G protein-coupled bile acid receptor plays a key role in bile acid metabolism and fasting-induced hepatic steatosis

Abstract

Bile acids are signaling molecules that play a critical role in regulation of hepatic metabolic homeostasis by activating nuclear farnesoid X receptor (Fxr) and membrane G-protein-coupled receptor (Tgr5). The role of FXR in regulation of bile acid synthesis and hepatic metabolism has been studied extensively. However, the role of TGR5 in hepatic metabolism has not been explored. The liver plays a central role in lipid metabolism and impaired response to fasting and feeding contributes to steatosis and non-alcoholic fatty liver and obesity. We have performed a detailed analysis of gallbladder bile acid and lipid metabolism in Tgr5^−/− mice in both free-fed and fasted conditions. Lipid profiles of serum, liver and adipose tissues, bile acid composition, energy metabolism, and mRNA and protein expression of the genes involved in lipid metabolism were analyzed. Results showed that deficiency of the Tgr5 gene in mice alleviated fasting-induced hepatic lipid accumulation. Expression of liver oxysterol 7α-hydroxylase (Cyp7b1) in the alternative bile acid synthesis pathway was reduced. Analysis of gallbladder bile acid composition showed marked increase of tauro-cholic acid and decrease of tauro-α and β-muricholic acid in Tgr5^−/− mice. Tgr5^−/− mice had increased hepatic fatty acid oxidation rate and decreased hepatic fatty acid uptake. Interestingly, fasting induction of fibroblast growth factor 21 (Fgf21) in liver was attenuated. In addition, fasted Tgr5^−/− mice had increased activation of hepatic growth hormone-signal transducer and activator of transcription 5 (GH-Stat5) signaling compared to wild type mice.

Conclusion

This study suggests that TGR5 may play a role in determining bile acid composition and in fasting-induced hepatic steatosis through a novel mechanism involving activation of the GH-Stat5 signaling pathway.

Introduction

Bile acids are signaling molecules that play a key role in regulation of metabolic homeostasis via activation of a nuclear bile acid receptor farnesoid- X- receptor (Fxr) and membrane G protein-coupled bile acid receptor (Gpbar-1, aka Tgr5) ^{1, 2}. Bile acids are synthesized from cholesterol exclusively in the liver, secreted into bile and stored in the gallbladder and released to the intestinal tract for emulsifying fats and lipid-soluble vitamins. The classic bile acid synthesis pathway is initiated by cholesterol 7α-hydroxylase (Cyp7a1) for synthesis of two primary bile acids, cholic acid (CA), and chenodeoxycholic acid (CDCA). Sterol 12α-hydroxylase (Cyp8b1) is required for synthesis of CA and sterol 27-hydroxylase (Cyp27a1) is involved in steroid side-chain oxidation ³. In mouse liver, CDCA is 6β-hydroxylated to α-muricholic acid (α-MCA), which is epimerized to β-muricholic acid (β-MCA). The alternative bile acid synthesis pathway is initiated by Cyp27a1 and involved oxysterol 7α-hydroxylase (Cyp7b1). Bile acids are conjugated to the amino acids taurine or glycine before biliary secretion. In the intestine, conjugated CA and CDCA are de-conjugated by bacterial bile salt hydrolases and then de-hydroxylated by bacterial 7α-dehydroxylase to form deoxycholic acid (DCA) and lithocholic acid (LCA), respectively. Accumulation of high amounts of toxic bile acids could cause liver inflammation and injury and cholestatic liver diseases. Thus, bile acid synthesis is tightly regulated by bile acids returning from the ileum to hepatocytes to inhibit Cyp7a1 and Cyp8b1 gene transcription by two mechanisms. In the intestine, bile acids activate FXR to induce fibroblast growth factor 15 (Fgf15), which is circulated to the liver to activate the membrane Fgf receptor 4 signaling to inhibit Cyp7a1 and Cyp8b1 gene transcription. In the liver, bile acids activate FXR to induce the negative receptor small heterodimer partner (Shp) to inhibit Cyp7a1 and Cyp8b1 gene transcription.

The liver plays a central role in lipid metabolism during fasting and re-feeding cycles. During the postprandial state, the liver uptakes dietary fats to synthesize triglycerides for VLDL-mediated secretion and transport to adipose tissue for storage and to muscles for energy metabolism. During fasting, fatty acids are mobilized from adipocytes to the liver for energy metabolism. Prolonged fasting is known to induce hepatic steatosis and changes in endocrine and paracrine factors such as thyroid hormone, insulin, glucagon, growth hormone (GH) and fibroblast growth factor 21 (Fgf21) and these hormones regulate hepatic glucose, lipid and energy metabolism and maintain metabolic homeostasis ^4-6. The role of Fxr in regulation of metabolic homeostasis has been studied extensively. However, the role of Tgr5 in liver metabolism has not been explored. Tgr5 is expressed in many tissues including intestine, gallbladder, liver, and brain ^7-11. In the liver, Tgr5 is expressed in Kupffer cells and sinusoidal endothelial cells, but not in hepatocytes ^{10, 11}. In the gastrointestinal tract, activation of Tgr5 by its specific agonists protects intestinal barrier function and inflammation, stimulates gallbladder refilling, stimulates glucagon like peptide-1 (GLP-1) secretion from enteroendocrine L-cells ¹², which stimulates insulin secretion from pancreatic β cells ¹³. In brown adipocytes, activation of Tgr5 induces thyroid hormone deiodinase 2 (Dio2), which activates thyroid hormone to stimulate energy metabolism ¹⁴. Secondary bile acids LCA and DCA are efficacious activators of Tgr5 to induce cAMP and activate protein kinase A signaling pathways ^{7, 8}. Activation of Tgr5 by specific Tgr5 agonists has been shown to alleviate obesity and hepatic steatosis by maintaining glucose homeostasis in diet-induced obese mice ¹⁵. However, Tgr5 deficient (Tgr5^−/−) mice do not have abnormal metabolic phenotypes ¹⁶ and are protected from lithogenic diet-induced cholesterol gallstone formation ¹⁷. Similar to Fxr^−/− mice, Tgr5^−/− mice have impaired liver regeneration suggesting that both bile acid receptors are important in lipid metabolism during liver regeneration ¹⁸.

To understand the role of Tgr5 in liver lipid metabolism, we have performed a detailed analysis of bile acid and lipid metabolism in Tgr5^−/− mice in both free-fed and fasted conditions. We hypothesized that Tgr5 might play a key role in bile acid synthesis and lipid homeostasis. Our data show that gallbladder bile acid composition of Tgr5^−/− mice was markedly altered. Surprisingly, we found that Tgr5^−/− mice were protected from fasting-induced hepatic steatosis by reducing hepatic fatty acid uptake, and increasing hepatic fatty acid oxidation to ameliorate hepatic steatosis, likely through activation of a GH-signal transducer and activator of transcription (Stat5) signaling.

Materials and Methods

Animal experiments

Wild type male C57BL/6J (WT) mice were purchased from Jackson Laboratories (Bar Harbor, ME) and were maintained on regular chow diet and water ad libitum (unless specified) in a temperature-controlled facility with a 12 h light/12 h dark schedule (6 am to 6 pm light). Tgr5 knock out (Tgr5^−/−) mice were obtained from Merck Research Laboratories (Kenilworth, NJ) and bred to C57BL/6J background. Twenty to twenty-one week old WT mice (n=12) or Tgr5^−/− mice (n=12) were separated to two groups (6 WT and 6 Tgr5−/− in each group). One group was free-fed chow diet and the other group was fasted for 24 h. Mice were euthanized at 12 pm and tissues were collected and stored in −80°C until analysis. To study the role of the Jak2/Stat pathway on hepatic lipid metabolism, mice were given a Jak2 inhibitor, tyrphostin Ag 490 (5mg/Kg) (Sigma-Aldrich, St. Louise, MO) or vehicle (phosphate buffered saline, PBS), intra-peritoneally. Mice were fasted for 21 h and were given a second dose of Ag490 (5mg/kg) or vehicle and sacrificed 3 h later. All animal experiments were performed at Northeast Ohio Medical University (NEOMED) and protocols approved by the Institutional Animal Care and Use Committee.

Hepatic β-oxidation rate analysis

WT and Tgr5^−/− male mice were fasted for 24 h, euthanized and livers were collected. Hepatic fatty acid oxidation rate was assessed using [1-¹⁴C] palmitic acid ¹⁹. Intact mitochondria were isolated from mouse liver tissue using sucrose Tris-EDTA buffer. [1-¹⁴C] Palmitic acid (0.4 μCi) was suspended in 40 μl of 7% BSA solution at 37°C and added to 330 μl incubation mixture (100 mM sucrose, 10 mM Tris HCl, pH 7.4, 5 mM KH₂PO₄, 0.2 mM EDTA, pH8, 80 mM KCl, 1 mM MgCl₂, 2 mM L-carnitine, 0.1 mM Malate, 0.05 mM Coenzyme A, 2 mM ATP, 1 mM DTT, pH 8.0) and 30 μl of mitochondria protein solution. The mixture were incubated at 37°C for 30 min, transferred to a new micro centrifuge tube with 200 μl of perchloric acid (1 M) and pre-incubated Whatman filter paper discs in the cap. Samples were incubated for 1 h to entrap CO₂ released from reaction mixture. After incubation, discs were transferred to scintillation vials and CO₂ released was analyzed using a scintillation counter. Micro centrifuge tubes were centrifuged at 14,000 rpm for 10 min and 400 μl supernatant was used to determine acid soluble metabolite (ASM) levels formed during reaction. Data generated was expressed in nmole/mg/min of palmitic acid converted to CO₂ and ASM.

Hepatic fatty acid uptake analysis

Hepatic fatty acid uptake assay was performed as described ^{20, 21}. WT and Tgr5^−/− male mice were anesthetized with ketamine and xyalzine following overnight fasting. ³H-labeled fatty acids (30 μCi, ³H-oleic acid, Perkin-Elmer, Waltham, MA) were dissolved in 200 μl of bovine serum albumin solution (2 mg/ml in saline), which was injected to mice via tail vein. At 1 and 10 min post injection, mice were sacrificed and liver, skeletal muscle, adipose tissue and blood were collected and radioactivity was determined using a scintillation counter. Radioactivity of liver was corrected for radioactivity present in serum during sampling and was normalized with tissue weight, whereas for skeletal muscle radioactivity of tissue was normalized with tissue weight, and expressed as percent of total radioactivity injected to each mice.

Comprehensive Laboratory Animal Monitoring System (CLAMS)

Comprehensive metabolic profiling of 20-21 week old male WT and Tgr5^−/− mice during 24 h fasting was performed using an Oxymax lab animal monitoring system (Columbus Instruments, Columbus, OH). Respiratory exchange ratio, O₂ consumption and CO₂ production, activity and energy expenditure of mice were measured during 24 h fasting.

Statistics

The statistical significance between groups was determined using one-way ANOVA followed by a Turkey post hoc test or t-test, using GraphPad Prism® software (GraphPad Software Inc., CA). Data are presented as mean ± SE, with p ≤0.05 considered statistically significant.

Results

Tgr5−/− mice were resistant to fasting-induced hepatic steatosis

Prolonged fasting is known to induce hepatic steatosis by accumulation of adipose tissue-derived free fatty acids (FFA) in hepatocytes ²². Extended fasting for 24 h significantly decreased body weights in both WT and Tgr5^−/− mice, but no significant change was observed between WT and Tgr5^−/− mice after fasting (Supplemental Fig.1A). No significant changes in liver weight were observed between free-fed and fasted WT mice, whereas liver weight in Tgr5^−/− mice was significantly decreased after fasting (Supplemental Fig.1A). White adipose tissue (epididymal fat) weight decreased in both fasted WT and Tgr5^−/− mice (Supplemental Fig.1A). Fasting did not affect serum triglyceride levels between WT and Tgr5^−/− mice. Fasting increased serum FFA by 3-fold in WT mice, but Tgr5^−/− mice had significantly lower circulating FFA compared to WT mice (Supplemental Fig.1B). Serum cholesterol levels in Tgr5^−/− mice were significantly higher compared to WT mice in both free-fed and fasted conditions (Supplemental Fig.1B). Fasting induced serum β-hydroxybutyrate levels in both WT and Tgr5^−/− mice by ~10-fold compared to free-fed mice (Supplemental Fig.1B).

Fasting significantly increased liver triglycerides and FFA levels in both WT and Tgr5^−/− mice (Fig. 1A). This effect was attenuated in fasted Tgr5^−/− mice compared to fasted WT mice. Hepatic cholesterol levels show a small but significant increase in fasted WT mice, but no significant change in Tgr5^−/− mice. Hepatic glycogen content was significantly higher in free-fed Tgr5^−/− mice compared to WT mice indicating increased glucose storage as glycogen in Tgr5^−/− mice, but no difference in fasted mice. Fasting significantly decreased hepatic glycogen content in both WT and Tgr5^−/− mice. Oil-red-O staining for neutral lipids (Fig.1B) showed fasting markedly increased lipids in WT mice but Tgr5^−/− mice had significantly lower lipid content. These results are surprising and may suggest that lack of Tgr5 protects the liver from fasting-induced hepatic steatosis. In white adipose tissue and skeletal muscle, triglyceride levels were decreased in fasted Tgr5^−/− mice, but not WT mice (Fig. 1C & 1D). FFAs were significantly decreased in skeletal muscle (Fig. 1D) but not white adipose tissue (Fig. 1C) of fasted Tgr5^−/− mice.

Fig. 1. Effect of Tgr5 on fasting-induced lipid metabolism.

(A) Hepatic triglycerides, free fatty acids, cholesterol and glycogen levels in free-fed (FF) and fasted WT and Tgr5^−/− mice. (B) Oil-Red-O staining and quantification of liver neural lipids of fed and fasted WT and Tgr5−/− mice. (C) White adipose tissue triglyceride and free fatty acid (FFA) levels in fed and fasted mice. (D) Skeletal muscle triglyceride and FFA levels in FF and fasted mice. An “*” indicates significant difference between FF and fasted condition (p<0.05). A “#” indicates significant difference between WT and Tgr5^−/− mice in either FF or fasted condition (p<0.05).

Tgr5^−/− mice switched to the classic bile acid synthesis pathway to generate a hydrophobic bile acid pool

To understand the role of Tgr5 in bile acid metabolism, we analyzed hepatic bile acid synthesis gene expression and bile acid composition in gallbladder bile. Fig. 2A shows that Cyp7a1 mRNA levels were the same in the fasted and free-fed states in both WT and Tgr5^−/− mice. Feeding should induce Cyp7a1 mRNA expression ²³. Unchanged Cyp7a1 expression in free-fed and fasted mice may be because mice were sacrificed at 12 pm, the zenith of the Cyp7a1 circadian rhythm. In free-fed mice, Cyp7b1 and Cyp27a1 mRNA levels were significantly decreased in Tgr5^−/− mice compared to WT mice, and fasting significantly decreased Cyp7b1 and Cyp27a1 expression in both WT and Tgr5^−/− mice (Fig. 2A). Fasting induced Cyp8b1 mRNA in WT mice as expected ²⁴. However, fasting induction of Cyp8b1 was significantly attenuated in Tgr5^−/− mice compared to WT mice (Fig. 2A). We performed immunoblot analysis to study if changes of mRNA expression levels of Cyp7b1 and Cyp8b1 are consistent to their protein expression levels. Fig 2B showed that Cyp8b1 protein levels were induced by fasting in both wild type and Tgr5^−/− mice. Expression of Cyp7b1 is known to exhibit sexual dimorphisms, with higher expression in male than female rodents ²⁵. Fasting reduced Cyp7b1 protein levels in both wild type and Tgr5^−/− mice (Fig 2B), consistent with reduced Cyp7b1 mRNA levels by fasting. During fasting, bile acids are stored in the gallbladder and bile acid composition in bile represents de novo synthesized bile acids and bile acids circulated back to the liver via enterohepatic circulation. We therefore analyzed bile acid composition in gallbladder bile. Data showed a marked increase of TCA and TDCA and decrease of Tα-MCA and Tβ-MCA in both free-fed and fasted Tgr5^−/− mice compared to WT mice (Fig. 2C). TCA, TDCA, and Tα-MCA and Tβ-MCA account for more than 90% of total bile acids in mouse bile. Other minor bile acid species were also increased in both free-fed and fasted Tgr5^−/− mice. These changes in bile acid composition are consistent with decreased expression of Cyp7b1 in Tgr5^−/− mice, indicating that the alternative bile acid synthesis was suppressed and the classic bile acid synthesis pathway was stimulated to produce more CA and less CDCA derived Tβ-MCA. As the result, the hydrophobicity index of bile acid pool increases from −0.418 in WT mice to −0.22 in Tgr5^−/− mice. The ratio of 12α-hydroxylated bile acids (TCA and TDCA) to non-12α-hydroxylated bile acids (TCDCA and Tα-MCA and Tβ-MCA) increases from 0.71 in WT mice to 1.68 in Tgr5^−/− mice. It is interesting to point out that bile acid composition seems to dependent on genotype, and fasting/feeding does not affect bile acid composition except TDCA (%) increased two-fold by fasting in both wild type and Tgr5^−/− mice. TDCA is a preferred ligand for TGR5 ⁸ indicating that TGR5 signaling may be stimulated in fasted mice.

Fig. 2. Effect of Tgr5 on hepatic bile acid metabolism.

(A) Relative mRNA expression of hepatic bile acid synthesis gene in free-fed (FF) and fasted WT and Tgr5^−/− mice. (B) Immunoblot analysis and quantification of Cyp8b1 and Cyp7b1 protein levels in FF and fasted male mice. Each lane represents one mouse (n=3 per group). (C) Bile acid composition of gallbladder bile in FF and fasted mice. Bile acids are quantified as % of total bile acids analyzed (100%), < 0.5% is rounded to 0. An “*” indicates significant difference between FF and fasted condition (p<0.05). A “#” indicates significant difference between WT and Tgr5^−/− mice in either FF or fasted condition (p<0.05). A “†” indicates significant difference compared to WT male mice.

Tgr5^−/− mice had increased fatty acid oxidation

Fasting significantly decreased expression of lipogenic genes, such as sterol regulatory element-binding protein-1 (Srebp-1), acetyl-CoA carboxylase alpha (Acc-α), fatty acid synthase (Fasn) and stearoyl-CoA decarboxylase-1 (Scd-1) in both WT and Tgr5^−/− mice compared to free-fed mice (Fig. 3A). Hepatic mRNA expression of Acc-α was increased 3-fold in free-fed Tgr5^−/− mice compared to WT mice. Fasting increased hepatic carnitine palmitoyltransferase 1 α (Cpt1α), involved in fatty acid β-oxidation, and decreased lipoprotein lipase (Lpl) expression in both WT and Tgr5^−/− mice compared to free-fed mice (Fig. 3B). Protein analysis showed significant increase in Cpt1 protein levels only in Tgr5^−/− mice (Fig. 3B) indicating increased hepatic fatty acid oxidation. Peroxisome proliferator-activated receptor α (Ppar-α) levels were not altered in either free-fed or fasted WT and Tgr5^−/− mice. Ppar-gamma co-activator- 1α (Pgc-1α) levels were induced in fasted WT mice, and further increased in fasted Tgr5^−/− mice (Fig. 3B). Tgr5^−/− mice had significantly increased rate of palmitic acid oxidation to CO₂ and acid soluble material (ASM) compared to WT mice in the fasted condition (Fig. 3C). Furthermore, expression levels of liver Fgf21 mRNA and serum Fgf21 were drastically induced during fasting in WT mice ⁵, but were significantly lower in Tgr5^−/− mice (Fig. 3D). Since hepatic autophagy plays an important role in lipolysis during fasting ²⁶, we analyzed expression of the hepatic autophagy marker microtubule-associated protein 1 light chain 3 II (Lc3II), which increased 3.5-fold in fasted Tgr5^−/− mice (Fig. 3E).

Fig. 3. Effect of Tgr5 on fasting-induced hepatic lipid metabolism.

(A) Relative mRNA expression of hepatic fatty acid synthesis genes in free-fed (FF) and fasted WT and Tgr5^−/− mice. (B) Relative mRNA expression of hepatic lipolysis genes in FF and fasted mice, and western blot analysis and quantification of Cpt1 protein levels in FF and fasted mice. Each lane represents one mouse (n=3 per group). (C) Hepatic β-oxidation rates in fasted mice. Data was presented as nmoles of ¹⁴C-palmitic acid converted into CO₂ and acid soluble material (ASM). (D) Relative expression of hepatic Fgf21 mRNA and Western blot analysis of serum Fgf21 protein levels in FF and fasted mice. (E) Western blot analysis of hepatic autophagy Lc3 I and Lc3 II protein and quantitation of Lc3II levels in FF and fasted mice. Each lane represents one mouse (n=3 per group). An “*” indicates significant difference between FF and fasted condition (p<0.05). A “#” indicates significant difference between WT and Tgr5^−/− mice in either FF or fasted condition (p<0.05).

Tgr5^−/− mice had altered energy expenditure

As Tgr5 is known to regulate energy metabolism and hormones levels, we analyzed serum hormone levels in fed and fasted mice. Tgr5^−/− mice have increased serum insulin levels only during free-fed conditions and decreased thyroxin (T₃) hormone levels only during fasting conditions compared to WT mice (Supplementary Fig. 2A and 2B). No significant difference in fasting glucagon levels were observed between WT and Tgr5^−/− fasted mice (Supplementary Fig. 2C).

To understand energy metabolism involved in attenuation of hepatic steatosis in Tgr5^−/− mice, we examined respiratory exchange ratio (RER), energy expenditure and physical activity using CLAMS. Oxygen consumption rate (VO₂) and carbon dioxide production rate (VCO₂) were significantly increased in fasted Tgr5^−/− mice during both day and night compared to WT mice (Fig. 4A). Although both VO₂ and VCO₂ are increasing in Tgr5^−/− mice at both day and night times, RER (VCO₂/VO₂) increased significantly only in daytime and significantly decreased in nighttime in Tgr5^−/− mice compared to WT mice (Fig. 4B). Energy expenditure was increased in fasted Tgr5^−/− mice only during the night (Fig. 4C). Physical activity was increased during both day and night in fasted Tgr5^−/− mice compared to WT mice (Fig. 4D). These data indicate that fasted Tgr5^−/− mice are more active and used more lipids as an energy source (decreased RER) during the active period (night) compared to WT mice. These results are consistent with decreases in hepatic FFA and triglyceride levels and circulating FFA in fasted Tgr5^−/− mice.

Fig. 4. Effect of Tgr5 on oxygen consumption and physiological activity during fasting condition.

(A) VO₂ and VCO₂ levels during fasting in WT and Tgr5^−/− mice. (B) Respiratory exchange ratio during fasting in WT and Tgr5^−/− mice. (C) Energy expenditure during fasting in WT and Tgr5^−/− mice. (D) Activity levels during fasting in WT and Tgr5^−/− mice. A “#” indicates significant difference between WT and Tgr5^−/− mice in fasted condition (p<0.05).

Lack of Tgr5 signaling impaired hepatic fatty acid uptake

To determine whether attenuation of fasting-induced steatosis could be due to altered hepatic lipid uptake in Tgr5^−/− mice, we analyzed hepatic fatty acid translocase (Cd36) and fatty acid transport protein (Fatp) mRNA levels. Fasting significantly increased hepatic Cd36 and Fatp4 mRNA levels in WT mice (Fig. 5A). In Tgr5^−/− mice, CD36 mRNA expression was reduced compared to wild type mice in free fed state. Interestingly, hepatic Cd36 and Fatp4 mRNA levels were significantly down regulated in fasted Tgr5^−/− mice compared to fasted WT mice (Fig. 5A). Fatp5 and Fabp1 mRNA levels were significantly decreased in fasted Tgr5^−/− compared to WT mice. Western blot analysis showed Tgr5^−/− mice had significantly decreased hepatic Cd36 protein levels in both fed and fasted conditions (Fig. 5B). We further determined hepatic VLDL secretion and lipid uptake. Hepatic VLDL secretion was only significantly increased after 3 h in Tgr5^−/− mice compared to WT mice (Fig. 5C). Fatty acid uptake assay showed significant decrease in hepatic fatty acid uptake in Tgr5^−/− mice compared to WT mice at 1 min and 10 min, but increased fatty acid uptake in skeletal muscle. (Fig. 5D). These results suggest that Tgr5^−/− mice may have impaired hepatic fatty acid uptake, but increased fatty acid oxidation and VLDL secretion and increased fatty acid uptake into skeletal muscle to reduce hepatic triglyceride and FFA during fasting.

Fig. 5. Effect of Tgr5 on hepatic lipid uptake and disposition.

(A) Relative mRNA expression of hepatic fatty acid uptake genes in free-fed (FF) and fasted mice. (B) Western blots analysis and quantification of hepatic Cd36 protein levels in FF and fasted mice. Each lane represents one mouse (n=3 per group). (C) Hepatic VLDL secretion in WT and Tgr5^−/− mice. (D) Hepatic and muscle lipid uptake in WT and Tgr5^−/− mice injected ³H-oleic acid. An “*” indicates significant difference between FF and fasted condition (p<0.05). A “#” indicates significant difference between WT and Tgr5^−/− mice in either FF or fasted condition (p<0.05).

Adipose tissue lipolysis in Tgr5^−/− mice

Prolonged fasting activates lipolysis in white adipose tissue resulting in increased circulating FFA and hepatic steatosis; therefore, we analyzed genes involved in lipid metabolism in white adipose tissue. Fasting significantly induced adipose triglyceride lipase (Atgl) mRNA expression, but did not change hormone sensitive lipase (Hsl) and monoglyceride lipase (Mgll) mRNA levels in WT and Tgr5^−/− mice (Fig. 6A). Fasting significantly induced uncoupling protein-1 (Ucp-1) and Ucp-3 mRNA levels in WT mice compared to free-fed mice (Fig. 6A). In Tgr5^−/− mice, fasting significantly increased Ucp-2 mRNA levels compared to free-fed mice. In both WT and Tgr5^−/− mice, fasting did not affect adipose tissue Ppar-γ or its target genes Cd36, Pgc-1α, fatty acid binding protein 4 (Fabp4) expression (Fig. 6A). In fasted wild type mice, Ucp-1, 2 and 3, Cpt1α and Dio2 mRNA levels were increased in brown adipose tissue indicating increase thermogenesis. However, fasting-induced increases of thermogenesis gene expression was lowered in fasted Tgr5^−/− mice compared to fasted wild type mice (Fig. 6B).

Fig. 6. Effect of Tgr5 on fasting-induced adipose tissue lipolysis.

(A) Relative mRNA expression of the genes involved in fatty acid metabolism in white adipose tissue. (B) Relative mRNA expression of brown adipose-specific genes in free-fed (FF) and fasted mice. An “*” indicates significant difference between FF and fasted condition (p<0.05). A “#” indicates significant difference between WT and Tgr5^−/− mice in either FF or fasted condition (p<0.05).

Tgr5^−/− mice had increased GH-Stat5 signaling

It has been reported that Cd36 and hepatic steatosis are negatively regulated by GH-Stat5 signaling ²⁷. We analyzed p-Stat5 and Stat5 levels in WT and Tgr5^−/− mice in fed and fasted conditions. Fasting significantly decreased the p-Stat5/Stat5 ratio in both WT and Tgr5^−/− mice compared to free-fed mice. However, the p-Stat5/Stat5 ratio was significantly higher in Tgr5^−/− mice compared to WT mice in fasting state (Fig. 7A). Fasted Tgr5^−/− mice have significantly higher nuclear Stat5 levels compared to WT fasted mice (Fig. 7B). We also analyzed Ppar-γ target genes involved in lipid droplet formation since Stat5 activation is known to decrease Ppar-γ target gene expression ²⁷. Hepatic aP2 (Fabp4) and cell death-inducing DEFA-like effector c (Cidec) mRNA levels were significantly down regulated in fasted Tgr5^−/− mice compared to WT mice (Fig. 7C). We also analyzed serum GH levels in these mice. Fig. 7D shows that serum GH levels were markedly increased in both fasted WT and Tgr5^−/− mice. However, Tgr5^−/− mice have significantly higher serum GH levels than WT mice (Fig. 7D). These results indicate that TGR5 signaling affect GH levels in both fasting and free fed state. However, Tgr5^−/− mice may have activated hepatic GH-Stat5 signaling to reduce fatty acid transport to hepatocytes, thus reduces lipid droplets accumulation in Tgr5^−/− mice during fasting. GH is known to activate the Janus Kinase 2 (Jak2)/Stat pathway. To further study the GH-Jak2/Stat pathway in hepatic lipid metabolism, we injected mice with a Jak2 inhibitor, Ag490, and fasted fro 24 h. Fig. 7E showed that Ag490 significantly decreased hepatic triglycerides and free fatty acids in fasted WT mice, but did not further reduce hepatic lipids in fasted Tgr5^−/− mice. Ag490 reduced p-Stat5 in hepatocytes (Suppl Fig 3A), tend to increase hepatic Cd36, Fgf21 and Cidec mRNA expression (Supplemental Fig. 3B), and increased phosphorylation of adipocyte hormone-sensitive lipase (p-HSL) (Suppl Fig 3C). Thus Ag490 may reduce adipocyte lipolysis, consistent with a report that deficiency of Jak2 impaired adipose lipolysis ²⁸. These results suggest that similar to adipose-specific Jak2^−/− mice, increased GH in Tgr5−/− mice may also improve hepatic lipid metabolism in fasting.

Fig. 7. Effect of Tgr5 on growth hormone signaling and growth hormone levels during Free-fed (FF) and fasted conditions.

(A) Western blot analysis and quantification of hepatic p-Stat5 and Stat5 protein levels in both WT and Tgr5^−/− mice during FF and fasting conditions. Each lane represents one mouse (n=3 per group). (B) Western blot analysis and quantification of Stat5 hepatic nuclear protein expression in fasted WT and Tgr5^−/− mice. Each lane represents one mouse (n=4 per group). (C) Relative mRNA expression of hepatic Ppar-γ target gene in WT and Tgr5^−/− FF and fasted mice. (D) Serum growth hormone levels in both WT and Tgr5^−/− mice during FF and fasted conditions. (E) Hepatic triglyceride and free fatty acids levels in fasted WT and Tgr5^−/− mice treated with either vehicle or Ag 490. An “*” indicates significant difference between FF and fasted condition (p<0.05). A “#” indicates significant difference between WT and Tgr5^−/− mice in either FF or fasted condition (p<0.05). A “$” Indicates significance between vehicle and Jak2 inhibitor treatment (p<0.05). A “+” indicates significance between WT and Tgr5^−/− mice treated with either vehicle or Jak2 inhibitor (p<0.05).

Discussion

Detailed analysis of bile acid composition in Tgr5^−/− mice revealed a significant increase in TCA and decrease in T-MCAs in bile, suggesting stimulation of the classic bile acid synthesis pathway and suppression of the alternative pathway. This is consistent with decreased expression of the key regulatory genes Cyp27a1 and Cyp7b1 in the alternative bile acid synthesis pathway. Increasing TCA and decreasing TMCAs increases the bile acid hydrophobicity indices in Tgr5^−/− mice. In mice, TCA is the major Fxr ligand and tauro α/β-MCAs are potent Fxr antagonists ²⁹. This study suggests that Tgr5 may play an important role in altering bile acid composition in bile acid pool and lowering of 12α-hydroxylated bile acids, the levels of which have been linked to insulin resistance in diabetic patients ³⁰ and improving glucose sensitivity after vertical sleeve gastrectomy by modulating bile acid composition in circulating bile acids ³¹.

Recent studies show that activation of Fxr decreases fasting-induced hepatic autophagy, a key pathway involved in maintaining nutrient homeostasis during fasting ³². The increase of hepatic Lc3II levels in Tgr5^−/− mice may indicate increase of lipophagy to decrease hepatic triglyceride levels ²⁶ and increased Lc3II levels are correlated to increased GH-Stat5 activation ³³. During fasting, Tgr5^−/− mice have increased hepatic Pgc-1α and Cpt1 and increased palmitic acid oxidation. Fasting increases FFA to stimulate Ppar-α activity and induce hepatic Fgf21 production. This study demonstrates a significant increase of Pgc-1α and decrease of hepatic Fgf21 production in Tgr5^−/− mice. Pgc-1α is known to inhibit hepatic Fgf21 production ³⁴ and a recent study reports that Fgf21 inhibits hepatic GH-Stat5 signaling ³⁵. Therefore, reducing hepatic Fgf21 production stimulates hepatic Stat5 signaling in fasted Tgr5^−/− mice.

The liver is a major target of GH action, and Stat5 signaling activates Jak2/Stat5 signaling to regulate fatty acid oxidation and steroid metabolism ²⁷. Mice expressing mutant GH receptor or deletion of Stat5 had increased hepatic Cd36 expression and hepatic steatosis when challenged with high-fat diet ²⁷. Our study shows that Tgr5^−/− mice have increased serum GH levels compared to WT mice. Chronic infusion of GH activates Ppar-α and prevents hepatic steatosis in diabetic rats ³⁶. It is possible that this GH-Stat5 mechanism may be activated in Tgr5^−/− mice to stimulate hepatic lipolysis. Moreover, plasma GH levels show pulsatile secretion patterns in male rats, mice and humans, but GH is constantly secreted at much lower levels in females. Pulsatile GH secretion results in sexual dimorphic expression of many male-specific liver genes including Cyp7b1 ²⁵. In hepatocyte-specific Stat5-deficient mice, sexual dimorphic Cyp7b1 expression is reduced and hypophysectomy reduces Cyp7b1 expression ²⁵. Pulsatile infusion of GH to hypophysectomized mice increases Cyp7b1 expression, but continuous GH infusion for 5-7 days or disruption of GH pulsatile patterns decreases Cyp7b1 expression in WT male mice ^{25, 37}. It is interesting to note that high-fat diet-fed female Tgr5−/− mice accumulated fat, gained weight and had impaired insulin sensitivity, and male mice are not obese but had increased steatosis ^{16, 38}. It is possible that GH signaling may play a role in gender-specific regulation of bile acid synthesis by Tgr5. Since TGR5 is not expressed in hepatocytes ¹⁰, TGR5 may regulate hepatic metabolism through paracrine or endocrine mechanisms. Bile acids have been detected in the brain ³⁹. TGR5 is expressed in the hypothalamus and may regulate neuroendocrine hormone secretions in cholestasis ^{40, 41}. It has been reported that pituitary controls bile acid metabolism ⁴² and GH plays a role in regulation of bile acid synthesis ⁴³. Thus bile acids may activate pituitary TGR5 to regulate GH secretion in hypothalamus. Further studies are needed to explore the brain to liver axis in TGR5 regulation of liver bile acid metabolism.

It is interesting to point out that TGR5 has both positive and negative effect in hepatic metabolism in fasting and feeding response and hepatic steatosis. Our current study shows Tgr5^−/− mice are protected against fasting-induced hepatic steatosis. Previous studies show Tgr5^−/− mice are protected against lithogenic diet-induced gallstone and insulin tolerance is improved in male Tgr5^−/− mice on chow diet but impaired insulin sensitivity when fed a high fat diet ^{17, 38}. Tgr5^−/− mice have increased serum GH, similar to adipose-specific Jak2^−/− mice with high GH and reduced hepatic fatty liver ²⁸. On the other hand, postprandial activation of TGR5 by bile acids is known to improve insulin secretion from β-cells via stimulating glucagon like peptide-1 (GLP-1) secretion in enteroendocrine L cells ¹², whereas growth hormone counteracts insulin action in fasting. In this study we report altered lipid metabolism in a model, which has no or low insulin signaling. Beneficial effects we observe during fasting condition are due to lack of insulin signaling and increasing GH-Stat5 signaling. Lack of insulin signaling in fasting condition is similar to diabetic condition.

Fig 8 illustrates a mechanistic link of Tgr5 signaling to regulation of bile acid synthesis and lipid metabolism. Pulsatile GH secretion is known to activate Stat5 to induce male-specific Cyp7b1 expression in mice. High levels of serum GH in fed Tgr5^−/− mice may inhibit Cyp7b1 in the alternative bile acid synthesis pathway and results in reducing TMCAs and increasing CA in bile acid pool. GH is known to stimulate lipolysis in adipose tissues and mobilize FFA. In hepatocytes, GH activates GH receptor, which recruits JAK2 to activate Stat5 by phosphorylation. In fasted Tgr5^−/− mice, increased GH activates Stat5 signaling to decrease Cd36 expression and thus reduce FFA uptake into hepatocytes. FFA induces Pgc-1α, which increases fatty acid oxidation, but increase of Pgc-1α in Tgr5^−/− mice dampens fasting-induced production of Fgf21, which is known to reduce hepatic Stat5 signaling and lipophagy to reduce hepatic steatosis in Tgr5^−/− mice. Thus lower Fgf21 levels may also stimulate hepatic Stat5 activation during fasting. This Cd36-Fgf21-Stat5 regulatory loop may be induced in fasting to regulate lipid metabolism.

Fig. 8. Deletion of Tgr5 alters bile acid metabolism and alleviates fasting-induced hepatic steatosis via altering hepatic GH-Stat5 signaling.

Pulsatile release of GH may activate Stat5 to induce male-specific Cyp7b1 expression and the alternative bile acid synthesis pathway. In Tgr5−/− mice, high levels of non-pulsatile GH secretion (GH resistance) may inhibit Cyp7b1 and Cyp27a1 to reduce CDCA and MCA synthesized by the alternative bile acid synthesis pathway and increase CA synthesis. In fasted Tgr5^−/− mice, GH resistance activates Stat5 signaling to decrease Cd36 expression and uptake of free fatty acids into hepatocytes. Free fatty acids induce Pgc-1α, which increases fatty acid oxidation, but increase of Pgc-1α in Tgr5^−/− mice dampens fasting-induced Fgf21 production. Lower Fgf21 levels stimulate hepatic Stat5 activation and reduce hepatic steatosis. This Cd36-Fgf21-Stat5 regulatory loop may regulate lipid metabolism during fasting and feeding cycles.

In conclusion, deficiency of Tgr5 in mice alleviates fasting-induced hepatic steatosis by increasing hepatic lipolysis and GH-Stat5 signaling. This study uncovered a novel mechanism that bile acid-activated Tgr5 signaling may alter GH signaling and bile acid synthesis to regulate hepatic bile acid and lipid metabolism. Further study of GH-Stat5 signaling in Tgr5^−/− mice is needed to understand the underlying mechanisms for increasing serum GH-Stat5 signaling during fasting, and for GH-Stat5 signaling in modulation of Cyp7b1 expression levels and bile acid metabolism.

Abbreviations

Acc-α

acetyl-CoA carboxylase α

aP2

adipocyte 2 (or fatty acid binding protein 4, Fabp4)

ASM

acid soluble metabolite

Cpt1α

carnitine palmitoyltransferase 1α

CDCA

chenodeoxycholic acid

CA

cholic acid

Cd36

cluster of differentiation 36, (or FAT, fatty acid translocase)

CLAMS

comprehensive laboratory animal monitoring system

Cyp7a1

cholesterol 7α-hydroxylase

Cyp8b1

sterol 12α-hydroxylase

Cyp27a1

sterol 27-hydroxylase

Cyp7b1

oxysterol 7α-hydroxylase

Dio2

Deiodinase 2

DCA

deoxycholic acid

Fatp

fatty acid transport proteins

Fas

fatty acid synthase

FFA

free fatty acid

Fgf15

fibroblast growth factor 15 (human ortholog Fgf19)

Fgf21

fibroblast growth factor 21

Fxr

farnesoid- X- receptor

GH

growth hormone

GLP-1

glucagon like peptide-1

LCA

lithocholic acid

α/β-MCA

α- and β-muricholic acids

Ppar-α

peroxisome proliferator-activated receptor α

Pgc-1α

peroxisome proliferator-activated receptor γ co-activator 1-α

Shp

small heterodimer partner

Stat5

signal transducer and activator of transcription 5

Socs2

suppressors of cytokine signaling 2

T₃

thyroxin

Tgr5

Takeda G protein-coupled receptor 5 (aka G protein coupled bile acid receptor (Gpbar-1)

Ucp

uncoupling protein