This Article Abstract Full Text (PDF) All Versions of this Article: 25/6/899    most recent Final Manuscript Author Manuscript Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Knouff, C. Articles by Auwerx, J. Search for Related Content PubMed PubMed Citation Articles by Knouff, C. Articles by Auwerx, J.

Peroxisome Proliferator-Activated Receptor- Calls for Activation in Moderation: Lessons from Genetics and Pharmacology

Institut de Génétique et de Biologie Moléculaire et Cellulaire (C.K., J.A.), Centre National de la Recherche Scientifique/Institut National de la Santé et de la Recherche Médicale/Université Louis Pasteur, and Institut Clinique de la Souris (J.A.), Génopole Strasbourg, 67404 Illkirch, France

Correspondence: Address all correspondence and requests for reprints to: Johan H. Auwerz, M.D., Ph.D., Centre National de la Recherche Scientifique, Institut National de la Santé et de la Recherche Médicale, IGBMC, Parc d’Innovation, 1, rue Laurent Fries, 67404 Illkirch, France. E-mail: auwerx{at}igbmc.u-strasbg.fr

	Abstract

Top Abstract I. Introduction: The Biology... II. What Mutations in... III. Releasing the Power... IV. The Pharmacology of... V. Overview and Perspectives References

 
The peroxisome proliferator-activated receptor (PPAR) is a prototypical member of the nuclear receptor superfamily and integrates the control of energy, lipid, and glucose homeostasis. PPAR can bind a variety of small lipophilic compounds derived from metabolism and nutrition. These ligands, in turn, determine cofactor recruitment to PPAR, regulating the transcription of genes in a variety of metabolic pathways. PPAR is the main target of the thiazolidinedione class of insulin-sensitizing drugs, which are currently a mainstay of therapy for type 2 diabetes. However, this therapy has a number of side effects. Here, we review the clinical consequences of PPAR polymorphisms in humans, as well as several studies in mice using general or tissue-specific knockout techniques. We also discuss the recent pharmacological literature describing a variety of new PPAR partial agonists and antagonists, as well as pan-PPAR agonists. The results of these studies have added to the understanding of PPAR function, allowing us to hypothesize a general mechanism of PPAR action and speculate on future trends in the use of PPAR as a target in the treatment of type II diabetes.

I. Introduction: The Biology of Peroxisome Proliferator-Activated Receptor- (PPAR) in a Nutshell

II. What Mutations in the Human PPAR Gene Teach Us

A. Hypomorphic and loss-of-function alleles: Pro12Ala

B. Dominant-negative PPAR alleles

C. Gain-of-function PPAR mutations: Pro113Gln

D. Other mutations

E. Summary

III. Releasing the Power of Mouse Genetics on PPAR (Table 1)

A. Germline mutations of PPAR

B. Tissue-specific mutations of PPAR

IV. The Pharmacology of PPAR: Drugs to Mimic Mutations (Table 2)

A. Full agonists

B. Partial agonists

C. Antagonists

D. PPAR coagonists

E. Summary

V. Overview and Perspectives

View this table: [in this window] [in a new window]   TABLE 1. Mouse models for PPAR

View this table: [in this window] [in a new window]   TABLE 2. Summary of the properties of newer PPAR compounds

	I. Introduction: The Biology of Peroxisome Proliferator-Activated Receptor- (PPAR) in a Nutshell

Top Abstract I. Introduction: The Biology... II. What Mutations in... III. Releasing the Power... IV. The Pharmacology of... V. Overview and Perspectives References

The actions of PPAR are mediated by two protein isoforms, the widely expressed PPAR1 and adipose tissue-restricted PPAR2 (33, 34). Expression of each isoform is driven by a specific promoter that confers the distinct tissue expression patterns. These isoforms are produced from a single gene by alternative splicing and differ only by an additional 30 amino acids (28 in mice) in the N terminus of PPAR2 (35, 36, 37). The two additional amino acids in human PPAR2 are due to translation initiation in human PPAR1 at a methionine codon two residues downstream from the start codon used in mouse PPAR1. This addition of 30 N-terminal amino acids results in a 5- to 6-fold increase in the activation function of the N-terminal ligand-independent activation domain (AF-1) (38). There are also two other mRNA variants of PPAR, which differ in the 5'-untranslated region but give rise to proteins identical to PPAR1: PPAR3, which is restricted to macrophages, adipose tissue, and colon (39), and PPAR4, the tissue distribution of which is unclear at this time (40).

PPAR is the master regulator of adipocyte differentiation and energy storage and thus is central in the control of whole-body metabolism. Consistent with this, PPAR increases the expression of genes that promote fatty acid storage, whereas it represses genes that induce lipolysis in adipocytes (reviewed in Ref.41). The central role of PPAR in adipogenesis and metabolic regulation has been shown by a number of cellular, genetic, and pharmacological studies that we will discuss in more detail below.

White adipose tissue (WAT) is required for proper glucose homeostasis, because lipodystrophy is associated with severe insulin resistance (42). The observation that PPAR agonists increase fat mass along with improving glycemic control supports the notion that WAT mediates at least some of their effects on glucose homeostasis (43). A reduction in circulating FFA, due to fatty acid repartitioning toward fat rather than muscle and pancreas, is an early consequence of PPAR activation and precedes the decrease in glucose and triglyceride levels (41). Consistent with this, the level of insulin sensitization upon PPAR activation is inversely correlated with lipid accumulation in skeletal muscle (44). In addition to fatty acids, WAT secretes several adipokines that affect insulin signaling in other tissues. Examples are TNF (45), leptin (46, 47), and resistin (48), all secreted proportionally to WAT mass; and adiponectin, the levels of which are inversely related to the amount of WAT (49, 50, 51, 52).

The role of PPAR in the control of glucose homeostasis likely extends beyond its primary effects in WAT, because effects of PPAR ligands have been reported in muscle (53, 54, 55, 56, 57), pancreatic ß-cells (58, 59), and liver (60, 61). However, these data are potentially compromised by the lack of complete specificity of these ligands for PPAR. The importance of hepatic and muscle PPAR was recently underscored by the characterization of mice with a liver-specific or muscle-specific disruption of the PPAR gene (62, 63, 64). These tissue-specific PPAR-deficient models indicate that not only WAT, but also hepatic and muscle PPAR, contributes to metabolic homeostasis.

In addition to its involvement in insulin sensitization and adipose cell differentiation, PPAR may also play a critical role in the pathogenesis of atherosclerosis. For example, PPAR has the potential not only to regulate cardiac metabolism indirectly through its influence on circulating lipid and glucose levels, it also has been implicated directly in atherogenesis by modulating macrophage functions and foam cell formation (65). PPAR is highly expressed in the foam cells of early atherosclerotic lesions (66, 67), and expression is induced in human monocytes after exposure to oxidized low-density lipoproteins (oxLDLs). The ligand-activated PPAR induces the expression of the scavenger receptor CD36, leading to an increased uptake of oxLDL. Therefore, PPAR regulates the uptake of its ligand in an autoregulatory loop in foam cells (68). PPAR ligands also have been shown to stimulate the expression of proinflammatory receptors, such as CD14 and CD11b/CD18 (68), suggesting that PPAR can mediate the pathological response of the vessel. Thus, it was postulated that PPAR has a proatherogenic effect. However, PPAR has been shown to also exert antiatherogenic properties under certain circumstances. For instance, PPAR ligands reduce inflammatory cytokine production by macrophages (60, 67, 69) and inhibit the development of atherosclerosis in mouse models (70, 71, 72, 73). In addition, PPAR activation leads to a down-regulation of expression of the scavenger receptor SR-A in macrophages, whereas the expression of ATP-binding cassette (ABC)A1, a transporter involved in apolipoprotein A1-mediated cholesterol efflux, is activated, although potentially, through an indirect pathway involving liver X receptor (LXR) and RXR (73, 74, 75, 76). Thus, PPAR ligands regulate both influx and efflux of cholesterol and oxidized lipids in macrophages, counteracting their potentially atherogenic effect of CD36 induction.

	II. What Mutations in the Human PPAR Gene Teach Us

Top Abstract I. Introduction: The Biology... II. What Mutations in... III. Releasing the Power... IV. The Pharmacology of... V. Overview and Perspectives References

 
Human genetic studies support an important role of PPAR in adipogenesis. First, one of the loci with suggestive linkage to obesity maps close to the location of PPAR on chromosome 3p25-p24 (77). Second, the detailed study of a number of specific mutations in the PPAR gene has confirmed the role of this receptor in adipocyte differentiation and homeostasis (Fig. 1). Due to the presence of two PPAR isoforms, there has been some discrepancy in the numbering system for the various human mutations. In this review we will number all mutations based on PPAR2, which has a total length of 505 amino acids.

View larger version (13K): [in this window] [in a new window]   FIG. 1. Schematic of PPAR secondary structure and human polymorphisms. The three principal domains of PPAR, the N-terminal domain (gray), DNA-binding domain (white), and C-terminal ligand-binding domain (black), are highlighted. PPAR1 is widely expressed, whereas PPAR2 is restricted to adipocytes. PPAR2 contains an additional 30 residues at its N terminus in humans (28 in mice). The common Pro12Ala polymorphism is within these extra residues of PPAR2. All other mutations affect both isoforms. The constitutively active Pro113Gln mutation is located near serine 112 and interferes with phosphorylation. All the dominant-negative mutations are found in the ligand-binding domain and may interfere with ligand binding and/or heterodimerization with RXR. No mutations have yet been found in the DNA-binding domain.

1. Modifying factors.
Importantly, recent data have begun to show an interaction of modifying factors, including diet and exercise, with the Pro12Ala mutation in the development and treatment of type 2 diabetes. In the Quebec Family Study, carriers of the Pro12 allele had lower BMI, waist circumference, and fat mass (both sc adipose tissue and visceral adipose tissue) than carriers of the Ala12 allele. However, carriers of the Ala12 allele did not respond to a higher fat intake, whereas carriers of the Pro12 allele responded to high dietary fat with a gradual deterioration of metabolic parameters, as well as an increase in BMI and waist circumference (99). In a study by Kahara et al. (100), an exercise regimen for 3 months in healthy Japanese men improved insulin resistance in carriers of the Ala12 mutation, but not of the Pro12 allele. These results are also supported by the Finnish Diabetes Prevention Study, which showed that whereas the Ala12 allele may predispose to development of type 2 diabetes in individuals with insulin resistance, these individuals respond to dietary and exercise therapy with a significantly greater weight loss than subjects that carry Pro12 (88). Furthermore, none of the carriers of the Ala12 allele developed type 2 diabetes after dietary and exercise therapy, as opposed to carriers of the Pro12 allele. This is also supported by Nicklas et al. (101), who showed that women with the Ala12 allele who lost weight became more insulin sensitive than carriers of the Pro12 allele; however, women with the Ala12 allele were also more prone to weight regain. Another dietary intervention study has shown greater response to n-3 fatty acid supplementation in carriers of the Ala12 allele, who have a greater reduction in serum triglyceride levels after 3 months of treatment (102). Altogether, these studies suggest that the presence of the Ala12 allele protects carriers against negative environmental influences, such as high-fat diet and lack of exercise. Although these studies provide tantalizing evidence suggesting that there is an interaction between environmental factors and the Pro12Ala mutation in the risk to develop the metabolic syndrome, much further study needs to be done to consolidate this evidence.

Reduced fetal growth is associated with increased risk of type 2 diabetes and insulin resistance in adulthood (103). Furthermore, subjects born small for gestational age (SGA) display "catch-up" growth in childhood, with increased adiposity compared with children born appropriate for gestational age (AGA) (104). Jaquet et al. (105) investigated the role of the Pro12Ala polymorphism (among others) in this phenomenon. Interestingly, whereas AGA subjects with the Ala12 allele showed no difference in insulin resistance parameters compared with carriers of Pro12 (study subjects were young, healthy adults), higher insulin levels while fasting and after oral glucose tolerance test were observed in SGA subjects carrying the Ala12 allele compared with SGA subjects with Pro12. The authors speculate that this unexpected result may be due to the interaction of the genetic modifier (Ala12) with a specific environmental risk (impaired maternal-child environment in SGA subjects). Perhaps the increased childhood adipogenesis in SGA subjects results in altered function of Ala12 PPAR as compared with AGA subjects. Recently, another aspect of the metabolic syndrome, i.e., dyslipidemia, with increased atherogenic lipoprotein cholesterol, was found to be associated with Ala12 in SGA subjects (106).

Another potential way to study the impact of fetal growth on the development of insulin resistance is to examine dizygotic twins (DZ) vs. monozygotic twins (MZ) (MZ tend to have a more deleterious fetal environment than DZ). Interestingly, the silent C1431T mutation in exon 6 (discussed below) was found to be more common in DZ compared with MZ twins, leading the authors to speculate that this mutation may be linked to DZ twinning and intrauterine survival (107). This association was not supported in a recent study by Poulsen et al. (108), who examined the relation of the Pro12Ala mutation and twinning to diabetes risk. Although MZ twins had increased insulin resistance compared with DZ twins, subjects carrying Ala12 in both groups had improved IR parameters. Whereas these results may appear to contradict the results of Jaquet et al. (105), it is impossible to compare the studies because birth weight was not included in the analysis by Poulsen et al. (108); nevertheless, these results demonstrate a complex interaction between intrauterine environment and the Pro12Ala polymorphism in the development of diabetes.

2. Mechanism.
The Pro12Ala mutation represents the first genetic variant with a broad impact on the risk and complications of type 2 diabetes. Therefore, identifying the underlying mechanism of how it influences the multifactorial pathway that leads to the development of insulin resistance is essential. The localized expression of Pro12Ala PPAR2 in the adipocyte has underscored the fundamental role of adipose tissue in insulin sensitization, but the effects of this mutation on the muscle, liver, and other metabolic tissues are not well understood. A recent report in humans showed only a minor influence of the Pro12Ala mutation on the transcription of PPAR target genes in adipose tissue of massively obese individuals (109). However, other studies have shown interesting results of the Pro12Ala mutation on FFA metabolism. Using three different protocols, Tschritter et al. (110) observed a nearly 2-fold reduction of circulating FFA in the Ala12 group during the euglycemic hyperinsulinemic clamp. Furthermore, low lipoprotein lipase activity has been linked to coronary artery disease, and Schneider et al. (111) observed a nearly 20% reduction in postheparin plasma lipoprotein lipase activity in diabetic and coronary artery disease patients carrying the Ala12 mutation. However, no association between the Pro12Ala allele and fasting FFA levels was found in a healthy population in another study, although postprandial differences in FFA were not investigated (112). A more detailed analysis of the metabolism of FFA in a variety of nutritional states in carriers of this mutation is warranted to determine the exact role of alterations in FFA delivery to liver in the increased insulin sensitivity of Ala12 carriers.

Another postulated mechanism through which the Pro12Ala mutation may confer different susceptibilities to obesity and insulin resistance may be through alterations in adipocyte-derived peptide hormone levels. PPAR has been shown to transcriptionally regulate TNF (113), leptin (114), resistin (48), and adiponectin (115) (for review see Ref.41). A recent report from Yamamoto et al. (116) observed a decreased serum adiponectin level in Japanese subjects with the Ala12 allele. This result seems to contradict the observed decreased insulin resistance in individuals with the Ala12 mutation, as adiponectin has been shown to improve insulin sensitivity, but is consistent with a positive influence of PPAR on adiponectin transcription. However, a more recent report from Thamer et al. (117) examining European subjects showed no difference in adiponectin levels between carriers of the Ala12 or Pro12 alleles. Although no studies have shown any association between the Pro12Ala polymorphism and TNF or resistin levels, one study observed increased leptin levels in type 2 diabetic women possessing the Ala12 allele (118). It is hypothesized that increased leptin function in women carrying the Ala12 allele may result in decreased food intake and lowered BMI. Based on the conflicting evidence described above, no definitive correlation between the Pro12Ala polymorphism and adipocyte-derived hormones can be established at present, warranting further study.

3. Caveats.
As with any genetic association study, many of the above studies can have inherent errors that make interpretation difficult. False-positive association can occur due to stratification of the study population, as well as false negatives due to inadequate power of the sample size. These studies also can be prone to type I error of multiple comparisons, as well as publication bias for positive association. There is always the potential that there will be differences of the effect of the variant between different ethnic groups and genetic backgrounds. Additionally, a possible confounding factor in all of these studies is that the missense Pro12Ala mutation may itself not be the etiology for the protection from insulin resistance but could be in linkage disequilibrium with another mutation in the promoter region of the PPAR gene, resulting in reduced transcription and hence activity of PPAR in adipocytes of subjects possessing this allele (discussed in Ref.119). However, levels of PPAR mRNA were found to be increased in individuals possessing the Ala12 allele in visceral adipose in one study of massively obese subjects, although in vivo measurement of the expression of PPAR target genes between the different Pro12Ala genotypes failed to confirm that the Ala12 allele has reduced transcriptional activity (109). Although some studies have shown no associated polymorphisms in the promoter associated with the Ala12, a recent study has found a single nucleotide polymorphism (SNP) in a putative E2 box in the PPAR promoter in strong linkage disequilibrium with the Ala12 allele in the Pima Indian population (120). This SNP was shown to significantly alter transcriptional activity in vitro from a luciferase reporter construct. Although not in the promoter region, a silent SNP, C1431T in exon 6 (see below, His477His), has also been found to be in linkage disequilibrium with the Ala12 mutation, with Ala12 and T1431 occurring together in 70% of carriers (121). Furthermore, these results showed that whereas the Ala12 allele alone was associated with lower BMI, the T1431 allele alone was consistently associated with a higher BMI. Again, these studies highlight the complex nature of the Pro12Ala mutation and underscore the need for a clean genetic in vivo model to unravel the role of this PPAR mutation in the etiology of obesity and insulin resistance.

4. Other diseases.
Given the central role of PPAR in adipogenesis and the wide-ranging role of adipocytes as regulators of metabolism, it is not surprising to see several reports on an association of the Ala12 allele with other metabolic diseases, such as atherosclerosis. A recent large study showed that the Ala12 mutation may confer resistance to atherosclerosis, even in nondiabetic subjects and when traditional cardiovascular risk factors were taken into account (89). The only data to date on the relation of Ala12 on other atherosclerosis risk factors show independent (but conflicting) effects of the Ala12 allele on cholesterol homeostasis (122, 123), as well as oxLDL metabolism (124). A recent report demonstrates lower diastolic blood pressure in male type 2 diabetic subjects harboring the Ala12 allele (125).

Additionally, several groups have investigated whether there is any association of this allele with carcinogenesis. Decreased PPAR expression (via spontaneous somatic mutation) has been associated with more advanced breast cancer (126), multiple types of thyroid neoplasia (127), and prostate carcinoma (128). However, a recent study showed no correlation with PPAR expression and prostate cancer development in a mouse model (129). Although no link has been found between Pro12Ala status and prostate (130) or breast (131) tumorigenesis, there appears to be a clear link between the hypomorphic Ala12 allele and a reduced susceptibility to colorectal cancer (132). This is inconsistent with the observations of somatic mutations that inactivate PPAR in sporadic colorectal carcinoma (133) and studies in animals with xenografts of human colon carcinoma cells in which tumor development was inhibited by PPAR agonists (134). However, the idea that the hypomorphic allele may protect against colon carcinoma goes hand in hand with two studies in an adenomatous polyposis coli mutant animal model that documented a link between PPAR activation and the development of colon polyps (135, 136). This apparent paradox of the role of PPAR in colon carcinogenesis was explained by the observation that the protective role of PPAR may only be apparent if the adenomatous polyposis coli gene is intact (137). Additionally, the T1431 PPAR variant has been associated with increased endometrial cancer (138). Therefore, preliminary evidence demonstrates that the PPAR gene may also be involved in tumorigenesis, especially with tumors epidemiologically related to obesity and fat intake.

Another interesting loss-of-function mutation has been recently discovered in PPAR that results in a premature stop codon at residue 185 (139). This mutant is truncated in the second zinc-finger in the DNA-binding domain and hence PPAR cannot bind to PPREs. Interestingly, this mutation does not display dominant-negative characteristics, but results in severe metabolic syndrome only when digenically inherited with another mutant gene. Whereas several members of this pedigree were found to be heterozygous for the truncated PPAR, only those that were compound heterozygous for a premature stop codon in protein phosphatase 1, regulatory subunit 3 (the muscle-specific regulatory subunit of protein phosphatase 1), were found to have extreme hyperinsulinemia and characteristics of the metabolic syndrome (140). Those heterozygous for only one of either truncated allele had normal biochemical profiles.

B. Dominant-negative PPAR alleles
Several dominant-negative mutations in PPAR have been described. To date, the Pro495Leu (Pro467Leu) and Val318Met (Val290Met) nomenclature mutations are the best characterized in the literature. The Pro495Leu mutation involves a residue located in helix 12 in the AF2 motif of the ligand-binding domain of PPAR that is critical for mediating ligand-dependent transcription and coactivator recruitment (141, 142). The Val318Met mutation is more proximally located within the ligand-binding domain on helix 3 (Fig. 2). The net effect of both of these mutations is to disrupt the orientation of helix 12, which is important for the interaction of PPAR with ligands and coactivators, and hence has important effects on PPAR signaling. This has been shown in vitro by the markedly impaired response of both mutants to a PPAR agonist-mediated transcriptional transactivation (141). Loss of PPAR function in human subjects with either of these mutations is reflected by insulin resistance (141) and partial lipodystrophy (140). These human studies demonstrated that the Leu495 mutation was associated with an opposite redistribution of adipose observed in subjects treated with PPAR agonists, i.e., a loss of limb and buttock sc adipose tissue, and preserved abdominal sc and visceral adipose tissue, as well as hepatic steatosis (41). The adipose tissue in these subjects has clear abnormalities, especially the inability to trap postprandial FFA. This is associated with severely reduced plasma levels of adiponectin (140). Resistance of cultured monocytes from these subjects to PPAR agonist treatment was shown by the impaired increase in FABP4 (human homolog of murine aP-2) (140). Interestingly, although the mutated PPAR protein responds less effectively to the addition of agonists, metabolic consequences of PPAR agonist treatment have been reported in two subjects, demonstrating that TZDs have targets other than PPAR that mediate the observed metabolic affects (140).

View larger version (73K): [in this window] [in a new window]   FIG. 2. Structure of the C-terminal domain of PPAR with known dominant-negative mutations. A, The positions of three dominant-negative mutations, Pro495Leu, Val318Met, and Phe388Leu, are shown. Mutations in these residues will alter hydrophobic interactions near the ligand binding pocket (LBP, broken circle), altering normal ligand binding. B, The position of Arg425Cys substitution, together with other residues with which it interacts, is shown. Mutation of this residue to a cystine will disrupt multiple interhelical salt bridges (represented by broken lines), potentially disrupting the heterodimerization interface (DI, broken oval). Mutated residues (and residues that interact with these mutated residues) are highlighted in green and labeled in yellow. -Helices are shown in red and labeled in blue. ß-Sheets are shown in blue.

Recently, two other potential dominant-negative mutations have been discovered. A Canadian kindred with autosomal-dominant partial lipodystrophy was shown to have a T to A mutation at nucleotide position 1164 in PPAR, resulting in the (predicted) substitution Phe388Leu in PPAR2 (position 358 in PPAR1, Fig. 2) (148). Phe388 is found in helix 8 of the ligand-binding domain, at the core of a hydrophobic region between helices 8, 9, and 10. Replacement of this phenylalanine residue with a smaller leucine will presumably destabilize the packing of these helices and potentially affect ligand-binding domain architecture, resulting in impaired ligand binding and/or heterodimerization with RXR because helices 9 and 10 are part of the heterodimeric interface. Similar to the dominant-negative PPAR alleles described above, this mutant has significantly decreased basal transcriptional PPAR activity and impaired in vitro stimulation by a synthetic PPAR ligand.

Another patient with autosomal-dominant partial lipodystrophy was found to have a C to T mutation at nucleotide 1273 of exon 6. This patient had type 2 diabetes and hypertriglyceridemia and developed a lipodystrophy marked by severe loss of sc adipose tissue. This mutation results in an Arg425Cys substitution in PPAR2 (position 395 in PPAR1, Fig. 2) (149). This residue, found in the loop between helices 8 and 9, forms a salt bridge with glutamate 352 from helix 5 (position 322 in PPAR1), arginine 471 from helix 10 (position 441 in PPAR1), and aspartate 424 in loop 8–9 (position 394 in PPAR1). Mutation of Arg425 to a cysteine will, at the same time, create a hole between helices 5, 8, 9, and 10 and suppress the local charge compensation, resulting in a global rearrangement difficult to predict but that would probably involve a conformational change in loop 8–9 and neighboring side chains, including Glu352, Asp424, and Arg471. Because the latter two side chains are part of the heterodimeric interface, it is reasonable to assume that this interface will be affected by the Arg425Cys mutation. This would be expected to result in impaired ligand binding and/or heterodimerization with RXR. However, as yet, there are no data on the effect of this mutation on PPAR transcriptional activity in vitro.

In vitro evidence has confirmed and extended the data obtained in vivo in some of these mutations in the PPAR ligand-binding domain. Similar to the results of Savage et al. (140), Tamori et al. (150) overexpressed a different dominant-negative form of PPAR, in this case one that lacks the 16 COOH-terminal amino acids. This results in prevention of TZD-induced differentiation of 3T3-L1 cells into adipocytes. Furthermore, overexpression of this variant in mature adipocytes resulted in decreased adipocyte cell size and triglyceride content, increased lipolysis, and reduced FFA uptake as well as down-regulation of many genes involved in the maintenance of the adipocyte phenotype, including the glucose transporter-4, insulin receptor, insulin receptor substrate, and CCAAT/enhancer-binding protein-. Nugent et al. (151), who overexpressed a compound human PPAR1 mutant (Leu496Ala and Glu499Ala, positions 468 and 471 in PPAR1, respectively) in 3T3-L1 cells, showed that whereas this mutant strongly inhibited adipogenesis, upon activation of PPAR by TZD and non-TZD treatment, these cells still exhibited increased glucose uptake. These last results provide further evidence (along with much pharmacological data discussed below) that the main biological effects of PPAR stimulation, i.e., adipogenesis and insulin sensitivity, are separable.

C. Gain-of-function PPAR mutations: Pro113Gln
PPAR can be inactivated by phosphorylation of a serine residue at position 112 (also referred to as Ser114 in the literature) nomenclature, which affects both the AF-1 and AF-2 activation functions (152, 153, 154, 155). PPAR phosphorylation by MAPK induces the recruitment of the corepressor, silencing mediator of retinoid and thyroid receptors, and reduces receptor activity (17). This was proposed as a mechanism by which growth factors inhibit PPAR and adipocyte differentiation (152). A very rare human PPAR point mutation leading to an amino acid exchange at position 113 (Pro113Gln, also referred to as Pro115Gln in the literature) was shown to prevent serine 112 phosphorylation, resulting in accelerated adipocyte differentiation and cellular accumulation of triglyceride in vitro (156). Four humans possessing the PPAR Gln113 mutation were reported to be markedly obese, with a mean BMI of 41.9 ± 4.5, 25% higher than 117 other subjects in the obese group (156). Interestingly, whereas the subjects carrying the Gln113 mutation were markedly obese, they had nearly 2-fold lower fasting plasma insulin concentration compared with control obese subjects, suggesting lower insulin resistance. In contrast to this original report in four subjects, another study of one obese human with the Gln113 allele found increased fasting insulin levels and insulin resistance (157), whereas several other human studies underscored the extreme rarity of this allele (90, 158, 159, 160). A mouse model of the inhibition of PPAR phosphorylation, similar to the human mutation, was recently published (161). Interestingly, although inhibition of PPAR phosphorylation did not result in obesity, it did improve insulin sensitivity in mice with diet-induced obesity. The precise molecular mechanism underlying mutations in the AF-1 function on PPAR activity remains to be defined.

In-depth analysis of the in vivo actions of the constitutively active PPAR Gln113 mutation may inform on the development of therapeutic strategies aimed to increase PPAR activity. For instance, the insulin signaling pathway has been implicated in aging, but it is unclear whether this is due to alterations in the insulin pathway itself or is the result of changes in adipocyte function. Administration of PPAR ligands to mice can inhibit atherosclerosis (70, 71, 72), whereas mice lacking PPAR activity in macrophages have increased atherogenesis (76). It is therefore tempting to speculate that mice (and humans) with a constitutively active PPAR may be protected against atherosclerosis.

D. Other mutations
In addition to these well-defined mutations in the PPAR gene, a silent polymorphism, His477His, has been found in the human population. In the literature this mutation is referred to by various names, including C161T (for the nucleotide 161 in human exon 6), CAC477CAT (for codon 477 of human PPAR2), His449His (for codon 449 of PPAR1), and C1431T (for nucleotide 1431 of PPAR2). This silent polymorphism is a relatively common mutation, with the T allele frequency at 16.3% in an Australian Caucasian population (162) and 14.0% in a French population (163). The initial report showed that whereas this polymorphism was not associated with any differences in BMI or waist-hip ratio, obese subjects possessing at least one T allele had higher plasma leptin levels than those possessing the C allele (163). In a Finnish population, the T allele was associated with increased BMI in obese women, especially when coinherited with the Ala12 PPAR variant (93). Although no association between the T allele and BMI was found in an Australian population, subjects possessing the T allele had a nearly 2-fold reduction in coronary artery disease risk, particularly among CT heterozygotes (162). These subjects also were shown to have favorable plasma lipoprotein profiles, with reduced apoB and total cholesterol to high-density lipoprotein (HDL) cholesterol ratios. Another study in a Japanese pediatric population failed to find any effect of this allele on BMI (164). Interestingly, the T allele was also found to be enriched in a cohort of American patients with glioblastoma multiforme (165) and may also be associated with increased risk for endometrial cancer (138).

The mechanism by which this silent mutation in PPAR affects its activity remains unclear, but one possibility may be that it is in linkage disequilibrium with mutations in other regions of the gene that regulate the activity of PPAR [for instance, the Pro12Ala mutation as described above (121)]. These could also include mutations in regions affecting the transcription of the PPAR gene as well as mRNA or protein stability. Although no evidence for functional mutations in regions controlling stability of PPAR mRNA or protein have yet been found, evidence is emerging that there may be polymorphisms in the promoter region of PPAR with functional consequences. As outlined above, the PPAR gene produces four different mRNA isoforms, with PPAR1, -3, and -4 giving rise to identical protein products. By specifically examining the sequence of the PPAR3 promoter, Meirhaeghe et al. (166) identified a polymorphism at position –681 from exon A2. This polymorphism was found to be within a consensus binding site for the signal transducer and activator of transcription 5B and was found to abolish binding of this transcription factor to the site. Signal transducer and activator of transcription 5B acts in the GH signaling pathway, and this mutation was found to be associated with increased height in a French population, as well as increased plasma LDL cholesterol levels. This promoter mutation, together with the discovery of a SNP in the E2 box of the PPAR2 promoter in Pima Indians described above (120), underlines the importance of mutations that occur outside coding regions in the normal function of PPAR.

E. Summary
In conclusion, subjects with the partial "loss-of-function" Pro12Ala mutation in the PPAR2-specific B exon seem, in general, to have a lower BMI, greater insulin sensitivity, and an improved lipid profile (80, 85). The physiological consequences of the Pro12Ala polymorphism are, however, largely dependent on confounding genetic and environmental factors. It is also important to note that this is the only adipose-restricted mutation, and that all other known mutations will occur in PPAR1–4, with a far wider tissue distribution. In contrast to this hypomorphic allele, the rare Pro113Gln substitution renders PPAR constitutively active through the modulation of the phosphorylation of PPAR Ser112 by MAPK (152, 153, 154, 156). In keeping with the fact that PPAR stimulates adipocyte differentiation, carriers of the Pro113Gln gain-of function mutation are extremely obese and insulin sensitive (156). In contrast, several dominant-negative mutations, i.e., Pro495Leu, Val318Met, Phe388Leu, and Arg425Cys, have been associated with partial lipodystrophy, severe insulin resistance, diabetes, and hypertension (140, 141, 148, 149). An interesting point is that two of these mutations, i.e., Pro12Ala and Pro113Gln, map to the ill-understood NH₂-terminal ligand-independent AF-1 function of PPAR. Despite the fact that these mutations are not within the ligand-dependent AF-2 domain, they have a strong impact on PPAR activity, perhaps as great as the Pro495Leu mutation in the AF-2 domain, underscoring the importance of the AF-1 domain for normal PPAR activity.

	III. Releasing the Power of Mouse Genetics on PPAR (Table 1)

Top Abstract I. Introduction: The Biology... II. What Mutations in... III. Releasing the Power... IV. The Pharmacology of... V. Overview and Perspectives References

 
A. Germline mutations of PPAR
PPAR exerts pleiotropic functions, and alterations of PPAR activity have wide-ranging consequences affecting not only metabolism, but a variety of processes ranging from cell proliferation to inflammation. Mouse genetic studies offer excellent opportunities to unravel this complex physiology in a well-controlled genetic background and environmental setting. Initial studies showed that germline deficiency of PPAR results in embryonic lethality, with mice dying at embryonic d 10 from myocardial thinning due to deficiencies in placental vascularization and terminal trophoblast differentiation (167), similar to RXR mutant mice (168). A single PPAR-deficient mutant that lived to term via tetraploid-rescue subsequently died due to multiple hemorrhages and severe lipodystrophy. This critical dependence of WAT development on PPAR was confirmed in mice chimeric for wild-type and PPAR null cells (169).

Perhaps a more interesting phenotype was seen in mice that were heterozygous for the disrupted PPAR allele. Unexpectedly, these mice, with reduced PPAR activity, exhibit increased insulin sensitivity, with reduced steady-state insulin levels, increased insulin-induced glucose disposal rate, and suppression of hepatic glucose production compared with wild-type mice (170). High-fat diet has been shown to lead to adipocyte hypertrophy and insulin resistance in C57/Bl6 mice. Feeding high-fat diet to PPAR +/– mice resulted in less insulin resistance and no adipocyte hypertrophy compared with wild-type mice (171). A more recent study, however, is in apparent conflict with this result, showing that feeding a high-fat diet results in adipocyte hypertrophy and similar insulin resistance in both PPAR +/+ and PPAR +/– mice (172). However, whereas insulin sensitivity declined with age for both wild-type and PPAR +/– mice fed a normal chow, the decline was substantially less in PPAR +/– mice, such that at the age of 8 months the heterozygous mice were more insulin sensitive than wild type. Paradoxically, PPAR agonist treatment resulted in reversion to the wild-type phenotype in PPAR +/– mice fed a high-fat diet, with increased adipocyte hypertrophy and decreased insulin sensitivity (172). The conclusion of this last study was that partial PPAR deficiency partially protected against age-related insulin resistance, but not high-fat diet-induced insulin resistance.

B. Tissue-specific mutations of PPAR
One of the drawbacks of the characterization of such germline mutations is that they are often lethal in the homozygous state and affect all tissues. To avoid these complications, researchers have been increasingly resorting to the use of other gene knockout strategies, e.g., the Cre-loxP conditional knockout system. Although the following systems do not all utilize temporally controlled gene expression, they do allow a tissue-specific PPAR disruption, alleviating the in utero mortality.

1. Adipose.
Koutnikova et al. (173) generated by homologous recombination a WAT knockdown of PPAR. Although the birth weight of homozygous PPAR^hyp/hyp mice is similar to that of wild-type mice, the PPAR^hyp/hyp animals are growth retarded and develop severe lipodystrophy and hyperlipidemia. Almost half of these PPAR^hyp/hyp mice die before adulthood, whereas the surviving PPAR^hyp/hyp animals overcome the growth retardation, yet remain lipodystrophic. In contrast to most lipodystrophic models, the adult PPAR^hyp/hyp mice have only mild glucose intolerance and show no signs of a fatty liver. These metabolic consequences of the lipodystrophy are relatively benign because of the induction of a compensatory gene expression program in the muscle that enables efficient oxidation of excess lipids. These PPAR^hyp/hyp mice provide further evidence that adipose tissue PPAR is absolutely required for adipose tissue development, whereas lipid and glucose homeostasis can be relatively well maintained in the absence of WAT.

This model is supported by He et al. (174), who disrupted the PPAR gene specifically in WAT. Deletion of PPAR in this model also resulted in adipocyte hypotrophy and lipodystrophy, with elevated plasma triglycerides and FFA. In addition, these authors observed that there was no difference in insulin sensitivity in these mice, with normal muscle glucose uptake, yet impaired adipose and hepatic insulin sensitivity. In addition, these mice were more susceptible to high-fat diet-induced steatosis and insulin resistance. This model further supports the hypothesis that lipid and glucose homeostasis can be maintained in the absence of WAT.

To overcome these effects of constitutively deleted PPAR on adipocyte differentiation in early development, Imai et al. (175) used a new technique to temporally induce Cre-recombinase in the adult mouse, thereby allowing normal development of adipocytes. In this model, the Cre-recombinase is fused to the ligand-binding domain of the estrogen receptor, allowing induction of recombinase activity after tamoxifen administration. Imai et al. (175) found that mature PPAR null white and brown adipocytes die within a few days and are replaced by newly formed PPAR-positive adipocytes. This clearly demonstrates that PPAR is essential for the in vivo survival of mature adipocytes, in addition to its role in adipocyte differentiation.

2. Liver.
Matsusue et al. (62) disrupted the PPAR gene specifically in liver in both wild-type and leptin-deficient mice through the use of Cre-recombinase under control of the liver-specific albumin promoter. Lack of PPAR in the liver of leptin-deficient mice normalized hepatic triglyceride levels as a result of a decrease in expression of lipogenic PPAR target genes. However, plasma triglycerides and FFA significantly increased, resulting in an aggravation of insulin resistance in these mice. Using a similar approach, Gavrilova et al. (64) disrupted the PPAR gene specifically in liver in both wild-type and lipoatrophic A-ZIP/F-1 mice. Wild-type mice lacking liver PPAR developed fat intolerance, increased adiposity, hyperlipidemia, and insulin resistance. However, these mice maintained responsiveness to TZD PPAR agonists as measured by reduced serum glucose, insulin, and triglycerides and increased adiposity after treatment. In contrast, lipoatrophic A-ZIP/F-1 mice lacking liver PPAR had a significant decrease in hepatic steatosis but a deterioration of their hyperlipidemia, triglyceride clearance, and muscle insulin resistance. Ablation of PPAR in the liver of these mice lacking adipose tissue also abolished the hypoglycemic and hypolipidemic effects of rosiglitizone. Hence, the authors conclude that hepatic PPAR contributes to triglyceride homeostasis, regulating both triglyceride clearance and lipogenesis. The lack of hepatic PPAR results in reduced hepatic uptake of plasma triglycerides and triglyceride deposition in adipose tissue and muscle, contributing to the development of insulin resistance.

3. Muscle.
Skeletal muscle is known to be the main target tissue of insulin, with insulin resistance being mainly a result of impaired insulin-stimulated glucose disposal to this tissue (176, 177). Norris et al. (63) generated mice lacking PPAR in skeletal muscle using the muscle creatinine kinase promoter to drive the expression of Cre-recombinase. These mice became obese despite decreased caloric intake. Interestingly, although the muscle lacked PPAR, insulin-stimulated glucose uptake was normal. However, whole-body insulin resistance (due to impaired insulin function in liver and possibly fat) increased in these mice compared with wild type (similar to a muscle-specific insulin receptor knockout). When placed on a high-fat diet, mice lacking muscle PPAR developed hyperinsulinemia and impaired glucose homeostasis similar to wild-type controls. Both genotypes responded identically to treatment with PPAR agonists. The authors speculate that impaired hepatic insulin sensitivity may be a result of alterations in adipokine levels secondary to increased adiposity. How obesity developed in these mice in the presence of adipose tissue with reduced insulin sensitivity remains unclear. Possibly these mice have either reduced lipid oxidation as a result of decreased PPAR coactivator-1 activity or increased shunting of lipid to WAT from muscle. These last results also support the idea that PPAR agonists act primarily on adipose tissue, with increased insulin sensitivity in muscle a secondary effect caused by altered release of adipocyte-derived signaling factors, such as adipokines or fatty acids.

Hevener et al. (178) used a similar approach to create a muscle-specific deletion of PPAR. This group also reported increased adiposity in mice lacking muscle PPAR, with decreased hepatic insulin sensitivity. In contrast to Norris, Hevener et al. (178) found decreased action of insulin in muscle in vivo and in vitro in mice lacking muscle PPAR. Furthermore, this study showed a somewhat greater insulin resistance, with an 80% reduction in insulin-stimulated glucose disposal rate. However, these studies were done on older mice [12 months of age vs. 4 months in Norris et al. (63)]. Importantly, TZD treatment had no effect on insulin sensitization in skeletal muscle, yet did decrease plasma glucose, insulin, triglyceride, and FFA levels. This is a result similar to Norris et al. (63), who did not examine muscle-specific insulin sensitivity after TZD treatment, but did observe a whole-body hypoinsulinemic effect. Overall, these two studies agree, but the important difference in muscle insulin sensitivity between these models needs to be resolved and may involve age effects (see above) and strain differences, because Norris et al. (63) studied younger mice on more mixed background.

4. Macrophage.
Akiyama et al. (179) have selectively deleted PPAR in macrophages using an inducible (MX) promoter driving Cre-recombinase. This group reported a nearly complete loss of PPAR transcript in macrophages derived from these mice after MX promoter induction. PPAR-deficient macrophages exhibited a severe reduction in basal and TZD-induced expression of PPAR target genes, including lipoprotein lipase, CD36, LXR, and ABCs A1 (ABCA1) and G1 (ABCG1). Importantly, basal cholesterol efflux from cholesterol-loaded macrophages to HDL was significantly reduced from macrophages lacking PPAR. This group also showed that troglitazone, but not other TZD PPAR agonists, further inhibited ABCA1 and cholesterol efflux in both wild-type and PPAR null macrophages. These data are similar to the data of Chawla et al. (180), who transplanted PPAR-null and wild-type bone marrow to mice lacking the LDL receptor. Absence of PPAR in the macrophages resulted in a significant increase in atherosclerosis, further supporting the hypothesis that PPAR regulates LXR and ABCA1 expression, with increased PPAR activity increasing cholesterol efflux from foam cells and protecting against atherosclerosis.

5. Pancreas.
A pancreas-specific deletion has also been generated recently by using a rat insulin-Cre driver (181). The pancreas-specific PPAR-deficient mice did not manifest an obvious metabolic phenotype but had 2-fold increase pancreatic islet size, due to ß-cell hyperplasia. However, when placed on a high-fat diet, the opposite effect was observed, with a 25% reduction of the diet-induced ß-cell hyperplasia normally observed. These mice reveal an important proliferative effect of PPAR in pancreatic islets in obesity, yet also show that in these mice ß-cells lacking PPAR can compensate for their reduced number through increased function.

6. Other tissues.
Cui et al. (182) used the Cre-loxP system to disrupt the PPAR gene either in mammary tissue (with the Cre recombinase driven by the whey acidic protein promoter) or epithelial cells, ovary, B cells, and T cells (using the mouse mammary tumor virus promoter). Whereas these mice showed that PPAR is not necessary for functional development of mammary gland or the establishment of B and T cells, loss of the PPAR gene in oocytes and granulosa cells impaired fertility.

7. Other models.
Recently, Rangwala et al. (161) generated a mouse model with a mutation S112A that prevents PPAR phosphorylation at serine 112. This mutation will prevent the reduction in activity of PPAR by MAPK and should be similar to the gain-of-function mutation Pro113Gln found in humans. In contrast to some (but not all) reports of the human mutation (see above), the mouse does not have increased adipogenesis. However, in the setting of diet-induced obesity, these mice exhibit increased insulin sensitivity compared with wild-type mice. This is also associated with reduced adipocyte size, elevated serum adiponectin, and reduced FFA levels.

To date, only one study has examined the effects of overexpression of PPAR in mice. Using an adenovirus, PPAR1 was overexpressed in the liver of mice deficient for PPAR (183). This resulted in a large increase in the expression of adipocyte-specific genes, including adiponectin, aP2, and CD36, among others, which may explain the marked steatosis seen in the infected livers. Interestingly, hepatic steatosis induced by other methods in PPAR-deficient mice, e.g., by feeding choline-deficient diet or fasting, did not result in induction of these adipocyte-specific genes, suggesting that only PPAR overexpression can lead to partial transformation of hepatocytes to adipocytes, at least in mice lacking PPAR.

It is clear that the use of mouse models to study PPAR remains at an early stage. To date there are no published accounts of mice expressing the various mutated forms of the PPAR gene found in humans. Furthermore, the development of other mouse lines with inducible forms of the Cre recombinase under the control of a variety of cell- and tissue-specific promoters should open the door to a new set of studies that examine the homeostatic role of PPAR in multiple tissue types, both in developing and in adult animals.

	IV. The Pharmacology of PPAR: Drugs to Mimic Mutations (Table 2)

Top Abstract I. Introduction: The Biology... II. What Mutations in... III. Releasing the Power... IV. The Pharmacology of... V. Overview and Perspectives References

 
PPAR agonists, such as the TZDs, invariably increase WAT mass, redistribute WAT from visceral to sc depots, and induce the appearance of small, newly differentiated adipocytes at the expense of large, mature adipocytes (25, 184, 185). Whereas enhanced PPAR activity is associated with an increase in WAT, suboptimal PPAR activation is neutral or even reverses weight gain. Some partial PPAR agonists are less adipogenic but lower glucose more effectively than full agonists, an effect that is linked to the recruitment of a distinct set of coactivators to PPAR (reviewed in Ref.25). Furthermore, inhibition of PPAR/RXR activity by antagonists also translates into improved insulin sensitivity in vivo.

A. Full agonists
A variety of agonists for PPAR have been developed with the TZDs, rosiglitazone and pioglitazone, currently being used clinically in the treatment of type 2 diabetes (22). A thorough examination of the pharmacology of the PPAR agonists is beyond the scope of this review, and the reader is referred to more specialized reviews on this topic (25, 186, 187, 188, 189). PPAR agonists increase insulin sensitivity and exert antidiabetic effects. However, classical full PPAR agonists have a variety of side effects, chiefly weight gain, due to edema and increased fat mass. Troglitazone use was also associated with significant liver toxicity. Recently, new PPAR ligands have been developed, with improved affinities for PPAR, as well as differing transcriptional activities and pharmacological profiles. A brief review of recent data is informative for an understanding of PPAR function, although it must be stressed that the activities of many of these compounds may be due, in part, to PPAR-independent mechanisms.

B. Partial agonists
One can divide PPAR agonists into two groups, classical "full agonists," which are represented by the TZDs, such as rosiglitizone and pioglitazone, and newer "partial agonists," which were developed, in large part, to reduce the side effect of weight gain observed with the full agonists. Partial agonists are compounds that, at saturating concentrations, produce activity below that of saturating concentrations of a full agonist. GW0072 was the first partial agonist to be discovered but remains poorly characterized in the literature, with reports of decreased adipocyte differentiation in vivo compared with rosiglitazone (190). We recently demonstrated that FMOC-L-Leucine, a non-TZD partial PPAR agonist, activates PPAR with lower potency as the TZD agonists (191). It improves insulin sensitivity in normal as well as diet-induced and db/db glucose-intolerant mice. Interestingly, incubation of PPAR with FMOC-L-Leucine results in increased recruitment of the coactivator SRC-1 (steroid receptor coactivator 1), whereas TZD agonists promote association of PPAR with transcriptional intermediary factor 2 (TIF-2). TIF-2 is an important determinant of adiposity and can inhibit adaptive thermogenesis and lipid oxidation. This is supported by the fact that mice lacking TIF-2 have reduced weight gain despite increased caloric intake, a hyperactive brown adipose tissue, and an increase in adaptive thermogenesis (29). Therefore, partial agonists that result in PPAR association with SRC-1 rather than TIF-2 may be desirable.

MCC555 is a TZD with very potent antidiabetic activity in vivo yet only one tenth the binding affinity for PPAR compared with rosiglitizone (192). This compound also had impaired adipogenic activity as assessed in vitro using the 3T3-L1 cell differentiation system and aP-2 expression as a marker for adipogenesis. MCC-555 also induced different coactivator recruitment to PPAR, in this case impaired SRC-1 recruitment compared with similar levels of rosiglitizone. The other p160 coactivators, TIF-2 and SRC-3, were not investigated with MCC-555. CLX-0921, derived from the bark of Pterocarpus genus and used for treatment of diabetes mellitus in the Indian Ayurvedic system of medicine, is a weak activator of PPAR compared with rosiglitazone (193). However, this drug had a glucose-lowering activity in C57/BL6 mice as potent as rosiglitazone, with a 10-fold reduction in adipogenic drive. This compound also increased glycogen synthesis, which has not been observed for other PPAR agonists. PAT5A is another new, chemically distinct non-TZD compound with reduced capacity to activate PPAR compared with rosiglitazone (194) but with comparable insulin-sensitizing effects. As with CLX-0921, this partial agonist exhibited some pharmacological effects distinct from the full agonists, inhibiting cholesterol and fatty acid biosynthesis. This effect was also associated with a distinct cofactor recruitment profile from the full agonists, with decreased SRC-1 recruitment yet similar PPAR coactivator-1 recruitment as rosiglitazone. L-764406 is a non-TZD agonist that covalently binds Cys313 (195). This results in an increased, coactivator-dependent transcriptional activity of PPAR, although not nearly as much as observed by TZD treatment. Another non-TZD partial PPAR agonist ameliorated hyperglycemia and hyperinsulinemia, yet reduced weight gain and adipose depot size when administered chronically to C57/BL6 mice fed a high-fat diet (196). Furthermore, binding of non-TZD partial PPAR agonist to PPAR resulted in altered protein conformational stability as well as qualitative differences in affected gene expression compared with TZD agonists. This compound also has been shown to inhibit vascular smooth muscle growth and RB protein phosphorylation [unphosphorylated RB is a corepressor of PPAR (14) as well as E2F], further demonstrating its promise as a potential therapy for the metabolic syndrome (197). A synthetic triterpenoid, 2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid, induces adipocyte differentiation of 3T3-L1 cells, but not as much as rosiglitazone (198). Compared with rosiglitazone, 2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid had a reduced ability to recruit the cofactor CREB-binding protein (198), which is important for adipocyte differentiation (199).

Several NSAIDs, including indomethacin, fenoprofen, ibuprofen, and flufenamic acid, were shown to be low-affinity PPAR ligands (24). These same compounds were shown to activate PPAR at concentrations of 50–500 µM, which are above therapeutic levels required for cyclooxygenase inhibition and effects on the production of inflammatory cytokines (200). A recent report has shown that indomethacin can inhibit rosiglitizone-induced PPAR reporter gene activation in vascular smooth muscle cells, suggesting that, under certain conditions, these NSAIDs can act as a partial PPAR