Prostaglandins as PPARγ Modulators in Adipogenesis

Adipocytes and fat cells play critical roles in the regulation of energy homeostasis. Adipogenesis (adipocyte differentiation) is regulated via a complex process including coordinated changes in hormone sensitivity and gene expression. PPARγ is a ligand-dependent transcription factor and important in adipogenesis, as it enhances the expression of numerous adipogenic and lipogenic genes in adipocytes. Prostaglandins (PGs), which are lipid mediators, are associated with the regulation of PPARγ function in adipocytes. Prostacyclin promotes the differentiation of adipocyte-precursor cells to adipose cells via activation of the expression of C/EBPβ and δ. These proteins are important transcription factors in the activation of the early phase of adipogenesis, and they activate the expression of PPARγ, which event precedes the maturation of adipocytes. PGE2 and PGF2α strongly suppress the early phase of adipocyte differentiation by enhancing their own production via receptor-mediated elevation of the expression of cycloxygenase-2, and they also suppress the function of PPARγ. In contrast, PGD2 and its non-enzymatic metabolite, Δ12-PGJ2, activate the middle-late phase of adipocyte differentiation through both DP2 receptors and PPARγ. This paper focuses on potential roles of PGs as PPARγ modulators in adipogenesis and regulators of obesity.


Introduction
Obesity is a major health concern worldwide [1] and is associated with the development of a number of pathological disorders such as type 2 diabetes, hypertension, and cardiovascular disease [2][3][4]. Excess adipose tissue can be the consequence of both an increased number (hyperplasia) and an enlarged size (hypertrophy) of adipose cells. A major role of adipocytes is to store large amounts of triglycerides during periods of energy excess and to mobilize these depots during periods of nutritional deprivation [2][3][4]. Moreover, adipocytes are highly specialized cells that secrete various adipocytokines, whose release largely reflects the amounts of stored triglyceride [2,[5][6][7][8].
The regulation of adipocyte differentiation (adipogenesis) is complex and this process includes alteration of the sensitivity to hormones and the expression of a number of genes in response to various stimuli including lipid mediators. Peroxisome proliferator-activated receptor (PPAR) γ and CCAAT/enhancer-binding proteins (C/EBPs) are the most important transcription factors involved in the activation of adipogenesis, and they induce the expression of a number of adipogenic and lipogenic genes that participate in the control of adipogenesis [9,10].
PPARs are members of the nuclear receptor superfamily and play critical roles in the regulation of storage and catabolism of lipids [11,12]. To date, three types of PPAR subtypes have been identified, that is, PPARα, PPARβ/δ, and PPARγ [11,12]. PPARs increase the expression of a variety of genes in various cells through heterodimerization with retinoic acid receptors or retinoid X receptors in a ligand-dependent manner [12][13][14][15][16]. Among them, PPARγ is expressed predominantly in adipose tissue and macrophages, is closely related to the regulation of lipid and glucose metabolisms, and is associated with the control of obesity and related diseases [11,12]. Until now, many natural and synthetic ligands for PPARγ have been identified [17][18][19]. 15-Deoxy-Δ 12,14 -prostaglandin (PG) J 2 (15d-PGJ 2 ) was the first identified endogenous ligand for PPARγ, and it activates adipogenesis in cultured cells [20,21]. Moreover, fatty acids such as lauric acid (C12:0) and petroselinic acid (C18:1) of the saturated fatty acids [22], linolenic acid (C18:3), eicosapentaenoic acid (C20:5), and docosahexaenoic acid (C22:6) of the ω3 (n-3) family [23], arachidonic acid of the ω6 (n-6) 2 PPAR Research family [22,23], and very-long chain fatty acids [24] were later identified as other endogenous PPARγ ligands that activate PPARγ functions. In addition, 9-hydroxy and 13-hydroxy octadecadienoic acids (HODE), the components of oxidized low-density lipoprotein (ox-LDL), were also identified as endogenous ligands for PPARγ [25,26]. However, whether these natural molecules can function as physiological ligands of PPARγ in vivo remains unknown. In addition to natural ligands, many synthetic ligands have been identified. For example, thiazolidinediones (TZDs) such as Troglitazone, Rosiglitazone, Ciglitazone, and Pioglitazone are used for the treatment of type 2 diabetes mellitus; and these ligands affect insulin resistance and glucose homeostasis by activating PPARγ functions [12,18]. However, these TZDs increase hepatic toxicity and cardiovascular risk. Finally, Troglitazone was withdrawn from the market [27]. It is still unknown whether the toxicities associated with TZDs are derived from the binding with PPARγ.
PGs are lipid mediators that play a number of physiological roles in a variety of cells. PGs are synthesized through the following three enzymatic steps ( Figure 1). First, arachidonic acid is liberated from the membrane phospholipids by the action of cytosolic phospholipase A 2 (cPLA 2 ) [28]. Second, arachidonic acid is converted to PGH 2 , which is a common precursor of all prostanoids, by either cyclooxygenase-(COX-) 1 or COX-2 [29]. The activity of these enzymes is critical to determine the production rate of PGs. Third, PGH 2 is metabolized to various PGs, that is, PGD 2 , PGE 2 , PGF 2α , prostacyclin (PGI 2 ), and thromboxane A 2 (TXA 2 ), by the action of specific PG synthases [29]. PGs exert a wide range of actions through their binding to specific PG receptors that belong to the G protein-coupled receptors (GPCRs) gene family [30]. GPCRs span cell membranes via seven transmembrane-spanning segments and are the most important therapeutic targets. In this decade, the functions of PGs in the regulation of adipogenesis have been extensively investigated. Elucidation of the molecular mechanisms underlying adipogenesis may provide strategies for reducing the prevalence of obesity. This paper focuses on the recent advances in our understanding of the function of PGs as modulators of PPARγ in the regulation of adipogenesis.

Roles of COXs in Adipocytes
COX consists of two isozymes, COX-1 and COX-2, and is the rate-limiting enzyme in the PG biosynthesis [29]. COX-1 is constitutively expressed in most cells including adipocytes, whereas COX-2 expression is induced by various stimuli [29] and transiently activated in the early phase of adipogenesis, followed by lowered expression during adipogenesis [31]. There have been a number of reports regarding the contribution of COX isozymes to the regulation of adipocyte differentiation. However, the roles that COX-2 plays during adipogenesis are still controversial.
In cell-based studies, Yan et al. demonstrated that inhibition of COX activities by their selective inhibitors, for example, SC-560 for COX-1, and NS-398 and Celecoxib for COX-2, enhances adipocyte differentiation via an increase in the mRNA levels of PPARγ and C/EBPα. Thus, both COX-1 and COX-2 participate in the regulation of adipogenesis [32]. Moreover, in 3T3-L1 cells stably expressing COX-2 in the antisense direction, lipid accumulation is enhanced during adipogenesis with elevated expression of adipogenic genes such as PPARγ and C/EBPα. In addition, this enhancement of lipid accumulation in antisense COX-2-expressing cells can be reversed by cotreatment with either antiadipogenic PGE 2 or PGF 2α [33].
In contrast, when 3T3-L1 cells are pretreated before the initiation of adipocyte differentiation or treated during the clonal expansion phase with SC-58236, a selective COX-2 inhibitor, and then caused to differentiate into adipocytes, lipid accumulation is reduced along with repressed expression of the adipogenic fatty acid-binding protein 4 (FABP4, also called aP2) gene [34]. In contrast, a selective COX-1 inhibitor, SC-58560 does not have any effect on adipogenesis. Additionally, when 3T3-L1 cells are caused to differentiate into adipocyte in a medium containing each of two selective COX-1 and COX-2 inhibitors that are added after the clonal expansion phase, adipogenesis is not affected. Thus, inhibition of COX-2 activity suppresses adipocyte differentiation by repressing the clonal expansion phase [34].
In in vivo studies, overexpression of COX-2 in white adipose tissue (WAT) increases de novo recruitment of brown adipose tissue (BAT) and energy expenditure, while suppressing the high fat diet-induced gain in body weight [35]. Also, Ghoshal et al. reported that in COX-2 geneknock-out mice, their total body weight is significantly lower than that of wild-type mice, along with reduced expression of adipogenic genes such as those of PPARγ and lipoprotein lipase [36]. In addition, PGD 2 and 15d-PGJ 2 levels in cells prepared from adipose tissues of COX-2 gene-knock-out mice and placed in primary culture are reduced as compared with those in wild-type mice [36]. Thus, further studies are needed to elucidate the precise functions of COXs in the regulation of adipogenesis.
AKR1B3-produced PGF 2α is detected in preadipocytes and its level is enhanced with a peak at 3 h after the initiation of adipogenesis and then decreases [50], indicating that PGF 2α suppresses an early phase of adipogenesis.  Fluprostenol, an FP receptor agonist, clearly reduces the expression of PPARγ and its target genes [31,50]. Moreover, this Fluprostenol-mediated suppression of the gene expression is cleared by cotreatment with AL8810, an FP receptor antagonist, indicating that PGF 2α inhibits adipocyte differentiation of 3T3-L1 cells by acting through an FP receptor. AKR1B7 gene-knock-out mice display excessive adiposity resulting from adipocyte hyperplasia/hypertrophy and exhibit high sensitivity to diet-induced obesity. Treatment of 3T3-L1 cells or AKR1B7 gene-knock-out mice with Cloprostenol, an FP receptor agonist, decreases adipocyte size and inhibits the expression of lipogenic genes [49].
The precise molecular mechanism of PGF 2α -mediated suppression of adipogenesis has been investigated. PGF 2α represses the function of PPARγ by causing its phosphorylation via FP receptors [50]. In addition, Fluprostenol enhances the expression of COX-2 via activation of the mitogen-activated protein kinase (MEK)/extracellular signal-regulated kinase (ERK) 1/2 pathway. Moreover, promoter-luciferase and chromatin immunoprecipitation assays demonstrated that PGF 2α -derived COX-2 expression is activated by the binding of cAMP-responsive element binding protein (CREB) to the promoter region of the COX-2 gene in 3T3-L1 cells [50]. Thus, the MEK/ERK-CREB cascade forms a positive feedback loop, one that probably plays a critical role in the suppression of the early phase of adipogenesis by elevating the de novo production of antiadipogenic PGF 2α .

Suppression of the Early Phase of
Adipogenesis by PGE 2 PGE 2 is also known to suppress adipogenesis through suppression of PPARγ function. PGE 2 and an EP4 agonist, AE1-329, increase the intracellular cAMP levels in 4 PPAR Research preadipocytes in a dose-dependent manner [42]. Moreover, AE1-329 decreases the expression of adipogenic genes such as PPARγ and C/EBPα [51]. The inhibitory effect of PGE 2 , but not that of Fluprostenol, is reversed by the addition of an EP4 antagonist, AE3-208 [42], indicating that PGE 2 suppresses adipogenesis through the EP4 receptor. Although the functions of PGE 2 and the expression of the functions of PGE 2 and the expression of PGESs have been investigated in adipocytes [27,52,53], the PGE 2 -producing enzyme in adipocytes has never been identified. To date, three major PGESs have been identified [54,55]. Microsomal PGES-1 (mPGES-1) is a member of the membrane-associated proteins in eicosanoid and glutathione metabolism (MAPEG) protein family [56] and produces PGE 2 in response to various stimuli [57]. Microsomal PGES-2 (mPGES-2) has also been identified and its expression is high in the heart and brain [58]. Cytosolic PGES (cPGES) is constitutively and ubiquitously expressed in various cells [59]. PGE 2 production is detected in preadipocytes and increases during the early phase of adipogenesis with a peak at 3 h after the initiation of adipogenesis; and mPGES-1 is expressed in these cells, with its mRNA and protein levels being consistently detected during adipogenesis. Finally, we found that mPGES-1 is responsible for the production of PGE 2 in adipocytes [60]. This result is consistent with results showing that treatment of mouse embryonic fibroblast (MEF) cells with PGE 2 for the first two days of adipocyte differentiation is enough to suppress adipocyte differentiation, with reduced expression of the PPARγ2 gene and reduced accumulation of intracellular lipids [43].
In wild-type mouse MEF cells, inhibition of endogenous PG synthesis by indomethacin enhances adipocyte differentiation, and this enhancement is reversed by the addition of PGE 2 . In MEF cells prepared from EP4 receptor geneknock-out mice, adipocyte differentiation is elevated, and no more enhancement of adipocyte differentiation is observed following treatment with indomethacin. Thus, PGE 2 -EP4 receptor signaling suppresses the early phase of adipocyte differentiation in MEF cells [43].

Synergistic Suppression of Early Phase of Adipogenesis by PGF 2α and PGE 2
Both PGF 2α and PGE 2 suppress the early phase of adipogenesis, and so we investigated the synergistic regulation of these PGs in 3T3-L1 cells. The increased production of PGF 2α and PGE 2 in the early phase of adipogenesis is a consequence of the elevated expression of the COX-2 gene [61]. PGF 2α forms a positive feedback loop that coordinately suppresses the early phase of adipogenesis through the increased production of antiadipogenic PGF 2α and PGE 2 , both of which inhibit PPARγ function. In addition, PGE 2 also enhances the production of PGF 2α and itself through the elevation of the expression of the COX-2 gene in an EP4 receptor-mediated fashion. Moreover, when the cells are caused to differentiate into adipocytes in medium containing both PGF 2α and PGE 2 , the expression of the adipogenic genes is decreased to a greater extent than when the cells are cultured in a medium containing either of them. Thus, PGE 2 and PGF 2α synergistically suppress the early phase of adipogenesis through a self-amplifying loop, triggered by PGF 2α -FP receptor and PGE 2 -EP4 receptor couplings and activation of the COX-2 gene expression in 3T3-L1 cells [61]. However, Inazumi et al. demonstrated that the differentiation of MEF cells prepared from FP receptor gene-knockout mice is almost the same as that in these cells from wildtype mice and still shows sensitivity to indomethacin, indicating that FP receptor-mediated suppression is not directly associated with the regulation of adipocyte differentiation in MEF cells [43]. Therefore, the regulation of suppression of adipogenesis by PGE 2 and PGF 2α might occur in a cell-typedependent manner.

Acceleration of Adipocyte Differentiation by PGD 2 and Its Metabolites
PGD 2 acts as an allergic and inflammatory mediators and is produced in a variety of cells such as mast cells, macrophages, and adipose cells [62,63]. PGD 2 is produced from PGH 2 by the action of PGD synthases (PGDSs), enzymes that catalyze the isomerization of the 9,11-endoperoxide group of PGH 2 to PGD 2 . Two distinct types of PGDSs have been identified. One is hematopoietic PGDS (H-PGDS), which is abundantly expressed in mast cells and Th2 cells [64]. The other is L-PGDS, which is detected abundantly in the brain, male genital organs, and heart [62,63]. PGD 2 has been considered a candidate for a molecule that acts as an endogenous inducer of adipogenesis, basically because 15d-PGJ 2 , one of its metabolites, has been identified as a ligand for PPARγ and activates adipogenesis in vitro [20,21]. PGD 2 is nonenzymatically metabolized to PGs of the J series, that is, PGJ 2 , Δ 12 -PGJ 2 , and 15d-PGJ 2 . However, the concentrations of 15d-PGJ 2 required for the activation of PPARγ reported in most of the literature are much higher (2.5-100 μmol/L) than those of conventional PGs (pmol/L range); and 15d-PGJ 2 is present in vivo at a low level that is insufficient for activation of adipocyte differentiation [65], whose finding is consistent with our current results indicating that 15d-PGJ 2 is not detectable in adipocytes [60]. Recently, we identified Δ 12 -PGJ 2 as being the dominant PGD 2 metabolite in differentiated adipocytes [60], in good agreement with recent results showing that Δ 12 -PGJ 2 is produced in adipocytes and activates the expression of adipogenic genes in 3T3-L1 cells [66]. PGD 2 is synthesized by the action of L-PGDS in adipocytes [67]. However, another PGDS, H-PGDS may not be involved in the production of PGD 2 in adipocytes, because the expression level of H-PGDS is very low during adipogenesis. Although the function of PGD 2 or L-PGDS in vitro has been extensively investigated, the in vivo function is still controversial. Ragolia et al. reported that adipose size is increased in L-PGDS gene-knock-out mice under normal and high-fat diet feeding. Moreover, L-PGDS gene-knockout mice become glucose intolerant and insulinresistant. Also the serum adiponectin level is decreased in such mice [68]. Adipocytes isolated from L-PGDS gene-knockout mice are significantly less sensitive to insulin-stimulated glucose transport. Thus, L-PGDS is an important mediator of muscle and adipose glucose transport which is modulated by glycemic conditions and plays a significant role in the glucose intolerance associated with type 2 diabetes [69]. Furthermore, Tanaka et al. showed that L-PGDS geneknock-out mice have a significantly increased body weight when fed high-fat diet and the size of adipocytes in the subcutaneous and visceral fat tissues is significantly enlarged [70]. In contrast, Fujitani et al. demonstrated that transgenic mice overexpressing human H-PGDS, which produce plenty of PGD 2 in every tissue including adipose, become obese under high-fat diet feeding but that obesity is not observed under normal diet feeding [71]. Serum leptin, insulin, and adiponectin levels are increased in these PGD 2overproducing mice. Moreover, their triglyceride level is decreased by about 50% as compared with that in WT mice. Moreover, the PGD 2 -overproducing mice show increased insulin sensitivity [71]. Furthermore, the epididymal adipose tissue mass of COX-2 gene-knock-out mice is decreased. PGD 2 and the levels of PGD 2 metabolites are also decreased in the adipose tissue of these mice. Thus, reduced adiposity in COX-2 gene-knock-out mice results from the inhibition of the production of PGD 2 and its metabolites required for PPARγ activation [36]. This discrepancy may be derived from a variety of physiological functions of PGD 2 in the body. Therefore, the adipocyte-specific function of PGD 2 and/or L-PGDS in the regulation of obesity should be further clarified.

Activation of Adipogenesis in
Adipose-Precursor Cells by PGI 2 PGI 2 activates the protein kinase A (PKA) pathway by binding to its IP receptor and enhances the differentiation of adipose precursor cells [72,73]. The activation of IP receptors upregulates the expression of C/EBPβ and C/EBPδ, both of which are critical for the progression of the early phase of adipogenesis and directly activate the expression of the PPARγ and C/EBPα genes for maturation of adipocytes [9,10]. Moreover, IP receptor gene-knock-out mice fed a high-fat diet do not show any changes in body weight, fat mass, or adipose size [74,75]. Therefore, PGI 2 activates the progression of adipogenesis in the adipose precursor cells through the enhancement of the expression C/EBPβ and C/EBPδ via the cAMP-PKA pathway.

Conclusion
PGs are involved in the regulation of adipogenesis and act as modulators of PPARγ functions. The regulation of adipogenesis by PGs is very complex, because PGs regulate adipogenesis both positively and negatively. In the early phase of adipogenesis, PGF 2α and PGE 2 suppress the progression of adipogenesis, and their receptor-mediated mechanisms leading to suppressed PPARγ function have been well elucidated. In contrast, PGD 2 and its metabolites activate the middle-late phase of adipogenesis ( Figure 2). In addition, recently we found that PGD 2 and its metabolite Δ 12 -PGJ 2 accelerate adipogenesis by acting through DP2 (CRTH2; chemoattractant receptor homologous molecule of Th2 cells) receptors and PPARγ, thus, indicating that when elucidating the function of a given PG, the roles of not only it but also those of its metabolites should be considered. All PGs function through their specific G proteincoupled receptors and PPARγ. Although their receptor agonists and antagonists are functional in the cultured adipocytes (in vitro), in vivo studies do not show clear effects of PGs in the regulation of obesity. Moreover, PG receptor gene-knock-out mice are not affected like the cells observed in in vitro studies. The explanation of the problems is quite difficult. As PGs have a variety of physiological functions, studies using gene-knock-out mice might not be appropriate to elucidate the functions of PGs in obesity. The precise in vivo functions of PGs especially those of 15d-PGJ 2 required further clarification. Tissue-(cell-)specific gene-knock-out mice might be a powerful tool to identify the in vivo function of PGs. Understanding of the mechanisms of PG-mediated regulation of adipogenesis may lead to a novel therapeutic strategy for the treatment of obesity.