The Transcriptional Corepressor RIP140 Regulates Oxidative Metabolism in Skeletal Muscle

Summary Nuclear receptor signaling plays an important role in energy metabolism. In this study we demonstrate that the nuclear receptor corepressor RIP140 is a key regulator of metabolism in skeletal muscle. RIP140 is expressed in a fiber type-specific manner, and manipulation of its levels in null, heterozygous, and transgenic mice demonstrate that low levels promote while increased expression suppresses the formation of oxidative fibers. Expression profiling reveals global changes in the expression of genes implicated in both myofiber phenotype and metabolic functions. Genes involved in fatty-acid oxidation, oxidative phosphorylation, and mitochondrial biogenesis are upregulated in the absence of RIP140. Analysis of cultured myofibers demonstrates that the changes in expression are intrinsic to muscle cells and that nuclear receptor-regulated genes are direct targets for repression by RIP140. Therefore RIP140 is an important signaling factor in the regulation of skeletal muscle function and physiology.


Affymetrix microarray hybridisation and data analysis.
Total RNA was isolated from the gastrocnemius of wild type or null animals fed either control diet (10% kcal) or high fat diet (45% kcal) for 3 months (n=5,6). 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 (Li and Wong, 2001). P-values were generated in d-CHIP by a 2 tailed unpaired student's t-test to use as a ranking and filtering device. The p-values were not adjusted for multiple testing.
Cell culture of H2K cells. Conditionally immortal myogenic clonal cell lines were derived from the edl of H2K x RIP140+/+ and -/-mice as previously described (Morgan et al., 1994). Cell proliferation and differentiation of these cells are temperature and interferon (IFN)-γ-dependent. Myoblasts were maintained in the presence of heat-inactivated fetal calf serum (20%), chick embryo extract (2%), gelatin (0.01%) and IFN-γ at 33˚C. In order to differentiate into myotubes cells were plated densely, switched to media without IFN-γ and low serum (5% horse serum) and the temperature was raised to 37˚C. This resulted in sarcomere formation within 3-4 days.

Ligand Treatment.
Wild type and null cells were differentiated into myotubes for 4 days and the medium was changed to 2% horse serum supplemented with the agonists for PPARδ (GW501516, 1µM), RXR (9-cis retinoic acid (100 nM Sigma)) or the vehicle (0.1% DMSO and 0.01% EtOH) as control. Cells were harvested after 24 hrs. With ERRα ligand cells were differentiated for 3 days and the medium was changed to 2% horse serum supplemented with XCT790 (10µM, a gift from Tim Willson and William J. Zuercher, GlaxoSmithKline) or the vehicle (0.1% DMSO) as control. Cells were harvested after 48 hrs.

Chromatin immunoprecipitation assay
Fully differentiated cells (day 5) were incubated in Dimethyl adipimidate 2 (DMA) in media (final concentration 10mM) for 30 mins at room temperature to cross-link protein-protein interactions. The cells were put in 1% formaldehyde in PBS for 15 mins at 37 ºC to cross link protein and DNA. Cross-linked cells were lysed, sonicated and immunoprecipitated with protein A/G PLUS-agarose (Santa Cruz, SC-2003) according to the manufacturers instructions using rabbit-polyclonal anti-mouse RIP140 (a gift from Dr D. Chen) or control normal rabbit IgG (Santa Cruz SC-2027). DNA fragments were purified with a QIAquick PCR purification kit (QIAGEN) and used as templates for PCR. Primer sequences are available on request.

Figure S1. A. SDH staining in wt, het and null gastrocnemius sections
Histochemical staining for SDH in muscle sections at lower and higher magnification (scale bars=100µm at low magnification and 40µm at high magnification) Figure S2. A. Semi-quantitative evaluation of mitochondria in wt, heterozygous and null muscle Scanning Electron Microscopy was carried out on muscle sections taken from the EDL of wt, het and null mice. A scoring system was devised to semi-quantify the number of mitochondria per skeletal muscle fibre with 1 being the least (pairs of mitochondria uniformly distributed near the z line) to 4 (skeletal muscle fibres with long strings of enlarged mitochondria found throughout the field of view). The scores are an average of 20 different skeletal muscle fibres scored blind by 4 individuals.

B. Quantification of mitochondrial copy number
The EDL was dissected from adult, age-matched wt and RIP140 null mice and DNA was extracted using a DNAeasy 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. Values are shown as normalised to levels of the nucleus-encoded UCP1 promoter (n=3).

Figure S3. Analysis of muscle fibre types by metachromatic ATPase and myosin heavy chain immunostaining.
Serial sections of soleus muscle from wild type and RIP140 null mice were examined by metachromatic ATPase and myosin heavy chain immunostaining to determine the numbers of fibres corresponding to different types of fibres. The areas marked are for myofibre identification purposes only. Immunostaining for type I and type IIA fibres is shown with a line drawn through the type I fibres to facilitate the distinction with type IIA fibres. Figure S4. Analysis of fibre type composition and SDH staining in TA muscle after exercise. TA muscle from wild type and RIP140Tg mice either untrained or allowed free access to voluntary exercise wheels for 6 weeks was exanibed for MyHC expression and SDH activity. The histograms show the proportional change in fibre type as determined by detection of IIA, IIX and IIB stained fibres using specific antibodies Figure S5. Light microscopy of proliferating wild type and RIP140 null myoblasts and differentiated myotubes. Proliferating myoblasts were grown at low density in 20% fetal bovine serum and induced to differentiate by culturing in 5% horse serum for five days.

Table S1. Relative proportions of muscle fibre types in the soleus and extensor digitorum longus (EDL) muscles of wild-type and RIP140-null mice
Serial sections were analysed by metachromatic ATPase and myosin heavy chain immunostaining as shown in Supplementary Fig 1 to quantitate type I, IIA and IIB fibres while the proportion of type IIX fibres was deduced from the difference between total type II and type IIA plus IIB fibres and confirmed using the specific IIX antibody 6H1. Values are the means + SEM from 3 or 4 independent analyses as indicated. Changes in the proportion of type I fibres in the soleus and type II fibres in the EDL were significant at P < 0.05.

Table S2. Muscle-specific genes significantly altered by absence of RIP140
Affymetrix microarray analysis of mRNA levels in wt vs null gastrocnemius muscle in adult mice fed a normal chow diet (n=6). Affymetrix microarray and real-time PCR analysis of mRNA levels in wt vs null gastrocnemius muscle in adult mice fed a normal chow diet (n=6). PPAR δ target genes as referenced in (Dressel et al. 2003;Tanaka et al. 2003;Hummasti and Tontonoz. 2006).

Table S5. ERRα target genes upregulated by the absence of RIP140
Affymetrix microarray analysis of mRNA levels in wt vs null gastrocnemius muscle in adult mice fed a normal chow diet (n=6). ERRα target genes as identified and referenced in (Luo et al. 2003;Huss et al. 2004;Herzog et al. 2006)

Table S6. Expression profile of early and late myogenic markers
Affymetrix microarray analysis of mRNA levels in wt vs null gastrocnemius muscle in adult mice fed a normal chow diet (n=6).