The Orphan Nuclear Receptor Rev-Erbα Is a Peroxisome Proliferator-activated Receptor (PPAR) γ Target Gene and Promotes PPARγ-induced Adipocyte Differentiation*

Rev-Erbα (NR1D1) is an orphan nuclear receptor encoded on the opposite strand of the thyroid receptor α gene. Rev-Erbα mRNA is induced during adipocyte differentiation of 3T3-L1 cells, and its expression is abundant in rat adipose tissue. Peroxisome proliferator-activated receptor γ (PPARγ) (NR1C3) is a nuclear receptor controlling adipocyte differentiation and insulin sensitivity. Here we show that Rev-Erbα expression is induced by PPARγ activation with rosiglitazone in rat epididymal and perirenal adipose tissues in vivo as well as in 3T3-L1 adipocytes in vitro. Furthermore, activated PPARγ induces Rev-Erbα promoter activity by binding to the direct repeat (DR)-2 response element Rev-DR2. Mutations of the 5′ or 3′ half-sites of the response element totally abrogated PPARγ binding and transcriptional activation, identifying this site as a novel type of functional PPARγ response element. Finally, ectopic expression of Rev-Erbα in 3T3-L1 preadipocytes potentiated adipocyte differentiation induced by the PPARγ ligand rosiglitazone. These results identify Rev-Erbα as a target gene of PPARγ in adipose tissue and demonstrate a role for this nuclear receptor as a promoter of adipocyte differentiation.

appears to promote adipocyte differentiation by activating the expression of PPAR␥ and increasing the synthesis of endogenous PPAR␥ ligands (13)(14)(15). Members of the PPAR family bind as heterodimers with the retinoid X receptors (RXR) to specific response elements termed peroxisome proliferator response elements (PPRE) (for review see Ref. 16). These PPREs usually consist of a direct repeat of the PuGGTCA motif spaced by one nucleotide (DR1). The transcriptional activity of the PPARs is activated by a number of different fatty acid metabolites, most notably products of the cycloxygenase and lipoxygenase pathways. In addition, a large number of synthetic compounds are known to be potent and subtype specific PPAR ligands. For example, thiazolidinedione compounds used as insulin sensitizers in the treatment of type II diabetes are high affinity PPAR␥ ligands (17).
Rev-Erb␣ (NR1D1) is another nuclear receptor, the expression of which is induced during adipocyte differentiation (18). Rev-Erb␣ is highly expressed in adipose tissue but also in skeletal muscle, liver and brain (18 -21). Since no ligand has been identified so far, Rev-Erb␣ is considered as an orphan member of the nuclear receptor superfamily. Rev-Erb␣ has been shown to act as a negative regulator of transcription (22) binding either as monomer on nuclear receptor half-site motifs flanked 5Ј by an A/T rich sequence (A/T PuGGTCA), or as a homodimer to a direct repeat of the PuGGTCA motif spaced by two nucleotides (DR2).
We have previously shown that PPAR␣ activates the expression of Rev-Erb␣ through an atypical PPRE, a DR-2 element, in the Rev-Erb␣ promoter (23). Transcriptional activation by PPAR␥ through a DR-2 element has so far not been reported. However, since Rev-Erb␣ is induced during the course of adipocyte differentiation, we decided to investigate whether PPAR␥ could be involved in transcriptional induction of Rev-Erb␣ expression in adipocytes. Furthermore, we wanted to investigate whether Rev-Erb␣ plays a modulatory role in the process of adipogenesis. Our results from both in vivo and in vitro studies demonstrate that treatment with the PPAR␥ agonist rosiglitazone increases Rev-Erb␣ gene expression and that PPAR␥ activates Rev-Erb␣ transcription via the Rev-DR2 response element present in the human Rev-Erb␣ promoter. Finally, we show that ectopic expression of Rev-Erb␣ in 3T3-L1 preadipocytes significantly augments the adipogenic activity of the PPAR␥ selective ligand rosiglitazone.

MATERIALS AND METHODS
Animals-Male Sprague-Dawley rats (10 weeks old) were treated for 14 days by gavage with rosiglitazone (10 mg/kg/d) suspended in 1% carboxymethylcellulose solution. Control animals received an equal volume (5 ml/kg/d) of carboxymethylcellulose solution. At the end of the experiments, animals were sacrified by exsanguination under ether anesthesia. Adipose tissues were removed immediately and frozen in liquid nitrogen.
Reverse transcription was performed with MMLV reverse transcriptase starting with 1 g of total RNA following the manufacturer's instructions (Invitrogen). For quantitative PCR, reverse transcribed were quantified by real time PCR on a MX4000 apparatus (Stratagene, La Jolla, CA), using specific oligonucleotide primers indicated in Table  I. PCR amplification was performed in a volume of 25 l containing 100 nM of each primers, 4 mM MgCl 2 , the Brilliant Quantitative PCR Core reagent Kit mix as recommended by the manufacturer (Stratagene, La Jolla, CA). The conditions were 95°C for 10 min, followed by 40 cycles of 30 s at 95°C, 30 s at 55°C, and 30 s at 72°C. mRNA levels were normalized to 28S rRNA. The values presented are means Ϯ S.D. of triplicates.
Transfection Experiments-The human Rev-Erb␣ promoter constructs were described previously (29), 3T3-L1 and COS cells were obtained from ATCC (Manassas, VA). Cells were grown in DMEM, supplemented with 2 mM glutamine and 10% (v/v) fetal calf serum in a 5% CO 2 humidified atmosphere at 37°C. All transfections were carried out as described previously (23). 3T3-L1 preadipocytes were transfected using the cationic lipid RPR 120535B as described previously (30). Luciferase activities were determined on total cell extracts using a luciferase assay system (Promega, Madison, WI). Transfection experiments were performed in triplicate and repeated at least three times.
In Vitro Translation and EMSAs-pSG5hPPAR␥, pSG5mRXR␣, and pSG5hRev-Erb␣ were in vitro transcribed with T7 polymerase and translated using the rabbit reticulocyte lysate system (Promega, Madison, WI). Electrophoretic mobility shift assays (EMSAs) with Rev-Erb␣, PPAR␥, and/or RXR␣ were performed exactly as described previously (23). For competition experiments, increasing amounts of the indicated cold probe were added just before the labeled oligonucleotide. The complexes were resolved on 5% polyacrylamide gels in 0.25ϫ TBE buffer (90 mM Tris borate, 2.5 mM EDTA, pH 8.3) at room temperature. Gels were dried and exposed overnight at Ϫ70°C to x-ray film (XO-MAT-AR, Eastman Kodak, Rochester, NY).
Viral Production and Infection-GPϩE86 virus-encapsidating cells (31) were cultured in DMEM (4.5 g/liter glucose) containing 10% heatinactivated calf serum (Hyclone), 8 g/ml gentamycin, 50 units/ml penicillin, 50 g/ml streptomycin at 37°C in a 5% CO 2 /95% air humidified atmosphere. In order to generate cell lines that constitutively overexpress Rev-Erb␣, the coding sequence of human Rev-Erb␣ was inserted in front of the internal ribosome entry site and the neomycin resistance gene (pCITE, Novagen) of the MFG retrovirus plasmid (32) using the NcoI-BamHI sites to generate pMFG-Rev-Erb␣. A similar construction lacking the Rev-Erb␣ sequence was made and used throughout the study as a negative control (pMFG-Neo). The bicistronic construct was designed to allow separate expression of the Rev-Erb␣ protein and the neomycin resistance by the infected cells. GPϩE86 cells (15,000/cm 2 ) were transfected with the MFG plasmid constructs (2 g) using LipofectAMINE (Invitrogen) and selected for resistance using the geneticin analog G418 (0.8 mg/ml, Invitrogen). 3T3-L1 cells were infected with the GPϩE86-produced MFG-Neo or MFG-Rev viruses essentially as previously described (33). Geneticin-resistant infected cell pools were used for the studies within three passages after infection.
Cell Culture and Differentiation-3T3-L1 cells were cultured in growth medium containing DMEM and 10% calf serum. The cells were differentiated by the method of Bernlohr et al. (34). 2 days post-confluent cells (designated day 0) were switched to differentiation medium (DMEM, 10% fetal calf serum, 1 M dexamethasone, 10 g/ml insulin, and 0.5 mM 3-methyl-1-isobutylxanthine (IBMX) (Sigma)) for 2 days. Thereafter, the cells were cultured in post-differentiation medium (DMEM, 10% fetal calf serum, insulin) with or without rosiglitazone (1 M). The medium was changed every day. Retroviral infected 3T3-L1 preadipocytes were cultured under the same conditions, but differentiated either with (standard differentiation conditions) or without (rosiglitazone-dependent differentiation conditions) dexamethasone. After treatment the cells were fixed with 10% formaldehyde in phosphate-buffered saline and stained with Oil Red O (Sigma) or total RNA was extracted as described above (35).

Activation of PPAR␥ Increases Rev-Erb␣ mRNA Levels in
Adipose Tissue-In order to determine whether PPAR␥ activation affects Rev-Erb␣ expression in vivo, rats were treated for 14 days with rosiglitazone, a highly specific and potent PPAR␥ ligand (36). The expression of Rev-Erb␣ was analyzed in epididymal and perirenal adipose tissue by Northern blot analysis. Compared with control treated rats, rosiglitazone treatment strongly increased Rev-Erb␣ mRNA levels both in epididymal and perirenal adipose tissue ( Fig. 1), whereas no change in control ␤-actin mRNA levels was observed. These experiments demonstrate that PPAR␥ activators increase Rev-Erb␣ mRNA levels in adipose tissue in vivo.
PPAR␥ Activation Induces Rev-Erb␣ mRNA in 3T3-L1 Preadipocytes-To study the molecular and cellular mechanisms of this induction, we next investigated the regulation of Rev-Erb␣ mRNA expression by rosiglitazone in the 3T3-L1 preadipocyte cell line. Cells were grown until confluency in medium containing calf serum. 2 days post-confluent cells (designated day 0), cells were transferred to medium containing fetal calf serum and differentiated with the classic differentiation mixture containing dexamethasone, IBMX, and insulin. From day 2, either 1 M rosiglitazone or vehicle was added. RNA was harvested at day 0, 2, 4, 6, and 8 and used for Northern analysis. As previously reported (18), Rev-Erb␣ mRNA levels increased upon differentiation of preadipocytes into adipocytes (Fig. 2, A and  B). However, compared with the standard differentiation treat- ment, Rev-Erb␣ mRNA levels were induced earlier in the presence of rosiglitazone. Moreover, Rev-Erb␣ mRNA levels were higher after 8 days in fully differentiated 3T3-L1 adipocytes treated with rosiglitazone compared with controls ( Fig. 2, A and B). Rosiglitazone is known to be a potent inducer of differentiation, thus in order to distinguish between direct effects of rosiglitazone on Rev-Erb␣ gene expression and indirect effects mediated via increased differentiation, we investigated whether rosiglitazone was able to induce Rev-Erb␣ in fully differentiated day 10 adipocytes. As shown in Fig. 2, C and D, rosiglitazone activated the expression of Rev-Erb␣ as well as the adipocyte lipid-binding protein (ALBP/aP2), a well-characterized PPAR␥ target gene (5), in mature adipocytes. Inhibition of protein synthesis by cycloheximide caused a superinduction of Rev-Erb␣ and could therefore not be used to investigate whether the induction of Rev-Erb␣ was mediated directly by PPAR␥ (data not shown). Nevertheless, these experiments demonstrate that activation of PPAR␥ increases Rev-Erb␣ mRNA during in vitro differentiation of 3T3-L1 preadipocytes as well as in mature adipocytes. PPAR␥ Induces Rev-Erb␣ Expression at the Transcriptional Level-PPAR␣ has previously been shown to activate the human Rev-Erb␣ promoter in hepatocytes via a DR-2 element (23), through which Rev-Erb␣ represses also its own transcription (29). To determine whether this site also mediates the activation of the Rev-Erb␣ promoter by PPAR␥ in adipocytes, transfection assays were carried out using luciferase reporter constructs driven by the Rev-Erb␣ promoter. 3T3-L1 cells were cotransfected with the PPAR␥ expression vector (pSG5-hP-PAR␥) or empty vector (pSG5) and treated with rosiglitazone or FIG. 2. Induction of Rev-Erb␣ mRNA expression in differentiating and mature 3T3-L1 adipocytes by rosiglitazone. A, 3T3-L1 preadipocytes were grown until confluence in medium containing 10% calf serum. Upon confluency cells were differentiated with IBMX, dexamethasone, and insulin for 2 days. Cells were subsequently treated with or without RSG (1 M in H 2 O) for 6 days. RNA was isolated and 10 g was analyzed by Northern blot analysis using 32 P-labeled rat Rev-Erb␣, mouse PPAR␥, and rat 36B4 cDNA probe fragments. B, quantification of Rev-Erb␣ mRNA levels normalized to 36B4 mRNA levels in RSG-stimulated and control cells. C, day 10 3T3-L1 adipocytes were incubated in media containing RSG (1 M in H 2 O) or vehicle. RNA was isolated and analyzed by Northern blotting using 32 P-labeled rat Rev-Erb␣ or rat aP2/ALBP cDNA fragments as probes. D, quantification of Rev-Erb␣ and aP2/ALBP mRNA levels normalized to 28S rRNA probe. The relative mRNA level of day 10 is set as 1. vehicle. Rev-Erb␣ promoter activity was induced by PPAR␥ cotransfection, an effect that was enhanced in the presence of rosiglitazone (Fig. 3). By contrast, Rev-Erb␣ promoter activity was unaffected by rosiglitazone in the absence of overexpressed PPAR␥, probably due to the insufficient levels of endogenous PPAR␥ in non-confluent preadipocytes. These data indicate that Rev-Erb␣ gene transcription is induced by rosiglitazone via PPAR␥.
To further investigate the importance of the DR-2 element in the induction by PPAR␥, Rev-Erb␣ promoter constructs in which the Rev-DR2 site was mutated, were tested next. Mutations affecting the 5Ј-AGGTCA motif (pGL2hRev-Erb␣ ⌬ as described (29) (Fig. 4A)) of the Rev-DR2 site resulted in the loss of Rev-Erb␣ promoter inducibility by rosiglitazone and PPAR␥ (Fig. 4B). These results indicate that the Rev-DR2 site mediates the transcriptional induction of Rev-Erb␣ by PPAR␥.
To assess whether the Rev-DR2 site could also function as a PPAR␥-responsive element in front of a heterologous promoter, transient transfection experiments were performed using constructs containing the wild-type or mutated versions of the Rev-DR2 site (Fig. 4B) cloned in front of the heterologous SV40 promoter (Rev-DR2, Rev-DR2M5Ј and Rev-DR2M3Ј). In COS cells, cotransfection of pSG5-hPPAR␥ on the Rev-DR2 driven SV40 reporter vector led to a 2.5-fold induction of transcription activity compared with empty pSG5 vector cotransfection (Fig.  4C). This effect was enhanced in the presence of rosiglitazone. By contrast, PPAR␥ did not activate the Rev-DR2 site mutated in its 5Ј-AGGTCA half-site (Rev-DR2M5Ј). Furthermore, mutation of the 3Ј-half-site (Rev-DR2M3Ј) of the DR2 site also abolished transactivation by PPAR␥. These results clearly demonstrate that the Rev-Erb␣ human promoter is regulated by PPAR␥ and that this induction is mediated via the Rev-DR2 site.
PPAR␥ Binds as a Heterodimer with RXR␣ to the Rev-DR2 Site-To investigate whether PPAR␥ binds directly to the Rev-DR2 site, electrophoretic mobility shift assays (EMSAs) were performed using in vitro synthesized PPAR␥ and RXR␣ proteins. Since the Rev-DR2 site was previously described as a Rev-Erb␣ response element, binding of Rev-Erb␣ was assayed as a control (Fig. 5A). As expected (29), Rev-Erb␣ bound both as homo-and as monomer to the Rev-DR2 site. Furthermore, no binding was observed on the Rev-DR2 oligonucleotide carrying a mutation in the 5Ј-half-site (M5Ј), whereas Rev-Erb␣ bound only as a monomer to the Rev-DR2 carrying a mutation in the 3Ј-half-site (M3Ј) in accordance with previous observations (29). RXR␣ or PPAR␥ alone did not bind to either wild-type or mutated Rev-DR2 sites confirming that PPAR␥ cannot bind as a monomer. By contrast, binding to the Rev-DR2 site was observed when PPAR␥ was incubated in the presence of RXR␣. The binding was specific since it was competed out by excess of unlabeled oligonucleotide (Fig. 5A). The specificity of the binding complex was verified by the addition of a specific anti-PPAR␥ antibody (38), which inhibited formation of the PPAR␥/ RXR␣ complex. The binding was prevented by mutation of both the 5Ј-(M5Ј) or 3Ј-(M3Ј) half-sites of the Rev-DR2 element. To determine the relative binding affinity of the PPAR␥/RXR␣ heterodimer for the Rev-DR2 site, cross-competition EMSA experiments were performed comparing binding of PPAR␥/ RXR␣ to the natural DR-1 site in the aP2 promoter (5) and the Rev-DR2 site. As shown in Fig. 5B, competition with the cold Rev-DR2 site decreased PPAR␥/RXR␣ binding to both the radiolabeled Rev-DR2 and aP2 DR-1 sites. These experiments demonstrate that PPAR␥ binds as a heterodimer with RXR␣ to the Rev-DR2 site of the Rev-Erb␣ promoter, albeit with significantly lower affinity compared with the aP2 PPRE site (Fig. 5B).

Rev-Erb␣ Increases the Adipogenic Activity of PPAR␥ Ago-
nists-To assess a potential role of Rev-Erb␣ in adipogenesis, full-length human Rev-Erb␣ was cloned into a retroviral vector, and 3T3-L1 preadipocytes were infected with the resulting virus (MFG-Rev) or the control MFG-Neo virus. Pools of cells stably transduced, but not clonal selected 3T3-L1 cells were subsequently cultured with an incomplete differentiation mixture, which requires PPAR␥ activation with rosiglitazone for optimal differentiation, and either rosiglitazone (1 M) or vehicle was added from day 2 to day 8. The presence of Rev-Erb␣ in MFG-Neo or MFG-Rev infected cells was analyzed by Western blot (Fig. 6A) and Northern blot analysis (Fig. 6B) of 3T3-L1 cells treated for 6 days with the differentiation mixture. Western blot analysis demonstrated the presence of ectopic expressed human Rev-Erb␣ protein in MFG-Rev infected, but not in the control cells (Fig. 6A). Since the antibody used was raised against a peptide, its affinity was too low to detect endogenous Rev-Erb␣ protein. However, both ectopic and endogenous Rev-Erb␣ expression was detected by Northern blot analysis in MFG-Rev cells, whereas MFG-Neo cells only expressed endogenous Rev-Erb␣ (Fig. 6B).
The effect of ectopic Rev-Erb␣ on adipocyte differentiation FIG. 5. PPAR␥ binds as a heterodimer with RXR␣ to the Rev-DR2 site of the Rev-Erb␣ promoter. Electrophoretic mobility shift assays were performed using the indicated end-labeled oligonucleotides in the presence of in vitro translated hPPAR␥, mRXR␣, hRev-Erb␣ or unprogrammed reticulocyte lysate. A, inhibition of PPAR␥/RXR␣ complex formation was obtained with 1 g of anti-PPAR␥ antibody (38). Competition experiments for binding were performed by adding 10-, 100-, or 200-fold excess of cold Rev-DR2 oligonucleotide (See Fig. 4A for the sequences of the used oligonucleotides). B, competition experiments for binding were performed by adding 10-, 100-, or 200-fold excess of cold aP2 DR-1 (5) or Rev-DR2 oligonucleotides to the reaction mixture.
was investigated using Oil Red O staining to assess triglyceride accumulation in 3T3-L1 cells induced to differentiate in incomplete medium (Insulin and IBMX, but no dexamethasone). Under these conditions, full differentiation is dependent on the presence of the PPAR␥ agonist rosiglitazone. In the absence of rosiglitazone, ectopic expression of Rev-Erb␣ induced only a slight increase in triglyceride accumulation (Fig. 6, C and D). However, when rosiglitazone was added, a major increase in triglyceride accumulation was observed in the cells expressing Rev-Erb␣ compared with control cells. These morphological changes were accompanied by a pronounced induction of mRNA levels of the adipocyte-specific marker, and PPAR␥ target gene, aP2, whose expression was strongly induced by rosiglitazone and Rev-Erb␣ (Fig. 7). Moreover, a strong increase in PPAR␥ mRNA levels was observed in MFG-Rev compared with MFG-Neo cells, which may be a reflection of the differentiation status of the cells. In addition, rosiglitazone treatment induced PPAR␥ expression due to PPAR␥ auto-induction (39). Similarly, mRNA levels of C/EBP␣, another differentiation marker, were also induced by Rev-Erb␣ over-expression. Moreover, as expected, C/EBP␣ mRNA level was induced by rosiglitazone in MFG-Neo cells, likely due to the cross regulation of PPAR␥ and C/EBP␣ (40). However, in Rev-Erb␣ expressing cells, rosiglitazone treatment did not further enhance C/EBP␣ expression. Although the expression levels of C/EBP␤, an early inducer of adipocyte differentiation, and SREBP-1/ADD1 were higher upon rosiglitazone treatment, likely due to the optimal differentiation status of the cells under these conditions, the mRNA levels of these transcription factors were not influenced by Rev-erb␣ expression (Fig. 7).
Next, the influence of Rev-Erb␣ on the expression of aP2, PPAR␥ and C/EBP␣ was determined when 3T3-L1 cells were differentiated under classical conditions (insulin, IBMX, and dexamethasone). As expected, induction levels of aP2, PPAR␥, C/EBP␣, and endogenous Rev-Erb␣ were more pronounced in the complete compared with the incomplete mixture without dexamethasone (compare Fig. 7 to Fig. 8). Induction levels of these adipogenic markers were identical between non-infected cells and MFG-Neo cells indicating that the retrovirus-infected cells differentiate normally (data not shown). Interestingly, Rev-Erb␣ infected cells expressed Rev-Erb␣ mRNA levels similar to those present in 3T3-L1 cells differentiated in complete mixture containing rosiglitazone, indicating that the Rev-Erb␣ retrovirus system produced Rev-Erb␣ levels within the physiological range (Fig. 8). Rev-Erb␣ over-expression further enhanced aP2, PPAR␥, and C/EBP␣ expression although the effect was less pronounced than in the presence of incomplete mixture. Finally, no change was observed in C/EBP␤ and SREBP-1/ADD1 mRNA levels under these conditions (Fig. 8). Altogether, these data support a role of Rev-Erb␣ as an enhancer of adipogenesis likely acting downstream PPAR␥.

DISCUSSION
The expression of the nuclear receptor Rev-Erb␣ has previously been reported to be transcriptionally up-regulated during adipocyte differentiation (18). However, the mechanisms of this induction and the physiological role of Rev-Erb␣ in adipogenesis remained unexplored. In the present report, we demonstrate that PPAR␥ activation by rosiglitazone induces Rev-Erb␣ mRNA expression both in vivo in adipose tissue of rats and in vitro in 3T3-L1 adipocytes. Transfection experiments revealed that the regulation of human Rev-Erb␣ expression by rosiglitazone occurs at the transcriptional level via activation of PPAR␥. Mutation of the Rev-DR2 site located in the Rev-Erb␣ promoter prevented PPAR␥-induced Rev-Erb␣ transcription identifying this site as a novel type of PPRE. Finally, we show that ectopic Rev-Erb␣ promotes adipocyte differentiation induced by PPAR␥ activators.
Recent structure-function analysis and three-dimensional modeling of Rev-Erb␣ indicated that the structure of the putative ligand cavity is occupied by side chains, suggesting that this receptor may not have any endogenous ligands (41). Therefore, the regulation of Rev-Erb␣ expression constitutes a crucial level of control of receptor activity. Rev-Erb␣ expression has been reported to be controlled by a variety of stimuli. Human Rev-Erb␣ represses its own expression via a Rev-DR2 site located in its promoter (29). In rat liver, as well as in cultures of human primary hepatocytes (42) and rat fibroblasts (43), Rev-Erb␣ expression oscillates in a circadian rhythm. In addition, rat and human hepatic Rev-Erb␣ expression is also under control of other nuclear receptors, such as the glucocorticoid receptor (42). Interestingly, we previously demonstrated that fibrates induce human Rev-Erb␣ expression in liver via PPAR␣ binding to the Rev-DR2 site (23). Here, we find that activation of PPAR␥ induces Rev-Erb␣ gene expression in adipose tissue. This induction occurs at the transcriptional level through binding of PPAR␥ to the Rev-DR2 site, in a manner as described for PPAR␣. This identifies Rev-Erb␣ as a new PPAR␥ target gene in adipose tissue. Moreover, since all PPAR␥ elements so far identified are DR-1 elements composed of a direct repeat of the canonical AGGTCA sequence separated by one base, our data identify the DR-2 element as a novel type of PPAR␥ response element. PPAR␣ and PPAR␥ share a number of common target genes. Indeed fibrates, which activate preferentially PPAR␣, induce LPL, ACS, and CD36 expression in liver, whereas glitazones, which selectively activate PPAR␥, have little effect on liver, but induce similar target genes in adipose tissue of rats (44 -46). Therefore, Rev-Erb␣ is another example of a gene whose expression is induced specifically by PPAR␣ in liver and by PPAR␥ in adipose tissue.
The molecular mechanisms regulating the initial steps of adipogenesis have been thoroughly studied. The transcription cascade triggering adipogenesis requires a fine orchestration of sequential expression of different transcription factors. The sequential activation of C/EBPs and PPARs are central in the adipogenic process. C/EBP␤ and C/EBP␦ are induced early during differentiation in 3T3-L1 preadipocytes and therefore appear to play an important role in the initiation of the adipogenic cascade. C/EBP␤ and C/EBP␦ induce the expression of PPAR␥ (9, 10) via two C/EBP binding sites in the promoter of PPAR␥2 (38,47). PPAR␥ and C/EBP␣ play a crucial role in further promoting adipocyte differentiation. PPAR␥ and C/EBP␣ mutually activate the expression of the other, leading to a sustained expression of these transcription factors (40,48). Our results show that Rev-Erb␣ is a target gene of PPAR␥ (Fig.  9). Whether Rev-Erb␣ is also a target gene for C/EBPs awaits further studies.
Since Rev-Erb␣ is a transcription factor, we studied whether FIG. 7. Ectopically expressed Rev-Erb␣ synergizes with rosiglitazone to induce expression of late adipogenic markers in 3T3-L1 cells differentiated in incomplete differentiation medium. 3T3-L1 cells were infected with a retrovirus containing either Rev-Erb␣ (MFG-Rev) or not (MFG-Neo). Cells were induced to differentiate in the presence of an incomplete differentiation mixture (insulin, IBMX but without dexamethasone) with or without RSG (1 M). Total RNA was isolated at day 8 and RNA levels were analyzed by quantitative RT-PCR as described under "Materials and Methods." it might participate in the adipogenic program. Rev-Erb␣ significantly promoted adipogenesis especially under conditions where the differentiation is dependent on PPAR␥ activation. Thus, Rev-Erb␣ is not only a PPAR␥ target gene, but also potentiates the adipogenic activity of this receptor. The observation that ectopic expression of Rev-Erb␣ increases the expression of the PPAR␥ target genes, aP2, and C/EBP␣, indicates that Rev-Erb␣ enhances the adipogenic function of PPAR␥. The fact that Rev-Erb␣ also induces PPAR␥ expression could be an indirect reflection of the pronounced stimulation of adipocyte differentiation, since this effect is also observed under conditions independent of rosiglitazone treatment (Fig. 8). Alternatively, Rev-Erb␣ could indirectly enhance PPAR␥ expression by interfering negatively with the expression or activity of a transcription factor that represses PPAR␥ expression. Since, Rev-Erb␣ over-expression does not change C/EBP␤ and SREBP-1/ADD1 mRNA levels, it appears to play a role as an enhancer of adipogenesis acting downstream of PPAR␥ (Fig. 9).
In conclusion, we have identified Rev-Erb␣ as a new target gene for PPAR␥ in the adipogenic cascade of transcription factors and as an important factor modulating adipocyte function, at least in part, by enhancing the adipogenic action of PPAR␥. The exact molecular mechanisms by which Rev-Erb␣ enhances adipogenesis await further studies.