Activation of Peroxisome Proliferator-activated Receptor g Bypasses the Function of the Retinoblastoma Protein in Adipocyte Differentiation*

The retinoblastoma protein (pRB) is an important regulator of development, proliferation, and cellular differentiation. pRB was recently shown to play a pivotal role in adipocyte differentiation, to interact physically with adipogenic CCAAT/enhancer-binding proteins (C/ EBPs), and to positively regulate transactivation by C/EBP b . We show that PPAR g -mediated transactivation is pRB-independent, and that ligand-induced transactivation by PPAR g 1 present in RB 1 / 1 and RB 2 / 2 mouse embryo fibroblasts is sufficient to bypass the differentiation block imposed by the absence of pRB. The differentiated RB 2 / 2 cells accumulate lipid and express adipocyte markers, including C/EBP a and PPAR g 2. Interestingly, adipose conversion of pRB-deficient cells occurs in the absence of compensatory up-regulations of the other pRB family members p107 and p130. RB 1 / 1 as well as RB 2 / 2 cells efficiently exit from the cell cycle after completion of clonal expansion following stimulation with adipogenic inducers. We conclude that ligand-induced activation of endogenous PPAR g 1 in mouse embryo fibroblasts is sufficient to initiate a transcriptional cascade resulting in induction of PPAR g 2 and C/EBP a expression, withdrawal from the cell

The retinoblastoma protein (pRB) 1 is a key regulator of the mammalian cell cycle. Through repression of the growth-promoting E2F transcription factors, pRB controls the transition from the G 1 to the S phase (1). pRB function is regulated by cyclin-dependent kinases, which phosphorylate pRB in a char-acteristic cell cycle-dependent manner (2). In addition, pRB plays a pivotal role during development and differentiation. The multifunctional character of pRB has been demonstrated by targeted disruption of the retinoblastoma gene in mice. Homozygous mutant embryos die in utero and show abnormalities in hematopoiesis and neurogenesis (3).
Numerous ex vivo studies have established the importance of pRB in myocyte differentiation (4). pRB has been shown to interact physically and functionally with members of the myogenic MyoD family of basic helix-loop-helix transcription factors (5), and pRB-deficient cells fail to undergo terminal myogenesis (6,7). This included a defect in expression of late differentiation markers, reduced myoblast fusion, a failure to terminally withdraw from the cell cycle, and an increased incidence of apoptosis (6 -8). These observations have also been seen in vivo when pRB is expressed at subphysiological levels (9).
Adipocyte differentiation ex vivo requires growth arrest, usually obtained by growing cells to confluence. Following stimulation with adipogenic factors, density-arrested preadipocytes undergo several rounds of postconfluent cell divisions (clonal expansion), followed by terminal withdrawal from the cell cycle, expression of adipocyte markers, and accumulation of intracellular lipid (24). The importance of pRB in adipocyte differentiation has been amply demonstrated. It was shown that the ability of a truncated simian virus 40 large T antigen to block adipocyte differentiation is dependent on its ability to sequester the pRB family (pRB, p107, p130) (25), and recently it was demonstrated that lung fibroblasts from RB Ϫ/Ϫ mouse embryos are unable to undergo adipose conversion unless rescued by an RB transgene (26). Furthermore, pRB was shown to physically interact with C/EBPs, promote the binding of C/EBP␤ to its cognate DNA response element, and increase its transactivation capacity (26,27). The functional interaction with C/EBP␤ suggests that pRB plays an important role early in the adipocyte differentiation program. Finally, regulated phosphorylation and expression of the three pRB family members during adipose conversion have recently been demonstrated (24,28).
In this study, we used fibroblasts from normal and RB Ϫ/Ϫ mouse embryos to further characterize the importance and functions of pRB in adipocyte differentiation. We show that transactivation by PPAR␥ is not dependent on pRB. Mouse embryo fibroblasts (MEFs) express PPAR␥1 in the predifferentiated state, and the inability of pRB-deficient MEFs to differentiate is efficiently bypassed by addition of PPAR␥ ligands. The differentiated RB Ϫ/Ϫ MEFs accumulate lipid and express adipocyte markers. Surprisingly, adipocyte differentiation of RB Ϫ/Ϫ MEFs was found not to be accompanied by compensatory up-regulation of p107 and p130 expression, and RB ϩ/ϩ as well as RB Ϫ/Ϫ MEFs effectively withdraw from the cell cycle following clonal expansion.
PCR Analysis of RB Gene Status-RB gene disruption via insertion of the hygromycin resistance gene in exon 19 was detected as described previously (29). Briefly, 50 ng of genomic DNA from individual MEFs were used for PCR amplification. To detect disruption of RB alleles, the following upstream and downstream primers were used: CGATCT-TAGCCAGACGAGCG(within the hygromycin resistance gene) and TGAGGCTGCTTGTGTCTGTG (within exon 19 of RB). To detect wildtype RB alleles, the downstream exon 19 primer was used in combination with the following upstream primer: GACTAGGTGAAGGAATG-CAGAG (within intron 18 of RB). As a control, we amplified part of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene using the following primers: ATTGGGCGCCTGGTCAC and CCAGAGGGGC-CATCCAC. Following a 10-min denaturation/activation of DNA polymerase (AmpliTaq Gold, Perkin Elmer), 35 cycles were performed as follows: 94°C for 20 s, 58°C for 20 s (62°C for GAPDH), 72°C for 60 s. PCR products were resolved on 1.5% agarose gels.
Oil Red O Staining-Dishes were washed in PBS and cells fixed in 3.7% formaldehyde for 1 h, followed by staining with Oil Red O for 1 h. Oil Red O was prepared by diluting a stock solution (0.5 g of Oil Red O (Sigma) in 100 ml of isopropanol) with water (6:4) followed by filtration. After staining, plates were washed twice in water and photographed.
Plasmids and Transfections-The PPREx3-tk-luc reporter containing three copies of the peroxisome proliferator-activated receptor response element (PPRE) from the acyl-CoA oxidase promoter was kindly provided by Ronald M. Evans (31). The pSPORT-mPPAR␥2 expression vector was kindly provided by Bruce M. Spiegelman (32). CMV-RB (33) and CMV-HA-E2F-1 (34) expression vectors and the 6xE2F-luc reporter containing six E2F binding sites (34) were kindly provided by Kristian Helin. The CMV-RB(H209) expression vector was kindly provided by Sibylle Mittnacht. It encodes a pRB mutant with a cysteine-to-phenylalanine substitution at amino acid 706 which abolishes the function of the pocket. The CMV-hBrm expression vector was kindly provided by Christian Muchardt (35). The CMV-r42-C/EBP␣ expression vector was kindly provided by M. Daniel Lane. The reporter containing the proximal part of the PPAR␥2 promoter cloned in front of the luciferase gene was kindly provided by Jeffrey M. Gimble (36). The MSV-C/EBP␤ vector was kindly provided by Steven L. McKnight (37). The CMV-␤galactosidase vector used for normalization is from CLONTECH. The human cervix carcinoma cell line C33A was grown in DMEM containing 10% FBS. Cells were transfected with the DC-Chol method as described (38). Cells were harvested approximately 48 h after transfection. The luciferase and ␤-galactosidase activities in cell lysates were determined by standard techniques.
Whole Cell Extracts-Plates were washed twice in TBS and cells were lysed on the plates by addition of an SDS sample buffer containing 2.5% SDS, 10% glycerol, 50 mM Tris-HCl (pH 6.8), 10 mM dithioerythritol, 10 mM ␤-glycerophosphate, 10 mM NaF, 0.1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and the Complete protease inhibitor mixture (1/50 tablet per ml) (Boehringer Mannheim). Lysis of cells was immediately followed by 3 min of boiling. Lysates were subsequently treated with benzon nuclease (Merck). Whole cell extracts were stored at Ϫ80°C. Protein concentrations were determined by the Bradford method (Bio-Rad).
Western Blotting-One hundred g of protein were loaded in each lane. After SDS-polyacrylamide gel electrophoresis, proteins were blotted onto polyvinylidene difluoride membranes (Micron Separation) using a Kem-En-Tec semidry blotter. Equal loading/transfer was confirmed by Ponceau S staining of membranes. Membranes were blocked overnight in PBS (or TBS) containing 5% nonfat dry milk and 0.1% Tween 20 (Sigma). Incubation with primary and secondary antibodies was performed in PBS (or TBS) containing 5% nonfat dry milk for 1-2 h. After incubation with antibodies, membranes were washed in PBS (or TBS) containing 0.1% Tween 20. PBS was used in all experiments except for those in which the mouse anti-human pRB antibody (G3-245, PharMingen) was used. Here, TBS was used instead. Other primary antibodies used were rabbit anti-human p107 (C-18, Santa Cruz Biotechnology), mouse anti-human p130 (Transduction Laboratories), rabbit anti-human TATA-binding protein (TBP) (Santa Cruz Biotechnology), rabbit anti-mouse aP2/adipocyte lipid-binding protein (ALBP) (kindly provided by David A. Bernlohr), rabbit anti-PPAR␥ antibody recognizing both PPAR␥ isoforms (kindly provided by Mitchell A. Lazar), and rabbit antibodies against mouse C/EBP␣ and C/EBP␤ (kindly provided by M. Daniel Lane). Secondary antibodies were horseradish peroxidase-conjugated anti-mouse or anti-rabbit antibodies (DAKO). Enhanced chemiluminescence (ECL) (Amersham Pharmacia Biotech) was used for detection. Stripping of membranes was done by boiling for 5-10 min in water.
RNA Purification and Reverse Transcription-Total RNA was purified as described (39). The integrity of all RNA samples was confirmed by electrophoresis in denaturing formaldehyde-containing gels. Reverse transcriptions were performed in 25-l reactions containing 1 g of total RNA, 3 g of random hexamers (Amersham Pharmacia Biotech), 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl 2 , 10 mM dithiothreitol, 40 units of RNAguard (Amersham Pharmacia Biotech), 0.9 mM dNTPs (Amersham Pharmacia Biotech), and 200 units of Moloney murine leukemia virus reverse transcriptase (Life Technologies). Reactions were left 10 min at room temperature, followed by incubation at 37°C for 1 h. After cDNA synthesis, reactions were diluted with 50 l of water and frozen at Ϫ80°C.
Preparation of Cells for Flow Cytometry-At the indicated time points, bromodeoxyuridine (BrdUrd) (Sigma) was added to the plates to a final concentration of 10 M, and incubation was continued for 20 min. Cells were then harvested by trypsinization, washed in 0.9% NaCl, fixed in 75% ethanol, and stored at 4°C until further analysis. Cells were treated with pepsin before incubation with a monoclonal anti-BrdUrd antibody (Becton Dickinson), followed by incubation with a fluorescein isothiocyanate-conjugated rabbit anti-mouse secondary antibody (DAKO). Cells were RNase-treated and stained with propidium iodide before loading onto an Epics Profile I flow cytometer.

RESULTS
Transactivation by PPAR␥ Is Independent of pRB-Transcriptional activation by nuclear receptors is dependent on recruitment of coactivator proteins. pRB has recently been shown to modulate the activity of the thyroid hormone and glucocorticoid receptors by interaction with coactivators (41,42). Trip230 was shown to be a thyroid hormone receptor (TR) coactivator and a pRB-interacting protein (41). pRB was able to sequester Trip230 from TR, thereby down-regulating the activity of TR. Contrary to the effect on TR, pRB was found to potentiate glucocorticoid receptor (GR)-mediated transactivation by direct interaction with hBrm (42), a previously identified GR coactivator (35). To examine whether the transcriptional activity of PPAR␥ was modulated by pRB, we analyzed the transactivation potential of full-length PPAR␥2 in the human cell line C33A which does not express functional pRB. C33A cells have previously been used to demonstrate potentiation of GR transactivation by pRB (42). As shown in Fig. 1A, PPAR␥ is transactivating the reporter in C33A cells (column 3), and addition of the high affinity PPAR␥ ligand BRL49653 further enhances its activity (column 4). Coexpression of pRB has little or no effect on either BRL49653-dependent or -independent transactivation by PPAR␥ (compare columns 3 and 7 and columns 4 and 8). As expected, no effect on PPAR␥ transactivation was observed by coexpressing the nonfunctional pRB mutant pRB(H209) (compare columns 3 and 11 and columns 4 and 12). Since C33A cells express no hBrm (35), we wanted to rule out the possibility that pRB affected PPAR␥ transactivation via this coactivator. This appeared not to be the case, as we observed no effect on PPAR␥ transactivation by coexpression of hBrm, neither in the presence nor in the absence of pRB (data not shown). As a positive control, we tested the effect of pRB on E2F-mediated transactivation. Fig. 1B shows that pRB represses basal reporter activity, probably by repressing endogenous E2F in C33A cells (compare columns 1 and 3). Furthermore, E2F-1-induced transactivation of the reporter (column 2) was partially repressed by coexpression of pRB (column 4). The pRB mutant failed to repress either basal reporter activity or E2F-1-induced reporter activity (compare columns 1 and 5 and columns 2 and 6). From these experiments, we conclude that transactivation by PPAR␥ is independent of pRB in C33A cells.
Activation of PPAR␥ Bypasses the Function of pRB in Adipocyte Differentiation-The use of embryonic fibroblasts (MEFs) from mice with targeted disruptions of specific genes is a powerful tool in deciphering the importance and functions of proteins in cellular differentiation. By using lung fibroblasts from mouse embryos with targeted disruption of the RB gene, the importance of pRB in adipocyte differentiation was demonstrated (26). pRB and C/EBPs were shown to interact, and it was demonstrated that pRB potentiated transactivation by C/EBP␤ (26,27).
To further characterize the importance and function of pRB in adipose conversion, we examined the potential of different known adipogenic inducers to support adipocyte differentiation of fibroblasts from normal and RB Ϫ/Ϫ mouse embryos (29). These experiments were performed with MEFs from one wildtype (MEFA) and two pRB-deficient (ME3 and ME8) mouse embryos. Genotypes were validated by genomic PCR (Fig. 2A), and the absence of pRB expression in RB Ϫ/Ϫ MEFs was confirmed by Western blotting (Fig. 2B) and immunostaining (data not shown). Using a standard differentiation protocol including treatment with dexamethasone, methylisobutylxanthine, and insulin (DMI treatment), only the RB ϩ/ϩ MEFs differentiated to a significant degree (Fig. 3, A and B (a and c)). This is in agreement with previous results (26). However, we consistently observed some pRB-deficient cells accumulating lipid in response to the DMI treatment (Fig. 3B, e). In one of the RB Ϫ/Ϫ MEFs (ME8), only a few cells accumulated lipid in response to standard inducers, whereas, in the other (ME3), approximately 1% of the cells accumulated lipid. By RT-PCR and Western blotting, we found that all three MEFs express PPAR␥1 mRNA and protein in the predifferentiated state (see Fig. 4). Therefore, we hypothesized that addition of a high affinity ligand for PPAR␥ might be able to bypass the block in adipose conversion imposed by the absence of pRB. Addition of BRL49653 to the standard differentiation medium efficiently promoted differentiation of RB Ϫ/Ϫ as well as RB ϩ/ϩ MEFs (Fig. 3, A and B (b and  d)). The ligand concentration needed to bypass the defective differentiation in RB Ϫ/Ϫ MEFs was in agreement with the K d of BRL49653 binding to PPAR␥ (17) in that differentiation was promoted with 50 nM BRL49653 (data not shown). Furthermore, even though predifferentiated MEFs express low levels of PPAR␦ and PPAR␣ mRNA (data not shown), the concentration of BRL49653 used in this report (0.5 M) is sufficient to activate only the PPAR␥ subtype (17,18). Therefore, we conclude that PPAR␥ is the target receptor in the BRL49653induced differentiation of pRB-deficient MEFs.
To characterize the differentiation of RB ϩ/ϩ and RB Ϫ/Ϫ MEFs in more detail, gene expression was examined by multiplex RT-PCR and Western blotting. Treatment of MEFs with adipogenic inducers (DMI) resulted in a transient induction of C/EBP␤ with expression levels peaking on day 1, irrespective of RB status and supplementation of BRL49653 (Fig. 4B). A transient up-regulation of C/EBP␤ is also seen during differentiation of 3T3-L1 cells (15,43). Treatment of RB ϩ/ϩ MEFs with either DMI or DMI together with BRL49653 resulted in the induction of PPAR␥2, C/EBP␣, and GPDH mRNAs (Fig. 4A,  left). However, induction was accelerated and expression levels were higher when cells were treated with the PPAR␥ ligand. Western blotting showed that the induction of PPAR␥2 mRNA was accompanied by synthesis of PPAR␥2 protein. Similarly, a robust induction of aP2/ALBP was detected (Fig. 4B, left). Of interest, even though C/EBP␣ mRNA was induced in the dif-  ferentiating RB ϩ/ϩ cells in absence of the PPAR␥ ligand, C/EBP␣ protein was detected only in cells treated with BRL49653 (Fig. 4B, left). Even extended exposure of the blots of protein from the DMI-treated RB ϩ/ϩ cells challenged with antibodies against C/EBP␣ revealed no signals above background (data not shown). Considering the relatively strong induction of C/EBP␣ mRNA on day 6 in the DMI-treated cells, the absence of detectable C/EBP␣ protein suggests a posttranscriptional regulation of C/EBP␣ expression in MEFs. In pRB-deficient MEFs, PPAR␥2, C/EBP␣, and GPDH mRNAs were very weakly induced when cells were treated with DMI in the absence of BRL49653 (Fig. 4A, right). Treatment of RB Ϫ/Ϫ MEFs with DMI plus BRL49653, however, led to an induction of PPAR␥2, C/EBP␣, and GPDH mRNAs similar to that observed in RB ϩ/ϩ MEFs (Fig. 4A). Robust induction of PPAR␥2, C/EBP␣, and aP2/ALBP proteins in pRB-deficient cells was also dependent on the PPAR␥ ligand (Fig. 4B, right). As mentioned above, PPAR␥1 (but not PPAR␥2) mRNA and protein were expressed in confluent MEFs (day 0 in Fig. 4, A and B). Therefore, it appears that ligand-activation of endogenous PPAR␥1 induces PPAR␥2 expression and differentiation of RB Ϫ/Ϫ MEFs.

The Effect of pRB on C/EBP␣-and C/EBP␤-mediated Transactivation of the Proximal PPAR␥2
Promoter-pRB has been demonstrated to potentiate C/EBP␤-mediated transactivation of reporter plasmids containing multimeric C/EBP binding sites, possibly by acting as a chaperone to induce binding of C/EBP␤ to its cognate DNA response element (26,27). Whether pRB was capable of regulating the activity of natural promoters via C/EBP sites was not addressed in these studies. To investigate this, we analyzed the importance of pRB in the transactivation of a C/EBP-regulated gene which is induced during adipose conversion. The proximal part of the PPAR␥2 promoter contains two C/EBP sites, which confer C/EBP-dependent activation in transient transfection studies (36). Using the proximal PPAR␥2 promoter as a reporter plasmid (36), cotransfection with both C/EBP␣ and C/EBP␤ expression vectors was found to transactivate the reporter in the pRB-deficient C33A cells (Fig. 5, columns 3 and 7). The effect of coexpression of pRB on C/EBP␣-or C/EBP␤-mediated transactivation is shown in Fig. 5. In the case of C/EBP␣, coexpression of pRB was found to have little or no effect (compare columns 3 and 4) in accordance with the previously noted pRB insensitivity of C/EBP␣-dependent transactivation (44), and similarly, pRB only modestly increased the C/EBP␤-mediated transactivation (compare columns 7 and 8). The pRB pocket mutant pRB(H209) did not significantly affect transactivation mediated by the C/EBPs (Fig. 5, compare columns 3 and 6 and columns 7 and 9). The absent or very moderate effect of pRB on C/EBP-mediated transactivation of the proximal PPAR␥2 promoter is in contrast to the pronounced effect on reporters containing multimeric C/EBP binding sites (26,27). The experiments in Fig. 5 were performed with the same ratio of transcription factor to pRB expression vectors as in Fig. 1, where pRB significantly repressed E2F-mediated transactivation. pRB has been shown to potentiate GR-mediated transactivation in C33A cells (42), but it cannot be excluded that the chaperone-like effect of pRB on C/EBP proteins may be sensitive to relative protein levels. Furthermore, cell lines may differ in their ability to support a functional pRB-C/EBP interaction. In conclusion, however, our results suggest that pRB may regulate adipogenesis through pathways in addition to those controlled by C/EBP proteins.
Regulation of pRB Family Members during Adipocyte Differentiation of RB ϩ/ϩ and RB Ϫ/Ϫ MEFs-During myocyte differentiation of RB ϩ/ϩ cells, the levels of p107 and p130 are inversely regulated, with p107 being down-regulated (6, 7) and p130 up-regulated (7). Myogenic conversion of RB Ϫ/Ϫ cells, however, causes induction of both p107 (6, 7) and p130 (7). The pRB-deficient cells are impaired in expression of late myocyte differentiation markers and fail to terminally withdraw from the cell cycle (6,7). However, the RB Ϫ/Ϫ cells do express early myocyte markers (6,7), and it is therefore conceivable that the up-regulation of p107 in RB Ϫ/Ϫ cells promotes the early steps in the differentiation program.
To study possible compensatory regulations of p107 and p130 during adipose conversion of pRB-deficient MEFs, we compared expression profiles of the three pRB family members in differentiating RB ϩ/ϩ and RB Ϫ/Ϫ MEFs by Western blotting. Fig. 6 (left) shows that pRB is present mainly in the hypophosphorylated state before stimulation with adipogenic factors (day 0). After 24 h, a significant fraction is hyperphosphorylated as seen by the reduced migration. After day 1, the majority of pRB is hypophosphorylated (Fig. 6). This pattern of phosphorylation is independent of the presence of BRL49653. Furthermore, the overall level of pRB does not change in these experiments. We found no significant differences in the expression profiles of p130 and p107 between RB ϩ/ϩ and RB Ϫ/Ϫ MEFs (Fig. 6). The level of p130 was high at day 0, transiently reduced following stimulation with adipogenic inducers, and restored on days 4 -6 (Fig. 6). The regulation of p130 was not significantly affected by addition of BRL49653. p107 showed two peaks of induction (days 1 and 3), which were also associated with an increased phosphorylation. Interestingly, addition of BRL49653 significantly reduced the level of p107 from day 4 in the differentiation program compared with control cells (Fig.  6). This was particularly prominent in wild-type cells, but a similar tendency was observed in both pRB-deficient MEFs ( Fig. 6 and data not shown). Since BRL49653 effectively increases the number of differentiating cells (see Fig. 3), this indicates that adipose conversion is associated with decreased p107 expression, even in pRB-deficient cells.
To examine the importance of pRB in the cell divisions taking place after stimulation of confluent cells with adipogenic inducers, we used BrdUrd labeling to measure the percentage of cells in S phase. Fig. 7 shows the percentage of BrdUrdpositive RB ϩ/ϩ and RB Ϫ/Ϫ MEFs every 12 h after stimulation with standard inducers (DMI treatment). The figure shows the result of one of two independent experiments. Two rounds of FIG. 5. The effect of pRB on C/EBP␣-and C/EBP␤-mediated transactivation of the proximal PPAR␥2 promoter. C33A cells were transfected with the proximal PPAR␥2 promoter cloned in front of a luciferase reporter gene (36) (0.7 g) together with combinations of expression vectors for C/EBP␣ (0.7 g), C/EBP␤ (0.7 g), pRB (0.7 g), and pRB(H209) (0.7 g). Empty expression vector was added to ensure equal promoter load. Luciferase values were not normalized to ␤-galactosidase values in these experiments since coexpression of C/EBPs significantly increased expression from the CMV-␤-galactosidase vector (data not shown). Transfections were performed in triplicate, measured in duplicate, and repeated two times.
DNA synthesis are apparent, the first peaking on day 1 and the second peaking on day 2.5. Supplementation of 5 M BRL49653 to the standard inducers did not significantly affect the distribution of neither RB ϩ/ϩ nor RB Ϫ/Ϫ MEFs in S phase (data not shown). A similar distribution of cells in S phase was seen during adipocyte differentiation of 3T3-L1 cells, again with peaks on days 1 and 2.5. 2 The percentage of cells in S phase was consistently higher in wild-type cells compared with RB Ϫ/Ϫ cells (Fig. 7). The flow cytometric analysis was performed with only one of the RB Ϫ/Ϫ MEFs (ME8), so whether the reduced number of cells in S phase extend to other pRB-deficient MEFs is not known at present. However, the fact that both RB ϩ/ϩ and RB Ϫ/Ϫ MEFs underwent two rounds of DNA replication with approximately the same time course indicates that pRB is not critical for the timing of the clonal expansion phase. In addition, Fig. 7 shows that pRB is not essential for cell cycle exit during adipose conversion.
A comparison of the time course of DNA synthesis (Fig. 7) and the expression profiles of pRB, p107, and p130 (Fig. 6) indicates that hyperphosphorylation of pRB in RB ϩ/ϩ MEFs is coinciding with the first round of DNA replication on day 1. The peaks of DNA synthesis coincide approximately with the induction and hyperphosphorylation of p107 on days 1 and 3 in both pRB-positive and pRB-negative cells. The transient down-regulation of p130 after stimulation with adipogenic inducers in both RB ϩ/ϩ and RB Ϫ/Ϫ MEFs indicate that the level of p130 is low during clonal expansion, followed by an up-regulation after the clonal expansion phase. Similar results have been reported for differentiating 3T3-L1 cells (24,28). DISCUSSION In this report we show that a high affinity PPAR␥ ligand effectively bypasses the block in adipocyte differentiation imposed by pRB-deficiency. To show this we used fibroblasts from normal and RB Ϫ/Ϫ mouse embryos. A significant fraction of the RB ϩ/ϩ cells differentiated in response to a standard differentiation protocol as determined by lipid accumulation and expression of adipocyte markers, whereas only few RB Ϫ/Ϫ cells differentiated when subjected to the same treatment. This is in agreement with previous work showing the importance of pRB in adipose conversion (25,26). Addition of the high affinity PPAR␥ ligand BRL49653 dramatically increased adipose conversion of RB Ϫ/Ϫ as well as RB ϩ/ϩ MEFs. At the molecular level this was accompanied by induction of adipocyte markers, including the key transcription factors C/EBP␣ and PPAR␥2. The ability of BRL49653 to induce differentiation in pRB-deficient cells was not strictly dependent on the standard inducers (dexamethasone, methylisobutylxanthine, and insulin) since exposure to PPAR␥ ligand alone induced significant lipid accumulation (data not shown).
C/EBP␤ along with C/EBP␦ play crucial roles in adipocyte differentiation, as revealed by targeted disruptions (16), and are considered important for induction of PPAR␥2 and C/EBP␣ expression (36,45). In our experiments, BRL49653 did not significantly affect the level of C/EBP␤ protein but was required for induction of both PPAR␥2 and C/EBP␣ mRNAs in RB Ϫ/Ϫ MEFs. Wild-type and pRB-deficient MEFs express PPAR␥1, but no PPAR␥2, in the predifferentiated state. Therefore, it is conceivable that the critical steps regulated by pRB early in the differentiation program are bypassed by ligand activation of PPAR␥1. The differentiation-promoting effect of ligand-induced PPAR␥ activation in the pRB-deficient fibroblasts is in agreement with our observation that PPAR␥ transactivation is independent of pRB in C33A cells.
pRB was shown to stimulate binding of C/EBP␤ to DNA without being present in the C/EBP-DNA complex (26,27). This indicates that pRB acts as a chaperone to enhance specific DNA binding of C/EBPs. Furthermore, pRB was shown to potentiate C/EBP␤-mediated transactivation of a reporter construct containing multimerized C/EBP binding elements in the promoter (26,27). Thus, it could be hypothesized that the lack of adipocyte differentiation of RB Ϫ/Ϫ MEFs was related to a severely reduced level of C/EBP␤-dependent transactivation.
We found that C/EBP␤-mediated transactivation of the proximal PPAR␥2 promoter was rather insensitive to the level of pRB expression. This indicates that the inability of pRB-deficient MEFs to undergo adipose conversion in response to a treatment that is sufficient to induce adipocyte differentiation of normal embryo fibroblasts may reflect impairment of additional processes involving pRB. Glucocorticoids play decisive roles during differentiation of most preadipocyte cell lines (15,43), but surprisingly little is known about the molecular func- tions of GR in adipocyte differentiation. pRB has been shown to potentiate transactivation mediated by GR (42,46,47), suggesting that impaired GR function in the pRB-deficient cells may contribute to the refractoriness of these cells to undergo adipocyte conversion. Finally, our finding that addition of a high affinity PPAR␥ ligand is required to induce adipose conversion of RB Ϫ/Ϫ MEFs leaves open the possibility that pRB participates in a pathway leading to the production of an endogenous ligand for PPAR␥. Such a pathway may involve ADD1/SREBP1, which was recently shown to play an important role in the production of an unidentified PPAR␥ ligand (23). Both RB ϩ/ϩ and RB Ϫ/Ϫ MEFs express ADD1/SREBP1 mRNA (data not shown), but whether pRB modulates the activity of ADD1/SREBP1 is not known.
To examine whether BRL49653-induced differentiation of RB Ϫ/Ϫ MEFs was accompanied by a compensatory regulation of p107 and p130, we compared the expression of these genes during differentiation of both RB ϩ/ϩ and RB Ϫ/Ϫ cells. We found little or no difference in the expression pattern of p107 and p130 between normal and RB Ϫ/Ϫ MEFs. In the later stages of the differentiation program, we found that the level of p107 was lower in cells treated with BRL49653 compared with control cells. This indicates that down-regulation of p107 is related to the degree of adipose conversion, even in RB Ϫ/Ϫ MEFs. This is in contrast to the observed up-regulation of p107 during myocyte differentiation of RB Ϫ/Ϫ cells, an up-regulation not seen in RB ϩ/ϩ cells (6, 7). However, even though p107 is downregulated in the terminal stages of adipose conversion, it is transiently up-regulated during clonal expansion (Ref. 28 and this study). Furthermore, p130 is up-regulated during the late stages of adipocyte differentiation (Ref. 28 and this study). This raises the question as to whether p107 and p130 are important regulators of adipose conversion. Recent evidence from other differentiation systems suggests that members of the pRB family differ in their ability to regulate differentiation. Using the myocyte differentiation system, cells from wild-type, RB Ϫ/Ϫ , p107 Ϫ/Ϫ , and p130 Ϫ/Ϫ mouse embryos were compared (7). Only RB Ϫ/Ϫ cells had defects in expression of late differentiation markers and terminal cell cycle withdrawal. Furthermore, pRB was significantly more potent in activating MyoD-mediated transactivation than p107 and p130 (7). A similar increased activity of pRB compared with p107 and p130 was observed in flat cell formation of Saos-2 cells, a phenotype indicative of osteoblast differentiation (46). The in vivo importance of the three pRB family members has been addressed by gene targeting in mice. Whereas RB Ϫ/Ϫ embryos die in utero with defects in neurogenesis and erythropoiesis (3), p107 Ϫ/Ϫ and p130 Ϫ/Ϫ mice are viable, fertile, and show no apparent abnormalities (48,49). These observations show that pRB is unique among the pRB family members in the regulation of differentiation of many lineages. Whether pRB is the key pocket protein positively regulating adipogenesis remains to be established, but evidence obtained so far indicates that this may very well be the case.
Terminal withdrawal from the cell cycle is an essential step in differentiation of many cell lineages. Little is known about the regulation of cell cycle withdrawal in adipocyte differentiation. Recent evidence suggests that hypophosphorylation of pRB is important for the commitment of cells to undergo adipose conversion (24,50). Both of the major regulators of adipocyte differentiation, C/EBP␣ and PPAR␥, have been shown to inhibit cell proliferation (51,52). Inhibition of proliferation by C/EBP␣ does not require the presence of pRB but is dependent on a functional activation domain (44). C/EBP␣ inhibits proliferation via transcriptional stimulation and posttranslational stabilization of the p21 cyclin-dependent kinase inhibitor (53,54). Activation of PPAR␥ has been shown to inhibit proliferation by down-regulation of the PP2A phosphatase, which in turn is accompanied by a decrease in E2F activity (52). The inhibition of cell proliferation by PPAR␥ was also observed in cells expressing the simian virus 40 large T antigen, indicating that PPAR␥-mediated growth arrest does not require a functional pRB (52). In adipocyte differentiation, both C/EBP␣ and PPAR␥ are present at the time when clonal expansion ceases, and therefore, they are both possible effectors of the cell cycle withdrawal. Evidently, the PPAR␥/C/EBP␣ initiated cell cycle withdrawal and adipocyte differentiation of MEFs may proceed in the absence of a functional pRB-dependent pathway, a notion in keeping with the finding that a certain cell cycle control prevails in pRB deficient cells (29,55). How PPAR␥ and C/EBP␣ function in such regulatory circuits remains to be established.