Peroxisome Proliferator-activated Receptor γ Regulates Expression of the Anti-lipolytic G-protein-coupled Receptor 81 (GPR81/Gpr81)*

The ligand-inducible nuclear receptor peroxisome proliferator-activated receptor γ (PPARγ) plays a key role in the differentiation, maintenance, and function of adipocytes and is the molecular target for the insulin-sensitizing thiazoledinediones (TZDs). Although a number of PPARγ target genes that may contribute to the reduction of circulating free fatty acids after TZD treatment have been identified, the relevant PPARγ target genes that may exert the anti-lipolytic effect of TZDs are unknown. Here we identified the anti-lipolytic human G-protein-coupled receptor 81 (GPR81), GPR109A, and the (human-specific) GPR109B genes as well as the mouse Gpr81 and Gpr109A genes as novel TZD-induced genes in mature adipocytes. GPR81/Gpr81 is a direct PPARγ target gene, because mRNA expression of GPR81/Gpr81 (and GPR109A/Gpr109A) increased in mature human and murine adipocytes as well as in vivo in epididymal fat pads of mice upon rosiglitazone stimulation, whereas small interfering RNA-mediated knockdown of PPARγ in differentiated 3T3-L1 adipocytes showed a significant decrease in Gpr81 protein expression. In addition, chromatin immunoprecipitation sequencing analysis in differentiated 3T3-L1 cells revealed a conserved PPAR:retinoid X receptor-binding site in the proximal promoter of the Gpr81 gene, which was proven to be functional by electromobility shift assay and reporter assays. Importantly, small interfering RNA-mediated knockdown of Gpr81 partly reversed the inhibitory effect of TZDs on lipolysis in 3T3-L1 adipocytes. The coordinated PPARγ-mediated regulation of the GPR81/Gpr81 and GPR109A/Gpr109A genes (and GPR109B in humans) presents a novel mechanism by which TZDs may reduce circulating free fatty acid levels and perhaps ameliorate insulin resistance in obese patients.

The ligand-inducible nuclear receptor peroxisome proliferator-activated receptor ␥ (PPAR␥) plays a key role in the differentiation, maintenance, and function of adipocytes and is the molecular target for the insulin-sensitizing thiazoledinediones (TZDs). Although a number of PPAR␥ target genes that may contribute to the reduction of circulating free fatty acids after TZD treatment have been identified, the relevant PPAR␥ target genes that may exert the anti-lipolytic effect of TZDs are unknown. Here we identified the anti-lipolytic human G-protein-coupled receptor 81 (GPR81), GPR109A, and the (humanspecific) GPR109B genes as well as the mouse Gpr81 and Gpr109A genes as novel TZD-induced genes in mature adipocytes. GPR81/Gpr81 is a direct PPAR␥ target gene, because mRNA expression of GPR81/Gpr81 (and GPR109A/Gpr109A) increased in mature human and murine adipocytes as well as in vivo in epididymal fat pads of mice upon rosiglitazone stimulation, whereas small interfering RNA-mediated knockdown of PPAR␥ in differentiated 3T3-L1 adipocytes showed a significant decrease in Gpr81 protein expression. In addition, chromatin immunoprecipitation sequencing analysis in differentiated 3T3-L1 cells revealed a conserved PPAR:retinoid X receptorbinding site in the proximal promoter of the Gpr81 gene, which was proven to be functional by electromobility shift assay and reporter assays. Importantly, small interfering RNA-mediated knockdown of Gpr81 partly reversed the inhibitory effect of TZDs on lipolysis in 3T3-L1 adipocytes. The coordinated PPAR␥-mediated regulation of the GPR81/Gpr81 and GPR109A/Gpr109A genes (and GPR109B in humans) presents a novel mechanism by which TZDs may reduce circulating free fatty acid levels and perhaps ameliorate insulin resistance in obese patients.
Because of a high calorie diet and a sedentary lifestyle, obesity and its associated co-morbidities like hypertension, type II diabetes, and atherosclerosis rapidly increase worldwide (1). Adipose tissue is the major site of lipid storage in the body and plays a pivotal role in the regulation of whole body metabolic homeostasis and therefore in the pathophysiology of obesity (2). After a meal, excess fuel substrates are partitioned to adipose tissue where they are processed and stored as triglycerides (TAG). 2 Conversely, during fasting TAGs are hydrolyzed to free fatty acids (FFA) and glycerol, and the FFA released into the bloodstream can subsequently be used by other organs as energy substrates. The latter process, termed lipolysis, is tightly regulated by hormones and cytokines (3). The three main hormones that regulate lipolysis in humans are insulin, which inhibits lipolysis, and catecholamines (adrenaline and noradrenaline) and glucagon, which stimulate lipolysis. In rodents, inhibition of lipolysis by adenosine presents an additional regulatory pathway. Lipolysis is deregulated in obesity; basal lipolysis rates are increased (4), whereas the stimulation of lipolysis by catecholamines (5) as well as the anti-lipolytic action of insulin (6) are inhibited. The impairment of hormonal control of lipolysis may be due to high levels of tumor necrosis factor-␣, which is overproduced by adipose tissue in obese humans and rodents (7). Deregulated lipolysis results in increased circulating FFA levels and lipid accumulation in nonadipose tissues, * This work was supported in part by a short term fellowship of the Federation ultimately contributing to insulin resistance and other obesityrelated metabolic disorders (8).
One of the key regulators of adipocyte differentiation, maintenance, and function is peroxisome proliferator-activated receptor ␥ (PPAR␥), a member of the nuclear hormone receptor superfamily of ligand-inducible transcription factors (9). PPAR␥ exists in two isoforms, PPAR␥1 and PPAR␥2. PPAR␥2 has an additional 30 amino acids at the N terminus, and its expression is restricted to adipose tissue, while PPAR␥1 is more widely distributed (e.g. adipocytes, lower intestine, monocytes, and macrophages). In vitro and in vivo studies showed that PPAR␥ is both necessary and sufficient to induce adipogenesis (9). PPAR␥ bind as an obligate heterodimer with the retinoic acid X receptors (RXRs) to PPAR-responsive elements (PPREs), which consist of two direct repeats of six nucleotides (AGGTCA) interspaced by one nucleotide (DR-1). Upon binding of ligand these proteins undergo a conformational change, which allows the interaction with so-called coactivators, starting a cascade of protein interactions and modifications that finally results in the induction of specific target genes (10). Although the endogenous ligands for PPAR␥ have not been firmly established, natural compounds like polyunsaturated fatty acids and eicosanoids have been shown to activate PPAR␥. In addition, the antidiabetic drugs, such as thiazolidinediones (TZDs) act as high affinity PPAR␥ ligands (11). Administration of these TZDs to obese and/or insulin-resistant patients has been shown to reduce circulating FFAs and thereby improve insulin sensitivity. Part of these effects may be explained by the stimulatory effect of TZDs on adipocyte differentiation, thereby increasing lipid storage capacity in adipose tissue. In addition, PPAR␥ also regulates a number of genes essential for the adipocytic phenotype, such as genes involved in lipid uptake, lipid synthesis, lipid droplet stabilization, glycerol/FA recycling, and FA oxidation (12). Because elevated levels of serum FFAs promote insulin resistance (13), an important potential mechanism for the beneficial effects of TZDs is therefore the net partitioning of lipids in adipose tissue. Consistent with this notion, genes encoding proteins involved in lipid uptake in adipocytes, such as lipoprotein lipase, CD36, and the oxidized LDL receptor have been reported to be directly regulated by PPAR␥ (9). In addition, PPAR␥ directly regulates the expression of genes directly involved in lipid storage, like the lipiddroplet proteins perilipin and S3-12 (14). PPAR␥ also regulates genes (potentially) involved in the "futile cycle" (9,15,16), the re-esterification of fatty acids and glycerol to triglycerides. Several findings suggest that PPAR␥ and TZDs may also be implicated in the regulation of lipolysis. First, the TZD troglitazone has been shown to lower basal lipolysis rates in differentiated adipocytes (this study and see Refs. (17)(18)(19)) as well as tumor necrosis factor-␣-activated lipolysis (20,21). Second, introduction of a dominant-negative form of PPAR␥ in mature adipocytes resulted in increased lipolysis, suggesting that PPAR␥ normally inhibits this process (22). Finally, treatment of diabetic patients with TZDs has been shown to restore insulin-mediated suppression of lipolysis (23)(24)(25)(26). However, the relevant PPAR␥ target genes that may exert the anti-lipolytic effect of TZDs are unknown.
To identify novel target genes that may play a role in the effects of TZDs on lipid metabolism, we performed a transcriptome analysis in human adipocytes treated with the TZD rosiglitazone. In this study we show that TZDs induce the expression of two anti-lipolytic G-protein-coupled receptors, GPR81/ Gpr81 and GPR109A/Gpr109A, in human and murine adipocytes. In addition, a third anti-lipolytic GPR, the humanspecific GPR109B, is also induced by rosiglitazone. This PPAR␥-mediated activation may occur through a conserved PPRE located in the GPR81 promoter. The coordinated PPAR␥-mediated regulation of the GPR81/Gpr81 and GPR109A/Gpr109A genes (and GPR109B in humans) presents a novel mechanism by which TZDs may reduce circulating FFA levels and perhaps ameliorate insulin resistance in obese patients.
Plasmids-All recombinant DNA work was performed according to standard procedures (27). The murine Gpr81 reporter, Gpr81(Ϫ1059/ϩ28)-luc, was generated by inserting the respective promoter into the XhoI/HindIII site of the PGL3-basic vector (Promega). All mutations were generated by QuickChange mutagenesis (Stratagene) and verified by sequencing. All other plasmids have been described earlier (28).
Reporter assays were performed exactly as described (28).
Microarray Analysis-3T3-L1, SGBS, and hMADs were differentiated as described above and at days 6, 8, and 17, respectively, treated with rosiglitazone or DMSO for 6 h. Micro-array experiments were performed as described before (31). In short, total RNA was isolated using TRIzol reagent. Concentrations and purity were determined on a NanoDrop ND-1000 spectrophotometer (Isogen). RNA integrity was checked on an Agilent 2100 Bioanalyzer (Agilent Technologies) with 6000 Nano-Chips. Part of the RNA samples from four 6-cm dishes was used for quantitative RT-PCR (see under "RNA Isolation and Real Time PCR"). Remaining RNA samples from four 6-cm dishes were pooled and used for microarray analysis. Samples were hybridized on human NUGO arrays from Affymetrix. A detailed description of the analysis method is available on request.
Animal Study-Animal study was performed as described earlier (31). In short, Sv129 male mice were purchased at The Jackson Laboratory (Bar Harbor, ME). At 20 weeks of age, the diet of half of the mice group was supplemented with rosiglita-zone (0.01% w/w) for a week. At the end of the experiment epididymal white adipose tissue was dissected, weighed, and used for RNA isolation. The animal experiments were approved by the animal experimentation committee of Wageningen University.
RNA Isolation and Real Time PCR-3T3-L1 fibroblasts were differentiated as described above. Three independent samples of total RNA were extracted at different time points using TRIzol reagent (Invitrogen). cDNA was synthesized using the superscript first strand synthesis system (Invitrogen) according to manufacturer's protocol. Gene expression levels were determined by quantitative real time PCR with the MyIq cycler (Bio-Rad) using SYBR-green (Bio-Rad) and normalized to TFIIb expression.

RESULTS
GPR109A, GPR109B, and GPR81 Are Regulated by Rosiglitazone in Mature Adipocytes-To identify novel TZD-regulated genes in mature human adipocytes, we performed transcriptome analysis in differentiated hMADs (30). Out of 361 genes that were up-regulated after 6 h of treatment with the TZD rosiglitazone (data not shown), we selected the human-specific G-protein-coupled receptor 109B (GPR109B) to explore in more detail. Together with GPR109A and GPR81, GPR109B belongs to the class A rhodopsin-like G-protein-coupled receptors. GPR109A (also called puma-g) and the human-specific GPR109B are 96% homologous (33) and expressed in adipose tissue, spleen, and immune cells (34 -36), whereas GPR81 expression is almost exclusively restricted to adipose tissue (37,38). GPR109A has been identified as the receptor for the antilipolytic drug nicotinic acid, and in GPR109A knock-out mice it has been shown that GPR109A was the receptor mediating the lipid-lowering effects of nicotinic acid (34 -36). Recently, the ketone body ␤-hydroxybutyrate was reported as an endogenous agonist for GPR109A (39), whereas aromatic D-amino acids can activate GPR109B (40). In addition, very recently two reports (41,42) showed that GPR81 functions as a receptor for lactate, which reduces lipolysis in vitro and in vivo (43,44). Interestingly, the GPR81, GPR109A, and GPR109B genes are colocalized on human chromosome 12 and share synteny with murine Gpr81 and Gpr109A on mouse chromosome 5 (Fig. 1A) (45). For this reason, expression of the GPR109A and GPR81 genes, which were not represented on the microarray, was determined together with the GPR109B gene in differentiated hMADs cells. Using quantitative RT-PCR, mRNA expression of these three genes was found to increase 4 -5-fold after treat-ment with rosiglitazone for 6 h (Fig. 1B). The same experiment was performed in another human adipocyte cell line, the SGBS cell line (29). In these cells a similar mRNA expression profile was observed (Fig. 1C). To investigate whether conserved mechanisms of regulation exist in mouse adipocytes, we examined the effect of rosiglitazone treatment on Gpr81 and Gpr109A mRNA expression in differentiated 3T3-L1 adipocytes. As was observed for the human adipocytes, treatment of murine adipocytes with rosiglitazone stimulated the mRNA expression levels of Gpr81 and Gpr109A up to 4-fold (Fig. 1D). Finally, we examined the effect of rosiglitazone treatment on the mRNA expression of Gpr81 and Gpr109A in mouse adipose tissue in vivo. For this Sv129 male mice received a diet supplemented with rosiglitazone for 1 week; RNA was isolated from epididymal fat pads and subjected to quantitative RT-PCR analysis. As shown in Fig. 1E, mRNA expression of both Gpr109A and Gpr81 was up-regulated in epididymal fat pads of the mice administered rosiglitazone compared with control mice.
In summary, rosiglitazone treatment induces the mRNA expression of the human GPR109A, GPR109B, and GPR81 genes in differentiated adipocytes. A conserved regulatory mechanism may underlie this induction, because the mouse Gpr109A and Gpr81 genes were also significantly up-regulated, in vitro and in vivo, upon rosiglitazone treatment.
GPR81 and GPR109A mRNA and Protein Expression Increase during 3T3-L1 Differentiation-Because PPAR␥ plays an essential role in adipogenesis and the expression of known PPAR␥ target genes increases during differentiation, we studied the protein and mRNA expression levels of Gpr81 and Gpr109A during differentiation of 3T3-L1 pre-adipocytes into mature adipocytes. At different time points of differentiation, cells were harvested to determine mRNA and protein expression levels. Quantitative RT-PCR showed that mRNA expression of both Gpr81 and Gpr109A steadily increased during adipocyte differentiation, starting at day 2 (Fig. 2). In addition, protein expression levels of GPR81 and GPR109A were determined during adipogenesis. Protein expression of GPR81 could be detected from day 4 onward, whereas expression of the Gpr109A protein was observed slightly earlier (day 3). As a control, protein expression levels of PPAR␥ and the well established PPAR␥ target gene Fabp4 were determined and showed a similar increase during adipocyte differentiation (Fig. 2). In conclusion, Gpr81 and Gpr109A mRNA and protein expression clearly increased during differentiation of 3T3-L1 adipocytes.
PPAR␥ Directly Regulates GPR81 Protein Expression in Mature Adipocytes-To investigate whether the activation of the Gpr81 and Gpr109A genes by rosiglitazone is directly reg-  ulated by PPAR␥, we reduced expression of the PPAR␥ protein in mature 3T3-L1 adipocytes by siRNA-mediated knockdown. As described earlier (46), rosiglitazone treatment resulted in reduced PPAR␥ protein levels (Fig. 3, lanes 1 and 2). Electroporation of siRNA oligonucleotides directed against PPAR␥ significantly reduced, but did not completely abolish, PPAR␥ protein expression in the absence and presence of rosiglitazone (Fig. 3). In agreement with the mRNA expression data (Fig. 1B), protein expression of Gpr81 increased 3-fold in the presence of rosiglitazone in these cells (Fig. 3A, lanes 1 and 2). Knockdown of PPAR␥ resulted in a significant reduction (4-fold, as determined by densitometry) of both basal and rosiglitazone-induced Gpr81 protein expression compared with 3T3-L1 adipocytes treated with nontargeting siRNA oligonucleotides. In the case of Gpr109A protein, a 1.5-fold induction was observed upon rosiglitazone treatment, which was reduced 2-fold upon knockdown of the PPAR␥ protein under both basal and treated conditions. In addition, the ability of rosiglitazone to induce Gpr81 protein expression was examined in the presence of the PPAR␥-specific antagonist GW9662. As depicted in Fig. 3B, treatment with GW9662 inhibited the induction of Gpr81 by rosiglitazone. Taken together, these results indicate that the PPAR␥ protein is essential for the activation of the Gpr81 gene, and to a lesser extent the Gpr109A gene, by rosiglitazone.
Endogenous PPAR␥ and RXR Bind to the Proximal Promoter of GPR81-The rapid activation of the Gpr81 and Gpr109A genes by rosiglitazone (Fig. 1) together with the essential role of the PPAR␥ protein in this process (Fig. 3) prompted us to examine whether PPAR␥ and its heterodimeric partner RXR are recruited to the proximal promoter of the Gpr81 and/or Gpr109A genes. Very recently, a genome-wide analysis of PPAR␥ and RXR binding during 3T3-L1 differentiation by ChIP sequencing technology was reported (12). Detailed analysis of the chromosomal region surrounding the Gpr81 and Gpr109A genes revealed clear PPAR␥ and RXR binding in the proximal promoter (Ϫ294/Ϫ55) of Gpr81, suggesting a PPAR␥: RXR-binding site at this location. Interestingly, no significant peaks in close proximity of the Gpr109A gene were observed (Fig. 4A). The recruitment of PPAR␥ and RXR to the proximal promoter of the Gpr81 gene in mature 3T3-L1 adipocytes (day 6) was confirmed by ChIP-PCR (Fig. 4B). In addition, the recruitment in preadipocytes (day 0), in which PPAR␥ expression is low, was negligible, and neither PPAR␥ nor RXR was detected on an arbitrary gene region of the ␤-globin gene on chromosome 7, which served as a negative control (Fig. 4B).
Because the proximal promoter region of the Gpr81 gene is well conserved between human and mouse, we investigated if PPAR␥/RXR also binds to the proximal promoter of the human GPR81 gene in SGBS preadipocytes (day 0) and mature SGBS adipocytes (day 8). Interestingly, both PPAR␥ and RXR were recruited to the proximal promoter of GPR81 in differentiated SGBS cells but not in preadipocytes (day 0) (Fig. 4C). Also, in this case binding of PPAR␥ and RXR was absent in the negative control (Fig. 4C, human beta-globin). Taken together these results indicate that a PPAR␥/RXR heterodimer binds to the proximal promoter of the mouse Gpr81 gene as well as the human GPR81 gene.
Identification of a Functional PPRE in the Proximal Promoter of GprR81-To identify the PPRE in the proximal promoter of Gpr81, we subjected the sequence underneath the peaks (Ϫ294/Ϫ55) of the ChIP-seq data (Fig. 4A) to an in silico promoter analysis (Nuclear Hormone Receptor scan (47)). A potential PPRE was detected, which was conserved between humans and mice (Fig. 5A). To assess if PPAR␥/RXR␣ binds to this potential PPRE, we first performed electrophoretic mobility shift assays. A 32 P-labeled probe containing the PPRE was incubated with in vitro translated PPAR␥2 and/or RXR␣. As shown in Fig. 5B, a specific PPAR␥2-RXR␣ heterodimeric complex was formed on the Gpr81 PPRE, as formation of this protein-DNA complex could be diminished by addition of an excess of unlabeled wild type PPRE but not by an excess of mutant PPRE (Fig. 5B). Specific antibodies against PPAR␥ and RXR␣, but not an antibody directed against an irrelevant protein (Gal4), supershifted the protein-DNA complex, confirming the heterodimeric composition of the complex (Fig. 5C).
Next, we fused the 5Ј-flanking region and start site of the mouse Gpr81 gene (Ϫ1059/ϩ28) to a luciferase gene. The activity of this reporter was determined in human osteosarcoma (U2OS) cells, which express negligible levels of endogenous PPAR␥ protein but display a robust response upon over- FIGURE 3. PPAR␥ directly regulates GPR81 and GPR109A protein expression. A, differentiated 3T3-L1 adipocytes (day 6) were electroporated with control or PPAR␥ siRNA. Twenty hours after electroporation medium was replaced by medium with or without 1 M rosiglitazone and incubated for an additional 28 h. Cells were lysed and subjected to Western blot analysis. ␣-Tubulin was used as a loading control. This blot is a representative of at least three independently performed experiments. B, differentiated 3T3-L1 adipocytes (day 6) were treated with or without 1 M rosiglitazone and with or without 1 M GW9662 for 24 h. Western blot analysis was performed as described under A. expression of the protein (28). As shown in Fig. 6A, transfection of cells with an expression vector encoding PPAR␥2 markedly activated the reporter gene compared with empty vector control (pCDNA). Activation of PPAR␥ by rosiglitazone further enhanced the PPAR␥-mediated activation of the reporter. Mutation of the potential PPRE completely abolished the PPAR␥-mediated activation of the reporter, both in the absence and presence of rosiglitazone (Fig. 6A).
To examine the regulation of the Gpr81 promoter in more detail, we tested the ability of the PPAR␥1 isoform, as well as three PPAR␥2 mutants to activate this reporter. As shown in Fig. 6B, PPAR␥1 activated the reporter to a comparable level as PPAR␥2, suggesting there is no isoform specificity for this PPRE. The two natural PPAR␥2 mutants R425C and P495L displayed reduced and negligible activity, respectively, in agreement with their activity on other promoters (28,48). The heterodimerization defective mutant L464R failed to activate the reporter (Fig. 6B), confirming that dimerization of PPAR␥ and RXR␣ is a prerequisite for binding to the Gpr81 PPRE (Fig. 5, B and C). Taken together, these results indicate that PPAR␥ activates transcription of the Gpr81 gene by binding of a PPAR␥/RXR heterodimer to a conserved PPRE located in the proximal promoter (Ϫ141/Ϫ129) of the gene.
siRNA-mediated Knockdown of Both Gpr81 and PPAR␥ in Mature Adipocytes Increased Lipolysis-Next, we wished to establish the relevance of the PPAR␥-mediated upregulation of the Gpr81 gene in the anti-lipolytic action of TZDs. For this, we reduced Gpr81 or PPAR␥ protein expression by siRNA-mediated knockdown in mature 3T3-L1 adipocytes and determined glycerol levels as a measure for lipolyis. In agreement with previous studies (17)(18)(19), TZD treatment decreased glycerol levels by 2-fold (Fig. 7B). TZD treatment also inhibited glycerol release when lipolysis was stimulated with the ␤-adrenergic agonist isoprotenerol (supplemental Fig. 1). Partial knockdown of PPAR␥ increased lipolysis and reduced the effect of rosiglitazone treatment (35% reduction; Fig. 7B, right panel). Similarly, knockdown of Gpr81, which was also partial (Fig. 7A), resulted in a slight increase in glycerol levels and a 35% reduction of the rosiglitazone-mediated inhibition (Fig. 7B, right panel). Taken together, these data suggest that the anti-lipolytic action of rosiglitazone is partly mediated through the PPAR␥-mediated regulation of the Gpr81 gene in mature adipocytes.

DISCUSSION
PPAR␥ plays a key role in (pre)adipocyte biology by regulating their differentiation, maintenance, and lipid metabolism. The insulin-sensitizing TZDs have been shown to be high affinity ligands for PPAR␥ and are administered to patients with insulin resistance. This class of antidiabetic drugs increases systemic insulin sensitivity in diabetic animal models and humans (49). The number of target genes that help to explain the beneficial effects of these ligands is limited, however. Here we present compelling evidence that the anti-lipolytic G-protein-coupled receptor 81 (GPR81/Gpr81) is a novel direct PPAR␥ target gene in human and murine adipocytes. Interestingly, in addition to the Gpr81/GPR81 gene, expression of the anti-lipolytic Gpr109A/GPR109A gene (and the GPR109B gene in human adipocytes) was also stimulated by TZD-activated PPAR␥, but a functional PPRE could only be identified in the proximal promoter of the Gpr81 gene ( Fig. 1A and Fig. 4A). It is possible that this site also controls the Gpr109A/GPR109A promoter, which is 16 kb further downstream. Of note, the genome-wide profiles of PPAR␥:RXR in 3T3-L1 have unequivocally shown that only a very small percentage of PPAR␥:RXR target sites lie in the proximal promoters (12,50). Distal gene regulation has been proposed to occur via a mechanism by which a transcription factor bound to a distal site directs looping, thereby bringing coactivators and chromatin remodelers at the distal sites in proximity of transcription start site of target genes and facilitating recruitment of RNA polymerase II (51). The rapid increase (within 6 h) and synchronous induction of GPR81/Gpr81 and GPR109A/ Gpr109A in mature adipocytes observed in our studies support the direct regulation of both genes by PPAR␥. Alternatively, PPAR␥ could regulate the GPR109A/Gpr109A gene (and the GPR109B gene in humans) in an indirect manner. Future studies are therefore required to establish the exact molecular mechanisms underlying the regulation of the GPR109A/ Gpr109A gene.
Based on the data presented here showing that PPAR␥ directly regulates the anti-lipolytic GPR81, GPR109A, GPR109B (this study), and GPR43 genes (52,53) in adipocytes together with previous reports showing that TZDs reduce lipolysis in these cells (17)(18)(19)(20)(21), we propose the following model for the anti-lipolytic effect of TZDs in adipocytes. Administration of TZDs to mature adipocytes activates PPAR␥ leading to increased transcription of GPR81 and GPR109A (and GPR109B in humans) and subsequent increase in protein expression of both receptors. GPR109A and GPR81 couple to members of the G i family of G-proteins (38,45). In adipocytes activation of G i preferentially results in the inhibition of adenylyl cyclases, which counteracts the activity of pro-lipolytic receptors (e.g. ␤-adrenergic and glucagon receptors). As a result, intracellular cAMP levels will be lowered and protein kinase A will be less active. Protein kinase A phosphorylates a number of proteins, most notably hormone-sensitive lipase and perilipin, which are required for hydrolysis of TAGs. Phosphorylation of perilipin allows access to the TAG-containing droplets by the now activated hormone-sensitive lipase and a second lipase, adipose triglyceride lipase, which hydrolyzes the TAG in FFA and glycerol. Our data in mature 3T3-L1 adipocytes suggest that the PPAR␥-mediated up-regulation of Gpr81 and Gpr109A con- Electrophoretic mobility shift assay using a 32 P-labeled fragment from the proximal Gpr81 promoter containing the PPRE incubated with in vitro translated PPAR-␥2 and/or RXR␣ proteins in the presence or absence of an excess of unlabeled wild type (wt) or mutant (mut) fragment (B) or ␣-Gal4, ␣-PPAR␥, or ␣-RXR␣ antibody, respectively (C). Protein-DNA complexes were separated from unbound DNA on nondenaturing polyacrylamide gels and visualized by autoradiography of dried gels. FIGURE 6. Identified PPRE in the proximal promoter of Gpr81 is functional in cells. A, U2OS cells were cotransfected with a reporter construct (pGL3-Gpr81(Ϫ1059/ϩ28)) containing wild type (wt) or mutant (mut) PPRE together with empty (pCDNA) or PPAR␥-encoding expression vectors. Activation of the luciferase reporter in the absence or presence of 1 M rosiglitazone (Rosi) is expressed as fold induction over that with empty reporter cotransfected with pCDNA in the absence of rosiglitazone after normalization for Renilla luciferase activity. B, U2OS cells were transfected with the Gpr81(Ϫ1059/ϩ28) reporter and expression vectors encoding wild type PPAR␥1, PPAR␥2, or mutant PPAR␥2. Activation of the luciferase reporter is expressed as described above. Results are averages of at least three independently performed experiments assayed in duplicate means Ϯ S.E.
tributes to the anti-lipolytic action of TZDs in vitro. It is therefore tantalizing to speculate that also in vivo the insulin-sensitizing, antidiabetic and hypolipidemic actions of TZDs are partly mediated via PPAR␥-mediated up-regulation of the Gpr81 and Gpr109A genes in rodents and the GPR81, GPR109A, and GPR109B genes in humans. It should be noted, however, that adipocytes also express at least one additional anti-lipolytic GPR, Gpr43, which is also transcriptionally controlled by PPAR␥ (53). The presence of numerous anti-lipolytic GPRs in adipocytes may explain the relatively modest effect of Gpr81 knockdown on rosiglitazone-mediated inhibition of lipolysis observed in our experiments (Fig. 7). Furthermore, inhibition of lipolysis is clearly not the only mechanism by which TZDs reduce circulating FFA, as these drugs also stimulate adipogenesis and increase lipid uptake, lipid synthesis, lipid droplet stabilization, glycerol/FA recycling, and FA oxidation in adipose tissue (12). Investigations on the effects of TZDs in Gpr81 knock-out mice, Gpr109A knock-out mice, and double knock-out animals will help to establish the relative importance of these GPRs in mediating the lipid-lowering effects of these drugs.