane was immunoblotted with polyclonal Nrf2 (sc-365949, 1:500), superoxide dismutase 1 (SOD 1, sc-101523, 1:500), catalase (CAT, sc-271803, 1:500), p-ERK (sc-81492), p-AKT (sc-514032), and dihydrofolate reductase (DHFR, sc-33184, 1:500) from Santa Cruz (Santa Cruz Biotechnology, Dallas, TX), p-PI3K (4228s), p-Tau (12885s), GSK-3β (12456T), p-GSK-3β (9336s), total caspase 3 (14220T), cleaved caspase 3 (9661s) and β-catenin (9562s) (1:1,000, Cell Signaling, Danvers, MA), nNOS α (ab229785, 1:1,000, NH₂-terminus; Abcam, Cambridge, MA), primary antibody and anti-rabbit/anti-mouse IgG conjugated with horseradish peroxidase were used as secondary antibodies (1:10,000) (Sigma-Aldrich). Binding of antibodies to the blots was detected using an enhanced chemiluminescence (ECL) system (Amersham Pharmacia Biotech, Piscataway, NJ) following the manufacturer’s instructions. Stripped blots were reprobed with β-actin-specific polyclonal antibodies (Sigma Chemicals, St. Louis, MO) to enable normalization of signals between samples. Band intensities were analyzed using ImageQuant LAS 500 (GE Health Sciences, Pittsburgh, PA).

Primary cultures of adult mouse enteric neuronal cell culture.

Primary cultures of enteric neuronal cell (ENC) were performed as previously described using adult mouse gut tissue (74, 76). In brief, myenteric neuronal cells were isolated from colon and ileum myenteric plexus of WT-ND mice and were seeded on Matrigel-coated plates and cultured at 37°C with 5% CO₂ in complete neuronal basal media containing B-27 serum-free supplement, 1 mmol/L glutamine, 1% FBS, and 10 ng/mL GDNF with half of the medium replaced every day for 5–7 days. The cells were then fixed in 4% paraformaldehyde in PBS and processed using immunofluorescent staining procedures by employing anti-Nrf2 (1:100) and anti-GSK-3β (1:500), as described above. Nuclei were counterstained with DAPI. β-tubulin (Tuj-1) and SB 100 were used as neuronal and glial cell markers, respectively. Secondary antibodies conjugated to anti-mouse IgG (Alexa Fluor 488, Molecular Probes, Eugene, OR) or anti-rabbit IgG (Alexa Fluor 594; Molecular Probes, Eugene, OR) diluted in PBS were incubated for 1 h in the dark. Slides were washed and mounted using FluorSave reagent (Millipore, Burlington, MA). Mounted slides were viewed on a Leica DM LB2 microscope with Nikon Digital Sight DS-U2 camera, using ×40 objectives. Images were taken using the software NIS-Elements version 3.0 (Nikon, Japan).

Furthermore, in a separate set of experiments, the primary neuronal cell cultures were incubated with GSK-3β inhibitor (10 µM) for 24 h, and the cells were spun down, the supernatant removed completely, and the cells were washed in ice-cold PBS. Lysis buffer (50–80 μL containing 0.2% protease inhibitor cocktail) was added to the pellet, kept on ice for 5 min, sonicated 4 times for 5 s each on ice, and microcentrifuged for 10 min at maximum speed (15,700 g). Finally, the supernatant cell lysate was transferred to a new Eppendorf tube and stored at –80°C. Protein was measured by the Lowry method (67).

Statistical analysis.

Data were presented as the means ± SE. Statistical comparisons between groups were performed using the Student’s t test or Tukey’s test after one-way and two-way ANOVA with GraphPad Prism Version 5.0 (GraphPad Software, San Diego, CA). A P value of less than 0.05 was considered to be statistically significant.

RESULTS

Effect of SB 216763 treatment on whole animal body and adipose and liver tissue weights.

As shown in Fig. 1A, when compared with the mice of the ND group, mice fed with HFD exhibited greater body weight (BW) gain (P < 0.05) beginning from week 3 until week 16. The HFD was supplemented with SB starting at week 14. SB treatment continued until experimental period (end of week 16). Our data show that SB treatment did not attenuate BW gain in mice while fed with HFD (Fig. 1A). To further investigate the effect of SB on obesity, the weight of epididymis fat pads (an indicator of visceral fat mass) was measured. We observed an eightfold increase in the weight of epididymal fat pads in HFD group when compared with ND mice (ND: 0.21 ± 0.07 g vs. HFD: 1.65 ± 0.12 g, P < 0.001, Fig. 1B). Treatment with SB significantly reduced the weight of epididymis fat in mice fed on HFD (HFD + SB: 1.19 ± 0.2 g vs. HFD: 1.65 ± 0.12 g, P < 0.05, Fig. 1B). Liver weights among all of these experimental groups were analyzed. The data show that there was no statistical difference among these three groups (Fig. 1C).

Fig. 1.

Effects of SB216763 (SB) treatment on weights of whole body, epididymal fat tissue, and liver. A: body weight. The body weight of mice was measured weekly following a normal diet or high-fat diet feeding during the period of 16 wk. The arrow indicates the starting time of SB 216763 treatment. B: weight of epididymal adipose tissue. C: liver weight. The mice were fed with either a normal diet (ND) or high-fat diet (HFD; n = 20 mice) for 16 wk. The HFD-fed mice then received SB 216763 (10 mg/kg) or vehicle (5% DMSO) for 3 wk. The ND-fed mice were also treated with the vehicle. Data are expressed as means ± SE; n = 6 mice. *P < 0.01 compared with ND group. #P < 0.05 compared with HFD group.

SB 216763 improves glucose disposal and insulin sensitivity.

Data depicted in Fig. 2A shows that fasting blood glucose levels were elevated in HFD-fed group compared with the ND group. Supplementation of SB rapidly (P < 0.05) attenuated elevated fasting glucose levels in the HFD-fed group. Previous studies demonstrate that long-term supplementation of SB significantly prevented HFD-induced insulin resistance (66). To assess whether SB could reverse insulin insensitivity in an established chronic diabetes model or to test its treatment efficacy, we performed an IPGTT. As shown in Fig. 2B, glucose levels following the intraperitoneal injection were significantly higher at 16 wk feeding of HFD. The analysis of area under curve (AUC) of the IPGTT during the tested 120 min also showed that HFD induced about a twofold increase in the AUC index, while SB normalized this elevation by more than 50% (Fig. 2C). Furthermore, to assess whether insulin sensitivity was enhanced by SB treatment, we measured fasting glucose and insulin levels (Fig. 2, D and E) and calculated the homeostasis model assessment-insulin resistance (HOMA-IR) index (Fig. 2F). We found that HFD induced almost a three-fold increase in this index, while SB significantly (P < 0.05) reversed this abnormality. These results indicate that the effect of SB on improving glucose utilization occurs via enhancing insulin action and the inhibition of GSK-3β in HFD-fed mice.

Fig. 2.

Effects of SB216763 (SB) treatment on fasting blood glucose levels, glucose tolerance, and insulin sensitivity. A: fasting blood glucose levels every alternate week during the study period. B: intraperitoneal glucose tolerance test in the normal-diet group, high-fat diet group, and high-fat diet plus SB216763 treatment group (n = 6 mice). C: area under the curve of glucose tolerance test. D: serum glucose levels. E: serum insulin levels. F: homeostasis model assessment - insulin resistance index values. G: serum estradiol (E₂). H: serum nitrate levels. At the end of 16 wk following normal diet (ND+Veh), high-fat diet (HFD+Veh), and HFD+SB216763 treatments, intraperitoneal glucose tolerance tests were performed, and blood glucose levels were measured. Data are expressed as means ± SE; n = 6 mice. *P < 0.01 compared with ND group. #P < 0.05 compared with HFD group. AUC, area under the curve; HOMA-IR, homeostasis model assessment-insulin resistance index.

SB treatment restored serum E₂ and nitrite levels.

Earlier reports have demonstrated that estrogen deficiency or impaired estrogen signaling is associated with insulin resistance and faulty regulation of metabolic homeostasis, which contributes to the development of Type 2 diabetes and obesity in both human and animal models (72). Further studies indicate that E₂ elevates NO levels and regulates gastric motility function (55). As shown in Fig. 2G, serum E₂ levels were elevated in HFD-fed group in comparison to ND, whereas serum nitrite levels were significantly (P < 0.05) reduced in HFD-fed groups (Fig. 2H). SB treatment normalized both E₂ and nitrite levels in HFD-fed mice group.

Inhibition of GSK-3β reversed obesity/diabetes-induced delay in GE.

Next, we investigated whether inhibition of GSK-3β attenuated delayed GE in HFD-induced chronic diabetes. HFD-fed WT mice showed a higher gastric retention, which was significantly different between ND-fed WT mice, whereas SB treatment in HFD group showed no significant difference compared with ND-fed WT mice. As shown in Fig. 3, HFD-fed WT mice showed a significant delay (P < 0.05) in solid GE compared with ND-fed WT mice (22.5 ± 4.9% vs. 77.3 ± 3.4%, P < 0.05). SB treatment completely restored GE (81.3 ± 2.8%, P < 0.05) in WT HFD-fed mice (Fig. 3).

Fig. 3.

Effect of SB216763 (SB) treatment on solid gastric emptying in high-fat diet (HFD)-fed female mice. The mice were fed with either a normal diet (ND+Veh) or high-fat diet (HFD+Veh) for 16 wk. The HFD-fed mice then received SB 216763 (50 mg·kg⁻¹·day⁻¹; n = 10 mice) or vehicle (5% DMSO; n = 10 mice) for 3 wk. Data were analyzed using one-way ANOVA by using GraphPad Prism software. Data are expressed as means ± SE; n = 6 mice. *P < 0.05 compared with ND mice. #P < 0.05 compared with HFD mice.

Inhibition of GSK-3β reversed of PI3K, ERK, and Akt protein expression in obesity/diabetes-induced mice.

Since GSK-3β is inhibited by phosphorylation at Ser-9 by Ser/Thr protein kinases via AKT, it has been suggested that Nrf2 might be upregulated through activation of AKT and inactivation of GSK-3 (11). To further investigate whether the PI3K/ERK/Akt pathway altered in gastric antrum specimens, we measured the protein expression levels of phosphorylated PI3K, ERK, and Akt in HFD-fed mice. Compared with the normal diet-fed group, the expression levels of p-PI3K, p-ERK, p-Akt, and p-GSK3β markedly decreased in HFD-fed mice. Treatment with SB normalized expression levels of all these proteins in the HFD-fed group (Fig. 4).

Fig. 4.

Effect of SB216763 (SB) treatment on protein expressions in gastric neuromuscular tissues. Representative immunoblot and densitometry analysis data for p-PI3K (A), p-ERK (B), p-AKT (C), total glycogen synthase kinase 3β GSK-3β (D), p-GSK 3β (E), total caspase 3 (F), cleaved caspase 3 (G), and β-catenin (H) protein expression in female mice gastric neuromuscular tissue. Stripped blots were reprobed with β-actin. Data were normalized for housekeeping gene or protein (β-actin). Bar graphs showed a ratio of target gene or protein with β-actin. Data were analyzed using one-way ANOVA by using GraphPad Prism software. Values are expressed as means ± SD; n = 4 mice). *P < 0.05 compared with ND. #P < 0.05 compared with HFD-fed mice.

Inhibition of GSK-3β attenuates gastric muscle apoptosis by decreasing caspase-3-associated apoptosis.

Caspase-3 is an important regulator of apoptosis (13). Caspase-3 activation, a crucial event of neuronal cell death program, is also a feature of many chronic neurodegenerative and metabolic diseases (13). In the present study, the caspase-3 apoptotic pathway was investigated in gastric antrum specimens. The protein expression levels of caspase-3 were increased in the HFD-fed mice, but were reduced following SB treatment (P < 0.001; Fig. 4E).

Attenuation of a diabetes-induced decrease in the Nrf2 and phase II enzyme protein expression.

Inhibition of GSK-3β has been shown to activate the Nrf2 signaling in various disease models, but not in HFD-induced diabetic gastric specimens (71). To further explore the potential mechanism of an antioxidative effect of SB, we determined whether treatment with SB stimulates gastric Nrf2 signaling in vivo. The protein expressions of Nrf2, as well as phase II enzymes from the three groups of mice, were examined by Western blot analysis. As shown in Fig. 5A, HFD feeding reduced Nrf2 protein level ~50% in mice on a high-fat diet, whereas SB treatment significantly (P < 0.05) attenuated the repressive effect of an HFD. Furthermore, HFD feeding reduced protein levels of catalase (Cat) and superoxide dismutase-1 (SOD-1) whereas SB significantly (P < 0.05) reversed this reduction (Fig. 5, B and C). These results collectively suggest that inhibition of GSK-3β by SB significantly (P < 0.05) activates the Nrf2-Phase II pathway in gastric antrum of HFD-fed mice.

Fig. 5.

Effect of SB216763 (SB) treatment on protein expressions of Nrf2 (nuclear factor (erythroid-derived 2)-like 2)and phase II enzymes in gastric neuromuscular tissues. Representative immunoblot and densitometry analysis data for Nrf2 (A), Cat (B), and SOD-1 (C) protein expression in gastric neuromuscular tissue. Stripped blots were reprobed with β-actin. Data were normalized for housekeeping gene or protein (β-actin). Bar graphs showed a ratio of target gene or protein with β-actin. Data were analyzed using one-way ANOVA by using GraphPad Prism software. Values are expressed as means ± SE; n = 4. *P < 0.05 compared with normal diet (ND). #P < 0.05 compared with high-fat diet (HFD)-fed mice.

GSK-3β inhibition normalized gastric nNOSα and dimerization in HFD-fed mice.

As depicted in Fig. 6A, gastric nNOSα expression was significantly (P < 0.05) reduced in HFD-fed mice, whereas SB treatment enhanced the expression of nNOSα expression. To measure whether the decrease in the nNOSα was due to altered nNOS dimer (an indirect measurement for enzyme activity) levels, we performed the dimerization study by LT-PAGE gel. As shown in Fig. 6B, a significant (P < 0.01) decrease in the dimer/monomer ratio of gastric nNOSα was noticed in HFD-fed mice. Treatment with SB significantly (P < 0.05) restored dimerization in HFD-fed mice.

Fig. 6.

Effect of SB 216763 (SB) treatment on the expression of nitric oxide (NO) signaling proteins in gastric antrum specimens. Representative immunoblots and densitometric analysis data for the following in the gastric antrum neuromuscular specimens: neuronal nitric oxide synthase (nNOSα) (A), nNOSα dimerization (B), and dihydrofolate reductase (DHFR) (C). Blots for β-actin were stripped and reprobed. Data were normalized with band intensities for β-actin. Bar graphs depict ratios of target proteins to β-actin. Data were analyzed with one-way ANOVA using the GraphPad Prism software. The values are expressed as means ± SE; n = 4 mice. *P < 0.05 compared with normal diet (ND)-fed mice. #P < 0.05 compared with the high-fat diet (HFD)-fed mice.

GSK-3β inhibition attenuated diabetes induced decrease in gastric DHFR expression.

Guanylate cyclohydrolase-1 (GCH-1) is responsible for BH₄ biosynthesis via the de novo pathway. Whereas dihydrofolate reductase (DHFR) play a role in BH₄ biosynthesis via the salvage pathway (9). No changes in GCH-1 protein expression were noticed either in mice fed with HFD (data not shown). As depicted in Fig. 6C, treatment with SB significantly (P < 0.05) restored HFD-induced impairment of DHFR protein expression in gastric antrum specimens.

GSK-3β inhibition restored hyperglycemia-induced impairment of gastric nitrergic relaxation in vitro.

As shown in Fig. 7, a statistically significant (P < 0.05) decrease in nitrergic relaxation was observed in gastric antrum tissues exposed to in vitro hyperglycemic conditions when compared with the control group. Our data further show that preincubation with SB (10 μM, 30 min) in the tissues exposed to HG, resulted in significant (P < 0.05) reversal of gastric nitrergic relaxation (Fig. 7).

Fig. 7.

Nitrergic relaxation of in vitro following electric field stimulation (EFS; 2 Hz) in circular gastric antrum strips. The nitric oxide (NO) dependence of the nonadrenergic, noncholinergic (NANC) relaxations was confirmed by preincubation (30 min) with the NO inhibitor nitro-l-arginine methyl ester (l-NAME; 100 μM). In the absence (control) or presence of either high glucose (50 mM) and preincubation of SB216763 (SB; 10 μM) for 30 min. Data were analyzed using one-way ANOVA by using GraphPad Prism software. The values are expressed as means ± SE; n = 4 mice. *P < 0.05 compared with the response in the absence of l-NAME. #P < 0.05 compared with the high-glucose (HG).

Inhibition of GSK-3β in primary enteric neuronal cells.

Immunofluorescence staining revealed that GSK-3β and Nrf2 were expressed both in enteric glial and neuronal cell bodies (Fig. 8). In addition, our study shows that both of them were colocalized in cytosolic compartment of these cells. Next, we determined whether suppression of GSK-3β under normal glycemic conditions alters the protein expression levels of downstream GSK signaling, such as Tau and β-catenin, in enteric neuronal cells. As shown in Fig. 9, the protein expressions of Tau and β-catenin were significantly (P < 0.05) reduced.

Fig. 8.

Triple immunofluorescent staining. Representative images of primary enteric neuronal cells fluorescent staining (from the top) for DAPI (nuclear staining), Nrf2 [nuclear factor (erythroid-derived 2)-like 2], glycogen synthase kinase 3β (GSK-3β). Cells were stained with either β-tubulin (Tuj-1) for neurons and SB 100 for glial cells. Bottom: colocalization of Nrf2/GSK-3β in both glial and neuronal subtypes. Cells were visualized with ×40 magnification. Scale bars: 50 µm. The arrows indicate the positive staining for either Nrf2 or GSK-3β in neuronal and glial cells.

Fig. 9.

Inhibition of glycogen synthase kinase 3β (GSK-3β) by SB 216763 (SB) on the expression of downstream GSK-3β signaling in adult mice primary enteric neuronal cells in normal glycemic conditions. Representative immunoblots and densitometric analysis data for p-Tau (A) and β-catenin (B). Blots for β-actin were stripped and reprobed. Data were normalized with band intensities for β-actin. Bar graphs depict ratios of target proteins to β-actin. Data were analyzed with one-way ANOVA using the GraphPad Prism software. The values are expressed as means ± SE; n = 4 mice. *P < 0.05 compared with Control.

DISCUSSION

GSK-3β inhibitors, such as arylindolemaleimide SB 216763, known to be potent selective small-molecule, cell-permeable compound (6). Previous studies have shown that inhibition of GSK-3β by SB‐216763 play a central role in the survival of neuronal cell populations (10). As reported by Pandey and DeGrado (49), SB-216763 has been shown to inhibit GSK-3β kinase activity with high affinity by 96% in an ATP-competitive manner. These studies further show that this compound did not inhibit other serine/threonine and tyrosine protein kinases, suggesting that it is a selective small-molecule inhibitor for GSK-3β (6). Most importantly, GSK-3β is involved in the downregulation of Nrf2 and control of its subcellular distribution (11). Hence, the interest in using GSK-3β inhibitor on activating Nrf2 in diabetic gastroparesis will be more valuable.

T2DM is known to cause an impairment of both basal and insulin-stimulated glucose metabolism in insulin-responsive, peripheral tissues, including skeletal muscle and liver (48, 52). Previous studies have shown an elevation of GSK-3 expression and activity in the skeletal muscle of patients with T2DM and in adipose tissues of obese diabetic mice (22, 48). Furthermore, transgenic overexpression of GSK-3β in the skeletal muscle of mice results in impaired glucose tolerance, elevated plasma insulin levels, and reduced glycogen content (51). Conversely, GSK-3 inhibitors can mimic insulin action in various cell lines and tissues (6, 50). In addition, in vivo administration of GSK-3 inhibitors to rodent models of obesity and T2DM improves insulin sensitivity and glucose homeostasis by increasing glycogen synthesis and coordinately reducing glucose output by inhibiting hepatic gluconeogenesis (30, 36). GSK-3β is broadly expressed in most of the tissues, and recent evidence starts to address an important role of GSK-3 in regulating insulin sensitivity in multiple types of tissues (69). GSK-3β activities have been found in liver, skeletal muscle, and pancreas to be positively associated with insulin resistance and T2DM in obese and diabetic humans and rodents (35, 48, 50). Development of efficacious GSK-3β inhibitors, with excellent safety profiles, is an ongoing effort and represents potential new and promising therapeutics for T2DM and other diseases with elevated GSK-3 activity, such as neurodegeneration and inflammation (49). Inhibition of GSK-3 in obese mouse models resulted in reductions in body weight and serum glucose levels and increases in serum insulin, C-peptide, and adiponectin (42). Our previous study revealed that mice fed an HFD for 12 wk showed a higher GSK-3β activity in gastric tissue (57). So, on the basis of those results, we used selective small-molecule GSK-3β inhibitor SB 216763 (SB) in our experimental model. Our data show that SB treatment was unable to reduce body weights in HFD group; however, it reduced fat mass and improved glucose homeostasis. The possibility is that inhibition of GSK-3β evoked Nrf2 signaling and modulated lipid metabolism, adipogenesis, adipocyte differentiation, indirectly affecting glucose metabolism (31, 56, 59).

Emerging research suggests that hyperglycemia has an adverse effect on gastric emptying (68). Although both delayed and rapid gastric emptying can be observed in diabetes mellitus, delayed gastric emptying occurs in the majority of patients (33). The stimulation of local pyloric contractions and inhibition of antral contractions contribute to the hyperglycemia-induced delayed gastric emptying (18). The disease affects females more than males in an approximate 4:1 ratio, perhaps due to altered sex hormones, such as estrogens (64). Our previous study revealed that circulatory estrogen levels were altered in HFD-induced diabetic female mice (57). The results herein demonstrated that SB treatment attenuated elevated serum E₂ levels and delayed gastric emptying in female mice fed an HFD, suggesting that inhibition of GSK-3β plays a key role in maintaining circulatory estrogen levels and normal gastric emptying in a setting of obesity-induced diabetes. Mai et al. (37) reported that aberrant GSK-3β expression and activity inhibit the survival of gastrointestinal, pancreatic, and liver cancer cells. Furthermore, these studies suggest that exposure to GSK-3β inhibitor enhanced the survival rate of these cells by phosphorylating GSK-3β (37). Thotala et al. (62) demonstrated improved survival of intestinal crypt cells and increased latency to murine GI-related death from irradiation with SB 216763 in the gastrointestinal system. Our study for the first time shows that inhibition of GSK-3β with SB (SB-216763) reversed delayed gastric emptying. Collectively, the above data suggest that suppression of GSK-3β is critical in maintaining normal gastric emptying rate in HFD-induced diabetic rodents.

GSK-3 has emerged as an important target for drug development for various diseases because of its association with numerous different signaling pathways, which are involved in many discrete biological functions and human diseases (65, 66, 69). The phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB; also known as Akt) signaling pathway is recognized as playing a central role in the survival of diverse cell types (38). It has been reported that the novel potent and selective small-molecule inhibitors of GSK-3, SB-415286 and SB-216763, protect both central and peripheral nervous system neurons in culture from death induced by reduced PI3K pathway activity (10). On the basis of accumulated evidence, the PI3K/AKT signaling pathway is required for normal metabolism because of its characteristics, and its imbalance leads to the development of obesity and Type 2 diabetes mellitus (24). AKT exerts an inhibitory effect on GSK3 by phosphorylation of GSK3 (16). The PI3K enzyme promotes the phosphorylation of Akt, and its activation inhibits GSK-3β, which regulates the levels of β-catenin (3). Inhibition of GSK-3β activity, therefore, leads to stabilization and accumulation of β-catenin in the cytosol, which is shuttled into the nucleus, where it functions to regulate gene expression (26). Our data in adult mouse primary enteric neuronal cells showed an accumulation of GSK-3β in the cytosol as well as coexpressed with Nrf2. In addition, our data indicate that GSK-3β inhibition activates the PI3K/AKT signaling pathway. Mwangi et al. (44) found that enteric neuronal cell survival is dependent on glial cell line-derived neurotrophic factor (GDNF) via activation of the PI3K/Akt signaling pathway. In addition, these studies demonstrated that GDNF promoted enteric neuronal survival by modulating GSK-3β and its downstream target tau expression. Our data show that inhibition of GSK-3β reduced the phosphorylation of Tau. Taken together, the above studies suggest that activation of PI3K/AKT and Nrf2 signaling pathway via inhibition of GSK-3β protects against apoptosis in enteric neuronal cells. Additional studies are warranted to dissect underlying cellular and molecular mechanisms of SB in ENC.

Our data demonstrate that HFD repressed Nrf2-phase II protein expression in gastric tissues by overexpression of GSK-3β, whereas the GSK-3β inhibitor markedly reversed these defects, coincident with activation of PI3K/Akt in HFD-induced DG. Oxidative stress is also a plausible etiologic factor for underlying loss of nitrergic function, because it is well known that diabetes induces a high oxidative stress state that can target various tissues (64). Oxidative stress can be caused by increased reactive oxygen species and loss of antioxidant protection, accounting for the aggravation of diabetic gastroparesis (32). Nrf2 is a key transcription factor that regulates a large group of antioxidant and detoxifying enzymes. Nrf2 deficiency results in increased reactive oxygen species (ROS) level, leading to higher blood glucose level, and impaired insulin signaling in murine models (47). Nrf2 activation represents a crucial cellular defense mechanism in ameliorating oxidative damage and provides substantial therapeutic benefits in DG (46, 58). Previously, Nrf2 upregulation has been proven to protect against diabetes-induced oxidative damage (40). Previous studies from our group demonstrate that activation of Nrf2-phase II pathway restored DGE by attenuating altered gastric GSK-3β, ERK/Akt, JNK/p38 MAPK, BH₄, and nNOSα in HFD-induced T2D female mice (57). Therefore, the application of Nrf2 activator and/or GSK-3β inhibitor may be a viable treatment option to prevent the oxidative damage, mitochondrial dysfunction, and subsequent tissue destruction in DG.

In vivo and in vitro results demonstrated that naturally occurring polyphenolic compounds regulate the phosphorylated (p)-PI3K, Akt, and GSK3β pathways and, consequently, attenuate amyloid β oligomer (AβO)-induced elevations in ROS level and oxidative stress (14). The changes in the above pathways were shown to occur via stimulating the master endogenous antioxidant system of Nrf2 and heme oxygenase-1 (Nrf2/HO-1) enzymes (14). This results in a reduction in apoptosis and neurodegeneration by suppressing the apoptotic and neurodegenerative markers, such as activation of caspase-3 and Poly [ADP-ribose] polymerase 1 expression (14). Previous studies from our laboratory have demonstrated that activation of Nrf2 by cinnamaldehyde improved phosphorylating GSK-3β in the gastric tissue of HFD female mice (57). Collectively, our data suggest that inhibition of GSK-3β restores Nrf2-Phase II enzymes and regulate normal gastric emptying in HFD female mice.

Our data further show that treatment with SB restored impaired BH₄ biosynthesis enzyme, DHFR, and nNOSα dimerization in HFD mice. In addition, SB attenuated hyperglycemia-induced impairment of nNOSα-mediated gastric relaxation. Several lines of evidence suggest that abnormalities in the enteric nervous system (ENS) play a significant role in DG induced gut dysmotility (54). A major component of the ENS is the myenteric plexus, a network of nerves comprising glial and neuronal cells that is integrated between the longitudinal and circular muscle layer of the gut and coordinates gastric motor function. The myenteric plexus comprises excitatory (cholinergic and purinergic) and inhibitory (nitrergic and purinergic) motor neurons, as well as primary afferent neurons and several classes of interneurons. The excitatory motor neurons induce muscle contractions via release of neurotransmitters, such as ACh and substance P, whereas the inhibitory neurons will relax the muscle tissue via release of nitric oxide and also ATP and vasoactive intestinal peptide (1). Pathological changes in these pathways, especially the nitrergic nerves, will affect motor control and may contribute to problems, such as delayed emptying, impaired accommodation, and gastric dysrhythmia (18). Evidence for a role of nitrergic nerves was already obtained in early studies that demonstrated that nNOS knockout mice developed a dilated stomach with hypertrophy of the circular muscle layer (45). In other animal experiments, decreased expression of nNOS by disease or pharmacologic interference with NOS is also able to induce impaired gastric emptying (53, 63). Patient data on loss of the enteric neurons or nervous function are limited, although immunohistochemistry data in a small case series revealed a decrease in nNOS and substance-P expression in the ENS in the stomach of diabetic patients when compared with controls (61). nNOS activity represents a critical signaling node for regulating gastric motor function. nNOS catalyzes the formation of NO, which initiates smooth muscle relaxation. nNOS activity, in turn, is regulated by BH₄, a cofactor for nNOS dimerization and enzyme activity (25). Loss of Nrf2 plays a significant role on the BH₄ levels, and, in turn, on nNOS function (57). It has been reported that activation of Nrf2 transcriptionally upregulates the gene expression of GCH-1 and DHFR, which are responsible for the de novo biosynthesis and salvage pathways of BH₄; such a coordinated gene induction results in increased BH₄ and NO bioavailability, thus protecting endothelial-dependent vascular function (70). Taken together, the results from our studies suggest that inhibition of GSK-3β promotes activation of gastric Nrf2 and enhances the expression of BH₄ enzyme DHFR, nNOS dimerization, and nitrite levels in the serum and normalizes delayed gastric emptying in obese/T2D female mice.

In conclusion, for the first time, our data demonstrate that overexpression of GSK-3β causes delayed gastric emptying in HFD female mice. Our data further demonstrate that elevation of GSK-3β reduces gastric PI3K/Akt/Nrf2-Phase II enzymes, BH₄-nNOSα, and increases caspase 3, Tau, and β-catenin protein expressions in these mice. Finally, our data indicate that suppression of GSK-3β by SB 216763 attenuated the above pathways and normalizes NO-mediated gastric emptying in HFD female mice (Fig. 10).

Fig. 10.

The schematic diagram represents the mechanistic role of GSK-3β inhibitor SB216763 (SB) on the regulation of Nrf2 [nuclear factor (erythroid-derived 2)-like 2] (via PI3K/Akt-mediated pathway, neuronal nitric oxide synthase (nNOS) induced gastric motility and gastric emptying in obesity/type 2 diabetes (T2D) female mice. Treatment with SB restores delayed gastric emptying via 1) PI3K/AKT, 2) increasing phosphorylation of GSK-3β, 3) Nrf2 and phase II antioxidant enzymes, 4) BH₄ (cofactor of nNOS) biosynthesis enzyme DHFR (salvage), and 4) nNOSα protein and dimerization in Obese/T2D female mice. Inactivation of GSK-3β leads to a reduction in β-catenin degradation. Activation of AKT phosphorylation by SB treatment prevents the caspase-3 expression in ENC. Arrows (↑) indicate activation, whereas red bars (┴) indicate inhibition.

GRANTS

The National Institute of General Medical Sciences of the National Institutes of Health (NIH) under award number SC1GM121282 (to P. Gangula) supported the research reported in this publication. This work also was partly supported by NIH grants R01DK080684 (to S. Sampath) and a Veterans Affairs Research and Development Merit Review Award BX000136- 08 (to S. Sampath).

DISCLOSURES

P. Gangula filed a patent application on BH₄ use for diabetic gastroparesis from University of Texas Medical Branch at Galveston, TX. Otherwise, no conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

P.R.G. conceived and designed research; C.S. performed experiments; C.S. analyzed data; C.S. and P.R.G. interpreted results of experiments; C.S. prepared figures; C.S. and P.R.G. drafted manuscript; S.S., M.L.F., and P.R.G. edited and revised manuscript; P.R.G. approved final version of manuscript.