Orphan Nuclear Receptor Estrogen-Related Receptor γ (ERRγ) Is Key Regulator of Hepatic Gluconeogenesis*

Background: Dysregulation of glucose homeostasis is often associated with insulin resistance and diabetes. Results: Hepatic ERRγ expression is increased by fasting-dependent activation of the CREB-CRTC2 pathway, which leads to the induction of hepatic gluconeogenesis. Conclusion: Orphan nuclear receptor ERRγ is a novel transcriptional regulator of hepatic gluconeogenesis. Significance: An ERRγ inverse agonist could be a new potential therapeutic approach for the treatment of type 2 diabetes. Glucose homeostasis is tightly controlled by hormonal regulation of hepatic glucose production. Dysregulation of this system is often associated with insulin resistance and diabetes, resulting in hyperglycemia in mammals. Here, we show that the orphan nuclear receptor estrogen-related receptor γ (ERRγ) is a novel downstream mediator of glucagon action in hepatic gluconeogenesis and demonstrate a beneficial impact of the inverse agonist GSK5182. Hepatic ERRγ expression was increased by fasting-dependent activation of the cAMP-response element-binding protein-CRTC2 pathway. Overexpression of ERRγ induced Pck1 and G6PC gene expression and glucose production in primary hepatocytes, whereas abolition of ERRγ gene expression attenuated forskolin-mediated induction of gluconeogenic gene expression. Deletion and mutation analyses of the Pck1 promoter showed that ERRγ directly regulates the Pck1 gene transcription via ERR response elements of the Pck1 promoter as confirmed by ChIP assay and in vivo imaging analysis. We also demonstrate that GSK5182, an inverse agonist of ERRγ, specifically inhibits the transcriptional activity of ERRγ in a PGC-1α dependent manner. Finally, the ERRγ inverse agonist ameliorated hyperglycemia through inhibition of hepatic gluconeogenesis in db/db mice. Control of hepatic glucose production by an ERRγ-specific inverse agonist is a new potential therapeutic approach for the treatment of type 2 diabetes.

The estrogen receptor-related receptor subfamily consists of three members, ERR␣, -␤, and␥ (NR3B1-3), which bind to classic estrogen response elements (EREs) and to extended half-site core sequences (TNAAGGTCA; ERR response element (ERRE)) as either monomers or dimers (11). Structural studies indicate that ERR␥ is constitutively active in the absence of endogenous ligands, but small molecule ligands that could further transcriptionally activate or repress ERR␥ have been reported to date (12)(13)(14)(15). The ligand-independent transcriptional activity of ERR␥ depends on nuclear receptor coregulators, such as steroid receptor coactivator 2, PGC-1␣, receptorinteracting protein 140, and small heterodimer partner (SHP), all of which are involved in the regulation of hepatic glucose metabolism (16 -20). It has been reported that ERRs are expressed in tissues with high metabolic demand and regulated by peripheral circadian clock in key metabolic tissues, such as white and brown adipose tissues, muscle, and liver (21). Interestingly, ERR␥ has been reported to play an important role in the regulation of a nucleus-encoded mitochondrial genetic network that coordinates postnatal metabolic transition in cardiac muscle as evidenced by phenotype analyses of perinatally lethal ERR␥-null mice (22). Recently, chromatin immunoprecipitation (ChIP)-on-chip analysis in cardiomyocytes has shown that ERR␣ and ERR␥ have the potential to regulate mitochondrial programs involved in oxidative phosphorylation, and gene expression analysis in ERR␥-null mice has demonstrated that this receptor plays an important role in regulation of potassium homeostasis in the heart, kidney, and stomach (23,24). However, the role of ERR␥ in hepatic glucose metabolism in adults remains largely unknown. In this study, we have demonstrated that orphan nuclear receptor ERR␥ is a novel transcriptional regulator of hepatic gluconeogenesis, and its inverse agonist could ameliorate hyperglycemia in mouse models of type 2 diabetes.

EXPERIMENTAL PROCEDURES
Chemicals-8-Br-cAMP and insulin were purchased from Sigma-Aldrich and dissolved in the recommended solvents. GSK5182 was synthesized as described previously (15). Synthetic method of D4 is described in the supplemental Methods. GSK5182 was used as in HCl salt form and dissolved in sterile filtered 30% PEG-400 aqueous solution to give a 40 mg/kg concentration for in vivo experiments.
Cell Culture and Transient Transfection Assay-HepG2, H4IIE, and AML12 cells were maintained as described previously (25). Transient transfection was performed using Lipofectamine 2000 (Invitrogen) or SuperFect (Qiagen, Hilden, Germany) according to the manufacturers' instructions. The cells were treated with 10 M forskolin for 6 h, 500 M 8-Br-cAMP for 6 h, 10 M dexamethasone for 18 h, and/or 10 M GSK5182 for 24 h unless noted otherwise. The cells then were harvested 48 h after transfection, and luciferase activity was measured. Luciferase activity was normalized to ␤-galactosidase activity. The data are representative of at least three to five independent experiments.
Culture of Primary Hepatocytes-Primary hepatocytes were isolated from Sprague-Dawley rats (male, 180 -300 g) by collagenase perfusion (28) and seeded with Medium 199 (Cellgro). After 3-6 h of attachment, cells were infected with the indicated adenoviruses for overexpression and/or treated with 10 M forskolin for the final 6 h and 10 M GSK5182 for the final 24 h unless noted otherwise.
Western Blot Analysis-Whole cell extracts were prepared using radioimmune precipitation assay buffer (Elpis-Biotech). Proteins from whole cell lysates were separated by 10% SDS-PAGE and then transferred to nitrocellulose membranes. The membranes were probed with monoclonal ERR␥ antibodies (Perseus Proteomics, Tokyo, Japan), PGC-1␣ (Santa Cruz Biotechnology, Santa Cruz, CA), and SHP (Santa Cruz Biotechnology). Immunoreactive proteins were visualized using an Amersham Biosciences ECL kit (GE Healthcare) according to the manufacturer's instructions.
ChIP Assay-Nuclear isolation and cross-linking on cell lines, primary hepatocytes, and liver samples were performed as described previously (23,29). After sonication, soluble chromatin was subjected to immunoprecipitation using anti-FLAG M2 (Stratagene), anti-ERR␥ (Perseus Proteomics), anti-CRTC2 (Santa Cruz Biotechnology), and anti-PGC-1␣ (Santa Cruz Biotechnology). DNA was recovered by phenol/chloroform extraction and analyzed by PCR or Q-PCR using primers against relevant promoters.
In Vivo GST Pulldown Assay-HepG2 cells were transfected with pEBG (GST), pEBG-ERR␥, and HA-PGC-1␣ vectors and then treated with 1 and 10 M GSK5182 for 24 h. Cell lysates were co-immunoprecipitated with GST beads and anti-HA (Roche Applied Science) and anti-GST (Santa Cruz Biotechnology) antibodies as described previously (29).
Animal Experiments-Male 7-12-week-old C57BL/6J and db/db mice (Charles River Laboratories) were maintained on a 12-h/12-h light/dark cycle and fed ad libitum. GSK5182 (40 mg/kg/day as a final dose) and corn oil emulsion were sonicated again immediately before injection of db/db mice. After 14 h of fasting, intraperitoneal injections were performed for 5 days. After the injections, blood glucose levels were monitored after 4 h of fasting. All experiments were conducted following the guidelines of the Sungkyunkwan University School of Medicine Institutional Animal Care And Use Committee.
Glucose Output Assay-Twenty-four hours after seeding of rat primary hepatocytes, the medium were replaced with Krebs-Ringer buffer (115 mM NaCl, 5.9 mM KCl, 1.2 m MgCl 2 , 1.2 mM NAH 2 PO 4 , 2.5 mM CaCl 2 , 25 mM NaHCO 3 , pH 7.4) supplemented with 10 mM lactate and 1 mM pyruvate. The cells were treated with 10 M forskolin for 12 h, 100 nM insulin for 18 h, and/or 10 M GSK5182 for 18 h. The glucose level in the medium was measured using a QuantiChrom glucose assay kit (Bioassay Systems, Hayward, CA).
Electrophoretic Mobility Shift Assay (EMSA)-Doublestranded oligonucleotides containing a CRE site on ERR␥ promoter or the ERR␥ binding sites (ERREs) on Pck1 promoter were generated and labeled with [␣-32 P]dCTP using the Klenow fragment of DNA polymerase I. The oligonucleotide sequences were as follows: sense 5Ј-GGGGCGGCT-CGCGTCGCCTCCCTCC-3Ј and antisense 5Ј-GGGGGG-AGGGAGGCGACGCGAGCCG-3Ј for CRE; sense 5Ј-GGGG-TGGACCTCCAGGTCATTTCGT-3Ј and antisense 5Ј-GGG-GACGAAATGACCTGGAGGTCCA-3Ј for ERRE1; and sense 5Ј-GGGGGCCTCCCTGACCTAAGGGA-3Ј and antisense 5Ј-GGGGTCCCTTAGGTCAGGGAGGC-3Ј for ERRE2. Underlined sequences were substituted to AA for mutant CRE and TT for mutant ERRE1 and ERRE2. Purified recombinant protein for CREB or ERR␥ was generated by in vitro translation using the TNT-coupled reticulocyte lysate system (Promega). Unprogrammed TNT-coupled reticulocyte lysate was used as a negative control, and 32 P-labeled double-stranded oligonucleotides containing the ERR␥ binding site on DAX-1 promoter were used as a positive control (31). Unlabeled oligonucleotides were added as cold competitors at 50 -100-fold molar excess where indicated. DNA-protein complexes were separated on a 4% polyacrylamide gel in 0.5ϫ Tris borate-EDTA. The gels were dried and then analyzed by autoradiography.
Surface Plasmon Resonance Analysis-The dissociation constant of GSK5182 with ERR␥ ligand binding domain was determined by surface plasmon resonance spectroscopy using a Biacore T100 instrument (GE Healthcare). The surface carboxyl group of a CM5 sensor chip was activated with a mixture of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and N-hydroxysuccinimide in flow cells 3 and 4 to generate the reactive succinimide ester on the CM5 chip. ERR␥ ligand binding domain (100 g/ml in 10 mM NaOAc, pH 5.0) was passed through flow cell 4 (5200 resonance units) and immobilized via amide bond formation with succinimide ester on the CM5 chip. The remaining succinimide ester on flow cells 3 and 4 was quenched by an injection of 1 M ethanolamine HCl, pH 8.0. Phosphate-buffered saline (PBS) was used as a running buffer throughout the immobilization process. After the immobilization, various concentrations of GSK5182 ranging from 50 nM to 5 M were injected for 60 s with a flow rate of 30 l/min, and the dissociation of GSK5182 from ERR␥ ligand binding domain immobilized on the sensor chip surface was monitored for 600 s at the same flow rate. The running buffer was 10 mM HEPES buffer, pH 7.5 containing 5% DMSO, 50 mM NaCl, 2 mM MgCl 2 , 1 mM EDTA, 1 mM DTT, and 0.005% P20. The binding events were measured at 25°C, and a DMSO correction was performed during the binding assay. The data were analyzed using Biacore T100 Evaluation software. Final sensorgrams were obtained by eliminating responses from flow cell 3 and bufferonly control. The dissociation constant (K D ) was calculated by fitting the sensorgrams to a 1:1 binding model.
Quantitative PCR-Total RNA from either primary hepatocytes or liver was extracted using an RNeasy minikit (Qiagen). cDNA generated by Superscript II enzyme (Invitrogen) was analyzed by Q-PCR using a SYBR Green PCR kit and a TP800 Thermal Cycler DICE Real Time system (Takara). All data were normalized to ribosomal L32 expression.
Statistical Analyses-All values are expressed as means Ϯ S.E. The significance between mean values was evaluated by two-tailed unpaired Student's t test.

Hepatic ERR␥ Expression Is Induced by cAMP Signaling under
Fasting-It has been reported that hepatic expression of ERR␥ rhythmically oscillates in the daily light/dark cycle and is induced during fasting (21,32), suggesting that it could be regulated by nutritional status. Because glucagon stimulates hepatic glucose production mainly through the cAMP signaling pathway under fasting (33), we first investigated whether ERR␥ expression is induced by the adenylate cyclase activator forskolin (FSK) in AML12 cells and rat primary hepatocytes. The mRNA and protein levels of ERR␥ were rapidly increased 1 and 3 h, respectively, after the addition of FSK (Fig. 1, A and B), whereas the induction of Pck1 mRNA occurred after 1-h treatment with FSK. However, expression of ERR␣ was not significantly changed (Fig. 1A). Notably, the expression of Pck1 and G6PC occurred after 30 and 45 min, respectively, whereas the induction of ERR␥ mRNA was only increased after 1-h treatment with FSK (Fig. 1C). These results indicate that FSK stimulation led to the initial induction of Pck1 and G6PC, which is faster than that of ERR␥. To further test the correlation between ERR␥ and gluconeogenic gene expression in vivo, we analyzed mRNA levels for these genes in wild type mice under fasting over time. Similar to the results in cultured cell lines, Pck1 and G6PC mRNA levels were rapidly induced after 1-h of fasting and were strongly induced after 6-h fasting conditions, whereas ERR␥ gene expression was only enhanced 3 h after fasting and further elevated until 12 h post-food deprivation (Fig. 1D). These results suggest that, in addition to the initial induction of gluconeogenic gene expression by fasting, ERR␥ expression precedes that of additional gluconeogenic gene expression in vivo.
In an attempt to determine whether ERR␥ expression is involved in the additive effect of cAMP on Pck1 gene expression, H4IIE cells were treated with cAMP alone or cAMP plus cycloheximide (CHX), a protein synthesis inhibitor, in a timedependent manner. Pck1 gene expression was significantly increased by cAMP stimulation for 1 h and was further enhanced by 6-h stimulation with cAMP (Fig. 1E). Interestingly, Pck1 gene induction by 1-h treatment with cAMP was not affected by CHX treatment, whereas the additional Pck1 gene induction by 6-h treatment with cAMP was blocked by CHX treatment, clearly suggesting that cAMP-mediated ERR␥ expression contributes to the additional pathway of cAMP-mediated Pck1 gene expression.
Regulation of ERR␥ Expression Is Mediated by CREB-CRTC2-We next explored the potential mechanisms underlying the induction of ERR␥ by cAMP during fasting. We first investigated the role of CRTC2, a mediator of the cAMP-dependent transcriptional program in hepatocytes (6,34), in FSKmediated elevation of ERR␥ and its target genes. Hepatic expression of constitutively active CRTC2 (CRTC2 S171A) significantly increased mRNA levels of ERR␥, Pck1, PGC-1␣, and CPT-1␣ but not of SCD-1 and ERR␣ ( Fig. 2A and data not shown). Conversely, shRNA-mediated knockdown of CRTC2 in mouse liver considerably reduced ERR␥ and gluconeogenic gene expression (Fig. 2B). Adenovirus-mediated expression of SIK1, a known inhibitor of CRTC2 (6), also reduced hepatic ERR␥ and gluconeogenic gene expression in mouse primary hepatocytes, but kinase-inactive SIK1 T182A did not (Fig. 2C), further supporting the notion that endogenous CRTC2 induces ERR␥ expression.
Investigation of human, mouse, and rat ERR␥ promoter sequences revealed the presence of a potential CRE (Fig. 2E,  bottom). In transient transfection, we demonstrated that FSK and CREB-CRTC2 significantly increased the ERR␥ promoter activity, and this effect was abolished either by the co-transfection of A-CREB or SIK kinases or by mutation of the ERR␥ promoter CRE (Fig. 2, D and E). Furthermore, ChIP demonstrated that CRTC2 was recruited to the CRE of ERR␥ promoter as well as that of Pck1 or G6PC promoters in the presence of FSK in mouse primary hepatocytes (Fig. 2F). In addition, EMSA revealed that CREB directly binds to CRE on ERR␥ promoter (Fig. 2G), showing that CREB-CRTC2 directly regulates ERR␥ at the transcriptional level.
ERR␥ Regulates Gluconeogenic Gene Expression-The response of ERR␥ to fasting in liver suggested that this receptor might directly regulate hepatic gluconeogenesis. Indeed, infection with adenovirus expressing ERR␥ (Ad-ERR␥) significantly induced the mRNA levels of Pck1 and G6PC as well as their promoter activities in cultured cells (Fig. 3, A and B). Moreover, ERR␥ expression led to the induction of PGC-1␣ expression ( Fig. 3A, right). Interestingly, the induction of Pck1 promoter activity was observed with co-transfection of expression vector for ERR␥ but not with that of other ERR isoforms (Fig. 3B). However, ERR␥-mediated Pck1 promoter activity was inhibited by ERR␣ (Fig. 3C). Consistent with previous reports regarding the roles of PGC-1␣ and SHP in the transcriptional activity of ERR␥ (17,19), PGC-1␣ potentiated and SHP inhibited ERR␥mediated induction of gluconeogenic gene expression (Fig. 3A). We confirmed the competition between PGC-1␣ and SHP for the transcriptional activity of ERR␥ on Pck1 promoter (Fig. 3D), suggesting that the regulation of hepatic gluconeogenesis by PGC-1␣ and SHP may be achieved in part by competition between these factors for the association with ERR␥ on gluconeogenic promoters.
ERR␥ Directly Activates Pck1 Gene Expression at Transcription Level-PEPCK is the key enzyme controlling the rate of hepatic gluconeogenesis and is regulated at the transcriptional level (35). We identified two potential conserved ERREs in the human, mouse, and rat Pck1 promoters (supplemental Fig. 1). We confirmed the functional significance of these sites using serial deletion (Fig. 4A) or ERRE mutants of the Pck1 promoter in transfection assays (Fig. 4B). In addition, double ERRE mutations of the Pck1 promoter largely blunted cAMP-or PGC-1␣-mediated induction of wild type promoter activity (Fig. 4, C and E), whereas CRE-mutated Pck1 promoter activity was significantly induced by ERR␥, suggesting a significant role of the Pck1 ERREs in cAMP-or PGC-1␣-mediated induction of Pck1 promoter activity. However, interestingly, Ad-shERR␥ led to marked reduction of basal and FSK-induced gluconeogenic gene expression in rat primary hepatocytes (Fig. 4D), indicating that ERR␥ has indirect effects on Pck1 gene expression because it completely abolished the effect of cAMP, whereas mutation of the ERRE did not.
FSK-dependent occupancy of ERR␥, but not ERR␣, on the Pck1 ERREs was confirmed by EMSA and ChIP assay (Fig. 4, F  and G). An in vivo ChIP assay demonstrated that PGC-1␣ was recruited to the Pck1 ERREs in livers of fasted mice but not of fed mice (Fig. 4H). To further examine the importance of ERREs in the Pck1 transcription by fasting signals in vivo, we performed in vivo imaging analysis with an adenoviral reporter construct carrying either wild type or ERRE mutant Pck1 promoter fused to luciferase. Fasting increased wild type Pck1 promoter activity 45-fold over feeding controls (Fig. 4I). However, the stimulatory effect of fasting was largely ablated in mice with the double ERRE mutant promoter, strongly indicating that ERR␥, rather than the CRE, is a critical downstream mediator of Finally, to elucidate whether ERR␥-dependent induction of gluconeogenic gene expression results in increased glucose production, we assayed glucose output in rat primary hepatocytes. As expected, Ad-ERR␥ increased glucose production in primary hepatocytes, and its effect was greatly decreased by the addition of insulin (Fig. 4J). Taken together, these data suggest that ERR␥ exerts its effects on hepatic glucose production through direct transcriptional regulation of gluconeogenic genes.
GSK5182 Specifically Inhibits Transcriptional Activity of ERR␥-GSK5182 is a 4-hydroxytamoxifen (4-OHT) analog and a selective inverse agonist of ERR␥ relative to ER␣ due to its additional non-covalent interaction with Tyr-326 and Asn-346 of ERR␥ (15) (supplemental Fig. 2). Indeed, using the mammalian one-hybrid assay, we confirmed that GSK5182, but not D4 (a non-functional synthetic analog of 4-OHT), directly inhibited transcriptional activity of ERR␥ (Fig. 5A). Interestingly, this inhibitory effect of GSK5182 depends on the interaction with Tyr-326 rather than Asn-346 of ERR␥. We then performed biophysical binding analysis of GSK5182 with ERR␥ ligand binding domain using surface plasmon resonance spectroscopy and observed high binding affinity (K D ϭ 65 nM), which confirms the direct and specific interaction of GSK5182 with ERR␥ (Fig.  5B). Unlike 4-OHT, GSK5182 had no effect on estradiol-in-duced transactivation by ER␣ or other nuclear receptors (Fig. 5,  C and D), indicating that the inverse agonist GSK5182 specifically inhibits the transcriptional activity of ERR␥.
Previous studies revealed that binding of 4-OHT to ERR␥ leads to a conformational change in the AF-2 domain that blocks PGC-1␣ binding (13,36). Similarly, transient transfection and in vivo GST pulldown assays showed that treatment with GSK5182 inhibited the PGC-1␣-potentiated ERR␥ transcriptional activity and disrupted the interaction of ERR␥ with PGC-1␣, effects not observed with D4 ( Fig. 5E and supplemental Fig. 3). cAMP-or PGC-1␣-induced Pck1 promoter activity was also significantly inhibited by GSK5182 (Fig. 5F). To test whether GSK5182 specifically inhibits the recruitment of PGC-1␣ to ERREs on the Pck1 promoter, ChIP assays were performed in HepG2 cells treated with GSK5182. Indeed, GSK5182 inhibited the occupancy of PGC-1␣ over ERREs on Pck1 promoter without affecting the binding of ERR␥ on the same region (Fig. 5G). These results demonstrate that GSK5182 suppresses the expression of gluconeogenic genes by disrupting the interaction between ERR␥ and PGC-1␣ without affecting the DNA binding ability of ERR␥.
Inverse Agonist of ERR␥ Improves Hyperglycemia in db/db Mice-Based on the ability of ERR␥ to regulate the gluconeogenic program, we next examined the effect of this inverse agonist on gluconeogenic gene expression and glucose output in rat primary hepatocytes. Indeed, GSK5182 treatment decreased FSK-induced gluconeogenic gene expression to 40% without a change in ERR␥ mRNA levels in rat primary hepatocytes (Fig. 6A). Consistent with the reduction of gluconeogenic genes, GSK5182 also attenuated FSK-induced glucose production in primary hepatocytes (Fig. 6B).
Finally, to assess whether GSK5182 directly affects hepatic glucose metabolism in vivo, fasting blood glucose levels were measured in db/db mice that were injected intraperitoneally with GSK5182. Indeed, GSK5182-injected mice showed a marked reduction in fasting blood glucose levels compared with control groups (Fig. 6C). Consistent with this result, the expression of gluconeogenic genes and PGC-1␣ was markedly decreased in GSK5182-treated db/db mice (Fig. 6D). In addition, GSK5182 significantly reduced the occupancy of PGC-1␣ on the Pck1 HNF4 regulatory element as well as the Pck1 ERREs (Fig. 6E). No significant changes were shown in plasma insulin, triglyceride, or total cholesterol levels with GSK5182 treatment (Fig. 6, F-H). Taken together, these results suggest that inverse agonist-mediated inactivation of ERR␥ ameliorates the hyperglycemic phenotype in type 2 diabetic mice via direct regulation of hepatic gluconeogenesis.

DISCUSSION
In the current study, we identified the orphan nuclear receptor ERR␥ as a novel alternative downstream mediator of cAMP-CREB in hepatic glucose metabolism (Fig. 7), which is evidenced by the results showing that FSK-mediated induction ofPck1 and G6PC mRNA expression was faster than the induction of ERR␥ mRNA expression in vivo and in vitro. In addition, CHX treatment did not block the initial induction of Pck1  mRNA expression, whereas the additional Pck1 gene expression by 6-h treatment with cAMP was blocked by CHX treatment. These results are further supported by the ChIP assay showing that ERR␥ occupancy was weakly increased by 1-h stimulation with FSK, and the occupancy on Pck1 ERREs was strongly enhanced by 6-h treatment with FSK. We also found that mutation of the Pck1 CRE markedly reduces the effect of cAMP, whereas mutation of the Pck1 ERREs did not abolish the effect of cAMP. Overall, cAMP-mediated activation of CREB-CRTC2 triggers the initial induction of hepatic gluconeogenic gene expression in response to fasting, and then ERR␥ expression contributes to the additional gluconeogenic gene expression.
We have also demonstrated that ERR␥ contributes significantly to the fasting-mediated glucose production through hepatic gluconeogenesis. FSK-induced Pck1 and G6PC gene expression was reduced up to 80% by knockdown of ERR␥ in hepatocytes. Moreover, the fasting-mediated induction of Pck1 promoter activity was completely abolished by double ERRE mutations compared with that of wild type promoter activity in vivo, suggesting a more important role for ERR␥ induction, rather than the CRE, in mediating the effect of fasting on Pck1 promoter activity in vivo. In addition, GSK5182-mediated inhibition of ERR␥ transcriptional activity decreased gluconeogenic gene expression and improved hyperglycemia in db/db mice. On the other hand, it is known that ERR␥ induces expres-sion of pyruvate dehydrogenase kinase 4 (PDK4), which inhibits the activity of pyruvate dehydrogenase complex, a key regulatory enzyme in the oxidation of glucose to acetyl-CoA (37,38). It has been reported that expression of PDK4 is elevated in diabetes and starvation, leading to decreased oxidation of pyruvate to acetyl-CoA (39,40), and inhibition of PDKs by their inhibitors reduces hyperglycemia in type 2 diabetic rats (41).  . ERR␥ regulates hepatic gluconeogenesis. Glucagon triggers hepatic gluconeogenic gene expression though activation of CREB-CRTC2 in response to fasting. In addition, hepatic ERR␥ expression is also increased by glucagon-mediated activation of CREB-CRTC2, which in turn leads to additional gluconeogenic gene expression through cooperation with PGC-1␣. ERR␥-dependent induction of gluconeogenic genes is directly inhibited by its inverse agonist GSK5182 in a PGC-1␣-dependent manner.
Taken together, these observations suggest that ERR␥ may have a dual role by inhibiting glucose oxidation and inducing gluconeogenic flux.
It has been reported that PGC-1␣, a master regulator of hepatic glucose metabolism, is closely associated with the transcriptional activity of ERR␣ and ERR␥ (42). For example, the regulation of key metabolic processes, such as mitochondrial biogenesis and oxidative phosphorylation by PGC-1␣, was shown to be dependent on the presence of ERR␣ (43). Several lines of evidence have also shown that the transcriptional regulation of PDK4 expression by PGC-1␣ is mediated by ERR␣ or ERR␥ (37,44). Similar to these reports, we also found that PGC-1␣ plays an important role in ERR␥-mediated regulation of gluconeogenic gene expression as evidenced by an in vivo ChIP assay showing strong recruitment of PGC-1␣ on ERREs of Pck1 promoter during fasting. Moreover, PGC-1␣ expression was also induced by ERR␥, suggesting that they reciprocally cooperate for the regulation of gluconeogenesis. Notably, mutation of the ERREs of the Pck1 promoter largely attenuated PGC-1␣-mediated induction of wild type promoter activity, further supporting the important role of the ERREs for the effect of PGC-1␣. To date, most studies have been focused on the factors that control the Ϫ500 region from the start site of Pck1 gene transcription because this region contains many critical elements that regulate the response of the gene to diet and hormones (1,2). However, a number of researches have suggested that a part of Pck1 promoter that extends considerably upstream from the better characterized downstream region is involved in both the hormonal and tissue-specific control of Pck1 gene transcription (35,45). Indeed, a new extended glucocorticoid regulatory unit that is present at Ϫ1365 in the Pck1 promoter has been characterized, and this extended glucocorticoid regulatory unit was shown to be liver-specific and to play an important role in the regulation of Pck1 gene transcription by glucocorticoids (46). ERR␣ has been shown to repress the PGC-1␣-mediated gluconeogenic program through inhibition of recruitment of PGC-1␣ to Pck1 promoter (47). We found that ERR␣ inhibits the ERR␥ transcriptional activity for Pck1 promoter, suggesting that ERR␣ could also inhibit the association between ERR␥ and PGC-1␣. However, ERR␣ did not affect the FSK-stimulated DNA binding ability of ERR␥ on Pck1 ERREs. Similarly, it has been shown that ERR␣ can inhibit ERR␥ transcriptional activity without affecting the DNA binding ability of ERR␥ and suggested that a mechanism for ERR␣ repression is the heterodimerization between the ERRs as evidenced by the reports showing that heterodimerization between ERR␣ and ERR␥ inhibits transactivation of each other, whereas homodimerization is needed for their transcriptional activity (32,48,49). On the other hand, their functions in terms of transcriptional regulation of each ERR for downstream targets are more complicated. This notion is supported by previous reports that the transcriptional induction of SHP gene expression is regulated by ERR␥ but not ERR␣ or ERR␤ (17) and that ERR␣ and ERR␥ could directly control the same target genes (23). In addition, it has been reported that ERR␣ and ERR␥ in breast cancer exhibited correlative opposing functions as negative and positive biomarkers and lead to different responses of specific metabolic genes (50,51). Therefore, the molecular mechanism for the different transcriptional regulation or output of each ERR isoform needs further characterization.
ERR␣ and ERR␥ are known to have the potential to regulate mitochondrial programs involved in fatty acid oxidation in cardiac muscle (23). It has been also reported that muscle-specific ERR␥ transgenic mice exhibited increased muscle mitochondrial activity and oxidative capacity through induction of slow twitch muscle fiber properties characterized by high mitochondrial content, fatigue-resistant fibers, and dense vascularity (52,53). Consequently, these mice showed enhanced oxygen consumption and running endurance and reduced weight gain upon high fat diet compared with control. Interestingly, muscle-specific PEPCK transgenic mice also displayed greatly enhanced physical activity, running endurance, and longevity due in part to an increased the number of mitochondria (54), a phenotype similar to that of ERR␥ transgenic mice. Based on our results that ERR␥ regulates Pck1 gene expression as demonstrated by in vitro and in vivo studies, the alteration of energy metabolism by overexpression of ERR␥ in skeletal muscle may occur through induction of PEPCK.
A number of other nuclear receptors directly regulate hepatic gluconeogenesis. The orphan nuclear receptor HNF4␣ has been recognized to stimulate the gluconeogenic genes in a PGC-1␣-dependent manner in response to fasting (5), whereas the NR4A pathway is PGC-1␣-independent (55). The orphan nuclear receptors TR4 and retinoid-related orphan receptor ␣ have also been reported to induce hepatic gluconeogenesis during fasting (56,57). However, none of these orphan receptors are ligand-responsive. We have confirmed that GSK5182, an inverse agonist of ERR␥, strongly suppresses hepatic gluconeogenesis via direct and specific inhibition of the transcriptional activity of ERR␥ by blocking its association with PGC-1␣. Recently, we also reported that GSK5182 inhibited the transcriptional activity of ERR␥ on PDK4 expression by recruitment of the transcriptional corepressor SMILE (small heterodimer partner-interacting leucine zipper protein) (38,58), suggesting that GSK5182 actively mediates an exchange of coactivator for corepressor on transcriptional activity of ERR␥.
In conclusion, we have demonstrated that the reduction of ERR␥ activity by inverse agonist treatment inhibits gluconeogenic gene expression and lowers blood glucose levels in diabetic mice. Moreover, we have revealed that GSK5182 improves the impaired hepatic insulin signaling induced by diacylglycerol-mediated protein kinase C ⑀ (PKC⑀) activation (59). Inhibition of hepatic gluconeogenesis is emerging as a promising intervention strategy in type 2 diabetes, and the generation of ERR␥-specific inverse agonists provides a new therapeutic strategy for the treatment of type 2 diabetic patients.