Peroxisome Proliferator-activated Receptor a Activates Transcription of the Brown Fat Uncoupling Protein-1 Gene A LINK BETWEEN REGULATION OF THE THERMOGENIC AND LIPID OXIDATION PATHWAYS IN THE BROWN FAT CELL*

,

High expression of the peroxisome proliferator-activated receptor ␣ (PPAR␣) differentiates brown fat from white, and is related to its high capacity of lipid oxidation. We analyzed the effects of PPAR␣ activation on expression of the brown fat-specific uncoupling protein-1 (ucp-1) gene. Activators of PPAR␣ increased UCP-1 mRNA levels severalfold both in primary brown adipocytes and in brown fat in vivo. Transient transfection assays indicated that the (؊4551)UCP1-CAT construct, containing the 5-regulatory region of the rat ucp-1 gene, was activated by PPAR␣ co-transfection in a dose-dependent manner and this activation was potentiated by Wy 14,643 and retinoid X receptor ␣. The coactivators CBP and PPAR␥-coactivator-1 (PGC-1), which is highly expressed in brown fat, also enhanced the PPAR␣dependent regulation of the ucp-1 gene. Deletion and point-mutation mapping analysis indicated that the PPAR␣-responsive element was located in the upstream enhancer region of the ucp-1 gene. This ؊2485/؊2458 element bound PPAR␣ and PPAR␥ from brown fat nuclei. Moreover, this element behaved as a promiscuous responsive site to either PPAR␣ or PPAR␥ activation, and we propose that it mediates ucp-1 gene up-regulation associated with adipogenic differentiation (via PPAR␥) or in coordination with gene expression for the fatty acid oxidation machinery required for active thermogenesis (via PPAR␣).
The peroxisome proliferator-activated receptor ␣ (PPAR␣) 1 is a fatty acid-activated transcription factor that plays a key role in the transcriptional regulation of genes involved in cellular lipid metabolism (1). PPAR␣ together with PPAR␥ and PPAR␦/␤ belong to a subgroup of the nuclear hormone receptor superfamily that heterodimerizes with the 9-cis-retinoic acid receptors (RXRs) (2)(3)(4)(5). The PPAR-RXR heterodimer binds to specific response elements (PPREs), which consist of a direct repeat of the consensus half-site motif spaced by one nucleotide (DR-1) (6). Fatty acids, peroxisome proliferators, and fibrate hypolipidemic drugs can activate PPAR␣ (1,4), and natural (leukotrine B 4 ) or synthetic (fibrate Wy 14,643) specific ligands for PPAR␣ have been identified (7). In contrast, 15-deoxy-⌬ 12,14 -prostaglandin J 2 and thiazolidinedione antidiabetic agents are selective ligands for PPAR␥ (8 -10). In addition to ligand selectivity, PPAR subtypes have been involved in different biological functions. PPAR␣ is mostly expressed in tissues with high rates of fatty acid oxidation and peroxisomal metabolism, such as brown fat, liver, or heart (1,11). Recent studies of PPAR␣-null mice have confirmed that PPAR␣ is necessary in vivo for hepatic fatty acid oxidation and ketone body synthesis during starvation (12). PPAR␦, which is ubiquitously expressed, seems to be involved in basic lipid metabolism (11). High expression of PPAR␥ is mainly restricted to white (WAT) and brown (BAT) adipose tissue (13). Hence, in contrast to the role of PPAR␣ in cellular lipid catabolism, PPAR␥ regulates adipogenesis (i.e. lipid deposition) (13,14).
BAT is a major site for nonshivering thermogenesis in mammals. Its thermogenic capacity relies on the presence of an inner mitochondrial protein uniquely expressed in brown adipocytes, the uncoupling protein (UCP) (15), now referred to as UCP-1 since the discovery of the more widely expressed UCP-2 and UCP-3 (for review, see Ref. 16). Brown fat thermogenesis is mainly controlled by norepinephrine released from sympathetic terminals innervating the tissue, although nuclear receptor-mediated pathways have also been described. Thus, activation of PPAR␥ promotes HIB-1B brown adipocyte differentiation (17), and up-regulates ucp-1 gene expression (18). Furthermore, we demonstrated that retinoic acid is a powerful inducer of ucp-1 gene transcription, acting through retinoic acid receptors and RXRs (19,20). The 5Ј-flanking region of the rat ucp-1 gene contains the proximal regulatory promoter, including C/EBP-regulated sites (21) and the main cAMP-regulatory element (22), and an upstream enhancer involved in complex regulation by retinoic acid receptors, RXR, and thyroid hormone nuclear receptors (19,20,23). A site responsive to PPAR␥ activators has also been located in the upstream enhancer of the murine ucp-1 gene (18).
BAT highly coexpresses not only PPAR␥ and PPAR␦ subtypes but also PPAR␣ (24). BAT stores triglycerides but, in contrast to WAT, it uses lipids as oxidative substrates to generate heat. Since PPAR␣ induces the expression of fatty acid oxidation enzymes in tissues other than BAT (6), it may do so in BAT in association with thermogenic requirements. Here we report that PPAR␣ activators induce ucp-1 gene expression in brown adipocytes and in BAT in vivo, acting through a PPRE located in the upstream enhancer of the ucp-1 gene that is also responsible for PPAR␥-dependent regulation. PPAR␣ is proposed to coordinate the activation of lipid oxidation and thermogenic activity in brown fat.
Cell Culture-Primary culture of differentiated brown adipocytes was performed as described previously (19), and grown in 5 ml of Dulbecco's modified Eagle's medium-Ham's F-12 medium (1:1) supplemented with 10% fetal calf serum, 20 nM insulin, 2 nM 3,5,3Ј-triiodothyronine, and 100 M ascorbate. Experiments were performed on day 9 of culture when 80 -90% of the cells were considered to be differentiated on the basis of lipid accumulation and acquisition of brown adipocyte morphology. Brown adipocytes were exposed to 10 M Wy 14,643 for 24 h, or at the concentrations and times indicated in the experiments. Cells were also exposed to various PPAR agonists for 24 h, except for 15-deoxy-⌬ 12,14 -prostaglandin J 2 which was added at a final concentration of 10 M for 6 h. As indicated, cycloheximide (Sigma) was used at a dose of 5 g/ml as reported (19).
RNA Isolation and Northern Blot Analysis-Total RNA was extracted using the RNeasy Mini Kit (Quiagen). Northern blot analysis and hybridization were carried out as described (24). Blots were hybridized using as probes the full-length cDNA for rat UCP-1 (25) and 0.5 kb of the cDNA for mouse mitochondrial-genome-encoded cytochrome oxidase subunit II (COII) (26), which was used as a control. Hybridization signals were quantified using Molecular Image System GS-525 (Bio-Rad). Statistical analysis was performed by Student's t test.
Transfection Assays-Murine primary brown adipocytes differentiated in culture were transiently transfected by the calcium phosphate precipitation method on day 9 of culture (22). Each transfection con-tained 12 g of (Ϫ4551)UCP1-CAT and included or not 3 g of the expression vector pSG5-PPAR␣. When indicated 10 M Wy 14,643, 10 M BRL 49653, or 30 M Ly171883 was added after transfection. 1 g of cytomegalovirus-␤-galactosidase was also included to assess the efficiency of separate transfections. The cells were incubated for 24 h and, for each condition, at least three plates were pooled.
HepG2 and HIB-1B cells were transfected using the FuGENE 6 Transfection Reagent (Roche Molecular Biochemicals) for 16 h and cells were harvested 24 h later. Unless otherwise indicated, each transfection contained between 0.5 and 1 g of UCP1-CAT vector, 0.1 g of cytomegalovirus-␤-galactosidase, and included or not 0.3 g of pSG5-PPAR␣ or pSG5-PPAR␥ expression vector, and/or 0.1 g of pRSV-RXR␣. When indicated, 0.1 g of the expression vector pCMX-CBP or pSV-PGC1 was added.
Analysis of CAT activity was determined by thin layer chromatography (22) and quantified by radioactivity counting (AMBIS). The amount of cell extract used was adjusted to maintain a percentage conversion of chloramphenicol between 1 and 20%. The CAT activity was normalized for variation in transfection efficiency using the ␤-galactosidase activity measured for each sample as a standard.
DNA Binding Experiments-Nuclear proteins were isolated from rat BAT or differentiated primary brown adipocytes as described elsewhere (21,22). cDNAs for mPPAR␣, mPPAR␥, and hRXR␣ were transcribed and translated in vitro by using the TNT Quick Coupled Transcription/ translation Systems (Promega) according to the manufacturer's instructions.
Tissue Samples-BAT was extracted from two-month-old female, 15-day lactating, or newborn Swiss mice. Adult mice were treated with a single intraperitoneal injection of Wy 14,643 (50 g/g body weight) or troglitazone (100 g/g body weight) in 50% dimethyl sulfoxide/saline. Controls were given equivalent volumes of the vehicle and mice were studied 6 h after injections. Neonates were placed in a humidified thermostated chamber at 28°C, and injected intraperitoneally 2 h after birth with Wy 14,643 (50 g/g body weight), BRL 49653 (50 g/g body weight), or equivalent volumes of the 20% dimethyl sulfoxide/saline vehicle solution. Pups were studied 15 h after treatment.

Activators of PPAR␣ Induce the Expression of the ucp-1 Gene in Differentiated Brown
Adipocytes-To analyze whether PPAR␣ agonists modulate the expression of the ucp-1 gene, primary cultures of murine brown adipocytes were used since they express all three PPAR subtypes (24). As shown in Fig. 1, exposure of brown adipocytes differentiated in culture (day 9) to PPAR␥ activators resulted in a 2-fold (15-deoxy-⌬ 12,14 -prostaglandin J 2 ) to 8-fold (10 M BRL 49653) increase in UCP-1 mRNA levels. When PPAR␣ activators, such as several fibrates and the PPAR␣-specific ligand Wy 14,643, were tested an even higher (3-12-fold) increase in UCP-1 mRNA expression was detected. In contrast, COII mRNA expression did not respond to PPAR activators, thus indicating that the effect of PPAR activators is specific for UCP-1 mRNA.
Exposure to Wy 14,643 led to a dose-dependent increase in UCP-1 mRNA expression ( Fig. 2A) and maximum induction was attained at 10 M, a concentration at which it selectively activates PPAR␣ (4), whereas at 100 M it activates all three PPAR subtypes (10). The effects of 10 M Wy 14,643 were maximal after 12 h of exposure to the PPAR␣ ligand, and maintained after 24 h (Fig. 2B). The maximal effect of 10 M Wy 14,643 on UCP-1 mRNA levels resulted in an induction that was around 40% that of 0.5 M norepinephrine (20-fold Ϯ 2.7) and 180% that of 0.5 mM 8-bromo-cAMP, a nonmetabolizable cAMP derivative (4.3-fold Ϯ 0.6).
The Stimulation of ucp-1 Gene Expression by the PPAR␣ Ligand Wy 14,643 Is Independent of Protein Synthesis and Synergizes with the Effects of an RXR-specific Agonist-Brown adipocytes were exposed for 12 h to 10 M Wy 14,643 in the absence or presence of 5 g/ml cycloheximide, an inhibitor of protein synthesis (Fig. 3A). Cycloheximide treatment led to lower basal expression of UCP-1 mRNA, as already described (19), but it did not affect the ability of Wy 14,643 to increase UCP-1 mRNA.
When the effects of the RXR-specific agonist methoprene were analyzed (Fig. 3B), results showed that besides its reported direct action upon UCP-1 mRNA expression (20), there was a synergistic effect when both the PPAR␣ and the RXR ligands were added, suggesting a PPAR␣-RXR heterodimermediated effect on ucp-1 gene expression.
PPAR␣ Induces the Rat ucp-1 Gene Promoter Activity-Primary brown adipocytes were transiently transfected with a plasmid containing the upstream 4.5 kb of the rat ucp-1 gene fused to a CAT reporter gene. As shown in Fig. 4, PPAR␣ activators increased the (Ϫ4551)UCP1-CAT activity at least 2-fold, in the same range of the effect caused by BRL 49653. Responsiveness of (Ϫ4551)UCP1-CAT to PPAR activators was enhanced 6-fold by co-transfection of the expression vector for PPAR␣. Thus, expression of both endogenous ucp-1 gene and transfected ucp-1 gene promoter are up-regulated by PPAR␣ activators in primary brown adipocytes.
RXR␣ Enhances the PPAR␣-dependent Induction of the ucp-1 Gene Promoter-To further investigate the transcriptional regulation by PPAR␣ of the ucp-1 gene promoter, we used the brown adipocyte-derived HIB-1B cell line. These cells express PPAR␥ and PPAR␦ but not PPAR␣ (24). Thus, HIB-1B cells provide a useful model of brown fat-derived cell in which PPAR␣dependent regulation rely on transfected receptor. In agreement, Wy 14,643 did not modify (Ϫ4551)UCP1-CAT activity (Fig. 5A). However, co-transfection of pSG5-PPAR␣ induced (Ϫ4551)UCP1-CAT activity 3-fold in the absence and nearly 7-fold in the presence of 10 M Wy 14,643. Co-transfection of pRSV-RXR␣ caused a synergistic increase in the PPAR␣-dependent effect upon (Ϫ4551)UCP1-CAT activity. We next performed co-transfection experiments using HepG2 cells to avoid any interference of PPAR␥ in the observed effects. The HepG2 cell line was chosen because, in contrast to HIB-1B cells, does not express PPAR␥ nor PPAR␣ (31), and has been widely used to analyze PPAR␣ regulation of gene transcription (31)(32)(33). As shown in Fig. 5B, co-transfection of pSG5-PPAR␣ enhanced (Ϫ4551)UCP1-CAT activity and its responsiveness to Wy 14,643 in a dose-dependent manner, and maximal effects were observed at 0.3 g of pSG5-PPAR␣. This amount of vector was the same at which the maximum synergistic enhancement by co-transfection of pRSV-RXR␣ was found (Fig. 5C), nearly 200fold in the absence and 350-fold in the presence of 10 M Wy 14,643. When the other PPAR subtypes were tested, co-transfection of PPAR␥ in the presence of RXR␣ caused a similar increase in (Ϫ4551)UCP1-CAT activity than that of PPAR␣, but no effect was observed due to PPAR␦ co-transfection (data not shown).
CBP and PGC-1 Coactivate the PPAR␣-dependent Activation of the ucp-1 Gene Promoter-We next analyze whether coregulators CBP and/or PGC-1 were involved in mediating PPAR␣ transcriptional regulation of (Ϫ4551)UCP1-CAT. Cotransfection of pCMX-CBP or pSV-PGC1 alone enhanced basal (Ϫ4551)UCP1-CAT activity 5-and 7-fold, respectively (Fig. 6). When co-transfected together with pSG5-PPAR␣, an additive effect was observed in the absence of PPAR␣ ligand, but when 10 M Wy 14,643 was added, a synergistic activation was detected (30-fold for CBP and nearly 40-fold for PGC-1). When both coregulators were co-transfected in the presence of PPAR␣, a further increase in (Ϫ4551)UCP1-CAT activity was observed. These results point to an involvement of both CBP and PGC-1 in coactivating PPAR␣ and further increasing responsiveness of the ucp-1 gene promoter to PPAR␣-ligand.
PPAR␣and PPAR␥-dependent Regulation Require the Same Element in the Upstream Region of the ucp-1 Gene Enhancer-To determine the site in the 5Ј-region of the rat ucp-1 gene responsible for PPAR␣ action, the effects of PPAR␣ co-transfection on different deletion and double-point mutants of (Ϫ4551)UCP1-CAT were studied in transfected HepG2 and HIB-1B cells (Fig. 7). For comparative purposes, parallel cotransfection experiments were performed with PPAR␥. Results in both cell lines and for each PPAR subtype, indicated that both PPAR subtypes share a responsive site located in the Ϫ2494/Ϫ2318 enhancer region of the ucp-1 gene. When a double-point mutation, in which the CC at positions Ϫ2472 and Ϫ2473 were changed to AA (see Fig. 8A), was introduced in both (Ϫ3628/Ϫ2283)UCP1-CAT and (Ϫ2494/Ϫ2318)UCP1-CAT vectors, responsiveness to both PPAR␣ and PPAR␥ was abolished. Furthermore, when the Ϫ2494 to Ϫ2445 fragment was placed upstream (Ϫ172)UCP1-CAT or the HSV thymidine kinase promoter in pBLCAT2, it conferred 6-and 3-fold responsiveness, respectively, to PPAR␣ and PPAR␥ (data not shown).
The Ϫ2485/Ϫ2458 Site in the ucp-1 Gene Binds PPAR␣ and PPAR␥-Analysis of the sequence required for PPAR responsiveness in the rat ucp-1 gene promoter indicated the presence of a direct repeat with 1-base pair spacing related to a consensus PPRE (34) (Fig. 8A, arrows indicate half-site-related motifs). This Ϫ2485/Ϫ2458 sequence (UCP1-PPRE) in the rat ucp-1 gene promoter is highly conserved when compared with the previously reported PPAR␥-responsive element in the murine ucp-1 gene (18) and to the corresponding sequence in the human ucp-1 gene promoter (35) (Fig. 8A). Electrophoretic gel mobility shift assays were performed using the UCP1-PPRE as labeled probe. As shown in Fig. 8B, in vitro transcribed/translated RXR␣ alone (lane 2) did not bind significantly to this sequence although two nonspecific bands were detected as with the reticulocyte lysate (lane 1, n.s.). However, incubation with a mixture of PPAR␣ or PPAR␥ with RXR␣ resulted in the formation of the respective heterodimer complexes (lanes 3 and 4, respectively). To further assess the interaction of UCP1-PPRE with PPAR␣-RXR␣ or PPAR␥-RXR␣ heterodimers found in nuclear extracts from differentiated brown adipocytes in primary culture (Fig. 8C) or from BAT (Fig. 8D), supershift assays were performed using specific antibodies against RXR␣, PPAR␣, or PPAR␥. Arrows indicate the supershifted complexes formed (that contain RXR␣ and PPAR␣ or PPAR␥). Incubation with an antibody against ETS transcription factors, used as negative control, did not result in any change in the pattern of bands. Competition experiments performed in Fig. 8D with a 100-fold molar excess of specific (UCP1-PPRE) or its mutated version (mutUCP1-PPRE, see legend of Fig. 8A) confirmed the presence of a nonspecific band (n.s.). Taken together, these findings demonstrate that both PPAR␣ and PPAR␥ are present in brown fat cell nuclei and bind to UCP1-PPRE as heterodimers with RXR␣.

The PPAR␣ Ligand Wy 14,643 Induces ucp-1 Gene Expression in Brown Adipose Tissue in Vivo in Different Physiological
Situations-To assess the in vivo significance of PPAR activators on the expression of the ucp-1 gene, mice at different physiological situations were injected with single doses of the PPAR␣-specific ligand Wy 14,643 or, for comparative purposes, of the PPAR␥ activator troglitazone. We have previously reported that sensitivity of gene expression to PPAR␣ activators in acute treatments in vivo depends on the status of lipid metabolism able to provide endogenous PPAR␣ ligands (36). In adult mice (Fig. 9), Wy 14,643 caused a moderate 1.5-fold increase in UCP-1 mRNA abundance in BAT. When lactating mice were analyzed, Wy 14,643 significantly increased (5-fold) UCP-1 mRNA levels. During lactation, functional atrophy of BAT, including diminished lipolytic and lipoprotein lipase activities, and reduced expression of the ucp-1 gene contribute to energy sparing (37,38). In contrast, troglitazone only had a moderate effect on brown fat UCP-1 mRNA abundance.
When newborn mice at thermoneutrality were analyzed, injection of pups with Wy 14,643 caused a significant 3-fold rise in UCP-1 mRNA levels whereas injection of the PPAR␥-ligand BRL 49653 did not significantly change UCP-1 mRNA expression. The action of PPAR agonists was specific for the ucp-1 gene since COII mRNA levels were essentially unaffected by PPAR activators in BAT (see Fig. 9, bottom). Present results demonstrate an acute regulation of the ucp-1 gene in vivo by the PPAR␣-ligand Wy 14,643 that is more potent than that observed for PPAR␥ ligands.

FIG. 4. Effects of PPAR␣ and PPAR␥ agonists on (؊4551)UCP1-CAT expression in transiently transfected brown adipocytes.
Brown adipocytes differentiated in culture (day 9) were transfected with 12 g of (Ϫ4551)UCP1-CAT. When indicated, 3 g of the expression vector pSG5-PPAR␣ was co-transfected. After transfection, cells were exposed or not to 10 M Wy 14,643, 10 M BRL 49653, or 30 M Ly171883. Results are expressed as CAT activity relative to control, which is set to 1, and are means of two independent experiments, each one performed in triplicate.

DISCUSSION
Here we have established that PPAR␣ activators regulate the expression of the ucp-1 gene both in primary brown adipocytes and in BAT in vivo. Brown adipocytes differentiated in primary culture were used since they highly coexpress all PPAR subtypes, equally to BAT (24). In contrast, the HIB-1B brown adipocyte cell line lacks PPAR␣ expression (24), and therefore, the results of previous studies using HIB-1B cells to determine the effects of PPAR activators on the expression of the ucp-1 gene must be viewed with caution. Present results also demonstrate that PPAR␣ induces the rat ucp-1 gene promoter activity upon treatment with its specific ligand Wy 14,643, but it can also activate transcription in the absence of exogenously added ligand. This has been widely described for other PPAR␣-responsive gene promoters (32,39), and could be explained by either the presence of endogenous activators, such as fatty acids or their metabolites, or by ligand-independent activity of these nuclear receptors (40). The responsiveness of the ucp-1 gene promoter to PPAR␣-ligand is increased by co-transfection with expression vectors for either coactivator CBP or PGC-1. Furthermore, the synergistic effect observed when adding both coactivators points to the involvement at the same time of CBP and PGC-1 in coactivating PPAR␣. In this way, PPAR␣ can interact directly with CBP (41) and also with PGC-1 (42). In addition, CBP can form a complex with PGC-1 (43), thus providing multiple contact points to stabilize the complex assembly. Furthermore, CBP can also interact with other transcription factors, such as CREB and C/EBP, known to regulate transcription of the rat ucp-1 gene through its proximal regulatory region (22,21).
By deletion and mutation analysis we have identified the PPAR␣-responsive element in the upstream enhancer region of the rat ucp-1 gene. This Ϫ2485/Ϫ2458 region contains a potential PPRE consensus formed by two direct repeats separated by one nucleotide (DR-1). Highly comparable elements are also found in the human and mouse ucp-1 genes (see Fig. 8A), indicating that these sequences may have an important regulatory role in response to PPAR␣. In fact, the murine element has been described to mediate PPAR␥ responsiveness (18). Our present results further demonstrate that the Ϫ2485/Ϫ2458 element in the rat ucp-1 gene behaves as a promiscuous responsive site to either PPAR␣ and PPAR␥ activation, but not PPAR␦. From the analysis of various natural PPREs, it has been reported that the binding strength and functional transactivation for each PPAR subtype on the same PPRE was similar (33). Only some significant PPAR␥ specificity was described, and it was related to the 5Ј-flanking sequence with respect to the DR-1 element, which is essential for PPAR␣ binding (33). However, present results indicate a similar capacity of PPAR␣ and PPAR␥ to bind and activate ucp-1 transcription through the UCP1-PPRE. The predominant role of any subtype at any one time may thus depend on: 1) the relative amount of each subtype. For instance, PPAR␣ and PPAR␥ gene expression in brown adipocytes are under opposite regulation by their ligands and retinoic acid: up-regulation of PPAR␣ but down-regulation of PPAR␥ (24). 2) Cross-talk with other signaling pathways, like regulation of PPAR transcriptional activity by MAP kinase-dependent phosphorylation, which enhances PPAR␣ (44) but decreases PPAR␥ activity (45).
3) Ligand availability. Several PPAR ligands have been described to be highly subtype-specific (6), although identification

FIG. 5. PPAR␣-dependent induction of (؊4551)UCP1-CAT expression in transiently transfected HIB-1B and HepG2 cells: influence of RXR cotransfection.
A, HIB-1B cells were transfected with 1 g of (Ϫ4551)UCP1-CAT vector, and included or not 0.3 g of pSG5-PPAR␣, and/or 0.1 g of pRSV-RXR␣. After transfection, cells were exposed (dark bars) or not exposed (open bars) to 10 M Wy 14,643 for 24 h. Results are shown as relative to the basal expression of (Ϫ4551)UCP1-CAT, which is set to 1. Bars are means of at least two independent experiments, each one done in duplicate. B, HepG2 cells were transfected with 1 g of (Ϫ4551)UCP1-CAT vector together with increasing amounts of the expression vector pSG5-PPAR␣. After transfection, cells were exposed (l) or not exposed (E) to 10 M Wy 14,643 for 24 h. C, as in B, but 0.1 g of pRSV-RXR␣ was also co-transfected. Points are means of at least two independent experiments, each one done in duplicate.

FIG. 6. Effects of CBP and/or PGC-1 co-transfection on the PPAR␣-dependent induction of (؊4551)UCP1-CAT expression.
HepG2 cells were co-transfected twith 1 g of (Ϫ4551)UCP1-CAT vector, and included or not 0.3 g of pSG5-PPAR␣. When indicated, 0.1 g of pCMX-CBP and/or pSV-PGC1 were also co-transfected. After transfection, cells were exposed (dark bars) or not exposed (open bars) to 10 M Wy 14,643 for 24 h. Results are shown as relative to the basal expression of (Ϫ4551)UCP1-CAT, which is set to 1. Bars are means of at least two independent experiments, each one done in duplicate. of endogenous ligands and how their synthesis is regulated, is far from being established. 4) Interaction with coregulators. The interaction of PGC-1 with PPAR␣ is ligand-dependent whereas that with PPAR␥ is not (42,30). These and other possible events may determine which PPAR subtype activates transcription of ucp-1 in response to brown adipocyte physiological condition, mainly PPAR␥ in association with differentiation-dependent events or PPAR␣ in coordination with increased lipid catabolism in active BAT.
Other PPAR target genes have been described to be induced by both PPAR␣ and ␥ activators through the same PPRE (39,46). However, since they have been studied in tissues such as liver, which highly expresses PPAR␣ but not PPAR␥, or WAT, which predominantly expresses PPAR␥, tissue-specific regulation has been suggested. In contrast, BAT provides a model to study whether PPAR subtypes specifically regulate a PPRE in a target gene or whether a unique element behaves as a common site, as shown by our present findings in the ucp-1 gene promoter. For instance, the lipoprotein lipase (LPL) gene is up-regulated by PPAR␣ (in liver) and PPAR␥ (in WAT) through the same PPRE (46). During BAT differentiation, induction of LPL allows for increased fatty acids delivery to brown adipocytes, which results in triglyceride accumulation, thus promoting the adipocyte phenotype. However, thermogenic stimulus  (34) and to the analogous regions in the murine (Ϫ2499/Ϫ2472) and human (Ϫ3732/Ϫ3705) ucp-1 gene promoters (18,35). Asterisks indicated the double-point mutant derivative version (mutUCP1-PPRE) in which the CC at positions Ϫ2472 and Ϫ2473 were changed to AA. The upper arrows show the putative alignments of three motifs closely related to an idealized half-site. B, gel mobility shift assay: the double-stranded oligonucleotide Ϫ2485/Ϫ2458 was end-labeled and incubated with 5 l of in vitro transcribed/translated RXR␣, alone or together with PPAR␣ or PPAR␥. Arrows indicate the corresponding heterodimers bound to the probe. Lane 1 showed that the mock lysate produced two nonspecific bands when incubated with the probe. C, super-shift assay: the labeled UCP1-PPRE probe was incubated with 5 g of nuclear protein extract from differentiated primary brown adipocytes. When indicated, 1 l of antiserum against RXR␣, PPAR␣, PPAR␥, or ETS (as negative control) were added. Arrows indicate the super-shifted bands. D, protein extracts from rat brown adipose tissue nuclei (5 g) were incubated with the labeled UCP1-PPRE probe. Super-shift analysis was performed by incubation with 1 l of antiserum against PPAR␣, PPAR␥, or ETS. Oligonucleotide competitors, UCP1-PPRE (WT) and mutUCP1-PPRE (MUT), were added at a 100-fold molar excess relative to probe concentration. A nonspecific-binding band was detected (n.s.). Bracket indicates the specific-binding bands and arrows the super-shifted bands. also up-regulates LPL to increase fatty acids uptake, which increases the supply of substrate for oxidation. Expression of LPL mRNA is increased by PPAR␣ and ␥ activators in differentiated brown adipocytes, 2 suggesting that LPL gene transcription in BAT could be activated by both PPAR␣ and -␥. Other genes, such as the fatty acid transport protein and the acyl-CoA synthetase genes, which also regulate cell uptake of fatty acids, might be similarly regulated in BAT since they are induced by PPAR␣ and -␥ activators (39,47).
Here we also demonstrate that in vivo activation of PPAR␣ by Wy 14,643 up-regulates UCP-1 mRNA expression in BAT. The effects of the acute administration of this synthetic ligand are higher in those physiological situations (lactating dams and newborn pups at thermoneutrality) in which endogenous PPAR␣-ligands are expected to be low, in agreement with previous findings that PPAR␣ sensitivity in vivo depends on the status of lipid metabolism (36). Furthermore, the higher ucp-1 gene responsiveness to acute treatments with PPAR␣ than PPAR␥ agonists underlines the in vivo significance of PPAR␣dependent regulation of ucp-1 gene expression. In contrast, it has been reported that chronic exposure to PPAR␣ or PPAR␥ activators led to opposite results: long-term oral treatment of rats with Wy 14,643 did not change UCP-1 mRNA levels and thiazolidinedione administration resulted in a slight up-regulation of UCP-1 mRNA (48). This behavior of ucp-1 is similar to other bona fide PPAR-target genes in BAT, which remain unchanged by chronic exposure to PPAR␣ agonists (49). Positive effects of long-term treatment with thiazolidinediones on ucp-1 gene expression may be a consequence of their reported action promoting overall BAT differentiation (17,48,50).
Activation of BAT thermogenesis has been classically recognized to be mediated by norepinephrine. Among other regulatory effects, there is a cAMP-dependent activation of hormone sensitive-lipase, which rapidly hydrolyzes the stored triglycerides and releases high concentrations of fatty acids. These fatty acids, in addition to be the major substrate for thermogenesis and the inducers of UCP-1 uncoupling activity, may also act as PPAR activators. Accordingly, cold exposure and ␤-adrenergic stimulation of BAT result in activation of the PPAR pathway (51). We have previously reported that norepinephrine directly up-regulates transcription of the ucp-1 gene promoter, mainly through a cAMP responsive region in the proximal promoter region (22). However, the upstream enhancer region of the rat ucp-1 gene is also responsive to norepinephrine, although it lacks a defined cAMP responsive region. Several lines of evidence suggest a role for PPAR␣ in mediating this regulation, although the involvement of PPAR␥ cannot be ruled out. Mutation of the UCP1-PPRE affects the response of the ucp-1 gene promoter to norepinephrine (17). 3 Furthermore, the mitogenactivated protein kinase pathway is activated in BAT by adrenergic stimulation (52). This may result in activation of PPAR␣ but inactivation of PPAR␥, as discussed above. The coactivator PGC-1 is rapidly induced by cold-exposure through ␤-adrenergic pathways in BAT (30). Present data demonstrate that PGC-1 coactivates PPAR␣ and further increases ucp-1 gene responsiveness to PPAR␣-ligand dependent activation. Likewise, PGC-1 cooperates with PPAR␣ in the transcriptional control of nuclear genes encoding mitochondrial fatty acid oxidation enzymes (42), and it also induces mitochondrial gene expression by regulating the nuclear respiratory factor system (53). Taken together, these data point to PGC-1/PPAR␣ interaction as playing an important role in mediating changes in gene expression in response to BAT thermogenic requirements. Although basal expression of UCP-1 mRNA in PPAR␣-null mice has been reported to be unaltered (12), as also reported for other bona fide PPAR-target genes in liver (6), further studies are in course to determine whether ucp-1 gene expression is altered in these mice in response to thermogenic stimulus.
In conclusion, PPAR␣ directly regulates ucp-1 gene transcription and we propose that this transcriptional regulatory mechanism is a component of the coordinate control of thermogenic and lipid oxidation pathways in active BAT. Recently, PPAR␣ has been implicated in obesity (54) and selective PPAR␣ activators have been described to improve insulin sensitivity and reduce WAT mass (55). Part of these effects could be due to an increase in energy expenditure in BAT, and the positive action of PPAR␣ on ucp-1 gene expression opens new perspectives on the molecular targets of PPAR␣ involved in mediating these effects. FIG. 9. Effects of PPAR activators on UCP-1 mRNA expression in brown fat of adult, lactating, or neonate mice. A, representation of the relative abundance of UCP-1 mRNA in brown fat from adult female control and 15-day lactating dams after 6 h of being injected intraperitoneally with Wy 14,643 (50 g/g body weight), troglitazone (100 g/g body weight), or vehicle solution, and neonates that were injected intraperitoneally with Wy 14,643 (50 g/g body weight), BRL 49653 (50 g/g body weight), or vehicle solution (see "Experimental Procedures" for details). Data are expressed as relative to the adult female control which was set as 1. Statistical significance of comparisons between groups of treated mice and their respective vehicle treated controls are shown by: *, p Յ 0.05; **, p Յ 0.01. Comparison between Wy 14,643 and troglitazone treatment is shown by ⌬, p Յ 0.05. B, representative Northern blot analysis of equal amounts of brown fat RNA (20 g/lane) hybridized with the UCP-1 and COII probes, as described in the legend to Fig. 1.