RIP140-Targeted Repression of Gene Expression in Adipocytes

Animals.The generation of RIP140-null mice has previously been described (39). Mice used in this study were backcrossed six generations to a C57BL/6J background, maintained under standard conditions with controlled light and temperature, and fed a chow diet ad libitum. All experiments were performed according to United Kingdom Home Office guidelines.

Morphological and immunohistochemical analysis.Tissue sections prepared as described previously (17) were incubated with polyclonal goat anti-mouse primary antibody to Cidea (sc-8732; Santa Cruz Biotechnology), diluted to a ratio of 1:200 in phosphate-buffered saline (PBS), and detected using a Vectastain Elite ABC kit (Vector Laboratories). Sections were counterstained with hematoxylin.

Cell culturing.Primary WAT cultures were prepared from inguinal fat depots. The tissue was finely minced and digested with collagenase (500 μg/ml) and DNase (100 μg/ml), followed by centrifugation at 1,000 rpm for 5 min. The cells were passed through a 70-μm filter and cultured in Dulbecco's modified Eagle's medium (DMEM)-F12 medium supplemented with 10% fetal bovine serum (Invitrogen). Mouse embryonic fibroblasts (MEFs) from embryos at embryonic day 11.5 were isolated and cultured in DMEM-F12 medium supplemented with 10% fetal bovine serum (Invitrogen). The RIPKO-1 cell line was generated by continuous culturing of MEFs. Differentiation of the cells was performed as previously described (35) in the absence or presence of 2.5 μM rosiglitazone (Ro) as indicated below. Differentiated cells were visualized with oil red O staining.

Lentiviral transduction of RIPKO-1 cells.The PCR-amplified coding sequence for human RIP140 was cloned into pLenti6/V5-D-TOPO (Invitrogen). This vector was cotransfected into 293FT cells with ViraPower packaging mix to generate the lentivirus. RIPKO-1 cells were transduced with the lentivirus and stable cell lines (RIPKO-L) generated by selecting with blasticidin.

Transient transfection. Ucp1 promoter reporter constructs were generated by cloning a 4-kb fragment or 220-bp enhancer element (bp −2530 to −2310 relative to the transcriptional start site) of the 5′-flanking region of the murine Ucp1 gene into pGL3/basic vector (Promega). A 4-kb promoter construct lacking the enhancer region (ΔEnh; deletion of bp −2615 to −2228) was generated by PCR. RIPKO-1 cells were transfected in 24-well plates by using Fugene6 (Roche) with 1 μg of reporter gene, 250 ng of pRL-CMV, and/or 500 ng of pCI-RIP. Cells were harvested for luciferase assay 1, 3, and 5 days following the addition of the differentiation cocktail. Renilla luciferase activity was used to correct for differences in transfection efficiencies.

Expression analysis.RNA extraction from cell lines and tissue and cDNA preparation were performed as described previously (17). RIP140, L19, and UCP1 gene expression levels were determined using specific primers and TaqMan probes. Expression levels for all other genes were determined with SYBR green reagent by using specific primers. Expression levels for all genes were correlated to that for the L19 ribosomal coding gene. Primer sequences may be obtained on request.

Quantitation of mitochondrial DNA.DNA was extracted from untreated and differentiated primary MEFs using a DNeasy kit (QIAGEN). Real-time PCR with SYBR green reagent and specific primers was used to monitor levels of the mitochondrial cytochrome c oxidase subunit II gene and normalized to levels of the nucleus-encoded Ucp1 promoter.

Western blots.Sodium dodecyl sulfate-polyacrylamide gel electrophoresis, Western blotting, and immunodetection were performed as described previously (6) using rabbit polyclonal anti-mouse UCP1 (AB1426; Chemicon International) at a ratio of 1/1,000 and mouse monoclonal anti-β-actin (ab6276; Abcam Ltd., Cambridge, United Kingdom) at 1/5,000. Bands were visualized using secondary peroxidase-conjugated antibody and enhanced chemiluminescence.

Immunocytochemical staining.RIPKO-1 cells cultured on chamber slides were fixed in methanol and permeabilized in 0.2% Triton. UCP1 was detected using a rabbit polyclonal anti-mouse UCP1 antibody (AB3038; Chemicon International) diluted 1/100, a procedure followed by incubation with fluorescein isothiocyanate-conjugated swine anti-rabbit immunoglobulin G (IgG) (DAKO) diluted 1/50. Cell numbers were assessed by counting nuclei stained in the presence of a DAPI (4′,6′diamidino-2-phenylindole)-containing Vectashield mountant (Vector Labs).

Mitochondrial stain.Cells were incubated for 30 min in MitoTracker Green (Molecular Probes) at 100 nM, washed in PBS, and mounted in Vectashield medium (Vector Labs).

[³H]palmitate oxidation assay.Following differentiation, primary MEFs and RIPKO-1, RIPKO-L, and 3T3-L1 cells were assayed for ³H₂O production from excess [³H]palmitic acid as described previously (38).

Chromatin immunoprecipitation (ChIP) assay.Cells were incubated in the protein-protein cross-linking reagent dimethyl adipimidate · 2HCl (Pierce) at 10 mM in PBS for 30 min, followed by incubation in 1% formaldehyde in DMEM for 5 min at 37°C. The cross-linked cells were lysed, sonicated, and immunoprecipitated with protein A/G PLUS-agarose (Santa Cruz) according to the manufacturer's instructions using mouse anti-V5 antibody (Invitrogen) or control normal mouse IgG (Santa Cruz). DNA fragments were purified with a QIAquick PCR purification kit (QIAGEN) and used for templates in PCRs.

Affymetrix array hybridization and data analysis.Total RNA was isolated from three samples each of RIPKO-1 and RIPKO-L cells, both undifferentiated and after adipocyte differentiation (10 days) in the presence of the PPARγ ligand GW1929 (2.5 μm) (Sigma). Two RIP140-lentivirus-transduced cell lines were used for profiling in order to avoid variation due to differences in the sites of incorporation of the virus. Affymetrix array hybridization and scanning were performed by the CSC/IC Microarray Centre, Imperial College London, Hammersmith Campus, using murine 430 2.0 chips. Array data were analyzed with d-CHIP software (19). The microarray data are available at http://www.ebi.ac.uk/arrayexpress/ under accession number E-MIMR-42 .

Adipogenesis of wild-type and RIP140-null cells.To investigate the intrinsic role of RIP140 in adipocytes, primary cultures of WAT and MEFs were induced to differentiate with a standard hormone cocktail in the absence or presence of the PPARγ agonist rosiglitazone. In these primary cell systems, both wild-type and RIP140-null cells underwent adipogenesis, as judged morphologically by the accumulation of cytosolic fat droplets (data not shown). In addition to these changes, the induction of the adipocyte marker aP2 gene and the adipocyte transcription factor PPARγ2 was detected in all cell types following differentiation (Fig. 1A and B).

• Open in new tab
• Download powerpoint

FIG. 1.

Gene expression in wild-type and RIP140-null cells. Results from real-time PCR analysis of mRNAs for factors required for adipocyte function and energy dissipation are shown. (A and B) Pri-mary cultures of WAT and MEFs were untreated (Co) or differentiated in the absence (D) or presence (D+Ro) of rosiglitazone. Cells were harvested after 9 days (WAT) or 6 days (MEFs) of treatment. (C) RIPKO-1 cells preconfluent (PC) and differentiated in the presence of Ro up to day 10. Data are expressed relative to results for untreated RIP140-null cells.

The absence of RIP140 in mice results in altered expression in adipose tissue of key genes involved in energy metabolism, including Ucp1 and the carnitine palmitoyltransferase 1b gene (Cpt1b); however, the signaling mechanisms in vivo are complex, involving intrinsic and systemic factors. We therefore analyzed the expression levels of these genes in the cell culture systems. Ucp1 and Cpt1b were more highly expressed in primary cultures of WAT and MEFs devoid of RIP140, with expression levels being dependent on differentiation (Fig. 1A and B).

We generated a MEF line lacking the RIP140 gene (termed RIPKO-1) by continuous culturing of the primary MEFs. This cell line retained the potential to differentiate into adipocytes, as judged morphologically by cytosolic fat droplet accumulation (data not shown). Consistent with these changes, aP2 and PPARγ were induced in the cells, with maximal expression being reached by day 6 of treatment (Fig. 1C, upper panel). Ucp1 and Cpt1b displayed progressive increases in expression in the RIPKO-1 cells (Fig. 1C, lower panel). The expression of Ucp1 in BAT is normally associated with the transcription factor PGC1α. However, expression levels of PGC1α or of the related gene PGC1β were similar in wild-type and RIP140-null primary WAT adipocytes (data not shown).

Identification of RIP140 target genes following reexpression of RIP140 in RIPKO-1 cells.To verify an essential role for RIP140 in gene regulation, it was reexpressed in the RIPKO-1 cells using a lentiviral vector. Adipocyte differentiation of stably expressing cell lines (termed RIPKO-L cells) was unaltered compared to that of the parental null cells, as determined by oil red O staining (Fig. 2A). Furthermore, both aP2 and PPARγ2 gene expression levels were progressively increased during differentiation, as observed in both RIPKO-1 and 3T3-L1 cells (data not shown). The expression levels of RIP140 in the RIPKO-L and 3T3-L1 cells were similar at days 5 and 14 following differentiation (Fig. 2B). However, the patterns of expression differed between the cell lines; expression increased in 3T3-L1 cells following differentiation, whereas in RIPKO-L cells, expression was reduced. This difference is due to the reintroduced RIP140 being under the control of the cytomegalovirus promoter rather than that of its endogenous regulatory mechanisms.

• Open in new tab
• Download powerpoint

FIG.2.

Identification of RIP140 target genes following reexpression of RIP140 in RIPKO-1 adipocytes. (A) Oil red O staining of undifferentiated and differentiated RIPKO-1 and RIPKO-L cells. (B) Real-time PCR analysis of RIP140 expression in RIPKO-1, RIPKO-L, and 3T3-L1 cells harvested untreated (day 0) or differentiated in the presence of rosiglitazone (days 5 and 14). Data are expressed relative to results for untreated 3T3-L1 cells. (C) Cluster analysis of gene expression in undifferentiated and differentiated RIPKO-1 and RIPKO-L cells. Adipogenesis-related genes are listed in the upper panel. The middle panel lists genes that are upregulated by at least twofold in RIPKO-1 cells upon differentiation yet repressed fivefold upon RIP140 reexpression. Genes associated with BAT are listed in the lower panel. S. pombe, Schizosaccharomyces pombe.

To identify genes that, in adipocytes, are targeted for regulation by RIP140, we performed expression profiling using Affymetrix microarrays. Gene expression in RIPKO-1 cells was compared to that in cells in which RIP140 had been reintroduced with a lentiviral vector. The transcription factors C/EBPα, PPARγ, and SREBP1, which are essential for normal adipocyte differentiation and function, were induced following differentiation in both RIPKO-1 and RIPKO-L cells (Fig. 2C, upper panel). Genes that are associated with adipocyte function, such as those for SCD1, aP2, and adipsin, were also induced.

The microarray data were analyzed to identify genes that are induced in mature adipocytes and may be targeted for repression by RIP140. We selected genes that are upregulated by at least twofold in RIPKO-1 cells upon differentiation yet are repressed fivefold upon RIP140 reexpression (Fig. 2C, middle panel). The Ucp1 gene was identified within this gene cluster, showing that the reexpression of RIP140 in RIPKO-1 cells ablates the adipocyte-dependent expression of this gene. This analysis identified a number of additional genes repressed by the reexpression of RIP140 (Fig. 2C, middle panel). These include genes implicated in adipocyte function, namely, the Cidea (cell-death-inducing DFF45-like effector A) (41), Aquaporin 7 (AQPap) (13), β3-adrenergic receptor (2), carboxylesterase 3 (34), acetyl coenzyme A synthetase 2 (33), and membrane metalloendopeptidase (31) genes. In addition, a number of other genes are similarly repressed by RIP140 expression (Fig. 2C, middle panel). In contrast, the expression levels of key regulators and markers of BAT function (37) are unaltered or only moderately affected (Fig. 2C, lower panel). For example, PGC1α and FOXC2 were expressed at lower levels in RIP140-null cells.

Cidea expression is altered in RIP140^−/− cells in vitro and in vivo.We next investigated the regulation of the Cidea gene, identified by the microarray study as highly expressed only in differentiated RIPKO-1 cells, which has been implicated in thermogenesis through the direct modulation of Ucp1 activity (41). Cidea mRNA expression was undetectable in RIPKO-1, RIPKO-L, and 3T3-L1 adipocytes prior to differentiation and, following adipogenesis, increased only in RIP140-null cells (Fig. 3A, left panel); the expression profile closely resembled that of Ucp1 (Fig. 4A).

• Open in new tab
• Download powerpoint

FIG. 3.

RIP140 represses transcription of the Cidea gene. Results from real-time PCR analysis of Cidea expression are shown. (A) Cell lines, as indicated, were harvested untreated (day 0), or differentiated in the presence of rosiglitazone (days 5 and 14). Primary MEFs that were untreated (Co) or differentiated in the absence (D) or presence of rosiglitazone (D+Ro) were harvested after 6 days of treatment. (B) Expression in WAT from wild-type (n = 4) and RIP140-null (n = 4) mice. Real-time PCR data are expressed relative to maximal expression. Detection of Cidea in WAT from wild-type and RIP140-null mice by immunohistochemistry is shown (right panel).

• Open in new tab
• Download powerpoint

FIG. 4.

RIP140 targets the Ucp1 promoter to repress transcription. (A) Real-time PCR analysis of Ucp1 expression in cell lines that were either untreated (day 0) or differentiated in the presence of rosiglitazone (days 5 and 14), as indicated. Data are expressed relative to results for untreated RIPKO-1 cells. (B) Immunoblot for UCP1 and β-actin (loading control) in RIPKO-1 and RIPKO-L cells untreated and differentiated in the presence of rosiglitazone (D+Ro) (day 14) (left panel), and immunocytochemical staining for UCP1 (right panel) in untreated cells (day 0) and cells differentiated in the presence of Ro (days 7, 9, and 14). (C) Transient transfection of RIPKO-1 cells with reporter constructs pGL3-Basic, Ucp1(4kb)/luc, or Ucp1(4kbΔEnh)/luc in the absence and presence of RIP140, differentiated for 5 days in the presence of rosiglitazone. (D) Transient transfection of RIPKO-1 cells with full-length 4-kb or 220-bp enhancer element Ucp1 promoter-luciferase reporter genes in the absence and presence of RIP140. Cells were harvested 1, 3, and 5 days following control or differentiation-plus Ro treatment (D+Ro). (E) Chromatin immunoprecipitation of the Ucp1 enhancer or upstream control region in RIPKO-1 and RIPKO-L16 cells with control (IgG) or anti-V5 antibody.

To confirm these changes, we determined the expression patterns of Cidea in primary adipocyte culture systems and in vivo. Cidea was undetectable in undifferentiated primary MEFs and induced only in the RIP140-null cells (Fig. 3A, right panel). A similar pattern was also observed in primary WAT cultures (data not shown). Quantitative analysis of Cidea expression showed a 15-fold increase in WAT in RIP140-null mice relative to wild-type mice (Fig. 3B). Immunohistochemical staining revealed that the expression of Cidea in unilocular adipocytes in WAT of RIP140^−/− mice (Fig. 3B) localized predominantly in the cytoplasmic area near the plasma membrane, similarly to Ucp1.

RIP140 targets the Ucp1 promoter to repress transcription.To investigate further the mechanism of action of RIP140, we analyzed its ability to repress transcription from the Ucp1 promoter. Real-time PCR showed that Ucp1 expression was reduced in RIPKO-1 cells upon exogenous RIP140 expression to generate the RIPKO-L cells (Fig. 4A). Western blot analysis showed an induction of a 32-kDa immunoreactive band corresponding to UCP1 in fully differentiated RIP140-null cells (Fig. 4B). Immunocytochemical staining of RIPKO-1 cells showed a progressive increase in UCP1 as differentiation proceeded (Fig. 4B). This staining verified that UCP1 was confined to the lipid-containing differentiated cells and localized to the cytoplasm in accordance with its role as a protein confined to the mitochondrial compartment.

In mouse brown adipocytes, the transcription from the Ucp1 gene promoter is under the control of a region up to 4 kb upstream of the transcriptional start site (4, 16). Subsequent studies with transgenic mice have identified a 220-bp enhancer from kb −2.5 to −2.3 that controls tissue-specific expression (32). RIPKO-1 cells were transiently transfected with a luciferase reporter gene under the control of either the wild-type 4-kb Ucp1 promoter or a construct with the 220-bp enhancer region deleted (4kbΔEnh). Reporter activity was reduced by 50% with the deletion of the enhancer element (Fig. 4C). Exogenous expression of RIP140 inhibited luciferase activity from both the wild-type and 4kbΔEnh constructs. Following the demonstration that the enhancer region was necessary for the full induction of Ucp1 promoter activity, RIPKO-1 cells were transiently transfected with a luciferase reporter gene under the control of either the 4-kb Ucp1 promoter or the 220-bp enhancer element linked to the thymidine kinase promoter from −105 to +50. There was a progressive increase in promoter activity as the cells differentiated. Coexpression of RIP140 inhibited the Ucp1 promoter activity from both the full-length promoter and the 220-bp enhancer element, thus identifying this specific region as a target for RIP140 suppression of Ucp1 gene transcription (Fig. 4D).

Direct association of RIP140 with the Ucp1 enhancer element was determined using ChIP assays. The expression of RIP140-V5 in the RIPKO-L cells does not exceed that of the endogenous RIP140 in differentiated 3T3-L1 adipocytes (Fig. 2B). An antibody specific for V5 precipitated the Ucp1 enhancer element in differentiated RIPKO-L cells but not a control region 15 kb upstream of the mouse Ucp1 gene (Fig. 4E). Enrichment of the Ucp1 enhancer element was not observed with the V5 antibody in RIPKO-1 cells devoid of RIP140. Thus, RIP140 is recruited to the Ucp1 enhancer element in differentiated RIPKO-L cells.

Increased β-oxidation in RIP140-null adipocytes.Following the identification of the Ucp1 gene as a target for RIP140, we investigated the functional consequences of increased Ucp1 expression. The induction of factors that facilitate fatty acid oxidation and uncouple respiration leads to an increased capacity for β-oxidation. Measurement of the rate of ³H₂O production derived from a [³H]palmitic acid substrate showed that RIP140-null adipocytes derived from primary MEFs exhibited a total fatty acid oxidation 1.5-fold higher than that of wild-type adipocytes (Fig. 5A). The cells had differentiated to similar extents, as judged by the expression of aP2; however, Ucp1 levels were elevated in the RIP140-null adipocytes (Fig. 5A). Measurement of the ratio of mitochondrial DNA to nuclear DNA in primary MEFs showed that mitochondrial biogenesis occurred in both wild-type and RIP140-null cells following differentiation (Fig. 5A). However, the absence of RIP140 did not significantly affect the level of mitochondrial DNA in the differentiated cells. The specific uptake of a MitoTracker mitochondrion-selective probe revealed that both RIPKO-1 and RIPKO-L differentiated cells contained more mitochondria than undifferentiated cells did (Fig. 5B).

• Open in new tab
• Download powerpoint

FIG. 5.

Mitochondrial biogenesis and β-oxidation in RIP140-null adipocytes. (A) Wild-type (WT) and RIP140^−/− (KO) primary MEFs differentiated for 9 days were assayed for β-oxidation of [³H]palmitic acid (left panel; data are expressed as cpm/cell relative to results for the wild type), and expression levels of aP2 and Ucp1 were assayed by real-time PCR (center panel) and by measurement of the ratio of mitochondrial DNA to nuclear DNA (right panel). (B) MitoTracker staining of RIPKO-1 and RIPKO-L cells. (C) β-Oxidation of [³H]palmitic acid in differentiated cells (right panel) and aP2 expression in untreated cells (Co) and in cells differentiated in the presence of Ro (D+Ro) (left panel).

The rate of ³H₂O production derived from [³H]palmitic acid was determined in RIPKO-1, RIPKO-L, and 3T3-L1 cells that were differentiated to similar extents, as judged by aP2 expression (Fig. 5C). RIP140-null adipocytes exhibited total fatty acid oxidation levels twofold higher than those of wild-type 3T3-L1 adipocytes, and this level was repressed by the reexpression of RIP140 (Fig. 5C). The potential to prevent the induction of the expression of genes involved in energy dissipation demonstrates a primary role for RIP140 in this process and provides a mechanism for regulating β-oxidation within adipose cells.