Peroxisome proliferator-activated receptor-γ (PPAR-γ) is a nuclear receptor that is activated by the binding of an appropriate ligand. Several studies have demonstrated that certain ligands can also induce the expression of PPAR-γ. In the present study, we examined the mechanism whereby this induction occurs by specifically addressing whether potentiation of the transactivation function of PPAR-γ per se leads to induction of expression. We observed that thiazolidinediones, a group of insulin-sensitizing drugs, had differential effects, with troglitazone inducing protein levels of PPAR-γ, while rosiglitazone, englitazone, and ciglitazone were without effect. Similarly, the prostaglandin metabolite 15-deoxy-Δ12,14-prostaglandin J2 and the potent synthetic ligand GW1929 (N-(2-benzoyl phenyl)-l-tyrosine) also had no effect, as did ligands for other isoforms of PPAR. Since troglitazone has antioxidant properties, we also examined the effect of α-tocopherol and observed that it induced PPAR-γ expression in a dose-dependent fashion. Finally, we found that mice fed troglitazone as a dietary admixture displayed an up-regulation of hepatic PPAR-γ mRNA and protein, indicating that the mechanism of action is at the level of gene expression and not protein stability. These data indicate that 1) up-regulation of the transactivation function of PPAR-γ does not alone account for the induction of expression of PPAR-γ by troglitazone, and 2) an antioxidant-related mechanism may be involved.
Peroxisome proliferator-activated receptors (PPARs) are a family of nuclear receptors that regulate genes by binding to specificcis-elements in the promoters of genes as heterodimers with retinoid X receptor (reviewed by Mangelsdorf et al., 1995). There are three PPAR isoforms that have highly conserved DNA-binding domains but differ in their transactivation domains as well as in their relative tissue distribution (Mangelsdorf et al., 1995; Braissant et al., 1996). The transcriptional activity of PPARs is induced by the binding of ligands to a specific domain within the receptor. The ligands that have been identified thus far are generally lipid metabolites, with many showing some degree of cross-reactivity between isoforms (Bocos et al., 1995; Krey et al., 1997). However, some relatively isoform-specific ligands have been identified, a number of which have been synthesized chemically. These have greatly facilitated the characterization of the biological properties and roles of the individual PPAR isoforms.
The γ-isoform of PPAR was initially characterized as an adipose-specific factor that played a role in the differentiation of this tissue, and was found to regulate the expression of a number of adipose-specific genes (reviewed by Spiegelman, 1998). Additional roles for this transcription factor were suggested following the discovery that a class of insulin-sensitizing drugs used in the treatment of type 2 diabetes, termed thiazolidinediones (TZDs), were specific ligands for PPAR-γ (Lehmann et al., 1995; Lambe and Tugwood, 1996). Although this discovery suggested a possible mechanism whereby TZDs exerted their insulin-sensitizing effects (Reginato and Lazar, 1999), the adipose-specific distribution of PPAR-γ was difficult to reconcile with the insulin-sensitizing effects of TZDs that could be observed in nonadipose tissues such as muscle and liver (Spiegelman, 1998). A possible explanation for this apparent contradiction has been suggested by studies showing that troglitazone, a specific TZD, can induce the expression of PPAR-γ in nonadipose tissues and cell lines (Park et al., 1998; Davies et al., 1999a). The resulting increase in nuclear receptor levels that occurs in response to this drug could in turn provide the permissive environment necessary for the intracellular effects of these drugs to be exerted.
The mechanism behind the reprogramming of the relative PPAR isoform abundance that occurs in cells in response to troglitazone has not been addressed. Since troglitazone is a specific ligand for PPAR-γ, we previously hypothesized that troglitazone binds to and activates the small amount of PPAR-γ that is present in nonadipose cells, and this complex in turn binds to cis-elements in the promoter of the PPAR-γ gene that leads to induction of transcription and eventually protein levels (Davies et al., 1999a). This hypothesis predicts that other activating ligands for PPAR-γ should also induce expression of this nuclear factor. In the present study, we have tested this hypothesis by examining the ability of a variety of PPAR ligands, synthetic and natural, to induce the expression of PPAR-γ in isolated hepatocytes.
Troglitazone and englitazone were gifts from Pfizer Central Research (Groton, CT). Rosiglitazone was a gift from GlaxoSmithKline (Welwyn Garden City, Hertfordshire, UK), and GW1929 (N-(2-benzoyl phenyl)-l-tyrosine) was a gift from Glaxo Wellcome Inc.(Research Triangle Park, NC). 15-PGJ2, LY171,883 (1-(2-hydroxy-3-propyl-4-(4-(1H-tetrazol-5-yl)butoxy)phenyl)ethane),WY14,643 (4-chloro-6-(2,3-xylidino)-2-pyrimidinylthioacetic acid), ETYA (5,8,11,14-eicosatetraynoic acid), and ciglitazone were purchased from BIOMOL Research Laboratories (Plymouth Meeting, PA). Collagenase, α-tocopherol, and Williams E medium were purchased from Sigma-Aldrich Canada (Oakville, ON, Canada).
Preparation of Primary Hepatocytes.
Primary hepatocytes were prepared from male Sprague-Dawley rats, as previously described (Davies et al., 2001), which used the collagenase method originally described by Cascales et al. (1984). Principles of laboratory care were followed, and all protocols were approved by the University of Saskatchewan Committee on Animal Care. Cell viability was determined by the trypan blue exclusion method and was typically greater than 90%. Initially, hepatocytes were cultured on collagen-coated plates and incubated in Williams E medium containing 10% fetal bovine serum for 1 h. The medium was changed and cells were incubated for an additional 4 h. The medium was changed again, and the cells were incubated for 8 h in the presence of various compounds as indicated in the figure legends.
Western Blot Analysis.
The preparation of protein lysates from cultured cells has been described previously (Davies et al., 1995), as has the method for Western blot analysis (Davies et al., 1999a). Primary antibodies against PPAR-γ and β-actin were obtained from BIOMOL Research Laboratories and Sigma-Aldrich, respectively. A chemiluminescent detection system (PerkinElmer Life Sciences, Boston, MA) was used for identification of the antigen-antibody complex signal.
Human hepatoma HepG2 cells were transfected by the calcium phosphate precipitate method as described previously (Roesler et al., 1992). The PEPCK promoter reporter gene plasmid (−2086 PCK-CAT) contains promoter sequences extending from −2086 to +76 (Tontonoz et al., 1995), driving expression of the chloramphenicol acetyltransferase (CAT) as an indicator of promoter activity. The expression plasmids for PPAR-γ and RXR have been described previously (Tontonoz et al., 1995). A β-galactosidase reporter gene plasmid (RSV-βgal) was used as an indicator of transfection efficiency. HepG2 cells were cotransfected with the above plasmids, and the following morning cells were washed with phosphate-buffered saline, refed with fresh medium, and treated with the compounds as indicated in the figure legends for 24 h. Cells were harvested and extracts were prepared and assayed for CAT activity, β-galactosidase activity, and protein content (Bio-Rad reagent) as previously described (Roesler et al., 1992).
C57BL/6 male mice were fed and watered ad libitum, and kept under a 12-h light/dark cycle. Three mice served as controls and were fed an isocaloric diet (Pugazhenthi et al., 1993), and three mice were fed the same diet containing a 0.2% admixture of troglitazone. After 10 days on the diets, the livers were collected. Protein lysates were prepared and analyzed by Western blot as previously described (Davies et al., 1999a). Total RNA was also isolated from the livers and analyzed by Northern blot (Davies et al., 1999b), incorporating the use of UltraHyb reagent (Ambion, Austin, TX) to enhance the intensity of the signal. The32P-labeled cDNA probes for PPAR-γ (Tontonoz et al., 1995) and ribosomal phosphoprotein PO (RPPO) (Laborda, 1991) were prepared using the Random Priming Hexanucleotide mix using from Roche Molecular Biochemicals (Laval, PQ, Canada).
We recently observed that treatment of hepatocytes with troglitazone induces expression of PPAR-γ, both at the level of mRNA and protein (Davies et al., 1999a). Since troglitazone is a specific ligand for the γ-isoform of PPAR (Forman et al., 1995; Lehmann et al., 1995), we examined whether other ligands for this nuclear receptor could also induce expression. In Fig. 1and Table 1, the results of experiments testing the ability of other TZDs to increase the levels of PPAR-γ in hepatocytes are shown. In comparison to the effects of troglitazone, which increases levels of this protein by approximately 3-fold at the highest dose tested, neither ciglitazone, englitazone, nor rosiglitazone had any substantive effect. The lack of effect by rosiglitazone is particularly significant in light of the fact that this TZD has a binding affinity for PPAR-γ that is approximately two orders of magnitude greater than that of troglitazone (Young et al., 1998).
We next examined other classes of PPAR-γ ligands for their effect on the expression of their receptor in hepatocytes. 15-PGJ2 is a prostaglandin metabolite that is considered to be a “natural” ligand for PPAR-γ (Forman et al., 1995; Kliewer et al., 1995). Increasing concentrations of this compound had no effect on expression of PPAR-γ (Fig.2; Table 1). GW1929 is anN-aryl tyrosine agonist of PPAR-γ with a potency similar to that of rosiglitazone (Brown et al., 1999). As observed with 15-PGJ2, incubation of hepatocytes with increasing concentrations of this compound had no effect on PPAR-γ levels (Fig. 2; Table 1).
The lack of effect of the PPAR-γ ligands tested above, with the exception of troglitazone, led us to verify their ligand activity through the use of a reporter gene assay. We transfected human hepatoma HepG2 cells with a reporter containing the PEPCK promoter-driving expression of the CAT reporter gene. The promoter used has been shown to contain PPAR response elements through which PPAR-γ can activate transcription (Tontonoz et al., 1995). We also cotransfected with expression plasmids for PPAR-γ and RXR to ensure an adequate supply of heterodimeric receptor so that a sensitive response to the ligands would be obtained. As shown in Fig. 3, overexpression of PPAR-γ and RXR resulted in an approximately 6-fold activation of promoter activity. When transfected HepG2 cells were subsequently treated with troglitazone, 15-PGJ2, rosiglitazone, or GW1929, an additional enhancement of promoter activity was observed (Fig. 3). The greater responses observed with rosiglitazone and GW1929 are consistent with their higher binding affinities for PPAR-γ (Young et al., 1998; Brown et al., 1999). Treatment of HepG2 cells with ETYA produced no further enhancement of the promoter activity measured in the presence of overexpressed PPAR-γ, which is consistent with its specificity as a ligand for PPAR-α (Yu et al., 1995).
Several other common ligands for various PPAR isoforms were also tested. LY171,883, which is a broad-specificity ligand that activates all three isoforms (Cannon and Eacho, 1991), had no significant effect on PPAR-γ levels in hepatocytes (data not shown). A similar lack of effect was observed with WY14,643 and ETYA (data not shown), which are specific ligands for the α-isoform of PPAR (Kliewer et al., 1994; Yu et al., 1995).
Since troglitazone was the only ligand that demonstrated an ability to induce the expression of PPAR-γ, it was of interest to address what aspect of this compound was responsible for this activity. Troglitazone can act as an antioxidant due to the α-tocopherol moiety in its chemical structure (Inoue et al., 1997), and there is evidence to suggest that some of its biological activities are mediated through its antioxidant activity (Davies et al., 2001). To address whether troglitazone up-regulates PPAR-γ expression through an antioxidant-related mechanism, we assessed the effect of α-tocopherol on PPAR-γ levels in hepatocytes. As shown in Fig.4 and Table 1, treatment of cells with this antioxidant induced the levels of PPAR-γ in a dose-dependent manner.
Finally, the ability of troglitazone to up-regulate PPAR-γ expression in vivo was examined. C57 mice were fed a diet containing troglitazone as a 0.2% admixture for 10 days, following which the livers were processed and assessed by Western blot analysis for the expression of PPAR-γ. As shown in Fig. 5 (upper panel), the levels of hepatic PPAR-γ mRNA were significantly induced (5.8 ± 0.7-fold) in all three of the mice tested compared with control mice fed a diet lacking troglitazone, and a 2.1 ± 0.1-fold increase in PPAR-γ protein levels was also observed (Fig. 5, lower panel). These data confirm previous results which showed that rats fed troglitazone also led to up-regulation of PPAR-γ protein levels (Davies et al., 1999a) and further indicate that this effect is not exerted at the level of protein stability but rather on the level of gene expression, as evidenced by the elevated mRNA levels.
PPAR-γ is an adipose-enriched nuclear receptor that appears to play an important role in regulating overall energy homeostasis, and has been called the “ultimate thrifty gene” (Auwerx, 1999). This nuclear factor is important for adipose differentiation and is an intracellular receptor for thiazolidinediones, which are compounds that increase insulin sensitivity in a number of cell types. Several TZDs are currently in clinical use for the treatment of type 2 diabetes, and are particularly effective since they target the underlying insulin resistance. The pharmacological evidence that PPAR-γ is a target for thiazolidinediones is compelling (reviewed by Spiegelman, 1998), and include the observations that 1) all of the thiazolidinediones tested to date bind and activate PPAR-γ at concentrations that parallel their effective antidiabetic dose (Lehmann et al., 1995; Willson et al., 1996), and 2) the rank order of potency of the various thiazolidinediones is similar to their relative affinities for PPAR-γ (Lehmann et al., 1995; Young et al., 1998).
However, the hypothesis that the major intracellular target for TZDs is PPAR-γ might appear inconsistent with the observations that TZDs exert their insulin-sensitizing effects in tissues such as muscle and liver that have low levels of expression of PPAR-γ. One possible explanation, for which there is experimental support, is that TZDs somehow induce expression of PPAR-γ in those tissues where levels are normally low, thereby providing sufficient levels of the receptor necessary to mediate the cellular effects of these drugs. For example, troglitazone induces the expression of PPAR-γ in hepatocytes at the level of mRNA and protein, as well in the liver of rodents fed troglitazone in their diet (Davies et al., 1999a; this study). Treatment of human skeletal muscle cultures with troglitazone also leads to up-regulation of PPAR-γ mRNA and protein (Park et al., 1998). However, neither troglitazone nor pioglitazone has any effect on PPAR-γ mRNA levels in rat adipocytes (Tanaka et al., 1999), and rats treated with rosiglitazone show no observable increase in PPAR-γ mRNA levels in adipose tissue unless they are fed a high-fat diet; no effect was observed in control fed or high carbohydrate fed animals (Pearson et al., 1996). These studies, which used a variety of experimental models and assessed PPAR-γ expression in different tissues and/or cell types, did not allow for a conclusion to be made concerning the role of TZDs in PPAR-γ expression and the mechanism whereby such regulation occurs.
In this paper, we specifically examined the issue of whether up-regulation of PPAR-γ transactivation potential by agonists, including several TZDs, can induce the expression of this nuclear receptor in hepatocytes. Our observation that troglitazone was the only agonist that led to increased expression, while other more potent agonists such as rosiglitazone and GW1929, and general PPAR agonists such as LY171,883, had no effect, suggest that induction of PPAR-γ by troglitazone is not achieved through potentiation of the transactivation efficacy of PPAR-γ or other PPAR isoform. However, it is possible that the binding of troglitazone to PPAR-γ induces a different conformational change in the receptor compared with other ligands, even other TZDs, such that it recruits a unique coactivator complex, which in conjunction with other specific attributes of the PPAR-γ gene promoter, promotes transactivation in hepatocytes. In support of this hypothesis, Camp et al. (2000) showed that conformational differences occur in PPAR-γ when bound with troglitazone versus rosiglitazone. It has also been shown that the relative binding affinities of PPAR-γ agonists differ between cell types (Camp et al., 2000). Thus, simple analysis of relative binding affinities of ligands in vitro, or relative strength of agonist activity within a single cell type, may not provide a truly comprehensive picture of the biological activity of a specific compound.
We also examined whether this up-regulation occurred via an antioxidant mechanism. This possibility was explored based on the knowledge that troglitazone has antioxidant activity, due to its α-tocopherol moiety, whereas other TZDs lack this particular characteristic (Inoue et al., 1997; Davies et al., 2001). Recently, we demonstrated that troglitazone and α-tocopherol both inhibit PEPCK gene expression in hepatocytes, whereas other TZDs have no effect (Davies et al., 2001). Given the specific ability of troglitazone to up-regulate PPAR-γ levels, it was hypothesized that this might be due to its antioxidant potential. In the present study, α-tocopherol was also able to induce PPAR-γ expression, making this antioxidant and troglitazone the only two compounds we have observed to possess this activity. Precisely how an antioxidant could lead to up-regulation of PPAR-γ expression is not clear. It is known that alterations in the redox state within the cell can result in changes in gene expression patterns, and several transcription factors, including activator protein-1 (AP-1) and nuclear factor-κB (NF-κB) (Sen and Packer, 1996), have been identified which can mediate these effects. Current studies are aimed at examining whether either of these factors regulate transcription of the PPAR-γ gene.
It is also possible that troglitazone induces PPAR-γ expression through a mechanism distinct from those mentioned above, and indeed several studies suggest other mechanisms of action for TZDs that are PPAR-γ- and antioxidant-independent. For example, TZDs have been shown to directly inhibit 3β-hydroxysteroid dehydrogenase type II, with troglitazone showing greater activity relative to rosiglitazone (Arit et al., 2001). An additional mechanism of action of TZDs was recently demonstrated when it was shown that these drugs can lead to increased ubiquitination and subsequent degradation of PPAR-γ (Hauser et al., 2000). However, neither of these alternate mechanisms appear to provide a reasonable explanation for the enhanced expression of PPAR-γ in hepatocytes in response to troglitazone. Moreover, we have shown previously (Davies et al., 1999) and in this study that the up-regulation of PPAR-γ protein in cultured hepatocytes and in the livers of animals treated with troglitazone is associated with increased mRNA levels. Thus, the effect of troglitazone is exerted either at the level of gene transcription or mRNA stability and likely not on protein turnover. Ongoing studies examining the response of the PPAR-γ gene promoter to troglitazone and other PPAR-γ ligands in different cell types should provide further insight into the regulatory aspects of this important nuclear receptor.
We would like to thank Elmus Beale and Bruce Spiegelman for plasmids used in this study.
This work was supported by a grant from the Canadian Diabetes Association (to W.J.R. and R.L.K.).
- peroxisome proliferator-activated receptor
- chloramphenicol acetyltransferase
- 15-deoxy-Δ12,14-prostaglandin J2
- 4-chloro-6-(2,3-xylidino)-2-pyrimidinylthioacetic acid
- 5,8,11,14-eicosatetraynoic acid
- phosphoenolpyruvate carboxykinase
- ribosomal phosphoprotein PO
- retinoid X receptor
- Received June 6, 2001.
- Accepted September 26, 2001.
- The American Society for Pharmacology and Experimental Therapeutics