Estrogen-related Receptor α Is a Repressor of Phosphoenolpyruvate Carboxykinase Gene Transcription*

The orphan nuclear receptor estrogen-related receptor (ERR) α is a downstream effector of the transcriptional coactivator PGC-1α in the regulation of genes important for mitochondrial oxidative capacity. PGC-1α is also a potent activator of the transcriptional program required for hepatic gluconeogenesis, and in particular of the key gluconeogenic enzyme phosphoenolpyruvate carboxykinase (PEPCK). We report here that the regulatory sequences of the PEPCK gene harbor a functional ERRα binding site. However, in contrast to the co-stimulating effects of ERRα and PGC-1α on mitochondrial gene expression, ERRα acts as a transcriptional repressor of the PEPCK gene. Suppression of ERRα expression by small interfering RNA leads to reduced binding of ERRα to the endogenous PEPCK gene, and an increase in promoter occupancy by PGC-1α, suggesting that part of the ERRα function at this gene is to antagonize the action of PGC-1α. In agreement with the in vitro studies, animals that lack ERRα show increased expression of gluconeogenic genes, including PEPCK and glycerol kinase, but decreased expression of mitochondrial genes, such as ATP synthase subunit β and cytochrome c-1. Our findings suggest that ERRα has opposing effects on genes important for mitochondrial oxidative capacity and gluconeogenesis. The different functions of ERRα in the regulation of these pathways suggest that enhancing ERRα activity could have beneficial effects on glucose metabolism in diabetic subjects by two distinct mechanisms: increasing mitochondrial oxidative capacity in peripheral tissues and liver, and suppressing hepatic glucose production.

Nuclear receptors mediate the effects of many hormonal and dietary signals. These receptors bind to specific genomic sequences, recruit coactivators or corepressors of transcription, and regulate accordingly the expression of genes important for a wide range of biological processes, including development, reproduction, and metabolism (for reviews, see Refs. 1 and 2 and references therein). The estrogen-related receptor (ERR) 2 ␣ is a nuclear receptor with high sequence similarity to the estrogen receptors, and the founding member of a small family of orphan receptors that also includes ERR␤ and ERR␥ (3,4). Despite their similarity to estrogen receptors, ERRs are not activated by estrogens or other known natural agonists (reviewed in Refs. 5 and 6). Structural studies of ERR␣ and ERR␥ indicate that these receptors can achieve a transcriptionally active conformation in the absence of a ligand, and suggest that ERR activity may not be subject to regulation by small lipophilic molecules (7,8). As an alternative mechanism of regulation, ERR␣ activity is controlled by the availability of specific coactivators that act as protein ligands (9 -11). Notably, the transactivation function of ERR␣ is weak in many cells where other nuclear receptors are active, and is greatly enhanced by expression of the transcriptional coactivators PGC-1␣ or PGC-1␤ (9 -11). PGC-1␣ not only activates the transcriptional function of ERR␣ but also induces the expression of this receptor (9). Consistent with the ability of PGC-1␣ to induce ERR␣ activity and expression, PGC-1␣ and ERR␣ show similar spatial and temporal expression patterns in vivo. They are co-expressed at high levels in tissues with high energy demands, and are co-induced in a tissue-specific manner in response to signals like fasting, exposure to cold, and physical exercise (9,12,13).
Several recent studies support a role for ERR␣ in the regulation of mitochondrial oxidative capacity. Binding sites for ERR␣ (ERR␣ response elements) are present in genes with roles in fatty acid oxidation (PPARA and ACADM), mitochondrial biogenesis (NRF2 and NRF1), the citric acid cycle (IDH3A), the respiratory chain (CYCS and ATP5B), and mitochondrial dynamics and function (MFN2) (13)(14)(15)(16)(17)(18). At these targets, ERR␣ co-operates with PGC-1␣ to induce gene expression (16 -18). Consistent with the ERR␣ functions observed in cell culture systems, ERR␣ null mice show defects in lipid metabolism and decreased expression of genes coding for fatty acid oxidation enzymes and oxidative phosphorylation components (18,19).
The transcriptional coactivator PGC-1␣ coordinates, in a signal-and tissue-specific manner, the induction of genes important for multiple adaptive metabolic responses (reviewed in Ref. 20). PGC-1␣ functions in adaptive thermogenesis in brown adipose tissue (21)(22)(23), and fibertype specification in skeletal muscle (24), two processes that involve the regulation of mitochondrial biogenesis and function. PGC-1␣ also regulates gluconeogenesis in liver (25)(26)(27). The induction of hepatic gluconeogenesis during periods of fasting is an essential adaptive response that requires the enhanced transcription of genes encoding rate-determining gluconeogenic enzymes, such as phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase (28 -31). Transcrip-tion of the PEPCK gene is tightly regulated by hormones, including glucagon, insulin, and glucocorticoids, via complex mechanisms that involve several transcription factors, such as HNF-4␣, COUP-TFI, forkhead factors, the glucocorticoid receptor, C/EBPs, cAMP-response element-binding protein, and AP-1 (32)(33)(34)(35). PGC-1␣, whose hepatic expression is increased upon fasting, contributes to the induction of the gluconeogenic genes PEPCK and glucose-6-phosphatase, via its interactions with HNF-4␣, glucocorticoid receptor, and the forkhead factor FOXO1 (26, 36 -38). Importantly, increased expression of PGC-1␣ in the liver of fed animals is sufficient to inappropriately induce gluconeogenic enzymes and elevate blood glucose levels, a finding that underscores the possible involvement of this coactivator in the elevation of hepatic glucose production in the diabetic state (26).
Whereas a positive role of ERR␣ in PGC-1␣-induced expression of mitochondrial genes has been established, the role of ERR␣ in the regulation of genes involved in hepatic glucose metabolism, including PGC-1␣-stimulated gluconeogenesis, is not clear. ERR␣ is induced in the liver in response to fasting, but has been proposed to either be a repressor of general PGC-1␣ transcriptional activity (12) or to have no effect on gluconeogenic enzymes (17). Notably, the PEPCK gene promoter contains a sequence that resembles an ERRE and that overlaps with a site (glucocorticoid accessory factor 3 (gAF3)) critical for the activation of PEPCK transcription by PGC-1␣ (45). In this study, we address in vitro and in vivo the role of ERR␣ in the regulation of PEPCK gene expression. Our findings demonstrate that ERR␣ inhibits PEPCK gene expression, at least in part by restricting the interaction of PGC-1␣ with the PEPCK promoter. Furthermore, the repressive effect of ERR␣ is specific for gluconeogenic genes, whereas genes involved in mitochondrial respiratory function are activated by hepatic ERR␣. The potential benefits of an ERR␣ role in reducing hepatic glucose production while increasing oxidative capacity are discussed.

MATERIALS AND METHODS
Animals-Generation of ERR␣ null mice in the C57BL/6 background has been described elsewhere (19). Mice were housed at 21°C on a 12-h light-dark cycle and fed ad libitum with a standard diet containing 5% fat (Harland Tekland LM-485, Indianapolis, IN). For gene expression analysis, adult (8 -10 weeks) ERR␣ null females and wild type littermates were fasted for 24 h and then fed for 5 h with standard diet. After the refeeding period, mice were euthanized, and tissues were removed and stored for subsequent analysis. All procedures were performed in accordance with the guidelines for animal care and use of The Scripps Research Institute.
Cell Culture, Transfections, Luciferase Assays, and Adenovirus Infections-H4IIE rat hepatoma cells were maintained as described (42). HepG2 and COS-7 cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. H4IIE cells were transfected with calcium phosphate-precipitated DNA (42). COS-7 and HepG2 cells were transfected using the FuGENE 6 reagent (Roche) according to the manufacturer's instructions. Luciferase reporter activities were determined ϳ18 h after transfections using the Dual Luciferase Reporter Assay System (Promega). For adenovirus infection, cells were infected first with the siERR␣ adenovirus or a control virus in 10-cm dishes and at a multiplicity of infection of 100, and then (3 days later) with GFP or PGC-1␣ expressing viruses and a second dose of the siERR␣ or control viruses (multiplicity of infection 50 each). Cells were harvested 24 or 48 h after the second infection.

ERR␣ Binds to the Regulatory Sequences of the PEPCK Gene-
The sequence of the gAF3 binding site in the PEPCK gene promoter shows a high degree of identity with known response elements for ERR␣ (14,16,46). This putative ERRE is conserved in the mouse, rat, and human PEPCK regulatory sequences (Fig. 1A). To test whether ERR␣ binds to this site, we used gel mobility shift assays. Incubation of nuclear extracts from COS-7 cells expressing ERR␣ with a labeled oligonucleotide containing the gAF3 sequence gave rise to a complex that could be supershifted with an anti-ERR␣ antibody but not with a nonspecific antibody (Fig. 1B). A 50-fold molar excess of an oligonucleotide containing a known ERRE from the TR␣ promoter (TR␣) or the gAF3 sequence (AF3) effectively competed for the formation of the complex. However, oligonucleotides with mutations in the TR␣-ERRE (TR␣M4) or the gAF3-sequence (AF3m␤ and AF3m␥) failed to compete for the formation of the complex, demonstrating that ERR␣ binds specifically the gAF3 site (Fig. 1B). Next, we asked if endogenous ERR␣ in H4IIE rat hepatoma cells recognizes the gAF3 site. Incubation of nuclear extracts of H4IIE cells with the gAF3 probe led to the formation of multiple complexes (Fig. 1C). As shown previously, an antibody against COUP-TFI, which also binds the gAF3 site, resulted in a supershift (39). An antibody specific for ERR␣ also caused a supershift (Fig. 1C), whereas a control antibody had no effect, demonstrating that endogenous ERR␣ binds the gAF3 site of the PEPCK promoter. Finally, we tested whether ERR␣ binds the endogenous PEPCK promoter in its chromatin context, using chromatin immunoprecipitation assays. H4IIE cells were treated with no hormone or dexamethasone for 2 h. Cross-linked and sheared chromatin fragments were precipitated with either no antibody as control, an antibody against the glucocorticoid receptor or the anti-ERR␣ serum. The PEPCK promoter sequences containing the gAF3 site could be amplified from the precipitates with the ERR␣-specific antibody but not from the control sample (Fig. 1D). In contrast to the binding of the glucocorticoid receptor, which requires the presence of hormone, ERR␣ binds to the PEPCK promoter independent of glucocorticoid treatment.
ERR␣ Inhibits the Induction of PEPCK Gene Expression by PGC-1␣-ERR␣ mediates the stimulatory effect of PGC-1␣ on the expression of several genes involved in mitochondrial function (16,17). Because PGC-1␣ also stimulates the expression of PEPCK (25,26), we tested whether ERR␣ could mediate the enhancing effect of PGC-1␣ on the PEPCK gene. We used adenoviral vectors to express ERR␣ and PGC-1␣ in HepG2 hepatoma cells. Expression of ERR␣ alone had no effect on basal PEPCK expression in these cells and PGC-1␣-induced PEPCK expression as expected (26), however, coexpression of ERR␣ with PGC-1␣ repressed the stimulatory effect of PGC-1␣ ( Fig. 2A). The decreased induction of PEPCK expression could not be accounted for by changes in PGC-1␣ expression (Fig. 2B). Next, we used a PEPCK luciferase reporter to test if ERR␣ acts through the promoter sequence present in this construct. Expression of PGC-1␣ in H4IIE hepatoma cells enhanced the expression of a PEPCK luciferase reporter, as shown previously (26,45) (Fig. 2C). Increasing the cellular levels of ERR␣, by transfecting increasing amounts of an ERR␣ expression plasmid, repressed the activity of the PEPCK luciferase reporter in a dose-dependent manner (Fig. 2C). These results indicate that ERR␣ exerts an inhibitory effect on the expression of PEPCK, and show that the regulatory sequence that drives the expression of the PEPCK luciferase reporter is sufficient to mediate both the induction by PGC-1␣ and the inhibition by ERR␣.
A Constitutively Activating Form of ERR␣ Induces PEPCK Gene Expression-The observed inhibition of PEPCK expression by ERR␣ could be due to the ability of ERR␣ to bind the PEPCK promoter in a mode that supports repression of transcription. Alternatively, ERR␣ could be acting indirectly, e.g. by inducing a repressor of PEPCK expression or of PGC-1␣ function. We reasoned that, if ERR␣ acted as a bona fide repressor when bound to the PEPCK promoter, modification of the FIGURE 1. ERR␣ binds to the gAF3 element of the PEPCK regulatory sequences. A, alignment of the consensus ERRE, ERRE-like sequences in the regulatory regions of PEPCK and glycerol kinase (GlyK) genes, and other characterized ERREs (16,46). The position (5Ј end) of the ERREs relative to the transcription start site of the genes is indicated in parentheses. The nucleotides that have been mutated in the ERREs of the PEPCK gAF3 and TR␣ (in AF3 m␥/AF3m␤ and TR␣M4, respectively) are underlined. The PEPCK gene sequences are shown in the reverse-complementary orientation. B, endlabeled PEPCK AF3 oligonucleotides, containing the sequences from nucleotide Ϫ345 to Ϫ304 of the rat Pepck gene, were incubated with nuclear extracts from COS-7 transfected with the ERR␣ expression vector (pSG5/mERR␣), in the presence or absence of antibodies or of 50-fold molar excess of competitor oligonucleotides, as indicated. The resulting DNA-protein complexes were separated from the free probe in a polyacrylamide gel. C, labeled PEPCK AF3 oligonucleotides were incubated with nuclear extracts from H4IIE cells and the indicated antibodies. The DNA-protein complexes were separated by electrophoresis. The complexes supershifted by the antibodies are indicated by the arrow. D, H4IIE cells were treated for 2 h with vehicle (Ϫ) or 100 nM dexamethasone (ϩ), and occupancy of glucocorticoid receptor (GR) and ERR␣ at the PEPCK regulatory regions was determined by chromatin immunoprecipitation. PEPCK sequences harboring the gAF3/ERRE site (nucleotides Ϫ486 to Ϫ268, relative to the transcription start site) were amplified by PCR.
protein so that it functions as a constitutive activator of transcription should alter the regulatory effect on PEPCK. To test this, we expressed a chimeric protein consisting of the ERR␣ receptor and the potent activation domain VP16 (ERR␣-VP16) (14). The expression of ERR␣-VP16 induced endogenous PEPCK gene expression about 6-fold, compared with control GFP (Fig. 3A). These results support the notion that a transcriptional repressor function of wild type ERR␣ is required for the down-regulation of PEPCK expression. Moreover, ERR␣-VP16 activated the PEPCK luciferase reporter gene, suggesting that it acts through the regulatory sequences present in this construct (Fig. 3B). Finally, in this context, we tested the role of the gAF3 site, by introducing two different mutations that prevent ERR␣ binding to the gAF3 site (Fig. 1B). In contrast to the wild type reporter, ERR␣-VP16 did not enhance reporter activity from either one of these mutant reporter constructs. The inability of ERR␣-VP16 to act on PEPCK promoter with the mutated AF3 site supports the notion that ERR␣ regulation of the PEPCK promoter depends on this promoter element.

Suppression of Endogenous ERR␣ Leads to an Enhancement of PEPCK
Expression-To test the role of endogenous ERR␣ on the regulation of the PEPCK gene, we used an adenovirus that expresses a siRNA for ERR␣ and leads to efficient suppression of endogenous ERR␣ expression (Fig. 4A). Expression of PGC-1␣ in HepG2 cells led to the induction of endogenous ERR␣, as has been shown in other cell types (9,17,47), and the siRNA for ERR␣ effectively suppressed this induction (16) (Fig.  4, A and B). Suppression of ERR␣ expression in HepG2 cells had no significant effect on basal PEPCK expression, but led to an increased induction of PEPCK mRNA by PGC-1␣ (Fig. 4C). The increased induction of PEPCK mRNA was not because of changes in PGC-1␣ protein levels, which were not affected by the siRNA for ERR␣ (Fig. 4B), suggesting that ERR␣ antagonizes PGC-1␣ action on the PEPCK gene promoter. These findings demonstrate that endogenous ERR␣ functions as a repressor of PEPCK gene expression. Furthermore, they suggest a possible feedback regulation loop, with PGC-1␣ inducing ERR␣, and ERR␣ antagonizing the action of PGC-1␣.
ERR␣ Antagonizes PGC-1␣ Recruitment to the PEPCK Gene Promoter-To obtain insight into the possible mechanism(s) by which ERR␣ inhibits PEPCK expression, we next used chromatin immunoprecipitation assays to determine the occupancy of ERR␣ and PGC-1␣ at the endogenous PEPCK regulatory site shown in Fig. 1. As expected from the ability of PGC-1␣ to induce ERR␣ expression, we observed that viral introduction of PGC-1␣ led to an increased binding of not only PGC-1␣, as reported previously (45), but also of endogenous ERR␣ to the PEPCK gene promoter (Fig. 5A). Decreasing the amount of ERR␣ by siRNA effectively reduced the presence of ERR␣ at the PEPCK gene promoter, as expected. These results also underscored the specificity of the observed ERR␣ binding in this assay. Strikingly, the decrease in ERR␣ expression also led to an increase in the binding of endogenous PGC-1␣ to the PEPCK gene promoter region (Fig. 5, A and B). These  results are consistent with a model where the presence of ERR␣ antagonizes the recruitment of PGC-1␣ to the PEPCK gene promoter.
Endogenous ERR␣ Is Required for the Induction of Mitochondrial Gene Expression by PGC-1␣ in Hepatocytes-ERR␣ mediates the PGC-1␣-induced expression of genes involved in mitochondrial function in SAOS2 and C2C12 cells (16,17). Next we asked if the repressive effect of ERR␣ on PEPCK gene expression in hepatocytes reflects a distinct function of ERR␣ in different cell types (e.g. liver versus muscle) or at different types of genes (e.g. gluconeogenic versus mitochondrial genes). To address this point we determined the effect that suppression of ERR␣ expression in HepG2 cells had on genes encoding mitochondrial proteins of the citric acid cycle (IDH3A encoding isocitrate dehydrogenase subunit ␣) or with a role in oxidative phosphorylation (CYCS and ATP5B, encoding cytochrome c, somatic, and the ␤ subunit of ATP synthase). As shown in Fig. 6, the ability of PGC-1␣ to induce the expression of these 3 genes depends on the expression of endogenous ERR␣. Suppression of ERR␣ levels led to a diminished induction by PGC-1␣, despite the fact that PGC-1␣ was expressed comparably in the presence and absence of ERR␣ (protein levels for PGC-1␣ and ERR␣ shown in Fig. 4). These findings suggest that in the same cell type, ERR␣ mediates the enhancing effects of PGC-1␣ on one set of genes, while repressing a distinct set of PGC-1␣-regulated genes.

ERR␣ Has Distinct Functions on the Two Different Sets of PGC-1␣regulated Genes in Vivo-
To test the notion that ERR␣ acts as a downstream effector of PGC-1␣ on some targets, such as genes that encode mitochondrial proteins, while repressing other targets, such as genes encoding critical enzymes in hepatic gluconeogenesis, we compared the expression of representative genes in the liver of mice lacking ERR␣ (ERR␣ knock-out) and their wild type littermates. Consistent with the in vitro experiments, the expression of several genes involved in gluconeogenesis, including PEPCK and glycerol kinase, 3 was markedly increased in the liver of fed ERR␣ knock-out mice (Fig. 7A). The expression of glucose-6-phosphatase, another gluconeogenic gene, was also enhanced in ERR␣ knock-out animals, although the increase did not reach statistical significance (data not shown). Interestingly, there was no effect of ERR␣ on the induced levels of these genes in fasted animals, suggesting that the repressive function of ERR␣ is most apparent when PGC-1␣ levels are low 4 (data not shown). In contrast, genes important for oxidative phosphorylation, including Atp5b and cytochrome c-1,

Repression of Gluconeogenic PGC-1␣ Action by ERR␣
were expressed at reduced levels in the liver of ERR␣ knock-out, compared with wild type animals, supporting the positive role of ERR␣ for their expression (Fig. 7B). These results support a view in which ERR␣ represses gluconeogenic gene expression parallel to enhancing mitochondrial gene expression in liver. Moreover, we noticed that ERR␣ null animals also express higher levels of hepatic PGC-1␣ in the fed state (Fig. 7C). These differences in PGC-1␣ expression, which have not been seen in the in vitro systems (Fig. 4 and data not shown), suggest that the repressive effects of ERR␣ on gluconeogenic gene expression in vivo may be exerted at multiple levels, and not only by direct binding to the promoters of PGC-1␣ target genes like PEPCK.

DISCUSSION
Hepatic glucose production is tightly regulated by signals that relay the metabolic state of the organism. Gluconeogenesis is promoted at times of fasting (by glucagon and glucocorticoids), and suppressed during the fed state (by insulin and high levels of glucose). The stimulatory effects of glucagon (mediated by cAMP) and glucocorticoids on the transcription of key genes involved in gluconeogenesis have been studied in great detail (25,29,49). In contrast, much less is known about the mechanisms that repress the expression of gluconeogenic genes in the fed state (50,51). PEPCK, a rate-determining enzyme of gluconeogenesis, is tightly regulated at the level of gene expression by the signals that regulate gluconeogenesis (28). Consequently, PEPCK has served as a valuable model for dissecting the signal transduction pathways and transcription factors that control gluconeogenesis in liver (25)(26)(27).
Here we show evidence of a mechanism involving ERR␣ that suppresses PEPCK gene expression. Our findings suggest that one physiological role of ERR␣ in liver is to limit the expression of the gluconeogenic program in the fed state, possibly by antagonizing the stimulatory effects of PGC-1␣. Several lines of evidence support the repressive function of ERR␣ on PEPCK expression. First, ERR␣ binds to the PEPCK gene promoter. Second, overexpression of ERR␣ represses the induction of PEPCK by PGC-1␣, whereas inhibition of ERR␣ expression results in enhanced PEPCK gene expression. Third, disruption of the ability of ERR␣ to repress, by combining it with a potent viral transcriptional activation domain, converts ERR␣ to an activator of PEPCK expression. Finally, animals that lack ERR␣ show increased PEPCK gene expression in the liver, compared with wild type littermates.
Previous studies have shown that ERR␣ interacts with an inhibitory domain of PGC-1␣ and represses the transcriptional activity of PGC-1␣ in transfection assays (12). ERR␣ was proposed to affect the subnuclear distribution of PGC-1␣, and render PGC-1␣ less accessible to target promoters in general (12). Our findings in this study are consistent with these earlier observations in that ERR␣ indeed antagonizes the PGC-1␣  action at the PEPCK gene promoter. However, in the same cell where ERR␣ inhibits PGC-1␣ action at the PEPCK gene, ERR␣ also mediates the enhancing effects of PGC-1␣ on genes encoding mitochondrial proteins of the oxidative phosphorylation system, such as CYCS and ATP5B. Thus, ERR␣ represses in a promoter-specific manner, and is not a general inhibitor of PGC-1␣ activity.
Consistent with gene-specific repression by ERR␣, we find that ERR␣ binds at the previously characterized gAF3 site of the PEPCK promoter. It is not yet clear how the ERR␣ binding site at the PEPCK gene promoter is distinct from the similar ERR␣ binding sites at the promoters of ATP5B and CYCS (Fig. 1A). Notably, the gAF3 site at the PEPCK gene promoter is recognized by other nuclear receptors as well (e.g. RAR/ RXR, COUP-TFs), and is important for responsiveness to both glucocorticoids and PGC-1␣ (39,45). It is possible that ERR␣ bound at the gAF3 site fails to cooperate properly with adjacent transcription factors, as well as prevents the binding of another factor, perhaps a nuclear receptor, that promotes the formation of a transcriptionally active and PGC-1␣-responsive complex. Alternatively, ERR␣ bound at the gAF3 sequence may actively recruit co-repressors. Furthermore, the finding that mice lacking ERR␣ have increased PGC-1␣ expression suggests that ERR␣ in vivo affects gluconeogenic gene expression by multiple mechanisms, including ones that limit PGC-1␣ expression. Further studies will be required to understand fully the underlying molecular mechanisms by which ERR␣ regulates PEPCK gene expression.
The induction of PEPCK gene expression during fasting is a prerequisite for the activation of hepatic gluconeogenesis and maintenance of circulating glucose levels. Suppression of PEPCK expression in the fed state is equally important for proper glucose homeostasis. Multiple signals repress PEPCK transcription, including insulin and high glucose concentrations, but the mechanisms employed are not well understood. The ability of ERR␣ to repress PEPCK expression manifests in vivo in the fed state as inappropriately elevated hepatic PEPCK mRNA levels in ERR␣ null animals. Future studies will need to address if ERR␣ is part of a pathway that mediates the repressive effects of insulin or high glucose. Notably, despite the elevated expression of several rate-determining gluconeogenic genes, ERR␣ null mice do not have increased plasma glucose levels at the fed state (data not shown). The apparent lack of effect on glucose homeostasis may be explained by a decreased capacity of ERR␣ null hepatocytes for mitochondrial oxidative metabolism and supply of ATP, similarly to what has been proposed for PGC-1␣ null animals that have increased hepatic gluconeogenic gene expression and normal plasma glucose levels in the fed state (22). Moreover, possible decreases in mitochondrial oxidative metabolism in peripheral tissues may result in increased peripheral utilization of glucose and mask increases in hepatic glucose production.
Interestingly, we observed no differences in the induced, fasted levels of PEPCK mRNA in wild type and ERR␣ knock-out animals, even though ERR␣ levels are increased by fasting (12) (and data not shown). It is possible that hormonal signals (e.g. cAMP) regulate the transcriptional activity of ERR␣, and inactivate its repressor function at the PEPCK gene. The high ERR␣ levels that accumulate during fasting may serve to poise the system for fast and efficient repression as soon as food and associated repressive signals arrive. Additionally, it is likely that ERR␣ in the fasted state has other roles in hepatocytes, not addressed here.
Even though ERRs have no natural ligands yet, searches for synthetic ligands have identified inverse agonists, as well as partial agonists, that can modulate ERR signaling pathways (52)(53)(54)(55)(56)(57)(58). At present, it is not clear how such ligands will modulate ERR␣ activity in contexts where ERR␣ represses, such as at the PEPCK gene. Because ERR␣ regulates its own expression in an autoregulatory loop (17,47), putative agonist ERR␣ ligands are likely to increase ERR␣ expression (54), and thus may enhance the native functions of ERR␣, including enhancement of mitochondrial genes and suppression of gluconeogenic gene expression. If so, agonists of ERR␣ may exert beneficial effects in the treatment of diabetes by two mechanisms: (i) increasing mitochondrial function and oxidative capacity (16 -18), and (ii) inhibiting hepatic gluconeogenesis and lowering plasma glucose levels.