Irisin Is Regulated by CAR in Liver and Is a Mediator of Hepatic Glucose and Lipid Metabolism

Abstract

Irisin, a hormone proteolytically processed from fibronectin type III domain-containing protein 5 (FNDC5), has been reported to induce the browning of sc adipocytes by increasing the level of uncoupling protein 1. In this study, we showed that activation of the nuclear receptor constitutive androstane receptor induced FNDC5 mRNA expression in the liver and increased the circulating level of irisin in mice. FNDC5/irisin is a direct transcriptional target of constitutive androstane receptor. Hepatic-released irisin functioned as a paracrine/autocrine factor that inhibited lipogenesis and gluconeogenesis via the Adenosine 5′-monophosphate (AMP)-activated protein kinase pathway. Adenovirus-overexpressed irisin improved hepatic steatosis and insulin resistance in genetic-induced obese mice. Irisin transgenic mice were also protected against high-fat diet-induced obesity and insulin resistance. In conclusion, our results reveal a novel pathway in regulating FNDC5/irisin expression and identify a physiological role for this hepatic hormone in glucose and lipid homeostasis.

Hepatic steatosis is characterized by the accumulation of excessive triglyceride in the liver. If unmanaged, simple steatosis may progress from nonalcoholic fatty liver disease into nonalcoholic steatohepatitis (NASH), in which triglyceride accumulation is associated with inflammation (1). The mechanisms underlying the accumulation of fat in the liver may include excess dietary fat, increased delivery of free fatty acids to the liver, inadequate fatty acid oxidation, and increased de novo lipogenesis (2). Ectopic accumulation of hepatic lipids has been linked to the development of hepatic insulin resistance and type 2 diabetes (3).

Irisin, a novel polypeptide hormone, is proteolytically processed from fibronectin type III domain-containing protein 5 (FNDC5), which is highly expressed in skeletal muscle (4). Irisin has been reported to induce browning of sc adipocytes and thermogenesis by increasing uncoupling protein 1 (Ucp1) level (4). Initial reports demonstrated irisin was induced by endurance exercise through peroxisome proliferator-activated receptor-γ coactivator 1-α (4). However, subsequent studies had conflicting results and showed the induction of irisin by exercise was only observed in highly selected patients (5–8). These results suggest an alternative pathway might be involved in the regulation of FNDC5/irisin expression.

Recent studies showed FNDC5 was also expressed in other tissues such as heart, adipose tissue, and liver, which indicates additional functions of this hormone (6, 9). It has been reported that irisin may have a role in hepatic lipid metabolism (10, 11). Irisin level was decreased in patients with nonalcoholic fatty liver disease and NASH and was inversely associated with intrahepatic triglyceride content in obese adults (10, 11). Serum irisin was also decreased in patients with type 2 diabetes and inversely associated with newly diagnosed type 2 diabetes, suggesting that it may play a crucial role in glucose metabolism as well (12, 13).

The constitutive androstane receptor (CAR) is a xenobiotic sensor that controls the expression of genes involved in the metabolism of exogenous molecules and endogenous compounds such as bile acids, bilirubin, and thyroid hormones (14–16). Recent studies have also linked CAR to lipid metabolism and glucose metabolism (17). CAR can be activated by phenobarbital or 1,4-Bis[2-(3,5-dichloropyridyloxy)] benzene (TCPOBOP) (18). Activation of CAR inhibits hepatic steatosis by repression of lipogenesis and increases fatty acid β-oxidation (19, 20). CAR activation also prevents diet- and genetic-induced obesity and insulin resistance by reducing the expression of glucose-6-phosphatase (G6p) and phosphoenolpyruvate carboxykinase (Pepck), 2 key enzymes in gluconeogenesis (19, 20). Despite the known function of CAR in energy metabolism, how CAR activation may affect glucose and lipid metabolism remains unclear.

In this study, we investigated whether CAR could regulate FNDC5/irisin expression in the liver. Adenovirus overexpressed irisin and irisin transgenic mice were used to study the metabolic function of irisin in diet and genetic induced obesity. Our result demonstrate a novel pathway in regulation of FNDC5/irisin expression by CAR in the mouse liver. Furthermore, this work reveals an essential role of irisin in hepatic glucose and lipid metabolism and provides a potential therapeutic target for the treatment of metabolic disease.

Materials and Methods

Animals, drug treatment, cell culture, and histological evaluation

Animals were housed at West China Hospital, Sichuan University, in accordance with the guidelines of the animal care and utilization committee of the institute. Male C57BL/6J mice and ob/ob mice were purchased from The Jackson Laboratory. CAR−/− mice were maintained on a C57BL/6J background. When necessary, C57BL/6J mice received a single ip injection of TCPOBOP (0.5 mg/kg) and were killed 3 days after the injection. To activate pregnane X receptor (PXR), mice received 2 daily ip injections of pregnenolone 16α-carbonitrile (40 mg/kg) for 2 days and were killed 8 hours after the last dose. For histological analysis, tissues were fixed in 4% formaldehyde, embedded in paraffin, sectioned at 5 μm, and stained with hemotoxylin and eosin. Frozen sections (10 μm) were used for Oil-red O staining.

EMSA and transient transfection, luciferase reporter gene assay, and chromatin immunoprecipitation (ChIP) assay

EMSA was performed using ³²P-labeled oligonucleotides and receptor proteins prepared by the TNT method (21). The FNDC5 gene promoter (2 kb) was cloned into the pGL3. HepG2 cells were transfected with the reporter construct and CAR expression vector in 48-well plates as previously described (21). When necessary, cells were treated with TCPOBOP (250 nmol/L) for 24 hours before luciferase assay. The transfection efficiency was normalized against the β -galactosidase activities from a cotransfected CMX β-galactosidase. For in vivo luciferase assays, 40 μg of FNDC5-luc, mutant FNDC5-luc, and Cyp2b10-luc plasmids were administered to 6-week-old mice by tail vein in 7 seconds. The mice were administrated with vehicle or TCPOBOP 24 hours after the hydrodynamic injection. Twenty-four hours later, the mice were killed, and liver tissues were homogenized for luciferase assay (22). ChIP assay was performed essentially as described previously (23).

Adenovirus irisin (Ad-irisin) experiments

Eight-week-old male ob/ob mice received tail-vein injections of 1.2 × 10¹¹ adenovirus particles of Ad-GFP or Ad-irisin as described (24). At 2 weeks after injection, mice were killed. Tissues were collected and freshly frozen in liquid nitrogen or fixed in formalin solution.

Serum chemistry and hepatic lipid profile analysis

Blood samples were collected from overnight-fasted mice. Serum levels of total triglycerides and cholesterol (Zhongsheng), insulin (Crystal Chem), and irisin (Phoenix Biotech) were measured by using commercial assay kits according to the manufacturer's instructions. To measure hepatic lipid content, tissues were homogenized in 2-mL buffer containing 18mM Tris (pH 7.5), 300mM mannitol, 50mM EGTA, and 0.1mM phenylmethysulfonyl fluoride. An amount of 400-μL homogenate was mixed with 4-mL chloroform/methanol (2:1) and incubated overnight at room temperature with occasional shaking. An amount of 800-μL H₂O was then added, vortexed, and centrifuged for 5 minutes at 3000g, and the lower lipid phase was collected and dried under nitrogen gas. The lipid pellets were then dissolved in 1% Triton X-100 in absolute ethanol, and the triglyceride and cholesterol levels were measured by using commercial assay kits (25).

Glucose production assay

Hepatocytes isolated from mice fasted overnight were plated onto collagen-coated 24-well plates at 2 × 10⁵ cells/well. After cells had attached, the medium was changed to glucose and phenol red-free DMEM with the addition of 10mM lactate, 1mM pyruvate, and 5mM glutamine. Hepatocytes were pretreated with irisin for 30 minutes and then incubated with 10μM forskolin for 4 hours. Medium glucose was assayed by use of a hexokinase-based glucose assay (Sigma-Aldrich).

Glucose tolerance test (GTT) and insulin tolerance test (ITT)

For GTT, mice were fasted for 16 or 4 hours before receiving an ip injection of D-glucose at 1- or 2-g/kg body weight. Blood samples were taken at different times, and glucose was measured. For ITT, mice were fasted for 4 hours before receiving an ip injection of human insulin (Novolin) from Novo Nordisk at 0.5- or 1.5-U/kg body weight. Blood samples were taken, and blood glucose were measured.

RT-PCR analysis

For real-time PCR analysis, reverse transcription involved use of random hexamer primers and Superscript RT III enzyme from Invitrogen. SYBR Green-based real-time PCR was performed with the ABI 7300 real-time PCR system (21). Data were normalized against the control of cyclophilin.

Statistical analysis

Experiments were repeated at least 3 times, and data were expressed as mean ± SE. Statistical significance was determined by Student's unpaired 2-tailed t test or one-way ANOVA for multiple comparisons with Tukey's test. P < .05 was considered statistically significant.

Results

CAR regulates FNDC5 expression in the liver

In preliminary experiments, transcriptional profiling of liver in mice treated with CAR agonist suggested that FNDC5 was regulated by CAR. To confirm this result, 8-week-old male C57BL/6J and CAR−/− mice were treated with a single dose of CAR ligand TCPOBOP or vehicle for 3 days. As expected, TCPOBOP increased the mRNA expression of Cyp2b10, a known CAR target gene in wild-type mice (Figure 1A). FNDC5 mRNA was also robustly induced in the liver of TCPOBOP-treated wild-type mice, and more importantly, this induction was abolished in CAR−/− mice (Figure 1B). With the induction of FNDC5 expression in CAR-activated mice, the serum level of irisin was increased 2.5-fold in wild-type mice (Figure 1C). Mouse primary hepatocytes treated with TCPOBOP also showed induced FNDC5 expression (Figure 1D). FNDC5 gene expression was not affected by CAR activation in skeletal muscle and adipose tissue, suggesting that elevated serum level of irisin was not attributable to its expression in these tissues, which are known to express FNDC5 (Supplemental Figure 1, A and B) (4, 26). Pregnenolone X receptor (PXR) is a sister receptor of CAR and shares many of its target genes with CAR (17). However, FNDC5 expression was not regulated by PXR activation (Supplemental Figure 1C). Previous study reported that irisin increased Ucp1 expression in sc adipose tissue (4). Immunochemistry revealed a clear increase in the number of Ucp1-positive multilocular adipocytes in TCPOBOP-treated sc adipose tissue (Figure 1, E and F). An mild increase in peroxisome proliferator-activated receptor-γ coactivator 1-α mRNA expression was also observed, but there were no changes in expression of cell death-inducing DFFA-like effector A and PR domain-containing protein 16 (Supplemental Figure 1D).

Figure 1.

Activation of CAR regulated FNDC5/irisin expression in liver. Wild-type (WT) and CAR−/− mice were ip injected with vehicle (Veh) or CAR agonist TCPOBOP (TC), 05 mg/kg. Mice were killed 3 days after the injection, blood and liver tissue were collected. Real-time PCR analysis of hepatic mRNA expression of Cyp2b10 (A) and FNDC5 (B). C, EIA analysis of serum irisin level. D, Primary hepatocytes were isolated from C57BL/6J mice and were treated with TC (0.25 μmol/L) overnight. Real-time PCR analysis of FNDC5 mRNA expression. Photograph (E) and immunostaining of Ucp1 (F) in sc adipose tissue from Veh and TC-treated mice. Data are mean ± SEM of n = 5–7; *, P < .05.

FNDC5 is a direct transcriptional target of CAR

Having demonstrated that CAR is necessary and sufficient for FNDC5 regulation in liver, we determined whether FNDC5 is a direct transcriptional target of CAR. Inspection of the FNDC5 promoter revealed a DR5-type (direct repeat spaced by 5 nucleotides) of nuclear receptor-response element (TGACCTctatcTGACCC) (Figure 2A). EMSA revealed that CAR/RXR heterodimers could bind to this DR5 site (Figure 2A). The binding of FNDC5/DR5 to CAR was specific, as shown by the efficient competition of binding by excess unlabeled FNDC5/DR5 or supershift by CAR antibody (Figure 2A).

Figure 2.

FNDC5/irisin is a transcriptional target gene of CAR. A, EMSA of CAR/RXRα heterodimers binding to DR-5. The unlabeled DR-5 probe showed efficient competition. Supershift assay was used to evaluate the specific binding of CAR/RXRα to DR-5. Antibody specific to CAR (lane 5) was added to CAR/RXRα heterodimers for 20 minutes at room temperature, followed by the shift reaction. B, Luciferase assay of transient transfection in HepG2 cells with the FNDC5 natural promoter reporter, Cyp2b10 was included as a positive control. C and D, ChIP assay of recruitment of CAR to the FNDC5 promoter. C, Primary mouse hepatocytes were infected with Ad-GFP and Ad-CAR-FLAG adenovirus with or without TCPOBOP (TC) treatment. ChIP involved an anti-FLAG antibody with IgG as a control. D, The recruitment of CAR to FNDC5 promoter was also confirmed in vivo in the liver of TC-treated mice. Data are mean ± SEM of n = 6; *, P < .05.

Luciferase reporter assays were used to determine whether CAR could regulate FNDC5 promoter activity with a reporter gene that contains the natural promoter of FNDC5 (−2116 to +53 bp). In transient transfection assays of HepG2 cells, reporter activity was induced about 2-fold by CAR in the presence of TCPOBOP and mutation of DR5-binding site resulted in the loss of CAR effect on reporter activity (Figure 2B). The luciferase reporters were further introduced to mouse liver in vivo by hydrodynamic injection through the tail vein. Upon activation by TCPOBOP, CAR was able to produce a 10-fold increase in FNDC5 promoter activity. Mutation of DR5 in FNDC5 promoter abolished CAR effect (Supplemental Figure 1E). Both in vitro and in vivo ChIP assay was used to demonstrate the recruitment of CAR onto the FNDC5 promoter. Mouse primary hepatocytes were infected with Ad-CAR-FLAG for 12 hours, then treated with TCPOBOP for 16 hours. ChIP assay was performed with the use of an anti-FLAG antibody or IgG as a control. As shown in Figure 2C, treatment with TCPOBOP increased the recruitment of CAR onto the FNDC5 promoter (Figure 2C). In the ChIP assay, Cyp2b10, a well-known CAR target gene was included as a positive control. The recruitment of CAR to FNDC5 promoter was also confirmed in vivo in the liver of TCPOBOP-treated mice (Figure 2D).

Irisin inhibited gluconeogenesis and lipogenesis in primary human hepatocytes by activating the Adenosine 5′-monophosphate (AMP)-activated protein kinase (AMPK) pathway

To understand the paracrine/autocrine function of hepatic-released irisin, primary human hepatocytes were treated with recombinant irisin, then gluconeogenesis/lipogenesis-related gene levels were measured. We found that irisin inhibited basal and cAMP-stimulated gluconeogenic gene expression (Figure 3A). Pepck and G6p are 2 rate-limiting enzymes involved in gluconeogenesis. The expression of both genes was significantly decreased by irisin under basal and cAMP-stimulated conditions (Figure 3A). Also, irisin modestly but significantly reduced forskolin-stimulated glucose production in primary human hepatocytes (Figure 3B). Similar to gluconeogenesis, lipogenic gene expression was repressed by irisin. Irisin inhibited the basal and liver X receptor-α-induced expression of lipogenic genes such as sterol regulatory element-binding protein (Srebp)-1c, fatty-acid synthase (Fas), and stearoyl-CoA desaturase-1 (Scd-1) (Figure 3C).

Figure 3.

Irisin suppressed gluconeogenesis and lipogenesis by activating AMPK pathway. A, Primary human hepatocytes were pretreated with 50nM irisin for 15 minutes, then incubated with vehicle (Veh) or forskolin (Forsk) for 4 hours. Real-time PCR analysis of G6p and Pepck mRNA expression. B, Glucose production from hepatocytes with or without irisin treatment. C, Cultured human hepatocytes were pretreated with 50nM irisin for 15 minutes, then incubated with Veh or liver X receptor (LXR) agonist, 10μM GW3965 (GW) overnight. Real-time PCR analysis of lipogenic gene expression. D and E, Western blot analysis of AMPK, Erk, Akt, and JNK phosphorylation in primary hepatocytes treated with irisin at the indicated time or concentration. Acc, acetyl-CoA-carboxylase 1. Data are mean ± SEM of n = 6; *, P < .05.

AMPK is a key regulator in glucose and lipid metabolism. Irisin-suppressed gluconeogenesis and lipogenesis prompted us to determine whether AMPK was activated by this hepatic-released factor. Indeed, we found that irisin dose dependently activated AMPK, with maximum effect occurring at 60nM–80nM (Figure 3D). Activation of AMPK was observed as soon as 15 minutes after irisin treatment (Figure 3E), and this effect remained detectable for 4 hours. Other signaling pathways, such as Akt, Erk, and c-Jun N-terminal kinase, were not changed (Figure 3E). These results suggested the inhibitory effect of irisin on gluconeogenesis and lipogenesis may be partially mediated by AMPK activation.

Ad-irisin treatment improved hepatic steatosis and insulin resistance in ob/ob mice

The metabolic function of irisin on hepatocytes led us to determine whether it confers beneficial effects in diabetic mice. A human FNDC5 cDNA was cloned into an empty shuttle adenoviral vector under the control of the cytomegalovirus promoter. Intravenous adenovirus injection has been used for efficient transduction of adenovirus into the vast majority of liver hepatocytes (24). Eight-week-old ob/ob mice were injected with Ad-irisin or Ad-GFP through the tail vein. At 2 weeks after injection, Ad-irisin-treated mice showed a significant increase in serum irisin levels as compared with Ad-GFP injection (18.8 ± 3.12 vs 12.8 ± 2.18 nmol/L, P < .05). Ad-irisin-and Ad-GFP-treated mice showed similar body weight. However, the ratio of sc and brown adipose tissue weight to body weight was lower in Ad-irisin- than Ad-GFP-treated mice (Figure 4A).

Figure 4.

Adenovirus-overexpressed irisin improved hepatic steatosis and insulin resistance in ob/ob mice. Ob/ob mice received tail vein injections of 1.2 × 10¹¹ adenovirus particles of Ad-GFP or Ad-irisin. At 2 weeks after injection, mice were killed. A, The ratio of fat weight to body weight in 3 different fat tissues. B, Oil-red O (top) and H&E staining (bottom) of liver sections. C, Hepatic cholesterol and triglyceride content. D, Real-time PCR analysis of hepatic expression of lipogenic genes. E, GTT in ob/ob mice after Ad-GFP or Ad-irisin injection. For GTT, mice fasted for 16 hours before receiving an ip injection of D-glucose at 1-g/kg body weight. F, Real-time PCR analysis of hepatic mRNA expression of gluconeogenic genes. Data are mean ± SEM of n = 5–7; *, P < .05.

Obesity is known to be associated with hepatic steatosis. Oil-red O staining showed decreased lipid staining in the liver of Ad-irisin-treated mice and histology showed reduced lipid droplets in livers of Ad-irisin-treated ob/ob mice (Figure 4B). The inhibition of steatosis was also confirmed by decreased hepatic levels of triglycerides but not cholesterol (Figure 4C). Consistent with suppressed lipid accumulation in liver, adenovirus overexpressed irisin significantly inhibited lipogenic gene expression such as Srebp-1c, Fas, acetyl-CoA-carboxylase, and Scd-1 (Figure 4D). Interestingly, the serum level of triglycerides was not significantly affected by irisin (Supplemental Figure 2).

The suppressed hepatic steatosis by irisin was associated with improved insulin sensitivity. Ad-irisin-treated ob/ob mice showed significantly improved glucose tolerance (Figure 4E). The mRNA expression of the hepatic gluconeogenic genes G6p and Pepck was lower in Ad-irisin- than Ad-GFP-treated mice (Figure 4F). The blood insulin level was also significantly reduced with irisin treatment (Supplemental Figure 3). Consistent with results from cultured human hepatocytes, irisin induced AMPK phosphorylation in ob/ob mouse liver (Supplemental Figure 4).

Irisin transgenic mice are protected against high-fat diet (HFD)-induced obesity and insulin resistance

To further understand the function of irisin in HFD-induced obesity and insulin resistance, irisin transgenic mice were created under the control of the chicken β-actin promoter and cytomegalovirus enhancer. FNDC5 mRNA level in liver was 35-fold higher in irisin transgenic than wild-type mice (Supplemental Figure 5A). Also, Ucp1-positive cells were higher in irisin transgenic mice in sc fat tissue (Supplemental Figure 5B). Under chow diet, wild-type and irisin transgenic mice did not differ in their body weights (Supplemental Figure 5B). However, after 12 weeks of HFD feeding, transgenic mice were smaller and had 19% lower body weight than their wild-type counterparts (Figure 5A). Necropsy revealed that irisin transgenic mice had a significantly decreased mass of visceral and sc adipose tissue than wild-type mice (Figure 5, B and C). H&E staining showed decreased adipocyte size in visceral, sc, and brown fat in transgenic mice (Figure 5D). The decreased adipocyte size in visceral fat of transgenic mice was confirmed by cell size quantification (Figure 5E).

Figure 5.

Irisin transgenic mice are protected against HFD-induced obesity. Wild-type (WT) and irisin transgenic mice were fed with HFD for 12 weeks. A, Growth curve of WT and irisin transgenic mice over 12 weeks HFD. B, The ratio of fat weight to body weight in epididymal and sc fat tissue. C, Representative appearance of epididymal fat tissues. D and E, Histological analysis of 3 fat tissues (D) and adipocyte size in epididymal fat (E). Top is WT, bottom is irisin. Data are mean ± SEM of n = 6–7; *, P < .05.

We also examined hepatic steatosis in wild-type and irisin transgenic mice. Oil-red O staining revealed decreased hepatic steatosis in transgenic mice (Figure 6A). The inhibition of steatosis in irisin transgenic mice was confirmed by the decreased hepatic levels of triglycerides but not cholesterol (Figure 6B). Consistently, transgenic mice also showed decreased lipogenic gene expression of Srebp-1c, Fas, and Scd-1 (Figure 6C). When the insulin sensitivity was examined, we found that transgenic mice showed improved performance on the GTT (Figure 6D) and ITT (Figure 6E) as compared with wild-type mice. Similar to results from Ad-irisin-treated ob/ob mice, serum triglyceride level was not changed in transgenic mice (Supplemental Figure 6).

Figure 6.

Irisin transgenic mice are protected against HFD-induced hepatic steatosis and insulin resistance. Wild-type (WT) and irisin transgenic mice were fed a HFD for 12 weeks. A, Oil-red O (top) and H&E staining (bottom) of liver sections. B, Hepatic cholesterol and triglyceride content. C, Real-time PCR analysis of hepatic expression of lipogenic genes. GTT (D) and ITT (E) in WT and irisin transgenic mice fed with HFD for 12 weeks. For GTT, mice fasted for 16 hours before receiving an ip injection of D-glucose at 2-g/kg body weight. For ITT, mice fasted for 4 hours before receiving an ip injection of insulin at 0.5-U/kg body weight. F, Real-time PCR analysis of hepatic mRNA expression of gluconeogenic genes. Data are mean ± SEM of n = 5–7; *, P < .05.

Discussion

In this study, we reported that activation of CAR induced FNDC5 mRNA level in the liver and increased circulating irisin level in mice. We also showed FNDC5 is a direct transcriptional target of CAR. Hepatic-released irisin functions as a paracrine/autocrine factor by inhibiting lipogenesis and gluconeogenesis via the AMPK pathway. Adenovirus-overexpressed irisin improved hepatic steatosis and insulin resistance in ob/ob mice.

Irisin was initially reported to be regulated by endurance exercise (4). Irisin level was elevated in mice after a 3-week free-wheel running period and in healthy human adults after 10-week endurance training (4). However, in contrast to the initial report, Timmons et al (5) reported a lack of change in muscle FNDC5 mRNA expression in either endurance or resistance-trained subjects. Similarly, a study of young healthy males did not find any changes in circulating irisin level after an 8-week exercise program (6). In line with other reports describing negative data, a German randomized controlled trial did not detect any changes in serum irisin level after 6 months of aerobic or strength exercise training (7). These conflicting results suggested an alternative pathway in the regulation of FNDC5/irisin expression. Indeed, we found that the basal level of FNDC5 in mouse liver was low but highly inducible by CAR. CAR activation increased both hepatic FNDC5 mRNA and circulating irisin level in mice.

We found FNDC5/irisin was regulated by CAR in the mouse liver, but its expression was not changed in skeletal muscle or adipose tissue (Supplemental Figure 1, A and B). This finding is consistent with a previous report that CAR is predominantly expressed in liver and intestine (17). CAR, originally identified as a xenobiotic receptor, has been reported to suppress lipogenesis and gluconeogenesis and protect against hepatic steatosis (19, 20). However, the target genes involved in the beneficial effect of CAR have not been identified. The antiobesity and antidiabetic effect of irisin suggests that it might participate the beneficial effect of CAR. Future studies are necessary to determine whether irisin is required for the metabolic benefit of CAR.

PXR is a sister receptor of CAR that shares many functions in xenobiotic regulation. However, CAR and PXR appear to have opposite effects on metabolic disease. We and others have reported that activation of PXR promoted hepatic steatosis and insulin resistance (21, 27–29). PXR transgenic mice showed hyperlipidemia, fatty liver and impaired glucose and insulin tolerance (21, 27–29). The antidiabetic effect of CAR is opposite to the prodiabetic effect of PXR activation. This observation suggests that CAR and PXR have overlapping but distinct spectra of target genes.

Our finding that irisin inhibited hepatic steatosis was interesting in light of recent reports indicating the association of irisin and hepatic lipid metabolism. Polyzos et al (11) found irisin levels were significantly lower in subjects with nonalcoholic fatty liver and NASH than lean controls. Zhang et al (10) showed that irisin level was inversely correlated with hepatic triglyceride level. In this study, irisin level gradually decreased with increasing intrahepatic triglyceride content (10). Although irisin has been linked to fatty liver disease, the underlying mechanism is unclear. We showed that irisin inhibited hepatic lipogenic gene expression and suppressed triglyceride accumulation. Our results provide direct evidence that irisin is involved in hepatic lipid metabolism. This is also consistent with recent report that irisin inhibited the fatty acid-induced lipogenic markers acetyl-CoA-carboxylase and Fas, and prevented the fatty acid -induced lipid accumulation in vitro (30). The suppressed gluconeogenesis also suggested irisin might have an important role in diabetes. Indeed, overexpression of irisin protected mice against both diet- and genetic-induced obesity. This finding is also supported by studies showing a down-regulation of irisin was observed under conditions of impaired glucose control or diabetes mellitus (12, 13).

Although several metabolic effect has been proposed, the signaling pathway of irisin remains elusive. Zhang et al (31) demonstrated that irisin induced phosphorylation of the p38 and Erk signaling pathways. Inhibition of p38 MAPK or Erk abolished the up-regulatory effect of irisin on Ucp1 expression (31). This signaling pathway was also shown in human umbilical vein endothelial cells (32). However, we did not find the Erk pathway nor the Akt and JNK pathways changed by irisin treatment in hepatocytes. Instead, we found that irisin dose and time dependently activated AMPK phosphorylation. AMPK is an evolutionarily conserved protein kinase that serves as a primary cellular monitor of energy charge. Activation of AMPK leads to inhibition of hepatic lipogenesis and gluconeogenesis, and reduced obesity induces insulin resistance (33). Our suggested the metabolic functions of irisin are likely mediated by AMPK. While this manuscript was in review, a similar result was found that AMPK was activated by irisin in skeletal muscle cells (34). These results uncover a downstream signaling pathway of irisin that likely mediates its effects.

In conclusion, we established a novel pathway in regulating FNDC5/irisin expression. Hepatic-released irisin functions as a paracrine/autocrine hormone that inhibits lipogenesis and improves hepatic steatosis, and protects mice against genetic- and diet-induced obesity and insulin resistance. The induction of FNDC5/irisin by CAR also provides a plausible mechanism for this nuclear receptor-mediated beneficial metabolic function. Induction of FNDC5/irisin expression in liver might be a potential therapeutic strategy for the treatment of metabolic disease.

Acknowledgments

Author contributions: L.M., J.S., and J.H. designed and performed experiments and wrote the manuscript; Y.K., Q.L., S.C.; and Y.Z. helped with experiments; M.Z., W.J., C.J., and A.Q. contributed to the discussion and review of the manuscript; J.H. obtained funding, designed experiments, and wrote the manuscript; and L.M., J.S., and J.H. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

This work was supported by National Natural Science Foundation of China Grants 81500045, 81270926, and 81471068, the Distinguished Young Scientists of Sichuan Province Grant 2014JQ0034, and the Sichuan University Young Scientist Fellowship 2013SCU04A17.

Disclosure Summary: The authors have nothing to disclose.

Abbreviations

 
• Ad-irisin

  adenovirus irisin

 
• AMPK

  Adenosine 5′-monophosphate (AMP)-activated protein kinase

 
• CAR

  constitutive androstane receptor

 
• ChIP

  chromatin immunoprecipitation

 
• DR5

  direct repeat spaced by 5 nucleotides

 
• FNDC5

  fibronectin type III domain-containing protein 5

 
• G6p

  glucose-6-phosphatase

 
• GTT

  glucose tolerance test

 
• HFD

  high-fat diet

 
• ITT

  insulin tolerance test

 
• NASH

  nonalcoholic steatohepatitis

 
• Pepck

  phosphoenolpyruvate carboxykinase

 
• PXR

  pregnane X receptor

 
• Scd-1

  stearoyl-CoA desaturase-1

 
• Srebp

  sterol regulatory element-binding protein

 
• TCPOBOP

  1,4-Bis[2-(3,5-dichloropyridyloxy)] benzene

 
• Ucp1

  uncoupling protein 1.

Author notes

^*

L.M. and J.S. contributed equally to this work.