/03/$15.00/0 Molecular Endocrinology 17(12): Copyright 2003 by The Endocrine Society doi: /me

Transcription

1 /03/$15.00/0 Molecular Endocrinology 17(12): Printed in U.S.A. Copyright 2003 by The Endocrine Society doi: /me The Peroxisome Proliferator-Activated Receptor / Agonist, GW501516, Regulates the Expression of Genes Involved in Lipid Catabolism and Energy Uncoupling in Skeletal Muscle Cells UWE DRESSEL, TAMARA L. ALLEN, JYOTSNA B. PIPPAL, PAUL R. ROHDE, PATRICK LAU, AND GEORGE E. O. MUSCAT Institute for Molecular Bioscience, The University of Queensland, St. Lucia, Queensland 4072, Australia Lipid homeostasis is controlled by the peroxisome proliferator-activated receptors (PPAR, - /, and - ) that function as fatty acid-dependent DNAbinding proteins that regulate lipid metabolism. In vitro and in vivo genetic and pharmacological studies have demonstrated PPAR regulates lipid catabolism. In contrast, PPAR regulates the conflicting process of lipid storage. However, relatively little is known about PPAR / in the context of target tissues, target genes, lipid homeostasis, and functional overlap with PPAR and -. PPAR /, a very low-density lipoprotein sensor, is abundantly expressed in skeletal muscle, a major mass peripheral tissue that accounts for approximately 40% of total body weight. Skeletal muscle is a metabolically active tissue, and a primary site of glucose metabolism, fatty acid oxidation, and cholesterol efflux. Consequently, it has a significant role in insulin sensitivity, the blood-lipid profile, and lipid homeostasis. Surprisingly, the role of PPAR / in skeletal muscle has not been investigated. We utilize selective PPAR, - /, -, and liver X receptor agonists in skeletal muscle cells to understand the functional role of PPAR /, and the complementary and/or contrasting roles of PPARs in this major mass peripheral tissue. Activation of PPAR / by GW in skeletal muscle cells LIPID HOMEOSTASIS IS regulated by dietary intake, de novo synthesis, catabolism, and lifestyle. Disorders of lipid metabolism are associated with hyperinsulinemia, and anomalous levels of the lipid triad, Abbreviations: ABC, ATP binding cassette; ACS, acyl-coa synthetase; ADRP, adipocyte differentiation-related protein; ApoE, apolipoprotein; CD36/FAT, FA translocase; CMB, confluent myoblasts; CoA, coenzyme A; CPT, carnitine palmitoyl transferase; DMSO, dimethylsulfoxide; FA, fatty acid; FABP, FA binding protein; FAS, fatty acid synthase; FFA, free fatty acids; G-6-P, glucose-6-phosphate; GLUT, glucose transporters; GYG1, glycogenin; HDL, high-density lipoprotein; LDL, low-density lipoprotein; LPL, lipoprotein lipase; LUC, luciferase; LXR, liver X receptor; M-CPT1, muscle carnitinepalmitoyl-transferase-1 (M-CPT1); MUFAs, monounsaturated fatty acids; NR, nuclear hormone receptor; PDC, pyruvate dehydroxygenase complex; PDK, pyruvate dehydroxygenase induces the expression of genes involved in preferential lipid utilization, -oxidation, cholesterol efflux, and energy uncoupling. Furthermore, we show that treatment of muscle cells with GW increases apolipoprotein-a1 specific efflux of intracellular cholesterol, thus identifying this tissue as an important target of PPAR / agonists. Interestingly, fenofibrate induces genes involved in fructose uptake, and glycogen formation. In contrast, rosiglitazone-mediated activation of PPAR induces gene expression associated with glucose uptake, fatty acid synthesis, and lipid storage. Furthermore, we show that the PPAR-dependent reporter in the muscle carnitine palmitoyltransferase-1 promoter is directly regulated by PPAR /, and not PPAR in skeletal muscle cells in a PPAR coactivator-1-dependent manner. This study demonstrates that PPARs have distinct roles in skeletal muscle cells with respect to the regulation of lipid, carbohydrate, and energy homeostasis. Moreover, we surmise that PPAR / agonists would increase fatty acid catabolism, cholesterol efflux, and energy expenditure in muscle, and speculate selective activators of PPAR / may have therapeutic utility in the treatment of hyperlipidemia, atherosclerosis, and obesity. (Molecular Endocrinology 17: , 2003) i.e. low high-density lipopoprotein (HDL) cholesterol, high low-density lipoprotein (LDL) cholesterol, and elevated serum triglycerides. The increased incidence of cardiovascular disease has been linked to dyslipidemias associated with diet and lifestyle. Insulin resistance, diabetes, atherosclerosis, obesity, and hypertension are comorbidities with these lipid disorders, which are collectively described as syndrome X. HDLs have a defensive role in the prevention of athero- kinases; PGC-1, PPAR coactivator-1; PMB, proliferating C2C12 myoblasts; PPAR, peroxisome proliferator-activated receptor; PPRE, PPAR-dependent reporter; RXR, retinoid X receptor; SCD, stearoyl CoA desaturase; SFAs, saturated fatty acids; TCA, tricarboxylic acid; tk, thymidine kinase; UCPs, uncoupling proteins; VLDL, very low-density lipoprotein. 2477

2 2478 Mol Endocrinol, December 2003, 17(12): Dressel et al. PPAR / Target Genes in Muscle genic dyslipidemia by mediating cholesterol efflux from peripheral tissues. In contrast, the LDLs accumulate in the arterial wall leading to atherosclerotic cholesterol-laden foam cells (1). Research demonstrates the evolution of a multilayered autoregulated system involving nuclear hormone receptors (NRs) for sensing and metabolizing biologically active lipids. NRs involved in control of lipid and cholesterol homeostasis, include the liver X receptors (LXRs), farnesoid X receptor, peroxisome proliferator-activated receptors (PPARs), - /, and - [NR1C1, -2, -3, respectively (2)] liver receptor homolog-1 and the small heterodimeric partner (3, 4). PPARs regulate the transcription of genes involved in lipid homeostasis, carbohydrate metabolism, energy expenditure and reverse cholesterol transport in a subtype- and tissue-specific manner. They are activated by a wide range of dietary factors, including saturated and unsaturated fatty acids (FAs), oxidized FA metabolites derived through the lipoxygenase and cyclo-oxygenase pathways, and selective synthetic compounds (e.g. hypolipidemic fibrates, and antidiabetic thiazolidinediones). From the viewpoint of PPAR /, the putative natural agonists are prostanoids, which are produced by the regulated conversion of poly-unsaturated FAs. In addition, PPAR, - /, and - form obligate and permissive heterodimers with the retinoid X receptors (RXR) that can also be activated by the RXR agonists 9-cis-retinoic acid, and/or specific synthetic agonists called rexinoids (e.g. LG101305). PPAR and - are predominantly expressed in liver and adipose tissue, respectively. The expression of PPAR / is ubiquitous. Moreover, it is very abundantly expressed in brain, intestine, skeletal muscle, spleen, macrophages, lung, and adrenals (5 7). Mouse transgenic, knockout, and knock-in studies coupled to pharmacological investigations have exposed the discrete physiological functions of the PPAR and - isoforms in lipid and carbohydrate metabolism. For example, PPAR promotes adipogenesis and increases lipid storage. In contrast, PPAR enhances the conflicting process of lipid catabolism/fa oxidation in the liver (5, 6). These physiological functions correlate with the hypolipidemic and antidiabetic (type II) effects of the synthetic and selective fibrate and glitazone drugs, which activate PPAR and PPAR, respectively. Relatively less is known about PPAR /, which has been implicated in bone and fat metabolism (8 10). Recently, the potent, synthetic and selective PPAR / agonist, GW501516, a phenoxyacetic acid derivative, has been reported (11). It was demonstrated that the triglyceride component of native very low-density lipoproteins (VLDLs) activate PPAR /. GW corrects hyperinsulinemia in insulin-resistant and obese primates. Furthermore, it raises ABCA1 mrna expression, and serum HDL cholesterol, while lowering triglycerides. However, PPAR / agonists also promote lipid absorption and storage in macrophages. Moreover, serum apolipoprotein (Apo) CIII levels and total cholesterol are raised. Hence, the overall effect of PPAR / agonists on whole body cholesterol homeostasis, lipid metabolism, target tissues and mode of action remain unclear (10). The PPAR -mediated FA oxidation in the liver plays a major role in ketosis that supports fuel requirements during fasting. Similar but distinct mechanisms must exist within peripheral tissues to implement localized responses to energy requirements and burdens in these tissues. For example, one would hypothesize that a PPAR knockout would have major consequences on skeletal muscle fuel metabolism and gene expression. However, Muoio et al. (12) observed the skeletal muscle metabolic/ -oxidation phenotype was not compromised in PPAR / mice, in contrast to the dramatic deleterious effects in liver and heart tissue. A plausible hypothesis suggests that PPAR / regulates fuel metabolism in skeletal muscle, a major mass peripheral tissue that accounts for 40% of the total body mass. Skeletal muscle is one of the most metabolically demanding tissues that relies heavily on FAs as an energy source. PPAR / is the most abundant PPAR in muscle tissue (12 14). It was first implicated in FA metabolism from studies using the knockout animals. Most PPAR / / embryos die at an early stage due to a placental defect, the small number that survive exhibit a reduction in fat mass/adiposity (8, 15). However, this phenotype is absent in an adipocyte-specific PPAR / knockout model, suggesting a complex autonomous action regulating systemic lipid metabolism (15, 16). This idea was further strengthened by the observation that treatment with the synthetic compound GW in insulin resistant primates dramatically improves the serum lipid profile, and improves hyperinsulinemia. However, it is unclear which tissue is the major target for this activity. The classification of PPAR / as sensor of dietary triglyceride in native VLDLs released by lipoprotein lipase (LPL) activity suggests skeletal muscle is a potential target tissue (17, 18). In addition, exercise and/or starvation induced up-regulation of FA oxidation genes in muscle remains intact in PPAR / mice. Muscle is a major site of glucose metabolism and FA oxidation. Furthermore, it is an important regulator of cholesterol homeostasis and HDL levels (19). Consequently, it has a significant role in insulin sensitivity, the blood lipid profile, and lipid metabolism. This underscores the need to define the contribution of this major mass tissue to PPAR /. Surprisingly, the fundamental role of PPAR / in skeletal muscle cholesterol, lipid, glucose, and energy homeostasis has not been examined. Correspondingly, the objective of this study is to examine the functional role of PPAR / in skeletal muscle, and to investigate the genes and regulatory genetic programs activated by PPAR / involved in the control of lipid and energy homeostasis. In summary, we demonstrate that PPAR / directly and/or indirectly regulates genes involved in triglycer-

3 Dressel et al. PPAR / Target Genes in Muscle Mol Endocrinol, December 2003, 17(12): ide-hydrolysis and FA oxidation [LPL, acyl-coenzyme A (CoA) synthetase 4 (ACS4), carnitine-palmitoyltransferase (CPT1)], preferential lipid utilization (PDK4), energy expenditure [uncoupling protein (UCP)-1, -2, and -3], and lipid efflux (ABCA1/G1). Furthermore, we show that the muscle carnitine-palmitoyl-transferase-1 (M-CPT1) is directly regulated by PPAR / in skeletal muscle, in a PPAR coactivator-1 (PGC-1)-dependent manner. In summary, we show that PPAR / activates the entire cascade of gene expression involved in lipid-uptake to FA oxidation, and in addition, activates the UCPs, thereby uncoupling oxidation from the production of energy, and increasing energy expenditure and thermogenesis. This provides the molecular basis for the lipid lowering effects of PPAR / agonists previously described in obese primates (11), and we speculate that PPAR / agonists would have therapeutic utility against a highfat diet and obesity. RESULTS GW Is a Potent Agonist for PPAR / in Skeletal Muscle C2C12 Cells The yet unclear distinct physiological role of PPAR / in major mass peripheral tissues led us to investigate the role of this PPAR subtype in skeletal muscle. We therefore performed RT-PCR using total RNA from C2C12-muscle cells to synthesize mouse PPAR / cdna into the expression vector psg5 as a tool for our studies. Furthermore, to verify the integrity of the cloned PPAR / constructs after sequencing, and the efficacy of the PPAR / agonist GW (11), we used the GAL4-hybrid system. Full-length PPAR /, the C-terminal D/E-domain (that encodes the ligandbinding-domain; LBD), and the N-terminal A/B-region (that encodes the AF-1 domain) were fused to the DNA-binding domain (DBD) of the yeast transcription factor GAL4 (Fig. 1A). If these regions encode functional transcriptional activation domains they will induce the GAL4-responsive reporter construct, G5E1b- LUC [containing a basal E1b-promoter with five 17- oligomer GAL4-binding sites linked to a luciferase (LUC) reporter gene] in an agonist-dependent manner in skeletal muscle C2C12-cells and nonmuscle CV1-cells. We cotransfected the GAL4-PPAR / -LBD together with the G5E1b-LUC reporter into CV1-cells and examined the dose-dependent activation by GW A maximum activity is reached at approximately 150 nm (with an EC nm), in agreement with the study of Oliver et al. (11) (who reported an EC 50 of 24 nm for mouse PPAR / ), thereby validating the potency and efficacy of our agonist preparation, and the integrity of the PPAR / -cdna (data not shown). GW (1 M) was used in all latter experiments, as used in all cell culture experiments by Oliver et al. (11) to induce a reproducible maximal PPAR / response. Moreover, Fig. 1. GW Is a Potent Agonist in Muscle- and Nonmuscle Cells A, Schematic representation of the PPAR / constructs used in this study. psg5-ppar / encoding for full-length PPAR /, encompassing the entire coding region [440 amino acids (aa)]; psv40-gal4-ppar /, encoding the GAL4- DNA-binding domain fused to full-length PPAR / ; psv40- GAL4-PPAR / -LBD, encoding the GAL4-DBD fused to the Hinge/LBD-Region (aa ); psv40gal4-ppar / - AF1, encoding the GAL4-DBD fused to the AF1 (aa 1 72). B, The indicated constructs were transiently transfected together with the G5E1B-LUC reporter into C2C12-muscle cells. Twelve hours after transfection, cells were treated for 24 h with GW (1 M), or DMSO as control. LUC activity is shown as relative light units (RLU). at this concentration Oliver et al. demonstrated that GW was highly selective and did not activate or bind RXR and other nuclear receptors. Subsequently, we examined the ability of this agonist to activate the different PPAR / -constructs in CV1 (data not shown) and C2C12 cells (Fig. 1B). We cotransfected the various GAL4-PPAR / constructs (full length, LBD, AF1) together with the G5E1b-LUC reporter into CV1 (data not shown), and C2C12-cells (Fig. 1B), in the presence or absence of GW (1 M). The AF1-domain of PPAR / inefficiently activated the LUC reporter, and did not respond to agonist treatment. In contrast, both GAL4-PPAR /, and GAL4-PPAR / -LBD-transactivated gene expression in an efficacious and agonist-dependent manner, in muscle (Fig. 1B) and nonmuscle cells (data not shown). Similar transactivation patterns were observed when using another GAL4-dependent reporter construct, tkmh100-luc, which utilizes the thymidine kinase (tk)-promoter backbone (20) instead of E1b (data not shown). Subsequently, we examined the ability of the PPAR / agonist GW to activate a PPAR-dependent reporter (PPRE) in muscle cells. Moreover we

4 2480 Mol Endocrinol, December 2003, 17(12): Dressel et al. PPAR / Target Genes in Muscle examined the ability of the cofactors PGC-1, p300 and SRC-2/GRIP-1 to coactivate GW dependent activation of gene expression in skeletal muscle C2C12 cells. We used the PPRE-tk-LUC reporter that contains three copies of a consensus binding site cloned upstream of the heterologous herpes simplex virus tk promoter linked to the LUC reporter gene. Furthermore, these experiments were performed in the absence of exogenous/ectopic PPAR / expression vector (because these cells contain endogenous PPAR / ). As shown in Fig. 2A the PPAR / agonist GW activated the expression of the PPREcontaining reporter approximately 2-fold in skeletal muscle C2C12 cells. No response was observed when the tk-luc-backbone, lacking the PPRE, was used (data not shown). Furthermore, GW dependent PPRE activation was enhanced when PGC-1, relative to p300 and SRC-2/GRIP-1, was cotransfected. For example, GW activated the expression of the PPRE-reporter approximately 2.5-fold, and approximately 4-fold in the presence of the RXR agonist. Subsequently, we validated the ability of the cofactor PGC-1 to coactivate GW PPAR / dependent transactivation of the PPAR-dependent PPRE-containing reporter in nonmuscle CV-1 cells. Therefore, we cotransfected PPRE-tk-LUC, psg5- PPAR / and the expression vector encoding PGC-1 in the presence or absence of the PPAR / agonist into nonmuscle CV1-cells (Fig. 2B). A PPAR / - and GW dependent, 4-fold activation of the PPREtk-LUC reporter is clearly observed. Furthermore, the experiment demonstrates the specific coactivation of PPRE-tk-Luc expression by PGC-1 expression. These studies demonstrate the integrity of the cloned PPAR / constructs, and verify the potent and efficacious function of the GW agonist preparation in nonmuscle and skeletal muscle cells. Furthermore, it demonstrates the selective coactivation of PPAR / mediated gene expression by PGC-1 in skeletal muscle cells. In addition, we further examined the ability of PPAR and PPAR agonists (fenofibrate and rosiglitazone, respectively) to regulate PPAR-dependent PPRE-containing reporter in muscle cells (Fig. 2C) and nonmuscle CV1 cells (data not shown) in the presence and absence of the coactivators PGC-1, p300, and SRC-2/GRIP-1. These experiments were performed to demonstrate these agonists regulate gene expression in skeletal muscle cells, as we subsequently wished to Fig. 2. PGC-1 Acts as a Transcriptional Coactivator for PPAR / in Muscle and Nonmuscle Cells A, Skeletal muscle C2C12 cells which endogenously express PPAR / were transfected with PPRE-tk-LUC and the indicated coactivators (or cdna3.1 as control) in the absence (Vehicle), or presence of the indicated agonists (RXR: LG101305, 0.1 M; PPAR /, GW501516, 1 M), or both together (LG & GW). Fold activation is shown relative to the LUC activity obtained after cotransfection of PPRE-tk-LUC and cdna3.1 in the absence of agonists. B, Nonmuscle CV1 cells were transiently transfected with PPRE-tk-LUC, psg5-ppar /, and cdna-pgc-1 in the presence, or absence of GW LUC activity is shown as relative light units (RLU). C and D, C2C12 cells were transfected with PPRE-tk-LUC and the indicated coactivators in the absence (Vehicle), or presence of the indicated agonists [PPAR : Fenofibrate (FF); 100 M, PPAR : Rosiglitazone (Rosi); 10 M, RXR: 9-cis retinoic acid (9cRA); 100 nm]. Fold activation is shown relative to the LUC activity obtained after cotransfection of PPRE-tk-LUC and cdna3.1 in the absence of agonists.

5 Dressel et al. PPAR / Target Genes in Muscle Mol Endocrinol, December 2003, 17(12): compare the relative effects of PPAR, - /, and - agonists on the expression of the genes involved in skeletal muscle lipid and carbohydrate metabolism. Figure 2, C and D, clearly demonstrates that PPAR and - agonists efficiently regulate gene expression in skeletal muscle cells. In addition, we show that SRC- 2/GRIP-1 and PGC-1 selectively coactivate PPAR and -, respectively, in skeletal muscle cells (Fig. 2, C and D). Finally, these experiments demonstrate that the C2C12 skeletal muscle cells express functional PPAR, - /, and - receptors that support the activation of PPAR-dependent gene expression by selective agonists. In summary, we demonstrate that PGC-1 expression in C2C12 cells selectively coactivates GW and Rosiglitazone mediated activation of the PPAR-dependent PPRE. The selective coactivation of PPRE expression by cofactors in the presence of the selective PPAR agonists in the skeletal muscle cells was performed in the absence of exogenous/ectopic (high level) receptor expression, and provides an unbiased demonstration of cofactor selectivity, and receptor functionality in skeletal muscle cells. Clearly, PGC-1 expression in skeletal muscle cells increases GW and Rosiglitazone inducibility, and the absolute level of PPREdependent expression. In contrast, SRC-2/GRIP1 expression preferentially increases Fenofibratemediated activation. Regulation of Gene Expression in Skeletal Muscle Cells by PPAR, - /, and - Agonists We investigated the expression of the genes involved in skeletal muscle lipid and carbohydrate metabolism (see Table 1) in the presence and absence of the PPAR, - /, and - agonists. We undertook these studies in the C2C12 skeletal muscle cell culture model. In this system, proliferating C2C12 skeletal myoblasts differentiate into post-mitotic multinucleated myotubes that acquire a muscle-specific, contractile phenotype. This in vitro system has been used to investigate the regulation of cholesterol homeostasis and lipid metabolism by LXR agonists (19). Muscat et al. demonstrated that the selective and synthetic LXR agonist, T induced similar effects on mrnas encoding ABCA1/G1, ApoE, stearoyl CoA desaturase (SCD-1), SREBP-1c, etc. in differentiated C2C12-myotubes and Mus musculus quadriceps skeletal muscle tissue. The physiological validation of the cell culture model in the mouse corroborates the utility of this model system. This evidence, coupled to the flexibility and utility of cell culture in terms of cost, agonist treatment, RNA extraction, and target validation provides an ideal platform to identify the PPAR / -dependent regulation of metabolism. In addition, and more importantly this cell line (16, 21 23) and other rodent skeletal muscle cell lines (13, 14) have been demonstrated to express functional PPAR, Table 1. Key Target Genes in this Study ABCA1 and ABCG8 ACS4 ADRP/Adipophilin ApoE CPT-1 FAS FAT/CD36 and FABP3 GLUT-4 and -5 Glycogenin/GYG1 LPL PDK-2 and -4 SCD-1 and -2 SREBP-1c UCP-1, -2, and -3 ATP binding cassette. Transporters that transfer cholesterol to the HDL acceptors, i.e. reverse cholesterol efflux. Acyl-CoA synthetase-4. Enhances the uptake of fatty acids by catalyzing their activation to acyl-coa esters for subsequent use in catabolic fatty acid oxidation pathways. Adipocyte differentiation-related protein. Involved in lipid storage. Apolipoprotein-E. Facilitates cholesterol and lipid efflux. Carnitine palmitoyl transferase-1. Transfers the long-chain fatty acyl group from coenzyme A to carnitine, the initial reaction of mitochondrial import of long-chain fatty acids and their subsequent oxidation. Fatty acid synthase. Involved in de novo fatty acid production. Fatty acid translocase, and fatty acid binding protein. Facilitate uptake of long chain fatty acids (LFCAs) and LDLs. Glucose transporters. GLUT4 facilitates glucose uptake in response to insulin stimulation. GLUT5 catalyzes uptake of fructose. Initiates the synthesis of glycogen, the principal storage form of glucose in skeletal muscle. Lipoprotein lipase. Hydrolysis of lipoprotein triglycerides into free fatty acids and responsible for the uptake of free fatty acids. Pyruvate dehydrogenase kinases. Inhibiting the pyruvate dehydroxygenase complex, thereby controlling glucose oxidation and maintaining pyruvate for gluconeogenesis. Stearoyl CoA desaturase-1 and -2. Enzymes associated with adiposity, i.e. storage and esterification of cholesterol, and responsible for the cis saturation of stearoyl and palmitoyl-coa converting them to oleate and palmitoleate, which are the monounsaturated fatty acids of triglycerides. Sterol regulatory element binding protein-1c, the hierarchical transcriptional activator of lipogenesis. Uncoupling proteins. Mitochondrial proteins that uncouple metabolic fuel-oxidation from ATP synthesis, regulating energy expenditure.

6 2482 Mol Endocrinol, December 2003, 17(12): Dressel et al. PPAR / Target Genes in Muscle Fig. 3. PPAR / and RXR Agonists Do Not Affect Myogenic Differentiation A, Schematic illustration of the experimental procedure: PMB were grown in DMEM supplemented with 20% fetal calf serum (FCS). After reaching confluency (CMB) cells were differentiated by changing the medium into DMEM supplemented with 2% adult horse serum (HS) for 4 d. Subsequently, the myotubes were treated with agonists for RXR (LG101305; 0.1 M), PPAR / (GW501516; 1 M), both together (LG & GW), or the vehicle (DMSO) as control. After 24 h, total RNA was harvested and analyzed using Northern blot experiments or quantitative real-time PCR. B, Northern blot analysis. After blotting, RNA was hybridized with 32 P- radiolabeled cdnas for key-indicators for myogenic differentiation (myogenin, sarcomeric -actin, cytoskeletal -, and -actin), cell-cycle exit (cyclin-d1), and terminal differentiation (p21). C, Quantitative real-time PCR analysis of PPAR expression levels in muscle cells. Total RNA from differentiated myotubes was analyzed for the expression of PPAR, - /, and -. Expression levels are normalized to GAPDH. The primers for PPAR detect all isoforms for PPAR. - /, and - receptors. Our quantitative real time analysis in Fig. 3C verifies the published reports that the PPAR mrnas are expressed in skeletal muscle C2C12 cells. Our analysis demonstrates PPAR and / are expressed at similar levels in 96 h differentiated myotube cells. PPAR mrna is abundantly expressed, however, the primers reflect mrna expression from all three PPAR isotypes (i.e ). Proliferating C2C12 myoblasts (PMB), cultured in DMEM supplemented with 20% FCS were grown to confluency (confluent myoblasts; CMB) and induced to differentiate into postmitotic multinucleated myotubes by serum withdrawal in culture over a 96-h period. This transition from a nonmuscle phenotype to contractile phenotype is associated with the repression of nonmuscle proteins concurrent with the activation of the contractile apparatus and metabolic enzymes (Fig. 3). We examined the consequences of 24 h treatment with agonists for PPAR / (GW501516; 1 M), RXR (LG101305; 100 nm), PPAR (fenofibrate; 100 M), PPAR (Rosiglitazone; 10 M) or the vehicle [dimethylsulfoxide (DMSO)] on these predifferentiated myotubes. We isolated total RNA from differentiated myotubes, which were treated with GW for 24 h (compare Fig. 3A), and analyzed the expression levels of several mrnas. Northern blot analysis demonstrated the induction of myogenin, repression of the cytoskeletal nonmuscle -, -actin, and the activation of the sarcomeric -actins which confirmed that these cells had differentiated into myotubes (Fig. 3B). Moreover, the repression of cyclin D1, and activation of p21 confirmed that these cells were exiting the cell cycle and differentiated terminally. This expression pattern is not altered by treatment with the RXR or PPAR / agonists (LG or GW501516, respectively), nor cotreatment with both agonists. Similarly, PPAR and - agonists had minimal (1.5-fold) effects on the myogenic expression patterns after 24 h treatments (data not shown). This suggests that these agonists do not significantly effect proliferation, cell cycle withdrawal and/or differentiation of these skeletal muscle cells. Subsequently, we used quantitative real-time PCR to investigate the expression pattern of genes involved in lipid/cholesterol absorption (CD36/FAT, FABP3; Fig. 4A), lipogenesis (SREBP-1c, FAS, SCD-1 and -2; Fig. 4B), triglyceride hydrolysis, and FA oxidation (LPL, M-CPT1, ACS4; Fig. 4C), glucose/fructose absorbtion and utilization (Glut-4, and -5; Fig. 4D, PDK-2 and -4; Fig. 4E), lipid efflux (ABCA1 and -G1, ApoE; Fig. 4F), energy expenditure (UCP-1, -2, and -3; Fig. 4G), and glucose and lipid storage (Glycogenin1/GYG1, adipophilin/adrp; Fig. 4H). The majority of candidate target genes investigated showed modest response ( 2-fold) to either the RXR, or the PPAR / agonist alone. However, the obvious exceptions were the mrnas for UCP-1 and -2, which encode UCPs involved in thermogenesis and energyexpenditure (Fig. 4G). UCP-1 and -2 were induced approximately 7- and 4-fold, respectively, by treatment with the PPAR / agonist GW More, importantly, these mrnas were not induced by agonists for PPAR or -. Although not abundantly expressed in adult muscle tissue, UCP-1 has been reported to be expressed in C2C12-cells (24). Other candidate mrnas that showed a modest, but significant increase in expression ( 2-fold) upon treatment with the GW were FABP3 (lipid uptake; Fig. 4A), LPL and M-CPT1 (triglyceride-hydrolysis and FA oxidation, respectively; Fig. 4C), UCP-3 (another member of the UCP family, involved in energy expenditure; Fig. 4G), and ADRP (lipid storage; Fig. 4H). All of these mrnas also responded to treatment with the RXR agonist, LG101305, and were synergistically activated upon treatment with both agonists. Noteworthy, the level of mrna encoding for UCP-2 was only marginally activated by LG101305, when compared with treatment with the PPAR / agonist. ADRP

7 Dressel et al. PPAR / Target Genes in Muscle Mol Endocrinol, December 2003, 17(12): Fig. 4. Regulation of Gene Expression in Skeletal Muscle Cells by PPAR, - /, and - Agonists Differentiated myotubes were treated as described in Fig. 3A for 24 h with agonists for RXR (LG101305; 100 nm), PPAR / (GW501516; 1 M), PPAR (Fenofibrate; 100 M), PPAR (Rosiglitazone; 10 M), or PPAR agonists together with the RXR agonist. After extraction, total RNA was analyzed by quantitative real-time PCR for the expression of genes involved in (A) lipid/cholesterol absorption (CD36/FAT, FABP3), (B) lipogenesis (SREBP-1c, FAS, SCD-1 and -2), (C) triglyceride hydrolysis, FA transport and oxidation (LPL, M-CPT1, ACS4), (D) glucose transport (Glut-4 and -5), (E) fuel utilization (PDK-2 and -4), (F) lipid efflux (ABCA1 and -G1, ApoE), (G) energy expenditure (UCP-1, -2, and -3), and (H) glucose and lipid storage (glycogenin1/gyg1, adipophilin/ ADRP). Results are shown as fold induction relative to the respective mrna level (normalized to GAPDH) in the absence of agonists.

8 2484 Mol Endocrinol, December 2003, 17(12): Dressel et al. PPAR / Target Genes in Muscle mrna expression was also induced by rosiglitazone treatment. The increase in mrna expression level subsequent to treatment with both agonists was also observed with a number of candidate target genes investigated. In the context of lipid and FA uptake, we observed a 6-fold increase in the mrna encoding FABP3, and a 2-fold increase in CD36 (Fig. 4A). Some regulators and markers of lipogenesis (SREBP-1c, SCD-1 and -2; Fig. 4B) were relatively refractory to treatment with one of the agonists, but showed significant induction after cotreatment ( 2- to 3-fold). Interestingly, FAS was reproducibly repressed upon treatment with the PPAR / agonist. In contrast, rosiglitazone increased FAS mrna expression approximately 2-fold. Interestingly, SREBP-1c induction did not result in the induction of the downstream targets, FAS, SCD-1 and -2. In muscle, PPAR / activation of SREB1c may be uncoupled from FA metabolism, similar to the uncoupling of LXR activity and FA metabolism observed in quadricep tissue (19). The transcripts encoding LPL, M-CPT1, ACS4, and PDK4 that are involved in triglyceride-hydrolysis, FA oxidation and preferential fuel utilization were induced approximately 7-, 4-, 3-, and 7-fold by cotreatment, respectively (Fig. 4C). The significance of the synergistic activation of LPL and CPT1 by cotreatment with the PPAR / and RXR agonists, are highlighted by the observation that the cotreatment with agonists for PPAR and - in the presence of an RXR agonist does not induce LPL and CPT1 expression (Fig. 4C). Interestingly, PDK4 mrna was activated by PPAR, - /, and - agonists. The glucose, and fructose transporters (Glut-4 and -5) were induced by rexinoid treatment, but completely refractory to the PPAR / agonist (Fig. 4D). In contrast, we observed Glut-4, and -5 were activated by PPAR and PPAR agonists, respectively. As mentioned earlier, cotreatment with agonists for PPAR / and RXR led to a dramatic increase in the level of mrnas encoding the UCPs that regulate energy expenditure. UCP-1, -2, and -3 were activated approximately 23-, 8-, and 16-fold, respectively (Fig. 4G). The significance, and specificity of the UCP-1 to -3 response by cotreatment with PPAR / and RXR agonists, are further highlighted by the observation that the cotreatment with the agonists for PPAR and - in the presence of an RXR agonist relatively poorly induced UCP-2 and -3 mrna expression (Fig. 4G). We also examined the expression of genes involved in lipid efflux, lipid storage, and glycogen deposition (Fig. 4F). ABCA1 mrna was synergistically induced approximately 11-fold by cotreatment with both agonists, ABCG1 approximately 4-fold. ApoE was induced by rexinoid treatment, but refractory to GW Cotreatment did not lead to further activation. Finally ADRP/adipophilin was synergistically induced approximately 7-fold by PPAR / and RXR agonist cotreatment. Furthermore, ADRP mrna increased approximately 2- to 3-fold to treatment with PPAR / and -, and the RXR agonists alone (Fig. 4H). The observed synergistic effect of cotreatment by both agonists is consistent with the fact that PPARs bind to DNA as heterodimers with RXR. Glycogenin was only induced approximately 2-fold by cotreatment. However, fenofibrate treatment induced glycogenin-1 mrna levels approximately 4-fold. To verify some of the results obtained from real-time PCR analysis, we performed Northern blot analysis using RNA extracted from C2C12-myotubes differentiated for 96 h and subsequently treated for 24 h with PPAR / and/or RXR agonists (Fig. 5, A and B). These results unconditionally confirm that the mrnas of UCP-2/-3 and LPL are induced upon treatment with GW501516, and validate the real-time PCR analysis. Figure 5B also demonstrates that the activation of UCP-2 occurs after a short time, such as 4 h. Furthermore, we explored whether some of the significant effects we observed only after cotreatment with GW and LG were specific to the PPAR / agonist, and not due to another RXR partner. For example, we examined the expression of ADRP, ACS4, glycogenin/gyg1 (Fig. 6), and ABCA1/G1 (Fig. 7, A and B) mrna expression in the presence of the LXR agonist, T Clearly, the LXR agonist does not activate the expression of mrnas encoding ACS4, glycogenin-1 and ADRP (Fig. 5, C E). However, LXR (a Fig. 5. Activation of UCP-2, -3, and LPL Is Confirmed in Northern Blot Analysis A, Total RNA isolated from differentiated myotubes treated with RXR and/or PPAR / agonists, as described in Fig. 3 was analyzed using Northern analysis. After blotting, RNA was hybridized with 32 P-radiolabeled cdnas encoding GAPDH, UCP-2 and -3, and LPL. B, Myotubes differentiated for 4 d (MT4s) were subsequently treated with GW (1 M) for the indicated time points. Cells treated for 24 h with DMSO (MT5s) were used as control.

9 Dressel et al. PPAR / Target Genes in Muscle Mol Endocrinol, December 2003, 17(12): Fig. 6. LXR Agonists Do Not Induce the mrnas Encoding for ADRP, ACS4, or GYG1 Total RNA from differentiated myotubes treated as described in Fig. 3A with agonists for RXR (LG101305; 0.1 M), LXR (T ; 1 M), or both together (RXR & LXR) was analyzed by quantitative real-time PCR for the expression of ADRP, ACS4, and GYG1. Results are shown as fold induction relative to the respective mrna level (normalized to GAPDH) in the absence of agonists. Fig. 7. PPAR / and LXR Agonists Induce ABCA1 mrna and ApoA1-Dependent Cholesterol Efflux in Skeletal Muscle Cells A and B, Total RNA isolated from differentiated myotubes treated with agonists for RXR (LG101305; 100 nm), PPAR / (GW501516; 1 M), both together (LG & GW), LXR (T ; 1 M), or RXR and LXR together (LG & T09) was analyzed by quantitative real-time PCR for the expression of ABCA1 (A) and ABCG1 (B). Results are shown as fold induction relative to the respective mrna level (normalized to GAPDH) in the absence of agonists. C, PPAR / -mediated activation of reverse cholesterol transport in differentiated myotubes. Confluent C2C12 cells were allowed to differentiate into myotubes in the absence of serum for 72 h. After differentiation, cells were cultured for an additional 24 h in the absence or presence of agonists. After ligand treatment, ApoA1- dependent cholesterol efflux was measured as described in Materials and Methods. demonstrable efficacious and potent ABCA1 activator) dramatically induced ABCA1 mrna expression in the absence and presence of RXR agonists (Fig. 6, A and B). Obviously, LXR is a more potent activator of ABCA1 mrna expression than PPAR /. However, Oliver et al. (11) demonstrated also that GW increases ABCA1 mrna expression and induces ApoA1-dependent cholesterol efflux in macrophages (although not as efficaciously as the LXR agonist). Moreover they observed a dramatic dose dependent rise in serum high density lipoprotein cholesterol. Hence, we examined whether GW promoted ApoA1-dependent reverse efflux in skeletal muscle cells, relative to the LXR agonist, T (Fig. 7C). LXR agonists are potent activators of ABCA1 mrna expression and ApoA1-specific cholesterol efflux in peripheral tissues and cells including macrophages, adipose and skeletal muscle (Ref. 19 and references therein). We treated differentiated skeletal muscle cells with LXR, RXR and PPAR / specific agonists. The LXR agonist, T induced approximately 3.5-fold increase in efflux relative to vehicle alone, and cotreatment resulted in a approximately 10-fold increase in ApoA1- specific efflux. Although not as effective as T , the PPAR / agonist produced an approximately 2.5- fold increase in reverse efflux to ApoA1, and cotreatment with the RXR agonist produced a 4.5-fold increase. The relative levels of PPAR / and LXR ApoA1-dependent efflux (Fig. 7C) are entirely consistent with the ABCA1 mrna levels in skeletal muscle cells (Fig. 7A), and those reported for LXR and PPAR / agonist in macrophages. In summary, we observed that the PPAR / agonist GW dramatically activates the mrnas encoding the UCPs, suggesting that PPAR / has an important role in energy uncoupling. Furthermore it activates the expression of genes involved in preferential lipid utilization, FA catabolism, and energy expenditure. Interestingly, fenofibrate induces genes involved in fructose uptake, and glycogen

10 2486 Mol Endocrinol, December 2003, 17(12): Dressel et al. PPAR / Target Genes in Muscle formation in skeletal muscle. In contrast, rosiglitazone-mediated activation of PPAR induces gene expression associated with glucose uptake, FA synthesis and lipid storage. This demonstrates that PPARs have distinct, complementary, and opposing roles in skeletal muscle. M-CPT1 Is a Primary Target of PPAR / in Skeletal Muscle We further explored the molecular basis of PPAR / mediated gene activation in skeletal muscle cells by evaluating whether direct, or indirect mechanisms mediated the observed increase in mrna levels. We investigated whether the promoters of selected target genes were active in skeletal muscle cells, and tested the responsiveness of the promoters to PPAR / and RXR agonists in a cell-based reporter assay. Because the promoters were introduced into skeletal muscle cells in the absence of exogenous receptors, the ligand-dependent responses reflect the functional properties of the endogenous receptors. We transiently transfected C2C12 cells with the regulatory sequences of selected target genes that were accessible to us, including ABCA1 (19), CD36/FAT (25), LPL (26), M-CPT1 (27), SREBP-1c (19), and UCP-2 (28), cloned in front of the pgl2/3-basic LUC backbone and examined the response after treatment with GW and/or LG Interestingly, only the M-CPT1 promoter responded to the PPAR / agonist and was further activated by cotreatment with both agonists (Fig. 8A). All other promoters tested, even though active in C2C12-skeletal muscle cells, did not respond to treatment with the PPAR / agonist, GW (data not shown). The transcriptional coactivator PGC-1 is expressed in skeletal muscle and has been demonstrated to induce mitochondrial biogenesis, oxidative metabolism, and thermogenesis (29 32). Furthermore, we had demonstrated that PGC-1 selectively coactivates GW induced PPRE expression in skeletal muscle cells (Fig. 2A), and PGC-1 has been shown to coactivate the liver-specific isoform of CPT1 (L-CPT1) (33). Hence, we investigated the ability of PGC-1 to coactivate the observed activation of the muscle- CPT1 promoter. We observed that PGC-1 significantly enhanced the transactivation of the M-CPT1 promoter after treatment with PPAR / agonist in skeletal muscle cells (Fig. 8B). Furthermore, the synergistic activation after treatment with both agonists was also significantly increased. We then investigated the regulation of the M-CPT1 promoter in CV1 cells, which do not endogenously express PPAR /. The M-CPT1 promoter was cotransfected into CV1 cells with PPAR /, PGC-1, or Fig. 8. The M-CPT1 Promoter Is Activated by PPAR / in a PGC-1-Dependent Manner A, pgl2-basic, or pgl2-mcpt1 (-1025/-12) were transfected into skeletal muscle C2C12-cells which endogenously express PPAR / and subsequently treated with agonists for RXR (LG101305), PPAR / (GW501516), or both together (LG & GW). B, The same constructs were transfected into C2C12-cells with, or without cdna-pgc-1, and treated as described above. C, Nonmuscle CV1-cells that lack endogenous PPAR / were transfected with PPRE-tk-LUC, psg5-ppar /, cdna-pgc-1 in the absence or presence of agonists for PPAR / and RXR. D, A similar experiment was carried out to compare the effect of PGC-1 with that of the coactivators SRC2/GRIP1 and p300 in nonmuscle CV1 cells. Reporter activity is shown as relative light units (RLU).

11 Dressel et al. PPAR / Target Genes in Muscle Mol Endocrinol, December 2003, 17(12): both together. As seen in Fig. 8C, PPAR / activated the M-CPT1 promoter in nonmuscle CV1 cells significantly only in the presence of agonists and exogenous PGC-1. We also demonstrated that PGC-1, relative to SRC-2/GRIP-1 and p300 most efficiently coactivated the M-CPT1 promoter (Fig. 8D). In summary, these results clearly demonstrate that M-CPT1 is a target for PPAR /, and selectively coactivated by PGC-1. Cofactor expression increases GW inducibility and the absolute levels of CPT1 expression. To rigorously define that M-CPT1 was regulated by GW in a PPAR / -PPRE-dependent manner, we mutated the previously defined PPAR response element in the M-CPT1 promoter between 775 and 763 bp upstream of the initiator codon (27, 34). We showed that the wild-type M-CPT1 promoter and not the PPRE mutant M-CPT1m1 was specifically activated by the PPAR / -specific agonist (Fig. 9A). This demonstrated that GW mediated activation is dependent on the M-CPT1 PPRE. This element was previously defined as a PPAR -regulated motif in cardiac muscle and primary cardiomyocytes (27, 34). Consequently, we examined the ability of the synthetic and selective PPAR, - /, and - agonists to activate the expression of the wild type M-CPT1 promoter and the PPRE mutant M-CPT1m1 in skeletal muscle cells in the presence and absence of the coactivator, PGC-1. We observed that in skeletal muscle cells that M-CPT1 was regulated preferentially by the selective PPAR / and not the PPAR agonist (Fig. 9 B/C). In summary, M-CPT1 is an established target for PPAR in cardiac muscle (27, 34, 35). However, we clearly demonstrate by transfection in the presence of PPAR, - /, and - agonists that the M-CPT1 promoter in skeletal muscle cells responds preferentially to PPAR / agonist activation (and not a selective PPAR agonist) (Fig. 9, B and C). This is reminiscent of the differential cell specific regulation of the PPRE in the LPL promoter by PPAR and - agonists in adipose and liver (32). Moreover, we showed that the wild type M-CPT1 promoter and not the PPRE mutant M-CPT1m1 was specifically activated by the PPAR / -specific agonist (Fig. 9, A C). This demonstrated that GW mediated activation was dependent on the M-CPT1 PPRE. This element was previously defined as a PPAR -regulated motif in cardiac muscle (27, 34, 35). Moreover, we demonstrate that the native LPL promoter responds to PPAR, not PPAR /, agonists in muscle cells in the absence of exogenous PPAR (Fig. 10A), further validating the specificity of the PPAR / response on the CPT1 promoter in skeletal muscle cells. The previous literature demonstrates that these cells express functional PPAR, /, and - receptors (16, 21 23). In addition, our data demonstrate the multimerized DR-1 PPRE reporter is efficiently activated by PPAR, - /, and - agonists in the absence of exogenous receptors. The LPL promoter data, transfection data in the absence of exogenous receptors and the previous reports above clearly demonstrate the selective and specific activation of M-CPT1 by PPAR / and not - agonists, is not due to lack of PPAR expression, and/or nonfunctional PPAR. Finally, to rigorously demonstrate that the selective ac- Fig. 9. M-CPT1 Is a Direct Target of PPAR / in Skeletal Muscle A, Schematic of the characterized PPRE in the M-CPT1 promoter ( 1025/ 12) and the introduced point mutation M-CPT1-m1, according to Brandt et al. (27), is shown. pgl2-basic, pgl2-mcpt1-wt, or pgl2-mcpt1-m1 were transfected into C2C12-cells and subsequently treated with agonists for RXR (LG101305), PPAR / (GW501516), both together (LG & GW), or DMSO. Reporter activity is shown as relative light units (RLU). B and C, pgl2-mcpt1-wt (B), or pgl2-m-cpt1-m1 (C) were transfected into C2C12-cells and subsequently treated with agonists for RXR (LG101305), PPAR / (GW501516), PPAR (rosiglitazone, 10 M), PPAR (Wy14634, 10 M), or the vehicle (DMSO). Reporter activity is shown as RLU.

12 2488 Mol Endocrinol, December 2003, 17(12): Dressel et al. PPAR / Target Genes in Muscle 24 h RXR agonist, PPAR / agonist, and cotreatment (Fig. 10B). As observed earlier, PPAR and - / mrnas are similarly expressed relative to GAPDH mrna before agonist treatment; however, RXR and PPAR / agonist treatment preferentially induced PPAR mrna expression (Fig. 10B). This definitively demonstrates that the selective activation of the M-CPT1 promoter by the PPAR / agonist (and not the PPAR agonist) in skeletal muscle cells is not due to lack of PPAR expression. Lastly, we observe that the induction of M-CPT1 mrna expression by GW was also observed with cycloheximide treatment, suggesting that this effect is independent of de novo protein synthesis (Fig. 10C). In summary, we have shown that GW directly regulates the M-CPT1 promoter in a PPAR / /PPRE-dependent manner. DISCUSSION Fig. 10. PPAR Is Expressed and Fuctional in Muscle Cells A, pgl2e-lpl ( 565/ 181), or the vector backbone pgl- Enhancer were transfected into muscle C2C12 cells (which endogenously express PPARs) and treated with selective agonists for PPAR (fenofibrates), PPAR (rosiglitazone), PPAR / (GW501516), or the vehicle DMSO as control. Reporter activity is shown as relative light units (RLU). B, Total RNA from differentiated C2C12 myotubes treated as described in Fig. 3A with agonists for RXR (LG101305), PPAR / (GW501516), or cotreatment (LG & GW) was analyzed by quantitative real-time PCR for the expression of PPAR and - / mrnas. Expression levels are normalized to GAPDH. C, Differentiated myotubes were treated for 8 h with the agonist for PPAR / (GW501516) in the absence or presence of cycloheximide (10 g/ml). Total RNA was analyzed by quantitative real-time PCR for the expression of M-CPT1. Results are shown as fold induction relative to the respective mrna level (normalized to GAPDH) in the absence of drugs. tivation of the M-CPT1 promoter is not due to the elevated expression of PPAR /, relative to PPAR mrna, after agonist treatment we examined the expression of PPAR and - / mrna expression after Studies with selective agonists and knockout mice demonstrate that PPAR regulates FA catabolism, and that PPAR controls lipid storage. The function of the ubiquitously expressed PPAR / remained elusive in major mass peripheral tissues (36, 37). In this investigation, we demonstrate that the PPAR / agonist, GW501516, induces the expression of genes involved in lipid absorption, preferential lipid utilization, -oxidation, cholesterol efflux, and energy uncoupling in skeletal muscle cells. Similar effects in lipid metabolism are observed after exercise training in human skeletal muscle (38). Interestingly, the PPAR agonist fenofibrate induces genes involved in fructose uptake, and glycogen formation in skeletal muscle. In contrast, rosiglitazone-mediated activation of PPAR induces gene expression associated with glucose uptake, FA synthesis and lipid storage. This study demonstrates that PPARs have distinct complementary, and contrasting roles in skeletal muscle with respect to the regulation of gene expression involved in lipid, carbohydrate and energy homeostasis (see Fig. 11). PPAR / has been implicated in fat (8 10) and bone (39) metabolism. Recently, the potent, synthetic and selective PPAR / agonist GW501516, a phenoxyacetic acid derivative, has been reported, which corrects hyperinsulinemia and hypertriglyceridemia in insulinresistant and obese primates (11). However, PPAR / target genes, target cells/tissues and mode of action remained unclear. Very recently, Evans and colleagues (16) demonstrated that expression of activated PPAR / in adipose tissue leads to a lean phenotype, with a normo-phagic diet. They showed the phenotype is associated with increased FA oxidation and energy uncoupling in adipose tissue. Our investigation demonstrates that the PPAR / agonist activates gene expression in skeletal muscle cells, which is involved in preferential lipid utilization,

13 Dressel et al. PPAR / Target Genes in Muscle Mol Endocrinol, December 2003, 17(12): Fig. 11. Schematic Overview of Metabolic PPAR / Action Enzymes and functions found to be activated by PPAR / agonist are marked in green. Red indicates inhibition of pathways. The green arrows underline FA uptake and oxidation, followed by uncoupling oxidation from ATP synthesis. FA catabolism and energy uncoupling. Muoio et al. (12) observed that skeletal muscle metabolism/ -oxidation and regulation of three well-characterized PPAR target genes UCP-3, CPT1, and PDK4 (in other tissues) were almost identical in skeletal muscle from either wild type or PPAR / mice (40 45). Furthermore, metabolism and gene expression in skeletal muscle were not compromised in PPAR / mice, in contrast to the dramatic deleterious effects in liver and heart tissue. Our data account for these observations and suggests PPAR / in peripheral tissues functions in a complimentary manner to PPAR in the liver and the heart. In the context of the data from Evans and colleagues (16) in adipose tissue, we provide further data that suggest that PPAR / targets skeletal muscle cells. Furthermore, our demonstration that M-CPT1 is preferentially regulated by PPAR /, and not PPAR in skeletal muscle cells illustrates distinct mechanisms exist within different cell types to implement localized responses to energy requirements and burdens in these tissues. The Oliver et al study (11) also demonstrated that GW raised cholesterol efflux in macrophages and serum HDL cholesterol. We demonstrate that GW activated ABCA1 mrna expression with the subsequent metabolic consequence of increased cholesterol efflux from skeletal muscle cells. Therefore, the effects of PPAR / agonists on skeletal muscle cell gene expression is entirely consistent with the beneficial impact of GW on dyslipidemia and hyperinsulinemia in obese primates, especially when one considers that muscle is a metabolically demanding tissue that accounts for 40% of the total body mass. Interestingly, in contrast to the role of PPAR in the liver and the heart, fenofibrate induces genes involved in fructose uptake, and glycogen formation in skeletal muscle. Furthermore, we observed a repression in SREBP-1c expression, similar to the effect observed in hepatic cells (46). However, gene expression involved in preferential FA catabolism was not activated by the PPAR agonist. In congruence with the observations that -oxidation in skeletal muscle was not compromised in PPAR / mice. Furthermore, the induction of fructose uptake by fenofibrates is in accordance with the amelioration of high fructose induced insulin resistance, fat accumulation and hyperlipidemia in rats by fenofibrate treatment (47). This suggests that PPAR /, not PPAR, regulates lipid catabolism in skeletal muscle cells. We speculate the / isoform also activates FA oxidation in skeletal muscle, in vivo. Rosiglitazone-mediated activation of PPAR induces gene expression associated with glucose uptake, FA synthesis, and lipid storage, consistent with previous studies. We did not observe robust changes in gene expression after agonist treatment. Thiazolidinediones induce dramatic changes in diseased, not normal healthy animals (48).