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, 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 atherogenic 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 multi-layered 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) α, -β/δ, and -γ [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 PPARβ/δ, the putative natural agonists are prostanoids, which are produced by the regulated conversion of poly-unsaturated FAs. In addition, PPARα, -β/δ, and -γ 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).
PPARα and -γ are predominantly expressed in liver and adipose tissue, respectively. The expression of PPARβ/δ 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 PPARα and -γ isoforms in lipid and carbohydrate metabolism. For example, PPARγ promotes adipogenesis and increases lipid storage. In contrast, PPARα 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 PPARα and PPARγ, respectively.
Relatively less is known about PPARβ/δ, which has been implicated in bone and fat metabolism (8–10). Recently, the potent, synthetic and selective PPARβ/δ 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 PPARβ/δ. GW501516 corrects hyperinsulinemia in insulin-resistant and obese primates. Furthermore, it raises ABCA1 mRNA expression, and serum HDL cholesterol, while lowering triglycerides. However, PPARβ/δ 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 PPARβ/δ agonists on whole body cholesterol homeostasis, lipid metabolism, target tissues and mode of action remain unclear (10).
The PPARα-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 PPARα knockout would have major consequences on skeletal muscle fuel metabolism and gene expression. However, Muoio et al. (12) observed the skeletal muscle metabolic/β-oxidation phenotype was not compromised in PPARα−/− mice, in contrast to the dramatic deleterious effects in liver and heart tissue. A plausible hypothesis suggests that PPARβ/δ 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. PPARβ/δ is the most abundant PPAR in muscle tissue (12–14). It was first implicated in FA metabolism from studies using the knockout animals. Most PPARβ/δ−/− 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 PPARβ/δ 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 GW501516 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 PPARβ/δ 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 PPARα−/− 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 PPARβ/δ. Surprisingly, the fundamental role of PPARβ/δ 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 PPARβ/δ in skeletal muscle, and to investigate the genes and regulatory genetic programs activated by PPARβ/δ involved in the control of lipid and energy homeostasis.
In summary, we demonstrate that PPARβ/δ directly and/or indirectly regulates genes involved in triglyceride-hydrolysis and FA oxidation [LPL, acyl-coenzyme A (CoA) synthetase 4 (ACS4), carnitine-palmitoyl-transferase (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 PPARβ/δ in skeletal muscle, in a PPARγ coactivator-1 (PGC-1)-dependent manner. In summary, we show that PPARβ/δ 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 PPARβ/δ agonists previously described in obese primates (11), and we speculate that PPARβ/δ agonists would have therapeutic utility against a high-fat diet and obesity.
RESULTS
GW501516 Is a Potent Agonist for PPARβ/δ in Skeletal Muscle C2C12 Cells
The yet unclear distinct physiological role of PPARβ/δ 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 PPARβ/δ cDNA into the expression vector pSG5 as a tool for our studies. Furthermore, to verify the integrity of the cloned PPARβ/δ constructs after sequencing, and the efficacy of the PPARβ/δ agonist GW501516 (11), we used the GAL4-hybrid system. Full-length PPARβ/δ, the C-terminal D/E-domain (that encodes the ligand-binding-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.
Fig. 1. Open in new tabDownload slide GW501516 Is a Potent Agonist in Muscle- and Nonmuscle Cells A, Schematic representation of the PPARβ/δ constructs used in this study. pSG5-PPARβ/δ encoding for full-length PPARβ/δ, encompassing the entire coding region [440 amino acids (aa)]; pSV40-GAL4-PPARβ/δ, encoding the GAL4-DNA-binding domain fused to full-length PPARβ/δ; pSV40-GAL4-PPARβ/δ-LBD, encoding the GAL4-DBD fused to the Hinge/LBD-Region (aa 138–440); pSV40GAL4-PPARβ/δ-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 GW501516 (1 μM), or DMSO as control. LUC activity is shown as relative light units (RLU).
We cotransfected the GAL4-PPARβ/δ-LBD together with the G5E1b-LUC reporter into CV1-cells and examined the dose-dependent activation by GW501516. A maximum activity is reached at approximately 150 nM (with an EC50 ∼20 nM), in agreement with the study of Oliver et al. (11) (who reported an EC50 of 24 nM for mouse PPARβ/δ), thereby validating the potency and efficacy of our agonist preparation, and the integrity of the PPARβ/δ-cDNA (data not shown). GW501516 (1 μM) was used in all latter experiments, as used in all cell culture experiments by Oliver et al. (11) to induce a reproducible maximal PPARβ/δ response. Moreover, at this concentration Oliver et al. demonstrated that GW501516 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 PPARβ/δ-constructs in CV1 (data not shown) and C2C12 cells (Fig. 1B). We cotransfected the various GAL4-PPARβ/δ 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 GW501516 (1 μM). The AF1-domain of PPARβ/δ inefficiently activated the LUC reporter, and did not respond to agonist treatment. In contrast, both GAL4-PPARβ/δ, and GAL4-PPARβ/δ-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 PPARβ/δ agonist GW501516 to activate a PPAR-dependent reporter (PPRE) in muscle cells. Moreover we examined the ability of the cofactors PGC-1, p300 and SRC-2/GRIP-1 to coactivate GW501516 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 PPARβ/δ expression vector (because these cells contain endogenous PPARβ/δ). As shown in Fig. 2A the PPARβ/δ agonist GW501516 activated the expression of the PPRE-containing 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, GW501516-dependent PPRE activation was enhanced when PGC-1, relative to p300 and SRC-2/GRIP-1, was cotransfected. For example, GW501516 activated the expression of the PPRE-reporter approximately 2.5-fold, and approximately 4-fold in the presence of the RXR agonist.
Fig. 2. Open in new tabDownload slide PGC-1 Acts as a Transcriptional Coactivator for PPARβ/δ in Muscle and Nonmuscle Cells A, Skeletal muscle C2C12 cells which endogenously express PPARβ/δ 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 μM; PPARβ/δ, GW501516, 1 μM), 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-PPARβ/δ, and cDNA-PGC-1 in the presence, or absence of GW501516. 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 [PPARα: Fenofibrate (FF); 100 μM, PPARγ: Rosiglitazone (Rosi); 10 μM, 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.
Subsequently, we validated the ability of the cofactor PGC-1 to coactivate GW501516-PPARβ/δ-dependent transactivation of the PPAR-dependent PPRE-containing reporter in nonmuscle CV-1 cells. Therefore, we cotransfected PPRE-tk-LUC, pSG5-PPARβ/δ and the expression vector encoding PGC-1 in the presence or absence of the PPARβ/δ agonist into nonmuscle CV1-cells (Fig. 2B). A PPARβ/δ- and GW501516-dependent, 4-fold activation of the PPRE-tk-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 PPARβ/δ constructs, and verify the potent and efficacious function of the GW501516 agonist preparation in nonmuscle and skeletal muscle cells. Furthermore, it demonstrates the selective coactivation of PPARβ/δ-mediated gene expression by PGC-1 in skeletal muscle cells.
In addition, we further examined the ability of PPARα and PPARγ 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 compare the relative effects of PPARα, -β/δ, and -γ agonists on the expression of the genes involved in skeletal muscle lipid and carbohydrate metabolism. Figure 2, C and D, clearly demonstrates that PPARα and -γ agonists efficiently regulate gene expression in skeletal muscle cells. In addition, we show that SRC-2/GRIP-1 and PGC-1 selectively coactivate PPARα and -γ, respectively, in skeletal muscle cells (Fig. 2, C and D). Finally, these experiments demonstrate that the C2C12 skeletal muscle cells express functional PPARα, -β/δ, and -γ 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 GW501516 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 GW501516 and Rosiglitazone inducibility, and the absolute level of PPRE-dependent expression. In contrast, SRC-2/GRIP1 expression preferentially increases Fenofibrate-mediated activation.
Regulation of Gene Expression in Skeletal Muscle Cells by PPARα, -β/δ, and -γ 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 PPARα, -β/δ, and -γ 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, T0901317 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 PPARβ/δ-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 PPARα, -β/δ, and -γ 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 PPARα and β/δ are expressed at similar levels in 96 h differentiated myotube cells. PPARγ mRNA is abundantly expressed, however, the primers reflect mRNA expression from all three PPARγ isotypes (i.e. γ1 + γ2 + γ3).
Fig. 3. Open in new tabDownload slide PPARβ/δ 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 μM), PPARβ/δ (GW501516; 1 μM), 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 32P-radiolabeled cDNAs for key-indicators for myogenic differentiation (myogenin, sarcomeric α-actin, cytoskeletal β-, and γ-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 PPARα, -β/δ, and -γ. Expression levels are normalized to GAPDH. The primers for PPARγ detect all isoforms for PPARγ.
Table 1.Key Target Genes in this Study Open in new tab
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 PPARβ/δ (GW501516; 1 μM), RXR (LG101305; 100 nM), PPARα (fenofibrate; 100 μM), PPARγ (Rosiglitazone; 10 μM) or the vehicle [dimethylsulfoxide (DMSO)] on these predifferentiated myotubes. We isolated total RNA from differentiated myotubes, which were treated with GW501516 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 β-, γ-actin, and the activation of the sarcomeric α-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 PPARβ/δ agonists (LG101305 or GW501516, respectively), nor cotreatment with both agonists. Similarly, PPARα and -γ 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).
Fig. 4. Open in new tabDownload slide Regulation of Gene Expression in Skeletal Muscle Cells by PPARα, -β/δ, and -γ Agonists Differentiated myotubes were treated as described in Fig. 3A for 24 h with agonists for RXR (LG101305; 100 nM), PPARβ/δ (GW501516; 1 μM), PPARα (Fenofibrate; 100 μM), PPARγ (Rosiglitazone; 10 μM), 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.
The majority of candidate target genes investigated showed modest response (<2-fold) to either the RXR, or the PPARβ/δ agonist alone. However, the obvious exceptions were the mRNAs for UCP-1 and -2, which encode UCPs involved in thermogenesis and energy-expenditure (Fig. 4G). UCP-1 and -2 were induced approximately 7- and 4-fold, respectively, by treatment with the PPARβ/δ agonist GW501516. More, importantly, these mRNAs were not induced by agonists for PPARα or -γ. 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 GW501516 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 PPARβ/δ agonist. ADRP 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 PPARβ/δ 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, PPARβ/δ 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 PPARβ/δ and RXR agonists, are highlighted by the observation that the cotreatment with agonists for PPARα and -γ in the presence of an RXR agonist does not induce LPL and CPT1 expression (Fig. 4C). Interestingly, PDK4 mRNA was activated by PPARα, -β/δ, and -γ agonists. The glucose, and fructose transporters (Glut-4 and -5) were induced by rexinoid treatment, but completely refractory to the PPARβ/δ agonist (Fig. 4D). In contrast, we observed Glut-4, and -5 were activated by PPARγ and PPARα agonists, respectively.
As mentioned earlier, cotreatment with agonists for PPARβ/δ 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 PPARβ/δ and RXR agonists, are further highlighted by the observation that the cotreatment with the agonists for PPARα and -γ 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 GW501516. Cotreatment did not lead to further activation. Finally ADRP/adipophilin was synergistically induced approximately 7-fold by PPARβ/δ and RXR agonist cotreatment. Furthermore, ADRP mRNA increased approximately 2- to 3-fold to treatment with PPARβ/δ and -γ, 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 PPARβ/δ 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.
Fig. 5. Open in new tabDownload slide 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 PPARβ/δ agonists, as described in Fig. 3 was analyzed using Northern analysis. After blotting, RNA was hybridized with 32P-radiolabeled cDNAs encoding GAPDH, UCP-2 and -3, and LPL. B, Myotubes differentiated for 4 d (MT4s) were subsequently treated with GW501516 (1 μM) for the indicated time points. Cells treated for 24 h with DMSO (MT5s) were used as control.
Furthermore, we explored whether some of the significant effects we observed only after cotreatment with GW501516 and LG101305 were specific to the PPARβ/δ 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, T0901317. Clearly, the LXR agonist does not activate the expression of mRNAs encoding ACS4, glycogenin-1 and ADRP (Fig. 5, C–E). However, LXR (a demonstrable efficacious and potent ABCA1 activator) dramatically induced ABCA1 mRNA expression in the absence and presence of RXR agonists (Fig. 6, A and B).
Fig. 6. Open in new tabDownload slide 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 μM), LXR (T0901317; 1 μM), 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. Open in new tabDownload slide PPARβ/δ 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), PPARβ/δ (GW501516; 1 μM), both together (LG & GW), LXR (T0901317; 1 μM), 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, PPARβ/δ-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.
Obviously, LXR is a more potent activator of ABCA1 mRNA expression than PPARβ/δ. However, Oliver et al. (11) demonstrated also that GW501516 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 GW501516 promoted ApoA1-dependent reverse efflux in skeletal muscle cells, relative to the LXR agonist, T0901317 (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 PPARβ/δ specific agonists. The LXR agonist, T0901317 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 T0901317, the PPARβ/δ 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 PPARβ/δ 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 PPARβ/δ agonist in macrophages.
In summary, we observed that the PPARβ/δ agonist GW101516 dramatically activates the mRNAs encoding the UCPs, suggesting that PPARβ/δ 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 formation in skeletal muscle. In contrast, rosiglitazone-mediated activation of PPARγ 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 PPARβ/δ in Skeletal Muscle
We further explored the molecular basis of PPARβ/δ-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 PPARβ/δ 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 GW501516 and/or LG101305. Interestingly, only the M-CPT1 promoter responded to the PPARβ/δ 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 PPARβ/δ agonist, GW501516 (data not shown).
Fig. 8. Open in new tabDownload slide The M-CPT1 Promoter Is Activated by PPARβ/δ in a PGC-1-Dependent Manner A, pGL2-basic, or pGL2-MCPT1(-1025/-12) were transfected into skeletal muscle C2C12-cells which endogenously express PPARβ/δ and subsequently treated with agonists for RXR (LG101305), PPARβ/δ (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 PPARβ/δ were transfected with PPRE-tk-LUC, pSG5-PPARβ/δ, cDNA-PGC-1 in the absence or presence of agonists for PPARβ/δ 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).
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 GW501516 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 PPARβ/δ 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 PPARβ/δ. The M-CPT1 promoter was cotransfected into CV1 cells with PPARβ/δ, PGC-1, or both together. As seen in Fig. 8C, PPARβ/δ 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 PPARβ/δ, and selectively coactivated by PGC-1. Cofactor expression increases GW501516 inducibility and the absolute levels of CPT1 expression.
To rigorously define that M-CPT1 was regulated by GW501516 in a PPARβ/δ-PPRE-dependent manner, we mutated the previously defined PPARα 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 PPARβ/δ-specific agonist (Fig. 9A). This demonstrated that GW501516 mediated activation is dependent on the M-CPT1 PPRE. This element was previously defined as a PPARα-regulated motif in cardiac muscle and primary cardiomyocytes (27, 34).
Fig. 9. Open in new tabDownload slide M-CPT1 Is a Direct Target of PPARβ/δ 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), PPARβ/δ (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), PPARβ/δ (GW501516), PPARγ (rosiglitazone, 10 μM), PPARα (Wy14634, 10 μM), or the vehicle (DMSO). Reporter activity is shown as RLU.
Consequently, we examined the ability of the synthetic and selective PPARα, -β/δ, and -γ 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 PPARβ/δ and not the PPARα agonist (Fig. 9B/C).
In summary, M-CPT1 is an established target for PPARα in cardiac muscle (27, 34, 35). However, we clearly demonstrate by transfection in the presence of PPARα, -β/δ, and -γ agonists that the M-CPT1 promoter in skeletal muscle cells responds preferentially to PPARβ/δ agonist activation (and not a selective PPARα agonist) (Fig. 9, B and C). This is reminiscent of the differential cell specific regulation of the PPRE in the LPL promoter by PPARα and -γ 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 PPARβ/δ-specific agonist (Fig. 9, A–C). This demonstrated that GW501516-mediated activation was dependent on the M-CPT1 PPRE. This element was previously defined as a PPARα-regulated motif in cardiac muscle (27, 34, 35).
Moreover, we demonstrate that the native LPL promoter responds to PPARα, not PPARβ/δ, agonists in muscle cells in the absence of exogenous PPARα (Fig. 10A), further validating the specificity of the PPARβ/δ response on the CPT1 promoter in skeletal muscle cells. The previous literature demonstrates that these cells express functional PPARα, β/δ, and -γ receptors (16, 21–23). In addition, our data demonstrate the multimerized DR-1 PPRE reporter is efficiently activated by PPARα, -β/δ, and -γ 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 PPARβ/δ and not -α agonists, is not due to lack of PPARα expression, and/or nonfunctional PPARα. Finally, to rigorously demonstrate that the selective activation of the M-CPT1 promoter is not due to the elevated expression of PPARβ/δ, relative to PPARα mRNA, after agonist treatment we examined the expression of PPARα and -β/δ mRNA expression after 24 h RXR agonist, PPARβ/δ agonist, and cotreatment (Fig. 10B). As observed earlier, PPARα and -β/δ mRNAs are similarly expressed relative to GAPDH mRNA before agonist treatment; however, RXR and PPARβ/δ agonist treatment preferentially induced PPARα mRNA expression (Fig. 10B). This definitively demonstrates that the selective activation of the M-CPT1 promoter by the PPARβ/δ agonist (and not the PPARα agonist) in skeletal muscle cells is not due to lack of PPARα expression.
Fig. 10. Open in new tabDownload slide PPARα 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 PPARα (fenofibrates), PPARγ (rosiglitazone), PPARβ/δ (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), PPARβ/δ (GW501516), or cotreatment (LG & GW) was analyzed by quantitative real-time PCR for the expression of PPARα and -β/δ mRNAs. Expression levels are normalized to GAPDH. C, Differentiated myotubes were treated for 8 h with the agonist for PPARβ/δ (GW501516) in the absence or presence of cycloheximide (10 μg/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.
Lastly, we observe that the induction of M-CPT1 mRNA expression by GW501516 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 GW501516 directly regulates the M-CPT1 promoter in a PPARβ/δ/PPRE-dependent manner.
DISCUSSION
Studies with selective agonists and knockout mice demonstrate that PPARα regulates FA catabolism, and that PPARγ controls lipid storage. The function of the ubiquitously expressed PPARβ/δ remained elusive in major mass peripheral tissues (36, 37). In this investigation, we demonstrate that the PPARβ/δ agonist, GW501516, induces the expression of genes involved in lipid absorption, preferential lipid utilization, β-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 PPARα agonist fenofibrate induces genes involved in fructose uptake, and glycogen formation in skeletal muscle. In contrast, rosiglitazone-mediated activation of PPARγ 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).
Fig. 11. Open in new tabDownload slide Schematic Overview of Metabolic PPARβ/δ Action Enzymes and functions found to be activated by PPARβ/δ agonist are marked in green. Red indicates inhibition of pathways. The green arrowsunderline FA uptake and oxidation, followed by uncoupling oxidation from ATP synthesis.
PPARβ/δ has been implicated in fat (8–10) and bone (39) metabolism. Recently, the potent, synthetic and selective PPARβ/δ agonist GW501516, a phenoxyacetic acid derivative, has been reported, which corrects hyperinsulinemia and hypertriglyceridemia in insulin-resistant and obese primates (11). However, PPARβ/δ target genes, target cells/tissues and mode of action remained unclear. Very recently, Evans and colleagues (16) demonstrated that expression of activated PPARβ/δ 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 PPARβ/δ agonist activates gene expression in skeletal muscle cells, which is involved in preferential lipid utilization, FA catabolism and energy uncoupling. Muoio et al. (12) observed that skeletal muscle metabolism/β-oxidation and regulation of three well-characterized PPARα target genes UCP-3, CPT1, and PDK4 (in other tissues) were almost identical in skeletal muscle from either wild type or PPARα−/− mice (40–45). Furthermore, metabolism and gene expression in skeletal muscle were not compromised in PPARα−/− mice, in contrast to the dramatic deleterious effects in liver and heart tissue. Our data account for these observations and suggests PPARβ/δ in peripheral tissues functions in a complimentary manner to PPARα 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 PPARβ/δ targets skeletal muscle cells. Furthermore, our demonstration that M-CPT1 is preferentially regulated by PPARβ/δ, and not PPARα 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. 2001 study (11) also demonstrated that GW501516 raised cholesterol efflux in macrophages and serum HDL cholesterol. We demonstrate that GW501516 activated ABCA1 mRNA expression with the subsequent metabolic consequence of increased cholesterol efflux from skeletal muscle cells. Therefore, the effects of PPARβ/δ agonists on skeletal muscle cell gene expression is entirely consistent with the beneficial impact of GW501516 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 PPARα 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 PPARα agonist. In congruence with the observations that β-oxidation in skeletal muscle was not compromised in PPARα−/− 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 PPARβ/δ, not PPARα, regulates lipid catabolism in skeletal muscle cells. We speculate the β/δ isoform also activates FA oxidation in skeletal muscle, in vivo.
Rosiglitazone-mediated activation of PPARγ 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).
We demonstrate that PPARβ/δ agonists have a significant role in the regulation of the mRNAs encoding the UCPs (UCP-1 to -3, mitochondrial proton carriers) that control metabolic efficiency, energy expenditure, adaptation to nutrient (i.e. preferential lipid utilization) and thermogenesis by uncoupling oxidation/respiration from ATP synthesis (see Fig. 11). These data are consistent with the observations that mRNA expression associated with selective utilization of lipid substrates is augmented during exercise, starvation, and physiological states that are associated with increased systemic delivery and utilization of FAs. In adult organisms, UCP-1 is almost exclusively expressed in brown adipose tissue, whereas UCP-2 and UCP-3 are expressed in adipose and skeletal muscle tissue, although UCP-3 is predominantly found in muscle. However, UCP-1 mRNA expression has been observed previously in muscle cells (24) and may be an artifact of the myogenesis (muscle differentiation) program in cell culture, and/or reflect an expression profile associated with differentiation during embryogenesis.
A number of studies have investigated ectopic and muscle specific overexpression of UCP-2 and -3 in transgenic mice, and in cell culture. These investigations have reported: 1) reduced metabolic efficiency and increased rates of energy expenditure; 2) preferential FA oxidation vs. glucose utilization/oxidation; 3) resistance to high fat diet induced weight gain, and obesity in the context of hyperphagic behavior; 4) lower fasting plasma glucose and insulin levels, and increased glucose tolerance and clearance rate; and 5) adaptive thermogenesis. These studies emphasize the regulatory role of UCP-2 and -3 in metabolic efficiency/energy expenditure, thermogenesis, and in preferential substrate utilization (49–55). Similarly, when UCP-1 (normally expressed in brown adipose tissue) is overexpressed in the muscle, the transgenic mice have a lower body weight, increased food intake accompanied by energy expenditure (56). We hypothesize that the effects of PPARβ/δ agonists on skeletal muscle, a major mass peripheral tissue would have utility and protect against diet-induced obesity and glucose intolerance.
In addition, we confirm that M-CPT1 (and the PPRE), an established target for PPARα in cardiac muscle (27, 34, 35) responds preferentially to PPARβ/δ agonist activation (and not selective PPARα agonist) in skeletal muscle cells. This is reminiscent of the differential cell specific regulation of the LPL promoter by PPARα and -γ agonists in adipose and liver. This observation highlights that PPARs have distinct tissue specific functions, and that a single DNA motif can mediate a cell-specific transcriptional phenotype. Furthermore, it suggests that PPARα and PPARβ/δ function in a complementary but tissue-specific manner.
Furthermore, we demonstrate the PGC-1-dependent nature of muscle-CPT1 transcriptional activation, and PPRE trans-activation by GW501516-PPARβ/δ in skeletal muscle cells. The studies by Spiegelman and colleagues (31, 57) exquisitely demonstrate that PGC-1 regulates adaptive thermogenesis, and mitochondrial biogenesis. Moreover, PGC-1 is regulated by exercise (57, 58). Our studies in cells are in consonance with these in vivo studies.
During the preparation of this manuscript, a number of manuscripts appeared in the literature describing the role of PPARβ/δ agonists in cardiac myocytes and macrophages. For example, Gilde et al. (59) published work describing the role of PPARβ/δ in cardiac lipid metabolism. They demonstrated that the long chain FA induced regulation of gene expression in primary cardiomyocytes is controlled by PPARα and PPARβ/δ. Lipid catabolism was activated in response to PPARα and -β/δ agonists, concluding that PPARα and PPARβ/δ have overlapping functions in the control of cardiac lipid homeostasis. Our investigation, and the study by Wang et al.(16) suggest that PPARα and -β/δ have distinct functions in adipose and muscle. Moreover, work from Vosper et al. (1) demonstrated that PPARβ/δ agonists induce lipid absorption and storage in macrophages. Paradoxically, ApoE and cyp27 mRNA expression is repressed, and in contrast, ABCA1 mRNA expression and ApoA1-dependent cholesterol efflux is induced. This is entirely consistent with the observations reported in this study. Furthermore, increased lipid absorption and ApoA1-dependent cholesterol efflux in macrophages and skeletal muscle cells are entirely consistent with the profound effects observed by Oliver et al. (11) in lowering circulating levels of triglycerides and LDL, with a corresponding increase in HDL cholesterol. We comment that this compound is in clinical trials, and the Oliver et al.2001 manuscript reported that the beneficial effects of GW501516 on HDL cholesterol and triglycerides were observed at 1 and 3 mg/kg with corresponding circulating serum concentrations of approximately 265 and 700 ng/ml, or 0.5 and 1.5 μM, respectively. This is consistent with the concentration used in this study. Finally, Chawla et al. (17) state that triglyceride-enriched VLDLs activate PPARβ/δ in macrophages and lead to the induction of ADRP/adipophilin (lipid storage droplets), which is consistent with the induction of ADRP mRNA expression we observed in skeletal muscle cells after agonist treatment. In conclusion, we suggest the activation of PPARβ/δ in skeletal muscle cells programs a cascade of gene expression designed to activate catabolism, and energy expenditure.
MATERIALS AND METHODS
Plasmids
Mouse PPARβ/δ was reverse transcribed from differentiated C2C12 myotubes, using total RNA, and the following primers: 5′-full-length and -AF1: GCGGGATCCTCACCATGGAACAGCCACAGGAGGAGACC (BamHI); 5′-hinge/LBD: GCGGGATCCTCACCATGTCGCACAACGCTATCCGC (BamHI); 3′-full-length and hinge/LBD: GCGGGATCCTTAGTACATGTCCTTGTAGATTTCC (BamHI); 3′-AF1: GCGTCTAGACATGTTGAGGCTGCCGCCTGAGGCC (XbaI). The PCR products were cut with the respective enzymes and cloned into the corresponding sites of pSG5, and/or pSV40-GAL0. CDNA4-PGC-1 was kindly provided by B. M. Spiegelman (32). PPRE-tkLUC was kindly provided by X-Ceptor Therapeutics. The CPT1 promoter (−1025/−12) was kindly provided by D. P. Kelly (27). cDNA-GRIP1/SRC2 and cDNA-p300 have been described previously (60).
RNA Extraction and Northern Hybridization
Total RNA used for RT-PCR cloning and Northern blot analysis was extracted using TRI-Reagent (Sigma Aldrich Australia Pty Ltd, Castle Hill, New South Wales, Australia), according to manufacturer’s protocol. For quantitative real-time RT-PCR, RNA was further purified using RNeasy (QIAGEN, Clifton Hill, Victoria, Australia), after manufacturer’s instructions. Northern blot hybridization was carried out as described previously (61).
Cell Culture, Transient Transfections, and Cholesterol Efflux Assay
Mouse myogenic C2C12 cells were cultured in growth medium [DMEM supplemented with 10% Serum Supreme (BioWhittaker, Edward Keller Pty Ltd, Hallam, Victoria, Australia)] in 6% CO2. For differentiation assays, cells were grown to confluency, at which point media was changed into differentiation medium (DMEM supplemented with 2% horse serum). Cells were harvested at indicated time points. For drug assays, cells were differentiated into myotubes for 4 d (MT4s), and the medium was changed into phenol red-free differentiation medium supplemented with the agonists for PPARβ/δ (GW501516, 1 μM), RXR (LG101305, 100 nM), PPARα (Fenofibrate, 100 μM or Wyeth14643, 10 μM), PPARγ (Rosiglitazone, 10 μM), or the vehicle (DMSO) as control. Cells were harvested at the indicated time points (usually 24 h, if not indicated differently). African green monkey kidney CV1 cells were grown in DMEM supplemented with 10% heat-inactivated fetal calf serum.
Transfections were carried out using a DOTAP (N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methylsulfate) (Biontex Laboratories GmbH, Munich, Germany)/DOSPER (1,3-di-oleoyloxy-2-(6-carboxy-spermyl)-propylamid) (Roche Diagnostics Pty Ltd, Castle Hill, New South Wales, Australia) 3:1 liposome mixture in HEPES buffered saline [42 mM HEPES, 275 mM NaCl, 10 mM KCl, 0.4 mM Na2HPO4, 11 mM dextrose (pH 7.1)], with 1μg of total DNA per well. Medium was replaced after 16 h with the respective fresh medium, supplemented with agonists for PPARs and/RXR (as described above), harvested after 24 h and assayed for LUC activity using the Luclite kit (PerkinElmer Life Science, Knoxfield, Victoria, Australia) according to manufacturer’s protocol.
ApoA1-dependent cholesterol efflux was performed as described previously (19).
Quantitative RT-PCR
Target cDNA levels were quantitated by real-time RT-PCR using an ABI Prism 7700 Sequence Detector system utilizing SYBRE green I (Molecular Probes, Eugene, OR; catalog no. S-7562, used at 0.8×) as a nonspecific PCR product fluorescence label. Quantitation was over 45 cycles of 95 C for 15 sec and 60 C for 1 min two-step thermal cycling preceded by an initial 95 C for 2 min for activation of 0.75 U Platinum Taq DNA polymerase (Invitrogen Australia Pty Ltd, Mulgrave, Victoria, Australia). The 25-μl reaction also contained 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 5 mM MgCl2, 200 μM each of deoxy (d) GTP, dATP, dCTP, 400 μM deoxyuridine triphosphate, 0.5 U uracil-N-glycosylase, 500 nM ROX reference dye (Invitrogen) and 200 nM each forward and reverse primers. Mus musculus primer sequences (forward and reverse, respectively): ABCA1: GCTCTCAGGTGGGATGCAG, GGCTCGTCCAGAATGACAAC;
ABCG1: CTGAGGGATCTGGGTCTGA, CCTGATGCCACTTCCATGA;
ACS4: GGTTTGGTAACAGATGCCTTCAA, CCCATACATTCGCTCAATGTCTT; ADRP: CCCTGGTTCTAAGAAGCTGCTTT, GGCCAGATGACCCCTTTTG;
ApoE: GCTGTTGGTCACATTGCTGA, TGCCACTCGAGCTGATCTG;
CD36: GGCCAAGCTATTGCGACAT, CAGATCCGAACACAGCGTAGA;
CPT1: ATCATGTATCGCCGCAAACT, CCATCTGGTAGGAGCACATGG;
FABP3: CCCCTCAGCTCAGCACCAT, CAGAAAAATCCCAACCCAAGAAT;
FAS: CGGAAACTTCAGGAAATGTCC, TCAGAGACGTGTCACTCCTGG;
GAPDH: GTGTCCGTCGTGGATCTGA, CCTGCTTCACCACCTTCTTG;
Glut4: ATGGCTGTCGCTGGTTTCTC, ACCCATACGATCCGCAACAT;
Glut5: CTTGCCTTTACCGGGTTGAC, CATCTGGTCTTGCAGCAACTCT;
GYG1: CCCAAACCCCTCATCTGATG, GCACGTTTCCATACATAGTATGTGAA;
LPL: CCAATGGAGGCACTTTCCA, TGGTCCACGTCTCCGAGTC;
PDK2: TGCTCCGGCTTGCCTTAT, CACTCCATCCTTCTTAACATTGACA;
PDK4: AAAGGACAGGATGGAAGGAATCA, TTTTCCTCTGGGTTTGCACAT;
PPARα: TCTTCACGATGCTGTCCTCCT, GGAACTCGCCTGTGATAAAGC
PPARβ/δ: TCCAGAAGAAGAACCGCAACA, GGATAGCGTTGTGCGACATG;
PPARγ: CAGGCCGAGAAGGAGAAGCT, GGCTCGCAGATCAGCAGACT
SCD1: TGTACGGGATCATACTGGTTCC, CCCGGCTGTGATGCC;
SCD2: ACTGTGACTCAAGTTCAACTCTTGAAA, TGCCCACAAATTGAGGATAGC;
SREBP-1c: CGTCTGCACGCCCTAGG, CTGGAGCATGTCTTCAAATGTG;
UCP-1: ACAGAAGGATTGCCGAAAC, AGCTGATTTGCCTCTGAATG;
UCP-2: GTTCCTCTGTCTCGTCTTGC, GGCCTTGAAACCAACCA;
UCP-3: TGACCTGCGCCCAGC, CCCAGGCGTATCATGGCT.
Amplification specificity was verified by visualizing PCR products on an ethidium bromide-stained 2.5% agarose gel. GAPDH was used for normalization between samples for quantitation.
Acknowledgments
We thank Dr. Richard Heyman and X-Ceptor Therapeutics Inc. for kindly providing the PPARβ/δ, and RXR agonists, GW501516, and LG101305, respectively. Rosiglitazone was kindly provided by Thomas A. Gustafson (Metabolex Inc.). We thank Shayama Wijedasa and Rachel Burrow for excellent technical assistance.
U.D., T.L.A., and J.B.P. contributed equally to this study.
This work was supported by the National Health and Medical Research Council (NHMRC) of Australia. G.E.O.M. is an NHMRC Principal Research Fellow, and U.D. is a University of Queensland Postoctoral Research Fellow.
Uwe Dressel,Tamara L. Allen,Jyotsna B. Pippal,Paul R. Rohde,Patrick Lau,George E. O. Muscat. "The Peroxisome Proliferator-Activated Receptor β/δ Agonist, GW501516, Regulates the Expression of Genes Involved in Lipid Catabolism and Energy Uncoupling in Skeletal Muscle Cells" Molecular Endocrinology, Volume 17, Issue 12, 1 December 2003, Pages 2477–2493.