Estrogen-related Receptor γ (ERRγ) Is a Novel Transcriptional Regulator of Phosphatidic Acid Phosphatase, LIPIN1, and Inhibits Hepatic Insulin Signaling*

Background: The PAP function of LIPINs is involved in the regulation of intracellular lipid levels and hepatic insulin receptor signaling. Results: ERRγ-mediated induction of LIPIN1 results in the perturbation of hepatic insulin signaling through DAG-mediated activation of PKCϵ. Conclusion: ERRγ is a novel transcriptional regulator of LIPIN1. Significance: An ERRγ inverse agonist could ameliorate LIPIN1-mediated perturbation of hepatic insulin signaling. LIPINs have been reported to perform important roles in the regulation of intracellular lipid levels. Their mutations induce lipodystrophy, myoglobinuria, and inflammatory disorders. Recently, the phosphatidic acid phosphatase function of LIPINs has been associated with the perturbation of hepatic insulin receptor signaling via the diacylglycerol-mediated stimulation of PKCϵ activity. Here, we report that nuclear estrogen-related receptor (ERR) γ is a novel transcriptional regulator of LIPIN1. Overexpression of ERRγ significantly increased LIPIN1 expression in primary hepatocytes, whereas the abolition of ERRγ gene expression attenuated the expression of LIPIN1. Deletion and mutation analyses of the LIPIN1 promoter showed that ERRγ exerts its effect on the transcriptional regulation of LIPIN1 via ERRE1 of the LIPIN1 promoter, as confirmed by ChIP assay. We also determined that the gene transcription of LIPIN1 by ERRγ is controlled by the competition between PGC-1α and small heterodimer partner. Additionally, ERRγ leads to the induction of hepatic LIPIN1 expression and diacylglycerol production in vivo. Finally, an inverse agonist of ERRγ, GSK5182, restores the impaired insulin signaling induced by LIPIN1-mediated PKCϵ activation. Our findings indicate that the selective control of ERRγ transcriptional activity by its specific inverse agonist could provide a novel therapeutic approach to the amelioration of impaired hepatic insulin signaling induced by LIPIN1-mediated PKCϵ activation.

LIPIN1 was originally identified via positional cloning as the mutated gene in fatty liver dystrophy (fld) mice, a mouse model of lipodystrophy (1,2). LIPIN1 was later shown to be involved in lipid and glucose metabolism in a variety of tissues, including adipose tissue, liver, and skeletal muscle (3)(4)(5)(6). LIPIN1 mutations in fld mice lead to significant reductions in body fat, insulin resistance, and increased susceptibility to atherosclerosis (7). Conversely, high levels of LIPIN1 in adipose tissue or skeletal muscle of transgenic mice result in the obesity phenotype (6).
Early biochemical studies have shown that LIPIN1 indicates soluble Mg 2ϩ -dependent phosphatidic acid phosphatase type 1 (PAP1) activity. PAP1 has been shown to perform a key function in the catalysis of PA 2 conversion to DAG, a direct precursor of triglycerides and phospholipids (8,9). It has been reported that lipid infusions designed to increase plasma fatty acid in skeletal muscles and liver can result in the accumulation of fatty acid metabolites, such as DAGs, which function as second messengers for signaling cascades (10). DAG, in particular, has been shown to be a major activator of protein kinase C⑀ (PKC⑀), which accounts for the link between lipid accumulation and insulin signaling in the liver (11,12). Recently, the PAP function of hepatic LIPIN1 has been reported to disrupt insulin receptor signaling via the DAG-mediated induction of PKC⑀ activity in mice (4). In addition, LIPIN1 has been reported to be induced by PGC-1␣, which in turn functions as a transcriptional coactivator of the hepatic PGC-1␣/peroxisome proliferator-activated receptor ␣ regulatory pathway via its direct interaction (3).
The estrogen receptor-related receptor subfamily consists of three members, ERR␣, -␤, and -␥ (NR3B1-3), which bind as either monomers or dimers (13). In adult mice, both ERR␣ and ERR␥ are ubiquitously expressed, whereas ERR␤ expression is restricted to the brain, kidney, and heart (14). The results of structural studies have indicated that the ERRs are constitutively active without endogenous ligands (15,16), whereas several synthetic ligands that could either stimulate or repress the transcriptional activity of the ERRs by recruiting or disrupting the ERRs-coactivator interactions have been also identified (15). The ligand-independent transcriptional activities of ERR␣ and -␥ depend on nuclear receptor coregulators, such as steroid receptor coactivator 2 (SRC-2), PGC-1␣, receptor interacting protein 140 (RIP140), and small heterodimer partner (SHP), all of which are known regulators relevant to various metabolic programs (15,(17)(18)(19)(20)(21)(22)(23)(24)(25)(26). Interestingly, ERRs are regulated by the peripheral circadian clock in key metabolic tissues, such as white and brown adipose tissues, muscle, and liver, thereby suggesting that these receptors may function as molecular links between the circadian oscillator and energy metabolism (27). Recently, ChIP-on-chip and expression analysis studies have demonstrated that ERR␣ and ERR␥ are involved in the regulation of mitochondrial programs in cardiomyocytes (28). In particular, ERR␥ is known to regulate a nuclearly encoded mitochondrial genetic network that coordinates postnatal metabolic transition in the heart (29). However, the role of ERR␥ in the liver remains to be clearly determined.
In this study, we demonstrated that ERR␥ is a previously unrecognized transcriptional regulator of hepatic LIPIN1 that leads to the induction of DAG and PKC⑀ activation, which inhibits insulin signaling. The control of ERR␥ transactivity by an inverse agonist could provide an effective approach to the LIPIN1-mediated regulation of DAG generation, which induces impaired insulin signaling.
Cell Culture and Transient Transfection Assay-Maintenance of HepG2, AML12, and 293T cells was carried out as mentioned previously (34). In brief, the cells were maintained in DMEM (Hyclone) supplemented with 10% fetal bovine serum (FBS, Hyclone) and antibiotics (Hyclone) in a humidified atmosphere containing 5% CO 2 at 37°C. Transient transfections were conducted as described previously (34). In brief, cells were transfected with the indicated reporter plasmids together with expression vectors encoding for various genes using Lipofectamine 2000 TM (Invitrogen), in accordance with the manufacturer's instructions. Total cDNA used for each transfection was adjusted to 1 g/well via the addition of appropriate amounts of empty vector and cytomegalovirus (CMV)-␤-galactosidase plasmids as an internal control. The cells were harvested at 40 -48 h post-transfection for luciferase and ␤-galactosidase assays. The luciferase activity was normalized to ␤-galactosidase activity.
Culture of Primary Hepatocytes-Primary hepatocytes were isolated from Sprague-Dawley rats (male, 8 weeks) via the collagenase perfusion method as described previously (4). Cells were maintained in M199 media (Cellgro) overnight for attachment and adenoviral infection, or chemical treatments were carried out as described in the figure legends.
Animal Experiments-Male 8-week-old C57BL/6J mice (The Jackson Laboratory) were used for this study. All mice were acclimatized to a 12-h light/dark cycle at 22 Ϯ 2°C with free access to food and water in a specific pathogen-free facility. Ad-GFP and Ad-FLAG-ERR␥ were injected into the tail veins of mice, and the mice were sacrificed on day 3 after the injection. All animal experiments were approved and performed by the Institutional Animal Use and Care Committee of the Korea Research Institute of Bioscience and Biotechnology.
RNA Isolation and Analysis-Total RNAs were isolated from cell lines or primary hepatocytes using TRIzol reagent (Invitrogen), in accordance with the manufacturer's instructions, and semi-quantitative and real time quantitative PCR analysis were conducted using primers for ERR␣, ERR␥, LIPIN1, LIPIN2, SHP, PGC-1␣ and PDK4 as described previously (4,30,32). All data were normalized to ribosomal L32 expression.
Measurement of DAGs-DAGs were extracted from cell pellets or liver tissues with chloroform/methanol (2:1, v/v) containing 0.01% butylated hydroxytoluene. A known amount of 1,2-dinonadecanoin (19:0 DAG) was added as an internal standard. After vortexing and centrifugation, the lower organic phase was collected. Samples in organic phase were evaporated and dissolved in hexane/methylene chloride/methyl tert-butyl ether for loading onto the diolbonded solid phase extraction column (Waters, Inc., Milford, MA) under vacuum. DAGs were eluted as described previously by Pacheco et al. (35). Extracted lipids were dried under nitrogen gas and redissolved in methanol. Then the contents of DAGs among extracted lipids were measured via liquid chromatography-tandem mass spectrometry (LC-MS/MS) using a method described previously (36) with some modifications, employing a bench-top tandem mass spectrometer, API 4000 Q-trap (Applied Biosystem, Framingham, MA) interfaced with an atmospheric pressure chemical ionization source and an Agilent series 1200 micro-pump equipped with an autosampler. Five selected DAG molecular species (di-16:0, di-18:1, 16:0 -18:1, 18:0 -18:2, and 18:0 -20:4) were separated by a Gemini C6-phenyl column (50 ϫ 2.0 mm inner diameter, 3 m, Phenomenex, Torrance, CA) and ionized in positive atmospheric pressure chemical ionization mode. The mobile phase was 98% methanol with a flow rate of 0.25 ml/min, and 10 l of sample was analyzed.
Statistical Analysis-Data were expressed as means Ϯ S.E. Statistical analysis was conducted via Student's t test and oneway analysis of variance among groups. Differences were considered statistically significant at p Ͻ 0.05.

ERR␥ Positively Regulates the Gene Expression of LIPIN1 in
Hepatocytes-To assess the possibility that the expression of the LIPIN1 gene is regulated by nuclear receptor ERRs, we carried out transient transfection with vectors expressing ERRs and the mLIPIN1 promoter in 293T and HepG2 cells. Interestingly, ERR␥, but not ERR␣ and ERR␤, significantly increased mLIPIN1 promoter activity in both 293T and HepG2 cells (Fig. 1, A and B). In an effort to further examine the transcriptional regulation of LIPIN1 by ERR␥, we conducted adenoviral overexpression of GFP (Ad-GFP) or ERR␥ (Ad-ERR␥) in a nontransformed mouse liver (AML12) cell line and rat primary hepatocytes. As anticipated, Ad-ERR␥ elicited a marked increase in LIPIN1 mRNA and protein levels ( Fig. 1, C and D). Expression of pyruvate dehydrogenase 4 (PDK4), a known ERR␥ target gene, was also higher in Ad-ERR␥-infected AML12 cells compared with controls. Interestingly, LIPIN2, an isoform of LIPIN, also evidenced a 2-fold induction by Ad-ERR␥ in rat primary hepatocytes. Conversely, the reduction of endogenous ERR␥ via the adenoviral overexpression of short hairpin RNA for ERR␥ (Ad-shERR␥) in rat primary hepatocytes resulted in a considerable reduction of basal LIPIN1 expression (Fig. 1E). However, Ad-ERR␣ had no measurable effect on LIPIN1 expression (Fig. 1F). We conclude that ERR␥ is a transcriptional inducer of hepatic LIPIN1 expression.
LIPIN1 Is a Direct Target of ERR␥-Next, the molecular mechanisms by which ERR␥ regulates LIPIN1 gene transcription were explored. Close investigation of the LIPIN1 promoter revealed two putative ERRE motifs (ERRE1 and ERRE2) conserved in both humans and mice. To determine whether ERR␥ could directly regulate LIPIN1 transcription through these putative ERREs, we generated serial deletion constructs of LIPIN1 promoter containing either ERRE1 or ERRE2 and performed transient transfection in HepG2 cells. As is shown in Fig. 2A, the activity of the ERRE2-deleted LIPIN1 promoter still evidenced a 7-fold induction by cotransfection with ERR␥, whereas that of the ERRE1-deleted LIPIN1 promoter was profoundly reduced. Moreover, the dose-dependent activation of wild-type LIPIN1 promoter by ERR␥ was almost wholly blocked in the ERRE1mutated LIPIN1 promoter (mLIPIN1 ERREmut-luc) (Fig.  2B). Finally, a ChIP assay in AML12 and rat primary hepatocytes demonstrated that ERR␥ was strongly recruited to the ERRE1 region of the LIPIN1 promoter but not to the upstream region (Fig. 2C). These findings strongly indicate that ERR␥ is a direct transcriptional regulator of LIPIN1.
Transcriptional Activity of ERR␥ on LIPIN1 Is Controlled by Coregulators-The transcriptional activity of ERR␥ is known to be regulated by coregulators such as PGC-1␣ and SHP (18,21), suggesting that these coregulators could be involved in the regulation of ERR␥-mediated LIPIN1 expression. As anticipated, PGC-1␣ potentiated and SHP inhibited the ERR␥-mediated induction of LIPIN1 promoter activity in a dose-dependent manner (Fig. 3A). Moreover, SHP significantly attenuated the potentiation of PGC-1␣ for the ERR␥-mediated activation of LIPIN1 promoter. Consistent with the observed promoter activity, the ERR␥-mediated induction of LIPIN1 mRNA was increased further by PGC-1␣ in rat primary hepatocytes, which was largely inhibited by SHP (Fig. 3B). To further evaluate the involvement of these coregulators in ERR␥-mediated LIPIN1 transcription, we conducted a ChIP assay using AML12 cells infected with either Ad-SHP or Ad-PGC-1␣. As anticipated, the recruitment of ERR␥ and PGC-1␣ on the ERRE1 of the LIPIN1 promoter was increased significantly, whereas Ad-SHP reduced the PGC-1␣ occupancy on ERRE1 without changing ERR␥ occupancy (Fig. 3C). We subsequently evaluated the role of ERR␥ in the PGC-1␣-mediated induction of LIPIN1. The ablation of endogenous ERR␥ significantly reduced PGC-1␣induced LIPIN1 mRNA levels in rat primary hepatocytes (Fig.  3D). Additionally, the knockdown of endogenous ERR␥ significantly reduced both basal and PGC-1␣-induced LIPIN1 protein levels in AML12 cells (Fig. 3E). These results demonstrate that the transcriptional regulation of LIPIN1 by ERR␥ depends on the coregulators.
Inverse Agonist of ERR␥ Inhibits LIPIN1-mediated DAG Production-Based on the molecular mechanism by which ERR␥ regulates LIPIN1 gene expression, we also attempted to determine whether ERR␥ results in the induction of hepatic LIPIN1 expression and DAG production in vivo, because LIPIN1 has a PAP1 function to catalyze the conversion of PA to DAG, which is a precursor of triglyceride (8,9). Consistent with the in vitro results, the hepatic overexpression of ERR␥ via the tail vein injection of Ad-ERR␥ resulted in a marked induction of LIPIN1 compared with controls (Fig. 4A). Additionally, hepatic DAG levels were also significantly increased by overexpression (Fig. 4B), thereby suggesting that ERR␥ does indeed regulate hepatic LIPIN1 expression, resulting in DAG production both in vitro and in vivo.
GSK5182, a 4-hydroxytamoxifen analog, is an inverse agonist that selectively inhibits the transactivity of ERR␥ relative to estrogen receptor ␣ due to additional noncovalent interaction with Tyr-326 and Asn-346 (37). Based on the ability of ERR␥ to stimulate LIPIN1 expression, we subsequently evaluated the effects of GSK5182 on the regulation of LIPIN1 by ERR␥ in HepG2 cells. As shown in Fig. 4C, GSK5182 resulted in a marked reduction of ERR␥-induced LIPIN1 promoter activity in a dose-dependent manner. Additionally, the ERR␥-mediated induction of LIPIN1 mRNA was significantly reduced by GSK5182 treatment in rat primary hepatocytes but not that of ERR␥ Y326A, which cannot interact with GSK5182; this suggests that the inverse agonist selectively inhibits the transcriptional activity of ERR␥ (Fig. 4D).
Next, we attempted to determine whether the inverse agonist of ERR␥ could inhibit LIPIN1-mediated DAG production. As shown in Fig. 4E, the adenoviral overexpression of ERR␥ in AML12 cells significantly induced cellular DAG levels as compared with controls, which was blocked by Ad- shLIPIN1, thereby indicating that LIPIN1 is a downstream mediator of ERR␥ in DAG generation. Furthermore, the ERR␥-mediated induction of DAG was largely attenuated by GSK5182 treatment, thus suggesting that the inhibition of transcriptional activity of ERR␥ by GSK5182 could be an attractive means for the suppression of LIPIN1-mediated DAG production.
GSK5182 Restores Impaired Insulin Signaling via LIPIN1mediated PKC⑀ Activation-Abundant evidence indicates that increased production is associated with obesity-related insulin resistance via the activation of PKC in peripheral tissues (10 -12). Recently, it has been reported that the induction of LIPIN1 by TORC2 stimulates DAG production through its PAP activity and disrupts hepatic insulin receptor signaling via the DAG-mediated stimulation of PKC⑀ activity in the livers of mice (4). Therefore, we subsequently attempted to determine whether the ERR␥-mediated induction of LIPIN1 expression interrupts insulin signaling via the stimulation of PKC⑀ activity. AML12 cells were infected with Ad-GFP and Ad-ERR␥ in the absence or presence of GSK5282. As expected, the adenoviral overexpression of ERR␥ significantly increased LIPIN1 expression and the phosphorylation of PKC⑀ (Ser-729) as compared with GFP, which was dramatically reduced by GSK5182 treatment (Fig.  5, A and B). Considering the inhibitory effect of GSK5182 on PKC⑀ activation, we next attempted to determine whether the inverse agonist of ERR␥ restores the inhibition of insulin-mediated AKT phosphorylation (Ser-473) via the ERR␥dependent induction of LIPIN1 expression and PKC⑀ activation. Surprisingly, the insulin-mediated phosphorylation of AKT was significantly diminished by the ERR␥-mediated increase of LIPIN1 expression and PKC⑀ phosphorylation, which was completely restored by GSK5182 treatment (Fig.  5B). These results indicate that the inactivation of ERR␥ by its specific inverse agonist could ameliorate the LIPIN1-mediated dysregulation of hepatic insulin receptor signaling.

DISCUSSION
In this study, we demonstrated that the nuclear receptor ERR␥ is a previously unrecognized transcriptional regulator of hepatic LIPIN1, which is known to have a PAP function, catalyzing the conversion of PA to DAG. PA and DAG, major precursors of phospholipid biosynthesis, perform a pivotal role in cellular signaling cascades, energy storage, and lipid biosynthetic pathways. Recently, DAG, as a second messenger for signaling cascades, has been shown to activate PKC and PKC⑀ in muscle and liver, respectively, thereby inhibiting insulin receptor signaling, indicating that DAG functions as a link between lipid accumulation and insulin signaling (11,12). Consistent with these results, it has also been reported that the induction of LIPIN1 by TORC2 results in the perturbation of hepatic insulin receptor signaling via the induction of DAG and PKC⑀ activity in the livers of mice, thereby suggesting that the control of hepatic LIPIN1 expression is important for the amelioration of impaired insulin receptor signaling via DAG-mediated PKC⑀ activation (4). Therefore, based on our findings, the inhibition of transcriptional activity of ERR␥ by its inverse agonist could constitute an attractive means for ameliorating the impaired insulin receptor signaling induced by LIPIN1-mediated DAG production.
In addition to the cytosolic function of LIPIN1 as a PAP, it evidences a putative nuclear function as a transcriptional coactivator. Finck et al. (3) have demonstrated that hepatic LIPIN1 expression is induced by PGC-1␣ under fasting conditions and functions as an amplifier of nuclear receptors peroxisome proliferator-activated receptor ␣ and PGC-1␣ in a complex that regulates fatty acid oxidation gene expression. However, the molecular mechanism by which PGC-1␣ induces LIPIN1 expression remains unclear. Here, we have demonstrated that the transcriptional regulation of PGC-1␣ for LIPIN1 is achieved through the ERR␥-binding site on the LIPIN1 promoter (Fig. 4C) and that the induction of LIPIN1 by PGC-1␣ was almost completely blocked by the knockdown of ERR␥ (Fig. 4, D and E), thereby suggesting that ERR␥ is a key transcriptional mediator of the PGC-1␣-mediated induction of LIPIN1.
However, unlike the function of LIPIN1 as a coactivator in the nucleus, our data indicate that hepatic LIPIN1 has a cytosolic function, activating DAG production. These results are consistent with previous findings showing that hepatic LIPIN1 functions as a cytosolic PAP in the regulation of the insulin signaling cascade (4). Recent studies also demonstrated that hepatic LIPIN1 is associated with triglyceride synthesis and secretion, although the results between in vitro and in vivo conditions are somewhat inconsistent (38,39), thereby suggesting that the role of hepatic LIPIN1 appears to involve a cytosolic function related to de novo lipogenesis rather than a nuclear function related to fatty acid oxidation. However, functional differences in hepatic LIPIN1 according to its localization will require further characterization.
Studies of the molecular mechanism underlying the transcriptional regulation of LIPIN1 expression have progressed greatly in recent years. It is known to be induced by glucocorticoids in adipose tissues and the liver and by ␤-adrenergic signaling in skeletal muscle through the glucocorticoid receptor and the orphan nuclear receptor NOR-1 (40,41). It has also been reported that LIPIN1 is a target of SREBP-1 (sterol regulatory element-binding protein 1) and nuclear factor Y in hepatoblastoma cells (42). While this study was being carried out, one report was published regarding the role of LIPIN1 in cardiac myocytes. In a study conducted using cultured ventricular myocytes and cardiac tissue, LIPIN1 gene expression and its PAP activity were shown to be up-regulated by the PGC-1␣-dependent activation of ERR␣ or ERR␥ via the ERR-response elements within the first intron of the LIPIN1 gene (43). These results are similar to those of our study conducted in hepatocytes, with the exception of the ERR␥-specific regulation of hepatic LIPIN1 transcription among the ERR subfamily via ERRE1 in the LIPIN1 promoter. However, the SHP promoter is activated by ERR␥ but not ERR␣ or ERR␤ (18), and Dufour et al. (28) demonstrated that ERR␣ and ERR␥ could regulate the same direct target genes, thereby suggesting the complexity of their function in terms of the transcriptional output of each ERR.
In this study, we have revealed a molecular mechanism by which the nuclear receptor ERR␥ regulates hepatic LIPIN1 expression, resulting in the production of DAG and the activation of PKC⑀ (Fig. 5C). Interestingly, the regulation of hepatic LIPIN1 by ERRs is ERR␥-specific. The overexpression of ERR␥ resulted in the induction of LIPIN1 expression and promoter activity in hepatocytes. We also demonstrated that the ERR␥mediated transcriptional regulation of LIPIN1 is achieved via direct binding of the LIPIN1 promoter; competition between PGC-1␣ and SHP is also involved in the regulation of LIPIN1 by ERR␥. Additionally, we determined that GSK5182, the inverse agonist of ERR␥, resulted in a marked reduction of ERR␥-induced LIPIN1 expression and DAG production via the inhibition of ERR␥ transactivity. Finally, we have demonstrated here that the inverse agonist attenuates LIPIN1-mediated DAG production, leading to the inhibition of insulin receptor signaling, and it also restores the PKC⑀-mediated perturbation of insulin signaling. These findings indicate that the control of ERR␥ by its inverse agonist might provide an attractive means for the regulation of LIPIN1-mediated DAG generation and PKC⑀ activation, thereby disrupting hepatic insulin receptor signaling.