TGR5-mediated bile acid sensing controls glucose homeostasis

Abstract

TGR5 is a G protein-coupled receptor expressed in brown adipose tissue and muscle, where its activation by bile acids triggers an increase in energy expenditure and attenuates diet-induced obesity. Using a combination of pharmacological and genetic gain- and loss-of-function studies in vivo, we show here that TGR5 signaling induces intestinal glucagon-like peptide-1 (GLP-1) release, leading to improved liver and pancreatic function and enhanced glucose tolerance in obese mice. In addition, we show that the induction of GLP-1 release in enteroendocrine cells by 6alpha-ethyl-23(S)-methyl-cholic acid (EMCA, INT-777), a specific TGR5 agonist, is linked to an increase of the intracellular ATP/ADP ratio and a subsequent rise in intracellular calcium mobilization. Altogether, these data show that the TGR5 signaling pathway is critical in regulating intestinal GLP-1 secretion in vivo, and suggest that pharmacological targeting of TGR5 may constitute a promising incretin-based strategy for the treatment of diabesity and associated metabolic disorders.

Figure 1. Activation of TGR5 signaling pathway activates mitochondrial oxidative phosphorylation and increases ATP/ADP ratio in enteroendocrine L cells

A. Activity of 6α-ethyl-23(S)-methyl cholic acid (INT-777) and cholic acid (CA) on TGR5 in CHO cells transiently transfected with human TGR5 expression vector and a Cre-Luc reporter vector. EC₅₀ values are expressed as percent of the activity of 10 µM of LCA (n=3). B. Chemical structure of CA and INT-777 and respective TGR5 EC₅₀. C. Intracellular cAMP levels in STC-1 cells transfected with control or mTGR5 shRNA for 36 hr and treated with INT-777 at the concentrations indicated (n=3). D. Correlation plots for liver mRNA expression of TGR5 and CoxVI1a in the mouse BxD genetic reference population (n=41) as found in www.genenetwork.org and described in the supplemental methods. E. Cytochrome c oxidase activity in STC-1 cells treated for 1hr with INT-777 at the concentration indicated. Vehicle or adenylate cyclase inhibitor MDL-12330-A (MDL) (1 µM) was added 15min prior to treatment (n=3). F. Oxygen consumption in STC-1 cells was measured using the XF24 extracellular flux analyzer (Seahorse Biosciences). The first vertical dotted line indicates the addition of vehicle or MDL-12330-A (MDL) to culture medium, the second dotted line depicts the treatment with INT-777 at 1 µM (n=10). G. ATP/ADP ratio in STC-1 cells treated as in panel (n=3). H. Correlation plots for liver mRNA expression of TGR5 and Kir6.2 in the mouse BxD genetic reference population according to a similar strategy as described in D. I. mRNA expression levels of TGR5, CoxIV, and Kir6.2 in STC-1 cells transfected for 36h with control or mTGR5 shRNA was measured by real-time quantitative PCR. Target mRNA levels were normalized to 36B4 mRNA levels (n=3). The data are represented as mean±SE. Student’s unpaired t-test. * P < 0.05.

Figure 2. Activation of TGR5 signaling pathway increases intracellular calcium level and stimulates GLP-1 release in enteroendocrine L cells

A. Correlation plots for liver mRNA expression of TGR5 and Cav2.2 in the mouse BxD genetic reference population (n=41) as found in www.genenetwork.org and as described in the supplemental methods. B–C. Intracellular calcium level in NCI-H716 cells transfected with mock vector, hTGR5 expression vector or hTGR5 siRNA for 36 hr and treated with 1 µM (B) or 10 µM (C) of INT-777. The arrow represents INT-777 treatment (n=3). D. Intracellular calcium level in NCI-H716 cells treated with 3μM of INT-777 (indicated by the arrow) in the presence of vehicle or adenylate cyclase inhibitor MDL-12330-A (MDL) (10 µM). MDL or vehicle were added 15 min prior to INT-777 treatment (n=3). E. Intracellular calcium level in NCI-H716 cells treated with 1% glucose and then with 1 µM of INT-777 (n=3). F. GLP-1 release in NCI-H716 cells treated with 1% glucose or 1 µM of INT-777, or a combination of both agents (n=3). G: GLP-1 release in STC-1 cells transfected for 36h with control, mTGR5 expression vector or mTGR5 shRNA and then exposed 30 min to INT-777 at the indicated concentration. A DPP4 inhibitor (Millipore) was added into culture medium at 0.1% (n=3). H. Impact of 30 min of INT-777 treatment on GLP-1 release in STC-1 cells transfected with mTGR5 expression vector in the presence of vehicle or adenylate cyclase inhibitor MDL-12330-A (10 µM). MDL or vehicle were added 15 min prior to INT-777 treatment. A DPP4 inhibitor (Millipore) was added into culture medium at 0.1% (n=3). The data are represented as mean±SE. Student’s unpaired t-test. * P < 0.05 vehicle vs INT-777 treatment, ^# P < 0.05 vehicle vs MDL-12330-A treatment.

Figure 3. TGR5 signaling pathway modulates GLP-1 secretion in vivo

A. Oral glucose tolerance test (OGTT) in 10-weeks HF-fed TGR5-Tg male mice and CD/HF-fed age-matched male littermates. Body weight of TGR5-Tg and control littermates was 37.9±1.7g and 37.0±1.8g, respectively (n=8; not statistically different). The adjacent bar graph represents the average area under the curve (AUC) (n=8). B–C. Plasma levels of insulin (top panel) and GLP-1 (bottom panel) during OGTT (B) or before and after a test meal challenge (C) (n=8). D. GLP-1 release from ileal explants isolated from 18-weeks HF-fed control and TGR5-Tg male mice and exposed for 1hr to the indicated concentrations of LCA (n=4). E. Representative immunofluorescent insulin-stained pancreatic sections from CD and 20-weeks HF-fed control and TGR5-Tg male mice. F. Distribution profile of pancreatic islets from 20-weeks HF-fed control and TGR5-Tg male mice. Islets were counted and sized by the ImageJ analysis software on 4 hematoxylin eosin-stained alternated pancreatic sections spaced each by 150 µM (n=5). G. Insulin content in collagenase-isolated pancreatic islets from 20-weeks HF-fed control and TGR5-Tg male mice (n=5). H. OGTT in 8-weeks HF-fed TGR5^−/− and TGR5^+/+ mice. The inset represents the average AUC. Body weight of TGR5^+/+ and TGR5^−/− male mice at time of analysis was 46.3±3.9g and 51.9±2.0g, respectively (n=8; not statistically different). I–J. Plasma GLP-1 levels in CD-fed TGR5^+/+ (I) and TGR5^−/− mice (J) after an oral glucose challenge preceded 30min before by the oral administration of saline or INT-777 (30 mg/kg), alone or in combination with a dipeptidyl-peptidase-4 inhibitor (DPP4i, 3mg/kg) (n=6). The data are represented as mean±SE. Student’s unpaired t-test. * (P < 0.05) HF-fed compared to HF-fed INT-777-treated mice and ^# (P < 0.05) HF-fed vs CD-fed mice except for panels I and J where * assessed saline or DPP4i treated mice vs INT-777 or INT-777 + DPP4i treated mice and ^# saline vs DPP4i treated mice.

Figure 4. The TGR5 agonist INT-777 increases energy expenditure and reduces hepatic steatosis and adiposity in DIO mice

A. Measurement by HPLC of plasma INT-777 levels in CD-, HF- and HF-fed INT-777-treated male C57BL6/J mice. B. Dietary intervention with INT-777 (30mg/kg/d) was started after a 14-week period of HF feeding at the time indicated by the arrow. Body weight evolution in all groups was followed throughout the study (n=8). C. Body composition was assessed by qNMR after 8 weeks of dietary intervention (n=8). D. Organ mass was expressed as percent of the weight of CD-fed control mice. E. Food intake (n=8). F. Spontaneous horizontal activity and energy expenditure, evaluated by the measurement of oxygen consumption (VO₂) and carbon dioxide release (VCO₂) were monitored over a 18hr period 6 weeks after the initiation of the dietary intervention. The respiratory quotient (RQ) was calculated as the ratio VCO₂/VO₂. Bar graphs represent the average AUC. For the RQ, bar graphs represent the average (n=8). G. Gene expression in BAT by real-time quantitative PCR. Target mRNA levels were normalized to 36B4 mRNA levels (n=8). H. Primary brown adipocytes isolated from CD-fed C57BL/6J male mice were cultured for 12hr with vehicle or 3 µM of INT-777 and O₂ consumption was measured using the XF24 extracellular flux analyzer (Seahorse Biosciences) (n=5). The dotted lines illustrate the addition of the uncoupling agent FCCP at successive doses of 250 nM and 500 nM. I. Representative pictures of oil-red-O (ORO) staining of cryosections (top panel) and sirius red staining of paraffin-embedded sections (bottom panel) of the liver at the end of the dietary intervention. Fibrosis is indicated by the arrow J. Lipid content in liver samples extracted according to the Folch’s method (n=8). K–L. Plasma levels of liver enzymes (K) and lipids (L) at the end of the dietary intervention (n=8). The data are represented as the mean ± SE. Student’s unpaired t-test. * (P < 0.05) HF-fed compared to HF-fed INT-777 treated mice and ^# (P < 0.05) HF-fed vs CD-fed mice.

Figure 5. The TGR5 agonist INT-777 improves insulin sensitivity in obese mice

A. OGTT in CD- and HF-fed male C57BL6/J mice supplemented with 30mg/kg/d INT-777 for 8 weeks following the onset of obesity induced by feeding a HF diet during 10 weeks. The inset represents the average AUC. Body weight of vehicle and INT-777 treated mice was 38.08±1.83g and 32.26±0.95g, respectively (n=8; P < 0.05). B. Fasting glycemia and insulinemia (4h fasting) in DIO mice after 3 weeks of dietary intervention with INT-777 (top panel). Plasma insulin levels during OGTT in DIO mice (bottom panel). C. OGTT in 14-week-old CD-fed db/db male mice treated with 30mg/kg/d INT-777 for 6 weeks. The inset shows the average AUC (n=8). D. Fasting (4h) glycemia and insulinemia in db/db mice after 6 weeks of treatment with INT-777 (top panel). Plasma insulin levels during OGTT in DIO mice (bottom panel). E. Insulin sensitivity evaluated through the average glucose infusion rate at equilibrium (euglycemia) in a hyperinsulinemic euglycemic clamp (10mU insulin/min/kg) in DIO mice (following the onset of obesity induced by feeding a HF diet during 10 weeks) after 10 weeks of dietary intervention with INT-777 (30mg/kg/d) (n=5). The evaluation of liver glucose production and its suppression by insulin, as well as the rate of glucose disappearance, was assessed at equilibrium using ³H-glucose (n=5). F. Insulin-stimulated glucose uptake in the indicated tissues was measured using ¹⁴C-2-deoxy-glucose tracers (n=5). G. Gene expression profiling in liver was performed by real-time quantitative PCR. Target mRNA levels were normalized to 36B4 levels (n=8). The data are represented as mean±SE. Student unpaired t-test. * (P < 0.05) HF-fed compared to HF-fed INT-777 treated mice and ^# (P < 0.05) HF-fed vs CD-fed mice.

Figure 6. INT-777 mediated improvement of glucose tolerance is TGR5-dependent

A. TGR5^+/+ and TGR5^−/− male mice were fed a HF diet for 9 weeks and a first OGTT was performed thereafter. HF was then supplemented with INT-777 at 30mg/kg/d. A second OGTT was performed 4 weeks after treatment with INT-777 was initiated. Curves represent glucose tolerance before and after 4 weeks treatment with INT-777 in TGR5^+/+ (left panel) and TGR5^−/− (right panel) mice. The inset represents the average AUC. In TGR5^+/+ mice, body weight before and after INT-777 treatment was 46.86±3.54g and 43.50±3.47g, respectively (n=8; not statistically different). In TGR5^−/− mice, body weight before and after INT-777 treatment was 54.34±2.23g and 52.30±2.72g, respectively (n=8; not statistically different). B. Plasma insulin levels were concurrently measured during the OGTT in DIO in TGR5^+/+ (left panel) and TGR5^−/− (right panel) mice before and after 4 weeks treatment with INT-777. The inset represents the average AUC (n=8). The data are represented as mean±SE. Student’s unpaired t-test. * (P < 0.05) Vehicle compared to INT-777 treated mice

Figure 7. TGR5-mediated bile acid sensing maintains metabolic function

In brown adipose tissue (mouse) and in skeletal muscle (human), TGR5 activation triggers an increase in mitochondrial oxidative phosphorylation (OXPHOS) which results in energy expenditure and helps prevent obesity in mice treated with TGR5 agonists. Here, we demonstrated that in enteroendocrine L-cells, TGR5 activation also triggers an increase in mitochondrial OXPHOS, which is associated to a rise in the ATP/ADP ratio and a subsequent closure of the ATP-dependent potassium channel (K_ATP) and calcium mobilization (Ca_v). As a consequence, release of the incretin glucagon-like peptide-1 (GLP-1) is increased which helps explain the improvement of glucose homeostasis in obese mice treated with a TGR5 agonist.