Glucagon Receptor Signaling Regulates Energy Metabolism via Hepatic Farnesoid X Receptor and Fibroblast Growth Factor 21

Glucagon is secreted from pancreatic α-cells in response to hypoglycemia and is the primary counterregulatory hormone to insulin action, increasing glycogenolysis and gluconeogenesis while simultaneously inhibiting glycogen synthesis (1). These actions, although beneficial in the context of hypoglycemia, may contribute to pathophysiological hyperglycemia in the setting of diabetes (2). Glucagon receptor (GCGR) agonism also modulates bile acid (BA) metabolism, stimulates fatty acid utilization, and reduces dyslipidemia, characteristics clearly desirable in antiobesity therapeutics (1). It is now accepted that GCGR agonism, when coupled with glucagon-like peptide 1 (GLP-1) agonism, offers potential opportunities for the therapy of metabolic syndrome (3,4).

We have reported that fibroblast growth factor 21 (FGF21), secreted in response to GCGR agonism, mediates many glucagon actions, including the prevention of diet-induced obesity (DIO) (5). Like glucagon, FGF21 regulates cholesterol (CHL) and BA metabolism (6,7). Similarly, the BA nuclear receptor farnesoid X receptor (FXR) is a regulator of energy metabolism, mitochondrial function, and FGF21 gene expression (8). In this study, we investigated the roles of hepatic GCGR, FGF21, and FXR in the antiobesity effects of the GCGR agonist IUB288.

All studies were approved by and performed according to the guidelines of the Institutional Animal Care and Use Committee of the University of Alabama at Birmingham or the University of Cincinnati. Mice were single or group housed on a 12:12-h light/dark cycle at 22°C and constant humidity with free access to food and water, except as noted. Gcgr- and Fxr-floxed mice were obtained from the original investigators (9,10), whereas Fgf21-floxed and Albumin-Cre mice were obtained from The Jackson Laboratory (Bar Harbor, ME). All models were validated for tissue-specific, target gene knockout (Supplementary Fig. 1A–E). All mice maintained in our facilities are on a C57Bl/6J background. Mice were fed standard chow (5.8% fat; Teklad LM-485; Harlan Teklad) for colony maintenance and high-fat diet (HFD) (58.0 kcal% fat; D12331; Research Diets, New Brunswick, NJ) for DIO studies. For sacrifice, isoflurane anesthesia was used, torso blood was collected, and plasma was collected by centrifugation of whole blood at 3,000g for 10 min.

IUB288 was synthesized as previously described (5) and native glucagon obtained from American Peptide Company.

Body weight (BW) and food intake measurements were collected twice a week. Body composition was measured using MRS (EchoMRI; Echo Medical Systems). Combined indirect calorimetry was conducted as previously described (Comprehensive Lab Animal Monitoring System; Columbus Instruments) (11).

Intraperitoneal glucose (1.5 g/kg, 20% weight for volume d-glucose in 0.9% weight for volume saline; Sigma-Aldrich, St. Louis, MO) tolerance tests were conducted in 5-h–fasted mice as previously published (12). Tail vein blood glucose was assessed using a glucometer (Therasense FreeStyle glucometer; Abbott Laboratories, Abbott Park, IL).

Lipids in plasma and tissue samples from 2-h–fasted mice were determined using Infinity Triglycerides (#TR22421; Thermo Fisher Scientific), Infinity Cholesterol (#TR13421; Thermo Fisher Scientific), Total Bile Acids Assay Kit (#80259; Crystal Chem), and β-Hydroxybutyrate (Ketone Body) Colorimetric Assay Kit (#700190; Cayman Chemical).

Plasma aliquots (50 μL) were extracted to recover BAs. Diluted extracts (1.25 μL plasma equivalent) were resolved by reverse-phase gradient liquid chromatography and analyzed by negative electrospray ionization mass spectrometry using multiple-reaction monitoring. BA peak areas were analyzed by MultiQuant 3.0.1 (SCIEX) and compared with peak-area concentration standard curves of individual BAs. Plasma hormones from 2-h–fasted mice were determined by Bio-Plex Pro Mouse Diabetes 8-Plex Assay (Bio-Rad Laboratories) in plasma samples collected in the presence of protease and phosphatase inhibitors (Halt; Thermo Fisher Scientific).

Liver RNA was isolated from 2-h–fasted mice using the RNeasy Lipid Mini-Kit (Qiagen, Valencia, CA), and cDNA was synthesized by RT-PCR using SuperScriptIII, DNase treatment, and anti-RNase treatment according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA). Single-gene quantitative PCR was performed as previously described (11). Data were normalized to housekeeping genes Hprt, Rps18, or Ppia using the ΔΔ threshold cycle calculation. See Supplementary Table 1 for a list of primer sets. High-throughput RNA sequencing was performed in the Heflin Genomics Core at the University of Alabama at Birmingham (UAB). Gene network associations and differentially expressed genes were identified via unpaired two-tailed and Bonferroni-adjusted P values (Q value) <0.05, respectively. Sequencing data have been deposited within the Gene Expression Omnibus repository (https://www.ncbi.nlm.nih.gov/geo). Gene set enrichment analysis, functional and network analyses, and candidate upstream regulators were identified via Qiagen’s Ingenuity Pathway Analysis, where fold change >1.5, P < 0.05, and fragments per kilobase of transcript per million mapped reads >2.

Primary hepatocytes were prepared from anesthetized mice as previously described (13). Perfusion (Krebs Ringer with glucose and 0.1 mmol/L EGTA) followed by digestion buffer (Krebs Ringer with glucose, 1.4 mmol/L CaCl₂, and 50 µg/mL liberase [05401119001; Roche]) was infused into the vena cava via peristaltic pump. Viable hepatocytes were recovered by Percoll gradient centrifugation (350g for 5 min) followed by washing (50g for 3 min, three times) and seeded on rat tail type 1 collagen-coated plates in DMEM (10% FBS and 1% penicillin/streptomycin) with all experiments conducted <24 h postisolation.

All data are represented as mean ± SEM. Statistical significance was determined using unpaired Student t tests or, where appropriate, one- and two-way ANOVA with multiple-comparison Tukey and Sidak posttest, respectively. Statistics were completed using Prism version 7.0 for Macintosh and Windows (GraphPad Software, San Diego, CA). Statistical significance was assigned when P was <0.05.

We have previously reported that GCGR agonism reduces body and fat mass in DIO mice (5). Considering the high level of GcgR expression in liver tissue, we reasoned that the antiobesity signal may be hepatic in origin and tested this hypothesis using mice deficient for hepatic Gcgr (Gcgr^∆liver) (9). Six- to 8-week-old male Gcgr^∆liver mice and their littermate controls were placed on an HFD for 10 weeks to induce obesity. High-fat (HF) feeding stimulated similar food intake and accumulation of BW in both genotypes (Fig. 1A and B). Gcgr^∆liver mice exhibited slightly less fat mass and a trend for more lean mass (Fig. 1C), with profoundly enhanced glucose tolerance as compared with their HF-fed littermate controls (Fig. 1D). Following HF feeding, mice were matched for BW and fat mass within each genotype and treated for 17 days with vehicle (saline) or IUB288 (10 nmol/kg/day). Gcgr^∆liver mice were protected from hyperglycemia following GCGR agonism (Fig. 2A). Chronic GCGR agonism significantly reduced BW (Fig. 2B and Supplementary Fig. 1F) in wild-type (WT) mice, an effect mainly driven by fat mass loss with a modest decrease in lean mass (Fig. 2C). In contrast, BW, fat, and lean mass were preserved in IUB288-treated Gcgr^∆liver mice (Fig. 2B and C). IUB288 treatment reduced food intake in both genotypes, yet food consumption was not different between Gcgr^∆liver mice and their littermate controls (Fig. 2D).

We next examined the impact of GCGR agonism on circulating lipids. Chronic GCGR agonism significantly reduced circulating CHL in WT but not in Gcgr^∆liver mice, with no effect on circulating triglycerides (TGs) (Fig. 2E). Conversely, chronic GCGR agonism significantly reduced hepatic TG levels in WT but not Gcgr^∆liver mice, whereas liver CHL was unaffected by either genotype or treatment (Fig. 2F). Altogether, these data demonstrate the regulatory role of hepatic GCGR in whole-body energy balance, glucose, and lipid metabolism.

We and others have reported that glucagon stimulates FGF21 secretion in hepatocytes (5,14). To address the role of FGF21, obesity was induced via 16 weeks of HF feeding in 20-week-old male, liver-specific, Fgf21-deficient (Fgf21^∆liver) and WT mice. Mice from each genotype were matched for BW and fat mass and treated for 16 days with vehicle (saline) or IUB288 (10 nmol/kg/day). Chronic GCGR agonism reduced BW (Fig. 3A and Supplementary Fig. 1G), food intake, and fat and lean mass in WT mice (Fig. 3B and C). However, BW reduction in Fgf21^∆liver mice was significantly blunted, and GCGR-stimulated effects on body composition and food intake did not reach statistical significance (Fig. 3). Consistent with our prior findings (5), these data suggest that FGF21-dependent and -independent mechanisms mediate body and fat mass loss following GCGR agonism.

Glucagon regulates BA metabolism (1), and BAs are known metabolic modulators (15). We sought to determine the contribution of BA metabolism in the effect of glucagon on BW. Circulating BAs are suppressed in DIO mice (P < 0.01), yet rescued following chronic GCGR agonism (Fig. 4A), regulation that is absent in Gcgr^∆liver mice (Supplementary Fig. 1H). IUB288 likewise reduced mRNA expression of BA regulators Slc10a1, Cyp27a1, Hmgcr, and Cyp7a1 (Fig. 4B) (16) and elevated total and cholic BAs while decreasing taurodeoxycholic acids in DIO mice (Fig. 4C).

BAs are ligands of the FXR (FXR/Nr1h4) (15), leading us to investigate FXR signaling in glucagon action. Because both GCGR signaling (5) and FXR (8) are known to regulate Fgf21 expression, we assessed hepatic Fgf21 mRNA expression in response to GCGR agonist in WT and FXR-deficient (Fxr^−/−) mice. Intriguingly, hepatic gene expression and circulating levels of FGF21 were similarly upregulated in Fxr^−/− and in WT control mice (Supplementary Fig. 2A and B) (two-way ANOVA, main effect of treatment, P < 0.01), suggesting GCGR agonism independently stimulates Fxr and Fgf21 expression.

Six- to 8-week-old male WT and Fxr^−/− mice were treated for 25 days with IUB288 concomitant with HF feeding to assess the role of FXR in GCGR-mediated prevention of HFD-induced metabolic defects. GCGR activation prevented HFD-induced BW and fat mass gain in WT but not in FXR-deficient mice (Supplementary Fig. 2C–F), whereas lean mass and food intake remained unaffected in this study (Supplementary Fig. 2G and H). These results indicate that FXR action is a necessary mediator of the GCGR signaling on BW.

Because our findings demonstrate that chronic glucagon action reduces BW via the liver, we generated liver-specific Fxr-knockout mice (Fxr^∆liver) to test the organ-specific contribution of FXR signaling. Six- to 8-week-old WT and Fxr^∆liver mice exhibited similar BW and body composition while fed with standard chow (Supplementary Fig. 3A). However, Fxr^∆liver mice were DIO resistant compared with WT mice (Supplementary Fig. 3B), despite similar caloric intake (Supplementary Fig. 3C). After the 10-week HF-feeding period, BW-matched mice from each genotype (WT, 38.3 ± 1.2 g; Fxr^∆liver, 35.3 ± 1.3 g) received daily injections of vehicle or the GCGR agonist. IUB288-treated WT mice lost 17% of their original BW (P < 0.001) (Fig. 4D and Supplementary Fig. 3D), including reductions of fat and lean mass (Fig. 4E and F). GCGR agonism increased intestinal, but not liver, Gpbar1 mRNA expression in control but not Fxr^∆liver mice (Supplementary Fig. 3E). Although GPBAR1/TGR5 signaling induces Fgf21 (17) and Glp-1 (i.e., Gcg) (18), neither was differentially regulated in Fxr^∆liver mice (Table 1 and Supplementary Fig. 1A, B, and D). IUB288 efficacy was blunted in Fxr^∆liver mice, which lost significantly less BW when compared with IUB288-treated WT controls (Fig. 4D and Supplementary Fig. 3D). Furthermore, we failed to detect significant changes in either fat or lean mass in IUB288-treated Fxr^∆liver mice compared with vehicle counterparts (Fig. 4E). Notably, WT IUB288-treated mice displayed a small (16%) reduction in food intake over the treatment period that was not observed in Fxr^∆liver mice (Fig. 4F). The antiobesity effects of GCGR agonism were also associated with reduced epididymal and inguinal adipocyte size, as well as decreased lipid infiltration in brown adipose tissue (BAT) (Supplementary Fig. 4).

We assessed plasma samples from these mice to identify systems/pathways that were altered by GCGR agonism in an FXR-dependent manner (Table 1). Plasma GLP-1, insulin, plasminogen activator inhibitor 1, and glucagon levels were not altered by GCGR agonism. However, IUB288 treatment significantly decreased glucose-dependent insulinotropic polypeptide (GIP), leptin, and thyrotropin (TSH) levels in WT mice (P < 0.05), but not Fxr^∆liver mice. Resistin and ghrelin levels were significantly decreased, whereas thyroxine (T4) (but not triiodothyronine [T3]) levels were significantly increased in both genotypes upon IUB288 treatment. Unlike WT mice, Fxr^∆liver mice exhibit plasma BA accumulation as described (19) and were resistant to IUB288-induced regulation of BAs (Supplementary Fig. 3F). Together, these data highlight hepatic FXR as a critical mediator of glucagon’s antiobesity action.

To address mechanisms underlying the differential effects of GCGR agonism, we conducted indirect calorimetry in IUB288-treated Fxr^∆liver and WT mice. Food intake was not significantly reduced by this short-term GCGR agonism (Supplementary Fig. 5E and F). Nonetheless, and consistent with our prior reports (5), BW reduction following GCGR agonism in DIO WT mice associated with an increase in light- and dark-phase energy expenditure (EE) (Fig. 5A, B, and E and Supplementary Fig. 5A). In contrast, IUB288 had no effect on EE in mice lacking hepatic FXR (Fig. 5C–E and Supplementary Fig. 5B). Likewise, GCGR agonism reduced respiratory quotient (RQ) in WT but not Fxr^∆liver mice (Fig. 5F and G and Supplementary Fig. 5C and D), particularly during the light phase (Fig. 5H). Although EE was elevated in IUB288-treated WT mice, locomotor activity was not augmented by GCGR agonism in either genotype (Supplementary Fig. 5G and H). Altogether, these data suggest that GCGR agonism stimulates EE and fatty acid oxidation (FAOx), and this regulation is dependent upon hepatic FXR.

Plasma TGs trended higher after IUB288 treatment in both WT and Fxr^∆liver mice (Fig. 6A). Conversely, plasma CHL was considerably reduced and plasma β-hydroxybutyrate was elevated by IUB288 treatment in both genotypes (Fig. 6B and C). As in previous studies (Fig. 2 and ref. 5), IUB288-treated WT mice exhibited significantly reduced hepatosteatosis, whereas this effect was blunted in Fxr^∆liver mice (Fig. 6D and Supplementary Fig. 4). Consistent with a greater reduction in hepatic TG content, we also observed increased hepatic Ppargc1a expression concomitant with decreased Ppara, Scd1, and Srebp1c expression in WT mice, but not Fxr^∆liver mice (Fig. 6E).

To elucidate potential pathways that may mediate the antiobesity action of the GCGR–FXR signaling axis, we conducted RNA-sequence analysis on liver samples from IUB288-treated WT and Fxr^∆liver mice. This uncovered 953 genes differentially regulated by IUB288 treatment in an FXR-dependent manner, as well as 12 genes for which regulation was inverted in Fxr deficiency (Supplementary Fig. 6A and B). Top gene ontology–enriched pathways included oxidative phosphorylation, Eif2, p70S6K, sirtuin, and mammalian target of rapamycin signaling (Fig. 7A). Chip-sequencing enrichment analysis (20) of our data set identified retinoid X receptor (RXR), liver X receptor (LXR), and peroxisome proliferator–activated receptor α (PPARα) as likely upstream regulators (Fig. 7B). This analysis uncovered that genes related to BA (e.g., Cyp7b1, Fgfr4, and Nr1h3) and fatty acid metabolism (e.g., Nr1h2, Fasn, and Apoa4) were significantly regulated by GCGR activation (Fig. 7C). Consistent with a cell-autonomous FXR-dependent regulation, IUB288 or glucagon treatment stimulated FAOx in WT primary hepatocytes, but this activation was blunted in hepatocytes from Fxr^∆liver mice (Fig. 6F and Supplementary Fig. 6C). Likewise, liver homogenates from WT mice previously treated with IUB288 displayed enhanced FAOx when compared with vehicle-treated controls, whereas this effect is lost in liver homogenates from IUB288-treated Fxr^∆liver mice (Fig. 6G).

Glucagon, and by extension GCGR signaling, is a potent regulator of energy balance and glucose and lipid metabolism (1). Attempts to antagonize this critical metabolic pathway and thus reverse hyperglycemia have resulted in unexpected dyslipidemia, questioning whether attenuating or enhancing glucagon action is the appropriate therapeutic approach (21,22). Thus, an important and emerging question revolves around the identification of downstream mechanisms mediating GCGR action and potential segregation of GCGR-induced hyperglycemia from its antiobesity actions. In this study, we have investigated the thermogenic and antiobesity effects of GCGR signaling using IUB288 (5). We identified liver as the tissue of origin for these effects, demonstrating a role for FGF21 as a downstream regulator, and uncovered FXR signaling as an additional pathway that mediates some of the antiobesity actions of GCGR agonism. We likewise identified increased hepatocyte FAOx as a downstream action stimulated by GCGR agonism in an FXR-dependent manner. We further investigated the contributions of GCGR-mediated regulation of BA metabolism, a crucial regulator of whole-body energy balance.

Whole-body germline disruption (23) or tamoxifen-induced conditional whole-body loss of GCGR (24) function results in protection from DIO upon HFD feeding. Interestingly, our data demonstrate that mice with congenital loss of hepatic Gcgr expression were not protected from DIO. Nonetheless, our results demonstrate that IUB288-stimulated BW loss in HFD-fed mice requires intact hepatic Gcgr expression. Although it is possible that hypothalamic, GCGR-dependent inhibition of food intake (25) or secondary effects of hepatic factors in other tissues, such as BAT and white adipose tissue (WAT), could be contributing to the weight loss, our observations suggest that GCGR-increased EE is predominantly due to a hepatic effect.

As with its antiobesity effects, the beneficial effects of GCGR signaling on dyslipidemia are well known (1); however, our studies clearly identify hepatic GCGRs as the drivers of reduced plasma CHL and liver TGs. It is important to note that lack of hepatic GCGR signaling is sufficient to drive increased hepatic TG accumulation and is consistent with increased dyslipidemia following GCGR antagonism (26). These data highlight both the potential for GCGR agonists as anti–nonalcoholic fatty liver disease therapeutics as well as cautioning against GCGR antagonism.

We previously identified the hepatokine FGF21 as a crucial factor in GCGR-mediated energy metabolism (5). FGF21 null mice fail to respond to GCGR-stimulated prevention of DIO (5). However, in this study, HF feeding was initiated concurrent with GCGR agonism, and thus, FGF21 was only tested in the context of obesity prevention. In this paradigm, FGF21 was responsible for the entirety of GCGR-mediated energy balance. However, when these studies were moved to an obesity treatment paradigm, a more complex regulatory network emerged. These new studies in an obese model of liver FGF21 deficiency clearly show a blunted BW response to chronic GCGR agonism. It is possible that in pre-existing obesity, GCGR agonism stimulates FGF21 secretion from extrahepatic tissues. However, our findings in Gcgr^∆liver mice support reports suggesting that the vast majority of circulating FGF21 is hepatically derived (27). Thus, we can surmise that if FGF21 is an important downstream regulator of GCGR action, it must be hepatic in origin. Although FGF21 is a potent antiobesity signal, it is clear that in the context of GCGR signaling, there are both FGF21-dependent and -independent pathways engaged, and we must look beyond the FGF21 signaling pathway.

Glucagon, via cAMP-dependent protein kinase–dependent regulation of HNF4α, modulates hepatocyte Cyp7a1 expression, the rate-limiting enzyme in BA synthesis (28). Although this would predictably result in suppression of BA synthesis, we also observed suppression of Slc10a in IUB288-treated mice. Thus, it is possible that the elevated levels of BAs observed in circulation are the result, at least in part, of reduced hepatocyte transport at the basolateral membrane (29). Interruption of GCGR signaling (genetic or pharmacological) elevates primary and secondary plasma BAs (9,26,30,31). As compensatory effects of either genetic ablation or pharmacology could underlie these effects, our strategy to combine genetic and pharmacological interventions may provide a more complete view of these GCGR effects. Moreover, fasting, which was not controlled for in the cited reports (9,26,30,31), has a profound effect on BA levels.

We observed an elevation in the cholic acid species of BAs after GCGR agonism. This species is a potent activator of FXR (32) and suggests that glucagon signaling may regulate FXR signaling via BA metabolism. The interplay between BAs and FXR in the regulation of EE has yet to be fully elucidated. BAs increase EE in humans and rodent models of obesity (33–36), but these effects are often attributed to GPBAR1/TGR5, not FXR (37). Of interest, we observed an FXR-dependent increase in intestinal Gpbar1 expression following GCGR agonism, providing a line of future focus. Conversely, BA-binding resins reduce serum BAs and are effective to prevent and treat DIO (38). Likewise, Fxr^−/− mice are resistant to DIO (39). Consistent with this observation, chronic treatment with a synthetic FXR agonist GW4063 accentuated DIO (40), whereas FXR inhibition via tauro-β-muricholic acid (41) or glycine-β-muricholic acid (42) correlates with improved metabolic function. Thus, the role of BAs, FXR, and GPBAR1/TGR5 signaling in metabolic regulation warrants continued investigation.

Both GCGR (5) and FXR signaling (8) regulate the expression of FGF21. In this study, we report similar Fgf21 expression and circulating FGF21 levels in Fxr^−/−, Fxr^∆liver, and WT control mice, demonstrating that FXR signaling is dispensable for GCGR-induced FGF21. Therefore, it is plausible that the intermediate effect on BW observed in both Fgf21^∆liver and Fxr^∆liver mice, as compared with their appropriate controls, is a reciprocal component of the GCGR effect. Although not directly tested in this study, studies are underway to assess the combined contributions of these two pathways. We also assayed endocrine pathways known to regulate energy balance (i.e., ghrelin, GIP, leptin, resistin, TSH, T3, and T4). However, all of these factors were regulated in a similar manner between WT and Fxr^∆liver mice. This suggests that the liver is largely responsible for the FXR-dependent metabolic actions observed during GCGR agonism. Beyond FXR, GCGR activation may increase whole-body EE in part via thyroid hormone. Moreover, increased T4 levels are likely suppressing TSH in these mice. The suppressed ghrelin observed was a bit unexpected, as glucagon administration on isolated rat stomach has been reported to increase ghrelin secretion (43). Whether the observed decrease in ghrelin levels contributes to IUB288-induced BW loss cannot be completely discarded based on our experiments.

Glucagon increases oxygen consumption, body temperature, and EE in rodents (5,43) and likewise increases EE and fat oxidation in humans (44,45). Similarly, FXR regulates EE (46) and mitochondrial function (47). Our studies suggest that at least a portion of glucagon’s antiobesity action is mediated via hepatic FXR and involves an increase in EE. Consistent with accumulation of circulating IUB288 (an acylated peptide), EE increased with each subsequent dose and was most evident in the final days of indirect calorimetry. DIO Fxr^∆liver mice were unresponsive to GCGR agonism, even after 5 days of treatment. Moreover, increased EE was independent of changes in locomotor activity, suggesting that GCGR agonism stimulates basal metabolic rate in an FXR-dependent manner. Substrate preference (RQ) was also altered by GCGR agonism. RQ in all mice was suppressed (∼0.74) and indicative of the HF feeding. However, GCGR agonism was sufficient to further reduce RQ in WT but not Fxr^∆liver mice, suggesting that GCGR signaling stimulates FAOx in an FXR-dependent manner. This, along with the potent reduction in fat mass observed after IUB288 treatment, also suggests that the energetic demands induced by GCGR agonism are met via increased FAOx. Lipolysis may represent one of the main effects of GCGR activation (i.e., to fuel fat utilization). The amount of free fatty acid released from BAT by glucagon treatment is 10 times higher than that of WAT (48). Therefore, it is plausible that BAT intracellular lipid provides the first source for glucagon-stimulated FAOx, whereas WAT may represent a later source. Studies are currently underway to address these questions; however, the results described in this study confirm prior reports that GCGR agonism stimulates hepatocyte FAOx (49). Regarding the mechanisms underlying this elevated oxidative state, we observed an increase in expression of hepatic oxidative phosphorylation genes and specifically Ppargc1a. Of note, overexpression of hepatic peroxisome proliferator–activated receptor γ coactivator 1α is sufficient to increase hepatic mitochondrial respiration and whole-body fat oxidation (50), suggesting that this critical transcriptional coregulator may also contribute to FAOx and fat mass loss in our system. Likewise, elevated cAMP (as in GCGR signaling) and Ppargc1a overexpression both induce Fxr (51). Furthermore, peroxisome proliferator–activated receptor γ coactivator 1α interacts with the FXR DNA-binding domain to enhance subsequent FXR target gene induction (51). Thus, future studies will focus on the interaction of these crucial transcriptional regulators in the context of GCGR signaling.

In conclusion, we report that hepatic FXR is a critical regulator of glucagon’s antiobesity effects. The metabolic benefits of IUB288 appear to be liver cell autonomous, GCGR dependent, and mediated through parallel FGF21 and FXR pathways (Fig. 7D). These discoveries serve to further highlight the emerging value of fasting hormone pathways as superior target pathways for the treatment of metabolic disease. Additional dissection of the detailed molecular interactions connecting GCGR activation with FXR signaling and FGF21 induction may provide novel drug targets for the treatment of metabolic diseases.

Funding. The project described was supported by National Institutes of Health National Institute of Diabetes and Digestive and Kidney Diseases grants 5K01-DK-098319 and 1R01-DK-112934 (to K.M.H.), R01-DK-077975 (to D.P.-T.), and P30-DK-079626, as well as the UAB Small Animal Phenotyping Core supported by the National Institutes of Health Nutrition Obesity Research Centers (P30-DK-056336), Diabetes Research Center (P30-DK-079626), and the UAB Nathan Shock Center (PAG-050886A). This work was also supported by American Diabetes Association grant 1-13-JF-21 (K.M.H.), Canadian Institutes of Health Research grants 136942 and 154321, and the Canada Research Chairs Program (to D.J.D.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. T.K., S.N., C.H., L.W., T.F.B., and C.S. generated experimental data. T.K. and K.M.H. were responsible for study conception and design, data analyses and interpretation, and drafting the article. M.P. and A.R.W. were responsible for RNA-sequencing analyses. M.E.Y., S.B., D.J.D., B.F., R.D., D.P.-T., and M.T. advised on study concept and critical revision of the article. K.M.H. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented at the 76th Scientific Sessions of the American Diabetes Association, New Orleans, LA, 10–14 June 2016, and the 77th Scientific Sessions of the American Diabetes Association, San Diego, CA, 9–13 June 2017.