Nuclear receptor cofactor receptor interacting protein 140 controls hepatic triglyceride metabolism during wasting in mice


  • Mauricio Berriel Diaz,

    1. Emmy Noether and Marie Curie Research Group, Molecular Metabolic Control, German Cancer Research Center Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Anja Krones-Herzig,

    1. Emmy Noether and Marie Curie Research Group, Molecular Metabolic Control, German Cancer Research Center Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Dagmar Metzger,

    1. Emmy Noether and Marie Curie Research Group, Molecular Metabolic Control, German Cancer Research Center Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Anja Ziegler,

    1. Emmy Noether and Marie Curie Research Group, Molecular Metabolic Control, German Cancer Research Center Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Alexandros Vegiopoulos,

    1. Emmy Noether and Marie Curie Research Group, Molecular Metabolic Control, German Cancer Research Center Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Martin Klingenspor,

    1. Technical University Munich, Department of Molecular Nutritional Medicine, Freising-Weihenstephan, Germany
    Search for more papers by this author
  • Karin Müller-Decker,

    1. Core Facility Tumor Models, German Cancer Research Center Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Stephan Herzig

    Corresponding author
    1. Emmy Noether and Marie Curie Research Group, Molecular Metabolic Control, German Cancer Research Center Heidelberg, Heidelberg, Germany
    • Molecular Metabolic Control, German Cancer Research Center, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany
    Search for more papers by this author
    • fax: (49) 6221-423595.

  • Potential conflict of interest: Nothing to report.


In mammals, triglycerides (TG) represent the most concentrated form of energy. Aberrant TG storage and availability are intimately linked to the negative energy balance under severe clinical conditions, such as starvation, sepsis, or cancer cachexia. Despite its crucial role for energy homeostasis, molecular key determinants of TG metabolism remain enigmatic. Here we show that the expression of nuclear receptor cofactor receptor interacting protein (RIP) 140 was induced in livers of starved, septic, and tumor-bearing mice. Liver-specific knockdown of RIP140 led to increased hepatic TG release and alleviated hepatic steatosis in tumor-bearing, cachectic animals. Indeed, hepatic RIP140 was found to control the expression of lipid-metabolizing genes in liver. Conclusion: By preventing the mobilization of hepatic TG stores, the induction of RIP140 in liver provides a molecular rationale for hepatic steatosis in starvation, sepsis, or cancer cachexia. Inhibition of hepatic RIP140 transcriptional activity might, thereby, provide an attractive adjunct scheme in the treatment of these conditions. (HEPATOLOGY 2008.)

Excessive uptake and storage of energy substrates is a prerequisite for severe metabolic disorders, such as obesity, type II diabetes, and other components of the so-called metabolic syndrome.1

Conversely, catabolic conditions such as starvation, sepsis, and cancer cachexia are characterized by a negative energy balance, leading to a gradual loss of peripheral organ function and eventually to the death of affected subjects.2

Triglycerides (TGs) are the most efficient energy carriers in mammals. Given the crucial role of TGs as major energy supply for biological tissues, storage as well as mobilization of TG is tightly controlled by insulin and counterregulatory hormonal pathways.1, 3 Under pathological conditions of excessive adipose tissue lipolysis, the liver acts as a major sink for adipose tissue–derived free fatty acids (FFA).4 Aberrant accumulation of TGs in the liver (fatty liver) consequently represents a major hallmark of starvation, sepsis, and tumor-induced adipose tissue wasting and may contribute to liver failure.2, 5, 6 In contrast to the increase in hepatic TG levels, circulating serum TG levels were found to decline on starvation and cancer cachexia in mouse models.7, 8 Despite their critical importance for overall body function under energy-deprived conditions, molecular components involved in the control of inter-compartment/tissue TG distribution remain largely unknown.

Numerous studies within the past years have uncovered the critical role of nuclear receptor cofactor complexes in the control of hormone-dependent metabolism, thereby representing attractive molecular target sites for intervention strategies in metabolic diseases.9 In this respect, nuclear receptor coactivator peroxisome proliferator-activated receptor coactivator (PGC)-1α has been identified by us and others as a key regulator of gluconeogenic gene expression in the liver. PGC-1α gene expression is induced by fasting hormonal cues,10, 11 thereby representing a prototype for inducible cofactor activity in the control of energy homeostasis.

By investigating the expression of various transcriptional cofactors during catabolic metabolism, this study discovered an unforeseen regulatory role of receptor interacting protein (RIP) 140 for cachectic TG metabolism.


Apo, apolipoprotein; FAS, fatty acid synthase; FFA, free fatty acid; LPL, lipoprotein lipase; LXR, liver X receptor; mRNA, messenger RNA; PCR, polymerase chain reaction; PGC, proliferator-activated receptor coactivator; RIP140, receptor interacting protein 140; RNAi, RNA interference; RXR, retinoid X receptor; SEM, standard error of the mean; shRNA, short hairpin RNA; SRC, steroid hormone receptor coactivator; SREBP, sterol regulatory element binding protein; TG, triglyceride; VLDL, very-low-density lipoprotein.

Materials and Methods

Recombinant Adenoviruses.

Adenoviruses expressing RIP140-specific or nonspecific control short hairpin RNAs (shRNAs) were produced using the BLOCK-iT Adenoviral RNA interference (RNAi) expression system (Invitrogen, Karlsruhe, Germany) according to the manufacturer's instructions and purified by CsCl gradients.12

Animal Experiments.

Male 8-week-old to 12-week-old wild-type C57Bl/6J mice were obtained from Charles River Laboratories (Brussels, Belgium) and maintained on a 12-hour light–dark cycle with regular unrestricted diet. For starvation experiments, animals were fasted for 8, 24, or 48 hours with free access to water or fasted and refed for the following 6 or 24 hours. Sepsis was induced by intraperitoneal injection of 20 mg lipopolysaccharide/kg body weight (Sigma, Munich, Germany). For virus injections, 1 × 109 plaque-forming units per recombinant virus were administered via tail vein injection. In each experiment, 7 to 10 animals received identical treatments. Mice were sacrificed 7 to 9 days after adenovirus injection in the fasted (24 hours) or ad libitum fed [Colon 26 (C26) cachexia experiments] state. For tumor induction in cachexia models, 5 × 106 Lewis lung carcinoma cells or 1.5 × 106 C26 cells in phosphate-buffered saline were injected subcutaneously into 10-week-old C57BL/6J or CD2F1 mice (Charles River Laboratories, Brussels, Belgium), respectively. Control mice were injected with phosphate-buffered saline. In high-fat diet experiments, C57Bl6 mice were either fed a standard chow diet (10% energy from fat, Research diets D12450B, New Brunswick, NJ) or a high-fat diet (45% energy from fat, Research diets D12451) for a period of 16 weeks (samples provided by S. Kersten, Wageningen). Very-low-density lipoprotein (VLDL) production was determined after tyloxapol (SIGMA, Munich, Germany) injection as described.7 All animal procedures have been approved by local authorities and are in accordance with National Institutes of Health guidelines.

Blood Metabolites.

Serum levels of glucose, TG, cholesterol, ketone bodies, and FFA were determined by using an automatic glucose monitor (One Touch, Lifescan) or commercial kits, respectively (Sigma, Munich, Germany; RANDOX, Crumlin, Northern Ireland; WAKO, Neuss, Germany, respectively). Insulin levels were determined using a mouse insulin enzyme-linked immunosorbent assay (Mercodia, Uppsala, Sweden).

Tissue Lipid Extraction.

Hepatic lipids were extracted as described previously,13 and TG and total cholesterol contents were determined using commercial kits as described. Values were calculated as micromoles (triglycerides) or mg (cholesterol) per gram wet tissue.

Quantitative TaqMan Reverse Transcription Polymerase Chain Reaction.

Total RNA was extracted from homogenized mouse liver using the Qiazol reagent (Qiagen, Hilden, Germany) kit. Complementary was prepared by reverse transcription using Oligo dT primer (Fermentas, St. Leon-Rot, Germany). Complementary DNAs were amplified using assay-on-demand kits and an ABI-Prism 7700 Sequence detector (Applied Biosystems, Darmstadt, Germany). RNA expression data were normalized to levels of TATA-box binding protein RNA.

Protein Analysis.

Protein was extracted from frozen liver samples, loaded onto a 10% sodium dodecyl sulfate polyacrylamide gel and blotted onto nitrocellulose membrane. Western blot assays were performed as described10 using antibodies specific for RIP14014 or beta-actin (Abcam, Cambridge, UK).

Lipoprotein Lipase Activity.

Lipoprotein lipase (LPL) activity measurements were performed as described,15 using frozen adipose tissue and skeletal muscle samples.


Liver tissue was embedded in Tissue Tek optimal cutting temperature compound (Sakura, Torrance, CA). Five-micrometer cryosections were stained with Oil Red O as described.16

RNA Interference.

Oligonucleotides targeting mouse RIP140 (5′-GCTGCAGAAACTAAAGCTTCT-3′) were annealed and cloned into pENTRY RNAi vector (Invitrogen, Karlsruhe, Germany). Nonspecific oligonucleotides (5′-GATCTGATCGACACTGTAATG-3′) with no significant homology to any mammalian gene sequence were used as nonsilencing controls in all experiments.


Promoter luciferase constructs containing the fatty acid synthase (FAS), the sterol regulatory element binding protein (SREBP)-1c 5′-flanking regions,17 or an isolated liver X receptor (LXR) element,18 and expression plasmids for RIP140,19 LXR, retinoid X receptor (RXR),20 SREBP-1c,21 and dominant-negative LXR22 have been described previously. The murine apolipoprotein (Apo) A5 promoter region from −1028 base pairs (bp) to +94 bp was cloned into the pGL3 luciferase reporter vector (Promega, Mannheim, Germany) by standard polymerase chain reaction (PCR) methods.

Cell Culture and Transient Transfection Assays.

HepG2, Hepa1c1, and human embryonic kidney (HEK) 293 cells were transfected by the calcium phosphate precipitation method. Cell extracts were prepared 48 hours after transfection, and luciferase assays were performed as described,10 normalizing to activity from cotransfected beta-galactosidase expression plasmid. Primary mouse hepatocytes were isolated and cultured as described.23 Twenty-four hours after plating, the cells were infected with control or RIP140-specific shRNA adenoviruses at a multiplicity of infection of 100. Twenty-four hours after virus infection, cells were exposed to 22-hydroxycholesterol (10 μM) or the T0901317 compound (10 μM) for 24 hours and harvested for messenger RNA (mRNA) analysis thereafter.

Statistical Methods.

Results are presented as means ± standard error of the mean (SEM). Significance was assessed by two-tailed t test for two-group comparisons or two-way analysis of variance with Bonferroni post tests at the P = 0.05 level.


The identification of nuclear receptor cofactor PGC-1α as an essential component of the molecular fasting response in liver10, 11 prompted us to investigate the expression patterns of other transcriptional cofactors under food-deprived conditions. Consistent with previous studies,10, 11 mRNA levels of PGC-1α were induced by threefold in livers of starved wild-type C57Bl6 mice as compared with refed controls (Fig. 1A) by quantitative PCR. In contrast, expression levels of distinct p160 cofactors, steroid hormone receptor coactivator (SRC)-1 and SRC-2,9 remained unchanged under these conditions (Fig. 1A). Intriguingly, hepatic mRNA expression of nuclear receptor cofactor RIP1 140 was found to be elevated more than six-fold on 48 hours' starvation (Fig. 1A). This induction of RIP140 expression was already detectable after 8 and 24 hours of fasting (Supplementary Fig. S1), although to a lesser extent, indicating that RIP140 mRNA levels gradually adapt to prolonged starvation periods. RIP140 has been initially defined as a critical factor for female fertility and oxidative metabolism in skeletal muscle and adipose tissue,24, 25 acting as a corepressor of various nuclear receptors on metabolic target genes.26 To test whether the hepatic induction of RIP140 represents a more common feature of diverse catabolic conditions, we employed wild-type mice treated with bacterial lipopolysaccharide as a standard model for sepsis-induced tissue wasting and anorexia.27 Indeed, RIP140 mRNA levels were found to be induced by threefold to fourfold in livers of lipopolysaccharide-treated mice as compared with control littermates (Fig. 1B). In contrast, mRNA levels of PGC-1α, SRC-1, and SRC-2 remained unchanged (Fig. 1B), demonstrating the specificity of these effects.

Figure 1.

Hepatic RIP140 expression is determined by the systemic energy status. (A,B,C) Quantitative PCR analysis of peroxisome proliferator-activated receptor coactivator (PGC)-1α, steroid hormone receptor coactivator (SRC)-1, SRC-2, and receptor interacting protein (RIP) 140 mRNA levels in livers of (A) re-fed or 48-hour starved mice (n = 4). (B) Lipopolysaccharide (LPS)-injected or phosphate-buffered saline (PBS)-injected mice sacrificed 48 hours after LPS injection (n = 3), or (C) after 16 weeks of low-fat (LF) or high-fat (HF) diet feeding (n = 6) (means ± SEM). *P ≤ 0.05.

To next explore RIP140 expression under opposed conditions of energy excess, we performed expression analysis in a mouse model for diet-induced obesity. Intriguingly, in contrast to the marked up-regulation of RIP140 under catabolic conditions, RIP140 mRNA expression was significantly inhibited in mice kept on a high-fat diet for 16 weeks when compared with chow-fed littermates (Fig. 1C). Expression levels for PGC-1α, SRC-1, and SRC-2 were not changed in response to high-fat feeding (Fig. 1C). These results indicated that the expression of RIP140 in liver is modulated by the systemic energy status and that the hepatic induction of RIP140 represents a common feature of catabolic metabolism as exemplified by starvation or sepsis.

Based on studies on whole-body knockout mice, RIP140 has recently been implicated in the regulation of hepatic glucose and lipid metabolism.14 However, no tissue-specific loss-of-function model for RIP140 has been available, preventing the exact determination of the metabolic role of RIP140 on a tissue-by-tissue basis. Therefore, we specifically disrupted the activity of RIP140 in livers of wild-type C57Bl6 mice by delivering an adenovirus expressing a RIP140-specific shRNA construct via tail vein injection. Short hairpin RNA treatment efficiently down-regulated hepatic RIP140 protein as well as mRNA levels as compared with control shRNA-injected littermates (Fig. 2A and Supplementary Fig. S2). No effect of adenovirus injection on RIP140 mRNA expression was observed in skeletal muscle and adipose tissue (Supplementary Fig. S3), verifying the previously established liver specificity of the experimental system.10

Figure 2.

Hepatic RIP140 controls triglyceride metabolism. (A) Western blot of liver extracts from four representative control (Ctrl-RNAi) or RIP140 shRNA-injected (RIP140 RNAi) wild-type mice using RIP140 or beta-actin antibodies. Mice were analyzed after 24 hours of fasting. (B) Serum triglyceride (TG) levels in control or RIP140 shRNA-injected wild-type mice (means ± SEM, n = 5). (C) Liver triglyceride (TG) levels in same mice as in B (means ± SEM, n = 5). (D) Serum total ketone body levels in same mice as in B (means ± SEM, n = 5). (E) Hepatic very-low-density lipoprotein (VLDL) release in control or RIP140 shRNA-injected wild-type mice. Time after tyloxapol injection indicated (means ± SEM, n = 6). (F) White adipose tissue (WAT) lipoprotein lipase (LPL) activity in control or RIP140 shRNA-injected wild-type mice (means ± SEM, n = 7). *P ≤ 0.05.

Remarkably, mice infected with RIP140 shRNA adenovirus showed a more than twofold elevation of serum TG levels during starvation when compared with control animals (Fig. 2B), whereas hepatic TG and cholesterol content were significantly reduced on RIP140 depletion (Fig. 2C and Supplementary Fig. S4). In contrast, serum cholesterol, FFA, insulin, and glucose levels remained unchanged (Supplementary Figs. S5-S8), suggesting that hepatic RIP140 specifically affects serum and liver TG metabolism.

The substantial shift from liver to serum TG on hepatic depletion of RIP140 prompted us to explore the mechanisms behind this effect. Hepatic lipid stores are determined by the relative balance of lipid uptake and release, de novo TG formation (lipogenesis), and FFA combustion via the beta-oxidation pathway.5 As an output measure for the latter pathway, we determined circulating levels of ketone bodies in control and RIP140-depleted animals. Consistent with lower hepatic TG content, mice deficient in hepatic RIP140 showed markedly elevated levels of ketone bodies in the serum, indicating that hepatic FFA oxidation was increased in these animals (Fig. 2D). Conversely, the liver contributes to circulating TG levels through the production and release of VLDL complexes, carrying TG as a major lipid component.3 Tail vein injection of tyloxapol inhibits systemic TG clearance through the inhibition of systemic LPL activity, allowing the determination of hepatic VLDL release.7 As shown in Fig. 2E, liver-specific RIP140 deficiency triggered a significant increase in hepatic VLDL production as compared with controls (Fig. 2E), suggesting that hepatic knockdown of RIP140—at least in part—promotes elevation of serum TG levels through this metabolic pathway. Apart from hepatic VLDL production, serum TG levels are critically determined by the rate of LPL-mediated systemic clearance.28 To this end, hepatic RIP140 knockdown led to a significant reduction of LPL activity in adipose tissue and tended to also decrease LPL activity in skeletal muscle (Fig. 2F and Supplemental Fig. S9), suggesting that RIP140 in liver controls TG clearance in peripheral tissues. Taken together, these data indicated that loss of hepatic RIP140 determines hepatic and serum TG levels by a combinatorial mechanism comprising increased FFA oxidation, enhanced VLDL secretion, and reduced peripheral LPL activity.

To explore the molecular basis for the observed TG phenotype in liver-specific RIP140 knockdown animals, we performed expression analysis of key regulatory metabolic enzyme genes in RIP140 and control shRNA-treated animals. Consistent with the view that de novo TG synthesis—at least to a certain degree—feeds hepatic VLDL production,3 mRNA levels for key enzymes involved in TG biosynthesis, such as FAS, adenosine triphosphate citrate lyase, and glucokinase, were induced in livers of starved RIP140-deficient mice compared with controls (Fig. 3A). The genetic program of hepatic lipogenesis is predominantly controlled by the activity of insulin-responsive SREBP-1c.29 Indeed, RIP140 deficiency strongly up-regulated hepatic SREBP-1c levels (Fig. 3A), indicating that RIP140 normally contributes to the previously described suppression of SREBP-1c expression under starving conditions.29 To validate this finding in a cell autonomous manner, we infected primary mouse hepatocytes with RIP140-specific or control shRNA adenoviruses and exposed these cells to different ligands for nuclear receptor LXR, a potent inducer of SREBP-1c gene expression.30 Treatment of primary hepatocytes with LXR ligands T0901317 or 22-hydroxycholesterol induced SREBP-1c mRNA levels significantly (Fig. 3B and Supplementary Fig. S10). Consistent with the in vivo animal data, knockdown of RIP140 gene expression further increased SREBP-1c levels under both basal as well as stimulated conditions (Fig. 3B and Supplementary Fig. S10), supporting the notion that RIP140 serves as a repressor of SREBP-1c gene expression. Indeed, RIP140-deficient primary hepatocytes also displayed elevated levels of the SREBP-1c target gene FAS (Supplementary Fig. S11). In transient transfections of HepG2 hepatocytes, RIP140 potently inhibited SREBP-1c promoter activity driven by LXR ligand activation (Fig. 3C). Deletion of the LXR binding site within the SREBP-1c promoter (Fig. 3C, −148 bp), or cotransfection of a dominant-negative LXR mutant (data not shown) completely abolished the effect of RIP140 on reporter gene activity. To additionally test the impact of RIP140 on LXR-mediated transcription in an independent system, we cotransfected a reporter gene carrying an isolated LXR response element with expression plasmids for LXR, its heterodimerization partner RXR, and RIP140 into HEK cells. Similar to its effects on the SREBP-1c promoter, RIP140 overexpression completely blunted the stimulatory effects of ligand-activated LXR:RXR heterodimers on promoter activity (Supplementary Fig. S12), further strengthening the conclusion that RIP140 primarily acts as a corepressor for LXR. Consistent with the notion that RIP140 regulates lipogenic genes at the level of SREBP-1c promoter inhibition, SREBP-1c–mediated transactivation of the FAS promoter was unaffected by RIP140 overexpression (Supplementary Fig. S13), indicating that RIP140 does not act as a direct SREBP-1c cofactor under these conditions.

Figure 3.

RIP140 controls expression of key lipid-metabolizing genes. (A) Quantitative PCR analysis of fatty acid synthase (FAS), adenosine triphosphate citrate lyase (ACL4), glucokinase (GK), and sterol regulatory element binding protein (SREBP)-1c mRNA levels in livers of control or RIP140 shRNA-injected (RIP140 RNAi) wild-type mice (means ± SEM, n = 5). (B) Quantitative PCR analysis of SREBP-1c mRNA levels in primary mouse hepatocytes infected with control or RIP140-specific shRNA adenoviruses. Cells were treated with T0901317 (10 μM) or dimethylsulfoxide vehicle for 24 hours (means ± SEM, n = 4). (C) Transient transfection assay of HepG2 hepatocytes transfected with −408 or −148 bp SREBP-1c reporters. Cotransfection of mRIP140 and 22-hydroxycholesterol (22-OH) treatment (10 μM) indicated (means ± SEM, n = 9). (D) Quantitative PCR analysis of CD36, apolipoprotein (Apo) A5, acetyl-Coenzyme A acyltransferase 2 (ACAA2), acetyl-Coenzyme A carboxylase 2 (ACC2), and carnitine palmitoyl transferase 1α (CPT1α) mRNA levels in livers of control or RIP140 shRNA-injected wild-type mice (means ± SEM, n = 5). (E) Transient transfection assay of Hepa1c1 hepatocytes transfected with −1028 bp ApoA5 promoter construct. Cotransfection of control or RIP140 shRNA vectors indicated (means ± SEM, n = 9). *P ≤ 0.05.

In contrast to lipogenic genes, expression of fatty acid uptake transporter CD36 was inhibited in RIP140 shRNA-injected animals (Fig. 3D), suggesting reduced fatty acid flux into the liver. Consistent with the regulation of systemic LPL activity, RIP140 shRNA treatment also diminished the expression of ApoA5, a liver-derived activator of LPL activity31 (Fig. 3D), and simultaneously induced the LPL inhibitor, Apo C3, in mice (data not shown). In hepatocytes, RIP140 knockdown also significantly inhibited ApoA5 promoter activity in transient promoter assays (Fig. 3E), whereas overexpression of RIP140 induced ApoA5 transcription by two-fold (data not shown). Interestingly, mRNA levels of other prominent regulators of LPL, the angiopoietin-like proteins 3 and 4, as well as ApoC2 remained unchanged on RIP140 knockdown (Fig. 3D). Finally, reflecting the increased ketone body levels, mRNA expression of fatty acid oxidation gene acetyl-coenzyme A acyltransferase 232 was found to be elevated in RIP140-deficient livers as compared with controls, whereas expression of the mitochondrial enzyme acetyl-coenzyme A carboxylase 2, representing a critical negative regulator of fatty acid oxidation,33 was decreased by twofold on RIP140 knockdown. In contrast, carnitine palmitoyltransferase-1α remained unchanged (Fig. 3D).

Given the regulatory function of RIP140 for hepatic TG metabolism, we finally sought to characterize the impact of the observed induction of RIP140 on catabolic metabolism. To this end, we employed the C26 mouse model for tumor-induced cachexia and anorexia.34 Subcutaneous implantation of C26 cells into wild-type mice promoted severe loss of body weight, skeletal muscle, as well as adipose tissue mass (Fig. 4A). In addition, C26 mice displayed reduced levels of serum TG (Fig. S14) and substantial hepatic steatosis (Fig. 4A), a common hallmark of starvation, sepsis, and anorexia nervosa.7, 35 Remarkably, hepatic gene expression analysis revealed a down-regulation of the lipogenic pathway (FAS, adenosine triphosphate citrate lyase, SREBP-1c) along with an induction of lipid transporter CD36. Also, A carboxylase 2 mRNA levels substantially decreased in tumor-bearing animals, indicating enhanced fatty acid oxidation in livers of cachectic mice (Fig. 4B). Correlating with the regulation of a subset of RIP140 target genes in C26 mice (Fig. 4B) and consistent with the induction of hepatic RIP140 expression during starvation and sepsis (Fig. 1), cancer cachexia triggered the up-regulation of RIP140 gene expression in livers of C26 mice as compared with controls (Fig. 4C). Indeed, RIP140 mRNA levels also tended to be increased in an independent Lewis-Lung-Carcinoma model for cancer cachexia,36 (Fig. 4C), further arguing for the induction of hepatic RIP140 expression as a common feature of catabolic metabolism.

Figure 4.

RIP140 triggers hepatic steatosis during cancer cachexia. (A) Body weight, epididymal fat (epid. fat) pad mass, gastrocnemius (GC) muscle mass, and hepatic TG content in control (Ctrl) and Colon26 cachectic mice (means ± SEM, n = 5). (B) Quantitative PCR analysis of lipid-metabolizing gene mRNA levels in livers of same mice as in panel A (means ± SEM, n = 5). (C) Quantitative PCR analysis of RIP140 mRNA levels in livers of control and Colon26 or Lewis-Lung-Carcinoma (LLC) cachectic mice (means ± SEM, n = 4-5). (D) Representative pictures of Oil Red O stained liver sections from control (Ctrl) or Colon26 (C26) cachectic mice injected with control (Ctrl-RNAi) or RIP140 shRNA (RIP140 RNAi) adenovirus. (E) Quantitative PCR analysis of SREBP-1c, ApoA5, and CD36 mRNA levels in livers of same mice as in panel D (means ± SEM, n = 8). *P ≤ 0.05.

Remarkably, liver-specific disruption of RIP140 expression by adenoviral shRNA transfer in C26 cachectic animals alleviated the C26 fatty liver phenotype (Fig. 4D) and increased serum TG as well as ketone body levels in C26 mice (Supplementary Figs. S15 and S16). Consistently, RIP140 deficiency in livers of C26 animals induced the expression of SREBP-1c by threefold and inhibited CD36 and ApoA5 mRNA levels (Fig. 4E), suggesting that the induction of RIP140 in cachectic livers indeed represents a functional checkpoint for aberrant TG metabolism under these conditions.


Detailed molecular knowledge about the pathogenesis of clinically severe disorders, such as starvation, sepsis, or cancer cachexia, is still rare. Despite the clear negative correlation between the degree of wasting and the prognosis of affected patients, effective anti-cachectic treatment strategies are not available.2 Here, we identify an unexpected function of the nuclear receptor corepressor RIP140 in the control of hepatic triglyceride metabolism under catabolic conditions.

Whole-body RIP140 knockout animals are lean, display resistance against diet-induced obesity, and maintain insulin sensitivity, suggesting that RIP140 globally promotes fat accumulation and insulin resistance.37 Indeed, our results are consistent with the assumption that RIP140 action in the liver contributes to hepatic TG storage by triggering lipid transporter (CD36) expression, inhibiting hepatic FFA oxidation and VLDL release. The decrease in hepatic lipid levels in our liver-specific knockdown models is consistent with the physiological data from genetically engineered whole-body knockout mice that are also characterized by lower intrahepatic lipid stores.37

However, based on the treatment of the whole-body knockout animals with the LXR ligand T0901317, RIP140 had been suggested to serve as a coactivator for this nuclear receptor in the liver, leading to decreased expression of LXR-activated lipogenic genes, particularly SREBP-1c, on RIP140 knockout.14 In contrast, our studies involving RIP140 knockdown experiments in animals and primary hepatocytes as well as functional SREBP-1c and LXR element promoter studies argue for an LXR corepressor function of RIP140 in the liver. The latter findings are consistent with other reports, demonstrating a default RIP140 corepressor function for LXR38–40 as well as with the correlation between elevated RIP140 levels and concomitant inhibition of SREBP-1c expression during cancer cachexia (Fig. 4). The major difference between the genetic and the shRNA knockdown models is represented by the chronic versus the acute loss of RIP140 function. Therefore, it is conceivable that the long-term, whole-body genetic knockout of RIP140 already during embryonic development might trigger intrahepatic compensatory molecular mechanisms, leading to differential effects of RIP140 deficiency on LXR activity and lipogenic gene expression in the liver.

The inhibition of CD36 expression upon RIP140 depletion indicates that RIP140 action is not restricted to trans-repression and can differ within a single cell type in a gene-specific manner. This assumption is supported by genome-wide expression profiling studies, showing a roughly equal number of genes inhibited or activated by RIP140 in adipocytes.41 Indeed, the distinct composition of RIP140-associated transcriptional complexes within specific promoter contexts can switch RIP140 from a default corepressor into a coactivator, as found by Miyata et al.40 and our own unpublished observations). The identity of factors coactivated by RIP140 on the CD36 promoter remains elusive. However, given the importance of peroxisome proliferator-activated receptor nuclear receptors for CD36 gene regulation,42 it is tempting to speculate that RIP140 directs lipid uptake via coactivation of these nuclear receptor family members, while simultaneously serving as a corepressor for distinct LXR nuclear receptors in the control of lipogenesis.

In addition to the observed regulatory function of RIP140 for intrahepatic lipid stores, our studies also uncover an unprecedented role of this nuclear receptor cofactor for circulating TG levels. Whereas the reduction of hepatic lipids is also seen in whole-body RIP140 knockout mice,37 these animals lack the marked elevation of serum TG levels as observed in our liver-specific knockdown model (Fig. 2). Likely, enhanced FFA dissipation in adipose tissue and skeletal muscle in the whole-body knockout24, 25 masks these aspects of RIP140's liver-specific functions. To this end, the observed regulation of ApoA5 as long-distance mediator of LPL enzyme activity31 through RIP140 might—at least in part—explain the observed effects of hepatic RIP140 on systemic TG availability. Intriguingly, SREBP-1c has been identified as a repressor for ApoA5 gene expression,43 thereby providing a potential explanation for the inhibitory effect of RIP140 deficiency on ApoA5 promoter activity. The regulation of serum TG levels by apolipoprotein/LPL-dependent mechanisms might be additionally complemented by the stimulatory impact of RIP140 deficiency on lipogenic gene expression and the subsequent feeding of newly synthesized FA into the VLDL secretion pathway.3 However, because studies on the contribution of de novo lipogenesis to hepatic VLDL release yielded ambiguous results,43–48 the relative importance of this pathway for the observed phenotype is difficult to estimate. In any case, the decrease in hepatic TG stores despite elevated lipogenic gene expression is likely explained by the concomitant induction of FFA oxidation, thereby overriding any additional storage of newly synthesized TG in RIP140-deficient livers.

Interestingly, the inhibitory impact of aberrant induction of RIP140 on VLDL secretion and FA oxidation pathways under cancer cachexia conditions can synergistically aggravate the steatotic phenotype in the tumor-bearing state. This study establishes the hepatic induction of RIP140 as a common hallmark of wasting metabolism as observed during starvation, sepsis, and cancer cachexia. In this respect, we hypothesize that the seemingly inconsistent correlation between inhibition of RIP140 expression and elevated TG levels under high-fat diet conditions reflects a counter-regulatory circuit to antagonize further lipid accumulation under these conditions. RIP140 expression is controlled by distinct promoter regions, acting in a partly autoregulatory,49 tissue-specific, and context-specific manner.50 It will be interesting to determine the signaling pathways controlling the induction of hepatic RIP140 expression under catabolic conditions as opposed to situations of energy excess, for example, high-fat diet, in the future. In this regard, the development of compounds that modulate RIP140 function in the liver may provide useful adjunct anti-steatotic therapy for food-deprived, septic, or tumor-bearing patients.


The authors thank S. Stöhr and A. Pohl-Arnold for technical assistance, S. Bohl and U. Klingmüller for primary hepatocyte isolation, and M. Downes (La Jolla, CA), S. Kersten (Wageningen, NL), T. Osborne (Irvine, CA), M. G. Parker (London, UK), and L. N. Wei (Minneapolis, MN) for providing reagents.