Cyclin D1 Inhibits Peroxisome Proliferator-activated Receptor γ-mediated Adipogenesis through Histone Deacetylase Recruitment*

The cyclin D1 gene encodes the labile serum-inducible regulatory subunit of a holoenzyme that phosphorylates and inactivates the retinoblastoma protein. Overexpression of cyclin D1 promotes cellular proliferation and normal physiological levels of cyclin D1 function to inhibit adipocyte differentiation in vivo. We have previously shown that cyclin D1 inhibits peroxisome proliferator-activated receptor (PPAR)γ-dependent activity through a cyclin-dependent kinase- and retinoblastoma protein-binding-independent mechanism. In this study, we determined the molecular mechanism by which cyclin D1 regulated PPARγ function. Herein, murine embryonic fibroblast (MEF) differentiation by PPARγ ligand was associated with a reduction in histone deacetylase (HDAC1) activity. Cyclin D1–/– MEFs showed an increased propensity to undergo differentiation into adipocytes. Genetic deletion of cyclin D1 reduced HDAC1 activity. Reconstitution of cyclin D1 into the cyclin D1–/– MEFs increased HDAC1 activity and blocked PPARγ-mediated adipogenesis. PPARγ activity was enhanced in cyclin D1–/– cells. Reintroduction of cyclin D1 inhibited basal and ligand-induced PPARγ activity and enhanced HDAC repression of PPARγ activity. Cyclin D1 bound HDAC in vivo and preferentially physically associated with HDAC1, HDAC2, HDAC3, and HDAC5. Chromatin immunoprecipitation assay demonstrated that cyclin D1 enhanced recruitment of HDAC1 and HDAC3 and histone methyltransferase SUV39H1 to the PPAR response element of the lipoprotein lipase promoter and decreased acetylation of total histone H3 and histone H3 lysine 9. Collectively, these studies suggest an important role of cyclin D1 in regulation of PPARγ-mediated adipocyte differentiation through recruitment of HDACs to regulate PPAR response element local chromatin structure and PPARγ function.

The cyclin D1 gene was cloned as a breakpoint rearrangement in parathyroid adenoma (1) and as a macrophage colonystimulating factor-1-responsive gene in the mouse (2). Cyclin D1 encodes a labile growth factor-inducible regulatory subunit of the holoenzyme that phosphorylates and inactivates the retinoblastoma protein (Rb). 1 Cyclin D1 overexpression promotes G 1 phase progression in cultured cells and immunoneutralizing experiments have shown a requirement for cyclin D1 in fibroblast, mammary, and epithelial cell proliferation (3,4). Furthermore, cyclin D1 overexpression was shown to induce mammary tumorigenesis and to collaborate with the c-myc oncogene to induce lymphomagenesis (6). Deletion of the cyclin D1 gene in mice has demonstrated a key role for cyclin D1 in several distinct processes, including retinal and mammary gland development (7), cellular migration (8), cellular proliferation and survival (9), angiogenesis (10), and adipocyte differentiation (5). Cyclin D1 Ϫ/Ϫ MEFs have enhanced adipocyte differentiation in response to PPAR␥ ligands, which is reversed by cyclin D1 reintroduction (5).
In addition to promoting DNA synthesis and cellular proliferation, cyclin D1 has been shown to inhibit the activity of several transcription factors (V-Myb, MyoD) and nuclear receptors (androgen receptor, PPAR␥) (5,11). Cyclin D1 repression of PPAR␥ activity is known to have important physiological consequences. Although the mechanisms remain to be determined, several lines of evidence imply a role for histone acetyltransferase activity in cyclin D1 regulation of nuclear receptor function. For example, repression of AR activity by cyclin D1 occurs independently of the cyclin D1 Cdk-binding domain (11,12) and is reversed by co-expression of either P/CAF or p300 (11). P/CAF physically associates with cyclin D1 (11,13) through the Ada 2 region (11). Cyclin D1 also binds the histone acetyltransferase p300 (14,15), a co-activator for PPAR␥. HDAC1, a class 1 histone deacetylase, has been implicated in regulation of adipocyte differentiation. Treatment with histone deacetylase inhibitors promote preadipocyte differentiation (16). HDAC1/mSin3A has been shown to be recruited to the CCAAT/enhancer-binding protein promoter with CCAAT/enhancer-binding protein-␤ and promoted the deacetylation of histone H4. HDAC3, another member of the HDAC family, associates with PPAR␥ and Rb to form a PPAR␥⅐Rb⅐HDAC3 repressor complex and attenuated PPAR␥-mediated adipocyte differentiation (17).
Herein, studies were conducted to examine the mechanisms underlying the role of cyclin D1 as an inhibitor of PPAR␥ function and adipocyte differentiation. As HDAC inhibitors induce adipocyte differentiation of MEFs (16,17), we assessed the potential role of cyclin D1 in regulating HDAC activity and adipocyte differentiation. Detailed analyses demonstrated that cyclin D1 induces HDAC activity and blocks adipocyte differentiation. Cyclin D1 physically associates with HDACs and inhibits basal and ligand-induced PPAR␥ activity through recruitment of HDACs and histone methyltransferase to the PPAR␥ response element (PPARE) of the lipoprotein lipase (LPL) promoter.

EXPERIMENTAL PROCEDURES
Reagents, Reporter Genes, Expression Vectors, DNA Transfection, and Luciferase Assays-The acyl-coenzyme A oxidase triple PPARE luciferase (AOX) 3 LUC reporter gene, pCMX-PPAR␥, and the HDAC expression vectors (19) were described previously (5,20). Cyclin D1 ϩ/ϩ and cyclin D1 Ϫ/Ϫ MEFs and 3T3 cells (cyclin D1 Ϫ/Ϫ and cyclin D1 ϩ/ϩ ) were described previously (5,21). Cells were transfected by Superfect Transfection reagent (Qiagen, Valencia, CA) as described elsewhere (22). The medium was changed after 5 h, cells were treated with ligand or vehicle as indicated in the figure legends, and luciferase activity was determined after 24 h. Luciferase activity was normalized for transfection efficiency with ␤-galactosidase or Renilla reporters as an internal control. Luciferase assays were performed at room temperature with an Autolumat LB 953 (EG&G Berthold) (20). The -fold effect was determined by comparison to the empty expression vector cassette, and statistical analyses were performed using the Mann Whitney U test. The PPAR␥ ligand troglitazone was purchased from Calbiochem.
Retroviral Production and Infection-Retroviral production was described elsewhere (8). The coding region of the murine cyclin D1 cDNA (GenBank TM S78355) was inserted into the MSCV-IRESGFP vector at the EcoRI site upstream of the IRES driving expression of GFP. MSCV retroviruses were prepared by transient co-transfection with helper virus into 293T cells, using calcium phosphate precipitation. The retroviral supernatants were harvested 48 h after transfection (23) and filtered through a 0.45-m filter. Cyclin D1 ϩ/ϩ and cyclin D1 Ϫ/Ϫ MEFs were incubated with fresh retroviral supernatants in the presence of 4 g/ml Polybrene for 24 h, cultured for 6 days, and subjected to fluorescence-activated cell sorting (FACSVantage SE, BD Biosciences) for GFP-positive cells.
HDAC Assays-HDAC assays were performed using [ 3 H]acetateincorporated histones isolated from HeLa cells treated with sodium butyrate exactly as described previously (24). Hepatic extracts (600 g) were immunoprecipitated with saturating amounts of anti-HDAC1 antibodies (10 g, Santa Cruz) and then incubated with 1 ml of [ 3 H]acetate-labeled HeLa histones (10,000 dpm) for 2 h at 37°C, and acetylase activity was determined as described previously (24). Alternatively, HDAC1 assay was performed using the HDAC assay kit (Fluorometric Detection, Upstate, NY, catalog number 17-356) according to the manufacturer's instructions. Briefly, hepatic extracts (600 g dissolved in 600 l of cell lysate buffer) from cyclin D1 ϩ/ϩ , and cyclin D1 Ϫ/Ϫ mice were immunoprecipitated with saturating amounts of anti-HDAC1 antibodies (10 g) mixed with 30 l of protein A beads (50% slurry) at 4°C for 3-6 h. The immunoprecipitate and the protein A beads were collected by centrifugation (30 s in a microfuge at 8,000 ϫ g) and washed twice with 500 l of ice-cold Tris-buffered saline and once with 200 l of ice cold HDAC assay buffer (25 mM Tris, pH 8.0, 137 mM NaCl, 2.7 mM KCl, 1 mM MgCl 2 ). The beads were resuspended with 60 l of HDAC assay buffer containing 100 M HDAC assay substrate and then incubated at 30°C for 30 -60 min. After brief centrifugation, 40 l of the supernatant was transferred to 96-well plate and incubated with 20 l of the diluted activator solution at room temperature for 10 -15 min. The immunoprecipitated HDAC1 activity was measured in a fluorescence plate reader (excitation ϭ 350 nm, emission ϭ 460 nm) within 60 min.
Induction of Adipocyte Differentiation of MEFs-Induction of adipocyte differentiation of MEFs was performed as previously described (5). Primary MEFs were isolated from 14-day-postcoitus mouse embryos and were maintained at confluence for 1 day before being switched to basal differentiation medium (Dulbecco's modified Eagle's medium supplemented with 10% charcoal-stripped serum and 10 mg of insulin/ liter). Differentiation was induced by serum supplemented with 0.2 mM methylisobutylxanthine (Sigma), 5 M dexamethasone (Sigma), and 10 g/ml insulin (Sigma) for 3 days. Subsequently, cells were maintained in basal differentiation medium supplemented with troglitazone (5 M) (Calbiochem) or vehicle as indicated. Retroviral infection was conducted as described previously (5). For Oil Red-O staining, cells were fixed in 10% paraformaldehyde in phosphate-buffered saline for 15 min and rinsed briefly with water and ethanol. Cells were then stained with freshly prepared Oil Red-O solution (6 parts saturated Oil Red-O dye in isopropanol plus 4 parts water) at 37°C for 15 min, washed with 70% ethanol, and then rinsed with phosphate-buffered saline. Cells were inspected by microscopy.
Immunoprecipitation and Western Blot-293T cells were transfected with an expression vector for cyclin D1. Thirty hours after transfection, the cells lysates were prepared in 600 l of cell lysis buffer (10 mM HEPES, pH 7.5, 100 mM KCl, 0.4 mM EDTA, 10 mM sodium fluoride, 0.2% Nonidet P-40, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride with proteinase inhibitors (Roche Diagnostics GmbH, catalog number 1836145). 600 g of cellular lysate was subjected to immunoprecipitation with 10 l of anti-HDAC1 antibody (Upstate Biotechnology) and 30 l of protein A-agarose beads at 4°C overnight. Normal rabbit IgG was used as a negative control. The beads were washed with 800 l of cell lysis buffer five times, resuspended in 30 l of cell lysis buffer plus 6 l of SDS-PAGE loading buffer, and denatured by heating at 95°C for 5 min. Proteins were dissolved in 10% SDS-PAGE. The membrane was blotted with either anti-HDAC1 or anti-cyclin D1 antibody (Ab-3, Neomarker) at room temperature for 1 h, then washed three times with 0.05% Tween 20 phosphate-buffered saline. The membrane was then incubated with horseradish peroxidase-conjugated anti-rabbit antibody. The immunoreactive proteins were visualized by an enhanced FIG. 1. Cyclin D1 increases HDAC activity in vivo and blocks adipocyte differentiation. A, MEFs derived from either cyclin D1 ϩ/ϩ or cyclin D1 Ϫ/Ϫ were treated with either Me 2 SO (DMSO) or differentiation medium plus troglitazone (DM ϩ Trog) (5 M) for 10 days. Cells were then stained for Oil Red-O. Adipocytes differentiated from cyclin D1 Ϫ/Ϫ MEFs are also shown as an enlarged inset. B and C, HDAC1 activity was assayed using equal amounts of HDAC1 immunoprecipitated from either cyclin D1 ϩ/ϩ or cyclin D1 Ϫ/Ϫ MEFs treated with either control Me 2 SO or differentiation medium plus troglitazone. Nuclear extracts (NE) from HeLa cells served as a positive control and IgG was applied as a negative control for immunoprecipitation (IP).
chemiluminescence system (Amersham Biosciences). For immunoprecipitation of cyclin D1 with HDAC1-HDAC5, the expression vectors for cyclin D1 and HDAC1-HDAC5 were transfected into 293T cells. The cell lysates were then immunoprecipitated with the antibody (M2, Sigma), and Western blotting was conducted with an anti-cyclin D1 antibody (Ab-3).
Chromatin Immunoprecipitation (ChIP) assay-ChIP analysis was performed as previously described (25). 2 ϫ 10 7 3T3 cyclin D1 ϩ/ϩ or cyclin D1 Ϫ/Ϫ cells were grown in Dulbecco's modified Eagle's medium with 10% charcoal-dextran stripped serum for 3 days. Upon treatment, the cells were cross-linked by adding 1.0% formaldehyde buffer containing 100 mM sodium chloride, 1 mM EDTA-Na, pH 8.0, 0.5 mM EGTA-Na, Tris-HCl, pH 8.0, directly to culture medium for 10 min at 37°C. The medium was aspirated, and the cells washed twice using ice-cold phosphate-buffered saline containing 10 mM dithiothreitol and protease inhibitors. The cells were then lysed with 1% SDS lysis buffer and incubated for 10 min on ice. The cell lysates were sonicated to shear DNA to lengths between 200 and 500 bp, and the samples were diluted 10-fold in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris, pH 8.1, 167 mM NaCl). To reduce nonspecific background, the cell pellet suspension was precleared with 60 l of salmon sperm DNA/protein-A-agarose-50% slurry (Upstate Biotechnology) for 2 h at 4°C with agitation. Chromatin solutions were precipitated overnight at 4°C with rotation using 4 g of antibodies to either cyclin D1 (HD-11, Santa Cruz), PPAR␥ (H-100, Santa Cruz), FLAG (Sigma), acetyl histone H3, acetyl histone H3 (lysine 9), methyl histone H3 (lysine 4), HDAC1 (2E10) (Upstate Biotechnology), SUV39H1, or HP1␣ (Upstate Technology). For a negative control, rabbit or mouse IgG was incubated with the supernatant fraction for 1 h at 4°C with rotation. 60 l of salmon sperm DNA/protein A-agarose slurry was added for 2 h at 4°C with rotation to collect the antibody/histone complex and washed extensively following the manufacturer's protocol. Input and immunoprecipitated chromatin were incubated at 65°C overnight to reverse cross-linking. After proteinase K digestion for 1 h, DNA was extracted using a Qiagen spin column kit. Precipitated DNAs were analyzed by PCR of 30 cycles. The following oligonucleotides were used for PCR to identify the PPAR⌭ in the mouse LPL promoter 5Ј-AAAC-CCCTCCTCTCTGCCTC-3Ј and 5Ј-CCTCGGAGGAGGAGTAGGAG-3Ј or human LPL promoter, 5Ј-GGGCCCCCGGGTAGAGTGG-3Ј and 5Ј-CACGCCAAGGCTGCTTATGTGACT-3Ј. The oligonucleotides to identify the PPAR⌭ in the mouse adipocyte fatty acid binding protein (aP2) promoter are 5Ј-CAAGCCATGCGACAAAGGCA-3Ј and 5Ј-TA-GAAGTCGCTCAGGCCACA-3Ј (26).

Cyclin D1 Inhibits Adipocyte Differentiation of MEFs and
Regulates HDAC1 Activity-In our previous studies, cyclin D1 Ϫ/Ϫ MEFs exhibited enhanced differentiation in response to PPAR␥ ligands (5) indicating that cyclin D1 repression of PPAR␥ is an important physiological function of cyclin D1 in vivo. Several distinct HDAC inhibitors induce differentiation of MEFs into Oil Red-O-positive adipocytes (16,17). We therefore assessed the role of cyclin D1 in regulating cellular HDAC activity during adipocyte differentiation of MEFs. MEFs derived from either cyclin D1 wild type or cyclin D1 Ϫ/Ϫ mice were treated with differentiation medium and troglitazone, a PPAR␥ agonist, as described previously (5). The differentiation medium increased the abundance of Oil Red-O-positive lipid droplets in the cyclin D1 Ϫ/Ϫ MEFs compared with wild type cells (Fig. 1A). This result is consistent with previous observations demonstrating that there is enhanced induction of adipogenesis in cyclin D1 Ϫ/Ϫ MEFs with differentiation medium and PPAR␥ ligand (5). HDAC1 activity assays of these cellular extracts were conducted (Fig. 1C). HeLa cell nuclear extracts were used as a positive control (Fig. 1B). HDAC1 activity decreased during differentiation. Cyclin D1-deficient cells showed reduced HDAC activity, compared with cyclin D1 wild type cells (Fig. 1C, lane 1 versus 3), with a further reduction upon FIG. 2. Cyclin D1 augments HDAC activity. HDAC1 activity was assayed upon HDAC1 immunoprecipitation (IP) from murine liver lysates. A, Western blotting (WB) of HDAC1 demonstrating similar amounts of HDAC1 by Western blotting. IgG was applied as a negative control for immunoprecipitation. B and C, relative HDAC1 activity assayed by HDAC1 immunoprecipitation using hepatocellular extracts from cyclin D1 ϩ/ϩ or littermate cyclin D1 Ϫ/Ϫ mice and incubated with fluorometric-labeled HDAC substrates (B) or 3 H-labeled histone mix (C). Relative HDAC1 activity is shown. differentiation (Fig. 1C, lane 2 versus 4).
Reconstitution of Cyclin D1 into cyclin D1 Ϫ/Ϫ MEFs Increases HDAC Activity and Blocks Adipocyte Differentiation-As reduced cyclin D1 levels correlated with reduced HDAC1 enzyme activities, further experiments were conducted to determine whether reintroduction of cyclin D1 in turn was capable of increasing HDAC1 activity in cyclin D1 Ϫ/Ϫ MEFs. Cyclin D1 Ϫ/Ϫ MEFs were infected with either an expression vector for cyclin D1 (MSCV-cyclin D1-IRESGFP) or equal amounts of empty expression vector cassette (MSCV-IRESGFP). GFP-positive cells were selected through GFP fluorescence-activated cell sorting as described previously (8) (Fig. 3A, right panel). MEFs infected with the cyclin D1 expression vector showed increased relative HDAC1 activity compared with vector control (Fig.  3B). In addition, reintroduction of cyclin D1 into cyclin D1 Ϫ/Ϫ MEFs blocked adipocyte differentiation induced by differentiation medium and PPAR␥ ligand (Fig. 3A). Together, these studies suggest cyclin D1 contributes to cellular HDAC1 activity and regulation of HDAC activity by cyclin D1 plays a role in adipocyte differentiation.
Cyclin D1 Associates with HDACs in Vivo-We have previously shown that PPAR␥ transactivation induced by ligand was inhibited by cyclin D1 through a Rb-and Cdk-independent mechanism, requiring a region predicted to form a helix-loophelix structure (5). As HDAC1 is a known inhibitor for adipocyte differentiation, we hypothesized that cyclin D1 might confer repression of PPAR␥ through association with HDACs. Cyclin D1 may regulate HDAC activity either indirectly or through co-association with HDACs. To determine whether cyclin D1 co-associated with HDACs, immunoprecipitation/ Western blotting was conducted. A cyclin D1 expression plasmid was transfected into 293T cells. Immunoprecipitation with an HDAC1 antibody co-precipitated human cyclin D1 with endogenous HDAC1 (Fig. 5A, lane 2). In contrast, control IgG did not co-precipitate cyclin D1 or HDAC1 (Fig. 5A, lane 1). To examine whether cyclin D1 associated with HDAC1 in vivo, mouse liver lysate was subjected to immunoprecipitation with an anti-HDAC1 antibody. The immunoprecipitate was then resolved in a 7% SDS-PAGE and blotted with anti-cyclin D1 antibody. Immunoprecipitation with the HDAC1 antibody coprecipitated murine cyclin D1 from mouse liver lysate (Fig. 5B,  lane 2 versus 1), indicating that cyclin D1 and HDAC1 are associated in vivo.
Several trichostatin A-dependent histone deacetylases were next assessed to examine the possibility that cyclin D1 may co-precipitate with other members of the HDAC family. Immunoprecipitation and Western blotting were conducted with human embryonic kidney 293 cells transfected with expression vectors for either cyclin D1 or FLAG-tagged HDAC1-HDAC5 (Fig. 5C). Immunoprecipitation with anti-FLAG antibody for FLAG-HDAC1-HDAC5 and sequential Western blotting of cyclin D1 showed that HDAC1, -2, -3 and -5 co-precipitated cyclin D1. It was also noted that the relative abundance of HDAC1 and HDAC3 bound to cyclin D1 was greater than HDAC2 and HDAC5 (Fig. 5, D and E).
Cyclin D1 Enhances Recruitment of HDAC1 and HDAC3 to the PPARE of LPL Promoter-As cyclin D1 inhibited basal and ligand-induced activity of the synthetic PPARE, we examined the protein complexes recruited to the PPARE of the endogenous murine LPL promoter using ChIP assays. A comparison was made between cyclin D1 ϩ/ϩ and cyclin D1 Ϫ/Ϫ MEFs treated with differentiation medium and troglitazone (5 M) or equal amounts of control vehicle (Fig. 6A). The relative abundance of PPAR␥ at the PPARE was enhanced in the cyclin D1 Ϫ/Ϫ MEFs compared with cyclin D1 ϩ/ϩ MEFs (Fig. 6A, lanes  1 versus 2 and 3 versus 4). The relative abundance of HDAC1 at the PPARE in cyclin D1 ϩ/ϩ MEFs was increased 50% compared with the cyclin D1 Ϫ/Ϫ cells in the differentiated state (Fig. 6A,  lane 3 versus 4). HDAC3 recruitment at the PPARE was 4-fold greater in the cyclin D1 ϩ/ϩ compared with cyclin D1 Ϫ/Ϫ cells in the differentiated state (Fig. 6A, lane 3 versus 4), consistent with a model in which cyclin D1 enhances recruitment of HDAC1 and HDAC3 to the PPARE. To determine whether cyclin D1 was sufficient for the enhanced recruitment of HDAC1 to the PPARE, cyclin D1 was reintroduced into cyclin D1 Ϫ/Ϫ MEFs by transfecting with either the MSCV-cyclin D1-IRESGFP or MSCV-IRESGFP control virus. The cells were then treated with either differentiation medium and troglitazone or vehicle control. Reintroduction of cyclin D1, which is identified through the FLAG epitope, into cyclin D1 Ϫ/Ϫ cells reduced PPAR␥ recruitment in the presence of differentiation medium (Fig. 6B, lane 3 versus 4). The relative abundance of HDAC1 at the PPARE was increased in cells expressing cyclin D1 (Fig. 6B, lanes 1 versus 2, and 3 versus 4). Together these studies suggest that cyclin D1 enhances recruitment of HDAC1 and HDAC3 to a PPARE in vivo.
Cyclin D1 Deficiency Enhances Histone H3 Acetylation-To further confirm the effects of cyclin D1 abundance on the recruitment of HDAC to a PPARE, the relative amount of HDAC1 at a PPARE in the cyclin D1 Ϫ/Ϫ and cyclin D1 ϩ/ϩ cells was determined by ChIP assay of the murine LPL promoter (Fig. 7A) or aP2 promoter (Fig. 7B) in randomly cycling 3T3 cells. The relative abundance of HDAC1 at the PPARE of either the LPL (Fig. 7A, lane 1 versus 2) or the aP2 promoter (Fig. 7B,  lane 1 versus 2) was increased in cyclin D1 ϩ/ϩ cells compared with those of cyclin D1 Ϫ/Ϫ cells. PPAR␥ recruitment to the PPARE of the LPL (Fig. 7A, lane 1 versus 2) or aP2 promoter (Fig. 7B, lane 1 versus 2) was reduced in cyclin D1-expressing cells. These data are consistent with a model in which cyclin D1 enhances HDAC1 recruitment to PPAREs.
To examine the functional consequences of the association of cyclin D1 with HDACs, we assessed histone acetylation at the local chromatin structure of the PPARE of the LPL promoter. A comparison was made between cyclin D1 ϩ/ϩ and cyclin D1 Ϫ/Ϫ 3T3 cells. The cells were serum-starved for 48 h and then

FIG. 5. Cyclin D1 binds HDACs in vitro and in vivo. A, immunoprecipitation (IP) Western blot (WB) analysis of 293T cells transfected with
FLAG-tagged cyclin D1. Control IgG or antibody to endogenous HDAC1 were used for immunoprecipitation. Western blot analysis for HDAC1 and cyclin D1 is shown. B, HDAC1 immunoprecipitation Western blotting of murine hepatocellular extracts. IgG serves as a negative control for immunoprecipitation. Western blot analysis for HDAC1 and cyclin D1 is shown. C and D, relative binding affinity of cyclin D1 with HDAC1-HDAC5. C, Western blot analysis of 293T cells transfected with expression vectors for FLAG-tagged HDAC1-HDAC5 using antibodies to cyclin D1 (Ab-3) or anti-FLAG antibody. Guanine nucleotide dissociation inhibitor serves as a loading control for total protein. D, immunoprecipitation and Western blotting of the cell extracts for endogenous cyclin D1 and FLAG-tagged HDAC1-HDAC5. The cellular lysates were subjected to immunoprecipitation with anti-FLAG antibody for HDAC1-HDAC5, and Western blot was performed for endogenous cyclin D1. E, relative affinity of cyclin D1 with HDAC1-5. treated with 10% serum for 16 h. Endogenous cyclin D1 was detected at the PPARE by ChIP analysis in cyclin D1 ϩ/ϩ cells (Fig. 7C, lane 2). PPAR␥ levels were reduced and the abundance of both HDAC1 and HDAC3 were increased in the cyclin D1 ϩ/ϩ cells (Fig. 7C, lane 1 versus 2). Increased HDAC recruitment in cyclin D1 ϩ/ϩ cells was associated with decreased histone H3 (acetylated Lys-9) (Fig. 7C), consistent with the findings that histone H3-Lys-9 deacetylation associates with inactive promoters (27). In keeping with previous studies (28), a reduction in acetylated histone H3 lysine 9 was associated with increased H3 dimethyl lysine 9 and reduced H3 dimethyl Lys-4 (Fig. 7C). As the methylation of histone H3 lysine 9 provides binding sites for HP1 proteins (29), we examined the components of the methylase complex known to methylate histone H3 lysine 9. Compared with the cyclin D1 Ϫ/Ϫ cells, increased abundance of SUV39H1 and HP1␣ were identified in the cyclin D1 ϩ/ϩ cells (Fig. 7C), suggesting that cyclin D1 abundance regulated the recruitment of both histone deacetylases and histone methyltransferase to the local chromatin of the LPL promoter.
To determine whether a physiological change in cyclin D1 abundance was capable of altering HDAC recruitment to a PPARE, 3T3 cells were serum-starved for 48 h to reduce cyclin D1 levels in the cyclin D1 ϩ/ϩ 3T3 cells to barely detectable levels by Western blotting. Under these conditions, the recruitment of PPAR␥ was only modestly increased in the cyclin D1 ϩ/ϩ cells, and HDAC1 and HDAC3 recruitment was similar at the PPARE between cyclin D1 Ϫ/Ϫ and cyclin D1 ϩ/ϩ cells (Fig. 7D).
To determine whether the observed changes in HDAC recruitment to a PPARE was observed in epithelial cells, MCF7 cells transfected with an expression vector for FLAG-tagged PPAR␥ were treated with either trichostatin A or troglitazone, and the human LPL promoter was assessed with ChIP assays. Trichostatin A and troglitazone reduced endogenous HDAC3 recruitment to the PPARE of the human LPL promoter associated with increased acetylation of histone H3 and H4. Troglitazone enhanced PPAR␥ recruitment as evidenced by the anti-FLAG antibody chromatin immunoprecipitation of the PPARE (Fig. 7E). DISCUSSION Cyclin D1 has been increasingly linked to Cdk-independent transcriptional repression (30,31). As prior analysis of tissues and cells from cyclin D1 Ϫ/Ϫ mice demonstrated an important physiological role for cyclin D1 as an inhibitor of adipocyte differentiation through repression of PPAR␥ function (5), herein we investigated the mechanisms by which cyclin D1 repressed PPAR␥ function. HDAC inhibitors are also known to enhance the differentiation of MEFs into adipocytes (16,17). Thus, both cyclin D1 and histone deacetylases inhibit PPAR␥ activity and adipogenesis. Cyclin D1-deficient cells exhibited reduced HDAC1 activity, which was increased by the introduction of cyclin D1. Consistent with the finding that cyclin D1 bound HDACs, cyclin D1-deficient cells exhibited increased acetylation of histone H3 Lys-9, a marker of open chromatin configuration found at transcriptionally active promoters (32,33). Increased cyclin D1 expression, either through viral transduction of a cyclin D1 expression vector or through serum induction, increased HDAC recruitment to a PPARE commensurate with increased deacetylation of histone H3 (Lys-9). The recruitment of HDAC by cyclin D1 may contribute to the Cdkindependent repression of a subset of transcriptional factors including nuclear receptors such as AR and PPAR␥ by cyclin D1.
In the current studies cyclin D1 co-precipitated HDAC1 in vivo. Cyclin D1 preferentially bound HDAC1 and HDAC3. Several lines of evidence demonstrate that cyclin D1 association with HDACs is functionally relevant. Firstly, the target of HDAC function, acetylation of histone H3 lysine 9, was regulated by cyclin D1 abundance in the context of the local chromatin of an endogenous PPAR␥-responsive promoter. Secondly, cyclin D1 facilitated the recruitment of HDAC1 to a PPAR␥-responsive element in ChIP assays. The acetylation of histone H3 Lys-9, which marks the open chromatin configuration of transcriptionally active promoters, was increased in cyclin D1-deficient cells, consistent with the finding that cyclin D1 recruited HDACs. Thirdly, adipocyte differentiation of MEFs decreased HDAC1 activity, and the reintroduction of cyclin D1 into cyclin D1 Ϫ/Ϫ cells increased HDAC1 activity and inhibited adipocyte differentiation.
Previous studies have demonstrated that the recruitment of co-repressors, with associated histone deacetylase activity, are involved in the regulation of nuclear receptor function, either in the basal or ligand-activated state. The repression of ligandinduced receptor activation was previously reported for corepressor complexes coordinated by the SMRT/HDAC1-associated repressor protein (SHARP) (34), the transcription intermediary factor TIF1 (35), and the metastases-associated protein 1 co-repressor MTA (36). The binding of cyclin D1 to HDAC1 and the inhibition of PPAR␥-mediated differentiation by cyclin D1 is also consistent with previous studies on the role of HDAC1 in regulation of nuclear receptor function (30).
Adipogenesis involves two distinct phases, the first involving clonal expansion and a second phase involving cell cycle exit FIG. 6. Cyclin D1 recruits HDAC1 to the PPARE of the LPL promoter. A, ChIP assay analysis of the PPAR␥ response element of the murine LPL promoter of cyclin D1 ϩ/ϩ or cyclin D1 Ϫ/Ϫ MEF cells treated with differentiation medium (DM) plus either Me 2 SO or troglitazone (Trog). Immunoprecipitation was conducted with antibodies to PPAR␥, cyclin D1, HDAC1, or HDAC3, as indicated. B, ChIP analysis of the murine LPL promoter using cyclin D1 Ϫ/Ϫ MEFs, infected with either MSCV-IRESGFP control virus or with MSCV-Cyclin D1-IRES-GFP and treated with differentiation medium plus Me 2 SO or troglitazone. ChIP assays were conducted with the antibodies as indicated. and differentiation (37)(38)(39). Cyclin D1 expression inhibits cellular differentiation, the second phase of adipogenesis. This function is consistent with several previous studies in which cyclin D1 inhibited cellular differentiation of other cell types including myocytes (40 -42). Our findings are consistent with the previous studies of cell cycle proteins in adipogenesis. Rb inactivation by SV40 large T antigen inhibits adipogenesis (43), and Rb-deficient fibroblasts fail to differentiate into adipocytes when properly stimulated (44). In addition, the cyclin-dependent kinase inhibitors p18 and p21 are up-regulated during adipogenesis, consistent with their role in promoting cell cycle exit (45). Furthermore, studies suggest PPAR␥ up-regulates the Cdk inhibitors (45), which promote differentiation, consistent with our finding that cyclin D1 inhibits PPAR␥.
It is important to distinguish studies of adipocyte differentiation conducted with primary murine embryonic fibroblast (primary MEFs) from those conducted with clonal derivatives of MEFs or immortalized cell lines. It has been shown that 3T3L1 cells and CHO cells undergo differentiation in response to PPAR␥ ligands and differentiation medium. However, the differentiation medium that induced 3T3-L1 adipogenesis was not sufficient to induce differentiation in wild type MEFs (46), consistent with our findings that cyclin D1 wild type MEFs are relatively resistant to induction of adipocyte differentiation (5). When primary MEFs (Fig. 6, A and B) were treated with differentiation medium and PPAR␥ ligand, the relative amount of HDAC1 and HDAC3 recruited to the PPARE of the LPL promoter in cyclin D1-deficient cells was significantly reduced upon differentiation compared with those seen in cyclin D1 wild type cells (Fig. 6, A and B). Collectively, these results support the notion that the relative amount of HDAC recruited to the endogenous LPL promoter during adipocyte differentiation is regulated by cyclin D1.
Rb has been shown to recruit HDAC3 to PPAR␥ target genes and attenuate PPAR␥-mediated adipocyte differentiation. Disruption of the PPAR␥⅐Rb⅐HDAC3 complex by phosphorylation of Rb or inhibition of HDAC activity stimulates adipocyte differentiation (17). In contrast, other groups have suggested a FIG. 7. Cyclin D1 regulates acetylation of histone H3 at the LPL promoter region. ChIP assay of HDAC1 recruitment to the PPAR⌭ of the murine LPL promoter (A) or the aP2 promoter (B) in randomly cycling cyclin D1 ϩ/ϩ or cyclin D1 Ϫ/Ϫ 3T3 cells. IgG serves as an antibody negative control. C, cyclin D1 ϩ/ϩ or cyclin D1 Ϫ/Ϫ 3T3 cells were serumstarved for 48 h and subjected to 10% serum treatment for 16 h. ChIP analysis of the murine LPL promoter PPARE are shown using the indicated antibodies. D, ChIP analysis of the murine LPL promoter in serum-starved cyclin D1 ϩ/ϩ or cyclin D1 Ϫ/Ϫ 3T3 cells are shown. Antibodies used for immunoprecipitation are indicated. E, ChIP analysis of the human LPL promoter in MCF7 cells transfected with the FLAG-tagged PPAR␥ expression vector. Cells were treated with either vehicle (Veh), trichostatin A (TSA), or troglitazone as indicated in lanes 1, 2 and 3. DMSO, Me 2 SO. role of Rb in facilitating the differentiation of preadipocytes and MEFs into adipocytes (43,44,47). There are also controversial reports on the role of HDACs during adipogenesis (16 -18). Our results demonstrate a role of cyclin D1 in repression of the function of PPAR␥ through recruitment of HDAC1 and HDAC3. In our previous studies, PPAR␥ transactivation induced by the ligand BRL49653 was inhibited by cyclin D1 through a Rb-and Cdk-independent mechanism, requiring a region predicted to form a helix-loop-helix structure (5). It is likely that cyclin D1 plays a dual role by promoting cell proliferation and inhibiting cellular differentiation. On one hand, cyclin D1 is up-regulated by mitogenic signaling pathways, as seen in most cancer cells, resulting in phosphorylation of Rb and G 1 -S progression through association with and activation of Cdk4/6. On the other hand, cyclin D1 can regulate a subset of transcription factors, including nuclear receptors, through interacting with histone deacetylase activity independent of its Cdk-activation function (31). Further studies are necessary to elucidate the molecular mechanism by which cyclin D1 coordinates its Cdk-dependent and Cdk-independent functions.